Four single nucleotide polymorphisms in genes involved in neuronal signaling are associated with Opioid Use Disorder in West Virginia

Abstract

Objective:

Pilot study to assess utility in opioid use disorder (OUD) of a panel of single nucleotide polymorphisms in genes previously related to substance use disorder (SUD) and/or phenotypes that predispose individuals to OUD/SUD.

Design:

Genetic association study.

Setting:

West Virginia University’s Chestnut Ridge Center Comprehensive Opioid Abuse Treatment (COAT) clinic for individuals diagnosed with OUD.

Patients:

Sixty patients 18 years of age or older with OUD undergoing medication (buprenorphine/naloxone)-assisted treatment (MAT); all sixty patients recruited contributed samples for genetic analysis.

Outcome Measure(s):

Minor allele frequencies for single nucleotide polymorphisms.

Results:

Four of the fourteen single nucleotide polymorphisms examined were present at frequencies that are statistically significantly different than in a demographically-matched general population.

Conclusions:

For the purposes of testing WV individuals via genetic means for predisposition to OUD, at least four single nucleotide polymorphisms in three genes are likely to have utility in predicting susceptibility. Additional studies with larger populations will need to be conducted to confirm these results before use in a clinical setting.

INTRODUCTION

Opioid use disorder (OUD) has reached epidemic proportions in America.^{1, 2} Non-medical use of opioid analgesics and heroin is on the rise.^3–5 Opioid overdose is now the leading cause of death of people under 50 years old.⁶ Mindful of these developments, many addiction researchers are focusing combatting the epidemic with prevention and treatment.^7–10 A shared aspiration is to obtain an individual’s genetic “fingerprint” and identify whether they are at increased risk for developing OUD (e.g., ref. ¹¹). Such testing is likely to interrogate several genes, as a multitude of factors^12–23 contribute to OUD predisposition. Several gene variations are associated with predisposition to OUD and/or OUD treatment response^12–16. Environmental factors also contribute to OUD^19–22, and some environmental factors can affect the influence of genetic factors on predisposition to developing addiction²³. The ability to identify a priori individuals at increased OUD risk would help greatly in prevention. For instance, using opioid analgesics with these individuals could be avoided or strictly monitored.

West Virginia is the epicenter of this opioid crisis, leading in opioid-related overdose deaths²⁶. The state is also demographically homogenous^27*, being predominantly whites (94%) of European Caucasian ancestry (2010 US Census Data; http://censusviewer.com/state/WV). These factors make West Virginia ideal for genetic studies aimed at curbing the opioid epidemic. West Virginia University’s Chestnut Ridge Center is home to a successful medication-assisted treatment program, known as the Comprehensive Opioid Addiction Treatment (COAT) clinic^{7, 28–30}. Given access to this WV patient population, we sought to address whether there are single nucleotide polymorphisms (SNPs) more or less prevalent in WV persons with OUD than in the general population.

METHODS

Recruitment

60 volunteers, diagnosed with OUD using DSM-V criteria³¹, were recruited from WVU’s COAT clinic under IRB 1506733605. All participants were patients of the COAT clinic in Morgantown, West Virginia. The COAT clinic’s service area (and hence our research subject area) is geographically diverse, as it draws from the entire state.

Tissue Collection

Consenting volunteers provided buccal swabs, which were placed in 15-ml conical centrifuge tubes (Corning) and stored at −80ºC.

Genomic DNA Extraction

DNA was extracted using QIAamp DNA Mini Kits according to manufacturer’s protocol. Yield was assessed by measuring 260-nm absorbance of DNA extracts on a QIAxpert microfluidic reader. Purity was assessed by measuring 260-nm and 280-nm absorbance; all samples had a 260-nm/280-nm quotient greater than 1.8 units.

Genotyping

Genotyping was performed using TaqMan (FisherSci #4351379) primer-probe sets for each SNP (Table 1) and Type-it Fast SNP Probe PCR Kits (Qiagen) according to manufacturer’s protocol. Samples were run on a Rotor-Gene Q (Qiagen) in duplicate, and no-DNA negative controls were also performed for each TaqMan primer-probe set.

Table 1.

Genetic markers and minor allele frequency (MAF) data obtained from public databases (1000 Genomes Project and/or ExAc) and WV COAT clinic patients (N = 60).

Gene	dbSNP ID	Minor Allele	TaqMan Probe	MAF via 1000Genomes*	MAF via ExAc†	MAF from COAT clinic‡
OPRM1	rs9479757	A	C_25472011_10	0.097	0.082	0.500 p < 0.001
OPRM1	rs1799971	G	C_8950074_1	0.162	0.185	0.083 p = 0.018
RGS2	rs4606	G	C_2498717_10	0.276	n.a.	0.533 p < 0.001
HTR2B	rs6736017	C	C_25614588_20	n.a.	0.0078	0.500 p < 0.001
OPRM1	rs2075572	G	C_1691815_1	0.413	n.a.	0.483 p = 0.12
RGS17	rs596359	T	C_7830523_10	0.517	n.a.	0.467 p = 0.27
DRD2 / ANKK1	rs1800497	A	C_7486676_10	0.188	0.256	0.217 p = 0.41
DRD2	rs1799978	C	C_7486599_20	0.060	n.a.	0.075 p = 0.44
BDNF	rs6265	T	C_11592758_10	0.197	0.194	0.167 p = 0.49
KCNC1	rs60349741	C	C_89088414_10	0.001	n.a.	0.000
HTR2B	rs77982984	A	C_99996051_10	0.003	0.0012	0.000
HTR2B	rs78484969	C	C_99996081_10	0.001 (N = 1030 Finns)	n.a.	0.000
HTR2B	rs77570025	G	C_99996068_10	0.006	0.002	0.000
CREB1	rs35349697	A	C_25636228_10	n.a.	0.000008	0.000

^*Minor allele frequency reported on N = 1006 Europeans via 1000 Genomes (except as otherwise noted); n.a. = not available.

^†Minor allele frequence reported on N ~ 60,000 people via ExAc database of the Broad Institute; n.a. = not available.

^‡p-value, denoting significance of difference between MAFs, is derived by comparison between WV COAT clinic and European (if not, then ExAc) MAF using binomial test (null hypothesis is “no difference between WV COAT patients and Europeans”; alternative is “there is difference between WV COAT patients and Europeans”)

Statistical Analysis

Allele frequencies were compared to those of 1006 individuals of European descent (1000 Genomes Project³²); if no information was available from the 1000 Genomes Project, allele frequencies were compared to ExAc database³³ of ~60,000 genomes. Statistical analyses for differences of SNP variant frequency between groups were performed using Pearson’s chi-square statistical test. The null hypothesis was that the allele frequencies from WV COAT clinic participants were the same as those from public databases. Therefore, p-value < 0.05 implies that the allele frequencies from WV COAT clinic participants are significantly different from the general public.

RESULTS

A literature search for SNPs associated with OUD/SUD and/or related traits (e.g., anxiety) identified fourteen of known or suspected relevance to the biological actions of opioids (i.e., variations within genes encoding G protein-coupled receptors [the GPCRs dopamine D₂ receptor^{34, 35}, serotonin 5-HT_2B receptor^36–38, and μ-opioid receptor^39–62]; Regulators of G protein Signaling [RGS2, RGS17]^63–70; a nerve growth factor [brain-derived neurotrophic factor]^{71, 72}; a voltage-gated potassium channel [KCNC1]⁷³; and, a second messenger-regulated transcription factor [cyclic-AMP responsive element binding protein-1]⁷⁴; Table 1). We isolated buccal swabs from 60 WV COAT clinic participants, prepared genomic DNA, and performed genotyping of the following SNPs.

OPRM1 rs9479757 (mu opioid receptor gene variant^39–41)

Xu et al.⁴⁰ identified rs9479757 as linked to heroin addiction severity among Han Chinese males; the G-to-A transition facilitates exon 2 skipping, leading to altered OPRM1 splice-variant mRNAs and encoded mu opioid receptor (MOR) isoforms. All sixty participants were heterozygous for OPRM1 rs9479757; in contrast, the minor allele frequency in the general European population is only 0.097 (Table 1).

OPRM1 rs1799971 (mu opioid receptor gene variant^41–62)

Woodcock et al.⁴⁶ published that Caucasian male carriers of the non-synonymous OPRM1 rs1799971 minor allele (G) reported significantly more heroin use-related consequences and quit attempts, and were more likely to seek OUD treatment, than individuals homozygous for the ancestral A allele. Schwantes-An et al.⁴⁸, in a meta-analysis of ~28,000 Europeans, investigated non-specific risk for SUD and compared cases dependent on any substance to controls not dependent on all assessed substances. The rs1799971 G allele showed a modest protective effect on general substance dependence⁴⁸. The rs1799971 G allele was significantly less frequent in WV COAT clinic participants than in the population at large, by a factor of ~2 (Table 1; 0.083 vs 0.162, p = 0.018).

RGS2 rs4606 (Regulator of G protein Signaling-2 gene variant^63–69)

One of us⁷⁵ previously reported that loss of RGS2 in mice engenders heightened anxiety and diminished aggression. Subsequently, Smoller et al.⁶³ found that RGS2 variations, including rs4606, were associated with introversion (a core personality trait in social anxiety disorder)⁶³. Associations between the RGS2 SNP rs4606 and panic disorder⁶⁸, suicidal ideation⁶⁴, post-traumatic stress disorder⁶⁹, and generalized anxiety disorder⁶⁵ have also been reported. We found that rs4606 was more frequent in COAT clinic volunteers than in the general population by a factor of ~2 (Table 1; 0.533 vs 0.276, p < 0.001).

HTR2B rs6736017 (serotonin [5-HT] receptor-2B gene^36–38 variant)

There are presently no published details on the HTR2B SNP rs6736017. However, HTR2B variations likely play a role in drug abuse. Doly et al.^{36, 37} showed that the serotonin 5-HT_2B receptor, encoded by the HTR2B gene, is required for 3,4-methylene-dioxymethamphetamine (MDMA)-induced hyperlocomotion, serotonin and dopamine release, and conditioned place preference, the latter a rodent-based metric of reinforcing properties. Furthermore, Lin et al.³⁸ conducted a genome-scan identifying the human HTR2B gene as a candidate for drug abuse liability. The HTR2B SNP rs6736017 was markedly more frequent in COAT clinic participants than in the population at large, when comparing to the 60,000 genomes of the ExAc database (Table 1; 0.500 vs 0.0078, p < 0.01).

None of the other 10 SNPs investigated exhibited a significantly different frequency in the COAT clinic volunteers versus the population at large (Table 1).

DISCUSSION

We identified four SNPs present at different frequencies in persons with OUD compared with a demographically-matched general population. It is currently difficult to quantify the degree of significance of each of these altered frequencies on OUD in the COAT population. One reason for this difficulty is that many gene variants are likely to play an integrated role in predisposition to OUD, and these variants may have additive or synergistic effects; here, we have only interrogated 14 SNPs in 8 candidate genes. Also, possible gene x environment effects were not assessed in the present study.

Two SNPs were in the OPRM1 gene encoding MOR -- the main molecular target for the euphoric effects of opioids, as well as for the therapeutic effects of buprenorphine^{7, 30} and methadone in OUD treatment. Therefore, it is not surprising that OPRM1 variations would be associated with OUD.

We also identified an RGS2 variant associated with OUD. This finding is novel, but it is also not surprising. Opioids like oxycodone and buprenorphine generate neuronal signals by activating opioid-binding GPCRs, altering receptor conformation to cause nucleotide exchange⁷⁶ by the associated G protein heterotrimer (Galpha·GDP/Gbeta/Ggamma). GTP-loaded G-alpha then dissociates from the Gbeta/gamma heterodimer; both subunits become free to modulate enzymes that generate intracellular second messengers (e.g., cyclic AMP, calcium) or stimulate potassium channels, leading to neuronal membrane hyperpolarization and thus action potential inhibition. G-alpha proteins have intrinsic GTPase activity, leading to self-inactivation and reassociation with Gbeta/gamma subunits⁷⁶. RGS proteins like RGS2 greatly enhance the GTP hydrolysis rate of G-alpha subunits^76–78, thereby inhibiting GPCR signaling. Thus, one might indeed expect that a polymorphism in an RGS gene – encoding a negative regulator of GPCR signaling -- would be implicated in OUD. Moreover, RGS2 function is particularly germane to multiple anxiety disorders^63–69 that could predispose an individual toward OUD.

HTR2B encodes the serotonin 5-HT_2B receptor, implicated in the rewarding actions of MDMA^{36, 37} and cocaine⁷⁹, and thought to be involved in impulsivity⁸⁰ and resistance to antidepressants⁸¹, conditions that could predispose to OUD. Other polymorphisms in HTR2B have also been associated with vulnerability to illegal drug use³⁸. Therefore, our identification of a HTR2B polymorphism significantly more frequent in persons with OUD compared with a demographically-matched general population is not surprising.

There are several goals for identifying SNPs associated with OUD. One of these goals is to identify markers that, in genetic screens, could identify people at risk for developing OUD (e.g., before prescribing opioid analgesics). Another goal is to identify potential novel “druggable” targets for preventing and/or treating OUD. For instance, the published role of the 5-HT_2B receptor in the actions of illicit drugs, coupled with our finding that rs6736017 is so frequent in OUD individuals, suggests that 5-HT_2B receptor-active pharmaceuticals (of which many are already approved for human use) might have efficacy in preventing/treating OUD. Future studies with larger cohorts are required to validate our findings, both in West Virginia and in other, more demographically heterogeneous regions. Such studies should identify genes to interrogate in order to predict individuals with susceptibility to OUD/SUD and in doing so provide an opportunity for the prevention of the disease.