Marker for predicting sensitivity of gastric cancer immunotherapy and use thereof

Abstract

The application discloses a marker for predicting the sensitivity of gastric cancer immunotherapy and application thereof, and relates to the technical field of biological medicine. The marker comprises GJA1 and CD274, and when the sensitivity of gastric cancer immunotherapy is predicted, the prediction method is based on the difference between the expression levels of GJA1 and CD274. The marker has a synergistic prediction effect that cannot be realized by a single marker, can effectively solve the clinical problem that the existing PD-L1 CPS score cannot accurately distinguish immunotherapy responders from non-responders, provides a reliable basis for clinical medication decision-making, and is simple and easy to operate, low in cost and easy to popularize and apply.

Description

Technical Field

[0001] This invention belongs to the field of molecular biology technology, specifically relating to a biomarker for predicting the sensitivity of gastric cancer to immunotherapy and its application. Background Technology

[0002] Gastric cancer is a common malignant tumor of the digestive tract. Immunotherapy for gastric cancer activates / remodels the body's own immune system, enabling immune cells to actively recognize and kill gastric cancer cells, providing a new treatment option for gastric cancer patients. Tumor cells can disguise themselves and achieve "immune escape" by inhibiting immune cells through immune checkpoints. Immune checkpoint inhibitors (ICIs) can block the immunosuppressive signals of tumor cells against T cells, reactivating the body's anti-tumor immune response to kill tumor cells. Currently, the most widely used ICIs in clinical practice are monoclonal antibodies targeting programmed death-1 (PD-1) and its ligand PD-L1, such as pembrolizumab and nivolumab, which target PD-1. Chemotherapy combined with PD-1 / PD-L1 inhibitors has become the first-line standard treatment for HER2-negative advanced gastric cancer. For HER2-positive patients, the combination of chemotherapy with anti-HER2 targeted therapy and PD-1 / PD-L1 inhibitors has significantly improved the prognosis of some patients.

[0003] However, gastric cancer exhibits significant inter-tumor and intra-tumor heterogeneity in histological type, molecular characteristics, tumor mutational burden, and tumor microenvironment composition. This high degree of heterogeneity severely impacts clinical efficacy, resulting in an objective response rate of only around 20-40%. Against this backdrop, relying on biomarkers for detection to accurately screen potential beneficiaries and to deeply elucidate the molecular mechanisms of immunotherapy sensitivity and resistance have become core issues urgently needing to be addressed in the field of gastric cancer immunotherapy.

[0004] Currently, the most commonly used biomarkers in clinical practice include PD-L1 protein expression level, microsatellite instability (MSI), deficient mismatch repair (dMMR), and tumor mutational burden (TMB). These biomarkers each have their limitations. For example: (1) Regarding PD-L1 protein expression level, clinical studies have shown that even PD-L1-negative patients may still benefit from PD-L1 inhibitor therapy. However, the US Food and Drug Administration (FDA) has approved nivolumab for third-line treatment of advanced gastric cancer without PD-L1 expression, which weakens the reliability of PD-L1 as a screening criterion; (2) Regarding MSI and dMMR, the MSI-H / dMMR status has been proven to be a positive indicator in various solid tumors. A powerful predictor of the efficacy of immunotherapy, a key study showed that the objective response rate of MSI-H gastric cancer patients was as high as 57.1%, while that of microsatellite stable (MSS) patients was only 9.0%. However, MSI-H gastric cancer accounts for only about 10-20% of all gastric cancer patients, and its application scope is extremely limited. About 80-90% of MSS patients lack effective predictive indicators; (3) For TMB, whole exome sequencing is required, the detection cost is high and the detection cycle is long. Its predictive efficacy is limited by complex factors such as the tumor microenvironment, making clinical promotion difficult.

[0005] Other potential biomarkers, such as Epstein-Barr virus (EBV) infection status and dynamic changes in the tumor microenvironment, may also serve as predictive indicators of response to anti-PD-1 immunotherapy. However, EBV-related gastric cancer is usually accompanied by high-density immune cell infiltration and may be highly sensitive to immunotherapy, but the conclusions of different clinical trials vary greatly, and there is no unified consensus on the predictive value of efficacy, which still needs further verification.

[0006] Meanwhile, existing biomarkers are mostly single-indicator detection methods, which are insufficient to offset errors caused by the tumor microenvironment and stromal interference in gastric cancer. Therefore, developing a method for predicting immunotherapy sensitivity that requires less tissue, has low detection costs, is easy to operate, has high predictive accuracy, and is applicable to a wide range of populations has significant clinical value and application prospects.

[0007] GJA1 (Gap Junction Protein Alpha 1) encodes Connexin 43 (Cx43), a core protein constituting intercellular gap junction channels, playing a dominant role in intercellular communication, metabolic coupling, and signal transduction. In gastric cancer tissues, GJA1 is mainly expressed in endothelial cells and tumor-associated fibroblasts, while its expression is significantly reduced or absent in gastric cancer cells. However, current research has not revealed its association with immunotherapy sensitivity. CD274 (Cluster of Differentiation 274) encodes PD-L1, a key molecule mediating tumor immune escape. By binding to PD-1 on the surface of T cells, it inhibits T cell activation, leading to immune tolerance. Therefore, GJA1 participates in the construction of the tumor microenvironment, while CD274 reflects the state of immune escape. Their functions are interrelated yet each has its own emphasis. This invention utilizes the expression characteristics of these two genes to jointly assess the immunotherapy sensitivity of gastric cancer, thereby overcoming the shortcomings of existing single biomarkers and improving predictive accuracy. Summary of the Invention

[0008] To address the problems mentioned in the background section, this invention provides a biomarker for predicting the sensitivity of gastric cancer to immunotherapy and its application.

[0009] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides biomarkers for predicting the sensitivity of gastric cancer to immunotherapy, including GJA1 and CD274.

[0010] Secondly, the present invention provides a method for predicting the sensitivity of gastric cancer to immunotherapy, which uses the aforementioned biomarkers and the prediction method is based on the difference between the expression levels of GJA1 and CD274.

[0011] The prediction method includes the following steps: 1. Obtain biological samples from gastric cancer patients; Tumor tissue samples were collected from gastric cancer patients before they received immunotherapy. These samples could be obtained via endoscopic biopsy or surgical resection, and could be fresh tissue, frozen tissue, or formalin-fixed paraffin-embedded (FFPE) tissue. The samples needed to be pathologically confirmed to have an adequate tumor cell content (recommended tumor cell percentage ≥20%) to ensure the quality and quantity of subsequent RNA extraction.

[0012] 2. Detect the expression levels of the GJA1 and CD274 genes in the samples; Total RNA was extracted from the samples and reverse transcribed into complementary DNA (cDNA). The mRNA expression levels of the GJA1 and CD274 genes were detected using quantitative reverse transcription polymerase chain reaction (qRT-PCR) or RNA sequencing (RNA-Seq), respectively. The specific procedures are as follows: (1) RNA extraction and quality control: Total RNA was extracted using commercial RNA extraction kits (such as the TRIzol method or centrifugation column method), and the RNA purity (A260 / A280 ratio between 1.8 and 2.0) and concentration were detected by spectrophotometer; (2) Reverse transcription: Take an appropriate amount of total RNA (1 μg) and use primers and reverse transcriptase to synthesize the first-strand cDNA; (3) Expression level detection: mRNA levels can be detected by RNA-seq or qRT-PCR.

[0013] 1) RNA-Seq method: a) Construct sequencing libraries from the extracted RNA and perform high-throughput sequencing (e.g., using the Illumina platform). After quality control and alignment to a human reference genome (e.g., hg38), the TPM (Transcripts Per Million) value for each gene is quantified. The TPM values ​​are then logarithmically transformed: X gene = log2(TPM gene + 1).

[0014] b) Calculate the difference between GJA1 expression level and CD274 expression level by directly substituting the gene TPM value into the formula: Δ = log2(TPM_GJA1 + 1) - log2(TPM_CD274 + 1).

[0015] c) The preset threshold is 1.568; if the difference Δ> 1.568, the patient is judged to have low sensitivity to immunotherapy and is predicted to be a non-responder; if the difference Δ≦1.568, the patient is judged to have high sensitivity to immunotherapy and is predicted to be a responder.

[0016] 2) qRT-PCR method (recommended for rapid clinical testing):

[0017] a) The specific primers for GJA1, CD274, and the internal reference gene GAPDH are designed as shown in the table below:

[0018] b) Establish the following reaction system using the BeyoFast™ SYBR Green One-Step qRT-PCR Kit (Beyotime):

[0019] Run the reaction on a real-time quantitative PCR instrument under the following conditions:

[0020] c) Analyze the results using the software provided by the real-time PCR instrument: obtain the cycle threshold (Ct value) for each sample. Using 2... -ΔΔCt The method calculates the standardized expression level of the target gene relative to the internal reference gene.

[0021] d) Calculate the difference between GJA1 expression level and CD274 expression level; first use 2 -ΔΔCt The relative expression levels of GJA1 and CD274 were calculated using the following formula, and the difference Δ was calculated: Δ = GJA1- CD274.

[0022] e) Compare the difference Δ with a preset threshold to determine the patient's sensitivity to immunotherapy; specifically, the preset threshold is 2.550; if the difference Δ > 2.550, the patient is judged to have low sensitivity to immunotherapy and is predicted to be a non-responder; if the difference Δ ≦ 2.550, the patient is judged to have high sensitivity to immunotherapy and is predicted to be a responder.

[0023] Thirdly, the present invention provides the application of a biomarker for predicting the sensitivity of gastric cancer to immunotherapy in the preparation of a detection product, wherein the detection product includes an in vitro diagnostic kit, a software product containing a prognostic / predictive model, or a clinical decision support system.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention provides a biomarker combining GJA1 and CD274, which can effectively solve the clinical problem that the existing PD-L1CPS score cannot accurately distinguish between immunotherapy responders and non-responders, and provides a reliable basis for clinical medication decisions.

[0025] (2) This invention integrates the expression information of GJA1 and CD274 through a specific logarithmic difference calculation method, which produces a synergistic prediction effect that cannot be achieved by a single marker, and has unexpected technical effects.

[0026] (3) The detection method described in this invention is based on mature qRT-PCR technology. The primer sequences used are well-defined and the amplification conditions are clear. It has the advantages of simple operation, low cost and easy promotion and application. Attached Figure Description

[0027] Figure 1 : Expression distribution of GJA1 in drug-resistant and treatment-sensitive groups.

[0028] Figure 2 : Expression distribution of CD274 in the drug-resistant and treatment-sensitive groups.

[0029] Figure 3 : Expression distribution of PD-L1 CPS in drug-resistant and treatment-sensitive groups.

[0030] Figure 4 Expression and predictive analysis of GJA1 in a cohort of gastric cancer patients treated with pembrolizumab.

[0031] Figure 5 : Expression and predictive analysis of CD274 in a cohort of gastric cancer patients treated with pembrolizumab.

[0032] Figure 6 Expression and predictive analysis of GJA1 in a pembrolizumab treatment cohort containing multiple tumor types.

[0033] Figure 7 : Expression and predictive analysis of CD274 in a pembrolizumab treatment cohort containing multiple tumor types.

[0034] Figure 8 : Expression and predictive analysis of GJA1 in a broader cohort including all tumor types treated with nivolumab and pembrolizumab.

[0035] Figure 9 : Expression and predictive analysis of CD274 in a broader cohort of all tumor types treated with nivolumab and pembrolizumab.

[0036] Figure 10 The expression distribution of the combined biomarkers of this invention in the drug-resistant group and the treatment-sensitive group.

[0037] Figure 11 ROC curve comparison of the combined marker of the present invention with that of existing markers.

[0038] Figure 12 Comparison of the DCA decision curves of the combined markers of this invention with those of existing markers.

[0039] Figure 13 The expression distribution of GJA1, CD274, and the difference between GJA1 and CD274 in the drug-resistant and treatment-sensitive groups in clinical data.

[0040] Figure 14 Comparison of ROC curves for GJA1, CD274, and the difference between GJA1 and CD274 in clinical data.

[0041] Figure 15 Comparison of DCA curves for GJA1, CD274, and the difference between GJA1 and CD274 in clinical data.

[0042] In the above figures, GJA1_CD274 represents the difference in expression levels between GJA1 and CD274. Detailed Implementation

[0043] The following section, with specific implementation methods, elaborates in detail on biomarkers for predicting the sensitivity of gastric cancer to immunotherapy and their applications.

[0044] Example 1: Investigating the expression characteristics of GJA1 and CD274 in gastric cancer immunotherapy Pretreatment transcriptome data and clinical outcome data of 45 patients with metastatic gastric cancer treated with pembrolizumab were obtained through the Tumor Immunotherapy Gene Expression Resource (TIGER; http: / / tiger.canceromics.org) (Project No.: STAD-PRJEB25780). This cohort was divided into 12 responders (complete response / partial response, CR / PR) and 33 resistant patients (non-responders, stable disease / progressive disease, SD / PD) according to the Recognition of Solid Tumor Efficacy in the Study of Patients with Solid Tumors (RECIST).

[0045] like Figure 1-2 As shown, the expression distribution of GJA1 and CD274 in the drug-resistant group and the treatment-sensitive group were compared; Figure 3 The expression distribution of PD-L1 CPS is shown in the figure. It can be seen that GJA1 expression is significantly higher in the drug-resistant population than in the drug-sensitive population, while CD274 is mainly significantly highly expressed in the treatment-sensitive population. Figure 1-3 The p-values ​​were all much less than 0.05, indicating that the differences were extremely significant.

[0046] Example 2: Investigating the predictive power of single genes GJA1 and CD274 We evaluated the potential of GJA1 and CD274 as predictive biomarkers for immunotherapy sensitivity using the ROC plotter in the Kaplan-Meier plotter online database (http: / / rocplot.com / immune).

[0047] First, ROC curve analysis was performed separately for both pembrolizumab and non-pembrolizumab in a cohort of gastric cancer patients (n=45, responder=12, non-responder=33). Figure 4-5 As shown in the figure. The results showed that the AUC of CD274 (PD-L1) was 0.833, while the AUC of GJA1 was 0.793, indicating that CD274 is a relatively better predictive indicator in the specific tumor type of gastric cancer.

[0048] Then, to further validate the pan-cancer predictive potential of these two genes, an expanded analysis was performed in a pembrolizumab treatment cohort containing multiple tumor types (n=525, responder=178, non-responder=374), such as... Figure 6-7 As shown, the results indicated that at the pan-cancer level, GJA1's predictive power (AUC=0.675) was significantly superior to CD274 (AUC=0.559); the expanded analysis in a broader cohort including all tumor types treated with nivolumab and pembrolizumab (n=761, responder=257, non-responder=504) also yielded positive results. Figure 8-9 This trend still exists: GJA1's AUC = 0.619, and CD274's AUC = 0.549.

[0049] The above results indicate that while CD274 demonstrates superior predictive efficacy compared to GJA1 in gastric cancer alone, GJA1 shows significantly better predictive efficacy for immunotherapy sensitivity at the pan-cancer level. This finding suggests that GJA1 may be a more widely applicable predictive biomarker for immunotherapy, while CD274's predictive efficacy may exhibit stronger tumor-specificity. Therefore, combining GJA1 and CD274 could potentially achieve synergistic and complementary predictive effects across different tumor types, leading to a more robust predictive model.

[0050] Example 3: Evaluating the predictive power of combined biomarkers Based on Examples 1 and 2, a combined marker was constructed, denoted as the difference between GJA1 and CD274, and the calculation formula is as follows: Δ= log2(TPM_GJA1 + 1) - log2(TPM_CD274 + 1); The expression levels of both genes were normalized using log2(TPM+1) to ensure comparability of expression levels between different samples and different genes.

[0051] The following methods are used to evaluate the predictive efficacy of the combined biomarkers of the present invention: (1) To assess the expression differences of the combined biomarkers in the sensitive and resistant groups, STAD-PRJEB25780 gastric cancer transcriptomics data were analyzed. The results are as follows: Figure 10 As shown, the results indicate that the expression level of the combined biomarker in the drug-resistant group was significantly higher than that in the sensitive group (Wilcoxon rank-sum test, p < 0.05), demonstrating that the combined biomarker of the present invention can serve as an effective indicator for distinguishing the response status of immunotherapy.

[0052] (2) To comprehensively evaluate the predictive performance of the combined biomarkers, a head-to-head comparison was performed with CD274, GJA1, and commonly used clinical predictive biomarkers (TMB, MSI status, PD-L1 CPS). The AUC comparison results are shown in Table 1, and the ROC curve analysis results are shown in Table 2. Figure 11 As shown.

[0053] Table 1. Comparison of AUC between the combined marker of the present invention and existing markers.

[0054] Note: AUC_lower represents the lower limit of the 95% confidence interval for AUC; AUC_upper represents the lower limit of the 95% confidence interval for AUC.

[0055] The results showed that the predictive efficacy of the combined biomarkers of this invention (AUC=0.891) was comparable to that of PD-L1 CPS (AUC=0.910), and significantly superior to MSI status (AUC=0.693) and TMB (AUC=0.798), with statistically significant differences. Furthermore, compared to single genes, the predictive efficacy of the combined biomarkers was superior to that of GJA1 (AUC=0.814) and CD274 (AUC=0.838) as independent biomarkers, indicating a synergistic effect. Therefore, the combined biomarkers of this invention combine the robust predictive ability of GJA1 at the pan-cancer level with the specific predictive advantage of CD274 in gastric cancer, and are expected to overcome the limitations of single biomarkers.

[0056] (3) To evaluate the clinical net benefit and application value of combined biomarkers, decision curve analysis (DCA, which quantifies the net benefit at different threshold probabilities to more comprehensively reflect the actual value of biomarkers in clinical decision-making) was used to compare multiple biomarkers, such as... Figure 12 As shown in the figure. The results show that, within the clinical decision-related range of threshold probabilities of 0.1-0.5, the net benefit of the combined biomarker of the present invention is significantly better than that of PD-L1 CPS, indicating that the combined biomarker of the present invention has higher practical value in providing guidance for clinical treatment decisions; at the same time, the combined biomarker of the present invention can maintain a high clinical net benefit over a wide range of threshold probabilities, indicating that it has good adaptability to clinicians with different decision preferences.

[0057] (4) To apply the combined biomarkers of this invention to clinical decision-making, various methods were used to analyze the cutoff values ​​of the ROC curves, including: maximizing the Youden index and minimizing the top left corner distance (closest.topleft), and manual traversal calculations were performed. The results showed that the above three methods obtained completely consistent cutoff values ​​in the gastric cancer cohort (STAD-PRJEB25780), verifying the robustness of the threshold. When the threshold of the combined biomarkers of this invention (Δ= log2(TPM_GJA1+1)- log2(TPM_CD274+1)) was set to 1.568 (rounded from 1.567730028), the Youden index reached its maximum value of 0.742, corresponding to a sensitivity of 83.3% (i.e., true positive rate) and a specificity of 90.9% (i.e., true negative rate). The calculation results of each method are summarized in Table 2.

[0058] Table 2 Comparison of optimal cutoff values ​​calculated by different methods

[0059] These results indicate that when the difference between GJA1 and CD274 in a patient's tumor tissue exceeds 1.568 (after standardization based on log2(TPM+1)), the predictive specificity for immunotherapy resistance can reach over 90%, demonstrating high clinical diagnostic value. In practical applications, this threshold can be used as a reference cutoff value for assessing the sensitivity of gastric cancer patients to immunotherapy: if Δ > 1.568, the patient is considered to have low sensitivity to immunotherapy and is predicted to be a non-responder; if Δ ≤ 1.568, the patient is considered to have high sensitivity to immunotherapy and is predicted to be a responder.

[0060] The above results suggest that the predictive model based on the difference between GJA1 and CD274 using the combined biomarkers of this invention not only has high discriminative power (AUC), but also has superior clinical decision support value, and has good prospects for clinical application and translation.

[0061] Example 4: Clinical Data Validation This study retrospectively collected clinical data from 30 gastric cancer patients who received pembrolizumab immunotherapy. Based on RECIST 1.1 criteria, 15 patients who achieved complete remission (CR) or partial remission (PR) were included in the sensitive group, and 15 patients who achieved stable disease (SD) or progressive disease (PD) were included in the resistant group (i.e., primary resistance).

[0062] All specimens were formalin-fixed and paraffin-embedded (FFPE) tissues; pathological review confirmed that the proportion of tumor cells exceeded 20%. Total RNA was extracted from the tissues, and the mRNA expression levels of GJA1 and CD274 were detected by qRT-PCR after reverse transcription.-ΔΔCt The relative expression levels of each were calculated, and the difference Δ = GJA1 - CD274 was calculated.

[0063] The results show that: Figure 13 The expression level of GJA1 in the drug-resistant group was significantly higher than that in the sensitive group, and the expression level of CD274 in the sensitive group was significantly higher than that in the drug-resistant group. The difference between GJA1 and CD274 was also significantly higher in the drug-resistant group, and the difference between the groups was more obvious than that of a single indicator.

[0064] Further evaluation of the predictive efficacy of three indicators for the efficacy of immunotherapy in gastric cancer was conducted using ROC curves, such as... Figure 14 As shown in Tables 3-4, the results indicate that the AUC of the difference between GJA1 and CD274 is 0.929, which is higher than that of GJA1 (AUC=0.724) and CD274 (AUC=0.893).

[0065] The Delong test showed that the AUC of the difference between GJA1 and CD274 was significantly better than that of GJA1 (P=0.0289); however, it did not reach a statistically significant difference compared with CD274 (P=0.5596). However, considering that the sample size of this validation cohort was small (n=30) and the test power was limited, and that the sensitivity (0.867) and specificity (1) of the difference between GJA1 and CD274 were both better than those of CD274 (sensitivity was 0.8 and specificity was 0.933), it can be seen that the overall performance of the difference between GJA1 and CD274 was better.

[0066] In addition, such as Figure 15 As shown, decision curve analysis (DCA) also suggests that the difference between GJA1 and CD274 has a better net clinical benefit.

[0067] Table 3. Predictive efficacy of various indicators for sensitivity to immunotherapy in gastric cancer during clinical trials.

[0068] Note: AUC is the area under the curve.

[0069] Table 4. Pairwise comparisons of AUC on ROC curves (Delong test)

[0070] Clinical data show that the combined biomarker based on the difference between GJA1 and CD274 in this invention has excellent predictive ability and has practical application prospects.

[0071] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A biomarker for predicting the sensitivity to immunotherapy in gastric cancer, characterized in that: Including GJA1 and CD274.

2. A method for predicting the sensitivity of gastric cancer to immunotherapy, characterized in that: The biomarker as described in claim 1 was used, and the prediction method was based on the difference between the expression levels of GJA1 and CD274.

3. The method as described in claim 2, characterized in that: Includes the following steps: (1) Obtaining biological samples from gastric cancer patients; (2) Detect the expression levels of GJA1 and CD274 in the samples; (3) Calculate the difference in expression levels between GJA1 and CD274; (4) Compare the difference with the preset threshold to determine the sensitivity of gastric cancer patients to immunotherapy.

4. The method as described in claim 3, characterized in that: The expression levels of GJA1 and CD274 in the sample were detected using RNA-Seq. The difference was calculated using the formula: Δ = log2(TPM_GJA1 + 1) - log2(TPM_CD274 + 1).

5. The method as described in claim 4, characterized in that: In step (4), the preset threshold is 1.568; if the difference Δ>1.568, the patient is judged to have low sensitivity to immunotherapy and is predicted to be a non-responder. If the difference Δ≦1.568, the patient is considered to have high sensitivity to immunotherapy and is predicted to be a responder.

6. The method as described in claim 3, characterized in that: The expression levels of GJA1 and CD274 in the sample were detected by qRT-PCR. The difference was calculated using the formula: Δ = GJA1 - CD274.

7. The method as described in claim 6, characterized in that: In step (4), the preset threshold is 2.550; if the difference Δ>2.550, the patient is judged to have low sensitivity to immunotherapy and is predicted to be a non-responder; if the difference Δ≦2.550, the patient is judged to have high sensitivity to immunotherapy and is predicted to be a responder.

8. The application of the biomarker for predicting the sensitivity of gastric cancer to immunotherapy as described in claim 1 in the preparation of a detection product, characterized in that: The testing products include in vitro diagnostic reagent kits, software products containing prognostic / predictive models, or clinical decision support systems.

Images