Use of a Novel Chimeric Mouse Model with a Functionally Active Human Immune System To Study Human Immunodeficiency Virus Type 1 Infection^▿

Generation of HIS-Rag2^−/−γ_c^−/− mice.

H-2^d Rag2^−/− mice were originally obtained from Anton Rolink and Shunichi Takeda (Basel Institute for Immunology, Basel, Switzerland) and were crossed with mice with a knockout of the interleukin-2 receptor gamma chain (γ_c^−/− mice) to obtain H-2^d Rag2^−/−γ_c^−/− mice (15). These double-knockout mice have no functional T and B cells and no NK cells, and in contrast to NOD/SCID mice they do not develop thymus sarcomas. Rag2^−/−γ_c^−/− mice were originally bred at The Netherlands Cancer Institute and transferred to the University of California-Los Angeles for the HIV-1 studies. Mice were maintained in isolators and fed autoclaved food and water, and all manipulations were done under laminar flow.

As described previously, HIS-Rag2^−/−γ_c^−/− mice were generated by injecting CD34⁺ cells isolated from human fetal liver tissue into the mice in the first week of life, when the conditions for the engraftment, expansion, and reconstitution of a human immune system are optimal (9). We and others have shown that the CD34⁺ cells develop into T lymphocytes on the mouse thymus stromal remnant that becomes a thymic organ and gives rise to human T cells (αβ and γδ, as well as regulatory T cells) (9, 16, 27). In addition, B cells, NK cells, monocytes, plasmacytoid dendritic cells, granulocytes, and low but consistent percentages of erythrocytes can be found in the peripheral blood and organs of the Rag2^−/−γ_c^−/− mice (9, 27). At 1 to 3 days after birth, mice were irradiated (3.5 Gy) once and received human CD34⁺ cells (5 × 10⁵) through intraperitoneal injection. Peripheral blood was taken biweekly from the HIS-Rag2^−/−γ_c^−/− mice starting 4 weeks after injection to determine the presence of human cells.

HIV-1 infection of HIS-Rag2^−/−γ_c^−/− mice.

CCR5-tropic molecular clone HIV-1_NFN-SX(SL9) stocks were prepared as previously described (24). The CCR5 tropism of the virus stocks was confirmed with CCR5⁺CD4⁺ GHOST cells (National Institutes of Health AIDS Research and Reference Program) by using green fluorescent protein expression with flow cytometry and cytopathicity, as described by Morner et al. (19). No signs of infection with HIV-1_NFN-SX(SL9) in parental or CXCR4⁺CD4⁺ GHOST cells were observed. All infections were standardized by determining 50% tissue culture infective doses (TCID₅₀) (24). HIS-Rag2^−/−γ_c^−/− mice were infected by intraperitoneal injection with small amounts of HIV-1 (500 or 5,000 TCID₅₀ of HIV-1) at 14 to 36 weeks of age (ages were as follows: experiment 1, 14 weeks; experiment 2, 14 weeks; experiment 3, 17 weeks; and experiment 4, 36 weeks). The 500-TCID₅₀ viral inoculum contained 375 pg of p24, and the 5,000-TCID₅₀ viral inoculum contained 3,750 pg of p24. The volume of each viral inoculum was 250 μl and contained 2% fetal bovine serum. Controls consisted of mock-infected mice or mice infected with heat-inactivated HIV-1. Peripheral blood, thymuses, spleens, and bone marrow were taken from the HIV-1-infected and control HIS-Rag2^−/−γ_c^−/− mice for immunophenotyping of the human cells and coculturing with phytohemagglutinin- and interleukin-2-activated CD8-depleted human peripheral blood mononuclear cells (PBMC; 5 ×10⁵ cells) for 10 days to detect HIV-1. Virus replication was assessed by measuring the level of the p24 antigen in the supernatants of the cocultures by using a specific p24 antigen enzyme-linked immunosorbent assay (ELISA; Coulter, Hialeah, FL).

Surface and intracellular immunophenotyping.

Surface immunophenotyping of human thymocytes and peripheral lymphoid cells with directly conjugated antibodies was performed as previously described (26). Fluorescent conjugated monoclonal antibodies to CD3, CD4, CD8, CD19, and CD45 and isotype control antibodies of mouse immunoglobulin G1 (IgG1) and mouse IgG2 conjugated with fluorescein isothiocyanate, phycoerythrin, and/or allophycocyanin were obtained from Becton Dickinson Immunocytometry Systems (BDIS; San Jose, CA). Monoclonal antibodies to CCR5 (clone 2D7) and CXCR4 (clone 12G5) conjugated to fluorescein isothiocyanate, phycoerythrin, or allophycocyanin were obtained from BD-Pharmingen (La Jolla, CA). Cells were acquired on a dual-laser FACSCalibur flow cytometer (BDIS, San Jose, CA). Multiparameter data acquisition and analysis were performed with Cell Quest software (BDIS).

ELISA and Western blotting.

Sera obtained from mice injected with human fetal liver CD34⁺ cells were tested for the presence of total human IgM and IgG starting at 14 to 16 weeks after injection by using a standard ELISA (Alerchek). Sera from uninfected and HIV-1-infected mice were also examined for HIV-specific IgM and IgG at 6, 9, and 11 weeks postinfection by Western blotting. Supernatant from 293T cells [mock-transfected or transfected with HIV_NFN-SX(SL9)] was harvested at 48 h posttransfection and filtered through a 0.22-μm-pore-size filter. Equivalent volumes of samples (5 μl) were mixed with 2× Laemmli buffer containing 100 mM dithiothreitol and separated on an 8 to 16% Tris-glycine gel (Cambrex). Sera from mock-infected or HIV-1-infected mice (1:200 dilution) were used for Western blotting, followed by secondary horseradish peroxidase-conjugated goat anti-human IgG or IgM antibody (Sigma), and blots were developed with the enhanced chemiluminescence assay (Amersham, Arlington Heights, IL).

ELISPOT assays.

Enzyme-linked immunospot (ELISPOT) assays were performed as described by Helms et al. (13). In brief, immunospot plates (96 wells; Cellular Technology, Cleveland, OH) were coated overnight at 4°C with capture antibodies dissolved in phosphate-buffered saline (PBS) that were specific for gamma interferon. The wells were blocked with PBS-bovine serum albumin for 1 h and washed three times with PBS. Lymphocytes were plated in medium at 2 × 10⁵ to 3 × 10⁵ cells per well, and stimulant CD2/CD2R plus CD28 (BDIS) or HIV peptides (National Institutes of Health AIDS Research and Reference Reagent Program) were added to duplicate wells. After 24 h, the plates were washed and biotinylated detection antibodies were added for 12 h at 4°C, followed by streptavidin-horseradish peroxidase in PBS-bovine serum albumin-Tween for 2 h at room temperature. The spots were developed using AEC (3-amino-9-ethyl-carbazole) solution. Quantitative analysis of the ELISPOT assay results by using computerized image analysis to allow background subtraction and artifact exclusion was performed with the immunospot image analyzer (Cellular Technology). A positive response was defined as an increase of twofold or more above the background.

Statistics.

The nonparametric Mann-Whitney test was used to statistically analyze differences in the ratios of CD4⁺ T cells to CD8⁺ T cells in HIS- Rag2^−/−γ_c^−/− mice that were productively infected and the ratios of these cells in uninfected and nonproductively infected mice.

Reconstitution of human hematopoietic system in neonatal Rag2^−/−γ_c^−/− mice.

As we and others previously reported (9, 27), the injection of human CD34⁺ cells into neonatal Rag2^−/−γ_c^−/− mice leads to the reconstitution of a human hematopoietic system. The percentages of CD45⁺ cells (including CD19⁺ B cells and CD4⁺ and CD8⁺ T cells) in our present study performed at the University of California-Los Angeles ranged from 9.5 to 74.2% (average ± standard deviation, 48.5% ± 23%) at 8 to 32 weeks after injection with human CD34⁺ cells. Figure 1 shows results from a representative experiment comparing the immunophenotype of CD45⁺ cells in the peripheral blood of HIS-Rag2^−/−γ_c^−/− mice with that of CD45⁺ cells among normal human PBMC. The percentages of human cells expressing HIV-1 coreceptors CCR5 and CXCR4 were similar to those in normal human peripheral blood (Fig. 1). Ratios of CD4⁺ T cells to CD8⁺ T cells in the peripheral blood of the HIS-Rag2^−/−γ_c^−/− mice increased or were relatively stable over a 2- to 10-week time period after reconstitution was observed. Human lymphoid cells were found in the thymuses, spleens, and bone marrow of the mice, similar to the results in our previous report (9). These data indicate that these Rag2^−/−γ_c^−/− mice bred at different institutions show similar patterns of reconstitution after injection with human CD34⁺ fetal liver cells.

As we detected cells that expressed the B-cell marker CD19, we examined human immunoglobulin production in animals with reconstituted systems by measuring the total human IgG and IgM levels in the mouse sera by using a sandwich ELISA. The total human IgG levels ranged from 10 to 980 μg/ml in the mice with reconstituted systems, and the levels of total human IgM ranged from 5 to 116 μg/ml (data not shown). Thus, our mice with reconstituted HIS were capable of producing human IgG and IgM.

HIS-Rag2^−/−γ_c^−/− mice can be productively infected with small amounts of R5 HIV-1.

The infection of the HIS-Rag2^−/−γ_c^−/− mice with 5,000 TCID₅₀ of R5 HIV-1_NFN-SX(SL9) (24) resulted in productive infection in three of three animals as determined by coculturing (Table 1). HIV-1 replication was detected in peripheral blood, thymuses, and spleens. Bone marrow, tested from two animals, was also positive for HIV-1. HIV-1 infection was independent of the level of reconstitution of human CD45⁺ cells in the peripheral blood of the HIS mice immediately before HIV-1 infection. HIS-Rag2^−/−γ_c^−/− mice with proportions of CD45⁺ cells in the peripheral blood of less than or more than 50% were productively infected (Table 1). Likewise, the percentages of human CD4⁺ T cells in the peripheral blood before HIV-1 infection did not predict a productive HIV-1 infection (data not shown).

To determine whether even a 10-fold-smaller amount of HIV-1 could be used for the infection of HIS-Rag2^−/−γ_c^−/− mice, the mice were infected with 500 TCID₅₀ of R5 HIV-1_NFN-SX(SL9). Two out of five mice were found to be productively infected with HIV-1. As observed in the mice infected with the 10-fold-larger amount of HIV-1, peripheral blood and all lymphoid organs in productively infected mice were positive for HIV-1 by coculture (Table 1). The virus was undetectable in the mouse sera.

HIV-1-infected HIS-Rag2^−/−γ_c^−/− mice show a decrease in the ratios of CD4⁺ T cells to CD8⁺ T cells.

Immunophenotyping of peripheral blood and lymphoid organs of the HIV-1-infected and uninfected mice was done to determine the extent of CD4⁺-T-cell depletion among the human CD45⁺ cells. Four out of five productively infected mice showed a decline in ratios of CD4⁺ T cells to CD8⁺ T cells in the peripheral blood 2 to 11 weeks postinfection (Fig. 2; Table 2). In contrast, the CD4⁺-T-cell/CD8⁺-T-cell ratios in uninfected mice or mice injected with heat-inactivated virus did not decrease. There was also no decline in CD4⁺-T-cell/CD8⁺-T-cell ratios in HIV-1-inoculated mice that did not show productive infection (Fig. 2; Table 2). However, at weeks 4 to 6 postinfection, there was a statistically significant difference in the CD4⁺-T-cell/CD8⁺-T-cell ratios in mice that were productively infected with HIV-1 and the ratios in uninfected and nonproductively infected mice (P = 0.042). Although the patterns of change in CD4⁺-T-cell levels were not so consistent among the mice, the decline in CD4⁺-T-cell/CD8⁺-T-cell ratios was likely due to CD4⁺-T-cell depletion.

Immune responses in HIS-Rag2^−/−γ_c^−/− mice.

Immune responses to foreign antigens in this mouse model have been reported previously (27). We examined mice from experiments 2 and 3 for antigen-specific immune responses (mice from experiments 1 and 4 were not tested for antibody responses). We did not observe antibody responses specific to HIV-1, as assayed by Western blotting, either to viral proteins prepared from virions or to recombinant pr55 Gag. However, in four out of four HIV-1-inoculated animals that were not productively infected (mice 4, 5, 8, and 9), we observed an antibody response (IgM and IgG) directed to an unidentified 51-kDa antigen of fetal bovine serum present in virus preparations (representative data from mouse 5 are shown in Fig. 3). The antibody response was detectable starting at week 6 postinfection and continued until week 11 postinfection, when the mice were sacrificed. Antibody responses to the fetal bovine serum antigen were not detected in two out of two mice that were productively infected (mice 6 and 10) or in uninfected mice (mice 1, 3, and 7). It is noteworthy that antibody responses to the fetal bovine serum antigen were detected only in those inoculated animals in which T-cell depletion and productive viral infection were not observed, suggesting that active viral replication leading to the depletion of the lymphoid system inhibits the formation of functional immune responses.

Cellular T-cell responses were measured using gamma interferon ELISPOT assays. No cellular T-cell responses to HIV-1 Gag and Nef peptides were detected by gamma interferon ELISPOT assays (data not shown).