Methods for cell-type specific profiling to identify drug target

Abstract

To provide methods of profiling gene and protein expression of a plurality of nuclei from a single cell type.SOLUTION: Provided is a method of profiling nuclear epigenetic expression and / or nuclear structure, the method comprising: (a) processing a tissue sample containing a plurality of cell types to expose nuclei, the tissue sample comprising brain tissue, brain stem tissue, and / or spinal cord tissue; (b) contacting the nuclei with two or more affinity labels that bind to different nucleic acid transcripts or proteins that are specific to the nuclei of a cell type in the tissue sample to label the nuclei, the nucleic acid transcripts or proteins being encoded by genes selected from a specific group; (c) purifying the nuclei labeled with two or more affinity labels; and (d) (i) collecting information regarding histone modifications, binding of transcription factors, and DNA modifications for the labeled nuclei to profile the epigenetic expression of the nuclei, and / or (ii) collecting information on the chromosomal structure of the labeled nuclei to profile the nuclear structure.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 62 / 457,420, filed February 10, 2017, and No. 62 / 558,396, filed September 14, 2017, each of which is incorporated by reference in its entirety.

[0002] Sequence Listing This application contains a "long" sequence listing, which has been submitted on a CD-R in lieu of a paper printout and is incorporated herein by reference in its entirety. The CD-R contains the following sequence listing text file as part of the newly filed subject matter: Filename: 2111_1000PCT_SL.txt; File Size: 123,473,920 bytes; Creation Date: February 9, 2018. These CD-Rs are labeled "CRF", "Copy 1" and "Copy 2", respectively, and each contains only one identical file as described in the immediately preceding sentence. The machine-readable format of each CD-R is IBM-PC format, and the operating system of each compact disc is Microsoft-Windows.

[0003] FIELD OF THEINVENTION The present invention relates to gene and protein expression profiling of multiple nuclei from a single cell type to identify drug targets. [Background technology]

[0004] The mammalian central nervous system (CNS) is a complex organ that contains hundreds of different interwoven cell types. The ability to genetically target (Gong, et al. (2003) Nature 425, 917-925) and molecularly profile (Heiman, et al. (2008) Cell 135, 738-748) specific cell types is beginning to allow insight into the essential features of the mammalian brain discovered in the foundational work of Ramon y Cajal over a century ago (Ramon y Cajal, S., et al. (1899). Texture of the Nervous System of Man and the Vertebrates (Wein New York: Springer)). Each anatomically distinct, classically defined cell type expresses a set of characteristic genes (Dougherty,et al.(2010)Nucleic Acids Res 38,4218-4230 and Doyle,et al.(2008)Cell 135,749-762) that confer essential properties for specialized cellular functions (Kim,I.-J.,et al.(2008)Nature 452,478-482;Nakajima,M.,et al.(2014)Cell 159,295-305). Expression of these genes depends on the maintenance of cell-specific epigenetic states that orchestrate nuclear functions (Kriaucionis,et al.(2009)Science 324,929-930;Mellen,et al.(2012)Cell 151,1417-1430). Furthermore, application of this profiling technique in mouse models has led to the recognition that environmental influences (Heiman, et al. (2008) Cell 135, 738-748; Shrestha, et al. (2015) eLife 4, 289), internal physiological cues (Schmidt, et al. (2012) Cell 149, 1152-1163), and diseases causing genetic disorders (Fyffe, et al. (2008) Neuron 59, 947-958; Ingram, et al. (2016) Neuron 89, 1194-1207) alter gene expression in affected cell types.Despite the pace of progress in experimental systems, fundamental questions about the complexity of the human brain remain unanswered: for example, it is not clear how many distinct cell types exist in the human brain, how these cell types vary between individuals or species, whether the process of brain ageing is comparable across cell types, and why mutations in ubiquitously expressed genes can have devastating consequences in one or a few select cell types.

[0005] Although a wealth of information exists regarding the circuits affected in psychiatric and neurodegenerative diseases, the ability to specifically modulate these circuits requires knowledge of targets that are found exclusively in the neurons of the circuit.

[0006] A method called translating ribosome affinity purification (TRAP) was previously developed to allow identification of all genes expressed in a cell type of interest. TRAP profiling studies have led to the development of several drugs that are in clinical trials. Although TRAP technology has enabled several drug developments, a major limitation of TRAP is that target identification is first performed using transgenic mice. Although mice and humans are similar, there are known differences in the CNS circuitry of the two species, and there may be differences in gene expression between orthologous cell types as well. Furthermore, it is not clear how many gene expressions in specific cells vary between individual subjects and whether these differences affect disease development or response to treatment. In TRAP, a BAC vector is used to generate a transgenic mouse line. See Doyle et al., Cell, 135:749-762 (2008) and Heiman et al., Cell, 135:738-748 (2008). This document is incorporated herein by reference in its entirety, and hence the method is referred to as bacTRAP.

[0007] In this method, EGFP-L10a ribosomal fusion protein is targeted to the cell type of interest, e.g., Purkinje cells. Affinity purification is performed on cell-specific polysomal RNA using anti-EGFP coated magnetic beads as shown in Heiman et al. Purified samples may be analyzed using microarrays. BACs are often selected from the Gene Expression Neural System Atlas (GENSAT) project (www.gensat.org), developed by Dr. Nathaniel Heintz at Rockefeller University. In some cases, the transgene may be expressed in more than one cell type. For example, immunofluorescence in Doyle et al. showed that EGFP-L10a is expressed in both cerebellar molecular layer Pvalb positive and NeuN negative interneurons, evidence that both astrocytes and basket cells are targeted. mRNA is immunoprecipitated and genome-wide transcriptional profiling is performed on the mRNA samples. These results identify cell type specific markers compared to profiling of non-immunoprecipitated samples.

[0008] To develop more efficient and specific drugs, candidate targets need to be identified directly from humans and human tissues, rather than from transgenic mice, and the extent to which these targets vary between individuals needs to be determined. Thus, there is a long-felt need to develop generalizable methods that allow the molecular study of certain cell types in wild-type animals and humans. Furthermore, non-genetic techniques are needed to complement the discovery of specific features of neural cell types and circuits in model systems and to gain insight into the potential unique properties of the human nervous system. Summary of the Invention

[0009] Various embodiments of the invention herein provide methods for profiling gene expression of nuclei, the methods comprising: processing a tissue sample comprising a plurality of cell types to expose nuclei; labeling at least one nucleus from the processed tissue sample by contacting the processed tissue sample with an affinity label that binds to a nucleic acid transcript or protein encoded by a gene selected from Tables 19-24 that is specific to a nucleus of at least one cell type contained in the tissue sample; purifying the at least one labeled nucleus; and collecting nucleic acid transcript and / or protein expression data for the labeled nucleus, thereby profiling gene expression of the nucleus.

[0010] In some embodiments, the cell type originates from cerebellar tissue and is selected from the group consisting of granule cells, Purkinje cells, glia, Bergmann glia, dopaminergic neurons, basket / astrocytes, astrocytes, brainstem motor neurons, oligodendrocytes, upper motor neurons, lower motor neurons, and deep cerebellar nuclei. In some embodiments, the protein is a transcription factor localized in the nucleus. In some embodiments, the cell type originates from basal ganglia tissue and is selected from the group consisting of striatonigral medium spiny neurons (MSNs), striatopallidal MSNs, striatal cholinergic interneurons, subthalamic nucleus, dopaminergic neurons, and bed nucleus of the stria terminalis (BNST) neurons. In some embodiments, the cell type originates from thalamic tissue and is selected from the group consisting of thalamocortical neurons, thalamostriatal neurons, and thalamic reticular nucleus neurons. In some embodiments, the cell type originates from cortical tissue and is selected from the group consisting of corticostriatal neurons, entorhinal cortex layer 2 / 3 neurons, fast-firing cortical interneurons, and layer 2 / 3 pyramidal cells from prefrontal cortical tissue. The cell type is a cholinergic projection neuron from medial habenula tissue of the pineal gland. In some embodiments, the cell type originates from hippocampal tissue and is selected from the group consisting of cornu ammonis area 1 (CA1), cornu ammonis area 2 (CA2), cornu ammonis area 3 (CA3), and dentate gyrus cells. In some embodiments, the cell type originates from at least one tissue selected from the group consisting of brain tissue, brainstem tissue, and spinal cord tissue.

[0011] Various embodiments of the invention described herein provide a method of isolating granule cells, the method comprising: processing a tissue sample containing multiple cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with at least one affinity label that binds to a nucleic acid transcript or protein specific to the nuclei of granule cells, encoded by one of the genes Itpr, NeuN, Cdh15, Calb2, Rbfox3, Neurod1, and Reln; and purifying the at least one labeled nucleus. In some embodiments, the at least one labeled nucleus does not contain a nucleic acid transcript encoded by Olig2.

[0012] Various embodiments of the invention described herein provide methods of isolating Purkinje cells, the method comprising processing a tissue sample containing a plurality of cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with an affinity label that binds to a nucleic acid transcript or protein encoded by one of the genes Pcp2, Pvalb, Cabl1, and Itpr1; and purifying at least one labeled nucleus.

[0013] Various embodiments of the invention described herein provide methods of isolating Purkinje cells, the method comprising: processing a tissue sample containing a plurality of cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with an affinity label that binds to a nucleic acid transcript or protein encoded by one of the genes Lypd6, Pvalb, Kit, NeuN, Itpr, and Sorcs3; and purifying at least one labeled nucleus.

[0014] Various embodiments of the invention described herein provide methods for isolating astrocytes and oligodendrocytes, the methods comprising processing a tissue sample containing a plurality of cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with an affinity label that binds to a nucleic acid transcript or protein encoded by one of the genes Olig2, Pdga, Cspg4, Mag, Mbp, and Mog; and purifying at least one labeled nucleus.

[0015] Various embodiments of the invention described herein provide methods for isolating astrocytes, the method comprising processing a tissue sample containing multiple cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with an affinity label that binds to a nucleic acid transcript or protein encoded by one of the genes Aldh1a1, Gfap, S110b, and Slc1a3; and purifying at least one labeled nucleus.

[0016] Various embodiments of the invention described herein provide a method of isolating basket cells, the method comprising: processing a tissue sample containing a plurality of cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with at least one affinity label that binds to a nucleic acid transcript or protein encoded by Sorcs3 and NeuN, where the combination of nucleic acid transcripts or proteins is encoded by Sorcs3 and is unique to basket cell nuclei; and purifying the at least one labeled nucleus.

[0017] Various embodiments of the invention described herein provide methods of isolating dopaminergic neurons, the methods comprising: processing a tissue sample containing multiple cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with at least one affinity label that binds to a nucleic acid transcript or protein specific to dopaminergic neuronal nuclei, such as those encoded by genes FoxA1, Slc6a3, TH, FoxA2, ​​or Drd2; and purifying the at least one labeled nucleus.

[0018] Various embodiments of the invention described herein provide methods of isolating brain stem motor neurons, the methods comprising: processing a tissue sample containing multiple cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with at least one affinity label that binds to a nucleic acid transcript or protein specific to brain stem motor neuron nuclei, such as those encoded by genes VaChT or ErrB; and purifying the at least one labeled nucleus.

[0019] In some embodiments, the affinity label is selected from the group consisting of an antibody, an RNA probe, and a DNA probe. In some embodiments, the method further comprises collecting nucleic acid transcript and / or protein transcript data for the labeled nuclei, thereby profiling the gene expression of the nuclei. In some embodiments, the method further comprises comparing the gene expression in the nuclei with the gene expression in at least one nucleus of a cell type from a different tissue sample to identify changes in gene expression to obtain a drug target. Alternatively, in some embodiments, the method further comprises comparing the gene expression in the nuclei with the gene expression in at least one nucleus of a different cell type to identify changes in gene expression to obtain a drug target.

[0020] In some embodiments, the affinity label further comprises a fluorescent tag. In some embodiments, purification of the nuclei is performed by fluorescence activated cell sorting (FACS). In some embodiments, the tissue sample is from a mammal. In some embodiments, the mammal is selected from the group consisting of a mouse, a human, a rat, and another non-human primate. In some embodiments, the tissue sample or the different tissue sample is frozen prior to isolation of the nuclei. In some embodiments, the tissue sample or the different tissue sample is virgin prior to isolation of the nuclei. In some embodiments, the tissue sample or the different tissue sample is from a female. In some embodiments, the tissue sample or the different tissue sample is from a male. In some embodiments, the nuclei is adjacent to the endoplasmic reticulum. In some embodiments, the protein is a membrane protein. In some embodiments, the membrane protein is synthesized in the endoplasmic reticulum. In some embodiments, the protein is localized in the endoplasmic reticulum. In some embodiments, the nucleic acid transcript is localized in the nucleus. In some embodiments, the RNA probe specifically binds to a chromosome associated transcript (CAT) or a polyA transcript. In some embodiments, the DNA probe specifically binds to a transporter gene. In some embodiments, the affinity labels are two or more antibodies, RNA probes, or DNA probes. In some embodiments, each affinity label binds to a different agent. In some embodiments, the two or more antibodies, RNA probes, or DNA probes bind to the same agent.

[0021] In some embodiments, the tissue sample is post-mortem. In some embodiments, the tissue sample or the different tissue sample is from a diseased subject. In some embodiments, the tissue sample or the different tissue sample is from a healthy subject. In some embodiments, the diseased subject is afflicted with at least one condition selected from the group consisting of ataxia, Parkinson's disease, Alzheimer's disease, and amyotrophic lateral sclerosis (ALS), and Huntington's disease. In some embodiments, the cell type is associated with ataxia and is selected from the group consisting of Purkinje cells, granule cells, Bergmann glia, basket / astrocytes, astrocytes, oligodendrocytes, and deep cerebellar nuclei. In some embodiments, the cell type is associated with Parkinson's disease and is selected from the group consisting of substantia nigra and VTA dopaminergic neurons. In some embodiments, the cell type is associated with Alzheimer's disease and is selected from layer 2 / 3 entorhinal cortex, CA1 hippocampus, and CA2 / 3 hippocampus. In some embodiments, the cell type is associated with ALS and is selected from the group consisting of brainstem, cortical motor neurons, and spinal motor neurons. In some embodiments, the diseased subject expresses a disease phenotype. In some embodiments, the diseased subject does not express a disease phenotype. In some embodiments, the method further comprises comparing profiling of tissue samples from the diseased subject and tissue samples from healthy subjects. In some embodiments, the method further comprises comparing profiling of tissue samples from at least two different healthy subjects. In some embodiments, the method further comprises identifying changes between disease stages by comparing profiling of tissue samples from diseased subjects expressing a disease phenotype and profiling of tissue samples from diseased subjects not expressing a disease phenotype.

[0022] In some embodiments, the method further comprises gene expression profiling performed on multiple nuclei from a single cell type. In some embodiments, the method further comprises gene expression profiling performed on multiple nuclei from different cell types. In some embodiments, the method further comprises comparing profiling of tissue samples from two or more different subjects to identify changes between different subjects. In some embodiments, the drug target is encoded by a gene selected from Tables 19-24.

[0023] Various embodiments of the invention described herein provide methods for profiling epigenetic expression of nuclei, the methods comprising: processing a tissue sample containing multiple cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue with an affinity label that binds to a factor specific to at least one cell type, where the affinity label is selected from the group consisting of an antibody, an RNA probe, and a DNA probe; purifying at least one labeled nucleus; and collecting information regarding histone modifications, transcription factor binding, and DNA modifications to the nuclei, thereby profiling epigenetic expression.

[0024] Various embodiments of the invention described herein provide a method for profiling the nuclear structure of a nucleus, the method comprising: processing a tissue sample containing multiple cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue with an affinity label that binds to a factor specific to at least one cell type, where the affinity label is selected from the group consisting of an antibody, an RNA probe, and a DNA probe; purifying at least one labeled nucleus; and collecting information regarding the chromosomal structure of the labeled nucleus, thereby profiling the nuclear structure.

[0025] Various embodiments of the invention described herein provide a method for profiling the nuclear structure of nuclei, the method comprising: processing a tissue sample containing multiple cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue with an affinity label that binds to a factor specific to at least one cell type, where the affinity label is an antibody; isolating gDNA from the nuclei; performing oxidative bisulfite sequencing; and collecting chromosomal structures for the labeled nuclei, thereby profiling the nuclear structure.

[0026] Various embodiments of the invention described herein provide methods of isolating nuclei from a single cell type, the method comprising: processing a tissue sample containing a plurality of cell types to expose nuclei; isolating nuclei from the processed tissue sample by contacting the processed tissue sample with three or more affinity labels specific for a unique combination of nucleic acid transcripts and / or proteins for the cell type, wherein each affinity label specifically binds to a nucleic acid transcript comprising a sequence selected from SEQ ID NOs: 1-28,815, or each affinity label specifically binds to a protein comprising an amino acid sequence selected from SEQ ID NOs: 28,816-38,643; and purifying at least one labeled nucleus, thereby isolating nuclei from the cell type. [Brief description of the drawings]

[0027] [Figure 1] 1 shows an embodiment of a sorting method that may be used in the cell profiling techniques described herein, where "A" represents a nucleus of a cell type of interest and "B" represents a nucleus of another cell type. [Figure 2A]Flow cytometry sorting of three cell types from the cerebellum using Itpr1 and NeuN. Using a combination of Itpr1 and NeuN, three populations were isolated: Purkinje cells, granule cells, and glia. The x-axis is Itpr1 and the y-axis is NeuN. The Itpr1+ population were Purkinje cell nuclei and the NeuN+ population were glial nuclei. The NeuN+ / Itpr1- nuclei (largest group) represent granule cells. [Figure 2B] Gene expression profiling from RNAseq data comparing sorting of Purkinje nuclei RNA from transgenic animals (GFP sorted) with Itpr1+ nuclei. Pcp4 is a Purkinje marker (enriched in sorted Purkinje nuclei compared to unsorted nuclei in the cerebellum) and Calb2 is a granule cell marker (enriched in unsorted nuclei compared to Purkinje). [Figure 3A] Flow cytometry sorting results of nuclear staining and sorting of dopaminergic nuclei are shown. [Figure 3B] Comparison of RNAseq recordings from unsorted midbrain nuclei and sorted dopaminergic nuclei is shown. [Figure 4] Genome browser diagram showing three recordings - nuclear RNA levels from granule cells, cytoplasmic RNA levels from whole cerebellum, and ATACseq DNA accessibility from whole cerebellum - revealing species- and cell type-specific differences between mouse, rat, and human. [Figure 5A-B] Flow cytometric sorting of nuclei from two different human subjects is shown. [Figure 6] Flow cytometry sorting results are shown, using a combination of forward scatter (FSC) and side scatter (SSC) to approximate nuclear size. [Figure 7A-D] 1 shows flow cytometry results of four different staining methods for human cerebellar nuclei. [Figure 8] Boxplots showing pairwise Pearson correlation coefficients within and between cell types for different individuals. [Figure 9]13 is a dot plot showing pairwise Pearson correlation coefficients grouped by cell type and species for each sample. [Figure 10] Graph showing changes in expression of FosB in granules, basket, and glial cells during autolysis times from 5 to approximately 25 hours. [Figure 11A-C] Scatter plots showing the longitudinal gene expression of three cell type-specific age-decreased genes. Figure 11A shows the results for Aqp7, Figure 11B shows the results for Asic1, and Figure 11C shows the results for Robo2. [Figure 12] Expression of stress response and immediate early genes in three individuals between three cell types in Fos+ and Fos- individuals. Gene records show combined expression from three Fos+ and three age- and sex-matched Fos- individuals. [Figure 13A] Gene expression of fos and immediate early genes in granule cells, astrocytes, and glia is shown. [Figure 13B] Gene expression of long-term stress response genes in these cells is shown. [Figure 13C] We show expression of stress-related genes in glia, such as the activated astroglial marker GFAP, S100B, exportin 1 (also known as Xpo1, leucine-rich nuclear export signal (Nes)), and Musashi RNA-binding protein 1 (Msi1), separate from markers of longer-term stress. [Figure 14] RNAseq data of cell type-specific chromosome-associated transcripts (CATs) are shown. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] I. Introduction Complex tissues are made up of many cell types. So far, the translating ribosome affinity purification (TRAP) method has been used to identify molecular signatures of diverse cell types in a single tissue. The TRAP method requires the use of transgenic animals, which poses many challenges. For example, some disease models are difficult to breed as animals, and animal models do not always recapitulate all aspects of human diseases, such as the effects on human neurons. Furthermore, when studying aging, it is necessary to maintain specific transgenic lines for several years. In addition, the TRAP method requires fresh samples containing high-quality RNA for accurate analysis.

[0029] Determination of molecular characteristics of targeted cell types by genes has provided fundamental insights into mouse brain function and dysfunction. The embodiments of the cell profiling technology described herein provide a method for searching gene expression events in specific cell types in diverse species for thorough and reproducible gene expression studies aimed at identifying markers for each cell type or condition. Results are presented herein for mouse, rat, and human neurons and glia, characterizing homologous human cell types. Classically defined homologous neuronal and glial cell types have been shown to differ between rodents and humans by the expression of hundreds of orthologous, cell-specific genes. Using a combination of epigenetic mapping, RNAseq, quantitative PCR, and immunofluorescence localization, evidence was obtained that these genes are differentially active. Studies on 16 human postmortem brains demonstrated cell-specific molecular responses to aging. The response of human neurons and glia to aging was observed to be cell type specific, and analysis of postmortem human brain tissue also revealed the induction of a common, robust response to an unknown external event experienced by 3 of 16 donors. The compositions and methods described herein create a comprehensive approach for the analysis of unique molecular events associated with specific pathways and cell types in a wide variety of human conditions.

[0030] Recent advances in human genome sequencing and analysis technologies have increased knowledge of the complex genetic causes of human psychiatric and neurological disorders (Burguiere, et al. (2015) Curr Opin Neurobiol 30, 59-65; Hinz, FI, et al. (2017). Molecular Genetics of Neurodegenerative Dementias. PubMed-NCBI. Cold Spring Harb Perspect Biol 9, a023705; Vorstman, et al. (2017) Nat Rev Genet 18, 362-376, each of which is incorporated herein by reference in its entirety). In some cases, recapitulation of the causative mutations in the mouse genome has provided experimental models accurate enough to study the molecular mechanisms of the disorder (Lombardi, et al. (2015) J. Clin. Invest. 125, 2914-2923; Lyst, et al. (2015). Nat Rev Genet 16, 261-275; Orr, HT (2012) J. Cell Biol. 197, 167-177, each of which is incorporated herein by reference in its entirety), and these models have led to the discovery of unexpected features of the disease (Guy, al. (2007) Science 315, 1143-1147, each of which is incorporated herein by reference in its entirety). In other cases, informative animal models remain elusive (Lavin, MF (2013). DNA Repair 12, 612-619, which is incorporated herein by reference in its entirety) and the molecular mechanisms of disease remain difficult to study (Biton, et al. (2008) DNA Repair 7, 1028-1038; Medina, M., et al. (2014) Front. Pharmacol. 5, 270, which is incorporated herein by reference in its entirety).

[0031] Prior to the development of the cell profiling technology described herein, techniques used to attempt genome-wide profiling of specific human cell types include laser capture and single-cell sequencing.None of these techniques can generate reliable deep profiles of specific cell types.Laser capture requires dedicated equipment and training, and results are highly dependent on precise dissection method and tissue anatomy, and therefore there is a large variation in results across samples.

[0032] Data from single-cell RNAseq also varies widely between individual cells. Single-cell RNAseq is also problematic in the nervous system because it is difficult to separate neurons from each other. In studies that isolate neurons, only the cell body is captured and the neurites (axons and dendrites) are lost. It is well known in the art that the processes in neurons are extremely large compared to the size of the cell body and also have their own local translation machinery. Thus, gene expression analysis of only the cell body is missing genes expressed in the neurites, which are some of the most important genes for determining the identity of a neuron. This bias can be avoided by profiling gene expression in the nucleus, which synthesizes all the RNA of the cell. The depth of data from any individual cell is not sufficient to achieve a comprehensive profile, so this method has not been accepted by many in the art of cell profiling.

[0033] Human diseases are problems experienced across entire cell populations. Cell types from single-cell analysis must be determined post hoc, and this analysis may be obscured by disease state. Furthermore, sequencing from single-cell RNAseq is not deep enough to detect subtle changes that may occur with disease. For this reason, cell types should be considered as populations. To do so, single-cell methods require sequencing of thousands of individual cells, and since some cell types are extremely rare, enrichment is necessary to investigate diseases that only affect these rare cell types. Distinguishing between technical and biological changes would be nearly impossible without enrichment.

[0034] To allow for the identification of human-specific targets, the cell profiling techniques described herein have been developed for the molecular profiling of post-mortem human tissues of specific cell types. The cell profiling techniques described herein have a variety of applications, including the study of gene expression and epigenetic changes to identify targets for drugs or other therapeutics.

[0035] Because cells in the CNS have complex morphology and cannot be effectively separated from one another, the cell profiling techniques described herein sort nuclei rather than whole cells. In some embodiments of the method, after isolating nuclei from the brain region of interest, the nuclei can be immunostained with antibodies specific for transcription factors or transcripts that specifically label the cell type of interest. Immunofluorescence clarifies whether nuclei can be labeled with antibodies against membrane proteins. Staining was observed in the endoplasmic reticulum (ER) around the nucleus. Alternatively, RNA or DNA probes may be used to label the cell type of interest.

[0036] In some embodiments, after purification of the labeled population using a fluorescence activated cell sorter (FACS), the isolated nuclei can be used for gene expression profiling. Figure 1 provides an embodiment of a sorting method that can be used with the cell profiling techniques described herein. "A" represents a nucleus of a cell type of interest, and "B" represents a nucleus from another cell type.

[0037] In certain embodiments, an algorithm may be used to identify cell-specific and enriched gene expression for a single cell type. For example, one analytical method to identify and quantify cell-specific and enriched mRNA across multiple cell populations is called the Specificity Index (SI) and is described in Dougherty et al., Nucleic Acids Research, 38(13):4218-4230, which is incorporated herein by reference in its entirety. Gene expression was measured in multiple cell types. These values ​​were first normalized within replicate treatment groups and then normalized across cell types. Non-specific background was filtered out using negative controls. Filtered values ​​were compared with other unfiltered samples in the data set in replicates, and ratios were calculated for each set of samples. The sets were ranked from highest to lowest to prevent extreme outliers from distorting subsequent analysis and to make the analysis of difficult-to-normalize data sets more robust. These ranks were averaged to obtain the SI. A permutation test was used to assign p-values ​​to each SI value, resulting in a list of genes that were significantly enriched in the cell type of interest. The lower the SI, the more specific the gene is for that cell type.

[0038] The cell profiling techniques described herein have a range of applications, including identifying human targets for therapeutics, understanding what variations between people lead to variable penetrance of therapeutics, investigating early changes occurring in human degenerative diseases, identifying drug targets, and developing therapeutics based on comparative analysis of vulnerable and spared cell populations.The cell profiling techniques described herein can be used to profile any cell type of interest.For example, to identify putative drug targets, gene expression can be compared for different cell types to identify genes that are specifically expressed in the cell type of interest.Comparing heterogeneity between different patients can reveal disease biomarkers that can be used to identify patients before symptoms appear or to select a subset of patients for whom therapeutic intervention is likely to be more effective.

[0039] In other embodiments, gene expression between species (e.g., between mouse and human) is compared using the cell profiling techniques described herein to identify common and species-specific genes. This information also fuels drug discovery, since mouse-specific targets are less commercially valuable. Furthermore, species-common targets allow basic research in mice that are more likely to be translatable to humans. Finally, human-specific targets may be the reason why humans develop some diseases, but it is difficult to induce mice to develop the same diseases. Pursuit of common targets may provide novel drug targets that cannot be identified in mice alone.

[0040] Another application of the cell profiling technology described herein is to compare gene expression between the same cell type in the same tissue across different individuals to identify changes across individuals.Currently, genome-wide association studies (GWAS) are used to examine changes between individuals at the DNA level, but these techniques do not show whether DNA changes affect gene expression and function.The data generated by the cell profiling technology described herein can be used for patient stratification, and can be combined with GWAS data to determine why there are changes in response to treatment across individuals.

[0041] In contrast to the TRAP method, the cell profiling techniques described herein do not require the use of transgenic animals. The cell profiling techniques described herein can be used with postmortem tissue from any species, including non-model organisms such as humans and other primates.

[0042] In a particular embodiment of the present invention, tissue samples are processed and antibodies specific to proteins expressed in the cell type of interest are added to the sample. Fluorescent labeling of the antibodies is then added to the sample. Nuclei from the cell type of interest are isolated from the tissue sample by fluorescence. For example, FACS is used to isolate nuclei based on fluorescence intensity. Sorting can be performed using antibodies specific to transcription factors localized in the nucleus (e.g., Matevossian et al., J. Vis. Exp. (13), e717, doi: 10.3791 / 717, (2008), which is incorporated herein by reference in its entirety) or antibodies specific to proteins localized in the endoplasmic reticulum (ER). Since many neurons have been identified based on membrane markers, antibodies specific to membrane proteins synthesized in the ER can be used to investigate the nervous system. Furthermore, it has been observed that a portion of the endoplasmic reticulum remains attached to the isolated nuclei.

[0043] The antibody used for labeling and sorting nuclei from a particular cell type can be selected from the previously described TRAP method investigation to identify factors that are specifically expressed in our cell type of interest. Instead of antibodies, RNA or DNA probes specific for nucleic acid transcripts localized in the nucleus or ER of the cell type of interest are used. For example, Mo et al. (Neuron, 86:1364-1384 (2015), incorporated herein by reference in its entirety) analyzed excitatory neurons, PV interneurons, and VIP neurons of transgenic mice using Camk2a, PV, and VIP as targets, respectively. According to Mo et al., about 50% of cortical nuclei were Camk2a positive, about 3% were PV positive, and about 2% were VIP positive.

[0044] The methods described herein provide circuit-based molecular characterization of cell types involved in the study of cellular vulnerability in human neurodegenerative diseases, obsessive-compulsive disorder, drug addiction, autism spectrum disorder, and age-dependent cognitive impairment. In some embodiments, the goal of the cell profiling techniques described herein is to obtain human CNS cell type data to establish characterization of brain circuits or pathways involved in important human neurological disorders and neurodegenerative diseases. For example, cell-specific data from mice was obtained using antibodies specific to proteins identified from nuclear-enriched, fully processed, polyA+ RNA and chromosomal associated transcripts (CAT). RNA / DNA probes for these can be designed for additional labeling and sorting. Specifically, purified GFP+ nuclei and isolated RNA from transgenic animals developed for the TRAP method were analyzed using polyDT beads that capture mature RNA, polyA+ RNA. Enrichment of RNA was observed in the ER, adjacent to the nuclear membrane. These data provide details about transcripts that are enriched in the ER in close proximity to the nucleus and are putative targets of antibodies and RNA / DNA probes that can be used to perform the sorting step of the cell profiling techniques described herein. The methods described herein result in the generation of highly reproducible cell type specific data and validation of cell specific expression profiles by in situ hybridization or immunofluorescence of newly discovered cell specific markers in human tissue samples, providing validation that transcript profiles from bulk tissues, such as those from the brain, can be used to identify novel cell specific probes.

[0045] II. Gene Nomenclature Gene symbols are used herein along with the ensemble gene ID to refer to genes from mouse, rat, and human. Unless otherwise noted, the gene name corresponding to each gene symbol is as shown in Table 1. [Table 1] TIFF2025032204000003.tif245165TIFF2025032204000004.tif245164TIFF2025032204000005.tif245165TIFF2025032204000006.tif245162TIFF2025032204000007.tif245164TIFF2025032204000008.tif245164TIFF2025032204000009.tif245165TIFF2025032204000010.tif245164TIFF2025032204000011.tif245164TIFF2025032204000012.tif245164TIFF2025032204000013.tif245161TIFF2025032204000014.tif245164TIFF2025032204000015.tif245162TIFF2025032204000016.tif245165TIFF2025032204000017.tif245164TIFF2025032204000018.tif245161TIFF2025032204000019.tif245164TIFF2025032204000020.tif245165TIFF2025032204000021.tif245164TIFF2025032204000022.tif245162TIFF2025032204000023.tif245164TIFF2025032204000024.tif245164TIFF2025032204000025.tif245165TIFF2025032204000026.tif245165TIFF2025032204000027.tif245162TIFF2025032204000028.tif245164TIFF2025032204000029.tif245164TIFF2025032204000030.tif245163TIFF2025032204000031.tif245164TIFF2025032204000032.tif245164TIFF2025032204000033.tif245165TIFF2025032204000034.tif245165TIFF2025032204000035.tif245165TIFF2025032204000036.tif245162TIFF2025032204000037.tif245165TIFF202503220400 0038.tif245164TIFF2025032204000039.tif245162TIFF2025032204000040.tif245162TIFF2025032 204000041.tif245163TIFF2025032204000042.tif245165TIFF2025032204000043.tif245165TIFF20 25032204000044.tif245165TIFF2025032204000045.tif245165TIFF2025032204000046.tif127165.

[0046] For mouse, Tables 2 and 2b provide unique ensemble identifiers corresponding to the mouse genes (MUSG), transcripts (MUST), and proteins (when available) (MUSP) analyzed in the experiments herein. The unique identifiers for each ensemble item were corrected to remove leading zeros (0) from the first five following identifiers for the ENSMUSG, ENSMUST, and ENSMUSP labels. If the ensemble transcript or protein identifier is not available, include the GenBank transcript (nucleic acid or protein). [Table 2] TIFF2025032204000048.tif245164TIFF2025032204000049.tif245164TIFF2025032204000050.tif245165TIFF2025032204000051.tif245165TIFF2025032204000052.tif245165TIFF2025032204000053.tif245164TIFF2025032204000054.tif245165TIFF2025032204000055.tif245164TIFF2025032204000056.tif245164TIFF2025032204000057.tif245165TIFF2025032204000058.tif245164TIFF2025032204000059.tif245165TIFF2025032204000060.tif245165TIFF2025032204000061.tif245164TIFF2025032204000062.tif245164TIFF2025032204000063.tif245165TIFF2025032204000064.tif245165TIFF2025032204000065.tif245165TIFF2025032204000066.tif245164TIFF2025032204000067.tif245165TIFF2025032204000068.tif245165TIFF2025032204000069.tif245165TIFF2025032204000070.tif245165TIFF2025032204000071.tif245165TIFF2025032204000072.tif245165TIFF2025032204000073.tif245164TIFF2025032204000074.tif245164TIFF2025032204000075.tif245165TIFF2025032204000076.tif245164TIFF2025032204000077.tif245164TIFF2025032204000078.tif245164TIFF2025032204000079.tif245165TIFF2025032204000080.tif245164TIFF2025032204000081.tif245165TIFF2025032204000082.tif245165TIFF2025032204000083.tif245164TIFF2025032204000084.tif245164TIFF2025032204000085.tif245165TIFF2025032204000086.tif245165TIFF2025032204000087.tif245165TIFF2025032204000088.tif245165TIFF2025032204000089.tif245164TIFF2025032204000090.tif245164TIFF2025032204000091.tif245164TIFF2025032204000092.tif245164TIFF2025032204000093.tif245164TIFF2025032204000094.tif245164TIFF2025032204000095.tif245164TIFF2025032204000096.tif245165TIFF2025032204000097.tif245165TIFF2025032204000098.tif245165TIFF2025032204000099.tif245165TIFF2025032204000100.tif245165TIFF2025032204000101.tif245165TIFF2025032204000102.tif245165TIFF2025032204000103.tif245165TIFF2025032204000104.tif245165TIFF2025032204000105.tif245165TIFF2025032204000106.tif245164TIFF2025032204000107.tif245164TIFF2025032204000108.tif245165TIFF2025032204000109.tif245164TIFF2025032204000110.tif83165.

[0047] Table 2b includes additional genes for mouse. This table identifies unique ensemble identifiers corresponding to the mouse genes (MUSG), transcripts (MUST), and proteins (when available) (MUSP) analyzed in the experiments herein. The unique identifiers for each ensemble item were corrected to remove leading zeros (0) from the first five following identifiers for the ENSMUSG, ENSMUST, and ENSMUSP labels. If the ensemble transcript or protein identifier is not available, include the GenBank transcript (nucleic acid or protein). [Table 3] TIFF2025032204000112.tif245163TIFF2025032204000113.tif245162TIFF2025032204 000114.tif245163TIFF2025032204000115.tif245165TIFF2025032204000116.tif24516 3TIFF2025032204000117.tif245165TIFF2025032204000118.tif245163TIFF2025032204 000119.tif245164TIFF2025032204000120.tif245162TIFF2025032204000121.tif52165

[0048] For the rat, unique ensemble identifiers corresponding to the rat genes (MUSG), transcripts (MUST), and proteins (where available) (MUSP) analyzed in the experiments herein are provided in Table 3. The unique identifiers for each ensemble item were modified to remove leading zeros (0) from the first five following identifiers for the ENSMUSG, ENSMUST, and ENSMUSP labels. [Table 4] TIFF2025032204000123.tif237166TIFF2025032204000124.tif245164TIFF2025032204000125.tif245164TIFF2025032204000126.tif245164TIFF2025032204000127.tif245165TIFF2025032204000128.tif245163TIFF2025032204000129.tif245163TIFF2025032204000130.tif245164TIFF2025032204000131.tif245163TIFF2025032204000132.tif245163TIFF2025032204000133.tif245163TIFF2025032204000134.tif245163TIFF2025032204000135.tif245163TIFF2025032204000136.tif245163TIFF2025032204000137.tif245163TIFF2025032204000138.tif245163TIFF2025032204000139.tif245164TIFF2025032204000140.tif245164TIFF2025032204000141.tif245163TIFF2025032204000142.tif245163TIFF2025032204000143.tif245164TIFF2025032204000144.tif245161TIFF2025032204000145.tif245163TIFF2025032204000146.tif245164TIFF2025032204000147.tif245164TIFF2025032204000148.tif245163TIFF2025032204000149.tif245163TIFF2025032204000150.tif245163TIFF2025032204000151.tif245163TIFF2025032204000152.tif245163TIFF2025032204000153.tif245162TIFF2025032204000154.tif245162TIFF2025032204000155.tif245162TIFF2025032204000156.tif245162TIFF2025032204000157.tif245163TIFF2025032204000158.tif245162TIFF2025032204000159.tif245163TIFF2025032204000160.tif245163TIFF2025032204000161.tif245163TIFF2025032204000162.tif245162TIFF2025032204000163.tif245163TIFF2025032204000164.tif245163TIFF2025032204000165.tif245164TIFF2025032204000166.tif245164TIFF2025032204000167.tif245162TIFF2025032204000168.tif245162TIFF2025032204000169.tif245162TIFF2025032204000170.tif245162TIFF2025032204000171.tif245163TIFF2025032204000172.tif245163TIFF2025032204000173.tif245164TIFF2025032204000174.tif245162TIFF2025032204000175.tif245162TIFF2025032204000176.tif245163TIFF2025032204000177.tif245163TIFF2025032204000178.tif245163TIFF2025032204000179.tif245162TIFF2025032204000180.tif245162TIFF2025032204000181.tif245163TIFF2025032204000182.tif245163TIFF2025032204000183.tif245163TIFF2025032204000184.tif245162TIFF2025032204000185.tif245163TIFF2025032204000186.tif91166.

[0049] For human samples, unique ensemble identifiers corresponding to the human genes (ENSG), transcripts (ENST), and proteins (where available) (ENSP) analyzed in the experiments herein are provided in Table 4. The unique identifiers for each ensemble item were corrected to remove leading zeros (0) from the first five following identifiers for the ENSG, ENST, and ENSP labels. [Table 5] TIFF2025032204000188.tif245164TIFF2025032204000189.tif245162TIFF2025032204000190.tif245163TIFF2025032204000191.tif245163TIFF2025032204000192.tif245163TIFF2025032204000193.tif245164TIFF2025032204000194.tif245163TIFF2025032204000195.tif245163TIFF2025032204000196.tif245161TIFF2025032204000197.tif245162TIFF2025032204000198.tif245162TIFF2025032204000199.tif245161TIFF2025032204000200.tif245163TIFF2025032204000201.tif245162TIFF2025032204000202.tif245163TIFF2025032204000203.tif245162TIFF2025032204000204.tif245163TIFF2025032204000205.tif245164TIFF2025032204000206.tif245162TIFF2025032204000207.tif245162TIFF2025032204000208.tif245163TIFF2025032204000209.tif245163TIFF2025032204000210.tif245162TIFF2025032204000211.tif245162TIFF2025032204000212.tif245163TIFF2025032204000213.tif245163TIFF2025032204000214.tif245161TIFF2025032204000215.tif245162TIFF2025032204000216.tif245162TIFF2025032204000217.tif245162TIFF2025032204000218.tif245163TIFF2025032204000219.tif245163TIFF2025032204000220.tif245162TIFF2025032204000221.tif245161TIFF2025032204000222.tif245162TIFF2025032204000223.tif245162TIFF2025032204000224.tif245162TIFF2025032204000225.tif245162TIFF2025032204000226.tif245162TIFF2025032204000227.tif245162TIFF2025032204000228.tif245161TIFF2025032204000229.tif245162TIFF2025032204000230.tif245162TIFF2025032204000231.tif245162TIFF2025032204000232.tif245162TIFF2025032204000233.tif245163TIFF2025032204000234.tif245163TIFF2025032204000235.tif245163TIFF2025032204000236.tif245162TIFF2025032204000237.tif245164TIFF2025032204000238.tif245164TIFF2025032204000239.tif245163TIFF2025032204000240.tif245163TIFF2025032204000241.tif245162TIFF2025032204000242.tif245162TIFF2025032204000243.tif245162TIFF2025032204000244.tif245163TIFF2025032204000245.tif245162TIFF2025032204000246.tif245162TIFF2025032204000247.tif245163TIFF2025032204000248.tif245161TIFF2025032204000249.tif245162TIFF2025032204000250.tif245164TIFF2025032204000251.tif245162TIFF2025032204000252.tif245162TIFF2025032204000253.tif245161TIFF2025032204000254.tif245161TIFF2025032204000255.tif245162TIFF2025032204000256.tif245161TIFF2025032204000257.tif245162TIFF2025032204000258.tif245162TIFF2025032204000259.tif245162TIFF2025032204000260.tif245161TIFF2025032204000261.tif245162TIFF2025032204000262.tif245161TIFF2025032204000263.tif245161TIFF2025032204000264.tif245163TIFF2025032204000265.tif245164TIFF2025032204000266.tif245162TIFF2025032204000267.tif245162TIFF2025032204000268.tif245162TIFF2025032204000269.tif245162TIFF2025032204000270.tif245162TIFF2025032204000271.tif245162TIFF2025032204000272.tif245164TIFF2025032204000273.tif245165TIFF2025032204000274.tif245164TIFF2025032204000275.tif245163TIFF2025032204000276.tif245165TIFF2025032204000277.tif245162TIFF2025032204000278.tif245162TIFF2025032204000279.tif245162TIFF2025032204000280.tif245163TIFF2025032204000281.tif245163TIFF2025032204000282.tif245162TIFF2025032204000283.tif245164TIFF2025032204000284.tif245161TIFF2025032204000285.tif245163TIFF2025032204000286.tif245163TIFF2025032204000287.tif245162TIFF2025032204000288.tif245164TIFF2025032204000289.tif245163TIFF2025032204000290.tif245162TIFF2025032204000291.tif245162TIFF2025032204000292.tif245161TIFF2025032204000293.tif245164TIFF2025032204000294.tif245162TIFF2025032204000295.tif245163TIFF2025032204000296.tif245164TIFF2025032204000297.tif245163TIFF2025032204000298.tif245163TIFF2025032204000299.tif245162TIFF2025032204000300.tif245162TIFF2025032204000301.tif245162TIFF2025032204000302.tif245162TIFF2025032204000303.tif245164TIFF2025032204000304.tif245162TIFF2025032204000305.tif245163TIFF2025032204000306.tif245161TIFF2025032204000307.tif245162TIFF2025032204000308.tif245162TIFF2025032204000309.tif245162TIFF2025032204000310.tif245163TIFF2025032204000311.tif245162TIFF2025032204000312.tif245162TIFF2025032204000313.tif245163TIFF2025032204000314.tif245163TIFF2025032204000315.tif245162TIFF2025032204000316.tif245163TIFF2025032204000317.tif245165TIFF2025032204000318.tif245163TIFF2025032204000319.tif245163TIFF2025032204000320.tif245162TIFF2025032204000321.tif245163TIFF2025032204000322.tif245164TIFF2025032204000323.tif245163TIFF2025032204000324.tif245163TIFF2025032204000325.tif245164TIFF2025032204000326.tif245162TIFF2025032204000327.tif245165TIFF2025032204000328.tif245162TIFF2025032204000329.tif245163TIFF2025032204000330.tif245163TIFF2025032204000331.tif245163TIFF2025032204000332.tif245162TIFF2025032204000333.tif245163TIFF2025032204000334.tif245163TIFF2025032204000335.tif245162TIFF2025032204000336.tif245162TIFF2025032204000337.tif245162TIFF2025032204000338.tif245163TIFF2025032204000339.tif245162TIFF2025032204000340.tif245163TIFF2025032204000341.tif245162TIFF2025032204000342.tif245162TIFF2025032204000343.tif245163TIFF2025032204000344.tif245161TIFF2025032204000345.tif245163TIFF2025032204000346.tif245163TIFF2025032204000347.tif245162TIFF2025032204000348.tif245163TIFF2025032204000349.tif245162TIFF2025032204000350.tif245163TIFF2025032204000351.tif245162TIFF2025032204000352.tif245163TIFF2025032204000353.tif245162TIFF2025032204000354.tif245163TIFF2025032204000355.tif245162TIFF2025032204000356.tif245162TIFF2025032204000357.tif245162TIFF2025032204000358.tif245162TIFF2025032204000359.tif245163TIFF2025032204000360.tif245163TIFF2025032204000361.tif245163TIFF2025032204000362.tif245161TIFF2025032204000363.tif245163TIFF2025032204000364.tif245162TIFF2025032204000365.tif245163TIFF2025032204000366.tif245162TIFF2025032204000367.tif245162TIFF2025032204000368.tif245162TIFF2025032204000369.tif245162TIFF2025032204000370.tif245162TIFF2025032204000371.tif245161TIFF2025032204000372.tif245162TIFF2025032204000373.tif245162TIFF2025032204000374.tif245163TIFF2025032204000375.tif245162TIFF2025032204000376.tif245162TIFF2025032204000377.tif245162TIFF2025032204000378.tif245163TIFF2025032204000379.tif245163TIFF2025032204000380.tif245162TIFF2025032204000381.tif245163TIFF2025032204000382.tif245164TIFF2025032204000383.tif245162TIFF2025032204000384.tif245163TIFF2025032204000385.tif245162TIFF2025032204000386.tif245162TIFF2025032204000387.tif245163TIFF2025032 204000388.tif245163TIFF2025032204000389.tif245161TIFF2025032204000390.tif221162.

[0050] III. Isolation and Profiling Methods of Nuclei from Cell Types In certain cell populations, many questions of human biology can be productively studied. For example, understanding the molecular events associated with human neurodegenerative events will provide insight into the genetic complexity of these conditions (Burguiere, E., et al. (2015) Curr Opin Neurobiol 30, 59-65.; Hinz, et al. (2017) Cold Spring Harb Perspect Biol 9, a023705; Vorstman, et al. (2017) Nat Rev Genet 18, 362-376, each of which is incorporated herein by reference in its entirety), the stochastic nature of neuronal loss (Clarke, et al. (2000). Nature 406, 195-199 and Clarke, et al. (2005) Brain Research Bulletin 65, 59-67, each of which is incorporated herein by reference in its entirety), and the changing effects of aging on both disease onset and progression (Corrada, et al. (2010) Ann. Neurol. 67, 114-121 and Niccoli, et al. (2012) Current Biology 22, R741-R752, each of which is incorporated herein by reference in its entirety), a significant number of different cell types from human samples can be obtained through longitudinal studies. With this in mind, a method herein was developed to precisely determine the molecular characteristics of specific brain cell types. The method is robust enough to employ postmortem tissue from common brain tissue banks without the need for transgenic animals.

[0051] The cell profiling techniques described herein were developed as approaches for cell type-specific analysis that exploit the unique complement of nuclear localization and nuclear associated proteins within each cell type to allow for robust and reproducible study of specific cell types in the mammalian brain. The cell profiling techniques described herein allow for cell type-specific gene expression profiling of mouse, rat, and human postmortem brain, and these methods are extendable to any organ or tissue sample.

[0052] Comparative examination of data from the cell profiling techniques described herein from rodent and human brain cell types reveals evolutionary divergence of gene expression profiles from highly conserved cerebellar cell types determined by classical methods (D'Angelo, E. (2013) Handbook of the Cerebellum and Cerebellar Disorders, (Dordrecht: Springer Netherlands), pp. 765-791; Eccles, JC (1967) The Cerebellum as a Neuronal Machine (Berlin / Heidelberg / New York: Springer); Llinas, RR (1969). Neurobiology of Cerebellar Evolution and Development (Chicago: American Medical Association); each of which is incorporated herein by reference in its entirety). The data also reveal molecular mechanisms of brain aging differently in each human cell type. Finally, the data show distinct molecular phenotypes indicating that common external events can be identified in control postmortem human brain. Comparative studies of human genetic and clinical data with expression and epigenetic investigations using the cell profiling techniques described herein will provide insight into additional, but unrecognized, molecular characteristics of human brain function and dysfunction.

[0053] Transgenic mice expressing EGFP-L10a ribosomal protein fusion under the control of cell type-specific promoters have been previously developed in Doyle, et al. (2008) Cell 135, 749-762 and Heiman, et al. (2008) Cell 135, 738-748, each of which is incorporated herein by reference in its entirety. Using these five transgenic animal lines developed for use with the TRAP method, EGFP+ nuclei (Kriaucionis, et al. (2009) Science 324, 929-930 and Mellen, M., et al. (2012) Cell 151, 1417-1430, each of which is incorporated herein by reference in its entirety) were purified from specific cell types by gene expression using the TRAP method or by epigenetic characterization after the use of fluorescence-activated cell sorting (FACS). Previous studies have confirmed that nuclear RNA profiles are also unique to each cell type and that they are informative regarding cellular function (Deal, et al. (2011) Nat Protoc 6, 56-68; Henry et al., Nucleic Acids Res. 2012 Oct; 40(19):9691-704; Mo, et al. (2015) Neuron 86, 1369-1384; Steiner, et al. (2012) Genome Res 22, 766-777; each of which is incorporated by reference in its entirety herein).

[0054] An important limitation of cell type-specific molecular profiling methods developed prior to the cell profiling techniques described herein is that they require the use of transgenic animals to genetically target the cell type of interest.To overcome this limitation, we have investigated the cell type-specific expression of nuclear and endoplasmic reticulum proteins in detail, purifying and characterizing nuclei from specific cell types in various species, including human postmortem brain samples.

[0055] In some embodiments of the invention, nuclei are prepared from wild type or transgenic tissue, fixed with formaldehyde, stained with fluorescent antibodies specific for a given cell type in the brain region of interest, sorted by flow cytometry (e.g., FACS), and then analyzed using RNAseq to characterize species-, cell type-, or environment-specific expression profiles.

[0056] The cell profiling technique described herein may be applied to nuclei from any cell type or tissue that is related to a particular disease, disorder, or condition. For example, it may be applied to nuclei from cell types originating from basal ganglia tissue, including striatonigral medium spiny neurons (MSNs), striatopallidal MSNs (caudate and nucleus accumbens), striatal cholinergic interneurons, subthalamic nucleus, dopaminergic neurons, and bed nucleus of stria terminalis (BNST) neurons. For example, it may be applied to nuclei from cell types originating from thalamic tissue, including thalamocortical neurons, thalamostriatal neurons, and thalamic reticular nucleus neurons. For example, it may be applied to nuclei from cell types originating from cortical tissue, including entorhinal cortex layer 2 / 3 neurons, fast-firing cortical interneurons, and layer 2 / 3 pyramidal cells from prefrontal cortical tissue. For example, it may be applied to nuclei from cell types originating from hippocampal tissue, including cornu ammonis area 1 (CA1), cornu ammonis area 2 (CA2), cornu ammonis area 3 (CA3), and dentate gyrus cells.

[0057] Cell types may also be associated with the following tissues: medial habenula (cholinergic projection neurons, projecting to the interpeduncular nucleus), brainstem (upper motor neurons), spinal cord (lower motor neurons), or all tissues (various types of interneurons, astrocytes and oligodendrocytes). Human brain cell types may also be associated with neurodegenerative conditions such as ataxia, Parkinson's disease, Alzheimer's disease, and amyotrophic lateral sclerosis (ALS). For example, cell types are profiled that are associated with ataxia, such as Purkinje cells, granule cells, Bergmann glia, basket / astrocytes, and deep cerebellar nuclei. Also profiled are substantia nigra ventral tegmental area (VTA) dopaminergic neurons associated with Parkinson's disease or layer 2 / 3 entorhinal cortex, CA1 hippocampus, CA2 / 3 hippocampal neurons associated with Alzheimer's disease. Alternatively, brainstem and spinal motor neurons associated with ALS are profiled.

[0058] A. Cell Type Sorting Using Nuclear Transcriptomes Nuclei may be sorted using antibodies, RNA probes, or DNA probes specific for endoplasmic reticulum (ER) proteins, membrane proteins synthesized in the ER, or nuclear transcripts. To enrich for ER transcripts translated near the nuclear membrane, polyA pull-down of nuclear transcripts may be performed. In some embodiments, nuclei from cell types of interest were isolated from tissues of transgenic mice previously profiled using the TRAP method to compare the identity of cells obtained from each technique. The profiling results of transgenic animals developed for use with the TRAP method were based on the fact that GFP-tagged ribosomal subunits are integrated into the nucleoli. Alternatively, the tissue from which the sorted nuclei originate is derived from a wild-type mammal, such as a human, non-human primate, mouse, or rat.

[0059] Figure 1 provides an embodiment of a sorting method that may be used in the cell profiling techniques described herein, also referred to as NETSseq™ technology or INSPECTION™. "A" represents a cell type of interest and "B" represents the nucleus of the cell type of interest in Figure 1.

[0060] In some embodiments, the strategy to isolate three cell types from the cerebellum in mice includes antibody binding using Itpr1 (inositol 1,4,5-triphosphate receptor type 1; an intracellular receptor in the endoplasmic reticulum (ER)) as a target to label Purkinje cells, NeuN (a gene encoded by a member of the RNA-binding Fox protein family; NeuN is a neuroantigen that is often used as a biomarker for neurons) as a target to label granule cells, and Gfap (glial fibrillary acidic protein; one of the major intermediate filament proteins of mature astrocytes used to distinguish astrocytes from glial cells) as a target to label glia to confirm the presence of these cell types in the sample. A combination of the Itpr1 and NeuN results can then be used to isolate all three populations. Alternatively, nuclei from wild-type mouse granule cells can be sorted as positive for Itpr (Itpr+) and NeuN (NeuN+), and negative for Olig2 (Olig2-) and Sorcs3 (Sorcs3-). Nuclei from Purkinje cells can be sorted as Itpr1+ and NeuN+. Nuclei from basket cells can be identified as Sorcs3+ and intermediate NeuN-. Nuclei from astrocytes can be identified as Sorcs3- and NeuN-.

[0061] Alternatively, neuronal differentiation 1 or Neurod1 (Neurod1+) may be used to sort granule cells; Purkinje cell protein 2 or Pcp2 (Pcp2+) may be used to sort Purkinje cells; Septin 4 or Sept4 (Sept4+) may be used to sort Bergmann glia; Collagen beta(1-O) galactosyltransferase 2 or Glt25d2, also known as Colgalt2, (Glt25d2+) was used to sort corticopontin pyramidal cells; and Neurotensin receptor 1 or Ntsr1 (Ntsr+) was used to sort corticothalamic pyramidal cells. In some embodiments, the following nuclear transcripts may be used to identify the nuclei of each cell type: alpha-2-macroglobulin (A2m) for Sept4 positive cells (Bergmann glia), Pde1a and Colgalt2 for Colgalt2 positive cells (corticopontin pyramids), Pde1a and heparan sulfate-glucosamine 3-sulfotransferase 4 (Hs3st4) for Excit positive cells (excitatory neurons), Pde1a and Hs3st4 for Ntsr1 positive cells (corticothalamic pyramids), aristares-related homeobox (Arx) and parvalbumin (Pvalb) for PV positive cells (parvalbumin interneurons), vasoactive intestinal peptide (Vip) for VIP positive cells (VIP interneurons), solute transporter family 9 member A3 (Slc9a3) for Pcp2 positive cells (Purkinje), and Grin2c for Neurod1 positive cells (granules).

[0062] In some embodiments, dopaminergic neuronal nuclei from mice were sorted using Forkhead box A1 (FoxA1) and dopamine transporter solute carrier family 6 member 3 (Dat, also known as Slc6a3). Dopaminergic neurons have been observed to be positive for both FoxA1 and Dat. In some embodiments, additional markers of dopaminergic neurons that can be used for sorting include: tyrosine hydrolase (TH), forkhead box A2 (FoxA2), and dopamine receptor 2 (Drd2).

[0063] In some embodiments, genes that are downregulated in a sorted population of nuclei may also be used as evidence for the cell type specificity of the nuclei. For example, dopamine receptor D1 (Drd1) and glutamic acid decarboxylase (Gad) are not known to be expressed in dopaminergic neurons and therefore should be downregulated in a sorted population of nuclei from dopaminergic neurons.

[0064] In some embodiments, wolframin ER transmembrane glycoprotein (Wfs1) and B-cell CLL / lymphoma 11B (Ctip2, also known as Bcl11b) may be used to select the nuclei of cortical cell types layer 2 / 3 pyramidal neurons and layer 5 and 6 pyramidal neurons, respectively, from mice. In some embodiments, ETS variant 1 (Etv1, also known as ER81) may be used to select the nuclei of layer 5 pyramidal neurons. Etv1 has previously been observed to label both corticostriatal neurons, corticopontin neurons, and possibly others, in layer 5 of the cortex. In some embodiments, Wfs1 may be used as a marker for carbonic anhydrase 1 (CA1)-positive neurons, and Purkinje cell protein 4 (Pcp4) may be used as a marker for carbonic anhydrase 1 (CA2)-positive dentate gyrus.

[0065] In some embodiments, solute transporter family 18 member A3 (VaChT, also known as Slc18a3), chondrolectin (Chodl), and estrogen-related receptor beta (EsrrB) may be used as markers for nuclei from mouse brainstem motor neurons. Affinity labels specific to VaChT and ErrB specifically label motor neurons, while Chodl labels everything but these neurons. Antibodies specific to Chodl may be substituted for those that directly label brainstem motor neurons.

[0066] In some embodiments, rat nuclei are sorted using at least one membrane protein or nuclear-associated transcript. For example, nuclei from granule cells can be identified as Itpr negative (Itpr-), Olig2 negative (Olig2-), Sorcs negative (Sorcs-), and NeuN positive (NeuN+). Nuclei from Purkinje cells can be identified as Itpr positive (Itpr+) and NeuN+. Nuclei from basket cells can be identified as Itpr+ and moderate NeuN negative (med NeuN-). Nuclei from astrocytes can be identified as Sorcs3 positive (Sorcs3+) and med NeuN-. Nuclei from mature oligodendrocytes can be identified as Sorcs3- and NeuN-. Nuclei from a combination of mature oligodendrocytes and OPCs can be identified as Olig2 positive (Olig2+) and high NeuN-.

[0067] In some embodiments, nuclei from Purkinje cells from human samples may be identified as Itpr+ and Id2 positive (Id2+) or FoxP4 positive (FoxP4+) and Itpr1+. In some embodiments, nuclei from granule cells from human samples were identified as FoxP4+ and NeuN+. In some embodiments, nuclei from mature oligodendrocytes and OPCs may be distinguished by the expression level of Olig2. For example, low levels of Olig2 expression (Olig2+low) identify nuclei as mature oligodendrocytes and high levels of Olig2 expression (Olig2+high) identify nuclei as OPCs. The nuclear sorting strategy described herein can be applied to any species.

[0068] B. Profiling of cell types in sorted nuclei of different species The genes expressed in each cell type define their function, their response to internal and external cues, and their occurrence across species. The cell profiling techniques described herein have been demonstrated to be useful for accurate cell type-specific gene expression profiling of neurons and glia from mouse, rat, and human brains.

[0069] In Mo et al., (2015) Neuron 86, 1369-1384 (incorporated herein by reference in its entirety), nuclei from transgenic mice expressing tagged nuclear membrane proteins with the Cre-loxP system were profiled using RNAseq. Comparison of RNAseq profiles of nuclei isolated from cell types of transgenic animals and wild-type animals of different origins provides evidence that nuclear RNA profiles can be used to determine cell identity and measure the association of different cell types. In some embodiments, transgenic animals can express increased GFP (EGFP / L10a fusion protein) in targeted cell types. In some embodiments, the nuclei sorted and characterized in Mo et al. provide RNAseq-derived hierarchical clustering, and the cell profiling techniques described herein can be used to group by cell type, such as interneurons (e.g., VIP and parvalbumin interneurons) versus pyramidal cells (e.g., cortical pyramidal, cortical thalamic, and stimulatory).

[0070] Even homogeneous cerebellar cell types, defined by traditional methods, were observed to differ between species, with expression of hundreds of orthologous genes, and sequential expression of human granule cell-expressed genes was also commonly observed in other regions of the mouse brain. Indeed, nuclear RNA profiles are sensitive enough to distinguish between different neuronal subtypes, such as mature oligodendrocytes (also called corticopontin pyramidal neurons), and oligodendrocyte precursor cells or OPCs (also called corticothalamic pyramidal neurons).

[0071] 1. Differences in RNA profiles of sorted nuclei from mice In some embodiments, gene expression is analyzed by RNAseq to generate RNA profile.In some embodiments, the results are normalized and the expression of each gene is averaged across all samples.For example, the log2 fold change results of RNAseq from the nuclei purified by FACS of granule cells, Purkinje cells, Bergmann glia, oligodendrocytes and OPCs can be compared to the normalized RNAseq results of the 250 most changing genes across all nuclei from all cell types. In mice, the most altered genes can be selected from the following: 1700047M11Rik, 3110039M20Rik, 2410124H12Rik, 8030451O07Rik, 9530059O14Rik, A230077H06Rik, A2m, A730036I17Rik, Aass, Acsm5, Actbl2, Adra2b, Afp, Aqp4, Arhgap25, Arx, Atp13a4, Atp1a2, Atp2a 3, B230209E15Rik, B3gnt5, Barhl2, Bcl11b, Btbd17, Cacng3, Camk1g, Camkv, Car8, Casp12, Casq2, Cbln2, Cbln3, Cckb r, Cd70, Cd9, Cdc42ep1, Cdh19, Chrm1, Chrm3, Chsy3, Clic6, Cml5, Cnpy1, Cobl, Col1a2, Col4a5, Col6a1, Colgalt2, Cor in, Coro6, Cpne7, Crym, Csta1, Ctxn3, Cxcl14, Cxcl5, Cxcr4, Cyp27a1, Cyp2j9, D430019H16Rik, Ddn, Ddx3y, Dlgap2, D lx4os, Dlx6os1, Dmp1, E030003E18Rik, Ebf1, Ebf2, Ebf3, Ednrb, Egfr, Egr3, Eif2s3y, Elfn1, Emp2, En2, Enc1, Etnppl, Eva1a, F3, Fam107b, Fam19a1, Fat2, Fbln2, Fezf2, Fgd3, Flrt2, Folh1, Foxg1, Frem1, Gabra6, Galnt14, Gas1, Gda, Gdf 10, Gja1, Gjb6, Gjc3, Gli1, Gli3, Glt25d2, Gm11549, Gm266, Gm5083, Gm5089, Gm5468, Gm5607, Gm8179, Gpr37l1, Grik1,Grm5, Hepacam, Hes3, Hs3st2, Hs3st4, Htr2a, Icam5, Icosl, Il22, Iltifb, Irx1, Irx2, Irx5, Itih3, Itpripl2, Kank1, Kcnc2, Kcnf1, Kc nj10, Kcnj16, Kcnj4, Kcnq5, Kcnt2, Kcnv1, Kctd12b, Kdm5d, Lama2, Lamp5, Lcat, Ldb2, Lfng, Lgr6, Lhx1, Lhx1os, Lhx5, Linc-md1, Lrrc 7, Lyzl4, Mal2, March1, Megf10, Meis2, Mertk, Metrn, Mgst1, Miat, Mlc1, Mmp14, Mpped1, Msx2, Mybpc1, Myh6, Neto1, Nid1, Nkx2-2, Npr 3, Npy1r, Nrgn, Nrk, Ntsr2, Nup62cl, Oprd1, P2ry1, Paqr6, Pax3, Pcsk6, Pde1a, Pde5a, Pdgfd, Pdzph1, Pih1h3b, Pkp3, Pla2g7, Plekhd1 , Plekho2, Plpp3, Plscr1, Plscr4, Ppp1r17, Prkag3, Procr, Prr5, Ptk2b, Ptpn5, Ptprz1, Rab11fip1, Rab27b, Robo3, Rtn4rl2, Rxfp1, S 1pr1, Scube1, Sdc4, Serpinh1, Skor2, Slc14a1, Slc1a3, Slc24a4, Slc2a10, Slc2a4, Slc30a3, Slc39a12, Slc6a7, Slc7a10, Slc9a3, Slc o4c1, Smpx, Sod3, Sparc, Stac, Stk17b, Stk32a, Sycp1, Syt10, Tbr1, Tex36, Tlcd1, Tlr3, Tmem132b, Tmem179, Tmem200a, Tnc, Tox3, Tril, Trim30a, Trp53cor1, Ttpa, Uty, Vip, Vsig2, Vstm2b, Wif1, Zfp385c, Zfp831, Zic1, Zic2, Zic3, Zic4, and Zic5 or others identified herein.

[0072] In mice, genes whose expression can be used as markers to identify granule cells or their nuclei include 2510003B16Rik, 4930449E18Rik, 6430548M08Rik, 6530403H02Rik, 9230112J17Rik, Ablim1, Ablim3, Ackr1, Adamts18, Adcy1, Airn, Ak4, Als2, Atp1a1, Atp2b3, Barhl2, Bcl2l15, Boc, Brinp1, Bsn, Btg1, C130030K03Rik, Cacna1c, Cacna1e, Cacna1c, Cacna1d, Cacna1e ... 2d1, Cadm3, Cadps2, Calb2, Caln1, Camk4, Camkk2, Car10, Car15, Car4, Cbln1, Cbln3, Ccdc120, Cd300a, Cdh15, Cdh7, Cdh8, Celf4, Cerkl, Cers6, Chgb, Chn2, Chrd, Cnksr2, Cnnm4, Cntn6, Cntnap1, Cntnap4, Cplx2, Crhr1, Crtam, Dab2ip, Des, Dgkd, Diras2, Dlgap3, Dpp6, Dpysl4, Dusp5, Eif4e3, Epb41, E pb41l4b, Epha3, Erc1, Ets2, Etv1, Exph5, Faap20, Fam135b, Fat2, Fgf14, Fzd7, Gabra6, Gabrb3, Gabrd, Galnt13, Gap43, Ggact, Gm2694, Gm7854, Gm989 9, Gm996, Golga7b, Gprc5c, Grik2, Grin2a, Grin2c, Grm4, Ica1l, Igfbp5, Il16, Il20rb, Inadl, Itga7, Jph3, Jph4, Kank2, Kbtbd11, Kcnc1, Kcnd2, Kcnh 1, Kcnip4, Kcnj12, Kcnj6, Kcnk1, Kcnk10, Kcnk12, Kcnk3, Kcnk9, Kcnt1, Khnyn, Kif26b, Lgi1, Lhfp, Lin7a, Lrrc3b, Lurap1, Maml3, Mapk11, Mapk12, Ma rveld2, Mctp1, Mir670, Mmp24, Mpp4, Msra, Ndrg3, Ndst3, Nedd4l, Negr1, Neto2, Neurl1a, Neurod1, Nhsl2, Nrep, Nrip2, Nrn1, Ntn4, Nxn, Olfm1, Olfm3,Pagr1a, Panx2, Pclo, Pcsk2, Pde10a, Pde3b, Pgm2l1, Pkib, Plch2, Plcl2, Plcxd3, Pld5, Ppargc1b, Ppfia4, P qlc3, Prkce, Prr16, Prss52, Ptchd1, Ptpn22, Pxylp1, Rab15, Rab37, Ralyl, Rapgef4, Rbfox1, Rbfox3, Rcan3, Reln, Rims1, Rims2, Rims3, Rnf112, Rnf152, Ror1, Rps6ka1, Rtn4r, Rtn4rl1, Ryr2, Scn2a1, Sel1l3, Sema4f, S ergef, Sgpp2, Sgsm1, Sh3gl2, Shf, Shisa8, Sidt1, Slc16a10, Slc17a7, Slc25a22, Slc29a4, Slc2a13, Slc35f3 , Slc8a2, Slc9a5, Slit3, Slitrk4, Snap25, Snrk, Specc1, Speg, Sphkap, Sptb, St3gal5, Stap2, Stmn2, Stxbp 1, Stxbp5l, Sv2b, Svep1, Svop, Syndig1l, Syt1, Syt12, Syt4, Tas2r143, Tbata, Tbc1d8, Tenm1, Tesc, Tiam1, T Genes with the greatest changes in expression compared to normalized expression in unsorted nuclei include ll1, Tmem163, Tmem178, Tmem181c-ps, Tmem266, Tmem56, Tmem63c, Tmtc4, Trhde, Ubtd2, Unc13c, Uncx, Usp3, Vat1l, Vsnl1, Wscd2, Xkr6, Xkr7, Zdhhc2, Zfp385b, and Zic2.

[0073] In mice, genes whose expression can be used as markers to identify Purkinje or their nuclei include 1190002N15Rik, 1700008O03Rik, 1700124L16Rik, 2410124H12Rik, 9530052E02Rik, 9530059O14Rik, A330050F15Rik, Abhd3, Adam23, Adamts3, Adgrl2, Aff1, Ank1, Ankrd33b, Ankrd53, Anks1b, Apip, Arap1, Arhgap20, Arhgap26, Arhgef33, Arnt l, Atl2, Atp2a3, Atp2b2, Atp6ap1l, Atp6v1b1, Atrnl1, Auts2, B3gnt5, Baiap2, Bcl11a, Bean1, Brinp2, Bzrap1, Cacna1a, Cacna1g, Cacna2d2, Calb1, C amk2a, Car7, Car8, Casq2, Ccdc107, Ccdc85a, Cck, Cds1, Cep126, Cep76, Clec2l, Clic6, Clstn3, Cntn3, Cntnap5b, Cntnap5c, Col18a1, Corin, Cpeb1, C pne9, Csta1, Cyth3, D430036J16Rik, Dach1, Dagla, Defb23, Dgkg, Dgkh, ​​Dgkz, Dlg2, Dlx4os, Dmd, Dner, Doc2b, Dpp10, Ebf1, Ebf2, Ece1, Etl4, Fam107b , Fam117a, Fam174b, Fam184b, Fam21, Fam78b, Far2, Fbxl21, Fgd3, Fhl5, Flt3, Fmnl1, Foxp2, Foxp4, Frmpd3, Gabbr2, Garnl3, Gfra2, Gm14169, Gm5083, Gm5803, Gng13, Gpr63, Grid2, Grid2ip, Grik1, Grip1, Grm1, H2-D1, Heatr5a, Hes3, Homer3, Hpcal1, Hrh2, Hspa12a, Htr1b, Icmt, Id2, Ifngr2, Il22, Il tifb, Inpp4a, Inpp5a, Itga3, Itpka, Itpr1, Kalrn, Kcnab1, Kcnab2, Kcnh7, Kcnip1, Kcnma1, Kctd12, Kitl, Ksr2, L3mbtl1, Large, Ldhd, Lhx1, Lhx1os,Lhx5, Linc-md1, Lpcat4, Lrfn5, Lrp8, Lrrfip1, Lsmem1, Magoh, Mcemp1, Mdfi, Mir124a-1hg, Mir138-1, Mir3470b, Mtss1, Myh10, Myo10, Nefh, Nefm, Nek2, Nell1, Nexn, Ninl, Nkiras2, Npr1, Nr2f2, Nrk, Nsg1, Nup93, Ociad2, Orai2, Paxbp1, Pcp2, Pcp4, Pde2a, Pde5a, Pde9a, Pih1h3b, Plcb3, Plekhd1, Plekhd1os, Plpp6, Plxdc1, Pnpla3, Polr1b, P pp1r17, Ppp4r4, Prkag3, Prkcg, Prkg1, Prmt8, Psd2, Ptchd4, Ptprr, Rab43, Rbms1, Reep2, Retn, Rgs8, Rmdn3, Rnf19 b, Rreb1, Rtel1, Ryr1, Scn4b, Serinc2, Sestd1, Sh2d4b, Shank1, Shisa6, Skor2, Slc16a9, Slc1a6, Slc20a1, Slc9a 3, Smpx, Snhg11, Sorl1, Sos1, Sox5os3, Sptan1, Sptbn2, Stac, Steap2, Stil, Stk17b, Strip2, Strn3, Stxbp2, Svil, Genes with the greatest changes in expression compared to normalized expression in unsorted nuclei include Sycp1, Syndig1, Syt7, Tenm4, Tex261, Thy1, Tm6sf1, Tmem255a, Trabd2b, Trpc3, Tspan11, Ttll5, Tuba8, Unc5d, Vax2, Vwa5b2, Ywhah, Zbtb46, Zfand4, Zfp385a, Zfp385c, and Znhit1.

[0074] In mice, genes whose expression can be used as markers to identify basket cells or their nuclei include 1700024F13Rik, 1700080N15Rik, 9530026P05Rik, A530058N18Rik, Acot10, Adamts15, Adamts16, Adamts2, Adarb2, Adgrl3, Adrb2, Agl, AI504432, Aldh18a1, Aldh1l2, Aldob, Ankrd13b, Ar, Arl4a, Arl4c, Arrdc1, Asgr1, Asic2, Asl, Asns, At p1b1, Atp2b1, Bex1, Bhlhe22, Btbd11, Bzw2, C1qtnf4, Cabp1, Cacna1b, Cacna1d, Cacna2d3, Cacng2, Camk1g, Camk2n1, Cars, Ccdc136, Ccdc74a, Cd59a , Cdkl1, Cebpb, Cep70, Cgn, Chac1, Chchd7, Chd3os, Chst1, Chst9, Clmp, Cnpy1, Cnr1, Cntn4, Coro6, Cryba2, Csmd1, Csrnp3, Cx3cl1, Cyld, Cyp4x1, Cyp 7b1, Dnah11, Dock7, Efna5, Elavl4, Elmod1, Eml5, Enpp1, Ephb2, Erc2, Esrrg, Fam84a, Fev, Fgf13, Flrt2, Frmd3, Fuca2, Gabra1, Gabrg2, Gad1, Gad2, Galnt18, Gars, Gdpd5, Gfra4, Gjd2, Glce, Gldc, Gm11780, Gm14204, Gm15631, Gm15663, Gm2762, Gpr12, Gprasp2, Grb10, Gria3, Grid1, Grik3, Grin2d, G rm8, H2afz, H2-Bl, Haus4, Hcn1, Hecw1, Hnrnpa0, Hpca, Hpcal4, Hunk, Igsf3, Inhba, Inpp4b, Kcna2, Kcnab3, Kcns2, Kirrel, Kirrel3, Kit, Klhl33, Lam b1, Lrrc38, Lsm11, Ly6e, Lypd6, Maged1, Man2a1, March11, Mgat4c, Mgat5, Mir193b, Mir384, Mkx, Mpp7, Msantd4, Mtfp1, Mthfd2, Myo16, Myrip, Myt1l,Nap1l2, Nars, Ndn, Nefl, Neurod2, Neurod6, Nln, Nmt2, Nos1, Nos1ap, Nrxn3, Ntn1, Osbpl10, Osbpl5, Pam, Pa rm1, Penk, Pfkp, Pgrmc1, Phactr2, Phf24, Phyhip, Pja1, Pkp4, Pla2g4e, Plch1, Plcxd2, Plppr4, Plvap, Plxna 4, Plxnc1, Pnck, Ppp1cc, Prdm8, Prkcd, Prrg3, Ptchd2, Purb, Rab10os, Rab3c, Rai2, Rasgrf1, Rasl10b, Rec8 , Ret, Rmst, Rora, Rph3a, Rpl15, Scn9a, Sdk2, Sema3e, Serpini1, Sesn2, Sez6, Sgk1, Sgtb, Shank3, Skor1, Slc 12a5, Slc19a2, Slc1a4, Slc24a3, Slc2a3, Slc32a1, Slc4a8, Slc6a7, Slc7a1, Slc7a3, Snph, Socs2, Sod2, Sor bs2, Sorcs3, Sp5, Stac2, Stard5, Stc2, Syt13, Tbc1d30, Tbc1d4, Tfap2a, Tfap2b, Tmem117, Tmem132e, Tmem15 Genes with the greatest changes in expression compared to normalized expression in unsorted nuclei include: 1b, Tmem169, Tmem25, Tram1l1, Trib3, Trim67, Tro, Trpc5, Trpc7, Tspyl4, Ttc39b, Ttc9, Unc5c, Usp1, Vmn2r2, Wasf1, Wbscr17, Wfs1, Xkr4, Zbtb18, Zcchc16, and Zfp423.

[0075] In mice, genes whose expression can be used as markers to identify astrocytes or their nuclei include 1700066O22Rik, 2810032G03Rik, 3110082J24Rik, 9330188P03Rik, A2m, A330033J07Rik, Abi3bp, Acot11, Acsbg1, Actr3b, Adra1a, AI464131, Akt2, Aldh1l1, Alms1-ps2, Apoe, Aqp4, Arrb1, Atp13a4, Atp1a2, Atp1b2, AU022751, Axl, Baalc, Bcar 3, Btbd17, Cables1, Cacnb1, Cacng5, Cacng8, Camk1, Camk2g, Caskin1, Cbs, Ccdc190, Ccdc78, Cd70, Cdc42ep1, Cdc42ep4, Cdh22, Cdh4, Celrr, Clcn2, C lu, Cml5, Cmya5, Cnn3, Cpne2, Ctxn3, Cxcl14, Cyp2d22, Cyp4f14, D330050G23Rik, Dao, Ddah1, Ddo, Dmp1, E030003E18Rik, Ednrb, Efhd2, Efnb2, Egfr, E lmo2, Elovl2, Emid1, Erich5, Etnppl, Etv4, Eva1a, F830045P16Rik, Fads2, Fam107a, Fam173a, Fam20a, Fam43a, Fam92b, Fgd6, Fgfr3, Frem1, Frmpd1, F xyd1, Fxyd7, Fzd1, Gabra2, Gabra4, Gabrb1, Gareml, Gas2l3, Gdf10, Gja1, Gjb6, Gli1, Gli2, Glipr2, Gm5089, Gm6277, Gpr153, Greb1l, Gria1, Gsap, Gs tm1, Hif3a, Hk2, Hopx, Hrh1, Htra3, Icosl, Id4, Igdcc3, Igdcc4, Itih3, Itpkb, Kifc3, Klf15, Lama2, Lcat, Lfng, Lgi4, Lgr6, Lrig1, Lrp3, Lrrc8a, Man 1c1, Map2k6, Mertk, Metrn, Mfge8, Micalcl, Mir3093, Mir6375, Mir6390, Mlc1, Mmd2, Mpp6, Mro, Msi1, Msx2, Mt1, Mt2, Mt3, Mxra8, Mybpc1, N4bp3, Nat8,Nbl1, Ndrg2, Nhsl1, Nkain4, Nr1d1, Ntsr2, Nudt14, Nwd1, Oplah, Paqr6, Paqr8, Pax3, Pbxip1, Pde8b, Pdlim 4, Phf21b, Phka1, Phkg1, Phyhd1, Pitpnc1, Pitpnm2, Pla2g7, Plce1, Plekho2, Plpp3, Plscr4, Pltp, Plxnb1, Pnky, Pnpla7, Por, Ppm1m, Prex2, Proca1, Prodh, Prr5, Ptch1, Ptch2, Pxmp2, Pyroxd2, Rab34, Rab36, Ramp2 , Rasl11a, Rbm24, Rdh5, Rfx2, Rfx4, Rgl1, Rgma, Rpph1, Rpusd3, Rras, S1pr1, Sdc4, Sfxn1, Sfxn5, Sh3pxd2b, Shisa9, Sirpa, Slc12a4, Slc13a5, Slc1a2, Slc1a3, Slc25a18, Slc27a1, Slc38a1, Slc38a3, Slc39a12, Slc4 1a1, Slc4a4, Slc7a10, Slc8b1, Slco4a1, Smad3, Smim23, Sned1, Sod3, Sox9, Sparc, Sparcl1, Stk32a, Synpo2 Genes with the greatest changes in expression compared to normalized expression in unsorted nuclei include: Syt10, Tapbpl, Tcf7l1, Tmem198b, Tnc, Tom1l1, Traf1, Trib2, Tril, Trp53cor1, Ttll3, Ttll8, Ttyh1, Ttyh3, Ucp2, Vim, Wwc1, Zfp467, Zfp641, Zic1, Zic4, and Zim1.

[0076] In mice, genes whose expression can be used as markers to identify oligodendrocytes or their nuclei include 1500015L24Rik, 1700063D05Rik, 4930474G06Rik, 4933413G19Rik, 5031410I06Rik, 5031439G07Rik, 5430431A17Rik, 9330111N05Rik, 9330117O12Rik, A230001M10Rik, A330049N07Rik, Aatk, Abca8a, Adamtsl4, Adap1, Adipor2, Agpat4, and Ankub. 1, Ano4, Apod, Arc, Arhgap23, Arhgef10, Arl4d, Arrdc2, Arrdc3, Arsg, Aspa, Asphd1, B3galt5, Bche, Bcl2l1, Bicd2, Bin1, Bpgm, C030029H02Rik, C630 043F03Rik, Car14, Car2, Carns1, Ccdc152, Ccp110, Cdh19, Cdk18, Cdk19, Cep97, Cers2, Chdh, Cldn11, Cnp, Cntn2, Cpm, Cpox, Creb5, Cryab, Cyp2j12, D 16Ertd472e, D1Ertd622e, D7Ertd443e, Daam2, Ddit4, Ddr1, Degs1, Depdc7, Dnajb2, Dock10, Dock5, Dpy19l1, Dusp26, Edil3, Efnb3, Elovl7, Endod1, E npp2, Enpp4, Erbb2ip, Erbb3, Ermn, Fa2h, Fam134b, Fam46a, Fbxo32, Fbxw15, Folh1, Frmd4b, Fth1, Gab1, Galnt5, Galnt6, Gatm, Gbp11, Gjb1, Gjc2, Glu l, Gm10471, Gm1979, Gm21671, Gm5535, Gm5862, Gm7361, Gm9895, Gng11, Gng8, Gpr37, Gpt, Grm3, Gstm7, Hapln2, Hebp1, I730030J21Rik, Il33, Inf2, Ins c, Josd2, Kcnj2, Kcnk13, Kctd13, Kctd3, Kif13b, Klhl2, Klk6, Kndc1, Lap3, Larp6, Lipa, Litaf, Lrrc8b, Lzts2, Mag, Mal, Map6d1, Map7, Mast3, Mast4,Mboat1, Mbp, Mcam, Mobp, Mog, Myo1d, Myrf, Ndrg1, Neat1, Nfe2l3, Nhlrc4, Nipal4, Nkain1, Nkain2, Nmral1, Olfml1, Opa lin, Osbpl7, Pacrg, Pacs2, Pak7, Pde4b, Pde8a, Pex5l, Phlpp1, Piga, Pigz, Pik3c2b, Pim3, Pip4k2a, Pkd2l1, Pla2g16, Pl cl1, Plekhg3, Plekhh1, Plp1, Pls1, Plxnb3, Ppfibp2, Ppp1r14a, Prickle2, Prima1, Prox1, Prr18, Prr5l, Prrg1, Psat1, P stpip2, Ptgds, Ptma, Ptprd, Qdpr, Rasgrp3, Rcbtb1, Rffl, Rhob, Rhou, Rnf220, Rnf7, Rtn4, S1pr5, Sall1, Sccpdh, Sec11c , Sec14l5, Serpinb1a, Sez6l2, Sgk2, Sh3tc2, Shisa4, Shtn1, Slain1, Slc12a2, Slc25a13, Slc38a2, Slc48a1, Slc7a15, Sm ad7, Sox2ot, Speer4a, Spg20, St18, St6galnac3, Stmn1, Stmn4, Stxbp6, Synj2, Sytl2, Tbc1d5, Thumpd1, Tmeff1, Tmeff2, Genes with the greatest changes in expression compared to normalized expression in unsorted nuclei include Tmem125, Tmem151a, Tmem229a, Tmem258, Tmem63a, Tmem88b, Tmem98, Tmprss5, Tnfaip6, Tnni1, Tppp, Trf, Trim36, Trim59, Tspan2, Ttll7, Tubb4a, Ugt8a, Usp31, Usp54, Vmp1, Xaf1, and Zdhhc20.

[0077] In mice, genes whose expression can be used as markers to identify oligodendrocyte precursor cells (OPCs) or their nuclei include 1700086L19Rik, 1810041L15Rik, 5730559C18Rik, 6330403K07Rik, 8030451O07Rik, A230077H06Rik, A730017C20Rik, Abcg2, Abhd2, Abtb2, Adam12, Adm, Alcam, Amz1, Arhgap24, Asap3, Ascl1, B3gat2, B9d1, Basp1, Bcas1, and Bmp4. , Brinp3, C1ql1, C1ql2, Cacng4, Calcrl, Cav1, Cav2, Ccnd1, Cd200, Cdh13, Cdh24, Cdo1, Cfap20, Chadl, Chst11, Chst2, Chst3, Chst5, Cobl, Col11a2, C ol27a1, Csgalnact1, Csnk2a2, Cspg4, Cspg5, Ctxn1, Cysrt1, Dbpht2, Dcc, Dpm3, Dpysl3, Dscam, Ebf4, Eid2b, Elfn1, Elfn2, Enpp6, Epb41l2, Epn2, Eya 1, Fam19a2, Fam3c, Fam89a, Fbn2, Fdps, Fhl3, Frrs1, Ftsj2, Fzd9, G0s2, Galnt3, Gfpt2, Gfra1, Gng4, Grm5, Gsx1, H2-K1, Has2, Hip1r, Hrk, Ick, Ifitm 7, Igfbp3, Itga9, Itpr2, Kcnh5, Kcnh8, Khdrbs3, Klhl5, Lbh, Lhfpl3, Limd1, Lims2, Lmcd1, Lpcat2, Lrfn2, Lrrc4c, Lrrtm4, Marcks, Matn1, Matn4, Mex 3b, Mfsd2a, Miat, Mir17, Mmp15, Mmp16, Mmp2, Mycl, Myt1, Ncan, Neto1, Neu4, Nova1, Nxph1, Olfm2, Olig2, Opcml, Oprl1, Pcdh15, Pcdh17, Pcdh7, Pdgfr a, Pdzd2, Pfn2, Phlda1, Pmepa1, Ppfibp1, Ppp2r2b, Prr36, Prss23, Prss48, Ptgfrn, Ptpro, Pxdc1, Qpct, Rapgef3, Rasgef1b, Rcn1, Rep15, Rgcc, Rin2,Rlbp1, Rnd2, Rnf122, Rpl13, Rprm, S100a3, S100a4, S1pr2, Sapcd2, Scel, Scrg1, Sema3d, Sema5a, Sema5b, Serpine2, Sertm 1, Sh3bp4, Sh3rf1, Slc1a1, Slc22a23, Slc35f1, Slc44a5, Slitrk3, Slitrk5, Slitrk6, Snhg5, Snx1, Snx22, Sox10, Sox4, Sox Genes with the greatest changes in expression compared to normalized expression in unsorted nuclei include Spry4, Sstr1, Sulf2, Susd5, Tac1, Tes, Tgfa, Tmem100, Tmem132b, Tmem176b, Tnr, Tns3, Tox3, Tspan6, Vash1, Vcan, Vmn1r55, Vpreb1, Vstm2b, Vwc2, Xylt1, Zc3hav1l, Zdhhc23, Zfp365, and Zfp488.

[0078] In some embodiments, one, two, three or more, including four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty-five, thirty, forty, fifty or more sets, subsets of the described markers may be used to identify any one or more nuclei of granule cells, basket cells, astrocytes, oligodendrocytes, and OPCs from either human, mouse, rat, or non-human primates. Hierarchical clustering may be used to visualize the degree of similarity in RNAseq profiles.

[0079] 2. Differences in RNA profiles in sorted nuclei from rats In some embodiments, nuclei from rat tissue are sorted into groups of nuclei from granule cells, Purkinje cells, basket cells, astrocytes, oligodendrocytes and OPCs. Genes that are differentially expressed compared to normalized expression across granule cells, Purkinje cells, astrocytes, oligodendrocytes and OPCs may be used as markers.For example, AABR07001623.1, AABR07001734.1, AABR07006310.1, AABR07006713.1, AABR07006713.2, AABR07006727.1, AABR07013140.1, AABR07026032.2, AABR07030473.1, AABR07030880.1, AABR07046961.1, AABR07047823.1, AABR07049491.1, AABR07052897.1, AABR07058658.1, AABR07061428.1, AABR 07070161.1, AABR07070161.6, AABR07070578.1, Abi3, Aqp9, Arhgef26, Arhgef33, Arl4a, B3gnt5, Bgn, C1ql1, C1ql2, Cacng5, Cadps2, Calb1, Calb2, Car4, Car8, Cav1, Cbln1, Ccnd1, Cdh15, Cdk2, Cep76, Cldn11, Clmp, Col12a1, Col5a3, Cspg4, Cyp2d5, Enpp1, Ermn, Fam107b, Fam89a, Fat2, Gabra6, Ga brd, Gdf10, Gjd2, Gldc, Gli1, Gpr37, Gpr63, Grm8, Hapln2, Hpcal4, Id4, Il16, Itga7, Itga9, Itih3, Itpr1, Kcnh1, Kcnj12, Klk6, Lama3, Lfng, Lgi4, L ims2, Lypd6, Mag, Mal, March11, Mobp, Mog, Msx2, Mt2A, Nkain4, Opalin, Pdgfra, Plch1, Plk2, Plxdc1, Pmel, Pou3f2, Ppfibp1, Ppp1r17, Prr5l, Prrt2 Expression or lack of expression of at least one of the differentially expressed genes selected from: Qdpr, Rasgrp3, Rbfox3, RGD1561557, RGD1561849, Rlbp1, Ryr1, S1pr1, Scara3, Selplg, Serpine2, Slc1a6, Slc25a18, Slc5a11, Slc7a10, Slc9a3, Snx22, Sorcs3, Sv2b, Tesc, Tmem255b, Tmem63a, Tmem88b, Tnc, and Zfp385c may be used as a species- and cell type-specific marker.

[0080] In rats, genes whose expression may be used as markers to identify granule cells or their nuclei include significantly enriched expression of Fat2 and Rbfox3 compared to normalized expression in unsorted nuclei. Genes that may be used as markers to identify Purkinje cells or their nuclei include significantly enriched expression of Car8 and Calb1. Genes that may be used as markers to identify Bergmann glia or their nuclei include significantly enriched expression of Aldh1L1 and Slc1a3. Genes that may be used as markers to identify oligodendrocytes (corticopontin pyramidal cells) or their nuclei include significantly enriched expression of Colgalt2. Genes that may be used as markers to identify Bergmann glia or their nuclei include significantly enriched expression of Hs3st4. Hs3st4 was observed to be specific to corticothalamic pyramidal cells, while Pde1a and Csmd1 were observed to be expressed in oligodendrocytes and OPCs (both cortical pyramidal cell types).

[0081] In some embodiments, expression of Itih3 may be used as a marker for both cerebellar astrocyte type Bergmann glia and cerebellar non-Bergmann glial astrocyte type. Additional markers for the nuclei of the six cell types may include: Tescalcin (Tesc) for granule cell nuclei, Carbonic anhydrase 8 (Car8) for Purkinje cell nuclei, March11 for basket cell nuclei, Inter-Alpha-Trypsin Inhibitor Heavy Chain 3 (Itih3) for astrocyte nuclei, Mog for oligodendrocyte nuclei, and Platelet-Derived Growth Factor Receptor Alpha (Pdgfra) for oligodendrocyte progenitor cell nuclei.

[0082] In some embodiments, marker localization was analyzed to confirm cell type using the Allen Mouse Brain Atlas in situ hybridization database.

[0083] In some embodiments, one, two, three or more, including four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty-five, thirty, forty, fifty or more sets, subsets of the described markers may be used to identify any one or more nuclei of granule cells, basket cells, astrocytes, oligodendrocytes, and OPCs from either human, mouse, rat, or non-human primates. Hierarchical clustering may be used to visualize the degree of similarity in RNAseq profiles.

[0084] 3. Differences in RNA profiles in selected nuclei from humans Analysis of samples from 16 human postmortem brains revealed that the specific molecular consequences of ageing, sex, autolysis time of tissue samples, and stress differed between cell types, but in each case the expression of genes involved in synapse development and maintenance was reduced. Robust, cell type-specific molecular pathways indicated pathophysiological responses that occurred in the brains of three of the 16 donors. Cerebellar granule cells from the three brain samples were observed to show a robust induction of a set of 224 genes, including many immediate early genes, and were enriched for GO categories (protein folding / refolding, apoptosis, transcriptional response to stress, response to external stimuli, ATPase activity, etc.), indicating acute responses to several external influences, but the common clinical condition of these donors remains unclear. As mentioned in the examples, no correlation with autolysis time and no induction of glial markers, indicative of stroke or brain injury, was found. Although activity-dependent gene expression changes have been extensively characterized in cultured mouse granule cells, the responses identified in these three human samples were clearly distinct, albeit overlapping. The results herein provide a guidepost for experimental studies of model systems or iPS cells aimed at identifying the signals that induce these responses in the human brain. Molecular profiling using the cell profiling techniques described herein can be applied to any species and provides a means for the study of molecular events associated with human cell type function and dysfunction.

[0085] In some embodiments, increased expression of Rbfox3 and Fat2 may be used as a marker for human granule cell-derived nuclei. Alternatively, increased expression of March11 may be used as a marker for human basket / astrocyte-derived nuclei. Alternatively, increased expression of Aldh1a1 and Slc1a3 may be used as a marker for human astrocyte nuclei. Alternatively, increased expression of Mog may be used as a marker for mature oligodendrocytes. Alternatively, increased expression of Pdgfra and chondroitin sulfate proteoglycan 4 (Cspg4) may be used as a marker for OPCs.

[0086] In some embodiments, one, two, three or more, including four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty-five, thirty, forty, fifty or more sets, subsets of the described markers may be used to identify any one or more nuclei of granule cells, basket cells, astrocytes, oligodendrocytes, and OPCs from either human, mouse, rat, or non-human primates. Hierarchical clustering may be used to visualize the degree of similarity in RNAseq profiles.

[0087] C. Direct Determination of Human Cell Types The cell profiling techniques described herein may also be used independently of the cell types characterized in mice. This is because there may be differences between neurons from mice and humans, and human-specific neuronal subtypes may exist. Examination of the transcriptional profile of unsorted nuclei from regions of interest identified putative cell type-specific genes for sorting. In mice, chromosome-associated transcripts (CAT) and polyA+ transcripts observed to be expressed at low levels in unsorted nuclei were often expressed at high levels in one or several cell types. Species differences were explored for different cell types, and the findings were supported by in situ hybridization and immunofluorescence. Human nuclear RNA from specific cell types was profiled using the same strategy as the cell profiling methods described herein.

[0088] IV. Comparative analysis of nuclear profiles Nuclei isolation and transcriptional profiling of multiple cell types from mouse, rat, and human are described herein. For example, cell types originate from basal ganglia, thalamus, cerebellum, or hippocampus tissue. Cell types originating from the cerebellum include Purkinje cells, granule cells, Bergmann glia, basket / astrocytes, astrocytes, oligodendrocytes, and deep cerebellar nuclei cells. These samples were sequenced for analysis to determine differences between individuals and across species. In some embodiments, results are confirmed by in situ hybridization using tissue sections.

[0089] The impetus for the development of accurate and efficient non-genetic methods to characterize the molecular properties of specific CNS cell types is to allow direct interrogation of human biology. Major differences are observed between gene expression in the same cell types between mice and humans. Given the explosion of research identifying the genetic causes of human disease and the striking demonstration in both animal models and humans that these may involve lesions with simple changes in gene dosage, exploration of the details of cell types in normal and diseased brains is essential. Thus, comparative analysis of data showing human cell type-specific properties obtained in experimental systems using the cell profiling techniques described herein will advance our understanding of a wide range of human disorders that affect CNS circuits.

[0090] A recent evolutionary model (Arendt, D. (2008). Nat Rev Genet 9, 868-882 and Arendt, et al. (2016) Nat Rev Genet 17, 744-757, each of which is incorporated herein by reference in its entirety) was a consensus definition of cell types that helped to consider the data generated using the cell profiling techniques described herein. According to this model, certain properties of homologous cell types can change as long as they remain defined by their unique shared regulatory apparatus. For example, Purkinje cell, granule cell, or astrocyte cell gene expression profiles can change between species or within individual cell types without losing their cell type uniqueness. This definition fits both functional changes of cell types between species and expression of altered genes within cell types due to mutations in cis-regulatory sequences. The data herein demonstrate major differences in the expression of orthologous genes in each cell type between rodent and human brains. These hundreds of differences are cell type specific, with most of these genes remaining expressed in granule neurons compared with other regions of the brain. These data indicate that the fine-tuned biochemistry of homologous human and mouse CNS cell types differs markedly.

[0091] The results provided here provide two major insights into the evolution of the mammalian brain. First, while the results are consistent with previously published examples of cis-regulatory sequence divergence as a mechanism of evolution (Maricic, et al. (2012) Molecular Biology and Evolution 30, 844-852; Prud'homme, et al. (2006) Nature 440, 1050-1053; Weyer et al., Mol Biol Evol. 2016 Feb; 33(2): 316-322, 2015; each of which is incorporated herein by reference in its entirety), the lack of highly enriched GO categories for most of the cell-specific events demonstrated herein strongly suggests that regulatory changes are a major driver of phenotypic variation between homologous cell types in the mammalian brain. This is consistent with the evolutionary model presented above. This is because it allows substantial evolution of the functional properties of cells, while at the same time maintaining the core regulatory complexes (CoRCs) that define a given cell type (Arendt, et al. (2016) Nat Rev Genet 17, 744-757, which is incorporated herein by reference in its entirety). Second, the expression differences demonstrated here are large, and the analysis was limited to high-confidence orthologous genes. This data complements previous studies of progenitor cells that reported considerable changes in expression of human genes lacking mouse orthologs (Florio et al, Science. 2015 Mar 27;347(6229):1465-70, which is incorporated herein by reference in its entirety). Taken together, these data sets demonstrate that expression differences between mouse and human cell types are entirely cell type specific and significant from a phenotypic perspective.

[0092] In certain embodiments of profiling human samples using the methods herein, correlation coefficients are (r=0.98-1.0) between samples, indicating that results are highly reproducible and that autolysis time has a small effect on nuclear RNA expression profiles. The initial indication that sex and age differentially affect gene expression in CNS cell types indicates that the cell profiling techniques described herein are sufficiently accurate for detailed investigation of both sexually dimorphic human behavior and aging of specific brain circuits. For example, comparative analysis of human brains by the cell profiling techniques described herein allows insight into the effects of aging on cell types selectively vulnerable to late-onset human disorders.

[0093] A. Identifying cell type-specific functions that differ between species Although the organization of cells in the cerebellum is similar between mice and humans, many differences remain. For example, Purkinje cells are much more widely distributed in human cells compared to mouse cells, and NeuN cells are spread throughout the Purkinje layer, which is evidence that the human cerebellum has a different organization compared to mice. Furthermore, the size of Purkinje cell bodies and nuclei is larger in humans compared to mice. In some embodiments, RNAseq can be used to compare gene expression between species to identify species-specific markers. For example, protein tyrosine kinase 2 beta (Ptk2b), solute carrier family 43 member 2 (Slc43a2), glutamate ionotropic receptor kainate type subunit 3 (Grik3), Itpr1, protocadherin 9 (Pcdh9), and RAS guanyl-releasing protein 1 (Rasgrp1) nuclei from granule, stellate, and glial cells can be used to distinguish between human and mouse nuclei. Because the cerebellum is one of the most conserved brain structures across species, it is more likely that more structural and molecular differences will be observed in other brain regions.

[0094] Although samples from the same species may show heterogeneity, the most specific genes for each cell type of different species can be used as the markers specific to that cell type.In some embodiments, the expression of two or more genes can be used to identify a particular cell type.Compared to the normalized expression of all cell types, such as granule cells, Purkinje cells, astrocytes, oligodendrocytes, and OPCs, differentially expressed genes can be removed from the analysis to identify at least one marker for the cell type of interest.

[0095] In some embodiments, the most specific genes are calculated using the specificity index (SI) algorithm described herein. For example, the most specific genes (in order of decreasing specificity) in granule cell nuclei from human samples are Cadps2, Reln, Synpr, Galnt5, Ccm2l, Srrm4, Tmem266, Cckbr, Cerkl, Chn2, Grik2, Mal2, Fat2, Adamts16, Lcn8, Fstl5, Tll1, Vsnl1, Rit2, Slc8a2, Slc17a7, Zp 2, Tspan18, Fgf5, Prag1, Slc6a7, Rims1, Cdh7, L1cam, Mical2, Ptk2b, Kcnh1, Arhgap29, Cdk15, Pde3b, I tga4, Etv1, Tmem178, Chgb, Col13a1, Rnf152, Kcnk1, Il16, Sstr2, Als2, Barhl1, Scn2a, Cbln1, Dusp5, N dst3, Lurap1l, Kcnj12, Snai2, Ifnlr1, Neurod1, Cdyl2, Ppp1r1c, Jkamp, ​​Neurl1a, Trhde, Mfsd9, Xkr7, Barhl2, Cbln3, Uncx, Ptchd1, Cblb, Runx1t1, Sphkap, Gabra6, Coro2a, Bmper, Dkk4, Camk4, Unc13c, Trp 53i11, Ephb1, Glb1l3, Mdga1, Kcnc1, Tmem252, Mpp7, Hipk4, Tmem51, Cacnb3, Cyb5a, Tmem2, 6430573F11Rik, Rasgrp1, Rab37, Adarb1, Svop, Pgm5, Mthfd1l, Cdh15, Vwc2, Disp2, Tspan15, Slc26a5, and Lhfp.For example, the most specific genes (in order of decreasing specificity) in granule cell nuclei from mouse samples were Gabra6, Grin2c, Fat2, Cadps2, Ptpn22, Chn2, Neurod1, Calb2, Trhde, Il16, Ppfia4, Cbln1, Sidt1, Tmem266, Gabrd, Reln, Kcnj12, Gprc5c, Rims3, Tll1, Cb ln3, Olfm3, Slc17a7, Chgb, Kcnh1, Cdh15, Barhl2, Dusp5, Kcnc1, Camk4, Des, Cerkl, Svep1, Uncx, Slc8 a2, Cacna2d1, Pxylp1, Ptchd1, Sel1l3, Rtn4r, Cacna1e, Patj, Atp2b3, Marveld2, Kcnk10, Rab15, Rapge f4, Rims1, Tmem178, Ak4, Lin7a, Igfbp5, Tesc, Neurl1a, Vat1l, Msra, Vsnl1, Rps6ka1, Lhfp, Kcnk3, Ry r2, Exph5, Kif26b, Adamts18, Cntnap4, Tspan18, Mmp24, Cdh8, Ablim3, Cadm3, Ntf3, Pld5, Syndig1l, G alnt15, Etv1, Sh2d1a, Scn2a, Panx2, Speg, Bsn, Rnf112, Il20rb, Slc16a10, Ntn4, Adcy1, Mapk12, Boc, Syt12, Fstl5, Tead4, Clcn1, Pde3b, Pgm2l1, Tmtc4, Abcc8, Olfm1, Nedd4l, Lurap1, Krt24, and Als2.For example, the most specific genes (in order of decreasing specificity) in granule cell nuclei from rat samples were Tesc, Itga7, Rbfox3, AABR07013140.1, Calb2, Kcnh1, Cdh15, AABR07030880.1, Prrt2, Gabrd, Il16, Scara3, Cbln1, AABR07026032.2, Cadps2, Sv2b, Fat2, Gabra6, Car4, Kcnj12, and Nr ep, AABR07043098.3, Rtn4r, Shisa8, Sel1l3, Grb7, Rgl3, Dusp5, Chn2, Lamp5, Pagr1, Cemip, Gprc5c, Shf, Wdr66, Tm em44, Reln, Tll1, Slc17a7, Pax6, Capn1, Calhm3, AABR07043098.1, Camk4, Selm, Sema3g, LOC100360491, Diras2, Ag bl2, Mctp1, Mvp, Tmem51, Vsnl1, Plk5, Mrgprf, LOC100910401, Ak4, Wscd2, Grin2a, Cfap74, Rnf39, Cbln3, Fzd7, Gol ga7b, AABR07033679.1, Ptpn22, Grin2c, Ablim1, Kcnk1, Arpp21, C1rl, Zscan30, Maf, Rtn4rl1, Barhl2, Kcnk12, Act l6b, AABR07053850.1, Sdc1, Camkk2, Nudcd3, Ccdc155, Cbfa2t3, AABR07037151.1, Etv1, Jph3, Itga11, Rnf182, RGD1560784, Plcxd3, Ablim3, Rims1, Nyx, Uncx, AABR07034445.1, AABR07025051.3, Kcnk10, Kcnk3, Nol3, and Als2.

[0096] For example, the most specific genes (in order of decreasing specificity) in Purkinje cell nuclei from mouse samples were Car8, Calb1, Arhgef33, Atp2a3, Itpr1, Ppp1r17, Slc1a6, Slc9a3, Trpc3, Casq2, Clic6, Plekhd1, Ryr1, Cacna1g, Pcp2, Sycp1, Pde5a, Tspan11, Gpr63, and Ne ll1, Cep76, Bcl11a, Grid2ip, Htr1b, Pet100, Strip2, Plxdc1, Skor2, Dagla, Garnl3, Itpka, Myo10, Gng 13, Kcnip1, Trabd2b, Rnf207, Scn4b, Arhgap20, Grip1, Pde9a, Ppp4r4, Cck, Ebf1, Bean1, Tuba8, Smpx, E bf2, Slc19a1, Ccdc85a, Flt3, Camk2a, Sh2d4b, Homer3, Stk17b, Clec2l, Hrh2, Ksr2, Chtf18, Vax2, Foxp 4, Ankrd33b, Cpne9, Corin, Il22, Prkag3, Cep126, Slc20a1, Akain1, Nrk, Pih1h3b, Inpp5a, Lhx1, Sptbn 2, Paxbp1, Dner, Doc2b, Tspoap1, Gfra2, Fhl5, Dmd, Kcnab2, Serinc2, Car7, Dnah6, Prmt8, Prkcg, Mdfi, Dgkh, ​​Fam117a, Nek2, Fmnl1, Atl2, Cfap161, Lin28b, Cacna2d2, Ptchd4, Pkp3, Shank1, Grid2, and Nefh.For example, the most specific genes (in order of decreasing specificity) in Purkinje cell nuclei from rat samples were Arhgef33, Car8, Slc1a6, Calb1, Ppp1r17, Itpr1, Zfp385c, Ryr1, AABR07030473.1, Slc9a3, AABR07006713.1, AABR07006713.2, RGD1561557, Gpr63, AABR07049491.1, AABR070067271.1, Plxdc1, Cep76, Fam107b, B3g nt5, Far2, AC108261.1, Grid2ip, AABR07072628.1, Kcnip1, AABR07006724.1, Trpc3, Phka2, Car7, Fam117a, Hes3, Casq2, Dgkh, ​​Arhgap 26, Trabd2b, Nell1, Bcl11a, Col18a1, LOC100125362, Ptchd4, Camk2a, Dyrk4, Kitlg, Kcnab1, Nxnl2, Ltk, LOC689927, AABR07056156.1 , Cpne8, Htr1b, AABR07033570.1, Icmt, AABR07047598.1, Cstf2t, Grik1, Itpka, LOC100910996, Hpcal1, Efna5, Myo10, Rgs8, Skor2, Il 22, Slc12a3, Cpne9, AABR07007980.2, Cacna1g, Doc2b, Sptbn2, LOC108351705, AABR07035926.1, AABR07037943.1, Pcp2, AABR0706186 0.1, AABR07024820.1, Impg1, Nkiras2, Krtap12-2, Tmem123, AABR07044009.1, Ptprr, Homer3, Garnl3, Sos1, Kcnab1, Cntn3, LOC304725, Stk17b, AABR07047604.1, Zdhhc23, Scn4b, AC126640.1, Pde9a, Ptprb, Prkcg, Pcp4, AABR07056374.1, Fstl4, Ppp4r4, and Nup93.

[0097] For example, the most specific genes (in order of decreasing specificity) in basket cell nuclei from human samples were Bhlhe22, March11, Kit, Tfap2b, Clmp, Lbx1, Slc38a5, Frmd3, Btbd11, Lipg, Rab3b, Lrrc38, Galnt14, Cnr1, Tcerg1l, Cbln2, Siah3, Stac2, Scgn, Rspo 1, Lmcd1, Lpl, Adra1b, Gad2, Chrna3, Socs2, Trpc6, Kcna3, Syn3, Eml5, Nefl, Fam84a, Trpc5, KIAA164 4, Hrh3, Disp3, Tfap2a, Gjd2, Adamts2, Sorcs3, Ptprk, Timp3, Gjb4, Ehd3, Cxcl12, Actc1, Pnma2, Fam4 3a, Adam11, Dact2, Rgs16, Slc5a8, Exph5, Cfap221, Myo16, Megf10, Cacna1d, Pax2, Efna5, Plch1, Gab rg2, Prrt4, Itga5, Sp5, Nefm, Grik3, Cx3cl1, Fbxo24, Lect2, Nefh, Efnb3, Grk3, Mpz, Syt2, Scrt1, Cd c42ep3, Trim67, Grin2b, Rab3c, Plxna4, Fzd8, Elmod1, Smim17, Slc8a3, Nrip3, Plekhg5, Arl4a, Grm8, Cntnap5a, Maats1, Loxl2, Arap3, Dcx, Nyap2, Skor1, Pvalb, Ppm1j, Slc5a4a, Tmem132e, and Cygb.For example, the most specific genes (in order of decreasing specificity) in basket cell nuclei from mouse samples were Adamts15, March11, Tfap2b, Socs2, Penk, Bhlhe22, Plch1, Adamts16, Adrb2, Frmd3, Clmp, Cacna2d3, Trpc7, Galnt18, Tbc1d4, Asgr1, Kit, Tfap2a, and Fam84. a, Gjd2, Grik3, Lrrc38, Chst9, Flrt2, Chac1, Grm8, Esrrg, Slc6a7, Pla2g4e, Stac2, Cabp1, Phf24, Camk 1g, Btbd11, Stc2, Sorcs3, Lamb1, Trpc5, Coro6, Slc32a1, Kirrel, Cdkl1, Chst1, Asns, Sp5, Cryba2, Ald h1l2, Kcnab3, Tg, Plxna4, Prkcd, Cgn, Tcte1, Trim67, Osbpl10, Serpini1, Slc7a1, Gpr22, Cntn4, Skor1 , Slc7a3, Atp1b1, Kcna2, Cebpb, Parm1, Dock7, Eml5, Cnih3, Asic2, Hpcal4, Adamts2, Masp1, Plxnc1, Ga brg2, Mtfp1, Ntn1, Ret, Zfp423, Nln, Man2a1, Neurod6, Grin2d, Angptl3, Cnr1, Tmem132e, Apcdd1, Rasgrf1, Hcn1, Scnn1g, Ly6e, Pax2, Rab3c, Rph3a, Rai2, Gad2, Mpp7, Arl4a, Igsf3, Tmem179, and Sorbs2.For example, the most specific genes (in order of decreasing specificity) in basket cell nuclei from rat samples were Plch1, AABR07001734.1, Lypd6, Lama3, March11, Gldc, Grm8, AABR07047823.1, AABR07001623.1, Gjd2, Arl4a, AABR07070578.1, Hpcal4, AABR07061428.1, Col12a1, Enpp1, Sorcs3, Clmp, Plk2, Arhgef26, RGD15 61667, Asic2, AABR07026472.1, AC124896.1, Trpc5, AABR07052084.1, AABR07059478.1, Kit, Nt5dc2, AABR07041885.1, Hcn1, Skor1 , Prdm8, Phf24, Esrrg, Grik3, Bhlhe22, Mir346, Tmem200a, AABR07003600.7, Scn9a, Inpp4b, Ace, Kcna2, Adra1d, Slc32a1, Cntnap4, Grid1, Trpc7, Begain, AABR07049535.3, Frmpd4, Tfap2b, Cnr1, AABR07054088.3, ​​Zcchc16, Trpv6, Grb10, Pax2, Nrxn3, Tyms, Grid1, Myt1l, Disp3, AABR07063359.1, Btn1a1, Rn60_10_0565.6, AABR07041972.1, Osbpl10, Six4, Mir384, Tbc1d4, Cabp1, AABR07003600. 2, AABR07057590.1, Adrb2, Cacna1d, Plxnc1, Dusp22, Hecw1, Gdpd4, rno-mir-344a-1, Lgi2, AABR07060291.2, Adarb2, Angpt1, Rspo4, Rn50_7_1158.3, Dcx, AABR07065531.31, Onecut1, Mkx, Rgs4, Kcnc2, Hgf, Fgf13, Itga8, AABR07070578.2, Apcdd1, and Trim15.

[0098] For example, the most specific genes (in order of decreasing specificity) in astrocyte nuclei from human samples were S1pr1, Gja1, Lgi4, Efemp1, Etnppl, Aqp4, Rgma, Lgr6, Lcat, Tril, Serpini2, Pgghg, Slc1a3, Gfap, Aldh1a1, Pla2g5, Fxyd1, Wdr49, Gli1, Nfatc4, Scara 3, Slc4a4, Lama2, Cdc42ep4, Gabrg3, P2ry1, Gabra2, Colec12, Mlc1, Slc7a11, Aldh1l1, Sox9, Lgals3 , Mgst1, Gatsl3, Ahnak, Pipox, Rhoj, Arhgef26, Arhgef28, Srpx, Mt3, Atp1a2, Fxyd7, Lfng, Lamb2, Gdf 10, F3, Sdc4, Hspb1, Hes1, Slc12a4, Mertk, C3, AI464131, Id4, Vit, Fjx1, Ccdc80, Adora2b, Ryr3, Mfg e8, Plekho2, Elovl2, Fam179a, Cebpb, Acss1, Slc13a5, Ednrb, Pltp, Il33, Tlr4, Fam198b, Isyna1, Gl is3, Dao, Fgfr3, Ncan, Msx2, Spatc1, Pdgfrb, Tubb2b, Fat3, Fam167a, Col22a1, Tcf7l1, Ndrg2, Nhsl1, Ddit4l, Pbxip1, Rasl12, Sh3bp2, Aplnr, Robo3, Rubcnl, Rab34, Gtf2a1l, Fgfr2, Kank2, and Rfx4.For example, the most specific genes (in order of decreasing specificity) in astrocyte nuclei from mouse samples were Gdf10, Slc1a3, Fxyd1, Dao, A2m, Lfng, Lgi4, Mybpc1, Cdc42ep4, Slco4a1, Fam20a, Lcat, Acot11, Vim, Nwd1, S1pr1, Tnc, Rbm24, Acsbg1, Ntsr2, Pre x2, Cables1, Gli1, Etnppl, Pla2g7, Phkg1, Nhsl1, Slc14a1, Plekho2, Gjb6, Plce1, Itih3, Sdc4, Me rtk, Prr5, Sparcl1, Id4, Mlc1, Pax3, Gpr153, Abi3bp, Pltp, Fxyd7, Rpusd3, Aqp4, Igdcc3, Fgfr3, Mt 3, Atp1a2, Gabrb1, Ndrg2, Ucp2, Plpp3, Frem1, Elovl2, Gli2, Efhd2, Zic4, Oplah, Cxcl14, Mfge8, S ox9, Atp13a4, AI464131, Pdlim4, Rfx4, Traf1, Dmp1, Gabra4, Gja1, Etv4, Map2k6, Wwc1, Syt10, Spar c, Tril, Slc25a18, Aldh1l1, Eva1a, Mxra8, Lgr6, Rgma, Egfr, Slc39a12, Slc8b1, Glipr2, Axl, Aass , Hk2, Itpkb, Ramp2, Pyroxd2, Gabra2, Tcf7l1, Garem2, Wnt6, Dfna5, Fam92b, Plxnb1, and Vamp1.For example, the most specific genes (in order of decreasing specificity) in astrocyte nuclei from rat samples were AABR07070161.1, Itih3, Gdf10, Gli1, Slc25a18, Nkain4, Tnc, Msx2, S1pr1, Slc7a10, Aqp9, Mt2A, Lgi4, Pou3f2, Id4, Cacng5, Cyp2d5, AABR07058658.1, AABR07070161.6, Lfng, and AABR07058658. 070161.3, Fam107a, Nat8f3, Ctxn3, Aldh1l1, Entpd2, Rbm24, Slc1a3, Mlc1, Cxcl14, Hif3a, Fxyd1, Acsbg1, AABR07070 161.7, Notch3, Il33, Nwd1, Cdc42ep4, Samd9l, Gabrb1, Plpp3, Cxcr4, Rgma, Itm2a, Ednrb, Pdlim4, Sgca, Sncaip, Prod h1, Agt, Slc14a1, AABR07032343.1, AABR07035782.1, Marc1, Fgfr3, AABR07070161.5, AC109886.1, Prr5, Aqp4, Plekh o2, AABR07035780.2, Sh2b2, F3, Sparcl1, Eva1a, Pltp, Gfap, Cyp2d4, Crlf1, Sdc4, Sod3, Dao, Chst1, AABR07001555.1 , AABR07063346.1, Mpp6, Ptch2, AABR07031767.1, AC105604.1, Grin2b, Pygm, Etnppl, Slc15a2, Casp4, Gria1, Slc4a4, Camk2g, Cbs, AABR07035175.1, Ntsr2, Gja1, Aspg, Sfxn5, Ccbe1, Ahnak, Gramd3, Ndrg2, Mt3, Shroom3, and Myh11.

[0099] For example, the most specific genes (in order of decreasing specificity) in oligodendrocyte nuclei from mouse samples were Klk6, Mog, Gldn, Cndp1, Gm20425, Tmem98, Slc5a11, Mag, Carns1, Gpr37, Cdk18, Hhatl, Anln, Enpp2, Folh1, Sec14l5, Myrf, Plp1, Mal, Gjb1, Itga2, Cp m, Plpp2, Cntn2, Ermn, Opalin, Abca8b, Lpar1, Pld1, Sall1, Cercam, Dock5, Pi16, Man2a1, Pex5l, Tme m63a, Tmem125, Gipr, Sh3tc2, D7Ertd443e, Kel, Ldb3, Ccdc152, Cldn11, Synj2, Lrp2, Ninj2, Psrc1, N drg1, Actn2, Sgk2, Trim59, Plekhg3, Larp6, Slc45a3, Cnp, Col4a5, Pde8a, Rhobtb1, D16Ertd472e, Qd pr, Enpp6, Rhou, Sema3c, Gsn, Cyr61, Capn3, Hoxd1, Frmd4b, Tmem144, Galnt6, Pkmyt1, Fbxo32, Pla2g 16, Dpyd, Sh3gl3, Prima1, Rasal1, Fa2h, Fam124a, Treh, Paqr4, Azgp1, Plekhh1, Dapk2, Knop1, Sema3b, Elovl1, Shroom4, Car2, Necab1, Aspa, Gab1, Kif6, Hhip, Ugt8a, Iqgap1, Map6d1, Eln, and Pllp. For example, the most specific genes (in order of decreasing specificity) in oligodendrocyte nuclei from mouse samples were: Gpr37, Pex5l, Hapln2, Apod, Ermn, Prr5l, Aspa, Efnb3, Anln, Mog, Plp1, Plekhh1, Tmem63a, Il33, D7Ertd443e, Sec14l5, Carns1, Tnfaip6, Galnt6, Galnt5, Ndrg1, Cldn11, Gjc2, Pla2g16, Fth1, Opalin, Car2, Dock5, Gm20425, Sall1, Pkd2l1, Kcnk13, Ppp1r14a, Gatm, Serpinb1a, Pde8a, Tnni1, Glul, Tmprss5, Nipal4, Trim59, Mal, Slc12a2,Qdpr, Litaf, Gjb1, Edil3, Pls1, Mboat1, St6galnac3, Gng11, Enpp2, Cpm, Pstpip2, Trim36, Klk6, Myo1d, Tmem9 8, Synj2, 4933413G19Rik, Mag, Insc, Pigh, Pik3c2b, Ninj2, Trpv3, Car14, Tmem125, Map7, Cdk18, Depdc7, Aatk, Myrf, Cnp, Stmn4, Chdh, Nek4, Frmd4b, Ugt8a, Palm2, Arsg, Ccdc152, Efemp1, Piga, Rftn1, Itgb4, Emilin2, Enpp4, Nkain2, Adamtsl4, Hcn2, Plxnb3, Dpy19l1, Arhgap23, Prima1, Gipr, Sh3tc2, Tubb4a, Map6d1, and Kctd13. For example, the most specific genes in oligodendrocyte nuclei from rat samples (in order of decreasing specificity) were Gpr37, Pex5l, Hapln2, Apod, Ermn, Prr5l, Aspa, Efnb3, Anln, Mog, Plp1, Plekhh1, Tmem63a, Il33, D7Ertd443e, Sec14l5, Carns1, Tnfaip6, Galnt6, Galnt5, Ndrg1, Cldn11, Gjc2, Pla2g16, Fth1, Opalin, Car2, Dock5, Gm20425, Sall1, Pkd2l1, Kcnk13, Ppp1r14a, Gatm, Serpinb1a, Pde8a, Tnni1, Glul, Tmprss5, Nipal4, Trim59, Mal, Slc12a2, Qdpr, Litaf, Gjb1, Edil3, Pls1, Mboat1, St6galnac3, Gng11, Enpp2, Cpm, Pstpip2, Trim36, Klk6, Myo1d, Tmem98, Synj2, 4933413G19Rik, Mag, Insc, Pigh, Pik3c2b, Ninj2, Tr pv3, Car14, Tmem125, Map7, Cdk18, Depdc7, Aatk, Myrf, Cnp, Stmn4, Chdh, Nek4, Frmd4b, Ugt8a , Palm2, Arsg, Ccdc152, Efemp1, Piga, Rftn1, Itgb4, Emilin2, Enpp4, Nkain2, Adamtsl4, Hcn2,Plxnb3、Dpy19l1、Arhgap23、Prima1、Gipr、Sh3tc2、Tubb4a、Map6d1、Kctd13、Mog、Opalin、Mobp、Tmem63a、AABR0700631 0.1、Hapln2、Klk6、Qdpr、Selplg、Cldn11、Mal、Ermn、AABR07046961.1、Abi3、Gpr37、Tmem88b、Rasgrp3、Prr5l、Slc5a11 、Mag、LOC100362909、Tmem176b、LOC100302465、Cx3cr1、Ctss、Carns1、Tmem125、Tmem176a、Sec14l5、Tf、C1qb、Apod、S1 pr5、Rn60_5_1374.5、Plp1、Insc、Mbp、Pls1、Tmem98、Anln、Ndrg1、Pafah1B1、Pex5l、LOC361016、Galnt6、Trim36、AABR0 7017693.1、Gpatch4、Plekhg3、Hamp、Csf1r、Gjc2、Car14、Trim59、AABR07053870.1、Gpr84、Plekhh1、Ghr、Kndc1、Slco2 b1、Cd74、Aspa、Trpv3、AABR07008030.1、Aplp1、Tgfbr2、Gjb1、Ptp4a3、Prima1、Pik3c2b、AABR07036035.1、Sall1、Plpp 2、Pld4、Fam102b、Cd33、Cdkn1a、Nkd1、Elovl7、Piga、Inf2、Blnk、Enpp4、Il18、Spata13、Mettl7a、Inpp5d、Myrf、Gjc3、T rim2、AABR07012039.1、AABR07033887.1、Enpp2、Pacs2、C1qa、Trem2、Slc45a3、Hhip、Anxa3、and AABR07022098.1。、

[0100] For example, the most specific genes (in order of decreasing specificity) in nuclei of OPCs from human samples were Fermt1, Lims2, C1ql1, Sstr1, Olig1, Bambi, Usp43, B3gnt7, Pdgfra, C1ql2, Olig2, Fmo3, Cspg5, Galr1, Gpc2, Fibin, Cspg4, Col11a1, Bche, Col9a1, Afap1l2, Cld n1, Plpp4, Sox13, Plat, Neu4, D630023F18Rik, Col20a1, Pxylp1, Asic4, Spsb4, Prelp, Gpr34, Best3, Stk 32b, Akr1c14, Susd5, Ophn1, Mmp16, Sulf2, Prkg2, Tns3, B3galt4, Fgfbp3, Nkain3, Ntn1, Ascl1, Sema5a, Hrasls, Gal3st4, Ebf4, Tmem132d, Nt5e, Jag1, Megf11, Ildr2, Plppr1, Dcc, Ntn4, Sox6, Cacng4, Pxdn, Sa pcd2, Tnk2, Slc43a3, Adamts6, Vsig8, Itm2a, Marcksl1, Alk, Smoc1, Cacna2d3, Tm4sf1, Bgn, Pdgfc, Atp2 c2, Aldh1a3, Prrx1, Meis3, Cxxc4, Cd27, Sema3e, Adamts17, Akr1c20, Arhgap10, Crispld2, Sipa1l2, Ppp1r36, Gng12, Syt17, Midn, Gadd45a, Khdrbs3, Thbs4, Vipr2, Ccnd1, Iffo1, Atoh8, Exosc4, and Pde7b.For example, the most specific genes (in order of decreasing specificity) in nuclei of OPCs from mouse samples were Neu4, C1ql1, Ppfibp1, 3110035E14Rik, Pdgfra, Rgcc, Susd5, Itga9, Tes, Rprm, Rep15, Cspg4, Matn4, Amz1, 1810041L15Rik, Pxdc1, Tox3, Sh3bp4, Elfn1, Gal nt3, Lmcd1, Serpine2, C1ql2, Neto1, Grm5, Mmp2, Col11a2, Sema3d, Lims2, Cav1, Sapcd2, Rcn1, Ptpro, C cnd1, Ptgfrn, Ncan, 5730559C18Rik, Lpcat2, Olig2, Fhl3, Tspan6, Rin2, Ascl1, Basp1, Pmepa1, Fzd9, F am114a1, Abhd2, Ctxn1, Cspg5, Cfap20, Vwc2, Cdo1, Cdh13, Marcks, Gsx1, Sertm1, Slc1a1, Matn1, Has2, Lama4, Mycl, Klhl5, Kcnh5, Kcnh8, Sema5a, Enpp6, Slc5a7, Cav2, Lbh, Adm, Adam12, S100a4, Mfsd2a, Igf bp3, Midn, Asap3, Prkg2, Rlbp1, Abcg2, Vstm2b, Dock6, Fbn1, Frrs1, Khdrbs3, Slitrk6, Chst11, Ptpn14, Dpysl3, Slc44a5, Eya1, Sox6, Phox2a, Col11a1, Swap70, Fam89a, Mrm2, Lypd1, Shisa7, and Pcdh17.For example, the most specific genes (in order of decreasing specificity) in nuclei of OPCs from rat samples were Fam89a, Pdgfra, RGD1561849, Rlbp1, Ccnd1, Tmem255b, C1ql1, Bgn, Itga9, Ppfibp1, Snx22, Col5a3, Pmel, Cav1, Serpine2, Cspg4, Lims2, AABR07052897.1, C1ql2, Cdk2, AABR07010022.1, Cacng4, and Matn4. , AABR07049948.1, Susd5, Prkg2, Pxdc1, RGD1566029, AABR07003304.1, Sapcd2, Cspg5, Mmp15, Rgcc, Slc22a6, Sema5b, Qprt, Pnlip, Slc6a12, RGD1311892, AABR07003306.1, AABR07044671.1, Rbpjl, Chst5, Traf4, Sox6, Cox6b2, Pcdh15, Cd101, Xylt1, AABR07003304.2, Calcrl, Col1a1, Lama4, Elfn1, Ampd3, Fam212b, Slc6a13, Sstr1, Sema3d, AC141997.1, Cpne7, Pt gfrn, Wbscr28, Ryr3, Gpsm2, Neto1, Chst11, Slc22a8, Pstpip2, AABR07044668.1, Bcas1, Igf2, Lbh, Olfm2, Ndnf, V stm2b, AABR07058656.1, Rep15, Vtn, Eln, Ramp1, Pmepa1, Gpnmb, Ptpro, Limd1, Cdh13, Nrxn2, Mgp, Myt1, Marcks, LOC100362216, AABR07043601.4, LOC102553018, Masp1, Sulf2, Tle6, Fam114a1, Tmem100, Ccnd2, and Rasgef1b.

[0101] In some embodiments, it is observed that the specificity rank is more conserved between mouse and rat than between mouse and human. In some embodiments, Geno Ontology (GO) analysis can be performed on the genes considered specific for each cell type or species to determine whether the genes function in the same pathway. In some embodiments, the change in expression of cell type or species specific markers in a given cell type is observed in different family members. For example, Pde1c can be used for cerebellar mouse cell types or their nuclei, and Pde1a expression can be used for cerebellar human cell types or their nuclei.

[0102] In some embodiments, differential expression levels of protein tyrosine kinase 2 beta (Ptk2b), solute carrier family 43 member 2 (Slc43a2), glutamate ionotropic receptor kainate subunit 3 (Grik3), Itpr1, protocadherin 9 (Pcdh9), and RaS guanyl releasing protein 1 (Rasgrp1) are used to distinguish between nuclei derived from granule, glia, and basket / astrocytes from humans and mice.

[0103] B. Analysis of cell type- and species-specific gene expression by cross-species analysis In some embodiments, the most changing genes are identified from the comparison of RNAseq results from selected nuclei from each cell type compared to normalized RNAseq results across the three species. In some embodiments, cell type-specific and species-specific genes can be identified by the genes with the largest changes in cell type and / or the largest changes in the species analyzed. In some embodiments, the genes observed to have the largest expression changes in species can be further differentiated by cell type. For example, nuclei can have greater differentiation in species than in cell types. In some embodiments, hierarchical clustering can be used to observe the degree of separation obtained by each variation component.

[0104] In some embodiments, Pcsk2, Pcsk6, and Pcsk8 are used to classify nuclei by both cell type and species, or Pcsk1 and Pcsk3 are used to classify nuclei by cell type.

[0105] In some embodiments, nuclear structure can be compared to differences in gene expression regulation or disease susceptibility across cell types.

[0106] C. Effect of Clinical Attributes of Human Samples In some embodiments, nuclear profile differences according to clinical attributes may be analyzed to identify markers of the attribute, for example, profile differences observed for different periods of autolysis time, gender, age, and stress.

[0107] Autolysis was observed to have only a minor effect on gene expression in previous studies. (Gupta et al., 2012, BMC Genomics 13, 26, which is incorporated herein by reference in its entirety). In some embodiments, expression of at least four genes in sorted nuclei may be used to identify granule cells from human samples at various stages of autolysis. In some embodiments, expression of at least one gene may be used to identify basket / astrocytic cells or glia from human samples at various stages of autolysis. In some embodiments, expression levels of kinesin family member 19 (Kif19) may be measured to determine autolysis time. For example, higher levels of Kif19 expression indicate longer autolysis times and lower levels of Kif19 expression indicate shorter autolysis times. Expression levels may be compared to normalized expression across samples. In some embodiments, higher expression levels of FosB proto-oncogene, AP-1 transcription factor subunit (FosB) indicate intermediate autolysis times of about 5 hours to about 24 hours.

[0108] We can also identify sex- and cell type-specific gene expression to determine the effect of sex on gene expression. Xist is a gene on the X chromosome that is involved in X chromosome inactivation and is expressed only in females. We then analyzed whether sex affects gene expression in specific cell types. We observed that Tsix, a gene that acts as an antisense to Xist, is expressed only in female glia. This is an example of sex- and cell type-specific gene expression.

[0109] In some embodiments, the profiles of nuclei isolated from male and female samples can be compared to determine sex-specific markers. For example, increased expression of KDM5D, RPS4Y1, ZFY, DDX3Y, TTTY15, TTTY14, USP9Y, UTY, GYG2P1, NLGN4Y, TXLNGY, LINC00278, PRKY, and / or TTTY10 can be used as markers for nuclei from male samples compared to normalized expression of whole unsorted nuclei. Furthermore, decreased expression of ZFX, PUDP, KDM5C, LOC389906, PIN4, MTRNR2L8, MTRNR2L6, COL4A1, DHFR, LOC101929541, TSIX, MIR6723, and / or XIST can be used as markers for nuclei from female samples. In some embodiments, the expression results are further analyzed to identify cell type and sex-specific markers. For example, increased expression of long intergenic non-protein-coding RNA 278 (LINC00278) may be used to distinguish glia from female samples. Additionally, increased expression of Family 8 member A4, pseudogene (Fam8a4p), Arylsulfatase D pseudogene 1 (Arsdp1), and Gyg2p1 with sequence similarity relative to normalized expression in total unsorted nuclei may be used as markers of nuclei from astrocytes derived from male samples.

[0110] To determine whether age affects cell types uniformly, genes that are differentially expressed in samples from humans of various ages can be identified. In some embodiments, the following genes in the nucleus: ABCC10, ACAP3, ACHE, ATF6B, ATP6V0B, CACNA2D3, CUL7, DNAJB5, EGR1, FZR1, GBAP1, GPR107, GRK5, KCTD21-AS1, KLHL22, LINC00499, LINGO2, LOC100132077, LOC100506990, LOC101927592, MI Differences in expression of R9-3HG, MPND, MTRF1, POLR2J3, PPP6R2, PRSS55, PTMS, RNF17, SNAI3-AS1, SPHK2, ST3GAL2, TAOK2, TRDN, WASH7P, ZFPM1, ZNF404, ZNF496, ZNF580, ZNF585A, ZNF607, ZNF84, and ZSCAN29 can be used to determine the age of the subject from which the sample was obtained.

[0111] In some embodiments, further analysis of the profile may provide age- and cell type-specific markers, such as the following genes, compared to normalized expression across samples of various ages: KCNQ3, SCHLAP1, UGGT2, PLXDC2, ITGB5, DCLK1, AQP7, PTPN3, FAT1, CMTM8, FARP1, PIP4K2A, NEDD4, CATSPERG, LOC101926, 941, WWC2, MPP6, SDK1, NAB2, NTNG2, LMNTD2, PLCH2, PTPRJ, MRAP2, PTMS, CDH23, SYT7, SPRED3, PLXNA1, GRK5, ZYX, FGF3, LINC01 544, PCYT1B, RHBG, HCN2, ACAP3, MGAT3, C1orf61, SURF2, AHRR, FAM131A, CHGB, PLXNA4, ACHE, DYRK1B, CDK5R2, GDPD5, ZNF703, FHDC1, LOC10192 7, 078, CREB3L2, AP1S3, DYSF, MAMSTR, KIAA0895L, TP73, PDE3A, LINC00963, ELAVL3, TTYH3, HPN, PTPRE, SLC17A7, GRIN2C, PXN, TMEM130, LRPAP 1, GRM4, STEAP3, CHRD, SBNO2, ZDHHC23, GACAT2, MAP6D1, FAM53A, KIAA1549L, SLC38A7, CDK16, CALY, LPCAT4, SLC25A6, SRCIN1, ELMO1, GRAMD4, RIMS3, CPLX2, PIP5K1C, SDC3, CBLN3, PDZD7, VWA5B2, ZNF219, FAM212B, RASIP1, GYS1, ATP6V0B, NKD1, ST3GAL2, PTK7, ACOT7, ATXN7L2, DAO, SYT 12, ADAM19, HIP1, GABBR1, SYT3, TP53INP2, KCNK3, TPM4, RNF152, KCNK10, EPB41L1, ABCC10, NAV2, ERGIC1, BAD, KIF3B, RAI2, XKR7, PTK2B, CHST 12, TRIM67, DHDDS, NUMBL, KCND3, LRRC1, RAB37, KCNIP2, RAB6B, PC, TCEAL2, MEG8, FMNL3, MAPK4, DGCR5, REEP1, GTPBP2, KIF26B, NRXN2, ZNRF1,C14orf132, FLNB, EFR3B, RBFOX3, PPP2R2C, KLHDC4, PILRA, CLSTN3, CABIN1, NHSL2, LYRM1, KCNA2, SLC38A10, DGKA, TMEM266, FAM135B, ATP2B2, TMEM16 3, UNC5B, TPM3, MICAL2, GNAO1, GABBR2, KCNK1, USP4, TEX2, PRKAG2, APBA2, SUSD6, HDHD2, NAA20, CRIPT, KNTC1, RPS6KA5, PHF10, PCGF5, DPY19L4, APH1 B, ZNF546, ABCA17P, CDH26, CBWD6, ZNF736, CFAP70, UBE4A, TTC32, CFAP53, PXDNL, CPVL, GLIPR1, COL25A1, MTHFD2L, LSM5, LOC100505, 715, NR3C2, TOM 1L1, ANKRD42, SLC8A1, -AS1, ATAD2, GAB1, CCDC146, ELMOD1, CFAP74, C1orf146, P4HA2, RANBP3L, OTUD6B, -AS1, EYS, ALS2CR11, LECT1, BHLHE40, -AS1, STAB2, PXYLP1, LOC101929, 710, LINC00499, RHOBTB3, CRYBG3, NPY6R, RGS20, C2orf73, YIPF7, BUB1B, CFAP43, LOC102724, 623, LRRC69, STK3, CLIC5, C ACNA2D3, C2orf27A, LOC105378, 385, BEST3, EPHA6, LOC100506, 393, MEF2C, GTSCR1, ELOVL2, SPATA9, PLSCR4, GCK, LINC01032, MEF2C-AS1, BMS1P21, C Differential expression of ASP12, CLMP, ERBIN, TMC3-AS1, CXXC4, DCBLD1, GSTO2, PLOD2, P4HA2-AS1, TTC23L, CPLX3, SOX13, PRDM1, RSPO2, TSG1, MIR31HG, CTB-12O2.1, PLCH1, JAML, SEMA3G, LOC101928, 203, GLRA1, SYT16, SYN3, ARHGAP32, MIR646HG, ALCAM, SCGN, LINC01515, and LOC101929415 may be used as markers to identify granule cells.

[0112] For example, compared to normalized expression across samples of various ages, the following genes: ACER3, ACSS3, ADAMTS2, ADAMTS9, ADAMTSL3, ADD2, ADGRL3, AKAP13, ALOX5, ALS2CR11, ARHGEF10L, ARHGEF6, ASIC1, ASIC2, ATP6V0B, BMP3, BTBD11, C11orf87, C14orf1, C4orf22, C9orf135, ​​CA7, CACNA1D, CACNA2D3, CACNB4, CADM3, CADPS, CALR, CDK15, CFAP46, CLSTN 3, CNTN3, CPNE8, CYYR1, DBH, DCC, DDO, DGKH, DISP3, DLG1, DLG4, DMGDH, DNAH6, DNER, E2F3, ECHDC2, ELOVL2, ENPP1, EPHB1, ERC2, FAM135B, FAM46C, FBX L21, FGF5, FHOD3, FKBP5, FRMPD4, GABBR2, GALNT18, GAP43, GBAP1, GDF11, GPR161, GPR63, GREB1, HCN4, HECTD2-AS1, IQSEC3, KCNA4, KIF26B, KRT222, LA CTB2-AS1, LHFPL3, LINC00457, LINC00499, LINC01158, LOC105377448, LOC285692, LOC653602, LPCAT4, MARC2, MARCH11, MASP1, MEGF10, MTHFSD, MYRI P, NHS, NOS1, NPL, NPY6R, NTNG1, NXNL2, PAQR6, PARM1, PCP4, PEBP4, PHACTR2-AS1, PHKA1, PLPPR4, PLXNA4, PRKCG, RGMB, RNF212, RNF215, RORA, RPH3A, R The differential expression of PS6KA5, SCN2B, SEPSECS, SEPSECS-AS1, SHANK1, SHANK2, SLC12A5, SLC17A4, SLC25A29, SLC26A8, SLC35F4, SLC8A3, SPARCL1, SPHK2, SPON2, ST6GAL2, ST6GALNAC5, STAT5A, STK33, SV2C, SYT17, TEX9, TFAP2B, THSD7B, TMEM178B, TRAF5, TRIM67, TTC23L, TUNAR, UBASH3B, UNC5D, ZNF124, and ZYG11A was analyzed in the 144 mice with 144 mice.It can be used as a marker to identify basket cells.

[0113] For example, differential expression of the following genes: ADAMTS7, ANKRD33B, AOX1, C1orf95, CACNG4, CACNG8, CAPN9, CRTAC1, DACH1, EFHD2, EGF, ETS1, FRAS1, IMPA2, LINC00499, LINC00844, LINC01208, LOC101927078, LOC102724360, PKDCC, PLXNA4, PRR5, RHCG, ROBO2, SEMA5B, SHROOM2, SLIT2, TRDN, TXNIP, VWA5A, and WNT7B, relative to normalized expression across samples of various ages, may be used as markers to identify glia.

[0114] In some embodiments, gene expression in the nucleus

[0115] D. Marker and Target Selection In some embodiments, analysis of altered expression associated with sorting nuclei between gene orthologs in different species is performed, for example, comparing expression of 1:1 orthologs to ensure that results are due to altered expression and not differences in gene annotations in species.

[0116] In some embodiments, further analysis may be performed to ensure that altered expression is not due to annotation differences in the species. For example, only genes with 1:1 orthologs may be used to determine expression levels. Analysis may be further limited to genes greater than 1 kb in length and / or with less than 2-fold difference in length in species such as mouse, rat, and human. In some embodiments, cell type-specific or species-specific genes include the following: 1810041L15Rik, 2810459M11Rik, 3110035E14Rik, A2m, Adamts15, Adra1a, AI593442, Aldh1a1, Anln, Apcdd1, Apod, Aqp4, Arhgef33, Ascl1, Asgr1, Aspa, Atp1a2, Atp2a3, B3gnt5, B3gnt7, Bcl11a, Bhlhe22, C 1ql1, C1ql2, Cacna1g, Cacng4, Calb1, Camk1g, Car8, Carns1, Casq2, Cbln2, Ccdc152, Ccdc180, Ccnd2, Cd82, Cd9, Cdh19, Cdh r1, Cldn11, Clec7a, Clmn, Cmtm5, Cntn3, Cntnap5a, Cobl, Col5a3, Col6a1, Cpm, Cryab, Cspg4, Csrp1, D7Ertd443e, Ddx3y, Doc k5, Dpy19l2, Ebf2, Echdc2, Efna5, Enpp2, Erbb3, Ermn, Etnppl, Fa2h, Fam46a, Fgfr2, Flrt2, Fmo3, Folh1, Foxh1, Fsip2, Fxy d1, Fxyd7, Gad2, Gal3st1, Galnt5, Galnt6, Galr1, Gdf10, Gfap, Gja1, Gjb1, Gjc2, Gldn, Gli1, Glp2r, Glul, Gm136, Gnb3, Gng1 1, Gpr37, Gpr37l1, Gpr63, Gpx2, Gramd3, Grb14, Gria3, Grik3, Grm1, Grm5, Gsn, Hapln2, Hepacam, Hhatl, Hspa1b, Htr1b, Id4, Igsf11, Il22, Il33, Insc, Itgb8, Itih3, Itpr1, Kcnc2, Kcnj10, Kif19a, Kit, Kitl, Klk6, Leng9, Lfng, Lgi4, Lims2, Mag, Mak,Mal, March11, Marcksl1, Mcam, Mff, Mfge8, Mlc1, Mog, mt-Co1, mt-Nd4, Mybpc1, Myot, Myrf, Ndrg1, Nell1, Neto1, Neu4, Ninj 2, Nkx2-2, Notch1, Nrk, Olig1, Olig2, Opalin, Pax3, Pcp2, Pdc, Pdgfra, Pdzd3, Penk, Pex5l, Pgghg, Phyhip, Pkp3, Pla2g16, P lch1, Plekhd1, Plekhg3, Pllp, Plp1, Plpp3, Plpp4, Plscr1, Plxnb3, Pmp2, Pmp22, Ppfibp1, Ppp1r14a, Ppp1r17, Prex1, Prex 2, Prima1, Prkcg, Prr5, Prrg3, Prrx1, Ptchd4, Ptprk, Ptprz1, Pttg1, Pvalb, Pxdc1, Qk, Rapgef3, Rassf2, Rbp2, Rbpjl, Rgcc, Rgma, Ryr1, S100b, S1pr1, Sall1, Sec14l5, Serpine2, Shisa6, Skor2, Slc1a3, Slc1a6, Slc26a3, Slc32a1, Slc4a4, Slc5a11, Slc9a3, Slitrk6, Socs2, Sorcs3, Sox10, Sox2, Sox6, Sox8, Stag3, Stk32a, Sulf2, Susd5, Tceal7, Tfap2b, Thbs2, Tmem125, Tm The 250 most changed genes in the entire mouse, rat, and human nucleus may be selected, including em132d, Tmem63a, Tmem98, Tnc, Tnfrsf13c, Tns3, Trabd2b, Tril, Trpc3, Trpc7, Tshb, Tspan11, Tspan2, Tuba8, Tubb2b, Txnip, Ube2b, Ugt8a, Upp1, Vstm2b, Wfdc1, Xrra1, Zcchc24, Zeb2, and Zfp36l1.

[0117] V. Modulation of Identified Targets In some embodiments, at least one gene selected from one of Tables 2-4 is associated with at least one disease, disorder, or condition. A gene associated with a disease, disorder, or condition or its expression is also referred to as a disease target. In some embodiments, a method of treating a disease, disorder, or condition includes the use of a therapy to modulate expression of a disease target. In some embodiments, the therapy targets at least one transcript selected from Tables 2-4. In some embodiments, the therapy targets at least one protein selected from Tables 2-4. In some embodiments, the therapy is a pharmaceutical composition, such as a small molecule drug or an RNA interference (RNAi) molecule (such as an siRNA), a gene therapy, or a cell therapy.

[0118] VI.Definition Below is a non-limiting list of term definitions.

[0119] As used herein, the term "about" as applied to one or more values ​​of interest refers to a value similar to the indicated reference value. In certain embodiments, the term "approximately" or "about" refers to a range of values ​​that falls within 25%, 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, 0.9%, 0.8%, 0.7%, 0.6%, 0.5%, 0.4%, 0.3%, 0.2%, 0.1%, or less in either direction (greater than or less than) the indicated reference value, unless otherwise specified or otherwise clear from the context (unless such number is deemed to exceed 100% of possible values).

[0120] As used herein, the term "antibody" refers to an immunoglobulin molecule, e.g., polynucleotide, polypeptide, protein, etc., capable of specifically binding to a target through at least one antigen recognition site located in the variable region of the immunoglobulin molecule. The term "antibody" encompasses intact (e.g., full-length) polyclonal or monoclonal antibodies, as well as antigen-binding fragments thereof (Fab, Fab', F(ab')2, Fv, etc.), single chain (scFv), variants thereof, fusion proteins containing antibody portions, humanized antibodies, chimeric antibodies, diabodies, linear antibodies, single chain antibodies, multispecific antibodies (e.g., bispecific antibodies), and any other modified configuration of immunoglobulin molecule that contains an antigen recognition site of the required specificity, including glycosylation variants of antibodies, amino acid sequence variants of antibodies, and covalently modified antibodies. Antibodies include any class of antibodies, e.g., IgD, IgE, IgG, IgA, or IgM (or subclasses thereof), and the antibody need not be of any particular class.

[0121] As used herein, the term "cell-specific profiling" refers to the use of the nuclear transcriptome to precisely define the identity of a cell.

[0122] As used herein, the term "chromosome-associated transcript (CAT)" refers to a transcript that is carried on the chromosome during synthesis.

[0123] As used herein, the term "disease tissue" or "disease sample" refers to a sample taken from a subject experiencing a disease state.

[0124] As used herein, the term "DNA probe" or "RNA probe" refers to a DNA or RNA fragment (usually 100-1000 bases long) that hybridizes to a nucleotide sequence complementary to the sequence of the probe. The probe may be radio-, fluorescently, or chemically (e.g., biotin) labeled for visualization.

[0125] As used herein, the term "drug target" or "target for therapy" refers to a nucleotide sequence that can be acted upon by a drug or therapy to regulate gene expression in a cell type.

[0126] As used herein, the term "factor" refers to a nucleotide sequence that can be hybridized to a DNA or RNA probe or specifically bound by an antibody to allow precise selection of nuclei of a particular cell type.

[0127] As used herein, the term "gene expression" or "expression" refers to the level of one or more of the following events: (1) production of an RNA template from a DNA sequence (e.g., by transcription); (2) processing of the RNA transcript (e.g., splicing, editing, 5' capping, and / or 3' end processing); (3) translation of the RNA into a polypeptide or protein; (4) folding of the polypeptide or protein; and (5) post-translational modification of the polypeptide or protein.

[0128] As used herein, the term "cell profiling techniques" refers to non-genetic methods for the molecular profiling of populations of cell types in post-mortem tissue.

[0129] As used herein, the term "polyA transcript" refers to a stretch of adenine nucleotides that is added to a messenger RNA (mRNA) molecule during RNA processing to increase the stability of the molecule.

[0130] As used herein, the term "label" refers to one or more markers, signals, or moieties that are attached, incorporated, or bound to another entity, and that are easily detected by methods known in the art, including X-ray, fluorescence, chemiluminescence, enzyme activity, absorbance, and the like. Detectable affinity labels include radioisotopes, fluorophores, chromophores, enzymes, dyes, metal ions, ligands such as biotin, avidin, streptavidin, and haptens, quantum dots, and the like. Detectable affinity labels may be located at any position in the entity to which they are attached, incorporated, or bound. For example, when attached, incorporated, or bound to a peptide or protein, they may be within an amino acid, peptide, or protein, or may be located at the N- or C-terminus.

[0131] As used herein, the term "most specific gene" refers to a gene that has a lower specificity index (SI) value compared to the SI values ​​of the other analyzed genes, the SI value being calculated using the algorithm described in Example 1.

[0132] As used herein, the term "tissue sample" refers to an aliquot or portion taken from a source and provided for analysis or processing. In some embodiments, the sample is from a biological source such as tissue, cells, or components (e.g., bodily fluids, including but not limited to blood, mucus, lymph, synovial fluid, cerebrospinal fluid, saliva, amniotic fluid, amniotic cord blood, urine, vaginal fluid, and semen). In some embodiments, the sample may be or include a homogenate, lysate, or extract prepared from a whole organism or a subset of its tissues, cells, or components, or a fraction or portion thereof, including but not limited to, for example, plasma, serum, cerebrospinal fluid, lymph, skin, external parts of the respiratory, intestinal, and genitourinary tracts, tears, saliva, milk, blood cells, tumors, organs. In some embodiments, the sample is or includes a medium, such as a nutrient broth or gel, which may contain cellular components such as proteins or nucleic acid molecules. In some embodiments, the sample is subjected to one or more processing (eg, separation, purification, etc.) steps to prepare the sample for analysis or other uses.

[0133] As used herein, the term "transgenic animal" means an animal that has a foreign gene inserted into its genome.

[0134] As used herein, the term "Translating Ribosome Affinity Purification (TRAP)" refers to a method to identify the molecular signatures of diverse cell types in a single tissue from a transgenic animal.

[0135] Described herein is a method for molecular profiling of a population of cell types in postmortem tissue. Details of one or more embodiments of the present invention are described in the following description. Although any materials and methods similar or equivalent to those described herein can be used to carry out or test the present invention, preferred materials and methods are described below. Other features, objects, and advantages of the present invention will be apparent from the description. In the description, singular forms include plural forms unless otherwise clearly indicated by the context. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs. In the event of any discrepancy, the description in the present specification shall prevail.

[0136] The invention is further illustrated by the following non-limiting examples.

[0137] Working Example Example 1. Experimental Procedure The following experimental procedures were used to carry out the experiments described in Examples 2-23 herein. A. Buffer Preparation Buffers were prepared as described in Kriaucionis et al., Science, 324(5929):929-930 (2009). RNase-free purified water was used when included in the buffers involved in the protocol for isolating RNA, otherwise Milli-Q water was used.

[0138] Buffer A used in the examples herein is 60% iodixanol (Optiprep, Sigma-Aldrich, St. Louis, MO; Catalog No. D1556-250ML). Buffer A was stored covered in foil until use. Buffer B used in the examples herein contains 900 mM KCl, 30 mM MgCl2, 120 mM Tricine-KOH (pH 7.8). Buffers A and B were stored at room temperature (RT).

[0139] Buffer C used in the examples herein contained 50% iodixanol (5 volumes of Buffer A and 1 volume of Buffer B), 0.5 mM spermidine, 0.15 mM spermine, EDTA-free protease inhibitor cocktail (Roche Holding AG, Basel, Switzerland; distributed by Sigma-Aldrich, St. Louis, Mo.; catalog number 11836170001), 1 mM DTT, and 200 units / mL Superasin (Thermo Fisher Scientific, Waltham, Mass.; catalog number AM2696). Buffer D used in the examples herein contained 0.25 M sucrose, 150 mM KCl, 5 mM MgCl2, 20 mM Tricine-KOH (pH 7.8), 0.5 mM spermidine, 0.15 mM spermine, EDTA-free protease inhibitor cocktail, 50 μL of 1 mM DTT, 200 units / mL Superasin, and 400 units / mL RNasin (Promega, Madison, WI; catalog number N2515).

[0140] Buffer E used in the examples herein contained 27% iodixanol (1.174 volumes of buffer C to 1 volume of buffer D). Wash / blocking buffers used for analysis in the examples herein contained 3% BSA in PBS with 0.05% Triton X-100; BSA was obtained from Jackson ImmunoResearch Laboratories, Inc., code number 001-000-162. Wash buffers for RNA used in the examples herein contained 50 ng / mL RNase-free BSA, 1 mM DTT, and 100 units / mL Superasin in PBS with 0.05% Triton X-100. Blocking buffers for RNA used in the examples herein included RNA wash buffer supplemented with an additional 50 ng / mL BSA. The cushion buffer used in the examples herein contained 500 μL of Buffer D, 250 μL of 50% glycerol, and 5 μL of 10% NP-40.

[0141] B. Nuclear isolation Nuclear isolation was adapted from protocols described in previous publications (Kriaucionis, et al. (2009) Science 324, 929-930 and Mellen, et al. (2012) Cell 151, 1417-1430, each of which is incorporated herein by reference in its entirety). For mouse experiments, cortices and cerebellum were dissected as previously described. All tissues from C57BL / 6J animals were flash frozen using liquid nitrogen. Tissues from mice developed for use in the bacTRAP method were freshly dissected and used directly for homogenization. Rat and human tissues were obtained frozen. All frozen tissues were thawed on ice for a minimum of 30 minutes before use. To isolate nuclei, fresh or thawed tissue was transferred to 5 mL of homogenization medium (0.25 M sucrose, 150 mM KCl, 5 mM MgCl2, 20 mM Tricine pH 7.8, 0.15 mM spermine, 0.5 mM spermidine, EDTA-free protease inhibitor cocktail, 1 mM DTT, 20 U / mL Superase-In RNase inhibitor, 40 U / mL RNasin ribonuclease inhibitor). Tissue was homogenized with 30 strokes of a loose (A) glass Dance homogenizer followed by 30 strokes of a tight (B) glass Dance homogenizer. The homogenate was supplemented with 4.6 mL of 50% iodixanol solution (50% iodixanol / Optiprep, 150 mM KCl, 5 mM MgCl2, 20 mM Tricine pH 7.8, 0.15 mM spermine, 0.5 mM spermidine, EDTA-free protease inhibitor cocktail, 1 mM DTT, 20 U / mL Superase-In RNase inhibitor, 40 U / mL RNasin ribonuclease inhibitor) and placed on a 27% iodixanol cushion. Nuclei were pelleted by centrifugation at 10,000 rpm for 30 min at 4°C in a swinging bucket rotor (SW41) of a Beckmann Coulter XL-70 ultracentrifuge. For mouse, rat, and human cytoplasmic fractions, 175 μL of the top layer was added to 175 μL of QIAGEN® Buffer RLT and stored at −80° C. The nuclear pellet was resuspended in homogenization buffer.

[0142] C.Antibodies Primary and secondary antibodies were diluted in blocking buffer. Secondary antibodies were purchased from Life Technologies or Jackson Immunoresearch and used at a dilution of 1:500. Primary antibodies and dilution factors are shown in Table 5. In this table, NeuN means neuronal nuclear antigen, Itpr means inositol 1,4,5-trisphosphate receptor type 1, Sorcs3 means sortilin-related VPS10 domain-containing receptor 3, EAT1 means excitatory amino acid transporter 1, and Olig2 means oligodendrocyte transcription factor 2. [Table 6] TIFF2025032204000392.tif40165

[0143] The cell type specificity of the antibodies used to isolate all cell types is shown in Table 6. Unless otherwise indicated, NeuN refers to rabbit antibody. [Table 7]

[0144] D. Nuclear fixation and staining After isolation, nuclei were fixed by adding formaldehyde (methanol-free formaldehyde; Electron Microscopy Sciences, Hatfield, PA; catalog number 15710) to a final concentration of 1% and incubated at room temperature for 8 minutes with gentle rotation. Formaldehyde was quenched by adding glycine to a final concentration of 125 mM. Samples were incubated at room temperature for 5 minutes with gentle rotation. Samples were centrifuged at 1000 g for 4 minutes at 4° C. to pellet the nuclei. The supernatant was poured off and the pellet resuspended in Buffer D. Samples were centrifuged again at 1000 g for 4 minutes at 4° C. The supernatant was poured off and the pellet resuspended in wash buffer (PBS, 0.05% Triton X-100, 50 ng / mL BSA, 1 mM DTT, 10 U / μL Superase-In RNase inhibitor). Nuclei were then transferred to a new tube and centrifuged at 1000g for 4 minutes at 4°C. The supernatant was poured off and nuclei were resuspended in blocking buffer and blocked for 30 minutes at room temperature (RT) with gentle rotation. Antibodies were added to the nuclei in blocking buffer and incubated for 1 hour at RT with gentle rotation. Alternatively, this incubation step was performed overnight (O / N) at 4°C.

[0145] A small aliquot of nuclei was taken at this stage to serve as an unstained control for flow cytometry. Nuclei were pelleted at 1000g for 3 min at 4°C and washed with wash buffer (PBS, 0.05% Triton X-100, 50ng / mL BSA, 1mM DTT, 10U / μL Superase-In RNase inhibitor). This was repeated two more times. After the final wash, the appropriate secondary antibody was added at a 1:500 dilution in blocking buffer. Nuclei were covered with foil and incubated at RT for 30 min with gentle rotation. After incubation, nuclei were pelleted at 1000g for 3 min at 4°C and washed with wash buffer. This was repeated two more times. The pellet was then resuspended in the appropriate volume of blocking buffer for analysis or sorting. If no immediate analysis was planned, nuclei were stored on ice.

[0146] E. Flow Cytometry Analysis / Sorting DyeCycle™ Ruby (dcRuby) (Thermo Fisher Scientific, Waltham, MA; catalog number V10273) was added at 1-2 μL per mouse cerebellum sample and 5-7 μL per 200 mg human tissue sample. This amount was based on the number of nuclei in the sample, the volume of buffer, and whether the nuclei were of mouse or human origin. Varying concentrations of dcRuby were allowed depending on the voltage of the flow cytometer.

[0147] Prior to flow cytometry, nuclei were co-stained with DyeCycle™ Ruby to a final concentration of 20 μM. Nuclei were analyzed using a BD LSRII (BD Biosciences, San Jose, CA, USA) flow cytometer with 488 nm, 561 nm, and 640 nm lasers. Nuclei were sorted using a BD FACS Aria cell sorter with 488 nm, 561 nm, and 635 / 640 nm lasers. All samples were first gated to identify singlets using DyeCycle™ Ruby. Analysis was performed using FACS Diva (BD) or FlowJo software. QIAGEN® Buffer RLT (unfixed samples) or QIAGEN® Buffer PKD (fixed samples) was added to sorted nuclei and samples were stored at -80°C.

[0148] After sorting, the nuclei were stored for RNA isolation. For storage, Proteinase K digestion buffer (QIAGEN® AllPrep® FFPE RNA Kit, QIAGEN®, Cat. No. 80234) was added to the sorted nuclei and the resulting mixture was stored at -80°C.

[0149] Forward scatter (FSC) measurements were plotted against side scatter (SSC) measurements and gating was performed on total scatter. dcRuby intensity was plotted linearly and voltage adjusted for singlets. Singlets were gated. Samples were plotted by color and populations were gated based on predicted staining patterns. Percentages of gated populations were compared to predicted percentages of cell types of interest.

[0150] F. RNA and DNA isolation and purification Both DNA and RNA were isolated from the same samples using the QIAGEN® AllPrep® FFPE Kit. Isolation was performed according to the kit protocol with the following modification: RNA was incubated at 65° C. for 30 minutes prior to the 80° C. heating step, followed by an additional 30 minutes at 70° C.

[0151] RNA from cytoplasmic fractions or fresh nuclei was purified using the QIAGEN® RNeasy® Micro kit with on-column DNase digestion. RNA from fixed nuclei was purified using the protocol QIAGEN® RNeasy® FFPE kit with the following modification: after proteinase K digestion, the RNA was centrifuged at maximum speed for 15 min. The supernatant was removed and incubated at 65°C for 30 min, 70°C for 30 min, and 80°C for 15 min. On-column DNase digestion was performed instead of in-solution DNAse digestion. RNA quantity was measured using the QUBIT® RNA HS Assay Kit and RNA quality was measured using an Agilent 2100 Bioanalyzer with an RNA6000 pico chip.

[0152] Purified DNA and RNA were quantified using QUBIT® (Thermo Fisher Scientific, Waltham, MA; Catalog No. Q32854 and Q32855). Purified RNA samples were also run on an AGILENT® Bioanalyzer chip (Agilent Technologies, Santa Clara, CA; Catalog No. 5067-1513) for quantification and size determination. DNA was stored at -20°C and used for sequencing or epigenetic analysis. RNA was stored at -80°C and used for cDNA synthesis.

[0153] G. cDNA Creation and Library Generation using Nugen Ovation® RNAseq V2 Kit cDNA was prepared from purified nuclear RNA using the Nugen Ovation® RNAseq V2 kit according to the kit protocol. The final cDNA was purified using the QIAGEN® MinElute® Reaction Cleanup Kit (catalog number 28206) according to the Nugen Ovation® RNAseq V2 kit protocol. cDNA quality was assessed using an AGILENT® TapeStation and D1000 ScreenTape (Agilent Technologies; Santa Clara, CA; catalog number 5067-5582, ScreenTape catalog number 5067-5583). A 1:40 dilution of cDNA was prepared for qPCR to check for enrichment or depletion of predictive markers. cDNA was quantified using NanoDrop™.

[0154] Libraries were prepared using the Illumina TruSeq DNA LT Library Preparation Kit or the NEBNext Ultra DNA Library Prep Kit for Illumina with NEBNext Multiplex Oligos for Illumina. Library quality was assessed using an Agilent 2200 TapeStation system and a D1000 High Sensitivity ScreenTape. Libraries were sequenced at the Rockefeller University Genomics Resource Center using an Illumina HiSeq2500 instrument to obtain 50 bp single-end reads or an Illumina NextSeq500 to obtain 75 bp paired-end reads. All datasets have been deposited with GEO (accession number pending).

[0155] H. Oxidative Bisulfite (oxBS) Conversion Reaction Genomic DNA (gDNA) is isolated from sorted nuclei as described in the user guide for the TrueMethyl® Whole Genome Kit (Cambridge Epigenetix Limited).

[0156] I. Immunofluorescence Mice were deeply anesthetized and perfused transcardially with phosphate-buffered saline (PBS) followed by 4% formaldehyde (w / v) in PBS. Brains were dissected and post-fixed overnight at 4°C, cryoprotected in 30% sucrose in PBS, embedded in OCT TissueTeck, and cut into 20 μm sections on a Leica CM3050 S cryostat, which were directly mounted on slides. Slides were stored at -20°C. Antigen retrieval was performed by immersing slides in sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20, pH 6.0) at 95–100°C and boiling in a microwave for 10 min. Slides were cooled to room temperature and then washed with PBS. Immunofluorescence was performed by blocking in IF blocking buffer (3% BSA in PBS with 0.1% Triton X-100) for 30 minutes, incubating with primary antibody overnight, washing 3 times with PBS, incubating with secondary antibody for 1 hour, washing with PBS, staining with DAPI solution (1ug / mL in PBS) for 15 minutes, washing with PBS for 2 minutes, and coverslipping with Prolong Diamond mounting media. For some samples, tyramide signal amplification was performed as follows: slides were incubated with secondary antibody conjugated to horseradish peroxidase, washed 3 times with PBS, incubated with Cy3 in amplification buffer from the TSA Cyanine 3 detection kit for 10 minutes, washed with PBS, and stained with DAPI as described above. All steps were performed at room temperature. Slides were imaged on a Zeiss LSM700 confocal microscope using the same image acquisition settings for mouse and human slides. Brightness and contrast adjustments were performed in ImageJ with the same adjustments applied for mouse and human images after acquisition. In addition to the above antibodies for nuclear staining, the following primary antibodies shown in Table 7 were used for immunostaining.

[0157] In Table 7, GFAP means glial fibrillary acidic protein, Mog means myelin oligodendrocyte glycoprotein, Mag means myelin associated glycoprotein, Pde1a means phosphodiesterase 1A, and Pde1c means phosphodiesterase 1C. [Table 8]

[0158] The antibodies described herein may be generated, tested, and / or characterized using a variety of methods. Such methods may be used to determine a variety of properties, which may include, but are not limited to, antibody affinity; specificity; and activity (e.g., activation or inhibition of intracellular signaling pathways or other cellular or biological activity).

[0159] Once an antibody molecule of the disclosure has been produced by an animal, by a cell line, by chemical synthesis, or by recombinant expression, the antibody molecule can be purified (i.e., isolated) by any method known in the art for the purification of immunoglobulin or polypeptide molecules, for example, by chromatography (e.g., ion exchange chromatography, affinity chromatography, particularly for the specific antigen, Protein A, and sizing column chromatography), centrifugation, differential solubility, or any other standard technique for the purification of proteins. Additionally, antibodies of the disclosure or fragments thereof can be fused to heterologous polypeptide sequences described herein, or otherwise known in the art, to facilitate purification.

[0160] The affinity between an antibody and a target or ligand (e.g., the antigen used to generate a given antibody) can be measured in terms of KD using one or more binding assays known in the art. Depending on the target of interest for a given antibody, variable KD values ​​may be desirable. High affinity antibodies typically have an affinity of about 10 -5 M or less, e.g., about 10 -6 M or less, about 10 -7 M or less, about 10 -8 M or less, about 10 -9 M or less, about 10 -10 M or less, about 10 -11 M or less, or about 10 -12 Ligand binding occurs with a KD of less than or equal to M.

[0161] J. DNA and RNA Probes Alternatively, or in combination with antibodies, RNA and DNA probes can be designed and made using the polymerase chain reaction (PCR) to generate antisense fragments of regions of interest in the genome according to David et al. Biotechniques 30, 769-772, 2001.

[0162] K. Real-time PCR and primer list Separate RNA samples were purified from cytoplasmic fractions from mouse and human cerebellum, and purified RNA was converted to cDNA and amplified using Nugen Ovation® RNAseq System V2 as described above. Quantitative real-time PCR was performed on a LIGHTCYCLER® 480II system (Roche) using TAQMAN® probes and the LIGHTCYCLER® 480 probe master mix. Two biological replicates were used for both mouse and human, and three technical replicates were performed for each biological replicate. Cq values ​​were calculated using the second derivative maximum method in the LIGHTCYCLER® 480 software. Normalization was performed with the delta Cq method using the geometric mean of three housekeeping genes: actin beta (Actb), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and RNA polymerase II subunit A (Polr2a). Technical replicates were averaged (arithmetic mean) before performing statistical analysis using a two-tailed t-test assuming equal variances. Primers are listed in Table 8. In Table 8 and as used herein, the genes referred to are Actb, GAPDH, Polr2a, von Willebrand factor C domain containing 2 (Vwc2), coiled-coil domain containing 175 (Ccdc175), Clavesin 2 (Clvs2), Pde1a, Pde1c, endothelin-converting enzyme 1 (Ece1), CNKSR family member 3 (Cnksr3). Also given in this table are mouse and human IDs from Integrated DNA Technologies (IDT) and Thermo Fisher (Thermo). [Table 9]

[0163] L. Software used for analysis Data processing steps utilized Linux tools in addition to custom Perl scripts. In addition to standard Linux commands (e.g., awk, cut, sort, uniq), the following packages were used: Trim Galore! (v0.4.1) for adapter trimming of ATACseq samples, STAR (v2.4.2a) and Bowtie2 (v2.1.0) for read alignment, SAMtools (v0.1.19-44428cd) for indexing and removal of duplicates, igvtools (v2.3.32) for generating files for browser visualization, and featureCounts (v1.5.2) for read summarization. Data analysis was performed using R Studio (v1.0.143, R:v3.4.0). In addition to base R (for data wrangling, Pearson correlation, hierarchical clustering, etc.) and custum R functions, the following packages were used extensively: tidyverse (v1.1.1, which includes ggplot2, reshape2, etc.) for data wrangling and visualization, DESeq2 (v1.16.1) for raw count normalization, differential expression analysis, and principal component analysis, gplot (v3.0.1) and pheatmap (v1.0.8) for heatmaps, RColorBrewer (v1.1-2) for color palettes, and clusterProfiler (v3.4.3) for GO analysis.

[0164] M.ATACseq library construction and sequencing Nuclei from mouse, rat and human cerebellum were isolated as described above and 50,000 were washed in ice-cold PBS. ATACseq libraries were prepared as previously described by Buenrostro, et al. (2013). Nat Methods 10, 1213-1218 and Buenrostro, et al. (2015). Curr Protoc Mol Biol 109, 21.29.1-.29.9 (both of which are incorporated herein by reference in their entireties) and according to Chen, et al. (2016). Nature Methods. 13(12):1013-1024 (which are incorporated herein by reference in their entireties) with the following modifications: nuclei were resuspended in ice-cold lysis buffer, the incubation step on ice was omitted, and centrifugation was performed. Libraries were generated using the Nextera DNA Library Preparation Kit. Transposition reactions were performed at 37°C for 30 min using Tn5 transposase from Nextera. 200 μL of reverse crosslinking solution (50 mM Tris-Cl, 1 mM EDTA, 1% SDS, 0.2 M NaCl, 5 ng / ml proteinase K) was added and the mixture was incubated overnight at 65 °C in a heat block with 1000 rpm shaking, followed by purification with a QIAGEN® MINELUTE® kit. DNA fragments were purified and size-selected using AMPure XP beads to obtain 100-700 bp fragments. ATACseq libraries from mouse and human cerebellum were sequenced on an Illumina HiSeq2500 to obtain 50 bp paired-end reads, and libraries generated from rat cerebellum were sequenced on an Illumina NextSeq500 to obtain 75 bp paired-end reads. All libraries were sequenced, obtaining at least 55M reads per sample.

[0165] N.ATACseq read mapping and visualization Adapter sequences were trimmed and reads were filtered according to quality using Trim Galore! (trim_galore --stringency 3 --fastqc --paired *_R1_2.fastq.gz *_R2_2.fastq.gz). Filtered reads were aligned to hg38, mm10 or custom rat genomes using Bowtie2 (Langmead, et al. (2012). Nat Methods 9, 357-359, which is incorporated by reference in its entirety). Duplicate reads were removed using SAMtools (Li, et al. (2009) Bioinformatics 25, 2078-2079, which is incorporated by reference in its entirety). For data visualization, tdf files were generated using igvtools.

[0166] O. RNAseq read mapping, visualization, and read summarization RNAseq reads were aligned using STAR (Dobin, et al. (2013). Bioinformatics 29, 15-21, which is incorporated herein by reference in its entirety) and genome assemblies from Ensemble. In addition to the default STAR parameters, the following parameters were used: (--outFilterMismatchNmax 999 --outFilterScoreMinOverLread 0 -- outFilterMatchNminOverLread 0 --outFilterMatchNmin 35 --outFilterMismatchNoverLmax 0.05) for single-end and (--outFilterMismatchNmax 999 --alignMatesGapMax 1000000 -- outFilterScoreMinOverLread 0 --outFilterMatchNminOverLread 0 --outFilterMatchNmin 60 -- outFilterMismatchNoverLmax 0.05) for paired-end data. Aligned bam files were converted to tdf format for visualization using igvtools. Raw counts were generated using featureCounts (Liao, et al. (2014) Bioinformatics 30, 923-930, which is incorporated herein by reference in its entirety). Refseq or ensemble gene model annotations for all genes were downloaded using the UCSC Table Browser tool. Refseq (Zhao, et al. (2015). BMC Genomics 16, 97, which is incorporated herein by reference in its entirety) annotations were used. Ensemble was selected for annotation for intra-species comparison. For comparative analysis across species, ensemble annotations were used for all species and matched with the list of orthologous transcripts obtained from the ensemble.In addition to the default parameters for featureCounts, the following parameters were used: for paired-end data, count fragments instead of reads (-p), and do not count chimeric fragments (-C). For comparative analyses across species, allow reads to be assigned to more than one matching metafeature (-O) to avoid problems when genes overlap in one species but not in another. For interspecies comparisons, reads that map to more than one matching metafeature are not counted (default). Annotations are shown in Table 9. [Table 10]

[0167] P. Generation of orthologous guannotations The method for generating annotations for mouse, rat, and human orthologs, and mouse-human high-confidence orthologs can be summarized as follows: 1) species-wide ortholog annotations were downloaded from Ensemble BioMart (release 88), 2) filtered to include only: high-confidence pairs (ensemble ortholog confidence = 0), 1:1 orthologs (transcripts annotated as orthologous to two or more transcripts in another species), genes over 1 kb in length (including untranslated regions), and genes that change in length less than 2-fold across species. For each gene, the longest orthologous transcript was used for annotation.

[0168] Due to the incomplete annotation of the rat genome, nearly 3,000 more genes were identified as orthologous between mouse and human than between mouse, human, and rat.

[0169] Q. Gene count normalization, differential expression analysis, and principal component analysis Raw gene count tables were loaded into R and differential expression analysis was performed using DESeq2 (Love, et al. (2014). Genome Biology 15, 550, which is incorporated herein by reference in its entirety). Numerical covariates were used in modeling for age and autolysis time. Statistically significant genes were defined as those with adjusted p-values ​​less than 0.01. No fold change cutoff was used unless otherwise noted. Normalized numbers were generated for downstream analysis (e.g., heatmaps, clustering) using the rlog function. Principal component analysis was performed using the plotPCA function. No method was used to remove batch effects. Predictive biology-based clustering was observed throughout but not by batch. For example, clustering by cell type was observed between EGfP+ and sorted nuclei using the cell profiling techniques described herein, despite numerous differences: preparation date, tissue condition (fresh vs. flash frozen + thawed), nuclei condition (unfixed vs. fixed), nuclei treatment (none vs. antibody stained), library prep kit (Illumina vs. NEB), sequencer (HiSeq2500 vs. NextSeq500), and read type (50bp single-end vs. 75bp paired-end).

[0170] R. Hierarchical Clustering Hierarchical clustering was performed using normalized gene counts generated by the rlog function from DESeq2 (Love, et al. (2014) Genome Biology 15,550, which is incorporated by reference in its entirety). Unless otherwise specified, the top 250 most variable genes across samples, calculated by the rowVar function, were used for clustering. Clusters were generated using the dist and hclust functions with default parameters (Euclidean distance and complete linkage).

[0171] S. Specificity Index Algorithm The specificity index (SI) algorithm was previously developed to find genes that are specific in a dataset of interest compared to other datasets (Dougherty, et al. (2010) Nucleic Acids Res 38, 4218-4230, which is incorporated herein by reference in its entirety). Since Dougherty's algorithm was designed for microarray data, it was modified to accept RNAseq data as input. For the calculation, we added the ability to incorporate information from biological replicates. Fpkm was calculated from raw counts to avoid bias towards long genes, and genes less than 1 kb were excluded to avoid bias towards extremely short genes. The fpkm values ​​were then recorded and the value of log10(fpkm) was minimized to -2 to avoid bias towards long non-expressed genes. The mean and standard deviation were calculated for all genes within a cell type, taking into account the number of replicates of a sample.

[0172] SI values ​​were calculated as described above, but instead of using the average gene expression value as input, samples were randomly selected from a normal distribution with the mean and standard deviation of gene expression. This process was performed 1000 times, and the final rank was averaged across the iterations. To analyze the distribution of highly specific genes in the mouse cerebellum, images of each gene were examined from the Allen Mouse Brain Atlas (Lein, et al. (2007) Nature 445, 168-176, which is incorporated herein by reference in its entirety). Genes that were not in the Atlas at the time of analysis or did not show staining in any region of the brain were excluded from the analysis. The percentage of genes showing the appropriate distribution of cell types was calculated and presented herein.

[0173] T.GO Analysis Significantly differentially expressed genes were divided into up- and down-expressed (or mouse-enriched and human-enriched) genes, and gene ontology enrichment analysis was performed for granule cells, basket cells, and glia using clusterProfiler (Yu, et al. (2012) Journal of Integrative Biology 16, 284-287, which is incorporated herein by reference in its entirety). Significant ontologies were defined as those with a q-value less than 0.01 (default). GO analysis results containing significant ontologies were collected for three domains: biological process (BP), molecular function (MF), and cellular component. For simplicity of visualization, only biological process ontologies were used. Furthermore, when many ontologies were found, the simplify function was used to remove duplicate items for visualization. Network-based enrichment maps were generated using the enrichMap function. To simplify the visualization of the enrichment map, related ontologies were combined into one label.

[0174] Example 2. Defining cell identity using the nuclear transcriptome Nuclei can be sorted using antibodies specific for constitutive endoplasmic reticulum (ER) proteins, membrane proteins synthesized in the ER. First, nuclear RNA was isolated from GFP+ nuclei as described in Example 1. To enrich for ER transcripts translated near the nuclear membrane of wild type mice, polyA pulldown of nuclear transcripts was performed. The majority of these transcripts were not observed to be nuclear, but were fully mature nucleic acid transcripts that were translated in the nucleus around the ER. Although not all transcripts of membrane proteins are found in this nuclear polyA+ fraction, knowledge of what is synthesized near the nucleus has dramatically increased the number of factors that can be targeted for nuclear labeling. After performing immunofluorescence, staining was observed in the ER around the nucleus, confirming that the nucleus can be labeled using antibodies against membrane proteins.

[0175] An initial strategy to isolate the three cell types from the cerebellum included antibody binding using Itpr1 (inositol 1,4,5-triphosphate receptor type 1; an intracellular receptor in the endoplasmic reticulum (ER)) as a target to label Purkinje cells, NeuN (a gene encoded by a member of the RNA-binding FOX protein family; NeuN is a neural antigen often used as a biomarker for neurons) as a target to label granule cells, and GFAP (glial fibrillary acidic protein; one of the major intermediate filament proteins of mature astrocytes used to distinguish astrocytes from glial cells) as a target to label glia to confirm the presence of these cell types in the sample. A combination of Itpr1 and NeuN results was then used to isolate all three populations.

[0176] Nuclei from cell types of interest were isolated from tissues of transgenic mice previously profiled using the TRAP method to determine whether the nuclear transcriptome could be used to precisely define cell identity. Profiling results from transgenic animals developed for use with the TRAP method were based on the fact that GFP-tagged ribosomal subunits were incorporated into nucleoli. Using immunofluorescence, GFP was observed to be distributed primarily in the cytoplasm of tissue sections of cerebellar tissue from transgenic animals with Purkinje cell protein 2 (Pcp2)-GFP expressed only in Purkinje cells, but also in one spot in the nucleus (nucleolus). Furthermore, staining results showed that isolated nuclei from the cerebellum of Pcp2-GFP animals could be labeled with DAPI, and one of seven nuclei was observed to be a Purkinje nucleus (GFP+).

[0177] Example 3. Use of flow cytometry to profile cells derived from transgenic animals developed for use in the TRAP method Table 10 shows the flow cytometry results of EGFP+ nuclei from each of the five cell lines derived from transgenic animals developed for use in the TRAP method. [Table 11]

[0178] Additional flow cytometry was performed on nuclei isolated from five different transgenic mouse strains previously profiled using the TRAP method. From the previous profiles purified by the TRAP method, we used Purkinje cell protein 2 (Pcp2) as a target to identify nuclei derived from Purkinje cells and observed that approximately 0.4% of nuclei were positive for Pcp2. We used Neurod1 (neuronal differentiation 1) as a target to identify nuclei derived from granule cells and observed that approximately 91% of nuclei were positive for Neurod1. We used Sept4 (septin 4) as a target to identify nuclei derived from Bergmann glia and observed that approximately 2% of nuclei were positive for Sept4. We used Glt25d2 or Colgalt2 (collagen beta(1-O) galactosyltransferase 2) as targets to identify nuclei derived from layer 5 cortical-pontine pyramidal neurons and observed that approximately 0.5% of nuclei were positive for Glt25d2. Using neurotensin receptor 1 (Ntsr1) as a target to identify nuclei derived from layer 6 corticothalamic projection pyramidal neurons, we observed that approximately 1.8% of nuclei were Ntsr1 positive.

[0179] The values ​​in Table 10 are based on conservative gating of the positive population for sorting during flow cytometry, while the percentages in the previous paragraph are based on a less conservative analysis of the flow data.

[0180] Example 4. Expression profiles of cell lines derived from transgenic animals developed for use in the TRAP method To determine whether nuclear RNA profiles are sensitive enough to distinguish between different subtypes of neurons, we used RNAseq to generate nuclear expression profiles from FACS-purified EGFP+ nuclei from transgenic mouse lines expressing EGFP-L10a ribosomal fusion proteins targeted separately to five cell types: cerebellar Purkinje cells, granule cells, Bergmann glia, corticopontin pyramidal neurons, or corticothalamic pyramidal neurons. Neurod1, Pcp2, and Sept4 were known to promote expression in cerebellar granule cells, Purkinje cells, and Bergmann glia, respectively. Colgalt2 (or Glt25d2) and Ntsr1 were known to promote expression in cortical corticopontin and corticothalamic pyramidal cells. Known markers were confirmed to be specific for the nuclei of each cell type.

[0181] Table 11 provides a summary of the nuclear RNAseq data sets from transgenic mice developed for use with the TRAP method. [Table 12]

[0182] We performed hierarchical clustering of the nuclear RNAseq data for the nuclei of the five cell types presented in Table 11. In addition, three published nuclear RNAseq datasets from Mo et al. (Neuron 2015) were included in this hierarchical clustering: excitatory neurons from Camk2a-cre animals (labeled excit), PV interneurons from PV-cre animals (labeled pv), and VIP interneurons from VIP-cre animals (labeled vip). We observed that pyramidal neurons clustered together.

[0183] Expression calculated from the RNAseq results for each gene was normalized to the average expression across samples. The log2 fold change results of RNAseq from FACS-purified nuclei of each of the five cell types are shown in Table 12, compared to the normalized RNAseq results of the 250 most variable genes. The log2 fold change of expression was calculated for nuclei from each cell type in triplicate. Negative values ​​indicate reduced expression in the nuclei of that cell type compared to the normalized value across nuclei from all analyzed cell types, and positive values ​​indicate increased expression compared to the normalized value across nuclei from all analyzed cell types. [Table 13] TIFF2025032204000400.tif249165TIFF2025032204000401.tif248165TIFF2025032204000402.tif248165TIFF20250322040 00403.tif249165TIFF2025032204000404.tif248165TIFF2025032204000405.tif248165TIFF2025032204000406.tif177165

[0184] For RNAseq data from Mo et al. (2015) Neuron 86, 1369-1384 (incorporated herein by reference in its entirety), the log2 fold change in expression from the RNAseq results for the same 250 most variable genes found in Table 12, compared to normalized expression in nuclei from all analyzed cell types, is shown in Table 13. Negative values ​​indicate reduced expression in the nuclei of that cell type compared to the normalized value for whole nuclei from all analyzed cell types, and positive values ​​indicate increased expression compared to the normalized value for whole nuclei from all analyzed cell types. [Table 14] TIFF2025032204000408.tif250165TIFF2025032204000409.tif249165TIFF2025032204000410.tif175165

[0185] Additional cell type-specific nuclear transcripts were identified from genes observed to have increased expression: alpha-2-macroglobulin (A2m) for Sept4-positive cells (Bergmann glia), Pde1a and Colgalt2 for Colgalt2-positive cells (corticopontin pyramids), Pde1a and heparan sulfate-glucosamine 3-sulfotransferase 4 (Hs3st4) for Excit-positive cells (excitatory neurons), Pde1a and Hs3st4 for Ntsr1-positive cells (corticothalamic pyramids), aristares-related homeobox (Arx) and parvalbumin (Pvalb) for PV-positive cells (parvalbumin interneurons), vasoactive intestinal peptide (Vip) for VIP-positive cells (VIP interneurons), solute transporter family 9 member A3 (Slc9a3) for Pcp2-positive cells (Purkinje), and Grin2c for NeuroD1-positive cells (granules). The results of this Example herein provide evidence that nuclear RNA is sufficient to determine cell type.

[0186] Hierarchical clustering was performed using genes from all eight of these cell types to determine whether the nuclear RNA profiles could classify samples based on predictive biology. Clustering was performed not only by cell type, but also by groups of related cell types. For example, two classes of interneurons (VIP and PV) clustered with each other, while three pyramidal cell types (corticopontin and corticothalamus from the data herein and all excitatory pyramidal cell types from Mo et al.) clustered together. The robustness of this clustering demonstrates the utility of nuclear RNA profiles to specify cellular identities and measure relatedness of different cell types, even across samples prepared using different methods and from different laboratories.

[0187] Example 5. Development of a method to isolate specific nuclei from wild-type animals Previous cell profiling techniques required the breeding of disease models in transgenic animals. Some disease models are difficult to breed, and when aging studies are performed, it is necessary to maintain specific transgenic lines for several years. Such methods are not applicable to organisms where genetic modification is difficult or unethical, as in the case of humans. Therefore, there is a need to develop generalizable methods that allow molecular studies of specific cell types in wild-type animals.

[0188] Figure 2A shows the sorting of nuclei from mouse cerebellum using the strategy described in Example 2. Itpr1 is on the x-axis and NeuN is on the y-axis. The Itpr1+ population were Purkinje cell nuclei and the NeuN+ population were glial nuclei. The NeuN+ / Itpr1- nuclei (largest group) represent granule cells.

[0189] Figure 2B shows gene expression profiling from RNAseq data comparing a panel of Purkinje nuclei RNA from transgenic animals from GFP sorting to Itpr1+ nuclei sorted with the cell profiling technique described herein. Pcp4 is a Purkinje marker that was observed to be enriched in Pcp2-GFP+ nuclei and wild type (WT) Itpr1+ nuclei compared to cerebellar unsorted nuclei (Pcp2 unsorted and WT unsorted). Calb2 is a granule cell marker that was observed to be enriched in unsorted samples (Pcp2 unsorted and WT unsorted) compared to sorted Purkinje nuclei (Pcp2-GFP+ or WT Itpr1+).

[0190] Table 14 provides the Pearson correlation showing that the number of Purkinje nuclei from GFP and antibody sorting are highly correlated (R=0.99). [Table 15] TIFF2025032204000412.tif8165

[0191] Antibodies specific for these factors would allow an increase in the number of cell types expressing these factors to be isolated.

[0192] Midbrain dopaminergic neurons were also stained with a combination of two antibodies: one against the transcription factor FoxA1 and the other against the membrane-localized dopamine transporter (DAT). Staining for FoxA1 was observed in the nuclei of dopaminergic neurons. Staining for DAT was observed mainly in the axons, but at high magnification it was also observed in the cell bodies, where it was probably synthesized in the endoplasmic reticulum.

[0193] Nuclear staining and sorting of dopaminergic nuclei by flow cytometry is shown in Figure 3A. The double positive (DP) population (positive for FoxA1 and DAT) was isolated for RNAseq analysis. Figure 3B shows a comparison of RNAseq recordings from unsorted midbrain nuclei and sorted dopaminergic nuclei. Dopaminergic nuclei were observed to be positive for known markers-solute transporter family 6 member 3 (DAT, also known as Slc6a3), tyrosine hydroxylase (TH), forkhead box A1 (FoxA1), forkhead box A2 (FoxA2), and dopamine receptor D2 (Drd2). Meanwhile, genes known not to be expressed in these neurons (Drd1 and glutamic acid decarboxylase (Gad)) were observed to be depleted in the unsorted population. Thus, dopaminergic neurons were observed to be successfully isolated in mice. This method can also be used for human samples.

[0194] Wolframin ER transmembrane glycoprotein (Wfs1) and B-cell CLL / lymphoma 11B (Ctip2, also known as Bcl11b) were observed to be putative markers of the cortical cell types layer 2 / 3 pyramidal neurons and layer 5 and 6 pyramidal neurons, respectively, from staining. Another putative marker of cortical cell type observed from staining is ETS variant 1 (Etv1, also known as ER81), which labels layer 5 pyramidal neurons. Etv1 has previously been observed to label both corticostriatal neurons, corticopontin neurons, and possible others in layer 5 of the cortex. Two putative markers of the hippocampus were also identified from staining: Wfs1 for carbonic anhydrase 1 (CA1)-positive neurons and Purkinje cell protein 4 (Pcp4) for carbonic anhydrase 1 (CA2)-positive dentate gyrus.

[0195] Putative markers for brainstem motor neurons were observed to be solute transporter family 18 member A3 (VaChT, also known as Slc18a3), chondrolectin (Chodl), and EsrrB. Antibodies specific for the three genes VaChT, Chodl, and ErrB were observed to label this population of neurons in the brainstem by in situ hybridization. The Allen Brain Atlas gene expression database was used as a reference. VaChT and ErrB specifically label motor neurons, while Chodl labels everything but these neurons. Antibodies specific for Chodl may be substituted for those that directly label brainstem motor neurons.

[0196] Example 6. Application of antibody selection method to wild-type derived nuclei To determine whether the cell profiling techniques described herein could be used productively for cell type-specific interrogation in wild-type animals, the cell profiling techniques described herein were applied to processed tissue from the mouse cerebellum, a structure composed of several classically defined, neuronal and glial cell types that exhibit distinctive morphologies and distributions.

[0197] Candidate antibody specificities were selected from expression profiles from each cell type generated from data obtained from transgenic animals expressing an EGFP / L10a fusion protein and developed for use in the TRAP method (Dougherty, et al. (2010) Nucleic Acids Res 38, 4218-4230 and Doyle, et al. (2008) Cell 135, 749-762, each of which is incorporated herein by reference in its entirety). Data were collected from three neuronal cell types (granule, Purkinje, and basket) and two glial cell types (astrocytes and oligodendrocytes) in the cerebellum of wild-type mice.

[0198] Immunofluorescence staining was performed on five separate cell types in the cerebellum using methods similar to those applied to cells from transgenic animals. The specificities of the antibodies used for each cell type to identify differences between nuclei were NeuN for granule cells, Itpr1 for Purkinje cells, Sorcs3 for basket cells, Gfap labeling for astrocyte cell bodies and processes, and MoG labeling for oligodendrocyte cell bodies and processes.

[0199] To screen, a combination of at least two antibodies was used: NeuN, a splicing factor localized to the euchromatin inside the nucleus, which labels neuronal nuclei, and cell type specific antibodies: Itpr1, an endoplasmic reticulum membrane protein localized to the nuclear envelope, as well as the basket cell marker Sorcs3 and the astrocyte marker Eeat1 (also called Glast), two plasma membrane proteins that are labeled in the nuclear envelope. Antibodies against an oligodendrocyte marker and the transcription factor Olig2 showed labeling in euchromatin.

[0200] It is known in the art that in cerebellar tissue, NeuN labels the nuclei of granule neurons, but not Purkinje and basket neurons, or glia. Immunofluorescence results on tissue sections with NeuN confirmed this result. However, staining of nuclei with NeuN revealed that basket cell, astrocyte, and oligodendrocyte nuclei have lower levels of NeuN, as expected, while Purkinje nuclei are unexpectedly observed to have high levels. This is not due to spectral overlap of Purkinje-specific markers, since confocal imaging showed bright expression of NeuN in Purkinje nuclei in euchromatin - the expected distribution for splicing factors. Thus, the exposure of the NeuN epitope was observed to be different in isolated nuclei and tissue sections.

[0201] To isolate nuclei from oligodendrocytes, antibodies against the transcription factor Olig2 were used. Immunofluorescence was used to confirm the validity of these antibodies. To isolate nuclei from granule cells, a combination of Itpr1, Sorcs3, and Olig2, plus NeuN, was used to remove contaminating nuclei from other cell types. Nuclei were counterstained with DAPI, which is also a marker for heterochromatin. FACS results are shown in Table 15. [Table 16]

[0202] Thus, the sorting methods used to isolate nuclei in cell types from transgenic animals developed for use in the TRAP method were observed to be effective on nuclei from the same cell types in wild-type mice.

[0203] Example 7. Correlation of results from different cell sorting methods Results were analyzed at a genome-wide level by calculating pairwise Pearson's correlation coefficients (r) of normalized gene expression as shown in Table 16 in GFP+ sorted nuclear samples from animals developed for use with the TRAP method and nuclear samples sorted using the cell profiling techniques described herein. The ".g" samples were GFP+ sorted samples from animals developed for use with the TRAP method and the ".a" samples were sorted using the cell profiling techniques described herein. [Table 17] TIFF2025032204000415.tif245162

[0204] From this analysis, correlations between different cell types in the cerebellum were observed to be high: r=0.91-0.94, and correlations between biological replicates were observed to be extremely high: r=0.98-1. For gene expression profiles from EGFP-L10a labeled nuclei from granule cells, Purkinje cells, and Bergmann glia, extremely high correlations were observed between genetically identified and cell profiling specific cell types: r=0.98 for granule cells, r=0.99 for Purkinje cells, and r=0.96-0.97 between Bergmann glia purified from Sept4 / EGFP-L10a animals and astrocytes purified by the cell profiling technique described herein. Since Bergmann glia are a subpopulation of cerebellar astrocytes, it makes sense that these two data sets are not perfectly correlated with each other.

[0205] Hierarchical clustering of samples revealed clustering according to known biology. For example, in the first division, neuronal and glial samples were separated from each other. In the second division, the five cell types were separated from each other, regardless of whether the nuclei were purified using genetic or antibody labeling. Taken together, the results herein show that the cell profiling techniques described herein allow cell type-specific gene expression profiling of mice without the need for transgenic animals.

[0206] Example 8. Antibody-based selection of cell types in rats Nuclei were isolated from rat cerebellum, fixed, and stained with the same antibodies used in mice, and the results were analyzed by flow cytometry. The FACS results of nuclear samples from Sprague-Dawley rats are shown in Table 17. [Table 18]

[0207] The staining profile of rat nuclei was similar to that of mouse. However, there were differences, for example in the rat cerebellum, where NeuN staining yielded more distinct positive and negative populations than were evident in the mouse cerebellum. As a result, additional populations of nuclei were revealed by flow cytometry analysis, which were isolated and characterized by RNAseq. Nuclei were isolated from basket cells, astrocytes, and oligodendrocytes by staining with antibodies specific for Sorcs3 and NeuN. This flexibility allowed the purification of all six cell types using four staining methods. For cell types isolated from more than one type of staining, the population identified by RNAseq as most enriched for known markers and depleted for markers of other cerebellar cell types was selected.

[0208] To determine the set of genes specific to each rat cell type, differential gene expression analysis (DESeq2) was performed as described in Example 1 to compare the nuclear expression profile of each cell type to unsorted cerebellar nuclei. The log2 fold change (adjusted p<0.01, fold change>2) of expression of 120 differentially expressed genes between the normalized RNAseq results from the nuclei of the six cell types compared to the RNAseq results of sorted nuclei from each individual cell type is shown in Table 18. Negative values ​​indicate reduced expression in the nuclei of that cell type compared to the normalized value of the whole nuclei from all analyzed cell types, and positive values ​​indicate increased expression compared to the normalized value of the whole nuclei from all analyzed cell types. Duplicate samples of nuclei from each cell type were analyzed. [Table 19] TIFF2025032204000418.tif245164TIFF2025032204000419.tif245164TIFF2025032204000420.tif23165

[0209] In Table 18, an overall enrichment of known markers for each cell type and depletion of markers for all other cell types was observed. Nuclear expression of Fat2 and Rbfox3 was observed to be significantly enriched in granule cells. Nuclear expression of Car8 and Calb1 was observed to be specific to Purkinje cells. Nuclear expression of Aldh1L1 and Slc1a3 was observed to be specific to Bergmann glia. Nuclear expression of Colgalt2 was observed to be specific to corticopontin pyramidal cells. Nuclear expression of Hs3st4 was observed to be specific to corticothalamic pyramidal cells, while Pde1a and Csmd1 were both observed to be expressed in cortical pyramidal cell types. In addition to the 120 genes in Table 18, the data also identified 700 to 1,600 expressed genes that were significantly enriched in both rat cell types.

[0210] The most specific genes for each cell type were also identified by implementing the latest version of the Specificity Index (SI) algorithm (Dougherty, et al. (2010) Nucleic Acids Res 38, 4218-4230, which is incorporated herein by reference in its entirety). The advantage of SI is that it can identify differentially expressed genes even in highly abundant cell types in a region of interest (e.g., cerebellar granule cells).

[0211] To validate both the cell type-specific rat data and the current SI algorithm, the expression of genes identified as specific to each cell type was confirmed using the Allen Mouse Brain Atlas in situ hybridization database. For the top 20 most specific genes calculated by the SI algorithm in Example 1 for all cell types, labeling was observed specifically in the determined cell type with an average probability of 93%. The percentages of the top 20 SI genes from the Allen Mouse Brain Atlas images showing the expected distribution of each cell type were: granule 93%, Purkinje 100%, basket 85%, astrocyte 93%, oligodendrocyte 100%, and OPC 80%. The distribution of the astrocyte-specific gene Itih3 was also observed to be characteristic of both the astrocyte type Bergmann glia and the non-Bergmann glia astrocyte type of the cerebellum. The following genes were selected as additional markers for the nuclei of the six cell types: tescalcin (Tesc) for granule cell nuclei, carbonic anhydrase 8 (Car8) for Purkinje cell nuclei, March11 for basket cell nuclei, Inter-Alpha-Trypsin Inhibitor Heavy Chain 3 (Itih3) for astrocyte nuclei, Mog for oligodendrocyte nuclei, and platelet-derived growth factor receptor alpha (Pdgfra) for oligodendrocyte progenitor cell nuclei.

[0212] These results demonstrate that the cellular profiling technique described herein is a generalizable method that can be readily adapted for use in any species.

[0213] Example 9. Identification and comparison of the most specific genes in each species To gain insight into cell type specific functions that changed across species, a specificity index was calculated for all genes in each species. The 100 most specific genes for each species are shown in Tables 19-24 for each cell type. Lower SI values ​​indicate genes that are more specifically expressed in one species than another.

[0214] Table 19 provides the 100 most specific genes, as measured by the specificity index (SI), in granule cells in human, rat, and mouse samples, as described in Example 1. [Table 20] TIFF2025032204000422.tif254165TIFF2025032204000423.tif165165

[0215] Table 20 provides the 100 most specific genes for Purkinje cells in human, rat, and mouse samples as measured by specificity index (SI). Table 20. The 100 most specific genes by measuring SI for each species in Purkinje cells [Table 21] TIFF2025032204000425.tif245160TIFF2025032204000426.tif245163

[0216] Table 21 provides the 100 most specific genes for basket cells in human, rat, and mouse samples as measured by specificity index (SI). [Table 22] TIFF2025032204000428.tif245159TIFF2025032204000429.tif77164

[0217] Table 22 provides the 100 most specific genes for astrocytes in human, rat, and mouse samples as measured by specificity index (SI). [Table 23] TIFF2025032204000431.tif245164TIFF2025032204000432.tif149165

[0218] Table 23 provides the 100 most specific genes for oligodendrocytes in human, rat, and mouse samples as measured by specificity index (SI). [Table 24] TIFF2025032204000434.tif245162TIFF2025032204000435.tif223162

[0219] Table 24 provides the 100 most specific genes for OPCs in human, rat, and mouse samples as measured by specificity index (SI). [Table 25] TIFF2025032204000437.tif245163TIFF2025032204000438.tif49165

[0220] The rank of the 100 most specific genes in each cell type in mouse was compared to their rank in rat or human, as shown in Table 25. [Table 26] TIFF2025032204000440.tif245164TIFF2025032204000441.tif129165

[0221] It was observed that the specificity ranks were more conserved between mouse and rat than between mouse and human.

[0222] Example 10. Determining the extent to which cell type-specific genes vary between species To further explore the extent to which cell type-specific genes vary between species, comparative analyses were focused on data from mouse and human nuclear samples because of the low completeness of the annotation of the rat genome. By excluding rat genes, the number of high-confidence orthologs available for comparison will increase from 11,443 to 14,273 by selecting ensemble orthologs.

[0223] Additionally, to focus on differentially expressed genes that are most likely to impact cellular function, genes that passed a stringent adjusted p-value cutoff of <0.01, changed at least 4-fold between species, and were expressed at significant levels in each cell type (mean counts >400 and fpkm >0) were selected. The numbers retained after sorting by expression level (baseMean >50, log10(fpkm) >0) and the numbers retained after sorting by differentially expressed genes across all cell types between mouse and human were broken down into human and mouse enriched genes as shown in Table 26. [Table 27]

[0224] According to Table 26, 490-823 genes were identified that were enriched in mouse or enriched in human in these cell types.

[0225] To identify cell type and species differences rather than general differences between mouse and human, genes observed to be differentially expressed in all cell types analyzed (regulated p<0.01, fold change >2) were excluded from the analysis. Such selection yielded 133-293 cell type- and species-specific differentially expressed genes that were approximately evenly distributed between those enriched in mouse and those enriched in human.

[0226] To confirm these data, the abundance of granule cell neurons relative to all other cell types in the cerebellum allowed two types of validation not available for less abundant cell types: first, in addition to assaying expression of nascent nuclear RNAs, expression of mature cytoplasmic RNAs was investigated; second, given the presence of a high proportion of granule cell genomes in cerebellar nuclei preparations, chromatin accessibility was assessed using unsorted cerebellar nuclei. Because previous studies have shown that ATACseq peaks are typically found in promoters and within genomic regions of expressed genes and are greatly reduced or absent in non-expressed genes (Buenrostro, et al. (2013) Nat Methods 10, 1213-1218; Mo, et al. (2015) Neuron 86, 1369-1384; Su, et al. (2017) Nat. Neurosci. 20, 476-483, each of which is incorporated herein by reference in its entirety), ATACseq data from mouse, rat and human cerebellar nuclei were used to confirm the expression data.

[0227] As expected, protein encoding genes identified as expressed in data from the cell profiling techniques described herein from all three species were accompanied by the presence of spliced ​​cytoplasmic mRNA, and ATACseq peaks were evident in their promoters and within the region of the gene. For example, the housekeeping gene Gapdh and the granule-specific gene Etv1 are strongly expressed in mouse, rat, and human granule cells. As expected, the presence of nuclear RNA, cytoplasmic RNA, and ATACseq peaks over these genes was observed. Apart from the human enriched genes Clvs2 and Vwc2, these three independent measures of active expression are present in the human data, but no expression or chromatin accessibility was observed in the mouse data. Conversely, for the mouse enriched genes Cnksr3 and Ece1, nuclear and cytoplasmic RNA expression and ATACseq peaks were observed in the mouse data, but not in the human data. As expected, expression of these genes in rat was similar to mouse, although low levels of the human enriched genes Clvs2 and Vwc2 were observed to be present in rat.

[0228] Two additional features of expression between mice and humans were identified. Changes in expression can involve more than one gene at a given genomic span. For example, at a locus containing G protein-coupled receptor 135 (Gpr135), trans-L-3-hydroxyproline dehydratase (L3hypdh), JNK1 / MAPK8-associated membrane protein (Jkamp), Ccdc175, and reticulon 1 (Rtn1), all five genes were expressed in human granule cells, as revealed by the nuclear and cytoplasmic gene expression and the presence of ATACseq peaks in Figure 4. Jkamp, ​​L3hypdh, and Gpr135 were also observed to be expressed in mice and rats, as supported by the presence of promoter DNA accessibility sites indicated by ATACseq peaks, but at much lower levels compared to their human orthologues. Rtn1 was observed to be expressed at lower levels in mice than in rats and humans, but different patterns of ATACseq peaks were observed in each sample. In mice and rats, Ccdc175 was not observed to be expressed, and no ATACseq peaks were observed, indicating that the genes in this region form two distinct domains, one containing Rtn1 and the other containing Gpr135, L3hypdh, Jkamp, ​​and Ccdc175. In the rodent data, Gpr135, L3hypdh, and Jkamp are equally accessible to ATACseq. Therefore, the reduced nuclear and cytoplasmic RNA levels of these genes were likely due to altered chromatin structure of the adjacent and non-expressed gene, Ccdc175.

[0229] A second characterization of expression between mouse and human data showed that changes can involve the expression of alternative family members in a given cell type. For example, the Pde1 gene family contains three calcium- and calmodulin-dependent phosphodiesterases, Pde1a, Pde1b, and Pde1c. These enzymes catalyze the hydrolysis of cAMP or cGMP and play important roles in regulating neuronal activity. Using ATACseq, it was confirmed that mouse granule cells express only Pde1c, whereas granule cells from humans express high levels of Pde1a and low levels of Pde1c. These results were consistent with previous findings in Loughney, K., & Ferguson, K. (1996) Phosphodiesterase inhibitors (pp. 1-19); Loughney, et al. (1996) J Biol Chem 271, 796-806; and Sonnenburg, et al. (1993) J Biol Chem 268, 645-652 (each of which is incorporated herein by reference in its entirety), that Pde1a is expressed in humans but not in mouse cerebellum. Because Pde1a preferentially hydrolyzes cGMP, whereas Pde1c hydrolyzes both cAMP and cGMP with equal efficiency (Takimoto, E. (2009) Circ. Res. 105, 931-933, which is incorporated herein by reference in its entirety), these differences may result in slightly different biochemical outcomes in mouse and human cerebellum.

[0230] To confirm the presence or absence of these mRNAs in the mouse brain, gene expression data from the publicly available Allen Mouse Brain Atlas and GENSAT projects (Gong, et al. (2003) Nature 425, 917-925; Lein, et al. (2007) Nature 445, 168-176, each of which is incorporated herein by reference in its entirety) were analyzed. These data confirmed the expression of the granule-specific marker Etv1 and mouse enriched genes Cnksr3, Ece1, and Pde1c in the granular layer of the cerebellum, and the absence of expression of the human enriched genes Ccdc175, Clvs2, Vwc2, and Pde1a. Interestingly, although Clvs2 and Pde1a are absent from the cerebellum, they are expressed in other regions of the mouse brain, suggesting that alterations in cell-specific regulatory sequences may underlie these species differences (Carroll, SB (2005) PLoS Biol 3, e245; Carroll, SB (2008) Cell 134(11), 25-36, each of which is incorporated herein by reference in its entirety). Ece1 expression was also observed in the Purkinje layer of the cerebellum. Expression of the human enriched genes Ccdc175, Clvs2, Vwc2, and Pde1a was observed in the mouse brain. No staining was detected in any of the regions for Ccdc175 and Vwc2. Strong staining for Clvs2 was observed in the dentate gyrus, and for Pde1a in layers 5 and 6 of the cortex and CA1-3 of the hippocampus.

[0231] To further validate the data, qPCR was used to measure the relative expression of mouse-enriched and human-enriched genes. qPCR analysis of gene expression was performed for housekeeping, human-enriched, and mouse-enriched genes using human and mouse cerebellar cytoplasmic cDNA. For each species, two biological replicates and three technical replicates were used. Expression was normalized to the arithmetic mean of expression from the three housekeeping genes. Fold changes were calculated relative to expression in human XK samples. p-values ​​were calculated using a two-tailed Student's t-test on the mean normalized values ​​of the biological replicates (mean of the technical replicates) assuming unequal variances as follows: 1) housekeeping genes-Actb=0.334, Gapdh=0.141, and Polr2a=0.109; 2) human enriched genes-Clvs2=0.0195, Vwc2=0.0161, and Pde1a=0.00086; and mouse enriched genes-Pde1c=0.0696. p-values ​​were not calculated for Ccdc175, Cnksr3, and Ece1 because all technical replicates failed to amplify at least one biological replicate.

[0232] In all but one case, qPCR results corroborated the RNAseq and ATACseq data. In the outlier case, Pde1c showed no significant change in expression between mouse and human. To understand this difference, Pde1a and Pde1c expression was assayed by immunofluorescence in mouse and human cerebellar sections. At least two stainings were performed each on sections from two separate mice and two donors. NeuN was used as a marker for granule cells. In line with the RNAseq and ATACseq data, we observed that members of both Pde1c and Pde1a were primarily localized to the granule layer, with Pde1c expression dominating in mouse cerebellum and Pde1a expression dominating in human cerebellum. Mouse cerebellum specifically showed Pde1c in a background labeling granule cells and Pde1a, whereas human cerebellum specifically showed Pde1a in a background labeling granule cells and Pde1c.

[0233] To understand whether these genes may reflect expression changes occurring independently or transcriptional programs that vary between species, a gene ontology (GO) analysis was performed on 10 categories of species-specific genes. Four gene groups yielded significant GO clusters: human enriched astrocytes, human enriched oligodendrocytes, human enriched OPCs, and mouse enriched granule genes. The remaining six genes were not observed to have significantly associated GO categories, suggesting that species differences in these cell types contribute to a wide range of pathways and functions. GO analysis of human enriched genes from astrocytes, oligodendrocytes, or OPCs revealed that these genes are involved in cell structure, size regulation, and nervous system function, potentially reflecting known differences in morphology between mouse and human glial cells. Granule cell mouse enriched genes are involved in polysaccharide metabolism, which may play a role in nervous system function and cell-cell interactions.

[0234] Taken together, these data indicate that the cell profiling techniques described herein can be employed in the nucleus to accurately assess cell type-specific gene expression in rodent and human brains, and in the nervous system. Gene expression in the brain was observed to be highly conserved in specific cell types across species. Thus, the most abundant and cell-specific markers of well-characterized cell types were observed to be shared in rodent and human brains. Despite this shared uniqueness, there were a significant number of genes in each cell type whose expression was not conserved across species. This set of genes typically does not fit known GO categories, and their expression is often found in other brain regions. Cell type and species-specific expression can reflect altered regulation of neighboring genes within a locus, or selective expression of functionally related but distinct members of a given gene family. Although the consequences of these gene expression changes need to be interrogated in future studies, the data herein indicate that functionally important differences in gene expression in specific CNS cell types occur across species, and that they may result in important differences in the biochemical function of rodent and human cell types.

[0235] Example 11. Comparison of mouse and human nuclear transcriptomes Immunofluorescence was used to determine whether the sorting yielded a pure population of human samples. Nuclear structure was observed to be quite different between the three populations, and the size of the nuclei in each population varied. NEUN+ granule neurons were observed to be small and highly heterochromatic. NeuN- nuclei were observed to be a mixture of larger nuclei likely derived from Bergmann glia and smaller nuclei likely derived from oligodendrocytes. ITPR1+ nuclei were observed to be larger and more euchromatic than the nuclei of the other two groups. Figure 6 shows the flow cytometry sorting results, which approximate the size of the nuclei with a combination of forward scatter (FSC) and side scatter (SSC). (DP=double positive). These results indicate that nuclei from different cell types are distinct in nuclear structure, which may result in differences across each cell type in gene expression regulation or disease susceptibility.

[0236] However, heterogeneity of nuclei sorted from two different human subjects (person 1 and person 2) was observed in the flow cytometry results in Figures 5A and 5B, respectively. Species-specific markers can be identified to distinguish nuclei from different cell types.

[0237] Figure 7A, Figure 7B, Figure 7C, and Figure 7D provide flow cytometry results of four different staining methods of human cerebellar nuclei developed from the method used to sort mouse nuclei. Figure 7A shows flow cytometry results by gating on Itpr1 and Id2 of Purkinje cells (i.e., Itpr+ and Id2+). Figure 7B shows results by gating on forkhead box P4 (FoxP4) and Itpr1 of Purkinje cells (i.e., FoxP4+ and Itpr+). Figure 7C shows results by gating on NeuN and Itpr1 of granule cells (i.e., Itpr- and NeuN+) using the same sorting method as mouse. Figure 7D shows results by gating on FoxP4 and NeuN. The lightest grey populations of all four staining methods give the same population of nuclei as shown by gene expression profiling from RNAseq shown in Table 37.

[0238] Gene expression profiles from RNAseq results from mouse and human sorted nuclei are shown in Examples 15 and 17, respectively. Granule cell nuclei populations in Table 37 of Example 17 were observed to have a similar profile to granule cells from mouse, as shown in Table 31 of Example 15. Glial nuclei populations in Table 37 were observed to have a similar profile to glia from mouse (Tables 33 and 34). Basket nuclei populations from human samples in Table 37 were observed to have a similar profile to basket / astrocytes from mouse in Table 32. These similarities were particularly noticeable in the RNAseq results for Rbfox3 (also known as NeuN), Fat2, ryanodine receptor 1 (Ryr1), Pvalb, glutathione S-transferase mu 1 (Gstm1) and pleiotrophin (Ptn).

[0239] In contrast, in the RNAseq data in Tables 31-34 and 37, significant changes were observed between mouse and human for protein tyrosine kinase 2 beta (Ptk2b), solute transporter family 43 member 2 (Slc43a2), glutamate ionotropic receptor kainate type subunit 3 (Grik3), Itpr1, protocadherin 9 (Pcdh9), and RAS guanyl releasing protein 1 (Rasgrp1)-positive nuclei from granule, basket / stellate, and glial cells.

[0240] Hierarchical clustering of the nuclear RNAseq data (first by species and then by cell type) from the top 500 genes that contributed the most variation between samples, and therefore best discriminated one sample from another, revealed several gene groups that discriminated between cell types and other gene groups that discriminated between species. This analysis revealed conserved and divergent genes between mouse and human.

[0241] Principal component analysis of mouse and human cerebellar samples allows visualization of clustering by cell type and by species. The first principal component (x-axis) represents the largest variation between samples, and the second principal component represents the second largest source of variation between samples and how they are separated by cell type. The analysis herein showed that cell type contributes X% of the variance between samples, while species contributes Y% of the variance between samples.

[0242] Example 12. Identification of cell type and species specific genes through cross-species comparison Given the high quality of the cell-specific expression profiles described above, we analyzed the extent of gene expression changes between species in these cerebellar cell types. To ensure that results reflected expression changes and not annotation differences between species, we limited the analysis to high-confidence 1:1 orthologs between mouse, rat, and human. To define orthologous annotations for mouse, rat, and human genes, orthologous annotations across species were downloaded from the ensemble (304,147 transcripts vs 19,197 genes) and filtered to include only high-confidence pairwise (76,172 transcripts vs 14,541 genes) 1:1 orthologs (67,522 transcripts vs 13,242 genes) genes with total length greater than 1 kb and less than 2-fold change in length between all species (30,497 transcripts vs 11,443 genes). For each gene, the longest orthologous transcript was used for annotation, resulting in 11,443 transcripts and 11,443 genes.

[0243] The log2 fold change in expression of the 250 most variable genes in whole mouse, rat, and human nuclei among the normalized RNAseq results from the nuclei of the seven cell types, compared to the RNAseq results of selected nuclei from each individual cell type, is shown in Table 27. Negative values ​​indicate reduced expression in the nuclei of that cell type compared to the normalized value of whole nuclei from all analyzed cell types, and positive values ​​indicate increased expression compared to the normalized value of whole nuclei from all analyzed cell types. Duplicate samples of nuclei from each cell type were analyzed. [Table 28] TIFF2025032204000444.tif245162TIFF2025032204000445.tif245160TIFF20250322040 00446.tif245160TIFF2025032204000447.tif245160TIFF2025032204000448.tif216166

[0244] The log2 fold change in expression of the 250 most variable genes in mouse, rat, and human whole nuclei among the normalized RNAseq results from the nuclei of the seven cell types, compared to the RNAseq results of selected nuclei from each individual cell type, is shown in Table 28. Negative values ​​indicate reduced expression in the nuclei of that cell type compared to the normalized value of whole nuclei from all analyzed cell types, and positive values ​​indicate increased expression compared to the normalized value of whole nuclei from all analyzed cell types. Two replicate samples of nuclei from each cell type were analyzed. [Table 29] TIFF2025032204000450.tif245164TIFF2025032204000451.tif245164TIFF20250322040 00452.tif245164TIFF2025032204000453.tif245164TIFF2025032204000454.tif116166

[0245] Examination of the first eight principal components: proprotein convertase subtilisin / kexin type 1 (PCSK1, also known as PC1) through proprotein convertase subtilisin / kexin type 8, (PCSK8, also known as (Pc1-Pc8)), which collectively account for 90.5% of all sample variance, revealed that expression of Pc1 and proprotein convertase subtilisin / kexin type 3 (PCSK3, also known as PC3) classify samples by cell type, whereas Pc2, proprotein convertase subtilisin / kexin type 6 (PCSK6, also known as PC6) and PC8 divide samples by species. The other principal components organize samples by both cell type and species. Since Pc1 and Pc3 accounted for 47% of the variation across samples, while Pc2, PCc6, and Pc8 accounted for only 22% of the variation, it was determined that the most significant differences between profiles arise from cell type-specific expressed genes shared between species. However, it is clear that genome-wide quantitative differences in expression profiles between species should also be considered when evaluating the functional properties of a given cell type in various species.

[0246] Hierarchical clustering was used to examine the similarities and differences across species for mouse, rat, and human cell type-specific datasets based on the expression of all genes except the 250 most variable genes, showing clustering primarily by species, however, when clustering was performed using only the 250 most variable genes across species, clustering was observed to be primarily by cell type.

[0247] Example 13. Profiling of three cell types from 16 postmortem human brains Taking into account the genetic heterogeneity of human populations, the variation in processing time of human tissues, and the difficulties noted in previous investigations of total RNA isolated from human tissues (McCall, et al. (2016) Am. J. Hum. Genet. 99, 624-635.; Webster, MJ (2006). Functional Genomics and Proteomics in the Clinical Neurosciences, (Elsevier), pp. 3-14, each of which is incorporated herein by reference in its entirety), the source of variation across samples from different individuals was also of interest to assess whether the changes observed in disease were due to the disease itself or the result of variation between individuals. To better understand this source of variation, cell type-specific gene expression was analyzed in an additional 14 human brain samples using the cell profiling techniques described herein. Overall, samples were obtained from 16 control donors, split approximately evenly between genders (7 males and 9 females) and across ages (4 individuals each in the following age groups: 20-30, 40-50, 60-70, and 80-90 years). All samples were obtained from unaffected subjects with a neuropathological diagnosis of "normal adult brain." Table 29 provides additional details about each donor. [Table 30]

[0248] Nuclei were isolated from all 16 samples and stained with antibodies against ITPR1 and NEUN, and nuclei were purified from three different cell types in one sort: granule cells, basket cells, and total glia. Table 30 shows the percentage of each cell type by FACS analysis of nuclei. [Table 31]

[0249] Although staining patterns varied slightly between samples, the granule cell-containing ITPR1-NEUN+ population, the basket cell-containing ITPR1+NEUN- population, and the total glia-containing ITPR1-NEUN- were easily distinguished. The only exceptions were the glial populations from individuals WC and WO, where the ITPR1-NEUN- population did not separate well from the ITPR1-NEUN+ population. Gene expression analysis of these two samples demonstrated that sorting failed to efficiently separate granule cells from glia, and the samples were excluded from further analysis. Using RNAseq, we analyzed 46 human datasets: 16 from granule cells, 16 from basket cell shells, and 14 from glia.

[0250] It has been observed that the cell profiling techniques described herein reduce the variation due to differences in cell type composition, which reduces the overall variation. Little variation was observed between the samples analyzed herein, because in the normal analysis of whole tissue, the variation is due to both differences in tissue composition and variations in each cell type.

[0251] As shown in Examples 8-14, RNAseq analysis of these datasets demonstrated high purity of each nuclear sample because each sample was enriched for the appropriate marker for that cell type and depleted for markers of other cerebellar cell types. For example, only samples from female individuals expressed X-inactive specific transcript (XIST) and only samples from males expressed Y chromosome genes, such as lysine demethylase 5D (KDM5D). Furthermore, pairwise Pearson correlation coefficients between biological replicates of each cell type were observed to be between 0.98 and 1 (Figure 8). These data demonstrated that cell type-specific human expression data obtained by the cell profiling techniques described herein are reproducible, despite small inter-individual differences in the abundance of each cell type.

[0252] Example 14. Summary of differences in cell type-specific nuclear RNA profiles between mice and humans Gene expression profiles were generated for five cerebellar cell types each in mice and humans (granule, Purkinje, basket, astrocytes, and oligodendrocytes) by isolating RNA from sorted nuclei and performing RNAseq.

[0253] Tables 31-34 show gene expression differences between mouse and human. To further focus on differentially expressed genes that may have the most impact on cellular functions, we selected genes that had an adjusted p-value <10E-5, changed at least 4-fold between species, and were expressed at significant levels in each cell type (mean normalized counts / DESeq2 baseMean>400 and log2(fpkm)>0). We observed 119-410 genes that were enriched in mouse or enriched in human. To identify cell type and species differences rather than general differences between mouse and human, genes that were differentially expressed (adjusted p<10e-5, fold change>2) between mouse and human in all cell types analyzed were excluded from the analysis. Such selection yielded 67-277 cell type- and species-specific differentially expressed genes that were approximately evenly distributed between those enriched in mouse and those enriched in human.

[0254] The log2 fold changes of granule cell markers in sorted nuclei from human granule cells compared to normalized RNAseq results from granule cell nuclei from mice are shown in Table 31. Negative values ​​indicate reduced expression in the nuclei of that cell type compared to the normalized value of granule cell nuclei from the mouse analyzed, and positive values ​​indicate increased expression compared to the normalized value of granule cell nuclei from the mouse analyzed. [Table 32] TIFF2025032204000458.tif225160

[0255] Granule cell markers, RNA-binding protein, Fox-1 homolog 3 (Rbfox3, also known as NeuN) and atypical cadherin 2 (Fat2), were observed to be highly expressed in granule cells and not significantly expressed in the other four cell types. Genes with larger values ​​of log2 fold change in expression between human and mouse granule cell nuclei would be more effective as markers of human granule cell-derived nuclei.

[0256] The log2 fold change in expression in sorted nuclei from human basket cells compared to the normalized expression results for sorted nuclei from mouse basket cells is shown in Table 32. Negative values ​​indicate decreased expression in the nuclei of human basket cells compared to the normalized value for nuclei from mouse basket cells, and positive values ​​indicate increased expression compared to the normalized value for nuclei from mouse basket cells. [Table 33] TIFF2025032204000460.tif193165

[0257] The basket cell markers, Sorcs3 and membrane-associated Ring-CH-Type finger 11 (March11), were observed to be enriched in basket cells and not significantly expressed in the other four cell types.

[0258] The log2 fold change in expression in sorted nuclei from human astrocytes compared to the normalized expression results for sorted nuclei from mouse astrocytes is shown in Table 33. Negative values ​​indicate decreased expression in human astrocyte nuclei compared to the normalized value for nuclei from mouse astrocytes, and positive values ​​indicate increased expression compared to the normalized value for nuclei from mouse astrocytes. [Table 34]

[0259] Aldehyde dehydrogenase 1 family members L1 (Aldh1L1) and Slc1a3 were found to be enriched in astrocytes and not significantly expressed in the other four cell types. The log2 fold change in expression in sorted nuclei from human oligodendrocytes compared to normalized expression results for sorted nuclei from mouse oligodendrocytes is shown in Table 34. Negative values ​​indicate reduced expression in human oligodendrocyte nuclei compared to the normalized value for nuclei from mouse oligodendrocytes, and positive values ​​indicate increased expression compared to the normalized value for nuclei from mouse oligodendrocytes. [Table 35]

[0260] The log2 fold change in gene expression of genes differentially expressed in nuclei from human OPCs compared to normalized gene expression of OPCs in mouse nuclei is shown in Table 35. Negative values ​​indicate decreased gene expression in nuclei of human OPCs and positive values ​​indicate increased gene expression in nuclei of human OPCs compared to nuclei from mouse OPCs. [Table 36] TIFF2025032204000464.tif98164

[0261] Olig2 and Mog were observed to be enriched in both mature oligodendrocytes and oligodendrocyte precursor cells (OPCs) and were not observed to be significantly expressed in the other four cell types. Hierarchical clustering of nuclear RNAseq profiles from all cerebellar cell types confirmed that the oligodendrocyte lineage was divided into two groups, mature oligodendrocytes and OPCs, with all oligodendrocyte clustering sets containing mature oligodendrocytes.

[0262] Example 15. Identification of factors that influence human cell type-specific gene expression The main reason for developing the cell profiling technology described herein was to enable direct testing of human cell type-specific expression and its association with biological and clinical factors that may affect cell function. The nuclei to be characterized were obtained from normal cerebellum samples, but differences in gender, age, and time between death and tissue storage (autolysis time) may affect expression data. Table 36 shows the number of genes observed to be significantly changed (adjusted p-value<0.01, baseMean>50) in nuclei as a result of autolysis, gender, and age. [Table 37]

[0263] Age was observed to confer significant changes in the expression of most genes.

[0264] Example 16. Application of antibody-based cell type sorting to postmortem human tissue Previously, inbred animals were used for these types of studies, but heterogeneity was observed across human individuals. To determine whether the cell profiling techniques described herein were effective for analyzing human brain cell types, postmortem human cerebellar tissue from two donors (codes XK and PK) was isolated and nuclei were stained with previously determined cell type specific antibodies. FACS was performed on granule, basket, total glial, astrocyte, mature oligodendrocyte, and oligodendrocyte progenitor cell (OPC) nuclei from the cerebella of two individuals (XK and PK). The percentages of the populations in each gating are shown in Table 37. [Table 38]

[0265] Further analysis of the expression profiling from the results in Table 37 revealed two distinct populations: OLIG2+ low nuclei derived from mature oligodendrocytes, and OLIG2+ high nuclei derived from both oligodendrocyte precursor cells (OPCs) and mature oligodendrocytes.

[0266] After further flow cytometric analysis of oligodendrocytes, the OLIG2+ population appeared to be further divisible based on its levels: approximately 20% of all OLIG2+ nuclei had high levels of OLIG2 (high) and 80% had lower levels of OLIG2 (low).

[0267] The stained profiles of human cerebellar nuclei were generally similar to the rodent data, although differences existed. For example, conditions used to sort granule and basket neurons, astrocytes, oligodendrocytes, and OPCs were directly transferable to human postmortem tissue, whereas conditions sorting mouse and rat Purkinje cell nuclei were not successful in isolating human Purkinje cell nuclei. Subtle differences in staining were observed between the two postmortem human samples: in XK, oligodendrocyte and OPC populations were successfully separated, whereas these cells are more difficult to identify in the adult PK.

[0268] To assess whether these differences might interfere with cell-type specific analysis of human postmortem samples, five cell types (granule, basket, astrocyte, oligodendrocyte, OPC) were analyzed using RNAseq in these two individual brains. Examination of known markers from each of these cell types demonstrated that the cell-specific markers enriched each cell type and depleted the other cell types.

[0269] The log2 fold change between normalized RNAseq results from unsorted nuclei from mouse, rat, and human samples compared to RNAseq results for sorted nuclei from granule cells, basket cells, astrocytes, oligodendrocytes, and OPC cells is shown in Table 38. Negative values ​​indicate reduced expression in the nuclei of that cell type compared to the normalized value for the whole nuclei from all analyzed cell types, and positive values ​​indicate increased expression compared to the normalized value for the whole nuclei from all analyzed cell types. Table 38 shows the RNAseq results for the 250 most differentially expressed genes across all species of two postmortem human samples. [Table 39] TIFF2025032204000468.tif245163TIFF2025032204000469.tif245162TIFF20250322040 00470.tif245164TIFF2025032204000471.tif245164TIFF2025032204000472.tif211166

[0270] RBFOX3 (also known as NeuN) and FAT2 were observed to be significantly enriched in granule cells of both samples. MARCH11 was observed to be significantly enriched in basket cells of both samples. ALDH1L1 and SLC1A3 were observed to be significantly enriched in astrocytes of both samples. Olig2 was observed to be significantly enriched in oligodendrocytes and OPCs of both samples. MOG was observed to be significantly enriched in oligodendrocytes of both samples. PDGFRA was observed to be significantly enriched in OPCs of both samples. Only Olig2+ high nuclei express genes that mark oligodendrocyte precursor cells (OPCs), such as Pdgfra and chondroitin sulfate proteoglycan 4 (Cspg4). Olig2+ low nuclei were derived from mature oligodendrocytes, and Olig2+ high nuclei were derived from a mixture of mature oligodendrocytes and immature OPCs. In each case, high expression of markers was observed only in the predicted cell types. Thus, Olig2 (all oligodendrocytes), Pdgra and Cspg4 (OPCs), Mag and Mog (mature oligodendrocytes) were identified as effective for implementing the cell profiling techniques described herein.

[0271] These results highlight that the cell profiling technique described herein is useful for identifying unexpected subpopulations as a result of its highly quantitative nature. The heterogeneity of results can be used to identify patient biomarkers. For example, heterogeneity can predict how different individuals will respond to drugs or identify therapeutic targets.

[0272] Example 17. Correlation between RNAseq results of each sample Table 39 shows the Pearson coefficients for each cell type for two postmortem human samples, XK and PK. [Table 40]

[0273] Figure 9 compares the results between human, rat, and mouse samples, split by cell type, with pairwise Pearson correlation coefficients for each sample. Calculation of pairwise Pearson correlation coefficients between the 10 data sets revealed that the two neuronal cell types (granule and basket) were highly correlated between samples (r=0.99 for both), while the glial data was slightly less correlated (r=0.96-0.97). Furthermore, hierarchical clustering showed separation by cell type, indicating that the variation across cell types was greater than the variation between individual samples. The range of correlation coefficients between samples of the same cell type across species was observed to be similar to the clustering of cell types isolated from the same species, suggesting that both cell type and species are important properties of the expression profile.

[0274] Example 18. Determining the effect of autolysis on human cell type-specific gene expression Prior literature has shown that autolysis time, defined as the time between death and tissue freezing, does not correlate with RNA quality and only has a minor effect on gene expression (Gupta et al., 2012, BMC Genomics 13, 26, which is incorporated herein by reference in its entirety). Postmortem human samples have previously been observed to vary widely in autolysis time, and therefore the effect of autolysis time on gene expression in nuclear samples from 16 human subjects was previously characterized to determine whether clinical attributes affect gene expression.

[0275] As shown in Table 40, the shortest autolysis time among the samples was 5.13 hours, and the longest autolysis time among the samples was 24 hours. [Table 41]

[0276] Only a few genes were observed to have significantly altered expression with autolysis time: 12 in all postmortem human tissue samples, 4 in granule cells, 1 in basket cells, and 1 in glia. Log2 fold change of differentially expressed genes in all human samples with different autolysis times. Table 41 shows differentially expressed genes (p<0.01, adjusted to baseMean>50) with autolysis time in all samples. [Table 42] TIFF2025032204000476.tif140165

[0277] Npas4 was observed to have expression with the greatest fold change in all nuclei compared to normalized expression in unsorted nuclei from the same sample.

[0278] Table 42 shows genes in granule cells that were differentially expressed (p<0.01, fold change>2 regulation) with various autolysis times compared to normalized expression in all human samples. [Table 43]

[0279] As shown in Table 42, the four genes observed to be differentially expressed in granule cell nuclei were also differentially expressed in unsorted nuclei from all samples.

[0280] Inhibin beta A subunit (Inhba) was observed to be the only differentially expressed gene in the nuclei of basket cells (baseMean: 6.673; log2 fold change: 0.466; p-value: 1.15E-08; p-regulation: 2.18E-04).

[0281] Only kinesin family member 19 (Kif19) was observed to change linearly with autolysis time in glia (baseMean: 126.311; log2 fold change: -0.160; p-value: 7.76E-09; p-reg: 1.25E-04). Additional genes, such as FosB proto-oncogene, AP-1 transcription factor subunit (FosB) (p-reg = 3.51E-12), were observed to be increased in expression at intermediate autolysis times in all three cell types (Figure 10).

[0282] Example 19. Determining the effect of gender on human cell type-specific gene expression To test the effect of gender on expression profiles, gene expression was compared between samples from males and females to identify genes that are differentially expressed (adjusted p-value<0.01, baseMean>50). As expected, most of the gender-specific genes identified herein are present on the X or Y chromosome, with X chromosome genes enriched in female samples and Y chromosome genes enriched in male samples. 27 genes that show gender-specific expression differences in males (adjusted p-value<0.01, baseMean>50) compared to females are shown in Table 43. [Table 44] TIFF2025032204000479.tif43166

[0283] Seven of these genes are located on the X chromosome, 14 on the Y chromosome, and 6 on the autosomes. In some cases, sex-specific genes may be differentially expressed in some cell types more than others. For example, long intergenic non-protein-coding RNA 278 (LINC00278) was observed to be enriched in males and was also differentially expressed more in glia than in granule or basket cells. Furthermore, a protein kinase, Y-linked pseudogene (Prky), was observed to have the highest levels in granule cells, intermediate levels in glia, and low levels in basket cells.

[0284] In contrast, glycogenin 2 pseudogene 1 (GYG2P1) was identified as a male-enriched gene (p=3.74E-42), but this gene was more male-enriched in basket cells than in granule cells and glia. This was not due to mismapped reads due to duplication, because glycogenin 2 (GYG2) is significantly expressed only in glia and does not show sex-specific expression. Examination of the Y chromosome locus containing GYG2P1 revealed male-specific basket cell expression over a large region that includes GYG2P1 and a region approximately 300 kb downstream. Nearby genes testis-specific transcript Y15 (Ttty15) and ubiquitin-specific peptidase 9, Y-linked (Usp9y) were also enriched in males, but no cell type specificity was observed in results from glia, basket, and granule cells. Additionally, Y-linked finger protein (Zfy) and KDM5D were also observed to be enriched in males, but not cell type specific, and XIST was observed to be enriched in females in Table 43, but not cell type specific among glial, basket, and granule cells. Genes showing sex- and cell type-specific expression in both glial and granule cells are also evident. The results herein provide a potential source of sexually dimorphic cell-specific functions in the mammalian brain.

[0285] Transcription was observed only in males and only in astrocytes / basket cells for the genes: family 8 member A4 with sequence similarity, pseudogene (Fam8a4p), arylsulfatase D pseudogene 1 (Arsdp1), and Gyg2p1.

[0286] Thus, sex-specific gene expression was observed to confirm the known sex of an individual, and cell type-specific sex-specific gene expression was also observed.

[0287] Example 20. Determining the effect of age on human cell type-specific gene expression A fundamental question in aging research is whether aging occurs uniformly throughout the brain, or whether some cell types are more susceptible to the effects of aging than others. To address this question, we investigated the effect of age on gene expression in each of the three cerebellar cell types. DESeq2 (Love, et al. (2014) Genome Biology 15,550, which is incorporated herein by reference in its entirety) was used as described in Example 1 to identify genes that were differentially expressed (regulated at p<0.01, fold change>2) by age, or specifically in granule cells, basket cells, or glia, in all samples. 274 genes with differential expression were identified in granule cells, 139 in basket cells, 31 in glia, and 42 in all samples regardless of cell type.

[0288] RNAseq results for 274 genes in granule cells whose expression changed significantly with age are shown in Table 44. The log2 fold change of sorted nuclei from granule cells from each human sample compared to the normalized expression of sorted nuclei from granule cells from all samples is shown in Table 44. Negative values ​​indicate decreased expression in the nucleus of interest compared to the normalized value for whole nuclei from all samples, and positive values ​​indicate increased expression compared to the normalized value for whole nuclei from all analyzed cell types. [Table 45] TIFF2025032204000481.tif241166TIFF2025032204000482.tif242166TIFF2025032204000483.tif24216 6TIFF2025032204000484.tif241166TIFF2025032204000485.tif241166TIFF2025032204000486.tif24216 6TIFF2025032204000487.tif242166TIFF2025032204000488.tif241166TIFF2025032204000489.tif24216 6TIFF2025032204000490.tif240166TIFF2025032204000491.tif242166TIFF2025032204000492.tif11166

[0289] RNAseq results for differentially expressed genes (regulated at p<0.01, baseMean>50) in granule cells across all samples are shown in Table 45. [Table 46] TIFF2025032204000494.tif243166TIFF2025032204000495.tif244166TIFF2025032204000496.tif38166

[0290] RNAseq results for 139 genes in basket cells whose expression changed significantly with age are shown in Table 46. The log2 fold change of sorted nuclei from basket cells from each human sample compared to the normalized expression of sorted nuclei from granule cells in all samples is shown in Table 46. Negative values ​​indicate decreased expression in the nuclei of interest compared to the normalized value of whole nuclei from all samples, and positive values ​​indicate increased expression compared to the normalized value of whole nuclei from all analyzed cell types. [Table 47] TIFF2025032204000498.tif245165TIFF2025032204000499.tif245165TIFF2025032204000500.tif166166

[0291] RNAseq results of differentially expressed genes (regulated at p<0.01, baseMean>50) in basket cells in all samples are shown in Table 47. [Table 48] TIFF2025032204000502.tif241166TIFF2025032204000503.tif94166

[0292] RNAseq results for 31 genes in glia whose expression changed significantly with age are shown in Table 48. The log2 fold change of sorted nuclei from glia (astrocytes and oligodendrocytes) from each human sample compared to normalized expression data from sorted glia from all samples is shown in Table 48. Negative values ​​indicate decreased expression in the nucleus of interest compared to the normalized value for whole nuclei from all samples, and positive values ​​indicate increased expression compared to the normalized value for whole nuclei from all analyzed cell types. [Table 49] TIFF2025032204000505.tif238166

[0293] RNAseq results for differentially expressed genes (regulated at p<0.01, baseMean>50) in glial cells in all samples are shown in Table 49. [Table 50]

[0294] The results showed less change in expression with age compared to genes with differential expression with other clinical attributes analyzed herein. Previous investigations of cerebellar aging found fewer genes showing differential expression, suggesting that gene expression profiling in the nuclei of specific cell types provides much greater sensitivity compared to using bulk tissue.

[0295] Example 21. Overlapping changes in expression differences with age As shown in Table 50, the differentially expressed genes (adjusted p<0.01, baseMean>50) are approximately evenly divided between those whose expression increases with age and those whose expression decreases. Table 50 shows the number of genes that are common among the eight groups of differentially expressed aging genes. [Table 51] JPEG2025032204000508.jpg26165

[0296] No overlap was observed between age-upregulated genes from any cell type and downregulated genes from any other cell type.

[0297] Expression of only a few genes changes significantly with age in more than one cell type. For example, the 42 genes identified as differentially expressed with age in all cell types appear heterogeneous between conditions, and most of these do not reach significance when tested in each cell type separately (Table 51). [Table 52] TIFF2025032204000510.tif244166TIFF2025032204000511.tif32166

[0298] RNAseq results for 42 genes in granule cells with differential expression across the three cell types with age are shown in Table 52. The log2 fold change in expression of sorted nuclei from glia from each human sample compared to the normalized expression of sorted nuclei from granule cells in all samples is shown in Table 52. Negative values ​​indicate reduced expression in the nucleus of interest compared to the normalized value for whole nuclei from all samples, and positive values ​​indicate increased expression compared to the normalized value for whole nuclei from all analyzed cell types. [Table 53] TIFF2025032204000513.tif78166

[0299] The RNAseq results of the 42 most age-differentially expressed genes in basket cells are shown in Table 53. The log2 fold change between the RNAseq results of sorted nuclei from basket cells from each human sample compared to the normalized expression of unsorted nuclei from all samples is shown in Table 53. Negative values ​​indicate reduced expression in the nucleus of interest compared to the normalized value for whole nuclei from all samples, and positive values ​​indicate increased expression compared to the normalized value for whole nuclei from all analyzed cell types. [Table 54] TIFF2025032204000515.tif239166TIFF2025032204000516.tif241166TIFF2025032204000517.tif241166 TIFF2025032204000518.tif240166TIFF2025032204000519.tif240166TIFF2025032204000520.tif101166

[0300] RNAseq results for 42 genes in glial cells that were significantly altered in expression with age in the three cell types are shown in Table 54. The log2 fold change between the RNAseq results of sorted nuclei from glia from each human sample compared to the normalized expression of unsorted nuclei from all samples is shown in Table 54. Negative values ​​indicate decreased expression in the nucleus of interest compared to the normalized value for whole nuclei from all samples, and positive values ​​indicate increased expression compared to the normalized value for whole nuclei from all analyzed cell types. [Table 55] TIFF2025032204000522.tif244166TIFF2025032204000523.tif147166

[0301] These data suggest that, at least in the cerebellum, certain molecular consequences of aging differ depending on the cell type. Regardless of the differences in the genes affected by aging in each cell type, it remains possible that the cellular processes altered by these changes may be related across cell types.

[0302] In all three cell types, genes that were downregulated with age were enriched for synaptic genes, regardless of the inclusion of different specific genes. For example, aquaporin 7 (Aqp7) (preg=3.78E-9) and acid-sensing ion channel subunit 1 (Asic1) (preg=4.07E-7) were observed to be downregulated with age in granule cells, as shown in Figure 11A and Figure 11B, respectively. Figure 11C showed that roundabout guidance receptor 2 (Robo2) was downregulated in both glial and basket cells. The finding that genes encoding synaptic components decline in expression with age is consistent with previous findings of axonal and dendritic atrophy in the brains of older individuals (Burke, et al. (2006) Nat Rev Neurosci 7, pages 30-40 (2006); Dickstein, et al. (2013) Neuroscience 251, 21-32; Freeman, et al. (2008) J. Neuropathol. Exp. Neurol. 67, 1205-1212; each of which is incorporated herein by reference in its entirety). However, the data herein extend these observations to establish that this general process affects distinct genes in each cell type.

[0303] To test this hypothesis, GO analysis was performed in each cell type for genes that were up- or down-regulated with age. No significant GO categories were observed for any group of genes that were up-regulated with age, and no significant GO categories were identified for genes that were down-regulated with age from the "all cell types" group.

[0304] Example 22. Determining the effects of stress on human cell type-specific gene expression Transcriptional induction of immediate early genes, such as the FOS proto-oncogene, AP-1 transcription factor subunit (Fos, also known as c-FOS) and JunB proto-oncogene, AP-1 transcription factor subunit (Junb), is associated with neuronal activation and can indicate normal behavioral responses or pathological effects (Okuno, H. (2011) Neuroscience Research 69, 175-186 and Perez-Cadahia, et al. (2011) Biochem. Cell Biol. 89, 61-73, each of which is incorporated herein by reference in its entirety). The immediate early response genes associated with these events are often common, and the target genes and pathways induced by these factors reflect the nature of the event experienced by the cell or circuit (Duclot, et al. (2017) Front. Behav. Neurosci. 11, 717, each of which is incorporated herein by reference in its entirety). The stress response was observed to induce a group of genes in 3 of 16 human samples, which was not explained by race, autolysis time, sex, or age. Table 55 shows the differentially expressed genes (p<0.01, adjusted to baseMean>50) in all samples compared to normalized expression in all cell types. [Table 56] TIFF2025032204000525.tif244166TIFF2025032204000526.tif38166

[0305] c-FOS was observed to be strongly induced in granule cells from three of the 16 individuals (WI, VM, and ZH) (FIG. 12). To further explore this difference, differential expression (adjusted p<0.01, baseMean>50) analysis was performed to identify additional genes that were significantly up- or down-expressed in the three individuals compared to the other individuals, i.e., differential gene expression analysis of granule cells between the three fos+ individuals vs. the other 13 samples was performed. As shown in Table 56, a very strong response was evident in granule cells from individuals WI, VM, and TM, characterized by induction of 224 genes and reduced expression of 10 genes. [Table 57] TIFF2025032204000528.tif242166TIFF2025032204000529.tif227166

[0306] There was an overlapping but weaker response in glial cells that included 136 differentially induced genes (p<0.01, regulated at base Mean>50). The results are shown in Table 57. [Table 58] TIFF2025032204000531.tif146166

[0307] Induction of these genes in glia cannot be attributed to contamination by granule cells, because highly expressed markers of granule cells (e.g., FAT2) were not present in the glial data. This is interesting because glutamate-induced c-FOS expression was thought to occur through a different mechanism than induction by increased activity of cultured granule neurons (Edling, et al. (2007) Glia 55, 328-340, which is incorporated herein by reference in its entirety). Only two differentially expressed genes were evident in the basket cell data: heat shock protein family A (Hsp70) member 1A (HSPA1A) (base Mean: 102.303, log2 fold change: -3.234, p-value: 1.07E-12, p-regulation: 1.06E-08) and ring finger protein 122 (RNF122) (base Mean: 96.812, log2 fold change: -1.552, p-value: 4.20E-06, p-regulation: 0.008).

[0308] The stress response was observed to induce a group of genes in 3 of 16 samples, which was not explained by race, autolysis time, sex, or age. Gene expression is shown in Figure 13A for activity-regulated cytoskeleton-associated protein (Arc), early growth response factor 1 (Egr1), c-Fos, FosB, Jun proto-oncogene, AP-1 transcription factor subunit (Jun), JunB, and JunD proto-oncogene, AP-1 transcription factor subunit (JunD). Induction of fos genes and stress-related genes was observed to be slightly cell type specific, occurring primarily in granule cells, but also in glia. Little change in expression was observed in astrocytes.

[0309] Approximately 355 genes were observed to be induced in the three samples of granule cells, while 25 genes were observed to be downregulated. Many of these genes were intermediate early genes that mark neuronal activity, but there were also many long-term stress response genes that were induced, as shown in Figure 13B. Examples of affected gene groups include those involved in cellular metabolic processes, positive regulation of macromolecules, the unfolded protein response, response to morphologically incorrect proteins, protein folding, protein refolding, regulation of cellular response to heat, regulation of apoptotic processes, positive regulation of biological processes, and response to chemical stimuli.

[0310] Stress response genes in glia, such as the activated astrocyte marker GFAP, S100B, and Musashi RNA-binding protein 1 (Msi1), are shown in Figure 13 C. The only significant change in expression observed was in GFAP in the fos nucleus.

[0311] Samples WI, VM, and ZH were observed to have fos and stress response genes induced in glia and granule neurons, and to a lesser extent in stellate neurons. The three individuals observed to have induced expression of the fos gene did not appear to have much in common in their clinical histories. The data show reproducible activation of this stress response pathway, providing details on the biology associated with the disease.

[0312] Given the robust response evident in the granule cell data, a GO analysis was performed to determine whether the 224 induced genes in these neurons were enriched for specific functional categories. Significantly enriched GO categories were identified for upregulated genes in fos+ individuals, showing enrichment for genes involved in stress response and protein folding. As shown in Table 58, this analysis revealed categories related to response to external stimuli, protein folding / refolding, apoptosis, transcriptional response to stress, ATPase activity, and RNA splicing / metabolism. [Table 59] TIFF2025032204000533.tif243166TIFF2025032204000534.tif245165TIFF2025032204000535.tif18166

[0313] Analysis of immediate-early gene expression showed that their induction was also selective: for example, induction of EGR1, FOS, and FOSB occurred, but no differential expression of ARC, a well-characterized immediate-early gene, was evident (Fig. S12). Furthermore, while expression of immediate early genes such as c-FOS can be induced in response to injury such as ischemic stroke, no evidence was found for induction of glial genes such as GFAP and S100B, which mark astrocyte activation in response to stroke and other brain injuries (Choudhury, et al. (2016) Neurobiology of Disease 85, 234-244; Ding, S. (2014) Neural Regen Res 9, 2048-2052; Dirnagl, et al. (1999) Trends in Neurosciences 22, 391-397; Kajihara, et al. (2001) Brain Research 909, 92-101; Pekny, M., et al. (2005) Glia 50, 427-434; each of which is incorporated herein by reference in its entirety).

[0314] Finally, review of clinical records failed to reveal features (sex, age, cause of death) that distinguish these three individuals in Figure 12 from the 13 other control donors. Although the cause or causes of these molecular events remain unclear, the discovery of this robust and highly reproducible stress response in selected postmortem human brain samples suggests the existence of a subclinical condition, and analysis of specific cell types using the cell profiling techniques described herein can be used to discover pathways and biomarkers associated with pathophysiological conditions in humans.

[0315] Example 23. Staining and sorting of nuclei with RNA or DNA probes Antibodies often have non-specific targets and may be rate limiting for extending the cell profiling techniques described herein to more cell types. An alternative to using antibodies for labeling is to use RNA probes against cell type specific nucleic acid transcripts. RNA probes are faster and easier to synthesize than antibodies. RNA probes are designed against two classes of transcripts. The first are chromosome associated transcripts (CAT) - transcripts that are retained on chromosomes during synthesis. The second class of nucleic acid transcripts are polyA+ transcripts that are abundant in the ER adjacent to the nucleus. To characterize the variability between individuals, candidate CAT and polyA+ transcripts were identified after generating gene expression profiles for unsorted nuclei from several regions of the human brain. Using in situ hybridization on tissue sections, genes that delineate specific cell types were identified by anatomy and morphology. Antibodies or RNA probes against these candidates were used to label and sort nuclei from human cell types to determine the amount of change in gene expression between populations, without the need for prior knowledge from mouse data.

[0316] CAT has a unique labeling pattern of two spots at two chromosome synthesis sites, which was observed to be highly specific with low background using in situ hybridization. Figure 14 shows the RNAseq results of CAT. A2m is a Bergmann glia-specific CAT, Itpr1 is a Purkinje cell-specific CAT, and Fat2 is a granule cell-specific CAT. Thus, cell type-specific CATs can be detected from the RNAseq profile from all cerebellar nuclei. RNA Probes Based on the in situ hybridization results of labeled CAT and polyA+ transcripts with extensive knowledge of nuclear labeling for RNA probes and FACS, RNA probes for any cell type can be designed.

[0317] Equivalence and Scope Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific inventive embodiments described herein. The scope of the present invention is not intended to be limited to the above description, but rather is as set forth in the appended claims.

[0318] In the claims, articles such as "a," "an," and "the" may mean one or more, unless specifically indicated to the contrary or otherwise clear from the context. A claim or description including "or" between one or more group members is deemed to be satisfied if one, more than one, or all of the group members are present in, employed in, or otherwise relevant to a given product or process, unless specifically indicated to the contrary or otherwise clear from the context. The invention includes embodiments in which a member of the group is present in, employed in, or otherwise relevant to a given product or process. The invention includes embodiments in which one or more or all of the group members are present in, employed in, or otherwise relevant to a given product or process.

[0319] It should also be noted that the term "comprising" is intended to be an open term, allowing, but not requiring, the inclusion of additional elements or steps. When the term "comprising" is used herein, the term "consisting of" is also encompassed and disclosed.

[0320] Where ranges are given, the endpoints are included. Additionally, unless otherwise indicated or apparent from the context and the understanding of one of ordinary skill in the art, it is to be understood that values ​​expressed in ranges can take any particular value or subrange within the ranges set forth in alternative embodiments of the invention, to the tenth of the unit of the lower limit of that range, unless the context clearly dictates otherwise.

[0321] Furthermore, it should be understood that any particular embodiment of the present invention that falls within the scope of the prior art may be expressly excluded from any one or more of the claims. Such embodiments may be excluded even if the exclusion is not expressly stated herein, since they are deemed to be known to those skilled in the art. Any particular embodiment of the composition of the present invention (e.g., any antibiotic, therapeutic agent or active ingredient; any method of manufacture; any method of use; etc.) may be excluded from any one or more of the claims for any reason, whether related to the existence of prior art or not.

[0322] The words used are words of description rather than of limitation, and changes may be made within the purview of the appended claims without departing from the true scope and spirit of the invention in its broader aspects.

[0323] While the present invention has been described in considerable detail and with respect to certain illustrated embodiments, it is not intended that the invention be limited to any such details or embodiments, but rather should be construed with reference to the appended claims in order to give the claims the broadest possible interpretation in view of the prior art, and therefore in order to effectively encompass the intended scope of the invention.

Claims

1. 1. A method for profiling nuclear epigenetic expression and / or nuclear structure, comprising: (a) treating a tissue sample comprising a plurality of cell types to expose nuclei, said tissue sample comprising brain tissue, brain stem tissue, and / or spinal cord tissue; (b) labeling at least one nucleus from said processed tissue sample by contacting said processed tissue sample with two or more affinity labels, each of which binds a different nucleic acid transcript or protein unique to a nucleus of at least one cell type in the tissue sample, wherein each of said nucleic acid transcripts or proteins is encoded by a gene selected from Tables 19-24, and said affinity labels are selected from the group consisting of an antibody, an RNA probe, and a DNA probe; (c) purifying at least one labeled nucleus using said two or more affinity labels; and (d) (i) collecting information regarding histone modifications, transcription factor binding, and DNA modifications of said labeled nuclei, thereby profiling the epigenetic expression of said nuclei; and / or (ii) collecting information about the chromosomal structure of the labeled nuclei, thereby profiling the nuclear structure. A method comprising:

2. The cell type is (i) originating from cerebellar tissue and selected from the group consisting of granule cells, Purkinje cells, glia, Bergmann glia, basket / astrocyte cells, astrocytes, brain stem motor neurons, oligodendrocytes, upper motor neurons, lower motor neurons, and deep cerebellar nuclei; or (ii) originating from basal ganglia tissue and selected from the group consisting of striatonigral medium spiny neurons (MSNs), striatopallidal MSNs, striatal cholinergic interneurons, subthalamic nucleus, dopaminergic neurons, and bed nucleus of the stria terminalis (BNST) neurons; (iii) originating from thalamic tissue and selected from the group consisting of thalamocortical neurons, thalamostriatal neurons, and thalamic reticular nucleus neurons; or (iv) originating from cortical tissue and selected from the group consisting of corticostriatal neurons, entorhinal cortex layer 2 / 3 neurons, fast firing cortical interneurons, and layer 2 / 3 pyramidal cells from prefrontal cortical tissue; or (v) pineal gland medial habenular tissue-derived cholinergic projection neurons; or (vi) The method of claim 1, wherein the cells originate from hippocampal tissue and are selected from the group consisting of cornu ammonis area 1 (CA1), cornu ammonis area 2 (CA2), cornu ammonis area 3 (CA3), and dentate gyrus cells.

3. (e) epigenetic expression and / or nuclear structure of the nucleus; (i) cell types from different tissue samples, or (ii) Different cell types The method of claim 1 or 2, further comprising comparing the epigenetic expression and / or nuclear structure in at least one nucleus of said subject to identify changes in epigenetic expression and / or nuclear structure to obtain a drug target.

4. 2. The method of claim 1, wherein the affinity label further comprises a fluorescent tag, and optionally the nuclei are purified using fluorescence activated cell sorting (FACS).

5. (i) the tissue sample or the different tissue sample is a frozen tissue sample; (ii) the tissue sample or the different tissue sample is a new tissue sample; (iii) the tissue sample or the different tissue sample is from a female; (iv) the tissue sample or the different tissue sample is from a male; (v) the tissue sample or the different tissue sample is from a diseased subject; Optionally, the diseased subject expresses or does not express a disease phenotype, or the diseased subject suffers from at least one disease selected from the group consisting of ataxia, Parkinson's disease, Alzheimer's disease, and amyotrophic lateral sclerosis (ALS), and Huntington's disease; Optionally, the cell type is: (a) associated with ataxia and selected from the group consisting of Purkinje cells, granule cells, Bergmann glia, basket / astrocytes, astrocytes, oligodendrocytes, and deep cerebellar nuclei; or (b) associated with Parkinson's disease and selected from the group consisting of substantia nigra and VTA dopaminergic neurons; (c) associated with Alzheimer's disease and selected from layer 2 / 3 entorhinal cortex, CA1 hippocampus, and CA2 / 3 hippocampus; or (d) associated with ALS and selected from the group consisting of brainstem, cortical motor neurons, and spinal motor neurons; (vi) the tissue sample or the different tissue sample is from a healthy subject, or (vii) the drug target is encoded by a gene selected from Tables 19-24; The method according to claim 3.

6. (i) the nucleus is adjacent to the endoplasmic reticulum; or (ii) the protein is a membrane protein, optionally wherein the membrane protein is synthesized in the endoplasmic reticulum, and optionally wherein the membrane protein is localized in the endoplasmic reticulum; The method according to claim 1 or 2.

7. The method of claim 1 or 2, wherein the nucleic acid transcript is localized in the nucleus.

8. (i) the RNA probe specifically binds to a chromosome-associated transcript (CAT) or a polyA transcript; or (ii) The method of claim 1, wherein the DNA probe specifically binds to a transporter gene.

9. (i) comparing the profiling of tissue samples from diseased subjects with tissue samples from healthy subjects; or 3. The method of claim 1 or 2, further comprising (ii) comparing the profiling of tissue samples from at least two different healthy subjects.

10. said epigenetic expression and / or nuclear structure profiling comprising: (i) performed on multiple nuclei from a single cell type; or (ii) performed on multiple nuclei from different cell types; The method according to claim 1 or 2.

11. 2. The method of claim 1, further comprising: (e) identifying changes between disease stages by comparing profiling of tissue samples from diseased subjects expressing the disease phenotype with profiling of tissue samples from diseased subjects not expressing the disease phenotype.

12. 4. The method of claim 3, further comprising, in (e)(i), comparing profiling of two or more tissue samples, each from a different subject, to identify alterations between different subjects.

13. Profiling the nuclear structure of nuclei comprises labeling at least one nucleus from the processed tissue sample by contacting the processed tissue sample with two or more affinity labels, wherein the affinity labels are antibodies, and (d)(ii) comprises: (1) isolating gDNA from the nuclei; (2) performing oxidative bisulfite sequencing; and 3. The method of claim 1, further comprising: (3) collecting information about chromosomal structure for the labeled nuclei, thereby profiling the nuclear structure.

Images