Inhibitor of apoptosis, IAP, genes play a critical role in the survival of HIV-infected macrophages

Latent viral reservoirs of HIV-1 that persist despite antiretroviral therapy (ART) are major barriers for a successful cure. Macrophages serve as viral reservoirs due to their resistance to apoptosis and HIV-cytopathic effects. We have previously shown that inhibitor of apoptosis proteins (IAPs) confer resistance to HIV-Vpr-induced apoptosis in normal macrophages. Herein, we show that second mitochondrial activator of caspases (SMAC)-mimetics (SM) specifically induce apoptosis of monocyte-derived macrophages (MDMs) infected in vitro with a R5-tropic laboratory strain expressing heat stable antigen, and GFP-expressing HIV, chronically infected U1 cells, and ex-vivo derived MDMs from naïve and ART-treated HIV patients. SM-induced cell death was found to be mediated by IAPs using IAP siRNAs, was independent of endogenously produced TNFα and was attributed to the concomitant downregulation of IAP-1/2 and the receptor interacting protein kinase-1 degradation following HIV infection. Altogether, modulation of the IAP pathways may be a potential strategy for selective killing of HIV-infected macrophages in vivo. Summary After more than 30 years of rigorous and intensive research since the identification of HIV-1, much progress has been made in understanding and controlling the pathogenesis of the virus. However, successful cure is currently unavailable. HIV-1 can remain undetected in various cell types, including memory T cells and macrophages, which make it difficult to achieve viral clearance without inciting cell death in infected cells. The “shock and kill” approach aims to reawaken dormant integrated virus and boosts host’s immune system for viral clearance in latently infected CD4+ T cells. However, to completely eradicate HIV in infected individuals, it is imperative to eliminate both CD4+ T cells and myeloid tissue reservoirs. Here we show that inhibition of the inhibitor of apoptosis (IAP) pathway, a cellular signalling pathway responsible for controlling cell death, by IAP inhibitors, smac mimetics can be utilized to kill HIV-infected macrophages. Deletion of cellular IAP proteins using smac mimetic, a synthetic anti-cancer compound currently being tested in several clinical trials, rendered HIV-infected macrophages susceptible to cell death. Herein, our results suggest that modulation of the IAP-associated signaling pathways may be a potential strategy for selective killing of HIV-infected macrophages.


Summary
After more than 30 years of rigorous and intensive research since the identification of HIV-1, much progress has been made in understanding and controlling the pathogenesis of the virus.
However, successful cure is currently unavailable. HIV-1 can remain undetected in various cell types, including memory T cells and macrophages, which make it difficult to achieve viral clearance without inciting cell death in infected cells. The "shock and kill" approach aims to reawaken dormant integrated virus and boosts host's immune system for viral clearance in latently infected CD4+ T cells. However, to completely eradicate HIV in infected individuals, it is imperative to eliminate both CD4+ T cells and myeloid tissue reservoirs. Here we show that inhibition of the inhibitor of apoptosis (IAP) pathway, a cellular signalling pathway responsible for controlling cell death, by IAP inhibitors, smac mimetics can be utilized to kill HIV-infected macrophages. Deletion of cellular IAP proteins using smac mimetic, a synthetic anti-cancer compound currently being tested in several clinical trials, rendered HIV-infected macrophages susceptible to cell death. Herein, our results suggest that modulation of the IAP-associated signaling pathways may be a potential strategy for selective killing of HIV-infected macrophages.

Introduction
Macrophages (Mφ) are permissive to productive infection with HIV and a source of viral progeny for transmission to other cell types such as T cells [1][2][3][4][5][6][7]. HIV-infected Mφ are widely distributed in tissues such as gastrointestinal and other mucosal tissues, lymph nodes and within the central nervous system where they have a life span extending from months to years [8][9][10][11][12][13].
While several recent studies support that Mφ serve as a major non-T cell HIV reservoir [30][31][32][33][34][35][36][37][38], the role of Mφ in HIV infection and persistence has been conclusively demonstrated by employing humanized BLT and myeloid only mice (MoM mice containing myeloid cells devoid of T cells). Honeycutt et al show that replication competent virus could be recovered from tissue Mφ, and the transfer of infected Mφ into uninfected animals resulted in sustained infection demonstrating that Mφ are genuine targets for HIV infection in vivo [3]. Further, they demonstrated that HIV persists in Mφ following suppressive ART in vivo in MoM model [39]. Therefore, to completely eradicate HIV in individuals on ART, it is imperative to eliminate both CD4+ T cells and myeloid tissue reservoirs. Most research to date has focused on eliminating the latent reservoir of CD4+ T cells by employing strategies to reactivate HIV in T cells and elimination of reactivated HIV-infected cells by host immunity [40][41][42]. However, approaches towards killing of HIV-infected Mφ in vitro or in vivo are not well studied. Two recent studies have attempted to clear Mφ reservoir by targeting infected Mφ with CSF-1 receptor antagonists [43] and galactin-3 [44] with some success.
In order to devise strategies to eliminate HIV-infected Mφ, it is imperative to identify apoptosis-related genes and signaling proteins involved in resistance of HIV-infected Mφ to apoptosis. The mechanism underlying resistance of infected Mφ to HIV-induced apoptosis may relate to the differential expression of pro-and anti-apoptotic genes including inhibitors of apoptosis (IAP) proteins [15,45]. The role of IAPs has been studied by employing antagonists of second mitochondria-derived activator of caspases (Smac), Smac mimetics (SM). SMs are small peptides that competitively inhibit Smac-IAP-1/2 interactions and repress anti-apoptotic functions of IAP proteins. Recently, IAP1/2 and survivin, another member of the IAP family were suggested to be involved in survival of HIV-infected CD4+ T cells [46,47]. In addition, IAPs have been implicated in protection against hepatitis B infection and in the reversal of HIV latency in CD4+ T cells [48,49]. Using HIV-Vpr as an apoptosis-inducing agent, we have shown a protective role for IAP genes in resistance to cell death in Mφ [50][51][52]. CpG-induced protection against apoptosis and mitochondrial depolarization in monocytic cells was shown to be mediated by c-IAP-2 induction [50,52]. Moreover, down regulation of IAP-1/2, by using siRNAs and SMs, sensitized Mφ to Vpr-induced apoptosis [51]. Therefore, strategies based on suppressing IAPs by employing SMs, may be useful in killing HIV-infected Mφ. Herein, we show that SMs induced apoptosis in in vitro HIV-infected Mφ and that this may occur through the concomitant down regulation of both IAPs and receptor interacting protein kinase-1 (RIPK-1).

U937 cells
SMs bind to cIAP1/2 and promote their E3 ligase activity which leads to their autoubiquitination, subsequent proteasome degradation and apoptosis [53,54]. We have previously shown that cIAP1/2 genes play a protective role in mediating survival of Mϕ in response to Vprinduced cell death [50][51][52]. To determine whether SMs impact apoptosis in HIV-infected Mϕ, chronically infected U1 cells and uninfected counterpart U937 cells were treated with SM-LCL161 followed by assessment of cell death by PI staining and flow cytometry. SM treatment induced significant cell death in U1 cells but not in U937 cells (Fig 1A). To determine whether differentiation of U937 render these cells susceptible to SM-induced apoptosis, U937 and U1 cells were differentiated with PMA. Similar to the effect of SM on undifferentiated U1 cells, SM -LCL161 induced significant cell death in differentiated U1 cells but not in differentiated U937 cells ( Fig 1B). Specific killing of HIV-infected U1 cells was further confirmed by showing cleavage of caspase-3 in U1 but not in U937 cells (Fig 1C).

SMs induce cell death in in vitro HIV-infected MDMs and MDMs derived from HIVinfected patients
To validate above findings in primary MDMs, we first verified the functional activity of SM by treating HIV-infected MDMs with LCL161 and observed degradation of both cIAP1 and cIAP2 (Fig 2A) as reported earlier [51,55]. The in-vitro HIV CS204-infected MDMs were treated with SM-LCL161 followed by assessment of cell death by PI staining and flow cytometry. SM-LCL161 induced significant cell death of HIV CS204 -infected MDMs but not in mock-infected MDMs ( Fig   2B). Representative histograms of the intracellular PI staining are shown (Fig. 2C) Apoptosis has been shown to induce viral activation and replication in latently infected U1 and ACH2 cell lines [56]. In addition, Pache et al have shown that SMs can affect viral transcription in infected CD4 + T cells via NF-κB dependent signalling [49]. To determine if SMs affect HIV replication in Mϕ, in vitro HIV-infected MDM were treated with SM -LCL161 for 48 hr followed by analysis of p24 secretion. Interestingly, virus replication in primary HIV-infected MDM ( Fig 2F) and in HIV-infected U1 cells (Fig 2G) was not affected by SM treatment.

Smac mimetics specifically kill HIV-infected MDMs
Based on above results, it is unclear if SMs are killing HIV-infected and/or bystander uninfected HIV-exposed MDM. To examine this, we employed a R5 laboratory strain of HIV-1, HIV-Bal-HSA, expressing mouse HSA (CD24). Expression of CD24 by HIV-infected cells can be used to identify infected cells by flow cytometry using FITC-conjugated anti-mouse HSA antibody [57]. To determine whether SMs killed uninfected HIV-exposed bystander cells, MDM were infected with HIV-Bal-HSA for 7 days followed by treatment with either SM-AEG40730 or SM-LCL161 for another two days. Specific killing of HSA-expressing (ie HIV-infected) and HSAnegative (HIV uninfected) cells by SM-AEG 40730 or SM-LCL161 was quantified by counter staining with BV-711 labelled Annexin-V as above. SM-AEG40730 and SM-LCL161 killed significantly high numbers of HIV-HSA-expressing (HIV-infected) cells compared to either the mock or HSA-negative (HIV-uninfected/bystander, HIV-exposed) cells (Fig 3 D). However, killing of HSA-negative (HIV-uninfected/bystander) cells was relatively higher than mock-

SM-induced cell death in HIV-infected MDM is mediated by apoptosis
To determine whether SM-induced cell death in in vitro HIV-infected MDM is due to apoptosis, caspase activation was quantified based on the fluorescent signal of cleaved caspase substrates.
Treatment of HIV cs204 -infected MDM with SM-LCL161 showed activation of caspases 3, 8, and 9 in contrast to the mock-infected MDM (Fig 6A-6C). Moreover, prior treatment with zVAD-FMK, a pan-caspase inhibitor, reduced the activation of caspase-8 and 9 after SM-LCL161 treatment (Fig 6B-C). A representative histogram for the induction of caspase 3, 8 and 9 following SM treatment of HIV-infected MDM is shown (supp. Fig 1).

TNF-α mediates SM-induced apoptosis in U1 cells but not in primary HIV-infected MDM
SM-induced cell death of various tumor cells is mediated by endogenously produced TNF- following SM treatment through the activation of the non-canonical NF-κB pathway [58,59]. To determine if SM-induced apoptosis in HIV-infected MDM is due to endogenous TNFα production, SM-LCL161-treated U937, U1 cells and in vitro HIV-infected primary MDM were analyzed for TNF-α secretion. SM-LCL161 treatment resulted in low level although significant TNF-α production in undifferentiated and differentiated U937 and U1 cells (Fig 7A-D) in contrast to both in vitro mock-and HIV-infected MDM (Fig 7E). Similarly, ex vivo derived MDM from HIV-infected patients did not produce significantly higher levels of TNF-α following SM-LCL161 treatment compared to the untreated negative controls (Fig 7F).
To evaluate the impact of TNF-α in SM-induced apoptosis of primary MDM, SM-LCL161-treated MDM were stimulated with rTNF-α followed by analysis of cell death by PI staining. Treatment of MDMs with SM-LCL161 and TNFα did not result in cell death ( Fig 7G).
In contrast, rTNF-α either alone or in combination with SM-LCL161 induced significant cell death in U937 and U1 cells (Fig 7H, I) similar to that observed in various tumor cells [60,61].
These results suggest that SM-mediated killing of HIV-infected MDM is independent of TNFα.

HIV-infected MDM do not develop M1 phenotype before or after SM treatment
Macrophages polarized with IFNγ develop a M1 phenotype which is highly susceptible to SMinduced cell death (Supp. Fig 2). Therefore, it is possible that SM-induced cell death of HIV- There was no significant difference in the secretion of IL-10, IL-21, IL-13, and IL-23 between the HIV-infected and mock-infected controls (Supp. Fig 3). Remaining cytokines were not detected in either group suggesting that HIV infection of MDMs does not result in the upregulation of cytokines related to M1 phenotype. SM treatment did not affect the secretion of above mentioned cytokines including CCL20/MIP3α, IL-6, IL-23, IL-10, IL-21, IL-13, and TNFα in in-vitro HIV-infected MDMs (Supp. Fig 4) or in ex-vivo derived MDMs from HIVinfected patients (Supp. Fig 5). These results suggest that in-vitro HIV-infected MDM either before or after SM treatment did not express M1 phenotype and SM-mediated apoptosis of HIVinfected MDM is independent of M1-polarization.

HIV-infection downregulates RIPK1 in MDMs
SM-induced apoptosis of HIV-infected macrophages may be ascribed to the impaired expression of IAP-associated signalling kinases such as RIPK-1 [62,63]. RIPK-1 plays a key role in the regulation of various cellular processes such as NF-κB signalling and apoptosis [64]. Moreover RIPK-1 is a target substrate for HIV protease, a viral protein that is synthesized late in the viral life cycle and inactivates RIPK1 in HIV-infected primary CD4+ T cells [65]. To determine whether RIPK1 is similarly cleaved and inactivated in HIV-infected MDMs, in vitro mock and HIV CS204 -infected MDMs for 7 days were treated with SM -LCL161 for 2 days followed by immunoprobing for RIPK-1. HIV infection resulted in the downregulation of RIPK-1 in the presence and absence of SM-LCL161 compared to the mock infected controls (Fig 8A). This was also demonstrated by in vitro infection of MDMs with HIV CS204 for 2-8 days. Infection with HIV CS204 resulted in cleavage of RIPK1 with a relative decrease in full length RIPK1 while the cleaved RIPK1product gradually increased over time (Fig 8B).
To confirm the downregulation of RIPK-1 in HIV-infected MDM, MDMs were infected with HIV-Bal-HSA. After 9 days of infection, HIV-infected HSA-expressing MDMs were harvested by magnetic column separation based on HSA expression followed by immunoblotting for RIPK-1 analysis [57]. The negative fraction represents HIV-exposed uninfected cells that do not express HSA on their surface, and hence get eluted after the first passing of the labelled cells.

SM treatment of HIV-infected MDMs downregulates apoptosis associated signalling molecules TRAF-1 and Bid
In addition to RIPK-1, the process of apoptosis requires the fine-tuned functionality of several signalling molecules including TRAF-1/2, as well as proteins that regulate homeostasis of mitochondria such as Bid and Bax [64,[66][67][68]. We determined the expression of these signalling molecules in response to SM-LCL161 treatment of in vitro HIV CS204 -infected MDM. HIV infection resulted in the down regulation of TRAF-1 (Fig 8E, lanes 1 and 4). Treatment of HIVinfected MDM with SM-LCL161 also resulted in the downregulation of TRAF-1 compared to the mock infected and SM-treated MDMs (Fig 8E lanes 2, 3 and 5). However, TRAF-2 and Bax did not show a significant change in their expression in mock-and HIV-infected MDM as well as between SM-LCL161 treated mock and HIV-infected MDMs (Fig 8E). Bid was downregulated with increasing concentration of SM-LCL161 in the HIV-infected MDM but not in mock-infected MDM (Fig 8E). Overall, these results suggest that SM dysregulates the expression of apoptosis-associated TRAF-1 and Bid in HIV-infected MDM. To achieve eradication of HIV-1 in patients undergoing suppressive ART, it is imperative to devise strategies to eliminate HIV reservoirs in cell targets other than T cells such as Mφ.
Recently, IAPs were shown as a potent negative regulator of LTR-dependent HIV-1 transcription and leading to the reversal of HIV latency in JLat latency model system and primary T cells [49].
In addition, IAP1/2 and survivin, another member of the IAP family were suggested to be involved in survival of HIV-infected CD4+ T cells [46,47]. SM activate the non-canonical NF-κB pathway by virtue of RelA:p50 and RelB:p52 transcription factors which bind to the HIV-1 LTR region and results in the induction of virus transcription in latently infected JLat cell lines [49,69]. In addition, XIAP down regulation by flavopiridol, a cyclin-dependent kinase 9 (CDK-9) inhibitor, resulted in increased apoptosis of ACH2 cells (a chronically HIV-infected T cell line) [70]. Additionally, ablation of cIAP1/2 by SMs cleared hepatitis B virus in immune competent mouse models [48]. We and others have previously shown that ablation of cIAP1/2 by SMs does not affect survival of normal primary human Mϕ [51,71]. However, resistance of Mϕ to apoptogenic HIV-Vpr was attributed to cIAP1/2 [51]. Our results suggest that the mechanism of SM-mediated killing of HIV-infected MDMs is independent of endogenous TNF-α. SM-mediated killing has been attributed to endogenous TNF-α in cancer cells [58,59]; however, it has been reported to be independent of TNF-α in some cancer cells [78]. Given that Mϕ produce high levels of TNFα, the possibility that SM- phenotype making these cells susceptible to SM-induced apoptosis was studied. We show that in vitro infected and ex vivo derived MDM exposed to SM were not polarized into M1 phenotype suggesting that SM-mediated killing of HIV infected Mϕ was not due to M1 polarization.
Our results suggest that the mechanism of SM-mediated selective killing of U1 cells and primary MDM infected with the clinical strain, HIV cs204 is via apoptosis. The pathways of apoptosis are regulated by RIPK-1 [64,67]. In TNFα-mediated signalling, RIPK-1 is recruited in a multiprotein complex I along with TRADD, TRAF2, and cIAP1/2 to promote transcription of genes with anti-apoptotic properties such as cIAP1/2 [67]. RIPK1 is also recruited in a protein complex composed of TRADD, FADD, and caspase-8, which depending on additional proteins recruited, can induce apoptosis or necroptosis [67]. TRAF1 is an important receptor interacting protein that forms a complex with TRAF2 to transduce TNFα-induced MAPK and NF-κB activation [82]. TRAF2 is also a key determinant for SM-induced degradation of cIAP1/2 [82].

Generation of human monocyte-derived macrophages (MDM), cell lines and reagents
Human peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Ficoll Paque (GE Healthcare Life Sciences Buckingmshire, UK) from the blood of healthy donors. Human MDMs were generated from monocytes via adherence methods as previously described [55]. Briefly, 2.0x10 6 PBMCs/well were allowed to adhere for 3 hr and

HIV-1 production and infection of MDMs
The dual tropic HIV-CS204 was a gift from Dr. J. Angel (The Ottawa Hospital, Ottawa, ON, Canada). HIV CS204 stocks were produced in CD8 + depleted PBMCs from healthy donors as described earlier [83].

TNF-α ELISA and cytokine ELISA array
Human TNF-α duo set (R&D System) was used to quantify TNF-α as per the manufacturer recommendations. Briefly, the 96-well plates were preincubated with TNF-α capture antibody for 16 hr followed by blocking with 1% FBS. TNFα (1-1000 pg/mL) was used as standards. The samples were added to the plates for 16 hr followed by the detection antibodies for two hr. Next,

Western immunoblot analysis
The lysates were subjected to SDS-PAGE electrophoresis as described earlier [51,55,84].

Statistical analysis
Data was plotted using Graphpad Prism 5. Statistical significance was calculated using student t test or One-way Anova, followed by Dunnett post test. Plotted data represent the mean ± SD.

Ethics statement
Healthy participants involved in the study gave informed written consent and the protocol for