CCR7 high expression leads to cetuximab resistance by cross-talking with EGFR pathway in PI3K/AKT signals in colorectal cancer

Patients and follow-up

One hundred and ninety patients with colorectal cancer (CRC) were enrolled in our study, who had received surgical treatment at Zhongshan Hospital (Shanghai, China) from June 2008 to September 2015. Tumor tissue and paired normal tissue samples that had been formalin-fixed and paraffin-embedded and clinicopathologic data were retrieved from our prospectively constructed CRC database. The tumor stage was determined according to the seventh edition of the International Union Against Cancer (UICC)/American Joint Committee on Cancer (AJCC) TNM classification. The patients were followed every 3 months in the first year after surgery. The overall survival (OS) was calculated from the day of surgery to the date of death due to CRC or last follow-up. Disease-free survival (DFS) was measured from the date of surgery to the date of documented recurrence or metastasis.

Samples for evaluating the therapeutic effect of Cetuximab were from 14 patients with wild-type KRAS synchronous non-resectable liver-limited metastases from colorectal cancer, which had demonstrated EGFR expression. After resection of their primary tumors, the patients received chemotherapy (mFOLFOX6 [modified fluorouracil, leucovorin, and oxaliplatin]) plus Cetuximab. After 8 weeks, tumor response was assessed by a multidisciplinary team with the use of contrast-enhanced computed tomography or magnetic resonance imaging (and optional positron emission tomography scan), according to RECIST criteria [21]. Ethical approval was obtained from the Clinical Research Ethics Committee of Zhongshan Hospital of Fudan University (Shanghai, China). Signed informed consent was obtained from all of the patients.

Cell lines

The human CRC cell lines HCT116, Caco-2, DLD-1, SW620, SW480, HT-29 and LoVo were all from the Institute of Biochemistry and Cell Biology (SIBS, CAS) and maintained in a 5% CO₂-humidified atmosphere at 37°C. HCT116, DLD-1, SW620, SW480 and LoVo cells were maintained in RPMI-1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, USA). Caco-2 cells were maintained in MEM medium (Gibco, USA) supplemented with 10% FBS. HT-29 cells were maintained in McCoy’s 5A medium (Gibco, USA) supplemented with 10% FBS.

Drugs

Cetuximab (Erbitux, C225), an anti-EGFR human-mouse commercial mAb, was kindly provided by Merck Serono (USA). LY29004, a selective PI3K/AKT signaling pathway inhibitor, was purchased from Cell Signaling Technology (CST, USA). LY29004 was dissolved in sterile dimethyl sulfoxide, and a 10 mM working solution was prepared. All solutions were stored in aliquots at -20°C and diluted to the working solution before use.

Bioinformatics analysis

The network composed of CCR7 and its downstream genes was constructed using Toppgene and Exploratory Gene Association Networks (EGAN) software [22]. Pathway enrichment was based on the Kyoto Encyclopedia of Genes and Genomes pathways(KEGG, http://www.genome.jp/kegg/) database using DAVID Bioinformatics Resources 6.7 (http://david.abcc.ncifcrf.gov/home.jsp) [23].

Western blotting

Isolated proteins were analyzed by immunoblotting with primary antibodies against CCR7 (1:2000), p-EGFR, AKT, p-AKT (Ser473) (1:1000) (all from CST, USA); EGFR (1:1000) (Pierce, USA); and GAPDH (1:10000; KangCheng Bio-tech Inc., China); E-cadherin (1:1000), Snail (1:500), β-catenin (1:4000), and Twist (1:100) (all from Abcam, USA); N-cadherin (1:4000) (Epitomics, USA).

Tissue microarrays and immunohistochemistry

Tissue microarrays were constructed as previously described [24], Immunohistocytochemistry of the chip was performed using a two-step protocol (Novolink Polymer Detection System, Novocastra, Newcastle, UK). Briefly, sections were deparaffinized with xylene and rehydrated with graded ethanol solutions. After the endogenous peroxide activity in the slides was blocked by incubation in 0.3% H₂O₂, retrieval was performed by placing the sections in boiling citrate or ethylenediamine tetraacetic acid (EDTA) buffer for 10 min, and nonspecific binding sites were blocked with 5% goat serum (Beyotime, China) in PBS. The slides were incubated with a primary antibody against CCR7 (1:200), horseradish peroxidase (HRP)-conjugated secondary antibodies (1:400) (KPL, MD, USA), and 3’-diaminobenzidine (DAB) solution (Sigma-Aldrich, MO, USA). Negative control slides without primary antibodies were included in all assays.

Evaluation of immunohistochemical staining

All specimens were counterstained with hematoxylin and examined under light microscopy (TCS-SP5; Leica Camera AG; Germany). Staining intensity was referred as the standard of results judgment. The intensity of peroxidase deposits, ranging from light beige to dark brown, was assessed visually in the tumor cell membrane and scored as 0 (negative), 1 (weak), 2 (moderate), or 3 (strong). Two pathologists in gastroenterology who are blinded to the clinical data pathologically assessed all specimens independently.

Cell cycle analysis

PI staining was used to analyze DNA content. Cells were plated in 6-well culture plate at concentrations determined to yield 40-50% confluence within 24 h. Cells were pre-treated with SLC (200 ng/ml) or Ne-CCR7 (4 μg/ml) for 12 h, and then treated with Cetuximab 48 h. Cells were collected and resuspended in PBS, fixed with 70% ethanol, labeled with PI (0.05 mg/ml), incubated at room temperature in the dark for 30 min. DNA content was then analyzed by BD flow cytometry.

Immunofluorescent staining

For immunofluorescent staining, cells were seeded in 24-well plates with coverslips. After the cells attached, the medium was changed to medium with or without SLC (Peprotech, USA), and the cells were cultured for 8 h for HCT116 cells and 12 h for Caco-2 cells. Then, the cells were fixed, permeabilized, blocked and incubated with primary antibodies overnight at 4°C. Cell monolayers were then incubated with fluorescein isothiocyanate 594-labeled secondary antibody (1:1000) (KPL, USA) for 2 h and stained with 1 μg/μL 4’,6-diamidino-2-phenylindole (DAPI, Sigma, USA). Slides were mounted, covered with coverslips, and analyzed using a laser-scanning confocal microscope (magnification 400 ×, TCS-SP5, Leica Camera AG, Germany).

Cell viability assay

Cell proliferation was quantitatively evaluated using the Cell Count kit-8 (CCK-8, Dojindo, Japan). Cells were plated in 96-well plates and incubated for 72 h. CCK-8 solution (10 μL) was added to each well. The absorbance at 450 nm (OD450) was determined using a microplate reader (Model 680; Bio-Rad, CA, USA). The inhibition rate was calculated using the formula: (OD of control group-OD of test group)/ (OD of control group-OD of blank group) × 100%.

Migration assays

The migration capability of cells was assessed using a scratch-wound assay. The cells were seeded into 6-well tissue culture dishes at a concentration of 1 × 10⁶ cells/mL and cultured in medium containing 10% FBS to achieve confluent cell monolayers. Then, a cross-shaped linear wound was generated in the monolayer with a sterile 200 μL plastic pipette tip. Any cellular debris was removed by washing the coverslips with phosphate buffer saline (PBS). Serum-free medium with 100 ng/mL SLC was added to the experimental group, while the control group was covered with serum-free medium only and incubated at 37°C with 5% CO₂. Images of cell migration into the wounds were captured with bright field microscopy at 24 h and 48 h. The results were analyzed with ImagePro Plus (USA). An independent T-test was conducted to determine the significance of the migratory effect. The wound healing effect was calculated as the percentage of the remaining cell-free area compared with the area of the initial wound. The experiments were performed at least in duplicate.

Statistical analysis

Experimental data were presented as the Mean ± SEM of at least three independent experiments performed in duplicate. The actual overall survival rates were calculated by the Kaplan-Meier method and analyzed by the log-rank test. A secondary analysis was performed to assess the relationship among the variables and clinicopathologic characteristics. Drug inhibition rate was analyzed by One-Way ANOVA. All analyses were performed using SPSS 20.0 software (SPSS, Chicago, IL) and GraphPad prism 5 (version 5.01).

High expression of CCR7 in CRC tissue was correlated with poor prognostic and cetuximab reaction

The clinical and pathologic characteristics of the 190 patients with CRC were summarized in Supplementary Table 1. Paired normal and tumor tissues were taken from each patient and stained by IHC with CCR7 antibody. According to the expression intensity of CCR7 in tumor tissues, 190 patients were classified into CCR7-positive group (n = 113) and CCR7-negative group (n = 77). The relationship between the clinical characteristics and the level of CCR7 expression was summarized in Table 1. The data shown that CCR7 expression was significantly correlated with the infiltration depth (P = 0.03) and regional lymph node metastasis (P = 0.006). There was no significant correlation of CCR7 expression with the patients’ age, gender, tumor differentiation, primary tumor parameters, distant metastasis, and TNM stage grouping.

At the same time, as shown in Figure 1, CCR7 was mainly expressed on the membrane of CRC cells (Figure 1A), and significantly higher in the paired tumor tissues (P < 0.001, Figure 1B). Furthermore, based on Kaplan-Meier method analysis, there was a significant difference both in overall survival (OS) (P < 0.001, Figure 1C) and Disease-free survival (DFS) (P < 0.001, Figure 1D) between the patients of CCR7-positive group and CCR7-negative group.

To explore whether CCR7 expression involves in Cetuximab (C225) reaction, CRC patients with Cetuximab treatment were taken into consideration and classified into PD (progressive disease) group (n = 7) and PR (partial remission) group (n = 7), based on their reactions to Cetuximab treatment. As shown in Figure 2, we found that the expression of CCR7 was significantly higher in the patients of PD group compared with the patients of PR group (Figure 2A and 2B, P = 0.014). More interesting, we found that EGFR also highly expressed and co-localized with CCR7 in PD group by double immunofluorescent staining (Figure 2C). These data implied CCR7 expression closely related to EGFR expression in the CRC patients who are resistant to Cetuximab treatment.

SLC/CCR7 axis lead to cetuximab drug resistance in CRC cells

In order to further explore the relationship of CCR7 high expression and Cetuximab drug resistance, we explored Cetuximab inhibition rates and CCR7 expression in seven different CRC cell lines. Seven CRC cell lines displayed various levels of CCR7 expression (Supplementary Figure 1A) and higher CCR7 expression lead to lower sensitive to Cetuximab treatment (Supplementary Figure 1B). HCTT-116 harbors a mutated KRAS gene while Caco-2 has a wide-type one, we chose CCR7-high expression cell lines HCT116 (KRAS^G13D) and Caco-2 (KRAS^wt) to do further research. As shown in Figure 3, after SLC stimulation, the Cetuximab inhibition rate was significantly decreased. For Caco-2 cells, the inhibition rate decreased from 50.37% to 28.7% at 20 ug/ml node (Figure 3A and 3B, **P = 0.005), and 14% to 6% in HCT116 cells at 100 ug/ml node (*P = 0.01, Figure 3C and 3D). However, pre-treating cells with Ne-CCR7 reverse SLC induced Cetuximab resistance, and the inhibition rate in HCT116 cells increased more than one-fold compare Ne-CCR7 pre-treating group to control group. Meanwhile, cell cycle analysis of HCT116 showed that SLC-stimulated cells re-accumulate in S phase in response to treatment with Cetuximab and Ne-CCR7 (Supplementary Figure 2).

PI3K/AKT pathway was activated by SLC/CCR7 axis in CRC cells

To explore the underlying mechanisms of SLC induced drug resistance, CCR7, EGFR and theirs downstream genes were firstly screened by bioinformatics analysis. The data showed that PIK3CA and PIK3CB were involved in the signaling cascades of both CCR7 and drug resistance (Supplementary Figure 3). More importantly, PI3K/AKT signaling not only was the key downstream pathway of CCR7, but also the activated signaling during the process of drug resistance. Therefore, the expression of AKT was checked. SLC treatment induced phosphorylation of AKT (p-AKT) (Figure 4A) while Ne-CCR7 down-regulated p-AKT (Ser473) in the both groups with or without SLC treatment. We next treated CRC cells with Cetuximab. The data shown p-AKT was up-regulated by SLC and SLC plus Cetumimab treatment groups, respectively (Figure 4B). Surprisingly, EGFR were also activated in these two groups (Supplementary Figure 4).

Selective PI3K/AKT inhibitor sensitizes colorectal cancer cells to cetuximab

To further confirm SLC/CCR7 induced Cetuximab resistance via PI3K/AKT pathway, we next treated cells with Cetuximab and/or LY294002. As shown in Figure 5, compared with Cetuximab plus SLC group, Cetuximab plus LY294002 combined treatment significantly reversed Cetuximab resistance that was induced by SLC (Figure 5A-D). The Ne-CCR7 treatment also resulted in an increase of inhibition rate compared with Cetuximab plus LY294002 in CRC cells.

Activation of SLC/CCR7 axis facilitated CRC cell lines undergoing EMT

As long as PI3K/AKT, the downstream pathway of CCR7 and EGFR was activated after SLC treatment, how they facilitated tumor cells metastasis seems to be interesting. Our bioinformatic analysis identified CCR7 and EGFR were closely related to EMT (Supplementary Figure 3). It showed that the migration ability of the two cell lines were significantly promoted by the treatment of SLC after 48 hours (***P < 0.001) (Figure 6A-D). And the EMT transcription factor Twist1 and β-catenin was up regulated after treated with SLC (Figure 6E and 6F). We next applied Ne-CCR7 to see whether blocking the pathways would prevent EMT of tumor cells. As expected, the expressions of Twist1 and β-catenin significantly decreased in the circumstance with SLC or not (Figure 6E and 6F). These results indicated that activation of SLC/CCR7 downstream pathway PI3K/AKT contributes to the tumor cells undergoing EMT.