Compositions and methods for inhibiting seizures
By administering agents that modulate KD-associated gene products or pathways, the method addresses the challenges of KD implementation and variability, enhancing seizure resistance and offering a compliant alternative for treating epilepsy and related disorders.
Patent Information
- Application Number
- PCT/US2024/028328
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-10-12
- Filing Date
- 2024-05-08
- Publication Date
- 2024-11-14
AI Technical Summary
The low-carbohydrate, high-fat ketogenic diet (KD) is effective in treating refractory epilepsy but faces challenges due to difficulties in implementation, dietary compliance, and adverse side effects, with limited understanding of its mechanisms and molecular targets, leading to low patient retention and variability in responsiveness.
Administering agents that modulate gene products or pathways associated with the KD, such as those listed in Tables 1 and 2, to mimic its effects and treat conditions like seizures, epilepsy, and other neurodevelopmental disorders, using compositions that can be orally administered and formulated with pharmaceutically acceptable carriers.
The proposed method effectively mimics the KD's benefits, increasing seizure resistance and altering brain transcriptomes, providing a potential alternative or adjunct to dietary therapy with improved compliance and reduced side effects.
Smart Images

Figure IMGF000029_0001 
Figure IMGF000030_0001 
Figure IMGF000031_0001
Abstract
Description
[0001] COMPOSITIONS AND METHODS FOR INHIBITING SEIZURES
[0002] RELATED APPLICATIONS
[0003] This application claims the benefit of U.S. Provisional Application No. 63 / 465,088, filed May 9, 2023, and U.S. Provisional Application No. 63 / 543,759, filed October 12, 2023, each of which is hereby incorporated by reference in its entirety.
[0004] GOVERNMENT SUPPORT
[0005] This invention was made with government support under NS115537 awarded by the National Institutes of Health. The government has certain rights in the invention.
[0006] BACKGROUND
[0007] Epilepsy is characterized by recurrent seizures that can lead to loss of awareness, loss of consciousness, and / or disturbances of movement, autonomic function, sensation (including vision, hearing and taste), mood, and / or mental function. Epilepsy afflicts 1-2% of the population in the developed world.
[0008] The low-carbohydrate, high-fat ketogenic diet (KD) is a treatment for refractory epilepsy, wherein more than one-third of epileptic individuals do not respond to existing anticonvulsant medications. The efficacy of the KD is supported by multiple retrospective and prospective studies, which estimate that -30% of patients become seizure-free, and -60% experience significant benefit. However, despite its value for treating epilepsy and its increasing application to other disorders, including autism, Alzheimer’s disease, Parkinson’s disease, metabolic syndrome and cancer, use of the KD remains low due to difficulties with implementation, dietary compliance and adverse side effects. In fact, even with successful seizure reduction, epileptic patient retention on the KD is only an estimated 12% by the third year of dietary therapy. Moreover, mechanisms underlying the beneficial effects of the KD are poorly understood, and molecular and / or cellular targets for intervention are lacking. That the diet succeeds in controlling various types of symptoms in cases when drugs fail suggests that it enhances endogenous neuroprotective pathways that are not targeted by existing medications. Compositions and methods for replicating these effects would be highly useful.
[0009] SUMMARY
[0010] In some aspects, provided herein are methods and compositions for preventing or treating a condition responsive to a ketogenic diet in a subject, comprising administering to the subject an agent that modulates a gene product of any one of the genes listed in Table 1. In some aspects, provided herein are methods and compositions for preventing or treating a condition responsive to a ketogenic diet in a subject, comprising administering to the subject an agent that modulates a gene product of any one of the pathways listed in Table 2.
[0011] In some embodiments, the condition is seizures, optionally wherein subject has a neurodevelopmental condition, e.g., selected from epilepsy, autism spectrum disorder, Rett syndrome, attention deficit disorder, and fragile X syndrome.
[0012] In some embodiments, the subject has a condition selected from Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), cancer, stroke, a metabolic disease, a mitochondrial disorder, depression, migraines, and traumatic brain injury (TBI).
[0013] In some embodiments, the gene is selected from one or more of Acaa2, Abcc9, Aldh3bl, Ap000350.4, Ap000350.5, Ap000350.6, Arf3, Arf6, Arfipl, Atf2, Atm, Atp5h, Atp5j2, Atp5o, Atp8a2, Aurkaipl, Bazlb, Bmprlb, Briplos, Calr, Caprinl, Carmil2, Ccdcl36, Ccl6, Ccnyll, Ccnq, Cchcrl, Cdh4, Cenpu, Clic5, Col8a2, Cox5b, Crybgl, Cwc22, Cxxld, Ddx3y, Def8, Diexf, DocklO, Did, Dmkn, Dnajcl8, Duspl2, Ehmt2, Ei24, Ecm2, Ecsit, Egfir, Eloa, Emidl, Eme2, Ep300, Epn3, Epsl5, Etfl, Faml07a, Faml85a, Fbxo8, Fcerlg, Fgfrlop, Flotl, F5, Gabarapll, Gas5, Gm527, Gmcll, Gpx2, Grebl, Gstol, Gstt2, Gstpl, Gtpbp6, Hars, Hif3a, Hnmpab, Hscb, Hylsl, Iqsec3, Ino80b, Ingl, Irf2, Jmy, Kctd20, Kdm5d, Kdm6b, Lampl, Lhfpl4, Lrrcl, Lrrtml, Lonp2, Loxl3, Mif4gd, Mcf2, Mettll8, Mettl23, Mnsl, MrpllO, Mrpll8, Mrpl30, Mrpl42, Mrps24, Mybll, MyhlO, Myb, Ndorl, Ndufal2, Ndufb7, Ndufb9, Ndufv2, Nebl, Nfe211, Nfya, Nfxl, Nrlh2, Nr4a3, Nrbpl, Nrg2, Ntpcr, Nubl, Nupl53, Nuprl, Nusapl, Odf21, Ogfod3, 01fm3, Oligl, Pcedla, Pdgfa, Pdgfb, Pedsl, Pesl, Phfllb, Piercel, Phip, Ppargcla, Ppil6, Pole3, Pramell2, Prkce, Proml, Psd, Qtrtl, Rabl2, RanbplO, Rbbp4, Rbbp8, Rbpj, Rere, Rgs8, Rnfl4, Rnfl38, Rprd2, Roml, Runxltl, Rwdd3, SlOOalO, Scaf4, Serbpl, Serpinbla, Setmar, Shrooml, Slc5a3, Slc6al2, Slitl, Smiml9, Smim24, Tbx2, Tbx3osl, Tbxl8, Tceb3, Tfblm, Tmem69, Tmemll4, Tmeml84b, Tmem237, Ttil, Tpbg, Tpm3, Txnl, Ubash3b, Ufspl, Usp6nl, Uty, Utpl4a, Utpl4b, Vars2, Wdr6, Zc3havl, Zcchc2, Zfpl42, Zfp383, Zfp397, Zfp87, and Zkscanl4.
[0014] In some embodiments, the pathway is selected from one or more of Formaldehyde assimilation; Guanosine nucleotide degradation; Amino acid biosynthesis; L-proline biosynthesis; proline biosynthesis; Sugar acid biosynthesis; GDP-mannose biosynthesis; monosaccharide biosynthesis; 2-methylcitrate cycle; methylcitrate cycle I; methylcitrate cycle II; Glycol degradation; glycol metabolism; Fatty acid beta-oxidation; fatty acid oxidation; fatty acid biosynthesis; hydroxy fatty acid biosynthesis; Methionine metabolism; Betahydroxybutyrate (BHBA) metabolism; N-substituted nicotinamide metabolism; Alphalinolenic acid and linoleic acid metabolism; Phenylpropanoic acid biosynthesis; Glycine and serine metabolism; Glutathione metabolism; Betaine metabolism; Adenine and adenosine salvage; dTDP-L-rhamnose biosynthesis; Folate transformations; L-valine biosynthesis; sulfoglycolysis; Acetyl-CoA fermentation; Pyruvate fermentation; Branched-chain amino acid biosynthesis; Valine, leucine, and isoleucine biosynthesis; Phenylalanine, tyrosine, and tryptophan biosynthesis; Glycerophospholipid metabolism; Aerobactin biosynthesis; Sidophore biosynthesis; Polyamine biosynthesis; Monosaccharide metabolism; Biotin biosynthesis; Fatty acid metabolism; Hydroxy fatty acid metabolism; and Phenylpropanoic metabolism.
[0015] In some embodiments, the agent increases the activity or expression of a gene product of any one of the genes listed in Table 1. In some embodiments, the agent increases the activity or expression of a gene product of any one of the pathways listed in Table 2.
[0016] In some embodiments, the agent is a potentiator in Table 1. In some embodiments, the agent is selected from one or more of 1,1, 1,3,3, 3-hexafluoro-2-{4-[(2, 2,2- trifluoroethyl)amino]phenyl}propan-2-ol; 3-Hydroxymethyl-5-aziridinyl-lmethyl-2-[lH- indole-4, 7-dione] -propanol; 3-(hydroxymethyl)-l-methyl-5-(2-methylaziridin-l-yl)-2-phenyl- lH-indole-4, 7-dione; 5'-Guanosine-diphosphate-monothiophosphate; 5 -Methoxy- 1,2- dimethyl-3-(4-nitrophenoxymethyl)indole-4, 7-dione; Adenine; Benzylpenicilloyl Poly lysine; Benzenesulfonyl; Bishy droxy[2H-l-benzopyran-2-one,l,2-benzopyrone]; Canfosfamide; Cibacron Blue; Colforsin; Dicoumarol; Glutathione; Glutathione disulfide; Guanosine-5'- diphosphateguanosine-5'-diphosphate; Hyodeoxy cholic acid; Lauryl dimethylamine-N-oxide; L-histidine; Melatonin; Menadione; Myristic acid; N-(2,2,2-trifluoroethyl)-N-{4-[2,2,2- trifluoro-l-hydroxy-l-(trifluoromethyl)ethyl]phenyl}benzenesulfonamide; N- formylmethionine; NADH; NAV-2729; p-Nitrophenol; Pseudoephedrine; Rosiglitazon; Succinic acid; S-(4-bromobenzyl)cysteine; S-hexylglutathione; Ubiquinone-1; and Vitamin C.
[0017] In some embodiments, the agent decreases the activity or expression of a gene product of any one of the genes listed in Table 1. In some embodiments, the agent decreases the activity or expression of a gene product of any one of the pathways listed in Table 2.
[0018] In some embodiments, the agent is an inhibitor listed as an inhibitor in Table 1. In some embodiments, the agent is selected from one or more of adenosine triphosphate, Afatinib, Axitinib, alpha-Tocopherol, alpha-Tocopherol acetate, alpha-Tocopherol succinate, Amitriptyline, Anisomycin, Benzoyl peroxide, Brigatinib, Caffeine, Cannabidiol, Cholic Acid, Cladribine, Clomipramine, Curcumin, Dacomitinib, Deoxycholic acid, Dequalinium, Diazoxide, DL-alpha-Tocopherol, Drotrecogin alfa, Erlotinib, Etacrynic acid, Ezatiostat hydrochloride, Fostamatinib, Garcinol, Gefitinib, glyburide, Guvacine, Hexachlorophene, Lapatinib, Lidocaine, Midostaurin, Metformin, Omacetaxine mepesuccinate, Osimertinib, Puromycin, Racephedrine, Rhein, Roxadustat, Sgc-cbp30, Sunitinib, Tamoxifen, Trimetazidine, Vandetanib, and Zanubrutinib.
[0019] As used herein, “gene product” includes a nucleic acid (e.g., an mRNA) or a peptide encoded by the gene.
[0020] The subject may be a pediatric subject (e.g., the subject is under 18 years of age). The subject may be an adult (e.g., 18 years old or older). In some embodiments, the subject is no more than 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, or 65 years of age. In some embodiments, the condition disclosed herein may be a condition in children. Thus, the condition may be a pediatric condition. The child may be less than about 1 week old. The child may be less than about 1 month old. The child may be less than about 6 months old. The child may be less than about 12 months old. The child may be less than about 2 years old. The child may be less than about 3 years old. The child may be less than about 4 years old. The child may be less than about 5 years old. The child may be less than about 6 years old. The child may be less than about 7 years old. The child may be less than about 8 years old. The child may be less than about 9 years old. The child may be less than aboFut 10 years old. The child may be less than about 12 years old.
[0021] In some embodiments, the subject is on a diet, and the diet may be a control diet, a ketogenic diet, a high fat diet, or a low carbohydrate diet.
[0022] In some embodiments, the agent is formulated for oral delivery. In some embodiments, the agent is self-administered.
[0023] Also provided herein are compositions (e.g., compositions comprising any one or more of the agents listed in Table 1 and a pharmaceutically acceptable carrier). In some embodiments, the composition is formulated for oral delivery. In some embodiments, the composition is self-administered. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 shows that clinical KD is associated with alterations in the functional potential, but not composition, of the gut microbiota in a cohort of children with refractory epilepsy, Related to Figure 2. Figure 1A shows alpha diversity as measured by rarefaction curve (left) and observed OTUs (right) of matched donor pre-KD (n=10) and post-KD (n=10) fecal microbiota samples showing no difference in alpha-diversity, (two-way ANOVA with Sidak (left); two-tailed, Wilcoxon matched-pairs signed rank test (right)). Figure IB (left) shows the difference in total abundance of the 26 most abundant pathways between matched donor pre-KD (n=10) and post-KD(n=10) fecal microbiota samples (two-tailed, Wilcoxon matched-pairs signed rank test). Right shows the total number of observed MetaCyc functional pathways in matched donor pre-KD (n=10) and post-KD (n=10) fecal microbiota samples (two-tailed, Wilcoxon matched-pairs signed rank test). Data is displayed as mean ± SEM, unless otherwise noted. *p < 0.05, n.s = no statistical significance; KD, ketogenic diet.
[0025] Figure 2 shows the transfer of the clinical KD-associated gut microbiota from pediatric epilepsy patients to mice confers resistance to 6-Hz seizures. Figure 2A shows the experimental schematic for transplantation of human donor fecal microbiota samples into germ-free (GF) Swiss Webster (SW) mice for 6-Hz psychomotor seizure testing. Figure 2B shows 6-Hz seizure thresholds for replicate mice transplanted with human microbiota from paired donor pre-KD and post-KD samples (One-way ANOVA with Tukey’s, n = 13-15 mice per patient sample, with # denoting statistical differences when considering within-patient recipient mice as technical replicates). Figure 2C shows the average seizure thresholds of recipient mouse cohorts per patient donor sample. (Two-tailed, unpaired Welch’s t-test. n=10 patient samples per group). Figure 2D shows latency to exploration for all individual pre-KD (n=140) and post-KD (n=141) recipient mice. Figure 2E shows average current (mA) at which remission seizures were observed per patient donor sample (Two-tailed, unpaired Welch’s t-test. n = 10 patient samples per group). Data is displayed as mean ± SEM, unless otherwise noted. *** < 0.001, **** / ? < 0.0001,####p <0.0001 (within-patient mouse recipients); KD, ketogenic diet.
[0026] Figure 3 shows that antibiotic treatment abrogates the seizure protective effects of inoculation with the clinical KD-associated human gut microbiome, Related to Figure 2. Figure 3 A shows the experimental schematic for transfer of human donor fecal microbiota sample into conventionally colonized (SPF) mice pre-treated with oral antibiotic (Abx) for 7 days, followed by 6-Hz psychomotor seizure testing. Figure 3B shows 6-Hz seizure thresholds for Abx-treated mice inoculated with patient 3 post-KD human microbiota (n=14) compared to GF mice inoculated with patient 3 post-KD (n=14) and patient 3 pre-KD (n=14) relative control fecal microbiota from Figure 2B. (One-way ANOVA with Tukey’s, with # denoting statistical differences when considering within-patient recipient mice as technical replicates). Figure 3C shows the latency to exploration for inoculated Abx-treated (n=14) vs inoculated GF mice from Figure 2B. Figure 3D shows the experimental schematic for transfer of human donor fecal microbiota samples to germ-free (GF) mice, followed by 5 days of oral Abx or vehicle (Veh) treatment, and then 6-Hz psychomotor seizure testing. Figure 3E shows bacterial loads as measured by quantitative PCR of the 16S rRNA gene from fecal pellets collected once daily before and during Abx or Veh treatment (two-way ANOVA with Sidak, n=3 cages of 3 mice each). Figure 3F shows 6-Hz seizure thresholds for mice inoculated with patient 1 post-KD human microbiota treated with Abx (n=12) or Veh (n=14). Reference lines denote seizure thresholds for mice inoculated with patient 1 post-KD (n=14) and pre-KD (n=14) relative control fecal microbiota from Figure 2B (Oneway ANOVA with Tukey’s, with # denoting statistical differences when considering within- patient recipient mice as technical replicates). Figure 3G shows latency to exploration for each Abx (n=12) and Veh (n=14) mouse that underwent 6-Hz psychomotor seizure testing. Data is displayed as mean ± SEM, unless otherwise noted. (One-way ANOVA with Tukey’s, with # denoting statistical differences when considering within-patient recipient mice as technical replicates) ****p < 0.0001.####p <0.0001 (for within-patient mouse recipients); SPF, specific pathogen free; SW, Swiss Webster; Abx, antibiotics; KD, ketogenic diet; GF, germ free; Veh, vehicle.
[0027] Figure 4 shows that sterile filtration prevents the seizure protective effects of transfer of the clinical KD-associated human gut microbiome, Related to Figure 2. Figure 4A shows the experimental design for administration of human donor fecal filtrate samples to germ-free (GF) mice, followed by 6-Hz seizure testing 4 days later. Figure 4B shows the seizure thresholds for mice treated with sterile filtered patient 2 pre-KD (n=14) filtrate and sterile filtered patient 2 post-KD (n=13) filtrate fecal samples compared to seizure thresholds for mice transplanted with unfiltered patient 2 post-KD (n=14) and unfiltered patient 2 pre-KD (n=14) relative control fecal microbiota from Figure 2B (One-way ANOVA with Tukey’s, with # denoting statistical differences when considering within-patient recipient mice as technical replicates). Figure 4C shows latency to exploration for mice treated with sterile filtered pre-KD (n=14) and sterile filtered post-KD (n=13) that underwent 6-Hz psychomotor seizure testing. Figure 4D shows the experimental design for administration of human donor fecal filtrate samples to germ-free (GF) mice, followed by 6-Hz seizure testing 2 hours later. Figure 4E shows the seizure thresholds for mice treated with sterile filtered patient 6 pre-KD filtrate (n=13) and sterile filtered patient 6 post-KD filtrate (n=14) fecal samples. Reference lines denote seizure thresholds for mice transplanted with unfiltered patient 6 post-KD (n=14) and unfiltered patient 6 pre-KD (n=14) relative control fecal microbiota from Figure 2B. (One-way ANOVA with Tukey’s, with # denoting statistical differences when considering within-patient recipient mice as technical replicates). Figure 4F shows the latency to exploration for mice treated with sterile filtered pre-KD (n=13) and sterile filtered post-KD (n=14) that underwent 6-Hz psychomotor seizure testing. Figure 4G shows the seizure threshold comparisons for mouse cohorts colonized with pre-KD or post-KD human fecal microbiota or treated with sterile filtered post-KD filtrate from Patients 4, 6, and 7. (One-way ANOVA with Tukey’s). Data is displayed as mean ± SEM, unless otherwise noted. *p<0.05, **p<0.01, n.s = no statistical significance; ####p <0.0001 (for within-patient mouse recipients); KD, ketogenic diet.
[0028] Figure 5 shows the taxonomic fidelity of human microbiota transfer to mice, Related to Figure 2. Figure 5A shows the Shannon index alpha-diversity of fecal microbiota from human donor pre-KD and post-KD samples (left) and matched mouse recipient pre-KD and post-KD samples (right) (two-tailed, Wilcoxon matched-pairs signed rank test, donors: n=10 patients, recipients: n=10 per patient condition, where each n is an average from 4-5 cages per patient). Figure 5B shows the Shannon index alpha-diversity of fecal microbiota from all human donor samples (n=20 patients) and all matched mouse recipient samples (two-tailed, Wilcoxon matched-pairs signed rank test; n=20 patient conditions, where each n is an average from 4-5 cages per patient). Figure 5C shows the alpha diversity as measured by rarefaction curve (left) and observed OTUs (right) of mouse recipient pre-KD (n=10, where each sample is the average of 4-5 recipient mice per donor patient sample) and post-KD (n=10, where each sample is the average of 4-5 recipient mice per donor patient sample) fecal microbiota samples showing no difference in alpha-diversity, (two-way ANOVA with Sidak (left); two- tailed, Wilcoxon matched-pairs signed rank test (right)).
[0029] Figure 6 shows the clinical KD-associated human microbiome exhibits functional alterations that are phenocopied in seizure-protected recipient mice. Microbial functional pathways were differentially abundant as determined by MaAsLin2 analysis comparing post- KD (samples relative to pre-KD controls for either donor fecal samples or recipient fecal samples (donors: n=10 per diet condition; recipients: n=10 per donor diet condition [50 total, where each n reflects average of 5 technical replicate recipient mice per donor patient sample]). Red font denotes pathways that are commonly differentially abundant in the same direction in both donors fed KD and recipient mice fed CD (all pathways listed had minimum p<0.1; *denotes pathways with p<0.05). Figure 6A shows microbial functional pathways that are commonly differentially abundant by MaAsLin2 analysis in the same direction in post- KD donor and recipient samples (donors: n=10 per diet condition; recipients: n=10 per donor diet condition [50 total, where each n reflects average of 5 technical replicate recipient mice per donor patient sample]). Figure 6B shows beta-hydroxybutyrate (BHBA) and glucose levels in human donor (left) and mouse recipient (right) pre-KD and post-KD feces and serum (two-tailed Wilcoxon matched-pairs signed rank test; donors: n=10 per diet condition; recipients: n=10 per donor diet condition [50 total, where each n reflects average of 5 technical replicate recipient mice per donor patient sample]). Figure 6C shows metabolite set enrichment analysis showing the top 25 enriched chemical subclasses ordered by p-value for the set of differentially abundant metabolites in human donor (left) post-KD vs pre-KD fecal samples (p<0.05, two-tailed, matched pairs Student’s t-test, n=10 per diet condition).
[0030] Metabolite set enrichment analysis showing enriched chemical subclasses ordered by p-value for the set of differentially abundant metabolites in recipient mouse (right) post-KD vs pre- KD fecal samples (p<0.05, matched pairs Student’s t-test, n=10 per patient diet condition, where each sample is pooled from 5 recipient mice per donor patient sample). Red font denotes chemical subclasses altered in post-KD vs pre-KD human donor feces that are shared with those differentially regulated in post-KD vs pre-KD mouse recipient feces. Orange asterisks (*) denote additional chemical subclasses that are relevant to KD based on existing literature. Figure 6D shows metabolite set enrichment analysis showing the top 25 enriched SMPDB pathways by p-value for the set of differentially abundant metabolites in human donor (left) post-KD vs pre-KD fecal samples (p<0.05, matched pairs Student’s t-test, n=10 per patient diet condition). Metabolite set enrichment analysis showing enriched SMPDB pathways by p-value for the set of differentially abundant metabolites in recipient mouse (right) post-KD vs pre-KD fecal samples (p<0.05, matched pairs Student’s t-test; n=10 per patient diet condition, where each sample is pooled from 5 recipient mice per donor patient sample). Red font denotes metabolic pathways altered in post-KD vs pre-KD human donor feces that are shared with those differentially regulated in post-KD vs pre-KD mouse recipient feces. Orange asterisks (*) denote additional chemical subclasses that are relevant to KD based on existing literature. Data is displayed as mean ± SEM, unless otherwise noted. *p < 0.05, ***p < 0.001, n.s.=not statistically significant; KD, ketogenic diet; BHBA, betahydroxybutyrate; CD, control diet; SMPDB, The Small Molecule Pathway Database.
[0031] Figure 7 shows the clinical KD alters metabolomic profiles in human fecal samples and in fecal and serum samples of mice inoculated with human microbiota, Related to 6. Figure 7A shows differentially abundant metabolites (p<0.05) in post-KD compared to pre- KD samples of human donor feces, mouse recipient feces, and mouse recipient serum (Two- tailed matched pairs Student’s t-test, n=10 per condition, where each recipient sample is pooled from 5 recipient mice per donor patient sample). Figure 7B shows differentially abundant metabolites (p<0.05) in human feces (post-KD compared to pre-KD) and feces of mice fed KD vs. CD chow for 14 days (Human fecal: Two-tailed matched pairs Student’s t- test, n=10 per condition, where each recipient sample is pooled from 5 recipient mice per donor patient sample; Mouse fecal: two-way ANOVA with contrasts, n=8 per condition; Fisher’s Exact Test). Figure 7C shows differentially abundant metabolites (p<0.05) in post- KD compared to pre-KD samples of human donor feces, which were also significantly altered in conventional mice (SPF) fed KD chow or vitamin- and mineral- matched control diet (CD) for 14 days. Red font denotes the subset of metabolites that were further altered by pretreating KD chow-fed mice with antibiotics (Abx) to deplete gut bacteria, (human: Two- tailed matched pairs Student’s t-test, n=10 per condition; mouse: ANOVA contrasts, n=8 per condition). Figure 7D shows metabolite set enrichment analysis of chemical subclass for the 20 differentially abundant metabolites (p<0.05, matched pairs Student’s t-test) found in both human post-KD vs pre-KD fecal samples and SFP mouse KD vs CD fecal samples (left) (human: n=10 per condition, where each sample is pooled from 5 recipient mice per donor patient sample; mouse: n=8 per condition). Metabolite set enrichment analysis of SMPDB pathways for the 20 differentially abundant metabolites (p<0.05, matched pairs Student’s t- test) found in both human post-KD vs pre-KD fecal samples and SFP mouse KD vs CD fecal samples (right) (human: n=10 per condition, where each sample is pooled from 5 recipient mice per donor patient sample; mouse: n=8 per condition). Figure 7E shows metabolite set enrichment analysis of chemical subclass for differentially abundant metabolites (p<0.05, matched pairs Student’s t-test) in recipient mouse post-KD vs pre-KD serum samples (left) (n=10 per condition, where each sample is pooled from 5 recipient mice per donor patient sample). Metabolite set enrichment analysis of SMPDB pathways for differentially abundant metabolites (p<0.05, matched pairs Student’s t-test) in recipient mouse post-KD vs pre-KD serum samples (right) (n=10 per condition, where each sample is pooled from 5 recipient mice per donor patient sample). Red font denotes metabolic pathways altered in post-KD vs pre-KD mouse serum that are shared with those differentially regulated in post-KD vs pre- KD mouse feces and / or human feces. Orange asterisks (*) denote additional chemical subclasses that are relevant to KD based on existing literature. Data is displayed as mean ± SEM, unless otherwise noted. KD, ketogenic diet; SPF, specific pathogen free; CD, control diet; Abx, antibiotics; SMPDB, The Small Molecule Pathway Database.
[0032] Figure 8 shows the seizure resistance in mice inoculated with the post-KD microbiota is associated with alterations in the brain transcriptome. Figure 8A shows GO: Biological Process gene ontology of differentially expressed genes (p<0.05) in recipient mouse post-KD (n=10, where each sample is pooled from 6 recipient mice per donor patient sample) compared to pre-KD hippocampal samples, top 20 ranked by p-value (n=10 per patient diet condition, where each sample is pooled from 6 recipient mice per donor patient sample). Figure 8B shows a heatmap of Euclidian row and column clustered top 25 differentially expressed genes in recipient mouse post-KD compared to pre-KD hippocampus ranked by p- value, smallest to largest, and with log2 fold-change >2 (n=10 per patient diet condition, where each sample is pooled from 6 recipient mice per donor patient sample). Figure 8C shows functional enrichment of top MCL sub-network clusters from hippocampal transcriptomics STRING network analysis, proteins are colored based on their overall log2FC. Circles marked with an X indicate negative log2FC. If log2FC >3 or <-3, the value is listed next to the node name (n=10 per patient diet condition, where each sample is pooled from 6 recipient mice per donor patient sample). KD, ketogenic diet; GO, gene ontology; MCL, Markov Cluster Algorithm.
[0033] Figure 9 shows mice inoculated with the post-KD microbiota exhibit alterations in the frontal cortical transcriptome, Related to Figure 8. Figure 9A shows GO: Biological Process gene ontology of differentially expressed genes (p<0.05) in recipient mouse post-KD compared to pre-KD frontal cortex samples, top 20 ranked by p-value (n=10 per patient diet condition, where each sample is pooled from 6 recipient mice per donor patient sample). Figure 9B shows a heatmap of top 25 differentially expressed genes in recipient mouse post- KD compared to pre-KD frontal cortex ranked by p-value, smallest to largest, with log2-fold change >2 (n=10 per patient diet condition, where each sample is pooled from 6 recipient mice per donor patient sample). Figure 9C shows functional enrichment of top MCL subnetwork clusters from frontal cortex transcriptomics STRING network analysis, proteins are colored based on their overall log2FC. Circles marked with an X indicate negative log2FC. If log2FC >3 or <-3, the value is listed next to the node name (n=10 per patient diet condition, where each sample is pooled from 6 recipient mice per donor patient sample). KD, ketogenic diet; GO, gene ontology; MCL, Markov Cluster Algorithm.
[0034] Figure 10 shows multi’omic network analysis identifies key microbial genomic pathways and microbially modulated metabolites directly and indirectly associated with differential expression of hippocampal transcripts. MMVEC based co-occurrence network constructed from =human donor pre-KD and post-KD fecal metagenomic and fecal metabolomic datasets and mouse recipient pre-KD and post-KD fecal metagenomic, fecal metabolomic, serum metabolomic, hippocampal transcriptomic, and frontal cortical transcriptomic datasets. wKDA analyses was performed on the network. Figure 10A shows differentially regulated genes in each area. Figure 10B It shows a table of top associated genes from epilepsy GWAS mapping onto mouse recipient hippocampal and frontal cortical DEGs (n=10 per patient diet condition, where each sample is pooled from 6 recipient mice per donor patient sample). MMVEC, microbe-metabolite vectors; KD, ketogenic diet; wKDA, weighted key driver analysis; GWAS genome-wide association study; DEGs, differentially expressed genes.
[0035] DETAILED DESCRIPTION
[0036] Provided herein are methods and compositions for mimicking the effects of a ketogenic diet by administering compositions provided herein.
[0037] In some aspects, provided herein are methods and compositions for preventing or treating a condition responsive to a ketogenic diet in a subject, comprising administering to the subject an agent that modulates a gene product of any one of the genes listed in Table 1.
[0038] Also provided herein are compositions (e.g., compositions comprising any one of the agents listed in Table 1 and a pharmaceutically acceptable carrier).
[0039] In certain embodiments, the methods and compositions are for the treatment or prevention of seizures in a subject (e.g., a subject with a neurodevelopmental disorder, such as an autism spectrum disorder, Rett syndrome, fragile X, attention deficit disorder (ADD), attention-deficit / hyperactivity disorder (ADHD), refractory epilepsy, such as pediatric refractory epilepsy, and / or non-refractory epilepsy). In other embodiments, the methods and compositions are for preventing or treating a condition (e.g., autism spectrum disorder, epilepsy, seizures, Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), cancer, stroke, a metabolic disease (e.g., obesity or diabetes), a mitochondrial disorder, depression, migraines (e.g., chronic migraines), Rett syndrome, attention deficit disorder, fragile X syndrome, or traumatic brain injury (TBI)) in a subject.
[0040] Definitions
[0041] As used herein in the specification, "a" or "an" may mean one or more. As used herein in the claim(s), when used in conjunction with the word "comprising", the words "a" or "an" may mean one or more than one. As used herein “another” may mean at least a second or more.
[0042] The term “agent” is used herein to denote a chemical compound, a small molecule, a mixture of chemical compounds and / or a biological macromolecule (such as a nucleic acid, an antibody, an antibody fragment, a protein or a peptide) Agents, as disclosed herein, modulate a gene product of any one of the genes listed in Table 1. The activity of such agents may render them suitable as a “therapeutic agent” which is a biologically, physiologically, or pharmacologically active substance (or substances) that acts locally or systemically in a subject.
[0043] Unless otherwise specified here within, the terms “antibody” and “antibodies” broadly encompass naturally-occurring forms of antibodies (e.g. IgG, IgA, IgM, IgE) and recombinant antibodies such as single-chain antibodies, chimeric and humanized antibodies and multi-specific antibodies, as well as fragments and derivatives of all of the foregoing, which fragments and derivatives have at least an antigenic binding site. Antibody derivatives may comprise a protein or chemical moiety conjugated to an antibody. Antibodies and antigen binding fragments disclosed herein may, for example, specifically bind and modulate a peptide gene product of any one of the genes listed in Table 1.
[0044] The term “antibody” as used herein also includes an “antigen-binding portion” of an antibody (or simply “antibody portion” . The term “antigen-binding portion”, as used herein, refers to one or more fragments of an antibody that retain the ability to specifically bind to an antigen (e.g., a biomarker polypeptide or fragment thereol). It has been shown that the antigen-binding function of an antibody can be performed by fragments of a full-length antibody. Examples of binding fragments encompassed within the term “antigen-binding portion” of an antibody include (i) a Fab fragment, a monovalent fragment consisting of the VL, VH, CL and CHI domains; (ii) a F(ab')2 fragment, a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region; (iii) a Fd fragment consisting of the VH and CHI domains; (iv) a Fv fragment consisting of the VL and VH domains of a single arm of an antibody, (v) a dAb fragment (Ward et al., (1989) Nature 341:544-546), which consists of a VH domain; and (vi) an isolated complementarity determining region (CDR). Furthermore, although the two domains of the Fv fragment, VL and VH, are coded for by separate genes, they can be joined, using recombinant methods, by a synthetic linker that enables them to be made as a single protein chain in which the VL and VH regions pair to form monovalent polypeptides (known as single chain Fv (scFv); see e.g., Bird et al.
[0045] (1988) Science 242:423-426; and Huston et al. (1988) Proc. Natl. Acad. Sci. USA 85:5879- 5883; and Osbourn et al. 1998, Nature Biotechnology 16: 778). Such single chain antibodies are also intended to be encompassed within the term “antigen-binding portion” of an antibody. Any VH and VL sequences of specific scFv can be linked to human immunoglobulin constant region cDNA or genomic sequences, in order to generate expression vectors encoding complete IgG polypeptides or other isotypes. VH and VL can also be used in the generation of Fab, Fv or other fragments of immunoglobulins using either protein chemistry or recombinant DNA technology. Other forms of single chain antibodies, such as diabodies are also encompassed. Diabodies are bivalent, bispecific antibodies in which VH and VL domains are expressed on a single polypeptide chain, but using a linker that is too short to allow for pairing between the two domains on the same chain, thereby forcing the domains to pair with complementary domains of another chain and creating two antigen binding sites (see e.g., Holliger et al. (1993) Proc. Natl. Acad. Sci. U.SA. 90:6444- 6448; Poljak c / o / . (1994) Structure 2:1121-1123).
[0046] An antibody for use in the instant invention may be a bispecific antibody. A bispecific antibody has binding sites for two different antigens within a single antibody polypeptide. Antigen binding may be simultaneous or sequential. Triomas and hybrid hybridomas are two examples of cell lines that can secrete bispecific antibodies. Examples of bispecific antibodies produced by a hybrid hybridoma or a trioma are disclosed in U.S. Patent 4,474,893. Bispecific antibodies have been constructed by chemical means (Staerz et al. (1985) Nature 314:628, and Perez et al. (1985) Nature 316:354) and hybridoma technology (Staerz and Bevan (1986) Proc. Natl. Acad. Sci. USA, 83:1453, and Staerz and Bevan (1986) Immunol. Today 7:241). Bispecific antibodies are also described in U.S. Patent 5,959,084. Fragments of bispecific antibodies are described in U.S. Patent 5,798,229.
[0047] Bispecific agents can also be generated by making heterohybridomas by fusing hybridomas or other cells making different antibodies, followed by identification of clones producing and co-assembling both antibodies. They can also be generated by chemical or genetic conjugation of complete immunoglobulin chains or portions thereof such as Fab and Fv sequences.
[0048] Still further, an antibody or antigen-binding portion thereof may be part of larger immunoadhesion polypeptides, formed by covalent or noncovalent association of the antibody or antibody portion with one or more other proteins or peptides. Examples of such immunoadhesion polypeptides include use of the streptavidin core region to make a tetrameric scFv polypeptide (Kipriyanov et al. (1995) Human Antibodies and Hybridomas 6:93-101) and use of a cysteine residue, biomarker peptide and a C-terminal polyhistidine tag to make bivalent and biotinylated scFv polypeptides (Kipriyanov et al. (1994) o / . Immunol. 31:1047-1058). Antibody portions, such as Fab and F(ab')2 fragments, can be prepared from whole antibodies using conventional techniques, such as papain or pepsin digestion, respectively, of whole antibodies. Moreover, antibodies, antibody portions and immunoadhesion polypeptides can be obtained using standard recombinant DNA techniques, as described herein.
[0049] Antibodies may also be “humanized, ” which is intended to include antibodies made by a non-human cell having variable and constant regions which have been altered to more closely resemble antibodies that would be made by a human cell. For example, by altering the non-human antibody amino acid sequence to incorporate amino acids found in human germline immunoglobulin sequences. The humanized antibodies of the disclosure may include amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutation in vivo), for example in the CDRs. The term “humanized antibody”, as used herein, also includes antibodies in which CDR sequences derived from the germline of another mammalian species, such as a mouse, have been grafted onto human framework sequences.
[0050] The terms “antigen-binding fragment” and “antigen-binding portion” of an antibody, as used herein, refers to one or more fragments of an antibody that retain the ability to bind to an antigen. Examples of binding fragments encompassed within the term "antigen-binding fragment" of an antibody include Fab, Fab', F(ab')2, Fv, scFv, disulfide linked Fv, Fd, diabodies, single-chain antibodies, NANOBODIES®, isolated CDRH3, and other antibody fragments that retain at least a portion of the variable region of an intact antibody. These antibody fragments can be obtained using conventional recombinant and / or enzymatic techniques and can be screened for antigen binding in the same manner as intact antibodies. The terms “CDR”, and its plural “CDRs”, refer to a complementarity determining region (CDR) of an antibody or antibody fragment, which determine the binding character of an antibody or antibody fragment. In most instances, three CDRs are present in a light chain variable region (CDRL1, CDRL2 and CDRL3) and three CDRs are present in a heavy chain variable region (CDRH1, CDRH2 and CDRH3). CDRs contribute to the functional activity of an antibody molecule and are separated by amino acid sequences that comprise scaffolding or framework regions. Among the various CDRs, the CDR3 sequences, and particularly CDRH3, are the most diverse and therefore have the strongest contribution to antibody specificity. There are at least two techniques for determining CDRs: (1) an approach based on cross-species sequence variability (i.e., Kabat et al., Sequences of Proteins of Immunological Interest (National Institute of Health, Bethesda, Md. (1987), incorporated by reference in its entirety); and (2) an approach based on crystallographic studies of antigen-antibody complexes (Chothia et al., Nature, 342:877 (1989), incorporated by reference in its entirety).
[0051] As used herein, the term “humanized antibody' refers to an antibody that has at least one CDR derived from a mammal other than a human, and a FR region and the constant region of a human antibody. A humanized antibody is useful as an effective component in a therapeutic agent since antigenicity of the humanized antibody in human body is lowered.
[0052] The phrase “pharmaceutically-acceptable carrier” as used herein means a pharmaceutically-acceptable material, composition or vehicle, such as a liquid or solid filler, diluent, excipient, solvent or encapsulating material. Each carrier must be “acceptable” in the sense of being compatible with the other ingredients of the formulation and not injurious to the subject. Some examples of materials which can serve as pharmaceutically-acceptable carriers include: (1) sugars, such as lactose, glucose and sucrose; (2) starches, such as com starch and potato starch; (3) cellulose, and its derivatives, such as sodium carboxymethyl cellulose, ethyl cellulose and cellulose acetate; (4) powdered tragacanth; (5) malt; (6) gelatin; (7) talc; (8) excipients, such as cocoa butter and suppository waxes; (9) oils, such as peanut oil, cottonseed oil, safflower oil, sesame oil, olive oil, com oil and soybean oil; (10) glycols, such as propylene glycol; (11) polyols, such as glycerin, sorbitol, mannitol and polyethylene glycol; (12) esters, such as ethyl oleate and ethyl laurate; (13) agar; (14) buffering agents, such as magnesium hydroxide and aluminum hydroxide; (15) alginic acid; (16) pyrogen-free water; (17) isotonic saline; (18) Ringer's solution; (19) ethyl alcohol; (20) phosphate buffer solutions; and (21) other non-toxic compatible substances employed in pharmaceutical formulations. The term “preventing” is art-recognized, and when used in relation to a condition, such as a local recurrence, is well understood in the art, and includes administration of a composition which reduces the frequency of, or delays the onset of, symptoms of a medical condition in a subject relative to a subject which does not receive the composition. Thus, prevention of seizures includes, for example, reducing the number of seizures in a population of patients receiving a prophylactic treatment relative to an untreated control population, e.g., by a statistically and / or clinically significant amount.
[0053] A “guide RNA” or “gRNA” is an RNA molecule that binds to a Cas protein (e.g., Cas9 protein) and targets the Cas protein to a specific location within a target DNA. Guide RNAs can comprise two segments: a “DNA-targeting segment” and a “protein-binding segment.” “Segment” includes a section or region of a molecule, such as a contiguous stretch of nucleotides in an RNA. Some gRNAs, such as those for Cas9, can comprise two separate RNA molecules: an “activator-RNA” (e.g, tracrRNA) and a “targeter-RNA” (e.g, CRISPR RNA or crRNA). Other gRNAs are a single RNA molecule (single RNA polynucleotide), which can also be called a “single-molecule gRNA,” a “single-guide RNA,” or an “sgRNA.” See, e.g, WO 2013 / 176772, WO 2014 / 065596, WO 2014 / 089290, WO 2014 / 093622, WO 2014 / 099750, WO 2013 / 142578, and WO 2014 / 131833, each of which is herein incorporated by reference in its entirety for all purposes. For Cas9, for example, a single-guide RNA can comprise a crRNA fused to a tracrRNA (e.g. , via a linker). In some embodiments, an agent disclosed herein is a gRNA (e.g., a gRNA that binds and modulates any one of the genes in Table 1).
[0054] The term “guide RNA target sequence” as used herein refers specifically to the sequence on the non-compl ementary strand corresponding to (i.e., the reverse complement of) the sequence to which the guide RNA hybridizes on the complementary strand. That is, the guide RNA target sequence refers to the sequence on the non-complementary strand adjacent to the PAM (e.g, upstream or 5’ of the PAM in the case of Cas9). A guide RNA target sequence is equivalent to the DNA-targeting segment of a guide RNA, but with thymines instead of uracils. As one example, a guide RNA target sequence for a SpCas9 enzyme can refer to the sequence upstream of the 5’-NGG-3’ PAM on the non- complementary strand.
[0055] The terms “polynucleotide”, and “nucleic acid” are used interchangeably. They refer to a polymeric form of nucleotides of any length, either deoxyribonucleotides or ribonucleotides, or analogs thereof. Polynucleotides may have any three-dimensional structure, and may perform any function, known or unknown. The following are non-limiting examples of polynucleotides: coding or non-coding regions of a gene or gene fragment, loci (locus) defined from linkage analysis, exons, introns, messenger RNA (mRNA), transfer RNA, ribosomal RNA, ribozymes, cDNA, recombinant polynucleotides, branched polynucleotides, plasmids, vectors, isolated DNA of any sequence, isolated RNA of any sequence, nucleic acid probes, and primers. A polynucleotide may comprise modified nucleotides, such as methylated nucleotides and nucleotide analogs. If present, modifications to the nucleotide structure may be imparted before or after assembly of the polymer. The sequence of nucleotides may be interrupted by non-nucleotide components. A polynucleotide may be further modified, such as by conjugation with a labeling component. The term “recombinant” polynucleotide means a polynucleotide of genomic, cDNA, semisynthetic, or synthetic origin which either does not occur in nature or is linked to another polynucleotide in a non-natural arrangement.
[0056] As used herein, “interfering nucleic acids" generally include a sequence of cyclic subunits, each bearing a base-pairing moiety, linked by intersubunit linkages that allow the base-pairing moieties to hybridize to a target sequence in a nucleic acid (typically an RNA) by Watson-Crick base pairing, to form a nucleic acid: oligomer heteroduplex within the target sequence. Interfering RNA molecules include, but are not limited to, antisense molecules, siRNA molecules, single-stranded siRNA molecules, miRNA molecules and shRNA molecules. In some embodiments, an agent described herein may be an interfering nucleic acid agent that binds and inhibits a gene product (e.g., an mRNA product) of any one of the genes listed in Table 1.
[0057] Typically at least 17, 18, 19, 20, 21, 22 or 23 nucleotides of the complement of the target mRNA sequence are sufficient to mediate inhibition of a target transcript. Perfect complementarity is not necessary. In some embodiments, the interfering nucleic acid molecule is double-stranded RNA. The double-stranded RNA molecule may have a 2 nucleotide 3’ overhang. In some embodiments, the two RNA strands are connected via a hairpin structure, forming a shRNA molecule. shRNA molecules can contain hairpins derived from microRNA molecules. For example, an RNAi vector can be constructed by cloning the interfering RNA sequence into a pCAG-miR30 construct containing the hairpin from the miR30 miRNA. RNA interference molecules may include DNA residues, as well as RNA residues. Interfering nucleic acid molecules provided herein can contain RNA bases, non-RNA bases or a mixture of RNA bases and non-RNA bases. For example, interfering nucleic acid molecules provided herein can be primarily composed of RNA bases but also contain DNA bases or non-naturally occurring nucleotides.
[0058] The interfering nucleic acids can employ a variety of oligonucleotide chemistries. Examples of oligonucleotide chemistries include, without limitation, peptide nucleic acid (PNA), linked nucleic acid (LNA), phosphorothioate, 2’0-Me-modified oligonucleotides, and morpholino chemistries, including combinations of any of the foregoing. In general, PNA and LNA chemistries can utilize shorter targeting sequences because of their relatively high target binding strength relative to 2’0-Me oligonucleotides. Phosphorothioate and 2’0-Me- modified chemistries are often combined to generate 2’0-Me-modified oligonucleotides having a phosphorothioate backbone. See, e.g, PCT Publication Nos. WO / 2013 / 112053 and WO / 2009 / 008725, incorporated by reference in their entireties.
[0059] Peptide nucleic acids (PNAs) are analogs of DNA in which the backbone is structurally homomorphous with a deoxyribose backbone, consisting of N-(2-aminoethyl) glycine units to which pyrimidine or purine bases are attached. PNAs containing natural pyrimidine and purine bases hybridize to complementary oligonucleotides obeying Watson- Crick base-pairing rules, and mimic DNA in terms of base pair recognition (Egholm, Buchardt et al. 1993). The backbone of PNAs is formed by peptide bonds rather than phosphodiester bonds, making them well-suited for antisense applications (see structure below). The backbone is uncharged, resulting in PNA / DNA or PNA / RNA duplexes that exhibit greater than normal thermal stability. PNAs are not recognized by nucleases or proteases.
[0060] Despite a radical structural change to the natural structure, PNAs are capable of sequence-specific binding in a helix form to DNA or RNA. Characteristics of PNAs include a high binding affinity to complementary DNA or RNA, a destabilizing effect caused by single-base mismatch, resistance to nucleases and proteases, hybridization with DNA or RNA independent of salt concentration and triplex formation with homopurine DNA. PANAGENE™ has developed its proprietary Bts PNA monomers (Bts; benzothiazole-2- sulfonyl group) and proprietary oligomerization process. The PNA oligomerization using Bts PNA monomers is composed of repetitive cycles of deprotection, coupling and capping. PNAs can be produced synthetically using any technique known in the art. See, e.g, U.S. Pat. Nos. 6,969,766, 7,211,668, 7,022,851, 7,125,994, 7,145,006 and 7,179,896. See also U.S. Pat. Nos. 5,539,082; 5,714,331; and 5,719,262 for the preparation of PNAs. Further teaching of PNA compounds can be found in Nielsen et al., Science, 254:1497-1500, 1991. Each of the foregoing is incorporated by reference in its entirety.
[0061] Interfering nucleic acids may also contain “locked nucleic acid” subunits (LNAs). “LNAs” are a member of a class of modifications called bridged nucleic acid (BNA). BNA is characterized by a covalent linkage that locks the conformation of the ribose ring in a C30- endo (northern) sugar pucker. For LN A, the bridge is composed of a methylene between the 2’-0 and the 4’-C positions. LNA enhances backbone preorganization and base stacking to increase hybridization and thermal stability.
[0062] The structures of LNAs can be found, for example, in Wengel, et al., Chemical Communications (1998) 455; Tetrahedron (1998) 54:3607, and Accounts of Chem. Research (1999) 32:301); Obika, et al., Tetrahedron Letters (1997) 38:8735; (1998) 39:5401, and Bioorganic Medicinal Chemistry (2008) 16:9230. Compounds provided herein may incorporate one or more LNAs; in some cases, the compounds may be entirely composed of LNAs. Methods for the synthesis of individual LNA nucleoside subunits and their incorporation into oligonucleotides are described, for example, in U.S. Pat. Nos. 7,572,582, 7,569,575, 7,084,125, 7,060,809, 7,053,207, 7,034,133, 6,794,499, and 6,670,461, each of which is incorporated by reference in its entirety. Typical intersubunit linkers include phosphodiester and phosphorothioate moieties; alternatively, non-phosphorous containing linkers may be employed. One embodiment is an LNA containing compound where each LNA subunit is separated by a DNA subunit. Certain compounds are composed of alternating LNA and DNA subunits where the intersubunit linker is phosphorothioate.
[0063] “Phosphorothioates” (or S-oligos) are a variant of normal DNA in which one of the nonbridging oxygens is replaced by a sulfur. The sulfurization of the intemucleotide bond reduces the action of endo-and exonucleases including 5’ to 3’ and 3’ to 5’ DNA POL 1 exonuclease, nucleases SI and Pl, RNases, serum nucleases and snake venom phosphodiesterase. Phosphorothioates are made by two principal routes: by the action of a solution of elemental sulfur in carbon disulfide on a hydrogen phosphonate, or by the method of sulfurizing phosphite triesters with either tetraethylthiuram disulfide (TETD) or 3H-1,2- bensodithiol-3-one 1,1-dioxide (BDTD) (see, e.g, Iyer et al., J. Org. Chem. 55, 4693-4699, 1990). The latter methods avoid the problem of elemental sulfur’s insolubility in most organic solvents and the toxicity of carbon disulfide. The TETD and BDTD methods also yield higher purity phosphorothioates. “2’0-Me oligonucleotides” molecules carry a methyl group at the 2’-OH residue of the ribose molecule. 2’-O-Me-RNAs show the same (or similar) behavior as DNA, but are protected against nuclease degradation. 2’-O-Me-RNAs can also be combined with phosphorothioate oligonucleotides (PTOs) for further stabilization. 2’0-Me oligonucleotides (phosphodiester or phosphorothioate) can be synthesized according to routine techniques in the art (see, e.g., Yoo et al., Nucleic Acids Res. 32:2008-16, 2004).
[0064] The interfering nucleic acids described herein may be contacted with a cell or administered to an organism (e.g, a human). Alternatively, constructs and / or vectors encoding the interfering RNA molecules may be contacted with or introduced into a cell or organism. In certain embodiments, a viral, retroviral or lentiviral vector is used. In some embodiments, the vector has a tropism for cardiac tissue. In some embodiments the vector is an adeno-associated virus.
[0065] Typically at least 17, 18, 19, 20, 21, 22 or 23 nucleotides of the complement of the target mRNA sequence are sufficient to mediate inhibition of a target transcript. Perfect complementarity is not necessary. In some embodiments, the interfering nucleic acids contains a 1, 2 or 3 nucleotide mismatch with the target sequence. The interfering nucleic acid molecule may have a 2 nucleotide 3’ overhang. If the interfering nucleic acid molecule is expressed in a cell from a construct, for example from a hairpin molecule or from an inverted repeat of the desired sequence, then the endogenous cellular machinery will create the overhangs. shRNA molecules can contain hairpins derived from microRNA molecules. For example, an RNAi vector can be constructed by cloning the interfering RNA sequence into a pCAG-miR30 construct containing the hairpin from the miR30 miRNA. RNA interference molecules may include DNA residues, as well as RNA residues.
[0066] The interfering nucleic acid molecule is a siRNA molecule. Such siRNA molecules should include a region of sufficient homology to the target region, and be of sufficient length in terms of nucleotides, such that the siRNA molecule down-regulate target RNA. The term “ribonucleotide” or “nucleotide” can, in the case of a modified RNA or nucleotide surrogate, also refer to a modified nucleotide, or surrogate replacement moiety at one or more positions. It is not necessary that there be perfect complementarity between the siRNA molecule and the target, but the correspondence must be sufficient to enable the siRNA molecule to direct sequence-specific silencing, such as by RNAi cleavage of the target RNA. In some embodiments, the sense strand need only be sufficiently complementary with the antisense strand to maintain the overall double-strand character of the molecule. In addition, an siRNA molecule may be modified or include nucleoside surrogates. Single stranded regions of an siRNA molecule may be modified or include nucleoside surrogates, e.g, the unpaired region or regions of a hairpin structure, e.g, a region which links two complementary regions, can have modifications or nucleoside surrogates. Modification to stabilize one or more 3'- or 5'-terminus of an siRNA molecule, e.g, against exonucleases, or to favor the antisense siRNA agent to enter into RISC are also useful. Modifications can include C3 (or C6, C7, C12) amino linkers, thiol linkers, carboxyl linkers, non-nucleotidic spacers (C3, C6, C9, C12, abasic, triethylene glycol, hexaethylene glycol), special biotin or fluorescein reagents that come as phosphoramidites and that have another DMT-protected hydroxyl group, allowing multiple couplings during RNA synthesis.
[0067] Each strand of an siRNA molecule can be equal to or less than 35, 30, 25, 24, 23, 22, 21, or 20 nucleotides in length. In some embodiments, the strand is at least 19 nucleotides in length. For example, each strand can be between 21 and 25 nucleotides in length. In some embodiments, siRNA agents have a duplex region of 17, 18, 19, 29, 21, 22, 23, 24, or 25 nucleotide pairs, and one or more overhangs, such as one or two 3' overhangs, of 2-3 nucleotides.
[0068] A “small hairpin RNA" or “short hairpin RNA" or " hRNA" includes a short RNA sequence that makes a tight hairpin turn that can be used to silence gene expression via RNA interference. The shRNAs provided herein may be chemically synthesized or transcribed from a transcriptional cassette in a DNA plasmid. The shRNA hairpin structure is cleaved by the cellular machinery into siRNA, which is then bound to the RNA-induced silencing complex (RISC). In some embodiments, shRNAs are about 15-60, 15-50, or 15-40 (duplex) nucleotides in length, about 15-30, 15-25, or 19-25 (duplex) nucleotides in length, or are about 20-24, 21-22, or 21-23 (duplex) nucleotides in length (e.g, each complementary sequence of the double-stranded shRNA is 15-60, 15-50, 15-40, 15-30, 15-25, or 19-25 nucleotides in length, or about 20-24, 21-22, or 21-23 nucleotides in length, and the doublestranded shRNA is about 15-60, 15-50, 15-40, 15-30, 15-25, or 19-25 base pairs in length, or about 18-22, 19-20, or 19-21 base pairs in length). shRNA duplexes may comprise 3’ overhangs of about 1 to about 4 nucleotides or about 2 to about 3 nucleotides on the antisense strand and / or 5 ’-phosphate termini on the sense strand. In some embodiments, the shRNA comprises a sense strand and / or antisense strand sequence of from about 15 to about 60 nucleotides in length (e.g, about 15-60, 15-55, 15-50, 15-45, 15-40, 15-35, 15-30, or 15-25 nucleotides in length), or from about 19 to about 40 nucleotides in length (e.g, about 19-40, 19-35, 19-30, or 19-25 nucleotides in length), or from about 19 to about 23 nucleotides in length (e.g., 19, 20, 21, 22, or 23 nucleotides in length).
[0069] Non-limiting examples of shRNA include a double-stranded polynucleotide molecule assembled from a single-stranded molecule, where the sense and antisense regions are linked by a nucleic acid-based or non-nucleic acid-based linker; and a double-stranded polynucleotide molecule with a hairpin secondary structure having self-complementary sense and antisense regions. In some embodiments, the sense and antisense strands of the shRNA are linked by a loop structure comprising from about 1 to about 25 nucleotides, from about 2 to about 20 nucleotides, from about 4 to about 15 nucleotides, from about 5 to about 12 nucleotides, or 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, or more nucleotides.
[0070] Additional embodiments related to the shRNAs, as well as methods of designing and synthesizing such shRNAs, are described in U.S. patent application publication number 2011 / 0071208, the disclosure of which is herein incorporated by reference in its entirety for all purposes.
[0071] Interfering nucleic acid molecules can be prepared, for example, by chemical synthesis, in vitro transcription, or digestion of long dsRNA by Rnase III or Dicer. These can be introduced into cells by transfection, electroporation, or other methods known in the art. See Hannon, GJ, 2002, RNA Interference, Nature 418: 244-251; Bernstein E et al., 2002, The rest is silence. RNA 7: 1509-1521; Hutvagner G et al., RNAi: Nature abhors a double-strand. Curr. Opin. Genetics & Development 12: 225-232; Brummelkamp, 2002, A system for stable expression of short interfering RNAs in mammalian cells. Science 296: 550-553; Lee NS, Dohjima T, Bauer G, Li H, Li M-J, Ehsani A, Salvaterra P, and Rossi J. (2002). Expression of small interfering RNAs targeted against HIV-1 rev transcripts in human cells. Nature Biotechnol. 20:500-505; Miyagishi M, and Taira K. (2002). U6-promoter-driven siRNAs with four uridine 3' overhangs efficiently suppress targeted gene expression in mammalian cells. Nature Biotechnol. 20:497-500; Paddison PJ, Caudy AA, Bernstein E, Hannon GJ, and Conklin DS. (2002). Short hairpin RNAs (shRNAs) induce sequence-specific silencing in mammalian cells. Genes & Dev. 16:948-958; Paul CP, Good PD, Winer I, and Engelke DR. (2002). Effective expression of small interfering RNA in human cells. Nature Biotechnol. 20:505-508; Sui G, Soohoo C, Affar E-B, Gay F, Shi Y, Forrester WC, and Shi Y. (2002). A DNA vector-based RNAi technology to suppress gene expression in mammalian cells. Proc. Natl. Acad. Sci. USA 99(6):5515-5520; Yu J-Y, DeRuiter SL, and Turner DL. (2002). RNA interference by expression of short-interfering RNAs and hairpin RNAs in mammalian cells. Proc. Natl. Acad. Sci. USA 99(9): 6047-6052.
[0072] An interfering nucleic acid molecule or an interfering nucleic acid encoding polynucleotide can be administered to the subject, for example, as naked nucleic acid, in combination with a delivery reagent, and / or as a nucleic acid comprising sequences that express an interfering nucleic acid molecule. In some embodiments the nucleic acid comprising sequences that express the interfering nucleic acid molecules are delivered within vectors, e.g. plasmid, viral and bacterial vectors. Any nucleic acid delivery method known in the art can be used in the methods described herein. Suitable delivery reagents include, but are not limited to, e.g., the Mirus Transit TKO lipophilic reagent; lipofectin; lipofectamine; cellfectin; polycations (e.g, polylysine), atelocollagen, nanoplexes and liposomes. The use of atelocollagen as a delivery vehicle for nucleic acid molecules is described in Minakuchi et al. Nucleic Acids Res., 32(13):el09 (2004); Hanai et al. Ann NY Acad Sci., 1082:9-17 (2006); and Kawata et al. Mol Cancer Then, 7(9):2904-12 (2008); each of which is incorporated herein in their entirety. Exemplary interfering nucleic acid delivery systems are provided in U.S. Patent Nos. 8,283,461, 8,313,772, 8,501,930. 8,426,554, 8,268,798 and 8,324,366, each of which is hereby incorporated by reference in its entirety.
[0073] The term “prophylactic” or "therapeutic" treatment is art-recognized and includes administration to the host of one or more of the subject compositions. If it is administered prior to clinical manifestation of the unwanted condition (e.g., disease or other unwanted state of the host animal) then the treatment is prophylactic (i.e., it protects the host against developing the unwanted condition), whereas if it is administered after manifestation of the unwanted condition, the treatment is therapeutic (i.e., it is intended to diminish, ameliorate, or stabilize the existing unwanted condition or side effects thereof).
[0074] The term “small molecule" is a term of the art and includes molecules that are less than about 1000 molecular weight or less than about 500 molecular weight. In one embodiment, small molecules do not exclusively comprise peptide bonds. In another embodiment, small molecules are not oligomeric. Exemplary small molecule compounds which can be screened for activity include, but are not limited to, peptides, peptidomimetics, nucleic acids, carbohydrates, small organic molecules (e.g., polyketides) (Cane et al. (1998) Science 282:63), and natural product extract libraries. In another embodiment, the compounds are small, organic non-peptidic compounds. In a further embodiment, a small molecule is not biosynthetic. The term “subject” refers to a mammal, including, but not limited to, a human or nonhuman mammal, such as a bovine, equine, canine, ovine, or feline.
[0075] A “therapeutically effective amount” of a compound with respect to the subject method of treatment refers to an amount of the compound(s) in a preparation which, when administered as part of a desired dosage regimen (to a mammal, preferably a human) alleviates a symptom, ameliorates a condition, or slows the onset of disease conditions according to clinically acceptable standards for the disorder or condition to be treated or the cosmetic purpose, e.g., at a reasonable benefit / risk ratio applicable to any medical treatment.
[0076] As used herein, the term “treating” or “treatment” includes reversing, reducing, or arresting the symptoms, clinical signs, and underlying pathology of a condition in a manner to improve or stabilize a subject's condition.
[0077] Therapeutic Methods
[0078] A ketogenic diet (KD) induces substantial changes in the gut microbiome, and that enriching KD-associated bacteria via probiotic administration, fecal transplant, or selective microbial reconstitution of the native microbiome mimics the beneficial effects of the KD. Provided herein are methods and compositions that can replace the KD diet in the treatment or prevention of a condition as described. The methods and compositions described herein can be used separately or in conjunction with the KD diet in the treatment or prevention of a condition described herein.
[0079] In some embodiments, the subject is on a diet, and the diet may be a control diet, a ketogenic diet, a high fat diet, or a low carbohydrate diet. In some embodiments, the subject is given antibiotics to deplete the subject’s gut microbiota.
[0080] In certain embodiments, the methods treat or prevent seizures in a subject. In some aspects, provided herein are methods for preventing or treating a condition responsive to a ketogenic diet in a subject, comprising administering to the subject a composition comprising a an agent listed in Table 1. The composition may comprise at least, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9 or at least 10 of the compounds listed in Table 1.
[0081] In some embodiments, the method comprises administering to the subject an agent that modulates a gene product of any one of the genes listed in Table 1.
[0082] In some embodiments, the agent decreases the activity or expression of a gene product or gene. In some embodiments, the agent is an inhibitor of the gene or gene product. In some embodiments, the agent is selected from one or more of adenosine triphosphate, Afatinib, Axitinib, alpha-Tocopherol, alpha-Tocopherol acetate, alpha-Tocopherol succinate, Amitriptyline, Anisomycin, Benzoyl peroxide, Brigatinib, Caffeine, Cannabidiol, Cholic Acid, Cladribine, Clomipramine, Curcumin, Dacomitinib, Deoxycholic acid, Dequalinium, Diazoxide, DL-alpha-Tocopherol, Drotrecogin alfa, Erlotinib, Etacrynic acid, Ezatiostat hydrochloride, Fostamatinib, Garcinol, Gefitinib, glyburide, Guvacine, Hexachlorophene, Lapatinib, Lidocaine, Midostaurin, Metformin, Omacetaxine mepesuccinate, Osimertinib, Puromycin, Racephedrine, Rhein, Roxadustat, Sgc-cbp30, Sunitinib, Tamoxifen, Trimetazidine, Vandetanib, and Zanubrutinib.
[0083] In some embodiments, the agent increases the activity or expression of a gene product or gene. In some embodiments, the agent is a potentiator of the gene or gene product. In some embodiments, the agent is selected from one or more of 1,1, 1,3,3, 3-hexafluoro-2-{4-[(2, 2,2- trifluoroethyl)amino]phenyl}propan-2-ol; 3-Hydroxymethyl-5-aziridinyl-lmethyl-2-[lH- indole-4, 7-dione] -propanol; 3-(hydroxymethyl)-l-methyl-5-(2-methylaziridin-l-yl)-2-phenyl- lH-indole-4, 7-dione; 5'-Guanosine-diphosphate-monothiophosphate; 5 -Methoxy- 1,2- dimethyl-3-(4-nitrophenoxymethyl)indole-4, 7-dione; Adenine; Benzylpenicilloyl Poly lysine; Benzenesulfonyl; Bishy droxy[2H-l-benzopyran-2-one,l,2-benzopyrone]; Canfosfamide; Cibacron Blue; Colforsin; Dicoumarol; Glutathione; Glutathione disulfide; Guanosine-5'- diphosphateguanosine-5'-diphosphate; Hyodeoxy cholic acid; Lauryl dimethylamine-N-oxide; L-histidine; Melatonin; Menadione; Myristic acid; N-(2,2,2-trifluoroethyl)-N-{4-[2,2,2- trifluoro-l-hydroxy-l-(trifluoromethyl)ethyl]phenyl}benzenesulfonamide; N- formylmethionine; NADH; NAV-2729; p-Nitrophenol; Pseudoephedrine; Rosiglitazon; Succinic acid; S-(4-bromobenzyl)cysteine; S-hexylglutathione; Ubiquinone-1; and Vitamin C.
[0084] In some embodiments, the method comprises administering to the subject an agent that modulates a gene product of any one of the pathways listed in Table 2.
[0085] In some embodiments, the agent increases the activity or expression of a gene product or pathway. In some embodiments, the agent is a potentiator of the pathway or gene product.
[0086] In some embodiments, the pathway is selected from one or more of Formaldehyde assimilation; Guanosine nucleotide degradation; Amino acid biosynthesis; L-proline biosynthesis; proline biosynthesis; Sugar acid biosynthesis; GDP-mannose biosynthesis; monosaccharide biosynthesis; 2-methylcitrate cycle; methylcitrate cycle I; methylcitrate cycle II; Glycol degradation; glycol metabolism; Fatty acid beta-oxidation; fatty acid oxidation; fatty acid biosynthesis; hydroxy fatty acid biosynthesis; Methionine metabolism; Beta- hydroxybutyrate (BHBA) metabolism; N-substituted nicotinamide metabolism; Alphalinolenic acid and linoleic acid metabolism; Phenylpropanoic acid biosynthesis; Glycine and serine metabolism; Glutathione metabolism; Betaine metabolism; Adenine and adenosine salvage; dTDP-L-rhamnose biosynthesis; Folate transformations; L-valine biosynthesis; sulfoglycolysis; Acetyl-CoA fermentation; Pyruvate fermentation; Branched-chain amino acid biosynthesis; Valine, leucine, and isoleucine biosynthesis; Phenylalanine, tyrosine, and tryptophan biosynthesis; and Glycerophospholipid metabolism.
[0087] In some embodiments, the agent decreases the activity or expression of a gene product or pathway. In some embodiments, the agent is an inhibitor of the pathway or gene product.
[0088] In some embodiments, the pathway is selected from one or more of Aerobactin biosynthesis; Sidophore biosynthesis; Polyamine biosynthesis; Monosaccharide metabolism; Biotin biosynthesis; Fatty acid metabolism; Hydroxy fatty acid metabolism; and Phenylpropanoic metabolism.
[0089] In some embodiments, the subject has epilepsy (e.g., pediatric epilepsy; refractory or non-refractory epilepsy). The disclosed epilepsy disorder may be benign Rolandic epilepsy, frontal lobe epilepsy, infantile spasms, juvenile myoclonic epilepsy (JME), juvenile absence epilepsy, childhood absence epilepsy (e.g. pyknolepsy), febrile seizures, progressive myoclonus epilepsy of Lafora, Lennox-Gastaut syndrome, Landau-Kleffher syndrome, Dravet syndrome, Generalized Epilepsy with Febrile Seizures (GEFS+), Severe Myoclonic Epilepsy of Infancy (SMEI), Benign Neonatal Familial Convulsions (BFNC), West Syndrome, Ohtahara Syndrome, early myoclonic encephalopathies, migrating partial epilepsy, infantile epileptic encephalopathies, Tuberous Sclerosis Complex (TSC), focal cortical dysplasia, Type I Lissencephaly, Miller-Dieker Syndrome, Angelman's syndrome, Fragile X syndrome, epilepsy in autism spectrum disorders, subcortical band heterotopia, Walker- Warburg syndrome, Alzheimer's disease, posttraumatic epilepsy, progressive myoclonus epilepsies, reflex epilepsy, Rasmussen's syndrome, temporal lobe epilepsy, limbic epilepsy, status epilepticus, abdominal epilepsy, massive bilateral myoclonus, catamenial epilepsy, Jacksonian seizure disorder, Unverricht-Lundborg disease, or photosensitive epilepsy. The epilepsy may include generalized seizures or partial (i.e. focal) seizures.
[0090] Epilepsy disorder, as disclosed herein, may be Dravet Syndrome, Lennox-Gastaut Syndrome, infantile spasm, or Ohtahara Syndrome. The epilepsy disorder may be Dravet Syndrome, Lennox-Gastaut Syndrome, infantile spasm, or Ohtahara Syndrome, or a pediatric epilepsy disorder. The pediatric epilepsy disorder may be benign childhood epilepsy, Benign Neonatal Familial Convulsions (BFNC), febrile seizures, Dravet Syndrome, Lennox-Gastaut Syndrome, infantile spasm, Ohtahara Syndromejuvenile myoclonic epilepsy uvenile absence epilepsy, childhood absence epilepsy (e.g. pyknolepsy), infantile spasms. In embodiments the epilepsy disorder is Dravet Syndrome. According to the invention, the epilepsy disorder is Dravet syndrome or Lennox-Gastaut syndrome.
[0091] The disclosed pediatric epilepsy disorder may be benign childhood epilepsy. The disclosed pediatric epilepsy disorder may be Benign Neonatal Familial Convulsions (BFNC). The disclosed pediatric epilepsy disorder may be febrile seizures. In embodiments according to the invention the pediatric epilepsy disorder is Dravet Syndrome. In embodiments according to the invention the pediatric epilepsy disorder is Lennox-Gastaut Syndrome. The disclosed pediatric epilepsy disorder may be infantile spasm. The disclosed pediatric epilepsy disorder may be Ohtahara Syndrome. The disclosed pediatric epilepsy disorder may be juvenile myoclonic epilepsy. The disclosed pediatric epilepsy disorder may be juvenile absence epilepsy. The disclosed pediatric epilepsy disorder may be childhood absence epilepsy (e.g. pyknolepsy). The disclosed pediatric epilepsy disorder may be infantile spasms.
[0092] In some embodiments, the epilepsy disorder may be a result of a neurological disease or injury such as, for example, encephalitis, cerebritis, abscess, stroke, tumor, trauma, genetic, tuberous sclerosis, cerebral dysgenesis, or hypoxic-ischemic encephalophathy. The epilepsy disorder may be associated with a neurodegenerative disease such as, for example, Alzheimer's disease or Parkinson's Disease. The epilepsy disorder may be associated with autism. The epilepsy disorder may be associated with a single gene mutation. The epilepsy disease may be associated with compulsive behaviors or electrographic seizures.
[0093] The epilepsy disorder may be an epilepsy disorder which is non-responsive to treatment with an antiepileptic drug (AED). The AED may be acetazolamide. The AED may be benzodiazepine. The AED may be cannabadiols. The AED may be carbamazepine. The AED may be clobazam. The AED may be clonazepam. The AED may be eslicarbazepine acetate. The AED may be ethosuximide. The AED may be ethotoin. The AED may be felbamate. The AED may be fenfluramine. The AED may be fosphenytoin. The AED may be gabapentin. The AED may be ganaxolone. The AED may be huperzine A. The AED may be lacosamide. The AED may be lamotrigine. The AED may be levetiracetam. The AED may be nitrazepam. The AED may be oxcarbazepine. The AED may be perampanel. The AED may be piracetam. The AED may be phenobarbital. The AED may be phenytoin. The AED may be potassium bromide. The AED may be pregabalin. The AED may be primidone. The AED may be retigabine. The AED may be rufmamide. The AED may be valproic acid. The AED may be sodium valproate. The AED may be stiripentol. The AED may be tiagabine. The AED may be topiramate. The AED may be vigabatrin. The AED may be zonisamide.
[0094] In some embodiments, the subject has a neurodevelopmental disorder. Representative neurodevelopmental disorders include autism spectrum disorder, Rett syndrome, fragile X, attention deficit disorder, and attention-deficit / hyperactivity disorder. In some embodiments, the neurodevelopmental disorder is a disorder known to be comorbid with seizures. As used herein, a condition “responsive to a ketogenic diet” includes, but is not limited to, epilepsy, autism spectrum disorder, Rett syndrome, fragile X, attention deficit disorder, attention- deficit / hyperactivity disorder, seizures, autism spectrum disorder, Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), cancer, stroke, a metabolic disease (e.g., diabetes or obesity), a mitochondrial disorder, depression, migraines (e.g., chronic migraines), or traumatic brain injury (TBI).
[0095] In some embodiments, the subject is refractory to anti-conversant drug or anti-seizure drug. An "anti-seizure drug", "anti-epilepsy drug", “anti-epileptic drug”, “anti-seizure medication”, "AED", or "anticonvulsant" are used interchangeably herein and according to their common and ordinary meaning and include compositions for reducing or eliminating seizures. Anticonvulsants include, but are not limited to acetazolamide, benzodiazepine, cannabidiols, carbamazepine, clobazam, clonazepam, eslicarbazepine acetate, ethosuximide, ethotoin, felbamate, fenfluramine, fosphenytoin, gabapentin, ganaxolone, huperzine A, lacosamide, lamotrigine, levetiracetam, nitrazepam, oxcarbazepine, perampanel, piracetam, phenobarbital, phenytoin, potassium bromide, pregabalin, primidone, retigabine, rufmamide, valproic acid, sodium valproate, stiripentol, tiagabine, topiramate, vigabatrin, or zonisamide.
[0096] In some embodiments, the agents and / or compositions described herein may be administered conjointly. In some embodiments, a composition comprising a compound of Table 1 may be conjointly administered with an anticonvulsant.
[0097] Table 1: Genes and modulating agents
[0098]
[0099] Table 2: Pathways associated with changes to Ketogenic Diet or KD-microbiomes
[0100]
[0101] In other embodiments, the subject has a condition responsive to a ketogenic diet. The condition may be Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), cancer, stroke, a metabolic disease, a mitochondrial disorder, depression, migraines (e.g., chronic migraines), or traumatic brain injury (TBI). In some embodiments, the methods and compositions comprise administering to the subject a composition provided herein. In some embodiments, the condition is epilepsy, seizures, autism spectrum disorder, Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), cancer, stroke, a metabolic disease (e.g., obesity or diabetes), a mitochondrial disorder, depression, migraines (e.g., chronic migraines), Rett syndrome, attention deficit disorder, fragile X syndrome, or traumatic brain injury (TBI). In some embodiments, the compositions and methods provided herein are useful in treating or preventing aging or aging-associating conditions. In some embodiments, the compositions and methods provided herein can replace the ketogenic diet in the treatment or prevention of a condition described herein; in other embodiments, the compositions and methods provided herein can be combined with the ketogenic diet. More information on conditions may be found in Stafstrom et al. (2012) Front. Pharmacol. 3:59, hereby incorporated in its entirety.
[0102] The composition may be formulated for oral delivery. In some embodiments, the composition may comprise probiotics. In some embodiments, the compositions disclosed herein are food products. The composition may be in the form of a pill, tablet, or capsule. In some embodiments, the subject may be a mammal (e.g., a human). In some embodiments, the composition is self-administered.
[0103] In some embodiments, the above methods include reducing the amount of pathogenic bacteria in a subject (i.e., in the gastrointestinal tract of the subject) prior to administration of an agent disclosed herein. In some embodiments, this includes any such therapy that achieves the same goal of reducing the number of pathogenic organisms, when used in combination with the compositions described herein, would lead to replacement of the pathogenic microflora involved in the diseased state with microflora associated with a nondiseased state, or less pathogenic species occupying the same ecological niche as the type causing a disease state. For example, a subject may undergo treatment with antibiotics (e.g., antimicrobial compounds) or a composition comprising antibiotics to target and decrease the prevalence of pathogenic organisms, and subsequently be treated with a composition described herein. The treatment may also comprise an antifungal or anti-viral compound.
[0104] Suitable antimicrobial compounds include capreomycins, including capreomycin IA, capreomycin IB, capreomycin IIA and capreomycin IIB; carbomycins, including carbomycin A; carumonam; cefaclor, cefadroxil, cefamandole, cefatrizine, cefazedone, cefazolin, cefbuperazone, cefcapene pivoxil, cefclidin, cefdinir, cefditoren, cefime, ceftamet, cefmenoxime, cefmetzole, cefminox, cefodizime, cefonicid, cefoperazone, ceforanide, cefotaxime, cefotetan, cefotiam, cefoxitin, cefpimizole, cefpiramide, cefpirome, cefprozil, cefroxadine, cefsulodin, ceftazidime, cefteram, ceftezole, ceftibuten, ceftiofur, ceftizoxime, ceftriaxone, cefuroxime, cefuzonam, cephalexin, cephalogycin, cephaloridine, cephalosporin C, cephalothin, cephapirin, cephamycins, such as cephamycin C, cephradine, chi ortetracy cline; chlarithromycin, clindamycin, clometocillin, clomocycline, cloxacillin, cyclacillin, danofloxacin, demeclocyclin, destomycin A, dicloxacillin, dirithromycin, doxycyclin, epicillin, erythromycin A, ethanbutol, fenbenicillin, flomoxef, florfenicol, floxacillin, flumequine, fortimicin A, fortimicin B, forfomycin, foraltadone, fusidic acid, gentamycin, glyconiazide, guamecycline, hetacillin, idarubicin, imipenem, isepamicin, josamycin, kanamycin, leumycins such as leumycin Al, lincomycin, lomefloxacin, loracarbef, lymecycline, meropenam, metampicillin, methacycline, methicillin, mezlocillin, micronomicin, midecamycins such as midecamycin Al, mikamycin, minocycline, mitomycins such as mitomycin C, moxalactam, mupirocin, nafcillin, netilicin, norcardians such as norcardian A, oleandomycin, oxytetracycline, panipenam, pazufloxacin, penamecillin, penicillins such as penicillin G, penicillin N and penicillin O, penillic acid, pentylpenicillin, peplomycin, phenethicillin, pipacyclin, piperacilin, pirlimycin, pivampicillin, pivcefalexin, porfiromycin, propiallin, quinacillin, ribostamycin, rifabutin, rifamide, rifampin, rifamycin SV, rifapentine, rifaximin, ritipenem, rekitamycin, rolitetracycline, rosaramicin, roxithromycin, sancycline, sisomicin, sparfloxacin, spectinomycin, streptozocin, sulbenicillin, sultamicillin, talampicillin, teicoplanin, temocillin, tetracyclin, thostrepton, tiamulin, ticarcillin, tigemonam, tilmicosin, tobramycin, tropospectromycin, trovafloxacin, tylosin, and vancomycin, and analogs, derivatives, pharmaceutically acceptable salts, esters, prodrugs, and protected forms thereof.
[0105] Suitable anti-fungal compounds include ketoconazole, miconazole, fluconazole, clotrimazole, undecylenic acid, sertaconazole, terbinafine, butenafme, clioquinol, haloprogin, nystatin, naftifine, tolnaftate, ciclopirox, amphotericin B, or tea tree oil and analogs, derivatives, pharmaceutically acceptable salts, esters, prodrugs, and protected forms thereof.
[0106] Compositions
[0107] Provided herein are compositions (e.g., compositions comprising an agent disclosed herien and a pharmaceutically acceptable carrier). The composition may comprise at least one, at least two, at least three, at least four, at least five, at least six at least seven, at least eight, at least nine, or at least ten of the compounds listed in Table 1.
[0108] The composition may comprise a pharmaceutically acceptable carrier. The composition may comprise probiotics. The pharmaceutical compositions disclosed herein may be delivered by any suitable route of administration, including orally, bucally, sublingually, parenterally, and rectally, as by powders, ointments, drops, liquids, gels, tablets, capsules, pills, or creams. In certain embodiments, the pharmaceutical compositions are delivered generally (e.g., via oral administration). In certain other embodiments, the compositions disclosed herein are delivered rectally.
[0109] The pharmaceutically acceptable carriers can be either solid or liquid. Solid form preparations include powders, tablets, pills, capsules, cachets, suppositories, and dispersible granules. A solid carrier may be one or more substance that may also act as diluents, flavoring agents, binders, preservatives, tablet disintegrating agents, or an encapsulating material.
[0110] In powders, the carrier may be a finely divided solid in a mixture with the finely divided active component. In tablets, the active component may be mixed with the carrier having the necessary binding properties in suitable proportions and compacted in the shape and size desired.
[0111] Actual dosage levels of the active ingredients in the pharmaceutical compositions may be varied so as to obtain an amount of the active ingredient which is effective to achieve the desired therapeutic response for a particular patient, composition, and mode of administration, without being toxic to the patient.
[0112] The selected dosage level will depend upon a variety of factors including the activity of the particular agent employed, the route of administration, the time of administration, the rate of excretion or metabolism of the particular compound being employed, the duration of the treatment, other drugs, compounds and / or materials used in combination with the particular compound employed, the age, sex, weight, condition, general health and prior medical history of the patient being treated, and like factors well known in the medical arts.
[0113] A physician or veterinarian having ordinary skill in the art can readily determine and prescribe the effective amount of the pharmaceutical composition required. For example, the physician or veterinarian could prescribe and / or administer doses of the compounds employed in the pharmaceutical composition at levels lower than that required in order to achieve the desired therapeutic effect and gradually increase the dosage until the desired effect is achieved.
[0114] EXEMPLIFICATION
[0115] An exemplification of this application by the inventors was published as Gregory Lum et al. “Ketogenic diet therapy for pediatric epilepsy is associated with alterations in the human gut microbiome that confer seizure resistance in mice.” Cell Reports, 42(12): 113521, 26 December 2023, and is hereby incorporated by reference in its entirety.
[0116] The gut microbiome modulates seizure susceptibility and the anti-seizure effects of the ketogenic diet (KD) in animal models, but whether these relationships translate to KD therapies for human drug-resistant epilepsy is unclear. Herein, we find that the clinical KD shifts the function of the gut microbiome in children with refractory epilepsy. Colonizing mice with KD-associated human gut microbes confers increased resistance to 6-Hz psychomotor seizures, as compared to colonization with gut microbes from matched pretreatment controls. Parallel analysis of human donor and mouse recipient metagenomic and metabolomic profiles identifies subsets of shared functional features that are seen in response to KD treatment in humans and preserved upon transfer to mice fed a standard diet. These include enriched representation of microbial genes and metabolites related to anaplerosis, fatty acid beta-oxidation, and amino acid metabolism. Mice colonized with KD-associated human gut microbes further exhibit altered hippocampal and frontal cortical transcriptomic profiles relative to colonized pre-treatment controls, including differential expression of genes related to ATP synthesis, glutathione metabolism, oxidative phosphorylation, and translation. Integrative co-occurrence network analysis of the metagenomic, metabolomic, and brain transcriptomic datasets identifies features that are shared between human and mouse networks, and select microbial functional pathways and metabolites that are candidate primary drivers of hippocampal expression signatures related to epilepsy. Together, these findings reveal key microbial functions and biological pathways that are altered by clinical KD therapies for pediatric refractory epilepsy and further linked to microbiome-induced alterations in brain gene expression and seizure protection in mice.
[0117] The low-carbohydrate, high-fat ketogenic diet (KD), is a mainstay treatment for refractory epilepsy, particularly in children who do not respond to existing anti-seizure medications. The efficacy of the KD is supported by multiple retrospective and prospective studies, which estimate that -30% of pediatric patients become seizure-free and -60% experience substantial benefit with >50% reduction in seizures (Coppola et al., 2002; Freeman et al., 1998; Hoon et al., 2005; Neal et al., 2008). However, use of the KD for treating pharmacoresistant epilepsy remains low due to difficulties with implementation, dietary compliance, and adverse side effects (Kossoff et al., 2018). Even with successful seizure reduction, retention of epileptic children on the KD is a reported 13% by the third year of dietary therapy (Hemingway et al., 2001). The primary reasons cited for discontinuation include “diet restrictiveness” and “diet side effects,” in addition to low diet responsiveness. While many pioneering studies have proposed important roles for immunosuppression, ketone bodies, anaplerosis, and gamma-aminobutyric acid (GABA) modulation in mediating the neuroprotective effects of the KD, they do not fully account for the clinical heterogeneity in patient responsiveness. Exactly how the KD confers protection against epilepsy in individuals with varied seizure semiologies remains unclear, and the biological determinants of patient responsiveness to the KD are poorly understood.
[0118] The gut microbiome plays an integral role in mediating effects of diet on multiple aspects of host physiology, including metabolism, neural activity, and behavior (Singh et al., 2017; Sonnenburg & Backhed, 2016). To date, a few clinical studies have reported associations between the KD regimens and alterations in the composition and / or functional potential of the gut microbiota in epileptic individuals (Lindefeldt et al., 2019; G. Xie et al., 2017; Zhang et al., 2018). While promising, thus far there is little consistency across these different reports in the specific microbial taxa or gene pathways that correlate with the KD. Moreover, functional consequences of the KD-associated human epilepsy microbiome on host seizure susceptibility remain unknown. KD-induced alterations in the gut microbiome were necessary and sufficient for mediating the seizure protective effects of KD chow in two mouse models for refractory epilepsy - the 6-Hz psychomotor seizure model and the Kenai deficiency model for sudden unexpected death in epilepsy (SUDEP) (Olson et al., 2018). Similarly, in a rat injury model of infantile spasms, transfer of the KD-induced gut microbiota into naive animals fed a control diet reduced spasms (Mu et al., 2022). In addition, taxonomic differences in the gut microbiome were correlated with seizure severity and seizure protection in response to KD chow in the Senia deficiency model for Dravet syndrome (Milj anovic & Potschka, 2021). Together, these findings across various seizure models provide proof-of- principle that the KD alters the gut microbiome in ways that can promote seizure protection. However, whether these results from rodent studies apply to human epilepsy, the human gut microbiome, and clinical KD regimens used to treat pediatric epilepsy is still unknown, and the core microbial functions that impact seizure susceptibility are unclear. Herein, a prospective study of KD interventions in children with refractory epilepsy and test causal effects of the human gut microbiome before and after initiating clinical KD regimens on seizure susceptibility in mice is described. Functional changes in the human gut microbiome that are associated with KD treatment in pediatric epilepsy patients were evaluated. Further, select features of the clinical KD-associated gut microbiome that are shared across both human donors and inoculated mouse recipients were identified that correlate with microbiome-dependent seizure protection in mice. Finally, key direct and indirect network interactions between the gut microbiome, metabolites, and brain transcriptome were identified that may contribute to the ability of the clinical KD-associated human gut microbiome from pediatric epilepsy patients to promote seizure protection in mice.
[0119] Clinical KD regimens elicit shared functional features of the gut microbiome in a cohort of children with refractory epilepsy
[0120] The KD is commonly prescribed for pediatric refractory epilepsy, wherein children consume commercial ketogenic infant formula and / or fat-rich, carbohydrate-restricted meals with dietary guidance from clinicians and registered dieticians (Kossoff et al., 2018). Notably, treatment regimens for the KD vary from patient to patient. KD composition depends on patient tolerability, which dictates the ratio of fat intake relative to carbohydrate and protein. Additionally, variable food sources determine the specific macro- and micronutrients that comprise ketogenic meals. Moreover, the KD is prescribed broadly for various forms of refractory epilepsy, the treatment population varies in genetic risk, seizure semiology, and past anti-seizure medication exposures, among other factors. In order to assess effects of clinically-relevant KD treatments for refractory epilepsy on the gut microbiome, a prospective study of 10 children with pediatric refractory epilepsy who were newly enrolled to the Ketogenic Diet Program at UCLA Mattel Children’s Hospital was conducted. From each patient, stool samples were collected within 1 day before initiating a KD regimen (pre-KD sample) and after approximately 1 month of adherence to a clinically- guided KD (post-KD sample). 1 month was chosen as a time point at which stabilized microbial responses to the dietary regimen were expected to be observed (David et al., 2014).
[0121] Data from 16S rRNA gene amplicon sequencing of fecal samples indicated no significant difference in bacterial a-diversity in the post-KD fecal microbiota from pediatric epilepsy patients relative to their matched pre-KD internal controls (Figure 1 A). Principal coordinates analysis of unweighted and weighted Unifrac distances revealed substantial variation across individuals in baseline composition of the pre-KD microbiota (Figure IB). Additionally, the clinical KD elicited differential shifts in bacterial P-diversity and varied responses across post-KD samples relative to their matching pre-KD controls, which were not significantly associated with demographic or clinical measures, such as age, sex, and prior anti-seizure medication exposure (Figure IB, Yassour et al., 2016). Consistent with the interindividual variation in microbial taxonomic profiles, ANCOM and ANOVA analyses (paired or unpaired) identified no significant differences in relative abundances in particular bacterial taxa when considering all post-KD sample relative to their matched pre-KD controls. These results indicate that, within this particular study cohort, there are no shared effects of the clinical KD on the microbial composition of the gut microbiota of children with refractory epilepsy.
[0122] Functional redundancy is common across different microbial species of the human gut microbiota (Tian et al., 2020). In light of the varied bacterial taxonomic profiles at baseline and in response to dietary treatment, it was hypothesized that the clinical KD is associated with shared alterations in the functional potential of the gut microbiota from children with pediatric epilepsy. Shotgun metagenomic profiling and pathway analysis indicated that compared to pre-KD samples, post-KD samples shared a significant decrease in relative abundance of genes belonging to the top 26 most abundant functional pathways, which together comprised >94% of the pathway diversity detected. There were no reported changes in the total observed pathways in post-KD samples compared to their respective pre-KD controls. These observations suggest that the clinical KD restricts the relative membership of various types of microbial taxa that harbor genes related to prevalent functions and / or enriches for existing microbial taxa that harbor genes related to previously rare or underrepresented functions. In particular, post-KD samples exhibited significant enrichment of genes related to formaldehyde assimilation, guanosine nucleotide degradation, and L- proline biosynthesis, and decreased representation of genes related to aerobactin biosynthesis, as compared to pre-KD controls. There were also modest increases in genes related to GDP- mannose biosynthesis, 2-methylcitrate cycle, and glycol metabolism and degradation, and decreases in genes related to polyamine biosynthesis and biotin biosynthesis, subsets of which will be discussed in greater detail in the following sections. Taken together, these data suggest that treatment with KD regimens that differ in KD ratio and specific nutritional composition elicit broad shifts in the functional potential of the gut microbiome that are shared across children with varied subtypes of refractory epilepsy. Transferring the fecal microbiota from KD-treated pediatric epilepsy patients to mice confers seizure resistance
[0123] Causal influences of the human microbiome can be effectively studied in gnotobiotic mice, wherein transferring microbes in a clinical sample into microbiota-deficient mice is used to recapitulate the taxonomic and functional diversity of the donor human microbiota. To evaluate whether gut microbes associated with the clinical KD impact seizure susceptibility, individual cohorts of germ-free (GF) mice were inoculated with matched pre- KD and post-KD stool samples collected from children with refractory epilepsy and maintained the colonized mice on standard (non-ketogenic) mouse chow (control diet, CD). Each human donor sample (pre-KD and post-KD from 10 individuals, as biological replicates) was inoculated into 13-15 GF mice (as technical replicates) to enable cohort-level testing of susceptibility to 6-Hz psychomotor seizures (Figure 2A). The 6-Hz seizure model involves low-frequency comeal stimulation to induce acute complex focal-onset seizures reminiscent of human limbic epilepsy (Barton et al., 2001). Consistent with refractory epilepsy, the 6-Hz model is resistant to several anti-seizure medications, but treatment with KD chow effectively protects against 6-Hz seizures in rodents (Hartman et al., 2008), raising the intensity of current required to elicit a seizure in 50% of the subjects tested (CC50, seizure threshold). The 4-day time point was chosen as the maximum duration of time that a KD-induced microbiota could be maintained in mice fed CD (Olson et al., 2018).
[0124] Despite the variation in bacterial diversity across patient gut microbiota (Figure 1), it was observed that GF mice colonized with microbes from the post-KD microbiota required greater intensity of current to induce 6-Hz seizures (Figure 2B) as compared to controls colonized with microbes from the pre-KD microbiota. This effect was seen when comparing post-KD vs. pre-KD microbiota transfer for individual technical replicates per patient (Figure 2B), as well as when data were averaged across all patients (Figures 2C and 2D). Similar increases in seizure threshold were seen by transferring a randomly selected post-KD donor sample into mice pre-treated with broad-spectrum antibiotics (Abx) to deplete the microbiome (Figures 3A-3C), suggesting that the increases in seizure resistance are not dependent upon the GF background of the recipient mice. In addition, compared to pre-KD controls, mice colonized with microbes from the post-KD microbiota required increased intensity of current to elicit one or more recurred seizures observed after the initial stimulus- induced seizure (Figure 2E), indicating that transfer of the post-KD human microbiota promotes resistance to both primary induced seizures and remission seizures in mice. On average, the post-KD samples raised seizure thresholds by 22.4% ± 6.4% relative to matched pre-KD controls (Figures 2C and 2D). This aligns with both our previously published data on the average effect size of KD chow on wildtype mice tested in the 6-Hz seizure assay (24.5%, Olson et al., 2018), and the observed 24.0% increase in seizure threshold seen in GF mice colonized with a conventional adult mouse microbiota (GF-conv) and fed a 6: 1 KD chow, as compared to conventionalized controls fed a standard vitamin- and mineral-matched control diet (Figure 2B). Discrepancies in effect size across patient samples were largely driven by differences between responses for pre-KD controls (Figure 2B), suggesting that the comparatively low microbial diversity resulting from cross-host species transfer increases seizure susceptibility. Consistent with this, it was previously observed that decreasing microbial diversity via antibiotic treatment reduced seizure threshold in the 6-Hz assay (Olson et al., 2018). Overall, these results indicate that inoculating mice with the clinical KD- associated human gut microbiota increases 6-Hz seizure threshold to levels similar to the effect sizes seen with direct consumption of the experimental 6: 1 KD.
[0125] Human microbiota transplantation to mice involves oral inoculation with a human stool suspension, which is comprised of microbial biomass, as well as undigested food matter and secreted molecules from the host and microbiota. As such, effects seen in response to the transfer procedure could be due to the KD-associated gut microbiota or microbiotaindependent dietary or host factors. To gain insight into whether bacteria from the gut microbiota are required for mediating the increases in seizure protection seen with inoculation of the human post-KD microbiota into mice, GF mice inoculated with a randomly selected post-KD donor sample were post-treated with Abx to deplete the microbiota, or with vehicle (Veh) as negative control (Figures 3D and 3E). Mice that were inoculated with the post-KD sample and treated with Veh displayed seizure thresholds that were comparable to that seen previously in recipient mice without the added Veh treatment (Figures 3F, 3G, and IB). This suggests that the post-KD microbiota induced increases in seizure resistance that were maintained beyond the 4-day time point used in our initial transfer experiments (Figures 2B and 2C), extending out to 12 days post inoculation in Veh-treated mice fed CD (Figures 3F and 3G). In contrast, depletion of gut bacteria in mice that were colonized with the post- KD microbiota decreased seizure thresholds to levels that were lower than previously seen in pre-KD colonized controls (Figures 3F, 3G, and 2B). These results indicate that bacterial members of the post-KD microbiota are necessary for mediating the increases in seizure threshold seen in response to transfer of the clinical-KD associated microbiota from a pediatric epilepsy patient into mice.
[0126] Administration of microbial metabolites or other microbiome-dependent molecules, in lieu of viable microbiota, has been reported to ameliorate symptoms of recurrent Clostridiodes difficile infection, inflammatory bowel disease, and multiple sclerosis, among other conditions (Cekanaviciute et al., 2017; Levy et al., 2015; Ott et al., 2017). To gain insight into whether administration of clinical KD-associated intestinal small molecules is sufficient to confer seizure protection in mice, a post-KD donor sample selected at random was sterile filtered and then administered to a cohort of GF recipient mice (Figure 4A), alongside controls that were administered the unfiltered post-KD suspension, as was done previously for human microbiota inoculation (Figures 4A and 4B). At 4 days post inoculation, mice that were treated with the post-KD filtrate exhibited lower seizure threshold compared to controls that were treated with the corresponding unfiltered post-KD suspension (Figures 4B and 4C). These data indicate that clinical KD-associated small molecules in the post-KD fecal sample from a pediatric epilepsy patient are not sufficient to confer persistent seizure protection in mice.
[0127] Orally administered microbial metabolites can be rapidly absorbed and cleared from systemic circulation within a few hours of administration (Abrams & Bishop, 1967; Williams et al., 2020). To further assess whether clinical KD-associated intestinal small molecules, including microbial metabolites, acutely modulate seizure susceptibility, mice were orally gavaged with a sterile-filtered post-KD sample and assessed 2 hours later for 6-Hz seizure threshold, rather than 4 days later as in the previous experiments (Figure 4D). Mice treated with post-KD filtrate exhibited significantly increased seizure protection compared to controls treated with pre-KD filtrate (Figures 4E and 4F), with seizure thresholds that approached those seen after inoculation of the post-KD suspension (Figures 4E, 4G, and 2B). These data indicate that administration of clinical KD-associated intestinal small molecules can acutely confer seizure protection in mice over short timescales (i.e., 2 hours, Figures 4D, 4E, 4F, and 4G), which diminishes by 4 days post treatment (Figures 4A, 4B, and 4C). Taken together, the results presented in these series of experiments suggest that the clinical KD for pediatric refractory epilepsy is associated with alterations in metabolic activities of the gut microbiota that promote seizure resistance in mice.
[0128] While the “humanization” of mice with microbiota from clinical stool samples is a powerful tool for translational microbiome research (Tumbaugh et al., 2009), the approach has technical and biological limitations that warrant careful consideration (Walter et al., 2020). Namely, while much of the taxonomic and functional diversity of the donor inoculum can be recapitulated in recipient mice (Bokoliya et al., 2021), developmental influences and host-specific selection (Rawls et al., 2006), among other factors, preclude full “engraftment” of the human gut microbiota in GF mice (Walter et al., 2020). To evaluate the fidelity of fecal microbiota “transplantation” from pediatric epilepsy patients to GF mice, both the donor pre- KD and post-KD stool samples and corresponding recipient mouse fecal pellets collected at 4 days post-inoculation (the day of seizure testing) were subjected to 16S rRNA gene amplicon sequencing (Figures 2A, 5A-5C). However, a significant reduction in a-diversity was observed, with an average decrease of 38% for all mouse recipient microbiota relative to all human donor microbiota , indicating incomplete transfer or engraftment of the human microbiota in mice. These results align with several previous reports of reduced bacterial a- diversity in mice inoculated with human microbiota, with estimated decreases of 35%, 38%, and 50% (Blanton et al., 2016; Sharon et al., 2019; Staley et al., 2017), suggesting that levels of transfer fidelity were achieved that are consistent with those in the field. However, the inability to fully recapitulate the taxonomic diversity of the human gut microbiota from pediatric epilepsy patients in mice draws into question whether the increases in seizure resistance seen in mice inoculated with post-KD microbiota are relevant to the actual clinical condition. Therefore subsequent experiments focused on identifying and evaluating the subset of functional features of the KD-associated human gut microbiome that are recapitulated in recipient mice, and the microbiome-dependent alterations in host physiology that correspond with seizure protection in mice.
[0129] Select functional features of the clinical KD-associated human microbiome are recapitulated in colonized recipient mice and correlate with seizure protection
[0130] Given the widespread use of the clinical KD for treating epilepsy, and an increasing number of other neurodevelopmental and neurodegenerative disorders, elucidating how the activity of the gut microbiome is altered by the clinical KD could reveal important insights into its physiological effects. To identify microbiome associations with the clinical KD and further determine which of the associations, if any, may modify seizure risk, the gut microbiome from pediatric epilepsy patients before and after treatment with the clinical KD, as well as from gnotobiotic mice that were inoculated with the patient samples, was functionally characterized and tested for causal outcomes on seizure susceptibility. Metagenomic sequencing and analysis revealed microbial gene pathways that were differentially abundant in post-KD samples relative pre-KD controls, and shared between both human donor samples and mouse recipient samples . In particular, microbial genes relevant to fatty acid -oxidation, glycol metabolism and degradation, methylcitrate cycle I, methylcitrate cycle II, and proline biosynthesis were similarly elevated in post-KD human samples and post-KD-inoculated mice compared to their respective pre-KD controls (Figure 6A). These findings align with reported influences of the KD on fatty acid oxidation (A. R. Kennedy et al., 2007), of carbohydrate restriction on promoting the glyoxylate cycle (Puckett et al., 2017), and of fatty acid P-oxidation on the initiation of the methylcitrate cycle (Clark & Cronan, 2005). Proline metabolism involves reactions with glutamine, glutamate, ornithine, and arginine, which might relate to reported effects of KD on amino acid metabolism, particularly of glutamine and glutamate (Yudkoff et al., 2007). In addition, both post-KD human donor and mouse recipient samples exhibited reductions in microbial genes relevant to poly amine biosynthesis and aerobactin biosynthesis (6A). The main role of polyamine biosynthesis is generation of putrescine, mainly using the glucogenic amino acid L-arginine which is consumed in reduced amounts while on the KD. Aerobactin, a sidophore, biosynthesis uses the ketogenic amino acid L-lysine, which is also essential for acetyl-CoA synthesis and energy production during ketosis. These data suggest that the consumption of a clinical KD by children with refractory epilepsy enriches for gut microbes that have the functional capacity to metabolize dietary fats and to perform anaplerotic reactions when dietary carbohydrates are restricted. The findings further indicate that these general features of the KD-associated human gut microbiome are phenocopied in recipient mice that exhibit microbiome-dependent protection against 6-Hz seizures.
[0131] The observed metagenomic signatures reveal clinical KD-associated changes in the functional potential of the gut microbiome that are preserved upon transfer to GF mice. To identify clinical KD-induced alterations in the functional activity of the gut microbiome, untargeted metabolomic profiling of aliquots of the same donor fecal samples from pediatric epilepsy patients collected before and after initiating the KD regimen, and of both fecal and serum samples from recipient mice that were inoculated with the pre-KD or post-KD human fecal microbiota and fed CD was performed. Results from clinical laboratory testing of human blood samples confirmed that the month-long clinical KD regimen elevated serum P- hydroxybutyrate (BHBA) levels and reduced serum glucose levels, relative to pre-KD concentrations, in pediatric refractory epilepsy patients (6B). Decreases in glucose, but not BHBA, were similarly seen in human post-KD stool samples relative to matched pre-KD controls (6B), which is consistent with dietary carbohydrate restriction and KD-induced BHBA synthesis by the liver to elevate systemic, but not fecal, BHBA levels (Westman et al., 2007). Transfer of the post-KD human microbiota into mice yielded no significant differences in serum BHBA or glucose relative to pre-KD recipient controls (6B), indicating that the clinical KD-associated microbiota does not sufficiently promote key systemic features of ketosis in mice fed the standard CD. Notably, however, mice that were inoculated with post-KD human microbiota and fed CD exhibited statistically significant increases in fecal BHBA levels relative to matched pre-KD recipient controls (6B). This could reflect alterations in intestinal synthesis of BHBA (Mierziak et al., 2021) and / or in microbial utilization of host-derived BHBA (Ang et al., 2020). Since this effect was not seen in the donor human fecal samples, we reasoned that this phenotype is likely an artifact of the “transplantation” approach and / or experimental design, and therefore not relevant to the clinical condition. These results suggest that transfer of the clinical KD-associated human gut microbiota into mice promotes resistance to 6-Hz seizures (Figure 2B) through mechanisms that act independently of ketosis.
[0132] Results from untargeted metabolomic profiling were assessed to identify metabolites that were differentially abundant in post-KD samples relative to pre-KD controls and patterns that were shared across human donor and mouse recipient samples (6C, 2E, and Figure 7). Despite heterogeneity in the patient population and specific clinical KD regimens, 79 metabolites were significantly differentially abundant in fecal samples from post-KD human fecal samples relative to their matched pre-KD controls (Figure 7A). 336 metabolites were identified in both the human fecal samples and mice fed the 6: 1 KD chow vs. vitamin- and mineral-matched control chow for 2 weeks samples, previously published by our group (Table S14, Olson et al., 2018). 35 metabolites were differentially regulated in human fecal samples and 169 metabolites were differentially regulated in mouse fecal samples (Figure 7B). Of these significantly altered metabolites, 20 were found to be changed in the same direction across human and mouse samples (Figures 7B and 7C). These included KD-induced increases in levels of metabolites related to fatty acid beta-oxidation, such as palmitoleoylcamitine (C16: l) and oleoylcamitine (C18:l), and a decrease in kynurenine which have previously been associated with seizure susceptibility (Zamowska et al., 2019). This statistically significant overlap suggests that there are biochemical changes that are shared across clinical KD treatments for pediatric epilepsy and mouse models of KD, and that some of the fecal metabolomic alterations observed in KD-treated epilepsy patients are a direct consequence (rather than correlate) of dietary intervention. Of the 20 significantly differentially abundant metabolites shared in human and mouse, 14 (~70%) were further significantly altered by antibiotic treatment to deplete gut bacteria in KD-fed mice (Figures 7B, 7C, and 7D , Olson et al., 2018). Taken together, these data indicate that clinical KD regimens alter fecal metabolites in children with refractory epilepsy, a subset of which have the potential to be microbiome-dependent.
[0133] Although there was substantial variability in taxonomic composition of microbiota within mouse recipients of the same experimental condition (Figure 1 and Figure 5), fecal samples from mouse recipient cohorts exhibited statistically significant alterations in 45 metabolites that were shared when considering all post-KD mouse recipient fecal samples relative to their pre-KD controls (Figure 7A). Notably, however, none of these 45 differentially abundant metabolites in mouse feces were identical to the 79 differentially abundant metabolites seen in human donor fecal samples (Figure 7A), which could reflect host specific metabolite utilization and the fact that recipient mice were fed standard chow (CD), while human donors were consuming a clinical KD at the time of sample collection. To gain insight into whether the differentially abundant metabolites relate to similar biological functions, metabolite set enrichment analysis (MSEA) of the significantly altered metabolites in human vs. mouse was performed (Pang et al., 2021). MSEA of the significantly altered metabolites identified select chemical classes (6C) and metabolic pathways (6D) that were similarly enriched in both human donor and mouse recipient post-KD samples relative to pre- KD controls. In particular, there was shared enrichment of amino acid, hydroxy fatty acid, sugar acid, phenylpropanoic acid, and monosaccharide-related metabolites across post-KD conditions for both human donors and mouse recipients (6C). Differentially abundant metabolites from human post-KD fecal samples also exhibited enrichment of bile acids and other fatty acid derivatives (6C, left), which might reflect KD- and / or microbiome-driven alterations in lipid metabolism (Joyce et al., 2014).
[0134] For metabolic pathways, post-KD samples for both human donor and mouse recipient conditions exhibited differential abundance of metabolites related to methionine metabolism, glycine and serine metabolism, and betaine metabolism (6D). While the biological relevance to KD is unclear, one possibility is that these pathways reflect known influences of the KD on one-carbon (1C) metabolism, a series of interlinking metabolic pathways that control levels of methionine, serine, and glycine, and that integrate nutrient availability with cellular nutritional status (Ducker & Rabinowitz, 2017). In addition, differentially abundant fecal metabolites from mouse post-KD recipients mapped to pathways related to alpha linolenic acid and linoleic acid metabolism, fatty acid biosynthesis, and beta-oxidation of very long chain fatty acids (6D, right), which aligns with the observed metagenomic enrichment in microbial genes related to fatty acid metabolism in response to the clinical KD (6A). Some of the differential metabolite chemical subclasses and metabolic pathways in mouse fecal samples were similarly seen in matched mouse serum samples - in particular, metabolites representing amino acid, hydroxy fatty acid, and unsaturated fatty acid subclasses, and related to alpha linolenic acid and linoleic acid metabolism, betaine metabolism, and beta-oxidation of fatty acids were altered in both feces and serum of mice receiving post-KD samples relative to pre-KD control samples (Figure 7E). Taken together, these results suggest that the clinical KD induces alterations in the function of the gut microbiome of pediatric epilepsy patients, and that a subset of these functional characteristics may be phenocopied upon microbial transfer to mice, which develop microbiome-dependent resistance to 6-Hz seizures.
[0135] Transferring the fecal microbiota from KD-treated pediatric epilepsy patients to mice induces alterations in brain gene expression
[0136] Seizures result from atypical neural function related to discharge of electrical signals or failure to constrain the spread of these signals. To gain insight into how colonization with microbes derived from the fecal microbiota of KD-treated individuals may alter brain function to modify seizure susceptibility, transcriptomic profiling of brain tissues from cohorts of mice colonized with microbes from the post-KD human microbiota or pre-KD controls was performed, focused on the hippocampus and frontal cortex based on their relevance to human epilepsy, their involvement in initiating psychomotor seizures in the 6-Hz seizure assay, and evidence that the microbiome can alter gene expression and metabolites in these brain regions (Chauhan et al., 2022; Suarez et al., 2018). RNA sequencing of hippocampal tissues revealed many differentially expressed genes that were seen in post-KD samples relative to pre-KD controls, including those related to core cell biological processes relating to RNA processing, translation, cellular stress response, TORC1 signaling, regulation of long-term synaptic potentiation, neuronal development, and response to nutrient levels (Figure 8A). The most drastic alterations included upregulation of Duspl2, Bmprlb, and Cmya5 and downregulation of Abcc9, Ufspl, and Tbx2 transcripts (Figure 8B). Duspl2 is a dual specificity phosphatase that can dephosphorylate phosphothreonine and phosphoserine (Muda et al., 1999), Bmprlb, a member of the bone morphogenic receptor family, is a serine / threonine kinase influencing neuronal cell fate (Venugopal et al., 2012), and Cmya5 encodes for myospyrn which is essential for structural integrity during neuritogenesis (Hsiung et al., 2019). Abcc9 is an ATP-binding cassette transporter encoding the sulfonylurea receptor 2 subunit for potassium channels (Nelson et al., 2015), Ufspl is a Ufml specific protease that regulates ubiquitin-like conjugation and has been linked to seizures (Millrine et al., 2022), and Tbx2 is a transcription factor linked to neuronal cell cycle control and neuroinflammation (Reinhardt et al., 2019). STRING network analysis additionally revealed top protein interaction clusters enriched for essential biological processes including RNA processing, oxidative phosphorylation, and cell cycle regulation, consistent with results from GO enrichment analysis (Figure 8A, 8C, 8D), as well as endocytosis and glutathione metabolism (Figure 8D).
[0137] Some differentially expressed genes were also identified in frontal cortical tissues of post-KD recipients relative to pre-KD controls, which similarly to hippocampus, included those related to core cell biological processes for RNA surveillance and catabolism, cellular stress responses, TORC1 signaling, and further included genes related to potassium ion transport, and core carbohydrate metabolism (Figure 9A). The most drastic alterations included upregulation of Serpinbla, Nqol, and Slc6al2 transcripts, and downregulation of AldhSbl, Setmar, and Tfblm transcripts (Figure 9B). Serpinbla is a serine / cysteine protease inhibitor (Huasong et al., 2015), Nqol encodes an antioxidant enzyme that primarily catalyzes the reduction of quinones (Ross & Siegel, 2021), and Slc6al2 encodes for a betaine-GABA transporter (Zhou et al., 2012). Aldh3bl is an aldehyde dehydrogenase linked to oxidative stress reduction (Marchitti et al., 2007), Setmar encodes a histone-lysine N- methyltransferase (Cordaux et al., 2006), and Tfblm has been shown to function as methyltransferase (Metodiev et al., 2009). STRING network clustering analysis additionally revealed top protein interaction clusters enriched for transcription regulation, translation, and oxidative phosphorylation, also seen in frontal cortex GO enrichment analysis and in the hippocampal STRING network, as well as clusters enriched for calcium signaling, transcriptional regulation, and translation (Figure 9C). Differentially expressed gene sets from both hippocampus and frontal cortex were enriched for TORC1 signaling, cellular response to stress, and oxidate phosphorylation through GO enrichment and STRING clustering, which have all been shown to affect seizure susceptibility (Chan, 2001; Nguyen & Bordey, 2021). The similarities between transcriptomic results from hippocampus and frontal cortex suggest that colonization with post-KD microbes elicits key alterations in host metabolism that impact core biological processes that are generally consistent across different brain regions. Overall, these results indicate that mice that acquire seizure resistance in response to colonization with microbes from the post-KD human gut microbiota exhibit alterations in hippocampal and frontal cortical gene expression, relative to pre-KD recipient controls.
[0138] Multi ’omics analysis reveals network connections linking microbial genomic pathways and metabolites to hippocampal transcripts related to epilepsy
[0139] To further identify key gut microbial functions that may drive particular brain gene expression signatures, microbe-metabolite vectors (MMVEC) (Morton et al., 2019) were used to build an integrated network of fecal metagenomic, fecal metabolomic, serum metabolomic, hippocampal transcriptomic, and frontal cortical transcriptomic datasets from mice inoculated with the human pre-KD or post-KD microbiota. A parallel network comprised of fecal metagenomic and fecal metabolomic datasets from human pre-KD and post-KD donor stool samples was generated to identify features similarly underscored in both human and mouse networks, suggesting their clinical relevance. The human donor and mouse recipient networks were linked by 6 common nodes - metagenomic pathways describing adenine and adenosine salvage (PWY-6609), dTDP-L-rhamnose biosynthesis (DTDPRHAMSYN-PWY), folate transformations (PWY-3841), L-valine biosynthesis (VALSYN-PWY), preQO biosynthesis (PWY-6703), and sulfoglycolysis (PWY-7446) . The shared dTDP-L-rhamnose biosynthesis pathway was identified by weighted key driver analysis in both the human donor and mouse recipient networks, and adenine and adenosine salvage, folate transformations, L-valine biosynthesis, and sulfoglycolysis nodes in the human donor network, were further identified by weighted key driver analysis as highly interconnected across the ‘omics datasets and essential regulator nodes of the network (Ding et al., 2021) . The human donor network also contained additional key driver metagenomic nodes of interest including acetyl-CoA fermentation to butanoate II (PWY-5676, connected to shared nodes VALSYN-PWY, PWY6609, and DTDPRHAMSYN-PWY), pyruvate fermentation to acetate and lactate II (PWY-5100, connected to shared node PWY-7446), branched chain amino acid (BCAA) biosynthesis (BRANCHED-CHAIN-AA-SYN-PWY, connected to shared node PWY-6609), and super pathway of fatty acid biosynthesis (FASYN-INITIAL-PWY, connected to shared node PWY-6609) - each of which has been previously linked to the microbiome, KD, and / or neurotransmission (Dalile et al., 2019; Larsson & Markus, 2017; McKenna et al., 2019). Additionally, the mouse recipient network contained L-isoleucine biosynthesis (PWY-5101), which aligns with the human donor metagenomic node relating to BCAA biosynthesis. Consistent with the shared metagenomic nodes and key drivers between mouse and human networks, the human fecal metabolomic module was enriched for nodes related to valine, leucine, and isoleucine (BCAAs) biosynthesis and phenylalanine, tyrosine and tryptophan biosynthesis. In the mouse network, the fecal metabolomic module included several nodes related to glycerophospholipid metabolism for fecal metabolites. Nodes for fecal l-(l-enyl-palmitoyl)-2-linoleoyl-GPC (P- 16:0 / 18:2)*, l-(l-enyl-palmitoyl)-2-palmitoyl-GPC (P-16:0 / 16:0)*, l-(l-enyl-palmitoyl)-2- arachidonoyl-GPC (P-16:0 / 20:4)*, and myo-inositol were similarly identified as differentially abundant in individual metabolomic analyses for recipient post-KD fecal samples relative to pre-KD controls. These mouse metagenomic modules were linked directly to 8 transcriptomic modules for hippocampal genes and indirectly to 1 frontal cortex module. The hippocampal transcript modules were enriched for nodes related to cellular response to nerve growth factor stimulus, Wnt signaling pathway and regulation of mitochondrial fission, neuron migration and generation, cellular response to amino acid starvation, long-chain fatty-acyl-CoA metabolic process, glycosylphosphatidylinositol (GPI) anchor biosynthetic process, GPI anchor metabolic process and glycolipid biosynthetic process, and regulation of telomerase RNA localization to Cajal body. The frontal cortical transcripts were enriched for nodes related to positive regulation of transcription of Notch receptor target. This suggests that these particular biological processes are most closely associated with the microbial functional features identified in the network. The transcript nodes included 38 hippocampal genes that were similarly identified in individual transcriptomic analyses as differentially expressed in post-KD recipient mice relative to pre-KD controls . The higher direct number of connections between metagenomic modules and hippocampal transcripts suggests that the gut microbiome may exhibit a greater regulatory role for the hippocampus than for the frontal cortex in post- KD recipient mice compared to pre-KD controls. Of particular interest are the indirect links between fecal metabolites related to glycerophospholipid metabolism, which are regulated by the microbiome (Zheng et al., 2021), hippocampal transcript modules enriched for cellular response to nerve growth factor stimulus, Wnt signaling, neuron migration, cellular response to amino acid starvation, and GPI anchor biosynthetic and metabolic processes implicated in seizure susceptibility (Gall et al., 1991; Hodges & Lugo, 2018; Qin et al., 2017; Takeuchi et al., 2021; Wu et al., 2020), and the frontal cortical transcript module enriched for transcription of Notch receptor target pathway also implicated in seizure susceptibility (Hodges & Lugo, 2018; Wu et al., 2020). Overall, several genes were identified as having differential regulation associated with the microbial functional features identified in the network (Figure 10A), including: Mcf2, Nfxl, Odf21, Pedsl, Snordl3, Zfpl42, Cfdpl, Dnajcl8, Eapp, Golga2, Oligl, Qtrtl, Slitl, Tbxl8, Tmeml84b, Acaa2, Arhgap21, Bazlb, Ccnq, Cdh4, Faml07a, Fbxo8, Gnaz, Ino80b, Lhfpl4, Loxl3, Mnsl, Mpg, MyhlO, 01fm3, Phip, Rgs8, Runxltl, Slc5a3, Atp8a2, Clic5, Cln8, Ctxn2, Dhx32, Fam204a, Fcerlg, Gmcll, Irf2, Lampl, Mpdz, Rnfl38, Scaf4, Slc37a4, Smim24, Tpbg, Def8, Mro, Rprd2, Bloclsl, Faml85a, DocklO, Kctd20, Mettl23, Rbml8, Tbx3osl, Zfp87, ArD, Bapl, Calr, Caprinl, Ccdcl36, Ehmt2, Eloa, Gtpbp6, Hars, Hnrnpab, Ingl, Lrrtml, Nebl, Nfe211, Nrbpl, Nubl, Nupl53, Ppplrla, Rabl2, Rell2, Tbllxrl, Zcchc2, Zfp397, Ecsit, Gabarapll, Hlf4, Kdm6b, Lonp2, Meigl, Ntpcr, Pcedla, Prkce, Stmpl, Tmem237.
[0140] To gain insight into whether the hippocampal and frontal cortical transcripts that cooccur directly and indirectly with microbial metagenomic and metabolomic features have been implicated in human epilepsy, single nucleotide polymorphisms (SNPs) identified from epilepsy genome-wide association studies (GWAS) were mapped to genes using hippocampus and frontal cortex splicing quantitative trait loci (sQTLs) and expression quantitative trait loci (eQTLs) to represent epilepsy-associated genes informed by GWAS. The mouse orthologs of these human GWAS disease genes were then compared with hippocampal and frontal cortical transcriptomic results to identify DEGs in post-KD vs pre- KD recipients that have been implicated in genetic risk for human epilepsy. There was a statistically significant enrichment of the hippocampal DEGs in the epilepsy GWAS (p=0.003), but no significant enrichment of the frontal cortical DEGs in the epilepsy GWAS (p=0.26) (Figure 10B). These results suggest that microbial alterations in hippocampal gene expression may contribute to the microbiome-dependent increases in seizure resistance seen in post-KD recipient mice compared to pre-KD controls. From the co-occurrence network, 7 hippocampal DEGs were linked to epilepsy GWAS results: Atf4, which encodes a cAMP- response element binding protein 2 belonging to the AP-1 family of transcription factors; Taf9b, which encodes for the 9b subunit of TFIID coordinating initiation of RNA polymerase II transcription; Pxk, which encodes a phox domain-containing Serine / Threonine kinase that regulates synaptic transmission through binding brain Na-ATPase and K-ATPase; Rprd2, which encodes a transcriptional repressor that modulates RNA polymerase II activity; Gnaz, which encodes G protein alpha subunit that regulates ion equilibrium and chaperone- mediated folding; Cfdpl, which encodes a subunit of the chromatin remodeling complex and is important for cell division, and Mro, which encodes a nucleolus protein proposed to be testis-determining in the reproductive tract, but expressed in the nervous system with as yet unknown function. Overall, the multi’omics analysis of human donor and mouse recipient datasets together with epilepsy GWAS mapping to hippocampal and frontal cortical DEGs identified key microbial genomic pathways and microbially modulated metabolites that may directly and indirectly contribute to alterations in the expression of particular hippocampal genes in mice that exhibit microbiome-induced protection against 6-Hz seizures.
[0141] DISCUSSION
[0142] Results from this research provide evidence from a treatment study of children with refractory epilepsy, coupled with functional testing in gnotobiotic mice, that clinical KD regimens alter the function of the gut microbiome in ways that could contribute to seizure protection. Microbiome composition and function in 10 children with refractory epilepsy was assessed shortly before initiating and approximately one month after adherence to classical KD regimens. Following clinical practice, the patient cohort was heterogeneous in type and underlying cause of refractory epilepsy, as well as the ratio of fat to carbohydrate and protein and specific nutritional composition of the KD they consumed. This highlights the diversity of epilepsies that resist current antiepileptic drugs and the broad range of KD interventions that are administered to treat pediatric refractory epilepsy. Consistent with this heterogeneity, it was observed that participants varied substantially in the composition of the fecal microbiota at baseline and in response to KD treatment. There was no clear KD-induced taxonomic signature of the gut microbiota that was shared across the study population, which contrasts prior studies of KD treatments for epilepsy that each reported alterations in the gut microbiota in response to a KD. Our results, however, support the finding that little to no consistency in specific microbial taxa affected exists across studies (Ozcan et al., 2022).
[0143] Despite variation in microbiota composition, evidence of shared functional features of the gut microbiome was seen with KD treatment across participants in the study. This aligns with the notion of functional redundancy of gut microbes, wherein phylogenetically unrelated species can exhibit the same genetically-encoded biological activities (Tian et al., 2020). Results from metagenomic sequencing indicated that microbial genes related to fatty acid P- oxidation, 2-methylcitrate cycle, glycol metabolism, and proline biosynthesis were more highly represented in the gut microbiota of epileptic children after treatment with the KD compared to their internal pre-treatment controls. -oxidation by select microbes in anaerobic environments enables them to utilize fatty acids from the diet as energy sources, wherein saturated and unsaturated fatty acids are oxidized into acetyl-CoA (Yao & Rock, 2017). P- oxidation of dietary odd-chain fatty acids additionally produces propionyl-CoA, which can be toxic to cells, so the methylcitrate cycle enables microbes to further catabolize propionyl- CoA into pyruvate and succinate (Dolan et al., 2018). Glycol, including glycolate and glyoxylate, metabolism allows microbes to use products from fatty acid oxidation to fuel gluconeogenesis (Ahn et al., 2016). Proline synthesis from the central metabolite glutamate, via intermediates amino acids arginine and ornithine, is widely upregulated in bacteria to counteract growth in osmotically unfavorable conditions (Stecker et al., 2022). The elevated representation of genes related to these pathways in the post-KD samples suggests that the clinical KD shapes the gut microbiome to enrich microbial taxa that digest fat and synthesize carbohydrates under fat-rich, carbohydrate-limited conditions. These metagenomic features of the human microbiome from pediatric epilepsy patients consuming a clinical KD were preserved upon transfer to GF mice that were fed a standard diet, suggesting that the source microbes are maintained under non-ketogenic dietary conditions.
[0144] Aligning with results from metagenomic sequencing, metabolomic profding of fecal samples from KD-treated epileptic children revealed statistically significant alterations in several metabolites, including subsets of amino acids, sugar acids, hydroxy fatty acids, bile acids, and other fatty acid derivatives, which reflect KD-, and potentially microbiome-, induced alterations in lipid and amino acid metabolism. In particular, glutamate and ornithine, both precursors of proline, were significantly decreased in post-KD human samples, relative to pre-KD controls, which may align with the observed metagenomic alterations in proline biosynthesis pathways. These metabolite alterations were induced by KD consumption in mice, and modified by microbiota depletion in mice, suggesting a causal response to the clinical KD in the human cohort that is dependent on the gut microbiome. Microbially modulated increases in palmitoleoylcamitine (Cl 6:1) were also seen in KD-fed mice and in post-KD human samples, alongside several other lipid species, aligning with the high fat content of the KD and roles for the gut microbiome in lipid metabolism (Joyce et al., 2014). Notably, however, there were no significant effects of the clinical KDs on fecal levels of gamma-glutamylated amino acids in our human cohort, which contrasts our previous observations in mice fed KD chow (Olson et al., 2018). This highlights key differences in the effects of chow-based versus clinical KDs in mice versus humans and further indicates that studying KDs in different animal species will yield both shared and distinct molecular signatures.
[0145] The individual metabolite changes seen in human donors, including those induced by KD in a microbiome-dependent manner, were not specifically recapitulated by microbiome transfer to GF mice that were fed standard chow. This is perhaps not surprising given the important role of dietary composition in driving microbial activity (David et al., 2014). While no specific metabolite shifts were shared, a few pathway-level metabolomic changes were consistent between post-KD fecal samples from human donor (consuming the clinical KD) and mouse recipients (fed standard chow), relative to their respective pre-KD controls. Namely, differentially abundant metabolites related to metabolism of methionine, glycine, serine, and betaine were shared across post-KD conditions for human donor and microbiotarecipient mice. Methionine metabolism involves the production of homocysteine, adenosine, cysteine, and alpha-ketobutyrate, which can then be routed to glucogenic pathways by conversion to propionyl- and succinyl-CoA. Serine, synthesized via glycerate, is used to create glycine (and cysteine) via the homocysteine cycle, which can undergo microbial conversion into pyruvate or glyoxylate. Betaine (trimethylglycine), derived from diet or synthesized from choline, is metabolized by the gut microbiome (Koistinen et al., 2019) and functions as a methyl donor in transmethylation reactions, including those involved in methionine metabolism. While the relevance to KD and seizure protection is unclear, alterations in peripheral and central amino acid metabolism have been widely implicated in mediating the anti-seizure effects of the KD (Yudkoff et al., 2001). In addition, post-KD samples from both human donors and mouse recipients exhibited alterations in chemicals related to lipid metabolism, such as hydroxy fatty acids. The shared metabolite pathway- and chemical subclass-level features may reflect changes that are induced by KD in humans and generally phenocopied by gut microbes upon transfer to mice reared under non-ketogenic conditions. This suggests that transfer of clinical KD-induced gut microbes to mice maintained under non-ketogenic conditions could result in molecular outputs that are distinct, but functionally similar, to those seen in the donor human sample.
[0146] It was observed that inoculating mice with human fecal samples collected after clinical KD treatment conferred resistance to 6-Hz seizures across all post-KD donor conditions compared to controls that received the baseline pre-treatment (pre-KD) microbiota. There was no correlation with patient responsiveness to diet, as indicated in clinician notes taken at 1 month after adherence to the clinical KD. This may be due to the unreliability of the metric, which was based on parental reporting, as well as the cross- sectional nature of the assessment, given inter-individual variation in latency to respond to KD treatments and the patient’s peak KD ratio. These concerns aside, the results highlight the importance of host determinants of KD responsiveness, some of which may mask or block any beneficial influences of the KD-associated microbiota. Many patients included in this study exhibited genetic bases for refractory epilepsy, some of which could be epistatic to functional genomic changes in the KD-associated gut microbiome. Large human studies that subclassify different types of epilepsies and seizure semiologies and comparisons to healthy human controls are warranted to study potential roles for the gut microbiome in modifying or predicting responsiveness to the KD.
[0147] Following clinical procedure, epilepsy is deemed “refractory” after failed response to adequate trials of 2-3 ASMs, upon which the KD may be considered as a therapeutic option. In some cases, individuals will continue ASM treatment during the course of KD. In this study cohort, 7 out of 10 participants continued ASM regimens that remained unchanged during the course of the experiment from collection of pre-KD to post-KD samples. All mouse cohorts colonized with post-KD donor fecal sample displayed increased seizure threshold compared to their paired pre-KD colonized cohort (Figure 2B). Additionally, all pre-KD colonized mouse cohorts did not display increased seizure thresholds compared to controls. In fact, pre-KD colonized mouse cohorts displayed similar or decreased seizure thresholds compared to control groups (Figure 2B), leaving open the possibility of the pre-KD associated microbiota contributes to seizure risk in these cohorts. As such, it is unlikely that residual ASM content (presumed to be present in both pre-KD and post-KD samples) can sufficiently account for these differential effects on seizure susceptibility. Moreover, post-treatment with antibiotics abrogated the increased seizure thresholds seen in post-KD recipient mice, as compared to vehicle-treated controls (Figure 3F). This suggests that bacterial members of the gut microbiome are required for the seizure protective effects of the post-KD microbiota, and further suggests that they are not mediated by any residual ASM content in the stool. Lastly, many ASMs display short half-lives in vivo. For example, levetiracetam, which was taken by 8 out of 10 participants in our study has a half-life of approximately 1.5 hours in mice (Markowitz et al., 2010). In our microbiota transfer experiments, 6-Hz seizure testing was performed at 4 days after inoculation with fecal suspension, a time point at which any residual ASMs like levetiracetam was expected to be fully cleared from the host. Taken together these data suggest that any residual ASMs in donor fecal samples do not account for the increases in seizure threshold seen upon transfer of the post-KD microbiota relative to pre-KD controls.
[0148] The microbiota-dependent increases in seizure protection were associated with brain transcriptomic alterations. In particular, both hippocampus and frontal cortex from post-KD recipient mice exhibited enrichment of differentially expressed genes related to i) RNA processing, transcriptional regulation, and translation ii) T0RC1 signaling and cell cycle, and iii) oxidative phosphorylation and cellular stress response (i.e., nitric oxide, reactive oxygen species), when compared to controls colonized with pre-KD microbiota from both GO enrichment and STRING network analyses. Neuronal excitability requires protein synthesis in response to altered neuronal stimulation, and risk factors for various epilepsies include dysregulation of RNA processing, RNA stability, transcription, and translation (Malone & Kaczmarek, 2022). TORC1 is a major nutrient- and energy-sensing serine / threonine kinase complex that controls cell growth and differentiation by coordinating core processes of transcription, translation, and autophagy. Abnormal regulation of TORC1 signaling has been implicated in a wide variety of epilepsies, and as such, is a therapeutic target of interest for treating refractory epilepsies (Nguyen & Bordey, 2021). Previous studies have reported inhibitory effects of the KD and select fatty acids on TORC1 activity (McDaniel et al., 2011; Warren et al., 2020), suggesting that it may contribute to the anti-seizure effects of the KD. Oxidative phosphorylation is a central process for cellular energy metabolism from nutrients, that generates as a byproduct reactive oxygen species (ROS) (Rowley & Patel, 2013) and is regulated by the retrograde glutamatergic neurotransmitter nitric oxide (NO). In animal epilepsy models, both ROS and NO are elevated during seizure activity due to oxidative stress-associated neuronal death (Zhu et al., 2017), which can further contribute to epileptogenesis (Chan, 2001). The KD has been previously reported to reduce oxidative stress by promoting antioxidant enzymatic activity and scavenging ROS (Greco et al., 2016). Overall, these results suggest that the KD-associated human gut microbiota alters brain transcriptional pathways that may contribute to protection against 6-Hz seizures in mice.
[0149] Integration of multi ’omics datasets across human donor and mouse recipients revealed direct and indirect network associations between select gut microbial metagenomic pathways, fecal metabolites, and hippocampal transcripts, suggesting that they may contribute to the microbiome-dependent increases in seizure protection seen in mice inoculated with human post-KD microbiota, compared to pre-KD controls. The human donor and mouse recipient co-occurrence networks were linked by shared metagenomic pathway nodes related to adenine and adenosine salvage, dTDP-L-rhamnose biosynthesis, folate transformations, L- valine biosynthesis, preQO biosynthesis, and sulfoglycolysis. Shared key driver for dTDP-L- rhamnose biosynthesis and shared nodes for L-valine biosynthesis and adenine and adenosine salvage were linked to mouse recipient network hippocampal transcript modules enriched for genes related to neuron generation and migration, cellular response to amino acid starvation, and Wnt signaling and regulation of mitochondrial fission. Catabolism of L-rhamnose produces L-lactaldehyde under aerobic conditions and dihydroxyacetone phosphate (DHAP) under anaerobic conditions; both can be further catabolized to pyruvate, a major precursor for BCAA synthesis and input for the TCA cycle (Foster et al., 2014; van Maris et al., 2006). Additionally in the human donor network, the shared node related to adenine and adenosine salvage was linked to BCAA biosynthesis while the fecal metabolomic module was enriched for BCAA biosynthesis. BCAAs modulate brain import of precursors required for synthesis of monoamine transmitters (Larsson & Markus, 2017; Salcedo et al., 2021; Song et al., 2017). BCAAs also serve as nitrogen donors for synthesis of glutamate vs. GABA, and as such regulates synaptic balance between excitation and inhibition, a key determinant of seizure susceptibility (McKenna et al., 2019). Wnt signaling regulates calcium pathways that are important for hippocampal neurogenesis and dendrite formation, and is increasingly linked to early epileptogenesis (Hodges and Lugo, 2018). Additionally, mapping GWAS-based risk genes to the co-occurrence network identified seven hippocampal nodes as linked to epilepsy. Of particular interest was Gnaz. which encodes G protein alpha-Z, a protein that mediates neuronal signal transduction within the hippocampus (Jang et al., 2018) and is proposed to modulate seizure susceptibility (Hultman et al., 2019). BCAA derivatives have been reported to promote phosphorylation of G-proteins, and abnormalities in GPCR mediated neuronal signaling can contribute to increased susceptibility to seizure (Shellhammer et al., 2017; Yu et al., 2019). Altogether, results from this study reveal that the clinical KD regimens used to treat pediatric refractory epilepsy are associated with alterations in the function of the child microbiome, which causally modify brain function and seizure susceptibility upon transfer to mice. Further research is warranted to define the mechanisms by which the human KD- associated microbiome signals across the gut-brain axis to modify seizure risk, and to further assess the potential for identifying microbiome-based interventions that could increase the efficacy of KD treatment, alleviate dietary side effects, and / or ease clinical implementation.
[0150] EXPERIMENTAL MODELS AND SUBJECT DETAILS
[0151] Human Subjects
[0152] This study was approved by UCLA’s Institutional Review Board (IRB protocol #15- 000453). Pediatric refractory epilepsy patients were screened and enrolled in collaboration with the Ketogenic Diet Program at UCLA Mattel Children’s Hospital. Prospective participants who met study criteria were provided information detailing this study by phone and email 1-2 weeks before their pre-diet initiation visit. Prior to enrollment, informed signed consent was provided by all participants and their guardians to the program clinical coordinator during the pre-diet initiation appointment. Subjects were enrolled across diverse seizure semiology and prior medical histories. Inclusion criteria: enrolled in UCLA’s program for classical 4:1 KD, children aged 1-10 with refractory epilepsy, any gender, any ethnicity, any previous exposure to AEDs, any seizure semiology. Exclusion criteria: use of antibiotics or probiotics within 4 weeks prior to enrollment, existing diagnosis of gastrointestinal, immunological, or metabolic disorder. Human donor stool samples were collected from 10 participants, each providing 2 stool samples. The first sample was collected within 1 day before starting KD treatment (pre-KD) and the second sample was collected after maintaining on the clinical KD for 1 month (post-KD). Clinical metadata from the medical record were coded and stripped of identifiers before being shared, and included participant demographic data, medical history, AED exposure history, additional medications take during this study, laboratory blood glucose and bloody ketone body levels, seizure severity, seizure frequency, seizure semiology, and dietary regimen.
[0153] Human stool sample collection
[0154] For in-patient fecal sample collection, once a study participant was admitted to the hospital during the pre-diet initiation visit, they were given a coded stool collection kit and sterile specimen container. Stool samples were freshly collected within 1 day prior to starting the clinical KD treatment (pre-KD). Fresh stool samples were immediately placed on dry ice for short term storage and transportation and were freshly frozen at -80°C for long-term storage. Post-KD stool samples were collected in the same manner as stated above when the study participant returned for the 1 -month follow-up visit. For out-patient collection of the post-KD stool sample, which was necessitated because of hospital pandemic policies, a deidentified stool sample collection kit and sterile specimen cup was provided to the patient and guardian along with a pre-labeled return shipping box. After 1 month of the clinical KD treatment, stool samples were collected in a sterile specimen cup, immediately placed in an at home freezer, and the next day either (1) shipped back overnight to UCLA on dry ice or (2) brought with the patient to their 1 month follow-up appointment. Fresh frozen fecal samples were homogenized under liquid nitrogen and 3 -500 mg aliquots were made per sample by sterile storage in anaerobic Balch tubes to be used for transplantation, metagenomic, and metabolomic studies.
[0155] Mice 6-8 week old wild-type germ-free Swiss Webster mice (Taconic Farms), were bred in UCLA’s Center for Health Sciences Barrier Facility. Breeding animals were fed “breeder” chow (Lab Diets 5K52). Experimental animals were fed vitamin- and mineral-matched control diet (Harlan Teklad TD.150300). Juvenile mice were used to mimic the age range of the human donor population (<10 years old). All animal experiments were approved by the UCLA Animal Care and Use Committee.
[0156] METHOD DETAILS
[0157] 16S rRNA Gene Sequencing and Analysis
[0158] Bacterial genomic DNA was extracted from human or mouse fecal samples using the Qiagen PowerSoil Kit. For human samples, the n reflects one donor sample. For mouse samples, the n reflects independent cages containing 3 mice per cage to preclude effects of co-housing on microbiota composition. The sequencing library was generated in line with (Caporaso et al., 2011). PCR amplification, run in triplicate, of the V4 region of the 16S rRNA gene was completed using individually barcoded universal primers and 30 ng of the extracted genomic DNA. The PCR product triplicates were pooled and purified using the Qiaquick PCR purification kit (Qiagen). Samples were sequenced using the Illumina MiSeq platform and 2 x 250bp reagent kit for paired-end sequencing at Laragen, Inc. Amplicon sequence variants (AS Vs) were chosen by closed reference clustering based on 99% sequence similarity to the SILVA138 database. Taxonomy assignment, rarefaction, and differential abundance testing were performed using QIIME22022.2 (Bolyen et al., 2019; Mandal et al., 2015). ANCOM analysis were performed using feature table outputs generated from QIIME2. Feature tables were prepared using standard methodology recommended by QIIME2 implementing add-pseudocount imputation to address all zero-values followed by a log-transformation before standard ANCOM analysis (Bolyen et al., 2019). Multivariate association analysis of collected donor patient clinical metadata was performed using MaAslin2 with recommended and appropriate normalization (total sum squares), transformation (log transformation), and analysis model (linear model) parameters (Mallick et al., 2021).
[0159] Fecal Shotgun Metagenomics
[0160] Bacterial genomic DNA was extracted from human or mouse fecal samples using the Qiagen PowerSoil Kit. 1 ng of DNA was used to prepare DNA libraries using the Nextera XT DNA Library Preparation Kit (Illumina) and genomic DNA was fragmented with Illumina Nextera XT fragmentation enzyme. IDT Unique Dual Indexes were added to each sample before 12 cycles of PCR amplification. AMpure magnetic Beads (Beckman Coulter) were used to purify DNA libraries which were eluted in QIAGEN EB buffer. Qubit 4 fluorometer and Qubit dsDNA HS Assay Kit were used for DNA library quantification. Libraries were then sequenced on Illumina HiSeq 4000 platform 2xl50bp at a 6M read depth using by CosmosID. Metagenomic data was analyzed using HUMAnN 3.0 (Beghini et al., 2021) and MetaCyc database to profile gene families and pathway abundance. File2meco R package was used for MetaCyc pathway hierarchical classification (Liu et al., 2022). MaAsLin 2.0 (Mallick et al., 2021) was used to assess significant pathway associations between pre-KD and post-KD with a non-adjusted p-value cutoff of 0.1, where non-adjusted p-value < 0.05 pathways are indicated in the figure by asterisk. Heatmaps were generated using the pheatmap vl.0.12 package for R.
[0161] Human Donor Fecal Microbiota Transfer
[0162] To prepare collected human stool samples for transplantation studies, the frozen stool sample was pulverized into a powder under liquid nitrogen stream in a sterile heavy-duty foil covered mortar and pestle, aliquoted at 500 mg per tube into 2mL screw cap tubes, and frozen at -80C. A single 500 mg aliquot of human stool sample was entered into a Coy anaerobic chamber and resuspended in pre-reduced lx PBS + 0.05% L-cysteine. The sample was homogenized using sterile borosilicate glass beads and passed through a lOOum filter. GF Swiss Webster mice were colonized by oral gavage of 200 ul fecal suspension. Excess fecal suspension was resuspended and stored at -80C in pre-reduced lx PBS + 0.05% L-cysteine + 15% glycerol. For administration of fecal filtrates, the fecal suspension was passed through a sterile 0.2 urn filter before colonization via oral gavage using 200 ul fecal filtrate.
[0163] 6-Hz Psychomotor Seizure Assay
[0164] 6-Hz psychomotor seizure assay testing was conducted following Samala et al., 2008. One drop (~50 ul) of 0.5% tetracaine hydrochloride ophthalmic solution was applied to the corneas of each mouse 15 min before stimulation. A thin layer of electrode gel (Parker Signagel) was applied directly to the comeal electrodes and was reapplied before each trial. A constant-current current device (ECT Unit 57800, Ugo Basile) was used to deliver current through the comeal electrodes at 3s duration, 0.2 ms pulse-width and 6 pulses / s frequency. CC50 (the milliamp intensity of current required to elicit seizures in 50% of the mouse cohort) was measured as a metric for seizure susceptibility. Pilot experiments were conducted to identify 28 mA as the CC50 for SPF wild-type Swiss Webster mice, aged 6-8 weeks. Each mouse was seizure-tested only once, and thus at least n > 14 mice were used to adequately power each cohort. To determine CC50s for each tested cohort, 28 mA of current was administered to the first mouse per cohort, followed by stepwise fixed increases or decreases by 2 mA intervals. Mice were restrained manually during stimulation and then released into a new cage for behavioral observation. Quantitative measures for falling, tail dorsiflexion (Straub tail), forelimb clonus, eye / vibrissae twitching and behavioral remission were scored manually. For each behavioral parameter, we observed no correlation between percentage incidence during 28+ mA seizures between pre-KD or post-KD microbiota status, suggesting a primary effect of the microbiota on seizure incidence rather than presentation or form. Latency to exploration (time elapsed from when an experimental mouse is released into the observation cage (after comeal stimulation) to its first lateral movement) was scored manually with an electronic timer. Mice were blindly scored as protected from seizures if they did not show seizure behavior and resumed normal exploratory behavior within 10 s. Seizure threshold (CC50) was determined as previously described (Kimball et al., 1957), using the average log interval of current steps per experimental group, where sample n is defined as the subset of animals displaying the less frequent seizure behavior. Data used to calculate CC50 are also displayed as latency to explore for each current intensity, where n represents the total number of biological replicates per group regardless of seizure outcome.
[0165] Antibiotic Treatment
[0166] Transplanted mice were gavaged with a solution of vancomycin (50 mg / kg), neomycin (100 mg / kg) and metronidazole (100 mg / kg) every 12 hours daily for 5 days, as adapted from (Reikvam et al., 2011). Ampicillin (1 mg / ml) was provided ad libitum in drinking water. For mock treatment, mice were gavaged with a similar volume of lx PBS (vehicle) water every 12 hours daily for 7 days. Antibiotic-treated mice were maintained in sterile caging with sterile food and water and handled aseptically for the remainder of the experiments.
[0167] Fecal and Serum Metabolomics
[0168] Previously collected human donor fecal samples were aliquoted as described in section “Human Donor Fecal Microbiota Transfer”. Mouse fecal samples were collected from mice housed across independent cages, with four cages housing 3 mice and one cage housing 2 mice. Mouse serum samples were collected by cardiac puncture and separated using SST vacutainer tubes, then frozen at -80C. Samples were prepared using the automated MicroLab STAR system (Hamilton Company) and analyzed on GC / MS, LC / MS and LC / MS / MS platforms by Metabolon, Inc. Protein fractions were removed by serial extractions with organic aqueous solvents, concentrated using a TurboVap system (Zymark) and vacuum dried. For LC / MS and LC-MS / MS, samples were reconstituted in acidic or basic LC- compatible solvents containing > 11 injection standards and run on a Waters ACQUITY UPLC and Thermo-Finnigan LTQ mass spectrometer, with a linear ion-trap frontend and a Fourier transform ion cyclotron resonance mass spectrometer back-end. For GC / MS, samples were derivatized under dried nitrogen using bistrimethyl-silyl-trifluoroacetamide and analyzed on a Thermo-Finnigan Trace DSQ fast-scanning single-quadrupole mass spectrometer using electron impact ionization. Chemical entities were identified by comparison to metabolomic library entries of purified standards. Following log transformation and imputation with minimum observed values for each compound, post-KD vs. pre-KD comparisons for human fecal, and mouse serum and fecal data were analyzed by paired t-test. Metabolomic data from SPF or antibiotic-treated mice fed KD vs. CD chow were acquired from (Olson et al., 2018), as log transformed and imputed with minimum observed values for each compound. Data were analyzed using two-way ANOVA to test for group effects. P and q-values were calculated based on two-way ANOVA contrasts. Principal components analysis was used to visualize variance distributions. Supervised Random Forest analysis was conducted to identify metabolomics prediction accuracies. Metabolite set enrichment analysis (MSEA) using the Metaboanalyst 5.0 platform (Pang et al., 2021) was performed on human fecal, mouse fecal, and mouse serum metabolites statistically significantly altered in post-KD compared to pre-KD (non-adjusted p-value <0.05). Metabolite sets were analyzed for chemical sub-class enrichment and metabolite pathway enrichment, using The Small Molecule Pathway Database (SMPDB).
[0169] Transcriptomics
[0170] Recipient mice were sacrificed on day 4 post-colonization. Hippocampus and frontal cortex were dissected from pre-KD and post-KD recipient mice (n=6 per cohort) and immediately placed in Trizol. RNA was extracted using the PureLink RNA Mini kit with Turbo DNAse treatment. RNA was prepared using the TruSeq RNA Library Prep kit and 2 A~ 69-bp paired-end sequencing was performed using the Illumina HiSeq 4000 platform by the UCLA Neuroscience Genomics Core. FastQC v0.11.5, bbduk v35.92, and RSeQC v2.6.4 were used for quality filtering, trimming, and mapping. Reads were aligned to UCSC Genome Browser assembly ID: mmlO using STAR v2.5.2a, indexed using samtools vl.3, and aligned using HTSeq-count v0.6.0. Differential expression analysis was conducted using DESeq2 vl.24.041. Heatmaps were generated using the pheatmap vl.0.12 package for R. GO term enrichment analysis of differentially expressed genes with non-adjusted p-value < 0.05 was conducted using enrichR v3.1. Protein interaction networks were generated using STRING vlO.5. Functional enrichments of network nodes were categorized by GO: biological process, molecular function, and cellular component.
[0171] Multi-omics Integration
[0172] To assess the relationships across omics layers, we first carried out dimension reduction for each data set using weighted gene co-expression network analysis (WGCNA vl.72.1) (Langfelder & Horvath, 2008). Metabolomics for human donors and mouse recipients and RNA-seq for mouse recipients (hippocampus and frontal cortex) were used to build WGCNA modules within each dataset, where modules represent clusters of highly co- regulated / expressed molecules which are typically involved in similar biological functions. For metabolomics data, goodSamplesGenes function was first used with default parameters to filter out sparse metabolites across samples before constructing networks; this step was not used for RNAseq data. Standard WGCNA steps were then carried out for the filtered metabolomics and RNAseq data. Module eigengenes (MEs), or the first axis of principal component were calculated from each module. MEs were then targeted for correlation analysis with the metadata (pre-KD vs. post-KD and responder vs. non-responder). Modules that had significant correlation (p-val <0.05) with the metadata were chosen for subsequent integrative analysis.
[0173] A systematic network that combined all omics data was inferred based on the probability of co-occurrence (POC) between molecules from different omics data. To calculate POC, we leveraged a neural-net based tool called MMVEC vl.0.6 with default parameters (Morton et al., 2019). The subset of raw data that contains module components that were selected from WGCNA analysis were log normalized and combined based on sample ID. This combined data matrix was then used as input for MMVEC. For example, on donor side, modules from fecal metagenome and metabolomics were added together and, on the recipient side, the combined matrix contained the raw data from metagenome, metabolomics, and RNAseq. Due to high density of the overall network generated from MMVEC, the top 10% of POC connections were retrieved to minimize overall complexity of the network for both donors and recipients using in-house python script (https: / / github.com / smhal 18 / keto_diet_pediatric_epilepsy). The networks of modules from individual omics layers comprising of pre-KD (n=10) and post-KD (n=10) donor metagenome and metabolomic samples and recipient pre-KD (n=10, where each n reflects average of 5 technical replicate recipient mice per donor patient sample]) and post-KD (n=10, where each n reflects average of 5 technical replicate recipient mice per donor patient sample]) metagenome, pre-KD (n=10, where each sample is pooled from 5 recipient mice per donor patient sample) and post-KD (n=10, where each sample is pooled from 5 recipient mice per donor patient sample) metabolomics, and pre-KD (n=10, where each sample is pooled from 6 recipient mice per donor patient sample) and post-KD RNAseq (n=10, where each sample is pooled from 6 recipient mice per donor patient sample), as well as differentially expressed / abundance molecules were then seeded into Mergeomics v3.16 pipeline along with the integrated network generated with MMVEC for weighted key driver analysis (wKDA) to identify key drivers of the networks (Ding et al., 2021). wKDA uses a %2 -like statistic to identify molecules that are connected to significant larger module components than what would be expected by random chance. The analysis was done on the human and mouse networks separately. To further look into the network that are relevant to ketogenic diet and epilepsy, we selected key drivers (KDs) based on i) the number of modules that a key driver was invoked related to, ii) their relation to the Ketogenic diet or epilepsy. A subset of nodes in each module that were connected to the KDs were collected. These nodes were retrieved with following priority i) they are part of differentially regulated molecules ii) POC value with KDs. Finally, the network was visualized using Cytoscape (Shannon et al., 2003). To minimize overall density of the network, we chose to show the key drivers Mergeomics with the highest occurrence in their respective MEs and with > 5 degrees of connectivity.
[0174] Marker set enrichment analysis (MSEA) to connect hippocampus and frontal cortex DEGs with epilepsy GWAS
[0175] To assess the potential role of the DEGs from the hippocampus and frontal cortex in epilepsy, we collected the summary statistics of the latest epilepsy GWAS (Abou-Khalil et al., 2018). Single nucleotide polymorphisms (SNPs) that had a linkage disequilibrium of r2> 0.5 were filtered to remove redundancies. To map the epilepsy GWAS SNPs to genes, we used GTEx version 8 eQTL and sQTL data for brain hippocampus and brain frontal cortex (Aguet et al., 2020), which help us derive genes likely to be regulated by the SNPs. We next used the MSEA function of the Mergeomics package (Ding et al., 2021) to compare epilepsy disease association p-values of the SNPs representing the DEGs (hippocampus or frontal cortex) with those of the SNPs mapped to random genes to assess whether the DEGs contain SNPs that show stronger epilepsy association than random genes using a %2-like statistic.
[0176] Quantification And Statistical Analysis
[0177] Statistical analyses were conducted using Prism8 software v8.2.1 (GraphPad). Before statistical analysis, data was assessed for distribution to determine appropriate statistical tests to use. Data were plotted in figures as mean ± SEM. For figures: IB, 3B, 3F, 4B, and 4E n = the number of technical replicates. For all other figures, n = the number of biological replicates. No samples or animals were excluded from data analysis. Differences between two sample conditions from parametric data sets were analyzed using two-tailed, paired Student’s t-test. Differences between two sample conditions from nonparametric data sets were analyzed using two-tailed, Wilcoxon matched-pairs signed rank test. For differences among >2 groups when analyzing one variable, a one-way ANOVA with Tukey’s post hoc test was used. For differences among >2 groups with two variables, a two-way ANOVA with Sidak’s post hoc test was used. For technical replicates from within-patient analysis (Figures: IB, 3B, 3F, 4B, and 4E), differences from the above tests are denoted by:#p<0.05;##p<0.01;###p<0.001;####p<0.0001. For biological replicates (all other figures), differences from the above tests are denoted by: *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. Non-significant differences are denoted in the figures using “n.s”.
[0178] Incorporation by Reference
[0179] All publications, patents, and patent applications mentioned herein are hereby incorporated by reference in their entirety as if each individual publication, patent or patent application was specifically and individually indicated to be incorporated by reference. In case of conflict, the present application, including any definitions herein, will control.
[0180] Equivalents
[0181] Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the following claims.
Claims
We Claim:
1. A method of preventing or treating a condition responsive to a ketogenic diet in a subject, comprising administering to the subject an agent that modulates a gene product of any one of the genes listed in Table 1 and / or any one of the pathways listed in Table 2.
2. The method of claim 1, wherein the condition is seizures, optionally wherein subject has a neurodevelopmental condition, e.g., selected from epilepsy, autism spectrum disorder, Rett syndrome, attention deficit disorder, and fragile X syndrome.
3. The method of claim 1, wherein the subject has a condition selected from Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), cancer, stroke, a metabolic disease, a mitochondrial disorder, depression, migraines, and traumatic brain injury (TBI).
4. The method of any one of claims 1 to 3, wherein the gene is selected from one or more of Acaa2, Abcc9, Aldh3bl, Ap000350.4, Ap000350.5, Ap000350.6, Arfi, Arf6, Arfipl, Atf2, Atm, Atp5h, Atp5j2, Atp5o, Atp8a2, Aurkaipl, Bazlb, Bmprlb, Briplos, Calr, Caprinl, Carmil2, Ccdcl36, Ccl6, Ccnyll, Ccnq, Cchcrl, Cdh4, Cenpu, Clic5, Col8a2, Cox5b, Crybgl, Cwc22, Cxxld, Ddx3y, Def8, Diexf, DocklO, Did, Dmkn, Dnajcl8, Duspl2, Ehmt2, Ei24, Ecm2, Ecsit, Egfir, Eloa, Emidl, Eme2, Ep300, Epn3, Epsl5, Etfl, Faml07a, Faml85a, Fbxo8, Fcerlg, Fgfrlop, Flotl, F5, Gabarapll, Gas5, Gm527, Gmcll, Gpx2, Grebl, Gstol, Gstt2, Gstpl, Gtpbp6, Hars, Hif3a, Hnmpab, Hscb, Hylsl, Iqsec3, Ino80b, Ingl, Irf2, Jmy, Kctd20, Kdm5d, Kdm6b, Lampl, Lhfpl4, Lrrcl, Lrrtml, Lonp2, Loxl3, Mif4gd, Mcf2, Mettll8, Mettl23, Mnsl, MrpllO, Mrpll8, Mrpl30, Mrpl42, Mrps24, Mybll, MyhlO, Myb, Ndorl, Ndufal2, Ndufb7, Ndufb9, Ndufv2, Nebl, Nfe211, Nfya, Nfxl, Nrlh2, Nr4a3, Nrbpl, Nrg2, Ntpcr, Nubl, Nupl53, Nuprl, Nusapl, Odf21, Ogfod3, 01fm3, Oligl, Pcedla, Pdgfa, Pdgfb, Pedsl, Pesl, Phfl lb, Piercel, Phip, Ppargcla, Ppil6, Pole3, Pramell2, Prkce, Proml, Psd, Qtrtl, Rabl2, RanbplO, Rbbp4, Rbbp8, Rbpj, Rere, Rgs8, Rnfl4, Rnfl38, Rprd2, Roml, Runxltl, Rwdd3, SlOOalO, Scaf4, Serbpl, Serpinbla, Setmar, Shrooml, Slc5a3, Slc6al2, Slitl, Smiml9, Smim24, Tbx2, Tbx3osl, Tbxl8, Tceb3, Tfblm, Tmem69, Tmeml l4, Tmeml84b, Tmem237, Ttil, Tpbg,Tpm3, Txnl, Ubash3b, Ufspl, Usp6nl, Uty, Utpl4a, Utpl4b, Vars2, Wdr6, Zc3havl, Zcchc2, Zfpl42, Zfp383, Zfp397, Zfp87, and Zkscanl4.
5. The method of any one of claims 1 to 4, wherein the pathway is selected from one or more of Formaldehyde assimilation; Guanosine nucleotide degradation; Amino acid biosynthesis; L-proline biosynthesis; proline biosynthesis; Sugar acid biosynthesis; GDP- mannose biosynthesis; monosaccharide biosynthesis; 2-methylcitrate cycle; methylcitrate cycle I; methylcitrate cycle II; Glycol degradation; glycol metabolism; Fatty acid beta-oxidation; fatty acid oxidation; fatty acid biosynthesis; hydroxy fatty acid biosynthesis; Methionine metabolism; Beta-hydroxybutyrate (BHBA) metabolism; N-substituted nicotinamide metabolism; Alphalinolenic acid and linoleic acid metabolism; Phenylpropanoic acid biosynthesis; Glycine and serine metabolism; Glutathione metabolism; Betaine metabolism; Adenine and adenosine salvage; dTDP-L-rhamnose biosynthesis; Folate transformations; L-valine biosynthesis; sulfoglycolysis; Acetyl-CoA fermentation; Pyruvate fermentation; Branched-chain amino acid biosynthesis; Valine, leucine, and isoleucine biosynthesis; Phenylalanine, tyrosine, and tryptophan biosynthesis; Glycerophospholipid metabolism; Aerobactin biosynthesis; Sidophore biosynthesis; Polyamine biosynthesis; Monosaccharide metabolism; Biotin biosynthesis; Fatty acid metabolism; Hydroxy fatty acid metabolism; and Phenylpropanoic metabolism.
6. The method of any one of claims 1 to 5, wherein the agent increases the activity or expression of a gene product of any one of the genes listed in Table 1 and / or any one of the pathways listed in Table 2.
7. The method of any one of claims 1 to 6, wherein the agent is a potentiator listed in Table 1.
8. The method of claim 6 or 7, wherein the agent is selected from one or more of l,l,l,3,3,3-hexafluoro-2-{4-[(2,2,2-trifluoroethyl)amino]phenyl}propan-2-ol; 3-Hydroxymethyl- 5-aziridinyl-lmethyl-2-[lH-indole-4,7-dione]-propanol; 3-(hydroxymethyl)-l-methyl-5-(2- methylaziridin-l-yl)-2-phenyl-lH-indole-4, 7-dione; 5'-Guanosine-diphosphate- monothiophosphate; 5-Methoxy-l,2-dimethyl-3-(4-nitrophenoxymethyl)indole-4, 7-dione;Adenine; Benzylpenicilloyl Polylysine; Benzenesulfonyl; Bishydroxy[2H-l-benzopyran-2- one,l,2-benzopyrone]; Canfosfamide; Cibacron Blue; Colforsin; Dicoumarol; Glutathione; Glutathione disulfide; Guanosine-5'-diphosphateguanosine-5'-diphosphate; Hyodeoxycholic acid; Lauryl dimethylamine-N-oxide; L-histidine; Melatonin; Menadione; Myristic acid; N- (2,2,2-trifluoroethyl)-N-{4-[2,2,2-trifluoro-l-hydroxy-l- (trifluoromethyl)ethyl]phenyl}benzenesulfonamide; N-formylmethionine; NADH; NAV-2729; p-Nitrophenol; Pseudoephedrine; Rosiglitazon; Succinic acid; S-(4-bromobenzyl)cysteine; S- hexylglutathione; Ubiquinone-1; and Vitamin C.
9. The method of any one of claims 1 to 5, wherein the agent decreases the activity or expression of a gene product of any one of the genes listed in Table 1 or any one of the pathways listed in Table 2.
10. The method of any one of claims 1 to 5 and 9, wherein the agent is an inhibitor listed in Table 1.
11. The method of claim 9 or 10, wherein the agent is selected from one or more of adenosine triphosphate, Afatinib, Axitinib, alpha-Tocopherol, alpha-Tocopherol acetate, alpha- Tocopherol succinate, Amitriptyline, Anisomycin, Benzoyl peroxide, Brigatinib, Caffeine, Cannabidiol, Cholic Acid, Cladribine, Clomipramine, Curcumin, Dacomitinib, Deoxycholic acid, Dequalinium, Diazoxide, DL-alpha-Tocopherol, Drotrecogin alfa, Erlotinib, Etacrynic acid, Ezatiostat hydrochloride, Fostamatinib, Garcinol, Gefitinib, glyburide, Guvacine, Hexachlorophene, Lapatinib, Lidocaine, Midostaurin, Metformin, Omacetaxine mepesuccinate, Osimertinib, Puromycin, Racephedrine, Rhein, Roxadustat, Sgc-cbp30, Sunitinib, Tamoxifen, Trimetazidine, Vandetanib, and Zanubrutinib.
12. The method of any one of claims 1 to 11, wherein the subject is on a diet, and the diet is a control diet.
13. The method of any one of claims 1 to 11, wherein the subject is on a diet, and the diet is a ketogenic diet.
14. The method of any one of claims 1 to 11, wherein the subject is on a diet, and the diet is a high fat diet.
15. The method of any one of claims 1 to 11, wherein the subject is on a diet, and the diet is a low carbohydrate diet.
16. The method of any one of claims 1 to 15, wherein the agent is formulated for oral delivery.
17. The method of any one of claims 1 to 16, wherein the agent is self-administered.
18. The method of any one of claims 1 to 17, wherein the subject is a pediatric subject.
19. The method of any one of claims 1 to 18, where the subject is an adult.
Citation Information
Patent Citations
Compositions and methods for modulating cognitive behavior
US20220218762A1
Methods, compositions, and kits for treating and preventing neurological conditions
WO2012149472A2
