Prostaglandin E2 promotes features of replicative senescence in chronically activated human CD8+ T cells

Abstract

Prostaglandin E2 (PGE2), a pleiotropic immunomodulatory molecule, and its free radical catalyzed isoform, iso-PGE2, are frequently elevated in the context of cancer and chronic infection. Previous studies have documented the effects of PGE2 on the various CD4+ T cell functions, but little is known about its impact on cytotoxic CD8+ T lymphocytes, the immune cells responsible for eliminating virally infected and tumor cells. Here we provide the first demonstration of the dramatic effects of PGE2 on the progression of human CD8+ T cells toward replicative senescence, a terminal dysfunctional state associated multiple pathologies during aging and chronic HIV-1 infection. Our data show that exposure of chronically activated CD8+ T cells to physiological levels of PGE2 and iso-PGE2 promotes accelerated acquisition of markers of senescence, including loss of CD28 expression, increased expression of p16 cell cycle inhibitor, reduced telomerase activity, telomere shortening and diminished production of key cytotoxic and survival cytokines. Moreover, the CD8+ T cells also produced higher levels of reactive oxygen species, suggesting that the resultant oxidative stress may have further enhanced telomere loss. Interestingly, we observed that even chronic activation per se resulted in increased CD8+ T cell production of PGE2, mediated by higher COX-2 activity, thus inducing a negative feedback loop that further inhibits effector function. Collectively, our data suggest that the elevated levels of PGE2 and iso-PGE2, seen in various cancers and HIV-1 infection, may accelerate progression of CD8+ T cells towards replicative senescence in vivo. Inhibition of COX-2 activity may, therefore, provide a strategy to counteract this effect.

Figure 1. EP receptor expression and cAMP upregulation in CD8+ T cells.

CD8+ T cells were freshly isolated from PBMC from whole blood derived from healthy donors or HIV+ persons. (A) (Left) EP2 and EP4 transcripts were evaluated by quantitative PCR in ex vivo samples and in T cells activated with anti-CD2/CD3/CD28 microbeads for 24 hours. 36B4 was used as the housekeeping gene and data represents 3 healthy donors performed on a single plate (*p = 0.05). (Right) EP2 and EP4 surface expression was also evaluated in healthy at 2 hours and 24 hours post activation, or with no Ab-coated bead activation. Flow cytometric histogram shows one representative donor from a healthy person stained with PE–anti-human EP2 or PE–anti-EP receptor antibodies (Cayman Chemical). (B) Intracellular cAMP was evaluated using a direct cAMP ELISA kit (Enzo Biosceinces) in T cells treated with PGE₂, isoPGE₂, and known EP agonists, misoprostol (EP2, EP3, and EP4) and butaprost (EP2) for 72 h (n = 3; p<0.005 by Kruskal Wallis for the comparison of all treatment groups to control).

Figure 2. Proliferative potential decreases and p16 transcripts increase in the presence of PGE₂ and iso-PGE₂.

T cells were activated with anti-CD2/CD3/CD28 microbeads with or without PGE₂ or iso-PGE₂ and population doublings (PD) was calculated by the formula PD = log₂ (final cell concentration/initial cell concentration) (A) Long term culture of one representative donor. (B) p16 transcripts were quantified by qPCR during early (PD4–8) and late (PD12–16) time points in the presence of PGE₂, iso-PGE₂ or diluent (n = 5; *p = 0.031 compared to control by the paired permutation test). 36B4 was used as the housekeeping gene.

Figure 3. PGE₂ and iso-PGE₂ inhibit telomerase activity while increasing intracellular ROS.

CD8+ T cells were negatively selected after 72 hours post activation with anti-CD2/CD3/CD28 microbeads with or without PGE₂ or iso-PGE₂, and inhibitors of PKA pathway. (A) (Top Left) hTERT expression was quantified at equivalent PDs by qPCR (n = 5; *p = 0.0312; **p = 0.031 using one-sided T test because of decreased sample size for these conditions only; n = 4). (Top Middle) Representative gel showing effects of PGE₂ (10⁻⁶ to 5⁻⁷M) and iso-PGE₂ (10⁻⁶ to 5⁻⁷M) on telomerase activity of CD8+ T cells. Band intensity per lane correlates with relative telomerase activity of 1,250 CD8+ T cells in each treatment group. (B) Telomerase activity was measured as described in (A) with PGE₂ or iso-PGE₂ in the presence of 1 µM of a PKA inhibitor, H89 dihydrochloride. (C) CD8+ T cell cultures were established and chronically activated as previously described in the presence of the immune modulators. Telomere lengths were evaluated by Real-Time PCR and expressed as a percentage of telomere length of a human tumor cell line, SAOS (∼23Kb). Data represent telomere lengths over the lifetime from 3 representative donor cultures. (D) (Top) The relative amount of intracellular ROS was determined by the mean fluorescence intensity of DCFDA–stained, live CD8+ T cells after 24 h of culture with media alone or in the presence of PGE₂, isoPGE₂, ox-LDL, or H₂O₂ (pos control). (Bottom) Representative flow cytometry profile of MitoSOX red oxidation in PGE₂- and iso-PGE₂-treated T cells.

Figure 4. Key features of T cell function are modulated by PGE₂ and isoPGE₂.

(A) (Top Left) CD8+ T cells treated with iso-PGE₂ (top) and PGE₂ (bottom) were immunostained with FITC–anti-CD25, PE–anti-CD28, and gated on PerCP–anti-CD8 and APC–anti-CD3+ T cells (all BD Biosciences). Samples were compared at the same PD. (Top Middle) The %CD28+cells from different treatment samples over the lifetime of the culture in five donors; *p = 0.031 (Top right) CD28 and CTLA-4 expression was determined by qPCR during both early (PD4–8) and late (PD12–16) culture stages in CD8+ T cells treated with the immunomodulators (n = 5; *p = 0.031). All samples were tested in triplicate and normalized to the housekeeping gene 36B4 (B) IL-2 message was similarly quantified by qPCR as in (A) after treatment (n = 5; *p = 0.031) (C) CD8+ T cells were stimulated with Ab-coated microbeads for 72 h in the presence of PGE₂, iso-PGE_2, or were treated with diluent (DMSO). Intracellular IFN-γ and TNF-α was analyzed flow cytometrically using FITC–anti-TNF-α, PE–anti–IFN-γ, PerCP–anti-CD8 and APC–anti-CD3. The frequencies of the IFN-γ– and TNF-α producing cells in T-cell fractions gated on CD3+CD8+ are shown as percentages.

Figure 5. COX-2 activity and EP4 expression increases in T cells during chronic activation.

(A) (Top left) COX-2 transcripts were quantified by qPCR in ex vivo and activated (anti-CD2/CD3/CD28) CD8+ T cell samples from healthy donors. Each sample was tested in triplicate and normalized to the housekeeping gene 36B4 (n = 5; *p = 0.031). Supernatants from early and late cultures from 3 donors were collected 72 h post activation with Ab-coated microbeads and their average PGE₂ concentration was calculated from triplicate wells, *p = 0.05. (Top Right) COX-2 and EP4 transcripts were quantified 72 h post each round of activation in four healthy donors. Each sample was tested in triplicate and normalized to the housekeeping gene, 36B4. In addition, COX-2 activity was determined in CD8+ T cells at early and late time points 72 h post activation and at quiescence (15–17 d post stimulation) in three healthy donors using the COX Activity Assay (Cayman Chemical). (B) CD8+ T cells were pre-incubated with a highly specific COX-2 inhibitor CAY10404 (1 µM) (Cayman Chemical) or diluent (DMSO) for 30 min and then activated with Ab-coated microbeads as described. 1×10⁶ CD8+ T cells were collected 24h post activation during early and late PDs and intracellular cAMP was measured in triplicate using a direct cAMP ELISA kit (Enzo Life Sciences) (n = 4; *p = 0.05). (C) CD8+ T cells were cultured in the presence of 1 µM CAY10404 or DMSO for 30 min, and then activated with Ab-coated microbeads. 72 h after each round of activation, cells were collected and IL-2 and CD28 transcripts were quantified by qPCR. Each sample was tested in triplicate and normalized to the housekeeping gene 36B4 (n = 5; *p = 0.031).