Transient SNAIL1 expression is necessary for metastatic competence in breast cancer

Abstract

SNAIL1 has been suggested to regulate breast cancer metastasis based on analyses of human breast tumor transcriptomes and experiments using cancer cell lines and xenografts. However, in vivo genetic experimental support for a role for SNAIL1 in breast cancer metastasis that develops in an immunocompetent tumor microenvironment has not been determined. To address this question, we created a genetic SNAIL1 model by coupling an endogenous SNAIL1 reporter with an inducible SNAIL1 transgene. Using multiple genetic models of breast cancer, we demonstrated that endogenous SNAIL1 expression was restricted to primary tumors that ultimately disseminate. SNAIL1 gene deletion either during the premalignant phase or after primary tumors have reached a palpable size blunted metastasis, indicating that late metastasis was the main driver of metastasis and that this was dependent on SNAIL1. Importantly, SNAIL1 expression during breast cancer metastasis was transient and forced transient, but not continuous. SNAIL1 expression in breast tumors was sufficient to increase metastasis.

Figure 1. Generation and characterization of SNAIL1 reporter and tet-inducible SNAIL1 mice (SNAIL1-CBR)

A, A diagram of the modified SNAIL1 locus in SNAIL1-CBR mice. B, Targeted ES cells were screened by PCR using indicated primers. C, SNAIL1-CBR was properly regulated in targeted ES cells. Representatives of 3 experiments are shown. D, The TRE-SNAIL1-3Flag transgene was expressed in ES cells only when both rtTA and dox were present. E, Uterine SNAIL1-CBR signal in a pregnant female at 7.5 dpc in vivo (left panel), ex vivo (middle 2 panels) and after uterine removal (right panel). N=5 mice. F, SNAIL1-CBR was upregulated in CCl₄-induced liver fibrosis as seen in vivo (left panel) and ex vivo (right panel). N=5 mice G, CCl₄-induced bridging fibrosis (arrows) as revealed by picrosirius stain of liver sections. N=5 mice. H, Whole breast mount stain of SNAIL1-CBR and SNAIL1-CBR; MTA mice treated with dox for 1 week starting at age 4 weeks. N=5 mice. I, Mammary SNAIL1-3Flag mRNA expression was detected by RT-PCR only when both MTA and dox were present.

Figure 2. SNAIL1 expression in primary breast tumors strongly correlates with increased metastasis and an invasive phenotype in MMTV-neuNT mice

N=23 mice without and 15 mice with lung metastasis for all panels. A, A representative image of SNAIL1-CBR bioluminescence in MMTV-NeuNT mice with SNAIL1-CBR positive or negative tumors. B, Representative images of superficial lung metastases as seen grossly (top panel, black arrow) and after Indian ink stain as white lesions on a black background (bottom panel, white arrow). C, A histogram of incidence of lung metastasis in mice with SNAIL1-CBR positive or negative primary tumors. D, Representative images of BM DTCs as detected by Her2Neu IF in mice with SNAIL1-CBR positive or negative primary breast tumors. E, A histogram of number of BM DTCs per 10⁶ BM cells in mice with SNAIL1-CBR positive or negative primary breast tumors. *p=7.4×10⁻⁷. F, Representative images of SNAIL1-CBR positive and negative tumors as seen in situ (upper panels) and ex vivo (lower panels). G, A histogram of distribution of invasive carcinomas and adenomas by gross appearance of SNAIL1-CBR positive and negative tumors. H, Representative images of H&E stained primary tumor sections revealing a typical protrusion into surrounding stroma in a Snail1-CBR positive tumor (black arrow) compared to the smooth contour in a Snail1-CBR negative tumor. (I and J), Representative images of SNAIL1 and E-cadherin IHC of sections from SNAIL1-CBR positive (I) and negative (J) primary tumors, revealing SNAIL1-positive cells (red arrowheads) in an invasive front of SNAIL1-CBR positive tumor where E-cadherin expression was decreased. Mus denotes Muscle.

Figure 3. SNAIL1 expression in primary breast tumors precedes detection of DTCs, but SNAIL1 is undetectable in DTCs and lung metastases in MTA; TAN mice

A, Important milestones for the MTA; TAN breast cancer model. (B–D), Histograms of relative SNAIL1 mRNA levels in developing primary breast tumors as determined by RT-PCR and normalized to GAPDH mRNA (B; *p=0.01), number of BM DTCs per 10⁶ BM cells as quantified by Her2Neu IF (C; **p=0.03), and tumor burden (D) at indicated times. N=5 mice at each time point. E, A histogram of relative SNAIL1 mRNA levels in a set of primary tumor, BM DTCs and isolated lung metastasis from the same mouse as determined by RT-PCR and normalized to GAPDH mRNA (***p=0.0001; ****p=0.002). Representatives of 3 experiments are shown. F, Representative images of the same H&E and IHC stained sections as in Table 1, revealing SNAIL1-positive cells (green arrows) at an invasive front of a primary tumor where E-cadherin was low and Vimentin high (top panel). SNAIL1 was undetectable in BM DTCs, which exhibited mixed phenotype (middle panel), and lung metastasis, which appeared epithelial (bottom panel).

Figure 4. SNAIL1 is necessary for breast tumor metastasis in MMTV-PyMT mice

A, A diagram of breast cancer progression and genetic mouse models used. B, Representative images of SNAIL1 IHC stained sections of primary tumors from Early SNAIL1 KO and control mice, confirming efficient SNAIL1 deletion. C, Representative images of lungs as seen grossly (upper panel) and after Indian ink stain (lower panel) of Control and Early SNAIL1 KO mice. (D and E), Histograms of number of lung metastases as measured by Indian ink stain (D; *p=0.0038), and number of BM DTCs per 10⁶ total BM cells as measured by panCK IF (E; *p=0.00087) in Control and Early SNAIL1 KO mice. N=8 for Control and 6 for Early SNAIL1 KO. (F and G), Histograms of number of lung metastases as measured by Indian ink stain (F; *p=0.04), and number of BM DTCs per 10⁶ total BM cells as measured by panCK IF (G; *p=0.03) in Control and Late SNAIL1 KO mice. N=10 for Control and 6 for Late SNAIL1 KO.

Figure 5. Transient, not continuous, overexpression of SNAIL1 in primary breast tumors increases breast cancer metastasis in MMTV-NeuNT mice

A, A diagram of breast cancer progression and genetic mouse models used. (B and C), Histograms of incidence of lung metastasis as determined by Indian ink stain and histology (B) and number of BM DTCs per 10⁶ total BM cells as determined by Her2Neu IF (C; *p=0.0013; **p=5.28×10⁻⁸; ***p=2.5×10⁻⁵) in Control, Continuous SNAIL1 and Transient SNAIL1 mice. N=32, 23 and 8 for Control, Continuous and Transient SNAIL1, respectively.