SEARCH

SEARCH BY CITATION

Keywords:

  • transferrin;
  • polymorphism;
  • iron deficiency anaemia;
  • haemochromatosis;
  • mutation

Abstract

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Numerous polymorphisms of the transferrin gene result in a range of electrophoretic variants. We show that one of these mutations has a functional consequence. A G[RIGHTWARDS ARROW]A mutation at cDNA nucleotide 829 (G277S) was associated with a reduction in total iron binding capacity (TIBC). In menstruating white women, the G277S genotype was a risk factor for iron deficiency anaemia: iron deficiency anaemia was present in 27% of homozygous G277S/G277S women, 10% of G277G/G277S heterozygous women and 5% of homozygous wild-type G277G/G277G women.

Transferrin (TF) (OMIM 190000) is the most important biological carrier of iron in blood plasma. Severe mutations in transferrin leading to atransferrinaemia cause microcytic anaemia accompanied by hepatic accumulation of iron in humans (Beutler et al, 2000a) and mice (Trenor et al, 2000). This raises the question of whether less severe mutations in transferrin, which cause some loss in iron binding capacity or kinetic abnormalities in iron binding, can cause mild disorders of iron metabolism. Furthermore, could mild mutations in transferrin affect the degree of iron loading in haemochromatosis patients?

Previously, we have demonstrated that individuals with either the transferrin C1 or C2 variant (P570P or P570S) had serum iron, iron binding capacity and ferritin levels that were indistinguishable from each other (Lee et al, 1999). We have continued our search for polymorphisms in transferrin that may affect the severity of haemochromatosis in persons carrying mutations in the haemochromatosis gene, HFE. In the course of this study, we found that a common polymorphism in exon 7 of the transferrin gene, G[RIGHTWARDS ARROW]A at nucleotide 829 (G277S) was associated with a reduction in total iron binding capacity (TIBC). Although the loss in total iron binding capacity does not compromise the iron status of men and post-menopausal women, it predisposes menstruating women to iron deficiency anaemia.

Patients and methods

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Patient samples DNA samples were obtained with informed consent from patients undergoing health appraisal screening at Kaiser Permanente, San Diego (Beutler et al, 2000b). Data collected on all patients participating in the study included DNA analysis for mutations in the haemochromatosis HFE gene, measurements of serum ferritin, serum iron, total iron binding capacity and haemoglobin concentration. All patients were asked to fill out a questionnaire, which included data on their ethnicity and, for females, menstrual status. Only participants who indicated a single ethnic background are included in this analysis. At the time of this analysis, there were 28 598 patients participating in the Kaiser Study Group of whom 2373 (8·3%) are white menstruating women with a normal HFE genotype (C282C/C282C and H63H/H63H or H63H/H63D). Of these women 103 (4·34%) exhibited iron deficiency anaemia defined as a haemoglobin (Hb) level < 12 g/dl and a ferritin level < 20 µg/l.

Gene frequencies were obtained from 1342 randomly selected subjects and 250 subjects selected for African-American origin. In subsequent studies, we analysed 1271 white men and 1987 white women who were homozygous for the wild-type HFE genotype (C282C and H63H). Of the 1987 white women homozygous for the wild-type HFE genotype, 1227 were non-iron deficient as defined as having Hb levels geqslant R: gt-or-equal, slanted 12 g/dl and ferritin values geqslant R: gt-or-equal, slanted 20 µg/l. Studies examining the risk of the G277S mutation for iron deficiency anaemia were performed on 88 iron deficient (Hb < 12 g/dl and ferritin < 20 µg/l) menstruating women < 60 years old and 182 non-iron deficient (Hb geqslant R: gt-or-equal, slanted 12 g/dl and ferritin geqslant R: gt-or-equal, slanted 20 µg/l) menstruating women < 60 years old. The effect of the G277S mutation was also examined in 94 subjects homozygous for the mutant HFE C282Y/C282Y mutant genotype.

Sequence analysis Previously, we have identified intronic sequence flanking each exon (GenBank accession AF288139-11, AF294270-1) and described primers and conditions that were used to amplify and sequence each exon (Beutler et al, 2000a).

Allele specific oligonucleotide hybridization (ASOH) Exon 7 was amplified using primers described previously (Beutler et al, 2000a). Amplification of exon 7 was performed using polymerase chain reaction in a 50 µl reaction mix containing 33·5 mmol/l Tris pH 8·8, 8·3 mmol/l (NH4)2SO4, 3·35 mmol/l MgCl2, 85 µg/ml bovine serum albumin, 0·2 mmol/l dNTPs, 150 ng of each primer and 1 U of Taq polymerase (Perkin Elmer Cetus, Boston, MA, USA) and 50–100 ng of genomic DNA. After an initial denaturation at 95°C for 4 min, amplification was performed for 30 cycles at 95°C for 1 min, 60°C for 30 s and 72°C for 30 s. Allele-specific oligonucleotide hybridization (ASOH) was performed using the mutant oligomer 5′-GTATGGGCAGCAAGGAGG-3′ and the wild-type oligomer 5′-GTATGGGCGGCAAGGAGG-3′. Oligomers corresponding to each allele were radiolabelled with [32P]-γATP using polynucleotide kinase and purified by gel filtration on a G-50 spin column. The labelled oligomers were used at a final concentration of 1–2 × 106 c.p.m./ml of pre-hybridization solution. Hybridization was performed overnight at 42°C. The membranes were then washed in 6X Saline-sodium citrate (SSC), 0·1% sodium dodecyl sulphate (SDS) initially at room temperature and then at 1°C under the oligomer's Tm for 10 min.

GenBank accession numbers for the transferrin protein sequences were: human, NP_001054; rabbit, TFRBP; cow, AAA96735; pig, S01384; horse, S33761; rat, NP_058751; medaka, BAA10901; salmon, T11749; frog, S12100; drosophila, AAF48831; cockroach, A47275; lactoferrin, P02788.

Results

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Sequencing of the transferrin gene.

We sequenced the entire coding region of the transferrin (TF) gene in 28 individuals in order to identify mutations that might effect iron metabolism and possibly contribute to the penetrance of the HFE mutant genotype C282Y. The subjects included five homozygous for the HFE C282Y mutant haemochromatosis genotype with iron overload, six homozygous for the HFE C282Y mutant genotype without iron overload, five with a normal HFE gene and iron overload, six with a low total iron-binding capacity (TIBC < 31·33 µmol/l) and six normal subjects. All subjects were white except for one African-American and one Asian subject in the low TIBC group. All subjects had normal transferrin sequence in the coding region, varying primarily at nt 1765, the C1/C2 (P589P or P589S, respectively, OMIM Tf variant 4) genetic locus. Seven of the 28 individuals were heterozygous for the Tf C2 variant. One individual was heterozygous for the Tf B2 (G671E, OMIM Tf v3) variant. None of the 28 individuals exhibited the Tf D1 (D296G, OMIM TFv1), Tf chi (H319R, OMIM TFv2) or Tf Bv (K646E, OMIM Tf v5) variants. One normal subject and two white subjects with low TIBC were found to be heterozygous at nt 829 (G[RIGHTWARDS ARROW]A), which results in a G277S change, numbering from the initiator methionine.

Transferrin G277S variant

The Tf G277S variant was present at a frequency of 0·0634 in the normal white population (111 of 1752 alleles), 0·009 in the African-American population (5 of 568 alleles), 0·005 in the Hispanic population (1 of 222 alleles) and absent in the Asian population (n = 122 alleles). The Tf G277S allele did not appear tightly linked to any of the other polymorphisms although the existence of a low degree of linkage equilibrium could not be ruled out with the small numbers available.

Because Tf G277S variant was found primarily in persons of European ancestry, further studies were restricted to the white population. Furthermore, because mutations in the haemochromatosis gene, HFE, influence serum transferrin saturation and ferritin levels (Beutler et al, 2000b), individuals carrying mutations in HFE, except those who were simple heterozygotes for the phenotypically very mild H63D mutation were also excluded. The effect of the G277S variant on blood serum iron, transferrin saturation, and ferritin in 3258 white subjects are shown in Table I. Both men and women heterozygous for the Tf G277S variant were found to have a lower mean total iron binding capacity (TIBC). Men homozygous for the G277S variant were found to have a further reduction in the mean TIBC. Surprisingly, however, women who were homozygous for the G277S variant were found to have a higher mean TIBC than either heterozygotes or homozygous normal subjects. As iron deficiency causes an increase in TIBC, we examined the effect of Tf G277S on TIBC on non-iron deficient women (Hb levels geqslant R: gt-or-equal, slanted 12 g/dl and ferritin levels geqslant R: gt-or-equal, slanted 20 µg/l). Non-iron deficient women who were homozygous for Tf G277S had a lower mean TIBC than wild-type Tf G277G women. The presence of Tf G277S variant was found to increase the mean transferrin saturation in men and non-iron deficient women. There was no statistically significant effect of Tf G277S on mean corpuscular volume, Hb level, serum iron level, or serum ferritin in men or women. Nevertheless, there was a trend in women for the Hb, and ferritin levels to decrease with the presence of the G277S mutant allele.

Table I.   Effect of the transferrin G277S variant in white men and women on mean plasma iron parameters, total iron binding capacity (TIBC) (µmol/l), transferrin saturation (%), ferritin (µg/l), serum iron (µmol/l), mean corpuscular volume (MCV)(fl), and Hb (g/dl) levels.
 G277S/G277SG277G/G277SG277G/G277G
  • Ferritin values are geometric means. Values in parenthesis are lower, upper 95% confidence limits. Statistical analyses were performed on G277S homozygous and heterozygous values compared with G277G homozygous values.

  • *

    P < 0·001;

  • **

    P < 0·01,

  • ***

    P < 0·05.

  • n, the number of individuals in the group. All individuals were homozygous for the HFE wild-type genotype (C282C and H63H). Non-iron deficient women have Hb geqslant R: gt-or-equal, slanted 12g/dl and ferritin geqslant R: gt-or-equal, slanted 20µg/l).

White men
 Number81221141
 TIBC56·9 (52·3, 61·5)58·7 (57·4, 60·1)*61·8 (61·4, 62·3)
 Transferrin saturation33·8 (21·9, 45·7)29·0 (27·4, 30·7)**26·3 (25·7, 26·9)
 Ferritin111·7 (68·1, 183·3)118·6 (103·1, 136·6)113·7 (108·6, 119·0)
 Serum iron19·4 (11·9, 27·0)16·8 (15·9, 17·7)16·1 (15·7, 16·4)
 MCV89·8 (85·2, 92·3)89·6 (88·9, 90·4)89·7 (89·4, 89·9)
 Haemoglobin15·4 (14·3, 16·4)15·2 (15·0, 15·3)15·1 (15·0, 15·1)
White women
 Number171941776
 TIBC68·0 (60·9, 75·1)63·9 (62·7, 65·2)*67·0 (66·5, 67·4)
 Transferrin saturation20·5 (15·0, 26·1)22·4 (21·0, 23·8)21·3 (20·9, 21·8)
 Ferritin23·7 (13·1, 42·6)33·4 (29·1, 38·3)34·2 (32·8, 35·7)
 Serum iron13·1 (9·9, 16·3)13·9 (13·1, 14·7)13·9 (13·6, 14·2)
 MCV89·6 (86·4, 92·9)88·6 (87·8, 89·4)89·0 (88·8, 89·3)
 Haemoglobin12·7 (12·1, 13·3)13·1 (13·0, 13·3)13·2 (13·1, 13·2)
Non-iron deficient white women
 Number81131106
 TIBC61·2 (56·7, 65·6)62·4 (60·9, 63·9)***64·7 (64·2, 65·3)
 Transferrin saturation27·5 (19·8, 35·2)26·3 (24·6, 28·1)***23·8 (23·2, 24·3)

Three of the 17 white women who were homozygous for Tf G277S/G277S were iron deficient judged by the finding that they had Hb levels < 12 g/dl and ferritin levels < 20 µg/l (data not shown). Six of the 17 white women homozygous for G277S/G277S were non-menstruating and were therefore less likely to be iron-deficient. This raised the question as to whether the Tf G277S variant pre-disposed menstruating women to iron deficiency anaemia. Of the 1530 menstruating women 88 (5·8%) exhibited iron deficiency anaemia (Hb levels < 12·1 g/dl and ferritin levels < 20 µg/l). Three of 11 homozygous mutant G277S/G277S women (27·3%) were iron deficient, 15 of 151 heterozygous G277G/G277S women (10%) were iron deficient and 70 out of 1349 wild-type homozygous G277G/G277G women (5·2%) were iron deficient (χ2 = 14·9; P < 0·001) (Fig 1).

image

Figure 1.  Percentage of iron-deficient menstruating white women within each transferrin genotype, Tf G277G/G277G, G277G/G277S and G277S/G277S. Numbers above the bars indicate the number of individuals with that genotype.

Download figure to PowerPoint

A comparison was also made of iron deficient women, as defined above, with those with Hb levels > 13·5 g/dl and ferritin levels > 50 µg/l (non-iron deficient women). The frequency of the mutant allele in the iron deficient menstruating population was twice as high as in the non-iron deficient population (χ2 = 10·8; p < 0·005) (Table II). The relative risk of anaemia in patients carrying the mutant allele was 2·2 with a 95% confidence interval of 1·187–4·039.

Table II.   Percentage of each genotype within the group of menstruating iron deficient women < 60years or group of menstruating non-iron deficient women < 60 years.
 G277S/G277S (%)G277G/G277S (%)G277G/G277G (%)
  1. n, the number of individuals in the group.

Iron deficient women (n = 88)3·4117·0579·55
Non-iron deficient women (n = 182)0·558·7990·66

The effect of the Tf G277S genotype on individuals homozygous for the mutant HFE C282Y genotype was also examined to determine if the mutation was associated with an increase in transferrin saturation and serum ferritin levels. The presence of the G277S allele in the heterozygous state in HFE patients reduced mean TIBC levels to a degree similar to that of white individuals with a normal HFE gene. Nevertheless, owing to the small number of subjects, there was no significant effect of the G277S mutation on transferrin saturation or serum ferritin levels in haemochromatosis patients (Table III).

Table III.   Effect of the transferrin G277S variant in HFE C282Y/C282Y homozygous individuals on mean plasma iron parameters, total iron binding capacity (TIBC) (µmol/l), transferrin saturation (%), ferritin (µg/l), serum iron (µmol/l), mean corpuscular volume (MCV)(fl) and haemoglobin (g/dl) levels.
 G277S/G277SG277S/G277GG277G/G277G
  1. Ferritin values are geometric means. Values in parenthesis are the lower, upper 95% confidence limits. Statistical analyses were performed on G277S homozygous and heterozygous values compared with G277G homozygous values where n > 2. None of the values were statistically significant.

Males C282Y/C282Y
n2640
TIBC44·846·6 (42·9, 50·2)49·0 (46·7, 51·4)
Transferrin saturation67·143·2 (19·3, 67·1)63·3 (56·8, 69·7)
Ferritin965·8187·4 (25·5, 1377·5)341·7 (214·2, 545·1)
Serum Iron19·620·5 (8·0, 33·0)27·0 (23·4, 30·5)
MCV96·297·0 (94·8, 99·3)93·7 (91·8, 95·7)
Haemoglobin13·315·5 (14·5, 16·6)14·8 (14·3, 15·4)
Females C282Y/C282Y
n0343
TIBC 50·3 (36·8, 63·8)52·4 (49·2, 55·4)
Transferrin saturation 47·7 (−1·7, 97·1)51·2 (44·4, 58·0)
Ferritin 312·3 (56·7, 1719·9)179 (117·2, 274·2)
Serum Iron 13·3 (−5·0, 31·6)22·3 (18·5, 25·6)
MCV 93·5 (68·5, 118·6)93·1 (91·2, 95·1)
Haemoglobin 12·5 (6·9, 18·2)13·2 (12·9, 13·7)

Discussion

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Serum transferrin is rich in mutations. Extensive electrophoretic polymorphisms have been known since the late 1950s (Smithies, 1957, 1958), but none of these seemed to have functional abnormalities except for the vary rare atransferrinaemia mutations. Because atransferrinaemia and hypotransferrinaemia result in hepatic iron overload, we questioned whether milder mutations in transferrin might lead to a disruption of iron homeostasis and contribute to the iron overload seen in haemochromatosis patients. Mutations were not found in the transferrin gene of five individuals with iron overload and a normal HFE genotype. This suggests that abnormalities in transferrin are not responsible for the iron accumulation seen in these patients. Furthermore, there were no mutations found in the transferrin gene that could account for the difference in the degree of iron overload in individuals homozygous for the HFE mutant C282Y genotype. Eleven individuals in our study were homozygous for the HFE C282Y mutant genotype and carried the G277S transferrin mutation. Given this small number, it was not possible to draw any conclusions as to whether presence of the Tf G277S allele influenced iron uptake and accumulation in HFE patients as measured by increased ferritin and transferrin saturation levels.

Beckman et al (1998) had previously suggested that the G277S polymorphism was identical to the transferrin electrophoretic C3 variant. They were unable to confirm this hypothesis as they had not sequenced the entire transferrin gene from DNA isolated from a transferrin C3 individual. We have sequenced the entire transferrin coding region of three individuals heterozygous for the G277S variant and found no other mutations. Because of this, and because the gene frequency is the same as that reported for the transferrin C3 variant (Roychoudhury & Nei, 1999), we believe that the G277S mutation is, in fact, the molecular basis of the C3 transferrin electrophoretic variant.

This variant is found at the highest frequency in the white population and at significantly lower frequency in the African-American and Hispanic populations. Both men and women with the Tf G277S allele had a lower mean plasma TIBC. This suggests that the G277S mutation results in the formation of a transferrin molecule that is either less stable, resulting in reduced plasma transferrin levels, or that it has a lower affinity for iron. Plasma iron levels are normal; hence the presence of the G277S mutant allele, even in the heterozygous state, results in a higher mean transferrin saturation because there is a reduced amount of total or ‘available’ transferrin. Among white menstruating women with a normal HFE genotype, we found that the frequency of iron deficiency anaemia was fivefold higher among women homozygous for the G277S mutation than among those homozygous for the wild-type G277G genotype. Furthermore, the frequency of the G277S mutant allele was twofold higher in the population of iron deficient women than in the non-iron deficient population. Although the mechanism by which this mutation causes an increased risk of iron deficiency is not entirely clear, it is notable that an increase in the transferrin concentration is one of the most prominent responses to the development of iron deficiency (Fairbanks et al, 1971). Presumably, the increase in transferrin concentration occurs in response to iron deficiency in an effort to increase iron uptake.

It is notable that the glycine at amino acid position 277 is conserved between human, rat, cow, pig, horse, rat and frog but is a valine in drosophila (Fig 2). Thus, the level of conservation across species suggests that the glycine in amino acid position 277 is important in maintaining the biological activity and/or structure of transferrin.

image

Figure 2.  Comparison of the transferrin amino acid sequence among species in the region of the G277S variant. The human transferrin sequence represents amino acids 275–294. The GenBank accession numbers are described in Patients and Methods.

Download figure to PowerPoint

Acknowledgments

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

This is manuscript 13679-MEM from The Scripps Research Institute. This work was supported by NIH grants DK53505-02, HL 25552-10 and RR00833 and funds from the Stein Endowment Fund.

References

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  • Beckman, L.E., Van Landeghem, G.F., Sikström, C., Beckman, L. (1998) DNA polymorphism and haplotypes in the human transferrin gene. Human Genetics, 102, 141144.DOI: 10.1007/s004390050667
  • Beutler, E., Gelbart, T., Lee, P., Trevino, R., Fernandez, M.A., Fairbanks, V.F. (2000a) Molecular characterization of a case of atransferrinemia. Blood, 96, 40714074.
  • Beutler, E., Felitti, V., Gelbart, T., Ho, N. (2000b) The effect of HFE genotypes in patients attending a health appraisal clinic. Annals of Internal Medicine, 133, 329337.
  • Fairbanks, V.F., Fahey, J.L., Beutler, E. (1971) Clinical Disorders of Iron Metabolism, 2 edn. Grune & Stratton, New York.
  • Lee, P.L., Ho, N.J., Olson, R., Beutler, E. (1999) The effect of transferrin polymorphisms on iron metabolism. Blood Cells, Molecules, and Diseases, 25, 374379.
  • Roychoudhury, A.K. & Nei, M. (1999) Transferrin. In: Human Polymorphic Genes. World Distribution, pp. 177181. Oxford University Press, New York.
  • Smithies, O. (1957) Variations in human β-globulins. Nature, 180, 14821483.
  • Smithies, O. (1958) Third allele at the serum β -globulin locus in humans. Nature, 181, 12031204.
  • Trenor, C.C., Campagna, D.R., Sellers, V.M., Andrews, N.C., Fleming, M.D. (2000) The molecular defect in hypotransferrinemic mice. Blood, 96, 11131118.