Phenotypic and functional separation of memory and effector human CD8+ T cells

Abstract

Human CD8+ memory- and effector-type T cells are poorly defined. We show here that, next to a naive compartment, two discrete primed subpopulations can be found within the circulating human CD8+ T cell subset. First, CD45RA-CD45R0(+) cells are reminiscent of memory-type T cells in that they express elevated levels of CD95 (Fas) and the integrin family members CD11a, CD18, CD29, CD49d, and CD49e, compared to naive CD8+ T cells, and are able to secrete not only interleukin (IL) 2 but also interferon gamma, tumor necrosis factor alpha, and IL-4. This subset does not exert cytolytic activity without prior in vitro stimulation but does contain virus-specific cytotoxic T lymphocyte (CTL) precursors. A second primed population is characterized by CD45RA expression with concomitant absence of expression of the costimulatory molecules CD27 and CD28. The CD8+CD45RA+CD27- population contains T cells expressing high levels of CD11a, CD11b, CD18, and CD49d, whereas CD62L (L-selectin) is not expressed. These T cells do not secrete IL-2 or -4 but can produce IFN-gamma and TNF-alpha. In accordance with this finding, cells contained within this subpopulation depend for proliferation on exogenous growth factors such as IL-2 and -15. Interestingly, CD8+CD45RA+CD27- cells parallel effector CTLs, as they abundantly express Fas-ligand mRNA, contain perforin and granzyme B, and have high cytolytic activity without in vitro prestimulation. Based on both phenotypic and functional properties, we conclude that memory- and effector-type T cells can be separated as distinct entities within the human CD8+ T cell subset.

Figure 1

Triple color immunofluorescence analysis of cord blood lymphocytes and PBMCs from an adult donor. Cells were simultaneously stained with either CD4-PerCP or PerCP-labeled CD8 in combination with PE-conjugated CD45RA and FITC-conjugated CD27 mAbs. Cells were gated on forward scatter/side scatter and PerCP-fluorescence. Four distinct populations can be defined within CD8⁺ cells from adult blood: (1) CD45RA⁺CD27⁺ cells, (2) CD45RA⁻CD27⁺ cells, (3) CD45RA⁺ CD27⁻ cells, and (4) CD45RA⁻CD27⁻ cells. Note that there is also a transitional population with dull CD45RA expression. As in the CD4⁺ subset (23), these cells have an intermediate phenotype for the studied surface markers (data not shown). Functionally, this subset was not analyzed.

Figure 2

(A) CD8⁺ T cell subsets differ in Fas expression. (B) Fas ligand (FasL) and HPRT (hypoxanthine-guanine phosphoribosyltransferase) mRNA expression in sorted CD8⁺ T cell subsets as assessed by RT-PCR. Lane A, CD8 total; lane B, CD8⁺ CD45RA⁺CD27⁺; lane C, CD8⁺ CD45RA⁺CD27⁻; lane D, CD8⁺ CD45RA⁻CD27⁺.

Figure 2

(A) CD8⁺ T cell subsets differ in Fas expression. (B) Fas ligand (FasL) and HPRT (hypoxanthine-guanine phosphoribosyltransferase) mRNA expression in sorted CD8⁺ T cell subsets as assessed by RT-PCR. Lane A, CD8 total; lane B, CD8⁺ CD45RA⁺CD27⁺; lane C, CD8⁺ CD45RA⁺CD27⁻; lane D, CD8⁺ CD45RA⁻CD27⁺.

Figure 3

Intracellular measurement of cytokines in CD8⁺ T cells. Purified CD8⁺ cells were stimulated for 4 h (IFN-γ and IL-4) or 8 h (IL-2) with PMA and ionomycin in the presence of the protein secretion inhibitor monensin. After surface staining with CD45 and CD27 mAbs, cells were fixated and permeabilized and intracellular accumulated cytokines were detected with specific mAbs. The histograms show the cytokine staining. Dotted lines indicate the negative controls; the positive cell fraction is painted black; and the dot-plots show the distribution of positive cells (black) among the total cell population (gray).

Figure 4

Proliferative capacity of CD8⁺ T cell subsets. Purified CD8⁺ T cells were sorted into the indicated subsets and stimulated with a combination of three CD2 mAbs in the presence of different cytokines. Proliferation was measured on day 4 by adding [³H]thymidine during the last 4 h of the culture. Results from one out of three experiments are shown.

Figure 5

Cytotoxic capacities of CD8⁺ T cell subsets. Purified CD8⁺ T cells were sorted into the indicated subsets and cytotoxicity was directly analyzed against P815 target cells in the presence of CD3 mAb in a 4-h ⁵¹Cr-release assay. All effector populations used were unable to lyse P815 cells in the absence of CD3 mAb (lysis <3%), thus excluding NK cell–like activity in this system (data not shown). Results are the mean ± SD of three independent experiments. The distribution of cells in the different subsets was as follows: CD45RA⁺CD27⁺, 18–37%; CD45RA⁺ CD27⁻, 7–23%; and CD45RA⁻ CD27⁺, 30–38%. (CD8⁺ T cells, squares; CD8⁺CD45RA⁺CD27⁺, triangles; CD8⁺CD45RA⁻CD27⁺; plus; CD8⁺ CD45RA⁺CD27⁻, circles.)

Figure 6

Expression of mediators of cytotoxicity in CD8⁺ T cell subsets. After surface staining of freshly isolated CD8⁺ cells with CD45RA and CD27 mAbs, cells were fixated and permeabilized and intracellularly present perforin and granzyme B were detected with specific mAbs. The histograms show the staining for the intracellular proteins. Dotted lines indicate the negative controls. The positive cell fraction is painted black. The dot-plots show the distribution of positive cells (black) among the total cell population (gray). Note that the different dot-plot staining is due to the use of either biotinylated CD27 mAb in combination with red 670 or FITC-labeled CD27 mAb.

Figure 7

Using standard limiting dilution analysis, the CTLp frequencies for the HIV-1_LAI RT-derived peptide aa 244-252 (IVLPEKDSW) were quantified in CD8⁺ T cell subsets from an HIV-infected individual. Peptide-specific CTLp frequencies were calculated by subtracting CTLp frequencies determined with targets without peptide from CTLp frequencies determined with targets pulsed with 5 μg/ml of RT_LAI peptide aa 244-252. Mean + SD of three separate experiments.

Figure 8

Helper and suppressor capacities of CD8⁺ T cell subsets. B cells were cocultured with the indicated T cell subsets in CD3-coated wells for 14 d. Ig content of the supernatant was measured with specific ELISAs. (A) Cells of the different T cell populations were cultured with B cells alone to determine their helper capacities. (B) Cells of the distinct CD8⁺ subsets were added to the coculture of autologous B and CD4⁺ T cells to measure their suppressor activity. One out of two comparable experiments is shown. The distribution of cells in the different subsets was as follows: CD45RA⁺CD27⁺, 50%; CD45RA⁺CD27⁻, 5%; and CD45RA⁻ CD27⁺, 35%.

Figure 8

Helper and suppressor capacities of CD8⁺ T cell subsets. B cells were cocultured with the indicated T cell subsets in CD3-coated wells for 14 d. Ig content of the supernatant was measured with specific ELISAs. (A) Cells of the different T cell populations were cultured with B cells alone to determine their helper capacities. (B) Cells of the distinct CD8⁺ subsets were added to the coculture of autologous B and CD4⁺ T cells to measure their suppressor activity. One out of two comparable experiments is shown. The distribution of cells in the different subsets was as follows: CD45RA⁺CD27⁺, 50%; CD45RA⁺CD27⁻, 5%; and CD45RA⁻ CD27⁺, 35%.