Iron speciation in hypotransferrinaemic mouse serum.

A mouse model of human disease was recently described by Bemstein I1 I . Hypotransferrinaernic mice have 5O-fold decreased serum transfenin levels, hypochromic. riiicrocytic (iron deficient) anaemia but develop liver iron overload I1.21. These mice parallel for the rare human disease atrmsferrlnaemia 13.41 and may be a useful model for hiiman iron overload disorders. Great interest is currently focussed on (he non transferrin-bound iron reported to be present in scriiiii from pallenls with iron overload 15-81, as the nature and reactivity of the non transferrin-bound iron species may be o l significance in determining tissue uptake of iron as bvell a s tissue damage. Hypotransferrinaemic niice provide a usef~tl model for the study of iron species produced in the seruiii during iron overload. without the complicating presence of Iransferrin. Serum from homozygous hypotransferrinaemic mice (data art. given for a nilxed group of males and females aged 6-7weeksl was found to contain apparently very low levels of seniiii iron (mean 1.5+0.5uM (SEM. n=5). range <0.5-4uM). This iron fraction includes all species reactive with the Fe(II1 chelator ferrozine at pH 4.5 in the presence of hydroxylaxnine. Model compounds such as transferrin-Fe. FeNTA2. Fe/citrate ( 1 :2) or Fe/alburnin (prepared by mixing 50vM FeC13 with IOIiig/iiil bovine serum albumin (pH711 are all detectable by this assay. but fenitln iron or FeEDTA (1:2) are relatively unreactive. Study of hypokansferrinaemic mouse serum with a bleomycin-dependent assay 161. sensitive to iron ions and low molecular mass complexes revealed 28+9&1M (n=3) iron (wildtype controls showed 3.3+3.3~M. n=7). sugjiesting the presence of non transfenin-bound iron species. Determination 0 1 nonhaem iron by boiling serum with 12.5% trlchloroacetic acid/ 2% Naqpyrophosphate (TCA/PPi) (191 the iron being detected with ferrozine/ ascorbate [lo]) demonstrated large concentrations of iron in hypotransferrinaernic serum. The quantitiy of iron detectable was found to increase progressively with repeated extraction of the serum (11.7+2.3&1M. n=7. with a single extraction. 39.329.4vM. n=9 with 3 extractions. p<O.O11. Even higher quantities of iron were detectable by atomic absoqlion spectrophotonietry (65.3+16.5vM. n=9. p<O.Ol). Serum nonhaem and total iron were positively correlated with liver nonhaem Fe (Figure 11. Two possible explanations for these observations are the presence of haeni iron or ferritin iron in the serum. Haenioglobin iron was detectable at approx. l O u M by spectrophotometry. based on the alpha absorption band near 540nm. but test experinients showed that less than 20% of haemoglobin-Fe is released by boiling 3 times in TCA/PPi. Serum ferritln was detemiined by fluodgenlc ELISA 1111 using anti-mouse liver ferritin with values of 641+128pg/l (n=14) in hypotransferrlnaeniic mice (wild-types 44+6pg/I. n= 14). Even if highly loaded with iron, this ferritin could not explain the observed levels of iron in the hypotransferrinaernic serum. FerriUn was found to 80-90% extractable by a single boiling with TCA/PPi. a s was the iron in overloaded livers from hypotransferrinameic mice. This suggests that known tissue iron storage proteins and haem cannot account for most of the iron present in hypotransferrlnaeniic serum. Serum was fractionated by Sephadex (2200 chromatography (elution with lOmM HEPES/ 0.15M NaC1. pH 7.0) and fractions were assayed for iron by atomic absorption spectrophotometry. Three distinct species were observed with apparent niolecular weights of ~ 2 0 0 k d . 30-50kd and 1-5kd. The non transferrinbound Fe therefore consists of several distlnct species. possibly with dlfTering toxicity and availability for tissue uptake. Further i: O l o $ *