HGF-induced invasion by prostate tumor cells requires anterograde lysosome trafficking and activity of Na+-H+ exchangers

Abstract

Hepatocyte growth factor (HGF) is found in tumor microenvironments, and interaction with its tyrosine kinase receptor Met triggers cell invasion and metastasis. It was previously shown that acidic extracellular pH stimulated peripheral lysosome trafficking, resulting in increased cathepsin B secretion and tumor cell invasion, which was dependent upon sodium-proton exchanger (NHE) activity. We now demonstrate that HGF induced the trafficking of lysosomes to the cell periphery, independent of HGF-induced epithelial-mesenchymal transition. HGF-induced anterograde lysosome trafficking depended upon the PI3K pathway, microtubules and RhoA, resulting in increased cathepsin B secretion and invasion by the cells. HGF-induced NHE activity via increased net acid production, and inhibition of NHE activity with 5-(N-ethyl-N-isopropyl)-amiloride (EIPA), or a combination of the NHE1-specific drug cariporide and the NHE3-specific drug s3226 prevented HGF-induced anterograde trafficking and induced retrograde trafficking in HGF-overexpressing cells. EIPA treatment reduced cathepsin B secretion and HGF-induced invasion by the tumor cells. Lysosomes were located more peripherally in Rab7-shRNA-expressing cells and these cells were more invasive than control cells. Overexpression of the Rab7 effector protein, RILP, resulted in a juxtanuclear location of lysosomes and reduced HGF-induced invasion. Together, these results suggest that the location of lysosomes is an inherently important aspect of invasion by tumor cells.

Fig. 1.

Overnight treatment with HGF induces cell scattering and the trafficking of lysosomes to the cell periphery. (A,B) Representative merged IF images depicting the trafficking of lysosomes (red) before and after overnight HGF treatment; shown also is actin (green) and nuclei (blue; top row). (C) Quantification of lysosomal distribution, shown as mean distance from individual cell nuclei. Error bars represent the s.e.m. of 30 cells from at least three independent experiments; *denotes statistical significance (P<0.0001) versus the control (Cntl). (D-I) Cathepsin B (green; D,G) and LAMP-1 (red; E,H) colocalize (F and I) in the presence (G-I) or absence (D-F) of HGF. Arrows indicate the same vesicles in each image. Scale bars: 10 μm.

Fig. 2.

The PI3K pathway is required for HGF-induced lysosome trafficking. (A) Western blot analysis of DU145 cells indicates that HGF rapidly induced a sustained phosphorylation of Met (Tyr1234/1235), Akt (Ser483), Erk (The202/Tyr204) and Jnk (The183/Tyr185); actin is shown as a protein load control. (B) Cells were pretreated with LY294002 for 30 minutes before the addition of HGF overnight and the location of lysosomes (red) were observed using IF microscopy; shown also are actin (green) and nuclei (blue). (C) Quantification of lysosomal distribution is shown as mean distance from individual cell nuclei. Error bars represent s.e.m. of 30 cells from at least three independent experiments; *statistical significance (P<0.0001) versus control (Cntl). Scale bars: 10 μm.

Fig. 3.

Effect of HGF on pH_i and pH_e. Treatment with HGF overnight (A) increases net acid production, causing acidification of the extracellular pH (C), but not the intracellular pH (B). This suggests the activity of sodium-proton exchangers in mediating intracellular pH homeostasis in the presence of HGF. See Materials and Methods for more information.

Fig. 4.

EIPA, a broad inhibitor of sodium-proton exchangers, prevents and reverses HGF-induced lysosome trafficking, cathepsin B secretion, and invasion. (A) DU145 cells were pretreated with EIPA for 30 minutes before the addition of HGF overnight, followed by IF microscopy to visualize lysosomes (red); also shown are actin (green) and nuclei (blue). (B) IF images of a DU145 HGF-overexpressing cell line after two days in culture. Lysosomes have undergone anterograde trafficking in these cells, which did not occur in vector control cells, and EIPA was able to reverse this effect and induce lysosomes to traffic toward the cell nucleus. (C) Quantification of lysosomal distribution is shown as mean distance from individual cell nuclei. Xs indicate the addition of HGF and/or EIPA; O/E, overexpressing cells. Error bars represent s.e.m. of 30 cells from at least three independent experiments. (D) DU145 cells were treated for 24 hours with EIPA and/or HGF and the amount of cathepsin B secreted into the culture medium was determined. (E) DU145 cells were seeded onto Matrigel-coated transwell chambers and allowed to invade for 24 hours in the presence of EIPA and/or HGF. See Materials and Methods for more information. *Statistical significance (P<0.0001) versus –HGF control; **statistical significance (P<0.001) versus vector control. Scale bars: 10 μm.

Fig. 5.

NHE1 and NHE3 are both involved in HGF-induced lysosome trafficking. (A) DU145 cells were pretreated for 30 minutes with 10 μM cariporide and/or 10 μM s3226 before the addition of HGF overnight, followed by IF microscopy to visualize the distribution of lysosomes (red); also shown are actin (green) and nuclei (blue). Cariporide and s3226 work additively to prevent HGF-induced lysosome trafficking. (B) Quantification of lysosomal distribution is shown as mean distance from individual cell nuclei. Error bars represent s.e.m. of 30 cells from at least three independent experiments; *Statistical significance (P<0.0001) versus control; **statistical significance (P<0.001) versus HGF; ^$statistical significance (P<0.01) versus HGF. Scale bars: 10 μm.

Fig. 6.

Rab7 is required for EIPA-mediated prevention of HGF-induced lysosome redistribution, cathepsin B secretion and invasion. (A) Representative merged IF images depicting the effects on the distribution of lysosomes (red); also shown are actin (green) and nuclei (blue) in vector control and Rab7-shRNA-expressing cells. EIPA requires Rab7 to prevent HGF-induced lysosome trafficking. HGF was added overnight after a 30 minute pretreatment with EIPA. (B) Western blot analysis indicates efficient Rab7 knockdown in one of the cell pools expressing Rab7 shRNA. Tubulin is shown as a protein load control. (C) Quantification of lysosomal distribution is shown as mean distance from individual cell nuclei. Error bars represent s.e.m. of 30 cells from at least three independent experiments or s.e.m. from three independent assays. (D) Rab7-shRNA-expressing cells secrete more cathepsin B than non-target (NT)-shRNA-expressing cells and EIPA no longer reduces secretion in Rab7-shRNA-expressing cells. See Materials and Methods for secretion assay details. (E) Invasion assay indicates that Rab7-shRNA-expressing cells are more invasive and EIPA does not prevent HGF-induced invasion when Rab7 is downregulated. See Materials and Methods for more assay details. (F) IF microscopy shows that lysosomes in cells expressing WT-RILP-GFP are clustered near the cell nucleus. Lysosomes are shown in red and transfected cells in green. Yellow color represents WT-RILP-GFP and lysosome colocalization. (G) Invasion assays indicate that RILP overexpression prevents HGF-induced invasion. *Statistical significance (P<0.0001) versus NT cntl; **statistical significance (P<0.001) versus NT cntl; ^$statistical significance (P<0.0001) versus NT HGF treated. Scale bars: 10 μm.

Fig. 7.

EIPA increases vesicular-associated cholesterol, which colocalizes with LAMP-1. (A) Filipin staining demonstrates that EIPA, but not HGF induces an increase in vesicular cholesterol as demonstrated by punctate staining. (B) IF microscopy shows that cholesterol-rich vesicles (flipin staining; green) colocalize with LAMP-1 (red). EIPA induces these cholesterol- and LAMP-1-positive vesicles to enlarge and accumulate in the perinuclear region. Scale bars: 10 μm (A); 5 μm (B).