Introduction Hereditary haemochromatosis (HH) caused by a homozygous p.C282Y mutation in haemochromatosis (HFE) gene has been well documented. However, less is known about the causative non-HFE mutation. We aimed to assess mutation patterns of haemochromatosis-related genes in Chinese patients with primary iron overload.
Methods Patients were preanalysed for mutations in the classic HH-related genes: HFE, HJV, HAMP, TFR2 and SLC40A1. Whole exome sequencing was conducted for cases with variants in HJV signal peptide region. Representative variants were analysed for biological function.
Results None of the cases analysed harboured the HFE p.C282Y; however, 21 of 22 primary iron-overload cases harboured at least one non-synonymous variant in the non-HFE genes. Specifically, p.E3D or p.Q6H variants in the HJV signal peptide region were identified in nine cases (40.9%). In two of three probands with the HJV p.E3D, exome sequencing identified accompanying variants in BMP/SMAD pathway genes, including TMPRSS6 p.T331M and BMP4 p.R269Q, and interestingly, SUGP2 p.R639Q was identified in all the three cases. Pedigree analysis showed a similar pattern of combination of heterozygous mutations in cases with HJV p.E3D or p.Q6H, with SUGP2 p.R639Q or HJV p.C321X being common mutation. In vitro siRNA interference of SUGP2 showed a novel role of downregulating the BMP/SMAD pathway. Site-directed mutagenesis of HJV p.Q6H/p.C321X in cell lines resulted in loss of membrane localisation of mutant HJV, and downregulation of p-SMAD1/5 and HAMP.
Conclusion Compound heterozygous mutations of HJV or combined heterozygous mutations of BMP/SMAD pathway genes, marked by HJV variants in the signal peptide region, may represent a novel aetiological factor for HH.
- hereditary haemochromatosis
- heterozygous mutation
- non-HFE gene
- signal peptide region
Statistics from Altmetric.com
Hereditary haemochromatosis (HH) is an inherited disease characterised by excessive absorption and toxic accumulation of iron in various organs such as liver, pancreas, myocardium and skin. This causes lesions in the affected organs.1 The liver is the most affected organ where the toxic accumulation of iron may lead to cirrhosis and hepatocellular carcinoma.2
Type I, II and III HH are all diseases that show autosomal recessive inherence; while type IV is dominantly inherited.1 3 The most frequent type of HH (type I) is caused by a homozygous p.C282Y mutation in haemochromatosis (HFE) gene or compound heterozygous HFE p.C282Y/H63D mutations with a morbidity of 1/220–250 in the Caucasian population.1 These mutations occur in 85%–90% and 3%–5% of Caucasian patients with HH, respectively.1 3 Mutations in other types of HH include HJV p.G320V and p.I222N in type IIA HH,4 HAMP p.C78T in type IIB HH,5 TFR2 p.L99V, p.A75V and p.V277L in type III HH6 and SLC40A1 p.V162del in type IV HH.7 Mutations in BMP/SMAD pathway gene BMP6, which is ultimately leading to the downregulation of hepcidin gene transcription, such as BMP6 p.L96P, have also been reported recently.8 In addition, next generation sequencing has identified novel genetic variants involved in iron metabolism, such as HEPHL1 p.D136V and GNPAT p.D519G.9 10
In contrast, in East Asian countries, the morbidity of HH is less reported and the HFE p.C282Y mutation is rarely identified. There has been only one isolated case with p.C282Y homozygous HH reported in a Japanese case.11 Also, some individual case reports have described variants in the four classic non-HFE genes. HJV homozygous mutation p.D249H and p.Q312X were identified in two cases from Japan,12–14 while a p.I287S was identified in a Chinese case.15 HJV heterozygote mutation p.D249H, p.Q312X and A75V were also identified in a Japanese case,16 while a compound heterozygote for HJV p.Q6H/C321X and I281T was also identified in a Chinese case.17 However, one case harboured only heterozygote HJV p.Q6H/C321X, and it was proposed that previously unrecognised environmental or other genetic factors may have interacted with the heterozygous genotype in the patient.18 Mutation in HAMP was rarely identified although a p.R75X in homozygosity was identified in a Japanese case.19 TFR2 homozygous mutation p.L490R, p.V561X and p. AVAQ594-597del were identified in Japanese cases,13 20 21 whereas heterozygous mutation p.L367R/2008-9delAC and p.A356fs/p.G430R were identified in a Japanese and Chinese case, respectively.13 15 However, single TFR2 p.R481H heterozygous mutation was identified in a Taiwanese case.22 Regarding the SLC40A1 mutation, p.D157A, p.V162del, p.D175A, p.R489S and p.H507R were identified in Japanese cases,13 16 23 24 whereas p.W158C, p.S209L, p.G267D, p.C326F and p.G490D were identified in Chinese cases.15 25–28 However, no variant has been identified in the other HH-related genes.
Although these sporadic case reports have documented a series of variants in the classical non-HFE genes, little is known about the general patterns of HH-related gene mutation in Chinese patients with HH. There were still many unexplained cases with iron overload, particularly those in whom no causative mutation in the classic known HH-related genes can be identified.
Therefore, in the present multicentre clinicogenetic study, we assessed the mutation patterns of haemochromatosis-related genes and we analysed the function and explored the mechanisms of novel mutations in a Chinese cohort with primary iron overload.
Patients with iron overload enrolled at the China Registry of Genetic/Metabolic Liver Diseases (CR-GMLD, ClinicalTrials.gov: NCT03131427) since October 2014 were selected to screen for mutation in HH-related genes.
Primary iron overload was diagnosed based on the American Association for the Study of Liver Diseases 2011 practice guidelines on haemochromatosis1 29: (1) transferrin saturation ≥45% and/or elevated ferritin (>300 ng/mL in men and postmenopausal women and >200 ng/mL in premenopausal women); (2) iron overload in liver and/or spleen on MRI of liver or on liver histology; (3) exclude causes of secondary iron overload, such as alcoholic or other chronic liver disease, iron-overloading anaemia and parenteral iron overload. The end-organ manifestations of liver diseases and clinical characteristics of patients were diagnosed and determined according to related guidelines.
Sanger sequencing and in silico analysis of mutations in the classic haemochromatosis-related genes
Genomic DNA was extracted from whole blood using a Genomic DNA Purification Kit (Qiagen, Valencia, CA). All exons of HFE, HJV, HAMP, TFR2 and SLC40A1 were PCR amplified with their associated boundary regions using primers designed with primer3 software (online supplementary table 1). PCR amplification was performed in an ABI Veriti 96 PCR cycler (Applied Biosystems, MA, USA). PCR products were sequenced in forward and reverse orientations using an automated ABI 3730 DNA sequencer (ABI).
Supplementary file 1
Three predictors, Polyphen-2 (http://genetics.bwh.harvard.edu), SIFT (http://sift.jcvi.org/) and Mutation Taster (http://www.mutationtaster.org/), were used to predict the functional consequence of the identified novel variants, and the result ‘deleterious’ had to be predicted by at least two of the three predictors.
Whole exome sequencing to identify mutations accompanying HJV p.E3D
Using approximately 1 µg of genomic DNA, a targeted exome library with an insert size of 150–200 bp was constructed by an exome capture strategy, using a GenCap custom exome enrichment kit (MyGenostics, Beijing, China). An Illumina HiSeq 2000 platform was used to generate paired-end 100 bp raw reads from each enriched library according to the manufacturer’s protocol. The 100 bp paired-end reads were aligned against NCBI build 37 of the human genome using Burrows Wheeler Aligner. Duplicate reads were marked, local indel realignment performed and base-quality scores recalibrated for each sample with the Genome Analysis Toolkit (GATK).
Novel point mutations were identified using MuTect, while indel variants were identified using Somatic Indel Detector in the GATK. The identified potential pathogenic variants were confirmed by Sanger sequencing.
Site-directed mutagenesis for HJV and DENND3 variants
To create an HJV-Flag plasmid, we amplified HJV from human cDNA and cloned the fragment into the XhoI/KpnI sites of the CMV-MCS-3FLAG-SV40-Neomycin vector. An HJV-Flag p.Q6H construct was generated using the Gene Tailor Site-Directed Mutagenesis System (Invitrogen, Carlsbad, CA, USA). To create the HJV-Flag p.C321X construct, we amplified the 1–321 amino acids fragment of HJV from a human cDNA and cloned the fragment into the XhoI/KpnI sites of the CMV-MCS-3FLAG-SV40-Neomycin vector. The HJV-Flag p.Q6H+C321X construct was generated from the HJV-Flag p.C321X variant using the Gene Tailor Site-Directed Mutagenesis System.
To create the DENND3 plasmid, we amplified DENND3 from human cDNA and cloned the fragment into the XhoI/KpnI sites of the CMV-MCS-3FLAG-SV40-Neomycin vector. A DENND3 L708V construct was generated using the Gene Tailor Site-Directed Mutagenesis System (Invitrogen).
siRNA interference of SUGP2
Sense sequences for SUGP2 siRNA1 (siRNA ID s19758), SUGP2 siRNA2 (siRNA ID s19757) and negative control siRNA were purchased from Thermo Fisher (Thermo Fisher Scientific, MA, USA).
Cell culture and transfection
Human kidney cell line 293T, human liver cell line QSG and hepatocellular carcinoma cell lines Hep3B, Huh7 and HepG2 were obtained from the Cell Resource Center of the Chinese Academy of Medical Science (Beijing, China). The cell lines were cultured as described previously.30 Cells were cultured to 80% confluency and were then transfected using Lipofectamine 3000 following the manufacturer’s instructions (Invitrogen). The 293T, QSG and Hep3B cells were transfected respectively with HJV-Flag: HJV-Flag p.Q6H, HJV-Flag p.C321X and HJV-Flag p.Q6H+C321X constructs. Both Huh7 and HepG2 were transfected with DENND3 and DENND3 L708V constructs, and SUGP2 siRNA. The culture medium was changed 6 hours after transfection. Cells were harvested 24 or 48 hours after transfection for different assays.
Immunofluorescence staining for cellular localisation of mutant HJV proteins
Immunofluorescence analysis was conducted as described previously.30 Cells were incubated with a primary antibody directed against rabbit anti-HJV (1:200, Invitrogen) or mouse anti-Flag (1:1000, Santa Cruz Biotechnology, CA, USA) at 4°C overnight. After three 5 min washes with phosphate buffered saline (PBS), cells were incubated with anti-mouse Alex 488 (1:500, Invitrogen) or a mixture of anti-mouse Alex 488 and anti-rabbit Alex 594 (1:500, Invitrogen) conjugated secondary antibodies for 1 hour at room temperature. For the F-actin, following three 5 min washes with PBS, cells were incubated with Rhodamine conjugated phalloidin at 5 U/mL (Molecular Probes, USA) for 30 min. After additional PBS washes, cells were mounted on a slide in mounting medium (Molecular Probes). Cells were examined and photographed using an FV 300 confocal microscope (Olympus, Japan).
Western blot analysis of HJV, p-SMAD1/5 and TFR2
Western blot analysis was conducted as described previously.30 31 Membranes were incubated with rabbit anti-p-SMAD1/5 (1:1000, Invitrogen), rabbit anti-HJV (1:1000, Invitrogen), mouse anti-TFR2 (1:1000, Santa Cruz) or mouse anti-GAPDH antibodies (1:5000, Santa Cruz) overnight at 4°C, followed by incubation with horseradish peroxidase (HRP)-conjugated goat anti-mouse (1:5000 dilution, Santa Cruz Biotechnology) for 1 hour at 37°C. Target proteins were detected using Immobilon Western Chemiluminescent HRP Substrate (Millipore, USA).
Real-time PCR to determine HJV, HAMP, SUGP2 and DENND3 expression
The isolation of total RNA from cell lines and the real-time PCR assays were conducted as described previously.31 The sequences of the primers used were as follows: 5′-TTTTCCCACAACAGACGGGA-3′ and 5′-CTCCTTCGCCTCTGGAACAT-3′, or 5′-TGTTTTCCCACAACAGACGGG-3′ and 5′-CGCAGCAGAAAATGCAGATGG-3′ (137 bp, for electrophoresis analysis only) for HAMP, 5′-GCTAACCCTGGGAACCATGTG-3′ and 5′-CCCAACACAGAGCTGCAGGT-3′ for HJV, 5′-CATCGACCAGCTTGTGAAAC-3′ and 5′-CTGCATTTCTGCCAACTTCA-3′ for SUGP2, 5′-CTGTTCGAGGCCTTGACTGT-3′ and 5′-TTGTCTTGACGGAGCTGGAC-3′ for DENND3, and 5′-GAGTCAACGGATTTGGTCGT-3′ and 5′-GAGTCAACGGATTTGGTCGT-3′ for GAPDH as control.
We used SPSS software V.18.0 (SPSS) to conduct all statistical comparisons. The data are presented as the means±SD. Non-parametric statistics were applied for comparison of results in all experiments unless stated otherwise. The Wilcoxon signed-rank test was applied to paired data, and the Mann-Whitney test to unpaired data. P values of less than 0.05 were considered to be statistically significant.
Patients and clinical data
Twenty-two patients with primary iron overload from the CR-GMLD, including 19 probands, were selected to screen for mutations in HH-related genes (online supplementary figure 1). All the probands with primary iron overload were validated by liver biopsy and/or MRI examinations. Clinical data of primary iron-overload cases and relativesare presented in table 1 and online supplementary tables 2 and 3.
Non-HFE variants identified in primary iron-overload cases
In 21 of the 22 primary iron-overload cases (95.5%), Sanger and exome sequencing identified at least one non-synonymous non-HFE variant in any one of the HH-related genes. No HFE [NM_000410.3] p.C282Y or HAMP [NM_021175.3] variants were identified (table 2).
Seven cases out of the 22 cases can be explained by diallelic mutations in the known genes: two cases (H5 and H46) carried homozygous HJV [NM_213653.3] (c.309C>G, p.F103L) or p.Q312X; five cases (H13, H13-1, H22, H40 and H31) carried compound heterozygous variants in HJV or TFR2 [NM_003227.3]. Four out of the 22 cases harboured single variant in SLC40A1 [NM_014585.5], representing a dominantly inherited HH of type 4 (H9, H11, H34 and H49). However, seven out of the 22 cases were identified with combined heterozygous variants in different HH-related genes (H1, H1-2, H22-1, H25, H35, H35-1 and H31), and only one single variant or no variant in HH-related gene was identified in the remaining three or one cases (H2, H10, H45 and H19), respectively (table 2).
Mutations identified in the known HH-related gene
Previously reported mutations in the known non-HFE genes, that is, HJV p.Q312X, p.C321X and p.I281T, TFR2 p.A75V, and SLC40A1 p.V162del, were identified in seven primary iron-overload cases (7/22, 31.8%) (table 2).
Meanwhile, variants which were identified for the first time in HH, including HJV (c.9G>C, p.E3D), p.F103L (c.311A>G, p.H104R) and (c.820G>A, p.V274M), TFR2 (c.905C>A, p.A302E) and (c.2234T>G, p.L745R), and SLC40A1 (c.997T>C, p.Y333H), IVS3+10delGTT and IVS1-8C/G, TMPRSS6 [NM_001289000] (c.992C>T, p.T331M) and BMP4 [NM_001202] (c.806G>A, p.R269Q), were identified in 13 primary iron-overload cases (13/22, 59.1%) (table 2). In silico analysis indicated HJV p.F103L, p.H104R and p.V274M, TFR2 p.L745R, and SLC40A1 p.Y333H to be ‘deleterious’ (table 3). The two SLC40A1 variants, IVS3+10delGTT and IVS1-8C/G, are typical splicing variants, with base alteration within 1–12 bp of adjacent intronic areas.
Variants identified in the novel HH-related genes
Whole exome sequencing identified variants in two novel HH-related genes, which may be associated with HH phenotype, that is, SUGP2 [NM_001017392] (c.1916G>A, p.R639Q) in all the three cases with HJV p.E3D (3/22, 13.6%), and DENND3 [NM_014957] (c.2122C>G, p.L708V) in four cases (4/22, 18.2%) (table 2). In silico analysis indicated the SUGP2 p.R639Q and DENND3 p.L708V to be ‘deleterious’ (table 3).
Genetic variants in the signal peptide region of HJV might indicate a variant hot spot for HH
Variants in the signal peptide region of HJV, p.Q6H and p.E3D, were identified in four cases (4/22, 18.2%) and five cases (5/22, 22.7%) of primary iron overload, respectively (table 2, figure 1, online supplementary figure 2A). As these two variants were identified in nearly half of the cases of primary iron overload (40.9%), they may represent a variant hot spot. HJV p.Q6H has not been identified in the general population.18 However, based on the 1000 Genomes Project, HJV p.E3D has a general population minor allele frequency of 0.26% (rs12025510, p<0.0001, table 3).
In addition, a synonymous variant, HFE p.P7P (CCG→CCA), was identified in the signal peptide region of HFE in a primary iron-overload case (H31) (online supplementary figure 2B).
The HJV p.Q6H variant was accompanied by compound heterozygous HJV variants
Pedigree analysis of cases with HJV p.Q6H showed that HJV p.Q6H and p.C321X variants were located in the same allele, but that the single allele with HJV p.Q6H/C321X did not have the iron-overload phenotype (figure 1A). The disease phenotype was identified in cases carrying a third variant in the second allele of the HJV gene, such as HJV p.V274M (case H40), p.I281T (cases H13 and H13-1) and p.H104R (case H22) (figure 1A). Cases H13, H13-1 and H22 showed the classical phenotype of type II HH with an early age of disease onset (18, 26 and 27 years, respectively).
The HJV p.E3D variant was accompanied by combined heterozygous variants in BMP/SMAD pathway genes
To analyse if HJV p.E3D is accompanied by any other potential pathogenic variants in other iron metabolic related genes (as in the case of HJV p.Q6H), whole exome sequencing was conducted on the three probands carrying HJV p.E3D. Two of the three probands carried a novel predicted deleterious variant in BMP/SMAD pathway genes: heterozygous p.T331M in TMPRSS6 (case H25) and p.R269Q in BMP4 (cases H35 and H35-1)9 (tables 2 and 3, figure 1B, online supplementary figure 3, supplementary tables 4–7). For case H1, no variant was identified in the known BMP/SMAD pathway genes, but a homozygous p.L708V in DENND3 was identified. DENND3 was reported to function in TFR regulation32 33 (tables 2 and 3, online supplementary figure 3, supplementary tables 4–8).
Interestingly, one novel variant, p.R639Q in SURP and G-Patch Domain Containing 2 (SUGP2), which functions in mRNA splicing,34 was identified in all cases with HJV p.E3D, but not in any other cases without HJV p.E3D (figure 1B, online supplementary figure 3, supplementary table 7). In silico analysis showed the SUGP2 p.R639Q variant to be deleterious (table 3). Pedigree analysis of the probands showed that the combined heterozygous mutations of SUGP2 p.R639Q with HJV p.E3D/BMP4 p.R269Q (H35) or HJV p.E3D/TMPRSS6 p.331M (H25) may be associated with the HH phenotype (figure 1B).
The HJV p.Q6H+C321X variants affect the membrane localisation of mutant HJV and regulation of the BMP/SMAD pathway
Immunofluorescence analysis showed that the HJV p.C321X and HJV p.Q6H+C321X mutants lost almost all of their membrane localisation and that their levels were increased in the cytoplasm. However, no or only slightly alterations were observed for the HJV p.Q6H mutant in 293 T cells in which endogenous HJV was not expressed (figure 2A), or QSG cells with weak endogenous HJV expression (figure 2B).
To evaluate the effects of the HJV p.Q6H and p.C321X variants on the BMP/SMAD pathway, Western blot analysis of p-SMAD1/5, a molecule downstream of HJV,35 36 was conducted. Because HJV mutations may lead to a soluble form of HJV, which competitively interacts with BMP receptors,35 36 the effect of endogenous HJV on the detection of p-SMAD1/5 levels and HAMP expression may not to be ruled out.
In the two cell lines, QSG and Hep3B, HJV p.C321X and HJV p.Q6H+C321X variants consistently led to a decrease in the level of p-SMAD1/5 and HAMP expression. However, for the HJV p.Q6H variant, a decreased p-SMAD1/5 level was observed, but only slightly decreased HAMP expression was detected, suggesting the HJV signal peptide variant p.Q6H may have minor effects on the BMP/SMAD pathway (figure 2C–E).
Silencing SUGP2 expression downregulated the level of p-SMAD1/5 and HAMP expression: a novel function in the downregulation of the BMP/SMAD pathway
To evaluate if SUGP2 functions in downregulation of the BMP/SMAD pathway, SUGP2 expression was silenced by transfection of siRNA into cell lines, and the levels of p-SMAD1/5 and HAMP were determined. In the two cell lines, Huh-7 and HepG2 with relatively high HAMP expression, the silencing of SUGP2 consistently led to a decrease in the level of p-SMAD1/5 and HAMP expression (figure 3A,B), suggesting that SUGP2 functions as a BMP/SMAD pathway gene.
In addition, a decrease in the level of p-SMAD1/5 and TFR2 was observed in the Huh-7 cell line transfected with the DENND3 and further in DENND3 p.L708V constructs, while the decrease in HAMP expression was not observed in the Huh-7 cell line, but observed in HepG2 cell line (figure 3C).
Reports of HH morbidity in Asian countries are rare and the general pattern of mutation in haemochromatosis-related genes in these populations remains unclear. Through the mutation analysis of 22 primary iron-overload cases, we were able to assess the mutation pattern of haemochromatosis-related genes in China. We found none of the investigated cases carried the classical HFE p.C282Y mutation, whereas known mutations associated with typical HH phenotypes, and novel potential pathogenic variants were frequently identified in non-HFE genes. Our results indicate that the compound heterozygous mutations of HJV or combined heterozygous mutations of BMP/SMAD pathway genes, marked by HJV variants in the signal peptide region, may constitute a major aetiological factor for HH in China. Notably, independently of identified mutation, the patient with haemochromatosis must be treated by phlebotomy or chelation therapy when phlebotomies are medically contraindicated according to recently published guidelines.29 It would be more helpful if we have specific recommendations for patients with non-HFE haemochromatosis in the future.
Recent studies have revealed a diversity of genetic characteristics in patients with HH of different races, and a significant and growing number of patients with non-HFE-related HH are being identified in Europe and South America.37–39 Similar to China, some countries like Brazil have low percentage of HFE p.C282Y homozygous patients but considerable number of patients with non-HFE mutations such as HJV p.G320V and HAMP 5′-UTR (g.47G>A).40 41 In East Asian countries, some individual case reports, mainly from Japan, have described variants in each of the four classic non-HFE genes, respectively.12–28 In the present study, we identified novel homozygous HJV p.F103L in HH and the known homozygous HJV p.Q312X mutation, which is first identified in Chinese iron-overload cases.42 The present study also identified novel heterozygous mutations in SLC40A1 for type IV HH, which were dominantly inherited. These included SLC40A1 p.Y333H, IVS1-8C/G and the IVS 3+10delGTT, which is the first SLC40A1 splicing variant identified in patients with type IVB HH.43 In addition, the known SLC40A1 p.V162del mutation,42 which is novel in Chinese iron-overload cases, was also identified.
However, unlike typical monogenic disorders with homozygous mutations in target genes in autosomal recessive inherited disease, the majority of cases in the present study carried combined heterozygous, potentially pathogenic variants in HJV gene located in different alleles or several BMP/SMAD pathway genes in different genomes of parents (figure 1). All cases with HJV p.Q6H had a common HJV p.C321X mutation and a potential pathogenic variant in the other HJV allele, such as the known p.I281T,17 or one of the two novel identified variants, p.V274M and p.H104R. These findings indicate a correlation between compound heterozygous mutations in HJV and the HH phenotype, as described in a previous case report.17
Similarly, we also found by whole exome sequencing that in all the three HH cases with HJV p.E3D, the variant was accompanied by novel potentially pathogenic variants in BMP/SMAD pathway genes (TMPRSS6 p.T331M and BMP4 p.R269Q) and by a common SUGP2 p.R639Q variant. TMPRSS6 has been reported as a BMP/SMAD pathway gene and has a role in the release of soluble form of HJV leading to the downregulation of hepcidin.9 18 TMPRSS6 mutations have been reported in iron-refractory iron deficiency anaemia.44 In the present study, TMPRSS6 p.T331M was first identified in HH and may act as an active mutation, although further biological functional study is needed. Cellular and animal models have identified BMPs, particularly BMP6, as compelling candidates for the class HH-related genes, but until recently, evidence in humans was limited to the possible role of common SNPs in BMP2 and BMP4 as modifiers of classical HFE-related HH.45 Most recently, BMP6 mutations were identified in HH, such as BMP6 p.L96P, p.E112Q and p.R257H.8 In the present study, BMP4 p.R269Q was first identified in an HH case. SUGP2 has been known functioning in mRNA splicing,34 but its role in the pathogenesis of HH remains unknown. Through SUGP2 siRNA cell models, the present study showed for the first time that the SUGP2 may have a novel role in downregulation of the BMP/SMAD pathway and hepcidin expression. It is interesting that heterozygous SUGP2 p.R639Q was only identified in cases with HJV p.E3D, suggesting a similar link as for HJV p.Q6H/p.C321X.
For the case H1, except for HJV p.E3D and SUGP2 p.R639Q, no variant in known BMP/SMAD pathway genes was identified; however, a homozygous DENND3 p.L708V was identified. DENND3 has been reported functioning in the degradation of TFR1 in mouse embryonic fibroblast through the downregulation of rab12 and may play a role in the cellular iron homeostasis.32 33 In the present study, we identified for the first time a homozygous mutation DENND3 p.L708V in one case, and heterozygous mutation in three cases (4/22, 18.2%). Our functional studies have also shown the overexpression of DENND3, and the activating mutation DENND3 p.L708V, may affect the regulation of p-SMAD1/5 levels, TFR2 or HAMP expression, suggesting a possible role in the BMP/SMAD pathway in addition to TFR regulation. However, heterozygous variants in other currently unknown BMP/SMAD pathway genes cannot be ruled out, and the combination of altered BMP/SMAD and TFR pathways may play a role in the phenotype of case H1.
Thus, our results indicate a possible correlation of combined heterozygous mutations in HJV or BMP/SMAD pathway genes with the HH phenotype, and may provide a genetic basis for the diagnosis of HH in China. Consistently, recent studies have shown that autosomal recessive diseases can be caused by compound heterozygous genotypes.46 47
Another feature of the present study is that genetic variants in the signal peptide region of HJV, which is a transmembrane protein, were frequently identified. Localisation of a transmembrane protein within a membrane is critical for its function and usually depends on an N-terminal signal peptide, which is normally composed of 5–30 amino acids.48 Variants in the signal peptide of some transmembrane proteins may lead to loss of membrane localisation and are associated with the pathogenesis of several diseases that show autosomal recessive inheritance.48 49 HJV p.Q6H and p.E3D variants are not located in a significant region of the signal peptide as predicted, and our functional analysis showed no or only slight alterations of the cellular localisation of the HJV p.Q6H mutant protein, as well as slight downregulation of p-SMAD1/5 and HAMP expression. However, there is a clear correlation between HJV p.Q6H and p.E3D variants and HH, suggesting an involvement of mutations in the HJV or BMP/SMAD pathway genes with HH.
It is worth noting that no variant was identified in any of the HH-related genes in one case (1/22, 4.6%), and that three cases (3/22, 13.6%) only had one single variant in HFE (p.H63D) or TFR2 (p.I238M), suggesting that pathogenic variants in other HH-related genes may exist in those cases. In recent years, next generation sequencing analysis of HH cases has identified genetic variants in a series of novel iron metabolism-related genes,9 50 and in the current study two novel HH-related genes, SUGP2 and DENND3, were identified by exome sequencing. However, genome-wide studies require a sufficient number of cases to identify common variants associated with diseases such as HH; therefore, we plan to conduct further exome sequencing with more cases based on our CR-GMLD network.
In summary, the compound heterozygous mutations of HJV or combined heterozygous mutations in BMP/SMAD pathway genes, marked by HJV variants in the signal peptide region, may represent a novel aetiological factor for HH.
TL, WZ and AX contributed equally.
Contributors JH, XO and JJ designed the study, analysed and interpreted the data, and wrote the manuscript. TL, WZ and AX provided patients' samples and clinical data, analysed and interpreted the data, and wrote the paper. TL, WZ, AX, YL, DZ, BZ and XL did the experiments, and analysed and interpreted the data. XZ, YW, XW, WD and QW provided, analysed and interpreted patients' samples and clinical data. HX, JZ, RZ, LZ, YD, LL, YC, JL, SZ and WW provided patients' samples, clinical data or both. XO, JJ and HY advised on the conception, designed the study and supervised the project. JH conceptualised and designed the study, supervised the project and revised the paper. All authors vouch for the respective data and analysis, approved the final version and agreed to publish the manuscript.
Funding This study was supported by grants from the National Key Technologies R&D Program (No 2015BAI13B09), the Beijing Natural Science Foundation (No 7132058, 7182039) and the National Natural Science Foundation of China (No 81650014).
Competing interests None declared.
Patient consent Obtained.
Ethics approval Clinical Research Ethics Committee of Beijing Friendship Hospital, Capital Medical University (No 2016-P2-061-01).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.