Use of circulating cell-free methylated DNA to detect tissue damage

Abstract

A method for determining whether a subject suffers from tissue damage resulting from exposure to a toxic agent, comprising the steps of: sequencing cell-free DNA (cfDNA) in a biological sample from a subject; determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; measuring the amount of the determined cfDNA of cellular origin, and comparing the measured amount of the determined cfDNA of cellular origin with a normal amount of the determined cfDNA of cellular origin. A higher amount of the measured cfDNA of the determined cfDNA of cellular origin indicates that the subject suffers from tissue damage.

Description

[Technical field]

[0001] (CROSS REFERENCE TO RELATED APPLICATIONS) This application claims the benefit of U.S. Provisional Application No. 63 / 224,873, filed July 23, 2022, and U.S. Provisional Application No. 63 / 324,112, filed March 27, 2022, each of which is incorporated by reference in its entirety.

[0002] (STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT) This invention was made with Government support under Grant Nos. T32 CA009686, F30 CA250307, ​​and R01 CA231291 awarded by the National Institutes of Health. The United States Government has certain rights in this invention. [Background technology]

[0003] 2. Background of the Invention The human body is frequently exposed to agents that can have damaging effects on tissues. Such agents can be, for example, pathogenic agents such as bacteria or viruses, environmental agents such as sunlight, or therapeutic agents such as medicines that have side effects.

[0004] Another type of therapy that can potentially result in tissue damage is the therapy used to treat cancer, such as surgery, chemotherapy, radiation therapy, targeted therapy, and immunotherapy. Each of these therapeutic interventions can have significant systemic effects. For example, radiation therapy uses ionizing radiation to target tumor cells (Haussmann et al., 2020; Xu et al., 2008), but normal tissues are also affected, resulting in tissue damage and remodeling (Ruysscher et al., 2019; Hubenak et al., 2014). For breast cancer patients, the heart and lungs are the most common organs affected by radiation toxicity, with a linear increase in cardiovascular disease risk of 7.4% per gray mean dose to the heart reported (Darby et al., 2013; White and Joiner, 2006). Furthermore, radiation-induced lung injury, manifested as radiation pneumonitis or fibrosis, is a severe complication reported in 5–20% of cases ( Giuramno et al., 2019 ; Arroyo-Hernaendez et al., 2021 ).

[0005] The ability to distinguish different cell types involved and potentially contributing to toxicity using cell-free DNA (cfDNA) in serial blood samples can significantly impact therapeutic decision making. Imaging methods can be used as an indirect method to measure treatment efficacy, but these results are often unreliable and difficult to interpret. Imaging results can be clouded by the expression of pseudoprogression, making them an ineffective or crude means of monitoring the simultaneous changes necessary to guide treatment decisions. In view of the risk of tissue damage from radiation therapy or exposure to other toxic factors, a means is essential to effectively assess tissue damage and monitor the impact of treatment. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Haussmann et al., 2020 [Non-Patent Document 2] Xu et al., 2008 [Non-Patent Document 3] Ruysscher et al., 2019 [Non-Patent Document 4] Hubenak et al., 2014 [Non-Patent Document 5] Darby et al., 2013 [Non-Patent Document 6] White and Joiner, 2006 [Non-Patent Document 7] Giuramno et al., 2019 [Non-Patent Document 8] Arroyo-Hernaendez et al., 2021 Summary of the Invention [Means for solving the problem]

[0007] Summary of the Invention Some of the main aspects of the present invention are summarized below. Further aspects are described in the Detailed Description, Examples, Figures and Claims sections of this disclosure. The statements in each section of this disclosure are intended to be read in conjunction with the other sections. Furthermore, the various embodiments described in each section of this disclosure can be combined in various different ways, and all such combinations are intended to fall within the scope of the present invention.

[0008] The present invention provides a novel method for detecting tissue damage resulting from exposure to a toxic agent.

[0009] In one aspect, the present invention relates to a method for determining whether a subject suffers from tissue damage resulting from exposure to a toxic agent. In some embodiments, the method includes the steps of: (a) sequencing cfDNA in a biological sample from a subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, and the cellular origin of the cfDNA is determined when the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; (c) measuring the amount of the determined cfDNA of cellular origin; and (d) comparing the measured amount of the determined cfDNA of cellular origin with the normal amount of the determined cfDNA of cellular origin. An increase in the measured amount of the determined cfDNA of cellular origin over the normal amount of the determined cfDNA of cellular origin indicates that the subject suffers from or will suffer from tissue damage resulting from the exposure. In other embodiments, the method includes the steps of: (a) sequencing cfDNA in a biological sample from the subject; determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain a methylation site, where the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; and (c) measuring an amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point. An increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point indicates that the subject is or will be affected by tissue damage resulting from the exposure.

[0010] In another aspect, the present invention also relates to a method for treating a subject suffering from tissue damage resulting from exposure to toxic agent.In some embodiments, the method comprises administering a treatment for tissue damage to a subject, wherein the subject is determined to suffer from tissue damage by a method comprising: (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more parts of the sequence of the cfDNA that includes methylation sites, and when the methylation pattern in the one or more parts is the same as a known cell type specific methylation pattern, the cellular origin of the cfDNA is determined; (c) measuring the amount of the cfDNA of the determined cellular origin; and (d) comparing the measured amount of the cfDNA of the determined cellular origin with the normal amount of the cfDNA of the determined cellular origin.An increase in the measured amount of the cfDNA of the determined cellular origin over the normal amount of the cfDNA of the determined cellular origin indicates that the subject suffers from tissue damage. In another embodiment, the method includes administering a treatment for tissue injury to a subject, wherein the subject is determined to be suffering from tissue injury by a method comprising, at two or more time points, (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain a methylation site, wherein the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; and (c) measuring the amount of cfDNA of the determined cellular origin. An increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point indicates that the subject is suffering from tissue injury.

[0011] In yet another aspect, the present invention further relates to a method for treating tissue damage in a subject. In some embodiments, the method includes administering a treatment for tissue damage to the subject and monitoring the tissue damage, wherein the monitoring includes (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cfDNA is determined when the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; (c) measuring the amount of the determined cellular origin cfDNA; and (d) comparing the measured amount of the determined cellular origin cfDNA with a normal amount of the determined cellular origin cfDNA. A decrease in the measured amount of the determined cellular origin cfDNA compared to the normal amount of the determined cellular origin cfDNA indicates that the treatment is effective, and an increase or no change in the measured amount of the determined cellular origin cfDNA above the normal amount of the determined cellular origin cfDNA indicates that the treatment is not effective. In other embodiments, the method includes administering to a subject a treatment for tissue damage and monitoring the tissue damage, the monitoring comprising: (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; and (c) measuring an amount of cfDNA of the determined cellular origin. A decrease in the measured amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point indicates that the treatment is efficacious, and an increase or no change in the measured amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point indicates that the treatment is not efficacious.

[0012] In some embodiments, the tissue damage is caused by exposure to a toxic agent, hi certain embodiments, the toxic agent comprises radiation.

[0013] The radiation may be radiation for therapeutic purposes, accidental radiation, or environmental radiation. In some embodiments, the radiation comprises a radioactive material. The radioactive material may be ingested by the subject, inhaled by the subject, or absorbed through body surface contamination by the subject.

[0014] In other embodiments, the virulence factor comprises a microorganism. The microorganism may comprise a pathogen such as a bacterium or a virus.

[0015] In some embodiments, the toxic agent is from a chemical source or from a biological source.

[0016] In some embodiments, the toxic factors include drug therapy.

[0017] In some embodiments, the toxic agent comprises a chemical or biological or radiological agent used as a weapon.

[0018] In a further aspect, the present invention relates to a method for treating a subject in need of treatment. In some embodiments, the method comprises administering a treatment to the subject and monitoring whether the treatment causes tissue damage in the subject, the monitoring comprises: (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, and when the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern, the cellular origin of the cfDNA is determined; (c) measuring the amount of the determined cfDNA of cellular origin; and (d) comparing the measured amount of the determined cfDNA of cellular origin with a normal amount of the determined cfDNA of cellular origin. An increase in the measured amount of the determined cfDNA of cellular origin over the normal amount of the determined cfDNA of cellular origin indicates that the treatment causes tissue damage. In other embodiments, the method includes administering a treatment to the subject and monitoring whether the treatment causes tissue damage in the subject, wherein the monitoring includes, at two or more time points, (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell-type-specific methylation pattern; and (c) measuring an amount of cfDNA of the determined cellular origin, wherein an increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to a previous time point indicates that the treatment is causing tissue damage.

[0019] In some embodiments, the methods further comprise adjusting the treatment administered to the subject if the treatment is shown to be ineffective or causing tissue damage.

[0020] In some embodiments, the normal amount of cfDNA includes the amount of cfDNA for the above-determined cell origin generated in a population of individuals that have not been exposed to a toxic agent or administered a treatment.

[0021] In yet another aspect, the present invention relates to a method of treating a subject having a tumor. In some embodiments, the method includes the steps of: (A) monitoring a response to a first treatment, an adverse reaction to a first treatment, or a combination thereof, wherein the monitoring comprises: (i) determining whether there is an adverse reaction to a first treatment, comprising: (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain a methylation site, wherein the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; (c) measuring an amount of the determined cellular origin cfDNA, and (d) comparing the measured amount of the determined cellular origin cfDNA with a normal amount of the determined cellular origin cfDNA, wherein an increase in the measured amount of the determined cellular origin cfDNA over the normal amount of the determined cellular origin cfDNA indicates an adverse reaction; (ii) determining whether there is a response to a first treatment, comprising: (a) determining whether there is an adverse reaction to a first treatment, comprising: (a) determining whether there is an adverse reaction to a first treatment, and (b) determining whether there is an adverse reaction to a first treatment, wherein the determining comprises ... and (c) determining whether there is an adverse reaction to a first treatment, and (d) determining whether there is an adverse reaction to a first treatment, and The method includes the steps of: (A) sequencing circulating tumor DNA (ctDNA) in a biological sample from a subject; determining clonal heterogeneity of cells of the tumor by genotyping the ctDNA, wherein the presence of two or more clones of the tumor cells or the presence of a tumor cell clone not previously identified in the subject indicates an ineffective response to the first treatment; and (B) administering the same treatment as the first treatment if it is determined that there is no adverse reaction, there is no ineffective response, or a combination thereof; or administering an adjusted treatment if it is determined that there is an adverse reaction, there is an ineffective response, or a combination thereof.In other embodiments, the method includes the steps of: (A) monitoring a response to a first treatment, an adverse reaction to a first treatment, or a combination thereof, wherein the monitoring comprises, at two or more time points, (i) determining whether there is an adverse reaction to a first treatment, comprising: (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of a sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) measuring an amount of cfDNA of the determined cellular origin, wherein an increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point is indicative of an adverse reaction; and (ii) determining whether there is a response to a first treatment, comprising: (a) sequencing circulating tumor (ctDNA) ... contains methylation sites, wherein the cellular origin of the cfDNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; The method includes the steps of: (A) determining the clonal heterogeneity of cells of the tumor by genotyping the ctDNA, wherein the presence of two or more clones of the tumor cells or the presence of a tumor cell clone at a later time point not identified at an earlier time point indicates an ineffective response to the first treatment; and (B) administering the same treatment as the first treatment if it is determined that there is no adverse reaction, there is no ineffective response, or a combination thereof; or administering an adjusted treatment if it is determined that there is an adverse reaction, there is an ineffective response, or a combination thereof.

[0022] In some embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin generated in a population of individuals without tumors.In other embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin generated in a population of individuals who have not received the first treatment.

[0023] In some embodiments, the biological sample comprises a bodily fluid. In certain embodiments, the bodily fluid is selected from blood, serum, plasma, cerebrospinal fluid, saliva, urine, and sputum. In preferred embodiments, the bodily fluid comprises blood, serum, or plasma.

[0024] In some embodiments, the methylation pattern comprises a segment of a nucleotide sequence that includes at least three CpG dinucleotides.

[0025] In some embodiments, the known methylation patterns are shown in Table 2. [Brief description of the drawings]

[0026] [Figure 1] Figure 1 shows an example of the use of predictive treatment response and therapy-related toxicity obtained from combined genetic and epigenetic analysis of cfDNA. Predictive treatment response and therapy-related toxicity obtained from combined genetic and epigenetic analysis of cfDNA. The minimally invasive nature of liquid biopsies allows for serial sampling to monitor changes over time, especially under selective pressure from ongoing therapy. Circulating tumor DNA (ctDNA) can be used to track clonal heterogeneity over time, assess treatment response and detect treatment-resistant clones. Normal cell-specific cfDNA methylation patterns can be used in combination with ctDNA to assess the impact of treatment on the surrounding tumor microenvironment and monitor therapy-related toxicity in somatic cell types. [ctDNA = circulating tumor DNA; cme-DNA = circulating methylated cell-free DNA]. [Diagram 2]FIG. 2 shows a global analysis of cell-free methylated DNA in blood to identify the origin of radiation-induced cell damage, as described in the Examples. Serial serum samples were collected from human breast cancer patients treated with radiation. Concurrently, paired serum and tissue samples were collected from mice receiving radiation at 3 Gy or 8 Gy doses compared to sham controls. Methylome profiling of liquid biopsy samples was performed using a bisulfite-based capture sequencing method optimized for cfDNA input. Differential cell type-specific methylation blocks were identified from reference WGBS data collected from healthy cell types and tissues in humans and mice. A methylation map was created highlighting the cell types that constitute the target risk organs from radiation exposure, such as lung, heart, and liver. Deconvolution analysis of cfDNA using fragment-level CpG methylation patterns in these identified cell type-specific blocks was used to decipher the origin of radiation-induced cell damage. [Figure 3A]FIG. 3 shows the sensitivity and specificity of the identified mouse cell type-specific differential methylation blocks, as described in the Examples. In panels A-D, the top images are heat maps of all cell type-specific methylation blocks selected for each target cell type. All blocks contain 3+ CpG sites and have a beta difference margin of 0.4 or greater that distinguishes the target cell type from all other cell types included in the reference map. All methylation blocks identified for the following mouse cell types were hypomethylated: pulmonary endothelial (n=1546), hepatocytes (n=616), and cardiomyocytes (n=2,917). In contrast, all identified immune cell-specific blocks (n=148) were hypermethylated compared to other solid organ cell types in mice. In panels A-D, the right images show in-silico mix-in validation of the fragment-level probabilistic deconvolution model. Target cell type read pairs were in-silico mixed in a background of lymphocyte or buffy coat read pairs at various known percentages (0, 0.5, 1, 2, 5, 10, 15%) with 10 replicates per percentage. The deconvolution models were validated on these in-silico mixed samples of known cell type percentages in selected blocks. The average predicted % target was plotted against the known % mixture to evaluate the sensitivity and specificity of the identified cell type specific blocks and the deconvolution models. Data are presented as mean ± SD; n=3 replicates per percentage. Reference WGBS samples with less than 3 replicates were split into a "0.8 train" for selecting methylation blocks and a "0.2 test" for creating in-silico mixed samples. When available, in-silico mixed samples of the same cell type from mice of different ages were tested. Additionally, bulk tissues of each cell type were also tested. [Figure 3B] Continued from Figure 3. [Figure 3C] Continued from Figure 3. [Figure 3D] Continued from Figure 3. [Figure 4A]FIG. 4 shows the sensitivity and specificity of the identified human cell type-specific differential methylation blocks, as described in the Examples. In panels A-F, the top images are heat maps of all cell type-specific methylation blocks selected for each target cell type. All blocks contain 3+ CpG sites and have a beta difference margin of 0.4 or greater that distinguishes the target cell type from all other cell types included in the reference map. In panels A-F, the bottom images show in-silico mix-in validation of the fragment-level probabilistic deconvolution model. Target cell type read pairs were in-silico mixed in a background of lymphocyte or buffy coat read pairs at various known percentages (0, 0.5, 1, 2, 5, 10, 15%). The deconvolution model was validated on samples of known cell type proportions in these in-silico mixed selected blocks. The average predicted % target is plotted against the known % mix to assess the sensitivity and specificity of the identified cell type-specific blocks and the deconvolution model. Data are expressed as mean ± standard deviation; n = 3 replicates per ratio. [Figure 4B] Continued from Figure 4. [Figure 4C] Continued from Figure 4. [Figure 4D] Continued from Figure 4. [Figure 4E] Continued from Figure 4. [Figure 4F] Continued from Figure 4. [Figure 5A] Figure 5 shows the characterization of human and mouse cell type specific reference methylation data as described in the Examples. Panel A shows a tree dendrogram depicting the relationships between the human reference whole genome bisulfite sequencing (WGBS) datasets included in the analysis. Methylation states in the top 30,000 variable blocks were used as input data for unsupervised hierarchical clustering. Samples from >n=3 replicate cell types were pooled. [Figure 5B] Panel B shows UMAP predictions of the human WGBS reference dataset colored by tissue and cell type. [Figure 5C]Panel C shows the UMAP predictions of the mouse WGBS reference dataset. [Acronyms: HUVEV = human umbilical vein endothelial cells, PAEC = pulmonary artery endothelial cells, CAEC = coronary artery endothelial cells, PMEC = pulmonary microvascular endothelial cells, CMEC = cardiac microvascular endothelial cells, CPEC = combined cardiopulmonary endothelial cells, LSEC = liver sinusoidal endothelial cells, NK = natural killer cells, MK = megakaryocytes.] [Figure 6A] Figure 6 shows characterization of mouse cell type-specific reference methylation data, as described in the Examples. Panel A shows a dendrogram depicting the relationships between the mouse reference WGBS datasets included in the analysis. The methylation states in the top 30,000 variable blocks were used as input data for unsupervised hierarchical clustering. [Figure 6B] Panel B shows a heat map of differentially methylated cell type-specific blocks identified from reference WGBS data collected from healthy cell types and tissues in mice. Each cell in the plot shows the average methylation of one genomic region (row) in each of nine mouse tissues and cell types (columns). The 100 highest methylation score blocks per cell type are shown. Differential blocks identified from cell types that comprise the target risk organs of radiation (lung, heart, and liver) were selected for the creation of a radiation-specific methylation map that distinguishes these solid organ cell types from all other immune cell types. [Figure 7A-B]Figure 7 shows the identification and biological validation of cell type specific DNA methylation blocks in humans and mice as described in the Examples. Panels A and B show heat maps of differentially methylated cell type specific blocks identified from reference WGBS data collected from healthy cell types and tissues in humans (panel A) and mice (panel B). Each cell in the plot shows the methylation score of one genomic region (row) in each of 20 cell types in humans and 9 cell types (columns) in mice. The 100 or fewer highest methylation score blocks per cell type are shown. Methylation scores represent the number of fully unmethylated read pairs / total coverage or fully methylated read pairs / total coverage for hypomethylated and hypermethylated blocks, respectively. [Figure 7C] Panel C shows a heatmap of distance scores between gene set pathways identified from GeneSetCluster. Genes flanking human cell type specific methylation blocks were identified using HOMER and pathway analysis was performed using both Ingenuity Pathway Analysis (IPA) and GREAT. Gene set pathways significantly enriched (P<0.05) from differential methylation blocks identified in immune, cardiomyocyte, hepatocyte, and lung epithelial cell types were analyzed using GeneSetCluster. Cluster analysis was performed to determine the distance between all identified gene set pathways based on the degree of overlapping genes from each individual gene set compared to all other gene sets. Overrepresentation analysis was performed in WebgestaltR (ORAperGeneSet) plugin to interpret and functionally label the identified gene set clusters. [Acronyms: HUVEV = human umbilical vein endothelial cells, CPEC = cardiopulmonary endothelial cells, LSEC = liver sinusoidal endothelial cells, NK = natural killer cells.] [Figure 8]Figure 8 shows the biological functions of mouse cell type specific methylation blocks, as described in the Examples. Heatmap of distance scores between gene set pathways identified from GeneSetCluster. Genes adjacent to cell type specific methylation blocks were identified using HOMER, and pathway analysis was performed using both Ingenuity Pathway Analysis (IPA) and GREAT. Gene set pathways significantly enriched (P<0.05) from differential methylation blocks identified in immune, cardiomyocyte, hepatocyte, and pulmonary endothelial cell types were analyzed using GeneSetCluster. Cluster analysis was performed to determine the distance between all identified gene set pathways based on the degree of overlapping genes from each individual gene set compared to all other gene sets. Overrepresentation analysis was performed in WebgestaltR (ORAperGeneSet) plugin to interpret and functionally label the identified gene set clusters. [Figure 9A] Figure 9 shows that cell type-specific DNA methylation is predominantly hypomethylated and enriched in intragenic regions and developmental transcription factor (TF) binding motifs, as described in the Examples. Panel A shows a schematic diagram depicting the location of human cell type-specific hypo- and hypermethylated blocks. Genomic annotation of cell type-specific methylation blocks was determined by analysis using HOMER. [Figure 9B-C] Panels B and C show the distribution of human (panel B) and mouse (panel C) cell type specific methylation blocks compared to the genomic regions used in the hybridization capture probes. Captured blocks with less than 5% variance across cell types represent blocks with no cell type specificity and were used as background. [Figure 9D] Panel D shows the top 5 TF binding sites enriched among cell type-specific hypo- and hypermethylated blocks identified in human (top) and mouse (bottom) using HOMER motif analysis. Captured blocks with less than 5% variance among the same cell types as above were used as background. [Figure 10A] Figure 10 shows that methylation profiling of human endothelial cell types reveals tissue-specific differences that coincide with changes in RNA expression levels and biological function, as described in the Examples. Panel A shows pathways that support the biological significance of endothelial-specific methylation blockade (all P<0.05). [Figure 10B] Panel B shows the important functions of genes flanking endothelial-specific methylation blocks. Genes marked with an asterisk have flanking hypermethylated regulatory blocks. Genes without an asterisk have flanking hypomethylated regulatory blocks. [Figure 10C] Panel C shows gene expression in genes adjacent to tissue-specific endothelial-specific methylation blocks. Expression data was generated from RNA sequencing of the same paired cardiopulmonary endothelial cells (CPEC) and liver sinusoidal endothelial cells (LSEC) used to generate the methylation reference data. Pan-endothelial genes upregulated in both populations (ALL) are identified as endothelial-specific methylation blocks common to both LSEC and CPEC populations. [Figure 10D] Panel D shows the top 5 transcription factor binding sites enriched among the identified endothelial-specific hypomethylated blocks using HOMER de novo motif analysis and known motif analysis. The background for the HOMER analysis consisted of the other 3,574 identified cell type-specific hypomethylated blocks in all cell types except endothelial. [Figure 10E]Panel E shows an example of the NOS3 locus that is specifically unmethylated in endothelial cells. This endothelial-specific differentially methylated block (DMB) is 157 bp long (7 CpGs) and is located in the NOS3 gene, an endothelial-specific gene (GTEx inset as well as paired RNA sequencing data, upregulated in vascular endothelial cells). The alignment (top) from the UCSC genome browser provides the genomic locus organization and is aligned by average methylation across cardiomyocyte samples, lung epithelial samples, liver sinusoidal endothelial (LSEC) samples, cardiopulmonary endothelial (CPEC) samples, hepatocyte samples, and immune (PBMC) samples (n=3 / cell type group). Results from RNA sequencing generated from paired cell types as well as peak intensities from H3K27ac and H3K4me3 public ChIP-seq data generated in endothelial cells are represented [acronyms: HUVEV = human umbilical vein endothelial cells, CPEC = cardiopulmonary endothelial cells, LSEC = liver sinusoidal endothelial cells.] [Figure 11A] Figure 11 shows the development of a radiation-specific methylation map focused on cell types from target organs at risk (OARs) as described in the Examples. Panel A shows a representative three-dimensional conformal radiation therapy (3D-CRT) treatment planned for a right-sided (i and ii) and left-sided (iii and iv) breast cancer patient, respectively. Coronal and sagittal images of computed tomography simulation show the anatomical location of the target volume relative to adjacent organs. The map shows different radiation dose levels or isodose lines (95% of the prescribed dose, 90% isodose line, 80% isodose line, 70% isodose line, 50% isodose line). [Figure 11B]Panel B shows a heat map of differentially methylated cell type-specific blocks identified from all reference WGBS data collected from healthy human cell types and tissues. Each cell in the plot shows the average methylation (row) of one genomic region in each of 20 human cell types (columns). Blocks with the 100 highest methylation scores per cell type are shown. Differential blocks identified from cell types that comprise the target risk organs of radiation (lung, heart, and liver) were selected for the creation of a radiation-specific methylation map that distinguishes these solid organ cell types from all other immune cell types. [Acronyms: HUVEV = human umbilical vein endothelial cells, PAEC = pulmonary artery endothelial cells, CAEC = coronary artery endothelial cells, PMEC = pulmonary microvascular endothelial cells, CMEC = cardiac microvascular endothelial cells, CPEC = combined cardiopulmonary endothelial cells, LSEC = liver sinusoidal endothelial cells, NK = natural killer cells, MK = megakaryocytes.] [Figure 12A] Figure 12 shows that dose-dependent radiation damage in mouse tissues correlates with the origin of circulating methylated cfDNA, as described in the Examples. Panel A shows representative hematoxylin and eosin (H&E) staining of lung, heart, and liver tissues from mice treated with 3Gy and 8Gy radiation compared to sham controls. Scale bar, 200 μm. [Figure 12B] Panel B shows quantitative polymerase chain reaction (qPCR) analysis of CDKN1A (p21) marker of apoptosis in mouse tissues treated with 3 Gy and 8 Gy radiation compared to sham controls. Expression in each sample was normalized to the expression of the housekeeping gene ACTB (actin) and shown relative to expression in sham controls. Data are presented as mean ± SD (n = 3). Kruskal-Wallis test was used for between-group comparison; lung tissue p = 0.004, heart tissue p = 0.025, liver tissue p = 0.004. [Figure 12C-E]Panels C-F show methylated cfDNA in the circulation of lung endothelium, cardiomyocytes and hepatocytes from mice treated with 3 Gy and 8 Gy radiation compared to sham controls, expressed as genome equivalents (Geq). cfDNA was extracted from 18 mice (n=6 in each group) and cfDNA from two mice was pooled for each methylome preparation. Mean ± SD; n=3 independent methylome preparations. Kruskal-Wallis test was used for between-group comparisons. ns (not significant), P≥0.05; *, P<0.05; lung endothelium p=0.01, cardiomyocytes p=0.01, hepatocytes p=0.13. [Figure 13A-B] Figure 13 shows radiation-induced apoptotic damage in mouse tissues, as described in the Examples. qPCR analysis of apoptotic markers (Trp53, Gadd45a, Aifm3, and Bad) in lung, heart, and liver tissues of mice treated with 3Gy and 8Gy radiation compared to sham controls. Expression of each sample was normalized to the expression of the housekeeping gene ACTB (actin). Data are presented as mean ± SD (n=3). [Figure 13C-D] Figure 13 shows radiation-induced apoptotic damage in mouse tissues, as described in the Examples. qPCR analysis of apoptotic markers (Trp53, Gadd45a, Aifm3, and Bad) in lung, heart, and liver tissues of mice treated with 3Gy and 8Gy radiation compared to sham controls. Expression of each sample was normalized to the expression of the housekeeping gene ACTB (actin). Data are presented as mean ± SD (n=3). [Figure 14A-B]FIG. 14 shows radiation-induced effects on immune and solid organ cfDNA, as described in the Examples. Panels A-C show radiation-induced effects in humans, and panels D and E show radiation-induced effects in mice. Panel A shows predicted human immune-derived cfDNA in Geq. Human Geq is calculated by multiplying the relative proportion of cell type-specific cfDNA x the initial concentration cfDNAng / mL x the weight of the haploid human genome. Immune cfDNA was assessed in n=222 methylation blocks that were found to distinguish immune cell types from solid organ cell types (g1=B cells, CD4 T cells, CD8 T cells, NK, MK, erythroblasts, monocytes, macrophages, neutrophils; g2=breast basal / luminal epithelium, lung epithelium, hepatocytes, kidney podocytes, pancreatic islets, colonic epithelium, cardiomyocytes, LSECs, CPECs, HUVECs, nerves, and skeletal muscle). Panel B shows predicted human solid organ derived cfDNA by Geq, where % solid organ is defined as 100-% immune using these same n=222 methylation blocks. Friedman test was performed for between-group comparisons for panels A and B. ns, P>0.05; *, P<0.05; immune p=0.07, solid organ p=0.008. [Fig. 14C-D] Panel C shows fold change in human immune vs solid organ Geq at EOT and recovery compared to baseline. Data are presented as mean ± SD; n=15. Panel D shows predicted mouse immune-derived cfDNA in Geq. Mouse Geq is calculated by multiplying the relative proportion of cell type specific cfDNA x initial concentration cfDNAng / mL x weight of haploid mouse genome. Immune cfDNA was assessed in n=148 methylation blocks found to distinguish immune cell types from solid organ cell types. (g1=B cells, CD4 T cells, CD8 T cells, neutrophils; g2=breast epithelium, cardiomyocytes, hepatocytes, lung endothelium, cerebellum, hypothalamus, colon, intestine, kidney). For panel D, mean ± SD; n=3 independent methylome preparations. Kruskal-Wallis test was used for between-group comparison. ns, P>0.05;*, P<0.05; immune p=0.20, solid organs p=0.01. [Figure 14E]Panel E shows predicted mouse solid organ-derived cfDNA by Geq. For Panel E, mean ± SD; n = 3 independent methylome preparations. Kruskal-Wallis test was used for group comparison. ns, P>0.05; *, P<0.05; immune p=0.20, solid organ p=0.01. [Figure 15A-B] Figure 15 shows radiation-induced hepatocyte cfDNA and liver endothelial cfDNA in patients with right-sided versus left-sided breast cancer, as described in the Examples. Panels A and B show hepatocyte cfDNA (Geq / mL) in serum samples taken at different times. Fragment-level deconvolution using hepatocyte-specific methylation blocks (n=200). Wilcoxon matched pairs signed rank test was used for between-group comparisons, and results were considered significant if *P<0.05; ns, P≧0.05; right-sided p=0.02, left-sided p=0.81. [Figure 15C-D] Panel C shows fold change in hepatocyte cfDNA at end of treatment (EOT) and recovery compared to baseline. Mean ± SD; n=8 right, n=7 left. Panels D and E show LSEC cfDNA (Geq / mL) in the same serum samples. Fragment-level deconvolution used LSEC-specific methylation blocks (n=89). Wilcoxon matched-pairs signed-rank test was performed between groups and results were considered significant if *P<0.05; ns, P≥0.05; right p=0.02, left p=0.93. [Fig. 15E-F] Panels D and E show LSEC cfDNA (Geq / mL) in the same serum samples. Fragment-level deconvolution used LSEC-specific methylation blocks (n=89). Wilcoxon matched-pairs signed-rank test was performed between groups and results were considered significant when *P<0.05; ns, P>0.05; right p=0.02, left p=0.93. Panel F shows fold change in LSEC cfDNA Geq at EOT and recovery compared to baseline levels. Mean ± SD; n=8 right, n=7 left. [Figure 16A-B] FIG. 16 shows that radiation-induced cardiopulmonary cfDNA in patients correlates with radiation dose and indicates persistent damage to cardiomyocytes, as described in the Examples. Panel A shows lung epithelial cfDNA (Geq / mL) in serum samples taken at different times. Fragment level deconvolution used the lung epithelial specific methylation block (n=69). For panel A, Friedman tests were performed comparing paired results at baseline, EOT, and recovery time points. Results were considered significant when *P<0.05; ns, P≧0.05; lung epithelium p=0.98, cardiopulmonary endothelium p=0.02, cardiomyocytes p=0.03. Panel B shows the correlation between lung epithelial cfDNA and dosimetric data. EOT / baseline represents the percentage of lung epithelial cfDNA after radiation at the end of treatment (EOT) compared to baseline levels. The volume of the lung receiving the 20 Gy dose is expressed as lung V20 (%) and the average radiation dose to the whole body is expressed as total body average (Gy). For panel B, Pearson correlation r was calculated and linear correlation was considered significant when *P<0.05. [Fig. 16C-D] Panel C shows fold change in lung epithelial cfDNA at EOT and recovery compared to baseline. For panel C, a Wilcoxon matched-pairs signed-rank test was performed between groups and results were considered significant if *P<0.05. Data are presented as mean ± SD; n=15. Panel D shows CPEC cfDNA (Geq / mL). Fragment-level deconvolution used CPEC-specific methylation blocks (n=132). For panel D, a Friedman test was performed comparing paired results at baseline, EOT, and recovery time points. Results were considered significant if *P<0.05; ns, P≧0.05; lung epithelium p=0.98, cardiopulmonary endothelium p=0.02, cardiomyocytes p=0.03. [Fig. 16E-F]Panel E shows the correlation between CPEC cfDNA and dosimetry data. The volume of lung receiving a 5 Gy dose is represented by Lung V5 (%). For panel E, Pearson correlation r was calculated and linear correlation was considered significant if *P<0.05. Panel F shows the fold change of CPEC cfDNA at EOT and recovery compared to baseline levels. For panel F, Wilcoxon matched-pairs signed rank test was performed between groups and results were considered significant if *P<0.05. Data are presented as mean ± SD; n=15. [Fig. 16G-H] Panel G shows cardiomyocyte cfDNA (Geq / mL). Fragment level deconvolution used cardiomyocyte specific methylation block (n=375). For panel G, Friedman test was performed comparing paired results at baseline, EOT, and recovery time points. Results were considered significant if *P<0.05; ns, P>0.05; lung epithelium p=0.98, cardiopulmonary endothelium p=0.02, cardiomyocyte p=0.03. Panel H shows correlation between cardiomyocyte cfDNA and maximum heart dose (Gy). For panel H, Pearson correlation r was calculated and linear correlation was considered significant if *P<0.05. [Figure 16I] Panel I shows the fold change of cardiomyocyte cfDNA at EOT and recovery compared to baseline. For Panel I, Wilcoxon matched-pairs signed rank test was performed between groups and results were considered significant when *P<0.05. Data are presented as mean ± SD; n=15. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] Detailed Description of the Invention The practice of the present invention may employ, unless otherwise indicated, conventional techniques of genetics, molecular biology, computational biology, genomics, epigenomics, mass spectrometry, and bioinformatics, which are within the skill of those in the art.

[0028] In order that the present invention may be more readily understood, certain terms are first defined. Further definitions are provided throughout this disclosure. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention pertains.

[0029] Any headings provided herein are not limitations of the various aspects or embodiments of the invention, which can be seen by reference to the specification in its entirety, and therefore the terms defined immediately below are more fully defined by reference to the specification in its entirety.

[0030] All references cited in this disclosure are incorporated herein by reference in their entirety. Additionally, any manufacturer's instructions or any product catalogs cited or referred to herein are incorporated herein by reference. Any document incorporated by reference into this document or any teaching therein may be used in the practice of the present invention. Documents incorporated by reference into this document are not admitted to be prior art.

[0031] definition The phrases or terms in this disclosure are for purposes of description, and not of limitation, such that the terms or phrases herein will be interpreted in the context of teaching and guidance by those skilled in the art.

[0032] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. The term "a" (or "an") and the terms "one or more" and "at least one" can be used interchangeably.

[0033] Furthermore, "and / or" shall be considered as a specific disclosure of each of the two specified features or components with or without the other. Thus, the term "and / or" used in a phrase such as "A and / or B" is intended to include A and B, A or B, A (alone), and B (alone). Similarly, the term "and / or" used in a phrase such as "A, B, and / or C" is intended to include A, B, and C; A, B, or C; A or B; A or C; B or C; A and B; A and C; B and C; A (alone); B (alone); and C (alone).

[0034] Wherever an embodiment is described with the words "comprising," similar embodiments are encompassed except where described in terms of "consisting of" and / or "consisting essentially of."

[0035] Units, prefixes, and symbols are given in the form accepted by their International System of Units (SI). Numerical ranges include the numbers defining the range, and any individual value provided herein can serve as an endpoint of a range that includes other individual values ​​provided herein. For example, a set of values ​​such as 1, 2, 3, 8, 9, and 10 is also a disclosure of numerical ranges such as 1-10, 1-8, 3-9, etc. Similarly, a disclosed range is a disclosure of each individual value (i.e., intermediate values) encompassed within said range, including integers and fractions. For example, a stated range of 5-10 is also a disclosure of 5, 6, 7, 8, 9, and 10 individually, as well as 5.2, 7.5, 8.7, etc.

[0036] Unless otherwise indicated, the terms "at least" or "about" preceding a series of elements should be understood to refer to every element in that series. The term "about" preceding a numerical value includes ±10% of the cited value. For example, a concentration of about 1 mg / mL includes 0.9 mg / mL to 1.1 mg / mL. Similarly, a concentration range of about 1% to 10% (w / v) includes 0.9% (w / v) to 11% (w / v).

[0037] The term "cell-free DNA" or "cfDNA" or "circulating cell-free DNA" as used herein refers to DNA circulating in the peripheral blood of a subject. The DNA molecules in cfDNA may have a median size of 1 kb or less (e.g., about 50 bp to 500 bp, or about 80 bp to 400 bp, or about 100 bp to 1 kb), although fragments with median sizes outside this range may exist. This term is intended to encompass free DNA molecules circulating in the bloodstream as well as DNA molecules present in extracellular vesicles (e.g., exosomes).

[0038] "Methylation site" refers to a CpG dinucleotide.

[0039] "Methylation pattern" refers to the pattern produced by the presence of methylated or unmethylated CpGs in a segment of DNA. For example, in a segment of DNA containing three CpGs, one methylation pattern is where all three CpGs are methylated; a different methylation pattern is where all three CpGs are unmethylated; another methylation pattern is where only the first CpG is methylated; yet another methylation pattern is where only the second CpG is methylated; yet another methylation pattern is where both the first and second CpGs are methylated, etc.

[0040] "Methylation status" refers to whether a CpG dinucleotide is methylated or unmethylated.

[0041] As used herein, "hypermethylation" refers to the presence of methylated CpGs. For example, a hypermethylated genomic region means that every CpG in the genomic region is methylated.

[0042] As used herein, "hypomethylation" refers to the presence of unmethylated CpGs. For example, a hypomethylated genomic region means that each CpG in the genomic region is unmethylated.

[0043] As used herein, the term "sequencing" refers to a method by which the identity of at least 10 consecutive nucleotides of a polynucleotide is obtained, e.g., the identity of at least 20, at least 50, at least 100, or at least 200, or more consecutive nucleotides.

[0044] The term "next-generation sequencing" as used herein refers to parallelized sequencing-by-synthesis or sequencing by ligation platforms currently utilized by companies such as Illumina, Life Technologies, and Roche. Next-generation sequencing methods may also include nanopore sequencing methods such as those commercialized by Oxford Nanopore Technologies, electronic detection-based methods such as Ion Torrent technology commercialized by Life Technologies, or single-molecule fluorescence-based methods such as those commercialized by Pacific Biosciences.

[0045] A "subject" or "individual" or "patient" is any subject, particularly a mammalian subject, for which a diagnosis, prognosis, or treatment is desired. Mammalian subjects include humans, domestic animals, livestock, sports animals, and laboratory animals, such as humans, non-human primates, dogs, cats, pigs, cows, horses, rodents (such as rats and mice), rabbits, and the like.

[0046] An "effective amount" of an active agent (factor) is an amount sufficient to carry out a specifically stated purpose.

[0047] Terms such as "treating" or "treatment" or "to treat" or "alleviating" or "to alleviate" refer to therapeutic measures that cure, slow, reduce symptoms, and / or halt the progression of a diagnosed pathological condition or disorder. In certain embodiments, a subject is successfully "treated" for a disease or disorder when the patient exhibits a total, partial, or transient alleviation or elimination of at least one symptom or measurable physical parameter associated with the disease or disorder.

[0048] How to use cfDNA to determine tissue damage The present invention relates to a method for determining tissue damage using circulating cfDNA. The majority of cfDNA fragments are maximized at around 167bp, which corresponds to the length of DNA including nucleosomes (147bp) and linker fragments (20bp). The footprint of nucleosomes in this cfDNA reflects degradation by nucleases as a by-product of cell death (Heitzer et al., 2020).

[0049] DNA methylation typically involves the covalent addition of a methyl group to the 5-carbon of cytosine (5mc), and the human and mouse genomes contain 28 million and 13 million CpG sites, respectively (Greenberg and Bourc'his, 2019; Michalak et al., 2019). Stable cell-type-specific patterns of DNA methylation are preserved during DNA replication, thus providing a major mechanism for genetic cellular memory during cell proliferation (Kim & Costello, 2017; Dor & Cedar, 2018). DNA methylation changes associated with disease and physiological aging occur at all locations in the epigenome, distinct from regions critical for cell type identity, making methylated cfDNA a robust cell-type-specific readout across diverse patient populations (Michalak et al., 2019; Dor & Cedar, 2018).

[0050] Recent studies have demonstrated the feasibility of tissue of origin (TOO) analysis using cfDNA methylation, but such studies have traditionally averaged the methylation state across a population of fragments at a single CpG site (Barefoot, et al., 2021; Barefoot et al., 2020). The present invention involves sequencing a portion of the cfDNA to identify patterns of differential methylation and using these patterns of differential methylation to determine the cellular origin of the cfDNA.

[0051] The use of differential methylation patterns to determine the cellular origin of cfDNA can be applied to a method for determining whether a subject suffers from tissue damage resulting from exposure to a toxic agent. In some embodiments, the method includes: (a) sequencing cfDNA in a biological sample from a subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, and when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern, the cellular origin of the cell-free DNA is determined; (c) measuring the amount of the determined cfDNA of cellular origin; and (d) comparing the measured amount of the determined cfDNA of cellular origin with a normal amount of the determined cfDNA of cellular origin. An increase in the measured amount of the determined cfDNA of cellular origin over the normal amount of the determined cfDNA of cellular origin indicates that the subject suffers from or will suffer from tissue damage resulting from the exposure.

[0052] In some embodiments, a method for determining whether a subject is suffering from tissue damage resulting from exposure to a toxic agent comprises the steps of: (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain a methylation site, wherein the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell-type specific methylation pattern; and (c) measuring the amount of the determined cfDNA of cellular origin at a later time point compared to an earlier time point, which indicates that the subject is suffering from or will suffer from tissue damage resulting from the exposure.

[0053] The use of differential methylation patterns to determine the cellular origin of cfDNA can also be applied to a method of treating a subject suffering from tissue damage resulting from exposure to a toxic agent. In some embodiments, these methods include administering a treatment for tissue damage to a subject, wherein the subject is indicated to be suffering from tissue damage by a method comprising: (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cell-free DNA is determined when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) measuring the amount of the determined cfDNA of cellular origin; and (d) comparing the measured amount of the determined cfDNA of cellular origin with a normal amount of the determined cfDNA of cellular origin. An increase in the measured amount of the determined cfDNA of cellular origin over the normal amount of the determined cfDNA of cellular origin indicates that the subject is suffering from tissue damage.

[0054] In some embodiments, a method of treating a subject suffering from tissue damage resulting from exposure to a toxic agent comprises administering a treatment for tissue damage to a subject, wherein the subject has been shown to be suffering from tissue damage by a method comprising the steps of: (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain a methylation site, wherein the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell-type specific methylation pattern; and (c) measuring the amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point. An increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point indicates that the subject is suffering from tissue damage.

[0055] In another embodiment, the method is a method for treating tissue damage in a subject. The method includes administering a treatment for tissue damage to the subject and monitoring the effectiveness of the treatment. The monitoring includes (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; (c) measuring the amount of the determined cellular origin cfDNA; and (d) comparing the measured amount of the determined cellular origin cfDNA with a normal amount of the determined cellular origin cfDNA. A decrease in the measured amount of the determined cellular origin cfDNA compared to the normal amount of the determined cellular origin cfDNA indicates that the treatment is effective. An increase or no change in the measured amount of the determined cellular origin cfDNA above the normal amount of the determined cellular origin cfDNA indicates that the treatment is not effective.

[0056] In some embodiments, a method for treating tissue damage comprises administering a treatment for tissue damage to a subject and monitoring the effectiveness of the treatment. The monitoring comprises: (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; and (c) measuring the amount of the determined cellular origin cfDNA. A decrease in the measured amount of the determined cellular origin cfDNA at a later time point compared to a previous time point indicates that the treatment is effective. An increase or no change in the measured amount of the determined cellular origin cfDNA at a later time point compared to a previous time point indicates that the treatment is not effective.

[0057] In some embodiments, the method may further include administering an adjusted treatment if the first treatment is determined to be ineffective. In some embodiments, the tissue damage is caused by exposure to a toxic agent.

[0058] In some embodiments, the toxic factor comprises radiation, which may be radiation for therapeutic purposes, accidental radiation, or environmental radiation.

[0059] In some embodiments, the toxic agent is radiation therapy, and in certain embodiments, radiation therapy comprises external beam radiation therapy. Examples of external beam radiation therapy include, but are not limited to, conventional external beam radiation therapy, stereotactic radiation therapy, three-dimensional conformal radiation therapy, intensity modulated radiation therapy, intensity modulated arc therapy, episodic feathered radiation therapy, particle therapy, and Auger therapy.

[0060] In certain embodiments, radiation therapy includes brachytherapy, where the radiation is in a sealed source. Brachytherapy may be interstitial brachytherapy, where the radiation source is placed directly into the diseased target tissue; or brachytherapy may be contact brachytherapy, where the radiation source is placed in the space adjacent to the target tissue, for example in a body cavity (intracavitary brachytherapy), a body lumen (intraluminal brachytherapy), or externally (surface brachytherapy).

[0061] In certain embodiments, radiation therapy includes systemic radioisotope therapy, which delivers radiation to a target site, for example, using the chemical properties of the isotope or binding of the isotope to another molecule or antibody that directs the isotope to the target site.

[0062] In some embodiments, the toxic factor is accidental radiation, for example, work-related exposure to radiation.

[0063] In some embodiments, the toxic agent is environmental radiation, including, by way of non-limiting example, exposure to radiation resulting from high altitude flight and space travel.

[0064] In some embodiments, the toxic agent comprises a radioactive material that is ingested by the subject, inhaled by the subject, or absorbed through body surface contamination by the subject.

[0065] In some embodiments, the virulence factor comprises a microorganism. In certain embodiments, the virulence factor comprises a pathogen, such as a bacterium or a virus. Specific examples of pathogens include, but are not limited to, species of the following genera: Bacillus, Brucella, Clostridium, Corynebacterium, Enterococcus, Escherichia, Klebsiella, Leptospira, Listeria, Mycobacterium, Mycoplasma, Neisseria, Pseudomonas, Staphylococcus, Treponema, Vibrio, and Yersinia.

[0066] In some embodiments, the toxic agent comprises a toxin from a chemical source or a toxin from a biological source.

[0067] In some embodiments, toxic factors include medications, such as chemicals, used for therapeutic purposes.

[0068] In some embodiments, the toxic agent comprises a weapon, for example a chemical or biological or radiological agent used as a weapon in a terrorist attack or war.

[0069] In yet another embodiment, a method of treating a subject includes administering a treatment to the subject and monitoring whether the treatment causes tissue damage in the subject. The monitoring includes (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type specific methylation pattern; (c) measuring the amount of the determined cellular origin cfDNA; and (d) comparing the measured amount of the determined cellular origin cfDNA with a normal amount of the determined cellular origin cfDNA. An increase in the measured amount of the determined cellular origin cfDNA over the normal amount of the determined cellular origin cfDNA indicates that the treatment is causing tissue damage.

[0070] In another embodiment, a method of treating a subject includes administering a treatment to the subject and monitoring whether the treatment causes tissue damage in the subject. The monitoring includes: (a) sequencing cfDNA in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain a methylation site, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; and (c) measuring the amount of cfDNA of the determined cellular origin at the later time point compared to the earlier time point. An increase in the measured amount of cfDNA of the determined cellular origin at the later time point indicates that the treatment causes tissue damage.

[0071] In some embodiments, the method may further include administering an adjusted treatment if the first treatment is determined to cause tissue damage.

[0072] In some embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin that is generated in a population of individuals that have not been exposed to a toxic agent. In other embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin that is generated in a population of individuals that have not been administered a treatment.

[0073] Another aspect of the present invention is a method for determining organ, tissue, or cell type damage induced by a substance administered to a subject. The method includes the steps of: (a) sequencing cfDNA in a biological sample from a subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) measuring the amount of the determined cellular origin cfDNA; and (d) comparing the measured amount of the determined cellular origin cfDNA with a normal amount of the determined cellular origin cfDNA. An increase in the measured amount of the determined cellular origin cfDNA over the normal amount of the determined cellular origin cfDNA indicates that an organ or tissue of the cell type, or the cell type itself, is suffering from damage. In some embodiments, the substance administered to the subject can be a pharmaceutical product, such as an investigational new drug.

[0074] Yet another aspect of the present invention is a method for determining organ, tissue, or cell type damage induced by a substance administered to a subject. The method includes the steps of: (a) sequencing cfDNA in a biological sample from a subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; and (c) measuring the amount of cfDNA of the determined cellular origin at a later time point compared to a previous time point. An increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to a previous time point indicates that an organ or tissue of the cell type, or the cell type itself, is suffering from damage. In some embodiments, the substance administered to the subject can be a pharmaceutical product, such as an investigational new drug.

[0075] A further aspect of the present invention is a method for determining the organ, tissue, or cell target of a substance administered to a subject. The method includes the steps of: (a) sequencing cfDNA in a biological sample from a subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) measuring the amount of the determined cellular origin cfDNA; and (d) comparing the measured amount of the determined cellular origin cfDNA with a normal amount of the determined cellular origin cfDNA. An increase in the measured amount of the determined cellular origin cfDNA over the normal amount of the determined cellular origin cfDNA indicates that the organ or tissue of the cell type, or the cell type itself, is a target of the substance. In an embodiment, the substance administered to a subject can be a pharmaceutical substance, such as an investigational new drug.

[0076] Yet another aspect of the present invention is a method for determining the organ, tissue, or cell target of a substance administered to a subject. The method includes the steps of: (a) sequencing cfDNA in a biological sample from a subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, where the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; and (c) measuring the amount of cfDNA of the determined cellular origin. An increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to a previous time point indicates that the organ or tissue of the cell type, or the cell type itself, is a target of the substance. In an embodiment, the substance administered to a subject can be a pharmaceutical product, such as an investigational new drug.

[0077] In some embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin that is generated in a population of individuals that have not been exposed to a toxic agent. In other embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin that is generated in a population of individuals that have not been administered a treatment.

[0078] In some embodiments, the normal amount of cfDNA for a determined cellular origin is the amount of cfDNA for said determined cellular origin that is expected for said determined cellular origin.

[0079] In some embodiments, the two or more time points can all be after exposure to a treatment or toxic agent, hi some embodiments, at least one of the two or more time points can be before exposure to a treatment or toxic agent.

[0080] The above time points may be, for example, one or more days apart, and may be, for example, daily, every 2, 3, 4, 5, 6 days, weekly, every 2 weeks, every 3 weeks, every 4 weeks, monthly, every 2 months, every 3 months, every 4 months, every 5 months, every 6 months, every 7 months, every 8 months, every 9 months, every 10 months, every 11 months, yearly, or any time point therebetween.

[0081] The increase in the measured amount of cfDNA of the determined cell origin over the normal amount of cfDNA of the determined cell origin or over the previously measured amount of cfDNA of the determined cell origin can be, for example, a percent increase of about 0.1% to 100%, such as about 0.1%, 0.5%, 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100%; or a fold increase of at least about 2-fold, such as about 2-fold, or 3-fold, or 4-fold, or 5-fold, or 6-fold, or 7-fold, or 8-fold, or 9-fold, or 10-fold. In some embodiments, the increase can be any increase determined to be statistically significant (e.g., p≦0.05, p≦0.01, etc.) as calculated by statistical methods known in the art.

[0082] In some embodiments, the subject has cancer.

[0083] The biological sample can be a bodily fluid obtained from a subject, such as, but not limited to, whole blood, plasma, serum, urine, or any other bodily fluid sample produced by a subject, such as saliva, cerebrospinal fluid, urine, or sputum, etc. In certain embodiments, the biological sample is whole blood, plasma, or serum.

[0084] Methods for quantifying cfDNA are known in the art and include, but are not limited to, PCR; fluorescence-based quantification methods (e.g., Qubit); chromatographic techniques, such as gas chromatography, supercritical fluid chromatography, and liquid chromatography, such as partition chromatography, adsorption chromatography, ion exchange chromatography, size exclusion chromatography, thin layer chromatography, and affinity chromatography; electrophoretic techniques, such as capillary electrophoresis, capillary zone electrophoresis, capillary isoelectric focusing, capillary electrochromatography, micellar electrokinetic capillary chromatography, isotachophoresis, transient isotachophoresis, and capillary gel electrophoresis; comparative genomic hybridization; microarrays; and bead arrays.

[0085] How to combine epigenetic and genetic analysis The use of differential methylation patterns to determine the cellular origin of cfDNA can be combined with genetic analysis of cfDNA. Such a combination can be applied to therapeutic methods that include monitoring treatment response and therapy-related adverse events. The combination of changes in mutant ctDNA and changes in the proportion of cell type-specific cfDNA can reflect therapeutic intervention-based changes. The half-life of cfDNA is 15 minutes to 2 hours. Rapid clearance allows for serial analysis of disease progression over time, especially under selective pressure from ongoing therapy. The method of the present invention allows for serial sampling, including baseline comparisons, from which therapy-related relative changes can be evaluated, taking into account patient-specific comorbidities at an individualized level.

[0086] The combination of genetic and epigenetic analysis of cell-free DNA has many inherent advantages when applied to precision therapy in cancer. Liquid biopsies have been shown to accurately characterize tumor genotypes, enable molecular subtyping, and provide a comprehensive view of intratumoral heterogeneity. High sampling frequency allows modeling of the evolutionary dynamics of tumor progression. Similarly, molecular changes identified after the initiation of therapy can provide insight into therapy response and even track tumor subclones that may lead to the emergence of therapy resistance. The whole-body view provided by serial liquid biopsies is ideal to monitor widespread changes that may better inform clinical decisions despite uncertainties. For example, in the case of surgical removal of a tumor or successful treatment, liquid biopsies can be used to monitor for minimal residual disease and recurrence. While ctDNA can be used to track molecular changes in circulation, there is the benefit of monitoring changes in the host microenvironment associated with cancer in parallel, requiring the combination of genetic and epigenetic analysis. Cell-specific cfDNA methylation patterns of normal cells can also be used in combination with ctDNA to assess the impact of therapy on the surrounding tumor microenvironment. This is particularly useful for monitoring metastatic disease in tissue types distant from the primary tumor, as well as monitoring therapy-related toxicity in somatic cell types. Furthermore, liquid biopsies can help to reveal factors underlying clinical outcomes, providing a basis for recommending different treatments based on predicted patient benefit. Liquid biopsies can identify predictive biomarkers to guide treatment selection, recognize off-target effects, and develop personalized treatment plans for patients. These applications provide a more complete picture of treatment response as well as tissue-specific cytotoxicity, better informing clinical care and management throughout the treatment process.

[0087] The minimally invasive nature of liquid biopsies allows for serial sampling to monitor changes over time, especially under selection pressure from ongoing therapies. ctDNA can be used to track clonal heterogeneity over time, assess treatment response, and detect treatment-resistant clones. Normal cell-specific cfDNA methylation patterns can be used in combination with ctDNA to assess the impact of treatment on the surrounding tumor microenvironment and monitor therapy-associated toxicity in somatic cell types (Figure 1).

[0088] The use of differential methylation patterns to determine the cellular origin of cfDNA combined with genetic analysis can be applied to a method of treating a subject with a tumor. In some embodiments, the method includes: (a) monitoring the response to a first treatment, the adverse reaction to a first treatment, or a combination thereof, wherein the monitoring comprises performing genetic and epigenetic analysis of cfDNA, ctDNA, or a combination thereof at two or more time points, and optionally comparing with normal cfDNA, ctDNA, or a combination thereof to determine whether to change the first treatment; and (b) administering a treatment adjusted according to the genetic and epigenetic analysis or continuing the first treatment.

[0089] In other embodiments, the method includes the steps of: (A) monitoring a response to a first treatment, an adverse reaction to a first treatment, or a combination thereof, wherein the monitoring comprises: (i) determining whether there is an adverse reaction to a first treatment, comprising: (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of a sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cell-free DNA is determined when the methylation pattern in the one or more portions is the same as a known cell-type specific methylation pattern; (c) measuring the amount of the determined cellular origin cfDNA, and (d) comparing the measured amount of the determined cellular origin cfDNA with a normal amount of the determined cellular origin cfDNA, wherein an increase in the measured amount of the determined cellular origin cfDNA above the normal amount of the determined cellular origin cfDNA indicates an adverse reaction; and (ii) determining whether there is a response to a first treatment, comprising: (a) determining whether there is an adverse reaction to a first treatment, wherein the determining comprises: The method includes the steps of: (a) sequencing ctDNA in a biological sample from the subject; and (b) determining the clonal heterogeneity of cells of the tumor by genotyping the ctDNA, wherein the presence of two or more clones of the tumor cells or the presence of a tumor cell clone not previously identified in the subject indicates an ineffective response to the first treatment; and (B) administering the same treatment as the first treatment if it is determined that there is no adverse reaction, there is no ineffective response, or a combination thereof; or administering an adjusted treatment if it is determined that there is an adverse reaction, there is an ineffective response, or a combination thereof.

[0090] In some embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin generated in a population of individuals who have not received the first treatment.In other embodiments, the normal amount of cfDNA comprises the amount of cfDNA for the determined cell origin generated in a population of individuals who do not have a tumor.

[0091] In some embodiments, the normal amount of cfDNA for a determined cell origin is the amount of cfDNA for said determined cell origin that is expected for said determined cell origin.

[0092] In yet other embodiments, the method includes the steps of: (A) monitoring a response to a first treatment, an adverse reaction to a first treatment, or a combination thereof, the monitoring comprising, at two or more time points, (i) determining whether there is an adverse reaction to a first treatment, comprising: (a) sequencing cfDNA in a biological sample from the subject; (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of a sequence of the cfDNA that contain a methylation site, wherein the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell-type specific methylation pattern; and (c) measuring an amount of cfDNA of the determined cellular origin, wherein an increase in the measured amount of cfDNA of the determined cellular origin measured at a later time point compared to an earlier time point indicates an adverse reaction; and (ii) determining whether there is a response to a first treatment, comprising: (a) sequencing ctDNA in a biological sample from the subject; and (b) determining a cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of a sequence of the cfDNA that contains a methylation site, wherein the cellular origin of the cell-free DNA is determined if the methylation pattern in the one or more portions is the same as a known cell-type specific methylation pattern; and (c) measuring an amount of cfDNA of the determined cellular origin, wherein an increase in the measured amount of cfDNA of the determined cellular origin measured at a later time point compared to an earlier time point indicates an adverse reaction; and (ii) determining whether there is a response to a first treatment, comprising: (a) sequencing ctDNA in a biological sample from the subject; and (b) The method includes the steps of: (A) determining the clonal heterogeneity of the cells of the tumor by genotyping the ctDNA, wherein the presence of two or more clones of the tumor cells or the presence of a tumor cell clone at a later time point not identified at a previous time point indicates an ineffective response to the first treatment; and (B) administering the same treatment as the first treatment if it is determined that there is no adverse reaction, there is no ineffective response, or a combination thereof; or administering an adjusted treatment if it is determined that there is an adverse reaction, there is an ineffective response, or a combination thereof.

[0093] In some embodiments, the subject has a tumor associated with cancer. Examples of cancer include, but are not limited to, colorectal cancer, brain cancer, ovarian cancer, prostate cancer, pancreatic cancer, breast cancer, renal cancer, nasopharyngeal cancer, hepatocellular carcinoma, melanoma, skin cancer, oral cancer, head and neck cancer, esophageal cancer, gastric cancer, cervical cancer, bladder cancer, lymphoma, chronic or acute leukemia (e.g., B, T, and myeloid origin), sarcoma, lung cancer, and multidrug resistant cancer. Other examples are diseases that require drug treatment with chemicals (small molecules) or proteins such as insulin or antibodies. Such diseases can be metabolic diseases such as diabetes, or infectious diseases such as bacterial or viral infections, e.g., hepatitis, or cardiovascular diseases, e.g., but are not limited to, hypertension, coronary artery disease, cerebrovascular disease, or peripheral vascular disease.

[0094] In some embodiments, cfDNA is used to compare damage to cells from a first treatment to undamaged normal cells from the same tissue.

[0095] In some embodiments, the methylation pattern in the cfDNA is evaluated. In certain embodiments, the methylation pattern of cfDNA from damaged cells and cfDNA from healthy cells is compared.

[0096] In some embodiments, the analysis includes comparing damaged cells to healthy cells to ascertain where the damage occurred.

[0097] In some embodiments, the treatment comprises chemotherapy, radiation therapy, targeted therapy, immunotherapy, or a combination thereof.

[0098] In some embodiments, the two or more time points can all be after the first treatment. In some embodiments, at least one of the two or more time points can be before the first treatment.

[0099] The time points may be, for example, one or more days apart, and may be, for example, daily, every 2, 3, 4, 5, 6 days, weekly, every 2 weeks, every 3 weeks, every 4 weeks, monthly, every 2 months, every 3 months, every 4 months, every 5 months, every 6 months, every 7 months, every 8 months, every 9 months, every 10 months, every 11 months, yearly, or any time point therebetween.

[0100] The increase in the measured amount of cfDNA of the determined cell origin over the normal amount of cfDNA of the determined cell origin or over the previously measured amount of cfDNA of the determined cell origin can be, for example, a percent increase of about 0.1% to 100%, such as about 0.1%, 0.5%, 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100%; or a fold increase of at least about 2-fold, such as about 2-fold, or 3-fold, or 4-fold, or 5-fold, or 6-fold, or 7-fold, or 8-fold, or 9-fold, or 10-fold. In some embodiments, the increase can be any increase determined to be statistically significant (e.g., p≦0.05, p≦0.01, etc.) as calculated by statistical methods known in the art.

[0101] The biological sample can be a bodily fluid obtained from a subject, such as, but not limited to, whole blood, plasma, serum, urine, or any other bodily fluid sample produced by a subject, such as saliva, cerebrospinal fluid, urine, or sputum, etc. In certain embodiments, the biological sample is whole blood, plasma, or serum.

[0102] Methods for quantifying cfDNA are known in the art and include, but are not limited to, PCR; fluorescence-based quantification methods (e.g., Qubit); chromatographic techniques, such as gas chromatography, supercritical fluid chromatography, and liquid chromatography, such as partition chromatography, adsorption chromatography, ion exchange chromatography, size exclusion chromatography, thin layer chromatography, and affinity chromatography; electrophoretic techniques, such as capillary electrophoresis, capillary zone electrophoresis, capillary isoelectric focusing, capillary electrochromatography, micellar electrokinetic capillary chromatography, isotachophoresis, transient isotachophoresis, and capillary gel electrophoresis; comparative genomic hybridization; microarrays; and bead arrays.

[0103] Another aspect of the invention relates to a method for detecting and / or quantifying changes in circulating methylated DNA in a patient undergoing treatment.

[0104] A further aspect of the invention relates to probes designed for any tissue and / or cell type within a tissue to detect changes in the abundance of tissue-specific DNA fragments in the circulation.

[0105] Analysis of cfDNA The present invention encompasses the analysis of cfDNA to determine the cellular origin of cfDNA, which includes identifying methylation patterns in the sequence of cfDNA and comparing the methylation patterns in the sequence of cfDNA to known methylation patterns associated with different cell types.

[0106] Table 1 provides examples of cell origins associated with different types of tissue.

[0107] [Table 1]

[0108] cfDNA can be obtained by centrifuging a bodily fluid, such as whole blood, to remove all cells, and then isolating the DNA from the remaining plasma or serum. Such methods are well known (see, for example, Lo et al., 1998). Circulating cfDNA and ctDNA can be double-stranded or single-stranded DNA.

[0109] In the present invention, different DNA methylation detection techniques can be used, including, but not limited to, restriction enzyme digestion approaches, which include a step of cutting DNA at enzyme-specific CpG sites; affinity enrichment methods, such as methylated DNA immunoprecipitation sequencing (MeDIP-seq) or methyl-CpG-binding domain sequencing (MBD-seq); bisulfite conversion methods, such as whole genome bisulfite sequencing (WGBS), reduced representation bisulfite sequencing (RRBS), methylated CpG tandem amplification and sequencing (MCTA-seq) and methylation arrays; enzymatic approaches, such as enzymatic methyl sequencing (EM-seq) or ten-eleven translocation (TET)-assisted pyridine borane sequencing (TAPS); and other methods that do not require DNA processing, such as nanopore sequencing from Oxford Nanopore Technologies (ONT) and single molecule real-time (SMRT) sequencing from Pacific Biosciences (PacBio).

[0110] The comparison of the methylation pattern in the sequence of the cfDNA with the known methylation pattern may include identifying the presence of a methylation pattern in the sequence of the cfDNA or a portion thereof that originates from a particular cell type. In some embodiments, the presence of a methylation pattern is performed by hybridization capture sequencing of the cfDNA. In other embodiments, the presence of a methylation pattern is performed using bisulfite amplicon sequencing.

[0111] The methylation pattern may comprise a segment of a nucleotide sequence comprising at least one CpG dinucleotide, or at least about two CpG dinucleotides, or at least about three CpG dinucleotides, in some embodiments, the methylation pattern may comprise a segment of a nucleotide sequence comprising at least about four CpG dinucleotides, or at least about five CpG dinucleotides, or at least about six CpG dinucleotides, or at least about seven CpG dinucleotides, or at least about eight CpG dinucleotides, or at least about nine CpG dinucleotides, or at least about ten CpG dinucleotides.

[0112] Table 2 provides the methylation status of CpG dinucleotides in genomic regions that represent different cell types. The presence of the same methylation pattern between the sequence of cfDNA and the genomic region shown in Table 2 indicates the cell type from which the cfDNA is derived. Table 2 provides the continuous methylation status across multiple adjacent CpG sites (patterns) within a genomic region.

[0113] [Table 2] TIFF2024529192000003.tif250164TIFF2024529192000004.tif250164TIFF2024529192000005.tif250164TIFF20245291920 00006.tif250164TIFF2024529192000007.tif250164TIFF2024529192000008.tif250164TIFF2024529192000009.tif250164 TIFF2024529192000010.tif250164TIFF2024529192000011.tif250164TIFF2024529192000012.tif250164TIFF20245291920 00013.tif250164TIFF2024529192000014.tif250164TIFF2024529192000015.tif250164TIFF2024529192000016.tif250164 TIFF2024529192000017.tif250164TIFF2024529192000018.tif250164TIFF2024529192000019.tif250164TIFF20245291920 00020.tif249167TIFF2024529192000021.tif249167TIFF2024529192000022.tif249167TIFF2024529192000023.tif249167 TIFF2024529192000024.tif249167TIFF2024529192000025.tif249167TIFF2024529192000026.tif249167TIFF20245291920 00027.tif249167TIFF2024529192000028.tif249167TIFF2024529192000029.tif249167TIFF2024529192000030.tif186168

[0114] Analysis of ctDNA An embodiment of the present invention includes the analysis of ctDNA to determine the clonal heterogeneity of tumor cells. The determination of tumor cell heterogeneity includes genotyping ctDNA to obtain a genotypic profile of ctDNA. The genotypic profile of ctDNA can be compared with the genotypic profile of ctDNA previously obtained from the subject, and is well established in cancer genotyping for signature mutations or previously unknown mutations. These mutations can be point mutations, methylation changes, tumor-specific rearrangements (e.g., inversions, translocations, insertions and deletions), or cancer-derived viral sequences.

[0115] Examples of methods that may be used in genotyping include, but are not limited to, sequencing, such as whole genome or whole exome sequencing; PCR; Sanger-based ctDNA detection methods (Newman et al., 2014); BEAMing (beads, emulsion, amplification, and magnetics) developed by Diehl et al. (2008); and Cancer Personalized Profiling by Deep Sequencing (CAPP-seq) (Newman et al., 2014). EXAMPLES

[0116] We conducted a study to establish sequencing-based, cell-type-specific DNA methylation reference maps of human and mouse tissues that allow the assignment of DNA released into the circulation from dying cells back to its cellular origin. This study showed that cell-free methylated DNA in blood samples revealed tissue-specific cellular damage derived from radiation therapy.

[0117] method Human serum sample collection. Breast cancer patients undergoing adjuvant radiotherapy participated in the study. For serum isolation, peripheral blood (approximately 8–12 ml) was collected, allowed to clot for 30 min at room temperature, and then centrifuged at 1500 x g for 20 min at 4 °C to separate the serum fraction. This serum was aliquoted in 0.5 mL fractions and stored at -80 °C until use. Serial serum samples were collected from 15 breast cancer patients at baseline (before radiotherapy), at the end of treatment (EOT; 30 min after the last radiotherapy), and at recovery (1 month after cessation of radiotherapy), thus preparing an intra-patient internal standard and a baseline. A schematic diagram of the time series of sample collection can be seen in Figure 2. Patients received either three-dimensional conformal RT (3D-CRT) or proton beam therapy (PBT) combined with 3D-CR. Patient characteristics and treatment details, including radiation dosimetry, are summarized in Table 3 and Barefoot et al., 2022, Supplementary Table 8.

[0118] Mouse serum and tissue collection. C57Bl6 mice (n=18) were irradiated in the upper thorax at different doses (sham control, 3 Gy, 8 Gy) for three consecutive treatments. Serum and tissues were collected 24 hours after the last radiation dose. For serum isolation, blood was collected via cardiac puncture (approximately 1 mL) and allowed to clot for 30 minutes at room temperature, then centrifuged at 1500 xg for 20 minutes at 4°C to separate the serum fraction. Heart, lung, and liver tissues were excised, sectioned, flash frozen, and formalin fixed for subsequent analysis.

[0119] Cell isolation. Reference methylomes were generated for mouse immune cell types and human endothelial cell types to augment publicly available datasets. Peripheral blood and bone marrow were isolated and spleens from healthy C57Bl6 mice were dissociated into single cells and FACS sorted using cell type specific antibodies. Methylomes were generated for buffy coat (n=4), bone marrow (n=3), CD19+ B cells (n=1), CD4 T cells (n=1), CD8 T cells (n=1) and Gr1+ neutrophils (n=1) using the following antibodies: FITC anti-mouse CD45, Alexa Fluor 647 anti-mouse CD3, Brilliant Violet 711 anti-mouse CD4, Brilliant Violet 421 anti-mouse CD8a, PE anti-mouse CD19, PE / Cy7 anti-mouse Ly-6G / Ly-6C (Gr-1) (all BioLegend 1:20). Cryopreserved human liver sinusoidal endothelial cells (LSECs) at passage 1 were purchased. Purity was determined by immunofluorescence using antibodies specific for vWF / Factor VIII and CD31 (PECAM). Cryopreserved human coronary artery, cardiac microvascular, pulmonary artery, and pulmonary microvascular endothelial cells at passage 2 were isolated from purchased healthy human tissue from a single donor. All endothelial cell populations were CD31 positive and Dil-Ac-LDL uptake positive. Paired RNA-seq data were generated from the same cell populations used for DNA methylome profiling to verify the identity of the purchased cell populations through analysis of cell type expression markers.

[0120] RNA isolation, RNA sequencing, and RT-qPCR analysis. RNA was isolated from tissues or sorted cells using the RNeasy kit after a homogenization step using MagNA Lyser according to the manufacturer's protocol and quantified by Qubit RNA BR assay. Total RNA samples were validated using the Agilent RNA 6000 Nano assay on a 2100 Bioanalyzer TapeStation. The resulting RNA Integrity number (RIN) of samples selected for downstream qPCR or RNAseq analysis was at least 7. Reverse transcription was performed using the iScript cDNA Synthesis kit according to the manufacturer's protocol. Real-time quantitative RT-PCR was performed with iQ SYBR Green Supermix. Primers used for RT-qPCR were purchased from Integrated DNA Tehnologies. Fold changes were calculated as percentages normalized to the housekeeping gene human actin (ACTB) using the delta Ct method. All RT-qPCR assays were performed in triplicate. RNA sequencing libraries were generated using Novogene Corporation Inc.'s TruSeq Total RNA library Prep Kit, and 150-bp paired-end sequencing was performed on an Illumina Hiseq 4000 at a depth of 50,000,000 paired reads per sample. Reference indexes were generated using GTF annotations from GENCODEv28. Raw FASTQ files were aligned to GRCh38 or GRCm38 by HISAT2. Derived counts per million and P-values ​​were used to generate rank-ordered lists, which were then used for subsequent analysis and confirmation of the identity of isolated cell types for methylome analysis. The expression levels of known cell type markers obtained from a single-cell expression database were used to validate the identity of isolated cell type populations for methylome analysis (Khan et al., 2018).

[0121] Isolation of circulating cfDNA. Circulating cfDNA was extracted from 3–4 mL of human serum and 0.5 mL of mouse serum using the QIAamp Circulating Nucleic Acid kit according to the manufacturer's instructions. cfDNA was quantified using both the dsDNA High Sensitivity Assay Kit and the Qubit fluorometer. As a quality control, the fragment size distribution of isolated cfDNA was verified based on analysis using a 2100 Bioanalyzer TapeStation. Further purification using Beckman Coulter beads was performed to remove high molecular weight DNA reflecting cell lysis and leukocyte contamination, as previously described (Maggi et al., 2018). The size distribution of cfDNA fragments was re-verified after purification using 2100 Bioanalyzer TapeStation analysis.

[0122] Genomic DNA isolation and fragmentation. Genomic DNA from tissues was extracted with the DNeasy Blood and Tissue Kit according to the manufacturer's instructions and quantified with the Qubit fluorometer dsDNA BR assay kit. Prior to library construction, genomic DNA was fragmented to the recommended 150-200 base pairs by sonication using a Covaris E220 instrument. Lambda phage DNA was also fragmented and included at 0.5% w / w as a spike-in to all DNA samples to serve as an internal unmethylated standard. Bisulfite conversion efficiency was calculated by assessing the number of unconverted Cs on unmethylated lambda phage DNA. The SeqCap Epi capture pool contains probes to capture the lambda genomic region from base 4500-6500. Conversion rate was calculated as follows: conversion rate = 1 - (sum(C_count) / sum(CT_count)) over the captured lambda genomic region.

[0123] Bisulfite capture sequencing library construction. Bisulfite capture sequencing libraries were constructed from either cfDNA or reference DNA input following the same protocol. As a first step, WGBS libraries were constructed using Zymo Research Pico Methyl-Seq Library Prep Kit (D5455) with the following modifications. Bisulfite conversion was performed using Zymo EZ DNA Methylation Gold Kit instead of EZ DNA Methylation-Lightning Kit. For mouse samples, cfDNA from two mice in the same group was pooled as input for library construction. Two additional PCR cycles were added to the recommended cycle number based on the total input cfDNA amount. WGBS libraries were eluted in 15 μL 10 mM Tris-HCl buffer, pH 8. Library quality control was performed using an Agilent 2100 Bioanalyzer and quantity was determined by KAPA Library Quantification Kit.

[0124] Cell-free WGBS libraries were pooled to meet the required 1 μg DNA input required for targeted enrichment. However, no more than four WGBS libraries were pooled in a single hybridization reaction, and the 1 μg input DNA was divided equally between the libraries being multiplexed. Hybridization capture was performed using SeqCap Epi CpGiant probe pool for human samples and SeqCap Epi Developer probe for mouse samples with xGen Universal Blocker-TS Mix as blocking reagent according to the SeqCap Epi Enrichment System protocol. Washing and recovery of captured libraries, as well as PCR amplification and final purification, were performed as recommended by the manufacturer. Capture library products were evaluated by the Agilent Bioanalyzer DNA 1000 assay. Bisulfite capture sequencing libraries, including the 15–20% spike-in PhiX Control v3 library, were clustered on an Illumina Novaseq 6000 S4 flow cell and subsequently subjected to 150 bp paired-end sequencing.

[0125] Bisulfite sequencing data alignment and preprocessing. Paired-end FASTQ files were aligned using Trim Galore (https: / / github.com / FelixKrueger / TrimGalore) with the parameters "--paid -q 20 --clip_R1 10 --clip_R2 10 --three_prime_clip_R1 10 --three_prime_clip_R2 10" (https: / / github.com / FelixKrueger / Bismark). Aligned paired-end FASTQ reads were mapped to the human genome (assembled GRCh37 / hg) using Bismark (V 0.22.3) with the parameters "--non-directional" and then converted to BAM files using Santools (V. 1.12). BAM files were sorted and indexed using Santools (V1.12). Reads were cleaned from non-CpG nucleotides and converted to BETA and PAT files using webstools (V 0.1.0) ( https: / / github.com / nloyfer / wgbs_tools ), a tool suitable for working with WGBS data while maintaining inherent read-specific dependencies ( Loyfer et al., 2022 ; Loyfer & Kaplan).

[0126] Reference DNA methylation data from healthy tissues and cells. We requested controlled access to reference WGBS data from normal human tissues and cell types from public consortia participating in the International Human Epigenome Consortium (IHEC) and, once approved, downloaded them from the European Genome-Phenome Archive (EGA), Japanese Genotype-phenotype Archive (JGA), and database of Genotypes and Phenotypes (dbGAP) data repositories (Table 4; Barefoot et al., 2022, see also Supplementary Table 1). Reference mouse WGBS data from normal tissues and cell types were downloaded from selected GEO and SRA datasets (Table 5). The downloaded FASTQs were processed and realigned in the same way as the locally generated bisulfite sequencing libraries described above. However, parameters were adjusted to account for each WGBS library type in both the alignment and refinement steps as previously described in the Bismark User Guide (http: / / felixKrueger.github.io / Bismark / Docs / ). WBGS libraries were deduplicated using deduplicate_bismark (v 0.22.3). For samples prepared by the μWGBS protocol, bisulfite conversion efficiency was specifically considered, and reads with less than 90% bisulfite conversion or fewer than 3 cytosines outside a CpG context were removed.

[0127] Segmentation and clustering analysis. The genome was segmented into blocks of homogeneous methylation using wgbstools (with parameters segment --max_bp 5000) as previously described in Loyfer et al. 2022 (Loyfer et al., 2022; Loyfer & Kaplan). Briefly, a multi-channel Dynamic Programming segmentation algorithm was used to divide the genome into contiguous genomic regions (blocks) that exhibit homogeneous methylation levels across multiple CpGs for each sample. The above segmentation algorithm was applied to 278 human reference WGBS methylomes, retaining 351,395 blocks covered by the hybridization capture panel (80 Mb, capturing approximately 20% of the CpGs) used in the analysis of cfDNA in human serum. Similarly, segmentation of 103 mouse WGBS datasets from healthy cell types and tissues identified 1,344,889 blocks covered by the mouse hybridization capture panel (210 Mb, capturing approximately 75% of the CpGs). The hierarchical relationships between the reference tissue and cell type WGBS datasets were visualized through the creation of a dendrogram. The top 30,000 most divergently methylated blocks containing at least three CpG sites and coverage across 90% of the samples were selected. The average methylation for each block and sample was computed using wgbstools (--beta_to_table). Trees were constructed using the unweighted pair-group method with arithmetic mean (UPGMA) using scipy (V 1.7.1) and L1 distances, and then visualized in R with the ggtree package (V 2.4.1). Similarity between samples was assessed by the variability of distances between samples of the same cell type (average 23,056) compared to the variability of distances between samples of different cell types (average 273,018). Dimensionality reduction was also performed on selected blocks using the UMAP package (V 0.2.8.2.0).Default UMAP parameters were used (15 neighbors, 2 components, Euclidean metric, minimum distance 0.1).

[0128] Identification of cell type-specific methylation blocks. The initial 278 human WGBS samples were reduced to a final set of 104 samples to identify differentially methylated cell type-specific blocks. Samples from bulk tissues and samples that did not have sufficient coverage (missing values ​​in >50% methylation blocks) were excluded. Outlier replicates or clustering with these fibroblast or stromal cell types were excluded due to potential contamination. DMBs were identified using only immune cell methylomes reprocessed from raw sequencing data into PAT files. The final 104 human reference samples were clustered into 20 cell type groups (see Table 4 and Barefoot et al., 2022, Supplementary Table 1). Similarly, the starting 103 mouse WGBS samples were reduced to a final set of 44 samples, which were clustered into 9 cell type and tissue final groups (see Table 5 and Barefoot et al., 2022, Supplementary Table 2). Tissue- and cell type-specific methylation blocks were identified from the final reduced reference WGBS data using custom scripts. A one-to-all comparison was performed to identify differentially methylated blocks unique to each group. This was done separately for human and mouse. First, blocks covering a minimum of 3 CpG sites with a length of less than 2 Kb and at least 10 observations were identified. The average methylation per block / sample was then calculated as the proportion of methylated CpG observations across all reads sequenced from that block. Differential blocks were sorted by a separation margin called "delta beta", defined as the minimal difference between the average methylation in any sample vs. all other samples from the target group. Blocks with delta-beta ≥ 0.4 in human and delta-beta ≥ 0.35 in mouse were then selected. This resulted in a variable number of cell type-specific blocks available for each tissue and cell type. Each DNA fragment was characterized as U (mostly unmethylated), M (mostly methylated), or X (mixed) based on the percentage of methylated CpG sites, as previously described ( Loyfer et al., 2022 ).A threshold of ≦33% methylated CpGs for U reads and ≧66% methylated CpGs for M reads was used. Methylation scores for each identified cell type-specific block were calculated based on the ratio of U / X / M reads among all reads. U ratios were used to define hypomethylated blocks and M ratios were used to define hypermethylated blocks. Human and mouse blocks selected for cell types of interest can be found in Barefoot et al., 2022, Supplementary Tables 3 and 4. Heatmaps were generated using the pretty heatmap function in the RStudio package for R Bioconductor.

[0129] A likelihood-based probabilistic model for fragment-level deconvolution. A probabilistic fragment-level deconvolution algorithm was used to determine the cell type origin of cfDNA. The model was used to calculate the likelihood of each cfDNA molecule using a fourth-order Markov model that considered the combined methylation state of 5 or fewer adjacent CpG sites. Within individual tissue- and cell-type-specific blocks, the model is used to predict whether each molecule is classified as belonging to the tissue of interest or as background. The posterior probability of each cfDNA molecule is calculated based on the log likelihood that the origin of a particular read pair is from the target cell type multiplied by the prior knowledge of the probability that any read should have originated from that target cell type. The model was trained on reference bisulfite sequencing data from normal cells and tissues to learn the distribution of each marker in the target tissue / cell type of interest compared to background. The model was then utilized to test the cfDNA methylome for a binary classification of the origin of each cfDNA molecule. The percentages of molecules assigned to the tissue of interest across all cell type-specific blocks were then summed and used to determine the relative abundance of cfDNA derived from that tissue origin in each sample. The resulting percentages were adjusted to have a sum of 1 by multiplying by a normalization constant. The relative tissue of origin percentages were converted to genome equivalents and reported as an absolute measure that takes into account the initial cfDNA concentration (Geq / mL) [i.e., percentage of cell type-specific cfDNA x initial concentration cfDNAng / mL x 3.3 x 10-12 grams / human haploid genome equivalent (or x 3.0 x 10-12 grams / mouse haploid genome equivalent)].

[0130] In-silico simulated WGBS deconvolution. In-silico mix-in simulations were performed to validate the fragment-level deconvolution algorithm on the identified cell-type-specific blocks contained in the radiation-specific methylation map (Figures 3 and 4). Reference data with more than three replicates per cell type were split into independent training and test sets, with at least one replicate left out for testing. Mouse cardiomyocyte reference WGBS data had less than three replicates, so fragments were integrated across replicates for this cell type and split into a training set (80%) and a test set (20%). For each cell type profiled, a known proportion of target fragments was mixed into a background of leukocyte fragments (leukocyte fragments obtained from n=4 buffy coat samples in mouse and n=10 buffy coat samples in human) spanning the identified cell-type-specific methylation blocks. Ten replicates were performed for each mixture ratio evaluated (0.001, 0.005, 0.01, 0.02, 0.05, 0.1, 0.15) and the average predicted percentage and standard deviation across replicates are presented. Model accuracy was assessed by correct classification of the actual percent target mixture, and the relative degree of evolution with increasing amounts of mixed target reads was used to assess accuracy in estimating proportional changes across groups (mice) and time points (humans) from consecutive samples. Cell type-specific blocks included in the radiation-specific methylation map were constructed using only the training set fragments. Read merging, splitting, and mixing were performed using wgbstools (Loyfer & Kaplan).

[0131] Longitudinal Analysis of Serial Serum Samples. Longitudinal analysis was performed on serial serum samples collected from breast cancer patients. Changes in cell type percentages in cfDNA at the end of treatment (EOT) and recovery were evaluated compared to baseline levels (Baseline) before the start of therapy. Fold change (FC) from baseline was used to express percent cell type cfDNA at EOT and recovery compared to baseline in the same individuals. Exploratory correlation analysis was performed to evaluate the linear relationship of the change in cell type percentages obtained from EOT to baseline using Pearson's correlation coefficient.

[0132] Functional annotation and pathway analysis. The identified cell type-specific methylation blocks were provided as input for analysis in HOMER (http: / / homer.ucsd.edu / homer / ). Each block was associated with its nearest neighboring genes to provide genomic annotation. By default, TSS (transcription start site) was defined from -1kb to +100bp, TTS (transcription termination site) was defined from -100bp to +1kb, and CpG islands were defined as genomic segments with GC content ≥ 50%, genomic length > 200bp, and ratio of observed / expected CpG number > 0.6. Predictions of known and novel transcription factor binding motifs were also assessed by HOMER. The top 5 motifs based on p-value were selected from each analysis. Pathway analysis of the identified tissue- and cell type-specific methylation blocks was performed using Ingenuity Pathway Analysis (IPA) and Genomic Regions Enrichment of Annotations Tool (GREAT) (McLean et al., 2010). GeneSetCluster was used to cluster the identified gene set pathways based on shared genes (Ewing et al., 2020). The identified gene set clusters were interpreted and functionally labeled by collapsing all identified significant gene set pathways to the top representative one using the WebgestaltR (ORAperGeneSet) plugin. Integration of methylome and transcriptome data generated from tissue-specific endothelial cells was performed using an expanded set of cell type-specific blocks (--bg.quant 0.2) compared to the more limited set of blocks used for the deconvolution analysis in circulation (--bg.quant 0.1). The extended endothelial-specific methylation blocks can be found in Barefoot et al., 2022, Supplementary Table 10.

[0133] Cluster analysis and visualization techniques. Hierarchical relationships between the reference tissue and cell type WGBS datasets were visualized by constructing a dendrogram. The top 30,000 most divergently methylated blocks containing at least 3 CpG sites and coverage across 90% of the samples were selected. The average methylation for each block and sample was computed using wgbstools (--beta_to_table). Trees were constructed using the unweighted pair-group method with arithmetic mean (UPGMA) and visualized in R with the ggtree package. Dimensionality reduction was also performed on the selected blocks using the UMAP algorithm with default UMAP parameters (15 neighbors, 2 components, Euclidean metric, minimum distance 0.1). Heatmaps were generated using the prettyheatmap function in the RStudio package in R bioconductor (RStudioTeam, 2015). Statistical analysis for group comparisons and correlations was performed using Prism and R. Sequencing reads were visualized using the Integrative Genomics Viewer (IGV) (Robinson et al., 2011) using the bisulfite CG mode for alignment coloring. The BEDTools suite and AWK programming were used to overlay sequencing data across samples and to perform comparisons across sample groups and replicates. Python was used to operate the WGBS tool and also generate visualization plots.

[0134] result DNA methylation is highly cell type specific and reflects cell lineage specification. We gained access to reference human and mouse WGBS datasets from publicly available databases and identified cell type-specific differential DNA methylation patterns preferentially from primary cells isolated from healthy human and mouse tissues. Additionally, we created cell type-specific methylomes for purified mouse immune cell types (CD19+ B cells, Gr1+ neutrophils, CD4+ T cells, and CD8+ T cells) as well as human tissue-specific endothelial cell types (coronary artery, pulmonary artery, cardiac microvessels, pulmonary microvessels, and liver sinusoidal endothelium). Due to limited cell type-specific data available for mice, we included reference data from mouse bulk tissues when none was available from purified cell types in those tissues. This resulted in the curation of methylation data from 10 different cell types and 18 tissues for mice and over 30 distinct cell types for humans (Tables 4 and 5; see also Barefoot et al., 2022, Supplementary Table 10).

[0135] To better understand the epigenomic landscape of cell types in these healthy human and mouse tissues, we characterized the methylome by first segmenting the above data into uniformly methylated blocks, where DNA methylation status at adjacent CpG sites is highly co-regulated by methylation enzyme processivity (Loyfer et al., 2022). Exploration of epigenetic diversity between cell types at the block level increased the robustness of downstream analyses and proved more resistant to noise induced as a by-product of bisulfite sequencing. We applied segmentation to human WGBS datasets from 275 publicly available purified cell types and identified 351,395 blocks contained in probes used for hybridization capture sequencing to enrich for cfDNA in human serum (Table 4). Segmentation of WGBS datasets from normal cell types and tissues in 83 mice identified 1,344,889 blocks contained in mouse hybridization capture probes (Table 5). On average, each block was over 300 bp with 4–8 CpG sites per block. Unsupervised hierarchical clustering analysis of the top 30,000 most divergently methylated blocks in human and mouse shows the relationships between samples as dendrograms and UMAP predictions, respectively (Figures 5 and 6). The closely correlated relationships between methylomes of the same cell type observed from the cluster analysis reinforces the notion that methylation status is conserved in regions crucial for cell type identification. Diversity within cell types is significantly reduced compared to diversity between cell types. This stability allows methylated DNA to serve as a robust biomarker that can be generated across diverse patient populations despite patient heterogeneity. In most cases, cell types that constitute distinct lineages, such as immune, epithelial, muscle, neuronal, endothelial, and stromal cell types, remain closely related. Examples include tissue-specific endothelial cells and tissue-resident immune cells that cluster with endothelial cells or immune cells, respectively, regardless of the germ layer origin of their resident tissue.Similarly, some cell types cluster separately from their bulk tissue counterparts. For example, cardiomyocytes cluster separately from cardiac tissue in the mouse dendrogram, indicating a heterogeneous composition of different cell types contributing to the organ and distinct embryonic origins (Figure 6, Panel A). Surprisingly, large epigenetic distances were observed between immune cells of hematopoietic origin and solid organ cells from other lineages (Figure 5, Panels A and B). This is important for tissue of origin analysis of circulating cfDNA to distinguish DNA of solid organ origin from DNA of hematopoietic origin. Quite unexpectedly, we also found numerous epigenetic signatures that could distinguish between immune cells, clustering separately from lymphoid and myeloid cell types. Within the immune cell cohort, we observed increased separation of terminally differentiated cells compared to precursors, with naïve B and T cells clustering separately from their more mature central and effector memory counterparts (Figure 5, Panel B). Collectively, these findings confirm that DNA methylation is highly cell type specific and reflects cell lineage specification.

[0136] Differential DNA methylation distinguishes between cell types in healthy human and mouse tissues. Based on the unsupervised clustering analysis described above, the inclusion / exclusion criteria were further refined to select the final set of reference methylomes used to identify differentially methylated cell type-specific blocks. Low coverage WGBS samples were excluded from bulk tissues. Similarly, samples that did not cluster with other replicates were excluded from the same cell type and instead clustered with fibroblasts and other stromal cell types. This resulted in a reduction of the starting 278 human WGBS samples to a final set of 104 samples, which were clustered together in groups of 20 cell types. Similarly, the starting 103 mouse WGBS samples were reduced to a final set of 44 samples, which were clustered together in final groups of 9 cell types and tissues. Several related subsets of cell types were considered together to form final groups (i.e., monocytes clustered together with macrophages and colon clustered together with small intestine). The final combination of the above groups was found to best represent the entire cell-specific epigenetic diversity without overlaps using this publicly available data. Cell type-specific differentially methylated blocks (DMBs) that contained a minimum of three CpG sites were identified. The co-methylation status of adjacent CpG sites in these blocks was able to distinguish between all cell types included in the final group. We identified 4,502 human DMBs and 7,344 mouse DMBs (Barefoot et al., 2022, see Supplementary Tables 3 and 4) with lower separation margins for mouse (0.35) vs. human (0.40) due to more limited data. A complete overview of the identified human and mouse cell type-specific methylation blocks can be found in Tables 6 and 7. A variable number of blocks was required to obtain the same specificity for each cell type based on the depth of coverage, purity, and degree of separation from other tissues and cell types included in the map. Similar to others, we found enhanced separation of reference datasets using methylomes of purified cell types as opposed to more heterogeneous mixtures from bulk tissues ( Moss et al., 2018 ).This is evident from the low number of DMBs identified from mouse bulk tissues compared to mouse purified cell types (average 310 DMBs / tissue vs. 1,488 DMBs / cell type). Although over 85% of cell type-specific DMBs are hypomethylated, the blocks were represented as a heatmap with a methylation score that is independent of the directionality of methylation status and highlights the degree of separation of both hypomethylated and hypermethylated blocks in the targeted group compared to all other groups. The methylation score is calculated by dividing the number of read pairs that are completely unmethylated or methylated by the total coverage for the hypomethylated and hypermethylated blocks, respectively. The heatmap in Figure 7 represents the 100 highest methylation score blocks for each cell type group.

[0137] Differential DNA methylation is closely linked to regulating cell type-specific functions. The role of cell type-specific methylation in shaping cell identity and cell function was investigated. Genes adjacent to cell type-specific methylation blocks were identified using HOMER, and pathway analysis of annotated genes was performed using both Ingenuity Pathway Analysis (IPA) and GREAT. GeneSetCluster was used to group significantly enriched pathways based on shared genes, and each cluster was WebgestaltR functionally labeled by its top-defined biological process (Figure 7, Panel C; and Figure 8). Gene set pathways largely clustered in independent cell type groups, reinforcing that cell-specific differential methylation occurs adjacent to unique genes essential for cell type-specific functions. Taken together, cell type-specific methylation was preferentially located adjacent to genes with biological functions involved in cell development, migration, proliferation, differentiation, and morphology. Furthermore, transcriptional apparatus genes, including transcription factors and coregulators, were significantly associated with cell type-specific DNA methylation, particularly those involved in the assembly of the RNA polymerase III complex and the pre-mRNA degradation process (see Table 11). However, despite these commonalities, important biological differences were also observed in the gene sets identified based on specific processes unique to the profiled cell types. For example, the biological functions of genes associated with immune cell type-specific methylation reflect the processes of leukocyte cell-cell adhesion, immune response regulation signaling, and hematopoietic system development (Figure 7, Panel C). In contrast, for hepatocytes, we identified fatty acid metabolic pathways, lipid metabolism, and acute phase response signaling. These findings suggest that cell type-specific methylation is involved in the regulation of these cellular processes. The biological pathways and functions significantly enriched for genes associated with differential methylation in each cell type investigated are provided in Table 11.

[0138] Cell type-specific DNA methylation is largely hypomethylated and enriched in intragenic regions containing developmental TF binding motifs. Most identified human and mouse cell type-specific blocks were hypomethylated, consistent with proposed mechanisms of methylation resetting during embryonic development that result in highly regulated cell type-specific differences (Greenberg & Bourc'his, 2019; Dor & Cedar, 2018). In human samples, 86% of cell type-specific DMBs were found to be hypomethylated and only 14% were hypermethylated. Notably, in mouse samples, 98% of cell type-specific DMBs were hypomethylated and only 2% were hypermethylated. In the schematic diagram in Figure 9, panel A represents the location of the identified human cell type-specific hypomethylated and hypermethylated blocks. Interestingly, regardless of orientation, the majority of cell type-specific blocks were located in intragenic regions. To confirm whether this distribution was enriched, the genomic loci of the cell type-specific blocks were compared to blocks that did not vary between cell types (Figure 9, panels B and C; Table 8). It was found that there was a significant enrichment of cell type-specific blocks in intragenic regions compared to other captured regions for both human and mouse (P<0.05). Furthermore, the intragenic distribution of cell type-specific blocks showed a significant increase in locations in exons and a decrease in promoter-TSS segments (P<0.05). There was also a significant relationship between directionality and intragenic distribution, with most cell type-specific blocks being hypermethylated in exons and hypomethylated in introns (P<0.05). The similar distribution of cell type-specific methylation blocks in human and mouse suggests a conserved biological function of these genomic regions across species.

[0139] To further explore what common purpose these identified regions may have in human and mouse development, we performed motif analysis using HOMER to determine whether there were any commonly enriched transcription factor binding sites (TFBS). MADS motifs bound by MEF2 transcription factor were significantly enriched in both human and mouse cell type-specific hypomethylated blocks (Figure 9, Panel D, left). MEF2 transcription factor is a well-established developmental regulator with a role in the differentiation of many cell types derived from different lineages. In comparison, homeobox motifs bound by several different HOX TFs were enriched in human cell type-specific hypermethylated blocks (Figure 9, Panel D, right). Notably, HOXB13 was the top TF associated with binding at sites in human hypermethylated DMBs. Recently, HOXB13 has been found to control cell state through binding to super-enhancer regions, suggesting a novel regulatory function for cell type-specific hypermethylation. In addition to the common TFBS enriched by all cell type-specific blocks, endothelial-specific TFs were found to be enriched in endothelial cell hypomethylated blocks, such as EWS, ERG, Fli1, ETV2 / 4, and SOX6 (see FIG. 10, panel D). Overall, this data reveals unknown functions of these cell type-specific blocks that represent cell-specific biology.

[0140] Methylation profiling of tissue-specific endothelial cell types reveals epigenetic heterogeneity associated with differential gene expression. Radiation-induced endothelial damage is a major complication of radiotherapy and is believed to be the main cause of the development of late-onset cardiovascular disease (Tapio, 2016; Wagner & Dimmeler, 2019). The microvasculature is particularly sensitive to radiation, and dysfunction of these cells may contribute to injury in various tissues (Wijerathne et al., 2021; Park et al., 2012). To this end, we generated tissue-specific endothelial methylomes and paired transcriptomes to profile injury originating from distinct populations of microvascular and macrovascular endothelial cell types, e.g., coronary arteries, pulmonary arteries, cardiac microvessels, pulmonary microvessels, and liver sinusoidal endothelium. Similarly, we also utilized the publicly available umbilical vein endothelial methylome obtained from the Blueprint Epigenome Consortium to complete our data (Table 4; see also Barefoot et al., 2022, Supplementary Table 1). Previous studies support modeling the heart and lungs as an integrated system in the development of radiation injury, as they are coupled by the cardiopulmonary circulation (Barazzuol et al., 2020). Thus, we integrated cardiac and pulmonary endothelial cell types together to generate an integrated cardiopulmonary endothelial signal to identify specific methylation blocks for cardiopulmonary (CPEC, n=132), liver sinusoidal endothelial (LSEC, n=89), and umbilical vein endothelial (HUVEC, n=116) cell types. Pathway analysis of genes associated with these methylation blocks confirmed endothelial cell identity and revealed genes involved in regulating vasculogenesis, angiogenesis, and vascular development (Figure 10, Panel B). Additionally, we identified unique pathways that capture the tissue-specific epigenetic diversity of these distinct endothelial cell populations. For example, liver fibrosis signaling was found to be LSEC-specific, cardiac hypertrophy signaling was identified to be CPEC-specific, and thioredoxin pathway activity was specific to HUVEC (Figure 10, Panel A). The identity of the starting material used to create these human endothelial methylomes was verified by paired RNA sequencing analysis. Integrated analysis of DNA methylation and paired RNA expression allowed for a better understanding of the relationship between cell type-specific DNA methylation and corresponding gene expression changes. Methylation status in several identified blocks was found to correspond to RNA expression of known endothelial-specific genes, corroborating the identity of the isolated LSEC and CPEC populations (Figure 10, panels C and E; Barefoot et al., 2022, Supplementary Table 10). For example, hypomethylation was associated with increased expression in several pan-endothelial genes, such as NOTCH1, ACVRL1, FLT1, MMRN2, NOS3, and SOX7. Similarly, hypomethylation in CPEC- and LSEC-specific genes resulted in differential expression when comparing the two populations, reflecting tissue-specific differences. CPEC- and LSEC-specific expression of selected genes has been reported in previous studies investigating vascular heterogeneity at the transcriptome level (Feng et al., 2019; Sabbagh et al., 2018; Nolan et al., 2013; Cleuren et al., 2019). However, correlating these expression patterns with cell type-specific methylation is a novel property. Although the majority of endothelial-specific methylation blocks were hypomethylated, select hypermethylated blocks were also identified, such as CCM2L in CPECs, which corresponded to reduced gene expression compared to LSECs. Due to the relative abundance of cell types in the circulation, the ability to noninvasively detect distinguishable damage to various types of endothelial cell populations could prove useful for monitoring tissue-specific damage.

[0141] Development of radiation-specific methylation maps focused on cell types derived from target organs at risk (OARs). After ensuring the specificity of the identified cell type-specific methylation blocks by comparison with all other cell types using available WGBS data, the evaluation of circulating cfDNA origin was limited to select cell types derived from target organs at risk for radiation damage. Limiting to a focused radiation-specific methylation map helped to maintain the sensitivity of radiation-induced damage to cell types of interest based on prior knowledge of the target, damaged organs with existing clinical correlations. A representative treatment plan for a breast cancer patient receiving adjuvant radiation provides the estimated organ volumes affected and radiation dose levels for target risk organs for radiation damage, such as the heart and lungs (Figure 11, Panel A). In addition to organs in close proximity to the target treatment area, the liver is another organ that may receive a substantial dose from radiation, especially in right-sided breast cancer patients. Differential blocks identified from cell types including these radiation target risk organs (lung, heart, and liver) were selected for the creation of radiation-specific methylation maps that distinguish these solid organ cell types of interest from all other immune cell types (Figure 11, Panel B; Figure 6, Panel B). These cell type specific human and mouse blocks can be found in Barefoot et al., 2022, Supplementary Tables 3 and 4. Due to the high degree of separation of the epigenetic signatures of hematopoietic cells from other solid organ cell lineages, all hematopoietic cell types were combined into one unified "immune" supergroup. This method also accounts for the hematopoietic origin of the majority of cfDNA at baseline and helps to reveal signals derived from solid organ cell types of interest. Focusing on these same target organs in both humans and mice, based on available reference cell type data, gave rise to a final curation of six groups for humans (immune, pulmonary epithelial, cardiopulmonary endothelial, cardiomyocytes, hepatocytes, and liver sinusoidal endothelial) and four groups for mice (immune cells, pulmonary endothelial cells, cardiomyocytes, hepatocytes).

[0142] Cell-free methylated DNA in blood identifies the origin of radiation-induced cellular damage in tissues. Serial serum samples were collected from breast cancer patients undergoing standard radiotherapy. Additionally, paired serum and tissue samples were collected from irradiated mice. Unbiased methylome-wide hybridization capture sequencing of DNA from human or mouse serum samples was performed. Deconvolution analysis was used to trace the origin of cfDNA fragments, allowing minimally invasive monitoring of radiation-induced cytotoxicity from blood samples (Figure 2). Compared to previous studies using single CpG sites, sequencing-based methods allow fragment-level cfDNA analysis with CpG methylation patterns (Scott et al., 2020; Li et al., 2018). To this end, we modeled the co-methylation status of adjacent CpG sites on the same molecule, which was performed by a novel probabilistic deconvolution method. The model was applied using cell type-specific blocks obtained from the human and mouse radiation-specific methylation maps described above. The predictive accuracy of fragment-level deconvolution was validated by in-silico mix-in simulations for each tissue and cell type of interest (Figures 3 and 4).

[0143] Dose-dependent indicators of radiation damage in mice. To explore the relationship between radiation-induced damage in tissues and changes in the percentage of circulating cfDNA origin, mice were used to model exposure from different radiation doses. Mice receiving upper thoracic radiation at 3 Gy or 8 Gy doses compared to sham controls formed three groups for comparison (Figure 2). Tissues and serum were harvested 24 hours after the last part of treatment, and tissues along the path of the radiation beam (heart, lung, and liver) were targeted for subsequent analysis. Through histological analysis, dysregulated tissue architecture corresponding to higher doses of radiation was observed (Figure 12, Panel A). These changes were most evident in lung tissue sections showing prominent alveolar collapse with increasing radiation doses. Liver tissue showed increased fibrosis with increasing radiation doses, and only minor changes were seen in heart tissue, consistent with its resilience to higher radiation. Tissue effects were also assessed by qPCR analysis of established indicators of radiation effect, such as expression of CDKN1A(p21), which showed a dose-dependent increase in expression in response to radiation in all tissues (Figure 12, panel B; Figure 13) (Hyduke et al., 2013).

[0144] Data obtained from capture sequencing of methylated cfDNA were analyzed to evaluate indicators of cardiac, pulmonary, and hepatic injury in serum samples (Figure 2). For analysis, mouse cardiomyocyte (n=2,917), pulmonary endothelial (n=1546), hepatocyte (n=616), and immune (n=148) cell type-specific methylation blocks derived from radiation maps for target risk organs as described above were used. Combining signals from mice treated with 3 Gy and 8 Gy, we found significant increases in percent pulmonary endothelial, cardiomyocyte, and hepatocyte cfDNA in radiation-treated groups compared to sham controls, which correlated with apoptotic cell death in the corresponding tissues. Furthermore, we observed significant dose-dependent increases in percent pulmonary endothelial, cardiomyocyte, and combined solid organ cfDNA across all three treatment groups, which correlated with radiation-induced cell death in the corresponding tissues (P<0.05, Kruskal-Wallis test) (Figure 12, panels C and D; Figure 14, panel E). However, there was no dose-dependent increase in hepatocyte or immune cfDNA (Figure 12, panel E; Figure 14, panel D). As proof of principle, this confirms that methylated DNA in blood may represent a source of radiation-induced cellular damage in tissues.

[0145] Radiation treatment of patients with breast cancer. To evaluate whether changes in cfDNA patterns could indicate damage to tissues in patients after radiation, serum samples were collected from breast cancer patients at three time points during their post-surgery standard of care radiation therapy (Figure 2). A baseline sample was collected for each patient before the start of radiation therapy, and a second end-of-treatment (EOT) sample was collected 30 minutes after the last treatment, after a total of 20-30 treatments. Finally, a recovery sample was collected 1 month after the completion of radiation therapy. The demographic information and clinical characteristics of the patients enrolled in this study are listed in Table 3 and in Barefoot et al., 2022, Supplementary Table 8. For the analysis of cfDNA, we focused on the cell types that make up the heart, lung, and liver tissues.

[0146] Radiation-induced liver injury. Although liver injury is not a common radiation-induced toxicity experienced by breast cancer patients, substantial doses can still be administered to the liver, especially right-sided tumors (Figure 11, Panel A). The top hepatocyte methylation block (n=200) and liver sinusoidal endothelial methylation block (n=89) were used to evaluate the sequence data for the presence of liver-derived cfDNA. Surprisingly, in patients undergoing radiation treatment for right-sided breast cancer, increases in circulating hepatocyte methylated DNA and liver sinusoidal endothelial methylated DNA indicated significant radiation-induced cell damage in the liver (P<0.05, Wilcoxon matched-pairs signed-rank test) (Figure 15, Panels A-F). High levels of either hepatocyte cfDNA and / or liver sinusoidal endothelial cfDNA were detected in seven of eight breast cancer patients with right-sided tumors. In contrast, in patients with left-sided breast cancer, there was no significant increase in hepatocyte cfDNA or liver sinusoidal endothelial cfDNA.

[0147] Radiation-induced cardiac and pulmonary injury. The heart and lungs are common risk organs for breast cancer patients undergoing radiation therapy due to their close proximity to the target treatment area. To evaluate radiation-induced lung injury, serum-derived cfDNA was examined for the presence of lung epithelial methylated DNA blocks (n=69). Interestingly, no significant increase in lung epithelial cfDNA across all patients was observed (P≧0.05, Friedman test) (FIG. 16, Panel A). However, several patients showed an increase in lung epithelial cfDNA, indicating lung injury that correlated with increasing dose and targeted lung volume (FIG. 16, Panel B). Notably, long-term changes in lung epithelial cfDNA after radiation were found to correlate with the volume of the ipsilateral lung receiving a 20 Gy dose (Lung V20) (Pearson's r=0.67, P<0.05) and the mean whole-body dose (Pearson's r=0.90, P<0.05). In addition to lung injury, cardiovascular disease is one of the most severe complications from radiation exposure, with increased morbidity and mortality (White & Joiner, 2006; Brownlee et al., 2018). Through deconvolution with cardiopulmonary endothelial (CPEC, n=132) and cardiomyocyte-specific (n=375) DNA methylation blocks, we found increased CPEC cfDNA and cardiomyocyte cfDNA in serum samples, indicating significant cardiovascular cell injury across all breast cancer patients (P<0.05, Friedman test) (Figure 16, Panels D and G). Strikingly, circulating cardiomyocyte-specific methylated DNA correlated with maximum radiation dose to the heart (Pearson's r=0.63, P<0.05), but not with mean dose to the heart (Pearson's r=0.09, P≧0.05) (Figure 16, Panel H). This suggests that cardiomyocyte sensitivity to radiation-induced damage required a sufficiently high dose to enhance the resilience of this cell type to radiation damage compared with corresponding epithelial and endothelial cell types from the heart and lung.

[0148] Distinct endothelial and epithelial damage from radiation. Distinct epithelial and endothelial cell type responses to radiation were observed across the different tissues profiled. Differential responses to radiation were observed when comparing hepatocellular damage to lung epithelial damage (Figure 15, Panels A-C vs. Figure 16, Panels A-C), demonstrating the ability of methylated DNA from serum samples to distinguish between tissue-specific epithelial cell types. Similarly, analysis of tissue-specific endothelial populations reveals differences in cardiopulmonary microvascular and hepatic sinusoidal endothelial responses to radiation (Figure 15, Panels D-F vs. Figure 16, Panels D-F). In general, there was a greater magnitude of damage to the endothelium compared to the epithelium in the different organs. The endothelium forms a layer of cells that lines blood vessels as well as lymphatic vessels. As a result, it is likely that turnover from this cell type could contribute to the high amplitude signal detected from serum (Moss et al., 2018). However, this could also be a result of the differential susceptibility of endothelial versus epithelial cell types to radiation-induced damage. There was a 5-fold higher signal from CPEC cfDNA compared to lung epithelial cfDNA. Similarly, there was a 2-fold increase in LSEC cfDNA compared to hepatocytes in right-sided cases. Similarly, higher cardiomyocyte cfDNA indicates sustained injury and delayed recovery compared to epithelial- and endothelial-derived cfDNA (Figure 16, panels C, F, and I). This may reflect important differences in cell turnover rates, resulting in different regeneration and repair processes in these cell types. Notably, the increased turnover of endothelial and cardiomyocyte cells indicates long-lasting tissue remodeling, but one month after the completion of radiotherapy, the endothelial injury signature detected from cfDNA returned to baseline levels. Overall, these findings demonstrate the applicability of this method to discover distinct cellular injury in different tissues during the course of treatment in a minimally invasive manner.

[0149] Comparison of results in humans and mice. Comparing cfDNA origins after radiation, similar radiation-related changes were observed in both human and mouse serum samples. There was a significant increase in lung endothelial cfDNA and cardiomyocyte cfDNA after radiation in both humans and mice. Similarly, there was an overall increase in cfDNA derived from any solid organ tissue after radiation in both irradiated breast cancer patients and mice (Figure 14). Total cfDNA concentrations were also elevated at EOT in some breast cancer patients, suggesting an overall increase in cfDNA immediately after radiation treatment (Table 9). Changes in mouse cfDNA concentrations with increasing radiation dose were not significant (Table 13), as also reported in previous studies78,79.

[0150] This study demonstrated the ability of tissue-of-origin analysis of cell-free methylated DNA to monitor the whole-body response to radiation therapy. The assignment of DNA fragments extracted from serum samples from treated patients as well as from experimental animals to specific cell types required an in-depth analysis of tissue and cell type methylation patterns. It was surprising that there was a significant association between cell type-specific DNA methylation blocks and cell type-specific gene expression, transcription factor binding motifs and signaling pathway regulation. This study led to the development of a methylation map that included cell type-specific methylation patterns from target risk organs of radiation injury, such as the heart, lungs and liver. Methylated DNA in blood samples was found to be an indicator of radiation damage that could be useful for predicting patients who are more likely to develop severe adverse effects.

[0151] [Table 3]

[0152] [Table 4] TIFF2024529192000033.tif215168

[0153]

Table 5

[0154]

Table 6

[0155]

Table 7

[0156]

Table 8

[0157]

Table 9

[0158]

Table 10

[0159]

Table 11

[0160] References TIFF2024529192000066.tif210162TIFF2024529192000067.tif201161

Claims

1. A method for determining whether a subject is suffering from tissue damage resulting from exposure to a toxic agent, the method comprising: (a) sequencing cell-free DNA (cfDNA) in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cfDNA is determined when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) measuring the amount of cfDNA of the determined cellular origin; and (d) comparing the measured amount of cfDNA of the determined cellular origin with either (i) the normal amount of cfDNA of the determined cellular origin or (ii) the amount of cfDNA of the determined cellular origin measured at a previous time point wherein an increase in the measured amount of cfDNA of the determined cellular origin that exceeds the normal amount of cfDNA of the determined cellular origin or the amount of cfDNA of the determined cellular origin measured at a previous time point indicates that the subject is suffering from or is at risk of suffering from tissue damage resulting from the exposure. The method.

2. The toxic agent is (i) including radiation, (ii) including microorganisms, (iii) derived from a chemical source or a biological source, (iv) including drug therapy, or (v) including a chemical substance or a biological substance or a radioactive substance used as a weapon, The method according to claim 1.

3. A composition comprising a substance for use in the treatment of tissue damage in a subject caused by exposure to a toxic agent, wherein the subject has undergone: (a) sequencing cell-free DNA (cfDNA) in a biological sample from the subject; (b) determining the cellular origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that contain methylation sites, wherein the cellular origin of the cfDNA is determined when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) measuring the amount of cfDNA of the determined cellular origin; and (d) comparing the measured amount of cfDNA of the determined cellular origin with either (i) the normal amount of cfDNA of the determined cellular origin or (ii) the amount of cfDNA of the determined cellular origin measured at a previous time point (d) comparing the measured amount of cfDNA of the determined cell origin with either (i) the normal amount of cfDNA of the determined cell origin or (ii) the amount of cfDNA of the determined cell origin measured at a previous time point determined to be suffering from tissue damage by a method comprising wherein an increase in the measured amount of cfDNA of the determined cell origin that exceeds the normal amount of cfDNA of the determined cell origin or exceeds the amount of cfDNA of the determined cell origin measured at a previous time point indicates that the subject is suffering from tissue damage the composition [

4. ] wherein the toxic factor is (i) including radiation (ii) including microorganisms (iii) derived from a chemical source or a biological source (iv) including drug therapy, or (v) including a chemical substance or a biological substance or a radioactive substance used as a weapon The composition according to claim 3. [

5. ] A composition comprising a substance for use in a method of treating tissue damage in a subject, the method comprising a step of monitoring tissue damage, wherein the monitoring is as follows: (a) sequencing cell-free DNA (cfDNA) in a biological sample from the subject; (b) determining the cell origin of the cfDNA by identifying a methylation pattern in one or more portions of the sequence of the cfDNA that includes methylation sites, wherein the cell origin of the cfDNA is determined when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) measuring the amount of cfDNA of the determined cell origin, and (d) comparing the measured amount of cfDNA of the determined cell origin with either (i) the normal amount of cfDNA of the determined cell origin or (ii) the amount of cfDNA of the determined cell origin measured at a previous time point comprising A decrease in the measured amount of cfDNA of the determined cell origin, compared to the normal amount of cfDNA of the determined cell origin or measured at a previous time point, indicates that the treatment is effective, and an increase or no change in the measured amount of cfDNA of the determined cell origin, which is above the normal amount of cfDNA of the determined cell origin or above the amount of cfDNA of the determined cell origin measured at a previous time point, indicates that the treatment is not effective. Said composition. Claim 6 The composition according to claim 5, wherein the tissue damage is caused by exposure to a toxic factor. Claim 7 The toxic factor is (i) including radiation, (ii) including microorganisms, (iii) derived from a chemical source or a biological source, (iv) including drug therapy, or (v) including chemical, biological or radioactive substances used as weapons, The composition according to claim 6. Claim 8 The method further includes the step of adjusting the treatment administered to the subject when it is shown that the treatment is not effective, the composition according to claim 5. Claim 9 The normal amount of the cfDNA includes the amount of cfDNA for the determined cell origin generated in a population of individuals not exposed to the toxic factor, the composition according to claim 5. Claim 10 A composition comprising a chemotherapeutic agent, radiation therapy, targeted therapy, immunotherapy or a combination thereof for use in a method of treating a subject in need thereof, the method including the step of monitoring whether the chemotherapeutic agent, radiation therapy, targeted therapy, immunotherapy or a combination thereof causes tissue damage in the subject, wherein the monitoring is as follows: (a) sequencing cell-free DNA (cfDNA) in a biological sample from the subject; (b) determining the cell origin of the cfDNA by identifying the methylation pattern in one or more portions of the sequence of the cfDNA containing methylation sites, the step being such that when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern, the cell origin of the cfDNA is determined; (c) measuring the amount of cfDNA of the determined cell origin, and (d) Comparing the measured amount of cfDNA of the determined cell origin with either (i) the normal amount of cfDNA of the determined cell origin or (ii) the amount of cfDNA of the determined cell origin measured at a previous time point comprising wherein an increase in the measured amount of cfDNA of the determined cell origin that exceeds the normal amount of cfDNA of the determined cell origin or exceeds the amount of cfDNA of the determined cell origin measured at a previous time point indicates that a chemotherapeutic agent, radiation therapy, targeted therapy, immunotherapy, or a combination thereof causes tissue damage the composition **Claim 11** The composition according to claim 10, further comprising the step of adjusting the treatment administered to the subject when it is indicated that the treatment is not effective or causes tissue damage **Claim 12** The composition according to claim 10, wherein the normal amount of the cfDNA comprises the amount of cfDNA for the determined cell origin generated in a population of individuals not administered a chemotherapeutic agent, radiation therapy, targeted therapy, immunotherapy, or a combination thereof **Claim 13** A composition comprising a chemotherapeutic agent, radiation therapy, targeted therapy, immunotherapy, or a combination thereof for use in a method of treating a subject having a tumor, the method comprising (A) Monitoring the response, adverse reaction, or combination thereof to a chemotherapeutic agent, radiation therapy, targeted therapy, immunotherapy, or a combination thereof, wherein the monitoring comprises the following (i) Determining whether there is an adverse reaction to a chemotherapeutic agent, radiation therapy, targeted therapy, immunotherapy, or a combination thereof, the step comprising (a) Sequencing circulating tumor DNA (cfDNA) in a biological sample from the subject (b) Determining the cell origin of the cfDNA by identifying the methylation pattern in one or more portions of the sequence of the cfDNA comprising methylation sites, wherein the cell origin of the cfDNA is determined when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern (c) Measuring the amount of cfDNA of the determined cell origin, and (d) A step of comparing the measured amount of cfDNA of the determined cell origin with the normal amount of cfDNA of the determined cell origin, wherein an increase in the measured amount of cfDNA of the determined cell origin exceeding the normal amount of cfDNA of the determined cell origin indicates a harmful reaction, said step; Said step including (ii) A step of determining whether there is a response to a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof, comprising: (a) A step of sequencing circulating tumor DNA (ctDNA) in a biological sample from a subject, (b) A step of determining the clonal heterogeneity of the cells of the tumor by genotyping said ctDNA, wherein the presence of two or more clones of said tumor cells or the presence of a tumor cell clone not previously identified in the subject indicates an ineffective response to a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof, said step; Said step including, and (B) A step of continuing the administration of a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof when it is determined that there is no harmful reaction, no ineffective response, or a combination thereof; or a step of administering an adjusted chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof when it is determined that there is a harmful reaction, an ineffective response, or a combination thereof, either Said composition including.

14. The composition according to claim 13, wherein the normal amount of said cfDNA includes the amount of cfDNA for the determined cell origin generated in a population of individuals who have no tumor or have not received a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof.

15. A composition comprising a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof for use in a method of treating a subject having a tumor, said method comprising (A) A step of monitoring a response, harmful reaction, or a combination thereof to a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof, said monitoring being at two or more time points, the following (i) A step of determining whether there is a harmful reaction to a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof, the following (a) Sequencing cell-free (cfDNA) in a biological sample derived from a subject; (b) Determining the cellular origin of the cfDNA by identifying the methylation pattern in one or more portions of the sequence of the cfDNA containing methylation sites, wherein the cellular origin of the cfDNA is determined when the methylation pattern in the one or more portions is the same as a known cell type-specific methylation pattern; (c) Measuring the amount of cfDNA of the determined cellular origin, wherein an increase in the measured amount of cfDNA of the determined cellular origin at a later time point compared to an earlier time point indicates a harmful reaction; including the above steps; and (ii) Determining whether there is a response to a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof, comprising: (a) Sequencing circulating tumor (ctDNA) in a biological sample derived from a subject; (b) Determining the clonal heterogeneity of tumor cells by genotyping the ctDNA, wherein the presence of two or more clones of the tumor cells or the presence of tumor cell clones at a later time point not identified at an earlier time point indicates a non-effective response to a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof; including the above steps; including the above steps, and (B) Continuing the administration of a chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof if it is determined that there is no harmful reaction, no non-effective response, or a combination thereof; or administering an adjusted chemotherapeutic agent, radiotherapy, targeted therapy, immunotherapy or a combination thereof if it is determined that there is a harmful reaction, a non-effective response, or a combination thereof. (including the above composition).

16. The method according to claim 1 or 2, wherein the biological sample comprises a body fluid.

17. The method according to claim 16, wherein the body fluid is selected from blood, serum, plasma, cerebrospinal fluid, saliva, urine, and sputum.

18. The method according to claim 16, wherein the body fluid comprises blood, serum, or plasma.

19. The method according to claim 1 or 2, wherein the methylation pattern comprises a segment of a nucleotide sequence comprising at least 3 CpG dinucleotides.

20. The method according to claim 1 or 2, wherein the known methylation pattern is shown in Table 2.

Images