Analysis of prognostic and drug resistance markers in lung cancer

Davidson, Scott Mitchell (2007) Analysis of prognostic and drug resistance markers in lung cancer. MD thesis, University of Glasgow.

Abstract

Lung cancer continues to have by far the highest cancer mortality in the UK. Beyond stage of disease and the patient’s Performance Status there are no other robust clinical or molecular markers of prognosis available today. One major reason for the high mortality rate of this disease is the significant proportion of patients who present with advanced incurable stage disease. A second significant problem in the management of patients with lung cancer is chemotherapy resistance. In patients with NSCLC (80% of cases) initial response rates to cisplatin-based chemotherapy are modest at best. In the case of SCLC (20% of cases) the initial response rates to cisplatin-based chemotherapy are high. However relapse, often with chemotherapy resistant disease, is all too common.

There is a significant body of evidence that demonstrates the role of the loss of mismatch repair activity in the mechanism of chemotherapy resistance. Studies to date have demonstrated that the loss of mismatch repair protein expression is a consequence of 2 main mechanisms: gene mutation and the epigenetic phenomenon of methylation of the MLH1 (a major mismatch repair protein) gene promoter region. Studies have shown that the loss of MLH1 expression is associated with acquired resistance in ovarian and breast cancers. Allelic imbalance (loss of heterozygosity) of chromosome 3p is common in lung cancer samples and the MLH1 gene locus resides here. This loss of heterozygosity (LOH) has been demonstrated to be a poor prognostic indicator in patients with primary adenocarcinoma of the lung. In this translational research project the role of mismatch repair (MMR) and LOH in patients with lung cancer has been further investigated.
In the first study using 2 separate cohorts of patients the potential role of mismatch repair proteins as a marker of prognosis in patients with NSCLC was investigated.

The first cohort of patients had received either cisplatin based or non-cisplatin based chemotherapy for predominantly advanced (stage IIIb/IV) disease at Stobhill Hospital in Glasgow. Pre-treatment paraffin-embedded bronchoscopic samples were collected retrospectively. The aims of this study were to examine the level of expression of the MLH1, MSH2, p53 and Ki67 proteins in these small samples and assess whether these results correlated with any clinicopathological variables or with prognosis. In addition differences in overall survival between the cisplatin and the non-cisplatin chemotherapy treated patients were assessed relative to the level of protein expression. Despite the small size of the biopsies good inter-observer kappa scores were demonstrated between 2 independent scorers for each protein immunohistochemistry score (IHC) (MLH1 - 0.6062, MSH2 – 0.4313, p53 – 0.591). Although this demonstrates that historic bronchoscopy samples can be used for such studies, the number of markers studied was limited due to small sample size and in this study there was insufficient tissue to assess expression levels of Ki67. No significant correlations were demonstrated between IHC score and overall survival for any of the proteins studied, nor was there any difference between the different chemotherapy regimens. Similarly there was no correlation between IHC score of the studied proteins and any clinicopathological variables.

The second cohort of patients all had resection of their primary NSCLC tumour at Aberdeen Royal Infirmary. Of these 50 patients, 10 had received pre-operative cisplatin-based chemotherapy. Fresh frozen surgical samples collected at the time of surgery were analysed retrospectively. The aims of this study were to examine the level of expression of the MLH1, MSH2, p53 and Ki67 proteins in these samples and assess whether these results correlated with any clinicopathological variables or prognosis. Whether any difference existed in these variables between the patients receiving pre-operative chemotherapy and those that did not was also investigated. This study failed to show any significant correlation between the level of protein expression and overall survival or any other clinicopathological variable. Further there was no difference in the level of protein expression between those patients who had received pre-operative chemotherapy and those who had not. However given that only a small number of patients had received pre-operative chemotherapy, further large studies would be required to validate these results.

In the second study we investigated the status of CpG island methylation (using methylation specific PCR) and its role as a marker of prognosis in patients with NSCLC. The samples were the same surgical samples as described above as well as normal adjacent lung tissue. The markers studied were MLH1, p16, DAPK, TIMP 3, HIC 1, MINT 25, MINT 31 and RASSF1A. In this study 30 (60%) of samples exhibited methylation of at least one promoter site with 19 (38%) at 1 site, 5 (10%) at 2 sites, 2 (4%) at 3 sites and 4 (8%) at 4 sites. Twenty (40%) of the tumour samples exhibited no methylation at any promoter sites. Methylation rates in normal adjacent lung tissue were low. There was no significant correlation between the number of methylated sites and either overall survival or any other studied clinicopathological variable. The investigation of methylation at individual sites demonstrated an association between HIC 1 methylation and stage of disease (p = 0.020) and methylation of MINT 31 was associated with a better overall survival (P = 0.030). This remained the case when analysis was performed excluding those patients who had received pre-operative chemotherapy. This is the first report of MINT 31 methylation being studied in NSCLC and therefore further independent studies would be required to validate these results and confirm that the associations had not occurred by chance due to multiple testing. No significant differences in any of the studied variables were demonstrated when comparing patients who had received pre-operative chemotherapy with those who had not.

A third study was performed to validate the findings of previous studies that loss of heterozygosity (LOH) of chromosome 3p is a common occurrence in patients with NSCLC (Mitsudomi et al., 1996). This study also aimed to investigate any correlations between 3p LOH (D3S1289, D3S1300, D3S1304) or hMLH1 promoter methylation and level of MLH1 expression. In addition this study attempted to correlate the presence of any molecular changes found in the serum DNA taken preoperatively from patients with those of their primary tumour samples. For this study prospective collection of surgical tumour and normal adjacent lung tissue samples as well as a pre-operative whole blood sample was collected from patients undergoing resection of their primary disease at the Western Infirmary in Glasgow. Numbers in this study were small thus making any attempt at statistical analysis inappropriate. Observations demonstrated that 3p LOH was common in the primary tumour with 4/8 samples demonstrating LOH at D3S1289, 5/6 at D3S1300 and 5/9 at D3S1304. Corresponding changes were demonstrated in the preoperative serum samples in 2 of the 4 patients at D3S1289, 3/5 at D3S1300 and 2/5 at D3S1304. Loss of Heterozygosity at chromosome 3p did not appear to affect the level of MLH1 expression and hMLH1 methylation was not demonstrated in any of the studied tumour samples.

A final study, in collaboration with Dr J Plumb, was performed to investigate the role of the mismatch repair proteins in the chemotherapy sensitivity of a panel of small cell lung cancer cell lines. In this study it was demonstrated that there was a high correlation between cisplatin sensitivity and the mismatch repair proteins MLH1 (r2 = 0.83) and MSH2 (r2 = 0.87) but not PMS2 (r2 = 0.22). Two of the cell lines originated from metastatic biopsies from the same patient, one pre-chemotherapy treatment (LS274) and one post-chemotherapy (LS310). It was shown that LS310 is 2.3 times more resistant to cisplatin and shows a 50% reduction in MLH1 expression when compared to LS274 (p < 0.001). It was demonstrated that the hMLH1 promoter region of LS310 exhibited methylation whereas the LS274 promoter region did not. Neither of these lines exhibited methylation of the p16, MINT 25 or DAPK loci suggesting that de novo methylation was not responsible for the methylation specific PCR results. Further work demonstrated that treatment of the LS310 cell line with the demethylating agent decitabine increased its cisplatin sensitivity as well as increasing the level of MLH1 expression of the cell line. No such changes were demonstrated in the LS274 cell line after treatment with decitabine.

In summary, this research project was limited by the availability of samples. However it has demonstrated that collaborative multidisciplinary prospective planned translational research can be done and emphasises the need for a translational component to be an integral part of future lung cancer studies.

• Item Type: Thesis (MD)
• Qualification Level: Doctoral
• Keywords: lung cancer, prognosis, drug resistance, mismatch repair, MLH1, DNA methylation, methylation specific PCR, allelic imbalance, loss of heterozygosity
• Subjects: R Medicine > R Medicine (General); R Medicine > RC Internal medicine > RC0254 Neoplasms. Tumors. Oncology (including Cancer)
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cancer Sciences
• Supervisor's Name: Brown, Professor Robert and Milroy, Dr Robert
• Date of Award: September 2007
• Depositing User: Dr Scott M Davidson
• Unique ID: glathesis:2007-101
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 10 Mar 2008
• Last Modified: 30 Jan 2024 11:54
• URI: https://theses.gla.ac.uk/id/eprint/101