Morphological Clues from Multilegged Frogs: Are Retinoids to Blame?

See allHide authors and affiliations

Science  30 Apr 1999:
Vol. 284, Issue 5415, pp. 800-802
DOI: 10.1126/science.284.5415.800


Morphological analysis was performed on multilegged deformed frogs representing five species from 12 different localities in California, Oregon, Arizona, and New York. The pattern of duplicated limbs was consistent with mechanical perturbation by trematode infestation but not with the effects of retinoids.

Reports of amphibians with supernumerary limbs from natural populations have been in the scientific literature for centuries, but the causes remain unclear (1–3). Recent reports of deformed frogs from Minnesota (4) and elsewhere have generated a resurgence of interest. Suggested causes include environmental pollution, ultraviolet irradiation (3, 5), genetic mutation, microbes, parasites, and some other disease (2,6). Most reported malformations involve missing or extra hind limbs (2–4). Missing limbs are difficult to interpret, but supernumerary limbs often contain important morphological clues to the underlying mechanism (2,7).

Supernumerary limbs can be experimentally induced in tadpoles by treatment with retinoids such as retinoic acid (RA) or by mechanical perturbation of developing or regenerating limbs (8–12). The possible involvement of retinoids in recent reports of deformed amphibians is of interest because retinoids interact with steroid hormone-like receptors and include some of the most powerful known human teratogens (9). One common pesticide is a suspected retinoid mimic (13). The effects of retinoids and mechanical perturbation on amphibian limbs allow predictions of specific kinds of morphological abnormalities (2,7, 8–12). By comparing the morphological patterns found in multilegged amphibians from natural populations with these predicted patterns, it should be possible to identify the most parsimonious explanation for this kind of deformity.

RA inhibits intact developing limb buds and causes two kinds of duplications in regenerating limbs: serial duplications along the proximal-distal axis (PD duplications) (Fig. 1, A and B) and, less commonly, mirror-image duplications along the anterior-posterior axis (14). RA biochemically changes the positional values of cells by inducing ectopic gene expression in all three primary axes of the limb bud: anterior-to-posterior, dorsal-to-ventral, and distal-to-proximal (9, 15–17). RA-induced mirror image duplications in the anterior-posterior axis are always double posterior mirror-image duplications (PMIDs) (Fig. 1C); this is true even in RA-induced ectopic limbs from amputated tails of frog tadpoles (18). Likewise, RA proximalizes distal cells (17), which explains the predominance of PD limb duplications in experimentally treated amphibians (19, 20).

Figure 1

Mirror-image limb duplications (A, C, E, G) (digits are numbered from anterior to posterior) and actual mirror-image duplications seen in an RA-treated forelimb of A. mexicanum (B) and field-caught multilegged H. regilla treefrogs (D, F, H). (A and B) PD duplication (Fe = femur, T/F = tibia/fibula, Fe' = serially duplicated femur, T/F' = serially duplicated tibia/fibula, t/f = tibulare/fibulare; arrow indicates amputation site through the tibia/fibula); (C and D) PMID; (E and F) AMID; (G and H) MIT.

Limb duplications can also be produced by any mechanical perturbation that causes physical rearrangement of cells in the developing or regenerating limb (2, 7, 12, 21). Surgical rotation of limb buds in amphibians, for example, results in the outgrowth of several limbs from one original limb bud (7,12). A less dramatic result can be obtained by moving small groups of cells between different regions within the limb bud (22). The effects of such mechanical perturbation include double anterior mirror-image duplications (AMIDs) (Fig. 1E), as well as PMIDs, mirror-image triplications (MITs) (Fig. 1G), and “bony triangles” (12). Notably absent from the range of effects of mechanical perturbations are PD duplications.

The results of a morphological analysis of 391 specimens of cleared and stained multilegged Pacific treefrogs (Hyla regilla) from eight different localities are summarized in Table 1. We have also examined small samples of multilegged specimens of four other anuran species: cascades frog (Rana cascadae) from Oregon, wood frog (Rana sylvatica) from New York, green frog (Rana clamitans) from New York, and leopard frog (Rana pipiens) from Arizona. All these specimens are infested with trematode cysts located in and around the deformities (23).

Table 1

Limb duplications observed in multilegged Pacific treefrogs (H. regilla) examined in this study. PDD = proximal-distal duplication. Each deformity is expressed as percent of total number of mirror-image duplications observed. n = number of specimens examined. Last row is a summary of limb duplications produced by RA treatment of three species of amphibians (10, 11).

View this table:

The supernumerary limbs and related deformities in these samples of frogs are confined to the hind limbs in all but five of the specimens. The total number of hind limbs ranges from 1 to 12. The observed morphological patterns (Table 1) include a wide range of deformities—bony triangles (Fig. 1F) and both kinds of anterior-posterior mirror-image duplications, PMIDs (Fig. 1D) and AMIDs (Fig. 1F), as well as MITs (Fig. 1H). In addition, about 1686 specimens of long-toed salamanders (Ambystoma macrodactylum) of various ages were found with similar limb deformities at the same site that yielded large numbers of multilegged H. regilla(2). Not a single example of a PD duplication was found in any of these specimens. Similar results have recently been obtained from a sample of more than 300 multilegged H. regilla(24). These observed morphologies are not consistent with the effects of retinoids on amphibian limbs, but they are within the range of limb abnormalities expected from mechanical perturbation (Table 1). Recent studies have shown that trematode cyst infestation in the laboratory can induce the outgrowth of supernumerary limbs in amphibians (25). These studies, combined with the evidence presented here, argue against retinoids as an explanation for the apparent increase in reports of this kind of deformity in natural populations. Instead, evidence points to the need for additional research on environmental factors that affect trematode infestation rates.

  • * To whom correspondence should be addressed. E-mail: sessionss{at}


View Abstract

Navigate This Article