TRIM29 promotes bladder cancer invasion by regulating the intermediate filament network and focal adhesion

Abstract

Bladder cancer is a common malignancy whose lethality is determined by invasive potential. We have previously shown that TRIM29, also known as ATDC, is transcriptionally regulated by TP63 in basal bladder cancers where it promotes invasive progression and metastasis, but the molecular events which promote invasion and metastasis downstream of TRIM29 remained poorly understood. Here we identify stimulation of bladder cancer migration as the specific role of TRIM29 during invasion. We show that TRIM29 physically interacts with K14+ intermediate filaments which, in turn, regulates focal adhesion stability. Further, we find that both K14 and the focal adhesion protein, ZYX are required for bladder cancer migration and invasion. Taken together, these results establish a role for TRIM29 in the regulation of cytoskeleton and focal adhesions during invasion and identify a pathway with therapeutic potential.

Fig. 1. TRIM29 regulates invasion and migration of bladder cancer cells.

A Knockout (KO) or knockdown (KD) of TRIM29 expression assessed by immunoblot. NT nontargeting siRNA control. B KO or KD of TRIM29 decreases invasion in a transwell invasion assay. WT wildtype, TKO TRIM29-KO, TKD TRIM29-KD. n = 6. *p < 0.001 by t-test. C Timelapse images of 2D migration assay demonstrate that TRIM29 knockout (TKO) or knockdown (TKD) decreases migration of human bladder cancer cells. White dashed lines indicate the leading edges at start timepoint. Scale bar = 100 µm. D Quantitative analysis of 2D migration assay. Data represent the mean ± STD. n = 50. *p < 0.01, **p < 0.0001, ns = not significant by t-test.

Fig. 2. TRIM29 and K14 are upregulated in migratory cells during bladder cancer spheroid invasion.

A Immunofluorescence images of UM-UC5 spheroids show enriched expression of TRIM29 and K14 as compared to noninvasive cells (center dashed circle, original outline of the spheroid). Scale bar = 100 µm. B Colocalization (white arrows) of filamentous TRIM29 and K14 in multiple human bladder cancer cell lines. Scale bar = 10 µm. C Co-immunoprecipitation of mCherry-tagged K14 and TRIM29 demonstrate a physical association between K14 and TRIM29. Protein loading for input blots: 20 µg/lane. Arrow heads: endogenous K14 (bottom) and mCherry-tagged K14 (top). D TRIM29-KO alters distribution of K14-containing IF and cell morphology in migrating cells. Scale bar = 25 µm. E, F TRIM29-KO in UM-UC5 and UM-UC14 results in altered cell morphology and disordered migration in a modified scratch assay. Dashed white arrow = direction of migration. Scale bar = 50 µm. WT wildtype, TKO TRIM29-KO.

Fig. 3. TRIM29 does not modulate K14 proteasomal degradation or ubiquitination.

A Immunoblots of input for K14-mCherry, TRIM29, ubiquitin-HA, ZYX and PXN in UM-UC14 cells treated with DMSO or MG132 that were used in immunoprecipitations. 30 µg protein/lane. Arrow indicates expected band for K14-mCherry. B Immunoprecipitation performed with anti-mCherry antibody shows pull-down of mCherry-K14 and HA-ubiquitin. C Immunoprecipitation of HA-ubiquitin resulted in pulldown of K14-mCherry.

Fig. 4. TRIM29+ IF terminate at focal adhesions in migrating cells.

TRIM29+ IFs terminate in PXN+ (A, C) and ZYX+ (B, D) focal adhesion sites in filipodia and lamellipodia in invasive UM-UC5 (A, B) and UM-UC14 (C, D) bladder cancer cells. Inset images show higher magnification view. Scale bar = 10 µm.

Fig. 5. TRIM29 regulates focal adhesion.

A–D TRIM29-KO (TKO) reduces ZYX+ focal adhesion sites during UM-UC5 and UM-UC14 bladder cancer spheroid invasion. Magenta: ZYX. Yellow: Actin. Blue: Hoechst 33342+ Nuclei. Scale bar = 10 µm. E Quantitative analysis of ZYX+ focal adhesions (UM-UC5-WT, n = 80; UM-UC5-TKO, n = 51; UM-UC14-WT, n = 33; UM-UC14-TKO, n = 18) or PXN+ focal adhesions (UM-UC5-WT, n = 168; UM-UC5-TKO, n = 194; UM-UC14-WT, n = 261; UM-UC14-TKO, n = 340). Data represent the mean ± STD. *p < 0.05 for t-test.

Fig. 6. The expression of TRIM29 affects the turnover of focal adhesion sites during cell migration.

A Timelapse images of UM-UC5 cells expressing mCherry-Paxillin during cell migration. Arrowhead indicates monitored focal adhesion site. Dashed arrow shows the direction of cell movement. Scale bar = 20 µm. B TRIM29 KO had no effect on the rate of assembly (Ka) but increased the rate of disassembly (Kd) of focal adhesion sites in UM-UC5 and UM-UC14 bladder cancer cells (UM-UC5, n = 21; UM-UC14, n = 24). Data represent the mean + STD. *p < 0.05 for t-test.

Fig. 7. K14 regulates TRIM29-induced focal adhesions and bladder cancer migration.

A, B Re-expression of TRIM29 in the UM-UC14 TRIM29 KO cell line allows recovery of ZYX+ and PXN+ focal adhesion sites, but this effect is abrogated by K14 KD. Vector: expression vector control. T29: TRIM29 re-expression vector. NT: nontargeting siRNA control. K14KD: siRNA targeting K14. Magenta: ZYX or PXN. Yellow: F-Actin. Green: Reintroduced TRIM29. Blue: Hoechst 33342+ Nuclei. Overexpression of TRIM29 increased migration (C) and invasion (D) which was abrogated by knockdown of KRT14 or ZYX. Top row: UM-UC10; bottom row: UM-UC14. Vector: empty expression vector control. OE: overexpression. KD: siRNA mediated knockdown. Data represent the mean + STD. *p < 0.01. ns: not significant for one-way ANOVA.

Fig. 8

Model of the role of TRIM29, K14 and focal adhesion complexes during bladder cancer migration and invasion.