Sex-Specific Genetic Loci for Femoral Neck Bone Mass and Strength Identified in Inbred COP and DA Rats

Abstract

Introduction

Hip fracture is the most devastating osteoporotic fracture type with significant morbidity and mortality. Several studies in humans identified chromosomal regions linked to hip size and bone mass. Animal models, particularly the inbred rat, serve as complementary approaches for studying the genetic influence on hip fragility. The purpose of this study is to identify sex-independent and sex-specific quantitative trait loci (QTLs) for femoral neck density, structure, and strength in inbred Copenhagen 2331 (COP) and Dark Agouti (DA) rats.

Materials and Methods

A total of 828 (405 males and 423 females) F₂ progeny derived from the inbred COP and DA strains of rats were phenotyped for femoral neck volumetric BMD (vBMD), cross-sectional area, polar moment of inertia (Ip), neck width, ultimate force, and energy to break. A whole genome screen was performed using 93 microsatellite markers with an average intermarker distance of 20 cM. Recombination-based marker maps were generated using MAPMAKER/EXP from the COP × DA F₂ data and compared with published Rat Genome Database (RGD) maps. These maps were used for genome-wide linkage analyses to detect sex-independent and sex-specific QTLs.

Results

Significant evidence of linkage (p < 0.01) for sex-independent QTLs were detected for (1) femoral neck vBMD on chromosomes (Chrs) 1, 6, 10, and 12, (2) femoral neck structure on Chrs 5, 7, 10, and 18, and (3) biomechanical properties on Chrs 1 and 4. Male-specific QTLs were discovered on Chrs 2, 9, and 18 for total vBMD, on Chr 17 for trabecular vBMD, on Chr 9 for total bone area, and on Chr 15 for ultimate force. A female-specific QTL was discovered on Chr 2 for ultimate force. The effect size of the individual QTL varied between 1% and 4%.

Conclusions

We detected evidence that sex-independent and sex-specific QTLs contribute to hip fragility in the inbred rat. Several QTLs regions identified in this study are homologous to human chromosomal regions previously linked to QTLs contributing to femoral neck and related phenotypes.

INTRODUCTION

Osteoporosis is a common polygenic disorder with reduced BMD and increased susceptibility to fracture at multiple skeletal sites.(1) Hip fracture is one of the most devastating osteoporotic fracture types causing significant morbidity and mortality(2) and has a greater incidence in women then men. BMD, structure, and strength are the primary skeletal determinants of hip fracture risk.(3–6) These phenotypes are under substantial genetic control,(7–9) and there is evidence that the heritability of bone structure at the hip is higher in women than men.(8)

In humans, several studies have shown linkage for hip BMD, axis length, and bone size to multiple chromosomal regions.(10–30) Some recent studies also provide evidence of sex-specific bone mass regulation at different skeletal sites including the hip in humans.(21,24,30) Unfortunately, none of these studies has yet identified a gene or genes underlying the variation in these hip phenotypes. To expedite the gene discovery process, it would be most effective to use extensive skeletal phenotyping along with gene expression and proteomic analyses to prioritize potential candidate genes. Unfortunately, such studies are extremely difficult, if not impossible, in humans but they can be undertaken in animal models.

Previously, we showed that skeletal mass, structure, and strength varies among inbred strains of rats.(31) We also showed that, despite similar body size, substantial variation exists in bone geometry and biomechanical properties between adult Copenhagen 2331 (COP) and Dark Agouti (DA) rats.(32) Femoral neck work to failure in COP rats was >3-fold greater than DA rats because of thicker femoral neck and shorter femur axis length in COP rats. Thus, COP rats seem to have alleles that substantially affect the hip structure, and these inbred rats may provide a unique resource to identify the genetics underlying the phenotypic variation at the femoral neck.

We generated a large number of second filial (F₂) progeny derived from a cross of COP and DA inbred rats having significant differences in their hip structure. We obtained multiple hip phenotypes, many of which could only be obtained in a rodent model. The purpose of this study was to identify the quantitative trait loci (QTLs) contributing to the variation in hip structure. In addition, by testing male and female F₂, we were able to also test for sex-specific genetic regulation of hip bone mass and strength.

MATERIALS AND METHODS

Animal breeding

Reciprocal mating of COP rats with DA rats was performed to first create a hybrid F₁ population, and the F₁ rats were intercrossed to create a total of 828 (405 males and 423 females) second filial (F₂) offspring. Once an F₂ population was created, the rats were allowed to grow to 6 mo (26 wk) of age before they were killed. Rat identities were recorded on data chips implanted subcutaneously and were verified using a scanner from Biomedic Data System (Seaford, DE, USA). The rats were housed at Indiana University's Laboratory Animal Resource Center (LARC), two rats per cage, and provided standard rat chow (NIH-31 Mouse/Rat diet 7017; Harlan Teklad, Madison, WI, USA) and water ad libitum. After death, rat left and right femora attached with muscle were dissected out, and the right femora were immediately stored at −20°C for biomechanical testing. The left femora were stripped of the muscle and soft tissues and transferred to 70% ethyl alcohol and stored at 4°C for densitometry analyses. The excised spleens were immediately stored in liquid nitrogen before transferring them to −80°C. The procedures performed throughout the experiment followed the guidelines of the Indiana University Animal Care and Use Committee (IACUC).

pQCT

The proximal ends of the left femora were placed in plastic tubes filled with 70% ethyl alcohol and centered in the gantry of a Norland Stratec XCT Research SA+pQCT (Stratec Electronics, Pforzheim, Germany). Single slice measurements of 0.26 mm thickness and a voxel size of 0.07 mm were taken perpendicularly through the femoral neck. Five consecutive scans perpendicular to the neck axis were obtained 0.25 mm apart from each other starting at the base of the femoral head and ending at the greater trochanter. For each slice, the X-ray source was rotated through 180° of projection. Volumetric BMD (vBMD; mg/cm³), cross-sectional area (CSA; mm²), and polar moment of inertia (Ip; mm⁴) of the femoral neck were measured from the average values of all five pQCT slices using the XCT Research SA Plus, software version 5.40. A density threshold of 900 was used to identify the boundary between trabecular bone and cortical bone, and the lower threshold of 300 was used to distinguish trabecular bone from soft tissue. The accuracy and repeatability of the pQCT measurements was confirmed using a method described previously.(33)

Femoral neck width measurement

The femoral neck width of each rat was measured in the anterior–posterior direction using digital calipers accurate to 0.01 mm, with a precision of +0.005 mm (Mitutoyo, Aurora, IL, USA). The shortest transverse distance of the femoral neck was considered the width of the femoral neck.

Biomechanical testing

The frozen right femora were brought to room temperature slowly in a saline bath and sectioned in the femur shaft at 25% of their total length from the distal end. The proximal ends of the femora were mounted vertically in a special chuck that clamped the femoral shaft to the lower platen of the material testing machine (Alliance RT/5; MTS Systems, Eden Prairie, MN, USA). The bones were held in place by a small (1 N) preload, and load was applied directly downward at a cross-head speed of 20 mm/min onto the femoral head at room temperature in monotonic axial compression until fracture. Force and displacement measurements were collected every 0.05 s. From the force versus displacement curves, ultimate force (Fu; N) and energy to break (U; mJ) were calculated in TestWorks software, version 4.06. Fu reflects the strength of the bone or maximum load that the bone can support before failing; U reflects the amount of energy the specimen can absorb before fracture.

DNA isolation and microsatellite marker genotyping

Genomic DNA was isolated from the individual rat spleen using the Puregene DNA isolation kit (Gentra Systems, Minneapolis, MN, USA) according to the manufacturer's protocol. Genotyping for each animal was accomplished using PCR with microsatellite markers (Research Genetics, Birmingham, AL, USA) previously shown to be polymorphic for COP and DA rats. A total of 93 markers, with an average intermarker spacing of 20 cM, were genotyped in 828 F₂ progeny to provide coverage of the entire genome (chromosomes [Chrs] 1–20, X). The PCR primer sequences were designed for multiplexing using fluorescent dyes. The Qiagen Multiplex PCR Kit was used for all multiplex PCR reactions (Qiagen, Valencia, CA, USA). Allele sizes of the fluorescently labeled PCR products were determined on an ABI 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) using the program, Genotyper, version 3.6.

Heritability measurement

Heritability of the phenotypic measures was estimated as unity minus the ratio of the pooled variance in the parental animals (COP and DA) to the phenotypic variance in the F₂ progeny. These variance estimates provided indices of the environmental and genetic variances, respectively, for each trait.

Quantitative linkage analysis

Quality assessment of the data was performed. Genotyping data for each marker were tested to ensure the expected 1:2:1 mendelian ratio. Recombination-based marker maps were generated using MAPMARKER/EXP with genotype data from all 828 F₂ animals.(34) Marker order and distances were compared with those previously published at the Rat Genome Database (RGD) website (http://rgd.mcw.edu/; June 2007). The following femoral neck phenotypes were prioritized for analysis: femoral neck total vBMD, cortical vBMD, trabecular vBMD, neck width, polar Ip, total area, cortical area, trabecular area, ultimate force, energy to break, stiffness, and elongation. The distribution of each phenotype was reviewed, and all outliers were verified or removed from further analysis. In addition, phenotypic distributions were also examined to detect sex differences. When sex differences were found (e.g., body weight), analyses were performed in the male and female F₂ separately, and combined analyses of the male and female F₂ were not performed. For the vast majority of phenotypes, different distributions were not observed for the two sexes, and analyses were performed in the male and female F₂ rats and the combined sample.

The recombination-based marker map generated using the data from the 828 F₂ animals was used to perform multipoint linkage analyses using the program R/qtl.(35) Permutation tests were performed to obtain appropriate genome-wide significance levels for the linkage results.(36) Because the phenotypes are correlated, the set of phenotypes for each rat was kept together and randomly reassigned (permuted) to another rat in the sample. This phenotypic reassignment was performed for each of the 828 F₂ animals to generate a permutated sample. This process was repeated to generate 5000 permutated datasets, and genome-wide linkage analysis was performed in each of the 5000 permutated datasets.(37) All phenotypes measured for each rat were included in the permutation approach. Thus, genome-wide significance thresholds appropriate for each phenotype were obtained. The maximum LOD scores for linkage for each phenotype was recorded in each permutated dataset. In this manner, the LOD significance thresholds for the 95th (suggestive) and 99th (significant) percentile of the maximum genome-wide LOD scores across all phenotypes were found to be 3.5 and 4.3, respectively. ANOVA was performed using the most significant marker in each QTL region to further characterize significant genotypic group differences and obtain effect size estimates. For measures resulting in detection of multiple QTLs, these markers were also included in an overall ANOVA model to estimate their combined contribution to phenotypic variation.

Sex specificity of each QTL detected in our primary analysis was evaluated according to the method described by Solberg et al.,(38) which compares two linkage models that differ only in a QTL-by-sex interaction effect. The full model (model 5) contains terms for QTL, sex, and QTL-by-sex interaction effect, whereas the reduced model (model 4) contains terms for QTL and sex only. At the position of each QTL achieving genome-wide significance in the primary screen (LOD > 3.5), a traditional likelihood ratio test (1-df χ²) was performed using the full and reduced models. This represents a test for sex specificity of the QTLs for the phenotype and chromosomal position considered.

RESULTS

Phenotypic heritability

Heritability of femoral neck total vBMD, cortical vBMD, trabecular vBMD, total area, cortical area, ultimate force, and energy to break was 0.74, 0.32, 0.79, 0.32, 0.32, 0.46, and 0.65, respectively.

Genetic loci for body weight

Significant evidence of linkage (p < 0.01) of body weight was observed on Chrs 2 (LOD = 6.8) and 12 (LOD = 7.1) in male and Chrs 7 (LOD = 6.5), 10 (LOD = 5.0), and 12 (LOD = 9.4) in female rats (Fig. 1A). In addition, suggestive linkage (p < 0.05) for body mass was observed on Chrs 10, 13, and 17 in male and on Chrs 6 and 13 in female rats (Fig. 1A; Table 1). All bone phenotypes reported below were adjusted to correct for variation in body weight.

FIG. 1.

Genome-wide linkage analysis for body mass (A), total vBMD (B), cortical vBMD (C), trabecular vBMD (D), and ultimate force (E) for rat Chrs 1–20 (excluding X and Y chromosomes). Multipoint LOD scores plotted on the y axis vs. the relative location along each chromosome on the x axis. The horizontal lines indicate the threshold value for genome-wide significance at p = 0.05 (bottom line) and p = 0.01 (top line). The blue linkage line indicates male, the red linkage line indicates female, and the black linkage line indicates the whole F₂ sample.

Table 1.

Summary of Genome Screen LOD Score for Density, Biomechanics, and Structure of Femoral Neck*

	Density				Biomechanics				Structure
		LOD (position in cM)†				LOD (position in cM)†				LOD (position in cM)†
Chromosome	Phenotype	All	Male	Female	Phenotype	All	Male	Female	Phenotype	All	Male	Female
1	Total vBMD	4.0 (102)			Ultimate force	7.6 (114)		8.3 (114)	Neck width	4.0 (113)		3.5 (113)
	Trabecular vBMD	6.6 (127)	3.6 (108)	4.5 (130)
2	Total vBMD	10.7 (31)	6.3 (33)‡		Ultimate force	3.6 (36)		4.1 (38)‡
	Cortical vBMD	8.9 (32)	6.0 (33)		Energy to break		3.6 (88)
4					Ultimate force	4.5 (96)	3.5 (87)
5									Cortical area	4.4 (3)
6	Total vBMD	4.3 (10)
7	Total vBMD	4.2 (53)							Total area	5.4 (52)
	Cortical vBMD	3.8 (51)							Ip	5.1 (53)
8	Total vBMD	3.5 (82)
9	Total vBMD		4.0 (27)‡
10	Total vBMD	4.9 (44)		4.3 (43)					Total area		3.5 (13)‡
	Cortical vBMD	7.1 (45)	3.5 (38)	5.4 (44)					Ip	4.5 (43)
12	Cortical vBMD	4.4 (30)							Total area	5.2 (43)		3.6 (44)
15					Ultimate force	7.4 (33)	6.6 (36)‡
					Energy to break	6.6 (35)	4.1 (36)
17	Trabecular vBMD	4.7 (18)	5.1 (19)‡						Cortical area	4.5 (21)
18	Total vBMD	7.3 (31)	5.6 (14)‡
	Cortical vBMD	3.9 (32)

* All LOD scores above the significance thresholds for the 95th percentile (LOD = 3.5) are reported.

^† Position in centimorgans is based on the RGD map.

^‡ Represents male sex-specific QTLs.

Genetic loci for femoral neck vBMD

Evidence of significant linkage (p < 0.01) of femoral neck total vBMD was observed on Chrs 2, 6, 10, and 18 in the full sample of 828 F₂ animals (LOD score of 10.7, 4.3, 4.9, and 7.3, respectively; Fig. 1B; Table 1). Significant linkage for total vBMD was observed on Chrs 2 (LOD = 6.3) and 18 (LOD = 5.6) in male and Chr 10 (LOD = 4.3) in female rats. Suggestive linkage (p < 0.05) for total vBMD was observed on Chrs 1, 7, and 8 in the full F₂ sample. Linkage of male-specific total vBMD QTLs was detected on Chrs 2 (sex-specific p = 0.04), 9 (p = 0.00002), and 18 (p = 0.002), and of trabecular vBMD on Chr 17 (p = 0.002). Figure 2 shows the sex-specific linkage for total vBMD on Chrs 2 (A), 9 (B), and 18 (C) and trabecular vBMD on Chr 17 (D).

FIG. 2.

Sex-specific linkage for femoral neck total vBMD on Chr 2 (A), Chr 9 (B), and Chr 18 (C) and trabecular vBMD on Chr 17 (D).

The compartmentalized analysis of neck density showed that cortical and trabecular density at the femoral neck was linked to several QTLs. Evidence of significant linkage (p < 0.01) of femoral neck cortical vBMD was observed on Chrs 2 (LOD = 8.9), 10 (LOD = 7.1), and 12 (LOD = 4.4) in the full sample (Fig. 1C; Table 1). Suggestive linkage (p < 0.05) for cortical vBMD was observed on Chrs 7 and 18 in the full sample and Chr 10 in male rats. Significant linkage for cortical vBMD was observed on Chr 2 (LOD = 6.0) in male and Chr 10 (LOD = 5.4) in female rats. Two major femoral neck trabecular vBMD QTLs (p < 0.01) were detected on Chrs 1 (LOD = 6.6) and 17 (LOD = 4.7) in the full sample of F₂ rats (Fig. 1D; Table 1). Significant linkage (p < 0.01) was detected on Chr 17 (LOD = 5.1) in male and on Chr 1 (LOD = 4.5) in female rats. Suggestive linkage (p < 0.05) for trabecular vBMD was observed on Chr 1 in male F₂.

Several cortical vBMD QTLs were found to be significantly affecting the QTLs for total density. The cortical density QTLs on Chr 2 in male and on Chr 10 in female were found on the same region of Chrs 2 and 10, where total density QTLs is located. On the other hand, we found no significant (p < 0.01) density QTLs affecting both sexes in any chromosomes. However, cortical density QTL on Chr 10 in female found to be suggestively linked to male cortical density. Similarly trabecular density QTL on Chr 1 in female was found to be suggestively linked to male trabecular density.

Genetic loci for femoral neck structure

Significant linkage (p < 0.01) of femur neck cross-sectional total area was observed on Chrs 7 and 10 in the combined sample with LOD score of 5.4 and 5.2, respectively (Table 1). Suggestive linkage (p < 0.05) of total area was detected on Chr 10 in female rats. Significant linkage (p < 0.01) for cortical area was found on Chrs 5 (LOD = 4.4) and 17 (LOD = 4.5) in the full sample. Significant linkage (p < 0.01) of polar moment of inertia (Ip) was observed on Chrs 7 and 10 with a LOD score of 5.1 and 4.5, respectively. In addition, suggestive linkage (p < 0.05) for neck width was observed on Chr 1 both the combined and female samples. Also, a male-specific QTL for total area was found on Chr 9 (sex-specific p = 0.0001). Figure 3A shows the sex-specific linkage for total bone area on Chr 9.

FIG. 3.

Sex-specific linkage for femoral neck total area on Chr 9 (A), ultimate force on Chr 2 (B), and ultimate force on Chr 15 (C).

Genetic loci for femoral neck strength

Significant linkage (p < 0.01) of femur neck ultimate force was observed on Chrs 1, 4, and 15 in the full sample with LOD scores of 7.6, 4.5, and 7.4, respectively (Fig. 1E; Table 1). A locus significantly linked (p < 0.01) to ultimate force was found on Chr 1 (LOD = 8.3) in female rats. Also, a significant male-specific QTL for ultimate force was found on Chr 15 (sex-specific p = 0.0006), and a female-specific ultimate force was detected on Chr 2 (p = 0.003). Figure 3 shows the sex-specific linkage for ultimate force on Chrs 2 (B) and 15 (C). Suggestive linkage (p < 0.05) of ultimate force was detected on Chr 2 in the full sample, on Chr 4 in male rats, and on Chr 2 in female rats. Significant linkage (p < 0.01) was observed for energy to break on Chr 15 (LOD = 6.6) in the combined sample (Table 1). In addition, suggestive linkage (p < 0.05) was observed for energy to break on Chrs 2 and 15 in male rats.

DISCUSSION

We identified several sex-independent and sex-specific chromosomal regions linked to femoral neck density, structure, and strength in inbred COP and DA F₂ rats. Analysis of the full sample of F₂ rats showed significant sex-independent linkage on Chrs 1, 6, 10, and 12 for density, Chrs 5, 7, 10, and 18 for structure, and Chrs 1 and 4 for strength. Subsequent analyses of the full sample by sex detected male-specific QTLs on Chrs 2, 9, and 18 for total vBMD, Chr 17 for trabecular vBMD, Chr 9 for total area, and Chr 15 for ultimate force. A female-specific QTL was found on Chr 2 for ultimate force. In addition, we observed allele-specific effects on several chromosomal regions. The c/c alleles contributed beneficially on Chrs 2, 10, 17, and 18 for neck density, Chrs 7 and 17 for structure, and Chrs 4 and 15 for biomechanics (Table 2). On the other hand, the c/c alleles had detrimental effects on Chrs 1, 6, 7, and 12 for neck density, Chr 10 for structure, and Chr 1 for biomechanics.

Table 2.

Genotypic Mean Values for Femoral Neck Bone Mass, Structure, and Strength QTLs in all F₂ Samples*

			Genotype			Effect of COP alleles	r2	ANOVA p value
Phenotype	Chromosome	Marker	c/c	c/d	d/d
Total vBMD (mg/mm3)	2	D2Rat280	1067.9 ± 3.7	1052.1 ± 2.7	1033.3 ± 3.8	A	0.03	<0.0001
	6	D6Rat46	1039.2 ± 4.0	1051.5 ± 2.7	1062.2 ± 3.7	A	0.01	0.0002
	7	D7Rat23	1037.3 ± 3.9	1055.8 ± 2.6	1056.0 ± 4.1	R	0.02	0.0003
	10	D10Rat162	1062.7 ± 3.9	1054.0 ± 2.7	1035.6 ± 4.0	A	0.02	<0.0001
	18	D18Rat13	1064.8 ± 3.6	1051.6 ± 2.8	1035.1 ± 3.8	A	0.02	<0.0001
Cortical vBMD (mg/mm3)	2	D2Rat280	1259.8 ± 2.7	1247.4 ± 2.0	1236.4 ± 2.8	A	0.03	<0.0001
	10	D10Rat162	1257.2 ± 2.8	1250.4 ± 2.0	1234.4 ± 2.9	A	0.04	<0.0001
	12	D12Rat91	1240.4 ± 2.7	1247.0 ± 2.0	1256.9 ± 2.9	R	0.01	0.0002
Trabecular vBMD (mg/mm3)	1	D1Rat169	735.1 ± 1.8	742.7 ± 1.2	748.3 ± 1.7	A	0.02	<0.0001
	17	D17Rat12	746.8 ± 1.7	743.2 ± 1.2	735.3 ± 1.8	A	0.02	<0.0001
Total area (mm2)	7	D7Rat23	5.5 ± 0.1	5.4 ± 0.1	5.2 ± 0.1	A	0.02	<0.0001
	10	D10Rat162	5.3 ± 0.1	5.3 ± 0.1	5.6 ± 0.1	R	0.02	<0.0001
Cortical area (mm2)	17	D17Rat12	3.5 ± 0.1	3.4 ± 0.1	3.3 ± 0.1	A	0.01	<0.0001
Ip (mm4)	7	D7Rat23	5.4 ± 0.1	5.1 ± 0.1	4.7 ± 0.1	A	0.02	<0.0001
	10	D10Rat162	4.9 ± 0.1	4.9 ± 0.1	5.5 ± 0.1	R	0.02	<0.0001
Ultimate force (N)	1	D1Rat169	102.6 ± 1.0	104.3 ± 0.6	109.9 ± 0.9	R	0.03	<0.0001
	4	D4Rat72	108.6 ± 0.9	105.0 ± 0.7	102.6 ± 0.9	A	0.02	<0.0001
	15	D15Rat117	109.2 ± 1.0	105.9 ± 0.7	101.6 ± 1.0	A	0.02	<0.0001
Energy to break (mJ)	15	D15Rat97	29.7 ± 0.9	26.2 ± 0.6	22.8 ± 0.9	A	0.03	<0.0001

* Values are mean ± SE in all F2 samples.

A, additive; R, recessive; c/c, homozygous for COP alleles; c/d, heterozygous; d/d, homozygous for DA alleles.

Significant QTLs affecting only density without influencing structure or strength were found on Chr 18 (Table 1). Although the QTL linked to total vBMD in male on Chr 18 is separated 17 cM from the QTL linked in all F₂ rats for the same phenotype, they seem to be a single QTL. Remarkably, syntenic regions corresponding to these QTLs regions were also linked to femur neck BMD on 5q32 and 18p11 in humans (Table 3).(13,29) A locus on Chr 5 was found affecting only neck structure without affecting density or strength. On the other hand, QTLs influencing neck strength were discovered on Chrs 4 and 15 with no linkage to either density or structures. The syntenic regions for rat Chr 15 QTLs were linked to hip and neck density and structure on 10q22, 13q21, and 14q23 in humans.(18–20)

Table 3.

Rat QTLs That Overlap With Homologous Human QTLs and Related Phenotypes*

Chromosome	Rat femoral neck phenotypes	Sex	Flanking markers	Human synteny and phenotype
Chr 1	Trabecular vBMD	All samples	D1Rat169–D1Rat82	9q21, Neck BMD(21)
		Female only	D1Rat169–D1Rat82	10q26, Hip BMD(15)
	Ultimate force	All samples	D1Rat169–D1Rat82	11q12–13, Neck BMD(13,22)
		Female only	D1Rat169–D1Rat82
Chr 2	Total vBMD	All samples	D2Rat198–D2Rat280	3q24–26, Pelvic axis length(23)
		Male only	D2Rat198–D2Rat280	5p12–15, Trochanter size(11)
	Cortical vBMD	All samples	D2Rat198–D2Rat280	5q11, Neck axis length(14)
		Male only	D2Rat198–D2Rat280	4q23–26, Neck BMD and axis length(14,24)
Chr 4	Ultimate force	All samples	D4Rat104–D4Rat72
Chr 5	Cortical area	All samples	D5Rat187–D5Rat3	1p35–36, Neck and hip BMD(16,17,25,29)
Chr 6	Total vBMD	All samples	D6Rat46–D6Rat42
Chr 7	Total vBMD	All samples	D7Rat23–D7Rat78	8q22–24, Hip bone size and Ward's BMD(26,28)
	Total area	All samples	D7Rat23–D7Rat78
	Ip	All samples	D7Rat23–D7Rat78
Chr 8	Energy to break	Female only	D8Rat111–D8Rat15	3q21–24, Pelvic axis length(23)
				11q22, Neck BMD(30)
				15q22–25, Neck and hip BMD, neck cross-sectional area(18)
Chr 10	Total vBMD	All samples	D10Rat162–D10Rat124	5q31–35, Neck BMD and axis length(13)
		Female only	D10Rat162–D10Rat124	17p11–13, Neck and hip BMD(10)
	Cortical vBMD	All samples	D10Rat162–D10Rat124
		Female only	D10Rat162–D10Rat124
	Ip	All samples	D10Rat162–D10Rat124
	Total area	All samples	D10Rat162–D10Rat124
Chr 12	Cortical vBMD	All samples	D12Rat91–D12Rat30	7p22–23, Neck BMD(27)
Chr 15	Ultimate force	All samples	D15Rat117–D15Rat97	10q22, Neck cross-sectional area(20)
		Male only	D15Rat117–D15Rat97	13q21, Hip BMD(19)
	Energy to break	All samples	D15Rat117–D15Rat97	14q23, Neck BMD(18)
Chr 17	Trabecular vBMD	All samples	D17Rat80–D17Rat12	9q21–22, Neck BMD(21)
		Male only	D17Rat80–D17Rat12
	Cortical area	All samples	D17Rat80–D17Rat12
Chr 18	Total vBMD	All samples	D18Rat50–D18Rat13	5q32, Neck BMD(13)
		Male only	D18Rat65–D18Rat50	18p11, Neck BMD(29)

* All rat QTLs with LOD scores above the significance thresholods for the 99th percentile (LOD = 4.3) are reported.

Sex-independent pleiotropic QTLs affecting both density and structure were found on Chrs 1, 7, and 10. The trabecular density QTL found on Chr 1 was also linked to neck width (Table 1). The corresponding human syntenic regions on 9q21 and 10q26 were previously linked to neck and hip BMD (Table 3).(15,21) The density QTLs discovered on Chr 7 in the full sample were also linked to neck structure phenotypes. A syntenic region in humans on 8q22–24 for this Chr 7 QTL in rat was linked to hip bone size and Ward's BMD.(26,28) Similarly, the QTLs linked to total and cortical vBMD on Chr 10 in the full sample and females were also linked to neck structure phenotypes. The corresponding human syntenic region on 5q31–35 and 17p11–13 were linked to neck BMD and axis length.(10,13)

Sex-independent pleiotropic QTLs affecting both density and strength were found on Chr 1. Trabecular density QTL on Chr 1 was also linked to ultimate force (Table 1). The syntenic chromosomal regions for these QTLs on Chr 1 in rats were linked to neck and hip BMD on 9q21, 10q26, and 11q12–13 in human (Table 3).(13,15,21,22) The only pleiotropic QTL that affected density, structure, and strength was found on Chr 1 (Table 1).

Sex-specific QTLs analyses showed that male-specific density QTL was found on Chrs 2, 9, and 18 for total vBMD (Fig. 2; Table 1). A male-specific trabecular vBMD was found on Chr 17. A male-specific neck structure QTL was found on Chr 9 for total area, and a male-specific strength QTL was detected on Chr 15 for ultimate force (Fig. 3). On the other hand, a female-specific QTL for ultimate force was found on Chr 2. This same QTL on Chr 2 also significantly linked to both total and cortical vBMD. The QTL linked to energy to break on Chr 2 is at a different location from the QTL for density or strength on the same chromosome, suggesting that this QTL influences bone strength independently. The corresponding human syntenic region for this Chr 2 QTL in rat was extensively linked to neck BMD and structure on 3q24–26, 4q23–26, 5p12–15, and 5q11.(11,14,23,24) The syntenic regions for male-specific ultimate force QTL on Chr 15 in rat coincides with the neck BMD on 14q23,(18) hip BMD on 13q21,(19) and neck cross-sectional area on 10q22(20) in humans (Table 3).

In this study, we analyzed each femoral neck trait independently as we did for our previous studies. However, the underlying correlation among the phenotypes can potentially be used to improve the ability to identify QTLs contributing to multiple traits. Therefore, we measured the correlation among the phenotypes when these traits map to a common QTL region. At the 1% genome-wide significance level, total vBMD and cortical vBMD are both linked to the same QTL regions on Chrs 2 and 10. The correlation coefficient between these two phenotypes is 0.77, supporting the localization of a QTL that affects common variability in both phenotypes. On the other hand, trabecular vBMD and ultimate force are both linked to the same QTL region on Chr 1. The correlation coefficient between these two phenotypes is −0.13. Thus, it would possible that within the QTL region there could be separate genes contributing to these traits, or there could be a common gene that has been identified that would not have been detected based on the correlation structure of these phenotypes.

The statistically significant individual QTLs detected in the combined male and female sample for femoral neck phenotypes in our study were of modest effect, with ANOVA r ² estimates between 1% and 4% (Table 2). We also evaluated the combined effect of the significant QTLs for each phenotype with multiple findings and found that all the QTLs that we report remained significant (p < 0.01) when considered simultaneously. The total variance explained in each phenotype was 11.4% for total vBMD, 8.6% for cortical vBMD, 4.9% for trabecular vBMD, 5.4% for total area, and 9.4% for ultimate force. These estimates are all approximately equal to the sum of the individual QTL effects, indicating that they act independently to influence the femoral neck phenotypes measured.

In a previous study, we identified several genetic loci that affect femoral neck BMD and a single locus on rat Chr 4 that strongly affects femoral neck structure and strength in inbred female Fischer 344 and Lewis rats.(33) However, in this study, we found only a couple of QTLs influencing ultimate force on Chr 4 in COP and DA rats, suggesting that COP and DA rats harbor different genetic variants that influence the femoral neck. Additionally, we found several chromosomal locations in both of Fischer 344 × Lewis and COP × DA inbred models overlapped for femoral neck phenotypes. A QTL for cortical vBMD in Fischer 344 and LEW rats coincides with the QTLs for total vBMD, cortical vBMD, and ultimate force in COP and DA rats on Chr 2. Similarly, a QTL on Chr 4 for cortical vBMD in the F344 × LEW study coincides with the QTL for ultimate force in the COP × DA study. A QTL for total vBMD, trabecular vBMD, and trabecular area on Chr 15 in the F344 and LEW F2 sample coincides with the QTL for ultimate force and energy to break in COP and DA F₂ rats. In addition, QTLs for total and cortical area on Chr 18 in F344 and LEW rats coincide with the QTLs for total and cortical vBMD in COP and DA rats.

In summary, we detected several sex-independent and sex-specific QTLs that contribute to hip fragility in inbred COP and DA rats. Several rat QTLs found in this study are syntenic to multiple human QTLs previously linked to different femoral neck phenotypes in both men and women. Besides these syntenic QTLs observed in this study, some new QTLs were also discovered for different femoral neck phenotypes. These QTLs harbor hundreds of candidate genes contributing genetic determinants of density, structure, and strength at the hip. Further studies involving fine mapping of individual QTLs through the development of congenic lines and cDNA microarray and proteomic analyses will be necessary to identify the key genes contributing to the risk of hip fracture.