TREM-1 Protects HIV-1-Infected Macrophages from Apoptosis through Maintenance of Mitochondrial Function

The major challenge to human immunodeficiency virus (HIV) treatment is the development of strategies that lead to viral eradication. A roadblock to accomplishing this goal is the lack of an approach that would safely eliminate HIV from all resting/latent reservoirs, including macrophages. Macrophages are a key part of the innate immune system and are responsible for recognizing invading microbes and sending appropriate signals to other immune cells. Here, we found that HIV induces the upregulation of the protein TREM1 (triggering receptor expressed on myeloid cells 1), which signals an increase in the expression of antiapoptotic proteins, thus promoting survival of HIV-infected macrophages.

the absence of CD4 ϩ T cells (4,5), and contain lower intracellular concentrations of antiretrovirals than CD4 ϩ T cells, resulting in ongoing HIV replication (6). Latently infected macrophages also harbor and transmit replication-competent virus to nearby CD4 ϩ T cells through virologic synapses that are resistant to antibody-mediated neutralization (7). Additionally, they survive for months to years and are resistant both to the cytopathic effects of HIV infection and to CD8 ϩ T cell-mediated killing (8)(9)(10).
Notwithstanding recent advances defining cell death signaling pathways, much is unknown about the mechanism(s) by which macrophages escape apoptosis during HIV infection. Despite this, studies suggest that HIV signals through multiple pathways to reprogram the transcriptome and proteome of human macrophages to render them resistant to HIV-mediated apoptosis. HIV Nef induces the phosphorylation and inactivation of proapoptotic BCL2-associated agonist of cell death (BAD) (11), and HIV Tat and gp120 induce the transcription of apoptosis regulatory proteins, including colonystimulating factor 1 (CSF1), caspase 8 (CASP8) and Fas associated via death domain (FADD)-like apoptosis regulator (CFLAR), BCL2 family members BCL2 and BCLXL, and inhibitor of apoptosis proteins (IAP) X-linked inhibitor of apoptosis (XIAP), baculoviral IAP repeat-containing 2 (BIRC2), and BIRC3 (12)(13)(14). Additionally, acute exposure to HIV and to HIV gp41 peptides upregulates triggering receptor expressed on myeloid cells 1 (TREM1) in peripheral blood mononuclear cells (PBMC) and macrophages (15)(16)(17).
TREM1 is a 30-kDa glycoprotein expressed on macrophages that belongs to the immunoglobulin variable (IgV) domain superfamily of proteins and consists of a positive-charged transmembrane domain and a short cytoplasmic domain that lacks any signaling motif. Thus, TREM1 requires TYRO protein tyrosine kinase binding protein (TYROBP) for signaling (18,19). Although the precise ligands for TREM1 are unknown, CD177, peptidoglycan recognition protein 1 (PGLYRP1), high-mobility group box 1 protein (HMGB1), and some damage-associated molecular patterns (DAMPs) and pathogen-associated molecular patterns (PAMPs) activate TREM1 signaling (20)(21)(22)(23)(24). TREM1 ligation leads to an increase in reactive oxygen species and the secretion of proinflammatory cytokines through the activation of NF-B. Additionally, TREM1 ligation synergistically amplifies Toll-like receptor (TLR)-mediated and Nod-like receptor (NLR)-mediated proinflammatory responses (19). TREM1 is also an antiapoptotic molecule that prolongs macrophage survival. Specific ligation or overexpression of TREM1 induces BCL2 expression and the subsequent depletion of caspase 3, preventing the cleavage of PARP1 (25). In addition, both ligation and overexpression of TREM1 lead to an increase in mitofusin expression levels, suggesting that TREM1 contributes to the maintenance of mitochondrial integrity, thus favoring cell survival (25). The involvement of and mechanisms by which TREM1 may alter macrophage survival during HIV infection are unknown. Therefore, we investigated whether TREM1 affects cell survival during HIV infection of human primary macrophages.

RESULTS
HIV infection of macrophages results in persistent infection and prolongs cell survival. We designed our initial experiments to determine whether HIV infection results in increased apoptosis of infected macrophages. We infected macrophages with HIV for 28 days, during which time we assessed HIV replication kinetics and apoptosis. HIV p24 antigen levels increased in culture supernatants between day 3 and day 18 followed by a decline after 18 days (Fig. 1A). Uninfected macrophages showed low levels of apoptosis, with an observed maximum of 13.5% of cells displaying formamidesensitive single-stranded DNA (ssDNA; a specific indicator of apoptosis [26]) by day 28 (Fig. 1B). HIV infection of macrophages was associated with a decrease in cell death, reaching a high of just 5.5% at day 28 (Ϯ0.3% standard error of the mean [SEM]; P ϭ 0.017) (Fig. 1B). Culture supernatants assessed for lactate dehydrogenase (LDH) release demonstrated the same profile (Fig. 1C). Collectively, these results suggest that HIV infection prolongs survival of macrophages.
HIV influences pro-and antiapoptotic protein expression in macrophages. To determine why there is an absence of apoptosis in HIV-infected macrophages, we (P Ͻ 0.01; Fig. 2B). Next, we determined if the HIV antigens gp120, Tat, and RNA40 (a 20-mer synthetic single-stranded RNA oligoribonucleotide derived from the HIV longterminal repeat and TLR8 ligand) influenced the expression of these pro-and antiapoptotic proteins in macrophages. Both gp120 and Tat increased the expression levels of BCL2 and BCLXL in a dose-dependent manner while simultaneously decreasing those of both BAD and BAX (see Fig. S1 in the supplemental material). In contrast, while RNA40 also dose dependently increased the expression of both BCL2 and BCLXL, it also increased the expression of BAD while having no effect on BAX expression levels (Fig. S1).
A key event in apoptosis is the translocation of BCL2L11 from the cytoskeleton to the mitochondrial membrane (27). In healthy cells, BCL2L11 is inactivated by its interaction with dynein light chain LC8-type 1 (DYNLL1), which recruits it to the microtubule-based dynein motor complex (28). During HIV infection of CD4 ϩ T cells, HIV Tat binds tubulin through a four-amino-acid subdomain of its conserved core region, altering microtubule dynamics (29,30) and leading to the mitogen-activated protein kinase 8 (MAPK8)-mediated phosphorylation of BCL2L11 and dissociation from DYNLL1 and the dynein motor complex. Phosphorylated BCL2L11 then translocates to the mitochondrial membrane, where it directly activates the BAX/BAK-dependent mitochondrial pathway of apoptosis and cell death (27,31,32). Therefore, we examined whether HIV infection also induces the translocation of BCL2L11 to mitochondria in macrophages despite the absence of HIV-mediated cytopathogenesis. Using cell fractionation, we determined that BCL2L11 translocates to mitochondria, suggesting that HIV is inhibiting the mitochondrial pathway of apoptosis (Fig. 2C).
HIV infection increases the expression of TREM1 in macrophages. As HIV infection resulted in increased expression of BCL2 and as overexpression of TREM1 can induce the expression of BCL2 (25), we examined whether macrophages productively infected with HIV express higher levels of TREM1. At 10 days after infection, HIV-infected macrophages displayed increased levels of TREM1 (P ϭ 0.02; Fig. 3A). We then investigated the mechanism by which HIV induces TREM1 expression. As NF-B transcriptionally regulates TREM1 in macrophages in response to LPS (33), we utilized RNA interference (RNAi) for the NF-B subunit RELA to investigate whether NF-B regulates TREM1 expression in response to HIV infection. RELA silencing resulted in the almost complete ablation of TREM1 expression in response to HIV (Fig. 3B), suggesting that NF-B regulates TREM1 expression in HIV-infected macrophages. To confirm the sequence-specific interaction of RELA with the TREM1 promoter, we analyzed the promoter region of TREM1 using Genomatix (version 3.11; Intrexon) and found consensus binding sites for NF-B. We performed an electrophoretic mobility shift assay using a biotinylated probe that encompassed the predicted binding site. This probe showed evidence of strong binding to nuclear proteins and was supershifted by antibodies against RELA and NFKB1 but not by antibodies against REL, NFKB2, RELB, or BCL3 (Fig. 3C). To investigate whether HIV triggers the classical NF-B signaling pathway, we analyzed the ability of RELA to physically interact with NFKB1 p50 and to translocate to the nucleus using coimmunoprecipitation. NFKB1 p50/RELA complexes were detected in both the cytoplasm and the nucleus, whereas no NFKB1 p50/RELB complexes were detected (Fig. 3D). Interestingly, although uninfected cells transfected with siRELA showed apoptosis levels similar to the levels seen with those transfected with control RNAi, RELA-silenced HIV-infected macrophages showed an increased number of apoptotic cells ( Fig. 3E; P Ͻ 0.001).
We next investigated whether HIV gp120, Tat, and RNA40 induce TREM1 expression in human macrophages. All three antigens induced a dose-dependent increase in TREM1 expression; HIV Tat and gp120 induced 18-fold (Ϯ3.9 SEM; P ϭ 0.0005) and 7.5-fold (Ϯ2.9 SEM; P ϭ 0.0075) increases, respectively, at 10 ng ml Ϫ1 , and HIV RNA40 induced a 42-fold increase (Ϯ4.6 SEM; P Ͻ 0.0001) at 5 g ml Ϫ1 (Fig. 4A). Similarly to HIV-infected macrophages, RELA silencing of uninfected macrophages ( Fig. 4B and C) significantly increased cell death in response to the presence of HIV Tat, gp120, or RNA40 ( Fig. 4D), while it also attenuated TREM1 expression (Fig. 4E). Electrophoretic mobility shift assay (EMSA) demonstrated that HIV Tat, gp120, and RNA40 rapidly induced NF-B DNA binding, with supershift analysis demonstrating that the NF-B complex consisted primarily of NFKB1 and RELA subunits (Fig. S2A). We then investigated whether these antigens trigger the classical NF-B signaling pathway. HIV Tat, gp120, and RNA40 all triggered the rapid processing of NFKB1 p105 to p50, the dimerization of NFKB1 p50 with RELA, and the subsequent nuclear translocation ( Fig. S2B and C).
TREM1 is required for HIV-infected macrophage survival. To address the role of TREM1 in macrophage survival during HIV infection, we used TREM1 small interfering RNA (siRNA). We reasoned that if TREM1 is involved in promoting cell survival, then silencing TREM1 should increase cell death during HIV infection. Macrophages infected with HIV for 10 days were transfected with control scrambled siRNA (siNS) or TREM1 siRNA (siTREM1) for 48 h and then either assessed for apoptotic ssDNA damage or stained with annexin A5 (ANXA5) and propidium iodide (PI) (Fig. 5A). Uninfected cells transfected with siTREM1 showed apoptosis levels similar to those transfected with siNS. In contrast, TREM1 siRNA-transfected HIV-infected macrophages showed an increased number of apoptotic cells (83.9% Ϯ 6%) compared with HIV-infected macrophages transfected with siNS (4.2 Ϯ 1.4%; P ϭ 0.02). Because TREM1 silencing induced Next, we analyzed the effect of HIV antigen-induced apoptosis on TREM1-silenced cells (Fig. 6A) after exposure of the cells to HIV antigens for 24 h. Importantly, exposure of siNS-transfected cells to HIV Tat, gp120, or RNA40 did not result in increased apoptosis (P Ͼ 0.4; Fig. 6B). In contrast, and similarly to HIV infection, TREM1-silenced macrophages treated with HIV antigens displayed an enhanced number of apoptotic cells compared with vehicle-treated cells (Fig. 6B). TREM1-silenced cells exposed to Tat exhibited the highest proportion of apoptotic cells (mean, 92.4 Ϯ 0.4%; P Ͻ 0.0001) closely followed by RNA40 (mean, 84.7% Ϯ 1.3%; P Ͻ 0.0001) and gp120 (mean, 73.5% Ϯ 2.2%; P Ͻ 0.0001). Importantly, staurosporine exposure similarly induced apoptosis in both siNS-transfected and siTREM1-transfected cells (mean, 94.5% Ϯ 1.4% versus 94.0% Ϯ 0.4%; P ϭ 0.4) (Fig. 6B). We then evaluated the effect of TREM1 silencing on HIV antigen-induced expression of antiapoptotic and proapoptotic proteins. TREM1 silencing decreased BCLXL and BCL2 expression in uninfected macrophages by more than 86% (P Ͻ 0.001) while having no noticeable effect on the expression of either BAX or BAD (Fig. 6C). Of note, neither HIV Tat nor gp120 nor RNA40 had a significant effect on the expression of BCLXL, BCL2, BAX, or BAD compared to the vehicle in these TREM1-silenced cells. Collectively, these data suggest that induction of TREM1 by HIV inhibits apoptosis and prolongs survival of HIV-infected macrophages. TREM1 Protects Macrophages from HIV-Mediated Apoptosis ® Next, we sought to define the expression of BCL2 in relation to specific ligation of TREM1 by the use of monoclonal antibodies (mTREM1). BCL2 expression was significantly increased in macrophages in response to mTREM1 compared with the isotype control antibody (P ϭ 0.0006; Fig. 7A). We then sought to define the relationship between BCL2 expression and cell survival during HIV infection of macrophages using RNAi for BCL2. BCL2 silencing of HIV-infected macrophages led to an increased number of apoptotic cells (86.3% Ϯ 5.4%; P ϭ 0.0001) (Fig. 7B). Similarly, BCL2 silencing also significantly increased uninfected macrophage cell death in response to exposure to HIV Tat, gp120, or RNA40 (P Ͻ 0.001; Fig. 7C).
TREM1 silencing leads to the induction of apoptosis through the intrinsic pathway. The interactions between anti-and proapoptotic proteins exist in a delicate balance at the mitochondrial membrane that determines cell fate, with mitochondrial fragmentation a central mechanism of apoptosis. Loss of balance between pro-and antiapoptotic BCL2 proteins leads to outer membrane permeabilization of mitochondria, calcium influx, and mitochondrial fragmentation, leading to the subsequent activation of effector caspases. In contrast, a fused mitochondrial phenotype protects cells from programmed cell death. Therefore, we examined the effect of TREM1 silencing on the mitochondrial membrane potential (Δm) using DiIC1(5) (1,1=,3,3,3=,3=hexamethylindodicarbo-cyanine iodide). Following TREM1 silencing, there was no difference in Δm levels between siNS-transfected HIV-infected and uninfected macro- phages, indicating that mitochondria are still functional during infection. Similarly, we observed no difference in Δm levels between uninfected cells transfected with siTREM1 and those transfected with siNS. In contrast, we observed a marked decrease in DiIC1(5) staining intensity in TREM1-silenced HIV-infected macrophages (Fig. 8A). These data suggest that TREM1 silencing results in the depolarization of the Δm in HIV-infected macrophages but not in uninfected macrophages.
A key event in apoptosis is the formation of the transition pore in the mitochondrial membrane and the release of mitochondrial factors such as cytochrome c (CYCS) into To determine whether TREM1 silencing in HIV-infected macrophages leads to the mitochondrial release of CYCS into the cytoplasm, we performed cell fractionation. Our data indicate that mitochondria retain CYCS in siNS-transfected HIV-infected and uninfected cells (Fig. 8B). In contrast, TREM1 silencing induces the release of CYCS into the cytoplasm in HIV-infected macrophages but not in uninfected macrophages (P ϭ 0.018; Fig. 8B). In the presence of ATP/dATP, CYCS released into the cytosol binds to and triggers the oligomerization of cytosolic apoptotic protease activating factor 1 (APAF1). The resultant complex recruits and activates multiple copies of the apical caspase in the mitochondrial pathway of apoptosis, caspase 9 (CASP9). TREM1 silencing resulted in the cleavage and activation of CASP9 in HIV-infected macrophages but had no effect on CASP9 activation in uninfected macrophages (Fig. 8C).
We next assessed the effect of HIV Tat, gp120, and RNA40 on Δm, CYCS release, and subsequent CASP9 activation in uninfected macrophages. There was no significant difference in Δm after Tat, gp120, and RNA40 exposure in siNS-transfected macrophages, indicating that Tat, gp120, and RNA40 have little effect on Δm in the presence of TREM1 (Fig. S3A). In contrast, we observed a marked decrease in DiIC1(5) staining intensity in TREM1-silenced macrophages after Tat, gp120, and RNA40 treatment, suggesting that TREM1 is essential for maintaining Δm after exposure to HIV antigens. Similarly, treatment of siNS-transfected macrophages with HIV Tat, gp120, or RNA40 did not induce the release of CYCS into the cytosol or induce CASP9 cleavage (Fig. S3B). In contrast, HIV Tat, gp120, and RNA40 exposure of TREM1-silenced cells resulted in CYCS release from the mitochondria and subsequent CASP9 cleavage. Together, these results indicate that TREM1 is essential for macrophage survival of acute HIV infection. TREM1 enhances mitochondrial integrity by inducing mitofusin expression. At the molecular level, mitochondrial fusion is a two-step process requiring the coordinated fusion of outer and inner mitochondrial membranes by separable sequential events. In humans, this process depends on distinct mitochondrial sublocalization of three fusogenic proteins: mitofusins 1 and 2 (MFN1 and MFN2) (embedded within the outer mitochondrial membrane) and optic atrophy 1 (located at the inner mitochondrial membrane). In addition to the well-characterized role of the profission dynamin- 1-like protein in mediating mitochondrial fission in cells undergoing cell death, inhibition of mitochondrial fusion is also an established consequence of apoptosis induction and occurs around the time of BAX activation. During oxygen glucose deprivation, MAPK1/2 phosphorylates MFN1, inhibiting its profusion function, and triggers its association with and oligomerization of BAK, facilitating the subsequent release of CYCS from mitochondria through BCL2L11 (34). Similarly, in response to cellular stresses, MAPK8 phosphorylates MFN2, leading to its ubiquitin-dependent proteasomal degradation and subsequently to mitochondrial fragmentation and enhanced apoptotic cell death (35). Therefore, we assessed mitofusin expression in HIV-infected macrophages. We found that at day 10 postinfection, the expression levels of both MFN1 and MFN2 were increased (P Ͻ 0.007; Fig. 9A). We next assessed the effect of HIV Tat, gp120, and RNA40 on mitofusin expression. All three antigens induced dose-dependent increases in both MFN1 and MFN2 (Fig. S4A).
We next assessed the effect of TREM1 silencing on mitochondrial integrity in HIV-infected cells by determining mitofusin expression post-TREM1 RNAi treatment. TREM1 silencing resulted in a significant reduction in HIV-induced MFN1 and MFN2 to levels similar to those seen with uninfected cells (P Ͼ 0.3; Fig. 9B). When we exposed TREM1-silenced uninfected macrophages to HIV Tat, gp120, or RNA40, each still increased MFN1 expression, although the increase was appreciably less than in the siNS cells (Fig. S4B). In contrast, TREM1 silencing reduced MFN2 expression in all treatments, including the vehicle treatment. Notably, this decrease was significantly enhanced post-HIV Tat treatment (mean, 87% Ϯ 1.5% reduction; P Ͻ 0.0001), gp120 treatment (mean, 89% Ϯ 0.5% reduction; P ϭ 0.0001), and RNA40 treatment (mean, 85% Ϯ 1.5% reduction; P ϭ 0.0007) compared with the vehicle. These data suggest that TREM1 expression is required for the HIV-mediated upregulation of mitofusins.

DISCUSSION
Efforts to purge the latent HIV reservoir predominantly involve reactivation of viral production from latently infected CD4 ϩ T cells followed by clearance of these cells through a combination of viral and cell-mediated cytotoxicity while ART prevents subsequent rounds of infection. Although this strategy has shown promise in cell lines, it has been ineffective on primary resting CD4 ϩ T cells from patients on suppressive ART and in in vivo studies (36)(37)(38)(39)(40). This approach is even less effective in macrophages, as HIV-infected macrophages are particularly resistant to HIV-mediated apoptosis and CD8 ϩ T cell-mediated killing (8)(9)(10). Thus, we require an improved understanding of the precise mechanisms underlying resistance of macrophage and microglia to cytopathic effects of HIV in order to develop approaches capable of efficiently and effectively eliminating infected cells. Depleting HIV-infected macrophages may result in the decline of cytokine and chemokine production and hence suppress the excessive inflammation and aberrant immune cell recruitment that persists despite suppressive ART and which has been associated with poor prognosis and ongoing viral replication. The present study demonstrated that HIV upregulates the expression of TREM1 in Infected macrophages display an increase in the levels of expression and recruitment of the proapoptotic protein BCL2L11 into mitochondria. Proapoptotic BH3-only proteins such as BCL2L11 bind to antiapoptotic proteins, allowing the proapoptotic multidomain proteins BAX and BAK1 to oligomerize and form channels on the mitochondrial membrane and leading to CYCS release and apoptosis. It is possible that the observed upregulation and sequestration of BCL2L11 into the mitochondria of HIVinfected macrophages are early steps in the apoptotic response to a viral infection, a response that HIV subsequently blocks downstream as we do not observe the release of CYCS from mitochondria nor an increase in apoptosis once infection is established. However, when we silenced TREM1, the expression levels of BCL2, BCLXL, and mitofusins all decreased, while the expression levels of BAX and BAD remained unchanged. This coincided with the release of CYCS from mitochondria and an increase in apoptosis.
Although TREM1 activity was initially described only in the context of bacterial or fungal infections, a role for TREM1 during viral infections is emerging. Early studies focused on and demonstrated that TREM1 signaling modulates pattern recognition receptor virus-associated inflammation and that TLR signaling upregulates TREM1 expression. TLR3, TLR7, TLR8, and TLR9 are all common TLRs that respond to viral PAMPs. For example, TREM1-activated murine bone marrow-derived dendritic cells display increased tumor necrosis factor (TNF) expression after treatment using CpG DNA (a TLR9 ligand) (41). Similarly, TREM1-activated PBMC display enhanced production of TNF following stimulation with either CpG DNA or the TLR3 ligand poly(I•C) (42); poly(I•C) induces transcription of TREM1 in human primary monocytes (43). In neutrophils, TREM1 activation acts synergistically with TLR7 and TLR8 ligand-mediated activation with regard to effector mechanisms but does so antagonistically for survival (44). In the present study, we demonstrated that RNA40, a TLR8 ligand, upregulates TREM1 expression and that TREM1 is required for macrophage survival. With regard to viral infections, West Nile virus (WNV) increases TREM1 expression in murine macrophages and dendritic cells (45) and enhances TREM1 signaling in murine livers (46). Consistent with this, exogenous activation of TREM1 increases expression levels of the interferon alpha 2 gene (IFNA2), TNF, and the interleukin-6 gene (IL-6) in WNV-infected murine embryonic fibroblasts, while antagonizing TREM1 leads to a reduction in their expression, suggesting that TREM1 plays a role in modulating the inflammatory response to WNV (45). Similarly, Zika virus also upregulates TREM1 signaling (47). Furthermore, the Filoviridae Marburg virus and Ebola virus both upregulate the expression and activation of TREM1 on neutrophils following internalization (20). Although Ebola virus does not replicate in these cells, it is possible that Ebola virus glycoprotein, which interacts with neutrophils, acts as a ligand for TREM1 and contributes to the cytokine storm associated with the terminal stage of Ebola virus infection (20,48,49). While we showed that HIV infection of macrophages increases TREM1 expression and is essential for cell survival, the role of TREM1 in the inflammatory response to HIV infection remains to be elucidated.
Cells also shed soluble TREM1 (sTREM1) from cell membranes. Although the role of sTREM1 in the context of viral pathogenesis is unknown, sTREM1 levels are increased during infection with a number of viruses, including Crimean-Congo hemorrhagic fever orthonairovirus (50), hepatitis B virus (51), WNV (45), dengue virus (52), and hepatitis C virus (51). It is possible that such increases in the levels of sTREM1 could represent a virally induced compensatory mechanism to counteract inflammatory processes. Alternatively, such increases could represent a host-induced mechanism to control tissue damage by attenuating downstream inflammatory signals. Studies are required to elucidate this function and to identify the potential of sTREM1 as a marker of disease severity in acute infections by such viruses as influenza virus and dengue virus or in chronic infections such as HIV.
Evading apoptosis by upregulation of antiapoptotic or downregulation of proapo-ptotic proteins is an important step in altering cell survival. Importantly, the inhibition of macrophage apoptosis may contribute to viral latency and persistence during HIV infection (10,53). In summary, this study demonstrated an important role and a mechanism for TREM1 in prolonging survival of HIV-infected macrophages. Our report also highlights TREM1 as a therapeutic target to clear the macrophage HIV reservoir and suggests that current HIV reservoir paradigms include the targeting of TREM1 in any effort to achieve HIV eradication.

MATERIALS AND METHODS
Macrophages. Venous blood was drawn from human subjects using protocols approved by the Human Research Protections Program of the University of California, San Diego, in accordance with the requirements of the Code of Federal Regulations (CFR) on the Protection of Human Subjects (45 CFR 46 and 21 CFR 50 and 56). All volunteers gave written informed consent prior to their participation. PBMC were isolated from whole blood by density gradient centrifugation over Ficoll-Paque Plus (GE Healthcare). Macrophages were prepared by incubating 6 ϫ 10 6 PBMC ml Ϫ1 in macrophage media (RPMI 1640 [Gibco] supplemented with 10% [vol/vol] heat-inactivated fetal bovine serum [FBS; Sigma]; 2 mM L-glutamine, 0.1 mg ml Ϫ1 streptomycin, and 100 U ml Ϫ1 penicillin [all Gibco]; and 10 ng ml Ϫ1 CSF1 (Peprotech]), after which nonadherent cells were removed by aspiration and washed with Dulbecco's phosphate-buffered saline (Gibco). Adherent cells were further incubated in macrophage media for 10 days at 37°C and 5% CO 2 , with medium changes performed every 2 days, before use. This protocol results in a high proportion of CD14 ϩ sialic acid binding Ig-like lectin 1 (SIGLEC1) Lo macrophages that are permissive to HIV infection (54,55).
HIV. HIV Ba-L was obtained through the NIH AIDS Reagent Program from Suzanne Gartner, Mikulas Popovic, and Robert Gallo (56,57). Virus stocks were prepared as previously described (58). HIV infectivity levels were calculated as the 50% tissue culture infectious doses (TCID 50 ) as described previously (59) and multiplicity of infection (MOI) levels confirmed using TZM-bl cells (from John C. Kappes, Xiaoyun Wu, and Tranzyme Inc.) (60). Macrophages were infected with HIV at an MOI of 0.04 unless otherwise stated.
Statistics. Sample sizes are stated in the figure legends and refer to the number of independent replicates (n). Data are represented as dot blots with means Ϯ 95% confidence intervals. Data were assessed for symmetry, or skewness, using Pearson's skewness coefficient. Normalized ratiometric data were log 2 transformed. When the level of protein expression in the reference sample was zero, we used proportion statistics to detect differences (58,66). Comparisons between groups were performed using the paired, two-tailed Student's t test. In all experiments, differences were considered significant when the P value was less than 0.05 (*, P Ͻ 0.05).

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/mBio .02638-19. of Mental Health (NIMH). The content is solely our responsibility and does not necessarily represent the official views of the NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.