Low TINAGL1 expression is a marker for poor prognosis in breast cancer

Tubulointerstitial nephritis antigen-like 1 (TINAGL1) was reported to suppress tumor metastasis and growth in triple-negative (TN) breast cancer. We aimed to determine the associations of TINAGL1 expression with clinicopathological factors and prognosis in breast cancer patients with long-term follow-up. A total of 599 consecutive primary invasive breast cancer patients with available tissue specimens from surgery in our hospital were included in the study. TINAGL1 mRNA expression was examined in all 599 tissue specimens using a TaqMan real-time PCR system. TINAGL1 protein expression was further examined in 299 patients with available tissue specimens for immunohistochemical staining. Survival analyses were performed using the Kaplan–Meier method and Cox proportional hazards models. The median follow-up period was 12.0 years. In the total patients, low TINAGL1 mRNA expression was associated with significantly shorter disease-free survival (DFS) and overall survival than high expression (P = 0.003 and P = 0.01, respectively). Furthermore, hormone receptor-positive/human epidermal growth factor receptor 2-negative breast cancer patients with low TINAGL1 mRNA expression had a worse prognosis. Multivariate analysis identified low TINAGL1 mRNA expression, combined with lymph node positivity, as an independent poor prognostic factor for DFS in invasive breast cancer patients (HR 1.41; 95% CI 1.02–1.96; P = 0.036). TINAGL1 mRNA expression also varied with menopausal status, with low TINAGL1 mRNA expression being positively associated with poor prognosis in premenopausal patients, but not in postmenopausal patients. Our findings demonstrate that TINAGL1 may be a promising candidate biomarker and therapeutic target in breast cancer patients.


Introduction
Breast cancer is the most frequently encountered malignancy in women worldwide. Although breast cancer is curable in patients with early-stage and non-metastatic disease, advanced breast cancer with distant organ metastases is considered incurable by currently available therapies.
Tubulointerstitial nephritis antigen-like 1 (TINAGL1) was identified as an extracellular matrix (ECM) protein that promoted integrin-mediated cell adhesion (Li et al. 2007). Detection of TINAGL1 expression has been reported in vascular smooth cells, skeletal muscle, aorta, heart, placenta, and kidney (Kaltezioti et al. 2021). However, the specific function of TINAGL1 remains to be elucidated. Recent 1 3 studies identified TINAGL1 expression as a prognostic factor in hepatocellular carcinoma (Sun et al. 2019), gastric cancer (Shan et al. 2021), and breast cancer (Shen et al. 2019). TINAGL1 was suggested to inhibit tumor growth and metastasis in breast cancer (Shen et al. 2019;Musetti and Huang 2021;Korpal et al. 2011), and TINAGL1 expression was dependent on Sec23 homolog A (SEC23A), an essential component of coat protein complex II vesicles that is indispensable for ECM protein secretion (Korpal et al. 2011). SEC23A was also reported to regulate the translocation of secretory and ECM proteins (Korpal et al. 2011;Zeng et al. 2021). Shen et al. (2019) reported that low Tinagl1 expression was associated with poor prognosis in patients with triplenegative (TN) breast cancer, defined by lack of estrogen receptor (ER), progesterone receptor (PgR), and human epidermal growth factor receptor 2 (HER2) expression. They showed that TINAGL1 inhibited the cell signaling pathways for epidermal growth factor receptor (EGFR) and focal adhesion kinase by blocking EGFR dimerization and binding of fibronectin to integrins. However, the long-term prognostic impact of TINAGL1 expression remained unknown because the follow-up period in their study was short at < 2 years. Furthermore, the relationships between TINAGL1 expression and breast cancers other than TN breast cancer also remained unknown. We hypothesized that other breast cancers would also show an association of TINAGL1 expression with prognosis, similar to the case for TN breast cancer.
We designed the present study to determine the associations of TINAGL1 mRNA and protein expression with clinicopathological factors and prognosis in primary invasive breast cancer patients with long-term follow-up of > 10 years. We also investigated the association between TINAGL1 and SEC23A expression in breast cancer patients.

Patients and samples
A total of 599 consecutive primary invasive breast cancer patients with available tissue specimens collected during surgery between 1992 and 2008 at the Department of Breast Surgery, Nagoya City University Hospital, Japan, were included in the study. All 599 tissue specimens were measured for TINAGL1 mRNA expression. Of these patients, 299 patients from 2000 to 2008 had invasive breast cancer tissue specimens available as tissue microarrays (TMAs), and these specimens were used to evaluate TINAGL1 protein expression. The tissue specimens obtained during surgery were snap-frozen in liquid nitrogen immediately after resection and stored at − 80 °C until RNA extraction or fixed in 10% buffered formalin and embedded in paraffin. Histological grade was estimated using the Bloom and Richardson method, as proposed by Elston and Ellis (1991). Adjuvant therapy was conducted at the physician's discretion according to each patient's clinicopathological features. Patients who underwent neoadjuvant chemotherapy were excluded from the study to eliminate any biological effects of neoadjuvant chemotherapy on the tumor tissues. Patients who underwent adjuvant trastuzumab therapy were also excluded to avoid mixing patients who did and did not receive adjuvant trastuzumab treatment for HER2-positive breast cancer. Disease-free survival (DFS) was defined as the interval from the date of primary surgery to the earliest occurrence of one of the following: loco-regional recurrence, distant metastasis, or death from any cause. Overall survival (OS) was defined as the interval from the date of primary surgery to death from any cause. Written informed consent forms for comprehensive research use were obtained from all patients before surgery. The study protocol was approved by the Institutional Review Board of Nagoya City University Graduate School of Medical Sciences (approval number: 70-00-0166) and conformed to the guidelines of the Declaration of Helsinki.

RNA extraction and reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
Total RNA was extracted from the frozen breast cancer tissue specimens using an RNeasy Mini Kit (Qiagen, Tokyo, Japan) according to the manufacturer's protocol. The quantity of total RNA was measured using a DS-11 Spectrophotometer (DeNovix, Wilmington, DE, USA). cDNA was reverse-transcribed from the total RNA samples using a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's protocol. TaqMan Gene Expression assays (Thermo Fisher Scientific) were used to measure the mRNA expression levels of TINAGL1, SEC23A, and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Duplex RT-qPCR assays were performed using a TaqMan StepOnePlus Real-time PCR System (Thermo Fisher Scientific). The reactions was analyzed using a FAMlabeled probe for TINAGL1 or SEC23A (Thermo Fisher Scientific) and a VIC-labeled probe for GAPDH (Thermo Fisher Scientific) as a single assay for each sample. The amplification reaction mixture comprised 2 μL of cDNA, 10 μL of TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific), 1 μL of TaqMan FAM-labeled probe, 1 µL of TaqMan VIC-labeled probe, and 6 μL of DEPC-treated water (Thermo Fisher Scientific) in a total volume of 20 μL. The RT-qPCR conditions were 95 °C for 20 s, followed by 40 cycles of denaturation at 95 °C for 1 s and annealing and extension at 60 °C for 20 s. The results were converted into relative concentrations using a standard curve. The mRNA expression levels of TINAGL1 or SEC23A were normalized by the expression level of the housekeeping gene GAPDH as an internal control. We determined the cut-off levels for TINAGL1 and SEC23A mRNA expression as the median expression.

Immunohistochemistry (IHC) of ERα, PgR, HER2, TINAGL1, and SEC23A
A single 4-μm-thick section from each paraffin-embedded specimen was initially stained with hematoxylin and eosin to verify that invasive carcinoma cells were present in sufficient numbers and that the quality of fixation was adequate for IHC analysis. Subsequently, 4 μm-thick serial sections were prepared from suitable tissue blocks and float-mounted on adhesive-coated glass slides for immunostaining of ERα (Dako Envision FLEX-ER; EP1; Agilent Technologies, Santa Clara, CA, USA), PgR (Dako Envision FLEX-ER; PgR636; Agilent Technologies), and HER2 (HercepTest II; Agilent Technologies) using an Autostainer Link 48 (Agilent Technologies). The immuno-stained specimens were scored after the entire sections had been evaluated by light microscopy (BX51; OLYMPUS, Tokyo, Japan).
Expression of ERα or PgR was evaluated by the proportion and intensity of positively stained tumor cells using Allred's procedure, with positivity defined as ≥ 1% of tumor cells exhibiting staining in the nucleus (Harvey et al. 1999). ERα-positive and/or PgR-positive specimens were considered hormone receptor-positive. HER2 expression was evaluated by the membrane staining pattern and scored on a scale of 0 to 3 + (Hicks and Tubbs 2005). Tumors with a score of 3 + were considered HER2-positive. Tumors with a score of 2 + were tested for gene amplification by fluorescence in situ hybridization (FISH) using the Path-Vysion assay (Abbott Laboratories, Abbott Park, IL, USA) according to the manufacturer's protocol. A ratio of > 2.0 for HER2 gene/chromosome 17 was considered HER2-positive. Scores of 0, 1 + , or 2 + /FISH-negative were considered HER2-negative.
For IHC analysis of TINAGL1, TMAs of 2-mm diameter were created on slides after confirming that a sufficient number of invasive carcinoma cells were present and that the quality of fixation was suitable for IHC analysis. The primary antibody was a rabbit polyclonal anti-TINAGL1 antibody (Atlas Antibodies, Stockholm, Sweden) used at 1:30 dilution. Immunostaining was performed using a Leica Bond-Max Automated System and a Leica Refine Detection Kit (Leica Biosystems, Wetzlar, Germany). The TMA slides were scanned at 20 × using Aperio Scanscope CS2 (Leica Biosystems). To ensure objectivity, the TINAGL1 protein expression level in the cytoplasm was evaluated by quantifying the staining intensity according to the H-score using a digital pathology system (eSlide Manager Application; Leica Biosystems). The H-score produced by this digital image analysis tool was previously shown to overcome the limitations of visual quantitative scoring by enabling objective and highly reproducible quantification of biomarkers (Bankhead et al. 2018;Ram et al. 2021). The H-score was calculated as follows: classification of cytoplasmic staining intensity into three categories (1, weak; 2, moderate; 3, strong), followed by multiplication of the percentage of stained cells by each category number, and finally addition of the values. In short, the H-score was calculated using the following formula: 1 × (% cells in weak category) + 2 × (% cells in moderate category) + 3 × (% cells in strong category). For survival analysis based on TINAGL1 protein expression, the median H-score was set as the cut-off level. The Atlas Antibodies assay number for Tinagl1 was HPA048695.
For the IHC analysis of SEC23A, the primary antibody was a rabbit polyclonal anti-SEC23A antibody (Abcam, Cambridge, UK) used at 1:400 dilution. Immunostaining was performed using the same methods as TINAGL1 by a Leica Bond-Max Automated System and a Leica Refine Detection Kit (Leica Biosystems). TMAs for SEC23A were also scanned at 20 × by Aperio Scanscope CS2 (Leica Biosystems). SEC23A protein expression levels in the cytoplasm were evaluated according to the H-score using eSlide Manager Application (Leica Biosystems), and the median H-score was set as the cut-off level. The Abcam assay number for SEC23A was ab137583.

Statistical analysis
The associations of TINAGL1 mRNA expression level with clinicopathological factors were assessed by Student's t test and Fisher's exact probability test. The Mann-Whitney U test was used to compare differences in TINAGL1 or SEC23A mRNA expression levels between two groups, and comparisons between three groups were adjusted by the Bonferroni method. Spearman's rank correlation coefficient was used to evaluate correlations between two factors, such as TINAGL1 mRNA and TINAGL1 protein expression levels. Survival analyses were performed using the Kaplan-Meier method and verified by the log-rank test. DFS was censored on the last follow-up date if patients were relapse-free and alive, and OS was censored on the date when patients were last confirmed alive. Cox proportional hazards regression models were utilized for univariate and multivariate analyses to identify prognostic factors. Values of P < 0.05 were considered statistically significant. All statistical analyses were performed with R version 4.0.2 (R Foundation for Statistical Computing, Vienna, Austria).

TINAGL1 mRNA expression and prognosis of breast cancer patients
First, we investigated the associations between TINAGL1 mRNA expression and clinicopathological parameters in the breast cancer patients. All 599 consecutive invasive breast cancer tissue specimens were subjected to TINAGL1 mRNA expression analysis. The median (range) follow-up period was 12.0 (0.02-26.9) years. The characteristics of the patients are shown in Table 1 and Fig. 1. Low TINAGL1 mRNA expression was positively associated with higher tumor grade, PgR-negativity, and HER2-positivity.
Next, we analyzed the association between TINAGL1 mRNA expression level and prognosis. As shown in Fig. 2, patients with low TINAGL1 mRNA expression had significantly shorter DFS (hazard ratio [HR] 1.59; 95% confidence interval 95% CI 1.17-2.16; P = 0.003) and shorter OS (HR 1.61; 95% CI 1.11-2.34; P = 0.01) in the total patients. We further investigated the relationships between TINAGL1 mRNA expression level and prognosis in patients with different subtypes of breast cancer classified by hormone receptor and HER2 status. Patients with low TINAGL1 mRNA expression had significantly shorter DFS and OS among hormone receptor-positive/HER2-negative breast cancer patients (Fig. 3a, b), while no association was observed between TINAGL1 mRNA expression and prognosis in HER2-positive or TN breast cancer patients (Fig. 3c-f). Among lymph node-positive patients, patients with low TIN-AGL1 mRNA expression had shorter DFS (HR 1.6; 95% CI 1.08-2.38; P = 0.02) and OS (HR 1.61; 95% CI 1.01-2.56; P = 0.04) than patients with high expression, while lymph node-negative patients showed no association between TIN-AGL1 mRNA expression and prognosis (Fig. 4). Because TINAGL1 mRNA expression level was associated with menopausal status (Table 1), we investigated the association between TINAGL1 mRNA expression and prognosis in 241 premenopausal patients and 353 postmenopausal patients. As shown in Fig. 5, patients with low TINAGL1 mRNA expression had a significantly worse prognosis among premenopausal patients, while no significant difference was observed in postmenopausal patients.
Univariate and multivariate Cox regression analyses were conducted on the associations of clinicopathological factors with prognosis in the total patients ( Table 2). The multivariate analysis identified low TINAGL1 mRNA expression and lymph node positivity as independent poor prognostic factors for DFS in the total patients (HR 1.41; 95% CI 1.02-1.96; P = 0.036).

TINAGL1 protein expression in breast cancer patients
Among the 599 invasive breast cancer patients analyzed in the study, 299 patients had available TMA specimens for evaluation of TINAGL1 protein expression by IHC. Representative images of TINAGL1 protein expression are shown in Fig. 6a. TINAGL1 protein expression was localized in the cytoplasm of cancer cells. The median (range) followup period was 11.9 (0.07-20.3) years. We first analyzed the protein expression by the H-score. The characteristics of the patients evaluated are shown in Supplementary Table 1. No association was found between TINAGL1 protein expression and TINAGL1 mRNA expression (Fig. 6b, c). The survival analysis also showed no association between TINAGL1 protein expression and prognosis (Fig. 6d, e). The results for the survival analysis based on the SI method reported by Shen et al. (2019) are shown in Supplementary Fig. 1. No association between TINAGL1 protein expression and prognosis was found using the SI method, similar to the findings for the H-score.

Association between TINAGL1 and SEC23A expression levels
TINAGL1 was reported to be a SEC23A-dependent metastasis suppressor (Korpal et al. 2011). The total 599 breast cancer tissue specimens were also analyzed for SEC23A mRNA expression. The correlation between TINAGL1 and SEC23A mRNA expression levels was evaluated by Spearman's rank correlation test. A positive association was found between TINAGL1 and SEC23A mRNA expression levels (r = 0.35; P < 0.001; Fig. 7). We further investigated the association between SEC23A mRNA expression levels and clinicopathological parameters. Low SEC23A mRNA expression was positively associated with higher tumor grade, ERαnegativity, PgR-negativity, and HER2-positivity (Supplementary Fig. 2a). Although the TINAGL1 mRNA expression level was associated with prognosis as described earlier, the SEC23A mRNA expression level was not ( Supplementary  Fig. 2b, c). No association was also shown between SEC23A mRNA expression and prognosis according to breast cancer subtype. We next evaluated the correlation between TIN-AGL1 and SEC23A protein expression levels, and identified  a positive association between them (r = 0.36; P < 0.001; Supplementary Fig. 3).

Discussion
In the present study, we investigated the associations of TINAGL1 mRNA and TINAGL1 protein expression levels with clinicopathological factors and prognosis in breast cancer patients with long-term follow-up. Our findings demonstrated that low TINAGL1 mRNA expression was significantly associated with poor prognosis in primary invasive breast cancer patients. Furthermore, low TINAGL1 mRNA expression was also associated with poor prognosis in hormone receptor-positive/HER2-negative patients and premenopausal patients. Our analyses further identified low TINAGL1 mRNA expression as an independent poor prognostic factor in the total invasive breast cancer patients evaluated in the study. In this study, low TINAGL1 mRNA expression was significantly associated with poor prognosis in hormone receptor-positive breast cancer patients, but not in TN breast cancer patients, while a previous study reported that low TINAGL1 mRNA expression was associated with poor prognosis only in TN breast cancer patients (Shen et al. 2019). This discrepancy between the present and previous studies may be due to the difference in follow-up periods. Specifically, the median follow-up period in the present study was > 10 years, while that in the previous study was only 1.7 years. Our findings also demonstrated that higher histological grade was positively associated with low TIN-AGL1 mRNA expression. These findings suggested that TINAGL1 expression levels were associated with highly malignant tumor status, consistent with a previous report (Shen et al. 2019).
Our data revealed that TINAGL1 mRNA expression varied depending on the menopausal status. Low TINAGL1 mRNA expression more frequently observed in postmenopausal patients compared with premenopausal patients. Interestingly, low TINAGL1 mRNA expression was significantly associated with poor prognosis in premenopausal patients, but not in postmenopausal patients. Although TINAGL1 was reported to be associated with ovulation in aged female mice in a mouse model (Akaiwa et al. 2020), there have been no reports of associations between female hormone levels and TINAGL1 expression or function to date. It is speculated that either low expression of TINAGL1 may promote tumor growth or high expression of TINAGL1 may inhibit tumor growth in breast cancer patients under situations involving high levels of female hormones, such as estrogen and/or progesterone in premenopausal patients.
We examined the association between TINAGL1 and SEC23A mRNA expression levels because Korpal et al. (2011) reported that SEC23A mediated secretion of TIN-AGL1. Our data demonstrated a positive correlation between these levels. However, no association was observed between SEC23A mRNA expression level and prognosis. These findings suggested that TINAGL1 may be clinically more important than SEC23A for the SEC23A-TINAGL1 axis in invasive breast cancer patients.
Although low TINAGL1 mRNA expression was positively associated with poor prognosis in breast cancer patients, no association was found between TINAGL1 mRNA and TINAGL1 protein expression. There were some possible reasons for these findings. First, TINAGL1 protein expression may be affected by the metabolic speed of proteolysis. If the in vivo half-life of TINAGL1 is short, it is reasonable that the results may be mismatched between the mRNA and protein expression levels (Greenbaum et al 2003). Second, the quality of the formalinfixed paraffin-embedded (FFPE) specimens may have deteriorated due to the long-term storage (Omilian et al.  Liu et al. 2016). This study has several limitations. It was a retrospective analysis at a single institute using archived materials. We could not completely exclude the possibility that the quality of the breast cancer tissue specimens may have affected the results because the specimens analyzed in the study were obtained more than 10 years ago. We used samples from macro-dissected surgical specimens that were cryopreserved immediately after resection. We did not confirm the amount of cancer tissues in the cryopreserved specimens, and thus the percentage of cancer cells was likely to vary. The study included 599 consecutive breast cancer tissue specimens collected between 1992 and 2008. During this period, adjuvant therapies for breast cancer progressed substantially, and we could not eliminate the effects of different adjuvant therapies in our study. Moreover, adjuvant anti-HER2 therapy was Regarding therapeutic use of TINAGL1, Shen et al. (2019) demonstrated that TINAGL1 protein therapy showed efficacy in preventing metastasis or inhibiting outgrowth of established metastases. Recently, Musetti and Huang (2021) reported a lipid nanoparticle-based gene therapy that directly targeted TINAGL1 expression in breast cancer cells as localized expression. They reported that the TINAGL1 gene therapy was able to slow breast cancer growth and increase tumor vasculature without increasing the tumor permeability or risk of metastasis. In the present study, low TINAGL1 mRNA expression was associated with significantly poor prognosis in lymph node-positive patients. Our data also suggest that lymph node-positive breast cancer patients with low TINAGL1 expression who have completed standard therapy may be candidates for TINAGLl1 gene therapy in future.

Conclusion
We demonstrated that low TINAGL1 mRNA expression was significantly associated with poor prognosis in primary invasive breast cancer patients with long-term follow-up, and identified low TINAGL1 mRNA expression as an independent poor prognostic factor in breast cancer patients. We also demonstrated that low TINAGL1 expression was associated with poor prognosis in hormone receptor-positive/HER2positive patients. Furthermore, TINAGL1 mRNA expression varied with menopausal status, and low TINAGL1 mRNA expression was only significantly associated with poor prognosis in premenopausal patients. Our findings suggest that TINAGL1 may play a role in the progression or proliferation of breast cancer and may be a promising candidate biomarker for prognosis and a therapeutic target in breast cancer patients.