Chronic Morphine Administration Differentially Modulates Viral Reservoirs in a Simian Immunodeficiency Virus SIVmac251-Infected Rhesus Macaque Model

Identification and clearance of human immunodeficiency virus (HIV) reservoirs is a major challenge in achieving a cure for HIV. This is further complicated by comorbidities that may alter the size of the reservoirs.

IMPORTANCE Identification and clearance of human immunodeficiency virus (HIV) reservoirs is a major challenge in achieving a cure for HIV. This is further complicated by comorbidities that may alter the size of the reservoirs. There is an overlap between the risk factors for HIV and opioid abuse. Opiate abuse has been recognized as a prominent comorbidity in HIV-infected populations. People infected with HIV also abusing opioids have immune modulatory effects and more severe neurological disease. However, the impact of opioid abuse on HIV reservoirs remains unclear. In this study, we used a morphine-dependent simian immunodeficiency virus SIVmac251-infected rhesus macaque (RM) model to study the impact of opioids on HIV reservoirs. Our studies suggested that people with HIV who abuse opioids have higher reservoirs in central nervous system (CNS) than in the lymphoid system. Extrapolating the macaque findings to humans suggests that such differential modulation of HIV reservoirs among people living with HIV abusing opioids could be considered for future HIV cure research efforts. examine the impact of morphine dependency on the size of simian immunodeficiency virus (SIV) reservoirs using the SIVmac251-infected rhesus macaque (RM) model. SIV reservoirs were found to persist in all the major cellular and tissue locations in ART-suppressed, morphine-dependent, and saline-administered rhesus macaques. Our findings suggest that morphine differentially modulates the size of viral reservoirs in lymphoid tissues versus those in the CNS. To the best of our knowledge, this is the first study to report differential regulation of viral reservoir dynamics in the context of opioid dependence and warrants further investigation to dissect the molecular mechanism(s).

RESULTS
Dynamics of plasma and cerebrospinal fluid viral loads in morphinedependent versus control SIVmac251-infected rhesus macaques with and without antiretroviral therapy. In this study, we included a total of 19 macaques. Ten RMs were ramped up over 2 weeks to a final dosage of 6 mg/kg body weight morphine administered twice daily, which was maintained for 7 weeks, while nine RMs received saline and served as the control. At week 9, all the macaques were inoculated with a single 200 50% tissue culture infective dose (TCID 50 ) of SIVmac251 by intravenous injection. Five weeks postinoculation, cART was initiated in six of ten RMs in the morphine group and in five of nine RMs in the control group. Four monkeys from each group were left untreated until the end of the study (Fig. 1). Within the control group (n = 9), the median peak plasma viral load was 1.25 Â 10 7 SIV copies/ml and median peak cerebrospinal fluid (CSF) viral load was 1.02 Â 10 5 SIV copies/ml. On the other hand, in the morphine-administered group (n = 10), the median peak plasma and CSF viral loads were 8.68 Â 10 7 SIV copies/ml and 4.62 Â 10 4 SIV copies/ml, respectively ( Fig. 2A, B, E, and F). All 11 animals from the antiretroviral-treated group became aviremic with completely suppressed plasma and CSF viral loads within 13 weeks The study included a total of 19 rhesus macaques. The study had two experimental arms, one of which contained eight monkeys that did not receive antiretroviral therapy (A), and the other group had 11 macaques that were treated with combinational antiretroviral therapy (B). (A) In this group of eight animals, four were ramped up with 6-mg/kg intramuscular (i.m.) injections of morphine twice daily for 2 weeks and then continued with morphine doses for 7 weeks, and the other four animals received similar doses of normal saline (control group). After 9 weeks, all the macaques were intravenously infected with 200 TCID 50 of SIVmac251, while the administration of morphine/saline continued until end of the study. (B) In this group of 11 animals, six were ramp up with 6-mg/kg intramuscular injections of morphine twice daily for 2 weeks and then continued with morphine doses for 7 weeks, and the other five animals received similar doses of normal saline (control group). After 9 weeks, all the macaques were intravenously infected with 200 TCID 50 of SIVmac251. Five weeks postinfection, antiretroviral therapy (ART) was initiated and continued until end of the study. ART regimen consisted of two reverse transcriptase inhibitors (FTC, 40 mg/ml, and TFV, 20 mg/ml) and one integrase inhibitor (DTG, 2.5 mg/ml). The antiretroviral drugs were administered subcutaneously once daily at 1 ml/kg body weight, while the administration of morphine/saline continued until end of the study.
FIG 2 Plasma and CSF viral loads of morphine-administered SIVmac251-infected rhesus macaques. Ten macaques were ramped-up with twice daily doses of morphine (6 mg/kg) for 8 weeks, and nine macaques served as controls (received saline); the macaques were then infected with SIVmac251. Five weeks postinfection, in six animals from the morphine group and five animals from the control group, daily single doses of combination antiretroviral therapy (cART) were initiated. Daily doses of cART and morphine/saline administration continued until the end of the study. Plasma and CSF SIV viral loads were (Continued on next page) postinoculation, which is equivalent to 8 weeks of antiretroviral therapy, and they remained virally suppressed until the end of this study ( Fig. 2E and F). The longitudinal geometric means of plasma and CSF viral loads between morphine and saline-treated RMs in all the groups are presented in Fig. 2C, D, G, and H.
Cell-associated SIV DNA/RNA in the blood, lymph nodes, and rectal mucosal tissues. To understand how chronic morphine administration could modulate the size of SIV reservoirs in various tissue compartments in the presence and/or absence of cART, we sought to quantify the cell-associated SIV DNA and RNA in various anatomical tissue sanctuaries collected during necropsy of all the RMs included in this study. Since CD4 1 T cells are known to be the major contributor to viral reservoirs in both the blood and LNs, CD4 1 T cells were purified from peripheral blood and LNs followed by isolation of DNA/RNA as described in Materials and Methods. In peripheral blood mononuclear cells (PBMCs) of ART-naive RMs, the median cell-associated SIV DNA load was 15,898 copies/10 6 CD4 1 T cells in the morphine-administered group versus 8,547 copies/10 6 CD4 1 T cells in saline controls (P = 0.600). In the PBMCs of ART-treated RMs, the median cell-associated SIV DNA load was 3,332 copies/10 6 CD4 1 T cells in the morphine-administered group versus 4,583 copies/10 6 CD4 1 T cells in saline controls (P = 0.055) (Fig. 3A). Similarly, in PBMCs of ART-naive RMs, the median cell-associated SIV RNA load was 30,533 copies/10 6 CD4 1 T cells in the morphine-administered group versus 18,890 copies/10 6 CD4 1 T cells in saline controls (P = 0.990). In the PBMCs of the ART-treated group, the median cell-associated SIV RNA load was 94 copies/10 6 CD4 1 T cells in the morphine-administered group versus 446 copies/10 6 CD4 1 T cells in controls (P = 0.120) (Fig. 3B).
Next, we measured the cell-associated DNA and RNA in CD4 1 T cells purified from LNs. In LNs of ART-naive RMs, the median cell-associated SIV DNA load was 11,295 copies/10 6 CD4 1 T cells in the morphine-administered group versus 14,962 copies/10 6 CD4 1 T cells in saline controls (P = 0.600). On the other hand, in LNs of the ART group, a significant difference was found, with a median cell-associated SIV DNA load of 4,392 copies/10 6 CD4 1 T cells in the morphine-administered group versus 7,847 copies/10 6 CD4 1 T cells in saline controls (P = 0.008) (Fig. 3C). In LNs of ART-naive RMs, the median cell-associated SIV RNA load was 330,802 copies/10 6 CD4 1 T cells in the morphineadministered group versus 968,610 copies/10 6 CD4 1 T cells in saline controls (P = 0.220). In the LNs of the ART-treated group, the median cell-associated SIV RNA load was 143 copies/10 6 CD4 1 T cells in the morphine-administered group versus 260 copies/10 6 CD4 1 T cells in saline controls (P = 0.360) (Fig. 3D).
Furthermore, we assessed cell-associated DNA and RNA in mucosal tissue collected from the rectum during necropsy. In the ART-naive RMs, the median cell-associated SIV DNA load was 1,201 copies/10 6 cells in the morphine-administered group versus 10,961 copies/10 6 cells in saline controls (P = 0.342). In the ART-treated group, a significant difference was found, with a median cell-associated SIV DNA load of 70 copies/ 10 6 cells in the morphine-administered group versus 272 copies/10 6 cells in saline controls (P = 0.036) (Fig. 3E). In ART-naive RMs, the median cell-associated SIV RNA load was 368,990 copies/10 6 cells in the morphine-administered group versus 341,883 copies/10 6 cells in saline controls (P = 0.885). In the ART-treated group, the median cellassociated SIV RNA load was 11 copies/10 6 cells in the morphine-administered group versus 39 copies/10 6 cells in saline controls (P = 0.080) (Fig. 3F). measured longitudinally in all rhesus macaques by using a quantitative PCR (qPCR) assay. (A) Longitudinal plasma viral loads in individual animals that did not receive cART (copies/ml). (B) Longitudinal CSF viral loads in individual animals that did not receive cART (copies/ml). (C) Geometric means of plasma viral loads in animals that did not receive cART (black line represents control group who received saline and red line represents morphine-treated group). (D) Geometric means of CSF viral loads in animals that did not receive cART (black line represents control group who received saline and red line represents morphine-treated group). (E) Longitudinal plasma viral loads in individual animals that received cART (copies/ml). (F) Longitudinal CSF viral loads in individual animals that received cART (copies/ml). (G) Geometric means of plasma viral loads in animals that received cART (black line represents control group who received saline and red line represents morphine-treated group). (H) Geometric means of CSF viral loads in animals that received cART (black line represents control group who received saline and red line represents morphine treated group). The shaded regions indicate ART phase of the study. The dashed lines represent the limit of detection of the assay (50 copies/ml).
We also measured the cell-associated DNA/RNA in both the spleens and lungs of macaques. There was no difference in cell-associated SIV DNA/RNA levels in the lungs or spleens of the morphine-administered versus saline controls in the presence of ART treatment as well as in the ART-naive RMs (Fig. 4). In summary, we found significant reduction in cell-associated DNA viral loads within LNs and rectal mucosal tissue in morphine-administered cART-treated animals compared with that in saline controls.
Quantitation of SIV reservoirs in CD4 + T cells of blood and lymph nodes. We used the Tat/rev induced limiting dilution assay (TILDA) to measure the size of inducible SIV reservoirs in the blood and LNs of morphine/saline-administered, SIV-infected cART-treated RMs (Fig. 1B). In peripheral blood, the median frequency of CD4 1 T cells producing inducible multiply spliced (ms)-Tat/rev transcript per million of CD4 1 T cells was 4.54 in the morphine-administered group versus 9.11 in the saline control group (P = 0.067) (Fig. 5A). In the LNs, the median frequency of CD4 1 T cells producing The green dots represent samples from control animals administered saline that did not receive ART; maroon dots represent samples from morphine-administered animals that did not receive ART; blue dots represent samples from control animals administered saline and treated with ART; and the red dots represent samples from morphine-administered animals treated with ART. The horizontal dashed lines represent limit of detection of the assay (20 copies/ml). All values below the limit of detection were brought to the limit of detection for better representation in the figure. In the ART-naive control group, for one RM we were not able to recover enough cells from PBMC and lymph nodes for measuring cellassociated DNA/RNA. inducible ms-Tat/rev transcript per million of CD4 1 T cells was 6.47 in the morphineadministered group versus 10.38 in the saline control group, which was statistically significant (P = 0.036) (Fig. 5B).
We next used the intact proviral DNA assay (IPDA) to estimate the size of replication-competent SIV reservoirs in both the blood and LNs. Recently Bruner et al. (46) developed a multiplex droplet digital PCR (ddPCR)-based HIV-1 proviral DNA quantification assay that can distinguish between intact versus defective proviruses and is referred to as IPDA. The authors have described a positive correlation and identical decay rate of viral reservoirs when measured by quantitative viral outgrowth assay (QVOA) and IPDA (46). Bender et al. have reported an adaptation of IPDA for quantification of the intact SIV genome in ART-suppressed rhesus macaques, implicating thereby its utility to serve as a surrogate marker to measure the size of latent reservoirs (47). Here, using IPDA, we measured the frequency of intact SIV genomes in both blood and LNs of morphine/saline-administered as well as SIV-infected cART-treated RMs (Fig. 1B). In CD4 1 T cells purified from peripheral blood, the median number of intact SIV genomes per million cells was 1,480 for the morphine-administered group versus 2,521 in the saline control group (P = 0.080) (Fig. 5C). However, in the CD4 1 T cells purified from LNs, the median intact SIV genomes per million cells was measured as 2,789 for the morphine-administered group versus 4,399 in the saline control group (P = 0.036) (Fig. 5D).
Changes in CD4 + T cell polarization in PBMCs and lymph nodes. Chronic morphine exposure depletes lymphoid cells and leads to a phenotypic switch of Th1 to Th2 in CD4 1 T cells (38). On the other hand, a recent report suggests that in chronic HIV patients, the majority of intact replication-competent proviruses persist in Th1polarized CD4 1 T cells (48). To understand the impact of chronic morphine administration The green dots represent samples from control animals administered saline that did not receive ART; maroon dots represent samples from morphine-administered animals that did not receive ART; blue dots represent samples from control animals administered saline and treated with ART; and the red dots represent samples from morphine-administered animals treated with ART. The horizontal dashed lines represent limit of detection of the assay (20 copies/ml). All values bellow the limit of detection were brought to the limit of detection for better representation in the figure. on T lymphocytes, flow cytometry was performed at necropsy to determine the CD4 1 T cell polarity in the PBMCs and LNs of cART-treated RMs. In PBMCs, to differentiate between Th1, Th2, and Th17 cells, intracellular cytokine staining was performed as described earlier (49). Due to poor cytokine production by LN germinal center T follicular helper (Tfh) cells, it is difficult to differentiate Tfh subpopulations by cytokine production assays (50). Therefore, we performed surface staining of chemokine receptors of Tfh cells (CD95 1 PD1 1 CXCR5 1 ) from LNs to identify Th1-like Tfh (CXCR3 1 ), Th2-like Tfh (CCR4 1 ), and Th17-like Tfh (CCR6 1 ) subpopulations (51). The gating strategies used for PBMCs and LNs are described in Fig. 6, 7, and 8. In PBMCs, the frequency of CD4 1 Th1 polarized cells was measured by quantifying levels of gamma interferon (IFN-g) and tumor necrosis factor alpha (TNF-a) cytokine-secreting cells. In morphine-administered RMs, IFN-g-secreting cells were a median of 0.87% versus 1.55% (P = 0.536) for controls (Fig. 9A), while TNF-a-secreting cells were a median of 5.61% versus 14.70% for controls, which was significantly different (P = 0.017) (Fig. 9B). The median of interleukin 4 (IL-4)-secreting Th2 polarized CD4 1 T cells in the morphine-administered macaques was 2.55% versus 1.76% in saline controls (P = 0.305) (Fig. 9C). The levels of IL-17-secreting CD4 1 Th17 lymphocytes in the morphine-treated group were a median of 0.47% versus 0.48% in saline controls (P = 0.930) (Fig. 9D). Remarkably, exposure to morphine in cART-treated SIV-infected RMs led to a significant elevation of CD4 1 Treg lymphocytes, with a median of 5.22% versus 2.95% in the saline control group (P = 0.004) (Fig. 9E).
Within the LNs, morphine-administered cART-treated RMs had lower frequencies of T follicular helper cells (Tfh) expressing PD-1, CXCR5, and CD95, with a median value of 7.61% versus 13.49% in saline control groups (P = 0.017) (Fig. 9F). On the other hand, in the morphine-administered group, the frequency of regulatory Tfh (CD25 1 FoxP3 1 Tfh) cells was elevated, with a median of 8.78% versus 2.87% in saline control group (P = 0.030) (Fig. 9G). Similarly, there was an elevation of CD4 1 CXCR5 2 Treg (CD25 1 FIG 5 Size of SIV reservoirs in CD4 1 T cells from peripheral blood and lymph nodes (LNs) in morphine-administered versus control group of SIVmac251-infected ART-suppressed rhesus macaques. (A and B) Sizes of inducible SIV latent reservoirs in CD4 1 T cells from peripheral blood and LNs of morphine-administered and control rhesus macaques that received saline were measured by Tat/rev induced limiting dilution assay (TILDA), which quantifies the frequency of CD4 1 T cells producing inducible ms-Tat/rev transcript per million of CD4 1 T cells. (C and D) Frequencies of intact SIV genomes in CD4 1 T cells isolated from blood and LNs of morphine/saline-administered, SIVmac251-infected cART-treated rhesus macaques were quantified using IPDA. FoxP3 1 ) cells in the morphine-administered group, with a median of 4.57% versus 1.25% in saline control group (P = 0.004) (Fig. 9H). Evaluation of Tfh cell polarity was also carried out as shown by the representative sample in Fig. 9I. There were differences in Tfh cell differentiation (Th1 versus Th2) in the morphine-administered group versus that in the saline controls. In the morphine group, while there was a reduction in CXCR3 1 Tfh cells (Th1), with a median of 47.70% versus 56.30% in the saline control group (P = 0.004) (Fig. 9J), there was an elevation in CCR4 1 Tfh cells (Th2), with a median of 18.70% in the morphine-administered group versus 8.49% in the saline control group (P = 0.004) (Fig. 9K). Furthermore, CCR6 1 Tfh cells (Th17) were a median of 55.95% in the morphine-administered RMs versus 60.40% in saline control RMs (P = 0.792) (Fig. 9L). Finally, differences in immune activation across different cell subsets were evaluated. Figure 9M denotes a representative sample for CD4 1 T cell activation. For CD4 1 T cell activation (denoted by CD4 1 CD38 1 HLA-DR 1 ), the morphine-administered RMs had a significantly lower frequency, with a median of 3.37% versus 6.13% in saline controls (P = 0.017) (Fig. 9N). Similarly, for B cell activation (denoted by CD20 1 CD38 1 HLA-DR 1 ), the morphine-administered macaques had a significantly reduced frequency, with a median of 27.10% versus 53.50% in saline controls (P = 0.017) (Fig.  9O). On the other hand, for NK cell activation (CD3 2 CD8a/CD38 1 HLA-DR 1 ), the versus side scatter area (SSC-A) gating, lymphocytes were later gated and CD4 1 T cells further obtained from total CD3 1 T cells. From the CD4 1 T cells, the extent of IL-4 cytokine secretion was then evaluated following stimulation with PMA/ionomycin. Finally, the frequencies of CD4 1 T regs were evaluated based on the expression of CD25 and absence of CD127. These cells were later evaluated for their expression of the transcription factor FoxP3 to give the final frequency of T regs denoted as percent CD4 1 CD25 1 CD127 2 FoxP3 1 cells. morphine-administered group had a median frequency of 11.45% versus 9.16% in the saline control group (P = 0.428) (Fig. 9P).
Quantification of SIV reservoirs in the CNS. To understand how chronic morphine administration impacted the seeding and persistence of viral reservoirs in the CNS, we analyzed CD11b 1 myeloid cells (microglia and perivascular macrophages) from the brains of the cART-treated morphine-administered and saline control SIV-infected monkeys, all of which had complete viral suppression in the plasma and CSF (,50 copies/ml). Immediately after necropsy, CD11b 1 cells were purified from brains. After purification, we obtained more than 95% cells positive for CD11b (Fig. 10). The median of cell-associated DNA load was 50.5 copies per million CD11b 1 microglia/macrophages in the morphine-administered group versus 0 in the saline control group (P = 0.350) (Fig. 11A). The median cell-associated RNA load was 32.5 copies per million CD11b 1 microglia/macrophages in the morphine-administered group versus 15 in the saline control group (P = 0.710) (Fig. 11B). We next sought to estimate the size of latent replication-competent SIV reservoirs in CD11b 1 microglia/macrophages using macrophage QVOA (MØ-QVOA). The number of CD11b 1 macrophages used for the MØ-QVOA for each animal is described in Table 1. On average, we detected 0.075% CD3 1 T cells FIG 8 Gating strategy for lymph node activation and Tfh polarization panel. CD45 1 leukocytes were gated and T and non-T lymphocytes discriminated based on CD3 expression. CD3 2 cells were segregated based on those that were CD20 1 (B cells) and CD8a 1 (NK cells). The activation profiles of these cells were further evaluated based on CD38 and HLA-DR coexpression. CD4 1 T cells were separately gated out of CD3 1 T cells and levels of activation determined. The CD95 1 memory cell marker was then included to obtain memory CD4 1 T cells and later delineate the Tfh population based on CXCR5 and PD-1. Additional chemokines were used to study Tfh polarity based on Tfh1 (CXCR3), Tfh2 (CCR4), Tfh17 (CCR6), and T follicular regulatory cells (CD25 and FoxP3).
within the enriched CD11b 1 cells from brain, which is in line with 0.06% reported by Avalos et al. (52). Based on percentage of CD4 1 T cells of the total CD3 1 T cells and the frequency of CD4 1 T cells having intact proviral genome (as determined by IPDA), the probability of the presence of CD4 1 T cells with intact proviral genome in any sample of macrophage quantitative outgrowth assay was found to be less than 1. The CD11b 1 microglia/macrophages obtained from morphine-administered macaques had a median infectious unit per million (IUPM) value of 0.89 versus 0.19 in saline control group (P = 0.014) (Fig. 11C), indicating the presence of a significantly higher number of cells carrying infectious SIV in the morphine group.

DISCUSSION
In order to achieve a functional cure of HIV-1, the primary goal is to identify all cell types and anatomical sanctuaries of viral reservoirs and to understand the molecular mechanism of their long-term persistence. Substantial work has already been done to characterize and understand the underlying mechanism of latent reservoirs in circulating CD4 1 T cells and in lymphoid tissues. Several lines of evidence suggest that within 3 to 7 days of infection, HIV enters the CNS as well as tissue resident macrophages and microglia, which likely serve as viral reservoirs and are a source of viral rebound in the cases of therapeutic interruptions (34,(53)(54)(55). The mechanism of persistence of HIV reservoirs in myeloid cells is still unknown (35). Substance use disorders (SUD) and, more specifically, OUD are an important comorbidity among PLWH (23). Opioids have immune modulatory effects and mostly render macrophages more permissive to HIV infection (21,24,56,57). However, the impact of opioids on seeding and persistence of HIV reservoirs remains underscored, and this gap poses a major roadblock in HIV cure research. To shed some light on this direction, in the present study, we used SIVmac251-infected rhesus macaques as a nonhuman primate (NHP) model of HIV infection to understand how opioid use could modulate the size of viral reservoirs. We chose the SIVmac251 stock for intravenous infection since it contains multiple quasispecies of various tropism (58), and intravenous drug users (IVDU) are often initially infected by multiple transmitted viral sequences (59). Recently, Abreu et al. showed that, indeed, ART-suppressed SIVmac251 rhesus macaques are an ideal model for studying the viral reservoirs from different anatomical tissue sanctuaries (60). FIG 10 Gating strategy to detect the purity of CD11b 1 macrophages isolated from brains of rhesus macaques. Briefly, comparisons in the subsequent purity of CD11b microglia showed that higher yields of brain macrophages were obtained following enrichment of brain cells with CD11b beads. The results of the present study indicate there was no difference in the geometric means of plasma and CSF viral loads between morphine-administered and control animals receiving saline in both untreated and ART-suppressed groups. However, in a previous study, Bokhari et al. (18) reported that morphine-administered rhesus macaques exhibited 1-log higher plasma and CSF viral loads than controls. It should be noted that there are a couple of differences between the study of Bokhari et al. (18) (62). Furthermore, as the rhesus macaques used in different studies are outbred from multiple colonies, it is expected that there would be a wide variation in their genetic backgrounds and, subsequently, their responses to morphine administration and SIV disease pathogenesis. These differences could have contributed to the differences of plasma and CSF viral loads observed in these studies.
Among the untreated RMs, we did not find any difference in cell-associated DNA and RNA loads at different tissue compartments between morphine-administered and saline control groups. On the other hand, in cART-suppressed RMs, we observed a significant reduction in cell-associated DNA load in CD4 1 T cells obtained from LNs as well as in tissue samples collected from rectal mucosa in morphine-administered macaques compared to that in saline controls. This finding is substantiated by IPDA and TILDA performed on CD4 1 T cells purified from LNs, where the number of intact SIV genomes per million CD4 1 T cells was less and the size of inducible SIV reservoirs in CD4 1 T cells was lowered significantly in morphine-administered RMs compared to that in saline controls, respectively. These findings indicate that chronic morphine administration in combination with ART has a positive impact on lowering SIV reservoirs in lymphoid tissues.
The CD4 1 T cells are a major source of latent reservoirs of HIV in lymphoid tissues such as LNs, GALT, and in peripheral blood, among others (63). In LNs, HIV reservoirs seeded during acute infection are associated with virus production and storage of viral particles in immune complexes, and Tfh cells constitute the major part of LN viral reservoirs (64)(65)(66). We observed depletions of Tfh cells in LNs of morphine-administered macaques that support earlier reports that chronic morphine abuse depletes the volume and total number of lymphoid cells from LNs (67, 68). Brown et al. reported that morphine induces an immunosuppressive effect in LNs and peripheral blood of African green monkeys and pigtailed macaques, with a substantial decrease in the abundance of several metabolic proteins involved in energy metabolism pathways accompanied by significant decreases in activated CD4 1 /CD8 1 T cells (69). Therefore, we speculate that in our experimental setting, chronic morphine administration may dampen the T cell activation, suppress their metabolic activity, and cause a reduction in seeding of SIV in CD4 1 T cells during acute stage of infection before initiation of cART, which may result in a smaller size of persistent latent SIV reservoirs in CD4 1 T cells of morphineadministered animals than in control macaques that received saline. Next, we observed that morphine administration resulted in the expansion of regulatory T cells in peripheral blood as well as in LNs, which have been previously reported to suppress T-cell activation (70). In LNs, we also observed lower level of activation of CD4 1 T cells and B cells in morphine-administered animals than in saline controls. This supports the proposed mechanism of morphine-mediated reduction of SIV reservoirs within the lymphoid tissue compartments. In the LNs of morphine-administered animals, we observed a reduction in frequency of CXCR3 1 Th1-like Tfh cells and expansion of CCR4 1 Th2-like Tfh cells, and it was reported earlier that Th1-like Tfh cells enter the LN germinal center during the acute stage of HIV/SIV infection, express a high level of CCR5, and constitute a major part of viral reservoirs in LNs (50,71,72). In combination, the findings as described above may lead to a depletion of functional SIV reservoirs in CD4 1 T cells of morphine-administered ART-suppressed rhesus macaques (Fig. 12A).
Contrary to the CD4 1 T cells, we observed an opposite effect of morphine in modulation of the viral reservoir in the CNS. The median size of replication-competent SIV reservoirs in CD11b 1 microglia/macrophages from the brain was significantly larger in morphine-administered RMs than in control animals receiving saline. There are several factors that may be responsible for the morphine-mediated modulation of CNS SIV reservoirs. It has been reported that morphine upregulates the expression of CCR5 coreceptors in macrophages, which, in turn, could increase the viral permissiveness of these cells (21,57,73). Morphine is known to also lower the CNS penetration of several antiretrovirals, which could result in ongoing viral replication in the CNS and be responsible for the larger reservoir (74). Again, it has been shown that morphine alone or in combination with HIV Tat increases the permeability of the blood brain barrier (BBB) and increases transendothelial migration of leucocytes by activation of proinflammatory cytokines, intracellular Ca 21 release, and activation of myosin light chain kinase that downregulate tight junction proteins and decrease transendothelial electric resistance (26,75,76). It also increases the release of CCL2, CCL5, and IL-6 by astrocytes and upregulates expression of ICAM and VCAM on brain vascular endothelial cells that promote trafficking of HIV-infected peripheral leucocytes in the CNS (77). Morphine is also shown to downregulate the expression of anti-HIV microRNAs and impair the function of anti-HIV restriction factors in monocytes (78,79). Taken together, these could lead to increased size of the viral reservoirs in the CNS of the morphine-administered group in comparison to that in the controls that received saline (Fig. 12B). We observed a similar analogy with our finding of differential modulation of SIV reservoirs by morphine in T cells from PBMC/LNs versus from microglia/macrophages in the CNS in some previous reports where phenelzine, a monoamine oxidase (MAO) inhibitor, suppressed reactivation of HIV in T cells (80) and reactivated HIV in human microglia cell lines (81). To the best of our knowledge, this is the first report of opioid-mediated differential modulation of HIV reservoirs using macaque models of HIV infection.
The major caveat of the present study is the number of RMs included in each group due to the prohibitive cost of conducting long-term studies with multiple daily injections. Other issues include the experimental design, in which animals were initially ramped-up with morphine, then infected, and then treated with cART. In a patient cohort, this may not be an ideal scenario, as most of the people suffering from substance use disorders take multiple forms of drugs. Furthermore, opioid use may follow infection; for example, in PLWH, opioids are prescribed to treat chronic pain, and subsequently they may become addicted to these substances. For this population, how opioid use modulates the dynamics of different viral reservoirs needs to be explored. In addition, given the different cellular nature of the reservoir (CD4 1 T cells in lymphoid tissues and macrophages in the CNS), the possible effects of opioids on viral tropism is unknown.
In conclusion, for the first time, we describe a morphine-dependent SIVmac251infected RM model to study the impact of opioid use disorder (OUD) on HIV reservoirs. Our results suggest that morphine differentially modulates SIV reservoirs in LNs and rectal mucosal tissue versus SIV reservoirs in myeloid lineage of cells within the CNS. Chronic morphine administration reduces the size of viral reservoirs in CD4 1 T cells in LNs and increases the size of SIV reservoirs in brain-resident CD11b 1 macrophages. We propose that these preclinical models will serve as a tool to discover the molecular mechanism of opioid-mediated differential regulation of viral reservoirs among PLWH and suffering from OUD.

MATERIALS AND METHODS
Reagents and cell lines. Antiretroviral drugs tenofovir alafenamide (TFV) and emtricitabine (FTC) were obtained from Gilead Sciences, Foster City, CA, USA, while dolutegravir (DTG) was procured from ViiV Healthcare, Research Triangle Park, NC, USA, as a material transfer agreement (MTA) with S. N. Byrareddy. All other molecular biology-grade fine chemicals used in the study were purchased from Sigma-Aldrich, St. Louis, MO, USA, unless otherwise mentioned. CEMx174 is a hybrid human lymphoid cell line generated from human B721.174 and T-CEM cell lines that was used for expansion of virus in quantitative viral outgrowth assays due to its enhanced susceptibility to SIV infection (82). This cell line was provided by J. Hoxie (University of Pennsylvania, Philadelphia, PA). CEMx174 cells were maintained in complete RPMI (RPMI 1640) medium (Gibco; cat. no. 21870076) with 10% heat-inactivated fetal bovine serum (FBS) (Gibco; cat. no. 10437-028), 2 mM L-glutamine (Gibco; cat. no. 35050079), and 100 U/ml penicillin and 100 mg/ml streptomycin (Gibco; cat. no. 15140-122) at 37°C and 5% CO 2 . At the end of the study, all RMs were humanely euthanized using a high dose of ketamine-xylazine, and then the thoracic cavity was opened and perfused/exsanguinated according to the guidelines of the American Veterinary Medical Association. This study was reviewed and approved by UNMC Institutional Animal Care and Use Committee (IACUC) and the Institutional biosafety Committee (IBC) under protocol number 16-073-07-FC titled "The effect of cART and drug of abuse on the establishment of CNS viral reservoirs" and 15-113-01-FC titled "The combinatorial effects of opiates and promoter-variant strains of HIV-1 subtype C on neuropathogenesis and latency." UNMC has been accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International.
Study design. The overall design of the study is illustrated in a schematic form in Fig. 1. These experiments were designed to investigate the effect of chronic morphine administration on the establishment of viral reservoir in SIV-infected RMs in different anatomical compartments of the body. The study included 19 juvenile RMs. The RMs were randomly divided into two groups. One group had 10 animals which were ramped-up over 2 weeks to a final 6-mg/kg intramuscular injection of morphine administered twice daily, which was then maintained for 7 weeks, and the other group (n = 9) received a similar dose of normal saline (control group). At this point, all the macaques were intravenously inoculated with 200 TCID 50 of SIVmac251 (viral stock was obtained from Mahesh Mohan from Tulane National Primate Research Center), while the administration of morphine/saline continued until the end of the study. At 5 weeks postinoculation, daily ART was initiated in six macaques from the morphine group and five macaques in the control group and continued until the end of the study. ART regimen consisted of two reverse transcriptase inhibitors (FTC, 40 mg/ml, and TFV, 20 mg/ml) and one integrase inhibitor (DTG, 2.5 mg/ml). All the drugs were dissolved in a vehicle made up of 15% Kleptose HPB (Roquette, parenteral grade) (wt/wt) in 0.1 N NaOH. The antiretroviral drugs were administered subcutaneously once daily at 1 ml/kg body weight. Peripheral blood from the femoral vein, CSF by direct puncture of the cisterna magna or by lumbar puncture, LNs, and colorectal mucosa biopsy specimens were collected longitudinally at different time points of the study after anesthetizing the monkeys with ketamine-HCl (5 to 20 mg/kg) or tiletamine and zolazepam (Telazol, 3 to 5 mg/kg) to monitor SIV viral loads and a series of immunologic and virologic parameters as described in the experimental schema.
SIV plasma and CSF viral load quantification. SIV RNA concentration in the plasma and CSF samples were measured by quantitative reverse transcription-PCR (qRT-PCR) as previously described (83). In brief, plasma was separated from blood samples collected in K2-EDTA vacutainer tubes (Becton, Dickinson, San Diego, CA, USA) within 4 h of collection. RNA was extracted from 140 ml of plasma and Purification of CD4 + T cells. Peripheral blood mononuclear cells (PBMCs) and LN cells were enriched for CD4 1 T cells using the EasySep NHP CD4 1 T cell isolation kit from STEMCELL Technologies Canada Inc. (cat. no. 19582) as per the manufacturer's instructions. Briefly, frozen cells were thawed at 37°C, washed with complete RPMI (composition described above), centrifuged for 6 min at 1,200 rpm, and resuspended in 1 ml of recommended medium (phosphate-buffered saline [PBS] containing 2% fetal bovine serum [FBS] and 1 mM EDTA). The cell suspension was transferred to a 12-by-75-mm polystyrene tube, 50 ml of EasySep negative selection cocktail was added and mixed well, and the cell suspension was incubated at room temperature (RT) for 10 min. Then 100 ml of EasySep magnetic particles were added, and the cells were incubated at RT for 5 min. After incubation, 2 ml of recommended medium was added to the solution, and then the tubes were placed in an EasySep magnet for 5 min. The negatively selected enriched CD4 1 T cells were collected for downstream applications.
Isolation of total myeloid-enriched brain cells. Myeloid-enriched brain isolation used a modification of the procedure described by Marcondes et al. (84). In brief, the brain was sectioned and meninges removed in Hanks' balanced salt solution (HBSS; Invitrogen, Carlsbad, CA). The cleaned tissue was homogenized with a dounce homogenizer and washed twice with HBSS. Brain homogenate was digested on a rotating platform at 37°C for 30 min in HBSS with 28 U/ml DNase1 and 8 U/ml papain (Sigma, St. Louis, MO). The samples were triturated after 15-and 30-min digestion. Enzymes were inactivated with addiion of 3% FBS. Sample was washed twice with HBSS and then spun for 15 min at 1,800 rpm at 4°C through 25% Percoll (GE HealthCare, Pittsburg, PA). The resulting fatty upper and fluid middle layers were removed from the pellet. The pellet was resuspended in 10 ml of ice-cold HBSS and filtered through a 40-mm screen. Brain isolates were counted on a hemocytometer, and viability was measured by Trypan blue exclusion.
Enrichment of CD11b myeloid cells. Cell isolates were washed in PBS, and reconsituted in MACS buffer with 0.1% bovine serum albumin (BSA) (Miltenyi, Gladbach, Germany). Cells were counted and the volume was adjusted for staining with nonhuman primate CD11b microbeads (Miltenyi). Forty million cells were reconstituted in 160 ml of MACS buffer and reacted with 80 ml of CD11b microbeads at 4°C for 15 min. After incubation, cells were washed with MACS buffer with 0.1% BSA, reconstituted into 500 ml of MACS buffer, and loaded onto MACS Separator LD columns. Both the negative flowthrough and the positve CD11b 1 cells were collected, and cells were counted on Coulter Counter Z1. Isolates were used for downstream processing and analyzed for epitope expression using a fluorescence-activated cell sorter (FACS).
Purification of alveolar macrophages from bronchoalveolar lavage fluid. At necropsy, the lung was lavaged with sterile saline, and the bronchoalveolar lavage (BAL) fluid was collected. Cells were collected through centrifugation of the BAL fluid at 250 Â g for 10 min at 4°C. The cell pellets were washed twice with PBS containing 2% FBS and 2 mM EDTA. Cells were resuspended in complete RPMI (composition described above) and used for downstream applications.
Isolation DNA/RNA from CD4 + /CD11b + cells. Enriched CD4 1 T cells and CD11b 1 macrophages were pelleted by centrifugation for 6 min at 1,200 rpm and stored at 280°C until time of processing. To extract the DNA/RNA from the cells, first, the pellets were thawed on ice and thoroughly loosened by flicking the tube. Then, 600 ml Buffer RLT Plus with 2-mercaptoethanol (b-ME) at 10 ml/1 ml concentration was added and thoroughly mixed by vortexing. Following lysis, the mixture was transferred to a QIAshredder (Qiagen, cat. no. 79656), and the AllPrep DNA/RNA minikit (Qiagen, cat. no. 80204) was used for DNA/RNA isolation. The remaining protocol followed was per the kit's instructions. RNA was eluted in 30 ml of RNase-free water after performing column DNase digestion, while DNA was eluted in 50 ml of elution buffer provided in the kit. DNA/RNA concentrations were measured using a SimpliNano spectrophotometer (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) and stored in a 280°C deep freezer for future use.
Isolation DNA/RNA from frozen tissues. For DNA/RNA isolation from tissues, the AllPrep DNA/RNA minikit (Qiagen, Germantown, MD, USA; cat. no. 80204) was used. Approximately 30 mg of frozen tissue was placed into a 2-ml flat-bottom centrifuge tube on dry ice with a 5-mm stainless steel bead (Qiagen; cat. no. 69989) at the bottom. Then, 600 ml buffer RLT plus with 2-mercaptoethanol (bME) at 10 ml bME/ ml of RLT plus was added before placing the tube into a TissueLyser LT (Qiagen, USA; cat. no. 69980) set at 50 oscillations/s for 5 min. The lysate was checked for homogeneity and transferred to a QIAshredder (Qiagen; cat. no. 79656). DNA/RNA was then isolated per the kit's instructions. RNA was eluted in 30 ml of RNase-free water after performing column DNase digestion, while DNA was eluted in 50 ml of buffer EB. DNA/RNA concentrations were measured using a SimpliNano spectrophotometer (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) and stored in a 280°C deep freezer for future use.
SIV cell-associated RNA/DNA quantification. Total cell-associated SIV RNA was quantified using RT-ddPCR using 100 ng of RNA, a 1-Step RT-ddPCR Advanced kit for probes (Bio-Rad; cat. no. 1864022), and the same set of primers and probe used for SIV plasma viral load quantification. ddPCR was carried out on a Bio-Rad QX200 AutoDG digital droplet PCR system. In brief, 22 ml of reaction mix was used for droplet generation using a QX200 droplet generator, the ddPCR plate having the emulsified samples was heat sealed with foil (Bio-Rad; cat. no. 181-4040), and amplification occurred in a C1000 Touch thermal cycler (Bio-Rad, CA, USA). After thermal cycling, ddPCR plates were transferred to the QX200 droplet reader (Bio-Rad) for droplet count and fluorescence measurement. Positive droplets with amplified products were separated from negative droplets without target amplicon by applying a fluorescence amplitude threshold, and the absolute quantity of RNA per sample (copies/ml) was determined using QuantaSoft software. For quantification of cell-associated SIV DNA, the same methodology was used as described above without using the reverse transcription step and with 2Â ddPCR Supermix for probes (no dUTP) (Bio-Rad; cat. no. 1863024) instead of the 1-Step RT-ddPCR Advanced kit.
Tat/rev induced limiting dilution assay. The Tat/rev induced limiting dilution assay (TILDA) was performed as described earlier (85,86). In brief, enriched CD4 1 T cells from frozen PBMCs or LN total cells were suspended in complete RPMI at a 2-million cells/ml concentration and then rested at 37°C and 5% CO 2 for a minimum of 2 h. After resting, cells were divided into two groups: one was stimulated with 100 ng/ml phorbol myristate acetate (PMA) and 1 mg/ml ionomycin, and the other group was left untreated as the control. After 16 h, the cells were counted and then centrifuged at 2,000 rpm for 10 min to stop the reaction. Cells were then serially diluted in complete RPMI medium to the following concentrations: 18,000, 9,000, 3,000, and 1,000 cells/ml. A 96-well PCR plate was divided column-wise in half for the two treatments and then into quarters for the four-dilution sets in 8 replicates. In each well of the plate, 5 ml of master mix was added, and then 1 ml of the appropriate dilution for the corresponding treatment was added to each well. The first PCR mix contained 0. Intact proviral DNA assay. The intact proviral DNA assay was performed as described by Bender et al. (47), including thermal cycling conditions, primers, and probes used in this assay. In brief, 250 to 1,000 ng of DNA was mixed with 10 ml of 2Â ddPCR Supermix for probes (no dUTP) (Bio-Rad; cat. no. 863024), 600 nM primers, and 200 nM probes. ddPCR was carried out on a Bio-Rad QX200 AutoDG digital droplet PCR system as described above. To quantify the input cell numbers, a parallel ddPCR reaction for the rhesus macaque RPP30 gene was carried out.
Macrophage quantitative viral outgrowth assay. In this study, a modified version of the macrophage (MØ) QVOA was performed on CD11b 1 macrophages isolated from brain, lung, and liver as described by Avalos et al. (52). In brief, cells were purified using a CD11b 1 isolation kit per the manufacturer's recommended protocol (Miltenyi Biotec, Auburn, CA; cat. no. 130-091-100) as described above. The presence of the contaminating CD3 1 T cells within the enriched CD11b 1 cells was estimated by quantification of TCRb as described by Avalos et al. (52). Prior to setting the culture, plates were coated with poly-L-lysine solution (Sigma; NC9778696) for 30 min and then washed twice with PBS. Purified cells were plated in triplicates and in a 10-fold serial dilution in the presence of 10 mM zidovudine (Sigma), 25 nM darunavir (DRV; Janssen), and 5 nM ritonavir (RTV; Merck ) was used for brain cells, and complete RPMI medium was used for lung and liver. The culture was then placed in a 37°C 5% CO 2 incubator for 72 h to permit proper cell adherence. After the 3-day incubation, the antiretroviral-containing medium was removed, and cells were washed with PBS (1Â) and then replenished with BrMw /complete RPMI medium containing 10 ng/ml tumor necrosis factor alpha (TNF-a) (ProSpec; cat. no. CYT-114), 1 mg/ml Pam3CSK4 (Sigma-Aldrich; cat. no. 506350), and 1 mg/ml prostaglandin (Sigma-Aldrich; cat. no. 538904). Additionally, approximately 10 5 CEMX-174 feeder cells were added to each well. Supernatants were collected on days 5, 7, 10, and 14 post-culture activation and replenished with medium containing activating agents TNFa, Pam3CSK4, and prostaglandin. RNA was extracted from the supernatants using the QIAamp Viral RNA minikit (Qiagen; cat. no. 52906). Prior to extraction, the supernatants were centrifuged at 2,000 rpm for 5 min, and then 140 ml was aliquoted and used per the kit's instructions to isolate RNA. The RNA was eluted in 50 ml of AVE buffer. Viral RNA was quantified in culture supernatant using RT-ddPCR as described above. The frequency of replication-competent latent reservoir cells was estimated using the IUPMStats v1.0 infection frequency calculator (87).
Flow cytometry evaluation of changes in T cell polarization in peripheral blood and lymph nodes. The details of all fluorochrome-conjugated monoclonal antibodies used in the flow cytometry experiments are listed in Table 3. The single cell suspensions of PBMC and lymph nodes were prepared from samples collected from cART-treated rhesus macaques (n = 11) during necropsy. Two million to four million PBMCs were stimulated with PMA and ionomycin at final concentrations of 50 ng/ml and 500 ng/ml, respectively, in complete RPMI 1640 medium (composition described above). An unstimulated condition was included as an experimental control. After 2 h of incubation at 37°C in a humidified 5% CO 2 incubator, 1 ml of GolgiPlug (BD Biosciences, USA; cat. no. 555029) containing brefeldin A was added to each well together with 1Â of monensin (BioLegend, San Diego, CA; cat. no. 420701) and incubated for an additional 4 h. After that, a 1:1,000 dilution of zombie aqua amine reactive dye (BioLegend; cat. no. 423101) was added to the cell suspensions and incubated for 30 min in the dark. The cells were then washed and Fc blockade was carried out using polyclonal anti-human Fc receptor binding inhibitor (eBioscience, USA; cat. no. 16-9161-71). Surface staining was performed using anti-CD45, anti-CD8, anti-CD4, and anti-NKG2A antibodies for panel 1 and using anti-CD25 and anti-CD127 antibodies for panel 2 (Treg panel); cells were incubated for 1 h at room temperature in the dark. Thereafter, for panel 1, cells were fixed using a 2% paraformaldehyde (PFA) solution for 30 min, and permeabilization was performed using a 1Â BD perm/wash solution (BD Biosciences; cat. no. 554723) for 15 min at 4°C. Then, anti-CD3, anti-IFN-g, anti-TNF-a, and anti-IL-17 antibodies were added, and cells were incubated at 4°C for 15 min, washed, and resuspended in PBS. For the Treg panel, fixation and permeabilization were performed using a 1Â solution of FoxP3/transcription factor Fix/Perm concentrate (4Â) (Tonbo Biosciences, San Diego, CA; cat. no. TNB-1020-L050) that was diluted using the 1Â FoxP3/transcription factor Fix/Perm diluent (Tonbo Biosciences; cat. no. TNB-1022-L160). Following this, anti-CD3, anti-IL-4, and anti-FoxP3 antibodies were added, and cells were incubated at 4°C for 15 min, washed, and resuspended in PBS.
For LN cell suspensions, 2 million to 4 million cells were taken, zombie UV viability dye was added, and Fc receptor blockade was performed as described above. Then, in the T cell polarization panel, anti-CXCR5, anti-CXCR3, anti-CCR4, and anti-CCR6 antibodies were added, and in the Treg panel, anti-CXCR5 antibody was added. After that, cells in both panels were incubated at 37°C for 30 min. Following this, ani-CD45, anti-CD3, anti-CD4, anti-CD8, anti-CD20, anti-CD95, anti-CD38, anti-PD1, and anti-human HLA-DR antibodies were added in the polarization panel, while in the Treg panel, anti-CD25 and anti-PD1 antibodies were added. After that, cells in both panels were incubated at room temperature for 30 min, washed, and fixed. For the Treg panel, intracellular staining was performed as described above for PBMCs using anti-FoxP3 and anti-CTLA4 antibodies. All events were acquired using the BD Fortessa X450, and data were analyzed using FlowJo version 10.6. Fluorescence-minus-one (FMO) controls were included as cutoffs for placement of gates when handling nondiscriminatory variables.
Statistical analysis. Graph Pad Prism 8.0 software was used to carryout statistical comparisons and plot the figures. Descriptive statistics were presented to summarize continuous variables. Wilcoxon rank sum tests were used to determine the statistically significant difference in continuous outcomes between morphine and saline groups for flow cytometry data and other experimental data. All P values of less than 0.05 are considered statistically significant.