Journal of Surgical Oncology 2014;109:320–326
Intralesional Immunotherapy for Melanoma PETER HERSEY, FRACP, D.Phil1,2*
AND
STUART GALLAGHER, PhD2
1
Melanoma Institute Australia, North Sydney, NSW, Australia 2 Kolling Institute, The University of Sydney, St Leonards, NSW, Australia
Intralesional immunotherapy of melanoma has two complementary aims. One is to cause regression of the injected metastasis. The other is to incite or modulate systemic immune responses in such a way that non‐injected metastases will also undergo regression. A number of phase 1 and phase II studies with cytokines, viral, or bacterial agents have been conducted but their use has remained sporadic and has not progressed to become established treatments. Two treatments have progressed to randomized phase III studies. The most promising of these is based on intralesional injection of a genetically modified herpes simplex virus (HSV) (T‐Vec). Initial results have shown a significant effect on durable response rates (DRR) but effects on overall survival remain under study. The second involved injection of plasmids coding for the HLA B7 antigen (Allovectin). Despite encouraging early results the treatment did not reach its endpoints and its use has been discontinued. A phase II study involving intralesional injection of oncolytic A21 coxsackie virus (Cavatak) is also under way and is showing promise.
J. Surg. Oncol. 2014;109:320–326. ß 2013 Wiley Periodicals, Inc.
KEY WORDS: melanoma; intralesional therapy; immune responses; oncolytic viruses; GM‐CSF
INTRODUCTION The publication in 2010 [1] and 2011 [2] of reports of improved survival of patients with metastatic melanoma who were treated with the anti‐CTLA4 monoclonal antibody ipilimumab, a checkpoint inhibitor, ushered in what may be referred to as the post‐DTIC era of melanoma treatment. Prior to this DTIC had proven superior to a number of other treatments but nevertheless was relatively ineffective in treatment of the disease [3]. As a consequence, initiatives based on intralesional injection of accessible tumors in patients with locoregionally advanced metastatic stage IIIB/C or stage IV M1a melanoma were the subject of a number studies exploring intralesional immunotherapy with a range of different agents. Initial studies were largely aimed at local ablation of the lesions with agents such as vaccinia virus [4,5], thiotepa or typhoid vaccines [6]. More recently there has been a greater focus on the systemic effects of intralesional immunotherapy and trial designs have evolved that include assessment of both the local effects on the injected lesions as well as non‐ injected lesions including those in visceral sites. In the following sections we review the earlier studies on intralesional treatments with BCG and cytokines and then the more recent attempts to introduce these treatment approaches into management of melanoma. In particular we discuss their role as management transitions from the DTIC era to the era of checkpoint inhibitors and BRAF/MEK inhibitors.
HISTORICALLY IMPORTANT INTRALESIONAL THERAPY STUDIES Intralesional Bacillus Calmette‐Guerin (BCG) During the 1960 and 1970 period there was much interest in the use of intralesional injection of BCG to treat cutaneous melanoma [7]. One of the strongest advocates was Dr Donald Morton at the John Wayne Cancer Institute who in 1974 reported a 7‐year experience with this approach in 151 patients. Injection of intradermal and cutaneous melanoma resulted in regression of 90% of the injected lesions and 17% of uninjected lesions. Twenty‐five percent of patients remained free of disease for periods ranging from 1 to 6 years. Injections into subcutaneous or visceral lesions were not so effective [8]. Studies on 27 patients with lower extremity recurrences reported complete or partial disease control in 20 of the patients. Responses were associated
ß 2013 Wiley Periodicals, Inc.
with good immune function assessed by skin test responses to dinitrochlorobenzene and PPD [9]. Similar results were reported in studies on 15 patients with metastatic melanoma by Mastrangelo et al. [10]. There were five responses that were limited to patients with dermal metastases. Particular attention was focused on the possible systemic effects of BCG by reports of regression of pulmonary metastases during intralesional injection of BCG [11]. Nevertheless there were reports of systemic infections with BCG [12] and some allergic reactions. Trials of BCG therapy were also negative leading to this form of therapy falling out of favor [13].
Intralesional Cytokine Injections With GM‐CSF or Interferons One variable that may limit the potency of the host response against melanoma is the mixture of cytokines present in the tumor microenvironment. Granulocyte‐macrophage colony stimulating factor (GM‐CSF) was identified as the most potent molecule for augmenting tumor immunity following gene transfer into melanoma cells. Vaccination with irradiated melanoma cells engineered to secrete GM‐CSF enhanced host responses through improved tumor antigen presentation by recruited dendritic cells and macrophages [14]. Injection of 15–50 mg doses of GM‐CSF into two s.c. metastases in 13 patients resulted in regressions of injected or non‐injected lesions in 3 of the patients. Regressions were associated with infiltration by CD4 T cells and dendritic cells [15]. Intralesional injection of interferon (IFN) alpha or beta also attracted attention. Injection of IFN alpha into s.c. metastases in 51 patients was carried out by von Wussow et al. [16]. Half the patients received IFN alpha 2b and the other half natural IFN given three times per week. There
*Correspondence to: Peter Hersey, FRACP, D.Phil, Kolling Institute, The University of Sydney, Pacific Hwy, St Leonards, NSW 2065, Australia. Fax: 612‐9926‐4034. E‐mail: [email protected] Received 30 July 2013; Accepted 26 September 2013 DOI 10.1002/jso.23494 Published online 3 December 2013 in Wiley Online Library (wileyonlinelibrary.com).
Therapies for Melanoma were systemic responses in 9 patients in non‐injected lesions and 24 patients had responses in the injected lesions. Exploratory studies with intratumoral injection of interferon beta concluded that higher doses were probably needed to induce regression of injected lesions [17].
CURRENT INTRALESIONAL IMMUNOTHERAPY STUDIES Patients Being Studied, Treatment Procedures and Evaluations Most of the patients entered into these studies had locally recurrent disease proximal to the primary and in‐transit disease either without LN metastases (IIIb) or with LN metastases (stage IIIc) or Stage IV disease with injectable s.c. or LN metastases. In general the lesions were injected either visually or with ultrasound control using a 25–30 gauge needle and a fanning technique by which the needle is moved back and reinserted in different directions within the lesion. The volume was usually adjusted on the basis of the surface area of the lesion and total volumes set to avoid possible upper limits of toxicity. Tumor assessments were commonly carried out on a weekly basis by ultrasound in the first 4 weeks and then at 6‐week intervals by CT scans. There is interest in the use of PET scans to assess early events in the injected lesions [18,19]. Immune response criteria have been used in most of the studies most commonly at the 12‐week period. These criteria allow the inclusion of new lesions in assessment of responses and are added to the total tumor burden in assessment of CR, PR, and stable disease status [20]. Comparisons of biopsies taken prior to and at intervals during therapy were included when possible in most of the studies.
Velimogene (Allovectin) Allovectin (velimogene aliplasmid) consisted of a bicistronic plasmid encoding a major histocompatibility complex (MHC) class I heavy and light chains of HLA‐B7 and b2‐microglobulin formulated with a cationic lipid‐based system for direct intratumoral administration. Following injection into a single lesion it was intended to induce anti‐
Fig. 1. Diagrammatic illustration of possible mechanisms of action of intralesional therapy with Allovectin. Journal of Surgical Oncology
321
TABLE I. Phase II Clinical Trial Results for Velimogene Aliplasmid (VCL‐1005‐208) Response in 127 patients Overall response rate (CR þ PR) (95% CI) Duration of response (K‐M median) (95% CI) Median time to response 2 cycles (95% CI) CR PR Stable disease (SD) Progressive disease (PD)
n (%) 15 (11.8%) (6.2–17.4%) 13.8 months (6–66 months) (4 months) (2–4 months) 4 (3.2%) 11 (8.7%) 32 (25%) 80 (63%)
tumor immune responses against both treated and untreated distal lesions. Both the plasmid and lipid components of Allovectin were considered to activate of leukocytes via toll receptors. Other actions included the induction of both cytotoxic T‐cell and innate immune responses against allogeneic as well as tumor‐derived targets due to the increased MHC class I expression on tumor cells and the induction of a localized immune/inflammatory response [21]. It was considered that immune responses against the allo‐antigen would attract T cells and macrophages that would recognize and destroy the injected melanoma and non‐injected lesions [22] (Fig. 1). In the phase 2 study based on 133 patients, injections were given in single or multiple lesions and responses assessed by RECIST (Response Evaluation Criteria in Solid Tumors). In 127 evaluable patients the response rate was 11.8% and median duration of response 13.8 months. In 42 patients with stage IV melanoma there were responses in non‐ injected lesions in 21% of the patients [23] (Table I). In view of these results a phase III trial was conducted in 375 patients comparing Velimogene Aliplasmid with physician’s choice of chemotherapy (DTIC or temozolomide) in a 2:1 randomization schema. The primary outcome measure was percent DRR at 24 weeks after randomization. The trial is closed to accrual and results are expected in the latter half of 2013. Much speculation occurred as to what the median OS maybe in the phase 3 trial. In the phase 2 trial
Fig. 2. Illustration of the genetic engineering carried out on the modified herpes simplex virus and the general concepts involved in treatment with oncolytic viruses.
322
Hersey and Gallagher
Fig. 3. Outline of the OPTiM study on treatment with the modified HSV (T‐Vec).
TABLE II. Durable and Objective Responses in the OPTiM Phase III Study of Talimogene Laherparepvec (T‐Vec) ITT set DRR Objective OR (95% CI)
GM‐CSF (n ¼ 141)
Talimogene laherparepvec (n ¼ 295)
Unadjusted OR
2.1% 5.7% (1.9, 9.5)
16.3% 26.4% (21.4, 31.5)
8.9, 95% CI, 2.7–29.2, P < 0.0001 56 of 78 responders in the talimogene laherparepvec arm (72%) and 4 of 8 responders in the GMCSF arm (50%) were still in response at the last available tumor assessment
0.7% 5.0%
10.8% 15.6%
CR PR
78 patients only had 1 injection due to strict adherence to RECIST criteria. Survival in this group had been used as a proxy for the control group in the phase 3 trial. Despite this speculation a press release in august 2013 revealed that the trial had not met its endpoints in increasing percent DDR or overall survival and that the development of this treatment had been abandoned (finance.yahoo.com/news/vical‐phase‐ 3‐trial‐allovectin‐100000138.html).
Oncolytic Viruses Oncolytic viruses are those that preferentially infect and destroy cancer cells. This occurs due to direct effects of the virus on the metabolic processes in the cell but also by inducing immune responses that target the cancer cell. The latter is believed to be aided by activation of NF‐kB and release of chemokines and cytokines from the cancer cell [24,25]. Genetic engineering of certain viruses has attempted to reduce their virulence for the host and in some cases their ability to incite immune responses. Attempts to infect cancer cells by systemic administration of the virus continue but are limited by immune responses of the host. In view of this most studies have used the intralesional route and rely on immune responses generated by release of cell products and inflammatory responses generated against the infected cell [26].
presentation. In addition, ICP47 deletion increases US11 expression and thereby enhances virus growth and replication in tumor cells. The coding sequence for ICP34.5 was replaced by the coding sequence for human GM‐CSF to enhance the immune response to tumor antigens released after virus replication [28,29] (Fig. 2). Phase II studies were carried out in 50 patients with either stage IIIc (10 patients) or stage IV (40) melanoma patients. Treatment involved intratumoral injection of up to 4 ml of 106 pfu/ml of JS1/34.5‐/47‐/GM‐ CSF followed 3 weeks later by up to 4 ml of 108 pfu/ml every 2 weeks for P-value*
All ITT Stage IIIB/C Stage IVM1a
T‐Vec was previously referred to as OncoVEX GM‐CSF when it was owned by BioVex (Woburn, MA). It is an immune‐enhanced, oncolytic herpes simplex virus type 1 (HSV‐1) that is deleted for the neurovirulence factor ICP34.5 [27] and this results in tumor‐selective replication. It is also deleted for ICP47, which normally blocks antigen Journal of Surgical Oncology
HSV-1 Negative HSV-1 Positive
18%
***
12% 3%
Sex- Men
**
10%
3% 0%
**
8% 24%
4%
Intralesional Immunotherapy for Melanoma PETER HERSEY, FRACP, D.Phil1,2*
AND
STUART GALLAGHER, PhD2
1
Melanoma Institute Australia, North Sydney, NSW, Australia 2 Kolling Institute, The University of Sydney, St Leonards, NSW, Australia
Intralesional immunotherapy of melanoma has two complementary aims. One is to cause regression of the injected metastasis. The other is to incite or modulate systemic immune responses in such a way that non‐injected metastases will also undergo regression. A number of phase 1 and phase II studies with cytokines, viral, or bacterial agents have been conducted but their use has remained sporadic and has not progressed to become established treatments. Two treatments have progressed to randomized phase III studies. The most promising of these is based on intralesional injection of a genetically modified herpes simplex virus (HSV) (T‐Vec). Initial results have shown a significant effect on durable response rates (DRR) but effects on overall survival remain under study. The second involved injection of plasmids coding for the HLA B7 antigen (Allovectin). Despite encouraging early results the treatment did not reach its endpoints and its use has been discontinued. A phase II study involving intralesional injection of oncolytic A21 coxsackie virus (Cavatak) is also under way and is showing promise.
J. Surg. Oncol. 2014;109:320–326. ß 2013 Wiley Periodicals, Inc.
KEY WORDS: melanoma; intralesional therapy; immune responses; oncolytic viruses; GM‐CSF
INTRODUCTION The publication in 2010 [1] and 2011 [2] of reports of improved survival of patients with metastatic melanoma who were treated with the anti‐CTLA4 monoclonal antibody ipilimumab, a checkpoint inhibitor, ushered in what may be referred to as the post‐DTIC era of melanoma treatment. Prior to this DTIC had proven superior to a number of other treatments but nevertheless was relatively ineffective in treatment of the disease [3]. As a consequence, initiatives based on intralesional injection of accessible tumors in patients with locoregionally advanced metastatic stage IIIB/C or stage IV M1a melanoma were the subject of a number studies exploring intralesional immunotherapy with a range of different agents. Initial studies were largely aimed at local ablation of the lesions with agents such as vaccinia virus [4,5], thiotepa or typhoid vaccines [6]. More recently there has been a greater focus on the systemic effects of intralesional immunotherapy and trial designs have evolved that include assessment of both the local effects on the injected lesions as well as non‐ injected lesions including those in visceral sites. In the following sections we review the earlier studies on intralesional treatments with BCG and cytokines and then the more recent attempts to introduce these treatment approaches into management of melanoma. In particular we discuss their role as management transitions from the DTIC era to the era of checkpoint inhibitors and BRAF/MEK inhibitors.
HISTORICALLY IMPORTANT INTRALESIONAL THERAPY STUDIES Intralesional Bacillus Calmette‐Guerin (BCG) During the 1960 and 1970 period there was much interest in the use of intralesional injection of BCG to treat cutaneous melanoma [7]. One of the strongest advocates was Dr Donald Morton at the John Wayne Cancer Institute who in 1974 reported a 7‐year experience with this approach in 151 patients. Injection of intradermal and cutaneous melanoma resulted in regression of 90% of the injected lesions and 17% of uninjected lesions. Twenty‐five percent of patients remained free of disease for periods ranging from 1 to 6 years. Injections into subcutaneous or visceral lesions were not so effective [8]. Studies on 27 patients with lower extremity recurrences reported complete or partial disease control in 20 of the patients. Responses were associated
ß 2013 Wiley Periodicals, Inc.
with good immune function assessed by skin test responses to dinitrochlorobenzene and PPD [9]. Similar results were reported in studies on 15 patients with metastatic melanoma by Mastrangelo et al. [10]. There were five responses that were limited to patients with dermal metastases. Particular attention was focused on the possible systemic effects of BCG by reports of regression of pulmonary metastases during intralesional injection of BCG [11]. Nevertheless there were reports of systemic infections with BCG [12] and some allergic reactions. Trials of BCG therapy were also negative leading to this form of therapy falling out of favor [13].
Intralesional Cytokine Injections With GM‐CSF or Interferons One variable that may limit the potency of the host response against melanoma is the mixture of cytokines present in the tumor microenvironment. Granulocyte‐macrophage colony stimulating factor (GM‐CSF) was identified as the most potent molecule for augmenting tumor immunity following gene transfer into melanoma cells. Vaccination with irradiated melanoma cells engineered to secrete GM‐CSF enhanced host responses through improved tumor antigen presentation by recruited dendritic cells and macrophages [14]. Injection of 15–50 mg doses of GM‐CSF into two s.c. metastases in 13 patients resulted in regressions of injected or non‐injected lesions in 3 of the patients. Regressions were associated with infiltration by CD4 T cells and dendritic cells [15]. Intralesional injection of interferon (IFN) alpha or beta also attracted attention. Injection of IFN alpha into s.c. metastases in 51 patients was carried out by von Wussow et al. [16]. Half the patients received IFN alpha 2b and the other half natural IFN given three times per week. There
*Correspondence to: Peter Hersey, FRACP, D.Phil, Kolling Institute, The University of Sydney, Pacific Hwy, St Leonards, NSW 2065, Australia. Fax: 612‐9926‐4034. E‐mail: [email protected] Received 30 July 2013; Accepted 26 September 2013 DOI 10.1002/jso.23494 Published online 3 December 2013 in Wiley Online Library (wileyonlinelibrary.com).
Therapies for Melanoma were systemic responses in 9 patients in non‐injected lesions and 24 patients had responses in the injected lesions. Exploratory studies with intratumoral injection of interferon beta concluded that higher doses were probably needed to induce regression of injected lesions [17].
CURRENT INTRALESIONAL IMMUNOTHERAPY STUDIES Patients Being Studied, Treatment Procedures and Evaluations Most of the patients entered into these studies had locally recurrent disease proximal to the primary and in‐transit disease either without LN metastases (IIIb) or with LN metastases (stage IIIc) or Stage IV disease with injectable s.c. or LN metastases. In general the lesions were injected either visually or with ultrasound control using a 25–30 gauge needle and a fanning technique by which the needle is moved back and reinserted in different directions within the lesion. The volume was usually adjusted on the basis of the surface area of the lesion and total volumes set to avoid possible upper limits of toxicity. Tumor assessments were commonly carried out on a weekly basis by ultrasound in the first 4 weeks and then at 6‐week intervals by CT scans. There is interest in the use of PET scans to assess early events in the injected lesions [18,19]. Immune response criteria have been used in most of the studies most commonly at the 12‐week period. These criteria allow the inclusion of new lesions in assessment of responses and are added to the total tumor burden in assessment of CR, PR, and stable disease status [20]. Comparisons of biopsies taken prior to and at intervals during therapy were included when possible in most of the studies.
Velimogene (Allovectin) Allovectin (velimogene aliplasmid) consisted of a bicistronic plasmid encoding a major histocompatibility complex (MHC) class I heavy and light chains of HLA‐B7 and b2‐microglobulin formulated with a cationic lipid‐based system for direct intratumoral administration. Following injection into a single lesion it was intended to induce anti‐
Fig. 1. Diagrammatic illustration of possible mechanisms of action of intralesional therapy with Allovectin. Journal of Surgical Oncology
321
TABLE I. Phase II Clinical Trial Results for Velimogene Aliplasmid (VCL‐1005‐208) Response in 127 patients Overall response rate (CR þ PR) (95% CI) Duration of response (K‐M median) (95% CI) Median time to response 2 cycles (95% CI) CR PR Stable disease (SD) Progressive disease (PD)
n (%) 15 (11.8%) (6.2–17.4%) 13.8 months (6–66 months) (4 months) (2–4 months) 4 (3.2%) 11 (8.7%) 32 (25%) 80 (63%)
tumor immune responses against both treated and untreated distal lesions. Both the plasmid and lipid components of Allovectin were considered to activate of leukocytes via toll receptors. Other actions included the induction of both cytotoxic T‐cell and innate immune responses against allogeneic as well as tumor‐derived targets due to the increased MHC class I expression on tumor cells and the induction of a localized immune/inflammatory response [21]. It was considered that immune responses against the allo‐antigen would attract T cells and macrophages that would recognize and destroy the injected melanoma and non‐injected lesions [22] (Fig. 1). In the phase 2 study based on 133 patients, injections were given in single or multiple lesions and responses assessed by RECIST (Response Evaluation Criteria in Solid Tumors). In 127 evaluable patients the response rate was 11.8% and median duration of response 13.8 months. In 42 patients with stage IV melanoma there were responses in non‐ injected lesions in 21% of the patients [23] (Table I). In view of these results a phase III trial was conducted in 375 patients comparing Velimogene Aliplasmid with physician’s choice of chemotherapy (DTIC or temozolomide) in a 2:1 randomization schema. The primary outcome measure was percent DRR at 24 weeks after randomization. The trial is closed to accrual and results are expected in the latter half of 2013. Much speculation occurred as to what the median OS maybe in the phase 3 trial. In the phase 2 trial
Fig. 2. Illustration of the genetic engineering carried out on the modified herpes simplex virus and the general concepts involved in treatment with oncolytic viruses.
322
Hersey and Gallagher
Fig. 3. Outline of the OPTiM study on treatment with the modified HSV (T‐Vec).
TABLE II. Durable and Objective Responses in the OPTiM Phase III Study of Talimogene Laherparepvec (T‐Vec) ITT set DRR Objective OR (95% CI)
GM‐CSF (n ¼ 141)
Talimogene laherparepvec (n ¼ 295)
Unadjusted OR
2.1% 5.7% (1.9, 9.5)
16.3% 26.4% (21.4, 31.5)
8.9, 95% CI, 2.7–29.2, P < 0.0001 56 of 78 responders in the talimogene laherparepvec arm (72%) and 4 of 8 responders in the GMCSF arm (50%) were still in response at the last available tumor assessment
0.7% 5.0%
10.8% 15.6%
CR PR
78 patients only had 1 injection due to strict adherence to RECIST criteria. Survival in this group had been used as a proxy for the control group in the phase 3 trial. Despite this speculation a press release in august 2013 revealed that the trial had not met its endpoints in increasing percent DDR or overall survival and that the development of this treatment had been abandoned (finance.yahoo.com/news/vical‐phase‐ 3‐trial‐allovectin‐100000138.html).
Oncolytic Viruses Oncolytic viruses are those that preferentially infect and destroy cancer cells. This occurs due to direct effects of the virus on the metabolic processes in the cell but also by inducing immune responses that target the cancer cell. The latter is believed to be aided by activation of NF‐kB and release of chemokines and cytokines from the cancer cell [24,25]. Genetic engineering of certain viruses has attempted to reduce their virulence for the host and in some cases their ability to incite immune responses. Attempts to infect cancer cells by systemic administration of the virus continue but are limited by immune responses of the host. In view of this most studies have used the intralesional route and rely on immune responses generated by release of cell products and inflammatory responses generated against the infected cell [26].
presentation. In addition, ICP47 deletion increases US11 expression and thereby enhances virus growth and replication in tumor cells. The coding sequence for ICP34.5 was replaced by the coding sequence for human GM‐CSF to enhance the immune response to tumor antigens released after virus replication [28,29] (Fig. 2). Phase II studies were carried out in 50 patients with either stage IIIc (10 patients) or stage IV (40) melanoma patients. Treatment involved intratumoral injection of up to 4 ml of 106 pfu/ml of JS1/34.5‐/47‐/GM‐ CSF followed 3 weeks later by up to 4 ml of 108 pfu/ml every 2 weeks for P-value*
All ITT Stage IIIB/C Stage IVM1a
T‐Vec was previously referred to as OncoVEX GM‐CSF when it was owned by BioVex (Woburn, MA). It is an immune‐enhanced, oncolytic herpes simplex virus type 1 (HSV‐1) that is deleted for the neurovirulence factor ICP34.5 [27] and this results in tumor‐selective replication. It is also deleted for ICP47, which normally blocks antigen Journal of Surgical Oncology
HSV-1 Negative HSV-1 Positive
18%
***
12% 3%
Sex- Men
**
10%
3% 0%
**
8% 24%
4%
