From January 2017 to March 2023, a total of 611 HCC patients received immunotherapy (anti-PD-1 or anti-PD-L1) at the First Affiliated Hospital of Sun Yat-sen University.

The inclusion criteria were as follows: 1) diagnosed as hepatocellular carcinoma in BCLC C stage; 2) at least 2 cycles of immunotherapy and continuous immunotherapy according to the treatment plan; 3) Patients in the present study were categorized according to the post-immunotherapy change in CA19-9 levels, specifically, whether there was an increase or no increase (cut-off value: 35U/mL); 4) Complete clinical and follow-up data; 5) At least one measurable lesion as defined according to RECIST V1.1; 6) Child-Pugh class A, and Eastern Co-operative Oncology Group (ECOG) score ≤ 1. On the contrary, the exclusion criteria were as follows: 1) Combined with other malignant tumors; 2) Less than 2 course immunotherapy; 3) Lack of imaging, CA19-9, AST, ALT or other clinical data; 4) End-stage liver and kidney function; 5) Patients with CA19-9 increased before immunotherapy; 6) Lost to follow-up. The flow chart of the patient’s selection is shown in Figure 1A . According to the above criteria, a total of 326 patients were included in this study.

All patients signed an informed consent before immunotherapy. The Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University approved the study, which was in line with the 2008 Helsinki Declaration of the World Medical Association.

The TISIDB database was used to explored the relationship between FUT3 expression and immune subtypes or immunomodulators in HCC (19). The association between FUT3 and the relative enrichment score in 24 types of immune cells was carried out by the R package GSVA (20). The relationship in the infiltration of immunocytes between the high and low expression FUT3 groups was performed in Spearman’s correlation analysis.

IHC staining were performed in formalin-fixed, paraffin-embedded liver sections, as previously described (21). The primary antibodies (CD80 (1:100, Proteintech #66406) and CD206 (1:200, CST #24595)) were used in this study. Horse-radish peroxidase-conjugated anti-rat or anti-rabbit antibodies were used for detection, and images were captured using ZEISS Axio microscopy.

Mouse HCC cell line Hepa1-6 were obtained from the Cell Bank of the Type Culture Collection of the Chinese Academy of Sciences in January 2018. All cells were tested for Mycoplasma contamination using the single-step PCR method. All cells were cultured at 37°C and 5% CO2 in DMEM supplemented with 10% FBS (Gibco) and 1% penicillin–streptomycin.

All wild-type C57BL/6 mice (5-6 weeks, male) were purchased from GemPharmatech (Nanjing, China). All animals were acclimatized for 1 week before experiments and housed in a specific pathogen-free environment with a 12-hour light-dark cycle and permitted ad libitum consumption of water and a standard chow diet unless otherwise stated. All mouse experiments were approved by the Institutional Care and Animal Use Committee of Sun Yat-sen University Cancer Center (20110M).

For subcutaneous tumor xenograft models, mice were randomly divided into different groups (n=3-5/each group) until the tumor volume reached approximately 80–130 mm³. Hepa1-6 cells (5 × 10⁶ cells resuspended in 150 μL PBS) were injected into the right flanks of the mice. Tumor growth was monitored every three days, and the mice were euthanized at the end of study using pentobarbital sodium (40 mg/kg). The tumor volume was measured using the following formula: 0.5 × (larger diameter) × (smaller diameter)².

For drug treatment, 200 μg anti-PD1 antibody (Bio X Cell, West Lebanon, NH, USA) or anti-IgG (Bio X Cell) was injected intraperitoneally every three days. And 200 μg anti-CA19-9 (NS19-9) was daily intraperitoneal injected.

Tumor specimens were retrieved from euthanized murine subjects and processed into single-cell suspensions. Prior to antibody staining, single-cell suspensions were incubated with Fc receptor-blocking solutions to minimize non-specific binding. Specific antibodies were used to label the cells while maintained on ice.

For intracellular analysis, the Biolegend True-Nuclear™ Transcription Buffer Set (Catalog No. 424401) was employed. Following careful washing steps, both extracellular and intracellular antibodies were applied simultaneously to the cells. Fluorochrome-conjugated antibodies were sourced from reputable vendors including Biolegend, eBioscience, and BD Biosciences (CD45-BV421, CD45-APC-Cy7, FVD-ef506, CD3-APC, CD4-PE-Cy7, CD8-APC-Cy7, CD11b-APC, F4/80-FITC, MHC-II-PE, CD206-PE-Cy7, Ly6G-BV421, Ly6C-BV605). Data acquisition was performed using a BD flow cytometer, and subsequent analysis was conducted utilizing FlowJo software.

All patients were collected blood samples before the first immunotherapy. All data included age, gender, maximal tumor diameter(MTD), number of tumors, portal vein tumor thrombus(PVTT), extrahepatic metastasis, hepatitis status, liver cirrhosis, CA19-9 levels, albumin (ALB), total bilirubin (TBIL), prothrombin time (PT). After that, CA19-9 levels should be collected as far as possible for each immunotherapy. Using the normal reference value established by our hospital as a benchmark, CA19-9 levels ≥35 U/mL are classified as the elevated group, whereas CA19-9 levels <35 U/mL are categorized as the normal group.

Usually, patients received anti-PD-1/PD-L1 treatment every 3 weeks, and the immunotherapy plan could be formulated and adjusted according to the tumor situation. Using computed tomography or magnetic resonance imaging, according to the response evaluation criteria in solid tumors (RECIST1.1), the tumor was evaluated every 2 treatment cycles after the first treatment and then it was evaluated every 3-6 months after half a year of the first anti-PD-1/PD-L1 treatment. The serum CA19-9 level was collected and updated synchronously. The follow-up data was based on inpatient or outpatient visits or telephone records, and the follow-up data collection was terminated on June 30, 2023.

The main research endpoint was overall survival (OS), which was defined as the time interval from the beginning of the first immunotherapy to the death of any cause or the survival of the last follow-up (deletion). The secondary research endpoint was progress-free survival (PFS), which was defined as the time from the beginning of the first treatment to disease progression (PD) or death or last follow-up (deletion).

Continuous variables used Wilcoxon rank sum test or student’s t-test, while classification variables used chi-square test or Fisher precision test to compare the correlation between CA19-9 level and clinical pathological characteristics. The Kaplan-Meier method was used to generate the OS and PFS curves, and the Log-rank test was used to analyze the differences between the two groups of different CA19-9 levels. Based on the COX proportional hazards model, the univariate and multivariate analysis of survival were carried out to confirm independent predictors for PFS and OS. Clinical pathological factors with P<0.05 in univariate COX analysis were included in the multivariate analysis.

All statistical tests were double-tailed tests, and the p<0.05 was considered to be statistically significant.

Figure 1 presents the flowchart of patient enrolment. According to the inclusion and exclusion criteria, we included 326 patients in this study. Of the 326 patients, after immunotherapy, 66 and 260 exhibited elevated and normal serum CA19-9 levels, respectively (hereafter referred to as the elevated and normal groups, respectively). Table 1 summarizes the baseline characteristics of the two groups. All the listed variables were comparable between these groups (all P > 0.05).

The median follow-up periods of the elevated and normal CA19-9 groups were 17.4 and 19.93 months, respectively. A total of 80 patients died during the follow-up. Kaplan–Meier analysis revealed that the median progression-free survival periods of the elevated and normal CA19-9 groups were 5.47 months (95% confidence interval [CI]: 3.96–6.98) and 13.23 months (95% CI: 11.30–15.16), respectively. The corresponding 1-year and 2-year PFS rates were 16.9% and 11.3% in the elevated CA19-9 group and 53.3% and 29.1% in the normal CA19-9 group (P < 0.001, Figure 1B ). These results indicated that the elevated CA19-9 levels were positively correlated with shorter PFS in the patients with HCC receiving immunotherapy.

The median overall survival (OS) periods in the elevated and normal groups were 16.60 months [95% CI: 12.42–20.78] and 37.23 months [95% CI: 30.34–44.12], respectively. The corresponding 1-year and 2-year OS were 64.0% and 36.5% in the elevated group and 90.5% and 75.5% in the normal group, respectively (P < 0.001, Figure 1C ). Survival analysis revealed that the elevated serum CA19-9 levels after immunotherapy were associated with poor survival.

Table 2 lists the predictors of PFS and OS identified through univariate and multivariate analyses. Univariate analysis revealed that CA19-9 levels (P < 0.001), extrahepatic metastasis (P = 0.001), and number of tumors (P = 0.020) were all significantly associated with PFS, whereas CA19-9 levels (P < 0.001), extrahepatic metastasis (P < 0.001), and PVTT (P = 0.049) were significantly related to OS. Multivariate analysis indicated that CA19-9 levels (HR: 2.68, 95% CI: 1.94–3.72, P< 0.001), extrahepatic metastasis (HR: 1.71, 95% CI: 1.26–2.32, P < 0.001), and number of tumors (HR: 0.69, 95% CI: 0.50–0.95, P = 0.022) were the significant prognostic factors for PFS. Furthermore, CA19-9 (HR: 3.31, 95% CI: 2.06–5.32, P < 0.001), HCC extrahepatic metastasis (HR: 2.64, 95%CI: 1.69–4.14, P < 0.001), and PVTT (HR: 1.67, 95% CI: 1.07–2.62, P = 0.024) were the significant prognostic factors for OS.

PD1/PD-L1 inhibitors enhance the immune response against cancer cells, and CA19-9 has been recognized as a potential prognostic factor in immunotherapy. We investigated the effect of CA19-9 on the tumor immune microenvironment. We determined the expression of FUT3, which is responsible for the synthesis of CA19-9 (22), and evaluated the infiltration of various immune cells in HCC tissues through the single-sample gene set enrichment analysis ( Figure 2A ). We assessed the infiltration of 24 types of immune cells in HCC tissues and observed correlations between the expression levels of FUT3 and the infiltration of specific immune cell subtypes ( Figure 2A ). Spearman correlation analysis was performed to investigate the correlation between the expression level of FUT3, measured in transcripts per million format, and the infiltration level of immune cells. The results demonstrated that the infiltration of NK CD56 bright cells (Spearman r = 0.277, p < 0.001), macrophages (Spearman r = 0.200, p < 0.001), Tem (Spearman r = 0.181, p 0.001), and Th2 cells (Spearman r = 0.146, p = 0.005) was positively correlated with FUT3 expression ( Figure 2B ), with markedly higher infiltration observed in the group exhibiting high FUT3 expression ( Figure 2C ). By contrast, we noted a negative correlation between FUT3 expression and Th17 cell infiltration (Spearman r = −0.233, p < 0.001). Macrophage infiltration plays a vital role in HCC (23–25). Figure 2D presents the typical computed tomography (CT) and magnetic resonance imaging (MRI) images of patients with HCC with normal or elevated serum CA19-9 levels before and after immunotherapy. Given the crucial role of macrophages in HCC, we investigated how CA19-9 levels and immunotherapy affect macrophage infiltration and polarization. We observed a substantial increase in the density of CD206+ tumor-associated macrophages (TAMs, indicative of M2 macrophages) in HCC samples with elevated CA19-9 levels compared with those without elevated CA19-9 levels post immunotherapy. However, the level of CD80+ TAMs (M1 macrophages) was lower in HCC samples with elevated CA19-9 levels than in those without elevated CA19-9 levels ( Figure 2D ). Altogether, the increased prevalence of M2 macrophages in HCC samples with elevated CA19-9 levels sheds light on potential implications for the prognosis of patients with HCC receiving immunotherapy.

To evaluate the effect of a CA19-9 inhibitor on HCC and the tumor immune microenvironment, we used a subcutaneous xenograft tumor model with Hepa1-6 cells. Our findings revealed that the treatment with an anti-PD1 antibody significantly reduced tumor growth compared with treatment with an anti-IgG antibody. The addition of an anti-CA19-9 antibody further suppressed tumor proliferation ( Figures 3A, B ). To understand how CA19-9 inhibition affects HCC, we analyzed the tumor immune microenvironment through flow cytometry. Macrophages identified as CD11b+F4/80+ within the CD45+ cell population were categorized into two groups based on MHCII expression levels, similar to observations in a colon adenoma model (26). Our results demonstrated an increase in M2-polarized macrophages and a decrease in M1-polarized macrophages in tumors treated with anti-PD1 compared with those treated with IgG. This effect was reversed following the administration of an anti-CA19-9 antibody ( Figures 3C, D ). In addition, the combination of anti-CA19-9 therapy with or without anti-PD-1 significantly enhanced the tumor infiltration of CD8+ T cells and reduced the presence of myeloid-derived suppressor cells (MDSCs), including both polymorphonuclear (PMN)-MDSCs and mononuclear (M)-MDSCs ( Figures 3C, D ).