Abnormalities of Flavin Monooxygenase as an Etiologyfor Sideroblastic Anemia

Matthew Barber, 1† Marcel E. Conrad, 1* Jay N. Umbreit, 1 James C. Barton, 2 andElizabeth G. Moore 1

1USA Cancer Center, University of South Alabama, Mobile, Alabama2Southern Iron Disorders Center, Birmingham, Alabama

We postulated that a deficiency of flavin monooxygenase (FMO)—a ferrireductase com-ponent of cells—could produce sideroblastic anemia. FMO is an intracellular ferrireduc-tase which may be responsible for the obligatory reduction of ferric to ferrous iron so thatreduced iron can be incorporated into heme by ferrochelatase. Abnormalities of thismechanism could result in accumulation of excess ferric iron in mitochondria of ery-throid cells to produce ringed sideroblasts and impair hemoglobin synthesis. To inves-tigate this hypothesis we obtained blood from patients with sideroblastic anemia andnormal subjects. Extracts of peripheral blood lymphocytes were used to measure ferrire-duction by utilization of NADPH. Lymphoid precursors are reported to accumulate iron inmitochondria similarly to erythroid precursors. Utilization of lymphoid precursorsavoided the need for bone marrow aspirations. We studied three patients with sidero-blastic anemia. One patient and his asymptomatic daughter had a significant decrease inferrireductase activity. They also had markedly diminished concentrations of FMO inlymphocyte protein extracts on Western blots. This was accompanied by increased con-centration of mobilferrin in the extracts. These results suggest that abnormalities of FMOand mobilferrin may cause sideroblastic anemia and erythropoietic hemochromatosis insome patients. Am. J. Hematol. 65:149–153, 2000. © 2000 Wiley-Liss, Inc.

Key words: paraferritin; NADPH; protoporphyrin; calreticulin; mobilferrin; lymphocytes;iron

INTRODUCTION

The sideroblastic anemias are a spectrum of disorderscaused by defects in the generation of heme in develop-ing erythroid cells [1,2]. They occur as a result of eitherdecreased production of protoporphyrin or impaired in-corporation of iron into protoporphyrin [3–13]. The clini-cal hallmark of the disorders is demonstration of “ringedsideroblasts” in erythroid precursors in bone marrow as-pirates; this is due to accumulation of ferric iron in themitochondria of these cells. The defect may be eitherhereditary or acquired, and patients may remain asymp-tomatic for many years with gradual worsening of ane-mia as the patient ages. Diverse mutations in the ery-throid 5-aminolevulinate synthase gene appear to beresponsible for about half of patients diagnosed with sid-eroblastic anemia [1,2]. However, ringed sideroblasts canoccur due to defects at other portions of the heme syn-thesis pathway [13,14]. Presently about half of patientswith irreversible sideroblastic anemia who were studiedhave an identified defect in heme synthesis [1].

Ferrous iron is required for incorporation into proto-porphyrin IX by ferrochelatase to synthesize heme [15].Failure to reduce iron could result in accumulation offerric iron within mitochondria. Flavin monooxygenase(FMO) has been isolated from cells as a component of alarge protein complex that was shown to act as a ferrire-ductase [16]. Thus, it was postulated that either defi-

Contract grant sponsor: National Institute of Diabetes and Digestiveand Kidney Diseases of the National Institutes of Health; Contractgrant number: Merit Award R37 DK36112; Contract grant sponsor:Southern Iron Disorders Center.

*Correspondence to: Marcel E. Conrad, M.D., USA Cancer Center,University of South Alabama, Mobile Alabama 36688.

†Matthew Barber performed the studies reported in this article whilehe was supported as a recipient of a 1999 ASH Medical StudentAward.

Received for publication 3 December 1999; Accepted 5 April 2000

American Journal of Hematology 65:149–153 (2000)

© 2000 Wiley-Liss, Inc.

ciency or abnormalities of FMO could result in sidero-blastic anemia.

Isolated lymphocytes were used in this study becausethey are more readily available from patients than ery-throid precursors. Lymphocytes can be immortalized forbasic studies, and Prussian blue-reacting ferruginousgranules have been identified in the mitochondria of lym-phoid cells in sideroblastic anemia, demonstrating phe-notypic expression of the genetic defect (17).

METHODS

We studied three unrelated patients with sideroblasticanemia. These included (1) an elderly man with transfu-sion dependent idiopathic sideroblastic anemia and hisasymptomatic daughter; (2) a patient with lymphomawho had received chemotherapy, (3) a woman with se-vere X-linked sideroblastic anemia due to a 5-aminolevu-linate synthase gene mutation (exon 10: C1601T, Arg517 Cys) and skewed X-inactivation; two of her threedaughters also inherited this mutation but had mild ane-mia. All patients had ringed sideroblasts in bone marrowaspirates. Preliminary studies revealed that none of ourpatients with sideroblastic anemia inherited either theC282Y or H63D mutations of the HFE gene associatedwith most classical hemochromatosis in Caucasians [18].

Forty milliliters of peripheral blood was obtained fromeach subject and from normal volunteers. It was antico-agulated with 100 USP units of heparin (Abbots Labo-ratories, North Chicago, IL). Portions of 20 ml were lay-ered on 15 ml of Histopaque (Sigma, St. Louis, MO) andcentrifuged for 30 min (400 × G) at room temperature.Aliquots (5 ml) of lymphocyte–Histopaque mixture wererecovered and placed in 50-ml plastic test tubes. Fortymilliliters of calcium-magnesium free phosphate-buffered saline (PBS) was added to each tube, and thetubes were centrifuged (400 × G) for 15 min at roomtemperature. Cell pellets from each tube were collectedinto a single test tube containing 0.675 ml of cold hypo-tonic buffer (10 mM HEPES, 10 mM NaC1, pH 7.4,4°C). This method yields 85% lymphocytes at the inter-face, and more than 95% of these cells excluded TrypanBlue. The buffered lymphocytes were homogenized withthree strokes on a Potter-Elvehjem grinder with a Teflonpestle (Fisher Scientific Co., Pittsburgh, PA). TritonX-100 was added (150ml, 10%) and incubated at 4°C for15 min. The mixture was centrifuged at 12,800 × G for30 min at 4°C. The supernatant was recovered for mea-surement of protein concentration (BCA, Pierce, Rock-ford, IL) and enzyme assay. Specimens were diluted tobring each tube to the same total protein concentration.

The rate of iron-dependent NADPH utilization wasdetermined using a Gilford spectrophotometer (ResponseII, Oberlin, OH). The reactive mixture contained 200mlof 0.5mM HEPES, 5ml of FeNTA (25 mM Fe, 100 mM

NTA, pH 7.0), and 5ml NADPH (32 mM) with varyingquantities of crude enzyme preparation for the lympho-cytes. Controls included the reaction mixture without thecrude lymphocyte extract and the reaction mixture with-out FeNTA (iron-nitrilotriacetic acid).

Aliquots of the Triton-solubilized protein were dilutedto bring them to equal protein concentrations. Then theywere electrophoresed on 7.5% SDS-PAGE gels (Protean2 vertical electrophoresis apparatus (Bio-Rad, Rich-mond, CA)). The gel was transblottted to Immobilon P(Millipore, Medford, MA) on a Bio-Rad transblot appa-ratus (10 volts, 0.1 amps) overnight with cooling and asodium carbonate (3 mM)–sodium bicarbonate (10 mM)buffer (pH 9.9).

Polyclonal antibodies to mobilferrin (calreticulin)were obtained from Affinity Bioreagents (Golden, CO).Antibodies to FMO-II were raised against a syntheticpeptide sequence of rabbit lung FMO (amino acids 309–318) [(H-KVRRVKELT 4)8-MAP] [19]. Antibodieswere purified on Immunopure Immobilized Protein A/Gbeads (Pierce, Rockford, IL). A goat anti-rabbit antibodywas used as a secondary antibody (HPR, Amersham,Piscataway, NJ). Antibodies were diluted 1:1,000 foruse. Western blots were developed by enhanced chemi-luminescence according to the manufacturers instruc-tions (Amersham). Primary antibodies showed a singleband on Western blots of whole lymphocyte extracts atan appropriate apparent molecular mass (∼56 kDa). Inaddition, each primary antibody reacted with the purifiedproteins and not with other purified proteins, and thisreaction was blocked by incubation of the FMO peptidewith the antibody before the mixture was incubated witha Western blot containing FMO.

RESULTS

Three patients with sideroblastic anemia were studiedto ascertain whether they had abnormalities of FMO as apotential etiology of their hematological disorder. One ofthe patients had a significant decrease in ferrireductaseactivity in extracts of peripheral lymphocytes (Table I).Further investigation showed a significant diminution ofimmunologically detectible FMO on Western blots (Fig.

TABLE I. Ferrireductase Activity in Sideroblastic Anemia

Unitsa

Control subjects (7) 7.2 ± 4.5Sideroblastic anemia(1) FMO decreased 1.1 ± 0.8b

Daughter: FMO decreased 0.9 ± 0.7b

(2) FMO unchanged (lymphoma) 9.4 ± 4.3(3) FMO pattern changed (X-linked) 28.7 ± 8.5b

aOD/mg − min 10−5 ± 1 SD.bP # 0.01.

150 Barber et al.

1). This was accompanied by an increase in the concen-tration of immunologically detectable mobilferrin in cellextracts (Fig. 2). His daughter had similar abnormalitiesof ferrireductase activity and FMO (Fig. 3, Table I).However, she was asymptomatic and not anemic. Thepatient had been asymptomatic until age 89 years, whenhe developed transfusion dependent anemia. The anemiawas unresponsive to pyridoxine therapy but the patienthad a decrease in his transfusion requirement followingadministration of erythropoietin. We can only speculatewhether the daughter will develop anemia as she ages,similarly to many other patients with hereditary forms ofsideroblastic anemia [1,2].

The patient with a characterized X-linked hereditarysideroblastic anemia had increased ferrireductase activityand an unusual FMO isoform pattern on Western blots oflymphocyte extracts. Whether this is a reactive secondaryeffect cannot be determined by these studies.

DISCUSSION

The sideroblastic anemias have been classified intohereditary sideroblastic anemia, acquired sideroblasticanemia, idiopathic sideroblastic anemia, and sideroblas-tic anemia associated with myelodysplastic syndromes,myelofibrosis, certain drugs, and excessive ethanol in-gestion [1,2,20,21]. Only the latter two forms are revers-ible. The irreversible forms of sideroblastic anemia (he-

reditary and acquired sideroblastic anemia) are iron-overloading disorders, and patients may develop thephenotypic clinical manifestations of familial hereditaryhemochromatosis. Our patients did not inherit the HFEmutations observed in most Caucasian patients with clas-sical hemochromatosis, and these mutations do not ap-pear with increased frequency among patients with sid-eroblastic anemia [19,20].

During recent years, investigators undertook intensivestudies of 5-aminolevulinate synthase in patients withsideroblastic anemia. Anemia was improved in some pa-tients by administration of pyridoxine because pyridoxal58-phosphatase is a cofactor of the enzyme and is re-quired for heme synthesis. Many abnormalities of 5-ami-nolevulate synthase have been reported in different pro-bands with sideroblastic anemia. Abnormalitiesinvolving the initial step of heme synthesis appear to beresponsible for about half of the irreversible forms ofsideroblastic anemia and most of the patients with X-linked sideroblastic anemia. A few patients are reportedwith abnormalities of ferrochelatase, the last step ofheme synthesis [13], and a defect in mitochondrial cyto-chrome C oxidase [14]. The X-linked sideroblastic ane-mias associated with ataxia appear to be due to a muta-tion of the mitochondrial iron transporter gene (ABC7)[6]. Thus, enzymatic abnormalities associated with sid-eroblastic anemia appear to affect the first and last stepsof heme synthesis which occur within the mitochondriawhere ferric iron accumulates to produce ringed sidero-blasts.

Fig. 1. Chemiluminescence of a Western blot using a pri-mary polyclonal antibody against flavin monooxygenase(FMO) in extracts of lymphocytes. There was a marked de-crease in FMO in extracts from a patient with sideroblasticanemia and iron overload (lane 1) by comparison to extractsfrom a normal volunteer (C) and a patient with sideroblasticanemia after treatment for lymphoma (lane 2) and X-linkedhereditary sideroblastic anemia (lane 3). The latter speci-men appears to have an additional band.

Fig. 2. Chemiluminescence of a Western blot using a pu-rified polyclonal antibody against mobilferrin (calreticulin).Specimens were electrophoresed on SDS-PAGE gels inpairs with specimens to the left containing one half the pro-tein concentration of the specimen to the right. The patientwith diminished FMO in Figure 1 had an increased concen-tration of mobilferrin in lymphocyte extracts (lane 1).

FMO: An Etiology of Sideroblastic Anemia 151

Transferrin bound iron enters the cell as ferric iron andthis iron must be reduced for utilization by ferrochelatasefor heme synthesis [15]. Thus impaired intracellular fer-rireduction would limit heme synthesis and could lead tomitochondrial accumulation of ferric iron. Pulse chaseexperiments showed that transferrin iron transits mobil-ferrin before it can be utilized for heme synthesis withincells and that mobilferrin is associated with a large pro-tein complex named paraferritin which includes flavinmonooxygenase and the complex produces ferrireduction[16,22].

Among three unrelated patients with sideroblastic ane-mia, we identified a patient with idiopathic sideroblasticanemia who had significantly diminished ferrireductaseactivity and decreased concentrations of FMO in lym-phocyte extracts (Table II). Similar findings were docu-mented in his daughter, suggesting that this is a geneticabnormality. While she was not anemic, delayed pheno-

typic expression of hereditary sideroblastic anemia iscommonplace [1,2]. These observations suggest thatsome patients with sideroblastic anemias have a defi-ciency of FMO. Our other patients with sideroblasticanemia had either normal or increased ferrireductase ac-tivity. Observations in a large number of patients withsideroblastic anemia would be necessary to determine theprevalence of FMO abnormalities as an etiology for sid-eroblastic anemia. Our patient studied here appeared tohave a quantitative decrease in FMO. Mutations of FMOcould produce the same functional effect. Whether thealterations in concentration of mobilferrin observed inthese patients are important in the etiology of the disorderor merely secondary effects cannot be determined in thisstudy. However, up-regulation of mobilferrin could beresponsible for the increased absorption of iron and ironoverload observed in some patients with sideroblasticanemia.

REFERENCES

1. Bottomley SS. Secondary iron overload disorders. Semin Hematol1998;35:77–86.

2. Bottomley SS. Sideroblastic anemias. In: Lee GR, et al., editors. Win-trobe’s clinical hematology. 10th edition. Baltimore: Williams &Wilkins; 1998. p 1022–1045.

3. Bottomley SS, May BK, Cox TC, Cotter PD, Bishop DF. Moleculardefects of erythroid 5-aminolevulinate synthase in X-linked sidero-blastic anemia. J Bioenerg Biomembr 1995;27:161–168.

4. May A, Bishop DF. The molecular biology and pyridoxine respon-siveness of X-linked sideroblastic anemia. Haematologica 1998;83:56–70.

5. Cotter PD, May A, Li L, Al-Sabah AI, Fitzsimmons EJ, Cazzola M,Bishop DF. Four new mutations in the erythroid specific 5-aminolevu-linate synthase (ALAS2) gene causing X-linked sideroblastic anemia.Blood 1999;93:1757–1769.

6. Allikmets R, Raskind WH, Hutchinson A, Schueck ND, Dean M,Koeller DM. Mutation of a putative mitochondrial iron transportergene (ABC7) in X-linked sideroblastic anemia and ataxia (XLS/A).Hum Mol Genet 1999;8:743–749.

7. Gattermann N, Retzlaff S, Wang Y-L, Hofhaus G, Heinisch J, Aul C,Schneider W. Heteroplasmic point mutations of acquired idiopathicsideroblastic anemia. Blood 1997;90:4961–4972.

8. Cotter PD, Baumann M, Bishop DF. Enzymatic defect in ‘X-linked’sideroblastic anemia: molecular evidence for erythroidd-aminolevu-linate synthase deficiency. Proc Natl Acad Sci USA 1992;89:4028–4032.

9. Cox TC, Bottomley SS, Wiley JS, Bawden J, Matthews CS, May BK.X-linked pyridoxine response sideroblastic anemia due to a Thr388-to-

Fig. 3. Lymphocyte extracts from the patient (P) shown inFigure 2 and Table I to have diminished concentrations ofFMO and increased concentrations of mobilferrin and speci-mens from his daughter (D) when examined using chemilu-minescence on Western blot for FMO. Little activity wasobserved by comparison to lymphocyte extracts from twonormal subjects (C).

TABLE II. Sideroblastic Anemias: Concentration of IronBinding Protein in Lymphocytes*

Ferrireductaseactivity FMO Mobilferrin

(1) Propositus ↓ ↓ ↑(2) Chemotherapy associated N N N(3) X-linked ↑ ↑ N

*Western blots of cellular extracts.

152 Barber et al.

Ser substitution in erythroid 5-aminolevulinate synthase. New Engl JMed 1994;330:675–679.

10. Cotter PD, Rucknagel DL, Bishop DF. X-linked sideroblastic anemia:identification of the mutation in the erythroid-specificd-aminolevuli-nate synthase gene (ALAS2) in the original family described byCooley. Blood 1994;84:3915–3924.

11. Prades E, Chambon C, Dailey TA, Dailey HA, Briere J, GrandchampB. A new mutation of the ALAS2 gene in a large family with X-linkedsideroblastic anemia. Hum Genet 1995;95:424–428.

12. Harigae H, Furuyama K, Kudo, K, Hayashi N, Yamamoto M, Sassa S,Sasaki T. A novel mutation of the erythroid-specificd-aminolevuli-nate synthase gene in a patient with non-inherited pyridoxine-responsive anemia. Am J Hematol 1999;62:112–114.

13. Rademakers LHPM, Koningsberger JC, Sorber CWJ, Baart de la FailleH, Van Hattum J, Marx JJ. Accumulation of iron in erythroblasts ofpatients with erythropoietic protoporphyria. Eur J Clin Invest 1993;23:130–138.

14. Broker S, Meunier B, Rich P, Gattermann N, Hofhaus G. MtDNAmutations associated with sideroblastic anaemia cause a defect of mi-tochondrial cytochrome C oxidase. Eur J Biochem 1998;258:132–138.

15. Porra RJ, Jones OTG. Studies on ferrochelatase I. Assay and propertiesof ferrochelatase from a pig liver mitochondrial extract. Biochem J1963;87:181–187.

16. Umbreit JN, Conrad ME, Moore EG, Desai MP, Turrens J. Paraferri-

tin: a protein complex with ferrireductase activity is associated withiron absorption in rats. Biochemistry 1996;35:6460–6469.

17. Barton JC, Conrad ME, Parmley RT. Acute lymphoblastic leukemia inidiopathic refractory sideroblastic anemia: evidence for a commonlymphoid and myeloid progenitor cell. Am J Hematol 1980;9:109–115.

18. Benacerraf B, Unanue ER. Textbook of immunology. Baltimore:Williams & Wilkins; 1986. p 253–255.

19. Barton JC, Shih WWH, Sawada-Hirai R, Acton RT, Harmon L, RiversC, Rothenberg BE. Genetic and clinical description of hemochroma-tosis probands and heterozygotes: evidence that multiple genes linkedto the major histocompatibility complex are responsible for hemochro-matosis. Blood Cells Mol Dis 1997;23:135–145.

20. Bottomley SS, Wise PD, Wasson EG, Barton JC. The hemochroma-tosis of sideroblastic anemias. World Congress on Iron Metabolism,Sorrento, Italy, May 1999.

21. Shibata K, Shimamoto Y, Nakazato S, Matsuzaki M, Tadano J. Re-fractory anaemia with ringed sideroblasts concurrent with multiplemyeloma: a brief review of the recent literature. Haematologia(Budapest) 1997;28:199–205.

22. Conrad ME, Umbreit JN, Moore EJ, Heiman D. Mobilferrin is anintermediate in iron transport between transferrin and hemoglobin inK562 cells. J Clin Invest 1996;98:1449–1454.

FMO: An Etiology of Sideroblastic Anemia 153