IL15 by Continuous Intravenous Infusion to Adult Patients with Solid Tumors in a Phase I Trial Induced Dramatic NK-Cell Subset Expansion

The goal of immunotherapy is to direct the immune system to attack patients' cancers. Initial attempts in clinical trials to enhance latent immune responses focused on stimulatory cytokines such as IL2 or IFNα (1–6). Results from multiple clinical trials led to FDA approval of high-dose IL2 (HDIL2) for treatment of patients with metastatic renal cell carcinoma and metastatic melanoma. The severity of systemic toxicities caused by intensive cytokine regimens, especially with HDIL2, was a key factor prompting the search for other immunotherapeutic cytokines with the benefits of IL2 but with fewer negative adverse events (AEs).

IL2 and IL15 both stimulate proliferation of T cells, induce generation of cytotoxic lymphocytes and memory phenotype CD8 T cells, and stimulate prolonged expansion of natural killer (NK) cells (7–30). In contrast to IL2, IL15 did not mediate activation-induced cell death (AICD), less consistently activated Tregs, and caused less capillary leak syndrome in mice and nonhuman primates (8). Furthermore, preclinical studies of IL15 showed expansion and activation of NK cells and CD8 memory T cells with superior antitumor efficacy in mice compared with IL2, IL7, and IL21 (7).

We performed a first-in-human trial of Escherichia coli–produced recombinant human IL15 (rhIL15) where treatment was given as daily 30-minute intravenous bolus (IVB) infusions. Unexpectedly, postinfusion toxicities limited dose escalation and clinical evaluation (31). There was a consistent temporal pattern of posttreatment AEs with fevers, rigors, and hypotension that overlapped with maximum serum concentrations of IL6 and IFNγ and led to the conclusion that rhIL15 was too difficult to administer by this schedule. The exceedingly high initial IL15 C_max levels were sufficient to contribute to toxicity observed, as well as to allow induction of NK and CD8 T cells—the biological effects sought. To reduce C_max, we evaluated alternative dosing strategies including subcutaneous and continuous intravenous infusion (CIV) administrations (32, 33). We report the results of our phase I dose-escalation trial with CIV rhIL15 in patients with advanced metastatic refractory solid tumors.

Recombinant human IL15 (rhIL15) was produced in Escherichia coli (E. coli; refs. 31, 34).

Patients with advanced metastatic solid tumors for which standard curative or palliative treatments did not exist were enrolled in this phase I, open-label, nonrandomized phase I dose-escalation trial to determine the safety and toxicity of IL15 in patients with metastatic malignancy. Patients with CIV for 10 consecutive days received IL15 at a starting dose of 0.1 μg/kg/day. Dose escalation proceeded in a 3 + 3 standard escalation to dose levels of 0.25, 0.5, 1, 2, and 4 μg/kg/day. Patients receiving the 10-day treatment schedule without evidence of ongoing response after any two consecutive cycles of treatment discontinued rhIL15. Patients manifesting an ongoing response defined as >15% decrease in sum of marker lesions and/or improvement or disappearance of some nonmeasurable lesions and/or >10% decrease in tumor markers received additional cycles. Cycles 1 and 2 were 42 days in length but all subsequent cycles were 28 days in length. Toxicities of only the first cycle were considered in selecting the MTD/RP2D. It is possible that cumulative effects may be important in future selection of the dose recommended. The study was approved by the Institutional Review Board of the NCI, NIH. The study was performed in accordance with the ethical guidelines of the Declaration of Helsinki ethical principles of medical research. All patients signed a written informed consent for participation in the clinical trial. rhIL15 was produced under current good manufacturing practice conditions in the Escherichia coli expression system, as described previously (1).

Supplementary Table S1 summarizes patient demographics and treatment history.

The rhIL15 was delivered intravenously by infusion or ambulatory pump for 10 consecutive days (240 hours). The rhIL15 was diluted to a concentration of 1 μg/mL with 0.1% human serum albumin to improve stability and was administered at doses of 0.125, 0.25, 0.5, 1, 2, or 4 μg/kg/day.

Pharmacokinetic (PK) analysis of serum IL15 levels was performed during cycle 1 on serum samples obtained; immediately prior to the first dose, at 10 minutes after the first dose, at 1, 2, 4, 8, 12, approximately 24, and approximately 48 hours after the first dose, once daily on days 7, 8, 9, and 10, at the completion of treatment, at 10 and 30 minutes after the completion of treatment, and at 1, 2, 4, and approximately 24 hours after completion of treatment. Patients were assessed for development of anti-IL15 antibodies using a technique described previously (31).

Patients had regular measurement of vital signs, oral intake and output, physical examination, daily chemistry and hematology laboratories. Patients had restaging CT scans and/or other appropriate radiographic procedures to evaluate their disease after cycle 1, cycle 2, and every two cycles thereafter with antitumor response assessed by RECIST 1.1 criteria (35).

From July 2012 through January 2017, 27 patients with a variety of refractory cancers were treated with CIV rhIL15. As shown in Supplementary Table S1, the most common diagnoses were head and neck cancer (n = 4), sarcoma (n = 4), pancreatic cancer (n = 3), and renal cell carcinoma (n = 3). The median age of patients was 60 years, eight patients were 65 years or older. The median number of prior therapeutic interventions was 4 (range, 2 to 10).

Patients treated with the CIV rhIL15 regimen experienced typical recombinant cytokine side effects. Patients treated at 1 μg/kg/day or higher levels frequently demonstrated a mild erythematous, sometimes pruritic rash. Interestingly, two patients treated at 2 μg/kg/day developed a malar “butterfly rash” (Supplementary Fig. S1) classically associated with systemic lupus erythematosus, although neither patient had other findings consistent with this disease, nor developed any autoantibodies.

There was a total of eight serious adverse events (SAEs) possibly, probably, or definitely drug related (Table 1). There were two significant bleeding events (duodenal hemorrhage in a 1 μg/kg–treated patient from erosions caused by a duodenal stent, and bronchial hemorrhage in a 0.25 μg/kg–treated patient from known intrabronchial tumor mass that became necrotic) both definitely related to disease. One patient treated at 0.5 μg/kg had delayed recurrence of her grade 2 gastritis, with biopsy showing a brisk infiltrate of CD4⁺ lymphocytes. A patient with melanoma treated at 1 μg/kg developed isolated grade 3 papilledema without evidence of elevated intracranial pressure with MRI scans used to determine new enhancing lesions. A patient with melanoma treated at 1 μg/kg dose level developed repeated episodes of blurred vision in the week after discontinuing treatment for disease progression. Because of the concern for new central nervous system metastases the patient had an MRI, which demonstrated no intracranial lesions but did show bilateral focally enhancing retinal abnormalities felt to represent isolated papilledema. These findings were assessed as uveitis (grade 3) related to his rhIL15 treatment and possibly related to his prior treatment with high-dose IL2 or ipilimumab or nivolumab, as was recently described (36). The patient's symptoms and uveitis resolved after treatment with high-dose corticosteroids. A patient with squamous cell head and neck cancer developed worsening laryngeal inflammation from rapid disease progression that required a tracheostomy. A patient with multifocal renal cell carcinoma and pulmonary metastases treated at 1 μg/kg dose level developed grade 3 hypoxia and presumed pneumonitis that required discontinuation of treatment. This patient's pulmonary status improved with initiation of high-dose corticosteroids and diuresis but still required supplementary oxygen. The patient died 1 week after stopping treatment after returning home; autopsy was not performed but death was presumed to be due to disease progression. Both patients treated at the 4 μg/kg dose level had SAEs requiring early discontinuation of treatment. The first patient had grade 2 diarrhea initially linked to a change in enteral feeding formulation and later antibiotic-related diarrhea from empiric amoxicillin–sulbactam treatment. The CIV rhIL15 infusion was interrupted for 24 hours until her diarrhea improved after initiation of antimotility treatment, but soon after rhIL15 infusion was restarted the patient developed abdominal pain, decreased bowel sounds, and tenderness upon palpation, followed soon after by progressively increasing rigidity and rebound pain. The patient was taken to surgery and was seen to have diffuse patchy ischemia that involved her entire enteral tract (stomach, large, and small intestines) without evidence of tumor in the peritoneum or serosa, arterial thrombi, or perforation. The patient was given appropriate supportive measures (intravenous fluids, vasopressors, and broad-spectrum antibiotic treatment) with the hope that reperfusion would occur after treatment was discontinued but the patient eventually expired from multiorgan failure a few days later. The second 4 μg/kg patient had appreciable involvement of the liver with metastatic GIST and developed hepatic encephalopathy. Treatment was discontinued on day 5 and the patient's AEs resolved within a few days of supportive care.

The most common clinical toxicities experienced by patients treated on the basis of treatment level and grade of the adverse event are listed, as well as the serious adverse events, and the most common abnormal laboratory measurements in the patients treated in the protocol. N: the number of patients treated in the dosing cohort. CTC, V4: NCI clinical toxicity criteria version 4 (Table 1). In addition to the lymphocytopenia, an expected effect of IL15 administration, anemia was the most common laboratory abnormality (24 of 27 patients), generally grade 1, except at higher rhIL15 doses. Leukopenia, neutropenia, and thrombocytopenia were common at higher doses but were not associated with infectious complications or bleeding. Transient elevations of liver functions tests (ALT, AST, and alkaline phosphatase) generally beginning the second or third day of treatment, occurred in nearly all patients treated at or above 0.5 μg/kg dose level and resolved spontaneously to baseline by the end of treatment or the week following completion of treatment (Supplementary Fig. S2A and S2B). Aside from the 4 μg/kg patient who developed the DLT of hepatic encephalopathy, no other patients had clinical AEs related to this finding.

Patients' response to therapy was assessed by RECIST 1.1 criteria (35). As seen in Fig. 1, at least 6 patients had some evidence of tumor regression. Seventeen of 27 patients had stable disease as the best response, for a disease control rate (DCR) of 63%. Two patients treated at 1 μg/kg/day dose fulfilled the criteria defined in Supplementary Materials required to receive additional cycles of treatment. In particular, patient No.15 had a 16% decrease in his marker lesions, and complete disappearance of a cutaneous tumor nodule that was shown to be necrotic and having lymphocytic infiltration when biopsied. Patient No.16 had a 9% decrease in marker lesions, and her CA-19-9 decreased from a baseline of 126 to 47 U/mL. They had pancreatic cancer, and in addition to radiographic findings, had substantial decrease in CA 19-9 from 125.7 U/mL baseline to 47.3 U/mL postcycle 1 and 56.9 U/mL post cycle 2. Patient No. 15 with chondrosarcoma involving the right chest wall developed ipsilateral pleural effusions during each of the six cycles, had a decrease in sum of marker lesions measurements, and complete regression of the rarely observed (37–39) cutaneous metastatic lesion during treatment. Biopsy of this lesion showed an infiltration of CD3 (2+), CD4 (2+), CD8 (1+) and CD56 (±)-positive cells. While no patient had a partial response, almost half demonstrated stabilization of disease that had been progressing prior to initiation of treatment.

Previously, we demonstrated that IV bolus treatment with IL15 resulted in acute lymphopenia followed by hyperproliferation of IL15-responsive cells with an increase in NK and CD8⁺ T cells sustained for more than 3 weeks after first treatment. Posttreatment, cells were hypoproliferative until homeostatic baseline proliferative rates were restored (31). Here, we show that administration of rhIL15 by continuous intravenous infusion had an even greater impact on CD8⁺ T and NK—particularly CD56^bright NK-cell expansions. A significant threshold dose effect was seen between patients receiving low (0.125, 0.25, 0.5 μg/kg/day) and high (1.0, 2.0 μg/kg/day) doses of rhIL15 (Fig. 2A). Changes were maximal 2 days following treatment cessation but were still dramatic 10 days later (Fig. 2B). As expected, all patients experienced generalized acute lymphopenia at initiation of treatment (Supplementary Fig. S3). Following this, blood lymphocytes returned with evidence of high levels of proliferation, lasting through the end of treatment (Fig. 3). At the cessation of treatment, high-dose patients had a dramatic 358-fold increase of CD56^bright NK cells, a 38-fold increase of total NK cells, and a 5.8-fold increase of CD8⁺ T cells (Fig. 3). γδ and CD4⁺ T cells also responded to rhIL15 treatment, although to a lesser degree than NK cells (Fig. 3; Supplementary Fig. S3). There was also evidence of CD8⁺ T-cell activation, evidenced by increased frequency of CD38⁺ and HLA-DR⁺ cells without a corresponding increase in CD25⁺ cells including Tregs (Supplementary Fig. S4).

Twenty-seven patients had PK samples collected, but only 22/27 had sufficient data above the lower level of quantitation (LLOQ) for analysis (Fig. 4). IL15 was below LLOQ for patients receiving the 0.125 and 0.25 μg/kg/day dose levels during the first 48 hours, and modest levels of IL15 (range, 30 to 40 pg/mL) were detected later (Fig. 4). IL15 levels were relatively constant throughout the 10-day CIV infusion for patients in the 0.5 and 1 μg/kg/day dose groups. Serum rhIL15 levels for the patients treated at the 2 μg/kg expansion dose were markedly lower during days 7 through 11 of the treatment cycle compared with earlier time points. There was a modest increase in the expression of IL15Rα per cell. In addition, there was a marked increase in the number of cells expressing high levels of IL2/IL15Rβ (CD122), which with gamma-c (CD132) binds IL15 to yield 10⁻⁹ mol/L. The expanded expression of such IL15-binding NK and CD8 T cells evident in Fig. 3 may act as a sink for the administered IL15 thereby providing a partial explanation for the decline of levels of serum IL15 during the last 3 days of the infusion.

Patients were assessed for the development of anti-IL15 antibodies three times during each treatment cycle. None of the samples obtained for the 27 patients treated in this trial showed evidence of antibodies against IL15.

Increases in serum IL1β levels were first noted within 60 minutes after the start of rhIL15 infusions (Supplementary Fig. S5A). Dose-dependent increases in IL6 were seen by 8 hours after treatment initiation (Supplementary Fig. S5B). Serum TNFα levels increased gradually over the first 48 hours and remained elevated throughout the course of treatment (Supplementary Fig. S5C). Serum IFNγ was not measurable for most dose levels until 24 or 48 hours of treatment (Supplementary Fig. S5D) and generally showed a late-cycle dose-dependent increase paralleling the TNFα trend. The dose-proportional increase in inflammatory cytokines seen in the final days of the infusion coincided with the increased number of activated, cytokine-producing CD8⁺ T and NK cells. Prior to infusion, serum PD-1 level (normal range nondetectable to 50 pg/mL) was below the lower limit of detection, 20 pg/mL in eight of nine patients. With rhIL15 infusions at 2 μg/kg/day, the arithmetic mean concentrations rose to 23.7 pg/mL at 24 hours, 27.8 pg/mL at 48 hours, 43.9 pg/mL at day 8, and 45.0 pg/mL at day 12 following initiation of infusions. This is in accord with our previous observations that IL15 induces PD-1 on CD8 T cells. The serum PD-L1 concentration range 44.5 to 106 pg/mL remained below the lower limit of detection, 200 pg/mL. The arithmetic mean of the serum PD-L2 concentration preinfusion was 7,453 pg/mL (normal range, 1,425–10,678 pg/mL) and only rose to a maximum of 8,419 pg/mL on day 8 of the 2 μg/day infusions.

Analysis of effector cell infiltration and checkpoint inhibitor expression was performed on pre-and posttreatment tumor biopsies from two patients treated at higher dose levels (Fig. 5). The patient shown in Fig. 5B and upper portion of the Table had a slight increase of CD8⁺ T cells in the periphery and a noticeable increase in CD56^+bright NK cells both peripherally and centrally in the colorectal tumor deposit. A patient with squamous cell carcinoma of the head and neck shown in Fig. 5A and in lower portion of the Table had a modestly increased infiltrate of CD3⁺ cells and an appreciably increased number of CD56^+bright NK cells in the central portion of the tumor.

Preclinical data demonstrated IL15′s considerable potential as an immunotherapeutic from the cytokine's ability to generate prolonged proliferation and activation of effector lymphocytes without AICD (7–31, 40), and low potential to activate Tregs compared with IL2. In addition to defining safety profiles and MTD for new agents, phase I trials should also establish an optimal schedule and route of administration. As predicted by the preclinical NHP toxicology studies, the CIV regimen sustained serum rhIL15 in the 1,000 to 5,000 pg/mL range and generated greater expansion of NK and CD8 T effector cells than the 3 μg/kg dose in the IV bolus regimen (C_max 43,900 pg/mL and t_1/2 of 2.37 hours) or the subcutaneous regimen (C_max 1,638 pg/mL and t_1/2 = 10 hours; refs. 31, 41, 42; Supplementary Table S2). We suspect that rhIL15 infusion causes lymphocytes and especially NK cells to marginate or efflux from the circulation, where they undergo proliferation throughout the treatment course. At the termination of treatment, there is a redistribution of lymphocytes back into the circulation, resulting in dramatic increases in the numbers of circulating cells (Supplementary Fig. S6).

Continuous intravenous infusion of rhIL15 resulted in the greatest increase in circulating effector cells compared with administration by IVB or subcutaneous regimens or with IL15 superagonist complex ALT-803 (Supplementary Table S2; refs. 31, 41–43). The CIV regimen at 2 μg/kg/day resulted in a 38-fold increase in the number of circulating NK cells and 358-fold increase in the number of CD56^bright NK cells. The IVB regimen at the MTD of 0.3 μg/kg/day resulted in only an approximately two- to threefold increase of NK cells. Even the highest dose tested by IVB (3.0 μg/kg/day) resulted in only a 10-fold increase in NK cells while causing dose-limiting toxicities (31). IL15 administered subcutaneously at the expansion dose of 2 μg/kg/day was well tolerated but only produced a 10.8-fold increase in NK cells and 29.7-fold increase in CD56^bright NK cells (42). ALT-803 (IL15/IL15Rα-IgFc) treatment elicited cytokine release–related AEs with intravenous treatment and skin toxicity with subcutaneous dosing limiting the dose that could be used (43). However, when administered by IVB or subcutaneously, ALT-803 at five to 10 times the dose (10–20 μg/kg/day) containing 2.6 μg/kg/day of IL15 induced a smaller expansion (eight times) of NK cells than the 38-fold increase with CIV rhIL15 at 2 μg/kg/day in this study. Because IL15 is only 26 % of the ALT-803 complex, 10 μg of ALT-803 contain 2.6 μg of IL15.

At the termination of the 2 μg/kg/day 10-day CIV infusions the NK cells including the CD56^bright NK cells were not exhausted but were vigorously proliferating. Ki-67⁺ was present in over 95% of CD56^bright NK cells on both days 8 and 12 with the 10-day CIV. Furthermore, as reported previously (41), evaluated on day 12 of the 10-day CIV infusions IL15 augmented the cytotoxic activity of all NK-cell subsets against three target cell lines, the killing of which involved different receptors: CD20 rituximab antibody-coated RAJI, K562, and C1R-MICA. These cell lines were recognized by the functionally active NK cells, with CD16 mediating antibody-dependent cell-mediated cytotoxicity (ADCC), NKp30 and NKp46 mediating natural cytotoxicity for K562, and NKG2D mediating it for CIR-MICA. In our previous publication (41), we discussed functional and phenotypic analysis of the NK cells prior to infusion and on day 12 of a 10-day infusion, that is, 2 days following termination of the infusion during the influx of NK cells at their maximal number. CD56^dim NK cells treated with IL15 acquired features normally observed with CD56^bright NK cells, that is, increased expression of NKG2D, Trail, and IL15R alpha. Furthermore, after IL15 infusions CD56^bright NK cells expressed increased amounts of CD16, NKp30, and NKp46. In addition, IL15 infusions increased perforin in CD56^bright NK cells. These data show that 10-day continuous IL15 infusions enable all subsets of NK cells to function with increased cytotoxicity at the termination of the cultures. Thus, among available agents and dosing regimens, rhIL15 administered CIV provided the greatest augmentation of circulating functional NK cells.

Even in phase I safety trials of new agents, there is some expectation of clinical activity. Although no patients achieved even a partial response, the data in Fig. 1 demonstrate a flattening or negative deflection in tumor burden over time for a number of patients with a 63% DCR in a heavily pretreated population. One patient had a cutaneous chondrosarcoma nodule—a rare metastasis (37–39) that disappeared during the course of treatment. Complete regression of this nodule in the patient with chondrosarcoma, minor radiographic improvements, and the threefold decrease in CA 19-9 and disease stabilization for nearly 6 months in one of the patients with pancreatic cancer also implies antitumor activity. The extensive prior treatment history of most patients and, the small number of classically immunosensitive tumor patients only two of which were treated at the more biologically active doses of rhIL15 must be considered when weighing lack of sustained significant responses.

This modest impact on tumors with IL15 monotherapy led to our conclusion that to have a major clinical contribution to cancer therapy, IL15 will have to be used in combination therapy. In particular, IL15 could be used in combination with agents such as mAbs that take advantage of the increase in NK cell numbers and that impart tumor specificity (44). In modern immunotherapy, there has been a great contrast between the relative lack of efficacy of NK cells infused as monotherapy for cancer and its critical link in the chain of effectors in combination therapy involving NK cells with those anticancer mAbs (e.g., rituximab) that function predominantly by ADCC (45–50). Nevertheless, in the clinical trials cytokine (IL12, IL15, and IL18)-induced memory-like NK cells exhibited enhanced responses against myeloid leukemia. Clinical responses were observed in five of nine evaluable patients with AML including four complete remissions (46). Furthermore, there has been efficacy in bone marrow transplant studies in this leukemia (46, 48–50). In addition, there has been a positive association in some studies of NK infiltrate not T-cell infiltrate in renal cancer (45, 47). Renal cell carcinoma infiltrating NK cells, but not T cells, have been shown to have the inherent ability to recognize transformed cells but require cytokine activation and are sensitive to inhibition by inhibitory receptor ligands (45, 47).

The dramatic increases in NK cells alone in this study were not sufficient to produce antitumor activity, likely because most tumors express self-MHC class I molecules that inhibit NK-effector functions. However, it should be noted that when NK cells in combination therapy are considered they are a pivotal link in the chain of effectors required for the efficacy of certain (e.g., rituximab) antitumor mAbs that function by ADCC alone. We investigated combination therapy of IL15 with rituximab in a syngeneic mouse model of lymphoma transfected with human CD20 and with alemtuzumab (CAMPATH-1H) and in a xenograft model of human adult T-cell leukemia (ATL; ref. 51). IL15 greatly enhanced the therapeutic efficacy of both rituximab and alemtuzumab in these tumor models. The additivity/synergy was shown to be associated with augmented ADCC. Both NK cells and macrophages were critical elements in the chain of interacting effectors involved in optimal therapeutic responses mediated by rituximab and IL15. We provided evidence supporting the hypothesis that NK cells interact with macrophages to augment the NK-cell activation and expression of Fc gamma receptors and the capacity of these cells to become effectors of ADCC. Therefore, a very attractive antitumor strategy is to use IL15 in conjunction with antitumor mAbs to augment NK ADCC and, therefore antitumor efficacy. Recent publications describe results from two clinical trials that demonstrated appreciable clinical activity from the combination of ALT-803 and rituximab or nivolumab to patients with relapsed refractory non-Hodgkin lymphoma or patients with relapsed refractory non–small cell lung cancer, respectively (52, 53).

These observations and the results of our preclinical animal model testing of the combination of IL15 with mAbs support combination clinical trials involving IL15 (54, 55). Preclinical studies of IL15 with anti–CTLA-4 and anti–PD-L1 supported our initiation of a clinical trial with this triple combination (NCT03388632) involving IL15 (54, 55). Furthermore, they provided the scientific basis for our initiation of a phase I trial of IL15 combined with alemtuzumab (anti-CD52) that has been opened for patients with refractory and relapsed adult T-cell leukemia (NCT02689453). Additional trials are being initiated in patients with refractory and relapsed chronic lymphocytic leukemia where rhIL15 will be administered in combination with obinutuzumab (anti-CD20, NCT03759184), and in patients with renal cancer where IL15 and avelumab (anti–PD-L1) will be coadministered.

No potential conflicts of interest were disclosed.

Conception and design: K.C. Conlon, T.A. Waldmann

Development of methodology: K.C. Conlon, L.P. Perera, W.D. Figg, J.L. Yovandich, T.A. Waldmann

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K.C. Conlon, E.L. Potter, C.-C.R. Lee, M.D. Miljkovic, T.A. Fleisher, M. Petrus, J. Hsu, W.D. Figg, J.L. Yovandich, M. Roederer, T.A. Waldmann

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K.C. Conlon, E.L. Potter, S. Pittaluga, C.-C.R. Lee, M.D. Miljkovic, S. Dubois, B.R. Bryant, M. Petrus, L.P. Perera, W.D. Figg, C.J. Peer, J.H. Shih, M. Roederer, T.A. Waldmann

Writing, review, and/or revision of the manuscript: K.C. Conlon, E.L. Potter, S. Pittaluga, M.D. Miljkovic, T.A. Fleisher, M. Petrus, W.D. Figg, J.H. Shih, M. Roederer, T.A. Waldmann

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E.L. Potter, B.R. Bryant, M. Petrus, J. Hsu, S.P. Creekmore, M. Roederer, T.A. Waldmann

Study supervision: K.C. Conlon, J. Hsu, T.A. Waldmann

Others (acquired photomicrographic images of histology slides): C.-C.R. Lee

Others (technical support role in development, manufacturing, and testing of IL15): S.P. Creekmore

This study was supported by the Intramural Research Programs of the Center for Cancer Research, NCI and the Vaccine Research Center, National Institute of Allergy and Infectious Diseases, NIH (Bethesda, MD). Leidos Biomedical Research, Inc. (Reston, VA), and the Biopharmaceutical Development Program, NCI Frederick (Frederick, MD) provided the recombinant human IL15 for this study. The clinical trial registration number is NCT 01572 493.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.