COMT Val158Met Polymorphism, Cardiometabolic Risk, and Nadir CD4 Synergistically Increase Risk for Neurocognitive Impairment in Men Living with HIV

Abstract

Objective:

The Val allele of the Val158Met single-nucleotide polymorphism of the catechol-o-methyltransferase gene (COMT) results in faster metabolism and reduced bioavailability of dopamine (DA). Among persons living with HIV (PLWH), Val carriers display neurocognitive deficits relative to Met carriers, presumably due to exacerbation of HIV-related depletion of DA. COMT may also impact neurocognition by modulating cardiometabolic function, which is often dysregulated among PLWH. We examined the interaction of COMT, cardiometabolic risk, and nadir CD4 on neurocognitive impairment (NCI) among HIV+ men.

Methods:

329 HIV+ men underwent COMT genotyping and neurocognitive and neuromedical assessments. Cohort-standardized z-scores for body mass index, systolic blood pressure, glucose, triglycerides, and high-density lipoprotein cholesterol were averaged to derive a cardiometabolic risk score (CMRS). NCI was defined as demographically-adjusted global deficit score≥0.5. Logistic regression modelled NCI as a function of COMT, CMRS, and their interaction, covarying for estimated premorbid function, race/ethnicity, and HIV-specific characteristics. Follow-up analysis included the 3-way interaction of COMT, CMRS, and nadir CD4.

Results:

Genotypes were 81 (24.6%) Met/Met, 147 (44.7%) Val/Met, and 101 (30.7%) Val/Val. COMT interacted with CMRS (p=0.02) such that higher CMRS increased risk of NCI among Val/Val (OR=2.13, p<.01), but not Val/Met (OR=0.93, p>.05) or Met/Met (OR=0.92, p>.05) carriers. Among Val/Val, nadir CD4 moderated the effect of CMRS (p<.01) such that higher CMRS increased likelihood of NCI only when nadir CD4<180.

Discussion:

Results suggest a tripartite model by which genetically-driven low DA reserve, cardiometabolic dysfunction, and historical immunosuppression synergistically enhance risk of NCI among HIV+ men, possibly due to neuroinflammation and oxidative stress.

Introduction

Combination antiretroviral therapy (cART) prolongs life expectancies for persons living with HIV (PLWH); however, neurocognitive impairment (NCI) remains highly prevalent^1,2. It is important to identify discrete neuropathological mechanisms that predict NCI in PLWH given that NCI can translate to negative and costly everyday functioning outcomes, including unemployment and poor cART adherence^3,4. However, identifying these mechanisms has proven challenging^5–7 because PLWH are a heterogeneous group whose neurocognition may be directly impacted by HIV, but also by comorbidities and genetic predispositions that enhance vulnerability to neural injury^6,8,9.

PLWH are at greater risk of cardiometabolic complications (i.e., hypertension, dyslipidemia, diabetes, and obesity) possibly due to HIV-related accelerated biological aging and iatrogenic consequences of cART^10,11. Cardiometabolic risk contributes to NCI by triggering an array of neuro-compromising processes, including increased blood-brain-barrier permeability, chronic inflammation, and endothelial dysfunction^12–16. Furthermore, cardiometabolic risk factors predict imaging biomarkers of white matter and neurochemical abnormalities among PLWH^17,18. While cardiometabolic risk factors increase risk of NCI in successfully treated PLWH, they are neither necessary nor sufficient to predict NCI. Thus, susceptibility to the contribution of cardiometabolic risks may be moderated by individual differences in resilience against HIV-related CNS dysfunction.

One factor that could partially explain NCI among PLWH is dopamine (DA) bioavailability. HIV results in exposure to neurotoxic proteins Tat and gp120 that damage frontostriatal regions rich in DA^19,20. Post-mortem studies demonstrate decreased frontostriatal concentrations of DA and downregulated gene expression of DA receptors in HIV-infected versus HIV-uninfected persons, and that dysregulation of DA pathways relates to greater HIV disease severity (e.g., HIV RNA level, nadir CD4) and neurocognitive deficits among PLWH^21–25.

The erosion of DA homeostasis in PLWH has prompted research on catechol-o-methyltransferase (COMT)⁶, a functionally diverse enzyme responsible for metabolism of DA, particularly in the prefrontal cortex (PFC). Neurons exposed to macrophage-propagated HIV show increased mRNA expression of COMT and decreased expression of neuronal and synaptic proteins, suggesting that higher levels of COMT in HIV may contribute to DA and neurocognitive dysfunction. In further support of this theory, treatment of HIV-exposed neurons with a COMT inhibitor, Tolcapone, effectively reduced COMT expression and restored neuronal and synaptic integrity²⁶. The Val158Met (rs4680) single nucleotide polymorphism (SNP) of the COMT gene encodes differential levels of COMT enzymatic activity, with the Met allele resulting in 40% less activity, and therefore greater DA bioavailability, than the Val allele²⁷. The Met allele has been linked to enhanced neural activation and neurocognitive functioning in PLWH^28,29, possibly due to resilience against HIV-related depletion of DA.

Some literature suggests that COMT also operates within cardiometabolic pathways. COMT catalyzes the methylation of catechol estrogens into 2-methoxyestradiol³⁰, which reduces risk of diabetes, hypertension, and obesity^31–34. However, COMT activity produces a precursor to homocysteine, which may increase risk of cardiovascular disease, and Val carriers have significantly higher blood levels of total homocysteine than Met homozygotes^35–37. Furthermore, lower central DA tone has been linked to obesity and markers of metabolic syndrome, and DA receptor neurotransmission is reduced in obese humans and animals^38–40. While greater COMT activity associated with the Val allele may protect against cardiometabolic disorders through increased production of 2-methoxyestradiol, reduced COMT activity associated with the Met allele may also confer protection against cardiometabolic disorders by enhancing DA tone and limiting the accumulation of homocysteine. Although the influence of COMT on cardiometabolic health is complex, these mechanisms suggest that COMT may serve a multifactorial role in the neuropathogenesis of HIV-related NCI through the modulation of cardiometabolic risk and DA bioavailability.

Our first aim is to examine the independent and interactive effects of COMT and cardiometabolic risk on NCI. Given the putative neuroprotective effect of the Met allele in PLWH, we hypothesize that Val/Val carriers with high cardiometabolic risk will have higher rates of NCI compared to others because they are most vulnerable to the deleterious impact of cardiometabolic risk factors on the brain. Because NCI has been more reliably linked to nadir versus current CD4 count in the cART era^1,41–43, we will also examine the role of nadir CD4 in the conditional relationships between COMT, cardiometabolic risk, and NCI. We hypothesize that lower nadir CD4 counts will increase likelihood of NCI regardless of COMT and cardiometabolic risk, but will make a larger contribution to NCI in Val/Val carriers with high cardiometabolic risk.

Methods

Participants and Procedure

Participants were 329 HIV+ men who underwent COMT genotyping through NIH-funded research studies coordinated by the HIV Neurobehavioral Research Program (HNRP) at the University of California, San Diego. Of the 329 participants, 76 were enrolled in the Translational Methamphetamine AIDS Research Center (TMARC), a NIDA-funded HNRP cohort study focusing on the central nervous system effects of HIV and methamphetamine. The other 253 participants were enrolled in the CNS HIV Anti-Retroviral Therapy Effects Research (CHARTER) study. All studies were approved by local Human Subjects Protection Committees and all participants provided written informed consent. Exclusion criteria were: 1) diagnosis of psychotic or mood disorder with psychotic features, neurological or medical condition that may impair neurocognitive functioning, such as traumatic brain injury, stroke, epilepsy, hepatitis C, or advanced liver disease; 2) lifetime diagnosis of methamphetamine or cocaine use disorder; 3) low verbal IQ as estimated by a Wide Range Achievement Test⁴⁴ (WRAT) reading subtest score <70; 4) evidence of intoxication by positive urine toxicology for illicit drugs (except marijuana) or Breathalyzer test for alcohol on the day of testing; and 5) being female. We restricted our sample to men because sexually dimorphic effects of COMT on brain function have been reported^45,46 and there were insufficient numbers of women participants in the parent studies to support separate analyses. Lifetime cocaine or methamphetamine use disorder, even remote, was exclusionary in order to eliminate potential confounding effects of stimulant-induced alteration of dopaminergic signaling.

Neuromedical Assessment

All participants underwent a comprehensive neuromedical assessment and non-fasting blood draw. Detailed history of medical and antiretroviral (ARV) use was collected and ARV treatment status was coded as currently on, past use, or never used. HIV infection was diagnosed by enzyme-linked immunosorbent assay with Western blot confirmation. Routine clinical chemistry panels, complete blood counts, rapid plasma reagin, hepatitis C virus antibody, and CD4+ T cell count (flow cytometry) were performed at each site’s certified clinical laboratory. HIV viral load in plasma and CSF were measured using reverse transcriptase-polymerase chain reaction (Amplicor, Roche Diagnostics, Indianapolis, IN), with a lower limit of quantitation (LLQ) of 50 copies/ml). HIV viral load was dichotomized as detectable vs. undetectable at the LLQ of 50 copies/ml.

Cardiometabolic Risk Assessment

Non-fasting levels of serum glucose, plasma triglycerides, and plasma total, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol levels were assayed by standard protocols at each site’s clinical laboratory. Blood pressure (BP) was measured in the seated position with an automated sphygmomanometer and body mass index (BMI) was calculated from measured weight and height. Our primary predictor of interest was a continuously scaled, composite cardiometabolic risk score (CMRS)⁴⁷ based on the five components of metabolic syndrome outlined by the Adult Treatment Panel⁴⁸ including obesity, elevated BP, elevate blood glucose, elevated triglycerides, and reduced (HDL) cholesterol. Cohort-standardized z-scores for BMI, systolic BP, glucose, triglycerides, and HDL cholesterol (inverse polarity) were averaged to generate our CMRS. A CMRS was derived from 4 of the 5 component z-scores for participants who were either missing HDL (n=50) or triglycerides (n=4) values.

COMT Genotyping

For participants enrolled in TMARC, DNA for genotyping was isolated from stored whole blood or peripheral blood mononuclear cells (PBMCs) using the Qiagen QIAamp DNA Mini Kit (Qiagen, Valencia, CA). COMT Val158Met (rs4680) SNP was assayed using an array that included SNPs associated with catecholaminergic genes⁴⁹. For participants enrolled in CHARTER, DNA for genotyping was extracted from PBMCs using PUREGENE (Gentra Systems, Inc, Minneapolis, MN). All samples were genotyped using the Affymetrix Genome-Wide Human SNP Array 6.0TM⁵⁰.

Neurocognitive Assessment

All participants completed a comprehensive and standardized neurocognitive assessment across seven neurocognitive domains commonly impacted by HIV^1,51. Test scores were adjusted for known demographic influences (i.e., age, education, and race/ethnicity) on neurocognitive performance^52–54. Deficit scores that give differential weight to impaired over normal performance were calculated for each domain and averaged to derive a global deficit score (GDS) ranging from 0 (normal) to 5 (severe). Consistent with prior studies, neurocognitive status was classified as impaired (NCI) vs. unimpaired using a validated cut-point of GDS ≥ 0.5^51,55 The GDS is easier to compute, more clearly operationalized (e.g., Frascati criteria do not specify how to apply a 1 SD cutoff to define “impairment” of ability domains with variable numbers of measures) and a more conservative approach to classifying NCI as compared to the clinical ratings algorithm used in Frascati criteria for HIV-associated Neurocognitive Disorders; however, an individual classified as impaired via GDS ≥ 0.5 is essentially guaranteed to meet the NCI aspect of Frascati criteria⁵⁵.

Psychiatric Assessment

Current mood symptoms were assessed using the Beck Depression Inventory (BDI) version one or two⁵⁶. The computer-based Composite International Diagnostic Interview (CIDI)⁵⁷ was administered to determine DSM-IV diagnoses of current and lifetime substance use disorders (SUD) and Major Depressive Disorder (MDD).

Statistical Analysis

COMT group differences in demographics, HIV disease, neuropsychiatric, cardiometabolic, and neurocognitive variables were examined using ANOVAs, Kruskal-Wallis tests, and Chi-square statistics as appropriate. To follow-up on significant omnibus results, pair-wise comparisons were conducted using Tukey’s Honest Significant Difference (HSD) tests for continuous outcomes or Bonferroni-corrections for categorical outcomes⁵⁸. Cohen’s d statistics are presented for estimates of effect size for statistically significant pair-wise differences. Logistic regression examined the univariate relationship between CMRS and NCI.

Next, we used multivariate logistic regression to model our NCI classification as a function of COMT, CMRS, and their interaction. COMT genotype was reference coded⁵⁹ with the high enzymatic activity Val/Val group as the reference. WRAT, race/ethnicity, nadir CD4, current CD4, plasma viral load detectability, and ARV status were entered as covariates because they either significantly differed across COMT genotype or are known to influence neurocognition in the cART era^1,60. A 3-way interaction term between nadir CD4*COMT*CMRS, as well as accompanying lower-order terms, was added in a follow-up model in order to explore the potential moderating effect of HIV-induced immunosuppression on the contributions of COMT and CMRS on NCI. Group differences and logistic regression analyses were performed using JMP Pro version 12.0.1 (JMP^®, Version <12.0.1>, SAS Institute Inc., Cary, NC, 1989–2007).

Exploratory analyses, stratified by COMT genotype, employed the Johnson-Neyman (J-N) technique^61,62 to compute any specific boundaries of nadir CD4 at which CMRS significantly predicted NCI. These boundaries are referred to as regions of significance. Compared to simple slope analyses that describe the effect of a predictor (i.e., CMRS) at fixed levels of a continuous moderator (i.e., nadir CD4), the J-N technique identifies the full range of moderator values for which the predictor slope is statistically significant. Region of significance analyses adjusted for false discovery rate⁶³ and were computed using the jtools package in R statistical software (version 3.4.4, R Foundation for Statistical Computing, Vienna, Austria).

Results

Participants

Table 1 presents COMT group differences on demographic and clinical characteristics. COMT distribution across the 329 participants (age: M=44.0, SD=8.97, education: M=14.1, SD=2.34) was 81 (24.6%) Met/Met, 147 (44.7%) Val/Met, and 101 (30.7%) Val/Val. Genotype distribution was consistent with Hardy-Weinberg equilibrium in the full sample χ² (χ²=3.49, p=0.06) and within each race/ethnicity group (ps>.11). However, genotype frequency differed significantly by race/ethnicity (χ²= 11.24, p=0.02) with non-Hispanic White participants more likely to carry a Met allele than non-Hispanic Black participants (χ²=8.75, p=0.003). COMT groups were comparable across most demographic, psychiatric, and HIV disease characteristics, with the exception of estimated premorbid verbal IQ for which Met/Met displayed significantly higher WRAT scores than Val/Val (d=0.40, p=0.007). Most participants experienced cART-induced immune reconstitution, as evidenced by active ARV use (76%) and markedly higher current CD4 counts (median=458 cells/mm³) compared to nadir CD4 counts (median=180 cells/mm³). Half the sample (51%) had detectable levels of plasma viral RNA at a limit of detection of 50 copies per ml.

Table 1.

Demographic and clinical characteristics by COMT genotype (N=329)

Variable	Met/Met (n=81)	Val/Met (n=147)	Val/Val (n=101)	p
Demographics
Age (years)	45.1 (9.69)	43.7 (9.09)	43.6 (8.16)	0.44
Education (years)	14.3 (2.54)	14.1 (2.39)	13.9 (2.1)	0.54
Estimated Verbal IQ (WRAT)	104.4 (11.41)	101 (11.3)	99.8 (11.76)	0.02a
Ethnicity				0.02a
Non-Hispanic White (n=245)	66 (81%)	110 (75%)	69 (68%)
Non-Hispanic Black (n=33)	3 (4%)	12 (8%)	18 (18%)
Hispanic (n=51)	12 (15%)	25 (17%)	14 (14%)
HIV Disease Characteristics
AIDS diagnosis	53 (65%)	89 (61%)	57 (56%)	0.47
Duration of HIV infection (years)	10.2 (7.83)	9.5 (6.44)	9.2 (6.3)	0.60
Current CD4 count (cells/mm3)	517 (255.5)	470 (261.1)	457 (261.5)	0.27
Nadir CD4 count (cells/mm3)	180 [63.5–305]	175 [44–297]	175 [55–297.5]	0.83
Plasma viral load
Copies/ml (log10)	1.8 [1.7–3.3]	1.9 [1.7–3.7]	2.3 [1.7–3.9]	0.45
Detectable	42 (52%)	69 (47%)	57 (56%)	0.33
CSF viral load
Copies/ml (log10)b	1.7 [1.7–1.7]	1.7 [1.7–2.2]	1.7 [1.7–2.4]	0.21
Detectableb	11 (18%)	34 (29%)	24 (29%)	0.24
ARV status				0.36
Currently on	67 (83%)	105 (71%)	77 (76%)
Past use only	7 (9%)	16 (11%)	10 (10%)
ARV naïve	7 (9%)	26 (18%)	14 (14%)
Current regimen type				0.21
PI-based	31 (46%)	54 (51%)	42 (55%)
NNRTI-based	27 (40%)	38 (36%)	30 (39%)
PI/NNRTI-based	4 (6%)	11 (10%)	5 (6%)
Other	5 (7%)	2 (2%)	0 (0%)
Duration of current regimen (months)	12 [3–33]	13 [4–27]	12 [4–33]	0.96
Neuropsychiatric characteristics
Lifetime any substance use disorder	38 (47%)	68 (46%)	51 (51%)	0.80
Alcohol	35 (43%)	57 (38%)	42 (42%)	0.79
Cannabis	16 (20%)	21 (14%)	18 (18%)	0.54
Opioid	0 (0%)	1 (1%)	0 (0%)	0.99
Major Depressive Disorder
Lifetime	44 (54%)	72 (49%)	51 (51%)	0.74
Current	15 (19%)	19 (13%)	12 (12%)	0.41
BDI score	11.8 (9.57)	12.6 (10.14)	11.2 (9.32)	0.52

Note. Values presented as mean (SD), median [IQR], or N (%). WRAT= Wide-Range Achievement reading subtest; ARV= antiretroviral therapy; PI= Protease-inhibitor; NNRTI= non-nucleoside reverse transcriptase inhibitor; BDI= Beck Depression Inventory

^aSignificant difference between Met/Met and Val/Val

^bCSF viral load values available for a subset of participants: Met/Met (n=60), Val/Met (n=118), Met/Met (n=82)

COMT, Cardiometabolic Risk, and NCI

Table 2 presents COMT group differences on cardiometabolic and neurocognitive variables. Met/Met had significantly higher diastolic blood pressure than Val/Met (d=0.37, p=0.02) but not Val/Val. COMT groups did not differ significantly on any other cardiometabolic risk parameters, including the composite CMRS (F=0.09, p=0.91). Similarly, COMT groups did not differ significantly on GDS (F=0.17, p=0.84) or frequency of NCI (χ2=2.39, p=0.30), with rates ranging from 32% (Val/Met) to 42% (Val/Val). Although greater CMRS increased likelihood of NCI, this relationship was also not significant (OR=1.12, 95%CI [0.89–1.40], p=0.34).

Table 2.

Cardiometabolic risk and neurocognitive performance by COMT genotype

Variable	Met/Met (n=81)	Val/Met (n=147)	Val/Val (n=101)	p
Cardiometabolic parameters
Cardiometabolic risk score	0.01 (0.489)	−0.02 (0.548)	−0.01 (0.467)	0.91
Body mass index (kg/m2)a	26.3 (4.42)	25.8 (4.37)	25.4 (4.46)	0.43
Systolic blood pressure (mm Hg)a	127 (14.13)	124.4 (15.41)	124.6 (14.76)	0.41
Diastolic blood pressure (mm Hg)a	79.3 (9.68)	75.8 (8.98)	76.2 (9.66)	0.02b
Total cholesterol (mg/dL)	196.8 (54.11)	184.6 (40.68)	192.2 (41.24)	0.12
High-density lipoprotein (mg/dL)a	41.8 (15.94)	42.2 (15.71)	44.5 (22.52)	0.59
Low-density lipoprotein (mg/dL)	111.8 (37.63)	100.2 (35.16)	106.8 (35.22)	0.10
Triglycerides (mg/dL)a	217.6 (149.22)	221.3 (140.91)	228.8 (158.63)	0.87
Glucose (mg/dL)a	94 [86–105.5]	92 [81–107]	92 [83–107.5]	0.87
Neurocognitive Performance
Global deficit score	0.51 (0.52)	0.49 (0.58)	0.47 (0.48)	0.84
Neurocognitive impairmentc	29 (36%)	47 (32%)	42 (42%)	0.30

Note. Values presented as mean (SD), median [IQR], or N (%).

^aCardiometabolic risk score component

^bSignificant difference between Met/Met and Val/Met

^cNeurocognitive impairment defined as global deficit score ≥ 0.5

COMT and CMRS Interaction

Table 3 presents estimates for the multivariate logistic regression modelling NCI as a function of COMT, CMRS, and NCI, covarying for WRAT scores, race/ethnicity, and HIV disease characteristics. The overall model was significant (χ²(13,329)=392.68, p=0.001). A significant omnibus interaction between COMT and CMRS was detected (χ²(2,329)=7.93, p=0.02) such that higher CMRS levels significantly increased likelihood of NCI among Val/Val (OR=2.13, p<.01), yet the deleterious effect of CMRS on NCI was significantly attenuated in Met/Met (interaction of CMRS × Met/Met [compared to Val/Val]: ORR=0.42, p<.05) and Val/Met (interaction of CMRS × Val/Met [compared to Val/Val]: ORR=0.43, p<.05). Specifically, CMRS did not significantly predict NCI among Met/Met (OR=0.92, p>.05) or Val/Met (OR=0.93, p>.05) carriers (Figure 1). Lower WRAT, lower nadir CD4, and detectable HIV RNA significantly increased probability of NCI (ps<.05).

Table 3.

Multivariable logistic regression predicting neurocognitive impairment

Predictor	B (SE)	p	ORa	OR 95% CI
Met/Metb	−0.15 (0.34)	0.664	0.87	[0.45, 1.67]
Val/Metb	−0.42 (0.29)	0.156	0.66	[0.37, 1.17]
Cardiometabolic risk scorec	0.76 (0.28)	0.007	2.13	[2.13, 3.68]
Cardiometabolic risk score × Met/Met	−0.87 (0.39)	0.025	0.42	[0.19, 0.90]
Cardiometabolic risk score × Val/Met	−0.83 (0.33)	0.012	0.43	[0.23, 0.83]
WRATd	−0.37 (0.11)	0.001	0.69	[0.55, 0.86]
Nadir CD4e	−0.13 (0.03)	0.017	0.88	[0.79, 0.98]
Detectable plasma viral load	0.59 (0.29)	0.040	1.80	[1.03, 3.16]
Current CD4e	0.03 (0.05)	0.338	1.05	[0.95, 1.11]
ARV status never used	−0.26 (0.47)	0.574	0.77	[0.31, 1.93]
ARV status past usef	0.15 (0.44)	0.736	1.16	[0.49, 2.72]
Non-Hispanic Blackg	−0.37 (0.44)	0.398	0.69	[0.29, 1.63]
Hispanich	0.30 (0.34)	0.374	1.36	[0.69, 2.64]

Note. Bolded predictors are significant at p<.05. ARV= antiretroviral therapy; OR= odds ratio; WRAT= Wide Range Achievement Test

^aInteraction terms reflect odds rate ratios

^bCompared to Val/Val

^cPer 1 standard deviation change

^dPer 10-unit change

^ePer 50-unit change;

^fCompared to ARV status currently using

^gCompared to non-Hispanic White

Figure 1.

Greater cardiometabolic risk predicts neurocognitive impairment among Val/Val only

Conditional Role of Nadir CD4

To determine whether HIV disease severity influenced the interactive effects of COMT and CMRS on NCI, we expanded our model with terms capturing the 3-way interaction between nadir CD4, COMT, and CMRS. A significant 3-way interaction was detected between nadir CD4, COMT, and CMRS (χ²(2,329)=8.86, p=0.01). Nadir CD4 did not significantly alter the null associations between CMRS and NCI among Val/Met and Met/Met. In Val/Val, nadir CD4 significantly moderated the deleterious effect of CMRS on NCI (for 50-unit increase: ORR=0.53, 95%CI [0.32, 0.81], p=0.008). In order to inspect changes in the slope of CMRS on NCI due to different nadir CD4 counts, we applied the J-N technique. In Val/Val, CMRS significantly increased likelihood of NCI (i.e., lower bound of CMRS slope>0) at nadir CD4 counts below 180 (Figure 2). Conversely, CMRS did not significantly predict NCI in Val/Val for nadir CD4 counts at or above 180.

Figure 2.

Greater cardiometabolic risk significantly increases likelihood of neurocognitive impairment in Val/Val with nadir CD4 below 180

To focus on a clinically relevant subgroup, we applied the 3-way interaction model in participants who were currently using cART and had undetectable HIV RNA (n=155). Results in this virologically suppressed subgroup did not differ from those in the entire study sample, with COMT, CMRS, and nadir CD4 significantly interacting to predict NCI (χ²(2,150)=10.25, p=0.006).

Discussion

Dopaminergic dysregulation, cardiometabolic dysfunction, and low nadir CD4 counts (i.e., advanced HIV-induced immunosuppression) are putative risk factors for NCI in the context of HIV. In contrast to most neuroAIDS studies that examine independent effects of risk factors on NCI, this study also examined interactions of cardiometabolic risk, COMT, and nadir CD4 in order to address the complex interplay of these risk factors. Although neither COMT nor CMRS correlated univariably with NCI, higher CMRS significantly increased probability of NCI in Val/Val individuals. Furthermore, CMRS effects on neurocognitive function in Val/Val carriers were moderated by nadir CD4: higher CMRS increased probability of NCI only in participants with nadir CD4 counts below 180. These findings remained significant in patients with fully-suppressed HIV on cART, underscoring the relevance of these genetic and environmental risk factors for NCI for even the most successfully treated PLWH.

Our group has previously reported better executive function among Met/Met, compared to Val-carriers, in HIV+ men²⁹. In contrast, the Multicenter AIDS Cohort Study found no main effects of COMT or interactive effects with stimulant use and current CD4 on neurocognitive performance in predominantly male (87%−100%) samples^64,65. In our study, COMT did not univariably relate to frequency of NCI, which, consistent with prior GDS-based estimates of NCI in PLWH^51,55, was 36% across the entire study sample. We may not have detected a univariable relationship between COMT and NCI because individual SNPs often exhibit modest associations with behavioral phenotypes⁶⁶. Given that COMT pleiotropically influences multiple neurobiological processes, the study of moderating environmental variables can help explain under which conditions the neurocognitive effects of COMT are most salient.

Our results support a gene × environment interaction approach, as the putatively harmful effect of the Val allele only appeared under environmental conditions of high cardiometabolic burden and low nadir CD4 counts. Similar interactive effects of COMT and cardiovascular risk have been previously reported in healthy adults, with Val carriers demonstrating steeper declines in episodic memory compared to Met/Met individuals only at elevated levels of pulse pressure⁶⁷. While the effects of cardiometabolic comorbidities on poor neurocognition are not unique to PLWH, cardiometabolic abnormalities may be particularly disruptive in PLWH given their high prevalence and potential to combine with other HIV-related brain insults to synergistically deplete cognitive reserve⁶⁸. Although studies of cardiometabolic risk often dichotomize individual cardiometabolic conditions based on thresholds of clinical laboratory values⁶⁹, we modelled cardiometabolic risk as a standardized, continuous variable. This approach is statistically advantageous because continuous predictors preserve power and precision⁷⁰. Moreover, thresholds for dichotomizing cardiometabolic conditions are imprecisely defined and overlook incremental relationships between cardiovascular disease outcomes (e.g., stroke) and health indicators (e.g., BP)⁷¹. Importantly, continuous cardiometabolic risk scores inversely relate to physical activity^47,72 and therefore, provide further evidence to advocate exercise in its neurocognitive benefits for PLWH.

With respect to conditional effects of HIV-induced immunosupression, Levine et al.⁶⁵ did not detect interactions between current CD4 count and DA-related genes, including COMT, on neurocognitive performance. However, as the authors note, current CD4 may be a suboptimal indicator of HIV disease severity given that nadir CD4 is a stronger predictor of NCI and brain integrity in the cART era^1,41,73. Furthermore, lower nadir CD4, but not current CD4, is correlated with reduced DA transporter availability in the ventral striatum²⁵. Our results indicating that lower nadir CD4 counts independently predict higher odds of NCI are consistent with the prior findings of Ellis et al.⁴¹ that describe a monotonic relationship between lower nadir CD4 counts and higher odds of NCI, even in PLWH with viral suppression and minimal-to-moderate comorbidity burden. Importantly, we also demonstrate that nadir CD4 moderates the interactive effects of COMT and CMRS. The clinical impact of these findings is highlighted by the substantial portion (55%) of Val/Val participants with nadir CD4 counts at or below 180, a range at which our region of significance analysis indicates a heightened vulnerability to the deleterious neurocognitive effects of cardiometabolic risk. In a prior investigation that implemented machine learning to generate predictive models for NCI, the predictive utility of viral load and duration of treatment were improved when nadir CD4 was less than 225⁷⁴, a threshold similar to that of the present study.

Some hypothesize that the relationship between cortical function and DA bioavailability follows an inverted U-shaped curve, with optimal neurocognition occurring at intermediate levels of DA signaling^75,76. Within this framework, severe immunosuppression in PLWH may lead to persistently suboptimal levels of bioavailable DA even after successful immune reconstitution with cART. The slow rate of DA clearance conferred by the Met allele may protect against this shift, whereas the neurocognitive deficits present in Val/Val PLWH with high cardiometabolic burden may reflect an inability to compensate for reduced cognitive resources associated with DA dysfunction.

Although limitations in our data prevents us from investigating biological mechanisms that may underlie the interactive effects of COMT, CMRS, and nadir CD4 on NCI, we offer several plausible neurobiological interpretations. Chronic neuroinflammation is a hallmark feature of HIV-related CNS dysfunction⁷⁷. Immunosuppression, cardiovascular disease, and metabolic syndrome are associated with increased levels of pro-inflammatory biomarkers in the periphery and in CSF, as well as neuroinflammation on magnetic resonance spectroscopy in PLWH^{11,17,78–80}. Limited bioavailability of DA and epinephrine due to high enzymatic activity of COMT may also result in poor neuroinflammatory regulation because catecholamines play a pivotal role in modulating lymphocyte and inflammasome activity^81,82. In addition to neuroinflammatory dysregulation, excessive formation of reactive oxygen species and endothelial dysfunction are neurotoxic processes associated with immunosuppression, cardiometabolic risk, and catecholamine metabolism^83–89. While our data reflect the neurobehavioral consequences of genetically-driven low DA signaling, poor cardiometabolic health, and more severe HIV disease, studies at a cellular level are needed to elucidate the mechanisms of their interactions.

We acknowledge several limitations to this study. First, the cross-sectional and associative nature of our data prevent us from drawing causal inferences, especially about the relationship between NCI and CMRS in the context of genetic and HIV disease history predictors. Second, we lacked an HIV- uninfected comparison group, hindering our ability to determine the specificity of our findings to PLWH. However, by exploring the conditional effect of nadir CD4 with the J-N technique, we demonstrate how HIV disease severity in a clinically-relevant and well-represented range (i.e., nadir CD4 <180) can modulate the effects of genetic and environmental risk factors for NCI. Third, our results focused on clinically informative risk factors for NCI that may inform future interventions targeting the maintenance of cardiometabolic health (e.g., exercise, diet, sleep) and dopaminergic therapy (e.g., COMT-inhibitors), but our sample comprised men only and therefore limits the applicability of our findings to women. Fourth, we studied the effects of only one variant of COMT, but not other DA-related genetic variants on NCI in PLWH. As is the case with theoretically-driven single SNP analyses, COMT may be in linkage disequilibrium with other SNPs in the DA pathway that could account for the presumed effects of COMT. “Omics” techniques that leverage rich, multi-leveled data (e.g., genetic, transcriptomic, epigenetic) may advance the characterization of complex and subtle HIV-related neurobiological phenotypes⁶. Given that multiple components of the DA signaling pathway (e.g., receptors, metabolic enzymes, transporters) can be quantified across multiple levels of analysis (e.g., host genetics, microRNA expression, imaging), an “omics” approach has the potential to uncover clusters of DA-related factors that track with intermediate biological phenotypes as well as disease and neurobehavioral endpoints (e.g., NCI).

Taken together, our findings suggest a tripartite model by which genetically-driven low DA reserve, cardiometabolic dysfunction, and history of immunosuppression synergistically enhance risk of NCI among HIV+ men. These interactive effects remain significant in virally suppressed participants receiving cART, suggesting that currently effective HIV treatment alone may not sufficiently mitigate neurocognitive dysfunction associated with the combination of genetic vulnerabilities, historically severe HIV-induced immunosupression, and cardiometabolic risk factors. In addition to cART, adjunctive behavioral therapies that target modifiable lifestyle factors could be considered as a means to address this diverse combination of risk factors⁹⁰. In particular, interventions to increase physical activity and/or reduce sleep disturbances may help stabilize DA function and cardiometabolic health, reduce inflammatory and oxidative stress burden, and subsequently ameliorate persisting neurocognitive deficits in HIV-infected patients^91–95.