Introduction
Melanoma is a highly immunogenic tumor and consequently, efforts have been centered on the development of immune-based treatments for this malignancy [
1]. Interferons were initially described in the mid-1950s as proteins that interfere with viral replication [
2,
3]. Interferons are cytokines that activate Janus kinases (Jak), which lead to phosphorylation and activation of transcription factors belonging to the signal transducer and activator of transcription (STAT) family [
4,
5]. Interferon-alpha (IFN-α) became available for use in clinical trials in the mid-1980s [
6]. Results suggested that IFN-α inhibited the proliferation of malignant cells and stimulated immune effectors; therefore, IFN-α was initially used in patients with advanced disease [
7,
8]. Since then, several meta-analyses have demonstrated that high-dose adjuvant IFN-α (daily 20 MU/m
2 intravenous induction therapy for 1 month followed by maintenance subcutaneous 10 MU/m
2 three times per week for at least 1 year) can prolong the disease-free interval in high-risk melanoma patients [
9].
The introduction of checkpoint inhibitor therapy has revolutionized the adjuvant therapy of melanoma; however, there remains a role for IFN-α in this setting based on the potential for cancer immune escape or autoimmune events with CTLA-4 and PD-1 blocking antibodies [
10‐
14]. There has also been significant advances in mitogen-activated protein kinase (MAPK) targeted therapies, particularly for BRAF (an intracellular signaling kinase) and MEK (signaling molecule downstream of BRAF). A recent clinical trial demonstrated significant improvement in both relapse-free survival and overall survival with adjuvant dabrafenib (BRAF inhibitor) plus trametinib (MEK inhibitor) in patients with stage III melanoma. These therapies are now approved for adjuvant therapy in BRAF mutated tumors [
15]. However, since only approximately 40–50% of melanoma cells harbor an activating BRAF mutation, there still remains a role for IFN-α in this setting as the remaining 50–60% of melanomas would not be susceptible to BRAF-targeted therapies.
IFN-α activates the Jak-STAT signaling pathway and induces synthesis of hundreds of different proteins [
4,
5]. Our group has shown that STAT1-mediated gene regulation within immune effectors is necessary for mediating the anti-tumor effects of IFN-α and also that the amount IFN-α administered to melanoma patients is likely in excess of the optimal biological dose [
4]. Indeed, high doses of IFN-α appear to be no more effective in the induction of phosphorylated STAT1 (p-STAT1) and in the transcription of interferon-stimulated genes (ISGs) than intermediate doses [
16,
17]. Our group’s studies in genetically manipulated mice have shown that suppressors of cytokine signaling-1 (SOCS1) and SOCS3 negatively regulate IFN-induced Jak-STAT signal transduction, gene regulation and anti-melanoma activity, and that high doses of IFN-α can induce SOCS proteins [
18,
19].
We hypothesized that lower doses of IFN-α would be superior for induction of IFN signal transduction in patient immune cells. A prospective clinical trial was performed wherein patients eligible for adjuvant IFN-α-2b received 1 month of standard intravenous high-dose IFN-α-2b (20 MU/m2) followed by subcutaneous IFN-α-2b at a dose of 10 MU/m2 with dose reductions at set intervals down to a level of 4 MU/m2. Jak-STAT signal transduction and transcription of ISGs in patient peripheral blood mononuclear cells (PBMCs) were monitored during the course of adjuvant IFN-α therapy. The objective of this pilot study was to determine if lower doses of IFN-α were as effective in the induction of IFN signal transduction and gene expression as the standard high dose regimen.
Materials and methods
Eligibility criteria
A prospective pilot study of IFN-α-2b dose-reduction in melanoma (NCT01460875) was conducted at The Ohio State University under institutional review board approval (OSU-07033) with support from Merck Inc. Eligible patients were candidates for adjuvant IFN-α-2b after having undergone successful surgery for high-risk melanoma (Breslow thickness > 4 mm or lymph node involvement) or complete resection of metastatic disease and completion of 20 treatments of standard intravenous IFN-α-2b within 2 months of beginning treatment on this study. Patients were required to meet the following criteria: definitive surgery performed not later than 90 days prior to start of intravenous IFN-α-2b treatment, no evidence of persistent/recurrent disease, Eastern Cooperative Oncology Group (ECOG) performance status ≤ 2, life expectancy > 6 months, normal organ and marrow function, and ability to provide written informed consent.
Treatment regimen
Prior to treatment, patients completed 20 treatments of standard intravenous IFN-α-2b (20 MU/m2 5 days a week for 4 weeks). Patients then began subcutaneous IFN-α-2b injections at the standard dose of 10 MU/m2 thrice weekly for 4 weeks. After 1 month of therapy at 10 MU/m2, IFN-α-2b dose reductions were initiated. The IFN-α-2b dose was reduced to 8, 6, and 4 MU/m2 at 2-week intervals. The first dose of IFN-α-2b at each dose level was administered in the outpatient clinic and subsequent doses were self-administered as an outpatient. At each clinic visit, patients were evaluated for toxicities and venous blood was obtained for correlative assays. Heparinized blood samples were obtained prior to administration of IFN-α and at 1 and 4 h after administration. Once a 4-MU/m2 IFN-α-2b dose was achieved, patients went on to receive subcutaneous therapy for a total of 11 months. Repeat blood draws were performed every 3 months to confirm the activity of this dose.
Clinical outcome assessment
History and physical examinations were performed every 3 months. Patients with recurrent disease were removed from trial therapy. Overall survival was defined as time to death due to any cause evaluated from time of surgery or from time of treatment initiation. Time to end of active treatment was defined as the time from start of treatment to the time patients were removed from therapy due to any cause, including completion of therapy per protocol. Patients who were event-free (e.g. alive) at their last evaluation was censored at that time point.
Flow cytometric analysis of phosphorylated STAT1
PBMCs were isolated from patient blood via centrifugation with Ficoll-Paque Plus (Amersham Pharmacia Biotech). The phosphorylated form of STAT1 (Tyr
701) in cryopreserved PBMCs was measured by intracellular flow cytometry, as previously described [
20,
21]. Anti-p-STAT1 (Tyr
701) conjugated antibody and isotype control antibody were obtained from BD Biosciences Pharmingen (San Jose, CA, USA).
Reverse transcription polymerase chain reaction (RT-PCR)
Total RNA from PBMCs was extracted using Trizol reagent (Life Technologies, Grand Island, NY, USA). Reverse transcription reactions were performed using 500 ng RNA in a 20-µl reaction with the high-capacity reverse transcription kit (Life Technologies). cDNA was used as a template to measure the expression of human SOCS1, OAS1, CXCL10, and CD69 genes by quantitative RT-PCR using pre-designed primers (Life Technologies). β-Actin served as an internal control (Life Technologies). RT-PCR reactions were performed in triplicate using the ABI PRISM 7900HT fast RT-PCR system (Applied Biosystems).
Statistical methods
Clinical characteristics were descriptively summarized for all evaluable patients. Changes in p-STAT1 with decrease in dose levels were evaluated graphically for each patient, focusing on absolute change in p-STAT1 between dose level 10 MU/m2 and dose level 4 MU/m2 and specific p-STAT1 levels at each of these dose levels. Clinical outcomes, including toxicity and tolerability measures, were summarized across all patients, and reasons for end of active treatment were also dichotomized as completed treatment per protocol vs. not. Comparisons between groups were analyzed using Wilcoxon rank sum tests and differences between dose levels in the same patients were evaluated using the paired nonparametric Wilcoxon signed rank test. Time-to-event outcomes were analyzed graphically using the standard Kaplan–Meier methods. Modeling on time-to-event outcomes was done using average hazard ratio estimates in the Cox regression model to accommodate non-proportional hazards.
Discussion
The present study demonstrates that dose reduction of IFN-α-2b from 10 to 4 MU/m
2 is feasible and well-tolerated. Only one grade 4 toxicity was encountered (lymphopenia) and the documented grade 3 toxicities were easily reversed and consistent with those observed in prior trials. This dose-reduction regimen was associated with estimated survival rates that were on par with the 5-year relapse-free survival (RFS) and OS rates reported in the high-dose IFN-α E1690 clinical trial (45% vs. 44% RFS and 64% vs. 52% OS, respectively) [
22] and those reported in the high-dose IFN-α E1684 clinical trial (median RFS of 3.7 vs. 1.7 years and median OS of 6.8 vs. 3.8 years, respectively) [
23]. Subcutaneous administration of IFN-α-2b led to the induction of p-STAT1 in circulating immune cells with no difference in the levels of p-STAT1 at the 10-MU/m
2 dose as compared to the 4-MU/m
2 for the group of patients as a whole. Additionally, following the reduction in IFN-α-2b dose from 10 to 4 MU/m
2, the expression of several well-characterized ISGs (OAS1, CXCL10, and CD69) was not significantly reduced. However, the higher IFN-α-2b dose induced greater transcription of SOCS1, a negative regulator of the IFN-α immune response. In a separate analysis, higher p-STAT1 levels were found to be associated with prolonged TTR for both the 10 MU/m
2 and 4 MU/m
2 doses.
This pilot study provides evidence that IFN-α-2b at a dose of 4 MU/m2 is effective in activating circulating host immune effector cells and stimulating IFN-α induced gene expression. These results suggest that an analysis of downstream signal transduction and gene regulation may have utility in the dose selection process for IFN-α therapies. However, when looking at markers of interest to risk stratify patient populations, independent validation is required. In this study, independent validation was not performed. Therefore, data are not presented as definitive results but rather to explore and evaluate the role of p-STAT1 levels and their predictive utility in relation to clinical outcomes, such as TTR. As such, this report does not represent a true training set, but rather provides results that inform and support future larger studies evaluating the role of p-STAT1 levels in relation to clinical outcomes. Similarly, with respect to p-STAT1 level cutoff points, these results need to be validated and verified in larger, confirmatory studies before being applied in clinical practice.
Multiple IFN-α regimens have been evaluated as adjuvant therapy for intermediate or high-risk melanomas, with mixed results. The Eastern Cooperative Oncology Group (ECOG) trials E1684 and E1694 demonstrated durable improvement in both RFS and OS utilizing a high-dose IFN-α regimens [
23,
24]. However, the E1690 high-dose IFN-α trial only demonstrated durable improvement in RFS [
22] and the Sunbelt Trial demonstrate no RFS or OS benefit with the use of high-dose IFN-α [
25]. Given the toxicity of high-dose IFN-α, several clinical trials have evaluates low- and intermediate-dose regimens with, again, mixed results. While evaluating the utilization of intermediate-doses of adjuvant IFN-α, the European Organization for Research and Treatment of Cancer (EORTC) 18991 and the Nordic IFN trials demonstrated improvement in RFS, [
26,
27], whereas the EORTC 18952 trial demonstrated no RFS or OS benefit [
28]. However, on subgroup analysis the EORTC 18952 trial did demonstrate a durable improvement in RFS and OS in patients with ulcerated melanomas [
28].
Likewise, several clinical trials have been performed to evaluate the efficacy of low-dose adjuvant IFN-α in high-risk melanoma patients [
22,
29‐
36]. The Dermatologic Cooperative Oncology Group (DeCOG) trial demonstrated durable improvement in both RFS and OS [
29], while the French Cooperative Group trial demonstrated significant extension of RFS and a clear trend towards increased overall survival (
p = 0.059) [
30]. The Italian Skin Cancer Foundation, Austrian Malignant Melanoma Cooperative Group, and Scottish Melanoma Group clinical trials demonstrated significant improvement in RFS, but no improvement in OS [
31‐
33]. However, no improvement in either RFS or OS was seen in the WHO Melanoma Programme, AIM HIGH Study-United Kingdom Coordinating Committee on Cancer Research, EORTC 18871, or ECOG 1690 trials [
22,
34‐
36]. The vast heterogeneity of treatment regimens used and mixed treatment responses seen within these trials makes interpretation of the results difficult. Most studies have failed to define clinical or demographic features that would identify patients more likely to respond to IFN-α treatment regimens. However, since STAT1-mediated gene regulation within immune effector cells is necessary for mediating the anti-tumor effects of IFN-α [
4], the variations in treatment responses seen in these studies may be the result of the varied ability of the administered low-dose IFN-α to activate the Jak-STAT signaling pathway on a patient-by-patient basis. Therefore, individual evaluation of Jak-STAT signaling in response to IFN-α treatment may be able to identify patients that are more likely to benefit from adjuvant treatment and aid in dose optimization.
The use of phosphorylation-state specific antibodies for intracellular flow cytometry has a unique potential for the evaluation of signaling events in immune effectors following the administration of immunomodulatory cytokines. Until the precise molecular determinants of IFN-α-responsiveness are identified, it seems reasonable to use signal transduction within immune effector cells as a surrogate marker of IFN-α action in patients undergoing immunotherapy. The present trial only evaluated the response of each patient to a pre-determined schedule of dose reduction. Given the inter-patient variability in p-STAT1 induction to a given dose of IFN-α, it is likely that the optimal dose of cytokine (i.e., that dose which induces the greatest activation of immune cells as measured by the induction of p-STAT1) would be different for each patient. We, therefore, anticipate that this method might be useful as a means of identifying the dose of IFN-α which produces optimal Jak-STAT signal transduction on a patient-by-patient basis.
A previous clinical trial of patients with metastatic melanoma evaluated the treatment regimen of bevacizumab in combination with escalating doses of IFN-α-2b (5 MU/m
2 for 2 weeks and then 10 MU/m
2 thereafter) [
37]. Levels of p-STAT1 at a dose of 5 MU/m
2 IFN-α-2b were greater or equivalent to those at a dose of 10 MU/m
2 for six of the seven patients studied. Similarly, the induction of ISGs within PBMCs at a dose of 5 MU/m
2 was greater or statistically equivalent to that observed for 10 MU/m
2 IFN-α-2b for six of the seven of the patients. Microarray analysis was performed on five patients with metastatic melanoma undergoing immunotherapy with escalating doses of IFN-α-2b. Analysis of the gene expression profile within PBMCs from these patients revealed that a total of 35 genes (e.g., CD69, CXCR6, IL8, PBEF) were induced to a greater extent with 5 MU/m
2 as compared to 10 MU/m
2 IFN-α-2b (fold induction ≥ 2). Notably, SOCS1 and SOCS3 transcripts were significantly higher in patient PBMC following the 10-MU/m
2 dose of IFN-α-2b. These results, along with the results presented in this study, suggest that lower doses of IFN-α-2b may be just as effective as higher doses with respect to the induction of Jak-STAT signal transduction and ISG expression within immune effector cells.
There have been significant advancements in the adjuvant therapy of melanoma with the use of checkpoint inhibitors and targeted therapies for BRAF mutant melanomas. However, there still remains a role for adjuvant IFN-α in the setting of BRAF wild-type melanomas and in patients at increased risk of cancer immune escape or autoimmune events with CTLA-4 and PD-1 blocking antibodies. Although high-dose IFN-α is approved for the adjuvant treatment of melanoma, the substantial treatment-related toxicities have impeded the adoption of this regimen. The present study demonstrates that IFN-α-2b dose reduction is feasible, well tolerated, and associated with reasonable response rates. Additionally, this study provides data to support the contention that the standard subcutaneous dose of IFN-α may be higher than is necessary for maximal activation of the immune system. This study offers a novel method for potential dose-optimization, on a patient-by-patient basis, through individual analyses of signal transduction and gene regulation. However, the clinical impact of dose-optimized adjuvant IFN-α in patients with high-risk melanoma on OS and RFS needs to be evaluated in a large, randomized controlled trial.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.