DACH1 suppresses breast cancer as a negative regulator of CD44

Dachshund homolog 1 (DACH1), a key cell fate determination factor, contributes to tumorigenesis, invasion, metastasis of human breast neoplasm. However, the exact molecular mechanisms for the anti-tumor roles of DACH1 in breast carcinoma are still lack of extensive understanding. Herein, we utilized immunohistochemistry (IHC) staining and public microarray data analysis showing that DACH1 was higher in normal breast, low-grade and luminal-type cancer in comparison with breast carcinoma, high-grade and basal-like tumors respectively. Additionally, both correlation analysis of public databases of human breast carcinoma and IHC analysis of mice xenograft tumors demonstrated that DACH1 inversely related to cancer stem cells (CSCs) markers, epithelial-mesenchymal transition (EMT) inducers and basal-enriched molecules, while cluster of differentiation 44 (CD44) behaved in an opposite manner. Furthermore, mice transplanted tumor model indicated that breast cancer cells Met-1 with up-regulation of DACH1 were endowed with remarkably reduced potential of tumorigenesis. Importantly, meta-analysis of 19 Gene Expression Omnibus (GEO) databases of breast cancer implicated that patients with higher DACH1 expression had prolonged time to death, recurrence and metastasis, while CD44 was a promising biomarker predicting worse overall survival (OS) and metastasis-free survival (MFS). Collectively, our study indicated that CD44 might be a novel target of DACH1 in breast carcinoma.


Results
DACH1 and CD44 associated with tumorigenesis and histological grade of breast cancer. In order to evaluate the expression of DACH1 and CD44 in normal breast and breast malignant tissues, we carried out IHC analysis on two TMAs (BR1502-97 and BR1502-98) with normal breast and human breast cancer tissues. DACH1 was majorly found in nucleus and CD44 was mostly detected on the membrane of breast cancer cells. Representative images of IHC staining for noncancerous and cancerous tissues were shown in Fig. 1a, showing that DACH1 decreased and CD44 increased in breast neoplasm tissues in comparison with normal breast.
Additionally, we also explored the correlation between the protein abundance of DACH1 and CD44 and histological grade. Representative images of IHC staining for low-grade and high-grade cancerous tissues were showed in Fig. 1b, which indicating that DACH1 was inversely correlated with tumor grade, while CD44 was positively associated with histological grade.
Expression of DACH1 and CD44 correlated with molecular subtypes of breast cancer. We also carried out IHC staining to assess the protein abundance of DACH1 and CD44 in luminal-type and basal-like breast cancer tissues. Representative images for the expression of DACH1 and CD44 in luminal and basal tissues were shown in Fig. 2a. Furthermore, expression analysis of GSE20711 including a total of 45 luminal and 22 basal-like breast tumor cases was also interrogated to evaluate the mRNA levels of DACH1 and CD44 in luminal and basal-like breast neoplasm tissues, which showed that DACH1 was enriched in luminal breast carcinoma in comparison with basal-like breast cancer (P < 0.001) (Fig. 2b), while CD44 exhibited an opposite tendency (P < 0.001) (Fig. 2c) at mRNA level. Altogether, our results indicated that luminal breast carcinoma was most likely to be DACH1 high /CD44 low type, and basal-like breast tumor tissues were majorly DACH1 low / CD44 high type.

DACH1 down-regulated some CSC and EMT markers in vitro, as well as blocked Met-1 tumor growth in vivo.
Breast cancer Met-1 cells were transducted with a DACH1 expression vector resulting in an ∼4.5-fold increase in DACH1 expression (Fig. 3a) and subsequent reduction of CSC markers CD44, KLF4 and MYC as well as EMT markers including FN1 and VIM (Fig. 3b) by mRNA analysis. Western blot also demonstrated the presence of the DACH1-tagged FLAG epitome, and the effects of DACH1 overexpression on the protein abundance of CD44, Fibronectin, Vimentin, p21 and p27, which were showed in Fig. 3c. Ectopic expression of DACH1 contributed to remarkable reduction of CD44, Fibronectin, Vimentin and significant up-regulation of p21 and p27 in Met-1 cells. Mammary tumor growth in vivo was assessed by subcutaneous implantation of Met-1 cells in nude mice (Fig. 3f). Met-1 cells with engineered expression of DACH1 were endowed with remarkably reduced potential of tumorigenesis in xenograft tumors. Up-regulation of DACH1 significantly reduced the volume of tumors by ∼90% and slowed down tumor growth (Fig. 3d). Tumor weight was also reduced by ∼90% in comparison with the control tumors (Fig. 3e). The results implicated that DACH1 suppressed the expression of some CSCs and EMT markers and serves as a potent anti-tumor factor in xenograft tumors. Representative images of immunohistochemistry staining of DACH1 and CD44 in luminal-type and basallike breast cancer tissues were shown. (b) Expression analysis of public microarray dataset GSE20711 showed that the mRNA level of DACH1 was significantly higher in luminal-type than in basal-like breast tumor. (c) Expression analysis of GSE 20711 also displayed that CD44 mRNA expression was remarkably lower in luminaltype than in basal-like breast cancer.

Discussion
DACH1, as an important member of RDGN, is widely expressed in epithelial cells and plays critical roles in normal organ development 45,46 . However, its absent or lower expression contributes to the initiation and progression of various tumor types, including lung adenocarcinoma 14 , pancreatic cancer 47 , breast cancer 10 and gastric cancer 48 . This study provided results that expression types of DACH1 and CD44 were opposite in breast cancer. Overexpression of DACH1 reduced the levels of CSC and EMT markers in breast cancer cell line and potently inhibited the ability of tumorgenesis in xenograft model. Besides, correlation analysis exhibited that DACH1 and CD44 behaved completely differently in the correlations with FN1, VIM, YBX1, FOXA1, EGFR and MKI67. Importantly, DACH1 serves as a protective factor, while CD44 is an unfavorable element for the prognosis of breast cancer patients. Thus, we concluded that DACH1 might exert inhibitory effects on the development of breast cancer partly by suppression of EMT inducers and CSCs markers, especially CD44.
Carcinogenesis may derive from the acquisition of a plethora of oncogenic mutations, sequential suppression of endogenous growth inhibitors 49 and the loss control over pivotal cellular functions 50 . DACH1 accounts for the carcinogenesis of various tumor types, including human breast cancer 51 . Previous studies have implicated that DACH1 negatively regulated cellular proliferation and tumor growth by repressing cell cycle protein cyclin D1 in both breast cancer 11 and renal clear cell cancer 16 . Restoration of DACH1 suppressed clone formation of renal clear cell cancer cells in vitro as well as tumor growth in vivo through the inhibition of cyclin D1 transcription via associating with AP-1 protein 16 . Furthermore, hypermethylation of DACH1 promoter region itself led to carcinogenesis and it also exerted inhibitory effects on tumor initiation through activating transforming growth factor-beta (TGF-beta) signaling 15 . Besides, DACH1 inhibited cellular growth in an NAD and p53-dependent manner by direct protein-protein association in breast cancer 52 . On the contrary, CD44, a family of transmembrane glycoproteins, exerted promoting roles in tumorigenesis of various cancer types 20 . Up-regulation of the cleaved intracellular domain of CD44 (CD44ICD) enriched the mammosphere formation, while the blockage of CD44ICD reversed the effects 53 . Immunohistochemistry analysis on human breast cancer tissues and expression analysis of GSE42568 indicated that CD44 remarkably increased in breast carcinoma in comparison with normal breast tissues 22 . CD44 functions in carcinogenesis through binding to extracellular matrix components and messenger molecules in tumor environment 54 . Several lines of evidence indicated that DACH1 expression correlated with tumor differentiation. Immunohistochemistry analysis of clear cell renal cancer tissues displayed that DACH1 was inversely correlated with tumor grade 16 . About 60% of cells in low-grade tumors expressed DACH1, and less than 20% cells in grade III tumors expressed DACH1 16 . Similar phenomenon was also found in breast cancer 13 . DACH1 was remarkably enriched in low-grade breast tumors 13 . In contrast, CD44 positively correlated with breast tumor grade 21,22 . Our previous study displayed a positive association between CD44 expression and breast tumor grade at both mRNA and protein levels 21,22 . This paper further supported this opposite tendency of DACH1 and CD44 in low-grade and high-grade breast carcinoma, respectively.
Majorly according to the status of ER, PR and Her2, breast cancer is grouped into four distinct subtypes including luminal, Her2-positive, basal-type and normal-like. Among these, luminal tumor patients accounted for the most part of breast carcinoma population and have relatively good clinical outcomes, while patients . DACH1 and CD44 were related to clinical outcomes of breast cancer patients. Meta-analysis of a total of 19 public databases showed that breast cancer patients with higher DACH1 mRNA level tended to have better overall survival (a), relapse-free survival (c), metastasis-free survival (e). On the contrary, patients with higher CD44 mRNA expression had shorter time to death (b) and metastasis (f), but did not acquire statistically significant worse relapse-free survival in comparison to patients with comparatively lower CD44 expression (d).
with basal-like tumor endowed with malignancy features have significantly poor prognosis. Previous study has demonstrated that nuclear DACH1 is a biomarker of luminal breast cancer 13 . Breast tumors with positive ERalpha and co-expressing PR-alpha were most likely to highly express DACH1, and nuclear DACH1 expression was positively correlated with luminal marker FOXA1 and inversely associated with basal-like markers EGFR 13 . In contrast, our previous meta-analysis showed that CD44 mRNA was remarkably enriched in basal-like breast cancer compared with luminal-type breast tumor 21 . CD44 was also positively correlated with basal markers EGFR, KRT5 and KRT17, and inversely associated with luminal marker FOXA1 22 . Our results supported that luminal breast neoplasm tended to be with high DACH1 expression and low CD44 level, while basal-like tumors were most likely to be the inverse type. The correlation analysis further indicated that DACH1 was significantly inversely associated with CD44.
CSCs were composed of a subset of tumor cells with the expression of stem cell-associated markers and enhanced capacity for tumorigenesis, metastasis and therapy-resistance 55 . It has been implicated that endogenous DACH1 participated in the negative regulation of CSCs 18 . DACH1 suppressed the expression of stem cell markers SOX2, Nanog and KLF4 through binding to the promoters of these genes 18 . On the contrary, CD44, as a well-known CSCs marker, positively monitored the levels of CSCs markers 20 . Nuclear location of cleaved CD44 intracellular domain transcriptionally activated stemness factors Nanog, Sox2 and Oct4 53 .
EMT is a complex and highly conserved process which enhances cellular invasiveness, being critical for metastasis of various solid tumors and considered to be promising therapeutic targets 56 . EMT was dysregulated by a complex network during tumor development. Knock-down of DACH1 in breast cancer cells MCF-7 and T47D promoted the morphology change from epithelial phenotype to mesenchymal pattern and interfered with cellcell contact, accompanied by down-regulation of epithelial marker E-cadherin, resulting in cell migration and invasion 10 . DACH1 also transcriptionally suppressed the activity of Snail, leading to the activation of E-cadherin in breast cancer cells 10 , but the complex of DACH1 and Snail could bind to the E-box of E-cadherin promoter then contributing to the reduction of E-cadherin 10 . In addition, DACH1 reduced the expression of the mesenchymal marker Snail through suppressing the activity of the Y box-binding protein, an important EMT inducer 9 . Besides, previous study has revealed that DACH1 decreased both in breast cancer cell lines and tissues accompanied by relatively high proportion of CSCs 18 . Inversely, CD44 not only promoted EMT, but also was upregulated by some mesenchymal markers 20 . Previous studies have demonstrated that mesenchymal genes including ZEB1, TWIST1, SNAI1 and SLUG were positively correlated with CD44 expression 20 . The switch from CD44 variant isoforms to its standard isoform is essential to undergo EMT for normal epithelial cells 57 .
DACH1 played important roles in invasion and migration of various neoplasms, such as lung adenocarcinoma 14 , gastric cancer 48 , pancreatic cancer 47 and breast cancer 12 . Several lines of evidence showed that DACH1 is absent or suppressed in poor prognosis breast cancer 58 . Breast cancer patients with reduced DACH1 expression had three years shorter time to death in comparison to those with normal DACH1 levels 11 . Immunohistochemistry analysis showed that higher nuclear level of DACH1 was predictive of longer disease-free interval, cancer relapse-free survival and distant metastasis-free survival over 5 years post diagnosis 13 . In consistence, our analysis of GSE databases showed that higher mRNA level of DACH1 was parallel with better OS, RFS and MFS. The protected effect of DACH1 in the prognosis of breast cancer patients could be explained partly by negatively interplaying CSCs and EMT 10,59 . Besides, DACH1 inhibited breast cancer migration and invasion also via suppressing oncogene function through targeting interleukin-8 12 . Altogether, these studies suggested that DACH1 could be a valuable molecular marker for prognosis, thereby detection of DACH1 level is useful for therapeutic stratification of breast cancer patients. In contrast, CD44 contributed to tumor formation and progression predictive of poor clinical outcome 20 . CD44 functioned as an oncoprotein and knockdown of CD44 remarkably attenuated the migration and invasion of breast cancer cells MDA-MB-231 and Hs578T by modulating c-Src transcription 60 . According to the analysis of 448 primary breast tumors, CD44 was parallel with enhancive distant recurrence and decreased disease-free survival 61 .
In this study, 19 public datasets were enrolled to assess the correlation between the mRNA levels of DACH1 and CD44 and the survival performance of breast cancer patients. However, there are still some limitations: 1) the overall sample size is limited. Some data were not available when the meta-analysis was conducted; 2) the platforms used to evaluate the mRNA expression of DACH1 and CD44 are different; 3) the results at the mRNA levels might not be consistent with those at protein abundance because there is a complex regulation network from mRNA to protein.
In conclusion, this study confirmed that DACH1 and CD44 inversely related in breast cancer, different grade tumors and different subtypes. DACH1 was negatively associated with CD44 in vitro and vivo. DACH1 served as a good prognostic marker, and CD44 was an unfavorable element of breast cancer patients. Thus, we concluded that CD44 might be a novel target of DACH1.

Materials and Methods
Immunohistochemical staining. To  Analysis and quantification of staining. For quantification, a total of six 200× magnifications of each kind of protein were selected for IHC scoring. Two experienced pathologists assessed the immunohistochemical score independently. Scores were calculated on intensity and proportion of positive staining tumor cells in the whole tissue stains according to the Fromowitz Standard as described above 63 . The staining intensity was scored as 0 (no staining), 1 (weak staining, light yellow), 2 (moderate staining, yellow brown) and 3 (strong staining, brown). The proportions of stained tumor cells were classified as 1 (0-25% positive cells), 2 (26-50% positive cells), 2 (51-75% positive cells) and 3 (76-100% positive cells). The multiplication for intensity and proportion was utilized to represent the protein levels of DACH1, CD44, Myc, Sox2, Fibronectin, Vimentin, EGFR and Ki-67.
Cell culture. The breast cancer cell line Met-1, was cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS, Life Technologies, Inc.). All cells were grown 37 °C in a humidified incubator with 5% CO 2 .
Western blot. Cells were washed with cold PBS, scraped into RIPA buffer and centrifuged. The cell lysates were subjected to 10% SDS-PAGE and transferred to a polyvinylidene fluoride (PVDF) hybridization transfer membrane. The primary antibodies used were as follows: HA (Sc-7392, Santa Cruze), CD44 (15675-1-AP, ProteinTech group), Fibronectin (sc-8422, Santa Cruze), Vimentin (5741, Cell Signaling Technology), p21cip1 (sc-6246, Santa Cruze), p27kip2 (sc-1641, Santa Cruze) and β-actin (Sc-47778, Santa Cruze). Secondary staining and detection were carried out in accordance with standard protocols 16, 64 . RNA profiling by microarray. DNA-free total RNA isolated from Met-1 cells expressing GFP or DACH1 were used to probe Affymetrix Gene 1.0 arrays (Affymetrix, Santa Clara, CA). RNA quality was determined by gel electrophoresis. Analysis of the arrays was performed using GeneSpring. Arrays were normalized using robust multi-array analysis, and the p value of 0.05 was applied as a statistical criterion for differentially expressed genes.

Meta-analysis for DACH1 on published Gene Expression Omnibus (GEO) databases.
We carried out a comprehensive search of relevant GEO databases for mRNA expression of DACH1 and CD44 through ArrayExpress and Oncomine. The datasets meeting the following criteria were included: 1) the datasets were about human breast cancer; 2) the mRNA expression of DACH1 and CD44 was measured in these databases; 3) clinical outcomes of patients were showed in these databases; 4) the sample capacity was more than 50. Only the latest and most complete datasets were included when several databases shared common patients. At last, a total of 19 independent human breast cancer microarray databases with the mRNA expression of DACH1 and CD44 and required the survival information of breast cancer patients were enrolled in this systematic analysis.
Cutoff value for DACH1 and CD44 was median expression. OS, RFS and MFS were evaluated by Cox proportional hazard ratio (HR) and 95% confidence interval (95% CI). HRs was employed to assess the survival outcome of breast cancer patients with high mRNA expression of DACH1 and CD44 and HR >1 indicated that high expression of DACH1 and CD44 predicted worse survival of patients. The random-effect model was employed when heterogeneity was present, and the fixed-effect model was used when homogeneity was demonstrated. Heterogeneity of publication was evaluated by means of the Chi-square-based-Q statistic and inconsistency index (I 2 ) statistic. Statistical analysis was performed based on the guidelines of Meta-Analysis of Observational Studies. The STATA software package (version 12.0) (Stata Corp LP, College Station, TX, USA) was employed to perform the meta-analysis.
Analysis of gene expression data. GSE20711, available through GEO databases and containing 45 luminal-type and 22 basal-like breast carcinoma cases, was analyzed to evaluate the mRNA expression of DACH1 and CD44 in basal-like carcinoma in comparison with that in luminal-type tumors.
GSE20685, containing a total of 327 patients of breast cancer, was interrogated to assess the correlation between DACH1 and CD44, FN1, VIM, FOXA1, EGFR, MKI67 as well as the correlation between CD44 and FN1, VIM, YBX1, FOXA1, EGFR and MKI67 in breast cancer tissues.
Breast cancer cell line data of 51 breast cancer cell lines of different molecular subtypes published by Neve RM, et al. 25 were also employed to evaluate whether the correlation in breast cancer cells was consistent with that in breast tumor tissues.
Nude Mice Study. The animal protocols were approved by the ethics committee of the Tongji Hospital of Huazhong University of Science and Technology. The methods used in this section were in accordance with the relevant guidelines and regulations. 1 × 10 5 Met-1 cells expressing GFP control or DACH1 were implanted subcutaneously into 4-6-week-old athymic female nude mice purchased from Wuhan Hamilton Biological Polytron Technologies Inc. The tumor growth was measured twice weekly for 3 weeks by using a digital caliper. Tumor weight was measured when mice were sacrificed on day 27 after cells implantation. Statistical analysis. Correlation analyses were performed using SPSS 20 statistical software (SPSS Inc., Chicago, IL, USA). The Student's t-test was applied to evaluate the differences in groups as appropriate and the significance level was set at 0.05. The correlation analysis was evaluated by a χ 2 -test. A two-tailed p value < 0.05 was considered statistically significant.