- Email: [email protected]
Mutation of macrophage colony stimulating factor (Csf1) causes osteopetrosis in the tl rat
BBRC Biochemical and Biophysical Research Communications 294 (2002) 1114–1120 www.academicpress.com
Mutation of macrophage colony stimulating factor (Csf1) causes osteopetrosis in the tl rat David E. Dobbins,a,* Raman Sood,b Akira Hashiramoto,c Carl T. Hansen,d Ronald L. Wilder,e and Elaine F. Remmersf a
Department of Anatomy, Physiology and Genetics, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814-4799, USA b Cancer Genetics Branch, National Human Genome Research Institute, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892, USA c National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892, USA d Veterinary Resources Program, Office of Research Services, NIH, 9000 Rockville Pike, Building 14F/Room 101, Bethesda, MD 20892, USA e MedImmune Inc., 35 West Watkins Mill Road, Gaithersburg, MD 20878, USA f National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892, USA Received 24 May 2002
Abstract Osteopetrosis results from a heterogeneous group of congenital bone diseases that display inadequate osteoclastic bone resorption. We recently mapped tl (toothless), a mutation that causes osteopetrosis in rats, to a genetic region predicted to include the rat Csf1 gene. In this study, we sequenced the coding sequence of the rat Csf1 gene to determine if a mutation in Csf1 could be responsible for the tl phenotype. Sequencing revealed a 10-base insertion in the coding sequence of mutant animals that produces a frameshift and generates a stop codon early in the mutant Csf1 coding sequence. The 41 amino acid polypeptide predicted to be produced from the Csf1 promoter would have only the first nine amino acids of the wild-type rat protein. These data suggest that osteopetrosis develops in tl/tl rats because they cannot produce functional mCsf, a growth factor required for osteoclast differentiation and activation. Ó 2002 Published by Elsevier Science (USA). Keywords: Osteopetrosis; Osteoclasts; Osteoblasts; Colony stimulating factor 1; Bone disease; tl rats; Bone homeostasis; Growth factors
Bone is an extremely dynamic tissue that undergoes remodeling throughout life. Control of bone homeostasis is achieved through a complex combination of influences involving numerous cytokines, growth factors, steroid hormones, and various calcitropic molecules [1,2]. These multi-functional effector molecules regulate both osteoclastic bone resorption and osteoblastic bone formation. Any disequilibrium in the control of osteoclasts and osteoblasts can lead to a variety of bone diseases in which inappropriate bone homeostasis is manifest. Osteopetrosis results from a heterogeneous group of congenital bone diseases that are characterized by a generalized increase in skeletal mass, resulting from inadequate osteoclastic bone resorption [2–4]. Four autosomal-recessive, non-allelic mutations cause osteo*
Corresponding author. Fax: +1-301-295-3566. E-mail address: [email protected] (D.E. Dobbins).
petrosis in the rat: (1) toothless (tl), (2) osteopetrotic (op), (3) incisors absent (ia), and (4) microphthalmic blanc (mib) [5,6]. Each of these models of failed bone homeostasis exhibits different phenotypical profiles as to the number of osteoclasts present, the ability of the osteoclasts to form ruffled borders, disease onset and/or spontaneous recovery from disease, and the ability of bone marrow transplantation to rescue the phenotype [7]. The tl model is characterized by an extremely low number of osteoclasts (about 1% of normal littermates) and cannot be cured by bone marrow transplantation from normal littermates [3,7,8]. The fact that this model of osteopetrosis is not cured by transplantation of hematopoietic precursors from normal littermates suggests that the failure in osteoclastic bone resorption lies not within the hematopoietic cell lineage, from which the osteoclasts are derived, but is likely due to a failure in the bone microenvironment necessary to support the
0006-291X/02/$ - see front matter Ó 2002 Published by Elsevier Science (USA). PII: S 0 0 0 6 - 2 9 1 X ( 0 2 ) 0 0 5 9 8 - 3
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
differentiation and activation of osteoclasts [8]. It is known that two essential factors, receptor and activator of NF-jB ligand (RANKL) and macrophage colony stimulating factor (mCsf, gene symbol Csf1), both of which are produced by the osteoblasts, are both necessary and sufficient for the differentiation and activation of osteoclasts [9–11]. We recently mapped the tl mutation to a 2.5 cM region of rat chromosome 2 [12]. Based on maps of the homologous regions in the mouse and human, the rat Csf1 gene would be predicted to map within the genetic critical region delineated for the tl gene. Considering the importance of this growth factor in both the differentiation and activation of osteoclasts, Csf1 was a logical functional candidate for the tl gene. In this study, we mapped the rat Csf1 gene to the tl region of rat chromosome 2 using radiation hybrid mapping techniques [13] and assessed the coding sequences of the Csf1 gene in LEW.tl/tl congenic rats in comparison with two wild-type strains (BN and LEW) and the genetic stock on which the mutation occurred to determine whether mutation of this gene is responsible for osteopetrosis in the tl rat.
Methods LEW.tl/tl rats. The tl mutation was previously transferred to the inbred LEW/N background from stock obtained from the Great Lakes Naval Training Station in the late 1970s [12]. Because osteopetrotic tl/tl
1115
rats do not survive to breed, the tl mutation was maintained by intercrossing congenic LEW.tl heterozygous tl carriers. Homozygous affected animals were identified at weaning by small size and the failure of tooth eruption. Carriers of the tl mutation and homozygous wildtype offspring were identified by test mating with known carriers. Radiation hybrid mapping of genes to the tl region. The whole genome rat/hamster RH panel of 106 hybrid DNAs was purchased from Research Genetics (Huntsville, Ala). Primers for Csf1, Adora3, and Gnai3 were designed based on rat or mouse cDNA sequences using the Whitehead Institute’s Primer 3 software (http://www-genome.wi. mit.edu/cgi-bin/primer/primer3_www.cgi). Of those primers screened initially by PCR, those amplifying different sized fragments from rat and hamster DNA or those amplifying only rat DNA were used to score the RH cell lines for retention of rat chromosomal fragments (Table 1). Each PCR was performed with a reaction volume of 25 l1. Magnesium concentrations and annealing temperatures were optimized for each primer pair. PCR products were resolved on 1.5% agarose gels and visualized by ethidium bromide staining. Separate duplicate PCR for each locus were scored independently by two observers as present or absent and marker vectors were analyzed using the online RH mapping server available at http:// www.rgd.mcw.edu/RHMAPSERVER/ and the markers were positioned on the framework radiation hybrid maps available at this site. Preparation of cDNA. Spleens from LEW/N, BN/SsN, and LEW.tl/ tl rats (all 5-week-old males) were washed in phosphate-buffered saline and thoroughly homogenized in TRIzol Reagent (Gibco BRL, Gaithersburg, MD) according to the modified acid guanidine thiocyanate– phenol–chloroform method. cDNA was made by reverse transcription of 250 ng of each RNA sample, using the ThermoScript RT-PCR System (Invitrogen, Gaithersburg, MD). In brief, random hexamers were annealed to denatured RNA samples. The first-strand cDNA synthesis reaction was catalyzed by 1 U reverse transcriptase. The reaction was performed in 3 steps (25 °C for 10 min, 42 °C for 50 min, and 70 °C for 15 min) on an automatic heat block (Model PJ2000
Table 1 Primer pairs used for RH mapping and Csf1 sequencing Primer name
Primer sequence, 50 -30
RH mapping primers D2Wox27 D2Usu15 D2Usu16 D2Usu17
GATAATTGACATGTCCAGTTCC TTCATCACCTGGATCCTCATC TTTGATGGGGTCTTTCCATC AACTGTGCATGCCTTTCCTT
Primer name
Primer sequence, 50 -30
Csf1 cDNA sequencing primers D2Usu1 AAAGTTTGCCTCGGTGCTCT D2Usu2 AAGGTCCTGCAGCAGTTGAT D2Usu3 CCTGATTGCAACTGCCTGTA D2Usu4 CTTGGCTTGGGATGATTCTC D2Usu5 ACTGCTTCCCAAAAGCCACT D2Usu6 GCCCGTTTTAATTCCATTCC D2Usu7 AAAGTTTGCCTCGGTGCTCT Primer name
Gene CTGGCTGATGGTAGGATGAG GGCACATTGCAATCTCTGGT GCTGTCCCACCCTTTGAATA TCACCGCGTTAATACTGCTG
Cd53 Adora3 Csf1 Gnai3 Bases
ATCATCCAGCTGTTCCTGGT TACAGGCAGTTGCAATCAGG TCTGTCAGTCTCTGCCTGGAT TAGGGTTCACCTCTGTCAACG GAAAGACCCCATCAAAGCTG TGTGTGCCCAGCATAGAATC TTGACTGTCGATCAACTGCTG
3–390 304–737 718–1173 819–1267 1419–1984 1555–1842 3–333
Primer sequence , 50 -30
Csf1 genomic DNA sequencing primers D2Usu8 CGCTGCCCGTATGACC D2Usu9 GCCATGCCTTTCTCCTGTTA D2Usu10 GTCATTCACCCCAAAACGTG D2Usu11 CATATAAAGCCTGGCCACCA D2Usu12 AGGCCCATGTTCATGCTTAC D2Usu13 TTATTACTGTCCTGCCCTCTCC D2Usu14 TTTTTGACCCAGGAAAGGAA
GCAATAAGGAAGCCAGAACC CACCAGGATAGCCATTATCCA CTGCTGCTTTCATGCAGTTC AGGCAAGGCAAGCATATGAG CAATCCCATATCCCACCTTG GGGAATGGAGAAGATGCTCA ACCCAGAAGACAAACGCAGA
* Bases of ‘rat sequence’ NM_023981 which, upon further analysis, is proven to be mouse sequence and not rat sequence. All sequencing primers included M13 tails (see Methods).
1116
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
DNA Thermal cycler; Perkin–Elmer, Branchburg, NJ). Finally, 1 U of RNase H was added to the reactions. Csf1 sequencing. M-13 tailed primer sets (forward ¼ TGTAAAAC GACGGCCAGT, reverse ¼ CAGGAAACAGCTATGACC) were designed based on rat mRNA sequence (NM_023981, a sequence which is reported in the database to be from rat but which we determined is the mouse sequence) to amplify fragments that overlap the entire coding region of the Csf1 gene. These primer sets, generally designed to cover 300–400 bases and to overlap one another by at least 30 bases, were amplified using Clontech Advantage-GC cDNA kits. The total reaction volume was 50 ll. PCR was carried out in Perkin– Elmer 9600 machines using the PCR amplification conditions found to be most effective for the individual primers as revealed by initial screening. Generally, this was a program consisting of 94 °C for 1 min, followed by 35 cycles of 94 °C for 30 s, 60 °C for 3 min, and 70 °C for 3 min, and then holding at 4 °C. An aliquot (10 ll) of the PCR was separated on 1.5% agarose gels and the DNA was visualized by ethidium bromide staining. PCR products were purified from unincorporated nucleotides and primers using the QIAquick PCR Purification Kit (Qiagen). Sequencing reactions were performed using 10 ng/100 bp fragment size of purified PCR products and the PE Biosystems FS+ and Big Dye terminator chemistry [14] and analyzed on the PE Biosystems 377 XL DNA sequencing machine. Sequence data were analyzed using Sequencher software (Gene Codes, Ann Arbor, MI). Amplification of genomic DNA from the three strains and from homozygous wild-type tl stock animals was also undertaken using the Clontech Advantage-GC genomic DNA kits. The intronic sequences flanking the rat Csf1 exons were identified by searching the rat whole genome shotgun trace archive (http://www.ncbi.nlm.nih.gov/blast/ mmtrace.html). The reactions were carried out using the recommended volumes of the supplied reagents in accordance with manufacturer’s directions. Purification of PCR products and sequencing were handled in the same manner as the cDNA samples. All the primer pairs used for sequencing are listed in Table 1.
Results Radiation hybrid mapping of genes within the tl critical region revealed that the Csf1 gene in the rat mapped in close apposition to the Cd53 gene, a gene that we have shown to be tightly linked to the tl locus in a linkage study [12], i.e., there were no recombinants in 40 mutant F2 animals. Csf1 is bracketed by Adora3 and Gnai3 and the relative positions of Cd53, Adora3, Csf1, and Gnai3 (Fig. 1) are consistent with the relative positions of these genes in both the human and mouse genomes. Sequencing of the wild-type rat (BN and LEW) Csf1 cDNA generated a sequence that did not match the published sequence. Searches of sequence databases revealed that the complete published rat Csf1 cDNA sequence (NM_023981), including the untranslated regions, is identical to the mouse cDNA sequence. Furthermore, comparison of the cDNA sequences with rat genomic sequence showed that the published cDNA sequence had several mismatches, whereas our sequence showed perfect match at corresponding exons. We concluded that the published rat Csf1 sequence is actually mouse sequence and submitted to GenBank the rat Csf1 complete coding sequence we constructed from sequencing the BN and LEW Csf1 cDNAs (GenBank Accession Nos. AF514355 and AF514356). We identi-
Fig. 1. Radiation hybrid map of the tl region. The 2.5 cM critical region for the tl gene is represented by the open rectangle bounded by D2Arb16 and D2Wox8 [12]. The darkened rectangle represents a 32 cR region of RNO2 that contains the rat Csf1 gene (in filled rectangle) bounded by Adora3 and Gnai3. Cd53 has been shown to be non-recombinant with the tl locus. The three anonymous markers contained in open boxes are framework markers from the Medical College of Wisconsin’s RH map. Numbers to the right of the black bar indicate the distance between adjacent markers in cR.
fied the intron/exon boundaries of the genomic sequence encompassing the complete coding regions of the rat Csf1 gene (Table 2) by comparison of the cDNA with rat genomic sequences. Sequencing of cDNA from the mutant and wild-type animals revealed a 10-base insertion in the Csf1 cDNA from mutant animals, from the 28th to the 37th base following the adenine residue in the start codon (Fig. 2, GenBank Accession No. AF514357). This insertion, CGCGGGGCGC, is a 10-base repeat that occurs twice in the wild-type animals but is repeated a third time in the mutant animals (Fig. 2). The frameshift caused by this 10-base insertion results in the formation of a stop codon at bases 123–125 from the initial base of the start codon. This sequence would produce an mRNA that encodes a 41 amino acid protein with only the first 9 amino acids identical to those found in the wild-type rat mCsf protein (Fig. 2). The presence of this 10-base insertion was confirmed by sequencing of genomic DNA
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
1117
Table 2 Rat Csf1 intron/exon junctions
Fig. 2. Sequence of the mutant, BN, LEW, and tl stock animals displaying the 10-base insertion, CGCGGGGCGC that is repeated twice in wild-type animals but three times in the mutant animals, producing a stop codon at bases 123–125 after the adenine residue of the start codon. The predicted amino acid structure of these sequences is shown at the bottom of the figure.
from this region in the mutant animals and the two wildtype strains. Sequence analysis of genomic DNA from non-carrier progeny of the Osborne–Mendel tl carriers indicated that the insertion seen in the affected mutant animals was not present in the genetic background on which the mutation arose. Analysis of the coding sequence from the rat, human, and mouse predicts that the gene product of Csf1 in wild-type animals would produce a protein of 552 and 553 amino acids in the mouse and human and 566 amino acids in the rat (Fig. 3). The rat sequence contains two insertions (1 and 26 amino acids), as well as three deletions (1, 5, and 7 amino acids) relative to the mouse protein (Fig. 3).
Discussion Macrophage colony stimulating factor (mCsf, gene symbol Csf1) is an essential growth factor that controls the survival, proliferation, and differentiation of mononuclear phagocytes [15–17]. Through the process of alternative mRNA splicing and differential posttranslational processing, mCsf can be secreted into the circulation as a glycoprotein or a chondroitin sulfatecontaining proteoglycan or can be expressed as a transmembrane glycoprotein on the surface of mCsf producing cells such as mature osteoblasts. Much of our current knowledge of the role of mCsf on osteoclastogenesis has come from the spontaneous osteopetrotic
1118
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
Fig. 3. Alignment of the predicted amino acid sequence of mouse, rat, and human mCsf proteins. Conserved amino acids (including conservative substitutions) are shown in gray, non-conservative substitutions are shown in white boxes, and dashed lines represent missing residues in the corresponding species.
mutation in the mouse known as op. These mice display severe osteopetrosis to the exclusion of marrow space, totally lack mature osteoclasts, are stunted, and display failed tooth eruption, all as a result of a failure of osteoclastic bone resorption. The op mouse is a null mutant for mCsf due to a thymidine insertion into exon 4 of the Csf1 gene, which causes a translational frameshift and introduction of a stop codon in helix B. The mutated gene encodes a 94 amino acid truncated form of mCsf, which is not biologically active [18,19]. Several groups have now shown that injection of human recombinant mCsf rescues much of the phenotype of these mutants and results in the production and survival of mature bone-resorbing osteoclasts [8,18]. Corroboration of the vital role of mCsf in osteoclastogenesis and survival has also come from rat models of osteopetrosis [15]. The tl (toothless) osteopetrotic rat is a spontaneous mutant with severe osteopetrosis and extremely low numbers of both osteoclasts and macrophages [15]. It has been shown that the daily administration of human recombinant mCsf, if started on the day of birth, can significantly improve the skeletal phenotype of these animals, sustain tooth eruption, and provide support for mature osteoclasts and bone resorption. Although the phenotype is significantly ame-
liorated by the administration of exogenous mCsf, complete rescue is not possible. Metaphyseal sclerosis is still seen in the treated animals and osteoclasts cannot be found in the subepiphyseal areas of the long bones [15]. However, these data provide additional support for the crucial role of mCsf in both the production and survival of mature osteoclasts. We recently mapped the tl mutation to a 2.5 cM region on rat chromosome 2 [12]. By comparison with the homologous regions of the mouse and human genomes, the genetic critical region for the tl gene is predicted to contain the rat Csf1 gene. In light of the known crucial role for this growth factor in osteoclastogenesis and the activation of mature osteoclasts, this gene was a logical functional candidate for the tl gene. In this study, we mapped the rat Csf1 gene using radiation hybrid techniques [13] and sequenced the coding region of the Csf1 gene from wild-type and tl/tl rats in an attempt to identify a mutation that could be responsible for the tl phenotype. We identified a 10-base insertion which follows the 27th nucleotide beyond the adenine residue of the start codon that results in a frameshift and a resultant formation of a stop codon at bases 123–125 (Fig. 2). This transcript would produce a 41 amino acid truncated protein with only the first nine amino acids of the
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
50 end of the protein identical to the wild-type rat mCsf protein. There is, however, an additional methionine (whose codon is located at 650–652 bases from the adenine residue of the start codon) with a consensus Kozak sequence within the open reading frame of the gene, which if used as an alternate transcription start site, could produce a 351 amino acid protein lacking the Nterminal 213 amino acids of the wild-type rat mCsf protein. That product, if produced, is unlikely to be biologically active. It has been shown in the mouse that the initial 177 amino acids of the mCsf protein are essential for biological activity of the protein [20]. Therefore, these data indicate that the mutation responsible for the osteopetrosis in the tl rat is similar to that discovered in the op mouse, i.e., a coding sequence insertion that results in a frameshift and production of a truncated protein that is not biologically active. There are both significant similarities and significant differences between the phenotypes of the op mouse and the tl rat. Although osteopetrosis occurs in both op mice and tl rats, it is severe and does not improve with time in tl rats, whereas in op mice the osteopetrosis ameliorates with time and the animals spontaneously recover over the first few months after birth. The tl rat also exhibits dystrophic growth plates not seen in op mice [8]. Control of osteoclast function is regulated by several parallel pathways. Clearly, one pathway that stimulates osteoclastogenesis could compensate for a defect in another pathway. Interestingly, in the op mouse, it has been shown that the injection of human recombinant vascular endothelial growth factor (hrVEGF) can also result in growth and support of mature bone-resorbing osteoclasts. These data would suggest that, at least in the mouse model of osteopetrosis lacking biologically active mCsf, VEGF displays overlapping functions in the support of osteoclasts and bone resorption [19]. It is known that as the op mouse mutants age the osteopetrosis declines in severity and some mature bone-resorbing osteoclasts can be identified in older animals. It is believed that the increased endogenous production of VEGF in the aging mutants is responsible for the recruitment and support of osteoclasts and the decreased severity of the osteopetrosis observed in older animals. Apparently, this mechanism is not operative in the tl rat as osteopetrosis in the affected animals does not improve with time. Both the op mouse and the tl rat phenotypes are significantly improved by the administration of recombinant human mCsf, but complete rescue is not possible. This may well be due to the fact that although administration of recombinant human mCsf restores the circulating levels of the protein, it may not sufficiently reestablish appropriate local concentrations of growth factor. The recombinant human protein may not exhibit full biologic activity in rodents and it cannot restore the membrane-bound forms of the protein in the stromal cells that support osteoclast development and activation [8].
1119
The results of this study clearly demonstrate that the osteopetrosis seen in the tl rat is the result of a 10-base insertion within the coding sequence of the mutant strain. The resultant translational frameshift generates a truncated Csf1 protein that would not be predicted to be biologically active. Thus, the osteopetrosis seen in the tl rat is not the result of a failure in the hematopoietic cell lineage from which the osteoclasts arise, but from an inappropriate microenvironment provided by the stromal cells. These results are consistent with the fact that this form of failed bone homeostasis cannot be corrected with bone marrow transplantation from normal littermates. Our results imply that the mechanism of failed bone homeostasis in the tl rat is similar to that seen in the op mouse. These data underscore the pivotal role of mCsf in the maintenance of normal bone homeostasis.
Acknowledgments The authors wish to acknowledge the assistance of Christiane Robbins and Erica Eddings for their assistance in sequencing and PCR product purification.
References [1] S.L. Teitelbaum, Bone resorption by osteoclasts, Science 289 (2000) 1504–1508. [2] L.C. Hofbauer, S. Khosla, C.R. Dunstan, D.L. Lacey, W.J. Boyle, B.L. Riggs, The roles of osteoprotegerin and osteoprotegerin ligand in the paracrine regulation of bone resorption, J. Bone Miner. Res. 15 (1) (2000) 2–12. [3] P.R. Odgren, D.C. Hermy, S.N. Popoff, S.C. Marks, Cellular and molecular strategies for studying the regulation of bone resorption using the toothless (osteopetrotic) mutation in the rat, Histol. Histopathol. 12 (1997) 1151–1157. [4] T. Itzuka, M. Cielinski, S.L. Aukerman, S.C. Marks, The effects of colony-stimulating factor-1 on tooth eruption in the toothless (osteopetrotic) rat in relation to the critical periods for bone resorption during tooth eruption, Archs. Oral Biol. 37 (8) (1992) 629–636. [5] M.F. Siefert, S.N. Popoff, M.E. Jackson, C.A. MacKay, M. Cielinski, S.C. Marks, Experimental studies of osteopetrosis in laboratory animals, Clin. Orthop. 294 (1993) 23–33. [6] O.D. Benichou, B. Benichou, C. de Vernejoul, Osteopetrosis as a model for studying bone resorption, Rev. Rhum. 65 (12) (1998) 778–787. [7] A. Wojtowicz, A. Dziedzic-Goclawska, A. Daminski, W. Stachowicz, K. Wojtowica, S.C. Marks, M. Yamauchi, Alteration of mineral crystallinity and collagen cross-linking of bones in osteopetrotic toothless (tl/tl) rats and their improvement after treatment with colony stimulating factor-1, Bone 20 (2) (1997) 127–132. [8] P.R. Odgren, N. Kim, L. Van Wesenbeeck, C. MacKay, A. Mason-Savas, F.F. Safadi, S.N. Popoff, C. Lengner, W. Van-Hul, Y. Choi, S.C. Marks, Evidence that the rat osteopetrotic mutation toothless (tl) is not in the TNFSF11 (TRANCE, RANKL, ODF, OPGL) gene, Int. J. Dev. Biol. 45 (2001) 853–859. [9] G.J. Atkins, D.R. Haynes, S.M. Geary, M. Loric, T.N. Crotti, D.N. Findlay, Coordinated cytokine expression by stromal and hematopoietic cells during human osteoclast formation, Bone 26 (6) (2000) 653–661.
1120
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
[10] J. Li, I. Sarosi, X.-Q. Yan, S. Morony, C. Capparelli, H.-L. Tan, S. McCabe, R. Elliott, S. Scully, G. Van, S. Kaufman, S-C. Juan, Y. Sun, J. Tarpley, L. Martin, K. Christensen, J. McCabe, P. Kostenuik, H. Hsu, F. Fletcher, C.R. Dunstan, D.L. Lacey, W.J. Boyle, RANK is the intrinsic hematopoietic cell surface receptor that controls osteoclastogenesis and regulation of bone mass and calcium metabolism, Proc. Natl. Acad. Sci. USA 97 (4) (2000) 1566–1571. [11] T. Miyazaki, H. Katagiri, Y. Kanegae, H. Takayanagi, Y. Sawada, A. Yamamoto, M.P. Pando, T. Asano, I.M. Verma, H. Oda, K. Nakamura, S. Tanaka, Reciprocal role of ERK and NFjB pathways in survival and activation of osteoclasts, J. Cell Biol. 148 (2) (2000) 333–342. [12] E.F. Remmers, C.T. Hansen, A. Hashiramoto, R.L. Wilder, D.E. Dobbins, Application of interval haplotype analysis facilitates efficient mapping of the mutation causing osteopetrosis in tl rats, Mamm. Genome 13 (2002) 299–301. [13] D.E. Dobbins, B. Joe, A. Hashiramoto, J.L. Salstrom, S. Dracheva, L. Ge, R.L. Wilder, E.F. Remmers, Localization of the mutation responsible for osteopetrosis in the op rat to a 1.5 cM genetic interval on rat chromosome 10: identification of positional candidate genes by radiation hybrid mapping, J. Bone Miner. Res. (in press). [14] C.M. Robbins, E. Hsu, P.M. Gillevet, Sequencing homopolymer tracts and repetitive elements, Biotechniques 20 (1966) 862–868. [15] S.C. Marks, C.A. Mackay, M.E. Jackson, E.F. Larson, M.J. Cielinski, E.R. Stanley, S.L. Aukerman, The skeletal effects of colony-stimulating factor-1 in toothless (osteopetrotic) rats: per-
[16]
[17]
[18]
[19]
[20]
sistent metaphyseal sclerosis and the failure to restore subepiphyseal osteoclasts, Bone 14 (1993) 675–680. S.C. Marks, C. Lundmark, T. Wurtz, P.R. Odgren, C.A. MacKay, A. Mason-Savas, S.N. Popoff, Facial development and type III collagen RNA expression: concurrent expression in the osteopetrotic (toothless, tl) rat and rescue after treatment with colony-stimulating factor-1, Dev. Dyn. 215 (1999) 117–125. H. Watanabe, C.A. MacKay, A. Mason-Savas, E.H. Kislauskis, S.C. Marks, Colony-stimulating factor-1 (CSF-1) rescues osteoblast attachment, survival and sorting of b-actin mRNA in the toothless (tl-osteopetrotic) mutation in the rat, Int. J. Dev. Biol. 44 (2000) 201–207. H. Yoshida, S.I. Hayashi, T. Kunisada, M. Ogawa, S. Nishikawa, H. Okamura, T. Sudo, L.D. Shultz, S.-I. Nishikawa, The murine mutation osteopetrosis is in the coding region of the macrophage colony stimulating factor gene, Nature 345 (1990) 442–443. S. Niida, M. Kaku, H. Amano, H. Yoshida, H. Kataoka, S. Nishikawa, K. Tanne, N. Maeda, S.-I. Nishikawa, H. Kodama, Vascular endothelial growth factor can substitute for macrophage colony-stimulating factor in the support of osteoclastic bone resorption, J. Exp. Med. 190 (2) (1999) 293–298. M. Takahashi, T. Hirato, M. Takano, T. Nishids, K. Nagamura, T. Kamogashira, S. Nakai, Y. Hirai, Amino-terminal region of human macrophage colony-stimulating factor (M-CSF) is sufficient for its in vitro biological activity: molecular cloning and expression of carboxy-terminal deletion mutants of human MCSF, Biochem. Biophys. Res. Commun. 116 (2) (1989) 892–901.
Mutation of macrophage colony stimulating factor (Csf1) causes osteopetrosis in the tl rat David E. Dobbins,a,* Raman Sood,b Akira Hashiramoto,c Carl T. Hansen,d Ronald L. Wilder,e and Elaine F. Remmersf a
Department of Anatomy, Physiology and Genetics, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814-4799, USA b Cancer Genetics Branch, National Human Genome Research Institute, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892, USA c National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892, USA d Veterinary Resources Program, Office of Research Services, NIH, 9000 Rockville Pike, Building 14F/Room 101, Bethesda, MD 20892, USA e MedImmune Inc., 35 West Watkins Mill Road, Gaithersburg, MD 20878, USA f National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892, USA Received 24 May 2002
Abstract Osteopetrosis results from a heterogeneous group of congenital bone diseases that display inadequate osteoclastic bone resorption. We recently mapped tl (toothless), a mutation that causes osteopetrosis in rats, to a genetic region predicted to include the rat Csf1 gene. In this study, we sequenced the coding sequence of the rat Csf1 gene to determine if a mutation in Csf1 could be responsible for the tl phenotype. Sequencing revealed a 10-base insertion in the coding sequence of mutant animals that produces a frameshift and generates a stop codon early in the mutant Csf1 coding sequence. The 41 amino acid polypeptide predicted to be produced from the Csf1 promoter would have only the first nine amino acids of the wild-type rat protein. These data suggest that osteopetrosis develops in tl/tl rats because they cannot produce functional mCsf, a growth factor required for osteoclast differentiation and activation. Ó 2002 Published by Elsevier Science (USA). Keywords: Osteopetrosis; Osteoclasts; Osteoblasts; Colony stimulating factor 1; Bone disease; tl rats; Bone homeostasis; Growth factors
Bone is an extremely dynamic tissue that undergoes remodeling throughout life. Control of bone homeostasis is achieved through a complex combination of influences involving numerous cytokines, growth factors, steroid hormones, and various calcitropic molecules [1,2]. These multi-functional effector molecules regulate both osteoclastic bone resorption and osteoblastic bone formation. Any disequilibrium in the control of osteoclasts and osteoblasts can lead to a variety of bone diseases in which inappropriate bone homeostasis is manifest. Osteopetrosis results from a heterogeneous group of congenital bone diseases that are characterized by a generalized increase in skeletal mass, resulting from inadequate osteoclastic bone resorption [2–4]. Four autosomal-recessive, non-allelic mutations cause osteo*
Corresponding author. Fax: +1-301-295-3566. E-mail address: [email protected] (D.E. Dobbins).
petrosis in the rat: (1) toothless (tl), (2) osteopetrotic (op), (3) incisors absent (ia), and (4) microphthalmic blanc (mib) [5,6]. Each of these models of failed bone homeostasis exhibits different phenotypical profiles as to the number of osteoclasts present, the ability of the osteoclasts to form ruffled borders, disease onset and/or spontaneous recovery from disease, and the ability of bone marrow transplantation to rescue the phenotype [7]. The tl model is characterized by an extremely low number of osteoclasts (about 1% of normal littermates) and cannot be cured by bone marrow transplantation from normal littermates [3,7,8]. The fact that this model of osteopetrosis is not cured by transplantation of hematopoietic precursors from normal littermates suggests that the failure in osteoclastic bone resorption lies not within the hematopoietic cell lineage, from which the osteoclasts are derived, but is likely due to a failure in the bone microenvironment necessary to support the
0006-291X/02/$ - see front matter Ó 2002 Published by Elsevier Science (USA). PII: S 0 0 0 6 - 2 9 1 X ( 0 2 ) 0 0 5 9 8 - 3
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
differentiation and activation of osteoclasts [8]. It is known that two essential factors, receptor and activator of NF-jB ligand (RANKL) and macrophage colony stimulating factor (mCsf, gene symbol Csf1), both of which are produced by the osteoblasts, are both necessary and sufficient for the differentiation and activation of osteoclasts [9–11]. We recently mapped the tl mutation to a 2.5 cM region of rat chromosome 2 [12]. Based on maps of the homologous regions in the mouse and human, the rat Csf1 gene would be predicted to map within the genetic critical region delineated for the tl gene. Considering the importance of this growth factor in both the differentiation and activation of osteoclasts, Csf1 was a logical functional candidate for the tl gene. In this study, we mapped the rat Csf1 gene to the tl region of rat chromosome 2 using radiation hybrid mapping techniques [13] and assessed the coding sequences of the Csf1 gene in LEW.tl/tl congenic rats in comparison with two wild-type strains (BN and LEW) and the genetic stock on which the mutation occurred to determine whether mutation of this gene is responsible for osteopetrosis in the tl rat.
Methods LEW.tl/tl rats. The tl mutation was previously transferred to the inbred LEW/N background from stock obtained from the Great Lakes Naval Training Station in the late 1970s [12]. Because osteopetrotic tl/tl
1115
rats do not survive to breed, the tl mutation was maintained by intercrossing congenic LEW.tl heterozygous tl carriers. Homozygous affected animals were identified at weaning by small size and the failure of tooth eruption. Carriers of the tl mutation and homozygous wildtype offspring were identified by test mating with known carriers. Radiation hybrid mapping of genes to the tl region. The whole genome rat/hamster RH panel of 106 hybrid DNAs was purchased from Research Genetics (Huntsville, Ala). Primers for Csf1, Adora3, and Gnai3 were designed based on rat or mouse cDNA sequences using the Whitehead Institute’s Primer 3 software (http://www-genome.wi. mit.edu/cgi-bin/primer/primer3_www.cgi). Of those primers screened initially by PCR, those amplifying different sized fragments from rat and hamster DNA or those amplifying only rat DNA were used to score the RH cell lines for retention of rat chromosomal fragments (Table 1). Each PCR was performed with a reaction volume of 25 l1. Magnesium concentrations and annealing temperatures were optimized for each primer pair. PCR products were resolved on 1.5% agarose gels and visualized by ethidium bromide staining. Separate duplicate PCR for each locus were scored independently by two observers as present or absent and marker vectors were analyzed using the online RH mapping server available at http:// www.rgd.mcw.edu/RHMAPSERVER/ and the markers were positioned on the framework radiation hybrid maps available at this site. Preparation of cDNA. Spleens from LEW/N, BN/SsN, and LEW.tl/ tl rats (all 5-week-old males) were washed in phosphate-buffered saline and thoroughly homogenized in TRIzol Reagent (Gibco BRL, Gaithersburg, MD) according to the modified acid guanidine thiocyanate– phenol–chloroform method. cDNA was made by reverse transcription of 250 ng of each RNA sample, using the ThermoScript RT-PCR System (Invitrogen, Gaithersburg, MD). In brief, random hexamers were annealed to denatured RNA samples. The first-strand cDNA synthesis reaction was catalyzed by 1 U reverse transcriptase. The reaction was performed in 3 steps (25 °C for 10 min, 42 °C for 50 min, and 70 °C for 15 min) on an automatic heat block (Model PJ2000
Table 1 Primer pairs used for RH mapping and Csf1 sequencing Primer name
Primer sequence, 50 -30
RH mapping primers D2Wox27 D2Usu15 D2Usu16 D2Usu17
GATAATTGACATGTCCAGTTCC TTCATCACCTGGATCCTCATC TTTGATGGGGTCTTTCCATC AACTGTGCATGCCTTTCCTT
Primer name
Primer sequence, 50 -30
Csf1 cDNA sequencing primers D2Usu1 AAAGTTTGCCTCGGTGCTCT D2Usu2 AAGGTCCTGCAGCAGTTGAT D2Usu3 CCTGATTGCAACTGCCTGTA D2Usu4 CTTGGCTTGGGATGATTCTC D2Usu5 ACTGCTTCCCAAAAGCCACT D2Usu6 GCCCGTTTTAATTCCATTCC D2Usu7 AAAGTTTGCCTCGGTGCTCT Primer name
Gene CTGGCTGATGGTAGGATGAG GGCACATTGCAATCTCTGGT GCTGTCCCACCCTTTGAATA TCACCGCGTTAATACTGCTG
Cd53 Adora3 Csf1 Gnai3 Bases
ATCATCCAGCTGTTCCTGGT TACAGGCAGTTGCAATCAGG TCTGTCAGTCTCTGCCTGGAT TAGGGTTCACCTCTGTCAACG GAAAGACCCCATCAAAGCTG TGTGTGCCCAGCATAGAATC TTGACTGTCGATCAACTGCTG
3–390 304–737 718–1173 819–1267 1419–1984 1555–1842 3–333
Primer sequence , 50 -30
Csf1 genomic DNA sequencing primers D2Usu8 CGCTGCCCGTATGACC D2Usu9 GCCATGCCTTTCTCCTGTTA D2Usu10 GTCATTCACCCCAAAACGTG D2Usu11 CATATAAAGCCTGGCCACCA D2Usu12 AGGCCCATGTTCATGCTTAC D2Usu13 TTATTACTGTCCTGCCCTCTCC D2Usu14 TTTTTGACCCAGGAAAGGAA
GCAATAAGGAAGCCAGAACC CACCAGGATAGCCATTATCCA CTGCTGCTTTCATGCAGTTC AGGCAAGGCAAGCATATGAG CAATCCCATATCCCACCTTG GGGAATGGAGAAGATGCTCA ACCCAGAAGACAAACGCAGA
* Bases of ‘rat sequence’ NM_023981 which, upon further analysis, is proven to be mouse sequence and not rat sequence. All sequencing primers included M13 tails (see Methods).
1116
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
DNA Thermal cycler; Perkin–Elmer, Branchburg, NJ). Finally, 1 U of RNase H was added to the reactions. Csf1 sequencing. M-13 tailed primer sets (forward ¼ TGTAAAAC GACGGCCAGT, reverse ¼ CAGGAAACAGCTATGACC) were designed based on rat mRNA sequence (NM_023981, a sequence which is reported in the database to be from rat but which we determined is the mouse sequence) to amplify fragments that overlap the entire coding region of the Csf1 gene. These primer sets, generally designed to cover 300–400 bases and to overlap one another by at least 30 bases, were amplified using Clontech Advantage-GC cDNA kits. The total reaction volume was 50 ll. PCR was carried out in Perkin– Elmer 9600 machines using the PCR amplification conditions found to be most effective for the individual primers as revealed by initial screening. Generally, this was a program consisting of 94 °C for 1 min, followed by 35 cycles of 94 °C for 30 s, 60 °C for 3 min, and 70 °C for 3 min, and then holding at 4 °C. An aliquot (10 ll) of the PCR was separated on 1.5% agarose gels and the DNA was visualized by ethidium bromide staining. PCR products were purified from unincorporated nucleotides and primers using the QIAquick PCR Purification Kit (Qiagen). Sequencing reactions were performed using 10 ng/100 bp fragment size of purified PCR products and the PE Biosystems FS+ and Big Dye terminator chemistry [14] and analyzed on the PE Biosystems 377 XL DNA sequencing machine. Sequence data were analyzed using Sequencher software (Gene Codes, Ann Arbor, MI). Amplification of genomic DNA from the three strains and from homozygous wild-type tl stock animals was also undertaken using the Clontech Advantage-GC genomic DNA kits. The intronic sequences flanking the rat Csf1 exons were identified by searching the rat whole genome shotgun trace archive (http://www.ncbi.nlm.nih.gov/blast/ mmtrace.html). The reactions were carried out using the recommended volumes of the supplied reagents in accordance with manufacturer’s directions. Purification of PCR products and sequencing were handled in the same manner as the cDNA samples. All the primer pairs used for sequencing are listed in Table 1.
Results Radiation hybrid mapping of genes within the tl critical region revealed that the Csf1 gene in the rat mapped in close apposition to the Cd53 gene, a gene that we have shown to be tightly linked to the tl locus in a linkage study [12], i.e., there were no recombinants in 40 mutant F2 animals. Csf1 is bracketed by Adora3 and Gnai3 and the relative positions of Cd53, Adora3, Csf1, and Gnai3 (Fig. 1) are consistent with the relative positions of these genes in both the human and mouse genomes. Sequencing of the wild-type rat (BN and LEW) Csf1 cDNA generated a sequence that did not match the published sequence. Searches of sequence databases revealed that the complete published rat Csf1 cDNA sequence (NM_023981), including the untranslated regions, is identical to the mouse cDNA sequence. Furthermore, comparison of the cDNA sequences with rat genomic sequence showed that the published cDNA sequence had several mismatches, whereas our sequence showed perfect match at corresponding exons. We concluded that the published rat Csf1 sequence is actually mouse sequence and submitted to GenBank the rat Csf1 complete coding sequence we constructed from sequencing the BN and LEW Csf1 cDNAs (GenBank Accession Nos. AF514355 and AF514356). We identi-
Fig. 1. Radiation hybrid map of the tl region. The 2.5 cM critical region for the tl gene is represented by the open rectangle bounded by D2Arb16 and D2Wox8 [12]. The darkened rectangle represents a 32 cR region of RNO2 that contains the rat Csf1 gene (in filled rectangle) bounded by Adora3 and Gnai3. Cd53 has been shown to be non-recombinant with the tl locus. The three anonymous markers contained in open boxes are framework markers from the Medical College of Wisconsin’s RH map. Numbers to the right of the black bar indicate the distance between adjacent markers in cR.
fied the intron/exon boundaries of the genomic sequence encompassing the complete coding regions of the rat Csf1 gene (Table 2) by comparison of the cDNA with rat genomic sequences. Sequencing of cDNA from the mutant and wild-type animals revealed a 10-base insertion in the Csf1 cDNA from mutant animals, from the 28th to the 37th base following the adenine residue in the start codon (Fig. 2, GenBank Accession No. AF514357). This insertion, CGCGGGGCGC, is a 10-base repeat that occurs twice in the wild-type animals but is repeated a third time in the mutant animals (Fig. 2). The frameshift caused by this 10-base insertion results in the formation of a stop codon at bases 123–125 from the initial base of the start codon. This sequence would produce an mRNA that encodes a 41 amino acid protein with only the first 9 amino acids identical to those found in the wild-type rat mCsf protein (Fig. 2). The presence of this 10-base insertion was confirmed by sequencing of genomic DNA
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
1117
Table 2 Rat Csf1 intron/exon junctions
Fig. 2. Sequence of the mutant, BN, LEW, and tl stock animals displaying the 10-base insertion, CGCGGGGCGC that is repeated twice in wild-type animals but three times in the mutant animals, producing a stop codon at bases 123–125 after the adenine residue of the start codon. The predicted amino acid structure of these sequences is shown at the bottom of the figure.
from this region in the mutant animals and the two wildtype strains. Sequence analysis of genomic DNA from non-carrier progeny of the Osborne–Mendel tl carriers indicated that the insertion seen in the affected mutant animals was not present in the genetic background on which the mutation arose. Analysis of the coding sequence from the rat, human, and mouse predicts that the gene product of Csf1 in wild-type animals would produce a protein of 552 and 553 amino acids in the mouse and human and 566 amino acids in the rat (Fig. 3). The rat sequence contains two insertions (1 and 26 amino acids), as well as three deletions (1, 5, and 7 amino acids) relative to the mouse protein (Fig. 3).
Discussion Macrophage colony stimulating factor (mCsf, gene symbol Csf1) is an essential growth factor that controls the survival, proliferation, and differentiation of mononuclear phagocytes [15–17]. Through the process of alternative mRNA splicing and differential posttranslational processing, mCsf can be secreted into the circulation as a glycoprotein or a chondroitin sulfatecontaining proteoglycan or can be expressed as a transmembrane glycoprotein on the surface of mCsf producing cells such as mature osteoblasts. Much of our current knowledge of the role of mCsf on osteoclastogenesis has come from the spontaneous osteopetrotic
1118
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
Fig. 3. Alignment of the predicted amino acid sequence of mouse, rat, and human mCsf proteins. Conserved amino acids (including conservative substitutions) are shown in gray, non-conservative substitutions are shown in white boxes, and dashed lines represent missing residues in the corresponding species.
mutation in the mouse known as op. These mice display severe osteopetrosis to the exclusion of marrow space, totally lack mature osteoclasts, are stunted, and display failed tooth eruption, all as a result of a failure of osteoclastic bone resorption. The op mouse is a null mutant for mCsf due to a thymidine insertion into exon 4 of the Csf1 gene, which causes a translational frameshift and introduction of a stop codon in helix B. The mutated gene encodes a 94 amino acid truncated form of mCsf, which is not biologically active [18,19]. Several groups have now shown that injection of human recombinant mCsf rescues much of the phenotype of these mutants and results in the production and survival of mature bone-resorbing osteoclasts [8,18]. Corroboration of the vital role of mCsf in osteoclastogenesis and survival has also come from rat models of osteopetrosis [15]. The tl (toothless) osteopetrotic rat is a spontaneous mutant with severe osteopetrosis and extremely low numbers of both osteoclasts and macrophages [15]. It has been shown that the daily administration of human recombinant mCsf, if started on the day of birth, can significantly improve the skeletal phenotype of these animals, sustain tooth eruption, and provide support for mature osteoclasts and bone resorption. Although the phenotype is significantly ame-
liorated by the administration of exogenous mCsf, complete rescue is not possible. Metaphyseal sclerosis is still seen in the treated animals and osteoclasts cannot be found in the subepiphyseal areas of the long bones [15]. However, these data provide additional support for the crucial role of mCsf in both the production and survival of mature osteoclasts. We recently mapped the tl mutation to a 2.5 cM region on rat chromosome 2 [12]. By comparison with the homologous regions of the mouse and human genomes, the genetic critical region for the tl gene is predicted to contain the rat Csf1 gene. In light of the known crucial role for this growth factor in osteoclastogenesis and the activation of mature osteoclasts, this gene was a logical functional candidate for the tl gene. In this study, we mapped the rat Csf1 gene using radiation hybrid techniques [13] and sequenced the coding region of the Csf1 gene from wild-type and tl/tl rats in an attempt to identify a mutation that could be responsible for the tl phenotype. We identified a 10-base insertion which follows the 27th nucleotide beyond the adenine residue of the start codon that results in a frameshift and a resultant formation of a stop codon at bases 123–125 (Fig. 2). This transcript would produce a 41 amino acid truncated protein with only the first nine amino acids of the
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
50 end of the protein identical to the wild-type rat mCsf protein. There is, however, an additional methionine (whose codon is located at 650–652 bases from the adenine residue of the start codon) with a consensus Kozak sequence within the open reading frame of the gene, which if used as an alternate transcription start site, could produce a 351 amino acid protein lacking the Nterminal 213 amino acids of the wild-type rat mCsf protein. That product, if produced, is unlikely to be biologically active. It has been shown in the mouse that the initial 177 amino acids of the mCsf protein are essential for biological activity of the protein [20]. Therefore, these data indicate that the mutation responsible for the osteopetrosis in the tl rat is similar to that discovered in the op mouse, i.e., a coding sequence insertion that results in a frameshift and production of a truncated protein that is not biologically active. There are both significant similarities and significant differences between the phenotypes of the op mouse and the tl rat. Although osteopetrosis occurs in both op mice and tl rats, it is severe and does not improve with time in tl rats, whereas in op mice the osteopetrosis ameliorates with time and the animals spontaneously recover over the first few months after birth. The tl rat also exhibits dystrophic growth plates not seen in op mice [8]. Control of osteoclast function is regulated by several parallel pathways. Clearly, one pathway that stimulates osteoclastogenesis could compensate for a defect in another pathway. Interestingly, in the op mouse, it has been shown that the injection of human recombinant vascular endothelial growth factor (hrVEGF) can also result in growth and support of mature bone-resorbing osteoclasts. These data would suggest that, at least in the mouse model of osteopetrosis lacking biologically active mCsf, VEGF displays overlapping functions in the support of osteoclasts and bone resorption [19]. It is known that as the op mouse mutants age the osteopetrosis declines in severity and some mature bone-resorbing osteoclasts can be identified in older animals. It is believed that the increased endogenous production of VEGF in the aging mutants is responsible for the recruitment and support of osteoclasts and the decreased severity of the osteopetrosis observed in older animals. Apparently, this mechanism is not operative in the tl rat as osteopetrosis in the affected animals does not improve with time. Both the op mouse and the tl rat phenotypes are significantly improved by the administration of recombinant human mCsf, but complete rescue is not possible. This may well be due to the fact that although administration of recombinant human mCsf restores the circulating levels of the protein, it may not sufficiently reestablish appropriate local concentrations of growth factor. The recombinant human protein may not exhibit full biologic activity in rodents and it cannot restore the membrane-bound forms of the protein in the stromal cells that support osteoclast development and activation [8].
1119
The results of this study clearly demonstrate that the osteopetrosis seen in the tl rat is the result of a 10-base insertion within the coding sequence of the mutant strain. The resultant translational frameshift generates a truncated Csf1 protein that would not be predicted to be biologically active. Thus, the osteopetrosis seen in the tl rat is not the result of a failure in the hematopoietic cell lineage from which the osteoclasts arise, but from an inappropriate microenvironment provided by the stromal cells. These results are consistent with the fact that this form of failed bone homeostasis cannot be corrected with bone marrow transplantation from normal littermates. Our results imply that the mechanism of failed bone homeostasis in the tl rat is similar to that seen in the op mouse. These data underscore the pivotal role of mCsf in the maintenance of normal bone homeostasis.
Acknowledgments The authors wish to acknowledge the assistance of Christiane Robbins and Erica Eddings for their assistance in sequencing and PCR product purification.
References [1] S.L. Teitelbaum, Bone resorption by osteoclasts, Science 289 (2000) 1504–1508. [2] L.C. Hofbauer, S. Khosla, C.R. Dunstan, D.L. Lacey, W.J. Boyle, B.L. Riggs, The roles of osteoprotegerin and osteoprotegerin ligand in the paracrine regulation of bone resorption, J. Bone Miner. Res. 15 (1) (2000) 2–12. [3] P.R. Odgren, D.C. Hermy, S.N. Popoff, S.C. Marks, Cellular and molecular strategies for studying the regulation of bone resorption using the toothless (osteopetrotic) mutation in the rat, Histol. Histopathol. 12 (1997) 1151–1157. [4] T. Itzuka, M. Cielinski, S.L. Aukerman, S.C. Marks, The effects of colony-stimulating factor-1 on tooth eruption in the toothless (osteopetrotic) rat in relation to the critical periods for bone resorption during tooth eruption, Archs. Oral Biol. 37 (8) (1992) 629–636. [5] M.F. Siefert, S.N. Popoff, M.E. Jackson, C.A. MacKay, M. Cielinski, S.C. Marks, Experimental studies of osteopetrosis in laboratory animals, Clin. Orthop. 294 (1993) 23–33. [6] O.D. Benichou, B. Benichou, C. de Vernejoul, Osteopetrosis as a model for studying bone resorption, Rev. Rhum. 65 (12) (1998) 778–787. [7] A. Wojtowicz, A. Dziedzic-Goclawska, A. Daminski, W. Stachowicz, K. Wojtowica, S.C. Marks, M. Yamauchi, Alteration of mineral crystallinity and collagen cross-linking of bones in osteopetrotic toothless (tl/tl) rats and their improvement after treatment with colony stimulating factor-1, Bone 20 (2) (1997) 127–132. [8] P.R. Odgren, N. Kim, L. Van Wesenbeeck, C. MacKay, A. Mason-Savas, F.F. Safadi, S.N. Popoff, C. Lengner, W. Van-Hul, Y. Choi, S.C. Marks, Evidence that the rat osteopetrotic mutation toothless (tl) is not in the TNFSF11 (TRANCE, RANKL, ODF, OPGL) gene, Int. J. Dev. Biol. 45 (2001) 853–859. [9] G.J. Atkins, D.R. Haynes, S.M. Geary, M. Loric, T.N. Crotti, D.N. Findlay, Coordinated cytokine expression by stromal and hematopoietic cells during human osteoclast formation, Bone 26 (6) (2000) 653–661.
1120
D.E. Dobbins et al. / Biochemical and Biophysical Research Communications 294 (2002) 1114–1120
[10] J. Li, I. Sarosi, X.-Q. Yan, S. Morony, C. Capparelli, H.-L. Tan, S. McCabe, R. Elliott, S. Scully, G. Van, S. Kaufman, S-C. Juan, Y. Sun, J. Tarpley, L. Martin, K. Christensen, J. McCabe, P. Kostenuik, H. Hsu, F. Fletcher, C.R. Dunstan, D.L. Lacey, W.J. Boyle, RANK is the intrinsic hematopoietic cell surface receptor that controls osteoclastogenesis and regulation of bone mass and calcium metabolism, Proc. Natl. Acad. Sci. USA 97 (4) (2000) 1566–1571. [11] T. Miyazaki, H. Katagiri, Y. Kanegae, H. Takayanagi, Y. Sawada, A. Yamamoto, M.P. Pando, T. Asano, I.M. Verma, H. Oda, K. Nakamura, S. Tanaka, Reciprocal role of ERK and NFjB pathways in survival and activation of osteoclasts, J. Cell Biol. 148 (2) (2000) 333–342. [12] E.F. Remmers, C.T. Hansen, A. Hashiramoto, R.L. Wilder, D.E. Dobbins, Application of interval haplotype analysis facilitates efficient mapping of the mutation causing osteopetrosis in tl rats, Mamm. Genome 13 (2002) 299–301. [13] D.E. Dobbins, B. Joe, A. Hashiramoto, J.L. Salstrom, S. Dracheva, L. Ge, R.L. Wilder, E.F. Remmers, Localization of the mutation responsible for osteopetrosis in the op rat to a 1.5 cM genetic interval on rat chromosome 10: identification of positional candidate genes by radiation hybrid mapping, J. Bone Miner. Res. (in press). [14] C.M. Robbins, E. Hsu, P.M. Gillevet, Sequencing homopolymer tracts and repetitive elements, Biotechniques 20 (1966) 862–868. [15] S.C. Marks, C.A. Mackay, M.E. Jackson, E.F. Larson, M.J. Cielinski, E.R. Stanley, S.L. Aukerman, The skeletal effects of colony-stimulating factor-1 in toothless (osteopetrotic) rats: per-
[16]
[17]
[18]
[19]
[20]
sistent metaphyseal sclerosis and the failure to restore subepiphyseal osteoclasts, Bone 14 (1993) 675–680. S.C. Marks, C. Lundmark, T. Wurtz, P.R. Odgren, C.A. MacKay, A. Mason-Savas, S.N. Popoff, Facial development and type III collagen RNA expression: concurrent expression in the osteopetrotic (toothless, tl) rat and rescue after treatment with colony-stimulating factor-1, Dev. Dyn. 215 (1999) 117–125. H. Watanabe, C.A. MacKay, A. Mason-Savas, E.H. Kislauskis, S.C. Marks, Colony-stimulating factor-1 (CSF-1) rescues osteoblast attachment, survival and sorting of b-actin mRNA in the toothless (tl-osteopetrotic) mutation in the rat, Int. J. Dev. Biol. 44 (2000) 201–207. H. Yoshida, S.I. Hayashi, T. Kunisada, M. Ogawa, S. Nishikawa, H. Okamura, T. Sudo, L.D. Shultz, S.-I. Nishikawa, The murine mutation osteopetrosis is in the coding region of the macrophage colony stimulating factor gene, Nature 345 (1990) 442–443. S. Niida, M. Kaku, H. Amano, H. Yoshida, H. Kataoka, S. Nishikawa, K. Tanne, N. Maeda, S.-I. Nishikawa, H. Kodama, Vascular endothelial growth factor can substitute for macrophage colony-stimulating factor in the support of osteoclastic bone resorption, J. Exp. Med. 190 (2) (1999) 293–298. M. Takahashi, T. Hirato, M. Takano, T. Nishids, K. Nagamura, T. Kamogashira, S. Nakai, Y. Hirai, Amino-terminal region of human macrophage colony-stimulating factor (M-CSF) is sufficient for its in vitro biological activity: molecular cloning and expression of carboxy-terminal deletion mutants of human MCSF, Biochem. Biophys. Res. Commun. 116 (2) (1989) 892–901.