Circulating IL-15 exists as heterodimeric complex with soluble IL-15Rα in human and mouse serum

Abstract

IL-15 is an important cytokine for the function of the immune system, but the form(s) of IL-15 produced in the human body are not fully characterized. Coexpression of the single-chain IL-15 and the IL-15 receptor alpha (IL-15Rα) in the same cell allows for efficient production, surface display, and eventual cleavage and secretion of the bioactive IL-15/IL-15Rα heterodimer in vivo, whereas the single-chain IL-15 is poorly secreted and unstable. This observation led to the hypothesis that IL-15 is produced and secreted only as a heterodimer with IL-15Rα. We purified human IL-15/IL-15Rα complexes from overproducing human cell lines and developed an ELISA specifically measuring the heterodimeric form of IL-15. Analysis of sera from melanoma patients after lymphodepletion revealed the presence of circulating IL-15/IL-15Rα complexes in amounts similar to the total IL-15 quantified by a commercial IL-15 ELISA that detects both the single-chain and the heterodimeric forms of the cytokine. Therefore, in lymphodepleted cancer patients, the serum IL-15 is exclusively present in its heterodimeric form. Analysis of the form of IL-15 present in either normal or lymphodepleted mice agrees with the human data. These results have important implications for development of assays and materials for clinical applications of IL-15.

Figure 1

Schematic representation of ELISAs for detection of human IL-15. (A) Total IL-15 was detected by a commercially available IL-15 ELISA (R&D Systems, Quantiglo kit). The capture and detection monoclonal anti–human IL-15 antibodies target different accessible regions of IL-15 molecule. (B) IL-15/IL-15Rα heterodimers were detected using a monoclonal anti–human IL-15 as capture antibody and a polyclonal anti–human IL-15Rα as detection antibody. Single-chain IL-15 or purified sIL-15Rα tested separately were not detected by this heterodimeric IL-15/IL-15Rα ELISA.

Figure 2

Polyacrylamide gel analysis of IL-15/sIL-15Rα complex purified from an overproducing human cell line. (A) HPLC-purified human IL-15 and sIL-15Rα produced by a HEK293-derived cell line were mixed in a molar ratio 1:1, and IL-15/IL-15Rα complexes were visualized in Coomassie blue-stained denaturing SDS-polyacrylamide gel (top panel) and in Coomassie blue-stained nondenaturing polyacrylamide gel (bottom panel). (B-C) IL-15/IL-15Rα shown in panel A was detected by Western immunoblot. Complexes were analyzed under denaturing condition (top panels) or under nondenaturing condition (bottom panels) using a goat anti–human IL-15Rα antibody (B) or a goat anti–human IL-15 antibody (C).

Figure 3

Soluble IL-15/IL-15Rα protein standard for the heterodimeric IL-15/IL-15Rα ELISA. Standard curve for human heterodimeric IL-15/IL-15Rα ELISA. HPLC-purified human IL-15 and sIL-15Rα were quantified by amino acid mass analysis and mixed in a molar ratio 1:1. Triplicates of 3-fold serial dilutions of IL-15/IL-15Rα complexes were assayed in the heterodimeric IL-15/IL-15Rα ELISA. Circles represent RLU average for each IL-15/IL-15Rα concentration tested from 5 independent assays; SD for the 5 assays is shown. The intra-assay variability was 3% to 12% coefficient of variation (CV), and the interassay variability was 18% to 33% CV. Open squares represent RLU values of the indicated amounts of single-chain IL-15 tested in the heterodimeric IL-15/IL-15Rα ELISA (negative control). Linear regression analysis showed no significant correlation (r² = 0.32, P = .32)

Figure 4

Detection of IL-15 and heterodimeric IL-15/IL-15Rα complexes in human sera. (A) Total IL-15 (filled bars) and sIL-15Rα–associated IL-15 (open bars) in sera of 10 lymphodepleted melanoma patients. Total IL-15 was evaluated using the commercially available IL-15 ELISA (Quantiglo Q1500B). sIL-15Rα–associated IL-15 was detected by the heterodimeric IL-15/IL-15Rα ELISA using capture anti–IL-15 antibody and detection anti–IL-15Rα antibody, as described in “IL-15 and IL-15/IL-15Rα ELISA.” Data are mean ± SEM of 2 to 4 independent measurements. *Value below the limit of detection of the assay. (B) Correlation of total IL-15 and sIL-15Rα–associated IL-15 levels in the sera of lymphodepleted patients. Graph represents mean ± SEM from 2 to 4 independent measurements for each sample. Data were fitted to a linear regression curve using Prism 4.0a software package (correlation coefficient r² = 0.92; slope = 0.92).

Figure 5

Circulating human IL-15 is exclusively present as heterodimeric IL-15/sIL-15Rα complex. Serum samples from lymphodepleted cancer patients were preincubated with exogenous sIL-15Rα to allow the in vitro formation of IL-15/sIL-15Rα complexes from any free IL-15 potentially present in the sample. Serum samples were next evaluated for the amount of sIL-15Rα–associated IL-15 by ELISA. Serum sample S8, S9, or S10 was preincubated with increasing sIL-15Rα concentrations (0-2250 pg/mL; left panel). Serum samples S1, S2, S4, S5, and S7 were preincubated with a fixed amount of sIL-15Rα (750 pg/mL; right panel). The assay was repeated twice.

Figure 6

Demonstration of IL-15/sIL-15Rα heterodimers in normal and lymphodepleted mice. (A) Serum levels of IL-15/IL-15Rα heterodimer in normal mice () and on CTX treatment (□). Mice were bled at different time points (days 1-19) after CTX or placebo administration. Circulating IL-15/IL-15Rα was assessed by ELISA (eBioscience) and reported as mean ± SEM of 6 to 33 mice per time point. (B-C) Serum samples from 6 untreated (B) and 9 CTX-treated (day 3 after treatment; C) mice were preincubated with exogenous recombinant mouse IL-15Rα-Fc (100-1000 pg/mL, as indicated) to allow the in vitro formation of IL-15/sIL-15Rα complexes with any free IL-15 potentially present in the sample. Serum samples were next evaluated for the amount of sIL-15Rα–associated IL-15 by ELISA.

Figure 7

Different kinetics of IL-15/IL-15Rα heterodimers after CTX administration or whole-body irradiation. Six- to 8-week-old C57BL/6 mice were treated with CTX (□) or exposed to 6 Gy of whole-body irradiation (▴). Arrows indicate day 7 after treatment. (A). Mice were bled at the indicated time points after treatment, and the circulating IL-15/IL-15Rα heterodimers were measured by ELISA (eBioscience) and reported as mean ± SEM of 2 to 33 mice per time point. represents the average level of circulating mouse IL-15/IL-15Rα from 10 normal mice. **P = .007. (B-C) Mice were killed at the indicated time points after treatment, and spleens were collected for analysis. Total number of NK cells (B) and CD8⁺ T cells (C) per spleen over time was determined by flow cytometry. (D) Absolute count of lymphocytes in blood at day 7 after cytoreductive treatments. Three mice per group were analyzed. Hematologic profile on 10 untreated mice was also performed as control. **P < .01. ns indicates not significant.