Improving the diagnosis of cobalamin and related defects by genomic analysis, plus functional and structural assessment of novel variants

Cellular cobalamin problems are caused by nutritional deficiency or genetic defects that affect either the absorption or cellular uptake of the vitamin or the synthesis from it of methylcobalamin (MeCbl) and adenosylcobalamin (AdoCbl). The latter are, respectively, a cofactor of methionine synthase (MS EC_2.1.1.13), which catalyzes the remethylation of homocysteine (Hcys) to methionine in the cytoplasm, and of methylmalonyl-CoA mutase (MUT EC_5.4.99.2), which catalyzes the mitochondrial isomerization of L-methylmalonyl-CoA (MMACoA) to succinyl-CoA. Defects in the synthesis of AdoCbl or the conversion of MMACoA to succinyl-CoA lead to elevated methylmalonic acid (MMA) concentrations, while defects in the synthesis of MeCbl or the remethylation of Hcys to methionine cause elevated Hcys (HC). Defects in the absorption and transport of vitamin B₁₂, and in the cytosolic synthesis of the above cofactors, cause methylmalonic aciduria with homocystinuria (MMA&HC) [1, 2].

After its intake, cobalamin is first bound to haptocorrin (encoded by TCN1) and then to the intrinsic factor (IF) encoded by GIF. To date, only one pathogenic mutation and one functional polymorphism (p.Asp301Tyr in TCN1) have been described [3, 4]. Cubilin and amnionless, encoded by CUBN and AMN respectively, form the cubam dimer which functions as the receptor of IF-Cbl in the ileum [4]. The IF is then degraded and vitamin B₁₂ appears in the blood stream associated with transcobalamin II (TCII) [2, 5]. Inherited malabsorption of cobalamin causes haematological and neurological abnormalities that can be fatal [6].

Still bound to TCII, the vitamin enters cells via transcobalamin receptor (TCblR)-mediated endocytosis. This receptor is encoded by CD320, for which only one mutation has been described [7]. The cobalamin is then released into the cytosol, an event impaired in people with defects in LMBRD1 (cblF complementation group, MIM #277380) [8]. LMBRD1 protein interacts with ABCD4 protein, which is also involved in the release of Cbl into the cytoplasm from the lysosomes (a defect which falls into complementation group cblJ) [9]. The most common form of cobalamin defect, complementation group cblC (MIM#277400), is caused by mutations affecting MMACHC [10]. Patients with mutations in the Host Cell Factor 1 (HCFC1) locus (MIM 309541), located on the X chromosome, have a phenocopy of cblC disease [11] since HCFC1 encodes a transcriptional co-regulator of MMACHC. Recently additional defects have been described that affect transcription factors involved in the regulation of Cbl pathway: ZNF143 [12] and THAP11 [13]. In addition new variants in PRDX1 also affect the expression of MMACHC, named epi-cblC cases. Thus, these patients result from pathogenic variants in MMACHC and one in PRDX1 which force the antisense transcription of MMACHC and thereby a possible methylation mark [14].

Cellular cobalamin processing occurs via two major pathways - the cytosolic and mitochondrial pathways. The protein thought to be responsible for sorting cobalamin for these pathways is MMADHC (cblD complementation group, MIM #611935). This complementation group is the most complex of all since patients present with biochemical heterogeneity ranging from isolated homocystinuria (cblD variant 1) or isolated methylmalonic aciduria (cblD variant 2) through to methylmalonic aciduria combined with homocystinuria (cblD) depending on the nature and location of the mutations present [15].

The cytosolic pathway is involved in the synthesis of MeCbl by MS, encoded by MTR (defects give rise to complementation group cblG), while the enzyme methionine synthase reductase encoded by MTRR is involved in the reactivation of MS (defects give rise to complementation group cblE). Both cblG (MIM#250940) and cblE (MIM#236270) defects, lead to elevated Hcys, hypomethioninaemia, and megaloblastic anaemia [1, 2].

Inactive cob(II)alamin entering the mitochondria is converted into AdoCbl via a reductive adenosylation reaction catalysed by an adenosyltransferase (ATR) encoded at the cblB locus (MIM#251110) [16]. MUT is encoded at the MUT locus and two different phenotypes have been described: mut⁰ and mut⁻ (MIM#251100). Finally MMAA is the gene associated with complementation group cblA defects (MIM# 607481). It encodes a protein of the same name which may be involved in transferring AdoCbl from ATR to MUT protein as well as maintaining MUT’s functional integrity [5, 17].

Mildly elevated MMA also occurs in patients with mutations affecting SUCLA2 (MIM#612073) and SUCLG1 (MIM#245400) which lead to infantile mitochondrial encephalopathic depletion syndrome. These genes encode either the ß-subunit of ADP-forming or the α-subunit of GDP-forming succinyl-CoA synthetases. The accumulated succinyl-CoA inhibits the metabolism of MMACoA to succinyl-CoA, leading to the accumulation of MMA in body fluids. Mutations in SUCLA2 give rise to typical (though rare) early onset dystonia combined with deafness [18, 19]. An increase in MMA is also seen in patients with mutations in ACSF3 (MIM# 614265), in whom malonic acid levels may also be elevated (combined malonic and methylmalonic aciduria, CMAMMA) [20]. ACSF3 encodes a mitochondrial acyl-CoA synthetase considered to have malonyl-CoA and MMACoA synthetase activities.

The diagnosis of an intracellular cobalamin metabolism disorder in a symptomatic individual is based on both clinical suspicion and biochemical analyses [1, 21]. Since expanded newborn screening potentially allows early detection of certain disorders of intracellular cobalamin metabolism, some affected individuals may be diagnosed prior to the onset of clinical symptoms. Diagnosis is based on elevated propionylcarnitine (C3), a high propionylcarnitine to acetylcarnitine ratio (C3/C2), high heptadecanoylcarnitine (C17), and/or reduced methionine concentrations [22]. After symptomatic or asymptomatic detection, a differential diagnosis is needed to help identify the gene that might be involved. Confirmatory testing is based on measuring Hcys and MMA in plasma and/or urine, enzyme analyses, the incorporation of [1-¹⁴C] propionate and [5-¹⁴C] methyl-THF into proteins in fibroblasts cultured in basal and hydroxocobalamin-supplemented media, and in some cases by cellular complementation assays [1]. However, biochemical analysis cannot pinpoint a genetic defect, making diagnosis complicated. Fortunately, recent developments in high-throughput sequence capture have made next generation sequencing (NGS) a rapid and accurate means of analysing genetic locus and allele heterogeneous disorders [23‐25].

The main purpose of clinical laboratory testing is to support medical decision-making. Clinical genetic testing is generally used to identify or confirm the cause of a disease and, if possible, to provide ideas for personalized treatment. Massive parallel sequencing is the starting point of a complex process for identifying pathogenic variants as well as the description of new functions for a gene product [25]. The aim of the present work was to assess the use of genetic studies as second-tier tests following the detection of cobalamin disorders in biochemical newborn screening. The usefulness of massive parallel sequencing in the identification of the mutated genes responsible for cobalamin and related defects was analyzed, and the structural and functional analysis of the novel variants identified undertaken.

The gold standard for coming to genetic diagnoses of cobalamin defects has for some time been gene-by-gene Sanger sequencing of individual DNA fragments. Enzymatic and cellular methods are employed before such sequencing to help in the selection of the gene defects to be sought, but this is time-consuming and laborious [1]. Further, no biochemical methods have been available to test for cobalamin absorption and transport defects, and while plasma B₁₂ concentrations are tentatively used to help in the selection of genes to be sequenced, this has probably led to the under-recognition of cases [34]. This costly, stepwise, and time-consuming methodology is gradually being replaced by NGS technologies, which offer higher throughput and scalability, cost less per sequenced nucleotide and have shorter turnaround times [35]. The present work reports the NGS analysis of a cohort of patients suspected of having cobalamin defects (plus the corresponding structural/functional analyses for some of the defects involved) and shows it to be quick and reliable.

A 100% correct diagnostic rate was returned with the validation cohort, with exonic and intronic nucleotide changes and small deletions (c.271dupA in MMACHC or c.57_64del8 in MMADHC) all successfully detected. With the clinical exome massive parallel sequencing system used, the depth and breadth of coverage of the cobalamin metabolism-related genes was > 30× and 99% respectively. The sensitivity and specificity of the method appeared to be very good and the results reproducible (several mutations were detected in more than one unrelated mutant allele i.e., c.271dupA in MMACHC, c.671_678dupAATTTATG in MUT, c.748C > T in MMADHC etc.). Thus, even though the use of smaller, specific panels is recommended [36] so that secondary incidental findings are avoided, the present work shows clinical exome sequencing to be successful. It could be used in sequencing analyses of single patients as well groups of patients whose members have other, unrelated disorders.

The LoF mutations detected in GIF and AMN were found to affect the absorption of cobalamin. GIF and AMN, in addition to CUBN and TCN1, are not usually captured in the available panels offered by genetic companies but the results reported in the present work have demonstrated they should be analyzed [19, 34].

The potential missense changes p.Ala302Pro and p.Ala676Thr, identified in MUT, were predicted to be damaging by several bioinformatic algorithms. The present detection of the associated mutations in combination with functional cellular studies provided a rational basis for tailored treatment. Indeed, the cellular uptake of ¹⁴C-propionate revealed that p.Ala302Pro is a mut⁰ mutation, while p.Ala676Thr located in the C-terminal of the cobalamin binding domain is likely a mut⁻ mutation. Based on these results, patient P6 might not be responsive to the pharmacological administration of the vitamin, while P9 would likely respond. Case P11 inherited a described LoF mutation in MUT and the functional polymorphism p.Asp301Tyr in TCN1, probably an incidental finding of no clinical significance. This asymptomatic individual was detected in newborn screening. No other mutations were detected in the coding region of MUT, not even in the deep intronic sequence reported to bear an intronic pathogenic mutation [30, 37]. Although massive parallel sequencing exome analysis cannot rule out the presence of other nucleotide changes in promoter or intronic sequences, both the absence of clinical symptoms and the slight abnormal levels of urinary MMA in case P11 suggested he/she was a carrier of the disease.

Variations in genes that do not encode cobalamin metabolism-related proteins (ASCF3 and SUCLA2) were also detected in the present work. Both were associated with slightly increased urinary MMA. The two new variants of SUCLA2 (p.Gly326Arg and p.Ile312Thr) detected are probably damaging changes; both were predicted to be so in bioinformatic and structural analyses. Neither was found to be present in database control populations nor in the in-house control samples. Structural analysis suggested that the replacement of Ile312 may have a mild functional impact, which agrees with the less strict conservation pattern seen in multiple sequence alignments. Importantly, the mitochondrial dysfunction in deficient fibroblasts from P24 was rescued via the stable expression of the SUCLA2 protein through lentiviral transduction. This suggests that SUCLA2 is the gene likely affected in this patient, although his/her symptoms were atypical [18, 38]. With respect to ASCF3, individuals P24, P25, P26 and P27 were diagnosed as having isolated MMA since the urinary malonic acid that identifies CMAMMA can be easily missed [39]. Given the difficulty in diagnosing CMAMMA, several of the present patients were originally diagnosed during genetic analyses, e.g., exome sequencing [40]. Even though a new method for the rapid metabolic diagnosis of CMAMMA has recently been reported [39], diagnoses by genetic analysis would be improved by second tier genetic testing after newborn screening. In some cases ACSF3 defects only cause biochemical conditions such as short-chain acyl-CoA dehydrogenase deficiency [41] or 3-methylcrotonylglycinuria [42], and symptoms appear only during viral infections etc. [39, 43]. In other cases, as reported in the present work, patients can show neurological damage [19]. Prompt diagnosis would avoid unnecessary treatment with vitamin B₁₂.

The use of massive parallel sequencing in a new era of newborn screening would avoid the time-consuming biochemical and enzymatic methods presently used to help identify which genes might be causing isolated MMA, MMA&HC, HC or CMAMMA. The rapid return of an accurate genetic diagnosis would allow the prescription of an appropriate treatment just a few days after birth. This report illustrates the clinical usefulness of massive parallel sequencing in the diagnosis of cobalamin and related defects. This powerful technology could improve the detection of these under-recognized or rare vitamin B₁₂ defect conditions.

This work provides evidence supporting the use of massive parallel sequencing as second tier test for confirming cobalamin defects following their detection in newborn screening. Biochemical and cellular analyses and additionally comprehensive phenotyping are needed for an accurate classification of variants for making fully useful clinical reports.