Can we accurately report PTEN status in advanced colorectal cancer?

BACKGROUND

Loss of phosphatase and tensin homologue (PTEN) function evaluated by loss of PTEN protein expression on immunohistochemistry (IHC) has been reported as both prognostic in metastatic colorectal cancer and predictive of response to anti-EGFR monoclonal antibodies although results remain uncertain. Difficulties in the methodological assessment of PTEN are likely to be a major contributor to recent conflicting results.

METHODS

We assessed loss of PTEN function in 51 colorectal cancer specimens using Taqman® copy number variation (CNV) and IHC. Two blinded pathologists performed independent IHC assessment on each specimen and inter-observer variability of IHC assessment and concordance of IHC versus Taqman® CNV was assessed.

RESULTS

Concordance between pathologists (PTEN loss vs no loss) on IHC assessment was 37/51 (73%). In specimens with concordant IHC assessment, concordance between IHC and Taqman® copy number in PTEN loss assessment was 25/37 (68%).

CONCLUSION

Assessment PTEN loss in colorectal cancer is limited by the inter-observer variability of IHC, and discordance of CNV with loss of protein expression. An understanding of the genetic mechanisms of PTEN loss and implementation of improved and standardized methodologies of PTEN assessment are required to clarify the role of PTEN as a biomarker in colorectal cancer.

Proangiogenic tumor proteins as potential predictive or prognostic biomarkers for bevacizumab therapy in metastatic colorectal cancer

Tumor biomarkers to more accurately predict a patient's response to a given therapy are much needed in oncology practice. For metastatic colorectal cancer the anti-vascular endothelial growth factor (VEGF) monoclonal antibody bevacizumab is now commonly included in first-line therapy regimens and has led to modest but significant improvements in patient outcomes compared with chemotherapy. Given the modest gains there is a pressing need for predictive biomarkers to better identify patients who would benefit from this targeted therapy. We used a multiplex protein assay to determine the tumor expression levels of the proangiogenic proteins IL-6, IL-8, bFGF, PDGF-BB and VEGF-A in formalin-fixed paraffin-embedded tumors from the MAX clinical trial patients with available tissue samples. Patients were dichotomized into "low" vs. "high" expression subgroups based on median baseline levels to correlate with objective response rate (ORR), progression-free survival (PFS) and overall survival (OS). "Low" tumor VEGF-A level was predictive of better ORR for bevacizumab [ORR (low) 53% vs. (high) 19%, interaction p = 0.03] but not for PFS [hazard ratio, HR (low) 0.73 vs. (high) 0.62, interaction p = 0.68] in the comparison of capecitabine (C) versus C and bevacizumab (CB) and CB plus mitomycin (M). When analyzed as a dichotomized variable, "high" VEGF-A was prognostic for shorter PFS (unadjusted HR 1.34, p = 0.06; adjusted HR 1.55, p = 0.008). The other four proteins were neither predictive of bevacizumab benefits nor prognostic for ORR, PFS or OS. "Low" tumor VEGF-A was associated with longer PFS after adjustment for other baseline factors. Proangiogenic proteins were not predictive of benefit with bevacizumab for PFS.

Assessment of IL-6, IL-8, bFGF, PDGF-BB, and VEGF-A as prognostic and predictive biomarkers for anti-VEGF in metastatic colorectal cancer (mCRC)

502

Background: There remains a need for predictive biomarkers in mCRC to better identify patients who clinically benefit from anti-VEGF therapy, in particular bevacizumab (BEV). The AGITG MAX trial compared capecitabine (C) with capecitabine (+/- mitomycin C (M)) and BEV and PFS was superior in the CB+/-M arm. Here we have taken a panel of pro-angiogenic proteins to assess whether there is a signal for a predictive factor for bevacizumab outcome in the MAX trial tissue population (TTP). Methods: Protein was isolated from FFPE tumour tissue and assayed using custom Bio-plex arrays to assess the expression levels of candidate tumour biomarkers IL-6, IL-8, bFGF, PDGF-BB and VEGF-A. As there are no accepted cutoff points for these analytes we chose the medians of distribution of each of these biomarkers as cutoff points for dichotomization into “high” v “low”. We then correlated levels of the biomarkers with objective response rate, disease-free and overall survival. Results: Total protein was isolated from FFPE tumour sections from 41.5% of patients (n=196) recruited into the MAX clinical trial and assayed for the proteins using a multiplex platform. Patients were generally balanced, tumour tissue population v intention to treat. Median follow up time is 30.2 mths. For patients receiving BEV there was no relationship between the level of IL-6, IL-8, bFGF, PDGF-BB and VEGF-A and PFS and OS. In patients treated with BEV/C vs. C alone, low tumor VEGF-A was predictive of better ORR as compared to those with a high level (ORR 53% vs 38%, p = 0.03). Multivariate analyses showed low VEGF-A expression level was a significant prognostic factor for longer progression free survival (median PFS 9.3 months low VEGF-A v 7.2 months high VEGF-A, multivariate adjusted HR 1.55 (95% CI 1.12-2.13), p = 0.008). Conclusions: We have shown that the chosen panel of pro-angiogenic proteins was not predictive for PFS or OS outcomes when B is added to C in mCRC. However low VEGF-A level did predict for higher ORR and was an independent prognostic marker for PFS in this cohort.

Analysis of PTEN in patients with advanced colorectal cancer (CRC) receiving capecitabine alone or in combination with bevacizumab with or without mitomycin C in the phase III AGITG MAX trial

3534

Background: There is an urgent need for potential biomarkers for anti-VEGF therapies. The tumour suppressor gene PTEN is thought to have a potential role as a biomarker for anti-EGFR therapy in CRC, but there is also evidence that decreased levels of PTEN leads to increased expression of VEGF suggesting a potential relationship to outcome with anti-VEGF therapy*. The prognostic value of PTEN also remains controversial. Methods: Patients enrolled in the Phase III MAX trial of capecitabine (C) +/- bevacizumab (B) (+/- mitomycin C (M)) with available tissue were analysed for PTEN expression (loss v no loss) as assessed using a Taqman copy number assay to measure copy number variation at the PTEN locus. PTEN status was correlated with progression-free survival (PFS) and overall survival (OS) outcomes. The predictive value of PTEN status for bevacizumab efficacy was also examined. Results: Of the original 471 patients enrolled in the trial, tissue from 302 (64.1%) patients were analysed. Baseline characteristics of those with and without tissue were comparable. PTEN loss was observed in 38.7% of patients. There was no relationship between PTEN loss and KRAS or BRAF mutation. Patients with PTEN loss were more likely to have rectal primary (p=0.01) and less likely to have lung metastasis (p=0.03). By using the comparison of C v CB+CBM, PTEN status was not significantly predictive of the effectiveness of B for PFS or OS (Table). PTEN status is also not prognostic for PFS or OS in multivariate analyses adjusting for other baseline factors (Table). Conclusions: PTEN status did not significantly predict different benefit with anti-VEGF therapy. PTEN status was also not prognostic for survival in advanced colorectal cancer. [Table: see text]

Measurement of an analog of insulin-like growth factor-I in blood plasma using a novel enzyme-linked immunosorbent assay

Long-Arg3-IGF-I (LR3IGF-I) is a synthetic analog of IGF-I that has much lower affinity for IGF-binding proteins than do native IGFs-I and -II. Comparisons of the effects of LR3IGF-I with those of IGFs-I and-II in in vitro and in vivo studies have proved useful in defining the functions of IGF-binding proteins. We have developed a sensitive noncompetitive nonisotopic assay of LR3IGF-I. Mouse IgG 1A7-F5-E5 binds an epitope that contains the substituted arginine3 in LR3IGF-I and was used as the solid phase antibody. The solution phase antibody was a rabbit immunoglobulin Nelson which binds to an epitope that is common to IGF-I and LR3IGF-I. The ELISA system was able to detect as little as 50 pg LR3IGF-I in 100 microliters and the native peptides IGFs-I and -II have less than 0.01% activity. Blood plasma from animals treated with pharmacologically active doses of this growth factor analog could be diluted 33.3-fold before assay, at which concentrations plasma had no significant effect on the assay. The ELISA response to LR3IGF-I was unaffected by the presence of IGF-binding proteins. The intra-assay and interassay coefficients of variation are 2.8 and 7.3% respectively. Recovery of LR3IGF-I added to blood plasma was approximately 90%. The ELISA was used to measure LR3IGF-I concentrations in plasma of cows treated with a pharmacologically active dose of this peptide and the results were compared with those obtained by a previously established LR3IGF-I RIA that requires size exclusion chromatography of plasma under acidic conditions to eliminate IGF-binding protein artefacts from the RIA. There was a positive correlation between results obtained by the two assays. The LR3IGF-I ELISA permits discrimination between the exogenous synthetic IGF-I analog and the endogenous native IGFs-I and -II in animals treated with this growth factor without the need for radioiodination of LR3IGF-I and elimination of the requirement for extraction of plasma before assay.

Identification of a cell surface protein with a role in stimulating human keratinocyte proliferation, expressed during development and carcinogenesis

In an attempt to define cell surface molecules with an important role in the development of squamous cell carcinomas (SCCs), we generated monoclonal antibodies (MoAbs) to a human keratinocyte cell line (FEP18-11-T1) capable of giving rise to SCCs in nude mice. MoAb 10G7 was selected for further study because it bound to a cell surface component preferentially expressed by this cell line as compared with normal human foreskin keratinocytes. This MoAb recognizes a cell surface protein (10G7 antigen) that is not detectable on normal keratinocytes in the foreskin in vivo, but whose expression is induced when the keratinocytes are dissociated from this tissue and placed in culture. Interestingly, the 10G7 antigen is downregulated upon keratinocyte differentiation in vitro. Consistent with its expression in hyper-proliferative epithelia in vitro, 10G7 antigen exhibited a classic oncofetal pattern of expression in vivo. Thus, although no reactivity was obtained with MoAb 10G7 in the epithelia of normal foreskin or cervical tissue, strong reactivity was detected in epithelia from genital lesions ranging from benign warts to invasive SCCs. Epidermis from developing fetal tissue also exhibited strong reactivity with MoAb 10G7. We have been able to demonstrate that this MoAb is capable of stimulating FEP18-11-T1 keratinocyte proliferation in vitro in a concentration-dependent manner in the absence of growth factors, suggesting that the 10G7 antigen may play an important role in regulating cellular proliferation during development and in carcinogenesis in epithelial tissues.

High-affinity monoclonal antibodies against P-glycoprotein

Cancer patients treated with one anticancer agent often develop resistance to a broad spectrum of chemotherapeutic agents. This type of multiple drug resistance (MDR) is often accompanied by a decrease in drug accumulation and an increase in expression of a 170,000-Da plasma membrane glycoprotein (P-170) that can effectively pump various anticancer agents out of cytoplasm. A panel of 12 IgG1, IgG2a, or IgG2b monoclonal antibodies was generated against the extracellular portion of P-glycoprotein by immunizing mice with a human MDR1 gene-transfected BA3T3 fibroblast line. We have characterized two of the anti-P-glycoprotein monoclonal antibodies, 15D3 and 17F9, in some detail. Both antibodies immunoprecipitate a 170- to 180-kDa protein from MDR cells, but do not block binding of the known anti-P-glycoprotein antibody MRK16, suggesting that 15D3 and 17F9 bind to a different epitope on the extracellular domain of P-glycoprotein than MRK16. Scatchard analysis revealed that 15D3 and 17F9 had association constants of 1.3 and 1.1 x 10(8) M-1, respectively. 15D3 and 17F9 had little effect on MDR cell growth except for a minor inhibition of KB-V1 cells when the cells were incubated in the presence of vinblastine. Neither antibody inhibited the efflux of P-glycoprotein substrates from MDR cells. Because of their strong binding activity, these antibodies may be useful for diagnostic detection of MDR in patients undergoing chemotherapy or as targeting components of immunotherapeutic agents.