Journal of Molecular Neuroscience

, Volume 43, Issue 3, pp 333–336

Population and Computational Analysis of the MGEA6 P521A Variation as a Risk Factor for Familial Idiopathic Basal Ganglia Calcification (Fahr’s Disease)


  • Roberta R. Lemos
    • Keizo Asami Laboratory (LIKA)Federal University of Pernambuco
  • Danyllo F. Oliveira
    • Keizo Asami Laboratory (LIKA)Federal University of Pernambuco
  • Mayana Zatz
    • Human Genome Study CenterUniversity of São Paulo
    • Keizo Asami Laboratory (LIKA)Federal University of Pernambuco
    • Neuropsychiatry DepartmentFederal University of Pernambuco (UFPE)

DOI: 10.1007/s12031-010-9445-7

Cite this article as:
Lemos, R.R., Oliveira, D.F., Zatz, M. et al. J Mol Neurosci (2011) 43: 333. doi:10.1007/s12031-010-9445-7


Familial idiopathic basal ganglia calcification, also known as “Fahr’s disease” (FD), is a neuropsychiatric disorder with autosomal dominant pattern of inheritance and characterized by symmetric basal ganglia calcifications and, occasionally, other brain regions. Currently, there are three loci linked to this devastating disease. The first one (IBGC1) is located in 14q11.2-21.3 and the other two have been identified in 2q37 (IBGC2) and 8p21.1-q11.13 (IBGC3). Further studies identified a heterozygous variation (rs36060072) which consists in the change of the cytosine to guanine located at MGEA6/CTAGE5 gene, present in all of the affected large American family linked to IBGC1. This missense substitution, which induces changes of a proline to alanine at the 521 position (P521A), in a proline-rich and highly conserved protein domain was considered a rare variation, with a minor allele frequency (MAF) of 0.0058 at the US population. Considering that the population frequency of a given variation is an indirect indicative of potential pathogenicity, we screened 200 chromosomes in a random control set of Brazilian samples and in two nuclear families, comparing with our previous analysis in a US population. In addition, we accomplished analyses through bioinformatics programs to predict the pathogenicity of such variation. Our genetic screen found no P521A carriers. Polling these data together with the previous study in the USA, we have now a MAF of 0.0036, showing that this mutation is very rare. On the other hand, the bioinformatics analysis provided conflicting findings. There are currently various candidate genes and loci that could be involved with the underlying molecular basis of FD etiology, and other groups suggested the possible role played by genes in 2q37, related to calcium metabolism, and at chromosome 8 (NRG1 and SNTG1). Additional mutagenesis and in vivo studies are necessary to confirm the pathogenicity for variation in the P521A MGEA6.


MGEA6P521A variationPolymorphismFamilial idiopathic basal ganglia calcificationFahr’s disease


Familial idiopathic basal ganglia calcification (IBGC), also known as “Fahr’s disease” (FD), is a complex condition which exhibits various neuropsychiatric symptoms such as psychosis, motor impairment, and mood disorders. However, the most striking signal of this condition is bilateral brain calcinosis in basal ganglia, and, occasionally other brain regions, with normal biochemical and endocrinological profiles (Manyam 2005).

The first Fahr’s disease locus (IBGC1) was described in a large multigenerational American family in a 13.3-cM region (14q11.2-21.3) with an autosomal dominant pattern of inheritance. Another small kindred from Spain was also reported as being possibly linked to this locus, narrowing the candidate region to 10.9 cM. (Geschwind et al. 1999; Oliveira et al. 2004).

However, this locus was excluded in other families from China, Canada, and Germany, suggesting genetic heterogeneity (Oliveira et al. 2004). So far, other two loci have been identified in families from Italy (Volpato et al. 2009) and from China (Dai et al. 2010). These are in 2q37 and 8p21.1-q11.23, respectively.

In a systematic search for a candidate mutation at the IBGC1 locus, we identified a heterozygous non-synonymous single nucleotide polymorphism (C>G) at the MGEA6/cTAGE5 gene (rs36060072) in every affected member, but not at the controls, of the family firstly associated with this locus. This missense substitution, which is responsible for the change of the proline to alanine at the 521 position (P521A) in a proline-rich and highly conserved domain, was considered a rare variation, with a minor allele frequency of 0.0058 in the US population (Oliveira et al. 2007).

Curiously, this variation is located at the exon which is commonly spliced at the MGEA6 gene, generating the isoform MGEA 11, also expressed in the brain (Usener et al. 2003). MGEA6 is a coil-coiled protein expressed in several tissues including the brain, highly expressed in meningioma, a commonly benign intracranial tumor habitually presenting calcification visible at neuroimaging studies (Comtesse et al. 2001, 2002; Usener et al. 2003).

Considering that the population frequency of a given variation is an indirect indicative of potential pathogenicity (Freudenberg-Hua et al. 2003), we present now the sequencing of 100 control samples from Brazil to ascertain a local minor allele frequency (MAF) for this variation at the MGEA6, comparing the results with our previous analysis in the USA. We also screened two nuclear families for this polymorphism. One of them has already been reported in literature and presents a 53-year-old male proband with IBGC and neuropsychiatric symptoms (Oliveira et al. 2009). This proband is the father of two 23-year-old asymptomatic identical twins with extensive brain calcinosis in basal ganglia, white matter, and cerebellum. The second family is inedit and also presents a similar pattern of calcification in cerebellum and basal ganglia over three generations, one affected (87 years old) and two unaffected (33 and 58 years old), all females. Altogether, these patients did not present any known metabolic disturbances on clinical examination (Fig. 1).
Fig. 1

Analysis of pedigrees in two nuclear families. a The first family has already been reported in literature and has two identical twins with bilateral brain calcification, but without any psychiatric symptoms at all. b The second family has a three-generational transmission of brain calcification and is inedit in literature

Methods and Sample Collection

The DNA was extracted by salting out procedure of blood samples, followed by proteinase K digestion and precipitation with ethanol. The polymorphic region was amplified by polymerase chain reaction (PCR) and the cycling program was 95°C for 5 min, followed by 35 cycles of 95°C for 30 s, 51.5°C for 30 s, 72°C for 45 s, followed by a final extension of 72°C for 10 min. The primers, manually designed, used in this procedure were 5′tttcagatgaaggccaacccaatggtgagg3′ (for) and 5′accatatggggaatgctctggaataattaa3′ (rev). PCR was performed in a 25-μl reaction volume containing 100 ng of DNA in a reaction mixture containing 2 μM of MgCl2, 2 μL of dNTPs, 0.75 U Taq DNA polymerase (Invitrogen), 1× PCR buffer (Invitrogen) 0.2 μM.

The PCR products were purified by enzymatic reaction of exoquinase/shrimp alkaline phosphatase (GE). Finally, the products were sequenced in a Mega Bace 1000 (Sunnyvale, CA) and ABI Prism 3100 genetic analyzer (Lincoln, CA).

Besides the sequencing analysis, we accomplished an in silico study through bioinformatics programs to perform multiple amino acid sequence alignments and to predict the pathogenicity of a given polymorphism based on the protein primary sequence or tertiary structure, or both. Because the three-dimensional structure of the MGEA6 genic product is not yet available, we focused only in the examination through the sequence.

Initially, the software CLCBio Workbench Combined®, version 3.6.2, was used to overlap and compare the candidate region where the P521A was found, between the human MGEA6 and its orthologous genes from Pan troglodytes, Canis lupus familiaris, Sus scrofa, Mus musculus, Gallus gallus, and Danio rerio.

The bioinformatics analysis was performed with five programs: PMUT, PolyPhen, SAP, SIFT, and I Mutant. This last one does not predict if a determined mutation is pathogenic, but analyzes the changes in the free energy levels caused by the mutation and if this event might disrupt the structural stability of this macromolecule.

This study was approved by the ethics committee from the Federal University of Pernambuco (CAEE-0296.0172.000-08) and the subjects signed informed consent.

Results and Discussion

We found no P521A carriers during the sequencing screening of the control samples (Table 1). Before our studies, the MAF of this polymorphism was 0.0058 (Oliveira et al. 2007).
Table 1

Prediction of the pathogenicity and thermodynamic changes caused by P521A MGEA6 polymorphism




Electronic address






Probably damaging





Decreased molecular stability∼emidio/I-Mutant/I-Mutant.htm



Not pathogenic

There is an important ethnical heterogeneity when comparing random populations between Brazil and the USA; however, polling these data together with the previous study in the USA, we have now a MAF of 0.0036. In previous studies, it was attested that possible damaging mutations have a MAF < 0.05 in a European population (Freudenberg-Hua et al. 2003). Thus, this result corroborates to point the P521A MGEA6 as a potentially pathogenic mutation. None of the patients from the two FD families presented the polymorphism, reinforcing the genetic heterogeneity to this disease (Oliveira et al. 2004).

The orthologous gene comparison shows that this variation is located in a highly conserved region, suggesting that a non-synonymous mutation at this spot is potentially deleterious (Fig. 2).
Fig. 2

Multiple alignment of the MGEA6 proline-rich region where the polymorphism is located (pointed by arrow) shows that this region is highly evolutionarily conserved

The bioinformatics analysis provided conflicting findings (Table 1). SAP, PolyPhen, and I Mutant suggest that this variation is predicted to cause probable molecular deleterious effects. I Mutant is the only program that does not predict the pathogenicity, but rather analyzes the mutation through a thermodynamic perspective, pointing that this polymorphism decreases protein stability. In addition, PolyPhen classified this mutation as “probably damaging.”

The SIFT program generated conflicting findings. The use of UniProt-SWISS PROT 56.6, and NCBI databases, differently used during the comparative predictions, pointed this polymorphism as pathogenic. However, the same analysis using the UniProt-TrEMBL 39.6 database considered this substitution as neutral. The PMUT considered this same mutation as neutral.

The growing number of annotated SNP in databases raises the need of in silico tools with a high predictive power because it is unfeasible to verify individually the ramping number of variations found during the current sequencing and resequencing projects worldwide. All of these programs have been reported in the literature as highly precise tools to screen candidate SNPs; however, other studies have pointed potential failures (Ferrer-Costa et al. 2005; Ng and Henikoff 2006; Hu and Yan 2008; Di et al. 2009; Dorfman et al. 2010).

The lack of consensus in the analysis might also be due to the limited skill of these programs in the prediction. In addition, the discrepancies might be due to the poor source of data available for some genes. In the MGEA6 case, there is no structural model to the proteic product, and it could be a restraint to these online programs.

Proline-rich domains are reported as a widespread motif in genomes of various organisms and play a critical role in the construction of preferential three-dimensional structures, recognizing and interacting with domains such as WW and Src homology 3, which are involved in a wide variety of cellular activities such as growth, cell cycle, transcription, synaptic signalization, and cell motility (Zarrinpar et al. 2003).

In addition, there are other suggestive candidate genes that could be the underlying molecular basis of FD etiology. Volpato et al. (2009) point out several genes related to calcium metabolism located in 2q37, such as INPP5 and EFHD1. Curiously, two families with Fried syndrome (X-linked mental retardation with hydrocephalus and calcifications in basal ganglia) have been reported in France and Scotland and bearing AP1S2 mutations (Saillour et al. 2007).

More recently, Dai et al. (2010) reported a Chinese family linked to chromosome 8, but an initial screening at the NRG1 and SNTG1 genes found no mutation.

Previous preliminary linkage studies pointed to other candidate regions at chromosomes 7, 8, and 9, but additional fine mapping analyses are pending (Oliveira et al. 2003; Wszolek et al. 2006)

Therefore, continuous familial and population screening, in vivo studies, and mutagenesis analysis are necessary to definitely comprehend the pathogenesis of Fahr’s disease


We are deeply indebted to Dr. Cíntia Rocha, Dr. José Luiz de Lima Filho (LIKA), and Dr Constância Ayres (Aggeu Magalhães Research Center) for the support with sequencing experiments. We also thank Dr. Maria Rita Passos Bueno and Lilian Calabró dos Santos (University of São Paulo) for providing the control DNA samples. This study was supported by the following Brazilian funding agencies and academic bureaus: CNPq (Universal 2008-480149/2008-9), PROPESQ-UFPE, CAPES, and FACEPE (APQ-0997-4.01/08) and PIBIC-UFPE. J.R.M.O. holds a research fellowship from the Brazilian National Research Council (CNPq, Brasília, Brazil) and from the John Simon Guggenheim Foundation (New York, USA).

Copyright information

© Springer Science+Business Media, LLC 2010