Clinical significance of “S” isoform of DCLK1 in different gastric cancer subtypes using newly produced monoclonal antibody

Abstract

Background

Doublecortin-like kinase 1 (DCLK1) isoforms play distinct roles in the progression of gastrointestinal cancers. For the first time ever, the current study aimed to generate DCLK1-S-specific monoclonal antibodies (mAbs) to evaluate the clinical value of DCLK1-S (short isoform) in gastric cancer (GC).

Materials and methods

Mice were immunized with a unique 7-mer synthetic peptide of DCLK1-S conjugated with keyhole limpet hemocyanin (KLH). Immunoreactivity of hybridomas and mAbs was determined by ELISA assays and immunohistochemistry (IHC). DCLK1-S expression in two GC cell lines was assessed by flow cytometry. After characterization, the expression pattern of DCLK1-S was investigated in different histological subtypes of GC (n=217) and adjacent normal tissues (n=28) using IHC on tissue microarrays. The association of clinical prognostic values with DCLK1-S expression was also investigated.

Results

ELISA findings demonstrated that the generated monoclonal antibody (mAb) exhibited strong immunoreactivity towards the immunizing peptide. Positive control tissues, including GC and colorectal cancer, showed strong positive immunoreactivity with anti-DCLK1-S mAb whereas negative reagent control sections represented no staining, demonstrating the specificity of produced mAb. Flow cytometry analysis confirmed that the newly developed mAbs effectively recognized DCLK1-S on the cell surface. A mixture pattern of membranous, cytoplasmic, and nuclear DCLK1-S expression in the GC cells was observed. A significant and inverse association was identified between the expression DCLK1-S in the cell membrane and cytoplasm and PT stage, muscolarispropia, subserosa, and perineural invasion in intestinal subtype, respectively. In signet ring cell type, however, nuclear DCLK1-S expression was adversely associated with tumor size and PT stage. Furthermore, patients with low DCLK1-S expression had a shorter survival than patients with high expression, however, without a statistically significant association.

Conclusion

An efficient and precise tool for detecting DCLK1-S in cancer tissues has been developed. Moreover, DCLK1-S overexpression might be considered a favorable clinical factor in GC patients.

Keywords

short isoform of doublecortin-like kinase 1 (DCLK1-S)hybridoma technique monoclonal antibody (mAb)gastric cancer immunohistochemistry tissue microarray

Introduction

Gastric cancer (GC) is the fifth most commonly diagnosed malignancy and third leading cause of cancer-related mortality worldwide,^1,2 with over 770,000 GC-related deaths, a five-year survival rate of less than 30%, and high prevalence in Asian, Eastern European, and South American countries.³ Despite important advances in the multimodal therapeutic approaches, the global burden of GC still remains considerable with a particularly dismal prognosis. In addition, late-stage diagnosis, lack of adequate predictive biomarkers, progression to aggressive and metastatic forms, and treatment failure are the main challenges that hinder prognosis.⁴ Of note, biomarkers represent valuable promising tools that could optimize early-stage diagnosis and prognosis, provide information regarding gastric malignancy pathophysiology, and predict clinical outcomes.⁵

A growing body of evidence has highlighted the pivotal function of cancer stem cells (CSCs) in cancer initiation, progression, metastasis, and chemoresistance; in light of this, multiple strategies have been developed to recognize and target specific CSC markers in different types of cancers, including GC.⁶ More recently, doublecortin-like kinase 1 (DCLK1), a microtubule-associated protein, has been suggested as a novel, tumor-specific stem cell marker in gastrointestinal cancers. Nakanishi et al. first discovered that DCLK1 can be used as a biomarker for distinguishing between cancer and normal stem cells in the intestine.⁷ Thereafter, various investigations indicated that downregulating DCLK1 expression results in growth arrest in colorectal,⁸ renal,⁹ and pancreatic carcinomas.¹⁰ Mechanistically, DCLK1 modulates NF-kB, NOTCH, AKT, and KRAS molecular signaling cascades that support tumor progression and promote stemness and EMT of cancer cells.^11–13 Additionally, many of the investigations performed by our and other research teams have identified DCLK1 as being highly expressed in several malignancies, including colorectal,¹⁴ pancreatic,¹⁵ hepatocellular,¹⁶ renal,⁹ and bladder cancers.¹⁷

The human DCLK1 gene consists of two promoters, a 5′ (α) promoter which drives the transcription of the long (L)-isoform (DCLK1-L, ∼ 80–82 KDa) and a 3′ (β) promoter which is responsible for the short (S)-isoform expression (DCLK1-S, ∼ 45–50 KDa).¹⁸ Greater than 98% homology exists at the C-terminal end of both short and long isoforms. In contrast, the L-isoform and S-isoform differ mostly at the N-terminal end, including the absence of the doublecortin domain as well as a different sequence of six amino acids at the N-terminal of the DCLK1-S isoform.¹⁹ Recent findings have shown the isoform particular role of DCLK1 in cancer progression. Evidence indicates that the α-promoter, which drives the expression of the DCLK1-L isoform, is epigenetically hypermethylated and silenced during the adenoma–carcinoma sequence of gastrointestinal cancers.¹⁸ Thus, it has been suggested that DCLK1-L marks only normal human cells and early stage cancer cells. In contrast, the alternate β-promoter is used for transcription of the short DCLK1 isoform at the advanced stages of malignancies, such as CRC.¹⁸ Importantly, several recent studies have shown the correlation of DCLK1-S overexpression with shorter survival outcomes in CRC²⁰ and esophageal squamous cell carcinoma (ESCC) patients.²¹ Our recent findings also strengthen the crucial role of DCLK1-S, as an independent prognostic factor, in cancer aggressiveness in colorectal patients.²²

Regarding the importance of DCLK1 in GC, DCLK1 was first recognized as a potential driver based on a comprehensive genomic and molecular analysis of 100 primary GCs.²³ Limited research by our and other groups has identified the significance of DCLK1 in GC patients.^14,24,25 However, all these studies used commercially available antibodies that target the C-terminal end of DCLK1 proteins, which is common to both long and short isoforms, or only the N-terminal doublecortin domains of L-isoform, which the short isoform lacks these domains. To the best of our knowledge, monoclonal antibodies (mAbs) specific for short isoforms with immunohistochemical application are not commercially available. The detection of DCLK1-S using mAb in cancerous and normal tissues can be of great value for cancer investigations. Additionally, the expression pattern and prognostic importance of DCLK1-S in most carcinomas, including gastric cancer, remain unclear. Therefore, these gaps prompted us to develop anti-DCLK1-S mAbs against the unique N-terminal six amino acids sequence present only in DCLK1-S. Afterwards, the expression and clinicopathological prognostic value of DCLK1-S were evaluated, using our laboratory produced anti-DCLK1-S mAbs, in a large series of GC tissue microarrays (TMAs) by immunohistochemistry (IHC) technique.

Material and methods

Generation and characterization of anti-DCLK1-S-specific monoclonal antibodies

Preparation of immunogen

High-purity synthetic peptide selected from the unique sequence of DCLK1-S isoform (NH2-Met-Leu-Glu-Leu-Ile-Glu-Cys-COOH) was purchased from BIO BASIC INC (Canada). Imject Maleimide-activated bovine serum albumin (BSA) (Thermo Scientific, Rockford, IL, USA) and Maleimide-activated keyhole limpet hemocyanin (mcKLH) (Thermo Scientific, Rockford, IL, USA) were utilized to be conjugated separately with peptide. To investigate the conjugation efficacy, the sodium salt-polyacrylamide gel electrophoresis (SDS-PAGE) pattern of the BSA conjugate was investigated.

Immunization protocol

Female BALB/c mice (6–8 weeks old) were purchased from Pasteur Institute of Iran (Tehran, Iran) and maintained in standard conditions at the animal facilities of Iran University of Medical Sciences (IUMS). All protocols were approved by the animal research committee of IUMS. Each mouse was immunized intraperitoneally five times at three-week interval between the first and second injection and two-week intervals between next injections. For the first immunization, 50 μg KLH-peptide emulsified in complete Freund's adjuvant was utilized for the intraperitoneal immunization of each BALB/c mouse. For the subsequent four booster immunizations, each mouse was immunized with 50 μg the KLH-peptide in incomplete Freund's adjuvant. Subsequently, blood from the tail vein was obtained, ten days after the last injection, and the antibody titer was determined in each mouse sera using enzyme-linked immunosorbent assay (ELISA) against DCLK1-S peptide, as described below.²⁶ According to ELISA results, three days prior to cell fusion, 20 μg of the antigen was injected intravenously into the mouse with the highest antibody titer.

Indirect ELISA

Indirect ELISA was applied to assess and confirm the reactivity of mice sera, secreting hybrids, and the purified mAbs against DCLK1-S. Briefly, 96-well ELISA plates were coated with a total volume of 100 μl DCLK1-S peptide (10 µg/ml) diluted in phosphate-buffered saline (PBS, 0.15 M, pH 7.2). Following incubation at 37 °C for 1 h, the plates were incubated at 4 °C overnight. The following day, after three times washing with PBS containing 0.05% Tween 20, the wells were blocked for one hour with 2.5% skimmed milk at 37 °C. Following that, the wells were incubated with serial dilutions of mice sera, secreting hybridomas, or serial dilutions of purified anti-DCLK1-S mAbs for 1.5 h at 37 °C. Following washing, 1:1000 dilution of horseradish peroxidase (HRP)-conjugated sheep anti-mouse immunoglobulin (Ig) (Avicenna Research Institute, Tehran, Iran) was added to all wells at 37 °C for 1.5 h. After washing, tetramethylbenzidine substrate (TMB) substrate (Sigma-Aldrich Chemie GmbH, Germany) was applied for each well and H2SO4 (20%) was added to stop the reaction. Finally, signals were measured on a microplate reader (Synergy HTXMulti-Mode Reader, BioTek, USA) at 450 nm.

Cell culture

Murine myeloma cell line (SP2/0) and gastric cancer cell lines (MKN-45 and AGS) (all from National Cell Bank of Iran, Pasteur Inst. of Iran, Tehran) were cultured in either RPMI 1640 (Gibco) (for SP2/0, MKN-45) containing 10% fetal bovine serum (FBS) (Gibco) or Dulbecco's modified Eagle medium (Gibco, Invitrogen, CA, USA) (for AGS) containing 10% FBS. Cultured cells were maintained within an incubator containing humidified 5% CO2 atmosphere at 37 °C.

Generation of anti-DCLK1-S monoclonal antibodies

mAbs were generated by standard hybridoma technology, as described elsewhere.^26,27 Briefly, spleen lymphocytes were isolated from immunized mice and fused with SP2/0 cell line through polyethylene glycol (PEG) 1500 (Sigma) at a 5:1 ratio. Then, RPMI medium containing hypoxanthine-aminopterin-thymidine (HAT) and 20% FBS was used for the selection of hybridisms. After 10–12 days of fusion, indirect ELISA was applied to screen the supernatants of hybridomas. Using limiting dilution method, hybridomas capable of producing specific antibody were cloned four times. According to the results obtained from screening steps, finally, two clones, including 2D5C2 and 2H3D5, were opted for the subsequent experiments. Protein G affinity chromatography (Amersham Biosciences, the Netherlands) was applied to purify the monoclonal antibodies. The concentration of purified mAbs was spectroscopically determined at 280 nm. The purity of the mAbs was assayed by electrophoresis and the reactivity of the purified antibodies was evaluated using indirect ELISA technique.

Monoclonal antibody characterization by immunohistochemistry

The specificity of anti-DCLK1-S mAb was determined by IHC. In this study, humans colorectal and gastric cancer tissues were utilized as positive control for our generated mAb against DCLK1-S. In isotype control slides, mouse IgG (Avicenna Research Institute, Tehran, Iran), instead of the primary antibody, was utilized to validate anti-DCLK1-S mAbs specific bindings. Additionally, prior to the application of the primary antibody solution to the TMA sections, primary antibody was preincubated with a saturating concentration of immunizing peptide (1:100 molar ratio) in negative reagent control slides. For further verification of the specificity of our mAb, according to protein atlas databases, normal tissue of testis as well as skin and ovary cancerous tissues as negative control tissues were also utilized.

Analysis of DCLK1-S expression in gastric cancer cell lines by flow cytometry

Flow cytometry was applied to investigate the DCLK1-S expression in GC cell lines. MKN-45 and AGS cells at 70%–80% confluency were harvested. Following three times wash with cold PBS (pH 7.2, 0.15 M), 5% sheep serum was added to cells for 20 min on ice. Sequentially, cells were treated with either anti-DCLK1-S mAb diluted in PBS-2%FBS (10 μg/mL) or equal concentration of non-immunized mouse Ig (as negative control) on ice for 45 min. After three times washing, cells were incubated with FITC-conjugated sheep anti-mouse Ig (1:75, Avicenna Research Institute, Iran) for 30 min. Finally, cells were washed three times. Data was obtained via an Attune NxT flow cytometer (Thermo Fisher Scientific, USA) and analyzed by FlowJo software VX.

Population study

Sample collection and patients characteristics

In the current cross-sectional study, a total of 217 formalin-fixed, paraffin-embedded (FFPE) specimens from GC cases were enrolled from university-based referral hospitals in Tehran, Iran. Cases did not receive any preoperative chemotherapy or radiotherapy prior to surgery. Patient's hematoxylin and eosin (H&E) stained pathology sections and medical archived data were evaluated to obtain the clinical and pathological features: gender, age, tumor size (maximum tumor diameter), histological grade, histological subtype, tumor site, PT stage, tumor extension, lamina propria involvement, muscularis mucosa involvement, muscularis propria involvement, subserosa involvement, perineural invasion (PI), lymphovascular invasion, distant metastasis, and tumor recurrence. In addition, 28 adjacent normal tissues were utilized for assessment of DCLK1-S expression in comparison to cancerous specimens. Additionally, the patient's survival outcomes, including disease-specific survival (DSS) and progression-free survival (PFS), were reviewed. DSS was considered as the interval from surgery until death from patient's cancer. The PFS was calculated as the length of time between the primary surgery and the last follow-up with no evidence of disease, distant metastasis, or recurrence. This research was performed with Ethical approval by the Ethics Committee of the Iran University of Medical Sciences (Ref No: IR.IUMS.REC-98–4-28-16306) and informed written consents were obtained from all patients before inclusion.

Tissue microarray (TMA) construction

The GC TMA blocks construction was conducted, as reported previously.^28,29 Briefly, to prepare TMA blocks, three most representative regions from different areas of the tumor and normal tissues were opted and marked after matching with H&E slides. Afterward, the marked tumor regions of each donor block with 0.6-mm diameter were punched out by tissue arrayer instrument (ALPHELYS, Plaisir, France) and transferred into the TMA recipient paraffin blocks. Finally, the sectioning of GC TMA blocks was performed at 4 μm thickness and prepared sections were transferred to microscope slides (SuperFrost plus, Thermo Scientifc™, Germany). In this research, each tumor sample was constructed in three cores to cover tumor heterogeneity challenge during the TMA process, and, then, scored individually for precise evaluation of the specimens. To increase the validity and accuracy of the data analysis, the final scores for tissue specimens were calculated from the mean of the three cores.

Immunohistochemistry

Immunostaining of GC TMA slides for DCLK1-S was conducted through our previously described laboratory protocol.^22,30 Briefly, after deparaffnization of all TMA slides (at 60 °C and then by xylene), the sections were rehydrated under serial dilutions of ethanol. Afterward, to block endogenous peroxidase activity, tissue slides were incubated in methanol containing 3% hydrogen peroxide (H2O2) at room temperature for 20 min. After three times wash of the slides in Tris Bufered Saline (TBS), antigen retrieval was conducted through autoclaving the slides in sodium citrate buffer (pH 6.0) for 10 min. Afterwards, following three times wash, the slides were treated with our generated primary anti-DCLK1-S mAb (5 µg/ml) at 4 °C overnight. Additionally, mouse IgG (Avicenna Research Institute, Iran) was used at 5 µg/ml for isotype control. The following day, TMA sections were washed three times with TBS, followed by incubation for 1 h with the EnVision™ + /HRP, DualLink Rabbit/Mouse (Dako, Denmark), as secondary antibody, at room temperature. Then, all sections were visualized through chromogen, named 3, 3′-diaminobenzidine substrate (DAB, Dako, Glostrup, Denmark) during 10 min, followed by counterstaining through Mayer's hematoxylin dye (Dako, Glostrup, Denmark) for 2 min. Finally, the slides were dehydrated in graded ethanol, cleared in xylenes, and mounted for assessment by pathologists. Images were photographed by an H550S microscope equipped with a Digital sight DS-LS camera (Nikon, Japan).

Scoring system of IHC slides

Our pathologist (NM) applied a semi-quantitative scoring system to assess the immunostaining of tissue arrays in a blinded manner. The expression level of DCLK1-S was evaluated by three scoring methods, namely the staining intensity, positively stained cells percentage, and the Histochemical score (H-score). The immunostaining intensity of DCLK1-S was scored in a 4-point scale (absent/negative: 0, weak: 1, moderate: 2, or strong: 3) and the percentage was valued based on 1–100% of positively stained cells. The H-score was acquired by multiplying the staining intensity by the percentage, given an overall score of 0–300 for each core.^31–33 The final H-score of each sample was achieved from the average of three replicates. In the present study, for analysis, the H-scores were categorized into two groups (low or high expression) based on the median DCLK1-S expression. The median H-score value was 180 in total tumor samples and intestinal type for membranous, cytoplasmic, and nucleolus localizations. However, in signet ring cell type, the median H-score values (cutoffs) were 140 for membranous staining, 170 for cytoplasmic staining, and 160 for nucleolus staining.

Statistical analysis

Pearson's chi-square (χ2) and Spearman's analyses have been applied to assess the association and correlation between DCLK1-S expression and clinicopathological characteristics, respectively. Kruskal–Wallis and Mann–Whitney U analyses were utilized for pairwise comparison of DCLK1-S expression across the groups. The Kaplan–Meier analysis was applied to generate the DSS and PFS curves. Furthermore, to compare survival outcomes across the groups (low and high expression of DCLK1-S), log-rank analysis was conducted. The classified data were reported by N (%), and quantitative data by mean with standard deviation (SD) and median with quartile (Q1, Q3). The Cox proportional hazards regression model was utilized to conduct both univariate and multivariate analyses. The SPSS statistical software package version 25 (SPSS, Chicago, IL, USA) was applied to statistically analyze all clinical data. P value less than 0.05 (P ≤0.05) was considered as statistically significant.

Results

Generation of anti-DCLK1-S monoclonal antibody and characterization

Preparation of immunogen

Electrophoresis of the BSA-conjugate was applied to assess the efficacy of immunogen-carrier conjugation. The smear pattern of the DCLK1-S-conjugated BSA and the lack of free peptide confirmed peptide-carrier conjugation (Figure 1A).

Figure 1.

Generation of anti-DCLK1-S monoclonal antibody. (A) DCLK1-S peptide was conjugated to a carrier protein and conjugation efficacy was confirmed by SDS-PAGE analysis. BSA (lane 1), BSA-linker (lane 2), and BSA-conjugated peptide (lane 3), Molecular weight markers (lane 4). (B) Titration of anti-DCLK1-S antibody in immunized mice serum by indirect ELISA. Hyperimmunized serum was serially titrated on a peptide-coated plate. (C) Reactivity assessment of purified anti-DCLK1-S monoclonal antibodys (mAbs) by indirect ELISA. Purified antibodies indicated excellent reactivity with immunizing peptide in an indirect ELISA. (D) The purity of the mAbs was confirmed by SDS-PAGE. Molecular weight markers (lane 1), Clone 2H3D5 (lane 2), and Clone 2D5C2 (lane 3).

Immunization of mice and selection of hybridomas

After last immunization of mice, the production of anti-DCLK1-S antibody was confirmed by an optimized indirect ELISA. Both sera of the immunized mice showed a relatively high titer of anti-DCLK1-S antibodies. The mouse showing higher level of antibody titer was selected for fusion (Figure 1B). Following cell fusion, the hybridoma culture supernatants were screened for DCLK1-S peptide-specific antibodies, as described for the serum ELISA above. After four successive cloning, finally, two stable clones, named 2H3D5 and 2D5C2, were obtained and further characterized. The protein G affinity columns were applied to purify both clones.

Reactivity assessment of purified mAbs

Indirect ELISA was used to verify the immunoreactivity of purified mAbs. The findings indicated high immunoreactivity of both mAbs against immunizing peptide (Figure 1C). Both 2H3D5 and 2D5C2 mAbs showed partly similar reactivity reaching to plateau as low an antibody concentration of 5 μg/mL. Additionally, the mAbs’ purity was confirmed using SDS-PAGE (Figure 1D).

Validation of anti-DCLK1-S monoclonal antibody

To evaluate the specificity and reactivity of generated anti-DCLK1-S mAbs, immunohistochemical experiments were performed on GC and CRC tissues as positive control samples and human normal testis, as well as skin and ovary cancer tissues as negative control samples (Figure 2). Additionally, as negative reagent controls, primary antibody pre-adsorbed with immunizing peptide and isotype control were applied (Figure 2). Immunostaining of GC tissues showed that 2H3D5 clone against DCLK1-S have more specificity of reactivity compared to 2D5C2 clone (supplementary Figure 1). Therefore, 2H3D5 mAb was used for the following investigations. Positive control tissues, including GC and CRC, showed strong positive immunoreactivity with anti-DCLK1-S mAb whereas negative reagent control sections represented no staining, demonstrating the specificity of the produced mAb (Figure 2). Additionally, negative control tissues did not show any staining, confirming the particular immunoreactivity of the generated mAb against DCLK1-S (Figure 2). GC cells showed a variable immunostaining pattern of expression of DCLK1-S in their membrane, cytoplasm, and nucleus (Figure 3).

Figure 2.

Characterization of anti-DCLK1-S monoclonal antibody. (A, B) The reactivity of the generated 2H3D5 mAb against DCLK1-S was examined by immunohistochemistry on colorectal (A) and gastric (B) cancer tissues as positive control samples. (C, D) Negative reagent controls, including sections incubated with non-immunized mouse IgG instead of primary antibody (C) and primary antibody was pre incubated with 100 fold molar concentration of immunizing peptide (D). (E, F, G) Immunohistochemical staining of DCLK1-S in negative control tissues, including Normal testis (E), ovarian cancer (F), and squamous cell carcinoma (SCC) tissues (G). (Images were taken at 200x magnification).

Figure 3.

Different DCLK1-S expression patterns in gastric cancer tissues using immunohistochemical staining: (A) cytoplasmic expression (black arrow), (B) membranous and cytoplasmic expression (black arrow), and (C) nuclear expression of DCLK1-S (black arrow) (all images were taken at 400x magnification).

Expression of DCLK1-S in gastric cancer cell lines by flow cytometry

Flow cytometry was applied to assess the surface expression of DCLK1-S in two gastric cancer cell lines, including MKN-45 and AGS. Our results showed the surface DCLK1-S expression in both cell lines with varying levels. AGS cells (58.2%) expressed higher surface DCLK1-S levels compared to MKN-45 (23.6%) (Figure 4).

Figure 4.

Expression profiling of DCLK1-S in human gastric cancer cell lines. Surface DCLK1-S expression in (A) AGS and (B) MKN-45 gastric cancer cell lines was carried out using flow cytometry.

Immunohistochemical evaluation by anti-DCLK1-S monoclonal antibody in gastric cancer tissues

Patient characteristics

To evaluate the clinicopathological importance of DCLK1-S, its expression was investigated in 217 GC tissues. Overall, the population median age was 64 years (range 24–85). The study contained 161 (74.2%) male and 56 (25.8%) female patients that male-to-female ratio was 2.8. In this study, first, total GC samples (N = 217) were analyzed and then, investigation was performed based on categorization to each gastric cancer histological subtype, including intestinal subtype (N = 155) and signet ring cell carcinoma subtype (N = 61). Clinical and pathological data of total samples and subtypes of GC are demonstrated in Table 1.

Table 1.

Patients and tumor clinicopathological characteristics of total gastric cancer samples and different histological subtypes.

		Gastric cancer subtypes
	Total samples (217 samples),	Intestinal type (155 samples),	Signet ring cell carcinoma (61 samples),
Patients and tumor characteristics	N (%)	N (%)	N (%)
Number of samples
Median age, years (Range)	64 (24–85)	66 (27–85)	60 (24–83)
≤Median age	112 (51.6)	82 (52.9)	30 (49.2)
> Median age	104 (47.9)	73 (47.1)	30 (49.2)
Not identified	1 (0.5)	0 (0.0)	1 (1.6)
sex
Male	161 (74.2)	120 (77.4)	41 (67.2)
Female	56 (25.8)	35 (22.6)	20 (32.8)
(Male/female)	2.8	3.4	2
Subtype
Diffuse subtype	61 (28.1)	0 (0.0)	61 (100.0)
Intestinal subtype	155 (71.4)	155 (100.0)	0 (0.0)
Not identified	1 (0.5)	0 (0.0)	0 (0.0)
Mean tumor size (cm) (Range)	4.9 (1–15)	4.8 (1–13)	5.4 (1–15)
≤Mean	104 (47.9)	78 (50.3)	31 (50.8)
> Mean	100 (46.1)	70 (45.2)	24 (39.3)
Not identified	13 (6.0)	7 (4.5)	6 (9.8)
Histological grade
Well differentiated	47 (21.7)	45 (29.0)	2 (3.3)
Moderate differentiated	53 (24.4)	52 (33.5)	0 (0)
Poor differentiated	100 (46.1)	41 (26.5)	59 (96.7)
Not identified	17 (7.8)	17 (11.0)	0 (0.0)
Primary tumor (PT) stage
pT1	49 (22.6)	38 (24.5)	10 (16.4)
pT2	70 (32.3)	53 (34.2)	17 (27.9)
pT3	79 (36.4)	50 (32.3)	29 (47.5)
pT4	8 (3.7)	5 (3.2)	3 (4.9)
Not identified	11 (5.1)	9 (5.8)	2 (3.3)
Tumor extension
Subserosa	97 (44.7)	68 (43.9)	29 (47.5)
Serosa	39 (18.0)	23 (14.8)	16 (26.2)
Others	77 (35.5)	61 (39.4)	15 (24.6)
Not identified	4 (1.8)	3 (1.9)	1 (1.6)
Lamina propria
Yes	213 (98.2)	152 (98.1)	60 (98.4)
No	0 (0.0)	0 (0.0)	0(0)
Not identified	4 (1.8)	3 (1.9)	1 (1.6)
Muscularis mucosa
Yes	211 (97.2)	150 (96.8)	60 (98.4)
No	2 (0.9)	2 (1.3)	0 (0)
Not identified	4 (1.8)	3 (1.9)	1 (1.6)
Submucosa
Yes	197 (90.8)	142 (91.6)	54 (88.5)
No	16 (7.4)	10 (6.5)	6 (9.8)
Not identified	4 (1.8)	3 (1.9)	1 (1.6)
Muscolarispropia
Yes	188 (86.6)	133 (85.8)	54 (88.5)
No	25 (11.5)	19 (12.3)	6 (9.8)
Not identified	4 (1.8)	3 (1.9)	1 (1.6)
Subserosa
Yes	150 (69.1)	102 (65.8)	48 (78.7)
No	63 (29.0)	50 (32.3)	12 (19.7)
Not identified	4 (1.8)	3 (1.9)	1 (1.6)
Serosa
Yes	54 (24.9)	35 (22.6)	19 (31.1)
No	159 (73.3)	117 (75.5)	41 (67.2)
Not identified	4 (1.8)	3 (1.9)	1 (1.6)
Margin
Yes	12 (5.5)	6 (3.9)	6 (9.8)
No	186 (85.7)	135 (87.1)	50 (82.0)
Not identified	19 (8.8)	14 (9.0)	5 (8.2)
Perineural invasion
Present	92 (42.4)	66 (42.6)	26 (42.6)
Absent	93 (42.9)	67 (43.2)	25 (41.0)
Not identified	32 (14.7)	22 (14.2)	10 (16.4)
Lymphovascular invasion
Present	90 (41.5)	62 (40.0)	28 (45.9)
Absent	115 (53.0)	87 (56.1)	27 (44.3)
Not identified	12 (5.5)	6 (3.9)	6 (9.8)
Surgery
Radical	93 (42.9)	61 (39.4)	32 (52.5)
Partial	123 (56.7)	94 (60.6)	28 (45.9)
Not identified	1 (0.5)	0 (0.0)	1 (1.6)
Recurrence
Yes	37 (17.1)	26 (16.8)	11 (18.0)
No	159 (73.3)	114 (73.5)	44 (72.1)
Not identified	21 (9.7)	15 (9.7)	6 (9.8)
Metastasis
Yes	69 (31.8)	46 (29.7)	23 (37.7)
No	127 (58.5)	94 (60.6)	32 (52.5)
Not identified	21 (9.7)	15 (9.7)	6 (9.8)

DCLK1-S protein expression in gastric cancer patients, histological subtypes and adjacent normal tissues samples

The level of DCLK1-S expression in each GC was evaluated based on TMA–immunohistochemistry staining via three scoring methods: staining intensity, percentage of positively stained tumor cells and H-score. The DCLK1-S expression was observed at various intensities in the membranous, cytoplasmic and nuclear compartments in total tumor samples, both histological subtypes of GCs, and adjacent normal tissues (Figure 5 and Table 2). First, total GC samples (N = 217) were analyzed and then, investigation was performed based on categorization to each gastric cancer histological subtype. For all three localizations, GC were classified into either low (≤median of H-scores) or high (> median of H-scores) expression using the median H-score value as the cutoff. The median H-score value was 180 in total tumor samples and intestinal type for membranous, cytoplasmic, and nucleolus localizations. However, in signet ring cell type, the median H-score values (cutoffs) were 140 for membrane, 170 for cytoplasm, and 160 for nucleolus to categorize the samples into either low (≤median of H-scores) or high (> median of H-scores) expression. A low expression of DCLK1-S was detected in 128 (59.0%), 122 (56.2%) and 126 (58.1%) of GC samples in the membrane, cytoplasm and nuclear, whereas higher DCLK1-S expression was found in 89 (41.0%), 95 (43.8%) and 91 (41.9%) samples, respectively (Table 2). The DCLK1-S expression in adjacent normal tissues was almost similar to that seen in GC samples based on staining intensity and H-score, as Kruskal-Wallis and Mann-Whitney U tests demonstrated non-significance difference between the median membranous, cytoplasmic and nuclear DCLK1-S expression of tumural and normal tissues (table 2). However, a statistically significant difference was distinguished in the percentage of positively stained cells for membranous and cytoplasmic expressions of DCLK1-S between gastric cancer and normal samples (for both; P<0.001) (Figure 6 and table 2).

Figure 5.

Immunohistochemical evaluation of DCLK1-S protein expression in gastric cancer samples. DCLK1-S expression in intestinal type: (A, A-1) low expression, (B, B-1) high expression. DCLK1-S expression in signet ring cell carcinoma: (C, C-1) low expression, (D, D-1) high expression. DCLK1-S expression in in adjacent normal tissue: (E, E-1) low expression, (F, F-1) high expression (Images were taken at 100x and 200x magnification).

Figure 6.

Box plot analysis of the DCLK1-S expression in gastric cancer (gc) patients versus adjacent normal tissues based on percentage of positively stained cells. The results of Mann-Whitney U test indicated that there is a statistically significant differences in the median expression DCLK1-S levels among GC and adjacent normal tissues in (A) membrane (P < 0.001) and (B) cytoplasm (P < 0.001).

Table 2.

Membranous, cytoplasmic, and nuclear expression of DCLK1-S (intensity of staining, percentage of positively stained cells, and h-score) in total samples of gastric cancer, histological subtypes and adjacent normal tissue samples.

				Gastric cancer subtypes
	Total Tumorsamples	Adjacent normaltissue samples		Intestinal type	Signet ring cell carcinoma
Scoring system	N (%)	N (%)	P value	N (%)	N (%)	P value
Membrane expression
Intensity of staining			0.545			0.002
Negative (0)	19 (8.8)	3 (10.7)		10 (6.5)	9 (14.8)
Weak (+ 1)	26 (12.0)	0 (0.0)		13 (8.4)	13 (21.3)
Moderate (+ 2)	118 (54.4)	18 (64.3)		89 (57.4)	29 (47.5)
Strong (+ 3)	54 (24.9)	7 (25.0)		43 (27.7)	10 (16.4)
Percentage of positive cells			< 0.001			0.041
< 25%	19 (8.8)	3 (10.7)		10 (6.5)	9 (14.8)
25–50%	10 (4.6)	4 (14.3)		8 (5.2)	2 (3.3)
51–75%	38 (17.5)	11 (39.3)		27 (17.4)	11 (18.0)
> 75%	150 (69.1)	10 (35.7)		110 (71.0)	39 (63.9)
H-score cut off	180	150	0.133	180	140	0.004
Low	128 (59.0)	13 (46.4)		86 (55.5)	31 (50.8)
High	89 (41.0)	15 (53.6)		69 (44.5)	30 (49.2)
Cytoplasmic expression
Intensity of staining			0.468			0.002
Negative (0)	0 (0)	0 (0.0)		0 (0.0)	0 (0)
Weak (+ 1)	38 (17.5)	2 (7.1)		19 (12.3)	19 (31.1)
Moderate (+ 2)	124 (57.1)	19 (67.9)		92 (59.4)	32 (52.5)
Strong (+ 3)	55 (25.3)	7 (25.0)		44 (28.4)	10 (16.4)
Percentage of positive cells			<0.001			0.016
< 25%	0 (0)	0 (0.0)		0 (0.0)	0 (0)
25–50%	16 (7.4)	4 (14.3)		9 (5.8)	7 (11.5)
51–75%	42 (19.4)	14 (50.0)		29 (18.7)	13 (21.3)
> 75%	159 (73.3)	10 (35.7)		117 (75.5)	41 (67.2)
H-score cut off	180	150	0.101	180	170	0.003
Low	122 (56.2)	15 (53.6)		81 (52.3)	31 (50.8)
High	95 (43.8)	13 (46.4)		74 (47.7)	30 (49.2)
Nuclear expression
Intensity of staining			0.392			0.773
Negative (0)	17 (7.8)	2 (7.1)		11 (7.1)	6 (9.8)
Weak (+ 1)	85 (39.2)	11 (39.3)		60 (38.7)	24 (39.3)
Moderate (+ 2)	58 (26.7)	13 (46.4)		44 (28.4)	14 (23.0)
Strong (+ 3)	57 (26.3)	2 (7.1)		40 (25.8)	17 (27.9)
Percentage of positive cells			0.332			0.058
<25%	17 (7.8)	2 (7.1)		11 (7.1)	6 (9.8)
25–50%	7 (3.2)	1 (3.6)		3 (1.9)	4 (6.6)
51–75%	4 (1.8)	1 (3.6)		2 (1.3)	2 (3.3)
> 75%	189 (87.1)	24 (85.7)		139 (89.7)	49 (80.3)
H-score cut off	180	160	0.333	180	160	0.329
Low	126 (58.1)	13 (46.4)		89 (57.4)	33 (54.1)
High	91 (41.9)	15 (53.6)		66 (42.6)	28 (45.9)
Total	217 (100)	28 (100)		155 (100)	61 (100)

P value is based on Mann–Whitney U test. H-score histological score.

Bold means P-value<0.05.

Mann-Whitney U analyses were also applied to compare the DCLK1-S protein expression between the histological subtypes of GCs. A statistically significant difference was identified for the cytoplasmic and membranous expressions of DCLK1-S between GC histological subtypes in all scoring methods (Table 2). The DCLK1-S expression data is summarized in GC patients, histological subtypes, and adjacent normal tissues in Table 2.

Association of DCLK1-S expression with clinicopathological characteristics in the total gastric cancer patients and histological subtypes

Total gastric cancer samples

To assess the association of expression of DCLK1-S with clinicopathological features, Pearson's χ2 analysis was performed. As demonstrated in Table 3, a statistically significant association was identified between the membranous expression of DCLK1-S and histological grade (P = 0.026), PT stage (P = 0.019), muscolaris propia (P = 0.043), subserosa (P = 0.006), and perineural invasion (P = 0.013) in GC patients (in terms of H-score). Furthermore, there were significant association between cytoplasmic DCLK1-S expression and important clinicopathological parameters, including mean tumor size (P = 0.046), subtype (P = 0.046), histological grade (P = 0.017), PT stage (P = 0.007), tumor extension (P = 0.029), subserosa (P = 0.005), and perineural invasion (P = 0.014). Nuclear DCLK1-S expression was only associated with PT stage (P = 0.015), and subserosa (P = 0.025).

Table 3.

The association between DCLK1-S protein expression and clinicopathological characteristics in total gastric cancer patients (p value; Pearson's χ2 test).

Characteristics of tumor	Total samples N (%)	Membranous expression N (%)		P value	Cytoplasmic expression N (%)		P value	Nuclear expression N (%)		P value
		Low	high		Low	high		Low	high
		(≤180)	(> 180)		(≤180)	(> 180)		(≤180)	(> 180)
Median age, years (Range)	64 (24–85)
≤Median age	112 (51.6)	72 (56.3)	40 (45.5)	0.119	69 (56.6)	43 (45.7)	0.115	64 (50.8)	48 (53.3)	0.713
>Median age	104 (47.9)	56 (43.8)	48 (54.5)		53 (43.4)	51 (54.3)		62 (49.2)	42 (46.7)
Not identified	1 (0.5)	0 (0.0)
sex
Male		96 (75.0)	65 (73.0)	0.745	92 (75.4)	69 (72.6)	0.643	94 (74.6)	67 (73.6)	0.871
Female		32 (25.0)	24 (27.0)		30 (24.6)	26 (27.4)		32 (25.4)	24 (26.4)
Mean tumor size (cm) (Range)
≤Mean	4.968 (1–15)	55 (46.6)	49 (57.0)	0.144	50 (44.6)	54 (58.7)	0.046	54 (47.4)	50 (55.6)	0.245
> Mean	104 (47.9)	63 (53.4)	37 (43.0)		62 (55.4)	38 (41.3)		60 (52.6)	40 (44.4)
Not identified	100 (46.1)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
	13 (6.0)
Subtype
Diffuse subtype	61 (28.1)	42 (32.8)	19 (21.6)	0.072	41 (33.6)	20 (21.3)	0.046	36 (28.8)	25 (27.5)	0.831
Intestinal subtype	155 (71.4)	86 (67.2)	69 (78.4)		81 (66.4)	74 (78.7)		89 (71.2)	66 (72.5)
Not identified	1 (0.5)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Histological grade
Well differentiated	47 (21.7)	20 (17.2)	27 (32.1)	0.026	19 (17.0)	28 (31.8)	0.017	21 (18.1)	26 (31.0)	0.065
Moderate differentiated	53 (24.4)	30 (25.9)	23 (27.4)		28 (25.0)	25 (28.4)		30 (25.9)	23 (27.4)
Poor differentiated	100 (46.1)	66 (56.9)	34 (40.5)		65 (58.0)	35 (39.8)		65 (56.0)	35 (41.7)
Not identified	17 (7.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Primary tumor (PT) stage
pT1	49 (22.6)	21 (17.2)	28 (33.3)	0.019	19 (16.2)	30 (33.7)	0.007	22 (18.6)	27 (30.7)	0.015
pT2	70 (32.3)	40 (32.8)	30 (35.7)		38 (32.5)	32 (36.0)		35 (29.7)	35 (39.8)
pT3	79 (36.4)	55 (45.1)	24 (28.6)		54 (46.2)	25 (28.1)		55 (46.6)	24 (27.3)
pT4	8 (3.7)	6 (4.9)	2 (2.4)		6 (5.1)	2 (2.2)		6 (5.1)	2 (2.3)
Not identified	11 (5.1)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Tumor extension
Subserosa	97 (44.7)	63 (50.4)	34 (38.6)	0.06	59 (49.6)	38 (40.4)	0.029	57 (46.3)	40 (44.4)	0.067
Serosa	39 (18.0)	25 (20.0)	14 (15.9)		26 (21.8)	13 (13.8)		28 (22.8)	11 (12.2)
Others	77 (35.5)	37 (29.6)	40 (45.5)		34 (28.6)	43 (45.7)		38 (30.9)	39 (43.3)
Not identified	4 (1.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Lamina propria
Yes	213 (98.2)	125 (100.0)	88 (100.0)	-*	119(100.0)	94 (100.0)	-*	123(100.0)	90(100.0)	-*
No	-	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Not identified	4 (1.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Muscularis mucosa
Yes	211 (97.2)	124 (99.2)	87 (98.9)	0.802	118 (99.2)	93 (98.9)	0.867	122 (99.2)	89 (98.9)	0.824
No	2 (0.9)	1 (0.8)	1 (1.1)		1 (0.8)	1 (1.1)		1 (0.8)	1 (1.1)
Not identified	4 (1.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Submucosa
Yes	197 (90.8)	117 (93.6)	80 (90.9)	0.463	111 (93.3)	86 (91.5)	0.623	116 (94.3)	81 (90.0)	0.239
No	16 (7.4)	8 (6.4)	8 (9.1)		8 (6.7)	8 (8.5)		7 (5.7)	9 (10.0)
Not identified	4 (1.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Muscolarispropia
Yes	188 (86.6)	115 (92.0)	73 (83.0)	0.043	109 (91.6)	79 (84.0)	0.089	113 (91.9)	75 (83.3)	0.056
No	25 (11.5)	10 (8.0)	15 (17.0)		10 (8.4)	15 (16.0)		10 (8.1)	15 (16.7)
Not identified	4 (1.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Subserosa
Yes	150 (69.1)	97 (77.6)	53 (60.2)	0.006	93 (78.2)	57 (60.6)	0.005	94 (76.4)	56 (62.2)	0.025
No	63 (29.0)	28 (22.4)	35 (39.8)		26 (21.8)	37 (39.4)		29 (23.6)	34 (37.8)
Not identified	4 (1.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Serosa
Yes	54 (24.9)	34 (27.2)	20 (22.7)	0.46	34 (28.6)	20 (21.3)	0.224	37 (30.1)	17 (18.9)	0.064
No	159 (73.3)	91 (72.8)	68 (77.3)		85 (71.4)	74 (78.7)		86 (69.9)	73 (81.1)
Not identified	4 (1.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Margin
Yes	12 (5.5)	9 (8.0)	3 (3.5)	0.195	9 (8.3)	3 (3.4)	0.152	8 (7.2)	4 (4.6)	0.445
No	186 (85.7)	104 (92.0)	82 (96.5)		100 (91.7)	86 (96.6)		103 (92.8)	83 (95.4)
Not identified	19 (8.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Perineural invasion
Present	92 (42.4)	62 (57.4)	30 (39.0)	0.013	60 (57.7)	32 (39.5)	0.014	57 (54.8)	35 (43.2)	0.118
Absent	93 (42.9)	46 (42.6)	47 (61.0)		44 (42.3)	49 (60.5)		47 (45.2)	46 (56.8)
Not identified	32 (14.7)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Lymphovascular invasion
Present	90 (41.5)	51 (43.2)	39 (44.8)	0.819	51 (44.7)	39 (42.9)	0.788	54 (46.2)	36 (40.9)	0.454
Absent	115 (53.0)	67 (56.8)	48 (55.2)		63 (55.3)	52 (57.1)		63 (53.8)	52 (59.1)
Not identified	12 (5.5)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Recurrence
Yes	37 (17.1)	21 (18.6)	16 (19.3)	0.902	17 (15.9)	20 (22.5)	0.241	22 (20.2)	15 (17.2)	0.601
No	159 (73.3)	92 (81.4)	67 (80.7)		90 (84.1)	69 (77.5)		87 (79.8)	72 (82.8)
Not identified	21 (9.7)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Metastasis
Yes	69 (31.8)	38 (33.6)	31 (37.3)	0.59	36 (33.6)	33 (37.1)	0.616	42 (38.5)	27 (31.0)	0.275
No	127 (58.5)	75 (66.4)	52 (62.7)		71 (66.4)	56 (62.9)		67 (61.5)	60 (69.0)
Not identified	21 (9.7)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)

No statistical are computed because the parameter is constant.

Bold means P-value<0.05.

The correlation between DCLK1-S expression and clinical features was evaluated using Spearman's correlation test. Spearman's correlation test results showed a significant inverse correlation between membranous, cytoplasmic, and nuclear expression of DCLK1-S and the above mentioned clinicopathological features, indicating that the higher DCLK1-S expression is more often found in GC tissues with well differentiation, lower PT stages and lack of perineural invasion of GC tissues.

Intestinal type

Pearson's χ2 test showed that membranous DCLK1-S expression is statistically associated with PT stage (P = 0.037), muscolarispropia (P = 0.007), subserosa (P = 0.008), and perineural invasion (P = 0.028) (Table 4). In addition, negative correlation was found between these clinicopathological parameters and membranous DCLK1-S expression.

Table 4.

The association between DCLK1-S protein expression and clinicopathological characteristics in intestinal type samples (p value; Pearson's χ2 test).

Characteristics of tumor		Membranous expression N (%)		P value	Cytoplasmic expression N (%)		P value	Nuclear expression		P value
		Membranous expression N (%)			Cytoplasmic expression N (%)			N (%)
	Total samples	Low	high		Low	high		Low	high
	N (%)	(≤180)	(> 180)		(≤180)	(> 180)		(≤180)	(> 180)
Median age, years (Range)	66 (27–85)
≤Median age	82 (52.9)	49 (57.0)	33 (47.8)	0.257	47 (58.0)	35 (47.3)	0.181	44 (49.4)	38 (57.6)	0.316
> Median age	73 (47.1)	37 (43.0)	36 (52.2)		34 (42.0)	39 (52.7)		45 (50.6)	28 (42.4)
sex
Male	120 (77.4)	69 (80.2)	51 (73.9)	0.35	66 (81.5)	54 (73.0)	0.206	71 (79.8)	49 (74.2)	0.415
Female	35 (22.6)	17 (19.8)	18 (26.1)		15 (18.5)	20 (27.0)		18 (20.2)	17 (25.8)
Mean tumor size (cm)(Range)	4.830 (1–13)
≤Mean	78 (50.3)	39 (48.8)	39 (57.4)	0.296	35 (46.7)	43 (58.9)	0.136	43 (52.4)	35 (53.0)	0.943
> Mean	70 (45.2)	41 (51.2)	29 (42.6)		40 (53.3)	30 (41.1)		39 (47.6)	31 (47.0)
Not identified	7 (4.5)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Histological grade
Well differentiated	45 (29.0)	18 (24.3)	27 (42.2)	0.071	18 (25.4)	27 (40.3)	0.137	21 (26.6)	24 (40.7)	0.076
Moderate differentiated	52 (33.5)	30 (40.5)	22 (34.4)		28 (39.4)	24 (35.8)		29 (36.7)	23 (39.0)
Poor differentiated	41 (26.5)	26 (35.1)	15 (23.4)		25 (35.2)	16 (23.9)		29 (36.7)	12 (20.3)
Not identified	17 (11.0)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Primary tumor (PT) stage
pT1	38 (24.5)	15 (18.50	23 (35.4)	0.037	13 (16.9)	25 (36.2)	0.016	18 (22.0)	20 (31.3)	0.179
pT2	53 (34.2)	28 (34.6)	25 (38.5)		27 (35.1)	26 (37.7)		27 (32.9)	26 (40.6)
pT3	50 (32.3)	34 (42.0)	16 (24.6)		33 (42.9)	17 (24.6)		33 (40.2)	17 (26.6)
pT4	5 (3.2)	4 (4.9)	1 (1.5)		4 (5.2)	1 (1.4)		4 (4.9)	1 (1.6)
Not identified	9 (5.8)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Tumor extension
Subserosa	68 (43.9)	42 (50.0	26 (38.2)	0.079	39 (49.4)	29 (39.7)	0.023	40 (46.0)	28 (43.1)	0.116
Serosa	23 (14.8)	15 (17.9)	8 (11.8)		16 (20.3)	7 (9.6)		17 (19.5)	6 (9.2)
Others	61 (39.4)	27 (32.1)	34 (50.0)		24 (30.4)	37 (50.7)		30 (34.5)	31 (47.7)
Not identified	3 (1.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Lamina propria
Yes	152 (98.1)	84(100.0)	68 (100.0)	-*	79 (100.0)	73 (100.0)	-*	87 (100.0)	65 (100.0)	-*
No	0 (0.0)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Not identified	3 (1.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Muscularis mucosa
Yes	150 (96.8)	83 (98.8)	67 (98.5)	0.88	78 (98.7)	72 (98.6)	0.955	86 (98.9)	64 (98.5)	0.835
No	2 (1.3)	1 (1.2)	1 (1.5)		1 (1.3)	1 (1.4)		1 (1.1)	1 (1.5)
Not identified	3 (1.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Submucosa
Yes	142 (91.6)	81 (96.4)	61 (89.7)	0.096	76 (96.2)	66 (90.4)	0.15	82 (94.3)	60 (92.3)	0.632
No	10 (6.5)	3 (3.6)	7 (10.3)		3 (3.8)	7 (9.6)		5 (5.7)	5 (7.7)
Not identified	3 (1.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Muscolarispropia
Yes	133 (85.8)	79 (94.0)	54 (79.4)	0.007	74 (93.7)	59 (80.8)	0.017	79 (90.8)	54 (83.1)	0.154
No	19 (12.3)	5 (6.0)	14 (20.6)		5 (6.3)	14 (19.2)		8 (9.2)	11 (16.9)
Not identified	3 (1.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Subserosa
Yes	102 (65.8)	64 (76.2)	38 (55.9)	0.008	61 (77.2)	41 (56.2)	0.006	64 (73.6)	38 (58.5)	0.058
No	50 (32.3)	20 (23.8)	30 (44.1)		18 (22.8)	32 (43.8)		23 (26.4)	27 (41.5)
Not identified	3 (1.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Serosa
Yes	35 (22.6)	22 (26.2)	13 (19.1)	0.303	22 (27.8)	13 (17.8)	0.142	24 (27.6)	11 (16.9)	0.122
No	117 (75.5)	62 (73.8)	55 (80.9)		57 (72.2)	60 (82.2)		63 (72.4)	54 (83.1)
Not identified	3 (1.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Margin
Yes	6 (3.9)	4 (5.3)	2 (3.1)	0.521	4 (5.5)	2 (2.9)	0.456	4 (5.1)	2 (3.2)	0.568
No	135 (87.1)	72 (94.7)	63 (96.9)		69 (94.5)	66 (97.1)		74 (94.9)	61 (96.8)
Not identified	14 (9.0)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Perineural invasion
Present	66 (42.6)	43 (58.1)	23 (39.0)	0.028	42 (59.2)	24 (38.7)	0.019	41 (54.7)	25 (43.1)	0.186
Absent	67 (43.2)	31 (41.9)	36 (61.0)		29 (40.8)	38 (61.3)		34 (45.3)	33 (56.9)
Not identified	22 (14.2)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Lymphovascular invasion
Present	62 (40.0)	35 (42.7)	27 (40.3)	0.769	35 (44.9)	27 (38.0)	0.397	38 (44.7)	24 (37.5)	0.377
Absent	87 (56.1)	47 (57.3)	40 (59.7)		43 (55.1)	44 (62.0)		47 (55.3)	40 (62.5)
Not identified	6 (3.9)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Surgery
Radical	61 (39.4)	39 (45.3)	22 (31.9)	0.088	37 (45.7)	24 (32.4)	0.092	40 (44.9)	21 (31.8)	0.098
Partial	94 (60.6)	47 (54.7)	47 (68.1)		44 (54.3)	50 (67.6)		49 (55.1)	45 (68.2)
Recurrence
Yes	26 (16.8)	13 (17.1)	13 (20.3)	0.627	10 (14.1)	16 (23.2)	0.166	14 (18.4)	12 (18.8)	0.96
No	114 (73.5)	63 (82.9)	51 (79.7)		61 (85.9)	53 (76.8)		62 (81.6)	52 (81.3)
Not identified	15 (9.7)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Metastasis
Yes	46 (29.7)	25 (32.9)	21 (32.8)	0.992	23 (32.4)	23 (33.3)	0.906	28 (36.8)	18 (28.1)	0.274
No	94 (60.6)	51 (67.1)	43 (67.2)		48 (67.6)	46 (66.7)		48 (63.2)	46 (71.9)
Not identified	15 (9.7)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)

No statistical are computed because the parameter is constant.

Bold means P-value<0.05.

Based on Pearson's χ2, we also observed a statistically significant association between cytoplasmic expression of DCLK1-S and PT stage (P = 0.016), tumor extension (P = 0.023), muscolarispropia (P = 0.017), subserosa (P = 0.006), and perineural invasion (P = 0.019) (Table 4) with an adverse correlation via Spearman's correlation analysis results.

As demonstrated in Table 4, there was no significant association between nuclear DCLK1-S expression and clinical parameters in the intestinal type cases.

Signet ring cell type

In Signet Ring Cell Type, a significant association was only identified between nuclear DCLK1-S expression and tumor size (P = 0.040) as well as PT stage of patients (P = 0.019) (Table 5). Spearman's correlation analysis results indicated an inverse correlation between these clinicopathological parameters and nuclear DCLK1-S expression.

Table 5.

The association between DCLK1-S protein expression and clinicopathological characteristics in signet ring cell carcinoma samples (p value; Pearson's χ2 test).

Characteristics of tumor	Total samples N (%)	Membranous expression N (%)		P value	Cytoplasmic expression N (%)		P value	Nuclear expression N (%)		P value
		Low	high		Low	high		Low	high
		(≤140)	(> 140)		(≤170)	(> 170)		(≤160)	(> 160)
Median age, years (Range)	60 (24–83)
≤Median age	30 (49.2)	20 (64.5)	10 (34.5)	0.061	18 (58.1)	12 (41.4)	0.196	18 (54.5)	12 (44.4)	0.436
> Median age	30 (49.2)	11 (35.5)	19 (65.5)		13 (41.9)	17 (58.6)		15 (45.5)	15 (55.6)
Not identified	1 (1.6)
sex
Male	41 (67.2)	21 (67.7)	20 (66.7)	0.929	21 (67.7)	20 (66.7)	0.929	22 (66.7)	19 (67.9)	0.921
Female	20 (32.8)	10 (32.3)	10 (33.3)		10 (32.3)	10 (33.3)		11 (33.3)	9 (32.1)
Mean tumor size (cm) (Range)	5.41 (1–15)
≤Mean	31 (50.8)	15 (53.6)	16 (59.3)	0.671	13 (48.1)	18 (64.3)	0.228	12 (42.9)	19 (70.4)	0.04
> Mean	24 (39.3)	13 (46.4)	11 (40.7)		14 (51.9)	10 (35.7)		16 (57.1)	8 (29.6)
	6 (9.8)
Histological grade
Well differentiated	2 (3.3)	1 (3.2)	1 (3.3)	0.981	0 (0.0)	2 (6.7)	0.144	0 (0.0)	2 (7.1)	0.118
Moderate differentiated	0 (0)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
Poor differentiated	59 (96.7)	30 (96.8)	29 (96.7)		31 (100.0)	28 (93.3)		33 (100.0)	26 (92.9)
	0 (0)
Primary tumor (PT) stage
pT1	10 (16.4)	2 (6.7)	8 (27.6)	0.19	2 (6.7)	8 (27.6)	0.184	2 (6.3)	8 (29.6)	0.019
pT2	17 (27.9)	10 (33.3)	7 (24.1)		9 (30.0)	8 (27.6)		7 (21.9)	10 (37.0)
pT3	29 (47.5)	16 (53.3)	13 (44.8)		17 (56.7)	12 (41.4)		21 (65.6)	8 (29.6)
pT4	3 (4.9)	2 (6.7)	1 (3.4)		2 (6.7)	1 (3.4)		2 (6.3)	1 (3.7)
	2 (3.3)
Tumor extension
Subserosa	29 (47.5)	16 (53.3)	13 (43.3)	0.328	15 (50.0)	14 (46.7)	0.643	15 (46.9)	14 (50.0)	0.269
Serosa	16 (26.2)	9 (30.0)	7 (23.3)		9 (30.0)	7 (23.3)		11 (34.4)	5 (17.9)
Others	15 (24.6)	5 (16.7)	10 (33.3)		6 (20.0)	9 (30.0)		6 (18.8)	9 (32.1)
	1 (1.6)
Lamina propria
Yes	60 (98.4)	30 (100.0)	30 (100.0)	-*	30 (100.0)	30 (100.0)	-*	32 (100.0)	28 (100.0)	-*
No	0(0)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
	1 (1.6)
Muscularis mucosa
Yes	60 (98.4)	30 (100.0)	30 (100.0)	-*	30 (100.0)	30 (100.0)	-*	32 (100.0)	28 (100.0)	-*
No	0 (0)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)
	1 (1.6)
Submucosa
Yes	54 (88.5)	28 (93.3)	26 (86.7)	0.389	28 (93.3)	26 (86.7)	0.389	30 (93.8)	24 (85.7)	0.301
No	6 (9.8)	2 (6.7)	4 (13.3)		2 (6.7)	4 (13.3)		2 (6.3)	4 (14.3)
	1 (1.6)
Muscolarispropia
Yes	54 (88.5)	28 (93.3)	26 (86.7)	0.389	28 (93.3)	26 (86.7)	0.389	30 (93.8)	24 (85.7)	0.301
No	6 (9.8)	2 (6.7)	4 (13.3)		2 (6.7)	4 (13.3)		2 (6.3)	4 (14.3)
	1 (1.6)
Subserosa
Yes	48 (78.7)	26 (86.7)	22 (73.3)	0.197	26 (86.7)	22 (73.3)	0.197	28 (87.5)	20 (71.4)	0.121
No	12 (19.7)	4 (13.3)	8 (26.7)		4 (13.3)	8 (26.7)		4 (12.5)	8 (28.6)
	1 (1.6)
Serosa
Yes	19 (31.1)	10 (33.3)	9 (30.0)	0.781	11 (36.7)	8 (26.7)	0.405	13 (40.6)	6 (21.4)	0.111
No	41 (67.2)	20 (66.7)	21 (70.0)		19 (63.3)	22 (73.3)		19 (59.4)	22 (78.6)
	1 (1.6)
Margin
Yes	6 (9.8)	3 (10.7)	3 (10.7)	1	3 (10.7)	3 (10.7)	1	3 (10.3)	3 (11.1)	0.926
No	50 (82.0)	25 (89.3)	25 (89.3)		25 (89.3)	25 (89.3)		26 (89.7)	24 (88.9)
	5 (8.2)
Perineural invasion
Present	26 (42.6)	14 (56.0)	12 (46.2)	0.482	14 (56.0)	12 (46.2)	0.482	16 (64.0)	10 (38.5)	0.068
Absent	25 (41.0)	11 (44.4)	14 (53.8)		11 (44.4)	14 (53.8)		9 (36.0)	16 (61.5)
	10 (16.4)
Lymphovascular invasion
Present	28 (45.9)	11 (42.3)	17 (58.6)	0.227	12 (44.4)	16 (57.1)	0.346	14 (50.0)	14 (51.9)	0.891
Absent	27 (44.3)	15 (57.7)	12 (41.4)		15 (55.6)	12 (42.9)		14 (50.0)	13 (48.1)
	6 (9.8)
Surgery
Radical	32 (52.5)	18 (58.1)	14 (48.3)	0.448	18 (58.1)	14 (48.3)	0.448	18 (54.5)	14 (51.9)	0.835
Partial	28 (45.9)	13 (41.9)	15 (51.7)		13 (41.9)	15 (51.7)		15 (45.5)	13 (48.1)
	1 (1.6)
Recurrence
Yes	11 (18.0)	7 (25.9)	4 (14.3)	0.281	7 (25.9)	4 (14.3)	0.281	8 (27.6)	3 (11.5)	0.137
No	44 (72.1)	20 (74.1)	24 (85.7)		20 (74.1)	24 (85.7)		21 (72.4)	23 (88.5)
	6 (9.8)
Metastasis
Yes	23 (37.7)	10 (37.0)	13 (46.4)	0.48	11 (40.7)	12 (42.9)	0.874	12 (41.4)	11 (42.3)	0.944
No	32 (52.5)	17 (63.0)	15 (53.6)		16 (59.3)	16 (57.1)		17 (58.6)	15 (57.7)
	6 (9.8)

No statistical are computed because the parameter is constant.

Bold means P-value<0.05.

However, membranous and cytoplasmic DCLK1-S expressions were not associated with clinicopathological features in signet ring cell type samples.

Prognostic value & survival outcomes of DCLK1-S expression in gastric cancer patients and histological subtypes

Kaplan–Meier curves (with log-rank test) were applied to analyze the association between DCLK1-S expression and DSS as well as PFS on 197 of 217 GC samples for whom follow-up data was accessible. For DSS, the median and mean follow-up duration was 40 (Q1, Q3; 19, 58) and 40 (SD = 25.7) months, respectively, ranging 1 to 114 months in total GC samples. For PFS, the median and mean duration of the follow-up was 36 (Q1, Q3; 15, 55) and 37 (SD = 25.3) months, ranging 1 to 105 months, respectively. During the follow-up period, 78 (35.9%) patients suffered disease-related death. For tissues stained with DCLK1-S, distant metastasis and tumor recurrence occurred in 69 (31.8%) and 37 (17.1%) patients, respectively. The main characteristics of survival outcomes were described in table 6 in the total GC samples and histological subtypes.

Table 6.

The main characteristics of survival outcomes in the total gastric cancer patients and histological subtypes.

Features	Total sample of GCN (%)	Histological subtypes of gastric cancer
Features	Total sample of GCN (%)	Intestinal typeN (%)	Signet ring cell carcinomaN (%)
Number of patients (N)	197	140	56
Not identified	20	15	5
Rang of follow-up duration for DSS or PFS (months)	1–114, 1–105	1–105, 1–105	1–114, 1–90
Mean duration of follow-up time for DSS and PFS (months) (SD)	40 (25.7), 37 (25.3)	40 (25.1), 38 (26.0)	40 (27.4), 36 (23.6)
Median duration of follow-up time for DSS and PFS (months) (Q1, Q3)	40 (19, 58), 36 (15, 55)	41 (20, 58), 37 (13, 57)	38 (17, 58), 36 (15, 53)
Cancer-related death (N %)	78 (35.9)	60 (38.7)	18 (29.5)
Other causes of death (N %)	11 (5.1)	9 (5.8)	2 (3.3)
Distant metastasis during follow-up (N %)	69 (31.8)	46 (29.7)	23 (37.7)
Tumor recurrence during follow-up (N %)	37 (17.1)	26 (16.8)	11 (18.0)
Alive patients without distant metastasis and tumor recurrence (N %)	110 (50.7)	83 (53.5)	26 (42.6)

Total gastric cancer samples

Based on the Kaplan–Meier analyses, there were no significant differences in DSS and PFS between cases with high and low membranous and cytoplasmic expression of DCLK1-S (DSS: P = 0.315, P = 0.433; PFS: P = 0.216, P = 0.377, respectively) (Supplementary Figure 2). Further, the Kaplan–Meier curve of nuclear DCLK1-S expression showed that cases with DCLK1-S overexpression have a longer DSS or PFS in comparison with those with low expression of DCLK1-S, however, without a statistically significant association (P = 0.196, P = 0.104, respectively) (Supplementary Figure 2). Univariate and multivariate analyses were also applied to assess the clinical significance of various parameters that might affect DSS in patients with SOC. In total gastric cancer patients, tumor size (P = 0.002) and tumor extension (P = 0.039) affected the DSS of patients in the univariate analysis. However, only tumor size was independent prognostic factor affecting the DSS based on multivariate analysis. Other variables did not show any significant results using the univariate analysis (Supplementary Table 1).

Intestinal type

In the intestinal type specimens, the Kaplan–Meier survival analyses showed no significant association between membranous, cytoplasmic and nuclear expression of DCLK1-S and DSS or PFS (DSS: P = 0.732, P = 0.978, P = 0.552; PFS: P = 0.545, P = 0.921, P = 0.334, respectively) (Figure 7). Additionally, tumor extension (P = 0.030) in intestinal type samples affected the DSS of patients in the univariate analysis. However, multivariate analysis was not performed because only one variable showed meaningful results (Supplementary Table 2).

Figure 7.

Survival analysis for membranous, cytoplasmic and nuclear DCLK1-S expression in intestinal type of gastric cancer patients. The Kaplan–Meier survival methods indicated no significant differences in disease-specific survival (DSS) and progression-free survival (PFS) between cases with high and low membranous (A, B), cytoplasmic (C, D) and nuclear (E, F) expression of DCLK1-S protein.

Signet ring cell type

In the cases of signet ring cell type, the Kaplan–Meier curves exhibited that cases with high membranous, cytoplasmic and nuclear DCLK1-S expression had a longer DSS or PFS than those with high DCLK1-S expression, however, without a statistically significant association (DSS: P = 0.233, P = 0.179, P = 0.129; PFS: P = 0.188, P = 0.152, P = 0.119, respectively) (Figure 8). Furthermore, the tumor size (P = 0.008) in samples of the signet ring cell carcinoma had an impact on the DSS of patients, as observed in the univariate analysis. However, the multivariate analysis could not be conducted due to the presence of only one variable that yielded significant findings (Supplementary Table 3).

Figure 8.

Survival analysis for membranous, cytoplasmic and nuclear DCLK1-S expression in signet ring cell type of gastric cancer patients. The Kaplan–Meier survival methods indicated differences in disease-specific survival (DSS) and progression-free survival (PFS) between cases with high and low membranous (A, B), cytoplasmic (C, D) and nuclear (E, F) expression of DCLK1-S protein, however, without a statistically significant association.

Discussion

Cancer biomarker discovery could be an essential step to understanding gastric cancer biology. In this regard, mAbs, with their specific pattern of reactivity, may be helpful tools for diagnostic methods and research. Evidence accumulated by our research group and others highlight the significance of DCLK1 expression as an independent predictor of prognosis, therapeutic resistance, and recurrence in diverse tumors.^{8,14,17,34–38} Human DCLK1 has two major isoforms, the long isoform (DCLK1-L) and the short isoform of DCLK1 (DCLK1-S). We have previously studied DCLK1 as a novel CSC marker with different commercial antibodies against DCLK1-L or against both DCLK1-S and -L isoforms.^17,24 Since there were no commercially available antibodies specifically targeting DCLK1-S at that time, we initially developed a polyclonal antibody against DCLK1-S.²² As no particular anti-DCLK1-S mAbs was generated, the current research focused on producing and characterizing mAb against the short isoform of DCLK1. This will provide more opportunities to researchers to acquire further reliable information regarding the DCLK1-S expression pattern.

The present study aimed to investigate expression, localization and clinical significance of DCLK1-S using our produced mAb on a large series of GC tissue samples as a preliminary study to find a potential suitable marker for immunotherapy.^32,39,40 Generated clones, even at low antibody concentrations, recognized mouse DCLK1-S peptides well in ELISA, displaying high binding affinity. However, the immunostaining pattern of the 2H3D5 clone was more specific with GC tissue specimens with no non-specific reactivity. In particular, the specificity of 2H3D5 mAb against DCLK1-S was proven through IHC staining of normal testis tissue, cancerous ovarian and skin tissues, and GC sections in which primary antibody was pre-adsorbed with immunizing peptide or replaced with mouse IgG. No positive staining was identified in these specimens at the protein level. These results indicated that 2H3D5 mAb specifically recognizes DCLK1-S in gastric cancer. Therefore, 2H3D5 mAb was selected to investigate its potential for monitoring cancer cells with DCLK1-S expression.

Regarding the specific functions of DCLK1 isoforms, two recent studies have shown that the DCLK1-S isoform plays a function in the progression of colon cancer and ESCC, however, similar findings for DCLK1-L were not acquired in these studies.^19,21 In addition, our previous study also showed the critical relationship of DCLK1-S with cancer aggressiveness and poor DSS in CRC.²² These results emphasize the significance of evaluating DCLK1 isoforms separately in various cancers where their expression and function may differ. Nonetheless, no studies have investigated DCLK1-S importance in gastric carcinoma. For the first time, the current study investigated the expression pattern, clinicopathological importance, and prognostic implications of DCLK1-S through IHC analysis of TMA cores in a large series of GC tissues. The assessment of stained GC tissue sections demonstrated the heterogeneous membranous, cytoplasmic, and nuclear patterns of DCLK1-S expression in cancer cells with a variety of intensities. This result is in line with the findings of Sarkar et al., who reported that in addition to membranous and cytoplasmic DCLK1-S expression, its expression was observed in the nuclei and mitochondria of neoplastic cells by immune-electron-microscopy (IEM) test in colorectal cell lines, unlike the L-isoform with primary accumulation in the plasma membranes and cytosol of the cells.^19,41,42 Differential localization of the two isoforms may be due to the lack of microtubule-binding domain in the short isoform. Variations in the biological activities of the DCLK1-L/S isoforms may be rooted in differences in subcellular locations of the two isoforms.¹⁹ In addition, according to the positively stained cells percentage, DCLK1-S expression was considerably higher in the membrane and cytoplasm of cancer cells in GC tissues in comparison with adjacent normal tissues, in accordance with other investigations.^21,22

According to the authors’ findings, the lower DCLK1-S expression in the membrane and cytosol of intestinal type samples was associated with increased tumor progression. However, for nuclear expression, such association was not observed. On the other hand, in the signet ring cell type, a significant association was identified between high nuclear expression of DCLK1-S and smaller tumor size as well as lower PT stage of patients. These findings indicate possible different functions of DCLK1-S in different subtypes of gastric carcinoma. Intestinal type and signet ring cell type have different macroscopic appearances and microscopic growth patterns, revealing notable differences in their molecular signaling pathways. Coordinating with this data, our results indicated that tumors with higher DCLK1-S expression, especially in nuclear, had longer DSS or PFS than tumors with low expression of DCLK1-S, although without a statistically significant association in either GC subtype. It should be stressed that this data could be reinforced with an extended follow-up time and more information from a larger cohort.

There is an absence of robust studies on DCLK1-S expression in the majority of cancers, especially gastric carcinoma. Some studies have analyzed the potential of DCLK1-S in colorectal and renal cancers, and their results can be used for the purpose of our comparison. However, there seems to be some controversy in the literature regarding the clinical value of DCLK1-S. In agreement with the findings from the current study, Ge et al. demonstrated a significant inverse relationship between high S-isoform expression and stage, tumor size, and lymph node status in clear cell renal cancer. Additionally, its expression predicted increased survival.⁴² By contrast, a study by our research group indicated the association of cytoplasmic DCLK1-S expression with cancer aggressiveness and worse DSS in CRC cancer.²² In addition, the positive clinical impact of high DCLK1-S expression in the authors’ GC tissue samples is in contrast to the report of a study on ESCC.²¹ There are other studies in accordance with our findings; however, these investigations used commercial antibodies against DCLK1 that identify sequence homology epitopes of both isoforms, DCLK1-S/L, or only the DCLK1-L isoform. For example, a study by our group revealed a significant correlation between high DCLK1 expression and early-stage GC tissues.²⁴ Tao et al. reported that DCLK1 expression has a high tendency to diminish with the progression of pancreatic cancer from early stages to advanced stages.^41,43

As the identification of surface markers is crucial for immunotherapy, our objective in this study, therefor, was to examine the cell membranous expression of DCLK1-S on selected GC cell lines (MKN-45 and AGS) using flow cytometry to determine whether cell lines with different invasive potentials display distinct expression patterns of DCLK1-S, comparable to what we observed in gastric cancer tissue using IHC. MKN-45 and AGS cell lines differ in terms of their invasive abilities.^44,45 Interestingly, we found that MKN-45 cells with higher invasive potential expressed a lower surface DCLK1-S level than AGS cells with less invasiveness. These findings in accordance with our clinicopathological results, indicating that GC specimens with a higher histological grade and PT stage expressed lower DCLK1-S levels. These findings may strengthen the concept that there is a negative correlation between DCLK1-S level and GC invasiveness.

This protective role may be ascribed in part to the fact that DCLK1 interaction is pivotal for ataxia-telangiectasia mutated (ATM) phosphorylation and activation in the epithelial cells to modulate ATM-mediated DNA damage repair (DDR).⁴⁶ Another study provided clear evidence that cells lacking DCLK1 are unable to restore the impaired quality of the tight junctions.⁴⁷ Because of the indispensable role of DCLK1-S in the generation of functional and productive DDR and stable tight junctions, the high expression of DCLK1-S can be used as a potential favorable clinical marker in GC patients. In addition, it can be assumed that the contradictory findings with regards to the clinical impact of DCLK1-S may be the result of differences in the organ specificity, cancer type, DCLK1-S-expressing cell type, antibodies used, criteria employed to determine the cutoff for classification of H-score, follow-up time, and sample sizes used in other studies.^18,21

Conclusion

This research is the first of its kind to examine the generation and characterization of a mAb against the short isoform of DCLK1. We demonstrated that our generated mAb could properly detect DCLK1-S in IHC in both subtypes of gastric cancer. For both histological subtypes of GC, DCLK1-S overexpression was associated with decreased tumor progression and early-stage disease. Moreover, our findings suggest that DCLK1-S expression, specially nuclear expression, might prove to be a favorable prognostic marker in GC cases if the follow-up time is prolonged and the number of cases is increased. Further studies are needed to decipher the contradictory results obtained from our study and other studies on DCLK1-S expression in GC and other cancer types.

Supplemental Material

sj-docx-1-cbm-10.1177_18758592241301691 - Supplemental material for Clinical significance of “S” isoform of DCLK1 in different gastric cancer subtypes using newly produced monoclonal antibody

Supplemental material, sj-docx-1-cbm-10.1177_18758592241301691 for Clinical significance of “S” isoform of DCLK1 in different gastric cancer subtypes using newly produced monoclonal antibody by Mahdieh Razmi, Ali-Ahmad Bayat, Nafiseh Mortazavi, Elham Kalantari, Leili Saeednejad Zanjani, Sima Saki, Roya Ghods and Zahra Madjd in Cancer Biomarkers

Footnotes

ORCID iD

Zahra Madjd

Ethics approval and consent to participate

All procedures performed in this study were approved by the Human Research Ethics Committee and the animal research committee of Iran University of Medical Sciences. Informed consent was obtained from all individual participants included in the study.

Author contributions

INTERPRETATION OR ANALYSIS OF DATA: MR, AB, NM, LS, EK, and SS.

PREPARATION OF THE MANUSCRIPT: MR.

REVISION FOR IMPORTANT INTELLECTUAL CONTENT: ZM and RGh.

SUPERVISION: ZM and RGh.

All authors read and approved the final manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work has received research support from Oncopathology Research Center, Iran University of Medical Sciences (No: 98-4-28-16306).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental material

Supplemental material for this article is available online.

References

Siegel

Miller

Fuchs

, et al. Cancer statistics. CA Cancer J Clin 2022; 72: 7–33.

Rawla

Barsouk

. Epidemiology of gastric cancer: global trends, risk factors and prevention. Gastroenterology Review 2019; 14: 26–38.

Morgan

Arnold

Camargo

, et al. The current and future incidence and mortality of gastric cancer in 185 countries, 2020–40: a population-based modelling study. eClinicalMedicine 2022; 47: 1–10.

Abbas

Faggian

Sintali

, et al. Current and future biomarkers in gastric cancer. Biomed Pharmacother 2018; 103: 1688–1700.

Niclauss

Gutgemann

Dohmen

, et al. Novel Biomarkers of Gastric Adenocarcinoma: current Research and Future Perspectives. Cancers (Basel) 2021; 13: 1–35.

Razmi

Ghods

Vafaei

, et al. Clinical and prognostic significances of cancer stem cell markers in gastric cancer patients: a systematic review and meta-analysis. Cancer Cell International 2021; 21: 1–20.

Nakanishi

Seno

Fukuoka

, et al. Dclk1 distinguishes between tumor and normal stem cells in the intestine. Nat Genet 2013; 45: 98–103.

Wang

Zhao

Lin

, et al. Targeting DCLK1 overcomes 5-fluorouracil resistance in colorectal cancer through inhibiting CCAR1/β-catenin pathway-mediated cancer stemness. Clin Transl Med 2022; 12: e743.

Ding

Yang

, et al. Inhibition of DCLK1 with DCLK1-IN-1 Suppresses Renal Cell Carcinoma Invasion and Stemness and Promotes Cytotoxic T-Cell-Mediated Anti-Tumor Immunity. Cancers (Basel) 2021; 13: 1–19.

10.

Ito

Tanaka

Akiyama

, et al. Dominant expression of DCLK1 in human pancreatic cancer stem cells accelerates tumor invasion and metastasis. PLoS One 2016; 11: e0146564.

11.

Chandrakesan

Yao

, et al. Dclk1, a tumor stem cell marker, regulates pro-survival signaling and self-renewal of intestinal tumor cells. Mol Cancer 2017; 16: 30.

12.

Roy

Ahmed

Stubbs

, et al. DCLK1 isoforms and aberrant Notch signaling in the regulation of human and murine colitis. Cell Death Discovery 2021; 7: 1–14.

13.

Weygant

Yao

, et al. Overexpression of DCLK1-AL increases tumor cell invasion, drug resistance, and KRAS activation and can be targeted to inhibit tumorigenesis in pancreatic cancer. J Oncol 2019; 2019: 6402925.

14.

Weygant

, et al. Cancer Stem Cell Marker DCLK1 Correlates with Tumorigenic Immune Infiltrates in the Colon and Gastric Adenocarcinoma Microenvironments. Cancers (Basel) 2020; 12: 1–19.

15.

Chandrakesan

Panneerselvam

May

, et al. DCLK1-Isoform2 Alternative splice variant promotes pancreatic tumor immunosuppressive M2-macrophage polarization. Mol Cancer Ther 2020; 19: 1539–1549.

16.

Fan

Qian

Dai

. Expression and prognostic significance of doublecortin-like kinase 1 in patients with hepatocellular carcinoma. Oncol Lett 2017; 14: 7529–7537.

17.

Shafiei

Kalantari

Saeednejad Zanjani

, et al. Increased expression of DCLK1, a novel putative CSC maker, is associated with tumor aggressiveness and worse disease-specific survival in patients with bladder carcinomas. Exp Mol Pathol 2019; 108: 164–172.

18.

O'Connell

Sarkar

Luthra

, et al. Epigenetic changes and alternate promoter usage by human colon cancers for expressing DCLK1-isoforms: clinical implications. Sci Rep 2015; 5: 14983.

19.

Sarkar

Popov

O'Connell

, et al. A novel antibody against cancer stem cell biomarker, DCLK1-S, is potentially useful for assessing colon cancer risk after screening colonoscopy. Lab Invest 2017; 97: 1245–1261.

20.

Sarkar

O'Connell

Okugawa

, et al. FOXD3 Regulates CSC marker, DCLK1-S, and invasive potential: prognostic implications in colon cancer. Mol Cancer Res 2017; 15: 1678–1691.

21.

Fan

Huang

, et al. DCLK1-Short Splice variant promotes esophageal squamous cell carcinoma progression via the MAPK/ERK/MMP2 pathway. Mol Cancer Res 2021; 19: 1980–1991.

22.

Kalantari

Ghods

Saeednejad Zanjani

, et al. Cytoplasmic expression of DCLK1-S, a novel DCLK1 isoform, is associated with tumor aggressiveness and worse disease-specific survival in colorectal cancer. Cancer Biomark 2022; 33: 277–289.

23.

Carli

ALE

Afshar-Sterle

Rai

, et al. Cancer stem cell marker DCLK1 reprograms small extracellular vesicles toward migratory phenotype in gastric cancer cells. Proteomics 2021; 21: e2000098.

24.

Kalantari

Asadi Lari

Roudi

, et al. Lgr5High/DCLK1High phenotype is more common in early stage and intestinal subtypes of gastric carcinomas. Cancer Biomark 2017; 20: 563–573.

25.

Mutoh

Sashikawa

Sakamoto

, et al. Cyclooxygenase 2 in gastric carcinoma is expressed in doublecortin- and CaM kinase-like-1-positive tuft cells. Gut Liver 2014; 8: 508–518.

26.

Ghods

Ghahremani

Darzi

, et al. Immunohistochemical characterization of novel murine monoclonal antibodies against human placenta-specific 1. Biotechnol Appl Biochem 2014; 61: 363–369.

27.

Niknam Lakpour

Sadeghi

Ranjbar

, et al. Production and characterization of a new specific monoclonal antibody against A-isoform of SALL4: A novel emerging testicular cancer marker. Andrologia 2022; 54: 1–10.

28.

Saeednejad Zanjani

Razmi

Fattahi

, et al. Overexpression of melanoma-associated antigen A2 has a clinical significance in embryonal carcinoma and is associated with tumor progression. J Cancer Res Clin Oncol 2022; 148: 609–631.

29.

Safaei

Sajed

Saeednejad Zanjani

, et al. Overexpression of cytoplasmic dynamin 2 is associated with worse outcomes in patients with clear cell renal cell carcinoma. Cancer Biomarkers 2022; Preprint: 1–19.

30.

Kalantari

Taheri

Fata

, et al. Significant co-expression of putative cancer stem cell markers, EpCAM and CD166, correlates with tumor stage and invasive behavior in colorectal cancer. World J Surg Oncol 2022; 20: 15.

31.

McCarty

Jr Miller

Cox

, et al. Correlation of biochemical and immunohistochemical methods using monoclonal antireceptor antibodies. Arch Pathol Lab Med 1985; 109: 716–721.

32.

Fedchenko

Reifenrath

. Different approaches for interpretation and reporting of immunohistochemistry analysis results in the bone tissue - a review. Diagn Pathol 2014; 9: 221.

33.

Mohsenzadegan

Razmi

Vafaei

, et al. Co-expression of cancer-testis antigens of MAGE-A6 and MAGE-A11 is associated with tumor aggressiveness in patients with bladder cancer. Sci Rep 2022; 12: 599.

34.

Mirzaei

Tavoosidana

Modarressi

, et al. Upregulation of circulating cancer stem cell marker, DCLK1 but not Lgr5, in chemoradiotherapy-treated colorectal cancer patients. Tumour Biol 2015; 36: 4801–4810.

35.

Mirzaei

Madjd

Kadijani

, et al. Evaluation of circulating cellular DCLK1 protein, as the most promising colorectal cancer stem cell marker, using immunoassay based methods. Cancer Biomark 2016; 17: 301–311.

36.

Sureban

Madhoun

May

, et al. Plasma DCLK1 is a marker of hepatocellular carcinoma (HCC): targeting DCLK1 prevents HCC tumor xenograft growth via a microRNA-dependent mechanism. Oncotarget 2015; 6: 37200–37215.

37.

Gao

Wang

, et al. DCLK1 Is up-regulated and associated with metastasis and prognosis in colorectal cancer. J Cancer Res Clin Oncol 2016; 142: 2131–2140.

38.

Kadletz

Thurnher

Wiebringhaus

, et al. Role of cancer stem-cell marker doublecortin-like kinase 1 in head and neck squamous cell carcinoma. Oral Oncol 2017; 67: 109–118.

39.

Shi

Taylor

. Antigen retrieval immunohistochemistry: review and future prospects in research and diagnosis over two decades. J Histochem Cytochem 2011; 59: 13–32.

40.

Porcario

Hall

Martucci

, et al. Evaluation of two sets of immunohistochemical and western blot confirmatory methods in the detection of typical and atypical BSE cases. BMC Res Notes 2011; 4: 376.

41.

Dai

, et al. Analysis of the clinical significance of DCLK1(+) colorectal cancer using novel monoclonal antibodies against DCLK1. Onco Targets Ther 2018; 11: 5047–5057.

42.

Weygant

, et al. Alternative splice variants of DCLK1 mark cancer stem cells, promote self-renewal and drug-resistance, and can be targeted to inhibit tumorigenesis in kidney cancer. Int J Cancer 2018; 143: 1162–1175.

43.

Johnson

Chandrakesan

, et al. Doublecortin-like kinase 1 is elevated serologically in pancreatic ductal adenocarcinoma and widely expressed on circulating tumor cells. PLoS One 2015; 10: e0118933.

44.

Busuttil

Liu

Di Costanzo

, et al. An orthotopic mouse model of gastric cancer invasion and metastasis. Sci Rep 2018; 8: 825.

45.

Kim

Choi

, et al. Establishment and characterization of six human gastric carcinoma cell lines, including one naturally infected with epstein-barr virus. Cell Oncol (Dordr) 2012; 35: 127–136.

46.

Chandrakesan

May

Weygant

, et al. Intestinal tuft cells regulate the ATM mediated DNA damage response via Dclk1 dependent mechanism for crypt restitution following radiation injury. Sci Rep 2016; 6: 37667.

47.

May

Weygant

, et al. Brief report: dclk1 deletion in tuft cells results in impaired epithelial repair after radiation injury. Stem Cells 2014; 32: 822–827.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.32 MB

0.00 MB