He refused to do any more traditional melanoma therapy. It’s made using a genetically engineered herpes virus — the same kind that causes cold sores — that carries an immune system compound into tumors, causing them to burst apart. In addition, the roles of p53, the Kaposi’s sarcoma-associated herpesvirus viral cyclin and nuclear factor-κB in the development and progression of Kaposi’s sarcoma are being further clarified, and therapeutic agents are being developed that may target these pathogenetic mechanisms. A natural “weapon” against cancers, these viruses have spent decades in the oblivion before emerging only recently as potential game changers in the world of conventional medicine that has suffered for decades from a severe drought in viable, selective cancer therapies. BRAK and MEK inhibitors have a high likelihood of reducing the size of the melanoma, but it is only for a short period of time. Some will not be eligible for systemic therapy. The transfection of the receptors into B78H1 cells did not induce a detectable in vivo immunogenicity to the tumors.
The luciferase expression level in melanoma masses was higher by pEBcLuc than by pcLuc, although Southern blot analysis showed the number of copies of pEBcLuc retained in the melanoma masses to be fewer than that of pcLuc. Complete remission was seen for 87.5% of the LM melanoma xenografts at 5 months after treatment termination. Two phase I trials of single agent intravenous administration of Reolysin® have been previously completed.15,16 Doses up to 3 × 1010 TCID50 days 1–5, of a 4-week cycle were well tolerated without dose-limiting toxicity being observed; this is the dose chosen for our study. Secondary objectives included assessment of progression-free and overall survival in melanoma patients treated with systemic Reolysin®, assessment of viral replication in metastatic melanoma deposits after intravenous administration of Reolysin®, and assessment of the impact of pre-existing anti-reovirus immunity on the efficacy and toxicity of Reolysin® treatment. She has been in remission for almost three years. For this to happen a technique would need to be developed to prevent the body from mounting an immune response to the virus prematurely. Another patient was found to be ineligible because of not meeting minimum size of metastatic lesions as per trial eligibility criteria.
Patient and tumor characteristics at registration of the remaining 21 eligible patients are presented in Table 1. Importantly, this approach avoids the problem of inadequate delivery created by high tumor interstitial pressures that are known to reduce efficacy of intravascular delivery of therapeutic agents.9 However, intratumoral injection is subject to limitations when used to address multiple small tumors, let alone those that are clinically and radiographically unapparent. Similar to other immunotherapy studies, survival rates appeared to ‘plateau’ after 1 year, with 58 and 52% of patients surviving to 1 and 2 years, respectively. Few severe (grade 3–4) treatment related toxicities were reported, with fatigue (9.5%), lymphopenia (9.5%), and hyponatremia (9.5%) being the most common. Nonhematologic and hematologic grade 1–2 toxicities most commonly reported were fatigue (66.7%), nausea (57.1%), fever (52.4%), and anemia (42.9%), respectively (Table 2). The rate of melanoma — the deadliest form of skin cancer — has been increasing in the United States for the past three decades. All adults should be screened for skin cancer once a year – or more frequently if they have certain risk factors.
Median time to progression was 45 days (range 13–96 days) and median survival was 165 days (range 15 days–15.8 months). Figure 1 displays the Kaplan–Meier curves for the distribution of progression-free survival and overall survival times. Adam Comprehensive Melanoma Research Center at the Moffitt Cancer Center in Tampa, Florida, who was not involved in the study. This patient had received two treatment cycles; although her cutaneous metastases remained stable in size, they became clinically necrotic with bleeding and symptomatic anemia which necessitated removal of two of these lesions. Pathology showed 75–90% tumor necrosis in these lesions, consistent with treatment effect (Figure 2). Overall only 6 patients remained on treatment with stable disease for >8 weeks. As such, the trial did not meet the previously defined efficacy rule in order to proceed to the second stage of accrual and closed to further enrollment.
Extensive necrosis of a resected metastatic melanoma lesion in a Reolysin®treated patient. A sheet of viable melanoma cells is seen in the lower right with abrupt transition to the necrotic area at left. This trial in second line has now been analysed and reported at ASCO in 2010 . Fifteen patients had tumor biopsies performed at ~1 week following treatment, timing in which reovirus replication has been detected by plaque assay and electron microscopy in a prior phase I trial.16 Of these biopsies, 13 contained metastatic tumor. Productive reoviral replication (viral antigen coexpression with tubulin) (Figure 3) was detected in two of the 13 tumors. Coexpression of the reovirus with p38 was also demonstrated in these two samples (Figure 3). Of note, both patients had a longer progression-free survival (80 and 87 days, respectively) as compared to the median survival of 45 days in this study, although the small sample size prevents any statistical conclusions.
Expression of reoviral proteins in post-treatment metastatic melanoma biopsies, as demonstrated with the Nuance system. (a) Hematoxylin and eosin (H&E) stain, (b) reovirus stain (fluorescent red), (c) tubulin stain (fluorescent green), (d) cells coexpressing both targets are fluorescent yellow, (e) higher magnification of an area of double positive cells indicative of productive reovirus infection (blue color, hematoxylin counterstain), (f) coexpression of reovirus (fluorescent red) with p38 (fluorescent green) in tumor cells results in fluorescent yellow. The KRAS and BRAF mutation status was examined in 11/13 biopsy patients for whom adequate tumor was available. KRAS was wild type in all 13 patients and BRAF mutations were detected in 2 of 13 patients. There was no association between detection of the V600E mutation and reoviral replication, but the patient sample size and the small number of positive samples prevents definitive conclusions. All 21 patients had baseline serum samples for determination of neutralizing anti-reovirus antibodies (NARA) titers, and 13 of the 21 patients had pre- and post-treatment samples available for evaluation of NARA response. Results per individual patient are given in Figure 4.
All 21 patients had pre-existing anti-reovirus neutralizing antibodies. Like human cyclin D, vCyc overexpression can drive cell proliferation by activating the cyclin-dependent kinases (Cdks), Cdk4 and Cdk6. In all 13 patients the post-treatment NARA titers exceeded 1/2,000, whereas in 9/13 patients the peak endpoint titer reached 1/10,000. The median fold increase from baseline was 247–729 with a range of 81 to 2,187. Electrochemotherapy is widely available in Europe. Increase in reoviral titers was observed in all 13 patients in whom baseline and post-treatment samples were available; titers reached their peak at 1 month following treatment initiation in the majority of patients. We report on the first phase II trial assessing the efficacy of reovirus as a single agent in the treatment of metastatic melanoma.
A modest median progression-free survival of 45 days and a median overall survival of 168 days were observed. Although the trial did not proceed to its second stage, due to not meeting the predetermined efficacy rule, it did convincingly demonstrate productive replication of reovirus in melanoma tissue following intravenous administration, the first such demonstration in this tumor type. Methodologies used in other clinical trials of different tumor types to assess reoviral replication, frequently included techniques such as recovery of live virus from tumor deposits or in situ hybridization which cannot convincingly distinguish between input and progeny virus.17 Reovirus is using the tubulin scaffold of the cell in order to replicate18,19; the reovirus core protein mu 2 determines the filamentous morphology of viral inclusion (“factories”) by interacting with and stabilizing microtubules.19 Coexpression of reovirus antigens and tubulin in biopsy samples from our study patients is therefore indicative of viral replication in these metastatic melanoma tumors. Preclinical experiments in melanoma models also indicate that p38 and MAPK signaling represent key molecular alterations promoting reoviral replication.13 Coexpression of p38 in the reovirus positive biopsy samples supports this hypothesis. Although our data do not support use of reovirus as monotherapy in the treatment of malignant melanoma, they do demonstrate productive reoviral replication in melanoma metastases (2/13 patients), following intravenous administration. Pockets of replication can easily be missed on biopsies either based on location in the tumor or relative timing of the biopsies and replication. Although this result may still underestimate the level and incidence of replication, it serves as a proof-of-principle upon which future treatment regimens incorporating reovirus can be based.
Preclinical data support that combination of reovirus with chemotherapy agents such as paclitaxel and cisplatin can lead into synergistic cytotoxicity by increasing apoptosis in treated tumors.20,21 Furthermore, phase I trials have demonstrated the safety of intravenous administration of reovirus in combination with chemotherapeutic agents such as docetaxel, gemcitabine, and carboplatin/paclitaxel.22,23,24 Based on these preclinical data and the clinical safety data, a phase II trial of reovirus in combination with paclitaxel and carboplatin in patients with metastatic melanoma is ongoing (clinicaltrials.gov NCT00984464). Velimogene aliplasmid is a lipid-based formulation containing a plasmid encoding major histocompatibility complex (MHC) class I and HLA-B7 and B2 microglobulin light chains. Although our study was not designed to incorporate an assessment of antitumor immune response in the context of reoviral treatment, the latter should represent an important correlative endpoint in future reovirus trials in melanoma patients. In summary, this phase II trial, although it did not meet its primary efficacy endpoint, convincingly demonstrated replication of reovirus in melanoma metastases following intravenous administration despite high NARA titers. These data support development of reovirus as a novel agent in treatment of malignant melanoma as part of combinatorial strategies. Eligible patients had to be 18 years of age or older and have histologic or cytologic confirmation of metastatic malignant melanoma with at least one lesion which could be accurately measured in at least one dimension as ≥20 mm with conventional techniques or ≥10 mm with spiral CT. Eligible patients should have failed at least one treatment regimen for metastatic disease.
They were also required to have an Eastern Cooperative Group Performance Score of 0–2; life expectancy of ≥12 weeks; acceptable hematologic function defined as absolute neutrophil count >1,500/µl, platelets ≥100,000/µl, and hemoglobin ≥9 g/dl; adequate hepatic and renal function defined as total bilirubin ≤1.5 × the institutional upper limit of normal, aspartate aminotransferase ≤2.5 × upper limit of normal, and creatinine ≤1.5 × upper limit of normal. Because of the theoretical concern that reovirus treatment might impact on cardiac function, patients were required to have a baseline left ventricular ejection fraction of ≥50% as assessed by echocardiogram or MUGA and normal troponin T levels. Patients were required to be ≥4 weeks from any prior chemotherapy (6 weeks for nitrosourea or mitomycin-C), and ≥2 weeks from radiation therapy, immunotherapy or treatment with small molecule cell cycle inhibitors. Patients with known brain metastases, known HIV positivity, and pregnant and nursing women were excluded. Study treatment. Reolysin® was administered at a dose of 3 × 1010 TCID50 per day in 250 ml 0.9% sodium chloride, infused intravenously over 60 minutes daily on days 1–5 of a 4-week cycle. Toxicity was graded according to the National Cancer Institute Common Terminology Criteria Version 3.0.
Reolysin® was decreased by one dose level in subsequent cycles for interim grade 4 hematologic toxicity, ≥grade 2 cardiac toxicity, and ≥grade 3 other nonhematologic toxicity. The C-KIT and BRAF inhibitors are furthest in their development at the moment. Grade 3 flu-like symptoms of ≤72 hours duration and grade 3 fever ≤72 hours duration did not require dose modification. Statistical analysis. A single arm two-stage phase II trial was designed to test the null hypothesis that the true clinical benefit rate with Reolysin® is at most 50% against the alternative hypothesis of the true benefit rate with Reolysin® at least 70%, with a 90% power at a 0.1 level of significance. Twenty patients were to be enrolled at the first stage of the trial. If at least 11 of these 20 patients remained on treatment for at least 8 weeks with 2 or more patients having a CR or PR, the trial was to proceed to the second accrual stage.
Determination of NARA titers. Patients’ serum for determination of NARA titers was collected at baseline, and before each subsequent treatment cycle (i.e., every 4 weeks if the patient remained on study). Dilutions of patient serum were treated with a constant dose of reovirus known to cause 80% cell death on L929 mouse cells. The serum:virus mix was incubated for 2–3 hours to allow any antibodies in the serum to bind and neutralize the virus, before transferring onto L929 cells. At 72 hours the cell survival was measured by MTT assay. An anti-reovirus rabbit antibody was used as a positive control for this assay. The NARA titer was expressed as the last dilution where any neutralization occurred before cell killing was resumed to that seen in the virus-only treated controls.
This trial established 5 × 106 IU of recombinant IFN-α2b as the maximum tolerated dose in combination with protease inhibitor-based HAART and although Kaposi’s sarcoma regression was observed, the study included only 14 patients and was not designed to estimate the response rate at the maximum tolerated dose nor was the mechanism of Kaposi’s sarcoma regression investigated. The first 15 patients enrolled in the trial had biopsies performed at 7 days ± 1 day after treatment initiation. Three core needle biopsies were obtained under CT or ultrasound guidance. Paraffin sections were examined to assess cytopathic effect, expression of reovirus antigens, tubulin and p38 by immunohistochemistry. Our immunohistochemistry protocol has been previously published.34,35 The Benchmark LT automated system from Ventana Medical Systems (Tucson, AZ) was employed according to the manufacturer’s specifications. In brief, optimal conditions were determined by comparing different antibody concentrations with no pretreatment, to protease pretreatment (Protease 1, 4 minutes) and to cell conditioning. Applying these protocols in reovirus infected cell lines (NH3, HN5, SIHN 5B, PJ41, CAL27, CHL, A375, DO4, SK-Me128, WM1791C, and corresponding uninfected control cells) tested and evaluated in a blinded fashion, it was determined that the optimal concentration conditions for antibodies used in this study were: reovirus (1:6,000, antigen retrieval for 30 minutes, Ab was provided by Dr Matt Coffey); tubulin (1:100, antigen retrieval for 30 minutes; Abcam, Cambridge, MA) and p38 (1:200 antigen, retrieval for 30 minutes; Abcam, Cambridge, MA).
Since the reovirus antibody is goat derived, tissue was incubated with the primary antibody for 20 minutes, followed by incubation with a rabbit/anti-goat antibody diluted at 1:1,000. The antigens were detected with the Ultraview Universal DAB or Fast Red system from Ventana with a counterstain of hematoxylin. The Nuance system (Cambridge Research Institutes) was employed for interpretation of colocalization. This is a microscope/computer based interface, which dissects the colorimetric based signal for the different chromogens, and then converts these color-based signals to fluorescence-based signals. This allows one to readily perform “fluorescence—mixing” combinations to determine if a given cell has neither, one, or two or more signals. KRAS and BRAF mutation analysis. Hematoxylin and eosin stained sections and 10 micron thick unstained sections were obtained from formalin-fixed, paraffin-embedded tissue.
J Clin Oncol. Approximately 1 cm of material or all available tissue was scraped and Proteinase K digested overnight. Extraction was performed using the automated column based QIAamp DNA Mini Kit on the QIA cube extraction system (Qiagen, Valencia, CA). This work was supported in part by Public Health Services grant N01 CM 62205-5, National Cancer Institute. The authors thank Renee Bradshaw for coordinating patient care and Raquel Ostby for her help with manuscript preparation. We also thank Ms. Kathleen Sergott of Ventana Medical System for providing some of the reagents used in assessment of reoviral replication.