If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
), the identification of novel blood biomarkers to monitor therapeutic response and disease recurrence is timely. MicroRNAs (miRNAs) are promising because they can be assayed directly from blood. Over 1,000 of these exist (
). Our hypothesis was that plasma miRNAs are biomarkers of melanoma burden. We used miR-21 as a proof of concept to test this because it has been widely studied in cancer.
We analyzed 160 melanocytic tumors (Supplementary Table S1 online) and 56 blood samples. First, miR-21 expression was measured in 51 melanomas using quantitative PCR, finding a significant association with Breslow thickness and ulceration, two important prognostic features (
), P=0.02, 0.024, respectively, Supplementary Table S2 online. To assess independent prognostic value, another set of 79 melanomas was analyzed (Supplementary Table S1 online), 40 having disease-free survival >5 years and 39 having metastasis within 5 years. Logistic regression showed that the stage (IB/IIA versus IIB/IIC) and miR-21 (dichotomized at median) both significantly predicted progression-free survival with an odds ratio of 4.83 (confidence interval (CI), 1.79–13.04), P=0.002 and 2.72 (CI, 1.01–7.34), P=0.048, respectively. The covariates explained between 16 and 21% of the total variation (Cox and Snell R2 and Nagelkerke R2, respectively). The addition of miR-21 to the American Joint Committee on Cancer (AJCC) stage increased the model accuracy (χ2=4.10, d.f.=1, P=0.043). These data suggest that tissue miR-21 has independent prognostic value. We next assessed miR-21 expression during tumor progression in 51 melanomas, 13 common nevi, and 11 congenital nevi using cultured melanocytes as calibrator (Figure 1a). Expression was significantly different (F=5.65, d.f.=2, P=0.005). Post hoc analysis revealed a trend of increasing expression from common nevi and congenital nevi to melanoma (F=11.05, d.f.=1, P=0.001). Colorimetric in situ hybridization confirmed tumor cell expression (Figure 1b–d). The relatively high expression in congenital nevi is intriguing, perhaps reflecting their increased risk of progressing to melanoma (
). We next looked at whether miR-21 expression in metastatic tumor tissue related to plasma levels, finding a strong correlation, n=5, r=0.997, P=0.0002 (Figure 2a). These data confirm that miR-21 is an important tissue biomarker in melanoma and that tissue expression reflects plasma level.
Figure 1MicroRNA-21 (miR-21) in melanocytic tumor tissue. (a) The relative expression of miR-21 in common acquired nevi (n=13), congenital nevi (n=11), and melanoma (n=51) tissue compared with neonatal normal melanocytes grown in culture. The mean for each group is shown. The results are expressed as ΔΔCT (see Supplementary methods online). (b) MiR-21 C-ISH in a nevus, which is negative. (c) Same nevus but C-ISH for U6, indicating good RNA integrity. (d) MiR-21 C-ISH in a melanoma showing strong staining. Bar in photomicrographs is 1mm. RT-PCR primers are shown in Supplementary Tables S3 and S4 online.
Figure 2MicroRNA-21 (miR-21) in plasma. (a) MiR-21 in paired melanoma and plasma samples. Results are expressed as relative expression; miR-21 expression was normalized to 18s ribosomal RNA to create a ΔCT. Relative expression as 2−ΔΔCT was calculated in relation to the ΔCT of normal human epidermal melanocytes. (b) MiR-21 in paired blood samples taken before and after removal of a melanocytic tumor. Results expressed as ΔCT, relative to miR-191 (see Supplementary methods online). (c) Six categories in which miR-21 was assessed: controls, cancer-free healthy people; N+DN, nevi and dysplastic nevi; MM DF, melanoma patients disease-free for a minimum of 3 years post excision of primary tumor; MM 0+I+II, melanoma patients with primary disease only, i.e., American Joint Committee on Cancer (AJCC) stage 0, I, or II; MM III, melanoma patients with AJCC stage III disease, i.e., regional metastases; and MM IV, melanoma patients with AJCC stage IV disease, i.e., distant metastases. (d) Ranked list of candidate plasma miRNAs after analysis of 740 mature miRNAs and endogenous controls in two matched RNA pools derived from preoperative and postoperative blood samples from patients having regional node resection for stage III melanoma; the list was filtered and ranked by preoperative CT<30 and a fold decrease of >5. (e) miR-211 levels in the same groups as panel c. (f) Cumulative z scores of miR-21 and miR-211, same groups as panel e.
We then looked at whether plasma miR-21 correlated with melanoma burden measured by the AJCC stage. We collected blood from 18 patients pending diagnostic excision whose subsequent histology showed benign nevus (n=2), dysplastic nevus (n=4), or stage 0–II primary melanoma (n=12). Blood was also obtained from 10 stage III and 4 stage IV melanoma patients. In addition, we obtained blood from four melanoma patients who had remained disease-free for at least 3 years post excision of their primary melanoma. Finally, we obtained paired preoperative and postoperative samples from nine of the above patients (one nevus, one dysplastic nevus, three primary melanomas, and four stage III melanomas). Postoperative samples were obtained 6 weeks after surgery to mitigate the effects of surgery. Eleven controls were healthy people undergoing screening colonoscopy. When preoperative and postoperative cases were compared, there was a drop in plasma miR-21 levels, t=2.91, d.f.=8, P=0.0195 (Figure 2b). Analysis of all samples from controls to advanced melanoma showed a positive trend of plasma miR-21 (F=14.32, d.f.=5, P<0.0001) (Figure 2c). This trend was also seen with the disease-free and active melanoma groups alone (F=10.57, d.f.=3, P<0.0001). These data suggest that the plasma miR-21 reflects tumor burden and therefore may have a role monitoring disease activity.
To maximize sensitivity and specificity, a panel of miRNAs is likely to be needed. Therefore, to identify additional candidate plasma miRNA biomarkers, we analyzed 740 mature miRNAs and endogenous controls using TaqMan Human Array Cards A and B, version 2, in two matched RNA pools. One pool comprised three preoperative blood samples from stage III melanoma patients and the other postoperative samples matched to the same patients. On the basis of the criteria of preoperative CT<30 and fold drop >5, miRNA-211 was top ranked (Figure 2d). Despite miR-211 being a putative tumor suppressor, there was a positive trend across all samples as well as in disease-free and active melanoma samples alone (F=14.79, d.f.=5, P<0.0001 and F=5.64, d=3, P=0.003, respectively) (Figure 2e). The tumor suppressor function of this miR-211 might be accounted for by its much lower absolute levels (i.e., higher CT values) compared with miR-21. MiR-21 and miR-211 levels were converted to z scores to normalize their measurement scales using the control sample mean and SD and then added together. The cumulative z scores had a positive trend among melanomas (F=8.57, d.f.=3, P<0.0001) (Figure 2f), but did not account for any more of the variation: R2 values for miR-21, miR-211, and miR-21/211 were 0.47, 0.25, and 0.46, respectively. A Supplementary data online file gives details of the methods. In summary, these data demonstrate that miR-21 in melanoma tissue is associated with important clinical parameters, has independent prognostic value, and correlates with plasma levels. We further show that plasma miR-21 correlates with melanoma burden and that miR-211 may be another useful plasma biomarker. Future work will require identification of miRNA combinations to enhance detection sensitivity and specificity, e.g., using large-scale assays (
36220-5&id=gr1.jpg){kind=link}
36220-5&id=gr2.jpg){kind=link}