Macrophage and Viral Infection © 1991 S. Karger AG. Basel i0i5-2008/9i/0594-0209$2.75/0

Macrophage: 1990 Annu Conf Upper Rhine Univ (II) Pathobiology 1991;59:209-213

Regulation of Cytokine and Viral Gene Expression in Monocytes Infected with the Human Immunodeficiency Virus Monte S. Meltzera, Lisa Baca3, Jim A. Turpina, D. Chester Kaltera-b, Carl Dieffenbachc, Robert M. Friedmanc, HowardE. Gendelman3b “Immunopathogenesis Program, Department of Cellular Immunology, Walter Reed Army Institute of Research, Washington, D.C.; b Henry M. Jackson Foundation for the Advancement of Military Medicine, Rockville, Md.; 'Department of Pathology, Uniformed Services University of the Health Sciences, Bethesda, Md., USA

Key Words. Monocytes • Human immunodeficiency virus • Interferon • Cytokines

There is well-established precedent for interferons (IFN) as dominant regulatory molecules in host defence reactions against such retrovial infections as murine and avian leukemia and several lentiviral diseases. For exam­ ple, in visna-maedi infection of sheep, virus replication in macrophages is reduced > 1,000-fold by IFN released from T cells [1,2]. Levels of IFN released by ovine lung leukocytes directly correlate with lentivirus infection of alveolar macrophages [3]. The role of IFN in human retroviral disease is more complex. In human immuno­ deficiency virus (HIV) infections of man, IFN activity is paradoxically found in sera of patients with late-stage HIV disease, and is an index of poor prognosis [4, 5]. But

strong antiviral activity is reported with addition of IFNa, IFN-ß and IFN-y to HIV-infected T cells and macro­ phages [for review, see ref. 6]. The time course for appearance and quantity of p24 antigen in culture fluids of phytohemagglutinin/interleukin-2 (IL-2)-treated lymphoblasts exposed to 500 IU/ml recombinant human IFN-a at the time of HIV infection and continuously thereafter were indistinguishable from those of infected control cells through 2 weeks [7], Anal­ ysis of control and IFN-treated HIV-infected lympho­ blasts by in situ hybridization for HIV-specific mRNA at 10 days showed no differences: about 20% of total cells expressed HIV mRNA in each instance. In contrast, lev­

Downloaded by: Univ. of California Santa Barbara 128.111.121.54 - 4/4/2018 7:59:01 PM

Abstract. Monocytes treated with interferon-a (IFN-a) at virus challenge show no evidence of human immuno­ deficiency virus (HIV) infection: no p24 antigen or reverse transcriptase (RT) activity, no viral mRNA and no proviral DNA. Levels of p24 antigen and RT activity in monocytes infected with HIV 1-3 weeks before IFN-a treatment gradually decrease to baseline. HIV-induced cytopathic changes are markedly reduced, as are levels of HIV mRNA: the frequency of productively infected cells is < 1%. But, levels of proviral DNA in the IFN-a-treated and control HIV-infected cells are indistinguishable, and remain so through 3 weeks. Large quantities of proviral DNA in IFN-a-treated cells with little active transcription suggest true microbiological latency. The major potential source for IFN-a in HIV-infected patients is the macrophage. With any of 15 virus isolates, tumor necrosis factor-a, interleukin-1p, interleukin-6, IFN-a) or IFN-p are not detected nor the mRNA expressed in HIV-infected or unin­ fected monocytes. Both uninfected and HIV-infected monocytes produce high levels of these cytokines after treat­ ment with synthetic double-stranded RNA (poly-I:C). Uninfected monocytes also produce high levels of IFN-a after treatment with Poly-I:C, Newcastle disease virus or herpes simplex virus. In marked contrast, HIV-infected mono­ cytes express no IFN-a activity or mRNA before or after treatment with any of these agents. The markedly dimin­ ished capacity of HIV-infected monocyte to produce IFN-a reflects a specific transcriptional block and may be an adaptive mechanism of virus to alter basic microbicidal functions of this cell. The inevitable result of this HIVinduced cytokine dysregulation is virus replication and persistence in mononuclear phagocytes.

els of virion-associated reverse transcriptase (RT) activ­ ity in culture fluids from these same IFN-treated cells were reduced 2- to 3-fold from those of infected control cells. Reduced levels of RT activity were matched by parallel reductions in infectious titer. In a series of 8 separate experiments with 3 different HIV isolates, max­ imal inhibition of RT activity in culture fluids of IFNtreated HIV-infected lymphoblasts was 50-70% that of infected control cells at 500-10,000 IU/ml IFN. No amount of IFN could prevent HIV infection or com­ pletely restrict virus replication. Coincident with this partial reduction in RT activity levels and infectious titer in the culture fluids of IFN-treated HIV-infected T lymphoblasts, RT activity levels in cell lysates from the identical cultures were increased about 2-fold. Thus, the major effect of IFN on the replication of HIV in T lym­ phoblasts is directed at virus assembly or release: the frequency of cells productively infected with virus and the levels of viral proteins produced by infected cells remain unchanged (or increased) by IFN treatment; the number of virions released from these infected cells, however, is reduced about 2-fold. In contrast to the relatively modest effect of IFN on the replication of HIV in T lymphoblasts, the effect of this cytokine on monocytes was quite dramatic. Simulta­ neous addition of 50-500 IU/ml recombinant human IFN-a with the HIV inoculum and continuous exposure of monocytes to IFN thereafter completely abrogated productive virus infection. No p24 antigen or RT activ­ ity was detected in culture fluids of IFN-treated infected monocytes at any time after infection. Analysis of IFNtreated HIV-infected monocytes by in situ hybridization for HIV-specific mRNA showed no evidence of infection 2 weeks after virus challenge. Twenty to 30% of control HIV-infected monocytes expressed HIV mRNA by this analysis at the same time point. Polymerase chain reac­ tion (PCR) amplification of DNA in cell lysates of monocyte cultures treated with 500 IU/ml IFN with HIV long terminal repeat (LTR) and gag primers showed no proviral DNA. Thus, monocytes treated with IFN at the time of virus challenge show no evidence of HIV infec­ tion: no viral protein, no viral mRNA and no proviral DNA. IFN interrupts one or more early event(s) in the virus replication cycle before the formation of proviral DNA: binding, uptake, uncoating or reverse transcrip­ tion. The exact mechanism(s) for this antiviral activity are not yet known but may include: (1) changes in virus receptor number or distribution (IFN-treated monocytes show marked changes in CD4, Fc receptor, CD1 la and mannosylated protein receptors, each of which is impli­

Meltzer/Baca/Turpin/Kalter/Dieffenbach/Friedman/Gendelman

cated in the uptake of HIV into monocytes); (2) changes in the monocyte plasma membrane that interrupt fusion or uptake of the virion into the cell (IFN-treated cells show alterations in membrane fluidity, microfilament organization and membrane proteases that could dam­ age bound virus), and (3) changes in subcellular com­ partments or cytosolic milieu that preclude reverse tran­ scription [IFN-treated cells synthesize 2',5'-(A)oligonucleotides that induce RNases, and directly inhibit re­ verse transcription]. Addition of IFN to monocytes previously infected with HIV for 7 days induced a dramatic decrease in both p24 antigen and RT activity. Quantitation of the infec­ tious virus released into culture fluids 2 days after IFN treatment showed a 1,000-fold decrease in infectious ti­ ter, but only a 3-fold decrease in RT activity at the same time point. Such discordance suggests that most progeny virions produced after IFN treatment are defective. However, analysis of these virions by transmission elec­ tron microscopy and by radioimmunoprécipitation with HIV-positive sera failed to document morphologically aberrant viral particles or abnormal proteins. Addition­ ally, patterns of virus budding at the monocyte plasma membrane or within intracytoplasmic vacuoles were no different from those of control infected cultures. Ten to 14 days after IFN treatment, no p24 antigen or RT activ­ ity was detected in culture fluids or cell lysates of HIVinfected monocyte cultures. IFN treatment also induced a dramatic reversal of the cytopathic changes associated with HIV infection. Multinucleated giant cells were evi­ dent 7 days after infection with a frequency of about 20% of total cells. Addition of 500 IU/ml IFN to these cultures at 7 days was associated with a gradual decrease in the frequency of multinucleated giant cells and ab­ sence of cell lysis through 3 weeks. In contrast, the fre­ quency of multinucleated giant cells in control infected cultured to 30-50% of total cells and cell lysis (decreased cellular density and increased debris) was common. In the absence of detectable levels of viral proteins (p24 antigen or RT activity), or of morphological evi­ dence for virus infection, proviral DNA was still de­ tected 2 and 3 weeks after virus challenge in monocyte cultures treated with 500 IU/ml IFN 7 days after infec­ tion. PCR amplification of DNA in cell lysates with LTR and gag primers showed levels of proviral DNA in the IFN-treated infected monocytes equivalent to that of control infected cells. Southern blot analysis of DNA from IFN-treated and control HIV-infected cells showed similar results. Levels of HIV-specific mRNA detected by Northern blot analysis in the IFN-treated HIV-

Downloaded by: Univ. of California Santa Barbara 128.111.121.54 - 4/4/2018 7:59:01 PM

210

infected monocytes were reduced by > 95% of those of control infected cells. Analysis of these same cell popula­ tions by in situ hybridization for HIV-specific mRNA showed productive infection in 244 of 1,164 untreated HIV-infected cells (a frequency of 21 ± 3 % infected cells in triplicate samples) versus 9 of 861 IFN-treated HIVinfected monocytes (a frequency of 1 ± 1% infected cells). The presence of large quantities of proviral DNA in cells with little or no evidence for active transcription suggests true microbiological latency... and this in a non­ replicating cell with no direct evidence for integrated virus. Such transciptional restriction of virus replication in the IFN-treated HIV-infected monocytes has no precedent in previously described retroviral systems. The preceding data present an apparent paradox. IFN shows potent antiviral activity in macrophage cultures. More impressively, IFN has significant antiviral activity in HIV-infected patients: in a randomized, placebo-controlled trial, asymptomatic HIV-seropositive individuals given IFN daily for < 12 weeks remained free of AIDSassociated opportunistic infections (0 of 16 patients) for up to 3 years; about a third (5 of 16 patients) of the placebo-treated control patients developed such infec­ tions over the same time interval [8]. Yet the major source of IFN during any infection is the macrophage. In HIV-infected patients, virtually all blood monocytes and most tissue macrophages (cells of the brain, lung, lymph nodes and skin expected) are free of virus [9, 10]. If IFTN is effective in HIV-infected patients, and most macro­ phages (the primary source of IFN) are free of virus infection, why is there HIV disease? The answer to this question may derive from an HIV-induced defect in the immune system as significant for progression of disease as the well-known depletion of CD4+ T cells. Monocytes from HIV-infected patients and monocytes from HIVseronegative donors infected in vitro with monocyte tropic HIV fail to produce IFN after any of numerous treatments with agents that induce high titers of IFN in cultures of control cells. No IFN activity (inhibition of murine encephalomyocarditis virus-induced cytopathic effects in human fore­ skin fibroblasts) was detected in culture fluids of unin­ fected or HIV-infected monocytes or T lymphoblasts with any of 15 different HIV isolates at any time through 3 weeks in culture. Similarly, levels of tumor necrosis factor-a (TNF-a), IL-1 (3or IL-6 in these culture fluids are < 1 pg/ml [11], Culture fluids from uninfected control monocytes treated with synthetic double-stranded RNA [polyinosinic acid-polycytodylic acid (poly-I:C)] contain high levels of IFN activity: > 1,000 IU/ml 24 h after

211

treatment. Neutralization studies with type-specific anti­ bodies show that > 99% of antiviral activity is IFN-a,. In striking contrast, poly-I:C-treated monocytes infected with HIV 14 days previously produce little or no IFN activity at any time through 3 days. HIV-associated inhi­ bition of poly-I:C-induced IFn activity in culture fluids of virus-infected monocytes increases with time after infection: < 10% by day 1; 50% by day 3; 80% by day 5; 90% by day 7, and > 99% by day 14. It is important to note that at 14 days, < 50% of monocytes in these cul­ tures are positive for HIV mRNA by in situ hybridiza­ tion. In replicate experiments with 3 different HIV iso­ lates, no IFN activity (< 10 IU/ml) was detected in cul­ ture fluids of HIV-infected monocytes treated with polyI:C, Newcastle disease virus or herpes simplex virus type 1 10-30 days after HIV inoculation. To further define the extent and specificity of the HIV-induced defect in IFN production in HIV-infected monocyte cultures, we examined IFN-a-specific mRNA using coupled reverse transcription/PCR analysis. While IFN-a mRNA was evident in cell lysates of uninfected monocyte cultures 8, 12 and 24 h after poly-I:C treat­ ment, no IFN-a mRNA was found in any cell lysate of HIV-infected monocyte cultures treated with identical conditions and for identical time intervals. Replicate experiments with primer pairs specific for the mRNA transcripts of 3 different members of the IFN-a family of genes (IFN-a 1, IFN-a2, IFN-a 13) showed significant lev­ els of IFN-a mRNA in cell lysates of uninfected polyI:C-treated monocytes, but none in the lysates of simi­ larly treated HIV-infected cells. These results confirm the near absence of IFN activity in culture fluids of polyI:C-treated, HIV-infected, monocyte cultures and localize this HIV-associated defect to a transcriptional block. The specificity of this HIV-associated transcriptional block in expression of IFN-a genes was explored by quantitation of mRNA for several other cytokines in­ duced in monocytes by poly-I:C. Coupled reverse tran­ scription/PCR analysis of mRNA for IL-ip, IL-6 and TNF-a in cell lysates of uninfected and HIV-infected monocyte cultures showed no cytokine mRNA. These results are consistent with the absence of these cytokines as detected by enzyme-linked immunosorbent assay (ELISA) in monocyte culture fluids and with published reports [12-14], After poly-I:C treatment, high and in­ distinguishable levels of mRNA of IL-1 P, IL-6 and TNFa were identified in cell lysates from both uninfected and HIV-infected monocyte cultures. High levels of mRNA were associated with similarly high concentrations of these same cytokines in monocyte culture fluids as de­

Downloaded by: Univ. of California Santa Barbara 128.111.121.54 - 4/4/2018 7:59:01 PM

Cytokine and Viral Gene Expression in HIV-Infected Monocytes

tected by ELISA. Thus, the strong HIV-associated tran­ scription block in the expression of IFN-a genes was not evident in expression of several other cytokine genes. Moreover, while IFN-a is the predominant IFN activity produced by monocytes, it is not the only IFN expressed by these cells. IFn-P and IFN-co are minor species of IFN genetically distinct from the IFN-a family of genes [15]. Both uninfected and HIV-infected monocyte cultures expressed IFN-P and IFN-co mRNA after poly-I:C treat­ ment: there were no differences in the amount or time course for appearance of these mRNA in cell lysates of uninfected or HIV-infected monocytes. These results, in toto, document a profound and highly specific transcrip­ tional block in the expression of IFN-a genes in mono­ cytes infected with HIV. The inability of HIV-infected monocytes to express IFN-a may represent an adaptive response by this virus to ensure its survival and un­ impeded replication in monocytes. The inevitable result of this HIV-induced cytokine dysrégulation is virus re­ plication and persistence in mononuclear phagocytes. In the HIV-infected patient, production of all IFN types is suppressed. Production of IFN-y by peripheralblood mononuclear cells (PBMC) of HIV-infected pa­ tients is significantly reduced [16]. Transcription of the IFN-y gene, but not the IL-2 gene, is impaired in HIVinfected T cells [17], Similarly, monocytes and PMBC from HIV-infected patients show a markedly reduced ability to secrete IFN-a or express IFN-a mRNA after exposure to vesicular stomatitis virus, influenza A virus or herpes virus type 1 infected fibroblasts [18-20]. The mechanism for the transcriptional block in IFN-a gene expression is not yet known. Remarkably, HIV-infected monocyte cultures that exhibited this near absolute IFNa defect showed a frequency for productively infected cells of < 50%. If the IFN-a defect is a direct sequela of infection, then all cells must harbor the HIV provirus which in turn is expressed in only half of the monocytes. Alternatively, the IFN-a defect may be mediated by a soluble factor released by HIV-infected cells that affects the entire cell population. This latter hypothesis is con­ sistent with observations on PBMC of HIV-infected pa­ tients in which a cell population that is > 99% virus-free shows a 1,000-fold decrease in ability to produce IFN-a [ 20], Poly-I:C induces strong antiviral activity in target cells through at least 2 different mechanisms. This syn­ thetic double-stranded RNA induces endogenous IFN production, but it also directly stimulates antiviral activ­ ity in treated cells in the absence of IFN. The first mech­ anism is apparently blocked in HIV-infected monocytes;

Meltzer/Baca/Turpin/Kalter/Dieffenbach/Friedman/Gendelman

however, the alternative direct pathway for antiviral activity may be intact. Analysis of viral gene expression in HIV-infected monocytes after poly-I:C treatment showed marked reduction in levels of HIV mRNA to baseline by 72 h. Reduction in viral mRNA was coinci­ dent with an equivalent reduction in p24 antigen and RT activity levels in the same cultures. It is possible that this strong antiviral activity represents the low levels of IFNP and IFN-co (