Regulatory function of cytomegalovirus-specific CD4+CD27-CD28- T cells

Abstract

CMV infection is characterized by high of frequencies of CD27-CD28- T cells. Here we demonstrate that CMV-specific CD4+CD27-CD28- cells are regulatory T cells (TR). CD4+CD27-CD28- cells sorted from CMV-stimulated PBMC of CMV-seropositive donors inhibited de novo CMV-specific proliferation of autologous PBMC in a dose-dependent fashion. Compared with the entire CMV-stimulated CD4+ T-cell population, higher proportions of CD4+CD27-CD28- TR expressed FoxP3, TGFbeta, granzyme B, perforin, GITR and PD-1, lower proportions expressed CD127 and PD1-L and similar proportions expressed CD25, CTLA4, Fas-L and GITR-L. CMV-CD4+CD27-CD28- TR expanded in response to IL-2, but not to CMV antigenic restimulation. The anti-proliferative effect of CMV-CD4+CD27-CD28- TR significantly decreased after granzyme B or TGFbeta inhibition. The CMV-CD4+CD27-CD28- TR of HIV-infected and uninfected donors had similar phenotypes and anti-proliferative potency, but HIV-infected individuals had higher proportions of CMV-CD4+CD27-CD28- TR. The CMV-CD4+CD27-CD28- TR may contribute to the downregulation of CMV-specific and nonspecific immune responses of CMV-infected individuals.

Figure 1. CMV-specific CD4⁺CD27⁻CD28⁻ T cells inhibit CMV-stimulated proliferation of autologous PBMC

Panel A: Gating strategy for CD4⁺CD27⁻CD28⁻ separation. CMV-stimulated PBMC were stained using the following conjugated mAbs: anti-CD3-AmCyan, anti-CD4-PE-Texas Red, anti-CD27-FITC and anti-CD28-FITC surface markers. Cells in the lymphocyte gate delineated by side and forward scatter (Left plot) were analyzed for expression of CD3 and CD4 (Middle plot) and subsequently combined CD27 and CD28 (Right plot). Using a cell sorter, CD4⁺CD27⁻CD28⁻ cells were separated from CD4⁺nonCD27⁻CD28⁻ cells. Panel B: Data were derived from 6 experiments using PBMC from independent donors. Freshly thawed autologous PBMC were treated with CMV-specific CD4⁺CD27⁻CD28⁻ cells at the indicated ratios and stimulated with CMV lysate for 6 days. % inhibition of CMV-stimulated proliferation was calculated by the following formula: (1 - median cpm in treated wells/median cpm in untreated wells)*100. Bars represent means and SEM. The inhibition of proliferation was dose-dependent as shown by linear trend analysis (p value and r² depicted in the graph). Panel C: Freshly thawed PBMC from 3 donors were treated with autologous CMV-specific CD4⁺nonCD27⁻CD28⁻ cells prior to de novo CMV stimulation. Ratios were similar to the experiments depicted in panel B. Untreated indicates no added cells. Bars represent means and SEM of median cpm for each condition. There was no evidence of a CD4⁺nonCD27⁻CD28⁻ inhibitory effect on proliferation of autologous PBMC.

Figure 2. Comparison of the inhibitory effect of CMV-specific CD4⁺CD27⁻CD28⁻ T_R with that of CD4⁺CD25⁺CD127⁻ natural T_R on CMV-stimulated proliferation of autologous PBMC

Light bars represent mean and SEM of 9 experiments that measured the inhibitory effect of CMV-specific CD4⁺CD27⁻CD28⁻ T_R at the indicated ratios. Dark bars represent mean and SEM of 3 experiments that measured the effect of CD4⁺CD25⁺CD127⁻ natural T_R on CMV-specific proliferation. The 2-way ANOVA analysis did not reveal any appreciable differences.

Figure 3. Frequency of CD4⁺CD27⁻CD28⁻ T_R in CMV-stimulated compared with mock- infected control and C. albicans-stimulated PBMC

Data were derived from 10 CMV-seropositive HIV-seropositive (CD4 of 200 to 957 cells/μl; plasma HIV RNA<400 copies/ml), 7 CMV-seropositive HIV-seronegative, and 4 CMV-seronegative HIV-seronegative individuals. The enumeration of CD4⁺CD27⁻CD28⁻ was performed after 6 days of in vitro PBMC stimulation with the indicated antigens. CMV stimulation significantly increased the proportion of CD4⁺CD27⁻CD28⁻ T_R in CMV-seropositive individuals, irrespective of their HIV status, but not in CMV-seronegative donors. Candida stimulation moderately increased CD4⁺CD27⁻CD28⁻ T_R.

Figure 4. Antigen specificity of the anti-proliferative effect of CD4⁺CD27⁻CD28⁻ T_R

Data were derived from 10 experiments including PBMC from 3 HIV-infected individuals (CD4 of 506 to 947 cells/μl; plasma HIV RNA<400 copies/ml) and 7 healthy subjects. Panel A shows inhibition of CMV-specific proliferation of PBMC after addition of autologous CMV-stimulated CD4⁺CD27⁻CD28⁻ T_R at a ratio of 1:10. Panel B shows C. albicans-specific proliferation in similar experimental conditions. Incorporated radioactivity was measured in counts per minute (cpm). P values were calculated by paired T test of log₁₀cpm. Panel C shows % inhibition of C. albicans- or CMV-stimulated PBMC calculated with the following formula: (1-median cpm in treated wells/median cpm in untreated wells)*100. Bars represent means and SEM. P values were obtained by paired T test analysis.

Figure 5. Phenotypic characteristics of the CMV-specific CD4⁺CD27⁻CD28⁻ T_R

PBMC were stimulated in vitro with CMV lysate for 6 days and stained for the indicated markers. Panel A shows an example of the gating strategy used to measure the expression of TGFβ on the CD4⁺ populations of interest. PBMC were stimulated for 6 days with CMV lysate. The frequency of cells expressing TGFβ or other markers (not depicted) was measured in the total CD4⁺ cell population (UL+UR) and in the individualized CD4⁺CD27⁻CD28⁻ subpopulation (UL). Panel B shows markers over-expressed by the CD4⁺CD27⁻CD28⁻ T_R (13 to 23 independent samples for each marker). Panel C shows markers under-expressed by CD4⁺CD27⁻CD28⁻ T_R (5 and 20 samples for CD127 and PD1-L, respectively). Panel D shows markers whose expression was not appreciably different between CD4⁺ cells and CD4⁺CD27⁻CD28⁻ T_R (6 to 17 samples/analyte).

Figure 5. Phenotypic characteristics of the CMV-specific CD4⁺CD27⁻CD28⁻ T_R

PBMC were stimulated in vitro with CMV lysate for 6 days and stained for the indicated markers. Panel A shows an example of the gating strategy used to measure the expression of TGFβ on the CD4⁺ populations of interest. PBMC were stimulated for 6 days with CMV lysate. The frequency of cells expressing TGFβ or other markers (not depicted) was measured in the total CD4⁺ cell population (UL+UR) and in the individualized CD4⁺CD27⁻CD28⁻ subpopulation (UL). Panel B shows markers over-expressed by the CD4⁺CD27⁻CD28⁻ T_R (13 to 23 independent samples for each marker). Panel C shows markers under-expressed by CD4⁺CD27⁻CD28⁻ T_R (5 and 20 samples for CD127 and PD1-L, respectively). Panel D shows markers whose expression was not appreciably different between CD4⁺ cells and CD4⁺CD27⁻CD28⁻ T_R (6 to 17 samples/analyte).

Figure 6. Co-expression of FoxP3, TGFβ, granzyme B and PD-1 on CMV-specific CD4⁺CD27⁻CD28⁻ T_R

PBMC from 8 donors were stimulated in vitro with CMV lysate for 6 days and stained for the indicated markers. Panel A shows examples of the gating strategy. In the upper group, gated CD4⁺CD27⁻CD28⁻ T_R (left plot) were analyzed for TGFβ expression (middle plot). Next, we measured the frequency of cells expressing granzyme B among the CD4⁺CD27⁻CD28⁻TGFβ⁻ and the CD4⁺CD27⁻CD28⁻TGFβ⁺ subpopulations. In the lower left and right areas, a similar strategy is depicted for measuring expression of TGFβ on granzyme B+ compared with granzyme B- cells and on PD-1+ compared with PD-1- cells, respectively. The bar graphs represent means and SEM of the 8 experiments and show that a higher proportion of CMV-specific CD4⁺CD27⁻CD28⁻FoxP3+ cells express TGFβ compared with CD4⁺CD27⁻CD28⁻FoxP3⁻ cells (panel B); CMV-specific CD4⁺CD27⁻CD28⁻TGFβ⁺ cells express FoxP3 (panel C), granzyme B (panel D) and PD-1(panel E) in a higher proportion comapred with CD4⁺CD27⁻CD28⁻TGFβ⁻ cells.

Figure 6. Co-expression of FoxP3, TGFβ, granzyme B and PD-1 on CMV-specific CD4⁺CD27⁻CD28⁻ T_R

PBMC from 8 donors were stimulated in vitro with CMV lysate for 6 days and stained for the indicated markers. Panel A shows examples of the gating strategy. In the upper group, gated CD4⁺CD27⁻CD28⁻ T_R (left plot) were analyzed for TGFβ expression (middle plot). Next, we measured the frequency of cells expressing granzyme B among the CD4⁺CD27⁻CD28⁻TGFβ⁻ and the CD4⁺CD27⁻CD28⁻TGFβ⁺ subpopulations. In the lower left and right areas, a similar strategy is depicted for measuring expression of TGFβ on granzyme B+ compared with granzyme B- cells and on PD-1+ compared with PD-1- cells, respectively. The bar graphs represent means and SEM of the 8 experiments and show that a higher proportion of CMV-specific CD4⁺CD27⁻CD28⁻FoxP3+ cells express TGFβ compared with CD4⁺CD27⁻CD28⁻FoxP3⁻ cells (panel B); CMV-specific CD4⁺CD27⁻CD28⁻TGFβ⁺ cells express FoxP3 (panel C), granzyme B (panel D) and PD-1(panel E) in a higher proportion comapred with CD4⁺CD27⁻CD28⁻TGFβ⁻ cells.

Figure 7. Growth characteristics of CMV-specific CD4⁺CD27⁻CD28⁻ T_R

Sorted CMV-specific CD4⁺CD27⁻CD28⁻ (panel A) and CD4⁺nonCD27⁻CD28⁻ (panel B) were restimulated for 6 days with mock-infected control antigen (mock), CMV lysate (CMV) or IL2. Data points represent ³H thymidine incorporation (cpm) measured at the end of the restimulation period. P values were calculated by paired T test of log₁₀cpm.

Figure 8. Mediators of the anti-proliferative effect of CMV-specific CD4⁺CD27⁻CD28⁻ T_R

CD4⁺CD27⁻CD28⁻ T_R sorted from CMV-stimulated PBMC were treated with anti-TGFβ neutralizing mouse mAbs (panel A), anti-PD-1 blocking mouse mAbs (panel B), granzyme B inhibitors (panel C) or an irrelevant control mouse mAb (panel D) before incubation with autologous PBMC. The bars represent inhibition of CMV-stimulated proliferation of PBMC incubated with treated or untreated T_R in paired experiments.