HMGB1 inhibitors for the treatment of APOE4-related tauopathies including Alzheimer's disease

Abstract

As described herein, inhibitors of high-mobility group box protein 1 (HMGB1) can significantly reduce nuclear-cytoplasmic transport of HMGB1, gliosis, neurodegeneration, tau pathology, and myelin loss, particularly in subjects having the APOE4 allele. Accordingly, methods are described herein that include administering to a subject having at least one genomic APOE4 allele, one or more inhibitors of high-mobility group box protein 1 (HMGB1).

Description

Technical Field

[0001] The present invention relates to HMGB1 inhibitors for the treatment of APOE4-related tauopathies including Alzheimer's disease.

Background Art

[0002] Alzheimer's disease is a chronic neurodegenerative disease that causes 60-70% of dementia cases. See, for example, Non-Patent Documents 1 and 2. There are approximately 30 to 35 million Alzheimer's disease patients worldwide. See the World Health Organization (Non-Patent Document 2). Alzheimer's disease affects approximately 6% of people aged 65 and over. Burns, A. et al. (Non-Patent Document 1). Alzheimer's disease is one of the most economically burdensome diseases in developed countries. See, for example, Non-Patent Document 3. However, at present, no treatment method for Alzheimer's disease is known.

Prior Art Documents

Non-Patent Documents

[0003]

Non-Patent Document 1

Non-Patent Document 2

Non-Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0004] Therefore, the treatment of Alzheimer's disease and other tauopathies remains in need.

Means for Solving the Problems

[0005] As described herein, inhibitors of High mobility group box protein 1 (HMGB1) can significantly reduce the nucleo-cytoplasmic translocation, gliosis, neurodegeneration, Tau pathologies, and myelin deficits of HMGB1, particularly in subjects with the APOE4 allele. Accordingly, methods are described herein that include administering one or more inhibitors of High mobility group box protein 1 (HMGB1) to a subject having at least one genomic APOE4 allele. In some cases, the subject has two genomic APOE4 alleles. For example, the subject may express detectable levels of APOE4 protein. Examples of HMGB1 inhibitors that can be used include glycyrrhizic acid, ethyl pyruvate, nicotine, (-)-epigallocatechin gallate (EGCG), tanshinone, chlorogenic acid, emodin-6-O-β-D-glucoside, rosmarinic acid, isorhamnetin-3-O-galactoside, persicarin, phloridzin, chloroquine, acteoside, cichonin, carbenoxolone, quercetin, lycopene, nafamostat mesilate, gabexate mesilate, sivelestat sodium, HMGB1 monoclonal antibodies (m2G7 or #10-22), recombinant HMGB1 box A protein, acetylcholine, nicotinic acetylcholine receptor subtype alpha7 agonist GTS-21, peptide P5779, resveratrol, metformin, or combinations thereof. In some cases, the subject may exhibit symptoms of HMGB1 nucleo-cytoplasmic translocation, gliosis, neurodegeneration, Tau pathology (accumulation of tau protein), or myelin deficit. For example, the subject may exhibit symptoms of at least one tauopathy. Such tauopathies are diseases characterized by abnormal deposition of tau protein in the brain.For example, tauopathy can be a neurodegenerative disease, Alzheimer's disease, Pick's disease, progressive supranuclear palsy, corticobasal degeneration, argyrophilic grain disease, primary age-related tauopathy, chronic traumatic encephalopathy, or frontotemporal dementia.

Brief Description of the Drawings

[0006]

Figure 1A

Figure 1B

Figure 1C

Figure 1D

Figure 1E

Figure 1F

Figure 1G

Figure 1H

Figure 1I

Figure 1J

Figure 1K

Figure 1L

Figure 1M

Figure 1N

Figure 2A

Figure 2B

Figure 3A

Figure 3B

Figure 3C

Figure 3D

Figure 3E

Figure 3F

Figure 3G

Figure 3H

Figure 3I

Figure 3J

Figure 4A

Figure 4B

Figure 4C

Figure 4D

Figure 4E

Figure 4F

Figure 4G

Figure 4H

Figure 4I

Figure 4J

Figure 4K

Figure 4L

Figure 4M

Figure 4N

Figure 4O

Figure 5A

Figure 5B

Figure 5C

Figure 5D

Figure 5E

Figure 5F

Figure 5G

Figure 5H

Figure 5I

Figure 5J

Figure 6A

Figure 6B

Figure 6C

Figure 6D

Figure 6E

Figure 6F

Figure 6G

Figure 6H

Figure 6I

Figure 6J

Figure 6K

Figure 6L

Figure 6M

Figure 6N

Figure 6O

Figure 6P

Figure 6Q

Figure 6R

Figure 7A

Figure 7B

Figure 7C

Figure 7D

Figure 8A

Figure 8B

Figure 8C

Figure 8D

Figure 8E

Figure 8F

Figure 9A

Figure 9B

Figure 10A

Figure 10B

Figure 10C

Figure 10D

Figure 10E

Figure 10F

Figure 11A

Figure 11B

Figure 11C

Figure 11D

Figure 11E

Figure 11F

Figure 11G

Figure 11H

Figure 11I

Figure 11J

Figure 12A

Figure 12B

Figure 12C

Figure 12D

Figure 12E

Figure 12F

Figure 12G

Figure 12H

Figure 12I

Figure 12J

Figure 12K

Figure 12L

Figure 12M

Figure 12N

Figure 12O

Figure 12P

Figure 12Q

Figure 12R

Figure 12S

Figure 13A

Figure 13B

Figure 13C

Figure 13D

Figure 13E

Figure 13F

Figure 13G

Figure 13H

Figure 13I

Figure 14A

Figure 14B

Figure 14C

Figure 14D

Figure 14E

Figure 14F

Figure 15A

Figure 15B

Figure 15C

Figure 15D

Figure 16A

Figure 16B

Figure 16C

Figure 16D

Figure 17A

Figure 17B

Figure 17C

Figure 17D

Figure 17E

Figure 17F

Figure 17G

Figure 17H

Figure 17I

Figure 17J

Figure 17K

Figure 17L

Figure 17M

Figure 18A

Figure 18B

Figure 19A

Figure 19B

Figure 19C

DETAILED DESCRIPTION OF THE INVENTION

[0007] As described herein, inhibitors of high mobility group box protein 1 (HMGB1) can significantly reduce nuclear-cytoplasmic translocation, gliosis, neurodegeneration, tau pathology, and myelin deficiency of HMGB1 in subjects having the APOE4 allele. For example, the method can be used for the treatment of tauopathy. Tauopathy is a neurodegenerative disorder characterized by abnormal deposition of tau protein in the brain. Such tauopathies include Alzheimer's disease, Pick's disease, progressive supranuclear palsy, corticobasal degeneration, and argyrophilic grain disease, primary age-related tauopathy, chronic traumatic encephalopathy, or frontotemporal dementia.

[0008] As exemplified herein, neuronal APOE4 promotes the release of HMGB1 from neurons, which initiates and / or exacerbates inflammation in the brain, induces gliosis, and results in neurodegeneration and myelin deficiency. In subjects having the APOE4 allele, the HMGB1 protein abnormally translocates from the nucleus to the cytosol of hippocampal neurons and is released from neurons into the hippocampal interstitial fluid. The experiments described herein show that administration of two HMGB1 inhibitors, ethyl pyruvate and glycyrrhizic acid, blocks the nuclear-cytoplasmic transport of HMGB1 in mice and dramatically reduces the extent of gliosis, neurodegeneration, tau pathology, and myelin deficiency in APOE4 mice, while having no distinguishable effect in APOE3 mice.

[0009] Methods are described herein that may include administering to a subject an HMGB1 inhibitor. The subject can be a subject whose genome contains the APOE4 allele. For example, the subject to be treated is a subject that expresses the APOE4 protein. In some cases, the method includes the step of administering to the subject a therapeutically effective amount of an HMGB1 inhibitor. Such a therapeutically effective amount of an HMGB1 inhibitor can reduce gliosis, neurodegeneration, tau deposition, and myelin loss. For example, such a therapeutically effective amount of an HMGB1 inhibitor can reduce one or more of gliosis, neurodegeneration, tau deposition, or myelin loss by at least 10%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 99%, or at least 99%.

[0010] High mobility group box protein 1 (HMGB1) High mobility group box protein 1 (HMGB1) is a ubiquitous nuclear protein released by glia and neurons upon inflammasome activation, and activates the receptor for advanced glycation end products (RAGE) and toll-like receptor (TLR) 4 on target cells. The HMGB1 / TLR4 axis is an important initiator of neuroinflammation. As shown herein, neuronal APOE4 promotes the release of HMGB1 from neurons, which initiates and / or exacerbates inflammation in the brain, induces gliosis, and results in neurodegeneration and myelin loss.

[0011] The sequence of the human HMGB1 protein is shown below as SEQ ID NO: 1 (UNIPROT accession number P09429).

[0012]

Chemical Structure

[0013] The cDNA sequence encoding such an HMGB1 protein is shown below as SEQ ID NO: 2 (NCBI accession number X12597.1).

[0014]

Chemical formula

[0015] The human HMGB1 gene is located on chromosome 13 (position 13q12.3; NC_000013.11 (30456704..30617597, complement)). Variants and homologs of any of the sequences described herein may also be relevant to the methods and compositions described herein. For example, such variants and homologs may have less than 100% sequence identity to any of the sequences described herein. Variants and homologs may have, approximately, at least 40% sequence identity, or at least 50% sequence identity, or at least 60% sequence identity, or at least 70% sequence identity, or at least 80% sequence identity, or at least 90% sequence identity, or at least 95% sequence identity, or at least 96% sequence identity, or at least 97% sequence identity, or at least 98% sequence identity, or at least 99% sequence identity, or 60 - 99% sequence identity, or 70 - 99% sequence identity, or 80 - 99% sequence identity, or 90 - 95% sequence identity, or 90 - 99% sequence identity, or 95 - 97% sequence identity, or 97 - 99% sequence identity, or 100% sequence identity to any of the sequences described herein.

[0016] HMGB1 inhibitor An HMGB1 inhibitor refers to any compound or molecule that can inhibit the expression, function, or extracellular release of HMGB1. Such HMGB1 inhibitors can include chemical compounds, antibodies, proteins, small RNA molecules, small DNA molecules, etc. that target the HMGB1 protein or HMGB1 nucleic acid.

[0017] The compound can be a synthesized compound or a naturally occurring compound. An example of an HMGB1 inhibitor that can be used is glycyrrhizin, also known as glycyrrhizic acid (the structure shown below).

[0018]

Chemical Structure

[0019] Glycyrrhizin is an HMGB1 inhibitor that can bind to one or both HMG boxes in HMGB1 and can inhibit the chemoattractant activity and mitogenic activity of HMGB1. After oral ingestion, glycyrrhizin is hydrolyzed by intestinal bacteria to 18β-glycyrrhetinic acid (enoxolone), and after absorption from the intestine, 18β-glycyrrhetinic acid is metabolized in the liver to 3β-monoglucuronyl-18β-glycyrrhetinic acid. The 3β-monoglucuronyl-18β-glycyrrhetinic acid metabolite circulates in the bloodstream. As a result, its oral bioavailability can be insufficient. Therefore, glycyrrhizin, 18β-glycyrrhetinic acid (enoxolone), and / or 3β-monoglucuronyl-18β-glycyrrhetinic acid can be administered intravenously or locally to a selected target tissue, such as the brain.

[0020] In other examples, HMGB1 inhibitors include, but are not limited to, ethyl pyruvate, nicotine, (-)-epigallocatechin gallate (EGCG), tanshinone, chlorogenic acid, emodin-6-O-β-D-glucoside, rosmarinic acid, isorhamnetin-3-O-galactoside, persicarin, forsythoside B, chloroquine, acteoside, cichonin, carbenoxolone, quercetin, lycopene, nafamostat mesilate, gabexate mesilate, sivelestat sodium, HMGB1 monoclonal antibodies (m2G7 or #10-22), recombinant HMGB1 box A protein, acetylcholine, nicotinic acetylcholine receptor subtype alpha7 agonist GTS-21, peptide P5779, resveratrol, metformin, and derivatives thereof, and may be used in the present invention. These compounds have been found to have an inhibitory effect on HMGB1 secretion and can be administered by any convenient method (e.g., orally, topically, or intravenously). One or more of these compounds can be administered in combination. Glycyrrhizin can also be administered together with one or more of these compounds.

[0021] In some cases, the HMGB1 inhibitor may be an RNA molecule such as siRNA, short hairpin RNA, microRNA, antisense RNA, or guide RNA. The RNA molecule can bind to an HMGB1 nucleic acid sequence such as SEQ ID NO: 2. The RNA molecule can be modified according to actual needs. Other suitable RNA molecules that can target the HMGB1 gene can also be used in the present invention. The following examples of HMGB1 siRNA are useful for targeting human HMGB1.

[0022]

Table 1

[0023] Apolipoprotein E4 The APOE gene encodes apolipoprotein E, a protein that binds to lipids including cholesterol to form lipoproteins. There are a number of different APOEs, but the major alleles are the e2, e3, and e4 alleles. The most common allele is e3, which is found in more than half of the general population. All apoE isoforms have approximately 299 amino acids, and the only difference is a single amino acid change. However, there are distinct functional differences as the common isoform apoE3 is not associated with Alzheimer's disease.

[0024] Apolipoprotein E4 (APOE4) is the strongest risk factor for sporadic late-onset Alzheimer's disease (AD) and accounts for the majority of Alzheimer's disease cases. APOE4 has a unique three-dimensional structure that affects its lipid-binding and receptor-binding properties and differs from APOE2 and APOE3 at amino acid positions 112 and 158. ApoE4 contains an arginine residue at position 112, whereas apoE3 has a cysteine at this position.

[0025] The sequence of the Homo sapiens apolipoprotein E isoform b precursor is shown below as SEQ ID NO:8 (NCBI accession number NP_000032.1).

[0026]

Chemical formula

[0027] The apolipoprotein E sequence of SEQ ID NO:8 contains an 18-amino acid signal peptide (bolded and underlined above). The amino acid sequence of the Homo sapiens apolipoprotein E isoform b precursor without the signal sequence is shown below as SEQ ID NO:9.

[0028]

Chemical formula

[0029] As described above, ApoE4 contains an arginine residue at position 112, whereas apoE3 has a cysteine at this position (bolded and underlined above). Accordingly, one sequence of Homo sapiens apolipoprotein E4 is shown below as SEQ ID NO: 10.

[0030]

Chemical formula

[0031] The human APOE gene is located on chromosome 19 (position 19q13.32; NC_000019.10(44905796..44909393) or NC_060943.1(47730492..47734089)). The APOE4 allele is found in approximately 10 - 15% of the population. Subjects with other APOE (non - APOE4) alleles typically have a lower probability of developing Alzheimer's disease or other tauopathies. Everyone has two copies of the APOE gene: people with the APOE E2 / E2 allele have the lowest overall risk of Alzheimer's disease, and those with APOE E4 / E4 have the highest risk. Other combinations of APOE - E2 / E3, E2 / E4, E3 / E3, and E3 / E4 fall in between.

[0032] To determine whether a subject has the APOE4 allele, various tests are available. Such tests can include genomic sequencing, polymerase chain reaction amplification, restriction enzyme mapping, single - nucleotide polymorphism detection (see, for example, any of the following websites: 23andme.com; empowerdxlab.com; nebula.org; alzheimersorganization.org; testing.com). Using any such test, it is possible to determine whether a subject has at least one APOE4 allele.

[0033] Dipeptidyl peptidase 10 (DPP10) Dipeptidyl peptidase 10 is a protein encoded by the DPP10 gene in humans. The DPP10 protein is a single-pass type II membrane protein that is a member of the S9B family in the clan SC of serine proteases. This protein does not have detectable protease activity, probably due to the absence of a conserved serine residue that is normally present in the catalytic domain of serine proteases. However, it binds to certain voltage-gated potassium channels and changes their expression and biophysical properties.

[0034] The sequence of the human DPP10 protein is available from the UNIPROT database under accession number Q8N608 and is shown below as SEQ ID NO: 11.

[0035]

Chem.

[0036] The cDNA encoding the above human DPP10 protein has the following nucleotide sequence (NCBI accession number NM_020868.6; SEQ ID NO: 12).

[0037]

Chem.

[0038]

Chem.

[0039]

Chem.

[0040]

Chem.

[0041] Any variant, isoform, and homolog of any of the DPP10 sequences described herein may also be relevant to the methods and compositions described herein. For example, such variants and homologs may have less than 100% sequence identity to any of the sequences described herein. Variants and homologs may have approximately at least 40% sequence identity, or at least 50% sequence identity, or at least 60% sequence identity, or at least 70% sequence identity, or at least 80% sequence identity, or at least 90% sequence identity, or at least 95% sequence identity, or at least 96% sequence identity, or at least 97% sequence identity, or at least 98% sequence identity, or at least 99% sequence identity, or 60-99% sequence identity, or 70-99% sequence identity, or 80-99% sequence identity, or 90-95% sequence identity, or 90-99% sequence identity, or 95-97% sequence identity, or 97-99% sequence identity, or 100% sequence identity to any of the sequences described herein.

[0042] As exemplified herein, an increase in the level of DPP10 can indicate and result in an increase in the level of APOE4. Thus, an increase in the DPP10 level in a sample from a subject indicates that the subject would benefit from treatment with at least one HMGB1 inhibitor, at least one DPP10 inhibitor, or a combination of one or more HMGB1 inhibitors and one or more DPP10 inhibitors. For example, if a sample from a subject has a DPP10 level that is increased by at least 10%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 100%, at least 1.1-fold, at least 1.2-fold, at least 1.25-fold, at least 1.5-fold, at least 2-fold, at least 2.5-fold, or at least 3-fold, the subject can benefit from treatment with such HMGB1 inhibitor(s) and / or DPP10 inhibitor(s).

[0043] DPP10 can be detected using an antibody that specifically binds to the DPP10 protein and / or a probe / primer that specifically binds to the DPP10 mRNA. Antibodies that bind to the DPP10 protein are available, for example, from Alomone Labs, Invitrogen, SigmaAldrich, and ThermoFisher Scientific. Probes and primers that specifically bind to the DPP10 mRNA can include a segment of about 15-100 nucleotides having at least 90% sequence identity or complementarity to the DPP10 coding region (e.g., the DPP10 cDNA having the sequence of SEQ ID NO: 12).

[0044] The DPP10 inhibitor comprises any compound or molecule capable of inhibiting the expression, function, or extracellular release of DPP10. Such DPP10 inhibitors can include chemical compounds, antibodies, proteins, small interfering RNAs, small DNAs, etc. that target the DPP10 protein or DPP10 nucleic acid.

[0045] Composition The present invention also relates to a composition containing one or more active agents such as any of the HMGB1 inhibitory compounds described herein. Such active agents can include polypeptides, nucleic acids encoding polypeptides (e.g., within expression cassettes or expression vectors), modified cells, inhibitory nucleic acids, small molecules, compounds identified by the methods described herein, or combinations thereof. The composition can be a pharmaceutical composition. In some embodiments, the composition can include a pharmaceutically acceptable carrier. "Pharmaceutically acceptable" means that the carrier, diluent, excipient, and / or salt is compatible with the other components of the formulation and not harmful to its recipient.

[0046] The composition can be formulated in any convenient form. In some embodiments, the active agent is administered in a "therapeutically effective amount." Such a therapeutically effective amount is an amount sufficient to obtain a desired physiological effect such as the alleviation of at least one symptom of tauopathy diseases. Such symptoms can include HMGB1 nuclear-cytoplasmic transport, gliosis, neurodegeneration, tau pathology, and myelin deficiency.

[0047] For example, the active agent can reduce the symptoms of HMGB1 nuclear-cytoplasmic transport, gliosis, neurodegeneration, tau pathology, and myelin deficiency by 5%, or 10%, or 15%, or 20%, or 25%, or 30%, or 35%, or 40%, or 45%, or 50%, or 55%, or 60%, or 65%, or 70%, or 80%, or 90%, 095%, or 97%, or 99%, or any numerical percentage between 5% and 100%.

[0048] To achieve the desired effect, the active agent can be administered as a single dose or divided doses. For example, the active agent can be administered at a dosage of at least about 0.01 mg / kg to about 500 - 750 mg / kg, at least about 0.01 mg / kg to about 300 - 500 mg / kg, at least about 0.1 mg / kg to about 100 - 300 mg / kg, or at least about 1 mg / kg to about 50 - 100 mg / kg per kg of body weight, although other dosages may also provide beneficial results. The amount administered will vary depending on various factors including, but not limited to, the type of small molecule or compound selected for administration, the mammalian disease, body weight, physical condition, health, and age. Such factors can be readily determined by a clinician using animal models or other test systems available in the art.

[0049] Administration of the active agent according to the present invention can be, for example, a single administration, multiple administrations, continuous or intermittent, depending on factors such as the physiological state of the recipient, whether the purpose of administration is therapeutic or prophylactic, and other factors known to those skilled in the art. Administration of the active agent and compositions of the present invention can be essentially continuous over a preselected period or can be a series of spaced doses. Both local and systemic administrations are contemplated.

[0050] To prepare the composition, the small molecule or compound is synthesized or otherwise obtained and purified as necessary or desired. These small molecules or compounds, and other agents, can be suspended in a pharmaceutically acceptable carrier and / or can be lyophilized or otherwise stabilized. These active agents are adjusted to appropriate concentrations and may be combined with other agents as needed. The absolute weight of a given small molecule or compound, and / or other agent, included in a unit dose can vary widely. For example, it is possible to administer at least one molecule, compound, and / or other agent, or a plurality of molecules, compounds, and / or other agents, in an amount of from about 0.01 to about 2 g, or from about 0.1 to about 500 mg. Alternatively, the unit dose can vary from about 0.01 g to about 50 g, from about 0.01 g to about 35 g, from about 0.1 g to about 25 g, from about 0.5 g to about 12 g, from about 0.5 g to about 8 g, from about 0.5 g to about 4 g, or from about 0.5 g to about 2 g.

[0051] The daily dose of the active agent of the present invention can similarly vary. Such a daily dose can, for example, range from about 0.1 g / day to about 50 g / day, from about 0.1 g / day to about 25 g / day, from about 0.1 g / day to about 12 g / day, from about 0.5 g / day to about 8 g / day, from about 0.5 g / day to about 4 g / day, and from about 0.5 g / day to about 2 g / day.

[0052] It is understood that the amount of the active agent for use in therapy will vary depending not only on the particular carrier selected, but also on the route of administration, the nature of the disease or condition being treated, and the age and condition of the subject. Ultimately, the responsible healthcare provider can determine the appropriate dosage. Further, the pharmaceutical composition can be formulated as a single unit dosage form.

[0053] Accordingly, one or more suitable unit dosage forms containing the active agent can be administered by a variety of routes including parenteral (including subcutaneous, intravenous, intramuscular, and intraperitoneal), oral, rectal, dermal, transdermal, intrathoracic, intrapulmonary, and intranasal (respiratory) routes. The active agent may also be formulated for sustained release (see, e.g., using microencapsulation, WO 94 / 07529 and U.S. Patent No. 4,962,091). The formulations can be conveniently provided in individual unit dosage forms, if desired, and can be prepared by any of the methods well known in the pharmaceutical art. Such methods can include mixing the active agent with a liquid carrier, solid matrix, semi-solid carrier, micronized solid carrier, or combinations thereof, and then, if desired, introducing or shaping the product into the desired delivery system. For example, the active agent can be bound to a convenient carrier such as nanoparticles, albumin, polyalkylene glycols, or can be supplied in prodrug form. The active agent and combinations thereof can be combined with carriers and / or encapsulated in vesicles such as liposomes.

[0054] The compositions of the present invention can be prepared in many forms including aqueous solutions, suspensions, tablets, hard or soft gelatin capsules, and other sustained release formulations such as liposomes and shaped polymeric gels. Administration of the inhibitor can also include parenteral or topical administration in an aqueous solution or sustained release vehicle.

[0055] Accordingly, the active agent and / or other agents can sometimes be administered in oral dosage forms, which can be formulated to protect small molecules, compounds, polypeptides, other agents, and combinations thereof from degradation or disintegration before the small molecules, compounds, other agents, and combinations thereof provide therapeutic utility. For example, in some cases, small molecules, compounds, and / or other agents can be formulated to be released into the intestine after passing through the stomach. Such formulations are described, for example, in U.S. Patent No. 6,306,434 and the references contained therein.

[0056] Liquid pharmaceutical compositions can be, for example, in the form of aqueous or oily suspensions, solutions, emulsions, syrups or elixirs, or in the form of dry powders for constitution with water or other suitable vehicles before use. Such liquid pharmaceutical compositions can contain conventional additives such as suspending agents, emulsifying agents, non-aqueous vehicles (which can include edible oils), or preservatives. The pharmaceutical composition may take the form of a suspension, solution, or emulsion in an oily or aqueous vehicle and may contain formulating agents such as suspending agents, stabilizers and / or dispersing agents. Suitable carriers include physiological saline, encapsulating agents (such as liposomes), and other materials. The active agent and / or other agents can be formulated in dry form (such as lyophilized form) in the presence or absence of a carrier. If a carrier is desired, the carrier can be included in the pharmaceutical formulation or can be separately packaged in a separate container and added to the inhibitor packaged in dry form, suspension or soluble concentrated form in a convenient liquid.

[0057] The active pharmaceutical agent and / or other agents can be formulated for parenteral administration (such as by injection, for example, by bolus injection or continuous infusion) and can be provided in unit dosage forms in ampoules, prefilled syringes, small volume infusion containers or multi-dose containers with added preservatives.

[0058] The composition can also contain other components such as active agents, antiviral agents, antibacterial agents, antimicrobial agents and / or preservatives. This specification is further illustrated by the following examples, which should in no way be construed as limiting.

Examples

[0059] Example 1: Materials and Methods This example illustrates the materials and methods used in the development of the present invention. Mouse Human LoxP-floxed APOE knock-in (fE) mice with conditional deletion of the human APOE gene were generated as described by Knoferle et al. (J. Neurosci., 34, 14069-14078 (2014)). Briefly, homozygous fE3 mice and fE4 mice (Wang et al., Neuron 109, 1657-1674.e7 (2021)) were crossed with synapsin 1-Cre transgenic mice [B6.Cg-Tg(Syn1-Cre)671Jxm / J] (The Jackson Laboratory) (Zhu et al., Genes Dev. 15, 859-876 (2001)) or GFAP-Cre transgenic mice [B6.Cg-Tg(GFAP-Cre)8Gtm] (National Cancer Institute Mouse Repository) (Bajenaru et al., Mol. Cell Biol., 22, 5100-5113 (2002); Uhlmann et al., Ann. Neurol., 52, 285-296 (2002)). fE / Cre mice were crossed with Tau-P301S (PS19) transgenic mice [B6;C3-Tg(Prnp-MAPT *They were further mated with (The Jackson Laboratory) P301S)PS19Vle / J] to generate PS19-fE4 and PS19-fE3 mice without Cre, with Syn1-Cre, or with GFAP-Cre. Littermates that were negative for Syn1-Cre or GFAP-Cre were used as PS19-fE controls. For the generation of the PS19-fE / Syn1-Cre line, only female Syn1-Cre mice were used for breeding purposes because germline recombination has been reported to occur in the progeny of male Syn1-Cre mice (Rempe et al., Genesis, 44, 44-49 (2006)). Wild-type (WT) mice [C57BL / 6J] were obtained from The Jackson Laboratory. All mice had a pure C57BL / 6 genetic background and were housed in a specific pathogen-free barrier facility at 19-23 °C and 30-70% humidity on a 12-hour light / dark cycle. Animals were identified by ear punch under short-term isoflurane anesthesia, and genotypes were determined by polymerase chain reaction (PCR) of a tail biopsy sample. All animals received no procedures other than those reported in this study. All animal experiments were performed in accordance with the guidelines and regulations of the National Institutes of Health, the University of California, and the Gladstone Institutes under protocol AN176773.

[0060] For brain tissue collection, mice were deeply anesthetized by intraperitoneal injection of avertin (Henry Schein) and perfused transcardially with 0.9% saline for 1 minute. Brains were fixed either as whole brains or hemispheres depending on the study. The right hemisphere was drop-fixed in 4% paraformaldehyde (16% PFA diluted in MilliQ H2O) (Electron Microscopy Sciences) for 48 hours, washed out in 1× PBS (Corning) for 24 hours, and cryoprotected in 30% sucrose (Sigma) at 4 °C for 48 hours. Fixed hemispheres were cut into 30-μm-thick coronal sections on a cryostat microtome (Leica) and stored at -20 °C in cryoprotectant solution (30% ethylene glycol, 30% glycerol, 40% 1× PBS). The left hemisphere was snap-frozen on dry ice and stored at -80 °C.

[0061] Immunohistochemistry For immunofluorescence staining, several sections (separated by approximately 300 μm) from each mouse were transferred to a 12-well plate in 1× PBS-T (PBS + 0.1% Tween®-20) (Millipore Sigma) and washed three times for 5 minutes each in PBS-T to remove the cryoprotectant solution. The sections were incubated in boiling antigen retrieval buffer (Tris buffer, pH 7.6) (TEKNOVA) for 5 minutes and washed twice for 5 minutes each in PBS-T. The sections were then incubated in blocking solution (5% normal donkey serum (Jackson Labs), 0.2% Triton-X (Millipore Sigma) in 1× PBS) for 1 hour at room temperature to prevent non-specific antibody binding. After blocking, the sections were washed once for 5 minutes in PBS-T and incubated in mouse-on-mouse (M.O.M.). Blocking buffer (1 drop of M.O.M IgG / 4 mL of PBS-T) (Vector Labs), 1 hour at room temperature. After M.O.M. blocking, the sections were diluted to the optimal concentration (anti-APOE 1:200 (Cell Signaling); anti-CD68 1:100 (Bio-Rad); anti-Cre 1:800 (Cell Signaling); anti-GFAP 1:800 (Millipore Sigma); anti-GFP 1:5000 (Thermofischer); anti-HMGB1 1:100 (Abcam); anti-Iba1 (rbt) 1:200 (Wako); anti-Iba1 (gt) 1:200 (Abcam); anti-MBP 1:500 (Abcam); anti-NeuN 1:500 (Millipore Sigma); anti-NG2 1:500 (Abcam); anti-S100β 1:200 (Abcam)) and then incubated overnight at 4 °C in the primary antibody. After primary antibody incubation, the sections were washed three times for 5 minutes each in PBS-T and then diluted in PBS-T and protected from light and incubated for 1 hour at room temperature in a fluorescently labeled secondary antibody (Abcam, Jackson Immuno, 1:1000 in PBS-T). The sections were then washed twice for 5 minutes each in PBS-T and incubated in DAPI (1:50,000 in PBS-T) (Thermofisher) for 8 minutes at room temperature protected from light.Next, the sections were washed twice in PBS-T for 5 minutes each, then mounted on microscope slides (Fisher Scientific), covered with a coverslip using ProLong Gold mounting media (Vector Laboratories), and sealed with clear nail polish. Images were taken at magnifications of 10x, 20x, 40x, or 60x depending on the staining using an FV3000 confocal laser scanning microscope (Olympus) or an Aperio VERSA slide scanning microscope (Leica). After setting a standard threshold value applied to all images, open-source Fiji (ImageJ) software was used to perform image analysis of the percent coverage area. The researchers were also blinded to the samples to eliminate the possibility of bias.

[0062] For DAB (3,3'-diaminobenzidine) staining, several sections (separated by approximately 300 μm) from each mouse were transferred to a 12-well plate in 1×PBS-T and then washed three times in PBS-T for 5 minutes each to remove the cryoprotectant solution. The sections were then incubated in boiling antigen retrieval buffer (1×PBS, 0.1 M sodium citrate, 0.1 M citric acid) (Fisher Scientific, Fluka) for 5 minutes and washed twice in PBS-T for 5 minutes each. Next, the sections were incubated in endogenous peroxidase buffer (1×PBS, 10% methanol (Fisher Scientific), 3% H2O2 (Sigma)) for 15 minutes and washed three times in PBS-T for 5 minutes each. The sections were then incubated in blocking solution (1×PBS-T, 5% normal donkey serum, 1% non-fat dry milk) for 1 hour at room temperature. After blocking, the sections were washed twice in PBS-T for 5 minutes each, then incubated in avidin / biotin blocking agent (4 drops per block) (Vector Laboratories) for 15 minutes and then washed twice in PBS-T for 5 minutes each. The sections were incubated in M.O.M. blocking buffer (1 drop of M.O.M IgG / 4 mL of PBS-T) (Vector Labs) for 1 hour at room temperature. After M.O.M. blocking, the sections were washed twice for 5 minutes each, diluted to the optimal concentration in PBS-T, and then incubated in primary antibody overnight at 4°C (anti-p tau (AT8) 1:100 (Invitrogen); anti-HT7 1:200 (Peter Davies)). After primary antibody incubation, the sections were washed three times in PBS-T for 5 minutes each and then incubated in biotinylated secondary antibody (1:200; Jackson Immuno) for 1 hour at room temperature. Next, the sections were washed three times in PBS-T for 5 minutes each and incubated in ABC buffer (Vector Laboratories) prepared 10 minutes before the incubation step. The sections were washed twice in PBS-T for 5 minutes each and once in Tris buffer (pH 7.6) for 5 minutes.The sections were incubated in DAB buffer (5 mL of 1×PBS, 2 drops of buffer stock solution, 2 drops of DAB, 2 drops of H2O2) (Vector Laboratories) for exactly 2 minutes. The staining was stopped by washing the sections 3 times in Tris buffer (pH 7.6) for 5 minutes each and 2 times in PBS-T for 5 minutes each. The sections were mounted on microscope slides and allowed to dry overnight at room temperature. Next, the mounted sections were immersed in xylene (Fisher Scientific) 2 times for 5 minutes each and then coverslipped using DPX mounting medium (Sigma-Aldrich). Images were taken using an Aperio VERSA slide scanning microscope (Leica) at 10× magnification.

[0063] Volumetric analysis. Serial coronal hippocampal brain sections (7 sections per mouse, 30 μm thick, 300 μm apart) were mounted on microscope slides (Fisher Scientific) and allowed to dry for 1 hour at room temperature. The 0.1% Sudan black solution was prepared by adding an appropriate amount of Sudan black powder (Sigma) to 70% ethanol (KOPTEC) and mixing the solution using a magnetic stirrer while protecting from light. The solution was then centrifuged at 3,000 RPM for 10 minutes, and the collected supernatant was filtered using a 0.2 μm filter syringe (Thermo Scientific) to remove undissolved dye. The sections were then stained with the 0.1% Sudan black solution for 10 minutes at room temperature and washed 3 times in 70% ethanol for 2 minutes each and 3 times in Milli-Q water for 5 minutes each. The sections were then coverslipped using ProLong Gold mounting medium (Invitrogen) and imaged using an Aperio VERSA slide scanning microscope (Leica) at 10× magnification. For volumetric analysis of the hippocampus and posterior lateral ventricle, the region of interest was traced in ImageJ using the segmented line tool, and the formula: volume = (sum of areas) *The volume was calculated using 0.3 mm (Uhlmann et al., Ann. Neurol., 52, 285 - 296 (2002)). The sum of the area values was obtained by taking the sum of the quantified area measurements of all seven brain sections per mouse, approximately between the coordinates AP = -1.2 and AP = -3.4.

[0064] Measurement of neuron layer thickness Two brain sections (30 μm thick, 300 μm apart) were immunofluorescently stained as described above using the primary antibody NeuN (1:500) to visualize the neuronal cell layer of the hippocampus. The sections were imaged at a magnification of 20× using an FV3000 confocal laser scanning microscope (Olympus). The thickness of the CA1 pyramidal cell layer and the dentate gyrus granule cell layer of the hippocampus was measured on Fiji (ImageJ) software by drawing a straight line perpendicular to the NeuN+ cell layer at two points per subfield of the hippocampus and taking the average value for each mouse.

[0065] Nuclear - cytoplasmic localization of HMGB1 measurement Two brain sections (30 μm thick, 300 μm apart) were immunostained with anti - HMGB1 (1:100) and DAPI (1:50,000) as described above. The sections were imaged at magnifications of 40× and 60× using an FV3000 confocal laser scanning microscope (Olympus). All image processing and quantification were performed using Fiji (ImageJ) software. Briefly, a median filter of 1 pixel was applied to the DAPI channel, and an appropriate threshold was set to create a mask of DAPI. Then, using the image calculator function, the DAPI mask and the HMGB1 channel were overlaid, resulting in HMGB1 staining localized only to the nucleus. After obtaining the integrated density and particle values, the image calculator was used to subtract the DAPI mask from HMGB1 to obtain the HMGB1 staining excluded from the nucleus.

[0066] Biochemical extraction of brain tissue The hippocampus was dissected from rapidly frozen mouse hemispheres after thawing on ice. The hippocampal tissue was weighed and homogenized at 10 μL / mg tissue using a Polytron® immersion disperser Polytron® homogenizer (Kinematica AG) in ice-cold RAB buffer (G Biosciences) supplemented with phosphatase inhibitor (Roche) and protease inhibitor (Roche). The samples were then centrifuged at 50,000 g for 20 min at 4 °C using an Optima TLX ultracentrifuge (Beckman Coulter), and the supernatant was collected as the RAB soluble fraction. The pellet was resuspended in ice-cold RIPA buffer (Thermo Scientific) at 10 μL / mg tissue and centrifuged at 50,000 g for 20 min at 4 °C. The supernatant was recovered as the RIPA soluble fraction, and the pellet was stored at -80 °C for further use. All fractions were stored at -80 °C until further analysis.

[0067] Western blot analysis Biochemically extracted mouse hippocampal tissue lysates were loaded onto a 12% Bis-Tris SDS-PAGE gel (Invitrogen) and separated by gel electrophoresis at 160 V using MOPS buffer. The separated proteins were transferred onto a nitrocellulose membrane at 18 V for 60 minutes (Trans-Blot Turbo Transfer System (Bio-rad)). The membrane was washed three times in PBS-T for 5 minutes each, then incubated in Intercept blocking buffer (LI-COR) for 1 hour at room temperature to block non-specific binding sites. After blocking, the membrane was washed three times in PBS-T for 5 minutes each and incubated overnight at 4 °C with primary antibodies (AT8 1:3,000 (Invitrogen), TUJ1 1:15,000 (Biolegend)). The membrane was washed three times in PBS-T for 5 minutes each and incubated in a fluorescently labeled secondary antibody (1:20,000; LI-COR) in the dark at room temperature for 1 hour. The resulting bands were detected using an Odyssey CLx infrared imaging system (LI-COR), and the fluorescence intensity of the bands was quantified as the ratio of the AT8:TUJ1 signals using Image Studio software.

[0068] Sandwich ELISA Biochemically extracted mouse hippocampal tissue lysates were diluted to an appropriate concentration in Milli-Q H2O and run according to the provided manufacturer's protocol (human APOE (Abcam); mouse HMGB1 (Novus Biologicals)). The reactions of the samples were read using a SpectraMaX® M5 spectrophotometer (Molecular Devices), the standard curve was interpolated, and the protein concentration was determined after adjusting the dilution factor.

[0069] Primary neuron cultures and recombinant protein treatment Primary cultures of neurons were prepared from prenatal E20 pups of various genotypes. After collecting the pups' brains, the cortex + hippocampus was isolated and placed in ice-cold dissociation medium + kynurenic acid medium (DM / KY) (in distilled H2O, DM: Na2SO4 (81.8 mM); K2SO4 (30 mM); MgCl2 (5.8 mM), CaCl2 (0.25 mM); HEPES (1 mM); glucose (20 mM); phenol red (0.001%); NaOH (0.16 mM)) (KY: kynurenic acid (10 mM); phenol red (0.0025%); HEPES (5 mM); MgCl2 (100 mM); NaOH (added dropwise to pH 7.4)). The resulting DM / KY medium was made by combining 90% DM with 10% KY medium. The isolated tissue was minced and then immersed in pre-warmed papain solution (1 mL / brain) for 13 minutes while gently inverting, and then immersed in trypsin inhibitor solution (5 mL for up to 10 brains) for 5 minutes while gently inverting. While gently inverting, the tissue pellet was washed with Optimem / glucose solution (20 mM glucose, 1 mL / brain). Then, fresh Optimem / glucose solution was added and the tissue was gently triturated until separated into single cells. After filtering the cells through a 40 μm cell strainer, the dissociated cells were plated in Neurobasal medium supplemented with B27, 100 U / mL -1 penicillin G, 100 μg / mL -1 streptomycin, and 1% GlutaMAX™ at 1 × 10 6 cells / well in a 12-well plate or 3 × 10 5 cells / well in a 24-well plate. Every 3 - 4 days, half of the medium was removed and replaced with fresh B27 / Neurobasal medium. In some experiments, primary neurons were treated in vitro for 24 hours on day 14 with either Dulbecco's PBS (dPBS) vehicle or the recombinant protein of interest (10 μg / mL for one well of a 12-well plate). After treatment, the medium was collected and the cultures were harvested for analysis. The total protein levels present in the cell lysates were obtained by BCA assay (Pierce).

[0070] Stereotaxic surgery of mice The mice were anesthetized by intraperitoneal injection of ketamine (60 mg / kg) and xylazine (30 mg / kg) and maintained with 0.8% - 1.0% isoflurane (Henry Schein). The mice were fixed to a stereotaxic alignment system model 940 (Kopf Instruments) using ear bars and a tooth bar. The scalp was prepared by removing hair using Nair® and sterilizing with 70% ethanol. The scalp was then incised using a scalpel and sterilized with 70% ethanol. The cranial suture was better visualized using 3% hydrogen peroxide. After identifying bregma, the unilateral stereotaxic site was drilled with a 0.5 mm microburr (Fine Science Tools) using coordinates X = +1.5, Y = -2.1, Z = -2.1 (Z is measured from the surface of the brain). Each virus (AAV2(Y444F)-SmCBA-human_P301S_tau-WPRE, 2.10E+13 vg / mL, Virovek; AAV2-synapsin-GFP, 1.0E+13 vg / mL, SignaGen) or the interstitial fluid (ISF) fraction was injected into the mice at a rate of 500 nL / min and diffused for 3 minutes. After the surgery, the mice were sutured with nylon monofilament non-absorbable 6-0 suture (Henry Schein), and analgesics buprenorphine (0.0375 mg / kg intraperitoneally), ketofen (5 mg / kg subcutaneously), and physiological saline (500 μL intraperitoneally) were administered. The mice were monitored on a heating pad until they were able to walk, and a hydrogel was provided for hydration.

[0071] Electrophysiological recording and data analysis of brain slices For electrophysiological recording studies, 8-month-old PS19-fE3 and PS19-fE4 mice without Cre, with Syn1-Cre or GFAP-Cre, were anesthetized with isoflurane and decapitated. The brains were quickly removed from the skulls and placed in ice-cold (2-5 °C) slicing solution. The slicing solution contained the following (in mM each): 110 choline chloride, 2.5 KCl, 26 NaHCO3, 10 MgCl2, 1.25 NaH2PO4, 0.5 CaCl2, 10 glucose, 3 sodium pyruvate, 1 L-ascorbic acid, pH 7.4. Sagittal slices 350 μm thick were cut from both hemispheres using a vibratome (VT1200, Leica) and transferred to a 95% O2-CO2 vapor interface holding chamber (BSK5, Scientific Systems Design) containing artificial cerebrospinal fluid (ACSF), where they were allowed to recover for 1 hour at 34 °C and then maintained at room temperature (20-22 °C). The ACSF contained (in mM each): 126 NaCl, 2.5 KCl, 1.5 CaCl2, 1.5 MgCl2, 26 NaHCO3, 1.25 NaH2PO4, 10 glucose, and 1.5 L-ascorbic acid, pH 7.4.

[0072] For input / output recording studies, orthodromic stimulation of the Schaffer collateral branches was used to evoke local field postsynaptic potentials (fPSPs). The stimulation was performed using a concentric bipolar stimulating electrode (FHC) connected to a constant-voltage isolation stimulator (DS2A-MKII, Digitimer North America) and placed in the CA2 radial layer. fPSPs were recorded using a glass borosilicate microelectrode filled with ACSF and placed in the CA1 radial layer. Signals were sampled and digitized by a MultiClamp 700B amplifier and Digidata 1550B1 acquisition system equipped with pClamp10 software (Molecular Devices), and analyzed using IgorPro6 software (Wavemetrics) that executed custom macros. The fPSP slope was analyzed as the linear fit slope value between 10% and 90% of the fPSP peak. The input-output relationship was recorded as the fPSP slope value in response to increasing stimulus intensities (20 - 60 μA), and the fPSP slope gain was calculated as the linear slope of the resulting input-output curve.

[0073] Microdialysis of the mouse hippocampus Brain interstitial fluid was collected using in vivo microdialysis of the hippocampus. Surgical procedures, including pre- and post-operative care, were performed as described above for stereotactic surgery. During surgery, a unilateral stereotactic site was drilled with a 1.2 mm bone drill bit (BASi), and an AtmosLM guide cannula PEG-4 (Amuza) was stereotactically implanted at coordinates X = +1.5, Y = -2.1, Z = -1.1 above the right hippocampus. The cannula was fixed in place using dental cement (GC America), a temporary PEG-4 AtmosLM dummy probe (Amuza) was inserted, and fixed with an AC-5 cap nut screw (Eicom). Two days after surgery, the mice were placed in a microdialysis-independent system (BASi) overnight for acclimation, and the next afternoon, a 1000 kDa AtmosLM collection probe (Eicom) was inserted into the hippocampus through the guide cannula. This extends 1 mm downward to Z = -2.1 to target the dentate gyrus. Artificial CSF (Harvard Apparatus) made with 0.15% BSA (Thermo Scientific) was circulated through the system at a rate of 0.5 μL / min using the push-pull method, and interstitial fluid (ISF) was collected approximately every 1 hour for 24 hours into a refrigerated fraction collector (BASI). To prevent tubing clogging, the pump was operated at 10 times the collection rate for the first 2 hours and then adjusted to a flow rate of 0.5 μL / min. After completion of ISF collection, the mice were euthanized and perfused with 0.9% saline as described above. The brain was sectioned into hemispheres, the right hemisphere was post-fixed in 4% PFA for 48 hours, and the left hemisphere was fresh frozen. The interstitial fluid (ISF) fraction was frozen at -80 °C for further analysis.

[0074] Treatment with HMGB1 inhibitor At 6.5 months of age, male and female PS19-fE4 mice and PS19-fE3 mice were randomly assigned to a control or treatment group. Mice were injected intraperitoneally with either sterile-grade 0.9% saline (Fisher Scientific) or a mixture of HMGB1 inhibitors: ethyl pyruvate (80 mg / kg) (Sigma-Aldrich) and glycyrrhizic acid (20 mg / kg) (Sigma-Aldrich) dissolved in 0.9% saline. Mice received three injections per week for 12 weeks starting at 6.5 months of age and reaching 9.5 months of age. During the experiment, the body weight changes, grooming changes, and posture of all mice were monitored, but no changes were observed. After treatment, the animals were perfused and their brain tissues were processed for histopathological analysis as described above.

[0075] Single-nucleus preparation for 10× loading Mouse hippocampi were dissected on ice and placed into a pre-chilled 2 mL Dounce containing 1 mL of cold 1× homogenization buffer (1× HB) (250 mM sucrose, 25 mM KCl, 5 mM MgCl2, 20 mM Tricine-KOH pH 7.8, 1 mM DTT, 0.5 mM Sermidine, 0.15 mM Sermine, 0.3% NP40, 0.2 units / μL RNase inhibitor, 0.2 units / μL protease inhibitor). Dounce with the loose pestle of “A” (about 10 strokes), then with the tight pestle of “B” (about 15 strokes). Filter the homogenate using a 70 μM Flowmi™ strainer (Eppendorf) and transfer it to a pre-chilled 2 mL LoBind® tube (Fischer Scientific). Pellet the nuclei by spinning at 350 RCF for 5 minutes at 4°C. Remove the supernatant and resuspend the nuclei in 400 μL of 1XHB. Next, add 400 μL of 50% iodixanol solution to the nuclei, then slowly layer 600 μL of 30% iodixanol solution under the 25% mixture, then layer 600 μL of 40% iodixanol solution under the 30% mixture. Then spin the nuclei in a pre-chilled swinging bucket centrifuge at 3,000 g for 20 minutes at 4°C. Collect 200 μL of the nuclear band at the 30%-40% interface and transfer it to a new tube. Next, add 800 μL of 2.5% BSA + 0.2 units / μL of RNase inhibitor in PBS to the nuclei and then spin at 500 RCF for 10 minutes at 4°C. Resuspend the nuclei in 2% BSA in PBS + 0.2 units / μL RNase inhibitor to reach about 500 nuclei / μL. Then filter the nuclei through a 40 μM Flowmi™ filter. Count the nuclei and then load about 13,000 nuclei per sample onto a 10× Genomics Next GEM Chip G. Prepare the snRNA-seq library using the Chromium Next GEM Single Cell 3’ Library and Gel Bead Kit v3.1 (10× Genomics) according to the manufacturer's instructions.The library was sequenced on an Illumina NovaSeq™ 6000 sequencer at the UCSF CAT Core.

[0076] Custom reference genome The PS19 tau mutant floxed APOE knock-in mouse model (Bien-Ly et al., J. Neurosci., 32, 4803-4811 (2012)) was used for single-nucleus RNA sequencing (snRNA-seq). Homo sapiens microtubule-associated protein tau (MAPT) (NCBI reference sequence: NM_001123066.4) (Agarwala et al., Nucleic Acids Res., 44, D7-D19 (2016)) and Homo sapiens APOE are the genes of interest in this study. Since these genes are not expected to be included in the mouse reference genome, a custom mouse reference genome was created using the reference mouse genome sequence (GRCm38) provided by Ensembl (release 98) (Howe et al., Ensembl 2021., Nucleic Acids Res., 49, D884-D891 (2021)) and the mouse gene annotation file provided by GENCODE (release M23) (Frankish et al., GENCODE 2021. Nucleic Acids Res., 49, D916-D923 (2021)) to quantify reads that align to these genes of interest. This method is similar to that used in the 10x Genomics Cell Ranger mouse reference package mm10 2020-A. The chromosome names included in the header of the Ensembl reference mouse genome sequence fasta file were modified to match the chromosome names included in the GENCODE fasta file. The annotation GTF file contains entries from non-polyA transcripts that overlap protein-coding genes. These reads were flagged as multi-mapped and not counted in the 10x Genomics Cell Ranger v6.1.1 counting pipeline (Zheng et al., Nat. Commun., 8, 14049 (2017)). To avoid this, the GTF file was modified to (1) remove the version suffix from transcript, gene, and exon ids to match the Cell Ranger reference package and (2) remove non-polyA transcripts.The Homo sapiens MAPT sequence and the Homo sapiens APOE sequence were added as separate chromosomes to the end of the mouse reference genome sequence, and the corresponding gene annotations were added to the filtered mouse reference gene annotation GTF file. A custom reference genome was constructed using the modified fasta and GTF files with the 10x Genomics Cell Ranger v6.1.1 mkref pipeline.

[0077] Pretreatment and clustering of mouse snRNA-seq samples The snRNA-seq samples included a total of 16 samples containing 4 mice from each of 4 genotype groups (PS19-fE4, PS19-fE4 Syn1-Cre, PS19-fE4 GFAP-Cre, and PS19-fE3). Each group of 4 mice consisted of 2 male and 2 female mice. The multiplexed fastq files for these samples were aligned using the 10x Genomics Cell Ranger v6.1.1 counting pipeline (Zheng et al., Nat. Commun., 8, 14049 (2017)) with a custom mouse reference genome (see Custom Reference Genome Methods for additional explanation). Details of this method are described in the Cell Ranger documentation. The intron inclusion flag for the counting pipeline was set to true to count reads mapping to intronic regions. The Cell Ranger count web summary indicated an error of "low fraction of reads in cells" for 2 samples (1 from the PS19-fE4 GFAP-Cre group and 1 from the PS19-fE3 group). For these 2 samples, only ~40% of the reads were assigned to cell-associated barcodes and <80% of the reads were mapped to the genome. These metrics were much higher than those of the other 14 samples. Checking the experimental records showed that these 2 samples had problems in the nuclear isolation step and that lower cDNA was recovered due to the use of an old batch of sample preparation reagents that had expired. The other 14 samples were all prepared using a new batch of sample preparation reagents. Therefore, these 2 samples were excluded and only the remaining 14 samples were used for downstream analysis by Seurat.

[0078] The filtered count matrices generated by the Cell Ranger count pipeline for 14 samples were processed using the R package Seurat v4.0.5 (Hao et al., Cell, 184, 3573 - 3587.e29 (2021)) for single - nucleus analysis. Each sample was pre - processed as a Seurat object, and the top 1% of cells per sample with a large number of unique genes, cells with 200 or fewer unique genes, and cells with 0.25% or more mitochondrial genes were filtered out for each sample. The 14 samples were integrated into one Seurat object, and normalization and variance stabilization were performed using sctransform (Hafemeister and Satija, Genome Biol., 20, 296 (2019)). At that time, the initial parameter estimation used the "glmGamPoi" (Bioconductor package version 1.6.0) method (Ahlmann - Eltze and Huber, Bioinformatics, 36, 5701 - 5702 (2021)).

[0079] Graph - based clustering was performed using the Seurat v4.0.5 functions FindNeighbors and FindClusters. First, the cells were embedded into a k - nearest neighbor (KNN) graph based on the Euclidean distance in the PCA space. The edge weights between two cells were further adjusted using the Jaccard similarity. Next, clustering was performed using the Louvain algorithm implemented in the Seurat function FindClusters. Clustering was performed for all combinations of 10, 15, 20 principal components (PCs) and resolutions of 0.4, 0.5, 0.6, 0.7, 0.8, 0.9. Clustering at 15 PCs and a resolution of 0.7 resulted in 33 distinct biologically relevant clusters, which were used for further analysis.

[0080] Cell type assignment Data visualization using Seurat v4.0.5 in the UMAP space for 14 samples did not reveal batch effects due to age, sex, genotype, date of birth, or date of nucleus isolation. Marker genes for each cluster were identified using the FindAllMarkers·Seurat function for SCT assay data. This algorithm uses the Wilcoxon rank-sum test to iteratively identify genes that are differentially expressed in the cluster relative to all other clusters. Marker genes were filtered to retain only positively expressed genes that were detected in at least 25% of the cells in any population and had a fold change of at least 0.5 log2. As previously reported by the inventors of the present application, the identity of the cell clusters was assigned by matching the cell clusters to known cell types using the expression of canonical cell type-specific genes, the expression of genes identified in publicly available mouse hippocampal single-cell RNA-seq datasets, and the expression of marker genes for each cluster in publicly available resources of in situ hybridization images of the whole brain (Zalocusky et al., Nat. Neurosci., 24, 786-798 (2021)).

[0081] Subclustering of sn-RNA-seq data of astrocytes and microglia Hippocampal cell cluster 10 was annotated as astrocyte cells, and hippocampal cell clusters 11, 21, and 29 were annotated as microglia cells. Both of these cell types were subclustered. Normalization and variance stabilization were performed using the "glmGamPoi" (Bioconductor package version 1.6.0) method (Ahlmann-Eltze and Huber, Bioinformatics, 36, 5701-5702 (2021)) for sctransform for initial parameter estimation 82It was performed using. Graph-based clustering was performed using the Seurat v4.0.5 functions FindNeighbors and FindClusters. First, cells were embedded into a k-nearest neighbor (KNN) graph based on Euclidean distances in the PCA space. The weights of the edges between two cells were further corrected using the Jaccard similarity. Next, clustering was performed using the Louvain algorithm implemented in the Seurat function FindClusters. Clustering was performed for all combinations of 10, 15, 20, 25, and 30 principal components (PCs) and resolutions of 0.4, 0.5, 0.6, 0.7, 0.8, 0.9. Subclustering at 15 PCs and a resolution of 0.9 resulted in 18 distinct biologically relevant subclusters for astrocytes. Subclustering at 15 PCs and a resolution of 0.9 resulted in 18 distinct biologically relevant microglia subclusters.

[0082] Gene Set Enrichment Analysis Genes differentially expressed between the target clusters were identified using the FindAllMarkers·Seurat function for the SCT assay data. This algorithm uses the Wilcoxon rank-sum test to identify genes differentially expressed between two populations. Differentially expressed genes were limited to those detected in at least 10% of the cells in either population and having at least a 0.1 log2 fold change. Overrepresentation analysis (or enrichment analysis) was performed using clusterProfiler v4.2.1 (Wu et al., Innov. (NY) 2, 100141 (2021)) and used to associate the gene sets of mice registered in the KEGG database (Kanehisa et al., Nucleic Acids Res., 44, D457-D462 (2016)) with the differentially expressed genes. The p-value is based on the hypergeometric test and was adjusted for multiple testing using the Benjamini-Hochberg method (Benjamini et al., J R Stat. Soc. Ser. B Stat. Methodol., 57, 289-300 (1995)). Significantly enriched gene sets were filtered to have an adjusted p-value of less than 0.8 and at least 10 differentially expressed genes present in the gene set. The same method was used for gene set enrichment analysis of astrocyte subclusters and microglia subclusters.

[0083] Association between clusters and genotypes A generalized linear mixed effects model (GLMM_AM) for evaluating the association with animal models was used to estimate the association between cluster membership and the mouse model by implementing the lme4 (v1.1 - 27.1) R package (Bates et al., J. Stat. Soft., 67, 1 - 48 (2015)). These models were run separately for each cluster of cells. The GLM models were run using the family argument set to a binomial probability distribution and the bobyqa control optimizer for maximum likelihood estimation. The cluster membership for each cell was modeled as a response variable ranging from 0 to 1 according to whether the cell belonged to the cluster under consideration. The corresponding mouse id from which the cell originated was set as a random effect variable, and the animal model for this mouse id was included as a fixed variable. The reference animal model was set to PS19 fE4. The p - values obtained for the three animal models (with respect to PS19 fE4) and the estimated log - odds ratios across clusters were adjusted for multiple testing using the Benjamini - Hochberg method (Benjamini et al., J R Stat. Soc. Ser. B Stat. methodol., 57, 289 - 300 (1995)). The same method was used to estimate the between - cluster association with genotypes for astrocyte sub - clusters and microglia sub - clusters.

[0084] Association between the proportion of cell types and histopathological parameters A generalized linear mixed effects model (GLMM_histopathology) for evaluating the association with histopathology was used to identify cell types whose proportion was significantly associated with histological changes across samples by implementing the lme4 (v1.1 - 27.1) R package (Bates et al., J. Stat. Soft., 67, 1 - 48 (2015)). These models were based on cell clusters and eight histological parameters: hippocampal volume (mm 3) Percent of AT8 coverage area, percent of GFAP coverage area, percent of S100β coverage area, percent of IBA1 coverage area, percent of CD68 coverage area, percent of MBP coverage area, and percent of oligodendrocyte progenitor cell (OPC) coverage area were each performed separately. The GLM model was run using a family argument set to the binomial probability distribution family and the bobyqa control optimizer for maximum likelihood estimation. The cluster membership for each cell was modeled as a response variable from 0 to 1 according to whether the cell belonged to the cluster under consideration. The corresponding mouse model from which the cell originated was included as a random effect, and further, the mouse id within a given mouse model was also modeled as a random effect. Note that this represents the hierarchical nature of this data in GLMM, where it is first assumed that the mouse models are sampled from the "universe of mouse models", followed by the individual mice being sampled within each mouse model. The choice to include the mouse models as random effects rather than fixed effects in the model means increasing the degrees of freedom (or maximizing the statistical power) to detect associations, especially considering that the number of replicates per animal model is relatively small (3 - 4). The histological parameters under consideration were modeled as fixed effects in this model.

[0085] A subset of the cell types of interest was selected, and after adjusting the distribution of p-values based on histopathological parameters between cell types, multiple testing correction by Benjamini-Hochberg (Benjamini et al., J R Stat.Soc. Ser.B Stat.Methodol., 57, 289-300 (1995)) was performed. The log odds ratio estimates (obtained from GLMM fitting) were visualized in a heatmap (Figure 4d) using the pheatmap package (version 1.0.12). The pipeline was applied to astrocyte and microglia subtypes, and the associations between the astrocyte and microglia subtypes of interest and eight histopathological parameters were visualized in Figures 4h and 4i, respectively. The first five principal component coordinates were estimated using eight log odds ratios for unit changes in histopathological parameters for each of the cell types of interest and the subcell types of astrocytes and microglia. This was implemented using prcomp(scale=T, center=T) of the stats R package. The inventors of the present application visualized the first two PCs using fviz_pca_ind() implemented in the factoextra 1.0.7 R package.

[0086] Mouse snRNA-seq dataset from GEO:GSE 164507 We reanalyzed mouse snRNAseq data available in the Gene Expression Omnibus database (accession number GSE164507) at ncbi.nlm.nih.gov / geo. For each sample, we downloaded publicly available datasets including a filtered matrix of genes by cell expression, a file with barcodes, and a file with expressed genes. The complete details of the study are available in the original publication (Wang et al., Neuron, 109, 1657 - 1674.e7 (2021)). Briefly, this study examined P301S mutant tau transgenic mice carrying floxed APOE-ε4 or APOE-ε3 alleles. These mice were crossed with mice expressing Cre recombinase under the regulation of a tamoxifen-inducible ER element and an Aldh1l1 astrocyte-specific promoter. Either tamoxifen or vehicle to induce Cre recombinase expression was administered to these Aldh1l1-CreERT2 mice at 5.5 months of age after the onset of tau pathology. Single nuclei isolated from the hippocampus of these mice were sequenced using 10×Genomics Chromium Single Cell sequencing, and the data were processed using the Cell Ranger Single Cell Software Suite (v3.0.2). The filtered count matrix generated by the Cell Ranger count pipeline for all eight samples was processed using Seurat v4.0.4 (Hao et al., Cell, 184:3573 - 3587.e29 (2021)). Samples were filtered to include only cells with 500 - 2000 detected genes and <5% mitochondrial reads. The filtered samples were integrated into a single Seurat object containing a matrix of 33,457 genes by 63,248 nuclei. Normalization and variance stabilization were performed using sctransform (Hafemeister and Satija, Genome Biol., 20, 296 (2019)).Clustering was determined to be performed in Seurat v4.0.4 using the RunPCA(), FindNeighbors(), and FindClusters() functions. The nearest neighbor distances were calculated using the first 15 PCs. This algorithm embeds cells into a k-nearest neighbor graph based on the Euclidean distance in the PCA space. The weights of the edges between any two cells are further refined using the Jaccard similarity. Clustering was performed using the Louvain algorithm with default settings and a resolution of 0.7, resulting in a set of 22 different clusters. The Seurat FindMarkers() function was used to detect differential gene expression by Wilcoxon rank sum test, with min.pct = 0.1, test.use = "wilcox", and logfc.threshold = 0.05.

[0087] General statistical analysis Differences between genotype groups were evaluated by ordinary one-way ANOVA using Tukey's multiple comparison test, which compares the mean of every other column to the mean of each column. All plotted data are shown as mean ± SEM. The correlation between two data points within the same genotype group was analyzed using simple linear regression and plotted as mean ± SEM. Analyses were performed and plots were created using GraphPad Prism version 9.2.0.

[0088] Example 2: Removal of the neuron- or astrocyte-specific APOE gene in human APOE-knockin mice expressing human mutant tau The inventors and co-researchers previously generated the following mouse lines: a mouse line expressing the floxed human APOE3 or APOE4 gene (Bien-Ly et al., J. Neurosci., 32, 4803-4811 (2012)), a mouse line expressing the Cre recombinase gene under the control of the neuron-specific synapsin-1 promoter (Syn1-Cre) (Zhu et al., Genes Dev. 15, 859-876 (2001)), or a mouse line expressing the Cre recombinase gene under the control of the astrocyte-specific glial fibrillary acidic protein promoter (GFAP-Cre) (Bajenaru et al., Mol. Cell Biol., 22, 5100-5113 (2002); Uhlmann et al., Ann. Neurol., 52, 285-296 (2002)).

[0089] These floxed APOE-KI (fE) mice express homozygous human APOE3 or APOE4 instead of the endogenous mouse Apoe. The human APOE gene is arranged in a form flanked by a pair of LoxP sites to enable its precise excision in the presence of cell type-specific Cre recombinase expression (Knoferle et al., J. Neurosci., 34, 14069-14078 (2014)). fE mice without Cre, with Syn1-Cre, or with GFAP-Cre were crossed with mice expressing the mutant 1N4R human microtubule-associated protein tau (MAPT) encoding the disease-related P301S mutation (PS19 strain), which is widely used as a tauopathy mouse model (Yoshiyama et al., Neuron 53, 337-351 (2007)). The resulting compound mice are designated as PS19-fE, PS19-fE / Syn1-Cre, or PS19-fE / GFAP-Cre mice.

[0090] Performed a rigorous characterization of fE / Syn1-Cre and fE / GFAP-Cre mice to verify the specificity of Cre recombinase expression under the neuron-specific Syn1 or astrocyte-specific GFAP promoter (Knoferle et al., J. Neurosci., 34, 14069-14078 (2014)). Further verified PS19-fE / Cre mice by immunohistochemical analysis of brain sections from 10-month-old PS19-fE4 / Syn1-Cre mice and PS19-fE4 / GFAP-Cre mice. Co-immunostaining with NeuN and GFAP antibodies was performed together with Cre recombinase antibody to confirm the specificity of Cre recombinase expression in hippocampal neurons and astrocytes when driven under the Syn1 or GFAP promoter, respectively. In PS19-fE4 / Syn1-Cre mice, Cre recombinase was expressed only in NeuN-positive neurons and not in GFAP-positive astrocytes. In PS19-fE4 / GFAP-Cre mice, Cre recombinase was expressed only in GFAP-positive astrocytes and not in NeuN-positive neurons. Co-immunostaining using antibodies against APOE, GFAP, and NeuN showed that PS19-fE4 / Syn1-Cre mice lacked APOE expression in neurons but had APOE expression in astrocytes, while PS19-fE4 / GFAP-Cre mice had APOE expression in some neurons but lacked APOE expression in astrocytes, compared to PS19-fE4 mice without Cre, which had APOE expression in GFAP-positive astrocytes and some NeuN-positive neurons. PS19-fE4 / GFAP-Cre mice showed some APOE-positive cells that were negative for NeuN and GFAP but positive for the microglial marker Iba1, indicating that they were APOE-expressing microglia.

[0091] To quantitatively determine the levels of APOE protein in these various mouse models, hippocampal lysates were analyzed by sandwich ELISA for human APOE in 10-month-old mice. PS19-fE4 / Syn1-Cre mice showed an approximately 20% decrease in APOE levels compared to PS19-fE4 mice (Figure 1A), which is consistent with previous data indicating that neuronal APOE contributes approximately 20-30% of the total APOE protein levels in the hippocampus and cortex. PS19-fE4 / GFAP-Cre mice showed an approximately 70% decrease in APOE levels compared to PS19-fE4 mice, which is consistent with the well-established role of astrocytes as the main producers of APOE within the CNS. Similarly, compared to PS19-fE3 mice, PS19-fE3 / Syn1-Cre mice showed an approximately 25% decrease in APOE levels, and PS19-fE3 / GFAP-Cre showed an approximately 67% decrease (Figure 1A). Collectively, these results provide strong evidence that in these compound mouse models, APOE gene expression is excluded in neurons or astrocytes when Cre recombinase expression is driven under the Syn1 or GFAP promoter, respectively.

[0092] Example 3: Removal of APOE4 from neurons rather than astrocytes dramatically reduces tau pathology. To determine whether the removal of neuronal or astrocytic APOE affects tau pathology, the inventors of the present application evaluated 10-month-old mice in which PS19 mice show extensive tau pathology throughout the hippocampus (Yoshiyama et al., Neuron, 53, 337-351 (2007)). The accumulation of p-tau in the hippocampus was analyzed by immunohistochemical staining using a p-tau specific AT8 antibody, and tau pathology was quantified as the percentage of the AT8-covered area in the hippocampus.

[0093] As previously reported (Shi et al., Nature, 549, 523-527 (2017)), PS19-fE4 mice showed extensive tau pathology throughout the hippocampus, and the degree of tau pathology was significantly lower in PS19-fE3 mice (Figs. 1B-1C). Compared with PS19-fE4 mice, PS19-fE4 / Syn1-Cre mice showed a significant ~81% reduction in tau pathology, while PS19-fE4 / GFAP-Cre mice showed a slight ~30% reduction that did not reach statistical significance. Interestingly, the reduction in p-tau coverage area in PS19-fE4 / Syn1-Cre mice was similar to the degree of tau pathology observed in PS19-fE3 mice (Figs. 1B-1C). There was no significant difference in tau pathology between PS19-fE3 with Cre and PS19-fE3 without Cre, probably because the tau pathology in PS19-fE3 mice was already low.

[0094] In addition, Western blot analysis was used to evaluate the levels of p-tau in mouse hippocampal tissues after sequential biochemical extraction using RAB and RIPA buffers containing highly soluble and less soluble tau proteins, respectively (Shi et al., Nature, 549, 523-527 (2017); Ishihara et al., Neuron, 24, 751-762 (1999)) (Figs. 1D-1G). There was no significant difference in p-tau levels among various genotype groups in the RAB fraction (Figs. 1D, 1F). However, PS19-fE4 / Syn1-Cre and PS19-fE3 mice showed a significant reduction in p-tau levels in the RIPA fraction compared with PS19-fE4 mice (Figs. 1E, 1G). PS19-fE4 / GFAP-Cre mice showed a tendency for a ~40% reduction in p-tau levels in the RIPA fraction compared with PS19-fE4 mice, which did not reach statistical significance. Collectively, these data indicate that neuronal APOE4 expression is a strong driver of tau pathology, while astrocytic APOE4 has a minimal effect on tau pathology.

[0095] Example 4: After the removal of APOE4 in neurons instead of astrocytes, the propagation of tau pathology is reduced. To further investigate the mechanism by which neuronal APOE4 drives tau pathology, we determined the effect of cell-type specific APOE4 expression on the propagation of tau pathology. Previous studies have shown that pathological tau can spread to anatomically connected brain regions after direct injection of various forms of tau protein into the mouse brain (Boluda et al., Acta Neuropathol., 129, 221-237 (2015); Rauch et al., Nature, 580, 381-385 (2020); Guo et al., J. Exp. Med., 213, 2635-2654 (2016); Kaufman et al., Neuron, 92, 796-812 (2016)).

[0096] The extent of tau propagation was analyzed after a single unilateral injection of adeno-associated virus-2 encoding human P301S mutant tau (AAV2-tau-P301S) into the right dorsal hippocampus of fE mice without Cre, with Syn1-Cre, or with GFAP-Cre (Fig. 1H-1I). fE mice lack human P301S mutant tau and instead express the endogenous mouse Mapt gene, and fE mice show minimal tau pathology, allowing for more accurate detection of human tau spread. Mice were injected with AAV2-tau-P301S virus at 10 months of age and evaluated at 13 months of age, 12 weeks after injection (Fig. 1H).

[0097] In contrast to the unexpected result that the virus encoding tau itself moves to the non-injected hippocampal side, to provide evidence that the observed tau propagation phenotype could truly be attributed to the spread of pathological human tau between neurons, unilateral injection into the right dorsal hippocampus of 10-month-old fE4 mice was evaluated for the same serotype of AAV2 encoding GFP (AAV2-GFP). Immunostaining with anti-GFP 2 weeks after injection revealed that the GFP signal remained localized to neurons within the injected hippocampal side (Figure 1J). The non-injected hippocampal side had no evidence of GFP signal in neuronal cell bodies, but there were some GFP-positive neuronal projections that were likely to have originated from neurons present on the injected hippocampal side. These results indicate that AAV2 itself does not spread between the right and left hippocampi after unilateral injection. Furthermore, for all immunohistochemical analyses, the number of neurons containing soma-positive tau was quantified to more accurately reflect the spread of human mutant tau between neurons, excluding confounding factors such as tau-positive cross-linked fibers from neurons originating from the injected side.

[0098] Immunohistochemical staining using the human tau antibody (HT7) revealed strong human tau propagation to the non-injected hippocampal side in PS19-fE4 mice and minimal tau propagation in PS19-fE3 mice (Figures 1K–1L), indicating that APOE4 promotes tau diffusion. Interestingly, PS19-fE4 / Syn1-Cre mice had limited human tau propagation to the non-injected hippocampal side, while PS19-fE4 / GFAP-Cre mice had extensive human tau propagation similar to the spread of tau pathology observed in PS19-fE4 mice (Figures 1K–1L). In particular, the number of HT7 + cells with tau propagation decreased by 55% after removal of neuronal APOE4, and there was no discernible difference in the number of HT7 + cells with tau propagation after removal of astrocytic APOE4 (Figures 1K–1N).

[0099] The degree of tau propagation was further analyzed by immunostaining of p-tau using the AT8 antibody. Both PS19-fE4 mice and PS19-fE4 / GFAP-Cre mice showed strong propagation of p-tau to the non-injected hippocampal side, while PS19-fE4 / Syn1-Cre mice and PS19-fE3 mice dramatically reduced p-tau propagation (Figs. 1M-1N).

[0100] Collectively, these data indicate that one mechanism by which neuronal APOE4 drives tau pathology is by stimulating the propagation of pathological tau across synaptically connected neurons. This data also further supports the idea that astrocytic APOE4 is not an important mediator of tau pathology, as evidenced by its lack of effect on tau propagation.

[0101] Example 5: Network hyperexcitability is eliminated after removal of neuronal APOE4 but not after removal of astrocytic APOE4 To determine the cell-type specific effects of APOE4 on neuronal function in the context of tauopathy, neuronal network excitability was measured by input-output curve analysis of network responses to incremental stimulation of Schaffer collaterals in the hippocampal cornu ammonis (CA1) region of PS19-fE3 and PS19-fE4 mice without Cre, with Syn1-Cre, or with GFAP-Cre (Fig. 2A). 54 。

[0102] PS19-fE4 mice had significant CA1 neuronal hyperexcitability compared to PS19-fE3 mice (Fig. 2B). Removal of neuronal APOE4 eliminated the hyperexcitability of the neuronal network, while removal of astrocytic APOE4 did not reach significance but resulted in a slight decrease (Fig. 2B). This indicates that neuronal APOE4 causes neurological dysfunction in the context of tauopathy.

[0103] Example 6: Neurodegeneration decreases after the removal of APOE4 from either neurons or astrocytes The degree of neurodegeneration in 10-month-old PS19-fE mice was evaluated after removing APOE from neurons or astrocytes. Analysis of hippocampal volume and posterior ventricular volume revealed that PS19-fE4 mice showed extensive neurodegeneration compared to PS19-fE3 mice (Figs. 3A - 3C). Neurodegeneration significantly decreased in both PS19-fE4 / Syn1-Cre mice and PS19-fE4 / GFAP-Cre mice, as indicated by an increase in hippocampal volume and a decrease in posterior ventricular volume (Figs. 3A - 3C), which was similar to the increase in hippocampal volume and decrease in posterior ventricular volume in PS19-fE3 mice. Removal of APOE3 from neurons or astrocytes had no significant effect on neurodegeneration (Figs. 3A - 3C).

[0104] Quantification of the degree of neuronal cell loss in various subfields of the hippocampus revealed that PS19-fE4 mice had extensive neuronal loss in CA1 and the dentate gyrus, while both PS19-fE4 / Syn1-Cre and PS19-fE4 / GFAP-Cre mice showed a significant decrease in neuronal cell loss in both the CA1 and dentate gyrus subfields of the hippocampus (Figs. 3D - 3G). When APOE3 was removed from neurons or astrocytes, there was no significant difference in the thickness of the neuronal cell layer in PS19-fE3 mice (Figs. 3D - 3G).

[0105] In PS19-fE4 mice, there was a weak but significant negative correlation between tau pathology and hippocampal volume (Figure 3H), suggesting that tau pathology contributes, at least to some extent, to the neurodegeneration occurring in these mice. There was also a strong negative correlation between hippocampal volume and the volume of the posterior ventricle, and a strong positive correlation between the thickness of the CA1 cell layer and hippocampal volume (Figures 3I - 3J). Overall, these data indicate that the removal of APOE4 from either neurons or astrocytes is protective against tau-mediated neurodegeneration and results in a reduction in neuronal cell and hippocampal volume loss. This result is interesting considering the tau pathology data indicating that astrocytic APOE4 has a minimal effect on the accumulation and propagation of tau pathology.

[0106] Example 7: After removing APOE4 from either neurons or astrocytes, gliosis dramatically decreases. To examine the role of neuronal or astrocytic APOE4 in inducing gliosis in relation to tauopathy, the degree of microgliosis and astrogliosis was investigated within the genotype groups at 10 months of age. Immunohistochemical staining of microglia using the Iba1 antibody revealed that PS19-fE4 mice had extensive microgliosis in the hippocampus, as demonstrated by a high percentage of Iba1-covered area (Figure 4A). Both PS19-fE4 / Syn1-Cre mice and PS19-fE4 / GFAP-Cre mice showed a dramatic reduction in microgliosis (Figure 4A). Furthermore, all PS19-fE3 mice with or without Cre, Syn1-Cre, or GFAP-Cre had a greatly reduced microgliosis compared to PS19-fE4 mice (Figure 4A). There was a strong negative correlation between the Iba1-covered area and hippocampal volume in PS19-fE4 mice, indicating that microgliosis is a good indicator of neurodegeneration and a potential contributing factor (Figure 4B).

[0107] This microgliosis phenotype was further defined by immunohistochemical staining of CD68, a marker of activated microglia. There were dramatic differences in the coverage area of activated microglia among the genotype groups, with PS19-fE4 mice showing extensive microglial activation and all other genotype groups showing minimal microglial activation (Figure 4C). Interestingly, among all the pathological correlations made in PS19-fE4 mice, the coverage area of CD68+ activated microglia had the strongest negative correlation with hippocampal volume (Figure 4D), suggesting that the degree of microglial activation is the strongest indicator and a potential contributing factor for APOE4-enhanced neurodegeneration in tauopathy. Collectively, these data indicate that microgliosis is strongly enhanced by APOE4 as compared to APOE3 in tauopathy and that removal of APOE4 from either neurons or astrocytes weakens the degree of microgliosis.

[0108] The degree of astrogliosis was also evaluated after removing APOE from either neurons or astrocytes. PS19-fE4 mice showed significant astrogliosis in the hippocampus compared to PS19-fE3 mice, as detected by immunohistochemical staining with an astrocyte GFAP antibody (Figure 4E). The degree of astrogliosis decreased significantly after removal of either neuronal APOE4 or astrocytic APOE4 (Figure 4E). These observations also showed the activated astrocyte marker S100β 56This was also supported by immunohistochemical staining for (Figure 4F). There was no obvious difference in the degree of astrogliosis after removing APOE3 from either cell type compared to PS19-fE3 mice (Figures 4E-4F). Surprisingly, and in stark contrast to microgliosis, neither the coverage area of astrocytes nor that of activated astrocytes significantly correlated with hippocampal volume (Figures 4G-4H). These data indicate that APOE4 potently enhances astrogliosis compared to APOE3 in tauopathy, and that removal of APOE4 from either neurons or astrocytes reduces astrocyte activation and / or recruitment. Furthermore, it is also shown that the degree of astrogliosis may not be a beneficial indicator or contributor to APOE4-enhanced neurodegeneration in tauopathy.

[0109] To further clarify the relationship between tau pathology and gliosis, the inventors of the present application examined whether tau propagation induces gliosis even on the non-injected hippocampal side (see the study in Figures 1H-1N). Indeed, fE4 mice showed higher microgliosis on the non-injected hippocampal side than fE3 mice, along with higher tau spread (Figure 4I). Interestingly, after removing either neuronal APOE4 or astrocytic APOE4, there was a significant decrease in microgliosis on the non-injected hippocampal side. This indicates that removal of neuronal APOE4 or astrocytic APOE4 reduces microgliosis regardless of the degree of tau spread, as fE4 / GFAP-Cre mice still had relatively high levels of tau spread on the non-injected hippocampal side.

[0110] These findings indicate that astrocytic APOE4 acts downstream of tau pathology to promote microgliosis. The degree of astrocytic coverage in the non-injected hippocampal side was also evaluated, and no distinguishable difference in astrogliosis was observed after the removal of neuronal APOE4 or astrocytic APOE4 (Figures 4I-4J), probably because astrogliosis in the non-injected hippocampal side is relatively low even in fE4 mice in this tau spreading model.

[0111] Example 8: Myelin deficiency and depletion of oligodendrocyte progenitor cells are reduced after removal of APOE4 from neurons, but not after removal of APOE4 from astrocytes Myelin degeneration and oligodendrocyte loss have been observed in the human AD brain (Benitez et al., Neuroimage Clin., 4:64-71 (2013); Dean et al., JAMA Neurol., 74, 41-49 (2017); Nasrabady et al., Acta Neuro.Comms., 6, 22 (2018)) and in mouse models of Alzheimer's disease and tauopathy (Desai et al., Glia, 57, 54-65 (2009); Zhang et al., Sci. Adv., 6, eabb8680 (2020); Shi et al., Neuron, 109, 2413-2426.e7 (2021)).

[0112] The inventors of the present application examined the effect of neuronal APOE4 and astrocytic APOE4 removal on the maintenance of myelin integrity and the density of hippocampal oligodendrocyte progenitor cells (OPCs). To determine the integrity of myelin in the hippocampus, myelin basic protein (MBP) was immunostained with an MBP-specific antibody. Quantification of the percentage of the covered area of MBP in the stratum radiatum under the pyramidal cell layer of CA1 revealed that PS19-fE4 mice had extensive myelin loss compared to PS19-fE3 mice (Figure 4K). Surprisingly, a significant reduction in myelin loss was observed in PS19-fE4 / Syn1-Cre mice, which were similar to the phenotype of PS19-fE3 mice, while PS19-fE4 / GFAP-Cre mice showed substantial myelin loss similar to the phenotype of PS19-fE4 mice (Figure 4K). These data indicate that neuronal APOE4 plays a major detrimental role in promoting myelin deficits in the context of tauopathy, while astrocytic APOE4 does not affect this phenotype. There was no substantial difference in myelin integrity in PS19-fE3 mice after removing APOE3 from neurons or astrocytes.

[0113] To further characterize the effects of neuronal APOE4 and astrocytic APOE4 on biological components involved in the myelin formation process, cells were immunostained with an NG2 antibody to search for oligodendrocyte progenitor cells (OPCs) that have been shown to help repair damaged myelin in the context of CNS injury and neurodegeneration (Shi et al., Glia, 57, 54 - 65 (2009); Tripathi et al., J. Neurosci, 30, 16383 - 16390 (2010)). A significant decrease in the percentage of OPC coverage area in the hippocampus of PS19-fE4 mice was observed compared to PS19-fE3 mice (Figure 4L). Surprisingly, removal of neuronal APOE4 significantly increased the percentage of OPC coverage area to levels similar to those in PS19-fE3 mice (Figure 4L). In contrast, PS19-fE4 / GFAP-Cre mice had a significantly lower OPC coverage area compared to PS19-fE4 / Syn1-Cre mice and showed a tendency for decreased OPCs compared to PS19-fE4 mice, but did not reach statistical significance (Figure 4L). After removal of neuronal or astrocytic APOE3 in PS19-fE3 mice, there was no obvious difference in the OPC coverage area (Figure 4L).

[0114] Immunohistochemical staining for myelin and OPCs in fE4 and fE3 mice revealed that APOE4 mice lacking human mutant tau-P301S did not show myelin deficiency and had OPC levels similar to fE3 mice in the hippocampus, indicating that the effect of APOE4 on these phenotypes is specific to the context of tauopathy.

[0115] Overall, these findings indicate that neuronal APOE4 plays a central role in depleting the hippocampal OPC pool and causing myelin deficiency in this complex tauopathy mouse model, while astrocytic APOE4 does not significantly contribute to either of these two pathological processes.

[0116] Example 9: By single-nucleus RNA sequencing (snRNA-seq), subpopulations of neurons and oligodendrocytes associated with neurodegenerative diseases were identified, and most of these were eliminated by removing neuronal APOE4 rather than astrocytic APOE4. To gain a deeper understanding of the cell-type specific effects of APOE4 at the transcriptome level across different types of hippocampal cells, single-nucleus RNA sequencing (snRNA-seq) was performed on hippocampi isolated from 10-month-old PS19-fE4 mice and PS19-fE3 mice without Cre, with Syn1-Cre, or with GFAP-Cre. The snRNA-seq dataset contained 119,317 nuclei covering 26,285 genes after normalization and filtering for quality control. Clustering by shared nearest neighbor (SNN) and visualization by uniform manifold approximation and projection (UMAP) revealed 33 different cell clusters. Based on the expression of marker genes, these clusters were assigned as follows: · 16 excitatory (Ex) neuron clusters (3 - 5, 7, 12, 15, 16, 18 - 20, 23, 25 - 28, 33); · 7 inhibitory (In) neuron clusters (6, 8, 9, 13, 22, 24, 30); · 3 oligodendrocyte clusters (1, 2, 17); · 1 astrocyte cluster (10); · 3 microglia clusters (11, 21, 29); · 2 oligodendrocyte progenitor cell (OPC) clusters (14, 32); and · 1 unknown cluster (31).

[0117] APOE was highly expressed in astrocytes of PS19-fE4 mice and PS19-fE3 mice, and its expression was dramatically reduced in astrocytes of PS19-fE4 / GFAP-Cre mice. The inventors and co-researchers of the present application have reported that some neurons also expressed APOE in PS19-fE4 mice and PS19-fE3 mice (Zalocusky et al., Nat. Neurosci., 24, 786-798 (2021)). As described herein, neuronal APOE expression was eliminated in PS19-fE4 / Syn1-Cre mice.

[0118] To evaluate the association with animal models, log odds ratio estimates from a Generalized Linear Mixed Effects Model (GLMM_AM) were used to identify cell clusters that changed in PS19-fE3, PS19-fE4 / Syn1-Cre, and PS19-fE4 / GFAP-Cre mice relative to PS19-fE4 mice. This analysis revealed that oligodendrocyte cluster 17 was less likely to have cells from PS19-fE4 / Syn1-Cre mice than from PS19-fE4 mice and was almost completely eliminated in PS19-fE4 / Syn1-Cre mice. Based on differentially expressed (DE) gene analysis, cells in cluster 17 showed dramatically upregulated expression of three major heat shock protein genes, as well as significantly downregulated expression of the myelin basic protein (Mbp) and myelin-associated oligodendrocyte basic protein (Mobp) genes, compared to other oligodendrocyte clusters. Differential expression (DE) pathway analysis revealed enrichment of Kyoto Encyclopedia of Gene and Genomes (KEGG) pathways related to general neurodegeneration, AD, and other neurodegenerative diseases, indicating that cluster 17 represents oligodendrocytes associated with neurodegenerative diseases.

[0119] Furthermore, the odds of having cells from PS19-fE4 / Syn1-Cre mice and PS19-fE3 mice in excitatory neuron cluster 25 were significantly lower than those from PS19-fE4 mice, and were almost completely excluded in PS19-fE4 / Syn1-Cre mice. Differential expression (DE) gene analysis revealed that cells in cluster 25 had significantly upregulated expression of three major heat shock protein genes, the human MAPT gene, and the amyloid precursor protein (App) gene. Differential expression (DE) pathway analysis revealed enrichment of KEGG pathways related to general neurodegeneration, Alzheimer's disease, and other neurodegenerative diseases, indicating that cluster 25 represents neurons related to neurodegenerative diseases.

[0120] Similarly, log odds ratio estimates from another generalized linear mixed effects model (GLMM_histopathology) for assessing the association with histopathology revealed that the proportion of cells in both clusters 17 and 25 showed a significant negative association with hippocampal volume, while showing a significant positive association with the coverage area of pTau, astrogliosis, and microgliosis. This association was mainly driven by the PS19-fE4 group and to a lesser extent by the PS19-fE4 / GFAP-Cre group. The proportion of cells in cluster 17 also had a significant negative association with the MBP coverage area.

[0121] In summary, these findings indicate that the removal of neuronal APOE4 rather than astrocytic APOE4 results in a dramatic decrease in the amounts of oligodendrocyte and neuron subpopulations related to neurodegenerative diseases in the hippocampus, and that these two cell populations have a strong association with tau pathology, gliosis, neurodegeneration, and oligodendrocyte deficits.

[0122] Example 10: Subpopulations of astrocytes and microglia related to neurodegenerative diseases are largely excluded by removing neuronal APOE4, but not by removing astrocytic APOE4.

[0123] Further subclustering of astrocytes identified 18 subpopulations. Log odds ratio estimates from GLMM_AM revealed that astrocyte subclusters 4 and 7 had lower odds of having cells from PS19-fE4 / Syn1-Cre mice than from PS19-fE4 mice, and that subcluster 7 was completely excluded in PS19-fE4 / Syn1-Cre mice. Interestingly, astrocyte subclusters 4 and 7 highly expressed APOE. Differential expression (DE) gene and pathway analysis revealed enrichment of KEGG pathways related to general neurodegeneration, Alzheimer's disease, and other neurodegenerative diseases in subcluster 7, indicating that subcluster 7 represents neurodegeneration disease-related astrocytes. Subcluster 4 showed enrichment of KEGG pathways related to cAMP signaling, synaptic function, and long-term potentiation, suggesting that subcluster 4 represents synapse dysfunction-related astrocytes.

[0124] Further subclustering of microglia also identified 18 subpopulations. Log odds ratio estimates from GLMM_AM revealed that microglia subclusters 4 and 8 had lower odds of having cells from PS19-fE4 / Syn1-Cre mice than from PS19-fE4 mice, and that subcluster 8 was completely excluded in PS19-fE4 / Syn1-Cre mice. Interestingly, microglia subclusters 4 and 8 highly expressed APOE. Differential expression (DE) gene and pathway analysis revealed enrichment of KEGG pathways related to general neurodegeneration, AD, and other neurodegenerative diseases in subcluster 8, indicating that subcluster 8 represents neurodegeneration disease-related microglia. Subcluster 4 showed enrichment of KEGG pathways related to synaptic function, cAMP signaling, and long-term potentiation, suggesting that subcluster 4 represents synapse dysfunction-related microglia.

[0125] Similarly, the log odds ratio estimates from GLMM_histopathological examinations revealed that the proportion of cells in both astrocyte subcluster 7 and microglia subcluster 8 showed a significant negative association with hippocampal volume, while showing a significant positive association with pTau coverage area, astrogliosis, and microgliosis. This association was mainly driven by the PS19-fE4 group and, to a lesser extent, by the PS19-fE4 / GFAP-Cre group. Interestingly, when oligodendrocyte cluster 17 and excitatory neuron cluster 25 were PCA clustered together with all astrocyte and microglia subclusters, subcluster 7 of disease-related astrocytes and subcluster 8 of microglia were shown to have similar contributions to the eight pathological parameters, similar to the two disease-related clusters of oligodendrocytes (cluster 17) and neurons (cluster 25).

[0126] In summary, these findings indicate that the removal of neuronal APOE4 rather than astrocytic APOE4 results in a dramatic reduction in the amount of neurodegenerative disease-related astrocyte and microglia subpopulations in the hippocampus, and that these two glial subpopulations have a strong association with tau pathology, gliosis, and neurodegeneration.

[0127] Example 11: Removal of astrocytic APOE4 secondarily reduces neuronal APOE4 expression During the analysis of the snRNA-seq data, the inventors of the present application noticed that the removal of astrocytic APOE also dramatically decreased APOE expression in neurons in addition to its predicted elimination in astrocytes. To confirm this observation, a publicly available snRNA-seq dataset (GEO: GSE164507) using a similar complex mouse model was further analyzed. This model holds a floxed knock-in type human APOE gene and an astrocyte-specific Cre (Aldh1l1-Cre / ERT2 BAC transgene) that deletes the human APOE gene in astrocytes in a tamoxifen-inducible manner during adulthood, and the presence or absence of the P301S mutant human Tau is incorporated (Wang et al., Neuron, 109, 1657-1674.e7 (2021)).

[0128] In this snRNA-seq dataset, neuronal APOE expression was observed in both FE3 mice and FE4 mice, which was further increased in PS19-FE4 (TAFE4_oil) mice and PS19-FE3 (TAFE3_oil) mice, indicating that tauopathy increases neuronal APOE expression. Removal of astrocytic APOE induced by tamoxifen treatment during adulthood also clearly observed a dramatic decrease in neuronal APOE in both excitatory and inhibitory neurons in PS19-fE4 / Aldh1l1-Cre / ERT2 (TAFE4_tam) mice and PS19-fE3 / Aldh1l1-Cre / ERT2 (TAFE3_tam) mice compared to mice treated with oil. The effect of the removal of astrocytic APOE on the decrease in neuronal APOE was observed in all subtypes of neurons, suggesting a general regulation of neuronal APOE expression by astrocytic APOE.

[0129] Therefore, the initial observation that the removal of astrocytic APOE secondarily reduces neuronal APOE expression is confirmed in different tauopathy models using human APOE, even when the removal of astrocytic APOE is driven under different astrocyte-specific promoters and its removal is induced at different time points in the lifespan of the mouse model.

[0130] To confirm whether this observation is verifiable at the protein level, the levels of APOE protein were quantitatively evaluated by sandwich ELISA using primary cultured neurons isolated from the cerebral cortex and hippocampus of prenatal mouse fetuses. Analysis revealed that primary neurons from PS19-fE4 / GFAP-Cre mice showed a significant 50% decrease in APOE protein levels compared to those from PS19-fE4 mice (Figure 4M).

[0131] These data provide strong evidence that the removal of astrocytic APOE4 results in a dramatic decrease in neuronal APOE4 expression and protein production, strengthening this regulatory relationship between astrocytic APOE4 and neuronal APOE4. Therefore, the limited protective effect of removing astrocytic APOE4 is likely due to a secondary decrease rather than a complete elimination of neuronal APOE4.

[0132] Initial studies by the inventors and co-researchers of the present application reported that neuronal APOE expression is regulated by astrocyte-released factor(s) via the regulation of the ERK pathway within neurons (Zhao et al., Nat. Commun., 11, 5540 (2020)). A detailed comparison of snRNA-seq data between astrocytes from PS19-fE4 / GFAP-Cre mice and astrocytes from PS19-fE4 mice revealed that several genes significantly downregulated after the removal of astrocytic APOE4 are known documented activators of the ERK / MAPK pathway, including dipeptidyl peptidase 10 (DPP10) and neuregulin-3 (Nrg3).

[0133] The inventors of the present application hypothesized that astrocytic APOE4 could regulate neuronal APOE4 expression by promoting the release of these factors that activate the neuronal ERK pathway, resulting in an increase in neuronal APOE4 expression. To examine this possibility, primary neurons isolated from PS19-fE4 mice and PS19-fE4 / GFAP-Cre mice were treated with 10 μg / mL of recombinant DPP10 or Nrg3 protein or dPBS control. Subsequently, APOE protein levels were quantitatively evaluated in primary neuron cell lysates by sandwich ELISA.

[0134] PS19-fE4 / GFAP-Cre primary neurons treated with DPP10 showed a significant 98% increase in APOE protein levels compared to the dPBS control (Figure 4N). On the other hand, there was no significant difference in APOE levels in PS19-fE4 primary neurons after DPP10 treatment (Figure 4O). These data indicate that primary neurons from PS19-fE4 / GFAP-Cre mice in which astrocytic APOE4 was removed are sensitive to DPP10 upregulation of APOE expression, while primary neurons from PS19-fE4 mice are not. The inventors of the present application did not observe a difference in APOE levels in PS19-fE4 / GFAP-Cre or PS19-fE4 primary neurons after treatment with Nrg3 (Figures 4N - 4O), indicating that Nrg3 is not involved in the regulation of neuronal APOE4 expression.

[0135] Example 12: Release of neuronal HMGB1 is completely suppressed by removing neuronal APOE4, not astrocytes In the search for inflammatory cytokines that can induce tau-induced and neurodegeneration-induced gliosis, the inventors of the present application observed that cells in neurodegenerative disease-related neuron cluster 25 dramatically increased the expression of the high-mobility group box 1 (Hmgb1) gene. The HMGB1 protein is required for glial cell activation (Paudel et al., Front. Neurosci., 12, 628 (2018)). Under pathological conditions, HMGB1 moves from the nucleus of stressed or dying cells to the cytoplasm and is then released to act as an inflammatory cytokine.

[0136] In 10-month-old mice, immunohistochemical staining for HMGB1 protein and the nuclear marker DAPI confirmed that most of the proteins in hippocampal neurons were outside the nucleus, revealing that in PS19-fE4 mice, the amount of HMGB1 protein translocating from the nucleus to the cytoplasm in the neurons of the hippocampal dentate gyrus was significantly higher. (Figs. 5A - 5D). On the other hand, in PS19-fE3 mice, most of the HMGB1 protein was retained in the nucleus and the translocation to the cytoplasm was minimal (Figs. 5A - 5D). Interestingly, removing APOE4 from either neurons or astrocytes attenuated this phenotype observed in PS19-fE4 mice, and most of the HMGB1 protein remained localized in the nucleus (Figs. 5A - 5D).

[0137] Furthermore, PS19-fE4 mice contained a greater number of HMGB1 - positive puncta in both the nucleus and cytoplasm compared to other genotype groups (Figs. 5E - 5F). Notably, fE4 and fE3 mice lacking human mutant tau - P301S had HMGB1 protein localized in the nucleus, indicating that the nuclear - cytoplasmic transport of HMGB1 in APOE4 mice requires the co - existence of both APOE4 and tauopathy.

[0138] To evaluate the extent of disease-related HMGB1 release from hippocampal neurons within these various genotypic groups, hippocampal interstitial fluid (ISF) from 8.5-month-old mice was collected over 24 hours using in vivo microdialysis (Yamada et al., J. Neurosci., 31:13110-13117 (2011)). The levels of HMGB1 protein in the interstitial fluid (ISF) were quantitatively determined using sandwich ELISA. PS19-fE4 mice showed high levels of HMGB1 protein within their hippocampal ISF, but removal of neuronal APOE4 decreased HMGB1 protein levels to undetectable levels in the ISF (Figures 5G-5H), indicating that neuronal APOE4 plays a role in controlling HMGB1 release from neurons. Removal of astrocytic APOE4 also resulted in an 86% decrease (not complete elimination) of HMGB1 protein in the ISF (Figures 5G-5H), which is likely due to a secondary decrease (not complete elimination) in neuronal APOE4 expression.

[0139] Experiments were also conducted to determine whether HMGB1 is a potential regulator of neuronal APOE4 expression. Thus, primary neurons isolated from PS19-fE4 / GFAP-Cre and PS19-fE4 mice were treated with 10 μg / mL of recombinant HMGB1 protein or dPBS control, and APOE protein levels were quantitatively evaluated in cell lysates by sandwich ELISA. No significant differences were observed in the APOE protein levels of primary neurons from PS19-fE4 / GFAP-Cre or PS19-fE4 after treatment with HMGB1, indicating that HMGB1 is not involved in the regulation of neuronal APOE4 expression.

[0140] To evaluate the importance of neuron-released HMGB1 in the induction of gliosis in APOE-related tauopathy, interstitial fluid (ISF) collected from the hippocampus of 8.5-month-old PS19-fE4 mice was injected into the hippocampus of 8.5-month-old wild-type mice. The injected ISF fractions were either enriched with relatively high levels of HMGB1 protein (fractions 19-22 in Fig. 5H) as determined by sandwich ELISA or had undetectable levels of HMGB1 protein (fractions 4-7 in Fig. 5H) as a control. Wild-type mice were unilaterally injected with either HMGB1 + or control ISF into the right dorsal hippocampus and analyzed on day 6 after injection to evaluate acute changes in gliosis.

[0141] Wild-type mice injected with control (HMGB1 - ) ISF showed relatively low levels of microgliosis and astrogliosis on both the injected and non-injected hippocampal sides, indicating that most of the gliosis induced by the injection surgery had resolved by 6 days after injection (Figs. 5I-5J). In contrast, wild-type mice injected with HMGB1 + ISF showed significantly increased levels of microgliosis on the injected hippocampal side compared to the non-injected side (Fig. 5I). Interestingly, injection of HMGB1 + ISF did not result in a significant increase in astrogliosis on the injected hippocampal side compared to the non-injected side (Fig. 5J). Collectively, these data indicate that neuron-derived release of HMGB1 represents a novel mechanism by which neuronal APOE4 promotes glial cell activation.

[0142] Example 13: Treatment with an HMGB1 inhibitor substantially reduces APOE4-driven gliosis, tau pathology, and degeneration Based on the finding that neuronal APOE4 is a potent driver of neuron-derived HMGB1 release, experiments were conducted to verify the therapeutic effect of an HMGB1 inhibitor in counteracting APOE4-driven pathologies associated with tauopathy.

[0143] Two HMGB1 inhibitors, ethyl pyruvate (EP) and glycyrrhizic acid (GA), were tested. These are selective inhibitors of HMGB1 translocation and release (Ulloa et al., Proc. Natl. Acad. Sci. USA, 99:12351 - 12356 (2002); Sun et al., Front. Immunol., 9, 1518 (2018); Mollica et al., Chem Biol., 14, 431 - 41 (2007); Dave et al., J. Leukoc. Biol., 86, 633 - 643 (2009)). A mixed solution of ethyl pyruvate (80 mg / kg) and glycyrrhizic acid (20 mg / kg) or a saline vehicle was administered to PS19 - fE4 mice and PS19 - fE3 mice via intraperitoneal injection in a regimen of 3 doses per week for 12 weeks (Figure 6A). As demonstrated in this study and other studies (Shi et al., Nature, Vol. 549, pp. 523 - 527 (2017); Yoshiyama et al., Neuron, Vol. 53:337 - 351 (2007)), treatment was initiated when the mice reached 6.5 months of age, when the harmful pathology was almost developed, and completed when the mice reached 9.5 months of age, when severe neurodegeneration and pathological changes were typically present.

[0144] The effectiveness of ethyl pyruvate (EP) and glycyrrhizic acid (GA) on blocking the nuclear - cytoplasmic transport of HMGB1 protein was first confirmed by immunostaining using anti - HMGB1 and DAPI. PS19 - fE4 mice treated with saline showed extensive translocation of HMGB1 from the nucleus, while PS19 - fE4 mice treated with HMGB1 inhibitors showed no distinguishable HMGB1 translocation (Figures 6B - 6D; 6P - 6Q). In PS19 - fE3 mice, there was no significant difference in HMGB1 translocation between the saline - treated group and the inhibitor - treated group.

[0145] Next, the degree of microgliosis and astrogliosis present in the treated mice was evaluated. PS19 - fE4 mice treated with saline had Iba1 +Microglia (Figs. 6E - 6F) and CD68 + As exemplified by the high coverage area of activated microglia (Fig. 9R), significant microgliosis was shown throughout the hippocampus. On the other hand, PS19 - fE4 mice treated with the HMGB1 inhibitor showed a significant decrease in microgliosis. Furthermore, immunostaining with anti - GFAP revealed that PS19 - fE4 mice treated with saline had extensive astrogliosis throughout the hippocampus, while PS19 - fE4 mice treated with the HMGB1 inhibitor had a significant decrease in astrogliosis (Figs. 6G - 6H). PS19 - fE3 mice showed no significant difference in microgliosis or astrogliosis whether treated with saline or the HMGB1 inhibitor.

[0146] Next, the degree of tau pathology was evaluated in these mice by immunostaining for p - tau. PS19 - fE4 mice treated with saline showed substantial tau pathology throughout the hippocampus, while PS19 - fE4 mice treated with the HMGB1 inhibitor showed a dramatic decrease in tau pathology (Figs. 6I - 6J). Furthermore, immunostaining for MBP was used to determine the effect of administration of the HMGB1 inhibitor on myelin deficiency. PS19 - fE4 mice treated with saline showed severe myelin loss in the radiate layer of CA1, while PS19 - fE4 mice treated with the HMGB1 inhibitor had a high coverage area of MBP and showed recovery of the deficiency (Figs. 6K - 6L). There was no distinguishable difference in myelin coverage area or tau pathology between PS19 - fE3 mice treated with saline and those treated with the inhibitor.

[0147] The effect of HMGB1 inhibitors on neurodegeneration was also determined by quantifying hippocampal volume and posterior ventricular volume. Saline-treated PS19-fE4 mice showed significant neurodegeneration, while HMGB1 inhibitor-treated PS19-fE4 mice showed recovery of neurodegeneration by a significant increase in hippocampal volume and a significant decrease in posterior ventricular volume compared to saline-treated PS19-fE4 mice (Figures 6M - 6O). There was no obvious difference in neurodegeneration between saline-treated PS19-fE3 mice and HMGB1 inhibitor-treated PS19-fE3 mice.

[0148] Collectively, these findings indicate that treatment of PS19-fE4 mice with HMGB1 inhibitors can effectively prevent the development of prominent APOE4-driven pathologies, including gliosis, tau pathology, neurodegeneration, and myelin deficits, by blocking nucleo-cytoplasmic transport and subsequent release of HMGB1 protein. The data support the conclusion that the translocation and release of HMGB1 play a role in the induction and exacerbation of these various APOE4-driven pathologies.

[0149] Example 14 Apolipoprotein E4 (APOE4) is a major genetic risk factor for Alzheimer's disease (AD), but the underlying cellular and molecular mechanisms remain elusive. In the brain, APOE is produced mainly by astrocytes and at lower levels by neurons. Described herein is a rigorous comparison of the neuronal and astrocytic APOE4 effects on AD-related pathologies by selectively removing APOE4 from either cell type in mice carrying the human P301S mutant tau transgene and the floxed knock-in human APOE gene. APOE4 mice had significantly more tau pathology, gliosis, neurodegeneration, neurological dysfunction, and myelin deficits compared to APOE3 mice. Single-nucleus RNA sequencing identified an enrichment of neurodegenerative disease-related subpopulations of neurons, oligodendrocytes, astrocytes, and microglia in APOE4 mice. Removal of neuronal APOE4 dramatically reduced all of these pathologies and substantially eliminated the disease-related cell subpopulations, whereas removal of astrocytic APOE4 reduced only gliosis and neurodegeneration. The limited protective effect of removing astrocytic APOE4 is due to a secondary reduction rather than a complete elimination of neuronal APOE4. Mechanistically, neuronal APOE4 promotes the release of HMGB1 from neurons, inducing gliosis and subsequent neurodegeneration and myelin deficits, which was effectively blocked by treatment with an HMGB1 inhibitor. Thus, neuronal APOE4 drives tau-mediated inflammation and degeneration by promoting neuronal HMGB1 release, and HMGB1 inhibitors represent a novel approach for treating APOE4-related AD and other tauopathies.

[0150] Introduction Tauopathy is a class of neurodegenerative disorders defined by the abnormal intracellular accumulation of hyperphosphorylated tau (p-tau) protein. Alzheimer's disease (AD), which represents the main type of tauopathy, is a very prevalent disorder that is pathologically characterized by the accumulation of amyloid plaques and tau neurofibrillary tangles and clinically characterized by the loss of memory and other cognitive functions. Of these two main pathological features of AD, only tau neurofibrillary tangles strongly correlate with neurodegeneration and cognitive decline. Other pathological features of AD under investigation include neuroinflammation and gliosis, which have recently been identified as major contributing factors to neurodegeneration. Furthermore, oligodendrocyte deficits and myelin degeneration have been observed in the human AD brain as well as in mouse models of AD and tauopathy. Thus, AD represents a complex series of pathologies, but the relationships between these pathologies remain unclear. Understanding the relationships between the pathologies and elucidating the underlying mechanisms that cause their induction or exacerbation are important for developing better treatment strategies that target these pathologies individually or in combination.

[0151] Epidemiological studies and genome-wide association studies have identified apolipoprotein E4 (APOE4) as a major genetic risk factor for AD. The human APOE gene has three common alleles, including ε2, ε3, and ε4. APOEε4 is considered the most deleterious allele as it increases the AD risk in a dose-dependent manner and decreases the disease onset age. Considerable efforts have been made to understand how APOE4 causes an increased AD risk, and a wide range of studies have shown that APOE4 exacerbates many prominent AD-related pathologies compared to APOE. In particular, APOE4 has been shown to accelerate hippocampal volume loss in human patients and increase neurodegeneration in mice with or without tauopathy. In addition to its well-studied role in promoting amyloid pathology, recent studies have found that APOE4 also increases tau burden in the human brain and promotes the accumulation of p-tau in mouse and human neuron models. Furthermore, APOE4 increases neuroinflammation and gliosis in human AD brains and tauopathy mouse models. APOE4 has also been reported to be associated with reduced myelination and decreased white matter integrity in the human brain. Collectively, these studies provide clear evidence that APOE4 is involved in promoting tau pathology, gliosis, neurodegeneration, and myelin degeneration in AD and other tauopathies. Nevertheless, the underlying mechanisms involved in the widespread effects of APOE4 on these various pathologies remain unknown.

[0152] Recently, there has been growing interest in establishing the cell-type specific effects of APOE4 in various aspects of the etiology of AD. Within the central nervous system (CNS), APOE acts as a primary lipid transporter and is mainly produced by astrocytes. Conditions of stress or injury can induce APOE expression in neurons and microglia. Previous studies have shown that the detrimental effects of APOE4 can depend on its cell source. However, there is a gap in the understanding of the exact role of neuronal and astrocytic APOE4 in the induction and / or exacerbation of various AD pathologies. In this study, to better understand the cell-type specific role of APOE4 in the etiology of AD and other tauopathies, a comprehensive analysis was performed to directly compare the effects of the removal of APOE4 from neurons or astrocytes on the development of major AD pathologies, including tau pathology, gliosis, neurodegeneration, neural dysfunction, and myelin loss. The aim was to elucidate the important cellular and molecular mechanisms involved in the induction of the AD-related pathology cascade and to determine whether these mechanisms are regulated by neuronal or astrocytic APOE4.

[0153] Results Neuronal or astrocyte-specific removal of the APOE gene in human APOE-KI mice expressing human mutant tau A mouse strain was generated in which the floxed human APOE3 or APOE4 gene and the Cre recombinase gene are expressed under the control of the neuron-specific synapsin-1 promoter (Syn1-Cre) or the astrocyte-specific glial fibrillary acidic protein promoter (GFAP-Cre). These floxed APOE-KI (fE) mice express homozygous human APOE3 or APOE4 instead of the endogenous mouse Apoe, and the human APOE gene is arranged in a form flanked by a pair of LoxP sites to enable its precise excision in the presence of cell-type specific Cre recombinase expression. fE mice without Cre, with Syn1-Cre, or with GFAP-Cre were mated with mice (PS19 strain) that express mutant 1N4R human microtubule-associated protein tau (MAPT) encoding the disease-related P301S mutation, which is widely used as a tauopathy mouse model. The resulting compound mice are designated as PS19-fE, PS19-fE / Syn1-Cre, or PS19-fE / GFAP-Cre mice.

[0154] We performed a rigorous characterization of fE / Syn1-Cre mice and fE / GFAP-Cre mice to verify the specificity of Cre recombinase expression under the neuron-specific Syn1 or astrocyte-specific GFAP promoter. We further evaluated PS19-fE / Cre mice by performing immunohistochemical analysis of brain sections from 10-month-old PS19-fE4 / Syn1-Cre mice and PS19-fE4 / GFAP-Cre mice. To confirm the specificity of Cre recombinase expression in hippocampal neurons and astrocytes when driven under the Syn1 or GFAP promoter, respectively, NeuN and GFAP antibodies were used together with the Cre recombinase antibody. In PS19-fE4 / Syn1-Cre mice, Cre recombinase was expressed only in NeuN-positive neurons and not in GFAP-positive astrocytes (Figure 7A). In PS19-fE4 / GFAP-Cre mice, Cre recombinase was exclusively expressed in GFAP-positive astrocytes and not in NeuN-positive neurons (Figure 7A). Co-immunostaining using antibodies against APOE, GFAP, and NeuN showed that PS19-fE4 / Syn1-Cre mice lacked APOE expression in neurons but had APOE expression in astrocytes, while PS19-fE4 / GFAP-Cre mice had APOE expression in some neurons but lacked APOE expression in astrocytes, compared with Cre-negative PS19-fE4 mice, which had APOE expression in GFAP-positive astrocytes and some NeuN-positive neurons (Figure 7B). PS19-fE4 / GFAP-Cre mice were negative for NeuN and GFAP (Figure 7B) but positive for the microglial marker Iba1, showing some APOE-positive cells that were APOE-expressing microglia (Figure 7C).

[0155] To quantitatively determine the levels of APOE protein in these various mouse models, hippocampal lysates from 10-month-old mice were analyzed by sandwich ELISA for human APOE. PS19-fE4 / Syn1-Cre mice showed a ~20% decrease in APOE levels compared to PS19-fE4 mice (Figure 7D), which is consistent with previous reports indicating that neuronal APOE contributes ~20–30% of the total APOE protein levels in the hippocampus and cortex26,42. PS19-fE4 / GFAP-Cre mice showed a dramatic ~70% decrease in APOE levels compared to PS19-fE4 mice, which is consistent with the well-established role of astrocytes as the major producers of APOE within the CNS. Similarly, compared to PS19-fE3 mice, PS19-fE3 / Syn1-Cre mice showed a ~25% decrease and PS19-fE3 / GFAP-Cre mice showed a ~67% decrease in APOE levels (Figure 7D). Collectively, these results provide strong evidence that in these compound mouse models, APOE gene expression is eliminated in neurons or astrocytes when Cre recombinase expression is driven under the Syn1 or GFAP promoters, respectively.

[0156] Removal of neuron-derived, rather than astrocyte-derived, APOE4 dramatically reduces tau pathology To determine whether the removal of APOE from neurons or astrocytes affects tau pathology, 10-month-old mice were evaluated when PS19 mice showed extensive tau pathology throughout the hippocampus. The accumulation of p-tau in the hippocampus was evaluated by immunohistochemical staining using a p-tau specific AT8 antibody, and tau pathology was quantified as the percentage of the AT8-covered area in the hippocampus. PS19-fE4 mice showed extensive tau pathology throughout the hippocampus, and the degree of tau pathology was significantly lower in PS19-fE3 mice. Compared with PS19-fE4 mice, PS19-fE4 / Syn1-Cre mice showed a significant decrease in tau pathology (about 81%), while PS19-fE4 / GFAP-Cre mice showed a slight decrease (about 30%) that did not reach statistical significance. Interestingly, the decrease in the p-tau covered area in PS19-fE4 / Syn1-Cre mice was similar to the degree of tau pathology observed in PS19-fE3 mice. There was no significant difference in tau pathology between PS19-fE3 with Cre and PS19-fE3 without Cre, probably because the tau pathology in PS19-fE3 mice was already low.

[0157] Western blot analysis was used to evaluate the levels of p-tau in mouse hippocampal tissues after sequential biochemical extraction with RAB buffer and RIPA buffer, which contain highly soluble tau protein and less soluble tau protein, respectively. There was no significant difference in p-tau levels among various genotype groups in the RAB fraction. However, PS19-fE4 / Syn1-Cre mice and PS19-fE3 mice showed a significant decrease in p-tau levels in the RIPA fraction compared with PS19-fE4 mice. PS19-fE4 / GFAP-Cre mice showed a tendency for a decrease in p-tau levels in the RIPA fraction (about 40%) compared with PS19-fE4 mice, which did not reach statistical significance. Collectively, these data indicate that APOE4 expression in neurons is a strong driver of tau pathology, while APOE4 in astrocytes has a minimal effect on tau pathology.

[0158] The propagation of tau pathology decreases after the removal of APOE4 in neurons but not in astrocytes. To further investigate the mechanism by which neuronal APOE4 drives tau pathology, we determined the effect of cell-type specific APOE4 expression on the propagation of tau pathology. Previous studies have shown that pathological tau can spread to anatomically connected brain regions after direct injection of various forms of tau protein into the mouse brain. We analyzed the extent of tau propagation after single unilateral injection of adeno-associated virus-2 encoding human P301S mutant tau (AAV2-tau-P301S) into the right dorsal hippocampus of fE mice that do not have Cre, have Syn1-Cre, or have GFAP-Cre. fE mice lack human P301S mutant tau and instead express the endogenous mouse Mapt gene, and since fE mice show minimal tau pathology, this allows for more accurate detection of human tau spread. Mice were injected with AAV2-tau-P301S virus at 10 months of age and evaluated at 13 months of age, 12 weeks after injection.

[0159] In contrast to the unexpected result that tau moves to the non-injected side of the hippocampus where the virus encoding tau itself was not injected, to provide evidence that the observed tau propagation phenotype could truly be due to the spread of pathological human tau between neurons, we tested the unilateral injection of AAV2 of the same serotype encoding GFP (AAV2-GFP) into the right dorsal hippocampus of 10-month-old fE4 mice. Immunostaining with anti-GFP 2 weeks after injection revealed that the GFP signal remained localized to neurons within the injected side of the hippocampus (Figures 8A–B). The non-injected side of the hippocampus did not have any obvious GFP signal in neuronal cell bodies, but there were some GFP-positive neuronal projections that were likely to have originated from neurons present on the injected side of the hippocampus. This indicates that AAV2 itself does not spread between the right and left hippocampi after unilateral injection. Furthermore, for all immunohistochemical analyses, the number of neurons containing cell body-positive tau was quantified to more accurately reflect the spread of human mutant tau between neurons, and cross-reactive factors, such as tau-positive cross-linked fibers from neurons originating from the injected side, were excluded.

[0160] Immunohistochemical staining with the human tau antibody (HT7) revealed strong human tau propagation to the non-injected hippocampal side in PS19-fE4 mice and minimal tau propagation in PS19-fE3 mice, indicating that APOE4 promotes tau diffusion. Interestingly, PS19-fE4 / Syn1-Cre mice had limited human tau propagation to the non-injected hippocampal side, while PS19-fE4 / GFAP-Cre mice had extensive human tau propagation similar to the spread of tau pathology observed in PS19-fE4 mice. In particular, there was a 55% decrease in the number of HT7+ cells associated with tau propagation after removal of neuronal APOE4, and no distinguishable difference in the number of HT7+ cells associated with tau propagation after removal of astrocytic APOE4.

[0161] The degree of tau propagation was analyzed by immunostaining for p-tau using the AT8 antibody. Both PS19-fE4 mice and PS19-fE4 / GFAP-Cre mice showed strong propagation of p-tau to the non-injected hippocampal side, while PS19-fE4 / Syn1-Cre mice and PS19-fE3 mice dramatically reduced p-tau propagation. Collectively, these data indicate that one mechanism by which neuronal APOE4 drives tau pathology is by stimulating the propagation of pathological tau across synapse-connected neurons. This data also further supports the idea that astrocytic APOE4 is not an important mediator of tau pathology, as evidenced by its lack of effect on tau propagation.

[0162] Network hyperexcitability is eliminated after removal of neuronal APOE4, but not after removal of astrocytic APOE4 To determine the cell-type specific effects of APOE4 on neuron function in the context of tauopathy, neuronal network excitability in the hippocampal Cornu Ammonis (CA1) region of PS19-fE3 and PS19-fE4 mice with or without Cre, Syn1-Cre, or GFAP-Cre was measured by input-output curve analysis of network responses to incremental stimulation of the Schaffer collateral (Figure 9A). PS19-fE4 mice had significant CA1 neuron hyperexcitability compared to PS19-fE3 mice (Figure 9B). Removal of neuronal APOE4 eliminated neuronal network hyperexcitability, while removal of astrocytic APOE4 resulted in a slight decrease that did not reach significance (Figure 9B), indicating that neuronal APOE4 drives neurological dysfunction in relation to tauopathy.

[0163] Neurodegeneration decreases after removal of either neuronal APOE4 or astrocytic APOE4 Next, the degree of neurodegeneration in 10-month-old PS19-fE mice was evaluated after removing APOE from neurons or astrocytes. Analysis of hippocampal and posterior ventricular volumes revealed that PS19-fE4 mice showed extensive neurodegeneration compared to PS19-fE3 mice. Neurodegeneration significantly decreased in both PS19-fE4 / Syn1-Cre and PS19-fE4 / GFAP-Cre mice, as indicated by an increase in hippocampal volume and a decrease in posterior ventricular volume, similar to the increase in hippocampal volume and decrease in posterior ventricular volume in PS19-fE3 mice. Removal of APOE3 from neurons or astrocytes did not significantly affect neurodegeneration.

[0164] Quantification of the degree of neuronal cell loss within various subfields of the hippocampus revealed that PS19-fE4 mice had extensive neuronal loss in the CA1 and dentate gyrus, while both PS19-fE4 / Syn1-Cre mice and PS19-fE4 / GFAP-Cre mice had significantly reduced neuronal cell loss in both the CA1 and dentate gyrus subfields of the hippocampus. When APOE3 was removed from neurons or astrocytes, there was no significant difference in the thickness of the neuronal cell layer in PS19-fE3 mice.

[0165] In PS19-fE4 mice, there was a weak but significant negative correlation between tau pathology and hippocampal volume, suggesting that tau pathology contributes, at least to some extent, to the neurodegeneration occurring in these mice. There was also a strong negative correlation between hippocampal volume and the volume of the posterior ventricle, and a strong positive correlation between the thickness of the CA1 cell layer and hippocampal volume. Overall, these data indicate that the removal of APOE4 from either neurons or astrocytes is protective against tau-mediated neurodegeneration and results in a reduction in neuronal cell and hippocampal volume loss. This result is interesting considering tau pathology data indicating that astrocytic APOE4 has a minimal effect on the accumulation and propagation of tau pathology.

[0166] Gliosis dramatically decreases after the removal of APOE4 from either neurons or astrocytes Next, the degree of microgliosis and astrogliosis within these various genotype groups at 10 months of age was investigated to examine the role of APOE4 in neurons or astrocytes in inducing gliosis in relation to tauopathy. Immunohistochemical staining of microglia using an Iba1 antibody revealed that PS19-fE4 mice had extensive microgliosis in the hippocampus, as demonstrated by a high percentage of Iba1-covered area. Both PS19-fE4 / Syn1-Cre mice and PS19-fE4 / GFAP-Cre mice showed a dramatic reduction in microgliosis. Furthermore, all PS19-fE3 mice without Cre, with Syn1-Cre, or with GFAP-Cre had a significantly reduced microgliosis compared to PS19-fE4 mice. There was a strong negative correlation between the Iba1-covered area and the hippocampal volume in PS19-fE4 mice, indicating that microgliosis is a good indicator of neurodegeneration and a potential contributing factor.

[0167] This microgliosis phenotype was further defined by immunohistochemical staining of CD68, a marker of activated microglia. There were dramatic differences in the covered area of activated microglia among the genotype groups, with PS19-fE4 mice showing extensive microglial activation and all other genotype groups showing minimal microglial activation. Interestingly, among all the pathological correlations made in PS19-fE4 mice, the covered area of CD68+ activated microglia had the strongest negative correlation with the hippocampal volume, suggesting that the degree of microglial activation is the strongest indicator of APOE4-enhanced neurodegeneration in tauopathy and a potential contributing factor. Collectively, these data indicate that APOE4 strongly enhances microgliosis compared to APOE3 in tauopathy, and removal of APOE4 from either neurons or astrocytes weakens the degree of microgliosis.

[0168] The degree of astrogliosis after removing APOE from neurons or astrocytes was evaluated. PS19-fE4 mice showed significant astrogliosis in the hippocampus compared to PS19-fE3 mice, as detected by immunohistochemical staining with an astrocyte GFAP antibody. The degree of astrogliosis decreased significantly after removing either neuronal APOE4 or astrocytic APOE4. These observations were also supported by immunohistochemical staining for S100, an activated astrocyte marker. There was no obvious difference in the degree of astrogliosis after removing APOE3 from either cell type compared to PS19-fE3 mice. Surprisingly, and in marked contrast to microgliosis, neither the astrocyte coverage area nor the activated astrocyte coverage area correlated significantly with hippocampal volume. These data indicate that APOE4 potently enhances astrogliosis compared to APOE3 in tauopathy, and that removal of APOE4 from either neurons or astrocytes reduces astrocyte activation and / or mobilization. Furthermore, it is also shown that the degree of astrogliosis may not be a beneficial indicator or contributing factor for APOE4-enhanced neurodegeneration in tauopathy.

[0169] To further clarify the relationship between tau pathology and gliosis, we examined whether tau propagation induces gliosis even in the non-injected hippocampus side. Indeed, the inventors of the present application observed that fE4 mice showed higher microgliosis in the non-injected hippocampus side than fE3 mice, along with higher tau diffusion (Figs. 8C-D). Interestingly, after removing either neuronal APOE4 or astrocytic APOE4, there was a significant decrease in microgliosis in the non-injected hippocampus side. Since fE4 / GFAP-Cre mice still had a relatively high level of tau diffusion in the non-injected hippocampus side, this indicates that the removal of neuronal APOE4 or astrocytic APOE4 reduces microgliosis regardless of the degree of tau diffusion. These findings suggest that astrocytic APOE4 acts downstream of tau pathology to promote microgliosis. Furthermore, the inventors of the present application evaluated the degree of astrocytic coverage area in the non-injected hippocampus side, and no distinguishable difference in astrogliosis was observed after removing neuronal or astrocytic APOE4, probably due to the relatively low astrogliosis in the non-injected hippocampus side even in fE4 mice in this tau diffusion model (Figs. 8E-F).

[0170] Myelin deficiency and depletion of oligodendrocyte progenitor cells are reduced after removal of APOE4 from neurons, but not after removal of APOE4 from astrocytes Since myelin degeneration and oligodendrocyte loss have been observed in human AD brains as well as in mouse models of AD and tauopathy, we investigated the effects of APOE4 removal from neurons and astrocytes on myelin integrity and the density of hippocampal oligodendrocyte progenitor cells (OPCs). To determine myelin integrity in the hippocampus, myelin basic protein (MBP) was stained with an MBP-specific antibody. Quantification of the percentage of MBP-covered area in the stratum radiatum beneath the CA1 pyramidal cell layer revealed that PS19-fE4 mice had extensive myelin loss compared to PS19-fE3 mice (Figures 10A–B). Surprisingly, a significant reduction in myelin loss in PS19-fE4 / Syn1-Cre mice was similar to the phenotype of PS19-fE3 mice, whereas PS19-fE4 / GFAP-Cre mice showed substantial myelin loss similar to the phenotype of PS19-fE4 mice (Figures 10A–B). This indicates that neuronal APOE4 plays a major detrimental role in promoting myelin loss in the context of tauopathy, while astrocytic APOE4 does not affect this phenotype. There was no substantial difference in myelin integrity in PS19-fE3 mice after removal of APOE3 from neurons or astrocytes.

[0171] To further characterize the effects of neuronal APOE4 and astrocytic APOE4 on biological components involved in the myelin formation process, NG2 antibodies were used to probe for OPCs (which have been suggested to assist in the repair of damaged myelin in CNS injury and neurodegenerative conditions). Compared to PS19-fE3 mice, there was a significant decrease in the percentage of OPC coverage area in the hippocampus of PS19-fE4 mice (Figures 10C - D). Surprisingly, the removal of neuronal APOE4 significantly increased the percentage of OPC coverage area to levels similar to those in PS19-fE3 mice (Figures 10C - D). In contrast, PS19-fE4 / GFAP-Cre mice had a significantly lower OPC coverage area compared to PS19-fE4 / Syn1-Cre mice and showed a tendency for decreased OPCs compared to PS19-fE4 mice, but did not reach statistical significance (Figures 10C - D). After the removal of neuronal or astrocytic APOE3 in PS19-fE3 mice, there was no obvious difference in the OPC coverage area (Figures 10C - D). Immunohistochemical staining for myelin and OPCs in fE4 and fE3 mice revealed that APOE4 mice lacking human mutant tau-P301S did not show myelin deficiency and had OPC levels similar to fE3 mice in the hippocampus, indicating that the effect of APOE4 on these phenotypes is specific to the context of tauopathy (Figures 10E - F). Overall, these findings suggest that neuronal APOE4 plays a central role in depleting the hippocampal OPC pool and causing myelin deficiency in this complex tauopathy mouse model, while astrocytic APOE4 does not significantly contribute to either of these two pathological processes.

[0172] snRNA-seq identified subpopulations of neurons and oligodendrocytes associated with neurodegenerative diseases, and most of these were eliminated by removing neuronal APOE4 rather than astrocytic APOE4 To gain a deeper understanding of the cell-type specific effects of APOE4 at the transcriptome level across different types of hippocampal cells, single-nucleus RNA sequencing (snRNA-seq) was performed on hippocampi isolated from 10-month-old PS19-fE4 and PS19-fE3 mice with no Cre, Syn1-Cre, or GFAP-Cre. The snRNA-seq dataset contained 119,317 nuclei covering 26,285 genes after normalization and filtering for quality control. Clustering by shared nearest neighbor (SNN) and visualization by uniform manifold approximation and projection (UMAP) revealed 33 different cell clusters. Based on their expression of marker genes, these clusters were assigned as follows: 16 excitatory (Ex) neuron clusters (3 - 5, 7, 12, 15, 16, 18 - 20, 23, 25 - 28, 33); 7 inhibitory (In) neuron clusters (6, 8, 9, 13, 22, 24, 30); 3 oligodendrocyte clusters (1, 2, 17); 1 astrocyte cluster (10); 3 microglia clusters (11, 21, 29); 2 oligodendrocyte progenitor cell (OPC) clusters (14, 32); and 1 unknown cluster (31) (Figure 11A). As predicted, APOE was highly expressed in astrocytes of PS19-fE4 and PS19-fE3 mice, and its expression was dramatically reduced in astrocytes of PS19-fE4 / GFAP-Cre mice (Figure 11B). As previously reported, some neurons also expressed APOE in PS19-fE4 and PS19-fE3 mice, and neuronal APOE expression was excluded in PS19-fE4 / Syn1-Cre mice (Figure 11B).

[0173] Using the log odds ratio estimates from the generalized linear mixed effects model (GLMM_AM), we evaluated the association with the animal model and identified cell clusters that changed in PS19-fE4 mice compared to PS19-fE3 mice, PS19-fE4 / Syn1-Cre mice, and PS19-fE4 / GFAP-Cre mice. This analysis revealed that in oligodendrocyte cluster 17, cells from PS19-fE4 / Syn1-Cre mice were less likely to be included compared to those from PS19-fE4 mice and were almost completely excluded in PS19-fE4 / Syn1-Cre mice (Figures 11C - D). Based on differentially expressed (DE) gene analysis, cells in cluster 17 showed dramatically upregulated expression of three major heat shock protein genes, as well as significantly downregulated expression of myelin basic protein (Mbp) and myelin-associated oligodendrocyte basic protein (Mobp) genes, compared to other oligodendrocyte clusters. DE pathway analysis revealed enrichment of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways related to general neurodegeneration, AD, and other neurodegenerative diseases (Figure 11E), indicating that cluster 17 represents oligodendrocytes associated with neurodegenerative diseases.

[0174] Furthermore, excitatory neuron cluster 25 had significantly lower odds of having cells from PS19-fE4 / Syn1-Cre mice and PS19-fE3 mice compared to those from PS19-fE4 mice and was almost completely excluded in PS19-fE4 / Syn1-Cre mice (Figure 11C, F). DE gene analysis revealed that cells in cluster 25 had significantly upregulated expression of three major heat shock protein genes (Figure 11A), the human MAPT gene, and the amyloid precursor protein (App) gene. DE pathway analysis revealed enrichment of KEGG pathways related to general neurodegeneration, AD, and other neurodegenerative diseases, indicating that cluster 25 represents neurons associated with neurodegenerative diseases.

[0175] Similarly, log odds ratio estimates from another generalized linear mixed effects model (GLMM_histopathology) to evaluate the association with histopathology revealed that the proportion of cells in both clusters 17 and 25 had a significant negative association with hippocampal volume and a significant positive association with the coverage area of p-tau, astrogliosis, and microgliosis. This association was mainly driven by the PS19-fE4 group and, to a lesser extent, by the PS19-fE4 / GFAP-Cre group (Figures 11G–J). The proportion of cells in cluster 17 also had a significant negative association with the MBP coverage area (Figure 11D). Together, these findings indicate that the removal of neuronal APOE4, but not astrocytic APOE4, results in a dramatic decrease in the amount of oligodendrocyte and neuronal subpopulations associated with neurodegenerative disease in the hippocampus, and that these two cell populations have a strong association with tau pathology, gliosis, neurodegeneration, and oligodendrocyte loss.

[0176] Subpopulations of astrocytes and microglia associated with neurodegenerative disease are largely eliminated by removing neuronal APOE4, but not by removing astrocytic APOE4 Further subclustering of astrocytes identified 18 subpopulations. Log odds ratio estimates from GLMM_AM revealed that astrocyte subclusters 4 and 7 had lower odds of having cells from PS19-fE4 / Syn1-Cre mice than from PS19-fE4 mice, and that subcluster 7 was completely excluded in PS19-fE4 / Syn1-Cre mice (Figs. 12A, C, F). Interestingly, astrocyte subclusters 4 and 7 highly expressed APOE (Fig. 12D). DE gene and pathway analysis revealed enrichment of KEGG pathways related to general neurodegeneration, AD, and other neurodegenerative diseases in subcluster 7, indicating that subcluster 7 represents neurodegeneration disease-related astrocytes. Subcluster 4 showed enrichment of KEGG pathways related to cAMP signaling, synaptic function, and long-term potentiation. This suggests that subcluster 4 represents synaptic dysfunction-related astrocytes.

[0177] Further subclustering of microglia also identified 18 subpopulations. Log odds ratio estimates from GLMM_AM revealed that microglia subclusters 4 and 8 had lower odds of having cells from PS19-fE4 / Syn1-Cre mice than from PS19-fE4 mice, and that subcluster 8 was completely excluded in PS19-fE4 / Syn1-Cre mice (Figs. 12I, K, N). Interestingly, microglia subclusters 4 and 8 highly expressed APOE (Fig. 12J). DE gene and pathway analysis revealed enrichment of KEGG pathways related to general neurodegeneration, AD, and other neurodegenerative diseases in subcluster 8, indicating that subcluster 8 represents neurodegeneration disease-related microglia. Subcluster 4 showed enrichment of KEGG pathways related to synaptic function, cAMP signaling, and long-term potentiation (Fig. 12R), suggesting that subcluster 4 represents synaptic dysfunction-related microglia.

[0178] Similarly, the log odds ratio estimates from GLMM_ histopathological examinations revealed that the proportion of cells in both astrocyte subcluster 7 and microglia subcluster 8 showed a significant negative association with hippocampal volume and a significant positive association with the coverage area of p-tau as well as astrogliosis and microgliosis. This association was mainly driven by the PS19-fE4 group and, to a lesser extent, by the PS19-fE4 / GFAP-Cre group (Figs. 12D–E, G–H, L–M, O–P). Interestingly, when oligodendrocyte cluster 17 and excitatory neuron cluster 25 were PCA clustered together with all astrocyte and microglia subclusters, subcluster 7 of disease-related astrocytes and subcluster 8 of microglia were shown to have similar contributions to eight pathological parameters, similar to the two disease-related clusters of oligodendrocytes (cluster 17) and neurons (cluster 25) (Fig. 12S). Collectively, these findings illustrate that the removal of neuronal APOE4 rather than astrocytic APOE4 results in a dramatic reduction in the amount of neurodegeneration disease-related astrocytes and microglia subpopulations within the hippocampus, and that these two glial subpopulations have a strong association with tau pathology, gliosis, and neurodegeneration.

[0179] Removal of astrocytic APOE4 secondarily reduces neuronal APOE4 expression During the analysis of snRNA-seq data, the removal of astrocytic APOE also dramatically reduced APOE expression in neurons in addition to its predicted elimination in astrocytes (Figure 10B). To confirm this observation, publicly available snRNA-seq datasets (GEO:GSE164507) were analyzed using a similar complex mouse model carrying a floxed knock-in human APOE gene and an astrocyte-specific Cre (Aldh11-Cre / ERT2 BAC transgene) that deletes the human APOE gene in astrocytes in a tamoxifen-inducible manner during adulthood, with or without the presence of the P301S mutant human tau. In this snRNA-seq dataset (Figure 13A), neuronal APOE expression was observed in both FE3 mice and FE4 mice (Figure 13B), which was further increased in PS19-FE4 (TAFE4_oil) mice and PS19-FE3 (TAFE3_oil) mice (Figure 13A, B), suggesting that tauopathy increases neuronal APOE expression. Removal of astrocytic APOE induced by tamoxifen treatment during adulthood led to a dramatic decrease in neuronal APOE in both excitatory and inhibitory neurons in PS19-fE4 / Aldh1l1-Cre / ERT2 (TAFE4_tam) mice and PS19-fE3 / Aldh1l1-Cre / ERT2 (TAFE3_tam) mice, as compared to oil-treated mice (Figure 13B, E, F). The effect of astrocytic APOE removal on the decrease in neuronal APOE was observed in all subtypes of neurons, suggesting a general regulation of neuronal APOE expression by astrocytic APOE. Thus, the initial observation that the removal of astrocytic APOE secondarily reduces neuronal APOE expression was confirmed in different tauopathy models using human APOE, even when the removal of astrocytic APOE was driven under different astrocyte-specific promoters and its removal was induced at different time points in the lifespan of the mouse model.

[0180] To confirm whether this observation is verifiable at the protein level, the amount of APOE protein was quantitatively evaluated by sandwich ELISA using primary cultured neurons isolated from the cerebral cortex and hippocampus of prenatal mouse fetuses. Analysis revealed that primary neurons from PS19-fE4 / GFAP-Cre mice showed a significant 50% decrease in APOE protein levels compared to those from PS19-fE4 mice (Figure 13G). These data provide strong evidence that the removal of astrocytic APOE4 results in a dramatic decrease in neuronal APOE4 expression and protein production, strengthening this regulatory relationship between astrocytic APOE4 and neuronal APOE4. Therefore, the limited protective effect of removing astrocytic APOE4 is likely due to a secondary decrease rather than a complete elimination of neuronal APOE4.

[0181] Initial studies reported that neuronal APOE expression is regulated by astrocyte-released factors via modulation of the ERK pathway within neurons. Detailed comparison of snRNA-seq data between astrocytes from PS19-fE4 / GFAP-Cre mice and astrocytes from PS19-fE4 mice revealed that several genes significantly downregulated after astrocytic APOE4 removal are known activators of the ERK / MAPK pathway, including dipeptidyl peptidase 10 (DPP10) and neuregulin-3 (Nrg3). Thus, it was hypothesized that astrocytic APOE4 may regulate neuronal APOE4 expression by promoting the release of these factors that activate the neuronal ERK pathway, leading to increased neuronal APOE4 expression. To test this possibility, primary neurons isolated from PS19-fE4 mice and PS19-fE4 / GFAP-Cre mice were treated with 10 μg / mL of recombinant DPP10 or Nrg3 protein or dPBS control, and then APOE protein levels in cell lysates were quantitatively evaluated by sandwich ELISA. It was observed that primary neurons from PS19-fE4 / GFAP-Cre mice with astrocytic APOE4 removed showed a significant 98% increase in APOE protein levels compared to the dPBS control (Figure 13H). In contrast, there was no significant difference in APOE levels in primary neurons from PS19-fE4 mice after DPP10 treatment (Figure 13I). This data indicates that primary neurons from PS19-fE4 / GFAP-Cre mice with astrocytic APOE4 removed are sensitive to DPP10 upregulation of APOE expression, while primary neurons from PS19-fE4 mice are not. No difference in APOE levels was observed in primary neurons from PS19-fE4 / GFAP-Cre or PS19-fE4 mice after treatment with Nrg3 (Figure 13H, I), indicating that Nrg3 is not involved in the regulation of neuronal APOE4 expression.

[0182] Release of neuronal HMGB1 is completely suppressed by removing neuronal APOE4, rather than astrocytes. In the search for inflammatory cytokines that can induce tau and neurodegeneration-induced gliosis, it was observed that cells in neurodegenerative disease-related neuron cluster 25 dramatically increased the expression of the high-mobility group box 1 (Hmgb1) gene (Figure 10A). The HMGB1 protein is widely recognized as a factor for glial cell activation. Under pathological conditions, HMGB1 moves from the nucleus of stressed or dying cells to the cytoplasm and is then released to act as a pro-inflammatory cytokine. As a result of immunohistochemical staining for HMGB1 protein and the nuclear marker DAPI in 10-month-old mice, it was revealed that in PS19-fE4 mice, the amount of HMGB1 protein translocating from the nucleus to the cytoplasm in the neurons of the hippocampal dentate gyrus was significantly high. This is due to the fact that most of the proteins in the hippocampal neurons were confirmed to be outside the nucleus. On the other hand, most of the HMGB1 protein in PS19-fE3 mice was retained in the nucleus, and the translocation to the cytoplasm was minimal. Interestingly, removing APOE4 from either neurons or astrocytes attenuated this phenotype observed in PS19-fE4 mice, and most of the HMGB1 protein remained localized in the nucleus.

[0183] Furthermore, PS19-fE4 mice contained a greater number of HMGB-1 positive dots in both the nucleus and cytoplasm compared to other genotype groups. Notably, fE4 and fE3 mice lacking human mutant tau-P301S had HMGB1 protein localized in the nucleus, indicating that the nuclear-cytoplasmic transport of HMGB1 in APOE4 mice requires the coexistence of both APOE4 and tauopathy (Figure 14).

[0184] To evaluate the extent of disease-related HMGB1 release from hippocampal neurons within these various genotype groups, hippocampal interstitial fluid (ISF) from 8.5-month-old mice was collected over 24 hours using in vivo microdialysis. The levels of HMGB1 protein in the ISF were quantitatively determined using sandwich ELISA. PS19-fE4 mice showed high levels of HMGB1 protein within their hippocampal ISF, but the removal of neuronal APOE4 decreased the HMGB1 protein levels in the ISF to undetectable levels, indicating that neuronal APOE4 plays a role in controlling HMGB1 release from neurons. The removal of astrocytic APOE4 also resulted in an 86% decrease (not complete elimination) of HMGB1 protein in the ISF, which is likely due to a secondary decrease (not complete elimination) in neuronal APOE4 expression (Figure 13B, F, G).

[0185] To determine whether HMGB1 is a potential regulator of neuronal APOE4 expression, primary neurons isolated from PS19-fE4 / GFAP-Cre mice and PS19-fE4 mice were treated with 10 μg / mL of recombinant HMGB1 protein or dPBS control, and the APOE protein levels in the cell lysates were quantitatively evaluated by sandwich ELISA. No significant differences in APOE protein levels were observed in PS19-fE4 / GFAP-Cre or PS19-fE4 primary neurons after treatment with HMGB1 (Figure 13H, I), indicating that HMGB1 is not involved in the regulation of neuronal APOE4 expression.

[0186] To evaluate the importance of neuronally released HMGB1 in the induction of gliosis in APOE4-related tauopathy, ISF collected from the hippocampus of 8.5-month-old PS19-fE4 mice was injected into the hippocampus of 8.5-month-old wild-type mice. The injected ISF fractions were either enriched with relatively high concentrations of HMGB1 protein (fractions 19 - 22) as determined by sandwich ELISA or had undetectable levels of HMGB1 protein as a control (fractions 4 - 7). Wild-type mice were unilaterally injected into the right dorsal hippocampus with either HMGB1 + or control ISF, and analyzed on day 6 after injection to evaluate acute changes in gliosis. Wild-type mice injected with control (HMGB1 - ) ISF showed relatively low levels of microgliosis and astrogliosis on both the injected and non-injected hippocampal sides, indicating that most of the gliosis induced by the injection surgery itself had resolved by 6 days after injection (Figures 14C - F). In contrast, wild-type mice injected with HMGB1 + ISF showed significantly increased levels of microgliosis on the injected hippocampal side compared to the non-injected side (Figures 14C - D). Interestingly, injection of HMGB1 + ISF did not result in a significant increase in astrogliosis on the injected hippocampal side compared to the non-injected side (Figures 14E - F). Collectively, these data suggest that neuronally derived release of HMGB1 represents a novel mechanism by which neuronal APOE4 promotes glial cell activation.

[0187] Treatment with an HMGB1 inhibitor substantially reduces APOE4-driven gliosis, tau pathology, and degeneration Based on the finding that neuronal APOE4 is a potent driver of neuron-derived HMGB1 release, experiments were conducted to verify the therapeutic effect of HMGB1 inhibitors in counteracting APOE4-driven pathologies in relation to tauopathy. For this purpose, two well-characterized HMGB1 inhibitors, ethyl pyruvate (EP) and glycyrrhizic acid (GA), which selectively inhibit the translocation and release of HMGB1, were tested. A mixed solution of EP (80 mg / kg) and GA (20 mg / kg) or a saline vehicle was administered to PS19-fE4 mice and PS19-fE3 mice by intraperitoneal injection at a dose of three times a week for 12 weeks. Treatment was initiated when the mice reached 6.5 months of age, when harmful pathologies had almost developed, as demonstrated in this and other studies, and completed when the mice reached 9.5 months of age, when severe neurodegeneration and pathological changes were typically present.

[0188] The effectiveness of EP and GA in blocking the nuclear-cytoplasmic transport of HMGB1 protein was first confirmed by immunostaining using anti-HMGB1 and DAPI. PS19-fE4 mice treated with saline showed extensive translocation of HMGB1 from the nucleus, while PS19-fE4 mice treated with HMGB1 inhibitors showed no distinguishable translocation of HMGB1 (Figures 15A, B). In PS19-fE3 mice, there was no significant difference in HMGB1 translocation between the saline-treated group and the inhibitor-treated group.

[0189] Next, the degree of microgliosis and astrogliosis present in the treated mice was evaluated. PS19-fE4 mice treated with saline were Iba1 + microglia and CD68 +As exemplified by the high coverage area of activated microglia, substantial microgliosis was shown throughout the hippocampus (Figures 15C, D). On the other hand, PS19-fE4 mice treated with the HMGB1 inhibitor showed a significant decrease in microgliosis. Furthermore, immunostaining with anti-GFAP revealed that PS19-fE4 mice treated with saline had extensive astrogliosis throughout the hippocampus, while PS19-fE4 mice treated with the HMGB1 inhibitor had a significant decrease in astrogliosis. PS19-fE3 mice showed no significant difference in microgliosis or astrogliosis regardless of treatment with saline or the HMGB1 inhibitor.

[0190] The degree of tau pathology in these mice was evaluated by immunostaining for p-tau. PS19-fE4 mice treated with saline showed substantial tau pathology throughout the hippocampus, while PS19-fE4 mice treated with the HMGB1 inhibitor had a dramatic decrease in tau pathology. Furthermore, immunostaining for MBP was used to evaluate the effect of administration of the HMGB1 inhibitor against myelin deficiency. PS19-fE4 mice treated with saline showed severe myelin loss in the stratum radiatum of CA1, while PS19-fE4 mice treated with the HMGB1 inhibitor had a high coverage area of MBP and showed recovery of the deficiency. There was no distinguishable difference in myelin coverage area or tau pathology between PS19-fE3 mice treated with saline and PS19-fE3 mice treated with the inhibitor.

[0191] The effect of the HMGB1 inhibitor on neurodegeneration was determined by quantifying hippocampal and posterior ventricular volumes. PS19-fE4 mice treated with saline showed substantial neurodegeneration, while PS19-fE4 mice treated with the HMGB1 inhibitor showed recovery of neurodegeneration by a significant increase in hippocampal volume and a significant decrease in posterior ventricular volume compared to PS19-fE4 mice treated with saline. There was no obvious difference in neurodegeneration between PS19-fE3 mice treated with saline and PS19-fE3 mice treated with the HMGB1 inhibitor.

[0192] In summary, these findings indicate that treatment of PS19-fE4 mice with an HMGB1 inhibitor can effectively prevent the development of prominent APOE4-driven pathologies, including gliosis, tau pathology, neurodegeneration, and myelin loss, by blocking nuclear-cytoplasmic transport and subsequent release of HMGB1 protein. This supports the idea that the translocation and release of HMGB1 play a role in the induction and exacerbation of these various APOE4-driven pathologies.

[0193] Discussion In the central nervous system, APOE is mainly produced by astrocytes, but stress or injury conditions can induce APOE expression in neurons. In this study, it was demonstrated that APOE4 leads to a significant increase in tau pathology, gliosis, neurodegeneration, neurological dysfunction, and myelin loss, as well as enrichment of neurodegenerative disease-related subpopulations of neurons, oligodendrocytes, astrocytes, and microglia. Removal of neuronal APOE4 dramatically reduces all of these observed pathologies and largely eliminates disease-related cell subpopulations, while removal of astrocytic APOE4 only reduces gliosis and neurodegeneration, indicating that neuronal APOE4 is a major driver of many of these pathologies.

[0194] Surprisingly, snRNA-seq and primary neuron culture studies have revealed that the removal of astrocytic APOE4 also results in a substantial but not complete decrease in neuronal APOE4 expression and protein production. This suggests that the limited protective effect of astrocytic APOE4 removal against gliosis and neurodegeneration is likely due to this partial decrease in neuronal APOE4. The presence of tau pathology, neurological dysfunction, and myelin loss after astrocytic APOE4 removal, as well as the subsequent decrease in neuronal APOE4, highly suggests that these lesions are sensitive to low levels of neuronal APOE4 and can be fully restored only after its complete elimination. Additionally, snRNA-seq analysis indicates that even after astrocytic APOE4 removal, the enrichment of disease-related subpopulations of neurons, oligodendrocytes, astrocytes, and microglia still exists. This finding shows that despite the overall decrease in gliosis and neurodegeneration after astrocytic APOE4 removal, these neurodegenerative disease-related cell subpopulations are sensitive to low levels of neuronal APOE4 and will still exist unless neuronal APOE4 is completely eliminated.

[0195] In addition, neurodegenerative disease-related neuronal subpopulations showed a significant increase in the expression of HMGB1, a nuclear protein known to induce neuroinflammation following translocation to the cytoplasm and subsequent extracellular release. Mechanistically, neuronal APOE4 was first demonstrated to promote HMGB1 release from neurons, as its removal results in the complete elimination of HMGB1 release. Conversely, astrocytic APOE4 removal results in a large decrease in HMGB1 release but cannot completely eliminate it, presumably due to the partial but incomplete decrease in neuronal APOE4. This neuronal APOE4-driven HMGB1 release induces gliosis and subsequent neurodegeneration and myelin loss, all of which can be effectively blocked by treatment with an HMGB1 inhibitor.

[0196] Based on all these findings, an “APOE4-HMGB1-inflammation-degeneration” cascade model for APOE4-related AD and other tauopathies is hypothesized (Figure 16G). As shown in PS19-fE4 mice (Figure 16A), in the presence of both neuronal APOE4 and astrocytic APOE4, the disease cascade starts with the neuronal expression of APOE4 that can be induced by various neuronal stressors. Neuronal APOE4 has a strong effect on the accumulation and propagation of tau pathology and can further induce the expression of neuronal APOE4. The elevation of neuronal APOE4, in concert with tau pathology, induces nucleo-cytoplasmic transport and the release of HMGB1 from neurons. Astrocytic APOE4 has no direct effect on the accumulation and proliferation of tau pathology but promotes the release of astrocytic factors such as DPP10 that can regulate neuronal APOE4 expression, and thus indirectly enhances APOE4 expression in neurons by secondarily promoting APOE4 / tau pathology-induced HMGB1 release from neurons. When released from neurons, HMGB1 acts as an inflammatory cytokine that induces gliosis. Widespread gliosis, particularly the accumulation of toxic astrocytic and microglial subtypes, causes abnormal phagocytosis of neurons, synapses, and myelin sheaths derived from oligodendrocytes, leading to subsequent neurodegeneration, neurological dysfunction, and oligodendrocyte degeneration.

[0197] As shown in PS19-fE4 / Syn1-Cre mice (Figure 16B), removal of neuronal APOE4 results in a dramatic reduction in the accumulation and propagation of tau pathology in neurons. The reduction of tau pathology and the absence of neuronal APOE4 together appear as a blockade of HMGB1 release from neurons. Since astrocytic APOE4 contributes to HMGB1 release by indirectly enhancing neuronal APOE4 expression, the absence of neuronal APOE4 blocks the indirect effect of astrocytic APOE4 on the induction of neuronal HMGB1 release. Without sufficient neuronal HMGB1 release, gliosis, particularly gliosis of toxic microglia and astrocyte subtypes, is not induced, and as a result, neurodegeneration, nerve dysfunction, and oligodendrocyte degeneration are reduced. Therefore, removal of neuronal APOE4 brings about complete prevention or relief of all the observed pathologies.

[0198] Removal of astrocytic APOE4 results in a significant but not complete decrease in neuronal APOE4 expression, as shown in PS19-fE4 / GFAP-Cre mice (Figure 16C). The reduced level of neuronal APOE4 is still sufficient to induce significant accumulation and propagation of tau pathology, which, together with the low level of neuronal APOE4, causes low-level HMGB1 release. Without a sufficiently high level of HMGB1 release, the overall level of gliosis is greatly reduced. Nevertheless, the low levels of neuronal APOE4 and HMGB1 release are sufficient to promote the accumulation of toxic astrocyte and microglia subtypes. The overall reduction of gliosis leads to a reduction in neurodegeneration, but due to the low level of neuronal APOE4, a subpopulation of disease-related neurons still exists. The presence of toxic glia subtypes probably still leads to oligodendrocyte deficits due to the higher sensitivity of oligodendrocytes to lower levels of toxic glia subtypes. The presence of high levels of tau pathology and toxic glia subtypes can continue to contribute to nerve dysfunction. Therefore, removal of astrocytic APOE4 brings about partial prevention or relief of the observed pathologies.

[0199] Importantly, pharmacological studies indicate that HMGB1 inhibitors represent a novel and effective approach for treating APOE4-related AD and other tauopathies. In the presence of HMGB1 inhibitors (Figure 16D), the release of HMGB1 from neurons is blocked. If HMGB1, which acts as an inflammatory cytokine, is not released, gliosis is significantly reduced, resulting in a significant reduction in neurodegeneration and oligodendrocyte degeneration. The absence of gliosis and a potential reduction in toxic glial subtypes also result in a significant reduction in tau pathology, which, together with the reduction in neurodegeneration, likely improves neurological dysfunction. Thus, HMGB1 inhibitor treatment also results in complete prevention or rescue recovery of all observed pathologies.

[0200] Inhibition of HMGB1 has been investigated as a viable therapeutic option for various neurological disorders, but this study provides the first evidence that treatment with HMGB1 inhibitors is a valuable and effective therapy to counteract the APOE4-driven effects on many prominent AD pathologies. Furthermore, the inventors' findings also suggest that the development of therapies targeting the removal of neuronal APOE4 is likely to provide a higher therapeutic value than the removal of astrocytic APOE4. Removal of either neuronal or astrocytic APOE4 dramatically reduces gliosis and neurodegeneration by the 10-month time point analyzed by the inventors, but as demonstrated by the inventors and other studies, the long-term consequences of the presence of high levels of tau pathology and neurodegenerative disease-related glial subpopulations that occur in the absence of astrocytic APOE4 are unknown and may lead to neurodegeneration at later time points or disrupt other important processes. Furthermore, removal of neuronal APOE4 provides the additional benefit of reducing tau pathology, neurological dysfunction, myelin degeneration, and OPC depletion. Overall, the inventors' study identifies neuronal APOE4 as a central player in the induction of the pathological cascade in APOE4-related AD and other tauopathies and provides potential targets for anti-AD and tauopathy drug development.

[0201] Example 15: Method Mouse: Human LoxP-floxed APOE knock-in (fE) mice with conditional deletion of the human APOE gene were generated. Briefly, homozygous fE3 and fE4 mice44 were crossed with synapsin 1-Cre transgenic mice [B6.Cg-Tg(Syn1-Cre)671Jxm / J] (The Jackson Laboratory)45 or GFAP-Cre transgenic mice [B6.Cg-Tg(GFAP-Cre)8Gtm] (National Cancer Institute Mouse Repository). fE / Cre mice were further crossed with Tau-P301S (PS19) transgenic mice [B6;C3-Tg(Prnp-MAPT*P301S)PS19Vle / J] (The Jackson Laboratory) that express human P301S 1N4R Tau driven by the PrP promoter to generate PS19-fE4 and PS19-fE3 mice without Cre, with Syn1-Cre, or with GFAP-Cre. Littermates negative for Syn1-Cre or GFAP-Cre were used as PS19-fE controls. For the generation of the PS19-fE / Syn1-Cre line, only female Syn1-Cre mice were used for breeding purposes because germline recombination has been reported to occur in the offspring of male Syn1-Cre mice. Wild-type (WT) mice [C57BL / 6J] were obtained from the Jackson Laboratory. All mice were on a pure C57BL / 6 genetic background and were housed in a pathogen-free barrier facility at 19–23 °C and 30–70% humidity on a 12-hour light cycle. Animals were identified by ear punch under brief isoflurane anesthesia and genotyped by polymerase chain reaction (PCR) of a tail biopsy sample. All animals received no procedures other than those reported in this study. All animal experiments were performed in accordance with the guidelines and regulations of the National Institutes of Health, the University of California, and Gladstone Institutes under protocol AN176773.

[0202] For brain tissue collection, mice were deeply anesthetized by intraperitoneal injection of avertin (Henry Schein) and perfused transcardially with 0.9% saline for 1 minute. Brains were fixed as either whole brains or hemispheres depending on the study. The right hemisphere was drop-fixed in 4% paraformaldehyde (16% PFA diluted in MilliQ H2O) (Electron Microscopy Sciences) for 48 hours, washed out in 1× PBS (Corning) for 24 hours, and cryoprotected in 30% sucrose (Sigma) at 4 °C for 48 hours. Fixed hemispheres were sectioned into 30-μm thick coronal sections on a cryostat slide microtome (Leica) and stored at -20 °C in cryoprotectant solution (30% ethylene glycol, 30% glycerol, 40% 1× PBS). The left hemisphere was snap-frozen on dry ice and stored at -80 °C.

[0203] Immunohistochemistry: For immunofluorescence staining, several sections (separated by approximately 300 μm) from each mouse were transferred to a 12-well plate in 1× PBS-T (PBS + 0.1% Tween®-20) (Millipore Sigma) and washed three times in PBS-T for 5 minutes each to remove the cryoprotectant solution. The sections were incubated in boiling antigen retrieval buffer (Tris buffer, pH 7.6) (TEKNOVA) for 5 minutes and washed twice in PBS-T for 5 minutes each. The sections were then incubated in blocking solution (5% normal donkey serum in 1× PBS (Jackson Labs), 0.2% Triton-X (Millipore Sigma)) for 1 hour at room temperature to prevent non-specific antibody binding. After blocking, the sections were washed once in PBS-T for 5 minutes and incubated in mouse-on-mouse (M.O.M.). Blocking buffer (1 drop of M.O.M IgG / 4 mL of PBS-T) (Vector Labs), at room temperature for 1 hour. After M.O.M. block, the sections were diluted to the optimal concentration (anti-APOE 1:200 (Cell Signaling); anti-CD68 1:100 (Bio-Rad); anti-Cre 1:800 (Cell Signaling); anti-GFAP 1:800 (Millipore Sigma); anti-GFP 1:5000 (Thermofischer); anti-HMGB1 1:100 (Abcam); anti-Iba1 (rbt) 1:200 (Wako); anti-Iba1 (gt) 1:200 (Abcam); anti-MBP 1:500 (Abcam); anti-NeuN 1:500 (Millipore Sigma); anti-NG2 1:500 (Abcam); anti-S100β 1:200 (Abcam)) and then incubated overnight at 4°C in the primary antibody. After primary antibody incubation, the sections were washed three times in PBS-T for 5 minutes each, then diluted in PBS-T and incubated for 1 hour at room temperature in the dark in a fluorescently labeled secondary antibody (Abcam, Jackson Immuno, 1:1000 in PBS-T). The sections were then washed twice in PBS-T for 5 minutes each and incubated in DAPI (1:50,000 in PBS-T) (Thermofisher) for 8 minutes at room temperature in the dark.Next, the sections were washed twice in PBS-T for 5 minutes each, then mounted on microscope slides (Fisher Scientific), covered with a coverslip using ProLong Gold mounting media (Vector Laboratories), and sealed with clear nail polish. Images were taken at magnifications of 10x, 20x, 40x, or 60x depending on the staining using an FV3000 confocal laser scanning microscope (Olympus) or an Aperio VERSA slide scanning microscope (Leica). After setting the standard threshold applied to all images, open-source Fiji (ImageJ) software was used to perform image analysis of the covered area percentage. The researchers were also blinded to the samples to eliminate the possibility of bias.

[0204] For DAB (3,3’-diaminobenzidine) staining, several sections (separated by approximately 300 μm) from each mouse were transferred to a 12-well plate in 1×PBS-T and then washed three times in PBS-T for 5 minutes each to remove the cryoprotectant solution. The sections were then incubated in boiling antigen retrieval buffer (1×PBS, 0.1 M sodium citrate, 0.1 M citric acid) (Fisher Scientific, Fluka) for 5 minutes and washed twice in PBS-T for 5 minutes each. Next, the sections were incubated in endogenous peroxidase buffer (1×PBS, 10% methanol (Fisher Scientific), 3% H2O2 (Sigma)) for 15 minutes and washed three times in PBS-T for 5 minutes each. The sections were then incubated in blocking solution (1×PBS-T, 5% normal donkey serum, 1% non-fat dry milk) for 1 hour at room temperature. After blocking, the sections were washed twice in PBS-T for 5 minutes each, then incubated in avidin / biotin blocking agent (4 drops per block) (Vector Laboratories) for 15 minutes and then washed twice in PBS-T for 5 minutes each. The sections were incubated in M.O.M. blocking buffer (1 drop of M.O.M IgG / 4 mL of PBS-T) (Vector Labs) for 1 hour at room temperature. After M.O.M. block, the sections were washed twice for 5 minutes each, diluted to the optimal concentration in PBS-T, and then incubated in primary antibody overnight at 4°C (anti-p tau (AT8) 1:100 (Invitrogen); anti-HT7 1:200 (Peter Davies)). After primary antibody incubation, the sections were washed three times in PBS-T for 5 minutes each and then incubated in biotinylated secondary antibody (1:200; Jackson Immuno) for 1 hour at room temperature. Next, the sections were washed three times in PBS-T for 5 minutes each and incubated in ABC buffer (Vector Laboratories) prepared 10 minutes before the incubation step. The sections were washed twice in PBS-T for 5 minutes each and once in Tris buffer (pH 7.6) for 5 minutes.The sections were incubated precisely for 2 minutes in DAB buffer (5 mL of 1× PBS, 2 drops of buffer stock solution, 2 drops of DAB, 2 drops of H2O2) (Vector Laboratories). The staining was stopped by washing the sections 3 times for 5 minutes each in Tris buffer (pH 7.6) and 2 times for 5 minutes each in PBS-T. The sections were mounted on microscope slides and dried overnight at room temperature. Next, the mounted sections were immersed 2 times for 5 minutes each in xylene (Fisher Scientific) and then mounted with a coverslip using DPX mounting medium (Sigma-Aldrich). Images were taken using an Aperio VERSA slide scanning microscope (Leica) at a magnification of 10×.

[0205] Volumetric analysis: Serial coronal hippocampal brain sections (7 sections per mouse, 30 μm thick, 300 μm apart) were mounted on microscope slides (Fisher Scientific) and dried for 1 hour at room temperature. The 0.1% Sudan Black solution was prepared by adding an appropriate amount of Sudan Black powder (Sigma) to 70% ethanol (KOPTEC) and mixing the solution using a magnetic stirrer while protecting from light. The solution was then centrifuged at 3,000 RPM for 10 minutes, and the collected supernatant was filtered using a 0.2 μm filter syringe (Thermo Scientific) to remove undissolved dye. The sections were then stained with the 0.1% Sudan Black solution for 10 minutes at room temperature and washed 3 times for 2 minutes each in 70% ethanol and 3 times for 5 minutes each in Milli-Q water. The sections were then coverslipped with ProLong Gold mounting medium (Invitrogen) and imaged at a magnification of 10× using an Aperio VERSA slide scanning microscope (Leica). For volumetric analysis of the hippocampus and posterior ventricle, the regions of interest were traced in ImageJ using the segmented line tool, and the formula: volume = (sum of the regions) *The volume was calculated using 0.3 mm23. The sum of the area values was obtained by taking the sum of the quantified area measurements of all seven brain sections per mouse, approximately between coordinates AP = -1.2 and AP = -3.4.

[0206] Measurement of the thickness of the neuronal layer: Two brain sections (30 μm thick, 300 μm apart) were immunofluorescently stained as described above using the primary antibody NeuN (1:500) to visualize the neuronal cell layer of the hippocampus. The sections were imaged at a magnification of 20x using an FV3000 confocal laser scanning microscope (Olympus). The thickness of the CA1 pyramidal cell layer and the dentate gyrus granule cell layer of the hippocampus was measured on Fiji (ImageJ) software by drawing a straight line perpendicular to the NeuN+ cell layer at two points per subfield of the hippocampus and taking the average value for each mouse.

[0207] Nuclear-cytoplasmic localization of HMGB1 measurement: Two brain sections (30 μm thick, 300 μm apart) were immunostained with anti-HMGB1 (1:100) and DAPI (1:50,000) as described above. The sections were imaged at magnifications of 40x and 60x using an FV3000 confocal laser scanning microscope (Olympus). All image processing and quantification were performed with Fiji (ImageJ) software. Briefly, a median filter of 1 pixel was applied to the DAPI channel, and an appropriate threshold was set to create a mask of DAPI. Then, the image calculator function was used to overlay the DAPI mask and the HMGB1 channel, resulting in HMGB1 staining localized only to the nucleus. After obtaining the integrated density and particle values, the image calculator was used to subtract the DAPI mask from HMGB1, providing HMGB1 staining excluded from the nucleus.

[0208] Biochemical extraction of brain tissue: After thawing the hippocampus on ice, it was dissected from the rapidly frozen mouse half-brain. The hippocampal tissue was weighed and homogenized at 10 μL / mg tissue using a Polytron® immersion disperser Polytron® homogenizer (Kinematica AG) in ice-cold RAB buffer (G Biosciences) supplemented with phosphatase inhibitor (Roche) and protease inhibitor (Roche). The samples were then centrifuged at 50,000 g for 20 minutes at 4 °C using an Optima TLX ultracentrifuge (Beckman Coulter), and the supernatant was collected as the RAB soluble fraction. The pellet was resuspended in ice-cold RIPA buffer (Thermo Scientific) at 10 μL / mg tissue and centrifuged at 50,000 g, 4 °C for 20 minutes. The supernatant was recovered as the RIPA soluble fraction, and the pellet was stored at -80 °C for further use. All fractions were stored at -80 °C until further analysis.

[0209] Western blot analysis: Biochemically extracted mouse hippocampal tissue lysates were loaded onto a 12% Bis-Tris SDS-PAGE gel (Invitrogen) and separated by gel electrophoresis at 160 V using MOPS buffer. The separated proteins were transferred onto a nitrocellulose membrane at 18 V for 60 minutes (Trans-Blot Turbo Transfer System (Bio-rad)). The membrane was washed three times in PBS-T for 5 minutes each, then incubated in Intercept blocking buffer (LI-COR) for 1 hour at room temperature to block non-specific binding sites. After blocking, the membrane was washed three times in PBS-T for 5 minutes each and incubated overnight at 4 °C with primary antibodies (AT8 1:3,000 (Invitrogen), TUJ1 1:15,000 (Biolegend)). The membrane was washed three times in PBS-T for 5 minutes each and incubated for 1 hour at room temperature in the dark with a fluorescently labeled secondary antibody (1:20,000; LI-COR). The resulting bands were detected using an Odyssey CLx infrared imaging system (LI-COR), and the fluorescence intensity of the bands was quantified as the ratio of the AT8:TUJ1 signals using Image Studio software.

[0210] Sandwich ELISA: Biochemically extracted mouse hippocampal tissue lysates were diluted to an appropriate concentration in Milli-Q H2O and performed according to the provided manufacturer's protocol (human APOE (Abcam); mouse HMGB1 (Novus Biologicals)). The reactions of the samples were read using a SpectraMaX® M5 spectrophotometer (Molecular Devices), the standard curve was interpolated, and after adjusting the dilution, the protein concentration was determined.

[0211] Primary neuron cultures and recombinant protein treatment: Primary cultures of neurons were prepared from prenatal E20 pups of various genotypes. After harvesting the pups' brains, the cortex + hippocampus was isolated and placed in ice-cold dissociation medium + kynurenic acid medium (DM / KY) (in distilled H2O, DM: Na2SO4 (81.8 mM); K2SO4 (30 mM); MgCl2 (5.8 mM), CaCl2 (0.25 mM); HEPES (1 mM); glucose (20 mM); phenol red (0.001%); NaOH (0.16 mM)) (KY: kynurenic acid (10 mM); phenol red (0.0025%); HEPES (5 mM); MgCl2 (100 mM); NaOH (adjusted dropwise to pH 7.4)). The resulting DM / KY medium was prepared by combining 90% DM with 10% KY medium. The isolated tissue was minced and then immersed in pre-warmed papain solution (1 mL / brain) for 13 minutes while gently inverting, and then immersed in trypsin inhibitor solution (5 mL for up to 10 brains) for 5 minutes while gently inverting. The tissue pellet was washed with Optimem / glucose solution (20 mM glucose, 1 mL / brain) while gently inverting. Then, fresh Optimem / glucose solution was added and the tissue was gently triturated until separated into single cells. After filtering the cells through a 40 μm cell strainer, the dissociated cells were plated in Neurobasal medium supplemented with B27, 100 U / mL -1 of penicillin G, 100 μg / mL -1 of streptomycin, and 1% GlutaMAX™ in 12-well plates at 1×10 6 cells / well or in 24-well plates at 3×10 5Cells were plated in wells. Every 3 - 4 days, half of the medium was removed and replaced with fresh B27 / Neurobasal medium. In some experiments, primary neurons were treated in vitro for 24 hours on day 14 with either Dulbecco's PBS (dPBS) vehicle or the recombinant protein of interest (10 μg / mL for one well of a 12 - well plate). After treatment, the medium was collected and the cultures were harvested for analysis. Total protein levels present in cell lysates were obtained by BCA assay (Pierce).

[0212] Mouse stereotaxic surgery: Mice were anesthetized by intraperitoneal injection of ketamine (60 mg / kg) and xylazine (30 mg / kg) and maintained with 0.8% - 1.0% isoflurane (Henry Schein). Mice were fixed in a stereotactic alignment system model 940 (Kopf Instruments) using ear bars and a tooth bar. The scalp was prepared by removing hair using Nair® and sterilizing with 70% ethanol. The scalp was then incised using a scalpel and sterilized with 70% ethanol. The cranial suture was better visualized using 3% hydrogen peroxide. After confirming bregma, the unilateral targeting site was drilled with a 0.5 - mm microdrill (Fine Science Tools) using coordinates X = +1.5, Y = -2.1, Z = -2.1 (Z is measured from the surface of the brain). Mice were injected with 2 μL of each virus (AAV2(Y444F)-SmCBA - human_P301S_tau - WPRE, 2.10E+13 vg / mL, Virovek; AAV2 - synapsin - GFP, 1.0E+13 vg / mL, SignaGen) or the ISF fraction at a rate of 500 nL / min and allowed to diffuse for 3 minutes. After surgery, mice were sutured with nylon monofilament non - absorbable 6 - 0 sutures (Henry Schein), and analgesics buprenorphine (0.0375 mg / kg intraperitoneal), ketofen (5 mg / kg subcutaneous), and saline (500 μL intraperitoneal) were administered. Mice were monitored on a heating pad until they were ambulatory and provided with hydrogel for hydration.

[0213] Electrophysiological recording and data analysis of brain slices: For electrophysiological recording studies, 8-month-old PS19-fE3 mice and PS19-fE4 mice without Cre, with Syn1-Cre, or with GFAP-Cre were anesthetized with isoflurane and decapitated. The brains were quickly removed from the skulls and placed in ice-cold (2-5 °C) slicing solution. The slicing solution contained the following (in mM each): 110 choline chloride, 2.5 KCl, 26 NaHCO3, 10 MgCl2, 1.25 NaH2PO4, 0.5 CaCl2, 10 glucose, 3 sodium pyruvate, 1 L-ascorbic acid, pH 7.4. Sagittal slices with a thickness of 350 μm were cut from both hemispheres using a vibratome (VT1200, Leica) and transferred to a 95% O2-CO2 vapor interface holding chamber (BSK5, Scientific Systems Design) containing artificial cerebrospinal fluid (ACSF), where they were allowed to recover for 1 hour at 34 °C and then maintained at room temperature (20-22 °C). The ACSF contained (in mM each): 126 NaCl, 2.5 KCl, 1.5 CaCl2, 1.5 MgCl2, 26 NaHCO3, 1.25 NaH2PO4, 10 glucose, and 1.5 L-ascorbic acid, pH 7.4.

[0214] For input / output recording studies, orthodromic stimulation of the Schaffer collateral branches was used to evoke local field postsynaptic potentials (fPSPs). A concentric bipolar stimulating electrode (FHC) connected to a constant-voltage isolation stimulator (DS2A-MKII, Digitimer North America) was used for stimulation and placed in the CA2 radial layer. Cultured fPSPs were recorded using a glass borosilicate microelectrode filled with ACSF and placed in the CA1 radial layer. Signals were sampled and digitized by a MultiClamp 700B amplifier and Digidata 1550B1 acquisition system equipped with pClamp10 software (Molecular Devices), and analyzed using IgorPro6 software (Wavemetrics) that executed a custom macro. The fPSP slope was analyzed as the linear fit slope value between 10% and 90% of the fPSP peak. The input-output relationship was recorded as the fPSP slope value corresponding to increasing stimulus intensities (20 - 60 μA), and the fPSP slope increase was calculated as the linear slope of the resulting input-output curve.

[0215] Microdialysis of mouse hippocampus: Interstitial fluid was collected using in vivo microdialysis of the hippocampus. Surgical procedures, including pre- and post-operative care, were performed as described above for stereotactic surgery. During surgery, a unilateral stereotactic site was drilled with a 1.2 mm bone drill bit (BASi), and an AtmosLM guide cannula PEG-4 (Amuza) was stereotactically implanted at coordinates X = +1.5, Y = -2.1, Z = -1.1 above the right hippocampus. The cannula was fixed in place using dental cement (GC America), a temporary PEG-4 AtmosLM dummy probe (Amuza) was inserted, and fixed with an AC-5 cap nut screw (Eicom). Two days after surgery, the mice were placed in a microdialysis-independent system (BASi) overnight for acclimation, and a 1000 kDa AtmosLM collection probe (Eicom) was inserted into the hippocampus through the guide cannula the following afternoon. This extended 1 mm downward to Z = -2.1 to target the dentate gyrus. Artificial CSF (Harvard Apparatus) made with 0.15% BSA (Thermo Scientific) was circulated through the system at a rate of 0.5 μL / min using the push-pull method, and ISF was collected approximately every 1 hour for 24 hours in a refrigerated fraction collector (BASI). To prevent tube clogging, the pump was operated at 10 times the collection rate for the first 2 hours and then adjusted to a flow rate of 0.5 μL / min. After completion of ISF collection, the mice were euthanized and perfused with 0.9% saline as described above. The brain was sectioned into hemispheres, the right hemisphere was post-fixed in 4% PFA for 48 hours, and the left hemisphere was freshly frozen. The ISF fraction was frozen at -80 °C for further analysis.

[0216] Treatment with HMGB1 inhibitor: At 6.5 months of age, male and female PS19-fE4 mice and PS19-fE3 mice were randomly assigned to a control group or a treatment group. The mice were injected intraperitoneally with either sterile grade 0.9% saline (Fisher Scientific) or a mixture of HMGB1 inhibitors dissolved in 0.9% saline: ethyl pyruvate (80 mg / kg) (Sigma-Aldrich) and glycyrrhizic acid (20 mg / kg) (Sigma-Aldrich). The mice received three injections per week for 12 weeks starting at 6.5 months of age until they reached 9.5 months of age. During the experiment, the body weight changes, grooming changes, and posture of all mice were monitored, but no changes were observed. After treatment, the animals were perfused and their brain tissues were processed for histopathological analysis as described above.

[0217] Single nucleus preparation for 10× loading: Mouse hippocampi were dissected on ice and placed into a pre-chilled 2 mL Dounce containing 1 mL of cold 1× homogenization buffer (1× HB) (250 mM sucrose, 25 mM KCl, 5 mM MgCl2, 20 mM Tricine-KOH pH 7.8, 1 mM DTT, 0.5 mM Sermidine, 0.15 mM Sermine, 0.3% NP40, 0.2 units / μL RNase inhibitor, 0.2 units / μL protease inhibitor). Dounce with the loose pestle of “A” (about 10 strokes), then with the tight pestle of “B” (about 15 strokes). Filter the homogenate using a 70 μM Flowmi™ strainer (Eppendorf) and transfer it to a pre-chilled 2 mL LoBind® tube (Fischer Scientific). Pellet the nuclei by spinning at 4°C, 350 RCF for 5 minutes. Remove the supernatant and resuspend the nuclei in 400 μL of 1X HB. Next, add 400 μL of 50% iodixanol solution to the nuclei, then slowly layer 600 μL of 30% iodixanol solution under the 25% mixture, then layer 600 μL of 40% iodixanol solution under the 30% mixture. Then spin the nuclei in a pre-chilled swinging bucket centrifuge at 4°C, 3,000 g for 20 minutes. Collect 200 μL of the nuclear band at the 30%-40% interface and transfer it to a new tube. Next, add 800 μL of 2.5% BSA + 0.2 units / L RNase inhibitor in PBS to the nuclei and then spin at 500 RCF, 4°C for 10 minutes. Resuspend the nuclei in 2% BSA in PBS + 0.2 units / L RNase inhibitor to reach about 500 nuclei / μL. Then filter the nuclei through a 40 μM Flowmi staining device™. Count the nuclei and then load about 13,000 nuclei per sample onto a 10× Genomics Next GEM Chip G. snRNA-seq libraries were prepared using the Chromium Next GEM Single Cell 3’ Library and Gel Bead Kit v3.1 (10× Genomics) according to the manufacturer's instructions.The library was sequenced on an Illumina NovaSeq™ 6000 sequencer at UCSF CAT Core.

[0218] Custom reference genome: The PS19 tau mutant floxed APOE knock-in mouse model was used for single-nucleus RNA sequencing (snRNA-seq). Homo sapiens microtubule-associated protein tau (MAPT) (NCBI reference sequence: NM_001123066.4) and Homo sapiens APOE were the genes of interest in this study. Since these genes were expected not to be included in the mouse reference genome, the reference mouse genome sequence (GRCm38) from Ensembl (release 98) and GENCODE (release M23), similar to those used in the 10x Genomics Cell Ranger mouse reference package mm10 2020-A, were used to quantify reads aligning to these genes of interest. 79A custom mouse reference genome was generated using the mouse gene annotation file from . The chromosome names included in the header of the Ensembl reference mouse genome sequence fasta file were modified to match those in the GENCODE fasta file. The annotation GTF file contains entries from non-polyA transcripts that overlap protein-coding genes. These reads were flagged as multi-mapped and not counted by the 10x Genomics Cell Ranger v6.1.1 counting pipeline. To avoid this, the GTF file was modified to (1) remove the version suffix from transcript, gene, and exon ids to match the Cell Ranger reference package, and (2) remove non-polyA transcripts. The Homo sapiens MAPT sequence and the Homo sapiens APOE sequence were added as separate chromosomes to the end of the mouse reference genome sequence, and the corresponding gene annotations were added to the filtered mouse reference gene annotation GTF file. A custom reference genome was constructed using the modified fasta and GTF files with the 10x Genomics Cell Ranger v6.1.1 mkref pipeline.

[0219] Pretreatment and Clustering of Mouse snRNA-seq Samples: The snRNA-seq samples included a total of 16 samples, each containing 4 mice from each of 4 genotype groups (PS19-fE4, PS19-fE4 Syn1-Cre, PS19-fE4 GFAP-Cre, and PS19-fE3). Each group of 4 mice consisted of 2 male and 2 female mice. The multiplexed fastq files for these samples were aligned using the 10x Genomics Cell Ranger v6.1.1 counting pipeline (Zheng et al., Nat. Commun., 8, 14049 (2017)) with a custom mouse reference genome (see the custom reference genome method for additional explanation). Details of this method are described in the Cell Ranger documentation. The intron inclusion flag for the counting pipeline was set to true to count reads mapping to intron regions. The Cell Ranger count web summary indicated an error of "low fraction of reads in cells" for 2 samples (1 from the PS19-fE4 GFAP-Cre group and 1 from the PS19-fE3 group). For these 2 samples, only about 40% of the reads were assigned to cell-associated barcodes and <80% of the reads were mapped to the genome. These metrics were much higher than those of the other 14 samples. Checking the experimental records showed that these 2 samples had problems in the nuclear isolation step and that lower cDNA was recovered due to the use of an old batch of expired sample preparation reagents. The other 14 samples were all prepared using a new batch of sample preparation reagents. Therefore, these 2 samples were excluded and only the remaining 14 samples were used for downstream analysis by Seurat.

[0220] The filtered count matrices generated by the Cell Ranger count pipeline for 14 samples were processed using the R package Seurat v4.0.581 for single-nucleus analysis. Each sample was preprocessed as a Seurat object, and the top 1% of cells per sample with a large number of unique genes, cells with 200 or fewer unique genes, and cells with 0.25% or more mitochondrial genes were excluded for each sample. The 14 samples were integrated into one Seurat object, and normalization and variance stabilization were performed using sctransform with the "glmGamPoi" (Bioconductor package version 1.6.0) method for initial parameter estimation.

[0221] Graph-based clustering was performed using the Seurat v4.0.5 functions FindNeighbors and FindClusters. First, cells were embedded into a k-nearest neighbor (KNN) graph based on Euclidean distances in the PCA space. The edge weights between two cells were further corrected using the Jaccard similarity. Next, clustering was performed using the Louvain algorithm implemented in the Seurat function FindClusters. Clustering was performed for all combinations of 10, 15, 20 principal components (PCs) and resolutions 0.4, 0.5, 0.6, 0.7, 0.8, 0.9. Clustering at 15 PCs and a resolution of 0.7 resulted in 33 distinct biologically relevant clusters, which were used for further analysis.

[0222] Cell type assignment: Data visualization using Seurat v4.0.5 in the UMAP space for 14 samples did not reveal batch effects by age, gender, genotype, date of birth, or date of nuclear isolation. Marker genes for each cluster were identified using the FindAllMarkers·Seurat function for SCT assay data. This algorithm uses the Wilcoxon rank sum test to iteratively identify genes that are differentially expressed in the cluster relative to all other clusters. Marker genes were filtered to retain only positively expressed genes that were detected in at least 25% of the cells in any population and had a fold change of at least 0.5 log2. The identity of cell clusters was assigned by matching cell clusters to known cell types using the expression of canonical cell type-specific genes, the expression of genes identified in publicly available mouse hippocampal single-cell RNA-seq datasets, and the expression of marker genes for each cluster in publicly available resources of in situ hybridization images of the whole brain.

[0223] Subclustering of astrocyte and microglia sn-RNA-seq data: Hippocampal cell cluster 10 was annotated as astrocyte cells, and hippocampal cell clusters 11, 21, and 29 were annotated as microglia cells. Both of these cell types were subclustered. Normalization and variance stabilization were performed using the "glmGamPoi" (Bioconductor package version 1.6.0) method for initial parameter estimation in sctransform 82This was performed using Seurat v4.0.5. Graph-based clustering was carried out using the functions FindNeighbors and FindClusters in Seurat v4.0.5. First, the cells were embedded into a k-nearest neighbor (KNN) graph based on the Euclidean distance in the PCA space. The weights of the edges between two cells were further corrected using the Jaccard similarity. Next, clustering was performed using the Louvain algorithm implemented in the Seurat function FindClusters. Clustering was performed for all combinations of 10, 15, 20, 25, and 30 principal components (PCs) and resolutions of 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9. Subclustering at 15 PCs and a resolution of 0.9 resulted in 18 distinct biologically relevant subclusters for astrocytes. Subclustering at 15 PCs and a resolution of 0.9 resulted in 18 distinct biologically relevant microglia subclusters.

[0224] Gene set enrichment analysis: Genes differentially expressed between the target clusters were identified using the FindAllMarkers Seurat function for the SCT assay data. This algorithm uses the Wilcoxon rank-sum test to identify genes differentially expressed between two populations. Differentially expressed genes were limited to those detected in at least 10% of the cells in either population and having at least a 0.1 log2 fold change. Overrepresentation (or enrichment) analysis was performed using clusterProfiler v4.2.1 to find gene sets in the KEGG database for mice related to the differentially expressed genes. The p-values were based on the hypergeometric test and adjusted for multiple testing using the Benjamini-Hochberg method. Significantly enriched gene sets were filtered to have an adjusted p-value less than 0.8 and at least 10 differentially expressed genes present in the gene set. The same method was used for gene set enrichment analysis of astrocyte subclusters and microglia subclusters.

[0225] Association between clusters and genotypes: A generalized linear mixed effects model (GLMM_AM) for assessing the association with animal models was implemented in the lme4 (v1.1 - 27.1) R package and used to estimate the association between cluster membership and the mouse model. These models were run separately for each cluster of cells. The GLM models were performed using the family argument set to a binomial probability distribution and the bobyqa control optimizer used for maximum likelihood estimation. The cluster membership for each cell was modeled as a response variable from 0 to 1 according to whether the cell belonged to the cluster under consideration. The corresponding mouse id from which the cell was derived was set as a random effect variable, and the animal model for this mouse id was included as a fixed variable. The reference animal model was set to PS19 fE4. The p-values obtained for the three animal models (with respect to PS19 fE4) and the log odds ratios estimated across clusters were adjusted for multiple testing using the Benjamini - Hochberg method. The same method was used to estimate the between-cluster association with genotypes for astrocyte subclusters and microglia subclusters.

[0226] Association between cell type proportions and histopathological parameters: A generalized linear mixed effects model (GLMM_histopathology) for assessing the association with histopathology was implemented in the lme4 (v1.1 - 27.1) R package and used to identify cell types whose proportions were significantly associated with histopathological changes across samples. These models were based on cell clusters and eight histological parameters: hippocampal volume (mm 3)、The percentages of AT8 coverage area, GFAP coverage area, S100β coverage area, IBA1 coverage area, CD68 coverage area, MBP coverage area, and OPC coverage area were each performed separately for each combination. The GLM model was run using a family argument set to the binomial probability distribution family and the bobyqa control optimizer for maximum likelihood estimation. The cluster membership for each cell was modeled as a response variable from 0 to 1 according to whether the cell belonged to the cluster under consideration. The corresponding mouse model from which the cell was derived was included as a random effect, and further, the mouse id within a given mouse model was also modeled as a random effect in the same way. Note that this represents the hierarchical nature of the data in the GLMM, first assuming that the mouse models were sampled from the "universe of all mouse models", and then followed by the procedure of sampling individual mice within each mouse model. The choice of including the mouse models as random effects rather than fixed effects in the model means increasing the degrees of freedom (or maximizing the statistical power) to detect associations, especially considering that the number of replicates per animal model was relatively small (3 - 4). The histological parameters under consideration were modeled as fixed effects in this model.

[0227] A subset of cell types of interest was selected, and the log odds ratio estimates (obtained from GLMM fitting) were visualized in a heatmap using the pheatmap package 1.0.12 after adjusting the p-value distribution among the pathological parameters for all cell types using multiple test correction by the Benjamini-Hochberg method. The pipeline was applied to astrocyte and microglia subtypes, and the associations between the astrocyte and microglia subtypes of interest and eight histopathological parameters were visualized in Figures 4h and 4i, respectively. The first five principal component coordinates were estimated using eight log odds ratios for unit changes in the histopathological parameters for each of the cell types of interest and the sub-cell types of astrocytes and microglia. This was implemented using prcomp(scale=T, center=T) of the stats R package. The first two PCs were visualized using fviz_pca_ind() implemented in the factoextra 1.0.7 R package.

[0228] Mouse snRNA-seq dataset from GEO:GSE164507: The mouse snRNAseq data available in the Gene Expression Omnibus database: www.ncbi.nlm.nih.gov / geo (accession number GSE164507) was reanalyzed. For each sample, a complete publicly available dataset was downloaded, including the filtered matrix of genes by cell expression, the file with barcodes, and the file with expressed genes. Briefly, this study examined P301S mutant tau transgenic mice carrying floxed APOE-ε4 or APOE-ε3 alleles. These mice were mated with mice expressing Cre recombinase under the regulation of a tamoxifen-inducible ER element and the Aldh1l1 astrocyte-specific promoter. Either tamoxifen or vehicle to induce Cre recombinase expression was administered to these Aldh1l1-CreERT2 mice at 5.5 months of age after the onset of tau pathology. Single nuclei isolated from the hippocampus of these mice were sequenced using 10×Genomics Chromium Single Cell sequencing, and the data was processed using the Cell Ranger Single Cell Software Suite (v3.0.2). The filtered count matrix generated by the Cell Ranger count pipeline for all 8 samples was processed using Seurat v4.0.481. Samples were filtered to include only cells with 500 - 2000 detected genes and <5% mitochondrial reads. The filtered samples were integrated into a single Seurat object containing a matrix of 33,457 genes by 63,248 nuclei. Normalization and variance stabilization were performed using sctransform. Clustering was determined to be performed in Seurat v4.0.4 using the RunPCA(), FindNeighbors(), and FindClusters() functions. The nearest neighbor distances were calculated using the first 15 PCs.This algorithm embeds cells into a k-nearest neighbor graph based on the Euclidean distance in the PCA space. The weights of the edges between any two cells are further refined using the Jaccard similarity. Clustering was performed using the Louvain algorithm with default settings and a resolution of 0.7, resulting in a set of 22 different clusters. The Seurat FindMarkers() function was used to detect differential gene expression by Wilcoxon rank-sum test, with min.pct = 0.1, test.use = "wilcox", and logfc.threshold = 0.05.

[0229] General statistical analysis: Differences between genotype groups were evaluated by ordinary one-way ANOVA using Tukey's multiple comparison test, which compares the mean of every other column to the mean of each column. All plotted data are shown as mean ± SEM. The correlation between two data within the same genotype group was analyzed using simple linear regression and plotted as mean ± SEM. Analyses were performed and plots were created using GraphPad Prism version 9.2.0.

[0230] References

[0231]

Table 2-1

[0232]

Table 2-2

[0233]

Table 2-3

[0234] All patents and publications referred to or mentioned in this specification indicate the technical level of those skilled in the art to which the present invention pertains, and each such referred patent or publication is specifically incorporated by reference herein to the same extent as if the whole thereof were individually incorporated by reference or the whole thereof were set forth herein. The applicant reserves the right to physically incorporate into this specification any and all materials and information from any such cited patent or publication.

[0235] The following description summarizes aspects and features of the present invention. Statement: 1. A method comprising administering one or more inhibitors of high mobility group box protein 1 (HMGB1) to a subject having at least one genomic APOE4 allele.

[0236] 2. The method of statement 1, wherein the subject has two genomic APOE4 alleles. 3. The method of statement 1 or 2, wherein the subject expresses a detectable level of APOE4 protein.

[0237] 4. The method of any one of statements 1, 2, or 3, wherein the cerebrospinal fluid of the subject has an increase in the level of high mobility group box protein 1 (HMGB1), dipeptidyl peptidase 10 (DPP10), or both HMGB1 and DPP10 as compared to a control.

[0238] 5. The method of any one of statements 1 to 4, wherein the cerebrospinal fluid of the subject has a level of HMGB1 or DPP10 that is increased by at least 25%, or at least 50%, or at least 90% as compared to a control.

[0239] 6. The method according to any one of statements 1 to 5, wherein one or more of the inhibitors of HMGB1 are glycyrrhizic acid, ethyl pyruvate, nicotine, (-)-epigallocatechin gallate (EGCG), tanshinone, chlorogenic acid, emodin-6-O-β-D-glucoside, rosmarinic acid, isorhamnetin-3-O-galactoside, persiculin, forsythoside B, chloroquinone, acteroside, cichonin, carbenoxolone, quercetin, lycopene, nafamostat mesilate, gabexate mesilate, sivelestat sodium, HMGB1 monoclonal antibody (m2G7 or #10-22), recombinant HMGB1 box A protein, acetylcholine, nicotinic acetylcholine receptor subtype alpha7 agonist GTS-21, peptide P5779, resveratrol, metformin, or a combination thereof.

[0240] 7. The method according to any one of statements 1 to 6, wherein the subject exhibits symptoms of HMGB1 nuclear-cytoplasmic transport, gliosis, neurodegeneration, tau pathology, or myelin deficiency. 8. The method according to any one of statements 1 to 7, wherein the subject exhibits symptoms of at least one tauopathy.

[0241] 9. The method according to either statement 7 or 8, wherein the tauopathy is a neurodegenerative disorder characterized by abnormal deposition of tau protein in the brain. 10. The method according to any one of statements 7 to 9, wherein the tauopathy is Alzheimer's disease, Pick's disease, progressive supranuclear palsy, corticobasal degeneration, argyrophilic grain disease, primary age-related tauopathy, chronic traumatic encephalopathy, or frontotemporal dementia.

[0242] 11. A method comprising the steps of measuring the levels of high mobility group box protein 1 (HMGB1) and dipeptidyl peptidase 10 (DPP10) in the cerebrospinal fluid of at least one subject, and administering one or more HMGB1 inhibitors to any subject in which the level of HMGB1 or DPP10 is increased as compared to a control.

[0243] 12. The method according to statement 11, wherein at least the subject has at least one genomic APOE4 allele. 13. The method according to statement 11 or 12, wherein the subject has two genomic APOE4 alleles.

[0244] 14. The method according to any one of statements 11 to 13, wherein the subject expresses a detectable level of APOE4 protein. 15. The method according to any one of statements 11 to 14, wherein the level of HMGB1 or DPP10 increased as compared to the control is a level of HMGB1 or DPP10 increased by at least 25%, or at least 50%, or at least 90% as compared to the control.

[0245] 16. One or more of the inhibitors of HMGB1 are glycyrrhizic acid, ethyl pyruvate, nicotine, (-)-epigallocatechin gallate (EGCG), tanshinone, chlorogenic acid, emodin-6-O-β-D-glucoside, rosmarinic acid, isorhamnetin-3-O-galactoside, persicarin, forsythoside B, chloroquine, acteroside, cichonin, carbenoxolone, quercetin, lycopene, nafamostat mesilate, gabexate mesilate, sivelestat sodium, HMGB1 monoclonal antibody (m2G7 or #10-22), recombinant HMGB1 box A protein, acetylcholine, nicotinic acetylcholine receptor subtype alpha7 agonist GTS-21, peptide P5779, resveratrol, metformin, or a combination thereof, according to any one of statements 11 to 15.

[0246] 17. The method according to any one of statements 11 to 16, wherein the subject exhibits symptoms of HMGB1 nuclear-cytoplasmic transport, gliosis, neurodegeneration, tau pathology, or myelin deficiency.

[0247] 18. The method according to any one of statements 11 to 17, wherein the subject exhibits symptoms of at least one tauopathy. 19. The method according to either statement 17 or 18, wherein the tauopathy is a neurodegenerative disorder characterized by abnormal deposition of tau protein in the brain.

[0248] 20. The method according to any one of statements 18 to 19, wherein the tauopathy is Alzheimer's disease, Pick's disease, progressive supranuclear palsy, corticobasal degeneration, argyrophilic grain disease, primary age-related tauopathy, chronic traumatic encephalopathy, or frontotemporal dementia.

[0249] The specific methods and compositions described herein are representative of preferred embodiments and are illustrative and are not intended as limitations on the scope of the invention. Other objects, aspects, and embodiments will occur to those skilled in the art upon consideration of this specification and are included within the spirit scope of the invention as defined by the claims. It will be readily apparent to those skilled in the art that various substitutions and modifications can be made to the invention disclosed herein without departing from the scope and spirit of the invention.

[0250] The invention as appropriately and illustratively described herein can be practiced in the absence of any element(s) or limitation(s) not specifically disclosed herein as essential. The methods and processes appropriately and illustratively described herein may be performed in a different order of steps and the methods and processes are not necessarily limited to the order of steps shown herein or in the claims.

[0251] As used in this specification and the appended claims, the singular forms "a", "an", and "the" include plural references unless the context clearly dictates otherwise. Thus, for example, reference to "a nucleic acid" or "a protein" or "a cell" includes a plurality of such nucleic acids, proteins, or cells (e.g., a solution or dried preparation of nucleic acids or expression cassettes, a solution of protein, or a population of cells). As used herein, the term "or" is used to mean non-exclusive, and "A or B" includes, unless otherwise indicated, "A but not B", "B but not A", and "A and B".

[0252] In no event shall this patent be construed as limited to the specific examples or embodiments or methods specifically disclosed herein. In no event shall this patent be construed as limited by any statement made by any examiner at the Patent and Trademark Office or any other official or employee, except where such statement is specifically and without qualification adopted in a response document by the applicant.

[0253] The terms and expressions employed are used as terms of description and not of limitation, and there is no intention, in the use of such terms and expressions, to exclude any equivalents of the features shown and described or portions thereof, it being recognized that various modifications are possible within the scope of the invention as claimed. Accordingly, while the invention has been specifically disclosed in preferred embodiments and with reference to any features, it is understood that modifications and variations of the concepts herein disclosed may be resorted to by those skilled in the art, and that such modifications and variations are considered to be within the scope of the invention as defined by the appended claims and the description of the invention.

[0254] The present invention has been described broadly and generally herein. Each of the narrower species and subgeneric classifications included in the general disclosure also forms part of the present invention. This includes the general description of the present invention having conditions or negative limitations that remove any subject matter from the genus, whether or not the deleted material is specifically described herein. Further, when features or aspects of the present invention are described with respect to Markush groups, those skilled in the art will recognize that the present invention is also described with respect to any individual member or subgroup of members of the Markush group.

[0255] Incorporation by reference of the sequence listing. The sequence listing is provided herein together with this specification as an XML file "2338683.xml" created on May 1, 2023 and having a size of 20,464 bytes. The contents of the XML file are incorporated herein by reference in their entirety.

Claims

Claim 1 A method comprising administering one or more inhibitors of high mobility group box protein 1 (HMGB1) to a subject having at least one genomic APOE4 allele. Claim 2 The method according to claim 1, wherein the subject has two genomic APOE4 alleles. Claim 3 The method according to claim 1 or 2, wherein the subject expresses a detectable level of APOE4 protein. Claim 4 The method according to any one of claims 1 to 3, wherein the cerebrospinal fluid of the subject has an increase in the level of high mobility group box protein 1 (HMGB1), dipeptidyl peptidase 10 (DPP10), or an increase in the levels of both HMGB1 and DPP10, compared to a control. Claim 5 The method according to any one of claims 1 to 4, wherein the cerebrospinal fluid of the subject has a level of HMGB1 or DPP10 that is increased by at least 25%, or at least 50%, or at least 90% compared to a control. Claim 6 One or more of the inhibitors of HMGB1 are glycyrrhizic acid, ethyl pyruvate, nicotine, (-)-epigallocatechin gallate (EGCG), tanshinone, chlorogenic acid, emodin-6-O-β-D-glucoside, rosmarinic acid, isorhamnetin-3-O-galactoside, persicarin, forsythoside B, chloroquine, acteoside, cichonin, carbenoxolone, quercetin, lycopene, nafamostat mesilate, gabexate mesilate, sivelestat sodium, an HMGB1 monoclonal antibody, such as m2G7 or #10-22, a recombinant HMGB1 box A protein, acetylcholine, a nicotinic acetylcholine receptor subtype alpha7 agonist GTS-21, peptide P5779, resveratrol, metformin, or a combination thereof, according to any one of claims 1 to 5. Claim 7 The method according to any one of claims 1 to 6, wherein the subject exhibits symptoms of HMGB1 nuclear-cytoplasmic transport, gliosis, neurodegeneration, tau pathology, or myelin deficiency. Claim 8 The method according to any one of claims 1 to 7, wherein the subject exhibits symptoms of at least one tauopathy. Claim 9 The method according to claim 7 or 8, wherein the tauopathy is a neurodegenerative disorder characterized by abnormal deposition of tau protein in the brain. Claim 10 The method according to any one of claims 7 to 9, wherein the tauopathy is Alzheimer's disease, Pick's disease, progressive supranuclear palsy, corticobasal degeneration, argyrophilic grain disease, primary age-related tauopathy, chronic traumatic encephalopathy, or frontotemporal dementia.

11. A method comprising measuring the levels of high mobility group box protein 1 (HMGB1) and dipeptidyl peptidase 10 (DPP10) in the cerebrospinal fluid of at least one subject, and administering one or more HMGB1 inhibitors to any subject in which the level of HMGB1 or DPP10 is increased as compared to a control.

12. The method according to claim 11, wherein the at least one subject has at least one genomic APOE4 allele.

13. The method according to claim 11 or 12, wherein the subject has two genomic APOE4 alleles.

14. The method according to any one of claims 11 to 13, wherein the subject expresses a detectable level of APOE4 protein.

15. The method according to any one of claims 11 to 14, wherein the level of HMGB1 or DPP10 increased as compared to the control is a level of HMGB1 or DPP10 increased by at least 25%, or at least 50%, or at least 90% as compared to the control.

16. One or more of the HMGB1 inhibitors are glycyrrhizic acid, ethyl pyruvate, nicotine, (-)-epigallocatechin gallate (EGCG), tanshinone, chlorogenic acid, emodin-6-O-β-D-glucoside, rosmarinic acid, isorhamnetin-3-O-galactoside, persicarin, forsythoside B, chloroquine, acteoside, cichonin, carbenoxolone, quercetin, lycopene, nafamostat mesilate, gabexate mesilate, sivelestat sodium, an HMGB1 monoclonal antibody, such as m2G7 or #10-22, a recombinant HMGB1 box A protein, acetylcholine, a nicotinic acetylcholine receptor subtype alpha7 agonist GTS-21, the peptide P5779, resveratrol, metformin, or a combination thereof, the method according to any one of claims 11 to 15.

17. The method according to any one of claims 11 to 16, wherein the subject exhibits symptoms of HMGB1 nuclear-cytoplasmic transport, gliosis, neurodegeneration, tau pathology, or myelin deficiency.

18. The method according to any one of claims 11 to 17, wherein the subject exhibits symptoms of at least one tauopathy.

19. The method according to claim 17 or claim 18, wherein the tauopathy is a neurodegenerative disorder characterized by abnormal deposition of tau protein in the brain.

20. The method according to claim 18 or 19, wherein the tauopathy is Alzheimer's disease, Pick's disease, progressive supranuclear palsy, corticobasal degeneration, argyrophilic grain disease, primary age-related tauopathy, chronic traumatic encephalopathy, or frontotemporal dementia.

Images