EVOLUTION  BY  LOSS

 

The permanent disuse of any organ imperceptibly weakens and deteriorates it, and progressively diminishes its functional capacity, until finally it disappears.

                                                                              J.B. Lamarck

 I think there can be little doubt that use in our domestic      animals strengthens and enlarges certain parts, and disuse diminishes them; and that such modifications are inherited.

                                                                                                                                          C. Darwin

Loss of structures in the kingdom Animalia is a widespread evolutionary phenomenon. Sometimes misnamed as “regressive evolution”, loss of structures is an adaptive evolutionary response to changed conditions of living, which make an organ or part functionally irrelevant or structurally disadvantageous. Consequently, under new conditions of living, an evolutionary pressure for getting rid of it arises. Disuse of the organ or part is associated with a corresponding change in behavior, which is the prelude to the vestigialization and loss of the structure. Loss of structures is generally a gradual process that starts with its vestigialization but sometimes, by evolutionary temporal standards, it may occur “suddenly”. No changes in genes or genetic information are involved in the best known and best investigated cases of the vestigialization and loss of structures. Experimental evidence from the field of developmental biology suggeststhat the interruption of developmental pathways leading to vestigialization and loss of structures is neurally determined.

 

Loss of Structures

Any organ, part, or other animal structure is necessary for performing specific function(s) but changes in environment or conditions of life may sometimes make some functions unnecessary and the organs used for performing these functions will be used less, not used at all, thus becoming an evolutionary encumbrance. The loss of a structure after it becomes unnecessary, is related to the evolutionary disadvantage that the cost of maintaining that structure represents. Obviously, an evolutionary pressure will arise for losing it.

In the “Origin” Darwin pointed out:  

I believe that disuse has been the main agency; that it has led in successive generations to the gradual reduction of various organs, until they have become rudimentary, - as in the case of the eyes of animals inhabiting dark caverns, and of the wings of birds inhabiting oceanic islands, which have seldom been forced to take flight, and have ultimately lost the power of flying. (Darwin, 1959j)

In “Recapitulation” Darwin reemphasizes: 

Disuse, aided sometimes by natural selection, will often tend to reduce an organ, when it has become useless by changed habits or under changed conditions of life. (Darwin, 1859k)

Animals respond to changes in conditions of living by appropriate changes in behavior, which, sometimes, make certain parts or organs useless or even disadvantageous, as seems to have occurred in cases of evolutionary loss of legs in subterranean animals or loss of lungs in some salamanders and fish. Hence, there is a consensus that loss of function precedes the loss of structure used to perform that function (loss of eyes or sight in cave dwelling animals, of limbs in snakes, of legs in aquatic mammals, of back fins in amphibians, etc). The fact that adaptive changes in behaviors precede changes in morphology, physiology, and life histories is neither surprising nor unpredictable: behavior is the most plastic component of animal phenotype.

As pointed out in chapter 9, animals respond immediately to environmental changes by changing their behavior. The changed behavior may contribute to their survival under changed conditions and to buy time for them until morphological, physiological, and life history characters might arise. So, the presumptive aquatic mammals had to switch from a walking to a swimming behavior before losing their limbs and acquiring general fish morphology; snakes’ ancestors adopted a burrowing style of life before losing their limbs; and birds in predator-free islands had to adopt a walking way of locomotion before evolving their reduced wings and losing the ability to fly.

Loss of structures as a result of disuse under changed conditions of life is a widespread phenomenon in the animal kingdom. J.B.S. Haldane believed that

Probably for every case of progressive evolution in the sense of descendants being more complex in structure and behavior than their ancestors, there have been ten cases of regressive evolution (Fong et al., 1995).

 Vestigialization of Structures

Commonly evolutionary loss of structures is not an “All-or- none” process but the end result of an orderly process of vestigialization of the part or organ.

From a Darwinian view, useless organs would not be lost if they would not be disadvantageous to the species. Charles Darwin believed that natural selection was not involved in vestigialization of organs in metazoans since useless organs would not be selected for or against: 

Rudimentary organs, from being useless, are not regulated by natural selection, and hence are variable. (Darwin, 1872c)

To a similar conclusion, but from a different perspective, comes Carl Gans:

Natural selection acts upon the totality of the organism, hence there should be no a priori reason for assuming that the organ that is vestigialized is indeed the primary target of selection. (Gans, 1975)

However, vestigial organs or parts of an organism sometimes may represent adaptive evolutionary modifications for performing new functions rather than stages in the process of their evolutionary loss. Such may be the case with the reduction of size of the reptilian jaw bones from which the mammalian middle ear ossicles (malleus, incus, and stapes) evolved. Indeed, the fact that the reduction of the reptilian jaw bones took place simultaneously with the process of their aboral displacement suggests that the reduction was a stage in the process of their adaptation for a new function as ear ossicles in mammals, rather than a stage in the process of their loss. 

Vestigialization of Limbs in Squamates 

Figure 15.1. Summary of the estimated number of origins of each of the ecomorphs of limb-reduced squamates in each major continental zoogeographic region (but also including Madagascar and the West Indies). Although the number of origins of each morph is similar on each continental region, different squamate clades evolve these morphs in different regions (e.g., all origins of the short-tailed burrowing morph occur within gymnophthalmids in South America but occur in lygosomine scincids and pygopodids in Australia)… The asterisk indicates that there is only a single lineage of the long-tailed morph (anguine anguids) which has dispersed among Asia, Europe, and North America. The number of origins of the short-tailed morph should be considered minimal estimates; these numbers may be considerably higher in some regions (e.g., Asia, Australia) as more detailed phylogenies for scincid lizards become available. Origins of the geographically widespread snakes and amphisbaenians are not included. Middle America seemingly lacks independent origins of either morph and is only questionably considered a separate biogeographic region, and is therefore not shown separately (From Wiens et al., 2006).

Simplification of the Brain and

Morphology in Plethodontid Salamanders

 

Most biologists used to believe, and still do, in the existence of an evolutionary trend that ascending the evolutionary tree, the morphology, the nervous system and the genome of metazoans become more complex. However, this rule does not seem to apply, at least to amphibians. Plethodontid salamanders of the tribe Bolitoglossini, comprising 180 species, display the highest degree of secondary simplification of the nervous system and of the body in general (Roth et al., 1993).

In many respects their brain is less differentiated than the brain of all salamander species and the brain of most fish, including lampreys and hagfish, a fact that is at variance with the position of the group in the evolutionary tree.

They have the lowest number of neurons per volume unit and the lowest level of differentiation and migration in the nervous system; they have also lost lungs, most of the larval stages in the egg, as well as an aquatic larval stage (Roth et al., 1992).

The simplification of the structure of the plethodontid bolitoglossini may be well related to the secondary simplification of their nervous system, which is in line with the basic tenet of the epigenetic theory on the nervous system as the controller of the development and evolution of metazoan morphology. Both the brain simplification and morphological simplification in bolitoglossini are secondary rather than plesiomorphic. This fact poses another extremely difficult problem to the neoDarwinian paradigm, for the neoDarwinian prediction that morphological evolution is related to evolution of genes and the number of genes is obviously contradicted by the “paradox” that the secondary simplification of the morphology in amphibians is associated with a huge expansion of their genome. It is questioned:

 

Why should these evolutionarily successful vertebrates have reduced the complexity of their brains and sense organs, when the trend has been toward increased complexity in other lineages? (Roth et al., 1992)

 

Without elaborating, attempts are made to relate this paradox with paedomorphosis:

 

Paedomorphosis commonly involves different degrees of retardation, reduction or absence of traits in otherwise fully developed organisms, as compared with phylogenetic outgroups. (Roth et al., 1992)

 

But paedomorphosis is a still poorly understood phenomenon, which needs itself a scientific explanation before being used for explaining why simplification of brain and morphology of these salamanders occurred. While it is clear that a correlation between the simplification of the nervous system, the morphology, and paedomorphosis exists, there is no evidence that the latter is the cause of the simplification of the nervous system in the Bolitoglossini group of plethodontid salamanders. The reverse may also be true. As a stage of metamorphosis, pedomorphosis is under control of the the CNS, not the other way around.

 

 

Loss of Animal Structures in Nature

 

One of the major chapters of the evolutionary loss of structures in animals, the so-called regressive evolution, is related to their switching to parasitic forms of living. Transition to parasitic mode of living often involved loss of whole organs or even systems of organs as is the case, e.g., with many parasitic worms that have lost their limbs, eyes, digestive tract and respiratory organs. However, the evolutionary loss of phenotypic traits in parasites is out of the scope of the present work.

 

 

Loss of Wings in Insects

 

Despite the clear evolutionary advantages of wings, loss of wings has occurred thousands of times in insects and wingless insects represent about 5% of the extant insect species (Whiting et al., 2003). The fact that GRNs (gene regulatory networks) for wing development are conserved in wingless insects for more than 300 million years suggests that the cause of the loss of wings has to do with an epigenetic inactivation of these GRNs in wingless insects rather than with any changes in genes which these networks consist of. We know for a fact that, most commonly, activation/inactivation of GRNs in insects is under hormonal control. And we also know that the basic hormones involved in the development and loss of wings in insects, ecdysteroids and JH (juvenile hormone), are under strict cerebral control via neuropeptides, PTTH (prothoracic hormone) and allatostatins/allatotropins, respectively.

 

Loss of Wings in Phasmids

 

An impressive case of loss of wings in insects is that of phasmids (order Phasmatodea). Although ancestral conditions of the order has been wingless, later in their evolution they independently developed wings on as many as 4 cases. Now, about 60% of the 3000 species of this group of 3 families and ~500 genera, have reduced wings or are wingless (Whiting et al., 2003).

Evolution of wings in phasmids is not a de novo event, but a reversion to the lost ancestral phasmid wings (Whiting et al., 2003). Having lost wings early in their phylogeny, later phasmids gained wings and again became wingless in a number of times. Now, most of them are wingless species that have lost fully (both fore- and hind wings) or partially (hind wings) their wings (Whiting et al., 2003):

 

Entomologists have long assumed that re-evolution of wings in apterous lineages was impossible, because functional wings require complex interactions among multiple structures, and the associated genes would be free to accumulate mutations in wingless lineages, effectively blocking the path for any future wing reacquisition. (Whiting et al., 2003)

 

Re-evolution of wings in this group would be unpredictable from the neoDarwinian view holding that evolution of morphology results from selection of randomly occurring mutations in genes, changes in allele frequencies, or genetic recombinations. There is no evidence that evolution and loss of flight in insects is related to changes affecting the function of any of the genes involved in wing development. Moreover, it has been repeatedly observed that even after the loss of the wings, not only GRNs and genes relevant to wing development but even the flight muscles and neural circuits determining the flight behavior are conserved in wingless insects:

 

Studies of flight motor patterns in flying and non-flying phasmids indicate that the non-flying phasmids have retained the neural structures and basic functional circuitry required for flight, as indicated by flight-specific neural activity in thoracic muscles, demonstrating that the loss of wings does not correlate with the loss of flight musculature and innervation… Our results support the hypothesis that the developmental pathway for wing formation evolved only once in insect diversification, but that wings evolved many times by silencing and re-expressing this pathway in different lineages during insect evolution. (Whiting et al., 2003)

 

That the loss or reversal of wings in insects involves no genetic changes in genes is also corroborated by the polyphenism observed in some insects, such as Lopaphus, which exhibit both partially winged and wingless condition in individuals of the same genotype (Whiting et al., 2003). Evidently, the fact that no changes in genes or genetic information have contributed to  the loss of wings in phasmids precludes any  neoDarwinian explanation of the evolutionary phenomenon.

 

Loss of the Gasbladder in Fish 

Most of the fish families that have no gasbladder live in the bottom of the water or in the deep sea where buoyancy is not needed (McCune and Carlson, 2004). The function of gasbladder in fish, however, is not restricted to buoyancy, but it is also used for hearing and sound production (McCune and Carlson, 2004).

Gasbladder loss occurred in 9 of the 14 extant teleost superorders. In 79 of the 425 extant families of teleost fish, the gasbladder is absent in at least one species. In most families there was either a single species or genus lacking a gasbladder or the bladder was absent in all members of the family. In 25 families there were multiple species in at least two genera that lacked the gas bladder. Most taxa (60 families) that lack a gas bladder are either benthic (live on or in the bottom) or deep sea fish. The 19 families that are neither benthic nor deep sea are either nested within clades that are entirely bladderless or have lifestyles that are not compatible with having a gasbladder. The loss of gassbladder in fish seems not to be related to gene mutations: 

All the bladderless mutations in this study are lethals. Thus, the mutations we have identified are not the actual mutations that have led to loss of the gas bladder in living teleosts. (McCune and Carlson, 2004)

There are teleost fish species, such as tuna, which sometimes, although sharing a common genotype, have reduced gasbladder or even lack it completely (McCune and Carlson, 2004). This very interesting fact adds to the empirical evidence that reduction of gasbladder and even the lack of it are not related to changes in genes or genetic information. It might be argued that this phenomenon in tuna may be a form of penetrance, but with penetrance being a descriptive term, i.e. that needs itself an explanation, the statement that the tuna phenomenon is a case of penetrance is no more than a tautology and, as such, cannot explain anything.

From a neoDarwinian view, unexplainable is not only the reduction or lack of gasbladder in a proportion of individuals that share the same genotype with the rest of population that have gasbladder. Because of the number of genes that have to “favorably” change and the long evolutionary time necessary for evolution of an organ such as the gasbladder, the repeated independent evolution and loss of this organ is unexplainable from that view.

 

 

Loss of the Pelvic Fin in Fish

 

Loss of the pelvic fin is the most frequent loss of a structure in fish. According to Nelson, 92 families of teleostean fish have lost the pelvic fin independently about 50 times, excluding multiple independent losses within these families (McCune and Carlson, 2004).

 

Loss of Teeth in Birds

 

The loss of dentition in birds is one of the most enigmatic and one of the major losses of organs (affecting the whole class Aves), that has occurred in the course of vertebrate evolution.  The loss is thought to have occurred about 60-80 million years ago (Chen et al., 2000; Mitsiadis et al., 2003).

 

NeoDarwinian Explanation

 

From a theoretical point of view, supporters of the neoDarwinian paradigm have failed to show what would be the selective advantages of losing teeth or why the natural selection would eliminate teethed individuals. The argument that weighty dentition would be disadvantageous for flying animals, is not convincing, for dentition has not been disadvantageous to other flying animals such as bats.

The standard neoDarwinian interpretation of the mechanism of loss of teeth in birds would be that it is result of accumulation of relevant mutations in odontogenic genes that, under the action of natural selection, led to the inactivation of the gene regulatory network for odontogenesis. However, no evidence has been presented that would suggest that such mutations occurred in genes involved in GRNs (gene regulatory networks) for tooth formation in birds or that GRNs for tooth development are absent or  nonfunctional in Aves.

Contrary to the neoDarwinian prediction on gene mutations being the cause of the loss of teeth, solid experimental evidence shows that  presently, ~80 million years after the loss of dentition in birds, embryonic bird epithelium is capable of forming teeth when supplied with neural crest cells from the mouse midbrain. This fact unequivocally shows that even during such an evolutionarily long period of time, despite unavoidable mutations that might have been accumulated, odontogenic genes and gene regulatory networks are fully functional in these teethless animals.

Marshall et al. (1994) have argued that gene function may be lost not only via mutations in the gene itself but by a mutation in the circuitry that controls its expression. However, this implies a change in the nucleotide sequence of another gene. But this contradicts their own estimation that, due to accumulation of spontaneous mutations, even this hypothetical unidentified gene would lose its regulatory function after such a long period of time. According to their estimation, a silenced gene might maintain the ancestral functional state for periods of time not longer than 0.5 to 6 million years, when the recent experimental evidence shows that all the genes necessary for the development of teeth in birds are still functional, presently, ~80 million years after being silenced in birds .

The failure of all the neoDarwinian arguments to rationalize the genetic hypothesis of the loss of teeth as a result of gene mutations in birds emphasizes the need to search for a possible epigenetic explanation.

 

EpigeneticExplanation

 

Teeth development results from interactions between oral epithelium and underlying ectomesenchyme cells of cranial neural crest origin. It has been observed that although birds have lost dentition, during ontogeny they go through initial stages of odontogenesis, similar to those observed during mammal tooth development, suggesting that they have retained the ancestral odontogenetic signaling pathway. Experimental evidence shows that they do not form teeth because somehow they are prevented from expressing Bmp4 and, hence genes Msx1 and Msx2 (Chen et al., 2000). Moreover, genes that are expressed during odontogenic activity of neural crest cells (Pax9, Msx1, Barx1, MK, etc.) in mammals such as mice, are functionally intact in Aves; only the ability of avian neural crest cells to express these genes is lost.

In a classic experiment of homotopic transplantation of the murine neural tube from the midbrain into chick embryos, it was observed that migration of the donor (mouse) neural crest cells to the mandibular and maxillar proceses of the developing chick embryos leads to formation of tooth-like germ structures in the latter. This clearly suggests not only that the murine neural crest cells are in possession of inducers of teeth formation in chick epithelial cells but also that the latter are in possession of functionally unchanged odontogenic genes. Indeed, expression of Msx1, Barx1, and MK genes by the transplanted murine neural crest cells induces expression of BMP4, Shh, and FGF8 and odontogenesis in chick epithelial cells of mandibular and maxillar processes (Mitsiadis et al., 2003; figure 14.44, in the previous chapter.).

All this suggests that the loss of dentition in Aves is result of a nongenetic, epigenetic-regulatory loss of ability of their neural crest cells to secrete signaling molecules necessary for chick epithelium to initiate odontogenesis.

What is the cause of the loss of the ability of the chick neural crest cells to induce tooth formation? There is no experimental data to give a scientifically reliable answer to this question. However, at a theortetical level, one might argue that the fact that it is the neural tube/CNS that provides neural crest cells with information on “what to do” in the sites of their migration, suggests that the chick neural tube/CNS ceased to provide odontogenic information to these cells. Since no changes in key odontogenic genes are involved it may be safely inferred that the change is determined by an epigenetic change in the chick neural tube/CNS.

 

Loss of Tetrapod Limbs

Loss of limbs has occurred in three of four tetrapod groups (amphibians, reptiles, and mammals). It represents one of the most extreme morphological changes in the history of tetrapods (Lande, 1978) and has been associated especially with elongation of the body and increase in the number of vertebrae.

It is believed that the loss of limbs occurred in response to new ways of locomotion as a result of a change in the life style of tetrapods. This seems to have been the case with transition of reptiles to a burrowing life style and reptant locomotion, which made their limbs useless. The loss and reduction in size of limbs in tetrapods was thought to have been a gradual process of sequential loss of limb components in the reverse order (distal-to-proximal) of their formation during the individual development (proximal-to-distal).

Latter studies, however, have shown that often evolutionary processes of body elongation, reduction of limb size, and reduction of digits, occurred almost simultaneously (Wiens and Slingluff 2001).

Loss of limbs has occurred repeatedly and independently in a large number of reptile species. Generally, forelimbs and pectoral girdle are lost before the hindlimbs and pelvic girdle. Loss of limbs in reptiles is associated with (Lande, 1978; Cohn and Tickle, 1999), and preceded by (Lande, 1978), body elongation. In turn, body elongation results from two different mechanisms: trunk elongation, related to subterranean dwelling, and tail elongation related to surface dwelling (Wiens and Slingluff, 2001) There is no consensus on the rates of evolution of limblessness in snakes. Two contrasting hypotheses have been proposed: one positing sudden loss of limbs (Cohn and Tickle, 1999) and the other stating that the loss has been gradual (Wiens and Slingluff, 2001).

Loss of Limbs in Amphibians and Reptiles 

Total loss of both pairs of limbs occurred several times in amphibians and reptiles. About 150 amphibian species of Caeciilidae family of the monophyletic order Gymnophiona (Apoda), are limbless. They populate tropical forests. The loss of limbs in this group is believed to have resulted from transition to subterranean mode of living (Summers and O’Reilly, 1997) and a number of caecilians presently are fossorial rather than aquatic species.

Among reptiles only the superorder Squamata has limbless species. Snakes are always functionally limbless. Four lizard families consist mainly of species with vestigialized limbs and three of the four families of reptiles of the suborder Amphisbaenia have extremely reduced limbs (Gans, 1975).

There is no evidence that the loss or  reduction of limbs in amphibians and reptiles is related with changes in any relevant genes.  

Loss of Limbs in Snakes 

Between 2700 (Coates and Ruta, 2000) and 3000 (Wiens et al., 2006) extant snake species are presently known. In a study on 261 species of squamate reptiles it was observed that snake-like body (short-tailed burrowers and long-tailed surface-dwellers) form evolved independently 25 times (Wiens et al.,  2006).

Snakes evolved from limbed terrestrial ancestors (Greene and Cundall, 2000; Tchernov et al., 2000). The possibility of a reverse, aquatic-to-terrestrial, origin of snakes evolving from marine voracious reptiles has also been suggested (Coates and Ruta, 2000), but a terrestrial-to-marine transition is more likely as a common theme of tetrapods switching to the aquatic mode of living (Greene and Cundall, 2000).

Most biologists consider the loss of limbs in snakes to be a result of adaptation to a burrowing or surface-dwelling style of life. Phylogenetic conclusions contradict the widely held “subterranean” theory of snake origins, and instead imply that burrowing snakes (scolecophidians and anilioids) acquired their fossorial adaptations after the evolution of the snake body form and jaw apparatus in a large aquatic or (surface-active) terrestrial ancestor (Scanlon and Lee, 2000). As pointed out by Gans (1975), the potential adaptive value of the transition to snake-like morphology has not been well established, but it is generally assumed that snakelike body shape facilitates locomotion underground and in dense grass (Wiens et al., 2006): 

Limb reduction proceeds by the loss of elements in a roughly distal to proximal sequence. The distal-proximal sequence of limb bone loss has also apparently occurred in aquatic mammals, as evidenced by the living and fossil forms, which retain only a femur or its proximal end, and the Greenland right-whale, Balaena mysticetus, which has internal remnants of a femur and tibia (Struthers, 1881). Living and extinct flightless birds show a similar pattern of reduction in wing size followed by loss of distal elements. (Lande, 1978) 

Elaborating on his idea that loss of limbs might have been a result of the body elongation, Gans argues that degeneration or loss of limbs was a secondary result rather than the direct response to a primary selective pressure. Thus, limb reduction followed, and was probably produced by, selective pressures established after bodily elongation had occurred: 

It is impossible at this moment to determine whether elongation was indeed primarily for the passage of crevices and perhaps for the capacity to traverse environments such as tuft grasses. It may also have been associated with lateral undulation, a generally more effective propulsive system in terms of energy cost than is tetrapody. (Gans, 1975) 

The last idea that a switch to an undulatory locomotion might have given rise to both body elongation and loss of limbs shows that Gans was way ahead of the biological thought of his time. He suggests the preeminence of the change in the locomotion behavior as a cause for both body elongation and the loss of limbs observed in so many species of amphibians, reptiles, and mammals.

The fact that tail loss, to various extents (from two thirds of its length to total loss), was observed in 58% of individuals of a large population of tiger snakes in Western Australia (Aubret et al., 2005) suggests that the species is in the process of losing the tail. Whether you call this a form of developmental polymorphism or even penetrance, is of little importance. What scientifically matters in this case is the fact that certain proportions of individuals of the same genotype, under the same environmental conditions, display different phenotypes. This clearly contradicts the basic tenet of the neoDarwinian paradigm that evolution of limblessness, as any other evolutionary change, requires accumulation of favorable mutations in relevant genes. Logically, this suggests that a nongenetic mechanism is inducing the loss of limbs in this snake species.

Recent studies have shown that multiple independently occurring loss/reduction of limbs in lizards of the Anguidae family have taken place as gradual, not as suddenly occurring processes (Wiens and Slingluff, 2001). 

Loss of Forelimbs in Pythons 

Pythons have no forelimbs but they develop reduced hind limbs. Anatomical transformations in python limbs have been sudden rather than gradual and are related to the progressive expansion of Hox gene expression patterns (Cohn and Tickle, 1999).

The loss of forelimbs in pythons is believed to be related to an anterior expansion of expression pattern of Hox genes. Hind limb buds are initiated in pythons but the ZPA (zone of polarizing activity) does not develop and the ectoderm does not form an AER (apical ectodermal ridge) in the region where the limb bud emerges in tetrapods, even though all the signaling genes responsible for their development are present. This is believed to be caused by changes in mesodermal Hox gene expression:

Progressive expansion of Hox gene expression domains along the body axis can account for the major morphological transitions in snake evolution. (Cohn and Tickle, 1999) 

In contrast to the forelimbs, pythons develop hind limb buds and rudimentary hind limbs with truncated pelvic girdle and femur. However, they are unable to express Shh because they have no AER (apical ectodermal ridge) and they do not express in their ectoderm AER-related genes, Dlx (Distal-less), Fgf2 and Msx (Cohn and Tickle, 1999). This does not mean that these genes are not functional for they are expressed in other organs of the python embryo. Remember that even in the presence of the AER and FGF8, Shh is not expressed if RA (retinoic acid) or RAR (retinoic acid receptor) is absent.

The python hindlimb mesenchyme can be experimentally induced to form an AER and express Shh by application of FGF. The fact that the python mesenchyme from the hind limb is functional when grafted to a chick embryo wing  (Cohn and Tickle, 1999) proves beyond doubt that the python hind limb bud is in possession of all genes involved in the initial development of tetrapod limbs, and the loss of hind limbs in pythons is not related to any change in the function of limb-inducing genes; all the genes (especially genes coding for transcription factors) and gene products essential for limb formation are present and functional.

It is known that expression of Hox genes is regulated by RA (retinoic acid). A glimpse at the embryonic expression domain of the HoxC-6 and HoxC-8 genes in chicks and python embryos shows that whereas in chicks expression of these genes takes place along the trunk with interruptions at the levels of the fore- and hind limbs, in python embryos these genes are expressed uninterruptedly anteriorly but are not expressed at the level of hind limbs (figure 15.2).  

 

Figure 15.2. The distribution of HoxC-8 and HoxC-6 in a limbed tetrapod (embryos of the common fowl) and a snake (embryos of the python). The expression boundaries are extended slightly more posteriorly and much more anteriorly in python than in chick embryos.

Abbreviations: FL, forelimb buds; HL, hindlimb buds (From Bejder and Hall, 2002).

 

Expression of Hoxc8 is under control of RA pathway and in the case of limb bud development the pattern and sites of HoxC-8 expression are negatively controlled by RA secreted by brachial spinal nerves that innervate the limb bud. This suggests an antagonistic relationship between the Hox gene expression along the trunk and limb development.

From a neoDarwinian view, it has been argued that small changes in the sequences of HoxC-8 gene enhacers between the mice and chicks may be the cause of differences in the region of the development of limb buds in the embryos of two species (Belting et al., 1998), but such changes in sequences will unavoidably accumulate over the time if they do not lead to the loss of gene function. Investigators have not concluded whether the changes in patterns of expression of HoxC-8 gene in mice and chicks are cause of the pattern of expression or are a normal result of the long divergent evolution of these species that did not affect the expression. It is this the reason why investigators themselves cautioned: “Additional experiments will be required to determine the specificity of nucleotide changes in the regulation of HoxC-8 expression pattern and correlated modifications of the body plan” (Belting et al., 1998). Besides, and predictably, differences in the enhancer are also observed between HoxC-8 enhancers in mice and whales (a 4 base pair deletion), but as two of the same group of investigators admit, they have found no correlation between the sequences of the baleen whale Hoxc8 enhancer and any specific morphological trait that evolved in this species (Shashikant et al., 1998).

With changes in genes excluded as cause of vestigialization and loss of limbs in pythons, the remaining alternative explanation is an epigenetic regulatory mechanism. Having shown that Hox gene expression domains along the body axis determine the absence of forelimbs and vestigialization of hindlimbs in pythons, now we have to remember that patterns of expression of Hox genes in vertebrates determined by the patterns of expression of RA along the body axis, in which the neural tube and motor neurons, as was shown earlier, play a crucial role (see for further information in section Role of the Nervous System in Limb Development, chapter 14).

 

Loss/Reduction of Limbs in Aquatic Mammals

 

Paleontological evidence shows that the ancestral forms of modern cetaceans, such as Pakicetus inachus of Early Eocene (~58-48Mya) in Pakistan may have been land tetrapods exhibiting all the typical features of terrestrial mammals (figure 15.3). A latter stage (~47 Mya) in the evolution of cetaceans in the fossil evidence is exemplified by Ambulocetus natans, which shows signs of transition to aquatic morphology characterized by reduction of forelimbs but still retains well-developed hind limbs with webbed feet, reminiscent of hind limbs of the sea otter. It probably swam by vertical axial undulations of the spine, while using hindlimbs like a fluke. The next stage (~40 Mya) in the evolution of cetaceans is the elongation of the body and increase in the number of vertebrae as well as marked vestigialization of hind limbs (Basilosaurus) indicating adaptation to a fully aquatic life (Thewissen and Bajpai, 2001). At a final stage of evolution of ceataceans, flukes evolved and the swimming by axial undulation was complemented by tail oscillations.

Evolution of terrestrial mammals into marine swimmers followed, and/or was correlated with, changes in locomotory and other behaviors that the aquatic life imposed. Adaptive changes in the locomotory behavior and accompanying changes in morphology in the course of evolution of aquatic mammals can be illustrated with eclectic examples of modern animals that presumably are in the process of the evolutionary adaptation to the aquatic life (figures 15.4 – 15.8).

I am tempted to illustrate these successive stages in the evolution of the locomotory behavior of marine mammals from their quadruped ancestors, with examples from extant quadrupeds that presently are in different stages of morphological, physiological, and behavioral adaptation to aquatic life.

In short, minks (Mustela vison) paddle quadrupedally (figure 15.5), and freshwater otters (Lontra canadensis) (figure 15.6) swim mainly with their hind limbs (pelvic paddling), although they derive some additional lift from the tail (pelvic undulation). Sea otter, Enhydra lutris (figure 15.7), uses its highly asymmetrical feet as the propelling surfaces, but most of the power for the movements comes from undulations of the vertebral column (pelvic undulation) rather than from the muscles of the hind limbs. The giant South American freshwater otter Pteronura brasiliensis uses caudal undulations: sinusoidal motions of the vertebral column, like a wave moving through the entire spine, power a long and narrow tail that is dorsoventrally flat (figure 15.8. No otter swims like a modern cetacean, but the swimming mode of Pteronura approximates whale swimming. Modern cetaceans differ from Pteronura in having a rigid body with most of the movement concentrated at one point: undulation, thus, became oscillation. In addition, modern cetaceans evolved a fluke (Thewissen and Bajpai, 2001).

The loss of limbs in tetrapods would have been impossible if these animals would not have been able to adopt a new form of locomotion.

Given that the loss of limbs is a process of adaptation to new conditions of living (aquatic, fossorial, or dense grass environment) animals first had to learn new modes of limbless locomotion (lateral undulation, swimming undulation, concertina, etc.). The process of learning may have been facilitated by the fact that ancestral motor patterns and FAPs (fixed motor patterns) during evolution are not lost and motor circuits generating these FAPs could be activated under stressfully changed habitat conditions.

Figure 15.3. Evolution of the changes in swimming mode during cetacean evolution. Modern whales comprise baleen and toothed whales. Modified from Berta and Sumich (1999) and Thewissen and Fish (1997) (From Bejder and Hall, 2002)

 

 

 

Figure 15.4. Hypothesis for the evolution of the caudal oscillation swimming mode of modern Cetacea, based on Thewissen and Fish (1997). Different swimming modes are listed in the left column, and arrows indicate transitions that can be predicted on the basis of efficiency considerations. Modern mustelids swim using various modes, and cetaceans probably went through these modes sequentially during their evolutionary history. Morphological study indicates that Ambulocetus was probably a pelvic paddler or caudal undulator and that Kutchicetus was mainly a caudal undulator (From Thewissen and Bajpai, 2001).

 

We must not forget that all the basic modes of locomotion, swimming (undulatory waves passing down the body), crawling and lateral undulation are functions of a single motor pattern circuit that evolved in invertebrate ancestors of vertebrates. The motor pattern for swimming was not lost in tetrapods adapted to terrestrial life and most tetrapods are still capable of learning to swim.

Concertina locomotion (figure 15.9 A) implies that some part of the body is fixed on the ground in order to push the rest of the body forward. While this form of locomotion is widespread among burrowing snakes another form the so-called “internal concertina” has been adopted by many caecilians, limbless snake-like amphibians (figure 15.9 B). This mode of locomotion consists in undulatory movements performed by vertebral column only (not the body as a whole). The ability to use internal concertina has been lost once in narrow-bodied caecilians, the typhlonectids, because this locomotion mode hinders vertebral waves in narrow-bodied animals (Summers and O’Reilly, 1997).

 

Figure 15.9. A comparison of normal concertina (A) and internal concertina (B). Concertina is shown in a snake traversing a smooth surface. Internal concertina is shown in a caecilian moving in a burrow. The vertebral column and skull are superimposed on the outline of the caecilian (From Summers and O’Reilly, 1997).

 

Changes in the locomotory behavior preceded  and facilitated vestigialization and loss of hind limbs in the evolution of these aquatic mammals from terrestrial tetrapods (figure 15.10). Two crucial steps in this process were a reduction of the time of expression of Shh in the hind limb bud and later a loss of ZPA in the hind limbs (Thewissen, 2006).

It is noteworthy that during the ontogeny, cetaceans develop hind limb buds showing all the initial steps of terrestrial mammal limb bud development, including cell differentiation, formation of both signaling centers, the AER (apical ectodermal ridge) and ZPA (zone of polarizing activity), innervation, secretion of FGF8, etc., before entering the regression stage, which is believed to result from suppression of the expression of Shh (Sonic hedgehog). It is believed that the evolutionary reduction of the expression of Shh in the limb bud of aquatic mammals (and the corresponding limb reduction) started ~41 million years ago, whereas the total loss of Shh expression (and resulting loss of hind limbs) occurred ~34 million years ago (Thewissen et al., 2006).

Embryos of the limbless river mammal, Stenella attenuata also develop limb buds. The limb bud grows to reach a length of 10-30 cm before starting regressive processes that lead to their reduction and total disappearance. This suggests that constraints might have existed that prevented direct elimination of the development of the limb bud and imposed later apoptotic elimination of the limb bud during the evolution of the species. Moreover, individuals with vestigial hind limbs are observed, at a low frequency, among the adult populations of the spotted dolphin (Sedmera et al., 1997). The sperm whale (Physeter catadon) is the only toothed whale with a hind limb skeleton (15 cm long), although its expression is variable (Hall, 1995). Reduction of limb development in whales is extreme but rudiments of the tetrapod bones are present and 37% of individuals of the Antarctic population of minke whale, which have ossified femoral rudiment. Sometimes

 

Atavistic skeletal elements can be surprisingly complete; 79 cm long bones in 125 cm long left and right “hindlimbs” in a female humpback whale. (Bejder and Hall, 2002)

 

In this case the “penetrance” may be an indicator of the a continuing process of the loss of legs in these whales.

As for evolution of flippers from terrestrial mammalian limbs in the spotted dolphin, like other cetaceans, these organs, apparently adaptations for aquatic life, have retained the inner mammalian limb structure, except for a marked increase in the number of falanges, which is clearly an adaptation for the aquatic life (Sedmera et al., 1997).

In order to have an idea on the possible mechanisms of the evolutionary loss of limbs in limbless vertebrate groups, let’s briefly review the normal development of limbs in tetrapods (see also subsection Role of the Nervous System in Limb Development in chapter 14).

 

Embryonic Development of Limbs in Tetrapods

 

As early as the late 80s it was demonstrated that RA (retinoic acid) and its receptors (RARs) are involved in the early limb development, starting from HH (Hamburger-Hamilton) stage 14, that is prior to wing bud formation (stage 17-18), and later. Although maternal RA is known to be present in early embryonic stages, later its main site of synthesis is the mesoderm and neuroectoderm (Niederreither et al., 1997), with the ventral spinal cord, as a rich source of RA. Two sites of highest expression of one of the enzymes responsible for RA synthesis (RA/RALDH-2) are motor neurons coming from the ventral spinal cord to innervate limb buds (Zhao et al., 1996). Even before the formation of motor neurons, the hot spots of RA synthesis appear with formation of limbs and persist until limb innervation is just about complete. This is clearly too late for an assumed role of the RA synthesis in the general dorso-ventral patterning of the spinal cord. However, as retinoic acid is known to substantially increase neuron survival and axon outgrowth in spinal cord cultures, the hot spots are a likely factor in the formation of the limb zones, by rescuing relatively more neurons in these regions from morphogenetic cell death and by stimulating neurite outgrowth in the developing limbs (McCaffery and Dräger, 1994). The “hot spots” of RA synthesis in the spinal cord at this early embryonic stage (E 12) correspond to the sites where limb buds form (figure 15.11).

 

 

Figure 15.11. Fluorescent view of RA released by the mouse embryo E12.8 spinal cord (From McCaffery and Dräger, 1994).

 

The nervous system is the main supplier of the RALDH-2, and consequently of RA during early development. In mice embryos, from E8.5 to E10.5, the spinal cord RA derives from the dorsal spinal cord but it also diffuses from adjacent somites.

 

The total amount of RAs varied by 29-fold across different tissues with the lowest in the heart and the highest in the neural tube. (Maden et al., 1998)

 

 

Even the myogenic cells coming from somites are not differentiated into muscle cells until the 25 H&H stage, when RALDH-2 accumulates around nerves and blood vessels (investigators do not elaborate on whether the nerves are also involved in the production of RALDH-2 by the wing vasculature). The observation that the limb bud vasculature also releases RALDH-2 may be explained with the presence of neurally-derived RA, which stimulates expression of RALDH-2, as suggested by the fact that RALDH-2 gene has a RA responsive element (Berggren et al., 2001).

The AER (apical ectodermal ridge) develops as a result of signals from underlying mesoderm and from ventral limb ectoderm, between the dorsal and ventral ectoderm (Pizette et al., 2001). This is a source of FGFs including FGF8 secretion that is involved in Shh expression. Its inactivation leads to reduction of the limb bud size and limb skeletal elements (Lewandoski et al., 2000), whereas absence of both Fgf8 and Fgf4, in the AER leads to arrest of the limb bud development and elimination of the bud limb via apoptosis (Boulet et al., 2004).

The early idea that the cause of the loss of limbs  was an “arrested development”, was shown to contradict some embryological observations on limbless snakes, lizards, and cetaceans. In general, during the early embryonic development some of these species develop limb buds and form an AER (apical ectodermal ridge), which later disappear gradually (Bejder and Hall, 2002). So, e.g., it  was observed that the gradual reduction of the limb bud in scincid reptiles with greatly reduced limbs is caused by “necrosis”, under which, at the time, was understood not only the pathological process of cell death in metazoans but the physiological, i.e. programmed cell death, or apoptosis, as well:

Figure 15.13. Model for RA action during limb bud development. A, two phases of RA action during limb bud development are shown in relationship to several other factors known to play important roles in limb development. In the early phase, RA acts upstream of dHand to initiate ZPA formation. At the same time, RA acts upstream of Meis2 and Tbx5 to initiate limb budding. In the late phase, RA is needed to form an AER structure that extends fully along the distal region of the limb; at this stage RA may function in parallel with Fgf8, which is needed to establish AER function (From Mic et al., 2004).

 

NeoDarwinian Explanation of Loss of Limbs

 

Hox genes exhibit remarkable conservation among metazoans with respect to their sequence, clustered genomic organization and collinear expression along the body axis. As shown, no changes in the function of these genes and other key limb-inducing genes are involved in the loss of limbs in vertebrates, as it is proven, among other things, by the fact that most of the limbless species initially activate the “limb-determining” genes, form AER and ZPA, and even develop limbs to advanced stages before arresting their development or starting the programmed cell death of limb tissues.

From the neoDarwinian view, the occurrence of such radical morphological differences as the presence and absence of limbs, anteriorization and posteriorization of limb buds etc., between species that have functionally unchanged all the limb-determining genes (including Hox genes) is unexplainable at best.

 

Epigenetic Explanation

 

A look at expression patterns of HoxC-8 gene shows that both in chicks and mice embryos it is expressed in the mid-thoracic mesoderm and in the brachial region of the neural tube. However, the anterior boundary of expression extends less anteriorly in chicks than in mice, determining thus the longer cervical region, more posterior appearance of limb bud as well as the smaller number of thoracic segments in chickens (figure 15.14). It is noteworthy that the anterior boundary of HoxC-8 expression in both species coincides with the site of origin of the brachial nerves that innervate limbs in both species (Bejder and Hall, 2002). Also remember: expression of Hox genes in general, and HoxC-8 in particular, are regulated by RA, which downregulates expression of posterior Hox genes

along the embryonic A-P and causes respective truncation of the embryo (Kessel, 1992).

Genes for enzymes for RA synthesis in vertebrates have not changed. What has changed is the spatio-temporal pattern of expression of RA in limbed and in limbless tetrapods as well as in chickens and mice, as shown in figure 15.14. This change is clearly nongenetic (all the limb-inducing genes are present and functional in both limbed and limbless species).

Where may be the source of the epigenetic information that is used for these adaptive changes in expression patterns of Hox and other genes involved in the development of limbs or leading to limblessness in tetrapods?

Figure 15.14. Schematic comparison of Hoxc8 expression in chicken and mouse in relationship to morphological landmarks. Cervical, thoracic, and lumbar regions of the vertebral column and the brachial region of the neural tube are indicated. Brachial spinal nerves C6, C7, C8, and T1 in mouse and C13, C14, C15, and T1 in chicken are shown. Shaded region in somites and neural tube represent Hoxc8 expression. Regions of highest expression are indicated in dark shades. The double-headed arrow indicates the anteroposterior orientation of the body axis.

Abbreviations: a, anterior; p, posterior; nt, neural tube; t, thoracic vertebrae; s, somites; sn, spinal nerves; v, vertebrae (From Belting et al., 1998).

 

The evidence presented in this chapter as well as in chapter 14 (section Role of the Nervous System in Limb Development) on the evolution of limbs in vertebrates shows that RA signals from the neural tube and local innervation are essential for the development of limbs in tetrapods.

In the process of vertebrate limb loss and reduction are also involved mechanisms of programmed cell death, which are epigenetically regulated as well. The process of apoptosis that leads to regression of the limb bud is known to be related to the fact that the AER does not secrete FGF, especially FGF-8 and FGF-4 (Boulet et al., 2004).

The fact that no changes have occurred in genes for the programmed cell death in limb tissues of tetrapods with reduced limbs, or that have lost their limbs, unequivocally shows that the cause of the programmed cell death is not genetic. As shown earlier (sections Apoptosis in Invertebrates and Neural Control of Apoptosis in chapter 6), the programmed cell death during the individual development is epigenetically determined via signal cascades that ultimately originate in the nervous system. Hence, evolution of the programmed cell death in limbless tetrapods has to start with changes in the activity or properties of neural circuits that produce signals that activating signal cascades for the programmed cell death.

 Loss of Characters in Cave-dwelling Animals

 

Life in dark caves usually leads to an evolutionary pressure for losing certain characters and acquiring troglomorphic (from ancient Gr. trogle, cave) characters. Cave-dwelling animals differ from their epigean (from ancient Gr. epi - at, on, over, and geo - the earth) conspecifics in a number of characters (table 15.1).

One of the most widespread evolutionary phenomena is loss of eyes in animal species living in dark caves, where normal photoreceptive eyes are of little use, if any.

Table 15.1. Catalogue of ‘troglomorphic’ features. These are the characters that frequently differ from those in closely related epigean organisms. Troglomorphic organisms may display only a few, some, or all of these characters. Some characters may differ in either direction (e.g., some troglomorphic fish display reduced metabolism, other species exhibit an increase) (From Romero and Green, 2005).

__________________________________________________________________________________

 

Morphological                                             Physiological                            Behavioral

__________________________________________________________________________________

Reduced, diminished, or lost

    Eyes, ocelli                                               Metabolism                               Photoresponse

    Visual brain centers                                  Circadian rhythms                      Aggregation

    Pigmentation                                             Fecundity                                  Response to alarm

    Pineal organ                                                                                               Aggression

    Body size

    Cuticles (terrestrial arthropods)

    Scales (fishes)

    Swimbladder (fishes)

Enlarged, enhanced, or exaggerated

    Chemo- and mechano-receptors               Life span

    Appendages                                              Lipid storage

    Body size                                                  Metabolism

                                                                     Egg volume

__________________________________________________________________________________

 

Most of the 50,000 to 100,000 obligate cave-dwelling species (arachnids, insects, crustaceans, fish, and salamanders) have lost their eyes (Fong et al., 1995)

  

Loss of Eyes in Astyanax faciatus (mexicanus) : Epigenetics of an Evolutionary Event

 

The Mexican teleost fish species, Astyanax mexicanus, exists in two forms, an eyed surface-dwelling (epigean) and an eyeless cave-dwelling (hypogean) form. Both morphs are interfertile although usually in nature they are spatially isolated. Over the last 10,000 years, at least 4 times, cavefish populations of Astyanax independently evolved various degrees of loss of eyes and at least 29 different populations of a blind/eye-reduced cave-dwelling (hypogean) morph are known (Dowling et al., 2002; Jeffery, 2005). Correlated with the loss of the eye structure and function and with reduction of the size of the optic tecta (the visual processing center in the brain of fish, amphibians, and reptiles), the cave-dwelling morph has also evolved new behaviors, various degrees of body depigmentation as well as several constructive characters in jaws, teeth, taste buds, mechanosensory system of cranial neuromasts, compensating for the lack of eyes (Teyke, 1990). Changes also occurred in the number of rib-bearing thoracic vertebrae in the axial skeleton (Dowling et al., 2002).

What takes place in the embryos of blind morphs is not complete prevention of oculogenesis. Initially, development of the eye Anlage in the embryos of the blind cavefish proceeds normally. There are no remarkable differences in the early development of the eye Anlagen between the blind hypogean and eyed epigean embryos besides the eye size and proportions. The divergence becomes apparent during the growth stage of the eye, when the embryos of the blind morph fail to enter that stage and the vestigial eye is covered by the growing regional skin (figure 15.15). Simultaneously, regressive processes start with the programmed cell death (apoptosis) taking place in the lens and later in the retina.

The development of the eye in invertebrates and vertebrates in general depends on a “conserved Pax-6 dependent mechanism” (Quiring et al., 1994) that is operative at early stages of development (Tomarev et al., 1997). Pax-6 gene is expressed in both sides of the midline of the anterior part of the neural plate. Anteriorly, the Pax6 expression domains fuse to form the forebrain and optic Anlagen. Secretion of Shh (Sonic hedgehog) by midline tissues is also essential for the development of ventral eye structures (Zhang and Yang, 2001). In any case, the initial signals for the development of the eye Anlagen originate in the neural plate/neural tube.

In cavefish, investigators found that all of oculogenic genes a re functional and all of them are expressed normally:Anlagen originate in the neural plate/neural tube.

In cavefish, investigators found that all of oculogenic genes are functional and all of them are expressed normally:

 

It appears that eye gene cascades are completely operational in cavefish embryos prior to the general transcriptional shutdown that occurs after the beginning of apoptosis. (Jeffery, 2005)

 

 

Figure 15.15. Eye development and degeneration in Astyanax mexicanus. Surface fish (A) and cavefish (B) adults. Diagram showing the timing of eye growth and development in surface fish (top) and eye degeneration in cavefish (bottom) (After Jeffery, 2005).

Figure 15.19. Morphogenesis of the lens. (A-D) show the stages of lens development in the mouse from E8.5 to E11.5 in daily intervals. The three tissue layers involved in eye development include the surface ectoderm (medium shaded) the mesenchyme (lightly shaded) and the neuroepithelium of the optic vesicle (dark shaded).

Abbreviations: lpl, lens placode; lv, lens vesicle; m, mesenchyme; oc, optic cup; ov, optic vesicle; pr, presumptive retina; lp, lens pit; pce, presumptive corneal ectoderm; ple, presumptive lens ectoderm; prpe, presumptive pigmented retinal epithelium (From Lang, 2004).

 

 

 

NeoDarwinian Explanation

 

NeoDarwinian paradigm has to deal with a huge difficulty: four times, within an evolutionary instant of about 10,000 years, hipogean forms of A. mexicanus, independently lost their eyes and pigmentation and additionally evolved several new “constructive” traits. This evolutionary change involved no changes in the function of relevant genes. It is believed that

 

The interesting aspect, and the rub, of evolutionary reductions is not that they are too difficult but rather that they are too easy to explain in theory. Distinguishing among various theories of regressive evolution is hampered by lack of empirical information and by experimental limitations posed by many of the organisms in question. (Fong et al.,  1995)

 

In the case of  A. mexicanus, after half a century of studies on the nature and origin of the loss of eyes in cave fish, this “lack of empirical information” suggests anything but an easy explanation. As an inherited character, the evolutionary loss of eyes in the hypogean form of Astyanax mexicanus requires, as a sine qua non, some new specific information to be transmitted from eyeless parents to the offspring. And, since that information is not genetic, i.e. no changes in genes are involved (Jeffery, 2005), the remaining alternative is that the information for this radical change in morphology is epigenetic. Any attempt to understand or explain the evolutionary loss of eyes in cavefish should basically deal with the fundamental problem of the origin of the information for the loss of eyes as a morphological novelty. Identifying that epigenetic information and its source essentially implies identifying the point where the eye developmental pathways of both forms (eyed and blind) of Astyanax diverge.

The neoDarwinian paradigm sees no other source of that information except mutations affecting the function of genes involved in the development of eyes in the fish, or the increase of the frequency of a preexisting allele (in such a case no new information would be necessary). But there is no evidence for relevant mutations to have occurred in genes related with eye development and there is no evidence that any allele for “eyelessness” existed in epigean forms of Astyanax. On the  contrary, experimental evidence shows that all of these genes are functionally normal in both the blind cave fish and its conspecific eyed form.

A number of investigators have argued against genetic mechanisms of the loss of eyes in cave fish:

 

Gene expression data suggest that loss of function mutations have not occurred in cavefish eye genes, including those structural genes that function at the bottom of regulatory cascades… lens transplantation indicates that cavefish have the capacity to form a complete eye and that they possess and are capable of using all the genetic factors necessary for later eye development… The developmental evidence does not support an evolutionary model that proposes loss of function of the genes involved in early eye development and/or eradication of the embryonic eye to conserve energy. (Jeffery, 2005)

 

The hypotheses of neutral mutations and energy conservancy also cannot account for the source of information for the loss of eyes, whereas the hypothesis that sees eye loss as a byproduct of modification of the feeding apparatus also fails to address that question.

The neutral mutations hypothesis is equally unfit for explaining loss of eyes in Astyanax mexicanus. According to that hypothesis, under conditions of darkness in caves, where the sight is not useful, mutations in genes that are involved in eye formation, but do not affect the development of other structures, might accumulate through the genetic drift. The latter would make it possible for neutral mutant alleles to be fixed in cavefish populations. But even theoretically the genetic drift would need evolutionarily long periods of time “to fix eyeless alleles”, whereas the loss of eyes in cave-fish, which occurred four times in A. mexicanus, took only a “moment” (~10,000 years) by evolutionary standards.

Even if, for the sake of argument, one would accept that theoretically it would be possible for neutral mutations for eyelessness to occur and accumulate, there is no evidence to suggest that alleles for eye loss are accumulated in the eyeless morph. On the contrary, as pointed out above (Jeffery et al., 2005), all the genes involved in the eye formation of the cave-fish have remained functional, as functional as in the epigean form. Hence,

 

Experiments provide evidence against the neutral mutation hypothesis as an evolutionary mechanism for eye degeneration. (Jeffery, 2005)

 

The hypothesis of energy conservancy is an hypothesis of indirect selection. It proposes that loss of eyes under conditions of darkness would offer a selective advantage by setting free energy for the development of sensory organs and other “constructive” traits that evolved in cavefish. To talk about the selective and evolutionary advantages that would offer a new trait is one of those “too easy” things, but the devil is in the details of the loss and transformation of the eye and the concurrent molding of these new phenotypic characters within an extraordinary short period of time. It is true that cave fish invest excessively matter and energy for the processes of initial development of eye structures and later for their regression via apoptosis. It is argued that

 

Most examples of evolutionary reduction are of interest because they resist explanations as adaptations per se. Most explanations of character reduction invoke indirect selection in terms of energy economy or antagonistic pleiotropy arguments, although what is meant by energy in such a context is usually unstated, and few, if any, such arguments are framed as testable hypotheses. (Fong et al ., 1995)

 

Summarizing the arguments rejecting that hypothesis, one of the leading  investigators of the loss of eyes in these cavefish writes:

 

Several lines of evidence argue against the possibility that cavefish eye development is blocked to conserve energy. First, cavefish males and females show the same degree of eye reduction, although the high cost of egg production might be expected to dictate a greater degree of eye reduction in females, as has been reported in cave-adapted beetles. Second, cave fish populations inhabiting pools under bat colonies do not appear to be food-limited, yet they show significant eye regression. Third, the manner of eye degeneration in Astyanax cavefish does not appear to be economical. Instead of undergoing eye loss at a very early stage, the cavefish eye develops to a relatively mature stage prior to the beginning of degeneration, presumably at high energetic cost. (Jeffery, 2005)

 

Finally, the hypothesis that sees the evolutionary eye loss as a byproduct of the need for better feeding apparatus in caves (Jeffery, 2005) explains the benefits of the evolutionary change but does not address the most essential fact of the evolutionary change, i.e. whether the new information for the eye loss is mutational, which the author of the hypothesis denies, or epigenetic.

A prediction of the neoDarwinian paradigm would be that the loss of eyes during the embryonic development would be associated with a downregulation of expression of oculogenic genes. Contrary to this prediction, many of these genes are upregulated in the cavefish rather than in the surface fish (Jeffery, 2005).

Now, summarizing, it may be said that all the neoDarwinian hypotheses presented above fail to account for the exceptionally rapid and repeated loss of eyes in A. mexicanus.

 

Epigenetic Explanation

 

The fact that the evolutionary change leading to eyelessness in A. mexicanus implies no changes in genes unequivocally tells us that the evolutionary change is transmitted to the offspring by nongenetic means.

Before considering the possible developmental mechanisms of evolution of eyelessness in caveshishes, let’s get a glimpse of the recent evidence suggesting that eyelessness in fish is an evolutionarily plastic trait.

Some populations of A. mexicanus, and other cave fish as well, show a remarkable polyphenism; within the some population blind, eyed, and intermediate eye morphologies exist (Romero and Green, 2005). Numerous observations have shown that cave fish exhibit not only complete loss of eyes but also various degrees of vestigialization of eyes.

Exposure of the larvae of the eyed, eyeless, and hybrid forms of A. mexicanus to light or darkness for one month leads to dramatic phenotypic changes such as development of eyes in the eyeless form and enlargement of eyes in the eyed form, suggesting that the photic stimulus influences the developmental pathways of eye formation (figure 15.20). Remember, the only known way light may influence developmental pathways is the neural way.

The observed degree of eye polyphenism might have been the raw matterial for evolution of eyelessness in cave fish.

The ability of troglomorphic individuals to regain some eye tissue and pigmentation when experimentally exposed to light illustrates the retention of a substantial capability for phenotypic plasticity even if under natural conditions they seem to represent an ecotype. (Romero and Green, 2005) The fact that initial stages of the development of the eye Anlage, including development of the crystalline lens, take place normally in the eyeless morphs suggests that all the basic genes involved in eye formation (Pax6, Shh, Sox 2, and Sox3) are normal and functionally unchanged (Jeffery, 2005). The other fact that in some other vertebrates, lens formation occurs only in the presence of the retina (Goss, 1969; Furuta and Hogan, 1998; Reza and Yasuda, 2004a; Reza and Yasuda, 2004b) and the fact that the best substitute for the eyecup in lens regeneration experiments is the adjacent brain (Goss, 1969), implies that a neural signal is necessary for lens development. It is likely that a failure of the retina to send that neural signal may be the proximate cause of the arrest of development of lens in the hypogean form. In turn, suppression of the development of the lens in the cavefish is the proximate cause of the arrest, or even lack, of development of the iris, cornea, and retinal pigment epithelium as is indicated by the fact that implantation of the epigean embryo lens in the optic cup of hypogean embryos induces formation of those optic structures in the presumptive eyeless cavefish.

Figure 15.20. Variation in developmental responses to light exposure of larval surface, cave, and hybrid Astyanax fasciatus. Larvae were reared in continuous darkness or continuous light for 30 days beginning when they were 24 h old. All three forms reveal an effect of light in the development of their eye tissues and the number of melanophores. The difference is particularly dramatic in the cave fish larvae (From Romero and Green, 2005).

 

Loss of Pigmentation in the Cavefish A. mexicanus

As a consequence of living in darkness, the hypogean morph of the teleost fish Astyanax mexicanus, has lost not only its eyes but its pigmentation as well.

The body pigmentation in this fish depends on the presence of pigment cells, melanophores, in the skin. These pigment cells, as well as two other Astyanax types of pigment cells, iridophores and xanthophores, originate from neural crest cells that form in the neural keel, a structure that forms by the infolding of the neural plate (Papan and Campos-Ortega, 1994) under the influence of Hh signaling that affects the medial and lateral neurogenesis (Takamiya and  Campos-Ortega, 2006).

Morphologically, melanoblasts in cavefish resemble melanophores and even are capable of producing melanin when provided with L-dopa (McCauley et al., 2004). Various cave fish populations differ widely from each other in the degree of depigmentation and in the proportion of melanophores to the total number of melanoblasts. While the pigmented epigean form of A. mexicanus has a 1:2 ratio of melanoblasts to melanophores, the Curva cave fish has a 8:1 ratio and the Pachón cavefish has no melanophores at all, and have the smaller number of melanoblasts than any cavefish (McCauley et al., 2004).

From a neoDarwinian view, as early as 1957 Sadoglu hypothesized that depigmentation in Astyanax was related to a mutation in an unidentified gene. Later, it was hypothesized that mutations in 2 unidentified genes might be involved in the evolutionary depigmentation of the cavefish. Both hypotheses are incompatible with the fact that depigmentation in Astyanax is not an “All-or-None” process (melanophores are still produced at a low proportion), as it would be expected when one or two unfunctional genes would be involved, but it is an ongoing epigenetic process of gradual loss of ability to differentiate melanoblasts into melanophores, as indicated by the wide range of variation of the melanophore to melanoblast ratio.

In an attempt to overcome such difficulties, later it was proposed that evolutionary depigmentation of Astyanax is result of accumulation of neutral mutations especially at a late step of the metabolic pathway of melanin synthesis. The fact that hypogean fish give birth to offspring that produce a proportion of melanophores rejects the hypothesis that neutral gene mutations may be involved in depigmentation of cavefish. Furthermore, melanoblasts in various populations of cave-dwelling Astyanax, including the one that produces no melanophores at all, synthesize melanin when provided with L-dopa, indicating that all of them have conserved the tyrosinase, which catalyzes various steps of melanin biosynthesis from tyrosine.

The fact that almost all of the depigmented Astyanax cave fish are capable of forming melanoblasts and melanophores and melanoblasts are capable of synthesizing melanin when provided with L-dopa clearly shows that there is no change in any gene that causes the evolutionary depigmentation and regression of pigment cells in these fish.

For these reasons, the new tendency is to see the depigmentation of the cave fish as an epigenetically determined evolutionary change. It is suggested that in cave morphs of A. mexicanus, the melanogenesis cascade is not blocked “because of a missing genetic component” but because of a nongenetic cause: 

A permanent block in tyrosine accessibility seems to have occurred during cavefish evolution. (McCauley et al. 2004)                   

 

Loss of a Sexually Selected Character in Lizards

Sexual dichromatisms, differences in body color according to the sex, have evolved in many phrynosomatid lizards, in which males have conspicuously blue colored throat and belly as  a courtship signal or as a warning to predators. Repeated losses, and less frequent gains, of sexually dichromatic coloration were found in a study on 130 lizard species. Loss of sexual dichromatism was found to be related to ground-dwelling, due to increased predation in such habitats.

The loss of male conspicuous coloration seems paradoxical given that the sexual selection would favor evolution and maintenance of male conspicuous coloration. It is argued that in this case, as well as in many other described cases of the loss of conspicuous coloration of plumage in birds, the cause of the loss of the male conspicuous coloration, is not genetic but is a consequence of changes (reduction or loss) in female mating preferences (Wiens, 1999).

But what is the cause of the reduction or loss of female mating preferences? The fact that these preferences may vary between individuals of the same genotype, between individuals of identical genotypes, and even may change during the lifetime of one and the same individual, clearly indicates that no changes in genes or genetic information are involved in the changes of female mating preferences. These preferences are determined by the mate recognition system, comprising sensory organs and their pathways to the CNS. As it will be explained in some details later (Mate Recognition System and Evolution of the Mate Recognition System in chapter 20), mate preferences are neurocognitive products of the activity of specific neural circuits. 

Loss of Sexual Dichromatism in Birds

In the northern hemisphere, some bird species are dichromatic and some – monochromatic. Monochromatism is believed to be a derived character. Phylogenetical evidence shows that loss of dichromatism has occurred repeatedly in ducks, with a stronger tendency for losing rather than gaining it. The same is true for passerine birds, where dichromatism is lost three times more often than gained and, according to Peterson (1996), in birds in general, dichromatism is lost 5 times more often than gained (Omland, K.E. 1997).

This seems to contradict the general neoDarwinian belief that dichromatism is related to sexual selection for colorful conspicuous plumage. For this reason some evolutionists like Mayr (1942) and Peterson (1996) resorted to gene (allelic) drift as a possible agent of the evolutionary loss of dichromatism. But the hypothesis that drift is a major player in the loss of dichromatism has not been substantiated and, consequently, lends no support for the neoDarwinian paradigm. Observational evidence, e.g., shows that ducks and mallard species that lose dichromatism retain the gene for the pigment in a functional state as may be inferred from the persistence of the yellow-green color of the bill after the loss of plumage dichromatism. Besides, “Monochromatic species in five of the six major clades of Anas (all except the green-wing clade) show evidence of vestigial features of the bright dichromatic plumage of their Northern relatives” (Omland, 1997).

A plausible mechanism of the loss of dichromatism in birds would be an epigenetic mechanism involving changes in female sensory biases followed by the action of natural selection. Not only has this hypothesis found greater empirical support, but it seems to rationally account for the frequent loss of female preferences (Ryan, 1998; Wiens, 2001), based on the relative evolutionary plasticity of the neural circuits determining animal behavior (see on the evolution of female preferences, in chapter 20).

 

Shedding of Teeth in the Mekong Giant Catfish

Mekong River giant catfish, Pangasianodon gigas (Teleostei) is the world’s biggests freshwater catfish reaching a body length of up to 2.5 meters. As juvenile it has three kinds of homodontic conical teeth: palatal, pharyngeal and jaw teeth, with general characteristics of the teeth of other teleosts including cap enameloid and tubular dentine, but showing a greater resemblance to bone tissue. In the later life a process of resorption of tooth tissue by osteoclast-like cells occurs. Both processes of resorbtion of the successional teeth and shedding of the functional teeth lead to the adult state of toothlessness in the Mekong River catfish (Kakizawa and Meenakarn, 2003).

Needless to say, teethed and toothless states occurring in the life of an individual, i.e. the loss of teeth in the fish, imply no changes in genes, no selection or drift, but just an epigenetic change in the behavior of osteoclasts, whose differentiation is under neural control via hormonal and neurohormonal mechanisms (Ohlsson et al., 1998; Canalis, 2003; Canalis and Delany, 2002; Weinstein et al., 2002; Burt-Pichat et al., 2005).

The fish is a living proof that no changes in genes are necessary for switching between toothed and toothless states. 

 

Loss of Life History Characters

Loss of particular life history stages has been a widespread phenomenon in the evolution of invertebrates and vertebrates. It has occurred often in the life cycle of animal parasites.

Many Coleoptera and Diptera have lost the imago stage. Calyptraeid gastropod species with feeding larvae lose that stage and transform into direct-developing species. The loss of the larval stage is rapid (Collin, 2004).

Direct development from large eggs evolved 11 times whereas from nurse eggs - eight times. Direct-developing species with nurse eggs have the potential of transition to an alternative mode of development but direct developing species with large yolky eggs may not be able to change the mode of development. Often direct-developing species lose some morphological features used for swimming and feeding in the water column (Collin, 2004).

Loss of life cycle stages is observed more frequently among the species with complex life cycles and is often correlated with the appearance of parthenogenesis and with expansion of the species’ geographic range (Moran, and Whitham, 1988). For instance, species of the aphid subfamily of Pemphiginae use two host plants in their life cycle, but most of genera of the subfamily skip the winter host plant. This enables the species to expand its geographic range by populating regions where the winter host plant is absent.

Among vertebrates, numerous cases of loss of the terrestrial stage are described in salamanders. These cases of paedomorphosis, and especially the facultative paedopmorphosis when, depending on environmental conditions, animals may metamorphose or not (the phenomenon is observed not only in amphibians but is also described in insects) demonstrate that no genetic changes are necessary for evolution of the loss of life history traits.  

Loss of Stages in Complex Life Cycles in Insects

Many aphids, including many species of Pemphiginae subfamily, show dispersal polymorphism. Pemphigus betae is an aphid with a complex life-history. It has a spring gall-forming phase on the narrowleaf cottonwood, Populus angustifolia, and a summer root phase on the secondary host plants of the genus Rumex. In autumn, with the drop in temperature as the only known cue, winged insects from root colonies fly to deposit their sexual generation on P. angustifolia. However, in response to crowding, P. betae may skip a phase (the first host, cottonwood tree) of its life cycle, by producing a wingless parthenogenetic generation that feeds on roots of Rumex and goosefoot plants of the genus Chenopodium. This is the phenomenon of anholocycly. The nonmigrating root colonies reproduce in the spring in the roots of the same Rumex plant.

Populations of this species in the Weber canyon, Utah, also show a clear tendency to switch to the reduced, one-host life cycle in the upper elevations of the canyon (Moran and Whitham, 1988; Moran et al., 1993).

It is observed that even clones with identical histories and genotype show very different natural tendencies for producing winged migrants (Moran et al., 1993). The mechanism of this radical change in the life history and in the morphology (winged/wingless individuals) is not known. What is certainly known is that no changes in genes are involved in producing it and that the development/suppression of wings in insects is ultimately neurally, i.e. epigenetically determined (see on the wing polyphenisms and experimental polyphenisms in insects in chapter 11, and on the evolution of wings in insects in chapter 14).   

Loss of Adult Stage of Development - Paedomorphosis in Insects

Many insects exhibit paedogenesis (neoteny), i.e., they reach sexual maturity during the larval stage and do not metamorphose into the adult form. Facultative paedogenesis in insects arose at least six times (four times in Diptera alone), twice independently in gall midges, Heteropeza pygmaea and Mycophila speyeri of the Cecidomyiidae (Diptera) family (Hodin and Riddiford, 2000). Female individuals of both species develop functioning ovaries and reproduce during the larval stage. The only detectable difference between the paedomorphic and metamorphic species is a larval expression of the functional ecdysone receptors, EcRs, and USP (ultraspiracle) in paedomorphic species.

From the neoDarwinian point of view, changes in genes responsible for metamorphosis would be necessary for the parallel evolution of paedomorphosis in these midge species. The fact that the ecdysone pathway responsible for entering metamorphosis is conserved not only in species of the family Cecidomyiidae but across insect taxa, refutes that neoDarwinian explanation.

An epigenetic explanation, based on the present knowledge of the neurohormonal mechanisms of metamorphosis in insects seems to be plausible. The functional receptor (EcR + USP) responsible for metamorphosis in insects is activated by ecdycone secretion by the prothoracic gland (and by the activity of nerve endings), which in turn is cerebrally regulated by secretion of the neurohormone PTTH (prothoracicotropic hormone).

 

Loss of Diphenism in Experiments on B. anynana  

Diphenic animals, under changed environmental conditions, can adaptively switch their offspring to monophenism. This implies an adaptive inactivation of one of the alternative developmental pathways for the trait.

Under natural conditions, the butterfly B. anynana produces offspring of two different seasonal phenotypes: with eyespots on their wings during the wet season, and plain wings in the dry season. This helps B. anynana to match the seasonal changes in the natural background and become less visible to its predators.

When two groups of the B. anynana butterflies were reared in different conditions (one group of in wet and the other in dry conditions) for twenty generations, each group evolved into a different race: butterflies of the group kept in wet environment (and high temperature) continued to produce only spotted offspring and those kept in dry and cool environment produced only plain wings when reared in each of the above alternative conditions.

Although we do not know the precise mechanism of this inherited transformation, we know with certainty that this inherited change in the butterfly phenotype involves no gene mutations (it occurs not randomly but systematically in the population, under laboratory conditions). It is likely that the processing of the sensory input on the environment received by the CNS for twenty generations is somehow involved in this adaptive change of the wing epigenetic program. This is not a mere theoretical inference. The production of the eyespotted wet-season morph results from an earlier and increased secretion of ecdysteroids in this morph (local application of ecdysteroids also induces formation of eyespots in wings). In turn, production of ecdysteroids is stimulated by the brain neuropeptide, PPTH (prothoracicotropic hormone), and by direct neural control (Chapman, 1998d).

Hence, not any change in the ecdysteroid genes  (these genes are unaffected) but a  neurally determined switch (on or off) of the secretion of the neurohormone PTTH in insect’s brain is responsible for the evolution of the monophenic forms of the East African butterfly in laboratory.

 

Loss of Terrestrial Mature Stage in Amphibians - Paedomorphosis

Loss of terrestrial stage by reaching reproductive maturity while still in a larval stage has occurred both in urodeles (salamanders and newts) and anurans (frogs). Salamanders of the genera Necturus and Siren, in North America and Proteus (subterranean cave salamanders) in Europe) have completely lost the ability to metamorphose, hence are known as obligatory paedomorphic. In distinction from them, most salamander species of the genus Ambystoma are facultatively paedomorphic, i.e. under certain environmental or laboratory conditions they can switch from paedomorphosis to full metamorphic development.

The Ambystoma tigrinum complex consists of species of salamanders that during the last few million years have independently evolved several times obligate and facultative paedomorphosis from the ancestral metamorphic state (Shaffer and Voss, 1996; figure 15.21).

Paedomorphic axolotl (Ambystoma mexicanum) reaches sexual maturity and reproductes while conserving larval traits, without undergoing metamorphosis. It retains external gills throughout life although it also develops lungs.

The mechanism of paedomorphosis can be understood only in the context of the general mechanism of metamorphosis. Metamorphosis in salamanders is stimulated by a surge in the level of the hormone thyroxine determined by a signal cascade that starts in the salamander’s brain (figure 15.22 ). The timing of the activation of the cascade is determined by the  

hypothalamic maturation comprising neurons of several regulatory centers and culminating at the time of the secretory surge. (Rosenkilde and Ussing, 1996)  

Paedomorphic salamanders fail to generate the characteristic burst of hypothalamic stimulation for activating the thyroid axis. This seems to be the main mechanism behind the axolotl paedomorphosis (Rosenkilde and Ussing, 1996). The hypothalamus regulates reproductive morphology and physiology, while evading its role as regulator of metamorphosis (figure 15.23).

The regulatory role of the brain in the process of metamorphosis in salamanders is not limited to the activation of the hypothalamic-pituitary-thyroid axis [thyrotropin-releasing hormone (TRH) à thyroid-stimulating hormone (TSH) à thyroid hormones (T3 and T4)].

 

Figure 15.21.  A reconstruction of the evolution of life history mode in the tiger salamander complex. Metamorphosis is treated as an unordered character with three states: transforming, facultative (both conditions found in a single population), and paedomorphic. Taxon names are the species or subspecies of Ambystoma, followed by the general locality of the sample (From Shaffer and Voss, 1996).

There is evidence suggesting that, via the hypothalamic-pituitary axis, the brain controls the antagonist effects of prolactin on metamorphosis and, via the hypothalamic-pituitary-adrenal axis, controls the agonist effect of corticoids (by increasing the number of T3 receptors) (Rosenkilde et al., 1996).

The action of thyroid hormones in the morphological transformation during metamorphosis is mediated by their nuclear receptors. It is observed that the highest levels of thyroxine in blood coincide with maximal synaptogenesis and other changes in hypothalamic neurons (See also Neural Control of Metamorphosis in Amphibians in chapter 6).

Metamorphosis has been experimentally induced in paedomorphic salamanders by administration of thyroid hormones but it can also be induced by manipulations at every level of the neurohormonal cascade. Thyroid hormone (T4) implanted in the brain is 10 times more active in inducing metamorphosis than when intravenously administered. However, morphological transformations may not be complete and increased mortality in metamorphic transformants is observed.

In addition to neurohormonal manipulations, experimental metamorphosis in paedomorphic salamanders is induced by stressful conditions (capture stress and conditions of captivity) that cause general disturbance in the central nervous system or by increasing the environmental temperature (Rosenkilde and Ussing, 1996).

Figure 15.22. Neurohormonal mechanism of metamorphosis in salamanders (From Rosenkilde and Ussing, 1996).

Figure 15.23. Diagrammatic representation of a mechanism of paedomorphosis in salamanders. The T4 surge occurring at the stage when toes differentiate shows (1) that TSH stimulating neurons have matured and are able to secrete; (2) that the TSH neurons are able to secrete and stimulate the thyroid; (3) that this gland is sensitive to TSH; and (4) able to secrete thyroxine to a high plasma level. The immersion experiments show that both young and older larvae are (5) able to respond to T3 with metamorphosis, but (6) the ability to activate thyroid hormone by deiodination of T4 to T 3 is delayed compared to metamorphosing species. Finally, two possible inhibitors are suggested by some experiments. Inhibition by prolactin, most probably (7) at the tissue level, or (8) a cerebral inhibitor, acting at the pituitary stimulating neurons.

Abbreviations: T4, thyroid prohormone, thyroxine; T3, the active deiodinated form of thyroxine; TSH, pituitary thyroid-stimulating hormone; PRL, prolactin (From Rosenkilde and Ussing, 1996).

Cases of spontaneous metamorphosis in paedomorphic salamanders have also been reported, corroborating the idea that no changes in genes are necessary for transition from metamorphosis to paedomorphosis and vice versa. 

NeoDarwinian Explanation 

The fact that paedomorphic salamanders spontaneously or under stressful conditions can revert to the ancestral state of metamorphosis unequivocally proves that they are in possession of the functionally intact ancestral “metamorphosis genes” and developmental mechanisms of metamorphosis, despite the long time since they abandoned that biphasic life history. Hence, any neoDarwinian mechanism of gene mutations, gene recombination, changes in allele frequencies and any other imaginable genetic mechanism are excluded from involvement in the loss of the life history stage and the appearance of paedomorphosis in salamanders. 

Epigenetic Explanation 

The essential question on evolution of paedomorphosis is: Where the signal cascade that determines metamorphosis is disrupted in paedomorphic salamanders?

The fact that all the hormones of the signal cascade for metamorphosis are normal and functional suggests that the disturbance may be at  the initial neural signals that activate the cascade. The theoretical inference that the disruption has occurred at a cerebral level is corroborated by empirical evidence:

1. Neurobiological disturbances in the brain, related to stressful conditions (capture and captivity ) induce paedomorphic individuals to perform metamorphosis

2. The extremely higher efficiency (10x higher) of brain implants of thyroid hormones in comparison with systemic administration of the hormone in inducing metamorphosis.

The hypothalamic neurons respond to the surge in thyroid hormone (figure 15.24) by removing an inhibitor, thus enabling them to secrete TRH (thyrotropin-releasing  hormone)

Why these neurons do not respond to the production of thyroxine in paedomorphic salamanders?

Hypothalamic neurons self-activate and secrete TRH in response to low premetamorphic levels of thyroxine. The fact that they do not respond that way in the case of paedomorphic salamanders suggests that the hypothalamus may have adaptively heightened the set point for responding to the hormone. The changes in set points are a well known epigenetic function of hypothalamus in vertebrates. 

Loss of Physiological Characters in Drosophila melanogaster

Loss of Resistance to Environmental Stressors in Drosophila melanogaster 

Under laboratory conditions, Drosophila melanogaster loses rapidly, within 3 years, the resistance to environmental stressors, starvation and desiccation. For starvation the mean time to 50% mortality declined from 50.1h to 35.9h and for desiccation it shifted from 14.3h to 9.8h.

The rapidity of the response suggested that mutation accumulation could not account for it. (Hoffmann et al., 2001)

Figure 15.24. Focal points of the activity increase in the thyroid axis leading up to metamorphosis (From Rosenkilde and Ussing, 1996).

Selection for early reproduction as well leads to the loss of these traits, although that character is inversely related to the resistance to environmental stressors (Hoffmann et al., 2001) indicating that selection does not act on genes or genetic material.

The extraordinary short time of the evolution of the above characters, under laboratory conditions, clearly suggests that a nongenetic mechanism is responsible for evolution of the above traits in as  little as three years.  

Loss of Behaviors

Loss of Dung Ball Rolling Behavior in Beetles 

Construction and rolling of dung balls for feeding and brooding as well as the nesting behavior in beetles arose at least 65 million years ago (Forgie et al., 2005) and it is suggested to have independently evolved several times in the Old World, in the genera of Scarabaeinae and Gymnopleurini as well as in the tribes Canthonini, Sisyphini, and Onthophagini. This behavior of rolling back portions of dung became a predominant mode of food relocation in horizontal tunnels. The dung ball rolling has been lost and reversion to the ancestral state of pushing and/or carrying in Scarabaeus galena and some Sceliages species has also been described (Forgie et al., 2005). 

Loss of the Acoustic Startle Response in Moths Endemic to Bat-free Habitats

 It might be predicted that moths that are no longer under bat predation threat, over time, will lose the ability to hear bat echolocation calls. The hypothesis is validated in a study on 7 species of day-flying moths (Notodontidae: Dioptinae) that have evolved from species sensitive to hearing echolocation calls in Venezuela. These diurnal species are presently in different stages of the reduction/loss of hearing: two of them have normal ears, two have reduced hearing at bat-specific frequencies and the remaining three exhibit advanced or complete loss of high-frequence hearing (Fullard et al., 1997).

From an evolutionary point of view very interesting are studies conducted for testing the hypothesis that in moths of bat-free areas, gradual decrease of sensitivity of the auditory system and, over time, deafness will evolve. Such studies have been conducted on noctuid moths in Pacific islands of French Polynesia, Tahiti and Moorea, where no gene flow from populations of bat-inhabited areas has occurred. These islands have been bat-free since they emerged 0.25-1.75 Mya (Fullard et al., 2007).

While moths that have recently immigrated to bat-free islands have normal auditory sensitivity and flight behavior, moths that have anciently migrated to these same bat-free islands initially had ears and were capable of ASR (acoustic startle response), i.e. to suddenly stop flying on detecting the presence of bats. Now, although still in possession of ears that are morphologically similar to the ears of recently arrived species, these moths have lost the auditory sensitivity, exhibit partial deafness and have lost the ASR. It is believed that the initial step in the process of the decline of the auditory sensitivity and loss of the flight interruption behavior has been “the decoupling of the sensory input (because of the absence of bats in their new habitat – N.C.) from the neural pathways that evoke behaviour”.

Neuroanatomical examinations of vestigial networks in other insects suggest that cellular events underlying this decoupling involve the sensory neurons [e.g. reduction in receptor cell terminal arborizations (Arbas, 1983a; Riede et al., 1990) and/or in the interneurons that process these inputs (Arbas, 1983b)]. Roeder (1974) proposed that the anti-bat flight defenses of noctuid moths are bimodal with 

the most sensitive auditory cell (A1) evoking controlled flight away from an approaching bat and the less sensitive cell (A2) activating the sudden erratic flight which constitutes the ASR. It is therefore possible that the extinction of ASR in Tahitian moths may be the result of a single regressive event at the level of A2 cell. (Fullard et al., 2004)

The ASR-evoking A2 neuron is not lost but is still present in the endemic Tahitian moths and the only difference of this neuron with A2 neurons of moths with normal hearing at bat-specific frequencies is that in Tahitian moths the A2 neuron has increased the auditory sensitivity set point (threshold) to ultrasounds from 25 to 30kHz so that it responds with reduced firing to the bat echolocation call stimulus, thus failing to perform the ancestral ASR behavior. Moths that arrived earlier in Tahitian islands are in more advanced stages of the process of the loss of ears (Fullard, 2007).

The decoupling of the sensory input from the neural pathways evoking the ASR, a phenomenon that is also observed in cases of the loss of flight in insects, suggests that inactivation of circuits determining specific behaviors is the first step in the process of evolutionary loss of morphological characters in metazoans and this is in line with the prediction of the epigenetic paradigm that evolution of a phenotype usually starts with changes in the behavior(s) related to that phenotype. In our particular case of moths in bat-free Tahiti islands, the loss of ASR behavior may be a prelude to an ongoing process of simplification or vestigialization of the morphology of the moth auditory system.

The fact that in numerous known and described cases, evolution of new behaviors, as products of evolution of new or modified circuits, is the first step in species recognition (see Mate Recognition System and Evolution of the Mate Recognition System in chapter 20) and evolutionary diversification suggests that the process of the decline of the auditory sensitivity and loss of flight interruption behavior in Tahiti and Moorea islands, French Polynesia, may have started with the degeneration of neural circuits. Neural circuits are the most malleable components of the auditory system. As Fullard points out:

It could also be that the most “expensive” components of a functional auditory system exist within the CNS circuits to which it connects. These circuits, and the behaviors they control, might be lost or inhibited at an earlier evolutionary stage than the cheaper peripheral sensory structures that activate them. (Fullard, 1994) 

The hearing loss in moths of the bat-free Pacific islands may be an illustration of the normal process of the regression that precedes the evolutionary loss of structures. It starts with the evolutionary loss of behaviors regulated by specific neural circuits, i.e. involves epigenetic changes in neural circuits but requires no changes in genetic information.  

NeoDarwinian Explanation of Loss of Structures and Behaviors 

In many studies it is implied or explicitly stated that the benefits from the loss of an inutile organ stimulate gradual selection for economy of energy that leads to the loss of the organ (Fong et al., 1995). While in principle he statement may be true, it only deals with the second phase of the evolutionary process alone, that is with selection and neglect the essential point of how that which will be selected arises. Just as the cart is useless without the horse, the paradigm of selection per se can make no inroads into the understanding of the origin and nature of the evolutionary loss of characters.

A neoDarwinian prediction would be that accumulation of appropriate changes at the genetic level would gradually lead to the vestigialization/loss of these structures. It may also be speculated that pseudogenes or mutations that would make genes nonfunctional or changes in introns may lead to vestigialization/loss of phenotypic characters, but, predictably, no substantiating evidence has been presented.

Validation of the neoDarwinian hypothesis that vestigialization/loss of phenotypical characters results from changes in the genetic information would require

- Identification of these changes at the gene level or at the level of gene products, and 

- Empirical evidence that these changes in genes have occurred before, not after, the vestigialization or loss of the particular structure or behavior.

In no single case of vestigialization/loss of structures, functions, or behaviors these requirements have been met; no causal relationship has been shown to exist between these phenotypic evolutionary changes and particular changes in genes or genetic information. On the contrary, recent evidence from experiments and nature on the reversion of structures that have been lost for up to tens of millions of years suggests that genes necessary for the development of these structures are generally conserved and functionally unchanged in species that have lost them. Successful experiments of induction of teeth in birds by transplanting appropriate mouse neural crest cells in the chick embryo epithelium (Mitsiadis et al., 2003; Mitsiadis et al., 2006) have proven that representatives of this class still now, ~80 million years after having lost their teeth (Chen et al., 2000), are still in possession of functionally unchanged odontogenic genes, including genes for enamel synthesis. Similarly, functional and conserved are oculogenic genes and opsin genes in eyeless cave fish.

All the cases of the loss of organs and behaviors presented in this chapter reject the neoDarwinian prediction that gene mutations, changes in allele frequencies or gene recombinations, genetic mechanisms in general, might have been involved in the loss of these phenotypic traits. 

 Epigenetic Explanation of Loss of Structures and Behaviors 

What clearly have occurred in some experimentally determined cases of the loss of structures (loss of limbs in tetrapods, loss of eyes in A. mexicanus, loss of teeth in birds, etc.) are epigenetic changes in expression patterns of specific genes and gene regulatory networks. Signals and signal cascades determining these  epigenetic changes are of neural origin and ultimately represent chemical outputs of the computational activity of neural circuits in animal brains.

Epigenetic mechanisms do not necessarily imply sudden evolutionary events. They may lead to both sudden and gradual evolution of phenotypes and natural selection also acts on the epigenetically evolving traits.