Effects of activin and TGFβ on p21 in colon cancer

Abstract

Activin and TGFβ share SMAD signaling and colon cancers can inactivate either pathway alone or simultaneously. The differential effects of activin and TGFβ signaling in colon cancer have not been previously dissected. A key downstream target of TGFβ signaling is the cdk2 inhibitor p21 (p21(cip1/waf1)). Here, we evaluate activin-specific effects on p21 regulation and resulting functions. We find that TGFβ is a more potent inducer of growth suppression, while activin is a more potent inducer of apoptosis. Further, growth suppression and apoptosis by both ligands are dependent on SMAD4. However, activin downregulates p21 protein in a SMAD4-independent fashion in conjunction with increased ubiquitination and proteasomal degradation to enhance migration, while TGFβ upregulates p21 in a SMAD4-dependent fashion to affect growth arrest. Activin-induced growth suppression and cell death are dependent on p21, while activin-induced migration is counteracted by p21. Further, primary colon cancers show differential p21 expression consistent with their ACVR2/TGFBR2 receptor status. In summary, we report p21 as a differentially affected activin/TGFβ target and mediator of ligand-specific functions in colon cancer, which may be exploited for future risk stratification and therapeutic intervention.

Figure 1. In the presence of SMAD4, TGFβ is a more potent inducer of growth suppression and activin a more potent inducer of apoptosis.

A) After ACVR2/TGFBR2/SMAD4-wild type FET, FET cells following transient SMAD4 knockdown, and ACVR2/TGFBR2-wild type/SMAD4-null SW480 cells were treated with vehicle (control), activin or TGFβ for 24 hours, the metabolic activity via MTT-growth assay was assessed. Growth suppression occurred only in the presence of SMAD4 following both activin and TGFβ treatment (***p<0.001). Further, TGFβ led to a significant increase in growth suppression compared to activin (***p<0.001). B) To determine the rate of apoptosis, a TUNEL assay was performed in SMAD4-wild type FET, SMAD4-KD FET, and SMAD4-null SW480 cells treated with control vehicle (C), activin (A), or TGFβ(T). Apoptosis was determined by TUNEL-labeling of apoptotic bodies. C) Normalization [% apoptotic bodies/nuclei] revealed that activin- and TGFβ-induced apoptosis occurred predominantly in SMAD4-wild type FET and not in SMAD4-KD FET or SMAD4-null SW480 colon cancer cells. In the SMAD4 expressing cells, activin induced apoptosis to a greater degree than TGFβ (*p<0.05). D) BrdU-labeled intracellular DNA fragments, indicative of apoptosis, were determined 24 hour after activin or TGFβ treatment of SMAD4-wild type FET colon cancer cells, FET cells with p21 KD, FET cells with SMAD4 KD and SMAD4-null SW480 colon cancer cells. Increase in DNA fragmentation was noted after activin and TGFβ treatment only in the SMAD4 wild type cells, with activin inducing more fragmentation compared to TGFβ. In the presence of SMAD4, p21KD lead to a basal increase in apoptosis, but activin treatment lead to no induction of apoptosis. SMAD4 knockdown resulted in loss of apoptosis in FET cells akin to effects observed in the SMAD4-null SW480 cells (*p<0.05, **p<0.01, ***p<0.001).

Figure 2. While TGFβ increases p21 expression in the presence of SMAD4, activin decreases nuclear and total p21 independent of SMAD4 status.

A) SMAD4-wild type FET and SMAD4-null SW480 colon cancer cells were treated with vehicle (control), activin, or TGFβ for 24 hours. p21-specific transactivation was determined using a dual luciferase assay with pWWP-luc and pRL-TK (left panel) and mRNA expression levels of p21 were quantified by qPCR and normalized to L19 (right panel). While TGFβ markedly induced both p21-specific transactivation and transcription in the presence of SMAD4, no increase in p21 transactivation and only a modest increase in transcription following activin treatment in the presence of SMAD4 were found (*p<0.05). B) SMAD4-wild type FET and SMAD4-null SW480 cells were treated with control vehicle (C), activin (A), TGFβ(T), or a combination of both ligands (A+T) for 24 hours prior to lysis for total protein, nuclear, and cytoplasmic preparation. Histone H3, α-tubulin, and GAPDH were used as loading controls for the respective fractions. While TGFβ markedly increased p21 levels in all three fractions in the SMAD4 positive cell line only, activin induced a decrease in nuclear and total p21 protein in SMAD4-positive and -negative cells (left panel). Densitometric analysis of all blots revealed statistically significant changes in p21 levels (right panel) (ns  =  non-significant, *p<0.05, **p<0.01, ***p<0.001). C) Initial upregulation of p21 protein is followed by downregulation by 24 h after activin treatment. SMAD4-wild type FET cells were treated with activin or vehicle (control) and harvested at various time points for quantification of p21 protein expression. GAPDH was used as loading control and relative expression was calculated via densitometry. D) While TGFβ-induced upregulation of p21 was SMAD4 dependent, activin-induced downregulation of p21 was still observed in the absence of SMAD4. SMAD4-wild type FET cells were treated with vehicle (CNT), activin or TGFβ in the presence of either scramble siRNA (SC) or SMAD4 siRNA (KD) and total p21 levels were determined. GAPDH was used as loading and C32 cell lysate as p21 positive control.

Figure 3. p21 mediates activin-induced growth suppression and counteracts activin-induced SMAD4-independent migration in the presence of SMAD4.

A) FET cells were treated with either scramble (SC) or p21 specific siRNA (KD). Growth suppression was assessed by MTT-metabolic assay following activin treatment. Activin induced cell growth inhibition in the presence of p21, but the effect was reversed in the absence of p21 (*p<0.05). B) Total viability is decreased in SMAD4 wild type colon cancers following activin treatment in the presence of p21. FET cells were treated with either scramble or p21 specific siRNA. Cell viability was assessed by trypan blue staining following activin treatment. Trypan blue positive cells after activin treatment were decreased in presence of p21, but increased after p21 knockdown (***p<0.001). C) Activin (A) induces cell migration in SMAD4-positive and SMAD4-negative cell lines. Cellular migration is induced in SMAD4-wild type FET cells and SMAD4-null SW480 cells following activin treatment, but more pronounced induction of migration is seen in the absence of SMAD4. Loss of p21 leads to an increase in baseline migration in SMAD4 expressing cells (*p<0.05, **p<0.01, ***p<0.001). D) p21 knockdown increases the overall pro-migratory effect of activin in FET cells. Loss of p21 in the absence of SMAD4 does further increase migratory induction (*p<0.05, ** p<0.01).

Figure 4. Schematic of proposed differential regulation and effects of activin and TGFβ signaling on p21 in colon cancer cells. * is indicative of total (cytoplasmatic + nuclear) p21.

Figure 5. Activin-induced p21 downregulation is associated with ubiquitination and counteracted by proteasomal degradation.

A) ACVR2/TGFBR2/SMAD4-wild type FET cells were were pretreated for 30 minutes with proteasomal inhibitor MG-132 and then treated with vehicle (control), activin, TGFβ for 24 hours and ubiquitination of total p21 was assessed via immunoprecipitation of p21 and blotting with a ubiquitin-specific antibody (upper panel) and reblotting of p21. Multiple bands indicative of polyubiquitination were seen only following activin treatment. B) Activin-induced p21 downregulation is dependent on the proteasome. SMAD4-wild type FET cells were pretreated for 30 minutes with proteasomal inhibitor MG-132 followed by treatment with vehicle (control) or activin for 24 hours and compared to cells treated accordingly without proteasomal inhibition. p21 expression was assessed and showed inhibition of p21 downregulation following activin treatment in conjunction with proteasomal inhibition.

Figure 6. Expression of p21 is lost in a subset of primary colon cancers correlating with the ACVR2/TGFBR2 receptor status.

Fifty-six colon cancers were stained for ACVR2, TGFBR2 and p21. Representative examples for p21 staining are shown: normal colon tissue with nuclear staining (left panel), colon cancer sample with maintained nuclear p21 staining (middle panel), and colon cancer sample with loss of nuclear p21 staining (right panel).