Homozygous HLA-A2.1 transgenic H-2Kb°Db° double knockout (KO) mice were created. Their potential to develop HLA-A2.1-restricted cytolytic responses was compared with that of their classical transgenic counterparts, which still express H-2Kb, Db molecules. On cell surfaces, both strains express similar amounts of chimeric (α1α2 domains of human, α3 cytoplasmic domains of mouse) HLA-A2.1 molecules in noncovalent association with mouse β2-microglobulin. Compared with mice that are totally deprived of histocompatibility class Ia molecules (H-2Kb°Db° double KO), the expression of HLA-A2.1 in transgenic/double KO mice resulted in sizeable increase in the periphery of CD8+ T cells with a normally diversified TCR repertoire. A biased education in favor of HLA-A2.1, ascribable to the absence of H-2 class Ia molecules, was evidenced in these transgenic/double KO mice by their improved capacity to mount HLA-restricted cytolytic responses, regardless of whether they were virally infected or injected with synthetic epitopic peptide. HLA class I transgenic, H-2 class Ia KO mice should represent useful animal models for the preclinical evaluation of vaccine formulations aiming at the induction of HLA class I-restricted CTL responses.

Since the feasibility of gene transfer between mammalian species was established (1) and the first HLA class I genes were isolated (2, 3), many laboratories have committed themselves to the creation of HLA class I transgenic mice (4, 5). These mice were expected to facilitate the study of HLA class I-restricted cytolytic responses and to be of interest for the design of vaccines aiming at the induction of CTLs specific for viral or tumoral Ags. Despite initial optimistic claims (6, 7), it was generally observed that HLA class I transgenic mice displayed poor usage of the human molecules (8, 9, 10, 11). By substituting the third domain of the heavy chain with a murine one, improved usage of HLA class I molecules was anticipated, because this domain is of key importance for the development of CD8-mediated stabilizing accessory interaction (12). Actually, transgenic mice expressing such chimeric molecules were successfully used for the study of HLA class I-restricted responses against certain viral and tumoral Ags (13, 14, 15, 16). Such improvement was not observed, however, with other HLA class I alleles (for example HLA-B7.1) and, even for those in which it was observed (HLA-A2.1 in particular), it remained incomplete, usually with preferential usage by mouse CTLs of H-2 class I molecules for Ag recognition (F.A.L., unpublished observation). As it became clear that this homospecific bias was not TCR-linked (17, 18) and could be overcome provided H-2-restricted responses were controlled, we decided to combine HLA class I transgenesis and H-2 class Ia gene invalidation to force the mouse CTLs to cooperate with HLA class I molecules.

We report here on the phenotypical and functional immunological characterization of mice lacking mouse H-2 class Ia (Kb and Db) molecules (which are the only H-2 class Ia molecules in the H-2b haplotype) and expressing a chimeric HLA-A2.1 heavy chain (α1α2 of human, α3 cytoplasmic domains of mouse) that is associated noncovalently with mouse β2-microglobulin (β2m).3

Homozygous double knockout (KO) (H-2Kb°Db°) mice that only express H-2 class Ib molecules were produced by crossing H-2Kb° and H-2Db° mice as described previously (19, 20). Homozygous HLA-A2.1 classical transgenic mice express, in addition to a full set of H-2 class Ia and b molecules, an HLA-A2.1 chimeric molecule (α1α2 domains of HLA-A2.1 and α3 cytoplasmic domains of H-2Kb) . These mice were obtained from Harlan Sprague-Dawley (Indianapolis, IN). Homozygous HLA-A2.1 transgenic/double KO (HLA-A2.1+, H-2Kb°Db°) mice were isolated by crossing double KO mice and classical HLA-A2.1 transgenic mice. In addition to the trangenic chimeric molecule, these mice also express a full set of H-2 class Ib but no H-2 class Ia molecules. All mice were bred in our animal facilities.

Untransfected and HHD-transfected RMA-S (TAP-deficient) as well as EL4 S3 Rob (mouse β2m-deficient) mice were used as targets in cytolytic assays. As described previously (19), HHD molecules are monochains with human β2m linked at their C terminus by a 15-aa peptidic arm to the N terminus of an HLA-A2.1 chimeric heavy chain (α1α2 domains of HLA-A2.1 and α3 cytoplasmic domains of H-2Db).

The expression of MHC class Ia (H-2Kb, H-2Db) or HLA-A2.1 was analyzed by indirect immunofluorescence using B8.24 3 (anti-H-2Kb), B22.249.R.19 (anti-H-2Db), BB7.2 (anti-HLA-A2) unlabeled mAb, and FITC-conjugated goat IgG F(ab′)2 anti-mouse Ig. The percentage of CD4+ and CD8+ T lymphocytes was determined by double staining using PE-labeled (CT-CD4) anti-mouse CD4 and FITC-labeled (CT-CD8β) or biotinylated (CT-CD8α) anti-mouse CD8 (Caltag Laboratories, South San Francisco, CA) mAb; the latter was detected with streptavidin-Perc-P (Becton Dickinson Immunocytometry Systems, San Jose, CA). Expression of the different Vβ TCRs was similarly analyzed using PE-labeled anti-CD8 mAb (Caltag) and purified, FITC-labeled Vβ2 (B.20.6), Vβ4 (KT.10.4), Vβ5.1,.2 (MR.9.4), Vβ6 (44.22), Vβ7 (TR 130), Vβ8.1,.2,.3 (F.23.1), Vβ9 (MR.10.2), Vβ10 (B.21.5), Vβ11 (RR3.15), Vβ12 (MR11.1), Vβ13 (MR12.4), Vβ14 (14/2), and Vβ17 (KJ.23.288.1) specific mAbs. RBC-depleted splenocytes of individual mice were enriched in T lymphocytes by wheat germ agglutinin (Sigma, St. Louis, MO) precipitation of B and NK cells as described previously (21). Cells (106 in 100 μl of PBS 1× with 0.02% sodium azide) were incubated for 30 min on ice with first layer mAb at saturating concentrations and, after three washings, incubated with the conjugates. A total of 10,000 paraformaldehyde-fixed cells per sample were subjected to cytofluorometric analysis on a FACScalibur (Becton Dickinson).

Acetone-fixed, rehydrated thymic cryosections were first incubated with either BB7.2 (anti-HLA-A2), B22.249.R.19 (anti-H-2Db), or B.1.23.2 (anti-HLA-B/C, negative control) mAb for 4 h at room temperature. After washes, cryosections were incubated overnight at room temperature with biotinylated goat anti-mouse Ig (Caltag). After washes, fixation of biotinilated Abs was revealed with FITC-conjugated streptavidin (Caltag). All rehydrations, reagent dilutions, and washes were performed using 1× PBS, 0.2% gelatin, and 0.1% Tween 20. Microscopic examinations were performed with a Leica (Wetzlar, Germany) TCS4D instrument. Photographs were taken with Ilford (Mobberley, U.K.) FP4 Plus 125 Asa film.

HLA-A2.1 classical and double KO transgenic mice were infected i.p. with 1000 hemagglutinating units of influenza A/PR/8/34. After 2 wk, 2.5 × 107 RBC-depleted splenocytes were restimulated in vitro in complete RPMI 1640 medium (10% FCS, 2 mM l-glutamine, 1 mM sodium pyruvate, 5.10−5 M 2-ME, 50 U/ml penicillin, and 50 μg/ml streptomycin) with an equal amount of syngeneic RBC-depleted splenocytes infected for 1.5 h with influenza A (10 hemagglutinating units for 106 cells) in FCS-free RPMI 1640 medium.

After 5 days, responder mice were individually tested in a standard cytolytic assay. Briefly, 106 cells that were uninfected or had been infected with 1000 hemagglutinating units of influenza A for 1.5 h in FCS-free RPMI 1640 medium and further incubated for 2 h in RPMI 1640 medium containing 5% FCS were subsequently labeled with 100 μCi of sodium [51Cr]chromate for 1.5 h at 37°C and washed three times. Cytolytic activity was determined in 4-h 51Cr release assays using V-bottom, 96-well plates containing 5 × 103 uninfected, influenza A matrix 58–66 (GILGFVFTL, Neosystem, Strasbourg, France) peptide-pulsed (10−6 M) or virus-infected target cells/well in the presence of effector cells from bulk cultures at different E:T ratios. Results are the mean of triplicates calculated as follows: 100 × ([experimental release − spontaneous release]/[total release − spontaneous release]), with maximal release being determined by the lysis of target cells in 1 M HCl.

The precursor frequencies of influenza virus-specific CTLs were estimated by limiting dilution analysis. Responder spleen cells derived from uninfected or influenza virus-infected mice were cultured in complete RPMI 1640 medium in 24 replicates at 6,250–200,000 spleen cells/well in round-bottom microplates. Stimulator influenza A virus-infected irradiated cells (1200 rad, 1.5 105 cells) were added to each well in T cell growth factor-supplemented medium. After 7 days of culture, each well was split and assayed for cytolytic activity on 51Cr-labeled, virus-infected, or peptide-pulsed uninfected target cells. Wells were considered positive when the 51Cr release exceeded the average spontaneous 51Cr release of 24 control wells from a parallel culture of uninfected mice by three SDs.

Each mouse was injected s.c. at the base of the tail with 100 μg of the HLA-A2.1-restricted epitopic peptide from HIV-1 reverse transcriptase (HIV reverse transcriptase 476–484 ILKEPVHGV, Neosystem) peptide, with or without 140 μg of an IAb-restricted helper peptide from the hepatitis B core protein (hepatitis B virus core 128–140 TPPAYRPPNAPIL, Neosystem) emulsified (v/v) in IFA (Difco, Detroit, MI). After 7 days, spleen cells were restimulated in vitro using irradiated (5000 rad), peptide-loaded (5.106 cells/ml, 10 μg/ml peptide, 2 h at room temperature in FCS-free RPMI 1640 medium), LPS-induced (25 μg/ml LPS, 7 μg/ml dextran sulfate in complete RPMI 1640 medium, 48 h of culture) lymphoblasts from syngeneic mice. After 6 days, lymphocytes were tested for cytolytic activity against HHD-transfected RMA-S targets loaded with relevant or negative control peptide.

Purified T splenocytes of classical transgenic (HLA-A2.1+, H-2Kb+H-2Db+), transgenic/double KO (HLA-A2.1+, H-2Kb°Db°), and control double KO (HLA-A2.1, H-2Kb°Db°) mice were analyzed by FACS in an indirect immunofluorescence assay using B8.24.3 (anti H-2Kb), B22.249 R9 (anti H-2Db), or BB7.2 (anti HLA-A2.1) first layer mAb.

The results are illustrated in Fig. 1 A. The absence of H-2Kb and Db molecules in double KO and expression of HLA-A2.1 molecules in transgenic animals were documented, with no significant difference in the levels of expression of the transgenic molecules, whether coexpressed or not with H-2 class I molecules. Two additional points are worthy of notice. First, the expression of HLA-A2.1 molecules was lower (roughly by a log) than that of H-2Kb/Db molecules. Second, we observed constantly (whichever the sex and generation of mice tested) the existence of an HLA-A2.1 negative cell population that was not limited to T lymphocytes, as also documented in B lymphocytes and dendritic cells (data not shown); this finding suggests that this population exists in all tissues. The size of this population, which is certainly explained by a variegation phenomenon (22), was variable between mice of the same litter, ranging from 10 to 40% of the total cells. Similar results were obtained by assaying cells with a monomorphic anti-HLA class I molecules mAb (H.F., data not shown).

FIGURE 1.

Phenotypical characterization of H-2Kb°Db° double KO mice, HLA-A2.1 transgenic/H-2Kb°Db° double KO mice, and HLA-A2.1 classical transgenic mice. A, Expression of HLA-A2.1, H-2Kb, and H-2Db molecules was detected on unactivated purified T splenocytes with BB7.2, B8.24.3, and B22.249.R.19 mAbs, respectively, and with FITC-conjugated goat IgG F(ab′)2 anti-mouse Ig (GAM). Negative controls did not have the first mAb. Results are expressed in fluorescence intensity in arbitrary units (x-axis, log scale) and in relative cell numbers (y-axis). B, Indirect immunofluorescence analysis of thymic cryosections from classical (upper row) and double KO (lower row) HLA-A2.1 transgenic mice using anti-H-2Db (1 and 4), anti-HLA-A2 (2 and 5), and negative control (3 and 6) mAb. The amplification of the fluorescent signal was the same for the cryosections incubated with the same mAb; however, this amplification was stronger for cryosections 2 and 5 (anti-HLA-A2) than for 1 and 4 (anti-H-2Db). Note the usual labeling pattern for histocompatibility class Ia molecules (strong for the medulla (M), faint and reticular for the cortex (C)) observed with both anti-HLA-A2 and anti-H-2Db mAb. C, Percentages of CD4+ and CD8+ T lymphocytes. Unactivated purified T splenocytes were double-stained with PE-labeled anti-CD4 mAb (x-axis) and biotinylated anti-CD8 mAb detected with streptavidin-Perc-P (y-axis). The percentages of CD4+ (bottom right) or CD8+ (upper left) are indicated.

FIGURE 1.

Phenotypical characterization of H-2Kb°Db° double KO mice, HLA-A2.1 transgenic/H-2Kb°Db° double KO mice, and HLA-A2.1 classical transgenic mice. A, Expression of HLA-A2.1, H-2Kb, and H-2Db molecules was detected on unactivated purified T splenocytes with BB7.2, B8.24.3, and B22.249.R.19 mAbs, respectively, and with FITC-conjugated goat IgG F(ab′)2 anti-mouse Ig (GAM). Negative controls did not have the first mAb. Results are expressed in fluorescence intensity in arbitrary units (x-axis, log scale) and in relative cell numbers (y-axis). B, Indirect immunofluorescence analysis of thymic cryosections from classical (upper row) and double KO (lower row) HLA-A2.1 transgenic mice using anti-H-2Db (1 and 4), anti-HLA-A2 (2 and 5), and negative control (3 and 6) mAb. The amplification of the fluorescent signal was the same for the cryosections incubated with the same mAb; however, this amplification was stronger for cryosections 2 and 5 (anti-HLA-A2) than for 1 and 4 (anti-H-2Db). Note the usual labeling pattern for histocompatibility class Ia molecules (strong for the medulla (M), faint and reticular for the cortex (C)) observed with both anti-HLA-A2 and anti-H-2Db mAb. C, Percentages of CD4+ and CD8+ T lymphocytes. Unactivated purified T splenocytes were double-stained with PE-labeled anti-CD4 mAb (x-axis) and biotinylated anti-CD8 mAb detected with streptavidin-Perc-P (y-axis). The percentages of CD4+ (bottom right) or CD8+ (upper left) are indicated.

Close modal

Thymic cryosections of classical and double KO HLA-A2.1 transgenics were comparatively analyzed using BB7.2, B22.249.R9, or negative control B.1.23.2 mAb. As illustrated in Fig. 1 B, similar levels and patterns of expression of the transgenic molecules were observed (strong in the medulla, faint and reticular in the cortex) whether coexpressed or not with H-2 class Ia molecules.

Purified T splenocytes were first double-stained with either FITC-conjugated anti-CD8β or biotinylated anti-CD8α and PE-conjugated anti-CD4 mAb to determine the percentages of CD8+ T cells in the periphery of classical HLA-A2.1 transgenic, HLA-A2.1 transgenic/double KO, and control double KO mice.

The results (identical with anti-CD8α and anti-CD8β mAb) are illustrated in the case of anti-CD8α in Fig. 1 C. Compared with classical transgenic mice, which possessed a normal (25.7) percentage of CD8+ peripheral T cells, a profound, however incomplete, reduction of this percentage was observed in double KO mice with a residual (2–3%) population of CD8+ T cells. The size of the CD8+ T cell population was augmented (≤6–7%) in HLA-A2.1 transgenic/double KO mice. When several mice were individually tested for both the expression of HLA-A2.1 molecules and CD8+ T cell numbers, no relationship was observed between the relative sizes of the HLA-A2.1 positive and negative populations of cells and the percentage of peripheral CD8+ T cells, with the latter always remaining in the 6–7% range.

Next, the Vβ diversity of the TCR expressed by CD8+ T cells was analyzed in a similar immunofluorescent assay in the three types of mice. The results are illustrated in Fig. 2. Almost superimposable patterns of Vβ expression were documented for classical transgenic mice, transgenic/double KO mice, and control double KO mice with, in the later case, a moderate and isolated increase of the Vβ5 population of T cells compared with HLA-A2.1 transgenics. Thus, the absence of H-2 class Ia molecules and expression of HLA-A2.1 transgenic molecules did not significantly alter the peripheral CD8+ T cell repertoire in terms of Vβ diversity.

FIGURE 2.

Peripheral CD8+ TCR repertoire. Double stainings were performed on unactivated purified T splenocytes with FITC-labeled Vβ-specific and PE-labeled anti-CD8 mAb. Results are the mean of at least three different experiments on six mice tested individually and are expressed as the percentage of total CD8+ T cells.

FIGURE 2.

Peripheral CD8+ TCR repertoire. Double stainings were performed on unactivated purified T splenocytes with FITC-labeled Vβ-specific and PE-labeled anti-CD8 mAb. Results are the mean of at least three different experiments on six mice tested individually and are expressed as the percentage of total CD8+ T cells.

Close modal

Classical and double KO HLA-A2.1 transgenic mice were infected i.p. with influenza A viruses. After 15 days, splenocytes were restimulated in vitro for 5 days; HLA-A2.1-restricted, influenza-specific cytolytic activity was evaluated in a 51Cr release assay using either influenza-infected or peptide-pulsed (influenza A matrix, 58–66) HLA-A2.1-transfected EL4 S3 Rob target cells.

The results are illustrated in Fig. 3. Specific although moderate lysis of either influenza-infected or matrix peptide-pulsed HLA-A2.1-transfected target cells was observed with classical HLA-A2.1 transgenic effector cells. More significant HLA-A2.1-restricted influenza-specific lytic activity was induced under similar experimental conditions in HLA-A2.1 transgenic/double KO mice; this lytic activity was, for the most part, directed at the matrix peptide, which is also the immunodominant T cell epitope in HLA-A2.1 humans. Similar results were obtained with all mice tested and suggested that the frequency of peripheral CD8+ T cells susceptible to be mobilized against an antigenic peptide presented by HLA-A2.1 molecules was significantly higher in transgenic/double KO mice than in classical transgenic mice.

FIGURE 3.

Influenza A virus-specific, HLA-A2.1-restricted CTL responses of HLA-A2.1 transgenic classical and double (H-2Kb°Db°) KO mice. At 2 wk postinfection, CTLs were restimulated in vitro for 5 days and 51Cr release assays were performed. Targets were untransfected (○) or HHD-transfected (•) β2m-deficient EL4 S3 cells. Targets were either uninfected, infected, or influenza A matrix (58–66) peptide-pulsed, as indicated. E:T ratios were 60:1, 20:1, 6.6:1, and 2.2:1. Spontaneous release was <15% for all targets.

FIGURE 3.

Influenza A virus-specific, HLA-A2.1-restricted CTL responses of HLA-A2.1 transgenic classical and double (H-2Kb°Db°) KO mice. At 2 wk postinfection, CTLs were restimulated in vitro for 5 days and 51Cr release assays were performed. Targets were untransfected (○) or HHD-transfected (•) β2m-deficient EL4 S3 cells. Targets were either uninfected, infected, or influenza A matrix (58–66) peptide-pulsed, as indicated. E:T ratios were 60:1, 20:1, 6.6:1, and 2.2:1. Spontaneous release was <15% for all targets.

Close modal

A split-well limiting dilution assay was performed to address this question. Responder cells of mice previously infected with influenza viruses were restimulated in vitro by coculture of decreasing numbers of effector cells (from 200,000 to 6,250 per well) with constant amounts of autologous influenza-infected, γ-irradiated stimulating cells. After 7 days, cells from each well were split and tested against influenza-infected or matrix peptide-pulsed HLA-A2.1-transfected 51Cr-labeled target cells. The results are illustrated in Fig. 4. Whereas influenza-specific lytic activity could only be observed with HLA-A2.1 classical transgenic mice in wells in which 50,000 effectors were seeded, clear lytic activity could be observed in wells in which as little as 6,250 HLA-A2.1 transgenic/double KO mouse effectors were seeded. Notice that, for both mice, the lytic activity was essentially directed at the matrix peptide since perfect square diagonals are obtained in the diagram plot. Thus, for this matrix peptide, it appeared that HLA-A2.1 transgenic/double KO mice had roughly 10 times more HLA-A2.1-restricted CTL precursors than classical transgenic animals.

FIGURE 4.

Split-well limiting dilution assay. At 3 wk after influenza infection of HLA-A2.1 transgenic/double H-2Kb°Db° KO mice (•) and HLA-A2.1 classical transgenic mice (○), responder spleen cells were cocultured for 7 days with infected syngeneic γ-irradiated spleen cells in T cell growth factor-supplemented medium at a concentration of 6,250 (a), 12,500 (b), 25,000 (c), 50,000 (d), 100,000 (e), and 200,000 (f) effectors per well, with 24 replicates being tested for each effector concentration. Wells were split and tested against infected or peptide-pulsed HHD-transfected EL4 S3 cells. Each point of the graph shows the levels of cytotoxicity in absolute cpm of 51Cr release. Wells were considered positive when the radioactivity exceeded the spontaneous release by three SDs as determined in parallel cultures and by assay with unprimed effector cells as indicated by the lines on the graphs. The split-well experiment was done twice with two mice in each group individually tested; similar results were seen in all cases.

FIGURE 4.

Split-well limiting dilution assay. At 3 wk after influenza infection of HLA-A2.1 transgenic/double H-2Kb°Db° KO mice (•) and HLA-A2.1 classical transgenic mice (○), responder spleen cells were cocultured for 7 days with infected syngeneic γ-irradiated spleen cells in T cell growth factor-supplemented medium at a concentration of 6,250 (a), 12,500 (b), 25,000 (c), 50,000 (d), 100,000 (e), and 200,000 (f) effectors per well, with 24 replicates being tested for each effector concentration. Wells were split and tested against infected or peptide-pulsed HHD-transfected EL4 S3 cells. Each point of the graph shows the levels of cytotoxicity in absolute cpm of 51Cr release. Wells were considered positive when the radioactivity exceeded the spontaneous release by three SDs as determined in parallel cultures and by assay with unprimed effector cells as indicated by the lines on the graphs. The split-well experiment was done twice with two mice in each group individually tested; similar results were seen in all cases.

Close modal

The results obtained infecting mice with influenza viruses could have meant that HLA-A2.1 classical transgenics did develop influenza-specific but H-2-restricted CTL responses that could preempt the HLA-A2.1-restricted ones. Immunizing mice with synthetic peptides binding specifically to HLA-A2.1 molecules would rule out such a possibility. Initial attempts to induce CTLs with the matrix peptide failed. Therefore, we selected an HIV-1-derived peptide (HIV-1 reverse transcriptase 476–484) that we knew to be an efficient CTL-inducer when injected into HHD mice in IFA (A.U.-V., unpublished observation). HLA-A2.1 classical transgenic and transgenic/double KO mice were injected with either synthetic HIV-1 reverse transcriptase 476–484 peptide alone or mixed with an H-2 I-Ab-restricted, hepatitis B-derived (hepatitis B virus core 128–140) helper peptide. After 7 days, splenocytes were restimulated in vitro for 5 days; lytic activity was evaluated in a 51Cr release assay.

The results are summarized in Table I. None of the 16 classical transgenic mice tested developed CTL responses against the HIV-1 reverse transcriptase 476–484 peptide, regardless of whether or not they had been coimmunized or not with the helper peptide. Similarly, transgenic/double KO mice immunized with HIV-1 reverse transcriptase 476–484 alone did not develop CTL responses. In contrast, six of seven tested mice responded to the HIV-1 reverse transcriptase 476–484 peptide when it was associated with the hepatitis B virus core 128–140 helper peptide.

Table I.

Cytolytic responses to HIV-1 reverse transcriptase 476-484 peptidea

MiceImmunizationResponses
A2A2Kb+/+, Kb+, Db+ HIV-1 reverse transcriptase 476–484 0 /9 
 HIV-1 reverse transcriptase 476–484+ hepatitis B virus core 128–140 0 /7 
A2A2Kb+/+, K, D HIV-1 reverse transcriptase 476–484 0 /10 
 HIV-1 reverse transcriptase 476–484+ hepatitis B virus core 128–140 6 /7 (15, 19, 26, 42, 100) 
MiceImmunizationResponses
A2A2Kb+/+, Kb+, Db+ HIV-1 reverse transcriptase 476–484 0 /9 
 HIV-1 reverse transcriptase 476–484+ hepatitis B virus core 128–140 0 /7 
A2A2Kb+/+, K, D HIV-1 reverse transcriptase 476–484 0 /10 
 HIV-1 reverse transcriptase 476–484+ hepatitis B virus core 128–140 6 /7 (15, 19, 26, 42, 100) 
a

HLA-A2.1 classical transgenic and HLA-A2.1 transgenic/double KO mice were injected s.c. with HIV-1 reverse transcriptase 476–484 peptide in IFA mixed or not with hepatitis B virus core 128–140 helper peptide. Following in vitro restimulation, effector cells were tested for specific cytolytic activity against peptide-loaded HHD-transfected RMA-S cells. The numbers of responder vs tested mice are given; the maximum levels of lysis for each individual responder mouse are indicated in parentheses and usually correspond to a 100:1 E:T ratio.

HLA class I transgenic mice were expected to provide us with a convenient animal model for the search of new T cell epitopes, the optimization of their immunogenicity, and the preclinical evaluation of the different strategies that can be proposed to induce CTL responses of potential vaccine interest. Classical transgenic mice, which still express their own H-2 class Ia and b molecules, have not fulfilled that goal satisfactorily. Because we realized that part of the problem was linked to the preferential development of H-2-restricted responses (17), we undertook the creation of strains of mice in which mouse class I genes would have been inactivated by homologous recombination (19, 20). The expression of HLA class I molecules in such mice was expected to force mouse CD8+ T cells to make use of the human molecules both at the thymic educational and peripheral effector levels.

The results presented in this paper establish that HLA-A2.1 transgenics lacking H-2 class Ia molecules have a larger (roughly 10-fold) HLA-A2.1 educated CD8+ T cell repertoire than classical transgenics. Thus, simultaneous expression of HLA and H-2 class Ia molecules results in reduced HLA education of mouse CD8+ T cells, despite the fact that such coexpression does not reduce the amount of HLA molecules on cell surfaces. In the absence of species-specific structural features of the variable segments of the TCR α- and β-chains (23), the most likely explanation of this reduced HLA education probably lies at the CD8 molecule level. It has been crystallographically documented that human CD8 molecules interact with the third but also the second and β2m domains of HLA-A2.1 molecules (24, 25). Therefore, mouse CD8 molecules might have retained a higher affinity for bona fide H-2 class Ia than HLA-A2.1 molecules, even with a mouse α3 domain. Assuming that thymic selection is a saturable process, this would bias education in favor of H-2 molecules in classical HLA-transgenics.

Homozygous expression of chimeric (mouse α3) HLA-A2.1 molecules in a H-2 Kb°Db° double KO context resulted in a sizeable increase (roughly a doubling) of the peripheral CD8+ T cell population size. Compared with the 30–40% of CD8+ T cells (among splenic T lymphocytes) usually found in wild-type C57BL/6 mice, it appears that the CD8+ T cell selection capacity of the chimeric HLA-A2.1 molecules in these transgenics remains relatively limited. Lower surface expression of the transgenic molecules relative to H-2 class Ia molecules and conformational alterations due to their heterospecific association with mouse β2m could account for such a limitation (26). However, this seems unlikely for two reasons. First, similar numbers of CD8+ T cells were found in the periphery of other independently derived, HLA-A2.1 transgenic/H-2 class I KO mice with an even lower cell surface expression of the transgenic molecules than documented in the strain of mice considered in this report. Thus, the HLA-A2.1 expression in the latter is above the level required for such molecules to express their full educational potential. Second, similar levels of peripheral CD8+ T cells are found in mice expressing the same chimeric (mouse α3) HLA-A2.1 molecules but in an H-Kb°Db°/mouse β2m°/human β2m+/+ context (H.F., unpublished observation). Therefore, we favor the possibility (as discussed above) that partial restoration of the CD8+ T cell pool reflects a suboptimal interaction between mouse CD8 and chimeric HLA-A2.1 (mouse α3) molecules.

Although limited, the number of peripheral CD8+ T cells in HLA-A2.1/double KO mice suffices for these mice to be a useful animal model for the study of HLA-A2.1-restricted CTL responses. It is conceivable, however, that with some other class I alleles the α3 substitution, even combined with H-2 class Ia gene destruction, will not suffice for a workable education of CD8+ T cells, in which case additional transgenesis with human CD8α and possibly CD8β genes would have to be considered.

We thank P. Ave for preparing thymus cryosections, J.-C. Manuguerra for providing influenza viruses, E. Perret for microscopic analysis, and N. Sauzet for secretarial assistance.

1

This work was supported by the Ligue Nationale contre le Cancer, the Association pour la Recherche contre le Cancer, the Société Française d’Expérimentation Animale, and the Institut Pasteur. A.U.-V. was the recipient of a fellowship from the Fondation pour la Recheche Redicale-Sidection.

3

Abbreviations used in this paper: β2m, β2-microglobulin; KO, knockout.

1
Palmiter, R. D., R. L. Brinster, R. E. Hammer, M. E. Trumbauer, M. G. Rosenfeld, N. C. Birnberg, R. M. Evans.
1982
. Dramatic growth of mice that develop from eggs microinjected with metallothionein-growth hormone fusion genes.
Nature
300
:
611
2
Barbosa, J. A., M. E. Kamarck, P. A. Biro, S. M. Weissman, F. H. Ruddle.
1982
. Identification of human genomic clones coding the major histocompatibility antigens HLA-a2 and HLA-B7 by DNA-mediated gene transfer.
Proc. Natl. Acad. Sci. USA
79
:
6327
3
Lemonnier, F. A., M. Malissen, P. Golstein, P. Le Bouteiller, N. Rebaï, M. Damotte, D. Birnbaum, D. Caillol, J. Trucy, B. R. Jordan.
1982
. Expression of human class I histocompatibility antigens at the surface of DNA-transformed mouse L cells.
Immunogenetics
16
:
355
4
Pérarnau, B., F. A. Lemonnier.
1997
. Immunologically humanized transgenic mice. L. A. Herzenberg, and D. M. Weir, and L. A. Herzenberg, and C. Blackwell, eds. 5th Ed.In
Weir’s Handbook of Experimental Immunology
Vol. 4
:
146
Blackwell Science, Inc, Malden, MD.
5
Arnold, B., G. J. Hämmerling.
1991
. MHC class-I transgenic mice.
Annu. Rev. Immunol.
9
:
297
6
Kievits, F., P. Ivanyi, P. Krimpenfort, A. Berns, H. L. Ploegh.
1987
. HLA-restricted recognition of viral antigens in HLA transgenic mice.
Nature
329
:
447
7
Kievits, F., W. Lokhorst, P. Ivanyi. Abnormal anti-viral immune response in mice is corrected in HLA-B27.2-transgenic mice.
Eur. J. Immunol.
20
:
1990a
1189
8
Barra, C., B. Pérarnau, P. Gerlinger, M. Lemeur, A. Gillet, P. Gibier, F. Lemonnier.
1989
. Analysis of the anti-HLA-Cw3 CTL response of HLA-B7 × human β2m double transgenic mice.
J. Immunol.
143
:
3117
9
Engelhard, V. H., E. Lacy, J. P. Ridge.
1991
. Influenza A-specific, HLA-A2.1-restricted CTLs from HLA-A2.1 transgenic mice recognize fragments of the M1 protein.
J. Immunol.
146
:
1226
10
Epstein, H., R. Hardy, J. S. May, M. H. Johnson, N. Holmes.
1989
. Expression and function of HLA-A2.1 in transgenic mice.
Eur. J. Immunol.
19
:
1575
11
Le, A.-X. T., E. J. Bernhard, M. J. Holterman, S. Strub, E. Lacy, V. Engelhard.
1989
. Cytotoxic T cell responses in HLA-A2.1 transgenic mice: recognition of HLA alloantigens and utilization of HLA-A2.1 as a restriction element.
J. Immunol.
142
:
1366
12
Connolly, J. M., T. A. Potter, E.-M. Wormstall, T. H. Hansen.
1988
. The Lyt-2 molecule recognizes residues in the class I α3 domain in allogeneic cytotoxic T cell responses.
J. Exp. Med.
168
:
325
13
Vitiello, A., D. Marchesini, J. Furze, L. A. Sherman, R. W. Chesnut.
1991
. Analysis of the HLA-restricted influenza-specific cytotoxic T lymphocyte response in transgenic mice carrying a chimeric human-mouse class I major histocompatibility complex.
J. Exp. Med.
173
:
1007
14
Wentworth, P., A. Vitiello, J. Sidney, E. Keogh, R. Chesnut, H. Grey, A. Sette.
1996
. Differences and similarities in the A2.1-restricted cytotoxic T cell repertoire in humans and human leukocyte antigen-transgenic mice.
Eur. J. Immunol.
26
:
97
15
Shirai, M., T. Arichi, M. Nishioka, T. Nomura, K. Ikeda, K. Kawanishi, V. H. Engelhard, S. M. Feinstone, J. A. Berzofsky.
1995
. CTL responses of HLA-A2.1-transgenic mice specific for hepatitis C viral peptides predict epitopes for CTL of humans carrying HLA-A2.1.
J. Immunol.
154
:
2733
16
Theobald, M., J. Biggs, J. Hernandez, J. Lustgarten, C. Labadie, L. A. Sherman.
1997
. Tolerance to p53 by A2.1-restricted cytotoxic T lymphocytes.
J. Exp. Med.
185
:
833
17
Barra, C., H. Gournier, Z. Garcia, P. N. Marche, E. Jouvin-Marche, P. Briand, P. Filippi, F. A. Lemonnier.
1993
. Abrogation of H-2-restricted responses and efficient recognition of HLA-A3 molecules in DBA/2 HLA/A24 responder mice.
J. Immunol.
150
:
3681
18
Man, S., J. P. Ridge, V. Engelhard.
1994
. Diversity and dominance among TCR recognizing HLA-A2.1+ influenza matrix peptide in human class I transgenic mice.
J. Immunol.
153
:
4458
19
Pascolo, S., N. Bervas, J. M. Ure, A. G. Smith, F. A. Lemonnier, B. Pérarnau.
1997
. HLA-A2.1-restricted education and cytolytic activity of CD8+ T lymphocytes from β2-microglobulin (β2m) HLA-A2.1 monochain transgenic H-2Db β2m double knockout mice.
J. Exp. Med.
185
:
2043
20
Pérarnau, B., M. F. Saron, B. Reina San Martin, N. Bervas, H. Ong, M. Soloski, A. Smith, J. Ure, J. E. Gairin, F. Lemonnier.
1999
. Single H2Kb, H2Db and double H2KbDb knockout mice: peripheral CD8+ T cell repertoire and anti-lymphocytic choriomeningitis virus cytolytic responses.
Eur. J. Immunol.
29
:
1243
21
Saron, M.-F., P. Truffa-Bachi, J.-C. Guillon.
1989
. Rapid enrichment of mouse natural killer cells by use of wheat germ agglutinin.
J. Immunol. Methods
59
:
151
22
Johansson, M. H., C. Bieberich, G. Jay, K. Kärre, P. Höglund.
1997
. Natural killer cell tolerance in mice with mosaic expression of major histocompatibility complex class I transgene.
J. Exp. Med.
186
:
353
23
Wilson, R. K., E. Lai, P. Concannon, R. K. Barth, L. E. Hood.
1988
. Structure, organisation, and polymorphism of murine and human T-cell receptor α and β chain gene families.
Immunol. Rev.
101
:
149
24
Sun, J., D. J. Leahy, P. B. Kavathas.
1995
. Interaction between CD8 and major histocompatibility complex (MHC) class I mediated by multiple contact surfaces that include the α2and α3 domains of MHC class I.
J. Exp. Med.
182
:
1275
25
Gao, G. F., J. Tormo, U. C. Gerth, J. R. Wyer, A. J. McMichael, D. I. Stuart, J. I. Bell, E. Y. Jones, B. K. Jakobsen.
1997
. Crystal structure of the complex between human CD8α(α) and HLA-A2.
Nature
387
:
630
26
Fukazawa, T., E. Hermann, M. Edidin, J. Wen, F. Huang, H. Kellner, J. Floege, D. Farahmandian, K. M. Williams, D. T. Yu.
1994
. The effect of mutant β2-microglobulins on the conformation of HLA-B27 detected by antibody and by CTL.
J. Immunol.
153
:
3543