Potential conflict of interest: Nothing to report.
Heterozygous deletion or mutation in hepatocyte nuclear factor 1 homeobox B/transcription factor 2 (HNF1B/TCF2) causes renal cyst and diabetes syndrome (OMIM #137920). Mice with homozygous liver-specific deletion of Hnf1β revealed that a complete lack of this factor leads to ductopenia and bile duct dysplasia, in addition to mild hepatocyte defects. However, little is known about the hepatic consequences of deficient HNF1B function in humans. Three patients with heterozygous HNF1B deficiency were found to have normal bile duct formation on radiology and routine liver pathology. Electron microscopy revealed a paucity or absence of normal primary cilia. Therefore, heterozygous HNF1B deficiency is associated with ciliary anomalies in cholangiocytes, and this may cause cholestasis. (HEPATOLOGY 2012;56:1178–1181)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
HNF1B (hepatocyte nuclear factor 1 homeobox B) deficiency is the underlying cause of the renal cyst and diabetes syndrome (OMIM #137920), characterized by defects in pancreatic islets leading to maturity onset diabetes of the young (MODY5), and deficiencies in the exocrine pancreas, urogenital tract, and liver.1 During bile duct morphogenesis, HNF1B belongs to a dynamic transcriptional network regulating bile duct formation. Homozygous liver-specific deletion of Hnf1β in mice leads to ductopenia and bile duct dysplasia.2 In humans, heterozygous mutation of HNF1B can result in ductal plate malformations and cholestasis.3
Normal cholangiocytes contain a 7-μm single nonmotile primary cilium, projecting into the lumen and essential for sensory functions. As in all primary cilia, the axoneme contains 9+0 microtubules (compared to 9+1 microtubules in motile cilia).4 Abnormalities in primary cilia lead to a wide variety of diseases affecting different organs, generally classified as ciliopathies.5
We recently showed that mice with homozygous liver-specific Hnf1β knockout mice have absent immunostaining of acetylated tubulin in developing bile ducts, suggesting aberrant ciliogenesis.6 However, the presence of cilia had not yet been investigated in patients or heterozygous knockout mice at the ultrastructural level.
Case 1: A 34-year-old woman was referred because of progressively increasing, mainly cholestatic, liver tests for more than 5 years. She was known to have had diabetes since age 14. Liver biopsy revealed only nonspecific changes and some steatosis. Ultrasound showed atrophy of the pancreas, renal cysts, and a bicornuate uterus, whereas fibroscan confirmed the absence of fibrosis.
Case 2: A 53-year-old man with mild mental retardation was followed for 9 years because of “alcohol-induced” chronic pancreatitis. After 7 years of follow-up he was diagnosed with diabetes mellitus, interpreted as the result of chronic pancreatitis. However, laboratory analysis also revealed mild renal insufficiency, elevated liver tests, and hypomagnesemia, whereas ultrasound revealed normal liver, atrophic pancreas, and renal cysts. Fibroscan suggested mild fibrosis. Despite relatively good diabetic control and alcohol stop for several years, there were increasing, mainly cholestatic, liver tests. A liver biopsy revealed only minor sinusoidal dilatation.
Case 3: A 30-year-old woman was followed with increasing cholestasis. She was known to have poorly controlled diabetes and renal insufficiency. Despite treatment with ursodeoxycholic acid, cholestatic liver tests were increasing. Ultrasound showed renal cysts and pancreatic atrophy. Liver biopsy showed thickened basal membranes around the bile ducts and minor sinusoidal dilatation.
At first sight, these three cases are classical poorly controlled diabetes patients with secondary renal insufficiency (see Table 1 for details). Based on clinical suspicion and ultrasound results, we hypothesized and documented underlying HNF1B deficiency in all three patients. As the bile ducts of these patients were structurally normal and as the mechanism by which HNF1B causes cholestasis is not completely understood, we performed electron microscopy and immunohistochemistry, focusing on bile duct epithelial cells.
Table 1. Patient Characteristics
BMI: body mass index; γGT: gamma-glutamyltransferase; AST: aspartate transaminase; ALT: alanine transaminase; GFR: glomerular filtration rate; HbA1c: hemoglobin A1c.
Age at diagnosis
18.5 – 25
Alkaline phosphatase (U/l)
760 ± 157
1577 ± 432
750 ± 350
188 ± 72
1500 ± 507
272 ± 124
58 ± 36
63 ± 25
37 ± 20
115 ± 54
106 ± 49
37 ± 19
1.39 ± 0.12
1.22 ± 0.12
1.53 ± 0.14
1.58 – 2.55
GFR (mL/min/1.73 m2)
57 ± 6
56 ± 10
41 ± 7
Range HbA1c (%) + mean
6.5 – 9.1 (8.1)
5.5 – 10.9 (7.1)
8.8 – 13.1 (11.1)
1423kbp deletion 17q12 (34.826.185 – 36.249.430)
1427kbp deletion 17q12 (34.822.461 – 36.249.430)
HNF1B/TCF2, hepatocyte nuclear factor 1 homeobox B/transcription factor 2; MLPA, multiplex ligation-dependent probe amplification; MODY5, maturity onset diabetes of the young 5; SOX9, sex determining region Y box 9.
Materials and Methods
Liver biopsies of the three patients and of mice with liver-specific heterozygous (Hnf1βflox/+-Alfp-Cre) or homozygous (Hnf1βdel/flox-Alfp-Cre) deletion of Hnf1β2 were analyzed as described.6 Mutation analysis of all coding exons (exon 1-9) including exon-intron borders was performed, as well as multiplex ligation-dependent probe amplification (MLPA) to detect deletions or duplications. The extent of the identified deletions was analyzed using the Affymetrix SNP array (2.7M platform) and Chromosome Analysis Suite software CytoB-N220.127.116.118.
Patient characteristics can be found in Table 1. All three patients were found to have significant cholestasis, renal impairment, and hypomagnesemia. Diabetic control was relatively poor despite adequate drug treatment, with lifestyle and dietary adjustments. Liver biopsies revealed no structural abnormalities of the bile ducts. Sex-determining region Y box 9 (SOX9) is a biliary-specific transcription factor, and cholangiocytes express high levels of E-cadherin. Immunohistochemistry revealed an accumulation of small SOX9+ E-cadherin+ bile ducts, surrounded by fibrosis, distant from the portal vein (Fig. 1). Electron microscopy showed the absence or paucity of primary cilia on bile duct epithelial cells, with deranged 9+0 microtubuli distribution in the few leftover primary cilia (Fig. 2).
Electron microscopy of heterozygous mice showed a paucity of cilia with an aberrant number of axonemes, or absence of normal cilia on bile duct epithelial cells (Fig. 2), whereas normal bile ducts or bile duct epithelial cells could not be visualized in homozygous knockout mice.
We have previously suggested that Hnf1β-deficient cholangiocytes in mouse embryos fail to develop normal primary cilia, based on immunostaining for acetylated tubulin.6 We now show for the first time that cholestasis in adult patients with heterozygous deletion or mutation of HNF1B is not related to structural defects of intra- or extrahepatic bile ducts, but is associated with the absence of normal primary cilia on cholangiocytes.
We conclude that HNF1B mutations should be considered in patients with unexplained chronic cholestasis, combined with any other feature of HNF1B deficiency syndrome.
The authors thank Prof. Dr. Gert Matthijs and Sigrun Jackmaert (Department of Human Genetics, University Hospitals Leuven) for help with genetic analysis.