2-Hydroxypropyl-β-Cyclodextrin Treatment Induces Modest Immune Activation in Healthy Rhesus Macaques

ABSTRACT

Experimental simian immunodeficiency virus (SIV) infection of Asian macaques is an excellent model for HIV disease progression and therapeutic development. Recent coformulations of nucleoside analogs and an integrase inhibitor have been used for parenteral antiretroviral (ARV) administration in SIV-infected macaques, successfully resulting in undetectable plasma SIV RNA. In a cohort of SIVmac239-infected macaques, we recently observed that administration of coformulated ARVs resulted in an unexpected increase in plasma levels of soluble CD14 (sCD14), associated with stimulation of myeloid cells. We hypothesized that the coformulation solubilizing agent Kleptose (2-hydroxypropyl-β-cyclodextrin [HPβCD]) may induce inflammation with myeloid cell activation and the release of sCD14. Herein, we stimulated peripheral blood mononuclear cells (PBMCs) from healthy macaques with HPβCD from different commercial sources and evaluated inflammatory cytokine production in vitro. Treatment of PBMCs resulted in increased sCD14 release and myeloid cell interleukin-1β (IL-1β) production—with stimulation varying significantly by HPβCD source—and destabilized lymphocyte CCR5 surface expression. We further treated healthy macaques with Kleptose alone. In vivo, we observed modestly increased myeloid cell activation in response to Kleptose treatment without significant perturbation of the immunological transcriptome or epigenome. Our results demonstrate a need for vehicle-only controls and highlight immunological perturbations that can occur when using HPβCD in pharmaceutical coformulations.

IMPORTANCE SIV infection of nonhuman primates is the principal model system for assessing HIV disease progression and therapeutic development. HPβCD has recently been incorporated as a solubilizing agent in coformulations of ARVs in SIV-infected nonhuman primates. Although HPβCD has historically been considered inert, recent findings suggest that HPβCD may contribute to inflammation. Herein, we investigate the contribution of HPβCD to healthy macaque inflammation in vitro and in vivo. We observe that HPβCD causes an induction of sCD14 and IL-1β from myeloid cells in vitro and demonstrate that HPβCD stimulatory capacity varies by commercial source. In vivo, we observe modest myeloid cell activation in blood and bronchoalveolar lavage specimens absent systemic immune activation. From our findings, it is unclear whether HPβCD stimulation may improve or diminish immune reconstitution in ARV-treated lentiviral infections. Our results demonstrate a need for vehicle-only controls and highlight immunological perturbations that can occur when using HPβCD in pharmaceutical coformulations.

INTRODUCTION

Antiretrovirals (ARVs) significantly improve the life span and quality of life of people living with HIV and have contributed to a significant reduction in the global HIV epidemic (1). Whereas the use of ARVs by people living with HIV reduces their viral loads and, in turn, their ability to transmit virus to uninfected sexual partners (2, 3), the use of ARVs by exposed, uninfected individuals reduces their risk of HIV infection between 50 and 90% (4, 5). The development of novel ARVs has been aided, in part, by the use of nonhuman primate models of HIV infection. Pharmacological studies in SIV-infected nonhuman primates have helped to determine whether investigational drugs can sufficiently reduce viral loads and whether their benefit/risk profile warrants preclinical testing, and furthermore, they allow researchers to identify persistent immunological defects in suppressed animals, to evaluate the effectiveness of adjunct postexposure prophylaxes, and to study the development of the viral reservoir from infection to end of life (6). Moreover, this model has been used to explore how immunological enhancements might reduce the reservoir in vivo (7–9).

Over the last decade, standard ARV regimens for use in simian immunodeficiency virus (SIV)-infected rhesus macaques (RMs)—that is, those that are not for the purposes of investigational drug use—–have undergone a shift in formulation. Whereas previous regimens largely involved the use of individually formulated and administered reverse transcriptase inhibitors emtricitabine and tenofovir, with an integrase inhibitor and/or protease inhibitor (10–13), the development of a coformulated regimen (14) has significantly eased combined ARV therapy preparation and administration for postprophylaxis nonhuman primate research. The ability to coformulate emtricitabine, tenofovir, and dolutegravir is possible using Kleptose (Roquette parenteral-grade 2-hydroxypropyl-β-cyclodextrin [HPβCD]) as the solvent vehicle to solubilize the coformulated ARVs. HPβCD is a chemically modified, cyclic oligosaccharide long considered a pharmaceutical excipient (15) and has been used to solubilize and stabilize many compounds for parenteral use in vivo (16). However, HPβCD also solubilizes lipids, including cellular cholesterol (17). Indeed, the administration of HPβCD alone has been shown to provide a significant therapeutic benefit in patients with Niemann-Pick disease type C, a lysosomal cholesterol storage disorder (18, 19). It is unclear how the lipid mobilization properties of HPβCD might influence inflammation in vivo. Whereas HIV impairs cholesterol efflux in macrophages (20), HPβCD administration improves macrophage cholesterol efflux (21). On the other hand, the ability of HPβCD treatment to destabilize cell surface CCR5 integrity (22) may impair or mask the recovery of target CD4⁺ T cells in treated macaques in vivo.

We recently explored if ARVs (coformulated with Kleptose) supplemented with sodium butyrate might improve immunological recovery in SIV-infected rhesus macaques treated with ARVs. In both butyrate-supplemented and ARV-only control animals, we observed an unexpected delay in the recovery of CCR5-expressing CD4⁺ T cells as well as a surprising increase in plasma soluble CD14 (sCD14) posttreatment (23). We considered that the solubilizing agent Kleptose may have contributed to these unexpected results. Here, to address this concern, we treated peripheral blood mononuclear cells (PBMCs) with HPβCD from various commercial sources in vitro and additionally treated healthy rhesus macaques with Kleptose (vehicle only) in vivo. In vitro, we observed that myeloid cells are exquisitely sensitive to HPβCD—exhibiting a significant increase in sCD14 release and interleukin-1β (IL-1β) expression—and that treated lymphocytes exhibit fluctuations in surface expression of CCR5 and HLA-DR. In treated macaques, we observed modest immunological changes, including an increase in the frequency of activated myeloid cells and an increase in sCD14. These results demonstrate that HPβCD can, indeed, alter immunological parameters of importance to ARV-mediated immunological recovery, within the nonhuman primate HIV model species. Long-term treatment with drugs coformulated with HPβCD may lead to low levels of inflammation.

RESULTS

HPβCD stimulates IL-1β production from macaque myeloid cells in vitro.

We and others recently observed that the use of Kleptose-containing ARV formulations in SIV-infected macaques is associated with significantly increased levels of sCD14 in plasma (23, 24), in contrast to previous observations using ARV formulations without Kleptose (12, 13). To extend our observations, we isolated banked plasma samples from control animals in our recent study (23) and assessed sCD14 levels in these animals. Indeed, we observe a significant and persistent increase in sCD14 coincident with the initiation of ARV treatment compared to both preinfection and chronic, untreated infection (Fig. 1A).

FIG 1.

HPβCD stimulates IL-1β production from myeloid cells in vitro. (A) Longitudinal analysis of plasma sCD14 by ELISA in SIV-uninfected, chronically infected, and chronically infected, ARV-treated rhesus macaques. Dots represent one biological replicate. Lines denote the mean. (B) Analysis of sCD14 by ELISA in the supernatant of PBMCs that were untreated or treated with LPS or HPβCD from the indicated sources. Dots represent one biological replicate. Bar heights denote the mean. (C) Representative flow plot overlay showing TNF-α and IL-1β expression in gated myeloid cells from PBMCs untreated (black) or treated with 10 mg/mL HPβCD from Sigma (blue). (D) Representative histogram overlays showing IL-1β staining Fluorescence in gated myeloid cells from PBMCs that were untreated, LPS treated, or treated with 1:2 serial dilutions of 10 mg/mL HPβCD from Sigma (blue), Roquette (green), APIChem (orange), or Cayman Chemicals (purple). Lines denote IL-1β fluorescence mode. (E) Frequency of IL-1β⁺ gated myeloid cells from PBMCs that were untreated or treated with serial dilutions of 10 mg/mL HPβCD from the indicated sources. Dots represent one biological replicate. Bar heights denote the mean. (F) Frequency of TNF-α⁺ gated myeloid cells from PBMCs that were untreated or treated with serial dilutions of 10 mg/mL HPβCD from the indicated sources. Dots represent one biological replicate. (G) Mean frequency (from panels E and F) of IL-1β⁺ gated myeloid cells from PBMCs that were untreated or treated with serial dilutions of 10 mg/mL HPβCD from the indicated sources. Lines connect mean frequency. All error bars represent the standard error of the mean (SEM). Significance in panels A, E, and F was assessed by one-way ANOVA, in panel B by one-way paired t test, and in panel G by two-way paired or unpaired t tests as appropriate.

We considered that the Kleptose agent—HPβCD—might itself stimulate myeloid cells. To assess this possibility, we obtained HPβCD from multiple commercial sources and tested stimulatory capacity in healthy macaque PBMCs in vitro. We first observed significant induction of sCD14 in the supernatant of PBMCs cultured with 10 mg/mL HPβCD sourced from Sigma (one-way t test; P = 0.0183), APIChem (P = 0.0294), and Cayman (P = 0.0363), but not Roquette (Fig. 1B). We further observed significant induction of IL-1β from myeloid cells. Assessed by flow cytometry (representative images in Fig. 1C and D), we observe that HPβCD induces a concentration-dependent increase in the fluorescence of antibody-stained IL-1β in myeloid cells compared to untreated controls. This effect was on par with lipopolysaccharide (LPS) stimulation from the Sigma and Cayman Chemical HPβCD, although only modest for the Roquette and APIChem HPβCD (Fig. 1). Despite differences in the magnitude of responsiveness to HPβCD treatment, the frequency of IL-1β⁺ myeloid cells was increased at 10 mg/mL of HPβCD for all compounds and demonstrated a significant dose-response effect (one-way repeated-measures analysis of variance [ANOVA]: Sigma, P = 0.0008; Roquette, P = 0.0142; APIChem, P = 0.0229; Cayman, P = 0.0002) (Fig. 1E) We also assessed tumor necrosis factor alpha (TNF-α) production in myeloid cells from the same cultures. Despite the significant production of IL-1β by myeloid cells in response to HPβCD from all sources, TNF-α production was significantly increased only in response to stimulation with the APIChem HPβCD (P = 0.0223) (Fig. 1F).

To determine whether HPβCD differed by source, we directly compared levels of IL-1β expression at matched concentrations across the different sources (Fig. 1G). We observe that the Sigma-derived HPβCD was significantly more immunostimulatory than HPβCD from all other sources, at all concentrations, and that the Cayman HPβCD was significantly more immunostimulatory than either the Roquette or APIChem HPβCD. Compared to the APIChem HPβCD, Roquette HPβCD showed a trend for increased IL-1β stimulatory capacity at all concentrations except for 2.5-mg/mL (1:4), where this difference was statistically significant. These results demonstrate that HPβCD from multiple sources has the capacity to stimulate myeloid IL-1β production in vitro, but differs in capacity by source.

HPβCD induces modest lymphocyte activation from macaque lymphocytes in vitro.

Cellular cholesterol plays a significant role in T-cell activation and polarization (25, 26). To investigate whether HPβCD might influence lymphocyte phenotype and function, we again cultured PBMCs with HPβCD obtained from multiple sources and measured lymphocyte surface markers of immune activation in vitro. Modest variations in CCR5 and HLA-DR expression were observed in response to the Sigma and Roquette HPβCD (Fig. 2). Lymphocyte CCR5 was increased in response to the Roquette HPβCD only (Fig. 2A), showing a drug-responsive effect in CD4⁺ memory T (TM) cells (one-way ANOVA; P = 0.0023), and a specific increase in expression at the 10-mg/mL treatment in both CD4⁺ TM cells (two-way paired t test; P < 0.0001) and B cells (P = 0.045). HLA-DR expression varied in response to HPβCD treatment for NK cells only (Fig. 2B), showing a specific expression increase at the 10-mg/mL treatment in response to Sigma HPβCD (P = 0.0091) and a dose-dependent response to the Roquette HPβCD (one-way repeated-measures ANOVA P = 0.0279). In PBMCs treated with the Cayman and APIChem HPβCD (see Fig. S1 in the supplemental material), only NK cells from the Cayman HPβCD-treated PBMCs showed a response, with the 10-mg/mL culture showing a specific expression increase in HLA-DR (paired two-way t test; P = 0.0313). In addition to lymphocyte activation markers, we examined cytokine expression in treated PBMCs. We saw neither a dose-dependent effect (data not shown) nor evidence of altered cytokine production in response to treatment with 10 mg/mL of HPβCD from any of our sources (Fig. 2C). These results demonstrate that HPβCD modestly influences lymphocyte activation phenotypes in vitro, with limited influence on lymphocyte function.

FIG 2.

HPβCD induces modest lymphocyte activation in vitro. (A and B) Frequency of CCR5⁺ (A) or HLA-DR⁺ (B) gated CD4⁺ TM, CD8⁺ TM, NK cells, or B cells from PBMCs that were untreated or treated with 1:2 serial dilutions of 10 mg/mL HPβCD from Sigma (top [blue]) or Roquette (bottom [green]). Dots represent one biological replicate. Solid black lines connect individual animals. Dashed red lines denote mean expression in PMA/ionomycin-treated positive controls. (C) Radar plot depicting log₂ expression values of the indicated cytokines in gated CD4⁺ TM, CD8⁺ TM, NK cells, or B cells from PBMCs that were untreated, treated with PMA/ionomycin, or treated with 10 mg/mL of HPβCD from the indicated commercial sources. Significance in panels A and B was assessed by one-way ANOVA and in panels A to C by two-way, paired t test between matched untreated and treated samples.

Kleptose treatment induces modest immune perturbations in healthy macaques.

Given our in vitro observations, we considered that HPβCD treatment in vivo may lead to immune perturbations in rhesus macaques, similar to previous observations in murine models (27, 28). To assess this possibility, we treated rhesus macaques with (n = 5) or without (n = 5) 1 mL/kg of body weight/day 15% (wt/wt) Roquette HPβCD (Kleptose)—in essence, testing a vehicle-only formulation of the ARVs used in many nonhuman primate SIV infection studies (23, 24, 29–31). At baseline and 14 days into treatment administration, we measured the phenotypes of PBMC cell subsets, including the following: myeloid and lymphocyte frequencies as a function of hematopoietic cells (Fig. 3A); B-cell, T-cell, and NK cell frequencies as a function of lymphocytes (Fig. 3B); and CD4⁺ and CD8⁺ T cells and memory T-cell (TM) and naive T-cell (TN) distributions as a function of total CD3⁺ T cells (Fig. 3C). Of these analyses, we saw significant differences regarding only T-cell distributions: longitudinally, we saw a difference in the frequency of total CD8⁺ T-cells within the Kleptose group (paired two-way t test; P = 0.0474).

FIG 3.

HPβCD treatment is associated with modest immune perturbations in vivo. (A) PBMC myeloid and lymphocyte frequencies in untreated control (black) and Kleptose (green) macaques before and after treatment, as assessed by flow cytometry. Dots represent one biological replicate. Bar heights denote mean. (B) Stacked bar charts depicting mean PBMC B-cell, T-cell, and NK cell frequencies in control and Kleptose macaques. Bar heights denote mean. (C) Stacked bar charts depicting mean PBMC CD4⁺ and CD8⁺ T-cell subset frequencies in control and Kleptose macaques. Bar heights denote the mean. (D) Radar plot depicting mean fold change from baseline (day 14 versus day 0) expression of the indicated markers/cytokines in PBMC subsets in control and Kleptose macaques. Cytokine expression measured in response to ex vivo PMA/ionomycin stimulation. (E) Violin plots depicting expression of the indicated markers/cytokines in PBMC subsets in control and Kleptose macaques 14 days posttreatment. Dots represent one biological replicate. (F) BAL specimen myeloid and lymphocyte frequencies as in panel A. (G) Stacked bar charts depicting mean BAL specimen B-cell, T-cell, and NK cell frequencies as in panel B; (H) stacked bar charts depicting mean BAL specimen CD4⁺ and CD8⁺ T-cell subset frequencies as in panel C; (I) radar plot depicting mean fold change from baseline (day 14 versus day 0) expression of the indicated markers/cytokines in BAL specimen myeloid cells as in panel D; (J) violin plots depicting expression of the indicated markers/cytokines in BAL specimen subsets as in panel E; (K) plasma sCD14 concentrations in control and Kleptose-treated macaques, assessed by ELISA. Lines connect individual animals. (L) Blood monocyte concentrations in untreated and treated macaques. Lines connect individual animals. All error bars denote SEM. Significance for all panels was assessed by paired or unpaired two-way t test on frequency-of-parent expression values.

Among our memory T-cell subsets and myeloid cells, we additionally studied markers of activation (CCR5, HLA-DR, Ki67, or CD11b as shown), and within T-cell subsets, cytokine production (gamma interferon [IFN-γ] and IL-2) in response to phorbol 12-myristate 13-acetate (PMA)/ionomycin stimulation. In comparing longitudinal expression values, we saw a difference only in the Kleptose group and only with regard to myeloid Ki67 expression (paired one-way t test; P = 0.0465) (Fig. 3D), which may suggest increased output from bone marrow. No differences between groups were seen in the expression of these markers and cytokines when we considered day 14 posttreatment alone (Fig. 3E).

As our in vitro results indicated that HPβCD more greatly stimulates myeloid cells, we considered that Kleptose might have a greater effect within the bronchoalveolar lavage (BAL) specimens, where myeloid cells are more abundant (Fig. 3A and F). As with our PBMC analyses, we sampled prior to and after treatment and measured the following: myeloid and lymphocyte frequencies (Fig. 3F); B-cell, T-cell, and NK cell frequencies (Fig. 3G); and CD4⁺ and CD8⁺ T cells and subset distributions (Fig. 3H). Of these analyses, we saw a difference only in lymphocyte frequencies from among hematopoietic cells within the Kleptose group, where lymphocytes exhibited a significant decline posttreatment (paired one-way t test; P = 0.0489). We further profiled T-cell and myeloid cell markers of activation (CCR5, HLA-DR, Ki67, or CD11b, as shown). In comparing longitudinal expression values, we again saw a difference only in the Kleptose group and only with regard to myeloid Ki67 expression (paired one-way t test; P = 0.0079) (Fig. 3I [lymphocyte data not shown]). No differences between groups were seen in the expression of these markers and cytokines when we considered day 14 posttreatment alone (Fig. 3J).

Monocyte activation—pursuant to either bacterial stimulation or nonspecific inflammation (32, 33)—initiates the cleavage and release of sCD14. Longitudinally, we observed an increase in sCD14 in the Kleptose group (paired one-way t test; P = 0.0554) that was absent in the control group, although no differences were seen between groups at day 14 posttreatment (Fig. 3K). Despite increased sCD14 and activation of PBMCs and BAL myeloid cells (Fig. 3D and I), in vivo Kleptose treatment was not accompanied by an increase in plasma monocyte numbers, with comparable monocyte concentrations evident in both groups at both time points (Fig. 3L). These results indicate that HPβCD modestly perturbs immune activation in vivo, both in the periphery as well as in the BAL fluid.

Kleptose treatment does not significantly perturb the immunological transcriptome or epigenome in healthy rhesus macaques.

Lipid biosynthesis and availability initiate signaling cascades that regulate cellular transcriptomes, thereby influencing inflammation and T-cell subset plasticity (25, 26), as well as potentially HIV or SIV infection susceptibility. We first assessed the PBMC, BAL, jejunal biopsy (Jej), and rectal biopsy (RB) specimen immunologic transcriptomes of our control and treated animals by NanoString (see Table S1 in the supplemental material). This targeted analysis revealed no clear clustering by either treatment group or time point (Fig. 4A). When we specifically focused on a longitudinal analysis of the transcripts by principal-component analysis (PCA), we observed no clear clustering by time point (Fig. 4B), suggesting that Kleptose treatment did not induce broad immunological dysfunction.

FIG 4.

In vivo HPβCD treatment does not significantly perturb the immunological transcriptome or epigenome. (A) Heat map depicting relative transcript abundance as assessed by NanoString in PBMCs and BAL, Jej, and RB homogenates. Samples are colored and filled by group and time point, respectively, and symbol shapes within each group pertain to an individual animal. Transcript values and samples are clustered by Euclidian distance. (B) PCA considering NanoString-enumerated transcripts from Kleptose group PBMCs and BAL, Jej, and RB homogenates. Samples are colored and filled by time point as in panel A. (C) PCA considering RNA-Seq-mapped transcripts from sorted PBMCs of 2 Kleptose-treated animals. Sample shapes denote sorted cell subsets and are filled by time point and colored by animal. (D) PCA considering the top 50 SIV integration site genes of RNA-Seq-mapped transcripts from sorted PBMCs of 2 Kleptose-treated animals. Sample symbols, colors, and filling are as in panel C. (E) PCA considering variation in reads across consensus ATAC-Seq peaks from sorted PBMCs of 2 Kleptose-treated animals. Sample symbols, colors, and filling are as in panel C.

To more closely ascertain a potential influence of Kleptose on the host, we continued treatment in our animals for 120 days and sorted CD4⁺ and CD8⁺ T-cell PBMC subsets from 2 Kleptose-treated animals in order to fully assess their transcriptomes and epigenomes by transcriptome sequencing (RNA-Seq) (gene fragments per kilobase per million [FPKM], as shown in Table S2) and assay for transposase-accessible chromatin sequencing (ATAC-Seq). PCA of all mapped genes identified by RNA-Seq did not reveal a clear separation of samples by time relative to treatment (Fig. 4C). We similarly did not observe a time-dependent separation of samples when we considered only expression of genes containing 50 well-characterized SIV integration sites (34) (Fig. 4D). Finally, we considered that although transcriptional signatures did not vary in response to Kleptose treatment, chromatin accessibility may be altered, which could influence HIV integration and proviral reservoir establishment by altering the epigenetic landscape within genes where HIV and SIV are frequently known to integrate (34, 35). No differences were seen in a PCA of variation in reads across consensus ATAC-Seq peaks, however, with clustering instead observed by cell type (Fig. 4E). Additionally, analysis of differential accessibility (DESeq2) using paired samples did not identify any peaks, suggesting that Kleptose treatment alone does not alter the epigenetic landscape.

DISCUSSION

Non-AIDS comorbidities are now a leading cause of death in people living with HIV and include increased incidence of diabetes, cardiovascular and renal disease, and neurocognitive decline (36, 37). The development of these comorbidities is strongly predicted by biomarkers associated with inflammation—including sCD14, LPS-binding protein (LBP), and C-reactive protein (CRP)—and they contribute to an early onset of aging-related sequelae, so-called “inflammaging” (36, 37). ARV therapy in rhesus macaques provides an invaluable research model for defining the circumstances leading to persistent inflammation, the evaluation of new therapeutics, and uncovering of mechanisms of HIV reservoir establishment and persistence (6). In macaques, recent ARV formulations have begun to incorporate HPβCD to improve drug solubility, allowing for the coformulation of multiple drugs for parenteral administration (14). Though largely considered an excipient (15), HPβCD should more aptly be considered an adjuvant or pharmaceutical in its own right. In macaques immunized with influenza virus hemagglutinin split vaccine, the addition of HPβCD to the vaccine formulation significantly improved hemagglutinin-specific IgG titers (27), and promising improvements have been demonstrated in Niemann-Pick type C disease patients receiving HPβCD alone (18, 19). Here, we investigated the potential effects of HPβCD on macaque PBMCs in vitro and in healthy macaques in vivo and demonstrate that HPβCD alone induces modest immune perturbations.

Our assessment of in vitro cultures of healthy macaque PBMCs with HPβCD demonstrated that HPβCD acquired from multiple sources led to a significant induction of sCD14 and IL-1β from myeloid cells, with TNF-α induction from only APIChem HPβCD-treated PBMCs (Fig. 1; see Fig. S1 in the supplemental material). An HPβCD, myeloid-specific IL-1β response is well documented (27, 28) and may be linked to increased cholesterol mobilization (21), rather than stimulation per se. In vitro, HPβCD has been shown to induce expression of TNFA in bone marrow-derived macrophages and macrophage-derived cell lines (28), but interestingly, this response is significantly abrogated in the presence of LPS stimulation (38), suggesting that HPβCD-associated regulation of TNF-α is sensitive to coregulatory effects. Importantly, we observe that the degree of IL-1β and TNF-α responsiveness in stimulated cells was not uniform across commercial sources. These differences in potency signal that caution is required when comparing/contrasting the results of studies utilizing formulations from different sources. Indeed, HPβCD formulations are not composed of a single molecular species, with hydroxypropylation varying among member species in a particular mixture (39). From cultured PBMCs, we further saw a dysregulation in lymphocyte CCR5 and HLA-DR in response to HPβCD, absent alterations in proinflammatory cytokine expression (Fig. 2). Lymphocyte activation may be secondary to myeloid cell costimulation after HPβCD treatment (40); however, as CCR5 and HLA-DR expression were destabilized rather than uniformly increased or decreased, it is likely that our observations reflect the conformational sensitivity of CCR5 (22) and HLA-DR (41) to cholesterol concentrations within plasma membrane lipid rafts.

In healthy mice, immune activation in response to HPβCD is administration route specific, with temporary DNA release and myeloid cell activation noted at the site of injection and prolonged MYD88 induction and IL1B induction observed in lymphoid tissues only (27). As a vaccine adjuvant, HPβCD is trapped by MARCO⁺ macrophages and CD11c⁺ dendritic cells in lymph nodes and boosts immune priming, with no evidence of systemic adjuvanticity (27). In our healthy macaques, we observed that subcutaneous HPβCD (Kleptose) administration increased Ki67 expression in peripheral blood and BAL fluid myeloid cells and increased circulating sCD14 in treated animals (Fig. 3). Although alterations in lymphocyte frequencies were also observed, a lack of consistent lymphocyte activation suggests that these perturbations were likely secondary to myeloid cell activation and costimulation. Analysis of the immunologic transcriptome at multiple sites further shows that treatment did not uniformly alter gene expression (Fig. 4), suggesting that the observed changes in myeloid activation did not result in robust or systemic immune activation.

HPβCD prevents HIV-1 and SIV infection of target cells in vitro at both at the cellular and viral level. Whereas HPβCD extracts cholesterol from plasma membrane lipid rafts which are necessary for HIV entry and egress (42, 43), HPβCD also depletes cholesterol from viral envelopes, leading to viral inactivation (44, 45). In vivo, however, the role of HPβCD in reducing viral infection and replication is less clear. Whereas the utilization of HPβCD-inactivated viral stocks or the coadministration of HPβCD with infectious stocks reduces susceptibility to vaginal infection in naive macaques, repeated vaginal exposure to HPβCD negates a protective (antiviral) effect, suggesting that HPβCD may induce a lentivirus-permissive milieu (46). Indeed, a proinflammatory cervicovaginal milieu—including high concentrations of IL-1β—is associated with a significantly increased risk of HIV acquisition in young women (47).

Clinically approved ARVs have an excellent risk/benefit profile, demonstrating minimal adverse reactions with prolonged preexposure prophylactic use (48); however, gastrointestinal complications are prevalent (49), and we have similarly documented modest perturbations to intestinal immunity and the intestinal microbiome in healthy macaques treated with short courses of ARVs (50). We did not explore how the combination of HPβCD with ARVs may influence immune homeostasis nor how HPβCD may influence immune reconstitution or potentiate non-AIDS-related comorbidities in the setting of persistent inflammation and immune dysregulation such as chronic HIV and SIV infection (36, 37, 51). Collectively, our findings do not negate the use of HPβCD as a valuable research tool in SIV nonhuman primate models of HIV infection. Chronic myeloid activation induced by HPβCD may accelerate immune exhaustion, expand the viral reservoir, or incite cardiovascular failure. On the other hand, myeloid activation may promote clearance of translocating bacteria or may enhance ARV efficacy, either by improved solubilization or by reactivation and clearance of the viral reservoir. Our results highlight the need for vehicle-only controls when using HPβCD in ARV formulations and suggest that long-term studies of HPβCD use in vivo are merited.

MATERIALS AND METHODS

Animals and Kleptose treatment.

Ten healthy, adult, male RMs (Macaca mulatta) were assigned to a control (n = 5) or Kleptose treatment (n = 5) group, with sample size based on previous studies of experimental manipulations of disease progression in the RM model. Groups were stratified by weight and genotype (Mamu-A*001, -A*002, -B*008, and -B*017), and animals were sampled as mixed populations. Kleptose-treated animals received 15% (wt/wt) Kleptose HBP parenteral grade (Roquette no. 346111109D) subcutaneously at 1 mL/kg of body weight/day for up to 120 days.

The NIAID Division of Intramural Research Animal Care and Use Program, as part of the NIH Intramural Research Program, approved all the experimental procedures (protocol LVD26E). The program complies with all applicable provisions of the Animal Welfare Act and other federal statutes and regulations relating to animals. Animals were housed and cared for at the NIH Animal Center, under the supervision of the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC)-accredited Division of Veterinary Resources and as recommended by the Office of Animal Care and Use Nonhuman Primate Management Plan. Husbandry and care met the standards set forth by the Animal Welfare Act, Animal Welfare Regulations, as well as the Guide for the Care and Use of Laboratory Animals (52). The physical conditions of the animals were monitored daily. Animals in this study were exempt from contact social housing due to scientific justification, per Institutional Animal Care and Use Committee protocol, and were housed in noncontact social housing where primary enclosures consisted of stainless-steel primate caging. Animals were provided continuous access to water and offered commercial monkey biscuits twice daily as well as fresh produce, eggs, and bread products twice weekly and a foraging mix consisting of raisins, nuts, and rice three times weekly. Enrichment to stimulate foraging and play activity was provided in the form of food puzzles, toys, cage furniture, and mirrors or television.

Sample collection.

Blood, BAL, and intestinal biopsy samples were collected longitudinally. Sampling occurred in random order. Neither the investigators nor the animal handlers were blind to group allocation to ensure multilateral supervision of design and palliative treatment. Animals were sedated with Telazol at 3 to 4 mg/kg intramuscularly (i.m.) and were further anesthetized with isoflurane gas by intubation, to effect. Successful anesthetization was monitored by response to stimuli. No animals met endpoint criteria as defined by (i) loss of 25% body weight from baseline weight when assigned to the protocol, (ii) major organ failure or medical conditions unresponsive to treatment, (iii) complete anorexia for 4 days or an inability to feed or drink sufficient nutrients to maintain body weight without assistance for 7 days, (iv) distress vocalization unresponsive to treatment or intervention for 7 days, or (v) tumors arising from other than experimental means that grew in excess of 10% of body weight, impaired movement, or ulcerated.

Whole blood was collected into EDTA. For BAL specimen collection, silicone tubing was directed into the trachea with the assistance of a laryngoscope, whereupon warmed normal saline was instilled and subsequently aspirated for collection. For rectal biopsies, fecal material was removed from the rectum and biopsy specimens obtained with biopsy forceps. Jejunal biopsy specimens were acquired by video-guided endoscopy, with samples collected by biopsy forceps. Ten pinch biopsy samples were obtained per animal at each time point. Biopsy samples were transported in RPMI and washed twice in phosphate-buffered saline (PBS) prior to processing.

Plasma was isolated from whole blood by centrifugation and stored at −80°C until accession. Mononuclear cells were isolated from blood by Ficoll gradient centrifugation and from BAL samples by straining samples through a 0.7-μm-pore cell strainer, followed by centrifugation. PBMCs and BAL cell pellets were either utilized immediately for in vitro stimulation or resuspended in freezing medium (fetal calf serum [FCS] supplemented with 10% dimethyl sulfoxide [DMSO]) and stored at −80°C until accession. Intestinal biopsy samples were transferred to PowerBead tubes (Qiagen) and homogenized in 1 mL TRIzol (Thermo Fisher Scientific) at room temperature on a Precellys 24 homogenizer (Bertin Technologies) at 5,000 rpm in 4 successive 20-s intervals. TRIzol-preserved homogenates were stored at −80°C until accession.

In vitro stimulations.

For cell analysis by flow cytometry, per condition, 2 × 10⁶ fresh PBMCs were plated with 1 μg/mL brefeldin A in 1 mL of R10 (RPMI supplemented with 10% fetal calf serum, 100 U/mL penicillin, 100 μg/mL streptomycin sulfate, and 1.7 mM sodium glutamate) and were untreated, treated with dilutions of HPβCD (CAS no. 128446-35-5) beginning at 10 mg/mL (Roquette no. 346111109D, Sigma no. H107-5G, APIChem Technology no. AC-3234, or Cayman Chemical no.16169), 1 μg/mL LPS (InvivoGen no. tlrl-peklps), or 2.5 ng/mL PMA with 1 μg/mL ionomycin. Cells were treated at 37°C for either 5 h for myeloid cell stimulation/analysis or 16 h for lymphocyte stimulation/analysis.

For sCD14 quantification by enzyme-linked immunosorbent assay (ELISA), per condition, 1 × 10⁶ fresh PBMCs were plated as described above, in the absence of brefeldin A. Cells were treated at 37°C for 5 h, and the supernatant was collected by centrifugation. Supernatant was stored at −80°C until accession.

Flow cytometry and sorting.

For PBMC and BAL specimen analysis by flow cytometry, frozen cells were thawed into R10 and washed twice in PBS. Antibodies against the following antigens were used for staining at predetermined concentrations: CCR5 (clone 3A9) phycoerythrin (PE), CD3 (SP34-2) Alexa Fluor 700, CD4 (L200) BV650, CD8 (SK1) BUV805 (or RPA-T8) Pacific Blue, CD11b (ICRF44) BV605, CD20 (2H7) allophycocyanin (APC)-H7 or BUV395, CD28 (CD28.2) PE-CF594, CD45 (D058-1283) BV786, CD95 (DX2) PE-Cy5, HLA-DR (G46-6/L243) APC, IL-1β (AS10) PE, Ki67 (B56) fluorescein isothiocyanate (FITC), and TNF-α (MAb11) FITC from Becton Dickinson; CD14 (M5E2) BV421, CD95 (DX2) PE-Cy5, HLA-DR (L243) BV711, IL-2 (MQ1-17H12) BV650, and TNF-α (MAb11) BV711 from Biolegend; CD4 (OKT4) eFluor-450, CD8 (SK1) peridinin chlorophyll protein (PerCP) eFluor-710, and IFN-γ (4S.B3) eFluor-450 from Thermo Fisher Scientific; and CD28 (CD28.2) ECD and NKG2a (Z199) PE-Cy7 from Beckman Coulter. Cell viability was assessed using the LIVE/DEAD Aqua fixable dead-cell stain (Thermo Fisher-Scientific). For phenotypic and functional analyses, cells were permeabilized with Cytofix/Cytoperm (BD) prior to intracellular staining.

For phenotypic and functional analyses, polychromatic flow cytometry was performed on stained cells, using a BD LSRII (FACSDiva v.8.0.1) or Cytek Aurora 5L (SpectroFlow v.3.0.3). A minimum threshold of 100 collected events in the parent population was utilized for phenotypic and functional expression analyses (FlowJo 10.7.1). For cell sorting, at least 20,000 cells were sorted into complete RPMI with 10% fetal bovine serum (FBS) using a BD FACSymphony S6 cell sorter (FACSDiva v.9.1.2). Cell populations were defined as in Fig. S2 in the supplemental material, with positive/negative gating based on clearly grouped populations, history-determined expression, and the use of internal controls.

ELISAs.

Concentrations of sCD14 were quantified from plasma or supernatant using a commercially available ELISA kit (R&D no. DC140) according to the manufacturer’s protocol. All samples were assessed as technical duplicates or triplicates and were independently confirmed.

RNA extraction and transcript quantification for NanoString.

For PBMC and BAL specimen transcript quantification, frozen aliquots were thawed into R10. RNA was isolated from cellular resuspensions using the RNeasy minikit (Qiagen) with on-column DNase treatment per the manufacturer’s protocol. For rectal and jejunal specimen transcript quantification, TRIzol-preserved samples were thawed and treated with 200 μL chloroform to separate nucleic acid into an aqueous phase. Following separation, Total RNA was isolated from the aqueous phase using the MagMAX-96 total RNA isolation kit (Thermo Fisher Scientific) per the manufacturer's protocol. RNA concentration and purity (A_260/280 ≥ 1.8) were assessed by spectrophotometer and normalized to 50 to 100 ng/μL in PCR-grade H₂O. For transcript quantification by NanoString (53), preparation, hybridization, and detection of RNA samples were carried out by following the NanoString manufacturer’s instructions (NanoString Technologies) using the nCounter NHP Immunology Panel. Subsequent analyses were performed using the nSolver analysis system v4.0.70 (NanoString Technologies). NanoString-generated reads were normalized to internal positive and negative controls and housekeeping genes.

RNA extraction and library preparation for RNA-Seq.

Sorted cells were pelleted and resuspended in Buffer RLT Plus (Qiagen) with 1% β-mercaptoethanol (BME). RNA was extracted using RNeasy Micro kit (Qiagen). RNA quality was verified by an RNA 6000 Pico kit (Agilent). mRNA selection and library preparation were performed using SMART-Seq v.4 ultralow-input RNA prep kit (TaKaRa).

DNA extraction, tagmentation, and library preparation for ATAC-Seq.

For tagmentation, 10,000 sorted cells from each population were pelleted and resuspended in 2× tagmentation buffer (Illumina), 0.02% digitonin, and 0.1% Tween 20, with Tagment DNA TDE1 enzyme (Illumina) and incubation at 37°C for 1 h. The mixture was then incubated with proteinase K for 30 min at 40°C. High-molecular-weight DNA was removed with a 0.7× selection ratio of Agencourt AMPureXP beads (Beckman Coulter), and low-molecular-weight DNA was isolated from the remainder with a 1.2× selection ratio. Library amplification was performed with KAPA HiFi HotStart ReadyMix (KAPA) and i5 and i7 indexing primers (Nextera DNA CD Indexes) for 11 to 15 total cycles, depending on the sample concentration reached after 5 cycles (KAPA library quantification kit). Library DNA was purified with a 1× selection ratio of AMPureXP beads. The final library was verified by a high-sensitivity DNA bioanalyzer (Agilent).

Sequencing and data processing.

RNA-Seq and ATAC-Seq pipelines were done utilizing the computational resources of the NIH HPC Biowulf cluster (http://hpc.nih.gov). RNA-Seq samples were sequenced as 100-bp paired-end libraries on a NovaSeq 6000 S4 (Illumina). All samples had over 120 million reads. RNA-Seq data were processed using the RNA-seek workflow v.1.2.1 (https://doi.org/10.5281/zenodo.5223025). In brief, reads were trimmed using Cutadapt v.1.18 (https://github.com/marcelm/cutadapt) and aligned to the Mmul_10/RheMac10 reference genome and Gencode (Ensembl release 108) using STAR v.2.7.6a in 2-pass basic mode (54). Expression levels were quantified with RSEM v.1.3.3 (55) before being converted into FPKM values.

ATAC-Seq samples were sequenced as 150-bp paired-end libraries on a Novaseq 6000 S4 (Illumina). All samples had over 130 million reads. ATAC-Seq data were processed as detailed at https://github.com/OpenOmics/ATAC-seq. In brief, samples were trimmed for adapters using Cutadapt before alignment. The trimmed reads were aligned to the RheMac10 reference using Bowtie2 v.2.3.4.1 (56) with flag -k 10. The peaks were called using Genrich v.0.6 (https://github.com/jsh58/Genrich) with the following flags: -j -y -r -v -d 150 -m 5 -e chrM, chrY. Differential peaks were analyzed using the DESeq2 (57) algorithm with paired samples, and no peaks were determined significant, with a false-discovery rate (FDR) of <0.05.

SIV integration site gene selection.

The Retrovirus Integration Database (34) was queried to find all integration sites for SIV in the rheMac8 genome, returning over 82,000 previously determined sites. These were sorted by the gene they are located within or closest to, and the top 50 genes with the most known integration sites were determined.

Statistical analyses.

One-way ANOVAs and unpaired or paired one-way or two-way t tests were used in statistical analyses of myeloid and lymphocyte phenotype and function and monocyte and sCD14 concentrations (Prism v.9.0; GraphPad Software, Inc.). Radar plots were generated from nonnormalized expression data using ggradar v.0.2 in RStudio v.1.1.463 using R v.3.6.2. NanoString-assessed transcriptome variability was visualized by PCA using the ggbiplot package v.0.55, and transcriptome significance was evaluated by adonis on derived Euclidian distances, in RStudio using R. Differential chromatin accessibility by ATAC-Seq sequencing was determined by DESeq2 as already discussed under “Sequencing and data processing.” No data that met minimum threshold requirements as outlined above were excluded.

Data availability.

RNA-Seq and ATAC-Seq sequencing files have been deposited in NCBI under GEO accession no. GSE233467. Data are available from this study from the corresponding author upon reasonable request. All utilized coding packages are publicly available as indicated.