|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2008.19.5578 on November 17 2008 © 2008 American Society of Clinical Oncology.
In Reply:
Vanderbilt University Medical Center and Vanderbilt-Ingram Cancer Center, Nashville, TN Nehlsen-Cannarella voices a number of important and scientifically valid concerns in her commentary on our report published in the May 10, 2008, issue of Journal of Clinical Oncology.1 Patients with melanoma enrolled onto these trials were typed by either serologic or molecular methods to define their HLA-A*02 allele type. As Nehlsen-Cannarella correctly states, the serologic typing of HLA-A*02 does not allow one to determine which of the more than 100 subtypes are present, and the HLA-A*0201 allele is the subtype that has been established as an effective binding molecule for the gp100 (210M) peptide. It would have been a scientifically cleaner study if we had required molecular typing for HLA-A*0201 on each and every patient. To save money, we did not select this requirement for typing, at the cost of good science, and missed the opportunity to provide valid conclusions. At the time the study was initially conceived and written in 1997 to 1998, we were concerned about all participating clinical centers having access to molecular typing. In addition, we encouraged all centers to have patients undergo molecular typing for HLA-A*0201 whenever available. In fact, the overwhelming number of patients did undergo molecular typing. Of the 131 patients, more than 100 underwent molecular typing, including nearly all of the patients who were accrued during the final 2 years of enrollment. In addition, before the study was initiated, we were aware that the frequency of white North Americans, who reflected more than 98% of patients with melanoma enrolled onto the trial, having an HLA-A*02 allele other than HLA-A*0201 is less than 5%.2 Therefore, it is likely that even if all 131 patients had been serotyped, only approximately seven patients may have had an HLA-A*02 allele other than *0201. In another study of patients without melanoma, of those with HLA-A*02 alleles, 96% of whites and 94% of Native Americans were *0201, whereas only 59% of African Americans and 53% of Asian/Pacific Islanders were *0201.3 However, as stated above, 129 of 131 enrolled patients with melanoma were white North Americans. Rivoltini et al.4 did demonstrate that Italian melanoma patients had a greater frequency of non-HLA-A*0201 subtypes of HLA-A*02, with up to 15% or greater of non-HLA-A*0201 subtypes. Some of the non-HLA-A*0201 subtypes poorly bound the gp100 and MART-1 melanoma peptides. In this population, the case for subtyping HLA-A*02 patients would be much stronger. Finally, Nehlsen-Cannarella mentions another article published on a completely separate group of patients with melanoma receiving a different vaccine who had been HLA class I typed for a different reason. Among stage II melanoma patients receiving a melanoma lysate vaccine (melaccine) or under observation, we (Southwest Oncology Group) did serologically type patients for HLA class I antigens.5 This was simply an exploratory analysis, and finding a better outcome in vaccinated patients who had HLA-A2/Cw3 generated a hypothesis to be tested prospectively. We had hoped to complete a study molecularly typing all patients prospectively. In summary, we do agree that any HLA allele-restricted vaccine trial would be improved and strengthened by molecular typing, especially for HLA-A*02-based trials vaccinating patients with immunogenic peptides. However, we strongly disagree that by allowing some of our patients to be serotyped, we may have invalidated our findings and posed an ethical concern for our patients. It is likely that of the 131 patients enrolled, only 30 or so patients were serotyped, and assuming a 95% frequency of HLA-A*0201, then one to two patients may have been HLA-A*02 and not HLA-A*0201. Obviously, these one to two patients would not have changed the results for the 131 patients, no matter what treatment they received. AUTHOR's DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. ACKNOWLEDGMENTS Written for the Cytokine Working Group. Supported by the Cancer Therapy Evaluation Program at the National Cancer Institute, a grant from Chiron Pharmaceuticals (Emeryville, CA), and K24 Grant No. 5K24-CA097588 (J.A.S.). NOTES published online ahead of print at www.jco.org on November 17, 2008. REFERENCES
1. Sosman JA, Carrillo C, Urba, et al: Three phase II cytokine working group trials of gp100 (210M) peptide plus high-dose interleukin-2 in patients with HLA-A2-positive advanced melanoma. J Clin Oncol 26:2292-2298, 2008 2. Player MA, Barracchini KC, Simonis TB, et al: Differences in frequency distribution of HLA-A2 subtypes between North American and Italian white melanoma patients: Relevance for epitope specific vaccination protocols. J Immunother Emphasis Tumor Immunol 19:357-363, 1996[Medline] 3. Ellis JM, Henson V, Slack R, et al: Frequencies of HLA-A2 alleles in five U.S. population groups: Predominance of A*02011 and identification of HLA-A*0231. Hum Immunol 61:334-340, 2000[CrossRef][Medline] 4. Rivoltini L, Loftus DJ, Barracchini K, et al: Binding and presentation of peptides derived from melanoma antigens MART-1 and glycoprotein-100 by HLA-A2 subtypes: Implications for peptide-based immunotherapy. J Immunol 156:3882-3891, 1996[Abstract] 5. Sosman JA, Unger JM, Liu PY, et al: Adjuvant immunotherapy of resected, intermediate-thickness, node-negative melanoma with an allogeneic tumor vaccine: Impact of HLA class I antigen expression on outcome. J Clin Oncol 20:2067-2075, 2002
Related Correspondence
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
