The authors have no conflict of interest
Congenic Strains of Mice for Verification and Genetic Decomposition of Quantitative Trait Loci for Femoral Bone Mineral Density†
Article first published online: 1 FEB 2003
Copyright © 2003 ASBMR
Journal of Bone and Mineral Research
Volume 18, Issue 2, pages 175–185, February 2003
How to Cite
Shultz, K. L., Donahue, L. R., Bouxsein, M. L., Baylink, D. J., Rosen, C. J. and Beamer, W. G. (2003), Congenic Strains of Mice for Verification and Genetic Decomposition of Quantitative Trait Loci for Femoral Bone Mineral Density. J Bone Miner Res, 18: 175–185. doi: 10.1359/jbmr.2003.18.2.175
- Issue published online: 2 DEC 2009
- Article first published online: 1 FEB 2003
- Manuscript Accepted: 25 OCT 2002
- Manuscript Revised: 10 SEP 2002
- Manuscript Received: 20 JUL 2002
- congenic mouse strains;
- bone mineral density;
- genetic loci;
- chromosome 1;
Peak femoral volumetric bone mineral density (femoral bone mineral density) in C57BL/6J (B6) 4-month-old female mice is 50% lower than in C3H/HeJ (C3H) and 34% lower than in CAST/EiJ (CAST) females. Genome-wide analyses of (B6 × C3H)F2 and (B6 × CAST)F2 4-month-old female progeny demonstrated that peak femoral bone mineral density is a complex quantitative trait associated with genetic loci (QTL) on numerous chromosomes (Chrs) and with trait heritabilities of 83% (C3H) and 57% (CAST). To test the effect of each QTL on femoral bone mineral density, two sets of loci (six each from C3H and CAST) were selected to make congenic strains by repeated backcrossing of donor mice carrying a given QTL-containing chromosomal region to recipient mice of the B6 progenitor strain. At the N6F1 generation, each B6.C3H and B6.CAST congenic strain (statistically 98% B6-like in genomic composition) was intercrossed to obtain N6F2 progeny for testing the effect of each QTL on femoral bone mineral density. In addition, the femoral bone mineral density QTL region on Chr 1 of C3H was selected for congenic subline development to facilitate fine mapping of this strong femoral bone mineral density locus. In 11 of 12 congenic strains, 6 B6.C3H and 5 B6.CAST, femoral bone mineral density in mice carrying c3h or cast alleles in the QTL regions was significantly different from that of littermates carrying b6 alleles. Differences also were observed in body weight, femoral length, and mid-diaphyseal periosteal circumference among these 11 congenic strains when compared with control littermates; however, these latter three phenotypes were not consistently correlated with femoral bone mineral density. Analyses of eight sublines derived from the B6.C3H-1T congenic region revealed two QTLs: one located between 36.9 and 49.7 centiMorgans (cM) and the other located between 73.2 and 100.0 cM distal to the centromere. In conclusion, these congenic strains provide proof of principle that many QTLs identified in the F2 analyses for femoral bone mineral density exert independent effects when transferred and expressed in a common genetic background. Furthermore, significant differences in femoral bone mineral density among the congenic strains were not consistently accompanied by changes in body weight, femur length, or periosteal circumference. Finally, decomposition of QTL regions by congenic sublines can reveal additional loci for phenotypes assigned to a QTL region and can markedly refine genomic locations of quantitative trait loci, providing the opportunity for candidate gene testing.
BONE MINERAL DENSITY (BMD) is an important component of bone strength and a recognized predictor of risk for osteoporotic fracture. As a phenotype, regardless of measurement methodology, BMD has a strong heritable component with estimates in humans(1) and in mice(2, 3) that genes account for 60–80% of variance in BMD. Recognition of a major genetic role in acquisition and maintenance of BMD has spurred much effort to identify genes that contribute to osteoporosis. In the human field, one genetic approach is to choose candidate genes and then test small populations for a relationship between allelic variation of the gene of choice and a clinically relevant phenotype, typically BMD of the spine or hip.(1) A companion approach uses sib-pairs analysis to reduce genetic variability, and to some extent environmental variation during development, thus improving assessment of association between bone density phenotypes and genotype information.(4, 5) A third approach is to identify affected families and look for Mendelian inheritance patterns, as was done for the high bone mass trait recently shown to be a mutation in the low-density lipoprotein (LDL) receptor-related protein 5 gene (LRP5).(6) Progress toward identifying critical genes has been frustratingly slow because of a myriad of issues such as methodologies (e.g., sample sizes, stratification, limited statistical tools, lack of consensus on detection levels), environmental variation (nutrition, lifestyles, cultural variables), subject health status (e.g., occult disease, comorbid conditions, use of pharmacologic agents), and the rudimentary state of our knowledge about genetic variability that underlies skeletal biology.
Experimental animal models are providing assistance in unraveling the complexity of skeletal genetics. Fortunately, many of the obstructing issues noted above for human studies are under investigational control, which assists with analyses and interpretation of experimental outcomes. One promising path for genetic investigations has focused on laboratory mice and use of mutations (spontaneous, transgenics, and knockouts) to disrupt bone functions and uncover BMD regulatory genes. Illustrative examples of spontaneous mutations are (1) spontaneous fracture (sfx) with distal femur fractures,(7) and (2) little (lit) with mutant GH-releasing hormone receptor, causing reduced skeletal growth and low BMD.(8, 9) Similarly, an example of transgenesis may be bound in mice overexpressing insulin-like growth factor I (IGF-I) that stimulates GH-deficient somatic growth and increased trabecular bone volume, but not osteoblast proliferation.(10, 11) Finally, two gene deletion examples are represented by (1) biglycan null (bgn−/−) mice with deficient bone formation and marrow stromal cell proliferation,(12, 13) and (2) core binding factor a1 null (cbfa1−/−) mice that fail to differentiate osteoblasts and mineralize bone.(14) These few examples selected from a growing literature in spontaneous and induced mutations do indeed point toward genes that, in dysfunctional states, can have a serious impact on bone acquisition and mineralization.
Another strategy of investigation approaches the genetics of skeletal BMD from the perspective that adult mice have complements of genes that yield normal adult skeletal status without obvious pathology. Such experimental studies have focused on inbred strains to ascertain numbers and linkage of loci affecting femoral BMD using genome-wide scanning methods for quantitative trait analyses. Genome-wide scanning methods for quantitative trait analyses have been applied to (1) the SAMP6 model of osteopenia by Shimizu et al.(15) and by Benes et al.(16) to reveal genetic loci (QTLs) on chromosomes (Chrs) 2, 7, 11, 13, and 16 for low bone mass; (2) crosses between C57BL/6J (B6) and DBA/2J (D2) inbred strains by Klein et al.(17, 18) to locate whole body areal BMD QTLs on Chrs 1, 2, 4, and 11; and (3) crosses between B6 and Castaneus (CAST) or C3H/HeJ (C3H) by Beamer et al.(2, 3) to locate femoral BMD QTLs on each of the seven Chrs already cited, plus six more (Chrs 3, 5, 6, 12, 14, and 18). It remains to be proven that femoral BMD QTLs on at least seven of the identified Chrs contain the same genetic loci as determine in other crosses assessed for whole body or bone site-specific BMD. Progress on skeletal BMD genetics in mice will greatly facilitate the daunting task facing human researchers because of the high homology (∼75%) between the human and mouse genomes (www.ensembl.org).
In this report, we present the results of testing QTLs responsible for differences in femoral BMD identified in related F2 crosses between two genetically diverse high femoral BMD strains, C3H or CAST, and a low femoral BMD strain, B6. Congenic strains for 12 significant QTLs were constructed, 6 from C3H and 6 from CAST. We found that 11 of the 12 congenics showed significant alterations in femoral BMD as predicted from genetic analyses of the B6C3HF2 (n = 1000) and B6CASTF2 (n = 700) progenies. Analyses of eight B6.C3H Chr 1 sublines sharpened the mapping for the Bmd5 QTL map on distal Chr 1 and revealed a second BMD QTL, Bmd19, in the proximal end of the B6.C3H-1T congenic segment.
MATERIALS AND METHODS
This study used three inbred strains of mice as sources of progenitor alleles—C57BL/6J (B6), CAST/EiJ (CAST), and C3H/HeJ (C3H). All mice were produced and maintained in our research colony under 14:10 h light:dark cycles. Mice were housed in polycarbonate cages (51 in2) in groups of three to five on sterilized Northern White Pine shavings bedding. Water was acidified with HCl to achieve a pH of 2.8–3.2 (to prevent bacterial growth) and was freely available. The diet used for all mice was autoclaved National Institutes of Health 31 (6% fat diet, Ca:P of 1.15:0.85, 19% protein, vitamin and mineral fortified; Purina Mills International, Richmond, IN, USA) and was freely available. Use of mice in this research project was reviewed and approved by the Institutional Animal Care and Use Committee of The Jackson Laboratory.
The female mice used for experimental analyses were necropsied at 4 months of age when acquisition of peak femoral total volumetric BMD was achieved.(2, 3) Body weights were recorded, and partial carcass preparations were preserved in 95% ethanol as previously described.(19) Femurs and lumbar vertebrae (data to be reported elsewhere) were isolated. Femoral lengths were measured by digital calipers (Stoelting, Wood Dale, IL, USA) before densitometry.
Generation of congenic strains
Each congenic strain was developed by transfer of a specific chromosomal region containing a bone density QTL from the high bone density strains, C3H or CAST, to the low bone density B6 strain. Transfer of the donor region was accomplished by first producing (B6 × C3H)N1F1 or (B6 × CAST)N1F1 offspring, and then backcrossing a N1F1 mouse to a recipient B6 strain mouse to obtain N2F1 progeny. Tail tip DNAs, made from female and male offspring by standard NaOH digestion method (see below), were genotyped to find carriers of the desired chromosomal regions. These carriers were mated to new B6 mice to generate N3F1 progeny for genotyping. This backcross mating system, followed by genotyping for carriers, was conducted for six cycles as previously described.(20) Six congenic strains were prepared from both C3H- and CAST-donated chromosomal segments. The congenic strain names and the QTL regions transferred from the donor C3H or CAST strains to the B6 recipient strain are given in Table 1.
Testing of phenotypic effects associated with each transferred femoral BMD QTL region was done at the N6 generation, when the genomes are estimated to be 98.4% B6 in composition.(21) N6F1mice were intercrossed to produce N6F2 progeny that were genotyped for alleles present in the transferred chromosomal regions. Mice in each congenic series—B6.C3H and B6.CAST—were retained if homozygous for C3H (c3h/c3h), CAST (cast/cast), or B6 (b6/b6) alleles within the QTL regions. Mice that were genotypically heterozygous (b6/c3 or b6/cast) for the congenic regions were not phenotyped for femoral BMD.
The Chr 1 femoral BMD QTL region present in the B6.C3H congenic set was selected to begin fine mapping of the Bmd5 QTL. After three to four more cycles of backcrossing to B6, genetic decomposition of the Chr 1 region was undertaken by genotype-based selection of progeny carrying various sized fragments of the transferred region derived from meiotic recombination during the backcross process. Males and females carrying the same small fragments were intercrossed, and their progenies were used to establish separate but overlapping sets of congenic sublines. Thus, mice within a subline were genetically homozygous for the Chr 1 segments from C3H with the remainder of their genomes 99+% homozygous for B6. The femoral BMD phenotype for each subline was compared with the progenitor B6 strain background. A total of eight B6.C3H-1 sublines carrying segments homozygous for C3H alleles were established for fine mapping of this QTL for femoral BMD. Congenic subline information is given in Table 2.
Genomic DNA was prepared by digestion of 1-mm tail tips in 0.5 ml of 50 mM NaOH for 10 minutes at 95°C and then pH was adjusted to 8.0 with 1M Tris-HCl. Genotyping of individual mouse DNAs was accomplished by polymerase chain reaction (PCR) using oligonucleotide primer pairs (Mit markers) from Research Genetics (Birmingham, AL, USA). These primer pairs amplify simple CA repeated sequences of anonymous genomic DNA that are of different sizes, and through gel electrophoresis, can uniquely discriminate between B6, C3H, or CAST genomes.(22) Details of standard PCR reaction conditions have been described previously.(2) PCR products from B6, C3H, CAST, and their F1 hybrids were used as electrophoretic standards in every gel to identify the genotypes of mice (i.e., b6/b6, b6/c3, c3h/c3h, b6/cast, cast/cast).
BMD measurements by peripheral quantitative computed tomography
Mice were necropsied at 4 months as described above. Isolated femora were assessed using peripheral quantitative computed tomography (pQCT; Stratec XCT 960M; Norland Medical Systems, Ft. Atkinson, WI, USA) as described previously.(19) Briefly, bones were isolated and stored in 95% EtOH until measured for bone parameters by pQCT. A threshold of 1.300 attenuation units was used to differentiate mouse bone from water, adipose tissue, muscle, and tendon. A threshold of 2.000 was used to differentiate high density cortical bone from lower density cancellous bone. Calibration of the densitometer was done with a set of hydroxy apatite standards (0.050–1.000 mg/mm3), yielding a correlation of 0.997 between standards and pQCT estimation of density. Daily confirmation of that calibration was confirmed using a phantom of known density. Precision of the XCT 960M for repeated measurement of femoral BMD was 1.2%. Isolated femora were scanned at 2-mm intervals over their entire lengths. Total and cortical BMD values were calculated by dividing the total mineral content by the appropriate bone volume and expressed as milligrams per cubic millimeter. The XCT 960M does not have sufficient resolution to accurately resolve trabecular bone volume; thus, such data are not presented. Femoral periosteal circumference were measured at the midpoint of the diaphysis. The CV for groups of progenitor strain femoral parameters at 4 months were (1) 2.8–3.1% for total BMD; (2) 8.4–9.8% for total mineral content; (3) 7.0–7.4% for total volume; (4) 2.1–3.6% for mid-diaphyseal periosteal circumference; and (5) 3.4–5.0% for mid-diaphyseal cortical thickness.
Statistical analyses of all data from congenic mice carrying the donated QTL regions (c3/c3 or cast/cast) were compared with littermates carrying B6 alleles (b6/b6) using StatView 4.5 software for Macintosh (Cary, NC, USA). These data were analyzed first by ANOVA to detect major genotype effects. Individual group means were assessed for significant differences by Fisher's protected least-squares difference (LSD) test. Correlation analyses were carried out as described.(23) Differences were considered statistically significant when p < 0.05.
The chromosomal regions containing femoral BMD QTLs in C3H and CAST progenitor strains were donated to the B6 progenitor recipient strain to develop the congenic strains and are presented in Table 1. Mit markers flanking the QTL region used to construct the congenic strains are given, with the centromeric end of each region listed first. Each congenic strain was given an abbreviated designation that includes the recipient strain, the donor strain, and the chromosome number containing the donated segment (e.g., B6.C3H-1T). The capital letter “T” indicates that the congenic strain carries the full or “total” sized donated segment. Twelve different congenic strains are shown, six carrying femoral BMD QTLs from C3H and six carrying femoral BMD QTLs from CAST.
Data are presented in Fig. 1 for body weight and three femoral phenotypes obtained from groups (n = 10–17) of 4-month-old N6F2 B6.C3H congenic strain littermate females that were genotypically either c3/c3 or b6/b6 in the QTL regions (see Table 1). Femoral BMD was the only phenotype that significantly differed within all six congenic strains when mice carrying the c3/c3 alleles were compared with littermate controls carrying the b6/b6 alleles in the femoral BMD QTL regions. Femoral BMD was significantly higher in B6.C3H-1T (7.0%), 4T (8.9%), 11T (4.1%), 13T (4.1%), and 18T (3.1%) mice when the c3/c3 alleles were present in the QTL region. In contrast, femoral BMD was 3.3% lower in B6.C3H-6T mice carrying c3/c3 alleles compared with mice carrying b6/b6 alleles in the femoral BMD QTL region of Chr 6. In three of the six congenic strains (B6.C3H-1T, 6T, and 18T), mice with the c3/c3 genotype had significant increases in body weight that paralleled the increases in BMD. When these three congenic strains were tested for correlation between femoral BMD and body weight, a significant correlation was found only within the B6.C3H-1T b6/b6 littermate controls (r = 0.74, p < 0.05), but not within the B6.C3H-1T c3h/c3h littermates. Significant correlations between femoral BMD and body weight were not found within the remaining three B6.C3H congenic strains (i.e., 4T, 11T, 13T).
With respect to femur size, two congenic strains carrying c3/c3 alleles, B6.C3H-4T and 6T, had significant but opposite differences in lengths, and three congenic strains carrying c3/c3 alleles, B6.C3H-4T, 6T, and 13T, showed significant but mixed differences in mid-diaphyseal periosteal circumferences compared with littermates carrying the b6/b6 alleles.
Figure 2 shows similar data on body weight and femoral phenotypes from the congenic strains derived from CAST donated chromosomal segments. Again, groups (n = 11–19) consisted of 4-month-old N6F2 congenic strain females that were genotypically either cast/cast or b6/b6 in the QTL regions described in Table 1. Four of the six B6.CAST congenic strains had significantly higher femoral BMD than their littermate controls. Specifically, femoral BMD was higher in B6.CAST-1T (8.1%), 3T (2.8%), 13T (3.5%), and 14T (5.6%) strain mice carrying cast/cast alleles compared with littermate mice carrying b6/b6 alleles. In contrast, femoral BMD was significantly lower in B6.CAST-5T (−4.5%) mice carrying cast/cast alleles, whereas B6.CAST-15T (−2.5%) showed a similar but nonsignificant trend (p = 0.14) for lower femoral BMD. With respect to body weight, only B6.CAST-14T mice carrying cast/cast alleles showed a significant change (−6.5%) compared with littermates carrying the b6/b6 alleles. No significant differences were observed in femur lengths among any B6.CAST congenics. Significant increases in mid-diaphyseal periosteal circumferences were found in B6.CAST-1T (3.2%) and 3T (3.5%) mice carrying cast/cast alleles compared with littermate controls carrying b6/b6 alleles.
The relative change in femoral BMD introduced by the donated chromosomal segments was similar in both B6.C3H and B6.CAST congenic strain sets. That is, c3h/c3h or cast/cast alleles with positive effects induced increases in femoral BMD that ranged from 2.8% to 8.9%, whereas c3h/c3h or cast/cast alleles with negative effects reduced mean femoral BMD from 2.5% to 4.5%. These changes are modest in size, but very similar to the percent of variance (1–10%) attributed to individual femoral BMD QTLs from analyses of the B6C3HF2 and B6CASTF2 progenies.(2, 3) Figure 3 presents the correlation between percent change in femoral BMD of both F2 progenies and the congenic strains of this report. The F2 femoral BMD data were obtained by calculating the percent change between mice that were c3h/c3h or cast/cast within a given QTL region and littermates that were b6/b6 within the same QTL region. The percent change in congenic strains' femoral BMD data were similarly calculated from means shown in Figs. 1 and 2. The regression of congenic femoral BMD on F2 femoral BMD for the six C3H and the six CAST QTLs are depicted in Fig. 3. The overall analysis revealed a highly significant correlation (r = 0.915, p < 0.0001), indicating that the magnitude of the QTL effect on femoral BMD observed in the original F2 data were predictive of the QTL effect in each respective congenic strain.
Donated chromosomal regions carrying femoral BMD QTLs range from 5.0 to 69.4 cM in size and require substantial reduction to discern the location of the gene(s) responsible for changes in femoral BMD noted above. Chr 1 was chosen for further analyses, because both C3H and CAST segments donated to the B6 strain background resulted in significantly increased femoral BMD. Furthermore, the composite interval map for Chr 1 presented in Fig. 4, derived from initial genetic analyses of B6CASTF2 and B6C3HF2 progeny,(2, 3) strongly indicated QTLs (Bmd1 from CAST; Bmd5 from C3H) located at the distal end of Chr 1, whereas the B6C3HF2 interval map shows a possible additional QTL at a more proximal Chr 1 location.
A nested set of Chr 1 congenic sublines was developed from the B6.C3H-1T congenic strain by additional backcrossing to the B6 background (five strains) or selective inbreeding of N6F2 carriers of smaller Chr 1 segments (three strains). Table 2 describes the eight sublines for B6.C3H-1 using markers for each region, the named BMD QTL thought to be present in each subline chromosomal region, and the current generation status. These sublines carry c3h/c3h alleles on the segments indicated in Table 2. Therefore, the controls used for comparison are bones from age- and gender-matched mice from the recipient B6 background strain.
Figure 5 presents the genetically defined set of nested congenic sublines for Chr 1 in the context of a Chr 1 schematic diagram. The original C3H Chr 1 segment donated to B6.C3H-1T is indicated, along with genetic distances and Mit markers used to define smaller chromosomal regions. The B6.C3H-1 sublines are indicated by vertical solid lines and labeled 1–1 to 1–10. Data on femoral phenotypes and body weight for the eight Chr 1 sublines are presented in Fig. 6. We found that five sublines (B6.C3H-1–1, 1–4, 1–5, 1–8, and 1–9) had significantly higher femoral BMD values than the B6 controls. Importantly, the proximal Chr 1 segment from C3H within B6.C3H-1–1 mice resulted in an increased femoral BMD compared with B6. Because the sublines B6.C3H-1–3, 1–7, and 1–10 did not show increased BMD, the portion of the region in subline 1–1 proximal to D1Mit334 and distal of D1Mit282 is highly likely to carry an additional femoral BMD QTL (Bmd19). This was suggested by the leading “shoulder” in the interval map for B6C3HF2 data in Fig. 4. The second femoral QTL, Bmd5, active in the four B6.C3H-1 sublines (1–4, 1–5, 1–8, and 1–9) lies toward the telomeric end of Chr 1 distal to D1Mit194. These latter four sublines, which share the Bmd5 QTL region from D1Mit194 to D1Mit403 (Fig. 5), also differ by ANOVA (p < 0.021) for mean femoral BMD levels. It may be important that B6.C3H-1–8 and 1–9 have the highest femoral BMD values and share a common region distal from D1Mit403 to D1mit511that is not present in either B6.C3H-1–4 or 1–5.
Figure 6 also presents data on phenotypes of body weight (Fig. 6B), femoral length (Fig. 6C), and mid-diaphyseal circumference (Fig. 6D) for the B6.C3H-1 sublines. We found that mean body weight was increased in each of the five sublines, showing significantly increased femoral BMD, and also in B6.C3H-1–7 subline mice that did not show a difference in femoral BMD. Regression analyses showed significant relationships between body weight and femoral BMD on within sublines B6.C3H-1–1 (r = 0.667; p < 0.001), B6.C3H-1–4 (r = 0.551; p < 0.001), and 1–8 (r = 0.644; p < 0.033), as well as 1–7 (r = 0.392; p < 0.001). On the other hand, no relationship for these parameters was found for sublines B6.C3H-1–5 and 1–9, which also showed increased femoral BMD. ANOVA of body weight from the four Chr 1 sublines sharing the distal Bmd5 QTL region showed a significant main effect (p = 0.006); however, there was no pattern of differences that could be related to the Chr 1 regions carried by each congenic subline.
In measures of femoral size seen in Fig. 6C, length was significantly increased only in the B6.C3H-1–4 subline; however, there was no statistically significant correlation of femoral length to femoral BMD. As would be anticipated, ANOVA of femur lengths in the four B6.C3H-1 sublines sharing the Bmd5 QTL region yielded a significant main effect (p = 0.002), wherein B6.C3H-1–4 femurs differed from all others. Although B6.C3H-1–4 carries C3H alleles proximal to D1Mit445 not found in B6.C3H-1–5, 1–8, or 1–9, neither subline 1–3 nor 1–10 showed increased femur length despite carrying the same proximal genetic region found in B6.C3H-1–4.
Finally, in mid-diaphyseal periosteal circumference data presented in Fig. 6D, this phenotype was significantly increased in four (B6.C3H-1–1, 1–4, 1–9, and 1–8) of the five sublines with increased femoral BMD. However, only the B6.C3H-1–1 line showed a correlation between BMD and periosteal circumference (r = 0.534; p < 0.011). ANOVA of the periosteal circumference in the four sublines sharing the Bmd5 QTL region showed a significant main effect (p = 0.008). Only in the B6.C3H-1–5 line was periosteal circumference different from 1–4, 1–8, and 1–9, yet all four sublines share the same region on Chr 1. Thus, cosegregation of periosteal circumference with BMD was not evident for the Bmd5 QTL region.
Congenic strains of mice are excellent tools for a variety of studies, including confirmation of the existence of individual QTLs for any number of phenotypes, defining allele effects on phenotypes, fine mapping and positional cloning, and biological studies of QTL effects. Our data on two sets of congenic strains show that QTL regions identified in genome-wide analyses of F2 intercross progeny did harbor loci that regulate femoral BMD. Moreover, there were few false positives, because only 1 of 12 QTLs chosen for congenic strain analyses proved not to significantly increase femoral BMD. At the N6 generation of backcrossing, congenic mice were statistically 98.4% B6 in genomic composition. The majority, but not all, of the remaining genomic material consists of the chromosomal QTL segments donated from either C3H or CAST progenitor strains. Any non-QTL genomic material declines with continued backcrossing, and by convention, a congenic strain is fully established after 10 cycles of backcrossing to the designated recipient strain. At the N6 generation, these mice might be more accurately considered to be “incipient” congenic strains. Nevertheless, assessment of N6F2 mice for presence of the phenotype(s) putatively associated with the QTL region proved enlightening. In addition, it is apparent that variation in normal alleles can account for a range of normal BMD values, and from a therapeutic perspective, manipulation of more than one genetic pathway may be required to obtain optimum BMD.
Given that the C3H and CAST progenitor strains have markedly higher femoral BMD than the B6 strain, it was anticipated that alleles at several loci would collectively act to increase femoral BMD when crossed to the B6 low BMD strain background. As shown in Fig. 3, the comparison of QTL effects on femoral BMD in the original F2 populations with those QTL effects on femoral BMD in the congenic strains were highly correlated. Overall in both sets of congenics, the predominant effect of the transferred QTLs was to increase BMD. However, in each set of congenic strains there was at least one QTL region from the C3H (B6.C3H-6T) and from the CAST (B6.CAST-5T) donor strains that significantly reduced femoral BMD. The B6.C3H-6T congenic is of special interest because this QTL region also carries a QTL for serum IGF-I.(20) In particular, the decreased femoral BMD seen at 4 months of age in the B6.C3H-6T congenic is accompanied by a 21% lower serum IGF-I, as well as reduced trabecular bone volume and microarchitecture. Although both phenotypes, femoral BMD and serum IGF-I, declined, it is not known yet whether the B6.C3H-6T carries a single genetic locus with pleiotropic effects or two distinct loci with independent effects on bone and serum IGF-I.
One exception to predictions from the F2 data sets is evidenced by the B6.CAST-15T congenic strain, which showed only a trend toward reduced femoral BMD, whereas the B6CASTF2 analyses predicted a significant decrease in BMD from this Chr 15 locus. It is possible that there is a significant interaction with an environmental factor or that multiple genetic loci are required to observe a Chr 15 femoral BMD QTL effect. It is also possible that further backcrossing to N10 will reduce the background genetic variation, and will permit a better assessment of QTL effects on femoral BMD in the B6.CAST-15T congenic strain mice. Overall, it may be concluded that allele effects yielding increased or decreased femoral BMD, predicted from the original F2 data sets, were recapitulated in the congenic strains, thus showing that the F2 analyses can predict both location of QTLs and effects of alleles on femoral BMD.
Inspection of the femoral BMD data in both Figs. 1 and 2 shows that there is variability among the mean values for littermate controls that are homozygous b6/b6 in the QTL regions. When the femoral BMD data for the six B6.C3H congenic strain littermate controls were examined by ANOVA, there was no significant strain effect. In contrast, similar analysis of the B6.CAST congenic strain littermate controls did yield a significant strain effect (p < 0.01). However, when controls for each congenic strain were compared with age- and gender-matched mice from the B6 progenitor colony, no significant differences in femoral BMD were detected. These data are indicative of subtle differences in phenotype that can be present among these N6F2 control groups because of continuing segregation of donor genes in the genome. Therefore, the best comparisons for genotype effects at the N6F2 stage of backcrossing are between littermates of differing genotypes (i.e., b6/b6 vs. c3h/c3h or cast/cast). With further backcrossing to the recipient strain to achieve a greater degree of homozygosity for the background alleles, one can then use the B6 recipient strain mice as appropriate controls.
Femoral BMD was the primary phenotype used to develop these congenic strains. Nevertheless, we were very interested in whether selection of femoral BMD was accompanied by selection of another phenotype, such as body weight, femur length, or mid-diaphyseal perisoteal circumference that could account for changes in femoral BMD assessed by pQCT. For example, in Fig. 1, the mean femoral BMD and body weight for B6.C3H-1T, −6T, and 18T congenics varied significantly in tandem. However, evidence of correlation between BMD and body weight was detected only within the littermate controls for B6.C3H-18T. There also was no consistent pattern for cosegregation of femoral BMD and body weight in the B6.CAST large region congenics, an observation that does not support a generic conclusion that BMD changes are driven by alterations in body mass. With respect to Chr 1 sublines, cosegregation of femoral BMD and body weight was observed in three of five B6.C3H-1 sublines. Yet, inspection of the Fig. 5 subline map for association of body weight with a common subregion was not informative. Femoral length was not correlated with BMD among the Chr 1 sublines sharing the Bmd5 QTL region. Interestingly, the B6.C3H-1–1 subline carrying the proximal Bmd19 QTL region did showed a modest correlation between periosteal circumference and femoral BMD. Overall, there were suggestive hints that femoral BMD and one or more of the body weight, femur length, or periosteal circumference phenotypes might be related. However, a consistent association between these ancillary phenotypes and the distal Chr 1 regions carrying Bmd5 was not demonstrable. These results led us to conclude us to conclude that either our numbers of observations were too low to detect actual relationships between BMD and these other phenotypes, or that in N6-N9 generation mice, these phenotypes vary independently and are caused by other donor genes that are still segregating in the congenic backgrounds, exerting strong effects on the overall phenotype.
The genome-wide analyses of F2 populations for loci underlying a complex phenotype often yield large regions of chromosomes with significant association between phenotype and genotype. Interval maps of such chromosomes may indicate that one-half or more of the genetic distance assigned to a chromosome could be involved with a QTL. Given the uncertainty of exactly where a QTL identified in an F2 cross might be located in a chromosomal region, we introgressed a large piece of each donor chromosome into the recipient strain background to avoid the risk of leaving the actual QTL behind. Once the phenotype has been established in a congenic strain, a critical question is whether a large QTL containing region consists of more than one locus. The value of the congenic strain as a genetic tool to address this question was clearly evident from the sublines developed for Chr 1. First, subline analyses led to the discovery that the initial Chr 1 QTL region in B6.C3H-1T harbors a second, subtle femoral BMD QTL that was designated Bmd19. This femoral BMD QTL resides within the 13-cM region defined by the proximal ends of both the B6.C3H-1T and B6.C3H-1–1 subline. We did not detect this proximal locus in the B6CASTF2 analysis, and the proximal region in question was not included in the initial CAST Chr 1-donated segment placed in B6.CAST-1T. We would speculate that CAST and B6 have the same alleles at this locus, and thus a phenotype difference was not detected. Second, the genetic decomposition of the B6.C3H-1T reduced the size of the region carrying the distal femoral BMD QTL by 2-fold. At this time, the B6.C3H-1 sublines define a large genetic region that extends proximally to D1Mit194 at 73.2 cM. It is possible but not proven that the distal Chr 1 femoral BMD QTL is the same in both C3H and CAST strains. If such proves to be the case, this locus is characterized by at least three (b6, c3h, cast) and perhaps more alleles, because Klein et al.(18) also report a DBA/2 total body areal BMD QTL in this distal region of Chr 1. Development of additional sublines carrying new recombinant chromosomal segments are needed to determine whether the distal Bmd5 locus is the same locus in both crosses. In fact, other loci affecting femoral BMD could yet be present in the distal part of Chr 1, because the four sublines B6.C3H-1–5, 1–4, 1–9, and 1–8 (Fig. 5) show a gradation in femoral BMD change that needs further investigation. Although we speculate that the femoral BMD variability will decline as the subline chromosomal segments become more uniformly homozygous at their ends, we cannot exclude the interpretation that the distal Bmd5 QTL may represent a cluster of linked genes.
During the time required to produce and age N6F2 mice for initial phenotype analyses, the process of backcrossing each congenic strain to the B6 progenitor genotype has continued. Moreover, we recently reported the results on the B6.C3H-6T congenic strain that have reached N9 generation. In B6.C3H-6T mice at both N6 and N9 generations, changes in serum IGF-I and femoral BMD phenotypes at 4 months were identical.(20) Preliminary femoral data from the other B6.C3H congenics at N9-N10 are fully supportive of the completed N6 generation data reported herein, although the B6.CAST congenics have not yet been tested. Importantly, the B6.C3H Chr 1 subline set, ranging from generation N6F6 through N9F8, show stable femoral BMD phenotype with establishment of homozygosity for donor segments in the B6 genomic background.
The distal region of Chr 1 that contains the mouse Bmd5 QTL carries numerous loci homologous to genes on human Chr 1q. Specifically, the distal region of mouse Chr 1 from 70 cM to the telomere at ∼110 cM shares extensive gene linkage homology with human 1q21-q43. Within this extensive domain, four reports from investigations of both unrelated and kindred human populations are relevant to genetic regulation of BMD. Reed et al.(24) reported a locus for familial absorptive hypercalciuria with low bone mass that mapped to 1q23.3-q24, proposing that functional polymorphisms in the soluble adenylate cyclase gene could be pathophysiologic for this disease.(25) Raymond et al.(26) reported a case control study of 140 postmenopausal white women wherein a modest, significant association was found between the osteocalcin gene polymorphisms on 1q25–31 and femoral neck BMD. On the other hand, Duncan et al.(27) compared 115 patients with low femoral neck or spinal BMD with 499 relatives with normal BMD for association with 23 candidate genes, but did not find an association with osteocalcin. Finally, Koller et al.(5) reported suggestive linkage of lumbar spinal BMD to 1q21–23 in a study of 595 healthy sister pairs. Irrespective of differences in statistical power of detection, it is reasonable to conclude that these homologous regions of Chr 1 in mouse and man harbor genes critical to adult bone density.
There are several limitations to the current study. First, as noted above, the data reported in this paper are limited to N6F2 congenic mice. The process of backcrossing to N10 is continuing with the very strong likelihood that the magnitude of effects noted at N6 will be at least that strong at N10. That belief is buttressed by recent work from our group showing that N9F2 6T congenics had nearly identical femoral BMD and serum IGF-I as N6F2 B6.C3H-6T congenic mice.(20) Second, it is important to note that there can be additional donor chromosomal material residing proximal or distal to the markers defining the transferred regions given in Tables 1 and 2. As the chromosomal segments carrying the femoral BMD QTLs are reduced in size by genetic recombination through continued backcrossing, exact definition of those segment ends becomes ever more critical. Use of two different mating systems, B6 with CAST and B6 with C3H, focused on distal Chr 1, increases the probability that adequate numbers of simple sequence length polymorphisms will be available to complete the fine mapping to 1 cM genetic distance or less and support positional cloning efforts. Third, there are subtle differences in the femoral BMD phenotype among N6F2 control (b6/b6) mice, especially for B6.CAST littermate controls (p = 0.01 among the controls). But we recognize these littermates remain the best control group for comparing phenotypic and genotypic differences (i.e., b6/b6 vs. c3h/c3h or cast/cast at a specific QTL). With further backcrossing to the recipient strain to achieve a greater degree of homozygosity for the background alleles, this variability should be further reduced. Fourth, it should be noted that we chose only one BMD phenotype for congenic development (i.e., femoral volumetric BMD as measured by pQCT). Clearly, from our previous work, we recognize that there can be major effect size differences when considering other skeletal phenotypes, including trabecular bone volume and microarchitecture.(28) Finally, the data in this report are based solely on female mice. Analyses of gender effects, recently reported for areal BMD in crosses between B6 and DBA/2 mice by Orwoll et al.,(29) are essential to full understanding of the BMD QTL effects observed in the B6.C3H congenic strains reported herein. At the present time, these congenics represent the strongest proof to date that QTLs for femoral BMD in F2 mice and are critical determinants of peak bone acquisition.
In summary, the two sets of congenic strains provided proof of principle that femoral BMD QTLs identified in genome-wide analyses of B6C3HF2 and of B6CASTF2 progeny could be individually isolated in a common genetic background, leading to significant changes in the femoral BMD phenotype. In congenics and sublines that have been tested at generation N6 through N10, the femoral BMD phenotype associated with the QTL region continues to be detectable. Sublines can be used to refine chromosomal locations of QTLs and will provide excellent tools for candidate gene identification and testing, as well as for biological studies of bone compartments, structure, and cellular functions associated with each femoral BMD QTL.
The authors thank H Coombs and A Silva for steadfast dedication and outstanding efforts in the animal quarters and laboratory. In addition, we are indebted to Drs G Cox and A Dorward for critical review and insightful comments directed toward improvement of the manuscript. This work was supported by grants from NIAMS (AR43618 and AR45433) and from the U.S. Army (DAMD17–96–1–6306 [WGB] and DAMD17–99–1–9571 [DJB]).
- 11999 Genetics of osteoporosis. Endocrine Rev 20:788–804.
- 21999 Quantitative trait loci for bone density in C57BL/6J and CAST/EiJ inbred mice. Mamm Genome 10:1043–1049., , , , , ,
- 32001 Quantitative trait loci for femoral and vertebral bone mineral density in C57BL/6J and C3H/HeJ inbred strains of mice. J Bone Miner Res 16:1195–1206., , , , , , ,
- 41998 Linkage of a QTL contributing to normal variation in bone mineral density to Chromosome 11q12–13. J Bone Miner Res 13:1903–1908., , , , , , , , , , , ,
- 52000 Genome screen for QTLs contributing to normal variation in bone mineral density and osteoporosis. J Clin Endocrinol Metab 85:3116–3120., , , , , , , , , , , ,
- 62002 A mutation in the LDL receptor-related protein 5 gene results in the autosomal dominant high-bone-mass trait. Am J Hum Genet 70:11–19., P CJ, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
- 72000 Spontaneous fracture (sfx): A mouse genetic model of defective peripubertal bone formation. Bone 27:619–626., , , , , , , ,
- 81976 Stimulation of growth in the Little mouse. J Endocrinol 71:37–45.,
- 92002 Genetic and GH/IGH-I effects on vertebral architecture. J Bone Miner Res 17:S1;S235., , , , , ,
- 101990 Expression of insulin-like growth factor I stimulates normal somatic growth in growth hormone-deficient transgenic mice. Endocrinology 127:1033–1040., , , , ,
- 112000 Targeted overexpression of insulin-like growth factor I to osteoblasts of transgenic mice: Increased trabecular bone volume without increased osteoblast proliferation. Endocrinology 141:2674–2682., , , , , , , , , , ,
- 121998 Targeted disruption of the biglycan gene leads to an osteoporosis-like phenotype in mice. Nat Genet 20:78–82., , , , , , , , , , , , , , ,
- 132002 Age-related osteoporosis in biglycan-deficient mice is related to defects in bone marrow stromal cells. J Bone Miner Res 17:331–340., , , ,
- 141997 Targeted disruption of cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell 89:755–764., , , , , , , , , , , , ,
- 151999 Identification of peak bone mass QTL in a spontaneously osteoporotic mouse strain. Mamm Genome 10:81–87., , , , , , , , , , , , , , , , ,
- 162000 Chromosomal mapping of osteopenia-associated quantitative trait loci using closely related mouse strains. J Bone Miner Res 15:626–633., , , , , ,
- 171998 Quantitative trait loci affecting peak bone mineral density in mice. J Bone Miner Res 13:1648–1656., , , ,
- 182001 Confirmation and fine mapping of chromosomal regions influencing peak bone mass in mice. J Bone Miner Res 16:1953–1961., , , , , ,
- 191996 Genetic variability in adult bone density among inbred strains of mice. Bone 18:397–403., , ,
- 202002 Generation of a new congenic mouse strain to test the relationships among serum IGF-I, bone mineral density and skeletal morphology in vivo. J Bone Miner Res 17:570–579., , , , , , , , , , ,
- 21SilverLM (ed.) 1995 Mouse Genetics. Oxford University Press, New York, NY, USA.
- 221996 A comprehensive genetic map of the mouse genome. Nature 380:149–152., , , , , , , , , , , , , , , , , , , ,
- 23SokalRR, RohlfFJ (eds.) 1995 Biometry: The Principles and Practice of Statistics in Biological Research, 3rd ed. W. H. Freeman, San Francisco, CA, USA.
- 241998 Location of a gene for absorptive hypercalciuria with bone loss to 1q24. Bone 23:S162., , , , , , ,
- 252002 Identification and characterization of a gene with base substitutions associated with the absorptive hypercalciuria phenotype and low spinal bone density. J Clin Endocrinol Metab 87:1476–1485., , , , , ,
- 261999 Osteocalcin: Genetic and physical mapping of the human gene BGLAP and its potential role in postmenopausal osteoporosis. Genomics 60:210–217., , , ,
- 271999 Suggestive linkage of the parathyroid receptor type 1 to osteoporosis. J Bone Miner Res 14:1993–1999., , , , , ,
- 282001 Variation in bone biomechanical properties, microstructure, and density in BXH recombinant inbred mice. J Bone Miner Res 16:206–213., , , , , , ,
- 292001 Gender specificity in the genetic determinants of peak bone mass. J Bone Miner Res 16:1962–1971., ,