IL-7 and IL-15 differentially regulate CD8⁺ T-cell subsets during contraction of the immune response

After acute infection or immunization, antigen-specific CD8⁺ T cells respond in 3 distinct phases: expansion, contraction, and memory.^1,,,–5  Antigen recognition initiates the expansion phase where individual CD8⁺ T-cell clones proliferate and acquire the ability to specifically lyse target cells and produce cytokines. After antigen clearance, the antigen-specific CD8⁺ T-cell population undergoes contraction, where the majority of effector cells die. A small number of effector CD8⁺ T cells survive, leading to stable levels of long-lived CD8⁺ memory T cells, which provide protection on reinfection

A number of studies suggest that cytokines, including interleukin-7 (IL-7) and IL-15, can play complimentary or overlapping roles in maintaining CD8⁺ T cells after antigen stimulation.^{4,,,,,,,,,,,,,,,–20}  Early after infection, CD8⁺ T-cell survival may be impaired in the absence of IL-7.²¹  Furthermore, it was reported that effector CD8⁺ T cells destined to become long-lived memory cells selectively express CD127, suggesting that these cells might be preferentially IL-7 dependent.²²  In support of this, it was shown that IL-7 plays a dominant role over IL-15 in supporting memory cell formation from recently activated CD8⁺ T cells. Whether such memory precursor cells are solely dependent on IL-7, however, is not clear, as CD127^hi effector cells develop after viral infection with the same kinetics in both IL-7-deficient and -sufficient mice, and IL-7-deficient mice develop memory CD8⁺ T cells.²³  In the case of IL-15, several groups have reported reduced maintenance of antigen-experienced CD8⁺ T cells in IL-15-deficient mice, an effect that was much more severe in some cases¹¹  than in others.^12,24  In addition, it was found that acute homeostatic proliferation of memory CD8⁺ T cells displayed overlapping dependency on both endogenous IL-7 and IL-15.^25,–27  To gain insight into these findings, we investigated whether IL-7 and IL-15 may differentially regulate specific CD8⁺ T-cell subsets. We focused our study on the contraction phase of the immune response, which ultimately dictates the seeding of the memory CD8⁺ T-cell compartment.

As a model to evaluate how IL-7 and IL-15 regulate the survival and proliferation of distinct subsets of CD8⁺ effector T cells during contraction, we followed the CD8⁺ T-cell response to viral or intracellular bacterial infection in mice with or without administration of exogenous cytokines. Of most significance, we found that IL-15, and the related cytokine IL-2, induced the preferential accumulation of KLRG1^hiCD127^lo CD8⁺ T cells, whereas provision of IL-7 favored accumulation of KLRG1^loCD127^hi cells, and to a lesser extent, KLRG1^hiCD127^hi CD8⁺ T cells. Interestingly, these respective phenotypes are thought to be indicative of whether a CD8⁺ effector T cell will transition to a long-lived memory T cell or die during or in the months after contraction.^22,28,29  Thus, KLRG1^hiCD127^lo CD8⁺ T cells are thought to be short-lived effector/memory cells, whereas KLRG1^loCD127^hi CD8⁺ T cells are thought to be long-lived memory precursor T cells. In the context of these CD8⁺ T-cell subsets, we find unexpected differences in the ability of IL-7, IL-15, and IL-2 to induce proliferation and survival.

Murine (m) IL-15 was purchased from eBioscience (San Diego, CA). Soluble mIL-15Rα-Fc and anti–human (h) IL-2 mAb (MAB602) were purchased from R&D Systems (Minneapolis, MN). Anti-hIL-7 mAb (500-M07) was purchased from PeproTech (Rocky Hill, NJ). Both hIL-2 and hIL-7 were provided by the NCI Biological Resources Branch Preclinical Repository (Frederick, MD), whereas hIL-15 was kindly provided by Amgen (Thousand Oaks, CA). Reagents and methodology used for routine flow cytometric analysis are described in “Flow cytometric analysis.”

Cytokine complexes were generated by incubation of cytokine and either soluble receptor or antibody for 20 minutes at 37°C and diluted in phosphate-buffered saline before intraperitoneal injection.^30,31  For generation of IL-15 complexes, recombinant mIL-15 or hIL-15 was used at doses that yielded similar biologic activity. The following concentrations were used per injection of cytokine complex: mIL-15 (1.5 μg) with msIL-15Rα-Fc (7 μg), hIL-15 (0.5 μg) with msIL-15Rα-Fc (2.3 μg), hIL-7 (1.5 μg) with anti-hIL-7 mAb (7.5 μg), or hIL-2 (1.5 μg) with anti-hIL-2 mAb (7.5 μg).

C57BL/6 (B6), B6.PL (Thy1.1), CD45.1, RAG1^−/−, and OT-I mice were obtained from The Jackson Laboratory (Bar Harbor, ME). IL-15^−/− mice were kind gifts of Amgen. All animal work was performed in accordance with the guidelines outlined by the Institutional Animal Care and Use Guidelines at the University of California–San Diego (San Diego, CA). For adoptive T-cell transfer, CD8⁺ T cells were injected intravenously one day before infection or cytokine treatment. For analysis of CD8⁺ T-cell contraction, mice were infected intravenously with either 5 × 10³ colony forming units of Listeria monocytogenes expressing ovalbumin (LM-OVA)³²  or 10⁵ plaque forming units of vesicular stomatitis virus expressing ovalbumin (VSV-OVA).³³  In a small percentage of mice, we observed rejection of donor OT-I⁺ CD8⁺ T cells after infection despite a pure B6 background, and were thus excluded from analysis. To ensure complete clearance of Listeria, where indicated, mice were injected intraperitoneally with 1 mg of ampicillin on day 5 of infection and maintained for 48 hours on drinking water with 2 mg/mL of ampicillin, as described.³⁴  To block IL-7-receptor, mice were treated with 500 μg anti–IL-7Rα mAb (A7R34) (intraperitoneally) on days 2, 4, 7, 10, and 13 after infection as previously described.²⁶ 

Before cell sorting, T cells were enriched by negative selection using a mouse T-cell enrichment column (MTCC-25, R&D Systems). Enriched T cells were labeled with antibodies, and CD45.1⁺KLRG1^hi or CD45.1⁺KLRG1^lo cell populations sorted to more than 98% purity using a BD FACSAria. In experiments involving the transfer of OT-I effector cells into IL-7^−/− mice, effector OT-I CD8⁺ T cells were enriched by positive selection, using anti-Thy1.1-biotin (H1S51), streptavidin particles, and an IMag magnet, following instructions provided by the manufacturer, BD Biosciences (San Jose, CA). For carboxyfluorescein diacetate succinimidyl ester (CFSE) labeling, T cells were labeled with 1.5 μM of CFSE (Invitrogen, Carlsbad, CA) following the manufacturer's directions.

Cells were analyzed by flow cytometry using standard techniques. Briefly, cells were washed in staining buffer containing 2% bovine growth serum and 0.01% sodium azide, and stained with fluorescently labeled antibodies as indicated. The antibodies used in these studies were as follows: B220-fluorescein isothiocyanate (FITC; RA3-6B2; eBioscience), CD8-FITC and -PerCP (53-6.7; BD Biosciences PharMingen, San Diego, CA), CD25-FITC and -phycoerythrin (PE PC61.5; eBioscience), CD27-PE (LG.7F9; eBioscience), CD43-FITC (1B11; BioLegend, San Diego, CA), CD44-FITC and -allophycocyanin (APC; IM7; eBioscience), CD62L-FITC and -PE (MEL-14; eBioscience), CD122-PE (TM-β1; eBioscience), CD127-FITC and -PE (A7R34; eBioscience), CXCR3-PE (220803; R&D Systems), interferon-γ (IFN-γ)–PE (XMG1.2; eBioscience), KLRG1-biotin and -APC (2F1; eBioscience), CD45.1-FITC, -PE, -PE-Cy7, and -APC (A20; eBioscience), CD45.2-FITC, -PE, and -APC (104; eBioscience), CD11b-PE (M1/70; eBioscience), CD49β-APC (DX5; eBioscience), Gr1-FITC (RB6-8C5; eBioscience), IFN-γ–PE mAb (XMG1.2; eBioscience), IgM-PE (II/41; eBioscience), Ly6C-FITC (AL-21; BD Biosciences PharMingen), and NK1.1-PE (PK136; eBioscience). We also used the polyclonal antibody, anti–IL-15Rα-biotin (BAF551; R&D Systems). Streptavidin-PE and -APC were obtained from eBioscience. H-2K^b tetramers bound with the OVA-peptide, SIINFEKL, and conjugated with PE, were purchased from Beckman Coulter (Fullerton, CA). Staining for intracellular bromodeoxyuridine (BrdU) was performed with reagents and according to the instructions outlined in the FITC BrdU flow kit (559619; BD Biosciences). Flow cytometry was performed with a BD FACSCalibur, BD FACSAria, or BD LSR II cytofluorometer (BD Biosciences). Data were analyzed using FlowJo software (TreeStar, San Carlos, CA).

To evaluate the effects of exogenous IL-7 and IL-15 during CD8⁺ T-cell contraction, ovalbumin-specific CD8⁺ T cells (OT-I RAG1^−/−CD45.1⁺) were transferred into B6 recipient mice. At the peak of the CD8⁺ T-cell response to VSV-OVA infection, mice were treated with IL-7 or IL-15 for 12 days (Figure 1A). Here we used cytokine complexes as these reagents provide dramatically enhanced in vivo delivery of cytokines^{30,31,35,,,–39}  (Figure S1, available on the Blood website; see the Supplemental Materials link at the top of the online article). Both IL-7 and IL-15 increased the number of antigen-specific CD8⁺ T cells by approximately 10-fold in the spleen by the end of treatment (Figure 1B) in agreement with previous studies.^{11,,,,,,–18}  Interestingly, in the spleen (data not shown) and peripheral blood lymphocytes (PBLs; Figure 1C), the percentage of donor T cells was much higher with IL-15 complexes than with IL-7 complexes. This difference highlights the ability of IL-7 complexes to increase total splenocyte cell numbers without as great a proportional impact on the percentage of antigen-specific CD8⁺ T cells, possibly the result of the distinct ability of IL-7 complexes to expand immature B-cell precursors (Figure S1C,D). Of note, the increased numbers and percentages of donor OT-I CD8⁺ T cells observed after cytokine treatment were sustained in the PBLs and spleen (Figure S2) at approximately 2 months after infection in agreement other studies.^{11,14,16,17,40}  However, the enhanced percentage of donor OT-I CD8⁺ T cells induced after cytokine administration did appear to wane over extended periods of time (Figure S2E), as has also been observed with other studies.^16,17,40 

We also evaluated the response of endogenous CD8⁺ T cells in mice infected with LM-OVA using H-2K^b tetramers bound with the OVA-peptide (OVAp; Figure 1D). Here, we were able to ensure that cytokine treatment did not affect pathogen clearance by pretreating mice with ampicillin. At the onset of contraction, infected mice were treated with IL-7 complexes, IL-15 complexes, or vehicle alone. Similar to experiments with donor OT-I cells, IL-7 induced approximately 2-fold and IL-15 induced approximately 7-fold increases in absolute numbers of endogenous, antigen-specific CD8⁺ T cells in the spleen on day 15 (Figure 1E) and increased the percentage of endogenous OVAp-specific CD8⁺ T cells in the PBLs (Figure 1F).

Interestingly, we found that IL-7 and IL-15 rescued different CD8⁺ T-cell subsets based on expression of KLRG1 and CD127 during the contraction phase after infection (Figures 2A, S3). However, statistically significant differential induction by IL-7 and IL-15 was not observed for other markers tested, including CD27, CD43, CD44, CD62L, CXCR3, and Ly6C. Administration of IL-15 complexes resulted in approximately a 2- to 3-fold increase in the percentages of KLRG1^hiCD127^lo CD8⁺ T cells and administration of IL-7 complexes induced increases in the percentages of the reciprocal KLRG1^loCD127^hi and KLRG1^hiCD127^hi CD8⁺ T-cell subsets, compared with control treatment (Figure 2A,B). These cytokine-induced trends were evident when we assessed the phenotype of either donor OT-I CD8⁺ T cells responding to VSV-OVA infection or endogenous CD8⁺ T cells responding to LM-OVA infection (Figure 2B, left and right), although for IL-7 complexes, the preferential induction of KLRG1^loCD127^hi CD8⁺ T cells was much more prevalent in experiments using donor OT-I CD8⁺ T cells. It is noteworthy that at the memory time point, we continued to see the preferential retention of the KLRG1^hiCD127^lo CD8⁺ T-cell subset after IL-15 administration (Figure S2D). However, these KLRG1^hiCD127^lo CD8⁺ T cells diminish over time,^22,28,29  perhaps explaining why administration of IL-15, and the related cytokine IL-2, fail to result in sustained increases in memory CD8⁺ T cell numbers.^11,16,17,40 

Accumulation of absolute numbers of antigen-specific cells during cytokine treatment similarly revealed that KLRG1^hiCD127^lo CD8⁺ T cells contracted approximately 10 times more than KLRG1^loCD127^high CD8⁺ T cells in the spleen (Figure 2C). Importantly, we found that administration of IL-15, but not IL-7 complexes, reduced the contraction of KLRG1^hiCD127^lo CD8⁺ T cells. Thus, in IL-15-treated mice, there was a less than 2-fold decrease in numbers of either donor or endogenous KLRG1^hiCD127^lo CD8⁺ T cells, whereas these cells decreased 90- and 20-fold, respectively, in mice not receiving cytokine treatment. In contrast, both IL-7 and IL-15 complexes helped reduce the more limited contraction of KLRG1^loCD127^hi CD8⁺ T cells. Interestingly, both IL-7 and IL-15 complexes also reduced the contraction, and even promoted, the accumulation of KLRG1^hiCD127^hi CD8⁺ T cells. Cumulatively, these results show that, whereas IL-15 complexes broadly promote the retention of all the relevant cell populations, IL-7 complexes selectively promote accumulation of KLRG1^loCD127^hi and KLRG1^hiCD127^hi CD8⁺ T cells.

Although our results show that administration of exogenous cytokines can modulate the distribution of CD8⁺ T-cell subsets during contraction, an important question was whether endogenous IL-7 and IL-15 play similar roles in regulating these subsets. Thus, we examined the contraction phase of the OT-I CD8⁺ T-cell response to VSV-OVA in wild-type or IL-15^−/− mice. In addition, in a subset of mice, IL-7Rα signaling was blocked with an anti–IL-7Rα antibody as previously described.^25,26  With these conditions, we found that contraction of total CD8⁺ effector T cells was greater in the absence of IL-15 than IL-7 (Figure 3A). Without endogenous IL-15, contraction of the KLRG1^hiCD127^lo CD8⁺ T-cell subset was particularly severe (Figure 3B), in agreement with recently published results.²⁹  In contrast, when we blocked IL-7 in wild-type hosts, KLRG1^hi CD8⁺ T cells were present at higher proportions (Figure 3B). (Note that CD127 levels during antibody administration cannot be determined as the antibody used for blocking cytokine signaling also blocks flow cytometric staining of the receptor.) In the IL-15–deficient hosts that received IL-7 blockade, we did not obtain sufficient donor cells to assess their phenotype, indicating that IL-7 and IL-15 collaborate in rescuing CD8⁺ T cells from contraction. Similar to experiments with donor OT-I T cells, we found accelerated contraction of OVA-specific tetramer⁺ KLRG1^hiCD127^lo CD8⁺ T cells in the absence of IL-15 during LM-OVA infection (Figure 3C,D). Treatment of IL-15^−/− mice with IL-15 complexes rescued the KLRG1^hiCD127^lo antigen-specific CD8⁺ T cells (Figure 3D).

Because there may be limitations in blocking IL-7 signaling by antibody administration, we also followed the contraction of antigen-specific CD8⁺ T cells in IL-7^−/− mice. Here, we adoptively transferred day 7 OT-I effector cells generated in B6 hosts, into wild-type or IL-7^−/− hosts. Of note, contraction of the OT-I T cells was not enhanced in IL-7^−/− mice (Figure 3E). However, KLRG1^hiCD127^lo CD8⁺ T cells dominated the percentage of surviving CD8⁺ effector T cells (Figure 3F), suggesting that, whereas the antigen-specific CD8⁺ T-cell population could survive in the absence of IL-7, the remaining cells were phenotypically distinguishable.

The substantial increase in percentage and absolute numbers of KLRG1^hiCD127^lo CD8⁺ T cells after treatment with IL-15 complexes suggested proliferation and/or survival were enhanced. To directly assess the proliferation of KLRG1 subsets, we adoptively transferred OT-I CD45.1⁺ CD8⁺ T cells into B6 mice and infected them with LM-OVA. Nine days after infection, BrdU was given for 3 days. We observed no BrdU incorporation by splenocytes in either KLRG1^hi or KLRG1^lo CD8⁺ T-cell subsets in the absence of cytokine treatment (Figure 4A). Treatment with IL-7 or IL-15 complexes induced 3- to 7-fold increases in BrdU incorporation by the donor KLRG1^lo CD8⁺ T cells. However, despite being present at a higher proportion after IL-15 complex administration, donor KLRG1^hi CD8⁺ T cells were essentially refractory to either IL-15- or IL-7–induced proliferation (Figure 4A). Although KLRG1^hi CD8⁺ T cells were resistant to cytokine-mediated proliferation in vivo, they produced IFN-γ after stimulation with OVAp in agreement with others (Figure S4).^28,29,41 

To rule out the possibility that KLRG1^hi CD8⁺ T cells convert to KLRG1^lo CD8⁺ T cells on cell division, we sorted KLRG1^lo and KLRG1^hi effector subsets on day 10 of LM-OVA infection. The sorted KLRG1^lo and KLRG1^hi CD8⁺ T cells were labeled with CFSE and adoptively transferred into secondary B6 recipients that were then treated with cytokine complexes for 6 days. As we observed with the BrdU analysis, both IL-7 and IL-15 complexes induced rapid proliferation of the KLRG1^lo CD8⁺ T cells, but neither induced significant proliferation of the KLRG1^hi subset as assessed by CFSE dilution (Figure 4B). Furthermore, there was no conversion in phenotype between KLRG1^lo and KLRG1^hi populations (Figure 4C). These results show that the accumulation of KLRG1^hi antigen-specific T cells during contraction with IL-15 complex treatment is the result of increased survival of KLRG1^hi CD8⁺ T cells, not the result of proliferation or conversion between the KLRG1^lo and KLRG1^hi populations.

The simplest explanation for why IL-7 and IL-15 differentially affect KLRG1^hi and KLRG1^lo CD8⁺ T cells is by virtue of their expression of the receptor subunits necessary for cytokine responsiveness. KLRG1^hi and KLRG1^lo CD8⁺ T cells both express the IL-2/IL-15 receptor components necessary for survival and cytokine signaling (CD122 and CD132; Figure S5). However, as shown earlier, KLRG1^hi CD8⁺ T cells, but not KRLG1^lo CD8⁺ T cells, are largely IL-7Rα (CD127) low. This differential expression of IL-7Rα provides a simple explanation for why KLRG1^hi CD8⁺ T cells do not respond functionally to IL-7, but it fails to explain why KLRG1^hi and KRLG1^lo CD8⁺ T cells respond differently to IL-15. Addressing this question will require elucidation of the intracellular signals involved in these 2 cell types, although it has been suggested that expression of p27^kip and Bmi-1 may be relevant.^15,42 

Given that the receptors for IL-15 also function in IL-2 signaling, our results predicted that administration of IL-2 would rescue cells during contraction in a manner similar to IL-15. To test this, we administered IL-2 complexes after infection and assessed the response of donor OT-I CD8⁺ T cells and endogenous antigen-specific CD8⁺ T cells after infection. In both cases, IL-2 not only increased the absolute numbers and percentages of responding antigen-specific CD8⁺ T cells, but also preferentially expanded the KLRG1^hiCD127^lo CD8⁺ T-cell subset (Figure 5A-E). IL-2 complexes also induced enhanced BrdU-incorporation in KLRG1^lo CD8⁺ T cells but not KLRG1^hi CD8⁺ T cells (Figure 5F). Thus, similar to IL-15, IL-2 also preferentially induces the accumulation of KLRG1^hiCD127^lo CD8⁺ T cells through enhanced survival as opposed to increased proliferation.

This study provides novel insights into cytokine-regulated survival of different CD8⁺ T-cell subsets during the contraction phase and seeding of the memory T-cell compartment after infection. Using expression of KLRG1 and CD127, we identify distinct populations of CD8⁺ T cells with differential responsiveness to cytokines. Administration of IL-15 or IL-2 complexes during contraction drives the preferential accumulation of KLRG1^hiCD127^lo CD8⁺ T cells while also promoting increased survival of other CD8⁺ T-cell subsets. In contrast, IL-7 complexes preferentially supported the accumulation of KLRG1^loCD127^hi and KLRG1^hiCD127^hi CD8⁺ T cells, with very limited impact on CD127^lo CD8⁺ T-cell subsets. The preferential ability of IL-15 and IL-2 complexes to expand KLRG1^hiCD127^lo CD8⁺ T cells was not expected as these populations express similar levels of receptor components, whereas the ability of IL-7 to selectively drive CD127^hi CD8⁺ T cells was more predictable based on IL-7Rα expression. Strikingly, we observed that KLRG1^hi but not KLRG1^lo CD8⁺ T cells exhibited a broad resistance to IL-15- and IL-2-mediated proliferation in vivo despite the use of cytokine complexes, which exhibit enhanced biologic activity. Combined with evidence that cytokine administration does not induce rapid conversion of KLRG1^lo to KLRG1^hi CD8⁺ T cells, our results show that IL-15 and IL-2 complexes dramatically improve the survival of KLRG1^hi CD8⁺ T cells. Consideration of these findings is useful in understanding the basic role of cytokines in promoting CD8⁺ T cell responses after infection and relevant to efforts that augment the efficacy of CD8⁺ T cell–directed therapies.

The functional properties and significance of CD8⁺ T-cell subsets based on KLRG1 and CD127 expression are not fully understood. For example, expression of KLRG1 on CD8⁺ T cells has been reported to be a marker of cellular senescence and an indicator of the inability to respond to antigenic challenge.⁴¹  A recent study found that forced expression of the IL-7Rα subunit on KLRG1^hiCD127^lo CD8⁺ T cells failed to restore IL-7–driven proliferation, suggesting these cells may have inherent deficiencies in cytokine-induced proliferation that extend beyond the simple absence of IL-7Rα expression.¹⁵  However, whereas some studies find that these cells exhibit a proliferative defect, others have reported that KLRG1^hi CD8⁺ T cells can undergo vigorous proliferation after antigen challenge.⁴³  Indeed, we find that day 10 KLRG1^hi effector CD8⁺ T cells can undergo significant proliferation after VSV-OVA infection (data not shown). The differences in these findings have yet to be explained; however, there is agreement that KLRG1^hi CD8⁺ T cells are not impaired in effector functions, such as IFN-γ production and cytotoxic killing compared with KLRG1^lo CD8⁺ T cells.^28,29,41  Relevant to our findings, it has been shown that the KLRG1^hiCD127^lo phenotype is associated with CD8⁺ T cells that are selectively lost over a period of months after the onset of contraction, as opposed to KLRG1^loCD127^hi phenotype that contains longer-lived memory CD8⁺ T cells.^22,28,29 

Our observation that IL-15 and IL-2 levels can preferentially impact the survival of these KLRG1^hi CD8⁺ T cells provides a new perspective on the significance of this T-cell subset. IL-15 and IL-2 can be elevated during periods of extended inflammation; our data suggest that such an environment would favor the maintenance of these KLRG1^hi CD8⁺ T cells. As these cells are functional in terms of both cytotoxicity and cytokine production, the increased numbers of KLRG1^hi CD8⁺ T cells could play a vital role in enhancing the clearance of prolonged infection. Importantly, as the survival of these KLRG1^hi CD8⁺ T cells appears tightly linked to the presence of elevated levels of IL-15 and IL-2, these cells would probably die soon after cytokine availability is reduced and the activity of these T cells might be harmful. In agreement, antibiotic treatment of Listeria-infected mice, which reduces inflammation, results in enhanced frequencies of KLRG1^loCD127^hi CD8⁺ T cells.⁴⁴  However, if antibiotic-treated mice are given the IL-15-inducing agent, CPG,⁴⁵  CD127^lo (and presumably KLRG1^hi) CD8⁺ T cells emerge. Similar results extended these findings and linked T-bet as a critical factor transcription factor in maintaining the functional characteristics of KLRG1^hi CD8⁺ T cells including expression of the IL-15 receptor subunit, CD122.²⁹ 

IL-2, IL-7, and IL-15 have all garnered interest as reagents for augmenting CD8⁺ T-cell immunity either in treating disease or enhancing the recovery of lymphocytes after cytoreductive therapy.^6,7,46,–48  In these regards, one critical issue is the selection of cytokine. Our results would suggest prioritizing the use of IL-7 where long-term immunity is the primary objective, as the KLRG1^lo CD8⁺ T-cell subset, which is thought to be long-lived and seed the memory compartment, is selectively expanded with this cytokine. In contrast, IL-2 or IL-15 should more efficiently increase effector cell numbers, adding to the functional arsenal, by supporting survival of KLRG1^hi CD8⁺ T cells. Indeed, the latter 2 cytokines have proven particularly useful at augmenting tumor immunotherapy, a situation where the expansion of effector function to eliminate tumor is important. It is worth noting, though, that vaccination or antigen-stimulation during IL-2 administration has in some cases been shown to be counterproductive.^49,50  Thus, it will be important to determine the optimal timing of vaccination in conjunction with cytokine administration to best enhance the desired CD8⁺ subset.

It is relevant to mention that, in our study, we administered cytokine complexes instead of free cytokines. Although the mechanism of how cytokine complexes exhibit improved biologic activity is not fully understood, it is thought that this may relate to improved half-life. Although it has been shown that administration of IL-15 cytokine complexes can result in improved donor CD8⁺ T-cell immune responses,³⁰  our study shows, for the first time, that the response of endogenous antigen-specific CD8⁺ T cells present at physiologic levels can also be dramatically augmented after administration of IL-2, IL-7, or IL-15 cytokine complexes. These findings represent an important step in translating the use of such reagents to the clinic where low numbers of antigen-specific T cells will probably be present.

It is becoming increasingly well recognized that the administration of cytokines to modulate immune responses has great potential in medicine. Although many cytokines are often viewed as having overlapping properties suggesting the potential for interchangeable use, our study finds distinct differences between IL-2, IL-7, and IL-15, not previously appreciated. Consideration of our findings, combined with a further understanding of the functional roles of these distinct CD8⁺ T-cell subsets, will aid in the development of therapies for augmenting T cell–based responses against infectious disease and cancer.

The authors thank Karen Ong for excellent help in flow cytometric analysis and cell sorting and K. Cheung, A. Doedens, W. D'Souza, L. D'Cruz, M. Salem, and M. Werneck for critical review of this manuscript.

This work was supported by the National Institutes of Health (NIH; grant RO1AI67545), Cancer Research Institute, V Foundation for Cancer Research, California Breast Cancer Research Program, and Pew Scholars Program (A.W.G.). M.P.R. is supported by a fellowship from the Prevent Cancer Foundation. J.A.B. is supported by the University of California, San Diego Immunology training grant from the NIH (National Research Service Award T32AI060536-02). J.F.P. is supported by a CJ Martin Fellowship from the Australian National Health and Medical Research Council.

National Institutes of Health

Contribution: M.P.R. designed and performed research, interpreted data, and wrote the paper; N.A.L., J.F.P., P.F., J.A.B., and P.A.M. designed and performed research and interpreted data; and C.D.S. and A.W.G. designed research, interpreted data, and wrote the paper.

Conflict-of-interest disclosure: C.D.S. is a shareholder in Nascent Biologics Inc. The remaining authors declare no competing financial interests.

Correspondence: Ananda W. Goldrath, 9500 Gilman Drive, University of California, San Diego, La Jolla, CA 92093-0377; e-mail: agoldrath@ucsd.edu.