Exploring opportunities for drug repurposing and precision medicine in cannabis use disorder using genetics

Abstract Cannabis use disorder (CUD) remains a significant public health issue globally, affecting up to one in five adults who use cannabis. Despite extensive research into the molecular underpinnings of the condition, there are no effective pharmacological treatment options available. Therefore, we sought to further explore genetic analyses to prioritise opportunities to repurpose existing drugs for CUD. Specifically, we aimed to identify druggable genes associated with the disorder, integrate transcriptomic/proteomic data and estimate genetic relationships with clinically actionable biochemical traits. Aggregating variants to genes based on genomic position, prioritised the phosphodiesterase gene PDE4B as an interesting target for drug repurposing in CUD. Credible causal PDE4B variants revealed by probabilistic finemapping in and around this locus demonstrated an association with inflammatory and other substance use phenotypes. Gene and protein expression data integrated with the GWAS data revealed a novel CUD associated gene, NPTX1, in whole blood and supported a role for hyaluronidase, a key enzyme in the extracellular matrix in the brain and other tissues. Finally, genetic correlation with biochemical traits revealed a genetic overlap between CUD and immune‐related markers such as lymphocyte count, as well as serum triglycerides.

The high heritability of CUD suggests a strong genetic component, which could be leveraged to inform new treatment options. Drug repurposing, where existing compounds are used in a new indication, is an attractive option to expediate changes in CUD clinical practice, as investment and success in de novo drug development for psychiatry remains comparatively limited.
Despite the potential of genetics, the tremendous heterogeneity among individuals with CUD makes current drug development challenging. To date, no pharmacotherapy has been clearly effective, and there are no approved medications for CUD treatment. 5,9 Various pharmacological approaches have been tested to assist people with CUD in reducing their cannabis use by addressing withdrawal symptoms, craving and other cognitive factors. However, most of these interventions have not progressed beyond small pilot trials. 10,11 Brezing and Levin's report highlights the need to consider individual patient characteristics for the treatment of CUD. Complex, polygenic disorders such as CUD do not have a one-size-fits-all approach, 5 and research over the past 20 years has shown that not everyone who uses cannabis is affected adversely in the same way.
Emerging vulnerability factors, including certain genes and personality characteristics, are being identified, although the mechanisms underlying the negative effects of cannabis use are not fully understood. 12 Given the heritability of CUD, genetics may provide a means to identify and prioritise novel treatment opportunities with greater specificity. Gene-based enrichment approaches, functional genomics and genetic correlation and causality analyses could be used to identify and refine opportunities for pharmacological interventions. Furthermore, genetic analyses of therapeutically actionable traits such as serum blood biomarkers could also inform drug repurposing opportunities. 13 In the present study, we explore these approaches to inform drug repurposing opportunities, expand on current literature and identify new therapeutic targets for those with CUD.  15 The cases in the PGC Substance Use Disorders working groups all met DSM-IV diagnostic criteria.
2.2 | Gene-based analyses to identify druggable targets associated with CUD Gene-based association analysis for CUD GWAS summary statistics 6 was performed using MAGMA version 1.09 (https://ctg.cncr.nl/ software/magma). MAGMA maps SNPs to genes by aggregating common variants (MAF > 0.01) at the gene level, which increases discovery power by using the linear combination of SNP-wise P-values as test statistic, reducing the burden of multiple testing correction seen in univariate GWAS. We used the 1000 genomes phase 3 European reference panel population for LD estimation and mapped variants to the NCBI hg19 genome assembly, which contained 18 297 autosomal protein-coding genes with SNPs mapped within the defined coordinates of 5 kb upstream and 1.5 kb downstream. Any genes from the major histocompatibility complex (MHC, chr6:28477797-33448354) on chromosome 6 were removed. Multiple testing correction was done using the Bonferroni method, where we divided the alpha level by the number of comparisons and set P < 2.7 Â 10 À6 as the P-value required for significance.

| Probabilistic finemapping
Given the association signal may be therapeutically actionable and to further support the MAGMA findings that PDE4B as the most significant gene in this locus, we subjected the PDE4B region to further analysis to evaluate whether it is a causal association. Firstly, we leveraged probabilistic finemapping to prioritise putatively causal genetic variation in this region. Specifically, we used a conventional Bayesian approach that was applied to all variants in the CUD GWAS within 3 MB of the defined PDE4B genic boundaries. Asymptotic Bayes' factors (ABF) for each SNP were approximated using Wakefield's method, assuming a prior variance of 0.2, 2 as outlined extensively elsewhere. 16 ABF were summed to define credible sets given that Bayes' factors are proportional to the posterior probability (PP) for causality for each variant. In other words, to define a 95% credible set of variants in this region, which contains a causal variant with 95% probability, variantwise PP were summed in descending order until 0.95 is exceeded. This method assumes a single causal variant so that we did not have to account for LD between variants, which has been demonstrated to be susceptible to false positives in finemapping studies with a prior of multiple causal variants that use references external to the GWAS or not otherwise very well matched at a population level. 17

| PDE4B phenome-wide association study
We then wished to further investigate the phenotypic relevance for the finemapped CUD association signal in PDE4B. For the variant with the highest PP, we performed a phenome-wide association study (pheWAS) using IEU open GWAS project database (https://gwas. mrcieu.ac.uk/phewas/) and the FinnGen release 7 resource (https:// r7.finngen.fi/about). The concept underlying this is to assess other traits to which this variant is linked.
2.5 | Colocalisation of the PDE4B association signals with expression quantitative trait loci We then attempted to infer using expression quantitative trait loci (eQTL) data whether upregulation or down-regulation of PDE4B was associated with liability to CUD and whether CUD and PDE4B expression displayed statistical colocalisation. PDE4B eQTLs were sourced from the multi-tissue eQTL catalogue resource (https://fivex.sph. umich.edu/). 18 We used cortical eQTL data assembled by the MetaBrain consortium (N = 2970) for colocalisation analyses via the coloc method. 19 The coloc approach infers the PP of five competing hypotheses (H) for a given region: H 0 = the region is associated with neither trait, H 1 = the region is associated with trait one, H 2 = the region is associated with trait two, H 3 = the region is associated with both traits but with a different underlying causal variant, and H 4 = both traits share a causal variant. This was implemented using default priors via version 4 of the coloc R package. We also performed a sensitivity analysis where we varied the SNP priors between (1 Â 10 À8 and 1 Â 10 À4 ), as implemented by the sensitivity function in the coloc package.

| Functional genomics analyses of CUD using transcriptome-and proteome-wide association studies
To further refine our understanding of CUD associated genes that may present opportunities for repurposing, we performed a transcriptome-wide association study (TWAS) and a proteome-wide association study (PWAS) via the FUSION framework. 20 We achieved this through leveraging genetically imputed models of mRNA and protein expression. For TWAS, we used tissue samples of whole blood (GTEx v7) and post-mortem brain (GTEx v7, PsychENCODE), 20,21 whereas protein expression weights of post-mortem brain and whole blood were yielded from Religious Orders Study and Memory and Aging Project (ROS/MAP) 22 and NHLBI's Atherosclerosis Risk in Communities (ARIC), 23 respectively. As this method integrates SNP effects from the model of genetically predicted expression with the effects of the same SNPs on CUD, TWAS/PWAS Z-scores, after accounting for linkage disequilibrium (LD), can be a conceptualised measure of genetic covariance between mRNA or protein expression of the gene and the GWAS trait of interest. As a result, the sign of the TWAS/ PWAS Z score is informative as to which direction of genetically predicted mRNA or protein expression is associated with increased odds of CUD. To ensure we captured only the most confidently associated genes that could constitute drug repurposing candidates, we utilised a conservative method for multiple-testing correction whereby the Bonferroni methodology was implemented to divide the alpha level (0.05) by the total number of significantly cis-heritable models of genetically regulated expression (GReX) tested from any brain tissue considered or whole blood. Finally, we implemented probabilistic finemapping of TWAS Z scores using FOCUS v0.6.10 to refine potential causal genes for which predicted expression is associated with CUD, as outlined extensively elsewhere. 24,25 Using the default prior and prior variance, we estimated marginal posterior inclusion probabilities (PIP) for membership of 90% credible set for each gene in the region in and around the implicated hyaluronidase gene cluster given its potential therapeutic relevance.

| Genetical correlation between CUD and biochemical traits
LD-score regression analysis (LDSR; v1.0.1) (https://github.com/ bulik/ldsc) 26 was used to estimate genetic correlation between CUD and GWAS on 50 biochemical traits from the UK Biobank (UKBB) as outlined previously by our group for other psychiatric disorders. 13 As the mode of action of many existing drugs involves modulating biochemical traits, for example, lipids and blood glucose, shared biology that may be indexed by genetic correlation could be informative as drug repurposing opportunities. In LDSR, the genetic covariance is estimated by regressing SNP-wise χ 2 , the product of the marginal SNP effects from both traits, on its LD score. The SNP heritability estimates for both traits are used to normalise the genetic covariance to obtain genetic correlation (r g ), which can be accurately estimated in the presence of any sample overlap only affects the LDSR intercept and not the slope. Bonferroni multiple testing correction was used for the 50 biochemical traits tested to define a significant r g . The CUD and biochemical trait GWAS summary statistics were cleaned to ensure they contain HapMap3 SNPs outside the MHC with minor allele frequency >0.05 for consistency. To evaluate evidence of a causal relationship, we used the latent causal variable (LCV) on all genetically correlated CUD and biochemical trait pairs as demonstrated elsewhere and in our previous work. 13,27 To estimate partial genetic causality, the LCV framework leverages the bivariate genomewide distribution of marginal SNP effects on both CUD and each of the biochemical traits and outputs the posterior mean genetic causality proportion metric (GCP), with GCP > 0 implying partial genetic causality of trait one on trait two, and vice versa. To calculate the GCP metric, the LCV model utilises the genome-wide SNP-trait association Z scores for two traits and the mixed fourth moments (cokurtosis) of the respective distributions to assess whether there is evidence for a causal effect of one trait on the other. To guard against false positives, partial genetic causality was defined using the recommended threshold of a significantly non-zerojGCPj > 0.6. 27 3 | RESULTS

| Exploring drug interactions of CUDassociated genes
Using aggregated gene-level association (MAGMA), we observed two genes that were associated with CUD-PDE4B (P = 2.09 Â 10 À6 ) and FOXP2 (P = 9.30 Â 10 À7 ) after Bonferroni correction (P < 2.7 Â 10 À6 ; Table S1). These genes have been previously reported in Johnson et al. 6 Further analysis using the Drug Gene Interaction Database (DGIdb) on the MAGMA significant genes for CUD revealed the highest interaction score for PDE4B was with dyphylline, a vasodilator and bronchodilator agent. Dyphylline is a phosphodiesterase 4 inhibitor clinically used for the prevention of bronchial asthma or other respiratory diseases. 28 It has similar pharmacological actions and safety profiles as other xanthine derivatives, such as caffeine and theobromine. 29 According to DrugBank, the mechanism of action of dyphylline in humans is a cAMP-specific 3 0 ,5 0 -cyclic phosphodiesterase 4A, 4B, 4C and 4D inhibitor, and an adenosine receptor A1 and A2a antagonist. [30][31][32][33] No drug interactions for FOXP2 were observed in DrugBank or in DGIdb. Interestingly, the PDE3 inhibitor, dipyridamole, also targets PDE4B and has been investigated therapeutically for bipolar disorders. 34 Using probabilistic finemapping, we prioritised one variant in the 95% credible set with a posterior probability greater than 40%

| Dysregulated hyaluronidase enzyme expression associated with liability to CUD through integration of transcriptomic and proteomic data
We utilised TWAS and PWAS methods to integrate genomic information into functionally relevant units that map to genes or proteins and their expression. Next, we assessed the statistical associations between CUD and predicted gene expression. Although TWAS/ PWAS associations do not necessarily imply a causal relationship, this approach can be implemented to identify candidate genes located at loci with an inferred mechanistic underpinning through expression. 35 In multi-brain region and blood TWAS of CUD, significantly down-regulated genetically proxied expression of the hyaluronidase gene HYAL3 was associated with the disorder in the whole blood, as well as several brain regions including the amygdala (Z = À4.294, P = 1.75 Â 10 À5 ), anterior cingulate cortex (Z = À5.103, P = 3.34 Â 10 À7 ), hippocampus (Z = À5.103, P = 3.34 Â 10 À7 ), hypothalamus (Z = À5.185, P = 1.05 Â 10 À7 ), nucleus accumbens (Z = À5.552, P = 2.82 Â 10 À8 ) and putamen (Z = À5.433, P = 5.51 Â 10 À8 ). The TWAS also revealed significantly up-regulated expression of NAT6, also known as HYAL1, in the cortex ( Figure 1A).
NAT6 is a hyaluronidase that is proximally located on chromosome 3, 36 similar to HYAL3. However, HYAL1 is considered to be a more canonically active hyaluronidase compared to HYAL3. 37 Additionally, a third gene TTC3 was identified using expression weights from the 3p21.3 putamen region (Z = 4.31, P = 1.63 Â 10 À5 ).
Overexpression of TTC3 has been associated with negative effects on cognitive function and neuronal health. 39,40 As a result, this 3p21.3 region contains several plausible CUD risk genes, and future in silico and experimental follow-up is warranted to disentangle true effects.
We expanded our analysis beyond mRNA expression by conducting a PWAS. In the brain, we again found that NAT6/HYAL1 was significantly up-regulated in terms of protein abundance (Z = 5.019, P = 5.19 Â 10 À7 ) ( Figure 1C). In the whole blood, we observed a novel significant association between predicted protein expression of NPTX1 and CUD (Z = À4.57, P = 4.88 Â 10 À6 ), which is critical for early human development 43 ( Figure 1D).
Although NPTX1 down-regulation was also detected in the TWAS, it did not reach statistical significance after correction for mRNA expression (Table S5).

| Exploring the therapeutic potential of hyaluronidase-1 enzyme inhibition
The TWAS/PWAS analyses revealed HYAL1 as the most promising target for drug repurposing, as it can be modified by some existing nutraceuticals and investigational compounds (Table 1). However, the other TWAS/PWAS associations did not have current approved therapies that target them, although PDE4B (from the MAGMA analyses) is an existing drug target. If hyaluronidase dysregulation is indeed associated with CUD, as suggested by the TWAS/PWAS, it may represent a promising opportunity for therapeutic investigation.
Hyaluronic acid, also known as hyaluronan, is a crucial component of the extracellular matrix and is known to play a role in the regulation of inflammatory processes and embryonic development. 44 In F I G U R E 1 Miami plot of cannabis use disorder (CUD) transcriptome-wide association (TWA) analysis of brain cerebral cortex (A) and whole blood (B) and proteome-wide association (PWA) analysis of brain (C) and whole blood (D). The orange horizontal lines are the significance threshold after Bonferroni correction for the total number of imputated models of expression (alpha/n). gene has a direct causal effect. However, this interpretation is conflicted by the association we found between down-regulation of another proximal hyaluronidase, HYAL3, and CUD. This requires further investigation, although there is evidence HYAL3 does not directly contribute to the metabolism of hyaluronic acid, 37 but rather it is believed to do so by augmenting the activity of HYAL1. 38 If hyaluronic acid catabolism were dysregulated in CUD, this finding may suggest the potential use of oral hyaluronic acid supplementation, which has been previously trialled for conditions like osteoarthritis. Additionally, the use of ascorbyl palmitate (L-ascorbic acid 6-hexadecanoate), an antioxidant that also inhibits hyaluronidase activity, may have therapeutic benefits as a neuroprotective agent.
However, it is important to note that ascorbyl palmitate is not currently registered with the Food and Drug Administration (FDA) for therapeutic use (Table 1). [46][47][48] To attempt to resolve the role of hyaluronic acid catabolism in CUD via effects at this locus, we applied probabilistic finemapping of TWAS test statistics on the region located at 3p21.3 that harbours both HYAL1 and HYAL3. Using a combined database of SNP weights from all GTEx tissues, as well some brain and blood datasets, HYAL1 was prioritised as the most likely causal gene for CUD, with a PIP of being in the 90% credible set of 89.3%, compared to 1.51% for HYAL3. However, probabilistic finemapping applied to the PsychEN-CODE SNP weight set, wherein HYAL1 did not have a model of genetically predicted expression available, prioritised HYAL3 instead (PIP = 78.3%). Given this disparity, further work is still required to reconcile how hyaluronidase activity may be involved in the aetiology of CUD.

| CUD genetically correlated with clinically significant metabolites and immune markers
We tested the genetic correlation between CUD and a panel of blood-based biomarkers from the UKBB to gain further insight into drug repurposing opportunities by exploring the interplay between circulating biochemical factors and the pathophysiology of CUD. After Bonferroni correction, we found that CUD was genetically correlated with 12 of the 50 blood-based biomarkers we tested. This included alanine aminotransferase (r g = 0.185, SE = 0.034, P = 7.25 Â 10 À7 ), CRP (r g = 0.206, SE = 0.55, P = 2.0 Â 10 À4 ), eosinophil count (r g = 0.122, SE = 0.032, P = 9.75 Â 10 À5 ), gamma glutamyltransferase (r g = 0.193, SE = 0.043, P = 7.94 Â 10 À6 ), lymphocyte count (r g = 0.178, SE = 0.033, P = 6.04 Â 10 À8 ), triglycerides (r g = 0.146, SE = 0.039, P = 2.0 Â 10 À4 ) and white blood cell (WBC) count (r g = 0.155, SE = 0.032, P = 1.89 Â 10 À6 ). It is worth nothing that inflammation of the CNS has long been linked to psychiatric disorders, including schizophrenia. 57 Although we did not find direct evidence of a causal effect, the shared biology between CUD and the biomarkers we tested using LDSC is still informative for future treatment opportunities (Tables S6 and S7).

| DISCUSSION
CUD has become the primary use disorder for which individuals seek treatment for globally, surpassing all other substances (UNODC, 2017). In this study, we used genetic approaches to explore drug repurposing opportunities for CUD, which currently has no approved treatments. Using gene-level association, we observed that SNPs localising at PDE4B had the greatest association with CUD. The causal variant in PDE4B, which was found using probabilistic finemapping and is associated with inflammation and substance use in a phenomewide analysis, is also believed to have a role in modulating the immune response of monocytes and neutrophils, given that PDE4B is the predominant isoform of the PDE4 protein family. 58,59 A role for inflammation in CUD is further supported by its positive genetic correlation T A B L E 1 Brief overview of compounds under investigation for inhibition of hyaluronidase-1.
This suggests that a more complex relationship may exist, perhaps through specific cytokine repertoires that were not able to be considered in this study. While these genetic correlations with CUD are not necessarily causal, they provide insight into potential shared biological mechanisms between biochemical traits that are targeted by drugs that warrant further investigation.
It is also plausible that PDE4B influences CUD more specifically in the brain given its high expression in the tissue. Inhibitors of type 4 phosphodiesterase (PDE) are known to have anti-inflammatory effects in various cells, including glia, by increasing cAMP and reducing inflammatory signalling 60 . One such PDE inhibitor, ibudilast, can cross the blood-brain barrier and suppresses TNF-alpha production and astrocyte and microglial activation, making it a potential treatment option for CUD. [61][62][63] We also conducted PWAS analyses that suggested that increasing genetically predicted neuronal pentraxin 1 (NPTX1) protein expression was protective for CUD. As a member of the pentraxin superfamily, NPTX1 shares structural homology with CRP. 64 NPTX1 is predominantly expressed in the brain and plays an important role in the regulation of synaptic strength and plasticity, as well as neurodegeneration. 65,66 Abnormal expression of neuronal pentraxins has been linked to various psychiatric disorders, including schizophrenia and bipolar disorder. 67,68 Interestingly, NPTX1 is a negative regulator of excitatory synapses, and its knock-down has been shown to increase the number of excitatory synapses, while elevated levels of NPTX1 in the plasma are linked to mild cognitive impairment. 69,70 If increased NPTX1 expression is protective for CUD, this could reflect a compensatory mechanism in response to the increased number of excitatory synapses in the brain, as previous research has shown that administration of addictive drugs enhances excitatory synaptic strength. 64 Further studies are needed to determine the causal relationship between NPTX1 expression and CUD, as well as to investigate the potential therapeutic implications of targeting NPTX1 in CUD treatment.
In this study, we extended the previously implicated role of HYAL1 and HYAL3 in CUD 6 by demonstrating that an increased genetically proxied protein expression of HYAL1 plausibly increases the risk for CUD beyond a mere association with mRNA expression.
Our finemapping analyses provided some support that HYAL1, one of the two proximally located hyaluronidase genes, is more likely a candidate causal gene, although this was not definitive. The role of HYAL1 is particularly interesting because it down-regulates the expression of hyaluronic acid. A lack of hyaluronic acid in the brain has been shown to cause a reduction in extracellular space (ECS) volume and induce an epileptic phenotype in mice. 71,72 This effect could be because HYAL1 digests high molecular weight hyaluronic acid into low molecular weight fragments intracellularly, and these small fragments are important for activating pathways involved in endothelial cell proliferation, adhesion and migration. 73 In addition to the repurposing opportunities provided by oral hyaluronic acid itself or the hyaluronidase inhibitor ascorbyl palmitate, inhibition of hyaluronic acid synthesis using the FDA-approved prescription drug 4-methylumbelliferone (4-MU) or genetic deletion of hyaluronan synthase genes has been shown to improve glucose homoeostasis and is already being explored for the treatment of obesity and diabetes. 74 This is particularly salient given the mounting evidence that metabolic dysfunction is an important component of substance use disorders. 75 However, additional research is needed to clarify the role of hyaluronic acid biology in CUD and to evaluate the potential of pharmacological agents for repurposing as CUD treatments. This could include in vivo investigation of these compounds in suitable animal models of addiction-associated behaviours, as well as characterising additional evidence of dysregulated hyaluronic acid biology in CUD using resources such as post-mortem brain. Moreover, follow-up studies should also be conducted to validate the potential risk genes prioritised in this study including functional investigations.