A humanized mouse model for a common β0-thalassemia mutation

Genomics 85 (2005) 453 – 461 www.elsevier.com/locate/ygeno

A humanized mouse model for a common h0-thalassemia mutation Duangporn Jamsaia,b, Faten Zaibaka, Wantana Khongniumb, Jim Vadolasa, Lucille Voullairea, Kerry J. Fowlerc, Sophie Gazeasc, Suthat Fucharoenb, Robert Williamsona, Panayiotis A. Ioannoua,d,* a

CAGT Research Group, The Murdoch Children’s Research Institute, Department of Paediatrics, The University of Melbourne, Royal Children’s Hospital, Flemington Road, Melbourne, VIC 3052, Australia b Thalassemia Research Center, Institute of Science and Technology for Research and Development and Institute of Molecular Biology and Genetics, Mahidol University, Nakornpathom 73170, Thailand c Department of Paediatrics, Disease Model Unit, The Murdoch Children’s Research Institute, The University of Melbourne, Royal Children’s Hospital, Flemington Road, Melbourne, VIC 3052, Australia d The Cyprus Institute of Neurology and Genetics, Nicosia, Cyprus Received 7 August 2004; accepted 29 November 2004 Available online 1 February 2005

Abstract Accurate animal models that recapitulate the phenotype and genotype of patients with h-thalassemia would enable the development of a range of possible therapeutic approaches. Here we report the generation of a mouse model carrying the codons 41–42 ( TTCT) h-thalassemia mutation in the intact human h-globin locus. This mutation accounts for approximately 40% of h-thalassemia mutations in southern China and Thailand. We demonstrate a low level of production of g-globins from the mutant locus in day 18 embryos, as well as production of mutant human h-globin mRNA. However, in contrast to transgenic mice carrying the normal human h-globin locus, 4-bp deletion mice fail to show any phenotypic complementation of the knockout mutation of both murine h-globin genes. Our studies suggest that this is a valuable model for gene correction in hemopoietic stem cells and for studying the effects of HbF inducers in vivo in a bhumanizedQ thalassemic environment. D 2004 Elsevier Inc. All rights reserved. Keywords: Bacterial artificial chromosomes; Hemoglobin; h-Thalassemia; 4-bp deletion mutation; Transgenic mice; Knockout mice; Gene correction; HbF induction

h-Thalassemia is one of the most common monogenic disorders in humans. It is characterized by reduced or absent h-globin synthesis due to mutations in the adult h-globin gene or its cis-acting regulatory elements [1,2]. These disorders can be divided into two major groups, h0 and h+, based on the quantity of h-globin produced; h0-thalassemia results from the complete absence of the h-globin chain, and h+-thalassemia results from a decrease in h-globin production. The phenotype of specific h-thalassemia mutations is * Corresponding author. CAGT Research Group, The Murdoch Children’s Research Institute, Department of Paediatrics, The University of Melbourne, Royal Children’s Hospital, Flemington Road, Melbourne, VIC 3052, Australia. Fax: +61 3 8341 6212. E-mail address: [email protected] (P.A. Ioannou). 0888-7543/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ygeno.2004.11.016

related directly to the quantitative deficit in h-globin gene production and the resulting imbalance in the a:h-globin chain synthesis ratio. Thalassemia major, or transfusiondependent h-thalassemia, is characterized by retarded growth and development resulting from a marked hemolytic anemia, with compensatory expansion of hemopoietic tissues and a consequent hypermetabolism state [3–5]. Current disease management consists of lifelong blood transfusions and iron chelation. However, chronic transfusion is complicated by high rates of alloimmunization and blood-borne infectious sequelae [6,7]. Allogeneic bone marrow transplantation is curative, but this option is limited to a minority of patients for whom a histocompatible donor can be identified [8]. In addition, the high cost involved means this treatment is not a feasible option in many

454

D. Jamsai et al. / Genomics 85 (2005) 453–461

developing countries where thalassemia is a common disease. There is clearly an urgent need for the development of new curative approaches for h-thalassemia. The h-thalassemias are caused by more than 200 different mutations in the functionally important regions of the h-globin gene [9]. Most of the mutations are point mutations that either decrease the level of mature mRNA or lead to nonfunctional mRNA, as is the case with nonsense and frameshift mutations. The codons 41–42 4-bp deletion ( TTCT) causes a frameshift and creates a stop codon at the new codon 59 position, thus resulting in h0-thalassemia. This particular mutation is the most frequent among patients in southern China and Thailand, accounting for 40% of h-thalassemia in some regions [10,11]. Five types of mouse models for h-thalassemia have been previously described. Heterozygotes for a naturally occurring deletion of the adult hmaj gene (muhth-1/+) show very mild thalassemia, with about 60% of mice homozygous for this deletion (muhth-1/th-1) surviving to adulthood [12]. In contrast, mice homozygous for an insertional disruption of the hmaj gene (muhth-2/th-2) do not survive more than a few hours after birth. The heterozygotes are anemic and have features of thalassemia similar to those found in human h-thalassemia intermedia [13]. No naturally occurring h0-thalassemia has been observed in mice but two additional knockout mouse models (muh0/+ and muhth-3/+) were produced by complete deletion of both the hmaj and the hmin genes [14,15]. The heterozygotes for these two models display microcytic anemia and splenomegaly, similar to h-thalassemia intermedia, while mice homozygous for these deletions (muh0/0 and mu th-3/th-3 h ) die during fetal development. These models recapitulate many aspects of the thalassemic phenotype and can be used in gene therapy studies that are based on gene supplementation [16,17]. Nonetheless, none of these models recapitulates h-thalassemia at the genotypic level, thus limiting the development of potential therapeutic strategies that are not based on gene supplementation (e.g., gene correction, restoration of splicing specificity by antisense oligonucleotides, or pharmacological upregulation of HbF expression). To overcome some of these limitations, a knock-in model carrying a single copy of the human h-globin gene with the IVS II-654 (C Y T) splicing mutation in place of the hmaj and hmin genes (muhth-4/+) was developed more recently [18]. Since this splicing mutation does not allow any significant amount of normal splicing, it has the same phenotypic effects as the muh0 and muhth-3 mutations. This represents the only animal model of a common hthalassemia splicing mutation in the context of the human h-globin gene sequence, allowing its use for the in vivo development of antisense oligonucleotides to restore splicing specificity as a potential approach to the therapy of common splicing defects in thalassemia and other genetic diseases. Studies in transgenic mice have demonstrated that the normal gene order and spatial organization of the members

of the human h-globin gene family are required for appropriate developmental and stage-specific expression of the genes [19]. As the cis-acting sequences that participate in the activation and silencing of the g- and h-globin genes are not yet fully defined, mouse models that preserve the normal structure of the locus are likely to have significant advantages for validating future therapies for h-thalassemias. We have used a 183-kb genomic fragment, originally isolated from the first human PAC library [20], to study expression of each of the globin genes in episomal and stable genomic reporter assays [21–23]. Our studies demonstrate regulated expression of the fetal and adult globin genes in these constructs in human K562-derived erythroleukemia cells lines. More recently, we have also demonstrated that the human h-globin locus in the unmodified 183-kb fragment can fully complement the embryonic lethality of homozygous knockout embryos (muhth-3/th-3), allowing the development of the first mouse models that are totally dependent on h-globin production from the normal human h-globin locus (Vadolas et al., manuscript in preparation). In preliminary studies to develop mouse models that accurately recapitulate thalassemia at both the phenotypic and the genotypic levels, we have also introduced the codon 26 HbE (G Y A) mutation [24], three common splicing mutations [IVS I-110 (G Y A), IVS I-5 (G Y C), and IVS II-654 (C Y T)] [25,26], the codon 39 (C Y T) stop mutation [27], and the codons 41–42 ( TTCT) deletion [24] into the same 183-kb genomic fragment. In this report, we describe the development and characterization of the first transgenic mouse model carrying a common h0-thalassemia mutation in the context of the intact human h-globin locus.

Results Characterization of transgene fragment The purified 183-kb fragment containing the human hglobin locus with the codons 41–42 ( TTCT) deletion mutation was microinjected into fertilized eggs to generate transgenic mice. We obtained 22 pups of which 2 (9%) were transgenic. Transgenic founder 1 failed to transmit the transgene to the F1 generation, presumably due to germ-line mosaicism. The overall integrity of the transgene fragment in the two founders was evaluated by PCR using primer pairs for 5VHS-5, 5VHS-2, and human q-, human g-, and human h-globin genes, as described elsewhere (Vadolas et al., manuscript in preparation). Both founders produced all the expected fragments (Fig. 1A, lanes L1 and L2). PCR analysis using primers specific for both the 5V and the 3V ends of the 183-kb genomic fragment also demonstrated that both lines contained intact 5Vand 3Vends (Fig. 1B, lanes L1 and L2). FISH analysis performed from fibroblasts using the pEBAC/148h probe showed a single site of integration of

D. Jamsai et al. / Genomics 85 (2005) 453–461

455

Fig. 1. Characterization of 4-bp deletion transgenic mice. (A) Transgene integrity examined using five regions within the human h-globin locus region. Lane M, molecular weight marker X (Roche); lane C, pEBAC/148h unmodified clone; lanes L1 and L2, founder mice L1 and L2. (B) Same as (A), using primers specific for the 5Vand 3Vends of the genomic fragment. (C) Southern blot analysis of EcoRI- and HpaI-digested genomic DNA from F1 transgenic mice of line L2, probed with a 447-bp probe from the human h-globin gene. The probe detects the corresponding sequence in the human h-globin gene (2.2 kb), as well as the homologous y-globin gene sequence (2.0 kb). Lane H1, human genomic DNA (1 Ag); lane H2, human genomic DNA (10 Ag); lane L2, 4-bp deletion transgenic line 2 genomic DNA (10 Ag). (D) RT-PCR analysis of total blood RNA using primers specific for the human h-globin gene sequence. M, molecular weight marker VIII (Roche); H, human patient blood (homozygous for the codons 41–42 4-bp deletion); WT, wild-type mouse; N, transgenic mouse carrying the normal human h-globin locus; L1 and L2, 4-bp deletion founder mice lines L1 and L2, respectively.

the transgene (data not shown). Transgene copy number estimated by Southern blot hybridization using a human h-globin probe and human genomic DNA as internal control showed that the transmitting line (line L2) contained six copies of the transgene (Fig. 1C, lane L2). Analysis of genomic DNA after digestion with rarecutting endonucleases, pulsed-field gel electrophoresis, and Southern blot hybridization is commonly used to detect long-range integrity. However, such studies are very uninformative regarding the overall integrity of large transgene fragments, especially when multiple copies of a transgene are integrated at the same site. We have therefore focused our subsequent studies on demonstrating the functionality of the transgene. Expression of the 4-bp deletion transgene RT-PCR was performed on peripheral blood samples of the 4-bp deletion transgenic mice using primers flanking exons 1 and 2 (G2-FW and G1-Rev primers) of the h-globin gene. These primers amplify only human h-globin sequences to give a 160-bp product of the normal h-globin cDNA and a 156-bp product of the hD4bp cDNA. Both founder mice produced the expected 156-bp product (Fig. 1D, lanes L1 and L2), suggesting regulated expression of the human h-globin gene in one or more copies of the mutated transgene. Sequencing of the 156-bp products confirmed the presence of the 4-bp deletion mutation at codons 41–42 (data not shown). Surprisingly, a 118-bp PCR product was also detected in both founder mice. This fragment could not be detected in a blood sample from a patient homozygous for the 4-bp deletion mutation (Fig. 1D, lane H) or in transgenic mice carrying the normal human h-globin locus

(lane N). Sequencing of this aberrant splicing product revealed a deletion of 38 bp of exon 1 (codons 18–30) (data not shown). It thus appears that the codons 41–42 deletion in exon 2 of the h-globin gene somehow activates an aberrant 5V cryptic site 38 bp upstream of the normal 5V donor splice site of intron I. Although the aberrant splice product was also not detectable in a second patient mRNA sample (data not shown), we have not examined for its presence in patient mRNA by allele-specific PCR. It is thus possible that it is a unique product of the way mouse spliceosomes recognize the mutant human h-globin sequence. On the other hand, the stability of the aberrant product may be different in the human and murine erythropoietic environments, while functional polymorphisms in spliceosome components could also induce significant variation in aberrant splicing patterns between different patients. It is hoped that further studies to resolve these issues in this model will lead to a better understanding of the mechanisms determining splicing specificity with normal and mutant pre-mRNA sequences. Hematological studies Breeding of transgenic mice carrying the 4-bp deletion (huhD4bp/0, muh+/+) with heterozygous h-globin knockout (muhth-3/+) mice allowed the generation of mice carrying the 4-bp deletion transgene on a heterozygous knockout background (huhD4bp/0, muhth-3/+) (abbreviated as DHD4bp mice). Hematological parameters in these mice were determined and compared with the corresponding values of wild-type (muh+/+) mice and heterozygous knockout (muhth-3/+) mice, as well as data from transgenic mice carrying the normal human h-globin locus on a hetero-

456

Genotype mu +/+

Wild-type ( h ) (n = 15) Heterozygous KO (muhth-3/+) (n = 12) DHN(F-6) a (n = 10) DHD4bp b (n = 12)

Hb (g/dl)

HCT (%)

RBC (109/ml)

MCV (fl)

MCH (pg)

RDW (%)

Retic (%)

Spleen (mg)

Bilirubin (mg/dl)

16.3 F 1.1 8.4 F 0.5

49.4 F 3.0 28.5 F 1.1

10.9 F 0.7 5.1 F 0.5

51.7 F 1.2 45.8 F 2.7

15.2 F 1.2 13.8 F 0.6

13.6 F 1.3 36.6 F 1.3

5.5 F 3.3 (n = 5) 37.4 F 6.3 (n = 7)

77 F 16 (n = 12) 556 F 52 (n = 11)

0.13 F 0.04 (n = 6) 0.41 F 0.17 (n = 6)

15.0 F 0.7 8.3 F 0.8

50.1 F 1.7 28.3 F 0.8

9.6 F 0.2 5.4 F 0.5

51.6 F 1.7 42.6 F 3.2

15.4 F 0.8 12.4 F 0.7

14.8 F 1.3 32.5 F 1.7

2.1 F 0.4 39.0 F 5.8 (n = 3)

ND 551 F 45 (n = 5)

ND 0.28 F 0.02 (n = 3)

Hematological values are expressed as means F SD. Hb, hemoglobin concentration; HCT, hematocrit; RBC, red blood cells; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; RDW, red cell distribution width; Retic, reticulocyte count. a Representative data from line F-6 carrying two copies of the normal human h-globin locus on a heterozygous knockout background (huh+/0, muhth-3/+) (Vadolas et al., manuscript in preparation). b DHD4bp denotes animals carrying the 4-bp deletion construct on a heterozygous knockout background.

D. Jamsai et al. / Genomics 85 (2005) 453–461

Table 1 Hematological data of transgenic mice carrying the human hD4bp-globin locus on a mouse h-globin knockout background

D. Jamsai et al. / Genomics 85 (2005) 453–461

zygous knockout background (Vadolas et al., manuscript in preparation). As previously reported [14,15] and confirmed in our studies, heterozygous knockout mice have dramatically decreased hematocrit, hemoglobin (Hb), and red blood cell counts, typical of the thalassemia intermedia phenotype of mice with this genotype. These changes are accompanied by large increases in reticulocyte counts, red cell distribution width, serum bilirubin, and spleen size, indicating active but ineffective erythropoiesis and hemolysis (Table 1). Our studies with four independent transgenic lines carrying the unmodified human h-globin locus on a heterozygous knockout background (Vadolas et al., manuscript in preparation) (representative data shown in Table 1 for only one transgenic line) have also demonstrated that expression of the normal human hglobin locus in the 183-kb transgene can fully complement all the hematological abnormalities of heterozygous knockout mice. In contrast, the hematological data for DHD4bp mice fail to show any hematological complementation, as they are essentially identical to the results of heterozygous knockout mice (Table 1). This is as expected, as the 4-bp deletion mutation should abolish all human h-globin protein production in adult transgenic mice.

457

Examination of peripheral blood smears from wild-type and DHD4bp mice demonstrated marked anisocytosis, poikilocytosis, and target cells (Fig. 2A), which was similar to the red cell morphology in heterozygous knockout mice. The data in Table 1 and Fig. 2B also demonstrate a large increase in spleen size of DHD4bp mice, similar to the increase in heterozygous knockout mice. Although visual examination of internal organs indicated that the increase in size was restricted to the spleen, a more systematic examination was carried out by analyzing spleens and livers of DHD4bp mice at 8 weeks of age, with age- and sexmatched controls. Spleen and liver weights were determined as a ratio to total body weight and compared to wild-type and heterozygous h-globin knockout mice. Body weights of DHD4bp mice at 2 months of age were also compared to those of wild-type animals, but no significant difference was found (data not shown). There was no significant difference in the ratio for liver between wild-type (51.2 F 0.20 mg/g), heterozygous knockout (51.6 F 0.15 mg/g), and DHD4bp mice (50.9 F 0.25 mg/g). There was, however, a highly significant increase ( p b 0.001) in the spleen weight/body weight ratio from the heterozygous h-globin knockout (20.23 F 2.17 mg/g) and DHD4bp mice (19.33 F 2.52

Fig. 2. Phenotypic analysis of the 4-bp deletion mice. (A) Peripheral blood smear examination from a wild-type mouse (WT) demonstrating normal erythrocyte morphology, in contrast to the thalassemia intermedia phenotype found in the blood smear from a DHD4bp mouse or a simple heterozygous knockout mouse (not shown). (B) Normal spleen from a WT mouse reveals the gross enlargement of the spleen from a DHD4bp mouse or a heterozygous knockout mouse (not shown). (C) Examination of embryos from a DHD4bp  DHD4bp cross at 18 days of gestation. WT, wild-type; T, 4-bp deletion transgenic mouse on a normal mouse background (huhD4bp/0, muh+/+); KO, heterozygous knockout mouse (muhth-3/+); DHD4bp, 4-bp deletion transgenic mouse on a heterozygous knockout background (huhD4bp/0, muhth-3/+); R, hemizygous or homozygous 4-bp deletion transgenic mouse on a homozygous knockout background (huhD4bp/0, muhth-3/th-3 or hu D4bp/D4bp mu th-3/th-3 h , h ).

458

D. Jamsai et al. / Genomics 85 (2005) 453–461

mg/g) compared to wild-type control mice (2.58 F 0.24 mg/ g) (n = 6). Characterization of globin gene expression from the 4-bp deletion transgene on a homozygous knockout background Crosses of DHD4bp  DHD4bp mice were set up. In contrast to similar crosses with DH mice carrying the normal human h-globin locus (Vadolas et al., manuscript in preparation), there were no transgenic animals on a homozygous knockout background (huhD4bp/0, muhthe-3/th-3 or huhD4bp/D4bp, muhthe-3/th-3) among the progeny. However, analysis of embryos at day 18 of gestation showed that homozygous knockout embryos carrying the transgene were surviving until late in pregnancy. Such embryos appeared to be severely anemic and smaller in size than wild-type, transgenic on a wild-type background, or DHD4bp littermates (Fig. 2C). To examine the globin chains being produced in these embryos, fetal liver hemolysates were prepared and analyzed by Triton X acid urea gel electrophoresis, in parallel with a HbE/4-bp deletion patient blood sample as control. Although there was significant variability in the protein content of these hemolysates, there was no detectable mouse or human h-globin expression in homozygous knockout embryos carrying the transgene at day 18 (Fig. 3, lanes 5 and 7). On the other hand, expression of both the Gg- and Ag-globin genes was detectable in all animals carrying the 4-bp deletion transgene (lane 2, transgenic embryo on normal background; lanes 3 and 6, DHD4bp embryos; lanes 5 and 7, transgenic embryos on a homozygous knockout background). There also appears to

Fig. 3. Globin chain analysis of 4-bp deletion transgenic mice. Analysis of globin chains by Triton X acid urea gel electrophoresis. Fetal liver hemolysates were prepared from embryos from a DHD4bp  DHD4bp cross at 18 days of gestation. Lane 1, wild-type mouse; lane 2, 4-bp deletion transgenic mouse on a normal mouse background (huhD4bp/0, muh+/+); lanes 3 and 6, 4-bp deletion transgenic mice on a heterozygous knockout background (huhD4bp/0, muhth-3/+); lane 4, heterozygous knockout mouse (muhth-3/+); lanes 5 and 7, hemizygous or homozygous transgenic mice on a homozygous knockout background (huhD4bp/0, muhth-3/th-3 or huhD4bp/D4bp, mu th-3/th-3 h ); lane 8, human patient blood (compound heterozygous for 4-bp deletion and HbE mutations). The positions of the main hemoglobin bands are indicated. Only low levels of Gg- and Ag-globins are detectable in transgenic mice at this stage of fetal development.

be a higher level of g-globin gene expression in DHD4bp embryos (lanes 3 and 6) than in embryos carrying the transgene on a wild-type background (lane 2), suggesting some delay in the switching off of g-globin gene expression on a heterozygous knockout background. Expression of the G g-globin gene was also higher than expression of the Agglobin gene, as is normally the situation during human fetal development. Despite these interesting observations that demonstrate regulated expression of the g-globin genes during fetal development from our 4-bp deletion transgene, the level of g-globin expression was low compared to the endogenous mouse a- and h-globin genes, due to the premature switching off of g-globin gene expression at day 13.5 [28,29] (Vadolas et al., manuscript in preparation). Therefore, the residual g-globin expression from the 4-bp deletion transgene at 18 days is not enough to compensate for the complete absence of human and mouse h-globin gene expression in homozygous knockout embryos carrying the human transgene.

Discussion Gene knockout strategies have been invaluable for elucidating the function of many genes and for creating cellular and animal models for gene therapy. However, simple gene knockout models are not useful for the development of therapeutic strategies that do not depend on gene supplementation. Inserting common disease-causing mutations into the corresponding mouse genes (knockin strategy) allows some of these limitations to be overcome. However, given the significant sequence variation that exists between mouse and human, it would be most desirable for therapeutic strategies in mouse models to be developed by targeting human sequences. This is greatly facilitated now by the availability of most human genes as intact functional loci in sequenced PAC/BAC clones from the Human Genome Project, the development of techniques for the insertion of disease-causing mutations and other fine modifications in such clones [21,24–27], and the efficient generation of transgenic mice by microinjection of large genomic DNA fragments into fertilized oocytes. In the case of h-thalassemia, potential therapeutic strategies other than gene supplementation with viral vectors include the pharmacological reactivation of HbF expression, the direct correction of mutations in the genomic DNA of hemopoietic stem cells, and the restoration of splicing specificity by antisense oligonucleotides for many of the splicing mutations. Additional therapeutic strategies may also become apparent as the factors that moderate the severity of h-thalassemia become better defined. The development of these approaches would be greatly facilitated by the availability of mouse models for thalassemia that accurately recapitulate the disease at both the phenotypic and the genotypic levels using human genomic sequences.

D. Jamsai et al. / Genomics 85 (2005) 453–461

We describe here the first mouse model carrying a common h-thalassemia mutation in the context of an otherwise normal human h-globin locus in a 183-kb transgene. In contrast to the normal 183-kb transgene that is able to complement fully the hematological abnormalities caused by heterozygosity or homozygosity for the mu th 3 h knockout mutation, the mutant transgene, carrying the codons 41–42 4-bp deletion, fails to show any complementation. We further demonstrate that the mutant transgene is functionally intact, in that it is able to drive the expression of g-globin during fetal development and of hglobin mRNA with the 4-bp deletion in adult mice. Thus the failure of the mutant transgene to show any complementation is specifically due to the failure to produce functional human h-globin protein from the mutant mRNA and not to any failure to express the globin genes as a result of integration site-dependent gene silencing. In humans, the switch from g- to h-globin is completed in the first few months after birth, after which the level of HbF is usually less than 1% of the total hemoglobin. However, in patients with homozygous h-thalassemia the silencing of the g-globin genes is often delayed, allowing the production of increased levels of HbF for a number of years. While in most untransfused patients the residual production of HbF is not enough to sustain life for more than a few years [3,4], a number of genetic factors can allow the production of substantial amounts of HbF in some patients, with a corresponding moderating impact on the severity of h-thalassemia [30,31]. In contrast, the g- to h-globin switch in transgenic mice is completed 6–8 days before birth (days 12–14 of gestation) [28,29]. Furthermore, in transgenic mice with the normal human h-globin locus, at least four copies of the transgene appear necessary to drive the production of sufficient human h-globin to allow homozygous knockout embryos to develop normally (Vadolas et al., manuscript in preparation). In the absence of any functional h-globin, as in embryos carrying the 4-bp deletion transgene on a homozygous knockout background, death in the final stages of fetal development is therefore inevitable. Although our studies suggest some delay in the silencing of g-globin expression in our thalassemic mice at E18 as in humans, the amount of residual g-globin production is too low to allow normal development of embryos through the last few days of pregnancy. Direct correction of mutations in genomic DNA has been the holy grail of gene therapy for many years, yet no convenient mouse model has been available to explore such a strategy in h-thalassemia. Our initial studies on gene correction (Zaibak and Ioannou, manuscript in preparation) suggest that our 4-bp deletion transgenic mice fulfill the key requirements for such a model, including: (a) easy maintenance of the animal colony, (b) readily available primary cells (bone marrow and also spleen in DHD4bp mice) for in vitro gene correction studies, (c) easy detection of gene correction by allele-specific PCR/RT-PCR assays, and (d)

459

zero background production of human h-globin protein and convenient detection of human h-globin by immunostaining and HPLC analysis, if a correction efficiency of 1% or more is attained. An alternative promising therapeutic strategy for hthalassemia aims at the pharmacological reactivation of HbF expression. Preclinical evaluation of promising agents is usually carried out in a human erythroid cell culture system [32,33] or in a YAC transgenic mouse model carrying the normal human h-globin locus on a normal mouse background [34]. This model does not reproduce the thalassemic phenotype of patients that often preconditions for a higher level of HbF expression. Our studies demonstrate that the DHD4bp model described here accurately recapitulates a thalassemia intermedia environment and should thus provide a more physiologically relevant response in the assessment of HbF inducers. Erythroid cell cultures from this model may also provide a more controlled and reproducible alternative to the human erythroid cell culture system. In conclusion, we describe the first transgenic mouse model that carries a common h0-thalassemia mutation in an otherwise intact human h-globin locus. Maintenance of the mutant transgene on a heterozygous knockout background produces a model that recapitulates the key features of thalassemia at both the phenotypic and the genotypic levels. This model should greatly facilitate studies for the direct correction of mutations in the genomic DNA of hemopoietic stem cells and for the identification of novel and safer inducers of HbF.

Materials and methods Introduction of the 4-bp deletion mutation into the 183-kb b-globin locus transgene The two-step GET Recombination system in Escherichia coli strain DH10B in combination with the EcoRI/KanR counterselection cassette [24] was used to introduce the codons 41–42 4-bp deletion ( TTCT) mutation into pEBAC148/h, a 205-kb BAC carrying the entire h-globin locus [35]. In the first stage of GET Recombination, the EcoRI/KanR cassette was inserted into intron I of the hglobin gene. This was subsequently knocked out in a second round of GET Recombination by a PCR fragment from patient DNA carrying the codons 41–42 4-bp deletion mutation. Generation of transgenic mice BAC DNA was purified by cesium chloride gradient ultracentrifugation. The 183-kb genomic insert containing the human h-globin locus was released from the vector backbone by digestion with NotI and separated by pulsedfield gel electrophoresis (PFGE). The 183-kb fragment was

460

D. Jamsai et al. / Genomics 85 (2005) 453–461

recovered by digestion with h-agarase I (New England Biolabs, Beverly, MA, USA). The DNA was concentrated by microdialysis in microinjection buffer (10 mM Tris–HCl, pH 7.4, 0.2 mM EDTA, 100 mM NaCl) overnight at 48C using a 0.05-Am Millipore filter (Millipore Corp., Bedford, MA, USA). DNA quality and quantity were carefully estimated by running a small amount of the purified insert by PFGE with E HindIII as a standard. The final concentration was adjusted to 0.4 ng/Al and microinjected into fertilized mouse oocytes (C57BL/6). Transgenic founders were identified by PCR of mouse tail DNA using primers LUG1A, 5V-ACAAGACAGGTTTAAGGAGACCA-3V, and LUG2A, 5V-GTCTGTTTCCCATTCTAAACTGTA-3V. These primers specifically amplify a 447-bp product from the human h-globin gene. Transgenic mice carrying the 4-bp deletion transgene (huhD4bp/0, muh+/+) were bred with heterozygous h-globin knockout mice (muhth-3/+) [15] to generate mice carrying the transgene on a heterozygous knockout background (huhD4bp/0, muhth-3/+). These mice are referred to as double heterozygous mice (denoted as DHD4bp mice). Crosses of DHD4bp  DHD4bp mice were used to generate embryos hemizygous or homozygous for the transgene on a homozygous knockout background (huhD4bp/0, muhth-3/th-3 or huhD4bp/D4bp, muhth-3/th-3). Identification of such embryos was performed by multiplex PCR using three primer pairs: HPRT-Fw, 5V-GATGGGAGGCCATCACATTGTAG-3V, HPRT-Rev, 5V-GCGACCTTGACCATCTTTGGATTA-3V (315-bp product); mouse-h Fw 5V-TGAGAAGGCTGCTGTCTCTTG-3V, mouse-h Rev, 5V-CAGAGGATAGGTCTCCAAAGCTA-3V(260-bp product); and the LUG1A/ LUG2A primer pair (447-bp PCR product). Reactions were performed in 25 Al for 30 cycles (948C for 30 s, 558C for 30 s, and 728C for 30 s). Transgene analysis The overall integrity of the human h-globin locus transgene was evaluated by PCR using primers as described elsewhere (Vadolas et al., manuscript in preparation). Fluorescence in situ hybridization analysis was similarly used to examine the number of integration sites on mouse chromosomes. The transgene copy number was determined by Southern blot hybridization using the human h-globin probe amplified by the LUG1A/LUG2A primers. Approximately 10-Ag genomic DNA samples from F1 generation transgenic mice and human genomic DNA (as a positive control) were digested with EcoRI and HpaI (Roche Diagnostic Systems, Indianapolis, IN, USA), separated on a 1% agarose gel for 10 h at 40 V, and transferred onto a Hybond-N+ membrane (Amersham, Arlington Heights, IL, USA) under alkaline conditions. A 32P-labeled h-globin probe was made using the RediPrime DNA labeling kit (Invitrogen, VIC, Australia). Hybridization was performed at 658C for 20 h. The high degree of homology exhibited by all the members of

the h-globin-like genes with respect to the region amplified by the LUG1A and LUG2A primers allows simultaneous detection of the 2.2-kb h-globin fragment and 2.0-kb of the y-globin gene. The transgene copy number was estimated by comparing the intensity of the human h-globin fragment from human genomic DNA (two copies/cell) to the intensity of the human h-globin fragment present in transgenic mice using Gene Tools (Syngene, UK). RT-PCR analysis Isolation of total RNA from blood and bone marrow samples was performed using the TRI-BD reagent (Molecular Research Center, OH, USA) and this was converted into cDNA using Superscript II reverse transcriptase (Gibco BRL, Grand Island, NY, USA). PCR primers G2-FW, 5V-ACCTGACTCCTGAGGAGAAG-3V, and G1-Rev, 5V-ATAACAGCATCAGGAGTGGAC-3V, were used to amplify specifically a 160-bp fragment of normal human h-globin mRNA (25-Al reaction volume for 30 cycles of 948C for 30 s, 558C for 30 s, and 728C for 30 s). Hematological analysis Freshly drawn EDTA-treated blood was collected by retro-orbital bleeding under anesthesia. Peripheral blood smears were stained with Wright’s stain (Sigma–Aldrich, NSW, Australia). Hematologic profiles were obtained with an automatic blood counter (Cobas Helios Hematogoly Analyzer, Roche Diagnostic Systems). For hemoglobin analysis, red cells were washed once in 20 vol of phosphate-buffered saline and lysed in 5 vol of cystaminecontaining buffer [36]. Samples containing approximately equal amounts of hemoglobin were then run on cellulose acetate membranes (Helena Laboratories, TX, USA). Hemoglobin bands were visualized by Ponceau S staining. Globin chain separation was performed by Triton X acid urea gel electrophoresis as previous described [37].

Acknowledgments This work was supported by grants to the CAGT Research Group from the Brockhoff Foundation and the National Health and Medical Research Council of Australia. Duangporn Jamsai was supported by a Royal Golden Jubilee Scholarship from the Thailand Research Fund and a grant from the Thalassemia Society of Victoria. Jim Vadolas is currently a recipient of a research fellowship from the Cooley’s Anemia Foundation, USA.

References [1] D.J. Weatherall, J.B. Clegg, Thalassemia—A global public health problem, Nat. Med. 2 (1996) 847 – 849.

D. Jamsai et al. / Genomics 85 (2005) 453–461 [2] D.J. Weatherall, Phenotype–genotype relationships in monogenic disease: lessons from the thalassaemias, Nat. Rev. Genet. 2 (2001) 245 – 255. [3] D.J. Weatherall, L. Pressley, W.G. Wood, D.R. Higgs, J.B. Clegg, Molecular basis for mild forms of homozygous beta-thalassaemia, Lancet 1 (1981) 527 – 529. [4] H.F. Bunn, B.G. Forget, Hemoglobin: Molecular, Genetic and Clinical Aspects, Saunders, Philadelphia, 1986. [5] N.F. Olivieri, The beta-thalassemias, N. Engl. J. Med. 341 (1999) 99 – 109. [6] S.M. Coles, H.G. Klein, P.V. Holland, Alloimmunization in two multitransfused patient populations, Transfusion 21 (1981) 462 – 466. [7] T. Spanos, et al., Red cell alloantibodies in patients with thalassemia, Vox Sang. 58 (1990) 50 – 55. [8] G. Lucarelli, C. Giardini, D. Baronciani, Bone marrow transplantation in thalassemia, Semin. Hematol. 32 (1995) 297 – 303. [9] R.C. Hardison, et al., HbVar: a relational database of human hemoglobin variants and thalassemia mutations at the globin gene server, Hum. Mutat. 19 (2002) 225 – 233. [10] V. Chan, T.K. Chan, F.F. Chebab, D. Todd, Distribution of betathalassemia mutations in south China and their association with haplotypes, Am. J. Hum. Genet. 41 (1987) 678 – 685. [11] S. Fucharoen, et al., Molecular basis of beta-thalassemia in Thailand: analysis of beta-thalassemia mutations using the polymerase chain reaction, Hum. Genet. 84 (1989) 41 – 46. [12] L.C. Skow, et al., A mouse model for beta-thalassemia, Cell 34 (1983) 1043 – 1052. [13] W.R. Shehee, P. Oliver, O. Smithies, Lethal thalassemia after insertional disruption of the mouse major adult beta-globin gene, Proc. Natl. Acad. Sci. USA 90 (1993) 3177 – 3181. [14] D.J. Ciavatta, T.M. Ryan, S.C. Farmer, T.M. Townes, Mouse model of human beta zero thalassemia: targeted deletion of the mouse beta majand beta min-globin genes in embryonic stem cells, Proc. Natl. Acad. Sci. USA 92 (1995) 9259 – 9263. [15] B. Yang, et al., A mouse model for beta 0-thalassemia, Proc. Natl. Acad. Sci. USA 92 (1995) 11608 – 11612. [16] C. May, S. Rivella, A. Chadburn, M. Sadelain, Successful treatment of murine beta-thalassemia intermedia by transfer of the human betaglobin gene, Blood 99 (2002) 1902 – 1908. [17] S. Rivella, C. May, A. Chadburn, I. Riviere, M. Sadelain, A novel murine model of Cooley anemia and its rescue by lentiviral-mediated human beta-globin gene transfer, Blood 101 (2003) 2932 – 2939. [18] J. Lewis, et al., A common human beta globin splicing mutation modeled in mice, Blood 91 (1998) 2152 – 2156. [19] W. de Laat, F. Grosveld, Spatial organization of gene expression: the active chromatin hub, Chromosome Res. 11 (2003) 447 – 459. [20] P.A. Ioannou, et al., A new bacteriophage P1-derived vector for the propagation of large human DNA fragments, Nat. Genet. 6 (1994) 84 – 89. [21] M. Orford, et al., Engineering EGFP reporter constructs into a 200 kb human beta-globin BAC clone using GET Recombination, Nucleic Acids Res. 28 (2000) E84.

461

[22] J. Vadolas, et al., Development of sensitive fluorescent assays for embryonic and fetal hemoglobin inducers using the human betaglobin locus in erythropoietic cells, Blood 100 (2002) 4209 – 4216. [23] J. Vadolas, H. Wardan, M. Orford, R. Williamson, P.A. Ioannou, Cellular genomic reporter assays for screening and evaluation of inducers of fetal hemoglobin, Hum. Mol. Genet. 13 (2004) 223 – 233. [24] D. Jamsai, et al., Insertion of common mutations into the human betaglobin locus using GET Recombination and an EcoRI endonuclease counterselection cassette, J. Biotechnol. 101 (2003) 1 – 9. [25] M. Nefedov, R. Williamson, P.A. Ioannou, Insertion of diseasecausing mutations in BACs by homologous recombination in Escherichia coli, Nucleic Acids Res. 28 (2000) E79. [26] D. Jamsai, et al., Targeted modification of a human beta-globin locus BAC clone using GET Recombination and an I-SceI counterselection cassette, Genomics 82 (2003) 68 – 77. [27] D. Jamsai, M. Orford, S. Fucharoen, R. Williamson, P.A. Ioannou, Insertion of modifications in the beta-globin locus using GET recombination with single-stranded oligonucleotides and denatured PCR fragments, Mol. Biotechnol. 23 (2003) 29 – 36. [28] K.M. Gaensler, M. Kitamura, Y.W. Kan, Germ-line transmission and developmental regulation of a 150-kb yeast artificial chromosome containing the human beta-globin locus in transgenic mice, Proc. Natl. Acad. Sci. USA 90 (1993) 11381 – 11385. [29] S. Porcu, et al., The human beta globin locus introduced by YAC transfer exhibits a specific and reproducible pattern of developmental regulation in transgenic mice, Blood 90 (1997) 4602 – 4609. [30] E. Gallo, et al., The importance of the genetic picture and globin synthesis in determining the clinical and haematological features of thalassaemia intermedia, Br. J. Haematol. 41 (1979) 211 – 221. [31] P.J. Ho, et al., Beta-thalassaemia intermedia: is it possible consistently to predict phenotype from genotype? Br. J. Haematol. 100 (1998) 70 – 78. [32] I. Lampronti, et al., Accumulation of gamma-globin mRNA in human erythroid cells treated with angelicin, Eur. J. Haematol. 71 (2003) 189 – 195. [33] E. Fibach, N. Bianchi, M. Borgatti, E. Prus, R. Gambari, Mithramycin induces fetal hemoglobin production in normal and thalassemic human erythroid precursor cells, Blood 102 (2003) 1276 – 1281. [34] B.S. Pace, et al., Short-chain fatty acid derivatives induce fetal globin expression and erythropoiesis in vivo, Blood 100 (2002) 4640 – 4648. [35] K. Narayanan, R. Williamson, Y. Zhang, A.F. Stewart, P.A. Ioannou, Efficient and precise engineering of a 200 kb beta-globin human/bacterial artificial chromosome in E. coli DH10B using an inducible homologous recombination system, Gene Ther. 6 (1999) 442 – 447. [36] J.B. Whitney, Simplified typing of mouse hemoglobin (Hbb) phenotypes using cystamine, Biochem. Genet. 16 (1978) 667 – 672. [37] B.P. Alter, S.C. Goff, G.D. Efremov, M.E. Gravely, T.H. Huisman, Globin chain electrophoresis: a new approach to the determination of the G gamma/A gamma ratio in fetal haemoglobin and to studies of globin synthesis, Br. J. Haematol. 44 (1980) 527 – 534.