Role of RIPK1 in SMAC mimetics-induced apoptosis in primary human HIV-infected macrophages

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) induce apoptosis of monocyte-derived macrophages (MDMs) infected in vitro with a R5-tropic laboratory strain expressing heat stable antigen, chronically infected U1 cells, and ex-vivo derived MDMs from HIV-infected individuals. To understand the mechanism governing SM-induced cell death, we show that SM-induced cell death of primary HIV-infected macrophages was independent of the acquisition of M1 phenotype following HIV infection of macrophages. Instead, SM-induced cell death was found to be mediated by IAPs as downregulation of IAPs by siRNAs induced cell death of HIV-infected macrophages. Moreover, HIV infection caused receptor interacting protein kinase-1 (RIPK1) degradation which in concert with IAP1/2 downregulation following SM treatment may result in apoptosis of macrophages. Altogether, our results show that SM selectively induce apoptosis in primary human macrophages infected in vitro with HIV possibly through RIPK1. Moreover, modulation of the IAP pathways may be a potential strategy for selective killing of HIV-infected macrophages in vivo.


SMs induce cell death in HIV-infected myeloid U1 cells but not in counterpart uninfected U937 cells.
We have previously shown that the cIAP1/2 play a protective role in mediating survival of macrophages in response to Vpr-induced cell death [35][36][37] . Here, we hypothesized that SMs, the IAPs antagonists, will induce apoptosis in HIV-infected macrophages. U1 cells are chronically infected with two copies of proviral HIV DNA and show minimal constitutive expression of the virus. PMA stimulation of U1 cells induces HIV production and terminal differentiation along the mononuclear phagocytic lineage 38 similar to HIV infection of primary macrophages 39 . Thus, we first evaluated the effects of SMs on apoptosis in undifferentiated and PMA-differentiated U937 and U1 cells. U1 cells and uninfected counterpart U937 cells were treated with SM-LCL161 (LCL161) followed by assessment of cell death by PI staining and flow cytometry. LCL161 treatment induced significant cell death in undifferentiated U1 cells (p-value = 0.008, DMSO vs LCL 1 μM) but not in U937 cells (Fig. 1a). Like the effect of SM on undifferentiated U1 cells, LCL161 induced significant cell death in PMA-differentiated primary macrophage-like U1 cells (p-value = 0.006, DMSO vs LCL 1 μM) but not in PMA-differentiated U937 cells (Fig. 1b). Apoptosis of HIV-infected U1 cells was further confirmed by showing cleavage of caspase-3 in U1 but not in U937 cells by immunoblotting (Fig. 1c).

SMs induce cell death in in vitro HIV-infected MDMs and MDMs derived from HIV-infected patients.
To validate the 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 previously reported 36,40 . LCL161 induced significant cell death of in vitro HIV-1 CS204-infected (p-value = 0.003, DMS0 vs LCL 1 μM) MDMs but not in mock-infected MDMs (Fig. 2b). Representative histograms of the intracellular PI staining are shown (Fig. 2c). The p24 values in MDMs infected with HIV-1 CS204 for 7 d are shown in Fig. 2d. SM-induced cell death experiments were conducted at 48 h post-treatment with LCL161 as higher % of cell death was observed at 48 h compared to 24 h post treatment (Supp Fig. 1a). To determine whether SMinduced cell death in in vitro HIV-infected MDM was due to apoptosis, caspase activation was quantified based on the fluorescent signal of cleaved caspase substrates. Treatment of HIV-1 CS204-infected MDM with LCL161 showed activation of caspases 3, 8, and 9 in contrast to the mock-infected MDM (Fig. 2e). A representative histogram for the induction of caspase 3, 8 and 9 is shown (Supp Fig. 2).
Further, to determine whether macrophages derived from HIV-infected patients are similarly prone to SM-induced cell death, MDMs were generated from ART-treated and ART-naïve HIV-infected patients and treated with LCL161. Consistent with in vitro infection studies, ex vivo derived MDMs from treatment-naïve (p-value = 0.03, DMS0 vs LCL 1 μM) and ART-treated patients (p-value = 0.03, DMS0 vs LCL 1 μM) showed a significantly increased susceptibility to LCL161-induced cell death in a dose-dependent manner (Fig. 2f, g). Moreover, ART treatment did not affect the expression of cIAP1 and cIAP2 in uninfected or HIV-infected macrophages (Supp Fig. 1b) suggesting that ART treatment did not affect the IAP pathway in uninfected or HIV-infected macrophages.

SMs specifically kill HIV-infected MDMs. The frequency of HIV-infected macrophages in in vitro HIV-
infected macrophages may not exceed 10%. However, the HIV-infected myeloid cells are rarely detected in the blood of HIV-infected individuals 41 ; hence the frequency of HIV-infected macrophages in ex vivo derived macrophages from the HIV-infected patients may be extremely low. But the above results show killing of 20-30% macrophages (Fig. 2b, f, g) following treatment with SMs. Therefore, it was important to ascertain whether SMs were killing HIV-infected cells alone, uninfected bystander cells alone or both. To examine this, we employed a R5 laboratory strain of HIV-1, HIV-Bal-HSA (HIV-HSA), expressing mouse HSA (CD24). Expression of HSA by HIV-infected cells can be used to quantify infected cells by flow cytometry using FITC-conjugated anti-mouse HSA antibody 42 . HIV-HSA infection rates in MDMs ranged from 5 to 30% over 15 d post infection depending upon the donor variability (Fig. 3a). Flow cytometric analysis of HIV-HSA-infected macrophages revealed two distinct HSA-positive (HSA +) and HSA-negative (HSA −) populations of macrophages (Fig. 3b). Following treatment with SM-AEG40730 (AEG40730), HIV-HSA-infected and HIV-HSA negative macrophages (unin- www.nature.com/scientificreports/ fected, HIV exposed bystander macrophages) were analyzed for infection rate (% HSA + cells) and apoptosis as quantified by counter staining with Annexin-V labelled with BV711 expression. The gating strategy is shown in Fig. 3b. Analysis of infection rate (% HIV-HSA FITC + cells) following AEG40730 treatment did not reveal significant differences between control DMSO and AEG40730 treated macrophages (Fig. 3c, left panel). As a control, AEG40730 treatment did not show significant increase in non-specific FITC-HSA autoflourescence in mock uninfected macrophages (Fig. 3c, left panel). Representative histograms are shown (Fig. 3c, right panel).    www.nature.com/scientificreports/ Quantification of Annexin-V positive, HSA + and HSA − MDMs revealed that AEG40730 killed a significantly higher number of HIV-HSA-infected macrophages compared to the control DMSO-treated HSA-infected macrophages (p-value = 0.0006, DMS0 vs AEG) (Fig. 3d, left panel). In contrast, there was no significant difference in cell death between DMSO-treated and AEG40730-treated mock uninfected and HIV-HSA-negative uninfected macrophages (Fig. 3d, left panel). Representative histograms showing killing of HSA + and HSA − macrophages is shown (Fig. 3d, right panel).

SMs kill significantly higher numbers of HIV-infected macrophages than uninfected bystander cells.
The HIV-exposed but uninfected bystander cells play a critical role in HIV pathogenesis and disease progression by causing selective depletion of CD4 + T cells, leading to immunodeficiency 43 . Although HIV infection of T cells and macrophages differ with respect to cell death in vivo and in vitro 1,14,17,18 , uninfected bystander macrophages may also play a crucial role in HIV pathogenesis. Therefore, we determined if SMs induce apoptosis of HIV-exposed uninfected bystander macrophages. AEG40730 and LCL161 killed significantly high numbers of HIV-HSA-expressing (HIV-infected) cells compared to either the mock or HSA-negative (HIVuninfected/bystander, HIV-exposed) cells (Fig. 3e,  . While killing of HSA-negative (HIV-uninfected/bystander) cells was somewhat higher than levels seen in mock-infected cells, these differences were not significant suggesting that SMs specifically kill HIV-infected macrophages (Fig. 3e, f).

Knocking down IAP genes results in specific killing of HIV-infected MDM.
To confirm the involvement of IAPs, we employed cIAP1/2 siRNAs as previously [35][36][37] . Flow cytometry of HIV-HSA-infected macrophages from healthy donors revealed two distinct HSA-positive (HSA +) and HSA-negative (HSA −) populations ( Fig. 4a, right panel). Following transfection with siRNAs for 72 h, HIV-HSA-infected and HIV-HSA negative uninfected macrophages (HIV exposed bystander macrophages) were quantified for infection rate (% HSA + cells) and apoptosis using Annexin-V labelled with BV711 antibodies as above. The gating strategy applied was same as shown in Fig. 3b. Analysis of infection rate (% HIV-HSA FITC + cells) following transfection with control non-targeting siRNA or cIAP1/2 siRNAs did not show significant differences in mock and HIV-HSA-infected macrophages (Fig. 4a). Furthermore, control or cIAP1/2 siRNAs did not show any increase in non-specific FITC-HSA autoflourescence in mock uninfected macrophages (Fig. 4a). Representative histograms are shown (Fig. 4a, right panel). Quantification of Annexin-V positive HSA + and HSA − MDMs revealed that knocking down cIAP1/2 by siRNAs killed significantly higher numbers of HIV-HSA+ macrophages compared to the HIV-HSA-infected cells treated with non-targeting siRNAs (Fig. 4b) (p-value = 0.0026, HIV-HSA+ , ctrl siRNA vs si IAPs). In contrast, there was no significant difference in cell death between control and IAP siRNA transfected mock uninfected and HIV-HSA-negative uninfected macrophages (Fig. 4b). Representative histograms are shown in Fig. 4b, right panel. The functional activity of the same cIAP1/2 siRNA has been shown by us previously [35][36][37] .

SM do not affect HIV replication in HIV-infected macrophages.
Apoptosis has been shown to induce viral activation and replication in latently infected U1 and ACH2 cell lines 44 (Fig. 5g, h) at any concentration, suggesting that LCL161 and AEG40730 treatment did not affect HIV replication following HIV infection. We did not observe a drop in the percentage of infected cells and p24 levels following exposure to SMs as Annexin-V detects early stage of apoptosis and the cells are still alive and intact and secreting p24. Annexin positive cells we observed will eventually die at a later stage. Consistent with this, when we analyzed cell death after 7 days of exposure to SMs, we did observe a significant drop in the percentage of infected cells following LCL161 (p-value = 0.0173, Media vs LCL 4 μM) and AEG40730 (p-value = 0.0021, Media vs AEG 5 μM) treatment of HIV-infected macrophages (Fig. 5i).

TNF-α mediates SM-induced apoptosis in U1 cells. SM-induced cell death of various tumor cells is
mediated by endogenously produced TNF-α (TNF) following SM treatment through the activation of the noncanonical NF-κB pathway 45,46 . To determine if SM-induced apoptosis in HIV-infected MDM is due to endogenous TNF production, LCL161-treated U937, U1 cells and in vitro HIV-infected MDM were analyzed for TNF secretion. LCL161 treatment resulted in low level although significant TNF production (p-value = 0.03, DMSO vs LCL 1-4 μM) in undifferentiated and differentiated U937 and U1 cells compared to DMSO-treated control cells ( Fig. 6a-d). In contrast, LCL161 treatment did not induce significant TNF production in both in vitro mock-and HIV-infected MDM (Fig. 6e). Similar to in vitro HIV-infected macrophages from healthy donors, ex vivo derived MDMs from HIV-infected patients did not produce significantly higher levels of TNF following LCL161 treatment compared to the untreated negative controls (Fig. 6f).
To evaluate the impact of TNF in SM-induced apoptosis of MDM, LCL161-treated MDM were stimulated with recombinant TNF (rTNF) followed by analysis of cell death using PI staining. Treatment of MDMs with www.nature.com/scientificreports/ LCL161 alone or with rTNF did not result in significant cell death (Fig. 6g). In contrast, rTNF either alone or in combination with LCL161 induced significant cell death in U937 and U1 cells (Fig. 6h, i), similar to that observed in various tumor cells 47,48 . These results suggest that SM-mediated killing of HIV-infected primary macrophages may be independent of endogenously produced TNF. However, the involvement of TNF in SM-induced cell death can be confirmed only following silencing of TNFR1 & TNFR2 in HIV-infected macrophages which may not be technically feasible at present.

HIV-infected MDM do not develop M1 phenotype before or after SMAC mimetics treatment.
We have shown that SMs selectively kill IFN-γ-generated M1 macrophages 49      www.nature.com/scientificreports/ infected MDM both before or after SM treatment did not express M1 phenotype and suggest that SM-mediated apoptosis of HIV-infected MDM is independent of M1 polarization.

RIPK1 levels are reduced in HIV-infected MDMs. SM-induced apoptosis of HIV-infected macrophages
may be ascribed to the impaired expression of IAP-associated signaling kinases such as RIPK1 27,50 . RIPK1 plays a key role in NF-κB signaling and apoptosis 51 . Moreover, RIPK1 is a target substrate for HIV protease, a viral protein synthesized late in the viral life cycle that inactivate RIPK1 in HIV-infected primary CD4 + T cells 52 . Examining changes in RIPK1 over the course of MDM infection in vitro, we found an upregulation of full length RIPK1 at day 2 post-infection until day 8 post-infection (Fig. 7a) that was associated with a gradual and significant accumulation of cleaved RIPK1 on days 2, 4, 6 and 8 post-infections (Fig. 7a). The full length RIPK1 expression gradually declined following day 2 post-infection until day 8 post-infection (Fig. 7a). This suggested that in vitro infection of MDMs with HIV-1 CS204 resulted in the initial upregulation and subsequent downregulation of full length RIPK1 and accumulation of cleaved RIPK1 over the period of HIV infection.
To determine whether RIPK1 is similarly cleaved and inactivated in HIV-infected MDMs treated with SM, in vitro mock and HIV CS204-infected MDMs (7 days post-infection) were treated with SM-LCL161 for 2 days followed by immunoprobing for RIPK1. HIV infection resulted in the downregulation of full length RIPK1 in the absence of LCL161 compared to the mock infected controls (densitometry ratio: compare DMSO lane 1 in both mock and HIV; Fig. 7b). SM-LCL161 treatment in both mock and HIV-infected MDM caused degradation of full length RIPK1 at all concentrations of LCL161 in both mock and HIV-infected macrophages (Fig. 7b). However, RIPK1 expression in HIV-infected LCL161-treated macrophages was lower compared to the mockinfected, LCL161-treated macrophages (Fig. 7b).
To confirm if the downregulation of RIPK1 occurs specifically in HIV-infected MDM, MDMs were infected with HIV-HSA for 9 days. HIV-infected HSA-expressing MDMs were harvested by magnetic column separation based on HSA expression 42 followed by western blotting for RIPK1 analysis. 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. The waste fraction represents cells that are eluted during the column wash prior to the collection of the HSA-selected MDM and contains both HSA+ and HSA negative macrophages. The positive fraction represents the HIV-infected HSA-expressing cells retained in the magnetic column that are eluted at the end of the HSA selection protocol. The gating strategy for detection of HIV-HSA infected macrophages is shown (Fig. 7c upper panel). The positively selected MDM infected with HIV-HSA had ~ 70% purity while the negatively selected HIV-uninfected cells and waste fractions contained ~ 10 and ~ 5% HSA-expressing macrophages, respectively (Fig. 7c, lower panel). The blots show that full length RIPK1 is significantly downregulated in the positively selected HIV-HSA enriched fraction compared to the mock infected macrophages (densitometric ratio: 1.0 vs 0.28; Fig. 7d). Overall, the results suggest that full length RIPK1 is degraded with a corresponding increase in cleaved RIPK1 in HIV-infected macrophages. cIAP1/2 and RIPK1 are essential for survival of HIV-infected MDM. The above results show that inactivation of RIPK1 in settings where cIAPs are absent, may affect the viability of macrophages. To evaluate the combined impact of cIAPs and RIPK1 inactivation on the survival of macrophages, MDM from healthy donors were pretreated with necrostatin-1 (nec-1), a specific RIPK1 inhibitor, for 2 h followed by treatment with LCL161 and analysis for cell death. Treatment with LCL161 or necrostatin-1 alone did not induce significant cell death in MDM. However, combined treatment of LCL161 and necrostatin-1 caused a significant increase in cell death of MDM (p-value = 0.03, LCL 2 μM vs LCL 2 μM + nec-1) (Fig. 8a). Figure 8b shows representative histograms of the intracellular PI staining. Furthermore, treatment with necrostatin-1 alone did not show cleavage of PARP or caspases-8 and 9 although treatment with LCL161 alone did show their minimal cleavage (Fig. 8c). However, treatment with both necrostatin-1 and LCL161 significantly enhanced cleavage of the three caspases as well as PARP (Fig. 8c). These results show that degradation of cIAP1/2 by LCL161 and inactivation of RIPK1 by necrostatin-1 results in death of normal macrophages and suggest that cIAP1/2 and RIPK1 play an important role in regulating viability of human macrophages. Since HIV infection down regulates RIPK1, and LCL161 causes IAPs degradation and death of HIV-infected macrophages, our results suggest that RIPK1 and IAPs may play key roles in SM-induced cell death of HIV-infected macrophages.

Discussion
In this study, we show that although cIAP1/2 are dispensable host factors for the viability of macrophages, these anti-apoptotic proteins play a critical role in the survival of HIV-infected macrophages. SMs induced apoptosis of chronically HIV-infected U1 cells, in vitro HIV-infected macrophages with a clinical strain and ex-vivo derived macrophages from naïve and ART-treated HIV patients. Furthermore, SMs were shown to specifically kill HIVinfected MDMs by employing a R5-tropic laboratory strain, HIV-HSA. SM-induced killing of HIV-infected macrophages is mediated by apoptosis, but is independent of the establishment of M1 polarization. In addition, SM-induced apoptosis of HIV-infected macrophages may be due to RIPK1 degradation which in concert with IAP1/2 degradation may result in apoptosis of HIV-infected macrophages.
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 macrophages. Recently, cIAP1/2 and survivin, another member of the IAP family were suggested to be involved in survival of HIV-infected CD4 + T cells 30,31 and macrophages 23 . In addition, XIAP down regulation by flavopiridol, a cyclin-dependent kinase 9 (CDK-9) inhibitor, caused increased apoptosis of a chronically HIV-infected ACH2 T cells 53 . We and others have previously shown that ablation of cIAP1/2 by SMs does not affect survival of primary human macrophages 36,54 . However, resistance of macrophages to apoptogenic HIV-Vpr was attributed to cIAP1/2 35 23 . The role of autophagy in LCL161-induced cell death in our study needs to be investigated. Our results on the effect of SM on macrophages derived from the HIV-infected patients show killing of 20-30% macrophages. However, HIV-infected myeloid cells are rarely detected in the blood of HIV-infected individuals 41 ; hence the frequency of HIV-infected macrophages in ex vivo derived macrophages from HIVinfected patients may be extremely low. Therefore, it is likely that SMs may be killing bystander cells (Fig. 2f, g) unlike the in vitro HIV-HSA-infected macrophages. It is extremely difficult to determine selective killing of very low numbers of HIV-infected macrophages in the patients and even in in vitro HIV-infected macrophages. The high variability in SM-mediated killing of ex vivo generated MDMs from monocytes of ART-treated and naïve untreated groups may also be due to the myeloid population in the patients which is exposed to a variety of inflammatory/anti-inflammatory cytokines and exposure to a variety of opportunistic infections 56 . Exposure to ARV drugs may be somewhat toxic to these cells and affect their ability to respond to stress by altering their mitochondrial function 57,58 .
The number of HIV-infected macrophages in in vitro infected macrophages with the clinical strain is around 5-10% and varies up to 30% with HIV-HSA due partly to a milieu of HIV restriction factors that limit the virus life cycle 12,42 . Although SMs killed high number of HIV-HSA-negative macrophages compared to mock-infected macrophages, the differences were not significant. Moreover, SMs killed significantly higher number of HIV-HSA-infected macrophages compared to the HIV-HSA-negative macrophages further suggesting a specificity of SMs towards killing of HIV-infected macrophages. We did not observe a drop in the percentage of infected cells following exposure to SMs as Annexin-V detects early stage of apoptosis. These Annexin positive cells we observed should eventually die at a later stage. Consistent with this, we did observe a significant drop in the percentage of infected cells after 7 days of exposure to SMs.
IAPs have been shown as a potent negative regulator of LTR-dependent HIV-1 transcription causing reversal of HIV latency in JLat latency model system and in latently infected memory T cells 32 . SMs activate the noncanonical NF-κB pathway by virtue of RelA:p50 and RelB:p52 transcription factors, which bind to the HIV-1 LTR region leading to the induction of virus transcription in latently infected JLat cells 32,59 . However, SM did not affect HIV transcription in U1 cells and in vitro HIV-infected macrophages. Our results are also in agreement with Campbell et al. on the effects of SM on HIV transcription in HIV-infected macrophages 23 . We have previously shown that SMs alone did not activate either classical or alternative NF-κB pathways in macrophages 40 , which may explain SM's inability to impact virus replication in macrophages.
SM-mediated killing has been attributed to endogenous TNF in cancer cells 45,46 ; however, it has been reported to be independent of TNF in some cancer cells 60 and HIV-infected T cells 32 . While TNF-mediated SM-induced killing in U1 undifferentiated and PMA-differentiated cells, it was not observed in primary MDMs. Although in vitro HIV-infected MDMs produced significant levels of TNF, SM treatment did not affect its secretion in either uninfected or HIV-infected macrophages. Moreover, rTNF did not induce cell death in SM-treated macrophages suggesting that SM-induced cell death in macrophages is independent of TNF. We have also shown that SM-induced apoptosis in M1 macrophages was not mediated by TNF although M1 macrophages produced high levels of TNF 61 . However, the involvement of TNF in SM-induced cell death cannot be conclusively ruled out without silencing of TNF receptors in HIV-infected macrophages.
HIV infection results in dysregulation of cytokine profile in-vivo and in-vitro 62 and can possibly affect the polarization state of macrophages. Since IFN-γ-generated M1 macrophages are susceptible to SM-mediated cell death 49 , HIV-1 infection may polarize macrophages into M1 phenotype and make them susceptible to SM-induced apoptosis. Since in vitro-infected and ex vivo-derived macrophages exposed to SM were not www.nature.com/scientificreports/ polarized into M1 phenotype suggested that SM-mediated killing of HIV-infected macrophages was not due to M1 polarization. The pathways of apoptosis are regulated by RIPK1 51,63 . In TNF-mediated signaling, RIPK1 is recruited in a multiprotein complex I along with TRADD, TRAF2, and cIAP1/2 to promote transcription of genes with antiapoptotic properties such as cIAP1/2 63 . 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 63 . Recently, HIV infection of primary activated CD4 + T cells was shown to downregulate RIPK1 through the HIV-1 protease 52 . RIPK1 modification in response to human rhinovirus and Newcastle disease virus infection has also been reported 64,65 . Herein, we show that infection of macrophages with HIV-1 CS204 or with HIV-HSA caused downregulation and cleavage of RIPK1. Given that down regulation of IAPs alone by LCL161 or of inactivation of RIPK1 alone by necrostatin-1 did not induce cell death in uninfected macrophages, suggests that RIPK1 and cIAP1/2 are dispensable in survival of macrophages. However, inactivation of RIPK1 by necrostatin-1 following IAP degradation by SM resulted in a dramatic increase in cell death, cleavage of caspases and PARP in normal macrophages suggesting that RIPK1 may play a key role in SM-induced killing of HIV-infected macrophages. The role of RIPK1 degradation during HIV-1 infection of macrophages needs further investigation.
In summary, the results of this study suggest a potential significance of SM in killing of HIV-infected macrophages in vivo. In support of these observations, recent reports using the combination of an LRA and SMAC mimetic 66 , and nanoengineered CD4 + T cell membrane-coated nanoparticles (TNP) encapsulating the SMAC mimetics LCL161 have been shown to be potential therapeutic agents to kill specifically HIV-1-infected cells as part of an HIV-1 cure strategy 67 . In the event SM are able to kill HIV-infected macrophages in vivo, they have the potential to be translated into clinical interventions aimed at eradicating HIV infection by directly targeting HIV-infected macrophages.

Generation of human MDMs, 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. MDMs were generated from monocytes via adherence methods as previously described 40 . Briefly, 2.0 × 10 6 PBMCs/well (12-well plate) were allowed to adhere for 3 h and non-adherent cells were washed off. Adherent monocytes were cultured for 7 days in complete DMEM (Wisent Inc., St. Bruno, Quebec) supplemented with 10% fetal bovine serum (FBS; GE Healthcare), penicillin and streptomycin and 10 ng/ml MCSF (R&D Systems, Minneapolis, MN, USA). MCSF-containing media was replaced every 2 days until the 7th day at which point the monocytes differentiated into macrophages. Purity of macrophages as assessed by measuring CD14 expression by flow cytometry was 100%. Using a similar protocol, MDMs were generated from PBMCs obtained from HIV-infected individuals receiving anti-retroviral treatment (typically one integrase strand transfer inhibitor plus two nucleosides/nucleotide analogues) and from HIVinfected individuals who never received any treatment (ART naïve).

HIV-1 production, in vitro infection of MDMs with HIV and HIV p24 ELISA. MDM were infected
with dual tropic HIV-1 CS204 viral stock supernatants containing 100 ng of HIV-p24 protein, supplemented with 8 µg/ml of polybrene (Sigma-Aldrich) for 2 h as described previously 68 . The dual tropic HIV-1 CS204 stocks were produced in CD8 + depleted PBMCs from healthy donors as previously described 69  MDMs were infected with 100 ng p24 per 1.0 × 10 6 cells supplemented with 5 μg/ml polybrene (Sigma Aldrich SKU: TR-1003-G) for 16 h followed by infection with HIV-1 CS204 or HIV-HSA for 7 days. The supernatants were assessed for p24 using HIV p24 CA ELISA capture kit as above. www.nature.com/scientificreports/ Treatment of MDM with SM and assessment of apoptosis by flow cytometry. MDMs were cultured in complete media without antibiotics for 2 h before treatment with various concentrations of AEG40730 or LCL161. MDMs were evaluated for cell death by using intracellular PI staining as described 35,40 . Briefly, cells were washed with DPBS and fixed with methanol for 15 min at 4 °C. Subsequently, cells were treated with 25 μl of 10 μg/ml RNase A, followed by staining with 25 μl of 1 mg/ml PI solution (Sigma-Aldrich) at 4 °C for 1 h. The DNA content was analyzed using a Fortessa X-20 cytometer (BD Biosciences, Franklin Lakes, NJ, USA) and the FlowJo v10 software. The subdiploid DNA peak (< 2 N DNA), immediately adjacent to the G 0 /G 1 peak (2 N DNA), represents apoptotic cells and was quantified by histogram analyses. www.nature.com/scientificreports/ IL-28A were detected using antibody-immobilized magnetic beads and were quantified by MAGPIX® multiplex with xPONENT® software (Luminex Corp.).

Statistical analysis.
Data was plotted using Graphpad Prism 10. The statistical significance was calculated using student paired t test or Mann-Whitney U test. Plotted data represent the mean ± SD.
Ethics statement. HIV-infected individuals and healthy participants involved in the study gave informed written consent and the protocol for obtaining blood samples was approved by the Review Ethics Board of the Ottawa General Hospital and the Children's Hospital of Eastern Ontario, Ottawa, ON, Canada. All methods were performed in accordance with relevant guidelines and regulations.