The osteopetrotic mutation toothless (tl) is a loss-of-function frameshift mutation in the rat Csf1 gene: Evidence of a crucial role for CSF-1 in osteoclastogenesis and endochondral ossification

Abstract

The toothless (tl) mutation in the rat is a naturally occurring, autosomal recessive mutation resulting in a profound deficiency of bone-resorbing osteoclasts and peritoneal macrophages. The failure to resorb bone produces severe, unrelenting osteopetrosis, with a highly sclerotic skeleton, lack of marrow spaces, failure of tooth eruption, and other pathologies. Injections of CSF-1 improve some, but not all, of these. In this report we have used polymorphism mapping, sequencing, and expression studies to identify the genetic lesion in the tl rat. We found a 10-base insertion near the beginning of the open reading of the Csf1 gene that yields a truncated, nonfunctional protein and an early stop codon, thus rendering the tl rat CSF-1(null). All mutants were homozygous for the mutation and all carriers were heterozygous. No CSF-1 transcripts were identified in rat mRNA that would avoid the mutation via alternative splicing. The biology and actions of CSF-1 have been elucidated by many studies that use another naturally occurring mutation, the op mouse, in which a single base insertion also disrupts the reading frame. The op mouse has milder osteoclastopenia and osteopetrosis than the tl rat and recovers spontaneously over the first few months of life. Thus, the tl rat provides a second model in which the functions of CSF-1 can be studied. Understanding the similarities and differences in the phenotypes of these two models will be important to advancing our knowledge of the many actions of CSF-1.

Fig 1.

Breeding scheme. Filled symbols, mutants; open symbols, wild type; spotted symbols, heterozygous mutants. Cosegregation of polymorphic markers and the mutation was examined in F2 animals.

Fig 2.

tl phenotype breeds true in BN/SSN strain background. Radiographs of long bones of a newborn normal rat (n) and a mutant F2 generation littermate (tl) show high bone density and lack of marrow spaces in the mutant (arrows). H, humerus.

Fig 3.

Ideogram of rat chromosome 2 and genetic map of markers used to delineate candidate region. The candidate region is shown as a gray bar on the first vertical line. Key recombinants are given on the other vertical lines. Gray bars represent chromosomal regions that contain the disease gene, and white bars indicate regions that recombined and therefore cannot contain the disease gene. The lines between represent uninformative regions. The gray regions include Csf1.

Fig 4.

Alignment of amino acid sequences of rat, mouse, and human CSF-1 full-length ORF show high conservation. The insertion in the rat from position 476–501 of the consensus is due to a nucleotide duplication unique to the rat. Excluding this stretch, there is >93% identity among the species, plus a number of conservative substitutions. The points of tl rat and op mouse frameshifting insertions are indicated. The normal rat cDNA and amino acid sequences have been deposited in GenBank (accession no. ).

Fig 5.

Organization of rat Csf1 gene. (A) The ORF of the gene is dispersed in 8 exons, shown with black boxes. 5′ and 3′ UTRs are shown with stripes. The alternative splice acceptor site in exon 6 is marked (asterisk). (B) Sequences of the intron/exon boundaries of the rat Csf1 gene and the length of each. Uppercase letters, exon sequences; lowercase letters, intron sequences.

Fig 6.

The tl mutation. (A) Nucleotide and amino acid sequence of the tl CSF-1 mutation. The 10-nt repeat, underlined, commences in the third codon. It occurs twice as a single repeat in the normal sequence, but in the tl mutation it occurs three times. The impact on the protein is seen in the translations, below, in which the mutant sequence diverges 9 residues from the initiation Met, continues for 33 residues, and then reaches a stop codon (*). The signal sequence of the wild-type protein, double underlined, is absent from the tl. Only the first 126 bases of the ORFs are shown. (B) Genotyping with PCR primers that flank the mutation. After separating on a polyacrylamide gel, wild-type (+/+) fragments can be distinguished from mutants (tl/tl) carrying the 10-base tl duplication. Heterozygotes (tl/+) clearly show two bands corresponding to the mutant and wild-type alleles.

Fig 7.

Alternative splicing of rat CSF-1 mRNA reveals no skipping of tl mutation-containing region. CSF-1 cDNAs were amplified from rat cDNA, separated in an agarose gel, blotted, and probed with ³²P-labeled rat CSF-1 cDNA. Lanes 1–8 represent skeletal muscle, liver, kidney, lung, brain, testis, pancreas, and heart, respectively. (A) RT-PCR product encompassing the complete coding sequence of rat CSF-1. Two alternative transcripts of 0.96 and 1.91 kb, respectively, can be distinguished. The shorter fragment can be explained by alternative splicing deletion of the major part of exon 6, as has been described in mouse and human. (B) PCR product containing the first 3 exons of rat CSF-1 cDNA. Only one transcript can be distinguished, indicating no alternative splicing in the 5′ part of the transcript that could bypass the tl mutation.

Fig 8.

Skeletal phenotype of tl rat compared with op mouse. Proximal tibiae from 4-week-old normal rat (n), tl rat (tl), and op mouse (op) stained histochemically for the osteoclast marker enzyme tartrate-resistant acid phosphatase (TRAP) (red). Sections are aligned along the top of the growth plate (GP). Arrows indicate lower margin of growth plate. Osteoclasts are evident in n, where remodeling is most active: 1, top of the GP; 2, bottom of the GP at chondroosseous junction; 3, near the ends of metaphyseal trabeculae. No TRAP-positive osteoclasts are seen in the CSF-1 mutant tl rat. In contrast, by 4 weeks, the op mouse has partially reestablished osteoclast populations in the metaphysis. (Scale bar, 310 μm.)