Three-generation female cohort with macrocytic anemia and iron overload

Abstract

A 16-year-old female (P-IIIA) with a past medical history of asthma presented to the hematology clinic with mild macrocytic anemia (hemoglobin 10 g/dL, red blood cells [RBC] 3.19 × 10⁶/μL, mean corpuscular volume [MCV] 103 femtoliters [fL], and red cell distribution width [RDW] of 12.9%). Platelet and white blood cell counts were normal. Two complete blood counts (CBC) tested within the previous 3 years on two occasions, once for a syncopal episode and the other for abdominal pain, had demonstrated similar findings. She was referred to hematology after her macrocytic anemia was noted during an attempted blood donation. At the hematology clinic, along with the CBC results noted above, she was found to have a negative direct antiglobulin test, a mildly elevated ferritin (158 ng/mL), with normal renal function, iron, total iron binding capacity (TIBC, 322 μg/dL), and iron saturation (30%). She reported mild occasional fatigue and intermittent thoracic back pain but was otherwise asymptomatic. She denied bleeding symptoms. Her examination was unremarkable, and she did not have evidence of splenomegaly. At the time of evaluation, she had transitioned to a pescatarian diet for 7 months, eating also vegetables and dairy. Her grandmother (P-I), mother (P-II), and younger sister (P-IIIB) were also noted to have mild macrocytic anemia, while P-II had also been found to have elevated ferritin of unclear etiology (reaching 646 ng/mL, but improved after phlebotomy every 6 weeks for a year).

Macrocytosis is defined as an MCV above the upper limit of normal for age, which varies from infants to adults. Increased MCV in an automatic complete blood count (CBC) may be artifactual, such as with a marked compensatory reticulocytosis in response to anemia; reticulocytes are normally larger than mature circulating red cells (up to 126 fL for healthy adults), while they can be even larger in conditions of stress erythropoiesis, raising the perceived MCV of the total erythrocytes (mature and non-mature) in automated CBC measurements.¹ Cold agglutinins may also cause spurious macrocytosis due to cell counting errors of red cell aggregates.² Obtaining the reticulocyte count and evaluating the blood smear are always valuable first steps to confirm a true macrocytosis and start the diagnostic evaluation.

True macrocytosis, typically associated with anemia, can have numerous causes based on clinical context, encompassing both acquired and congenital disorders. The most common pathophysiologic mechanisms involved are (1) impaired DNA synthesis in the erythroblast nuclei during terminal erythropoiesis causing megaloblastic bone marrow changes because of nuclear-cytoplasmic asynchrony, (2) altered red cell hydration increasing cell volume, or (3) altered composition of the lipid bilayer of the RBC membrane.

Impaired DNA synthesis in megaloblastic anemia is caused by vitamin B12 or folate deficiency, both critical components in the thymidine synthesis pathway. Deficiency of folate and/or vitamin B12 occurs either due to dietary restrictions, malabsorption syndromes like short-gut syndrome, or in autoimmune disorders associated with loss of intrinsic factor production (atrophic gastritis) or anti-intrinsic factor antibodies, which would prevent vitamin B12 absorption. Rare genetic disorders of vitamin B12 or folate transport and metabolism also exist, leading to syndromes that may present not only with megaloblastic anemia, but also with failure to thrive, other cytopenias, neurodevelopmental disorders, and/or thromboembolic events.^{3, 4} Medications, such as methotrexate (a folate antimetabolite), antiretrovirals, and chemotherapeutic nucleoside analogs, all of which interfere with DNA synthesis, may also cause macrocytosis with or without anemia. Notably, a very rare cause of macrocytosis is copper deficiency which also causes neutropenia and neurological manifestations like B12 deficiency.

Genetic disorders causing altered RBC hydration and hereditary hemolytic anemia (HHA) with macrocytosis include PIEZO1- or KCNN4-associated hereditary xerocytosis (HX), and RHAG-associated overhydrated stomatocytosis (OHSt). Increased reticulocyte count and characteristic osmotic gradient ektacytometry (OGE) curves point toward these diagnostic possibilities. OGE is a diagnostic assay measuring the deformability of RBCs as they are subjected to constant shear stress in a medium of increasing osmolality in a laser diffraction viscometer, allowing for differential diagnosis between the various RBC membrane disorders.⁵ Macrocytic stomatocytes and macrothrombocytopenia associated with increased phytosterol content in the membrane lipid bilayer are seen in the autosomal-recessive disorder of sitosterolemia.⁶ Increased cholesterol and phospholipids in the RBC membrane contribute to the macrocytic anemia associated with chronic alcoholic or nonalcoholic liver disease.⁷ Other causes of macrocytic anemia, not completely understood in their pathophysiology, include hypothyroidism and trisomy 21.

Macrocytic anemia is also a characteristic finding in inherited and acquired bone marrow failure syndromes (BMFS) affecting one or more lineages, such as Diamond-Blackfan anemia, Schwachman-Diamond syndrome, Fanconi anemia, and myelodysplastic syndromes (MDS).⁸ Congenital dyserythropoietic anemias (CDA), which are caused by pathogenic variants in a variety of genes directly or indirectly affecting cell division and leading to ineffective erythropoiesis, typically present as macrocytic anemia with suboptimal reticulocytosis.⁹ It is likely that the macrocytosis in BMFS as well as in CDAs is caused by a decreased average number of cell divisions during terminal erythropoiesis, producing fewer but larger erythrocytes.¹⁰

PIII-A had a normal absolute reticulocyte count (ARC) at 52 × 10⁹/L, suboptimal for her level of anemia. Peripheral blood smear revealed mild to moderate anisopoikilocytosis, normochromic macrocytes with occasional hypochromic red cells and notable ovalocytes (Figure S1), with no increase in polychromasia, suggesting inadequate RBC production rather than a purely hemolytic anemia (Figure 1). Vitamin B12, methylmalonic acid, and homocysteine levels in serum, RBC folate level, total bilirubin, haptoglobin, aspartate aminotransferase, copper, and thyroid stimulating hormone levels were normal. OGE was also normal (Figure S1), ruling out HX and OHSt.⁵ Similar CBC findings were noted 2 months later, so a bone marrow evaluation was performed. This showed normal marrow cellularity for age, erythroid hyperplasia with a myeloid:erythroid (M:E) ratio 2:1 (rather than the normal 3:1 ratio) with a mild left shift, that is, increased early erythroid precursors, and signs of dyserythropoiesis including binucleated erythroblasts, occasional cytoplasmic bridges, and karyorrhexis (Figure 1). Ring sideroblasts were not noted. Cytogenetic testing and immunophenotyping by flow cytometry did not show evidence of malignancy or MDS. Based on these findings along with the relatively high ferritin, suboptimal reticulocytosis, and family history of macrocytic anemia with iron overload, a possible diagnosis of CDA was considered, leading to next-generation sequencing (NGS) on a panel of genes associated with HHA/CDA (Table S1), which demonstrated a novel heterozygous variant of unknown significance (VUS) in exon 8 of ALAS2 (c.1066G>A, p.Val356Met) (Figure 2).

Missense pathogenic variants in ALAS2 are well known to cause congenital X-linked sideroblastic anemia (XLSA). These mutations are partial loss-of-function (LOF) variants leading to the non-syndromic form of sideroblastic anemia presenting with hypochromic, microcytic anemia in hemizygous males. Many variants decrease binding to the pyridoxal-5-phosphate (PLP) cofactor of the ALAS2 enzyme (Figure 2), and in many such cases, the disease responds to pyridoxine (vitamin B6) treatment. In XLSA, marrow studies reveal ring sideroblasts, since decreased ALAS2 activity limits protoporphyrin production, which would bind and utilize iron into heme, within the erythroblast mitochondria. Ring sideroblasts are erythroblasts with a minimum of five siderotic granules covering at least one-third of the nuclear circumference; these siderotic granules are iron-laden mitochondria visualized by Prussian blue staining in bone marrow aspirate smears.¹¹ Systemic iron overload with ALAS2 mutations develops independently of RBC transfusions due to ineffective erythropoiesis.^{12, 13} Females with such ALAS2 variants in heterozygous status are typically asymptomatic carriers but occasionally may have anemia characterized typically by dimorphic RBCs and increased RDW, because of unfavorably skewed X-chromosome inactivation.^{14, 15}

Over the last decade, rare cases of macrocytic anemia have been described in a few families, affecting only females, caused by certain ALAS2 missense pathogenic variants that lead to complete LOF of the produced enzyme. These variants appear to be embryonic lethal in hemizygous male offspring. The first such case (ALAS2 p.Tyr365His) was also noted to have dyserythropoiesis¹⁶; therefore, ALAS2 has been included in the HHA/CDA NGS panels.

The ALAS2 p.Val356Met variant has not been previously reported in Human Gene Mutation Database (HGMD), National Center for Biotechnology Information (NCBI), or gnomAD. It has been reported in ClinVar (2438964) as a variant of uncertain significance. In silico prediction programs (PolyPhen-2, SIFT, MutationTaster) suggest a damaging effect. Given P-IIIA's anemia and P-II's anemia with iron overload, follow-up targeted genetic testing was completed for other family members for three generations (Figure 3A). P-I, P-II, and P-IIIB, all female, also have the mutation. P-I and PIII-B have had borderline-low to normal hemoglobin levels (10–13.5 g/dL) and are asymptomatic, while P-II's phenotype is similar to P-IIIA, with evidence of occasional transfusion needs (hemoglobin nadir 7–8 g/dL) and elevated ferritin even before transfusions. A therapeutic trial of pyridoxine (maximum dose of 250 mg) was prescribed for 6 months to P-IIIA but failed to improve the anemia. Notably, no males were affected amongst siblings at any generational level, while there was a strong history of multiple miscarriages in the family. P-I has had four miscarriages and P-II has experienced three miscarriages, all occurring between 6 and 12 weeks of gestation.

This novel ALAS2 pathogenic variant appears to be similar to a short list of ALAS2 variants causing in males embryonal lethality and in females a form of CDA with macrocytosis and marrow dyserythropoiesis, sometimes with ring sideroblasts, especially when associated with significant iron overload, that resembles MDS with ring sideroblasts. This short list, in addition to p.Val356Met, includes p.Tyr365His,¹⁶ p.Arg227Cys,¹⁷ p.Arg163His,¹⁸ and p.Val208Phe.¹⁹ ALAS2 p.Val356Met appears to be detrimental for ALAS2's tertiary structure (Figure 2), causing complete LOF, rather than just decreased affinity to PLP; this corresponds with the failure of vitamin B6 treatment to improve P-IIIA's anemia.²⁰

As in classic microcytic XLSA, phenotypic heterogeneity in females may be due to variances in skewed X-chromosome inactivation, but the resulting RBCs are morphologically different. In XLSA, affected females have dimorphic distribution of erythrocytes: microcytic RBCs, produced by erythroblasts where the X-chromosome with the pathogenic ALAS2 is the activated one, mixed with normal RBCs; in contrast, in ALAS2-associated CDA, patients have uniformly macrocytic RBCs.¹⁶ In P-IIIA and P-II, we found that 90%–92% of the active X-chromosome was the maternally inherited one that contained the mutant ALAS2, as assessed by the human androgen receptor gene polymorphism assay (HUMARA) for X-chromosome inactivation in whole-blood genomic DNA. Family members with milder phenotypes (P-I and P-IIIB) had a lower degree of skewing of X-chromosome inactivation (73% and 69%, respectively) (Figure 3B), supporting the theory that X-chromosome inactivation patterns contribute to phenotypic severity.

P-IIIA has needed three transfusions for fatigue and acute hemoglobin reductions over the 3 years since diagnosis. She has had a slow but steady exacerbation of her anemia during that time frame, with current hemoglobin measurements 7–7.5 g/dL, MCV >110 fL, continued inappropriately low ARC (50–60 10⁹/L), and ferritin slowly rising >250 ng/mL.

Most individuals with congenital sideroblastic anemias are managed conservatively. Bone marrow transplantation (BMT) has been used in severe sideroblastic anemia cases, particularly for SLC25A38 mutations, which often require chronic transfusions and intensive chelation, following a transfusion-dependent-thalassemia-like management. However, BMT for ALAS2 mutations is rarely necessary.²¹ Interestingly, using patient-derived induced pluripotent stem cells (iPSCs) from a patient with ALAS2 p.Arg227Cys, also a complete LOF variant,¹⁷ it was demonstrated that treatment with the DNA demethylating agent azacitidine in vitro reactivated the silent, wild-type ALAS2 allele in the erythroid progenitors and ameliorated the erythroid differentiation defects, suggesting that DNA demethylation may provide a novel therapeutic approach for some patients.²²

The family described here was enrolled in the CDA Registry of North America (CDAR; NCT02964494). CDAR serves as a collaborative platform with the referring hematologists that aims to understand the natural history of this group of diseases, to identify novel genes and gene variants causing CDA, facilitate more rapid and accurate diagnoses, and investigate the molecular defects in CDA, unraveling pathways essential for erythropoiesis.

This case highlights both the complexities and nuances of ALAS2 structure and function along with the importance of X-chromosome skewing in hematologic pathophysiology. Vertebrates have two ALAS genes that encode δ-aminolevulinate synthase (ALAS), ALAS1 and ALAS2, which share about 60% homology; ALAS1 is a housekeeping gene while ALAS2 is erythroid-specific. ALAS2 is the first and rate-limiting enzyme in the heme biosynthesis pathway, being critical for the final goal of hemoglobin formation. It catalyzes the condensation of glycine and succinyl-coenzyme A (succinyl-CoA) to form δ-aminolevulinic acid (ALA) and requires PLP, the active coenzyme form of pyridoxine, for proper function.²³ Over 100 mutations have been described for ALAS2, located at chromosome Xp11.21; both LOF and gain-of-function (GOF) pathogenic variants have been described.^{20, 24}

The GOF variants causing X-linked erythropoietic protoporphyria (XL-EPP) truncate or alter the carboxyterminal end of the ALAS2 protein, which has been shown to play an inhibiting self-regulatory role for the enzyme, while they still leave intact a GOF domain earlier in the C-terminus of ALAS2.²⁵ This leads to the production of an ALAS2 enzyme with increased activity and consequently excessive protoporphyrin accumulation. The LOF pathogenic variants can now be divided to those well-known partial LOF variants causing XLSA with microcytic anemia in hemizygous males and in unfavorably skewed X-chromosome-inactivated heterozygous females and the recently recognized complete LOF variants causing embryonal lethality in hemizygous males and CDA in heterozygous females. Ring sideroblasts may be noted in the marrow especially after the 3rd-4th decade, associated with progressive iron overload. Interestingly, out of the more than 100 variants reported in HGMD as causing XLSA, only the variant p.Tyr365His is reported as causing macrocytic dyserythropoietic anemia.²⁶ The variants p.Arg227Cys, p.Arg163His, and p.Val208Phe are reported in HGMD as causing XLAS erroneously, since these variants have been reported to cause embryonic lethality in males, and a form of CDA with macrocytosis and marrow dyserythropoiesis, sometimes with ringed sideroblasts, in females.^17-19 We urge for the recognition of this second phenotype caused by complete LOF of the ALAS2 enzyme and inability to produce any heme in the mitochondria of the erythroid lineage in male fetuses with hemizygosity for such ALAS2 variants. This mechanism explain the multiple early miscarriages reported by P-I and P-II in this family with the novel p.Val356Met, similar to the previously reported pedigree with complete LOF of ALAS2 because of the p.Val208Phe mutation.¹⁷ Survival advantage of the wild-type ALAS2-containing erythroblasts has been demonstrated in erythropoiesis cultures in vitro starting from female patient-donated erythroid progenitors,¹⁶ as well as by demonstrating that cDNA generated from patients' reticulocyte mRNA was negative for the pathogenic variant.^{16, 17} These findings indicate that circulating macrocytes may be produced only by erythroblasts where the healthy X-chromosome is the active one, thus explaining why the resulting anemia is not microcytic. Instead, macrocytosis is seen due to stress erythropoiesis,¹⁰ with X-chromosome skewed inactivation in hematopoietic cells accounting for familial phenotypic heterogeneity, not unlike other reported cases of ALAS2-associated CDA.¹⁷

Ring sideroblasts were not identified in PIII-A, who had bone marrow studies at 18 years old, though it may have been too early; in contrast, RS have been noted in older patients with ALAS2-CDA and more significant iron overload.¹⁷ Genetic causes for macrocytic anemia caused by CDA or inherited BMFS are less common than acquired nutritional deficiencies or those secondary to medications. Moreover, acquired MDS with clonal hematopoiesis is more likely in middle-age or older adults like P-II and P-I, who would have been in risk of misdiagnosis with MDS with or without ringed sideroblasts if they had been evaluated before P-IIIA. Iron overload is a common comorbidity in macrocytic anemias with stress and especially ineffective, erythropoiesis such as CDAs and MDS, even without concurrent transfusions, since increased erythroferrone suppresses hepcidin.²⁷ Therefore, ferritin, TIBC, transferrin, and iron saturation monitoring are necessary, preferably with further evaluation of tissue iron with appropriate imaging (MRI T2*), if ferritin is persistently >400 ng/mL, in order to proceed to appropriate treatment.

After nutritional deficiencies and HHAs with reticulocytosis have been ruled out in macrocytic anemias, it is reasonable to proceed to bone marrow studies, since BMFS may underlie macrocytosis. Yet even marrow studies may not be adequate, either because CDA-associated dyserythropoiesis may be subtle or because it may be falsely interpreted as acquired MDS. This case of CDA highlights the value of genetic evaluation alongside marrow studies in similar macrocytic anemias and reflects the evolving landscape of understanding ALAS2 pathophysiology that had previously been primarily attributed to XLSA.

Alexander A. Boucher, Vanessa J. Dayton, and Theodosia A. Kalfa were involved in the clinical diagnostic evaluation. Alexander A. Boucher supervised the index patient's clinical care, coordinated genetic testing for all reported patients, and drafted the initial manuscript. Vanessa J. Dayton, Annaliisa R. Pratt, Yasmin Elgammal, and Theodosia A. Kalfa have all been involved in the hematopathology workup and Yasmin Elgammal and Theodosia A. Kalfa conducted the basic science laboratory analyses reported here. Nicolas N. Nassar conducted the protein spatial analyses and figure generation of the ALAS2 mutation and its proposed effects reported here. Theodosia A. Kalfa is the PI for the Congenital Hemolytic and Dyserythropoietic Anemia study and enrolled the involved patients. All authors read, edited, and approved the final manuscript.

This work was supported in part by the National Institutes of Health, National Heart, Lung, and Blood Institute grant R01HL152099, by the Cincinnati Children's Hospital Center for Pediatric Genomics, and by the Leukemia & Lymphoma Society (TRP-6664-23).

The authors have no competing conflicts of interests to declare.

All patients have been consented through the CDAR and are aware of this manuscript. In addition, all patients have signed consent for the genetic testing involved.