Developmental factors influencing bone strength in precocial mammals: An infant pig model

Abstract

Most vertebrates are precocial in locomotion, able to walk and run soon after birth. Precociality requires a bony skeleton of sufficient strength to resist mechanical loading during early locomotor efforts. The aim of this study was to use an animal model—the preterm infant pig—to investigate some of the proximate factors that might determine variation in bone strength in precocial animals. Based on the prior literature, we tested the null predictions that skeletal integrity would be significantly compromised by truncated gestation (i.e., preterm birth) and reduced body mass at birth. We generated a suite of both morphometric measures (tissue mineral density and cross‐sectional geometry) and performance‐related metrics (ability to resist loading, deformation, and fracture during three‐point bending tests) of the appendicular skeleton of preterm and full‐term infant pigs. Results showed that very few measures in our ontogenetic infant pig sample significantly varied with either gestation length or birth mass. Overall, our results contribute to a growing body of literature demonstrating the early functional capacity of the precocial infant musculoskeletal system and suggest that bone strength in perinatal precocial mammals may be robust to the factors shown to compromise skeletal integrity in more altricial taxa.

We collected a suite of morphometric and performance‐related metrics of appendicular bone strength in preterm and full‐term infant pigs, testing the hypothesis that infant skeletal integrity would be significantly compromised by truncated gestation (i.e., preterm birth) and reduced body mass at birth. Results showed that very few measures in our ontogenetic infant pig sample significantly varied with either gestation length or birth mass, suggesting that bone strength in infant precocial mammals may be robust to the factors shown to compromise skeletal integrity in more altricial taxa.

1. INTRODUCTION

Most vertebrates are precocial in locomotion, able to walk and run soon following birth. This is proposed to have evolved as the basal condition, as a response to high levels of predation upon infants and juveniles (Case, 1978; Dial & Carrier, 2012; Grand, 1992; Muir, 2000). Precocial infants require sufficiently well‐developed and integrated musculoskeletal systems at birth, capable of supporting locomotor movement and balance (Muir, 2000). As such, bone strength should be a primary factor leading to the ability of precocial species to walk soon after birth.

The aim of this study was to use an animal model—the preterm infant pig—to investigate some of the proximate factors that might determine variation in bone strength in precocial animals. As precocial mammals, pigs begin moving independently immediately after birth and show mature locomotor kinematics within the first four postnatal hours (Vanden Hole et al., 2017). Pigs are also an established model in which to investigate the impact of gestation length, birth mass, and other factors on various aspects of newborn mammalian physiology (Adjerid et al., 2021; Andersen et al., 2016; Eiby et al., 2013; Mayerl et al., 2019; Nielsen et al., 2018). Prior evidence, mostly arising from clinical research on humans, suggests that primary factors influencing infant bone strength include the duration of prenatal development and body size at birth. During the late human gestation, the musculoskeletal system undergoes substantial developmental acceleration, with approximately 80% of fetal bone mineralization occurring after 35 weeks of gestational age, and the diameter of muscle fibers substantially also increasing during the last five weeks of gestation (Ashmeade et al., 2007; Buttazzoni et al., 2016; Miller, 2003; Schloon et al., 1979). Early termination of these maturational processes via preterm birth can result in a relatively weak musculoskeletal system with compromised bone strength (i.e., “osteopenia of prematurity”; Ashmeade et al., 2007; Bowden et al., 1999; Rack et al., 2012; Ryan, 1998; Tong et al., 2018) Preterm osteopenia can persist into later childhood and even adulthood (Abou Samra et al., 2009; Buttazzoni et al., 2016), though there is debate in the literature about whether lower bone content in preterm individuals results from inadequate mineral stores or reduced biomechanical loading (Abou Samra et al., 2009; Chen et al., 2010; Chinoy et al., 2019; Miller, 2003; Ryan, 1998; Schulzke et al., 2014).

Similarly, low birthweight had been shown to independently be associated with compromised motor development (Boonzaaijer et al., 2021; de Kieviet et al., 2009; Pitcher et al., 2011) and low bone tissue mineral density, robusticity, and strength (Longhi et al., 2015). Even among term‐born human infants, a correlation between increased birthweight to increased bone tissue mineral density and muscle mass persists (Dodds et al., 2012; Longhi et al., 2015; Schlüssel et al., 2010).

Prior research, primarily on humans, therefore suggests that gestation length and body size at birth may influence bone strength in perinatal infants. However, humans—like most primates—are altricial animals, at least in terms of motor development (Grand, 1992). Research on the determinants of bone strength in more precocial animals is currently lacking. Given their relatively longer gestation durations and advanced perinatal motor skills, it is reasonable to hypothesize that the bone strength of precocial infants may be more robust to the developmental factors shown to compromise skeletal integrity in more altricial taxa. In this study, we use a dataset on limb bone tissue mineral density, cross‐sectional geometry, and performance in bending to quantify the bone strength phenotype of infant pigs, testing the following null predictions:

1. Bone strength should be directly correlated with gestation length, such that infant pigs delivered preterm will have reduced bone strength when compared to full‐term infants.

2. Increased birthweight should be directly correlated with higher bone strength.

2. MATERIALS AND METHODS

2.1. Specimen sampling and testing

Testing was performed on limb bones sampled from both full‐term and preterm Sus domesticus ranging in age from 3 to 42 days postnatal and in body mass at death from 0.86 to 10.2 kg (total N = 13 individuals, Table 1 for details). This age range places all of the animals in infant‐juvenile age period, given that domestic pigs typically reach sexual maturity from 4.5 to 6 months postnatal and skeletal maturity at 12–20 months (Cone et al., 2017). Preterm pigs were delivered via Caesarean section 6–7 days early (i.e., 94%–95% full‐term gestation) using standard aseptic technique (see Ballester et al., 2018 for a description of surgical methods and protocols for postnatal care). Bone samples were obtained from individuals that were euthanized for reasons unrelated to this study. There were no pathologies associated with the euthanized individuals. All live animal procedures received prior approval from the NEOMED Institutional Animal Care and Use Committee (protocol #17‐04‐071).

TABLE 1.

Metadata on infant pig sample, organized by gestational status and birth mass.

Identification number	Gestational status	Birth mass (kg)	Sex	Age at death (days)	Final mass (kg)
BM03	Term	1.24	Female	18	2.35
C86	Term	1.70	Male	17	3.00
MP07	Term	1.60	Male	41	10.2
T80	Term	1.25	Female	17	0.98
T85	Term	1.32	Female	17	2.79
TN26	Term	1.49	Female	20	3.3
P66	Preterm	0.42	Male	25	3.76
P69	Preterm	0.59	Male	40	7.48
P72	Preterm	0.88	Female	40	10.2
P75	Preterm	0.82	Male	25	6.4
PN03	Preterm	1.08	Female	3	1.14
PN04	Preterm	0.93	Female	3	0.86
PN12	Preterm	1.15	Female	5	1.31

Whole cadavers were stored frozen (−20°C) prior to tissue collection. Long bones from the forelimbs (humerus, radius, and ulna) and hindlimbs (femur, tibia, and fibula) were excised during gross dissections, swabbed clean of any soft tissue, wrapped in gauze‐soaked PBS, and placed in specimen bags before being stored frozen (−20°C) until material testing. For seven of the 13 individuals, bones were available from both the right and the left sides of the animals; for all others, only right limb bones were available. Our final sample size included 114 individual limb bones.

To measure bone cross‐sectional geometry, whole limb bones were thawed, and the midshaft region was imaged using a microcomputed tomography scanner (vivaCT 75 μCT scanner, Scanco USA, Inc., Southeastern, PA, USA), focusing on a region of interest centered at midshaft ±5% of total bone length. Bones were μCT scanned at 70kVp with a voxel size of 20.5 μm. Stacks of μCT slices (N slices: 115–556) were reconstructed using the Scanco software and imported into ImageJ (Schneider et al., 2012) for analysis using the add‐on plug‐in BoneJ (Doube et al., 2010). Cross‐sectional images were thresholded prior to analysis to include pixels that were sufficiently bright (i.e., dense) enough to qualify as bone tissue via the default automated algorithm in ImageJ (based on the method of Ridler & Calvard, 1978). Automated routines in BoneJ were then used to calculate standard measures of bone cross‐sectional geometry indicative of load resistance under standard beam theory (Figure 1a; Table 2).

FIGURE 1.

Methods used to establish bone strength in infant pigs. (a) Representative cross‐sectional image from a μCT scan of an infant pig femur (individual PN12, age: 5 days, body mass: 1.31 kg). The white scale bar on the bottom right indicates 1 cm. Orthogonal lines drawn over the image indicate the anatomical anterior–posterior (vertical line) and medial‐lateral (horizontal line) axes of the bone, as annotated by the BoneJ software (Doube et al., 2010). (b) Three‐point bending testing of long bone resistance to load. A uniaxial Instron Material Testing System (E3000) was used to load long bones to failure by advancing a cross head that was mounted to a 5kN load cell while bones were securely positioned in a bending jig. See text for additional details of the method. (c) Raw load–displacement data were used to quantify measures of whole‐bone performance in three‐point bending, including rigidity, bending moment at yield load, and bending moment at ultimate load (see Table 2 for definitions of measurements). The exemplar load–displacement curve is from the femur shown in panel A. Note that the short “toe” region of initial loading (prior to the dashed line indicating stiffness) results from the compression of small amounts of soft tissue still adherent to bone at the time of testing.

TABLE 2.

Dependent measures of bone strength.

Measure (abbreviation)	Definition
Tissue mineral density (TMD)	Grams of hydroxyapatite per cubic centimeter of bone; proportional to overall bone stiffness and strength (Currey, 1999; Currey, 2002)
Cortical area (CA)	Total area of mineralized bone tissue within the cross‐sectional region of interest; proportional to bone's average strength in axial loading
Polar section modulus (Zpol)	Quotient of the polar moment of area and the maximum radius of the cross section; proportional to the bone's average strength in bending
Rigidity	Ratio of change in load to change in displacement over the linear (i.e., elastic) region of the load–displacement curve (see Figure 1); measure of a bone's resistance to plastic deformation
Yield moment	Bending moment at the beginning of plastic deformation (i.e., at the end of the elastic period of loading), calculated as the product of yield load and one‐half the span distance between the supports of the bending jig (see Figure 1); measure of the strength of the bone in resisting plastic deformation
Ultimate moment	Maximum sustained bending moment prior to fracture, calculated as the product of ultimate load and one‐half the span distance between the supports of the bending jig (see Figure 1); measure of the strength of the bone in resisting fracture

Bending tests closely follow that of Mossor et al. (2022). Briefly, testing was performed with a uniaxial Instron Material Testing System (E3000) mounted with a 5kN load cell. Whole bones were placed in a 3‐point bending jig that consisted of two rest arms attached to the testing frame base. The rest arms were adjustable to accommodate different bone lengths. Each bone was positioned to have the metaphyses contacting the rest arms with the natural concavity of the bone facing downwards for enhanced stability during testing (Figure 1b). The crosshead anvil was attached to the load cell and was positioned to contact the bone diaphysis at midshaft. The distance between the anvil and rest arms of the jig was measured with digital calipers (Mitutoyo) prior to testing. Limb bones were loaded in bending at a strain rate of 2 mm min⁻¹ via a displacement control protocol consistent with previous studies (Carrier, 1983; Erickson et al., 2002; Young et al., 2014). For expediency, load resistance was tested in right limb bones only.

Bending tests were stopped once the bone fractured. Occasionally, some bones from the youngest individuals did not fracture. In these cases, testing was stopped once the load reached a constant plateau in the postyield range. After testing, bones were recovered, rewrapped in PBS‐soaked gauze, and returned to being stored frozen (−20°C). Records of load (in N) and displacement (in mm) were sampled at 10 Hz and saved after the test as .csv files. Raw bending load and displacement data were then imported into MATLAB (Mathworks, Natick, MA, USA) for analysis using custom‐written code. From load–displacement curves, we calculated several measures of whole‐bone performance in bending (Figure 1c; Table 2).

2.2. Statistical analyses

We used Mann–Whitney U tests to quantify differences in birthweight between full‐term and preterm pigs. We used linear models to test our predictions that metrics of bone strength would be significantly greater in full‐term infants and at higher birthweights. We decided to fit separate sets of models for each predictor variables (i.e., one set of models with birth status as the independent variable and another set of models with birthweight as the independent variable). Additionally, we fit separate models for each bone in the dataset. Fitting more gestalt models with multiple predictor variables or across multiple bones would have (1) reduced available degrees of freedom (already low, due to our limited number of individuals) and (2) may have collapsed across meaningful variation between bones (see Results below). We also checked for correlations between final body mass at death and each dependent variable and entered final body mass as a covariate in models where warranted. We did not include age at death as a variable in our models due to collinearity with final body mass.

Models testing for variation in cross‐sectional geometry (i.e., tissue mineral density, cortical area, and polar section modulus), where we occasionally sampled bones from both the left and the right sides of the same animal, were fit using a mixed‐effects approach to guard against pseudoreplication, specifying individual as a random factor in the analysis. Error term degrees of freedom for these models were adjusted using the Welch–Satterthwaite approximation to mitigate heteroscedasticity. Models testing for variation in metrics of load resistance (i.e., ultimate bending moment, yield bending moment, and stiffness) were fit using ordinary least squares, given that we only sampled bones from the right side of each animal, such that pseudoreplicaton was not an issue.

We made several accommodations to adjust statistical testing for our relatively low sample size of 13 individuals. First, we pooled analyses across sexes. Second, we only tested for main effects, eschewing tests of interactions given low degrees of freedom. Third, we rank‐transformed all variables prior to model fitting, facilitating nonparametric tests of our hypotheses (Conover & Iman, 1981). Fourth, to reduce the likelihood of committing a Type II error (given low statistical power), we (1) accepted tests as significant at p < 0.10 and (2) chose not to carry out any corrections for multiple comparison, as recommended by Nakagawa (2004). We recognize that our accommodations come at the expense of possibility committing a Type I error and that the patterns documented here might best be viewed as trends in need of later confirmation with a more robust statistical sample. Nevertheless, given the uniqueness of our sample, we feel that the current results are still of interest in their ability to speak to preterm mammal skeletal integrity at birth.

All statistical analyses were conducted in the R statistical platform (version 4.2.0) (R Core Team, 2022), including the add‐on packages car (Fox & Weisberg, 2011), dplyr (Wickham et al., 2021), lme4 (Bates et al., 2015), lmerTest (Kuznetsova et al., 2017), nlme (Pinheiro et al., 2019), and tidyr (Wickham & Henry, 2020). Summaries of all statistical tests are provided in Tables S1–S7. All data from this study are available at Magrini et al. (2022a). R scripts used in statistical analyses are available at Magrini et al. (2022b).

3. RESULTS AND DISCUSSION

Our final dataset included six full‐term infant pigs (two male and four female) and seven preterm infants (three male and four female). Birthweight (i.e., mass) across the sample ranged from 0.42 to 1.7 kg. Birthweights were significantly smaller in preterm infants than in full‐term infants (Mann–Whitney U test; p = 0.001; preterm $\overline{X}$ = 0.84 kg, full‐term $\overline{X}$ = 1.43 kg) (see also Adjerid et al., 2021; Andersen et al., 2016).

Several metrics increased with final body mass (i.e., body mass at death) in our cross‐sectional sample of infant pigs (Table S1). In general, these changes reflect increases in absolute bone robusticity and resistance to load as the pigs aged. For instance, CA, Z_pol, ultimate bending moment, and yield bending moment increased with body mass across all bones in the sample (all p ≤ 0.030). Similarly, bending rigidity significantly increased with body size in all bones except the femur and fibula (all p ≤ 0.069). In a microindentation study of ontogenetic changes in the cortical bone of the porcine femur, Rasoulian et al. (2013) found similar trends over a longer time frame, showing that average material stiffness increased over the first 48 postnatal months. Curiously, in the current sample, TMD significantly decreased with size in the ulna and tibia (p = 0.062). Both trends run counter to the general trends observed during mammalian bone ontogeny, where TMD tends to increase with age/size (Currey, 2001; Currey, 2002; Currey & Butler, 1975). It is possible that more standard ontogenetic patterns would have been observed had we extended our sampling into older ages. Prior studies have observed sharp increases in TMD in porcine femora from 1 to 3.5 months of postnatal age (Ambekar et al., 2012; Rasoulian et al., 2013).

Our null prediction was that preterm birth should result in decreased bone strength in infant pigs, as has been previously observed preterm (altricial) humans (Ashmeade et al., 2007; Bowden et al., 1999; Rack et al., 2012; Ryan, 1998; Tong et al., 2018). However, of the 36 individual measures examined (i.e., six metrics of bone strength across six bones), only one variable—femur yield moment—significantly varied between full‐term and preterm pigs in our sample (Table S7; p = 0.029; Figure 2a), with full‐term pigs showing significantly greater yield moments, as predicted. It may be that significant effects of preterm birth would have been observed had we delivered infants earlier in gestation. However, pigs born 6–7 days preterm broadly physiologically equate to a human infant born at 30–32 weeks (i.e., the beginning of the third trimester) (Eiby et al., 2013)—a degree of prematurity that could be expected to compromise bone strength in human infants. Rather, it appears that, in these precocial mammals, the musculoskeletal system is sufficiently developed by 95% of full‐term gestation to suffer few lasting negative effects from early termination from the intrauterine environment. Indeed, prior data of whole‐body composition collected using DEXA suggested that bone density values are similar between preterm and full‐term pigs (Andersen et al., 2016).

FIGURE 2.

Significant patterns of variation in infant pig bone strength. (a) Association between gestation length and femoral yield moment. Controlling for final body mass at death, full‐term pigs were characterized by significantly greater femoral yield moments versus preterm pigs. Both variables were rank transformed prior to analysis. (b) Association between birthweight and fibular TMD. Both variables were rank‐transformed prior to analysis. Symbols as in panel A. (c) Partial regression plot showing the association between birthweight and femoral yield moment, controlling for final mass at death (i.e., residual yield moment versus final body mass is plotted against residual birth mass controlling for final body mass). All variables were rank transformed prior to analysis. Symbols as in panel A.

We predicted that birthweight would be positively associated with all measures of bone strength. Previous studies and meta‐analyses of humans have found that higher mass at birth is significantly correlated with tissue mineral density, bone cross‐sectional robusticity, and even muscle strength (Longhi et al., 2015; Schlüssel et al., 2010). However, we found that birthweight was only significantly associated with two of 36 measures examined: TMD of the fibula and yield moment of the femur (Tables S2 and S6; p ≤ 0.044; Figures 2b,‐c). Both variables significantly increased with birthweight, as predicted. However, birthweight had very little effect on bone strength in the infant pig sample overall. Previous studies have found that birthweight per se is also a poor predictor of functional deficits during pig development (i.e., swallowing performance; Adjerid et al., 2021). Again, it may be that accelerated intrauterine development buffers infant pigs—and perhaps other precocial infants—against the deficits predicated by low birthweight in more altricial species.

There are some limitations to this study. Chiefly, the available sample of pig infants was necessarily small, due to both the difficulty of obtaining preterm pig infant carcasses, and the labor required to dissect and test many individual bones. Although our study lacks a robust sample size of individuals, it is nonetheless groundbreaking in the number of measures we examined and the degree to which sampled across the entire postcranial skeleton. Nevertheless, future studies with larger samples could elaborate on the current work by (1) testing predicted relationships with greater statistical power and (2) testing for interactions among predictor variables (not possible here due to insufficient degrees of freedom). Prior studies of infant pigs have suggested that bone strength may be most dramatically compromised by the interaction of both preterm birth and low body mass (Bæk et al., 2019).

Nonetheless, our results suggest that bone strength in perinatal precocial mammals may be robust to the factors shown to compromise skeletal integrity in more altricial taxa. This is undoubtedly due to the accelerated (or protracted) time line of prenatal development in precocial taxa. For instance, prior studies of prenatal skeletal development in several mammals have shown that secondary centers of ossification are well‐mineralized in precocial infants, compared with the primarily cartilaginous skeletons of more altricial species (da Silva et al., 2020; Gorissen et al., 2016; Watts, 1990). Recent studies of altricial and precocial birds have suggested that the advanced skeletal development of precocial infants can have direct functional consequence for skeletal integrity and performance. Wei and Zhang (2019) found that the femora of newborn Japanese quail (Coturnix coturnix japonica, a precocial ground‐dwelling bird) could withstand greater bending loads than those of older conspecifics, whereas femoral load resistance increased with positive allometry in pigeons, an altricial species more committed to arboreal nesting and powered flight. In a follow‐up simulation study using finite element modeling to investigate the functional consequences of these differences in material properties, the same authors showed that precocial femora incur lower predicted levels of strain than those of altricial pigeons (Wei & Zhang, 2021).

4. CONCLUSIONS

In this study, we used a comprehensive suite of structural‐ and performance‐based measures across the postcranial skeleton to investigate the factors that determine bone strength in an infant pig model of precocial mammals. Based on prior clinical research of (altricial) human development, we predicted that preterm birth and low birthweight would both be associated with compromised bone strength in our infant pig sample. Nevertheless, results consistently showed that neither truncated gestational length nor reduced body size at birth significantly impacted the infant skeletal phenotype, despite significant differences in body weight. Of 72 tested models (i.e., 6 measures × 6 bones × 2 predictor variables), only three tests indicated a significant relationship, though we do note that our predictions were validated in each of these cases. Overall, our results contribute to a growing body of literature demonstrating the early functional capacity of the precocial infant musculoskeletal system.

AUTHOR CONTRIBUTIONS

Samantha Magrini and Angela M. Mossor were involved in acquisition of data, data analysis/interpretation, drafting of the manuscript, critical revision of the manuscript, and approval of the article. Rebecca Z. German was involved in critical revision of the manuscript and approval of the article. Jesse W. Young was involved in concept/design, data analysis/interpretation, drafting of the manuscript, critical revision of the manuscript, and approval of the article.

DATA AVAILABILITY SYTATEMENT

All data from this study are available at https://doi.org/10.6084/m9.figshare.21539205.v3. R scripts used in statistical analyses are available at https://doi.org/10.6084/m9.figshare.21539229.v3.

Supporting information

Table S1‐S7

Magrini, S.H. , Mossor, A.M. , German, R.Z. & Young, J.W. (2023) Developmental factors influencing bone strength in precocial mammals: An infant pig model. Journal of Anatomy, 243, 174–181. Available from: 10.1111/joa.13848