MicroRNA response to hypoxic stress in soft tissue sarcoma cells: microRNA mediated regulation of HIF3α

Abstract

Background: Hypoxia is often encountered in solid tumors and known to contribute to aggressive tumor behavior, radiation- and chemotherapy resistance resulting in a poor prognosis for the cancer patient. MicroRNAs (miRNAs) play a role in the regulation of the tumor cell response to hypoxia, however, not much is known about the involvement of miRNAs in hypoxic signalling pathways in soft tissue sarcomas (STS).

Method: A panel of twelve STS cell lines was exposed to atmospheric oxygen concentrations (normoxia) or 1% oxygen (hypoxia) for up to 48 h. Hypoxic conditions were verified and miRNA expression profiles were assessed by LNA™ oligonucleotide microarrays and RT-PCR after 24 h. The expression of target genes regulated by hypoxia responsive miRNAs is examined by end-point PCR and validated by luciferase reporter constructs.

Results: Exposure of STS cell lines to hypoxic conditions gave rise to upregulation of Hypoxia Inducible Factor (HIF) 1α protein levels and increased mRNA expression of HIF1 target genes CA9 and VEGFA. Deregulation of miRNA expression after 24 h of hypoxia was observed. The most differentially expressed miRNAs (p<0.001) in response to hypoxia were miR-185-3p, miR-485-5p, miR-216a-5p (upregulated) and miR-625-5p (downregulated). The well-known hypoxia responsive miR-210-3p could not be reliably detected by the microarray platform most likely for technical reasons, however, its upregulation upon hypoxic stress was apparent by qPCR. Target prediction algorithms identified 11 potential binding sites for miR-485-5p and a single putative miR-210-3p binding site in the 3'UTR of HIF3α, the least studied member of the HIF family. We showed that HIF3α transcripts, expressing a 3'UTR containing the miR-485-5p and miR-210-3p target sites, are expressed in all sarcoma cell lines and upregulated upon hypoxia. Additionally, luciferase reporter constructs containing the 3'UTR of HIF3α were used to demonstrate regulation of HIF3α by miR-210-3p and miR-485-5p.

Conclusion: Here we provide evidence for the miRNA mediated regulation of HIF3α by hypoxia responsive miRNAs in STS, which may help to tightly regulate and fine-tune the hypoxic response. This provides a better insight into the mechanisms underlying the hypoxic response in STS and may ultimately yield information on novel prognostic and predictive markers or targets for treatment.

Figure 1

Hypoxia induces HIF1α protein levels and CA9 and VEGFA transcription in soft tissue sarcoma cell lines. (A) HIF1α protein levels are stabilized and increase during hypoxia. Cell lines were cultured under hypoxia (H) or normoxia (N), after the indicated times cell lysates were prepared and analysed by Western blotting for HIF1α expression. HIF1α levels peak at 6 hours of hypoxia. Upon prolonged hypoxia (24 and 48 hours) HIF1α protein levels decrease, but remain increased compared to the HIF1α levels detected in cells cultured under normoxic conditions. β-Actin was used as a loading control. (B,C) mRNA levels of HIF1 target genes CA9 (B) and VEGFA (C) are upregulated during hypoxia (grey bars) compared to normoxia (black bars). CA9 and VEGFA expression was determined by RT-PCR and normalized to the expression of HPRT. Bars indicate average fold change of mRNA expression ± S.D.(n = 2) compared to the levels detected at 0 h which are arbitrarily set at 1. Statistical significance between CA9 and VEGF expression in normoxic and hypoxic samples at the various time points was determined by two-sample t-tests: * = p < 0.05, ** = p < 0.005, *** = p < 0.0005. The complete panel of 12 soft tissue sarcoma cell lines was examined, the results obtained with five representative cell lines are shown.

Figure 2

Hypoxia induces changes in miRNA expression. (A) Unsupervised hierarchical clustering using miRNA expression data (407 miRNAs) of hypoxic (H) and normoxic (N) cell line samples does not discriminate the hypoxic and normoxic samples. The hypoxic samples cluster together with their normoxic counterparts. (B) The most significant differentially expressed miRNAs (two-sample t-test, p < 0.05) between cell lines that were cultured under normoxic and hypoxic conditions are used for a supervised hierarchical clustering. The expression of these 32 miRNAs can distinguish hypoxic from normoxic sarcoma cell line samples. (C) The four most significant (P < 0.001) differentially expressed miRNAs (miR-485-5p, miR-216a-5p, miR-185-3p and miR-625-5p) discriminate sarcoma cell lines that were cultured under normoxic conditions from those cultured under hypoxic conditions in a supervised hierarchical clustering. (D) MiR-210-3p is upregulated during hypoxia. RT-PCR was used to determine miR-210-3p levels in normoxic (black bars) and hypoxic (grey bars) samples from five representative sarcoma cell lines. Expression is normalized against RNU43 expression. Bars indicate average expression fold change ± S.D. (n = 2) of miR-210-3p compared to the expression at 0 h. Statistical significance between expression in normoxic and hypoxic samples at a time point were determined by two-sample t-tests: * = p < 0.05, ** = p < 0.005, *** = p < 0.0005. The different cell lines are indicated by a color code that identifies their tumor of origin: with synovial sarcoma (yellow); Ewing sarcoma (green); rhabdomyosarcoma (purple); leiomyosarcoma (blue); liposarcoma (orange) and fibrosarcoma (red).

Figure 3

The 3’UTR of HIF3α contains putative miR-210-3p and miR-485-5p binding sites and is expressed in sarcoma cell lines. (A) Schematic representation of the 3’UTR of the HIF3α transcript variants 003 and 201 (http://www.ensembl.org, release 72). Depicted are predicted target sites of miR-210-3p (open box) and miR-485-5p (black boxes). (B) Arrows indicate location of primers used to detect the presence of HIF3α 3’UTR transcripts. A 212 bp and 313 bp amplification product indicate expression of the 3’UTR of HIF3α transcription variants 003 and 201. (C) Lines designate the HIF3α 3’UTR fragments (HIF3α-short, HIF3α-long) which were cloned into the psiCHECK-2 luciferase reporter to verify regulation of HIF3α by miRNAs. (D) End-point RT-PCR was used to determine the presence of HIF3α 3’UTR transcripts and their induction upon hypoxia. Depicted are EtBr stained PCR amplification fragments of 212 and 313 bp derived from HIF3α 3’UTR cDNA and an amplified 336 bp HPRT fragment as input control. The results obtained with five representative cell lines are shown.

Figure 4

HIF3α is regulated by miR-210-3p and miR-485-5p. (A) MiR-210-3p and miR-485-5p mimic overexpression reduce HIF3α protein induction under hypoxic conditions. The sarcoma cell lines SW872 and SK-UT-1 were transfected with mimics of miR-210-3p (m210), miR-485-5p (m485) and a scrambled negative control mimic (mneg). 48 h post-transfection cells were cultured under hypoxic conditions (1% O₂) for 24 h. Subsequently total protein lysates were prepared and analysed by Western blotting for HIF3α protein expression. β-Actin is used as a loading control. (B) Hypoxia responsive miRNAs target HIF3α 3’UTR. SW872 and SK-UT-1 cell lines were transfected with psiCHECK 2 constructs containing short and long 3’UTR fragments of HIF3α. 24 h later the cells were exposed to hypoxia for 24 h. Bars indicate average luciferase activity ± SD (n = 3) measured in hypoxic cell lysates relative to the luciferase activity in normoxic cell lysates which is arbitrarily set at 100. (C) MiR-210-3p and miR-485-5p regulate HIF3α. SW872 and SK-UT-1 cell lines were transfected with mimics of miR-210-3p (m210), miR-485-5p (m485) or a scrambled control mimic (mneg) followed after 24 h by a transfection with psiCHECK 2 constructs containing short and long fragments of the HIF3α 3’UTR. Bars indicate average luciferase activity ± SD (n = 3) measured in cell lysates after 24 h. (D) MiR-210-3p and miR-485-5p regulate HIF3α. SW872 and SK-UT-1 cell lines were transfected with mimics of miR-210-3p (m210), miR-485-5p (m485) or a scrambled control mimic (mneg) followed, after 24 h, by a transfection with a psiCHECK 2-HIF3α-short construct containing either wild-type (WT) or mutated (mut) miR-210-3p and miR-485-5p binding sites. Bars indicate average luciferase activity ± SD (n = 3) measured in cell lysates. Statistical significance was determined by two-sample t-tests: * = p < 0.05, ** = p < 0.005, *** = p < 0.0005.