review
Genetic insights into the clinical diversity of b thalassaemia
The identification of defective genes underlying inherited
diseases has made it clear that patients with the same genotype
2 can have very variable patterns of clinical expression. The
profound phenotypic variability of the b thalassaemias,
including age of onset, varying involvement of different organs
and transfusion needs, is prototypical of how the wide
spectrum in disease severity of a monogenic disorder can be
generated. Relating phenotype to genotype is complicated not
only by the complex interaction of the environment with the
different allelic variants, but interaction with other genetic
factors at the secondary and tertiary levels is also involved.
Evidence for these modifier genes comes from the range of
phenotypes within families sharing the same genotypes. This
article presents an overview of the b globin gene structure,
function and expression, followed by a short description of the
clinical and haematological diversity encountered in b thalas-
saemia and the underlying pathophysiology. A discussion of
the genetic basis of the disease is presented with an overview of
the various genetic loci (modifier genes) that modulate the
effects of its clinical expression on different organs.
Introduction
b thalassaemia can be broadly defined as a syndrome of
inherited haemoglobin disorders characterized by a quanti-
tative deficiency of functional b globin chains. Although it is
defined as a reduction in the synthesis of b globin, some forms
result from structural haemoglobin variants that are ineffec-
tively synthesized or are so unstable that they result in a
functional deficiency of the b chains and a thalassaemia
phenotype (Weatherall & Clegg, 2001). The most common
forms are those that are prevalent in the malarial tropical and
subtropical regions, where a few mutations have reached high
gene frequencies because of the protection they provide against
malaria. In these countries where b thalassaemia is prevalent, a
limited number of alleles (four to six) account for more than
90% of the b thalassaemia (Flint et al, 1998), allowing a
targeted molecular diagnostic approach to be undertaken. In
other countries such as the UK, where there is an ethnic mix
because of recent population migrations, a screening approach
may be more appropriate and effective.
Correspondence: Professor Swee Lay Thein, Department of
Haematological Medicine Guy’s, King’s & St Thomas’ School of
Medicine, Denmark Hill Campus, Bessemer Road, London SE5 9PJ.
E-mail: sl.thein@kcl.ac.uk
264 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
b globin gene – structure, function and
expression
b globin is encoded by a structural gene found in a cluster with
the other b like genes on the short arm of chromosome 11. The
cluster contains five functional genes, 5¢--Gc-Ac-wb-d-b-3¢,
which are arranged in the order of their developmental
expression. Upstream of the entire b globin complex is the
locus control region (LCR), which consists of five DNase 1
hypersensitive (HS) sites (designated HS 1–5) distributed
between 6 and 20 kb 5¢ of the  gene. The LCR plays a critical
role in b globin gene expression by maintaining an open
chromatin state and acting as a powerful enhancer of globin
gene transcription; in its absence, the level of b globin gene
expression is low. Four of the sites (HS 1–4) are erythroid-
specific, encompassing binding sequences for erythroid-restric-
ted transcription factors (GATA-1 and NF-E2), while HS5 is
ubiquitous. There is one other hypersensitive site approxi-
mately 20 kb 3¢ to the b gene. The two extreme HS sites
flanking the b complex have been suggested to mark the
boundaries of the b globin gene domain. The b globin complex
is embedded in a cluster of olfactory receptor genes (ORG),
part of the family of approximately 1000 genes that are widely
distributed throughout the genome, and expressed in the
olfactory epithelium (Bulger et al, 2000).
The general structure of the b globin gene is typical of the
other globin loci (Efstratiadis et al, 1980). The genomic
sequence, which codes for 146 amino acids, spans 1600 bp;
the transcribed region is contained in three exons separated by
two introns or intervening sequences (IVSs). Exon 2 encodes
the residues involved in haem binding and ab dimer forma-
tion, while exons 1 and 3 encode for the non-haem-binding
regions of the b globin chain. Many of the amino acids
involved in globin subunit interactions required for the Bohr
effect, and 2,3-diphosphoglycerate binding, are found in exon
3. Conserved sequences important for b globin gene expression
are found in the 5¢ promoter region, at the exon–intron
junctions, and in the 3¢ untranslated region (3¢-UTR) at the
end of the mRNA sequences. The b globin gene promoter
includes three positive cis-acting elements: a TATA box
(positions )28 to )31), a CCAAT box (positions )2 to )76)
and duplicated CACCC motifs (proximal at positions )86 to
)90, and distal at position )101 to )105). While the CCAAT
and TATA elements are found in many eukaryotic promoters,
the CACCC sequence is found predominantly in erythroid cell-
specific promoters. Binding of the erythroid Krüppel like
factor (EKLF) to the CACCC motif appears to be crucial for

normal adult b globin expression. In addition to these motifs,
the region upstream of the b globin promoter contains two
binding motifs for the erythroid transcription factor GATA-1.
The importance of these various 5¢–flanking sequences for
normal gene expression is underscored by b thalassaemia
arising from point mutations in these sequences specifically in
and around the TATA box and the CACCC motifs in the )80
to )100 region. An enhancer is also found in intron 2
(Antoniou et al, 1988) and 3¢ of the globin gene, 600–900 bp
downstream of the poly (A) site (Trudel & Costantini, 1987).
The 5¢-UTR occupies a region of 50 nucleotides between the
CAP site, the start of transcription, and the initiation (ATG)
codon. There are two prominently conserved sequences in the
5¢-UTR of the various globin genes (both a and b) (Collins
& Weissman, 1984). One is the CTTCTG hexanucleotide found
8–13 nucleotides downstream from the CAP site, i.e. at positions
+8 to +13. The second conserved sequence is CACCATG, in
which the last three nucleotides form the initiation codon
(ATG). Again, the importance of these sequences in the
regulation of the b gene expression is exemplified by several
mutations in the 5¢-UTR causing b thalassaemia.
The 3¢-UTR constitutes the region of 132 nucleotides
between the termination codon (TAA) and the poly (A) tail
with one conserved sequence, AATAAA, located 20 nucleotides
upstream of the poly (A) tail. Several mutations affecting the
AATAAA sequence and other sequences in the 3¢-UTR causing
b thalassaemia have been described.
The developmental regulation of the globin genes reflect
their sequential activation in a 5¢)3¢ direction. Transcription
of the  gene in the embryonic yolk gene switches after the
sixth week of gestation to the transcription of the two c genes
in the fetal liver, and then around the prenatal period, to that
of the d (minor adult) and b (major adult) genes. At 6 months
after birth, Hb F comprises <5% of the total haemoglobin and
continues to fall reaching the adult level of <1% at 2 years of
age. It is at this stage that mutations affecting the b gene
become clinically apparent. The ‘switch’ from fetal (c) to adult
(b) haemoglobin production is not complete, and small
amounts of c expression persist in adult life. All adults have
residual amounts of fetal haemoglobin (Hb F, a2c2), present in
a subset of erythrocytes called F cells which also contain adult
(a2b2) haemoglobin (Boyer et al, 1975). These levels of Hb F
and F cells in adults vary considerably, and are largely
genetically controlled (Garner et al, 2000a).
The developmental expression of the individual globin genes
relies on two mechanisms, gene silencing and gene competi-
tion, governed by direct physical interactions between the
globin promoters and the b-LCR (Carter et al, 2002; Tolhuis
et al, 2002), which are dependent on the transcription
environment in embryonic, fetal and adult cells. While the
- and c globin genes are autonomously silenced at the
appropriate developmental stage, expression of the adult b
globin gene depends on the lack of competition from the c
gene for the LCR sequences. This is supported by the
concomitant down-regulation of the cis b gene when the c
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
Review
gene is up-regulated by promoter mutations, as illustrated by
the non-deletional hereditary persistence of fetal haemoglobin
(HPFH) (Wood, 2001). In addition, mutations that affect the
b promoter, which remove competition for the b-LCR, tend to
be associated with variable increases in the c and d gene
expression (Huisman, 1997; Thein, 1998).
Pathophysiology and clinical diversity
of b thalassaemia
The underlying pathophysiology of b thalassaemia relates to a
quantifiable deficiency of functional b globin chains, which
leads to an imbalanced globin chain production and an excess
of a globin chains (Weatherall & Clegg, 2001; Schrier, 2002)
(see Fig 1). The latter aggregate in red cell precursors, forming
inclusion bodies that cause mechanical damage and their
premature destruction in the bone marrow, i.e. ineffective
erythropoiesis. Red cells that survive to reach the peripheral
circulation are prematurely destroyed in the spleen. Anaemia
in b thalassaemia thus results from a combination of
ineffective erythropoiesis, peripheral haemolysis and an overall
reduction in haemoglobin synthesis. Factors which reduce the
degree of chain imbalance and the magnitude of a chain excess
in the red cell precursors, have an ameliorating effect on the b
thalassaemia phenotype.
A direct effect of the anaemia is the increased production of
erythropoietin, which leads to intense proliferation and
expansion of the bone marrow with the resulting skeletal
deformities. These secondary complications of bone disease,
splenomegaly, endocrine and cardiac damage can be related to
the severity of anaemia and the iron loading that results from
the increased gastrointestinal absorption and the blood
transfusions. Recently, it has become apparent that these
complications of b thalassaemia may also be genetically
modified by variability at other loci (tertiary modifiers).
The clinical manifestations of b thalassaemia are extremely
diverse, spanning a broad spectrum from the transfusion-
dependent state of thalassaemia major to the asymptomatic
state of thalassaemia trait. The most severe end of the spectrum
is characterized by the complete absence of b globin production
and results from the inheritance of two b thalassaemia alleles,
homozygous or compound heterozygous states. This combi-
nation of genotype usually results in b thalassaemia major and,
at their worst, the patients present within 6 months of life, and
if not treated with regular blood transfusions, die within their
first 2 years. Conversely, many patients who have inherited two
b thalassaemia alleles may have a milder disease, ranging from a
condition that is only slightly less severe than transfusion
dependence through a spectrum of decreasing severity to one
that is asymptomatic and often mistaken as b thalassaemia trait.
This diverse collection of phenotypes between the two extremes
of thalassaemia major and thalassaemia trait constitute the
clinical syndrome of thalassaemia intermedia. In a large
number of patients with thalassaemia intermedia, the reduced
disease severity can be explained by the inheritance of the
265
Review

Fig 1. Pathophysiology of b thalassaemia. Factors that modify the b thalassaemia phenotype act at three levels: (1) primary – generally these refer to
the nature of the mutation affecting the b globin gene itself, (2) secondary – loci that affect the a/b globin chain imbalance and include the a and c
globin genes. However, loci that affect the stability and amount of the a chain and loci that affect expression of the c globin should also be included;
(3) tertiary – loci that are not involved in globin imbalance but modify the complications of the disease in different ways. These include jaundice and
gallstones (UGTA1), iron loading (HFE) and bone disease (VDR, OesR, COL1A1).

milder b thalassaemia alleles (b++ and ‘silent’) that allow the
production of a significant proportion of b globin chains. A
substantial number, however, have b thalassaemia, and in such
cases, the absence of b globin chains is compensated by an
inherent ability to produce fetal haemoglobin (Hb F a2c2). Yet
other thalassaemia patients have inherited only one b thalas-
saemia allele. Most cases of unusually severe heterozygous b
thalassaemia are due to the co-inheritance of extra a globin
genes (see below) while others are due to the nature of the
underlying b thalassaemia mutation itself (see ‘Dominantly
inherited b thalassaemia’). Hence, the underlying genotypes of
b thalassaemia intermedia are extremely heterogeneous; the
genetic basis can be the inheritance of one or two b
thalassaemia alleles interacting with other genetic variables.
Carriers for b thalassaemia are typically clinically asympto-
matic; they may have a mild anaemia with characteristic
hypochromic microcytic red blood cells, elevated levels of
HbA2 and variable levels of Hb F. However, even the
heterozygous states for b thalassaemia show a phenotypic
diversity comparable with that of thalassaemia major. In some
cases, the b thalassaemia allele can be phenotypically ‘silent’,
with no anaemia or any haematological abnormalities. In
others, the heterozygous state causes a phenotype almost as
severe as the major forms, that is, the b thalassaemia allele is
dominantly inherited.
Although definition of the two extremes of the clinical
spectrum of b thalassaemia is easy, assigning the severity of the
intermediate form can be problematical. Criteria such as age
and level of haemoglobin at presentation, transfusion history
and the requirements for intermittent blood transfusion have
been used, but these have their inherent limitations and are
highly subjective and clinician dependent.
Mechanisms underlying phenotypic diversity of
b thalassaemia
Progress in our understanding of the mechanisms underlying
the remarkable phenotypic variability of b thalassaemia has
been made possible by a combination of the analysis of the
molecular basis of the different forms of thalassaemia, family

266 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274

studies and analysis of the genotype/phenotype relationship of
the thalassaemia intermedia.
Primary modifiers – heterogeneity and variable severity of
b thalassaemia alleles
With the exception of a few deletions, the vast majority of b
thalassaemias are caused by point mutations within the gene or
its immediate flanking sequences. A few b thalassaemia
mutations that segregate independently of the b globin cluster
have been described in several families (Thein, 1998); in such
cases, trans-acting regulatory factors have been implicated.
Examples of reduced b globin production caused by loci
outside the b globin complex include those mutations arising
in the general transcription factor TF11H (Viprakasit et al,
2001) and the erythroid-specific transcription factor GATA1
(Yu et al, 2002).
b versus b+ thalassaemia alleles. Functionally, the b
thalassaemia alleles can be classified as b or b+ reflecting the
resulting phenotype: b thalassaemia in which there is a
complete absence of b globin production and the most severe
possible, and b+ thalassaemia in which there is some, albeit
reduced b globin product. Mild b thalassaemia, sometimes
referred to as b++, alleles allow a moderate amount of b globin
to be produced, while in the ‘silent’ b thalassaemia, the deficit
in b chain production is minimal and carriers have minimally
reduced or normal red cell indices and their HbA2 levels are
normal.
The point mutations causing b thalassaemia result from
single base substitutions, minor insertions or deletions of a few
bases within the gene or its immediate flanking sequences
(Thein, 1998), and may affect any level of genetic regulation.
Mutations affecting transcription can either involve the
conserved DNA sequences that form the b globin promoter
or the stretch of 50 nucleotides in the 5¢-UTR. Generally they
result in a mild to minimal deficit of b globin output that
reflects the relatively mild phenotype of these b+ thalassaemias.
The C–T mutation at position )101 to the b globin gene
appears to cause an extremely mild deficit of b globin, such
that it is ‘silent’ in heterozygotes who have normal HbA2 levels
and normal red cell indices (Gonzalez-Redondo et al, 1989;
Maragoudaki et al, 1999). Several mutations in the 5¢-UTR,
e.g. CAP + 1A-C, also have a ‘silent’ phenotype (Wong et al,
1987).
Mutations that affect RNA processing can involve either of
the invariant dinucleotides (GT at 5¢ and AG at 3¢) in the splice
junction in which case normal splicing is completely abolished
with the resulting phenotype of b thalassaemia. Mutations
within the consensus sequences at the splice junctions reduce
the efficiency of normal splicing to varying degrees and
produce a b+ phenotype that ranges from mild to severe.
Mutations within introns or exons might also affect the
splicing pattern of the pre-mRNAs. For example, a cryptic
splice site that contains the sequence GT GGT GAG G has been
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
Review
found in exon 1 of the b globin gene, spanning codons 24–27.
Three mutations within this region activate this cryptic site,
which acts as an alternative donor site in RNA processing. The
mutations in codon 26 (GAG fi AAG) that gives rise to HbE
(b 26 Gln fi Lys) is one such mutation that activates this
cryptic splice site, with a reduction of the normal splicing that
produces the HbE variant. As HbE production is also
quantifiably reduced, the compound heterozygous state,
HbE/ b thalassaemia results in a clinical picture closely
resembling homozygous b thalassaemia, ranging from severe
anaemia and transfusion-dependency to thalassaemia inter-
media. Other RNA processing mutants affect the polyadeny-
lation signal (AATAAA) and the 3¢-UTR. These are generally
mild b+ thalassaemia alleles.
Mutations that abrogate mRNA translation either at the
initiation or extension phases of globin synthesis are all
associated with a b phenotype. Approximately half of the b
thalassaemia alleles are characterized by premature termin-
ation of b chain extension. They result from the introduction
of premature termination codons due to frameshifts or
nonsense mutations and nearly all terminate within exons 1
and 2. These mutations are associated with minimal steady-
state levels of mutant b mRNA in erythroid cells. In
heterozygotes for such cases, no b chain is produced from
the mutant allele and only half the normal b globin is present,
resulting in a typical asymptomatic phenotype.
The ‘silent’ mutations are normally identified in the
compound heterozygous states with a severe b thalassaemia
allele, which results in thalassaemia intermedia, or in homo-
zygotes who have a typical phenotype of b thalassaemia trait.
The ‘silent’ b thalassaemia alleles are not common, except for
the )101 C–T, which accounts for a large number of the milder
forms of b thalassaemia in the Mediterranean (Maragoudaki
et al, 1999).
The mild b thalassaemia (b++) alleles are associated with
clearly defined changes in heterozygotes and result in disorders
of intermediate severity in homozygotes. Interactions with the
severe alleles are less predictable because of the wider range of
b globin output, and extend from transfusion dependence to
intermediate forms of b thalassaemia at the mild end of the
spectrum (Camaschella et al, 1995; Ho et al, 1998).
Deletions causing b thalassaemia result in a complete
absence of b globin product and should cause a severe
phenotype. However, it is now apparent that these rare b
thalassaemia alleles are associated with unusually high Hb F
and HbA2 levels in heterozygotes. These deletions differ widely
in size, and apart from the Indian 619 bp deletion, commonly
remove a region (from positions )125 to +78 relative to the
mRNA cap site) in the 5¢-b promoter, which includes the
CACCC, CCAAT and TATA elements. The mechanism for
the unusually high levels of HbA2 and Hb F in heterozygotes
for these deletions appears to be related to the removal of
competition for the upstream LCR and limiting transcription
factors, resulting in increased interaction with the c and d
genes in cis, enhancing their expression. Although the increases
267
Review
in Hb F are variable and moderate in heterozygotes, they are
adequate to compensate for the complete absence of b globin
in homozygotes for these deletions (Schokker et al, 1966; Craig
et al, 1992). This mechanism may also explain the unusually
high HbA2 levels that accompany point mutations affecting the
b promoter region.
Transposable elements may occasionally disrupt human
genes and result in their inactivation. The insertion of such an
element, a retrotransposon of the family called L1, has been
reported with the phenotype of b+ thalassaemia. Despite the
insertion of 6–7 kb DNA into its IVS2, the affected gene
expresses full length b globin transcripts at a level corres-
ponding to approximately 15% of normal b globin mRNA
(Divoky et al, 1996).
The variable severity of the different b thalassaemia alleles is
reflected in their phenotypic effect in heterozygotes, in the
degree of hypochromia and microcytosis as indicated by the
mean cell haemoglobin and mean cell volume (MCV) values
respectively. Rund et al (1991) showed that the b thalassaemia
alleles which are associated with the most severe phenotype,
demonstrated a fairly tight range of MCVs (63Æ1 fl; SD ¼ 3Æ4)
while the b+ alleles were associated with a wider range of
MCVs (69Æ3 fl; SD ¼ 5Æ6). The broader range of MCV in b+
thalassaemia when compared with b thalassaemia, is not
surprising given the broad range in the deficit of b globin
production, from barely detectable levels at the severe end, to
moderately reduced (b++) and to just a little less than normal
in the ‘silent’ b thalassaemia alleles.
A more recent study has taken the correlation between the
severity of b thalassaemia alleles with haematological param-
eters to a finer level. Skarmoutsou et al (2003) measured a
series of haematological parameters, including reticulocyte
haemoglobin content (CHr), soluble transferrin receptor
(sTfR), reticulocytes and HbA2 and F levels in 57 iron-replete
individuals with heterozygous b thalassaemia. sTfR values, a
reliable quantifiable assessment of the erythropoietic activity,
were lowest in the very mild b thalassaemia (bsilent), and
highest in b thalassaemia heterozygotes. CHr, a product of
reticulocyte haemoglobin and volume, was lower in bsilent
thalassaemia compared with normals but the difference was
not statistically significant. However, the CHr values between
the bsilent and the severe groups (b+ and b thalassaemia alleles)
was significant, being much higher (27Æ0–32Æ0 pg) in the
former, compared with 19Æ5 to 25Æ3 pg in the latter group.
Furthermore, while sTfR values showed a positive correlation
with HbA2, there was a significant negative correlation between
CHr and HbA2 levels. This study confirms that all heterozy-
gous b thalassaemias have some degree of ineffective erythro-
poiesis that varies with the severity of the b thalassaemia
mutation.
Dominantly inherited b thalassaemia. The common b
thalassaemia alleles that are prevalent in the malarious
regions are inherited typically as Mendelian recessives;
heterozygotes are clinically asymptomatic and the inheritance
268 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
of two mutant alleles – as homozygotes or compound
heterozygotes – is required to produce clinical disease. Some
forms of b thalassaemia, however, are dominantly inherited, in
that inheritance of a single b thalassaemia allele in the presence
of a normal a globin genotype, results in a clinically detectable
disease (Thein, 1999; Thein, 2001). Heterozygotes have a
thalassaemia intermedia phenotype with moderate anaemia,
splenomegaly and a thalassaemic blood picture. Apart from the
usual features of heterozygous b thalassaemia, such as
increased levels of HbA2 and the imbalanced a/b globin
biosynthesis, large inclusion bodies similar to those seen in
thalassaemia major, are often observed in the red cell
precursors, hence the original term of ‘inclusion body b
thalassaemia’ (Fei et al, 1989).
This unusual form of b thalassaemia was probably first
described in an Irish family in 1973 (Weatherall et al, 1973);
several members of the family spanning three generations had
a thalassaemia intermedia phenotype that was clearly inherited
as a Mendelian dominant. Since the first description, more
than 30 dominantly inherited b thalassaemia alleles have now
been described (Thein, 1992, 1999); they include missense
mutations, minor deletions leading to the loss of intact codons,
frameshifts arising from minor insertions and deletions
resulting in elongated b variants with abnormal carboxy
terminal ends, and truncated b variants resulting from
nonsense mutations. The common denominator of these
mutations is the predicted synthesis of highly unstable b chain
variants, so unstable that in many cases, they are not detectable
and only implicated from the DNA sequence. The predicted
synthesis is supported by the presence of substantial amounts
of abnormal b mRNA in the peripheral reticulocytes (Hall &
Thein, 1994), comparable in amounts with that produced from
the normal b allele. Indeed, the large intra-erythroblastic
inclusions, which are so characteristic of this form of b
thalassaemia, have subsequently been shown to be composed
of both a and b globin chains (Ho et al, 1997). In contrast, the
inclusion bodies in homozygous b thalassaemia consisted only
of precipitated a globin.
It should be noted that some of these dominantly inherited
b thalassaemia alleles arise from premature termination
mutations similar to those that are recessively inherited.
How is it that some premature termination mutations cause
thalassaemia intermedia while others are clinically asympto-
matic in the heterozygous state? The answer appears to lie in
the differential effects of these in-phase termination mutants
on the accumulation of mutant mRNA. The stop codons and
frameshift mutations that are recessively inherited terminate in
exon 1 or 2, while those that are dominantly inherited,
terminate much later in the sequence of the b globin gene, in
exon 3 or beyond. Premature stop codons near the 3¢ end of
the gene, in exon 3 of the b gene, are less likely to trigger the
surveillance mechanism of nonsense mediated decay, leading
to an accumulation of the mutant b mRNA and to the
synthesis of truncated b chain variants (Maquat, 1995; Hentze
& Kulozik, 1999). These in-phase termination mutations

exemplify how shifting the position of a termination codon
can alter the phenotype of recessive inheritance caused by
haplo-insufficiency, to a dominant negative effect because of
the synthesis of an abnormal and deleterious protein.
The pathophysiology of these b chain variants relates to
their hyperinstability (Thein, 2001). The molecular mecha-
nisms include substitution of the critical amino acids in the
hydrophobic haem pocket displacing haem, leading to aggre-
gation of the globin variant; disruption of secondary structure
because of replacement of critical amino acids; substitution or
deletion of amino acids involved in ab dimer formation; and
elongation of subunits by a hydrophobic tail (Bunn & Forget,
1986). Again, a spectrum in phenotypic severity of this class of
b thalassaemia variants is observed, which can be related to
variation in the degree of instability of the b globin products
causing different degrees of ineffective erythropoiesis.
Mosaicism due to somatic deletion of b globin gene. This novel
mechanism was recently described in an individual who had
moderately severe thalassaemia intermedia despite being
constitutionally heterozygous for b thalassaemia with a
normal a genotype (Badens et al, 2002). Subsequent
investigations revealed that he had a somatic deletion of a
region of chromosome 11p15 including the b globin complex
giving rise to a mosaic of cells, 50% with one and 50% without
any b globin gene. The sum total of the b globin product is
approximately 25% less than the normally asymptomatic b
thalassaemia trait.
This unusual case once again illustrates that the severity of
anaemia of b thalassaemia reflects the defective b globin chain
production. Furthermore, with respect to potential gene
therapy, expression of a single b globin gene in a proportion
of the red blood cells appears to be sufficient to redress the
chain imbalance to produce a condition mild enough not to
need major medical intervention.
Secondary modifiers
The severity of anaemia in b thalassaemia reflects the degree of
globin chain imbalance and the excess of a globin chains with
all their deleterious effects on the red cell precursors. This
globin chain imbalance can be genetically modified by two
factors – variation in the amount of a globin production and
variation in fetal haemoglobin response.
a globin genotype. In many populations in which b
thalassaemia is prevalent, a thalassaemia also occurs at a
high frequency and hence it is not uncommon to co-inherit
both conditions. Homozygotes or compound heterozygotes for
b thalassaemia who co-inherit a thalassaemia will have less
redundant a globin and tend to have a less severe condition. As
with b thalassaemia, the different a thalassaemias that
predominate in different racial groups display a wide range
of severity. This interaction alone provides the basis for
considerable clinical heterogeneity; the degree of amelioration
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
Review
depends on the severity of the b thalassaemia alleles and the
number of functional a globin genes (Camaschella et al, 1995;
Ho et al, 1998). Co-inheritance of a single a gene deletion has
very little effect on the phenotype of b thalassaemia while
individuals with two a gene deletions and homozygous b+
thalassaemia may have a mild form of thalassaemia intermedia.
At the other extreme, patients who have co-inherited HbH
(equivalent of only one functioning a gene) and homozygous
b thalassaemia, also have thalassaemia intermedia.
Just as co-inheritance of a thalassaemia can reduce the
clinical severity of homozygous b thalassaemia, the presence of
increased a globin product in b thalassaemia heterozygotes tips
the globin chain imbalance further and crosses the threshold,
converting a typically clinical asymptomatic state to that of
thalassaemia intermedia. In the majority of cases, this is related
to the co-inheritance of triplicated a globin genes. Triplicated a
genes (aaa/) occur in most populations at a low frequency. The
co-inheritance of two extra a globin genes (aaa/aaa) or (aaaa/
aa) with heterozygous b thalassaemia results in the thalas-
saemia intermedia (Galanello et al, 1983; Thein et al, 1984).
However, the phenotype of a single extra a gene (aaa /aa) with
heterozygous b thalassaemia is more variable and depends on
the severity of the b thalassaemia allele (Traeger-Synodinos
et al, 1996; Camaschella et al, 1997). There appears to be a
critical threshold of globin chain imbalance in each individual,
above which clinical symptoms appear. This may be related to
the efficiency of the proteolytic mechanism of the erythroid
precursors and/or perhaps to the level of a-haemoglobin-
stabilizing protein, a chaperone of a globin (Kihm et al, 2002).
Variation in fetal haemoglobin production. The role of
increased Hb F response as an ameliorating factor becomes
evident in the group of homozygous b thalassaemia patients
who are not able to produce any haemoglobin A (a2b2) but yet
have a mild disease with a reasonable level of haemoglobin, all
of which is Hb F. Production of fetal haemoglobin after the
neonatal period in b thalassaemia is an extremely complex
process and still poorly understood. There appears to be a
genuine increase in c chain synthesis, presumably reflecting the
expansion of the ineffective erythroid mass. The effect is
augmented by the selective survival of the erythroid precursors
that synthesize relatively more c chains. Hence all b
thalassaemias, heterozygous or homozygous, have variable
increases in their levels of Hb F. Against this background, there
are undoubtedly genetic factors involved. Studies have shown
that approximately 90% of the variation in the level of Hb F
and F cells (subset of erythrocytes that contain Hb F) is
genetically controlled (Garner et al, 2000a). About one-third of
the genetic variance is the result of a common genetic variant,
C-T at position )158 of the Gc globin gene, also referred to as
the Xmn1-Gc polymorphism, but more than 50% of the
genetic variance in F-cell levels is caused by factors not linked
to the b chromosome (Garner et al, 2000b).
The Xmn1-Gc site is common in all population groups and
is present at a frequency of 0Æ32–0Æ35 (Garner et al, 2000b).
269
Review
Unlike the rare mutations in the c globin promoter that are
consistently associated with large discrete effects of increased
Hb F levels of 10–35% in heterozygotes, the so-called
pancellular HPFH (Wood, 2001), the change at Gc )158 does
not always raise the Hb F levels in otherwise normal
individuals. The rare non-deletion HPFH mutations occur in
transcription factor binding motifs in the c promoters,
clustered in three regions, around positions )114 to )117, at
)175, and from )195 to )202, while the sequence in the )158
region is not a recognized binding motif for any of the known
transcription factors. Nonetheless, although it has little effect
in normal individuals, clinical studies have shown that, under
conditions of haemopoietic stress, for example in homozygous
b thalassaemia and sickle cell disease, presence of the Xmn1-Gc
site favour a higher Hb F response (Labie et al, 1985; Thein
et al, 1987). This could explain why the same mutations on
different b chromosomal backgrounds (some with and others
without the Xmn1-Gc site) are associated with different
clinical severity. The increased Hb F output observed in
deletions or mutations that involve the promoter sequence of
the b globin gene reflect the competition between the c and b
globin gene promoters for interaction with the LCR or rate-
limiting transcription factors. Hence, although such deletions
cause a complete absence of b globin product, the severity of
the phenotype is offset by the concomitant increase in Hb F
(Craig et al, 1992).
Although presence of the cis Xmn1-Gc site is a modulating
factor, clearly there are some patients who have enhanced Hb F
response despite being Xmn1-Gc)/) (Galanello et al, 1989;
Ho et al, 1998). In many cases, family studies have shown that
there is an inherent capacity for producing Hb F and that the
genetic determinant is not linked to the b globin cluster
(Gianni et al, 1983; Thein & Weatherall, 1989). This is in
keeping with our genetic studies, which showed that >50% of
the F-cell variance in the general population is accounted for
by trans-acting factors. Indeed, linkage studies have mapped
loci controlling Hb F and F-cell levels to three regions of the
genome – chromosome 6q23 (Craig et al, 1996), Xp22 (Dover
et al, 1992) and 8q11 (Garner et al, 2002). The effects of the 8q
locus are conditional on the Xmn1-Gc site (Garner et al,
2002). As the genetic basis of the propensity to produce Hb F
becomes unravelled it is becoming clear that the conglomer-
ation of the Xmn1-Gc polymorphism, the quantitative trait
loci on 6q, Xp and 8q and others, linked and unlinked to the b
globin complex, contribute to the quantitative trait of Hb F
that constitute the loosely defined syndrome of heterocellular
HPFH (Thein & Craig, 1998). Until the different entities
become better defined, detection of an inherent capacity for
increased Hb F production is, at present, difficult and usually
inferred from family studies.
Tertiary modifiers
With the increasing lifespan of the b thalassaemia patients,
subtle variations in the phenotype with regard to some of the
270 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
complications have become apparent and evidence suggests
that they may be affected by genetic variants.
Hyperbilirubinaemia and a propensity to gallstone forma-
tion is a common complication of b thalassaemia and is
attributed to the rapid turnover of the red blood cells, bilirubin
being a breakdown product of haemoglobin. Varying degrees
of jaundice have often been observed in the thalassaemia
syndromes – from thalassaemia trait through to thalassaemia
major (Galanello et al, 1997; Sampietro et al, 1997; Galanello
et al, 2001). Studies have shown that the levels of bilirubin and
the incidence of gallstones in b thalassaemia is related to a
polymorphic variant (seven TA repeats) in the promoter of the
uridine diphosphate-glucoronyltransferase IA (UGTIA) gene,
also referred to as Gilbert’s syndrome. In vitro studies indicate
the variant causes a reduced expression of the UGTIA gene
(Bosma et al, 1995). Normal individuals who are homozygous
for the [TA]7 variant instead of the usual six, tend to have
higher levels of bilirubin (Bosma et al, 1995). The [TA]7
variant has also been shown to be associated with increased
bilirubin levels in sickle cell disease and other haemolytic
anaemia (Passon et al, 2001).
A common complication of b thalassaemia involves organ
damage from iron overload, not just from blood transfusions
but also from increased absorption. While the C282Y mutation
in the HFE1 gene, which causes hereditary haemochromatosis,
predisposes to iron loading in thalassaemia intermedia (Rees
et al, 1997), the co-existence of b thalassaemia trait aggravates
and accentuates iron loading in C282Y HFE homozygotes
(Piperno et al, 2000). As the C282Y mutation is rare in
populations in which b thalassaemia is common it has a
limited role in iron loading among these patients (Merry-
weather-Clarke et al, 1997). Much more common is the HFE
gene polymorphism, H63D, whose functional role is still being
investigated. Nonetheless, a recent study showed that b
thalassaemia carriers who are homozygotes for H63D have
higher serum ferritin levels than carriers without the poly-
morphism, suggesting that the H63D polymorphism may have
a modulating effect on iron absorption (Melis et al, 2002). As
other genes in iron homeostasis become uncovered, it is likely
there will be genetic variants in these loci that influence the
different degrees of iron loading in b thalassaemia (Andrews,
2000, 2002).
Progressive osteoporosis and osteopenia is another increas-
ingly common complication encountered in young adults with
b thalassaemia (Wonke, 1998). Several studies suggest that the
prevalence of bone disease in b thalassaemia is higher in men
than women, and that it is more severe in the spine than
femoral neck (Jensen et al, 1998; Dresner Pollack et al, 2000;
Perrotta et al, 2000). It is manifested by diffuse bone pain,
particularly in the lower back, vertebral fractures, cord
compression, spontaneous fractures and femoral head necrosis.
Bone disease is defined by reduced bone mineral density
(BMD), as measured by dual-energy X-ray absorptiometry
(DXA). Osteoporosis is present when the BMD is <2Æ5 SD of
the mean value in young adults (T score <)2Æ5), while

osteopenia is defined by a T score between )1 and )2Æ5. Bone
mass is determined by a combination of genetic and environ-
mental factors. Anaemia and bone marrow expansion, which is
prevalent in b thalassaemia, is a major contributor in
inadequately treated patients. The osteoporosis that occurs in
other patients who are reasonably well transfused but in which
there is severe iron loading, may be related to hypogonadism
(Anapliotou et al, 1995). However, it has become apparent
from recent studies that bone disease is increasingly common
even among the well-transfused and iron-chelated patients
(Wonke, 1998) and may be related to desferrioxamine toxicity.
Studies have suggested that, all things being equal, some
patients may be more susceptible to bone disease than others.
Bone mass is another quantitative trait known to be under
strong genetic control involving multiple loci (Pocock et al,
1987; Slemenda et al, 1991; Soroko et al, 1994). The genetic
loci implicated include oestrogen receptor gene, vitamin D
receptor (VDR), collagen type a1 (COL1A1) and COL1A2
genes, and transforming growth factor b1 (TGFb1). Poly-
morphism in TGFb1 has been associated with severe osteo-
porosis and increased bone turnover in women (Bertoldo et al,
2000), but a study of TGFb1 polymorphism failed to
demonstrate a statistical difference in b thalassaemia patients
with different BMDs. The VDR gene polymorphism in intron 8
(involving the Bsm1 site) was associated with osteopenia in
3 thalassaemia (Dresner Pollack et al, 2000). A G fi T poly-
morphism involving an Sp1 binding site in the COL1A1 gene
was strongly associated with reduced bone mass and osteo-
porosis (Grant et al, 1996). This same polymorphism in the
COL1A1 gene has also been shown to be strongly associated
with reduced BMD and osteoporosis in b thalassaemia
(Perrotta et al, 2000).
Cardiac complications are the main cause of death in b
thalassaemia. Many aspects of cardiac complications are still
poorly understood, and again it is clear that cardiac disease in
b thalassaemia is multifactorial, reflecting the chronic anaemia,
iron overload and pulmonary hypertension. The situation is
further confounded by the unpredictability of the severity of
cardiac iron deposition. A risk factor for left ventricular failure
in thalassaemia is decreased antioxidant activity of apolipo-
protein E4, related to the frequency of the apolipoprotein E4
allele (Economou-Peterson et al, 1998).
Genetic variants implicated in other complications of b
thalassaemia include specific human leucocyte antigen alleles
in the tendency to hepatitis and liver cirrhosis; genetic variants
in Factor V, prothrombin and methylenetetrahydrofolate
reductase, and the tendency to thrombosis. For a superb
review of the hypercoagulable state in thalassaemia and the
possible underlying genetic and environmental contribution,
the reader should consult Eldor and Rachmilewitz (2002).
Conclusion
There is a spectrum of phenotypes in b thalassaemia, the
severity of which relates directly to the degree of chain
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
Review
imbalance and the a globin excess. Much of the variation can
be explained by heterogeneity of the molecular lesions affecting
the b globin gene itself but it is also clear that variability at the
two loci – a and c globin genes – is important in determining
the phenotype, which is extremely encouraging for genetic
counselling. However, while genotyping at the b globin and a
globin loci is relatively easy to incorporate into the prenatal
diagnosis and counselling programme, detecting an inherent
ability to increase Hb F in response to haemopoietic stress is
still difficult. The presence of such heterocellular HPFH
determinants is usually implicated from studies of family
members who are often not available. Until the quantitative
trait loci for Hb F are better defined, it would appear that it is
still not possible to consistently predict phenotype from
genotype apart from the two categories of extra a globin genes
with heterozygous b thalassaemia, and the inheritance of
mild b+ thalassaemia alleles. Ethnicity and environment are
important factors in the analysis of genotype/phenotype
relationships. Studies have shown that all three categories of
genetic modifiers – primary, secondary and tertiary – are
population-specific. The tertiary loci include the many
different genetic polymorphisms that form the background
genes, some of which have been co-selected with the
thalassaemias.
Acknowledgments
I thank Claire Steward for help in preparation of the
manuscript and Dr Barnaby Clark for critical reading of the
manuscript.
Swee Lay Thein
Department of Haematological Medicine Guy’s, King’s and St Thomas’
School of Medicine, Denmark Hill Campus, Bessemer Road, London SE5
9PJ, UK
References
Anapliotou, M.L., Kastanias, I.T., Psara, P., Evangelou, E.A., Liparaki,
M. & Dimitriou, P. (1995) The contribution of hypogonadism to the
development of osteoporosis in thalassaemia major: new therapeutic
approaches. Clinical Endocrinology, 42, 279–287.
Andrews, N. (2000) Iron homeostasis: insights from genetics and
animal models. Nature Reviews Genetics, 1, 208–216.
Andrews, N.C. (2002) A genetic view of iron homeostasis. Seminars in
Hematology, 39, 227–234.
Antoniou, M., deBoer, E., Habets, G. & Grosveld, F. (1988) The human
b-globin gene contains multiple regulatory regions: identification of
one promoter and two downstream enhancers. EMBO Journal, 7,
377–384.
Badens, C., Mattei, M.G., Imbert, A.M., Lapoumérouliee, C., Martini,
N., Michel, G. & Lena-Russo, D. (2002) A novel mechanism for
thalassaemia intermedia. The Lancet, 359, 132–133.
Bertoldo, F., D’Agruma, L., Furlan, F., Colapietro, F., Lorenzi, M.T.,
Maiorano, N., Iolascon, A., Zelante, L., Locascio, V. & Gasparini, P.
(2000) Transforming growth factor-beta1 gene polymorphism, bone
271
Review
turnover, and bone mass in Italian postmenopausal women. Journal
of Bone and Mineral Research, 15, 634–639.
Bosma, P.J., Chowdhury, J.R., Bakker, C., Gantla, S., De Boer, A.,
Oostra, B.A., Lindhout, D., Tytgat, G.N.J., Jansen, P.L.M., Elferink,
R.P.J.O. & Chowdhury, N.R. (1995) The genetic basis of the
reduced expression of bilirubin UCP-glucuronosultransferase 1 in
Gilbert’s syndrome. New England Journal of Medicine, 333, 1171–
1175.
Boyer, S.H., Belding, T.K., Margolet, L. & Noyes, A.N. (1975) Fetal
hemoglobin restriction to a few erythrocytes (F cells) in normal
human adults. Science, 188, 361–363.
Bulger, M., Bender, M.A., van Doorninck, J.H., Wertman, B., Farrell,
C.M., Felsenfeld, G., Groudine, M. & Hardison, R. (2000) Com-
parative structural and functional analysis of the olfactory receptor
genes flanking the human and mouse beta-globin gene clusters.
PNAS, 97, 14560–14565.
Bunn, H.F. & Forget, B.G. (1986) Hemoglobin: Molecular, Genetic and
Clinical Aspects. W.B. Saunders Company, Philadelphia, PA.
Camaschella, C., Maza, U., Roetto, A., Gottardi, E., Parziale, A., Travi,
M., Fattore, S., Bacchiega, D., Fiorelli, G. & Cappellini, M.D. (1995)
Genetic interactions in thalassemia intermedia: analysis of beta-
mutations, alpha-genotype, gamma-promoters, and beta-LCR
hypersensitive sites 2 and 4 in Italian patients. American Journal of
Hematology, 48, 82–87.
Camaschella, C., Kattamis, A.C., Petroni, D., Roetto, A., Sivera, P.,
Sbaiz, L., Cohen, A., Ohene-Frempong, K., Trifillis, P., Surrey, S. &
Fortina, P. (1997) Different hematological phenotypes caused by the
interaction of triplicated alpha-globin genes and heterozygous beta-
thalassemia. American Journal of Hematology, 55, 83–88.
Carter, D., Chakalova, L., Osborne, C.S., Dai, Y.F. & Fraser, P. (2002)
Long-range chromatin regulatory interactions in vivo. Nature
Genetics, 32, 623–626.
Collins, F.S. & Weissman, S.M. (1984) The molecular Genetics of
human hemoglobin. Progress in Nucleic Acid Research and Molecular
Biology, 31, 315–465.
Craig, J.E., Kelly, S.J., Barnetson, R. & Thein, S.L. (1992) Molecular
characterization of a novel 10.3 kb deletion causing beta-thalas-
saemia with unusually high Hb A2. British Journal of Haematology,
82, 735–744.
Craig, J.E., Rochette, J., Fisher, C.A., Weatherall, D.J., Marc, S.,
Lathrop, G.M., Demenais, F. & Thein, S.L. (1996) Dissecting the loci
controlling fetal haemoglobin production on chromosomes 11p and
6q by the regressive approach. Nature Genetics, 12, 58–64.
Divoky, V., Indrak, K., Mrug, M., Brabec, V., Huisman, T.H.J. &
Prchal, J.T. (1996) A novel mechanism of ß thalassemia: The
insertion of L1 retrotransposable element into ß globin IVS II. Blood,
88, 148a.
Dover, G.J., Smith, K.D., Chang, Y.C., Purvis, S., Mays, A., Meyers,
D.A., Sheils, C. & Serjeant, G. (1992) Fetal hemoglobin levels in
sickle cell disease and normal individuals are partially controlled by
an X-linked gene located at Xp22.2. Blood, 80, 816–824.
Dresner Pollack, R., Rachmilewitz, E., Blumenfeld, A., Idelson, M. &
Goldfarb, A.W. (2000) Bone mineral metabolism in adults with
beta-thalassaemia major and intermedia. British Journal of Haema-
tology, 111, 902–907.
Economou-Peterson, E., Aesspopos, A., Kladi, A., Flevari, P., Kar-
abatsos, F., Fragodimitri, C., Nicolaidis, P., Vrettou, H., Vassilo-
poulos, D., Karagiorga-Lagana, M., kremastinos, D.T. & Petersen,
M.B. (1998) Apolipoprotein E epsilon4 allele as a genetic risk factor
272 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
for left ventricular failure in homozygous beta-thalassemia. Blood,
92, 3455–3459.
Efstratiadis, A., Posakony, J.W., Maniatis, T., Lawn, R.M., O’Connell,
C., Spritz, R.A., DeRiel, J.K., Forget, B.G., Weissman, S.M., Sligh-
tom, J.L., Blechl, A.E., Smithies, O., Baralle, F.E., Shoulders, C.C. &
Proudfoot, N.J. (1980) The structure and evolution of the human
beta-globin gene family. Cell, 21, 653–668.
Eldor, A. & Rachmilewitz, E.A. (2002) The hypercoagulable state in
thalassemia. Blood, 99, 36–43.
Fei, Y.J., Stoming, T.A., Kutlar, A., Huisman, T.H.J. & Stama-
toyannopoulos, G. (1989) One form of inclusion body ß-thalassemia
is due to a GAA->TAA mutation at codon 121 of the ß chain. Blood,
73, 1075–1077.
Flint, J., Harding, R.M., Boyce, A.J. & Clegg, J.B. (1998) The popula-
tion genetics of the haemoglobinopathies. In: Bailliére’s Clinical
Haematology (ed. by G.P. Rodgers), Vol. 11, pp. 1–52. Bailliere
Tindall, London.
Galanello, R., Ruggeri, R., Paglietti, E., Addis, M., Melis, M.A. & Cao,
A. (1983) A family with segregating triplicated alpha globin loci and
ß thalassaemia. Blood, 62, 1035–1040.
Galanello, R., Dessi, E., Melis, M.A., Addis, M., Sanna, M.A., Rosatelli,
C., Argiolu, F., Giagu, N., Turco, M.P., Cacace, E., Pirastu, M. &
Cao, A. (1989) Molecular analysis of beta zero-thalassemia inter-
media in Sardinia. Blood, 74, 823–827.
Galanello, R., Perseu, L., Melis, M.A., Cipollina, L., Barella, S., Giagu,
N., Turco, M.P., Maccioni, O. & Cao, A. (1997) Hyper-
bilirubinaemia in heterozygous beta-thalassaemia is related to
co-inherited Gilbert’s syndrome. British Journal of Haematology, 99,
433–436.
Galanello, R., Piras, S., Barella, S., Leoni, G.B., Cipollina, M.D., Perseu,
L. & Cao, A. (2001) Cholelithiasis and Gilbert’s syndrome in
homozygous beta-thalassaemia. British Journal of Haematology, 115,
926–928.
Garner, C., Tatu, T., Reittie, J.E., Littlewood, T., Darley, J., Cervino, S.,
Farrall, M., Kelly, P., Spector, T.D. & Thein, S.L. (2000a) Genetic
influences on F cells and other hematological variables: a twin
heritability study. Blood, 95, 342–346.
Garner, C., Tatu, T., Game, L., Cardon, L.R., Spector, T.D., Farrall, M.
& Thein, S.L. (2000b) A candidate gene study of F cell levels in
sibling pairs using a joint linkage and association analysis. GeneSc-
reen, 1, 9–14.
Garner, C.P., Tatu, T., Best, S., Creary, L. & Thein, S.L. (2002) Evi-
dence for Genetic Interaction between the beta-globin complex and
chromosome 8q in the expression of fetal hemoglobin. American
Journal of Human Genetics, 70, 793–799.
Gianni, A.M., Bregni, M., Cappellini, M.D., Fiorelli, G., Taramelli, R.,
Giglioni, B., Comi, P. & Ottolenghi, S. (1983) A gene controlling
fetal hemoglobin expression in adults is not linked to the non-alpha
globin cluster. EMBO Journal, 2, 921–925.
Gonzalez-Redondo, J.M., Stoming, T.A., Kutlar, A., Kutlar, F., Lanclos,
K.D., Howard, E.F., Fei, Y.J., Aksoy, M., Altay, C., Gurgey, A., Basak,
A.N., Efremov, G.D., Petkov, G. & Huisman, T.H.J. (1989) A
C fi T substitution at nt-101 in a conserved DNA sequence of the
promoter region of the ß-globin gene is associated with ‘silent’ ß-
thalassemia. Blood, 73, 1705–1711.
Grant, S.F., Reid, D.M., Blake, G., Herd, R., Fogelman, I. & Ralston,
S.H. (1996) Reduced bone density and osteoporosis associated with
a polymorphic Sp1 binding site in the collagen type I alpha 1 gene.
Nature Genetics, 14, 203–205.

Hall, G.W. & Thein, S.L. (1994) Nonsense codon mutations in the
terminal exon of the ß-globin gene are not associated with a re-
duction in ß-mRNA accumulation: a mechanism for the phenotype
of dominant ß-thalassemia. Blood, 83, 2031–2037.
Hentze, M.W. & Kulozik, A.E. (1999) A perfect message: RNA sur-
veillance and nonsense-mediated decay. Cell, 96, 307–310.
Ho, P.J., Wickramasinghe, S.N., Rees, D.C., Lee, M.J., Eden, A. &
Thein, S.L. (1997) Erythroblastic inclusions in dominantly inherited
ß thalassaemias. Blood, 89, 322–328.
Ho, P.J., Hall, G.W., Luo, L.Y., Weatherall, D.J. & Thein, S.L. (1998)
Beta thalassemia intermedia: is it possible to consistently predict
phenotype from genotype? British Journal of Haematology, 100, 70–
78.
Huisman, T.H.J. (1997) Levels of Hb A2 in heterozygotes and homo-
zygotes for beta-thalassemia mutations: Influence of mutations in
the CACCC and ATAAA motifs of the beta-globin gene promoter.
Acta Haematologica, 98, 187–194.
Jensen, C.E., Tuck, S.M., Agnew, J.E., Koneru, S., Morris, R.W.,
Yardumian, A., Prescott, E., Hoffbrand, A.V. & Wonke, B. (1998)
High prevalence of low bone mass in thalassaemia major. British
Journal of Haematology, 103, 911–915.
Kihm, A.J., Kong, Y., Hong, W., Russell, J.E., Rouda, S., Adachi, K.,
Simon, M.C., Blobel, G.A. & Weiss, M.J. (2002) An abundant ery-
throid protein that stabilizes free alpha-haemoglobin. Nature, 417,
758–763.
Labie, D., Pagnier, J., Lapoumeroulie, C., Rouabhi, F., Dunda-Bel-
khodja, O., Chardin, P., Beldjord, C., Wajcman, H., Fabry, M.E.
& Nagel, R.L. (1985) Common haplotype dependency of high G
gamma-globin gene expression and high Hb F levels in beta-
thalassemia and sickle cell anemia patients. Proceedings of the National
Academy of Sciences of the United States of America, 82, 2111–2114.
Maquat, L.E. (1995) When cells stop making sense: effects of nonsense
codons on RNA metabolism in vertebrate cells. RNA, 1, 453–465.
Maragoudaki, E., Kanavakis, E., Trager-Synodinos, J., Vrettou, C.,
Tzetis, M., Metxotou-Mavrommati, A. & Kattamis, C. (1999) Mo-
lecular, haematological and clinical studies of the -101 C fi T
substitution in the b-globin gene promoter in 25 b-thalassaemia
intermedia patients and 45 heterozygotes. British Journal of Hae-
matology, 107, 699–706.
Melis, M.A., Cau, M., Deidda, F., Barella, S., Cao, A. & Galanello, R.
(2002) H63D mutation in the HFE gene increases iron overload in
beta-thalassemia carriers. Haematologica, 87, 242–245.
Merryweather-Clarke, A.T., Pointon, J.J., Shearman, J.D. & Robson,
K.J.H. (1997) Global prevalence of putative haemochromatosis
mutations. Journal of Medical Genetics, 34, 275–278.
Passon, R.G., Howard, T.A., Zimmerman, S.A., Schultz, W.H. & Ware,
R.E. (2001) Influence of bilirubin uridine diphosphate-glucur-
onosyltransferase 1A promoter polymorphisms on serum bilirubin
levels and cholelithiasis in children with sickle cell anemia. American
Journal of Pediatric Hematology and Oncology, 23, 448–451.
Perrotta, S., Cappellini, M.D., Bertoldo, F., Servedio, V., Iolascon, G.,
D’Agruma, L., Gasparini, P., Siciliani, M.C. & Iolascon, A. (2000)
Osteoporosis in beta-thalassaemia major patients: analysis of the
genetic background. British Journal of Haematology, 111, 461–466.
Piperno, A., Mariani, R., Arosio, C., Vergani, A., Bosio, S., Fargion, S.,
Sampietro, M., Girelli, D., Fraquelli, M., Conte, D., Fiorelli, G. &
Camaschella, C. (2000) Haemochromatosis in patients with beta-
thalassaemia trait. British Journal of Haematology, 111, 908–914.
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
Review
Pocock, N.A., Eisman, J.A., Hopper, J.L., Yeates, M.G., Sambrook, P.N.
& Eberl, S. (1987) Genetic determinants of bone mass in adults. A
twin study. Journal of Clinical Investigation, 80, 706–710.
Rees, D.C., Luo, L.Y., Thein, S.L., Sing, B.M. & Wickramasinghe, S.
(1997) Nontransfusional iron overload in thalassemia: Association
with hereditary hemochromatosis. Blood, 90, 3234–3236.
Rund, D., Filon, D., Strauss, N., Rachmilewitz, E.A. & Oppenheim, A.
(1991) Mean corpuscular volume of heterozygotes for ß-thalassemia
correlates with the severity of mutations. Blood, 79, 238–243.
Sampietro, M., Lupica, L., Perrero, L., Comino, A. & Martinez di
Montemuros, F. (1997) The expression of uridine diphosphate
glucuronosyltransferase gene is a major determinant of bilirubin
level in heterozygous beta-thalassaemia and in glucose-6-phosphate.
British Journal of Haematology, 99, 437–439.
Schokker, R.C., Went, L.N. & Bok, J. (1966) A new genetic variant of
ß-thalassaemia. Nature, 209, 44–46.
Schrier, S.L. (2002) Pathophysiology of thalassemia. Current Opinion
in Hematology, 9, 123–126.
Skarmoutsou, C., Papassotiriou, I., Traeger-Synodinos, J., Stamou, H.,
Ladis, V., Metaxotou-Mavrommati, A., Stamoulakatou, A. & Ka-
navakis, E. (2003) Erythroid bone marrow activity and red cell he-
moglobinization in iron sufficient beta-thalassemia heterozygotes as
reflected by soluble transferrin receptor and reticulocyte hemoglobin
in content. Correlation with genotypes and Hb A2 levels. Haema-
tologica, 88, 631–636.
Slemenda, C.W., Christian, J.C., Williams, C.J., Norton, J.A. & John-
ston Jr, C.C. (1991) Genetic determinants of bone mass in adult
women: a re-evaluation of the twin model and the potential im-
portance of gene interaction on heritability estimates. Journal of
Bone and Mineral Research, 6, 561–567.
Soroko, S.B., Barrett-Connor, E., Edelstein, S.L. & Kritz-Silverstein, D.
(1994) Family history of osteoporosis and bone mineral density at
the axial skeleton: the Rancho Bernardo Study. Journal of Bone and
Mineral Research, 9, 761–769.
Thein, S.L. (1992) Dominant beta thalassaemia: molecular basis and
pathophysiology. British Journal of Haematology, 80, 273–277.
Thein, S.L. (1998) Beta thalassaemia. In: Baillière’s Clinical Haema-
tology, Sickle Cell Disease and Thalassaemia (ed. by G.P. Rodgers),
Vol. 11:1, pp. 91–126. Baillière Tindall, London.
Thein, S.L. (1999) Is it dominantly inherited beta thalassaemia or just a
beta-chain variant that is highly unstable? British Journal of Hae-
matology, 107, 12–21.
Thein, S.L. (2001) Structural variants with a beta-thalassemia pheno-
type. In: Disorders of Hemoglobin: Genetics, Pathophysiology, and
Clinical Management (ed. by M.H. Steinberg, B.G. Forget, D.R.
Higgs & R.L. Nagel), pp. 342–355. Cambridge University Press,
Cambridge, UK.
Thein, S.L. & Craig, J.E. (1998) Genetics of Hb F/F cell variance in
adults and heterocellular hereditary persistence of fetal hemoglobin.
Hemoglobin, 22, 401–414.
Thein, S.L. & Weatherall, D.J. (1989) A non-deletion hereditary per-
sistance of fetal hemoglobin (HPFH) determinant not linked to the
beta-globin gene complex. In: Hemoglobin Switching, Part B: Cellular
and Molecular Mechanisms (ed. by G. Stamatoyannopoulos & A.W.
Nienhuis), pp. 97–111. Alan R Liss, Inc., New York, NY.
Thein, S.L., Al-Hakim, I. & Hoffbrand, A.V. (1984) Thalassaemia
intermedia: a new molecular basis. British Journal of Haematology,
56, 333–337.
273
Review
Thein, S.L., Wainscoat, J.S., Sampietro, M., Old, J.M., Cappellini, D.,
Fiorelli, G., Modell, B. & Weatherall, D.J. (1987) Association of
thalassaemia intermedia with a beta-globin gene haplotype. British
Journal of Haematology, 65, 367–373.
Tolhuis, B., Palstra, R.J., Splinter, E., Grosveld, F. & de Laat, W. (2002)
Looping and interaction between hypersensitive sites in the active
beta-globin locus. Molecular Cell, 10, 1453–1465.
Traeger-Synodinos, J., Kanavakis, E., Vrettou, C., Maragoudaki, E.,
Michael, T. & Metaxotou-Mavromati, A. (1996) The triplicated al-
pha-globin gene locus in beta-thalassaemia heterozygotes: clinical,
haematological, biosynthetic and molecular studies. British Journal
of Haematology, 95, 467–471.
Trudel, M. & Costantini, F. (1987) A 3¢enhancer contributes to the
stage-specific expression of the human b-globin gene. Genes and
Development, 1, 954–961.
Viprakasit, V., Gibbons, R.J., Broughton, B.C., Tolmie, J.L., Brown, D.,
Lunt, P., Winter, R.M., Marinoni, S., Stefanini, M., Brueton, L.,
Lehmann, A.R. & Higgs, D.R. (2001) Mutations in the general
transcription factor TFIIH result in beta-thalassaemia in individuals
with trichothiodystrophy. Human Molecular Genetics, 10, 2797–2802.
Weatherall, D.J. & Clegg, J.B. (2001) The Thalassaemia Syndromes, 4th
edn. Blackwell Science, Oxford.
274 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 124, 264–274
Weatherall, D.J., Clegg, J.B., Knox-Macaulay, H.H.M., Bunch, C.,
Hopkins, C.R. & Temperley, I.J. (1973) A genetically determined
disorder with features both of thalassaemia and congenital dysery-
thropoietic anaemia. British Journal of Haematology, 24, 681–702.
Wong, C., Dowling, C.E., Saiki, R.K., Higuchi, R.G., Erlich, H.A. &
Kazazian, H.H.J. (1987) Characterization of beta-thalassaemia
mutations using direct genomic sequencing of amplified single copy
DNA. Nature, 330, 384–386.
Wonke, B. (1998) Bone disease in beta-thalassaemia major. British
Journal of Haematology, 103, 897–901.
Wood, W.G. (2001) Hereditary persistence of fetal hemoglobin and
delta beta thalassemia. In: Disorders of Hemoglobin: Genetics, Path-
ophysiology, and Clinical Management (ed. by M.H. Steinberg, B.G.
Forget, D.R. Higgs & R.L. Nagel), pp. 356–388. Cambridge Uni-
versity Press, Cambridge, UK.
Yu, C., Niakan, K.K., Matsushita, M., Stamatoyannopoulos, G., Orkin,
S.H. & Raskind, W.H. (2002) X-linked thrombocytopenia with
thalassemia from a mutation in the amino finger of GATA-1 af-
fecting DNA binding rather than FOG-1 interaction. Blood, 100,
2040–2045.
Keywords: genotype-phenotype correlation, b thalassaemia.