Potential treatment approaches for malignant peritoneal mesothelioma: in vivo and in vitro experimental study of natural killer cell immunotherapy

NK cell isolation, culture, and quality control

Mononuclear cells were isolated from 50 mL of umbilical cord blood that was collected at the Fifth Medical Center of Chinese PLA General Hospital (Beijing, China) using a Ficoll reagent kit (Dakewe Biotech Co., Ltd., Shenzhen, China) in accordance with the manufacturer’s guidelines. NK cells were expanded, cultured for 14 days using an NK cell culture reagent kit (Dakewe Biotech Co., Ltd.), and the CD3⁻ CD56⁺ phenotype was confirmed. Quality NK cells met the following six criteria (Figure S1): (1) > 95% CD3⁻ CD56⁺ NK cell population; (2) > 3 × 10⁹ cells on day 14; (3) negative for Mycoplasma, as determined using a commercial detection kit (Lonza, Basel, Switzerland); (4) negative for bacteria; (5) negative for hepatitis B, hepatitis C, syphilis, human immunodeficiency virus, and novel coronavirus (Shanghai Botuo Biotechnology Co., Ltd., Shanghai, China); and (6) endotoxin level < 0.5 EU/mL (Andesus, Zhanjiang, China). NK cells were co-cultured with colorectal cancer HCT-15 and RKO cells and breast cancer MCF-7 cells to assess cytotoxicity (PuNuoSai, Wuhan, China).

Extraction and identification of human MPM primary cells

MPM tumor tissues, passaged six times and preserved in liquid nitrogen, were thawed in a water bath at 37°C, cut into 1.0 mm³ pieces using a sterile surgical blade, and implanted subcutaneously in NOG mice (Beijing Vital River Laboratory Animal Technology Co., Ltd., Beijing, China). Upon reaching a size of approximately 500 mm³, the tumors were resected and cut into 2 sections to extract primary MPM cells and construct an intraperitoneal tumor model¹⁷. MPM tumor tissues and primary cells were identified based on the following four criteria: (1) staining with hematoxylin-eosin (H&E) and Wright–Giemsa stain (Biosharp, Hefei, China) to determine cell morphology; (2) positive calretinin expression (catalog no. ab16694; Abcam, Ltd., Cambridge, UK); (3) positive cytokeratin 5 expression (catalog no. ab52635; Abcam, Ltd.); and (4) positive Wilms’ tumor 1 (WT1) expression (catalog no. ab89901; Abcam, Ltd.). The extracted MPM primary cells were utilized in corollary studies.

NK cell strategies against different MPM cell densities

The cytotoxic effects of NK cells on MPM at various tumor-cell density levels (single cell, single tumor cell colony, and large tumor cell confluence) were systematically evaluated, as follows: (1) At the single cell level, MPM cells (3 × 10⁴) were seeded in a 24-well plate and cultured overnight. Randomly collected MPM cells from one well were counted to determine the quantity of NK cells to be added. (2) At the single tumor cell colony level, MPM cells (8 × 10³) were seeded in a 6-well plate, cultured for 14 days, and the number of individual clone cell clusters > 50 was quantified. (3) At the large tumor cell confluence level, MPM cells (1 × 10⁵) were seeded in a 24-well plate and cultured overnight to a cell confluence >80%. Based on different cell density levels, NK cells were added at a 20:1, 10:1, 5:1, and 0:1 ratio and co-cultured for 24 h.

High-content imaging analysis system

A high-content imaging analysis system (Operetta CLS; PerkinElmer, Inc., Waltham, MA, USA) was utilized to assess the optimal effector-to-target (E:T) ratio, which was calculated as the NK cell-to-MPM cell ratio. The specific parameters were set as follows: (1) 49 points were captured per well; (2) in situ images were captured every 6 min; and (3) the total imaging duration was 5 h. Therefore, 2,450 images of each well were generated for subsequent analysis of the experimental data.

Wright–Giemsa staining

Wright–Giemsa staining was performed to identify primary MPM cells and assess NK cell-mediated cytotoxicity against MPM cells at the single cell, single tumor cell colony, and large tumor cell confluence levels. The supernatants were removed from the wells after co-culturing and the cells were fixed with 4% paraformaldehyde (Beyotime Biotechnology, Shanghai, China) at room temperature for 20 min, aspirated, washed 3 times with phosphate-buffered saline (PBS; Gibco®, Thermo Fisher Scientific, Waltham, MA, USA), dyed with Wright–Giemsa stain at room temperature, washed 3 times with PBS, and imaged under a microscope (Axio Scope A1; Carl Zeiss AG, Jena, Germany).

Crystal violet staining

MPM cells were stained with crystal violet (Solarbio Science and Technology Co., Ltd., Beijing, China) to assess NK cell cytotoxic effect. Briefly, cultured MPM cells were fixed with 4% paraformaldehyde at room temperature for 20 min, washed 3 times with PBS, stained with crystal violet at room temperature for 10 min, washed 3 times with PBS, and imaged under an inverted microscope (Carl Zeiss AG) equipped with a camera (Canon Inc., Tokyo, Japan).

Cell Counting Kit-8 (CCK-8) assay

The CCK-8 assay (Biosharp) was used to assess the NK cell cytotoxic effect. Briefly, MPM cells were co-cultured with NK cells in a 24-well plate. After the addition of 10% CCK-8 solution to each well, the plate was incubated at 37°C for 1 h in 5% CO₂/95% air. The optical density at 450 nm was measured using a microplate reader (Thermo Fisher Scientific).

Enzyme-linked immunosorbent assay (ELISA)

The levels of effector molecule [granzyme B, perforin, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α] expression in MPM cells co-cultured with NK cells and tumor tissues were determined using commercial ELISA kits (Solarbio Science and Technology Co., Ltd.) in accordance with the manufacturer’s instructions.

Flow cytometry

Flow cytometry was used to assess NK cell purity, degranulation level, and activity. To determine purity, 1 × 10⁶ NK cells were suspended in PBS and probed with antibodies against CD3 and CD56 (catalog nos. 340662 and 340723, respectively; BD Biosciences, Franklin Lakes, NJ, USA). To determine NK cell degranulation level, 1 × 10⁵ MPM cells were incubated with 1 × 10⁶ NK cells at a 10:1 E:T ratio in a 96-well V-shaped plate for 4 h. At the start of incubation, a protein transport inhibitor (catalog no. 00-4980-93; Thermo Fisher Scientific) and antibody against CD107A (catalog no 561348; BD Biosciences) were added. After incubation, antibodies against CD3 and CD56 were added. To detect NK cell activity, antibodies against CD3, CD56, annexin V, and 7-aminoactinomycin D (catalog no. AP105; MultiSciences Biotech Co., Ltd., Hangzhou, China) were added. Following completion of all procedures, flow cytometry was performed.

Immunocytochemical analysis

Immunocytochemical analysis was used to detect calretinin, cytokeratin 5, WT1, and Ki-67 (catalog no. GB151142-100; Wuhan Servicebio Technology Co., Ltd., Wuhan, China) expression in primary MPM cells. Following the experimental procedure outlined previously, the molecular pathologic characteristics of MPM cells were determined using primary antibodies against WT1, calretinin, cytokeratin 5, and Ki-67 at 1:50, 1:100, 1:100, and 1:1000 dilutions, respectively. The slides were imaged with a digital pathology slide scanner (KF-PRO-400; Ningbo Jiangfeng Biological Information Technology Co. Ltd., Ningbo, China).

Live cell microscopy

Live cell microscopy was used to study the dynamic process of NK cells killing individual MPM cells. Briefly, 3 × 10⁴ MPM cells were seeded in confocal dishes (Corning, Inc., Corning, NY, USA) and cultured overnight. NK cells were added at a 10:1 E:T ratio. Using a Nikon microscope (Ti2-E; Nikon Corporation, Tokyo, Japan), in situ bright-field images were captured at 12 points every 2 min for 4 h.

Transmission electron microscopy

A transmission electron microscope (H-7650B; Hitachi High-Tech Corporation, Tokyo, Japan) was used to detect subcellular changes to NK and MPM cells. After co-culturing for 2 h, the samples were immediately fixed in 2.5% glutaraldehyde solution and processed at the Biomedical Center of Tsinghua University (Beijing, China).

Scanning electron microscopy

A scanning electron microscope (Quanta 200; FEI Company, Hillsboro, OR, USA) was used to observe changes to the surface structures of NK and MPM cells. Briefly, 3 × 10⁴ MPM cells were seeded in a 24-well plate (with built-in cell climbing slices) and cultured overnight. NK cells were added at a 10:1 E:T ratio. After co-culturing for 2 h, the samples were immediately fixed with osmium tetroxide and processed at the Biomedical Center of Tsinghua University.

Bioinformatics analysis

Utilizing public databases (10× Genomics website: https://support.10xgenomics.com/single-cell-gene-expression/datasets/3.1.0/5k pbmc protein v3 nextgem; GEO: https://www.ncbi.nlm.nih.gov/geo/, GSE172155), the differences between intertumoral infiltrating NK cells and peripheral blood were analyzed, as described in a previous report¹⁸.

Animal experimental protocol

The study protocol was approved by the Institutional Animal Care and Use Committee of the Tsinghua University Animal Ethics Committee [IACUC] (Approval No.: 23-LY2) and conducted in accordance with the guidelines of the “Guide for the Care and Use of Laboratory Animals.”

Twenty NOG mice (4–5 weeks old and weighing 16–18 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). Mice were injected intraperitoneally with 2 × 10⁶ MPM cells/100 μL into the left lower abdominal cavity on day 1 and randomly allocated to the control or NK cell treatment group (n = 10 mice/group)¹⁷. The NK group mice received an intraperitoneal NK cell suspension injection (2 × 10⁷ cells/100 μL) on day 2. NK cells were administered every 7 days for a total of 4 injections, while mice in the control group were injected intraperitoneally with an equivalent volume of saline¹⁹. Five mice from each group were randomly selected and the abdominal tumor burden was assessed on day 30. The remaining five mice in each group were observed for survival. All mice in the control group reached humane endpoints and the experiment was concluded on day 50 (Figure 5A). If mice were unable to move or near death, the experiment was terminated immediately in accordance with our humane endpoint criteria. The experimental peritoneal cancer index (ePCI) was used to evaluate the extent of tumor dissemination, as described in a previous study¹⁷. The abdominal-pelvic cavity was divided into four subareas and the lesion size (LS) score of each subarea was calculated as the diameter of the largest tumor, where LS-0 = no visible tumor, LS-1 = diameter ≤ 0.2 cm, LS-2 = 0.2 cm < diameter ≤ 0.5 cm, and LS-3 = diameter > 0.5 cm. The presence of malignant ascites was assigned one point. The accumulative ePCI score ranged from 0–13.

Magnetic resonance imaging

Magnetic resonance imaging (9.4T BioSpec system; Bruker, Billerica, MA, USA) was performed at Tsinghua University to evaluate the tumor burden in the abdominal cavity of mice anesthetized with 3% isoflurane for 5 min.

H&E staining and immunohistochemical (IHC) analysis

H&E staining and IHC analysis were performed to determine the histopathologic and molecular pathologic characteristics of mouse tumors and organs, as described in a previous study²⁰. IHC analysis was performed using primary antibodies (Abcam Ltd.) against calretinin, cytokeratin 5, WT1, and CD56 at 1:100, 1:200, 1:300, and 1:300 dilutions, respectively. Quantitative detection was performed, as described in a previous report²¹.

Hematologic testing

Routine blood, liver function, and kidney function tests were performed to assess NK cell toxicity to mice. Mice were anesthetized with isoflurane and blood was collected from the orbital venous sinus. Routine blood testing was performed using a fully automated blood cell analyzer (Thermo Fisher Scientific). Indicators of liver and kidney function were measured using a fully automated blood biochemical analyzer (Chemray 240; Rayto Life and Analytical Sciences Co., Ltd., Shenzhen, China).

Statistical analysis

The data are presented as the mean ± standard deviation. For continuous data following a normal distribution, the independent sample t-test was used for comparisons of two groups or one-way analysis of variance followed by Tukey’s or Games-Howell post hoc tests for comparisons of three or more groups. If the data did not follow a normal distribution, the independent samples rank-sum test was used for comparisons of two groups or the Kruskal–Wallis test for three or more groups. The chi-square test was used for comparisons of categorical data. Mouse survival analysis was conducted using the Kaplan–Meier method and log-rank test.

Optimal in vitro NK:MPM cytotoxicity ratio

A high-content imaging analysis system was used to assess the NK cell cytotoxic effect on MPM cells at different E:T ratios. Cytotoxicity was observed as early as 5 h after co-culturing. The killing rate reached 56.1% at a 5:1 E:T ratio (Figure 1A, B). Therefore, a 5:1 E:T ratio was selected to effectively kill MPM cells by NK cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Optimal in vitro NK:MPM cytotoxicity ratio. (A) The cytotoxic effect of NK cells on MPM cells at different E:T ratios were studied using a high-content imaging analysis system. The black arrows indicate dead MPM cells. Scale bar, 50 μm. (B) Statistical results from the high-content imaging analysis system for 5 h. (C) Specific lysis rate of MPM cells through the CCK-8 experiment for 24 h. (D) Morphologic changes in MPM cells were examined after 24 h of co-culture using Wright-Giemsa staining. Scale bar, 50 μm. The black and red arrows indicate NK and MPM cells, respectively. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, no significant difference.

Furthermore, morphologic changes to MPM cells were assessed after 24 h of co-culturing. Wright–Giemsa staining demonstrated the following morphologic changes: (1) at a high 20:1 E:T ratio, NK cells engulfed and destroyed MPM cells, leaving disintegrated cell debris; (2) at a medium 10:1 E:T ratio, numerous NK cells surrounded MPM cells, causing evident tumor cell damage but still visible cell morphology; and (3) at a low 5:1 E:T ratio, fewer NK cells gathered around MPM cells, which showed intact morphology. The killing rates at 20:1, 10:1, and 5:1 E:T ratios were 92.73%, 70.02%, and 49.74%, respectively (Figure 1C, D). The killing rate at a 10:1 E:T ratio exceeded the half-maximal inhibitory concentration, thus this ratio was selected for subsequent studies.

Dynamic cellular and subcellular changes to MPM cells after co-culturing with NK cells

Live cell imaging (Figure 2A) was performed to assess dynamic interactions between NK and MPM cells. Three distinguishable features were observed: (1) MPM cells exhibited typical apoptotic changes, such as cell membrane blebbing, cell shrinkage, and eventual cell disintegration (Figure 2B, E); (2) no morphologic changes were observed to MPM cells; and (3) MPM cells continued to proliferate, even after attack by NK cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Dynamic cellular and subcellular changes to MPM cells after co-culture with NK cells. (A) Dynamic interactions between NK and MPM cells were studied using live cell imaging techniques. Scale bar, 50 μm. Red circles represent MPM cells. (B) The process of NK cell-induced apoptosis in MPM cells. Scale bar, 50 μm. (C) Transmission electron microscopy was used to observe the internal structure of cells. Orange and red arrows represent NK and MPM cells, respectively. Yellow represents NK cell organelles. From left-to-right, top-to-bottom, the scale bars were 5 μm, 2 μm, 2 μm, 5 μm, 1 μm, 500 nm, 2 μm, and 2 μm, respectively. (D) The surface of cells was detected by scanning electron microscopy. Yellow and red represent NK and MPM cells, respectively. From left-to-right, top-to-bottom, the scale bars were 20 μm, 10 μm, 50 μm, 50 μm, 5 μm, 3 nm, 30 μm, and 5 μm, respectively. (E) Schematic diagram of the apoptosis process in MPM cells.

Transmission and scanning electron microscopy were used to assess subcellular (cell membrane and organelles) changes to MPM cells. NK cell-mediated killing of MPM cells was shown to involve four processes (Figure 2C, D): (1) chemotaxis, demonstrated by gradual migration of NK cells towards MPM cells, characterized by abundant membrane protrusions, and even distribution of the organelles in the cytoplasm of NK cells; (2) hitting of MPM cells by NK cells, resulting in significant depressions of the MPM cell membrane, with a reduction in the abundance of membrane protrusions of MPM cells, while lysosomes were clearly observed and the organelles of NK cells concentrated in the invaginations of MPM cells, even though the cell membranes of MPM cells remained intact at the attacking site; (3) adhesion of NK cells, which had the largest contact area, to MPM cells and the organelles of NK cells located on the adherent side; and (4) penetration of the MPM cell interior by NK cells via damage to the cell membrane.

Effector molecules released by NK cells at the optimal NK cytotoxicity level

The levels of the membrane protein CD107 (Figure 3A, B), granzyme B, perforin, IFN-γ, and TNF-α expression (Figure 3C–F) in NK cells were significantly increased after 24 h of co-culture.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Effector molecules released by NK cells at the optimal NK cytotoxicity level. (A) Level of NK cell membrane CD107A expression for 4 h at a 10:1 E:T ratio. (B) Statistical results of CD107A expression. (C–F) Levels of effector molecule expression for 24 h at a 10:1 E:T ratio. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

NK cell cytotoxicity gradually weakened with increasing tumor density

To better simulate clinical applications while maintaining the corresponding E:T ratio, the cell density of tumors was increased (Figure S2). At the single tumor cell colony level, the 24-h lysis rates at 20:1, 10:1, and 5:1 E:T ratios were 73.70%, 56.07%, and 31.26%, respectively (Figure 4A, B). At greater tumor cell confluence levels, the lysis rates were 48.91%, 35.19%, and 7.50%, respectively (Figure 4B, C). As the tumor density increased, the cytotoxicity of NK cells gradually decreased, possibly due to the following causes: (1) when aggregation of MPM cells exceeded 80%, the physical contact area decreased; (2) the tumor microenvironment (TME) was unsuitable for the survival of NK cells, resulting in significant death; and (3) the TME promoted dysfunction of NK cells.

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

NK cell cytotoxicity gradually decreased as tumor density increased. (A) Images of crystal violet staining after co-culturing NK and MPM cells (single tumor cell colony level) for 24 h. (B) CCK-8 experiment to detect the NK cell cytotoxicity against MPM cells at different densities for 24 h. (C) Images of Wright-Giemsa staining at the large tumor cell confluence level for 24 h. Scale bar, 50 μm. (D) Statistical results of the NK cell adhesion quantity at different tumor density levels for 24 h. (E) Flow cytometry analysis of NK cell activity after co-culturing with MPM cells (large tumor cell confluence level) for 24 h. (F) Statistical results of (E). (G) Immunofluorescence images of Ki-67 in MPM cells at the large tumor cell confluence level for 24 h. Yellow and red arrows represent NK and MPM cells, respectively. Scale bar, 20 μm. (H) Statistical results of Ki-67 expression in MPM cells for 24 h. (I–L) The level of different effector molecule expression (large tumor cell confluence level). (M) Potential mechanism diagram of infiltrating NK cell dysfunction.

To assess the effect of aggregation of MPM cells exceeding 80%, the number of NK cells adherent to a single MPM cell was calculated. The number of NK cells at the single cell level was higher than the large tumor cell confluence level (Figure 4D).

Flow cytometry was performed to assess the influence of the TME on NK cell activity. The results showed that > 95% of the NK cells were viable at the large tumor cell confluence level (Figure 4E, F).

Relevant analyses were performed using publicly available databases to determine if the TME promoted NK cell dysfunction. The results showed differential expression at the transcription level between NK cells infiltrating MPM tumor tissues and the peripheral blood (Figure S3A, Table S1). Gene Ontology (GO) function and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis concentrated on metabolic processes and NK cell-mediated cytotoxic signaling pathways, respectively (Figure S3B, C). The levels of granzyme B and perforin expression were clearly decreased in infiltrating cells compared to peripheral blood NK cells (Table S1). Furthermore, reduced levels of granzyme B and perforin expression were confirmed with an ELISA at the large tumor cell confluence level (Figure 4I, J), while the levels of IFN-γ and TNF-α expression were increased (Figure 4K, L). A potential mechanism underlying NK cell dysfunction is illustrated in Figure 4M.

The level of Ki-67 expression in MPM cells at the large tumor cell confluence level were assessed. Ki-67 expression was increased in residual MPM cells compared to the control group (Figure 4G, H).

NK cell adoptive transfer therapy reduced the tumor burden and prolonged survival of mice

Tumor recurrence and metastasis pose significant challenges for treatment of MPM patients. An intraperitoneal MPM tumor model was established to simulate recurrence (Figure S4). It is important to note that the day of tumor single cell injection was considered day 1. Mice were injected with NK cells at a 10:1 E:T ratio weekly for 4 consecutive weeks beginning on day 2 (Figure 5A).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

NK cell adoptive transfer therapy reduced the tumor burden and prolonged survival of mice. (A) The workflow of the experimental design. (B, C) Gross anatomy of mice. C, colon; D, diaphragm; K, kidney; L, liver; P, pancreas; SP, spleen; ST, stomach; U, uterus; X, xiphoid; Tumor tissue is indicated by yellow arrows. (D) Results of small animal magnetic resonance. (E) Results of ePCI score. (F) Ascites in mice. (G) Survival time of mice. (H) Body weight changes in different groups. NK injection was indicated by orange arrows. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Five mice from each group were randomly selected for gross anatomy examinations on day 30, which revealed that tumors of mice in the NK group were localized and mainly concentrated in the left abdominal area. In contrast, tumors of mice in the control group exhibited a diffusely progressive growth pattern with infiltration of various organs (Figure 5B, C). Magnetic resonance imaging demonstrated that tumors were significantly smaller in the NK group compared to the control group (Figure 5D). The ePCI was much lower in the NK group than the control group (Figure 5E, Table S2).

Importantly, within the 50-day follow-up period, bloody ascites was observed in 50% of mice in the control group (5/10), but none in the NK group (0/10; Figure 5F). All mice in the control group reached humane endpoints, while all mice in the NK group remained well and active (Figure 5G). Furthermore, tumors in the NK group were significantly smaller than the control group on day 50 (Figure S5). No significant changes in body weight were noted after NK cell injection (Figure 5H).

To further investigate tumor reduction in the NK group, tumor tissues were examined by IHC staining, which revealed apparent infiltration of the tumors by NK cells (Figure 6A), mainly located in the periphery. NK cells were not detected in the control group (Figure S6). Dead tumor cells were identified in the area of NK cell aggregation (Figure S7). Moreover, the ELISA results showed the levels of granzyme B and perforin expression in tumor tissues were higher in the NK group than the control group (Figure 6B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

NK cell releasing effector molecules inhibited tumor growth. (A) On day 30, IHC staining was performed to examine the infiltration of NK cells. NK cells infiltrated at the edge of the tumor tissue. Yellow and red arrows indicate NK and MPM cells, respectively (antibody: anti-CD56, species reaction: reacts with humans, but not with mice or rabbits). Scale bar, 50 μm. (B) On day 30, ELISA was used to detect the levels of effector molecule expression in tumor tissue. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, no significant difference.

Safety evaluation of NK cell therapy in vivo

Mice from both groups were euthanized on day 30 to assess the safety of NK cell therapy. Blood was collected for routine blood tests and to assess liver/kidney function. The heart, lung, liver, stomach, intestine, pancreas, spleen, kidney, testis, and uterus were collected for H&E staining to observe the presence of focal necrosis and lymphocyte infiltration. The hematologic results showed no significant differences in routine blood and liver/kidney function indicators between the NK and control groups (Figure 7A). The histopathologic results revealed no obvious focal necrosis or lymphocyte infiltration in either group (Figure 7B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

The safety evaluation of NK cell therapy in vivo. (A) Hematologic test results on day 30 in mice. HGB, hemoglobin; PLT, platelets; RBC, red blood cells; WBC, white blood cells; BUN, blood urea nitrogen; CREA, creatinine; UA, uric acid; TBIL, total bilirubin; DBIL, direct bilirubin; ALB, albumin; TBA, total biliary acid; γ-GT, γ-glutamyl transpeptidase. (B) H&E staining images of mouse organs on day 30. Scale bar, 50 μm.