Hypoxic induction of the regulator of G-protein signalling 4 gene is mediated by the hypoxia-inducible factor pathway

Abstract

The transcriptional response to hypoxia is largely dependent on the Hypoxia Inducible Factors (HIF-1 and HIF-2) in mammalian cells. Many target genes have been characterised for these heterodimeric transcription factors, yet there is evidence that the full range of HIF-regulated genes has not yet been described. We constructed a TetON overexpression system in the rat pheochromocytoma PC-12 cell line to search for novel HIF and hypoxia responsive genes. The Rgs4 gene encodes the Regulator of G-Protein Signalling 4 (RGS4) protein, an inhibitor of signalling from G-protein coupled receptors, and dysregulation of Rgs4 is linked to disease states such as schizophrenia and cardiomyopathy. Rgs4 was found to be responsive to HIF-2α overexpression, hypoxic treatment, and hypoxia mimetic drugs in PC-12 cells. Similar responses were observed in human neuroblastoma cell lines SK-N-SH and SK-N-BE(2)C, but not in endothelial cells, where Rgs4 transcript is readily detected but does not respond to hypoxia. Furthermore, this regulation was found to be dependent on transcription, and occurs in a manner consistent with direct HIF transactivation of Rgs4 transcription. However, no HIF binding site was detectable within 32 kb of the human Rgs4 gene locus, leading to the possibility of regulation by long-distance genomic interactions. Further research into Rgs4 regulation by hypoxia and HIF may result in better understanding of disease states such as schizophrenia, and also shed light on the other roles of HIF yet to be discovered.

Figure 1. Generation and verification of PC-12 TetON HIF-α inducible expression cell lines.

(A) PC-12/TetON/EYFP, PC-12/TetON/HIF-2α/EYFP and PC-12/TetON/HIF-1α/EYFP cells were transfected with a firefly luciferase reporter gene containing 4 copies of the HRE (pHRE4) or a control construct which lacks the HRE (pGL3). 6 hours post transfection, cells were exposed to 20% O₂ or <1% O₂ in the presence or absence of 2 µg/mL doxycycline (Dox). After 16 hours, cells were assayed for firefly luciferase activity against a cotransfected internal renilla luciferase control. Results representative of 3 independent experiments performed in triplicate. Error bars represent standard deviation. (B) Cell extracts were prepared from cells treated as in (A), and 20 µg of total protein from each was analysed by immunoblotting with antibodies to HIF-1α (upper immunoblot) or HIF-2α (lower immunoblot).

Figure 2. Rgs4 is responsive to hypoxia and hypoxia-mimetic chemicals.

(A) Northern blot of total RNA from parental PC-12 cells with or without 16 hours of hypoxic treatment, and the TetON-Control, TetON-HIF1α and TetON-HIF-2α PC-12 cell lines with or without 16 hours of doxycycline treatment (Dox). Blot shown is representative of either three (Rgs4 with doxycycline treatments), two (Rgs4 with hypoxia, Vegf with doxycycline) or one (Vegf with hypoxia) independent experiments. (B) SK-N-SH cells were treated with normoxia, hypoxia, 0.1% DMSO (vehicle), 1 mM DMOG or 100 µM DP for 16 hours before target gene analysis by qRT-PCR. Mean fold change with standard deviation for three independent experiments is shown. (C) PC-12 and (D) SK-N-BE(2)C cell lines were tested in two independent experiments as in 1(B). Representative experiments (from n = 3) are shown. (E) SK-N-BE(2)C and HUVEC cell cultures were treated with 0 (normoxia), 3 and 16 hours of hypoxia before quantification of Rgs4, Rgs5 and Vegf mRNA levels relative to Polr2a using qRT-PCR, normalised to normoxic levels. Mean fold change with standard deviation for three independent experiments is shown. Asterisks indicate that the normoxic control lies outside of the 99% confidence interval of relative mean target gene expression.

Figure 3. Knockdown of HIF-α subunits impairs response of Rgs4.

SK-N-BE(2)C cells were transfected twice with 50 nM of negative control siRNA, or siRNA targeted towards Hif1a or Epas1 mRNA (siHIF1_1541 or siHIF2_1599, respectively), or 50 nM each of both targeted siRNAs together. Cells were then grown for 8 hours with 0.1% DMSO vehicle (V) or 100 µM DP before collection and gene expression analysis for Rgs4, Vegf, Dec1, Hif1a and Epas1 expression levels relative to Polr2a by qRT-PCR. Mean fold change with standard deviation between three independent experiments is shown, each normalised to the DP-treated level of expression of each gene after negative control siRNA transfection, with asterisks indicating that the DP-treated control lies outside of the 99% confidence interval of relative target gene expression.

Figure 4. Rgs4 response is not due to mRNA stability or indirect transcriptional activation.

(A) SK-N-BE(2)C cells were treated with 4 µg/ml actinomycin D prior to treatment with 4, 8 or 12 hours of normoxia or hypoxia. Untreated cells were also used to give control normoxic (0 hours) and hypoxic (8^# hours) Rgs4 levels. Rgs4 mRNA levels were quantified using qRT-PCR relative to Polr2a, then normalised to normoxic levels. One representative of n = 2 is shown. Rgs4 (B) and Polr2a (C) mRNA levels relative to total RNA were also calculated from C(t) values without normalisation to reference gene relative to normoxia. The mean detection with standard error (n = 2) is plotted to give a one-phase decay using GraphPad Prism 5. Estimated mRNA half-lives are described in the text. (D) SK-N-BE(2)C cells were transfected twice with 100 ng of pCI_FL (control) or pCI_FL incorporating the 3′UTR of human Rgs4 (hRgs4_3′UTR). Mean relative luciferase units with standard deviation from 3 replicate wells of one experiment are shown as a representative of n = 3. (E) SK-N-BE(2)C cells were treated with 100 µM DP for 1, 2, 4 8 or 16 hours in comparison to an untreated (0 hours) control. Message levels were quantified by qRT-PCR for indicated transcripts relative to reference gene Polr2a, then normalised to normoxic levels. Data is presented as the mean and standard deviation from three independent experiments. Asterisks indicate that control level lies outside of the 99% confidence interval of the DP-treated target gene expression.

Figure 5. Functional hypoxia responsive elements cannot be detected within 33

kb of the human Rgs4 gene. (A) Diagram of the human Rgs4 locus, with HRE-like sequences depicted by crosses, and the regions (A-N) that were cloned by genomic PCR and inserted into pGL3 upstream of the luc ⁺ coding sequence. Primer sequences for PCR are listed in Table S1. hRgs4 splice variants and regions of high conservation within mammals as shown as depicted by UCSC Genome Browser. (B) SK-N-BE(2)C cells were transfected twice with 100 ng of pGL3promoter (-) or pGL3-based reporter plasmids incorporating the genomic sequences depicted in (A), while 25 ng pHRE4 was transfected separately as a positive control for hypoxia treatment. Cells were treated with hypoxia for 16 hours, before relative luciferase expression analysis. Mean relative luciferase units of triplicate wells with standard deviation are depicted, as a representative of three independent experiments.