Lymphangiogenesis in the liver of biliary atresia

Abstract

Background

Lymphatic vessels (LVs) play a crucial role in immune reactions by serving as the principal conduits for immune cells. However, to date, no study has analyzed the morphological changes in the LVs of patients with biliary atresia (BA). In this study, we aimed to determine the morphological changes in the LVs irrigating the liver in patients with BA, elucidate their correlations with the morphology of the portal vein (PV) branches, and discuss their etiopathogenetic significance.

Methods

Morphometric analyses of liver biopsy specimens from patients treated between 1986 and 2016 were performed. The parameters measured were as follows: the whole liver area of the specimen, fibrotic area, number of LVs, LVs without patent lumen (designated as Ly0) and PV branches, and diameters of the LVs with patent lumen and the PVs.

Results

The numbers of LVs, Ly0, and PV branches per unit area of the whole liver specimen were significantly higher in patients with BA than in control participants with liver disease and those with normal livers. However, no correlation was observed between the fibrotic area and the average diameter of LVs or PVs, and between the fibrotic area and the number of LVs or PV branches. Furthermore, no correlation was observed between the total number of LVs and the number of PV branches.

Conclusions

The present study showed a significant increase in the number of total LVs and Ly0, characterized by a high Ly0 to total LVs ratio, suggesting that lymphangiogenesis occurs in the liver of patients with BA.

Background

Biliary atresia (BA) is an immune-mediated disorder of the biliary tract characterized by abnormal narrowing, obstruction, or absence of extrahepatic bile ducts, leading to impaired bile flow, with potential progression to liver damage and cirrhosis [1]. Liver tissue with BA displays proliferation of the intrahepatic bile ducts, inflammatory infiltrates, and fibrous expansion of the portal tract that potentially progresses to cirrhosis [2]. Clinically, patients with BA present with jaundice and hepatomegaly, with or without splenomegaly. The color of stool is typically pigmented after birth, gradually becoming pale to clay-colored over the first couple of months in the postnatal period. Kasai portoenterostomy (KPE) is the standard surgical procedure for BA, which is typically performed between 1 to 3 months after birth. However, approximately half of the patients with BA eventually require liver transplantation by age 20 [3]. The underlying cause of this prognostic limitation and the mechanisms of BA pathogenesis remain unclear. Moreover, the etiology of BA is still controversial to date [4]. Various etiopathogenetic hypotheses have been proposed, including viral infections [5] and graft-versus-host disease (GvHD) associated with maternal microchimerism (MMc) [6].

Lymphatic vessels (LVs) play a crucial role in immune reactions by serving as the principal conduits for immune cells, including lymphocytes and antigen-presenting dendritic cells, to travel from peripheral tissues to regional lymph nodes, where antigen-specific immune responses are initiated. The liver is the largest lymph-producing organ, and a significant increase in the number of hepatic LVs (lymphangiogenesis) has been reported in various liver diseases [7, 8]. Despite its apparent relevance in liver pathology, the hepatic lymphatic system remains poorly studied, particularly in BA.

Masuya et al. reported an increase in the number of smaller portal vein (PV) branches and a decrease in the total luminal area of the PVs in the livers of patients with BA [9]. In embryos, the lymphatic endothelium arises from existing venous endothelial cells [10]. Despite having a common origin, lymphatic and venous endothelia are distinct at the functional and molecular levels. To our knowledge, no studies have analyzed the morphological changes in LVs or the relationship between LVs and PV in patients with BA. Therefore, in this study, we aimed to determine the morphological changes in LVs in the livers of patients with BA, elucidate their correlations with the morphology of PV branches, and discuss their etiopathogenetic significance.

Methods

Patients, stratification, and data collection

Between 1986 and 2016, liver tissue was collected from three groups of patients: 1) patients with BA (during KPE) at Ibaraki Children’s Hospital (n = 36) and all previously reported patients from Kagoshima University (n = 25) [9]; 2) control liver diseases (n = 42), including a) infants and children with choledochal cysts, samples of which were collected at the time of cyst excision (n = 24), b) neonates or infants with hepatitis (n = 7), c) infants with Alagille syndrome (n = 3), d) infants with intestinal failure-associated liver disease (IFALD) (n = 2), and e) infants with liver dysfunction requiring intraoperative liver biopsy (gastroschisis, n = 1; neuroblastoma, n = 1; portosystemic shunt, n = 1; gastroesophageal reflux, n = 1; trisomy 13, n = 1; and Wilson’s disease, n = 1); and 3) infants with normal livers obtained during fresh autopsy (n = 7).

Infants with choledochal cysts, IFALD, Alagille syndrome, hepatitis, or liver dysfunction were included as control liver disease participants for comparison with the BA group. Normal liver tissues obtained from infants at autopsy represented the normal liver control group.

Liver tissue sampling and pathological parameters

During KPE, tissues were obtained from the edges of the liver, fixed in 10% buffered formalin, and embedded in paraffin. Double chromogenic immunostaining for podoplanin (D2-40) (red) and CD34 (brown) was performed to distinguish LVs from PV branches as previously described (Fig. 1A) [9]. Elastica Masson-Goldner staining was performed to evaluate the degree of fibrosis.

Fig. 1

A Double chromogenic immunostaining for CD34 + D2-40. The vascular endothelium is stained brown, and the lymphatic endothelium is stained red. Circle: portal branch, dashed circle: lymphatic vessel. B The shortest diameters of lymphatic vessels and portal vein branches in histological specimen. C The lymphatic vessels (LVs) are differentiated from those without a patent lumen (designated as Ly0). Magnification, × 200

Morphometric analysis and correlation with preoperative laboratory data

Figure 1B shows histological images of the measured parameters. We measured the shortest diameters of the LVs and PV branches in histological specimens. Our study focus was on LVs without a patent lumen (designated as Ly0) (Fig. 1C). The parameters measured in this study were as follows: whole liver area of the specimen; total portal area; number of LVs, Ly0, and PV branches; the diameter of the LVs with a patent lumen (i.e., excluding Ly0) and the PVs; and the ratio of Ly0 to total LVs. The portal areas of BA were defined as fibrosis areas that included bile ducts. Measurements were performed using the image analysis software NIS-Elements BR 5.20.00 (Nikon, Japan). The total area of the specimen and the portal area were manually traced on a monitor using a computer tracking device. We also investigated the relationships between the morphometric parameters and preoperative laboratory data in patients with BA, including the levels of total bilirubin (T-Bil), direct bilirubin (D-Bil), aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyl transpeptidase (γ-GTP), white blood count (WBC), platelet count (PLT), and C-reactive protein (CRP).

Statistical analysis

Data are expressed as mean ± standard deviation. Differences between groups were tested for statistical significance using the Kruskal–Wallis test and the Steel–Dwass method.

Pearson’s correlation test was used to evaluate the correlation between the degree of fibrosis and the diameter of the LVs or PV as well as the numbers of LVs or PV branches. In addition, the correlation between the numbers of LVs and PV branches was assessed. P values < 0.05 were considered statistically significant. All statistical analyses were performed using JMP version 17 (SAS Institute, Cary, NA, USA).

Ethical approval

This study was approved by the Research Ethics Committee of Kagoshima University Hospital (registration number: 170347) and Ibaraki Children’s Hospital (registration number: 2023 IRB-37 K) and was performed in accordance with the Ethical Guidelines for Clinical Research from the Japanese Ministry of Health, Labor, and Welfare. The committee decided that use an opt-out system should be used instead of obtaining informed consent for this study.

Results

Patient background characteristics and clinical course

BA liver tissues (n = 61) were analyzed, and the results were compared with those of control liver disease participants (n = 42) and normal liver tissues (n = 7). Sixty-one patients (males, n = 26; females, n = 35) with BA underwent KPE at a median age of 62 days (range 30–143 days).

Morphological examination

Fibrosis

The fibrotic rate and number of total LVs, LVs without a patent lumen (Ly0), and PV branches in the whole liver area or portal area are presented in Table 1. The fibrotic rate in BA was significantly higher than that in the control liver disease and normal liver groups (Table 1) (21.77 ± 9.29 vs. 8.92 ± 6.77 vs. 5.52 ± 2.70%, respectively).

Table 1 Fibrotic rate and number of LVs or PV branches

Lymphatic vessels

The number of LVs per unit area of the whole liver specimen was significantly higher in the patients with BA than in the control liver disease and normal liver groups (Fig. 2A) (10.78 ± 5.14 vs. 7.31 ± 4.58 vs. 0.98 ± 0.23 /mm², respectively). However, the number of LVs per unit portal area was significantly higher in the control liver disease group than in the remaining groups, and it was significantly higher in livers with BA than in normal livers (57.09 ± 32.67 vs. 104.47 ± 59.7 vs. 20.54 ± 7.50 /mm², respectively).

Fig. 2

Morphometric analyses of lymphatic vessels. A A box plot comparing the numbers of lymphatic vessels per unit area of the whole liver specimens. B A box plot comparing the numbers of lymphatic vessels without a patent lumen per unit area of the whole liver specimens. C A histogram of the diameter of each lymphatic vessel in BA, control liver disease, and normal liver

The number of Ly0 per unit area of the whole liver specimen was also significantly higher in livers with BA than in the control liver disease and normal liver groups (Fig. 2B) (7.40 ± 3.38 vs. 4.47 ± 3.18 vs. 0.51 ± 0.08 /mm², respectively). The number of Ly0 per unit portal area was significantly higher in the control liver disease group than in livers with BA and normal livers (39.06 ± 22.7 vs. 62.76 ± 47.98 vs. 11.66 ± 6.12 /mm², respectively). The ratio of Ly0 to total LVs was 67.5% in the BA group, 56.6% in the control liver disease group, and 53.8% in the normal liver group, and the Ly0 ratio was significantly higher in the BA group than in the two other groups (p < 0.0005). Histograms of LVs diameters clearly showed that Ly0 was the most abundant in the livers with BA, and the distribution of the remaining LVs was similar to that of the control liver disease group (Fig. 2C).

Portal vein branches

The number of PV branches per unit area of the whole liver specimen was significantly larger in the livers with BA than in the control liver disease and normal liver group (Fig. 3A) (6.50 ± 5.36 vs. 2.54 ± 2.72 vs. 1.03 ± 0.34 /mm², respectively); however, no significant difference was observed in the number of PV branches per unit portal area among the groups (31.72 ± 24.34 vs. 35.96 ± 29.38 vs. 22.02 ± 11.81 /mm², respectively).

Fig. 3

Morphometric analyses of portal vein branches. A A box plot comparing the numbers of portal vein branches per unit area of the whole liver specimens. B A box plot comparing the average diameters of the portal vein

The average diameter of PV was significantly smaller in livers with BA than in the control liver disease and normal liver groups (Fig. 3B) (9.69 ± 5.37 vs. 21.38 ± 7.72 vs. 33.22 ± 8.36 μm, respectively).

No correlation was observed between the fibrotic area and the average diameter of LVs or PVs. Likewise, no correlation was found between the fibrotic area and the number of LVs or PV branches (Supplementary Figures. 1A, B, C, and D). Furthermore, no correlation was observed between the number of total LVs and the number of PV branches (Supplementary Figures. 1E and F).

Preoperative laboratory data

There were no significant correlations between the laboratory data (T-Bil, D-Bil, AST, ALT, γ-GTP, WBC, PLT, and CRP levels) and the number of LVs, Ly0, or PV branches.

Discussion

The present morphometric study highlighted two major vascular changes in BA livers: (1) lymphangiogenesis, characterized by a higher number of total LVs, especially those without the lumen (Ly0), and (2) changes in PV, including a decrease in PV size and an increase in the overall number of PV branches. The first change we observed provides a novel insight, while the second corroborated Masuya’s observations by adding more cases [9]. Importantly, these changes were not morphologically parallel to the fibrotic areas, suggesting that factors other than the fibrotic process could contribute to the increase in the number of total LVs, Ly0, and PV branches in the BA liver. In addition, the present study demonstrated that the number of total LVs, Ly0, and PV branches per unit portal area in control liver disease was higher than that in BA. This is attributed to the significantly larger portal area in BA, which made the distribution densities of total LVs, Ly0, and PV branches lower per unit portal area.

Regarding PV, diminished portal vasculature has been previously recognized in BA [11, 12]. The proliferation of fibrotic tissue compresses PVs, and PV branches potentially function as collateral vessels to supply more blood to the liver parenchymal cells [13]. However, the present study showed no correlation between fibrosis progression and the PV diameter, consistent with the results of Masuya et al. [9]. Therefore, PV narrowing may occur because of factors other than fibrosis.

As previously postulated by Muraji et al. [14], BA could be induced by maternal-derived immune cell attack on fetal bile duct epithelium expressing paternal antigens (maternal-fetal GvHD). We already speculated that this GvHD would also target portal vein endothelial cells [9] resulting in a marked decrease in the PVs and an increase in small branches of PVs. This mechanism would be applied to the present data on PVs.

The present study is the first to analyze the LVs in patients with BA. Lymphangiogenesis has been reported to occurs in the liver in various pathological conditions, including liver transplant, viral hepatitis, idiopathic portal hypertension, and GvHD [15,16,17,18]. Lymph production increases up to 30-fold in patients with cirrhosis, with concomitant increases in the formation of new LVs [7]. This phenomenon is thought to be due to a disturbance in the drainage of vascular flow from the sinusoid to the central/terminal hepatic veins associated with lobular distortion in patients with advanced liver disease or cirrhosis [7, 19]. Given this theory, both the number and diameter of LVs with a patent lumen would increase, and the degree of fibrosis would positively correlate with the luminal area of the LVs or the number of LVs with a patent lumen. However, in our study, the degree of fibrosis did not correlate with the average diameter of the LVs with a patent lumen or with the number of total LVs. Therefore, it is unlikely that lymphangiogenesis in BA is caused by vascular flow disturbances secondary to fibrosis. These findings indicate that factors other than fibrosis contribute to lymphangiogenesis in the BA liver.

The exact pathogenic mechanism underlying these changes in the LVs in BA remains unknown. However, it is reported that lymphangiogenesis in the liver is related to immune response [19], which can be caused by GvHD and viral infection. Ly0 could represent immature LVs, which we assume are secondary to probable immune responses such as GvHD [14].

We previously analyzed hepatitis-like findings, such as hepatocyte multinuclear changes, ballooning, and acidophilic bodies [20]. These changes are also observed in hepatic GvHD associated with hematopoietic stem cell transplantation [21,22,23]. We have already postulated that hepatitis-like findings in BA may be consequences of GvHD targeting hepatocytes [20]. Mertlitz et al., reported that lymphangiogenesis in GvHD can be ameliorated by antibodies against VEGFR-3 that LV endothelial cells express [16]. Considering these immune processes occurring in the BA liver, it can be speculated that lymphangiogenesis in BA is related to the immune responses in the BA liver [19].

As discussed, alterations of PVs and LVs in the BA liver may be related to immune mechanisms. However, we found no correlation between the two, indicating that their mechanisms may not be identical. Further analyses are required to clarify the mechanisms of alterations of LVs and PVs in the BA liver.

Conclusions

This is the first report descibing the LVs in the liver of patients with BA. The present study showed a significant increase in the total number of LVs and Ly0, characterized by a high Ly0 ratio to total LVs, suggesting that lymphangiogenesis occurs in the liver of patients with BA. Furthermore, the present study indicated the narrowing of PVs with an increase in the number of PV branches. These alterations in LVs and PVs may not be induced by liver fibrosis.