Approximately 20% of LARA’s are primary resistant tumors in which there is no reduction or just minimal reduction in the number of cancer cells after neoadjuvant therapy based on concurrent application of radiation and chemotherapy with fluoropyrimidines
[14–16]. Should there be valid predictive factors, these patients would be spared exposure to chemotherapy and radiation associated with substantial adverse effects and costs and surgery could be scheduled without delay.
MiRNAs constitute a robust regulatory network with post-transcription regulatory efficiency for almost one half of human coding genes, including important oncogenes, tumor suppressor genes, and genes associated with the invasion, dissemination, and therapy resistance of many tumors
. We have used a large-scale miRNA expression profiling and identified 8-miRNAs signature (miR-215, let-7e, miR-196b, miR-190b, miR-29b-2*, miR-450a, miR-450b-5p and miR-99a*) enabling correct classification of 90% of responders (9/10) and non-responders (9/10) to neoadjuvant chemoradiotherapy in patients with LARA. In 5 (miR-215, miR-99a*, miR-196b, miR-450b-5p and let-7e) out of 8 identified miRNAs a potential linkage was already established between their expression and radioresistance or chemoresistance to thymidylate synthetase (TS) inhibitors.
The most frequently studied and the most promising miRNA identified in our study is miR-215, because current knowledge partly enables mechanistical explanation of its association with chemoradioresistance. Song et al. provides direct evidence that miR-215 regulates the TS mRNA in HCT116 colon cancer cell line. Ectopic expression of miR-215 decreased the expression of TS mRNA and protein and at the same time miR-215 inhibits cell proliferation and increased chemoresistance to TS inhibitor raltitrexed. Inhibition of cell proliferation and subsequent chemoresistance was caused by the induction of G2-arrest
. Similar results were published by Boni et al., who used in their experiments 5-fluorouracil
. Cell cycle arrest as a result of increased expression of miR-215 was confirmed on model of colorectal cancer also by another independent study
 and our previous results
. A further and very important discovery was the fact that miR-215 regulates the cell cycle not only in colon cancer cells but also in colon cancer stem cells. Recent data suggest that colon cancer stem cells may utilize miR-215 to slow cell proliferation and avoid damage caused by chemotherapy and radiotherapy until receiving a proliferation and differentiation signal
Observations on in vitro models were confirmed in clinical practice as well. Karaayvaz et al. showed that high levels of miR-215 expression in cancer tissues are closely associated with poor overall survival of patients with colon cancer in stage II and III (HR 3.516; P=0.025)
. These results are in agreement with our results showing down-regulation of miR-215 in LARA responding to chemoradiotherapy.
The let-7 family of miRNAs (let-7a through let-7h) regulates expression of key oncogenes, such as RAS and MYC, and is specifically down-regulated in many cancer types. Weidhaas et al., reported that the let-7 family is over-represented in a class of miRNAs exhibiting altered expression in response to radiation. Using C. elegans–based in vivo model of radiation-induced reproductive cell death, they confirm the ability of let-7 family of miRNAs to increase radiosensitivity when over-expresed
. Accordingly, we have observed up-regulation of let-7e in tumors of therapy responders.
Another predictive miRNA identified in our study, miR-99a*, was previously observed in work of Bandres et al. as up-regulated in tumors of responders to neoadjuvant chemoradiotherapy
. Interestingly, among putative targets of miR-99a* are important proteins involved in DNA repair – RAD51C and RAD9B
. Up-regulation of miR-99a* in tumors could be associated with lower DNA repair capacity through down-regulation of these genes, which may lead to radiotherapy sensitization. One of the putative targets of miR-450b-5p, another up-regulated miRNA in tumors of patients responding to therapy, is a gene coding the X-linked inhibitor of apoptosis protein (XIAP). It was shown that knockdown of XIAP in vitro lead to sensitization of colon cancer cells to irradiation
. In case of miR-196b, our research group published a study in which we established that elevated expression of miR-196b was positively correlated with overall survival (HR 0.5470; P = 0.0492) in glioblastoma patients treated with concomitant chemoradiotherapy
Till now, there are only two studies that analyzed global miRNA expression profiles in LARA to find potential predictive miRNAs for response to neoadjuvant chemoradiotherapy. The first study was performed by Scarpati et al.
. They have found 14 miRNAs (miR-1183, miR-483-5p, miR-622, miR-125a-3p, miR-1224-5p, miR-188e5p, miR-1471, miR-671-5p, miR-1909, miR-630, miR-765, miR-1274b, miR-720, hsv1-miR-H1) differentially expressed between group of tumors in which, after neoadjuvant chemoradiotherapy, a complete remission occured (TRG1/pCR) vs. control group that contained all other tumors (TRG2, TRG3, TRG4). There is no overlap between miRNA signature identified in this study and our results. We suppose that the cause of this discrepancy is in a different and, in part, also unsuitable design of Scarpati’s study, as 55% of tumors in control group reached expressive regression of cancer cells classified by TRG2 grade. This evaluation practically means that there was almost complete remission of tumor cells. Moreover, patients in their study received only dose of 45 Gy and as chemotherapy oxaliplatine was used. From biological point of view, it can be expected that tumors in which total or almost total regression was achieved after neoadjuvant chemoradiation therapy, would share same or similar gene profiles and protein expression. It can also be expected that if tumors in Scarpati study were given radiation dose of 50 Gy, most of the tumors in group TRG2 would reach total eradication of tumor cells (TRG1/pCR).
Second study was published in the form of an abstract by Bandres et al. on ASCO 2012 Annual Meeting (28). Bandres’ and our studies are much more similar since we used the same therapy procedures for patients with LARA and we also used similar designs for the study. Bandres et al. identified a miRNA signature that correctly differentiated extreme-phenotype of responders (TRG1) and non-responders (TRG4). They found up-regulation of miR-21*, miR-99*, miR-125b, miR-125b1*, let-7c and miR-490 to be significantly correlated with a higher likelihood of achieving TRG1/TRG2 response, and down-regulation of miR-21* and miR-125a-3p to be associated with a TRG-4 response. Our results are in agreement at least in case of miR-99* and let-7 family. As far as the rest of miRNAs is concerned, only in case of miR-21* our data showed a difference between group of responders and non-reponders indicating trend but have not reached statistical significance (p=0,11).