|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 30 (October 20), 2007: pp. 4858-4860 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.0534
Prognostic Value of P-Glycoprotein in High-Grade OsteosarcomaLaboratorio di Ricerca Oncologica, Istituti Ortopedici Rizzoli, Bologna, Italy
Servizio di Chemioterapia, Istituti Ortopedici Rizzoli, Bologna, Italy To the Editor: In their article, Schwartz et al1 focus on the immunohistochemical evaluation of P-glycoprotein in a subgroup of high-grade osteosarcoma (OS) patients treated with the INT0133 protocol. The authors' main conclusion was that immunohistochemical evaluation of P-glycoprotein at diagnosis cannot predict the outcome of high-grade OS patients. From our point of view, the authors should be cautious in giving this conclusion on the basis of their study. The first major concerns are about the method used for immunohistochemistry. We would like to underline that in 1996 a consensus meeting was organized, from which emerged clear guidelines for immunohistochemical detection of P-glycoprotein in human tumor tissue samples.2 First of all, the trypsin and microwave pretreatment methods for antigen retrieval that were used in this study were not recommended for P-glycoprotein immunodetection, unless in the presence of a severe overfixation.2 Moreover, two of these guidelines stressed the importance of having tissue antigenicity controls to carefully evaluate fixation and antigen preservation, as well as adequate standardized drug resistant cell lines with different known levels of P-glycoprotein expression to calibrate any detection method used for subsequent analyses on clinical samples.2 These guidelines are very important and should be carefully followed when evaluating the clinical impact of P-glycoprotein expression by immunohistochemistry, especially with procedures that may alter sample antigenicity, like trypsin or microwave pretreatments. It is well known that the use of immunohistochemistry to detect P-glycoprotein in tumor specimens poses specific problems. Above all, it must be considered that rather than evaluating whether a sample contains or lacks P-glycoprotein, as is commonly done with diagnostic markers, the level of P-glycoprotein must also be quantified. In order to do this, a standardized and validated method with several control steps must be used. An adequate panel of drug resistant cell lines with different levels of P-glycoprotein expression is essential to calibrate the immunohistochemical method and antibody concentrations. This is mandatory to score as positive only those cases with P-glycoprotein expression correlated with drug resistance levels comparable to those that can be found in clinical settings. Furthermore, reference controls represented by standardized drug-sensitive and drug-resistant cell lines derived from the same tumor type, have proven to reflect the both physiological and pathological level of tumor-specific P-glycoprotein expression.2 Tissue antigenicity controls are absolutely necessary, in particular for OS tissue specimens, which are frequently decalcified and, consequently, may be affected by a possible protein degradation and loss of antigenicity. For example, for this purpose in our studies,3-5 we always used a section of each tumor specimen for immunodetection of vimentin with a monoclonal antibody specific for the epitope recognized by clone V9, which is very sensitive to protein degradation that may be induced by fixation or other treatments like decalcification or antigen retrieval procedures.6 As presented in Table 1, on the basis of this antigenicity control, we always excluded some V9-negative samples from our analyses, since their inclusion may have led to misinterpretation of the final results.
Unfortunately, none of the controls and standardization steps mentioned herein appear to have been evaluated or performed by Schwartz et al1, and from our point of view, this fact severely reduces the strength of their conclusions. In addition, it must be considered that microwave treatment may be, by itself, a potential source of artifacts, leading to a misidentification of P-glycoprotein at tissue level, as reported for other proteins.7 In particular, it is well known that all procedures of antigen retrieval, mostly those in which tissue samples are heated above 60°C, can cause artifacts leading to unexpected increase in sensitivity of immunoreactivity, nonspecific cross-reactions of several primary antibodies or epitopes loss.8 All of these possible problems may be even further enhanced by the fact that the authors preceded microwave treatment by a step of trypsin digestion, which can also contribute to create artifacts or generate uncontrolled and unpredictable proteolysis or protein conformation changes. Since the authors did not report the percentage of P-glycoprotein–positive cases, it is difficult to understand whether these pretreatments may have affected their immunohistochemical assessment, leading to different results compared with other similar analyses, in which microwave and protease treatments were not employed. Additional concerns about the conclusions of Schwartz et al are related to the number of patients who were analyzed for P-glycoprotein. The authors report that only 139 (20.3%; for C494 antibody) of 685 patients treated with the INT0133 protocol were included in the immunohistochemical study. Although the authors provide a Table in which they compare the clinical characteristics and outcome of patients either included or excluded from this study, this does not necessarily mean that also the outcome of such a small fraction of patients is representative for the entire group. In other words, it would be important to report any clinicopathologic feature correlated with clinical outcome in this small group of patients and verify whether it shows the same trend or evidence in the whole series of 685 patients. The INT0133 protocol de facto enrolled patients in four different arms. Data are once again not fully shown but, in case of a uniform distribution, less than 35 patients per arm should have been considered in the present analysis. With such a small number of patients, the goal of statistical significance is difficult to achieve. It is interesting to note that a different relative risk is reported in patients assigned to receive or not muramyl tripeptide phosphatidylethanolamine, regardless the type of chemotherapy treatment, while prognostic advantage was reported only for the arm treated with ifosfamide plus muramyl tripeptide phosphatidylethanolamine. How is this difference explained? Are the characteristics of the two groups different one from the other? Certainly, reporting the data in terms of event-free survival probability should have been more adequate and more informative. In localized OS, when a prognostic factor is evaluated, chemotherapy treatment should be taken into high consideration which is, besides surgery, the main factor influencing prognosis. This is a pooled evaluation, including four different treatment protocols, and separate analyses should be reported to clearly evaluate a possible role of P-glycoprotein in OS. Moreover, we would like to underline that the reference to the prospective analysis of Wunder et al9 is not appropriate. In fact, this important study was based on the evaluation of MDR1 gene expression and not on the assessment of P-glycoprotein protein level, and it has to be considered that the presence of MDR1 gene overexpression does not necessarily mean that the level of P-glycoprotein is also increased. Therefore, the results obtained in this relevant study are only partly or even not comparable to those in which immunohistochemistry was used. We also strongly disagree with the authors' statement that one the advantages of their prospective analysis was that "no bias in selection for patients with good or adverse outcome would have occurred using this approach."1 This statement appears to mean that the published retrospective studies have a bias in that patients were selected on the basis of their outcome and, therefore, the authors of these numerous reports have actually declared something untrue by stating that analyses were done blinded of patients' outcome. Finally, the authors state that "their study is unique in that it was prospective."1 We would like to underline that this is not true. In fact, our last study on P-glycoprotein evaluation in high-grade OS, which was published in December 20065, was also a prospective analysis, as clearly explained. However, we understand that the authors might not have been aware of this, since they submitted their manuscript in June 2006. AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest.
REFERENCES
1. Schwartz CL, Gorlick R, Teot L, et al: Multiple drug resistance in osteogenic sarcoma: INT0133 from the Children's Oncology Group. J Clin Oncol 25:2057-2062, 2007 2. Beck WT, Grogan TM, Willman CL, et al: Methods to detect P-glycoprotein-associated multidrug resistance in patients' tumors: Consensus recommendations. Cancer Res 56:3010-3020, 1996 3. Baldini N, Scotlandi K, Barbanti-Brodano G, et al: Expression of P-glycoprotein in high-grade osteosarcomas in relation to clinical outcome. N Engl J Med 333:1380-1385, 1995 4. Serra M, Scotlandi K, Reverter-Branchat G, et al: Value of P-glycoprotein and clinicopathologic factors as the basis for new treatment strategies in high-grade osteosarcoma of the extremities. J Clin Oncol 21:536-542, 2003 5. Serra M, Pasello M, Manara MC, et al: May P-glycoprotein status be used to stratify high-grade osteosarcoma patients? Results from the Italian/Scandinavian Sarcoma Group 1 treatment protocol. Int J Oncol 29:1459-1468, 2006[Medline] 6. Azumi N, Battifora H: The distribution of vimentin and keratin in epithelial and nonepithelial neoplasms: A comprehensive immunohistochemical study on formalin- and alcohol-fixed tumors. Am J Clin Pathol 88:286-296, 1987[Medline] 7. Shi SR, Cote RJ, Taylor CR: Antigen retrieval techniques: Current perspectives. J Histochem Cytochem 49:931-937, 2001 8. Shi SR, Imam A, Young L, et al: Antigen retrieval immunohistochemistry under the influence of pH using monoclonal antibodies. J Histochem Cytochem 43:193-201, 1995[Abstract] 9. Wunder JS, Bull SB, Aneliunas V, et al: MDR1 gene expression and outcome in osteosarcoma: A prospective, multicenter study. J Clin Oncol 18:2685-2694, 2000 Related Reply
Related Article
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
