The GNAS locus is an imprinted locus located in the long arm of chromosome 20 (20q13.2-20q13.3) (1). Due to its complex imprinted pattern, the expression of the multiple transcripts derived from the locus is dependent on the parental origin (2).

There are four first alternative exons, which are spliced onto exons 2-13, that are common for all the transcripts. These alternative first exons are regulated by four differentially methylated regions (DMRs), which are cis-acting elements that control the parent-of-origin expression: GNAS-NESP:TSS-DMR, GNAS-AS1:TSS-DMR, GNAS-XL:Ex1-DMR and GNAS A/B:TSS-DMR. The first one is methylated in the paternal allele and in consequence, is expressed from the maternal allele. In contrast, the other three are imprinted in the maternal allele and expressed from the paternal one (3–7).

When the methylation of the complex locus is abnormal in the DMRs, it gives rise to the disease known as pseudohypoparathyroidism type 1B (PHP1B, OMIM#603233) or according to the recent recommendations, inactivating PTH/PTHrP signaling disorder 3 (iPPSD3) (8). The methylation defect can affect one or multiple DMRs and it can be partial or complete (9), but all patients affected with iPPSD3 show loss of methylation (LOM) at the GNAS A/B:TSS-DMR (10). Clinically, iPPSD3 patients have been characterized by parathyroid hormone (PTH) resistance, leading to hypocalcemia and hyperphosphatemia. Sometimes, thyroid stimulating hormone (TSH) resistance is also present. In these patients, Albright’s Hereditary Osteodystrophy (AHO) phenotype is commonly absent, and if present, it is usually milder (10, 11). Some additional signs include excessive intrauterine growth (12, 13).

Of the iPPSD3 patients, 15%-20% of the cases are familial, with LOM at the GNAS A/B:TSS-DMR secondary to deletion of 3.3 kb in the STX16 gene. This genetic defect is inherited from the maternal lineage in an autosomal dominant manner (10), and is accompanied by a GNAS-AS2:TSS-DMR partial LOM (14). However, there are some rare familial cases with different deletions and duplications affecting only the GNAS A/B:TSS-DMR or all the DMRs (10). Apart from the familial cases, there are also sporadic iPPSD3 patients (15). In the majority of the sporadic cases, a methylation alteration usually affects the four GNAS-DMRs and the genetic cause of the methylation defect remains unknown. Nevertheless, in approximately 8%-20% of them, paternal uniparental isodisomy (iUPD) including 20q13 has been detected (15–17).

Uniparental disomy (UPD) is an abnormal genetic condition in which the individual carries a pair of chromosomes (or segments of chromosomes) from the same progenitor without a contribution from the other parent. Two different types of UPD can be distinguished: uniparental heterodisomy (hUPD), when both homologous chromosomes are inherited from one parent; and, uniparental isodisomy (iUPD), when a single chromosome is inherited from the parent and it duplicates (18, 19). Although it is very rare, translocations giving rise to paternal uniparental disomy of chromosome 20, upd(20)pat, have also been described (15).

For the detection of paternal iUPD in 20q13 [iUPD(20q)pat], current international guidelines for the diagnosis of PHP and related disorders recommend the analysis of microsatellites or short tandem repeats (STR) in the trio (if the samples are available) and/or single nucleotide polymorphism (SNP) array (15). However, if the samples of the parents are not available, STR analysis can be non-informative (20).

GNAS is a complex locus characterized by a specific imprinting pattern. When the methylation is altered, at least when hypomethylation in the GNAS A/B:TSS-DMR happens, patients develop iPPSD3 (10). The patient described in this work presented with a complete methylation defect affecting all four DMRs detected by the MS-MLPA specific for the GNAS locus, so he was classified as a sporadic case. In patients with complete alteration, paternal UPD should be tested, specifically, isodisomy of chromosome 20 [iUPD (20)] affecting the chromosomic region that includes the GNAS locus (10, 21–23).

Following the current guidelines (10), the genetic material of the proband was analyzed with an SNP array. An SNP array is a good technique to detect iUPD since in cases where a single homolog chromosome is transmitted in duplicate from a parent, whole (or a large part of) chromosomal homozygosity can be seen (23). As is shown in Figure 1 , there is a central zone with no heterozygosity in the patient. There are two possible reasons for this. First, when blocks of homozygosity are present in different chromosomes, parental consanguinity is suggested (24). However, at the genetic counseling consultation, the parents denied any relatedness and the SNP array results only showed loss-of-heterozygosity (LOH) in chromosome 20. Second, if large blocks of homozygosity are restricted to a single chromosome, it could represent UPD (25). Still, it was surprising that in the SNP array, there was a region with no LOH in which the GNAS locus was included. This could be a consequence of a chromosome with both isodisomy and heterodisomy. To confirm this, a custom MS-MLPA was designed and performed, confirming the loss of methylation at the four maternal DMRs located at 20q [MCTS2P:TSS-DMR (20q11.21), NNAT:TSS-DMR (20q11.2-q12), L3MBTL1:alt-TSS-DMR (20q13.12), and GNAS A/B:TSS-DMR (20q13.32)], which supported the possibility of UPD and was suggestive of a mixed disomy throughout chromosome 20.

Fortunately, to confirm this, samples from the parents were finally available and STR analysis was performed, with the STRs located around chromosome 20. Thus, as shown in Figure 1 ; Table 2 , the patient under study shows a combination of hetero- and isodisomy of chromosome 20. We would like to draw attention to this study of STRs. If, instead of the SNP array, we had initially analyzed the markers flanking GNAS (D20S496, D20S459, D20S443, D20S171, D20S173) in the absence of parental sample (as was in this situation), we would have ruled out iUPD as the patient carried a heterozygous genotype for some of them. And, since hUPD had been demonstrated once by STRs and the SNP array for the GNAS locus (17), this patient would have been characterized as spor-PHP1B/iPPSD3 with no underlying genetic cause. The correct (epi)genetic diagnosis of this disease allows for appropriate genetic counseling, as it is possible to establish the risk of recurrence and transmission of the disease in the presence of an underlying genetic alteration, whereas in the case of epimutation it would, for the time being, be unknown. Furthermore, if the patient carries iUPD, additional clinical manifestations secondary to an unmasked recessive disease may arise.

Moreover, from a research perspective, the correct subclassification of the different types of iPPSD3 cohorts [according to their (epi)genotype] could help to define clinical-epigenetic relationships and improve prognostic information. Thus, in the absence of parental samples, complete methylation defects at GNAS cannot be determined if it is due to an hUPD (20, 26), and, even for iUPD detection, a SNP array can be more informative than STRs (23), both have their limitations for hUPD identification. As an alternative, we propose the use of an MS-MLPA to examine additional DMRs at this chromosomal arm, as in our case.

It has already been well-demonstrated in the literature that this combination, called mixed UPD, is a molecular disease-causing mechanism described for several chromosomes (25, 27, 28). Regarding iPPSD3, researchers have also detected this genetic defect in chromosome 20, with an interstitial isodisomy in 20q affecting GNAS (29). Mixed UPD occurs when a chromosome contains both isodisomic and heterodisomic parts as a consequence of meiotic nondisjunction events. If the error occurs during meiosis I, a zone of homozygosity will be seen in the terminal region of the chromosomes. However, if the error is in meiosis II, the region of homozygosity is present in the centromeric region (23).

UPDs occur as a consequence of at least two independent errors during meiosis or after fertilization, so they do not follow a Mendelian inheritance (29) and are not necessarily disease-causing because no genetic material is lost or gained. However, some specific events lead to disease, as is the case in imprinting. The chromosomes that contain imprinted loci, as is the case for chromosome 20 and the GNAS locus, are sensitive to UPD. This happens because, due to UPD, the balance of methylation that leads to parental-dependent transcription is altered (23, 25).

In our study, we describe a patient with iPPSD3 with an heterodisomy in the GNAS locus. This is novel since even though iUPD(20q)pat is widely described, it is rare to find an heterodisomy as the cause of iPPSD3. In fact, only one case is described in the literature, and, as in our case, it was caused by a mixed UPD, and GNAS was found in the heterodisomic zone (17). In the reported patient, there were no informative STRs nor SNPs in the GNAS locus. There was one informative STR in the STX16 gene. The mutation rate of STR must be taken into account, which is quite a common event (30, 31) and could lead to a difficulty in heterodisomy confirmation. To solve this problem, we propose the inclusion of a novel custom-designed MS-MLPA to detect upd(20q). This custom-designed MS-MLPA includes probes that hybridize in different DMRs along the long arm of chromosome 20: MCTS2P:TSS-DMR, NNAT: TSS-DMR and L3MBTL1:alt-TSS-DMR. Thanks to the imprinted nature of these DMRs (32–34), it has been possible to detect an alteration in the methylation status ( Figure 2 ), showing complete hypomethylation. The combined information from the STRs/SNP array and this novel assay allowed us to confirm the mixed upd (20).

Although it is extremely rare, maternal UPD in chromosome 20 (upd(20)mat) has been detected in some patients. This genetic condition causes a poorly characterized syndrome known as Mulchandani-Bhoj-Conlin (MBCS, OMIM#617352). The main characteristics of this syndrome are growth failure (intrauterine/postnatal growth retardation) and feeding difficulties (35, 36). As the cause of MBCS is an UPD in chromosome 20, the MS-MLPA developed for the diagnosis of iPPSD3 can also be used for the detection of upd(20)mat and proceed to its diagnosis. At the moment, the specific (epi)genetic cause of MBCS still remains unknown, thus, a limitation of this custom MS-MLPA for the diagnosis of MBCS is that only 20q has been studied.

In conclusion, if the study of STRs is only done in the index case and it shows an heterozygote genotype, it is not enough to discard an UPD, as we present an extra confirmation that hUPD is possible at this locus. In this work, we have also presented a custom MS-MLPA for DMRs in the long arm of chromosome 20. Its efficiency in detecting upd(20q)pat as the underlying genetic cause in iPPSD3 cases has been demonstrated.