Human CD3γ, but not CD3δ, haploinsufficiency differentially impairs γδ versus αβ surface TCR expression

CD3G (γ^+/−) or CD3D (δ^+/−) haploinsufficient donors were uniformly healthy and showed abundant peripheral blood T lymphocytes with an essentially normal phenotype (Table 1). However, total T cell numbers were consistently lower than controls (Figure 1A) which correlated with a partial impairment of lymphocyte function (Table 1).

We have previously observed in γ^−/− individuals that CD3 expression levels are overestimated when T cells are defined using antibodies against TCR-associated epitopes [7], such as BMA031 (for TCRαβ) or Immu510 (for TCRγδ). To avoid a similar bias in haploinsufficient donors, TCR-independent electronic gates were first defined in order to identify αβ or γδ T cell subsets (Figure 1B). The results indicated that CD3⁺ cells within CD4⁺ or CD8^bright lymphocytes were >98% αβ T cells, whereas CD3⁺ double negative (DN) lymphocytes were 78 ± 6% γδ T cells. Accordingly, αβ and γδ T cells were gated as CD4⁺/CD8^bright and DN cells, respectively, for further analyses. Using several CD3-specific antibodies, analysis of surface TCR expression consistently showed reduced antibody binding in γ^+/− and δ^+/− T lymphocytes as compared to normal controls (50-90% as judged by their relative mean fluorescence intensity, Figure 2A, B). These results were confirmed in family members of two newly reported patients with a leaky mutation in CD3D (termed δ^+/leaky) [8]. Consistent with their relatively higher CD3δ content as compared to δ^+/− donors, δ^+/leaky donors showed a milder, but nevertheless clear decrease in surface TCR expression (Figure 2A, B). In order to establish if these results were associated with reduced availability of each CD3 chain, we measured intracellular CD3γ (iCD3γ) or CD3δ (iCD3δ) by flow cytometry in haploinsufficient γ^+/− and δ^+/leaky donors. The results showed that this was indeed the case (Figure 2C), confirming previous reports of decreased CD3γ protein in haploinsufficient donors [3].

Serial dilutions of CD3 mAb further confirmed the findings above (Figure 2D), since the reduced binding to γ^+/− T cells persisted in saturation conditions, but it was gradually lost near the endpoint, supporting the existence of less CD3 binding sites [9].

From these results we conclude that human CD3G or D haploinsufficient donors show reduced binding of CD3-specific mAb to the TCR of their γδ and αβ T cells.

Analysis of CD3 mAb surface binding to αβ and γδ T cells with the different CD3G and CD3D genotypes, relative to normal controls, revealed that binding of CD3 mAb to γ^+/− γδ T cells was unexpectedly poor (55 ± 3%) as compared with γ^+/− αβ T cells (82 ± 8%, Figure 2A). This discordant reduction was specific for γ^+/− donors, as it was not observed in δ^+/− or δ^+/leaky donors. Further support for this discordant reduction was provided by the γδ versus αβ CD3 expression ratio, which is normally 1.9 ± 0.22 [10, 11] but becomes significantly less in γ^+/− donors only (Figure 3). Taken together, these results indicate that normal surface γδ TCR expression in humans is more critically dependent on the relative availability of CD3γ, but not CD3δ, than that of the αβ TCR.

After obtaining informed consent following IRB authorization (Ethics Committee for Clinical Investigation of Clínico San Carlos Hospital, Madrid), we studied nine healthy heterozygous carriers of mutations in CD3γ (γ^+/−) [3, 7] of Spanish or Turkish descent and six healthy heterozygous carriers of mutations in CD3δ (δ^+/−, δ^+/leaky) [8, 16, 17] of Japanese, Mennonite or Colombian descent. Normal donors (termed^+/+) were used as controls. Their lymphocyte immunophenotype and functional features are summarized in Table 1. PBL were isolated by Ficoll-Hypaque (GE Healthcare) gradients and resuspended in RPMI 1640 medium (Gibco) supplemented with 10% FCS (PAA Laboratories), 1% L-glutamine and antibiotics-antymitotic (100 units/ml of penicillin G, 100 μg/ml of streptomycin sulfate, and 0.25 μg/ml of amphotericin B) from Gibco.

The expression of different surface markers was studied by flow cytometry using standard procedures in fresh whole blood or isolated PBL [18]. For intracellular stainings cells were fixed and permeabilized as previously reported [19].

The following anti-human mAb were used: anti-CD3ε (S4.1) from Caltag Laboratories (now Invitrogen), anti-CD3εγ/εδ (UCHT-1), anti-TCR αβ (BMA031), and anti-TCRγδ (Immu510) from Beckman Coulter, anti-CD3ε (SK7), anti-CD4 (Leu2a), anti-CD8 (Leu3a), anti-TCRγδ (11F2), and anti-CD8 (SK1) from BD Biosciences. Anti-CD3εγ/εδ (F101.01) hybridoma supernatant and anti-CD3δ (APA1/2) ascitic fluid were a generous gift from Dr. B. Rubin (CHU de Purpan, France). TG5 (an anti-CD3γ rabbit antiserum raised against the CD3γ C-terminal peptide GLQGNQLRRN) was kindly provided by D. Alexander (Babraham Institute, U.K.). The mAb were FITC-, PE-Cy5 or PE-conjugated, or purified, and for the latter a PE-conjugated goat anti-mouse IgG (H + L) or anti-rabbit (H + L) from Caltag Laboratories was used as a secondary reagent. Background fluorescence was defined with an isotype-matched irrelevant mAb from Caltag Laboratories. For comparative stainings we used the mean fluorescence intensity (MFI), defined as the average fluorescence value of the corresponding mAb referred to the logarithmic scale of fluorescence intensity along the x-axis of the histograms. Data were analyzed with FlowJo software (TreeStar).

1×10⁵ PBLs were placed in round-bottom microtitre wells and stimulated with 10 μg/ml anti-human CD3 (UCHT-1) or 10 μg/ml PHA. After 3 days of in vitro culture, wells were individually pulsed with 1 μCi of ³H-TdR (Amersham, Buckinghamshire, U.K.) for another 16 to 18 h and harvested onto glass fiber filters. Thymidine incorporation into cellular DNA was evaluated as cpm in a scintillation βcounter (Packard, Meriden, CT).

For CFSE (carboxyfluorescein diacetate succinimidyl ester) dilution experiments, cells were labeled with 1 μM CFSE in PBS for 10 min at 37°C at day 0, washed twice in cold PBS, plated and stimulated as above. CFSE dilution was subsequently determined by flow cytometry within CD3⁺ lymphocytes.

Cytotoxicity was measured using the nonradioactive Cytotoxicity Detection kit LDH (Roche). Cells were cocultured in a 96 V-well plates for 4 h at 25: 1 (Effector: Target) ratios and the percentage of specific lysis was determined from the amount of lactate dehydrogenase activity detected in culture supernatants.