Figure 2.1. Simplified diagram of the fashioning of a relationship between the external stimulus and a particular gene. An external stimulus is received from specific neurons of the sensory organ. Receptor neurons convert the stimulus into a train of electrical spikes and, in this form, transmit it to a specific neural circuit for processing or interpreting. The mechanism of processing is still almost a black box. What we know is the output of the processing which is an electrical/chemical signal that leads to selective activation of a specific algorithm or adaptive signal cascade out of  a number of cascades (1-6) the secretory neuron can potentially activate.

 

Thus, expression of genes in the CNS is not “stimulus-dependent”, as commonly described; its expression is processing-dependent, hence non-genetic, epigenetic. Selection of the gene to be expressed is normally adaptive, i.e. it tends to accommodate the organism to the stimulus-changed environment or even to the situations that the stimulus might presage. The connection between the gene and the stimulus is communicative rather than material and is computationally determined. As pointed out earlier, the processing in the nervous system establishes a causal relationship between the stimulus and expression of the gene, which otherwise would not exist.

I call this “nonclassical” expression of genes manipulative since the expression of the gene is contrived rather than determined by chemical affinity of the stimulus for the gene; no special thermodynamic and stereochemical relationships exist between the stimulus and the gene that is expressed in response to it. It is not the nature of the external/internal stimulus that determines expression of the specific gene in the CNS: in response to different stimuli, different organism may respond by turning on/off the same gene and, in response to the same stimulus, different species may respond by expressing (or not) different genes and activating or inactivating different signal cascades.

By processing external/stimuli in neural circuits, metazoans are able to figure both the gene and the signal cascade they have to activate. The function of the processing in the neural circuit is to interpret the stimulus, to give it a “meaning”, and find a solution to the problem it may present to the organism by selectively activating a specific signal cascade out of a number of available cascades.

 

External/Internal Stimuli Provide no Instructions to the CNS

 

It is commonplace in modern biology to speak of external stimuli inducing expression of genes or activation of developmental pathways, implying. that environmental stimuli are in possession of  specific information for inducing expression of specific genes in the CNS.

As a rule, the first link in the chain of events in a process of information transmission is the source of information. Hence, the fact that the causal chain leading to expression of genes that start signal cascades begins with external/internal stimuli arises the question whether the stimuli themselves may contain information for gene expression, despite the Lamarckian implications of recognition of an instructing role of the environment in metazoan development and evolution. Theoretical contemplations aside, the examples of expression of genes under the influence of external and internal stimuli to be presented in the next section show that none of them per se does or can induce gene expression.

Let’s consider the example of the switching off/on of the genes for melatonin biosynthesis by the pineal gland under the influence of the photic stimulus (day-night cycles) (see figure 2.2 and details in the next section). If day-nightcycles per se would be information for gene expression, it would be expected that dermis cells, which are directly affected by the stimulus (light/darkness), not pineal cells which are not, would synthesize melatonin. As we know, the opposite takes place in reality: dermis cells do not express those genes, while pineal cells, to which the stimulus has no access, do. Moreover, retinal neurons, which receive the stimulus and transmit it to the CNS, do not express genes for melatonin biosynthesis. The fact that the stimulus fails to induce expression of melatonin genes in dermis cells or retinal cells, which are directly affected by the stimulus, and induces expression of  these genes in pineal cells to which it has no access, excludes the possibility that the stimulus itself, the external agent, might represent information for expression of the melatonin genes. Similarly to all the experimentally and observationally determined cases, in the case of the night-day cycles, for instance, retinal neurons code the photic stimulus in the form of electrical spike trains. This initial input about the stimulus is transmitted for processing to the neurons of the specific circuit, what results in release of specific signal(s) (neurotransmitters?), which represents epigenetic information for activating melatonin genes in pineal cells. Thus, the information for expression of melatonin genes is generated in the process of computation of the retinal electrical signals in the neural circuit. The epigenetic information, in the form of the chemical output resulting from the processing of the stimulus in the neural circuit, is released on the pineal cells where it acts as a signal for turning on/off the genes for melatonin synthesis. Both the processes, the conversion of the photic stimulus into specific electrical spike trains and the processing in the melatonin neural circuit are computational, non-genetic processes. Hence, the information transmitted from the neural circuit to the pineal cells for inducing/suppressing genes for melatonin synthesis generated in the melatonin neural circuit and is epigenetic in origin.

Indeed, if external stimuli (light, darkness, temperature, humidity, crowding, other social stimuli, etc.) would provide information for selective activation of signal cascades, the whole costly processing of the stimuli would serve no purpose. Elementary knowledge on the way evolution works suggests that costly neural processing would have not evolved if it would not generate a benefit outweighing the cost.

The fact that the information for selective activation of signal cascades appears first after the processing of the stimuli in the central nervous system unambiguously suggests that the very purpose of the processing is to interpret the stimulus, which per se is “meaningless”, so that it becomes “intelligible” to particular gene(s) (out of thousands of genes in the genome), in particular groups of cells that receive that neurally generated information.

 

Epigenetic or Genetic? On the Nature of Information for Signal Cascades

 

The fact that signal cascades start in the CNS suggests that it is there where the information necessary for their activation originates. However, it might be argued that any talk about the CNS control of gene expression is irrelevant since, ultimately, brain signals that start signal cascades are genetic in origin. Is this really so?

Any dispute on the nature of information for selective activation of signal cascades in the CNS cannot be resolved by theoretical arguments. It can only empirically be determined by a closer examination of activation of signal cascades under the influence of external/internal stimuli in some paradigmatic examples of the so-called stimulus-dependent expression of genes in the brain.

Example 1. Most wild birds, from temperate to arctic zones are reproductively active, i.e. ovulate and lay eggs, in the spring time. The timing of the reproductive activity in these birds is clearly adaptive since spring is the best time for rearing the offspring. The beginning of the reproductive activity at that time of the year is mainly related to the prolongation of photoperiod (length of the day) and rise of temperature in the environment. Although the reproductive activity starts with expression of hypothalamic GnRH (gonadotropin-relesing hormone) and activation of the hypothalamic-pituitary-gonadal axis, neither the gene for the hypothalamic GnRH nor the gene for the pituitary gonadotropins or for gonadal hormones can be induced by the direct action of thermal or photic stimuli. In homeotherm animals, such as birds, neither the ovary/testis nor the egg cell/sperm cell (and less so their genes) receive those external stimuli. Nevertheless, birds perfectly respond to these stimuli by a precise spatio-tempporal activation/inactivation of the genes along the hypothalamic-pituitary-gonadal axis, inducing expression of thousands of genes in gametes as well as neighboring relevant cells (follicle cells, Sertoli cells, etc.), related to the preparation of the genital organs for gametogenesis, mating, and oviposition. Both the temperature and photic stimuli are “unintelligible” to genes, which generally do not  respond to the physical contact with these stimuli.

The fact that these genes are activated, despite being inaccessible to the stimuli, would be a paradox if we would not know what goes on “behind the scene”. The temperature is perceived in the hypothalamic POA (preoptic area), in a center containing warm-sensitive neurons and cold-sensitive neurons whereas the length of the day is sensed in another hypothalamic center, the SCN (suprachiasmatic nucleus). The perception of the temperature and day length in these brain centers comprises reception of the electrical input on the temperature and length of the day and processing of these stimuli in respective neural circuits. As a result of these computational processes neural circuits generate chemical outputs (=epigenetic information) for starting signal cascades that determine activation of hundreds and thousands of specific genes in the genital tract and the organism in general during the reproductive season.

One of the ways the photoperiod is used for the reproductive timing in vertebrates is by neural inhibition of melatonin biosynthesis and secretion in the pineal gland: melatonin injections in photoresponsive rats even during a stimulatory (long) photoperiod induce reproductive inhibition (Heideman et al., 2001).

The photic stimulus (shortening of the photoperiod) in the pineal gland induces expression of genes for the synthesis of melatonin, although no light, no photons reach those cells.

The suppression/induction of melatonin results not from a direct action of  the photic stimulus on the gland, but from the processing of the photic information in neural circuits, i.e. not from a genetic but from a computational epigenetic process. The photic stimulus is received by retinal photoreceptors, sent to the SCN (suprachiasmatic nucleus) in the form of electrical signals. By processing the photic stimulus (Larsen et al., 1998), the SCN (suprachiasmatic nucleus) sends to the PVN (paraventricular nucleus) a combination of stimulatory and inhibitory circadian signals (Perreau-Lenz et al., 2003) and the latter sends to the pineal signals for starting the synthesis of melatonin (figure 2.2 ).

 

 

Example 2. In response to seasonal cues and to a social auditory stimulus, the conspecific male mating call, female green treefrogs increase the release of the hypothalamic GnRH (gonadotropin releasing hormone). The auditory system of the  tree frog projects to the ventral hypothalamus (VH) and preoptic area (POA) (figures 2.3 and 2.4).

 

 

Figure 2.3. Primary projections of the ascending auditory system of anurans. The POA also receives projections from CT and SI, although these are less robust than those observed for VH.

Abbreviations: CT, central thalamus; DL, dorsolateral nucleus; TS, torus semicircularis; P,  pituitary; POA, preoptic area; SI, secondary isthmal nucleus; SO, superior olivary nucleus; VIII n., eighth cranial nerve; VH, ventral hypothalamus (From Burmeister and Wilczynsky, 2005).

 

Investigators have concluded that experimental results

 

Provide functional evidence for such a sensory-endocrine circuit by showing that acoustic signals influence GnRH neurons. (Burmeister and Wilczynski, 2005)

 

Obviously auditory signals per se cannot induce expression of any gene; it is the informational output of the processing of acoustic signals in the CNS, released in the form of a neurotransmitter/neuromodulator that via its receptors on the hypothalamic GnRH neurons activates a transduction pathway inducing  expression of the

GnRH gene. In turn GnRH, via the pituitary-gonadal axis, starts the signal cascade that stimulate the reproductive activity in female treefrogs.

Clearly, expression of GnRH gene in response to auditory stimuli is not genetically but epigenetically determined by the processing of auditory stimuli (species-specific mating calls).

 

 

Figure 2.4. A proposed model for regulation of the anuran hypothalamus-pituitary-gonad (HPG) axis during the breeding season. In the model, seasonal cues initiate the annual reactivation of the GnRH neurons which are further stimulated during the breeding s season by hearing mating calls. GnRH neurons regulate androgen levels through their action on gonadotropin (GT) release from the pituitary. Androgens, in turn, exert negative feedback on the HPG axis. Androgens also influence brain regions that increase the production of mating calls, creating a positive feedback loop between the communication system and the endocrine systemocrine system (From Burmeister and Wilczynsky, 2005).

 

Example 3. Detection of chilling in the Hawaiian cockroach Diploptera punctata, and in other insects, is function of antennal thermoreceptors, which send chilling signals to the pars intercerebralis and pars lateralis of the insect protocerebrum. Chilling experienced by the antennae has no specific effect on gene expression on the antennae but it, triggers a “long-range” action on another organ that is not directly affected by the chilling; it causes suppression of mitotic divisions in corpora allata, which precedes the increase in the volume of corpora allata and secretion of juvenile hormone by the glands of mated females (figure 2.5). Unilateral ablation of antenna does not prevent the contralateral corpus allatum from proceeding normally with the wave of mitotic divisions, while the mitotic divisions in the ipsilateral gland are suppressed. Unilateral disconnection of neurons of the pars intercerebralis of the insect brain from corpora allata suppresses the effect of chilling of antennae on the mitotic division of the contralateral corpus allatum, clearly demonstrating that a neural, not genetic, mechanism determines the suppressing effect of chilling on cell division in corpora allata. The information on the chilling of antennae is transmitted in the form of electric signals, perceived and processed in the pars intercerebralis of the insect’s brain, which send signals for suppressing the mitotic division in corpora allata (Pszczolkowski and Brown, 2003).

In another experiment it has been demonstrated that

 

Severance of the ventral nerve cord prior to chilling or to application of 20E (hydroecdysone) prevented suppression of CA cell division, indicating that effects of either chilling or 20E application are mediated by the ventral nerve cord. (Pszczolkowski and Gelman, 2004)

 

Example 4. The sound energy of the song of the male canary, Serinus canaria, in the form of mechanical waves is transformed into mechanical energy in the tympanum and transmitted through the middle and inner ears before it reaches the auditory receptor neurons, which transform it into electrical spike trains transmitted for processing in the auditory area of the forebrain. The sound waves induce no specific song-related expression of genes along the auditory pathway (tympanum, the outer, middle, and inner ear). The specific song-related expression of the transcriptional regulatory gene ZENK in female canaries occurs only in the auditory area of the forebrain after the processing of electrical signals (Jarvis and Nottebohm, 1997; figure 2.6) but not in the cells that are directly affected by sound waves of singing. Not only hearing of the conspecific song but their own singing in male canaries induces a 60-fold increase in expression of the transcriptional regulatory gene ZENK in the high vocal center and other song nuclei. At a histological level, the singing is associated with an increased electrical activity in these same centers (Jarvis and Nottebohm, 1997). A similar vocalizing-induced increase in expression of ZENK gene is also observed in a number of hummingbird species (Jarvis et al. 2000).

A recent study has shown that in caudomedial nidopallium (NCM), song-induced Zenk expression colocalizes with c-fos, c-jun and Arc expression. Arc is especially expressed in recently activated synapses. Mediators of the song-regulated  expression of genes are Ca²⁺ kinases, protein kinase A, and MAP kinase as well as MEK1/2(Velho et al., 2005; figure 2.7).

 

Figure 2.5. Schematic frontal view of Diploptera punctata protocerebrum depicting the location of neurons controlling corpora allata (CA) cell proliferation.

Abbreviations: PI, pars intercerebralis neurons, PL 2, pars lateralis neurons 2. Both PI and PL 2 neurons project to the contralateral gland (From Pszczolkowski and Brown, 2003).

 

 

Figure 2.6. Sagittal diagrams of songbird brain. (a) Anatomical relations between song nuclei (white) in which ZENK expression was induced by the act of singing. Black arrows show the direct motor pathway that innervates the vocal organs and produces learned song. Shaded arrows show links between other song system nuclei. (b) Anatomical relations between forebrain auditory relays (white) in which ZENK expression was induced by hearing song. Thick black arrows show the major ascending auditory pathway; thin black and gray arrows show some of the connections between forebrain auditory regions and the channeling of this input into HVC.

Abbreviations: HVC, high vocal center; NCM, the caudomedial neostriatum; DLM, medial nucleus of dorsolateral thalamus; RA, the robust nucleus of the archistriatum; MAN, magnocellular nucleus of the anterior neostriatum; X, area X of lobus parolfactorus; cHV, the caudal hyperstriatum ventrale; Av, nucleus avalanche; nXIIts, tracheosyringeal portion of the hypoglossal (12th) nucleus; Ov, nucleus ovoidalis  (From Jarvis and Nottebohm, 1997).

 

Example 5. Consumption of an unfamiliar taste in rats imprints a long-term memory on that testant, epigenetically encoded in the central gustatory area of the insular cortex (IC), which modulates expression of genes there, although gustatory stimuli do not reach these genes. This is result of a complex input of multiple neurotransmitters leading to the activation of ERK1-2 and the novelty-dependent gene expression (Berman et al., 2000; figure 2.8).

Example 6. This is an example of an internal stimulus that induces expression of a gene without having access to the cells that express the gene in the CNS. In ewes and other mammals, oestradiol is critical for the preovulatory LH (luteinizing hormone) surge. In non-neural cells, oestradiol forms complexes with its nuclear receptors and, in this form, it binds regulatory regions activating “estrogenic genes” (cyclin D (1), igf-1, etc.).  Nothing of the kind, no classical expression of “estrogenic genes”, has been observed so far in the hypothalamus, and there is no evidence whatsoever that oestradiol directly regulates GnRH neurons (Smith and Jennes, 2001).

Most of GnRH neurons do not even express nuclear estrogen receptors, mediators of the action of estrogen, and this is the reason why even the implantation of estrogen in regions of GnRH neurons does not induce LH surge (Herbison, 1998).

It is suggested that the steroid signal is conveyed to the GnRH neurones by a subset of neurones that are regulated by oestradiol. (Eyigor et al., 2004)

 

Figure 2.7. A song-induced gene expression program in NCM neurons. The activation of song-responsive auditory neurons in NCM activates intracellular signaling events (black arrows), including components of the MAP kinase pathway (MEK1⁄2 and ERK), that lead to the rapid induction of activity-dependent genes encoding early effectors and transcription factors, collectively known as immediate early genes (IEGs). Early transcription factors (e.g. zenk, c-jun, c-fos) act by regulating the expression of downstream targets that encode late effectors (gray arrows). Early effectors, by contrast, can have a more rapid effect on cell physiology. Arc (dashed grey lines) has a dendritic localization and is thought to act at the level of activated synapses.

Abbreviations: ERK, extracellular signal-regulated kinase; MEK1⁄2, mitogen-activated and extracellular signal-regulated protein kinase 1 and 2; NCM, caudomedial nidopallium (From Velho et al., 2005).

 

Figure 2.8. A highly simplified scheme of some elements in the processes that might subserve the encoding of taste memory in the IC. In the absence of taste input, the balance between the resting levels of glutamate (Glu; acting via AMPA/KR and mGluR), GABA, and dopamine (DA; via D1/D5 receptors) regulates the basal level of ERK1–2 activation. When a new taste is consumed, a hypothetical novelty detection system, e.g., thalamocortico-brainstem circuits, compares the on-line stimulus with taste representations in memory. According to this model, if a meaningful mismatch is detected, a signal is sent to the cholinergic and noradrenergic systems, resulting in release of acetylcholine and noradrenaline in the IC. Acquisition involves, in addition to glutamatergic transmission via the metabotropic and AMPA receptors, glutamatergic transmission via the NMDA receptor, as well as cholinergic and noradrenergic input. Acetylcholine activates ERK1–2, whereas noradrenaline β functions in this system independent of ERK1–2. Activation of ERK1–2 culminates in modulation of gene expression and ultimately in long-term representational changes.

Abbreviations: IC, insular cortex; ERK1–2, extracellular signal-regulated kinases 1–2; NMDA, N-methyl-D-aspartic acid (an amino acid derivative) (From Berman et al, 2000).

What acts on the hypothalamic GnRH neurons is not the estrogen but the chemical output of the processing of estrogen in estrogen-sensitive neurons in other parts of the brain and the hypothalamus. The rise in the oestradiol level is received as a stimulus and converted into an electrical signal by non-GnRH neurons, processed in estrogen-sensitive neural circuits inducing release of stimulating neurotransmitters [dopamine by the neural circuit of the ventromedial preoptic area (Anderson et al., 2001; Goubillon et al., 2002) and NA (noradrenaline) by a brain stem noradrenergic circuit (Scott et al., 1999)] and inhibitory neurotransmitters [GABA by a circuit in the rostral hypothalamus, and neuropeptide Y by two circuits in the brain stem and arcuate nucleus)]. It is the informational input generated by processing of the estrogen stimulus (a perceptible change in the level of the hormone) rather than estrogen itself that induces/suppresses expression of the GnRH gene (Herbison, 1998; Smith and Jennes, 2001; figures 2.9 and 2.10; Pompolo et al., 2003).

Numerous examples of manipulative expression of genes in response to external cues are also described in invertebrates. So, e.g., changes in the photoperiod cause the nerves originating in the subpedunculate lobe of the brain of cephalopods to stimulate or inhibit secretion of a gonadotropin by their optic glands, thus inducing or preventing the gonadal enlargement (Strand, 1999i).

 

Figure 2.11.  Calcium-activated signaling pathways that regulate gene transcription. In neurons, neurotransmitter reception and membrane depolarization lead to the opening of ligand- and voltage-gated calcium channels. Subsequent calcium influx across the plasma membrane drives the activation of a number of signaling molecules, including the calcium-sensitive adenylate cyclase, calcium/calmodulin-activated kinases, and Ras. Each of these molecules activates a cascade of signaling proteins that amplifies the calcium signal and carries it to the nucleus. Dashed lines represent the components of each pathway that are proposed to translocate into the nucleus. Nuclear kinases, including protein kinase A, CaMK-IV, and members of the Rsk family phosphorylate CREB at Ser-133, rendering it competent to mediate transcription of genes such as BDNF (brain-derived neurotrophic factor).

Figure 2.13. ERK pathway in the visual cortex. Neurotrophins and electrical activity modulate ERK through two separate pathways that converge on the ERK kinase MEK. The neurotrophin signal is conveyed from the tyrosine kinase receptors (trk) to Ras by the Shc-Grb2-sos intermediates (S-G2-s), leading to activation of ERK and CREB. Electrical activity is translated in two different intracellular signals: (1) an increase of Ca²⁺ caused by influx through voltage-gated channels and NMDA receptors, and (2) an increase of cAMP attributable to activation of metabotropic receptors (mR) or by Ca-dependent adenylate cyclase (AD). The cAMP-dependent activation of ERK is mediated by cAMP-GEF and CNrasGEF, whereas the direct Ca-dependent activation of ERK is primarily mediated by the Ca-dependent regulating factor RasGRF. The CRE promoter controls the expression of a large number of genes including BDNF (brain-derived neurotrophic factor).

Abbreviations: NT, Neurotrophins; AC, adenylate cyclase (From Cancedda et al. 2003).

 

Epigenetic Manipulation of Genetic Information in the CNS

Besides the manipulative expression of genes described so far, the CNS exhibits another equally, if not more, impressive capability, generally known as gene splicing, that I would characterize it as manipulation of genetic information, implying production of different proteins by selectively expressing not whole genes but specific segments of genes, thus enabling formation of numerous protein isoforms by the same gene.

From a theoretical point of view this defies a basic tenet of the classical genetics on the gene as carrier of information for a protein, RNA, or polypeptide. It seems to be more correct to say that the metazoans use the genetic information contained in a gene for manipulatively producing a varying number of proteins.

Neural wiring specificity in Drosophila melanogaster is thought to be determined by the fact that each neuron is able to produce a great number of specific protein isoforms. So, e.g., in the fruit fly, the Dscam (Down syndrome cell adhesion molecule) gene encodes a transmembrane receptor protein necessary for axon guidance. The role of the Dscam protein in axon guidance is proven by the fact that mutants for this protein die early during the larval stage and show defects in the ventral nerve cord development (Celotto and Graveley, 2001). The Dscam gene alone can potentially generate 38,016 different protein isoforms (Celotto and Graveley, 2001; Graveley et al., 2004; Kreahling and Graveley, 2004), i.e. almost twice as much transcripts than the total number of genes in the whole human genome.

Certainly, pre-mRNA splicing, excision of introns and assemblage of exons into a translation-ready mRNA, is a constitutive process that takes place in cells all over the animal body rather than in the CNS alone. However, in our context, it should be kept in mind that

Firstly, only eukaryotes and no prokaryote unicellulars are capable of m-RNA splicing.  Secondly, unlike the constitutive splicing, occurring in cells throughout the animal body, manipulative splicing is a characteristic of gene expression in the CNS.

In this work, I will use the term manipulative splicing instead of the common designation of the phenomenon as alternative splicing because it better expresses the process of the adaptive specification of pre-mRNA out of the great number of alternatives (in a standard English dictionary the verb manipulate is defined as follows: “to work with, or change information, systems, etc. to achieve the result that you want”).

Manipulative splicing, as a specialty of the nervous system, allows metazoans to adaptively produce different variants of proteins by the same gene through the combination of specific exons into the mature mRNA. Manipulative splicing is a widespread mechanism of generating protein diversity in metazoans. It is used for expression of more than one third of human genes (Hanke et al., 1999) and it may be the main mechanism for generating the great number of protein isoforms necessary for complex brain functions (Lipscombe, 2005).

Manipulative pre-mRNA splicing is regulated by extracellular signals, and is controlled in a tissue-specific manner (Celotto and Graveley, 2001), implying that it is an epigenetic phenomenon. Although manipulative splicing is observed in other organs (and the possible role of the local innervation in these organs is not incestigated), it is the central nervous system where this form of splicing is extensively used:

 

Alternative splicing in neurons was first discovered in the CA/CGRP gene encoding two peptides that differ in the terminal exon: calcitonin in thyroid cells and CGRP (calcitonin gene-related peptide) gene in specific neurons. Since then, alternative splicing has been found to be a common mechanism used in the generation of isoforms in a variety of functionally important neuronal genes. Although the biological significance of much of the alternative splicing is not always readily evident, in many instances it is thought to be important in the fine tuning of neuronal function. (Lisbin et al., 2001)

 

Manipulative splicing in the nervous system is regulated by a neuron-specific system of splicing. The electrical activity seems to be at the basis of production of different protein isoforms in the CNS (Mu et al., 2003;  Lipscombe, 2005). Calcium elevation resulting from the electrical activity induces specific changes in splicing and translation processes. Among the brain specific RNA binding proteins regulating manipulative splicing in metazoans are Nova, Fox1, Fox2, NaPOR, ELAV, etc. Some typical examples of manipulative splicing leading to formation of multiple protein isoforms involve  proteins like neurexins, STREX, Slo (Slowpoke), etc.

Neurexins are surface cell proteins encoded by three genes, but expressed exclusively in the brain (Ushkaryov et al., 1992). In the brain these genes are subject to intense manipulative splicing for producing specific neurexin isoforms that are involved in synaptogenesis as it has been concluded from the fact that their expression changes during synaptic remodeling (Gorecki et al., 1999). Estimates based on experimental work with expression of neurexins in various parts of the brain have shown that between 600 and 3000 distinct splicing-generated neurexins are produced by three neurexin genes in the brain of the developing Xenopus laevis (Ullrich et al., 1995).

The gene ania-6 is transcribed into two distinct mRNAs (long ania-6 mRNA and short ania-6 mRNA) coding for 2 different proteins via differential splicing in the process of their transcription. These proteins belong to a new family of cyclins. Sgambato et al. (2003) have shown that production of the long or short splice variants of ania-6 depends on the intracellular pathway: the long variant is produced when CaMK (calmodulin kinase) pathway is used and the short splice variant results from the activation of the ERK (extracellular signal regulated kinase) pathway. Which of these two pathways will be used is determined by the neurotransmitters and neurotrophins that are released from the presynaptic neuron, which, in turn, are computationally determined by processing the input of the electrical signals: release of glutamate activates CaMK pathway, while BDNF (brain-derived neurotropic factor) activates ERK pathway.

Experimental administration of neurotransmitters glutamate or dopamine, by inducing differential splicing of exons, leads to production of two different proteins by gene ania-6 (Berke et al. 2001). The role of neurotransmitters in determining the type of protein isoforms that are produced by splicing, in view of the neural epigenetic mechanisms of neurotransmitter synthesis and secretion, suggests that an epigenetic mechanism is responsible for the manipulative expression of protein isoforms in the nervous system.

Most of the proteins involved in manipulative splicing are neuron-specific. So, e.g., Nova-1 is a brain specific splicing factor involved in manipulative splicing of the pre-mRNA for glycine 2 exon and GABA_A exon γ2L receptor (Lisbin et al., 2001). Drosophila melanogaster ELAV, a product of the gene elav, is specifically expressed in all neurons and all the members of the ELAV family, with only one exception, are exclusively expressed in the nervous system. ELAV is implicated in the manipulative splicing of transcripts of genes nrg (neuroglia), erect wing, and armadillo resulting in production of neural-specific protein isoforms.

For a long time biologists wondered (and still do) how the synapses might generate the great molecular diversity necessary for recognizing each of their potential target neurons out of the myriad of neurons in the central nervous system. Now we know that may be enabled by the manipulative splicing, which generates numerous protein isoforms (with different binding affinities) from a single gene. So, e.g., it is found that three neurexin (nrxn) genes in mammalian neurons potentially encode over 1,000 isoforms of neuron membrane proteins (Zeng et al., 2006).

Some leading investigators in the field of the pre-mRNA splicing believe that by investing in the direction of generating greater numbers of protein isoforms in neurons, animals have increased chances that each neuron will produce protein isoforms that are different from neighboring neurons, enabling neuron’s dendrites and axons to distinguish between themselves and their neighbors. A more stringent regulation would require a tremendous investment of genetic regulation (Michalowski,  2005).

In view of the tremendous differences in morphology and function of neurons, manipulative splicing is very important for the development and function of the nervous system. As an illustration of this role of manipulative splicing would serve the example of the mechanism of hearing in chicken. Hair cells are mechanosensory cells that communicate with dendrites of efferent neurons by releasing (and receiving) neurotransmitters via calcium channels. Hair cells within cochlea, exhibit widely varying properties according to their position along the basilar papilla (figures 2.14, 2.15).

Figure 2.18.   Model of possible mechanisms of spine induction on short primary dendrites in the ventrolateral VMH (ventromedial hypothalamic nucleus). The effect of estrogen on cells lacking estrogen receptor suggests that estrogen acts transynaptically, rather than directly, to induce dendritic spines. The afferents that stimulate spine formation may be either extrinsic or intrinsic to the VMH. Extrinsic afferents are mainly found in the neuropil surrounding the VMH. Long primary dendrites extending to the ventrolateral border of the VMH may be innervated by extrinsic afferents in the neuropil. Intrinsic afferents may innervate short primary dendrites. Estrogen affects local intrinsic afferents to increase spine density on short primary dendrites. Open circles indicate excitatory synapses; arrows indicate direction of information flow through the circuit.

Reconfiguration of synaptic morphology, in turn, may change computational properties of neural circuits. Such is the case, e.g., in CHL1-deficient mice where synaptic reconfigurations leads to serious consequences in the information processing and computational properties of neural circuits (Montag-Sallaz et al., 2003).

There is a song-learning circuit in male zebra finches that connects DLM (medial dorsolateral nucleus of the thalamus) and lMAN (cortical lateral magnocellular nucleus of the anterior neostriatum) (Iyengar and Bottjer, 2002). This circuit underlies song learning from tutors by male birds. It is believed that DLM/lMAN circuit functions by detecting errors in the process of comparison of male’s own motor output with tutor’s song. During the sensitive period of vocal learning, between days 20 and 35, the shell subregion of lMAN and the DLM terminal field experience a three-fold increase in the volume that is followed by a sharp decrease between the day 35 and adulthood. The dramatic changes in the volume of the above structures related to the song learning reflect dynamic rearrangements of expansion and retraction of axonal structures rather than changes in the number of individual axons. Large scale experience-dependent pruning of axon branches occur in the course of song learning:

Figure 2.19. High-power camera lucida drawings of dendritic segments of dentate granule cells of control (A) and TTX-treated (B) entorhinohippocampal cocultures. All examples show dendritic segments with spines of various size and shape. Note the presence of long, filiform spines lacking a head (arrowheads) in controls, and their greater abundance in TTX-treated cultured slices. TTX, tetrodotoxin  (From Drakew et al., 1999).

The changes in complexity and spatial extent of DLM axon arbors described in this study indicate synaptic remodeling within the DLMàMAN circuit, which may represent structural correlates of vocal learning. For example, arbor regression may represent a morphological correlate of the auditory tuning of lMAN neurons to the bird’s own song and the decreased involvement of lMAN neurons in vocal learning. (Iyengar and Bottjer, 2002)

 

Let’s remember that song learning, as a learned behavior, as a phenotypic character takes place in the specific circuit in the bird’s brain. Singing a species-specific song in zebra finches implies the presence of specific information for performing the song. There is only one rational answer, that I can think of, to the question: Where that information is stored? The observed correlation between the changes in synaptic morphology and the song learning suggests that the information for performing the song is encoded in the specific synaptic morphology that develops in the brain song circuit of male zebra finches in the process of song learning.

A problem related to the information that may be encoded in the synaptic morphology is the mechanism by which the specific synaptic morphology, which changes easily under the influence of activity, is maintained for determined periods of time. It is believed that neural circuits have their own homeostatic regulatory mechanisms for stabilizing and maintaining their morphology under circumstances of increased or decreased activity (Turrigiano and Nelson, 2004; figure 2.20).

In experiments it has been shown that reconfiguration of synaptic connections and the resulting changes in computational properties of circuits induce specific changes in the behavioral or morphological output of the circuits (Getting, 1985). A number of investigators have reported cases of reconfiguration of synaptic morphology in response to changes in the embryonic structure during the individual development (Matsumoto and Arai, 1986; Kent and Levine, 1993). The transformation of the Bauplan during the metamorphosis in anuran amphibians is accompanied by a “complete reorganization” of neural circuits (Alley, 1990), which also implies reconfiguration of their synaptic morphology.

The synaptic morphology remains unchanged in absence of appropriate stimuli, i.e. as long as the perception of the environment remains unchanged. The reconfiguration of synaptic morphology which is function of large-protein signal-processing machines at the postsynaptic membrane (Kennedy, 2000) may modify the computational properties (von der Malsburg, 1999; von der Malsburg, 2002b) of the neural circuit. It is believed that changes in the structure of the postsynaptic membrane represent the fundamental mechanism for the processing and storage of information for learning and memory in the brain (Luscher et al., 2000).

Figure 2.20. Homeostatic regulation of the excitation–inhibition balance in cortical networks. Activity in recurrent cortical networks is strongly affected by feedback excitation and inhibition. Pyramidal neurons (white) make excitatory outputs (triangles) onto other pyramidal neurons, and also onto inhibitory interneurons (black). These inhibitory neurons in turn feed inhibition (black circles) back onto the pyramidal neurons. In cortical cultures, raising activity for two days produces a coordinated set of changes in synaptic strength that result in reduced feedback excitation and increased feedback inhibition onto pyramidal neurons (lower left). Conversely, blocking activity for two days increases the gain of excitatory feedback and decreases inhibitory feedback. Similar changes in the cortical excitation–inhibition balance are induced by sensory deprivation (From Turrigiano and Nelson, 2004).

 

 

Sometimes, neural networks of different configurations can produce very similar outputs (Prinz et al., 2004; Hooper, 2004). Such is the case with the pyloric rhythms of the crustacean stomatogastric ganglion: by using the most different combinations of synaptic strengths and neuron properties investigators succeeded in obtaining the same pyloric rhythm (Prinz et al., 2004). Other times, the same crustacean pyloric network, in response to different modulators, can produce different rhythms.

It is tempting to believe that this plasticity enables neural circuits to generate different and even opposite outputs by processing the same stimulus. That this is really the case is suggested by a number of well-established facts. So, e.g.,  in response to increased photoperiod cows respond by increasing reproductive activity, while the reproductive season in sheep starts with the shortening of the photoperiod.

The fact that, in principle, neural circuits are capable of relating any external stimulus to any signal cascade implies that they generate new information for developing new phenotypic characters. This neural information is generated in adaptive responses of the nervous system to the internal/external stimuli.

To say that neural circuits by modifying their synaptic connections and computational properties, release alternative chemical signals, and activate diverse signal cascades, which lead to alternative morphological results, is another way of saying that these circuits are generators of epigenetic information for animal morphology. Thus, in the course of evolution, the nervous system emerged as the generator of the huge amount of epigenetic information for evolution of metazoan phenotype.

The ability of metazoans to generate epigenetic information for the supracellular structure by reconfigurating synaptic morphology will not come as a surprise for a CNS that is able to similarly generate and store knowledge [“as is the reigning opinion, neural connections are the repository of knowledge in the brain” (von der Malsburg, 2002b)]. Far from a theoretical speculation, this is an established fact. So, for example, it has been determined that reconfiguration of synaptic morphology of a circuit in the hypothalamic suprachiasmatic nucleus “codes” the information for day length in rats (Schaap et al., 2003).

Embryonic  stage	Development of neural circuits	Processes driving formation of neural circuits	Source of information
Pre-phylotypic stage	Primitive neural circuits develop	Activity-independent and experience-independent processes	Prenatally provided epigene- tic information
Post-phylotypic stage (critical periods)	Refinement of connections and sculpting of new neural circuits	Activity-dependent (spontaneous or sensory) processes	Neurally generated epige- netic information in response to internal sensory input
Post-natal         development	Further refinement of neural connections	Activity-dependent and experience-dependent processes	Neurally generated epigene-tic information in response to external sensory input

 

The overwhelming majority of these connections are determined by computational activity of the CNS during the individual development in response to the input of internal and external stimuli. This is demonstrated by the fact that, in experimentally verified cases, blockage of afferent input in embryos prevents formation of neural circuits and dendritic spines (Katz and Shatz, 1996; Penn et al., 1996; Segal et al., 2003).

It is interesting to observe that after the phylotypic stage, when the reserve of the parental epigenetic information (parental cytoplasmic factors) is consumed, the embryonic development proceeds at an accelerated pace. Rapidly, all the tissues and organs begin to take shape.

At first sight, this explosive development at a time when the parental epigenetic information is exhausted seems to be paradoxical. Where that tremendous amount of information necessary for histogenesis and organogenesis, for erecting the immensely complex metazoan structure after the phylotypic stage might come from? The fact that exactly at this juncture, when parentally provided epigenetic information is exhausted, the embryonic CNS becomes functional, i.e. starts electrical activity, may not be a sheer coincidence.

Speculations aside, is there evidence that the CNS generates information for the postphylotypic development? From the phylotypic stage on, the CNS becomes functionally active; a continuous input of internal stimuli (changes in the rapidly developing embryonic structure) via afferents is communicated to the CNS. The embryonic CNS responds to that input by “spontaneous” electrical activity and by establishing specific neural circuits (Peinado, 2000; McAllister, 2000; Zhang and Poo, 2001).

It is not always clear what drives the spontaneous activity in the embryo, but in the spinal cord at least, spontaneous burstings of ~1 minute duration are followed by inter-episode pauses of minimal electrical activity. These “episodes” are preceded by motoneuron firing and experimental stimulation of motoneurons also induced R- (from Renshaw cells) interneurons in the process of starting an episode (Wenner and O’Donovan, 2001). This activity occurs spontaneously without any descending or afferent input with motoneurons playing a critical role in generating spontaneous activity (Hanson and Landmesser, 2003).

For example, the precise specification of projections from the retina to the lateral geniculate nucleus starts during embryogenesis at a time when there is no visual input from the environment and although the retina still cannot respond to photic stimuli, the retinal ganglion neurons show spontaneous activity (Penn and Shatz, 1999).

Adequate evidence suggests that “spontaneous electrical activity is responsible for sculpting circuits on the basis of the brain’s “best guess”” (Katz and Shatz, 1996), i.e. computationally, and that activity might represent a “self-organizing property” (Weliky, 1999). In the embryonic retina, e.g., this activity “can produce highly stereotyped patterns of connections before the onset of visual experience” (Penn et al., 1998), and “instruct” formation of eye-specific layers  (Shatz, 1996). By contrast, in absence of afferent excitatory input, normal neural circuits are not formed (Katz and Shatz, 1996; Penn et al., 1998)

and rat striatal neurons do not develop dendritic spines (Segal et al., 2003). Experimental deafferentiation causes retraction of postsynaptic dendrites and a particular level and/or pattern of afferent activity is necessary for postsynaptic dendritic morphology to develop.

 

Activity-dependent structural changes in postsynaptic cells act together with changes in presynaptic axonal arbors to shape specific patterns of connectivity in the nervous system. Thus, the growth of dendrites is a dynamic process influenced by, and integral to, the formation of connections in the nervous system. (McAllister, 2000)

 

The fact that in response to the input of afferent stimuli (changes in the developing embryonic structure) neural circuits modify their synaptic morphology suggests that synaptic connections are somehow related to the developing embryonic structure. Indeed, experimental delay of muscle development causes suspension of synaptic branching of respective motoneurons (Fernandes and Keshishian, 1998). Male rats castrated on day 1, implying that they do not receive testis input, have significantly reduced numbers of shaft and spine synapses in the ventro-lateral part of the VMH (ventromedial hypothalamus) (Matsumoto and Arai, 1986), and removal of ovaries, causes a profound decrease in the dendritic spine density of pyramidal cells in the hippocampus.

Xenopus laevis receives and processes the optic stimuli on its background color via the retinal-hypothalamic circuit. When in light-colored background, the SMIN (suprachiasmatic-melanotrope-inhibiting neurons) in the ventrolateral part of the hypothalamic SCN (suprachiasmatic nucleus) send inhibitory signals to the melanotrope cells in the pituitary, preventing secretion of melatonin, thus determining a lighter body color. In these cases, at the level of the retino-hypothalamic circuit, the clawed frog shows reduced innervation and synaptic contacts in SMINs, which results from retraction and degeneration of synapses that are present in dark background-adapted frogs (Kramer et al., 2002). This modification of the synaptic morphology seems to be causally related, via pituitary secretion of melanotropin, with the adaptive change in the body coloration of Xenopus laevis.

As shown, all the process of Xenopus background adaptation is epigenetically regulated.

It has been shown that the reconfiguration of synaptic morphology (neuronal plasticity) of the retino-hypothalamic circuit, taking place in response to the visual input on the environmental background, is responsible for generation of information for body repatterning in Xenopus laevis (Kramer et al., 2002). That reconfiguration is also responsible for the morphological changes in the fish dermis during background adaptation (van Eys and Peters, 1981; Tuinhof et al., 1994; Sugimoto et al., 2000; Sugimoto, 2002). This suggests that the reconfiguration of the synaptic morphology of the retino-hypothalamic circuit, generates the information necessary for the background adaptation in Xenopus laevis and this information is generated by processing the visual information in three brain circuits that send inhibitory signals (Tuinhof et al., 1994) and one in the raphe nucleus that sends stimulatory signals (Ubink et al., 1999) to the melanotrope cells of the pituitary.

Adaptation to white background is associated with synaptic retraction (Kramer et al., 2002) and changes in synaptic morphology of the SMIN (suprachiasmatic melanotrope-inhibiting)  neurons. It is observed that the process of exocytosis during secretion of POMC (proopiomelanocortin) and α-MSH by the pituitary is correlated with changes in the level of Ca²⁺ oscillations and so is correlated the expression of the POMC gene with Ca²⁺ oscillations. These facts led to the hypothesis that the patterns of Ca²⁺ oscillations may mediate effects of neurotransmitters in subcellular processes, including gene expression (Kramer et al., 2002; Jenks et al., 2003; Zhang et al., 2006).

Melanotrope cells in Xenopus laevis are found in the pars intermedia of the pituitary and they  synthesize and secrete the α-MSH (α-melanophore-stimulating hormone) as component of its precursor, POMC (proopiomelanocortin). The estimated number of melanotrope cells in Xenopus is 70,000. A number of neurons in the ventrolateral part of the hypothalamic SCN form synaptic connections with the pituitary melanotrope cells and release their neurotransmitters, dopamine, neuropeptide Y and GABA (γ-amino butyric acid). These neurotransmitters have inhibitory effect on the secretion of α-MSH, hence are known as suprachiasmatic melanotrope-inhibiting neurons (SMINs). SMINs respond to reception by the retina of visual input of the light background by releasing GABA, NPY (neuropeptide Y), and dopamine to the synaptic contacts with melanotropes, suppressing α-MSH secretion, leading thus to a lighter  body color in Xenopus. In contrast, when the background is dark, the SCN responds to the retinal visual input by activating a group of interneurons in the dorsal part of the SCN, which send signals that inhibit secretion of α-MSH-inhibiting neurotransmitters by the SMIN. In fact, the neural control of adaptation of color of  Xenopus laevis to the background is more complex: to the pars intermedia of the pituitary also project neurons from locus coeruleus, raphe nucleus and magnocellular nucleus releasing on the melanotrope cells their respective neurohormones/neurotransmitters: noradrenaline, serotonin, and thyrotropin-releasing hormone + corticotrophin-releasing hormone, all with stimulating effect on melanotrope cells (Kramer et al., 2002) (figure 2.23).

From a neuroendocrine viewpoint, the synthesis of α-MSH is determined by the release of stimulatory signals (neurohormones/ neurotransmitters) on the pituitary melanotrope cells but the mediators of these neural functions are two converging and interacting second messengers, Ca²⁺  oscillations and cAMP (cyclic adenosine monophosphate). Ca²⁺ oscillations are involved in the expression of the α-MSH gene but their pattern is determined by the nature of neurotransmitters on the melanotrope cells.

These oscillations are identified in a number of cell types (neurons, astrocytes, endocrine cells, cardiomyocytes and cells of the immune system) in vertebrates, but most of the knowledge on them derives from studies on the nervous system. As pointed out earlier, the background adaptation that is mediated via the action of the pituitary α-MSH in Xenopus laevis is also correlated with specific changes in the patterns of Ca²⁺ oscillations.

The change in color is produced not only via physiological transformations, i.e., via  the distribution of pigments in the pigment cells, but often it results from morphological changes, i.e., changes in the form and number of the chromatophores. The latter case is an unambiguous example of the translation of the visual information on the background into a camouflaging body patterning. The information for that patterning is computationally generated in specific neural circuits by processing the visual information.

Summarizing, it may be said that camouflage in Xenopus implies a complex computational processing of the visual input in several brain centers and culminates with the release from neurons of these centers of their chemical output via projections on the melanotrope cells of the pars nervosa in the pituitary.

If, as demonstrated in the above experiments, changes in the synaptic morphology of the retino-hypothalamic circuit in response to visual input on the background serve as information necessary for adaptive body pigmentation in Xenopus laevis, there is no visible or conceivable reason why analogous changes in patterning of neuronal connections would not code for other adaptive changes in animal morphology.

What Do Neural Circuits Do: Sum up Stimuli or Figure out Adaptive Responses?

Metazoans, like other living forms, are under influence of external and internal stimuli and appropriately respond to them. A stimulus is a challenge to which the organism tends to adaptively respond. After being received by the sensory organ the stimulus in an electrically encoded form is transmitted to a specific neural circuit, which takes it as a problem requiring solution.

The concept of stimulus is a relative one. In the meaning used in this work, a “stimulus” (stimulus - Latin for goad) is a neurally perceived, significant but not necessarily real, change in a variable, signal or condition of the external or internal environments, to which the CNS responds adaptively or otherwise. The mere presence of a condition, a signal, a variable, or a factor in the environment does not necessarily represent a stimulus. It is the perception by the CNS of a change in those entities that makes a stimulus out of them. A species-specific quantitative threshold or set point exists in the nervous system that determines when an agent will be perceived as a stimulus and will be appropriately responded to. Therefore, what may be stimulus for a species may not be for another.

But the fact that even false perceptions may lead to the same output that the real perception does, unambiguously shows that the product of processing depends not on the stimulus, but is a neural product of an intrinsic drive to adapt the organism to the stimulus or to its consequences or presagings.

Stimuli exist only as perceived changes. So, e.g., the mere presence of estrogen in the blood circulation is not a stimulus for the CNS; only a rise of hormone level above the upper limit of an estrogen set point, as determined in the CNS, will be received as a stimulus, processed in a specific neural circuit, and lead to expression of the GnRH (gonadotropin-releasing hormone) gene in hypothalamic neurons.

By determining set points, the CNS, in fact, sets standards for itself on what may and what may not be taken for stimuli. The thresholds are species-specific, but being neurally determined they may exhibit intraspecific variability. Individuals of the same species may exhibit different (within species limits) thresholds or set points and the same individual may change (reset) thresholds and set points within its life time or in response to particular environmental factors.

We take it for granted that a stimulus after being received by the sensory neurons/organs has to be transformed into trains of electric spikes and processed in neural circuits, sometimes very complex, for obtaining a chemical output that starts a signal cascade which leads to a specific phenotypic change. But if the causal law is relevant in biology we have to ask: Why should the stimuli be processed at all? Why should the stimulus and its electrical representation be led through the intricate mazes of brain circuits? What is the evolutionary logic and pressure behind such a highly energy-consuming processing of the external and internal stimuli in brain circuits? Natural selection would not favor evolution of neural processing of environmental stimuli if it would not offer advantages that overweigh the high cost of processing.

An answer to the above questions may be found by looking at the result the processing of the stimuli in neural circuits leads to. The output of the neural circuit leads to expression/suppression of a gene or a group of genes, which may not be predicted from viewpoint of the classical mechanism of gene expression. As argued earlier, there is no physical contact and no natural causal relation between the external/internal stimulus and the gene that will be expressed. Whether the gene will be expressed or which of genes will be expressed depends not on the nature of the stimulus but on the result of the processing of the stimulus in the neural circuit(s). The observed causal chain of events from the stimulus to the expression of the gene is not predetermined, it is determined by the neural processing and it cannot occur independently of the nervous system. Hence, it is the nervous system rather than the particular stimulus that determines which of the thousands of genes in the genome will be induced in response to the stimulus.

The processing of the stimulus in the nervous system bridges the physical gap between the stimulus and the gene, by relating in a contrived causal chain two elements, which otherwise would be causally unrelated.

The function of the processing in the nervous system is to convert the stimulus that per se is inert in relation to gene expression, into a chemical representation - a message for inducing its expression. The nervous system transforms the genetically meaningless stimulus in a way that makes it a genetically intelligible message. This is the most profound significance of the neural processing.

Each neuron in the circuit receives and integrates input from many other neurons in order to produce its own output. However, the integration of the input from many different neurons and generation of the output by the neuron is not a summation of the input for the input is modified when is received in dendrites. What is the processing of the stimulus essentially intended to do?

Let’s consider once again the case of the shortening of photoperiod. As a stimulus, length of the day is perceived in the CNS by special hypothalamic circuits. As pointed out earlier, the light or darkness per se do not affect expression of any gene in cells that are directly affected (dermis cells, retinal cells, e.g.), but it affects expression of the GnRH (gonadotropin-releasing hormone) in the hypothalamus in opposite ways in cows and ewes, which respond to the same stimulus, i.e. the shortening of the day length, in opposing ways by inhibiting and inducing expression of the GnRH gene in hypothalamic GnRH neurons respectively. Two different results from the same stimulus! This would be paradoxical if the stimulus per se would be the cause of expression/suppression of the GnRH gene. As argued in length earlier, the stimulus is not the cause, for were it a cause it would lead to the same result in both the above species. Far from a paradox, both opposite results are very adaptive as far as the reproduction success of both species is concerned: they are computed to produce offspring at the optimal time of the year for rearing them. These results show:

Firstly, that the external stimulus is not the cause of opposite results observed in the two species and contains no information or instruction for these organisms,

Secondly, that the information for timing sexual activity, mating and parturition is generated in the brain (hypothalamus), and

Thirdly, that the relevant computation in hypothalamic circuits does not consist simply in summation of external stimuli (otherwise the same result would be observed in both species).

Such diverse responses to the same stimulus show that what induces or inhibits expression of the GnRH gene is not the photoperiod per se (day light has no access to those neurons and genes, which are permanently in darkness), but the chemical output resulting from the manipulative, adaptation-oriented processing of the stimulus in specific neural circuits. Differences in the output result from differences in the processing of the same stimulus in both species, which, ultimately, are determined by differences in computational properties of their neural circuits.

Now let’s put the question more explicitly: is the computational processing in the nervous system a summation and integration of the input of stimuli or is it a solution-oriented complex calculation, estimation, or reckoning? If the first were the case, then it would be expected that the result of computation, i.e. the chemical signal released by neural circuits in different species, in response to the same stimulus, would be similar. But, as we all know, the chemical signal(s) released and the ensuing morphological, physiological, and behavioral characters induced in different metazoan species in response to the same stimulus often are different or even, in our case, opposite.

What computation of the stimulus in the neural circuit is intended to do is to start an adaptive response to the stimulus by establishing, via a signal cascade, a new naturally not existing relationship between the stimulus and a specific gene or group of genes. The function of the computation in neural circuits is to foreordain a morphological, physiological, behavioral, or life history modification that would adapt the organism to the new or changing element (stimulus) of the external or internal environment.

The end result of the complex processing of the stimulus in the neural circuit is secretion of a chemical that is necessary for starting a signal cascade for expression of a specific gene or group of genes. The processing, and the manipulative expression of genes it induces, represents a solution to the problem of adaptation of the organism to the stimulus.

Manipulative responses to external/internal stimuli evolved from a strong evolutionary pressure for both maintaining the continuously eroding metazoan structure as well as for rapid adaptation to the adversely changing environment. These responses became the predominant mode of gene expression in the CNS and a crucial element of developmental and adaptational processes in metazoans.

The ability of the CNS to manipulatively respond to various external and internal stimuli is a qualitatively new property of metazoans. By enabling the organism to produce different responses (signal outputs) to the same stimulus, and the reverse, to produce the same response to different stimuli, the CNS created a rich repertory of adaptive capabilities.

The computationally determined manipulative response of metazoans in response to external/internal stimuli is in sharp contrast with the genetic determinism of responses observed in unicellulars, where the relationship between internal/external stimuli is totally determined by thermodynamical-stereochemical properties of the stimuli (in some eukaryotic unicellulars also signal transduction pathways) and genes.

Characterizing the new intellectual situation von der Malsburg writes:

 

According to an old mode of thinking a Creator, or Mother Nature or a genetic program, are in control of every molecular reaction with exquisite algorithmic foresight. Although this thought pattern of “hetero-organization” under control of a preexisting plan is sub-consciously still governing many a thought, it now is quickly losing its dominance, giving way to models of self-organization on all levels – evolution, ontogenesis, learning and functional organization. (von der Malsburg, 2002a)

 

In chapter 6, adequate evidence is presented to show that by processing internal stimuli from the developing embryo, the embryonic CNS generates the epigenetic information necessary for the sequential stages of individual development.

 

 

 

 

<<PREVIOUS<<                                                                                                                                                      >>NEXT>>