Changes of circulating tumor cells expressing CD90 and EpCAM in early-phase of atezolizumab and bevacizumab for hepatocellular carcinoma

Circulating tumor cells (CTCs) are noninvasive biomarkers that can indicate the therapeutic response and prognosis. The study aimed to investigate the cellular characteristics of CTCs focusing on monitoring during atezolizumab and bevacizumab (Atezo-Bev) therapy in patients with hepatocellular carcinoma (HCC). Peripheral blood samples were collected from 10 healthy controls and 40 patients with HCC. CTCs enriched using RosetteSep™ Human CD45 depletion cocktail were analyzed by multiparametric flow cytometry. CTC isolation was based on PanCK(+)CD45(−) cells, and CTCs exhibiting markers CD90, CD133, EpCAM, or vimentin. The total number of CTCs and the number of CTCs expressing CD90, CD133, EpCAM, and vimentin were correlated with the BCLC stage of HCC. The change in total CTC count accurately reflected the initial response to Atezo-Bev therapy. The numbers and mean fluorescence intensity of the CTC subsets expressing CD90 and EpCAM molecules decreased in patients with partial response/stable disease, and increased in patients with progressive disease and were markedly correlated with overall survival. CD90(+) and EpCAM(+) CTCs may be candidate biomarkers for the early prediction of the treatment response and the overall survival of patients with HCC receiving Atezo-Bev therapy.


Introduction
Hepatocellular carcinoma (HCC) is a highly aggressive malignancy imposing a significant burden on health care systems worldwide [1].Despite advances in treatment options, the prognosis of advanced HCC remains poor, emphasizing the need for new therapeutic strategies [2].The combination of atezolizumab, an immune checkpoint inhibitor (ICI), and bevacizumab, a vascular endothelial growth factor inhibitor, has shown promising results in clinical trials as a first-line treatment for advanced HCC [3].However, the identification of reliable biomarkers to predict treatment response and monitor disease progression in patients undergoing combination therapy remains an area of active research.
Circulating tumor cells (CTCs), which are shed from primary tumors and circulate in peripheral blood, represent a dynamic and heterogeneous population of tumor cells that can provide valuable information about tumor biology [4].CTCs have emerged as potential non-invasive biomarkers for various solid tumors, providing insights into the treatment response and prognosis [5].Additionally, CTCs offer a unique opportunity for real-time monitoring of treatment response, providing an early indication of therapy effectiveness or the need for treatment modifications [6].In the context of HCC, the detection and characterization of CTCs have immense value, particularly during the course of treatment, as they can serve as real-time indicators of treatment response and disease dynamics [7].By longitudinal monitoring CTC levels and molecular characteristics during the course of atezolizumab and bevacizumab combination therapy, clinicians can gain valuable insights into treatment efficacy and disease progression [8].
In this study, we investigated the molecular and cellular characteristics of CTC, with a special emphasis on its detection and monitoring during the course of atezolizumab and bevacizumab combination therapy in patients with HCC.The therapeutic effects were correlated with changes in the expression of CD90 and EpCAM molecules observed in CTCs at the early timepoints of treatment and reflected the long-term prognosis of HCC patients.These results demonstrate that non-invasive monitoring of molecular alterations in CTCs may contribute to the early detection of predicted therapeutic responses in patients with HCC treated with ICI combination therapy.

Study protocol and patients
This retrospective study was performed in accordance with the Declaration of Helsinki and was approved by the Research Ethics Committee of University of Fukui (20200086).All participants signed a written informed consent form.From September 2020 to August 2023, peripheral venous blood samples were obtained from healthy controls and HCC patients at the University of Fukui Hospital.HCC was diagnosed radiologically in accordance with the American Association for the Study of Liver Diseases practice guidelines for the management of HCC, or histologically by needle biopsy of the liver tumor.ALBI (albumin-bilirubin) score was calculated using the following formula: (log10 bilirubin (μmol/L) × 0.66) + (albumin (g/L) × − 0.085).ALBI grade was defined by the following scores: ≤-2.60 = Grade 1, >-2.60 to ≤ -1.39 = Grade 2, >-1.39 = Grade 3 [9].Using reported cut-off value (− 2.270), ALBI grade 2 was subdivided into 2 grades (2a and 2b) and named as modified ALBI grade [10].

Basal hepatic disease
The etiology was defined as HBV positive in patients with positive hepatitis B virus (HBV) surface antigen (HBsAg).Hepatitis C virus (HCV) positive was determined in patients with positive anti-HCV antibody (HCV Ab).Patients negative for both the HBsAg anti-HCV Ab and were defined as non-B-non-C (NBNC).

Evaluation of the treatment response
Progression-free survival (PFS) was defined as the period from the start of treatment to disease progression or death, and overall survival (OS) was defined as the period from the start of treatment administration to death.

Enrichment of CTCs
Peripheral venous blood samples (10 mL each) were collected into ethylenediaminetetraacetic acid (EDTA)-2-Na anticoagulant tube.To avoid contamination by epithelial cells, the first 5 mL of blood was not used for CTC collection.CTCs were enriched using the RosetteSep Human CD45 Depletion Cocktail (StemCell Technologies, Vancouver, Canada).Blood samples were processed within 4 h after collection.After incubation with 50 μL Depletion Cocktail per 1 mL of blood for 20 min, the blood was diluted with an equal volume of Dulbecco's phosphate buffered saline (PBS; Gibco™, Waltham, MA, USA) and 2 % fetal bovine serum (FBS; Gibco™).The samples were then centrifuged at 1200×g for 15 min at RT using SepMate TM-50 mL tubes (StemCell Technologies) containing Lymphoprep™ (StemCell Technologies).The supernatant was transferred to a new standard tube and filled with buffer (PBS/2 % FBS).After centrifugation at 300×g for 10 min, cell pellets were suspended in PBS and collected.

RNA extraction and nested-PCR
Total RNA was extracted from cells enriched with RosetteSep using an RNeasy Mini Kit (Qiagen, Hilden, Germany).cDNA was synthesized from RNA using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) following the manufacturer's instructions.To increase the sensitivity of subsequent real-time PCR analysis, cDNA was pre-amplified in 14 PCR cycles with primers and the TaqMan Preamp Master Mix Kit (Applied Biosystems).To analyze mRNA expression, pre-amplified cDNA was evaluated using real-time PCR (StepOne Plus, Applied Biosystems).The primers and probes were obtained from Applied Biosystems (Table 2).The expression of target genes were analyzed using the ΔΔCt comparative threshold method.GAPDH was used as an internal control.

Statistical analyses
The Wilcoxon matched pair rank test, Mann-Whitney U test, or Tukey-Kramer post hoc test were conducted to analyze the statistical significance.Cumulative survival was operated with the Kaplan-Meier method, and analyzed by the log-rank test.Statistical analyses were conducted by GraphPad Prism software v10 (GraphPad Software Inc., San Diego, CA, USA).The results were considered statistically significant at p < 0.05.

Assay system for molecular biological characterization of circulating tumor cells
To investigate the molecular characteristics of CTCs in patients with HCC, we constructed an assay system using multiparametric flow cytometry after CTC enrichment.Peripheral venous blood samples were collected from 10 healthy controls and 40 patients with HCC.The clinical characteristics are summarized in Table 3.To capture and enrich CTCs from peripheral blood, we used the Roset-teSep™ Human CD45 Depletion Cocktail, which depletes white blood cells and red blood cells by density centrifugation with tetrameric antibody complexes recognizing CD45, CD66b, and glycophorin A (Fig. 1A).After CTC enrichment, PanCK(+) and CD45(− ) tumor cells were observed by immunofluorescent staining (Fig. 1B).In addition, PanCK(+) CTCs showing markers CD90, CD133, EpCAM, or vimentin were observed.Enriched CTCs were then analyzed in three steps by flow cytometry: (i) removing dead cells, (ii) isolation of CD45(− ) PanCK(+) cells as CTCs, and (iii) analysis of CTCs producing CD90, CD133, EpCAM, or vimentin, which are proteins associated with cancer progression (Fig. 1C).This CTC assay enables precise discrimination of CTCs in peripheral blood and the evaluation of their molecular biological characteristics in patients with HCC.

Number of CTCs expressing CD90, CD133, vimentin, and EpCAM molecules correlated with the BCLC stage
The number and characteristics of CTCs of patients with HCC in each BCLC stage were evaluated.The total number of CTCs in healthy controls (HCs) was significantly lower than the number of CTCs in HCC patients with BCLC stages B and C (Fig. 2A).In HCC patients, the number of CTCs was consistent with BCLC stage progression, with a significant increase in stage C compared to patients in stage A. CD90, CD133, and vimentin(+) CTC counts were significantly increased in patients with stage C compared to HCs (Fig. 2B).The number of EpCAM(+) CTCs increased with the progression of BCLC stage.These results indicate that the total number of CTC and the number of CTCs expressing CD90, CD133, vimentin and EpCAM correlated with the BCLC stage of HCC patients.

Clinical course of CTC in representative patients with HCC treated with atezolizumab and bevacizumab therapy
The clinical course of CTC and the therapeutic response are shown in representative cases of HCC treated with atezolizumab plus bevacizumab (Atezo-Bev) (Fig. 3).The first patient (case #13, 71 years old, female) was complicated by multinodular HCC (Fig. 3A).The number of total CTC, CD90(+) and EpCAM(+) CTC decreased 0.7 months after starting Atezo-Bev therapy and decreased in subsequent courses.AFP levels decreased and the tumor size was reduced by 17.3 % at 2.1 months.Subsequently, the tumor exhibited a sustained partial response (PR).The expression of Nanog homeobox (NANOG), OCT3/4, SRY-box transcription factor 2 (SOX2), stem cell-associated genes, N-cadherin, and epithelial-mesenchymal transition (EMT)-associated genes expressed by CTCs was reduced by Atezo-Bev therapy.
The second case (case #6, 48-year-old, male) was complicated by multinodular HCC (Fig. 3B).The total CTCs and CD90(+) and EpCAM(+) CTCs were increased after 0.6 months of Atezo-Bev therapy.Tumor size increased by 61 % after 2.0 months.The number of total and subtypes of CTC increased after treatment with the multikinase inhibitors lenvatinib and cabozantinib and tumor size increased.The expression of NANOG, OCT3/4, SOX2, and N-cadherin was up-regulated in CTCs.These results suggest that the total number of CTCs and CD90(+) and EpCAM(+) CTCs may be accurate biomarkers of clinical response during the course of treatment of patients with HCC receiving Atezo-Bev therapy.

Total number of CTCs reflects the early treatment effect of Atezo-Bev therapy
Changes in total CTC counts were investigated in patients with HCC treated with TACE or Atezo-Bev.The clinical characteristics are summarized in Table 4.After 1 week of treatment, the total CTC count was significantly reduced with TACE treatment but not with Atezo-Bev therapy (Fig. 4A).Among 40 patients with HCC, 21 patients were treated with Atezo-Bev.The change in the number of CTCs during Atezo-Bev therapy was evaluated in 19 patients, excluding 2 patients whose CTCs could not be collected at the time of the first treatment response evaluation.At the time of the first response evaluation to Atezo-Bev therapy according to RECIST 1.1, the change in total CTC counts was examined.In the PR/stable disease (SD) group, the total CTC counts decreased significantly (median CTC counts: baseline/response evaluation, 173.0/83.5 cells) (Fig. 4B).In patients with progressive disease (PD), the total number of CTC increased significantly (median CTC count: 112.0/374.0cells).The PR/SD group in the first response evaluation had better PFS and OS than the PD group (Fig. 4C).These results indicate that the changes in total CTC counts could accurately reflect the initial response to Atezo-Bev therapy in patients with HCC.

CTCs that express the CD90 and EpCAM molecules reflect the early treatment effects and OS of Atezo-Bev therapy
The association between changes in the number and characteristics of CTCs that express CD90, CD133, EpCAM, and vimentin during Atezo-Bev therapy and therapeutic response was analyzed in patients with HCC.In patients with PR/SD and PD as the initial response evaluation to Atezo-Bev, there was no change in the number of CTCs expressing CD90, CD133, EpCAM, and vimentin after 1 week (Fig. 5A).The numbers of CD90(+) and EpCAM(+) CTCs decreased in the PR/SD group and tended to increase in the PD group (Fig. 5B).The number of CD133(+) and vimentin (+) CTCs was not associated with the therapeutic response.Changes in the mean

Discussion
In this study, we isolated CTC from the peripheral blood of patients with HCC and monitored their molecular and cellular characteristics during ICI therapy.The total number of CTCs and the number of CTCs expressing CD90, CD133, vimentin, and EpCAM were correlated with the BCLC stage of HCC patients.Changes in total CTC counts accurately reflected the initial response of the treatment to Atezo-Bev.Furthermore, the number and MFI of the CTC subsets expressing CD90 and EpCAM decreased in patients with PR/SD and increased in patients with PD, and were markedly correlated with OS of the patients after treatment.Collectively, CD90(+) and EpCAM (+) CTCs may be candidate biomarkers for the early prediction of treatment response and OS in patients with HCC treated with Atezo-Bev.
Unlike conventional cytotoxic chemotherapy, ICIs may require a longer period to elicit a response, and some patients experience atypical radiological features, including delayed and dissociated responses [12].The RECIST criteria of responses to treatments in solid tumor patients relies primarily on unidimensional measurements of tumor size [13] and typically occurs at predefined time points, which may not capture the delayed responses seen with immunotherapy [14].This can lead to an underestimation of the efficacy of treatment with ICIs, due to the premature timing of therapeutic evaluation.
Liquid biopsy has emerged as a noninvasive diagnostic method for detecting and monitoring cancers, including HCC [15].It involves the analysis of circulating tumor-derived materials, such as extracellular vesicles (EVs), circulating tumor DNA (ctDNA), and CTCs in blood samples [16].Recent studies have identified specific genetic alterations in ctDNA samples from patients with HCC, including mutations of the TP53 and CTNNB1 genes and TERT promoter mutations [17,18].EV-based liquid biopsy holds promise for the diagnosis of HCC, as it allows the detection of tumor-specific molecules, such as proteins, RNAs and DNA [19].CTCs originating from solid tumors are related to hematogenous metastatic spread to distant sites, and patients with CTC-positive HCC have been reported to have a higher risk of recurrence and shorter recurrence-free survival [20].In addition, tumor cell dissemination may be an early event in the pathogenesis of HCC, as CTCs have been detected in 90 % of patients with HCC and more than half of patients with early stage HCC [21].Consistently, CTCs were detected in a high percentage of patients with HCC in this study.
Molecular expression analysis of CTCs using flow cytometry is clinically relevant to the prognosis and therapeutic efficacy of colorectal cancer [22,23], prostate cancer [24], and ovarian cancer [25,26].In this study, cell surface molecules of CTCs were stained with CD45 and CK antibodies as reported previously [27][28][29], and 7AAD-/CD45-/Pan CK + cells were determined as CTCs.However, flow cytometry must take into account technical difficulties in distinguishing between genuine CTCs and false positive events [25].The number of CTCs counted in this study was higher compared to the CTC studies of HCC counted by fluorescence microscopy [8,21].Since the number of CTCs in the HC was low in this study, it is possible that non-specific false bindings were counted as false positives in flow cytometry.Considering these possibilities, the protein expression in CTCs in patients with HCC was observed using fluorescence microscopy.
The clinical significance of molecular expression associated with EMT and Stem cell markers in CTCs has been reported [30,31].Analysis of CTCs expressing vimentin [21] as a mesenchymal marker and CD133 [21], CD90 [21,[32][33][34], and EpCAM [21,33,35] as stem cell markers in HCC has been reported to be useful for early diagnosis, prognosis prediction, and determination of therapeutic efficacy.Vimentin migrates from intracellular regions to the cell surface during EMT to produce cell surface vimentin [36,37].Cell surface vimentin has been used as a target to collect mesenchymal CTCs and has been reported to be an effective diagnostic and prognostic biomarker for gastric [38], colorectal [36], pancreatic [39], lung [40], and breast cancers [41].In this study, the clinical significance of CTCs expressing CD90, CD133, EpCAM, and vimentin in the course of Atezo-Bev was evaluated by serial collection of CTCs.Therapeutic effects were correlated with changes in CD90 and EpCAM expression observed in CTC at early time points of treatment in patients with HCC.CD90 is a marker of cancer stem cell (CSCs) that exhibits a comparable tendency in the association between circulating CSCs and clinical outcomes [42].CD90(+) HCC cells have been shown to have high lung metastatic potential [43].Gene expression profiling of CD90(+) HCC cells showed increased gene expression associated with inflammation, drug resistance, and cell proliferation [44].These support the increased number of CD90(+) CTCs in patients who progressed during Atezo-Bev therapy in this study.Consistent with current results, EpCAM(+) CTCs reportedly correlate with a poor outcome in patients with metastatic non-small cell lung cancer, prostate cancer, and HCC [45,46].A key event in the intravasation of HCC cells into the bloodstream is the EMT.EMT is a process in which epithelial tumor cells lose their adhesion capacity due to downregulation of EpCAM expression [47,48].To extravasate circulating EpCAM downregulated CTCs from the bloodstream and form metastases, they must become EpCAM positive again through the mesenchymal epithelial transition, the reversal of EMT [49,50].EpCAM(+) HCC cells are induced by activation of Wnt/β-catenin signaling [51].Activation of Wnt/β-catenin signaling in HCC has been reported to be associated with resistance to ICIs [52].These support the number and MFI of EpCAM(+) CTCs increased in patients with PD treated with Atezo-Bev.
There are several limitations to this study.First, although we performed the assay system using multiparametric flow cytometry after CTC enrichment in patients with HCC, only 19 patients received atezolizumab and bevacizumab treatment, which is a small fraction to draw definitive conclusions about the usefulness of CD90(+) and EpCAM(+) CTCs as predictive biomarkers of response to therapy.However, in these cases, CTCs were collected and evaluated at many points during the long-term course of Atezo-Bev therapy and correlated well with the image response determination of HCC.Second, this study was performed at a single institution.In the future, it may be important to increase the number of samples and analyze the molecular biological significance of CD90(+) CTCs and EpCAM(+) CTCs in Atezo-Bev therapy of HCC.
In conclusion, changes in the number and expression levels of CD90 and EpCAM molecules in CTCs at the early time point of Atezo-Bev therapy correlated with treatment effects and reflected in the long-term prognosis of HCC patients.Non-invasive monitoring of molecular alterations in CTC can contribute to the early detection of anticipated therapeutic responses in HCC patients treated with the immune checkpoint inhibitor combination therapy.T. Nosaka et al.

Fig. 1 .
Fig. 1. Assay system for molecular biological characterization of circulating tumor cells (CTCs) using multiparametric flow cytometry.(A) Peripheral blood of patients was collected and CTCs were enriched by Rosettesep.The expression of surface protein was then analyzed by flow cytometry.The gene expression was analyzed by qRT-PCR.(B) Immunofluorescence staining was performed in CTC.PanCK, CD45, CD90, CD133, EpCAM, vimentin and DAPI were stained.Representative images are shown.Bar, 10 μm.(C) Flowchart of CTC isolation and molecular analysis using multiparametric flow cytometry.(i) removing dead cells by Fixable Viability Stain, (ii) isolation of CD45(− ) PanCK(+) cells as CTCs, and (iii) CD90, CD133, EpCAM, or vimentin positive cells.

Fig. 2 .
Fig. 2. Analysis of CTC subtypes according to BCLC stage in HCC patients.(A) Total CTC counts in healthy controls (HCs) (n = 10) and HCC patients by BCLC stage (BCLC stage A/B/C; n = 10/12/18).(B) CD90, CD133, EpCAM, and vimentin-positive CTC counts in HCs and HCC patients by BCLC stage (upper panel) and percentage of each marker-positive CTCs to total CTCs in patients with HCC (lower panel).Centre lines of the box and whisker plot represent median values, whereas the box edges indicate the 25th and 75th percentiles, and the whiskers indicate the minimum and maximum values.Tukey-Kramer post hoc test.*p < 0.05; N.S., not significant.

Fig. 4 .Fig. 5 .
Fig. 4. The change in the total number of CTCs at the first response assessment to Atezo-Bev therapy in patients with HCC.(A) The change of total number of CTCs at baseline and 1 week after treatment in HCC patients treated with TACE (n = 15) and Atezo-Bev therapy(n = 12).(B, C) The change of total CTC counts at baseline and at the time of first response evaluation in HCC patients treated with Atezo-Bev therapy who were PR/SD (n = 12) and PD (n = 7) (B).Kaplan-Meier curves of progression free survival and survival rate of these patients (c).(A, B) Centre lines of the box and whisker plot represent median values, whereas the box edges indicate the 25th and 75th percentiles, and the whiskers indicate the minimum and maximum values.Wilcoxon matched-pairs rank test.(C) Log-rank test.*p < 0.05, CI, confidence interval; HR, hazard ratio; N.S., not significant.

Table 1
Antiobodies used in this study.

Table 2
Primers used in this study.

Table 3
Characteristics of patients with hepatocellular carcinoma and healthy controls.

Table 4
Characteristics of patients with hepatocellular carcinoma treated with TACE or atezolizumab + Bevacizumab.