Potential Clinical Significance of a Plasma-Based KRAS Mutation Analysis in Patients with Advanced Non–Small Cell Lung Cancer

Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKI) such as gefitinib and erlotinib are selective TKIs that can block the intracellular receptor binding sites of ATP, thus inhibiting the downstream signaling transmission. Several EGFR-TKIs have been approved as second- or third-line agents for advanced non–small cell lung cancer (NSCLC) patients who failed in platinum-based chemotherapy (1, 2).

The discovery that EGFR tyrosine kinase domain mutations were strongly associated with greater sensitivity of NSCLC to EGFR-TKIs in vitro and higher response rates in clinical studies provided rationale for using molecular markers to identify patients who are most likely to benefit from EGFR-TKI therapy. Subsequent prospective studies focusing on exploring the possibility of EGFR-TKIs as first-line therapy, such as IPASS (IRESSA Pan-Asia Study, a phase III, randomized, open-label, first-line study of gefitinib versus carboplatin/paclitaxel in clinically selected patients with advanced NSCLC in Asia) and the Spanish Lung Cancer Group trial (a multicenter prospective phase II trial of customized erlotinib for advanced NSCLC patients with EGFR mutations), have shown an outstanding survival benefit for patients with EGFR mutant tumors who received first-line EGFR-TKI therapy, which is superior to the outcomes of first-line chemotherapies, suggesting that EGFR mutation is a useful marker for selecting patients who may benefit from first-line EGFR-TKI treatment (3, 4).

KRAS is an important molecule in the downstream signaling network of EGFR. The mutation rate is about 15% to 30% in NSCLC (5–8). KRAS mutations are mostly on codon 12, 13, and 61 and result in the inhibition of GTPase activity, leading to the constitutive activation of Ras protein, which may render tumor cells independent of EGFR signaling. Occurrences of KRAS and EGFR mutations are often mutually exclusive and patients with KRAS mutations are nonresponsive to EGFR-TKIs. Several studies testified that KRAS mutations are associated with primary drug resistance to EGFR-TKIs (9–11). In a phase III trial (BR21.) comparing the effects of erlotinib and the best supportive care for advanced NSCLC patients who failed in platinum-based chemotherapy, patients with tumors carried KRAS mutation had lower response rate and poorer median survival time (MST) to erlotinib compared with patients without tumors KRAS mutation (12).

Although KRAS mutations can be detected in tumor tissues, obtaining such tissues is clinically difficult, particularly for patients with recurrent and refractory NSCLC. Even in prospectively conducted clinical trials, <40% of the patients had available tumor tissues for the mutation analysis. In our previous study, we have shown that DNA abnormalities such as EGFR mutations can be reliably detected in plasma samples of patients with stages IIIB to IV NSCLC and can be used as a biomarker to predict tumor response to EGFR-TKIs and progression-free survival (PFS; ref. 13), suggesting that plasma may be an alternative sample source for such genetic evaluation.

In this study, we compared KRAS mutation status between plasma DNA and the matched tumor tissues in 273 consecutive Chinese patients with stage IIIB/IV NSCLC in a hope to use plasma as an alternative source for KRAS mutation analysis.

All patients had pathologically confirmed stage IIIB or stage IV NSCLC, an Eastern Cooperative Oncology Group performance status of 0 to 2, has available plasma and primary tumor tissue, and received either chemotherapy or EGFR-TKI treatment. Two hundred and seventy-three consecutive patients who met the above criteria were identified in the Beijing University School of Oncology, Beijing Cancer Hospital from June 2005 to June 2008. The patient population consists of 115 females and 158 males. There were 198 cases of adenocarcinoma and 67 cases of squamous cell carcinoma and 8 other tumor types. There were 74 and 199 cases of stage IIIB and IV, respectively, with 113 cases of smokers (≥10 pack-year), 22 cases of exsmokers (100 cigarettes in a life time ≤ cigarettes < 10 pack-year), and 138 nonsmokers (<100 cigarettes in a lifetime). The laboratory data were obtained and recorded independently and blinded from clinical data until analyses by a biostatistician. The study was reviewed and approved by the Institutional Ethic Committee. All the patients signed an informed consent for participation of the study and the use of their biological tissues.

Objective tumor response was determined using the Response Evaluation Criteria in Solid Tumors. The patients had computed tomography scan covering target lesions every 6 wk (two cycles) during the course of chemotherapy and every 8 wk in the EGFR-TKI treatment phase. Brain or bone metastases were evaluated every 3 or 6 mo by magnetic resonance imaging or bone scintigraphy,

According to the Response Evaluation Criteria in Solid Tumors guideline, response to therapy was categorized into four group: complete response, partial response, stable disease, and progression of disease. Overall survival (OS) was calculated as the time from the beginning of therapy to death or last follow-up. The PFS of disease was calculated from the date of the beginning of chemotherapy to the date of tumor progression or death.

Each enrolled patient contributed one blood sample and one tumor tissue sample. The 4 mL anticoagulated venous blood from each patient was collected before TKI or second-line chemotherapy and placed at 4°C for 4 h, then centrifugalizated at 2,500 rpm for 10 min at low temperature. The plasma was aspirated to a new centrifuge tube and then high-speed centrifugation was carried out again. Plasma (1.5 mL) was obtained, and then added 15 μL protease K (20 mg/mL), 50 μL SDS (20%), and DNA was extracted by phenol/chloroform/isopentanol after water bath at 60°C for 2 h. Then, it was precipitated in alcohol and centrifugalizated the next day. After salt washing, the sediment was dissolved in quantity-sufficient Tris-EDTA buffer and then preserved at −20°C for spare use.

A total of six to eight pieces of 5-μm-thick slices were cut from paraffin-embedded tissues. The tumors were macrodissected and tumor contents were recorded for each sample by using immediately adjacent sections. About 80% of the samples had >30% tumor contents, whereas in 20% of the tumor samples, the tumor content was between 20% and 30%. After xylene dewaxing, we added lysate and protease K, then placed samples overnight in 60°C water bath. DNA was extracted by phenol/chloroform/isopentanol the next day, and the following steps were the same as the above process of plasma.

Plasma KRAS gene mutation was detected using PCR-restriction fragment length polymorphism in 120 NSCLC patients who were treated by EGFR-TKIs. The 10-μL PCR system of the first turn included 20 ng DNA, 1 × PCR buffer, 1.5 mmol/L MgCl₂, 200 nmol/L PCR primers (KRAS-A, KRAS-B; Table 1), 200 nmol deoxynucleotide triphosphates (Promega Corp.), and 0.1 unit of Go-Tag DNA polymeRase (Promega Corp.). This step was to introduce restriction sites of incision enzyme BstXI (5′-CCANNNNNNTGG-3′) and XcmI (5′-CCANNNNNNNNNTGG-3′) by primer KRAS-A. PCR instrument (2720 Thermal cycler, Applied Biosystems) was used for amplification. The reaction condition was as follows: predegeneration at 95°C for 5 min, degeneration at 95°C for 30 s, renaturation at 55°C for 30 s, and extension at 72°C for 30 s. There were a total of 38 circulations and a final extension for 5 min at 72°C. The product length of the first PCR was 166 bp. PCR product (2 μL) containing ∼150 ng DNA was digested by restriction enzymes BstXI (MBI Fermentas) and XcmI (New England Biolabs) at 37°C for 2 h. The total volume was 20 μL, and the operations were based on the instruction provided by the manufacturer. Then, 2 μL digested product was used for the second amplification, and the primers were KRAS-A and KRAS-C (Table 1), with the same reaction system as the first PCR amplification. Ten microliters of PCR product were then digested by restriction enzymes BstXI and XcmI again under the same condition. The digested products were examined by 18 μg agarose gel electrophoresis and stained by ethidium bromide, then the results were analyzed by a UV imaging system. The digested fragments of wild-type DNA included 106, 28, and 18 bp, and the digested fragments of mutant DNA included 134 and 18 bp. SW-480 and MDA-MB231 cell line containing mutant codon 12 and 13 were used as positive controls, and RKO and HT-29 containing wild-type codon 12 and 13 were used as negative controls. Meanwhile, the peripheral blood samples from normal subjects and patients with benign pulmonary diseases and autoclaving double-distilled water were used as negative controls.

Denaturing high-performance liquid chromatography was done to detect EGFR mutation by the TransgenomicWave Nucleic Acid 119 Fragment Analysis System with a DNASep column (Transgenomic) according to method described by Hua Bai and coworkers (13).

The SPSS 13.0 software was used for statistical analyses. The relationship between KRAS mutation and relevant factors such as sex, age, stage, histologic type, and smoking was examined by χ² test, with P < 0.05 as bilateral significant difference. The consistency between KRAS mutation in peripheral blood and matched tumor tissue, as well as the consistency between PCR-restriction fragment length polymorphism and sequencing, was analyzed by χ² test; κ was used as the consistency testing. The relationship between KRAS mutation and response of EGFR-TKIs was analyzed by χ² test, and the relationships between KRAS mutation and PFS as well as OS were analyzed by Kaplan-Meier survival curve. The prognostic factors on responses of EGFR-TKIs were discussed by logistic analysis, including age, sex, smoking history, tumor stage, and KRAS mutation. The prognostic influencing factors of PFS and OS after EGFR-TKIs treatment were analyzed by COX regression analysis.

Thirty (11.0%) KRAS mutations were detected in the 273 tumor tissues whereas 35 (12.8%) KRAS mutations were found in the 273 plasma samples. Identical mutations were observed in 23 (76.7%) of the 30 cases with KRAS mutations in the tumor tissues, which indicated a high correlation between the mutations detected in plasma DNA and the mutations detected in the corresponding tumor DNA (P <.001; correlation index, 0.668). Interestingly, 12 (4.4%) patients with plasma DNA mutations had no detectable KRAS mutations in the corresponding primary tumors. Similarly, 7 (2.6%) patients with tumor DNA mutations had no detectable KRAS mutation in the corresponding plasma DNA samples. The correlation between mutations detected in plasma DNA and tumor DNA is summarized in Table 2.

The relationships between KRAS mutations (in plasma and tissue) and clinical parameters were summarized in Table 3. No association was observed between KRAS mutation and smoking history, age, sex, tumor stage, histopathologic type, and EGFR mutation.

Among the 273 NSCLC patients, 120 received EGFR-TKI as second-line treatment or more. One hundred three and 17 of 120 patients received gefitinib and erlotinib therapy, respectively. Until July 10th, 2009, 77 of 120 patients died and 2 lost follow-up. The median follow-up time for the patient population was 23.8 mo. We subsequently analyzed the association between KRAS mutation (from plasma and tissues) and clinical outcomes of EGFR-TKI therapy for these 120 patients (Table 4). The responsive rates to EGFR-TKI treatment in patients with KRAS mutation in plasma and tissue were 5.3% and 7.1%, whereas the response rate of the patients with wild-type were 29.7% and 28.3% (P = 0.024, P = 0.112), respectively. Similarly, the disease control rate in patients with KRAS mutation in plasma and tissues were 15.8% and 27.3% compared with the corresponding 58.4% and 55.7% in patients with wild-type KRAS (P = 0.001, P = 0.022). The median PFS time of patients with KRAS mutation were 2.5 and 2.7 months compared with 8.8 and 8.1 months for patients with wild-type KRAS in plasma and tissues (P < 0.001, P = 0.006; Fig. 1). The MST of patients with KRAS mutation in plasma DNA and tissues were 16.9 and 13.3 months, which were shorter than those of patients with wild-type KRAS (20.3 and 20.6 months, respectively), but the difference did not reach statistic significance (P = 0.827, P = 0.151).

In an earlier study of the same cohort of 120 patients who received TKI treatment, we analyzed EGFR mutations in plasma and determined their clinical implications in patients with advanced NSCLC (13). We subsequently analyzed the association between KRAS mutation and EGFR mutation in plasma DNA. Among these patients, there were 5 patients with both KRAS and EGFR mutation (group A): 18 patients with KRAS mutation and wild-type EGFR (group B), 42 patients with wild-type KRAS and mutant EGFR (group C), and 55 patients with wild-type KRAS and EGFR (group D). Through comparing the PFS and OS by different KRAS and EGFR mutation status in plasma DNA, we found that median PFS in group A, B, C, and D were 2.70, 2.50, 13.25, and 3.53 months (P < 0.001; Fig. 2), and MST in group A, B, C, and D were 17.67, 12.70, 26.40, and 16.9 months (P = 0.103; Table 5). The effect of factors such as age, sex, histologic type, stage, smoking, KRAS (both in plasma and tumor tissues) and EGFR mutation status on PFS and OS were analyzed by Cox regression analysis, which showed that KRAS mutation in plasma, EGFR mutation, smoking status, and adenocarcinoma were independent predictive factors of PFS (P = 0.001, P = 0.004, P < 0.001, P = 0.041). In addition, only EGFR mutation and adenocarcinoma were independent prognostic factors for OS (P = 0.026 and 0.009, respectively). Logistic regression analysis was used to investigate the effects of these eight factors on response. The results showed that KRAS mutation in plasma, EGFR mutation, smoking status, and adenocarcinoma are independent predictors of response of EGFR-TKI therapy (P = 0.001, P = 0.036, P = 0.002, P = 0.012).

KRAS mutation has been linked to EGFR-TKI resistance in multiple studies (9–11), making KRAS mutation a possible marker of resistance to such treatment. However, it is difficult to test KRAS mutation in NSCLC in a routine basis because of the difficulty to obtain tumor tissues from patients with advanced NSCLC, particularly in a repeated manner. Therefore, to explore whether plasma can be used as an alternative source for KRAS mutation testing has important clinical significance. Our study found a positive correlation of KRAS mutation in plasma DNA and in cancer tissues (consistency of 0.668; P < 0.001). Plasma KRAS mutation status is associated with a poor outcome to EGFR-TKIs in NSCLC patients and may be used as a negative predictive marker in selecting patients for such treatment.

Recently, several studies have focused on investigating the association of KRAS mutation between plasma or serum-free DNA and tumor tissues, and their clinical significance in the patients with advanced NSCLC and other cancers, regardless of the type of cancer therapy (14–16). Kimura et al. (14) analyzed KRAS gene mutation in 25 plasma specimens from NSCLC patients with paclitaxel/carboplatin chemotherapy, among whom, there were 12 cases of matched tumor tissues after treatment. The results showed that KRAS mutation rate in peripheral blood was 20%, which was consistent with that in matched tumor tissues. There was no significant difference in MST (3.3 versus 11.4 months; P = 0.056) between patients with KRAS mutation and wild-type KRAS. The study of Ramirez et al. (15) showed that serum KRAS mutation rate in NSCLC patients was higher than that of matched tumor tissues (23% versus 18%), but the difference was not significant. A study by Gautschi and his colleagues (17) found matching codon 12 sequences in blood and tumor in 78% evaluable cases, which support that circulating KRAS mutations originate from tumors and are prognostically relevant in lung cancer. However, contrasting results were also reported (16, 18, 19). For example, Trombino et al. (16) failed to find a correlation between KRAS mutations in serum and the matched tumor tissues. These data together suggest that consistency of KRAS mutation in circulating DNA and tissue remains to be of great controversy. Therefore, standardizing the detection method, developing large samples, and testing it in a cohort of patients with specific cancer therapy are impending for learning its clinical application.

Although high consistency of KRAS mutation in tissues and matched plasma DNA was observed in our study, which is consistent with studies by Kimura et al. and Gautschi et al. (14, 17), it is notable that 4.4% KRAS mutation was identified in plasma DNA samples only and 2.6% in the tumor DNA samples only. This phenomenon has also been observed in previous studies with smaller sample sizes (15, 16). Possible explanations might include, first, the intratumor heterogeneity of genetic abnormalities. In such case, tumor biopsy specimens might not carry the EGFR mutations detected in circulating plasma DNA because these mutations could come from different parts of the tumor. Second, the lower tumor cell content in some of the tumors might also contribute to the lack of detectable mutations in plasma. Third, if the tumor parts that carried mutations shed less DNA than the other parts of the tumors into plasma, then one may miss such mutations in the plasma DNA. Fourth, the dilution of DNA derived from noncancerous tissues, such as inflamed tissues, might impede the detection of mutations in plasma DNA, despite the presence of mutations in tumors. The real mechanism for the discordant of molecular abnormalities between two kinds of samples remains to be resolved.

We observed that KRAS mutation in plasma DNA, but not in tissues, was a predictor factor for PFS and response of TKI therapy in our study, which may be attributed to the different times in obtaining two kinds of samples in a small fraction of our patients. Tissue samples were collected during the time of initial diagnosis, whereas blood samples were gained after the patients failed in one and in multiple regimen therapy and were ready to receive TKI therapy. At the moment, the detection of plasma DNA may more exactly reflect KRAS mutation status. In addition, although we have a decent sample size, due to the low mutation rate, the subgroup of patients with KRAS mutation was still small, especially for those with mutations in the tissue sample, which may result in bias in data analyzed.

Interestingly, our study found and further confirmed by using direct sequencing of individual clones that 5 of 120 patients with TKI therapy had coexisting EGFR mutation and KRAS mutation, which is different from previous reports (20–22). Several studies have indeed suggested that the occurrences of KRAS and EGFR mutation are mutually exclusive; however, occasional tumors with coexisting EGFR and KRAS mutation have also been reported (8, 23, 24). Moreover, we found that the patients with both KRAS and EGFR mutation in plasma DNA had very poor survival and PFS, which may suggest that KRAS mutation in plasma may possess stronger negative predicting role for TKI therapy.

There are several limitations from our study. First, this is a retrospective study in a single center and it needs more prospective study to verify the results. Second, samples might still be small and the mutation rate was low, which may result in some bias in the data analyzed. Third, time point to obtain blood and tissues samples was different in most patients in our study; this may also caused bias.

In summary, our study provides an evidence to suggest that circulating plasma DNA may be used as a surrogate tissue for KRAS mutation analysis in NSCLC. Plasma KRAS mutation status is associated with a poor tumor response to EGFR-TKIs in NSCLC patients and may be used as a predictive marker in selecting patients for such treatment, which needs more prospective, multicenter study to validate.

No potential conflicts of interest were disclosed.

We thank the Department of Statistics of the Beijing Cancer Hospital & Institute, Dr. Ning Wang, a radiologist, for his contribution to this study in response assessment, and Dr. Li Mao (Professor and Chair, Department of Oncology and Diagnosis Sciences University of Maryland Dental School; Adjunct Professor, Department of Thoracic/Head and Neck Medical Oncology, The University of Texas M.D. Anderson Cancer Center) and Xin Shelley Wang (MD, MPH, Associate Professor Department of Symptom Research University of Texas MD Anderson Cancer Center) for their thorough revision of the manuscript.

Grant Support: National High Technology Research and Development Program of China (863 Program 2006AA02A401), the Capital Development Foundation (30772472), and Peking University Program 985 (2-013-39).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.