HDS screening with patient-derived primary cells guided individualized therapy for esophageal squamous cell carcinoma–in vivo and vitro

Objective To analyze and evaluate the role of the High-throughput Drug Sensitivity (HDS) screening strategy in identifying highly sensitive drugs against esophageal squamous cell carcinoma (ESCC). Methods A total of 80 patients with progressive ESCC were randomly divided into the observation (40 cases) and the control groups (40 cases). In the observation group, primary ESCC cells were isolated from the tumor tissues with a gastroscope, and drug sensitivity screening was performed on cells derived from the 40 ESCC cases using the HDS method, followed by verification in a patient-derived tumor xenograft (PDX) mouse model. Finally, the differences in the therapeutic efficacy (levels of CEA, CYFRA21-1, SCCA after chemotherapy and the rates of overall survival, local progression, and distant metastasis at 12 months and 18 months time points after chemotherapy) were compared between the observation group (Screened drug-treated) and the control group (Paclitaxel combined with cisplatin regimen-treated). Results Forty ESCC patients were screened for nine different high-sensitive chemotherapeutics, with the majority showing sensitivity to Bortezomib. Experiments on animal models revealed that the tumor tissue mass of PDX mice treated with the HDS-screened drug was significantly lower than that of the Paclitaxel-treated mice (p < 0.05), and the therapeutic efficacy of the observation group was better than the control group (p < 0.05). Conclusion HDS screening technology can be beneficial in screening high-efficacy anticancer drugs for advanced-stage ESCC patients, thereby minimizing adverse drug toxicity in critically ill patients. Moreover, this study provides a new avenue for treating advanced ESCC patients with improved outcomes.


Introduction
Esophageal carcinoma (EC) is a highly prevalent and dangerous malignancy of the gastrointestinal (GI) tract (1,2), accounting for the seventh most common cancer worldwide as well as ranking sixth for cancer-related deaths (3). EC claims approximately 450,000 lives annually (4), with a 5-year survival rate of only 18% (5). Notably, the incidence of EC is increasing rapidly in several countries, including China, in recent times (6,7). In China, EC causes about 150,000 annual deaths, which equals 21.8% of all cancer deaths, ranking fourth in the list of deadliest carcinomas. The etiology of EC is complex, many factors such as drinking, smoking, diet, genetics, microorganisms and gender have been reported to be associated with elevated risk of EC by previous studies (5)(6)(7). Therefore, it is technically challenging and difficult to reduce the occurrence of EC. EC patients primarily undergo endoscopic diagnosis for GI symptoms. Esophagectomy is the major surgical intervention in EC patients, while other treatment modalities involve chemo alone or a combination of chemo and radiotherapy, depending on the clinical characteristic of the cancerous lesion (8, 9). Most often, patients are diagnosed with an advanced stage EC, leaving no option for radical surgery. Importantly, postoperative recurrence and metastasis are frequently detected in surgical EC patients, hence, chemotherapy is considered a potentially effective second line of treatment (10). Interestingly, there is a wide variation in EC tumor classification in different geographical regions, such that >90% of EC cases in Western countries are adenocarcinomas, while esophageal squamous cell carcinoma (ESCC) is the most prevalent (>90% of all cases) among the Chinese populations (11). Therefore, this wide range of variabilities hinders the universal application of the National Comprehensive Cancer Network (NCCN) guidelines to Chinese ESCC patients, directly affecting their treatment outcomes and tumor recurrence. Moreover, establishing an in vitro screening strategy for determining patient-specific sensitivity of chemotherapeutics may resolve the discrepancies in the generalized application of NCCN guidelines.
High-throughput drug sensitivity (HDS) screening strategy is a new method with the advantages of high efficiency, wide coverage, and potential adaptation to personalized therapy, which are promising for determining the antitumor drug sensitivity in liver cancer, head and neck cancer, and lymphomas (12)(13)(14). However, the application of this technique in antitumor drug screening of ESCC is rarely reported. Therefore, this study aimed to generate ESCC patient-derived cell models for HDS analyses and develop a patient-derived xenograft (PDX) model of human ESCC in rodents for in vivo investigation. This study provides optimized assay platform able to offer drug screening outcomes in ESCC samples with varying clinical features and guides the precision therapy for ESCC patients (see Figure 1).

Patient selection
A total of 80 outpatient and inpatient candidates with progressive ESCC were recruited from the Department of Gastroenterology of the 901 Hospital of the Joint Logistics and Security Forces of the Chinese People's Liberation Army from December 2018 to December 2019 and randomly assigned to either the observation group or the control group (n = 40 per group). There were no statistical differences in terms of age distribution, male-to-female ratio, tumor characteristics, and patient's Karnofsky score (15) between the two groups (p > 0.05). Patients' inclusion criteria were: (1) compliance with the diagnostic criteria of progressive ESCC (16); (2) availability of complete clinical data, and (3) no indication for surgery. On the other hand, patients were excluded if they had: (1) cardiovascular diseases; (2) penetrating ulcers; (3) hepatic and renal insufficiency; (4) coagulation disorder, and (5) severe immune system diseases. The study was approved by the Institutional Ethics Committee (Ethics Committee of the 901th Hospital of the Joint Support Force), and all participants provided their signed informed consent by themselves or their family members (Chinese clinical trial registration number: ChiCTR2300068566) (see Table 1).

Experimental animals
Severely immunodeficient (NCG) six-week-old female mice were purchased from Nanjing Model Animal Research Institute. The animal experiment protocol was approved by the Institutional Animal Care and Use Committee (Ethics Committee of the 901th Hospital of the Joint Support Force).
The main instruments and reagents that were used in this study are listed in Table 2.

Experimental methods
ESCC primary cell lines were derived from biopsied patient tumor tissue specimens and subsequently characterized as follows: (1) Gastroscopy and biopsy of ESCC tumors were performed using an Olympus CLV-260 electronic gastroscope. Hematoxylin and eosin (H&E) staining and immunohistochemical (IHC) analysis using anti-pCK, anti-p63, anti-Ki-67, and anti-Vim antibodies indicated that tumor tissues contained abundant ESCC cells (Figure 2A). (2) Isolation of ESCC primary cells was performed in the following steps: tissue washing, digestion, centrifugation, collection and resuspension of the cell pellet in the basal medium, centrifugation, removal of the supernatant and lysis of blood cells in appropriate buffer, and final collection of ESCC primary cells by centrifugation.  1 min, aspirated, and then 500 μL of 0.05% trypsin solution was added to each well and placed in an incubator at 37°C with 5% CO 2 for 10 min until complete cell detachment was observed under the microscope. After centrifugation at 1500 rpm for 4 min, the supernatant was discarded, and 500 μL of ESCC primary cell medium was added to resuspend and plate the cells ( Figure 2B and Figure 3C). Flowchart showing the entire work procedure of this study including sections of study object, core technology, animal experiment and clinical test. ESCC, esophageal squamous cell carcinoma; PDX, patient-derived xenograft; HDS, High-throughput drug sensitivity.

HDS screening for drug sensitivity in patient-derived ESCC lines
The patient-specific primary ESCC cells were separately treated with different first-line and second-line drug regimens as well as subjected to the FDA drug library screening using the HDS method.
Steps in this process include: (1) Preparation of cultured cells on petri dishes: ESCC primary cells were harvested by trypsin-digestion, centrifuged, and resuspended in DPBS buffer for cell counting by an image cytometer. The cell suspension was then diluted to 6 × 10 4 -1 × 10 5 cells/ml, and seeded into 384-well plates using 50 μL of cell suspension per well. (4) Cell viability testing: 72 h post-treatment, the cell viability in different treatment and control groups was measured using a plate reader-based CellTiter Glo assay and calculated by the formula, Cell viability (%) = luminescence value of the spiked wells/luminescence value of the control wells × 100%. Cell inhibition (%) = 100-Cell viability ( Figure 3A and Table 3).
In vivo validation of drug effects in PDX mouse model

Model construction
Approximately 5 × 10 6 ESCC progenitor cells were injected into the ESCC fat pad and the axillary area of the right forelimb of 6-weekold female NCG mice. The volume and growth rate of ESCC tumors in mice were routinely recorded every 3 days. Tumor formation was observed at both graft sites on day 15 post-administration. Tumor proliferation was evident from day 15 to day 30, indicating a successful model establishment.  Frontiers in Medicine 05 frontiersin.org

Dose optimization
The doses of the screening drugs were set at 25, 50, and 100 mg/kg/d. After the model was successfully constructed, the drug treatment group mice were administered with designated drug regimens either by oral gavage or intraperitoneal (i.p) injection for 20 days, while paclitaxel (70 mg/kg/d) (17) treated mice served as the positive control group, and vehicle (DMSO)-treated mice were included in the sham group.

Treatment protocol for ESCC subjects in the observation and control groups
Forty ESCC patients in the control group were treated with paclitaxel combined with a cisplatin regimen (18). On day 1, 135 mg/m 2 of paclitaxel injection (Jiangsu Aosikang Pharma., State Drug Quantification H20083848; 2 mL: 30 mg) was added to 500 mL of glucose injection (Jiangsu Hengrui Pharma., State Drug Quantification H32022368; 250 mL: 12.5 g) and diluted into intravenous (i.v) drip for 3 h/d. On days 2-4, subjects received i.v injections of 50 mg/m 2 of cisplatin (Dezhou Deyao Pharma., State Drug Quantification H20023236; 20 mg), dissolved in 500 mL of 0.9% injection-grade sodium chloride solution (Shijiazhuang IV Pharma., State Drug Quantification H13023200; 500 mL) for 2 h each time in a stretch of 3 weeks. The treatment lasted six cycles.
On the other hand, 40 ESCC patients in the observation group were treated according to screened therapeutic agents.
We compared several parameters between these two groups, such as the sensitivity of different anticancer drugs in ESCC cells by the HDS method, the difference in tumor growth curves before and after drug therapy in the PDX mice as well as relative expression levels of tumor markers, including carcinoembryonic antigen (CEA), serum cytokeratin protein fragment 19 (CYFRA21-1) and squamous epithelial cell carcinoma antigen (SCCA) in the ESCC patients. Furthermore, we compared the differences in overall survival (OS) rates, local progression (LP) rates, and distant metastasis (DM) rates between the two groups of patients at 12 and 18 months after chemotherapy. OS rate = number of survivors/total number × 100%. LP rate = number of local tumor diameter increase/total number × 100%. DM rate = number of new distant metastases/total number × 100%.
Statistical analysis SPSS v22.0 software (IBM Corp., Armonk, NY, United States) was used for all statistical analyses. Measured and counted data were expressed as xs and %, respectively. The t-test was used to compare the differences in expressions of CEA, CYFRA21-1, and SCCA in ESSC patients before and after treatment with different drugs as well as tumor volumes and masses in PDX models under various treatment conditions. Least significant difference (LSD) method was used for multiple comparisons. Chi-square (χ 2 ) test was used to compare the rate of local progression and distant metastasis between the two groups. A p-value of less than 0.05 was considered statistically significant.

Results
Identification of highly effective drugs against patient-derived ESCC cell models by HDS method Forty ESCC patients were subjected to treatment with 9 different highly sensitive chemotherapeutic agents, with the largest number of patients exhibiting significant sensitivities to bortezomib, as described in Table 4.

HDS-screened drugs show inhibition of tumor growth in the PDX mice
We found that HDS-screened drug-treated PDX mice had significantly smaller tumor Volumes and masses compared to that in the vehicle-treated sham (p < 0.05), and Paclitaxel treatment groups, as illustrated in Figure 4.

Treatment with antitumor agents alleviates the expression of tumor markers in ESCC patients
There were no statistical differences in the expression of CEA, CYFRA21-1, and SCCA between the two groups at the baseline (t = 0.327, 0.088, 0.276; p = 0.745, 0.930, 0.783), while significantly lower expression levels of these protein markers were detected in the observation group than those in the control group (t = 5.692, 5.748, 10.817; p < 0.001) after chemotherapy, as shown in Table 5.

HDS-guided treatment enhanced the clinical efficacy of chemotherapeutics
We found that the OS rates at 12 months and 18 months were significantly higher in the observation group than those in the control group (p < 0.05), while rates of LP and DM at 18 months were significantly lower in the observation group than those in the control group (p < 0.05), as shown in Table 6.

Discussion
Since the majority of patients with progressive ESCC are diagnosed at the advanced stage, most clinicians adopt palliative radiotherapy, chemotherapy, biological therapy, and traditional Chinese medicine (TCM) therapies (19-22). Notably, due to extensive heterogeneity in tumor characteristics across ESCC patients, individualized chemotherapeutic strategies are urgently needed to achieve the best therapeutic outcome and minimum drug toxicity (23, 24). Most clinical researchers exploit the patient-derived tumor cellderived PDX rodents to assess the cell growth inhibitory effect of different regimens of anticancer drugs to overcome the problem of adverse drug toxicity in critically ill ESCC patients (25,26). However, these in vitro testing models have their limitations, such as high cost Recently, high-throughput gene sequencing (HGS) technology has been used for the targeted therapy of ESCC. Some groups have reported the genetic landscape of human ESCCs using the whole genome and exome sequencing approaches (29)(30)(31), which has led to promising results in early studies of personalized and targeted therapies (32). However, due to an increase in the oncogenic mutational burden, differences in cellular signaling pathways, and complex interactions among tumor genes, the same tumor gene target may respond differentially to different drugs affecting the overall outcome of individualized therapy (33).
In contrast, the HDS technology used by our group extended the scope of screening with a large panel of FDA-approved drugs for determining the best possible drug regimen exerting maximum cell growth inhibitory effects. With this technically upgraded screening platform, we can rapidly identify highly sensitive and targeted chemotherapeutics for different ESCC patients, thereby solving most of the previously mentioned technical hurdles.
The primary aim of this study was to obtain high-quality primary tumor cells from ESCC patients and subsequently exploit them for in vitro drug sensitivity screening strategies as an alternative to in vivo models. We used fresh tumor tissue masses to isolate primary ESCC cells through trypsin digestion with a success rate as high as 90%, which was considerably higher than that reported in other studies (34). The HDS technique developed by our group could simultaneously measure the sensitivity of more than 300 drugs to patient-derived ESCC tumor cells, along with the screening of optimal drug concentrations to reduce the adverse drug toxicity effects. In this study, primary tumor cells from 40 ESCC patients were tested for drug sensitivity against 9 different chemotherapeutic agents, which fully illustrated the inter-patient heterogeneity of tumor cells (35). Moreover, the results of this study further signify the importance of individualized chemotherapeutic regimens in treating highly heterogeneous carcinomas like ESCC. Through this screening assay, we noticed that most ESCC patients were sensitive to Bortezomib.
Bortezomib is a 20S proteasome inhibitor that exerts its antitumor effects by interfering with cell signaling, causing cell cycle arrest, inducing apoptosis, and inhibiting angiogenesis and is currently used mainly for the treatment of multiple myeloma (36, 37). However, one study (38) reported its role in enhancing the sensitivity of ESCC to radiotherapy. Besides, there are few a studies on the application of Bortezomib in chemotherapy for advanced-stage ESCC (39). Here, we found that Bortezomib was effective in chemotherapy for esophageal squamous carcinoma by in vitro HDS assay, which might provide a new avenue for the treatment of ESCC.
To further verify the results of in vitro experiments, the previously reported PDX model (40) was used to determine the therapeutic effects of screened anticancer drugs in vivo. We exploited patient-derived primary ESCC cells for constructing the PDX mouse model to screen potential anticancer drugs. In addition, we used Paclitaxel (41) as the positive control in this animal study. The results showed that the tumor tissue mass and volume of PDX mice treated with screened    high-sensitivity drugs were significantly smaller than those treated with Paclitaxel or the vehicle control (p < 0.05). Finally, we determined the optimal drug concentration to achieve the best possible tumor growth inhibition in this PDX mouse model. After achieving encouraging results in animal experiments, we conducted individualized treatment for 40 ESCC patients with Bortezomib and compared the anticancer efficacy with 40 ESCC patients who received Paclitaxel combined with cisplatin chemotherapy in the same period of time. We first analyzed the differences in expressions of tumor markers CEA, CYFRA21-1, and SCCA between the two groups. CEA is one of the first tumor markers found to be widely present in tumor cells. CYFRA21-1 is also commonly detected in lung and esophageal tumor cells and exhibits a high sensitivity along with the diagnostic value in the evaluation of therapeutic effect against ESCC (42). SCCA is mainly present in tumor cells of the lung, pharynx, and esophagus and exhibits a high sensitivity for occurrence and progression of ESCC with high specificity (43). Some studies (44)(45)(46) have reported that the combination of these three biomarkers could be a valuable prediction tool for the precise diagnosis of EC and for predicting the disease progression rate as well. Here, we found that CEA, CYFRA21-1, and SCCA levels were significantly lower than those in the control group after chemotherapy treatment in the observation group (p < 0.05), suggesting that the HDS-screened anticancer drug was effective in controlling the ESCC tumor growth in the observation group. Furthermore, we analyzed the treatment efficacies in the two groups based on clinical symptoms and found that the 12-month and 18-month OS rates in the observation group were significantly higher than those in the control group (p < 0.05), while rates of LP and DM were significantly lower than those in the control group (p < 0.05), indicating promising efficacy of HDS-screened drugs against ESCC.
In summary, the HDS method is an efficient, reproducible, and easy-to-use drug sensitivity screening tool that can expedite the ESCC diagnosis, especially for terminally ill patients. Moreover, drugs for antitumor sensitivity on the HDS platform consistently exhibited significant inhibitory effects on ESCC tumor growth in disease animal models as well as human ESCC patients.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary materials, further inquiries can be directed to the corresponding authors.

Ethics statement
The studies involving human participants were reviewed and approved by Ethics Committee of the 901th Hospital of the Joint Support Force. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by Ethics Committee of the 901th Hospital of the Joint Support Force.