Table 1.
Clone, dilution and sources of primary antibodies.
Fig 1.
HE staining showing the histopathological features of two ACCs.
(A) Case 1. Tumor cells arranged in typical cribriform growth patterns. (B) Case 2. Tumor cells arranged in tubular growth patterns. (C) Immunohistochemical staining of α-SMA for Case1. Arrows pointed to the CAFs in the stroma of case 1. (D) Immunohistochemical staining of α-SMA for Case2. Arrows pointed to the CAFs in the stroma of case 2. Scale bar = 100μm
Fig 2.
Immunofluorescent staining for pan-CK, VIM, α-SMA, FAP, FSP-1 in NF, CAF-A1, and CAF-A2.
(A) NF was negative for pan-CK, α-SMA, FAP, FSP-1 and positive for VIM. (B) CAF-A1 was negative for pan-CK and positive for VIM, α-SMA, FAP, FSP-1.(C) CAF-A2was negative for pan-CK and positive for VIM, α-SMA, FAP, FSP-1. Scale bar = 100 μm
Fig 3.
Migration and invasion of NF, CAF-A1, and CAF-A2.
(A) Wound healing assay. Both CAF-A1 and CAF-A2 showed significantly increased migration activity compared to NF. (B) Transwell® migration assay. Significantly more CAF-A1 and CAF-A2 cells transmigrated through the pores of the Transwell® than NF. (C) Transwell® invasion assay. Significantly more CAF-A1 and CAF-A2 cells invaded through the matrix coating on the Transwell® membrane compared toNF. (n = 3)
Fig 4.
Microfluidic-based cell invasion assay.
(A) Illustration of the microfluidic device. (B) Photo of the microfluidic device filled with red stain. (C) Experimental design. Fibroblasts (red) were cultured in the cell culture channel of a microfluidic device.(D) Cell invasion assay using a microfluidic device. The invasion areas of CAF-A1 and CAF-A2 were significantly greater than that of NF. ** P< 0.01, *** P< 0.001, n = 5. Scale bar = 100 μm
Fig 5.
The effects of CM prepared from NF, CAF-A1, and CAF-A2 on the migration and invasion of SACC-LM and SACC-83 cells.
(A-B) ACC migration assay. CAF-A1-CM promoted SACC-LM migration significantly more than NF-CM. CAF-A2-CM also promoted SACC-LM migration more than NF-CM, but the difference was not significant (A). Both CAF-A1-CM and CAF-A2-CM promoted SACC-83 migration significantly more than NF-CM (B). (C-D) ACC invasion assay using Transwell® plates. CAF-A1-CM and CAF-A2-CM promoted SACC-LM (C) and SACC-83 (D) invasion significantly more than NF-CM. * P< 0.05, *** P< 0.001, n = 3. Scale bar = 100 μm
Fig 6.
The effects of NF, CAF-A1, and CAF-A2 on the invasion of SACC-LM and SACC-83 cells.
(A) Experimental design. Fibroblasts (red) and ACC cells (green) were co-cultured in the cell culture channel of a microfluidic device.(B-C) Co-culture of NF, CAF-A1, or CAF-A2 with ACC cells in a microfluidic device. CAF-A1 and CAF-A2 promoted SACC-LM (B) and SACC-83 (C) invasion. CAFs localized at the invasion front and ACC cells followed the CAFs (Inset). Inset represents the green dotted square. (D-E) The invasion areas of SACC-LM (D) and SACC-83(E) were significantly increased in CAF-A1 and CAF-A2 co-culture groups compared to that in the NF co-culture group. * P< 0.05, n = 3. Scale bar = 100 μm
Fig 7.
Inhibition of ACC invasion promoted by CAFs.
(A) GM6001 inhibited both SACC-LM and SACC-83 invasion promoted by CAF-A1 and CAF-A2. (B) CXCL12 expression was confirmed in CAF-A1 and CAF-A2. NF showed low CXCL12 expression. (C) AMD3100 inhibited SACC-LM and SACC-83 invasion led by CAFs. * P< 0.05, n = 3. Scale bar = 100 μm