346 Clinical report
Circulating microRNA expressions in colorectal cancer as predictors of response to chemotherapy Jian Zhang, KeJun Zhang, MeiSheng Bi, XueLong Jiao, DianLiang Zhang and Qian Dong Colorectal cancer (CRC) is the third most frequently diagnosed cancer and the third leading cause of cancer death in the Western world. Chemotherapy has been shown to improve outcomes in patients with CRC; however, only selected patients would benefit from this treatment. We aimed to identify predictors of response to chemotherapy in CRC using circulating microRNAs (miRNAs). We studied differential miRNA expression by miRNA array from serum of 253 patients who had chemotherapy treatment. We screened the differentially expressed serum miRNAs with TaqMan low-density arrays using pooled CRC patient serum samples. Differential expression was validated using hydrolysis probe-based stem–loop quantitative reverse transcription PCR in individual samples. We performed additional unsupervised cluster to analyse the differential expression of serum miRNA between the chemosensitive and chemoresistant patients. A distinct miRNA expression signature in response to chemotherapy was identified. The TaqMan low-density array results demonstrated that 17 serum miRNAs could predict chemosensitivity and chemoresistance. The quantitative reverse transcription PCR analysis further identified a profile of five serum
Introduction Colorectal cancer (CRC) has one of the highest rates of morbidity and mortality worldwide. Approximately 1.4 million cases of CRC are diagnosed yearly in the USA and 0.5 million CRC patients died of the disease in 2012 [1]. Chemotherapy is the most widespread treatment for CRC, especially metastatic CRC. However, the response rate of first-line chemotherapy is only about 50% [2], and for the second-line chemotherapy the response rate is related to the selection of the first-line regimen. In addition, chemoresistant patients suffer from adverse side effects following chemotherapy. To improve the treatment of CRC and to better select those patients who respond to chemotherapy, predictors are urgently required. Unfortunately, effective biomarkers that predict chemosensitivity for clinical use have not yet been identified.
miRNAs (miR-20a, miR-130, miR-145, miR-216 and miR-372) as a biomarker for predicting the chemosensitivity of CRC. The areas under the receiver operating characteristic curve of this five-serum miRNA signature were 0.841 and 0.918 for the two sets of serum samples, respectively. We identified a group of miRNA predictors in response to chemosensitivity for CRC patients. This could lead to a significant improvement in chemotherapy regimen selection strategy and personalized CRC management. Anti-Cancer Drugs c 2014 Wolters Kluwer Health | Lippincott 25:346–352 Williams & Wilkins. Anti-Cancer Drugs 2014, 25:346–352 Keywords: chemotherapy, colorectal cancer, microRNA Department of Surgery, Affiliated Hospital of Medical College Qingdao University, Qingdao, Shandong Province, People’s Republic of China Correspondence to Qian Dong, MD, Department of Surgery, the Affiliated Hospital of Medical College Qingdao University, 16 Jiangsu Road, Qingdao 266003, People’s Republic of China Tel: + 86 532 8291 1324; fax: + 86 532 8291 1630; e-mail: [email protected] Received 4 August 2013 Revised form accepted 19 October 2013
Recent advances in the field of RNA research have provided compelling evidence implicating microRNA (miRNA) in many diverse and substantial biological processes. Because of their capacity to regulate and thus fine-tune the expression of multiple target genes relevant in tumour progression, tumour genesis, angiogenesis, metastasis and sensitivity towards chemotherapy, they influence various pivotal cellular processes with diagnostic, prognostic and therapeutic relevance in CRC [3–5]. Thus, studies of miRNA may contribute to the discovery of possible biomarkers predicting the chemosensitivity of CRC. The aim of this study was to analyse global miRNA expression profiles in the clinical samples of CRC tumours in serum and to identify serum miRNA signatures specific to responders and nonresponders to chemotherapy in patients with CRC.
Materials and methods Study design and patients Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (www.anti-cancerdrugs.com). c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0959-4973
A multistage study was designed to identify a serum miRNA profile as a predictor for CRC chemotherapy sensitivity (Fig. 1). With consent from patients, all sample DOI: 10.1097/CAD.0000000000000049
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
MicroRNA expression and colorectal cancer Zhang et al. 347
Fig. 1
Table 1
Overview of design strategy Screening phase: TaqMan low-density arrays on pooled sample 20 responders and 20 nonresponders More than 20 copies in responders and nonresponders, two-fold altered expression Training phase: qRT-PCR on individual samples 20 responders and 20 nonresponders More than two-fold altered expression, P-value < 0.01 Validation phase: qRT-PCR on individual samples 93 responders and 80 nonresponders More than two-fold altered expression, P-value < 0.01 The proflile of five serum miRNAs serve as non-invasive biomarker for colorectal chemotherapy prediction Overview of design strategy. miRNA, microRNA; qRT-PCR, quantitative reverse transcription PCR.
collection was performed according to the protocols approved by the ethics committee of each participating institution. In total, 253 patients with CRC were enroled in our study. In the initial biomarker screening stage, CRC serum samples pooled from 20 responders and 20 nonresponders to chemotherapy were subjected to TaqMan low-density arrays (TLDA) to identify the miRNAs that were significantly differentially expressed. Subsequently, sequential validation was performed using a hydrolysis probe-based quantitative reverse transcription PCR (qRT-PCR) assay to refine the number of serum miRNAs as a CRC chemotherapy predictor. In the biomarker selection stage, 40 CRC serum samples formed a training set, whereas an additional 173 CRC serum samples formed a validation set. Histopathology of the patients was confirmed by surgical resection of the tumours and the tumour stage was determined on the basis of the surgical findings. For patients who were unsuitable for surgical treatment, histopathology and tumour stage were confirmed by histobiopsy and imaging technology. Tumours were staged according to the tumour-nodemetastasis staging system of the International Union Against Cancer [6]. The demographics and clinical features of the CRC patients are listed in Table 1. Serum preparation, RNA isolation and qRT-PCR assay
Blood samples (5 ml) were collected from each patient. Hydrolysis probes (Applied Biosystems, Foster City, California, USA) were used for the qRT-PCR analysis, according to the manufacturer’s instructions [7,8]. A detailed description of the experimental protocols is shown in the Supplementary data (Supplemental digital content 1, http://links.lww.com/ACD/A50). The miRNAs
Patient characteristics
Variables N Age [mean (SD)] Sex Male Female Stage III IV Chemotherapy Alcohol consumption Yes No Smoking Yes No Positive ratio CA19-9 CEA
Responders
Nonresponders
P-value
138 63.43 (11.48)
115 62.48 (10.97)
0.243a
75 63
65 50
79 59 Oxaliplatin-based
65 50 Oxaliplatin-based
72 66
61 54
0.542b
67 51
59 56
0.734b
48 79
42 61
0.312b
CA19-9, carbohydrate antigen19-9; CEA, carcinoembryonic antigen. Student’s t-test. Two-sided w2-test.
a
b
with a mean fold change of at least 2 and a P-value less than 0.01 were selected for further study. MiRNAs with an expression level lower than the threshold value (Cq > 35) and with a detection rate below 75% in both group samples were discarded. TaqMan low-density arrays
TLDA was analysed to identify the profile of differentially expressed miRNAs between the two sets of samples (20 responders and 20 nonresponders). In brief, total RNA was reverse-transcribed into cDNA by the TaqMan MicroRNA Reverse Transcription Kit and Megaplex RT set pool A and B version 3.0 (Applied Biosystems). The RT product was loaded into TaqMan Array Human MicroRNA A + B Cards Set v3.0 (Applied Biosystems), enabling simultaneous quantitation of 667 human miRNAs. TaqMan microRNA assays and analysis were performed on the ABI 7900HT Instrument (Applied Biosystems). All reactions were performed according to the standard manufacturers’ protocols. Quantitative miRNA expression data were acquired and normalized using ABI 7900HT SDS software (Applied Biosystems). Validation of miRNA
The training phase used serum samples from the 40 CRC cases to validate the results obtained from the TLDA phase, whereas the validation phase used serum samples from an additional 173 CRC cases to validate the significantly differentially expressed miRNA results from the training phase by qRT-PCR assay. Detection of CA19-9 and CEA
The serum values of carbohydrate antigen19-9 (CA19-9) and carcinoembryonic antigen (CEA) were measured using commercial kits (CA19-9 RIA; Abbott AxSYM System, Chicago, Illinois, USA; CEA fluoro-immunometric assay;
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
348 Anti-Cancer Drugs 2014, Vol 25 No 3
Beckman Coulter Inc., Fullerton, California, USA). The upper limits of normal values for CA19-9 and CEA were 37 and 5 U/ml, respectively. Statistical analysis
We used risk score to analyse and evaluate the association between chemotherapy sensitivity and miRNA expression levels. The risk score of each miRNA in the training set, denoted as s, was set as 1 if the expression level was greater than the upper 95% reference interval for the corresponding miRNA level in controls, and 0 otherwise. When taking into account the correlation of each miRNA with chemotherapy-resistant risk, each patient was assigned a risk score function (RSF) according to a linear combination of the expression level of the miRNA. The RSF for sample i using the information from the five miRNAs was as follows: X5 RSFi ¼ Wj sij : j1
In the equation above, sij is the risk score for miRNA j on sample i, and Wj is the weight of the risk score of miRNA j. Five univariate logistic regression models were fitted with the disease status with each of the risk scores to determine the W s. The regression coefficient of each risk score was used as the weight to indicate the contribution of each miRNA to the RSF. The receiver operating characteristic curves were then used to evaluate the predict effects of the profiling and to find an appropriate cutoff point. Validation of the procedure and the cutoffs were performed in the test sample [9]. All the statistical analyses were performed using the statistical analysis system software (v.9.1.3; SAS Institute, Cary, North Carolina, USA). Data are presented as median±SD. A P-value less than 0.01 was considered statistically significant.
with a mean fold change of at least 2 and a P-value less than 0.01 were selected for further study. MiRNAs with an expression level lower than the threshold value (Cq > 35) and with a detection rate below 75% in both group samples were discarded. A genome-wide expression profiling of serum miRNAs obtained by TLDA technology showed that serum from responders had 35-miRNA differential expression as compared with serum from the nonresponders, as shown in Supplementary Table 1 (Supplemental digital content 1, http://links.lww.com/ACD/A50). Individual miRNA was considered to be differentially expressed only if there was a larger than two-fold change in its expression level between the responders and nonresponders to chemotherapy. On the basis of these criteria, 14 miRNAs were found to be differentially expressed between the two groups, analysed by qRT-PCR for individuals of the 20 pool samples, as shown in Supplementary Table 1 (Supplemental digital content 1, http://links.lww.com/ACD/A50). The miRNAs were further analysed by qRT-PCR for individuals. These differentially expressed serum miRNAs were further examined by hydrolysis probe-based RT-qPCR in a training sample set including 20 responders and 20 nonresponders. In this phase, only those miRNAs with a mean fold change greater than 2 and a P-value less than 0.01 were retained. This phase generated a panel of five miRNAs (miR-20a, miR-130, miR-145, miR-216 and miR-372) that were significantly altered between the two group samples. In the validation set, the concentration of these five selected miRNAs was examined by qRT-PCR in a larger sample comprising 93 responders and 80 nonresponders to chemotherapy (Table 2). The miRNA expression pattern alterations in the validation set were consistent with the training set. The expression levels of the five miRNAs were significantly different between the responders and nonresponders. The differential expression of the five miRNAs is shown in Fig. 2.
Results Description and clinical features of the patients
All 253 patients enroled in this study were diagnosed with colorectal adenocarcinoma clinically and pathologically. As shown in Table 1, there were no significant differences in the distribution of smoking (P = 0.734), alcohol consumption (P = 0.953), age (P = 0.542) and sex (P = 0.312) between the two groups of patients. All patients underwent chemotherapy, on the basis of concurrent application of modified folfox6 (oxaliplatin 85 mg/m2, day 1; 5fluorouracil as a continuous intravenous infusion at 46 h dose 2400 mg/m2; 5-fluorouracil intravenous infusion at 30 min dose 300 mg/m2, day 1; CF 400 mg/m2 intravenous infusion at 2 h, day 1). All schedules were repeated every 14 days. Biomarker selection and validation phase
In the initial screening phase, serum samples were pooled from 20 responders and 20 nonresponders. The miRNAs
Separation of responders and nonresponders to chemotherapy by risk score analysis
Risk score analysis based on the five-miRNA expression profile was used to distinguish serum samples of the responders and nonresponders. First, the risk score of each serum sample was calculated, and on the basis of their scores, serum samples were then divided into a high-risk group, representing the possible nonresponder cases, and a low-risk group, representing the responder cases. At the optimal cutoff value of 3.46 with the value of sensitivity + specificity considered to be maximal, the positive predictive value and negative predictive value of the five-serum miRNA signature in the training set were 0.86 and 0.89, respectively. Similarly, when the same cutoff point was used to analyse the validation set, the positive predictive value and negative predictive value of the five-serum miRNA signature were 0.93 and 0.94, respectively (Table 3).
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
MicroRNA expression and colorectal cancer Zhang et al. 349
Overview of miRNAs with significantly different levels of expression in colorectal tumours of responders and nonresponders to chemotherapy (median±SD) (fmol/l)
Table 2
Training set
miR-20a miR-130 miR-145 miR-216 miR-372
Validation set
Responders (N = 20)
Nonresponders (N = 20)
Fold change
P-value
Responders (N = 93)
Nonresponders (N = 80)
Fold change
P-value
143.4±76.7 105.9±57.2 171.4±72.6 121.3±52.5 117.9±68.4
295.6±65.3 273.4±77.7 359.3±66.8 282.1±46.7 285.3±76.7
2.1 2.6 2.1 2.3 2.2
3.6 10–4 2.0 10–4 4.5 10–4 3.7 10–5 5.4 10–4
102.5±53.4 98.8±49.9 158.9±63.2 142.4±61.3 124.3±59.4
283.2±72.1 302.3±74.3 392.4±79.2 293.1±59.3 278.3±48.2
2.8 3.1 2.5 2.1 2.2
2.5 10–5 4.2 10–5 3.9 10–4 3.2 10–5 5.2 10–4
miRNAs, microRNAs.
Fig. 2
miR-20a ∗∗
(a)
Concentration (fM)
400
400
300
300
200
200
100
100 0
0 Responders (c)
Concentration (fM)
Nonresponders
miR-145 ∗∗
500 400
Responders
Nonresponders
miR-216
(d)
∗∗
400 300
300 200 200 100
100 0
0 Responder
Nonresponder
Responder
Nonresponder
miR-372 ∗∗
(e) 400 Concentration (fM)
miR-130 ∗∗
(b)
300 200 100 0 Responder
Nonresponder
(a–e) Detection of CRC using the five-serum miRNA profile as a biomarker. Serum levels of the five miRNAs were measured in 253 CRC patients (in both the training set and the validation set) using a hydrolysis probe-based qRT-PCR assay. **P < 0.01. CRC, colorectal cancer; miRNA, microRNA; qRT-PCR, quantitative reverse transcription PCR.
Receiver operating characteristic curves were then constructed to estimate the sensitivity and specificity of the five-serum miRNA signature. The areas under the curve (AUC) were 0.841 (95% confidence interval
0.707–0.975) and 0.918 (95% confidence interval 0.871–0.963) for the serum samples in the training set and validation set, respectively (Fig. 3a and b). Using the same serum samples, we compared the AUC of five
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
350
Anti-Cancer Drugs 2014, Vol 25 No 3
miRNAs with that of CEA and CA19-9, two currently available blood-based biomarkers for CRC detection. The AUC values of the five-serum miRNA signature were markedly higher than those of CEA (0.689) and CA19-9 (0.746) (Fig. 3c and d). The results indicate that the fiveserum miRNA signature is a more accurate biomarker than CEA and CA19-9 for CRC chemotherapy prediction. Table 3 Analysis of responders and nonresponders to chemotherapy by risk score
Training set Validation set
Score
0–3.46
3.46–13.23
PPV
NPV
Responders Nonresponders Responders Nonresponders
17 2 87 6
3 18 6 74
0.86
0.89
0.93
0.94
NPV, negative predictive value; PPV, positive predictive value.
Unsupervised cluster analysis
We also used the unsupervised cluster method to analyse the differential expression of miRNAs between the responders and nonresponders. The dendrogram generated by the cluster analysis showed a clear separation of the responders from the nonresponders on the basis of the five-serum miRNA profile (Fig. 3). In the training set, only one out of 20 responders and one out of 20 nonresponders were incorrectly classified (Fig. 4a). In the validation set, five nonresponders and four responders were misclassified (Fig. 4b).
Discussion The survival of patients with CRC has increased markedly in the last decade through the use of anticancer drugs in monotherapy or in combined regimens.
Fig. 3
ROC curve of training set (AUC = 0.841,95% CI 0.707−0.975)
ROC curve of CEA (AUC = 0.689,95% CI 0.618−0.760)
(c)
100
100
80
80
60
Sensitivity
Sensitivity
(a)
40 20
60 40 20
0
0 0
20
40 60 80 100% − specificity%
100
100
100
80
80
40 20
40 60 80 100% − specificity%
100
(AUC = 0.746,95% CI 0.682−0.851)
Sensitivity
Sensitivity
(AUC = 0.918,95% CI 0.871−0.963)
60
20
ROC curve of CA 19-9
(d)
ROC curve of validation set
(b)
0
60 40 20
0
0 0
40 20 60 80 100% − specificity%
100
0
60 40 80 20 100% − specificity%
100
(a, b) Receiver operating characteristic (ROC) curves for the five-serum miRNA profile to distinguish responders and nonresponders to chemotherapy in the training set (a) and the validation set (b). (c, d) ROC curves for CEA (c) and CA19-9 (d) to distinguish CRC serum samples in responders and nonresponders. AUC, area under the curve; CA19-9, carbohydrate antigen19-9; CEA, carcinoembryonic antigen; CI, confidence interval; CRC, colorectal cancer; miRNA, microRNA.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
MicroRNA expression and colorectal cancer Zhang et al. 351
Fig. 4
N–Responders10 N–Responders15 N–Responders9 N–Responders5 N–Responders13 N–Responders12 N–Responders3 N–Responders14 N–Responders11 N–Responders1 Responders14 N–Responders8 N–Responders4 N–Responders7 N–Responders20 N–Responders6 N–Responders19 N–Responders18 N–Responders17 N–Responders16 Responders13 Responders12 Responders3 Responders8 N–Responders2 Responders1 Responders4 Responders2 Responders9 Responders19 Responders10 Responders11 Responders6 Responders5 Responders7 Responders20 Responders18 Responders16 Responders15 Responders17
(a)
miR-216 miR-20a miR-145 miR-372 miR-130
(b)
Cluster analysis of miRNA differential expression between responders and nonresponders to chemotherapy serum samples. The concentration of the selected five serum miRNAs in samples was normalized, mean-centred, clustered and plotted as a heat map. (a) The training set (20 responders and 20 nonresponders). (b) The validation set (93 responders and 80 nonresponders). miRNA, microRNA.
Although the rate of success is higher than for other gastrointestinal tumours, chemotherapy response rates for CRC remain disappointingly low, and side effects and development of chemoresistance are severe limitations to pharmacological therapy. A major unresolved problem is that no predictors of response to these treatments are available. MiRNAs constitute a robust regulatory network with post-transcription regulatory efficiency for almost onehalf of human coding genes, including important oncogenes, tumour suppressor genes and genes associated with invasion, dissemination and therapy resistance [10]. Alterations in miRNA expression profiles could be useful in the prediction of individual response to chemotherapy and could even be targets for therapies to increase chemosensitivity. Svoboda et al. [11] demonstrated that miR-215, miR-99a*, miR-196b, miR-450b-5p and let-7e are involved in response of CRC to chemoradiotherapy. Nishida et al. [12] found that expression of miR-10b is a potential indicator of chemosensitivity to the common 5-FU-based chemotherapy regimen. Bitarte et al. [13] found that miR-451 is a novel candidate to circumvent recurrence and drug resistance in CRC and could be used as a marker to predict response to irinotecan in patients with colon carcinoma. Although tissue miRNAs can act as a predictor of chemotherapy for CRC, the collection of tissue samples is an invasive procedure and depends on surgical sections after the initial clinical classification. Recently, serum and plasma miRNAs have emerged as potential new blood-based markers for detecting cancers and other diseases [8,14–18]. In this study, the expression levels of serum miRNAs in patients with CRC have
been systematically determined and five serum miRNAs (miR-20a, miR-130, miR-145, miR-216 and miR-372) have been identified that are significantly upregulated in oxaliplatin-chemoresistant CRC patients compared with chemosensitive patients. We analysed global miRNA expression profiles in CRC serum to find potential predictive miRNAs for response to oxaliplatin-based chemotherapy. The discovery of serum miRNAs as a potential chemotherapy predictor can overcome the problems that exist in collecting tissue samples by an invasive process. Identification of a serum miRNAbased prediction biomarker from a genome-wide serum miRNA expression profile demonstrates that a combination of multiple serum miRNAs is a more comprehensive indicator for tumour chemotherapy than the conventional single protein-based or carbohydrate molecule-based biomarkers. This approach involves a screen of a pooled serum sample followed by individual RT-qPCR validation. Our results showed that the sensitivity and specificity of CRC prediction by this five-miRNA biomarker were 92 and 88%, respectively. This demonstrated that the expression profile of the five serum miRNAs could serve as an accurate predictive biomarker for CRC. Functional study of miRNAs in tumours is also helpful for evaluating serum miRNAs as indicators for various types of cancer. Among the five serum miRNAs identified in CRC patients, many are involved in general tumourigenesis. For example, miR-145 has been identified as a tumour suppressor in oesophageal squamous cell carcinoma [19]. Moreover, increased expression of miR-20a has been observed in colon cancer, pancreatic cancer and
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
352 Anti-Cancer Drugs 2014, Vol 25 No 3
other cancers [20–22]. These differential expression serum miRNAs could not only be useful as curative effect prediction markers but could also serve as a target for the development of novel therapeutic strategies to overcome drug resistance and tumour growth. The mechanisms and intracellular signal in pathways involved in the miRNA-mediated effects are still subject to investigation, and future studies are required to identify the target genes of these five serum miRNAs and the mechanism that regulates their biogenesis. Moreover, their predictive value in a true clinical setting remains undefined, alongside the potential side effects that systemic miRNA-based therapies may induce. Additional investigation is required to further determine the molecular functions and mechanisms of miRNAs, as well as their applicability as diagnostic and therapeutic tools. In sum, we have found five serum miRNAs that have the ability to separate complete remission from progression of disease and thus could be potential serum biomarkers for predicting response to the oxaliplatin-based chemotherapy in CRC. Given that the expression level of these miRNAs is important for CRC chemosensitivity and maintenance, novel therapies could be developed via changing the expression level of these miRNAs. miRNAbased therapy and predictor biomarkers are still in their infancy, but show great promise in the replacement or supplementation of current prognostic and predictive markers and anticancer treatment.
Acknowledgements
5 6
7
8
9
10
11
12
13
14
15
16 17
Conflicts of interest
There are no conflicts of interest.
18
References 1 2
3
4
Siegel R, Naishadham D, Jemal A. Cancer statistics. CA Cancer J Clin 2013; 63:11–30. Goldberg RM, Sargent DJ, Morton RF, Fuchs CS, Ramanathan RK, Williamson SK, et al. A randomized controlled trial of fluorouracil plus leucovorin, irinotecan, and oxaliplatin combinations in patients with previously untreated metastatic colorectal cancer. J Clin Oncol 2004; 22:23–30. Sengupta D, Bandyopadhyay S. Topological patterns in microRNA-gene regulatory network: studies in colorectal and breast cancer. Mol Biosyst 2013; 9:1360–1371. Liu H, Du L, Wen Z, Yang Y, Li J, Wang L, et al. Up-regulation of mir-182 expression in colorectal cancer tissues and its prognostic value. Int J Colorectal Dis 2013; 28:697–703.
19
20
21
22
Wang S, Xiang J, Li Z, Lu S, Hu J, Gao X, et al. A plasma microRNA panel for early detection of colorectal cancer. Int J Cancer 2013 [Epub ahead of print]. Hermanek P. The superiority of the new International Union Against Cancer and American Joint Committee on Cancer TNM staging of gastric carcinoma. Cancer 2000; 88:1763–1765. Mattie MD, Benz CC, Bowers J, Sensinger K, Wong L, Scott GK, et al. Optimized high-throughput microRNA expression profiling provides novel biomarker assessment of clinical prostate and breast cancer biopsies. Mol Cancer 2006; 5:24. Chen X, Ba Y, Ma L, Cai X, Yin Y, Wang K, et al. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res 2008; 18:997–1006. Liu R, Zhang C, Hu Z, Li G, Wang C, Yang C, et al. A five-microRNA signature identified from genome-wide serum microRNA expression profiling serves as a fingerprint for gastric cancer diagnosis. Eur J Cancer 2011; 47:784–791. Slaby O, Svoboda M, Michalek J, Vyzula R. MicroRNAs in colorectal cancer: translation of molecular biology into clinical application. Mol Cancer 2009; 8:102. Svoboda M, Sana J, Fabian P, Kocakova I, Gombosova J, Nekvindova J, et al. MicroRNA expression profile associated with response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer patients. Radiat Oncol 2012; 7:195. Nishida N, Yamashita S, Mimori K, Sudo T, Tanaka F, Shibata K, et al. MicroRNA-10b is a prognostic indicator in colorectal cancer and confers resistance to the chemotherapeutic agent 5-fluorouracil in colorectal cancer cells. Ann Surg Oncol 2012; 19:3065–3071. Bitarte N, Bandres E, Boni V, Zarate R, Rodriguez J, Gonzalez-Huarriz M, et al. MicroRNA-451 is involved in the self-renewal, tumorigenicity, and chemoresistance of colorectal cancer stem cells. Stem Cells 2011; 29:1661–1671. Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, PogosovaAgadjanyan EL, et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci USA 2008; 105:10513–10518. Resnick KE, Alder H, Hagan JP, Richardson DL, Croce CM, Cohn DE. The detection of differentially expressed microRNAs from the serum of ovarian cancer patients using a novel real-time PCR platform. Gynecol Oncol 2009; 112:55–59. Cho WC. Circulating microRNAs as minimally invasive biomarkers for cancer theragnosis and prognosis. Front Genet 2011; 2:7. Ng EK, Chong WW, Jin H, Lam EK, Shin VY, Yu J, et al. Differential expression of microRNAs in plasma of patients with colorectal cancer: a potential marker for colorectal cancer screening. Gut 2009; 58:1375–1381. Wang J, Chen J, Chang P, LeBlanc A, Li D, Abbruzzesse JL, et al. MicroRNAs in plasma of pancreatic ductal adenocarcinoma patients as novel blood-based biomarkers of disease. Cancer Prev Res (Phila) 2009; 2:807–813. Kano M, Seki N, Kikkawa N, Fujimura L, Hoshino I, Akutsu Y, et al. Mir-145, mir-133a and mir-133b: tumor-suppressive miRNAs target fscn1 in esophageal squamous cell carcinoma. Int J Cancer 2010; 127:2804–2814. Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F, et al. A microRNA expression signature of human solid tumors defines cancer gene targets. Proc Natl Acad Sci USA 2006; 103:2257–2261. Zeng X, Xiang J, Wu M, Xiong W, Tang H, Deng M, et al. Circulating mir-17, mir-20a, mir-29c, and mir-223 combined as non-invasive biomarkers in nasopharyngeal carcinoma. PLoS One 2012; 7:e46367. Wang YX, Zhang XY, Zhang BF, Yang CQ, Chen XM, Gao HJ. Initial study of microRNA expression profiles of colonic cancer without lymph node metastasis. J Dig Dis 2010; 11:50–54.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Circulating microRNA expressions in colorectal cancer as predictors of response to chemotherapy Jian Zhang, KeJun Zhang, MeiSheng Bi, XueLong Jiao, DianLiang Zhang and Qian Dong Colorectal cancer (CRC) is the third most frequently diagnosed cancer and the third leading cause of cancer death in the Western world. Chemotherapy has been shown to improve outcomes in patients with CRC; however, only selected patients would benefit from this treatment. We aimed to identify predictors of response to chemotherapy in CRC using circulating microRNAs (miRNAs). We studied differential miRNA expression by miRNA array from serum of 253 patients who had chemotherapy treatment. We screened the differentially expressed serum miRNAs with TaqMan low-density arrays using pooled CRC patient serum samples. Differential expression was validated using hydrolysis probe-based stem–loop quantitative reverse transcription PCR in individual samples. We performed additional unsupervised cluster to analyse the differential expression of serum miRNA between the chemosensitive and chemoresistant patients. A distinct miRNA expression signature in response to chemotherapy was identified. The TaqMan low-density array results demonstrated that 17 serum miRNAs could predict chemosensitivity and chemoresistance. The quantitative reverse transcription PCR analysis further identified a profile of five serum
Introduction Colorectal cancer (CRC) has one of the highest rates of morbidity and mortality worldwide. Approximately 1.4 million cases of CRC are diagnosed yearly in the USA and 0.5 million CRC patients died of the disease in 2012 [1]. Chemotherapy is the most widespread treatment for CRC, especially metastatic CRC. However, the response rate of first-line chemotherapy is only about 50% [2], and for the second-line chemotherapy the response rate is related to the selection of the first-line regimen. In addition, chemoresistant patients suffer from adverse side effects following chemotherapy. To improve the treatment of CRC and to better select those patients who respond to chemotherapy, predictors are urgently required. Unfortunately, effective biomarkers that predict chemosensitivity for clinical use have not yet been identified.
miRNAs (miR-20a, miR-130, miR-145, miR-216 and miR-372) as a biomarker for predicting the chemosensitivity of CRC. The areas under the receiver operating characteristic curve of this five-serum miRNA signature were 0.841 and 0.918 for the two sets of serum samples, respectively. We identified a group of miRNA predictors in response to chemosensitivity for CRC patients. This could lead to a significant improvement in chemotherapy regimen selection strategy and personalized CRC management. Anti-Cancer Drugs c 2014 Wolters Kluwer Health | Lippincott 25:346–352 Williams & Wilkins. Anti-Cancer Drugs 2014, 25:346–352 Keywords: chemotherapy, colorectal cancer, microRNA Department of Surgery, Affiliated Hospital of Medical College Qingdao University, Qingdao, Shandong Province, People’s Republic of China Correspondence to Qian Dong, MD, Department of Surgery, the Affiliated Hospital of Medical College Qingdao University, 16 Jiangsu Road, Qingdao 266003, People’s Republic of China Tel: + 86 532 8291 1324; fax: + 86 532 8291 1630; e-mail: [email protected] Received 4 August 2013 Revised form accepted 19 October 2013
Recent advances in the field of RNA research have provided compelling evidence implicating microRNA (miRNA) in many diverse and substantial biological processes. Because of their capacity to regulate and thus fine-tune the expression of multiple target genes relevant in tumour progression, tumour genesis, angiogenesis, metastasis and sensitivity towards chemotherapy, they influence various pivotal cellular processes with diagnostic, prognostic and therapeutic relevance in CRC [3–5]. Thus, studies of miRNA may contribute to the discovery of possible biomarkers predicting the chemosensitivity of CRC. The aim of this study was to analyse global miRNA expression profiles in the clinical samples of CRC tumours in serum and to identify serum miRNA signatures specific to responders and nonresponders to chemotherapy in patients with CRC.
Materials and methods Study design and patients Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (www.anti-cancerdrugs.com). c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0959-4973
A multistage study was designed to identify a serum miRNA profile as a predictor for CRC chemotherapy sensitivity (Fig. 1). With consent from patients, all sample DOI: 10.1097/CAD.0000000000000049
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
MicroRNA expression and colorectal cancer Zhang et al. 347
Fig. 1
Table 1
Overview of design strategy Screening phase: TaqMan low-density arrays on pooled sample 20 responders and 20 nonresponders More than 20 copies in responders and nonresponders, two-fold altered expression Training phase: qRT-PCR on individual samples 20 responders and 20 nonresponders More than two-fold altered expression, P-value < 0.01 Validation phase: qRT-PCR on individual samples 93 responders and 80 nonresponders More than two-fold altered expression, P-value < 0.01 The proflile of five serum miRNAs serve as non-invasive biomarker for colorectal chemotherapy prediction Overview of design strategy. miRNA, microRNA; qRT-PCR, quantitative reverse transcription PCR.
collection was performed according to the protocols approved by the ethics committee of each participating institution. In total, 253 patients with CRC were enroled in our study. In the initial biomarker screening stage, CRC serum samples pooled from 20 responders and 20 nonresponders to chemotherapy were subjected to TaqMan low-density arrays (TLDA) to identify the miRNAs that were significantly differentially expressed. Subsequently, sequential validation was performed using a hydrolysis probe-based quantitative reverse transcription PCR (qRT-PCR) assay to refine the number of serum miRNAs as a CRC chemotherapy predictor. In the biomarker selection stage, 40 CRC serum samples formed a training set, whereas an additional 173 CRC serum samples formed a validation set. Histopathology of the patients was confirmed by surgical resection of the tumours and the tumour stage was determined on the basis of the surgical findings. For patients who were unsuitable for surgical treatment, histopathology and tumour stage were confirmed by histobiopsy and imaging technology. Tumours were staged according to the tumour-nodemetastasis staging system of the International Union Against Cancer [6]. The demographics and clinical features of the CRC patients are listed in Table 1. Serum preparation, RNA isolation and qRT-PCR assay
Blood samples (5 ml) were collected from each patient. Hydrolysis probes (Applied Biosystems, Foster City, California, USA) were used for the qRT-PCR analysis, according to the manufacturer’s instructions [7,8]. A detailed description of the experimental protocols is shown in the Supplementary data (Supplemental digital content 1, http://links.lww.com/ACD/A50). The miRNAs
Patient characteristics
Variables N Age [mean (SD)] Sex Male Female Stage III IV Chemotherapy Alcohol consumption Yes No Smoking Yes No Positive ratio CA19-9 CEA
Responders
Nonresponders
P-value
138 63.43 (11.48)
115 62.48 (10.97)
0.243a
75 63
65 50
79 59 Oxaliplatin-based
65 50 Oxaliplatin-based
72 66
61 54
0.542b
67 51
59 56
0.734b
48 79
42 61
0.312b
CA19-9, carbohydrate antigen19-9; CEA, carcinoembryonic antigen. Student’s t-test. Two-sided w2-test.
a
b
with a mean fold change of at least 2 and a P-value less than 0.01 were selected for further study. MiRNAs with an expression level lower than the threshold value (Cq > 35) and with a detection rate below 75% in both group samples were discarded. TaqMan low-density arrays
TLDA was analysed to identify the profile of differentially expressed miRNAs between the two sets of samples (20 responders and 20 nonresponders). In brief, total RNA was reverse-transcribed into cDNA by the TaqMan MicroRNA Reverse Transcription Kit and Megaplex RT set pool A and B version 3.0 (Applied Biosystems). The RT product was loaded into TaqMan Array Human MicroRNA A + B Cards Set v3.0 (Applied Biosystems), enabling simultaneous quantitation of 667 human miRNAs. TaqMan microRNA assays and analysis were performed on the ABI 7900HT Instrument (Applied Biosystems). All reactions were performed according to the standard manufacturers’ protocols. Quantitative miRNA expression data were acquired and normalized using ABI 7900HT SDS software (Applied Biosystems). Validation of miRNA
The training phase used serum samples from the 40 CRC cases to validate the results obtained from the TLDA phase, whereas the validation phase used serum samples from an additional 173 CRC cases to validate the significantly differentially expressed miRNA results from the training phase by qRT-PCR assay. Detection of CA19-9 and CEA
The serum values of carbohydrate antigen19-9 (CA19-9) and carcinoembryonic antigen (CEA) were measured using commercial kits (CA19-9 RIA; Abbott AxSYM System, Chicago, Illinois, USA; CEA fluoro-immunometric assay;
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
348 Anti-Cancer Drugs 2014, Vol 25 No 3
Beckman Coulter Inc., Fullerton, California, USA). The upper limits of normal values for CA19-9 and CEA were 37 and 5 U/ml, respectively. Statistical analysis
We used risk score to analyse and evaluate the association between chemotherapy sensitivity and miRNA expression levels. The risk score of each miRNA in the training set, denoted as s, was set as 1 if the expression level was greater than the upper 95% reference interval for the corresponding miRNA level in controls, and 0 otherwise. When taking into account the correlation of each miRNA with chemotherapy-resistant risk, each patient was assigned a risk score function (RSF) according to a linear combination of the expression level of the miRNA. The RSF for sample i using the information from the five miRNAs was as follows: X5 RSFi ¼ Wj sij : j1
In the equation above, sij is the risk score for miRNA j on sample i, and Wj is the weight of the risk score of miRNA j. Five univariate logistic regression models were fitted with the disease status with each of the risk scores to determine the W s. The regression coefficient of each risk score was used as the weight to indicate the contribution of each miRNA to the RSF. The receiver operating characteristic curves were then used to evaluate the predict effects of the profiling and to find an appropriate cutoff point. Validation of the procedure and the cutoffs were performed in the test sample [9]. All the statistical analyses were performed using the statistical analysis system software (v.9.1.3; SAS Institute, Cary, North Carolina, USA). Data are presented as median±SD. A P-value less than 0.01 was considered statistically significant.
with a mean fold change of at least 2 and a P-value less than 0.01 were selected for further study. MiRNAs with an expression level lower than the threshold value (Cq > 35) and with a detection rate below 75% in both group samples were discarded. A genome-wide expression profiling of serum miRNAs obtained by TLDA technology showed that serum from responders had 35-miRNA differential expression as compared with serum from the nonresponders, as shown in Supplementary Table 1 (Supplemental digital content 1, http://links.lww.com/ACD/A50). Individual miRNA was considered to be differentially expressed only if there was a larger than two-fold change in its expression level between the responders and nonresponders to chemotherapy. On the basis of these criteria, 14 miRNAs were found to be differentially expressed between the two groups, analysed by qRT-PCR for individuals of the 20 pool samples, as shown in Supplementary Table 1 (Supplemental digital content 1, http://links.lww.com/ACD/A50). The miRNAs were further analysed by qRT-PCR for individuals. These differentially expressed serum miRNAs were further examined by hydrolysis probe-based RT-qPCR in a training sample set including 20 responders and 20 nonresponders. In this phase, only those miRNAs with a mean fold change greater than 2 and a P-value less than 0.01 were retained. This phase generated a panel of five miRNAs (miR-20a, miR-130, miR-145, miR-216 and miR-372) that were significantly altered between the two group samples. In the validation set, the concentration of these five selected miRNAs was examined by qRT-PCR in a larger sample comprising 93 responders and 80 nonresponders to chemotherapy (Table 2). The miRNA expression pattern alterations in the validation set were consistent with the training set. The expression levels of the five miRNAs were significantly different between the responders and nonresponders. The differential expression of the five miRNAs is shown in Fig. 2.
Results Description and clinical features of the patients
All 253 patients enroled in this study were diagnosed with colorectal adenocarcinoma clinically and pathologically. As shown in Table 1, there were no significant differences in the distribution of smoking (P = 0.734), alcohol consumption (P = 0.953), age (P = 0.542) and sex (P = 0.312) between the two groups of patients. All patients underwent chemotherapy, on the basis of concurrent application of modified folfox6 (oxaliplatin 85 mg/m2, day 1; 5fluorouracil as a continuous intravenous infusion at 46 h dose 2400 mg/m2; 5-fluorouracil intravenous infusion at 30 min dose 300 mg/m2, day 1; CF 400 mg/m2 intravenous infusion at 2 h, day 1). All schedules were repeated every 14 days. Biomarker selection and validation phase
In the initial screening phase, serum samples were pooled from 20 responders and 20 nonresponders. The miRNAs
Separation of responders and nonresponders to chemotherapy by risk score analysis
Risk score analysis based on the five-miRNA expression profile was used to distinguish serum samples of the responders and nonresponders. First, the risk score of each serum sample was calculated, and on the basis of their scores, serum samples were then divided into a high-risk group, representing the possible nonresponder cases, and a low-risk group, representing the responder cases. At the optimal cutoff value of 3.46 with the value of sensitivity + specificity considered to be maximal, the positive predictive value and negative predictive value of the five-serum miRNA signature in the training set were 0.86 and 0.89, respectively. Similarly, when the same cutoff point was used to analyse the validation set, the positive predictive value and negative predictive value of the five-serum miRNA signature were 0.93 and 0.94, respectively (Table 3).
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
MicroRNA expression and colorectal cancer Zhang et al. 349
Overview of miRNAs with significantly different levels of expression in colorectal tumours of responders and nonresponders to chemotherapy (median±SD) (fmol/l)
Table 2
Training set
miR-20a miR-130 miR-145 miR-216 miR-372
Validation set
Responders (N = 20)
Nonresponders (N = 20)
Fold change
P-value
Responders (N = 93)
Nonresponders (N = 80)
Fold change
P-value
143.4±76.7 105.9±57.2 171.4±72.6 121.3±52.5 117.9±68.4
295.6±65.3 273.4±77.7 359.3±66.8 282.1±46.7 285.3±76.7
2.1 2.6 2.1 2.3 2.2
3.6 10–4 2.0 10–4 4.5 10–4 3.7 10–5 5.4 10–4
102.5±53.4 98.8±49.9 158.9±63.2 142.4±61.3 124.3±59.4
283.2±72.1 302.3±74.3 392.4±79.2 293.1±59.3 278.3±48.2
2.8 3.1 2.5 2.1 2.2
2.5 10–5 4.2 10–5 3.9 10–4 3.2 10–5 5.2 10–4
miRNAs, microRNAs.
Fig. 2
miR-20a ∗∗
(a)
Concentration (fM)
400
400
300
300
200
200
100
100 0
0 Responders (c)
Concentration (fM)
Nonresponders
miR-145 ∗∗
500 400
Responders
Nonresponders
miR-216
(d)
∗∗
400 300
300 200 200 100
100 0
0 Responder
Nonresponder
Responder
Nonresponder
miR-372 ∗∗
(e) 400 Concentration (fM)
miR-130 ∗∗
(b)
300 200 100 0 Responder
Nonresponder
(a–e) Detection of CRC using the five-serum miRNA profile as a biomarker. Serum levels of the five miRNAs were measured in 253 CRC patients (in both the training set and the validation set) using a hydrolysis probe-based qRT-PCR assay. **P < 0.01. CRC, colorectal cancer; miRNA, microRNA; qRT-PCR, quantitative reverse transcription PCR.
Receiver operating characteristic curves were then constructed to estimate the sensitivity and specificity of the five-serum miRNA signature. The areas under the curve (AUC) were 0.841 (95% confidence interval
0.707–0.975) and 0.918 (95% confidence interval 0.871–0.963) for the serum samples in the training set and validation set, respectively (Fig. 3a and b). Using the same serum samples, we compared the AUC of five
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
350
Anti-Cancer Drugs 2014, Vol 25 No 3
miRNAs with that of CEA and CA19-9, two currently available blood-based biomarkers for CRC detection. The AUC values of the five-serum miRNA signature were markedly higher than those of CEA (0.689) and CA19-9 (0.746) (Fig. 3c and d). The results indicate that the fiveserum miRNA signature is a more accurate biomarker than CEA and CA19-9 for CRC chemotherapy prediction. Table 3 Analysis of responders and nonresponders to chemotherapy by risk score
Training set Validation set
Score
0–3.46
3.46–13.23
PPV
NPV
Responders Nonresponders Responders Nonresponders
17 2 87 6
3 18 6 74
0.86
0.89
0.93
0.94
NPV, negative predictive value; PPV, positive predictive value.
Unsupervised cluster analysis
We also used the unsupervised cluster method to analyse the differential expression of miRNAs between the responders and nonresponders. The dendrogram generated by the cluster analysis showed a clear separation of the responders from the nonresponders on the basis of the five-serum miRNA profile (Fig. 3). In the training set, only one out of 20 responders and one out of 20 nonresponders were incorrectly classified (Fig. 4a). In the validation set, five nonresponders and four responders were misclassified (Fig. 4b).
Discussion The survival of patients with CRC has increased markedly in the last decade through the use of anticancer drugs in monotherapy or in combined regimens.
Fig. 3
ROC curve of training set (AUC = 0.841,95% CI 0.707−0.975)
ROC curve of CEA (AUC = 0.689,95% CI 0.618−0.760)
(c)
100
100
80
80
60
Sensitivity
Sensitivity
(a)
40 20
60 40 20
0
0 0
20
40 60 80 100% − specificity%
100
100
100
80
80
40 20
40 60 80 100% − specificity%
100
(AUC = 0.746,95% CI 0.682−0.851)
Sensitivity
Sensitivity
(AUC = 0.918,95% CI 0.871−0.963)
60
20
ROC curve of CA 19-9
(d)
ROC curve of validation set
(b)
0
60 40 20
0
0 0
40 20 60 80 100% − specificity%
100
0
60 40 80 20 100% − specificity%
100
(a, b) Receiver operating characteristic (ROC) curves for the five-serum miRNA profile to distinguish responders and nonresponders to chemotherapy in the training set (a) and the validation set (b). (c, d) ROC curves for CEA (c) and CA19-9 (d) to distinguish CRC serum samples in responders and nonresponders. AUC, area under the curve; CA19-9, carbohydrate antigen19-9; CEA, carcinoembryonic antigen; CI, confidence interval; CRC, colorectal cancer; miRNA, microRNA.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
MicroRNA expression and colorectal cancer Zhang et al. 351
Fig. 4
N–Responders10 N–Responders15 N–Responders9 N–Responders5 N–Responders13 N–Responders12 N–Responders3 N–Responders14 N–Responders11 N–Responders1 Responders14 N–Responders8 N–Responders4 N–Responders7 N–Responders20 N–Responders6 N–Responders19 N–Responders18 N–Responders17 N–Responders16 Responders13 Responders12 Responders3 Responders8 N–Responders2 Responders1 Responders4 Responders2 Responders9 Responders19 Responders10 Responders11 Responders6 Responders5 Responders7 Responders20 Responders18 Responders16 Responders15 Responders17
(a)
miR-216 miR-20a miR-145 miR-372 miR-130
(b)
Cluster analysis of miRNA differential expression between responders and nonresponders to chemotherapy serum samples. The concentration of the selected five serum miRNAs in samples was normalized, mean-centred, clustered and plotted as a heat map. (a) The training set (20 responders and 20 nonresponders). (b) The validation set (93 responders and 80 nonresponders). miRNA, microRNA.
Although the rate of success is higher than for other gastrointestinal tumours, chemotherapy response rates for CRC remain disappointingly low, and side effects and development of chemoresistance are severe limitations to pharmacological therapy. A major unresolved problem is that no predictors of response to these treatments are available. MiRNAs constitute a robust regulatory network with post-transcription regulatory efficiency for almost onehalf of human coding genes, including important oncogenes, tumour suppressor genes and genes associated with invasion, dissemination and therapy resistance [10]. Alterations in miRNA expression profiles could be useful in the prediction of individual response to chemotherapy and could even be targets for therapies to increase chemosensitivity. Svoboda et al. [11] demonstrated that miR-215, miR-99a*, miR-196b, miR-450b-5p and let-7e are involved in response of CRC to chemoradiotherapy. Nishida et al. [12] found that expression of miR-10b is a potential indicator of chemosensitivity to the common 5-FU-based chemotherapy regimen. Bitarte et al. [13] found that miR-451 is a novel candidate to circumvent recurrence and drug resistance in CRC and could be used as a marker to predict response to irinotecan in patients with colon carcinoma. Although tissue miRNAs can act as a predictor of chemotherapy for CRC, the collection of tissue samples is an invasive procedure and depends on surgical sections after the initial clinical classification. Recently, serum and plasma miRNAs have emerged as potential new blood-based markers for detecting cancers and other diseases [8,14–18]. In this study, the expression levels of serum miRNAs in patients with CRC have
been systematically determined and five serum miRNAs (miR-20a, miR-130, miR-145, miR-216 and miR-372) have been identified that are significantly upregulated in oxaliplatin-chemoresistant CRC patients compared with chemosensitive patients. We analysed global miRNA expression profiles in CRC serum to find potential predictive miRNAs for response to oxaliplatin-based chemotherapy. The discovery of serum miRNAs as a potential chemotherapy predictor can overcome the problems that exist in collecting tissue samples by an invasive process. Identification of a serum miRNAbased prediction biomarker from a genome-wide serum miRNA expression profile demonstrates that a combination of multiple serum miRNAs is a more comprehensive indicator for tumour chemotherapy than the conventional single protein-based or carbohydrate molecule-based biomarkers. This approach involves a screen of a pooled serum sample followed by individual RT-qPCR validation. Our results showed that the sensitivity and specificity of CRC prediction by this five-miRNA biomarker were 92 and 88%, respectively. This demonstrated that the expression profile of the five serum miRNAs could serve as an accurate predictive biomarker for CRC. Functional study of miRNAs in tumours is also helpful for evaluating serum miRNAs as indicators for various types of cancer. Among the five serum miRNAs identified in CRC patients, many are involved in general tumourigenesis. For example, miR-145 has been identified as a tumour suppressor in oesophageal squamous cell carcinoma [19]. Moreover, increased expression of miR-20a has been observed in colon cancer, pancreatic cancer and
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
352 Anti-Cancer Drugs 2014, Vol 25 No 3
other cancers [20–22]. These differential expression serum miRNAs could not only be useful as curative effect prediction markers but could also serve as a target for the development of novel therapeutic strategies to overcome drug resistance and tumour growth. The mechanisms and intracellular signal in pathways involved in the miRNA-mediated effects are still subject to investigation, and future studies are required to identify the target genes of these five serum miRNAs and the mechanism that regulates their biogenesis. Moreover, their predictive value in a true clinical setting remains undefined, alongside the potential side effects that systemic miRNA-based therapies may induce. Additional investigation is required to further determine the molecular functions and mechanisms of miRNAs, as well as their applicability as diagnostic and therapeutic tools. In sum, we have found five serum miRNAs that have the ability to separate complete remission from progression of disease and thus could be potential serum biomarkers for predicting response to the oxaliplatin-based chemotherapy in CRC. Given that the expression level of these miRNAs is important for CRC chemosensitivity and maintenance, novel therapies could be developed via changing the expression level of these miRNAs. miRNAbased therapy and predictor biomarkers are still in their infancy, but show great promise in the replacement or supplementation of current prognostic and predictive markers and anticancer treatment.
Acknowledgements
5 6
7
8
9
10
11
12
13
14
15
16 17
Conflicts of interest
There are no conflicts of interest.
18
References 1 2
3
4
Siegel R, Naishadham D, Jemal A. Cancer statistics. CA Cancer J Clin 2013; 63:11–30. Goldberg RM, Sargent DJ, Morton RF, Fuchs CS, Ramanathan RK, Williamson SK, et al. A randomized controlled trial of fluorouracil plus leucovorin, irinotecan, and oxaliplatin combinations in patients with previously untreated metastatic colorectal cancer. J Clin Oncol 2004; 22:23–30. Sengupta D, Bandyopadhyay S. Topological patterns in microRNA-gene regulatory network: studies in colorectal and breast cancer. Mol Biosyst 2013; 9:1360–1371. Liu H, Du L, Wen Z, Yang Y, Li J, Wang L, et al. Up-regulation of mir-182 expression in colorectal cancer tissues and its prognostic value. Int J Colorectal Dis 2013; 28:697–703.
19
20
21
22
Wang S, Xiang J, Li Z, Lu S, Hu J, Gao X, et al. A plasma microRNA panel for early detection of colorectal cancer. Int J Cancer 2013 [Epub ahead of print]. Hermanek P. The superiority of the new International Union Against Cancer and American Joint Committee on Cancer TNM staging of gastric carcinoma. Cancer 2000; 88:1763–1765. Mattie MD, Benz CC, Bowers J, Sensinger K, Wong L, Scott GK, et al. Optimized high-throughput microRNA expression profiling provides novel biomarker assessment of clinical prostate and breast cancer biopsies. Mol Cancer 2006; 5:24. Chen X, Ba Y, Ma L, Cai X, Yin Y, Wang K, et al. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res 2008; 18:997–1006. Liu R, Zhang C, Hu Z, Li G, Wang C, Yang C, et al. A five-microRNA signature identified from genome-wide serum microRNA expression profiling serves as a fingerprint for gastric cancer diagnosis. Eur J Cancer 2011; 47:784–791. Slaby O, Svoboda M, Michalek J, Vyzula R. MicroRNAs in colorectal cancer: translation of molecular biology into clinical application. Mol Cancer 2009; 8:102. Svoboda M, Sana J, Fabian P, Kocakova I, Gombosova J, Nekvindova J, et al. MicroRNA expression profile associated with response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer patients. Radiat Oncol 2012; 7:195. Nishida N, Yamashita S, Mimori K, Sudo T, Tanaka F, Shibata K, et al. MicroRNA-10b is a prognostic indicator in colorectal cancer and confers resistance to the chemotherapeutic agent 5-fluorouracil in colorectal cancer cells. Ann Surg Oncol 2012; 19:3065–3071. Bitarte N, Bandres E, Boni V, Zarate R, Rodriguez J, Gonzalez-Huarriz M, et al. MicroRNA-451 is involved in the self-renewal, tumorigenicity, and chemoresistance of colorectal cancer stem cells. Stem Cells 2011; 29:1661–1671. Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, PogosovaAgadjanyan EL, et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci USA 2008; 105:10513–10518. Resnick KE, Alder H, Hagan JP, Richardson DL, Croce CM, Cohn DE. The detection of differentially expressed microRNAs from the serum of ovarian cancer patients using a novel real-time PCR platform. Gynecol Oncol 2009; 112:55–59. Cho WC. Circulating microRNAs as minimally invasive biomarkers for cancer theragnosis and prognosis. Front Genet 2011; 2:7. Ng EK, Chong WW, Jin H, Lam EK, Shin VY, Yu J, et al. Differential expression of microRNAs in plasma of patients with colorectal cancer: a potential marker for colorectal cancer screening. Gut 2009; 58:1375–1381. Wang J, Chen J, Chang P, LeBlanc A, Li D, Abbruzzesse JL, et al. MicroRNAs in plasma of pancreatic ductal adenocarcinoma patients as novel blood-based biomarkers of disease. Cancer Prev Res (Phila) 2009; 2:807–813. Kano M, Seki N, Kikkawa N, Fujimura L, Hoshino I, Akutsu Y, et al. Mir-145, mir-133a and mir-133b: tumor-suppressive miRNAs target fscn1 in esophageal squamous cell carcinoma. Int J Cancer 2010; 127:2804–2814. Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F, et al. A microRNA expression signature of human solid tumors defines cancer gene targets. Proc Natl Acad Sci USA 2006; 103:2257–2261. Zeng X, Xiang J, Wu M, Xiong W, Tang H, Deng M, et al. Circulating mir-17, mir-20a, mir-29c, and mir-223 combined as non-invasive biomarkers in nasopharyngeal carcinoma. PLoS One 2012; 7:e46367. Wang YX, Zhang XY, Zhang BF, Yang CQ, Chen XM, Gao HJ. Initial study of microRNA expression profiles of colonic cancer without lymph node metastasis. J Dig Dis 2010; 11:50–54.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
