Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 6 (8), 1375-84

The Multifunctional Protein glyceraldehyde-3-phosphate Dehydrogenase Is Both Regulated and Controls Colony-Stimulating factor-1 Messenger RNA Stability in Ovarian Cancer

Affiliations

The Multifunctional Protein glyceraldehyde-3-phosphate Dehydrogenase Is Both Regulated and Controls Colony-Stimulating factor-1 Messenger RNA Stability in Ovarian Cancer

Yi Zhou et al. Mol Cancer Res.

Abstract

Although glyceraldehyde-3-phosphate dehydrogenase's (GAPDH) predilection for AU-rich elements has long been known, the expected connection between GAPDH and control of mRNA stability has never been made. Recently, we described GAPDH binding the AU-rich terminal 144 nt of the colony-stimulating factor-1 (CSF-1) 3' untranslated region (UTR), which we showed to be an mRNA decay element in ovarian cancer cells. CSF-1 is strongly correlated with the poor prognosis of patients with ovarian cancer. We investigated the functional significance of GAPDH's association with CSF-1 mRNA and found that GAPDH small interfering RNA reduces both CSF-1 mRNA and protein levels by destabilizing CSF-1 mRNA. CSF-1 mRNA half-lives were decreased by 50% in the presence of GAPDH small interfering RNA. RNA footprinting analysis of the 144 nt CSF-1 sequence revealed that GAPDH associates with a large AU-rich-containing region. The effects of binding of GAPDH protein or ovarian extracts to mutations of the AU-rich regions within the footprint were consistent with this finding. In a tissue array containing 256 ovarian and fallopian tube cancer specimens, we found that GAPDH was regulated in these cancers, with almost 50% of specimens having no GAPDH staining. Furthermore, we found that low GAPDH staining was associated with a low CSF-1 score (P = 0.008). In summary, GAPDH, a multifunctional protein, now adds regulation of mRNA stability to its repertoire. We are the first to evaluate the clinical role of GAPDH protein in cancer. In ovarian cancers, we show that GAPDH expression is regulated, and we now recognize that one of the many functions of GAPDH is to promote mRNA stability of CSF-1, an important cytokine in tumor progression.

Figures

Figure 1
Figure 1. Silencing GAPDH and its effect on CSF-1 mRNA and protein expression
(A) Immunoblot analysis of total protein extracts from control or GAPDH siRNA treated NOSE.1 or Hey cells. The extracts were harvested 48 hours post-siRNA transfection. The membrane was probed with antibodies to GAPDH and β-actin. GAPDH siRNA treatment significantly reduced or abolished GAPDH protein. (B) RT-PCR analysis of the CSF-1 and β-actin RNA expression in GAPDH or control siRNA treated NOSE.1 or Hey cells. The RNAs were harvested 24 hours post-siRNA transfection. GAPDH siRNA significantly down-regulated CSF-1 RNA compared to controls. (C) Levels of CSF-1 secreted in the conditioned media of GAPDH or control siRNA treated NOSE.1 or Hey cells (ng/ml). The media was harvested 57 hours post-siRNA transfection. There was no significant effect of control siRNA on secreted CSF-1 levels compared to untreated Hey cells (p=0.15). GAPDH siRNA was able to significantly decrease secreted CSF-1 protein levels when compared to controls.
Figure 2
Figure 2. Effect of GAPDH siRNA treatment on CSF-1 mRNA half-life in Hey ovarian cancer cells
Hey cells were treated with 5 μg/ml actinomycin 48 hours after control or GAPDH siRNA transfection, and total RNA was isolated over the time period of the actinomycin-D chase for Northern Blot analysis of CSF-1 RNA and 18S RNA expression. After densitometric analysis, relative transcript levels were plotted versus time. A representative experiment is depicted. The mean difference in CSF-1 mRNA half-life is 2 fold between GAPDH and control siRNA treated Hey cells in 3 independent experiments.
Figure 3
Figure 3. RNA footprinting analysis of the interaction of GAPDH protein with the 3′ untranslated region of CSF-1 RNA
(A) 3′ end-labeled 3′ UTR CSF-1 RNA (3829-3972nt) sequenced chemically. The pattern derived was used to help interpret the footprinting experiments. (B) 3′ region of GAPDH binding to the 3′UTR CSF-1 RNA. The lane numbers are at the bottom of each panel. Lane 1, labeled CSF-1 RNAs (3828-3972nt) are partially hydrolyzed by RNase T1 (cleavage site 3′ of single-stranded G’s) to provide a reference ladder. CSF-1 RNA (3828-3972 nt) was subjected to a standard binding reaction in the presence (lane 3, 7, 8, 9) (+) or absence (lane 2, 4, 5, 6) (−) of 1.4 μg human GAPDH, followed by cleavage with RNase A (lanes 4–9). Partial cleavage with the decreasing amounts of RNase A (0.1ng/ml (lanes 4, 7), 0.02ng/ml (lanes 5, 8), 0.004ng/ml (lanes 6, 9)) or control reactions without RNase (lanes 2, 3) was performed. Triangle and arrow indicate the approximate boundaries of the 3′ region of the footprint between (3908-3916nt) and 5′ to 3955nt (the 3′ end of the footprint is better defined as 3939-3948nt in the text). (C) 5′ region of GAPDH binding to the 3′UTR CSF-1 RNA. The conditions for each lane are as described in (B). Triangle and arrow indicate the approximate boundaries of the 5′ region of the footprint between (3856-3868nt) and 3905nt. The 5′ region of the footprint can also be visualized in (B), but with less detail. (D) AU-rich footprint for GAPDH binding of 3′UTR CSF-1 RNA.
Figure 4
Figure 4. Predicted secondary structure of 3′UTR CSF-1 riboprobe and gel shift assays with cold competition by wildtype or mutant CSF-1 RNAs
(A) The Mulfold RNA folding program prediction of the secondary structure of the riboprobe containing the terminal 144nt 3′ UTR CSF-1 RNA (3829-3972nt). The dark shading indicates the omitted nucleotides in deletion I (23nt: U3885-U3907, UUUAUAAAGUCAUUUAAAUAUCU) created specifically to delete the ‘v’ structure) and the light shading the omitted nucleotides in deletion II (18nt: A3931-A3948, AUAUAUUUAAUAAUAAAA) at the 3′ end of the footprint). The U3856nt and A3948nt indicated by arrows represent the approximate start and end of the footprint. (B) RNA gel shift assay of GAPDH protein binding to CSF-1 RNA sequences using the wildtype 3′UTR CSF-1 (3829-3972nt) riboprobe. The results of competition with a molar excess (6.7-60 fold) of cold RNA sequences demonstrate successful competition for binding by excess identical cold CSF-1 RNA. However, excess identical cold CSF-1 RNA but containing either deletion I or II only partially competed for binding by the labeled wildtype CSF-1 riboprobe. Therefore both regions in CSF-1 RNA appear to contribute to binding of CSF-1 RNA to GAPDH protein.
Figure 5
Figure 5. RNA gel shift assays comparing binding of wild type and mutated form of CSF-1 144nt 3′UTR to protein extract from Hey or NOSE.1 cells
Mutations were created in 4 AU-rich regions within the terminal 144nt CSF-1 3′UTR as described in the text. Panels A&B represent results of binding of labeled wild type CSF-1 riboprobe to Hey (A) or NOSE.1 (B) protein extracts; Panels C&D represent results of binding of labeled mutant CSF-1 riboprobe to Hey (C) or NOSE.1 (D) protein extracts. Lanes 1, free riboprobe control; lanes 2, 1X probe mixed with 8 ug of protein extract; lanes 3, 3X probe mixed with 8ug of protein extract; lanes 4, 5 and 6, 1X probe mixed with 8 ug of protein extract in presence of 1500X, 770X or 390X excess cold probe, respectively.
Figure 6
Figure 6. Representative pictures of GAPDH from an ovarian and fallopian tube cancer tissue microarray
Panels A,C&E (original magnification, 40X) and B,D&F (magnified from squared area). Panels A&B: A case of ovarian serous carcinoma with both cytoplasmic and nuclear GAPDH score of zero. Hematoxlin was used as a counterstaining agent, therefore the blue color in the nuclei indicates negative staining. Panels C&D show an ovarian serous carcinoma case which had cytoplasmic (score 300) but no nuclear GAPDH staining. Panels E&F show an ovarian serous carcinoma case which had nuclear (score 100) but no cytoplasmic GAPDH staining.

Similar articles

See all similar articles

Cited by 40 PubMed Central articles

See all "Cited by" articles

Publication types

MeSH terms

Substances

Feedback