DISCUSSION
This is the largest series of patients with SLC25A38 associated CSA yet reported, describing 31 individuals from 24 different families and 11 novel mutations, expanding the total number of reported families and pathogenic alleles to 92 and 47, respectively.
Despite the diversity of mutations, there are several unexpected aspects of the SLC25A38 anemia revealed by these studies worth noting. First is the very limited evidence that there is a genotype-phenotype correlation. Essentially all patients present at birth or infancy with a severe hypochromic, microcytic anemia that eventually requires chronic transfusion. This is in stark contrast with the most common form of hypochromic microcytic, non-syndromic CSA, XLSA, which is the major differential diagnosis. Male patients with XLSA may present at birth to older adulthood. The most severe XLSA cases tend to present at an earlier age, but it is unusual for a patient to have transfusion dependent anemia as is typical of SLC25A38 disease. Indeed, the anemia in XLSA is frequently incidental and may be discovered only by screening or as a result of investigation of unexplained iron overload. There are, however, several exceptional cases of patients with SLC25A38 disease coming to medical attention in their teens or twenties. Three of these patients had homozygous mutations at codon 134 [p.Arg134His or p.Arg134Cys] (Fouquet et al., 2019; Hanina, Bain, Clark, & Layton, 2018; Le Rouzic et al., 2017). However, two other patients homozygous for the p.Arg134Cys allele presented at age 2 months and 2 years (W. An et al., 2015; W. B. An et al., 2019; Kannengiesser et al., 2011). In our own cohort, a patient with a homozygous variant at the only incompletely conserved +5 position of a splice donor site (c.792+5G>C) presented in his mid-teens. It is certainly possible that other genotype-phenotype correlations are masked by the clinical imperative to initiate transfusions in a patient with a HGB less than ~9 g/dL. Indeed, in the publications in which an initial diagnostic HGB is reported, only one individual, a neonate, had a hemoglobin >9 g/dL (Guernsey et al., 2009; Hanina et al., 2018; Kannengiesser et al., 2011; Liu et al., 2013; Wong et al., 2015).
Although clinical practice to transfuse these patients may disguise the subtle differences between SLC25A38 genotypes, it is abundantly evident that an SLC25A38 null genotype does not preclude some mitochondrial glycine being available for heme synthesis. This may be due to other, less specific transporters, possibly including the highly homologous protein SLC25A39, or pathways that produce glycine from other amino acids, such as serine (Amelio, Cutruzzola, Antonov, Agostini, & Melino, 2014; Kory et al., 2018). Leveraging these pathways may provide an avenue for therapy. However, based on limited data, it is unlikely that glycine supplementation alone will suffice (Fernandez-Murray et al., 2016; LeBlanc et al., 2016).
The SLC25A38 anemia is regarded as non-syndromic. Nevertheless, twelve of the 31 patients described here had developmental or intellectual disabilities. Similar abnormalities, including psychomotor delay, hypotonia, facial dysmorphism (Fouquet et al., 2019), Hypospadias (W. An et al., 2015), congenital myelomeningocele, patent ductus arteriosus and ventricular septal defects (Wong et al., 2015) have been reported in several other patients, but no abnormality is unusually prevalent across multiple families to suggest a specific syndromic association.
Three patients who underwent splenectomy developed thrombocytosis and/or recurrent thrombosis, further supporting the notion that splenectomy may be contraindicated in SLC25A38 CSA as has generally been advocated in other patients with microcytic CSAs (Bottomley & Fleming, 2014; Fouquet et al., 2019).
The 47 described SLC25A38 pathogenic mutations occur at 40 different codons. However, nearly one-third (27 of 92) families carry at least one copy of either the c.324_325del or the c.349C>T allele. The latter has been identified in families of Acadian (Guernsey et al., 2009), African American (current report), South Asian (Ravindra et al., 2020), Greek (Guernsey et al., 2009), and Northern European (this report) origin, suggesting that it has reoccurred on multiple occasions. In fact, in one case (21.1), the patient is homozygous for the c.349C>T allele, associated null allele, but one copy also carries a MS variant in a non-conserved residue in cis (c.[161G>A;349C>T]; p.[Arg54His;Arg117X]). Reanalysis of other SLC25A38 anemia patients previously reported by us (Guernsey et al., 2009) identified one other patient having this compound mutant chromosome (data not shown). Both of these variants occur at hypermutable cytosine-guanosine (CpG) dinucleotides. This would support the notion that the p.Arg117X allele has occurred on multiple occasions and that homozygosity for a disease-associated variant in SLC25A38 should not necessarily be taken as evidence of identity by descent.
Because of the severity and great similarity of the disorder to transfusion-dependent β-thalassemia (thalassemia major), all of the patients in our cohort were at one time managed with transfusion and iron chelation in a manner similar to thalassemia. This is similarly true of nearly all of the patients described in the literature. Just as we identified no consistent, distinctive syndromic aspects of the disease outside the anemia, we did not observe any complications or undue toxicity as a result of transfusion or iron overload. The oldest patient in this cohort was 39 years of age, and we are aware of at least two patients in their early fifties (Guernsey et al., 2009), suggesting that modern transfusion and chelation regimens support long-term survival.
Nonetheless, 9 of our patients underwent allogenic HSCT at varying times during their disease course. Ten other patients (Guernsey et al., 2009; Kannengiesser et al., 2011; Uminski et al., 2020) as well as the sibling of one patient included in this series (Kim et al., 2018) are reported to have been transplanted. Comprehensive details are unavailable for most of these patients, but when stated, similar to those described here, HSCT regimens have variously included matched-related and matched unrelated donors with fully myeloablative or reduced intensity conditioning regimens. In many cases, aggressive chelation was employed to reduce the iron burden pre-transplant and phlebotomy was often used post-HSCT to further normalize iron stores. In all, of 18 patients receiving HSCT, 14 have achieved transfusion independence, 3 grafts failed, and one patient died in the immediate post-transplant period with follow up ranging from months to 19 years. In no case has an unusual disease-specific transplant-related morbidity been reported. Thus, allogeneic HSCT provides a curative option in SLC25A38 CSA, and may be considered in young patients with appropriate donors prior to the development of sequelae of chronic transfusions such as hemosiderosis and alloimmunization.