High ferritin and low glycosylated ferritin may also be a marker of excessive macrophage activation.

OBJECTIVE: A high serum ferritin concentration with a low percentage of glycosylated ferritin (< 20%) have been reported to be a specific marker of active adult Still's disease (ASD). However, high ferritin levels are found during hemophagocytosis syndrome (HS). We investigated the ferritin level and the percentage of glycosylation in a HS series of various causes. METHODS: Diagnosis of HS was confirmed by erythrophagocytosis pictures on a bone marrow cytology or biopsy in all patients. Serum ferritin concentration was determined on a heterogenous immunoassay module. Glycosylated ferritin was separated using concanavalin A (Con-A) sepharose 4B chromatography. The nonglycosylated ferritin unbound to Con-A was recovered in the supernatant and quantified with the same procedure. Percentages of glycosylated ferritin less than 20% are considered to be usual in ASD, between 20 and 40% usual in inflammatory syndrome, and between 50 and 80% normal. RESULTS: In all cases tested during the acute phase of the disease, ferritin blood level was high and the percentage of glycosylated ferritin was low, less than 20%. CONCLUSION: The combination of high ferritin level and low percentage of glycosylation may be a marker of excessive macrophage activation.

Essential role of ferritin Pfr in Helicobacter pylori iron metabolism and gastric colonization.

The reactivity of the essential element iron necessitates a concerted expression of ferritins, which mediate iron storage in a nonreactive state. Here we have further established the role of the Helicobacter pylori ferritin Pfr in iron metabolism and gastric colonization. Iron stored in Pfr enabled H. pylori to multiply under severe iron starvation and protected the bacteria from acid-amplified iron toxicity, as inactivation of the pfr gene restricted growth of H. pylori under these conditions. The lowered total iron content in the pfr mutant, which is probably caused by decreased iron uptake rates, was also reflected by an increased resistance to superoxide stress. Iron induction of Pfr synthesis was clearly diminished in an H. pylori feoB mutant, which lacked high-affinity ferrous iron transport, confirming that Pfr expression is mediated by changes in the cytoplasmic iron pool and not by extracellular iron. This is well in agreement with the recent discovery that iron induces Pfr synthesis by abolishing Fur-mediated repression of pfr transcription, which was further confirmed here by the observation that iron inhibited the in vitro binding of recombinant H. pylori Fur to the pfr promoter region. The functions of H. pylori Pfr in iron metabolism are essential for survival in the gastric mucosa, as the pfr mutant was unable to colonize in a Mongolian gerbil-based animal model. In summary, the pfr phenotypes observed give new insights into prokaryotic ferritin functions and indicate that iron storage and homeostasis are of extraordinary importance for H. pylori to survive in its hostile natural environment.

A possible relation of the Helicobacter pylori pfr gene to iron deficiency anemia?

BACKGROUND: H. pylori infection is thought to contribute to iron-deficiency anemia, especially during puberty. The ferritin protein Pfr of H. pylori is homologous to eukaryotic and prokaryotic ferritins. The purpose of this study was to analyze the H. pylori pfr status in gastric biopsy specimens according to clinical data, including antral gastritis with or without iron-deficiency anemia. METHODS: A total of 26 H. pylori-positive patients aged from 10-18 years were categorized into subgroups based on the presence or absence of iron-deficiency anemia. All of them had antral gastritis. Sixteen patients were proved to have iron-deficiency anemia by hematological study, two of which had a duodenal ulcer. The other 10 patients showed normal hematological findings. DNA isolation was performed from each of the gastric biopsy specimens. PCR amplification of the pfr gene coding was done using two sets of primers. The pfr region, 501 bp, was generated by linking the sequences of the two PCR products. The nucleotide and protein sequences were compared between the pfr regions from Korean H. pylori strains and the NCTC 11638 strain, which was obtained from the Genbank. Sequence comparisons were also performed for the pfr regions between the iron-deficiency anemia (+) and (-) groups. RESULTS: Analysis of the complete coding region of the pfr gene revealed three sites of mutation. The Ser39Ala mutation was found in 100% (26/26), Gly111Asn in 26.9% (7/26), and Gly82Ser in 11.5% (3/26). There were no significant differences in the mutations of the pfr regions between the iron deficiency anemia (+) and (-) groups. CONCLUSION: The mutation in the pfr gene did not relate with the clinical phenotype, iron deficiency anemia. Further studies are needed on the aspects of host side or other complex factors to elucidate the mechanisms by which the H. pylori infection might lead to iron deficiency anemia.

Diagnostic value of ferritin and glycosylated ferritin in adult onset Still's disease.

OBJECTIVE: To determine the usefulness of serum ferritin and glycosylated ferritin (GF) levels in diagnosing adult onset Still's disease (AOSD). METHODS: We performed a retrospective multicenter study of 205 patients who had ferritin and GF assays in one hospital laboratory. Records of all patients were reviewed, and a standardized questionnaire used to extract all data available at the time of the assay. The clinicians' final diagnosis was also recorded. Patients were classified as having "certain AOSD" (based on Yamaguchi's criteria) or a control disease. The concordance of ferritin and GF levels with final diagnosis was evaluated. RESULTS: In total 49 AOSD and 120 control patients were eligible. The mean ferritin value was significantly higher in the AOSD group (4,752 +/- 9,599 microg/l) than in the control group (1,571 +/- 3,807 microg/l), p = 0.029. GF was significantly lower in AOSD patients (15.9 +/- 11.9%) than in the control group (31.5 +/- 18.7%), p < 0.001. The combination of a GF level of < or = 20% with ferritin above the upper limit of normal yielded a sensitivity of 70.5% and specificity of 83.2%. The combination of a GF level < or = 20% with ferritin 5 times normal produced a sensitivity of 43.2% and specificity of 92.9%. This latter combination allowed an AOSD diagnosis to be ruled out for 6 of the 8 control patients who met Yamaguchi's positive criteria. CONCLUSION: Ferritin and GF levels are powerful diagnostic markers of AOSD. They may be helpful in clinical practice for excluding differential diagnoses.

Characterization of two genes encoding ferritin-like protein in Actinobacillus actinomycetemcomitans.

Two genes encoding ferritin-like protein, designated afnA and afnB, were identified in the upstream region of actX on the Actinobacillus actinomycetemcomitans chromosomal DNA. The actX has been reported to be a regulatory gene homologous to the Escherichia coli fnr, which controls the growth and virulence of A. actinomycetemcomitans under anaerobic conditions. The afnB located 340 bp-upstream from the actX, and the afnA located just 15 bp-upstream from afnB. The afnA and afnB encoded 161 and 165 amino acid residues, respectively, which were similar to ferritin-like proteins of other microorganisms. Western immunoblotting using rabbit antiserum against E. coli ferritin showed these two proteins, which are reactive with the serum with 19-kDa molecular masses, are produced from A. actinomycetemcomitans. The N-terminal amino acid sequences of the two proteins were consequent with those deduced from afnA and afnB. Northern hybridization revealed that the afnA and afnB constituted a bicistronic operon and the accumulation of afnA and afnB mRNA was upregulated under aerobic conditions. These findings suggested that the operon was regulated by the presence of oxygen. The two ferritin-like proteins may have important roles in the adaptation of A. actinomycetemcomitans to oxidative environmental changes.

[High levels of serum ferritin in elderly patients with non-insulin-dependent diabetes mellitus].

We studied concentrations of serum ferritin, glycosylated ferritin, and non-glycosylated ferritin in elderly patients with diabetes. The subjects were 111 people who were at least 60 years old: 54 healthy controls, 14 diabetic patients without retinopathy, and 43 diabetic patients with retinopathy. The mean levels of ferritin, glycosylated ferritin, and non-glycosylated ferritin in serum were significantly higher in the patients with retinopathy than in healthy controls. The mean percent glycosylated ferritin did not differ between patients with retinopathy and healthy controls. The mean levels of serum ferritin, glycosylated ferritin, and non-glycosylated ferritin, and the percent glycosylated ferritin did not differ significantly between patients without retinopathy and health controls. None of these values differed between subjects with macroangiopathy and those without macroangiopathy, in both groups of patients. In patients with diabetes, none of the values measured was significantly related to fasting plasma glucose, HbA1c, or the duration of diabetes. These results suggest that diabetic microangiopathy is associated with abnormally high levels of ferritin in serum.

[Age-related changes in concentrations of ferritin, glyeosylated ferritin, and non-glycosylated ferritin].

We studied age-related changes in the concentrations in serum of ferritin, glycosylated ferritin, and non-glycosylated ferritin. The concentrations were determined in 95 healthy subjects: 39 men and 56 women, aged from 22 to 94 years. In the men, age correlated significantly with serum ferritin (r = 0.332, p < 0.05) and non-glycosylated serum ferritin (r = 0.628, p < 0.001) but not with glycosylated serum ferritin. In the women, age correlated significantly with serum ferritin (r = 0.456, p < 0.001), non-glycosylated serum ferritin (r = 0.439, p < 0.001), and glycosylated serum serum ferritin (r = 0.415, p < 0.01). The ratio of glycosylated serum ferritin to serum ferritin correlated negatively with age both in men and in women (men: r = -0.661, p < 0.001; women: r = -0.411, p < 0.01). Serum non-glycosylated ferritin levels were higher in older men. Both serum glycosylated ferritin and non-glycosylated ferritin levels were higher in older women, but this phenomenon was more pronounced with respect to the non-glycosylated form. These results suggest that hyperferritinemia in the elderly is mainly caused by an increase in the concentration of non-glycosylated ferritin, both in men and in women.

Biochemical analysis of ferritin subunits in sera from adult Still's disease patients.

To determine the origin of the increased serum ferritin that occurs in adult Still's disease (ASD), we analyzed subunits of the serum ferritin as follows. Gel filtration with Sepharose CL-6B demonstrated that the molecular weight of serum ferritin was about 490 kDa. Western blot analysis revealed only L-subunits (molecular weight 19 kDa) in patients with serum ferritin levels higher than 1,000 ng/ml. Patients with serum ferritin levels higher than 1,000 ng/ml, however, showed G-subunits (molecular weight 23 kDa) in addition to the L-subunits. When concanavalin A (Con-A) Sepharose 4B was used in an absorption test, the percentage absorption was extremely low in the patients with serum ferritin levels higher than 1,000 ng/ml. Isoferritin patterns of the patients determined by chromatofocusing revealed traces of acidic ferritin. The findings suggested that glycosylated ferritin does not account for the major portion of the increased serum ferritin.

A capillary electrophoresis method for studying apo, holo, recombinant, and subunit dissociated ferritins.

A capillary electrophoresis (CE) method is described for detecting and quantitating apo and holo ferritins from horse spleen (HoSF), rat liver (RLF), recombinant human light chain (rLF), recombinant human heavy chain (rHF), site-directed variants of human light chain, and Azotobacter vinelandii bacterial ferritin (AVBF). This procedure is carried out at pH 8.2, where the ferritin molecules are associated into their 24-mers. Protein mobilities as expressed as elution times were clearly resolved and could be used to distinguish one ferritin type from another, providing a means for detecting and quantitating various ferritin species in purified or partially purified states. Measurements of these and other ferritins were also conducted at pH 2.0, where dissociation into their respective subunits occurs. For HoSF and RLF, the individual L and H subunits were resolved and their relative concentrations were determined by integrating the areas of the elution peaks. HoSF gave 89.8% L and 10.2% H and RLF gave 70.7% L and 29.3% H, while rLF, rHF, and AVBF gave only a single subunit, all in agreement with reported values obtained by polyacrylamide gel electrophoresis. CE of HoSF, containing increasing amounts of iron in the interior, in general, showed that protein mobilities increased, reached a plateau, and then slowly decreased with increasing core size, although buffer effects altered this CE behavior to some extent. Such results indicate that species formed early during core formation have individual iron atoms present and differ from those formed later in which the oligomeric iron core has formed.(ABSTRACT TRUNCATED AT 250 WORDS)

Purification of a novel glycosylated ferritin from horse heart.

We have previously shown that mRNA coding for ferritin L subunit is present on both cytosolic ribosomes and endoplasmic reticulum-bound ribosomes in rat heart tissue [Campbell et al. (1989) Arch Biochem Biophys 273:89-98]; from this we infer that heart tissue is capable of making a secreted ferritin. We now report the purification from horse heart, of a ferritin that specifically binds to Concanavalin A-Sepharose and is immunologically cross-reactive with antibodies raised against both horse cellular ferritin and horse serum ferritin. Where cellular ferritin is 10 nm in diameter and contains primarily 21-kDa subunits (as determined by gel exclusion chromatography and electron microscopy), the glycosylated heart ferritin is smaller with diameters of 3-5 nm. Antisera raised against serum ferritin cross-reacted with the glycosylated heart ferritin did but did not show significant cross-reactivity with cellular ferritin thus indicating that serum ferritin and glycosylated heart ferritin have antigenic determinants which may not be present on cellular ferritin. The glycosylated ferritin also differs from cellular ferritin in subunit composition, with subunits of 66, 60.5, 53.5, 43.5, and 29.5 kDa, as shown by SDS-PAGE and Western blot analysis. Interestingly, ferritin purified from horse serum contains subunits of similar size.

Glycosylated serum ferritin in patients with hematological malignancies before and after bone marrow transplantation.

Glycosylated and total serum ferritin levels were monitored in patients with acute leukemia and lymphoma undergoing bone marrow transplantation (BMT). Serum ferritin was high in relapsing patients and normal in most patients in complete remission (CR). In patients with an uncomplicated course, levels of ferritin increased during the first month after BMT with subsequent decrease. Three patients with lymphoma and five with acute leukemia had high serum ferritin levels despite achieving apparent complete hematological remission which was of short duration. The results were compared with groups of lymphoma patients at presentation and during remission and with healthy normal controls. In all the lymphoma patients and in 3 of the 5 leukemia patients the percent of ferritin glycosylation was normal at CR. It was low at the time of diagnosis in all patients. Thus, the percent glycosylation proved a more reliable marker for clinical remission than total serum ferritin. During follow up after BMT in uncomplicated cases, the percent of glycosylated ferritin returned to normal levels earlier than the total serum ferritin. These findings indicate that the evaluation of the amount of glycosylated ferritin may provide useful information in hematological patients in whom there is a discrepancy between high serum ferritin levels and the clinical condition.

[The serum and intraerythrocyte concentration of ferritin in the anemia of rheumatoid arthritis]. / Concentración sérica e intraeritrocitaria de ferritina en la anemia de la artritis reumatoide.

Serum concentration of ferritin (Ft) and its glycosylated fraction (Ft-Gl) and intraerythrocytic ferritin (Ft-e) concentration were measured in 26 patients with anemia and active rheumatoid arthritis. Patients were divided into 2 groups according to the presence of anemia of chronic diseases (n = 13) or associated ferropenia. Unlike the first group, patients with associated ferropenia had lower concentration of the above parameters than 31 control subjects. The logarithmic value of FT (log FT) directly correlated with globular sedimentation velocity. Ft-Gl and log Ft-e correlated with transferrin saturation (r = 0.603, p less than 0.01 and r = 0.444, p less than 0.05). Log Ft-e also correlated with Ft (r = 0.504, p less than 0.01). The probability of ferropenia when Ft was 60 micrograms/l or lower was 0.91, and when Ft-e was 1.5 ag/cel or lower was 0.66. It is concluded that the ferropenic status in active rheumatoid anemia decreases the iron dependent synthesis of ferritin (Ft-Gl) more than that mediated by the acute phase response. The intraerythrocytic content is low due to the scanty iron supply to the erythroblast. Ft is more efficacious than Ft-e in the diagnosis of ferropenia.

Neonatal hemochromatosis. Genetic analysis of transferrin-receptor, H-apoferritin, and L-apoferritin loci and of the human leukocyte antigen class I region.

Neonatal hemochromatosis (NH), a generally fatal disorder of infancy, is characterized by severe hepatic insufficiency of intrauterine onset and by marked organ iron loading. Its cause is unknown. It has been suggested that NH may represent an unusual manifestation of hereditary hemochromatosis (HH), which is human leukocyte antigen (HLA) linked. Evidence for major rearrangements or deletions at the HLA class I region and at three loci directly involved in iron metabolism (H- and L-apoferritin and the transferrin receptor [TfR]) was sought. The population studied included five probands with NH and 14 first-degree family members in a total of six kindreds. Also sought were HLA associations with NH by collating the results of HLA serotyping in these 19 persons and in 17 members of 7 additional kindreds in which NH has occurred, including 5 probands with NH and 12 first-degree family members. We found no evidence for major rearrangements or deletions in H- or L-apoferritin genes, in TfR genes, or within the HLA locus. We found no evidence for linkage of NH to HLA serotypes. We conclude that while NH and HH are similar in their patterns of iron loading, they are not genetically related.

Selective adsorption of apoferritin on immobilized Fe(III): demonstration of Fe(III) binding sites.

Immobilized metal ion affinity chromatography has been used to demonstrate and partially characterize Fe(III) binding sites on apoferritin. Binding of Fe(III) to these sites is influenced by pH, but not affected by high ionic strength. These results suggest that both ionic and coordinate covalent interactions are important in the formation of the Fe(III): apoferritin complex. This is, to our knowledge, the first demonstration of direct Fe(III) binding to apoferritin. Other immobilized metal ions, including Zn(II), Ni(II), Cu(II), Cr(III), Co(II), and Tb(III), displayed little or no adsorption of apoferritin. The analytical technique of immobilized metal ion affinity chromatography also shows great promise in the purification of apoferritin, ferritin, and other iron-binding proteins.

Fe2+ binding to apo and holo mammalian ferritin.

The binding of Fe2+ to both apo and holo mammalian ferritin has been investigated under anaerobic conditions as a function of pH. In the pH range 6.0-7.5, 8.0 +/- 0.5 Fe2+ ions bind to each apoferritin molecule, but above pH 7.5, a pH-dependent Fe2+ binding profile is observed with up to 80 Fe2+ ions binding at pH 10.0. This Fe2+ binding is reversible and is accompanied by up to two H+ being released per Fe2+ bound at pH 10.0. The Fe2+ binding to apoferritin probably occurs in the 3-fold channels. A much larger and more complex pH-dependent Fe2+ binding stoichiometry was observed for holoferritin with up to 300 Fe2+ ions binding at pH 10.0. This pH-dependent Fe2+ binding was interpreted as Fe2+ interaction at the FeOOH mineral surface with displacement of H+ from -OH or phosphate surface groups by the incoming Fe2+ ions. Mossbauer spectroscopic measurements using 57Fe-labeled Fe2+ under anaerobic conditions showed that 57Fe2+ binding to holoferritin was accompanied by electron transfer to the core, yielding 57Fe3+, presumably bound to the mineral surface. Removal of added iron by Fe2+-specific chelating agents yielded 57Fe2+, demonstrating the reversibility of this electron-transfer process. The Fe2+ bound to apo- and holoferritin is readily converted to Fe3+ by exposure to O2 and strongly retained by the respective ferritin species.