Primary liver cancer, as a common malignant tumor, is the third leading cause of cancer-related death worldwide, and hepatocellular carcinoma accounts for 75-85% of all cancer-related deaths (1). Surgery is the main method for early-stage liver cancer. However, due to the insidious onset of liver cancer, most patients are in the middle and late stages when they are initially diagnosed, and they miss the opportunity of surgical operation. Moreover, HCC has a high recurrence rate after surgery, with a total recurrence rate of about 70% within 5 years, and most patients lose the chance of reoperation after recurrence (2). Therefore, systemic anti-tumor therapy, especially combination therapy based on immune checkpoint inhibitors (ICIs), is highly recommended. It has become the most used and the most important treatment for unresectable liver cancer. The combination therapy of PD-L1 inhibitors with antiangiogenic agents, specifically the atezolizumab plus bevacizumab regimen, has been recommended by major clinical guidelines as first-line treatment for advanced HCC patients (3). Notably, the immunotherapeutic combination of the PD-1 inhibitor Opdivo (nivolumab) and the CTLA-4-targeting antibody Yervoy (ipilimumab) represents the first approved dual immune checkpoint inhibitor regimen for the management of advanced HCC populations (4). ICIs for the treatment of HCC mainly include cytotoxic T lymphocyte-associated antigen (CTLA) monoclonal antibodies. Inhibitors of programmed death 1 (PD-1) and its ligand PD-L1, which kill tumor cells and inhibit their proliferation by reactivating the immune response of T cells to tumors (5–7). However, only 20 to 40% of cancer patients respond to immunotherapy (8). In a study of Asian HCC patients treated with a single PD-1 immunoagent such as nivolumab, pembrolizumab, or camrelizumab, the objective response rate (ORR) was found to be only approximately 15%, suggesting that the complex immunosuppressive microenvironment of HCC leads to ICIs evasion by unknown mechanisms (5, 9, 10). Their use has been hampered by limited response rates and a lack of predictive markers for clinical response (11, 12). Elucidation of the mechanisms underlying the immunosuppressive microenvironment and subsequent remodeling to guide rational combination therapy remains a major challenge for therapeutic intervention in HCC patients (13). There is an urgent need to overcome the intrinsic or adaptive resistance of HCC to immunotherapy.

Gut microbiota has been shown to play a regulatory role in the response to tumor immunotherapy (14). In recent years, several authoritative studies have shown that the number, type and composition of intestinal flora in cancer patients are closely related to the efficacy and survival of these patients treated with PD-1 inhibitors. The possible principle is that intestinal flora regulates the tumor microenvironment through microbial signals, thereby affecting the efficacy of immunotherapy (15–18). In 2015, two Science papers published startling results showing that gut microbiota plays a decisive role in the response of immunotherapy in mouse models (19, 20). At present, mouse models of different cancers have been studied to enhance the efficacy of ICIs after Fecal bacteria transplantation (FMT), including colorectal cancer (21), malignant melanoma (22), and renal cancer (23). However, in the clinical stage, FMT only achieved good results in patients with malignant melanoma: In 2021, FMT was tried in patients with immune-resistant malignant melanoma, and more than one third of patients experienced a re-response, using FMT from an effective population (22). In another 2023 study, patients with melanoma who received fecal microbiota from a healthy person had an approximately 20% improvement in response rate to first-line immunotherapy, with an objective response rate of 65% (24). In addition, in NSCLC patients, novel prebiotics have been tried to enhance the response to anti-PD-1 immunotherapy in NSCLC patients (25). In HCC, studies have found the relationship between the species diversity and abundance of gut microbiota and the clinical response and adverse effects of immunotherapy (26). Based on the above evidence, we can hypothesize that gut microbiota transplantation can also improve the efficacy of liver cancer immunotherapy.

Based on long-term experimental studies, we have identified specific gut bacteria associated with tumor immunity. In melanoma-bearing mice, the combination of oral Bifidobacterium (containing Bifidobacterium breve and Bifidobacterium longum) with PD-L1 inhibitors reduced tumor size by 80%, whereas PD-L1 inhibitors alone achieved only a 40% reduction (18). In melanoma patients, fecal samples from immunotherapy responders showed enrichment of probiotics like Bifidobacterium. Transplantation of these microbiota into mice enhanced the efficacy of immunotherapy (27). For Chinese non-small cell lung cancer patients receiving Bifidobacterium breve supplementation during immunotherapy, the objective response rate (ORR) reached 40%, disease control rate (DCR) 90%, and median progression-free survival (PFS) exceeded 500 days - significantly higher than in Bifidobacterium breve-negative patients (28). Bifidobacterium-derived signals were found to stabilize dendritic cell (DC) activation, thereby improving tumor-specific CD8⁺ T cell effector functions (19). Lactobacillus reuteri has also demonstrated immunotherapeutic enhancement. In melanoma mice, L. reuteri transplantation combined with PD-L1 inhibitors resulted in an additional 60% tumor reduction. In advanced melanoma patients, serum levels of the L. reuteri metabolite indole-3-aldehyde (I3A) significantly correlated with treatment outcomes, with high-I3A patients showing median PFS >50 months versus <10 months in low-I3A groups (29). Another species, Lactobacillus johnsonii, was shown to metabolize hypoxanthine and stimulate immunity via adenosine receptors. In colorectal cancer mice, L. johnsonii combined with CTLA-4 inhibitors reduced tumor volume by 85% compared to CTLA-4 inhibitors alone (30). Other immunomodulatory Lactobacillus species including L. casei and L. plantarum have also demonstrated potential for enhancing cancer immunotherapy (31). Streptococcus thermophilus exhibited anti-tumor activity through β-galactosidase secretion, significantly suppressing tumorigenesis in animal models. Notably, this effect disappeared when β-galactosidase-related genes were knocked out (32). Additional studies revealed that oral Lactobacillus rhamnosus GG enhanced anti-PD-1 efficacy by increasing tumor-infiltrating DCs and T cells (33).

The TSME used in this study incorporates these validated tumor-immunomodulatory gut bacteria and immune potentiators. Its formulation includes Bifidobacterium longum, Lactobacillus rhamnosus, Bifidobacterium animalis, Bifidobacterium adolescentis, Lactobacillus reuteri, Lactobacillus casei, Streptococcus thermophilus, Bifidobacterium bifidum, and Lactobacillus acidophilus. TSME has obtained China Food Standards certification (Food Production License SC10632117100037), ensuring safety for oral administration and clinical application.

This study aims to investigate whether TSME combined with PD-1 inhibitors can enhance cancer immunotherapy efficacy, particularly in immunotherapy-resistant patients.

This research complies with all relevant ethical regulations approved by the Institutional Animal Care and Use Committee (IACUC) of Huazhong University of Science and Technology.

This Investigator-initiated Trial (IIT) was performed under the guidance of Medical Ethics Committee of Taixing People’s Hospital (ethical lot. LS2023019). The subject has signed the informed consent form.

This study is the first to demonstrate the synergistic antitumor effects of TSME combined with PD-1/PD-L1 inhibitors in immunotherapy-resistant HCC. In animal experiments, TSME significantly enhanced the efficacy of αPD-1 monoclonal antibody by remodeling the tumor immune microenvironment (TME): The combination therapy group exhibited increased infiltration of CD8⁺ and CD4⁺ T cells ( Figures 2a, b ) and reduced proportions of immunosuppressive Treg cells ( Figure 2c ). This immune reprogramming correlated with a tumor growth inhibition (TGI) rate of 58.78%, surpassing monotherapy outcomes ( Supplementary Figure S3 ). Clinically, a patient with advanced HCC refractory to multiple therapies achieved a PFS of 7 months and alleviated symptoms (abdominal distension, fatigue) after TSME combined with atezolizumab treatment ( Figures 3 , 4 ). Despite eventual disease progression, the patient maintained a stable quality of life, suggesting that TSME may reverse immunotherapy resistance through multispecies synergy.

In this clinical case, the patient received carilizumab therapy for 17 months until disease progression, followed by combination therapy with atezolizumab and bevacizumab. Upon recurrence, the patient was treated with tegafur-gimeracil-oteracil (S-1) for 3 months. Subsequent recurrence prompted combination therapy with oral TSME, atezolizumab, and bevacizumab, which demonstrated favorable clinical efficacy. However, the potential delayed pseudo-progression effect of atezolizumab must be ruled out. Existing studies report varying pseudo-progression rates across tumor types: 2.78%-9.69% in melanoma, 1.81%-5.77% in non-small cell lung cancer, and 2.86%-8.82% in renal cell carcinoma, while HCC has limited documented cases (34). Pseudo-progression typically occurs within the first few weeks of immunotherapy (occasionally up to 12 weeks) but rarely manifests after prolonged treatment. In our case, disease progression was observed 16 weeks after initiating atezolizumab, prompting TSME combination therapy. Thus, the likelihood of atezolizumab-induced delayed pseudo-progression is minimal based on both temporal pattern and incidence rates. Notably, the patient reported moderate abdominal distension and fatigue prior to TSME therapy, which gradually resolved post-TSME intervention. This symptomatic improvement aligns with radiographic stabilization (7-month PFS), strongly suggesting that the observed clinical benefit was primarily driven by the synergistic effects of TSME combined with anti-PD-L1 inhibitors rather than delayed pseudo-progression.

FMT remains a primary strategy for modulating gut microbiota to enhance immunotherapy. However, its clinical utility is limited by donor dependency, batch heterogeneity, and infection risks. For instance, FMT trials in melanoma reported objective response rates (ORR) of only 20–35% (22, 24), with efficacy highly donor-dependent. In contrast, TSME comprises nine well-characterized probiotic strains (e.g., Bifidobacterium longum, Lactobacillus reuteri), validated in preclinical studies ( Figures 2d-f ). Standardized production (SC10632117100037) ensures batch consistency and food-grade safety for long-term oral administration. Moreover, TSME has a long shelf life of 24 months and is easy to store. The recommended storage conditions are to keep it in a cool, dry place, preferably refrigerated at 4°C, while avoiding direct sunlight. It is also convenient to use: the recommended administration method is to take it with warm water, 6 tablets twice daily. Importantly, the bacterial strains in TSME exhibit acid and bile salt resistance. TSME is prepared using a multi-layer encapsulation technology to prevent degradation by gastric acid and bile, significantly enhancing its suitability for clinical application and widespread promotion. Mechanistically, TSME targets complementary pathways through multispecies synergy: Bifidobacterium activates dendritic cells (DCs) via short-chain fatty acids (SCFAs) to promote CD8⁺ T cells activation (19), while Lactobacillus reuteri-derived indole-3-aldehyde (I3A) suppresses Treg differentiation and sustains effector T-cell function (29. Such precision is unattainable with FMT’s heterogeneous microbiota.

The resistance of (HCC) to (ICIs) is closely associated with T cell exhaustion and infiltration of immunosuppressive cells within the tumor microenvironment (TME) (35). Our study demonstrates that the combination of TSME with αPD-1 significantly enhances intra-tumoral CD8⁺/CD4⁺ T cell density (2.1-fold increase compared to monotherapy) while reducing the proportion of Treg cells ( Figures 2c, f ). These findings align with recent evidence suggesting that gut microbiota modulation, such as enrichment of Lachnospiraceae, improves HCC prognosis by promoting CD8⁺ T cell infiltration (26). Further supporting this mechanistic link, Mao et al. (26) identified specific bacterial taxa (e.g., Lachnospiraceae bacterium-GAM79 and Alistipes sp. Marseille-P5997) whose abundance correlates with prolonged PFS and overall survival (OS) in immunotherapy recipients (26). Importantly, our work extends these observations by revealing that multi-strain probiotic combinations can systemically remodel the immunosuppressive TME. This immunomodulatory effect may be partially mediated by Alistipes-derived metabolites, which have been shown to inhibit PD-1-induced T cell exhaustion (36).

Clinically, the observed radiological stabilization (SD-S) and symptomatic improvement in TSME-treated patients correlated with dynamic immune reconfiguration in the TME ( Figure 3 ). To advance these findings, future studies should employ integrated multi-omics approaches (metagenomics/metabolomics) to establish causal relationships within the “microbiota-metabolite-immune phenotype” axis and refine therapeutic protocols through mechanistic validation.

This study demonstrates that TSME combined with PD-1/PD-L1 inhibitors significantly improves outcomes in immunotherapy-resistant HCC. However, several limitations should be noted. First, clinical evidence relies on a single case, necessitating larger cohorts to validate TSME’s generalizability. Second, the specific contributions of individual TSME strains and their immunomodulatory metabolites require further exploration. Additionally, the H22 cell line-derived subcutaneous xenograft mouse model used in this study exhibits differences in tumor microenvironment characteristics (e.g., immune cell composition, stromal features) compared to human HCC orthotopic tumors, which may limit clinical translatability. Future studies should integrate multi-omics technologies (e.g., metabolomics) to establish causal “microbiota-metabolite-immune phenotype” relationships and explore TSME synergies with other ICIs (e.g., CTLA-4 inhibitors) or targeted therapies.

In conclusion, TSME combined with PD-1/PD-L1 inhibitors offers a novel strategy to overcome immunotherapy resistance in HCC by remodeling the TME. Its standardized formulation and mechanistic precision address FMT limitations, advancing gut microbiota interventions in precision oncology. The rigorous pseudo-progression assessment framework herein provides critical guidance for future clinical trials.