ABSTRACT
Human herpesvirus 6 (HHV-6) has been proposed as a potential cofactor in the progression of human immunodeficiency virus type 1 (HIV-1) disease. We used the SCID-hu Thy/Liv mouse model to evaluate the in vivo interactions between HHV-6 and HIV-1. Our results demonstrate that HHV-6 and HIV-1 can simultaneously replicate in the human thymus in vivo. In this model, however, the presence of one virus appears not to modify the replication or cytopathicity of the other.
Human herpesvirus 6 (HHV-6) is a betaherpesvirus first isolated from patients with lymphoproliferative disorders and immunosuppression due to either antineoplastic chemotherapy or human immunodeficiency virus type 1 (HIV-1) infection (29). All HHV-6 isolates can be classified into two subgroups, A and B. The B subgroup is predominant in the human population and is the etiological agent of exanthem subitum in infants. Although the A subgroup has not been definitively linked to any disease, most of the A isolates have been obtained from immunocompromised patients, including patients with AIDS (21). Both HHV-6 subgroups are characterized by a primary tropism for CD4+ T cells but can also infect, either productively or nonproductively, a wide variety of other cell types (18). Consistent with this broad tropism, a ubiquitous complement regulatory protein, CD46, was recently identified as a cellular receptor for HHV-6 (30).
Several lines of clinical and experimental evidence suggest that HHV-6 may play a role in the progression of HIV-1 disease (18,21). In particular, HHV-6 shares with HIV-1 a primary tropism for CD4+ T cells; the two viruses can simultaneously replicate in such cells, resulting in accelerated HIV-1 expression and cellular death in vitro (20). However, other studies have documented inhibition of HIV-1 replication by HHV-6, possibly as a consequence of rapid destruction of CD4+ T cells (5, 17). Moreover, HHV-6 A is able to induce expression of CD4 in cytotoxic effector cells, such as mature CD8+ T cells, NK cells, and γ/δ cells, rendering them susceptible to HIV-1 infection (21). A role of HHV-6 in HIV-1 infection is also suggested by clinical observations, including the demonstration of active and disseminated HHV-6 infection in symptomatic HIV-infected individuals (21), the rapid HIV-1 disease course documented in children with early acquisition of HHV-6 (15), and the dramatic acceleration of simian immunodeficiency virus infection in macaques coinfected with HHV-6 (19). However, the hypothesis that HHV-6 acts as a cofactor in AIDS remains unproven. Three different approaches have been suggested for clarifying the role played by HHV-6 in HIV-1 disease (10, 21): (i) controlled longitudinal studies of HHV-6 replication in HIV-1 infected patients, (ii) prospective clinical trials in HIV-1-infected patients using antiviral compounds with selective activity against HHV-6, and (iii) coinfection experiments in susceptible animal models.
We recently reported that SCID-hu Thy/Liv mice are permissive for HHV-6 replication and may represent a useful model to elucidate the immunosuppressive mechanisms of this virus (9). SCID-hu Thy/Liv mice are a well-established small-animal model for the study of HIV-1 infection (1, 4, 26, 33-35), and they have also been successfully used to investigate certain aspects of the pathogenesis of other human viral infections (2, 6, 8, 24, 25). SCID-hu Thy/Liv mice carry a human graft that supports long-term human lymphopoiesis and histologically resembles a human thymus (27). After direct intrathymic infection, HHV-6 isolates of both A and B subgroups were found to preferentially infect intrathymic T-progenitor cells (ITTPs) and to deplete thymocytes within the Thy/Liv implants (9). Notably, the ITTP is a rapidly cycling cell (16) that bears both CD4 and CXCR4 (3); thus, it represents a target for infection by CXCR4-utilizing isolates of HIV-1 (35). Infection and destruction of this cell subpopulation by HIV-1 (and perhaps also by HHV-6) may lead to thymocyte depletion because replenishment of more mature thymocytes cannot proceed (4,12, 13). Since both HIV-1 and HHV-6 can infect the ITTP subpopulation in the SCID-hu Thy/Liv mouse, we investigated the possibility that direct interactions between the two viruses in this model might alter their replication or pathogenicity.
SCID-hu Thy/Liv mice obtained from five different donors were coinfected by direct intrathymic injection of HHV-6 and HIV-1. Two isolates of HHV-6 (GS [subgroup A] and PL-1 [subgroup B]) (9) and two isolates of HIV-1 (NL4-3 and JD) (28) were tested. Controls included mice infected with either HHV-6 or HIV-1 alone. UV-inactivated viral stock or culture medium was used for mock inoculations. Mice were euthanatized by CO2 inhalation at various times after inoculation, and implants were removed and dispersed through a nylon mesh. Thymocyte subpopulations were analyzed by quantitative flow cytometry of total thymocyte suspensions stained with antibodies to CD3, CD4, and CD8, as previously described (9,35). The HHV-6 DNA load was quantitated with the real-time TaqMan fluorogenic detection system on an ABI PRISM 7700 Sequence Detector (Perkin-Elmer Applied Biosystems), as previously described (9; G. Locatelli, F. Santoro, A. Gobbi, F. Veglia, P. Lusso, and M. Malnati, submitted for publication). The limit of detection was 1.7 log10 DNA copies per 106cells, and this lower-limit value was used for mean calculation of implants with undetectable HHV-6 DNA. Quantitation of HIV-1 p24 was carried out with a commercial enzyme-linked immunosorbent assay kit (DuPont), and HIV-1 RNA load was measured by using the branched DNA test (Chiron). The limit of detection of the latter assay was 2.3 log10 RNA copies per 106 cells, and this lower-limit value was used for mean calculation of implants with undetectable viral RNA. Statistical analysis was performed by using the Mann-Whitney U test (StatView 5.0; Abacus Concepts). For immunohistochemistry, Thy/Liv implants were fixed in 10% buffered formalin, embedded in paraffin, and cut into 4-μm-thick sections. The sections were then stained with 9A5D12, a monoclonal antibody specific for the p41 early protein of HHV-6, as described elsewhere (9) or with a monoclonal antibody specific for HIV-1 p24 (Dako) by using a standard heat-induced epitope retrieval technique. All procedures and practices involving animals were approved by the University of California, San Francisco, Committee on Animal Research.
The first experiment was designed to study the interactions between HHV-6GS (subgroup A) and HIV-1NL4-3. These strains were chosen because they are both known to infect ITTPs in the SCID-hu Thy/Liv model (9, 35). Implants were infected by simultaneous inoculation with 104 50% tissue culture infective doses (TCID50) of HHV-6AGS and 103 TCID50 of HIV-1NL4-3 and were terminated 7 and 10 days after inoculation. Seven days after inoculation, no thymocyte depletion was observed in either singly infected or coinfected implants (Fig. 1A and data not shown), and no differences in HHV-6 DNA (Fig. 1B), HIV-1 RNA (Fig. 1C), or p24 (data not shown) were observed in coinfected animals compared to those infected with HHV-6AGS or HIV-1NL4-3 alone. At day 10, coinfected implants showed a moderate, statistically significant reduction in percentage of ITTPs (Fig. 1A) and a modest reduction of the HIV-1 RNA load (Fig. 1C).
Replication of HHV-6AGS and HIV-1NL4-3 in SCID-hu Thy/Liv implants. The percentage of ITTPs was quantified as a percentage of total live thymocytes and is shown as mean ± standard error of the mean (SEM) (A). Implant HHV-6 DNA (B) and HIV-1 RNA (C) loads are shown as means ± SEMs. The numbers at the base of each column indicate the number of animals in each cohort. When only two animals could be analyzed, the data are shown as mean without SEM. ∗, P < 0.05; ∗∗, P < 0.01 compared to implants infected with HIV-1 alone by Mann-Whitney U test. The arrows designate the limits of detection for viral nucleic acid (see text).
, mock-infected implants; ■, HIV-1-infected implants;
, HIV-1- and HHV-6-coinfected implants;
, HHV-6-infected implants.
In a second experiment, Thy/Liv implants were coinfected with HHV-6GS (104 TCID50) and HIV-1JD (103.7 TCID50), a strain characterized by a preferential tropism for CD4+thymocytes, which are more mature than ITTPs (35). Seven days after inoculation, ITTPs appeared to be depleted in coinfected implants obtained from donor A (Fig. 2A), which also showed a higher HHV-6 DNA load (Fig. 2B). However, neither ITTP depletion nor enhancement of HHV-6 replication was observed in mice prepared from a different donor (donor E), at either day 7 or day 10. There were no statistical differences between singly infected and coinfected implants in either HIV-1 RNA (Fig. 2C) or p24 (data not shown), suggesting that the coinfection had no effect on HIV-1JD replication. To better evaluate the levels of HHV-6 and HIV-1 expression in coinfected implants, immunohistochemistry with monoclonal antibodies specific for HHV-6 p41 and HIV-1 p24 was performed. At both day 7 (data not shown) and day 10 (Fig.3), a relatively high proportion of cells expressed HHV-6 antigens (Fig. 3A), and there was a distinct population of large macrophage-like cells intensely expressing HIV-1 p24 protein (Fig. 3B).
Replication of HHV-6AGS and HIV-1JD in SCID-hu Thy/Liv implants. The percentage of ITTPs was quantified as a percentage of total live thymocytes and is shown as mean ± SEM (A). Implant HHV-6 DNA (B) and HIV-1 RNA (C) loads are shown as means ± SEMs. ∗, P < 0.05 compared to implants infected with HIV-1 alone; §, P < 0.05 compared to implants infected with HHV-6 alone by Mann-Whitney U test. See the legend of Fig. 1 for definitions of symbols and numbers.
Immunohistochemistry of HHV-6AGS- and HIV-1JD-coinfected SCID-hu Thy/Liv implants. (A) Expression of HHV-6 p41 in an implant harvested 10 days after inoculation. Infected cells are scattered throughout the implant (arrowheads) (magnification, ×98). (B) Expression of HIV-1 p24 in the same implant (magnification, ×98).
A third experiment was designed to study the interactions between HHV-6 and HIV-1 at an earlier time point after inoculation (prior to the induction of thymocyte depletion), as well as to evaluate the effects of a smaller inoculum of HHV-6. Thy/Liv implants were injected simultaneously with 103 TCID50 of HHV-6AGS and with 103 TCID50 of HIV-1NL4-3. Viral replication was assessed 3 and 7 days after inoculation (Fig. 4). No significant differences in the viral loads of either virus were observed at either time point and, as expected, no evidence of thymocyte or ITTP depletion was observed (data not shown).
Early replication of HHV-6AGS and HIV-1NL4-3 in SCID-hu Thy/Liv implants. Implant HHV-6 DNA (A) and HIV-1 RNA (B) loads are shown as means ± SEMs. See the legend of Fig. 1 for definitions of symbols and numbers.
We then evaluated the effects of a second strain of HHV-6 (PL-1), a primary isolate belonging to subgroup B. Implants were infected with 104 TCID50 of HHV-6 and 103TCID50 of HIV-1NL4-3 and were terminated 10 days after inoculation (Fig. 5). As in the first experiment (Fig. 1), coinfected implants had somewhat lower levels of HIV-1 replication (Fig. 5C), but this occurred in the presence of slight to moderate levels of thymocyte depletion (data not shown) and severe depletion of ITTPs (Fig. 5A) in both singly infected and coinfected implants.
Replication of HHV-6BPL-1 and HIV-1NL4-3 in SCID-hu Thy/Liv implants. The percentage of ITTPs was quantified as a percentage of total live thymocytes and is shown as mean ± SEM (A). Implant HHV-6 DNA (B) and HIV-1 RNA (C) loads are shown as means ± SEMs. See the legend of Fig. 1 for definitions of symbols and numbers.
In summary, these experiments indicate that SCID-hu Thy/Liv mice are able to support the simultaneous replication of HHV-6 and HIV-1. However, aside from effects on ITTPs that were observed in some but not all experiments, infection with one virus appeared not to enhance the replication or pathogenicity of the other in this small-animal model. Although previous in vitro studies demonstrated an enhancing effect of HHV-6 on HIV-1 replication (21), the HIV-1 viral load was either unchanged or reduced in coinfected implants compared to those infected with HIV-1 alone. Because such reductions occurred at a time when HHV-6 had already started to induce thymocyte depletion in singly infected samples, the moderate inhibition of HIV-1NL4-3replication seen in coinfected implants might be related to depletion of the common target cells (i.e., ITTPs) for the two viruses. When the HHV-6 viral load was analyzed, a trend toward enhanced HHV-6 replication in implants coinfected with HIV-1 was observed in some of the conditions tested (see Fig. 1C and 2C). This result is consistent with the documented effect of HIV-1 Tat on the replication of HHV-6 in vitro (32). However, the overall lack of dramatic effects induced by coinfection with HHV-6 and HIV-1 in SCID-hu Thy/Liv mice suggests that the two viruses replicate simultaneously without affecting each other in a positive or negative way, possibly due to infection of different cells. Thus, at least under the conditions tested in this study, our results do not support the hypothesis that HHV-6 drives HIV-1 replication in vivo or vice versa.
It is, of course, important to note that the SCID-hu Thy/Liv mouse model may not accurately reflect the natural infection pattern occurring in humans. A major caveat related to this model system is the absence of both specific and innate immune mechanisms that may modulate the expression of one or both viruses. For example, accumulation of activated inflammatory cells could exert a dual effect in infected tissues by counteracting viral replication and spread while also providing new targets for infection. Additionally, it is possible that the load of each virus found in this model is different from that occurring within infected humans. For instance, 10 days after inoculation, the estimated HHV-6 DNA load was 0.3 to 10 copies per cell, a value markedly higher than previously reported in lymph nodes from both HIV-infected patients and uninfected controls (31). Similarly, the results of the immunohistochemical staining for HHV-6 were comparable to those reported for lymph nodes of HIV-infected patients (14). Regarding the HIV-1 load, we detected between 1 (Fig. 2C) and 0.001 (Fig. 3B) copies of HIV RNA per cell. These loads are comparable to those reported for human lymphoid tissue of HIV-1-infected patients (11). Moreover, the expression of HIV-1 p24 in coinfected Thy/Liv implants was generally much higher than that observed in lymphoid reservoirs of HIV patients, which is exceeded only in rare instances by patients with end-stage AIDS (B. G. Herndier, unpublished data). Thus, although it is possible that interactions between the two viruses might occur in situations wherein the proportion of HIV-1- and HHV-6-infected cells is higher, as may occur in coinfection experiments in vitro, the viral load values that we measured in coinfected Thy/Liv implants were comparable to or even higher than those reported for each virus in human lymphoid tissues.
Further studies with susceptible nonhuman primate models and clinical studies in HIV-1-infected patients will help to clarify the role of HHV-6 in AIDS. Meanwhile, and in consideration of the increasingly recognized role of the thymus in both children and adults with HIV-1 infection (7, 12, 22, 23), it seems important to investigate the mechanisms by which HHV-6 induces thymocyte depletion. Such depletion, for example, may be associated with acceleration of HIV-1 disease by affecting the thymic reserve or in another manner that is distinct from direct effects on HIV-1 replication.
ACKNOWLEDGMENTS
We thank Claudio Gattuso for figure editing.
This work was supported by European Union Biomed 2 grant PL951301 and Istituto Superiore di Sanita, Rome, grants 40A059 and 40B57 (to P. Lusso) and by grants from the National Institutes of Health (AI40312 and AI65309 to J. M. McCune and P30 MH59037 to B. G. Herndier). J. M. McCune is an Elizabeth Glaser Scientist supported by the Elizabeth Glaser Pediatric AIDS Foundation.
FOOTNOTES
- Received 22 February 2000.
- Accepted 9 June 2000.
↵* Corresponding author. Mailing address: Unit of Human Virology, DIBIT, San Raffaele Scientific Institute, Via Olgettina 58, Milan 20132, Italy. Phone: 39-02-2643-2821. Fax: 39-02-2643-4905. E-mail: Paolo.Lusso{at}hsr.it.
↵† Present address: Department of Experimental Oncology, European Institute of Oncology, 20141 Milan, Italy.
REFERENCES
- Copyright © 2000 American Society for Microbiology
