© 2006 European Society for Medical Oncology
editorial |
Transcriptional profiling of tumor biopsies in oncology trials—a ‘window’ of opportunity for evaluating new drugs in nasopharyngeal cancer?
Department of Clinical Oncology, Sir Y. K. Pao Center for Cancer, Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, New Territories, Hong Kong SAR, China
*(E-mail: anthony{at}clo.cuhk.edu.hk)
Historically, microarray-based expression profiling of tumor tissues has been used to identify molecular signatures that can promote the precise classification and prognostication of human cancers. There is a recent move toward applying this tool in the development of targeted drugs in oncology, where transcriptional profiling of tumor samples are performed for the purpose of elucidating the mechanisms of drug action and identifying biomakers of drug response [1]. In this issue of the Annals of Oncology, Dr Soo et al. [2] report an interesting study on the pharmacodynamic evaluation of celecoxib that focused on whether it could alter global gene expression, microvessel density (MVD) and proliferation index (Ki-67) in nasopharyngeal cancer (NPC). They designed a window study where subjects with untreated NPC were given 14 days of celecoxib before undergoing curative radiotherapy. Endoscopic biopsies of primary tumor were obtained at baseline and on the 14th day of taking celecoxib, while similarly scheduled biopsies were taken from other subjects with NPC who did not receive celecoxib (the ‘control’ group). The authors reported a statistically significant reduction in MVD in the samples taken from the celecoxib-treated group but not the control group. In eight out of the 15 paired samples obtained from the celecoxib-treated group, the authors reported an up-regulation of genes involved in cell cycle, metabolism, signaling and transcriptional regulation, whereas genes involved in immune response were down-regulated. No statistically significant change in Ki-67 following treatment with celecoxib was observed. The authors concluded by indicating that these results support the clinical evaluation of celecoxib in the treatment of NPC.
Pharmacodynamic evaluation of targeted drugs using tumor biopsies obtained before and after drug exposure is usually performed in oncology trials to confirm that drug-induced biological effects achievable in preclinical models can also be observed in vivo. These effects may include modulation of activity of the intended drug targets and of important biological processes such as apoptosis, proliferation and angiogenesis. In this study, the authors set themselves a difficult task of trying to support the case for clinical development of celecoxib in NPC—a drug without proven efficacy in established cancers to date—using pharmacodynamic end points. Their efforts should be credited as this study illustrates the practical challenges of this approach, which are the subjects for discussion in this editorial review.
The application of microarray-based profiling of global gene expression allows measurement of pharmacodynamic responses to drugs in oncology trials using only minute amount of tumor samples [1]. However, one of the drawbacks of this method is the molecular heterogeneity within and across tumor samples, and therefore the analytical precision of microarray analysis can be affected by how the samples are obtained and processed. In this study, the authors attributed changes in the expression levels of certain genes to the effect of celecoxib on cancer cells. This interpretation, however, is hampered by the uncertainty of whether microdissection was performed on the tumor samples before array analysis. Tumor-derived array ‘signals’ can be difficult to distinguish from ‘noise’ due to contamination from stromal and other cell types, thereby increasing the risk of creating false leads and erroneous data interpretation [3–5]. This is particularly relevant to NPC as it is characteristically associated with intense inflammatory infiltrates [6]. Furthermore, endoscopic biopsies of NPC often contain different populations of cells ranging from necrotic cells, normal epithelium, dysplastic epithelium to carcinoma, where each group has been shown to exhibit differing patterns of gene expression [3, 4]. This problem is compounded by the fact that endoscopic biopsies of NPC are often small in size, as indicated in this study where only 15 out of the 25 paired samples contained at least 50% of NPC components histologically. Therefore, it would be helpful if the authors could elaborate on the proportion of other cell types present in each sample [3, 4]. For instance, if each sample contained at least over 80%–90% of NPC components, then, one might argue microdissection was not essential for this study.
Another potential drawback of microarray is a decreased sensitivity in the detection of genes with low basal expression level (low-abundance genes). In previously published studies on the genome-wide transcriptional profiling of NPC that used microdissected specimens [3, 4], genes involved in apoptosis and cell structure were notably underexpressed, whereas genes associated with cell cycle aberrations, signal transduction and metastatic potential were highly expressed in tumor cells compared with normal cells. This trend is reflected in this study where genes related to apoptosis and angiogenesis (e.g. vascular endothelial growth factor receptor) were undetected or underexpressed at baseline. Thus, any possible effect of celecoxib on the expression of these ‘low-abundance’ genes might be beyond the detection limit of DNA microarray. Another issue is the reported magnitude of change in the gene expression level of the 35 gene transcripts. A fold change in the order of 0.7 to 1.9 seemed modest. Furthermore, the gene that displayed the greatest change (1.9-fold) was related to immune response (e.g. major histocompatibility complex, class II, DM beta) rather than cell cycle or transcription, as one would expect from the known preclinical activity of celecoxib in other cancers [7]. To illustrate this point further, Easdale et al. [8] reported in abstract form a study on the expression profiling of colon cancer cells before and after exposure to celecoxib, which found that celecoxib could alter the expression level of genes associated with the control of cell cycle, growth arrest, apoptosis and signal transduction.
Overall, it is difficult to be sure if the differential gene expression reported in this study is clinically significant or attributable to the effect of celecoxib on NPC cells or on other cell types, or whether the effect is due to celecoxib or other factors such as the process of biopsy itself. It would be interesting to see if these changes can be seen in the paired biopsies obtained from the five ‘control’ subjects who did not take celecoxib, however, it is unclear if microarray was performed in that group.
Another issue raised by this study is the use of window trials in drug development. Traditionally, window designs have been used in phase II oncology trials where a new anticancer drug is administered to previously untreated patients during a 4- to 8-week window period before starting standard therapy [9]. This provides a chance to evaluate the drug in patients who are often medically fitter and less likely to carry multiple-drug-resistant tumors. The study end point is usually clinical response and hence the drug is given with the hope that it may show activity against the disease. The adoption of window designs with the primary intent of evaluating the pharmacodynamic effect of a targeted drug is still a relatively novel concept. One potential advantage of this approach is that it provides an opportunity of validating the preclinical effects of a drug in vivo, free of any influence before chemotherapy or radiotherapy that may modify the molecular characteristics of tumors. However, some researchers and patients may be uncomfortable with the idea of delaying potentially curative treatment to patients with cancers that tend to run a progressive course (such as NPC), for the purpose of evaluating a drug with no proven activity in established cancers. This study by Soo et al. [2] was conducted just before the time when an association between celecoxib and cardiotoxicity became substantiated [10], therefore highlighting the unforeseen risks associated with the evaluation of new drugs using window studies. A third possible concern with window design is the need for repeat biopsies and the resultant limitation on overall sample size due to patient refusal, which in turn may undermine the statistical power of the study. However, the rate of patient refusal reported by Soo et al. [2] was relatively low where only nine out of 34 subjects (26%) refused repeat biopsies. Other investigators have also reported the feasibility of repeat biopsies for the pharmacodynamic evaluation of cytotoxic chemotherapy in breast and rectal cancer, and of targeted agents in acute leukemia [1, 11].
The anticancer properties of cyclooxygenase-2 (COX-2) inhibitors in the preventive setting have long been recognized before its mechanisms of action were understood. It is now known that celecoxib is more than just a COX-2 inhibitor. It is an apoptosis-inducing agent [12] that can also cause cell cycle arrest [13], and is an inhibitor of cell signaling [14] and angiogenesis [15] via COX-2 dependent and independent mechanisms in preclinical models of cancer. What puzzled investigators, however, was that the concentrations of celecoxib required to achieve most of these effects in vitro were quite high (ranging from 10 to over 100 µM) [7] for many cancer types, including NPC cells [16, 17]. This cast doubts over whether celecoxib could achieve these effects in vivo, until it was later found that the concentration of celecoxib required to induce apoptosis in animal models is actually many folds lower than that in vitro [7]. This study by Soo et al. [2] and another window study by Ferrandina et al. [18] (in cervical cancer) tried to address the question of whether celecoxib could exert similar preclinical anticancer effects in humans who took therapeutic doses of celecoxib. Like Ferrandina et al. [18], Soo et al. [2] used MVD as a surrogate marker of angiogenesis and reported a drop in MVD following celecoxib in NPC. However, the relatively small number of tumors analyzed in this study rendered the result less conclusive. Also, whether such a fall in MVD could be interpreted as an indicator of celecoxib's potential as a new therapy for NPC remains speculative, especially when there is little evidence to support a prognostic role of MVD in NPC [19]. As the authors have mentioned, the lack of changes in Ki-67 and apoptotic indices observed in NPC with short-term exposures to celecoxib indicate that longer exposure maybe required. This is supported by evidence in vitro where apoptosis could be induced at much lower concentrations of celecoxib following longer duration of drug exposure [20].
In conclusion, this study illustrates how the coupling of window trial design and microarray expression profiling of pre- and post-treatment tumor biopsies may provide a ‘window of opportunity’ for facilitating new drug development in oncology. However, this window is by no means unclouded and readily accessible because of its expense and other practical impediments as discussed in this review. With relevance to NPC research, this study's result does not seem to provide fertile grounds for the development of celecoxib in the treatment of NPC, especially in light of its association with cardiac toxicity and the lack of activity in clinical trials observed in other cancer types to date.
References
1. Clarke P, Poele R, Workman P. (2004) Gene expression microarray technologies in the development of new therapeutic agents. Eur J Cancer 40:2560–2591.[CrossRef][ISI][Medline]
2. Soo RA, Wu J, Aggarwal A, et al. (2006) Celecoxib reduces microvessel density in patients treated with nasopharyngeal carcinoma and induces changes in gene expression. Ann Oncol 17:1625–1630.
3. Sriuranpong V, Mutirangura A, Gillespie JW, et al. (2004) Global gene expression profile of nasopharyngeal carcinoma by laser capture microdissection and complementary DNA microarrays. Clin Cancer Res 10:4944–4958.
4. Zeng Z, Zhou Y, Xiong W, et al. (2006) Analysis of gene expression identifies candidate molecular markers in nasopharyngeal carcinoma using microdissection and cDNA microarray. J Cancer Res Clin Oncol [Epub ahead of print].
5. Sugiyama Y, Sugiyama K, Hirai Y, et al. (2002) Microdissection is essential for gene expression profiling of clinically resected cancer tissues. Am J Clin Pathol 117:109–116.[CrossRef][ISI][Medline]
6. Tang KF, Tang S, Chan SH, et al. (2001) A distinct expression of CC chemokines by macrophages in nasopharyngeal carcinoma: implication for the intense tumor infiltration by T lymphocytes and macrophages. Hum Pathol 32:42–49.[CrossRef][ISI][Medline]
7. Grosch S, Maier T, Schiffmann S, Geisslinger G. (2006) Cyclooxygenase-2 (COX-2)-independent anticarcinogenic effects of selective COX-2 inhibitors. J Natl Cancer Inst 98:736–747.
8. Easdale S, Clarke P, Titley J, et al. (2003) Gene expression and cell cycle analysis of HT29 human colon cancer cells, exposed to celecoxib, indomethacin, sulindac sulphide and sulindac sulphone. Proc Am Assoc Cancer Res 44:A914.
9. Smith MA and Anderson B. (2001) Phase II window studies: 10 years of experience and counting. J Pediatr Hematol Oncol 23:334–337.[CrossRef][ISI][Medline]
10. Solomon SD, McMurray J, Pfeffer MA, et al. (2005) Cardiovascular risk associated with celecoxib in a clinical trial for colorectal adenoma prevention. N Engl J Med 352:1071–1080.
11. Dowlati A, Haaga J, Remick SC, et al. (2001) Sequential tumor biopsies in early phase clinical trials of anticancer agents for pharmacodynamic evaluation. Clin Cancer Res 7:2971–2976.
12. Song X, Lin H, Johnson AJ, et al. (2002) Cyclooxygenase-2, player or spectator in cyclooxygenase-2 inhibitor-induced apoptosis in prostate cancer cells. J Natl Cancer Inst 94:585–591.
13. Grosch S, Tegeder I, Niederberger E, et al. (2001) COX-2 independent induction of cell cycle arrest and apoptosis in colon cancer cells by the selective COX-2 inhibitor celecoxib. FASEB J 15:2742–2744.
14. Kulp SK, Yang Y, Hung CC, et al. (2004) 3-phosphoinositide-dependent protein kinase-1/Akt signaling represents a major cyclooxygenase-2-independent target for celecoxib in prostate cancer cells. Cancer Res 64:1444–1451.
15. Masferrer JL, Leahy KM, Koki AT, et al. (2000) Antiangiogenic and antitumor activities of cyclooxygenase-2 inhibitors. Cancer Res 60:1306–1311.
16. Chen P and Long Q. (2004) Effects of cyclooxygenase 2 inhibitors on biological traits of nasopharyngeal carcinoma cells. Acta Pharmacol Sin 25:943–949.[ISI][Medline]
17. Chan CM, Ma B, Wong SC, Chan AT. (2005) Celecoxib induces dose dependent growth inhibition in nasopharyngeal carcinoma cell lines independent of cyclooxygenase-2 expression. Biomed Pharmacother 59:S268–S271.
18. Ferrandina G, Ranelletti FO, Legge F, et al. (2003) Celecoxib modulates the expression of cyclooxygenase-2, ki67, apoptosis-related marker, and microvessel density in human cervical cancer: a pilot study. Clin Cancer Res 9:4324–4331.
19. Foote RL, Harris J, Hammond E, et al. (2005) Evaluation of tumor angiogenesis measured with microvessel density (MVD) as a prognostic indicator in nasopharyngeal carcinoma: results of RTOG 9505. Int J Radiat Oncol Biol Phys 61:745–753.[CrossRef][ISI][Medline]
20. Patel M, Du B, Chang M, et al. (2005) Celecoxib inhibits prostate cancer growth: evidence of a cyclooxygenase-2-independent mechanism. Clin Cancer Res 11:1999–2007.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. J Randall and J. R. Foster The Demonstration of Immunohistochemical Biomarkers in Methyl Methacrylate-Embedded Plucked Human Hair Follicles Toxicol Pathol, December 1, 2007; 35(7): 952 - 957. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||