Inhibitory and Stimulatory Effects of IL-32 on HIV-1 Infection

IL-32 is a proinflammatory cytokine that induces TNF-α in a macrophage-like cell line and activates the NF-κB and p38 MAPK pathways (1). IL-32 does not share sequence homology with any known cytokine families (2). The receptor for IL-32 has not been identified, and no rodent orthologs of IL-32, and thereby no IL-32–knockout mice, have been reported (2). Despite these unconventional features, studies using transgenic mice overexpressing human IL-32 or recombinant human (rh)IL-32 have demonstrated that IL-32 is associated with inflammatory conditions and several diseases, such as tumors and infectious diseases, including HIV-1 (3).

The levels of IL-32 proteins in the plasma are known to be elevated in individuals infected with HIV-1 (4, 5) even after effective antiretroviral therapy (ART) (6), which is the case with the mRNA levels of IL-32 in PBMCs (7). T cells and NK cells highly expressed IL-32 mRNA (6, 7). Elevated IL-32 expression has also been observed in the gut and lymph nodes of HIV-1–infected individuals during the course of infection, with the highest levels at the stage of AIDS (8). Cytokeratin⁺ epithelial cells, CD4⁺ T cells, CD20⁺ B cells, CD11c⁺ dendritic cells, and CD163⁺ macrophages expressed IL-32 proteins (8). However, the role of IL-32 in HIV-1 infection remains a topic of controversy (3).

When added to HIV-1–infected macrophage-like cell line U1, the γ-type of rhIL-32 (rhIL-32γ), the largest IL-32 among nine alternatively spliced isoforms (2), reduced viral production (9). We also demonstrated the inhibitory effect of rhIL-32γ on HIV-1 production in monocyte-derived macrophages (MDMs) (10), although the underlying mechanisms remain to be elucidated. On the one hand, a similar antiviral effect of IL-32 has been reported for influenza virus, vesicular stomatitis virus (VSV), HSV type 2 (HSV-2), and hepatitis B virus (HBV) (11–14). On the other hand, when added to CD4⁺ T cells from ART-treated HIV-1–infected aviremic individuals, rhIL-32γ increased viral production (6). Palstra et al. also reported the stimulatory effect of rhIL-32γ on HIV-1 production in CD4⁺ T cells (15). A possible explanation for this is the inflammatory activation of CD4⁺ T cells by IL-32, which results in an increased susceptibility to HIV-1 (6, 15). However, the concentrations of IL-32 used in studies using CD4⁺ T cells were higher than those in studies using macrophages (250–500 ng/ml versus 10–100 ng/ml) (6, 9, 10, 15). Moreover, Mesquita et al. reported that when added at 100 ng/ml, rhIL-32γ reduced but did not increase viral production induced by PHA in CD4⁺ T cells of HIV-1–infected individuals (16). The reasons for these conflicting results remain unclear. The finding that IL-32 upregulates the expression of IDO1 (8), a well-known immunosuppressive molecule (17), makes it more difficult to understand the role of IL-32 in HIV-1 infection because it raises the possibility that IL-32 dampens the antiviral immune response of T cells through IDO1, thereby supporting HIV-1 replication in vivo (8, 17).

In this study, to gain insights into the role of IL-32 in HIV-1 infection, we sought to clarify the molecular mechanism by which IL-32 inhibits HIV-1 in MDMs, whether IL-32 upregulates other immunosuppressive molecules, and whether IL-32 has an additional effect beneficial for HIV-1 infection.

Human peripheral blood MDMs were prepared as previously described (10, 18). Briefly, PBMCs were seeded onto multiwell plates or dishes, and monocytes were enriched by allowing them to adhere. These monocytes were cultured for 5–7 d with RPMI 1640 with 10% FCS containing 100 ng/ml rhM-CSF (a gift from Morinaga Milk Industry) to induce the differentiation into MDMs. After washing with PBS to remove nonadherent cells, these MDMs were either cultured under various conditions or used in the experiments described below. Heparinized venous blood was collected from healthy donors after informed consent was obtained in accordance with the Declaration of Helsinki. Approval for this study was obtained from the Kumamoto University Medical Ethics Committee.

rhIL-32 (α-, β-, and γ-isoforms) were purchased from R&D Systems and used at a final concentration of 100 ng/ml unless otherwise stated. In this study, we mainly used the γ-isoform because it has more potent activity than other isoforms (14). rhIFN-α2 and rhIFN-γ were purchased from BioLegend and used at a final concentration of 10 ng/ml. Recombinant viral B18R proteins (inhibitor of type I IFN) and recombinant viral B8R proteins (inhibitor of type II IFN) were purchased from R&D Systems and used at a final concentration of 50 ng/ml.

The following chemical inhibitors were used: PD0332991 (inhibitor of cyclin-dependent kinase 4 [CDK4] and CDK6; Sigma-Aldrich), BMS-345541 (inhibitor of NF-κB kinase [IKK]-2 and IKK-1; Selleckchem), SB202190 (p38 MAPK inhibitor; Selleckchem), fludarabine (Stat1 inhibitor; Selleckchem), Stattic (Stat3 inhibitor; Selleckchem), and JQ1 (inhibitor of a BET bromodomain; Sigma). These inhibitors were dissolved in DMSO and added to the cultures at 0.1% v/v. The same volume of DMSO was used as the vehicle control. The final concentrations used were as follows: 0.1 or 1 μM for PD0332991; 10 μM for BMS-345541 and SB202190; and 5 μM for fludarabine, Stattic, and JQ1.

We used two recombinant HIV-1 viruses, the wild-type NL(AD8) strain (19) and the VSV envelope glycoprotein (VSV-G)-pseudotyped viruses, which are capable of infecting independently of the HIV-1 receptor CD4 (20). 293A cells (Invitrogen) cultured with DMEM with 10% FCS were employed as viral producer cells, as described previously (20). Briefly, the cells were transfected with HIV-1 proviral plasmids using the Lipofectamine 2000 reagent (Invitrogen). After 2 d, the supernatants containing viruses were collected and stored at −70°C in aliquots before use. To quantify the viral amount of the stock, the concentration of HIV-1 structural Gag proteins (p24 Gag) was measured by ELISA (MBL). The proviral plasmid for the CCR5-tropic NL(AD8) strain was obtained through the National Institutes of Health (NIH) AIDS Reagent Program (Division of AIDS, National Institute of Allergy and Infectious Diseases, NIH). To prepare the VSV-G–pseudotyped viruses, the 293A cells were cotransfected with an NL43 strain-based envelope (Env)-defective mutant proviral plasmid (pNL-Kp) and VSV-G expression plasmid (20).

The HIV-1 infection assay was performed as previously described (10, 18). MDMs were incubated with the viruses (p24 Gag concentration 10 ng/ml) at 37°C for 2 h. Then, the cells were washed with PBS to remove unbound viruses and cultured in the presence or absence of IL-32. Viral replication was monitored using several systems. The concentration of Gag in the culture supernatants was determined by ELISA (MBL). The percentage of infected cells was determined by flow cytometry (10): The cells were fixed, permeabilized, and stained with FITC-labeled anti-Gag Abs (KC57; Coulter) and analyzed on a FACSCanto II (BD Biosciences) using FlowJo software (BD Biosciences). The expression of HIV-1 Env proteins was quantified by immunofluorescence (21), and the cells were stained with anti-Env Abs (KD247) (provided by S. Matsushita, Kumamoto University, Kumamoto, Japan) followed by anti-human IgG–Alexa Fluor 488 (Molecular Probes). Signals were visualized using an FV1200 confocal laser-scanning microscope (Olympus), and image processing was performed using the FV Viewer (version 4.1) software (Olympus). The signal density of Env was quantified using ImageJ (1.52n) software (NIH). The number of copies of an early product of viral reverse transcription (R/U5 DNA) was also measured by real-time PCR (22, 23). The NL(AD8) viral stock was treated with Benzonase (Novagen) before the infection to digest proviral plasmids (19), and the DNA of infected MDM was isolated using DNeasy blood/tissue kit (Qiagen).

The expression of mRNAs encoding HIV-1 restriction factors (MX2, IFN-induced transmembrane 1 [IFITM1], IFITM2, IFITM3, and APOBEC3G), CDKs (CDK1, CDK2, CDK4, and CDK6), and immunosuppressive molecules (IDO1, IDO2, tryptophan 2,3-dioxygenase [TDO], PD-L1, and PD-L2) was analyzed by real-time PCR as described previously (24). Briefly, RNA was isolated using ISOGEN II reagent (Nippon Gene), and cDNA was prepared using Moloney murine leukemia virus RT (Invitrogen). Quantitative PCR was performed with SYBR Premix Ex Taq II (TaKaRa Bio) using a LightCycler (Roche). The expression level of each mRNA was normalized to that of β-actin. The sequences of the primers used are summarized in Supplemental Table I.

The expression of an intracellular molecule (IDO1) or cell surface molecules (BST-2, CD4, PD-L1, and PD-L2) was determined by flow cytometry (10). Briefly, the cells were detached using an enzyme-free cell dissociation buffer (13151-014; Life Technologies), stained with fluorescent dye–labeled Abs, and analyzed on the FACSCanto II using FlowJo software. To detect IDO1, the cells were fixed and permeabilized prior to staining. The Abs used were as follows: allophycocyanin-labeled anti-IDO1 (eyedio; eBioscience), PE-labeled anti-BST2 (RS38E; BioLegend), allophycocyanin-labeled anti-CD4 (RPA-T4; BioLegend), PE-labeled anti–PD-L1 (29E.2A3; BioLegend), and allophycocyanin-labeled anti–PD-L2 (MIH18; BioLegend).

The expression of HIV-1 restriction factors (MX2, IFITM1, IFITM3, APOBEC3G, SAMHD1, and phosphorylated SAMHD1) or a signaling molecule (Stat3 and phosphorylated Stat3) was determined by Western blotting (10). Briefly, MDMs were lysed on ice with Nonidet P-40 lysis buffer containing protease and phosphatase inhibitors. Total cell lysates were subjected to Western blot analysis. The Abs used were as follows: anti-MX2 (NBP1-81018; Novus Biologicals), anti-IFITM1 (60074-1-Ig; Proteintech), anti-IFITM3 (11714-1-AP; Proteintech), anti-APOBEC3G (HPA001812; Atlas Antibodies), anti-SAMHD1 (12586-1-AP; Proteintech), anti-phosphorylated (pThr592) SAMHD1 (D7O2M; Cell Signaling Technology), anti-IDO1 (13268-1-AP; Proteintech), anti-Stat3 (610189; BD Biosciences), anti-phosphorylated (pTyr705) Stat3 (612356; BD Biosciences), and anti-actin (EPR16769; Abcam). Detection was performed using HRP-labeled secondary Abs (GE Healthcare), the Immunostar LD detection reagent (Wako), and an ImageQuant LAS4000 image analyzer (GE Healthcare). The density of each band of interest on the blots was quantified using ImageJ software after normalization to the density of the control actin band.

To deplete SAMHD1 in MDMs, we used virus-like particles (VLPs) containing Vpx proteins that are encoded by HIV-2 or SIVs, which are well known to induce the degradation of SAMHD1 (25, 26).VSV-G–pseudotyped Vpx⁺ VLPs were prepared by transfecting 293A cells with pMDL-chp6, pVSV-G, pRSV-Rev, and myc-His–tagged Vpx plasmids (27). All plasmids were provided by N.R. Landau (New York University Grossman School of Medicine, New York, NY). After 2 d, the supernatants containing VLPs were collected and stored at −70°C in aliquots before use. Control VLPs lacking Vpr were also prepared. VLP stock was added to the cultures of MDMs at 10% v/v and incubated at 37°C for 6 h. After washing with PBS, NL(AD8) viral stock was added to the cultures, and infection assays were performed as described above (see HIV-1 infection and replication assays).

The number of cells was assessed using MTT reagent (Fig. 4A, lower panel; Fig. 7, middle panel). Alternatively, MDMs were subjected to repeated pipetting, and the number of floating cells was counted using the trypan blue dye exclusion method (Fig. 7B, top panel).

The morphologies of MDMs were analyzed by immunofluorescence (21). Briefly, the cells were cultured onto glass slides with rhM-CSF alone or with rhM-CSF and rhIL-32γ, fixed in paraformaldehyde, permeabilized with Triton X-100, and incubated with phalloidin conjugated to Alexa Fluor 488 and DAPI (both from Molecular Probes) to stain F-actin and nuclei, respectively. Signals were visualized using the FV1200 confocal laser-scanning microscope, and image processing was performed using the FV Viewer software.

The migration of MDMs was measured using a Transwell assay with 8-μm pore size inserts (Corning) (21). The inserts were placed into 24-well plates containing 600 μl of RPMI 1640 with 10% FCS. Then, MDMs detached using the enzyme-free cell dissociation buffer were added to the inserts (2 × 10⁵ cells in 100 μl of RPMI 1640 with 10% FCS containing rhM-CSF alone or rhM-CSF and rhIL-32γ) and incubated at 37°C for 24 h. The number of cells that migrated through the inserts was counted using MTT reagent. The wound healing assay was performed as described previously (21). A linear wound was generated in the sheet of the MDMs using a 200-μl pipette tip. Floating cells were removed by washing with media. Cells were then incubated for 24 h with rhM-CSF alone or with rhM-CSF and rhIL-32γ, and the number of cells in the wound area randomly selected was enumerated.

The statistical significance of intersample differences was determined using a paired Student t test. For multiple comparisons, two-way ANOVA with Sidak’s multiple comparisons test was used. The Mann-Whitney U test was used to compare data sets with nonnormal distributions using Prism 8 (GraphPad Software). Statistical significance was set at p < 0.05.

We previously reported that when MDMs were precultured with IL-32γ for 2 d followed by infection with HIV-1, viral integration and production were severely reduced (10). In this study, to clarify the inhibitory mechanism, we tested additional culture/infection conditions. When MDMs were infected with wild-type HIV-1 and then cultured with IL-32γ, viral production in the supernatants was still reduced (Fig. 1A, 1B, upper panel). An inhibitory effect was also observed for the percentage of intracellular Gag⁺ cells (Fig. 1B, lower panel) and the level of HIV-1 Env expression (Fig. 1C). The anti–HV-1 activity of IL-32γ was dose dependent (Fig. 1D), and 100 ng/ml, the concentration that we used in the previous study (10), was enough to observe the inhibitory effect in all donors tested (Fig. 1D). Thus, we added IL-32γ to cultures at a final concentration of 100 ng/ml also in this study. The anti–HIV-1 activity of IL-32γ was unlikely to be mediated by IFNs, because it was not neutralized by B18R or B8R proteins (Supplemental Fig. 1A, 1B), the vaccinia virus–encoded IFN inhibitors (28, 29). Indeed, IFN-α, but not IL-32γ, upregulated BST2 (Supplemental Fig. 1C, upper panel), a cell surface molecule that inhibits the release of HIV-1 particles (30). These results appear to explain the fact that IL-32γ reduced the extracellular Gag weakly when compared with IFN-α (Fig. 1B, upper panel), but IL-32γ and IFN-α similarly reduced intracellular Gag (Fig. 1B, lower panel).

As we previously reported (10), IL-32γ downregulated CD4, the receptor of HIV-1 (Supplemental Fig. 1C, lower panel). However, this may not be the main mechanism of the anti–HIV-1 activity of IL-32γ, because it still reduced the viral production even when MDMs were infected with VSV-G–pseudotyped HIV-1 viruses that infect independently of CD4 (Fig. 1E, 1F). In these experiments, IL-32γ was added to MDMs immediately after 2 h of incubation with the viruses. Interestingly, when added 9 h after incubation with viruses, IL-32γ lost a considerable part of its anti–HIV-1 activity (Fig. 1G, “Delayed-IL-32γ”), which was not observed for IFN-α (“Delayed-IFN-α”). These results suggest that IL-32γ inhibits HIV-1 during the early phase(s) after viral entry. The finding that IL-32γ reduced the synthesis of R/U5 DNA (Supplemental Fig. 1D), the early product of viral reverse transcription (22, 23), further supported this idea.

Next, we examined the effect of IL-32 on HIV-1 restriction factors that may inhibit the postentry early step, such as MX2, IFITMs, APOBEC3G, and SAMHD1 (25, 26, 30, 31). IL-32γ upregulated MX2, IFITM1, IFITM3, and APOBEC3G at the mRNA (Fig. 2A) and protein levels (Fig. 2B). However, the extent of upregulation was modest when compared with that of IFN-α (Fig. 2C). Meanwhile, IL-32γ appeared to promote the activity of SAMHD1 to restrict HIV-1 to a similar extent as IFN-α (Fig. 3A, 3B). SAMHD1 is a 2′-deoxynucleoside 5′-triphosphate triphosphohydrolase that inhibits viral reverse transcription (25, 26) but loses its inhibitory activity when phosphorylated by CDKs (32–34). In MDMs, SAMHD1 exists as a mixture of phosphorylated (inactive) and unphosphorylated (active) forms (35). In this study, we found that both IFN-α and IL-32γ reduced inactive SAMHD1 to an undetectable level at 12 h after stimulation (Fig. 3A, 3B). Consistent with this, IL-32γ downregulated CDK1, CDK4, and CDK6 (Fig. 3C). The downregulation of CDKs was dependent on the concentration of IL-32γ (Supplemental Fig. 2A) and was observed for IL-32β, but not IL-32α (Supplemental Fig. 2B), the least potent isoform (10, 14).

As reported by Pauls et al. (36), the dual inhibitor of CDK4 and CDK6 (PD0332991) severely reduced HIV-1 production in MDMs (Fig. 4A, upper panel) without affecting cell viability (Fig. 4A, lower panel), suggesting that the phosphorylation of SAMHD1 by CDK4 and/or CDK6 is important for efficient HIV-1 infection in MDMs. Combined with the finding that IL-32γ, but not IFN-α, strongly downregulated CDK6 (see (Fig. 3C, bottom panel), we hypothesized that IL-32γ inhibits HIV-1 in an SAMHD1-dependent manner. To test this hypothesis, we used Vpx proteins that are well known to induce the degradation of SAMHD1 (25–27). Under the SAMHD1-deficient conditions (Fig. 4B), IL-32γ, but not IFN-α, largely lost its inhibitory effect on HIV-1 production (Fig. 4C, 4D). Taken together, our results suggest that IL-32γ inhibits HIV-1 by downregulating CDKs, which leads to the activation of SAMHD1 and the inhibition of viral reverse transcription.

Despite its anti–HIV-1 activity, IL-32 has been proposed to play a supportive role in HIV-1 infection because it upregulates IDO1 (8). Indeed, IL-32γ–treated MDMs showed higher IDO1 expression than the control MDMs at both the protein (Fig. 5A, 5B) and mRNA levels (Fig. 5C, top panel). IDO1 produced by macrophages or dendritic cells catabolizes tryptophan into kynurenine, which is a key metabolic pathway that restricts the immune response of T cells (17, 37). It was also reported that elevated levels of IDO1 during HIV-1 infection skewed CD4⁺ T-cell differentiation into regulatory T cells (38). In this study, we found that IL-32γ also upregulated IDO2 (Fig. 5C), a paralog of IDO1 (37). IL-32γ did not affect the expression of IL4I1 (Fig. 5C), a recently identified tryptophan catabolic enzyme (39), but weakly upregulated TDO (Fig. 5C), another enzyme involved in tryptophan catabolism (37). Consistent with our previous finding that IL-32γ activates the NF-κB pathway and MAPKs, such as p38 in MDMs (10), BMS-345541 (inhibitor of IKK2 and IKK1) and SB202190 (inhibitor of p38 MAPK) diminished the upregulation of IDO1 and IDO2 by IL-32γ (Fig. 5D). Stattic (inhibitor of Stat3), but not fludarabine (inhibitor of Stat1), also diminished the upregulation of IDO1 and IDO2 by IL-32γ (Fig. 5D), which was consistent with the finding that IL-32γ activates Stat3 (Fig. 5E).

It has been demonstrated that the blockade of PD-1 expressed on T cells and its cognate ligand PD-L1 expressed on macrophages and other cells, a well-known immunosuppressive pathway, is effective against several tumors (40). In HIV-1 infection, the PD-1 signal in virus-specific CD8⁺ T cells was associated with T-cell exhaustion and disease progression (41, 42). The blockade of the PD-1/PD-L1 pathway in rhesus macaques infected with HIV-1–related virus SIV improved the functions of CD8⁺ T cells (43). Moreover, to enhance the efficacy of immunotherapy, a combined blockade of PD-1/PD-L1 and IDO1 has been evaluated in clinical trials (44). In this study, we found that IL-32γ upregulated PD-L1 mRNA (Fig. 6A, upper panel) but not another PD-1 ligand, PD-L2 (Fig. 6A, lower panel), in MDMs. The upregulation of IDO1 and IDO2 by IL-32γ was modest when compared with that by IFN-α or IFN-γ (Fig. 5C), whereas the upregulation of PD-L1 by IL-32γ was comparable to that by IFN-α or IFN-γ (Fig. 6A, upper panel, and (Fig. 6B), which was confirmed at the protein level (Fig. 6C). As observed for the upregulation of IDO1 and IDO2 (Fig. 5D), BMS-345541, SB202190, and Stattic diminished the upregulation of PD-L1 by IL-32γ (Fig. 6D). Recently, the BET bromodomain inhibitor JQ1 was shown to reduce the basal or IFN-γ–induced expression of PD-L1 in tumor cells (45), which enhanced the efficacy of PD-1/PD-L1–targeting immunotherapy (46). In this study, we found that JQ1 also diminished the upregulation of PD-L1 by IL-32γ in MDMs (Fig. 6E). Our results indicate that IL-32γ upregulates various immunosuppressive molecules, including IDO1, IDO2, TDO, and PD-L1.

Additionally, we attempted to clarify whether IL-32 has other functions beneficial for HIV-1 infection. MDMs treated with IL-32γ formed unique structures (Fig. 7A, white arrowheads), which resemble podosomes or invadopodia of tumor cells that mediate invasive cell motility (47). Thus, we analyzed the effects of IL-32γ on the cell motility–related phenotypes of MDMs because the motility of infected cells can increase the likelihood of encountering target cells, thereby facilitating viral transmission (21, 48). We found that IL-32γ increased the number of MDMs loosely adhered to dishes (Fig. 7B, top panel) and enhanced both the migratory (Fig. 7B, middle panel) and wound healing activities (Fig. 7C, bottom panel) of MDMs. These results indicate that IL-32γ enhances the motility of MDMs, which is consistent with the formation of podosomes or invadopodia-like structures and weak adhesive properties.

The present study demonstrated that the anti–HIV-1 activity of IL-32 in MDMs mainly depends on the CDK-SAMHD1 cascade. However, at the same time, this study revealed that IL-32 exerts multiple effects that support HIV-1 infection. Specifically, IL-32 upregulates IDO1, IDO2, TDO, and PD-L1 in MDMs and enhances the motility of MDMs. These findings indicate that IL-32 has both the direct inhibitory effect on HIV-1 production in MDMs and the indirect stimulatory effects through phenotypic modulation of MDMs, and they suggest that the stimulatory effects may outweigh the inhibitory effect because the window for IL-32 to inhibit HIV-1 is relatively confined to the suppression of SAMHD1-mediated reverse transcription in the life cycle of HIV-1.

Our finding that IL-32 requires SAMHD1 to inhibit HIV-1 (Fig. 4) may explain why IL-32 inhibits HIV-1 in MDMs (10) but not CD4⁺ T cells (6, 15), because SAMHD1 restricts HIV-1 in noncycling MDMs (25, 26) but not cycling CD4⁺ T cells (32–34). In cycling cells, SAMHD1 is constitutively phosphorylated by CDKs, thereby losing its anti–HIV-1 activity. The idea that the window for IL-32 to inhibit HIV-1 is relatively confined to the suppression of CDK/SAMHD1-mediated viral reverse transcription may explain why another study showed a very weak anti–HIV-1 activity of IL-32 in MDMs (49). Among CDKs tested, IL-32 downregulates CDK1, CDK4, and CDK6 (Fig. 3). Interestingly, Pauls et al. demonstrated that the knockdown of CDK6, but not CDK1 or CDK4, led to a severely reduced phosphorylation of SAMHD1 in MDMs (35). Thus, it is highly likely that the CDK6-SAMHD1 cascade is critical for the anti–HIV-1 activity of IL-32. Given that CDK6 and IL-32 inhibit and stimulate macrophage differentiation, respectively (50, 51), IL-32–mediated CDK6 downregulation may be important for the two distinct activities of IL-32, namely, macrophage differentiation-stimulating activity and anti–HIV-1 activity.

IL-32 has been shown to upregulate IDO1 expression in macrophages (8). In this study, we demonstrated that IL-32 also upregulates related enzymes, TDO and IDO2, in MDMs (Fig. 5). IDO1 and TDO have been well characterized to accelerate the immunosuppressive tryptophan catabolism along the kynurenine pathway (17, 37). Although IDO2 has a relatively weak enzymatic activity, it may also act as an immunosuppressive molecule, because the development of pancreatic ductal adenocarcinoma was notably decreased in IDO2-knockout mice (52). We also demonstrated that IL-32 upregulated PD-L1 expression (Fig. 6). Notably, across multiple tumor types, responses to PD-L1 inhibition have been observed in patients with tumors expressing high levels of PD-L1, particularly when PD-L1 is expressed by host cells, including macrophages (53). A similar result was obtained using various mouse tumor models (54, 55). These findings indicate that PD-L1 expression on macrophages is important for the immunosuppressive role of the PD-1/PD-L1 pathway. IL-32 upregulates PD-L1 to a similar extent as IFN-γ (Fig. 6), which is generally considered the most prominent inducer of PD-L1 (56). Thus, PD-L1 upregulation by IL-32 in macrophages may contribute to immunosuppression in HIV-1 infection.

IL-32 has been considered to induce proinflammatory M1-like macrophages (1). In fact, when added to M-CSF–derived MDMs, IL-32 enhanced the production of IL-8 or IL-6 and the expression of the costimulatory molecule CD80 (10). However, IL-32 also enhanced the anti-inflammatory M2-like characteristics, including phagocytic activity, and CD14 and CD163 expression (10). Recently, IL-32 was reported to induce detachment of Langerhans cells from the epidermal layer, allowing their migration in a human skin explant model (57). Consistent with this, we demonstrated that IL-32 induces the formation of invadopodia-like structures and weak adhesive properties and enhances the motility of macrophages (Fig. 7), which can facilitate the encounter of infected macrophages with potential target cells (21, 48).

IL-32 requires SAMHD1 to inhibit HIV-1 in MDMs (Fig. 4). MX2, which inhibits HIV-1 nuclear import (58), may also be involved in the anti–HIV-1 activity of IL-32 because it is upregulated by IL-32 (Fig. 2). In any case, the inhibitory effect of IL-32 is likely limited to early step(s) because the 9 h delayed addition of IL-32 to infected macrophages minimally inhibited HIV-1 (Fig. 1G). Given the two distinct functions of IL-32 in macrophages that are beneficial for HIV-1 infection or transmission, that is, the upregulation of many immunosuppressive molecules and enhancement of cell motility, our study suggests that IL-32 supports HIV-1 infection and transmission overall. However, further studies are needed to clarify how the upregulation of IDO1, IDO2, TDO, and PD-L1 in macrophages by IL-32 contributes to the pathogenesis of HIV-1 infection. IL-32 exists as several isoforms, such as α, β, δ, γ, ε, θ, and ζ (2, 3), and intracellular IL-32, but not extracellular IL-32, was shown to exert anti-HBV activity (14). Thus, it is also necessary to study how different isoforms and intracellular forms of IL-32 affect HIV-1 infection.

We thank N.R. Landau (New York University Grossman School of Medicine, New York, NY) for providing plasmids to prepare VSV-G-pseudotyped Vpx⁺ VLP. We also thank K. Tomoda and M. Tsunemasu for their secretarial assistance and Editage for the English language editing and review.

The authors have no financial conflicts of interest.