Circulating tumor cell detection in hepatocellular carcinoma based on karyoplasmic ratios using imaging flow cytometry

Abstract

Circulating tumor cells (CTCs) originate from tumor tissues and are associated with cancer prognosis. However, existing technologies for CTC detection are limited owing to a lack of specific or accurate biomarkers. Here, we developed a new method for CTC detection based on the karyoplasmic ratio, without biomarkers. Consecutive patients with liver cancer or non-cancer liver diseases were recruited. CTCs in blood samples were analyzed by imaging flow cytometry based on the karyoplasmic ratio as well as EpCAM and CD45. Microvascular invasion (MVI), tumor recurrence, and survival were recorded for all patients. A total of 56.2 ± 23.8/100,000 cells with high karyoplasmic ratios (HKR cells) were detected in cancer patients, which was higher than the number of HKR cells in the non-cancer group (7.6 ± 2.2/100,000). There was also a difference in HKR cells between liver cancer patients with and without MVI. Based on a receiver operating characteristic curve analysis, the threshold was 21.8 HKR cells per 100,000 peripheral blood mononuclear cells, and the area under the curve was higher than those of traditional methods (e.g., CD45 and EpCAM staining). These results indicate that the new CTC detection method was more sensitive and reliable than existing methods. Accordingly, it may improve clinical CTC detection.

Figure 1. Detection of cells with abnormal nuclei using imaging flow cytometry.

(a) Representative cell images with the same DAPI intensity. CD45, a lymphocyte biomarker, was labeled with PE-Cy5. Upper panel: CD45⁺ lymphocyte with a smaller DAPI area. Lower panel: CD45⁻ cells with a larger DAPI area, showing a looser structure of nuclei. Scale bars represent 20 μm. (b) Mean DAPI fluorescence intensity in PBMCs. Unlike the cells with a normal DAPI area (lower panel), a G1 and G2 peak could be detected in the cells with larger DAPI areas (upper panel). (c) Imaging flow cytometry test results for the peripheral blood samples from MVI patients. Left panel, basic results for peripheral blood mononuclear cells (PBMCs), the horizontal axis indicates the cell area and the vertical axis indicates the aspect ratio. The gate displays the group of single cells. Right panel: patients with MVI had 68.1 ± 14.8 cells with large DAPI areas. (d) Representative images for PBMCs from HCC patients. CD45, a lymphocyte biomarker, was labeled with PE-Cy5. EpCAM, a biomarker of circulating tumor cells (CTCs), was labeled with FITC. Left panel: cells with high karyoplasmic ratios (HKR cells), which were EpCAM-positive and CD45-negative. Right panel: cells with normal karyoplasmic ratios (normal cells), which were EpCAM-negative and CD45-positive. Scale bars represent 20 μm. (e) Flow cytometry test results for HKR cells from MVI patients. All HKR cells were CD45⁻, but only 8.7% of HKR cells were EpCAM⁺.(f) Mean DAPI fluorescence intensity in PBMCs. Unlike cells with normal DAPI areas (right panel), a G1 and G2 peak could be detected in the cells with larger DAPI areas. (g) Karyoplasmic ratio detection in PBMCs. Nuclei were marked using DAPI. Compared to the patients with MVI, normal individuals had more cells with large DAPI areas.

Figure 2. Differentiation of HCC based on HKR cell content.

(a) The source or composition of samples. (b) Representative results for the in vivo cellular localization of DAPI (Ex: 405 nm, Em: 454 nm) in peripheral blood mononuclear cells analyzed by real-time imaging flow cytometry 1 h after DAPI staining. Two groups of cells with different nuclear morphology are shown. Scale bars represent 20 μm. (c) Number of HKR cells in 100,000 peripheral blood mononuclear cells for three different groups. HKR cells were identified by imaging flow cytometry. Data are presented as means ± SD of three independent experiments, **P < 0.01. (d) Number of HKR cells in 100,000 peripheral blood mononuclear cells in the no MVI group (M0) and the MVI group (M1 & M2). HKR cells were identified by imaging flow cytometry. Data are presented as means ± SD of three independent experiments, **P < 0.01.

Figure 3. Diagnostic value of HKR cell content.

(a) Youden curve for HKR cells in 52 HCC patients and 12 normal individuals. Arrow refers to the maximum Youden index, which was defined as the threshold. (b) ROC curve of HKR cells for 23 HCC patients with MVI (M1 & M2) and 18 HCC patients without MVI (M0). HKR cells were identified by imaging flow cytometry. (c) Relative number of CTCs in 100,000 peripheral blood mononuclear cells in non-cancer liver disease patients and HCC patients. CTCs were defined as CD45⁻ & EpCAM⁺ cells in the peripheral blood and detected by imaging flow cytometry. Data are presented as means ± SD of independent experiments. (d) Relative number of CTCs in 100,000 peripheral blood mononuclear cells in HCC patients with MVI (M1 & M2) and HCC patients without MVI (M0). CTCs were defined as CD45⁻ & EpCAM⁺ cells in peripheral blood and detected by imaging flow cytometry. Data are presented as means ± SD of independent experiments. *P < 0.05. (e) ROC curve of CTCs for 23 HCC patients with MVI (M1 & M2) and 18 HCC patients without MVI (M0) in 52 total samples. CTCs were defined as CD45⁻ & EpCAM⁺ cells in peripheral blood and detected by imaging flow cytometry.

Figure 4. Association between prognosis and HKR cells.

(a) Recurrence-free survival curve of HCC patients. Different patient survival rates between the high HKR cell group (HKR cells > 57.3/100,000 PBMC) and low HKR cell group are shown. (b) Number of cancer cases in the high and low HKR groups. HKR cells were detected by imaging flow cytometry. A threshold of 21.8/100,000 PBMCs was used. *P < 0.05 C: Number of MVI (M0 or M1&M2) cases (HKR cells > 57.3/100,000 PBMCs) in HKR and other groups. HKR cells were defined and detected by imaging flow cytometry. A threshold of 57.3/100,000 PBMCs was used.