Knockdown of EWSR1::CREM and CREM expression in CHL-1 cell line induced senescence and reduced proliferation and migration
Loss of EWSR1::CREM fusion protein in CHL-1 has previously been found to be associated with reduced proliferation, senescence and impaired invasion3. To validate these findings, we investigated the cellular properties associated with endogenous expression of EWSR1::CREM in this cell line. For this, we studied the effect of EWSR1::CREM knockdown on cellular properties.
As we have earlier shown, the EWSR1::CREM fusion protein in CHL-1 can be detected as a band of ≈ 55 kDa size13. Western blotting showed that the EWSR1::CREM fusion protein was highly expressed, with a stronger band than those representing CREM isoforms. Using a pool of siRNAs targeting the DNA-binding domain of CREM, a striking reduction of EWSR1::CREM and CREM expression was seen in western blotting and cell staining (Fig. 2A). The expression of CREB was not altered. In line with EWSR1::CREM and CREM protein loss, reduction of respective mRNA expression was seen using RT-PCR using fusion specific and CREM primers (data not shown). After knockdown, cell shape shifted from spindle-like to a more enlarged and irregular shape, characteristic of cellular senescence14 (Fig. 2B).
Effects of CREM knockdown in the CHL-1 cell line. (A) Western blotting and cell staining after CREM siRNA treatment. The blots are cropped and original blots can be found in Supplementary information—Original blots Supplementary Figs. S2 and S3. (B) Cell morphology after CREM knockdown. (C) Senescence assessed with SA β‐galactosidase staining after CREM knockdown, confirming loss of proliferation from 84 to 4%. (D) Immunostaining for the proliferation marker Ki67 after CREM siRNA treatment. (E) Wound healing assay after CREM knockdown. ***P ≤ 0.001 (t-test).
Next, we measured SA β‐galactosidase positivity, which is a biomarker for senescence, and performed Ki67 staining to measure proliferation. In the CREM knockdown samples, SA β‐galactosidase positivity was substantially increased, indicating senescence (Fig. 2C) and the number of Ki67 positive cells was synchronously reduced (Fig. 2D). Migration was tested by wound healing assay, a significant delay in wound closure was seen after CREM knockdown (Fig. 2E).
mRNA sequencing analysis identified cell cycle/mitosis pathways regulated by ESWR1-CREM and CREM in CHL-1 cells
Ideally, the effects of endogenous EWSR1::CREM protein would be studied in a cancer cell line that does not express CREM. However, CREM is ubiquitously expressed, and no such cell line has to our knowledge been discovered. Therefore, we went on to identify genes and pathways primarily regulated by the highly expressed EWSR1::CREM fusion protein in CHL-1 cells considering that a minority of the transcriptional alteration is likely attributable to the wildtype CREM protein. We performed mRNA sequencing to compare the transcriptomes of ESWR1::CREM and CREM silenced CHL-1 cells with control siRNA treated CHL-1 cells. As a result, we found that as many as nearly 14 000 genes were differentially expressed after EWSR1::CREM and CREM knockdown (data available at Gene Expression Omnibus repository, accession number GSE210911).
To filter the gene list, we adopted criteria based on significant difference of P ≤ 0.05 and Log2 fold-change ≥ 2. With this scheme, we identified 712 genes for functional analysis. This selection of data is visualized by a volcano plot, Fig. 3A.
mRNA sequencing of CHL-1 cells treated with control and CREM siRNA. (A) Volcano plot for visualization of data selection: red for selected upregulated genes, blue for selected downregulated genes and grey for omitted genes based on significance and fold change. (B) Top 20 gene clusters enriched by CREM knockdown. Heatmaps of differentially expressed genes ≥ twofold change upon CREM Knockdown via CREM siRNA, with most being downregulated (upper rows) and small portion of upregulated genes (bottom rows) for (C) Cell cycle check points, (D) Positive Regulation of Cell cycle and (E) DNA replication. (F) Cell cycle markers presented in the heatmaps were investigated protein level in western blotting, GAPDH was used as protein loading. The blots are cropped and original blots/replicates can be found in Supplementary information—Original blots Supplementary Figs. S2, S4, S5, S6, S7. (G) Cell cycle analysis results presented by cell phases. **P ≤ 0.01 and ***P ≤ 0.001 (t-test).
To analyze pathways altered, we performed a Metascape functional analysis of these 712 genes. This analysis revealed that ESWR1-CREM and CREM knockdown altered pathways involved in cell cycle, DNA replication and cell cycle phase transition and other pathways (Fig. 3B).
Heatmaps of genes in representative pathways of cell cycle checkpoints, positive regulation of cell cycle, and DNA replication are shown in Fig. 3C,D and E. The vast majority of the genes were downregulated in these three pathways after EWSR1::CREM and CREM knockdown.
Next, we wanted to validate identified downstream genes at protein level using immunoblotting. We chose the following genes: two G2 checkpoint markers; WEE1 G2 checkpoint kinase (WEE1) and Cyclin B1 (CCNB1), G1 checkpoint marker Cyclin D1 (CCND1), G1 and G2 checkpoint marker: Cyclin A2, (CCNA2), Checkpoint kinase 1 (CHK1) that is responsible for checkpoint mediated cell cycle arrest in response to DNA damage or the presence of unreplicated DNA, representing one of the kinesins—Kinesin family member 11 (KIF11) and proliferation marker Proliferating cell nuclear antigen (PCNA). The results confirmed the reduction of all of the chosen gene products further validating our results. (Fig. 3F).
As cell cycle was the main process affected by EWSR1::CREM and CREM knockdown in CHL-1, we conducted cell cycle analysis to investigate the phase that would be altered. In this analysis, we found that the proportions of cells in S and G2/M phases were reduced after the knockdown of EWSR1::CREM and CREM, while the proportion of cells in the G0/G1 phase increased from 77 to 91% as cells were not able to progress from G1 or became senescent (Fig. 3G). FACS analysis showed that CREM knockdown inhibits S phase progression in CHL-1 cells.
Wild type CREM knockdown reduced proliferation in HEK293, PC-3, and WM164 cells
The cell cycle was the main cellular process affected by EWSR1::CREM and CREM knockdown in CHL-1 cells. We wondered if CREM knockdown in cell lines lacking the constitutionally active fusion protein would result in a similar effect. To interrogate the effect of CREM knockdown on non-malignant and malignant cells, we used three cell lines: Human embryonic kidney cell line HEK 293, human prostate carcinoma cell line PC-3, and human melanoma WM164 cell line.
Cells were transfected with CREM siRNA for 48 h. Reduced expression of CREM protein was confirmed by immunoblotting (Fig. 4A). Bands corresponding to CREM predicted size were clearly weaker after knockdown, with most efficient loss of CREM seen in PC-3 and WM164 cells. Morphologically, no significant alterations could be observed (Fig. 4B). Neither was a significant change observed in the frequency of senescent cells (data not shown). Proliferation was assessed using Ki67 staining and quantification of positively stained nuclei. The knockdown of CREM resulted in a modest reduction in proliferation in HEK-293, PC-3, and WM164 (Fig. 4C). In the wound healing assay, a slight delay in migration was observed solely in HEK-293 cells after CREM knockdown (Fig. 4D). No effect was observed in PC-3 or WM164 cells.
Wild type CREM knockdown. (A) Immunoblot showing Wild type CREM knocked down in HEK-293, PC3, and WM164 after treatment with CREM siRNA for 48 h. The blots are cropped and original blots can be found in Supplementary information—Original blots Supplementary Fig. S8. (B) Cell morphology in HEK-293, PC3, and WM164 cell lines after the reduction in CREM levels. (C) Proliferation assessed with Ki67 after CREM siRNA in HEK-293, PC3, and WM164 cell lines, it resulted with proliferation of loss from 82 to 78%, from 75 to 53%, and from 69 to 58%, respectively. (D) Wound healing assay to assess migration after CREM knockdown in HEK-293, PC3, and WM164 cell lines. (E) Downstream cell cycle related proteins and their expression measured by immunoblotting after CREM knockdown in HEK-293, PC3, and WM164 cell lines. The blots are cropped and CCNA2 was run in lanes with one empty in between so for this image, the empty lane is removed. Original blots/replicates can be found in Supplementary information—Original blots Supplementary Figs. S2, S4, S5, S6 and S7. (F) FACS analysis results after CREM siRNA treatment for HEK-293, PC3, and WM164 cell lines. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001 (t-test).
In order to investigate if the downstream cell cycle related proteins were altered as seen in the CHL-1 cells, immunoblotting for Wee1, CCNB1, CCND1, CCNA2, CHK1, KIF11 and PCNA was conducted. HEK-293 showed reduction of Wee1 expression. PC-3 showed reduction of Wee1 and CCNA2 expression. WM164 showed reduction of Wee1, CCNB1, CCNA2 and PCNA expression (Fig. 4E). The result reflects the lesser impact of CREM knockdown in these non-fusion bearing cell lines.
To analyze the effects of CREM knockdown on the cell cycle, we conducted FACS analysis. Here, results showed a reduction of cells in S phase after CREM siRNA treatment in HEK-293, PC-3, and WM164 cells (Fig. 4F). Only PC-3 cells demonstrated significant changes in all cell phase groups although all cell types had a similar effect as seen in CHL-1, but of smaller magnitude.
CREM target genes with altered expression after EWSR1::CREM and CREM loss in CHL-1 cells
Taken together, the results above indicate that the best part of alterations seen upon CREM knockdown in the CHL-1 cell line are attributable to the highly expressed fusion protein EWSR1::CREM, which by structure is predicted to enhance transcription from CREM target genes. The effects of CREM knockdown on cell lines lacking this fusion were, as expected, similar but of lesser amplitude.
The loss of EWSR1::CREM and CREM expression lead to a dramatic loss of proliferation and induction of senescence in CHL-1, but it remained unclear whether this was due to differential expression from direct CREM target genes, or alterations further downstream. In pursuance of direct CREM target genes responsible for the altered phenotype, we reassessed the differentially expressed gene list from our transcriptomic analysis, looking for direct CREM target genes. To identify CREM regulated genes, we used the TRANSFAC Predicted Transcription Factor Targets dataset that can be utilized to identify potential transcription factor binding sites. We found 660 predicted CREM target genes. Among these, 497 genes had been differentially expressed in our transcriptomic analysis of CHL-1 cells with decrease of EWSR1::CREM and CREM expression. From these 497 genes, 248 were significantly changed (P ≤ 0.05) (128 downregulated and 120 upregulated), when all fold-changes were considered. Adopting a fold-change cut off of Log2 ≥ 2, we found 23 genes; 18 downregulated and 5 upregulated.
Since the EWSR1::CREM fusion protein is a constitutively active transcription factor, the genes downregulated by EWSR1::CREM and CREM knockdown were of greatest interest. In addition, we listed upregulated genes, since specific isoforms of wild type CREM can repress expression of CREM target genes.
The majority of downregulated CREM target genes are involved in cell cycle related Gene Ontology biological processes according to the freely accessible UniProt website www.uniprot.org/15. Eleven out of the 23 predicted target genes are druggable according to The Drug Gene Interaction Database. The predicted CREM target genes differentially expressed upon EWSR1::CREM and CREM knockdown, their biological process and potential druggability are listed in Table 1.
Among the CREM target genes downregulated upon EWSR1::CREM and CREM knockdown, the highest relative mRNA expression before knockdown was noted for Ornithine decarboxylase 1 (ODC1). ODC1 is a well characterized enzyme responsible of decarboxylation of ornithine into putrescine16. This is the first phase of polyamine biosynthesis, a rate-limiting step that regulates cell proliferation17,18. Using western blotting, we established that the reduction of ODC1 mRNA is translated to loss of ODC1 protein (Fig. 5A). Another CREM target gene with reduced expression upon EWSR1::CREM and CREM knockdown also regulates polyamines: Spermidine/spermine N1-acetyltransferase 1 (SAT1, also known as SSAT). SAT1 is a catabolic enzyme that catalyzes the acetylation of two of the three polyamines: spermidine and spermine, leading to the degradation or transport of these polyamines out of cells19.The relative mRNA expression of SAT1 before knockdown was considerably lower than that of ODC1. Nevertheless, we were able to validate that the reduction of SAT1 mRNA also translates to reduced protein expression using western blotting (see Supplementary Fig. S1 online).
Effects of CREM and ODC1 silencing on cell lines CHL-1 and PC-3. (A) Immunoblot showing effects of CREM knockdown and ODC1 knockdown on ODC1 protein levels in in CHL-1, HEK-293, PC3, and WM164 cell lines. The blots are cropped and original blots can be found in Supplementary information—Original blots Supplementary Figs. S9 and S10. (B) Proliferation assessed with Ki67 after ODC1 siRNA in CHL-1 and PC3 cell lines. Loss of proliferation was 84–22% in CHL-1 and from 69 to 17% in PC3 cells. (C) Senescence assessed with SA β‐galactosidase staining after ODC1 knockdown in CHL-1 cell line, an increase from 15 to 71% was seen. (D) Wound healing assay to assess migration after ODC1 knockdown in CHL-1 and PC3 cell lines. (E) FACS analysis results after ODC1 siRNA treatment for CHL-1 and PC3 cell lines. An increase in G0/G1 in CHL-1cells from 82 to 91% and PC3 in cells from 83 to 95% (F) Downstream cell cycle related proteins and their expression measured by immunoblotting after ODC1 knockdown in CHL-1 and PC3 cell lines. The blots are cropped and original blots/replicates can be found in Supplementary information—Original blots Supplementary Figs. S11, S12 and S13. (G) Total polyamine concentration compared in CHL-1 and PC3 cell lines after CREM and ODC1 knockdown. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001 (t-test).
The extent of ODC1 loss, together with its known role in proliferation, suggested it may be the pivotal mediator of the cellular effects brought about by upon EWSR1::CREM and CREM knockdown.
ODC1 expression is regulated by the MYC proto-oncogene, bHLH transcription factor (MYC)20.To rule out a secondary ODC1 loss, through MYC, we checked the effect of upon EWSR1::CREM and CREM knockdown on MYC expression. In the transcriptomic analysis, MYC expression was unaltered. We further performed western blotting, where a slight reduction of c-myc protein expression was seen (see Supplementary Fig. S1 online).
ODC1 is essential for cell cycle progression in the fusion bearing CHL-1 cell line
We decided to focus on the exploration and validation of the newly discovered CREM-ODC1 axis. Our hypothesis was that ODC1-mediated polyamine synthesis has a pivotal role in mediating the EWSR1::CREM and CREM effects on cell cycle and migration.
First, we used western blot to study the possible reduction of ODC1 protein expression after CREM knockdown in the fusion bearing and control cell lines. The reduction is striking in the CHL-1 cell line. With the PC-3 cell line a reduction is clear but HEK-293 or WM164 cell lines have no detectable reduction of ODC1 expression (Fig. 5A). Based on these findings we decided to continue the experiments with CHL-1, using PC-3 as the control cell line.
Next, we tested whether ODC1 silencing could induce similar functional effects as detected upon CREM silencing. Using a pool of siRNAs targeting ODC1, a dramatic reduction of ODC1 expression was seen in western blotting and cell staining in the CHL-1 cell line. In the PC-3 cell line, ODC1 expression was initially low, and further reduced by knockdown (Fig. 5A). We measured SA β‐galactosidase positivity to assess senescence and performed Ki67 staining and quantification of positively stained nuclei to measure proliferation. The number of Ki67 positive cells was reduced in both CHL-1 and PC-3 cells, confirming decrease of proliferation (Fig. 5B). The ODC1 knockdown of CHL-1 cells substantially increased SA β‐galactosidase positivity indicating increased senescence (Fig. 5C). PC-3 cells showed no significant increase in SA β‐galactosidase positivity (data not shown). Migration was tested by wound healing assay. Interestingly, no significant change was noted in either cell line (Fig. 5D).
To analyze the effects of ODC1 knockdown on the cell cycle, we conducted FACS analysis. Here, results showed in both cell lines significant reduction of cells in S and G2/M phases after ODC1 siRNA treatment. Significant increase in G0/G1 phases was observed in CHL-1 cell line and in PC-3 cell line (Fig. 5E). The cell cycle alterations were of similarmagnitude as with CREM knockdown.
We were then interested to see how the CREM downstream genes involved in cell cycle were affected by the ODC1 knockdown. In both cell lines we saw reduction of Wee1, CCNB1, CCNA2, CHK1 and KIF11 expression (Fig. 5F).
Finally, we measured the total polyamine concentration in both cell lines and compared the effects of CREM siRNA and ODC1 siRNA knockdown on the total polyamine concentration. In CHL-1, both CREM siRNA and ODC1 siRNA knockdown produced a significant reduction in the total polyamine concentration. In the PC-3 cell line, only ODC1 siRNA knockdown produced modest, but statistically significant reduction (Fig. 5G). This suggests that PC3 has compensatory mechanisms maintaining intracellular polyamine levels, and is not as dependent on ODC1 as CHL-1 is.
All in all, we were able to demonstrate that CREM siRNA and ODC1 siRNA knockdown lead to very similar phenotypes. This suggests that the phenotype is largely brought upon by ODC1. The ODC1 mediated effect does not seem to extend to migration during the studied timeframe in these cell lines.
Read more here: Source link