Hydroxypropyl-Beta-Cyclodextrin Reduces Inflammatory Signaling from Monocytes: Possible Implications for Suppression of HIV Chronic Immune Activation

Chronic immune activation is a hallmark of HIV infection and is often not controlled even in patients under antiretroviral therapy. Indeed, chronic diseases with inflammatory pathogenesis are being reported as major causes of death for HIV-infected persons. Hydroxypropyl-beta cyclodextrin (HP-BCD) is a cholesterol-sequestering drug that inhibits HIV replication and infectivity in vitro and in vivo. Recent studies have demonstrated the importance of cholesterol metabolism and content in different inflammatory conditions; therefore, we investigated the potential of HP-BCD as an immunomodulatory drug, regulating the activation of cells from HIV-infected patients. Treatment of monocytes with HP-BCD inhibited the expression and secretion of receptors and mediators that are usually enhanced in HIV patients. Furthermore, we investigated the molecular mechanisms associated with the immunomodulatory effect of HP-BCD. Our results indicate that, besides reducing viral replication, HP-BCD treatment may contribute to modulation of chronic immune activation associated with AIDS.

lication, HP-BCD treatment may contribute to modulation of chronic immune activation associated with AIDS. KEYWORDS HIV, beta-cyclodextrin, monocytes, inflammation, human immunodeficiency virus H IV prevalence is rising in nearly every geographic region in the world due to the increased survival of patients treated with antiretroviral drugs (ARTs). The cause of death for most HIV-infected persons has changed from AIDS-related opportunistic infections to chronic diseases with inflammatory pathogenesis. Chronic immune activation is a hallmark of HIV infection, and increased frequency of activated immune cells is a strong predictor of disease progression (1)(2)(3)(4). Importantly, ART treatment has a limited effect on immunological control, even when efficient viral suppression is achieved (5,6).
The systemic inflammation observed in HIV-infected patients may result primarily from the activation of innate immune cells by the virus or by microbial products due to microbial translocation, impacting T cell activation (7)(8)(9). During HIV infection, a shift from classical monocytes (CD14 High CD16 Ϫ ) to inflammatory monocytes (CD14 ϩ CD16 ϩ ) and patrolling monocytes (CD14 High CD16 ϩ ) is often observed (10). The latter cells express high levels of activation markers, such as CD80, CD86, and HLA-DR, and exhibit enhanced TNF-␣ production upon stimulation with bacterial lipopolysaccharides (LPS) (11). Increased levels of TNF-␣ in plasma are a major indicator of immune activation, and it was strongly correlated with activation and depletion of T cells in HIV-positive patients (12,13). IL-10 levels, mainly produced by monocytes, are also augmented in HIV-infected patients and are correlated positively with viral load and negatively with CD4 T cell counts and function (14)(15)(16). Both IL-10 and TNF-␣ are biomarkers of early immune activation, and their levels are correlated with viral load and persistence (17).
Reciprocal regulation between immune and lipid metabolic function and content has been recently unraveling during monocyte/macrophage stimulation under different inflammatory conditions. Cholesterol crystals were associated with inflammasome activation in macrophages during atherogenesis (18), and oxidized cholesterol metabolites were strongly proinflammatory for different cell types, inducing the production of cytokines such as TNF-␣ and MIP-1␤ (19). Further, increased oxysterol levels, associated with repression of cholesterol synthesis, prevented inflammasome activation and restricted inflammation in experimental models of septic shock and autoimmunity (20,21).
Monocytes from HIVϩ patients express high levels of the lipid uptake receptor CD36, and it was inversely correlated with CD4 ϩ T cell counts (22). Also, dysregulation of genes involved in lipid metabolism have been reported to be involved in the sustained immune activation and macrophage inflammatory phenotype in HIV patients (23). Antigen-presenting cells (APCs) derived from HIV-infected nonprogressors (NP) showed lower levels of membrane cholesterol than cells from progressors, and this phenotype was correlated with a decreased ability of NP APCs to mediate HIV trans infection of T cells (24,25). These findings strengthened the fundamental role of cholesterol in viral replication and suggested that lower membrane cholesterol levels might contribute to delayed disease progression.
2-Hydroxypropyl-beta-cyclodextrin (HP-BCD) is a cholesterol-sequestering drug, which is being used in humans for several pharmaceutical applications and has also been tested in human clinical trials of Niemann Pieck C disease (NPCD), with a safe profile (26)(27)(28). Importantly, this agent has been extensively demonstrated to inactivate HIV in vitro, including ART-resistant virus (29)(30)(31)(32). HP-BCD treatment also decreased virus infectivity and partially prevented mucosal transmission of SIV in a macaque model (32). Given that diminished cholesterol levels are associated with decreased virus replication and lower inflammatory response in different models, we hypothesized that treatment of HIVϩ patients with HP-BCD could impair HIV disease progression by acting directly on virus particles and by downmodulating the inflammatory response.
Here, we investigated the impact of in vitro HP-BCD treatment on the activation profile of monocytes upon encounter with an inflammatory stimulus. Treatment of monocytes, obtained from HIV-positive and HIV-negative donors, with HP-BCD resulted in decreased expression of CD36 and remarkable inhibition of TNF-␣ and IL-10 secretion, independent of lipid raft disruption. The inhibition of TNF-␣ and IL-10 production was associated with decreased mRNA expression and downregulation of p38MAPK activation. The cells also showed decreased expression of HLA-DR upon HP-BCD treatment, suggesting that it could potentially affect T cell activation. Our data demonstrated that HP-BCD has an immunomodulatory effect, leading to a decreased inflammatory activation of antigen-presenting cells. Therefore, HP-BCD treatment may contribute to modulation of the chronic immune activation associated with AIDS.

RESULTS
HP-BCD decreased CD36 and TNF-␣ expression in monocytes obtained from chronic HIV patients. To investigate whether HP-BCD would modulate the activation threshold of monocytes obtained from HIV patients, primary cells were cultured with different concentrations of HP-BCD for 1 h, the cells were washed, and the medium was substituted for complete medium with no HP-BCD. The cells were then cultured for an additional 48 h before stimulation with LPS, which was used as a surrogate for microbial activation. Initially, we established the maximum nontoxic concentration of HP-BCD and determined the kinetics of cholesterol recovery after cell treatment. HP-BCD was not toxic for primary monocytes even at 10 mM (Fig. 1A). Cell treatment with 10 mM HP-BCD induced 70% cholesterol reduction after 1-h treatment, with almost 100% recovery after 48 h of culture (Fig. 1B). The addition of 1 mM HP-BCD did not significantly deplete cholesterol at any time point analyzed (Fig. 1B). Cholesterol and raft recovery after 48-h culture was confirmed by staining the cells with anti-CD59 and anti-CD45 as raft and nonraft markers and with anti-TLR4. As demonstrated in Fig. 1C to E, no alteration in the frequency or expression level of either receptor was detected at the investigated time point. We also measured the concentrations of several sterol intermediates, oxysterols, and sitosterols in the cells at 48 h after HP-BCD treatment, and no significant differences for any of these lipids were observed (Fig. 1F).
Monocytes were then pretreated or not with HP-BCD, and after culturing for 48 h in HP-BCD-free media, the cells were stimulated with LPS. Eight hours after LPS activation, the expression of CD36 and TNF-␣ were evaluated by flow cytometry. Pretreatment of the cells with HP-BCD decreased the frequency of CD36 ϩ cells from almost 100% to 85% and 65% average in HIVneg and HIVpos donors, respectively. All the analyzed samples also demonstrated a reduced expression level of CD36 (mean fluorescence intensity) in both HIVneg and HIVpos donors, with an average reduction of 44% and 45%, respectively ( Fig. 2A and B). Remarkably, HP-BCD treatment almost abolished intracellular TNF-␣ expression induced by LPS (Fig. 2C). Inhibition of TNF-␣ was detected even when HP-BCD was added at a concentration lower than 1 mM (Fig. 2D), supporting that this effect was not a simple consequence of membrane cholesterol depletion.
HP-BCD inhibited the production of TNF-␣ and IL-10 at transcriptional levels. We then addressed whether HP-BCD would impact the secretion of TNF-␣ and other pro-and anti-inflammatory mediators induced by LPS. Monocytes obtained from HIV patients were pretreated or not with HP-BCD and stimulated with LPS. After 48 h, the secretion of multiple cytokines was measured by multiplex analyses of the supernatants (Fig. 3). HP-BCD treatment inhibited LPS-induced secretion of TNF-␣ and IL-10 in all donors analyzed, and the differences were statistically different ( Fig. 3A and B). On the other hand, downregulation of IFN-␥ was observed in only 1 donor out of 11 donors; IL-4 was downregulated in 2 out of 10 donors; an IL-2R was reduced in 2 out of 6 patients, whereas the secretion of IL-1␤, IL-1RA, IL-2, IL-7, IL-12, IL-13, IP-10, MIG-1, and IFN-␣ were not affected in any of the donors analyzed ( Fig. 3C to N). These data demonstrate that the drug specifically modulated TNF-␣ and IL-10 secretion but did not induce an overall monocytic anergic phenotype.
The levels of expression of TNF-␣ and IL-10 were then accessed by measuring the respective cytokine mRNA levels after LPS stimulation of HP-BCD-pretreated cells. As expected, LPS induced the expression of TNF-␣ and IL-10 mRNAs, and both were greatly reduced when the cells were pretreated with HP-BCD ( Fig. 4A and B), indicating that HP-BCD reduced LPS-mediated inflammatory response by decreasing transcription of the cytokine genes.
HP-BCD treatment inhibits p38 MAPK activation induced by LPS. Monocyte stimulation by LPS-TLR4 engagement triggers distinct signaling pathways, resulting in increased expression of inflammatory and regulatory cytokines (33)(34)(35). Since PI3K activation is dependent on lipid raft recruitment (36), we initially investigated whether HP-BCD treatment would affect PI3K expression induced by LPS. Monocytes stimulated with LPS showed enhanced levels of PI3K, but it was not affected by HP-BCD ( Fig. 5A  and D). We then measured the levels of phosphorylated p38 (p-p38) and Erk (pErk), as an indication of MAPK activation. As expected, LPS stimulation induced increased expression of p-p38 and pErk. Activation of p38, but not Erk, was strongly inhibited in the cells previously treated with HP-BCD ( Fig. 5A to C). These data suggest that HP-BCD treatment down-modulates specific signaling pathways, which in turn may explain the selective inhibition of TNF-␣ and IL-10, but not other cytokines.
The expression of costimulatory molecules is impacted by HP-BCD treatment. Since chronic activated monocytes also impact T cell activation by direct stimulation through immunological synapses, we addressed whether HP-BCD affected the costimulatory phenotype of antigen-presenting cells. Flow cytometry analysis demonstrated that HP-BCD-treated monocytes showed decreased expression of HLA-DR upon LPS stimulation (Fig. 6A and B). We also investigated whether other myeloid cells would be affected by evaluating the expression of HLA-DR among BDCA1-, BDCA2-, and BDCA3-  expressing CD14-negative cells. Stimulation with LPS slightly enhanced HLA-DR expression in BDCA1ϩ myeloid DCs, and HP-BCD pretreatment abolished this effect (Fig. 6C). These findings suggest that HP-BCD may impair the ability of LPS-stimulated myeloid cells to activate T cells, which could further contribute to downregulate chronic immune activation.

DISCUSSION
This study demonstrated that treatment of monocytes from HIV-infected patients with the cholesterol-sequestering drug HP-BCD inhibits their stimulation by LPS, resulting in decreased secretion of TNF-␣ and IL-10. Membrane cholesterol and lipid raft organization are cellular elements essential for replication of many enveloped viruses, and HP-BCD has been demonstrated to inhibit HIV infection and to inactivate free HIV particles in in vitro and in vivo models (29)(30)(31)(32). In addition, HIV infection is associated with higher atherogenic risk and cardiovascular disease, which might be related to altered uptake, synthesis, or efflux of cholesterol, evidenced by accumulation of foam cells (37,38). Recent evidences have demonstrated that cellular lipid metabolism, specifically metabolism of cholesterol and other sterols impacts distinct signal transduction pathways during the activation of immune cells (18)(19)(20). These findings motivated us to investigate the effect of HP-BCD, in an in vitro model, as a potential anti-HIV drug able to act on the chronic immune activation experienced by HIV-infected patients.
Microbial translocation and circulating LPS are markers of HIV progression and are associated with the production of inflammatory mediators by monocytes, activation of T cells, and exacerbation of inflammation (39,40). LPS sensing through TLR4 induces the secretion of several inflammatory cytokines, including TNF-␣, which is one of the main mediators detected in plasma samples from chronic HIV patients. Activated monocytes were also associated with increased plasma IL-10 levels in HIV patients, which were correlated with CD4 T cell dysfunction and disease progression (14)(15)(16).
The importance of inflammatory mediators produced by monocytes/macrophages for the establishment of chronic immune activation during HIV progression was further evidenced in nonpathogenic SIV infection. Monocytes obtained from SIVsm-infected macaques showed lower TNF-␣ secretion in response to LPS compared to monocytes obtained from HIV-infected humans or SIVmac-infected rhesus macaques. Neutraliza- tion of TNF-␣ resulted in decreased activation of CD8 ϩ T cells, indicating that monocyte activation was associated with enhanced stimulation of T cells (41).
The treatment of monocytes with HP-BCD in vitro downregulated TNF-␣ and IL-10 production independently of raft disruption. These findings are in accordance with a previous report of an experimental model of LPS-induced fatal shock in mice, in which mice treated with cyclodextrin exhibited reduced blood TNF-␣ levels and mortality (42). Also, the addition of cyclodextrin to murine macrophage cell lines did not affect localization or expression of CD14 in lipid rafts or the membrane levels of TLR4-MD2 complexes but inhibited LPS-induced TNF-␣ secretion (42).
Recent evidences demonstrated that PRR immune signaling may cross talk with other metabolic pathways. The lipid uptake receptor CD36 was previously associated with upregulation of TLR2/4 and increased LPS responsiveness (43), and accumulation of saturated fatty acids enhanced TNF-␣ secretion (44). Consistent with previous reports, the cells obtained from HIV patients in this study presented increased basal expression of CD36 (22, 23) (Fig. 2B), and HP-BCD strongly inhibited CD36 expression, what may contribute to decreased TNF-␣ secretion. Interestingly, CD36 may also be upregulated by TNF-␣ (45). Therefore, decreased TNF-␣ production mediated by HP-BCD might explain lower CD36 expression. Importantly, decreased CD36 expression might be expected to modulate the dyslipidemia and attenuate the atherogenic risk in HIV patients. Hence, HP-BCD treatment could potentially modulate atherogenic and inflammatory pathways through its effect on cellular signaling pathways.
The impaired ability of HP-BCD-treated monocytes to respond to LPS resulted from transcriptional regulation of IL-10 and TNF-␣. LPS-induced upregulation of TNF-␣ and IL-10 mRNA in PBMCs was shown to be dependent on activation of different MAP kinases (34,35). MAPKs were also involved in CD36-mediated cell signaling (43). It was previously reported that monocytes cultured with high-density lipoprotein (HDL), which mediates cholesterol efflux, inhibited LPS-or IFN-␥-driven M1 polarization and decreased expression of TNF-␣, IL-6, and CCL2 genes (46). The HDL-mediated effect was shown to be dependent on inhibition of caveolin expression and Erk1/2 phosphorylation; however, unlike our model, HDL was added to the cultures simultaneously with LPS. We did not detect any alteration in Erk phosphorylation level, but HP-BCD strongly inhibited the phosphorylation of p38 induced by LPS, indicating that p38, but not all MAPK, were regulated by HP-BCD, and further demonstrating that cholesterol perturbation affects macrophage differentiation.
Despite evidence indicating that the PI3K/akt pathway is a major regulator of IL-10 production, we did not detect any significant alteration in PI3K expression in our model, indicating that other mechanisms might be involved in the regulation of LPS stimula- tion by HP-BCD. Secreted IL-10 may impair its own production in a negative-feedback loop by downregulating p38 phosphorylation (33). It is possible that a modest IL-10 secretion induced by HP-BCD downmodulated other signal transduction pathways controlling the production of TNF-␣ and IL-10 itself. Further analysis of cell signaling pathways induced by HP-BCD will be required to dissect the mechanisms by which the drug influences cell activation and inflammation.
Finally, we demonstrated that HP-BCD treatment affected the expression of HLA-DR in monocytes and myeloid DCs activated by LPS. HIV patients often present a lower frequency of certain DC subpopulations, but circulating DCs are able to vigorously respond to stimulation, what may impact chronic immune activation (47,48). HP-BCDmediated downregulation of HLA-DR could then contribute to the control of nonspecific T cell activation.
Previous studies demonstrated that BCD showed a direct effect on HIV replication by different mechanisms of action, including: (i) free virus inactivation (29,30); (ii) inhibition of HIV entry/fusion (29,49,50); (iii) inhibition of virus release (30). Similar to the experiments performed here, all these studies used BCD at concentrations up to 20 mM (typically 10 mM), which showed no toxicity to different cell types analyzed, including primary macrophages or PBL (49,50). A 20 mM dose of HP-BCD (or higher) was also tested in experimental mouse and macaque models, where topical application of the agent was demonstrated to prevent HIV and SIV transmission with no toxicity to the tissue (32,51). Importantly, HP-BCD has been extensively reported to be safe for human use as an oral, topical, and parenteral agent (26)(27)(28)(29). In addition, a phase 1/2a clinical trial designed to test the safety and effectiveness of intrathecal injections of HP-BCD for the treatment of NPCD patients was recently successfully completed (27,28). Therefore, the data presented in this study provide substantial support for the evaluation of HP-BCD as an antiretroviral drug potentially able to reduce viral replication (29)(30)(31)(32) and control HIV-associated chronic immune activation that aggravates HIV comorbidities and AIDS progression.

MATERIALS AND METHODS
Ethical statement and sample description. Blood samples from deidentified HIV-negative donors were obtained from the Hemotherapy Service at the Hospital Universitário Clementino Fraga Filho (HUCFF) of Universidade Federal do Rio de Janeiro (UFRJ). The study protocol was approved by the Experimental Ethics Committee of UFRJ (permit 105/07). Blood samples from HIV-infected subjects were obtained from University of California Davis Medical Center and included patients who were on antiretroviral therapy and had undetectable HIV RNA ( Table 1). The study protocol was approved by institutional review board at the University of California (protocol 219307).