Contents lists available at ScienceDirect

Molecular Genetics and Metabolism

journal homepage: www.elsevier.com/locate/ymgme

Regular Article

Digital PCR (dPCR) analysis reveals that the homozygous c.315–48T>Cvariant in the FECH gene might cause erythropoietic protoporphyria (EPP)

Valentina Brancaleonia,⁎, Francesca Granataa, Pasquale Missineob, Silvia Fustinonib,c,Giovanna Graziadeia, Elena Di Pierroa

a Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, U.O. Medicina Generale, Italyb Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, U.O. Tossicologia Ambientale e Industriale, ItalycUniversità degli Studi di Milano, Dipartimento Scienze Cliniche e di Comunità, Italy

A R T I C L E I N F O

Keywords:dPCRc.315-48 T > CFECHEPP

A B S T R A C T

Alterations in the ferrochelatase gene (FECH) are the basis of the phenotypic expressions in erythropoieticprotoporphyria. The phenotype is due to the presence of a mutation in the FECH gene associated in trans to thec.315–48 T > C variant in the intron 3. The latter is able to increase the physiological quota of alternativesplicing events in the intron 3. Other two variants in the FECH gene (c.1–252A > G and c.68–23C > T) havebeen found to be associated to the intron 3 variant in some populations and together, they constitute a haplotype(ACT/GTC), but eventually, their role in the alternative splicing event has never been elucidated.

The absolute number of the aberrantly spliced FECH mRNA molecules and the absolute expression of theFECH gene were evaluated by digital PCR technique in a comprehensive cohort. The number of splicing eventsthat rose in the presence of the c.315–48 T > C variant, both in the heterozygous and homozygous conditionwas reported for the first time. Also, the percentage of the inserted FECH mRNA increased, even doubled in theT/C cases, compared to T/T cases. The constant presence of variants in the promoter and intron 2 did notinfluence or modulate the aberrant splicing. The results of FECH gene expression suggested that the homo-zygosity for the c.315–48 T > C variant could be considered pathological.

Thus, this study identified the homozygotes for the c.315–48 T > C variant as pathological. By extension,when the samples were categorised according to the haplotypes, the GTC haplotype in homozygosis was pa-thological.

1. Introduction

Erythropoietic protoporphyria (EPP, MIM#17700) is a geneticallydetermined form of porphyria that is characterised by severe phototoxicmanifestations. The patients with EPP experience serious pain, er-ythaema, and swelling of the exposed skin upon exposure to the sun[2]. In almost 96% of the cases, the clinical manifestations of EPP aredue to a decreased activity of ferrochelatase (FECH), the enzyme in-volved in the last reaction of haeme biosynthesis [1]. FECH inserts theferrous iron into the protoporphyrin IX (PPIX) ring, giving rise to pro-tohaeme. As a result of a deficiency of the FECH enzyme, PPIX accu-mulates in the erythrocytes, vascular endothelium, and faeces. Being ahighly reactive molecule, PPIX is activated by UV-A of the visible lightand generates free radicals, leading to the cellular and tissue injuriesand severe pain, characteristic of EPP.

Intron 3 of the FECH gene pre-mRNA is subjected to alternativesplicing, thereby leading to the insertion of 63 intronic base pairs and

the nonsense-mediated decay of this aberrant mRNA molecule.Depending on the genotype of the c.315–48 T > C variant(rs2272783), the percentage of production of this aberrant mRNAvaries. In the presence of the wild-type T base, there is a 10–20%physiological production of the inserted FECH mRNA, whereas in thepresence of the C base – the cause of the so-called ‘hypomorphic FECHallele’ – this percentage increases up to 40% [11]. With the only ex-ception of the subjects with homozygous mutations in the FECH gene, anull mutation in the FECH gene combined in trans to the C base causesthe clinical expression of EPP in most of cases. On the other hand, thepresence of the T base at the position c.315–48 in trans to a mutationusually leads to a symptom-free condition, identifying the carrier sub-jects of EPP.

Thus, most of the studies about the molecular characterisation of theEPP patients report the presence of the c.315–48C polymorphism intrans to a mutant allele of the FECH gene. Nevertheless, different studieshave also reported the cases of EPP, harbouring only the c.315–418C

https://doi.org/10.1016/j.ymgme.2018.06.005Received 6 April 2018; Received in revised form 12 June 2018; Accepted 12 June 2018

⁎ Corresponding author at: Fond. IRCCS Ca' Granda Ospedale Maggiore Policlinico, Via F. Sforza 35, 20122 Milan, Italy.E-mail address: valentina.brancaleoni@policlinico.mi.it (V. Brancaleoni).

Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

1096-7192/ © 2018 Elsevier Inc. All rights reserved.

Please cite this article as: Brancaleoni, V., Molecular Genetics and Metabolism (2018), https://doi.org/10.1016/j.ymgme.2018.06.005

variant in homozygosis [15, 17–19]. One of the studies has reportedthat patients show only a mild EPP phenotype with slightly high ery-throcyte protoporphyrin and mild photosensitivity [18]. Similarly, arecent Japanese study has described three cases of the childhood onsetof mild EPP due to the presence of the homozygous C > C variant inintron 3 [15]. Considering these reports, it is not known if the differ-ences in the EPP phenotypes come from a differential expression of theFECH gene.

Moreover, there is evidence that other Single NucleotidePolymorphisms (SNPs) could also be involved in the phenotypic ex-pression or at least be constantly associated to the polymorphism inintron 3. In fact, the c.68–23C > T variant located in intron 1, (SNP,rs2269219) has been studied and it was demonstrated that it is able toalter the secondary structure of the mRNA [16]. In addition, thec.1–252A > G variant in the promoter region, (SNP, rs17063905) hasbeen associated with a reduced expression of the FECH gene [9]. Twomolecular studies in the Argentine population have reported the con-stant allelic association of these two SNPs with the c.315–48C variant.In particular, the EPP patients always show the c.1–252G, c.23–68 T,and c.315–48C alleles (GTC haplotype) in trans to the mutant allele [6].Previously, one independent study in two Italian families and anotherstudy, involving a single case of EPP in the English Scottish ancestryreported exactly the same association of the GTC haplotype in trans to amutant allele for the expression of clinical EPP [8, 13].

At present, it is still unclear whether these two polymorphisms in-fluence the splicing pattern of intron 3. At the same time, the previouslypublished studies do not give an absolute quantification of the aberrantFECH mRNA molecule but give only its percentage, relative to the totalFECH mRNA.

Recently, a new, highly sensitive, and precise technique called di-gital PCR (dPCR) has been made available as an alternative to theclassical quantitative PCR (qPCR) technique for the absolute quantifi-cation of nucleic acids. In this study, the technique is applied to eval-uate the alternative splicing events of the FECH gene and the absoluteexpression profile of the FECH gene in a cohort, composed of the EPPpatients, the carriers of the FECH gene mutation, and the normal sub-jects. The roles of the c.1–252A > G, c.68–23C > T, andc.315–48 T > C variants, when present in the homozygous conditionhas also been investigated. At the end, the correlations among the ex-pression of the FECH gene, the percentage of the inserted FECH mRNA,and the altered biochemical parameters are also evaluated in the EPPpatients.

2. Materials and methods

2.1. Study subjects

In this study, a total of 90 subjects, recruited at the Centre of RareMetabolic Diseases at Fondazione IRCCS Cà Granda from 2014 to 2017were investigated. The subjects consisted of 32 EPP patients, 12 carriersof the FECH gene mutation, 17 carriers of the c.315–48C > T variant(polymorphism), and 29 normal controls. All the EPP patients showed aclassical history of skin photosensitivity and high values of free proto-porphyrins in the erythrocytes. The carriers of the mutation and of thec.315–48C > T polymorphism were the relatives of the patients as andwhen they were available. The EPP patients as well as the carriers of themutation and of the c.315–48C > T polymorphism were already di-agnosed genetically and clinically.

Freshly drawn blood samples were collected in the EDTA tubes andimmediately processed for the molecular determinations. When pos-sible, the levels of the erythrocyte protoporphyrins and the faecalprotoporphyrins were determined.

Written informed consent was obtained from all the subjects inaccordance with the ethical guidelines of Fondazione IRCCS Cà GrandaOspedale Maggiore Policlinico and the Declaration of Helsinki by theWorld Medical Association (WMA).

2.2. The determination of the erythrocyte protoporphyrins and the totalfaecal protoporphyrins

Whenever possible, the concentrations of the erythrocyte proto-porphyrins and the faecal protoporphyrins were determined as a rou-tine analysis. The erythrocyte protoporphyrins were determined aspreviously reported [5]. The total faecal protoporphyrins were mea-sured according to a previously published method [14]. Briefly, a faecalsample of about 50mg was weighed and suspended in 1ml of 37% HCl.Following vigorous mixing, 3 ml of ethyl ether was added and mixedand then 3ml of water was added and mixed. The mixture was thencentrifuged and the upper phase was discarded, while the lower aqu-eous phase was submitted to a Lambda 25 UV–Vis spectrophotometerfrom 390 to 425 nm (Perkin-Elmer, Monza, Italy) for the quantificationof the total faecal protoporphyrins. The total faecal protoporphyrinswere expressed as nmol/g of the faecal dry weight, following the as-sessment of the content of humidity in the faeces by a Sartorius MA35moisture analyser balance (Sartorius, Muggiò MB, Italy). The referencerange of the total faecal protoporphyrins was up to 200 nmol/g of thefaecal dry weight and the limit of quantification of the assay was30 nmol/g of the faecal dry weight.

2.3. DNA isolation and genotyping

DNA was isolated from the peripheral blood samples with a Maxwell16 automated extractor (Promega Corporation, Madison, WI, USA)according to the previously published methods [10].

For the control subjects, the c.1–252A > G, c.68–23C > T, andc.315–48 T > C polymorphisms of the FECH gene were analysed. Anamount of 100 ng of DNA was amplified with the BIOTAQ DNA poly-merase (Bioline, London, UK), using the following PCR primer pairs: 5’-AGGGCACGCAACTAGGAGT-3′ and 5’-TCAGGGATCCTGGCCCTG-3′ forthe c.1–252A > G polymorphism, 5’-TGCCTGCAGAGAAATGCTAG-3′and 5′-(GC39) TCTTATTTGTACCTGATGTT-3′ for the c.68–23C > Tpolymorphism, and 5’-GTGTTGTGTGTCCTGAATCT-3′ and 5-‘CGAAAGAACTAATCTAGTTACATGT-3′ for the c.315–48C > T polymorphism,respectively. The PCR products were then subjected to automated directsequencing on an ABI Prism 310 Genetic Analyser (Thermo FisherCorporation Inc., San Francisco, CA, USA).

In order exclude a possible allele drop-off, DNA of homozygouspatients for the c.315-48 T > C variant was also amplified with asecond reverse primer, located in exon 4 (5′- TCCAGCAGCTTCACCATGCCCT- 3′), and then subjected to direct sequencing.

2.4. RNA isolation, reverse transcription PCR(RT-PCR), mRNA sequencingand digital PCR(dPCR)

For qualitative mRNA analysis purpose, total RNA was retro tran-scribed as previously described [8]. Two primer pairs allowed se-quencing FECH mRNA in two different overlapping fragments. The firstfragment covered from exon 1 to exon 4 and the second one from exon2 to exon 11. Thus, 100 ng of cDNA were amplified with the followingprimers pairs: IP-FW 5’-GGGGATCGCTACCCGGCTC-3′ with IP-RW-5’-TCCAGCAGCTTCACCATGCCCT-3′ and IIP-FW-5’-TCCAGCAGCTGGAGGGTC-3′ with IIP-RW-5’-ACACCCTCTCCACATCGGAGGTATC-3′. Thethermal cycling conditions were 94 °C for 5min, followed by 35 cycles:92 °C for 30 s, 60 °C for 30 s, 72 °C for 1min and a final elongation stepof 3min at 72 °C. The two fragments were then subjected to direct se-quencing in both orientations.

For digital PCR purpose, total RNA was isolated from freshly drawnbuffy coat samples with a Maxwell 16 automated extractor (PromegaCorporation, Madison, WI, USA) according to the previously publishedmethods [10].

An amount of 100 ng of total RNA was subjected to ReverseTranscription PCR (RT-PCR), using the ViLo Master Mix (Thermo FisherCorporation Inc., San Francisco, CA, USA).

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

2

For the absolute quantification of the normally spliced and the al-ternatively spliced mRNAs, dPCR was performed on a Quantstudio®3DDigital PCR System (Thermo Fisher Corporation Inc., San Francisco, CAUSA), using specific and differentially labelled TaqMan® probe mixes,in order to distinguish the two forms. For the quantification of the FECHgene, the TaqMan®MGB probe mix with FAM dye label, having the bestcoverage and localised in the boundary between the exons 3 and 4(Hs01555261_m1) was used.

The quantification of the aberrantly spliced mRNA was carried outusing a custom TaqMan®MGB probe mix with VIC dye label (the se-quences are available upon request).

An amount of 2.5 ng of the total RNA was loaded onto the 3D DigitalPCR Chip, using the QuantStudio®3D Digital PCR Chip Loader in a re-action mixture, consisting of the 2× Quantstudio®3D Digital PCRMaster Mix and 1× of each of the TaqMan® probes. The chips weresealed and loaded onto the ProFlex™ 2xFlat PCR System (Thermo FisherCorporation Inc., San Francisco, CA, USA) and cycled according to thefollowing parameters: initial denaturation at 96 °C for 10min, followedby 50 cycles of amplification as follows: denaturation at 98 °C for 30 s,annealing at 60 °C for 2min, and a final extension at 60 °C for 2min.After cycling, the end-point fluorescence of the partitions on the chipswas measured by transferring the chips to the QuantStudio®3D chipreader; secondary analysis was performed with QuantStudio®3DAnalysis Suite cloud software (application version 3.1.2; algorithmversion 4.4.10).

2.5. Statistical analysis

Unless otherwise stated, all statistical analyses were performedusing the GraphPad Prism software (version 7.0). The normality testswere assessed for each of the measurements. The data were expressedas logarithms to base 10. Using the one-way ordinary ANOVA, followedby the Tukey-Dunnett's post-hoc tests, the statistical analyses wereperformed. The nominal statistical significance was set at p < .05. Thecorrelations were assessed using the Pearson's correlation coefficient.

3. Results

3.1. The digital PCR (dPCR) assay

In this study, a dPCR assay was developed to evaluate the expressionof the FECH gene in vitro by means of an absolute quantification, bothfor the normally spliced and the aberrantly spliced species of the FECHmRNA. The assay was designed to work in duplex, introducing a dif-ferential label on the two TaqMan probes, used for the detection of thetwo species of the FECH mRNA.

For the experimental setup, different quantities of the FECH mRNA(25 ng, 5 ng, 2.5 ng, and 0.25 ng) were tested for defining the bestamount that could ensure a correctly loaded chip, the precision valuesin the desired range, and a homogenous fluorescence.

The quantity of 2.5 ng of the FECH mRNA gave better results forboth the assays: a correct loading of the chip with a homogenousfluorescence and sample distribution, and an optimal cluster divisionwith precision levels below the recommended threshold of 10%(Table 1). In order to get the precise quantification, a threshold of14,000 data points was set for the quality control of the chip (Suppl.Table 1 reports the copies/μl and CI for each target, the number ofdatapoints above the threshold and the number of filled partitions foreach subject in the study).

3.2. The alternative splicing events depend on the c.315–48C genotype

The dPCR assay was applied to a cohort of 90 subjects. Among theEPP patients, five were homozygous for the c.315–48C variant andlacked any mutation both in exons and exons-introns boundaries of theFECH gene (Table 2- Suppl. Table 1). The reverse primer for c.315–48

variant genotyping was designed on two low frequencies SNPs. Thepatients homozygous for the c.315–48C variant were evaluated alsowith another reverse primer, to exclude a possible allele drop-out effect.The homozygosity was confirmed also with the new PCR protocol. Thepresence of the mutations in the ALAS2 gene or long deletions in theFECH gene were excluded by Multiplex ligation probe-dependent am-plification analysis (MLPA), as previously described [7]. Also, FECHmRNA sequencing did not reveal any underling sequence in these pa-tients.

Firstly, this study cohort was classified according to the status of thec.315–48 T > C variant and for the presence of a mutation in the FECHgene without applying any filter with respect to the type of mutation.

The presence of the T/C genotype was able to double the number ofthe single events of aberrant splicing. The wild-type class presented amean number of aberrant splicing of 31 copies/μl, compared to about51 copies/μl of the T/C carriers and 65 copies/μl of the T/C genotypeplus mutation (Table 3). The number of the single events of aberrantsplicing showed a dependency on the genotype (Fig. 1.a). Differenceswere observed when the T/T genotype plus mutation was comparedwith the T/C genotype plus mutation. Notably, both the classes of theEPP patients, having the T/C genotype plus mutation and the C/Cgenotype presented the same mean number of the single events ofaberrant splicing. No differences were observed between the T/T gen-otype and the T/T genotype plus mutation. The results, expressed as thepercentage of the abnormal FECH mRNA, relative to the total FECHmRNA confirmed this tendency.

In the control group with the T/T genotype, the percentage of theinserted FECH mRNA was 2.23% of the total FECH mRNA, compared to4.4877% of the T/C genotype and 9.087% in the EPP patients (Table 4).Differences between the means were observed among the T/T, T/C, C/C, and T/C genotypes as well as a mutation in the FECH gene, indicatingthat the presence of the polymorphism was sufficient to increase thepercentage of the inserted FECH mRNA (Fig. 1B). The presence of amutation was not able to influence the percentage of the aberrantlyspliced mRNA (Fig. 1C-D). However, in conjunction with a mutation,this value raised from 2.9% to 9.087%, because the global FECH ex-pression is diminished. No differences are observed comparing per-centage of T/C plus mutation and C/C patients genotype.

3.3. The c.1–252A>G and c.68–23C>T variants do not affect thenumber or percentage of the aberrant splicing events

In order to evaluate the possible influence of the c.1–252A > G andc.68–23C > T variants on the aberrant splicing events of the FECHmRNA, probands were clustered according to the haplotypes of thesethree polymorphisms. The phases of the haplotypes were calculated byanalysing the family segregation of our EPP cohort families, which in-cluded 14 family control groups, and 256 normal controls, which wereanalysed previously (data not shown). The majority our EPP patientswere heterozygous for the GTC haplotype with a mutation in the FECHgene in trans or presented only the homozygous GTC haplotype withoutany other mutation on the FECH gene or the ALAS2 gene (Table 2).

The absolute quantity of the inserted FECH mRNA was analysed in

Table 1Precision values of the two targets at different mRNA concentration. The de-sired precision should be below 10%.

mRNAconcentration

FECH (FAM)Precision

FECH_Ins(VIC)

Precision

Datapointsqualified by

Quality Treshold(0,5)

n° of filled

25 ng 4.732% 2.427% 15,994 18,5605 ng 2.142% 5.833% 16,188 18,2552.5 ng 2.30% 6.11% 17,120 17,9720.25 ng 1.998% 4.858% 17,155 18,377

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

3

the patients, categorised according to the haplotypes and both the GTCand ATC haplotypes were able to raise the numbers of the splicingevents, as expected by the presence of the C base in the haplotype(Table 5).

The data obtained from the analysis, classified on the basis of thec.315–48 T > C variant were confirmed when the percentage of thehaplotypes with the aberrantly spliced mRNA was analysed. A differ-ence was noted between the wild-type ACT/ACT haplotype to theheterozygotes for the GTC haplotype and the patients (Fig. 2A). At thesame time, a difference was found between the carriers of the FECHgene mutation and the patients (Fig. 2B). The ACT/GTT haplotypebehaved similar to the ACT/ACT haplotype, showing a difference be-tween the heterozygotes for the GTC haplotype and the patients(Fig. 2C-D). The GTT haplotype did not show any influence on thenumber or the percentage of the aberrantly spliced mRNA.

The two haplotypes, composed by ACT and GTT, seemed to behavesimilarly with respect to the inserted FECH mRNA. Thus, the presence

of the G base at the position c.1–252 and the T base at the positionc.68–23 was not sufficient to influence the percentage of the insertedFECH mRNA.

3.4. The presence of the homozygous c.315–48 T/C variant or the GTChaplotype reduces the expression of the FECH gene to a pathological level

The absolute expression of the FECH gene in the EPP patients andcontrols was then evaluated by analysing the total number of the FECHgene transcripts and comparing it to the wild-type controls as well asthe carriers of the mutation or of the c.315–48 T/C variant.

We noticed that the expression of the FECH gene was highly variableamong the wild-type controls.

The expression of the FECH gene was found to be very similar be-tween the classical EPP patients with the C/T genotype plus mutationand the subjects with homozygosity at the C/C genotype for thec.315–48 T/C variant. Both the groups of subjects had a mean residual

Table 2Genetic and biochemical data of patients included in the study. Erythrocyte porphyrins n.v. < 3.00 μg/gHb; Fecal porphyrins n.v. < 200 nmol/g dry faeces.*Denotes new FECH gene mutations.

Patient FECH genotype FECH alleles Erytrocytes porphyrins Faecal porphyrins

c.1–252 A > G c.68-23C > T c.315-48 T > C

1. c.[=];[=] G > G T > T C > C 100.1 n.a.2. c.[=];[=] G > G T > T C > C 15 34.003. c.[=];[=] G > G T > T C > C 32.2 943.004. c.[=];[=] G > G T > T C > C 2.3 n.a.5. c.[=];[=] G > G T > T C > C 33 93.006. c.[1-251G > C;194+4350_463+1197del];[=] A > G C > T T > C 124 654.007. c.[1-251G > C;194+4350_463+1197del];[=] A > G C > T T > C 77.6 684.008. c.[67+5G > A];[=] A > A C > T T > C 26.6 397.009. c.[67+5G > A];[=] A > A C > T T > C 11.9 94.0010. c.[163G > T];[=] A > G C > T T > C 214.7 567.0011. c.[215dupT];[=] A > G C > T T > C n.a. 37.2012. c.[215dupT];[=] A > G C > T T > C 41 114.0013. c.[215dupT];[=] A > G C > T T > C 94.5 n.a.14. c.[215dupT];[=] A > G C > T T > C 70.7 257.0015. c.[215dupT];[=] A > G C > T T > C 90.9 679.0016. c.[215dupT];[=] A > G C > T T > C 38 444.0017. c.[215dupT];[=] A > G C > T T > C 129.7 n.a.18. c.[215dupT];[=] A > G C > T T > C 122.8 n.a.19. c.[215dupT];[=] A > G C > T T > C n.a. 63.2020. c.[215dupT];[=] A > G C > T T > C 122 n.a.21. c.[215dupT];[=] A > G C > T T > C 67.7 n.a.22. c.[343C > T];[=] G > G C > T T > C 10.5 147.0023 c.[544delC];[=]* A > G C > T T > C 69.5 581.0024. c.[706-3C > G];[=]* A > G C > T T > C 48.5 90.0025. c.[706-3C > G];[=]* A > G C > T T > C 43.4 35.0026. c.[757_761delAGAAG];[=] G > G C > T T > C 22.6 138.0027. c.[757_761delAGAAG];[=] G > G C > T T > C 20.7 399.0028. c.[801G > A];[=] G > G C > T T > C 1.7 24.0029. c.[899delTG];[=] A > G C > T T > C 122 56.0030. c.[942_945delAACA];[=]* A > G C > T T > C 165.3 640.0031. c.[1080_1081delTG];[=]* A > G C > T T > C 51.4 n.a.32. c.[1232G > A];[=] A > G C > T T > C 160.4 n.a.

Table 3Copies/μl of FECH aberrant inserted mRNA.

c.315–48 T/T c.315–48 T/T+MUT c.315–48 T/C c.315–48C/C c.315–48 T/C+MUT

Number of subjects 29 12 17 5 27

Minimum 12.61 14.89 24.44 47.21 20.4825% Percentile 20.74 19.15 33.86 52.13 36.72Median 29.56 37.23 48.3 61.83 67.4875% Percentile 41.06 50.15 65.38 72.68 84.17Maximum 76.81 58.09 105.4 72.84 164.1

Mean (n copies/μl) 31.36 35.97 51.51 62.29 65.61Std. Deviation 14.89 15.16 22.13 10.85 32.33Std. Error of Mean 2.766 4.377 5.367 4.852 6.221

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

4

expression of the FECH gene at 42–49% and presented a differencebetween the means of total expression of the FECH gene, compared tothe wild-type controls with the T/T genotype (Fig. 3A-B). The abovedata identified the two classes of the T/C genotype plus mutation andthe C/C genotype and was also responsible for a pathological decreasein the expression of the FECH gene.

After the categorisation of the samples according to the haplotypes,the expression of the FECH gene was analysed, as previously explained.Again, the homozygotes for the GTC haplotype showed a residual ac-tivity that was very similar to that of the classical EPP patients. All ofthem showed a mean residual expression of the FECH gene between

37.6 and 44% and a difference between the means of total expression ofthe FECH gene, compared to the wild-type haplotype ACT/ACT(Fig. 3B-C).

When the other possible haplotypes of our cohort were taken intoaccount, it was noticed that the expression of the FECH gene was dif-ferent in the GTC haplotype plus mutation, compared to the ACT/GTThaplotype (Fig. 3D). At the same time, no difference was found betweenthe ACT/ACT haplotype and the ACT/GTT haplotype or the GTT/GTThaplotype. Also in these cases, the presence of the G base at the positionc.1–252 and the T base at the position c.68–23 did not seem to influ-ence the total expression of the FECH gene.

Fig. 1. Percentage of aberrant inserted FECH mRNA. The categorisation is according to c.315–48T>C variant. ANOVA global p-value< .0001.

Table 4Percentage of aberrant inserted FECH mRNA. Samples are categorised according to the c.315-48T>C polymorphism and the presence in trans of a FECH mutation.

c.315-48 T > T c.315-48 T > T+Mut c.315-48 T > C c.315-48C > C c.315-48 T > C+Mut

Number of subjects 29 12 17 5 27

Minimum 1.27 0.88 2.26 8.167 4.31325% Percentile 1.635 2.178 3.406 8.941 7.209Median 2.18 3.034 4.29 10.33 9.1675% Percentile 2.5 3.692 5.765 13.41 11.28Maximum 4.89 4.4 7.21 14.42 15.37

Mean (%) 2.23 2.9 4.477 11.01 9.087Std. Deviation 0.8276 1.055 1.458 2.438 2.545Std. Error of Mean 0.1537 0.3046 0.3535 1.09 0.4897

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

5

3.5. The correlations between the inserted mRNA and the biochemicalparameters

Since the erythrocyte protoporphyrins and the faecal proto-porphyrins are the main biochemical parameters altered in the EPPpatients, their values were correlated to the total expression of theFECH gene and the percentage of the inserted FECH mRNA.

The correlations between the total expression of the FECH gene andthe percentage of the inserted FECH mRNA were significant, comparedto the erythrocyte protoporphyrins (Fig. 4A). In particular, the ex-pression of the FECH gene was negatively correlated to the total ery-throcyte protoporphyrins as a lower expression of the FECH gene

tended to have higher protoporphyrin values (Pearson's r=−0.4889;p= .0002; Table 6). On the contrary, the higher was the percentage ofthe inserted FECH mRNA, the higher was the accumulation of theprotoporphyrins, showing that this variable tended to increase alto-gether (Pearson's r=0.707; p > .0001; Table 6).

A similar trend was observed when these parameters were corre-lated to the faecal protoporphyrins (Fig. 4B). The expression of theFECH gene tended to be negatively correlated to the faecal proto-porphyrins (Pearson's r=−0.657; p= .0003) while the percentage ofinserted mRNA tend to increase at higher values of faecal porphyrins(Pearson's r=0.6; p= .0004). On the other hand, no correlation of thebiochemical parameters was noticed with the absolute number of the

Table 5Number of copies/μl of FECH aberrant inserted mRNA. Samples are categorised according to the phase haplotype and the presence of FECH mutation.

ACT/ACT ACT/MUT ACT/GTC ACT/ATC GTC/MUT ATC/MUT GTC/GTC ACT/GTT GTT/GTT

Number of subjects 16 10 16 1 25 2 5 9 2

Minimum 12.61 14.89 24.44 46.59 20.48 84.17 47.21 14.33 18.3725% Percentile 17.43 17.43 33.19 46.59 36.27 84.17 52.13 22.22 18.37Median 31.53 37.23 49.14 46.59 65.53 97.2 61.83 31.31 22.5575% Percentile 46.91 46.27 70.19 46.59 78.71 110.2 72.68 38.76 26.74Maximum 76.81 58.09 105.4 46.59 164.1 110.2 72.84 41.51 26.74

Mean (n copies/μl) 33.87 34.56 51.82 46.59 63.08 97.2 62.29 30.44 22.55Std. Deviation 18.46 15.64 22.82 32.06 18.43 10.85 9.265 5.92Std. Error of Mean 4.615 4.944 5.704 6.413 13.03 4.852 3.088 4.186

Fig. 2. Percentage of aberrant inserted FECH mRNA. The categorisation is according to the haplotypes formed by c.1–252A>G, c.68–23C>T and c.315–48T>Cvariants. ANOVA Global p-value< .0001.

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

6

alternative splicing events (Table 6).

4. Discussion

In this study, we applied a dPCR approach for the first time, in orderto study the expression of the FECH gene. In addition to its high levelsof sensitivity and precision, one of the most intriguing features of thedPCR approach lies in not requiring the presence of a standard re-ference [12] for the absolute quantification. We designed and set up adPCR protocol that was able to detect and, at the same time, quantify

the expression of the FECH gene and the events of alternative splicing towhich, intron 3 of the FECH gene is prone. In this system, we obtainedrobust and reproducible results, using 2.5 ng of the total FECH mRNA,which is a very low quantity that in qPCR experimental system did notpermit to obtain reliable results.

Non-sense mediated decay (NMD) was previously described as thecause of the low expression allele by Gouya et al. The c.315-48 T/Cvariant introduce a cryptic splice site and with the creation of a longermRNA, with insertion of a premature stop codon. The aberrant mRNA isphysiologically subjected to NMD, which causes a selective destroy of

Fig. 3. Absolute FECH gene expression, according to c.315–48T>C variant categorisation (A-B) and haplotypes (C-D). ANOVA global p-value< .0001.

Fig. 4. Correlations between erythrocyte (A) and faecal porphyrin (B) with absolute FECH expression, percentage and n° copies/μl of aberrant inserted FECH mRNA.EPP Patients, FECH mutation carriers and GTC carriers were included in the analysis, when samples were available.

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

7

the longer mRNA. Working with patients mRNA, probably NMD me-chanism lowers the measurable aberrant mRNA molecules. This me-chanism should be constant in all analysed subjects, thus it should notinterfere with the global measurements. In the past, with our real timePCR protocols we never had the possibility to find the aberrant mRNA.We choose a digital PCR approach, to have the chance to measure allavailable aberrant mRNA molecules.

The presence of the c.315–48 T > C variant in the FECH gene hasbeen reported to promote the use of a physiological site for the alter-native splicing events. In this study, we analysed the absolute numberof the aberrant splicing events in a comprehensive cohort for the firsttime. In the present study, it was reported that the number of theaberrant splicing events increases in the presence of thec.314–48 T > C variant, both in the heterozygous and homozygouscondition. The percentage of the inserted FECH mRNA also increased,even doubled in the T/C cases, compared to the T/T cases (2.23% vs.4.477%), which was lower compared to the previously reported studies[3, 11]. We also reported for the first time the percentage of the in-serted FECH mRNA in a large number of the EPP patients and foundthat in the pathological class, it rose up to about 10% (Fig. 1D) becausethe trans allele in the patients was hit by a mutation and thus the globalexpression of the FECH gene was lowered. The percentage of the ab-normal splicing events was calculated above the total events of theFECH gene, considering both the aberrantly spliced and the normallyspliced isoforms. It is natural that if the normally spliced isoform islowered by the presence of a mutation in the FECH gene, then the re-lative percentage of the aberrantly spliced isoform is increased. In fact,taking into account only the absolute number of the splicing events, wesaw that the only genetic element, influencing the count of the aberrantFECH mRNA was the presence of the C base at the position c.315–48.We confirmed the correlation of the genotype with the increase in thepercentage of the inserted FECH mRNA, as previously reported, al-though with different quantitative results [3, 11]. The above finding isprobably due to the fact that the dPCR method is more sensitive thanthe qPCR or semi quantitative methods, used in other studies and thus,it can lead to an overestimation in the cases of low sample input and/ortechniques that require the quantification of bands.

In this study cohort, most of the patients presented a mutation in theFECH gene with the c.315–48 T > C variant in trans, thus presentingthe genetics of classical EPP. We analysed the expression of the FECHgene in the patients, identifying levels that could be considered

pathological, compared to the control subjects. Although it is knownthat EPP is correlated to a lowered activity of the FECH enzyme, this isthe first study that reports the absolute levels of expression of the FECHgene. In the cases of the patients with classical EPP, a reductionof> 50% in the expression levels is necessary to have the phenotypicfeatures of the disease. Interestingly among our subjects, five did notpresent the genetics of classical EPP. Four out of the five cases pre-sented with the classical phenotypic and biochemical features of EPPbut they harboured only the c.315–48 T > C variant in homozygosiswithout any other mutations (point mutations or deletions) in the FECHgene (Table 2). All of them were the cases of classical EPP with theexception of only one and included both males and females (2 malesand 3 females) without any prevalence linked to their gender. Firstly,the analysis of the percentage of the inserted FECH mRNA in the case ofnon classical EPP showed levels that were comparable to those of theclassical form (9.087% vs. 11.01%). More importantly, the expressionof the FECH gene was completely comparable to the classical EPP pa-tients. Similar to the previous studies, the EPP cases in this study tooshowed the homozygous variant in intron 3 of the FECH gene as a causeof the phenotype (Table 7).

Taken together, all these cases showed variable phenotypic ex-pressions, ranging from not having any symptoms to the classical EPPphenotype. In some cases, the symptoms were categorised as only mild,although in the case of the Japanese subjects especially, they hadchildhood onset and their relatives showed a history of photosensitivity[15]. The results of the expression of the FECH gene and the previouslyreported cases suggest that the homozygosity for the c.315–48 T > Cvariant can be considered pathological to some extent. This in turns,could also explain why in some regions (such as South America and EastAsia, according to gnomAD), the frequency of homozygotes for c.315-48C variant is higher than other regions but without an apparent in-crease of EPP cases. The literature is still lacking comprehensive dataabout EPP phenotype from those regions, so it is not easy to give anexact evaluation of EPP prevalence. Thus it is not easy no correctlyaddress the real prevalence. We can suggest, that a variable phenotypecould be explained by the presence of a modifier gene locus that can actdirectly or indirectly, thereby influencing the expression of the FECHgene.

On analysing the expression of the FECH gene in all the categories, itwas noticed that the heterozygotes for the c.315–48 T > C variantshowed expression values that were similar to the mutation carriers.

Table 6Correlation table.

Pearson's r Erythrocyte Porphyrinsvs

FECH copies/μl Tot

Erythrocyte Porphyrinsvs

% FECH_ins

Erythrocyte Porphyrinsvs

FECH_ins copies/μl

R −0.4889 0.707 0.23795% confidence interval −0.6722 to −0.2493 0.5378 to 0.8214 −0.03841 to 0.4789R squared 0.2391 0.4998 0.05616

P valueP (two-tailed) 0.0002 <0.0001 0.0907P value summary *** **** nsSignificant? (alpha=0.05) Yes Yes NoNumber of XY Pairs 52 52 52

Pearson's r Faecal Porphyrinsvs

FECH copies/μl Tot

Faecal Porphyrinsvs

% FECH_ins

Faecal Porphyrinsvs

FECH_ins copies/μl

R −0.657 0.6394 0.0287495% confidence interval −0.8325 to −0.3617 0.335 to 0.8229 −0.3627 to 0.4115R squared 0.4316 0.4088 0.0008259

P valueP (two-tailed) 0.0003 0.0004 0.8892P value summary *** *** nsSignificant? (alpha=0.05) Yes Yes NoNumber of XY Pairs 26 26 26

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

8

Thus, this variant is able to reduce the global expression of the FECHgene as well as a classical mutation in the FECH gene to some extent.We can hypothesise that the inserted FECHmRNA can act in a dominantnegative manner towards the normal isoform, thereby reducing itsabsolute quantity, which in turn can lead to a diminished availability ofthe normal isoform for the posttranscriptional steps. Thus, this variantmay be able to promote the use of the alternative splicing events and atthe same time influence the expression of the FECH gene or more likely,the aberrant FECH mRNA can influence the total FECH gene transcriptby another molecular mechanism that needs to be elucidated.

Our study cohort not only showed the presence of thec.315–48 T > C variant in trans to the mutant allele of the FECH genebut two other variants (c.1–252A > G and c.68–23C > T) were alsofound to be constantly associated to the c.315–48 T > C variant. Atotal of four studies have already reported the same constant allelicassociation in trans to a mutant allele of the FECH gene—two molecularstudies in the Argentine population, a study of two Italian families, anda single case of EPP in the English Scottish ancestry [8, 13]. Previously,both these variants were functionally evaluated in vitro. Thec.1–252A > G variant was responsible for a decrease in the activity ofthe FECH gene promoter and the c.68–23C > T variant was able toalter the pre-mRNA structure [8, 16]. After the categorisation of ourcohort according to the haplotypes, all the patients presented the GTChaplotype in trans to the mutation and four of them were homozygousfor the GTC haplotype. Any additional difference was associated withthese variants when the aberrant splicing events were evaluated ac-cording to this categorisation. Thus, the constant presence of the var-iants in the promoter and intron 2 is not able to influence or modulatethe aberrant splicing events.

In our cohort, the homozygous GTC cases with the signs of classicalEPP were comparable to the classical EPP cases (GTC/Mut), when boththe percentage and/or the absolute number of splicing events wereconsidered. Both these categories showed a residual expression of theFECH gene that was lower than 50%, also identifying the GTC/GTCclass as pathological.

The absolute expression of the FECH gene also too analysed ac-cording to the haplotypes, in order to evaluate the role of thec.1–252A > G and c.68–23C > T variants in the expression of theFECH gene in vitro.

The presence of these variants in the GTT/ACT haplotype or theGTT/GTT haplotype did not show a reduction in the expression of theFECH gene. Therefore, it is conceivable that possibly a true significancecan be noticed by having a larger sample size. Thus, only the presenceof a mutation in trans to the GTC haplotype is able to significantly di-minish the total expression of the FECH gene; and the GTC haplotype inhomozygosis has the levels of expression of the FECH gene that are

comparable to the classical EPP genotype (GTC/Mut).Subsequently, we compared the altered biochemical parameters of

the EPP patients with the expression of the FECH gene and noticed thatthe erythrocyte protoporphyrins and the faecal protoporphyrins wereinversely correlated to the total expression of the FECH gene. The lowerexpression of the FECH gene was concomitant with a higher accumu-lation and excretion of these metabolites. Considering the aberrantsplicing events, it was interesting to note that no correlation was ob-served between the number of the FECH inserted mRNA molecules butcorrelations were found when the relative percentage of the insertedFECH mRNA molecules was taken into account. A higher percentagewas related to the higher accumulation and excretion of these proto-porphyrins.

For the first time, this study has reported such a correlation at themolecular level. The accumulation of these metabolites is the directconsequence of a lowered activity of the FECH enzyme, as alreadyknown [4]. However, to the best of our knowledge, no correlation hasever been evaluated with the expression of the FECH gene.

In conclusion, our data confirmed a dependency on the genotype.We used a more specific and sensible approach and reported betterresults than the previous studies. At the same time, we confirmed thatthe percentage of the inserted FECH mRNA was doubled in the het-erozygotes for the c.315–48 T > C variant with respect to the T/Tgenotype.

In our patient group the C/C genotype was identified as a patho-logical class, although limited to only a few samples.The GTC haplotypein homozygosis is pathological too, when samples are categorised ac-cording to the haplotypes. Notably, the c.315–48 T > C variant isshown to reduce the expression of the FECH gene but the mechanismsby which this variant act on the levels of FECH mRNA remain to beelucidated by further molecular studies.

Acknowledgements

The authors are grateful to all patients, their families and controlsubjects who donated samples for this study.

We thank Prof MD Cappellini for her constant and continuoussupport.

VB designed the study, planned and performed experiments, ana-lysed data and wrote the paper; FG collected RNA samples and per-formed experiments; PM and SF performed porphyrins determinationand wrote the corresponding section; GG recruited patients; EDP su-pervised research and critically revised the paper.

All authors revised and agreed to the final version of the manu-script.

Table 7Cases of homozygous c.315-48 T > C variant. Reference values: *30–80 μg/dl; **<80 μg/dl RBC; ***≤190 μg/dl RBC; ᵻ30–86 μg/dl; n.r.: not reported; n.d.: nodata avaiable.

Patient FECH alleles EPPSymptoms

Erytrocytes porphyrins Ref.

c.1–252 A > G c.68-23C > T c.315-48 T > C

1. – – C > C Classic 3668 μg/dl* Saruwatari H et al. [17]2. – – C > C Mild 338 μg/dl RBC** Schneider-Yin X. et al. [18]3. – – C > C Classic n.r. Whatley S.D. et al. [19]4. G > G T > T C > C No 160 μg/dl RBC*** Colombo F.P. et al. [6]5. G > G T > T C > C No 141 μg/dl RBC***6. G > G T > T C > C No 98 μg/dl RBC***7. G > G T > T C > C No 105 μg/dl RBC***8. G > G T > T C > C No 170 μg/dl RBC***9. G > G T > T C > C No n.d.10. – – C > C Mild 115 μg/dlᵻ Mizawa M. et al. [15]11. – – C > C Mild 103 μg/dlᵻ12. – – C > C Mild 122 μg/dlᵻ13. – – C > C Slight photosensitivity n.r.14. – – C > C Slight photosensitivity n.r

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

9

Funding

The Italian Ministry of Health, with grant numbersGR–2011–02347219 to EDP and RC–2017, supported this work.

Conflict of interest statement

The authors declare no conflict of interest to disclose.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ymgme.2018.06.005.

References

[1] M. Balwani, J. Bloomer, R. Desnick, Erythropoietic Protoporphyria, AutosomalRecessive, (1993).

[2] M. Balwani, H. Naik, K.E. Anderson, D.M. Bissell, J. Bloomer, H.L. Bonkovsky,J.D. Phillips, J.R. Overbey, B. Wang, A.K. Singal, L.U. Liu, R.J. Desnick, Clinical,biochemical, and genetic characterization of north American patients withErythropoietic Protoporphyria and X-linked Protoporphyria, JAMA Dermatol. 153(2017) 789–796.

[3] J. Barman-Aksozen, C. Beguin, A.M. Dogar, X. Schneider-Yin, E.I. Minder, Ironavailability modulates aberrant splicing of ferrochelatase through the iron- and 2-oxoglutarate dependent dioxygenase Jmjd6 and U2AF(65.), Blood Cells Mol. Dis. 51(2013) 151–161.

[4] D.M. Bissell, K.E. Anderson, H.L. Bonkovsky, Porphyria, N. Engl. J. Med. 377 (2017)862–872.

[5] V. Brancaleoni, M. Balwani, F. Granata, G. Graziadei, P. Missineo, V. Fiorentino,S. Fustinoni, M.D. Cappellini, H. Naik, R.J. Desnick, E. Di Pierro, X-chromosomalinactivation directly influences the phenotypic manifestation of X-linked proto-porphyria, Clin. Genet. 89 (2016) 20–26.

[6] F.P. Colombo, M.V. Rossetti, M. Mendez, J.E. Martinez, S.R. de Enriquez, C.B.A. Del,V.E. Parera, Functional associations of genetic variants involved in the clinicalmanifestation of erythropoietic protoporphyria in the Argentinean population, J.

Eur. Acad. Dermatol. Venereol. 27 (2013) 754–762.[7] E. Di Pierro, V. Brancaleoni, V. Besana, M.D. Cappellini, Multiplex ligation-de-

pendent probe amplification: a novel approach for genetic diagnosis of Porphyria, J.Hum. Genet. 54 (2009) 479–487.

[8] E. Di Pierro, V. Brancaleoni, V. Moriondo, V. Besana, M.D. Cappellini, Co-existenceof two functional mutations on the same allele of the human ferrochelatase gene inerythropoietic protoporphyria, Clin. Genet. 71 (2007) 84–88.

[9] E. Di Pierro, M.D. Cappellini, R. Mazzucchelli, V. Moriondo, D. Mologni,P.B. Zanone, A. Riva, A point mutation affecting an SP1 binding site in the promoterof the ferrochelatase gene impairs gene transcription and causes erythropoieticprotoporphyria, Exp. Hematol. 33 (2005) 584–591.

[10] V. Fiorentino, V. Brancaleoni, F. Granata, G. Graziadei, P.E. Di, The assessment ofnoncoding variant of PPOX gene in variegate porphyria reveals post-transcriptionalrole of the 5′ untranslated exon 1, Blood Cells Mol. Dis. 61 (2016) 48–53.

[11] L. Gouya, H. Puy, A.M. Robreau, M. Bourgeois, J. Lamoril, S. Da, B. Grandchamp,J.C. Deybach, The penetrance of dominant erythropoietic protoporphyria ismodulated by expression of wildtype FECH, Nat. Genet. 30 (2002) 27–28.

[12] J.F. Huggett, S. Cowen, C.A. Foy, Considerations for digital PCR as an accuratemolecular diagnostic tool, Clin. Chem. 61 (2015) 79–88.

[13] C. Li, P.E. Di, V. Brancaleoni, M.D. Cappellini, D.P. Steensma, A novel large deletionand three polymorphisms in the FECH gene associated with erythropoietic proto-porphyria, Clin. Chem. Lab. Med. 47 (2009) 44–46.

[14] W.H. Lockwood, V. Poulos, E. Rossi, D.H. Curnow, Rapid procedure for fecal por-phyrin assay, Clin. Chem. 31 (1985) 1163–1167.

[15] M. Mizawa, T. Makino, H. Nakano, D. Sawamura, T. Shimizu, Incomplete ery-thropoietic protoporphyria caused by a splice site modulator homozygous IVS3-48Cpolymorphism in the ferrochelatase gene, Br. J. Dermatol. 174 (2016) 172–175.

[16] Y. Nakahashi, H. Fujita, S. Taketani, N. Ishida, A. Kappas, S. Sassa, The moleculardefect of ferrochelatase in a patient with erythropoietic protoporphyria, Proc. Natl.Acad. Sci. U.S.A 89 (1992) 281–285.

[17] H. Saruwatari, Y. Ueki, S. Yotsumoto, T. Shimada, S. Fukumaru, T. Kanekura,T. Kanzaki, Genetic analysis of the ferrochelatase gene in eight Japanese patientsfrom seven families with erythropoietic protoporphyria, J. Dermatol. 33 (2006)603–608.

[18] X. Schneider-Yin, R. Mamet, E.I. Minder, N. Schoenfeld, Biochemical and moleculardiagnosis of erythropoietic protoporphyria in an Ashkenazi Jewish family, J.Inherit. Metab. Dis. 31 (Suppl. 2) (2008) S363–S367.

[19] S.D. Whatley, N.G. Mason, S.A. Holme, A.V. Anstey, G.H. Elder, M.N. Badminton,Molecular epidemiology of erythropoietic protoporphyria in the U.K, Br. J.Dermatol. 162 (2010) 642–646.

V. Brancaleoni et al. Molecular Genetics and Metabolism xxx (xxxx) xxx–xxx

10

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具