Advances In Thalassemia Research

738 Blood, Vol. 63, No. 4 (April), 1984: pp. 738-758
REVIEW
Advances in Thalassemia Research
By Arthur W. Nienhuis, Nicholas P. Anagnou, and Timothy J. Ley
T HE THALASSEMIAS are hereditary hemolytic
anemias characterized by decreased or absent
synthesis of one of the globin subunits of the hemogbo-
bin molecule.’ ln the a-thalassemias, decreased syn-
thesis of a-globin results in accelerated red cell
destruction because of the formation of insoluble HbH
(f34) inclusions in mature red cells. The greater clinical
severity of the fl-thalassemias reflects the extreme
insolubility of a-globin, which is present in relative
excess because of decreased 3-globin synthesis. Insolu-
ble a-globin precipitates in developing erythroblasts,
leading to marked ineffective erythropoiesis. Severe
disease does occur among the ct-thalassemias, but only
when no a-globin is produced. The thalassemias are
among the most common genetic diseases of man. The
high gene frequency results from a selective advantage
of the thalassemia phenotype in heterozygotes, where
it protects from severe malaria.
A quantitative decrease in globin synthesis without
a qualitative change in the gene product must reflect
an alteration in gene function or regulation. The recent
explosion in our understanding about gene organiza-
tion and structure has been matched by characteriza-
tion of several specific thalassemia mutations at the
level of DNA sequence. Many well studied examples of
mutations that alter specific steps in gene expression,
e.g., gene transcription, RNA processing, and mRNA
translation, will form the focus for this review.
The current status of several approaches to the
treatment of severe thalassemia will also be summa-
rized. Modifications of the transfusion regimen and
use of subcutaneous desferrioxamine have resulted in
improved conventional treatment, but the problem of
ultimately fatal iron overload may not yet be corn-
pletely solved. Bone marrow transplantation and phar-
macologic stimulation of l-IbF production are now
being explored as potentially useful, but still highly
experimental approaches, for treatment of severe 3-
thalassemia. Insertion ofnormal globin genes into bone
marrow stern cells of thalassernic individuals offers the
From the Clinical Hematology Branch, National Heart, Lung.
and Blood Institute, National Institutes ofHealth, Bethesda, MD.
Submitted October 7, 1 983; accepted December 5. 1983.
Address reprint requests to Dr. Arthur W. Nienhuis, Clinical
Hematology Branch, Building 10, Room 7C103, NHLBI. NIH.
Bethesda, MD 20205.
(C: J 984 by Grune & Stratton, Inc.
0006-4971/84/6304-0002$03.00/0
possibility of cure of this disease. Gene therapy is not
yet practical, although many successful strategies are
being employed to transfer genes into cells in tissue
culture; the application of these techniques to the
problem of thalassemia in the future seems quite
probable.
There have been major advances in the prenatal
diagnosis of the thalassemias and hernoglobinopathies
during the past several years. These advances are
beyond the scope of this review, but will be the subject
of detailed consideration in other issues of Blood.
ORGANIZATION AND STRUCTURE OF HUMAN
GLOBIN GENES
Organization
The human globin genes occur in multigene clusters;
the a-like genes are found on chromosome 16 and the
3-like genes on chromosome I I ,212 There are two fully
functional and nearly identical a and “y-genes on each
chromosome, reflecting their duplication during evolu-
tion and subsequent gene correction events (Fig. I).
The closely related t5 and fl-genes reflect gene duplica-
tion, but subsequent sequence “mutations” in the
regulatory region of the #{244}-gene
have rendered this gene
functionally insignificant. Therefore, there are only
two active /3-genes in each human diploid erythroid
cell. Both the a and fl-gene clusters include pseudo-
genes. These are thought to be byproducts of evolution,
representing previously functional genes in which
sequence alterations in coding or regulatory regions
have led to inactivation. The DNA between the mdi-
vidual globin genes includes copies of many repetitive
DNA sequence families, but no other functional genes
are known to occur within the a and /3-gene clusters.
Structure
A prototypical globin gene (Fig. 2B) includes three
coding blocks (exons) separated by two introns or
intervening sequences.7 ‘ Conserved sequences impor-
tant for gene function are found just before the mRNA
coding sequences, at the exon-intron boundaries, and
at the end of the mRNA coding sequences. The globin
gene promoter includes sequences at the 5’ end that are
essential for RNA polymerase binding and for accu-
rate and efficient initiation of RNA transcription.’4
Differences among the various fl-like globin genes with
respect to the arrangement of conserved sequences in
the promoter region (Fig. 2A) may be relevant to their
Downloaded
from
http://ashpublications.org/blood/article-pdf/63/4/738/1638660/738.pdf
by
guest
on
28
September
2021

l w aa2ai
a-IikeGenes 5’ #{149} 0-a I U 3
#{163} GyAy Hui 6 p
p -like Genes 5’ #{149} . . p #{149}
#{149} 3’ Chronxsome 11
0 10 20 30 40 50 6OKilobases
selective expression during development (discussed
Chromosome 16
below). The exon-intron boundary sequences are cru-
cial for the accurate removal of RNA transcribed from
the introns and for the subsequent precise splicing of
coding sequences to form a functional mRNA mole-
cule (Fig. 2C).’5 Transcription appears to continue
beyond the position on the DNA strand that corre-
sponds to the 3’ end of RNA coding sequences;’6
conserved sequences found at this position may be
involved in cutting the RNA transcript and the addi-
tion of adenosines to form the poly A tract.’7 The poly
A tract may be important for nuclear to cytoplasmic
mRNA transport and for mRNA stability in the
cytoplasm. 8
MOLECULAR MECHANISMS OF THALASSEMIA
The thalassemias must be considered with the reali-
zation that these are cis acting mutations.’ A cis acting
-.28
ATAAAA . EPSILON
-30
ATAAAA GAMMA
, ,,,, DELTA
POLYADENYLATION
SIGNAL
EXON 3
TIC) N CT AGG
-115 -83
CTCCACCC CCAAT
-148 -115 -88
/:
S. I
CONSENSUS SPLICE CIAGOT G/AAGT
SEOUENCES
C. GENE EXPRESSION
. 1 TRANSCRIPTION
THALASSEM IA 739
Fig. 1 . Map of the human globin gene clusters. The a-gene
cluster contains three functional genes. .a2. and a1, and two
pseudogenes. i/ a and 4’ ’. Such pseudogenes have sequence
homology to the functional genes but are defective in some
essential component. and thus. cannot be expressed in a globin
product. The fl-like gene cluster includes the functional genes. .
G,,, , A, , and 9. and one pseudogene, &fi. The and #{128}-genes
are
expressed during the first 3 mo of gestation. leading to synthesis
of Hb Gower 1 ( 22). The duplicated a and ‘y-genes are also
activated early in gestation; hence Hb Gower 2 (a22). Hb Portland
( 272). and HbF (a2”y2) may be found in fetal blood by the beginning
of the second trimester. Note that the two “y-globins differ by a
single amino acid; G7 encodes for glycine at position 136 and A,
encodes for alanine at that position. The $-gene also becomes
active very early in gestation. but the major switch from ‘y to /9
synthesis occurs in the perinatal period, when HbA (a2/92) becomes
the major hemoglobin.
A. HUMAN f3.UKE GLOBIN GENE PROMOTERS
-68
CCAAC - -----.-.
I -108 -93 -76 -31
I CCTCACCC CCACACCC CCAAT ATAAAA
LGP.THAL1, C G- THAL
I -- G=
B. PROTOTYPICAL GLOBIN GENE - -
PROMOTER EXON1INTRON I EXON 2
NUCLEUS
, CAPPING . .
2 RNA PROCESSING CAP POLY A ADDITION
3. TRANSPORT NUCLEAR MEMBRANE
4 TRANSLATION CAP iA.AAAA
CYTOPLASM
, GLOBA GLOBIN-” HEMOGLOBIN
Fig. 2. Structure and expression of the human /9-like genes. (A) Shown are the conserved blocks of DNA sequence in the promoter
region of the several genes; these are discussed in detail in the text. Solid boxes on the right of each line diagram indicate the position of
the coding sequences that are part of exon 1 . By convention. nucleotides upstream from the gene are numbered as minus the number from
the first transcribed nucleotide. Single nucleotide substitutions in the promoter region of four thalassemia genes have been described (see
text for discussion). (B) Shown is a diagram of the general structure of a /9-like globin gene. The three coding blocks (exons) are separated
by two introns. Exon 1 encodes for amino acids 1 -30, exon 2 for amino acids 31 -1 04, and exon 3 for amino acids 1 05-1 46. Intron I varies in
length from 1 25 to 1 30 nucleotides and intron II from 800 to 850 nucleotides. Sequences are found at the exon-intron boundaries that are
important for accurate and efficient splicing; shown are the consensus sequences derived by comparing the exon-intron boundary
sequences of more than 1 00 genes. The polyadenylation signal is thought to be required for correct cutting and addition of the poly A track,
although transcription apparently continues beyond the portion of the gene represented by the primary RNA transcript. (C) Shown are the
steps in the flow of information from gene to protein. The primary RNA transcript is processed by capping of the 5’ end. splicing to remove
RNA sequences transcribed from the introns. and by polyadenylation of the 3’ end. Transport to the cytoplasm may be subject to specific
regulation. Translation of the processed and transported mRNA in the cytoplasm is on polyribosomes (one mRNA and several ribosomes).
leading to synthesis of a globin product that associates with a-globin and hemin. spontaneously leading to hemoglobin formation.
Downloaded
from
by
guest
on
28
September
2021

740 NIENHUIS, ANAGNOU, AND LEY
mutation affects only the gene on the chromosome on
which the mutation occurs; a trans acting mutation
influences expression of both genes at a chromosomal
locus. The output of only one of the two f -genes in
erythroid cells is affected by a specific thalassemia
mutation. For example, in the cells of individuals who
are doubly heterozygous for a sickle (flS) and a thalas-
semia gene, total /3S globin synthesis is normal or even
increased, whereas fl-globin synthesis is decreased
(fl+) or absent (fl#{176})
As a general rule, cis acting
mutations affect gene structure directly, rather than
altering the function of another molecule involved in
gene expression. Indeed, all thalassemia genes that
have been carefully studied to date have been found to
contain mutations that directly alter gene structure
and, thereby, gene function. Gene deletion accounts
for many, but not all, a-thalassemia mutations. The
a-genes are apparently predisposed toward deletion
because of tandemly duplicated sequences that occur
in this cluster.4’5 Single nucleotide substitutions (or
deletion ofone or a few nucleotides) in otherwise intact
genes account for the vast majority of fl-thalassemia
mutations. The only exception, other than the exten-
sive deletions discussed below, is a mutation character-
ized by deletion of the 3’ portion of the /3-globin gene
and adjacent flanking sequences, which is observed in
a few patients of Indian ancestry.’92’
Promoter Mutations
The promoter of a gene is defined as those DNA
sequences that are essential for accurate and efficient
initiation of transcription. Promoter sequences are
found near the site at which initiation of transcription
begins and function in only a single orientation with
respect to the coding sequences. Comparison of DNA
sequences found just upstream to the coding portions
of the a and fl-globin genes of several species provided
the first clues as to the nature of DNA sequences that
function as a promoter.7”4 A conserved block of nucleo-
tides is found at a position just before the site corre-
sponding to the 5’ end of mRNA, and several addi-
tional blocks are found within 150 nucleotides
upstream. The position of the conserved blocks for the
human fl-like genes is shown in Fig. 2A. A conserved
block of adenine-thymine (AT)-rich sequences,
referred to as the TATA box, begins 28-3 1 basepairs
(bp) upstream from the modified methylated 5’ end of
mRNA (CAP) site (Fig. 2A). Forty-six to 58 nucleo-
tides upstream is a second conserved block that
includes the pentanucleotide segment, CCAAT, lead-
ing to the designation of this sequence as the “CAT”
box. Significant evolutionary divergence has occurred
in the DNA sequences between these conserved bhcks.
Note that the “CAT” box and I 8 nucleotides of
surrounding DNA sequence are duplicated in the
“y-gene promoter. The f3-globin gene promoter also
includes tandemly duplicated conserved blocks located
86 and 108 nucleotides upstream (Fig. 2A). This
sequence is not found upstream from the #{244}-gene
and is
found only once upstream from the ‘y and c-genes. The
a-globin gene promoter also lacks this conserved block
of nucleotides. The striking differences among the
sequences of the e, ‘y, and fl-globin gene promoters
invite the hypothesis that promoter structure is impor-
tant for developmental regulation of globin gene
expression.
Simple sequence comparison could not lead to iden-
tification of all sequences important for promoter
function. Hence, strategies have been devised for
removing DNA sequences to create genes that are
truncated at various positions in the promoter region.
Specific small deletions and single nucleotide substitu-
tions have also been made to create additional globin
gene “promoter mutants.” The function of these
mutated genes has been compared to that of the
normal gene after introduction of recombinant plas-
mids containing each into monkey kidney or HeLa
cells.’4’22’23 Globin genes are actively transcribed in
such cells; after 48 hr, RNA is harvested for analysis.
The relative amount of correctly initiated globin
mRNA provides a measure of promoter function.
These gene expression studies have provided a func-
tional profile of the 3-globin gene promoter.’4’23 The
promoter extends upstream from the CAP site for 1 10
bp (- I 10). The conserved CCAAT and TATA boxes
are indeed important-deletion of these blocks, single
nucleotide substitutions within these sequences, or
changes in the relative distance between the two cause
a marked quantitative decrease in promoter function.
The duplicated sequences between -85 and - 1 10 are
also important (Fig. 2B)-deletion of these sequences
abolishes promoter function, and single nucleotide
substitutions can cause a substantial decrease in the
amount of correctly initiated 3-globin mRNA. The
rabbit fl-gbobin gene has been used for most studies of
promoter function; the sequence of the human f3-globin
gene promoter7”4 suggests that it is quite similar in
functional organization.
Four fl-globin genes, isolated from individuals with
fl-thalassemia, have been shown to have single nucleo-
tide substitutions in the promoter region (Fig. 2A).2428
Transcription of three of these genes was markedly
down compared to a normal gene in a cellular expres-
sion assay.25’27’28 These observations established that
single nucleotide substitutions in the promoter region
can cause the thalassemia phenotype in human ery-
Downloaded
from
by
guest
on
28
September
2021

A. Heterocellular
Normal
6l3 Th&
and 5.12%
B. Pancellular
Kenya
Arj kb
6 7% ft #{149}
iI#{149}w
II.
HPFH 25 30%
THALASSEMIA 741
C. yDeletion
yd -ThaI No
>a t I ____
-/ K k X
throid cells. The definition of the f3-gbobin gene pro-
moter, derived from in vitro analysis, has thus been
validated by study of these thalassemia gbobin genes.
The #{244}-globingene may be considered “thalassemic”
in that it encodes for a normal globin that is synthe-
sized at a very low level. Functional studies have shown
that the “defect” in the #{244}-gene
can be localized to the
promoter region.29 Note that the “CAT” box is imper-
fect and the distance between it and the TATA box is
shorter than for functional genes (Fig. 2A). Further-
more, the #{244}-gene lacks the conserved sequences
between 86 and 108 nucleotides upstream from the
CAP site. Undoubtedly, this defective promoter struc-
ture largely accounts for the gene’s hypofunction,
although recent studies have suggested that t5-mRNA
is also somewhat less stable than f3-mRNA in reticulo-
cytes.3#{176}
Remote DNA Sequences That Influence Gene
Expression
Certain thalassemia mutations are characterized by
deletion of large segments of the fl-gene cluster (Fig.
3)2.3I 43 Among the more interesting of these is one
found in a family with ‘y f3-theIassemia; deficient syn-
thesis of each of these globins has been documented in
affected individuals. This deletion has resulted in loss
of more than 60 kilobases (kb) of DNA, including the
sy and s-genes, but has left the f3-gene intact, along with
3 kb of upstream flanking sequences.36 This fl-gene has
been cloned. Its DNA sequence is normal and it
functions normally in a cellular expression assay in
vitro despite its lack of function in vivo.44 These
observations provide evidence that sequences remote
Fig. 3. Gene deletions in the /9-like gene cluster.
The levels of HbF indicated on the left of the line
diagram of each deletion are those observed in hetero-
zygous individuals. The functional globin genes and the
pseudogene are labeled appropriately. The following
symbol (#{149})
indicates the position in the /9-cluster. of
members of the Alu moderately repetitive DNA
sequence family. Each member of this family is 300 bp
long. and there are roughly 500.000 members in the
human genome. A 6.4-kb repetitive DNA sequence
( = ) is found downstream from the /9-gene ; there
are roughly 4.000 members of this family. called the Kpn
family47 in human DNA. Note that a portion of this
sequence is represented in an opposite orientation
upstream from the G.y gene.4$ This portion is juxtaposed
to yet another repeated DNA sequence. Many additional
moderately repeated DNA sequences are found within
the cluster. The approximate size of each deletion is
given in kilobases (kb). A single arrow is used to indicate
the position of deletion when its exact size has been
defined. The two HPFH deletions are of equal size; the
endpoints are equally displaced on both ends of the
deletion. suggesting a common mechanism for genera-
tion of the two mutations.
from the promoter can markedly influence the tran-
scription of the globin gene in erythroid cells.
What are the possible mechanisms by which remote
DNA sequences could influence promoter function? In
cellular expression assays, the f3-globin gene promoter
requires an “enhancer” for function.28’49’5#{176} The
enhancer consists of a short segment of DNA that acts,
independent of orientation or distance from the gene,
to ensure accurate and efficient initiation of tran-
5cription.5’ Such enhancers were first found in viral
DNA, but recently, an enhancer element has been
found in the introns of immunoglobulin genes.5255 This
enhancer is thought to activate the variable gene
region promoter after the rearrangement that creates a
functional immunoglobulin gene. Perhaps loss of func-
tion of the intact /3-globin gene in the ‘y5f3-deletion
chromosome could reflect the loss of an essential
enhancer of the /3-gene promoter.
Alternatively, the large deletion could result in loss
of DNA sequences that are important for establishing
an active chromosomal conformation. Most DNA is
associated with histone octamers (nucleosomes) and is
tightly compacted into chromatin. Actively tran-
scribed genes are in an “open” conformation that
leaves them accessible for interaction with RNA poly-
merases and other proteins that may be involved in
transcriptional regulation ,56,57 Certain segments, re-
ferred to as hypersensitive sites because of their ex-
quisite sensitivity to digestion by nucleases, are
thought to consist of naked DNA that is unassociated
with histones. The importance of such hypersensitive
sites for promoter function has been shown for the
Drosophila glue protein gene.58 Deletion of upstream
HbFIeVeI G1A7 ,pp ____
<1% I #{149}
iII.fl III 1$
____________________ d p ______
Lepore 1-5% #{149}
OI Ito III 1
Sicilian 14 kb
#{149}1
U OI I’D HI K
Span h >70
SI U tI I’D #{149}t
Turkish 40 kb
I. #{149}
II
>60kb I
.00 I /,‘ IUSA1
I >60kb 1
#{149}1
#{149} lID If Ghana
Downloaded
from
by
guest
on
28
September
2021

DNA that is normally found in hypersensitive sites has
resulted in loss of function of the associated, structur-
ally intact promoter. Similar remote sequences might
also be involved in globin gene expression and, if
missing, lead to hypofunction of a structurally normal
gene. Indeed, the chromatin structure of the f. -globin
gene44 on the chromosome from which the e, ‘y, and
#{244}-geneshave been deleted (Fig. 3C, -y#{244}fl-thal) lacks
hypersensitive sites and is in a closed conformation in
erythroid cells, where the fl-gene on the normal chro-
mosome has the characteristic structure of an
expressed gene.
Also summarized in Fig. 3 are several other deletion
mutations that influence expression of the -y-globin
genes.343 These mutations have been carefully studied
with the hope that they would provide insights into the
mechanism(s) that regulate the developmental switch
from fetal (HbF = a2 y2) to adult hemoglobin produc-
tion (HbA = a2f3,) during the perinatal period. The
#{244} 3-thalassemia mutations lead to a modest increase in
HbF in adult life that is heterogeneously distributed
among red cells (heterocellular).’ In contrast, the
hereditary persistence of fetal hemoglobin (HPFH)
mutations lead to a more striking increase in HbF,
with a more even distribution among red cells (pancel-
lular). Huisman and Schroeder first formulated the
hypothesis that sequences between the A.y and t5-genes
might influence the developmental switch from HbF to
HbA synthesis.59 Furthermore, they suggested that
deletion of these sequences with HPFH mutations, but
not in the fl-thalassemia mutations, could account for
their differing phenotypes. Discovery of members of
the Alu class of moderately repetitive DNA
sequences45 in this region focused attention on these as
potential regulatory elements.60’6’ Indeed, with excep-
tion of the Turkish thalassemia deletion that also
removed the A..y..gene,34 the presence of at least one of
the two Alu sequences upstream from the t5-gene is
associated with the #{244}fl-thalassemia phenotype, whereas
deletion of all or a portion of both Alu sequences is
associated with the HPFI-I phenotype.42’43’62 However,
the two classes of mutations differ with respect to size
(Fig. 3), and hence, with respect to the nature of the
DNA sequences brought into the ‘y-gene region. These
factors might also be relevant to the different pheno-
types of the two classes of mutations.
There are several HPFH mutations that lead to
increased production of HbF in adults without altering
the general structure of the /3-like gene cluster (nonde-
letion HPFH).’ Certain of these mutations are not
tightly linked to the 3-like gene cluster,63’64 but others,
particularly those that increase G or A*y synthesis
selectively, may reflect sequence alteration in func-
tionally important parts of these genes. Molecular
studies in progress in several laboratories should ulti-
mately lead to characterization of these interesting
forms of HPFH.
Mutations That Affect RNA Processing
The study of RNA metabolism in erythroid cells
suggested that many patients with fl-thalassemia
would have defects in RNA processing.6568 This pre-
diction has been proven correct by the sequence analy-
sis and in vitro functional characterization of several
cloned thalassemic globin genes.25’27’6987 Indeed, the
analysis of these globin genes has been remarkably
fruitful in providing detailed insights about the RNA
sequences required for specific, selective, and efficient
RNA splicing.
The Consensus Splice Junction Sequences
Comparison of the sequences found at splice junc-
tions provided the first clues as to the sequence require-
ments for RNA processing, much in the way that study
of promoter sequences identified conserved blocks that
have subsequently been proven to be essential for
promoter function. First to be identified were the
dinucleotides, GT at the 5’ end and AG at the 3’ end of
introns.88 A recent comparison of more than a hundred
splice junctions has provided the expanded consensus
splicejunction sequence shown in Fig. 2B.’5 The splice
junction consensus sequences appear to be involved in
the formation of a base paired structure between U,
RNA and the precursor RNA.89’9#{176}
U, RNA is a small,
stable, abundant, nuclear RNA species that may func-
tion as a cofactor in this splicing reaction. Adjacent
portions of the 5’ end of U, RNA are complementary
to the 5’ and 3’ splicejunction sequences; by forming a
base paired structure with U,, the 5’ and 3’ ends of an
intron are juxtaposed in a way that could readily
facilitate accurate splicing.
Partial matches to the consensus sequences occur
frequently at positions other than at the exon-intron
boundaries. For example, GT occurs 44 times in the
f3-globin gene within a DNA segment that matches the
5’ consensus sequence at 5 of 9 positions, and AG
occurs 49 times within a sequence that is a 9 of 16
match. These cryptic splice sites are rarely (if ever)
used in normal RNA processing, but may be activated
by mutations that make them a better match with the
consensus sequence or by an effect of remote mutation.
Study ofvarious thalassemia mutant genes has yielded
insight into the functional significance of individual
nucleotides within the consensus sequences and has
provided clues as to how splicing at one site may
influence splicing at cryptic splice sites at other posi-
tions (Fig. 4).
Downloaded
from
by
guest
on
28
September
2021

2. AIph.
1CU : D.468IO,=,,
B MUTATIONS THAT CREATE AN ALTERNATE SITE
i-s. A-fl’15)
CTCTCT 0 CCI A TTGG T
. /1-
24 7 5. $ 27
-
5 GTGGTGAG 0
I -I l__..T-fi (HbK.os.ou)
L i”-’ A-fl (1880 2’
-5- A-fl
C. MUTATIONS THAT CREATE Aid ALTERNATE SITE AND
ACTIVATE A CRYPTiC SITS
CAGCTA:CA
0 A I GTAAG A
L
AAOG C PA TA
Fig. 4. Thalassemia mutations that alter the splicing pattern
of the /9-globin mRNA precursor. The DNA sequences printed
below the line diagrams are those that occur in the indicated
positions. The obligatory dinucleotides. GT and AG. are underlined
in each sequence. Lower case letters indicate nucleotides that are
not part of the splice junction sequence. Normal splices are
indicated by solid lines above the line diagrams. whereas abnormal
splices resulting from various mutations are shown as dashed
lines. The numbers in parentheses after several of the mutations
appear to permit crossreferencing to the specific haplotypes
shown in Fig. 7 (see text for discussion).
THALASSEMIA 743
A MUTATIONS THAT ALTER A NORMAL SITE AND ACTIVATE
CRYPTIC SITES
1. B88 r- ’ “fl do spIls.
I 1
TAT000TCTATTTTCCC A CCCTIAGOCTGCT
- ‘A-
CAOIGT TO
F ILc n’o t
I c-
L___ _,., A-#{216}POI “ “
A TIC), ACAOC
AG 0 GTGAGT
1:....A-094)
sequence region at the exon 1-intron I boundary, as
illustrated in Fig. 4A.24’25 These mutant splice sites
continue to function, as shown by the presence of
normally spliced mRNA in test cells in vitro, but cause
several abnormalities in mRNA processing. There is a
quantitative decrease in the amount of correctly
spliced mRNA, an increase in the amount of unspliced
globin RNA, and activation of a cryptic splice site in
coding sequence RNA. Thus, these consensus nucleo-
tides, although not absolutely essential for splice site
function, have a marked quantitative effect on RNA
processing efficiency.
Mutations That Create an Alternate Site
Mutations That Alter Splice Sites
Several mutations completely destroy the function
of splice sites (Fig. Usually, these
mutations lead to activation of otherwise cryptic splice
sites in coding or intervening sequence RNA. One
intriguing aspect of these mutations is the suggestion
that there is some “pressure” to complete a splice
within the globin gene transcript rather than leaving
the RNA molecule unspliced.
Two j3 -thalassemia genes have been found to have
a single nucleotide substitution in the consensus
Shown in Fig. 4B are several mutations that create
alternate splice site. The first to be described involved
replacement of G with A at a position 19 nucleotides
upstream from the normal intron I-exon 2 bounda-
ry.70’7’ This region ofthe normal transcript matches the
3’ consensus splice sequence very well, except for lack
of the essential dinucleotide, AG. The fl-thalassemia
mutation creates an AG, allowing the splice site to
function. Indeed, 90% of the mRNA molecules tran-
scribed from this f3-thalassemia gene in vitro are
incorrectly spliced at the alternate site.76’79 Such incor-
rectly spliced RNA molecules cannot be translated
into f3-globin.
Centered on the GT dinucleotide within codon 25 is
a block that matches the 5’ splice site consensus
sequence in 6 of 9 positions. Nonetheless, this site is
rarely if ever involved in splicing reactions during the
processing of normal 3-gbobin gene transcripts. How-
ever, this is one of the cryptic sites that is used when
the normal exon 1-intron I boundary sequence is
altered by mutation.25 Three separate mutations within
this cryptic splice site also result in its activation.8”92’93
Substitution of A for T in codon 24 is silent at the level
of protein sequence but leads to incorrect processing of
about 80% of fl-globin RNA molecules and thereby
causes a moderately severe f3 -thalassemia pheno-
type.8’ The other two mutations alter 3-gbobin protein
sequence and also lead to the use of this cryptic splice
site, although less frequently than with the gene con-
taming the codon 24 substitution.92 Hence, the f3E gene
and the Hb Knossos gene are associated with a mild
f3-thalassemia phenotype.
Mutations That Create an Alternate Site and Also
Activate a Cryptic Site
Three mutations within intron II create alternate 5’
splice sites.2425808593 ’ These are located near the 3’ end
of the intron, as shown in Fig. 4C. One of the more
interesting aspects of these mutations is the concurrent
activation of a cryptic 3’ splice site further upstream
Downloaded
from
by
guest
on
28
September
2021

744 NIENHUIS. ANAGNOU, AND LEY
within the intron. This sequence matches the 3’ consen-
sus splice sequence in 15 of 16 positions, but appar-
ently is rarely, if ever, used during normal processing.
The spliced product formed by use of this 3’ site within
intron II and the normal 5’ site at the exon 2-intron II
boundary cannot be spliced further, since there is not a
5’ splice sequence immediately downstream from the
internal 3’ site in the transcript. At least one of these
mutations appears to create a /3#{176}-gene,in that none of
the f. -globin RNA molecules found in test cells are
processed to link exon 2 and exon 3 correctly.93’
Splicing Mutations-Implicationsfor RNA
Processing Mechanism
A careful study of the various thalassemia muta-
tions may contribute to formulation of a general mech-
anism of RNA processing. Several lessons have
already been gained from the study of these mutant
genes. For example, the dinucleotides, GT and AG,
have been shown to be obligatory for function of splice
sites, and the consensus nucleotides have a quantitative
effect on splice site function. We have also learned that
the splicing pattern of one intron may influence the
removal of the other set of intervening sequences. For
example, the alternate splice site mutation in intron I
(Fig. 4B), in addition to affecting the splicing pattern,
also retards removal of intervening sequences tran-
scribed from both introns I and 11.7679 Similarly, the G
to A substitution at the exon 2-intron II boundary
retards processing of intron I RNA.82 This mutation
also causes aberrant splicing, involving exon 1-intron I
and intron Il-exon 3 boundary sequences to link exons
I and 3.
Finally, the mutations thus far described affect
splice sites directly. RNA molecules form a complex
secondary structure; such a secondary structure could
influence splicing by bringing two splice sites in prox-
imity or by “burying” a potential splice site within base
paired portions of the RNA molecule.94 Indeed, study
of mutant viral genes supports the influence of the
secondary structure on the processing pattern. Perhaps
activation ofthe cryptic 3’ site in intron II (Fig. 4C) by
the alternate splice site mutations could reflect an
alteration in the secondary structure of the RNA
involving the cryptic site.
A Polyadenylation Signal Mutation
RNA transcription apparently continues beyond the
position on DNA that corresponds to the 3’ end of
mature mRNA molecules. For example, the mouse
globin gene primary transcript has recently been
shown to terminate roughly 750 nucleotides (nt) down-
stream from mRNA coding sequence.’6 The sequence,
AAUAAA, in the RNA transcript is thought to act as
a recognition site for cutting the primary transcript
some 20 nt downstream from the signal, thereby
allowing addition of the poly A tract. In vitro mutagen-
esis of the corresponding DNA segment, AATAAA,
does indeed affect polyadenylation.95 Recently, an
a-globin gene, cloned from the DNA of an individual
with HbH disease, has been shown to have a single
nucleotide substitution in the polyadenylation signal
(Fig. 2B).96 This leads to a quantitative reduction in
a-mRNA and a-globin production in this patient who
is homozygous for this mutation in the a2-gene and for
a frameshift mutation in the a,-gene that abolishes
gene function completely. In the transient expression
assay in HeLa cells, the polyadenylation signal muta-
tion leads to a marked reduction in correctly processed
and polyadenylated RNA and gives rise to longer
transcripts, extended at the 3’ end.
Mutations That Affect RNA Translation
There are several ways by which specific nucleotide
substitutions could influence the efficiency with which
RNA is translated into protein. However, the only
mutations described to date are those that influence
termination of protein synthesis. Specific codons (3
nucleotides) are used to signal the stop position of
translation. Nucleotide substitutions that alter a codon
that specifies an amino acid to a terminator codon
cause premature termination of translation, whereas
substitutions in the normal terminator allows synthesis
of elongated globin chains.
Premature Termination Mutants
Several such mutations have been described.69’73
75,77.83,84 Each involves either a single nucleotide substi-
tution in a normal codon to create a termination codon
or a deletion that alters the reading frame of the
mRNA to put a termination codon into phase. Exam-
ples of both are shown in Fig. 5. Note that several of
the mutations that alter the splicing pattern of globin
RNA also result in formation of mRNA molecules
that have in-phase termination codons that lead to
premature termination.70-77’79-92 Mutations that cause
premature translation termination are sufficient to
cause the thalassemia phenotype. These mutations are
also associated with a variable, but usually a marked,
quantitative decrease in the amount of globin
mRNA.69’83’97’98 The traditional explanation for this
quantitative effect of premature termination muta-
tions is that the untranslated portion of the mRNA
becomes accessible to nuclease degradation in the
cytoplasm. However, recent studies suggest that pre-
mature termination mutations may affect intranuclear
Downloaded
from
by
guest
on
28
September
2021

THALASSEMIA 745
31 33 34 35 36 37 38
CTTAGG CTG CTG GTG GTC TAC CCT TGG ACC
Intron I Leu Leu Val Val Tyr Pro Tm Thr
Deleted in 13#{176}
- 44 Gene
Sec I Leu I Gly I lie I Cys I Pro ILeu
39 40 41 42 43 441 45 46 47 48 49 50
CAG AGG TTC Tn’ GAG u GGG GAT CTG TCC ACT
Gin Arg Phe Phe Glu Ser Phe Gly Asp Leu Ser Thr
T in #{176} - 38 Gene (7)
Ter
‘ Leu I Met I Leu I Leu I Trp I Ala I Thr I Leu I Arg ITer
51 52 53 54 55 56 57 58 59 60161
CCT GAT GCT GTT ATG GGC AAC CCT AAG G rA6 G ..
Pro Asp Ala Val Met Gly Asn Pro Lys Val I Lys
Fig. 5. Two of the several thalassemia mutations that destroy
gene function by introduction of a premature translation termina-
tion codon in /9-globin mRNA. The numbers above the individual
codons refer to the encoded amino acid’s position in /9-globin.
Replacement of C with T in codon 39 introduces the terminator
UAG in /9-globin mRNA? 7 Another /9#{176}-gene
has a deletion of
the third nucleotide (C) in codon 41 . This results in a shift in the
reading frame of the mRNA; the new amino acid sequence is
shown above the line. This new reading frame has an in-phase
terminator (UGA) in a position corresponding to codons 60-61.
The abnormal peptides resulting from translation of the mRNAs
derived from these two genes are presumed to be highly unstable.
as they cannot be demonstrated in the erythroid cells of patients
with such genes.
RNA stability or nuclear to cytoplasmic transport and
therefore imply a previously unrecognized link
between mRNA translation and nuclear RNA meta-
bolism.99’ ‘
Terminator Codon Mutants
The normal termination codon for both a and /3-
globin is UAA; a-globin mRNA contains 109 nucleo-
tides in the 3’ untranslated region, and fl-mRNA
contains 132 nucleotides. Replacement of a single
nucleotide in the normal terminator to give a codon
that specifies an amino acid rather than termination
leads to synthesis of elongated globin chains. As mdi-
cated in Table 1 B, several different substitutions in the
a-mRNA terminator codon have been discovered; the
resulting mutant hemoglobins are structurally dif-
ferent only in the amino acid at the terminator codon
position.’#{176}’ The first to be discovered was Hb Constant
Spring (CS). HbCS a-mRNA is read to the next
in-phase terminator to give a “globin” with 172 rather
than 141 amino acids. Shorter globin products of 169
and I 54 amino acids apparently reflect limited proteol-
ysis of the elongated a-globin. HbCS a-globin is
present in red cells of affected individuals in very low
quantities. Recently, the ability to measure a, and
a2-mRNA independently’#{176}2’#{176}4 has revealed that the
CS mutation occurs in the a2-gene and that its mRNA
is quantitatively deficient.’#{176}3
This deficiency is thought
Table 1 . Thalassemic Hemoglobinopathies
Molecular Mechanism Structural Abnormality
of Thalassemic Effect in Globin
(A) Activation of splice site (A) (1) HbE /3(26) GIu
Lys
(2) Hb Knossos /3(27)
Ala -‘ Ser
(B) Loss of normal terminator (B) Elongated globin chain
(1 1 Hb Constant Spring
a(142) ter GIn
(2) Hb Icaria a(142)
ter -‘ Lys
(3) Hb Koya Dora
a(142) ter -‘ Ser
(4) Kb Seal Rock a(142)
ter -‘ Glu
(5) HbTak/3(147)
ter -‘ Thr
(C) Globin or hemoglobin in- (C) (1) Hb Indianapolis
stability 3( 1 1 2) Cys -‘ Arg
(2) HbQuongSze
co(125) Leu -‘ Pro
(D) Unknown mchanism (D) (1 ) Hb Vicksburg /3(75)
(1) DecreasedmRNA Leu-’O
(2) Hb North Shore
/3(134) Val -‘ Glu
(2) Postulated posttrans- (1) Hb Suan-Dok a(109)
lational instability Leu -‘ Arg
(2) Hb Petah Tikvah
cr(1 10) Ala -‘ Asp
to reflect instability of the HbCS a-mRNA,’#{176}’ but this
mechanism has not been directly demonstrated.
Not all elongated globins are associated with the
thalassemia phenotype. For example, Hb Cranston
contains a /3-chain with 157 rather than 146 amino
acids.’#{176}tInsertion of two nucleotides into codon 145
results in a shift in the mRNA translation reading
frame; the normal terminator is out of phase, but a
terminator codon in the new reading frame stops
protein synthesis after addition of the 157th amino
acid. Heterozygotes for the /3cranston..gene have normal
red cells and a normal HbA2. In contrast, heterozy-
gotes for the /3Tak..gefle (which also encodes a globin of
1 57 amino acids because of insertion of 2 nucleotides
into codon I 47) have hypochromic, microcytic red cells
and normal or only slightly elevated HbA2 levels.’06
Because of the different sites of insertion, f3Tak and
/3Cranston differ in the amino acids found at positions 145
and 146; f3Tak has the normal amino acids Tyr and His,
respectively, while /3Cranston has Ser and lie. The molecu-
lar basis for the differing phenotypes of these mutants
has not been elucidated.
Alpha Thalassemias
Alpha thalassemia is most often due to gene dele-
tion, although with some mutations, the a-genes are
Downloaded
from
by
guest
on
28
September
2021

0 5 10 15 20 25 30 ,
. VARIABLE I. I ‘u 2 ‘1 VARIABLE HAPLOTYPE
TI --.-- ---- 3
25.30
-(, I I,-thal-2 Ileftward deletion)
-(, . a-thal-2 lr’ghtward
-k’l thal-1 (MedIterranean)
intact; these are referred to as the nondeletion types of
alpha thalassemia.10’
Alpha Gene Deletions
The advent of molecular cloning, DNA sequencing,
and restriction endonuclease mapping techniques has
led to an increased understanding of the molecular
basis for the DNA rearrangements that cause a-
thalassemia. DNA sequence analysis of the two a-
globin genes and study of their flanking sequences has
revealed several tandemly duplicated highly homolo-
gous segments within the a-gene cluster (Fig.
The presence of these highly homologous regions
predisposes the a-globin gene cluster to recombina-
tional events. Misalignment between two strands of
DNA from two separate chromosomes during meiosis
can effectively lead to deletion of one of the two
a-globin genes. One type of deletion involves the
homologous regions 3’ to the /‘a-gene on one chromo-
some and 3’ to the a2-gene on the other and leads to
deletion of 4.2 kb of DNA, whereas a second type
involves the homologous segments containing the a2
and a,-genes and leads to deletion of 3.7 kb of
DNA.’#{176}’“ These are referred to as the leftward and
rightward deletions, respectively (Fig. 6). The out-
come of these unequal crossovers is the generation of
new chromosomes containing either a single a-gene or
triplicated a-genes.’#{176} 23 All four possible chromo-
somes have been identified in various populations
where a-thalassemia occurs in high frequency. Further
support for this mechanism of unequal crossover is a
similar occurrence observed in vitro during the propa-
gation of cloned human DNA fragments containing
the a-genes in E. co/i. The deletions occurring in E.
co/i are indistinguishable from the two types of dele-
tions described above.4 Note that the actual crossover
sites could occur anywhere within the homologous
segments and that there may be more than one muta-
tional event that has led to single or triplicated a-gene
chromosomes in each case.
The leftward type of deletion is prevalent in Oriental
individuals’22 and also occurs in low frequency in Saudi
Arabia,’23 but is never encountered in Mediterraneans
or American blacks.”7 The rightward deletion has a
worldwide distribution among all racial groups (Orien-
tals, Mediterraneans, and blacks).’#{176} ”9 The single
a-gene chromosomes are much more common than the
triple a-gene chromosomes.41”2#{176}’23 Greek Cypriots
exhibit the highest frequency of a-thalassemia among
white people and also have a relative high frequency of
triple a-gene chromosomes’24
Alpha globin production is roughly proportional to
the number of functional a-globin genes,’ although
slightly more a2 than a,-mRNA is produced in normal
erythroid cells.’#{176}2’#{176}
Deletion of a single a-gene leads
to an insignificant decrease in a-globin production in
an individual having two functional a-globin genes on
the other chromosome. Phenotypically, such individu-
als have been classified as having a-thalassemia-2
(silent carrier), and the single a-gene chromosome
may be considered to be an f-thalassemia mutation.
Individuals with two chromosomes containing only a
single a-globin gene have a more significant decrease
in a-globin production; such patients have a-thalasse-
mia-l (mild hypochromic microcytic anemia) accord-
ing to the standard clinical classification.’
42
=
37
52
=
(, GLOBIN
PHENOTYPE SYNTHESIS
-k.)/ -thai-i (Mediterranean)
--I ,-thal-1 (Mediterr anean . hydrops) I’
--I .-tha)1 (Southeast Aslan.hvdrops)
--I (.-tha)-1 )Menta) retardation)
U functIonal genes
n pseudogenes
area of deletIon
::::: undefIned limjts
3 =- IregIons of homology
Fig. 6. A summary of deletions in the a-globin gene cluster. The line diagram indicates the relative position of the genes and
pseudogenes within the cluster. The two regions labeled “variable” are places where DNA length polymorphisms have been found; several
different alleles. characterized by DNA fragments of different lengths in this region. have been defined.’#{176}7”#{176}
The positions of several
tandemly duplicated DNA segments (X. V. and Z) are indicated.4’5 The extent of the deleted segments for each of the mutations is shown.
Haplotype” in this context refers to the number of a-genes or fragments of a gene found on each chromosome, whereas the phenotype is
that found in heterozygous individuals; a-thal-2 is a silent carrier state and a-thaI-i is thalassemia trait (mild hypochromic. microcytic
anemia). The interested reader should consult an excellent recent review on the a-thalassemias for further details and specific references
to each of these mutations.’#{176}’
Downloaded
from
by
guest
on
28
September
2021

THALASSEMIA 747
Several other deletions involve both a-globin genes
of a single chromosome, as shown in Fig. 6.lIlh12125l28
These are a#{176}-thalassemia mutations. The extent of
these deletions has been defined by restriction endonu-
clease mapping, but the mechanism of their occurrence
is not yet obvious. Inheritance of a chromosome lack-
ing both a-genes leads to the a-thalassemia-1 pheno-
type if the other chromosome is normal. If the other
chromosome 16 has only a single a-gene, HbH disease
(a moderately severe hemolytic anemia) occurs.
A novel and potentially important deletion in the
a-like globin gene cluster was characterized in two
patients of North European origin with an unusual
combination of acquired HbH disease and mental
retardation.’29 This phenotype apparently occurred
from the genetic combination of a standard “right-
ward” deletion chromosome inherited from the mother
and a de novo deletion of the whole ct-globin gene
cluster on the other chromosome (Fig. 6). The father
was clinically normal. This deletion must therefore
represent an acquired mutation that occurred either in
the germ cells of the father or at a very early stage of
fetal development. A third patient also apparently
inherited a de novo mutation from his father that
inactivated both a-genes; this mutation was apparently
of the nondeletion type.’29 The combination of these
unusual acquired mutations and mental retardation
might be coincidental or could imply the presence of a
locus within the a-gene cluster or within nearby DNA
sequences that influences mental development.
Nondeletion a- Thalassemia Mutations
These comprise a distinct and apparently heteroge-
neous group of mutations that lead to absent or greatly
reduced expression of a-genes that are structurally
intact, as determined by restriction endonuclease map-
ping.”2”3”3 ’33 The inheritance of such mutations, in
combination with either a or a#{176}-typedeletion muta-
tions, leads to a highly variable clinical pattern.’0’
Undoubtedly, the nondeletion types of a-thalassemia
represent the same types of mutations described in
detail above in the genes of patients with /3-thalasse-
mia. Indeed, a 5-nucleotide deletion at a splice junction
leads to abnormal splicing (Fig. 4A),9’ and a second
mutation (described below) leads to synthesis of a
structurally abnormal globin (1 25 Leu -k Pro) that is
highly unstable.’34 The interesting mutation, repre-
senting substitution of a single nucleotide in the poly-
adenylation signal (Fig. 2B), found in a nondeletion
type a-thalassemia gene96 has already been described.
The rather rare and intriguing syndrome of acquired
HbH disease is associated with a variety of myeloproli-
ferative disorders and has been described in 34
patients.’35 The a-globin genes are structurally intact
in the erythroid cells ofsuch patients, and furthermore,
they were found to be hypomethylated, a characteristic
of highly expressed genes.’36 A marked reduction in
transcription of all four a-globin genes has been dem-
onstrated, but no reactivation of c-gene expression was
observed.’35”36 The ratio of a, and a2-mRNAs was
normal, implying that the mutation causing this disor-
der has suppressed all four a-genes in a balanced
manner.’36 This result is most consistent with a muta-
tion affecting an intranuclear factor required for a-
gene transcription.
c-Gene Rearrangements
Because of the high sequence homology between the
1 - and the c-genes, it is not surprising that unequal
crossovers have led to generation of chromosomes
having single or triplicated -genes.’#{176}8”37 At least two
different sized deletions have been defined, reflecting
different recombinational events. Further studies are
needed to define the expression levels of these rear-
ranged genes and whether these rearrangements have
any significant physiologic effects during embryonic or
adult life.
USE OF RESTRICTION ENDONUCLEASE
POLYMORPHISMS TO STUDY BETA THALASSEMIA
The study of genomic variation at the level of DNA
in man has led to immediate applications to population
genetics, prenatal diagnosis, and characterization of
various diseases. With molecular probes and a vast
array of restriction endonucleases, a new body of
information has accumulated concerning DNA
sequence polymorphisms in the globin gene clusters,
and their linkage to various specific thalassemia muta-
tions and hemoglobinopathies has been established.
Polymorphisms are defined as genetic differences
among individuals.’38 The two possible alleles for a
restriction endonuclease cleavage site are either (+),
namely the enzyme cuts at that site, or (-), the
enzyme does not cut. A particular site is considered to
be polymorphic in a given population if the least
frequent allele is present in more than I % of individu-
als. The first restriction endonuclease site polymor-
phism found useful for study of hemoglobinopathies
was that involving an Hpa I site, 4 kb downstream
from the /3-globin gene.’39 In the first study, 87% of
chromosomes having a sickle (flS) globin gene also lack
this Hpa I cutting site (-), whereas the (-) allele is
very infrequently found in DNA of individuals who do
not have a flsglobmn gene. This linkage of the Hpa I
(-) allele to the f3S mutation has been useful for
prenatal diagnosis’39”#{176} and for population genetic
studies.’41”42 Discovery of the Hpa I polymorphism
and its use in sickle cell disease prompted a search for
Downloaded
from
by
guest
on
28
September
2021

other polymorphic sites within the /3-globin gene
cluster.’4415’ However, the linkage of a specific allele
at a single restriction endonuclease site with a particu-
lar globin gene mutation [such as the tight linkage of
the Hpa I ( - ) with the /3S gene] is quite rare.
By using several polymorphic restriction sites to
define the haplotype of an individual chromosome,
Orkin and Kazazian and their colleagues have found
an association of specific thalassemia mutations with
particular haplotypes (Fig. 7)#{149}24 This was initially
accomplished by cloning and sequencing globin genes
from DNA of thalassemic individuals with each of the
several haplotypes. Coupling of specific haplotypes to
particular mutations has subsequently permitted
restriction endonuclease analysis to be used for identi-
fication of specific mutations in individual patients.
The data shown in Fig. 7 apply only to the Mediter-
ranean population. In other ethnic groups, e.g., Amen-
can blacks and Asians, the same haplotypes are found
to be associated with completely different mutations
(Orkin and Kazazian, personal communication). This
is consistent with genetic isolation of the various
groups in whom thalassernia is common during the
period when malaria operated as a major selective
factor for the thalassemia trait phenotype. Study of
new patient populations by haplotype determination
provides an initial estimate of the number of different
mutations in that population and allows examples of
each to be selected for molecular characterization.
Coupling of a haplotype with a specific mutation is
5.LGyAyvPdfl3.
Hinc II Hd Ill Hd Ill Hinc II Barn HI
Haplotypes Ava II Mutation
Overall
Frequency
I + - - -- + + 5 47%
II - + + -s + + 17%
Ill - + - ++ + - 4 8%
lv - + - ++ - + 1%
V + - - -- + - 3 12%
VI - + + -- - + 2 6%
VII 4 - - -- - + 6 6%
Vlll - + - +- + - 1 1%
IX - + - ++ + + 7 3%
Fig. 7. Linkage of haplotypes. defined by several restriction
enzyme polymorphisms. to specific /9-thalassemia mutations. ( +)
Indicates that the enzyme cuts at that site. whereas ( - I indicates
that the enzyme does not. As shown. this combination of 7
polymorphic sites can be used to define 9 unique haplotypes. The
individual mutations linked to each haplotype can be identified by
reference to earlier figures. Mutation 1 is shown in Fig. 2A.
mutations 2-4 in Fig. 4A. mutation 5 in Fig. 48, mutation 6 in Fig.
4C. and mutation 7 in Fig. 5. Haplotype V is linked to a
/9-thalassemia gene having a two-nucleotide deletion in codon 6
that results in a frameshift in the translational reading frame and
therefore leads to premature termination.
not absolute, even within a single ethnic group. In their
initial study, Onkin et al. found the same mutation
(/3039 premature termination codon) associated with
two different haplotypes.24 Kan and colleagues have
recently found that the /3#{176}-39
mutation in the Sardin-
ian population is associated with five different haplo-
types.’59 Another exception was the finding of more
than one mutation associated with a single haplotype.
These findings may imply that the same mutation has
arisen more than once,’ #{176}or more likely, that DNA
sequence rearrangements in the /3-cluster occur with a
far higher frequency than was heretofore suspected.
From a practical perspective, haplotype analysis,
although useful in identifying thalassemia mutations,
is not absolutely accurate. Overall, the linkage of
specific mutations to a particular haplotype holds
about 80%-90% of the time.
A mutation that causes thalassemia or a hemo-
globinopathy also may result in loss of a restriction
endonuclease cutting site. For example, the /3 muta-
tion in the sixth codon of the /3-globin gene results in
loss ofcutting sites for Mnl I, Dde I, and Mst II. Use of
Mst II to recognize the f.3 mutation in amniotic fluid
cell DNA has provided an easily applied and practical
method for prenatal diagnosis.’6’’63 Among the thalas-
semia mutations that can be recognized by restriction
endonuclease analysis are those that occur at -87 in
the /3-globin gene promoter (Fig. 2A-mutation 1, Avr
II),24 at the exon 1-intron I splice junction of the
a-globin gene (Fig. 4A-mutation 4, Hph I),72 at the
exon 2-intron II splice junction of the /3-globin gene
(Fig. 4A-Hph I), ’ and within intron II of the /3-
globin gene (Fig. 4C-mutation 6, Rsa I).80 These
mutations are relatively infrequent in individuals with
/3-thalassemia, so that direct restriction endonuclease
analysis is not of great practical value.
CLINICAL AND GENETIC HETEROGENEITY OF THE
THALASSEMIAS
The clinical heterogeneity of the thalassemias is
based in part on the diversity of specific mutations and
their quantitatively variable effect on mRNA and
globin production. For example, individuals with two
/30 mutations usually will have thalassemia major
(transfusion-dependent). Similarly, individuals homo-
zygous for the alternate splice site mutation in intron I
(no. S in Fig. 4B and Fig. 7) also have severe disease
because precursor mRNA molecules transcribed from
the gene with this mutation are spliced incorrectly 90%
of the time. Hence, patients homozygous for this defect
have a marked quantitative deficiency of /3-globin
production.’52 In contrast, an individual homozygous
for a mutation in a consensus splice junction sequence
Downloaded
from
by
guest
on
28
September
2021

THALASSEMIA 749
(mutation 2, Fig. 4A) that decreases, but does not
abolish splicing, has milder disease-the thalassemia
intermedia phenotype. ‘ 52,1 52a
In certain populations, such as the American Greek
and Italian emigrants studied to obtain the data shown
in Fig. 7, several mutations are present, some of which
are found quite frequently. Hence, most patients with
severe thalassemia in this group, although functionally
homozygous in that both /3-globin genes are defective,
are, in fact, doubly heterozygous for two different
mutations. In such populations, homozygous thalasse-
mia will be extremely heterogeneous, and further
studies correlating molecular genotype with clinical
phenotype will be most informative. In other popula-
tions, only one mutation may be found and the thalas-
semia phenotype in homozygous individuals will be
quite homogeneous. Both haplotype analysis, using
restriction endonuclease polymorphisms,24 and direct
DNA analysis, with oligonucleotide probes specific for
a particular mutation,lS2b will provide genotype infor-
mation that can be correlated with clinical analysis.
Other modifiers of clinical phenotype in patients
with homozygous thalassemia include coincident
inheritance of an a-thalassemia mutation’53’56 or a
mutation that increases HbF synthesis.’57”58 Alpha
thalassemia reduces disease severity, as the pathophys-
iology of severe /3-thalassemia is based on the relative
excess of a-globin production. Any modifier that
reduces this relative excess will be beneficial. Similar-
ly, enhanced “y-globin synthesis reduces the imbalance
in globin production and has a beneficial effect. Con-
versely, coinheritance of a chromosome with three
a-genes increases the severity of /3 thalassemia.lS2c
The a-thalassemia syndromes are also extremely
clinically heterogeneous. Detailed analysis of genotype
has provided many insights into the molecular basis for
this heterogeneity. The reader is directed to an excel-
lent review on the a-thalassemias by Higgs and
Weatherall’#{176}’ for a comprehensive consideration of the
a-thalassemia syndromes.
THALASSEMIC HEMOGLOBINOPATHIES
Hemoglobinopathies are defined as abnormal hemo-
globins with one or more structural changes, usually a
single amino acid substitution in a globin chain that is
synthesized at a normal or nearly normal rate. In
contrast, thalassemia mutations are traditionally
defined as those that leave the globin structure
unchanged, but lead to a quantitative reduction in
synthesis of the globin. The simple categorization of
globin gene mutations has been violated by several that
lead both to alteration of globin structure and to
decreased synthesis. For example, as described above,
the HbE [/3(26) Glu- Lys]92 and Hb Knossos [/3(27)
Ala- SerJ93 mutations activate an alternate splice site,
leading to a quantitative deficiency of/3-globin mRNA
(Fig. 4B). HbCS is another example where both
structure and synthesis of a-globin is altered by a
single .l03 Similarly, the crossover globin,
Hb Lepore,’ arises from a recombinational deletion
that created a S’#{244}
3’/3 hybrid gene (Fig. 3), the product
of which is synthesized at a very low rate. Hence, the
Lepore gene causes thalassemia. Several other exam-
ples may be cited, as shown in Table 1 . These muta-
tions have been reviewed in detail recently.’64 Only
those in which the mechanism of the thalassemic effect
has been at least partially clarified will be discussed
here.
The /3-chain found in Hb Indianapolis has been
shown to be highly unstable by biosynthetic studies.’65
The /3Indchain apparently preferentially associates
with the red cell membrane, where it is also catabol-
ized, as shown by pulse-chase studies. Hb Indianapolis
is unusual in that it causes the severe thalassemia
phenotype in heterozygous individuals, perhaps due to
membrane damage caused by the /3lnd..chain. Other
unstable hemoglobin variants may be associated with
mild hypochromia, but ineffective erythropoiesis and
severe anemia (found in individuals with Hb Indianap-
ohs) is not usually observed. Heterozygosity for a mild
/3-thalassemia gene on the chromosome opposite the
/3Ind..gene, suggested by increased HbF levels (1 0% and
12%), could explain the clinical severity. However, the
affected individuals are father and daughter and there
is no other evidence of /3-thalassemia in this family.
The Hb Quong Sze mutation [a(125) Leu-’Pro]
was discovered upon sequencing of the two a-genes
from one chromosome of an individual with HbH
disease.’34 Both a-genes were missing from the other
chromosome. One of the two sequenced a-genes was
normal. Discovery of the sequence abnormality in the
other gene prompted the hypothesis that aQ 0I 8S2c chain
would be unstable, as leucine to proline substitutions
often have a profound effect on the stability of the
hemoglobin tetramer. This hypothesis was supported
by biosynthetic studies that demonstrated a normal
synthesis but instability of the mutant a-globin.’34
Direct measurements have shown that the /3-mRNA
for Hb Vicksburg [/3(75) Leu- ’0] is reduced in con-
centration,’66 whereas deficiency of the /3-mRNA for
Hb North Shore [/3( I 34) Val- GluJ has been inferred
by biosynthetic studies.’67 The deletion causing /3Vlcksburg
and the substitution causing /3North Shore do not alter
/3-globin RNA sequence so as to create or activate an
alternate splice site. The reason for the thalassemic
effects of these mutations therefore remains obscure.
Downloaded
from
by
guest
on
28
September
2021

Perhaps two mutations are present in each /3-gene: one
responsible for the structural change and the other
having a thalassemic effect. Alternatively, sequence
alterations in /3-globin RNA may alter transport or
processing by mechanisms not yet appreciated.
THE TREATMENT OF THALASSEMIA
The treatment of thalassemia pertains only to that
of severe (or homozygous) /3-thalassemia. Those
patients with thalassemia trait (heterozygous a or
/3-thalassemia) require no treatment. Patients with
HbH disease also usually do quite well with no specific
treatment. Because of their severe anemia, which
requires lifelong transfusional support, nonchelated
patients with severe /3-thalassemia die from iron over-
load between the ages of I 5 and 25 yr.
Conventional Therapy
Transfusions
Red cell transfusions were originally given to thalas-
semic patients to sustain life and then to relieve the
symptoms of anemia. “ Hypertransfusion,” a strategy
designed to maintain a minimal hemoglobin of 10 g/dl
and a mean of I 2 g/dl, was justified despite increased
blood requirements because of improved growth and
development during the first decade.’68 In addition,
maintaining the hemoglobin at higher levels tended to
reduce the hyperabsorption of dietary iron that occurs
with the dyserythropoietic 69 Splenectomy
when the red cell requirement exceeds 250 ml/kg/yr is
also a part of a standard conventional therapy.’7#{176}
With hypertransfusion, however, the bone marrow
mass and blood volume remain expanded. For this
reason, a more aggressive program of transfusions
(“supertransfusion”) was proposed.’7’ Maintenance of
a mean hemoglobin of 14-15 g/dl will almost com-
pletely suppress endogenous erythropoiesis, thereby
shrinking the mass of the bone marrow and reducing
the blood volume. Red cell transfusion requirements
therefore might be no greater in supertransfused
patients than in hypertransfused patients because of
this reduction in blood volume. This prediction
appeared to be supported in a large study of several
patients,’72 but as yet, there is insufficient clinical
experience to judge the long-term value of the “super-
transfusion” regimen.
The red cell requirements of thalassemic patients
could be reduced if young red cells (“neocytes”) with
prolonged survival in vivo could be given. Collection of
neocytes by continuous flow centnifugation,’7’ or by
fractionation of conventionally collected blood units
with a cell washer, is feasible.’73 Prolonged survival of
such young red cells can be demonstrated by measure-
ment of the 51Cr half-life. Well documented studies are
needed to quantify the decrease in red cell requirement
realized by use of neocyte units. A significant reduc-
tion in transfused iron must be demonstrated to justify
the added effort and expense required to obtain young
red cells.
Chelation Therapy
Desferrioxamine (Desferal#{174}) is the only iron chela-
tor that is currently in widespread clinical use. Its role
in the treatment of transfusional iron overload has
recently been reviewed.’74 This drug is effective in
inducing iron excretion and is relatively nontoxic;
unfortunately, it must be delivered as a continuous
8-12 hr subcutaneous infusion to be effectiye. Iron
balance can be maintained in most patients with
infusions of I ,500-2,000 mg of the drug given 5 or 6
days a week. Supplemental intravenous infusions of
larger amounts of the drug can be used to markedly
augment iron removal, particularly in older patients
with marked iron overload. A decrease in liver iron
concentration, an improvement in liver function tests,
and reduced fernitin levels have been demonstrated in
older patients on intensive desfernioxamine treatment.
Attainment of iron balance in younger patients may be
demonstrated by comparing transfused iron to iron
that is excreted. Iron balance is reflected by low serum
fernitin levels.
Prevention of cardiac damage by desfernioxamine
has not been demonstrated, however. Propper et al.’75
have recently reported that initiation of chelation in
thalassemic patients after the age of 10 does not
appear to prevent cardiac disease. However, we have
observed an adult with acquired pure red cell aplasia in
whom regular chelation was started shortly after the
transfusion requirement was established. This patient
reached an iron burden equivalent of 300 U of packed
red cells without significant cardiac iron deposition.’76
At present, despite the expense and cumbersome
nature of administration, the regular use of Desferal is
the only hope of forestalling the consequences of iron
overload in regularly transfused thalassemic patients.
Vitamin C is commonly employed to augment iron
removal in response to desfernioxamine. The ability of
this substance to mobilize iron from nontoxic stores, or
to enhance lipid peroxidation by iron in myocardial
cells, can apparently lead to cardiac deterioration in
heavily iron overloaded individuals treated with vita-
mm C.’77 Hence, a low dose should be used, the drug
should not be given without a concurrent desferniox-
amine infusion, and patients should be advised to avoid
large quantities of citrus fruits or vitamin C supple-
ments. Vitamin E deficiency, often present because of
the excessive catabolism of this substance in iron
Downloaded
from
by
guest
on
28
September
2021

THALASSEMIA 751
overloaded individuals, may be corrected by oral
administration.’78 This reduces the amount ofa peroxi-
dation product (malonyldialdehyde) in red cell mem-
branes, but prolongation of red cell survival has been
demonstrated only when vitamin E is given parenter-
ally to thalassemic patients.’79
Development of Experimental Chelators
The National Institute of Arthritis, Diabetes, and
Digestive and Kidney Diseases and the National
Heart, Lung, and Blood Institute have sponsored
efforts to develop new chelating agents that are active
orally. These efforts have led to identification of a
drug, ethylene diamine-N,N’-bis(2-hydroxyphenyl
acetic acid)-dimethyl ester dihydrochlonide (EDHPA),
that has considerable ability to chelate iron in animals
when administered orally (David Badman and W.F.
Anderson, personal communication). Toxicity studies
in animals have been initiated to determine whether
these compounds are suitable for human use. Efforts to
modify desfernioxamine chemically to make it active
by the oral route have been unsuccessful. All other
investigational chelating agents (e.g., rhodotorulic
acid and 2,3-dihydrobenzoic acid) have been found to
be excessively toxic or to be inefficacious in human
trials.’74
New Approaches to Treatment
Bone Marrow Transplantation
Although transplantation from an HLA identical
sibling involves an approximately 30% risk of death,
successful transplantation of individual thalassemic
patients can offer the chance for cure and perhaps a
normal life expectancy.’80”8’ On the other hand, con-
ventional therapy with regular blood transfusions can
be expected to provide 12-15 yr of fairly good quality
life before the complications of iron overload begin to
develop. Death usually occurs between the age of 15
and 25 yr.’74 The impact of chelation therapy on
survival is not certain.l75l8la Nonetheless, there is the
possibility that a newborn or very young child with
thalassemia might benefit from new developments in
genetic therapy (see below). Thus, the choice between
bone marrow transplantation and conventional then-
apy represents a fundamental ethical dilemma, as the
significant risk of death or chronic graft-versus-host
disease must be weighed against the probability of cure
in an individual patient.
To achieve a high probability of engraftment and to
reduce the risk of graft-versus-host disease, the Seattle
team has chosen to transplant very young thalassemic
children who have received minimal blood transfusions
prior to transplantation (“good-risk” patients). ‘ 8 1.182
Five patients transplanted to date were conditioned
with dimethyl busulfan and cyclophosphamide. Three
patients achieved engraftment with cure of their thal-
assemia and have done well during follow-up of 1 1 mo
to 2 yr. although two have died of transplant complica-
tions (E. Donnell Thomas, personal communication).
Lucarelli and his colleagues in Pesaro have also
achieved a high rate of engraftment in several good-
risk patients using the same conditioning regimen,
although their earlier efforts with a conditioning regi-
men including cyclophosphamide and total body irra-
diation (but no busulfan) were less successful. “High-
risk” patients-older individuals with an extensive
transfusion history-had a low frequency of engraft-
ment (1 in 7) (Guido Lucarelli, personal communica-
tion), with a conditioning regimen that included
cyclophosphamide and total body irradiation. From
experience with bone marrow transplantation for con-
ventional indications, it is highly likely that these
patients, because of prior sensitization, are at a rela-
tively high risk for graft rejection.’8#{176} About 25% of all
transplanted thalassemia patients had significant
graft-versus-host disease; and several patients have
died of complications related to bone marrow trans-
plantation.
Despite the small number of transplants thus far
performed, the data suggest that some young, mini-
mally transfused thalassemic patients may be cured of
their disease by successful transplantation; others will
die from complications of the transplant procedure.
Only patients with HLA identical donors are currently
candidates. Whether engraftment can be achieved in
more heavily transfused patients remains to be deter-
mined. Transplantation for this disease should still be
considered experimental and its application limited to
highly experienced clinical investigators.
Activation ofthe ‘y-Globin Genes
The clinical severity of /3-thalassemia is roughly
proportional to the imbalance in a and non-a-globin
synthesis.’ Those patients who also inherit a gene that
increases “y-globin synthesis during postnatal life
(HPFH)’57”58 have a milder clinical course than the
usual patient with /3-thalassemia. Hence, reactivation
of the developmentally repressed ‘y-globin genes has
long been proposed as a therapeutic goal for treatment
of patients with severe /3-thalassemia.
Demonstration that cytosine residues near the “y-
globin gene become methylated coincident with their
developmental 83 184 provided the rationale
for a strategy designed to reactivate these genes.’85”86
5-Azacytidine is an analog of cytidine that has nitro-
gen rather than carbon in the fifth position of the
pyrimidine ring; this is the position that is methylated
Downloaded
from
by
guest
on
28
September
2021

752 NIENHUIS. ANAGNOU. AND LEY
following DNA synthesis. Hence, 5-azacytidine, when
incorporated into DNA during cell division, cannot
undergo methylation and, more importantly, 5-azacy-
tidine residues in the newly synthesized DNA strand
apparently bind and inactivate the methyltransferase
responsible for adding methyl groups to other sites.’87
All DNA in cells exposed to 5-azacytidine, therefore,
becomes hypomethylated after a few cell divisions. As
5-azacytidine had been shown to reactivate repressed
genes in tissue culture cells, DeSimone, Heller, and
colleagues administered the compound to baboons in
an effort to reactivate the fetal globin genes.’85 These
investigators found a striking increase in y-globin
synthesis and a reciprocal decrease in /3-globin synthe-
sis in these animals, whose globin gene organization
and structure is similar to that of man.
These observations have prompted limited trials of
5-azacytidine in patients with severe /3-thalassemia
and sickle cell disease.’86”88”89 Four to seven-fold
increases in ‘y-globin synthesis have consistently been
observed. In the thalassemic patients, an increase in
the reticulocyte count and hemoglobin level suggested
that increased ‘y-globin synthesis led to more effective
erythropoiesis and improved red cell survival. The
effect of a 7-day course of 5-azacytidine disappears
after 2-3 wk, so that regular or cyclic administration
would be required to achieve a long-term increase in
fetal globin synthesis.
5-Azacytidine administration leads to generalized
hypomethylation of bone marrow cell DNA, but only
the ‘y-genes are activated. The embryonic t-globin
genes are also hypomethylated after treatment, but
very little mRNA for this globin was found in the bone
marrow cells of treated patients.’86”88 Several other
genes, including those for insulin, collagen, and the
myc cellular proto-oncogene, also become hypomethy-
lated without increased expression (T.J. Ley, unpub-
lished observation). The tendency of 5-azacytidine to
selectively reactivate the ‘y-globin gene (but not other
genes on human chromosome 1 1) has also been demon-
strated in vitro, using a somatic cell hybrid line derived
from mouse erythroleukemia cells and human fibro-
blasts.’ #{176}The remarkable specificity of 5-azacytidine
in activating the ‘y-genes, nonetheless, invites the sug-
gestion that the mechanism by which it activates this
gene is unrelated to its demethylating effect, but rather
is due to a subclinical cytotoxic effect on erythroid
progenitors. Bone marrow ablation with subsequent
regeneration is regularly accompanied by a transient
increase in HbF synthesis. The initial experimental
studies in baboons suggested that drugs other than
those that cause demethylation did not lead to rny-gene
activation,’85 but others have reported that cytosine
arabinoside or hydroxyurea can lead to increased HbF
production in primates.’90”’90 ’ Thus, the mechanism(s)
by which 5-azacytidine increases HbF production is
not yet established.
5-Azacytidine is a carcinogenic 192
although its administration to man or experimental
animals does not lead acutely to development of
tumors. No tumors were observed after several months
of administration of the drug to mice and rats,’93
although many of the older animals did ultimately
develop hematopoietic neoplasms. One study, involv-
ing a small number of Fischer rats, demonstrated an
increased incidence of tumors when compared to con-
trol animals several months after the drug was stopped
(Brian Carr, personal communication). More than
I ,000 patients have received the drug as treatment for
leukemia; no secondary tumors have yet been observed,
although the average follow-up period is relatively
short (information provided by the Investigational
Drug Branch, National Cancer Institute).
The long-term carcinogenic risk has led some to
argue that 5-azacytidine should not be given to thalas-
semic patients.’94 Many questions about its mechanism
and relative potency, compared to other drugs, in
increasing HbF synthesis can be resolved by animal
studies. Hence, human studies can be limited to those
designed to determine whether the drug can eliminate
or reduce the transfusion requirement of individual
patients. The long-term risk of the drug, in such
patients, should be weighed against the patient’s prog-
nosis with conventional treatment.
Gene Therapy
Recombinant DNA methodology and the availabil-
ity of cloned globin genes have provided the first
realistic opportunity to consider approaches to func-
tionally correct thalassemic defects by introducing new
genetic information into bone marrow cells. Transfer
of genes into tissue culture cells has become a routine
procedure. Cells deficient in an enzyme, e.g., thymi-
dine kinase, can be transformed to thymidine kinase
expression by introduction of purified DNA fragments
containing the thymidine kinase gene.’95 Because the
efficiency of transformation is low, a selective growth
media must be used to inhibit growth of enzyme-
deficient cells, while allowing growth of the trans-
formed cells. Transformation of metabolically normal
cells can also be achieved by use of genes for selectable
markers that confer a genetic advantage on such cells
in the presence of a particular drug. For example, the
gene for a mutant, methotrexate-insensitive, dihydro-
folate reductase has been cloned and will (if expressed)
allow cells to grow in high concentrations of metho-
96
Globin genes do not confer a growth advantage.
Downloaded
from
by
guest
on
28
September
2021

THALASSEMIA 753
Hence, their introduction in cells requires cotransfor-
mation: a gene for globin and one for a selectable
marker are introduced simultaneously. Using this
strategy, the human /3-globin gene has been introduced
into mouse erythroleukemia cells, and several trans-
formed clones have been shown to express the human
/3-globin gene.5#{176}
Exposure of these transformed eryth-
roleukemia cells to the inducer, dimethylsulfoxide,
leads to increased human /3-globin gene expression as
the cells mature into erythroblasts, in parallel with
increased expression of the mouse globin genes.
Although regulation of the human globin gene is
apparently “normal,” not all clones express the human
gene, and the level of its expression is considerably
lower than that of the native mouse globin genes.
The general strategy for introducing globin genes
into mouse hematopoietic stem cells was developed by
Cline et al.’97 Suspensions of hematopoietic cells were
exposed to DNA in vitro, reinfused into lethally irra-
diated animals, and the transformed cells selected by
treatment of the recipient animals with methotrexate.
These experiments employed genomic DNA from cells
that produced a methotrexate-insensitive dihydrofo-
late reductase. The availability of the cloned gene
sequences for this enzyme’96 should facilitate attempts
at gene transfer into hematopoietic stem cells of exper-
imental animals and allow cotransformation of such
cells with globin genes physicially linked to the select-
able marker gene. Such linkage, accomplished by
recombinant DNA techniques, increases the probabil-
ity of stable integration-and perhaps of expression-
of the newly introduced globin gene.
An alternate approach for functional correction of
premature termination mutations is the introduction of
a gene for a suppressor tRNA.’98 Such tRNAs “read”
termination codons and can be tailored to insert an
appropriate amino acid into globin. It is not yet clear
whether suppressor tRNA genes could be expressed at
an adequate level or whether suppression of normal
terminators would have deleterious effects on cells.
Many problems remain to be solved, in addition to
the variable and low level of globin gene expression in
transformed erythroid cells, before gene transfer can
be proposed with a reasonable probability of success.
Transformation frequency, still very low for tissue
culture cells, may be increased substantially by use of
retroviral 1 Recombinant retrovi ral genomes,
appropriately packaged into viral coat proteins, may be
introduced into target cell populations with very high
efficiency. Whether cotransformation of cells with
both the gene for a selectable enzyme marker and that
for globin can be accomplished with these retroviral
vectors has not yet been ascertained.
The ethical issues of gene therapy have been exten-
sively discussed.2#{176}#{176}202
Certainly, one could not reason-
ably attempt to introduce genes into human bone
marrow cells for therapeutic purposes until a proce-
dune had been shown to be safe, reproducible, and
effective in appropriate experimental animals. Despite
the significant problems that remain to be solved, this
approach does hold promise and may yet provide a
definite method of treatment for the severe /3-thalasse-
mias.
ACKNOWLEDGMENT
We wish to thank Drs. E. DonneH Thomas, G. Lucarelli, and B.
Carr for communication of unpublished results. We are grateful to
Cathy Magruder and Linda Garza for help in preparation of this
manuscript.
REFERENCES
I . Weatherall Di, Clegg JB: The Thalassaemia Syndromes (ed
3). Oxford, BlackweH, 1981
2. Maniatis T, Fritsch EF, Lauer J, Lawn RM: The molecular
genetics of human hemoglobins. Annu Rev Genet 14:145, 1980
3. Stamatoyannopoulos G, Nienhuis AW: Organization and
Expression ofGlobin Genes. New York, Alan R. Liss, 1981
4. Lauer J, Shen C-KJ, Maniatis T: The chromosomal arrange-
ment of human a-like globin genes: Sequence homology and a-
globin gene deletions. Cell 20: 1 19, 1980
5. Liebhaber SA, Goossens M, Kan YW: Homology and con-
certed evolution at the al and a2 Ioci of human a-gfobin. Nature
290:26, 1981
6. Fritsch EF, Lawn RM, Maniatis T: Molecular cloning and
characterization of the human /3-like globin gene cluster. Cell
19:959, 1980
7. Efstratiadis A, Posakony JW, Maniatis 1, Lawn RM, O’Con-
nell C, Spritz RA, DeRiel J, Forget B, Weissman SM, Slightom JL,
Smithies 0, Baralle FE, Shoulders CC, Proudfoot NJ: The structure
and evolution of the human /3-globin gene family. Cell 2 1:653, 1980
8. Baralle FE, Shoulders CC, Proudfoot NJ: The primary struc-
ture of the human -g1obin gene. Cell 21 :62 1, I 980
9. Slightom JL, Blechl AE, Smithies 0: Human fetal 0y.. and
A.y..globin genes: Complete nucleotide sequences suggest that DNA
can be exchanged between these duplicated genes. Cell 21:627,
I 980
I 0. Proudfoot NJ, Gil A, Maniatis T: The structure of the human
zeta-globin gene and a closely linked, nearly identical pseudogene.
Cell 31:553, 1982
I 1. Proudfoot NJ, Maniatis T: The structure of a human a-globin
pseudogene and its relationship to a-globin gene duplication. Cell
21:537, 1980
I 2. Kaufman RE, Kretschmer PJ, Adams JW, Coon HC, Ander-
son WF, Nienhuis AW: Cloning and characterization of DNA
sequences surrounding the human ‘y-, #{244}-,
and /3-globin genes. Proc
Natl Acad Sci USA 77:4229, 1980
13. Leder P, Hansen iN, Konkel D, Leder A, Nishioka Y,
Talkington C: Mouse globin system: A functional and evolutionary
analysis. Science 209:1336, 1980
14. Dierks P. van Ooyen A, Cochran MD, Dobkin C, Reiser J,
Weissman C: Three regions upstream from the Cap site are required
for efficient and accurate transcription of the rabbit /3-globin gene in
mouse 3T6 cells. Cell 32:695, 1983
Downloaded
from
by
guest
on
28
September
2021

I 5. Mount SM: A catalogue ofsplicejunction sequences. Nucleic
Acid Res 10:459 1982
I 6. Salditt-Georgieff M, Darnell JE: A precise termination site in
the mouse flmaJOt transcription unit. Proc NatI Acad Sci USA
80:4694, 1983
I 7. Proudfoot NJ, Brownlee GG: 3’ Non-coding region sequences
in eukaryotic messenger RNA. Nature 263:21 1, 1976
I 8. Darnell JE: Variety in the level of gene control in eukaryotic
cells. Nature 297:365, 1982
19. Flavell RA, Bernards R, Kooter JM, DeBoer E, Little PFR,
Annison G, Williamson R: The structure of the human /9-globin gene
in j3-thalassemia. Nucleic Acid Res 6:2749, 1979
20. Orkin SH, Kolodner R, Michelson A, Husson R: Cloning and
direct examination ofa structurally abnormal human /9#{176}-thalassemia
globin gene. Proc NatI Acad Sci USA 77:3558, 1980
21. Spritz RA, Orkin SH: Duplication followed by deletion
accounts for the structure ofan Indian deletion /3#{176}-thalassemia gene.
Nucleic Acid Res 10:8025, 1982
22. Mellon P. Parker V. Gluzman Y, Maniatis T: Identification
of DNA sequences required for transcription ofthe human al-globin
gene in a new SV4O host-vector system. Cell 27:279, 19 l
23. Grosveld GC, DeBoer E, Shewmaker CK, Flavell RA: DNA
sequences necessary for transcription of the rabbit /3-globin gene in
vivo. Nature 295:120, 1982
24. Orkin SH, Kazazian HH Jr, Antonarakis SE, Gofi’ SC,
Boehm CD, Sexton JP, Waber PG. Giardina PJV: Linkage of
/3-thalassemia mutations and j3-globin gene polymorphisms with
DNA polymorphisms in human /3-globin gene cluster. Nature
296:627, 1982
25. Treisman R, Orkin SH, Maniatis T: Specific transcription
and RNA splicing defects in five cloned f3-thalassaemia genes.
Nature 302:591, 1983
26. Poncz M, Ballantine M, Solowiejczyk D, Barak I, Schwartz
E, Surrey 5: /3-thalassemia in a Kurdish Jew: Single base changes in
the TATA box. J Biol Chem 257:5994, 1982
27. Antonarakis SE, Orkin SH, Cheng IC, Scott AF, Sexton JP,
Trusko 5, Charache 5, Kazazian HH Jr: f3-Thalassemia in Amen-
can blacks: Novel mutations in the TATA box and IVS-2 acceptor
splice site. Proc NatI Acad Sci USA (in press)
28. Orkin SH, Sexton JP, Cheng I-C, Goff SC, Giandina PJV,
Lee JI, Kazazian HH Jr: ATA box transcriptional mutation in
/3-thalassemia. Nucleic Acid Res I 1:4727, 1983
29. Humphnies RK, Ley I, Turner P. Davis-Moulton A, Nien-
huis AW: Differences in human a-, /3- and #{244}-globingene expression
in monkey kidney cells. Cell 30:173, 1982
30. Ross J, Pizzaro A: Human beta and delta messenger RNAs
turn over at different rates. J Mol Biol 167:607, 1983
3 1. Bernards R, Kooter JM, Flavell RA: Physical mapping of the
globin gene deletion in (#{244}/9)#{176}
thalassemia. Gene 6:265, 1979
32. Fnitsch EF, Lawn RM, Maniatis T: Characterization of
deletions which affect the expression of fetal globin genes in man.
Nature 279:598, 1979
33. Ottolenghi 5, Giglioni B, Comi P. Gianni AM, Polli E,
Acquaye CTA, Oldham JH, Masera G: Globin gene deletion in
HPFH, #{244}#{176}fl#{176}-thalassemia
and Hb Lepore disease. Nature 278:654,
I 979
34. Orkin SH, Alter BP, Altay C: Deletion of the Ay globin gene
in #{176} y#{244}flthalassemia J Clin Invest 64:866, 1979
35. Bernards R, Flavell RA: Physical mapping of the globin gene
deletion in hereditary persistence of fetal hemoglobin (HPFH).
Nucleic Acid Res 8:1521, 1980
36. Van Den Ploeg LHT, Donings A, Oort M, Roos D, Bennini L,
Flavell RA: -yj3-Thalassemia studies showing that deletion of the ‘y-
and #{244}-genesinfluences /3-globin gene expression in man. Nature
283:637, 1980
37. Jones RW, Old JM, Trent Ri, Clegg JB, Weathenall Di:
Restriction mapping of a new deletion responsible for G ( /3)G..
thalassemia. Nucleic Acid Res 9:6813, 1981
38. Jones RW, Old iM, Trent RJ, Clegg JB, Weatherall Di:
Major rearrangement in the human /9 globin gene cluster. Nature
291:39, 1981
39. Orkin SH, Goff SC, Nathan DG: Heterogeneity of DNA
deletion in -y#{244}f3-thalassemia. J Clin Invest 67:878, 1981
40. Ottolenghi S, Giglioni B, Taramelli R, Comi P, Mazza U,
Saglio G, Camaschella C, Izzo P, Cao A, Galanello R, Gimferrer E,
Baiget M, Gianni AM: Molecular comparison of #{244}/3-thalassemia and
hereditary persistence of fetal hemoglobin DNAs: Evidence of a
regulatory area? Proc Natl Acad Sci USA 79:2347, 1982
41. Weatherall Di, Clegg iG: Thalassemia revisited. Cell 29:7,
I982
42. Tuan D, Feingold E, Newman M, Weissman SM, Forget BG:
Different 3’-endpoints of deletions causing #{244}/3-thalessemia and
hereditary persistence of fetal hemoglobin: Implications for the
control of-y-globin gene expression in man. Proc NatI Acad Sci USA
80:6937, 1983
43. Vanin EF, Henthorn PS, Kioussis D, Grosveld F, Smithies 0:
Unexpected relationships between four large deletions in the human
f3-globin gene cluster. Cell 35:701, 1983
44. Kioussis D, Vanin E, deLange I, Flarell RA, Grosveld FG:
/3-globin gene inactivation by DNA translocation in a/3-thalassemia.
Nature 306:662, 1983
45. Schmid CW, Jelinek WR: The Alu family of dispersed
repetitive sequences. Science 216:1065, 1982
46. Adams JW, Kaufman RE, Kretschmer PJ, Harrison M,
Nienhuis AW: A family of long reiterated DNA sequences, one copy
of which is next to the human beta globin gene. Nucleic Acid Res
8:6113, 1980
47. Lerman MI, Thayer RE, Singer MF: Kpn I family of long
interspersed repeated DNA sequences in primates: Polymorphism of
family members and evidence for transcription. Proc Nail Acad Sci
USA 80:3966, 1983
48. Forget BG, Tuan D, Biro PA, Jagadeeswaran P. Weissman
SM: Structural features of the DNA flanking the human non-a
globin genes: Implications in the control of fetal hemoglobin switch-
ing. Trans Assoc Am Physicians 94:204, 1981
49. Banerji J, Rusconi 5, Schaffner W: Expression of a /3-globin
gene is enhanced by remote SV4O DNA sequences. Cell 27:299,
I981
50. Chao MV, Mellon P. Charney P, Maniatis I. Axel R: The
regulated expression of f3-globin genes introduced into mouse eryth-
roleukemia cells. Cell 32:483, 1983
51. Khoury G, Gruss P: Enhancer elements. Cell 33:313, 1983
52. Gillies SD, Morrison SL, Oi VT, Tonegawa 5: A tissue-
specific transcription enhancer element is located in the major intron
of a rearranged immunoglobulin heavy chain gene. Cell 33:7 17,
I983
53. Banerji J, Olson L, Schaffner W: A lymphocyte-specific
cellular enhancer is located downstream of the joining region in
immunoglobulin heavy chain genes. Cell 33:729, 1983
54. Queen C, Baltimore D: Immunoglobulin gene transcription is
activated by downstream sequence elements. Cell 33:741, 1983
55. Emonine L, Kuehl M, Weir L, Leder P. Max EE: A conserved
sequence in the immunoglobulin Jk-Ck intron: Possible enhancer
element. Nature 304:447, 1983
56. Weisbord 5: Active chromatin. Nature 297:289, 1982
57. Groudine M, Eisenman R, Gelinas R, Weintraub H: Devel-
opmental aspects of chromatin structure and gene expression, in
Stamatoyannopoulos G, Nienhuis AW (eds): Globin Gene Expres-
sion and Hematopoietic Differentiation. New York, Alan Liss, 1983,
p 159
Downloaded
from
by
guest
on
28
September
2021

THALASSEMIA 755
58. Shermoen AW, Beckendorf 5K: A complex of interacting
DNAse 1-hypersensitive sites near the Drosophila glue protein gene,
Sgs 4. Cell 29:601, 1982
59. Huisman THJ, Schroeder WA, Efremov GD, Duma H,
Mladenovski B, Hyman CB, Rachmilewitz EA, Bouver N, Miller A,
Brodie A, Shelton JR. Shelton iB, Apell G: The present status of the
heterogeneity of fetal hemoglobin in /3-thalassemia: An attempt to
unify some observations in thalassemia and related conditions. Ann
NY Acad Sci 232:107, 1974
60. Duncan C, Biro PA, Choudary PV, Elder iT, Wang RRC,
Forget BG, DeRiel JK, Weissman 5: RNA polymerase Ill tran-
scriptional units are interspersed among human non-a-globin genes.
Proc NatI Acad Sci USA 76:5095, 1979
6 I . Fnitsch EF, Shen CKJ, Lawn RM, Maniatis I: The organiza-
tion of repetitive sequences in mammalian globin gene clusters. Cold
Spring Harbor Symp Quant Biol 45:76 1, I 981
62. Ottolenghi 5, Giglioni B: The deletion in a type of #{176}-
/3#{176}-thalassemia begins in an inverted Alu I repeat. Nature 300:770,
I982
63. Gianni AM, Bregni M, Cappellini MD, Fiorelli G, Taramelli
R, Giglioni B, Comi P. Ottolenghi 5: A gene controlling fetal
hemoglobin expression in adults is not linked to the non-a globin
cluster. J Eur Mol Biol Org 2:921, 1983
64. Old JM, Ayyub H, Wood WG, Clegg iB, Weatherall DJ:
Linkage analysis of non-deletion hereditary persistence of fetal
hemoglobin. Science 21 5:98 1, 1982
65. Maquat LE, Kinnburgh Ai, Beach LR, Honig GR, Lazerson
J, Ershler WB, Ross J: Processing of human /3-globin mRNA
precursor to mRNA is defective in three patients with /3L thalasse
mia. Proc NatI Acad Sci USA 77:4287, 1980
66. Kantor iA, Turner PH, Nienhuis AW: Beta thalassemia:
Mutations which affect processing of the /3-globin mRNA precursor.
Cell2I:I49, 1980
67. Benz EJ Jr. Scarpa AL, Tonkonow BL, Pearson HA, Ritchey
AK: Post-transcriptional defects in /3 globin mRNA metabolism in
f3-thalassemia: Abnormal accumulation of /3 mRNA precursor
sequences. J Clin Invest 68:1529, 1981
68. Ley Ti, Anagnou NP, Pepe G, Nienhuis AW: RNA process-
ing errors in patients with /3-thalassemia. Proc NatI Acad Sci USA
79:4775, 1982
69. Chang JC, Kan YW: /3#{176}
Thalassemia: A nonsense mutation in
man. Proc NatI Acad Sci USA 76:2886, 1979
70. Spnitz RA, iagadeeswaran P. Choudary PV, Biro PA, Elder
iT, DeRieI iK, Manley iL, Gefter ML, Forget BG, Weissman SM:
Base substitution in an intervening sequence of a fl -thalassemic
human globin gene. Proc Natl Acad Sci USA 78:2455, 1981
7 1. Westaway D, Williamson R: An intron nucleotide sequence
variant in a cloned fl -thalassemia globin gene. Nucleic Acid Res
9:1777, 1981
72. Baird M, Dniscoll C, Schreiner H, Sciarratta GV, Sansone G,
Niazi G, Ramirez F, Bank A: A nucleotide change at a splice
junction in the human /3-globin gene is associated with /3#{176}-thalasse-
mia. Proc Nail Acad Sci USA 78:4218, 1981
73. Orkin SH, GoffSC: Nonsense and frameshift mutations in /3#{176}
thalassemia detected in cloned /3-globin genes. i Biol Chem
256:9782, 1981
74. Trecartin RF, Liebhaber SA, Chang iC, Lee KY, Kan YW:
f3#{176}
Thalassemia in Sardinia is caused by a nonsense mutation. i Clin
Invest 68:1012, 1981
75. Moschonas N, de Boer E, Grosveld FG, Dahl HHM, Wright
5, Shewmaker CK, Flavell RA: Structure and expression of a cloned
/3#{176}
thalassemia globin gene. Nucleic Acid Res 9:4391, 1981
76. Busslinger M, Moschonas N, Flavell RA: /3 Thalassemia:
Aberrant splicing results from a single point mutation in an intron.
Cell 27:289, 1981
77. Pergolizzi R, Spritz RA, Spence 5, Goossens M, Kan YW,
Bank A: Two cloned /3 thalassemia genes are associated with amber
mutations at codon 39. Nucleic Acid Res 9:7065, 1981
78. Gorski i, Fioni M, Mach B: A new nonsense mutation as the
molecular basis for /3#{176}
thalassemia. i Mol Biol 154:537, 1982
79. Fukumaki Y, Ghosh PK, Benz Ei in, Reddy VB, Lebowitz P.
Forget BG, Weissman SM: Abnormally spliced messenger RNA in
erythroid cells from patients with /3 thalassemia and monkey cells
expressing a cloned /3 thalassemia gene. Cell 28:585, 1982
80. Spence SE, Pergolizzi RG, Donovan-Peluso M, Kosche KA,
Dobkin CS, Bank A: Five nucleotide changes in the large intervening
sequence of a /3 globin gene in a fl thalassemia patient. Nucleic
Acid Res 10:1283, 1982
81. Goldsmith ME, Humphnies RK, Ley I, Cline A, Kantor i,
Nienhuis AW: Silent nucleotide substitution in a /3 -thaIassemia
globin gene activates splice site in coding sequence RNA. Proc NatI
Acad Sci USA 80:2318, 1983
82. Treisman R, Proudfoot NJ, Shander M, Maniatis I: A
single-base change at a splice site in a /3#{176}-thalassemic gene causes
abnormal RNA splicing. Cell 29:903, 1982
83. Kinniburgh AJ, Maquat LE, Schedl I, Rachmilewitz E, Ross
J: mRNA deficient /3#{176}-thalassemia results from a single nucleotide
deletion. Nucleic Acid Res 10:5421, 1982
84. Kimura A, Matsunaga E, Takihara Y, Nakamura I, Takagi
Y, Lin 5, Lee H: Structural analysis of a /3-thalassemia gene found
in Taiwan. i Biol Chem 258:2748, 1983
85. Dobkin C, Pergolizzi RG, Bahre P, Bank A: Abnormal splice
in a mutant human /3-globin gene not at the site of a mutation. Proc
NatI Acad Sci USA 80:1 184, 1983
86. Orkin SH, Sexton iP, GoffSC, Kazazian HH: Inactivation of
an acceptor RNA splice site by a short deletion in /3-thalassemia. i
Biol Chem 258:7249, 1983
87. Atweh GA, Anagnou NP, Kaufman RE: /3-thalassemia
caused by a 3’ splice site mutation in IVS. Blood 62(suppl):64a,
1983, (abstr)
88. Breathnach R, Benoist C, O’Hare K, Gannon F, Chambon P:
Ovalbumin gene: Evidence for a leader sequence in mRNA and
DNA sequences at the exon-intron boundaries. Proc NatI Acad Sci
USA 75:4853, 1978
89. Rogers i, Wall R: A mechanism for RNA splicing. Proc Nail
Acad Sci USA 77:1877, 1980
90. Mount SM, Pettersson I, Hinterberger M, Kurmas A, Steitz
JA: The U I small nuclear RNA-protein complex selectively binds a
r splice site in vitro. Cell 33:509, 1983
91. Felber BK, Orkin SH, Hamer DH: Abnormal RNA splicing
causes one form of thalassemia. Cell 29:895, I 982
92. Orkin SH, Kazazian HH, Antonarakis SE, Ostrer H, Goff
SC, Sexton iP: Abnormal RNA processing due to the exon mutation
of/3#{176}-globin gene. Nature 300:768, 1982
93. Arous N, Galacteros F, Fessas P. Loukopoulos D, Blouguit Y,
Komis G, Sellaye M, Boussiou M, Rosa i: Structural study of
hemoglobin Knossos, /3 27(B9) Ala-’.Ser. FEBS Lett 147:247,
I982
93a. Cheng I, Orkin SH, Antonarakis SE, Potter Mi, Sexton IP,
Markham AF, Giardina PJV, Li A, Kazazian HH: /3-Thalassemia in
Chinese: Use of in viva RNA analysis and olignonucleotide hybnidi-
zation in systematic characterization of molecular defects. Blood
62(suppl):64a, I 983 (abstr)
94. Khoury G, Gruss P, Dhar R, Lai C: Processing and expression
ofeanlySV4OmRNA.Cell 18:85, 1979
95. Fitzgerald M, Shenk I: The sequence 5’-AAUAAA-3’ forms
part of the recognition site for polyadenylation of late SV4O
mRNAs. Cell 24:251, 1981
96. Higgs DR. Goodbourn SEY, Lamb i, Clegg iB, Weatherall
Downloaded
from
by
guest
on
28
September
2021

Advances In Thalassemia Research

Advances In Thalassemia Research

Recommended

Recommended

More Related Content

Similar to Advances In Thalassemia Research

Similar to Advances In Thalassemia Research (20)

More from Kim Daniels

More from Kim Daniels (20)

Recently uploaded

Recently uploaded (20)

Advances In Thalassemia Research