Phenotypic integration of skeletal traits during growth buffers genetic variants affecting the slenderness of femora in inbred mouse strains

Abstract

Compensatory interactions among adult skeletal traits are critical for establishing strength but complicate the search for fracture susceptibility genes by allowing many genetic variants to exist in a population without loss of function. A better understanding of how these interactions arise during growth will provide new insight into genotype-phenotype relationships and the biological controls that establish skeletal strength. We tested the hypothesis that genetic variants affecting growth in width relative to growth in length (slenderness) are coordinated with movement of the inner bone surface and matrix mineralization to match stiffness with weight-bearing loads during postnatal growth. Midshaft femoral morphology and tissue-mineral density were quantified at ages of 1 day and at 4, 8, and 16 weeks for a panel of 20 female AXB/BXA recombinant inbred mouse strains. Path Analyses revealed significant compensatory interactions among outer-surface expansion rate, inner-surface expansion rate, and tissue-mineral density during postnatal growth, indicating that genetic variants affecting bone slenderness were buffered mechanically by the precise regulation of bone surface movements and matrix mineralization. Importantly, the covariation between morphology and mineralization resulted from a heritable constraint limiting the amount of tissue that could be used to construct a functional femur. The functional interactions during growth explained 56-99% of the variability in adult traits and mechanical properties. These functional interactions provide quantitative expectations of how genetic or environmental variants affecting one trait should be compensated by changes in other traits. Variants that impair this process or that cannot be fully compensated are expected to alter skeletal growth leading to underdesigned (weak) or overdesigned (bulky) structures.

Fig. 1

a The partially rendered three-dimensional micro-computed tomography image illustrates the region of analysis, and the midfemoral cross-section illustrates the morphologic traits that were quantified. b Schematic of the hypothesis illustrates how variation in outer surface (subperiosteal) expansion rate is functionally related to marrow expansion rate and the degree of matrix mineralization

Fig. 2

a Midfemoral cross-sections for representative members of the AXB/BXA RI panel are shown at four ages during growth. A/J and B6 are shown for comparison. b Slenderness measured at 4 weeks of age correlated significantly with slenderness measured at 8 and 16 weeks of age. c Slenderness correlated weakly with body weight at all ages (16 weeks shown below)

Fig. 3

Cortical bone area (Ct.Ar) increased linearly with body weight at 4, 8, and 16 weeks of age. Each data point represents the mean value for each RI strain

Fig. 4

Variation in the relationship between cortical area and body weight (Ct.Ar-BW) for a A/J, b B6, and c-f representative RI strains. Individual strains are compared to the average curve derived using mean values for all RI strains. All curves include data at 4, 8, and 16 weeks of age

Fig. 5

a Path Model A tested for functional relationships between outer-surface expansion rate, marrow expansion rate, and mineralization from 0 (day 1) to 4 weeks of age. b Path Model B tested for functional relationships between variation in slenderness relative to body size at 4 weeks of age and the set of adult traits (16 weeks) that contribute to whole-bone stiffness (Stiff)

Fig. 6

a Path Model C generalized the relationships observed in Path Models A and B by treating cortical area as an independent variable. No change in the overall fit of the model was observed. The structural equations derived from the generalized Path Model accurately predicted b cortical thickness and c tissue-mineral density for an independent panel of inbred mouse strains

Fig. 7

The effects of compensatory changes in morphology and mineralization on the development of bone stiffness (EI) were simulated for idealized femora with cylindrical cross-sections. The stiffness for slender femora with no covariation and only morphologic covariation was dramatically reduced compared to the stiffness achieved by slender and robust femora that were fully coadapted