In vitro human immunodeficiency virus-1 infection of purified hematopoietic progenitors in single-cell culture.

Uni- or multi-lineage suppression of hematopoiesis is observed in the majority of acquired immunodeficiency syndrome (AIDS) patients. The mechanism(s) underlying these abnormalities is not understood: particularly, the human immunodeficiency virus (HIV) infection of hematopoietic progenitor and stem cells (HPCs/HSCs) is highly controversial. We report that CD34+ HPCs from adult peripheral blood (PB) are in part CD4+ and susceptible to in vitro HIV infection. Primitive CD34+ HPCs were approximately 80% purified from PB. Double labeling for CD34 and CD4 membrane antigens was shown for 5% to 20% of the purified cells, thus suggesting their potential susceptibility to HIV-1 infection. The enriched HPC population, challenged with purified or unpurified HIV-1 strains, was cloned in unicellular methylcellulose culture. The single colonies generated by erythroid burst-forming units (BFU-E), granulocyte-macrophage colony-forming units (CFU-GM), and granulocyte-erythroid-macrophage-megakaryocyte colony-forming units (CFU-GEMM) were analyzed for the presence of HIV, ie, for gag DNA, tat mRNA, and p24 protein by PCR, reverse transcription PCR (RT-PCR), and enzyme-linked immunosorbent assay, respectively. In the first series of experiments incubation of HPCs with HIV-1 at multiplicities of infection (MOI) ranging from 0.01 to 10 TCID50/cell consistently yielded an 11% to 17% infection efficiency of BFU-E-generated colonies, thus indicating the sensitivity of HPCs to in vitro HIV infection. An extensive series of experiments was then performed on HPCs challenged with HIV at 0.1 MOI level. In the initial studies proviral gag sequences were detected in 9.2% of 121 analyzed CFU-GM colonies. In further experiments tat mRNA was monitored in 17% and 23% of BFU-E and CFU-GM colonies, respectively, but never in CFU-GEMM clones. Finally, 12% of CFU-GM clones and rare erythroid bursts were shown to be positive for the p24 viral protein. In control studies, purified HPCs grown in liquid suspension culture were induced to terminal unilineage erythroid, monocytic, or granulocytic differentiation: monocytes were consistently HIV-infected, whereas mature-terminal erythroblasts and granulocytes were not. Our observations indicate that a minority of primitive HPCs, but not of the multipotent type, is susceptible to in vitro HIV infection. These observations may reflect on the in vivo hematopoietic impairment in AIDS patients; more important, they provide an experimental model for studies on HIV hematopoietic infection and in vitro tests for anti-HIV HSC gene therapy.

Tumour necrosis factor-alpha increases the sensitivity of human immunodeficiency virus (HIV)-infected monocytic U937 cells to the complement-dependent cytotoxicity of sera from HIV type 1-infected individuals; role of the gp120 protein.

Sera of 40 intravenous drug addicts [25 seropositive and 15 seronegative for human immunodeficiency virus (HIV)] were tested for the presence of cytotoxic antibodies against uninfected and HIV-infected monocytic U937 cells. Six of the 25 seropositive samples proved to be cytotoxic for HIV-infected target cells in the presence of complement. The pretreatment of HIV-infected U937 cells with tumour necrosis factor (TNF)-alpha (which enhances virus production in these cells) increased the detection of serum cytotoxicity and 60% of these sera became cytotoxic. The percentage lysis was also increased after the TNF-alpha treatment of the target cells (from 16.2 +/- 4.5 to 71.2 +/- 4.9). The complement-dependent cytotoxic activity of these sera was significantly reduced by pretreatment with recombinant HIV gp120 antigen. This reduction was dose-dependent in the majority of cases. Immunofluorescence studies suggested that the cytotoxic sera mainly interacted with the viral antigens localized on the membrane of HIV-infected TNF-treated U937 cells. Moreover, comparative Western blot analyses using cellular extracts from untreated and TNF-treated U937 cells showed that there was a positive correlation between the cytotoxic phenotype and the capacity of sera to recognize the gp120 protein in extracts from TNF-treated HIV-infected cells. These results suggest that in some circumstances endogenous TNF-alpha can be a protective factor because it can render persistently infected cells highly sensitive to complement-dependent serum cytotoxicity as a result of increased expression of the relevant viral antigen (gp120) on the cell membrane.

Monocytotoxic antibodies in HIV-infected persons. TNF-alpha treatment of U937 cells increases the complement dependent cytotoxicity.

Sera of 40 intravenous drug addicts were tested for the presence of cytotoxic antibodies against uninfected and HIV-infected monocytic U937 cells. Twelve out of 31 seropositive samples proved to be cytotoxic for HIV-infected, untreated target cells in the presence of complement. The TNF-alpha treatment of HIV-infected U937 cells increased the detectability of cytotoxic effect of sera (21/31). The complement dependent cytotoxic activity of sera was reduced by pretreatment with recombinant HIV gp120. This reduction proved to be dose-dependent in the majority of cases. Immunofluorescence studies indicated that the cytotoxic sera interacted with antigens mostly localized on the cell membrane of HIV-infected TNF-alpha treated U937 cells. The specificity, the possible role and origin of monocytotoxic antibodies in HIV-infected persons is discussed.

Persistent infection of normal mice with human immunodeficiency virus.

In this article, we report the establishment of persistent HIV type 1 infection of normal Swiss mice after a single intraperitoneal injection with high-producing HIV-infected U937 cells. Anti-HIV antibodies were found more than 500 days after the original injection, and p24 antigenemia was detected in approximately 50% of the mice. By polymerase chain reaction (PCR) techniques, HIV-specific gag and env sequences were detected in DNA samples from peripheral blood mononuclear cells (PBMC) and peritoneal cells of seropositive mice 300 to 500 days after inoculation with HIV-infected cells. These DNA samples did not contain human DNA sequences, as determined by PCR analysis using primers and the probe for the HLA-DQ alpha gene. Low levels of p24 and detectable human reverse transcriptase activity were found in cultures of PBMC and peritoneal macrophages. Cocultivation of PBMC, peritoneal cells, and spleen cells with human uninfected U937 or CEM (a T lymphoma cell line) cells resulted in HIV infection of the target cells, as determined by PCR analysis and/or p24 assays. The intravenous injection of untreated Swiss mice with the PBMC from PCR-positive mice resulted in the development of an increasing antibody response to HIV in the recipient animals. Together these results indicate that cells from seropositive Swiss mice were persistently infected with HIV and were capable of producing infectious virus. The development of persistent HIV infection in an immunocompetent mouse may represent the starting point for further studies aimed at defining the host mechanisms involved in the restriction of virus replication, defining the pathogenesis of HIV infection, and testing antiviral compounds and vaccines.

Human immunodeficiency virus (HIV)-infected tumor xenografts as an in vivo model for antiviral therapy: role of alpha/beta interferon in restriction of tumor growth in nude mice injected with HIV-infected U937 tumor cells.

The host factors involved in the restriction of tumor growth were studied in nude mice transplanted with a cloned line of chronically human immunodeficiency virus (HIV)-infected U937 cells. HIV-infected and uninfected U937 cells exhibited the same growth patterns in culture. However, HIV-infected cells were not tumorigenic when injected subcutaneously in nude mice, whereas large solid tumors were observed in mice injected with uninfected U937 cells. Injection of nude mice with antibody to alpha/beta interferon (IFN-alpha/beta) enabled HIV-infected U937 cells to grow progressively in approximately 90 to 100% of mice. HIV-infected U937 cells formed solid tumors in the majority (60 to 90%) of either immunosuppressed (splenectomized, irradiated, and anti-asialo-GM1-treated) or genetically immunodeficient (bg/nu/xid) nude mice. In mice treated with antibodies to IFN-alpha/beta with established HIV-positive tumors, a direct correlation was found between p24 antigenemia and tumor size. Treatment of established HIV-positive U937 cell tumors with human IFN-alpha or mouse IFN-alpha/beta resulted in a clear-cut inhibition of both tumor growth and p24 HIV antigenemia. In contrast, treatment with tumor necrosis factor alpha markedly inhibited tumor growth but did not significantly decrease serum p24 levels. 3'-Azido-3'-deoxythymidine treatment did not affect either tumor growth or the levels of serum p24 antigen. These data indicate that endogenous IFN-alpha/beta is a crucial factor in the restriction of both tumor growth and p24 antigenemia in mice injected with HIV-infected tumor cells. Moreover, the results suggest that the development of HIV-1 p24 antigenemia in athymic immunosuppressed mice may represent an interesting in vivo model for anti-HIV therapy.

Increased human immunodeficiency virus (HIV) expression in chronically infected U937 cells upon in vitro differentiation by hydroxyvitamin D3: roles of interferon and tumor necrosis factor in regulation of HIV production.

We have investigated the roles of cytokines in the modulation of human immunodeficiency virus (HIV) production in chronically infected U937 cells upon in vitro differentiation by hydroxyvitamin D3. HIV-infected U937 cells exhibited markedly lower levels of CD4 and HLA-DR antigens than uninfected cells did. Vitamin D3 induced a time-dependent macrophagelike differentiation, as determined by monitoring the expression of some surface antigens by means of the monoclonal antibodies OKM1, OKM5, OKM13, OKM14, OKT4, anti-HLA-DR, TecMG2, TecMG3, LeuM3, LeuM1, anti-HLA-DP, and anti-HLA-DQ. Treatment with hydroxyvitamin D3 resulted in a marked increase in HIV production compared with control cultures. Interleukin 1 beta (IL-1 beta) and tumor necrosis factor alpha (TNF-alpha) were detected in the culture media, whereas interferon (IFN) was not generally found. Using the polymerase chain reaction technique, we found HIV-infected U937 cells to express detectable levels of mRNAs for alpha interferon (IFN-alpha), IFN-beta, TNF-alpha, and IL-1 beta. The addition of TNF resulted in a marked increase of HIV production, whereas IL-1 beta was ineffective. In contrast, both IFN-alpha and IFN-beta exerted some inhibitory effect on HIV production, which was more marked in vitamin D3-treated cultures than in untreated cultures. HIV production was significantly increased by antibodies to IFN-alpha in both untreated and vitamin D3-treated cultures. Anti-IFN-beta antibody increased HIV production only in vitamin D3-treated cells. In contrast, anti-TNF-alpha antibodies markedly decreased HIV production in both control and differentiating U937 cells. Vitamin D3 treatment resulted in a higher expression of TNF receptors in differentiating cells than in control HIV-infected cells. These data demonstrate a strong correlation between HIV production and macrophagelike differentiation in chronically infected U937 cells and suggest that endogenous IFN and TNF exert opposite effects in the regulation of virus production in both undifferentiated and vitamin D3-treated cell cultures.

Effects of different biological response modifiers on interferon expression in bacterial lipopolysaccharide (LPS)-responsive and LPS-hyporesponsive mouse peritoneal macrophages.

We have previously shown that the antiviral state of explanted mouse peritoneal macrophages (PM) decays during in vitro culture and that this decay is much more rapid in Lpsd PM than it is in Lpsn PM. Moreover, Lpsn PM can transfer the antiviral state to other cells, whereas Lpsd PM cannot. In vitro treatment of Lpsn PM with different agents [i.e., bacterial lipopolysaccharide (LPS), interferon (IFN)-gamma, tumour necrosis factor (TNF)-alpha, macrophage colony-stimulating factor (M-CSF) and antibody to Mac-1 antigen] induced an antiviral state to vesicular stomatitis virus (VSV) which was inhibited by antibodies to IFN-beta. Treatment of Lpsn PM with LPS or IFN-gamma resulted in greater accumulation of IFN-beta mRNA, whereas no change in the barely detectable levels of IFN-alpha mRNA was observed. Marked accumulation of IFN-beta mRNA was also observed in PM after TNF-alpha treatment. M-CSF and IFN-gamma (but not LPS) also induced an IFN-mediated antiviral state in Lpsd PM. Low levels of spontaneous transcription of IFN-beta mRNA were detected in nuclei from Lpsd PM. Treatment of Lpsd PM with IFN-gamma for 3 h resulted in the accumulation of IFN-beta mRNA without any concomitant increase in the transcription of the IFN-beta gene, as determined by run-on transcription assays with isolated nuclei. The addition of as little as I international unit/ml of IFN-gamma to PM resulted in a 100-fold inhibition of VSV yield. As antibodies to IFN-alpha/beta inhibited only a portion of the IFN-gamma-induced antiviral state, such an antiviral state might reflect the synergism between IFN-gamma and endogenous IFN-beta. In fact, the addition of low doses of both IFN-gamma and IFN-beta to either Lpsn or Lpsd PM resulted in synergistic antiviral effects. In vivo treatment of Lpsd mice with granulocyte-macrophage (GM)-CSF, M-CSF, IFN-gamma or Newcastle disease virus rendered peritoneal cells capable of transferring an antiviral state. These results indicate that (i) various stimuli can induce IFN-beta production by PM, (ii) Lpsd PM spontaneously transcribe low levels of IFN-beta mRNA, even though they cannot transfer an antiviral state, (iii) different stimuli, but not LPS, induce a normal IFN response in Lpsd PM, (iv) IFN-gamma increases the accumulation of IFN-beta mRNA in Lpsd PM by post-transcriptional mechanisms and (v) IFN-gamma may act synergistically with endogenous IFN-beta in inducing a potent antiviral state to VSV in PM.

Experimental design for the evaluation of the antitumor action of cytokines.

Cytokines are cellular proteins capable of exerting a variety of different biological effects both in vitro and in vivo. The availability of large amounts of recombinant highly purified cytokines now allows clinicians to explore the possible therapeutic use of these molecules. Some of these cytokines (such as interferons, "tumor necrosis factor" and interleukin-2) have been widely shown to exert antitumor effects in animal model systems and are now used in clinical trials to treat cancer patients. However, the mechanisms of these antitumor effects are poorly understood. In view of their multiple biological properties it appears very important to define suitable experimental systems for studying the antitumor effects of cytokines. In fact, one of the major problems facing researchers involved in the cytokine field is to evaluate the relevance of the different biological effects observed in in vitro cell systems with respect to the antitumor effects observed in vivo. Tumor bearing-mice injected with transplantable tumor cells can represent unique, preclinical, experimental systems to study the mechanisms of antitumor action of cytokines. In fact, only by combining the information derived from in vitro cell systems with data obtained from suitable animal models it is possible to achieve relevant insights on the mechanisms of antitumor action of cytokines. Such in vitro and in vivo studies should represent a basic support for a better use of cytokines in clinical trials with cancer patients. In this article we review the major mouse models for studying the mechanisms of antitumor action of cytokines. Furthermore, we briefly summarize our data on the antitumor effects of cytokines in mice injected with transplantable Friend leukemia cells and we discuss the advantages and the disadvantages in choosing specific animal models for studying the antitumor effects of cytokines.

Anti-tumor effects of interleukin-2 and interleukin-1 in mice transplanted with different syngeneic tumors.

We have studied the anti-tumor effects of human recombinant IL-2, alone or in association with LAK cells, in mice transplanted subcutaneously (s.c.) with the following syngeneic tumors: highly metastatic Friend leukemia cells (FLC), nonmetastatic FLC, lymphoma RBL-5 cells and HeJ16 fibrosarcoma cells. In these tumor models, peri-tumoral injections of IL-2 were more effective in inhibiting tumor growth than a systemic treatment. Although s.c. IL-2 treatment resulted in marked inhibition of tumor growth in mice injected s.c. with highly metastatic FLC, it was not effective in inhibiting growth of FLC in the liver and spleen. IL-2 therapy was more effective at increasing survival time in mice transplanted with non-metastatic FLC or with RBL-5 cells. In mice transplanted with HeJ16 fibrosarcomas, s.c. IL-2 treatment resulted in highly significant anti-tumor effect and survival of 70% of tumor-injected mice. No general correlation was found between in vitro sensitivity or resistance to the cytolytic activity of LAK cells and the anti-tumor effects observed in vivo. Subcutaneous injection of IL-1 beta in mice transplanted with highly metastatic FLC resulted in a marked increase in survival time and inhibition of metastatic tumor growth in liver and spleen. Combined treatment of IL-1 beta and IL-2 produced a synergistic anti-tumor effect: 60% of mice injected with highly metastatic FLC survived. Combined IL-1/IL-2 treatments exerted no anti-tumor activity either in DBA/2 mice injected with antibody to Thy 1.2 antigen or in nude mice, indicating that T cells play important roles during IL-1/IL-2 therapy. In vitro treatment of FLC with IL-1 beta resulted in a slight inhibition of cell multiplication, whereas even high doses of IL-2 did not affect FLC multiplication. Our results indicate that local combined treatments with IL-1 and IL-2 can induce potent, host-dependent (T cell-mediated) anti-tumor effects against highly malignant tumors.

Studies on the mechanism of the interferon-mediated antiviral state to vesicular stomatitis virus in resting mouse peritoneal macrophages.

We have analysed the expression of vesicular stomatitis virus (VSV) proteins in virus-infected freshly explanted mouse peritoneal macrophages (resistant to virus replication), macrophages aged in vitro (permissive for virus replication) and freshly explanted macrophages from mice treated with antibody to interferon (IFN) alpha/beta (permissive for VSV replication). Our data showed that some VSV proteins (i.e. N/NS and G) were synthesized in virus-infected (1 p.f.u/cell) freshly harvested macrophages at early times after infection (3 to 6 h); the expression of such viral proteins was subsequently inhibited at 18 h post-infection. In contrast, a progressive increase in the expression of VSV proteins was observed in the macrophages aged in vitro and infected with VSV at 1 p.f.u./cell. Infection with a higher m.o.i. (16 p.f.u./cell) resulted in similar viral protein electrophoresis patterns for both aged macrophages and freshly explanted macrophages. Even at low m.o.i. a marked and progressive expression of all VSV proteins was observed in freshly harvested macrophages from mice treated with antibody to mouse IFN-alpha/beta. Higher levels of oligo-2',5'-adenylate synthetase (2-5AS) were found in freshly harvested macrophages than in either aged macrophages or those from mice treated with antibody to IFN. No dsRNA-dependent 67K protein kinase was detected in freshly harvested macrophages or peritoneal cells from untreated mice or mice treated with poly(rI).poly(rC) or Newcastle disease virus. The following conclusions can be drawn from these results. Low levels of spontaneous IFN-alpha/beta are responsible for the time-dependent inhibition of VSV protein synthesis in virus-infected freshly harvested macrophages; high levels of 2-5AS (in the absence of detectable levels of 67K protein kinase) appear to correlate with the progressive inhibition of VSV proteins; this natural antiviral state is highly effective only at low m.o.i.

Role of endogenous alpha/beta interferon in the selection of virus nonproducer Friend leukemia cells after serial intraperitoneal passages in syngeneic mice.

Serial intraperitoneal passage of interferon (IFN)-sensitive Friend leukemia cells (FLCs) and L1210-S and RBL-5 tumor cells in syngeneic mice resulted in the selection of tumor cells exhibiting a marked decrease in the capacity to release reverse transcriptase (RT) activity. The virus nonproducer phenotype was a stable characteristic of clones derived from in vivo-passaged IFN-sensitive 745 FLCs. In contrast, in vivo passages of IFN-resistant 3Cl-8 FLCs and L1210-R cells did not result in any significant decrease in the capacity of these tumor cells to release in vitro RT activity. Although in vitro treatment of IFN-sensitive FLCs with mouse alpha/beta IFN (IFN-alpha/beta) for 1 or 10 passages resulted in a marked inhibition in the release of RT activity, these effects were completely reversible after removal of IFN from the culture medium. In addition, in vitro treatment of 745 FLCs with IFN resulted in a marked increase in the expression of H-2 (class I) and gp70 Friend virus antigens on the cell membrane. These effects were not observed in IFN-resistant 3Cl-8 cells. To investigate the possible role of endogenous IFN in the in vivo selection of virus nonproducer tumor cells, IFN-sensitive virus producer FLCs were serially passaged intraperitoneally in mice treated with antibodies to IFN-alpha/beta and in control mice, and the recovered tumor cells were cloned in vitro. Most (83 to 91%) of the clones derived from 745 cells recovered from control mice did not produce any detectable RT activity in the culture supernatants. In contrast, 96% of the clones (26 of 27) derived from 745 cells recovered from mice serially treated with antibodies to IFN-alpha/beta were still capable of releasing high levels of RT activity in the culture medium, indicating that endogenous IFN-alpha/beta was indeed an important host component for the in vivo selection of virus nonproducer tumor cell variants. The results reported in this article indicate that both direct effects of IFN on tumor cells and host-mediated effects are involved in this phenomenon.

Effect of mouse interferon alpha/beta on the expression of H-2 (class I) antigens and on the levels of 2'-5' oligoadenylate synthetase activity in interferon-sensitive and interferon-resistant Friend leukemia cell tumors in mice.

DBA/2 mice were injected intraperitoneally (i.p.) with interferon-sensitive 745 or interferon-resistant 3C1-8 Friend erythroleukemia cells (FLC) and then injected i.p. with mouse interferon alpha/beta. Interferon enhanced the expression of histocompatibility (H-2) antigens on individual 745 FLC within the peritoneum, but did not alter the expression of H-2 antigens on individual 3C1-8 FLC. Likewise, interferon treatment resulted in an increase in the level of 2'-5' oligo-adenylate (2-5A) synthetase activity in 745 FLC, but did not affect the level of activity in 3C1-8 FLC. These results provide evidence that the phenotype of interferon sensitivity or resistance of FLC does not change within the peritoneum. An incidental finding was that the basal level of 2-5A synthetase activity of in vivo passaged 745 cells was greater than that of 3C1-8 FLC. The finding that injection of mice bearing 745 FLC with antibody to mouse interferon alpha/beta reduced the level of 2-5A synthetase activity in these cells, but did not alter the level of 2-5A activity in 3C1-8 FLC, suggests that endogenous interferon in the peritoneum may have been the responsible factor.

Studies on the expression of spontaneous and induced interferons in mouse peritoneal macrophages by means of monoclonal antibodies to mouse interferons.

Monoclonal antibodies (MAbs) to mouse interferons (MuIFN) have been used to characterize the interferon-like activities spontaneously expressed in mouse peritoneal macrophages freshly explanted from normal pathogen-free mice. Injection of mice with MAbs to MuIFN-alpha or -beta resulted in a significant increase of vesicular stomatitis virus (VSV) multiplication in peritoneal macrophages. Addition of these MAbs to freshly explanted mouse macrophages accelerated the decay of the antiviral state to VSV during the 'ageing' in vitro of these macrophage cultures. Furthermore, these MAbs to MuIFN-alpha or -beta markedly inhibited the transfer of the antiviral state from freshly explanted peritoneal cells or macrophages to syngeneic macrophages 'aged' in vitro permissive for virus replication. These effects were not observed using a non-neutralizing antibody to MuIFN-alpha, nor with a MAb to MuIFN-gamma. In all experiments sheep polyclonal antibodies to MuIFN-alpha/beta were more effective than the corresponding amount of MAbs to MuIFN-alpha or -beta. A mixture of both these MAbs was more effective than either alone. Interferons produced after stimulation of peritoneal macrophages with Newcastle disease virus (NDV) and of total peritoneal cells with lipopolysaccharides (LPS) have also been characterized by means of MAbs to IFNs. The results of neutralization studies with these antibodies indicated that MuIFN-beta was the major component of peritoneal cell IFN (induced by both NDV and LPS) and MuIFN-alpha was a minor component (13 to 17%). These data indicate that both MuIFN-alpha and -beta, but not MuIFN-gamma, are spontaneously present in/on mouse peritoneal macrophages and are produced after stimulation with NDV or LPS.