Tracking of trastuzumab resistance in patients with HER2-positive metastatic gastric cancer by CTC liquid biopsy

Abstract

This study aimed to utilize circulating tumor cell-DNA (CTC-DNA) from liquid biopsies to monitor trastuzumab resistance in Gastric cancer (GC) and assess the limited response rate in HER2 metastatic gastric cancer. Given the heterogeneity of GC, we established a high-precision CTC detection system that effectively isolates tumor cells with high HER2 expression for downstream analysis. Targeted sequencing of 610 genes was conducted on 20 paired CTC and tissue samples to assess uniformity. A longitudinal analysis of CTC samples was then performed to monitor trastuzumab resistance throughout treatment. Targeted sequencing of the HER2 gene showed strong consistency with fluorescence in situ hybridization data. Detected HER2 Scna was superior in predicting tumor shrinkage and progression. Most patients with innate trastuzumab resistance exhibited elevated HER2 Scna levels during progression. PIK3CA mutations were significantly enriched, and ERBB2/4 gene mutations were predominant in patients with innate trastuzumab resistance. CTC-DNA sequencing provides new insights into gene alterations associated with trastuzumab resistance in HER2 mGC.

Introduction

Gastric cancer (GC) is the fifth most commonly diagnosed cancer globally and the third leading cause of cancer-related deaths. However, advancements in GC treatment, especially in advanced stages, have been gradual [1,2]. Human epidermal growth factor receptor 2 (HER2/ErbB2/neu) is a 185 kDa transmembrane receptor tyrosine kinase (RTK) and belongs to the epidermal growth factor receptor (EGFR) family [3-5]. HER2-positive GC constitutes approximately 6% to 30% of all GC cases. Given that overexpression of HER2 is observed in more than 15% of GC and is linked to a poor prognosis, it serves as a crucial therapeutic target for HER2-positive metastatic gastric cancer (mGC) [6,7]. Currently, tissue detection is the primary diagnostic method. As biopsies provide limited information and can be invasive, there’s a pressing need for non-invasive techniques to identify patients for anti-HER2 treatments and to monitor therapeutic outcomes, thereby improving the management of advanced GC [8,9].

Over the last few decades, numerous studies have highlighted the potential of circulating tumor cells (CTCs) as a novel blood biomarker for the diagnosis, prognosis, and treatment of various cancers, including GC [10-12]. CTCs, tumor cells from primary or metastatic foci that are shed into the peripheral circulation, carry the genetic and phenotypic information of the parental lesion, and thus have high research and clinical value. Liquid biopsy with CTC detection as a diagnostic tool has been widely studied in its application to GC and has progressed from a simple counting tool, to allowing for genotyping and molecular diagnostics. CTC detection and capture can aid in determining the genetic profiles of tumor cells over time without requiring a direct biopsy of the tumor mass. Wang and his colleagues performed the targeted sequencing of 146 clinically relevant genes in 78 paired plasma circulating tumor DNA (ctDNA) and biopsy samples to determine the consistency of plasma with tissue [13]. However, there are still gaps in comparing the consistency of DNA from CTC with tumor tissue. Since GC has high intrinsic molecular heterogeneity, the single-site immune lipid magnetic beads have a higher chance than multi-site beads to miss CTCs. Multi-site immune lipid magnetic beads based on HER2 and EpCAM can better capture CTCs of GC, improve the positive detection rate, and provide a foundation for the classification of GC and subsequent prognosis based on genetic signatures and CTC prevalence.

Trastuzumab, a targeted therapy for HER2, is effective both as a standalone treatment and in combination with other drugs. Evaluating GC tissues for HER2 overexpression is crucial to identify patients who would benefit from trastuzumab. Clinical trials have demonstrated that overall survival of HER2-positive GC patients significantly improved from trastuzumab treatment, leading to its approval as a first-line therapy [6]. Studies have shown that MET (hepatocyte growth factor receptor) mutations are present in GC patients, especially those with drug resistance. Some studies suggest that MET mutations contribute to resistance mechanisms against alfatinib and lapatinib in GC. We isolated GC CTCs using multi-site immune lipid magnetic beads targeting HER2 and EpCAM for early diagnosis and prognosis assessment. These captured CTCs were cultured to develop an animal model to study the impact of MET mutations on trastuzumab resistance in GC.

Materials and methods

Sample collection

Blood was collected from 50 patients, along with tissue samples of 20 paired patients, who had been pathologically diagnosed with GC in South Hospital of Renji Hospital, Shanghai Jiaotong University hospital between November 2020 and March 2022. All samples were obtained with informed patient consent, and the study was approved by the Ethics Committee of the Shanghai Jiaotong University Affiliated Renji Hospital (KY2021-013).

Materials and instruments

RPMI, fetal bovine serum (FBS), and trypsin were purchased from Gibco. BALB/C nude mice were purchased from Shanghai Jiesijie Laboratory Animal Co., Ltd. Anti-HER2 antibody (ab16901) and goat anti-rabbit secondary antibody (ab6721) were purchased from Abcam. The Path Vysion HER2 kit for fluorescence in situ hybridization (FISH) probe, RIPA lysate, and chemiluminescence kit were purchased from Shanghai Abbott Biotechnology Co., Ltd. 1-ethyl-3-(3-dimethylammonium propyl) ammonium carbohydrate (EDC), N-hydroxysuccinimide (NHS), EpCAM-lipid magnetic beads, Fe₃O₄, carboxymethyl chitosan hexadecane alkyl quaternary amine salt (HQCMC), 1,2-dioleoylphosphatidylcholine (DOPC) and dimethyl octadecyl epoxypropyl ammonium chloride (GHDC) were purchased from Huzhou Lieyuan Medical Laboratory Co., Ltd. OLYMPUS B×61 fluorescence microscope was purchased from Olympus Corp, a Japanese enterprise. BI-90Plus laser particle analyzer/Zeta potentiostat was purchased from American Brook Haven. XL-30 environmental scanning electron microscope (ESEM) was purchased from PHILIPS in the Netherlands.

Preparation of HER2- and EpCAM-lipid magnetic beads

HER2-lipid magnetic beads (H-LMB) and EpCAM-lipid magnetic beads (E-LMB) were synthesized as described in referenced literature. In brief, a mixture of cholesterol, DOPC, GHDC, HQCMC, and Fe₃O₄ in dichloromethane was combined with a 0.1 M PBS solution. The mixture was stirred vigorously to ensure emulsification and heated to 25°C to obtain LMBs. During this process, mechanical stirring was imperative to avoid agglomeration or aggregation of iron oxide nanoparticles and nanocolloids. 0.6 mg HER2 antibody was dissolved in 10 mL isopropanol. Coupling agents EDC and NHS were added respectively to the mixture of nanomagnetic beads. This mixture was stirred at a constant rate for 24 h to obtain lipid magnetic beads conjugated with anti-HER2 antibody. The preparation process for EpCAM-lipid magnetic beads was the same as above.

Characterization of HER2-LMB and Ep-LMB

10 μL H-LMB of the sample was diluted with 1 mL distilled water; a BI-90Plus laser particle analyzer/Zeta potentiostat was then used to detect the particle size and zeta potential. The atomistic characteristics of H-LMB were determined by an atomic force microscope (AFM). 10 μL H-LMB sample was diluted with 1 mL distilled water. Then, 50 μL of this solution was coated on a mica sheet for working in non-contact tap mode (frequency =325 Hz) at a scanning speed of 1.0 Hz. The mica sheet was observed when it was dried. 10 μL H-LMB and LMB samples were diluted with 1 ml distilled water, respectively, after which the absorbance (A) at 280 nm was measured with an ultraviolet (UV) spectrophotometer. The magnetic property of H-LMB was detected with a vibrating sample magnetometer. The characterization of E-LMB was the same as above.

Detection of expression of HER2 in tissues of GC patients

HER2 expression in GC tissues was confirmed using FISH. Tissues were treated at high temperature and high pressure for 3 min, digested with pepsin (0.2 mg/ml) for 20-40 min, then dehydrated and dried. 10 mL probe mixture was added to the target region of tissue slices, which were then covered and sealed. Tissue slices were placed in an in situ hybridization instrument and denatured at 83°C for 5 min then hybridization was performed at 45°C overnight (14-18 h). Tissue slices were rinsed quickly and dried at room temperature. 15 μL DAPI was added dropwise, slices were sealed, and staining was observed under a fluorescence microscope.

Gene detection

Second-generation sequencing by ligation and bioinformatic analysis were performed by Daen Biotechnology Co., Ltd. The ligation reactions of 4 fluorescent-labeled oligonucleotides were used for sequencing. Before sequencing, the DNA template was concentrated by Integrated vacuum concentrator (BIONOON VAC-AIO, Shanghai BANNUO BIOBECHNOLOGY CO., LTD.) and amplified by emulsification PCR, and the 3’-modified microbeads were deposited on the glass substrate. Sequencing by ligation was performed using an 8-base fluorescent probe mixture. Multiple rounds of hybridization were performed where fluorescent probes matching the reverse complement template at the 5’ end are ligated to the 3’ end of the anchor sequence. DNA at these ligation positions can then be determined by fluorescent readout.

Statistical analysis

Data were processed using SPSS 20.0. Comparisons between groups were made using the independent samples t-test, with a significance threshold set at α=0.05. A p-value less than 0.05 was deemed statistically significant. All reported data represent the mean of three independent experiments.

Results

Procedures and characterization of LMB

In this study, we aimed to separate and identify CTCs from peripheral blood (PB) in GC clinical samples, perform quantitative statistics, and conduct genetic tests based on the established lipid magnetic beads for HER2 and other antibodies (including EpCAM lipid magnetic beads; E-LMB is a routine general-purpose CTC sorting system). This approach was chosen to investigate the ability of CTC capture to help detect and monitor trastuzumab resistance in GC. Technical details are illustrated in Figure 1.

Figure 1.

Schematic of the preparation of H-LMB and CTC detection.

The performance of nanomagnetic bead sorting in the CTC separation and detection system is paramount for high efficiency at capturing GC CTC in the blood. The UV absorption spectrum is shown in Figure 2A. The nanolipid magnetic beads for HER2 antibody and EpCAM antibody exhibited distinct UV absorption peaks near 280 nm, indicating the presence of modified antibodies on the nano magnetic bead surface, and that antibody loading was controllable. Figure 2B shows the magnetic saturation curve for E-LMB and H-LMB, with empty LMB serving as the control. The results confirmed that the prepared magnetic beads possess high saturation magnetization and superparamagnetism. The saturation magnetization of E-LMB and H-LMB at 300k was 18.97 emu/g, which was about 73.4% of that of raw Fe₃O₄ magnetic beads. The particle size distribution diagram for E-LMB and H-LMB is displayed in Figure 2C and 2E, respectively. The average particle size was 204.1±5.7 nm and 193.7±6.4 nm, respectively. These findings were similar to the results of particle size shown by AFM (Figure 2G, 2H). The surface potentials of E-LMB and H-LMB were 19.4±3.1 mV and 25.8±2.6 mV in a neutral state, which reduces the nonspecific adsorption interference observed in previous studies (Figure 2D, 2F). These characterizations confirm the successful construction of the E-LMB and H-LMB capture systems for GC CTCs.

Figure 2.

A. UV absorption spectrum of E-LMB and H-LMB; B. Magnetic saturation curve of E-LMB and H-LMB; C-F. Zeta-potential distribution and particle size distribution of E-LMB and H-LMB; G, H. Atomic force image of E-LMB and H-LMB.

CTC detection and FISH

Since HER2 overexpression is the main marker of HER2+ GC, we analyzed CTC captured by H-LMB using FISH. Immunofluorescence characterization of CTC captured by E-LMB and H-LMB is shown in Figure 3A. GC patient CTCs exhibited distinct cellular morphology under white light and were strongly positive for CK19-FITC green fluorescence and DAPI blue luminescence. Captured cells were determined to be GC CTC by HER2+CK19+CD45- staining. FISH was used to analyze the blood of 20 GC patients. HER2 mutations were detected in the CTCs separated from the tissues of 5 GC patients and captured from 6 H-LMBs. Among 20 GC patient blood samples, 7 cases exhibited HER2 mutations, with 6 patients having CTCs with HER2 mutations. A subset of the detection results is presented in Figure 3B (refer to Table 1 for patient details).

Figure 3.

(A) Immunofluorescence identification results of CTCs of GC patients captured by E-LMB and H-LMB; and (B) FISH detection results of HER2 mutations in tissues of GC patients.

Table 1.

Clinical information and CTC count of patients

ID	Gender	Age	Tumor staging	CTC (+/-)
17201425	Male	55	II	6
17200334	Male	60	III	3
20117667	Male	71	IV	3
17201248	Female	65	I	2
17106909	Male	64	III	11
17102754	Male	52	III	12
17201002	Male	48	II	6
17202979	Female	68	IV	13
18201228	Male	49	I	4
20088540	Male	74	I	3
18201228	Male	49	II	10
18202862	Male	55	III	11
18202832	Male	55	III	6
19200731	Female	72	II	9
18201695	Female	64	III	8
19201069	Male	68	IV	17
17201866	Male	65	II	9
19200731	Female	72	III	11
19201326	Female	58	I	3
19201069	Male	68	II	6

To achieve precise labeling for both IF and FISH, we meticulously optimized the procedure to ensure effective labeling for both targets. All conditions, including temperature, duration, and reagent concentration, were refined through preliminary experiments using cell line-spiked peripheral blood (Figure 4). Specifically designed for the HER2 gene amplification, the FISH method was employed, and the results of the analysis of mean HER2 copy number/cell, mean CEP17 copy number/cell, and mean HER2 copy number/mean CEP17 copy number are presented in Figure S1. The results revealed that mean HER2 copy number/cell was the highest, while mean CEP17 copy number/cell was the second highest, and mean HER2 copy number/mean CEP17 copy number was the lowest.

Figure 4.

Immunofluorescence and FISH labeling in immunomagnetically enriched tumor cells: schematic of the specimen preparation and fluorescent labeling.

Molecular alterations in tumor and fluid biopsies of GC patients

To assess the consistency between molecular alterations detected in tumor and fluid biopsies, we sequenced 610 genes in 20 GC patients (14 CTC HER2+ and 6 CTC HER2-) and analyzed DNA copy number alterations, mutations, and structural variations in the tumor tissues of 20 patients paired with regular PB (detailed data of the 610 genes is provided in Table S1). Figure 5A displays the high-frequency (> 5%) molecular alterations detected in tissues and blood. The results revealed that ERBB2 (HER2) and TP53 are the most frequently altered genes in both tissues and blood, exhibiting notable heterogeneity. In terms of mutation identification, tissue-based detection methods generally identify more mutations. However, the number of mutations detected in tissues and CTC samples is highly correlated. Although the overall consistency was low, we observed increased consistency with advancing stages. Specifically, the detection ratio of TP53/HER2 mutation consistency between mGC (stage) tissues and CTC samples was 71.43% (Table 2). In conclusion, the molecular alterations detected in CTCs accurately reflect the state of tumor tissues, especially in advanced stages.

Figure 5.

A. High-frequency molecular alterations detected in CTCs in blood and alterations detected in paired tissue samples; B. Molecular alterations detected in known trastuzumab resistance-related genes at baseline and during progression of disease (PD).

Table 2.

The clinicopathological characteristics of patients (N=50)

Characteristics Number		TP53+		HER2+
		tissue	CTC	tissue	CTC
Gender
Male	39	16	9	18	13
Female	11	10	10	11	11
Age
<65	38	21	13	24	19
≥65	12	10	6	5	6
Tumor Stage
I Stage	8	5	6	4	4
II Stage	5	4	2	4	1
III Stage	16	14	10	15	13
IV Stage	11	8	7	6	7
Trastuzumab treatment
Yes	32	18	12	26	24
No	18	13	7	3	1

We defined molecular alterations associated with trastuzumab resistance as an increase in mutations or mutation during treatment. As depicted in Figure 5B, in 6 (76.5%) patients, the molecular alterations of these genes abruptly changed or mutated during treatment, potentially indicating trastuzumab resistance.

Considering the benefits of a larger sample size, we analyzed the sequencing results of 33 additional gastric cancer CTC-NGS to identify the primary CTC-derived mutant genes in gastric cancer. The results are presented in Figure S2C, which highlights several genes with high mutation frequency, including TP53, ERBB2, and CDH1. We also enumerated the primary mutation sites for these genes, as shown in Figure S2A, S2B, S2D, S2E. The primary mutation sites of TP53 were exon 5, exon 6, exon 7, and exon 8. The primary mutation sites of ERBB2 (HER2) were exon 3, exon 17, exon 20, exon 21, exon 22, and exon 26. The primary mutation sites of ARID1A were exon 1, exon 12, exon 18, and exon 20. The primary mutation sites of CDH1 were exon 2, exon 4, exon 8, and exon 16.

HER2 amplification in trastuzumab resistance

Of the enrolled patients, 16 had been treated with trastuzumab, while 9 were currently undergoing trastuzumab treatment. Among the 10 patients showing progression, 9 exhibited secondary resistance (PFS > 3 months), and 6 displayed primary resistance (PFS ≤ 3 months). Figure 6A demonstrates that the HER2 copy number and mutation number correlate strongly with tumor load and outperform CEA and CTC count in predicting tumor shrinkage. Since tumor heterogeneity plays a crucial role in resistance and HER2-amplified subclones are preferentially targeted by trastuzumab, we analyzed the potential progression of trastuzumab-resistant clones (Figure 6B). Among advanced GC patients, HER2 amplification was detected in 10 patients (59%) at both baseline and the PD stage. These patients might develop resistant clones from HER2 amplification, thereby overcoming trastuzumab treatment, which require tailored therapeutic regimens (Figure 6C). In another 4 patients (23%), HER2 amplification was detected at baseline but not during PD, suggesting that trastuzumab treatment might have eliminated HER2 amplification and clones, with resistance driven by other non-HER2 amplified clones (Figure 6D). Tracking of enrolled samples revealed changes in CTC number from Baseline to PR to SD to PD. The results indicated a decrease in CTC number during the transition from Baseline to PD (Figure S3A-C), followed by a slight increase (Figure S3D-F), and then a substantial increase (Figure S3G-I).

Figure 6.

(A) Consistency between tumor load and CEA, HER2 SCNA, Mutation Load, and CTC count indicators; (B) Distribution of potential trastuzumab resistance clones in tumor progression; (C) Copy number of HER2 detected at Baseline, during Treatment and during PD; and (D) Copy number of HER2 detected at Baseline and during PD.

Identification of patients with different resistance by CTC-DNA analysis

The total number of GC CTCs captured by conventional EpCAM magnetic beads and Vimentin magnetic beads was slightly higher than that captured by HER2 magnetic beads (Figure 7A). However, HER2 magnetic beads do have advantages in capturing GC CTCs. With further optimization in relation to variations in sample size and the distribution of enrolled patient stages, the capture efficiency of HER2 lipid magnetic beads for GC CTCs is expected to improve. ERBB2 amplification, along with ERBB2 amplification combined with PIK3CA/R1/C3 mutations, were the most common changes associated with trastuzumab resistance (Figure 7B). Specifically, PIK3CA mutations were more prevalent in patients with innate resistance (2/3, 66.6%) (Figure 7B), suggesting that detecting PIK3CA mutations may be an important biomarker to determine innate resistance.

Figure 7.

A. Number of CTCs captured from blood; B. Changes in ERBB2/4 and PIK3CA/R1/C3.

We reviewed additional clinical samples performing CTC counting and typing, with results summarized in Figure S4. The findings were consistent with the above trend: the number of GC CTCs captured by conventional EpCAM beads and Vimentin beads was slightly higher than that captured by HER2 beads.

Distribution of microsatellite status and gene mutation

CTC-DNA carries information on tumor molecular biology. High-throughput sequencing of CTC-DNA can reflect tumor TMB values and microsatellite stability. Summarized TMB values and gene mutations in CTC-NGS data are summarized in Figure 8. Specific gene mutation results are presented in Figure 8A, 8B, with the frequency of each gene mutation represented by the bar chart’s height. HER2 mutations are further delineated in Figure 8B. Among the evaluated GC patients, 11 (15.3%) were microsatellite instable (MSI), while 60 patients (83.3%) were microsatellite stable (MSS) (Figure 8D). Specific TMB values are shown in Figure 8C.

Figure 8.

A, B. Gene mutations in CTC-NGS sequencing data and specific results of HER2. C, D. Distribution of microsatellite status in CTC-NGS sequencing.

Effect of trastuzumab drug on gastric cancer cells

Given that GC samples with abnormal HER2 expression/mutation are currently targeted for disease remission using trastuzumab, we analyzed the functional impact of trastuzumab on GC cells in NCI-N87 and MGC803. After optimizing trastuzumab concentration, we found that cell proliferation significantly decreased at concentrations exceeding 25 μg/mL, while trastuzumab-resistant cell lines were less affected by trastuzumab concentration (Figure 9A, 9B). We also investigated the effects of trastuzumab on the invasive ability and apoptosis of GC cells at appropriate trastuzumab concentrations. The results revealed that the invasive ability of sensitive GC cells increased while their apoptotic ability decreased (Figure 9C, 9D). In contrast, the invasive ability of resistant GC cells decreased, and their apoptotic ability increased (Figure 9C, 9D). These findings suggest that trastuzumab interferes with GC cells. We also analyzed the proliferative capacity of GC cells in response to two other drugs (pyrotinib and apatinib), as shown in Figure S5. We further assessed changes in the proliferative capacity of GC cells after TPR-HER2, V10920, G10630, D12028, D12028, D12308, and Y12003 point mutations at seven drug concentrations.

Figure 9.

Determination of trastuzumab kill curve for GC cells: (A) NCI-N87, (B) MGC803. The effects of trastuzumab on the invasiveness (C) and apoptosis (D) of GC cells.

Differential gene expression analysis

The workflow for gene expression analysis, including the identification and validation of candidate biomarkers under trastuzumab treatment, is presented in Figure 10A. Regulated genes and biomarker candidates were identified following gene expression analysis. Biomarker candidates were validated in vitro or clinical specimens. The expressions of three biomarkers, ERBB2, ERBB4, and PIK3CA, in GC cell lines (NCI-N87) are depicted in Figure 10B-D. The expressions of these three biomarkers in clinical specimens are shown in Figure 10E.

Figure 10.

A. The workflow for gene expression analysis of candidate biomarkers under trastuzumab treatment. B-D. The expressions of ERBB2, ERBB4 and PIK3CA in NCI-N87 cell lines. E. The expressions of ERBB2, ERBB4 and PIK3CA in GC clinical specimens.

Discussion

Since GC is characterized by genomic heterogeneity, the clonal complexity of patients with multi-metastatic disease cannot be fully assessed by sampling from a single disease site due to its inherent complexity [14]. Recently, “liquid biopsies”, composed partially of CTCs, have emerged as a focal point in cancer research. CTCs hold significant clinical value in monitoring therapeutic responses and prognosticating disease outcomes [15,16]. Consequently, CTC-DNA is instrumental in detecting genomic alterations across heterogeneous genomic sites. This study leverages a GC-CTC sorting system for drug resistance analysis and monitoring disease progression. To enhance GC-CTC enrichment, we constructed a specific HER2-LMB tailored for capturing GC-CTC.

Prior research has demonstrated the utility of CTC-DNA for monitoring resistance and investigating mechanisms in malignant tumors. However, the background noise from cell-free DNA (cfDNA) can limit the detection sensitivity of CTC-DNA [17]. Shoda and his colleagues successfully detected the HER2 copy number using ctDNA-based ddPCR [18]. Subsequently, Wang et al. utilized high-throughput sequencing for ctDNA genotyping, demonstrating the efficacy of this technique in monitoring trastuzumab resistance and pinpointing resistance emergence [13]. Extracting DNA from CTCs and employing next-generation sequencing (NGS) technology offer a promising avenue for dynamic resistance monitoring [19]. At the time of our study, of the patients examined, 16 had undergone trastuzumab treatment, while 9 were currently receiving treatment. Among the 10 patients showing progression, 9 exhibited secondary resistance (PFS > 3 months), and 6 displayed primary resistance (PFS ≤ 3 months). Furthermore, our analysis of high-frequency (> 5%) molecular alterations in tissues and blood revealed that ERBB2 (HER2) and TP53 are the most commonly altered genes, exhibiting pronounced heterogeneity.

The advent of NGS has significantly advanced our understanding of tumor molecular characteristics. Genome projects like the Cancer Genome Atlas Network and the Asian Cancer Research Group (ACRG) have greatly contributed to our molecular understanding of GC [20,21]. Presently, NGS-based CTC-DNA genotyping has been validated as an effective method for monitoring trastuzumab resistance and discerning the resistance mechanisms in HER2+ GC [13,22]. Our findings suggest that CTC-DNA can predict trastuzumab responses, with the persistence or resurgence of HER2-amplified copies in blood signaling tumor progression linked to resistance. In our NGS-based CTC-DNA analysis, we identified multiple resistance mechanisms in 76.5% of patients. Notably, patients exhibiting multiple resistance mechanisms (PIK3CA/R1/C3 and ERBB2/4) experienced a significant reduction in progression-free survival (PFS) post-trastuzumab treatment. These findings underscore the potential of isolated CTC-DNA to effectively reflect the non-heterogeneity of acquired resistance mechanisms.

Recently, CTCs have garnered global attention as a means to surmount the challenges posed by the inherent limitations of tissue samples [23,24]. While CTC-DNA profiling may offer valuable insights into potential molecular changes associated with trastuzumab resistance, this technique has its limitations [25-27]. For instance, although our 610-gene panel encompasses most genes linked to solid tumors, there might be mutations in genes outside our panel. Additionally, gene functions regulated at transcriptional, translational, or post-translational levels might evade detection by DNA sequencing. The low abundance of CTCs in blood also poses challenges, as other cells might be inadvertently captured, potentially skewing detection results. However, the mechanism of CTC capture allows for the identification of proteins along with the co-localization of nucleic acid sequences by staining, demonstrating the techniques’ potential in resistance research. Moreover, our comprehensive analysis of cell biological functions in GC cell lines has provided further insights. By tracking and monitoring trastuzumab resistance in GC and focusing on HER2 (mGC) using CTC-DNA from 20 GC patient samples, we have highlighted the significance of CTC analysis in understanding disease heterogeneity, clonal evolution, and clonal dynamics.

Disclosure of conflict of interest

The author Xiaofei Liang is employed by Jukang (Shanghai) Biotechnology Co., LTD. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.