1. CAFs were isolated from LSCC and CAFs enhanced LSCC metastasis in vitro and vivo and ROCK1, not ROCK2 was highly expressed in CAFs.
In our study, CAFs, Cancer Para-laryngeal Fibroblasts (CPFs) and NFs were isolated and identified from laryngeal cancer, para-laryngeal cancer and normal tissues. It is well known that in comparison with NFs, the phenotype of CAFs was significantly different. CAFs express specific molecules, including α-SMA, FSP1, NG2 and PDGF-β receptor, et al [12]. The specific markers were determined by Western Blot, the expressions of FAP, α-SMA, FSP1, NG2 and PDGF-βwere significantly elevated in CAFs (**P<0.01, Figure 1A-1B). Here, we established a co-culture system in vitro (Figure 1C), in which fibroblasts were indirectly co-cultured with laryngeal cancer Hep2 cells line, which separated by a semi-permeable membrane (pore size of 0.6 µm). At the appropriate time point, Hep2 cells were either assayed by way of the wound healing, migration, invasion and vivo metastasis. As showed in Figure 1D-1E, Hep2 co-culture with CAFs cells (Hep2/CAFs cells) enhanced movement than Hep2 co-culture with NFs (Hep2/NFs cells, **P<0.01). Migration assay showed that Hep2/NFs cells (105±8.6) migrated less than Hep2/CAFs cells (286±15.2). Invasion assay also showed similar results, Hep2/NFs cells (51±6.5) invaded less than Hep2/CAFs cells (101±4.0, *P<0.05, **P<0.01, Figure 1F–1G). To further illustrate CAFs played a determinant role in metastasis. Hep2/CAFs cells and Hep2/NFs cells respectively inoculated via tail vein into 4-week-old male immunodeficient mice. Six weeks after inoculation, Hep2/CAFs cells (5.8±0.7) demonstrated larger and more frequently lung metastases as compared to Hep2/NFs cells (2±1.5, *P<0.05, Figure 1H–1I). In our study, ROCK1, not ROCK2 considerably high expressed in CAFs in comparison with CPFs and NFs (**P<0.01, Figure 1J–1K). Data above indicated that CAFs were successfully isolated from LSCC and CAFs enhanced metastasis in vivo and vitro and ROCK1, not ROCK2 was highly expressed in CAFs.
2.Deprivation of ROCK1 inhibited movement, migration and invasion and upregulation of ROCK1 possessed opposite results. CAFs with high expression of ROCK1 enhanced LSCC metastasis.
Considering that ROCK1 was highly expressed in CAFs, and lowly expressed in NFs, it was conceivable that ROCK1 might act as a positive regulator of metastasis. To explore the effect of ROCK1 in contribution on metastasis, Y27632 [13] was used to downregulate ROCK1 in CAFs at the appropriate time point, with ROCK1 expression modification confirmed, CAFs were co-cultured with Hep2 cells. Hep2 cells were either assayed by way of the immunofluorescence (IF), Western Blot, wound healing, migration and invasion assays. IF showed that the ROCK1 level was obviously reduced in Hep2 cells co-cultured with CAFs/Y27632 (Hep2/CAFs/Y27632 cells) in comparison with Hep2 cells co-cultured with CAFs/parental (Hep2/CAFs/parental, Figure 2A). The results of Western Blot also showed that levels of ROCK1, FAP, α-SMA, FSP1, NG2 and PDGF-β were significantly reduced (*P<0.05, **P<0.01, Figure 2D-2E). For wound-healing assay, Hep2/CAFs/Y27632 cells were less motile in comparison with Hep2/CAFs/parental cells (*P<0.05, Figure 2B–2C). Similarly, less Hep2/CAFs/Y27632 cells migrated through trans-well chambers (62±8.4) in comparison with Hep2/CAFs/parental cells (134±14.5). Ultimately, invasion assay showed that Hep2/CAFs/Y27632 cells (22±7.3) moved across Matrigel less frequently than Hep2/CAFs/parental cells (46±6.7, *P<0.05, **P<0.01, Figure 2F–2G). All data indicated that deprivation of ROCK1 inhibited movement, migration and invasion of LSCC.
Conversely, plasmid (ROCK1, Myc-ROCK1-Delta3 (1-727), p-CAG-Myc) transfection was used to increase ROCK1 expression in NFs at the appropriate time point, the effect of transfection was showed in Figure 2H. Hep2 cells cultured with NFs/ROCK1 (Hep2/NFs/ROCK1 cells) were the experimental group, and Hep2 cells cultured with NFs/vector (Hep2/NFs/vector cells) were the control group. Hep2 cells were either assayed by way of the Western Blot, wound healing, migration and invasion assays. Western Blot results suggested the levels of ROCK1, FAP, α-SMA, FSP1, NG2 and PDGF-β were markedly enhanced (*P<0.05, **P<0.01, Figure 2K-2L). For wound-healing assay, Hep2/NFs/ROCK1 cells were more motile in comparison with Hep2/NFs/vector cells (**P<0.01, Figure 2I–2J). Similarly, more Hep2/NFs/ROCK1 cells migrated through trans-well chambers (162±8.8) in comparison with Hep2/NFs/vector cells (89±11.4). At last, invasion assay suggested that Hep2/NFs/ROCK1 cells (79±4.24) moved across Matrigel more frequently than Hep2/NFs/vector cells (25±9.7, *P<0.05, **P<0.01, Figure 2M–2N). All data indicated that deprivation of ROCK1 possessed opposite results. These observations suggested CAFs with high expression of ROCK1 could enhance movement, migration and invasion of LSCC.
3. CAFs-derived ROCK1 promoted EMT of LSCC.
Accumulating evidences suggested that EMT, a well-characterized embryological process, had been identified to play a critical role in cancer metastasis [14]. In order to examine the role of CAFs-derived ROCK1 in mediating EMT, Hep2 cells were co-cultured with NFs/ROCK1 (Hep2/NFs/ROCK1 cells) or CAFs/Y27632 (Hep2/CAFs/Y27632 cells) in a previously described co-culture system respectively. At the appropriate time point, Hep2 cells were either assayed by way of the IF and Western Blots. As Figure3A-3B and 3D showed that the expression of E-cadherin was markedly decreased, while the expressions of N-cadherin, Slug and Vimentin were observably increased in Hep2/NFs/ROCK1 cells when compared to Hep2/NFs/vector cells (*P<0.05). In contrast, Figure 3A, 3C and 3E showed that expression of E-cadherin was markedly increased, while the expressions of N-cadherin, Slug and Vimentin were markedly decreased in Hep2/CAFs/Y27632 cells when compared to Hep2/CAFs/parental cells (*P<0.05, **P<0.01). All data indicated CAFs with high expression of ROCK1 promoted EMT in LSCC.
4. Signal molecules of JAK2, STAT3 and ERK1/2 were of great importance in LSCC metastasis.
Tumor metastasis is a multifactorial process [15], signal molecules might play a crucial role in LSCC metastasis. To interrogate the role of signal molecules in CAFs-induced ROCK1 mediating EMT to promote LSCC metastasis. First and foremost, Hep2 cells were co-cultured with NFs/ROCK1 (Hep2/NFs/ROCK1 cells), NFs/vector (Hep2/NFs/vector cells) or CAFs/Y27632 (Hep2/CAFs/Y27632 cells) in a previously described co-culture system respectively. At the appropriate time point, Hep2 cells were assayed by way of the Western Blot. As Figure4A-4B showed that the expressions of p-JAK2, p-STAT3 and p-ERK1/2 were markedly increased in Hep2/NFs/ROCK1 cells when compared to Hep2/NFs/vector cells, *P<0.05, **P<0.01). Consistent with these observations, Figure4C-4D showed that the expressions of p-JAK2, p-STAT3 and p-ERK1/2 were markedly decreased in Hep2/CAFs/Y27632 cells (*P<0.05). All data just indicated that signal molecules of JAK2, STAT3 and ERK1/2 were of great importance in LSCC metastasis, while the upstream and downstream relationship between them were not clear.
5. CAFs-derived ROCK1 mediated EMT to promote movement, migration and invasion via activating ROCK1/JAK2 axis.
To further verify the upstream and downstream relationship between ROCK1 and signal molecules of JAK2, STAT3 and ERK1/2. One of JAK2 inhibitors, AG490 [16], was used to treat CAFs for 0, 12, 24h to knockdown the activation of JAK2. With JAK2 expression modification confirmed, CAFs were co-cultured with Hep2 cells (Hep2/CAFs/AG490 cells) in a previously described co-culture system. Hep2 cells were either assayed by way of the Western Blot, wound healing, migration and invasion assays. As measured by Western Blot, levels of p-JAK2, p-STAT3, p-ERK1/2 were markedly decreased, while levels of ROCK1, ROCK2, JAK2, STAT3 and ERK1/2 were not changed. This indicated that JAK2 was the downstream of ROCK1, and the upstream of STAT3 and ERK1/2 (**P<0.01, Figure 5A-5B). We further examined whether blocking JAK2 by AG490 could also inhibit the tumor-promoting effects on LSCC. In wound-healing assay, Hep2/CAFs/AG490 cells were less motile at 48h in comparison with Hep2 cells co-cultured with CAFs/parental (Hep2/CAFs/parental cells, **P<0.01, Figure 5C-5D). Similarly, less Hep2/CAFs/AG490 cells migrated through trans-well chambers (67±6.8) in comparison with Hep2/CAFs/parental cells (143±7.2). In the end, invasion assay showed Hep2/CAFs/AG490 cells (32±10.5) moved less frequently than Hep2/CAFs/parental cells (63±12.7, *P<0.05, **P<0.01, Figure 5E-5F). These observations matched the significant increase in E-cadherin expression, while the expressions of N-cadherin, Slug and Vimentin were significantly decreased (Figure 5G-5H, *P<0.05, **P<0.01). All data illustrated that CAFs-derived ROCK1 mediated EMT to reinforce movement, migration and invasion via activating ROCK1/JAK2 axis.
6. CAFs-derived ROCK1 mediated EMT to promote movement, migration and invasion via activating ROCK1/JAK2/ STAT3 axis.
To elucidate the role of STAT3 in LSCC metastasis. In the same way, one STAT3 inhibitor, C188-9 [17], was used to treat CAFs for 0, 6, 12h to knockdown the activation of STAT3. With STAT3 expression modification confirmed, CAFs were co-cultured with Hep2 cells (Hep2/CAFs/C188-9 cells) in a previously described co-culture system. Hep2 cells were either assayed by way of the Western Blot, wound healing, migration and invasion assays. Western Blot showed that the levels of p-STAT3, p-ERK1/2 were markedly decreased, while levels of ROCK1, ROCK2, p-JAK2, JAK2, p-STAT3, STAT3, p-ERK1/2and ERK1/2 were not changed (*P<0.05, **P<0.01, Figure 6A-6B). This indicated that STAT3 was the downstream of ROCK1/JAK2, and was the upstream of ERK1/2. We further examined whether blocking STAT3 expression could restrain LSCC metastasis as well. In wound-healing assay, Hep2/CAFs/C188-9 cells were less motile at 48h in comparison with Hep2 cells co-cultured with CAFs/parental (Hep2/CAFs/parental cells, **P<0.01, Figure 6C-6D). Similarly, less Hep2/CAFs/C188-9 cells migrated through trans-well chambers (46±6.4) in comparison with Hep2/CAFs/parental cells (92±6.8). Finally, invasion assay indicated that Hep2/CAFs/C188-9 cells (29±6.1) moved through Matrigel less frequently than Hep2/CAFs/parental cells (44±4.3, *P<0.05, **P<0.01, Figure 6E-6F). Consistent with these observations, Figure 6G-6H showed that the level of E-cadherin was markedly increased, nevertheless the levels of N-cadherin, Slug and Vimentin were markedly decreased (*P<0.05, **P<0.01). Similarly, these data indicated CAFs-derived ROCK1 mediated EMT to promote movement, migration, invasion via activating ROCK1/JAK2/STAT3 signal pathway.
7. CAFs-derived ROCK1 mediated EMT to promote movement, migration and invasion via activating ROCK1/JAK2/STAT3/ERK1/2 axis.
Similarly, one ERK1/2 inhibitor, U0126 [18], was used to treat CAFs for 0, 24, 48h to block activation of ERK1/2. With ERK1/2 expression modification confirmed, CAFs were co-cultured with Hep2 cells (Hep2/CAFs/U0126 cells) in a previously described co-culture system. Hep2 cells were either assayed by way of the Western Blot, wound healing, migration and invasion assays. Western Blot showed that the level of p-ERK1/2 was markedly decreased, while the levels of ROCK1, ROCK2, p-JAK2, JAK2, p-STAT3, STAT3, and ERK1/2 were not changed (*P<0.05, Figure 7A-7B). This indicated that ERK1/2 was the downstream of ROCK1/JAK2/STAT3. We further examined whether blocking ERK1/2 expression could also inhibit the tumor-promoting effects on LSCC. In wound-healing assay, Hep2/CAFs/U0126 cells were less motile at 48h in comparison with Hep2 cells co-cultured with CAFs/parental (Hep2/CAFs/parental cells, **P<0.01, Figure 7C-7D). Similarly, less Hep2/CAFs/ U0126 cells migrated through trans-well chambers (67±10.3) in comparison with Hep2/CAFs/parental cells (91±11.7). Ultimately, the results of invasion assay manifested that Hep2/CAFs/U0126 cells (23±4.8) moved through Matrigel less frequently than Hep2/CAFs/parental cells (49±6.5, **P<0.01, Figure 7E-7F). These observations match what the expression of E-cadherin was markedly increased, while the expressions of N-cadherin, Slug and Vimentin were markedly decreased (*P<0.05, **P<0.01, Figure 7G-7H). These data indicated CAFs-derived ROCK1 mediated EMT to aggrandize the movement, migration and invasion ability of LSCC via activating ROCK1/JAK2/STAT3/ERK1/2 axis.
8. Blocking ROCK1/JAK2/STAT3/ERK1/2 axis impaired LSCC metastasis induced by CAFs in vivo.
In our study, we further estimated the contribution of ROCK1/JAK2/STAT3/ ERK1/2 axis on LSCC metastasis in vivo. Hep2 cells, co-cultured with CAFs/Y27632, CAFs/AG490, CAFs/C188-9 and CAFs/U0126 in a previously described co-culture system respectively, were experimental groups, Hep2 cells co-cultured with CAFs were control group. Then Hep2 cells were inoculated via tail vein into 4-week-old male immunodeficient mice. Six weeks after inoculation, as showed in Figure8A-8B Hep2/CAFs (5±1.7) attested larger and more frequently lung metastases with respect to Hep2/CAFs/Y27632 (1±1), Hep2/CAFs/AG490 (0.5±0.5), Hep2/CAFs/C188-9 (0.5±0.5) and Hep2/CAFs/U0126 cells (1±0.5, *P<0.05). All results suggested that CAFs-derived ROCK1 via JAK2/STAT3/ERK1/2 pathway promotes LSCC metastasis in vivo.