Down-regulation of hepcidin in porphyria cutanea tarda

Porphyria cutanea tarda (PCT) is due to a reduction in the activity of the heme biosynthetic enzyme uroporphyrinogen decarboxylase (UROD) in hepatocytes. Reduced activity of UROD leads to accumulation of the porphyrinogen substrates, which are then oxidized to the corresponding porphyrins. Approximately 25% of patients with PCT are heterozygous for UROD mutations and UROD activity and protein are half-normal in all tissues (familial PCT [F-PCT]).¹  Most heterozygotes do not express a disease phenotype. When the disease phenotype is expressed, activity of UROD in hepatocytes is decreased to rate-limiting levels, yet UROD protein levels do not change. The more common form of PCT is not associated with mutations in UROD, and UROD activity and UROD protein are normal in all tissues (sporadic PCT [S-PCT]). When the disease phenotype is expressed, activity of UROD in hepatocytes is decreased to the same rate-limiting levels seen in disease-expressing patients with F-PCT. The reduction in specific activity of UROD in both S-PCT and F-PCT is due to the generation of an inhibitor of UROD in hepatocytes.²  Generation of the inhibitor is highly dependent on the presence of excess hepatic iron.³  The key role of iron in the pathogenesis of PCT is emphasized by the finding that iron removal by phlebotomy therapy ameliorates all of the clinical manifestations of the disease, including production of the inhibitor. Approximately 20% of patients with PCT are homozygous for the HFE C282Y mutation, but the cause of excess hepatic iron in the remaining cases is not known.^1,4,5 

Homozygosity for a cysteine to tyrosine change at amino acid position 282 (C282Y) in the HFE protein⁶  is found in more than 90% of patients with hereditary hemochromatosis (hh).^7,8  Most remaining occurrences can be explained either by compound heterozygosity for the C282Y mutation and a second HFE mutation resulting in an H63D substitution, or mutations in HAMP, HJV, transferrin receptor 2 (TFR2), and ferroportin (FPN).^9–11  With the exception of FPN, all of these genetic abnormalities affect transcription of HAMP, resulting in failure to appropriately up-regulate HAMP expression as hepatic iron overload develops. The fully penetrant hh phenotype occurs only in a subpopulation of HFE C282Y homozygotes.^12–16  The reason for incomplete penetrance is unknown, but pedigree studies have suggested that other genes modify the phenotype, perhaps by further affecting HAMP expression.¹³  Several pedigrees have been identified in which heterozygosity for mutations of HAMP or HJV appeared to magnify the effect of homozygosity for the C282Y mutation of HFE.^17–20  Furthermore, mouse models clearly indicate that genetic background influences the degree of iron loading associated with mutations in the Hfe gene.^21,22 

We sought mutations in HJV and HAMP in 96 patients with PCT with and without HFE mutations and in 88 patients with hh with marked hepatic iron overload. All patients with hh were homozygous for the HFE C282Y mutation. We hypothesized that if polymorphisms in HJV, HAMP, or both were associated with a highly penetrant phenotype in patients with hh, the same alleles might be overrepresented in patients with PCT who had no HFE mutations. We further hypothesized that hepatic siderosis in PCT might be the result of dysregulation of genes involved in iron metabolism. To investigate this possibility, hepatic expression of HAMP and the iron-related genes FPN, SMAD4, neogenin (NEO), HJV, and transferrin receptors 1 and 2 (TFR1, TFR2) were compared between a subset of patients with PCT with and without HFE mutations and a control pool of patients with hh without PCT and with comparable hepatic iron stores.

Study subjects were selected from patients enrolled in the hemochromatosis and porphyria research clinics at the Huntsman General Clinical Research Center (GCRC) at the University of Utah. Approval for all studies was obtained from the University of Utah's Institutional Review Board and the Advisory Committee of the GCRC. All experiments were conducted in accordance with the Declaration of Helsinki.

Statistical analyses were performed using SPSS software (Chicago, IL). Independent sample t tests were performed using a 95% confidence interval. Differences in mean values were considered significant for P values less than .05. Correlations were determined using a Spearman 2-tailed correlation test. Correlations were considered significant if P values were less than .05.

All subjects with hh underwent percutaneous needle biopsies of the liver, as did 63 patients with PCT. Liver biopsies were obtained prior to phlebotomy therapy in all study participants. Portions of each liver biopsy specimen were fixed in formalin and stained with Perl stain; hepatic parenchymal cell stainable iron (HPCSI) scores were determined by the method of Scheuer and Williams.²³  Subjects with hh were selected for study if HPCSI scores were grade 3 or 4. Patients with PCT were studied regardless of HPCSI scores. Serum iron, transferrin (Tf) saturation, and serum ferritin values were determined as previously described.²⁴ 

Infection with hepatitis C (HepC) was determined serologically by the presence of anti-HepC IgG antibody as previously described.¹  Alcohol consumption was estimated and graded as previously described:¹  grade 0 indicates nondrinkers; grade 1, consumption of less than 20 g of alcohol per day; grade 2, consumption of between 20 and 70 g of alcohol per day for a minimum of 3 consecutive years; and grade 3, consumption of more than 70 g of alcohol per day for a minimum of 3 consecutive years.

DNA was prepared from peripheral blood leukocytes using an automated extraction protocol (Autopure; QIAGEN, Germantown, MD). DNA extraction and HFE genotyping were performed in the Core Laboratory of the GCRC. Genotyping was done using an allele-specific polymerase chain reaction (PCR) protocol.²⁵  Sequencing of HAMP and HJV was performed by PCR amplification and cycle sequencing using Applied Biosystems BigDye Terminator chemistry (Applied Biosystems, Foster City, CA). The University of California Santa Cruz (UCSC) genome browser²⁶  was used to extract positions of proximal promoters and RefSeq exons for HJV (NM_213653) and HAMP (NM_021175). Exons, proximal promoters, and genomic sequences conserved in mammals were sequenced using PCR amplification and fluorescent dideoxy DNA sequencing, as previously described.²⁷ Survey coordinates from the National Center for Biotechnology Information (NCBI; National Library of Medicine, Bethesda, MD) Build 35 for genomic ranges surveyed with a Phrap quality score of greater than 30 are: HAMP (chromosome 19) 40464702-40465702, 40467355-40468055; HJV (chromosome 1) 144124181-144125232; 144125488-144129119.

Internal primers were used for sequencing and cycle sequencing was carried out with an initial denaturation at 96°C for 30 seconds, followed by 45 cycles of 96°C for 10 seconds, 50°C for 5 seconds, and then 60°C for 4 minutes. Upon completion of sequencing, 20 μL 62.5% EtOH/1M KOAc (pH4.5) was added to each reaction, and the sequencing plates were centrifuged at 3500g at 4°C for 45 minutes. The samples were resuspended in 15 μL formamide and electrophoresed on an ABI 3700 capillary instrument (Applied Biosystems). Sequence trace files were evaluated using the Phred, Phrap, and Consed programs, and potential variants were identified by using the PolyPhred program.²⁸  Single nucleotide polymorphisms (SNPs) were verified by manual evaluation of the individual forward and reverse sequence traces. The amplification and sequencing primers used are presented in Table S1 (available on the Blood website; see the Supplemental Materials link at the top of the online article).

A portion of each liver biopsy sample was snap-frozen at the bedside, stored in liquid nitrogen, and later thawed for RNA preparation. Frozen liver tissues were disrupted in Trizol reagent (Invitrogen, Carlsbad, CA) using a Polytron homogenizer (PowerGene 700; Fisher Scientific, Pittsburgh, PA). Total RNA was isolated from 2 mg of liver according to the Trizol protocol with a yield of 1 to 5 μg RNA. To prepare target samples for microarray hybridization, total RNA samples were amplified and labeled as recommended by the microarray manufacturer (Agilent Technologies, Palo Alto, CA). The quality of amplified RNA was evaluated by capillary electrophoresis using an Agilent 2100 Bioanalyzer.

Microarray format, protocols for probe labeling, and array hybridization are described at http://expression.microslu.washington.edu. The primary data for the expression arrays are available at the same web address. A typical experiment was a comparison between 2 amplified and labeled cRNA samples, a test sample from one of the selected patients with PCT, and a common reference sample pooled from liver biopsies of 10 patients with hh with moderately increased HPCSI (grade 2) but without PCT. Each microarray experiment was done with 4 replicate human 22K oligonucleotide expression arrays (Agilent Technologies) using the dye label reverse technique. Spot quantitation, normalization, and application of a platform-specific error model was performed using Agilent's Feature Extractor software, and all data were then entered into a custom-designed database, Expression Array Manager, and then uploaded into Rosetta Resolver System 4.0.1.0.10 (Rosetta Biosoftware, Kirkland, WA) and Spotfire Decision Suite 7.1.1 (Spotfire, Somerville, MA). Selection of genes for data analysis was based on a greater than 95% probability of being differentially expressed (P ≤ .05) and a fold change of 2 or greater. The resultant false positive discovery rate was estimated in previous work as less than 1% by using semiquantitative reverse transcriptase–PCR.

HPCSI scores, Tf saturations, and serum ferritin concentrations are shown in Table 1. Among patients with PCT for whom DNA was sequenced, 17 (18%) were HFE C282Y homozygotes, 34 (35%) were HFE C282 heterozygotes, and 45 (47%) were wild-type homozygotes. Iron parameters in patients with PCT patients were compared by sex and HFE genotype using an independent sample t test (Table 2). Male and female C282Y homozygotes with PCT had elevated serum iron and Tf saturation compared with wild-type homozygotes. Mean serum ferritin values were higher in C282Y homozygotes with PCT compared with wild-type homozygotes but were not statistically different due to wide variation between individuals. Neither male nor female C282Y heterozygotes with PCT were statistically different than wild-type homozygotes for serum iron and Tf saturation values. Approximately 17% of male and 16% of female patients with PCT had mutations in the UROD gene (F-PCT), but iron parameters did not differ when compared with patients with S-PCT (data not shown). The UROD mutations in the patients with F-PCT have been reported previously.²⁹ 

HJV has been cloned and mapped to chromosome 1q. The gene includes a promoter region, 4 exons, 3 introns, and a 3′ flanking region.^30,31  We found no promoter or exonic polymorphisms in either study population. Noncoding polymorphisms and associated iron values are presented in Table 3. In the hh population, the C→T transition in intron 3 was in linkage disequilibrium with the T→C transition in the 3′ flanking region, as the 8 individuals with the C→T transition in intron 3 also had the 3′ flanking region T→C transition. The C→T, T→C haplotype was associated with higher serum ferritin values (2677 ± 535 μg/L vs 1361 ± 148 μg/L; P < .004 by independent sample t test; Table 3). A total of 9 patients with PCT also had the C→T, T→C haplotype, 3 of whom were HFE C282Y homozygotes. One was a 70-year-old woman with a serum ferritin value of 957 μg/L; the second, a 64-year-old man with a serum ferritin value of 1106 μg/L; and the third, a 5-year-old boy with a serum ferritin value of 91 μg/L (normal value for age, 14-80 μg/L). The mean serum ferritin value in patients with PCT who were HFE wild-type homozygotes or heterozygotes was 540 μg/L. The mean ferritin values were not different in HFE C282Y homozygote patients with PCT with or without the HJV polymorphisms. Collectively, these data suggest the HJV C→T, T→C haplotype is associated with increased penetrance of the HFE C282Y/C282Y genotype.

The HAMP gene is located on chromosome 19q13 and contains a promoter region, 3 exons, 2 introns, and a 3′ flanking region. Polymorphisms detected and iron measures in the subjects sequenced are summarized in Table S2.

A G→A transition in the third exon was found in a single patient with PCT (Table S2). This transition results in a previously identified G71D amino acid change, which affects iron loading when present with mutations in HFE.^17,18  The patient with PCT with this polymorphism was a 54-year-old man whose HFE genotype was wild-type/wild-type (wt/wt). He had undergone phlebotomy therapy elsewhere and was iron depleted at the time of our evaluation. An A→G transition in the third exon was also found in a 41-year-old man with PCT whose HFE genotype was wt/wt (Table S2). This transition results in a K83R change, which does not affect the function of the hepcidin peptide.³²  Iron measures in these 2 patients did not differ significantly when compared with other patients with PCT with the wt/wt HFE genotype (Table S2).

Individual liver samples from 14 patients with PCT were compared by microarray analysis to a pooled reference group from 10 patients with hh without PCT. The biopsy samples in the reference pool were from subjects selected to approximate the serum ferritin values and HPSCI scores of the 14 patients with PCT. None of the subjects in the reference pool had HepC, and none consumed alcohol. The mean values for Tf saturation and serum ferritin in the pooled reference group were 63.2% plus or minus 8.2% and 979.0 μg/L plus or minus 227.0 μg/L, respectively. The HFE genotypes of the 14 patients with PCT are shown in Table 4. Patients 1 through 8 were men and all but one (patient 3) was exposed to HepC, excessive alcohol consumption, or both. Patients 9 through 14 were women. All were taking medicinal estrogens and all had histologic evidence of hepatic steatosis. HAMP expression in patients with PCT was reduced compared with the reference pool of patients with hh (mean = −3.2-fold). Reduced values ranged from −1.2- to −10.3-fold (Table 4). Only 3 of the 14 patients with PCT did not have reduced values (Table 4). Patients 1 (HFE C282Y/wt) and 2 (HFE wt/wt) had the greatest reduction in HAMP expression. The next lowest patient (patient 13) was homozygous for the HFE 282Y mutation. These data indicate that reduced HAMP expression in PCT is independent of the HFE genotype. None of the other examined iron-related genes showed differential expression in the patients with PCT in comparison with patients with hh or in relation to HFE genotype among patients with PCT. A Spearman 2-tailed correlation analysis was performed on the expression data and negative correlations were found between HAMP and TFR1 expression (correlation coefficient [r] = −0.666; P < .009) and serum ferritin and TFR1 (r = −0.577; P < .039). There was also a positive correlation between neogenin and SMAD4 (r = 0.561; P < .037).

Iron values for 18 HepC⁺ patients with PCT were compared with 13 patients with PCT whose HepC serologies were negative. None of the iron values were different between the groups (Table 5). The frequency of homozygosity for the HFE C282Y mutation and alcohol use were similar between the 2 groups (data not shown).

The hepatic siderosis associated with PCT can be explained by homozygosity for the HFE C282Y mutation in a minority of patients, but most often the cause of the iron phenotype is not apparent. Several exonic HJV mutations appear to magnify the iron phenotype in HFE-associated hemochromatosis,^{19,20,33–35} but we detected no exonic mutation in either our patients with hh or our patients with PCT. In the hh group, however, heterozygosity for an HJV allele with an intron 3 C→T transition and a 3′ flank T→C transition was found in approximately 10% of patients, who had significantly higher serum ferritin values than those without this HJV allele (Table 3). Two adult patients with PCT homozygous for the HFE C282Y mutation were found to have the intron 3 C→T, 3′ flank T→C allele. Their serum ferritin values were higher than patients with PCT and hh but without these polymorphisms. The mechanism by which the intron 3 C→T, 3′ flank T→C allele affects penetrance of hh is not clear, as neither of the 2 polymorphisms affects a sequence predicted to alter transcription or message stability.

The G71D mutation in HAMP has been associated with abnormal iron values when combined with HFE mutations,¹⁸  but our single patients with PCT with this mutation had a wild-type HFE genotype. Our single patient with PCT with the K83R mutation also had a wild-type HFE genotype. Both were heterozygous for their respective HAMP mutations. Peptides containing the G71D and K83R substitutions have been synthesized and shown to be active in cell culture systems and to induce hypoferremia when injected into mice.³²  Whether these mutations affect transcription, protein expression, or processing of the preprotein is not known.

Gene expression studies offered some insight into the hepatic siderosis associated with PCT. Of the iron-related genes compared, only HAMP was differentially expressed in patients with PCT compared with a pooled liver RNA sample from patients with hh with similar HPCSI scores and serum ferritin values. HepC and alcohol abuse are the most common risk factors for the development of PCT,¹  and both of these risk factors may be associated with moderate to marked hepatic siderosis. Both HepC and alcohol may affect expression of HAMP. Suppressed hepatic expression of HAMP in patients with HepC was recently reported,³⁶  and serum prohepcidin levels are reduced compared with healthy controls.³⁷  Nishina et al created transgenic mice expressing HCV polyprotein; these animals displayed decreased expression of HAMP and liver iron loading.³⁸  It was proposed that HCV polyprotein increases production of reactive oxygen species and the related oxidative stress reduces HAMP expression. Others have suggested that oxidative stress may lead to a relative lowering of HAMP by attenuating the iron and inflammation-induced up-regulation of HAMP.³⁹ 

Iron is recognized as a potent cofactor in alcohol-induced liver injury,⁴⁰  and several groups have reported alcohol-induced down-regulation of HAMP. Harrison-Findik et al reported that alcohol abolishes iron-mediated regulation of HAMP,⁴¹  and that alcohol-mediated oxidative stress is responsible for down-regulation of HAMP transcription in mice.⁴²  Down-regulation of HAMP by alcohol in a rat model was also reported by Bridle et al.⁴³ 

Oral estrogen use is an independent risk factor for the development of PCT.⁴⁴  Estrogens may have a variety of adverse effects on the liver, including steatosis and steatohepatitis.^45,46  Evidence in humans and in animal models of nonalcoholic fatty liver disease suggests that oxidative cellular damage plays an important role in disease pathogenesis.^47–50 HAMP expression has not been previously studied in women with estrogen-induced hepatic steatosis. A total of 6 of the patients included in Table 4 were such women, and HAMP expression, compared with control subjects with hh, was reduced in most.

Oxidative stress is a finding associated with all of the risk factors for PCT (hh, HepC, alcohol and oral estrogen use). We propose that oxidative stress down-regulates HAMP expression, leading to increased iron export by enterocytes and macrophages. The resulting hepatic siderosis is required to generate the UROD inhibitor responsible for expression of the PCT phenotype.²  Our data suggest that as hepatic porphyrinogens accumulate and are oxidized to porphyrins, down-regulation of HAMP is accentuated. This view is supported by our finding that HAMP expression in patients with PCT is down-regulated beyond that of patients with hh whose HAMP expression is already impaired due to the HFE C282Y mutation.⁵¹  The risk factors for PCT are common, yet PCT occurs with a frequency of 1:5000,¹  suggesting that traits affecting response to oxidative stress influence disease susceptibility.

These studies were supported by National Institutes of Health grants DK062106, DK020503, and MO1 RR00064; National Institute on Drug Abuse grant 1P30DA01562; and Centers of Excellence in Molecular Hematology grant 5P30DK072437.

National Institutes of Health

Contribution: R.S.A., J.D.P., and J.P.K. acquired patients, analyzed the data, and wrote the manuscript; R.B.W. and D.M.D. performed sequencing and data analysis; and M.W.S., S.C.P., and M.G.K. performed microarray analysis.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Richard S. Ajioka, Division of Hematology, University of Utah School of Medicine, 50 North 1900 East, Salt Lake City, UT 84132; e-mail: richard.ajioka@hsc.utah.edu.