|
Selective suppression of inflammation could deplete HIV and control HIV activation - Novel mechanisms to inhibit HIV reservoir seeding using Jak inhibitors
| | |
Download the PDF here
"Our results show that ruxolitinib and tofacitinib can exert direct antiviral potency in infected cells isolated from viremic individuals and in primary CD4 T cells infected in vitro (Fig 2). These Jak inhibitors could confer a direct antiviral activity by blocking the phosphorylation of STATs and subsequent binding of this transcription factor to several STAT5 binding sites within the HIV-1 LTR (Fig 7E) [7, 22, 27]. Decreased virus production (summarized in Fig 7C-7F) could also be a consequence of the Jak-STAT induced down-regulation of T cell proliferaton. CD4 T cells exposed to these inhibitors should show decreased expression of transcription factors (NFAT-c, NF-κ-B) expressed by activated T cells [47] and which are required for HIV LTR mediated transcription [48]. Herein, we report that Jak inhibitors demonstrate nanomolar inhibition of viral replication in ex vivo CD4 T cells from viremic individuals, even in the presence of an EC99 of ART (Fig 2). Therefore, Jak inhibitors may block viral replication in ART treated subjects [33] which highlights their potential as a therapeutic modality, since they could block potential residual viral replication in pharmacological sanctuary sites in vivo where antiviral agents do not reach optimal concentrations [45]. Additionally, our results suggest that preferential cell death or shortened lifespan (down regulation of Bcl-2 across T cell subsets; S1 Fig) conferred by Jak inhibitor exposure lead to a reduction in the overall size of the viral reservoir. Dissemination of HIV to bystander cells is also inhibited by addition of Jak inhibitors further confirming the capacity of this class of molecules to prevent further seeding of the reservoir.....The effects of Jak inhibitors, described herein on mechanisms that promote HIV dissemination and persistence may be advantageous as an alternative therapeutic intervention that could be implemented in future curative strategies. The impact of ruxolitinib on inflammation associated with HIV infection, which plays an important role in seeding viral reservoirs, viral persistence and ongoing low-level replication in sanctuary sites is currently being evaluated in an NIH-ACTG sponsored 21-site Phase 2a study in humans that is currently enrolling."
The study results have strengthened rationale for an ongoing, AIDS Clinical Trials Group (ACTG) Phase 2a multi-site trial to evaluate the safety, tolerability and efficacy of ruxolitinib in HIV-infected individuals.
https://clinicaltrials.gov/ct2/show/NCT02475655?term=ruxolitinib&cond=Hiv&rank=1
----------------------
Novel mechanisms to inhibit HIV reservoir seeding using Jak inhibitors - Plos Pathogens
Selective suppression of inflammation could deplete HIV and control HIV activation
Woodruff Health Sciences Center | Dec. 21, 2017
http://news.emory.edu//stories/2017/12/schinazi_hiv_inflammation/index.html
A class of anti-inflammatory drugs already FDA-approved for rheumatoid arthritis could "purge" the reservoir of infected immune cells in people infected by HIV, according to new research.
When culturing cells from HIV-infected individuals, researchers found the medications tofacitinib and ruxolitinib block viral production from infected cells, prevent transmission to bystander cells, and decay the viral reservoir. The results were published in PLOS Pathogens.
"One of the major impediments to an HIV cure is the reservoir. These are a very small number of immune cells that have the virus integrated into their genomes. These cells are completely undetectable by the immune system because the virus is dormant. But as soon as you stop treatment, the virus reactivates," said the study's senior author Rafick-Pierre Sekaly, PhD, the Richard J. Fasenmyer Professor of Immunopathogenesis, co-director of the Center for AIDS Research Proteomics and Systems Biology Core, and professor of pathology at Case Western Reserve University School of Medicine. "Our results show you can kill these reservoir cells if you treat them with Jak inhibitors."
The new study included 37 people infected by HIV, but who have controlled the virus with antiretroviral drugs. The size of their residual reservoir - the number of cells with HIV integrated into their DNA - was linked to the activity of Jak enzymes in the cells. The research suggests blocking Jak enzymes could, in effect, impose a drought on the HIV reservoir. Laboratory experiments confirmed that Jak inhibitors prevent HIV spread to nearby, healthy cells.
Jak inhibitors block key pro-inflammatory cytokines that are dysregulated in certain diseases including autoimmune disorders such as rheumatoid arthritis. In blocking these inflammatory cytokines, Jak inhibitors restore the inflammatory state to a "normal" level observed in individuals without an inflammatory disorder.
Sekaly's foundational research into HIV reservoir persistence made him a natural collaborator for researchers at Emory University School of Medicine with expertise in small molecule Jak inhibitors. Co-corresponding author on the study is Raymond F. Schinazi, PhD, DSc, Frances Winship Walters Professor of Pediatrics and director of the Laboratory of Biochemical Pharmacology at Emory University. Schinazi is also the director of the HIV Cure Scientific Working Group at the Emory University Center for AIDS Research. One of the study's three first authors, Christina Gavegnano, PhD, is an assistant professor in Dr. Schinazi's group. Schinazi and Gavegnano previously reported that Jak inhibitors demonstrate antiviral potency in human immune cells, and that ruxolitinib can ameliorate HIV-associated encephalitis in a mouse model.
The researchers hope their collaborative study could result in a new approved indication for Jak inhibitors to treat HIV infected individuals. Said Schinazi, "Our results strongly suggest that monitoring and suppressing inflammation with Jak inhibitors will impact HIV reservoirs not only systemically, but also in the central nervous system where the virus hides."
A key safety feature of Jak inhibitors in the study is that they are not globally immunosuppressive, but instead immunomodulatory. Says Sekaly, "Our results show how Jak inhibitors do not prevent the reservoir cells from responding to new infections."
The PLOS Pathogen paper showed Jak inhibitors do not block functional immune responses to HIV, or normal immune cell function. These findings are mirrored by data collected in humans where global immunosuppression is not reported with Jak inhibitors.
This mechanism of action is unique due to the drugs' apparent specificity for HIV-infected cells, unlike global immunosuppressive agents used in transplants, which blunt all activation and result in reduced functional immunity.
Since Jak is overactive in HIV-infected cells, Jak inhibitors are able to specifically target inflammatory cytokines in the cells that cause them to activate and "reseed" nearby cells. HIV requires cellular activation to replicate efficiently. Blocking key inflammatory cytokines in HIV-infected cells with a Jak inhibitor creates an environment where the virus simply cannot replicate because it cannot properly activate the cell. The unique mechanism of action allows Jak inhibitors to block viral replication in infected cells, reduce the lifespan of reservoir cells (key markers such as Bcl-2 are down-regulated), and block expansion of the viral reservoir.
Explains Gavegnano, "Jak inhibitors are unique since they are highly selective agents that keep the lid on smoldering infection through anti-inflammatory properties. If the lid is on the smoldering fire long enough, it is our hope that the fire will be extinguished.
The study results have strengthened rationale for an ongoing, AIDS Clinical Trials Group (ACTG) Phase 2a multi-site trial to evaluate the safety, tolerability and efficacy of ruxolitinib in HIV-infected individuals.
"We are rigorously evaluating the effect of Jak inhibitors on key events that prevent eradication of HIV in culture, animal models, and humans. We are also carefully monitoring safety," says Gavegnano. "It is important to evaluate these agents from every possible angle, to ensure that we are delivering a safe and effective treatment. Our data to date provides reason to explore the indication of HIV for Jak inhibitors carefully. We look forward to understanding how blocking residual inflammation in HIV-infected cells can impact the ultimate goal of a cure."
The study was made possible by multiple grants to Sekaly from the National Institute of Mental Health and the National Institute of Allergy and Infectious Diseases. Schinazi's team is funded by the National Institute of Mental Health of the National Institute of Health, and by the Emory Center for AIDS Research.
----------------------------
Novel mechanisms to inhibit HIV reservoir seeding using Jak inhibitors
Plos Pathogens - Published: December 21, 2017
http://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1006740
Christina Gavegnano1, Jessica H. Brehm2¤, Franck P. Dupuy3, Aarthi Talla2, Susan Pereira Ribeiro2, Deanna A. Kulpa1, Cheryl Cameron2, Stephanie Santos4, Selwyn J. Hurwitz1, Vincent C. Marconi5, Jean-Pierre Routy6, Laurent Sabbagh7, Raymond F. Schinazi1*, Rafick Pierre SeAkaly2*
![ab](../images/010317/010317-1/ab.gif)
Abstract
Despite advances in the treatment of HIV infection with ART, elucidating strategies to overcome HIV persistence, including blockade of viral reservoir establishment, maintenance, and expansion, remains a challenge. T cell homeostasis is a major driver of HIV persistence. Cytokines involved in regulating homeostasis of memory T cells, the major hub of the HIV reservoir, trigger the Jak-STAT pathway. We evaluated the ability of tofacitinib and ruxolitinib, two FDA-approved Jak inhibitors, to block seeding and maintenance of the HIV reservoir in vitro. We provide direct demonstration for involvement of the Jak-STAT pathway in HIV persistence in vivo, ex vivo, and in vitro; pSTAT5 strongly correlates with increased levels of integrated viral DNA in vivo, and in vitro Jak inhibitors reduce the frequency of CD4+ T cells harboring integrated HIV DNA. We show that Jak inhibitors block viral production from infected cells, inhibit γ-C receptor cytokine (IL-15)-induced viral reactivation from latent stores thereby preventing transmission of infectious particles to bystander activated T cells. These results show that dysregulation of the Jak-STAT pathway is associated with viral persistence in vivo, and that Jak inhibitors target key events downstream of γ-C cytokine (IL-2, IL-7 and IL-15) ligation to their receptors, impacting the magnitude of the HIV reservoir in all memory CD4 T cell subsets in vitro and ex vivo. Jak inhibitors represent a therapeutic modality to prevent key events of T cell activation that regulate HIV persistence and together, specific, potent blockade of these events may be integrated to future curative strategies.
Author summary
HIV persists in infected hosts in a small number of CD4 + memory T cells as latently infected proviruses. Homeostatic cytokines, which play a major role in the maintenance of T cell memory, also enable the persistence of latently infected cells. The pathways downstream of these homeostatic cytokines are well known and drugs that target these pathways have been developed and have been safely used in inflammatory diseases and in myelofibrosis. We have used these drugs to inhibit the maintenance and the spread of HIV infected cells carrying latent forms of the virus. We show that ruxolitinib and tofacitinib will inhibit the expansion of these cells and their capacity to infect other cells upon reactivation. This class of drugs is currently being tested in clinical trials.
Introduction
Current antiretroviral therapy (ART) has yielded significant success in achieving long-term suppression of viral load and in improving survival of HIV infected subjects [1-4]. Even so, ART fails to eliminate a small number of cells harboring integrated, replication competent viral DNA. This HIV reservoir has represented a major limitation in eradicating HIV. The HIV reservoir has been shown to persist in central memory TCM, transitional memory TTM and effector memory TEM CD4 T cells which require exposure to γ-C receptor cytokines for their long-term persistence [5, 6]. IL-2, IL-7 and IL-15 are γ-C receptor, homeostatic cytokines involved in the maintenance of T cell memory and which activate STAT5 mediated signaling [7]. In addition, the Jak-STAT pathway is also triggered by type I and type II Interferons, two important mediators of inflammation in viral infections including HIV [8-11]. Initial attempts to purge HIV involved the use of IL-2; results of these studies while promising, since virus was undetectable, did not reach their objective as viral load rebounded upon cessation of therapy [12]. We have previously shown that IL-7 driven homeostatic proliferation contributes to HIV persistence by promoting the survival and proliferation of latently infected cells [13, 14]. Further highlighting the role of IL-7 in the expansion and maintenance of the viral reservoir, an ACTG sponsored trial (ACTG protocol number 5214; www.clinicaltrials.gov as # NCT00099671) demonstrated that IL-7, also leads to a 70% increase in the absolute numbers of CD4 T cells harboring integrated viral DNA [14], suggesting that this intervention would not be compatible with an HIV eradication strategy. IL-15, which also signals through STAT5, has also been demonstrated to induce homeostatic proliferation of CD4 T cell subsets [15-18]. Furthermore, recent in vitro and ex vivo studies with IL-15, the IL-15 superagonist (ALT-803) and IL-2 illustrated that not only did these γ-C cytokines increase viral reactivation, but they also primed the reservoir within CD4 T cells for recognition by autologous HIV-specific CD8 T cells [19].
Phosphorylation of STAT-5 (pSTAT5) is triggered following the engagement of IL-2, IL-7 or IL-15 cytokines with their receptors leading to pro-survival signaling and increased proliferation [15, 20, 21]. Given the presence of multiple binding sites for pSTAT5 within the HIV long terminal repeat (LTR) [22], IL-2, IL-7 and IL-15 enhanced viral expression from productively infected cells [14, 15, 20, 21, 23]. Interestingly, a dominant negative STAT5 inhibited Jak-induced HIV LTR activity and decreased productive HIV infection while overexpression of STAT5 enhanced virus production in unstimulated primary T cells [22]. Together, these events underscore the relationship between activation of the STAT5 pathway and production of HIV, including events that impact the establishment of latency, its maintenance, and /or expansion of the HIV viral reservoir [24-27].
Tofacitinib and ruxolitinib are two FDA-approved Jak inhibitors for long-term use for the treatment of rheumatoid arthritis, myelofibrosis, or polycythemia vera [28-30]. Ruxolitinib demonstrates potent inhibition of pro-inflammatory cytokines in vivo, including IL-6, IL-1α/β, and TNF-α [29-32], all of which have been shown to enhance HIV replication in vitro [33, 34]. Inhibition of Jak-STAT signaling by ruxolitinib was shown to significantly impede T cell homeostasis, reducing CD4 T cell numbers as well as decreasing numbers of T regulatory cells and activated (HLA-DR+) CD4 T cell populations after a few weeks of treatment [28]. Tofacitinib showed only small changes in CD3+, CD4+ and CD8+ counts and an increase in B cell counts after 24 weeks of treatment [29, 30, 35]. Attenuated activation and proliferation were not specific to CD4 T cells but were also reported for Natural Killer (NK) cells treated with ruxolitinib in vitro and also in ruxolitinib or tofacitinib treated patients where the number of mature NK cells was reduced [28, 35]. Ruxolitinib treatment was further shown to block monocyte-derived DC differentiation, DC-derived IL-12 production and activation marker expression triggered by exposure to lipopolysaccharide (LPS) [28], demonstrating the impact of Jak inhibitors on innate and adaptive immune responses.
We previously reported that ruxolitinib and tofacitinib blocked reactivation of HIV in a primary T cell latency model at physiologic concentrations, underscoring its potential to block HIV reservoir expansion and viral dissemination from latent stores [34]. Herein, we monitored the impact of these clinically approved and extensively evaluated Jak-STAT inhibitors on several stages of HIV persistence including seeding of bystander cells and HIV reactivation from latency. Overall, we show ex vivo and in vitro that these Jak inhibitors use several mechanisms to impede the seeding and maintenance of the HIV reservoir.
Results
Ex vivo markers of activation of the Jak-STAT pathway and of homeostatic proliferation are associated to the size of the HIV reservoir
We investigated the association between the Jak-STAT pathway and HIV persistence in a cohort (n = 37) of aviremic (<50 RNA copies/ml) ART-treated participants (S1 Table) to assess the in vivo relevance of Jak-STAT signaling in the establishment and maintenance of the HIV reservoir as monitored by frequencies of cells with integrated HIV DNA. Decreased CD4 numbers and immune activation are features of aberrant T cell homeostasis [13] and increased HIV reservoir size [36, 37], which was confirmed in the cohort studied here (S1 Table). We fit a linear model to identify univariate markers (Panel 1-activation markers, Panel 2-PD-1/IL-7R and Panel 3-STAT phosphorylation; S2 Table) associated with integrated HIV DNA (see methods). As previously described, the percent CD4+ T cells was significantly higher in immune responders (IR; median 34.7%) compared with non immune responders (NIR; median 22.4%) or recently treated (RT; median 24.7%) subjects (S3 Table) and increased integrated HIV DNA was associated with decreased frequencies of CD4+ T-cells (p = 0.019 [S2 Table]) [13, 14]. Indeed, we showed that increased levels of pSTAT5 expression in CD4 central memory (CM), transitional memory (TM) and in effector memory (EM) cells, all known to harbor HIV DNA, were associated with increased integrated HIV DNA (p = 0.043, 0.008, 0.001 respectively [S2 Table]). After adjusting for CD4 counts, CD4/CD8 ratio, nadir CD4 and pre-ART HIV-1 plasma RNA, CD4 CM, TM and EM pSTAT5 expression remained significantly associated with integrated HIV DNA (p = 0.031, 0.048 and 0.0048, respectively). As our cohort included immune responders (> 500 CD4 cells /mm3) and immune non responders (< 350 CD4 cells /mm3), successfully treated non-classified (SC NC) and recently treated subjects (< 1 year of treatment) (S1 Table), we determined the levels of pSTAT5 expression in each class. pSTAT5 in naïve CD4 cells was significantly lower in immune responders (median 1160 MFI) compared to immune non responders (median MFI 1500) or recently treated subjects (median 1523 MFI; p = 0.037 and 0.018 Wilcoxon rank test, respectively) (S17 Fig and S3 Table). There was no difference in STAT5 phosphorylation when comparing the different classes of subjects in CM, TM or EM CD4 cells, the subsets that encompass the bulk of HIV integrated DNA. The relationship between pSTAT5 and integrated viral DNA remains across immune responders and non-immune responders.
IL-7R engagement by its ligand leads to its internalization and to phosphorylation/activation of STAT5. Higher levels of surface IL-7R could reflect the lack of chronic receptor engagement, which would result in lower numbers of homeostatically proliferating cells and lower levels of the HIV reservoir [13, 14]. In line with this observation, we observed that increased surface expression of IL-7R on total CD4+ T-cells was associated with decreased integrated HIV DNA (p = 0.046, S2 Table), which was also significant in CD4 memory T cells (CD45RA-) (p = 0.032) and in CD4 CM cells (p = 0.032). For total CD4 T cells and CD4 memory T cells, IRs had significantly increased IL-7R levels (median 6916 MFI and 7163 MFI, respectively) compared with NIR (median 5314 MFI and 5288 MFI, respectively), RT (median 5536 MFI and 6091, respectively) or successfully treated non classified (median 6193 for CD4 memory T cells) classes of subjects; for the CD4 CM subset IL-7R levels in IRs (median 8647 MFI) was only significantly greater than ST NCs (median 7018 MFI, S3 Table). As cells from IRs are known to express lower levels of T cell activation and proliferation markers compared with NIRs [36, 37], increased expression of IL7R in IRs further supports our observation of decreased homeostatic proliferation and lower numbers of HIV infected cells in these subjects. We also monitored PD-1 expression, as it is up-regulated by γ-C receptor cytokines [38] and is a marker of cells undergoing homeostatic proliferation and the size of the HIV reservoir [37]. PD1 expression (MFI) on total CD4 (p = 0.027, S2 Table), naïve CD4 (p = 0.023), CD45RA- memory CD4 (p = 0.023), CD4 CM (p = 0.015) and CD4 TM (p = 0.037) cells was positively associated with increased integrated HIV DNA confirming previous reports [13, 14]. There was no difference in PD-1 levels (MFI) between IR and NIRs within this group of subjects however there was greater PD-1 levels on ST NC (median 9538 MFI) compared with IRs (median 7218 MFI, p = 0.048, S3 Table).
We also showed that an increase in frequencies of naïve CD4 T cells expressing Ki67 (p = 0.012, S2 Table), and cells expressing HLA-DR/CD38 (in total CD4 (p = 0.024), naïve CD4 (p = 0.008), CD45RA- memory CD4 (p = 0.03) and CD4 CM cells (p = 0.029)) were associated with increased frequencies of cells harboring integrated HIV DNA. The cells expressing HLA-DR/CD38 in naïve CD4 were correlated as well with increased frequencies of pSTAT5 positive CM, TM and EM subsets (p = 0.005, 0.027, 0.009 respectively). As observed in previous studies [36, 37], IRs had significantly lower levels of HLA-DR/CD38 in total CD4 T cells (median values for: IR-1.21%, NIR-2.36% and RT-2.22%), CD4 memory cells (CD45RA-) [median values for: IR- 1.83%, NIR-3.12%, RT- 3.55%] or CD4 CM cells (median values for: IR- 0.78%, RT- 1.38%) compared with NIRs and/or RT subjects; IRs also had lower levels of Ki67 in naive CD4 T cells (median 0.7%) than NIR (median 0.18%) (S3 Table). Heightened frequencies of cells with integrated HIV DNA were also associated with a decrease in frequencies of CD25+ cells in total CD4 (p = 0.017, S2 Table), memory (CD45RA-) CD4 (p = 0.002), as well as in CM (p = 0.0002) and TM (p = 0.009). In contrast to HLA-DR/CD38 and Ki67 expression, IRs had significantly greater levels of CD25 on CD4 T cells (median values for: IR- 40%, RT- 32%, ST NC- 29%), CD4 memory cells (CD45RA-) [median values for: IR- 56%, NIR- 42%, RT- 44%, ST NC- 42%], and CD4 CM cells (median values for: IR- 57%, NIR- 50%, ST NC- 46%) compared with NIR, RT and/or ST NC (S3 Table). These results demonstrated that up-regulated levels of pSTAT5 were associated with deregulated homeostatic proliferation and the size of the HIV reservoir and immune activation.
Within each of the panels (panel 1 -activation markers, panel 2 -PD-1/IL-7R or panel 3 -STAT phosphorylation), we next identified the combination of the univariate markers that could best predict the size of the HIV reservoir by building a multivariate regression model using feature selection [39] (see methods; Table 1). The activation markers panel showed that frequencies of total CD4+ T-cells, of naïve CD4+ cells expressing HLA-DR/CD38 (markers of immune activation) and Ki67, and of CD4+ CM cells expressing CD25 best predicted the size of the HIV reservoir (F-statistic p-value = 0.004 -Table 1). Within the PD-1/IL-7R panel, expression (MFI) of PD-1 by CM and TM cells, and of IL7R by CM cells best predicted the size of the HIV reservoir (F-statistic p-value = 0.04), while the STAT phosphorylation panel, showed only STAT5 phosphorylation on CD4+ EM cells as a predictor of the size of the HIV reservoir (F-statistic p-value = 0.001).
These multivariate models further confirm results of the univariate analyses showing the significant association between immune activation, T cell homeostasis and the HIV reservoir. These results provide further rationale for the use of interventions that target the Jak-STAT pathway i.e inhibitors of the Jak-STAT pathway that are known to inhibit immune activation and γ-C cytokine induced proliferation which could impact the size of the HIV reservoir.
Ruxolitinib and tofacitinib inhibit cytokine-induced STAT5 phosphorylation and Bcl-2 expression
We monitored, in vitro, STAT5 phosphorylation following γ-C receptor cytokine stimulation in total CD4 T cells and in the different memory subsets known to harbor the HIV reservoir in the presence or absence of the Jak inhibitors ruxolitinib and tofacitinib (Fig 1, S1 Fig and S4 Table). As expected, ruxolitinib significantly (p < 0.001) reduced IL-2, IL-7, and IL-15 induced pSTAT5 expression (Fig 1A) in a dose dependent fashion. Activity of ruxolitinib on STAT5 phosphorylation triggered by all three cytokines was completely abrogated at concentrations ≥ 0.1 μM (Fig 1A) although this inhibitor had a significantly stronger impact on IL-2 induced CD4+ STAT5 phosphorylation (10- fold reduction) when compared to IL-7 and IL-15 at 0.01 μM (1.9- and 2.7- fold reduction, respectively) (Fig 1A). These results were confirmed with tofacitinib where we also observed a similar dose dependent inhibition of the frequencies of pSTAT5+ cells (Fig 1C; p < 0.001). Importantly, ruxolitinib and tofacitinib completely abrogated frequencies of pSTAT5+ cells exposed to IL-2, IL-7 and IL-15, compared with DMSO for concentrations > 0.1 μM (p < 0.01) in all CD4 memory subsets (TCM,TTM and TEM) all known to harbor HIV integrated DNA (S1 Fig and S4 Table) [13]. Both Jak-STAT inhibitors showed a stronger impact on IL-2 induced STAT5 phosphorylation in all memory subsets as we had observed in total CD4+ T cells (S1 Fig). A linear regression model to study the association between the frequencies of pSTAT5+ cells within each CD4+ memory subset with a range of Jak-STAT inhibitor concentrations showed a negative correlation between these two variables, i.e. percent pSTAT5+ cells significantly decreased with increasing concentrations of ruxolitinib, specifically in CM and TM cells when exposed to IL-2 (p-values < 0.05) and with increasing concentrations of tofacitinib, in CM and TM cells when exposed to IL-15 and IL-2, while in the EM cells when exposed to IL-7 (at p-values < 0.05) (S4 Table and S1 Fig). These inhibitors also significantly (p < 0.01) decreased IFN-α stimulated STAT5 phosphorylation in CD4+ T-cells, IFN-α induced STAT1 phosphorylation in CD14+ monocytes and CD4+ T-cells and IL-10 induced STAT3 phosphorylation in CD14+ monocytes and CD4+ T-cells (S2 Fig) in a dose dependent fashion highlighting a possible role for these inhibitors in lowering the levels of hyper immune activation.
We also monitored for the upregulation of Bcl-2 expression triggered by these homeostatic cytokines; Bcl-2 is a transcriptional target of STAT5, which enhances the survival of cells exposed to IL-2, IL-7 and IL-15 [20, 40]. Both ruxolitinib (Fig 1B) and tofacitinib (Fig 1D) significantly (p < 0.01) reduced IL-2, IL-7, or IL-15-induced Bcl-2 levels (MFI) of expression in a dose dependent fashion to levels similar to the unstimulated control. Significant reduction of Bcl-2 expression (p < 0.01) was achieved at concentrations of the Jak-STAT inhibitors as low as 0.01 μM for IL-2 and IL-15 induced Bcl-2 expression (1.3- and 1.5- fold reduction, respectively). IL-7 induced Bcl-2 regulation required higher concentrations of these inhibitors as inhibition of IL-7 induced Bcl-2 levels (MFI) was only observed at concentrations ≥ 0.1 μM of ruxolitinib and tofacitinib (Fig 1B and 1D). Using a linear regression model that included Bcl-2 levels (MFI) within each CD4+ memory subset as the dependent variable and Jak-STAT inhibitor concentration as the independent variable, we showed that Bcl-2 expression (MFI) also significantly (p < 0.05) decreased with increasing concentrations of ruxolitinib or tofacitinib in all memory subsets (CM, TM and EM) upon exposure to IL-2, IL-15 and IL-7 (S4 Table and S1 Fig).
Altogether these results confirmed the inhibitory impact of ruxolitinib or tofacitinib on the activation of the Jak-STAT pathway in all CD4+ memory T cell subsets known to harbor the HIV reservoir.
Ex vivo and in vitro inhibition of HIV replication by ruxolitinib and tofacitinib
Our group previously demonstrated that ruxolitinib and tofacitinib confer submicromolar anti-HIV-1/2 activity in human PBMCs and in macrophages without demonstrable toxicity [34], however further mechanistic information about how these drugs modulate key events involved in the expansion and maintenance of the HIV viral reservoir have not been explored. Herein we show that both inhibitors significantly (p < 0.0001) reduced p24 production by CD4+ T cells isolated from viremic donors and stimulated with CD3/CD28 when compared to DMSO controls (Fig 2A and 2B and S3 and S4 Figs for data on individual donors). This inhibition was observed when tested in the absence of ART (to observe the effect of the Jak inhibitors on viral replication and de novo infection of cells) and in the presence of ART (to observe the effect of the Jak inhibitors when spreading of HIV infection is inhibited, hence on viral production). These results showed that Jak inhibitors prevented uninfected primary CD4 T cells from HIV infection (described in detail under bystander infection assay) and inhibited the production of HIV from infected cells (ruxolitinib antiviral potency with ART; EC50 0.17 μM; EC90 6.2 μM; ruxolitinib antiviral potency without ART; 0.007 μM; EC90 0.26 μM).
Since ruxolitinib and tofacitinib decreased γ-C-cytokine induced Bcl-2 expression (Fig 1), we monitored cell viability of CD4+ T cells isolated from viremic donors and cultured for 6 days with CD3/CD28 and increasing concentrations of Jak inhibitors. Decreased viability was observed in 3 of the 5 donors at 0.1 μM or greater ruxolitinib or tofacitinib (S15 Fig, Panel A); using a linear regression model, addition of either inhibitor to anti-CD3/CD28 stimulated cell cultures led to significantly lower frequency of viable cells (%AnnexinV-LIVE/DEAD-) compared to DMSO taking into account inhibitor concentration and adjusting for donor. In contrast, we did not observe a loss in viability of in vitro infected cells after 3 day culture with Jak inhibitors plus ART (S15 Fig, Panel B); and no loss in viability was observed after 6 day culture of uninfected cells treated with IL-2 or IL-7 (n = 3) and in 2 of 3 uninfected donors treated with IL-15 in the presence of Jak inhibitors (S15 Fig, Panel C). As HIV-1 infected subjects are known to express lower levels of Bcl-2 [41], decreased viability as a result of inhibition of Bcl-2 expression may be a potential mechanism of Jak inhibitors on blocking viral persistence along with HIV infection in CD4 cells from viremic donors.
The inhibition of viral replication and production mediated by ruxolitinib and tofacitinib ex vivo was confirmed in an in vitro HIV infection model using a CXCR4-tropic, GFP tagged virus (eGFP NL4-3 replication competent HIV-1 reporter virus) that was used to identify productively infected cells by flow cytometry. At day 6 post infection, ruxolitinib and tofacitinib (0.1, 1.0, and 10 μM) significantly (p < 0.05) reduced extracellular p24 production (Fig 2C) as well as the frequency of HIV-GFP+ CD4+ cells (Fig 2D). Both inhibitors significantly (p < 0.01) decreased CCR5 surface expression on CD4 T cells that was upregulated as a result of anti-CD3/CD28 stimulation in viremic subjects (n = 5) whereas the levels of CXCR4 were not impacted by the addition of these inhibitors (S5 and S6 Figs). Since CXCR4 expression remained unchanged, decreased viral production and viral spread may be mediated through a mechanism beyond entry such as reduced T-cell activation (S16 Fig). As CCR5 dependent viral strains are mostly prevalent in vivo [42-44], these results indicate that inhibition of anti-CD3/28 induced CCR5 expression by ruxolitinib could have an impact on viral spread and dissemination in an infected untreated host [42, 43] and specifically in tissues from ART treated subjects where HIV replication still prevails in spite of the presence of ART [33, 45].
Since γ-C cytokines promote the activation of the Jak-STAT pathway (Fig 1) [8, 9, 11], paracrine or autocrine inhibition of IL-2, IL-7, and IL-15 induced signaling by ruxolitinib could be in part responsible for the inhibitory effects exerted by this compound on p24 production. Addition of an exogenous source of IL-7 (30 ng/ml) in ex vivo CD4 T cell cultures from viremic donors reversed the block by ruxolitinib at an antiviral EC50 concentration of 33 nM conferred on extracellular HIV production (p < 0.01), measured by p24 after 6 days in culture (S7 Fig; n = 4). More specifically, exogenous IL-7, which signals largely through STAT5, reverses latency, underscoring the link between IL-7, the ability to control latency, and STAT5-mediated signaling.
Jak inhibitors reduce the frequency of cells harboring integrated HIV viral DNA ex vivo and block IL-15 induced reactivation of latent HIV ex vivo
We monitored the frequency of cells harboring integrated viral DNA in cultures of T cells obtained from viremic donors that were activated by T cell receptor (TCR) engagement. These experiments aimed at measuring the impact of ruxolitinib on the maintenance of the existing reservoir as cultures were generated in the presence of ART. Cultures were also performed in the absence of ART to measure the seeding of the viral reservoir, as under these conditions infected cells that can produce virions will infect new T cells. Ruxolitinib and tofacitinib significantly (p < 0.05) decreased the frequency of cells with integrated DNA in cultures of T cells activated by TCR in the presence or absence of ART (Fig 3A and 3B) with doses of ruxolitinib as low as 0.01 μM.
IL-15, which signals through STAT5, also regulates memory T cell homeostasis [17, 18, 21], and could be involved in modulating the HIV reservoir size by promoting the persistence of cells with integrated DNA or by enhancing HIV reactivation and dissemination. Since ruxolitinib and tofacitinib inhibited STAT5-mediated signaling triggered by IL-7 and IL-15 (Fig 1) we sought to define the impact of ruxolitinib on IL-15 induced reactivation of latent HIV in CD4 T cells from aviremic subjects (Fig 3C-3E). Memory CD4+ T cells were isolated from ART treated aviremic donors (n = 3), activated with a concentration of IL-15 that is known to activate Jak-STAT signaling (10 ng/ml) (Fig 3D) or CD3/CD28 (Fig 3E) and maintained with or without 1 μM ruxolitinib in the presence of ART. We showed in preliminary experiments that 10 ng/ml of IL-15 triggered optimal STAT5 phosphorylation and viral reactivation. Extracellular viral RNA copies were quantified by qRT-PCR six days post IL-15 activation of purified memory CD4 T cells. Indeed, IL-15 increased reactivation of latent HIV (200 to > 1,000-fold versus non stimulated controls), although to a lesser extent when compared to the CD3/28 control (> 30,000 fold versus non stimulated controls; Fig 3C-3E). However, ruxolitinib was found to significantly reduce (p < 0.01) IL-15 and CD3/28 induced reactivation of HIV from latently infected cells, resulting in values similar to unstimulated controls. Additionally, our results demonstrated that ruxolitinib and tofacitinib decreased anti-CD3/28 induced T cell proliferation (dilution of Cell Trace Violet; S8 and S9 Figs) and activation (CD25, CD38/HLA-DR and PD-1 expression; S8 Fig and S10-S12 Figs) and more importantly, led to a decrease in the frequency of p24+ cells (Fig 2) as well as a reduction in cells harboring integrated HIV provirus (Fig 3) and decreased IL-15 induced HIV reactivation (Fig 3). Each of these events mechanistically signals through STAT5, further highlighting the role of STAT5, and subsequent block by Jak inhibitors, in controlling these key events that drive viral persistence. Altogether these results indicate that Jak-STAT inhibitors can negatively impact de novo seeding and the maintenance of the HIV reservoir.
Ruxolitinib inhibits HIV bystander infection in human PBMCs
The impact of ruxolitinib on the magnitude of the HIV reservoir could result from the inhibition of TCR triggered CD4 T cell activation and proliferation (S8-S12 Figs). It could also result from the direct antiviral activity of the Jak-STAT inhibitors (Fig 2A and 2B). Finally, decreased frequencies of cells harboring HIV integrated DNA could result from the diminished infection of bystander activated T cells due to the effects of these compounds on T cell activation and viral replication. We sought to determine if the presence of ruxolitinib could prevent the transfer of infectious viral particles to uninfected bystander cells or upon the formation of a virological synapse [44]. We developed an in vitro model to assess the impact of ruxolitinib on infection of bystander cells (schematic: Fig 4A; representative dot plots: Fig 4B) where bystander cells were labeled with cell trace violet (CTV+) and stimulated with CD3/28 in the presence of increasing concentrations of ruxolitinib to inhibit their activation. Unlabeled cells, (CTV-), were stimulated with CD3/28 in the absence of ruxolitinib, prior to infection with eGFP NL4-3. Infected CTV-negative cells and bystander CTV-positive cells were then co-cultured for two days to determine the number of ruxolitinib treated bystander cells susceptible to infection by eGFP virus (CTV+GFP+). As expected Ruxolitinib blocked proliferation (CTV-lo) of bystander cells (range 23% - 94%; p < 0.01 compared to no drug controls; paired t test) at all concentrations tested (Fig 4D). Cultures that did not include ruxolitinib led to the infection of fifty percent of bystander CTV+ cells (CTV+GFP+ cells) while addition of ruxolitinib led to a 75-80 percent reduction (p < 0.001) in CTV+GFP+ infected bystander cells. Our results show that ruxolitinib inhibition of Jak-STAT signaling impacts the susceptibility of uninfected cells to infection by HIV-1 at concentrations ≥ 0.01 μM (Fig 4B: representative dot plots, and Fig 4C: graphical representation) which are comparable to the ex vivo EC50 of 0.007 μM for ruxolitinib in CD4 T cells from viremic donors and well below the EC90 of 0.26 μM (Fig 2A). These findings suggest that Jak-STAT inhibitors could synergize with ART and decrease seeding of the HIV reservoir [34] in uninfected bystander cells.
Ruxolitinib does not inhibit TCR downstream proximal signaling and function
An HIV cure will involve the elimination of residual persistently infected cells by immune effector mechanisms that include HIV specific T cells. Therefore therapeutic strategies such as Jak-STAT inhibition should not have an impact on T cell effector functions. We monitored the impact of ruxolitinib on the early and late signaling events downstream of TCR activation (n = 3). Fig 5A evaluated the effect of ruxolitinib on phosphorylation of SLP76 and CD3 zeta, two early events of the T cell receptor-signaling cascade [16, 46]. Fig 5A shows that ruxolitinib did not significantly alter CD3 zeta and SLP76 phosphorylation at all physiological concentrations (steady-state plasma concentrations found in vivo for all doses of ruxolitinib) tested (0.01, 0.1, and 1.0 μM) (S13 Fig- gating strategy and Fig 5A).
We next monitored the impact of ruxolitinib on late TCR signaling events such as the capacity of CD4+ T cell to produce effector cytokines (TNF-α, IL-2 and IFN-γ) when stimulated with CD3/CD28 (Fig 5B and S14 Fig) Ruxolitinib concentrations of 0.01, 0.1 and 1.0 μM did not alter cytokine production of IL-2, TNF-α, or IFN-γ single positive cells or IL2+TNFα+IFNγ+ polyfunctional CD4+ and CD8+ T-cells, as quantified by intracellular flow cytometry (Fig 5B and S14 Fig). Decreased cytokine production was observed only at 10 μM, which is above the Cmax and steady-state concentrations of ruxolitinib found in vivo [31]. Similarly, upstream events of TCR signaling and production of TNF-α, IL-2 and IFN-γ were inhibited only upon addition of a supra-physiological concentration (10 μM) of ruxolitinib, [32]. When we monitored cytokine production after gag peptide stimulation of PBMCs from stably treated, HIV-infected subjects, statistically significant differences in gag-specific antigenic responses were observed only at 10 μM ruxolitinib for TNF-α and at 0.1 μM for TNFα+IFNγ+ in CD4+ cells (Fig 5C and S18 Fig). There were no significant differences in gag-specific responses in CD8+ T-cells at the different concentration of ruxolitinib, however only 3 of the 6 subjects tested responded to gag-peptide stimulation and two subjects demonstrated decreased cytokine production at ruxolitinib concentrations ≥ 0.1 μM (Fig 5C and S18 Fig). These results demonstrate that ruxolitinib does not inhibit initial TCR function in both CD4 and CD8 T-cells nor HIV specific responses in CD4 T-cells with some donor-specific responses in CD8 T-cells, highlighting the ability of ruxolitinib to specifically inhibit signal transduction pathways that alter T cell proliferation and expansion of the HIV reservoir without interfering with the development of effector antiviral T cell functions.
Pharmacokinetic simulation for 10 mg and 20 mg bis in die (twice daily; bid) dosing of ruxolitinib in humans
Pharmacokinetic simulations were performed using reported plasma drug levels [31, 32] to determine if the plasma concentrations of drug during either 10 mg (Fig 6A), or 20 mg (Fig 6B) bid oral ruxolitinib treatment, could be correlated with in vitro efficacy. To assess PK/PD (pharmacokinetics/pharmacodynamics) relationship relative to the in vitro and in vivo efficacy reported herein, baseline population pharmacokinetics (PPK) model parameters without explicit patient covariates were used to construct a PPK model. Our model used δ2 as the ↔ subject variance (IIV) for that PK parameter, and ⌠2 = residual variance (within subjects). Log-normal error structure was used and 1,000 theoretical participants administered 10 mg ruxolitinib twice per day were modeled. Resulting computed percentile (P10, P25, P50 P75, P90) plasma concentrations versus time were used to mimic in vivo doses of 10 and 20 mg bid, which represent the low and high FDA-approved doses.
The model presented in Fig 6 shows that the pharmacodynamic effects (anti-HIV effects on viral reservoirs, antiviral, and HIV-induced activation/proliferation) of ruxolitinib falls within the Cmax and steady-state range for all FDA approved doses of the drug (dotted lines; 0.01-1.0 μM: Fig 6), highlighting that the concentrations required to inhibit these pro-HIV events are equivalent in vivo for individuals taking ruxolitinib, even for the 10 mg bid dosing, which represents the lowest effective dose in humans.
Discussion
The IL-7 and IL-15 γ-C cytokines are essential for long term maintenance of memory T cells by regulating homeostatic proliferation and enhancing cell survival which paradoxically leads to HIV persistence (Fig 7A and 7B) [13-15, 17, 18, 20]. Herein, we demonstrate that engagement of this pathway as shown by increased STAT5 phosphorylation is positively associated to the frequencies of cells with integrated HIV DNA and to immune activation which has also been shown to be positively associated with the size of the HIV reservoir. Conversely, we show an inverse relationship between IL-7R levels and integrated HIV DNA in CD4 T cells from ART-treated HIV-infected donors (Table 1; Fig 7A and 7B), confirming the role attributed to STAT5 signaling, in increasing the size of the HIV reservoir by promoting the survival of infected cells expressing receptors for these homeostatic cytokines and by triggering the homeostatic maintenance of these cells (Fig 7A and 7B).
Importantly, we show that ruxolitinib or tofacitinib inhibit signaling of the Jak-STAT pathway in all memory T cell subsets known to harbor the HIV reservoir. Our results show that Jak inhibitors impede T cell activation and proliferation both critical for HIV replication and seeding of the reservoir without impeding effector CD4 and CD8 HIV specific responses (Fig 7B-7D). Indeed we show that ex vivo and in vitro, exposure of T cells to Jak-STAT inhibitors prevents CD3/28-induced up-regulation of several cell surface markers (PD-1, HLADR/CD38 and CCR5) that are expressed by cells that harbor integrated DNA (Table 1, S8-S12 Figs; Fig 7C) [36, 37]. Although Jak inhibitors are not specific to the STAT5 pathway (S2 Fig), careful design of these experiments in the context of cytokines that signal almost exclusively through STAT5, allowed us to assess how blockade of STAT5 contributes to sentinel events driving reservoir establishment, maintenance, and expansion. We also demonstrated that ruxolitinib blocks IFN-α, induced pSTAT1 (CD4+ T cells and CD14+ monocytes) and IL-10 induced pSTAT3 (CD4+ T cells and CD14+ monocytes), underscoring that phosphorylation of other STATs may also confer the observed outcome, in combination with blockade of STAT5 (S2 Fig). The involvement of the STAT pathway was confirmed by the findings that exogenous IL-7 could overcome the Jak-inhibitor mediated block on latency reversal, and that Jak inhibitors block IL-15 induced viral reactivation. Importantly, these ex vivo and in vitro effects were obtained at concentrations (Jakafi.com; [31, 32]) that are achieved during steady-state of this drug in humans for all FDA approved doses. Of note, we previously reported that no apparent toxicity was observed for ruxolitinib or tofacitinib up to 50 μM in T cells and macrophages [34].
Furthermore, decreased viability was primarily observed after the cells from HIV infected subjects were treated with anti-CD3/28 which may be a consequence of HIV infected subjects having lower levels of Bcl-2 and consequently impaired cell survival (20, 44-45).
Our results show that ruxolitinib and tofacitinib can exert direct antiviral potency in infected cells isolated from viremic individuals and in primary CD4 T cells infected in vitro (Fig 2). These Jak inhibitors could confer a direct antiviral activity by blocking the phosphorylation of STATs and subsequent binding of this transcription factor to several STAT5 binding sites within the HIV-1 LTR (Fig 7E) [7, 22, 27]. Decreased virus production (summarized in Fig 7C-7F) could also be a consequence of the Jak-STAT induced down-regulation of T cell proliferaton. CD4 T cells exposed to these inhibitors should show decreased expression of transcription factors (NFAT-c, NF-κ-B) expressed by activated T cells [47] and which are required for HIV LTR mediated transcription [48]. Herein, we report that Jak inhibitors demonstrate nanomolar inhibition of viral replication in ex vivo CD4 T cells from viremic individuals, even in the presence of an EC99 of ART (Fig 2). Therefore, Jak inhibitors may block viral replication in ART treated subjects [33] which highlights their potential as a therapeutic modality, since they could block potential residual viral replication in pharmacological sanctuary sites in vivo where antiviral agents do not reach optimal concentrations [45]. Additionally, our results suggest that preferential cell death or shortened lifespan (down regulation of Bcl-2 across T cell subsets; S1 Fig) conferred by Jak inhibitor exposure lead to a reduction in the overall size of the viral reservoir. Dissemination of HIV to bystander cells is also inhibited by addition of Jak inhibitors further confirming the capacity of this class of molecules to prevent further seeding of the reservoir.
CD4 T regulatory cells (Tregs) are the major rheostat of immune homeostasis [49], in addition to suppressing cell mediated immunity to viral infections including HIV infection [50-52]. These cells are dependent on IL-2 and Jak-STAT signaling for their persistence [53]. STAT5-driven homeostatic proliferation is critical for the maintenance of absolute numbers of Tregs [53, 54] where increased number of Tregs could lead to further dysregulation of immune homeostasis (Fig 7, panel G). Jak inhibitors would potentially prevent the increase in Treg homeostatic proliferation in HIV infected subjects which was demonstrated in recent studies of subjects treated with these Jak Inhibitors for myeloproliferative disorders [55]. Inhibition of Treg function would trigger potent cell mediated immune responses that could control residual viral replication [5, 56, 57]. Indeed, ruxolitinib inhibits the production of IL-6, TNF-α IFN-α/β, D-dimer, IL-10, and IL-1α/β in vivo [32], markers associated with immune senescence and maintenance of the HIV reservoir. Together, these findings demonstrate that the anti-inflammatory properties of ruxolitinib that are associated with reduced activation and proliferation of T cells in an HIV-infected host could result in reduced bystander cell infection (Fig 7F), diminished levels of homeostatic proliferation (Fig 7B), and decreased permissiveness of uninfected cells by down-regulating the CCR5 co-receptor (Fig 7C).
Lastly, our results also show that this class of compounds can impact HIV reactivation without affecting TCR signaling and function. Indeed, we show that the most proximal events of TCR signaling as well as TCR induced effector function (i.e. cytokine production) are not affected by Jak inhibitors, although these molecules inhibit TCR induced proliferation. This is in line with previous observations showing that the threshold of MHC/peptide complexes required to trigger antigen specific effector functions is different than that required to induce T cell proliferation. Huang et. al. showed that a single peptide-major histocompatibility complex ligand interaction with a CD4+ T cell is enough to stimulate cytokine production in CD4+ T cells [58]. Weak signaling of the TCR still allows for phosphorylation of the TCR signaling pathway and production of cytokines, however only optimal TCR signaling may allow for T cell differentiation and proliferation [46, 58, 59]. This is consistent with our data, wherein steady-state plasma concentrations found in vivo (defined as physiological concentrations) did not impair TCR signaling or cytokine production while at the same time decreased CD4 T cell proliferation. Importantly, ruxolitinib did not block HIV gag-peptide T cell responses in CD4 T cells and only decreased CD8 response in 2 of 3 subjects at concentrations ≥ 0.1 μM, indicating that ruxolitinib will not negatively impact HIV specific responses in CD4 T-cells.
Jak inhibitors represent a potential therapeutic modality that addresses a clinical need which traditional direct-acting antiviral agents that interfere with steps in the viral replication cycle have not been successful. We demonstrate that Jak inhibitors confer STAT5-mediated block in homeostatic proliferation, and inflammation-driven HIV reservoir expansion. The effects of Jak inhibitors, described herein on mechanisms that promote HIV dissemination and persistence may be advantageous as an alternative therapeutic intervention that could be implemented in future curative strategies. The impact of ruxolitinib on inflammation associated with HIV infection, which plays an important role in seeding viral reservoirs, viral persistence and ongoing low-level replication in sanctuary sites is currently being evaluated in an NIH-ACTG sponsored 21-site Phase 2a study in humans that is currently enrolling.
|
| | | | |
|