Forelimb musculoskeletal-tendinous growth in frogs

The tendons unite and transmit the strength of the muscles to the bones, allowing movement dexterity, the distribution of the strength of the limbs to the digits, and an improved muscle performance for a wide range of locomotor activities. Tissue differentiation and maturation of the structures involved in locomotion are completed during the juvenile stage; however, few studies have investigated the ontogenetic variation of the musculoskeletal-tendinous system. We ask whether all those integrated tissues and limb structures growth synchronically between them and along with body length. We examined the ontogenetic variation in selected muscles, tendons and bones of the forelimbs in seventy-seven specimens belonging to seven anuran species of different clades and of three age categories, and investigate the relative growth of the forelimb musculoskeletal-tendinous structures throughout ontogeny. Ten muscles and nine tendons and their respective large bones (humerus and radioulna) were removed intact, and their length was measured and analyzed through a multivariate approach of allometry. We obtained an allometry coefficient, which indicates how the coefficient departures from isometry as well as allometric trends. Our data suggest that along with the post-metamorphic ontogeny, muscles tend to elongate proportionally to bone length, with a positive allometric trend. On the contrary, tendons show a negative allometric growth trend. Only two species show different patterns: Rhinella granulosa and Physalaemus biligonigerus, with an isometric and positive growth of muscles and bones, and most tendons being isometric.


INTRODUCTION
The musculotendinous system is particularly active in the general limb movements. The tendons unite and transmit the strength of the muscles to the bones, allowing movement dexterity, and the distribution of the strength of the limbs to the digits (Kardong, 2002). The release of the elastic energy of muscular aponeuroses and ligaments amplifies the power and reduces muscle work (Roberts, 2002;Biewener, 2003). The capacity for the differential jump between frog species is related to the relative amount of musculature of the hindlimb and the use of the energy stored in tendons and ligaments (Emerson, 1978). Further, tendons improve muscle performance for a wide range of locomotor activities (Roberts, 2002).
During ontogeny, the characteristics of muscle architecture and connective tissue vary according to the body length increase, and to the increase of daily activity functional demands. Thus, many of the evolutionary and developmental transformations in structures related to the locomotor function in anurans occur during larval stages and metamorphosis (Ročkova & Roček, 2005;Púgener & Maglia, 2009;Manzano et al., 2013;Fabrezi et al., 2014). Studies on anuran ontogeny reveal that locomotor modes (walking, jumping, and swimming) are achieved before the acquisition of the pelvis-sacral-urostil complex coordination, and the hindlimbs are acting as a unit (Fabrezi et al., 2014). Although metamorphosis has been considered as the period in which the most critical anatomical characteristics for adult locomotion are developed (for example, those related to girdles and limbs; Fabrezi et al., 2014), it is known that tissue differentiation and maturation of the structures involved in locomotion are completed during the juvenile stage (Vera, Ponssa & Abdala, 2015).
Here, we examined the ontogenetic variation in selected muscles, tendons, and bones of the forelimbs in seven anuran species. Our main goal was to investigate the relative growth of the forelimb musculoskeletal-tendinous structures throughout ontogeny. All the ontogenetic stages considered here belong to the fully functional category (Muntz, 1976), implying that all tissues and limb structures are integrated. We ask whether all those integrated tissues and limb structures growth synchronically between them and along with body length, and hypothesize that our data will show a general allometric pattern of increased growth rate with larger body length. The study of the effect of the ontogeny onto the longitudinal growth of the musculoskeletal-tendinous structures of frogs represents a starting point to a more global analysis considering other variables such as volume or cross-sectional areas.

MATERIAL AND METHODS
The right forelimb of seventy-seven specimens belonging to six species encompassing different clades of the anuran phylogeny (Duellman &Trueb, 1994 andPyron &Wiens, 2011) were dissected (Table 1). Then, 10 muscles and 9 tendons, and their respective large bones (humerus and radioulna) ( Table 2) were removed intact, and their length was measured (Fig. 1). The specimens were staged in three estimated categories according to Gosner (1960): metamorphic 46 (2); juvenil (3); and adult (4). Dissections of the anatomical traits were performed between the origin and insertion points under a binocular microscope (Nikon SMZ645), and were measured in mm with digital callipers (±0.01 mm; Mitutoyo CD-15B; Mitutoyo Corp., Kure, Japan). When referring to muscles, abbreviation ''m.'' before muscle names was added; otherwise names refer to tendons associated with those muscles ( Table 1) Table S1. All the examined specimens are deposited in systematic collections, and listed in Table 1.

Statistical analysis
To estimate the scaling of muscles and tendons throughout postnatal ontogeny of the forelimb we performed a multivariate allometric tests based on the generalized allometric equation proposed by Jolicoeur (1963). We performed a principal component analysis (PCA) to obtain the 1st PC eigenvector that expresses the scaling relationships among all variables with the latent size regarded as a latent variable affecting all measured variables  (Giannini, Abdala & Flores, 2004;Giannini et al., 2010). This eigenvector is extracted from a variance-covariance matrix of log 10 -transformed variables and scaled to unity (Jolicoeur, 1963). The significance of multivariate allometry coefficients was tested using a resampling strategy based on jackknife. Each specimen was removed from the sample at a time, generating n pseudovalues to calculate confidence intervals (CIs) for the original coefficients (Giannini, Abdala & Flores, 2004;Flores, Giannini & Abdala, 2006). If the interval excluded an expected value of isometry, the variable was considered positively or negatively allometric. For all multivariate coefficients of allometry, the expected value of isometry, which depends only on the number of variables (p), is calculated as 1/p 0.5 (0.21) for our set of 22 variables). Trimming the largest and smallest m pseudovalues (with m = 1) for each variable may significantly decrease the standard deviations calculated under jack-knife, and allow for more accurate allometric estimations; Giannini, Abdala & Flores, 2004). Here, untrimmed and trimmed calculations are reported, but the chosen results are those with either lower average standard deviation, or lower bias (with the latter defined as the difference between observed and jackknifed allometry coefficient; (Giannini, Abdala & Flores, 2004).

RESULTS
Individual values of the analysed morphological variables are shown in Table S1. Scaling analyses describing ontogenetic growth in length and width of forelimb muscles and tendons in seven frog species are shown in Table 3. Results of allometry multivariate analyses are given in Table 3. We report untrimmed (m = 0) as well as trimmed (m = 1) calculations of confidence intervals, opting for the results with lower average standard  deviation or bias, trimmed with CI 95% which combines the conservative safety of interval and estimated bias in these analyses .  Table 3 Summary of allometric trends in the seven species of anuran for 20 variables investigated. The used symbols are: ''+'' (accelerated with respect to overall size or positive allometric), ''='' (respect to overall size or negative allometric), ''='' (isometric). (HmedT), flexor digitorum communis tendon length (FdcT), flexor carpi ulnaris tendon length (FculT) and flexor carpi radialis tendon length (FcrT) were all negatively allometric (Table S2).

DISCUSSION
The main goal of the present study was to investigate the relative growth patterns of the forelimb musculoskeletal-tendinous system in seven anuran species of three age categories. The general allometric growth patterns inferred from our data indicate that along with the post-metamorphic ontogeny of most studied anuran species, muscles tend to elongate proportionally to bone length, with a positive allometric trend. On the contrary, tendons show a negative allometric growth trend. Only two species show different trends: Rhinella granulosa and Physalaemus biligonigerus, with an isometric and positive growth of muscles and bones, and most tendons being isometric. This trend represents a synchronic growth of all structures, which is an interesting pattern. Overall, the couple antagonist-agonist muscles do not present the same tendency in the surveyed species. For example, the triceps generally grows with positive allometry, and the coracoradialis (biceps) with negative allometry. On the contrary, in the forearm of Rhinella and Physalaemus, agonist and antagonist muscles presented the same trend. Taken together, these data might indicate that the differential growth of the musculoskeletal-tendinous system might be related to the intrinsic nature of each tissue e.g., tendons increase their lengths by apposition of collagen fibril segments in a complex and hierarchical process (Zhang et al., 2005), meanwhile bone growth in length is mainly achieved through the action of chondrocytes in the proliferative and hypertrophic zones Rauch, 2005). Manzano et al. (2013) stressed that the tendon is the last tissue to form during limb ontogeny in anurans and that it needs a fully functional limb to reach complete maturity. Our data shows that this delay in tendon appearance and growth occurs after reaching the fully functional stage, as evidenced by its negative allometric growth. On the contrary, muscles and bones show the same positive allometric trend. This coordinated growth trend contrasts with the previous morphogenetic processes in which muscular differentiation seems to be extremely fast compared to the differentiation of the limb skeletal element (Manzano et al., 2013). Our results on postnatal growth are concordant with the observations made by Huang et al. (2015) in mouse mutants (Splotch delayed (Spd) mice (Vogan et al., 1993)). These authors found that the first stage of tendon development-in which muscles span the zeugopodium anchor to autopodium induced tendons-might be better described through positive allometric muscle growth and a negative allometric tendon elongation, and a subsequent reversal of this trend. Thus, the correct assembly of the musculoskeletal-tendinous complex of a limb segment as unity is regulated by differential growth, in a similar way to that proposed by Eilam (1997). In that study, a heterochronic process was suggested as the critical factor to explain body morphology divergence in several rodent taxa.
Heterochrony is a central process driving morphological diversity in mammals (Ravosa, Meyers & Glander, 1993;Maunz & German, 1997;Richardson et al., 2009), which also seems to modulate musculoskeletal-tendinous growth in anurans. The synchronic muscle-bone growth combined with a negative allometric growth of tendon length results in a segment highly occupied by muscle fibers. On the contrary, a positive allometric growth of the tendons length combined with a synchronic and negative muscle-bone growth would result in a segment highly occupied by tendons. This process could explain, for example, the differences between tendon length of the forearm of a bat or a horse and that of a rat, and would provide a simple mechanism to account for their highly specialized locomotor types.
When the relative growth of the arm and forearm structures of the analyzed anuran species are compared, interesting trends emerge. In the forearm, there is a general trend of a positive bone and muscle allometry, including extensors and flexors, and the already reported delayed tendon growth. In the arm, there is a trend to a positive humerus and triceps allometry, combined with a negative coracoradialis allometry. Strikingly, the coracoradialis tendon presents isometric growth. The described growth of the coracoradialis tendon, combined with negative allometric muscle growth, indicates the presence of a forearm flexor layer with long tendons. Interestingly, the pattern highlighted by Bobbert (2001) as an intriguing design aspect of the human musculoskeletal system (distal muscletendon complexes spanning the distance between origin and insertion, with long tendons and very short muscle fibers) was only recorded for the coracoradialis. The longer tendon compared with the muscle length indicates a segment with less force but faster reaction, which could compensate for the great force with slower reaction indicated by the relative growth of the triceps.
In conclusion, our data indicate that the musculoskeletal-tendinous growth is different than posed in our initial hypothesis: limb bones and muscles tend to develop synchronically, with tendons exhibiting a delayed growth.

Data Availability
The following information was supplied regarding data availability: The raw muscles, bones and body measurements and the number of individuals used per species are available in Table S1.
Tables S2 -Table S7 contain a summary of results of multivariate allometry untrimmed (m = 0) as well as trimmed (m = 1) calculations of confidence intervals, opting for the results with lower average standard deviation or bias, trimmed with CI 95% which combines the conservative safety of interval and estimated bias in these analyses.

Supplemental Information
Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.8618#supplemental-information.