Background
After the discovery of
echinoderm microtubule associated protein like 4 (EML4)-ALK rearrangement in NSCLC,
ALK-positive NSCLC patients obtained remarkably improved progression-free survival (PFS) and overall survival (OS) with ALK-TKIs treatment [
1,
2]. However, due to the emergence of primary and acquired resistance to targeted therapy, clinical outcomes are heterogeneous among different patients [
2,
3]. Illuminating the molecular mechanism of drug resistance and searching for prognostic biomarkers are conducive to guiding next-line therapies.
A compelling body of evidence indicates that angiogenesis is involved in the occurrence and development of many solid tumors, including lung cancer [
4,
5]. Chemokines are small proteins (8–10 kDa) belonging to the family of chemoattractant cytokines, including 4 subtypes, C, CC, CXC, and CX3C [
6]. The chemokine network has been reported to participate in regulating the distribution of immune cells and angiogenic activity in the tumor microenvironment [
6,
7]. Among them, several studies demonstrated that the interaction of CCL20 and CCR6 promoted the tumor progression in melanoma, breast cancer, and hepatocellular carcinoma via enhancing angiogenesis [
8,
9]. Moreover, CCL20 induced cell proliferation and migration in lung cancer[
10]. The expression level of CCL20 in serum served as a crucial biomarker for prognostic prediction in NSCLC patients [
11]. However, whether CCL20 can serve as a prognostic marker for
ALK-positive NSCLC patients, and the role of CCL20-induced angiogenesis in crizotinib resistance remain unclear.
Anlotinib hydrochloride (AL3818) is a novel multi-target TKI to inhibit the angiogenesis of tumor [
12,
13]. It has been recommended as a third-line or further treatment for driver gene-positive advanced NSCLC in China [
14]. We hypothesized that anlotinib could improve the clinical outcomes via blocking the angiogenesis in
ALK-positive NSCLC patients, but the underlying molecular mechanism has not been fully clarified yet. Furthermore, recent clinical trials indicated that the combination of anti-angiogenic drugs (such as bevacizumab or anlotinib) with chemotherapy, epidermal growth factor receptor (EGFR)-TKIs targeted therapy, or immunotherapy prolonged the PFS and OS of NSCLC patients [
14‐
17]. However, whether the combination of anlotinib with ALK-TKIs can reverse crizotinib resistance and improve the response to ALK-TKIs has not been reported yet. In the current study, we identified plasma biomarkers to monitor and predict the drug resistance and clinical response of crizotinib, clarified its underlying mechanism, and explored the anti-tumor effect of the combination of anlotinib and crizotinib in crizotinib-resistant NSCLC.
Methods
Patient samples
76 plasma samples were collected from 61
EML4-ALK positive NSCLC patients treated with crizotinib in the Cancer Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College (CAMS & PUMC) from 2010 to 2015 (CHCAMS cohort). All baseline samples of 61 patients were obtained before the initiation of treatment, and the progression samples of 15 patients with PFS longer than 12 months were collected after crizotinib resistance. The cut-off date for the follow-up was February 2, 2021. Peripheral blood samples were collected in K
2EDTA tubes and centrifuged at 1600 g for 10 min after collecting. The isolated plasma was centrifuged again at 16,000 g for 10 min to remove cell debris. All extracted plasma samples were stored at – 80 ℃. All samples and data collected were with informed consent. This study was approved by the medical ethics committee of Cancer Hospital, CAMS & PUMC (No.19-019/1804). The patient characteristics of the CHCAMS cohort were shown in Table
1.
Table 1
Patient characteristics
Age, median (SD) | | 49 (13) |
Sex, % female | | 65.6% |
Smoking status (% never smokers)a | | 73.2% |
ECOG PS (%) at baselineb | 0 | 8 |
| 1 | 31 |
| 2 | 1 |
Histology (%) | Adenocarcinoma | 49/50c |
Chemotherapy | | 52/61 |
ALK TKI, patient number | Crizotinib | 61 |
Metastasis | Braind | 33/49 |
| Bonee | 31/55 |
Follow-up in months (median, [IQR]) | | 52.4 (38.0–82.43) |
Cases with baseline samples | | 61 |
Cases with disease progression samples | | 15 |
Cells lines and reagents
H3122 (
EML4-ALK, variant 1) and H2228 (
EML4-ALK, variant 3) cell lines were purchased from ATCC and PUMC-HUVEC-T1 cells were obtained from Cell Resource Center, Peking Union Medical College (Beijing, China). Crizotinib-resistant cell line (H3122CR) was developed from its parental cell line H3122 in our previous study [
18]. RPMI-1640 medium (HyClone, USA) with 10% FBS (Gibco, USA) was used to culture H3122 and H2228 cells. PUMC-HUVEC-T1 cells were cultured in DMEM containing 10% FBS and 1% NEAA (Cell Resource Center, Beijing). Additionally, crizotinib and STAT3 inhibitor Stattic were purchased from Sigma-Aldrich and MedChemExpress (MCE), respectively. Anlotinib was obtained from Chiatai Tianqing (Jiangsu, China).
Luminex liquid suspension chip detection
The expression levels of 40 chemokines in plasma samples and cell supernatants were detected using the Bio-Plex Pro Human Chemokine Panel 40-plex kit with Luminex 200 system performed by Wayen Biotechnologies (Shanghai, China). The 40 chemokines evaluated are shown in Additional file
5: Table S1.
RNA-seq and DEGs analysis
Total RNA from H2228, H3122, and H3122CR cells cultured for 48 h was extracted using the TRIzol reagent. The NEBNext Ultra RNA Library Prep Kit (NEB, USA) was used to construct cDNA library following the manufacturer’s directions. Paired-end sequencing with 150 bp reads was conducted on the Illumina Novaseq platform (Novogene, Beijing, China). After data cleaning, all clean reads were mapped to the reference genome (hg38) using Hisat2 v2.0.5. The gene expression level was measured by the fragments per kilobase per million (FPKM). The edgeR R package (3.18.1) was used to identify DEGs between H3122 and H3122CR cells. We selected DEGs with a corrected p-value < 0.05 and log2Fold Change > 1. The clusterProfiler R package was used to perform the Gene Ontology (GO) enrichment analysis and Gene Set Enrichment Analysis (GSEA) for DEGs.
RT-qPCR
Total RNA isolation was performed in cell lines after treatment for 24 h using the RNeasy Mini Kit (Qiagen, Germany). cDNA was synthesized by reverse transcription polymerase chain reaction (RT-PCR) with PrimeScript ™ RT reagent kit (TAKARA, Japan). Then the SYBR Premix Ex Taq ™ II kit (TAKARA, Japan) was used for quantitative real-time PCR (qPCR) assay on Roche LightCycler480 II platform. GAPDH was used as a reference control for normalization. All primer sequences used in RT-qPCR detection are listed in Additional file
5: Table S2.
Enzyme-linked immunosorbent assay (ELISA) and western blot (WB) analysis
CCL2 (SinoBiological, KIT10134), CCL20 (Abcam, ab269562), and CCL24 (Abcam, ab10050) levels in culture supernatant after treatment for 48 h were measured using ELISA kits. RIPA buffer with protease inhibitor and protein phosphatase inhibitor (APPLYGEN) was used for cell lysis. BCA reagent (APPLYGEN) was applied to detect the protein concentration. Cell lysates were loaded to 10% SDS-PAGE gel and transferred to PVDF membranes after separation. The following antibodies were used to detect proteins. GAPDH (CST, 14C10), VEGFA (Abcam, ab46154), STAT3 (CST, 79D7), p-STAT3 (Tyr705) (CST, D3A7), JAK2 (CST, D2E12), p-JAK2 (Tyr1007/1008) (CST, C80C3). Goat anti-rabbit IgG was used as the secondary antibody (CST).
Gene knockdown by siRNA
We plated 2 × 10
5/well H3122CR cells on 6-well plates for 24 h. Following the manufacturer’s instructions, cells were transfected with siRNA (Sangon Biotech) using Lipofectamine 3000 (Invitrogen). After transfection for 24–48 h, RT-qPCR or ELISA analysis was applied to verify the efficiency of gene knockdown. Cells transfected with nonsense siRNA duplexes were used as a control. Target sequences for siRNAs are listed in Additional file
5: Table S3.
Cell growth and viability assay
3 × 103/well cells were cultured for 0–72 h in 96-well plates to draw the cell growth curve. For dose–response curve, we plated cells in 96-well plates for 24 h and then treated with nine concentrations (0, 0.008, 0.04, 0.2, 1, 5, 10, 20, 40 μM) of crizotinib or anlotinib for 48 h. CCK-8 (Dojindo, Japan) assay was conducted to determine cell viability. Half maximal inhibitory concentration (IC50) values were calculated from the dose–response curves using GraphPad Prism 8 software.
Cell cycle and apoptosis assays
2 × 105 H3122CR cells were treated with siRNA after growing in 6-well plates for 24 h. To analyze the cell cycle phase, cells treated for 42 h were stained with propidium iodide (PI) (Dojindo, Japan). To evaluate cell apoptosis, Annexin V–fluorescein isothiocyanate (FITC) and PI were used to stain the cells after treatment for 48 h (Dojindo, Japan). Flow cytometry was used to analyze prepared samples (BD FACSCalibur).
H3122CR cells (500 cells per well) were cultured in 6-well plates for 24 h and then exposed to 1 μM anlotinib for 2 weeks. Due to the weaker colony-forming ability of H3122 and H2228 cells, H3122 and H2228 (4000 cells per well) were cultured in 6-well plates for 48 h and then exposed to 1 μM anlotinib for 10 days. 100% methanol and 0.1% crystal violet were used to fix and stain the cell colonies for 20 min, respectively.
Cell culture medium (CM) was collected after treatment for 24 h. 2 × 104 HUVECs (human umbilical vein endothelial cells) were suspended in 50 μl CM and planted on 96-well plates coated with 50 μl Matrigel (Corning, USA). The number of tubes and capillary length were assessed after seeding for 4 h using ImageJ software.
Animal experiments
Four-week-old female BALB/c nude mice were acquired from HFK Bioscience (Beijing, China) and maintained under specific pathogen-free conditions. To construct the H3122CR-derived xenograft model, 5 × 106 H3122CR cells were subcutaneously injected into the right flank of mice. When the tumor volume reached to 100 mm3, mice were randomized to four groups and treated daily (day 0) by oral gavage as follows: (a) Control group (n = 6): 0.2% CMC-Na; (b) Crizotinib monotherapy group (n = 6): crizotinib (50 mg/kg/d) alone; (c) Anlotinib monotherapy group (n = 6): anlotinib (3 mg/kg/d) alone for two consecutive weeks and then discontinued for one week; (d) Combination treatment (n = 6): crizotinib (50 mg/kg/d) combined with anlotinib (3 mg/kg/d) for two consecutive weeks and then crizotinib alone for one week. Body weight and tumor volume were measured every two or three days, and tumor volume was calculated from the following formula: tumor volume (mm3) = length × width^2/2. At the end of the experiment (day 26), mice were sacrificed and tumors were collected and weighed. All animal experiments were approved by the Animal Care and Use Committee of Cancer Hospital, Chinese Academy of Medical Sciences (No. NCC2018A026).
Analysis of The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) cohort
mRNA expression levels and clinical data for 497 lung adenocarcinoma (LUAD) patients were downloaded from TCGA dataset (The Pan-Cancer Atlas,
http://www.cbioportal.org,
https://gdc.cancer.gov/about-data/publications/pancanatlas). According to the median value of mRNA expression level, patients were divided into high expression group and low expression group to compare the disease-free survival (DFS) and OS. The comparison analysis of mRNA expression between lung squamous carcinoma (LUSC), LUAD, and normal tissue, and the correlation analysis between two genes were performed according to the Gene Expression Profiling Interactive Analysis (GEPIA,
http://gepia.cancer-pku.cn/). GSE94089 dataset was obtained from GEO database containing RNA-seq data of 4 cell lines, including H2228, H2228-crizotinib resistance (CRR), H2228-ceritinib resistance (CER), and H2228-alectinib resistance (ALR). GEO2R was used to screen the DEGs between H2228 and H2228-resistance cell lines. The thresholds for significant DEGs selection were p-value < 0.05 and log2
Fold Change > 1. Then 381 significant DEGs were used to perform GO enrichment analysis using DAVID Bioinformatics Resources 6.8 and Cytoscape v3.8.2 software (
https://david.ncifcrf.gov/,
https://cytoscape.org/).
Statistical analysis
In this study, data are presented as mean ± SD. Log-rank test was performed in univariable analysis using Kaplan–Meier survival analysis. COX regression model was conducted in multivariable analysis. R package “limma” was used for differential expression chemokines analysis. Three independent experiments of biological replicates were performed in all cell experiments excluding RNAseq. Student’s t-test was conducted in continuous variables. And Chi-square test was performed in categorical variables. SPSS 25.0, GraphPad Prism 8, and R packages were utilized for statistical analysis.
Discussion
Angiogenesis leads to tumor occurrence, progression, and drug resistance [
4,
5,
22]. Here, we reported that chemokine-induced angiogenesis drives resistance to crizotinib in
ALK-positive patients. This discovery has provided possible plasma chemokine biomarkers for response prediction in ALK-TKIs treated patients and guided us to explore the therapeutic effect of anlotinib monotherapy and combination therapy in ALK-TKIs resistant patients.
Although researches on prognostic markers for chemotherapy and immunotherapy have increased dramatically in NSCLC patients, biomarkers for ALK-targeted therapy are rarely reported. One of the clear evidence is that mutated
TP53 correlated to unfavorable crizotinib PFS in
ALK-positive patients [
23,
24]. Another prognostic marker for ALK-TKI therapy is
ALK variant type, but it remains contentious. Yoshida, et al. firstly found that
EML4-ALK variant 1 was correlated with superior prognosis versus non-v1, while Woo, et al. demonstrated that patients with v3a/b had shorter PFS after crizotinib treatment [
25,
26]. However, some other studies indicated no significant correlation between
ALK fusion variants and clinical outcomes [
27]. Due to the difficulty of tissue sample collection, circulating tumor cells and cell-free DNA in plasma samples were also investigated to monitor the duration and magnitude of clinical response of patients receiving ALK-TKIs [
28,
29]. In the current study, we showed that 3 chemokines (CCL24, CCL15, and CCL20) in baseline plasma samples were associated with PFS, and baseline CCL20 was an independent prognostic factor for OS in patients treated with crizotinib. Subsequently, we reported that baseline CCL20 and CCL24 were significantly elevated in patients with PFS < 6 months compared with patients with PFS > 12 months, indicating that CCL20 and CCL24 possibly contributed to the primary resistance to crizotinib. Furthermore, the detection of baseline and progression samples demonstrated that the dynamic changes of CCL20 and CCL15 can monitor the acquired resistance to crizotinib. TCGA and GEPIA datasets also verified that high expression of CCL20 and low expression of CCL15 were significantly related to the initiation of LUAD. Due to the convenience of plasma chemokines detection, CCL20, CCL24, and CCL15 can serve as prognostic markers to identify crizotinib resistance and predict the clinical outcomes in
ALK-positive patients. However, we failed to confirm the molecular function of CCL15 in cell experiments. Contrary to clinical results, CCL15 overexpressed in H3122CR instead of H3122 cells. Previous studies mainly reported that CCL15 was involved in the occurrence and development of hepatocellular carcinoma (HCC) [
30]. Highly expression of CCL15 was correlated with dismal survival in HCC patients [
30]. But the effect of CCL15 on lung cancer remains controversial. One previous study indicated that the CCL15 level significantly decreased when patients had a partial response after erlotinib and celecoxib treatment [
31]. However, according to our study, CCL15 decreased when patients acquired drug resistance and progressed. The complexity of the effect of CCL15 on NSCLC may explain the inconsistent results in vivo & vitro experiments in the current study.
Based on the observation of angiogenesis enriched both in clinical samples and cell lines, we explored the role of angiogenesis in the process of CCL20-induced resistance. We demonstrated that CCL20 boosted angiogenesis via JAK2/STAT3-CCL20-VEGFA/IL6 axis to confer resistance to crizotinib. Subsequently, anti-angiogenic drug anlotinib was applied to overcome resistance, showing an inhibitory effect on 4 chemokines and JAK2/STAT3-VEGFA/IL6 axis. It suggested that anlotinib not only suppressed angiogenesis in the tumor microenvironment but also inhibited angiogenesis pathways in tumor cells. Similarly, a previous study found the consistent phenomenon that anlotinib suppressed JAK2/STAT3/VEGFA pathway in NSCLC xenograft tumors [
32]. Moreover, Lu, et al. indicated that anlotinib inhibited tumor growth in an
EGFR-mutant NSCLC xenograft model by restraining the CCL2-induced angiogenesis, which further confirmed our results [
19]. Interestingly, the suppression effect of anlotinib on these 4 chemokines was not significant in crizotinib-sensitive H3122 and H2228 cells. This is possible because chemokines dramatically overexpressed in H3122CR compared with H3122, and anlotinib showed a stronger inhibitory effect on chemokines to reverse drug resistance. The underlying mechanism of anlotinib inhibiting H3122 and H2228 needs further exploration.
Preclinical researches and clinical trials have demonstrated that anti-angiogenic drugs combined with EGFR-TKIs is a viable strategy for
EGFR-mutant NSCLC patients [
33‐
36]. According to the results of an ongoing phase II clinical trial, impressive objective response rate and disease control rate were obtained with acceptable toxicity after the united medication [
33]. Li, et al. indicated that the combination of anlotinib with gefitinib enhanced the inhibition to cell proliferation in vitro and tumor angiogenesis in xenograft models [
35]. Here, based on the inhibitory effect of anlotinib on H3122CR cells and H3122CR-derived xenografts, we firstly reported a combination strategy of anlotinib with crizotinib for crizotinib-resistant
ALK-positive NSCLC. Results showed that anlotinib boosted the anti-tumor effect of crizotinib in vitro & in vivo. Similarly, a recent study has reported that the combination of anti-VEGFR2 antibody with crizotinib augmented the effect of anti-proliferative effects on tumor cells [
36]. Moreover, a phase II clinical trial combining alectinb with bevacizumab in
ALK-positive non-squamous NSCLC patients with alectinib resistance showed clinical efficacy and acceptable toxicity in vivo [
37]. Therefore, the combination therapy of anlotinib with ALK-TKIs could serve as a promising treatment strategy for
ALK-positive NSCLC patients.
Several limitations should be noted in our study. One limitation is the small sample size of the included patient cohort. And due to the limited sample size of EML4-ALK fusion NSCLC in TCGA dataset, we only validated the prognostic efficacy of CCL20 in LUAD patients but not in ALK-positive patients. Besides, we have found several chemokines related to crizotinib response in clinical samples but only the underlying mechanism of CCL20 was clarified in this study. Thirdly, clinical trials will be carried out to verify the anti-tumor effect of anlotinib monotherapy and combination therapy on ALK-positive patients in the future. Moreover, we mainly focused on the treatment strategies for ALK-TKIs-resistant NSCLC in the current study. Hence, the efficacy and molecular mechanism of anlotinib alone or combined with ALK-TKIs for the untreated ALK-positive patients still need further exploration.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.