Schizophrenia as variation in the sapiens-specific epigenetic instruction to the embryo


Timothy John Crow, Department of Psychiatry, University of Oxford, SANE-POWIC, Warneford Hospital, Oxford OX3 7JX, UK.
Tel.: +441865455918;
fax: +441865455922;


Crow TJ. Schizophrenia as variation in the sapiens-specific epigenetic instruction to the embryo.

The psychoses (schizophrenia and bipolar disorder) occur in all populations with approximately uniform incidence and sex-dependent age of onset. Core symptoms involve aspects of language; brain structural deviations are sex and hemisphere-related. Genetic predisposition is unaccounted for by linkage or association. The hypothesis is proposed that the ‘missing heritability’ is epigenetic in form and generated in meiosis on a species-specific XY chromosomal template. A duplication from Xq21.3 to Yp11.2 that occurred 6 million years ago is proposed as critical to hominin evolution. Within this block of homology the Protocadherin11XY gene pair is expressed as a cell surface adhesion factor in both X and Y forms; it has undergone a series of coding changes (16 in the Y sequence and 5 in the X including two to cysteines) in the hominin lineage. According to the hypothesis these sequence changes, together with one or more deletions and a paracentric inversion in the Y block, were successively selected; late events in this series established cerebral asymmetry (the ‘torque’) as the defining characteristic of the human brain. Built around this reference frame, an epigenetic message channels early development of the embryo in a sapiens-specific format. Diversity in meiotic pairing is postulated as the basis for species-specific deviations in development associated with psychosis.

Conflicts of interest

The author is aware of no conflicts of interest.

In the face of a fecundity disadvantage schizophrenic (1) and affective (2) psychoses continue to arise in all human populations with approximately uniform incidence. A genetic contribution is established by twin and adoption studies (3). Yet the hypothetical gene variants are not selected out. Resistance to wound shock and stress was suggested (4) as a balancing advantage, but no such association has been shown. An alternative explanation is that these conditions are in some way related to the human capacity for language, and may be associated with other variations that cross populations – for example, the incidence of right handedness, sex differences in verbal and spatial ability, and in age at procreation (5). Each characteristic distinguishes this from other mammalian species.

When the first genetic variations relating to neuropsychiatric disease were established in the late 1980s (6), there was widespread optimism that the same techniques could be applied to the problems of psychosis. All that was necessary was to ‘drain the pond dry’ to reveal the relevant genes. Claims for linkage were apparently replicated and genes for psychosis that were numerous and of small effect were widely discussed (7). But with a steady increase in sample size, particularly with the uncomplicated sibling pair strategy, none has proven robust (8). With more effective techniques and a substantial increase in the number of markers Genome Wide Association studies (GWAS) gained strength. However by 2007 at the World Congress of Psychiatric Genetics, it was apparent that no more replicable findings were emerging from GWAS than from linkage. No genetic association can yet be regarded as unequivocally established either by linkage or association.

That the symptoms are centred on the ability to communicate with conspecifics, i.e. to language (5) illustrates how the disorders discriminate certain brain areas and functions (9). Onsets occur throughout the reproductive period with a well-established sex difference in age of onset: earlier, with affective loss and worse outcome in males and, later, with affective exaggeration in females (10). Thus psychosis relates to the sequence of human brain development; arguably to precisely those variables that have changed in hominin evolution, and that determine the life-span of the species (11).

Deviations in brain structure include a degree of ventricular enlargement that is lateralized (12) as well as sex dependent (13). Recent meta-analyses suggest the primary target of the disease process is the insula, and following this the cingulate, and para-hippocampal gyri (14), structures that together make up ‘le grand lobe limbique’ of Paul Broca.

What is a species?

According to Darwin (15), the transition between species is gradual. The distinction between varieties and species is not well marked. On publication of the Origin of Species in 1859, TH Huxley wrote that he hoped Darwin had not loaded himself with ‘an unnecessary difficulty in adopting Natura non facit saltum so unreservedly’.

Thus, a debate was initiated to which hominin evolution is particularly relevant. It is widely assumed that Homo sapiens is a species distinct from Homo neanderthalensis and from Homo erectus. What is the essence of the differences?

One attempt at a definition is the Biological or Isolation Species Concept: a species is a ‘group of actually or potentially interbreeding natural populations that are reproductively isolated from other such groups' (16). A more specific, and functional, definition is ‘the most inclusive population of individual biparental organisms which share a common fertilization system’(17). Thus, according to Paterson, a species is defined by a ‘specific mate recognition system’ (note: such a recognition system is broadly defined to include pherormonal and behavioral mechanisms, as well as skin coloration and mating calls). Because the ‘mate recognition system’ distinguishes that species from all others Paterson's concept requires more than graduated change.

Natural selection and sexual selection

Few proposals have specified the nature of saltation. Any such attempt encounters the difficulty that the greater the magnitude of the saltation the less survival value it is likely to have, and the greater the difficulty the mutant will have in finding a mate. Is a mate with the same mutation required? – if so there is an obvious problem, such mutations by definition are rare; no such mate is available. Thus the challenge to Darwinian gradualism has been widely disregarded.

But Darwin's (18) juxtaposition of The Descent of Man and the theory of sexual selection introduces a new possibility. A role for sexual selection in modifying a primary change in a sexually dimorphic feature to establish a new species has been argued in relation to Hawaiian Drosophilid species (19, 20). Similar arguments have been applied to the case of the prolific speciation of cichlid fishes in the lakes of East Africa (21) and in birds (22).

Some authors favoring discontinuities have advocated chromosomal change as the mechanism (23, 24). Here, it is argued that it is not chromosomal change per se, but change specifically on the sex chromosomes that plays a role in speciation. Such changes include species-specific sexual dimorphisms necessary to the construction of a mate recognition system. Furthermore, non-recombining regions of X-Y homology can account (as in the case of lateralization in humans, see below) for quantitative differences in a characteristic between males and females. Such dimensions are plausible substrates for sexual selection.

The Y chromosome in mammals (like the W in birds) has a unique role, because it is not necessary for survival. While the X is the most stable chromosome across species [Ohno's law (25)] the Y is by far the most variable; it is a test-bed for evolutionary change. The hypothesis proposed here is that the primary change in speciation takes place on the Y, located in a region of homology with the X, thus allowing correlated but independent change in the two sexes. Differing ranges of variation in a single variable have the potential to explain the type of runaway sexual selection envisaged by Fisher (26). Thus a primary and saltational change on the heterogametic chromosome creates a target for selection by the other sex to define a new mate recognition system. According to this two-step theory each species is defined by a novel sexual dimorphism associated with a specific X-Y homologous gene pair.

Thus the theory accounts for radical differences between species in mechanisms of communication/ signaling between the sexes, by a qualitative saltational change (sometimes a rearrangement) on the heterogametic Y chromosome, associated with quantitative change mediated through related sequences on the homogametic (X) chromosome.

A role for the gene Prdm9 [that is involved in speciation (27) and species differences in recombination (28, 29)] in stabilizing the relationship between X and Y chromosomes in the XY body has been suggested (30). Through a zinc finger conformation that is highly variable between species the Prdm9 protein has DNA-binding capacity coupled with histone trimethylating potential. If these activities somehow combine to generate a lock and key compatibility between X and Y chromosomal structures in male meiosis this could constitute the template for the epigenetic message passed to the zygote. [‘Epigenetic’ here specifies modifications of gene expression without alteration of the DNA sequence itself; mechanisms include methylation of the DNA sequence, and acetylation, phosphorylation and methylation of the histone proteins with which the DNA sequence is associated in the chromosome]. If the conformation also generates a recombination motif, differences in C+G% that have been described between species (31) might be accounted for.

Broca (32) wrote ‘Man is of all the animals the one whose brain is most asymmetrical. He also possesses the most acquired faculties. The faculty of language distinguishes us most clearly from the animals'. The concept that asymmetry of the hemispheres is the defining feature of the human brain and the cerebral correlate of language (the mate recognition system) – referred to as the Broca-Annett axiom – implies that the genetic basis of this evolutionary innovation was the speciation event for modern Homo sapiens.

The genetics of cerebral dominance

A clue to the location of the cerebral dominance gene or ‘right shift factor’ comes from sex chromosome aneuploidies. Individuals (with a frequency of approximately 1 in 1000 in the population) who lack an X chromosome (XO, Turner's syndrome) have non-dominant hemisphere (spatial) deficits on cognitive testing. Individuals with an extra X (XXY, Klinefelter's, and XXX syndromes) have verbal or dominant hemisphere deficits (Table 1). A logical explanation is that an asymmetry determinant optimally expressed from both X chromosomes in normal females is associated with right hemisphere dysfunction when expressed from only one X, and left hemispheric dysfunction when expressed in three doses in XXX females. But then the question arises of why males, who only have one X chromosome do not have spatial deficits such as are seen in Turner's syndrome, and why Klinefelter's individuals who have two X chromosomes like normal females have left hemispheric deficits comparable to XXX syndrome individuals. The answer must be that the copy of the gene on the X chromosome is complemented by a copy on the Y, i.e. that the gene is in the X/Y homologous class (10). A hormonal explanation will not account for the similarity of the changes in XXY individuals, who are male, and XXX individuals, who are female. The case for location on the Y chromosome is substantially reinforced by the verbal deficits/delays that are observed in XYY individuals (33).

Table 1.  Neuropsychological impairments associated with sex chromosome aneuploidies
Normal femaleNormal maleTurner's syndromeKlinefelter's syndrome
Number of sex chromosomes221333
Verbal abilityNormalNormalNormalDelayedDelayedDelayed
Spatial abilityNormalNormalDecreasedNormalNormalNormal

The hypothesis is further strengthened by evidence that Turner's and Klinefelter's syndrome individuals have corresponding deviations in anatomical asymmetry (34) and by the demonstration in a family study of a same sex concordance effect – the tendency for handedness and sex to be associated above chance expectation – the hallmark of X-Y linkage (35). A role for an X-Y homologous gene is consistent with the presence of a sex difference – brain growth is faster (36) and lateralization to the right is stronger (37) in females. Females have greater mean verbal fluency and acquire words earlier (38, 39) than males. These facts are related, and they inform us of the nature of the genetic mechanism: the gene is present, but in a modified form, on both X and Y chromosomes. If the brain changes in psychosis are lateralized as earlier suggested this may also reflect variation in this X and Y linked gene.

The Xq21.3/Yp duplication

There is a lead to its identity. A major chromosomal rearrangement now dated at 6 million years (40) took place in hominin evolution. A 3.5 Mb contiguous block of sequences from the X chromosome long arm was duplicated onto the Y chromosome short arm. That event is therefore a candidate for the transition from a great ape/hominin precursor to Australopithecus. The homologous block thus created was subject to three subsequent deletions (small segments of the chromosome, that might influence the function of neighboring genes, were lost), and was split by a paracentric inversion [by a recombination, presently undated, of LINE-1 elements (41, 42)], a further rearrangement of the structure of the chromosome, to give two blocks of homology in Yp (Fig. 1). Two regions on the human Y chromosome short arm thus share homology with a single region on the human X chromosome long arm (Xq21.3) (43, 44). Genes within this region are therefore present on both the X and Y chromosomes in Homo sapiens but on the X alone in other great apes and primates (45).

Figure 1.

The Xq21.3/Yp duplication of 3.5 megabases from the X chromosome long arm to the Y chromosome short arm now dated at 6 million years, i.e. coincident with the separation of the chimpanzee, and hominin lineages. A second event (a ‘paracentric inversion’) reversed the direction of most of the duplication, and some of the pre-existing Y short arm, but has not been dated.

Three genes are located within this block; PABPC5, a poly (A)-binding protein; TGIF2LX and Y, (homeobox-containing genes with testis-specific expression) and the ProtocadherinX (PCDH11X) and ProtocadherinY (PCDH11Y) gene pair (all are located in the larger distal segment –Fig. 2). The Y gametologue of PABPC5 has been lost as a consequence of one of the deletions (c in Fig. 2) during hominin evolution, and TGIF2LY has been inactivated by a frameshift mutation. This leaves the PCDH11XY gene pair that codes for cell adhesion molecules of the cadherin superfamily as salient because both forms of the gene have been retained, and are highly expressed both in fetal and adult brain (46, 47) including the germinal layer of the cortex (Priddle, personal communication). The protein products of this gene pair are thus expected to play a role in intercellular communication (Figure 3). These proteins have been subject to change throughout hominin evolution (40) with 16 coding changes in the Y sequence and 5 coding changes including two to sulfur-containing cysteines, in the X sequence, all of which appear fixed across populations.

Figure 2.

Alignment of the homologous regions on the X long arm and Y chromosome short arm to show four deletions (a to d) on the Y and the content of genes.

Figure 3.

The Protocadherin11XY gene pair, each gene comprising seven extracellular cadherin motifs, a short transmembrane region and an intracellular cytoplasmic tail. The proteins encoded by this gene pair are the only Protocadherins that include both β-catenin and protein phosphatase 1α-binding sites. PCDH11Y has been described as the only ‘gain of function’ gene in Homo sapiens. It is plausible that the gene pair has a role in synaptogenesis, and as a sexually dimorphic axonal guidance factor. A 4-year-old male referred for severe language delay was found to have independent deletions in PCDH11X and PCDH11Y (54).

It is suggested that the presence of the homologous block on the Y chromosome created a field for genetic innovation throughout hominin evolution. Thus, while the duplication at 6 million years ago is a candidate for the Australopithecus speciation event, each subsequent deletion, or inversion influencing PCDH11Y expression is a potential initiator, and subsequent sequence change in PCDH11X a potential terminator, of a speciation event.

Species-specific variation is epigenetic

The central issue is the nature of the transition to modern Homo sapiens, now considered to have taken place 160 thousand years ago. What was the critical change? Although not dated the paracentric inversion may have been associated with an epigenetic revolution in this corner of the genome, and for this reason is the leading candidate.

In mammals, genes on one X chromosome are subject to the process of X inactivation, but gene sequences that are also represented on the Y chromosome are protected from this influence. Such genes are expressed from both X and Y in males and from both Xs in females, a similar dosage thus being maintained in each sex. The likely mechanism is by epigenetic suppression of unpaired chromosomes (48) in male meiosis. Gene sequences that have been transferred from the X to the Y are in a novel situation, and a phase of epigenetic equilibration must be assumed. If X-Y pairing plays a role, the orientation of the sequence on the Y is the key determinant of pairing, and the paracentric inversion initiated critical changes in quantitative expression of the PCDH11XY gene pair.

An epigenetic influence on cerebral asymmetry was reported in a magnetic resonance imaging study in which differences in handedness between monozygotic twins were correlated with differences in anatomical asymmetry of the planum temporale (49). Such variation can account for the stochastic element incorporated in genetic theories (50, 51). An epigenetic influence on transmission of psychosis is substantiated by an increase in risk with increasing age of the father, and more recently by an effect of equal size of the age of the maternal grandfather. These effects have been interpreted as mediated by the X chromosome. An approach to such epigenetic variation on the sex chromosomes is feasible through twin studies (52).

Epigenetic control of X and Y encoded sequences may have a role more fundamental than regulation of gene dosage in embryogenesis. According to the hypothesis developed here the X and the Y chromosomes encode a pattern of genetic activity that encompasses the most recently acquired sexual dimorphism in a particular species. Perhaps the primary function of ‘meiotic suppression of unpaired chromosomes' (MSUC) is to pass to the embryo a genetic message that defines the species. Although most apparent in late ontogeny such sexually dimorphic features may need to be imposed soon after fertilization at a time when gene expression in general is suppressed.

The scenario has a further implication. If the mechanism of imposition of the message is by MSUC there is scope for significant variation in pairing and thus for variation in the epigenetic message (53). Such variation will be species specific, in the case of Homo sapiens related to asymmetry and therefore to language. According to this theory, species-specific variation including that relating to pathologies of uniform incidence across populations is distinct from Mendelian variation. It has its origin in meiosis, and the transitions between species.