Targeting the PI3K/AKT/mTOR pathway offer a promising therapeutic strategy for cholangiocarcinoma patients with high doublecortin-like kinase 1 expression

The cholangiocarcinoma cell lines (RBE and HCCC 9810) and the human embryonic kidney cell line 293 T were acquired from the American Type Culture Collection (ATCC). RBE and HCCC 9810 were cultured in RPMI-1640 medium (Sigma, USA), while 293 T cells were cultured in DMEM medium (Sigma, USA). Both mediums were supplemented with 10% fetal bovine serum (FBS, Ausbian, Australia) and 1% penicillin/streptomycin (Solarbio, Beijing, China). Incubation of all cells took place in a humidified incubator at 37℃ with 5% CO₂.

For the construction of a DCLK1-knockout cell line, we employed the clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 technique. The sgRNA sequences utilized were listed as follows:

Oligo1-DCLK1- 5′-CACCGGAGTAGAGAGCTGACTACCA-3′

Oligo2-DCLK1- 5′-AAACTGGTAGTCAGCTCTCTACTCC-3′

These two sgRNA oligomers were cloned into the predigested lentiCRISPR V2 plasmid. Following transfection into 293 T cells alone with psPAX₂ and pMD₂G (packaging plasmids) using DNA transfection reagent (Neofect (Beijing) Biotech Co., Ltd), the lentivirus was harvested from the cell culture supernatant after 72 h and filtrated using a 0.45 μm filter. RBE cells underwent infection with lentivirus and polybrene, and subsequent stable transfection cells were screened using 2 μg/ml of puromycin.

To obtain a DCLK1-overexpressing cell line, the human DCLK1 overexpression lentivirus plasmid (pCDH-DCLK1-GFP) was purchased from Beijing AUGCT (China). The DCLK1-OE lentivirus transfection and infection procedures mirrored those described above. Ultimately, the target cells were screened using fluorescence microscopy.

To confirm the regulatory function of DCLK1 expression in CCA, sequence-modified DCLK1 was stably transfected using a linearized vector, thereby restoring DCLK1 expression in DCLK1-KO cell lines. Assess whether reinstating DCLK1 expression can counteract the tumor-suppressive effect induced by DCLK1 knockout. The verification of DCLK1 expression rescue was conducted through Western blot. Furthermore, we used two different DCLK1 inhibitors (LRRK2-IN-1 and DCLK1-IN-1) with a final concentration of 2.5 μM to inhibit the expression of DCLK1 in HCCC 9810 DCLK1-OE cell lines. The inhibition effect of these two compounds on DCLK1 expression was validated by Western blot.

To investigate the migratory and invasive capacities of cholangiocarcinoma cells, transwell assays were conducted using 8.0 μm pore membrane chambers placed in a 24-well plate. In order to perform the migration assay, we prepared cell suspensions using serum-free medium and filled the top chambers with 200μL cell suspensions (containing 2 × 10⁴ cells). In the invasion assay, Matrigel (BD Bioscience, USA), diluted at a 1:8 ratio with a serum-free media, was applied to the upper chambers before cell inoculation and allowed to incubate for 2 h at 37℃. In each chamber, 600μL of cell culture medium containing 10% FBS served as the chemoattractant at the bottom. Following a 24-h incubation period, the migrated and invaded cells were fixed for 10 min using 4% paraformaldehyde and then stained for 20–30 min with 0.1% crystal violet, and observed under the microscope. 4–5 fields/well were selected and analyzed with the Image J software.

RTCA, a cellular biosensor, operates on the principle of electrical impedance and utilizes electronic cell sensor arrays. This innovative cell detection technology enables real-time monitoring of cell dynamics. It provides insights into changes in cell survival, growth, and other pertinent information. E-plates were utilized, with 5 × 10³ cells/well seeded and incubated in an RTCA detector at 37 ℃. The automatic cell proliferation recording of each group occurred at 15-min intervals.

We planted 300 Cholangiocarcinoma cells into each well of a 6-well plate, and cells were cultured for 2-week intervals at 37 ℃ with 5% CO₂. After using 4% paraformaldehyde, the colonies were fixed for half-hour, and stained for 10 min with 0.1% crystal violet. The dyed colonies were observed and recorded under the light microscope.

Total cellular protein was extracted using RIPA lysis buffer (Beijing Solarbio Science and Technology Co., Ltd.), supplemented with 1% PMSF and 1% phosphatase inhibitor. After that, the BCA Protein Assay Reagent (Thermo Fischer, USA) was utilized to determine the concentrations of whole-cell protein. After boiling at 95 ℃ for 10 min, the fully denatured protein samples (50 μg) were loaded to the sample pores, separated using 10% SDS-PAGE gels, and subsequently transferred onto PVDF membranes (Millipore Corporation, USA). To block the membranes, 8% skim milk was applied for 1 h and incubated with primary antibodies at 4℃ overnight.

The primary antibodies used included: anti-DCLK1 was purchased from Abcam; anti-E-cad, anti-N-cad, anti-ZO-1, anti-Snail, anti-Slug, anti-Vimentin, anti-ZEB1 and anti-β-actin were obtained from Proteintech; anti-AKT, anti-Phospho-AKT, anti-PI3K, anti-Phospho-PI3K, anti-mTOR and anti-Phospho-mTOR were obtained from Cell Signaling Technology. Following three washes with 1 × TBST, membranes underwent a one-hour incubation with secondary antibodies and subsequent measurement using ECL reagents (E1080, LABLEAD). The specific information of antibodies, including dilution ratio, brand, and product numbers were attached to the supplement (Supplementary Table 1).

Total RNA extraction employed TRIzol reagent (Invitrogen, USA), followed by reverse transcription into cDNA using Hifair^® III 1st Strand cDNA Synthesis SuperMix for qPCR (gDNA digester plus) (11141ES60, Yeasen, China). Real-time PCR was conducted using SYBR Green Premix (Invitrogen, USA) on the 7500 Sequence Detection System (Applied Biosystems, China). GAPDH served as a reference gene, and the relative mRNA expression of target genes was calculated using the 2^−ΔΔC(t) method. Supplementary Table 2 contains a list of specific primer sequences used in this article.

To assess the expression of DCLK1 in cholangiocarcinoma and normal tissues, a CCA tissue microarray (including 79 CCA tissue) was purchased from Bioaitech. In addition, samples of primary tumor tissue from 49 CCA patients were obtained from Beijing Chao-Yang Hospital. All tissues included in this research obtained ethical approval by the ethics committee of Beijing Chaoyang Hospital, Capital Medical University. These tissue sections were roasted at 65℃ for two hours and dewaxed by soaking in xylene solution. After gradient hydration, antigen retrieval, endogenous peroxidase elimination, and goat serum blocking, all tissue sections were incubated with anti-DCLK1 (1:500, Abcam, ab31704) at 4 ℃ overnight. The secondary antibody was incubated after the tissue sections were returned to room temperature. DAB chromogen and hematoxylin were used to stain the slices and to assess the expression levels of DCLK1. After dehydration and sealing, all slides were observed and recorded under a microscope, and the IHC score of the staining density and proportion of DCLK1-positive cells in each slide was calculated using Image J software. Using the IHC score, CCA patients were categorized into groups based on DCLK1 expression levels, distinguishing between those with high and low DCLK1 expression.

To intuitively demonstrate the correlation between DCLK1 expression and the epithelial mesenchymal transformation (EMT) process, multiplex immunohistochemistry (mIHC) analysis was performed on the tissue sections mentioned above, using the Opal™ 4-colour Manual IHC kit (PerkinElmer, USA). All slides were incubated with anti-DCLK1 (ab31704, 1:500) opal 520, anti-E-cadherin (20874-1-AP, 1:3000) opal 690, and anti-Vimentin (10366–1-AP, 1:3000) opal 570, and the nuclear stain was conducted with DAPI. Multiple stained sections were scanned by Beijing Bodu Hengyi Technology company under the same exposure time conditions, and demonstrated through NDP. View 2 software. To further explore the expression of each marker, the Image J software was used to separately assess the mean fluorescence intensity of each channel.

CAA cells were seeded on a cover slide inside a 24-well plate until the cell density reached 70%. After fixation with 90% ethanol, 0.5% Triton X-100 was used to break the cell membrane. Primary antibodies were added separately and covered each slide at 4 ℃ overnight: anti-DCLK1 (ab31704, 1:150), anti-E-cadherin (20874–1-AP, 1:150) and anti-Vimentin (10366–1-AP, 1:150). The next day, Rhodamine-Conjugated Anti-Rabbit IgG (1:150) was incubated for one hour. All cover slides were sealed with a capping agent containing DAPI dye and observed under the fluorescence microscope.

Female BALB/c nude mice (5–6 weeks old), acquired from Charles River Laboratory (Beijing, China), were randomly allocated into two groups (6 mice per group) for the establishment of the Subcutaneous tumor model. Each mouse received a subcutaneous injection of 100μL RBE and RBE DCLK1-KO cells (containing 2 × 10⁶ cells). The tumor sizes were measured every 4–5 days and calculated via the formula:Volume = (Length × Width²)/2. Subcutaneous lumps were obtained and paraffin was embedded and fixed for mIHC analysis. All animal experiments conducted in this research received approval from the Institutional Animal Care and Use committee of Capital Medical University.

The level 3 mRNA expression data of DCLK1 (RNAseq (polyA + IlluminaHiSeq)) of 45 CCA patients in the TCGA database (Cohort: TCGA Bile Duct Cancer (CHOL)) were collected from UCSC Xena (http://xena.ucsc.edu/) in order to examine the differences in DCLK1 expression between the tumor and nearby non-tumor tissues. In addition, using the R package “DESeq2”, RNA-seq analysis was carried out to determine the differentially expressed genes (DEGs) across CCA cell lines: RBE vs. RBE DCLK1-KO, HCCC 9810 vs. HCCC 9810 DCLK1-OE. The principles for DEG screening were as follows: False Discovery Rate (FDR) < 0.05 and fold change of ≥ 2, or ≤ 2. The gene functions of DEGs were revealed by Gene ontology (GO) analysis, and the “clusterProfiler” R package was used to target the DEGs enrichment pathway in Kyoto Encyclopedia of Genes and Genomes (KEGG) study.

To assess DCLK1 expression in both tumor and non-tumor tissue, IHC detection was conducted on a cholangiocarcinoma tissue microarray (containing 79 tumors) and surgical tissues from cholecystitis patients (n = 10). As depicted in Fig. 1A, DCLK1 expression exhibited a notable upregulation in CCA samples compared to non-tumor tissues. Additionally, our findings from the TCGA-CCA cohort indicated a significant elevation in DCLK1 mRNA expression in CCA tissue (5.39 ± 1.40) compared to para-cancer tissue (3.52 ± 1.18) (t = 3.63, P < 0.001, Fig. 1 B), aligning with the results observed in our IHC analysis. In order to further examine the potential correlation between DCLK1 expression levels and patient survival, we examined a cohort comprising 49 CCA patients from Beijing Chao-Yang Hospital, with survival follow-ups. The specific clinical and tumor pathology information was provided in Supplementary Table 3. Paraffin sections from tumor tissues were acquired, and DCLK1 expression was assessed through IHC. Image J software was utilized to calculate the IHC score for each slide, defined as the number of positive cells in a 40 × field of view. Using the median IHC score (208.5) of all paraffin sections, we categorized the 49 CCA patients into two groups: high and low DCLK1 expression. The survival analysis, conducted using the Kaplan–Meier (K–M) curve, revealed that patients with elevated DCLK1 expression experienced a reduced overall survival (OS) compared to those with lower levels of DCLK1 expression (P = 0.007, Fig. 1C). These findings underscore the association between DCLK1 expression and cholangiocarcinoma progression, indicating that higher expression levels correlate with a poorer prognosis.

Bild vergrößern

To delve deeper into the distinct biological role of DCLK1 in driving CCA progression, a series of in vitro experiments were conducted. Initially, we assessed the baseline expression of DCLK1 in two cholangiocarcinoma cell lines, RBE and HCCC9810. Western blot results revealed a significantly higher expression level of DCLK1 in RBE compared to HCCC9810 (Fig. 2A). Subsequently, DCLK1 was knocked out in RBE through the CRISPR/Cas9 technique, while DCLK1 was overexpressed in HCCC9810 via lentivirus transfection. Successful establishment of stable RBE DCLK1-KO and HCCC9810 DCLK1-OE cell lines was confirmed using Western blot analysis (Fig. 2B, C). RTCA results showed that depletion of DCLK1 expression inhibited the proliferation of RBE cells. The cell proliferation index of RBE was 6.75 ± 0.24, compared to 4.98 ± 0.45 in RBE DCLK1-KO (t = 5.10, P = 0.015, Fig. 2D upper panel). In contrast, overexpression of DCLK1 enhanced the proliferation of HCCC9810 cells. The cell proliferation index of HCCC9810 was 6.28 ± 0.19, compared to 7.13 ± 0.23 in HCCC9810 DCLK1-OE (t = 3.69, P = 0.034, Fig. 2D lower panel). The same conclusion was also confirmed in colony formation assays, where cell lines with higher expression of DCLK1 (RBE and HCCC9810) performed more monoclonality. The clone counts of RBE was 14.00 ± 1.63, compared to 3.33 ± 0.47 in RBE DCLK1-KO (t = 8.88, P < 0.001). The clone counts of HCCC9810 was 12.00 ± 1.63, compared to 19.33 ± 2.87 in HCCC9810 DCLK1-OE (t = 3.14, P = 0.035, Fig. 2E). Moreover, transwell assays were employed to examine the impact of DCLK1 expression on the migration and invasion capacity of CCA cells. We found that the number of migrated cells with DCLK1 overexpressing HCCC9810 group was 431.67 ± 6.13, which was significantly higher than that of HCCC9810 control cells (66.67 ± 1.25, t = 82.54, P < 0.0001). Similarly, the migration number of RBE DCLK1-KO cells was 47.00 ± 3.74, which was significantly lower than that of the control group (255.00 ± 8.83, t = 30.67, P < 0.0001). The same trend was found in the invasion experiment. These outcomes indicated that overexpression of DCLK1 enhanced both the migration and invasion, and conversely, knockout of DCLK1 decreased both (Fig. 2F, G).

Bild vergrößern

Prior studies have highlighted DCLK1's impact in propelling the epithelial-mesenchymal transition (EMT) across a spectrum of malignancies. Nonetheless, the extent of its regulatory influence in cholangiocarcinoma is yet to be fully understood. To explore this aspect, we performed RT-PCR and Western blot of EMT-related genes in RBE vs. RBE DCLK1-KO and HCCC9810 vs. HCCC9810 DCLK1-OE. The results show that in RBE DCLK1-KO cells, mesenchymal markers including Vimentin, N-cadherin, and Snail were significantly down-regulated, while the epithelial marker (E-cadherin) was up-regulated (Fig. 2H). Conversely, DCLK1 overexpression was linked to EMT activation by inhibiting E-cadherin expression and enhancing ZEB1, N-cadherin, and Vimentin (Fig. 2I). Immunofluorescence (IF) staining of cells on glass slides further illustrated the negative correlation of DCLK1 expression with E-cadherin and its positive correlation with Vimentin (Fig. 2J). These findings underscore the pivotal role of DCLK1 in the EMT process, establishing it as a significant risk factor for the progression of CCA.

Subsequently, to further investigate the function of DCLK1 in augmenting tumor growth in vivo, we employed BALB/c nude mice to establish a subcutaneous transplant tumor model. Cells from both RBE CTRL and RBE DCLK1-KO groups, in the logarithmic growth phase, were subcutaneously injected into the abdomen of mice (6 mice per group). Notably, the RBE CTRL group displayed a significantly larger tumor size compared to the RBE DCLK1-KO group, indicating a substantial impairment in subcutaneous tumor development upon DCLK1 deactivation (Fig. 3A). Then, we obtained the subcutaneous tumor tissues for mIHC about the expression of E-cadherin, and Vimentin. The results demonstrated that tumor tissues from the RBE CTRL group exhibited lower E-cadherin expression and higher Vimentin expression compared with the RBE DCLK1-KO group (Fig. 3B). Therefore, we provided highly convincing evidence that DCLK1 promote the progression of CCA cells in vivo.

Bild vergrößern

In addition, we performed mIHC in samples from 49 CCA patients to assess the protein expression of DCLK1, E-cadherin, and Vimentin. Based on the median DCLK1 positive cell numbers (median IHC score = 208.5), patients were categorized into two groups: the DCLK1 high expression group and the DCLK1 low expression group. It was observed that individuals with elevated DCLK1 expression also exhibited heightened levels of Vimentin and reduced expression of E-cadherin (Fig. 3C). These outcomes provided compelling evidence supporting the role of DCLK1 in triggering the EMT process.

To unravel the specific biological mechanism through which DCLK1 promotes CCA progression, RNA transcriptome sequencing was conducted on HCCC9810, and HCCC9810 DCLK1-OE cells. The comparative analysis between HCCC9810 and HCCC9810 DCLK1-OE revealed 493 DEGs in HCCC9810 DCLK1-OE cells, including 38 down-regulated genes and 455 up-regulated genes (Fig. 4A and B). Notably, KEGG analysis highlighted the predominant enrichment of DEGs in the PI3K/AKT pathway (Fig. 4C). Correspondingly, we delved further into the relationship between DCLK1 and the PI3K/AKT/mTOR pathway utilizing the GeneMANIA database. The gene network apparently indicated that DCLK1 and PI3K/AKT/mTOR signaling pathway had network relationships based on the physical interactions, pathway, and genetic interactions (Fig. 4D). Therefore, based on the results of RNA-Seq analysis, we further explored whether DCLK1 promotes the development of cholangiocarcinoma by activating the PI3K/AKT/mTOR signaling pathway.

Bild vergrößern

Firstly, we conducted a Western blot to investigate the expression of PI3K/AKT/mTOR pathway-related markers in CCA cells. We found that the knockout of DCLK1 significantly reduced the protein levels of phosphor-mTOR, phosphor-PI3K, and phosphor-AKT. In contrast, overexpression of DCLK1 stimulated the PI3K/AKT/mTOR pathway by upregulating p-mTOR, p-PI3K, and p-AKT (Fig. 4E). In this way, we proposed that DCLK1 was strongly correlated with the activation of the PI3K/AKT/mTOR pathway. Subsequently, DCLK1 expression was rescued in RBE DCLK1-KO cells and then we extracted the total cell protein of RBE-DCLK1-OE and RBE DCLK1-KO-RES for Western blot to verify the expression of DCLK1 and other markers of PI3K/AKT/mTOR pathway. The results revealed that DCLK1 expression was successfully rescued in RBE DCLK1-KO cells. Besides, rescuing of DCLK1 expression could re-activate the PI3K/AKT pathway, including the upregulation of p-mTOR, p-PI3K, and p-AKT (Fig. 4 left panel). Subsequently, we used two distinct DCLK1 inhibitors, LRRK-2 (2.5 μM) and DCLK1-IN-1 (2.5 μM), to suppress DCLK1 expression in HCCC9810 DCLK1-OE cells. The activation of the PI3K/AKT/mTOR pathway in HCCC 9810 DCLK1-OE cells returned to baseline following treatment with DCLK1 inhibitors (Fig. 4F right panel). These findings indicate that elevated levels of DCLK1 in cholangiocarcinoma markedly stimulate the PI3K/AKT/mTOR signaling pathway.

To further verify whether DCLK1 promotes the malignant phenotype of CCA through stimulating the PI3K/AKT/mTOR pathway, we employed LY294002, a specific inhibitor of the PI3K/AKT/mTOR pathway, to evaluate the special impact of DCLK1 on CCA progression regulation. RTCA was used to record the proliferation curves of HCCC9810, HCCC9810 DCLK1-OE, HCCC9810 + LY294002, and HCCC9810 DCLK1-OE + LY294002. The outcomes revealed a significantly higher proliferation capacity in DCLK1-OE cells (5.80 ± 0.07) compared to control cells (5.06 ± 0.10, t = 9.86, P < 0.001). However, treatment with LY294002 (20 μM) led to a sharp decrease in the proliferation index in both DCLK1 overexpression cells (3.85 ± 0.34, t = 9.69, P < 0.0001) and control cells (2.35 ± 0.40) (Fig. 5 A). The clone counts of DCLK1 overexpressing HCCC9810 group was 61.33 ± 5.79, which was significantly higher than that of HCCC9810 control cells (23.33 ± 3.86, t = 7.72, P = 0.002) and DCLK1 overexpressing cells treated with LY294002 (5.33 ± 1.25, t = 13.37, P < 0.001).The colony formation assay also confirmed that the PI3K/AKT/mTOR pathway inhibitor could significantly reduce the proliferation of CCA cells mediated by high expression of DCLK1 (Fig. 5B). Moreover, transwell experiments were performed to compare the invasion and migration ability of each group, including HCCC9810, HCCC9810 DCLK1-OE, HCCC9810 + LY294002 and HCCC9810 DCLK1-OE + LY294002. The results demonstrated that the number of invaded and migrated cells in the DCLK1-overexpressed group was significantly higher than that in the control group and the LY294002 treatment group (P < 0.0001, Fig. 5C). Apparently, the PI3K/AKT/mTOR pathway inhibitor attenuated the promotion of invasion and migration by high DCLK1 expression in CCA cells. Additionally, protein levels of EMT-related markers in these four groups were assessed through Western blot and IF. As depicted in Fig. 5D, E, inhibition of the PI3K/AKT/mTOR pathway reversed the EMT process, manifesting as increased expression of E-cadherin and ZO1, and decreased expression of N-cadherin, Vimentin, Snail, and Slug. Together, these findings suggested that DCLK1 promotes CCA progression via the PI3K/AKT/mTOR pathway.

Bild vergrößern