Fig 2: ALCAM expression is silenced in MUM-2B cells via shRNA, and rescued by re-expression of ALCAM-GFP.Western blots of MUM-2B, ALCAM-silenced sh5, control sh6, and negative control MUM-2C reveal that ALCAM protein expression is nearly completely silenced in sh5, but not in sh6 (A). Tubulin is shown as a loading control. ALCAM expression in the rescue cell line (sh5rxd, containing an ALCAM-GFP construct with point mutations to evade knockdown) is similar to MUM-2B, as shown by western blot using an anti-ALCAM antibody (B; note that the GFP tag adds ∼27 kDa to the size of the protein; reprobing the blot with anti-GFP antibody detects only the rescue band, as expected). Panel (C) shows DAPI-stained nuclei of the sh5 cell line; immunostaining of sh5 confirms nearly complete knockdown of ALCAM protein expression (D, E). Panel (F) shows DAPI-stained sh6 nuclei; ALCAM expression in this cell line is robust, and localized to points of cell contact (G, H). Panel (I) shows DAPI-stained nuclei of the sh5rxd rescue cell line; ALCAM expression and localization in the rescue cells (J, K) is comparable to the sh6 cell line. Scale bar in (K) is 25 microns.

Fig 3: ALCAM-silenced cells display reduced MMP-2 activity.Levels of pro-MMP-2 were assayed in media conditioned by each cell line by gelatin zymography (A). Clear bands indicating pro-MMP-2 activity (which is activated by SDS; a faint “active” cleaved MMP2 band was present in some gels but often too weak for robust quantification) are present in MUM-2B, sh6, and the sh5rxd cell lines, but are reduced in the ALCAM-silenced sh5 cell line. MUM-2C and 2C-ALC display pro-MMP-2 levels that are lower than MUM-2B, and comparable to sh5 cells; the overexpression of ALCAM in 2C-ALC fails to increase pro-MMP-2 activity beyond that of MUM-2C. A minimum of three independent gelatin zymography trials (except 2C-ALC, 2 trials) are quantified in (B). (C) Western blots of cell lysates shows that the activation of pro-MMP-2 in the sh5 cell line is reduced (higher molecular weight band is pro-MMP-2; lower molecular weight band is active MMP-2) compared to MUM-2B, sh6, and the sh5rxd rescue cell line.

Fig 4: Expression of ALCAM in MUM-2C cells does not enhance wound-gap closure speed or invasive capacity.Immunostaining of MUM-2C cells (A) reveals that ALCAM expression is virtually undetectable in these cells. DAPI-stained nuclei are shown in blue; ALCAM staining is shown in red. In the 2C-ALC cell line (B), engineered to stably overexpress ALCAM, DAPI-stained nuclei are shown in blue, and ALCAM staining is shown in red. ALCAM localizes to points of contact between cells in the 2C-ALC cell line. The expression of ALCAM in 2C-ALC is confirmed by western blot in panel (C), and is undetectable in MUM-2C. Tubulin is shown as a loading control. The expression of ALCAM in 2C-ALC did not alter closure rate in a wound-gap assay (D), nor did it enhance invasive capacity of 2C-ALC cells when compared with MUM-2C in a transwell migration assay (t-test; p>0.05; the average number of invasive MUM-2C cells per three 10× fields was 16; E). Growth and survival of MUM-2C and 2C-ALC is similar, as assayed and described in Fig. 3 (F).

Fig 5: Uveal melanoma cell lines display different speeds of migration that correlate with ALCAM expression.Time-lapse imaging of a wound created with a pipet tip in a layer of confluent cells (A) reveals the difference in wound closure in MUM-2B vs. MUM-2C. The speed of 5 uveal melanoma cell lines (minimum 3 trials each) is quantified (B). Both OCM-1A and MUM-2C move at less than 0.5 microns per minute; MUM-2B, C918, and M619 move at approximately 1.2 microns per minute. The speed of OCM-1A and MUM-2C are significantly different when compared to MUM-2B (t-test; 1A vs. 2B, p = 0.0008; 2C vs. 2B, p<0.0001). ALCAM protein expression as assayed by western blot shows that the cell lines fall into two groups – those with detectable ALCAM, and those without detectable protein expression (C). Tubulin is shown as a loading control. RT-PCR analysis of 4 of the 5 cell lines reveals that ALCAM mRNA expression mirrors protein expression (D). GAPDH is shown as a positive control. Percent gap closure was tracked for MUM-2B (E) and MUM-2C (F), and revealed that these two representative cell lines differ in their modes of migration. While the tracking of the wound front is fairly linear in MUM-2B, the cell front of MUM-2C advances in a stop-and-start manner, due to loosely associated single cells progressing into the gap. Error bars are mean ± S.E.M.