Antibodies for flow cytometry were purchased from BD Biosciences or eBioscience. The CF1 antibody directed against a non-neutralizing epitope of IL-2Rβ was obtained from Beckman Coulter (Villepinte, France) and the blocking A41 anti-IL-2Rβ antibody, produced in house, has been previously characterized (49, 50). The NANTIL-15 molecules correspond to human IL-15 bearing the triple mutation L45D-N65K-L69R and linked to either a human Fc (IgG1) for in vitro experiments or a murine Fc (IgG2a) for in vivo experiments.

Recombinant 6xHis tagged human IL-15, IL-15Rα-Linker-IL-15 (RLI), IL-15_D8S, NANTIL-15, and Fc-IL-15 were produced by Evitria (Schlieren, Switzerland) and purified by HIS TRAP Excel column using an AKTA start system (Cytiva, Versailles, France). Fc-IL-15Rα and recombinant human IL-2Rβ were obtained from R&D Systems (Lille, France) Fc-IL-2Rβ/γc was kindly provided by Cytune Pharma (Nantes, France) and was produced using a Knobs-In-Holes strategy from CHO supernatant and purified on protein A, Nuvia HR-S, and gel filtration by Evitria (Switzerland). Recombinant human IL-2 was purchased from Chiron (Emeryville, CA, USA).

HEK-293 cell line was cultured in Dulbecco’s modified Eagle’s medium (DMEM) medium (Thermo Fisher, Illkirch, France) containing 10% of heat-inactivated fetal calf serum (FCS) (Gibco Illkirch, France), 2 mM glutamine, and 1 mg/ml glucose. Stably transfected HEK-293 cells expressing membrane-bound IL-15Rα (IL-15Rα-HEK-293 cells) (51) were cultured in the same medium supplemented with 50 ng/ml puromycin. Kit225 and Kit225-15RαKO (48) cells were cultured as described previously (51). The CTLL-2 mouse cytotoxic T-lymphocyte cell line was cultured in Roswell Park Memorial Institute (RPMI)-1640 medium (Sigma-Aldrich, Saint-Quentin-Fallavier, France) containing 8% heat-inactivated FCS, 2 mM glutamine, 10 µM 2-mercaptoethanol, and 15 ng/ml human rIL-2. All cell lines were maintained at 37°C, in a humidified, 5% CO₂ atmosphere.

C57BL/6 and DBA/1 mice were purchased from the Janvier Labs (Le Genest St. Isle, France). Mice were generally used between the age of 6 and 12 weeks. The procedures were approved by the French Ethics Committee for Animal Experimentation and performed according to European Union guidelines.

NANTIL-15 plasma levels were evaluated by ELISA using a human Fc-IL-15Rα (R&D Systems) as capture and an horseradish peroxidase (HRP) goat anti-mouse IgG (H+L) (INTERCHIM, Montluçon, France) as the revealing antibody. A standard curve was performed with NANTIL-15.

The biosensor used in this study was a Biacore T200 instrument (Cytiva). CM5 Research Grade Sensor Chips (carboxy-methyl-dextran derivatized with carboxyl group) and HBS-EP [0.01M HEPES, pH 7.4, 0.15M NaCl, 0.005% (v/v) surfactant P20, and 3 mM EDTA] running buffer were purchased from Cytiva. All surface plasmon resonance (SPR) experiments were performed at a flowrate of 40 µl/min at 25°C. Fc-IL-15Rα, recombinant human IL-2Rβ, or anti-IL-2Rβ CF1 antibody was coupled onto the CM5 surface following the standard amine coupling protocol (according to supplier’s procedures) to achieve a residual coupling response of 300, 250, and 8,000 RU, respectively.

SPR Kinetic Analysis on Fc-IL-15Rα and Recombinant Human IL-2Rβ. Recombinant NANTIL-15 protein was diluted in HBS-EP at concentrations ranging from 3.12 to 200 nM and injected over the Fc-IL-15Rα or the IL-2Rβ immobilized surface in a kinetic mode. Association and dissociation were monitored for 3 and 10 min, respectively. A 10 mM Glycin pH 1.5 solution was injected over the chips for 30 s for regeneration between each cycle. Rmax value (RU), k_on (M⁻¹ s⁻¹), k_off (s⁻¹), and Kd (M) were calculated from kinetics sensorgrams using the Langmuir 1:1 model.

SPR Kinetic Analysis on Recombinant Dimeric Fc-IL-2Rβ/γc Receptor. A capture strategy was used for SPR analysis on Fc-IL-2Rβ/γ. Briefly, recombinant Fc-IL-2Rβ/γc diluted at 200 nM in HBS-EP was injected over the CF1 immobilized chip for 1 min. After a 1-min stabilizing period, NANTIL-15 or Fc-IL-15 diluted in HBS-EP at concentrations ranging from 31.5 to 500 nM were injected over the Fc-IL-2Rβ/γ captured surface in a Single Cycle Kinetics (SCK) mode. Association and dissociation were monitored for 2 and 10 min, respectively. A 10 mM Glycin pH 1.5 solution was injected over the chips for 30 s for regeneration between each capture cycle. Rmax value (RU), k_on (M⁻¹ s⁻¹), k_off (s⁻¹), and Kd (M) were calculated from Single Cycle kinetics sensorgrams using the Langmuir 1:1 model.

The proliferation of Kit225, Kit225-15RαKO, and CTLL-2 cells in response to IL-2, IL-15, IL-15_D8S, and RLI was assessed by Alamar blue reduction assay (AbDSerotec, Marnes-la-Coquette, France). Cells were starved in the culture medium without cytokine for 4 h in case of CTLL-2 cells to reduce background or overnight in case of Kit225 and Kit225-15RαKO cells. Viability of cells was controlled by trypan blue staining. Then, cells (1 × 10⁴) were preincubated for 30 min with increasing concentrations of NANTIL-15 and then cultured for 2.5 days in the medium supplemented with IL-2 (100 pM), IL-15 (10 and 100 pM), IL-15_D8S (10 and 100 pM), or RLI (100 pM). Cell proliferation was revealed by adding Alamar blue (10 µl) to each well and, after a 6-h incubation period at 37°C, by measuring the emitted fluorescence at 590 nm under excitation at 560 nm using an EnSpire Multimode Plate Reader (PerkinElmer, Courtaboeuf, France).

Detection of pStat5 was carried out using either flow cytometry or the AlphaLISA SureFire Ultra™ assay kit (PerkinElmer). For both assays, exponentially growing Kit225 or Kit225-15RαKO cells were washed and starved overnight in the culture medium without cytokines to reduce basal phosphorylation. Cells were preincubated for 30 min with a fixed or increasing concentrations of NANTIL-15 or A41 anti-IL-2Rβ antibody and then stimulated at 37°C for 1 h with fixed concentrations of IL-15, RLI, or IL-15 mutants. For IL-15 trans-presentation experiments, Kit225 cells labeled with BD Horizon Violet Proliferation dye 450 (VPD-450; BD Biosciences, Le Pont de Chaix, France) were cocultured for 1 h with IL-15Rα HEK-293 cells preloaded for 30 min with IL-15.

For flow cytometry pStat5 staining, 3 × 10⁵ stimulated cells were fixed for 15 min at 37°C using Fix Buffer I (BD Biosciences) and permeabilized in Perm Buffer III (BD Biosciences) for 20 min at 4°C. Then, cells were labeled with phycoerythrin (PE)-conjugated mouse anti-human Stat5 (Y694) or PE-conjugated control isotype (BD Biosciences) for 1 h at room temperature. Cells were analyzed using BD FACS Calibur, or BD FACS Canto 2 (BD Biosciences) and FlowJo software.

For AlphaLISA assay, 2 × 10⁵ stimulated cells were lysed, and Stat5 phosphorylation was measured according to the manufacturer’s instructions using pStat5 AlphaScreen Surefire kit. Then, the signal in the wells was detected using an EnSpire Multimode Plate Reader (PerkinElmer).

Bioavailability of NANTIL-15 was evaluated in male C57BL/6 mice after a single intravenous injection of 10 µg of NANTIL-15. At each time point following injection, blood was collected (three mice per point) and centrifugated, and the plasma was frozen at −20°C. NANTIL-15 plasma levels were measured by ELISA (see above). Pharmacokinetic parameters were calculated using a one-compartment model with GraphPad Prism software.

The effect of repeated injections of NANTIL-15 was evaluated in vivo in male C57BL/6 mice. Mice received one daily i.p. injection of 1.5, 5, or 25 µg of NANTIL-15 or phosphate-buffered saline (PBS) over 4 or 14 days. The day following the last injection, mice were sacrificed, and spleens were collected for NK and CD8⁺ T cells by flow cytometric analysis.

Male DBA/1 mice 6–8 weeks of age (Janvier Labs, Le Genest St. Isle, France) were used in all experiments. Type II collagen from bovine tracheal cartilage (Sigma C1188) was dissolved in 0.1M acetic acid at 2 mg/ml at 4°C overnight. The collagen solution was then emulsified using a homogenizer with a small blade with an equal volume of Freund’s complete adjuvant (Difco 231131) as previously described (40). To induce CIA, 200 µg of collagen was injected intradermally at the base of the tail of DBA/1 mice. Twenty-one days after immunization, the animals were challenged with 200 µg of collagen i.p. in 0.05M acetic acid. Mice were monitored every day to observe the apparition of clinical signs of arthritis based on the following criteria: grade 0, normal joints and no swelling; grade 1, mild swelling and/or erythema; grade 2, pronounced edema or redness of the paw or several digits; and grade 3, severe swelling of entire paw and/or ankylosis. As each limb was individually graded, the maximal clinical score for each mouse was 12. The thickness of each paw was also measured using a caliper, and the mean paw thickness was calculated. Administration of NANTIL-15 or IgG2a (eBioscience 14-4724) treatment (daily i.p. injection at 1.5 or 5 µg/injection/day for 14 days, based on 39) was either started in a randomized manner after the establishment of the arthritis (i.e., clinical score at least equal to 1) or 1 day after the second immunization. Spleens were collected 1 h after the last injection of NANTIL-15 for flow cytometric analysis of NK and CD8 T cells. Paws were removed post-mortem, fixed in 1% paraformaldehyde for 3 days, and decalcified in 5% EDTA for 2 weeks before being embedded in paraffin for sectioning. Serial sagittal sections were made, mounted on glass slides, and stained with H&E and for CD8 marker. Tissue sections were analyzed by light microscopy.

Single-cell suspensions from spleen cleared from RBCs were obtained and stained. Before staining, Fc receptors were blocked using anti-CD16/32 Ab (BD Biosciences). Then, cells were incubated with Abs for cell surface staining. Spleen NK and CD8 T cells were determined by using a FACS Calibur flow cytometer (Becton Dickinson). For cell surface staining, mouse splenocytes were incubated with either fluorescein isothiocyanate (FITC)-conjugated anti-mouse CD3e mAb, APC-conjugated anti-mouse NKp46 mAb (52, 53), and FITC-conjugated anti-mouse CD11b mAb, or FITC-conjugated anti-mouse CD3e mAb, APC-conjugated anti-mouse CD8 mAb, PE-conjugated anti-mouse CD122 mAb, and FITC-conjugated anti-mouse CD44 mAb. Data were analyzed using FlowJo software (BD Biosciences).

Statistical analyses were performed using Mann–Whitney test and Kruskal–Wallis test with Dunn’s post-test (GraphPad Prism Software). Data are expressed as mean ± SEM. A p-value <0.05 was considered significant (*p < 0.05; **p < 0.01; ***p < 0.001).

We sought to selectively inhibit the action of IL-15 through the trimeric IL-15Rα/IL-2Rβ/γc receptor to assess whether the pro-inflammatory property of IL-15 is due to binding to its specific receptor chain, IL-15Rα. To this end, we designed NANTIL-15 that retains binding to IL-15Rα while not recruiting the IL-2Rβ transducing receptor chain.

We confirmed that NANTIL-15’s binding parameters to IL-15Rα were comparable to that of wild-type IL-15, whereas its binding properties to IL-2Rβ chain was completely impaired ( Figure 1B ) (54). We also observed that NANTIL-15 was not capable of binding to immobilized recombinant dimeric Fc-IL-2Rβ/γc receptor, suggesting that blocking the binding to IL-2Rβ chain is sufficient to fully impair the binding to the dimeric receptor ( Figure 1C ). We further investigated the IL-15Rα/IL-2Rβ/γc binding selectivity of NANTIL-15 by using Kit225 cells, expressing the three receptor chains and thus both trimeric IL-15Rα/IL-2Rβ/γc and dimeric IL-2Rβ/γc forms (48), and also Kit225-15RαKO cells in which il15rα gene was invalidated and therefore expressing only the dimeric IL-2Rβ/γc receptor (48). According to our previous results, NANTIL-15 binding was detected at the surface of Kit225 but not of Kit225-15RαKO cells, whereas Fc-IL-15 binding was detected at the surface of both cells ( Figure 1D ). Taken together, these data show that abrogating the recruitment of IL-15 to IL-2Rβ is sufficient to fully impair the binding to dimeric IL-2Rβ/γc receptor, while retaining the binding to IL-15Rα.

Then, we evaluated the functional inhibitory efficacy and the receptor specificity of NANTIL-15 ( Figure 2A ). First, we confirmed that NANTIL-15 was not capable of inducing either Stat5 phosphorylation or proliferation of Kit225-15RαKO cells, expressing only the dimeric IL-2Rβ/γc receptor ( Figures 2B , left panel, C, D ). As expected, NANTIL-15 was unable to inhibit IL-15-induced Stat5 phosphorylation in Kit225-15RαKO cells, whereas A41 anti-IL-2Rβ antibody was able to effectively inhibit this signal ( Figures 2B, C ). Then, we took advantage of the RLI fusion molecule that selectively signals through the dimeric IL-2Rβ/γc receptor, regardless of the expression of IL-15Rα. On Kit225-15RαKO cells, Stat5 phosphorylation and cell proliferation induced by a fixed concentration of IL-15 or RLI were challenged by increasing concentrations of NANTIL-15. NANTIL-15 did not affect these signals in either condition ( Figures 2C, D ). Taken together, these data show that NANTIL-15 does not affect IL-15 signaling through the dimeric IL-2Rβ/γc receptor.

Then, we wondered whether NANTIL-15 could affect IL-15 trans-presentation ( Figure 2E ). NANTIL-15 or Fc-IL-15 was loaded onto HEK-293 cells stably transfected with IL-15Rα and co-cultured in the presence of responding Kit225 cells (51). No Stat5 phosphorylation was observed when Kit225 cells were in the presence of NANTIL-15-loaded HEK-293-IL-15Rα cells, whereas pStat5 was detected when Kit225 cells were co-cultured with Fc-IL-15-loaded HEK-293-IL-15Rα cells (data not shown). These data indicate that NANTIL-15 does not signal in trans.

It has been shown that IL-15 trans-presentation requires the association of IL-15 with IL-15Rα inside the presenting cells, prior from emerging to the cell surface (14), and this mode of action is the dominant mechanism by which IL-15 delivers its signal in physiological conditions (55). To test whether, NANTIL-15 could inhibit IL-15 trans-presentation, wild-type IL-15 was first loaded to HEK-293-IL-15Rα and co-cultured with responding Kit225 cells, in the presence or in the absence of NANTIL-15. We observed that IL-15 trans-presentation was not inhibited by the presence of NANTIL-15, indicating that NANTIL-15 was not able to chase IL-15 bound to IL-15Rα ( Figure 2F ). Thus, NANTIL-15 does not affect IL-15’s action through the dimeric IL-2Rβ/γ receptor either in cis and in trans.

We further evaluated the inhibitory efficacy and the receptor specificity of NANTIL-15 using Kit225 cells expressing both types of receptors, including the trimeric IL-15Rα/IL-2Rβ/γc receptor ( Figure 3A ). Interestingly, NANTIL-15 partially inhibited IL-15-induced Stat5 phosphorylation, reaching a plateau at 60% of inhibition with an IC50 of 1 nM ( Figures 3B, C ). These results suggested that NANTIL-15 is capable of selectively impairing IL-15’s signaling through the trimeric IL-15Rα/IL-2Rβ/γc receptor without affecting that through the dimeric IL-2Rβ/γc receptor. To confirm that hypothesis, we took advantage of the IL-15_D8S mutant that is only able to signal through the trimeric IL-15Rα/IL-2Rβ/γc receptor (48). A fixed concentration of IL-15_D8S was submitted to increasing concentrations of NANTIL-15 on Kit225 cells. NANTIL-15 fully abrogated IL-15_D8S-induced pStat5 signal with an IC50 of 1 nM comparable to that of wild-type IL-15 ( Figure 3C ). Cell proliferation assays confirmed these observations, with an IC50 below 1 nM ( Figure 3D ). According to our previous results, NANTIL-15 was ineffective to inhibit pStat5 and the proliferation of Kit225 cells when stimulated with RLI, confirming that NANTIL-15 does not inhibit IL-15 signaling through the dimeric IL-2Rβ/γc receptor ( Figures 3C, D ). Taken together, these data show that NANTIL-15 is a selective potent antagonist of IL-15 acting in cis through the trimeric IL-15Rα/IL-2Rβ/γc receptor without affecting its cis action through the dimeric IL-2Rβ/γc receptor. Thus, NANTIL-15 inhibitory property is closely related to the expression of IL-15Rα on the cell surface of target cells.

To test the efficacy of NANTIL-15 to inhibit IL-15’s action in mice, we first took advantage of the CTLL-2 murine cell line, known to express murine IL-15Rα, IL-2Rα, IL-2Rβ, and γc receptor chains. As observed for human cells, NANTIL-15 was also able to inhibit IL-15-induced signal on CTLL-2 cells, whereas IL-2-induced signal remained unaffected ( Figure 4A ). In mice, we first determined NANTIL-15’s plasmatic half-life, estimated at around 6 h ( Figure 4B ). Then, to evaluate the relative impact of the administration of NANTIL-15 to IL-15’s dependent cells, such as NK and CD44⁺CD8 T cells (6), increasing doses (from 1.5 to 25 μg) of NANTIL-15 were injected in 4 consecutive days. No effect on NK and CD8 T-cell populations could be observed even at the highest dose (25 μg for a total amount of 100 μg) ( Figure 4C ). We also observed that repeated injections of 1.5 μg of NANTIL-15 every day during 2 weeks (total of 21 μg) did not affect either of these populations ( Figure 4D ). Moreover, NK cell maturation, revealed by the CD11b marker, was not affected either. Thus, NANTIL-15 injections do not disturb NK and CD8 T-cell populations in healthy mice.

IL-15 overexpression has been observed in human and murine model of rheumatoid arthritis. Thus, different strategies have been developed to block IL-15’s action reducing the severity of the disease (18). To further understand how IL-15 functions in this pathological condition, and which mode of action needs to be targeted, the efficacy of NANTIL-15 to attenuate inflammation was evaluated in the collagen-induced arthritis (CIA) model, in which IL-15 has been shown to participate to the inflammatory process (39). We performed two experimental models of CIA. In the first, mice were exposed to collagen twice 2 weeks apart. Treatment with NANTIL-15 began the day after the second immunization (day 22), when mice had already developed signs of inflammation for about 10 days. By measuring the clinical scores daily, we observed that mice treated with NANTIL-15 did not develop severe arthritis compared to those treated with Ig2a isotype control ( Figure 5A ). Histological examination of joints of NANTIL-15-treated animals revealed reduction in articular inflammation compared to controls. Immunohistology also showed that infiltration by CD8 T cells was reduced in mice treated with NANTIL-15 compared to controls ( Figure 5B ). In the second experimental design, NANTIL-15 was administrated every day for 2 weeks when the first inflammatory symptoms appeared. By measuring daily the clinical scores and paw thickness, we observed that NANTIL-15 treatment significantly reduced the inflammatory signs, compared to controls (mice treated with Ig2a isotype control) ( Figure 5C ). Interestingly, 5 days after the last injection of NANTIL-15, no increase in clinical signs was observed ( Figure 5C , left panel).

Splenic NK and CD8 T-cell populations were analyzed to evaluate the impact of NANTIL-15 treatment in the periphery of arthritic mice. No significant difference was observed between NANTIL-15-treated and control mice in terms of total NK cells, CD3, CD8, and CD44⁺CD122⁺CD8 T cells in the spleen ( Figures 6A, B ). Taken together, these results show that NANTIL-15 treatment reduces inflammation locally without any impact on NK and CD8 T cell homeostasis ( Figure 7 ), indicating that inhibition of IL-15’s action through the trimeric IL-15Rα/IL-2Rβ/γc receptor is sufficient to reduce inflammation while leaving IL-15’s target cell homeostasis unaffected.