Protective effector memory CD4 T cells depend on ICOS for survival

Abstract

Memory CD4 T cells play a vital role in protection against re-infection by pathogens as diverse as helminthes or influenza viruses. Inducible costimulator (ICOS) is highly expressed on memory CD4 T cells and has been shown to augment proliferation and survival of activated CD4 T cells. However, the role of ICOS costimulation on the development and maintenance of memory CD4 T cells remains controversial. Herein, we describe a significant defect in the number of effector memory (EM) phenotype cells in ICOS(-/-) and ICOSL(-/-) mice that becomes progressively more dramatic as the mice age. This decrease was not due to a defect in the homeostatic proliferation of EM phenotype CD4 T cells in ICOS(-/-) or ICOSL(-/-) mice. To determine whether ICOS regulated the development or survival of EM CD4 T cells, we utilized an adoptive transfer model. We found no defect in development of EM CD4 T cells, but long-term survival of ICOS(-/-) EM CD4 T cells was significantly compromised compared to wild-type cells. The defect in survival was specific to EM cells as the central memory (CM) ICOS(-/-) CD4 T cells persisted as well as wild type cells. To determine the physiological consequences of a specific defect in EM CD4 T cells, wild-type and ICOS(-/-) mice were infected with influenza virus. ICOS(-/-) mice developed significantly fewer influenza-specific EM CD4 T cells and were more susceptible to re-infection than wild-type mice. Collectively, our findings demonstrate a role for ICOS costimulation in the maintenance of EM but not CM CD4 T cells.

Figure 1. ICOS^−/− and ICOSL^−/− mice have fewer effector memory phenotype CD4 T cells.

Age-matched wild-type and ICOS^−/− (B) or ICOSL^−/− (C & D) mice were sacrificed. Splenocytes (A–C) or cells from the lungs, lymph nodes, and blood (D) were harvested and stained for CD3, CD4, CD44 and CD62L. Cells were gated on CD3⁺CD4⁺ cells then on CD44 and CD62L. Naïve cells are identified as CD44⁻CD62L⁺, effector memory cells as CD44⁺CD62L⁻, and central memory cells as CD44⁺CD62L⁺. Representative plots with gating on naïve, central memory and effector memory populations in wild-type, ICOS^−/−, and ICOSL^−/− mice are shown (A). The total numbers of naïve, effector memory phenotype and central memory phenotype cells in the organs of wild-type and ICOS^−/− or ICOSL^−/− mice were calculated from the percent of each population and total numbers of cells in the organ. N≥4 for each group.

Figure 2. Effector memory CD4 T cells accumulate at a significantly reduced rate in ICOS-deficient mice.

Wild-type and ICOS^−/− or ICOSL^−/− (ICOS-deficient) mice of various ages were sacrificed. Splenocytes were harvested and stained for flow cytometry as described in Figure 1. The total number of naïve, effector memory phenotype and central memory phenotype cells in the spleens of wild-type and ICOS-deficient mice of each age were calculated. For the effector memory phenotype cells, linear regression was performed and the p-value for effector memory phenotype cells (A) represents the likelihood of the slopes of the linear regressions being identical. 2-way ANOVA analysis of the data where there were precisely age-matched mice revealed similar significance.

Figure 3. ICOS-deficient and wild-type effector memory phenotype CD4 T cells undergo similar homeostatic proliferation.

A. Wild-type and ICOS^−/− mice were treated with BrdU for 21 days, then followed for a further 49 days without BrdU. Blood was drawn periodically and red blood cell-depleted. The lymphocytes were then stained for CD4, CD44, CD62L, and BrdU and analyzed by flow cytometry. The percent BrdU⁺ of CD44⁺CD62L⁻ CD4⁺ cells was assessed at each time point. B. Wild-type and ICOSL^−/− mice were treated with BrdU for 24 days, followed for 49 days, and analyzed similarly to A. N≥4 for all groups.

Figure 4. ICOS^−/− effector memory but not central memory CD4 T cells have a defect in survival.

Wild-type (filled squares, solid black line) and ICOS^−/− (open triangles, dotted lines) DO11.10 cells were adoptively transferred into separate hosts that were immunized the next day with OVAp and inactivated S. mansonii eggs, as previously described. On various days after immunization, hosts were sacrificed and the spleen and lymph nodes were harvested. Spleen and lymph node cells were stained with biotin-labeled KJ1-26 antibody, enriched with anti-biotin MACS beads, and then stained further for CD4, CD44, and CD62L and with a streptavidin-fluorochrome reagent to bind biotinylated KJ1-26 antibody. At some time-points, a fraction of the isolated cells were restimulated for 48 hours with OVAp to identify cytokine-producing cells (D). A. Total numbers of CD4⁺KJ1-26⁺ cells. B. Number of effector memory (CD44⁺CD62L^low) CD4⁺KJ1-26⁺ cells. C. Number of central memory (CD44⁺CD62L^high) CD4⁺KJ1-26⁺ cells. Significance (p-value) in part B. reflects the likelihood that the curves are the same (i.e. that one curve would represent all of the data as well as two separate curves for wild-type and ICOS^−/− cells). D. Total numbers of IL-4-producing KJ1-26-enriched cells on day 140. E. Sample plots of DO11.10 and ICOS^−/− DO11.0 memory populations at day 140. N≥4 for every group at each time-point.

Figure 5. ICOS^−/− mice develop fewer effector memory cells in response to influenza infection and have reduced recall responses and protection against re-infection.

ICOSL^−/− or ICOS^−/− and wild-type Balb/c mice were infected i.n. with either the PR8 of influenza then rested (Timeline shown in A). Lung, draining lymph node, and spleen cells were enriched for HA-specific cells and stained for CD4, HA-tetramer, and negatively stained for CD8, CD11c, and CD19 (in one fluorochrome). Data represents the total number of HA-specific CD4 T cells in all organs. A. Diagram of infection, resting time, and rechallenge time. B. Weight loss after primary infection. C. Sample plots of HA-tetramer staining on tetramer-enriched cells and CD44/CD62L staining on HA-tetramer⁺ CD4 T cells from day 39 post-infection (quantified in D). D. (left) Total HA-tetramer-specific CD4+ T cells at 39 days after PR8 infection, (middle) CD44⁺CD62L⁺ HA-tetramer⁺ CD4⁺ central memory cells, and (right) CD44+CD62L- HA-tetramer+ CD4+ effector memory cells. E. Total HA-tetramer-specific CD4 T cells in wild-type and ICOSL^−/− mice infected with PR8 then challenged 85 days later with re-infection with VSV-HA. F. Timeline for Hkx31 infection and PR8 challenge. G. Weight loss in ICOS^−/− and wild-type mice after infection with Hkx31 followed 60 days later by challenge with PR8 virus.