Abstract
Background The OPRM1 A118G gene variant (N40D) encoding the µ-opioid receptor (MOR) has been associated with dependence on opiates and other abused drugs but its mechanism is unknown. With opioid abuse-related deaths rising at unprecedented rates, understanding these mechanisms may provide a path to therapy.
Methods Seven human induced pluripotent stem (iPS) cell lines from homozygous N40D subjects (4 with N40 and 3 with D40 variants) were generated and human induced neuronal cells (iNs) were derived from these iPS cell lines. Morphological, gene expression as well as synaptic physiology analyses were conducted in human iN cells carrying N40D MOR variants; Two pairs of isogenic pluripotent stem cells carrying N40D were generated using CRISPR/Cas9 genome-editing and iN cells derived from them were analyzed.
Results Inhibitory human neurons generated from subjects carrying N40D MOR gene variants show mature properties in morphological and functional analyses. Gene expression revealed that they express mature neuronal marker and MORs. Activation of MORs suppressed inhibitory synaptic transmission in human neurons carrying both N40 or D40 MOR variants but D40 show stronger effects. To mitigate the confounding effects of background genetic variation on neuronal function, the regulatory effects of MORs on synaptic transmission were validated in two sets of independently generated isogenic N40D iNs.
Conclusions Activations of N40D MOR variants show different regulatory effects on synaptic transmission in inhibitory human neurons. This study identifies neurophysiological differences between human MOR variants that may predict altered opioid responsivity and/or dependence in this subset of individuals.
INTRODUCTION
Well over 46,000 Americans died of opioid overdose in 2016, with the sharp increase in 2014 – 2016 due to synthetic opioids (1), prompting a public health crisis whose biological underpinnings are poorly understood. The µ-opioid receptor (MOR) mediates the most powerful addictive properties of abused opiate alkaloids and much research has identified chemically diverse ligands of varying efficacies for pain relief or treatment of addiction. Because of its substantive role in mediating reward and positive reinforcement, MOR is also an indirect target of alcohol, nicotine, and other drugs of abuse (2, 3). MOR-mediated synaptic alterations in reward-associated brain regions may represent a key underlying mechanism of reinforcement in drug abuse (4), but our understanding of this process in human neurons is limited.
Human genetic studies suggest that MOR gene variants play key roles in susceptibility to opioid addiction in humans. Most prominently, the A118G single nucleotide polymorphism (SNP) in OPRM1, rs1799971, a non-synonymous gene variant which replaces asparagine at position 40 (N40) with aspartate (D40), is found in up to 50% of individuals in certain ethnic groups and is associated with drug dependence phenotypes (5). There have been a number of investigations (5–13) into the functional consequences of the MOR D40 variant on receptor activation in overexpression models, knock-in mice, and primate models, but no systematic investigations into the functional and electrophysiological consequences of OPRM1 A118G have been reported, specifically not in a human neuronal context. Understanding how the D40 variant affects MOR signaling and synaptic function when expressed at normal levels in human neurons may provide insight into mechanisms underlying drug abuse, at least in people carrying this variant.
In order to fill the gap in studies done in the mouse and heterologous systems, we generated human induced neuronal (iN) cells from induced pluripotent stem (iPS) cells derived from subjects carrying homozygous alleles for either MOR N40 or D40 in order to better dissect the role of MOR N40D in a physiologically relevant and human-specific model system. Strikingly, we found that D40 MOR human neurons exhibit a stronger suppression of inhibitory synaptic release in the presence of MOR-specific agonist DAMGO ([D-Ala2, N-MePhe4, Gly-ol]-enkephalin) compared to N40 human neurons. In order to control for the possibility of individual genetic background variation between subject cell lines, we used CRISPR/Cas9 gene targeting to generate two sets of isogenic human stem cell lines: one pair with a 118GG knock-in into a well-characterized human embryonic stem (ES) cell line and the other by converting a minor allele carrier (118GG, D40) into a major allele carrier (118AA, N40). Remarkably, the synaptic regulations of MOR activation in the isogenic lines recapitulate those of neurons generated from different human subjects. This study exemplifies the use of patient-specific iPS cells as well as gene targeted isogenic stem cell lines to advance our understanding of the fundamental cellular and synaptic alterations associated with MOR N40D in human neuronal context.
METHODS AND MATERIALS
Generation of human iPS cells from lymphocytes of subjects carrying MOR N40D
Human iPS cell lines were generated by RUCDR Infinite Biologics ® from human primary lymphocytes carrying either MOR N40 or D40 genotypes using Sendai viral vectors (CytoTuneTM, ThermoFisher Scientific), as previously described (14). Human iPS cells were cultured and maintained as described previously (15).
Human iPS cell maintenance
Human iPS cells were cultured in 37°C, 5% CO2 on Matrigel® Matrix (Corning Life Sciences)-coated plates in mTeSR medium (Stem Cell Technologies). For passaging and differentiation done weekly, iPS cells were dissociated using Accutase (Stem Cell Technologies), spun down at 1000 RPM for 5 minutes, and replated at a density of 20,000 cells/cm2 for maintenance cultures and 50,000 cells/cm2 for differentiation.
Lentivirus preparation
Lentiviruses were produced in HEK293T cells by co-transfection of the three envelope proteins REV, RRE and VSVG vectors with 22µg of either FUW-Tet-O-Ascl1-T2A-puromycin, FUW-Tet-O-Dlx2-IRES-hygromycin, or FUW-rtTA. For each transfection, 9.1µg of REV, 13.77 µg VsVg, 19.1 µg RRE with 22g of lentiviral vector was transfected into a 150mm dish of HEK293T cells of 60% confluency using calcium phosphate transfection technique. Media was changed 12 hours following transfection, and virus was harvested in the media 48 hours following transfection, pelleted using an ultra-centrifuge (25,000 RPM for 2 hours), resuspended in MEM and aliquoted. Virus was stored in −80°C until use.
Generations of isogenic human stem cell lines carrying N40D MOR gene variants
Two pairs of isogenic N40D MOR human stem cells lines were generated using CRISPR/Cas9 genome editing. Briefly, to convert H1 embryonic stem (ES) cells carrying homozygous AA118 major allele to GG118 homozygous minor alleles, a sgRNA designed from Optimized CRISPR Design Tool (http://crispr.mit.edu/) and Cas9 were expressed using the PX459 vector (Addgene plasmid #62988) and was transfected using Lipofectamine 3000 reagent (ThermoFisher Scientific, L300015) along with a single stranded oligodeoxynucleotide (ssODN) of 140 base pairs with homology arms flanking the mutation site carrying mutations for G118, a BamHI restriction enzyme site for screening, along with a mutation to mutate the PAM sequence. Individual clones were hand-picked for expansion and screening by PCR and sequencing. Heterozygous clone 9-2 was expanded and transfected for targeting the second allele of OPRM1 Exon 1. The two homozygous G118 knock-in clones were further subcloned before expansion and freezing.
To convert rs1799971 in the 03SF subject iPS cell line from homozygous minor allele (GG) to major allele (AA), a slightly different strategy was used. First, a CRISPR targeting site was found using ZiFit software (16). The target site (GGCAACCTGTCCGACCCATG) included the major allele sequence so the gRNA was designed to incorporate the minor allele (GGCgACCTGTCCGACCCATG). A 200 nt homologous recombination donor oligo was designed to convert minor to major allele, inactivate the CRISPR site, and introduce a HpaI site for screening. The gRNA was synthesized by PCR and in vitro transcription (GeneArt Precision gRNA Synthesis Kit, Life Technologies) (17), mixed with synthetic Cas9 protein (Life Technologies), donor oligo, and the mixture was electroporated into iPS cells (Amaxa nucleofector, Lonza) along with a GFP expression plasmid (pGFP-Max, Lonza). One day later, cells were dissociated with Accutase and GFP-expressing cells were collected by FACS and plated at about 5,000 cells per well in a 6-well plate on irradiated MEFs. By 7-10 days, colonies were visible and hand-picked for screening. Three iPS cell clones were selected: C12, which had no evidence of editing to be used as a negative control; D11 and A10, which both had homozygous edits to produce rs1799971 major allele (AA). In all gene-targeted cell lines, sequencing confirmed these edits and that all predicted off-target sites were unchanged.
Generation of GABAergic iN cells from human ES and iPS cells
The protocol of generating GABAergic human iN cells was described recently (18). Briefly, iPS cells and ES cells were plated as dissociated cells on Matrigel ® Matrix (Corning Life Sciences)-coated dishes in mTeSR (Stem Cell Technologies) medium with 2µM Y-27632 (Stemgent). The following day, the cells are infected with Ascl1, Dlx2 and rtTA lentiviruses for 10-12 hours upon which culture medium was replaced with Neurobasal medium (GIBCO by Life Technologies) with B27 and L-Glutamine supplemented with 2µg/mL Doxycycline (MP Biomedicals) and 2µM Y compound to induce TetO expression. The protocol for generating lentiviruses expressing different transcription factors was previously described (18). Puromycin and Hygromycin selection was conducted for the following 2 days, and on day 5, the iN cells are dissociated with Accutase and plated on glass coverslips with a monolayer of passage three primary astrocytes isolated from p1-3 pups, as described previously (15, 18, 19). Following plating, 50% of the culture medium was replaced every 2-3 days with fresh Neurobasal media containing B27, L-Glutamine, 100 ng/ml of BDNF, NT3 and GDNF.
Real-time RT-PCR (qPCR)
Total neuronal RNA from three independently generated batches of iN cells for each cell line was prepared using TRIzol ® Reagent (Thermo Fisher Scientific), Human-specific Taqman probes were purchased for OPRM1, MAP2, Tuj1, VGAT, GAD1, TH and PCR reaction conditions followed the manufacturer’s recommendations. Undifferentiated iPS cells, ES cells, and mouse astrocytes were used as negative controls. A sample of total RNA of a healthy human brain as well as Human Thalamus from Biochain ® was used as a positive control. Relative RQ values were obtained by normalizing expression levels to the C12 iN condition. Student’s t-test was used to compare grouped N40 and D40 means.
Immunocytochemistry and confocal imaging
Inhibitory human neurons were fixed for 15 minutes in 4% paraformaldehyde in PBS and permeabilized using 0.1% Triton X-100 in PBS for 10 minutes at room temperature. Cells were then incubated in blocking buffer (4% bovine serum albumin with 1% normal goat serum in PBS) for 1 hour at room temperature and then incubated with primary antibodies diluted in blocking buffer for 1 hour at room temperature, washed with PBS three times, and subsequently incubated in secondary antibodies for 1 hour at room temperature. Confocal imaging analysis was performed using a Zeiss LSM700. Primary Antibodies used include: mouse anti Oct4 (Millipore Sigma MAB4401, 1:2000), mouse anti Tra-1-60 (Millipore Sigma MAB4360, 1:1000), mouse anti MAP2 (Sigma-Aldrich M1406, 1:500), rabbit anti MAP2, (Sigma-Aldrich M3696, 1:500), rabbit anti Synapsin (e028, 1:3000), rabbit anti VGAT (Millipore Sigma AB5062P, 1:2000), mouse anti Gad-67 (Abcam ab26116, 1:500), mouse anti β3 Tubulin (BioLegend 801201, 1:2000).
Electrophysiology
Functional analyses of iN cells were conducted using whole cell patch clamp as described elsewhere (15, 20). Briefly, a K-Gluconate internal solution was used, which consisted of (in mM): 126 K-Gluconate, 4 KCl, 10 HEPES, 4 ATP-Mg, 0.3 GTP-Na2, 10 Phosphocreatine. The pH was adjusted to 7.2 and osmolarity was adjusted to 270-290 mOsm. The bath solution consisted of (in mM): 140 NaCl, 5 KCl, 2 CaCl2, 2 MgCl2, 10 HEPES, 10 Glucose. The pH was adjusted to 7.4. Spontaneous inhibitory postsynaptic currents (sIPSCs) were recorded at a holding potential of 0mV under voltage-clamp mode. Miniature IPSCs were recorded in the presence of tetrodotoxin (1 µM). Intrinsic action potential firing properties of the iN cells were recorded in a bath solution containing 50 µM Picrotoxin and 20 µM CNQX. Evoked synaptic currents were elicited using an extracellular concentric bipolar stimulating electrode positioned approximately 100 µm away from the cell soma. All recordings were performed at room temperature. Data presentation: All data are presented as mean ± S.E.M. Student’s t-test or 2-way ANOVA were used to assess statistical significance.
RESULTS
Generation of human inhibitory neurons carrying N40D MOR variants
To investigate the functional role of the MOR N40D variant in a human neuronal context, we obtained iPS cells from multiple individuals of European descent carrying homozygous alleles for either MOR N40 (n=4) or MOR D40 (n=3) (Supplemental Fig. 1A). The rs1799971 genotype and the pluripotency of all seven iPS cell lines are confirmed by sequencing and colocalized immunocytochemistry (ICC) for OCT4 and Tra-1-60 (Fig. 1A-B).
We derived inhibitory induced neuronal (iN) cells from all 7 iPS cell lines by lentiviral mediated ectopic expression of the transcription factors Ascl1 and Dlx233. These induced human neuronal cells express pan-neuronal makers including MAP2, β3-tubulin, and Synapsin (Fig. 1C, Supplemental Fig. 1B) as well as inhibitory neuronal markers GAD67 and VGAT (Fig. 1D, Supplemental Figs. 1C). Thus, the N40D SNP has no impact on MOR expression or inhibitory neuronal identity. To examine whether the N40 and D40 iN cells are functionally comparable under baseline conditions, we performed whole cell patch-clamp recordings of iN cells after 5-6 weeks of re-plating onto a monolayer of mouse astroglia. The iN cells of both genotypes exhibit similar intrinsic membrane excitability (Supplemental Fig. 1D-F) and can fire repetitive spontaneous action potentials (APs) at baseline levels (Supplemental Fig. 1G-I) and exhibit similar intrinsic excitability under baseline conditions (Fig. 1H-I). Similarly, no significant differences in spontaneous or miniature inhibitory postsynaptic currents (sIPSCs and mIPSCs, respectively) were observed by genotype (Figs. 1E-G, Supplemental Fig. 1J-L), indicating that the N40D variant does not affect passive or active membrane properties and that the neurons generated from subject iPS cell lines are of similar functional maturation and differentiation.
MOR D40 iN cells exhibit altered sensitivity to the MOR agonist DAMGO
There have been numerous studies7, 22-24, 28, 36-39 examining the functional consequences of MOR N40D on receptor activation in overexpression models and in knock-in mice harboring MOR N40D, but no functional or electrophysiological analyses on cultured neurons have been conducted, specifically not in a human neuronal context. To gauge whether N40 and D40 iN cells may respond differently to MOR activation, we used a MOR-specific agonist DAMGO ([D-Ala2, N-MePhe4, Gly-ol]-enkephalin) to study its role modulating synaptic release. In both N40 and D40 iN cells, DAMGO suppressed sIPSCs in a dose-dependent manner (Fig. 1K). However, the suppression of sIPSC frequency was more robust in D40 iN cells compared to N40 iN cells in multiple repeated experiments and multiple iPS cell lines (Fig. 1L), with no difference in sIPSC amplitude by genotype. To confirm that the observation is not due to a residual effect of prolonged agonist exposure, we applied a single concentration of 10µM DAMGO (Fig. 1M-N) and similarly found that D40 iN cells respond more robustly to MOR activation compared to N40 iN cells, illustrating genotype-dependent regulation of MOR signaling.
Generation of isogenic human pluripotent stem cell lines carrying MOR N40D SNP
To directly compare the two MOR genotypes in identical genetic backgrounds, eliminating the impact of secondary genetic variation, we generated two sets of isogenic human stem cell lines using CRISPR/Cas9 gene targeting. We first targeted the MOR locus in a well-characterized human H1 ES cell line, which carries only major allele (A118), using an sgRNA targeting the antisense DNA strand along with a 140bp single stranded oligodeoxynucleotide carrying G118 (Fig. 2A-B). Simultaneously, we converted one iPS cell line with the homozygous G118 allele to a homozygous A118 genotype with an alternative strategy utilizing direct transduction of guide RNA and Cas9 protein into the subject cell line (Fig. 2C-D). We isolated two clones (Supplemental Fig. 2A-C) with no detectable off-target effects from each targeting scheme. Inhibitory iN cells generated from isogenic lines stain positive for MAP2, Synapsin and VGAT (Fig. 2E-F) and exhibit similar intrinsic membrane properties (Supplemental Fig. 2D-E) as well as sIPSC and AP properties in both genotypes (Supplemental Figs. 2F-G, 3A-C). The densities and sizes of synapses were also not significantly different between genotypes (Supplemental Fig. 2H-I), illustrating that the N40D SNP does not alter synaptogenesis or functional maturation in the isogenic human neurons, and that the MOR N40D SNP has no consequence on iN cell maturation or synaptic transmission at baseline levels.
Isogenic human neurons recapitulate differential DAMGO response phenotype and exhibit altered synaptic function
In this highly controlled system of isogenic iN cells, we observed less culture-to-culture variability than the subject cell lines for OPRM1 mRNA and inhibitory neuronal markers (Fig. 2G-I). Furthermore, we observed a similar decrease in sIPSC frequency compared to subject iN cells following acute DAMGO application, with a stronger inhibition in D40 versus N40 iN cells, and no effect on amplitude (Fig 2J-O). Furthermore, to determine whether the effect of DAMGO was mediated by MOR, we applied Naltrexone, a broad spectrum MOR antagonist, and found that the DAMGO-induced synaptic suppression could be reversed (Supplemental Fig. 2J-K). Thus, the reproducibility of the DAMGO response phenotype illustrates that the D40 variant alone explains the differential signaling and it is not due to secondary genomic variation.
We focused the remaining analyses on one pair of isogenic cell lines, C12 and A10, on the basis of their consistent differentiation and maturation. We found that DAMGO application more robustly decreases mIPSC frequency in D40 versus N40 iN cells, which no change in mIPSC amplitude (Fig. 3A-C), which suggests DAMGO mediated decrease in synaptic release. Consistent with the decrease in mIPSC frequency, we observed that DAMGO decreases evoked IPSC amplitude more robustly in D40 iN cells compared to N40 iNs (Fig. 3D-E). This is consistent with the hypothesis that MOR activation by DAMGO in human iN cells more robustly decreases neurotransmitter release probability in D40 iN cells compared to N40 iN cells, suggesting that the A118G SNP directly regulates synaptic function.
D40 MOR-expressing neurons exhibit a more robust decrease in excitability following DAMGO compared to N40 iN cells
To understand whether the decreased synaptic release is compounded by decreased intrinsic excitability, we examined the effect of DAMGO on induced AP firing in N40 and D40 iN cells. We observed that 10 µM DAMGO induced D40 versus N40 iN cells to fire significantly fewer APs (Fig. 4A-B), with no effect on AP amplitude, firing threshold (Fig. 4C-D) or other properties including Time to reach peak or threshold (not shown). This is supported by an immediate and more robust decrease in spontaneous AP firing frequency following DAMGO application in D40 versus N40 iN cells (Fig. 4E), an effect which is sustained over the course of several minutes (Fig. 4F). This sustained decrease in AP frequency is paralleled by a rapid hyperpolarization of N40 and D40 iN cells (Fig. 4G). This effect was found to be significantly more robust in D40 versus N40 iN cells in the first minute following DAMGO application. The immediate drop in AP firing frequency and membrane potentials in iN cells suggests that this may be occurring through a G-protein mediated signaling mechanism, which is activated immediately following agonist binding (21). No differences in AP rise time, decay time or half width were detected by DAMGO application (not shown). However, we observed a slight trending increase in the after hyperpolarization potential in the D40 versus N40 iN cells (Fig. 4H-I), with no significant difference in firing threshold or AP half width (Fig. 4J-K). These data indicate the functional differences between the two genotypes are at least partly mediated by a preferential decrease in excitability in D40 versus N40 iN cells, likely mediated by alterations in the G-protein coupled signaling cascade. Overall, these data suggest that DAMGO-induced decrease in excitability is superimposed by a synapse-specific effect, i.e., a stronger reduction in synaptic release probability mediated by presynaptic MOR at the nerve terminal in D40 MOR inhibitory neurons.
DISCUSSION
Our study provides the first experimental evidence detailing the electrophysiological consequences of the N40D SNP on MOR activation in its endogenous human neuronal context. First, we generated iN cells from human subject-derived stem cells carrying homozygous alleles for N40 MOR or D40 MOR and found that D40 MOR expressing iN cells exhibit stronger inhibitory effects of MOR activation on synaptic release. Second, to validate the functional consequences of the SNP in a system highly controlled for background genetic variation, we used CRISPR/Cas9 mediated gene targeting to: 1) knock-in homozygous D40 alleles into H1ES cells; 2) correct the homozygous D40 alleles in 03SF iPS cell subject line into N40 alleles, and thus generated two sets of isogenic stem cell lines for highly controlled mechanistic analyses, The isogenic iN cells not only recapitulated the DAMGO response phenotype of the patient iN cells, but also revealed that the N40D SNP mediates a more robust decrease in excitability and synaptic release.
Despite previous studies in knock-in mouse models and heterologous expression systems, the precise molecular and cellular consequences of MOR N40D have remained unclear, primarily due to species-specific and context-specific mechanisms in the modulation of MOR signaling. For instance, rodent models and other expression systems have suggested that the D40 allele confers a “gain-of-function” effect by causing increased potency for DAMGO and other MOR agonists (7, 11, 12, 22). However, subsequent studies have reported that the D40 allele is associated with reduced mRNA and protein expression in multiple brain regions of knock-in mice (10, 13) along with reduced antinociceptive responses to morphine (8), providing support for a “loss-of-function” phenotype. These contradictory results strongly necessitate the need for a human neuronal model to understand MOR function.
The novelty of our approach using human iN cells to investigate the synaptic pathology of addiction is that these cells carry the genetic signatures of the subjects from whom they were derived. Identifying identical DAMGO-mediated responses across multiple subject-derived and CRISPR-edited cell lines generated using independently executed targeting strategies clearly demonstrates that the observed effect is a direct consequence of only the MOR N40D variant. Collectively, this overall approach of combining multiple patient lines with genome-engineered isogenic lines to assess addiction-associated electrophysiological phenotypes has not been fulfilled in previous work. Thus, we show the utility of disease modeling using stem cell derived disease-relevant cell types as a framework to the field of addiction for conducting future mechanistic analyses.
This study represents a significant advance in our understanding of the neurobiological mechanisms underlying the human N40D MOR variant in a human neuronal context. Our study provides direct evidence that common genetic variation encodes functional variation at the level of synaptic transmission. The use of patient-derived stem cells to unravel the impact of OPRM1 gene variants may ultimately provide the necessary insight to develop patient-specific, precision medical interventions for drug and alcohol dependence.
DISCLOSURES
The authors declare they have no competing financial interests.
ACKNOWLEDGMENTS
We thank RUCDR Infinite Biologics for generating the iPS cells from human subjects and assisting with CRISPR/Cas9 gene targeting on 03SF iPS cell line. Research is supported by grants from NIH-NIAAA R01 AA023797 as well as Collaborative Studies on the Genetics of Alcoholism/COGA 5U10AA008401-26. AH is supported by NIH-NIAAA NRSA F31AA024033. We are grateful to the members of the Collaborative Genetic Study of Nicotine Dependence (COGEND) for the selection of human subjects, and we are grateful to the de-identified individuals who contributed tissue to the study.