Cannabis Use Associates With Reduced Proviral Burden and Inflammatory Cytokine in Tissues From Men With Clade C HIV-1 on Suppressive Antiretroviral Therapy

Abstract

Background

Human immunodeficiency virus 1 (HIV-1) tissue reservoirs remain the main obstacle against an HIV cure. Limited information exists regarding cannabis’s effects on HIV-1 infections in vivo, and the impact of cannabis use on HIV-1 parenchymal tissue reservoirs is unexplored.

Methods

To investigate whether cannabis use alters HIV-1 tissue reservoirs, we systematically collected 21 postmortem brain and peripheral tissues from 20 men with subtype C HIV-1 and with suppressed viral load enrolled in Zambia, 10 of whom tested positive for cannabis use. The tissue distribution and copies of subtype C HIV-1 LTR, gag, env DNA and RNA, and the relative mRNA levels of cytokines IL-1β, IL-6, IL-10, and TGF-β1 were quantified using PCR-based approaches. Utilizing generalized linear mixed models we compared persons with HIV-1 and suppressed viral load, with and without cannabis use.

Results

The odds of tissues harboring HIV-1 DNA and the viral DNA copies in those tissues were significantly lower in persons using cannabis. Moreover, the transcription levels of proinflammatory cytokines IL-1β and IL-6 in lymphoid tissues of persons using cannabis were also significantly lower.

Conclusions

Our findings suggested that cannabis use is associated with reduced sizes and inflammatory cytokine expression of subtype C HIV-1 reservoirs in men with suppressed viral load.

cannabis use, subtype C HIV-1 tissue reservoirs, proviral DNA, inflammatory cytokines

Although antiretroviral therapy (ART) can suppress human immunodeficiency virus (HIV) plasma viral load (pVL) to undetectable levels (aviremia) and people with HIV (PWH) are living a more normal life [1], it is not a cure as ART cessation will lead to viral rebound [2]. This rebound is derived from integrated HIV-1 DNA (provirus) in different tissues—viral reservoirs [3, 4]. Most reservoir studies have focused on subtype B HIV-1, and central nervous system (CNS) and multiple peripheral tissues were documented as potential tissue reservoirs [5–8]. Few reservoir studies have focused on subtype C HIV-1, which predominates in sub-Saharan Africa and India and accounts for nearly 50% of global infection, with the premise that all HIV-1 subtypes infection would parallel findings from subtype B [9, 10]. Utilizing postmortem tissues from people with suppressed pVL enrolled in Zambia, we have demonstrated that subtype C DNA is variably detected in peripheral tissues and, unlike subtype B, it poorly accesses the CNS [11].

Currently, little is known about the impact of cofactors, including substance abuse, on the establishment and maintenance of HIV-1 reservoirs. Cannabis (Cannabis sativa) is widely consumed in the general population and PWH [12, 13]. Cannabis use prevalence in sub-Saharan Africa appears to be approximately 12% for adults and just under 8% for adolescents. Notably, regional differences exist in the prevalence of cannabis use, with a high prevalence of cannabis use reported in South Africa (16.7% [ranging from 9.1% to 26.0%]) and Zambia (36.5% [ranging from 34.3% to 38.7%]) [14–17]. Of the > 100 cannabinoids in C. sativa, Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD) are the primary psychoactive and nonpsychoactive ingredients, respectively [18]. To date, the effects of cannabis on HIV-1 have only been evaluated for subtype B infections and remained controversial. Some have reported cannabis either had no effect [19, 20], or detrimental effects [21, 22], or lower pVL in cannabinoid exposure [23], and reduced inflammation in experimental macaques [24–26]. Notably, most of these studies utilized subtype B-infected blood samples to study cellular reservoirs; whether cannabis use impacts the size of tissue reservoirs is unknown. Moreover, the impact of cannabis use in subtype C reservoirs is unexplored.

In this study, we systematically quantified HIV-1 LTR, gag, env DNA and RNA in 21 postmortem CNS and peripheral tissues from 20 men with subtype C HIV-1 and with suppressed pVL, 10 of whom were identified as persons using cannabis (THC+). We evaluated the effects of cannabis use on subtype C HIV-1 reservoir distribution, viral burden, and expression. Compared to men without cannabis use (THC−), THC+ men have significantly lower odds of a given tissue harboring subtype C HIV-1 proviral DNA, and lower proviral DNA burdens in tissues. Viral RNA was also expressed in fewer tissues in THC+ men. Moreover, in THC+ men, the relative mRNA levels of proinflammatory cytokines interleukin 1β (IL-1β) and IL-6 were also significantly lower in the tissues analyzed. Our findings suggest that there are potential beneficial effects of cannabinoids in reducing the subtype C HIV-1 tissue reservoir size and expression of proinflammatory mediators in men with suppressed pVL.

METHODS

Study Cohort and Tissue Collections

Deceased persons were consented by the next of kin for the postmortem collection of tissue specimens at the mortuary of the University Teaching Hospital in Zambia. Sociodemographic information as well as available medical history were collected whenever possible. Eventually, 381 deceased Zambians were enrolled and autopsied within 48 hours of their deaths. Different brain and peripheral tissues of these 381 persons were collected and cryopreserved at −80°C or processed to formalin-fixed, paraffin-embedded blocks. During autopsy, blood samples were also obtained from the carotids or the hearts of all persons. In cases of severe anemia, blood samples were obtained from the femoral vessels.

Screening of Persons With Suppressed pVL and With Cannabis Use

Postmortem plasma samples of the deceased were first utilized to determine HIV-1 serological status using HIV Rapid Test. For HIV-1 seropositive persons, plasma VL were quantified in triplicates using the RNA UltraSense One-Step reverse transcription quantitative polymerase chain reaction (RT-qPCR) System (Invitrogen) with AcroMetrix HIV-1 standard (Thermo Fisher Scientific) with HIV-1 LTR primers and probe (Supplementary Table 1). The cutoff of HIV-1 LTR RNA detection in our system was 70 copies/mL. In this study, PWH with pLV lower than 70 copies/mL were identified as persons with suppressed pVL. Plasma samples were also used in laboratory testing for 11-nor-9-carboxy-Δ9-tetrahydrocannabinol (THC-COOH), a stable main secondary metabolite of THC, using the Cannabinoid (THCA/CTHC) Direct enzyme-linked immunosorbent assay (ELISA) Kit (Immunalysis) to validate the information about cannabis use from surrogate interviewees (next of kin). Eventually, 10 male PWH with suppressed pVL and cannabis use (HIV+/THC+) were identified according to both ELISA and next-of-kin reports. To investigate the potential impacts of cannabis use on HIV-1 tissue reservoirs in persons with suppressed pVL, 10 age-matched male PWH with suppressed pVL who tested negative and were not reported by next of kin to use cannabis (HIV+/THC−) were included in this study as controls.

Quantifications of HIV-1 DNA/RNA in Different Tissues

For HIV-1 DNA quantification, 100 ng genomic DNA was used as template utilizing QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific). Genomic 8E5 cellular DNA, containing single proviral HIV-1 genome/cell, was used as standards. To determine viral RNA loads, 500 ng total RNA and AcroMetrix HIV-1 standards were used with the UltraSense One-Step RT-qPCR System (Invitrogen). Amplicons against HIV-1 LTR, gag, and env genes used for DNA copy determination were also used for viral cDNA (RNA) quantification (Supplementary Table 1).

Quantification of Inflammatory Cytokine mRNA Levels by SYBR Green qPCR

To identify the relative transcription levels of inflammatory cytokines in tissues, total RNA extracted from frozen tissues were used to synthesize first-strand cDNA using Invitrogen SuperScript III First-Strand Synthesis System (ThermoFisher), with 25 ng of cDNA as template in subsequent SYBR Green qPCR. PrimePCR SYBR Green Assays, using Bio-Rad primer pairs against inflammatory cytokine cDNAs (IL-1β assay ID, qHsaCID0022272; IL-6 assay ID, qHsaCID0020314; IL-10 assay ID, qHsaCED0003369; and transforming growth factor-β (TGF-β1) assay ID, qHsaCID0017026). Primers against the internal reference, GAPDH mRNA were also used (Supplementary Table 1). The Δcycle threshold value (ΔCt) value was calculated (cytokine Ct value − GAPDH CT value), and when the cytokine mRNA level Ct value was over 40, the ΔCt value was set as 99.

Statistical Analysis

Generalized linear mixed models (GLMMs) were used to compare THC+ and THC− persons. Fixed effects were included for cannabis use and tissue location, while random effects were used for each participant. Negative binomial families were assumed for predicting the copies of viral DNA and RNA, while binomial families were assumed for predicting the presence of any viral DNA or RNA. The functions glmer.nb and glmer from the lme4 package in R statistical software were used for the analyses. Confidence intervals (CIs) and associated P values were Wald-based, and no multiple correction adjustment was used. To tests whether there were THC differences within each specific tissue (no repeated measures), Firth corrected logistic regression (via the logistf function) was used for testing presence of any viral DNA/RNA while negative binomial (via the nb.glm function) was used for analyzing the counts directly. Firth correction was used here to correct for the separation that was observed in several tissue locations. The Mann-Whitney test was utilized in cytokine ΔCt comparisons. All tests were 2-tailed and P values < .05 were considered as significant.

RESULTS

Identification of THC+ Persons With Subtype C HIV-1 and Suppressed pVL

Figure 1A shows the experimental design and postmortem sampling. In this study, 381 deceased Zambians were enrolled and autopsied within 48 hours of their deaths. Among those persons, 82 persons with “undetectable” pVL (< 70 copies/mL) were identified. Among them 10 men (U1–U10) were identified as persons using cannabis via plasma THC metabolite detection (THC+), and 10 age-matched men with suppressed pVL (N1–N10) from the other 72 THC− persons with suppressed pVL were used as controls [11]. Eight CNS tissues (frontal lobe, parietal lobe, temporal lobe, occipital lobe, hippocampus, cerebellum, basal ganglia, and choroid plexus), and 13 peripheral tissues, as well as plasma sample, were collected from each person (Figure 1B).

Figure 1.

Postmortem sampling of persons with suppressed pVL, with or without cannabis use. A, Flow diagram showing the postmortem screening of persons with/without cannabis use. Plasma RNA preparations from HIV-1 seropositive persons were subjected to standard RT-qPCR to identify persons with suppressed pVL (< 70 copies/mL). The detection of THC-COOH by ELISA in plasma from persons with suppressed pVL was used as the criterion to group persons with cannabis use (THC+; U1–U10) and without cannabis use (THC−; N1–N10). HIV-1 subtyping by sequencing of env DNA confirmed subtype C HIV-1 infection. Viral DNA and RNA copies were determined by different PCR-based methods from frozen tissue genomic DNA and total RNA. The relative mRNA levels of different inflammatory cytokines were also measured in different tissues utilizing SYBR Green qPCR. B, Postmortem samples were collected within 48 hours of death. Multiple brain and peripheral tissues, and plasma samples were systematically collected. All samples were cryopreserved. This figure was made using BioRender. Abbreviations: ELISA, enzyme-linked immunosorbent assay; HIV-1, human immunodeficiency virus 1; pVL, plasma viral load; RT-qPCR, reverse transcription quantitative polymerase chain reaction; THC, Δ9-tetrahydrocannabinol.

Information including age, sex at birth, and ART treatment of our cohort are shown in Table 1. To confirm subtype C HIV-1 infection, genomic DNA was extracted from frozen tissues of each person and subjected to nested PCR to amplify the HIV-1 env gene. Amplified HIV-1 env products were sequenced to determine the HIV-1 subtypes using REGA HIV Subtyping Tool. HIV-1 env sequences were obtained from 17 persons with suppressed pVL, but not from 3 THC+ persons U4, U5, and U9 due to exceptionally low viral burden. These 17 persons with env sequences were demonstrated to harbor subtype C HIV-1 (Table 1). However, given that subtype C is responsible for approximately 99% of HIV infections in Zambia, persons U4, U5, and U9 are likely to be also infected with subtype C HIV-1.

Table 1.

Information of the Studied Autopsied Persons in This Study

Person ID	Age, y	Sex at Birth	THC Status	PVL, Copies/mL	Age at ART Initiation, y	ART Duration, y	ART Regimena	Postmortem Interval, h	HIV-1 Subtype
U1	28	M	THC+	<70	Unknown	Unknown	Atripla	5	C
U2	60	M	THC+	<70	Unknown	Unknown	Atripla	27	C
U3	59	M	THC+	<70	Unknown	Unknown	Atripla	40	C
U4	28	M	THC+	<70	Unknown	Unknown	Atripla	15	NAb
U5	25	M	THC+	<70	Unknown	Unknown	Atripla	45	NAb
U6	20	M	THC+	<70	Unknown	Unknown	Atripla	27	C
U7	28	M	THC+	<70	Unknown	Unknown	Atripla	14	C
U8	52	M	THC+	<70	Unknown	Unknown	Atripla	17	C
U9	35	M	THC+	<70	32	3	Atripla	12	NAb
U10	50	M	THC+	<70	Unknown	Unknown	Atripla	27	C
N1	40	M	THC−	<70	32	8	Atripla	30	C
N2	60	M	THC−	<70	52	8	Atripla	7	C
N3	45	M	THC−	<70	Unknown	Unknown	Atripla	3	C
N4	50	M	THC−	<70	40	10	Atripla	4	C
N5	53	M	THC−	<70	48	5	Atripla	< 48	C
N6	40	M	THC−	<70	36	4	Atripla	26	C
N7	47	M	THC−	<70	Unknown	Unknown	Atripla	18	C
N8	39	M	THC−	<70	34	5	Atripla	33	C
N9	36	M	THC−	<70	Unknown	Unknown	Atripla	18	C
N10	25	M	THC−	<70	Unknown	Unknown	Atripla	6	C

Abbreviations: ART, antiretroviral therapy; HIV-1, human immunodeficiency virus 1; NA, not available; PVL, plasma viral load; THC, Δ9-tetrahydrocannabinol;

^aAtripla is a combination of 3 different ART medications (efavirenz, emtricitabine, and tenofovir disoproxil fumarate) in single-tablet form.

^bNA due to the low viral burden in these persons.

Detectable HIV-1 DNA in Various Tissues of THC+ Persons With Suppressed pVL

Eight of 10 THC− persons with suppressed pVL (N1–N8) had already been previously characterized for viral tissue distribution and viral burden [11]. To identify the tissue distribution and viral burden of subtype C reservoirs in THC+ persons with suppressed pVL, HIV-1 LTR DNA copies in tissues of 10 THC+ men were quantified (Figure 2). Moreover, to identify which tissues might harbor potentially intact HIV-1 provirus, gag and env DNA were also quantified in LTR DNA-positive tissues.

Figure 2.

Heatmap of subtype C HIV-1 DNA in THC+ (U1–U10) and THC− (N1–N10). persons with suppressed pVL. The heatmap displays the abundance of 3 subtype C HIV-1 genes (LTR, gag, and env) DNA in different anatomical locations of 10 THC+ and 10 THC− men with suppressed pVL based on qPCR analysis. Blocks with color indicate viral copy numbers; undetectable are in blue to highest DNA copies in red (up to 1538 copies/1 × 10⁶ cells). The spleen of person N7 was not sampled (blank cells). Three subgroups of tissues are shown. HIV-1 DNA copies were determined by qPCR with 100 ng genomic DNA as template. All samples were analyzed in triplicate qPCR reactions, and viral DNA copy numbers were calculated as the means of triplicate PCR and normalized to 1 million cells. Abbreviations: GI, gastrointestinal; HIV-1, human immunodeficiency virus 1; LN, lymph node; pVL, plasma viral load; qPCR, quantitative polymerase chain reaction; THC, Δ9-tetrahydrocannabinol.

For the THC+ persons, HIV-1 DNA burden varied among different tissues and persons, and viral DNA was rarely detected in the brain tissues. Specifically, out of 80 total brain tissues from 10 THC+ persons, HIV-1 DNA was only detected in 12 tissues (15%), and only 3 brain tissues harbored LTR, gag, and env DNA (3.8%). Viral burdens in all brain tissues were lower than 100 copies/10⁶ cells, supporting our previously finding that brain is not a good reservoir for subtype C HIV-1 [11] (Figure 2 and Supplementary Table 2). In contrast, HIV-1 DNA was detected in 40 of 130 peripheral tissues (30.8%), and 6 of them harbored all 3 viral genes (Figure 2). However, the highest viral LTR DNA copy in THC+ tissues were quite low at 100 copies/10⁶ cells, as detected in the inguinal lymph node of person U2 (Figure 2 and Supplementary Table 3). In addition, the tissue distributions of detectable provirus also varied among the THC+ persons (Figure 2).

Fewer Tissue Samples Harboring Subtype C Provirus in THC+ Persons

To investigate the potential impact of cannabis use on subtype C HIV-1 reservoirs, we compared the tissue distribution of HIV-1 DNA between THC+ and THC− groups using a GLMM. After accounting for individual dependence and tissue location, THC+ persons had significantly lower odds of a tissue harboring HIV-1 DNA (Figure 3A). Specifically, tissues of THC+ persons showed 0.16 times, or approximately 6-fold lower odds of harboring HIV-1 LTR DNA (95% CI, .08–.32) and 0.07 times or approximately 14-fold lower odds of harboring all 3 LTR/gag/env DNAs (95% CI, .03–.17) (Figure 3A).

Figure 3.

Comparisons of tissue distribution of subtype C HIV-1 DNA between THC+ and THC− persons with suppressed pVL. Adjusted OR of harboring HIV-1 DNA in tissues for THC+ versus THC− groups were calculated from mixed logistic regression models with random effects for person ID and fixed effects for specific locations. A, Overall comparisons of THC effects across all tissues. Significantly lower odds of tissues harboring any proviral DNA (LTR DNA) or all 3 proviruses (LTR/gag/env DNA) in THC+ persons compared to THC− persons. B, Comparisons of THC effects within 3 tissue subgroups. In THC+ persons, tissues from lymphoid and GI tissues and other tissues subgroups, but not from brain tissues subgroup, had significantly lower odds of containing LTR DNA and all 3 viral DNAs compared to their counterparts in THC− persons. C, Comparison of THC effects within individual tissues. Forest plots showing the differences in the odds of presence of LTR DNA in tissues in subgroups lymphoid and GI tissues and other tissues between THC+ and THC− persons. Tissues are shown in order based on their P values, which were calculated from univariable logistic regression models (no repeated measurements within 1 tissue type). A–C, In all plots 95% CI and P values are shown. *P < .05, **P < .01, ***P < .001. Abbreviations: CI, confidence interval; GI, gastrointestinal; HIV-1, human immunodeficiency virus 1; LN, lymph node; OR, odds ratios; pVL, plasma viral load; THC, Δ9-tetrahydrocannabinol.

To better understand how this difference might be related to different organ/tissue systems, we stratified 21 tissues into 3 subgroups: (1) 8 CNS tissues were classified as brain tissues; (2) the 3 types of lymph nodes (LNs), bone marrow, spleen, ileum, and appendix were classified as lymphoid and gastrointestinal (GI) tract tissues; and (3) the remaining 6 peripheral tissues were calssified as the other tissues subgroup. The GLMM results showed that comparing THC+ to THC− persons, the odds of presence of any proviruses (LTR DNA) or potential intact proviruses (LTR/gag/env DNA) were significantly lower in lymphoid and GI tract tissues and other tissues subgroups, whereas the brain tissues subgroup showed no significant differences (Figure 3B). Furthermore, we also analyzed THC-dependent distribution differences of LTR DNA at single-tissue level using Firth corrected logistic regression. Here, there was only 1 observation per patient. Axillary LN, ileum, appendix, liver, testis, inguinal LN, and bone marrow showed significant differences, whereas belly fat, spleen, kidney, mesenteric LN, pancreas, and lung tissues did not (Figure 3C). For example, the odds of harboring LTR DNA in THC+ axillary LNs were 0.02 times or approximately 50-fold lower than that in THC− axillary LNs, and this difference is significant (P = .001). Although the odds of harboring LTR DNA in THC+ spleen tissues was 0.25 times or approximately 4-fold lower than that in THC− spleen tissues, this difference is not significant as the P value is .181, which is > .05 (Figure 3C).

Lower Viral Burden of Subtype C Reservoir in THC+ Persons

We also tested for significant magnitude of proviral burden (HIV-1 DNA copies) differentials between THC+ and THC− persons. Detectable DNA copies ranged from 4 to 1538 copies/10⁶ cells in THC− persons, and from 4 to 100 copies/10⁶ cells in THC+ persons (Figure 2, and Supplementary Tables 2 and 3). The GLMM results showed that THC+ persons had significantly lower copies of viral LTR, gag, and env DNA (Figure 4A). At the tissue subgroup level, copies of LTR DNA or env DNA were significantly lower in lymphoid and GI tract tissues and other tissues subgroups in THC+ persons, while the brain tissues subgroup showed no significant differences (Figure 4B). However, all peripheral tissues in THC+ persons were shown to harbor a lower level of viral LTR DNA compared to the THC− persons (Figure 4C). Specifically, the copies of LTR DNA in THC+ appendix tissues were 0.02 times or approximately 50-fold lower than that in THC− appendix tissues, and this difference is significant (P < .001). Moreover, the copies of LTR DNA in THC+ ileums were 0.08 times or approximately 12-fold lower than that in THC− ileums, and this difference is also significant (P = .011; Figure 4C).

Figure 4.

Comparisons of the levels of viral burden between THC+ and THC− persons with suppressed pVL. ME of THC+ versus THC− groups on the copies of viral DNA were calculated from mixed negative binomial regression models with random effects for individual ID and fixed effects for specific locations. A, Overall comparisons of THC effects across all tissues. Significantly lower copies of proviral LTR DNA (P < .001), gag DNA (P < .001), env DNA (P < .001) in THC+ persons compared to THC− persons. B, Comparisons of THC effects within 3 tissue subgroups. In THC+ persons, tissues from lymphoid and GI tissues and other tissues subgroups, but not from brain tissues subgroup, had significantly lower copies of LTR DNA and env DNA compared to their counterparts in THC− persons. C, Comparison of THC effects within individual tissues. Forest plot showing the differences in the levels of viral burden in tissues from subgroups lymphoid and GI tissues and other tissues between THC+ and THC− persons. Tissues are shown in order based on their P values, which were calculated from univariable negative binomial regression models (no repeated measurements within 1 tissue type). A–C, In all plots, 95% CI and P values are shown. **P < .01, ***P < .001. Abbreviations: CI, confidence interval; GI, gastrointestinal; ME, multiplicative effect; pVL, plasma viral load; THC, Δ9-tetrahydrocannabinol.

Few Tissues Express Viral RNA in THC+ Persons

To detect and quantify THC-dependent differences in viral RNA expression in different tissues, HIV-1 LTR, gag, and env RNA copies were quantified. In THC− persons, HIV-1 RNA was detected in 8 of 10 persons (80%). Viral RNA was detectable primarily in the lymphoid and GI tract tissues, sporadically in the other tissues, but not in the brain tissues [11] (Figure 5). Specifically, viral LTR RNA was detectable in 30.7% proviral DNA-positive tissues (35 of 114 tissues) with copy number ranging from 13 to 1609 copies/500 ng total RNA (Figure 5 and Supplementary Table 4). For THC+ persons, LTR RNA was only detected in 4 persons, compared to 8 in the THC− group. Moreover, viral RNA was only detected in the lymphoid and GI tract tissues subgroup and only 7 tissues expressed detectable viral RNA (Figure 5), with the copy number ranging from 52 to 278 copies/500 ng total RNA, suggesting a decrease in RNA expression compared to THC− group (Figure 5 and Supplementary Table 5).

Figure 5.

Heatmap of subtype C HIV-1 RNA in THC+ and THC− persons with suppressed pVL. A heatmap displaying viral RNA transcripts (LTR, gag, and env) abundance as determined by RT-qPCR analysis in 10 THC+ and 10 THC− persons with suppressed pVL. Blocks with color indicate viral copy numbers; undetectable are in blue to highest DNA copies in red (up to 1566 copies/500 ng total RNA). Blank indicates sample was not available. HIV-1 RNA copies were identified by RT-qPCR with 500 ng input RNA as template. All samples were analyzed in duplicate RT-qPCR reactions. Abbreviations: GI, gastrointestinal; HIV-1, human immunodeficiency virus 1; LN, lymph node; pVL, plasma viral load; RT-qPCR, reverse transcription quantitative polymerase chain reaction; THC, Δ9-tetrahydrocannabinol.

Although the GLMM results revealed that the odds of tissues expressing LTR RNA in THC+ persons were significantly lower than that in THC− persons (P = .041) (data not shown), the odds of tissues expressing gag or env RNA could not be analyzed because they were undetectable for the THC+ group (Figure 5).

Lower Levels of Relative mRNA of Proinflammatory Cytokines in the Lymph Node and Appendix Tissues of THC+ Persons

We then investigated the expressions of proinflammatory cytokines, IL-1β and IL-6, and anti-inflammatory cytokines, IL-10 and TGF-β1 in the HIV+/THC− and HIV+/THC+ groups. Ten HIV−/THC− persons comprised the control group. Cytokine mRNA levels were quantified from axillary lymph node and appendix, 2 tissues that showed lower viral burden in THC+ persons than in THC− persons (Figure 4). Relative mRNA levels of the 4 cytokines in axillary lymph node showed no significant differences between HIV−/THC− controls and HIV+/THC− persons. However, the axillary lymph nodes from the HIV+/THC+ group had significantly lower IL-1β and IL-6, but equivalent IL-10 and TGF-β1 relative mRNA levels (Figure 6A) when compared to the HIV+/THC− group. Similarly for appendix tissues, there were significantly lower relative mRNA levels of IL-1β and IL-6, but similar mRNA levels of IL-10 and TGF-β1 in the HIV+/THC+ group (Figure 6B). We also analyzed the cytokine mRNA levels in tissues (lung and basal ganglia) with very low levels of viral DNA and RNA. We did not observe any significant differences in any cytokine mRNA in either tissue among the groups (Figure 6C and 6D). Taken together, these results indicated that cannabis use associated with reduced proinflammatory cytokine expression but did not alter anti-inflammatory cytokine expression in subtype C HIV-1 lymphoid tissue reservoirs that harbored most of proviruses.

Figure 6.

Transcription levels of inflammatory cytokines in subtype C HIV-1 tissue reservoirs between THC+ and THC− persons with suppressed pVL. The relative mRNA levels of proinflammatory cytokines IL-1β and IL-6, and anti-inflammatory cytokines IL-10 and TGF-β1 were measured in axillary LN, appendix, lung, and basal ganglia among HIV−/THC−, HIV+/THC−, and HIV+/THC+ persons. A, Significantly lower relative transcription levels of proinflammatory cytokines IL-1β and IL-6 in axillary LN in HIV+/THC+ persons compared to HIV+/THC− persons (P < .05). B, Lower relative mRNA levels of proinflammatory cytokines IL-1β (P < .05) and IL-6 (P < .01) in appendix tissues in HIV+/THC+ persons compared to HIV+/THC− persons. C, There was no significant difference in inflammatory cytokine levels in lung tissues among the 3 groups. D, No significant difference in cytokine levels was observed in basal ganglia between any groups. The mRNA level of GAPDH was used as internal reference. Each data point represents a single tissue. −ΔCt value was used to have a better visualization (a higher y-axes location reflect a higher relative mRNA level). *P < .05, **P < .01. Abbreviations: Ct, cycle threshold; HIV-1, human immunodeficiency virus 1; IL, interleukin; LN, lymph node; ns, not significant; pVL, plasma viral load; TGF-β1, transforming growth factor-β1; THC, Δ9-tetrahydrocannabinol.

DISCUSSION

Despite the common use of cannabis by PWH, its potential effects on HIV-1 have not been extensively evaluated, even for subtype B HIV-1. Studies using blood samples from persons with subtype B HIV-1 showed that cannabis use is associated with lower pVLs [23], reduced frequency of activated immune cells [25], and accelerated viral DNA decay in peripheral blood mononuclear cells [26]. Whether cannabis use affects proviral DNA burden in parenchymal tissue reservoirs of people with subtype B HIV-1 requires further investigation. More importantly, whether cannabis use has any effects on non-B subtypes need to be explored. In this study we showed that the odds of tissue harboring subtype C HIV-1 DNA and proviral burden in peripheral tissues were significantly lower in THC+ men with suppressed pVL when compared to THC− men (Figure 3 and Figure 4). Our findings suggest that cannabis use has a beneficial effect on reducing size and the distribution of subtype C HIV-1 tissue reservoirs in men with suppressed pVL.

At this point it is not clear whether any of our identified subtype C tissue reservoirs harbor replication-competent proviruses. However, we are not able to perform quantitative viral outgrowth assay [27–30], 2-probe intact proviral DNA assay, or near full-length quadruplex PCR [31, 32] to determine whether there was full-length proviral DNA because of the postmortem tissue sampling and the low copy number of detectable viral DNA in the tissues. Given that the majority (> 90%) of HIV-1 proviruses were shown to be defective in subtype B reservoirs [3, 33], it is likely a similar proportion of intact proviruses may exist in subtype C infected tissues.

Subtype B HIV-1 studies have shown that cannabis use associates with reduced inflammatory cytokine production in cerebrospinal fluid and blood [34]. Consistent with that, the proinflammatory cytokines IL-1β and IL-6 transcript levels in tissues with relatively high subtype C proviral DNA burden (axillary lymph node and appendix) were reduced in THC+ persons whereas anti-inflammatory cytokine expression did not differ. Our findings with subtype C HIV-1 tissue reservoirs suggest cannabinoid negative regulation of inflammatory mediators in tissues reservoirs without concomitant positive regulation of anti-inflammatory pathways. Previous studies have indicated that cannabinoids might modulate inflammatory responses via endocannabinoid system with interactions with nuclear factor-κB (NF-κB) or toll-like receptor pathways [35–37]. The findings presented here suggest that specific mechanisms may be involved in cannabinoid modulation of inflammation in subtype C HIV-1 infection and will need further investigation.

One of the limitations in this study is the number of persons with suppressed pVL and with cannabis use available for postmortem sampling and comparison. We were readily able to identify and recruit THC+/HIV− persons but even though we have screened 381 autopsied persons, we were only able to identify 29 deceased THC+ PWH, and only 10 of whom were virally suppressed (ART-treated aviremic). This is due to common practices in Zambia where health care providers encourage PWH to cease drinking alcohol and smoking (including cannabis) due to their potential detrimental effects. Additionally, although all participants were autopsied within 48 hours of their deaths, there were variations of postmortem intervals among persons and possible degradation of nucleic acid. Nevertheless, our results clearly demonstrate significant differentials in subtype C HIV-1 tissue reservoirs with cannabis use in the cases analyzed. A further caveat of our study is that we were only able to identify people using cannabis by measuring the presence of THC metabolite in the deceased plasma using commercially available ELISA test. The test was not designed to be quantitative and can only provide a positive or negative result from recent cannabis use. We have been using it as confirmation of the report from next of kin. Similarly, given the limited resources we have on site we were not able to quantify the levels of CBD even though CBD was reported to have more significant anti-inflammatory effect than THC. Future similar study in a resource-rich setting will be needed to confirm our observations. In addition, due to lack of computerized clinical documentation in Zambia, and linkage to HIV/ART databases, information such as time since diagnosis of HIV infection, and duration and adherence to ART treatment, were also not available.

In summary, our work demonstrates that cannabis use is associated with reduced distribution and size of subtype C HIV-1 reservoirs in men who received ART and achieved pVL suppression. Our findings also have clinical implications; cannabinoids could have therapeutic benefit in conjunction with ART to further reduce viral burden and inflammation in HIV-1 proviral tissue reservoirs and mitigate susceptibility to a variety of chronic inflammation-associated conditions that disproportionately impact PWH. Our study also suggests that cannabis use should be further evaluated in the future randomized studies with larger PWH cohorts to study its potential effects.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Author contributions. Z. L. performed experiments, analyzed the data, and wrote the manuscript. P. J. led sample collection and cannabinoid detection. L. A. M. and C. M. H. aided with tissue collection and processing and obtaining clinical information. A. G. C. conducted data analysis. J. T. W. supported methodology, data analysis, and manuscript editing. C. W. conceived of and led the study.

Acknowledgments. We are grateful to the deceased and their families for participation in this study, and to the team members and investigators of University of Zambia/University Teaching Hospital.

Financial support. This work was supported by the National Institutes of Health (grant numbers R01DA044920 and U54CA221204 to C. W.); and the Fogarty International (D43 grant number TWO10354 to C. W. and fellowship to P. J.).

Author notes

Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.