Background
MicroRNAs (miRNAs, miRs) are a group of short, non-coding RNAs (~22 nucleotides in length) that have emerged as important (negative) regulators of gene expression. It has been shown that up to 100-200 mRNAs can be repressed by one miRNA [
1]. These molecules are considered key players in a variety of processes ranging from development, proliferation, morphogenesis and differentiation to cancer and apoptosis [
2,
3].
Roles of microRNAs in cancer development have been documented in several studies [
4,
5]. Typically, miRNAs involved in tumorigenesis are deregulated, and this deregulation is believed to alter the expression of protein-coding mRNA, thereby favoring uncontrolled tumor cell growth. The deregulation can be an under- or overexpression, suggesting that miRNAs may function as tumor suppressors or as oncogenes. The involvement of miRNAs in tumorigenesis is not the only topic of investigation. In addition the expression patterns of these regulators by cancer treatment modalities such as radiotherapy or chemotherapy are increasingly recognized. It has been shown for cancer cells that the expression of miRNAs may vary depending on parameters like cell type, post-radiation time and radiation dose [
6‐
8].
The tumor vessel system, and in turn endothelial cells as the characteristic parts of the vessel system, constitute critical targets for radiotherapy of tumors. However, to our best knowledge, the regulation of miRNAs in endothelial cells (EC) after radiation has not been investigated to date. EC are sensitive to ionizing radiation in proliferation and clonogenic assays in vitro and in vivo [
9] and may constitute critical targets in normal tissue such as in the gut microvasculature [
10]. In contrast, EC are also stimulated by radiation-induced indirect pro-angiogenic factor production including VEGF and bFGF [
9,
11]. Further ionizing radiation potently causes DNA damage, which has been shown to induce miRNA expression via the p53 network [
12]. Here we investigated the miRNA response in EC after ionizing radiation. To this end, human EC were irradiated and radiation-induced alterations of miRNA levels were analyzed by miRNA microarrays. The most stringently regulated miRNAs were then further analyzed. The effects of miRNA overexpression or inhibition were determined in functional assays including clonogenic assays with and without radiation in order to examine if the altered miRNA levels affected EC response to radiation.
Methods
Cell culture
Human dermal microvascular endothelial cells (HDMEC; PromoCell, Heidelberg, Germany) were cultured in modified PromoCell medium (for ref. see [
13]) for optimal growth results. Cells were cultured up to passage 7; for transfection cells of passage 3 to 5 were used.
Isolation of RNA
Cells were seeded in culture flasks until confluency of ~70% before 2 Gy photon irradiation (RT, 6 MeV; LINAC, Siemens). After RT they were transferred back to the incubator and after 6 hours lysed and stored at -80°C. Non-irradiated cells were used as controls. RNA was isolated from HDMEC using TRIzol LS reagent (Invitrogen, Karlsruhe, Germany) (3 biological replicates each for RT and control) as described by the manufacturers. Quality and quantity of isolated RNA were checked using Lab on Chip technology on Agilent 2100 bioanalyzer (Agilent technologies, CA, USA) and a Nanodrop spectrophotometer (ND-1000; Nanodrop technologies, DE, USA).
Locked nucleic acid (LNA) -based miRNA microarrays and data analysis
RNA samples from the three biological replicates were used for LNA-based array analysis. miRNA expression profiling was performed using a microarray platform which is based on locked nucleic acid (LNA)-modified capture probes which are immobilized on the chip surface. For detailed protocol and further details see Castoldi et al. [
14] and Exiqon (
http://www.exiqon.com). 361 miRNAs, including 315 human miRNAs were spotted in quadruplicates on the slides (see Additional file
1). Slides were scanned using the Genepix 4000B scanner (Axon instruments). Data analyses were done using 'Microsoft Excel' software and the 'SUMO' software package for microarray data evaluation (
http://www.oncoexpress.de/software/sumo/). For data normalization we developed a step-wise approach: First, normalization was performed on the background-subtracted mean intensity values against the intensity of the U6 snRNA spots on each chip. After this thresholding, the data underwent a two-class t-test. We then created a short-list of differentially expressed miRNAs as described in 'Results'. Microarray data were deposited in 'ArrayExpress' (accession no.: E-TABM-617).
Transfection
Transfection of primary HDMEC was performed using the siPORT Amine (Ambion, Texas, USA) transfection reagent. Transfection efficiency was analyzed using a GAPDH assay (KDalert, Ambion). The efficiency of transfection conditions was 30-50%. Furthermore a miR-1 transfection test system (Ambion) was used, which is known to down-regulate the PTK9 mRNA in human cells. Expression of PTK9 was measured by real-time PCR to verify our transfection conditions (see Additional file
2). In the experiments miRNA precursor (pre-miR) or inhibitor (anti-miR) molecules or the appropriate negative control molecules were added to the cells in a final concentration of 50 nM. The following pre- and anti-miRs were used: hsa-let-7g, hsa-miR-125a, hsa-miR-127, hsa-miR-148b, hsa-miR-189, hsa-miR-20a, pre-miR negative control #1, and anti-miR negative control #1 (all purchased from Ambion).
Clonogenic survival assay
HDMEC were pre-plated in 25 cm2 cell culture flasks; cell numbers varied depending on the treatment. In experimental settings with transfection medium and/or RT cell numbers were raised. After one day the transfection mixture was added for 6 hours, then cells were re-fed with normal growth medium. After 24 hours cells were irradiated with 2 Gy (6 MeV X-rays; LINAC, Siemens) and then returned to the incubator for 8-10 days. Untreated cells served as growth controls. For evaluation the number of counted colonies was normalized to the amount of pre-plated cells. At the end of the incubation period cells were stained with crystal violet (Sigma-Aldrich, Germany) and colonies were counted. All conditions were done in triplicate, the survival experiments for each miRNA were repeated three times.
Proliferation assay
The proliferation rate of cells was determined using a calcein assay (PromoKine, Heidelberg, Germany). The assay was performed in a 96 well-plate format. 2500 endothelial cells were seeded per well, after 1 day they were transfected for 6 hours, cultured with normal growth medium and incubated for another 24 hours. Then the cells were irradiated with 2 or 10 Gy - while controls were non-irradiated - and incubated for 3 days. Intracellular fluorescent calcein is directly proportional to the number of living cells and was measured using a plate reader (CytoFluor, PerSeptive Biosystems) with 485 nm excitation and 530 nm emission filters.
Statistical analysis
Statistical data evaluation was performed using two-tailed t-tests or in case of multiple comparisons using ANOVA along with Fisher's least significance difference test. The significance level was P < 0.05.
Discussion
The expression profiling by microarray showed that irradiation at the clinically relevant dose of 2 Gy induced significant changes in miRNA levels in human dermal microvascular endothelial cells (HDMEC). Six microRNAs (miRs) were chosen for subsequent functional analyses (let-7g, miR-125a, miR-127, miR-148b, miR-189, and miR-20a).
The proliferation and clonogenic assays documented that overexpression or inhibition of the here identified miRNAs is capable of reducing or enhancing endothelial cell proliferation and/or clonogenic survival. Moreover, we also found that overexpression or inhibition of selected miRNAs either enhanced or attenuated the radiation-induced reduction of clonogenicity or proliferation of HDMEC. This indicates that changes in radiation-induced miRNA expression alter the intrinsic functional cell properties and suggest that the radiosensitivity of endothelial cells itself is modified after irradiation. We conclude that radiation-induced miRNA expression alterations may play important roles in the desired and, potentially, also in side effects of radiotherapy of cancer and other applications of ionizing radiation.
One of the up-regulated miRNAs in response to radiation was a member of the let-7 family (let-7g). Other let-7 family members were not regulated upon radiation treatment (RT) or slightly up-regulated like let-7d and let-7f (see Additional file
3). A role of let-7 in cell growth has been described in normal and lung cancer cells, in which let-7 is down-regulated [
15]. Let-7 negatively regulates human Ras genes and it was reported that it is a negative regulator of cell proliferation pathways in human cells [
16]. An alteration in the expression of let-7 miRNAs in response to radiation was recently shown in human fibroblasts [
17]. Furthermore, a role of several let-7 miRNA family members for radiation sensitivity in lung cancer cells was reported by Weidhaas et al. The authors showed that overexpression of let-7g protected A549 cells from radiation. Corresponding to the let-7g up-regulation in our irradiated endothelial cells we found a reduction of clonogenic survival by overexpression of the miRNA. We found enhanced survival by inhibition of let-7g both in untreated and in irradiated cells. The data suggest that let-7g negatively regulates EC growth and furthermore sensitizes them to radiation. Since radiation up-regulates let-7g, the data also indicate that the miRNA up-regulation is correlatively or causatively associated with the direct anti-endothelial radiation effect as determined by clonogenic survival and proliferation inhibition.
miR-20a was also found to be up-regulated after radiation treatment. This microRNA sequence lies within the cluster miR-17-92, which is up-regulated in several human tumor types including lung, pancreas, prostate and colon cancer [
18]. Matsubara et al. could show that the inhibition of miR-20a can induce apoptosis in lung cancer cells over-expressing the miR-17-92 cluster [
19]. Furthermore, miR-20a is involved in cell cycle progression [
20]. Our own cell-based assays clearly show that miR-20a overexpression dramatically inhibits clonogenic survival, while the inhibition of the miRNA increased survival rates. Again, since radiation was clearly found to up-regulate miR-20a, this microRNA is another potential candidate in our system linking functional cell death with effects of radiation. Interestingly, our data showed that the radiation sensitivity itself in endothelial cells does not appear to be markedly dependent on the expression level of miR-20a.
The microRNAs miR-189, -125a, -127 and 148b were all found to be down-regulated after RT of 2 Gy. In the case of miR-189 the functional experiments revealed interesting opposite effects on cell growth with and without radiation: In survival assays we observed a strong decrease of clonogenic survival after overexpression of miR-189. In contrast, versus additional radiation as control sample, miR-189 over-expression increased clone number and miR-189 inhibition decreased clone number. These data suggest that miR-189 expression per se has negative effects on clonogenic survival and proliferation of endothelial cells. Radiation decreases the expression levels, which suggests that the anti-endothelial effects are associated with down-regulated miRNA expression. Moreover, and in line with these functional findings, miR-189 up-regulation seems to exert protective effects against radiation with an attenuation of radiation-induced growth inhibition.
In functional assays with miR-125a, -127 and -148b we observed weaker effects of miR overexpression or inhibition. According to our own results miR-125a is also differentially expressed in human fibroblasts by hydrogen peroxide (H
2O
2), which is like ionizing radiation a stressor for cells [
17]. When changing levels of miR-125a we found a decrease of clonogenic survival upon inhibition. Similar to miR-189, miR-125a had a positive effect on endothelial clonogenic survival after irradiation. Accordingly, we found that inhibition of miR-125a had the respective negative effects, comparable to non-RT conditions.
miR-127 also was found to be down-regulated in irradiated cells. Originally it had been described as a putative tumor suppressor. It is silenced in tumor cells, which causes the overexpression of the proto-oncogene bcl-6 [
21]. Likewise, we also found that the inhibition of miRNA-127 reduced clonogenic survival (and proliferation; data not shown), suggesting that the anti-endothelial radiation effects are associated with down-regulated miR-127 levels. Conversely, over-expressed miR-127 enhanced radiation sensitivity in clonogenic assays. Perhaps the executed signaling pathways are dependent on the expression levels of other parameters suggesting a functional 'switch' role of miRNA-127. Another explanation would be that the downregulation of miRNA-127 after radiation is not functionally in line but rather part of a negative feedback mechanism. Moreover, a dual role of miRNAs has recently been described, showing that a miRNA can repress or enhance mRNA translation, depending on the state of the cell cycle [
22].
Further, it has been described that ionizing radiation also may have dual roles with respect to endothelial cells, angiogenesis and the microenvironment: while radiation has dominantly direct anti-endothelial effects, it may also convey indirect pro-angiogenic effects with up-regulation of VEGF, PDGF or AKT signaling in endothelium. One might speculate that miR-127 is involved in such or similar pro-survival mechanisms [
13,
23].
In the case of miR-148b irradiation down-regulated expression levels. miR-148b inhibition itself slightly reduced clonogenic growth. In contrast, and similarly to the findings for miRNA-127, miR-148b inhibition might favor survival under radiation conditions.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MW-E designed experiments, performed experiments, analyzed data and wrote the manuscript. CS generated software for data evaluation and bio-statistical analyses. AA designed experiments and wrote the manuscript. UW designed experiments and analyzed data. PH designed experiments, analyzed data and wrote the manuscript. All authors read and approved the final manuscript.