4658 Vol. 6, 4658–4662, December 2000 Clinical Cancer Research

Vaccination with a Bivalent GM2 and GD2 Conjugate Vaccine: A Trial Comparing Doses of GD2-Keyhole Limpet Hemocyanin1

Paul B. Chapman,2 Donna Morrisey, INTRODUCTION Katherine S. Panageas, Linda Williams, expressed on melanoma are attractive targets Jonathan J. Lewis, Robert J. Israel, for immunotherapy. They are abundantly expressed, are not W. Bradley Hamilton, and Philip O. Livingston3 down-regulated when bound by antibody, do not require HLA molecule coexpression for the immune system to react to them, Department of Medicine [P. B. C., P. O. L.], Department of Epidemiology and Biostatistics [K. S. P.], Department of Nursing and may play an important role in melanoma-cell adhesion. In [L. W.], and Department of Surgery [J. J. L.], Memorial Sloan- addition, it is clear that patients with melanoma can generate Kettering Cancer Center, New York, New York 10021, and Progenics antibodies against ganglioside antigens (1–3). Among the gan- Pharmaceuticals, Inc., Tarrytown, New York 10591 [D. M., R. J. I., gliosides expressed by human melanomas, G appears to be W. B. H.] M2 the most immunogenic. Immunization with GM2 and BCG ad-

juvant induced anti-GM2 IgM in more than 85% of patients, and ABSTRACT the presence of either vaccine-induced or natural antibodies against G correlated with an improved relapse-free survival Immunization with GMK vaccine (GM2 ganglioside M2 conjugated to keyhole limpet hemocyanin mixed with QS-21 (4). Subsequent studies showed that immunization with GM2 conjugated to KLH4 and mixed with QS-21 (5, 6) adjuvant adjuvant) induces anti-GM2 antibodies in close to 100% of (GMK vaccine) induced anti-G IgM and IgG in close to patients. We found previously that anti-GD2 antibodies M2 could be induced in some patients using GD2-keyhole limpet 100% of patients (3, 7). On the basis of these observations, we -hemocyanin ؉ QS-21 (GDK). In this trial, we wished: (a)to determined that GMK vaccine resulted in optimal immunoge ␮ determine whether immunization with both GMK and GDK nicity at a GM2 dose of 30 g.

vaccines could induce antibodies against both GM2 and GD2; Because not all melanomas express GM2 (8), it seems likely and (b) to determine the optimal dose of GDK. Thirty-one that a maximally effective ganglioside vaccine will need to

patients with melanoma or sarcoma who had no evidence of include multiple gangliosides. GD2 ganglioside is expressed in disease after complete surgical resection were immunized many melanomas and sarcomas, and monoclonal antibodies ␮ with both GMK (30 gofGM2) and GDK on weeks 1, 2, 3, against GD2 can induce antitumor effects in patients with mel- 4, 12, 24, and 36. Patients were assigned to one of five GDK anoma or neuroblastoma. Although antibodies against GD2 can ␮ dose levels (3, 10, 30, 70, or 130 gofGD2). Anti-GM2 IgM be induced by immunizing patients with GD2-KLH plus QS-21 or IgG were induced in 97% of patients. The dose of GDK (9), the optimal GD2 dose has not been determined, and it is not

did not affect the anti-GM2 response, although at the highest known whether a bivalent ganglioside vaccine can induce anti- GDK dose level, 3 of 7 patients did not make anti-GM2 IgG. bodies against both GM2 and GD2. GDK was less immunogenic; overall 45% of patients devel- We carried out a trial using a mixed ganglioside conjugate

oped either IgM or IgG against GD2. At GDK doses of 30 or vaccine consisting of both GM2-KLH and GD2-KLH with the 70 ␮g, 8 of 11 patients (73%) made either IgM or IgG adjuvant QS-21. GM2-KLH was administered at a fixed dose of ␮ anti-GD2 antibodies. We conclude that both GMK and GDK 30 gofGM2/immunization with one of five GD2-KLH doses. vaccines can induce antibodies against GM2 and GD2 in a The primary goals were to assess the anti-GM2 antibody re- majority of patients and are safe. The optimal dose of GDK sponses induced, determine which GD2-KLH doses induced appears to be either 30 or 70 ␮g when administered with anti-GD2 antibodies, and to determine whether this mixed gan- GMK vaccine. glioside vaccine was safe.

MATERIALS AND METHODS Vaccine Preparation Received 6/3/00; revised 9/26/00; accepted 9/26/00. The costs of publication of this article were defrayed in part by the GM2-KLH and GD2-KLH were manufactured by Progenics payment of page charges. This article must therefore be hereby marked Pharmaceuticals, Inc. (Tarrytown, NY) as described previously advertisement in accordance with 18 U.S.C. Section 1734 solely to (3, 10). QS-21 was supplied by Aquila BioPharmaceuticals indicate this fact. (Framingham, MA). 1 This study was supported by National Cancer Institute PO1 CA33049 and by Progenics Pharmaceuticals, Inc. 2 To whom requests for reprints should be addressed, at Memorial Sloan-Kettering Cancer Center; 1275 York Avenue, Room K718, New York, NY 10021. Fax: (212) 794-4352. 3 P. O. L. is a paid consultant and stockholder in Progenics Pharmaceu- 4 The abbreviations used are: KLH, keyhole limpet hemocyanin; ITLC, ticals, Inc. immuno-thin layer chromatography.

Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2000 American Association for Cancer Research. Clinical Cancer Research 4659

Patient Eligibility on week 60. In patients who received GM2-KLH booster vacci- Patients with American Joint Committee on Cancer stage nations in year 2, serum was collected at the time of each

III or IV melanoma 2 weeks to 1 year after complete surgical vaccination and 1 month later. Anti-GM2 and anti-GD2 antibod- resection were eligible for the study. After the first 24 patients ies were measured using an ELISA method in which ganglioside had been entered, the protocol was amended to allow patients is adsorbed to 96-well polystyrene microtiter plates. Remaining with metastatic sarcoma. Patients had to be free of disease, older binding sites on the plate were blocked by PBS/Casein/Tween than 18 years of age, and have a Karnofsky performance status 20 buffer. Serially diluted patient sera or controls were added, Ն80%. Previous radiotherapy, chemotherapy, or immunother- and bound antibody was detected using a goat antihuman IgM or apy were permitted, but patients must have completed treatment IgG antibody (heavy chain-specific) conjugated to alkaline Ͼ4 weeks before starting the protocol. Patients were required to phosphatase. Plates were developed using p-nitrophenyl phos- have WBC Ն3.0 cells/mm3, platelets Ն100,000/mm3, total bil- phate substrate and absorbance read at 405 nm with a correction irubin Յ2.0 mg/dl, aspartate aminotransferase Յ74 U/l, lactate of 620 nm. Antibody titer was defined as the highest dilution of dehydrogenase Յ400 U/l, and alkaline phosphatase within nor- patient serum yielding a corrected absorbance of 0.1. Pooled mal limits. All patients signed written informed consent before human serum from previously vaccinated patients with a known participating in the study. anti-GM2 or anti-GD2 antibody titer or pooled normal human Patients were excluded if they had had another malignancy serum with no reactivity were used as positive and negative within the past 5 years (other than basal cell, squamous carci- controls, respectively. nomas of the skin, or cervical carcinoma in situ), or if they had Dot Blot Immunoassay. The specificity of the ganglio- a medical condition which might make it difficult to complete side antibody response was analyzed by dot blot immune stain- the full course of treatments or to respond immunologically. ing. Ganglioside standards GM2,GM3, and GD2 (Sigma Chem- ical Co., St. Louis, MO) were applied to polyvinylidene Treatment Plan difluoride membranes using a dot blot apparatus (Bio-Rad Lab- Before treatment, all patients had a chest X-ray or chest oratories, Inc., Hercules, CA). Nonspecific binding to the mem- computed tomography showing no evidence of disease. Addi- brane was blocked with PBS/casein/Tween 20 buffer. Immune tional radiographic tests were perfor