HIV-1 infection induces M1 polarization of macrophages

THP-1 cells or monocytes from healthy donors were respectively differentiated into macrophages in vitro, for the following infection experiments. The morphology of HIV-1–infected monocyte-derived macrophages (MDMs) under an inverted optical microscope showed that macrophages gradually developed an irregular shape and had prolonged long pseudopodia post–HIV-1 infection (Fig 1A), similar to LPS-stimulated M1 macrophages (Fig S1A). In addition, the mRNA expression of M1 markers, including TNF-α, IL-1β, and IL-6, was significantly elevated in HIV-1–infected cells compared with the control group (Figs 1B and S1B). Although the M2 markers CD163, CD206, and IL-10 increased simultaneously, the magnitude of the increase was much lower than that of the M1 markers (Fig 1B). At the protein level, we observed these M1 inflammatory factors (TNF-α, IL-1β, and IL-6) sustained the release to the supernatant during HIV-1 infection (Fig 1C). Furthermore, flow cytometric analyses revealed an increased proportion of CD68⁺CD86⁺ M1 macrophages after HIV-1 infection (33.29% versus 67.60%), whereas the proportion of CD68⁺CD163⁺ M2 macrophages was just slightly increased (7.18% versus 12.33%), and the proportion of CD68⁺CD163⁺ and CD68⁺IL-10⁺ M2 macrophages was decreased after HIV-1 infection (22.61% versus 11.56%; 2.18% versus 7.80%, respectively) (Fig 1D). Taken together, the above evidence confirms that HIV-1 infection induces M1 polarization of macrophages.

Changes in glucose metabolism in HIV-1–infected macrophages

Previous studies have shown that the metabolic reprogramming is involved in polarization of macrophages (Russell et al, 2019). Changes in the ATP/ADP ratio can indicate the primary energy metabolic pathway in the cell (Maldonado & Lemasters, 2014; Tan et al, 2022). Therefore, we used the ATP/ADP ratio to determine the metabolic state of HIV-1–infected macrophages. As a result, a lower ATP/ADP ratio was observed under HIV-1 infection, which indicated the cell metabolism is predominantly glycolytic (Fig 2A, left panel). Besides that, we measured the relative ATP production from glycolysis and OXPHOS in the context of HIV-1 infection in macrophages via inhibition of two ATP production pathways. Inhibitor 2-DG (10 mM), targeting hexokinase II (HK2), was used to block glycolysis (Pajak et al, 2019), and inhibitor oligomycin (40 μM) was used to block mitochondrial ATP production (Nieminen et al, 1994). Inhibition of glycolysis with 2-DG resulted in a greater ATP reduction in HIV-1–infected MDMs compared with control cells (HIV-1: 51.2% versus Control: 40.6%), whereas oligomycin inhibition of mitochondrial ATP production had a greater effect on reduction of ATP in uninfected control cells (Fig 2A, right panel), revealing that HIV-1 facilitates the use of glycolysis for energy supply.

To verify the changes in glucose metabolism pathways by HIV-1 infection, glucose uptake was detected. The results showed a significant increase in the proportion of glucose uptake in HIV-1–infected cells compared with control cells (HIV-1: 23.44% versus Control: 1.66%) (Fig 2B). As the final metabolic product of glycolysis, lactate accumulation means the increase in glycolysis. In the HIV-1–infected group, the level of lactate in the supernatant was about 1.7 times that in the control group (Fig 2C), and the results of the extracellular acidification assay yielded identical results (Fig 2D). More significantly, the latter showed a time-dependent effect. Meanwhile, we performed the oxygen consumption assay and the JC-1 mitochondrial membrane potential (MMP) assay to estimate the mitochondrial activity during HIV-1 infection. HIV infection led to a significant decrease in the oxygen consumption rate (OCR) and a 40% reduction in MMP, indicating impaired mitochondrial activity (Fig 2E and F). By calculating glycolytic indexes (lactate production × glucose uptake rate/OCR) (Valentín-Guillama et al, 2018), comprehensively, we revealed that glycolysis was enhanced in HIV-1–infected MDMs (Fig 2G).

In addition, we detected a few key genes for TCA and OXPHOS in HIV-1–infected MDMs. Intriguingly, the expression of all these genes was down-regulated. Among them, isocitrate dehydrogenase (IDH) was down-regulated by about two times (P < 0.01), and malate dehydrogenase was down-regulated by about 1.6 times (P < 0.05) (Fig 2H). IDH and malate dehydrogenase are both important rate-limiting enzymes in the TCA cycle. IDH catalyzes the synthesis of α-ketoglutarate, which is the intermediate product connecting carbon metabolism and nitrogen metabolism in the body.

The above results suggest that HIV-1 infection may alter the glucose metabolism of macrophages by promoting glycolysis and inhibiting TCA and OXPHOS.

Suppression of glycolysis blocks M1 polarization of macrophages in HIV-1 infection

In multiple reports, M1 macrophages are characterized by increased glycolytic activity, whereas M2 macrophages are characterized by increased fatty acid oxidation (FAO), glutaminolysis, and mitochondrial respiration (Viola et al, 2019). To investigate the impact of metabolic reprogramming in HIV-1–induced M1 polarization, we used 2-DG (targeting HK2) and heptelidic acid (targeting GAPDH) to block glycolysis. As shown in Figs 2I and S2A, both 2-DG and heptelidic acid treatment effectively reduced the elevated lactate release by HIV-1 infection. Under the condition of glycolysis inhibition, we observed that the expression of M1 markers TNF-α, IL-1β, and IL-6 was also reversed (Figs 2J and S2B–D), indicating that blocking glycolysis inhibits M1 polarization of macrophages driven by HIV-1 infection. In contrast, treatment of oligomycin promoted the release of pro-inflammatory cytokines, as evidenced in Fig S2.

HIV-1 infection leads to immunometabolic reprogramming in macrophages via activation of HIF-1α

To further study the molecular mechanism of immunometabolic reprogramming by HIV-1 infection, we investigated the expression of HIF-1α, a known strong inducer of glycolysis, in HIV-1–infected MDMs. Immunofluorescence results showed that HIF-1α expression was enhanced after HIV-1 infection and was mainly located in the nucleus (Fig 3A). HIF-1α–activating genes, including HK2, PDK1, and LDHA (which function vitally in the glycolytic pathway), were also up-regulated by HIV-1 infection. However, in the presence of LW6 (15 μM), an inhibitor of HIF-1α, the HIV-1–induced up-regulation of HIF-1α, as well as its downstream targets HK2, PDK1, and LDHA, decreased to levels comparable to those in the control group (Fig 3B). The changes in expression of these factors at the protein level were also revalidated by Western blot, as shown in Fig 8B. As for key genes of OXPHOS, LW6 treatment had little effect on the down-regulated expression of TCA key genes induced by HIV-1 infection (Fig 3C). Meanwhile, treatment of another HIF-1α inhibitor YC-1 (1 μM) obtained similar results (Fig 3D and E). Next, we performed a lactate test and the JC-1 assay to determine the role of HIF-1α in HIV-1–induced glucose metabolic reprogramming. As shown in Fig 4A and B, elevated lactate induced by HIV-1 infection was effectively restrained under LW6 treatment, whereas the impaired mitochondrial activity was not restored, suggesting that HIF-1α activation contributes to glycolysis enhancement, but may not be a direct factor for the OXPHOS reduction in HIV-1 infection. In addition, we observed decreased expressions of TNF-α, IL-1β, and IL-6 after HIF-1α inhibition, both at mRNA (Fig 4C) and protein (Fig 4D) levels. Collectively, the above evidence suggests that HIV-1 infection activates the HIF-1α pathway to promote glycolysis and cause immunometabolic reprogramming in macrophages.

HIV-1 regulates HIF-1α–dependent immunometabolic reprogramming via COX-2

Previous studies have demonstrated COX-2 activation motivates pro-inflammation, which is the main manifestation of M1 macrophages (Simon, 1999). In HIV-1–infected macrophages, we found COX-2 mRNA and protein expressions were markedly up-regulated (Fig 5A and B). Therefore, we sought to investigate whether the up-regulation of COX-2 is involved in mediating HIV-1–induced immunometabolic reprogramming in macrophages. Using meloxicam (50 μM), a selective inhibitor of COX-2 synthesis (Engelhardt, 1996), we found that meloxicam significantly reduced COX-2 expression induced by HIV-1 infection (Fig 5C). Key markers of glycolysis were subsequently detected. Results revealed that meloxicam treatment reversed the HIV-1–induced lactate production (Fig 5D), and also reversed the increased expression of HIF-1α and key enzymes in glycolysis (HK2, PDK1, and LDHA), resulting in a rate of glycolysis comparable to uninfected MDMs (Figs 5E and 8B). In addition, results of immunofluorescence exhibited obviously weakened fluorescence of HIF-1α after meloxicam treatment (Fig 5F and G), suggesting that COX-2 regulates the enhancement of glycolysis during HIV-1 infection through the HIF-1α pathway.

We subsequently correlated COX-2 activation with M1 polarization, showing that the expression of HIV-1–induced pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) was notably reversed by meloxicam treatment, particularly at the level of protein expression (Fig 5H and I).

To revalidate the role of COX-2 in regulating M1 polarization of macrophages, we constructed a COX-2–silenced THP-1 cell line using lentivirus-mediated RNAi (Fig 6A). Real-time quantitative PCR (RT-qPCR) and Western blot proved stable COX-2 knockdown expression in THP-1–derived macrophages (COX-2i group) (Fig 6B and C). In COX-2i cells, HIF-1α–dependent glycolysis and the expression of TNF-α, IL-1β, and IL-6 induced by HIV-1 were both reduced (Fig 6D–G), which was consistent with the results obtained from meloxicam treatment. These findings indicated that COX-2 plays an important role in mediating HIF-1α–dependent immunometabolic reprogramming of macrophages upon HIV-1 infection.

Phosphorylation of JNK involves in HIV-1–regulating immunometabolic reprogramming in macrophages

To further explore the upstream events of COX-2 activation, we analyzed the differential genes between HIV-1–infected macrophages and control macrophages using bioinformatics techniques. Based on our transcriptome sequencing (RNA-seq) data, we found the MAPK signal transduction pathway was significantly enriched (Fig 7A). Gene set enrichment analyses revealed that genes involved in the MAPK pathway were more expressed in the HIV-1 infection group than those in the control group (Fig 7B, Table S1), supporting that MAPK signaling is associated with response to HIV-1 infection in macrophages. To further investigate which subpathway response is more sensitive to HIV-1 infection, we measured the expression of JNK, ERK1/2, and p38. At the mRNA level, we found that only JNK expression in the infection group was higher than that in the control group (P < 0.05), but there was no significant difference in the expression of total JNK at the protein level (Fig 7C–E). However, at the protein phosphorylation level, we found the ratio of p-JNK to total JNK protein was significantly higher in the HIV-1 infection group compared with the control group, whereas the ratios of p-ERK to ERK and of p-p38 to p38 had no significant differences between two groups (Fig 7C–E). Therefore, it is plausible that HIV-1 causes immune response of macrophages via activating the MAPK/JNK signaling pathway.

We then used SP600125, an inhibitor of JNK, to assess the role of JNK in HIV-1–induced immunometabolic reprogramming in macrophages. In the presence of SP600125 (50 μM), HIV-1–induced phosphorylation of JNK (p-JNK/total JNK) was reduced to the control level (uninfected with HIV-1 and untreated with SP600125), along with a dramatic reduction of HIV-1–induced COX-2 overexpression (Fig 8A), demonstrating the ability of JNK in regulating COX-2 expression. Meanwhile, under inhibition of JNK phosphorylation, the expression of key factors in glycolysis was significantly decreased (Fig 8B). In addition, to verify the role of COX-2 and HIF-1α in p-JNK–mediating metabolic reprogramming and M1 polarization, we also used meloxicam (50 μM), LW6 (15 μM), and YC-1 (4 μM) to specifically inhibit the COX-2 pathway and HIF-1α pathway, respectively. The results confirmed the inhibition of COX-2 or HIF-1α led to the decreased expression of key factors (HIF-1α, HK1, HK2, and LDHA) in glycolysis (Fig 8B), which provided evidence at the protein level to support previous results in Figs 3 and 5. As a result of p-JNK inhibition, the HIV-1–induced lactate displayed a 30–50% reduction compared with HIV-1 infection alone, indicating the weakened glycolysis (Fig 8C). Furthermore, the inhibition of p-JNK, along with the suppression of COX-2 and HIF-1α, as well as the inhibition of glycolysis, resulted in decreased M1 polarization. Most notably, expressions of TNF-α and IL-6 were not increased under SP600125 treatment, despite the strong M1 induction from HIV-1 demonstrated above (Fig 8D and E).

Taken together, by analyzing changes in COX-2 expression, glycolysis level, and M1 polarization in the presence or absence of JNK inhibition, we found that phosphorylation of JNK promotes glycolysis enhancing, which is regulated by the COX-2/HIF-1α signaling axis, supporting M1 polarization of macrophages in HIV-1 infection.

Ethics statement

Guangxi Medical University Ethical Committee approved all experimental procedures and protocols used in this study (Approval No. 2022-0130).

Reagents and antibodies

PMA, LPS, IFN-γ, IL-4, IL-10, and SP600125 were purchased from Sigma-Aldrich. 2-Deoxy-D-glucose (2-DG), heptelidic acid (HA), oligomycin (Omy), LW6, and YC-1 were from MCE (United States). Meloxicam was from TCI (Japan). Immunofluorescence blocking buffer (#12411), and Alexa Fluor 488 goat anti-rabbit (#4412; RRID: AB_1904025) and Alexa Fluor 594 goat anti-mouse (#8890; RRID: AB_2714182) secondary antibodies were purchased from Cell Signaling Technology. Antibodies against β-actin (#3700; RRID: AB_2242334), phospho-SAPK/JNK (#4668; RRID: AB_823588), SAPK/JNK (#9252; RRID: AB_2250373), phospho-p44/42 MAPK (Erk1/2) (#4370; RRID: AB_2315112), p44/42 MAPK (Erk1/2) (#4695; RRID: AB_390779), phospho-p38 MAPK (#4511; RRID: AB_2139682), p38 MAPK (#8690; RRID: AB_10999090), COX-2 (#12282; RRID: AB_2571729), and HIF-1α (#36169; RRID: AB_2799095) and Glycolysis Antibody Sampler Kit (#8337; RRID: AB_10897509) were obtained from Cell Signaling Technology. The secondary antibodies of IRDye 680RD donkey anti-mouse IgG and IRDye 800CW donkey anti-rabbit IgG were purchased from LI-COR.

Cell culture and differentiation

The human monocyte THP-1 cell line (conserved in our laboratory, ATCC number: TIB-202; RRID: CVCL_0006) and primary human macrophages were used as cell models for HIV-1 infection. THP-1 cells were routinely cultured in 1640 medium (Gibco) containing 10% heat-inactivated FBS (Gibco) and 1% penicillin–streptomycin solution (Solarbio) at 37°C and 5% CO₂ atmosphere. To obtain THP-1–derived macrophages, THP-1 cells were seeded in a six-well culture plate (10⁶ cells per well) with DMEM (Gibco) containing 10% FBS and 1% penicillin–streptomycin solution, and stimulated with 50 ng/ml of PMA (48 h) for differentiation (Dunn et al, 2011). Primary human macrophages were derived from PBMCs of healthy donors as previously described (Fang et al, 2009). Briefly, PBMCs were isolated from whole blood by Ficoll (GE Healthcare) through density gradient centrifugation according to the manufacturer’s protocol (Riedhammer et al, 2016). PBMCs were then incubated in 75-cm² flasks precoated with gelatin (VETEC) in a 37°C, 5% CO₂ atmosphere for 30 min. Adhesive monocytes were harvested using EDTA (Solarbio). Collected monocytes were seeded at a density of 4 × 10⁶ cells/ml in complete 1,640 medium (containing 10% human A⁺B⁺ serum) and refreshed medium every 3 d until differentiation into macrophages for experiments.

Virus production and infection

HIV-1_Bal virions were originally from the NIH AIDS Reagent Program and reproduced in our laboratory. The modified method was based on Edward’s report (Barker et al, 1998). In brief, PHA (5 μg/ml) in complete 1,640 medium (containing 100 U/ml IL-2) stimulated human PBMCs for 3 d at 37°C, 5% CO₂. Afterward, polybrene (2 μg/ml) was added to enhance HIV-1 infection. After HIV-1_Bal infection, the supernatant was harvested every 3 d and viral titers were quantified using a HIV-1 p24 ELISA kit (Jianglaibio). We adjusted the viral titer to 100 ng p24/ml and stored it at −80°C.

HIV-1 infection was performed after differentiation of monocytes into macrophages. HIV-1 virions were added to culture medium (8 ng p24/10⁶ cells) for 6 h. Then, cells were washed with PBS (Solarbio) and cultured in fresh medium until harvest at the appointed time.

Lentiviral construction and transduction

Lentiviral particles were obtained from GeneChem. In detail, the siRNAs for COX-2 (NM_000963.4) (COX-2i) were chimeric into the GV248 plasmid. The sequence for the Vector was as follows: 5′-TTCTCCGAACGTGTCACGT-3′. The sequence for COX-2i was as follows: 5′-GCTGAATTTAACACCCTCTAT-3′. Titers of concentrated viral particles were 8.0 × 10⁸ units/ml. Lentiviral particles were transduced into THP-1 cells during the logarithmic growth phase. The supernatant was removed after 24 h of infection and replaced with complete 1,640 medium. To generate stable COX-2 knockdown cell lines, 2 μg/ml puromycin was added to medium for 48 h. A fluorescent microscope was used to observe successful transduction. RT-qPCR and Western blot were used to confirm the down-regulated expression of COX-2 in COX-2i THP-1 cells compared with Vector THP-1 cells.

RNA extraction and RT-qPCR

Total cellular RNA was extracted using TRIzol (Sigma-Aldrich). The cells were pipetted with TRIzol (40 μl/10⁵ cells) until completely lysed, and chloroform (Tedia) was added for stratification by the gradient. After centrifugation (4°C, 12,000 g/min, 15 min), the transparent supernatant was collected. Isopropanol (Tedia) was used to precipitate RNA, and 75% ethanol was for purification. The remaining RNA pellet was resuspended in enzyme-free water, followed by incubation at 55°C for 15 min. The quantification was performed using NanoDrop 2000C (United States). The cDNA was synthesized with PrimeScript RT Master Mix (Takara) using 1 μg RNA. cDNA samples were used for RT-qPCR with SYBR Premix Ex Taq^TM (Takara) according to the operation manual. The sequences of primer pairs are listed in Table S2. Relative mRNA expressions were analyzed based on the 2^−ΔΔCt method, normalizing to the housekeeping gene GAPDH.

Cytokine and lactate measurements

Culture supernatants were collected at the indicated time points, and protein concentrations of TNF-α, IL-1β, and IL-6 were quantified by ELISA according to the manufacturer’s instructions (NOVUS). Lactate production was measured by a lactate kit (Njjcbio) using colorimetry according to the manufacturer’s instructions.

Flow cytometry

For detection of polarization markers of macrophages, MDMs were harvested by scraping. The harvested cells were washed with PBS to remove the culture medium thoroughly and collected by centrifugation. For intracellular staining, the cells were fixated and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences), then stained with PE-linked anti-CD68 (#12-0689-41, RRID: AB_10805746; eBioscience), APC-linked anti-CD86 (#17-0869-41, RRID: AB_2802218; eBioscience), and APC-linked anti-CD206 (# 17-2069-42, RRID: AB_2573182; eBioscience) for 20 min at 4°C. The stained cells were washed with PBS and finally resuspended in PBS. Flow cytometry was performed using a CytoFLEX2 flow cytometer (Beckman Coulter) and analyzed by FlowJo software.

Western blot

Protein samples were prepared using lysis buffer (RIPA: PMSF = 100:1, RIPA: phosphatase inhibitors = 100:1). The total protein concentration of samples was quantified using the BCA Protein Assay kit (Beyotime) to make sure each sample had an equal amount of protein for SDS–PAGE. After transferring to PVDF membranes, the membranes were blocked with 5% skim milk for 1 h at RT and then incubated with primary antibody overnight at 4°C. Fluorescent secondary antibody was used for detection. Relative band intensities were measured by the LI-COR Odyssey CLx imaging system.

ATP assay and ATP/ADP ratio assay

ATP concentrations were measured using CellTiter-Glo Kit (Promega) following the manufacturer’s protocol. Differentiated macrophages were infected with or without HIV-1 for 48 h in refreshed medium, and incubated with detection reagents at 37°C for 10 min. The fluorescence intensity was read under a microplate reader (Synergy H1; BioTek). ATP concentrations were determined and normalized to the protein level of each sample.

Changes in the ATP/ADP ratio were detected using ATP/ADP-Lite Assay Kit (Vigorous Bio). The treated MDMs were lysed with lysis buffer for 10 min at RT with gentle shaking. Subsequently, 50 μl of assay reagent was mixed with 10 μl of the cell lysate, and the samples were then analyzed using a microplate reader to obtain the initial readings (Data A). After a 10-min incubation, another reading was taken (Data B). To each sample, 10 μl of ATP converting enzyme was added, followed by another reading after 10 min (Data C). The ATP/ADP ratio was calculated using the following formula:ATP/ ADP ratio=Data A/(Data C−Data B)

Glucose uptake assay

Glucose uptake was measured using 2-NBDG Glucose Uptake Assay Kit (BioVision) according to the manufacturer’s instructions. Briefly, MDMs (10⁵ cells) were seeded in a 24-well plate with 400 μl detection reagent (376 μl 0.5% serum 1,640 cell culture medium, 4 μl 2-NBDG reagent, and 20 μl glucose uptake enhancer) at 37°C for 30 min. Subsequently, cells were washed with iced analysis buffer three times and scraped off with rubber policeman. Fluorescence intensity was examined by a Beckman flow cytometer (excitation at 488 nm).

Extracellular oxygen consumption and acidification test

Changes in oxygen consumption and acidification after HIV-1 infection were analyzed using assays from Abcam (#ab197243 and #ab197244). Macrophages were seeded in a 96-well plate (black wall) at a density of 10⁵ cells/well. The cell culture medium was replaced with an extracellular oxygen consumption reagent or acidification assay reagent, and fluorescence was monitored using the time-resolved fluorescence (Synergy H1; BioTek). Results were normalized to the total DNA content, which was assessed with Hoechst dye (R&D).

MMP detection

MMP was measured by JC-1 (Beyotime). MDMs were seeded in a 48-well plate. JC-1 dye was prepared according to the instructions and added to the plate for incubation at 37°C for 20 min. The cells were then washed with iced washing buffer. Images were acquired using a fluorescence microscope, and image processing was performed using ImageJ software.

Immunofluorescence staining

MDMs were prepared for immunofluorescence staining. 4% formaldehyde and 100% cold methanol were used for fixation and permeation, respectively. Blocking was performed with blocking buffer at RT for 60 min. Cells were then incubated with rabbit anti-HIF-1α (1:500) or anti-β-actin (1:5,000) overnight at 4°C. The immune complexes were glowed using Alexa Fluor 488 goat anti-rabbit (1:1,000) or Alexa Fluor 594 goat anti-mouse (1:1,000) secondary antibodies before counterstaining with DAPI (Solarbio). Images were obtained using a fluorescence microscope (EVOS FL Auto2; Invitrogen), and image processing was performed using ImageJ software.

Statistical analysis

All experiments were repeated at least three times. Statistical analysis was performed using a t test with SPSS v25.0 and GraphPad Prism v8.0. Data were presented as the mean ± SD, and P < 0.05 was considered as the significance of differences between the two groups.