Distinct Pathogenic Sequela in Rhesus Macaques Infected with CCR5 or CXCR4 Utilizing SHIVs

See allHide authors and affiliations

Science  30 Apr 1999:
Vol. 284, Issue 5415, pp. 816-819
DOI: 10.1126/science.284.5415.816


Infection of macaques with chimeric simian–human immunodeficiency virus (SHIV) provides an excellent in vivo model for examining the influence of envelope on HIV-1 pathogenesis. Infection with a pathogenic CCR5 (R5)–specific enveloped virus, SHIVSF162P, was compared with infection with the CXCR4 (X4)–specific SHIVSF33A.2. Despite comparable levels of viral replication, animals infected with the R5 and X4 SHIV had distinct pathogenic outcomes. SHIVSF162P caused a dramatic loss of CD4+ intestinal T cells followed by a gradual depletion in peripheral CD4+ T cells, whereas infection with SHIVSF33A.2 caused a profound loss in peripheral T cells that was not paralleled in the intestine. These results suggest a critical role of co-receptor utilization in viral pathogenesis and provide a reliable in vivo model for preclinical examination of HIV-1 vaccines and therapeutic agents in the context of the HIV-1 envelope protein.

Despite recent advances in the treatment of AIDS, its severe global impact heightens the need for an efficacious vaccine against human immunodeficiency virus (HIV). Toward this end, significant efforts have been directed at developing animal models that faithfully mimic HIV-1 infection. Such systems provide a controlled setting not only for preclinical evaluation of antiviral compounds and vaccines but also for addressing issues related to viral pathogenesis. Several nonhuman primate models have been used to study HIV-1 infection. These include infection of macaques with simian immunodeficiency virus (SIV) (1); infection of chimpanzees and macaques with HIV-1 (2); infection of baboons and macaques with HIV-2 (3); and infection of macaques with chimeric envelope SHIV (simian–human immunodeficiency virus) constructed by substituting the tat, rev, andenv coding regions of the pathogenic molecular clone SIVmac239 with corresponding regions from HIV-1 clones (4). Although each model provides insight about host and viral determinants of pathogenesis, the fidelity of these models has been limited by various parameters (5). For example, infection with pathogenic env SHIVs generally results in a profound state of immunodeficiency characterized by a precipitous loss in peripheral CD4+ T cells (6). The kinetics of SHIV-induced lymphocyte loss differs significantly from that observed in either pathogenic SIV or HIV-1 infection (7), which raises doubts about the validity of this model system. The pathogenic SHIVs described thus far function with the CXCR4 (X4) co-receptor (8) or are dual-tropic, utilizing both CCR5 and CXCR4 (R5X4). The disease induced by these X4 and R5X4 SHIVs is more characteristic of late-stage HIV-1 infection or the rare infection of individuals who are homozygous for a 32–base pair (Δ32/Δ32) deletion in the gene encoding the CCR5 receptor (9). Development of a pathogenic CCR5 (R5)-specific SHIV might facilitate the preclinical analysis of candidate antiviral compounds and vaccines and reveal aspects of viral pathogenesis not observed in previously described animal model systems.

With the goal of developing an R5-specific pathogenic SHIV, we inoculated two rhesus macaques with a molecular clone SHIVSF162 followed by three sequential blood–bone marrow transfusions into naı̈ve macaques (10). This approach has been successfully used to generate X4 and dual-tropic pathogenic SHIVs (11). We collected longitudinal samples of peripheral blood and lymph node from each of the eight animals and analyzed the samples for plasma viremia and changes in T cell subsets (12). Significant amounts of viral replication were achieved during the acute phase of infection in passage 1 (P1) macaques with 106 viral RNA copies per milliliter of plasma. By 3 weeks postinfection (wpi), viremia began to decline and was not detected in either P1 animal after 6 weeks; nonetheless, proviral DNA and antiviral immune responses were still detected 11 months pi (13). Similar kinetics of viral replication were observed in passage 2 (P2) monkeys, albeit with slightly increased viral loads of 106to 107 viral RNA copies per milliliter of plasma. Additional successive in vivo passage of SHIVSF162, now designated SHIVSF162P, significantly enhanced viral replication as demonstrated by viral peaks between 107 and 108 viral RNA copies per milliliter of plasma in both P3 (T353 and T833) and P4 (T378 and T642) macaques, and two animals sustained plasma viremia between 104 and 105viral RNA copies per milliliter of plasma for over 11 months (Fig. 1A). Concomitant with peak viral load (2 wpi) was a moderate drop in the absolute number of CD4+peripheral blood mononuclear cells (PBMCs). However, this cell population rebounded to near baseline values at 4 to 6 wpi (Fig. 1B).

Figure 1

Peripheral blood was collected in EDTA at weekly intervals and separated into plasma and cellular compartments. (A) Copies of viral RNA in plasma per milliliter of plasma determined by bDNA analysis (Chiron). P3 macaques, T353 (solid circles) and T833 (open circles); P4 macaques, T378 (solid squares) and T642 (open squares). (B) Absolute number of CD3+ CD4+ cells per microliter of blood as determined by TruCount FACS analysis.

To determine whether SHIVSF162P maintained CCR5 specificity, we used PBMCs purified from a homozygous Δ32/Δ32 seronegative donor for in vitro studies. Results showed that, similar to the parental SHIVSF162, viruses isolated from P4 animals [T378, lymph node mononuclear cell (LNMC) isolate; T642, PBMC isolate] 2 weeks after infection did not infect human PBMCs lacking CCR5 surface expression, whereas the X4-specific SHIVSF33A.2 clone readily infected both wild-type and Δ32/Δ32 PBMCs (Fig. 2). Additionally, using HOS cells engineered to express the CD4 protein along with specific co-receptors, we found that P4 viruses were able to infect GHOST cells expressing the CCR5 protein but not those expressing CXCR4, BOB/GPR15, or Bonzo/STRL33 (13, 14). Collectively, these data strongly suggest that, despite vigorous in vivo replication of SHIVSF162P, viruses isolated from infected animals maintain CCR5 specificity with no evidence of expanded co-receptor utilization.

Figure 2

Virus recovered from P4 animals 2 weeks after infection was examined for replication in PBMCs isolated from a CCR5 wild-type donor (CCR5 +/+; solid squares) and a donor homozygous for Δ32 (CCR5 −/−; hatched squares). Culture supernatants were collected 10 days after infection and analyzed for p27gag content by antigen ELISA (Cellular Products, Buffalo, New York).

The transient loss in peripheral CD4+ T cells in P3 and P4 animals early after infection (Fig. 1B) is in sharp contrast to the profound and sustained depletion that occurs during primary infection with previously described X4- and dual-tropic SHIVs (6). A comparison of the viral load and peripheral T cell ratios between a representative macaque infected with SHIVSF33A (15) and the P3 SHIVSF162P-infected animal T353 illustrates a striking difference in CD4+ cell loss in these macaques despite comparable levels of viral replication (Fig. 3). In light of recent reports that indicate the gastrointestinal tract, and not peripheral lymphoid tissue, is the major site for replication of SIV during primary infection, we analyzed T cell populations within the gut-associated lymphoid tissue (GALT) in the P4 macaques (16). Jejunal biopsies and mesenteric nodes were obtained from P4 macaques by surgical resectioning of the intestinal tract 2 weeks after viral infection (17). Intestinal epithelia and lamina propria lymphocytes (LPLs) were isolated from fresh jejunal tissue (18) and LNMCs were isolated from both colonic and inguinal lymph nodes followed by three-color flow cytometry analysis for expression of CD3, CD4, and CD8. During acute infection, both P4 animals suffered depletion in the CD4+ LPL subset (T378, 8%; T642, 7%; Fig. 4A). This severe diminution in CD4+ T cells is unique to the GALT as lymphocytes isolated concurrently from blood or lymph nodes showed moderate changes in the percentage of CD4+ cells. To determine whether the CD4+ T cell subset was restored in jejunal tissue, we performed a follow-up analysis on the same P4 animals during chronic infection (26 wpi). After nearly 6 months of infection, the percentage of CD4+ T cells detected in the jejunum (T378, 24% CD4+; T642, 14% CD4+) had regenerated slightly compared with amounts during acute infection (Fig. 4, B and C); however, the percentage of CD4+ cells within LPLs was diminished compared with published uninfected control values (18). To further examine the pathogenicity of SHIVSF162P, we inoculated four naı̈ve animals intravenously with cell-free virus recovered from PBMCs of macaque T378 (P4) and then analyzed the tissue. These animals also had high levels of plasma viremia accompanied by depletion of CD4+ jejunal LPLs (9.7% ± 5.5%) with minimal changes in circulating CD4+ T cells at 2 wpi (13). Thus, similar to animals infected with CCR5 utilizing SIV, animals infected with SHIVSF162P had a profound loss of CD4+ T cells within the gastrointestinal tract while maintaining a relatively high number of peripheral CD4+ T cells. Similarly, analysis of asymptomatic HIV-1 patients suggests that CD4+ T cell depletion within the intestinal mucosa precedes the depletion in peripheral blood (19).

Figure 3

Comparison of plasma viremia (viral RNA per milliliter of plasma) (open circles) and CD4-CD8 T cell ratios (solid circles) during primary infection with SHIVSF162P(A) and SHIVSF33A (B).

Figure 4

(A) Three-color flow cytometry dot plots comparing lymphocytes isolated from the peripheral blood (PBMCs), colonic lymph node (LNMCs), and intestinal lamina propria (LPLs) 2 weeks after infection with SHIVSF162P. (Top) Cells obtained from P4 animal T378; (bottom) cells from P4 animal T642. Dot plots were generated by first selecting lymphocyte population and then selecting CD3+ T cells. Analysis of CD4-CD8 T cell ratios in lymphoid tissue at 2 weeks in peripheral blood (PBMCs; solid squares), colonic lymph node (LNMCs; hatched squares), and intestinal lamina propria (LPLs; striped squares) (B), and at 26 weeks in SHIVSF162P-infected macaques (C). The actual percentage of CD4+ T lymphocytes in each compartment is given in (B) and (C).

To further explore differential pathogenesis associated with X4 and R5 viruses, we determined T cell loss in GALT of X4-infected animals by means of a recently developed and characterized infectious molecular clone of SHIVSF33A designated SHIVSF33A.2(13). We inoculated two macaques intravenously with SHIVSF33A.2 and subjected them to the same blood, lymph node, and jejunal sampling procedures as described for the P4 SHIVSF162 animals. As observed with the parental isolate SHIVSF33A and other pathogenic X4 or R5X4 SHIVs (6), animals infected with the X4-specific SHIVSF33A.2 had a profound loss of circulating CD4+ T cells within 2 weeks of infection. Moreover, in marked contrast to SHIVSF162P is the absence of severe CD4+ T cell depletion in the GALT of SHIVSF33A.2-infected macaques (E239, 32%; G901, 49%) (Fig. 5). Although we did not determine the amount of viral infection in mucosal tissues, SHIV162P- and SHIVSF33A.2-infected animals had comparable levels of plasma viremia.

Figure 5

Comparison of T cell ratios in PBMC (solid squares) and jejunal LPL (striped squares) subsets 2 weeks after infection with SHIVSF162P (A) and SHIVSF33A.2 (B). The actual percentage of CD4+ T lymphocytes in each compartment is given.

Despite viral loads comparable to that of the pathogenic X4-utilizing SHIVSF33A and significant steady-state level of viral replication, the characteristic rapid and sustained inversion of the peripheral CD4-CD8 T cell ratio did not occur in late passaged SHIVSF162P-infected animals (Fig. 3). Rather, these animals had a gradual loss in the absolute number of peripheral CD4+ T cells reminiscent of most HIV-1 and SIV infections. However, analysis of LPLs purified from jejunal tissue revealed a severe depletion of mucosal CD4+ T cells within 2 weeks of inoculation, demonstrating that SHIVSF162Pis intrinsically pathogenic. Thus, productive infection with SHIVSF162P results in a pathogenic sequela similar to SIV infection in macaques rather than previously described pathogenic SHIV infections. This is further illustrated by our finding that, although macaques infected with the X4-specific molecular clone SHIVSF33A.2 suffer a profound loss of circulating CD4+ T cells, they do not undergo severe depletion in jejunal LPLs immediately after infection. These marked differences in viral pathogenesis occur despite similar levels of sustained plasma viremia. The most obvious difference between SHIVSF162P and SHIVSF33A.2 is co-receptor utilization that dictates the target cell of infection of each virus. It is known that CCR5 and CXCR4 expression varies in different cell types and at different stages of human T cell development (20). The depletion of CD4+ lymphocytes within the GALT during SIV infection may be due to the high percentage of activated memory CD4+T cells within the gut (16) that preferentially express CCR5 co-receptor (20). The absence of CD4+ T cell loss in jejunal LPLs of X4-infected animals therefore may reflect the lack of susceptible host target cells within the intestinal mucosa. In this regard, preliminary analysis of CXCR4 expression in various lymphoid compartments of SHIVSF33A.2-infected macaques indicates that, although comparable percentages of CD3+ CD4+CXCR4+ cells are present in these anatomically distinct compartments, the level of CXCR4 expression is significantly lower in the CD3+ LPLs than in LNMCs and PBMCs (13). Similar analysis with antibodies that recognize macaque CCR5 and markers of T cell development and differentiation will likely shed light on the mechanisms that underlie the preferential loss of CD4+ T cells from distinct compartments of SHIVSF162P-infected macaques. Thus, our findings suggest that target cell and tissue tropism, as influenced by differential co-receptor expression, may have profound implications on viral pathogenesis and subsequent host disease course.

The availability of pathogenic R5- and X4-specific envelope SHIVs that replicate with similar kinetics and to comparable levels provides the tools that are necessary to answer important questions related to viral pathogenesis, transmission, adaptation, and evolution. Although our data suggest that differential co-receptor utilization of SHIVSF162P and SHIVSF33A.2 significantly alters viral pathogenesis, it does not a priori indicate that co-receptor utilization is the sole determinant of viral pathogenesis. Co-receptors mediate viral entry into target cells, but productive HIV infection requires cellular activation (21). Thus, further phenotypic analysis of the host cells that are susceptible to productive infection with SHIVSF162P and SHIVSF33A.2 may provide insight into these important issues of pathogenesis. Additionally, the SHIVSF162P and SHIVSF33A animal model system described will be instrumental in addressing the long debated issues of preferential R5 transmission and phenotypic switch. Whether R5 viruses are exclusively selected or amplified during viral transmission and subsequently evolve into X4 viruses, perhaps to evade host immune responses or, alternatively, whether both R5 and X4 viruses are transmitted but X4 virus is sequestered early after transmission and eventually outgrows R5 virus can now be examined (22). An understanding of these and other interactions between HIV-1 proteins and host target cells should lead to improved understanding of viral pathogenesis and should help in development of effective vaccine and therapeutic candidates.

  • * To whom correspondence should be addressed. E-mail: cmayer{at}


View Abstract

Stay Connected to Science

Navigate This Article