Distinct mechanisms of TGF-beta1-mediated epithelial-to-mesenchymal transition and metastasis during skin carcinogenesis

Abstract

In the present study, we demonstrated that human skin cancers frequently overexpress TGF-beta1 but exhibit decreased expression of the TGF-beta type II receptor (TGF-(beta)RII). To understand how this combination affects cancer prognosis, we generated a transgenic mouse model that allowed inducible expression of TGF-beta(1) in keratinocytes expressing a dominant negative TGF-(beta)RII (Delta(beta)RII) in the epidermis. Without Delta(beta)RII expression, TGF-beta1 transgene induction in late-stage, chemically induced papillomas failed to inhibit tumor growth but increased metastasis and epithelial-to-mesenchymal transition (EMT), i.e., formation of spindle cell carcinomas. Interestingly, Delta(beta)RII expression abrogated TGF-beta1-mediated EMT and was accompanied by restoration of membrane-associated E-cadherin/catenin complex in TGF-beta1/Delta(beta)RII compound tumors. Furthermore, expression of molecules thought to mediate TGF-beta1-induced EMT was attenuated in TGF-beta1/Delta(beta)RII-transgenic tumors. However, TGF-beta1/Delta(beta)RII-transgenic tumors progressed to metastasis without losing expression of the membrane-associated E-cadherin/catenin complex and at a rate higher than those observed in nontransgenic, TGF-beta1-transgenic, or Delta(beta)RII-transgenic mice. Abrogation of Smad activation by Delta(beta)RII correlated with the blockade of EMT. However, Delta(beta)RII did not alter TGF-beta1-mediated expression of RhoA/Rac and MAPK, which contributed to increased metastasis. Our study provides evidence that TGF-beta1 induces EMT and invasion via distinct mechanisms. TGF-beta1-mediated EMT requires functional TGF-(beta)RII, whereas TGF-beta1-mediated tumor invasion cooperates with reduced TGF-(beta)RII signaling in tumor epithelia.

Figure 1

Immunostaining for TGF-β₁, TGF-βRII, and pSmad2 in human skin cancer samples revealed a patchy increase in TGF-β₁ and decrease in TGF-βRII staining in AK and CIS, and the same alterations were uniform in SCC. pSmad2-positive cells were detected in AK and CIS. Scale bar: 40 μm for all panels.

Figure 2

Skin tumor formation and tumor types in transgenic mice. (A) Schematic of the skin chemical carcinogenesis and TGF-β₁ transgene induction protocol. (B) Kinetics of tumor formation. Each point represents the average number of tumors per mouse. (C) H&E staining of TGF-β₁/ΔβRII–transgenic SCC. (D) TGF-β₁–transgenic SPCC. (E) Metastatic lesion in lymph node showing SCC cells surrounded by lymphocytes. (F) Lung metastasis that originated from TGF-β₁/ΔβRII–transgenic SCCs. The dotted line delineates lung tissue adjacent to the metastatic lesion. Scale bar in C: 40 μm for C–F.

Figure 3

TGF-β₁ expression levels in mouse tumors with different genotypes. (A) Results of RT-PCR for TGF-β₁ transgene expression in TGF-β₁– and TGF-β₁/ΔβRII–transgenic SCCs. (B) Results of TGF-β₁–specific ELISA. Each group contained 4–6 samples.

Figure 4

BrdU labeling in skin and papillomas 21 weeks after DMBA initiation and 1 week with (TGF-β₁) or without (Control) TGF-β₁ transgene induction. (A) Control skin adjacent to a papilloma. (B) TGF-β₁–transgenic skin adjacent to a papilloma. (C) Control papilloma. (D) TGF-β₁–transgenic papilloma. Scale bar in A: 20 μm for A and B; 40 μm for C and D.

Figure 5

Immunohistochemistry for pSmad2. Scale bar: 40 μm for all panels.

Figure 6

Immunohistochemical staining for E-cadherin and β- and γ-catenins in primary SCCs from ΔβRII-, TGF-β₁–, and TGF-β₁/ΔβRII–transgenic mice 25 weeks after DMBA initiation and in a lymph node metastasis from TGF-β₁/ΔβRII–transgenic SCCs. Note that all ΔβRII- and TGF-β₁/ΔβRII–transgenic SCCs and TGF-β₁/ΔβRII metastatic cells demonstrated staining for membrane-associated E-cadherin and β- and γ-catenins. Lung metastatic cells from TGF-β₁/ΔβRII–transgenic SCCs revealed a similar staining pattern (data not shown). However, these molecules appeared in the cytoplasm of cells in TGF-β₁–transgenic SCCs. Scale bar: 40 μm for all panels.

Figure 7

Angiogenesis and proteinase expression. (A) CD31 immunofluorescence. CD31 (green) highlights vessels. K14 (red) highlights the epithelial portion of tumors. Scale bar: 75 μm for all panels. (B) Percentage of the stromal area covered by vessels in TGF-β₁/ΔβRII–, TGF-β₁–, and ΔβRII-transgenic and control SCCs. The numbers in parentheses represents the number of tumors examined from each group. *P < 0.05. (C and D) Expression levels of VEGFR1 and VEGF (C) and MMP-2 and MMP-9 (D) detected by RPA in SCCs from TGF-β₁–, ΔβRII-, and TGF-β₁/ΔβRII–transgenic and control mice 25 weeks after DMBA initiation. L32 (C) or cyclophilin (Cyc.; D) was used to normalize the amount of RNA loaded in each lane.

Figure 8

Analysis of Notch, Rho/Rac, and MAPK signaling components in SCC samples with different genotypes. Data were averaged from 5 SCCs from each group. One control SCC with the lowest expression level of individual molecules was assigned the value of 1 arbitrary unit as determined by quantitative RT-PCR in A and B. *P < 0.05 compared with control SCCs; ^†P < 0.05 compared with TGF-β₁–transgenic SCCs. (C) Western blot analysis of MAPK components. A pair of samples from each group is presented. Four to six samples in each group were examined and exhibited patterns similar to the representative samples shown here. p-Erk1, phosphorylated Erk1.

Figure 9

Schematic depicting the potential mechanisms of uncoupling TGF-β₁–mediated EMT and tumor invasion by ΔβRII expression. The dotted lines indicate potential mechanisms of the cooperative effects between TGF-β₁ overexpression and ΔβRII expression on tumor invasion and metastasis.