Future directions in regional treatment strategies for melanoma and sarcoma

Abstract

Hyperthermic isolated limb perfusion (HILP) with melphalan and more recently isolated limb infusion (ILI) with melphalan +/− dactinomycin are common treatment modalities for both in-transit melanoma of the extremity and advanced extremity sarcoma. In order to further optimize treatment, future research should focus on selection of appropriate patients, verification of a technique that produces consistent results while maintaining acceptable toxicity, and development of novel strategies and agents. Development of these novel agents and strategies has potential to not only improve the efficacy of regional chemotherapy but may also help guide future strategies for systemic treatment.

• Regional chemotherapy
• melanoma

Introduction

The techniques of delivering regional chemotherapy to an isolated extremity via hyperthermic isolated limb perfusion (HILP) and more recently isolated limb infusion (ILI) have been relatively effective limb-sparing treatment modalities for patients with in-transit melanoma or advanced sarcoma of the extremity. While no study has demonstrated a survival benefit for therapeutic HILP, evidence does suggest that a survival benefit exists for melanoma patients achieving a complete response (CR) to therapy possibly independent of regional nodal status [1–5]. In order to further maximize the therapeutic effects of treatment and improve CR rates, several issues need additional consideration and expanded comprehensive evidence. Specifically, future efforts should focus on the identification of appropriate patients who will benefit from treatment, determination of a technique that produces consistent responses across similar populations, and the development of novel strategies and agents that can improve response. Clarifying these issues can optimize clinical response with the ultimate goal to improve survival while maintaining acceptable levels of toxicity.

Primary indications

HILP with melphalan has generally been the accepted treatment for patients with in-transit disease that is unresectable and for unresectable soft tissue sarcoma of the extremity. Additionally, HILP has been effective as a palliative treatment in selected stage IV melanoma patients with symptomatic advanced limb disease [6]. There have been no randomized controlled trials comparing HILP with other treatment options for patients with measurable in-transit disease. Assessing the results of HILP for truly unresectable disease is difficult because often potentially resectable recurrent lesions are left in situ to monitor the effect of HILP. Complete response (CR) rates from HILP have been reported to be 40–82% and averaged out to be 54% in a retrospective meta-analysis [7–10]. Modifications to the procedure as well as alternative agents to melphalan have been studied. The introduction of hyperthermia was based on evidence that hyperthermia can improve uptake of chemotherapy in tumor cells [11]. Currently, mild hyperthermia is advocated as a balance between optimizing response while maintaining acceptable levels of toxicity [12], [13]. Additionally, several different drugs have been tried in regional perfusion treatments including cisplatin [14–18], DTIC [19], and nitrogen mustard [20], with none being superior to melphalan. However, the addition of dactinomycin has been widely used in combination with melphalan for both HILP and ILI because of exceptionally good response rates (CR 73%) when the combination was administered by conventional HILP in a small number of patients without any apparent increase in toxicity [21]. However, there are no randomized control trials comparing the combination of melphalan plus dactinomycin to melphalan alone for either HILP or ILI. Currently, both HILP and ILI are primarily utilized in the treatment of patients with in-transit melanoma of the extremity although other subsets of patients including those with sarcomas, and Merkel cell carcinomas may benefit from treatment.

Adjuvant treatment

Selection of patients likely to benefit from HILP in melanoma has been examined in several clinical trials. A large multicenter prospective randomized trial demonstrated no benefit on time to systemic metastases or survival in patients undergoing adjuvant HILP with melphalan after excision of their primary melanoma [22]. This is not necessarily surprising as historically only 2–10% of extremity melanoma cases recur as in-transit metastasis [23]. HILP by design has the theoretical benefit of treating in-transit disease but not systemic or nodal disease. Therefore, to establish the potential efficacy, a very large trial would be necessary to try and demonstrate a difference in the small number of patients who theoretically could benefit from the intervention. The authors’ conclusion was that in view of the lack of a significant survival benefit in addition to the considerable morbidity and cost of adjuvant HILP, this procedure after excision of high risk primary melanoma could not be recommended. However, this practice does not take into consideration those subpopulations of patients who might benefit the most from early regional treatment. This population would include those individuals at highest risk of developing in-transit disease such as patients with thicker primary tumors, ulceration, regional lymph node metastasis, and lower extremity location of primary tumor [24–26]. While tumor thickness has traditionally been the most sensitive parameter for predicting the metastatic risk of cutaneous melanoma, other prognostic markers like the extent of lymphangiogenesis may provide more precise identification of patients at risk for recurrence [27]. In addition, future methods like microarray analysis may identify particular pathway deregulations or other aspects of tumor biology that can more accurately predict patients who will develop in-transit disease and would thus benefit from prophylactic HILP or ILI. Treatment to potentially decrease the occurrence of in-transit disease is important as the presence of in-transit metastases is associated with poor prognosis with 5-year survival rates ranging from 12% to 37% [28–30].

The value of HILP has also been examined in a prophylactic setting after excision of in-transit metastases. A randomized Phase III trial of 69 patients showed no overall survival benefit despite an improved disease-free survival [31]. While this study is limited by a low number of patients, the cost and morbidity of the procedure precluded a conclusion for the recommendation of prophylactic HILP. However, with the introduction of isolated limb infusion, novel well tolerated agents, and possible strategies to predict response, the delivery of regional chemotherapy to patients with high-risk melanoma of the extremity after excision of primary or in-transit disease may prove beneficial and should be the focus of future clinical trials.

Chemotherapy agents

Although melphalan has been the drug of choice for regional chemotherapy, the intravenous (IV) formulation of temozolomide (TMZ) has shown tremendous promise as a regional chemotherapy agent in preclinical models. TMZ is a DNA alkylating agent which is spontaneously converted to the active metabolite MTIC. In an animal model, regional infusion of TMZ was more effective than systemic therapy and frequently more effective than melphalan via regional infusion especially in tumors with low O6-alkylguanine-DNA alkyltransferase (AGT) activity [32]. Treatment with TMZ can be further enhanced with concomitant presence of hyperthermia during temozolomide infusion [33]. Another factor that makes ILI with TMZ a potentially promising approach is the ability to more accurately predict response. In recent work using a xenograft model of extremity melanoma, no correlation was identified between melphalan sensitivity and the GST/glutathione cellular detoxification pathway. In contrast, a strong correlation between the levels of AGT activity and percentage increase in tumor volume was noted for tumors treated with temozolomide; xenografts with low AGT activity showed much higher levels of response to regional TMZ [34]. Regional therapy with TMZ may ultimately prove to be more effective than melphalan in certain subsets of tumors. However, TMZ has not been used as a regional chemotherapy agent in humans due to its unavailability as an intravenous formulation. The IV formulation is currently under FDA review with phase I regional trials planned for melanoma patients pending FDA approval using both ILI and HILP.

Immunotherapy with the addition of cytokines to HILP with melphalan has also been explored for the potential to augment anti-tumor activity. Tumor necrosis factor α (TNFα) has been studied as an agent for regional chemotherapy. The combination of melphalan plus TNFα in HILP has been reported to achieve higher CR rates than melphalan alone [35], [36], [37]. A small randomized trial of 103 patients by Fraker et al. reported CR rates of 58% for the HILP with melphalan alone arm and 72% for HILP with melphalan plus interferon (INF) plus TNF- α [38]. However, a recently completed larger randomized prospective phase III trial run by the American College of Surgeons Oncology Group (ACOSOG) which compared melphalan alone to melphalan plus TNF was stopped at its interim analysis by the data safety monitoring committee because no response advantage to TNF was identified and there were three times the number of grade IV adverse events in the TNF arm including two treatment related amputations [39]. In addition, the single institution response rates of 80% using melphalan were not reproduced, with overall response rates in the ACOSOG trial being approximately 70% in both arms with only 25% of patients in each arm exhibiting a complete response. However, there was a trend toward more complete responders and more durable complete responses seen in the melphalan plus TNF group at the 6-month follow-up time point. Further exploration of the addition of TNF to melphalan in the reperfusion of patients who responded poorly to initial perfusion with melphalan alone may be appropriate.

In contrast to melanoma, the use of TNF appears to increase the effectiveness of melphalan when used in regional therapy for advanced extremity sarcoma. In 186 patients with locally advanced soft tissue sarcoma, major tumor responses were seen in 82% of patients after HILP with melphalan and TNF [40]. Treatment also rendered the large sarcomas resectable in most cases [40]. The study was conducted with modest systemic and loco regional toxicity. Thus the addition of TNF to melphalan has been advocated for use in regional therapy of advanced extremity sarcoma. Several single institution studies have subsequently reported similar high response rates [41], [42].

The cytokine interleukin 2 (IL-2) as well as the inflammatory mediator, histamine (Hi) have also been studied in combination with melphalan HILP in a rat experimental model. In a rat experimental model of melphalan-based HILP for lower limb soft tissue sarcoma, histamine appeared to augment tumor response with an overall response (OR) of 66% [43]. Similarly, HILP with IL-2 plus melphalan resulted in an OR of 67% compared to an OR of 17% with melphalan alone in the same experimental model [44]. The combination of histamine and IL-2 to melphalan HILP had no further benefit and was found to have an OR of 28%, less than each achieved alone when combined with melphalan HILP [45]. As new agents and modifications to treatment with regional chemotherapy are studied for their potential to improve response, the emphasis on maintaining acceptable treatment associated toxicity should be maintained.

Treatment modifications to decrease toxicity: Isolated limb infusion

Complications from HILP are common. Significant regional toxicity with skin necrosis, deep tissue necrosis (compartment syndrome) and peripheral neuropathy can occur. Vascular catastrophe requiring arterial reconstruction or amputation is uncommon but may occur [46]. In addition, approximately 1–3% of patients can suffer complications leading to amputation, 10–15% develop compartment syndrome, 30–40% develop lymphedema, and 5–8% can sustain long-term neuropathy [46], [47]. Repeated limb perfusions are technically difficult because of surgical scarring. The overall complication rate increases from 28% for a single procedure to 51% for a second procedure. As discussed previously, the addition of TNF was associated with three times the number of high grade adverse events and an observed CR in only 25% of patients in each study cohort. The disappointing results of this trial combined with the toxicity of HILP have led to the search for more effective, durable, and less toxic regional treatment strategies.

Isolated limb infusion for recurrent extremity melanoma and unresectable sarcoma is a strategy attempting to deliver regional melphalan in an equally effective but less toxic fashion than HILP. ILI, developed in Australia at the Sydney Melanoma Unit, differs from HILP in that ILI circulates blood in an isolated extremity at a much slower rate than HILP and for a duration of only 30 minutes [48]. In addition, during ILI the extremity is hypoxic which leads to marked acidosis in contrast to HILP where the pump oxygenator maintains the oxygenation and acid/base status of the extremity. ILI, which has not been as extensively studied as HILP, had in a preliminary report a single institution overall response rate of 85% with 41% being complete responders and 44% being partial responders [49]. The mean time to recurrence for a complete responder was 12 months. During ILI, melphalan is given in conjunction with dactinomycin, and the treatment is reportedly associated with a lower complication rate than HILP.

The initial experience in this country has not been as promising as the Australian experience. A recent report from Memorial Sloan-Kettering Cancer Center reported a 23% CR with an overall response rate of 50% while the combined experience of both Duke and MD Anderson Cancer Center described a 33% complete response rate and a 67% overall response rate in a cohort of 41 patients [50], [51]. Furthermore, we have recently analyzed data in 58 patients undergoing ILI with melphalan and dactinomycin at Duke University over a four-year period. CR was 30%, the partial response (PR) was 14%, and there was no response in 56% of patients at 3 months [52]. In comparison to the Duke experience with HILP (n = 59), a significantly better response rate at 3 months (CR 57%, PR 31%, NR 12%) was observed with perfusion. However, toxicity was significantly greater with HILP including 9 compartment syndromes and 2 amputations. One possible explanation for the discrepancy in response rates is the relatively low flow system of ILI compared to HILP may lead to lower levels of melphalan uptake by tumor cells although this has not been examined thoroughly. While ILI may not be as effective as HILP, the low toxicity and morbidity of ILI is an appealing option for patients as a first step in regional treatment reserving HILP for patients who are ILI failures or non-responders. In addition, the low toxicity allows for repeat ILIs without an apparent increase in toxicity. The potential to improve complete response with novel targeted agents and new chemotherapeutic agents may help further improve the therapeutic index of ILI making the technique the preferred treatment option for patients with in-transit melanoma the extremity.

Optimizing response

The ability to obtain a CR following HILP was the strongest predictor of survival in the Duke experience with HILP [1]. Similarly, in a recent review of the long term results of HILP, complete response was a positive prognostic indicator for survival [2] while the ability to achieve a CR in patients with a lower stage had an increased likelihood but was not statistically significant. In the SMU experience with ILI, median survival was longer in patients in whom a CR was achieved [49]. The finding that CR is a predictor of survival may suggest that tumors which are chemosensitive have a more favorable tumor biology as compared to tumors which possess or have acquired chemotherapy resistance. Recent work has explored resistance to chemotherapy and has shown in a nude rat human xenograft model that targeting resistance pathways with novel agents can improve tumor response to ILI with melphalan [53], [54]. We have also begun work with systemic administration of targeted therapies that have a low toxicity profile in combination with melphalan via ILI.

One approach to overcoming regional chemotherapy resistance has been to give a modulator of drug resistance proteins systemically pre, peri, and post regional chemotherapy treatment with melphalan and more recently regional therapy with temozolomide (TMZ). Buthathione sulfixime (BSO), is an inhibitor of gamma-glutamyl-cysteine-synthetase (c-GCS). C-GCS is a rate-limiting enzyme in GSH synthesis. Elevated intracellular GSH levels are associated with increased resistance to melphalan [55], [56]. Using a melanoma infusion animal model, short-term systemic BSO therapy was found to potentiate the effects of regional melphalan without increasing toxicity [57]. A phase I trial using a 3-day infusion of BSO around the time of a regional melphalan infusion has recently been approved by the Duke IRB and is designed to determine the efficacy of this approach in patients who have failed a previous melphalan-based therapy. In this trial, BSO will be given at its phase II dose and the melphalan will be dose escalated. Tumor biopsies before and during BSO therapy will attempt to document the degree of glutathione depletion before the regional melphalan treatment so that more can be learned about this type of treatment strategy.

We also examined whether modulation of resistance pathways would be an effective strategy for improving tumor response to other alkylating agents such as TMZ. O6-benzylguanine (O6BG) is an inhibitor of the DNA repair enzyme O6-alkylguanine-DNA alkyltransferase (AGT), the main resistance pathway for TMZ in melanoma. Using systemic O6BG with regional temozolomide via isolated limb infusion in an animal model has been shown to enhance tumor response [34]. While this demonstrates the importance of tumor resistance mechanisms in determining response for regionally treated tumors, the use of O6BG will not be an option until after the regional TMZ ILI phase I trial is completed. This trial is currently in preparation at Duke University.

Another potential method of treatment to improve CR rates with regional chemotherapy is administration of targeted therapies that may disrupt various cell signaling pathways that may make the tumors more susceptible to the effects of cytotoxic chemotherapies. Chemoresistance in melanoma may be an intrinsic property of the tumor cell potentially related to one of a number of molecular and genetic changes that leads to deregulation of signaling pathways. Preliminary in vitro data suggests that the inherent chemoresistance of melanoma may be overcome by using traditional chemotherapy agents in combination with a targeted drug that inhibits selected signaling pathways. One promising targeted therapy has been with the N-cadherin antagonist ADH-1 (Exherin) which is a cyclic pentapeptide that disrupts N-cadherin binding interactions. The cadherins are a large supergene family of proteins that are involved in cell to cell adhesion. Malignant transformation in melanoma is characterized by a ‘switch’ from E-cadherin to N-cadherin and is associated with changes in intracellular signaling pathways leading to increased proliferation, survival and angiogenesis and decreased apoptosis. The use of systemic ADH-1 in conjunction with regional melphalan shows marked increases in tumor responses even in melphalan resistant tumors [58]. A phase I dose escalation trial of systemic ADH-1 in combination with melphalan via ILI for patients who have AJCC Stage IIIB or IIIC extremity melanoma is currently open to accrual at Duke and MDACC.

Other potential targets involve the B-raf and N-ras signaling pathways. Sixty percent of melanoma tumors harbor an activating mutation in B-raf kinase [59].These genetic mutations lead to deregulation of multiple signaling pathways in the cell, but most noticeably increase activity of the MAP kinase signaling pathway which leads to increased proliferation and survival of tumor cells. As single agents, targeted therapies aimed at these proteins have been relatively unsuccessful in terms of producing tumor response. Interestingly, however, surprising responses were seen in a small number of clinical trials in melanoma where targeted therapy was utilized in conjunction with other chemotherapy agents [60], [61]. Initial clinical trials with the Bayer compound Sorafenib (Nexavar), a B-raf kinase inhibitor, suggest that as a single agent it may have limited activity. However, when used in combination with other chemotherapy agents like DTIC, sorafenib has shown some encouraging responses [62]. In contrast, the combination of sorafenib with carboplatin plus paclitaxel for patients with advanced melanoma did not demonstrate improvements in any clinical endpoints [63]. While these studies can potentially provide a basis for combinations that utilize regional chemotherapy, the ability to deliver regional chemotherapy at doses 10–15 times higher than what can safely be administered systemically may lead to higher levels of synergistic interaction and response. The interesting results from the trial of Sorafenib in combination with systemic DTIC has prompted a new chemosensitizing study using Sorafenib in patients with regionally advanced metatstatic melanoma. This trial, combining oral systemic Sorafenib before and after ILI with melphalan, has recently been IRB approved at Duke University and will also be carried out at MSKCC. Another potential target in melanoma is the src family of kinases. Stat3, a downstream target of src kinase, has been found to be activated in the majority (85%) of melanoma cell lines [64]. Stat3 has been shown to not only promote tumor growth and survival but also appears to up regulate VEGF and promote tumor angiogenesis [65]. Inhibition of src kinase blocked the growth of human melanoma cell lines with elevated Stat3 activity [65] and induced apoptosis by inhibiting the expression of anti-apoptotic genes Bcl-x_L and Mcl-1. Src inhibitors like Dasatinib (Sprycel) may also potentiate the effectiveness of traditional chemotherapy. A phase I trial combining Dasatinib with melphalan via ILI at Duke University is currently being planned. A summary of these phase I trials is shown in Figure 1.

Figure 1. Current status of various phase 1 regional therapy trials.

There are currently several other combination therapies currently under investigation for the treatment of metastatic melanoma which may also be applicable for combination with infusion chemotherapy where one can take advantage of the high dose of regionally administered chemotherapy. While the list of potential targets and agents is extensive, three particular agents are noteworthy. One possible targeted therapy agent is bevacizumab, a monoclonal antibody to vascular endothelial growth factor. This agent has shown survival benefits when used in combination with standard chemotherapies in patients with metastatic colorectal cancer [66]. There also appears to be anti-tumor activity with combination therapy in metastatic breast and lung cancer [67], [68]. Most recently, bevacizumab in combination with irinotecan has shown promise in the treatment of recurrent glioblastoma multiforme, demonstrating a high response rate and an improvement in the 6 month progression-free survival [69]. There have been a few case reports using bevacizumab in combination with chemotherapy for patients with metastatic melanoma that may suggest activity [70], [71]. In a recent phase II trial of bevacizumab alone versus bevacizumab plus IFN-α in 32 patients with metastatic melanoma, stable disease was achieved in 5 of 16 patients who received bevacizumab alone while there was one partial response observed in 1 of 16 patients treated with the combination [72]. While more preliminary evidence is still needed, the addition of bevacizumab to ILI with melphalan may provide improved response rates. Another potential chemosensitizing targeted therapy that has been explored is the multikinase inhibitor imatinib mesylate. While the frequency of KIT mutations in melanoma is unclear [73], [74], data suggests that melanomas with a lentiginous growth pattern have frequent genetic alterations of KIT and these patients may gain a therapeutic benefit from imatinib in combination with another chemotherapy [73], [74]. Finally, STA-4783 is a drug that enhances the therapeutic index of paclitaxel. The combination demonstrated anti-tumor activity in patients with metastatic solid tumors [75]. While combination therapies continue to be explored for metastatic melanoma, combination therapies utilizing ILI and HILP for in-transit melanoma also have potential to improve the therapeutic index of treatment.

In addition to exploring novel therapies, research is also being conducted into examining how gene expression signatures can optimize regional chemotherapy treatment in patients with in-transit malignant melanoma of the extremity. Specifically, recent work has demonstrated that gene expression signatures can predict response to various cytotoxic chemotherapeutic agents which may ultimately provide an opportunity to optimize the use of these drugs in patients [76]. In a similar fashion to methods described previously [76], [77], a melphalan sensitivity signature using microarray gene expression data combined with in vitro drug response data in melanoma cell lines has been described. This signature is currently being validated using biopsy samples from patients prior to treatment with ILI with melphalan to test if the signature correlates to clinical response as a measure of melphalan sensitivity. Application of a gene signature like this one in the context of targeted agents can help optimize both regional and systemic therapeutic strategies in the future. For example, the gene signature obtained from a biopsy sample could not only predict tumor response to chemotherapies allowing for optimal initial chemotherapy selection in ILI but also identify deregulated pathways and mechanisms of resistance that can subsequently be targeted with appropriate agents. While further development of this technology is needed, the potential exists to improve outcomes by individualizing treatment strategies.

Conclusion

The delivery of regional chemotherapy to an isolated limb continues to be an important treatment for patients with in-transit melanoma of the extremity and unresectable soft tissue sarcoma. The focus of the majority of these approaches is on melanoma which is a more homogenous and common entity presenting as advanced disease in an extremity. There are multiple strategies and therapies currently being explored that have the potential to improve response to regional therapy. The regional model of advanced extremity melanoma, with its easy accessibility of tumor tissue allows a unique opportunity to define how to utilize targeted agents and correlated alterations in functional proteins with tumor response and drug pharmacokinetics. The novel regional infusion trials discussed in this chapter should help facilitate our comprehension of how targeted therapies should be optimally utilized to overcome chemotherapy resistance. Additionally, while these trials may augment the therapeutic index of regional chemotherapy, they may also serve as a basis to more logically direct the development of novel approaches to systemic treatment strategies. The fact that most of these approaches focus on melanoma does not preclude their potential application to other malignancies like sarcomas, Merkel cell carcinomas, and squamous cell carcinomas that may recur in a multifocal pattern in the extremity.