The impact of Cysteine-Rich Intestinal Protein 1 (CRIP1) in human breast cancer

We found no or low CRIP1 expression in 79 tumors (37 were negative and 42 were classified as 1+), medium expression in 20 tumors (classified as 2+) and high CRIP1 expression in 14 tumors (classified as 3+). In breast cancer tissue, positive and negative staining of CRIP1 was frequently associated with HER2 staining (Figure 1). A significant correlation was found between CRIP1 and the expression of HER2 (p = 0.016), and an inverse correlation was found between CRIP1 expression and estrogen receptor (ER, p = 0.04). No significant association was identified between CRIP1 and lymph node status, tumor size, histological grade, or progesterone receptor expression.

In univariate analyses of the distant metastases-free survival of the patients, a significant positive correlation was found between CRIP1 expression (p = 0.039) and a more favorable prognosis for patients with positive CRIP1 expression (classified 1+ to 3+) (Figure 2A, Table 1). HER2 expression was not significantly associated with the clinical course of the disease (p = 0.8). In the tumor cohort analyzed, there was no significant association between CRIP1 expression and the lymph node status of the patients. However, when we considered only lymph node-positive tumors, a trend was observed between CRIP1 positivity and a better clinical course of the disease (p = 0.09) (Table 1).

The CRIP1 expression in our tumor cohort was associated with the expression of HER2 (p = 0.016). Considering the CRIP1 expression in only HER2-negative tumors (Table 1), no significant association was found with the clinical course. However, in HER2-positive tumors, two different prognostic groups could be identified according to the CRIP1 expression. CRIP1-positive tumors showed a better prognosis, with 39% of patients (23 out of 59) suffering distant metastases compared with 67% (14 of 21) of CRIP1-negative patients, within the follow-up period of more than 30 years (p = 0.021) (Figure 2B, Table 1). This result clearly indicates that CRIP1 expression may be a useful prognostic marker in HER2-positive tumors.

Remarkably, in multivariate Cox regression analysis, CRIP1 proved to be an independent prognostic factor, along with nodal status (pN) and tumor size (pT) (p = 0.039) (Table 1).

For functional in vitro analyses, appropriate breast cancer cell lines were identified that co-expressed both HER2 and CRIP1 at adequate levels. The adequate co-expression of both proteins was detected in the T47D, BT474 and MDA-MB-361 cell lines (out of seven analyzed breast cancer cell lines) (Figure 3A). In this study, we selected T47D and BT474 cells for CRIP1 knockdown and subsequent analyses because in these cells the protein expression levels of CRIP1 and HER2 were higher than in the MDA-MB-361 cells.

After the identification of small interfering RNAs (siRNAs) that showed specific and efficient CRIP1 downregulation (Figure 3B and Additional file 1), the effects of CRIP1 knockdown in the T47D and BT474 cells on the expression and phosphorylation of HER2 signaling-associated proteins were analyzed using immunoblotting. Following CRIP1 knockdown, no effects were observed for HER2 (human epidermal growth factor receptor 2), and HER2-related and proliferation-associated signaling proteins like MAPK (mitogen-activated protein kinase), STAT3 (signal transducer and activator of transcription 3), Akt, cdk2 (cyclin dependent kinase 2) or PTEN (phosphatase and tensin homologue deleted on chromosome ten) protein expression levels (Figure 4A). In contrast, the phosphorylation of MAPK at Thr202/Tyr204 was increased in both cell lines due to CRIP1 downregulation (Figure 4A). This mitogen-activated protein kinase is involved in cell proliferation, differentiation and growth [16]. Phosphorylated MAPK activates the downstream phosphorylation of its substrates in the cytoplasm, or it translocates to the nucleus and subsequently regulates gene expression through the phosphorylation of transcription factors. p38 MAPK regulates cell survival and apoptosis [16]. The phosphorylation of p38 MAPK at Thr180/Tyr182 leads to the activation of other MAPK and transcription factors that also regulate apoptosis. CRIP1 knockdown did not lead to an altered phosphorylation of p38 MAPK (data not shown). Akt activation leading to the regulation of survival and apoptosis is regulated by phosphorylation at Thr308 and Ser473 [17]. CRIP1 knockdown led to an increased phosphorylation of Akt at Thr308 (Figure 4A).

The phosphatase PTEN is a tumor suppressor that negatively regulates the PI3K/Akt pathway [18]. The phosphorylation of PTEN impairs its tumor suppressive function. CRIP1 silencing did not affect the phosphorylation of PTEN (data not shown).

STAT3 drives cell growth, survival, differentiation and gene expression via phosphorylation at Tyr705. The phosphorylation at Ser727 is associated with its role as a transcription factor [19]. After the siRNA-mediated downregulation of CRIP1, we did not observe an altered phosphorylation of STAT3 at Ser727, but the phosphorylation at Tyr705 was elevated in T47D cells (Figure 4A), in BT474 cells this phosphorylation site was not detectable. We further analyzed the expression and phosphorylation of cell cycle proteins in response to changes in CRIP1 expression. No altered expression was observed for cyclin E, cyclin D1, cyclin A proteins or the cyclin-dependent kinase 2 (data not shown). However, we observed a reduced phosphorylation of cdc2 at Tyr15 in both cell lines following CRIP1 silencing (Figure 4A).

In addition, we investigated the proliferation of T47D and BT474 cells following CRIP1 knockdown based on the enzymatic cleavage of tetrazolium salts into formazan (WST-1 proliferation assay). Compared with control cells (mock and cells transfected with siRNA targeting GAPDH) the proliferation was significantly elevated of approximately 40% when T47D cells were depleted of CRIP1 using the most efficient siRNA1 (Figure 4B). In BT474 cells, in both silencing approaches the proliferation index was elevated of over 40% or 60%, respectively (Figure 4B).

To further elucidate the functional role of CRIP1 in breast cancer, we analyzed the migration or invasion of transfected and control T47D and BT474 breast cancer cells. Due to a non-confluent cell formation, the BT474 cells are not suitable for a wound scratch assay. The migration of T47D cells was not affected by reduced CRIP1 protein levels (data not shown). In contrast, compared with control cells, the invasion of T47D cells was 2.7 fold higher after knockdown of the CRIP1 protein using the most efficient siRNA1 (Figure 5A). In addition, the invasion of BT474 cells was also elevated (approximately 2.3 fold higher) following CRIP1 knockdown. To further confirm this observation, we determined the activation of MMP9 (matrix metalloproteinase 9) with the immunoblotting of the supernatants of serum-starved cells. The activation of MMP9 (illustrated by the band at 84 kDa) was slightly increased following CRIP1 silencing in T47D cells (Figure 5B). In the BT474 cell line, the MMP9 protein was not detectable.

Ethical approval concerning the use of tumor tissues in this study was obtained from the Ethics Committee of the Medizinische Fakultät der Technischen Universität, Munich, Germany. All experimental research described here was performed on human tissue only and was in compliance with the Helsinki Declaration. Formalin-fixed and paraffin-embedded archival material was randomly collected from 113 patients with invasive ductal breast carcinomas. In total, 67 of the tumors were node-negative, and most of the tumors (n = 72) were less than 2 cm in size. According to the histological grade [37], most of the tumors were classified as grade 2 (n = 75), 9 as grade 1, and 29 as grade 3. In addition to the standard histopathological parameters (lymph node status, tumor size, histological type and grade), immunohistochemical data from the tumors were available for HER2 and estrogen and progesterone receptor. The median follow-up of patients was 134 months (max. 468 months), with 49 (44%) of the patients showing disease recurrence with distant metastases within the period of clinical follow-up.

Tissue microarrays (TMAs) were produced as previously described [38] using a tissue-arraying instrument (Beecher Instruments Inc., Silver Spring, MD, USA). Hematoxylin- and eosin-stained sections of the TMAs were examined, and the original paraffin blocks were re-examined to validate representative sampling.

Immunohistochemical staining was performed on 3 μm thick sections of the TMAs using an automated stainer (Discovery XT) and a DAB Map kit (both Ventana Medical Systems, Tucson, AZ, USA). The CRIP1 primary antibody (AbD Serotec, Oxford, UK) was diluted 1:100, and the staining intensities were scored by two independent observers using a 4-point scale as indicated: 0 (no staining) and from 1+ (light staining) to 3+ (strong staining).

The correlations between CRIP1, HER2, and the histopathological parameters were examined with Spearman's rank correlation test. For univariate survival analyses, Kaplan Meier curves were calculated, and the differences between strata were evaluated with the log-rank chi-squared test. A multivariate analysis was performed using Cox proportional hazards regression and a stepwise selection algorithm (SAS Institute, Cary, NC, USA). All of the parameters showing a significance level of p ≤ 0.15 in univariate analysis were analyzed with multivariate analysis. In all of the other tests, statistical significance was established if p ≤ 0.05.

The human T47D and MCF7 breast cancer cell lines were maintained in RPMI 1640 (Roswell Park Memorial Institute) medium. The human BT474, SKBR3, MDA-MB-231, MDA-MB-361 and JIMT breast cancer cell lines were maintained in DMEM (Dulbecco’s Modified Eagle Medium). The media were supplemented with 10% FBS, the antibiotics penicillin and streptomycin (0.5%), 10 μg/ml human insulin (for the T47D, MCF7 and JIMT-1 cells), and the cells were maintained at 37°C in 5% CO₂. To identify efficient and specific siRNAs for the knockdown of CRIP1, T47D and BT474 breast cancer cells were transiently transfected with four different siRNAs (Invitrogen, Carlsbad, CA, USA, and Santa Cruz Biotechnology, Heidelberg, DE) and positive and negative control siRNAs for 48 h and 72 h, as described previously [39]. Specific transfections were performed in three independent experiments.

For SDS-PAGE and Western blot analysis, T47D and BT474 breast cancer cells were treated as described previously [39]. The proteins were detected with primary antibodies targeting CRIP1 (AP4707b, Abgent, San Diego, CA, USA); HER2 (A0485, DAKO, Glostrup, DK); (phospho, 9554) PTEN (9559), (phospho, 4376) MAPK (4695), (phospho, 9211) p38 MAPK (9212), phospho-STAT3 (9131 and 9134), (phospho, 4056) Akt (9272), phospho-cdc2 (9111), and MMP9 (3852) (Cell Signaling Technology, Beverly, MA, USA); cdk2 (sc-6248) and GAPDH (sc-25778) (Santa Cruz Biotechnology, Heidelberg, DE); STAT3 (610190, BD Transduction Laboratories, Lexington, KY, USA); and actin (A5441) and tubulin (T5168, Sigma, St. Louis, MO, USA). Anti-rabbit (NA934) and anti-mouse (NA931) peroxidase-conjugated secondary antibodies were obtained from GE Healthcare (Chalfont St. Giles, Buckinghamshire, UK). All bands showing altered intensities after CRIP1 knockdown were quantified relative to the control bands using the Molecular Imager ChemiDoc™ XRS and the analysis software Quantity One® (Bio-Rad Laboratories, Hercules, CA, USA).

Cell proliferation was determined using water-soluble tetrazolium WST-1 (4-[3-(4-Iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzene disulfonate) for the spectrophotometric assay according to the manufacturer’s protocol (05015944001, Roche Diagnostics, Mannheim, DE). One day after transfection, T47D and BT474 cells were seeded at a concentration of 1 × 10⁴ cells per well in a 96-well tissue culture plate. After following 48 h, the WST-1 reagent was added and the cells were incubated for 0.5 h to 4 h at 37°C. The absorbance of the infected and the control cells was measured against a background control using a microplate ELISA reader (Bio-Rad, München, DE) at 450 nm (reference wavelength at 655 nm). Five independent experiments were performed.

A migration assay using transiently transfected and nontransfected T47D breast cancer cells was performed twice and quantified as described previously [39]. In brief, a confluent monolayer of T47D cells was scratched using a 1 mm pipette tip. The cells were washed, and serum-reduced medium was added at a concentration (0.1% FBS) that reduced proliferation but was sufficient to avoid apoptosis or cell detachment [40]. The cells were incubated at 37°C in 5% CO₂ and monitored.

Control and transfected T47D and BT474 breast cancer cells were seeded at a density of 5 × 10⁴ onto BD BioCoat Matrigel Invasion Chambers (BD, Bedford, MA, USA) in 24-well cell culture plates and incubated for 24 h (T47D cells) or 48 h (BT474 cells) at 37°C one day after transfection. For T47D cells, epidermal growth factor (EGF, 25 ng/ml in serum-reduced medium) was used as a chemoattractant in the lower chamber. For the BT474 cells in the lower chamber the complete medium was used and the invasion assays were performed according to the manufacturer’s instructions. After incubation, the non-invading cells were removed from the apical side of the membrane with a cotton swab. The invading cells were then fixed with methanol, stained with toluidine blue, and counted under a microscope. The assay was performed twice.