Novel missense mutation in VPS33B is associated with isolated low gamma-glutamyltransferase cholestasis: Attenuated, incomplete phenotype of arthrogryposis, renal dysfunction, and cholestasis syndrome

2.1 Subjects and genomic analysis

Subjects were enrolled from 2011 to 2017, with informed consent, under a clinical-diagnosis protocol approved by Children's Hospital and Jinshan Hospital of Fudan University (Shanghai, China) and according to the ethical guidelines of the 1975 Declaration of Helsinki. They underwent routine tests for liver diseases (Qiu et al., 2017). Genomic DNA for sequencing was extracted from ethylenediaminetetraacetic acid (EDTA)-treated peripheral blood cells (QIAamp DNA Blood Mini Kit, Cat. 51106; Qiagen, Germany). Before 2014, two ARC patients (P4, P5 in this study) were diagnosed by the three major symptoms, and then were confirmed biallelic pathogenic variations through Sanger sequencing of all VPS33B exons (Gissen et al., 2006; Li, Zhao, Chen, & Wang, 2014; J. S. Wang, Zhao, & Li, 2014). Since 2014, patients with suspected hereditary cholestasis receive targeted panel sequencing or whole exome sequencing to screen underlying pathogenic gene (Qiu et al., 2017); three more VPS33B mutated patients (P1, P2, and P6) and one VIPAS39 mutated patient (P7) were identified. P1 and P2 presented isolated cholestasis; P6 and P7 presented typical ARC phenotype. P3 was identified by Sanger sequencing within family members of P2 and presented isolated cholestasis.

Twenty patients with primary defect of bile salt export pump (BSEP) and 37 healthy controls in bile acid profiling investigations were described in former studies (Liu et al., 2018; Qiu et al., 2017).

2.2 In-silico variation prediction and genotype-phenotype correlation study

All the VPS33B/VIPAS39 variations in these patients were searched for frequency in public databases as well as in-house databases. Next, online tools were applied to predict their pathogenicity (Schwarz, Cooper, Schuelke, & Seelow, 2014). And then, we compared types of mutations between patients with typical ARC phenotype and isolated cholestasis. Websites of these online tools were presented as below.

Websites for public variant databases:

dbSNP. https://www.ncbi.nlm.nih.gov/SNP/

1000 genomes. http://www.1000genomes.org/

ExAC Browser. http://exac.broadinstitute.org/

Exome Variant Server (ESP6500). http://gvs-1.gs.washington.edu/EVS/

gnomAD. http://gnomad.broadinstitute.org/

ClinVar. http://www.clinvar.com

HGMD. http://www.hgmd.compareac.uk/ac/index.php

Websites for online variation prediction:

SIFT. http://sift.jcvi.org

PolyPhen2. http://genetics.bwh.harvard.edu/pph2/

MutationTaster. http://www.mutationtaster.org

PROVEAN. http://provean.jcvi.org/genome_submit_2.php

BDGP. http://fruitfly.org/seq_tools/splice.html

Swiss-model. https://www.swissmodel.expasy.org/

HOPE. http://www.cmbi.ru.nl/hope/input/

2.3 Construct of mutated VPS33B^{c.1726T>C, p.Cys576Arg} plasmid

To study the impact of the novel missense variation c.1726T>C, (p.Cys576Arg) on VPS33B, we bought wild-type human ORF cloned plasmids of VPS33B (NM_018668.4, pCMV3-C-Flag, Cat. HG16334CF, Sino Biological, Beijing, China) and VIPAS39 (NM_001193315.1 pCMV3-N-HA tag, Cat.HG22032-NY, Sino Biological, Beijing, China) that encoding its interacting protein. Mutated VPS33B plasmid was generated through point mutation by KOD -Plus- Mutagenesis Kit (Cat. 0811, TOYOBO, Osaka, Japan). Forward primer for point mutation was GTGGTCGTACATTCTCTGAGATCTC, and the reverse primer CCAAGAACACCACCAAGATGAGGCG.

2.4 Cell culture and transfection

For western blot study, HEK-293T cells were bred in 6-well plates. In each well, a total of 2 μg plasmids were transferred into when the cells grew near 90% (PolyJet In Vitro DNA Transfection Reagent, Cat. SL100688, SignaGen, Gaithersburg, MD). Plasmid pcDNA3.1 was set as the control. One microgram wild-type and mutated VPS33B plasmid were cotransfected with 1 μg pcDNA3.1, or with 1 μg Wild-type VIPAS39 plasmid, respectively. For the immunofluorescent study, HEK-293T and MDCK cells were bred in 24-well plates with glass coverslips respectively, with similar procedure but transfected with a total of 1 μg plasmids. Cells were harvested 48 hrs after transfection for following experiments.

2.5 Immunofluorescence microscopy

Cells were fixed with 4% paraformaldehyde for 15–30 min and gently washed by phosphate-buffered saline (PBS) 10 min for three times. And then, cells were incubated at 4°C overnight with mouse anti-Flag antibody (1:500, dilute with PBS containing 1% goat serum & 0.3% Triton, Cat. 8146S, CST, Danvers, MA) and rabbit anti-HA antibody (1:800, dilute with PBS containing 1% goat serum and 0.3% Triton, Cat. 3724S, CST, Danvers, MA). After gently washed by PBS containing 0.2% tween20 10 min for three times, cells were stained with fluorescent secondary antibodies for 1 hr (Goat antimouse Secondary Antibody, Alexa Fluor 594 conjugated, Cat. A-11032, Invitrogen, [Shanghai], Shanghai, China; Goat anti Rabbit Secondary Antibody, Alexa Fluor 488 conjugated, Cat.A0423, Beyotime, Shanghai, China), washed with 0.2% tween-20 PBS and stained with DAPI for 5 min (5 μg/ml, dilute with ddH2O, Cat. 1155GR001, NeoFROXX, Einhausen, Germany). Coverslips were put up-side-down on the microslide with Eukitt® Quick-hardening mounting medium for microscopy (Cat. 03989, Sigma-Aldrich, Osterode am Haz, Germany) and photographed with a fluorescent microscope and confocal microscope (Leica SP8).

2.6 Western blot

Five hundred microliters radioimmunoprecipitation assay (Cat. P0013B, Beyotime, Shanghai, China) with protease inhibitor were added to 6-well plates on ice. After 15-30 min, cell lysis was transferred to centrifuge tubes, centrifuged at the speed of 12000r for 5 min. After concentration measurements (Cat. 23227, Pierce™ BCA Protein Assay Kit - Reducing Agent Compatible, Thermo Fisher Scientific, [Shanghai], Shanghai, China), the supernatant was added with SDS loading and boiled for 10 min for sample preparation. The routine procedure of western blot study was performed, using the same primary antibodies in the immunofluorescent study (1:2,000), and the secondary antibodies from Beyotime (Cat. A0208 & A0216 respectively).

2.7 Bile acid analysis in serum

Serum sample was separated from blood by routine centrifugation, ali-quoted (50 μl), freeze-dried and stored at −80°C until bile acid analysis was performed by reversed-phase ultrahigh-performance liquid chromatography/multiple-reaction monitoring-mass spectrometry (UPLC/MRM-MS) with negative ion detection as described previously (Han et al., 2015; Liu et al., 2018; Qiu et al., 2017). The quantitation of bile acids utilized standard substances of 60 bile acids included all of the major bile acids and some minor species with 14 deuterium-labeled analogs of bile acids as internal standards. Serum bile acid profiling was carried out in all available VPS33B/VIPAS39 deficient patients, compared with healthy donors and patients with primary BSEP deficiency as controls.

3.1 Novel missense VPS33B variation c.1726T>C, p.Cys576Arg was shared in patients with isolated cholestatic phenotype

From 2011–2017, 6 VPS33B mutated patients (P1–P6) and one VIPAS39 mutated patient (P7) were identified in our center. They were all term babies from healthy unrelated Han parents (Figure 1). Of interest, P1–P3 presented only low-GGT cholestasis, but lack proof of other two typical manifestations of arthrogryposis and renal dysfunction. Moreover, P1–P3 shared a novel missense variation c.1726T>C, p.Cys576Arg in VPS33B (NM_018668.4 and NP_061138.3): P1 was homozygous, whereas P2 & P3 from another family were compound heterozygous. This novel variation has never been reported in public databases such as dbSNP, 1,000 genomes, ExAC, ESP6500, and gnomAD, nor does in our in-house database or other patients with typical ARC phenotype, predicted pathogenic by online software PROVEAN, SIFT, Polyphen2 and MutationTaster (Schwarz et al., 2014). Besides, no other disease-causing mutation was detected in genes associated with low GGT cholestasis.

For the rest patients (P4–P7) with complete ARC phenotype, they all carried biallelic mutations that were biologically pathogenic, including frameshift mutation, nonsense mutation, and classical splicing mutation (Figure 1). VPS33B mutations in P4-P6 were seen in reported patients, whereas VIPAS39 mutations in P7 were novel.

3.2 Clinical features of VPS33B mutant patients with isolated low-GGT cholestasis

Major clinical features of P1–P7 were shown in Table 1. Clinical course of P1–P3 were detailed below. P4–P7 had typical, complete phenotype of ARC syndrome and all died before the age of 2.

3.3 VPS33B^{c.1726T>C, p.Cys576Arg} caused remarkably declined VPS33B expression and hampered interaction with VIPAS39 in vitro

To further investigate the pathogenicity of the novel VPS33B mutation c.1726T>C, p.Cys576Arg in isolated cholestatic patients, we constructed Flag-tagged wild type VPS33B and mutant plasmids, and HA-tagged VIPAS39 plasmid. Compared with the wild-type, VPS33B^{c.1726T>C, p.Cys576Arg} caused significantly decreased protein expression, no matter overexpressed individually or co-overexpressed with VIPAS39 in HEK-293T cell lines (Figure 2a). Surprisingly, VPS33B^{c.1726T>C, p.Cys576Arg} also resulted in declined VIPAS39 expression when co-overexpressed (Figure 2b). Confocal microscopy repeated the observations above in both HEK-293T and MDCK cell lines and revealed that VPS33B^Cys576Arg cannot co-localize with VIPAS39, indicating hampered interaction (Figure 3).

3.4 Bile acid profiling suggested secondary BSEP dysfunction in VPS33B/VIPARS39 mutants

Previously studies have suggested that the defect of VPS33B/VIPAS39 complex leads to the dislocation of BSEP (or ABCB11) (Ackermann et al., 2014; Bull et al., 2006; Hanley et al., 2017), the main transporter for bile acid exocytosis on the canalicular membrane. In this study, the bile acid profiling in both ARC-isolated cholestasis group and ARC-typical phenotype group (ARC-IC and ARC-TPY, respectively, Figure 4) revealed “typically” cholestatic bile acid profiles with significantly increased total bile acids, total primary bile acids, and total conjugated bile acids(Figure 4a,b,d). The total secondary bile acids, which reflect the bile acid secretion through the canalicular membrane into the lumen via BSEP (Hylemon & Harder, 1998), were decreased in ARC group, resemble to those with primary BSEP defects. Of note, patients with isolated cholestasis had a much higher concentration of secondary bile acids than those with typical ARC symptoms.

4.1 Existence of incomplete phenotype of ARC syndrome with isolated low GGT cholestasis

In this study, we identified three VPS33B mutated patients (P1–P3) presenting isolated cholestasis. Different from previously reported ARC patients and P4–P7 with the typical, complete phenotype (Gissen et al., 2006), P1–P3 manifested only low-GGT cholestasis after birth, without other two major characteristics of arthrogryposis or kidney dysfunction. Besides, renal tubular function test of urine amino-acids profiling (Matsumoto Institute of Life Sciences, Beijing, China) revealed only one slight elevated amino-acid in P2 and P3 with isolated cholestasis, whilst multiple elevated amino-acids in P5 and P7 with typical ARC phenotype (10-fold over the normal controls, Table 1). Moreover, P1-P3 live much longer than that with typical ARC phenotype (Gissen et al., 2006; Smith et al., 2012), and have less accompanied symptoms such as platelet dysfunction or recurrent infection, indicating an attenuated phenotype.

4.2 VPS33B variation c.1726T>C, p.Cys576Arg is associated with the attenuated ARC phenotype of isolated cholestasis

As described, a missense VPS33B variation c.1726T>C, p.Cys576Arg was shared in all patients with isolated cholestasis without other explainable mutations: P1 was homozygous and P2 and P3 were compound heterozygous. This mutation was never reported before in public/in-house databases or patients with complete ARC phenotype proved pathogenic by online predicted software and in-vitro study. Like genotype-phenotype correlation between VPS33B mutation c.1225+5G>C and attenuated ARC phenotype, a strong association was suggested between VPS33B ^{c.1726T>C, p.Cys576Arg} and isolated cholestasis.

The previous study suggested that mutations in patients with severe, complete ARC phenotype caused absent protein expression or complete abolished interaction with VIPAS39, whereas mutations in patients with attenuated phenotype would be less severe with partially preserved protein expression and function (Cullinane et al., 2009b; Smith et al., 2012). The observation in our patients with complete ARC phenotype was consistent with the inference, all carrying biallelic recurrent mutations that predicted with absent protein expression. But strange for the shared mutation c.1726T>C, p.Cys576Arg in isolated cholestatic phenotype; it did result in partial protein expression but with completely abolished interaction with VIPAS39, which means loss of VPS33B function. Thus, further investigations were required.

4.3 Bile acid profile suggested secondary BSEP defect and more residual function of VPS33B^p.Cys576Arg mutant

Physically, VPS33B colocalizes and interacts with VIPAR to form the VPS33B/VIPAR complex, interacting with endosome marker RAB11A and playing a vital role in the intracellular trafficking and the hepatocyte polarity maintenance (Golachowska, Hoekstra, & van IJzendoorn, 2010; Wakabayashi, Dutt, Lippincott-Schwartz, & Arias, 2005; C. Wang et al., 2018). For VPS33B/VIPAR complex defect, aberrant location of canalicular membrane proteins of BSEP, MDR3, and CEA were observed in VPS33B/VIPAR deficient patients and Vps33b knockout mice (Ackermann et al., 2014; Bull et al., 2006; Hanley et al., 2017). In our study, bile acid profiling in six patients with ARC syndrome showed decreased total secondary bile acids, resemble to those with primary BSEP defects and secondary BSEP dysfunction due to MYO5B defect and consist with former studies. Another notable phenomenon is that patients with isolated cholestasis and longer lifetime had much higher secondary serum bile acids than those with typical ARC phenotype, reflecting more residue function of VPS33B^p.Cys576Arg and require further investigations.

4.4 Follow-up of kidneys and other possible involved organs were required for VPS33B mutated patients with isolated cholestasis

So far, we have not observed typical kidney and joint involvement in P1–P3 with isolate cholestasis. However, the slightly elevated urine microglobulin in P2 and P3 and progressive dextroscoliosis in the oldest patient P3 reminded us of the possibility that joint and renal damage might progress silently. Progressive kidney problems were indeed described in elder patients with ARC syndromes (Rosales et al., 2018), and VPS33B/VIPAS39 complex was reported essential for collagen homeostasis that is required for normal joint structure (Gruber et al., 2017; Rogerson & Gissen, 2018). With the extended life expectancy in patients with attenuated ARC phenotype of isolated cholestasis, follow-ups for kidney and other possible involved organs were indispensable.