Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Rare genetic variants in the endocannabinoid system genes CNR1 and DAGLA are associated with neurological phenotypes in humans

  • Douglas R. Smith ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – original draft

    drclsmith@comcast.net

    Affiliation Courtagen Life Sciences, Inc., Woburn, MA, United States of America

  • Christine M. Stanley,

    Roles Formal analysis, Validation, Writing – review & editing

    Current address: Wuxi Nextcode, Cambridge, MA, United States of America

    Affiliation Courtagen Life Sciences, Inc., Woburn, MA, United States of America

  • Theodore Foss,

    Roles Software, Validation, Writing – review & editing

    Current address: Wuxi Nextcode, Cambridge, MA, United States of America

    Affiliation Courtagen Life Sciences, Inc., Woburn, MA, United States of America

  • Richard G. Boles,

    Roles Formal analysis, Methodology, Writing – review & editing

    Current address: Mitochondrial and Molecular Medicine, Pasadena, CA, United States of America

    Affiliation Courtagen Life Sciences, Inc., Woburn, MA, United States of America

  • Kevin McKernan

    Roles Funding acquisition, Resources, Writing – review & editing

    Current address: Medicinal Genomics Corporation, Woburn, MA, United States of America

    Affiliation Courtagen Life Sciences, Inc., Woburn, MA, United States of America

Abstract

Rare genetic variants in the core endocannabinoid system genes CNR1, CNR2, DAGLA, MGLL and FAAH were identified in molecular testing data from 6,032 patients with a broad spectrum of neurological disorders. The variants were evaluated for association with phenotypes similar to those observed in the orthologous gene knockouts in mice. Heterozygous rare coding variants in CNR1, which encodes the type 1 cannabinoid receptor (CB1), were found to be significantly associated with pain sensitivity (especially migraine), sleep and memory disorders—alone or in combination with anxiety—compared to a set of controls without such CNR1 variants. Similarly, heterozygous rare variants in DAGLA, which encodes diacylglycerol lipase alpha, were found to be significantly associated with seizures and neurodevelopmental disorders, including autism and abnormalities of brain morphology, compared to controls. Rare variants in MGLL, FAAH and CNR2 were not associated with any neurological phenotypes in the patients tested. Diacylglycerol lipase alpha synthesizes the endocannabinoid 2-AG in the brain, which interacts with CB1 receptors. The phenotypes associated with rare CNR1 variants are reminiscent of those implicated in the theory of clinical endocannabinoid deficiency syndrome. The severe phenotypes associated with rare DAGLA variants underscore the critical role of rapid 2-AG synthesis and the endocannabinoid system in regulating neurological function and development. Mapping of the variants to the 3D structure of the type 1 cannabinoid receptor, or primary structure of diacylglycerol lipase alpha, reveals clustering of variants in certain structural regions and is consistent with impacts to function.

Introduction

The endocannabinoid system (ECS) plays an important role in the regulation of neurological activity throughout the central and peripheral nervous system [13], as well as in the regulation of cell division, metabolic, and immune processes in a variety of other tissues [46]. The type 1 cannabinoid receptor (CB1), encoded by the CNR1 gene, is a key component of the ECS. CB1 is the most abundant G-protein coupled receptor in the brain, and is present at high levels in the neocortex, hippocampus and cerebellum [7]. CB1 is activated by the natural endocannabinoid agonists N-arachidonoylethanolamine (AEA), 2-arachidonoyl glycerol (2-AG) and a variety of related compounds [8]. It also binds the phytocannabinoid Δ9-tetrahydrocannabinol (THC), and a wide variety of synthetic agonists and antagonists [9]. At the cellular level, CB1 is located primarily on presynaptic termini of GABAergic and glutamatergic neurons in the brain [10, 11], where it binds 2-AG released by postsynaptic termini to down-regulate neurotransmitter release [12].

The type II cannabinoid receptor (CB2), encoded by the CNR2 gene, is expressed in cells of the immune system (recently reviewed by Turcotte et al [13]) and is strongly induced in activated microglia in the brain [14]. CB2 is activated by the endocannabinoids AEA and 2-AG, and by the phytocannabinoids THC and β‒caryophyllene [13, 15].

2-AG is the most abundant endocannabinoid in the brain, and is generated through the action of diacylglycerol lipase α (encoded by the DAGLA gene) by hydrolysis of diacylglycerol [16]. 2-AG is a mediator of retrograde signaling to presynaptic CB1 receptors to regulate neurotransmitter release [12]. 2-AG acts over longer distances during early development due to low monoacylglycerol lipase levels to regulate neuronal development at axonal growth cones [17, 18]. 2-AG is degraded by the action of monoacylglycerol lipase (encoded by the MGLL gene). AEA is generated mainly through the activity of N-acyl phosphatidylethanolamine phospholipase D, encoded by the NAPEPLD gene, and is degraded mainly by fatty acid amide hydrolase, encoded by the FAAH gene [19, 20].

Common single nucleotide polymorphisms in or near CNR1, CNR2, FAAH and MGLL have been reported to be associated with a variety of clinical phenotypes in candidate gene association studies (substance abuse disorders, cardiovascular disease risk factors, irritable bowel syndrome, migraine, chronic pain and mood disorders) [2130]. The effect sizes are generally small, however, and replication studies in larger independent cohorts have been met with mixed results [3135]. In contrast, the impact of rare genetic variation in genes associated with the endocannabinoid system has not been studied systematically. There are several reports of pathogenic deletions and duplications involving the ECS genes CNR1, CNR2, DAGLA, MGLL and FAAH with associated developmental phenotypes in the Decipher [36] and Clinvar [37] databases. There is also a reported association of DAGLA duplications with spinocerebellar ataxia-20 in OMIM [38] (entry 608687). However, the size of those structural variants is generally large (24 Mb on average) and they all impact multiple genes.

In this study, we investigate the phenotypic impact of rare missense variants in the core ECS genes: CNR1, CNR2, DAGLA, MGLL and FAAH, which encode CB1, the type 2 cannabinoid receptor (CB2), diacylglycerol lipase alpha, monoglyceride lipase, and fatty acid amide hydrolase, respectively. Phenotypes for were selected for evaluation based on their presence in mouse knockout strains for each gene.

Methods

Two diagnostic gene panels were used in this study (S1 Table). These panels were designed to identify chromosomal alterations associated mitochondrial disorders (NucSEEK Comprehensive– 1207 genes including CNR1, MGLL and FAAH) and epilepsy (EpiSEEK Comprehensive– 471 genes including CNR1, CNR2, DAGLA, MGLL and FAAH). Genetic sequencing data was generated from over 6,000 individuals using one or both of those gene panels in our CLIA/CAP certified laboratory, and clinical diagnostic reports were created, as follows.

DNA was extracted from saliva or blood samples using Genfind v2 (Beckman Coulter). Sequence-ready libraries were prepared using HaloPlex® target capture kits (Agilent Technologies) in conjunction with DREAM PCR amplification [39, 40] to avoid sample contamination. Sequencing was performed on the Illumina MiSeq platform to generate paired 250 base reads at average coverage levels of 465 for NucSEEK and 565 for EpiSEEK. Sequence reads were trimmed to remove adapters and low quality bases using Cutadapt and FastQC before alignment to the hg19 (GRCh37) reference sequence using BWA-MEM [41]. Qualimap [42] and Picard were used to generate alignment QC metrics such as insert size distribution, mismatch rates, and GC bias. Variant calls were generated by an ensemble [43] approach using FreeBayes, Platypus [44] and the GATK [45] UnifiedGenotyper [46]. For annotation and clinical review, a custom platform integrating data from multiple tools and databases was used (Ensembl variant effect predictor [47], SnpEff [48], Human Phenotype Ontology (HPO) [49], OMIM [38], ClinVar [50], 1000 genomes [51], ExAc [52], and dbNSFP [53]). SmartPCA [54] was used to stratify patients into subgroups based on analysis of the variants with respect to 1000 Genomes data. These data were stored in a database and reviewed in the context of clinical phenotype data provided by the patient’s physician by Variant Scientists, Genetic Counselors, and Laboratory Directors. A comprehensive, user-friendly report was then generated for the patient’s physician including pathogenicity scores generated according to ACMG guidelines [55].

Retrospective, aggregate analysis of variants in the core ECS genes for this study utilized only the pre-existing clinical information provided at the time of testing. Individuals were excluded if they did not sign a consent form for research use of their clinical testing data, or if the available clinical information was absent or inadequate. Adequate clinical information included an informative clinical summary, a comprehensive evaluation, or a completed checklist of standardized terms by a treating physician. Clinical summaries were translated to HPO terms by genetic counselors using the PhenoTips [56] software tool. A limited data set, containing only the above-described information and no personal identifiers, was used for the present study. Families were not contacted in regards to the current study. Thus, per the Office of Human Research Protection of the U.S. National Institutes of Health, this study does not qualify as human subjects research, and informed consent is unnecessary. However, the study was approved by the Courtagen Life Sciences ethics committee.

Rare variants in CNR1, CNR2, DAGLA, MGLL and FAAH having an allele frequency of approximately 0.003 or less in the ExAC database were identified. A Pubmed search of the literature relating to mouse knockouts of each of the five genes was carried out, and a list of the phenotypes described in those studies was compiled (S2 Table). The mouse phenotypes were translated into general terms that could be related to human phenotypes in our test database (S2 Table). Comprehensive lists of HPO terms were generated from a query of all scored tests in our database in which the five ECS genes were present (S3 and S4 Tables). The terms were evaluated to generate subsets of HPO and text terms corresponding to each general phenotype for use as database queries (S3 and S4 Tables). Counts of subjects with individual or multiple neurological phenotypes were generated through text-based queries of summary clinical phenotype data associated with de-identified reports and stored in a database. The phenotypes were scored positive if any of the included terms matched one or more of the clinical terms provided by the patient’s physician. Non-neurological phenotypes were not evaluated because they were poorly represented in the clinical database, as almost all of the patients were being tested for neurological disorders.

A set of non-carrier controls were selected for each gene. The total number of non-carriers was 3,777 for DAGLA and CNR2, and 5,979 for CNR1, FAAH and MGLL. To simplify the analysis, 950 controls were selected at random for CNR1 and DAGLA, and ~500 controls were selected at random for CNR2, FAAH and MGLL. If the rare variant data for a particular gene was derived from two gene panels (this was the case for CNR1, FAAH, and MGLL), a proportional number of controls were selected from subjects sequenced using each of those panels. Counts of subjects with individual or multiple phenotypes were generated through text-based queries of summary clinical phenotype data as described above. Two-tailed Fisher exact tests were performed to evaluate the possible association of phenotypes present in carriers with rare variants, compared to controls, using a tool available at vassarstats.net (clinical research calculator number 3). The resulting P-values were Bonferroni corrected by multiplying by the number of general phenotypes tested (corrected P-values >1 were scored as 1). No correction was made for the number of genes tested, since they are all from a single pathway.

Results

The numbers of subjects harboring rare variants were 22 for CNR1, 11 for CNR2, 35 for DAGLA, 34 for MGLL and 53 for FAAH. Data on the rare variants detected, including genomic location, coding consequences, allele frequencies in the ExAC, GnomAD and CLS databases, Polyphen2 and SIFT scores, PhyloP conservation and ExAC gene constraint values are summarized in S5 Table. The human clinical phenotypes and HPO terms associated with each of the mouse knockout phenotypic categories detected are also presented in S5 Table.

The results of Fisher exact tests for phenotypes putatively associated rare variants in the genes CNR2, MGLL and FAAH did not reveal any significant association between rare variants and any phenotype or combination tested (S6 Table). The results for CNR1 and DAGLA, however, revealed several significant associations between the presence of heterozygous rare coding variants and neurological phenotypes as indicated in Table 1.

thumbnail
Table 1. Association of mouse knockout phenotypes with rare variants in CNR1 and DAGLA.

https://doi.org/10.1371/journal.pone.0187926.t001

The allele frequencies of the ECS gene variants in our database of genetic testing results (CLS database) were compared with those in the public reference databases ExAC and gnomAD [57]. Across all five genes, the rare variants had an average allele frequency of 0.0004 in ExAC and gnomAD. The rare variants were enriched in the CLS database compared to gnomAD by 4.5-fold for CNR1 and 6.7-fold for DAGLA (not including variants with zero frequency gnomAD). Eight of the 24 CNR1 variants, and three of the 36 DAGLA variants had zero frequency in gnomAD. The fraction of variants with zero frequency in gnomAD for the other genes was 1/10 for CNR2, 8/53 for FAAH and 3/40 for MGLL.

We looked for shared phenotypes amongst subjects carrying the same rare variant. Two of the CNR1 variants were observed in multiple samples (p.Glu93Lys in 2 samples, and p.Ala419Glu in 5 samples)—S5 Table. All subjects with p.Glu93Lys had seizures and developmental delay. Two of the subjects with p.Ala419Glu had the shared phenotypes: anxiety, sleep disorder, abnormality of the autonomic nervous system and a third had a sleep disorder combined with seizures. Those phenotypes were absent in the other subjects with p.Ala419Glu.

Four of the DAGLA variants were observed in multiple samples (p.His810Gln in 7 samples, p.Arg815His in 5 samples, p.Ala858Val in 2 samples, and p.Leu988Val in 3 samples)—S5 Table. Three of the subjects with p.His810Gln had infantile spasms and two had focal seizures. 5/6 of the subjects with p.Arg815His had seizures and developmental disorders and two had a brain abnormality. The subjects with p.Ala858Val did not have shared phenotypes. All three subjects with p.Leu988Val had seizures; two had developmental delay and the third had a brain abnormality in addition. One subject (ECS158) had two DAGLA variants, p.Arg197Trp and p.Ter1043GlnextTer29; the chromosomal phasing of those variants is unknown.

Additional genes with variants that could contribute to the observed phenotypes in the subjects, based on clinical reports, are noted in S5 Table.

Several 3D x-ray crystal structures of CB1 were published recently, by two groups [5860]. We mapped the position of the rare coding variants onto one of those structures with bound agonist, PDB ID: 5XRA, as well as onto the primary structure and domain organization of the enzyme (Fig 1). Of the variants that could be mapped, 6/7 were in the 7-transmembrane cannabinoid binding domain (Fig 1). The other CNR1 variants, located near the N- or C-terminus, could not be mapped since those regions were deleted from the crystalized proteins. A similar exercise was performed for DAGLA using the primary structure and domain organization of the enzyme as described by Reisenberg, et al. [61] and NCBI Reference Sequence: NP_006124.1, since a crystal structure is not available. Most of the DAGLA variants mapped to the C-terminal end of the protein, a region that contains many regulatory phosphorylation sites (Fig 2).

thumbnail
Fig 1.

A. Depiction of the x-ray crystal structure of the type I cannabinoid receptor with bound agonist AM11542 from PDB ID 5XRA. The location of rare coding variants that could be mapped to the structure are shown by magenta dots, with annotations indicating the amino acid substitutions and the cannabinoid binding site. B. The location of rare coding variants relative to domains of the enzyme and regulatory phosphorylation sites are indicated, based on NCBI Reference Sequence: NP_001153731.1. Red dot: rare variant; T (light green shading): transmembrane domain; P (blue vertical lines): phosphorylation sites.

https://doi.org/10.1371/journal.pone.0187926.g001

thumbnail
Fig 2. Mapping of rare variants to diacylglycerol lipase alpha.

The location of rare coding variants relative to domains of the enzyme and regulatory phosphorylation sites are indicated, based on NCBI Reference Sequence: NP_006124.1. Red dot: rare variant; purple dots: two variants present in one subject; T (light green shading): transmembrane domain; light blue shading: lipase domain; A (purple shading): Active site; N (magenta shading): Nucleophilic elbow; blue vertical lines: phosphorylation sites.

https://doi.org/10.1371/journal.pone.0187926.g002

Discussion

We report a set of 155 rare missense or short deletion variants in the core ECS genes CNR1, CNR2, DAGLA, FAAH and MGLL. A subset of those variants, in CNR1 and DAGLA, were found to be statistically associated with neurological phenotypes by case control analysis at the gene level. It is perhaps not surprising that CNR2, FAAH and MGLL were not associated with any phenotypes. Those genes are significantly more tolerant to missense and loss of function (LOF) variation than CNR1 and DAGLA, based on ExAC gene constraint values [57, 62]. Furthermore, our testing focuses on neurological phenotypes, so metabolic or immunological phenotypes associated with those genes may not have been reported.

The CNR1 variants were found to be associated with headache (including migraine), sleep and memory disorders, alone or in combination with anxiety, compared to a control set of 950 randomly selected patients without such CNR1 variants. The Fisher odds ratios for the associated variants ranged from 4.5 up to 150, with the highest ratios being associated with sleep disorder plus headache (18.9), sleep disorder plus anxiety (26.2), memory disorder with or without headache (150) and the combined phenotypes of headache, memory disorder and anxiety and/or sleep disorder, which did not occur in any of the controls (Table 1).

Based on ExAC gene constraint values [57, 62], the CNR1 gene is predicted to be somewhat tolerant of single missense and loss of function variants (missense z score = 2.7 and pLI = 0.15), but intolerant of dual heterozygous LOF variants (pRec = 0.78). All but one of the variants that could be mapped to the 3D structure of CB1 were in the 7-transmembrane domain. One additional variant, Ile339Val, would have mapped to the cytoplasmic domain if the region containing it had not been substituted in the 5XRA sequence. While we have no direct evidence that the variants we discovered impair function of the receptor, we note that three of the variants, p.Gly194Ala, p.Val249Met and p.Thr242Ile, map near the cannabinoid binding pocket in the 3-dimensional structure (Fig 1). In addition, two other variants, p.Arg148His and p.Val306Ile, map near the boundary between the 7-transmembrane and cytoplasmic signaling domains. Such variants could affect signal transduction of the receptor. There were four variants (at residues 7, 27,57 and 93) that could not be mapped since the N-terminus was deleted from the structure. The N-terminal region is not required for cannabinoid binding, but may play a role in regulating the stability and surface expression of CB1[63]. We also note that the PolyPhen 2 deleteriousness scores of the CNR1 variants are generally high (mean PP2 score = 0.85 with 11/24 ≥ 0.95).

Mice that are experimentally manipulated to harbor a non-functional CNR1 gene (CB1 knockout), or that are treated with CB1 antagonists, display behaviors that suggest impacts related to pain sensitivity, anxiety, memory and sleep, amongst others (S2 Table). The behaviors include elevated avoidance, freezing and risk-assessment behaviors [64], accelerated early learning and memory decline [65], food-seeking, and non-REM sleep alterations [66, 67]. There is a strong correlation between those phenotypes and those we observed in the present study (anxiety, sleep and memory disorders). Three of the eight subjects with rare CNR1 variants and headaches in our study also have memory disorders (mainly short term memory), and two of those have comorbid anxiety and sleep disorders.

Two of the subjects with migraine headaches harbor variants in other genes that have previously been associated with migraine (TRAP1 and CACNA1A) [68, 69]. However, the clinical correlation of the variants present in those subjects (TRAP1 p.Ile253Val in subject ECS031 and CACNA1A p.Ser2423_Gly2424del in subject ECS100) with migraine and anxiety is uncertain. Interestingly, 2/5 subjects with a CNR1 p.Ala419Glu variant (ExAC MAF = 0.00038) had the shared phenotypes: anxiety, sleep disorder, and abnormality of the autonomic nervous system. Only one of the other three subjects with that variant, however, had any of those phenotypes. These observations suggest that additional factors (genetic or environmental) may contribute to expression of the CNR1-associated phenotypes. Additional work will be required to resolve this.

The CB1 receptor encoded by CNR1 binds the natural endocannabinoid agonists AEA and 2-AG. A theory of Clinical Endocannabinoid Deficiency (CED), first described by Russo, posits that migraine, fibromyalgia, irritable bowel syndrome and comorbid anxiety may be related manifestations of reduced endocannabinoid tone, specifically AEA and 2-AG levels [70, 71]. These conditions frequently occur together, and a growing body of evidence suggests that they respond well to cannabinoid therapy, particularly THC [71]. Our results suggest that impaired CB1 signaling may be associated with increased susceptibility to a related set of clinical phenotypes: migraine, sleep and memory disorders with co-morbid anxiety. It would be interesting to explore whether migraine, anxiety and sleep disorders in patients with rare deleterious variants in CNR1 could be alleviated by treatment with CB1 agonists that could effectively stimulate the impaired and/or remaining functional CB1 receptors. Compounds that increase endocannabinoid levels such as CBD [72], or inhibitors of fatty acid amyl hydrolase, monoacylglycerol lipase, or fatty acid binding proteins [73, 74], might also be effective. We note that a recent observational study of 121 adults with migraine reported a decrease in mean headache frequency from 10.4 to 4.6 headaches per month (p<0.0001) for patients receiving medicinal cannabis [75]. However, the genetic basis (if any) of migraine in those patients was not investigated, and the therapeutic effect of THC could be a result of other mechanisms [75]. It is not clear whether high levels of THC could effectively remediate defective CB1 receptors because of its partial agonist nature and potential antagonistic effects on AEA and 2-AG signaling. Furthermore, the effects of THC may be different from AEA and 2-AG due to biased agonism [76, 77].

DAGLA encodes diacylglycerol lipase alpha, which is the major enzyme involved in 2-AG biosynthesis in the central nervous system. 2-AG levels have been reported to drop by ~80% in DAGLA knockout mice, resulting in pronounced anxiety and depression-like behaviors [78], and increased severity of kainate-induced seizures[79]. Developmental alterations observed in DAGLA knockout mice include altered cholinergic innervation of CA1 pyramidal cells of the hippocampus [18], and compromised adult neurogenesis in the hippocampus and subventricular zone [80]. In addition, altered axon formation in the midbrain-hindbrain region (associated with vision and locomotion) has been observed in response to transient morpholino DAGLA knockdown in zebrafish [81].

Heterozygous rare variants in DAGLA were found to be significantly associated with seizures, developmental disorders and abnormalities of brain morphology compared to controls. Based on ExAC gene constraint values, the DAGLA gene is predicted to be intolerant of missense and LOF variants (missense z score = 5.3 and pLI = 0.95 [57, 62]. We observed that 15/23 of the unique variants observed in our study map to the C-terminal third of the DAGLA coding sequence, a region that is predicted to contain numerous regulatory phosphorylation sites based on aggregate data from several biochemical studies [61]. This region appears to be more tolerant of variation than the catalytic domain, in which variation was sparse, despite its similar size. In contrast to mouse knockout models, anxiety phenotypes appear to be depleted in the clinical subjects—but the result was not significant. This raises the possibility that some of the variants might augment diacylglycerol lipase function. Further studies will be required to evaluate that possibility.

The frequency of seizures and developmental disorders in subjects with rare DAGLA variants were each elevated by a factor of approximately 4 compared to controls. The developmental disorders observed included several types of brain abnormalities: abnormalities of the cerebral white matter, brain malformations, posterior delayed myelination, heterotopia, and porencephaly with encephalomalacia. The prevalence of developmental disorders and brain abnormalities in these subjects aligns with the observations in animal models. Only two of the 35 samples with rare DAGLA variants harbored a likely pathogenic variant in another gene known to be associated with seizures and developmental delay (GRIN2A in subject 061 and MEF2C in subject 044). This represents about half the number of positive tests that would be expected based on the overall positive rate of the epiSEEK panel in our patient population (13%). Further work will be necessary to define whether rare variants in DAGLA are causatively associated with seizures and developmental disorders.

While the variants described in this study occur at extremely low frequencies in the general population, we suspect that patient populations with disorders of the nature studied here will be enriched for them. The allele frequencies observed in our test database for the rare CNR1 and DAGLA variants were approximately 4.5-fold higher, on average, than those observed in the ExAC database. It will be interesting to evaluate what fraction of patients with migraine, with or without anxiety, sleep, or memory disorders, harbor low frequency or rare variants in CNR1. Similarly, the prevalence of rare DAGLA variants in patients with developmental disorders should be investigated.

Supporting information

S1 Table. Column 1 lists the genes contained in the EpiSEEK gene panel.

Column 2 lists the genes contained in the NucSEEK gene panel.

https://doi.org/10.1371/journal.pone.0187926.s001

(XLSX)

S2 Table. The phenotypes associated with knockout mice are listed under each gene in column 1.

Column 2 gives the general phenotype that was used to select human phenotpes and HPO terms. Column 3 indicates whether the phenotypes are neurological. Column 4 provides PMIDs for corresponding mouse knockout publications.

https://doi.org/10.1371/journal.pone.0187926.s002

(XLSX)

S3 Table. Column 1 provides a comprehensive listing of HPO terms derived from clinical reports for 5464 patients tested using Courtagen's NucSEEK and EpiSEEK gene panels.

Column 2 provides the corresponding HPO descriptions. Columns 3–13 indicate the HPO terms corresponding to each of the general phenotypes derived from mouse knockouts according to the headers in row 3 (a "1" indicates that the term is included). Row 1 provides the strings of query terms used to evaluate the presence of general phenotypes in clinical reports of carrier and control subjects. Row 2 provides the frequencies of patients with the corresponding phenotypes amongst the 5464 patients.

https://doi.org/10.1371/journal.pone.0187926.s003

(XLSX)

S4 Table. Column 1 provides a comprehensive listing of HPO terms derived from clinical reports for 3728 patients tested using Courtagen's EpiSEEK gene panel.

Column 2 provides the corresponding HPO descriptions. Columns 3–8 indicate the HPO terms corresponding to each of the general phenotypes derived from mouse knockouts according to the headers in row 3 (a "1" indicates that the term is included). Row 1 provides the strings of query terms used to evaluate the presence of general phenotypes in clinical reports of carrier and control subjects. Row 2 provides the frequencies of patients with the corresponding phenotypes amongst the 5464 patients.

https://doi.org/10.1371/journal.pone.0187926.s004

(XLSX)

S5 Table. Data for rare variants discovered in this study.

Headers in row 1 are self explanatory.

https://doi.org/10.1371/journal.pone.0187926.s005

(XLSX)

S6 Table. Association data for CNR2, FAAH and MGLL with column headers: Gene, Phenotypes that were evaluated, Number of cases or controls in which that phenotype was present out of the total number, odds ratio (OR), 95% confidence intervals for the odds ratios, uncorrected two-tailed P-values, and P-value corrected for multiple testing.

https://doi.org/10.1371/journal.pone.0187926.s006

(XLSX)

S7 Table. This contains the same data as Table 1, with the addition of 95% confidence intervals for the odds ratios and the uncorrected two-tailed P-values.

Columns identical to Table 1: Gene, Phenotypes that were evaluated, Number of cases or controls in which that phenotype was present out of the total number, odds ratio (OR), and P-value corrected for multiple testing.

https://doi.org/10.1371/journal.pone.0187926.s007

(XLSX)

Acknowledgments

This study was supported by Courtagen Life Sciences, Inc. We thank Josee Dupuis for advice on the statistical analysis, and Ethan Russo and Anne Conyers-Hom for helpful comments on the manuscript.

References

  1. 1. Iannotti FA, Di Marzo V, Petrosino S. Endocannabinoids and endocannabinoid-related mediators: Targets, metabolism and role in neurological disorders. Prog Lipid Res. 2016;62:107–28. pmid:26965148
  2. 2. Lowin T, Straub RH. Cannabinoid-based drugs targeting CB1 and TRPV1, the sympathetic nervous system, and arthritis. Arthritis Res Ther. 2015;17:226. pmid:26343051
  3. 3. Maldonado R, Banos JE, Cabanero D. The endocannabinoid system and neuropathic pain. Pain. 2016;157 Suppl 1:S23–32.
  4. 4. Silvestri C, Di Marzo V. The endocannabinoid system in energy homeostasis and the etiopathology of metabolic disorders. Cell Metab. 2013;17(4):475–90. pmid:23562074
  5. 5. Cabral GA, Ferreira GA, Jamerson MJ. Endocannabinoids and the Immune System in Health and Disease. Handb Exp Pharmacol. 2015;231:185–211. pmid:26408161
  6. 6. Pyszniak M, Tabarkiewicz J, Luszczki JJ. Endocannabinoid system as a regulator of tumor cell malignancy—biological pathways and clinical significance. Onco Targets Ther. 2016;9:4323–36. pmid:27486335
  7. 7. Hu SS, Mackie K. Distribution of the Endocannabinoid System in the Central Nervous System. Handb Exp Pharmacol. 2015;231:59–93. pmid:26408158
  8. 8. Pertwee RG. Endocannabinoids and Their Pharmacological Actions. Handb Exp Pharmacol. 2015;231:1–37. pmid:26408156
  9. 9. Bow EW, Rimoldi JM. The Structure-Function Relationships of Classical Cannabinoids: CB1/CB2 Modulation. Perspect Medicin Chem. 2016;8:17–39. pmid:27398024
  10. 10. Ramirez-Franco J, Bartolome-Martin D, Alonso B, Torres M, Sanchez-Prieto J. Cannabinoid type 1 receptors transiently silence glutamatergic nerve terminals of cultured cerebellar granule cells. PLoS One. 2014;9(2):e88594. pmid:24533119
  11. 11. Lee SH, Ledri M, Toth B, Marchionni I, Henstridge CM, Dudok B, et al. Multiple Forms of Endocannabinoid and Endovanilloid Signaling Regulate the Tonic Control of GABA Release. J Neurosci. 2015;35(27):10039–57. pmid:26157003
  12. 12. Tanimura A, Yamazaki M, Hashimotodani Y, Uchigashima M, Kawata S, Abe M, et al. The endocannabinoid 2-arachidonoylglycerol produced by diacylglycerol lipase alpha mediates retrograde suppression of synaptic transmission. Neuron. 2010;65(3):320–7. pmid:20159446
  13. 13. Turcotte C, Blanchet MR, Laviolette M, Flamand N. The CB2 receptor and its role as a regulator of inflammation. Cell Mol Life Sci. 2016;73(23):4449–70. pmid:27402121
  14. 14. Maresz K, Carrier EJ, Ponomarev ED, Hillard CJ, Dittel BN. Modulation of the cannabinoid CB2 receptor in microglial cells in response to inflammatory stimuli. J Neurochem. 2005;95(2):437–45. pmid:16086683
  15. 15. Gertsch J, Leonti M, Raduner S, Racz I, Chen JZ, Xie XQ, et al. Beta-caryophyllene is a dietary cannabinoid. Proc Natl Acad Sci U S A. 2008;105(26):9099–104. pmid:18574142
  16. 16. Bisogno T, Howell F, Williams G, Minassi A, Cascio MG, Ligresti A, et al. Cloning of the first sn1-DAG lipases points to the spatial and temporal regulation of endocannabinoid signaling in the brain. J Cell Biol. 2003;163(3):463–8. pmid:14610053
  17. 17. Long LE, Lind J, Webster M, Weickert CS. Developmental trajectory of the endocannabinoid system in human dorsolateral prefrontal cortex. BMC Neurosci. 2012;13:87. pmid:22827915
  18. 18. Keimpema E, Alpar A, Howell F, Malenczyk K, Hobbs C, Hurd YL, et al. Diacylglycerol lipase alpha manipulation reveals developmental roles for intercellular endocannabinoid signaling. Sci Rep. 2013;3:2093. pmid:23806960
  19. 19. Fowler CJ, Jonsson KO, Tiger G. Fatty acid amide hydrolase: biochemistry, pharmacology, and therapeutic possibilities for an enzyme hydrolyzing anandamide, 2-arachidonoylglycerol, palmitoylethanolamide, and oleamide. Biochem Pharmacol. 2001;62(5):517–26. pmid:11585048
  20. 20. Basavarajappa BS. Critical enzymes involved in endocannabinoid metabolism. Protein Pept Lett. 2007;14(3):237–46. pmid:17346227
  21. 21. Herman AI, Kranzler HR, Cubells JF, Gelernter J, Covault J. Association study of the CNR1 gene exon 3 alternative promoter region polymorphisms and substance dependence. Am J Med Genet B Neuropsychiatr Genet. 2006;141B(5):499–503. pmid:16741937
  22. 22. Russo P, Strazzullo P, Cappuccio FP, Tregouet DA, Lauria F, Loguercio M, et al. Genetic variations at the endocannabinoid type 1 receptor gene (CNR1) are associated with obesity phenotypes in men. J Clin Endocrinol Metab. 2007;92(6):2382–6. pmid:17405839
  23. 23. Lu AT, Ogdie MN, Jarvelin MR, Moilanen IK, Loo SK, McCracken JT, et al. Association of the cannabinoid receptor gene (CNR1) with ADHD and post-traumatic stress disorder. Am J Med Genet B Neuropsychiatr Genet. 2008;147B(8):1488–94. pmid:18213623
  24. 24. Agrawal A, Wetherill L, Dick DM, Xuei X, Hinrichs A, Hesselbrock V, et al. Evidence for association between polymorphisms in the cannabinoid receptor 1 (CNR1) gene and cannabis dependence. Am J Med Genet B Neuropsychiatr Genet. 2009;150B(5):736–40. pmid:19016476
  25. 25. Proudnikov D, Kroslak T, Sipe JC, Randesi M, Li D, Hamon S, et al. Association of polymorphisms of the cannabinoid receptor (CNR1) and fatty acid amide hydrolase (FAAH) genes with heroin addiction: impact of long repeats of CNR1. Pharmacogenomics J. 2010;10(3):232–42. pmid:20010914
  26. 26. Carey CE, Agrawal A, Zhang B, Conley ED, Degenhardt L, Heath AC, et al. Monoacylglycerol lipase (MGLL) polymorphism rs604300 interacts with childhood adversity to predict cannabis dependence symptoms and amygdala habituation: Evidence from an endocannabinoid system-level analysis. J Abnorm Psychol. 2015;124(4):860–77. pmid:26595473
  27. 27. Sadhasivam S, Zhang X, Chidambaran V, Mavi J, Pilipenko V, Mersha TB, et al. Novel associations between FAAH genetic variants and postoperative central opioid-related adverse effects. Pharmacogenomics J. 2015;15(5):436–42. pmid:25558980
  28. 28. Wong BS, Camilleri M, Eckert D, Carlson P, Ryks M, Burton D, et al. Randomized pharmacodynamic and pharmacogenetic trial of dronabinol effects on colon transit in irritable bowel syndrome-diarrhea. Neurogastroenterol Motil. 2012;24(4):358–e169. pmid:22288893
  29. 29. Camilleri M, Kolar GJ, Vazquez-Roque MI, Carlson P, Burton DD, Zinsmeister AR. Cannabinoid receptor 1 gene and irritable bowel syndrome: phenotype and quantitative traits. Am J Physiol Gastrointest Liver Physiol. 2013;304(5):G553–60. pmid:23306084
  30. 30. Juhasz G, Lazary J, Chase D, Pegg E, Downey D, Toth ZG, et al. Variations in the cannabinoid receptor 1 gene predispose to migraine. Neurosci Lett. 2009;461(2):116–20. pmid:19539700
  31. 31. Lieb W, Manning AK, Florez JC, Dupuis J, Cupples LA, McAteer JB, et al. Variants in the CNR1 and the FAAH genes and adiposity traits in the community. Obesity (Silver Spring). 2009;17(4):755–60.
  32. 32. Hartman CA, Hopfer CJ, Haberstick B, Rhee SH, Crowley TJ, Corley RP, et al. The association between cannabinoid receptor 1 gene (CNR1) and cannabis dependence symptoms in adolescents and young adults. Drug Alcohol Depend. 2009;104(1–2):11–6. pmid:19443135
  33. 33. Clarke TK, Bloch PJ, Ambrose-Lanci LM, Ferraro TN, Berrettini WH, Kampman KM, et al. Further evidence for association of polymorphisms in the CNR1 gene with cocaine addiction: confirmation in an independent sample and meta-analysis. Addict Biol. 2013;18(4):702–8. pmid:21790903
  34. 34. de Miguel-Yanes JM, Manning AK, Shrader P, McAteer JB, Goel A, Hamsten A, et al. Variants at the endocannabinoid receptor CB1 gene (CNR1) and insulin sensitivity, type 2 diabetes, and coronary heart disease. Obesity (Silver Spring). 2011;19(10):2031–7.
  35. 35. Melroy-Greif WE, Wilhelmsen KC, Ehlers CL. Genetic variation in FAAH is associated with cannabis use disorders in a young adult sample of Mexican Americans. Drug Alcohol Depend. 2016;166:249–53. pmid:27394933
  36. 36. Swaminathan GJ, Bragin E, Chatzimichali EA, Corpas M, Bevan AP, Wright CF, et al. DECIPHER: web-based, community resource for clinical interpretation of rare variants in developmental disorders. Hum Mol Genet. 2012;21(R1):R37–44. pmid:22962312
  37. 37. Landrum MJ, Lee JM, Riley GR, Jang W, Rubinstein WS, Church DM, et al. ClinVar: public archive of relationships among sequence variation and human phenotype. Nucleic Acids Res. 2014;42(Database issue):D980–5. pmid:24234437
  38. 38. Amberger JS, Bocchini CA, Schiettecatte F, Scott AF, Hamosh A. OMIM.org: Online Mendelian Inheritance in Man (OMIM(R)), an online catalog of human genes and genetic disorders. Nucleic Acids Res. 2015;43(Database issue):D789–98. pmid:25428349
  39. 39. McKernan KJ, Spangler J, Helbert Y, Zhang L, Tadigotla V. DREAMing of a patent-free human genome for clinical sequencing. Nat Biotechnol. 2013;31(10):884–7. pmid:24104751
  40. 40. McKernan KJ, Spangler J, Zhang L, Tadigotla V, McLaughlin S, Warner J, et al. Expanded genetic codes in next generation sequencing enable decontamination and mitochondrial enrichment. PLoS One. 2014;9(5):e96492. pmid:24788618
  41. 41. Li H. Toward better understanding of artifacts in variant calling from high-coverage samples. Bioinformatics. 2014.
  42. 42. Garcia-Alcalde F, Okonechnikov K, Carbonell J, Cruz LM, Gotz S, Tarazona S, et al. Qualimap: evaluating next-generation sequencing alignment data. Bioinformatics. 2012;28(20):2678–9. pmid:22914218
  43. 43. Trubetskoy V, Rodriguez A, Dave U, Campbell N, Crawford EL, Cook EH, et al. Consensus Genotyper for Exome Sequencing (CGES): improving the quality of exome variant genotypes. Bioinformatics. 2015;31(2):187–93. pmid:25270638
  44. 44. Rimmer A, Phan H, Mathieson I, Iqbal Z, Twigg SR, Consortium WGS, et al. Integrating mapping-, assembly- and haplotype-based approaches for calling variants in clinical sequencing applications. Nat Genet. 2014;46(8):912–8. pmid:25017105
  45. 45. McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297–303. pmid:20644199
  46. 46. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011;43(5):491–8. pmid:21478889
  47. 47. Yourshaw M, Taylor SP, Rao AR, Martin MG, Nelson SF. Rich annotation of DNA sequencing variants by leveraging the Ensembl Variant Effect Predictor with plugins. Brief Bioinform. 2015;16(2):255–64. pmid:24626529
  48. 48. Cingolani P, Platts A, Wang le L, Coon M, Nguyen T, Wang L, et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly (Austin). 2012;6(2):80–92.
  49. 49. Kohler S, Doelken SC, Mungall CJ, Bauer S, Firth HV, Bailleul-Forestier I, et al. The Human Phenotype Ontology project: linking molecular biology and disease through phenotype data. Nucleic Acids Res. 2014;42(Database issue):D966–74. pmid:24217912
  50. 50. Landrum MJ, Lee JM, Benson M, Brown G, Chao C, Chitipiralla S, et al. ClinVar: public archive of interpretations of clinically relevant variants. Nucleic Acids Res. 2016;44(D1):D862–8. pmid:26582918
  51. 51. Genomes Project C, Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, et al. A global reference for human genetic variation. Nature. 2015;526(7571):68–74. pmid:26432245
  52. 52. Song W, Gardner SA, Hovhannisyan H, Natalizio A, Weymouth KS, Chen W, et al. Exploring the landscape of pathogenic genetic variation in the ExAC population database: insights of relevance to variant classification. Genet Med. 2015.
  53. 53. Liu X, Jian X, Boerwinkle E. dbNSFP v2.0: a database of human non-synonymous SNVs and their functional predictions and annotations. Hum Mutat. 2013;34(9):E2393–402. pmid:23843252
  54. 54. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–9. pmid:16862161
  55. 55. Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405–24. pmid:25741868
  56. 56. Girdea M, Dumitriu S, Fiume M, Bowdin S, Boycott KM, Chenier S, et al. PhenoTips: patient phenotyping software for clinical and research use. Hum Mutat. 2013;34(8):1057–65. pmid:23636887
  57. 57. Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fennell T, et al. Analysis of protein-coding genetic variation in 60,706 humans. Nature. 2016;536(7616):285–91. pmid:27535533
  58. 58. Shao Z, Yin J, Chapman K, Grzemska M, Clark L, Wang J, et al. High-resolution crystal structure of the human CB1 cannabinoid receptor. Nature. 2016.
  59. 59. Hua T, Vemuri K, Pu M, Qu L, Han GW, Wu Y, et al. Crystal Structure of the Human Cannabinoid Receptor CB1. Cell. 2016;167(3):750–62 e14. pmid:27768894
  60. 60. Hua T, Vemuri K, Nikas SP, Laprairie RB, Wu Y, Qu L, et al. Crystal structures of agonist-bound human cannabinoid receptor CB1. Nature. 2017;547(7664):468–71. pmid:28678776
  61. 61. Reisenberg M, Singh PK, Williams G, Doherty P. The diacylglycerol lipases: structure, regulation and roles in and beyond endocannabinoid signalling. Philos Trans R Soc Lond B Biol Sci. 2012;367(1607):3264–75. pmid:23108545
  62. 62. Samocha KE, Robinson EB, Sanders SJ, Stevens C, Sabo A, McGrath LM, et al. A framework for the interpretation of de novo mutation in human disease. Nat Genet. 2014;46(9):944–50. pmid:25086666
  63. 63. Andersson H D'Antona AM, Kendall DA, Von Heijne G, Chin CN. Membrane assembly of the cannabinoid receptor 1: impact of a long N-terminal tail. Mol Pharmacol. 2003;64(3):570–7. pmid:12920192
  64. 64. Litvin Y, Phan A, Hill MN, Pfaff DW, McEwen BS. CB1 receptor signaling regulates social anxiety and memory. Genes Brain Behav. 2013;12(5):479–89. pmid:23647582
  65. 65. Bilkei-Gorzo A, Drews E, Albayram O, Piyanova A, Gaffal E, Tueting T, et al. Early onset of aging-like changes is restricted to cognitive abilities and skin structure in Cnr1(-)/(-) mice. Neurobiol Aging. 2012;33(1):200 e11–22.
  66. 66. Silvani A, Berteotti C, Bastianini S, Lo Martire V, Mazza R, Pagotto U, et al. Multiple sleep alterations in mice lacking cannabinoid type 1 receptors. PLoS One. 2014;9(2):e89432. pmid:24586776
  67. 67. Pava MJ, Makriyannis A, Lovinger DM. Endocannabinoid Signaling Regulates Sleep Stability. PLoS One. 2016;11(3):e0152473. pmid:27031992
  68. 68. Boles RG, Hornung HA, Moody AE, Ortiz TB, Wong SA, Eggington JM, et al. Hurt, tired and queasy: Specific variants in the ATPase domain of the TRAP1 mitochondrial chaperone are associated with common, chronic "functional" symptomatology including pain, fatigue and gastrointestinal dysmotility. Mitochondrion. 2015.
  69. 69. Riant F, Ducros A, Ploton C, Barbance C, Depienne C, Tournier-Lasserve E. De novo mutations in ATP1A2 and CACNA1A are frequent in early-onset sporadic hemiplegic migraine. Neurology. 2010;75(11):967–72. pmid:20837964
  70. 70. Russo EB. Clinical endocannabinoid deficiency (CECD): can this concept explain therapeutic benefits of cannabis in migraine, fibromyalgia, irritable bowel syndrome and other treatment-resistant conditions? Neuro Endocrinol Lett. 2008;29(2):192–200. pmid:18404144
  71. 71. Russo EB. Clinical Endocannabinoid Deficiency Reconsidered: Current Research Supports the Theory in Migraine, Fibromyalgia, Irritable Bowel, and Other Treatment-Resistant Syndromes. Cannabis and Cannabinoid Research. 2016;1(1):154–65. pmid:28861491
  72. 72. Rimmerman N, Juknat A, Kozela E, Levy R, Bradshaw HB, Vogel Z. The non-psychoactive plant cannabinoid, cannabidiol affects cholesterol metabolism-related genes in microglial cells. Cell Mol Neurobiol. 2011;31(6):921–30. pmid:21533611
  73. 73. Kaczocha M, Rebecchi MJ, Ralph BP, Teng YH, Berger WT, Galbavy W, et al. Inhibition of fatty acid binding proteins elevates brain anandamide levels and produces analgesia. PLoS One. 2014;9(4):e94200. pmid:24705380
  74. 74. Pertwee RG. Elevating endocannabinoid levels: pharmacological strategies and potential therapeutic applications. Proc Nutr Soc. 2014;73(1):96–105. pmid:24135210
  75. 75. Rhyne DN, Anderson SL, Gedde M, Borgelt LM. Effects of Medical Marijuana on Migraine Headache Frequency in an Adult Population. Pharmacotherapy. 2016;36(5):505–10. pmid:26749285
  76. 76. Ibsen MS, Connor M, Glass M. Cannabinoid CB1 and CB2 Receptor Signaling and Bias. Cannabis Cannabinoid Res. 2017;2(1):48–60. pmid:28861504
  77. 77. Khajehali E, Malone DT, Glass M, Sexton PM, Christopoulos A, Leach K. Biased Agonism and Biased Allosteric Modulation at the CB1 Cannabinoid Receptor. Mol Pharmacol. 2015;88(2):368–79. pmid:26044547
  78. 78. Jenniches I, Ternes S, Albayram O, Otte DM, Bach K, Bindila L, et al. Anxiety, Stress, and Fear Response in Mice With Reduced Endocannabinoid Levels. Biol Psychiatry. 2016;79(10):858–68. pmid:25981172
  79. 79. Sugaya Y, Yamazaki M, Uchigashima M, Kobayashi K, Watanabe M, Sakimura K, et al. Crucial Roles of the Endocannabinoid 2-Arachidonoylglycerol in the Suppression of Epileptic Seizures. Cell Rep. 2016;16(5):1405–15. pmid:27452464
  80. 80. Gao Y, Vasilyev DV, Goncalves MB, Howell FV, Hobbs C, Reisenberg M, et al. Loss of retrograde endocannabinoid signaling and reduced adult neurogenesis in diacylglycerol lipase knock-out mice. J Neurosci. 2010;30(6):2017–24. pmid:20147530
  81. 81. Martella A, Sepe RM, Silvestri C, Zang J, Fasano G, Carnevali O, et al. Important role of endocannabinoid signaling in the development of functional vision and locomotion in zebrafish. FASEB J. 2016;30(12):4275–88. pmid:27623930