MiR-99b-5p expression and response to tyrosine kinase inhibitor treatment in clear cell renal cell carcinoma patients

A number of treatments targeting VEGF or mTOR pathways have been approved for metastatic clear cell Renal Cell Carcinoma (ccRCC), but the majority of patients show disease progression after first line therapy with a very low rate of complete or long-term responders. It has been shown that miRs may play a role in prediction of treatment response in various cancer types. The aim of our study was to identify a miR signature predictive for RCC patients' response to antiangiogenic tyrosine kinase inhibitor (TKI) treatment in the first line therapy. Sequencing of 40 paired normal/tumor formalin fixed and paraffin embedded ccRCC tissues revealed separate clustering via unsupervised dendrograms. With supervised analysis, the strongest differential expression was obtained with miR-99b-5p, which was significantly lower in patients with short progression free survival (<8 months) and TKI non-responders (progressive disease patients according to RECIST) (p<0.0001, each). Validation using RTqPCR and a second patient cohort compiled from three different hospitals (n=65) showed higher expression of miR-99b-5p in complete responders, but this trend did not reach statistical significance. It is concluded that low miR-99b-5p expression analyzed with sequencing methodology may correlate with tumor progression in TKI-treated ccRCC patients.


Exploratory sequencing step
Studies aimed at the identification of potential markers of TKI response in ccRCC patients included the sequencing of tumor and adjacent non-tumor tissues from 40 ccRCC patients (20 cases from Zurich cohort, 20 cases from the Vienna cohort). Patients were selected for this step based on the following criteria: i) availability of the PFS parameter (at the date of analysis); ii) reliable RECIST classification to represent each RECIST group: PD, SD, PR, CR; Moreover, miRs' expression was evaluated using the Vmatch software for sequencing data analysis with an application of miR v20 database (results of the Vmatch analysis are presented in supplementary table 6).
In order to define the control miRs' expression sequencing data were normalized with sum of reads (SOR) and relative value of tumor vs. non-tumor tissue was evaluated. For all but two patients, patient 10 and patient 16, the deregulation pattern was as expected (supplementary figure 4). To further validate the results for the aforementioned two patients, RTqPCR was performed. The results obtained with this method for patient 16 were consistent with the sequencing data. The down regulation of miR-155-5p in tumor vs non-tumor adjacent tissue was also observed (0.82-fold) (supplementary figure 5).
Therefore the results for patient 16 were regarded as valid. In contrast to the sequencing results, up regulation of miR-21-5p by RTqPCR was observed (12.2-fold) for patient 10. Due to the contradictory results of the RTqPCR and sequencing platform as well as due to low read number patient 10 was excluded from further analysis. ccRCC tissue testing using control miRs The general description and sequence of the experimental procedures are presented in Supplementary figure 6. To determine FFPE and frozen tissue comparability, the tumor tissue of one patient was analyzed for the expression levels of five key miRs known to be deregulated in ccRCC [1][2][3][4][5][6][7][8][9] :

Pilot sequencing
Tumor and non-tumor tissues from three patients, classified as PR, SD or PD, were selected to adjust the library preparation protocol to the FFPE samples for the pilot sequencing step. Samples were preceded according to the standard library preparation procedure (-R) 10 . In parallel for the same set of samples an additional ribosomal depletion step (+R) was applied in order to investigate its potential improvement in miR sequencing 11 . 488 miRs and 499 miRs out of 2044 investigated were detected with the -R and the +R version of the protocol, respectively. Since no significant improvement was observed after the ribosomal depletion step incorporation, the -R protocol was selected for the main sequencing step.  figure 7). Indeed, re-evaluation of histology revealed that the tissue was non-tumorous tissue.

Serum normalization method assessment
In order to define most reliable control miR we have evaluated 5 different miR candidates chosen based on the literature communications 12,13 . MiR-16 is the most often referred reference miR in the serum application 12 . MiR-191-5p used by some authors as a serum reference miR, proved to be the most stable in our data cohort if the sequencing data were analyzed. As an externally added and not expressed by human technical quality reference gene Caenorhabditis elegans miR-39 (cel-39), a spike in control, was added to the serum sample and used as a technical quality control miR. Moreover 2 miR were selected for the hemolysis control 13 : miR-451a and miR-23a-3p. MiR-23a-3p is hemolysis independent since it is not expressed in the red blood cells (RBC) and therefore no change was anticipated independent on the hemolysis intensity. On the contrary, miR-451a is highly expressed in RBC and therefore the increase of this miR in serum is expected with the higher hemolysis levels. In order to define the level of hemolysis in each sample the ∆Ct was calculated as presented in the formula below: ∆Ct = Ct miR-23a-3p -Ct miR-451a Providing the samples showed no hemolysis ∆Ct ≤5, the range of 5< ∆Ct ≤7 indicated low risk of hemolysis and ∆Ct >7 designated high risk of hemolysis.
For the validation purpose 4 randomly selected donors were chosen. Each miR was measured in triplicate in 4 independent experiment setups.
The results obtained (supplementary figure 9) indicated that the most stable miR was hsa-miR-191-5p that confirmed the sequencing results. Moreover, we proved that the experiments were performed with reliable and stable technical standard if based on the cel-39 results. Therefore, for the final miR evaluation in serum miR-191 should be used as an internal control, cel-39 as a technical quality control miR and a combination of miR-451a and miR-23a-3p as a hemolysis control miRs.
Supplementary Figure 1: Venn diagram summarizing RECIST differential expression analysis and PFS correlation analysis, each in tumor and normal tissue. Supplementary