Background: The most common site of metastasis for men with lethal prostate cancer (PCa) is the bone. In previous studies, Keratin 13 (KRT13), an intermediate filament cytokeratin found in stem-like prostate epithelial cells, was identified as a potential biomarker of bone metastatic progression. Tissue microarray analysis of over 300 localized PCa cases treated with radical prostatectomy at the Greater Los Angeles VA (GLA-VA) revealed 9% of cancer cores expressed KRT13. KRT13+ tumors were associated with significantly increased recurrence, metastases and poor overall survival. In contrast to the relatively small percentage of tumors expressing KRT13 in localized cases, nearly all cases of de novo metastatic PCa (clinical stage M1) were found to express KRT13 in diagnostic prostate needle biopsies. Evaluation of PCa bone metastases revealed KRT13+ tumors infiltrating the bone. Interestingly, megakaryocytes (MKs) within the bone marrow also expressed KRT13. MKs reside in the vascular niche and extend arm-like projections (podosomes) across the endothelium into the circulation to release platelets. Due to this location, MKs may be the first bone marrow residents to encounter circulating tumor cells seeking access to bone. Previous studies have demonstrated that MKs induce apoptosis in co-cultured tumor cells and may serve as a barrier to metastatic progression. We hypothesize that KRT13+ PCa with bone metastatic potential may display “MK-mimicry”, through a transcriptional program that enables podosome/invadopodia (P/I) to form and an MK-like phenotype to bypass the surveillance of the MKs, ultimately displacing them in order to access the bone marrow. The goal of this project is to begin to evaluate the role of KRT13 in P/I formation through a pathway involving the scaffolding protein, tyrosine kinase substrate with 5 Src homology domains, Tks-5.
Methods: A western blot was performed on KRT13+ (BPH-1 and A431) and KRT13- (LNCaP) cell lines in order to compare baseline KRT13 and Tks-5 expression. Each cell line (5x105 total cells) was seeded onto fluorescently-labeled gelatin coverslips and incubated for 4 hours. Coverslips were probed with KRT13, Phalloidin (detects F-actin), and Tks-5 fluorescent-linked antibodies, followed by multiplex imaging using confocal microscopy at 100X. P/I formation was quantitated through gelatin degradation as follows: 7.5x105 cells were seeded on fluorescently-labeled gelatin and incubated for 15 hours. Degradation areas were measured with ImageJ software and normalized to the number of cells imaged per view. 10 views were taken per coverslip with a total of 20 views per cell line.
Results: Western blot confirmed KRT13 expression in A431 and BPH-1 cells. These cells also expressed Tks5. KRT13 and Tks5 were not detected in LnCap cells. Confocal imaging revealed P/I formation and gelatin degradation in BPH-1 and A431. Phalloidin and Tks5 expression co-localized with areas of degradation, indicating Tks5-mediated formation of P/I in these cells. Although KRT13 and Tks-5 expression also co-localized in cytoskeletal structures, KRT13 was excluded from P/I areas overlying gel degradation. P/I quantification demonstrated significantly more activity in A431 and BPH-1 compared to LnCaP cells.
Conclusion: KRT13+ cell lines, A431 and BPH-1, co-express Tks-5 and form P/I. KRT13- LNCaP cells do not express Tks-5 and form significantly fewer P/I. Our data suggests a correlation between KRT13 expression and P/I formation via the Tks-5 pathway. Multiplex co-localization images indicate that although KRT13 expression generally mirrors Tks-5, P/I puncta appear to exclude KRT13. This finding suggests that KRT13 may not be required for P/I formation/function, although it remains to be seen if it contributes to the upstream signaling that initiates P/I formation. Further experiments to evaluate the role of KRT13 in P/I formation include deletion/overexpression of KRT13 in cell lines to evaluate if KRT13 is essential/sufficient for the process or simply indicates that cells possess P/I capability.