CD5 Suppresses IL-15–Induced Proliferation of Human Memory CD8⁺ T Cells by Inhibiting mTOR Pathways Free

Interleukin-15 belongs to a group of cytokines that share the common γ-chain (γ_c; CD132) receptor (1) and was initially discovered as a novel T cell growth factor (2, 3). Later, multifaceted roles of IL-15 were reported in both innate and adaptive immune cells (4). In particular, IL-15 stimulates NK cells and memory CD8⁺ T cells to proliferate and enhance effector functions.

The role of IL-15 in CD8⁺ T cells has focused mostly on the homeostatic proliferation of memory CD8⁺ T cells. Memory phenotype CD8⁺ T cells have been reported to be deficient in IL-15 or IL-15Rα knockout mice (5, 6). In addition, memory CD8⁺ T cells were demonstrated to be generated, but not homeostatically maintained, in IL-15 knockout mice (7, 8). Moreover, in vitro IL-15 stimulation induces profound proliferation of human memory CD8⁺ T cells in the absence of TCR engagement, which supports the significance of IL-15 in homeostatic proliferation of human memory CD8⁺ T cells (9, 10).

IL-15 plays a role not only in homeostatic proliferation, but also in bystander activation of memory CD8⁺ T cells (11–15). During acute viral infection, overproduced IL-15 can activate pre-existing bystander memory CD8⁺ T cells without TCR stimulation regardless of their Ag specificity (14). Bystander memory CD8⁺ T cells that are activated by IL-15 proliferate and upregulate the expression of cytotoxic molecules, such as perforin and granzyme B, and NK-activating receptors, such as NKG2D. Importantly, IL-15–stimulated bystander memory CD8⁺ T cells exert NKG2D-dependent innate-like cytotoxicity, which contributes to immunopathological host injury (14).

One interesting observation is that the IL-15–induced proliferation varies in human CD8⁺ T cells according to the stage of differentiation (9, 16). In general, memory CD8⁺ T cells in an advanced stage of differentiation have unique characteristics of replicative senescent cells, including short telomeres (17), low telomerase activity (18), and reduced proliferation in response to TCR stimulation (19). However, IL-15 stimulation elicits proliferation more vigorously in more differentiated memory CD8⁺ T cells than in less differentiated memory CD8⁺ T cells. The CD5 expression level that decreases in more differentiated memory CD8⁺ T cells was previously demonstrated to inversely correlate with IL-15 responsiveness, although whether CD5 indicates IL-15 hyporesponsiveness or directly regulates the IL-15 response in memory CD8⁺ T cells is not clear (20).

In the current study, we examined the mechanism by which more differentiated human memory CD8⁺ T cells would respond better to IL-15 stimulation. As reported previously, CD5 expression inversely correlated with IL-15–induced proliferation. In addition, the expression of CD5 decreased in more differentiated memory CD8⁺ T cells compared with less differentiated memory CD8⁺ T cells. Importantly, we demonstrated that IL-15–induced proliferation of human memory CD8⁺ T cells is significantly enhanced by knocking down CD5 expression or blocking CD5, indicating that CD5 is directly involved in the IL-15 response of human memory CD8⁺ T cells as a negative regulator.

Peripheral blood samples were obtained from healthy volunteers from the Korea Advanced Institute of Science and Technology Clinic Pappalardo Center. Donors did not have any comorbidities, were not on any immunosuppressive drugs, and retained physical mobility and lifestyle independence. PBMCs were isolated by density gradient centrifugation using lymphocyte separation medium (Corning Life Sciences, Manassas, VA). Some samples were cryopreserved in liquid nitrogen until use after HLA-A2 was determined by anti-HLA-A2 Ab.

PBMCs or cells sorted from the PBMCs were labeled with a cell division tracking dye, CellTrace Violet (Life Technologies, Gaithersburg, MD). Cells were incubated with 5 μM CellTrace Violet for 20 min at room temperature, washed with PBS supplemented with FBS, and cultured in RPMI 1640 medium (Welgene, Daegu, Korea) supplemented with 10% FBS and penicillin (10 U/ml)/streptomycin (10 g/ml) (Welgene) for 2–8 d in the presence of recombinant human IL-15 (10 ng/ml; PeproTech, Rocky Hill, NJ). Ki-67 staining was performed following fixation and permeabilization (fixation/permeabilization buffer kit; eBioscience, San Diego, CA).

For CD5-blocking experiments, cells were preincubated with 1.0 μg/ml anti-human CD5 (LT1; Abcam, Cambridge, MA) or mouse IgG1 (ICIGG1; Abcam or IS5-21F5; Miltenyi Biotec, Auburn, CA) for 1 h before being plated for IL-15 stimulation. In the experiments with chemical inhibitor, cells were pretreated with rapamycin (100 nM, Sigma-Aldrich, St. Louis, MO) or Akt inhibitor VIII (1 μM, Sigma-Aldrich) and cultured in the presence of IL-15. The stained cells were analyzed using an LSR II instrument (BD Biosciences, Franklin Lakes, NJ) and FlowJo v10.7 software (FlowJo, Ashland, OR).

To measure IL-15–induced BrdU incorporation, sorted CD45RA⁺ effector memory (CCR7⁻CD45RA⁺) T (T_EMRA) and effector memory T (T_EM) cells were cultured with IL-15 and BrdU (10 μM) for 4 d. Cells were stained using a BrdU flow kit (BD Biosciences) according to the manufacturer’s protocol.

Healthy donor PBMCs were labeled with anti-CD8 MicroBeads (Miltenyi Biotec) and sorted according to the manufacturer’s instructions. The sorted cells were further stained with anti–CCR7-FITC, anti–CD45RA-allophycocyanin, and DAPI (Life Technologies) for the isolation of CD45RA⁺ and CD45RA⁻ memory CD8⁺ T cells, or anti–CCR7-FITC, anti–CD45RA-allophycocyanin, anti–CD5-allophycocyanin-Cy7, and DAPI for the isolation of CD5^high and CD5^low memory CD8⁺ T cells. To isolate CCR7⁻, CD57⁺, or CD57⁻ CD8⁺ T cells from healthy donor PBMCs, CD8⁺ T cells were negatively isolated using the CD8⁺ T cell isolation kit (Miltenyi Biotec) according to the manufacturer’s instructions. The sorted CD8⁺ T cells were stained with anti–CD57-FITC, anti–CD8-allophycocyanin-H7, and 7-aminoactinomycin D for the isolation of CD57⁺ and CD57⁻ CD8⁺ T cells. The stained cells were sorted using a FACSAria II instrument (BD Biosciences), and the purity of all isolated populations was >99.0%. Isolated CD8⁺ T cells were stained with FITC-conjugated anti-CCR7 Abs (R&D Systems, Minneapolis, MN) and anti-FITC MicroBeads (Miltenyi Biotec) according to the manufacturer’s instructions for negatively isolating CCR7⁻CD8⁺ T cells.

For the in vitro CD5-blocking assay, memory CD8⁺ T cells were enriched from live PBMCs using a MagniSort human CD8 memory T cell enrichment kit (eBioscience) according to the manufacturer’s recommendation. The purity of enriched CD8⁺ T cells was >80%, and the purity of memory T cells in enriched CD8⁺ T cells was >99.5%.

In direct ex vivo pentamer staining, the FcR-blocked PBMCs were stained sequentially with Live/Dead fixable red stain dye and PE-conjugated MHC class I (MHC-I) pentamers specific for human CMV (HCMV), EBV, or influenza A virus (IAV) (all from ProImmune, Oxford, U.K.) and fluorochrome-conjugated mAbs for cell surface proteins. The following PE-conjugated HLA-A*0201 pentamers were used: HCMV pp65 495–504 (NLVPMVATV), HCMV IE1 81–89 (VLAELVKQI), EBV BALF-4 276–284 (FLDKGTYTL), EBV LMP-2 356–364 (FLYALALLL), EBV LMP-1 125–133 (YLLEMLWRL), EBV LMP-2 426–434 (CLGGLLTMV), EBV BMLF-1 259–267 (GLCTLVAML), and IAV MP 58–66 (GILGFVFTL).

In the evaluation of IL-15–induced proliferation and Ki-67 expression, PBMCs were stained with CellTrace Violet and cultured in the presence of IL-15 (10 ng/ml) for 4 d. The cultured PBMCs were harvested and treated with Fc receptor blocking reagents, followed by staining with Live/Dead fixable red stain dye, the virus-specific PE-conjugated MHC-I pentamers (ProImmune), and fluorochrome-conjugated mAbs. The cells were fixed and permeabilized using a fixation/permeabilization buffer kit (eBioscience) to stain Ki-67.

Cultured PBMCs were stained with Live/Dead fixable red stain dye (Life Technologies) and surface Abs, fixed with intracellular fixation buffer (eBioscience), and then permeabilized with 100% methanol on ice. These methanol-fixed cells were stained with phosphorylation-specific Abs and analyzed by multicolor flow cytometry.

Telomere length in CD8⁺ T cells was determined by the method described by Rufer et al. (21) with some modifications. PBMCs were stained with anti–CD3-BV510, anti–CD8-BV711 (BD Biosciences), anti–CD45RA-BV421 (BD Biosciences), anti–CCR7-Brilliant Violet 785 (BioLegend), and Live/Dead fixable red stain dye (Life Technologies). Stained PBMCs were fixed with BS3 solution (Thermo Scientific Pierce, Rockford, IL) and washed with permeabilization buffer (eBioscience). After incubating at 82°C for 10 min with a FITC-conjugated telomeric peptide nucleic acid probe (Panagene, Daejeon, Korea), the samples were cooled rapidly and incubated for 1 h at room temperature in the dark for hybridization. The stained samples were analyzed by flow cytometry after washing with permeabilization buffer and PBS. The telomere length was obtained as the mean fluorescence intensity.

PBMCs were transfected with CD5-specific small interfering RNA (siRNA) (Qiagen, Valencia, CA) or control siRNA (Qiagen) using the Neon transfection electroporation system (Life Technologies). Transfection was performed with 200 nM siRNA at 2000 V for three 10-ms pulses. After 4 h, transfected PBMCs were stained with CellTrace Violet (Life Technologies) and incubated with IL-15 (10 ng/ml). Cell proliferation and Ki-67 expression of memory CD8⁺ T cells were assessed after 4 d. IL-15–induced expression of perforin and granzyme B on memory CD8⁺ T cells was analyzed using a fixation and permeabilization buffer kit (eBioscience) after 48 h.

Human memory CD8⁺ T cells were enriched and transfected with either control siRNA or CD5 siRNA. The transfected memory CD8⁺ T cells were labeled with CellTrace Violet (Life Technologies) and transferred to NOD.SCID.γ_c-deficient (NSG) mice (5 × 10⁵ cells per mouse), which were provided by S.J. Kim (Sungkyunkwan University, Seoul, Korea). Subsequently, recombinant human IL-15 (PeproTech) was administered i.v. to NSG mice (20 μg per mouse). Proliferation and Ki-67 expression were determined in the liver 3 d later. All housing, breeding, and experimental procedures involving mice were approved by the Animal Care Committee of the Korea Advanced Institute of Science and Technology.

Memory CD8⁺ T cells were enriched using a MagniSort human CD8 memory T cell enrichment kit (eBioscience), pretreated with anti-CD5 blocking Abs or mouse IgG1 Abs, and stimulated with IL-15 (10 ng/ml). The IL-15–treated cells were lysed with RIPA buffer (Thermo Scientific, Rockford, IL) 4 d later, and 10 μg of each cell lysate was loaded onto gels for SDS-PAGE and analyzed using the following Abs: rabbit monoclonal anti-S6 (5G10; Cell Signaling Technology, Danvers, MA), rabbit monoclonal anti–phospho-S6 (2F9, Cell Signaling Technology), rabbit polyclonal anti–phospho-mTOR (Cell Signaling Technology), and HRP-conjugated anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA).

Differences between the two groups were analyzed by a paired t test. In the correlation analysis, the association between two parameters was tested by Pearson’s correlation. Statistical analyses were performed using SPSS version 20.0 (SPSS, Chicago, IL) and GraphPad Prism 5 software (GraphPad Software, San Diego, CA). Two-sided p values were determined in all analyses, and p < 0.05 was considered significant.

All samples were obtained from human volunteers with the approval of the Ethics Committee of the Korea Advanced Institute of Science and Technology (KH2018-118). Voluntary informed consent was obtained in accordance with the Declaration of Helsinki.

Human CD8⁺ T cells can be classified into different subsets according to the expression of CCR7 and CD45RA (or CD45RO), including naive (CCR7⁺CD45RA⁺) T cells, central memory (CCR7⁺CD45RA⁻) T cells, effector memory (CCR7⁻CD45RA⁻) T (T_EM) cells, and CD45RA⁺ effector memory (CCR7⁻CD45RA⁺) T (T_EMRA) cells (22–24). T_EMRA cells that re-express CD45RA, an isoform of CD45, are terminally differentiated memory cells, whereas T_EM cells are relatively less differentiated memory cells (25). In the current study, we analyzed PBMCs obtained from heathy donors and confirmed that T_EMRA cells are more differentiated than T_EM cells by demonstrating lower expression of CD27 and CD28 and higher expression of CD57 in T_EMRA cells than in T_EM cells (Supplemental Fig. 1A). We also confirmed that T_EMRA cells have shorter telomere lengths than for T_EM cells (Supplemental Fig. 1B).

First, we compared IL-15–induced proliferation between T_EMRA and T_EM cells sorted from healthy donor PBMCs. IL-15–induced proliferation evaluated by CellTrace Violet dilution assays was more vigorous in T_EMRA cells than in T_EM cells (Fig. 1A). Similar results were observed when IL-15–induced proliferation was assessed by Ki-67 expression (Fig. 1B) and BrdU incorporation (Supplemental Fig. 1C). We repeated this experiment with extension of the culture up to 8 d and found that IL-15 strikingly induced the proliferation of sorted memory CD8⁺ T cells at 6–8 d (Supplemental Fig. 1D, 1E). Importantly, the superior proliferation of T_EMRA cells compared with T_EM cells was maintained at 6–8 d. Interestingly, superior proliferation of T_EMRA cells was observed in sorted memory CD8⁺ T cells from fresh PBMCs (Supplemental Fig. 1D, 1E) but not from frozen/thawed PBMCs (Supplemental Fig. 1F). In addition, we compared IL-15–induced proliferation between sorted memory CD8⁺ T cells and flow cytometry–gated memory CD8⁺ T cells among unsorted PBMCs and found significantly less IL-15–induced proliferation in sorted memory CD8⁺ T cells compared with memory CD8⁺ T cells among unsorted PBMCs (Supplemental Fig. 2).

To corroborate that more differentiated memory cells respond better to IL-15, we compared IL-15–induced proliferation between CD8⁺ T cell populations with or without the expression of CD57, which is expressed by replicative senescent T cells. We sorted CD57⁺CD8⁺ and CD57⁻CD8⁺ T cells from healthy donor PBMCs and stimulated them with IL-15. We found that CD57⁺CD8⁺ T cells exhibited more proliferation in response to IL-15 treatment than did CD57⁻CD8⁺ T cells in CellTrace Violet dilution assays (Fig. 1C) and with regard to Ki-67 expression (Fig. 1D).

To determine the mechanism underlying enhanced IL-15 responsiveness in T_EMRA cells, we examined the expression of the IL-2R α-chain (IL-2Rα; CD25), IL-2/IL-15R β-chain (IL-2/IL-15Rβ; CD122), and a γ_c (CD132). However, we found no difference in the expression of CD25 and CD122 between T_EMRA and T_EM cells (Fig. 1E, 1F). Moreover, expression of CD132 was lower in T_EMRA cells than in T_EM cells (Fig. 1G).

To determine the mechanism underlying the varying IL-15 responsiveness between more and less differentiated memory CD8⁺ T cells, we performed correlation analyses with diverse T cell phenotype markers. For these analyses, we examined the expression of markers of the T cell phenotype at baseline among 21 different CD8⁺ T cell populations stained with virus-specific (HCMV, EBV, or IAV) MHC-I pentamers from seven healthy donors. In addition, we cultured CellTrace Violet–labeled PBMCs from the same donors in the presence of IL-15 for 4 d and evaluated IL-15–induced proliferation and Ki-67 expression of 21 different CD8⁺ T cell populations stained with the MHC-I pentamers (Supplemental Fig. 3A). Individual CD8⁺ T cell populations exhibited different degrees of IL-15–induced proliferation (Fig. 2A). Next, we performed correlation analyses between IL-15–induced proliferation and T cell phenotype markers examined at baseline in 21 different CD8⁺ T cell populations. Only CD5 expression was significantly associated with IL-15–induced proliferation, with an inverse correlation (Fig. 2B, 2C). When we examined Ki-67 expression to evaluate IL-15–-induced proliferation, CD5 expression was the top correlative parameter that exhibited an inverse association with IL-15–induced Ki-67 expression, although the expression of CD27, NKG2D, and CD28 also exhibited significant correlations (Fig. 2D–F).

We compared the expression of CD5 between T_EMRA and T_EM cells and found that T_EMRA cells expressed lower levels of CD5 than did T_EM cells (Fig. 2G). We also compared the expression of CD5 between CD57⁺CD8⁺ and CD57⁻CD8⁺ T cells and found that CD57⁺CD8⁺ T cells expressed lower levels of CD5 than did CD57⁻CD8⁺ T cells (Fig. 2H). These data show that the expression of CD5 decreases during differentiation of memory CD8⁺ T cells, and that decreased expression of CD5 is associated with the enhanced IL-15 responsiveness of more differentiated memory CD8⁺ T cells.

To validate that CD5 expression negatively correlates with IL-15–induced proliferation, we sorted CD5^high and CD5^low memory CD8⁺ T cell populations from healthy donor PBMCs (Supplemental Fig. 4A) and stimulated them with IL-15. As expected, IL-15–induced proliferation (Fig. 3A) and Ki-67 expression (Fig. 3B) were significantly higher in the CD5^low population than in the CD5^high population. Next, we examined whether CD5 is required for the negative regulation of IL-15–induced proliferation. We cultured memory CD8⁺ T cells in the presence of IL-15 with or without anti-CD5 blocking Abs and found that IL-15–induced proliferation (Fig. 3C) and Ki-67 expression (Fig. 3D) in memory CD8⁺ T cells were enhanced by treatment with anti-CD5 blocking Abs. We also investigated whether CD5 siRNA increases IL-15–induced proliferation. We partially silenced CD5 expression using siRNA (Fig. 3E) and found that IL-15–induced proliferation (Fig. 3F) and Ki-67 expression (Fig. 3G) in memory CD8⁺ T cells were significantly increased by transfection with CD5 siRNA. However, the IL-15–induced increase in the frequency of perforin⁺ or granzyme B⁺ cells was not further increased by transfection with CD5 siRNA (Supplemental Fig. 4B). When we sorted CD5^high and CD5^low memory CD8⁺ T cells and investigated IL-15–induced proliferation and Ki-67 expression after knocking down CD5 expression using CD5 siRNA, we observed enhanced proliferation following CD5 silencing only in CD5^high memory CD8⁺ T cells and not in CD5^low memory CD8⁺ T cells (Supplemental Fig. 4C).

We also examined the effect of CD5 silencing on the IL-15–induced proliferation of human memory CD8⁺ T cells using an in vivo model. We sorted human memory CD8⁺ T cells, transfected them with CD5 siRNA or control siRNA, and labeled them with CellTrace Violet dye (Fig. 4A). The siRNA-transfected, CellTrace Violet–labeled human memory CD8⁺ T cells were adoptively transferred to NSG mice and recombinant human IL-15 was administered i.v. Three days later, the mice were sacrificed and adoptively transferred cells were analyzed in the liver. IL-15–induced proliferation (Fig. 4B) and Ki-67 expression (Fig. 4C) were significantly enhanced by transfection of CD5 siRNA compared with control siRNA. As a result, the frequency of CD5 siRNA-transfected cells was ∼2-fold that of control siRNA-transfected cells in the livers of NSG mice (Fig. 4D).

Taken together, the results indicate that CD5 is not just a correlative marker for IL-15 hyporesponsiveness, but it is also directly involved in negative regulation of the IL-15 response of memory CD8⁺ T cells.

IL-15 induces the proliferation of memory CD8⁺ T or NK cells via the mTOR pathway (26, 27). We confirmed that rapamycin, an mTOR inhibitor, suppresses IL-15–induced proliferation of human memory CD8⁺ T cells (Fig. 5A). Next, we sorted CD5^high and CD5^low memory CD8⁺ T cell populations, stimulated them with IL-15, and examined the phosphorylation of mTOR and ribosomal protein S6, a downstream molecule of mTOR. The IL-15–induced phosphorylation of mTOR (Fig. 5B) and S6 (Fig. 5C) was higher in the CD5^low population than in the CD5^high population, and the phosphorylation of mTOR and S6 was abrogated by rapamycin treatment. In addition, IL-15–induced phosphorylation of mTOR and S6 was enhanced by treatment with anti-CD5 blocking Abs (Fig. 5D). This finding was validated by immunoblot analysis of sorted human memory CD8⁺ T cells (Fig. 5E). Moreover, IL-15–induced phosphorylation of mTOR (Fig. 5F) and S6 (Fig. 5G) in memory CD8⁺ T cells was significantly increased by transfection with CD5 siRNA. We also examined the AKT pathway in IL-15–stimulated memory CD8⁺ T cells. First, we found that Akt inhibitor VIII suppressed IL-15–induced proliferation of memory CD8⁺ T cells (Fig. 5H). Importantly, IL-15–induced phosphorylation of AKT was significantly increased by transfection with CD5 siRNA (Fig. 5I).

Thus, CD5 signals suppress IL-15–induced mTOR and AKT activation, leading to decreased IL-15–induced proliferation of human memory CD8⁺ T cells.

In the current study, we investigated the potential regulator of IL-15–induced proliferation of human memory CD8⁺ T cells. In correlation analyses with diverse T cell phenotype markers, CD5 was identified as the best correlative marker of IL-15 hyporesponsiveness. By performing CD5-blocking and CD5-silencing experiments, we demonstrated that CD5 is not only a correlative marker, but it is also a negative regulator of IL-15–induced proliferation of human memory CD8⁺ T cells. Finally, we showed that CD5 suppresses the IL-15–induced proliferative response by inhibiting the mTOR pathway.

The different proliferative responses in disparate subsets of human memory CD8⁺ T cells following IL-15 stimulation have been well reported (9, 16). However, the mechanism underlying enhanced proliferation in more differentiated memory CD8⁺ T cells remained to be clarified. The difference in the expression of CD122 was considered to be a possible cause of the disparate response to IL-15. In the current study, however, the expression of CD122 did not significantly correlate with IL-15–induced proliferation, implying that the expression of CD122 may not explain the varying IL-15 responsiveness among memory CD8⁺ T cell subsets, even though CD122 expression is indispensable for mediating the IL-15 response.

CD5 is a member of the highly conserved protein receptors belonging to a scavenger receptor cysteine-rich superfamily and is expressed in mature T cells, B-1a B cells, and leukemic B cells (28–30). CD5 is known to be a negative regulator for TCR signaling in T cells (31, 32), and it plays a crucial role during thymocyte development (33). In the thymus, increased expression of CD5 during thymic differentiation reflects strong avidity for self-peptide/MHC complexes (34, 35). In the periphery, naive T cells express high levels of CD5 that are related to the strength of the TCR–self MHC interaction (36–38). In memory T cells, however, the biological significance of CD5 expression remains to be elucidated.

Previous work demonstrated that CD5 expression is progressively downregulated during the process of human memory T cell differentiation, and the lowest expression of CD5 in terminally differentiated memory CD8⁺ T cells is phenotypically associated with enhanced responsiveness to IL-15 (20). However, whether CD5 serves only as a marker reflective of IL-15 hyporesponsiveness or is directly involved in mediating the IL-15 hyporesponse in memory CD8⁺ T cells was not fully elucidated. In the current study, we confirmed that IL-15–induced proliferation is significantly enhanced in human memory CD8⁺ T cells with lower CD5 expression. Furthermore, by abrogating the activity of CD5 with blocking Abs or siRNA, we revealed that CD5 directly suppresses the IL-15–induced proliferation of human memory CD8⁺ T cells.

In the current study, we did not examine which ligand molecule is responsible for CD5 receptor–mediated suppression of IL-15–induced CD8⁺ T cell proliferation. Previous studies reported several putative ligands for CD5 receptor (29). CD5 has been shown to interact with CD72, which is constitutively expressed on B cells (39). Another study described transient expression of activation-induced CD5 ligand (CD5L) on activated murine B cells and T cell clones as an inducible ligand of CD5 on activated lymphocytes (40). In addition, the IgVH framework region of Igs has demonstrated an ability to bind CD5 in rabbit experiments using the F(ab′)₂ fragments of Abs expressing VHa2 framework sequences (41). Investigating CD5 engagement in the cell line model has shown that CD5 itself can act as a ligand for CD5 (42). Furthermore, IL-6 has been reported to be a presumed ligand for CD5 that is expressed on B cells (43). However, it is unclear whether these ligands are involved in the suppression of IL-15–induced CD8⁺ T cell proliferation. Further study is necessary to identify a ligand for CD5 receptor that is responsible for the suppression of IL-15–induced proliferation of memory CD8⁺ T cells.

We also demonstrated that CD5 impedes IL-15–induced proliferation of memory CD8⁺ T cells in vivo using NSG mice. In this model, CD5 siRNA-transfected cells that were adoptively transferred to NSG mice proliferated more efficiently than did control siRNA-transfected cells in response to IL-15 administration. However, the magnitude of IL-15–induced proliferation of the transferred memory CD8⁺ T cells was small. The limited proliferation might be due to the short duration of IL-15–induced proliferation (i.e., 3 d). As shown in Supplemental Fig. 2, sorted memory CD8⁺ T cells require a longer duration for IL-15–induced proliferation. Further work is required to investigate IL-15–induced proliferation of CD5 siRNA-transfected memory CD8⁺ T cells for a longer time course in vivo.

CD5 contains both pseudo-ITAM sequences and pseudo-immunoreceptor tyrosine-based inhibition motif sequences, with multiple potential Ser/Thr phosphorylation sites in its cytoplasmic domain (44). It also has an ability to bind several intracellular signaling molecules, such as Lck, PI3K, CaMKII, protein kinase C, and CK2 activation-related molecules, as well as C-Cbl, RAS-GAP, and SHP-1 inhibition-related molecules (29). Furthermore, several studies have reported an association between CD5 and the mTOR pathway. One study reported that CD5–CK2 signaling in naive CD4⁺ T cells enhances the mTOR pathway upon TCR stimulation, leading to efficient Th17 differentiation (45). Other studies observed that CD5 inhibits the mTOR pathway in CD4⁺ T cells and facilitates the extrathymic regulatory T cell development induced by cytokines involved in effector cell differentiation (46, 47). In our results, CD5 inhibited IL-15–induced mTOR activation. More detailed mechanisms underlying CD5 signaling that inhibit IL-15–induced mTOR activation should be investigated in future studies.

The degree of IL-15–induced proliferation may also be influenced by the distinct metabolic status found in more differentiated human memory CD8⁺ T cells. In the previous study, CD8⁺ T_EMRA cells exhibited more metabolically active features than did T_EM cells in the steady state (48). CD8⁺ T_EMRA cells had greater expression of genes related to the glycolysis, glutaminolysis, and pentose phosphate pathways. Moreover, CD8⁺ T_EMRA cells were characterized by a larger ATP reservoir compared with T_EM cells. Therefore, these distinguished metabolic features found in CD8⁺ T_EMRA cells may contribute to their increased responsiveness to IL-15.

Terminally differentiated memory CD8⁺ T cells, including CD45RA⁺ or CD57⁺ memory CD8⁺ T cells, accumulate in humans with age, and their accumulation correlates with chronic graft-versus-host disease and renal graft rejection (49–52). However, the mechanisms underlying age-dependent accumulation of terminally differentiated memory CD8⁺ cells are poorly understood. The terminally differentiated memory CD8⁺ T cell population can be expanded by repetitive antigenic exposure during aging. However, enhanced IL-15 responsiveness may contribute to expansion of the terminally differentiated memory CD8⁺ T cell population because terminally differentiated memory CD8⁺ T cells express lower levels of CD5. Furthermore, from the aspect of shaping the immune repertoire of the memory CD8⁺ T cell population during aging, the uneven IL-15–induced homeostatic proliferation could result in clonal expansion of T cell clones with lower expression of CD5, which can lead to a compromised immune repertoire with age.

IL-15 was shown to induce bystander activation of memory CD8⁺ T cells, which contributes to disease pathogenesis (12–15, 53–55). Previously, our group reported that, in acute hepatitis A virus infection, excessive production of IL-15 activates pre-existing memory CD8⁺ T cells that are not specific for hepatitis A virus in an Ag-independent manner (14). This TCR-independent IL-15 stimulation activates memory CD8⁺ T cells to proliferate and upregulate the expression of cytotoxic molecules, including perforin, granzyme B, and NKG2D. Notably, these IL-15–induced bystander-activated memory CD8⁺ T cells exert innate-like cytotoxicity in an NKG2D-dependent manner, resulting in immunopathological liver injury (14). Although we only assessed IL-15–induced proliferation in the current study, it is possible that CD5 is associated with other functions induced by IL-15 stimulation. Further studies are required to determine whether CD5 suppresses the IL-15–induced innate-like cytotoxicity of bystander-activated memory CD8⁺ T cells and can be targeted for the alleviation of IL-15–induced T cell–mediated immunopathology.

In conclusion, we demonstrated that CD5 is not just a correlative marker of IL-15 hyporesponsiveness, but also directly suppresses IL-15–induced proliferation of human memory CD8⁺ T cells by inhibiting mTOR pathways. In the future, the roles of CD5 in IL-15–related biological events need to be investigated further, including the accumulation of memory CD8⁺ T cells during aging, the activation of bystander memory CD8⁺ T cells, and immunopathology during viral infections.

We thank Dr. S.J. Kim (Sungkyunkwan University, Seoul, Korea) for providing NSG mice.

The authors have no financial conflicts of interest.