Elevated levels of circulating tumor necrosis factor (TNF)-α and interleukin (IL)-6 have been detected in human immunodeficiency virus (HIV) type 1 infection. The overproduction of these cytokines could contribute to AIDS pathogenesis. Thus, the expression of TNF-α and IL-6 in human macrophages infected with HIV-1 was investigated. HIV-1 infection, per se, did not induce any TNF-α or IL-6 production or cytokine-specific mRNA expression. In contrast, HIV-1 primed macrophages to a prolonged TNF-α and IL-6 response to lipopolysaccharide (LPS) stimulation with respect to uninfected cells. Time-course analysis and flow cytometry demonstrated that cytokine production stopped at 6 h in uninfected macrophages but continued up to 24 h in HIV-1-infected cells. RNA studies suggested that HIV-1 interfered with late steps of cytokine synthesis. No modulation of membrane CD14 was found to account for the enhanced response to LPS. Finally, the effect of HIV-1 on cytokine response could not be abolished by the antiviral compound U75875.
The infection and destruction of T lymphocytes has been considered the striking manifestation of human immunodeficiency virus (HIV) type 1 infection [1, 2]. There is a growing body of evidence, however, that overproduction of certain cytokines plays a central role in the pathogenesis of the infection. In addition, HIV-infected patients have enhanced production of tumor necrosis factor (TNF)-α and interleukin (IL)-6 [3, 4]. Both of these cytokines can cause fever, contribute to a catabolic state, and, more importantly, induce HIV-1 expression in infected cells by acting at the transcriptional or posttranscriptional level [5, 6]. IL-6 production may contribute to increased serum immunoglobulins in HIV-infected patients, and enhanced serum levels of IL-6 have been associated with the subsequent development of B cell lymphomas .
In previous studies, HIV-1 was found to trigger the release of TNF-α and IL-6 from monocytes and macrophages, a major target for HIV-1 in vivo [8–11], and the possibility has been raised that this is a cause of the excess production of these cytokines in HIV-infected patients [12–16]. It is possible that strategies to reduce the production of these cytokines may be of benefit to patients with HIV infection. The present study was undertaken to investigate the mechanism by which HIV-1 modulates TNF-α and IL-6 release in macrophages and the ability of the antiviral compound U75875, a strong inhibitor of the HIV-1 protease, to normalize the production of cytokine in HIV-1-infected macrophages.
Materials and Methods
We used recombinant macrophage colony-stimulating factor (M-CSF; Genetic Institute, Cambridge, MA). The product concentration was 0.78 mg/mL, and the specific activity was 1.9 × 106 U/mg of protein (1 U equals half maximal stimulation in the murine bone marrow colony assay). Lipopolysaccharide (LPS) from Escherichia coli 0111/B4, Brefeldin A, saponin, and bovine serum albumin were from Sigma (St. Louis). U75875, a synthetic peptidomimetic inhibitor of the HIV-1 protease , was supplied by Upjohn Laboratories (Kalamazoo, MI); U75875 is for laboratory use only and is not to be administered to humans or food-producing animals or plants. Dynabeads M-450 CD 14 were obtained from Dynal (Oslo).
An HIV laboratory strain (Ba-L) and 3 HIV-1 clinical isolates were used to infect macrophages. Supernatants of infected macrophages were the source of HIV-1Ba-L. These were filtered and stored in liquid nitrogen before use. The HIV-1 clinical isolates were obtained from 3 HIV-seropositive persons. The strains were isolated from the plasma of the infected persons in peripheral blood mononuclear cell (PBMC) cultures, and the supernatants of the cultures were the source of the virus. Titration to determine infectivity of HIV-1Ba-L and primary isolates was done, respectively, in a primary macrophage system or in PBMC as previously described [18, 19]. The titer of the virus stocks, expressed as TCID50, was determined as previously described .
HIV p24 antigen production in supernatants was assessed by a sandwich ELISA (Abbott, Pomezia, Italy).
Peripheral blood from HIV-negative donors was enriched for PBMC by centrifugation over ficoll-hypaque. PBMC were further enriched for monocytes by elutriation as previously described . Cells obtained by this method are >90% monocytes as determined by flow cytometry (FACS) analysis. The cells were cultured in RPMI 1640 medium supplemented with 20% heat-inactivated fetal calf serum, 2 mM L-glutamine, 50 U/mL penicillin, 50 μg/mL streptomycin, and 1000 U/mL M-CSF, referred to as “complete medium.” M-CSF is a cytokine that regulates monocyte-macrophage differentiation and function in vitro and in vivo . We used the complete medium at 37°C in a humidified atmosphere of 5% CO2 in air in 48-well plates (Costar, Cambridge, MA) at a concentration of 5 × 105 cells/well/mL. Fetal calf serum was used instead of autologous serum to avoid uncontrolled stimulation of macrophages by cytokines and growth factors and interference with the ELISA due to the possible presence of TNF-α and IL-6. Each experiment used PBMC from a single donor. Blood donors had not been vaccinated with bacille Calmette-GÚerin.
Limulus amebocyte lysate test
All compounds and media used in this study were analyzed for endotoxin contamination by limulus amebocyte lysate test (QCL-1000; BioWhittaker, Walkersville, MD), and all were found to be free of endotoxin contamination (<0.1 EU/mL).
Infection of macrophages with HIV-1
Freshly elutriated monocytes were allowed to mature into macrophages for 7 days and then infected for 2 h with 300 TCID50 of the different HIV-1 isolates: 300 TCID50 is a virus load that enables most cells to be infected during the first cycle of infection (not shown). Further increase of the infectious dose did not result in any significant increase in the number of infected cells for any of the virus strains used. In contrast, a suboptimal infection could be seen by reducing the TCID50 from 300 to values 20%–50% lower, depending on the virus strain. Appropriate mock-infected cultures were run as controls. After infection, the cells were extensively washed to remove excess virus and cultured in complete medium. They were washed and fed every 3–4 days.
Assessment of IL-6 and TNF-α production by infected macrophages
In some experiments, at days 1, 3, 7, 10, and 14 after infection, the cultures were washed and refed with complete medium that may or may not have contained 100 ng/mL LPS. After 24 h of incubation, 500 μL of the supernatant from each sample was collected and stored at –80°C.
In other experiments, the cultures were kept until day 14 after infection and then stimulated with up to 0.1 μg/mL LPS. After 1.5 h, the supernatants were harvested and fresh complete medium was placed in the wells; this operation was repeated at 3, 4.5, 6, and 24 h. Cell counts by trypan blue exclusion at the end of each experiment did not reveal any cell detachment due to the wash procedure. The data represent cytokine production from each single monolayer throughout the 24-h observation period. All supernatants were stored at −80°C.
Commercially available sandwich ELISA kits (R&D Systems, Minneapolis) were used to determine the concentration of TNF-α and IL-6 in the supernatants. The detection limits of these ELISAs are 15.6 pg/mL (TNF-α) and 3.13 pg/mL (IL-6). According to the manufacturer's specifications, the ELISAs are specific for the relative cytokine. All samples were tested in duplicate in a single analytical set. Intraseries variation coefficient was <15%.
FACS analysis was done by use of a previously described method . Macrophages were cultured in 25-cm2 flasks and infected as described above. At day 14, the cultures were washed and refed with complete medium containing 0.1 μg/mL LPS. We added 2.5 μg/mL of protein transport inhibitor brefeldin A 30 min or 6 h later. At 3 h after brefeldin A was added, the cells were washed twice in PBS, detached by gentle scraping, collected by centrifugation, and stained for 15 min with R-phycoery-thrin-cyanin 5-conjugated anti-CD 14 monoclonal antibody (Immunotech, Marseille, France) for the determination of their surface phenotype. The cells were then washed twice in PBS and fixed in ice-cold PBS containing 4% paraformaldehyde. After 2 further washes in PBS, 2 × 105 cells were resuspended for 30 min at room temperature in 30 μL of PBS containing 0.1% saponin, 1% bovine serum albumin, and 0.5 μg/106 cells of the following monoclonal antibodies: phycoerythrin-conjugated mouse anti-HIV p24 antigen (Immunotech) plus either fluorescein isothiocyanate (FITC)-conjugated mouse anti-human TNF-α (PharMingen, San Diego) or FITC-conjugated mouse anti-IL-6 (PharMingen). Paired iso-type-specific control antibodies (PharMingen) were run with each sample. As a last step, the cells were washed twice in PBS containing 0.01% saponin, resuspended in PBS, and analyzed by FACScan flow cytometer (Becton Dickinson, Mountain View, CA). Five thousand cells were computed in list mode and analyzed by FACScan research software (Becton Dickinson). Macrophages were differentiated from lymphocytes and dead cells on the basis of forward angle and 90° scatter.
RNA polymerase chain reaction (PCR) of cytokine mRNA
The RNA PCR was performed as described elsewhere  with the following modifications. For the reverse-transcriptase reaction, 0.5 μg of total cellular RNA, extracted by the method of Chirgwin et al. , was mixed with 1 μg of oligo dT (12–18 oligomer; Pharmacia LKB Biotechnology, Uppsala, Sweden) and incubated for 10 min at 65°C. After being cooled on ice, the mixture was incubated in reverse-transcription buffer  containing 1 mM de-oxynucleoside 5′ triphosphates, and in the presence of 200 U of Moloney murine leukemia virus reverse transcriptase, for 1 h at 37°C. The cDNA obtained was amplified by use of 0.1 μg of primers specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), TNF-α, and IL-6 in a reaction mixture containing 200 μM dNTP and 0.1 U of Taq polymerase in PCR buffer . The sequences of TNF-α, IL-6, and GAPDH primers have been described [24, 26]. PCR was performed in a DNA thermal cycler (Cetus, Norwalk, CT) for 25 cycles of 40 s at 94°C, 40 s at 62°C, and 60 s at 72°C. The reaction products were analyzed as previously described .
We used Student's t test to analyze data.
Expression of surf ace antigens on elutriated macrophages cultured in the presence of M-CSF
The expression of surface antigens on elutriated macrophages was analyzed at different time points (table 1). Most of the fresh elutriated monocytes expressed CD14, CD11b, and HLA-DR. The expression of CD4 was quite variable in fresh elutriated macrophages (36%–57%) depending on the donor. Consistent with previous results , we found that the expression of CD4 increased somewhat in mature (7- and 14-day-old) macrophages cultured in the presence of M-CSF. In contrast, the expression of CD 14, CD1 1b, and HLA-DR did not substantially change during the culture period.
HIV-1 infection of macrophages
Elutriated macrophages were precultured for 7 days and then exposed to laboratory strain Ba-L and 3 HIV-1 clinical isolates (GS11, MD8, MD31). As shown in table 2, virus production (HIV p24 antigen production in the supernatants) was first observed at day 3 after infection and then increased steadily. At day 14 after infection, HIV p24 antigen levels ranged from 36.3 ng/mL in MD31-infected macrophages to 87.8 in HIV-1Ba-L-infected cells. These levels of HIV p24 antigen correlated with a viral infectivity titer of 300–3000 TCID50. FACS analysis showed that 14 days after infection, most of the macrophages in the cultures were actively producing HIV p24 antigen (table 2). Virus-induced cytopathic effect, characterized by multinucleated giant cell formation (>5 nuclei/cell), was readily apparent 7–14 days after infection in all experiments (data not shown).
HIV-1 does not modulate TNF-α and IL-6 production by un-stimulated macrophages
We analyzed the production of TNF-α and IL-6 in elutriated macrophages that were allowed to mature for 7 days and then exposed to HIV-1. In the absence of LPS stimulation, no TNF-α or IL-6 production could be detected by a specific ELISA in either control or HIV-1-infected macrophages 14 days after infection (all studies with HIV-1Ba-L macrophages: <15.6 [TNF-α] and <3.12 [IL-6] pg/mL at days 1, 3, 7, 10, and 14; similar results were obtained in cultures infected with HIV-1 GS11, MG8, and MD31). In certain experiments, the cultures were extended up to 35 days after infection, and even at this late time point, no TNF-α or IL-6 production could be detected (data not shown). In agreement with these data, several investigators have shown that TNF-α and IL-6 production is an inducible and not a constitutive event in macrophage cultures that exclude endotoxin contamination [28–30].
Productive HIV-1 infection primes macrophages for an enhanced TNF-α and IL-6 response to LPS
We analyzed the ability of HIV-1 to modulate the response of macrophages to LPS. Figure 1 shows TNF-α and IL-6 production by LPS-stimulated macrophages following exposure to 4 different HIV-1 strains. No significant induction of cytokine production was found 3 days after viral infection. However, starting at day 7, higher concentrations of both TNF-α and IL-6 were detected in infected cultures, which further increased at days 10 and 14. To analyze the effect of a suboptimal HIV-1 infection on cytokine production, we generated an infection curve by use of different TCID50 of HIV-1 to infect macrophages. As shown in table 3, there was a direct correlation between cytokine production and HIV-1 p24 antigen production and percentage of infected cells. These data indicate that the modulation of cytokine production in infected macrophages was strictly related to the development of a productive infection.
Because loss of cells in uninfected cells could have accounted for the differences in cytokine levels in the supernatants, mono-layer cell viability was evaluated by trypan blue dye exclusion and by a spectrophotometric method based on the reduction of the yellow-colored MTT by mitochondrial dehydrogenase of metabolically active cells to a blue formazan (MTT assay). This method has been used to measure the number of viable cells in macrophage cultures infected with HIV-1 [30–32]. Both trypan blue exclusion and the MTT assay showed that the number of viable cells did not differ between infected and uninfected cultures 1 and 3 days after infection. Then a cytopathic effect of the viral preparation took place, and the number of viable cells at days 7, 10, and 14 after virus exposure ranged up to, respectively, 0.3–0.5-fold, 0.7–1.3-fold, and 1.5–2.0-fold higher for uninfected macrophages compared with infected cells. No intracellular sequestration of cytokines could be detected in infected and uninfected cultures by ELISA testing of cells lysed by repeated cycles of freezing and thawing (a procedure that does not interfere with ELISA accuracy or sensitivity; data not shown).
Time course analysis
To further analyze the effect of HIV infection on the LPS-induced cytokine production, in certain experiments the cultures were run until day 14 after infection, stimulated by LPS, and analyzed for cytokine production at different time intervals for the following 24 h (figure 2). The production of both TNF-α and IL-6 was statistically not different between infected and control macrophages during the first 3 h (P ⩽.1) after LPS stimulation. Then, in the uninfected macrophages, cytokine production declined to very low levels at 3–4.5 h and was undetectable thereafter. In contrast, cytokine production was readily detectable in the infected cultures up to 24 h after LPS stimulation. Thus, these data indicate that the increase of cytokine production in HIV-1-infected cultures is due to a more protracted cytokine response to LPS stimulation than in uninfected cells.
Control and infected macrophages were anaitaliclyzed by flow cytometry for virus and cytokine intracellular production (figure 3). The percentage of cells expressing either TNF-α or IL-6 was similar in control and HIV-1-infected cultures 3 h after LPS stimulation: 15% ± 4% versus 14% ± 4%, P ⩽ .1, for TNF-α (figure 3 left, A and B) and 10% ± 2% versus 11% ± 4%, P ⩽ .1, for IL-6 (figure 3 left, C and D). After 6 h, no TNF-α or IL-6 production could be detected in uninfected macrophages, whereas significant intracellular accumulation of both TNF-α (16% ± 5%) and IL-6 (13% ± 3%) was evident in infected cells (figure 3 right). Similar results were obtained at 7.5 and 10.5 h (data not shown). These data strictly parallel those obtained by ELISA.
Expression of TNF-α and IL-6 mRNA in normal and HIV-1-infected macrophages
TNF-α- and IL-6-specific mRNAs were assayed in macrophages following HIV-1 infection by reverse transcription and PCR amplification (figure 4). In the absence of LPS stimulation, no amplification products were present in either infected or uninfected macrophages. Three hours after LPS stimulation, cytokine mRNA accumulation was observed in both uninfected and HIV-1-infected macrophages. This is consistent with data from ELISA testing and FACS analysis showing substantial TNF-α and IL-6 secretion at this time point. Similarly, cytokine mRNA was readily detectable in both uninfected and HIV-1-infected macrophages 6 h after LPS stimulation. This finding, in view of the absence of any detectable cytokine production at this time in uninfected macrophages, strongly suggests that a block of cytokine mRNA translation took place in uninfected cells, which was abolished in HIV-1-infected cells. GAPDH mRNA amplification confirmed the presence of cellular RNA and the efficiency of PCR amplification for all samples.
Membrane CD14 expression on uninfected and infected macrophages
We next investigated whether the difference in cytokine production between uninfected and infected cells was due to the ability of HIV-1 to modulate the expression of CD14, the cellular receptor for LPS [33, 34]. Since, under some experimental conditions, LPS may induce cytokine production in macrophages via CD 14-independent pathways , we studied the role of the CD 14 receptor in mediating LPS activity in uninfected and infected macrophages. Control and HIV-1-infected macrophages were depleted of CD14-positive cells by immunomagnetic beads and then stimulated with increasing concentrations of LPS for TNF-α and IL-6 production. As shown in table 4, only a very low cytokine production, possibly due to some residual CD14-positive cells, was obtained in control and HIV-1-infected CD14-negative macrophages with concentrations of LPS up to 10 μg/mL. Thus, the macrophage interaction with LPS at the concentrations used in our experimental system is largely CD14-dependent.
We analyzed the expression of the CD 14 receptor by flow cytometry at day 14 after infection. As shown in figure 5, no significant differences were observed between uninfected and infected cells. The percentage of cells expressing the CD14 receptor was 77.8% ± 11% versus 71% ± 9%, respectively, in uninfected and HIV-1-infected macrophages.
Effect of the HIV-1 protease inhibitor U75875 on HIV-1-induced modulation of cytokine production
Finally, we tested the ability of the antiviral compound U75875, a potent inhibitor of HIV-1 protease that suppresses HIV-1 replication in chronically infected macrophages , to affect cytokine production by HIV-1-infected macrophages. As shown in table 5, >99% inhibition of mature HIV-1 p24 antigen production and complete reduction of infectious virus production was achieved with 5 μg/mL U75875. However, no significant modulation of TNF-α and IL-6 secretion was observed in U75875-treated macrophages with respect to untreated cells.
There is considerable controversy over the ability of HIV-1 to induce cytokine production in macrophages. Some studies describe the induction of TNF-α and IL-6 during in vitro HIV-1 infection [12–16]; others suggest that HIV-1 affects cytokine production only when infected cells are activated with certain stimuli [28, 36–39]. Our data demonstrate that in the absence of activation by LPS, HIV-1 does not alter TNF-α and IL-6 production in human macrophages. The seemingly conflicting data on the ability of HIV-1 to modulate cytokine production in macrophages are probably due to technical differences in the methods used to isolate and cultivate these cells. Macrophages used in most previous studies were purified by adherence to plastic and matured in the absence of any growth factor. In contrast, in the present research, monocytes were obtained by elutriation and allowed to mature into macrophages in the presence of M-CSF. It is possible that these cells may have different patterns of cytokine production in response to stimuli such as HIV-1 infection. Elutriation is a gentle process that preserves normal cell viability and function  and gives rise to a highly purified monocyte population suitable for studies, such as those on cytokine induction where cell purity and unaltered cell function are critical. Moreover, M-CSF is a hematopoietic growth factor that plays an important role in macrophage homeostasis in vivo. It induces maturation of monocytes into macrophages and potentiates macrophage activity against bacteria and fungi and macrophage cytotoxicity against certain tumor cell targets [22, 40–44]. Also, one study demonstrated that M-CSF may prime macrophages to an increased in vitro TNF-α and IL-6 response to LPS stimulation . Finally, bioassays on blood have shown that endogenous M-CSF levels are very similar to those used here [46, 47], and, more importantly, that HIV-infected persons have increased levels of M-CSF .
Alternative explanations for the discrepancy between our results and those reported by others might be a different ability of the virus isolates used to induce cytokine production or a reduced efficiency to infect macrophages. This is improbable, however, since we used 1 laboratory-adapted strain and 3 different primary clinical isolates to infect macrophages, and all proved unable, per se, to induce any cytokine response, although they all gave rise to a highly productive infection.
We found that HIV-1 infection substantially enhanced TNF-α and IL-6 production following LPS stimulation. Several reports have suggested that certain HIV-1 proteins (e.g., gp120 and tat) can increase the gene expression of cytokines such as TNF-α, IL-10, IL-6, and IL-2 [49–51]. Our findings suggest that transcriptional modulation was not the leading mechanism by which HIV-1 affected cytokine production. However, since a “nonquantitative” PCR amplification technique was used, we cannot exclude that some modulation of cytokine mRNA might occur in HIV-1-infected macrophages. Rather, since we correlated cytokine secretion with mRNA expression in uninfected and infected macrophages, it is possible that, although cytokine mRNA was substantially expressed in uninfected macrophages 6 h after LPS stimulation, these cells did not synthesize any detectable amounts of TNF-α and IL-6 proteins. In contrast, at the same time point, the expression of cytokine mRNA in HIV-1-infected macrophages was associated with a substantial secretion of TNF-α and IL-6 proteins. This strongly suggests that translational repression of the cytokine mRNA took place in uninfected macrophages and was abolished in HIV-1-infected cells. Production of TNF-α and IL-6 by macrophages requires two signals: The first induces transcription and the second releases the translational repression of mRNA [52–54].
We hypothesize that the prolonged cytokine synthesis observed in infected macrophages is mediated via release of the translational repression of cytokine mRNA. Consistent with our findings, D'Addario et al.  demonstrated that HIV-1 infection of myelomonoblastic cells may alter translational mechanisms controlling cytokine expression. Thus, these results suggest that HIV-1 infection of macrophages may prime or sensitize cells such that subsequent endotoxin challenge leads to enhancement of cytokine production. It is possible that one or more viral proteins interfered with this process. It is of interest that the viral protein tat alters cytokine expression at the translational level . HIV-1 infection may indirectly affect TNF-α and IL-6 by promoting the secretion of a soluble factor able to increase cytokine mRNA translation. In this regard, interferon (IFN-γ) reportedly increases TNF-α mRNA translation in mouse macrophages . However, we did not find any IFN-γ production when we tested supernatants from both uninfected and infected macrophage (data not shown). Finally, it is possible that the preculture period with M-CSF affected cytokine production by potentiating the effect of HIV-1 infection on macrophages. This hypothesis is consistent with a recent report from our group that showed that M-CSF increases the susceptibility of macrophages to infection with HIV-1 .
Macrophages are a major target for HIV-1 infection in the body, as shown by infection of macrophages in spleen and lymph nodes, alveolar macrophages of the lung, and macrophage-like cells in the central nervous system [8–11]. Compelling evidence is accumulating that systemic release of TNF-α and IL-6 by HIV-infected macrophages may contribute to the maintenance of the inflammatory response, retroviral spreading, increased serum immunoglobulins, and development of B cell lymphomas [3–7]. Moreover, local release of TNF-α may cause tissue damage. Levels of TNF-α are increased in the brain during HIV-1-associated dementia and in the spinal cord during vacuolar myelopathy and may play a pathogenetic role in these diseases by inducing lesions of the deep white matter [57, 58]. In addition, production of proinflammatory cytokines by HIV-1-infected alveolar macrophages significantly contributes to the decline of pulmonary function in patients with AIDS by altering the lung endothelium, causing edema and interstitial damage [59, 60].
Our findings show that HIV-1 primes macrophages for an enhanced TNF-α and IL-6 response to LPS. Throughout the course of AIDS, gram-negative bacterial sepsis (and endotoxin production) is a major clinical problem that usually manifests itself as either pneumonia, bacteremia, or both at a frequency of 8–20 per 100 person-years, depending upon location, risk activity, and other factors [61, 62]. A recent study showed that the integrity of the large bowel wall in AIDS patients is compromised in a manner that favors the chronic translocation of bacteria or products of bacterial metabolism into the bloodstream . Thus, LPS is a likely stimulus in patients with HIV-1 infection. Finally, it is possible that HIV-1 may prime macrophages for an enhanced cytokine production in response to stimuli other than LPS that can induce TNF-α and IL-6 synthesis, such as gram-positive cell walls .
In conclusion, our findings suggest that HIV-1-infected macrophages can be an important source of proinflammatory cytokines in infected patients. More importantly, our data indicate that the HIV-1-induced enhancement of cytokine response to LPS cannot be abolished in infected macrophages by such compounds as the viral protease inhibitor U75875 at concentrations that can completely suppress viral replication in these cells. Thus, major efforts should be made to identify and eliminate possible sources of endotoxins in subjects infected with HIV-1.
We thank Upjohn Research Laboratories for providing U-75875, and M. Andreoni, Dept. of Public Health and Cellular Biology, University of Rome “Tor Vergata,” for the HIV-1 clinical isolates.