KIF4A Regulates the Progression of Pancreatic Ductal Adenocarcinoma through Proliferation and Invasion.

BACKGROUND: Pancreatic cancer is a malignant tumor of the digestive tract, which is difficult to diagnose and treat due to bad early diagnosis. We aimed to explore the role of kinesin superfamily 4A (KIF4A) in pancreatic ductal adenocarcinoma (PDAC).
METHODS: We first used the bioinformatic website to screen the data of pancreatic cancer in TCGA, and KIF4A protein was detected among the 86 specimens of patients in our hospital combined with clinic-pathological characteristics and survival analysis. KIF4A loss-expression cell lines were established by RNA interference (RNAi). In addition, we performed in vitro cell assays to detect the changes in cell proliferation, migration, and invasion. The proteins involved in the proliferation and metastasis of cancer cells were also detected by western blot. The above results could be proved in vivo. Further, the correlation between KIF4A and CDC5L was analyzed by TCGA and IHC data.
RESULTS: We first found a high expression of KIF4A in pancreatic cancer, suggesting a role of KIF4A in the development of pancreatic cancer. KIF4A was found to be differentially expressed (P < 0.05) among the 86 specimens of patients in our hospital and was significantly associated with PDAC TNM stages and tumor size. High KIF4A expression also significantly worsened overall survival (OS) and disease-free survival rate (DFS) (P < 0.05, respectively). In addition, cell proliferation, migration, and invasion were inhibited by the KIF4A-shRNA group compared with the control (P < 0.05, respectively). In the end, knockdown of KIF4A could inhibit tumor development and metastasis in vivo. Further, the positive correlation between KIF4A and CDC5L existed, and KIF4A might promote pancreatic cancer proliferation by affecting CDC5L expression.
CONCLUSION: In conclusion, the high expression level of KIF4A in PDAC was closely related to poor clinical and pathological status, lymphatic metastasis, and vascular invasion. KIF4A might be involved in promoting the development of PDAC in vitro and in vivo, which might be a new therapeutic target of PDAC.

1. Background

Pancreatic cancer is quite malignant and about 90% of the patients are pancreatic ductal adenocarcinomas (PDAC) that originate from the ductal epithelium. There will be 57600 new cases (male 30400 and female 27200 cases) of pancreatic cancer and 47050 (male 24640 and female 22410 cases) of them will die in 2020 [1]. The 5-year survival rate is very low, and its high surgical mortality and low cure rate have made it one of the worst prognosis of malignant tumors [2]. Therefore, for patients with PDAC, effective intervention and treatment are of great importance. Unfortunately, the therapy target after the surgery for the treatment of PDAC has not been well developed and potential mechanisms have not yet been fully explored, and there are still many areas of PDAC that are worth our in-depth discussion.

Kinesin superfamily 4A (KIF4A) had recently been found to play a role in a variety of tumors including lung and hepatocellular carcinoma and was proved to be closely related to tumor metastasis, recurrence, and many other clinical features associated with poor prognosis [3-5]. KIF4A had previously been shown to be highly expressed in PDAC [6]; however, the mechanism of KIF4A in tumor development remained unclear and worth discussing.

In this study, we used bioinformatics to screen information from the online database and then locked KIF4A as a subject and validated bioinformatic results by immunohistochemistry (IHC). We then explored the relationship between KIF4A and the clinical outcome of PDAC through a five-year follow-up and found that KIF4A is closely related to TNM staging, tumor grade, lymph node metastasis, and vascular tumor suppository. It was also confirmed by in vitro experiments that KIF4A could promote the proliferation, invasion, and metastasis of pancreatic cancer cells. In animal experiments, KIF4A was knocked down by shRNA and the tumor volume of nude mice was reduced.

2. Materials and Methods [7]

2.1. Bioinformatic Analysis

We used GEPIA (http://gepia.cancer-pku.cn/) to collate and analyze (The Cancer Genome Atlas) TCGA data with a threshold of P < 0.05 and LogFC > 1 or <-1 for differential genes, and the median was used as the basis for dividing the patients into two groups for Kaplan-Meier survival analysis. The 95% confidence interval is marked with a dotted line. In addition, the correlation between KIF4A and CDC5L, BRCA1, and BRCA2 was analyzed by GEPIA.

2.2. Specimen Collection

Based on the institutional review board's approval, this retrospective study identified 86 consecutive patients with PDAC who underwent pancreatic cancer resection in Tianjin Medical University Cancer Institute and Hospital from 2005 to 2009. Patients were classified according to the 2009 UICC TNM staging and the 2004 WHO/ISUP classification. All patients had no preoperative radiotherapy and/or chemotherapy and neoadjuvant chemotherapy. Attendees provided written informed consent. All patients showed no evidence of tumor metastasis confirmed by cross-sectional imaging.

2.3. IHC and Scoring

Specimens were fixed with formalin, then embedded in paraffin, and subsequently cut into 4 μm sections. The latter was dewaxed with xylene and hydrated with gradient ethanol, heated to 121°C in 1 mmol/l EDTA (pH 8.0), cooled to 92°C, and incubated for 5 minutes. The sections were then immersed in 3% H2O2 deionized water for 10 minutes to eliminate endogenous peroxidase activity and sealed with goat serum (ZSGB-BIO, China) for 30 minutes. Rabbit immunization was then performed on serial sections with primary mouse anti-human KIF4A antibody (1 : 1000 dilution, ab8226, Abcam plc, Cambridge, UK) or primary rabbit anti-CDC5L (1 : 250 dilution, ab129114, Abcam plc., Cambridge, UK) and goat anti-mouse IgG (ZSGB-BIO). Histochemistry was followed by instillation with diaminobenzidine (DAB) for 5 minutes. Next, sections were counterstained with hematoxylin and dehydrated in ethanol, clarified in xylene, and covered with glass slides. Primary antibodies were replaced with PBS as a negative control.

The scoring method of KIF4A expression is based on the stained area and staining intensity. The quantification is as follows: the intensity of KIF4A expression is negative, weak, moderate, and strong. And the IHC staining was evaluated by two separate pathologists. In this study, slides with negative/weak intensity were classified as low KIF4A expression, whereas moderate/strong intensity was classified as high KIF4A expression.

2.4. Cell Line and Cell Culture

The human pancreatic cancer PANC-1 and BxPC-3 cell line was obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured according to the instructions from the American Type Culture Collection (ATCC). PANC-1 and BxPC-3 cells were maintained in DMEM and RPMI-1640, respectively, supplemented with 10% fetal bovine serum (FBS, Gibco) and incubated at 37°C in a humidified chamber containing 5% CO2.

2.5. RT-PCR and Primers

RNA was extracted with the Trizol kit (Invitrogen, CA) according to standard protocols provided by the manufacturer. CDNA was synthesized with 2 μg total extracted RNA using Roche-Transcriptor First Strand cDNA Synthesis Kit (Roche, USA). The expression of KIF4A was quantified using real-time polymerase chain reaction (RT-PCR) using the SYBR Green PCR kit (Applied) using primers as follows: KIF4A forward 5′-TCTGTTTCAGGCTGCTTTCA-3′ and KIF4A reverse 5′-GCCCTGAAATATTTGATTGGAG-3′; GAPDH forward 5′-CGGCTACCACATCCAAGGAA-3′ and GAPDH reverse 5′-GCTGGAATTACCGCGGCT-3′.

Duplicate tests were performed on each sample, and relative mRNA expression was determined based on CT values and normalized by the same endogenous reference GAPDH expression levels.

2.6. RNA Interference and Transfection

We used specific short hairpin RNA (shRNA) with the sequence of AACAGGAAGAAGTCTTCAATACA (TL311908V) to target KIF4A, and scrambled sequence was used as negative control. Both two shRNA were synthesized by Origene Co., Ltd. (Beijing, CHINA 101111). Invitrogen Lipofectamine® 3000 (Thermo Fisher Scientific, Inc.) was used for shRNA transfection. There are 100,000 cells per well in six-well plates according to the manufacturer's protocol, and then, 3 groups were set, including the sh-KIF4A group, which transfected with shRNA targeting KIF4A; negative control group, which transfected with scrambled sequence; and mock group, which was treated without transfection (data not shown). Silence efficiency was measured by RT-PCR and western blot after 48 hours' transfection. These reduced cells were used to explore the association between KIF4A and cell proliferation and cell invasion.

2.7. Western Blot

A total of 50 μg of protein was transferred to nitrocellulose filters after 10% SDS-PAGE. The membranes were blocked and incubated with primary antibodies (including anti-KIF4A antibody (ab122227, 1 : 400)), mouse anti-β-actin (1 : 1000 dilution, ab8226, Abcam plc.) monoclonal antibodies overnight at 4°C, and then incubation with the secondary antibody (polyclonal goat anti-rabbit/mouse, 1 : 10,000 dilution) (Rockland Immunochemicals, Inc. Limerick, PA) for 1 hour. At last, the gray values were analyzed by using the Odyssey V3.0 software.

2.8. Cell Proliferation

Colony-forming assays were used to test proliferative capacity. A 6-well plate was used to grow the cell suspension, 300 cells per well. After incubation, cells are fixed in methyl hydrate for 10 minutes. Colonies were then stained and counted with light microscopy.

2.9. (4,5-Dimethyl-2-Thiazolyl)-2,5-Diphenyl-2-H-Tetrazolium Bromide (MTT) Assay

2 × 103 cells were seeded into 96-well plates, and 10 μl of 5 mg/ml MTT was added to each well, and incubation was continued for another 4 hours. After adding dimethylsulfoxide, the plate was placed on a microplate reader (Bio-Rad, Hercules). The optical density was read at a wavelength of 570 nm, and the cell growth curve was determined based on the optical density value.

2.10. Cell Scratch Analysis

Cells were cultured in 6 wells overnight. The next day, a scratch was introduced in the middle of the well using a sterile pipette tip. The medium is discarded and replaced with a fresh medium. Determine the speed of movement to the center of the wound at a specified time. The area of the wound closed was calculated at indicated time points.

2.11. Transwell Cell Invasion Assay

2 × 105 cells were isolated before being added to the upper chamber of transwell coated with matrigel (BD Biosciences, Mountain View, CA, USA) in 8.0-well microporous inserts in 24-well transwell chambers (Costar, USA). RPMI 1640 with 10% FBS was added to the lower chamber and incubated for 24 hours. Cells that migrated to the bottom of the filter were stained with a three-step staining kit (Thermo Scientific). Cells are counted from each chamber through a microscope.

2.12. In Vivo Xenograft Assays

We purchased the nude BalB/c female mice (6-8 weeks, 18-22 g) from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China), and then, the mice were housed in pathogen-free animal facilities and randomly assigned (n = 8 per group). 2 × 106 cells were injected subcutaneously into nude mice, and the tumor volume was measured. Mice were randomized into two groups and treated with PANC-1/control cells and PANC-1/KIF4A shRNA cells. Tumors were measured by a vernier caliper on days 14, 17, 21, 23, 26, and 29. Besides, PANC-1 cells transfected with KIF4A shRNA and controls were injected into mouse tail vein to investigate the association between KIF4A and metastasis of PDAC cells. The metastatic tumors of the lung were detected; after 4 weeks, the mice were euthanized and tumors were harvested. The following formula was used to calculate the tumor volume: Tumor volume (mm3) = Tumor length (mm) × Tumor width (mm)2/2.

2.13. Statistical Analysis

SPSS 22.0 software (SPSS Inc, IBM Corp., Armonk, NY) was used to analyze the data. For the immunohistochemistry experiments, associations between KIF4A expression and the clinicopathological features were evaluated by using χ2 tests. Survival analysis was estimated by the Kaplan-Meier method and log-rank tests. Student's t-test was used for statistical comparisons, and the data were shown as the mean ± standard deviation (SD) in vitro and in vivo experiments on PANC-1 and BxPC-3 cells. A value of P < 0.05 was set to be statistically significant.

3. Results

3.1. Bioinformatic Analysis of KIF4A in PDAC Patients

We first analyzed the expression level of KIF4A in pancreatic cancer patients using GEPIA [7], an online analytical database that analyzes cases from TCGA, and found that KIF4A expression was significantly increased in pancreatic tumor tissue compared with normal pancreatic tissue (Figure 1(a), P < 0.05). A total of 179 cases of TCGA patients were divided into high- and low-expression groups according to the median. The patients with higher level of KIF4A shared a significantly worse disease-free survival and overall survival compared with the lower expression cases (Figure 1(b), P = 0.015 < 0.001, P = 0.017 < 0.001, respectively). Meanwhile, we also examined the protein expression of KIF4A in peripheral blood and in the tumor and found that KIF4A is of guiding significance for the prognosis of pancreatic cancer, and the expression of KIF4A can also be found in many other kinds of tumors (Supplementary Figure S1). These results give a clue of a potential role of KIF4A in PDAC.

Figure 1

Bioinformatic analysis of KIF4A in cases from TCGA. (a) The expression of KIF4A between cancer (n = 179) and normal (n = 171) cases, cancer cases were filled in red, and normal cases were in blue: KIF4A expression was significantly increased in pancreatic tumor tissue compared with normal pancreatic tissue (P < 0.05). ∗P < 0.05. (b) The overall survival and disease-free survival rate of 178 TCGA cases with clear clinical information. High and low groups were divided by median and filled in red and blue, respectively. Comparing with the lower expression cases, patients with higher level of KIF4A shared a significantly worse disease-free survival and overall survival (P < 0.05, respectively).

3.2. KIF4A Is Highly Expressed in PDAC Patients and Associated with Various Clinic-Pathological Features

To confirm the results gotten from GEPIA, 86 specimens from patients were chosen, followed by immunohistochemistry and clinical analysis (Figure 2(a)). High KIF4A positive expression was found in 54/86 (62.8%) cases, and low KIF4A expression was observed in 32/86 (37.2%) cases, respectively (Table 1). From the results of our statistics in Table 1, the high KIF4A expression group had a significant difference in TNM staging (P = 0.003 < 0.001) and tumor size (P < 0.0001) when compared with the low-expression group, whereas no significant difference was observed in other features such as age, gender, or tumor grade.

Figure 2

Clinical features of 86 patients. (a) The immunohistochemistry results of the KIF4A high expression group and low expression groups. (b) The associations between KIF4A and prognosis of patients were performed by the Kaplan-Meier analysis of the survival time after surgery; high level of KIF4A was associated with short OS and P < 0.05.

Table 1

Relationships of KIF4A and clinicopathological characteristics in 86 patients with pancreatic ductal adenocarcinoma.

Feature	All n = 86	KIF4A expression		χ 2	P
		Low	High
		n = 32	n = 54
Age (year)				2.657	0.103
<65	50	15	35
≥65	36	17	19
Gender				2.335	0.127
Male	52	16	36
Female	34	16	18
pTNM stage				8.364	0.004∗
I-II	50	25	25
III	36	7	29
Tumor size				12.138	0.000∗
<5	22	15	7
≥5	64	17	47
Tumor grade				0.175	0.676
Low	54	21	33
High	32	11	21

3.3. The Associations between KIF4A and Clinical Outcomes in PDAC Patients

We then performed a 5-year overall survival (OS) and disease-free survival rate (DFS) for these 86 patients and showed the results by Kaplan-Meier survival analysis (Figure 2(b)). As shown in Figure 2, the OS and DFS after surgery with high expression of KIF4A were remarkably different from those in the low-expression group. The OS of patients with high expression of KIF4A was less than 20 months, while less than 20% of the patients in the low-expression group died at this time. In DFS, the high-expression group relapsed within 12 months, while the low-expression group just over half. The above results showed that the level of KIF4A expression and clinical outcomes have a significant relationship (P < 0.05 and P < 0.05, respectively).

3.4. Downregulation of KIF4A mRNA and Protein Expression Inhibits Cell Proliferation in PANC-1 and BxPC-3 Cells

The expression of KIF4A mRNA and protein in vitro was detected by RT-PCR and western blot. KIF4A was reduced in cancer cells by interference with transfection of the recombinant lentiviral vector LV-KIF4A shRNA. KIF4A mRNA and protein expression were significantly reduced in KIF4A shRNA cells compared to control cells (Figure 3).

Figure 3

KIF4A shRNA efficiency. (a) The shRNA efficiency measured by RT-PCR. (b) The shRNA efficiency measured by western blot. Quantified boxplot was shown in the following. The expression of KIF4A in the shRNA group was significantly decreased compared to controls (P < 0.05, respectively). All results were performed in three independent experiments. ∗P < 0.05.

We used colony formation and MTT assays to assess the possible impact of KIF4A silencing on pancreatic cancer cell proliferation. Stable silencing of KIF4A in KIF4A-shRNA cells obviously reduced the in vitro proliferative capacity of PANC-1 as well as BxPC cells compared to control cells (P < 0.05, Figures 4(a) and 4(b)). The above results suggested that silencing KIF4A significantly reduces the proliferation of pancreatic cancer cells. Taken together, our results suggested that silencing of KIF4A can reduce proliferation by regulating in vitro experiments.

Figure 4

KIF4A knockdown inhibits cell proliferation. (a) Colony formation assay in PANC-1 and BxPC-3 cells, quantified boxplot was shown on the right. The results of colony formation array found that colonies in the shRNA group were less than controls (P < 0.05, respectively). (b) MTT assay in PANC-1 and BxPC-3 cells, and the OD value was measured at 570 mm: OD values in the shRNA group were decreased (P < 0.05, respectively). All results were performed in three independent experiments. ∗P < 0.05.

3.5. KIF4A Silencing Inhibited Migration or Invasion in Pancreatic Cancer Cells

We used scratch assay to detect the changes in cell migration capacity. KIF4A-shRNA cells were found to show more obvious scratch space compared to control (P < 0.05, Figure 5(a), respectively). In addition, transwell invasion assays were also used to explore the possible role of KIF4A silencing in regulating cancer cell invasiveness. The results show that the invasion ability of KIF4A-shRNA cells decreased compared to control after 36 hours' culture (P < 0.05, Figure 5(b)). In a word, our results suggested that KIF4A silencing could inhibit cancer cell invasion and migration.

Figure 5

KIF4A knockdown inhibits cell migration or invasion. (a) Scratch test in PANC-1 and BxPC-3 cells. The distance between edges in the shRNA group was wider than controls (P < 0.05, respectively). (b) Transwell assay in PANC-1 and BxPC-3 cells. The ability of invasion in the shRNA group was decreased significantly (P < 0.05, respectively). All results were performed in three independent experiments. ∗P < 0.05.

3.6. Knockdown of KIF4A Restricted the Growth and Metastasis of Tumors in Mice

From the above results, we concluded that silencing KIF4A inhibits proliferation and suppresses invasion and metastasis by regulating related proteins in vitro. To further confirm this conclusion, we performed nude mouse in vivo experiments. Nude mice were randomly divided into two groups, respectively, treated with KIF4A-shRNA cancer cells or control pancreatic cancer cells. The results showed that after 3 weeks, the tumor volume of the KIF4A-shRNA group was significantly smaller than that of the control group (P < 0.05, Figure 6(a)). Furthermore, PANC-1 cells in two groups were injected into mouse tail vein to observe the difference of lung metastasis. The results suggested that the metastatic tumors in the KIF4A shRNA group were obviously smaller than controls (P < 0.05, Figure 6(b)). In addition, we evaluated KIF4A expression in subcutaneous mouse tumors by immunohistochemistry and showed that the expression of KIF4A was reduced by KIF4A-shRNA (P < 0.05, Figure 6(c)). Moreover, western blotting was performed to confirm the above result; from Figure 6(d), we could also find a significantly lower KIF4A level in the KIF4A knockdown group. These results were further proved that KIF4A-shRNA had these inhibitory effects in vivo.

Figure 6

In vivo KIF4A knockdown inhibits tumor growth. (a, b) Tumor volume in the KIF4A knockdown and control nude mouse group after 4 weeks, respectively. From growth curve, tumors in the shRNA group grew more slowly than controls and tumor volumes were smaller in the shRNA group (P < 0.05, respectively). (c, d) Knockdown efficiency tested by western blot and immunohistochemistry, and the boxplot results were shown in the following, respectively. The expression of KIF4A in mouse tumors was dramatically decreased by immunohistochemistry and western blot (P < 0.05, respectively). All results were performed in three independent experiments. ∗P < 0.05.

3.7. KIF4A Promotes Pancreatic Cancer Proliferation by Affecting CDC5L

We learned from the results shown above that KIF4A could regulate cell proliferation of pancreatic cancer both in vitro and in vivo; we then further sought to reveal the potential mechanism of KIF4A involved in tumor growth. Firstly through GeneCards, we found the positive gene interactions between KIF4A and CDC5L (https://www.genecards.org/cgi-bin/carddisp.pl?gene=KIF4A, Figure 7(a)). Then, the bioinformatic method was used to analyze the relationship between KIF4A and BRCA1, BRCA2, and CDC5L: http://gepia.cancer-pku.cn/detail.php?gene=KIF4A, and the results showed the existing strong positive correlation (R = 0.52, 0.42, and 0.34, P < 0.05, respectively, Figure 7(b)). Then, the KIF4A and CDC5L gene and protein expression were shown in the related website (https://www.proteinatlas.org/ENSG00000090889-KIF4A or ENSG0000096401-CDC5L/pathology), and we could find the results in the following: the expression level of KIF4A protein in peripheral blood of pancreatic cancer has guiding significance for prognosis (P < 0.05, Supplementary Figure S1), and there are more KIF4A and CDC5L proteins in tumor (Supplementary Figure S1 and S2). Therefore, we further consult GEPIA (http://gepia.cancer-pku.cn/) and found in the tumor that mRNA of CDC5L expression was obviously higher than that of normal pancreatic tissue (Supplementary Figure S2). After that, we performed immunohistochemical analysis of CDCC5L-expressed proteins and found that the staining was deep in the high-expression group and light in the low-expression group (Figure 8). In addition, we analyzed the relationship between CDC5L expression and the clinicopathological characteristics of pancreatic cancer and found that high expression was associated with tumor size (P = 0.026 < 0.05) and TNM stage (P = 0.014 < 0.05) but not with age (P = 0.259), gender (P = 0.186), or differentiation (P = 0.165) (Table 2). These results are very similar to those of KIF4A. Moreover, immunohistochemical analysis of the tumor tissues revealed that the CDC5L expression was low or high in the low- or high-expression KIF4A group (Figure 8). Correlation analysis by using our data confirmed the significant positive correlation between the expression of CDC5L and KIF4A (P < 0.05) (Table 3). According to these results, we investigated that KIF4A promoted pancreatic cancer by affecting the CDC5L signaling pathway.

Figure 7

Correlation analysis between KIF4A and CDC5L by bioinformatic analysis. (a) Mechanism diagram of gene interaction between KIF4A and BRCA1, BRCA2, and CDC5L by GeneCards. (b) Correlation analysis of the expression of KIF4A and BRCA1, BRCA2, and CDC5L by GEPIA.

Figure 8

The expression of KIF4A is correlated with CDC5L in tumor tissues of pancreatic cancer. Representative images of immunohistochemical staining of KIF4A and CDC5L expression in tumor tissues.

Table 2

Relationships of CDC5L and clinicopathological characteristics in 86 patients with pancreatic ductal adenocarcinoma.

Feature	All n = 86	CDC5L expression		χ 2	P
		Low	High
		n = 40	n = 46
Age (year)				0.303	0.582
<65	50	22	28
≥65	36	18	18
Gender				3.426	0.064
Male	52	20	32
Female	34	20	14
pTNM stage				8.735	0.003∗
I-II	50	30	20
III	36	10	26
Tumor size				8.167	0.004∗
<5	22	16	6
≥5	64	24	40
Tumor grade				0.249	0.618
Low	54	24	30
High	32	16	16

Table 3

Relationships of KIF4A and CDC5L in 86 patients with PDAC.

All n = 86	KIF4A		χ 2	P	Pearson	Spearman
	Low32	High54
CDC5L			10.131	0.001	0.343	0.343
Low40	22	18
High46	10	36

4. Discussion

PDAC is a malignant tumor of the digestive tract, which is very malignant and difficult to diagnose and treat [1]. PDAC is one of the most refractory cancers of chemotherapy, and the current therapies for its related molecular pathways have not yielded satisfactory results, due in part to the rapid survival of the cancer as a result of the rapid reimbursement pathway. Most of the available treatments were palliative to improve the quality of life, but they could not achieve the goal of inhibiting or nurturing PDAC. Currently available options were surgery, embolization, radiotherapy, chemotherapy, immunotherapy, and other more targeted drugs [8].

KIFs are motors for anterograde transport of mitochondria that are involved in cell cycle regulation [9]. KIF4A encodes a member of the kinesin 4 kinesin subfamily, which is an ATP-dependent tubulin-binding protein that is involved in the trafficking of cellular endocytic organelles. It is also associated with condensed chromosome arms and may be involved in maintaining chromosomal integrity during mitosis [10-16]. From the previous research, we could see that the KIF superfamily play an important role in various kinds of cancer [17-21], and the overexpression of mitotic kinesins like KIFs might lead to results including premature sister chromatid segregation, excessive spindle separation, overshooting before anaphase, and eventually bipolar or monopolar spindle formation [17], which then lead to cancer proliferation or metastasis phenotypes [18, 19]. On the opposite, the downregulation of KIFs might cause mitotic arrest, which could reduce the progression of cancer cells [20, 21].

Members in the kinesin 4 subfamily such as KIF14 were shown to correlate with tumor progression and poor prognosis in prostate cancer [22], and KIF2A overexpression was found in breast cancer and was closely associated with overall survival and lymph node metastasis [23, 24]. KIF4A functioned at a broad range, a recent research also indicated the role of KIF4A in DNA damage response [25]. In recent years, KIF4A has been shown to play a catalytic role in a variety of cancers [3–5, 12, 26–32]. For example, KIF4A promoted the proliferation of lung cancer cells by forming a complex with PHF14 and activation of KIF4A might be a prognostic biomarker and therapeutic target for lung cancer [5, 26], and KIF4A was closely related to the recurrence of liver cancer in hepatocellular carcinoma [4]. The overexpression of KIF4A results in the inhibition of proliferation in gastric carcinoma cells [27], and KIF4A could mediate the accumulation and reeducation of THP-1-derived macrophages via regulation of CCL2-CCR2 expression [28]. KIF4A was also found to play roles in kidney renal cancer [12], breast cancer [29], colorectal cancer [30], prostate cancer [31], and hepatocellular carcinoma [32] and has also been shown to be associated with chemoresistance mechanisms in tumors [3]. It can be seen that KIF4A is receiving more and more attention as a factor of poor prognosis of the tumor, but it is a pity that so far, no research on it has been seen in the field of pancreatic cancer. Therefore, we decided to analyze it first using a web-based database [7]. Interestingly, the results showed that KIF4A was obviously associated with overall survival and recurrence-free survival in patients with pancreatic cancer, and KIF4A expression was also significantly increased in PDAC patients. We were focused on this and subsequently collected tissue samples and clinical information from 86 patients and performed immunohistochemistry and Kaplan-Meier survival analysis, and the results were consistent with the network database, indicating that KIF4A might exist something of a potential in PDAC patients, but not yet discovered.

The published studies mainly focused on the effects of KIF4A in shaping mitotic chromosomes and chromosome segregation, where KIF4A functioned as an essential chromosome-associated molecular motor [33, 34]. KIF4A was also found to regulate the size of the central spindle [13, 35, 36]. In cancer, KIF4A also plays a role in many aspects, as described above. Based on these studies, we knocked out KIF4A using shRNA and conducted a series of in vitro experiments to investigate its function, including colony formation assay, MTT, cell scratch assay, and transwell assay. It was demonstrated that knocking out KIF4A in vitro could inhibit the proliferation of pancreatic cancer cells and inhibit the invasion and metastasis. On the other hand, we also conducted animal experiments in nude mice and found that the knockout of KIF4A results in tumor shrinkage in nude mice, which means the knockdown of KIF4A could inhibit tumor development and metastasis in vivo.

Based on our present data, we believed that the high expression of KIF4A could promote the proliferation and metastasis of PDAC in vitro and in vivo, thereby promoting the recurrence and metastasis of pancreatic cancer, shortening the survival of patients, and resulting in poor prognosis. PDAC is a very complex disease, in which many different substances are associated and play a decisive role. Many of these mechanisms have not yet been explored yet, and the research on PDAC needs to be further developed. Our study provided preliminary information for further study and understanding of the molecular mechanism of KIF4A involved in PDAC proliferation and metastasis potential and provided the basis for the diagnosis and treatment of PDAC in the future. The multiple roles of KIF4A in PDAC patients might be able to prompt the future treatment of PDAC. Thus, KIF4A might become a new therapeutic target of immunotherapy for patients with PDAC.

There are some limitations in this study; for that, it is retrospective and includes only a small number of patient samples from a single center. Few studies have been reported about the detailed mechanism of how KIF4A regulates tumor growth in PDAC. What is more, we only demonstrated the positive correlation between KIF4A and CDC5L by using TCGA data (GEPIA and GeneCards) and IHC (our clinical data), and we could say that KIF4A might promote pancreatic cancer proliferation by affecting the expression of CDC5L expression in the pathway, and further verification of this mechanism is needed in the future.

In summary, we firstly investigated the expression of KIF4A in PDAC tumor samples of GEPIA and the prognosis of survival and found that patients with high expression of KIF4A showed malignant progression and poor survival in PDAC. We then validated these results above with 86 collected samples. Then, we confirmed by cell experiments and animal experiments that KIF4A could promote PDAC in vitro and in vivo development, which was consistent with the findings of many other previous studies. Therefore, KIF4A might be a new target of PDAC. Further, the positive correlation between KIF4A and CDC5L existed, and KIF4A might promote pancreatic cancer proliferation by affecting CDC5L expression. These findings may contribute to more studies of KIF4A as a therapeutic target for the treatment of PDAC. However, these findings are preliminary and further studies are needed to confirm the results.