The Functional Requirement for CD69 in Establishment of Resident Memory CD8+ T Cells Varies with Tissue Location

Abstract

Recent studies have characterized populations of memory CD8⁺ T cells that do not recirculate through the blood but are, instead, retained in nonlymphoid tissues. Such CD8⁺ tissue resident memory T cells (T_RM) are critical for pathogen control at barrier sites. Identifying T_RM and defining the basis for their tissue residency is therefore of considerable importance for understanding protective immunity and improved vaccine design. Expression of the molecule CD69 is widely used as a definitive marker for T_RM, yet it is unclear whether CD69 is universally required for producing or retaining T_RM Using multiple mouse models of acute immunization, we found that the functional requirement for CD69 was highly variable, depending on the tissue examined, playing no detectable role in generation of T_RM at some sites (such as the small intestine), whereas CD69 was critical for establishing resident cells in the kidney. Likewise, forced expression of CD69 (but not expression of a CD69 mutant unable to bind the egress factor S1PR1) promoted CD8⁺ T_RM generation in the kidney but not in other tissues. Our findings indicate that the functional relevance of CD69 in generation and maintenance of CD8⁺ T_RM varies considerably, chiefly dependent on the specific nonlymphoid tissue studied. Together with previous reports that suggest uncoupling of CD69 expression and tissue residency, these findings prompt caution in reliance on CD69 expression as a consistent marker of CD8⁺ T_RM.

FIGURE 1.. CD69 plays a minimal role in promoting tissue residency after an LCMV infection.

Animals received a co-adoptive transfer of congenically distinct CD69^−/− and WT P14 CD8+ T cells. At both acute and memory timepoints lymphocytes were isolated from a variety of different nonlymphoid tissues (A). Graphed is the percent of IV- transferred cells that were Cd69^−/− over the percent WT, isolated from the indicated tissues days 30–50 (B) and 7 (C) p.i. The percent of CD69 WT co-transferred cells expressing CD69 (gates set on CD69^−/− cells) at the indicated timepoints, in the indicated tissues (D). The ratio presented on the Y axis of (B) and (C) is normalized across experimental repeats by dividing each individual data point by the known transfer ratio for that experimental repeat. 4 independent repeats for each of the timepoints are combined in the presented normalization. Error bars show mean ± SD. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001, one tailed t test against the known ratio of transferred cells on log₁₀ transformed data.

FIGURE 2.. Tissue resident CD69 deficient cells are functionally competent.

Animals received a co-adoptive transfer of congenically distinct Cd69^−/− and WT P14 CD8+ T cells followed by LCMV infection. At memory timepoints lymphocytes were isolated from the whole female reproductive tract and spleen of mice that did not receive additional treatment (A). A similar cohort of animals received transcervical re-challenge with gp33 peptide or PBS at memory timepoints with isolation of lymphocytes from the FRT 12hrs post treatment (B). Representative expression of Granzyme B, IFN-γ, CD44, and CD69 on CD69^−/− or WT cells isolated from the FRT of LCMV immune animals transcervically challenged with either gp33 peptide or PBS control (C). Quantified expression of Granzyme B (D) and IFN-γ (E). 2 independent repeats are presented for the normalization in (A) and the combined data in (D-E). Data points are normalized to the known ratio of co-adoptively transferred cells as in figure 1 for (A). Representative histograms shown in (C). Error bars show mean ± SD. *P ≤ 0.05, **P≤ 0.01, and ***P ≤ 0.001, one tailed t test against the known ratio of transferred cells on log10 transformed data for (A), paired t test for (D-E).

FIGURE 3.. CD69 is not required for tissue residency in the majority of tissues following infection with a variety of different models.

As in figure 1 animals received a co-adoptive transfer of Cd69^−/− and WT CD8+ T cells. P14 TCR transgenics were used in (A and C) while OT1 TCR transgenics were used for (B). The ratio of co-transferred, IV- cells isolated from a variety of different tissues between 14 and 21 days after infection with influenza-gp33 (A), 7 days after VSV-OVA (B), or 7 days after TriVax (C). Data points are normalized to the known ratio of co-adoptively transferred cells as in figure 1. 3 independent repeats for each model are combined in each of the presented normalizations. Error bars show mean ± SD. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001, one tailed t test against the known ratio of transferred cells on log₁₀ transformed data.

FIGURE 4.. The frequency of antigen specific cells present prior to infection correlates with the magnitude of the defect in the kidney, but not elsewhere.

A range of precursor cell numbers were co-adoptively transferred into recipient animals who were then infected with LCMV as in Figure 1. The normalized ratio (Cd69^−/−/Cd69 WT) of IV- transferred cells in each organ 7 days after infection with LCMV (A). CD69 impairs migration into the kidney within a 16 hour window. Cd69^−/− and WT P14s were co-adoptively transferred as in Fig. 1, followed by LCMV infection. 6 days later, 5 million CD8+ T cells were purified from spleens of infected animals by magnetic enrichment and adoptively transferred into infection matched recipient animals (B). Graphed is the normalized ratio of IV- cells in the indicated tissues 16 hours after the secondary co-transfer (C). Data points are normalized to the known ratio of co-adoptively transferred cells as in figure 1. 3 independent repeats for each model are combined in each of the presented normalizations. Error bars show mean ± SD. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001, ANOVA with multiple comparisons on log₁₀ transformed data for (A), one tailed t test against the known ratio of transferred cells on log₁₀ transformed data for (C).

FIGURE 5.. CD69 acts primarily via S1PR1 to promote tissue residency in the kidney.

In vitro activated P14 cells were transduced retroviral vectors containing either empty vector, CD69, or CD69Δ31, which cannot interact with S1PR1. Cells were transferred into hosts who were infected with LCMV the next day and nonlymphoid tissues were harvested 6 days post LCMV infection. Shown is the percentage of IV- transferred cells expressing the transduction marker CD90.1 on their cell surface, normalized to the percent expressing it in the average of the spleens for each experimental repeat in the lung (A), kidney (B), liver, (C) intestinal epithelium (D) intestinal lamina propria, (E) and inguinal lymph nodes (F). 3 independent repeats for each model are combined in each of the presented normalizations. Error bars show mean ± SD. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001, ANOVA with multiple comparisons.

FIGURE 6.. CD69-deficiency impairs tissue-residency and appearance of KLF2^low CD8⁺ T cells in the kidney.

Animals received a co-adoptive transfer of congenically distinct Cd69^−/− and WT P14 CD8+ T cells followed by LCMV infection. At a memory timepoint, these animals were surgically conjoined to infection matched animals and left to equilibrate for at least 30 days. At this point, cells were isolated from each of the indicated tissues with IV+ cells excluded from analysis for the nonlymphoid tissues. The number of cells in in the donor and recipient parabionts were used to calculate the percentage of cells that were tissue resident using methods from Steinert et al (A). CRISPR/Cas9 was used to knockout CD69 in KLF2-GFP expressing P14 T cells. Animals received a cotransfer of these cells alongside the congenically distinct KLF2-GFP P14s that used a scramble sgRNA as a control, followed by LCMV infection. Expression of KLF2-GFP on cells that received CD69 sgRNA and cells that received scramble sgRNA in the indicated tissues (B). The ratio of recovered CD69 sgRNA/scramble sgRNA receiving cells in KLF2-GFP hi and low cells in the IV- fraction of the indicated organs 7 d.p.i. is shown in (C). Data points are normalized to the known ratio of co-adoptively transferred cells as in figure 1 for (C). 2 independent repeats are presented for the combined data in (A) and the normalization in (C). Representative histograms shown in (B). Error bars show mean ± SEM for (A) and mean ± SD for (C). *P ≤ 0.05, **P≤ 0.01, and ***P ≤ 0.001, ANOVA with multiple comparisons for (A), one tailed t test against the known ratio of transferred cells on log10 transformed data for (C).