Systemic high-dose interleukin-2 (IL-2) achieved long-term survival in a subset of patients with advanced melanoma. The authors reported previously that intratumorally applied IL-2 induced complete local responses of all metastases in >60% of patients. The objectives of the current study were to confirm those results in a larger cohort and to identify patient or regimen characteristics associated with response.
Patients with melanoma who had a median of 12 injectable metastases received intratumoral IL-2 treatments 3 times weekly until they achieved clinical remission. The initial dose of 3 million international units was escalated, depending on the individual patient's tolerance.
Forty-eight of 51 patients were evaluable. Only grade 1/2 toxicity was recorded. A complete response that lasted ≥6 months was documented in 70% of all injected metastases. A complete local response of all treated metastases was achieved in 33 patients (69%), including 11 patients who had between 20 and 100 metastases. Response rates were higher for patients who had stage III disease compared with patients who had stage IV disease. No objective responses of distant untreated metastases were observed. The 2-year survival rate was 77% for patients with stage IIIB/IIIC disease and 53% for patients with stage IV disease. Efficacy and survival did not differ between patients who had ≥20 lesions and patients who had <20 lesions.
Metastatic melanoma has an unfavorable prognosis, and treatment options are limited.1, 2 In patients with stage III melanoma, the objective of treating locoregional metastases is the complete removal of any detectable tumor manifestation. In patients who have many in-transit metastases, surgery often is no longer feasible. In a subset of patients who have disease limited to 1 extremity, isolated limb perfusion can be applied but is associated with considerable toxicity.3 In patients with stage IV melanoma and for patients who have unresectable metastases, palliative systemic treatments generally are preferred. However, surgical treatment is the first choice in stage IV if the patient can be rendered free of disease at all known metastatic sites; and, in 10% to 20% of patients with stage IV disease, this therapeutic option seems to be associated with a clear survival benefit.4 Therefore, alternative direct treatment options are needed if disease is not completely resectable.
Interleukin-2 (IL-2) has been used in immunotherapy against cancer since the early 1980s, applied either as a single agent5 or in combination with adoptively transferred, lymphokine-activated killer cells.6 Another approach is the adoptive transfer of in vitro expanded, autologous tumor infiltrating lymphocytes7 followed by high-dose IL-2.8 A retrospective analysis of 8 trials that used systemic high-dose IL-2 in patients with melanoma demonstrated response rates of 16%, including 6% complete responses. Greater than 50% of the complete responders in those trials remained progression-free after 5 years.9 These data led to US Food and Drug Administration approval of systemic high-dose IL-2 for the treatment of advanced melanoma in 1998.
Intratumoral application of drugs is an appealing therapeutic concept, because high concentrations can be achieved within the tumor, which may be essential to attain the desired therapeutic effect. Furthermore, systemic concentrations of locally applied drugs are low in contrast to systemic treatments with the same agent, resulting in comparably low toxicity. Several cytokines have been used for intratumoral therapy with various results.10-12 The first report that described the regression of melanoma metastases after intratumoral IL-2 treatment was published in 1994.13 On the basis of those findings, we initiated a pilot study on 24 patients and observed complete responses of all treated metastases in >60%. In contrast to systemic, high-dose therapy, intratumoral application was associated with low toxicity.14 The objectives of the current study were to confirm our previous results in a larger cohort and to identify relevant patient or regimen characteristics associated with response to treatment and with overall survival.
MATERIALS AND METHODS
The study protocol (National Clinical Trials [NCT] clinicaltrials.gov identifier NCT00204581) was approved by the local ethics committees, and patients were treated only after written informed consent was obtained. Inclusion criteria comprised age >18 years, histopathologically proven malignant melanoma, the presence of injectable dermal or subcutaneous metastases either in clinical stage III or clinical stage IV, and an expected survival >3 months. Excluded were pregnant or lactating women and patients with severe cardiac disease (New York Heart Association III and IV), alanine and aspartate aminotransferase levels >3 times above the upper limit of normal, creatinine >1.5 times above the upper limit of normal, and concomitant systemic therapy with steroids. Patients who required systemic chemotherapeutic treatment for metastatic disease and patients who had locoregional lymph node metastases were excluded. Previous systemic chemotherapy was allowed.
Study Design and Treatment
The trial had the design of an open, prospective phase 2 study. Patients were recruited at 2 study sites (Tuebingen and Homburg, Germany). For preparation of IL-2, 18 million international units (MIU) of recombinant human IL-2 (Proleukin; Novartis, New York, NY) were dissolved in 6 mL glucose 5% supplemented with 0.2% human serum albumin. IL-2 was injected intratumorally using 30-gauge needles for superficial injections and 27-gauge needles for deep injections, and single doses ranged between 0.3 MIU and 6 MIU, depending on the lesion size. One injection per lesion was applied for doses per lesion up to 3 MIU, and 2 injections were applied for doses >3 MIU per lesion. Treatment was initiated at 3 MIU IL-2 daily, and the dose was escalated by 1.5 MIU each treatment day up to the desired total dose according to the number of lesions and treatment guidelines (Table 1), provided that tolerability was maintained. The total daily dose was divided between all injectable lesions. For deep soft tissue metastases, sonography was used to guide injections. Up to 25 lesions were treated simultaneously; however, if more lesions were present, then they were treated alternately or subsequently. The treatment schedule was 3 times weekly on an outpatient basis. Treatment was terminated when clinical regression and/or necrosis of metastases was evident or if progression occurred that no longer was manageable by ongoing IL-2 injections. Adverse events were graded according to the Common Toxicity Criteria (version 3).
Table 1. Treatment Guidelines
Minimal Single Dose of IL-2 per Lesion, MIU
Volume Stock Solution, mL
Duration of Treatment, wk
IL-2 indicates interleukin-2; MIU, million international units.
Size of Individual Lesion, mm
Every treated metastasis was evaluated separately regarding clinical response between 4 weeks and 8 weeks after stopping IL-2 treatment and every 3 months thereafter. The following definitions were used: a complete response (CR) of a treated lesion was defined as the disappearance of any evidence of vital tumor and lack of tumor growth after stopping injections over a period of at least 6 months. A partial response (PR) was defined as a decrease ≥30% in the greatest dimension (longest diameter [LD]). Stable disease (SD) was defined as neither sufficient shrinkage to qualify for a PR nor sufficient increase to qualify as progressive disease (PD). PD was defined as an increase ≥20% increase in the LD of the lesion. Subcutaneous metastases were evaluated by sonography. In nondistinctive cases (eg, if there was residual pigmentation), biopsies were taken for histopathologic confirmation of response.
Statistical analyses were conducted using the SPSS 15.0 software package (SPSS Inc., Chicago, Ill). Follow-up was defined as the time from the first IL-2 injection to the date of last contact or death. For survival analyses, only deaths that were caused by melanoma were considered. The last survival update was done for all patients in April 2009. Kaplan-Meier analyses were performed to estimate overall survival. Differences in response rates were calculated by using a 2-tailed Fisher exact test.
Analysis of Specific Cellular Immune Responses
For 1 patient, the frequency of antimelanoma T cells was measured by flow cytometric analysis at different time points. Briefly, peripheral blood mononuclear cells (PBMCs) were incubated for 11 days in X-Vivo 15 (Lonza Verviers, Braine-l'Alleud, Belgium) with mixtures of overlapping peptides (PepMix; JPT Technologies, Berlin, Germany) that spanned the whole sequence of Melan-A, cancer-testis antigen (NY-ESO1), melanoma-associated antigen 3 (MAGE-A3), Survivin, and membrane protein M1/nucleocapsid protein from influenza virus, each at a concentration of 1 μg/mL. On Day 4, IL-2 was added (40 U/mL; Chiron Behring GmbH, Marburg, Germany). On Day 11, autologous PBMCs as stimulator cells were labeled with 5 μM carboxy fluorescein succinyl ester (CFSE) (Invitrogen, Karlsruhe, Germany) and incubated for 6 hours with each peptide mixture mentioned above. Cultured T cells were harvested and incubated at a 1:2 ratio with unpulsed or peptide-pulsed stimulator cells in the presence of 1 μL/mL Golgi-Plug (BD Biosciences, Mississauga, Ontario, Canada) for 12 hours. Thereafter, the cells were incubated with ethidium monoazide (EMA) (Invitrogen) followed by fixation and permeabilization with CytoFix/CytoPerm (BD Biosciences) and staining with cluster of differentiation (CD) 3 antibody (anti-CD3)-Pacific Orange (Caltag; Invitrogen), anti-CD4-peridinin chlorophyll protein complex, and anti-CD8-allophycocyanin-indocyanine 7 (BD Biosciences). Intracellular IFN-γ staining was done with a phycoerythrin/indocyanine 7-conjugated antibody (BD Biosciences) according to the manufacturer's instructions and was measured immediately using a BD LSR II flow cytometer (BD Biosciences). Flow cytometric data analysis was performed using FlowJo software (Tree Star, San Carlos, Calif) after the exclusion of CFSE-positive stimulator cells and EMA-positive dead cells. IFN-γ secretion was considered antigen-specific only when the frequency of IFN-positive T cells that responded to peptide-pulsed PBMCs was at least 2 times greater than the frequency of IFN secretion in response to the negative control (unpulsed PBMCs).
Patients and Treatments
In total, 51 patients (44 in Tuebingen and 7 in Homburg) were enrolled between August 2003 and November 2007 and completed IL-2 treatment by December 2007. Two patients were excluded after they provided informed consent, because progressive visceral disease that was detected at baseline staging diagnostics required chemotherapy. For 1 patient, neither treatment nor follow-up data were available. The baseline data from the 48 remaining patients are summarized in Table 2. The median patient age was 69 years (range, 37-88 years). Thirty-three patients were treated in stage III, and 15 patients were treated in stage IV. Among the patients with stage IV disease, 5 patients had lymph node/visceral metastases that were not accessible to local IL-2 treatment. Among the patients with stage III disease and the remaining 10 patients with stage IV disease, all metastases were accessible for injections. Most patients had received previous extensive therapy. Twenty-eight patients developed locoregional recurrences after undergoing surgery for metastatic disease. The last surgery occurred ≤8 weeks before the start of IL-2 therapy for 14 of 28 patients, and 22 of 28 patients underwent ≥2 previous surgeries. In 5 patients, limb perfusion or radiotherapy also was performed. Previous systemic treatments with interferon alpha (IFN-α) (11 patients), chemotherapy (8 patients), or both modalities subsequently (3 patients) had been applied before IL-2 treatment was initiated. The median duration of IL-2 treatment was 6 weeks (range, 1-32 weeks; some patients received several subsequent treatments for newly developing metastases), and the applied median total dose was 68.5 MIU IL-2 (range, 13.5-548.1 MIU). The median number of metastases treated per patient was 12. The highest daily dose was 16 MIU IL-2.
Table 2. Patient Characteristics
No. of Patients
Visceral metastases present
Site of treated metastases
No. of treated metastases
Surgery in stage III/IV
Adjuvant interferon alpha
All 48 patients were included in the analysis of toxicity. The treatment generally was tolerated well, and only grade 1 and 2 toxicity was recorded. Intratumoral IL-2 therapy almost always caused an inflammatory injection site reaction (local swelling and erythema) followed by a selective necrosis of the tumor tissue that generally did not affect the surrounding normal tissue. Injection pain also was frequent but was manageable by the application of a local anesthetic cream and oral metamizole. The majority of patients experienced fever (58%) that could be controlled easily by acetaminophen. Fatigue (36%) and nausea (34%) usually were mild and of short duration. Frequent adverse events are presented in Figure 1. Adverse events that were observed in <10% of patients but that were at least possibly related to the treatment included stomach pain, myalgia, and headache (in 4 patients each); itching exanthema (in 3 patients); dry oral mucosa (in 2 patients); pruritus (in 2 patients); hair loss (in 1 patient); and diarrhea (in 1 patient). One patient presented with generalized urticaria, which was abated by antihistaminic treatment and did not recur after further IL-2 treatments under continued prophylactic medication. One patient observed worsening of a pre-existing atopic dermatitis, and 1 patient reported a single episode of mild cardiac arrythmia. One patient presented with vitiligo-like depigmentation around the treated metastases (Fig. 2).
Clinical Responses to IL-2 Treatment
In total, 894 of 917 separately treated metastases (97.5%) could be evaluated for local tumor response (Table 3). The analysis revealed a 78.7% CR rate, a 0.7% PR rate, 16.3% stable metastases, and 4.3% progressive lesions. Differences regarding the rate of complete local responses of injected metastases were detected between stage III versus stage IV (96.9% vs 54.8%, respectively; P < .0001) and the absence or presence of visceral metastases (92.5% vs 16.5%, respectively; P < .0001). Furthermore, efficacy differed significantly between dermal versus subcutaneous injected lesions both for stage III disease (CR rate: 97.9% vs 90.3%, respectively; P = .0034) and stage IV disease (CR rate: 56.7% vs 34.4%, respectively; P = .0247). No objective responses of noninjected distant lesions were observed in 5 patients with stage IV disease who had visceral metastases. Figure 3 illustrates the clinical course of 3 patients who received intratumoral IL-2 treatment.
Table 3. Treatment Response
In 33 patients (69%) a complete local response of all treated metastases was achieved with a better outcome for patients who had stage III disease compared with patients who had stage IV disease (82% vs 40%; P = .0067). Of these, 32 patients were completely free of recognizable tumor after treatment. The percentage of patients with a complete local response of all treated metastases did not depend on the number of treated lesions (73% for patients who had ≥20 treated metastases vs 66% for patients who had <20 treated metastases; P = .7458).
The median follow-up was 25 months from the start of IL-2 treatment (range, 4-68 months; median, 14 months for patients who died of their disease and 32.5 months for all others). Overall survival rates were promising for all patients (Fig. 4A). We did not record any deaths from melanoma later than 25 months after starting treatment, indicating a considerable chance for long-term survival in patients who survived the first 2 years. Patients who had stage III disease, as expected, had higher overall survival rates compared with patients who had stage IV disease (77% vs 53% after 2 years). It is noteworthy that overall survival for the subgroup of patients who had stage IV disease without visceral metastases was comparable to that for patients who had stage III disease (Fig. 4A, dotted line). The long-term outcome did not depend on the number of treated metastases, because the overall survival for 15 patients who had ≥20 metastases was as good as that for the other 33 patients who had <20 metastases (Fig. 4B).
Specific Cellular Immune Responses
For 1 additional patient with stage III disease who was treated in 2008, peripheral blood samples were collected before IL-2 injections, 4 weeks after starting treatment, and 2 months after ending treatment. This woman had 19 dermal metastases, received 11 treatments (cumulative dose 55 MIU IL-2), and had an ongoing complete clinical and histopathologic response after therapy. There was a significant increase (3 times greater than background) in IFN-γ–positive, CD8-positive T cells that were specific for MAGE-A3, Melan-A, and NY-ESO-1 during treatment, and the increased frequency of MAGE-A3–specific, CD8-positive T cells remained present even 2 months after the end of treatment; whereas the frequency of influenza-specific, IFN-γ–positive T cells remained constant at all 3 time points (data not shown). Similarly, there was a marked increase (almost 2 times greater than background) in the frequency of IFN-γ–positive, CD4-positive T cells that were specific for MAGE-A3 during and after therapy, but the frequency of CD4-positive T cells that were specific for other antigens did not change significantly (Fig. 5).
Confirming the results from our pilot study in the current, larger trial, we again observed a complete local response of all treated metastases in 69% of patients (82% of patients with stage III disease and 40% of patients with stage IV disease). It is well established that systemic, high-dose IL-2 treatment is effective in a subset of patients with melanoma, and this subgroup appears to represent most patients who have soft tissue metastases. Phan et al reported a 53% response rate to systemic IL-2 treatment in a subgroup of 28 patients with subcutaneous and dermal metastases.15 Thus, the observed efficacy in our current trial most likely was a result of the high intratumoral IL-2 concentration, which also has been described in animal models after intratumoral treatment.16
In contrast to systemic IL-2 treatment accompanied by severe toxicity, intratumoral IL-2 treatment is well tolerated. The good tolerability of this treatment obviously depends on the small amounts applied, with little systemic leakage, and on the application of treatment only 3 times weekly. This is reflected in the complete lack of any premature terminations of treatment because of toxicity in our study.
In the current study, we also tried to identify the patient characteristics that were associated with outcome. We observed that the presence of visceral metastases that could not be injected also was associated with a low local response rate of the injected metastases (CR, 16.7% of metastases; PR, 1.2% of metastases). This is an interesting observation and raises the issue of differences in the immune responsiveness of patients who have visceral metastases. These patients probably should be regarded as ineligible for this treatment. Efficacy also was greater for dermal metastases than for subcutaneous metastases in both stage III and stage IV disease. The greater size of subcutaneous lesions may help explain this finding, because we generally observed greater treatment efficacy in smaller metastases. Conversely, injections into subcutaneous metastases usually were applied after ultrasound-guided marking of the localization on the skin surface. This occasionally may have led to peritumoral injections instead of intratumoral injections, which are known to be less effective.17 Another possible explanation may be that the dermis provides a better environment for inflammatory reactions than subcutaneous tissues.
The 2-year overall survival probability for patients who had stage IIIB/IIIC disease was 77% (stage IIIB, 94%; stage IIIC, 61%), and it was 53% for patients who had stage IV disease after starting local IL-2. It is noteworthy that the presence of a large number of metastases was not associated with a poorer long-term outcome than the presence of fewer metastases. In the current study, we treated 13 patients who had ≥20 metastases but no visceral lesions. Remarkably, 11 or those patients have remained free of any residual tumor after stopping IL-2 treatment despite the initially high number of lesions. Seven patients have remained free of any recurrence over the follow-up period, and only 3 patients have died from melanoma to date (data not shown). This subgroup of patients often is regarded as unresectable because of the high number of metastases. Therefore, systemic chemotherapy (with response rates <30% and without any proven benefit on overall survival) otherwise generally would be the only alternative treatment option for such patients.2
The exact mode of action of IL-2 is unknown. One hypothesis is that lymphokine-activated killer cells are induced by IL-2 and subsequently destroy the tumor by direct lysis.6 Indeed, histopathologic evaluation of biopsies taken from IL-2–treated metastases revealed a dense intratumoral and peritumoral lymphocytic infiltrate. We observed that tumor cells were undergoing apoptosis and that the mononuclear infiltrate mainly consisted of CD8-positive, CD4-positive T cells, and (only to a minor extent) natural killer cells.14 The favorable long-term outcomes that we observed suggest a possible beneficial systemic mode of action induced by the local treatment. Systemic responses after locally applied intratumoral IL-2 have been described to date in various animal models. Maas et al treated lymphoma-bearing mice and reported the regression of distant uninjected lesions after intratumoral treatment. The same IL-2 doses given systemically were far less effective.18 Van Es et al studied a transplanted rabbit carcinoma model and demonstrated that peritumoral IL-2 induced complete regressions of untreated contralateral tumors. A second challenge of the cured animals resulted in tumor rejection, suggesting the generation of specific immunity.19 In accordance with these preclinical observations, we demonstrated vitiligo-like depigmentation in 1 patient as a sign for the induction of a specific immune response directed against differentiation antigens. In another patient, a marked increase in circulating antimelanoma T cells was observed, hinting at a possible systemic vaccination effect.20 The presented trial was planned as a local treatment option for unresectable patients that did not take into account possible systemic modes of action. We only performed specific immune monitoring in 1 patient. Therefore, the significance of these data are limited, and additional patients need to be analyzed in subsequent trials. We are aware that the design of the current study does not allow us to reach any definite conclusions regarding long-term outcome, because the trial did not include a control group and may have been influenced by a positive selection bias. Moreover, a potential positive systemic effect of locally applied IL-2 is questionable because of the inability of the treatment to induce regression of distant metastases that were not treated directly in 5 of our patients. Nevertheless, our observations strongly support further pursuit of this therapeutic approach in a randomized trial that includes a control (surgery-alone) group.
In conclusion, intratumoral IL-2 treatment elicited complete local responses of all injected lesions in 69% of patients, especially in patients without visceral metastases. A high proportion of patients could be rendered completely free of injected metastases using this approach. A large number of metastases was not associated with a poorer long-term outcome. Therefore, intratumoral IL-2 may be regarded as a promising therapeutic option for a selected subgroup of patients with melanoma and also may be feasible for the treatment of other accessible solid tumor entities. Further studies will be required to investigate the mode of action and the impact on survival of intratumoral treatment with IL-2.
CONFLICT OF INTEREST DISCLOSURES
Parts of this work were funded by Deutsche Forschungsgemeinschaft grant SFB685 and by a research grant of Novartis GmbH, Nuernberg, Germany.