A mouse model for β-thalassemia

Cell, Vol. 34, 1043-1052, October 1963, Copyright (D 1963 by MIT

A Mouse Model for ,&Thalassemia

L. C. Skew,* B. A. Burkhart,” F. M. Johnson,* Ft. A. Popp,+ D. M. Popp,+ S. Z. Goldberg,* W. F. Anderson,* L. B. Barnet&* and S. E. Lewis* *Laboratory of Genetics National Institute of Environmental Health Sciences Research Triangle Park, North Carolina 27709 ‘Biology Division Oak Ridge National Laboratory Oak Ridge, Tennessee 37830 *Laboratory of Molecular Hematology National Heart, Lung, and Blood Institute Bethesda, Maryland 20205 *Chemistry and Life Sciences Group Research Triangle Institute Research Triangle Park, North Carolina 27709

Summary

A mutation that produces an absolute deficiency of normal @major globin polypeptides has been re- covered from a DBA/2J male mouse. Most mice homozygous for the deficiency survived to adult- hood and reproduced but were smaller at birth than their littermates and demonstrated a hypochromic, microcytic anemia with severe anisocytosis, poikil- ocytosis, and reticulocytosis and the presence of inclusion bodies in a high proportion of circulating erythrocytes. Mice heterozygous for the deficiency demonstrated a mild reticulocytosis but were not clinically anemic. Analysis of globin chain synthesis in vitro by 3H-leucine incorporation revealed that B- globin synthesis was nearly normal (95%) in heter- ozygotes and about 75% of normal in deficiency homozygotes. Molecular characterization of the mu- tation by restriction analysis revealed a deletion of about 3.3 kb of DNA, including regulatory sequences and all coding blocks for B-major globin. Based on genetic and hematological criteria, mice homozy- gous for the mutant allele, designated /fbblh-‘, rep resent the first animal model of B-thalassemia (Cool- ey’s anemia), a severe genetic disease of humans.

Genes for mouse CY- and fi-globins are encoded in com- plexes (or haplotypes) of a-like and p-like loci on chromo- somes 11 (Russell and McFarland, 1974) and 7 (Hutton, 1969; Popp, 1969) respectively. Two @-globin genes are normally expressed in adult mice and are located toward the 3’ end of the @globin gene complex (Jahn et al., 1980; Edgell et al., 1981). These genes occur in polymorphic forms (Ranney et al., 1960; Russell and McFarland, 1974) that encode four types of B-globin polypeptides (Popp, 1973; Popp and Bailiff, 1973; Gilman, 1976). Mice of the Hbb” haplotype possess two genes that make an identical @globin subunit, /3single. Mice of the Hbbd haplotype

0092-8674/83/l 001043-10 !$0200/0

produce two distinct globins, p-major and o-minor, while mice of the HbbP haplotype make b-major and a variant form of p-minor. Since the electrophoretic variants com- monly observed in mouse hemoglobin are predominantly determined by differences in the 0-globin subunits, electro- phoretic forms of mouse tetrameric hemoglobin ((Y&) are designated by /3-globin type in this report. The organization of the mouse @globin gene complex is similar to that of humans (Lawn et al., 1978; Fritsch et al., 1979) goats (Haynes et al., 1980) sheep (Kretschmer et al., 1981) and rabbits (Lacy et al., 1979) and appears to be representative of mammals in general.

The extensive knowledge of mammalian globin gene structure and function makes it possible to readily evaluate mutations in globin genes of laboratory animals as models of human genetic diseases. A particularly significant group of hemoglobin disorders in humans is the B-thalassemias, a collection of genetically diverse hemoglobinopathies that are characterized by reduced synthesis of normal @globin subunits (Watherall and Clegg, 1972; Nathan and Gunn, 1966). In severe /?-thalassemias, the reduced amount of normal P-globin is associated with microcytic anemia, re- duced intracellular hemoglobin (hypochromia), and eryth- rocytic inclusion bodies composed of the insoluble chains of excess ol-globin (Benz and Forget, 1975). Also diagnos- tic of classical P-thalassemia is an elevated amount of HbA2, normally a very minor component of adult human hemoglobin. The hematologic syndromes observed in p- thalassemic individuals are variable in severity, ranging from asymptomatic hypochromia in heterozygotes to pro- foundly ineffective erythropoiesis, hemolytic anemia, and bone marrow hyperplasia resulting in skeletal deformities and retarded growth in homozygotes (/3-thalassemia major or Colley’s anemia). To date, P-thalassemia has been observed only in humans.

In this report we describe a deletion mutation in mice that results in a deficiency of P-globin polypeptides and produces associated hematologic abnormalities very sim- ilar to those observed in human /I-thalassemia.

Results

Inheritance The mutation was discovered during electrophoretic screening (Johnson and Lewis, 1981) for induced muta- tions in Fl progeny from matings of C57BL/6J female and DBA/2J male mice treated with ethylnitrosourea. Blood samples from 11 of 28 offspring in four litters demonstrated aberrant hemoglobin patterns evidenced by increased amounts of p-minor hemoglobin. Examination of breeding records revealed that all variant mice were sired by a single DBA/2J male. Subsequent electrophoretic analysis of he- moglobin from the DBA/2J sire also disclosed an elevated p-minor hemoglobin component, which established that the mutation was preexisting and not induced by the ENU treatment.

Figure 1 is a photograph of a portion of the original

Cd 1044

Origin 0 I234567

d major

/ “2 s”, + 42 62

Figure 1. Photograph of the Original Starch Gel Containing the First Two of Eleven ,%Thatassemic Fi Heterozygotes

Mutant heterozygotes (samples 3 and 6) were detected by electrophoresis to have an elevated p-minor hemoglobin component,

starch gel electropherogram showing hemoglobin samples from typical (C57BL/6J X DBA/2J)Fl animals and the first two variants identified from the progenitor. Electrophoresis of cystamine-modified hemoglobin on cellulose acetate plates gave better resolution of the three types of mouse hemoglobin (Figure 2) where bands corresponding to p- single, P-major, and B-minor hemoglobins were clearly resolved. Samples from normal (C57BL/6J X DBA/2J)Fl heterozygotes (Hbbs/Hbbd) (Figure 2, sample 5) showed three hemoglobin bands but samples from variant Fl animals (sample 2) lacked a detectable o-major compo- nent Hemoglobin from the progenitor (sample 3) con- tained relatively less p-major and more b-minor than sam- ples from control DBA/2J animals (sample 4). Only the p- minor hemoglobin was observed in samples from animals homozygous for the mutant Hbbd haplotype (sample 1). The mutation was judged to affect only the &major globin since the P-minor component of hemoglobin from variant homozygotes and heterozygotes was electrophoretically identical to the p-minor component of normal mice. Amino acid analysis of the tryptic peptides of B-minor globin produced in mutant homozygotes confirmed that it did not differ from that of the DBA/2J control (data not given). The simplest interpretation of these patterns is that the progen- itor and his 11 variant Fl progeny were heterozygous for a mutant allele of Hbbd that results in the absence of functional p-major globin. The mutant haplotype is here- after designated Hbbth-’ for &thalassemia.

Measurements of the hemoglobin fractions by scanning densitometry are presented in Table 1. These data confirm that B-major hemoglobin was missing in Hbbth-‘/Hbbth-’ homozygotes and HbbS/Hbbth-l heterozygotes and re- vealed that the proportion of p-minor hemoglobin in the blood of heterozygous mice was elevated to 24.6% f 1.9%, which is approximately 2.5 times the level of B-minor hemoglobin in normal Fl mice (9.7% f 1.9%). Analysis of the hemoglobin from the progenitor demonstrated that p-

+ T

Origin 0

dmajor -Ot? 82

Q 2 8: minor

123456 Figure 2. Cellulose Acetate Etectmphorettt Patterns of Hemoglobin

(1) HbbW’/Hbbh’ hcmozygous mutant; (2) Hbbm’/HW heterozygous mutant; (3) DBA/2J mate proposttus; (4) Hbbd/Hbbd normal DBA/2J, (5) Hbb’/Hbbd normal heterozygote; and (6) Hbb*/Hbb* normal C57BL/&t. Hemolysates from thatassemia homozygotes (1) and Hbb*/Hbb*’ hetero- zygotes (2) contained no detectable &ma@ hemoglobin. A sample from the original DBAPJ mutant (Hbbd/Hbb”‘) heterozygote (3) contained less @-major and more Bminor hemoglobin than control samples (4).

Table 1. Scanning Densitometry of Hemoglobin Types of Thatassemic and Nonal Littermates Resolved by Cellulose Acetate Electrophoresis and Expressed as 9% + S.E.M. of Total Hemoglobin

Genotypes of Mice

Hbb*/Hbb”

Hbbd/Hbbd

Hbb’/Hbbd

Hbbd/Hbbm’

Hbbs/Hbbh’

Hbbm’/Hbbn-’

Hemoglobin Components

No. &Single @-Major rW&ror

6 100 0 0

12 0 81.9& 1.5 17.6 + 2.1

9 53.8 + 1.9 36.7 f 0.3 9.7 f 1.9

10 0 55 45

19 75.3 f 2.1 0 24.6 + 1.9

9 0 0 100

’ Sample from DBA/PJ progenitor.

major hemoglobin was reduced from 81.9% f 1.5% in DBA/2J controls to about 55%.

Data defining the inheritance of Hbbm-’ are presented in Table 2A. In backcross and intercross litters, the mutant phenotype was inherited as a simple Mendelian trait. Chi-

A Mouse Model for PThafassemia 1045

Table 2A. lnherkance of the HbbL’ Allele in Backcross and Intercross Matings of (C57BL/6J x DBA/2J)Fl Animals Heterozygous for the Mutation

Frequencies of Genotypes at Birth

Matings Hbb”‘/Hbb”’ Hbb’/Hbb- Hbb’/Hbb’

Backcross C57BL/6J x 0 20 (0) 16 (‘3) (C57B46J x DBA/ 2J)Fl

Intercross (C57B46J x 27 (10) 48 (0) 28 (0) DBA/2J)Fl x Fl

The numbers of mice lost from birth to weaning are in parentheses.

Table 28. Linkage Relationships Expressed as Percentage Recombination among Gpi-I, Tam-l, and Hbb in Mouse Chromosome 7

Regron of Recombination

None (C57BL/6J parental type) (DBA/PJ parental type)

Single recombinants Gpi-I x Tam-l Tam-l x Hbb

Double recombinants Gpi-1 x Tam-l x Hbb

Frequency of Recombinant Gametes

35% (18152) 39% (20/52)

11% (6/52) 14% (7/52)

2% (l/52)

The gene order is centromere-Gpi-I-Tam-I-Hbb (Skow, 1976).

square analysis of the data demonstrates that the fre- quency of each phenotypic class at birth was not signifi- cantly different from the expected I:1 or 1:2:1 ratios for backcross and intercross generations, respectively, as- suming a two-allele system without dominance. However, ten of 27 (37%) Hbb”-‘/Hbbth-’ mice died between birth and weaning. Survival of mutant homozygotes after wean- ing (21 days) appears to be normal to at least five months of age. Linkage analysis confirmed that the mutation is located on chromosome 7 in the region of the fl-globin complex (Table 28) about 15 map units (8/52 recombi- nants) distal from Tam-l. Furthermore, recombination fre- quencies along the length of the chromosome are similar to previously reported values (Skow, 1978) indicating that chromosomal alterations such as inversions or transloca- tions probably are not associated with Hbb’“.‘.

The viability of the adult P-thalassemia homozygotes, as evidenced by normal survival after weaning, led us to attempt reciprocal matings of homozygotes with hetero- zygotes to determine whether homozygous mutants were fertile and when they might reach sexual maturity. Thirty matings were made and produced litters without evidence of delayed sexual maturity. However, the mean litter size for thalassemic females was somewhat lower (4.3 f 1.3 pups/litter) compared to heterozygous females (6.0 + 1.8 pups/litter) and normal females (6.2 f 2.0 pups/litter).

Hematology Figure 3 shows representative photographs of Wright’s stained blood films. Compared to controls (Figure 3A),

blood films from Hbbs/Hbbth” heterozygotes (Figure 38) revealed normal erythrocyte morphology, occasional dif- fusely basophilic erythrocytes, and the rare occurrence of normocytes with condensed nuclei. In contrast, blood films from Hbb*.‘/Hbb*-’ homozygotes (Figure 3C) demon- strated red blood cells of variable size (anisocytosis) and shape (poikilocytosis) with large numbers of microcytes, and the presence of cellular debris (erythrocyte dust). Also observed were large numbers of diffusely basophilic eryth- rocytes and normocytes and occasional very immature red cells, i.e., normoblasts and basophilic normoblasts. Inclusion bodies were observed in 20%-30% of circulating mature erythrocytes.

Additional data on the clinical hematology of mutant and normal littermate mice are shown in Table 3. The deficiency homozygotes exhibited significantly reduced hematocrits (37.4 + 0.5) and red blood cell counts (10.2 f 0.3 x IO61 mm3) when compared to control littermates. In mutant homozygotes, nucleated cell counts were very high and variable (25,870 -I- 5798/mm3) reflecting the presence of immature nucleated blood cells shown in blood films. Mutant homozygotes also exhibited reduced hemoglobin (10.8 + 0.3 g/dl), mean corpuscular hemoglobin, MCH (10.6 + 0.3 pg/rbc), and mean corpuscular volume, MCV (36.0 + 0.9 r3). The hematologic values for mutant heter- ozygotes were not significantly different from control val- ues. Peripheral blood stained with new methylene blue revealed that deficiency homozygotes had reticulocytosis (41.4% + 2.1%) compared to normal littermates (4.6% f 0.5%). Mutant heterozygotes had mildly elevated reticulo- cyte counts (6.1% + 0.4%).

Globin Synthesis Measurements of the relative rates of synthesis of (Y- and P-globin by 3H-leucine incorporation into reticulocytes re- vealed that cells from thalassemic heterozygotes (Hbb’/ Hbbth-‘) or homozygotes (Hbbth-‘/Hbbth-I) achieved sub- stantial modulation of gene expression to accommodate the genetic imbalance as evidenced by the j3/Bla-globin synthesis ratios. Pooled reticulocytes from control mice (DBA/2J) synthesized globins in a /~/LX ratio of 1.06. Retic- ulocytes from two thalassemic heterozygotes had j3/a synthesis ratios of 0.92 and 0.98 and pooled reticulocytes from two thalassemic homozygotes had /3/a synthesis ratios of 0.78. These data do not indicate whether the deficiency in p-major globin genes is moderated by in- creased transcription from the B-minor gene or by de- creased synthesis (or increased turnover) from the CX- genes. We have not exluded the possibility that proteolysis of the excess cY-chains contributes to the apparent in- crease in P-globin synthesis.

Restriction Analysis To investigate the nature of the Hbb*-’ mutation, we performed restriction mapping analysis of genomic DNA, using several probes specific for regions of the B-major gene and flanking sequences (Figure 4). Probe 48, a /3-

Cell 1046

Figure 3. Comparison of Air-dried Blood Films from F2 LIttermate Mice Stained with Wright’s Stain

(A) is from a normal mouse (Hbb*/Hbb’); (8) IS from a @-thalassemlc heterozygote (/-/bb*/Hbbh-‘) and (C) is from a @-thalassemic homozygote (/-/bbm’/Hbbh’). Note the severe hypochromia, anisocytosis, and poikilo- cflosis in the Hbbm’/Hbbh-’ homozygote.

major globin cDNA from BALB/c mice cloned in X phage (Blattner et al., 1977) was used to determine whether the transcriptional units were intact. Hybridization to endonu- clease-digested DNA from mice of strains DBA/ZJ and

BALB/cJ produced restriction patterns identical to those reported for BALB/cJ (Jahn et al., 1980; Edgell et al., 1981); however, no B-major sequences were observed to hvhririix with DNA with the /-/~/II’“-l/H&‘“-’ thalassemic ,, _. .____ _

A Mouse Model for o-Thalassemia 1047

Table 3. Hematologic Values of Blood from &Thafassemic and Normal Littermates

Hematologic Parameters

Mice Hct

No. Genotype w

14 Hbb’/Hbb’ 47.4 f 0.7’ (42-511)~

29 Hbb*/Hbb*- 46.8 + 0.6 (41-54)

16 Hbbm’/Hbbm” 37.4 f 0.5 (35-41)

a Standard error of the mean. b Range of values.

RBC (x 10B/mm3)

11.8 * 0.4 (9.3-14.1)

11.6 + 0.2 (9.1-13.9)

10.2 * 0.3 (8.5-12.3)

Nucleated Cells

(mm7

5370 + 541 (m-8680)

6230 + 499 (2200-12980

25870 + 5798 (9300-i 08rn)

Hb MCH

(g/d1 ) (w/rW

14.7 * 0.2 12.5 + 0.4 (13.5-15.7) (10.3-15.1)

14.1 * 0.2 12.3 + 0.2 (12.3-15.7) (9.9-14.9)

10.8 + 0.3 10.6 + 0.3 (8.9-12.6) (7.5-12.6)

MCV

(2)

40.5 f 1.1 (33.8-46.2)

40.7 + 0.7 (34.4-46.9)

36.0 + 0.9 (29.3-42.9)

Reties

(“4

4.6 2 0.5 (1.5-7.9)

6.1 k 0.4 (3.3-10.9)

41.4 + 2.1 (26.6-62.2)

EcoRI-Avo I I I

Figure 4. Mapping of Restriction Endonuclease Cleavage Sites in the B-Major Globin Region of DNA from Homozygous Thalassemic Mice and Control DBA/ 2J Mice

Organization of the mouse @globin complex, as described by Edgell et al. (1981). is presented at the top of the diagram. The locations of cleavage sites for Ava I, Barn HI, Bgl II, Eco RI, Hind III, and Xba I are shown for the B-major structural and flanking sequences. The deletion in thafassemia DNA is approximated by the single line and includes the three coding blocks; intact flanking sequences are represented by double lines. The different probes utilized in this study and the sequences that they detect are represented in the lower portion of the diagram.

mice. Similarly, we found no b-major sequences in thalas- semic DNA using pPK288, a probe for the first and second coding blocks and about 1 .O kb of 5’ flanking sequences. These data indicate that a deletion mutation extends from at least the 5’ Eco RI site 3’ through the coding blocks of ,&major, a distance of at least 3 kb. In both experiments, faint hybridization was observed to fragments containing the related p-minor gene and pseudogene /3h3, but no restriction differences in these sequences were observed between thalassemic and control DNA.

Sequences 3’ from the p-major structural gene were investigated using probes prepared from the plasmid pPK268, which is composed of the 7.5 kb Eco RI fragment containing the B-major globin gene inserted into the Eco RI site of pBR322 (P. Kretschmer, unpublished data). The

presence of a repetitive sequence near the 3’ end of the fragment made it necessary to remove the 3’ terminal sequences of the 7.5 kb mouse fragment in order to produce specific probes for &major flanking regions. Two probes, an Eco RI-Xba I fragment and an Eco RI-Ava I fragment, were prepared from double-digested pPK268 (see Figure 4) and detected identical restriction fragments when hybridized to Hbb’“~‘/Hbb’h~’ DNA. Only faint hybrid- ization was obtained with the Eco RI-Xba I probe, but using the Eco RI-Ava I probe we readily identified restric- tion fragments from Hbbth~‘/Hbbth-’ DNA that were larger (Barn HI, Eco RI, Bgl I), smaller (Hind Ill, Bgl II), or the same size (Xba I, Bgl II) as those seen in control DNA (Table 4 and Figure 5).

By utilizing the aberrant size fragments observed in

Cdl 1048

Table 4. Sizes (in kilobase pairs) of Normal and Aberrant Restriction Fragments Cleaved from DNA of DeA/2J (Hbbd/Hbbd) and Thslassemic (/-/bb”F’/Hbbn”) Mice

Source of DNA

Restriction Enzyme DBA/2J Thalsssemic

Bam HI

8911

Bgl II

Eco RI Hind Ill

Xba I

7.1 4.5 7.7 5.2 5.3 2.7 7.5 0.7 1.8 1.3 9.5 5.5

a.3

9.7

2.7 2.0 8.6 7.6

5.5

Fragments were detected by a probe specific for the @-major globin coding and flanking sequences extending from the 5’ Eco RI site to the 3’ Ava I site (see Figure 4).

Hbb”-‘/Hbbth-’ DNA, we have concluded that the deletion is 3.1-3.3 kb long, and spans the 5’ Eco RI site, all elements of the structural gene and 3’ flanking sequences proximate to but not including the Xba I site, about 400 bp downstream of the polyadenylated addition signal. Our reasoning is as follows (see also Figure 4): the internal Bam HI and Bgl I sites as well as the 5’ Eco RI site have been removed to produce novel restriction fragments larger than observed in control DNA. In both Barn HI and Bgl I digests, the size of respective aberrant fragments (8.3 and 9.7 kb) is about 3.3 kb shorter than the sum of the lengths of the two fragments seen in control DNA (11.6 for Barn HI and 12.9 kb for Bgl I). A single Eco RI fragment about 1 kb larger than the control fragment is the size expected for a new fragment generated by a 3.3 kb deletion spanning the 5’ Eco RI site with the new fragment arising from Eco RI sites 3’ from the pseudogene j3h3 and P-minor (see Edgell et al., 1981).

The aberrant fragments observed in Hind III (7.8 kb) and Bgl II (2.0 kb) digests are also consistent with a 3.3 kb deletion. In Hbbth-‘/Hbbth-’ DNA, the absence of the two small Hind Ill (1.8 and 1.3 kb) fragments and reduction in size of the normal 8.6 kb fragment indicates that the deletion includes the two Hind Ill sites within the Eco RI fragment and extends some distance 3’. The termini for the new Hind Ill fragment are then outside the reference Eco RI sites. Our data indicate that the 5’ Hind III site in control DNA is about 1 kb from the deleted Eco RI site. Therefore the size of this new Hind Ill fragment minus the deletion would be about 7.8 kb, consistent with our exper- imental observations. Similarly the normal 5.3 kb Bgl II fragment has been shortened by about 3.3 kb in Hbb’“‘/ Hbb’“” DNA. From our control data the 5’ Bgl II site can be mapped about 1.5 kb 5’ from the deleted Eco RI site. Based on the Hind Ill and Bgl II studies, the 5’ breakpoint of the deletion must fall betweeen the Hind Ill site and the 5’ Eco RI site.

The presence of normal sized 3’ fragments in Xba I and Bgl II digests (Figure 5) showed that the 3’ deletion breakpoint must fall between the Xba I site and the third coding block, since the structural gene sequences were missing in Hbb”.‘/Hbb’‘-’ DNA. No aberrant restriction fragments have been detected in Xba I digests of mutant DNA, suggesting that the 3’ breakpoint must be very near the Xba I site. Otherwise, a 5.8 kb fragment produced from the ph3 and ,&major Xba I sites would be expected. This interpretation assumes that the sequence remaining 5’ from the Xba I site is so small as to preclude effective hybridization under the conditions employed.

Discussion

The classical human fi-thalassemia, also called thalassemia major or Cooley’s anemia, is a homozygous genetic dis- order characterized by the absence or marked reduction of P-globin polypeptides. The pathophysiology of human @- thalassemia has been well defined (see review by Benz and Forget, 1980) and results from the accumulation, within erythroblasts, of excess cY-globin polypeptides that aggregate and precipitate within the developing red cells.

The results of the experiments reported here on the Hbbth-’ homozygote and heterozygote mice describe sev- eral features that are all comparable to the hematologic syndromes associated with human @thalassemias. In homozygous mice, these include the presence of a severe hypochromic anemia with reticulocytosis, anisocytosis, an elevated adult p-minor hemoglobin fraction that is structur- ally normal and large numbers of circulating red blood cells with inclusion aggregates. Heterozygous mice demon- strate only a mild reticulocytosis with elevation of the @- minor hemoglobin.

In untreated cases, human @-thalassemia is a lethal disease, usually resulting in death in early childhood. The reduced severity of /3-thalassemia in mice may be related to the proportionally greater contribution of mouse @-minor globin to functional hemoglobin as compared to the 6- globin chain of humans. In human /3-thalassemics, HbA2 ((~~6~) levels are elevated and variable but range from normal levels of about 2% to 6% of total hemoglobin (Aksoy and Erdem, 1969). In mice, B-minor is normally expressed at about 20% (in Hbbd/Hbbd mice) of total hemoglobin but is increased approximately 2.5 times in thalassemia heterozygotes, reflecting the apparent in- creased expression of the /I-minor gene to produce an almost normal ratio of p/cr globin synthesis. The in vitro synthesis experiments reported here suggest that Hbb”‘-‘/ Hbb’“’ mice may achieve an even greater expression of p-minor, at 75% of normal @-chain synthesis in fi-thalas- semic homozygotes. Data on fl/~lcu synthesis ratios are proportional and do not necessarily indicate an absolute increase in p-minor globin synthesis, but only an increase relative to a-chain synthesis. However, data on mean corpuscular hemoglobin (Table 3) reveal that B-minor he- moglobin synthesis in /?-thalassemic homozygotes is in-

A Mouse Model for &Thalassemia 1049

HindIII XbaI Hind III Eco RI Bal II Bal I BarnHI EcoRI -- x12222212- 2 268

23 5

9.7

6.6

4.3

2.2

2.1

Figure 5. Restriction Endonuclease Analysis of Mouse Genomic DNA and Identification of Fragments Containing B-Major and B-Minor Globin Sequences

DNA from homozygous thatassemic (1) and control DBA/2J (2) mice were digested with various endonucleases and the cleavage products were electrophoresed in a 1% agarose gel. Fragments from Hind Ill-digested X DNA, end-labeled wtth 9, were coelectrophoresed as size standards. Identification of @-globin related sequences was made after Southern transfer of DNA fragments to Genetran TM filters and hybridization to a “P-labeled Eco RI-Ava I probe (1rY’ cpm/pg) purified from plasmid pPK268 as described in Experimental Procedures. The sizes of B-major sequences in kb are indicated below the fragments. The faintly hybridizing fragments (no sizes given) are presumed to be @minor sequences based on data from Edgell et al. (1991) and do not differ between samples 1 and 2.

creased to 10.6 pglerythrocyte compared with 12.5 pg/ expression during development that p-minor may represent erythrocyte in DBA/2J mice (unpublished data). Fetal he- the product of a fetal-like P-globin gene. moglobin, HbF ((Y&, is typically observed in the blood of The reticulocytosis and presence of erythrocytic inclu- human thalassemics and is important in moderating the sion bodies in the blood of @thalassemic mice is in marked severity of the disease (Ley et al., 1982). Mice apparently contrast to the hematology of human @-thalassemics with do not express a fetal globin gene, although there is limited intact spleens; rather it resembles the hematology of sple- evidence (Whitney, 1977) based on changes in levels of nectomized human thalassemics. This difference between

Cell 1050

the human and mouse B-thalassemias might be attributable to differences in splenic function between the two species. In adult humans, splenic function is devoted entirely to the culling and pitting of senescent and deformed red cells whereas the spleen of the mouse is a major site of hematopoeisis, beginning late in gestation and continuing throughout life (Russell, 1979). Consequently, the mouse spleen may be less efficient in culling and pitting immature or deformed red cells.

The genetic lesions responsible for human @-thalasse- mias are diverse (see review by Weatherall and Clegg, 1982) and include small deletions and point mutations (Kan et al., 1975, 1977; Ramirez et al., 1976; Old et al., 1978; Benz et al., 1978; Baird et al., 1981; Orkin et al., 1982; Treisman et al., 1983) which are usually in regulatory and intervening sequences but only rarely involve coding sequences (Orkin et al., 1979). Therefore the deletion of all coding sequences and introns as well as 3’ and 5’ flanking sequences of the p-major globin gene in the /3- thalassemia mouse reported here differs somewhat from the types of mutations commonly reported for human thalassemias. The molecular basis for naturally occurring deletions is poorly understood. However, since the muta- tion apparently arose via natural processes, it is possible that an unequal crossing-over event produced the deletion of the /3-globin gene. Recent evidence (Jagadeeswaran et al., 1982; Ottolenghi and Gigliori, 1982) on a series of deletions responsible for human d-@thalassemia and he- reditary persistence of fetal hemoglobin (HPFH) revealed that the deletion breakpoints occurred in the regions of the Alu repetitive sequences, consistent with an unequal re- combination process being the mutagenic event. Six re- petitive sequences occur in and around the /3-globin gene complex in mice (Haigwood et al., 1981) but our data indicate that the deletion breakpoints in Hbbth-’ are not near these repetitive elements. Therefore it seems unlikely that unequal recombination via repetitive sequences was the mechanism that gave rise to the deletion in Hbbth.‘.

The discovery of a mouse model for human ,B-thalasse- mia is timely for the development of in vivo gene therapy. Several strategies have been proposed for incorporating normal genes into mutant genomes (Mercola and Cline, 1980) but many questions remain concerning proper reg- ulation and expression of exogenous genes introduced into animals (e.g., the importance of random vs. site- specific incorporation for correct gene function) (Anderson and Fletcher, 1980; Cline, 1982). It is our hope that the availability of the P-thalassemia mouse to the scientific community will facilitate the rapid resolution of these ques- tions

Experimental Procedures

Source of Animals Mice of the inbred strains C57BL/6J and DBAPJ were purchased from the Jackson Laboratory. Bar Habor, ME. As part of an ongoing experiment in mutagenesis, males of strain DBApJ were injected intaperitoneally with ethylnitrosourea (ENU) at 200 mg/kg-’ and mated to C57B46J females to produce (B6D2) Fl hybrid progeny, which were examined by electropho-

rests for Induced mutations (Johnson and Lewis, 1981). The mutatron described in this report arose in a DBA/2J male and was transmitted to 11 of his (B6D2)Fl progeny. These mice were the progenitors of the j3- thalassemic mice utilized in this study.

Genetic Studies Tested B6D2Fl carriers of the mutation were backcrossed to normal B6 mice to maintain the mutation and provide material for lknkage analysis. Lrnkage of the mutant phenotype with the @-globin complex on chromo- some 7 was established through a three-point linkage test using the gene markers hemoglobin @-chain (HBB), glucose phosphate isomerase (GPf-1, and tosylargrnine methyl esterase (TAM-l) as previously described (Skow, 1978). Tested carriers were also mated inter se to produce F2 litters of normal, heterozygous mutant and homozygous mutant mice for genetic, hematologc, and molecular analyses.

Collection and Hematologic Analysis of Blood Blood samples for the mutation screen were collected and analyzed by starch gel electrophoresis as described (Johnson and Lewis, 1981). Blood samples for cellulose acetate electrophoresis were collected into heparin- ized microhematocrit tubes from the tails of newborn mice or from the retroorbital sinus of weanling and adult mice. Erythrocytes were washed in 20 volumes of cold 0.85% NaCI, collected by centrifugation and lysed in 5 volumes of distrlled water. Hemoglobin phenotypes were determined by cellulose acetate electrophoresis (Titan Ill, Helena Laboratories) of cystam- me-modihed hemoglobrns (Whitney, 1978). Immediately after electrophore- sis, the surface of the cellulose acetate plates were covered with a clear mylar film and the hemoglobin fractions quantitated in a Flur-vis autoscan- ning densitometer (Helena Laboratones).

Hematology was performed on freshly drawn, heperinized blood col- lected from the retroorbrtal sinus of mice at 6-7 weeks of age. Hematocnt (microhematocrit method), red cell and nucleated cell (hemocytometer), hemoglobrn (cyanmethemoglobin calorimetric method), and reticulocyte (new methylene blue supra vrtal staining) (Brecher, 1949) values were determined by standard procedures (Wintrobe et al., 1974) and values for mean corpuscular hemoglobrn and mean corpuscular red cell volume were computed. The nucleated cell values represent the white cells plus the Immature nucleated red blood cells in the blood of B-thalassemic mice.

Globin Synthesis Relative rates of synthesis of 01. and &globrn chains were determined in vitro by 3H-leucine incorporation into globin of reticulocytes collected from mace that had recerved multiple injections of phenylhydrazine (Mar-knell et al., 1981). Freshly prepared reticulocytes were washed and collected by centrifugation and incubated in 5 ml of minimal essential medium (without leucrne) that contained 0.5% NaHC&, 0.6 mg of human transferrin, 1.25 mg of Fe(NH&(S0&.6H20 and 0.25 mCi of ?-f-leucine (pH 7.4). These conditions gave linear incorporation of ‘H-leucine for at least 3 hr. In this study, a representative time point of 90 min was chosen. Reticulocytes were incubated at 37’C, washed several times in isotonic saline, and lysed in 4 volumes of distilled water. Globin chains were precipitated with acidified acetone and separated by chromatography over carboxymethyl-cellulose (Whatman CM23) using a 5 to 30 mM gradient of sodium phosphate in 8 M urea (pH 6.7) (Clegg et al.. 1966). Absorbance at 280 nm of the protein in the eluate was recorded and samples of each fraction were analyzed for 3H in a Irqurd scrntillation counter.

Nucleic Acid Preparations Genomic DNA was purified from isolated liver nuclei by phenol-chloroform extractron followed by density gradient centrifugation in 5.3 M CsC12 solution containing 5 mg/ml propidium iodide (Radloff et al., 1965). Samples were centrifuged rn a Beckman vTi 50 rotor at 24tl,OOCI x g for 12 hr at 15°C. Punfred DNA was digested with a 5-fold excess of endonuclease (Bethesda Research Laboratones) overnight at 37°C and loaded on submerged 1% agarose gels in 0.089 M Tris EDTA Borate (pH 8.3) for electrophoresis at 4V/cm overnrght. After electrophoresis, the DNA was denatured and trans- ferred to GenetranTM 45 filters (D and L Filter Corporation) following the procedure of Southern (1975). After transfer, the filters were dried at 80°C for 2 hr in vacua and sealed in plastic bags for hybndrzation.

A Mouse Model for o-Thalassemia 1051

Hybridization Probes o-major cDNA was purified from the lambda derivative C3HAHb4 (Blattner et al., 1977). pPK268 and pPK288 are pBR322 derivatives (P, Kretschmer, unpublished data) containing p-major and flanking sequences from BALB/ CJ mace originally cloned into bacteriophage h (Tilghman et al., 1977). Both pPK266 and pPK288 were grown in E. coli C6OO host. A 5.6 kb Eco RI- Ava I fragment and a 3.4 kb Eco RI-Xba I fragment were prepared from pPK268 by double digestion with Eco RI and Ava I or Xba I and purified from plasmid DNA by electroelution.

Nick translations of probe DNA were performed as described by Rigby et al. (1977) using =P-o dCTP (3003 Ci/mmole) purchased from New England Nuclear. Labeled probes were separated from unincorporated nucleotides by chromatography over G-50 Sephadex columns. Specific activfties of probes were greater than 108 cpm/cg DNA.

Hybridization Baked filters were prehybridized for 4 hr at 42°C in 50% formamide, 5X SSC, 5x Denhardt’s, 1% SDS, 1 mM EDTA, in 10% dextran, 20 mM sodium phosphate (pH 7.0) buffer. Probe (10’ cpmpilter) and 290 rg of sonicated salmon sperm DNA were heat-denatured for 5 min at 90°C in 1 ml of prehybridrzation solution, added to the prehybridization mix, and hybndrzed to the filter for 60 hr at 42°C. Hybridized filters were washed three trmes at 55’C in 1 x SSC, 0.1% SDS, 2 hr per wash. Labeled fragments were visualized by autoradrography (DuPont Cronex X-ray film and Cronex intensifyrng screens) after 48 hr exposure at -80°C. Analysis of restriction patterns was performed using the NA2 automated nuclerc acid analyzer (Bethesda Research Laboratories).

We wash to thank Carolyn Felton, Frank Deal, Julret Drake, and Chris Worthy for their expert technrcal assistance. We are also grateful to Dr. Oliver Smithies for the generous grft of the p-major globs? cDNA. This work was supported in part by the National Institute of Environmental Health Scrences (Contract Number NOl-ES-25012) and by the Offrce of Health and Envi- ronmental Research, U. S. Department of Energy, under Contract Number W-7405.eng-26 with the Union Carbide Corporation.

The costs of publication of this artrcle were defrayed in part by the payment of page charges. Thus artfcle must therefore be hereby marked “advertisement” in accordance wrth 18 U.S.C. Sectron 1734 solely to rndrcate this fact.

Recerved May 23, 1983; revised July 18. 1983

References

Aksoy, M., and Ercten, S. (1969). Some problems of hemoglobin patterns in different thalassemic syndromes showing the heterogeneity of beta- thalassemia genes. Ann. N.Y. Acad. Sci. 765, 13-24.

Anderson, W. F., and Fletcher, J. C. (1980). Gene therapy in human beings: when is It ethical to begin? N. Eng J. Med. 303, 1273-1296.

Baird, M., Dnscoll, C., Schriener, H., Sciarratta, G. V., Sansone, G., Niazi, G., Ramrrez, F., and Bank, A. (1981). A nucleotide change at a splrce junction In the human fi-globin gene is associated with pBo-thalassemia. Proc. Nat. Acad. Sci USA 78. 4218-4221.

Benz, E.J., Jr., and Forget, B. G. (1975). The thafassemia syndromes: models for the molecular analysis of human disease. Prog. Hemato. 9, 107-155.

Benz, E.J., Jr., and Forget, 6. G. (1980). Pathogenesis of the thalassemia syndromes. Pathobrol. Ann. 10, l-32.

Benz, E. J., Jr., Forget, B. G., Hillman, D. G., Cohen-Solaf, M., Pritchard, J., Cavallesco, C., Prenksy. W.. and Housman, D. (1978). Variability in the amount of /3-globin mRNA in $ thalassemia. Cell 14, 299-312.

Blattner, F. R.. Wrllrams, B. G.. Blechl, A. E., Denniston-Thompson, K., Faber, H. E.. Ferlong, L.-A., Grunwald. D. J., Kiefer, D. O., Moore, D. D., Schumm, J. W., Sheldon, E. L., and Smrthies, 0. (1977). Charon phages: safer derivakves of bacteriophage lambda for DNA clonrng. Science 196, 161-169

Brecher, G. (1949). New methylene blue as a reticulocyte stain. Am. J. Clin. Pathol. 79, 299-308.

Clegg, J. B., Naughton, M. A., and Weatherall, D. J. (1966). Abnormal human haemoglobins. Separation and characterization of the alpha and beta chains by chromatography and the determination of two new variants, hb Chesapeake and hb J (Bangkok). J. Mol. Biol. IQ, 91-108.

Cline. M. J. (1982). Genetic engineering of mammalian cells: its potential application to genetic diseases of man. J. Lab. Clin. Med 99, 299-308.

Edgell, M. H.. Weaver, S.. Jahn, C. L., Padgett. A. W., Phillips, S. W.. Voliva, C. F.. Comer, M. B., Hardies, S. C., Haigwocd. N. L., Langley, C. H., Racine, R. R., and Hutchison, C. A. (1981). The mouse @-hemoglobin locus. In Organization and Expression of Globin Genes, G. Stamatoyannopoulos and A. W. Nienhuis. eds. (New York: Alan R. Liss, Inc.) pp. 69-88.

Fritsch, E. F., Lawn, R. M., and Maniatis. T. (1979). Characterization of deletions which affect the expression of fetal globin genes in man. Nature 279, 598-603.

Gilman, J. G. (1976). Mouse hemoglobin beta chains: comparative se- quence data on adult major and minor beta chain from two species. Biochem. J. 159, 43-53.

Haigwood. N. L., Jahn, C. L., Hutchinson, C. A.. Ill, and Edgell. M. H. (1981). Locations of three repetitive sequence families found in BALB/c adult 8-globin clones. Nucl. Acids Res. 9, 1133-l 150.

Haynes, J. R., Rosteck, P., and Lrngrel, J. B. (1980). Unusual sequence homology at the 5’ ends of the developmentally regulated beta A, beta C and gamma globin genes of the goat. Proc. Nat. Acad. Sci. USA 77, 7127- 7131.

Hrlse. K.. and Popp, R. A. (1968). Gene duplication as the basis for amino acid ambigurty in the alpha-chain polypeptides of mouse hemoglobins. Proc. Nat. Acad. Sci. USA 67, 930-936.

Hutton, J. J. (1969). Linkage analyses using biochemical variants in mice. Linkage of the hemoglobin beta-chain and glucose phosphate rosmerase loci. Biochem. Genet. 3, 507-515.

Jagadeeswaran, P., Tuan. D., Forget, B. G.. and Weissman, S. M. (1982). A gene deletron ending at the midpoint of a repetitive DNA sequence in one form of hereditary persistence of fetal haemoglobin. Nature 296, 469- 470.

Jahn. C. L., Hutchison, C. A., Ill, Phillips, S. J., Weaver, S.. Hargwood. N. L., Voliva, C. F., and Edgell, M. H. (1980). DNA sequence organization of the @globin complex in the BALB/c mouse. Cell 21, 159-168.

Johnson, F. M., and Lewis, S. E. (1981). Electrophoretically detected germrnal mutations induced rn the mouse by ethylnitrosourea. Proc. Nat. Acad. SCI. USA 78, 3138-3141.

Johnson, F. M.. Roberts, G. T., Sharma. R. K., Chasatow, F., Zweidinger. R.. Morgan, A., Hendren, R. W., and Lewis, S. E. (1981). The detection of mutants in mice by electrophoresis: results of a model induction experiment wrth procarbazrne. Genetics 97, 113-124.

Kan, Y. W., Dozy, A. M., Trecartin, R., and Todd, D. (1977). Identification of a non-deletion defect in a-thalassemia. N. Eng. J. Med. 297, 1081-1084.

Kan, Y. W., Dozy, A. M., Varmus, H. E., Taylor, J. M., Holland, J. P.. Lie- lnjo, L. E., Ganesan, J., and Todd, D. (1975). Deletion of a-globm genes in haemoglobrn-H drsease demonstrates multiple a-globins. Nature 255, 255- 256.

Kretschmer, P. J., Coon, H. C., Davis, A., Harrison, M., and Nienhuis, A. W. (1981). Hemoglobin switching In sheep: isolation of the fetal 8-globin gene and demonstration that the fetal 6 and adult pB*-globrn genes lie within 8 kb segments of homologous DNA. J. Biol. Chem. 256, 1975-1982.

Lacy, E., Hardison, A. C.. @on, D., and Maniatis, T. (1979). The linkage arrangement of four rabbit &like globin genes. Cell 18. 1273-1283.

Lawn, R. M., Fritsch. E. F., Parker, R. C., Blake, G., and Maniatis, T. (1978). The isolatron and characterization of linked & and @-globin genes from a cloned library of human DNA. Cell 75, 1157-I 174.

Ley, T. J.. DeSimone, J., Anagnou, N. P., Keller, G. H., Humphries, R. K., Turner, P. H., Young, N. S., Heller, P.. and Nienhuis. A. W. (1982). 5. azacytrdine selectively Increases y-globin synthesis rn a patient with p”- thalassemia. New Eng. J. Med. 307, 1469-1475.

Marsnell, J., Whrtney, J. B., Popp. R. A., Russell, L. 8.. and Anderson, W.

cell 1052

F. (1981). Three mouse models of human thalassemia. Proc. Nat. Acad. Sci. USA 78,505+586tl.

Mercola. K. E., and Cline, M. J. (1980). The potentials of inserting new genetic information. N. Eng. J. Med. 303. 1297-1300.

Nathan, D. G., and Gunn, R. B. (1988). Thaiassemia: the consequences of unbalanced hemoglobin synthesis. Am. J. Med. 47,815~830.

Old, J. M., Proudfoot, N. J., Wood, W. G.. Longley. J. I., Clegg, J. B., and Weatherall, D. J. (1978). Characterization of @globin in mRNA in the $ thalassemias. Cell 14,289~298

Whitney, J. B., III (1977). Differential control of the synthesis of two hemoglobin r!? chains in normal mice. Cell 72, 863-871,

Whitney, J. B., III (1978). Simplified typing of mouse hemoglobin (/-/be) phenotypes using cystamine. B&hem. Genet. 16, 887-872.

Witrobe, M. M., Lee. G. Ft., Bcgg, D. Ft.. Bithell, T. C., Athens, J. W., and Foerster, J. (1974). Clinical Hematology (Philadelphia: Lee and Febiier).

Orkin, S. H., Kazazian. H. H., Jr., Antonarakis, S. E.. Goff, S. C. Boehm, C. D., Sexton, J. P., Waber, P. G., and Giardina, P. J. V. (1982). Linkage of 6. thalassemia mutations and @globin gene polymorphisms with DNA poly morphisms in human @globin gene cluster. Nature 296, 827-831.

Orkin, S. H., Old, J. M., Weatherall, D. J., and Nathan, D. G. (1979). Partial deletion of @globin gene DNA in certain patients with #%thalassemia. Proc. Nat. Acad. Sci. USA 76, 24Cl&2404.

Ottolenghi, S., and Gigliori, B. (1982). The deletion of a type of 6-p thalassemia begins in an inverted A/u I repeat, Nature 300, 770-771.

Popp, Ft. A. (1987). Hemoglobins of mice: sequence and possible ambiguity at one position of the alpha-chain. J. Mol. Biol. 27, 9-18.

Popp, Ft. A. (1989). Studies on mouse hemoglobin loci. VIII. A fourth a- chain phenotype. J. Heredity 60. 126-133.

Popp, R. A. (1973). Sequence of amino acids in the &chain of single hemoglobins from C57BL, SWR, and NB mice. Biochim. Biophys. Acta. 303, 52-60.

Popp, R. A.. and Bailiff, E. G. (1973). Sequence of amino acids in the major and minor &chains of the diffuse hemoglobin from BALB/c mice. B&him. Biophys. Acta. 303. 61-67.

Radloff, R., Bauer, W., and Vinograd. J: (1985). A dye-buoyant density method for the detection and isolation of closed circular duplex DNA: the closed circular DNA in HeLa cells. Proc. Nat. Acad. Sci USA 57, 1514- 1521.

Ramirez, F., O’Donnell, J. V., Marks, P. A., Bank, A., Musumeci, S., Schiliro, G., Piuarelli, G.. Russo, G., Luppis, B., and Gambino, R. (1976). Abnormal or absent &mRNA in @Ferrara and gene deletion in 6.6.thalassemia. Nature 263,47I-475.

Ranney, H. M.. Smith, G. M., and Gluecksohn-Waelsch, S. (1960). Hae- moglobin differences in inbred strains of mice. Nature 788, 212-214.

Rigby, P. W. J., Dieckmann, M., Rhodes, C., and Berg, P. (1977). Labeling deoxyribonucleic acid to high specific activity by nick translation with DNA polymerase I. J. Mol. Biol. 7 13, 237-251.

Russell, E. S. (1979). Hereditary anemias of the mouse. A review for geneticists. Adv. Genet. 20, 357-459.

Russell, E. S., and McFarland, E. C. (1974). Genetics of mouse hemoglo- bins. Ann. NY Acad. Sci. 241, 25-38.

Skow, L. C. (1978). Genetic variation at a locus (Tam-l) for submaxillary gland protease in the mouse and its location on chromosome 7. Genetics 90, 713-724.

Southern, E. M. (1975). Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biil. 98, 563-517.

Tam, J. W. O., Kaufman, R. E., and Nienhuis, A. W. (1981). Analysis of globin gene structure in patients with &thalassemia by restriction endonu- clease mapping. Hemoglobin 5. 209-215.

Tilghman, S. M., Tiemeier. D. C., Polsky, F., Edgell. M. H., Seidman, J. G., Leder, A., Enquist, L. W., Norman, B., and Leder, P. (1977). Cloning specific segments of the mammalian genome: bacteriophage h containing mouse globin and surrounding gene sequences, Proc. Nat. Acad. Sci. USA 74, 4406-4410.

Treisman, R., Orkin, S. H., and Maniatis, T. (1988). Specific transcription and RNA splicing defects in five cloned 6thalassemia genes. Nature 302, 591-596.

Weatherall, D. J., and Clagg, J. B. (1972). The Thalassemia Syndrome, Second Edition (Oxford: Blackwell), pp. I-374.

Weatherall, D. J., and Clegg, J. B. (1982). Thalassemia revisited, Cell 29, 7-9.