Identification of Serum MicroRNA-21 as a Biomarker for Chemosensitivity and Prognosis in Human Osteosarcoma

Abstract

Objective:

The aim of this study was to investigate the serum level of microRNA (miR)-21 in patients with osteosarcoma and its correlation with chemosensitivity and prognosis.

Methods:

miR-21 levels in sera from 65 patients with osteosarcoma and 30 healthy controls were measured by real-time reverse transcription-polymerase chain reaction. Correlations between serum miR-21 and clinicopathological features in patients with osteosarcoma were determined. The prognostic significance of serum miR-21 was assessed using a Cox proportional hazards model.

Results:

The serum level of miR-21 was significantly higher in patients with osteosarcoma than in control subjects. High serum miR-21 was significantly correlated with advanced Enneking stage and chemotherapeutic resistance. Univariate and multivariate analyses for overall survival showed that upregulation of serum miR-21 was an independent, unfavourable prognostic factor for patients with osteosarcoma (hazard ratio, 2.325).

Conclusions:

miR-21 might be a good candidate for a therapeutic target, and a potential biomarker for the prediction of chemotherapeutic sensitivity and prognosis in patients with osteosarcoma.

Introduction

Osteosarcoma is the most common human primary malignant bone tumour in children and young adults. ¹ The tumour is located most often in the distal femur, the proximal tibia or the proximal humerus. ² Despite recent advances in therapeutic strategies such as wide tumour excision, adjuvant chemotherapy and radiotherapy, ∼80% of patients develop metastatic disease after surgical treatment ³ and 5-year survival rates are unsatisfactory. ² Although several oncogenes and tumour suppressor genes have been reported to be involved in the genesis of osteosarcoma, ⁴ the precise molecular mechanisms are still unclear and the prognosis in individual cases remains problematic. It would be helpful if objective means were available for predicting the chance of survival or response to chemotherapy, especially early in treatment and preferably before surgery.

MicroRNAs (miRNAs) are a class of small noncoding RNAs (18 – 24 nucleotides in length) that regulate a variety of cellular processes including cell differentiation, cell cycling and apoptosis.^5,6 Expression of miRNAs is altered in a variety of cancers, suggesting that miRNAs have the potential to be useful diagnostic and prognostic tools. ⁷ Several studies have found that the expression of specific miRNAs contributes to tumour growth, progression, metastasis and drug resistance.^{7 – 9} Furthermore, miRNAs have also been detected in human serum and plasma in remarkably stable forms, ¹⁰ raising the possibility that unique plasma/serum miRNA patterns might be used as noninvasive disease markers. In support of this possibility, differences between miRNA patterns have been found in sera or plasma from patients with several different malignancies, and in healthy controls. ¹¹

Among the large number of miRNAs that are known, miR-21 has received special attention because of its relationship with malignancies. High levels of miR-21 have been observed in gastric cancer, ¹² pancreatic cancer, ¹³ colorectal cancer, ¹⁴ breast cancer, ¹⁵ prostate cancer, ¹⁶ brain tumours, ¹⁷ ovarian cancer, ¹⁸ cholangiocarcinoma, ¹⁹ lung cancer, ²⁰ oesophageal cancer, ²¹ head and neck cancer ²² and hepatocellular cancer. ²³ Research has also demonstrated that miR-21 may play an important role in resistance to cancer drugs at the cellular level. ¹⁹ To our knowledge, however, no reports have focused on the serum miR-21 level in human osteosarcoma and its prognostic significance.

Here, the serum level of miR-21 was measured in patients with osteosarcoma and its correlations with clinicopathological features were assessed. The study also investigated whether serum miR-21 might be a useful molecular biomarker for predicting chemosensitivity and prognosis in osteosarcoma.

Patients and methods

Study Population

Patients with primary osteosarcoma, treated in the Department of Orthopaedics at The First Affiliated Hospital of Wenzhou Medical College, Wenzhou, China, between January 2005 and December 2007, were enrolled retrospectively in this study. Inclusion criteria were: no metastasis at presentation; no history of previous treatment; tumour excision with a wide or radical margin (histologically defined according to the system of Enneking ²⁴ ); adequate pre operative and postoperative chemo therapy (modified T12 protocol [high-dose methotrexate, cisplatin, adriamycin, bleomycin/cyclo -phosphamide/dactinomycin] ²⁵ ). Patients were excluded if they had a previous or secondary malignancy.

Serum samples (5 ml) were collected by peripheral venous puncture after surgery. Clinicopathological features were recorded at the time of serum sampling and tumours were staged according to Enneking's surgical staging system. ²⁶ Serum samples from healthy age- and sex-matched volunteers from the local population (Wenzhou, Zhejiang Province, China) were used as controls. Immediately after blood sample collection, each sample was centrifuged at 1200 g (KA-2200, Seromatic™; Kubota, Tokyo, Japan) for 10 min at 4 °C, and the serum-containing supernatant was transferred to a fresh tube and stored at -80 °C until further processing.

Written informed consent was obtained from all study participants. This study was approved by the Institutional Review Board of The First Affiliated Hospital of Wenzhou Medical College.

RNA Extraction and Quantitative Real-Time RT–PCR

Total RNA was isolated from 400 μl of serum using the mirVana™ kit (Ambion^®, Life Technologies, Paisley, UK) according to the manufacturer's instructions. The concentration of miR-21 in serum was determined with the quantitative real-time reverse transcription–polymerase chain reaction (RT–PCR) using the TaqMan^® miRNA assay system (Applied Biosystems, Foster City, CA, USA). To synthesize cDNA, 10 ng of total RNA from each serum sample was used for individual assays in a 15-μl reaction mixture containing 5 μl of RNA extract, 0.15 μl of 100 mM of each deoxynucleotide triphosphate, 1 μl of MultiScribe^® reverse transcriptase (50 U/μl), 1.5 μl of 10 × RT buffer, 0.19 μl of RNase inhibitor (20 U/ml), 1 μl of gene-specific TaqMan^® primer and 4.16 μl of nuclear-free water. The RT primers (Invitrogen, Carlsbad, CA, USA) were 5′-UAGCUUAUCAGACUGAUGUUGA-3′ (for miR-21) and 5′-CGTTCACGAATTTGCGT GTCAT-3′ (for RNA U6). The reaction mixture was incubated at 16°C for 30 min, 42 °C for 60 min and 85 °C for 5 min. Subsequently, 5 μl of the DNA template was amplified using 10 μl of LightCycler^® 480 Probes Master (Roche Diagnostics, Mannheim, Germany), 3 μl of nuclear-free water and 2 μl of gene-specific TaqMan^® primer/probe mix in a final volume of 20 μl. Quantitative real-time RT–PCR was run on the LightCycler^® 480 System II (Roche Diagnostics). The cycling programme comprised preliminary denaturation at 95 °C for 5 min followed by 40 cycles of denaturation at 95 °C for 10 s, annealing at 60 °C for 30 s and elongation at 72 °C for 1 s, followed by a final elongation step at 72 °C for 10 min. All samples were processed in triplicate. The quantitative endpoint for realtime PCR, threshold cycle (CT), was defined as the cycle number at which fluorescence passed a fixed threshold, determined automatically by the LightCycler^® system. Small nuclear RNA U6 (snRNA U6) was the internal control and was used to normalize miR-21 values in the PCR. The level of miR-21 relative to snRNA U6 was calculated using the equation 2^-ΔCT, where ΔCT = (CTmiR-21 -CTU6) ²⁷ and the ΔCT value was negatively correlated with the serum level of miR-21. ²⁸

Evaluation of Response to Chemotherapy

Response to chemotherapy was evaluated pathologically by examining resected specimens obtained at the time of osteosarcoma surgery (after two cycles of neoadjuvant chemotherapy). After fixation in 10% formalin and standard paraffin-wax embedding, all specimens were cross-sectioned serially at 0.5-cm intervals. The grading system described by Rosen et al. ²⁹ was used to assess the extent of tumour destruction resulting from chemotherapy: grade 1 indicates no evidence of necrosis; grade 2 indicates areas of necrotic material with other areas of histologically viable tumour; grade 3 indicates only scattered foci of viable tumour; grade 4 indicates no viable tumours seen in extensive sampling. All patients with grade 3 or 4 regression were considered to be pathological responders, while those with grade 1 or 2 regression were considered to be pathological nonresponders.

Statistical Analyses

All statistical analyses were performed with the SPSS^® statistical package, version 17.0 (SPSS Inc., Chicago, IL, USA) for Windows^®. Values were expressed as mean ± SD. Differences in serum miR-21 levels between groups were compared using Student's t-test. Survival time was calculated from the date of the first neoadjuvant chemotherapy to the date of death or the last follow-up visit. The association between levels of miR-21 and overall survival was analysed using Spearman's rank correlation coefficient. The joint effect of covariables was examined using the Cox proportional hazards regression model. All tests were two-tailed and a P-value < 0.05 was considered to be statistically significant.

Results

Peripheral blood samples from 65 patients (37 male, 28 female; mean age 20.4 years; range 9 – 68 years) with primary osteosarcoma were collected between January 2005 and December 2007. The locations of the tumour were the femur (38 cases), tibia (21 cases), proximal humerus (five cases) and pelvic bone (one case). The histological subtypes were osteoblastic in 53 patients and other subtypes in 12 patients. Nineteen cases were Enneking stage I, 26 cases were stage IIA and 20 cases were stage IIB. ²⁶ Blood samples were also obtained from 30 healthy controls during the same period (17 males, 13 females; mean age 19.2 years; range 9 – 69 years). None of the healthy controls had a history of any type of malignancy.

Using snRNA U6 as the internal reference, the serum level of miR-21 was detected by quantitative real-time RT–PCR. Serum miR-21 concentrations were increased in patients with osteosarcoma. The ΔCT value of miR-21 in the serum of patients with osteosarcoma was 4.95 ± 0.61, compared with 7.52 ± 0.86 in healthy controls (P < 0.001, Student's t-test).

Patient characteristics with respect to serum miR-21 levels are shown in Table 1. Increased serum miR-21 levels were significantly correlated with Enneking stage (P = 0.018) No correlation was demonstrated between serum miR-21 levels and age, sex, tumour location or tumour subtype.

Table 1:

Relationship between serum microRNA-21 (miR-21) level and clinicopathological features in patients with osteosarcoma (n = 65)

Clinicopathological feature	n	miR-21 level (ΔCT) a	Statistical significance b
Age, years
< 18	44	4.63 ± 0.58	NS
≥ 18	21	5.07 ± 0.64
Sex
Male	37	4.74 ± 0.49	NS
Female	28	4.28 ± 0.51
Tumour location
Femur or tibia	59	5.15 ± 0.60	NS
Others	6	4.80 ± 0.43
Subtype
Osteoblastic	53	5.24 ± 0.76	NS
Others	12	4.88 ± 0.36
Enneking stage
I	19	5.93 ± 0.72	P = 0.018
II	46	4.05 ± 0.55
Tumour response
Responder	35	6.19 ± 0.80	P < 0.001
Nonresponder	30	3.92 ± 0.33

Data presented as mean ± SD.

a

ΔC_T = C_TmiR-21 - C_TU6.

b

Student's t-test.

NS, no statistically significant difference (P ≥ 0.05).

After resection, a total of 35 patients were pathological responders (grade 3 or 4 tumour regression after chemotherapy) and 30 were nonresponders (grade 1 or 2 regression) (Table 1). The mean ΔCT value of miR-21 in the pathological responder group was significantly higher than that in the nonresponder group (P < 0.001).

Nineteen of the 65 patients were alive at the time of the scheduled analyses (January 2011). The median follow-up time was 46 months (range 13 – 82 months). Spearman's rank correlation coefficient analysis showed a significant correlation between serum miR-21 levels (ΔCT) and survival time (r = 0.445; P = 0.007).

Table 2 presents the results of the univariate and multivariate Cox proportional hazards regression analyses of the relationship between serum miR-21 levels, clinicopathological characteristics of osteosarcoma and prognosis. Univariate analysis revealed that serum miR-21 levels (hazard ratio [HR] 2.643; P = 0.019), Enneking stage (HR 3.489; P = 0.005) and histological response (HR 5.915; P < 0.001) were predictive factors for prognosis in osteosarcoma. Multivariate regression analysis confirmed that upregulation of miR-21 (HR 2.325; P = 0.026), advanced disease stage (HR 3.237; P = 0.012) and a poor tumour histological response (HR 4.196; P < 0.001) were independent unfavourable prognostic factors for overall survival.

Table 2:

Univariate and multivariate Cox regression analyses of the relationship between serum microRNA-21 (miR-21) level, clinicopathological characteristics and survival of patients with osteosarcoma

Variable	Subset	Hazard ratio	Statistical significance
Univariate analysis, n = 65
Age	≥ 18 versus < 18 years	0.668	NS
Sex	Male versus female	0.715	NS
Tumour location	Femur and tibia versus others	0.769	NS
Subtype	Osteoblastic versus others	0.677	NS
Enneking stage	I versus II	3.489	P = 0.005
Serum miR-21 levels	ΔCT a	2.643	P = 0.019
Histological response	Responder versus nonresponder	5.915	P < 0.001
Multivariate analysis, n = 65
Age	≥ 18 versus < 18 years	0.722	NS
Sex	Male versus female	0.576	NS
Tumour location	Femur and tibia versus others	0.494	NS
Subtype	Osteoblastic versus others	0.542	NS
Enneking stage	I versus II	3.237	P = 0.012
Serum miR-21 levels	ΔCT a	2.325	P = 0.026
Histological response	Responder versus nonresponder	4.196	P < 0.001

a

ΔC_T = C_TmiR-21 - C_TU6.

NS, no statistically significant difference (P ≥ 0.05).

Discussion

Research has demonstrated that circulating miRNAs might be used as potential molecular biomarkers for several disease conditions, including human malignancies. ³⁰ The present study reports for the first time that miR-21 was upregulated in osteosarcoma serum samples compared with normal control serum samples, and that a high serum miR-21 level was closely correlated with an advanced Enneking stage, a poor tumour response to neoadjuvant chemotherapy and a reduced overall survival rate. Data on the expression of miR-21 and its roles in diverse tumour entities have been reported,^{12 – 23} and the clinical significance of serum miR-21 expression in human cancers has also been described. For example, Asaga et al. ³¹ found that serum miR-21 concentrations had diagnostic and prognostic potential in breast cancer. Similarly, a study in patients with nonsmall cell lung cancer revealed that serum miR-21 was elevated, and that a high serum miR-21 concentration was significantly correlated with tumour–node–metastasis stage, lymph node metastasis and shorter survival time. ³² Elevated serum miR-21 levels were demonstrated in patients with hormone-refractory prostate cancer, suggesting its potential value in predicting the efficacy of docetaxel-based chemotherapy. ³³ These findings indicate that miR-21 may play an important role not only in tumour initiation but also in tumour development, and in the acquisition of drug resistance in many malignancies.

The mechanism by which miR-21 affects tumour behaviour has been investigated in a few published reports. These indicate that miR-21 can modulate growth, cell cycle progression, metastasis and chemosensitivity of tumour cells by targeting several tumour suppressor genes, including PTEN (phosphatase and tensin homolog protein), MARCKS (myristoylated alanine-rich protein kinase C substrate protein), PDCD4 (programmed cell death 4 protein) and CDC25A (cell division cycle 25 homolog A protein).^{34 – 37} Each miRNA can target more than 100 transcripts directly or indirectly; on the other hand, several miRNAs can converge on a single transcript target. ³⁷ As a consequence, the potential regulatory circuitry afforded by miR-21 may be enormous, and identification of the complex molecular network involved in its function will be an important focus of future investigations.

In conclusion, upregulation of miR-21 in the serum of patients with osteosarcoma was correlated with advanced Enneking stage, resistance to chemotherapy and poor survival. These findings indicate that miR-21 may be a good candidate as a therapeutic target, and a potential biomarker for the prediction of chemotherapeutic sensitivity and prognosis in patients with osteosarcoma.