Life in the Slow Lane: The Effect of Reduced Mobility on Tadpole Limb Development


  • Virginia Abdala,

    Corresponding author
    1. Instituto de Herpetología, Fundación Miguel Lillo-CONICET, Fac. de Cs. Naturales (UNT) Miguel Lillo 251, Tucumán, Argentina
    • Instituto de Herpetología, Fundación Miguel Lillo-CONICET, Fac. de Cs. Naturales (UNT) Miguel Lillo 251, 4000 Tucumán, Argentina.
    Search for more papers by this author
  • María Laura Ponssa

    1. CONICET, Instituto de Herpetología, Fundación Miguel Lillo, Miguel Lillo 251, Tucumán, Argentina
    Search for more papers by this author


Movement is thought to be a primary agent eliciting basic responses in the vertebrate body, such as the proper development of the musculoskeletal system. Embryos do not passively await hatching or birth but rather begin active movement very early on in their development. Most studies dealing with embryonic responses to changes in mobility have been performed in chickens or mammals. Herein, we investigate for the first time whether the embryos of organisms that are free-living during development demonstrate the same morphological responses to reduced mobility as embryos that undergo development in controlled environments such as in utero or in a shelled egg. We changed the viscosity of the environment in which free-living anuran tadpoles grow by rearing them in an agar medium. We thus increased the viscosity of the growth medium resulting in a decrease in larval movement. We predicted that a substantial increase in viscosity of the medium in which the larvae were reared would have at least two consequences: (1) a reduction of tadpole mobility and (2) a delayed onset of skeletogenesis thus producing shorter long bones. Our predictions were upheld and tadpoles reared in an agar medium remain immobile longer and showed a delayed onset of skeletogenesis compared with controls. We propose that the developmental responses to the same stimulus are similar throughout tetrapods, regardless of their developmental context (i.e., intrauterine, within an egg, or free-living). Anat Rec, 2012. © 2011 Wiley Periodicals, Inc.

Movement is thought to be a primary agent that elicits basic responses in the vertebrate body, such as the proper development of the musculoskeletal system. Thus, an embryo must move actively to ensure correct development of cartilage, bone, and joints, of muscles, tendons, and ligaments, and of connections in the central nervous system (Müller, 2003). Embryos do not passively await hatching or birth but begin active movement very early on in their development (Müller, 2003). Embryonic movement also is correlated with the appearance and growth of sesamoids (Carter et al., 1998; Sarin et al., 1999; Vickaryous and Olson, 2007; Kim et al., 2009). In addition, the turnover of the extracellular matrix (ECM) of connective tissues is influenced by physical activity, and both collagen synthesis and metalloprotease activity increase with mechanical loading (Kjaer et al., 2006). Mechanical stress induced by movement also modifies the cross-sectional geometry of long bones (Sylvester et al., 2006).

The ability to move can be significantly altered in tetrapods in several ways. Among the most common methods are denervation or neurectomy (Kim et al., 2009), immobilization of limbs in ovo by administration of botulin toxin or decamethonium bromide (DMB; Hosseini and Hogg, 1991; Pitsillides, 2006), and pharmacological paralysis of the embryo by a single injection of the postsynaptic blocking agent, decamethonium iodide (Sullivan, 1966; Hall, 1975; Müller, 2003). Application of these techniques in developing vertebrates produces dramatic effects on the differentiation of tissues, joints, etc. (Pitsillides, 2006).

Embryonic activity is correlated with chemical and physical environmental conditions, and changes in environmental parameters can strongly affect the motility patterns. However, little is known about the developmental effect of environmental influences altering embryonic activity (Müller, 2003). Most studies dealing with the embryonic responses to changes in mobility have been performed in chickens or mammals. Conducting such studies in vertebrates that undergo metamorphosis totally independent of maternal influences may elucidate the role and effects of external environmental factors on motility patterns. Herein, it is investigated whether the larvae of organisms that are free-living during development demonstrate the same morphological responses to stimuli reducing mobility as embryos that undergo development in controlled environments such as in a uterus or a shelled egg.

The viscosity of the environment in which free-living anuran tadpoles usually grow was changed by rearing them in an agar medium to explore the epigenetic effects (sensu Müller, 2003) of reduced motility on their development. Specifically, the relationship between motility and the differentiation of tissues such as tendons and muscles, and skeletogenesis was investigated. The focus was centered on the anatomy of the hind limb because of its importance to the locomotion of anurans. The limbs in younger metamorphosing tadpoles clearly are not used for locomotion, but still move (both passively, with the water currents generated by swimming, and actively, as the tadpoles start twitching, stretching, and kicking their developing limbs, see e.g., Muntz, 1975). It is predicted that a substantial increase in viscosity of the medium in which the larvae were reared would have at least two consequences: (1) A reduction of tadpole mobility; (2) a delayed onset of skeletogenesis that will produce a delayed appearance of the short bones, a loss of sesamoids, and shorter long bones. Hosseini and Hogg (1991) and Pitsillides (2006) stressed that immobilization diminished the rates of chondrocyte proliferation and recruitment. Hence, immobilizated chicks and mammals present cartilage formation markedly reduced resulting in a diminution in long bones length.


To achieve a medium that impairs movement in tadpoles agar in water was dissolved, and tadpoles were reared in this medium. Agar is typically used to modify the motility of microorganisms, especially bacteria (Greenberg and Canale-Parola, 1977; Hewitt, 2010). The mobility is directly related to the concentration of agarose in the medium, so various levels of effective viscosity can be selected, depending on the experimental objectives (Klitorinos et al., 1993).

Sixty tadpoles of Leptodactylus latinasus (Leptodactylidae) and 240 of Pleurodema borellii (Leiuperidae) were collected from temporary ponds in Lules and Yerba Buena (in the vicinity of Tucumán, Argentina) and maintained them under laboratory conditions. During the summers of 2008–2010, one experiment with L. latinasus was performed. In this experiment, tadpoles from Stages 36 to 40–42 were obtained. Four experiments with P. borellii obtaining tadpoles from Stage 26 to metamorphs were also performed.

Each experiment was conducted in the same way. Each of six containers was filled with one liter of water. Nine grams of agar were added to three containers, and the remaining three contained only water. The dissolved oxygen in the agar solution was measured by an oxygen meter (Hach sensION6) and it was 6.02 mg/L; in the water control tanks it was 7.22 mg/L. After successive measures, the difference of the dissolved oxygen between the agar and the water tank never went below 2 mg/L. In a typical fishpond, the critical oxygen concentration threshold is about 2 mg/L (Heargreaves and Tucker, 2002); therefore, the agar solution constituted a physiologically normal oxygenated medium.

Density of the medium was measured with a float-type densimeter. The agar solution had a density of 1.0 g/cm3, because of the fact that water-colloids have the density of the water. At 25°C, water has a viscosity of 0.008 Pa/sec−1. By adding agar, the viscosity of the medium was increased to 0.06 Pa/sec−1, thereby imposing resistance to larval movement. Viscosity was measured with an Ostwald viscometer.

To avoid contamination, excessive solidification and a drastic decrease of dissolved oxygen in the agar medium, the colloid was renewed three times a week. Ten tadpoles were placed in each container and fed with fish pellets ad libitum. Larvae placed in plain water were controls, and some of these control specimens were used to assess the normal anatomy and development of the tadpoles. To obtain ontogenetic series for each species entirely developed in the agar medium with their respective controls, control and treated tadpoles of all stages in Gosner (1960) were fixed in 10% formalin solution. All tadpoles that died during the experiments were also fixed in 10% formalin, and their data incorporated to our database. Ultimately, the ontogenetic series of larvae reared in the agar medium comprised specimens from different experiments. At least three complete developmental series of larvae raised in agar were obtained. Likewise, the morphometric data were recorded from selected samples of all experimental larvae; all tadpoles raised in agar were measured, along with 10 control tadpoles of the same stage. Snout-vent length (SVL), femur length (FeL), tibia length (TiL), and foot length (FL) were considered.

To estimate the mobility of the tadpoles, a digital video of 6 min was recorded for 10 experimental and 10 control containers selected at random. From the videos, one tadpole of each container was selected randomly. The time spent moving in each container was quantified and used as a measure of mobility. The videos were edited and analyzed using the program Windows Movie Maker (2006) version 6.0. Differences in the time spent moving by control and experimental tadpoles (N = 20) was analyzed with an analysis of variance (ANOVA).

Eighty-eight cleared-and-stained (Wassersug, 1976) larval specimens of both L. latinasus and P. borellii between Stages 34 and 46 were examined. Seven juvenile L. latinasus were cleared and stained, along with 18 adult L. latinasus and four adult P. borellii. The radioulna length and FeL of all juvenile and adult cleared-and-stained specimens were measured with digital calipers (Mitutoyo CD-30C and CD-15B; ±0.01 mm) and a stereoscopic microscope (Karl Zeiss Discovery V.8) equipped with an ocular micrometer.

ANOVA was used to test for differences in SVL and an analysis of covariance to test for differences of limb segments between experimental and control tadpoles, using SVL as a co-variable. An F-statistics for Stages 37–46 was calculated. All metric variables were log10-transformed before analyses to meet requirements of normality and homoscedascity (Zar, 1999). Additionally, two categories in relation to the hindlimb position were considered: the femur oriented to the posterior region of the body, or oriented perpendicular to the sagittal axis of the body. A Chi-square test was performed to determine whether the positions of the limbs differed significantly between the experimental and control groups.

Bones of control tadpoles and experimental tadpoles with obvious phenotypical modifications were sectioned to assess the effect of reduced mobility on tissue differentiation. Formalin-fixed specimens were treated with 10% neutral buffered formalin and dehydrated in a graded alcohol series. Serial sections were cuted on an measurement and scientific equipment sledge microtome at 6 μm along the long axis of the limbs, and stained them with alcian blue, hematoxylin-eosin (modified after Totty, 2002). Sections of four experimental and three control specimens of P. borellii, and two experimental and one control specimen of L. latinasus were prepared. Histological samples of L. latinasus used in the study of Ponssa et al. (2010) were also analyzed to provide additional data on the normal histology for comparison.

Throughout the experiment, the larval health was monitored by examining the skin, oral disk, and extremities of the tadpoles (Richards, 1962). In no case, did any tadpole have a malformation that could be associated with either parasites or chemical compounds (Meteyer, 2000; Meteyer et al., 2000).

All the experiments were approved by the Bioethics Committee at the Facultad de Medicina, Universidad Nacional de Tucumán, Argentina.


Tadpoles reared in an agar medium were less mobile than controls (they moved for a shorter of time and over shorter distances; see Supporting Information). Of the 6 min recorded, the selected experimental tadpoles moved for 71.064 ± 72.89 sec (N = 10); meanwhile, controls moved for 220.51 ± 51.56 sec (N = 10) (F = 28.0191; P < 0.0005). Both control and experimental tadpoles tended to move around the perimeters of the round containers (diameter of 20 cm). The distance travelled by the experimental tadpoles never exceeded one-sixth (∼10.4 cm) of the perimeter of the container, whereas the control larvae frequently swam along the entire perimeter (60.2 cm) (Fig. 1). Reduction of movement and frequent immobilization was progressive, becoming more obvious in older tadpoles. However, it is important to note that not all tadpoles were affected equally by the agar medium as some seemed to move in a more normal pattern with respect to the speed and distance of movements. Moreover, even tadpoles affected by the increased viscosity were able to feed regularly.

Figure 1.

Trajectory of the tadpoles in the containers. Blue arrow indicates the trajectory of a control tadpole. Yellow arrow shows the maximum distance that an experimental tadpole moves.

Experimental tadpoles tended to have smaller bodies than the control specimens (Table 1); however, the ANOVA analysis reveals no significant differences in the SVL in any of the developmental stages analyzed (P > 0.05). The limbs of the experimental tadpoles were particularly small (Fig. 2) (Table 1). The ANCOVA analysis revealed significant differences in lengths of limb segments beginning at Stages 43 and 44 (Table 2). The length differences in the feet noted between the experimental and control specimens began at Stage 43 and persisted across the rest of the developmental stages. Moreover, the limbs lie at an abnormal angle with respect to the longitudinal axis of the body, as indicated by characteristic relaxed hind-limb posture that may result from poor muscle tone (Fig. 2). The tails in the experimental specimens were laterally deflected distally, whereas in the controls, they were straight (Fig. 2).

Figure 2.

Tadpoles of P. borellii, Stage 42. The experimental tadpole is to the reader's right. The experimental tadpole shows an abnormal angle with the sagittal axis of the body and tail with distal bending, which rendered a well-recognized relaxed hind-limb posture, lacking muscular tone.

Table 1. Means and SDs of the six variables measured (mm) in the tadpoles of P. borellii of each trial (agar and control) for the quantitative analysis
  1. SVL, snou-vent length; TL, thigh length; TiL, tibia length; FL, feet length; R-U, radius-ulna length; FeL, femur length.

37/3813.03 ± 2.2414.03 ± 1.141.8 ± 0.41.58 ± 0.651.5 ± 0.81.66 ± 0.822.7 ± 1.082.76 ± 0.830.4 ± 00.4 ± 01.1 ± 0.561.73 ± 1.13
41/4214.08 ± 1.1614.5 ± 1.14.42 ± 1.194.47 ± 1.093.9 ± 1.184.37 ± 1.165.99 ± 1.596.5 ± 1.431.4 ± 0.31.8 ± 0.73.54 ± 0.634.4 ± 0.41
43/4413.23 ± 0.7413.67 ± 0.955.52 ± 0.386.1 ± 0.435.28 ± 0.555.95 ± 0.677.61 ± 0.919.07 ± 0.761.73 ± 0.252 ± 03.86 ± 0.114.5 ± 0
4513.77 ± 0.9813.92 ± 0.675.64 ± 0.435.99 ± 0.335.42 ± 0.475.56 ± 0.437.57 ± 0.98.64 ± 0.551.8 ± 02.4 ± 0.563.85 ± 0.214.75 ± 0.35
4614.56 ± 0.9314.65 ± 0.465.87 ± 0.985.77 ± 0.555.47 ± 0.755.72 ± 0.487.71 ± 1.188.57 ± 1.131.5 ± 0.52.75 ± 1.063 ± 14.25 ± 0.35
Table 2. Summary of ANCOVA results of the morphometric values of external morphology measured at different stages in tadpoles of P. borellii raised in agar and in normal conditions
  1. Bold values indicate significant F-values (<0.05).


Three features characterized the dyaphanized experimental specimens, regardless of their age (Table 3): (1) The long bones were shorter. This difference was apparent as the hind limbs of experimental and control P. borellii tadpoles were compared (Fig. 3), and was consistent throughout development. (2) The hind limbs had an abnormal angle to the midline of the body, such that the angle between the leading margin of the femur was obtuse in the experimental specimens, whereas it is 90 degree or less in the controls (Fig. 3). This difference was statistically significant (X = 5; P < 0.05). (3) Morphogenesis of the cartilaginous precursors of the carpals and tarsal bones was delayed (Fig. 4), and sesamoids tended to be absent. Delay in the development of sesamoids was most evident in specimens of L. latinasus. The graciella sesamoid was present in Stage 39 of control specimens in L. latinasus, as well as in all adults, but it was lacking in all experimental specimens. The palmar sesamoid, which was present in the Stage 41 controls and in all adults, was never found in the experimental sample. Because sesamoids were highly variable in their occurrence in P. borellii, no developmental trends were evident.

Figure 3.

Hind-limbs of tadpoles of P. borellii, Stage 42 raised in water (left) and in agar (right). The experimental tadpoles show shorter long bones and abnormal angle of the hind-limbs with the body.

Figure 4.

Fore-limbs (hand and radius-ulna) of tadpoles raised in water (to the reader's left) and in agar (to the reader's right) Stage 41 of P. borellii. The experimental tadpoles shows short bones with delayed development relative to the control tadpoles: the fusion of the elements has not yet occurred.

Table 3. Summary of the developmental differences in the limbs of tadpoles in raised in agar in contrast to those raised under normal conditions
CharacterStage 37Stage 38Stage 39Stage 41Stage 42Stage 43/44Stage 45Stage 46
  1. C, cartilaginous; O, ossified; Om, ossified in the middle area; U, undifferentiated.

Radioulna (mm)
 DiaphysesOmOmOmOmOmOmOmOmOO; shorterOO; shorterO; shorterO; shorterOO; shorter
Femur (mm)
Hind-limb position relative to the axial axis<90<90<90<90<90<90Straight<90Straight/ <90<90Straight/ <90<90<90<90StraightStraight/ <90°

At Stage 41/42 the experimental larvae of P. borellii exhibited completely deformed long bones of the hind limbs. The hind limbs were at this stage composed entirely of highly modified hypertrophied cartilage consisting of irregularly shaped cells with large lacunae that were flatter than normal (Figs. 5 and 6). The small amount of interlacunar matrix formed thin boundaries, resulting in a characteristic netlike appearance. The peripheral layer of the articular zone was thin, no more than one-cell thick (Figs. 5B and 6B). Zones of differentiated chondrocytes typical of anuran long bones (Felisbino and Carvalho, 1999, 2002; Shearman, 2008) were absent. There is no clear articulation area differentiated (Fig. 5B), and the cells of the osteochondral ligaments and the shapes of the epiphyses were highly deformed. All articular cartilages were deformed. Deformation was notable in the menisci, which were composed by hypertrophied cells with irregular boundaries. The shapes of these cells were different especially in the menisci external areas (Fig. 6B).

Figure 5.

Knee-joint histological section of tadpoles raised in water (left) and in agar (right) of P. borellii at Stage 42. A: Control specimen showing the articulation area with cells closely packed, small lacunae, and more intercellular matrix; the lateral cartilage and the osteochondral ligament are clearly formed. B: Experimental specimen showing hyaline cartilage, which consist of large lacunae, almost without interlacular matrix; there is no differentiated elements of the articulation. C: View of the complete hind-limb of P. borellii raised in agar. F: femur; lac: lateral articular cartilage; ocl: ostechondral ligament; TB: tibiafibula.

Figure 6.

Knee-joint histological section of tadpoles raised in water (left) and in agar (right) of P. borellii Stage 41. A: Control specimen showing the articulation area with cells closely packed, small lacunae, and more intercellular matrix; the lateral articular cartilage and the osteochondral ligament are clearly formed. B: Experimental specimen showing very flattened lacunae and nuclei of the chondrocytes.

In the control tadpoles, the articular areas had more matrix and were populated by cells that were more closely packed and that lay in swollen lacunae with homogeneous boundaries (Figs. 5A and 6A). The perichondrium was nearly 10 cells thick and each lateral articular cartilage was clearly formed with its correspondent osteochondral ligament. The epiphyseal shapes were entirely normal (Figs. 5A and 6A).

At Stage 41/42, the experimental tadpoles of L. latinasus exhibited the distal phalangeal epiphyses differentiated, although the interphalangeal joints of the fingers were still not differentiated. Although the interzone was visible, the articular boundaries of the epiphyses were not evident, and there were no signals of joint cavity formation (Fig. 7). When the phenotype of the knee of control tadpoles (Fig. 8A,B) was compared with that of the experimental tadpole in Stage 42 (Fig. 8C) a delayed development phenotpype was observed in the latter. In the experimental tadpoles, the graciella sesamoid was almost undifferentiated from the lateral articular cartilage, with which it shared the same anlage. The area between the femur and tibiofibula was composed of undifferentiated mesenchymatic tissue. The osteochondral ligament was not visible. The fibers of the gracilis major muscle had large nucleoli, and the tendinous tissue was undifferentiated (Fig. 8C). All these traits resembled those characterizing a control, Stage 37 L. latinasus tadpole (Fig. 8A). In a normal, Stage 43 L. latinasus tadpole, the osteochondral ligament was clear; the graciella sesamoid was entirely independent of the lateral articular cartilage, and the muscle fibers had indistinguishable nucleoli and the typical aspect of a mature muscle. A tendinous anlage was distinguishable (Fig. 8B).

Figure 7.

Histological section of the finger of tadpole of L. latinasus raised in agar at Stage 41 showing undifferentiated IP joints. The area with the cell condensation (interzone) of the forming finger skeleton is the site of the future joint. I: interzone.

Figure 8.

Knee-joint histological section of tadpoles raised in water (middle, left) and in agar (rigth) of L. latinasus. A: Control specimen at Stage 37. B: Control specimen at Stage 43. C: Experimental specimen at Stage 42. The specimens at Stage 42 raised in agar (right) are more similar to a control specimen at Stage 37 (left) than at a control specimen at Stage 43 (middle). F: femur; g: graciella; ga: graciella anlage; lac: lateral articular cartilage; m: gracilis major muscle; ocl: osteochondral ligament; ta: tendon anlage; TF: tibiafibula.

In the experimental specimens of Stage 43, the humeral shaft was composed of highly modified, hypertrophied cartilage like that described for the hind limbs of P. borellii. The lateral articular cartilage could not be distinguished, and the muscular fibers did not form muscular packages (Fig. 9). However, the pelvic girdles of experimental tadpoles were completely normal (Fig. 10).

Figure 9.

Shoulder-joint histological section of tadpoles of L. latinasus at Stage 41–42. The experimental tadpole is to the reader's right. The most affected area of the long bones is mainly the proliferation zone, although chondrocytes of the hypertrophied zone are also severely flattened. H: humerus; S: scapula.

Figure 10.

Pelvic-girdle histological section of tadpoles of L. latinasus, Stage 41–42. A: Control specimen. B: Experimental specimen. Both specimen (experimental and control) does not show differences. F: femur; A: acetabulum.


The most obvious effect of larval growth in the agar medium is the reduced mobility of the tadpoles. Agar is an effective way to reduce mobility. It is inexpensive, nontoxic, and easy to prepare. Assays using concentrations of agar that create an environment that is too viscous for organisms to move in have been frequently used in the context of bacteriological studies (Greenberg and Canale-Parola, 1977; Klitorinos et al., 1993; Hewitt, 2010). Because the agar medium density is normal, and the tadpoles move enough to feed normally, it can be inferred that the main factor resulting in reduced mobility of the tadpoles is the high viscosity of the medium. Thus, a fluid with a high-internal resistance to flow was obtained, impeding tadpole movement. Although our data show that the dissolved oxygen in the experimental containers was close to that measured in the control containers, both being normoxic, and that experimental tadpoles can swim to the surface to take oxygen, the possibility that a decrease in dissolved oxygen can be responsible of the described effects remains. Hypoxia has been considered a major stressor (Zuscik et al., 2008) and it has been proposed that tadpoles living in severely hypoxic water suffer from metabolic suppression (or a controlled reduction), which may produce a standby in their development (Wakeman and Ultsch, 1976; Feder and Wassersug, 1984; Ultsch et al., 1999). Our experimental results show that tadpoles raised in agar reach the final developmental stages in a similar time compared with the control ones, and thus no standby is observed. It should be also considered that most of the modifications described in this study are related to the cartilaginous structures and the articular areas. Among the cells best able to survive in hypoxic or nearly anoxic environments are articular chondrocytes. These cells are embedded in an avascular ECM with extremely long diffusion distance from the nourishing arteries (Schipani et al., 2001; Pfander et al., 2007). Moreover, Pinder and Burggren (1983) showed that tadpoles raised in chronic hypoxia suffered no significant change in their blood variables. Thus, if it is assumed that blood values are stable in tadpoles living in hypoxic mediums, and that chondrogenic tissue is rather independent of the oxygen supply, it could be inferred that oxygen deprivation is likely not the main cause to explain the morphological modifications observed in the cartilaginous tissue of the experimental tadpoles.

There are significant differences in limb sizes in experimental and control tadpoles, with the former having smaller limbs than the latter. This trend is particularly evident in the autopodial segments, as was reported by Pitsillides (2006) for chick embryos treated with DMB and pancuronium bromide. The correlation of reduced size with immobilization has been frequently reported (Sullivan, 1966; Murray and Drachman, 1969; Hall, 1975; Hall and Herring, 1990; Kim et al., 2009). Sohn and Kim (2007) found that the denervated limbs of the pipid frog Hymenochirus were slightly shorter than the contralateral, nondenervated limbs in the same individual. Hosseini and Hogg (1991) also reported that chick embryos paralyzed with a DMB solution have shorter tibia owing to the reduction of formation of perichondral bone and cartilage. Reduction in size of long bones seems to be correlated with a reduction of embryonic movement, which leads to a decrease in Type II collagen, aggrecan, and glycosaminoglycan expression, and the consequent reduction of cartilage-matrix synthesis (Newman and Müller, 2005). Because the cartilaginous precursors of the carpal and tarsal elements also are diminished in size, it is thought that the same processes affect long bones and elements of the manus and pes.

Another statistically significant pattern evident in the experimental tadpoles is the abnormal angle of the femur relative to the longitudinal axis of the body. This femur position is a normal part of the ontogeny of the anuran pelvic girdle (Ročková and Roček, 2005; Pomikal et al., 2011; personal observations), but only until Stage 41/42. In the experimental tadpoles, this abnormal angle persists to more advanced developmental stages (e.g., Stage 45/46). An abnormal femoral angle also occurs in tadpoles paralyzed by experimentally altering the thyroid metabolism (Brown et al., 2005). There are no differences in the histological samples of the pelvic girdle and its associated muscular complexes in any of the developmental stages of experimental and control larvae of L. latinasus and P. borellii (Fig. 10). Similarly, the general structure and configuration of pelvic joints are normal in both groups. Thus, the abnormal femoral angle seems to be correlated with less developed muscles resulting from reduced mobility. Muscular hypodevelopment may also account for the slightly relaxed positions of the forelimbs and the distal, lateral deflection of the tail. A relaxed body lacking muscular tone was reported in paralyzed chick embryos by Pitsillides (2006), and is commensurate with the results of Coutinho et al. (2002) in immobilized limb muscles of rats.

In the experimental specimens, the zones of epiphyseal proliferation are the most affected areas of the long bones. However, chondrocytes of the hypertrophied zone also are severely flattened. Quinn et al. (1998) stressed that mechanical loading through an increased static compression is associated with a decreased cell radii in the direction of compression, and this is the case in the cartilaginous cells of the long bones of the experimental tadpoles. The deformation illustrated by Quinn et al. (1998: Fig. 2Cd) resembles what was observed in this work. This cellular deformation suggests that the agar compresses the entire tadpole. Static compression of articular cartilage produced a zone-specific deformation of chondrocytes with the degree of deformation depending on the magnitude and duration of the load (Kääb et al., 2003). Given high-force and prolonged loading, chondrocytic deformation is considerable and correlated with the highly deformed collagen fibers (Kääb et al., 2003). As it was predicted, compression is a collateral effect of life in an agar medium, and it would be not anticipate that animals immobilized by drugs or de-enervation would experience the same kind of cartilage deformation. To our knowledge, such deformations have been not reported as a result of conventional immobilization experiments.

Our data reveal that in tadpoles, joints respond differently to a reduced mobility. For example, the articulation of the femur in the pelvic acetabulum does not vary between control and experimental tadpoles. In contrast, however, the knee, shoulder, and digital joints seem to be severely affected by reduced mobility in some experimental tadpoles. Although the epiphyses of the distal phalanges are differentiated at Stage 42, the interphalangeal joints are undifferentiated as late as Stage 41. Normally, tadpoles at this stage have perfectly formed phalangeal bones with normal articular surfaces as reported by Vickaryous and Olson (2007) and Fabrezi and Goldberg (2009). Likewise, in murine embryos immobilization does not affect all joints in the same way. Despite immobility, knee and finger joints remain intact in the mutant mice (Kahn et al., 2009). Our results support the current hypothesis that movement plays no role in joint specification (Kahn et al., 2009) because in the digits joints of the experimental tadpoles, the interzone formation was not affected, but joint cavity formation was disrupted. Thus, our data support the findings of other authors indicating that immobilization inhibits joint cavity formation without affecting earlier joint specifications (Pitsillides, 2006). The response of mesenchymal stem cells in the embryonic limb to biophysical stimuli may depend upon the location of the cells, suggesting that a complex interaction exists between mechanical forces and location-specific regulatory factors affecting bone and joint development (Nowland et al., 2010).

This study documents delay in the timing of three developmental processes: (1) differentiation of some muscular tissues, as evidenced by their late appearance (e.g., the gracilis major muscle); (2) morphogenesis of the carpal and tarsal bones; and (3) absence of some sesamoids (e.g., graciella and palmar sesamoids). Kim et al. (2009) reported that the lateral fabella, the distal os sesamoide tarsale and the cartilagines plantares disappear in denervated specimens of Hymenochirus boettgeri, and similar results were obtained by Mikic et al. (2000), who found that menisci and sesamoids of the joints were absent in late stage, immobilized chick embryos. Although the origin of sesamoid bones is controversial (Pearson and Davin, 1921a, b; Sarin et al., 1999; Doherty, 2007; Doherty et al., 2010; Jerez et al., 2010; Ponssa et al., 2010), sesamoids are thought to arise as a result of mechanical stress (Carter et al., 1998; Sarin et al., 1999; Sarin and Carter, 2000). Although sesamoids are small, their developmental response to mechanical stress differs from that of carpal and tarsal bones. Sesamoid development seems to respond to a lower threshold of mechanical load, because sesamoids are radically affected (e.g., some disappear) by reduced mobility, whereas for carpal and tarsal elements only the differentiation is delayed. It also could be also inferred that sesamoid development is genetically controlled (Doherty, 2007; Doherty et al., 2010; Ponssa et al., 2010); however, epigenetic stimuli related to movement also seem to be necessary. The anlage of the graciella sesamoid is visible in the knee of experimental larvae; however, control tadpoles at a similar stage have a clearly differentiated sesamoid. Thus, it could be argued that formation of the anlage of the sesamoid is genetically controlled, but extrinsic stimuli drive its differentiation (see also Sarin and Carter, 2000). Muscular movement is thought to stimulate bone formation (e.g., Palacios et al., 1992; Nowland et al., 2010), yet muscle maturation is delayed in experimental tadpoles. According to the sequence proposed by Muntz (1975), tadpoles in Stage 42 should have fully functional and fully differentiated limbs. However, the muscle development involved with the knees of Stage 42 experimental tadpoles is comparable with the premotile stage, when the limb trembles and muscles are just beginning to acquire striated fibrils (Muntz, 1975). Apparently, flaccidity, combined with delayed development, contributes to the absence of stimuli to produce differentiation of some sesamoids.

The initial developmental stages of normal and experimental tadpoles have no observable morphological differences. The observed anomalies in experimental tadpoles begin at Stage 41/42 and include remarkable delays in the differentiation of carpal and tarsal bones, joint cavity formation, the appearance of some muscles, the differentiation of some tendons, and a significant differential growth of the hind limbs. Immobilization experiments on embryonic chicks have demonstrated that the embryos respond differentially to mechanical stimuli, depending upon their developmental stage (Hall, 1977; Pitsillides, 2006). Based on our results, it is inferred that the onset of metamorphosis (Stage 41/42) represents the beginning of a phase in which the consequences of malformed tissues can be first observed. It is possible that intrinsic, genetic factors initiate bone formation and extrinsic, epigenetic factors, such as the mechanical consequence of reduced mobility, have their primary impact on later developmental stages (Pitsillides, 2006). Further studies are needed to elucidate these processes.

Relative to other developmental studies conducted in a context similar to ours, it is proposed that the developmental responses to the same stimulus are similar throughout tetrapods, regardless of their developmental scheme (i.e., intrauterine, within an egg, or free-living). The four main effects that are observed in tadpoles can be recognized in amniotes such as birds (chick embryos: Sullivan, 1966; Murray and Drachman, 1969; Hall, 1975; Hall and Herring, 1990; Quin et al., 1998; Pitsillides, 2006), and mammals (rat embryos: Quinn et al., 1998; Coutinho et al., 2002; Kahn, 2009). Based on these data (Fig. 11), it is predicted that the same pattern should occur in squamates, crocodiles, and turtles. The source of the stimulus also is irrelevant; thus, drug-induced immobilization or denervation should produce the same responses as those resulting from immobilization by an agar medium. Probably, the uniformity of response is primarily the result of the physical constraints imposed on the developing organisms. Thus, in all cases the affected tissues are connective (e.g., cartilages, bone, muscle, etc.), all of which are the same in vertebrates. In fact, many of the factors involved in vertebrate limb outgrowth and shaping (e.g., Dlx, Lmx, Hox, R-fng, Fgf, Shh, and Wnt) originated in ancestral metazoans, in which they also controlled tissue outgrowth and shaping (Newman and Müller, 2005). However, differences in the reactivity of the mesenchymal tissue to produce joints, and the threshold levels to respond to mechanical stress in sesamoid could constitute interesting variations to these highly conservative response patterns.

Figure 11.

Comparison of the known effects of immobilization in tetrapods. Both, amphibians and amniota show the similar main four consequences in this regard. However, more data on more amniote taxa should be gathered.


We are deeply indebted to Linda Trueb, Marissa Fabrezi, Bill Ryerson, Anthony Herrel, Pablo Mirande, Lucinda Backwell and Daniel Buchholz for their help in improving our work. Marisa and Franco Alcaide helped us in the interpretation of the histological data. Hector Maldonado and Cesar Moreno helped us with the viscosity and oxygen measures.