Rac1 activation in oral squamous cell carcinoma as a predictive factor associated with lymph node metastasis

Secondary lymph node metastasis (SLNM) indicates a poor prognosis, and limiting it can improve the survival rate in early-stage tongue squamous cell carcinoma (TSCC). Many factors have been identified as predictors of SLNM; however, there is no unified view. Ras-related C3 botulinum toxin substrate 1 (Rac1) was found to be a promoter of the epithelial–mesenchymal transition (EMT) and is also attracting attention as a new therapeutic target. This study aims to investigate the role of Rac1 in metastasis and its relationship with pathological findings in early-stage TSCC. Rac1 expression levels of 69 cases of stage I/II TSCC specimens and their association with clinicopathological characteristics were evaluated by immunohistochemical staining. The role of Rac1 in oral squamous cell carcinoma (OSCC) was examined after Rac1 in OSCC cell lines was silenced in vitro. High Rac1 expression was significantly associated with the depth of invasion (DOI), tumor budding (TB), vascular invasion, and SLNM (p < 0.05). Univariate analyses revealed that Rac1 expression, DOI, and TB were factors significantly associated with SLNM (p < 0.05). Moreover, our multivariate analysis suggested that Rac1 expression was the only independent determinant of SLNM. An in vitro study revealed that Rac1 downregulation tended to decrease cell migration and proliferation. Rac1 was suggested to be an important factor in the metastasis of OSCC, and it could be useful as a predictor of SLNM.


Introduction
Tongue squamous cell carcinoma (TSCC) is the most commonly diagnosed OSCC [1][2][3][4]. The main treatment for the primary lesion of stage I/ II (T1-2N0) TSCC is surgery, which has favorable outcomes; however, secondary lymph node metastasis (SLNM) is observed in 22-32% of cases [5,6]. It is known that the survival rates have been poor, even when neck dissection is performed after a primary operation [7]. A previous study showed that SLNM patients had significantly greater numbers of positive nodes, a higher incidence of extracapsular spread, and a lower survival rate than those who have undergone elective neck dissection (END) [8]. Many histopathological factors such as the depth of invasion (DOI), vascular invasion, and tumor budding (TB) have been reported to be associated with SLNM in clinical N0 early-stage TSCC [9][10][11]. We can expect to progress to the treatment plan for cases with a poor prognosis if we succeed to clarify the relationship between these factors and the metastatic potential of molecular biology in cancer cells.
The epithelial-mesenchymal transition (EMT) plays a critical role in promoting tumor invasion and metastasis in OSCC. During EMT, the epithelial marker called E-cadherin is downregulated, while the mesenchymal marker called N-cadherin is upregulated. EMT is associated with poor clinical outcomes in many cancers [12,13].
Ras-related C3 botulinum toxin substrate 1 (Rac1), which was found to be a promoter of EMT, is also attracting attention as a new therapeutic target [14,15]. Rac1 is a small GTPase member of the rho family that is involved in the regulation of various cellular processes, including cytoskeletal remodeling, cell adhesion, and transcriptional regulation [16][17][18]. The role of Rac1 in cancer is the regulation of cell polarity, proliferation, angiogenesis, and especially migration [16,19,20]. Rac1 overexpression has been detected in various types of cancer, such as head and neck cancer [21], esophageal cancer [22], breast cancer [23], pancreatic cancer [24], colorectal cancer [25], and stomach cancer [26]. Recently, it has been shown to be a metastasis-inducing gene [27] and has attracted attention as a target for molecular target drugs [19,28]. On the other hand, in OSCC, although the expression of Rac1 is higher than that in normal areas and its association with recurrence has been reported [29], to our knowledge, there are no sufficient reports of its association with SLNM.
This study aims to evaluate the correlation of Rac1 expression with clinicopathological features of OSCC.

Patient selection
We evaluated Rac1 expression and clinicopathological records in stage I and stage II of 69 TSCC patients at the Oral Cancer Center of Tokyo Dental College, Japan, between 2015 and 2019 (Fig. 1). The patients underwent glossectomy without neck dissection were observed for at least 1 year. Secondary cases, recurrent cases, and those with SCC in situ, patients with SCC of questionable invasion by postoperative histopathology were excluded. Clinical indicators such as sex, age, and T-factor were extracted from medical records. Tissue specimens were evaluated for histopathological endpoints such as the tumor size, SLNM, histological grade, infiltrative growth pattern (INF), DOI, TB, lymphatic infiltration, vascular invasion, perineural invasion, and worst pattern of invasion-5 (WPOI-5) as per the 8th edition of American Joint Committee on Cancer Staging Manual [30]. TB was evaluated in a 20 × 10 field of the deepest point of tumor invasion and the number of TB was counted after selecting one field of view with the largest TB. The study protocol conformed to the principles outlined in the Declaration of Helsinki and was approved by the Ethics Review Board of Tokyo Dental College Ichikawa General Hospital (Approval Number: I 20-06).

Immunohistochemistry
Resected TSCC samples were fixed with 10% formalin, embedded in paraffin, and cut into sections that are approximately 4 μm thick. Each section was dewaxed in dimethylbenzene, after which it was dehydrated in serial alcohols and washed in phosphate-buffered saline (PBS) three times (for 3 min each). The sections were put in 10 mM sodium citrate buffer (pH 8) and heated for 20 min in the microwave oven, after which they were cooled down to room temperature for 30 min. The slides were incubated with 3% hydrogen peroxide in PBS for 15 min to block endogenous peroxidase activity. Afterward, the sections were washed with PBS. The tissues were subsequently incubated for 30 min at 4 °C with mouse anti-human Rac1 (1:800; ab155938, Abcam Biotechnology, Cambridge, UK). After the slides were washed with PBS, appropriate biotinylated secondary antibodies with Nichirei Histofine SAB-PO (MULTI) kit for 10 min, exposed to streptavidin-HRP Fig. 1 Patient selection. We included cases of only partial glossectomy that were observed for over 2 years and excluded cases of questionable invasion, SCC in situ, recurrence, and secondary cases label for 30 min, at room temperature. A color reaction was induced by a 30-s incubation in diaminobenzidine (DAB) solution and counterstaining with hematoxylin. For the negative control, we used the PBS buffer instead of the primary antibody. The immunoreactivity scores (IRS) of Rac1 were determined by two observers who were blinded to the relevant clinicopathological information. The scores of them were compared, and if the scores differed, both pathologists re-evaluated the staining to achieve a consensus score. Rac-1 immunoreactivity was assessed in terms of the following: (a) proportion of positive tumor cells in the tumor tissue (0 = 0, 1 = 1-10, 2 = 11-50, 3 = 51-70, and 4 = 71-100%); and (b) signal intensity (0 = no signal, 1 = weak, 2 = moderate, 3 = strong). The IRS (a range of 0-12) was determined by multiplying the proportion score by the intensity score. The average IRS for each case was assigned as the staining result for the patient. The specimens were rescored if the difference between the scores determined by the two pathologists exceeded three points. The final score was stratified into four categories: absent: score 0, weak; score 1-4, moderate; score 5-8, strong; score 9-12. Then, absent and weak were considered low expression, while moderate and strong were considered high expression.

siRNA transfection
To investigate the role of Rac1 in OSCC lines, Rac1 was stably knocked down and named Ca9-22-Si, HSC3-M3-Si, and SAS-Si. The control cells were named Ca9-22-NC, HSC3-M3-NC, and SAS-NC. Rac1-specific siRNA (ID: s11712) and control-siRNA (Cats: 4,390,843) were purchased from Thermo Fisher Scientific. The cells were transfected with 10 μl siRNA using Lipofectamine RNAiMax (Thermo Fisher Scientific) as per the manufacturer's instructions. The expression level of Rac1 was confirmed by qRT-PCR and the protein level was examined via Western blot.

Western Blot analysis
The cells were lysed in an ice bath using a lysis buffer (50 mM Tris-Hcl + 150 mM ph7.4 Nacl + 1%Nonidet P-40). Protein concentrations were determined using a BCA protein assay kit (Thermo Scientific Cat# 23,225) with bovine serum albumin as the standard. The loading sample was prepared via mixing with the 4 × Sample buffer (Thermo Scientific Cat#NP0007) and the 10 × Sample Reducing agent (Thermo Scientific Cat#NP0009). Proteins (20 µg) were resolved on 16.5% polyacrylamide gels (p-PAGEL Cat# 2,332,260, ATTO, Tokyo, Japan) and transferred onto PVDF membranes (Q-Blot Kit ATTOCat#2,322,443). The membranes were blocked with 5% skim milk in TBST at room

Wound-healing assay
The cells were seeded in 60-mm dishes and allowed to grow to 100% confluence. Afterward, a scratch was made across the cell monolayer. After washing with PBS, some fresh growth medium was added. At that point and after 12 h of incubation, the closed area was photographed using a phase contrast microscope (× 100). The closed area of each time point was calculated using the Fiji/ImageJ software ver, 2.30 (NIH, USA).

Cell proliferation assay
The WST-1 assay was performed using the Premix WST-1 Cell Proliferation Assay System (Takara). OSCC cells were seeded into a 96-well plate. After 48 h, the WST-1 reagent (10 μl) was added to each well. After incubation for 4 h, the absorbance at 450 nm was measured using a microtiter plate reader (Bio-Rad Laboratories, Hercules, California, USA).

Statistics
Statistical analyses were performed using EZR (version 2.3.0, R Foundation for Statistical Computing). A receiver operating characteristics (ROC) curve was generated to determine the cutoff value for IRS, TB, and DOI. Relationships between Rac1 immunohistochemical staining and clinicopathological variables were analyzed using Fisher's exact test. The five-year survival probability and disease-free survival (DFS) were analyzed using the log-rank test via the Kaplan-Meier method. A Cox regression model was constructed to perform univariate analyses and multivariate analyses of factors associated with SLNM. All statistical tests were two sided, and P < 0.05 was considered statistically significant.

Clinical and pathological analysis
Rac1 was localized in the cytoplasm and the cytomembrane of TSCC (Fig. 2). The average IRS of Rac1 was 2.1 ± 2.0 in SLNM-negative, and 6.2 ± 2.9 in SLNM-positive, and the Rac1 expression level was significantly higher the SLNMpositive IRS than the SLNM-negative IRS (p < 0.01) (data not shown). From the ROC curve, each cutoff value was calculated: IRS-4, DOI-2.5 mm, and TB-6. The association between the clinicopathological factors and the Rac1 expression in the TSCC is shown in Table 1. There were significant correlations between Rac1 and SLNM (p = 0.002), DOI (p = 0.001), TB (p = 0.003), and vascular invasion (p = 0.013). Kaplan-Meier analyses revealed that there were significant differences between the low Rac1 group (6.8%) and the high Rac1 group (53.3%) with the 5-year DFS (P < 0.01) (Fig. 3A). However, Rac1 overexpression was not significantly associated with 5-year overall survival (OS) rates. Also, no significant correlation was shown between SLNM and 5-year OS rates (Fig. 3B, C).

mRNA and protein expression analysis by knockdown
To investigate the role of Rac1 in oral cancer cell lines, we established Ca-9-22-Si, HSC3-M3-Si, and SAS-Si with Rac1 silencing by siRNA. The mRNA levels of Rac1 were significantly reduced in Ca9-22-Si, HSC-M3-Si, and SAS-Si than in controls (p < 0.05) (Fig. 4A). and protein levels showed a trend (Fig. 4B). Thus, these cells are used in wound-healing and proliferation assays to elucidate the effects of Rac1 silencing on OSCC.

The functions of Rac1 in OSCC
Compared control cells, Ca9-22-Si and HSC-M3-Si showed significantly slower wound closure (p < 0.05), and in SAS-Si, they tended to be slow to close but did not differ significantly (p = 0.19) (Fig. 5A) by wound-healing assays. In proliferation assays, there were no significant differences between Rac1-silenced cells and controls (Fig. 5B).
They reported that wound closure was significantly slower with Rac1 silencing, which was similar to the results of our study. Rac1 is known to be required to promote actin polymerization at the cell periphery and lamellipodium extension by targeting WAVE and PAK during migration [20,34]. In various types of cancer, Rac1 is considered important for the regulation of cancer cell migration [14,15,26,35,36].
In addition, the present study showed similar results for the OSCC cell line. The existing literature suggests that Rac1 silencing reduces cell proliferation [15,26,36]. However, in this study, there was no significant decrease. Rac1 regulates cell proliferation through many different pathways, including PAK [33], mitogen-activated protein kinase (MAPK) [37], and NF-κB [38]. Adam et al. reported there were different results knockdown and inhibition of Rac1 on cell proliferation in non-small cell lung carcinoma [35]. When using Rac1 siRNA, NF-κB activity was suppressed in the absence of tumor necrosis factor-α (TNF-α), whereas in the presence of TNF-α, NF-κB showed slight activation. However, when using Rac1 inhibiter (NSC23766), NF-κB activity remained suppressed even in the presence of TNF-α. There are other pathways except Rac1 that contribute to cancer cell proliferation include the PI3K/Akt/mTOR pathway [39], the Ras/Raf/MEK pathway [40], and the Wnt/βCatenin pathway [41]. Since Rac1 suppresses cell proliferation in a signal-independent manner, we suggest that the association with other factors not only Rac1 could be important. The association of Rac1 with various pathologic factors and a poor prognosis has been studied in many types of cancer. Many reports have discussed and indicated a relationship between Rac1 expression and lymph node metastasis [14,15,22,26,42,43]. In this study, Rac1 overexpression was observed in the SLNM-positive group, and there were significant correlations between Rac1 and vascular invasion, TB, and DOI. Kamai et al. suggested that patients who have vascular invasion with higher Rac1 expression need to consider chemotherapy to prevent recurrence, distant metastasis, or postoperative lymph node metastasis in uroepithelial carcinoma [42]. In this study, six of eight cases with vascular invasion positivity and high Rac1 expression were SLNM-positive (data not shown), and it may be important to have careful follow-up in such cases. In rectal cancer, TB is defined as an invading cluster of up to five tumor cells at the invasion front, and it is associated with SLNM for a poor prognosis [44]. There was no reported association between Rac1 and TB; however, a previous study recommended considering that TB reflects the early stages of EMT in tumor cell metastasis [45]. TB may serve as an indicator of Rac1induced EMT. Although there are many unclear points about the effects of Rac1 overexpression on TB and vascular invasion, Rac1 might mediate these factors to facilitate SLNM.
Many studies have examined the predictors of lymph node metastasis. DOI is a recognized factor used for predicting SLNM in OSCC cases. The National Comprehensive Cancer Network (NCCN) guidelines recommend END in patients with oral cancer with a DOI of ≤ 4 mm [46]. The present study hints at a significant association between Rac1 and DOI. Previous studies also reported similar findings in esophageal cancer and gallbladder cancer [22,43]. While DOI is considered an indication for END, it has been reported that SLNM occurs even if the DOI is less than 4 mm [47]. Interestingly, of the seven cases with a DOI of < 4 mm and SLNM, four were cases of Rac1 overexpression (data not shown). This finding may suggest that even in cases with a shallow DOI, high Rac1 expression should be followed up in early-stage TSCC. Moreover, in our univariate analysis, the risk of SLNM was increased with a DOI of over 2.5 mm, vascular invasion positive, and high Rac1 expression. Finally, in our multivariate analysis, the Rac1 expression level was a significant risk factor. In this study, no effect of Rac1 expression level or the presence or absence of SLNM on survival was demonstrated. This may be due to the fact that the study was limited to early-stage tongue cancer. However, considering the significant association between Rac1 expression and DFS, Rac1 may be an important factor in postoperative follow-up.
The main limitation of the present study is that we have only been able to relate Rac1 mainly to tumor cell migration and proliferation. The mechanism by which Rac1 promotes metastasis should be better defined by examining the relationship between Rac1 and other capabilities such as cell invasion, adhesion, and apoptosis.
In the future, we intend to use a Rac1 inhibitor (NSC23766) to clarify the role of Rac1 and to elucidate the tumor-promoting functions of Rac1, including its interrelationships with neighboring genes and EMT. Moreover, by accumulating more cases, we will be able to better examine the correlation of Rac1 with local recurrence and distant metastasis. If the metastatic potential of Rac1 can be evaluated in the future, it could serve as a predictive marker of SLNM and may aid in the therapeutic management of earlystage TSCC.

Conclusion
In conclusion, we strongly suggested that Rac1 is an important determinant of OSCC metastasis. Rac1 as a prognosis factor might be a useful tool and a new therapeutic target.