The findings of Tkachev, et al

Oligodendrocyte Dysfunction in Schizophrenia and Bipolar Disorder

This is my response to "Oligodendrocyte Dysfunction in Schizophrenia and Bipolar Disorder, The Lancet 2003; 362: 798-805. It was sent to The Lancet and to the corresponding author. The "summary" and the "discussion" are reproduced at the end ot my response.

The findings of Tkachev, et al., may be explainable as results of low availability of dehydroepiandrosterone (DHEA). Tkachev reported "downregulation" of genes controlling myelin production in oligodendrocytes in schizophrenia. DHEA positively affects myelin in damaged nerves (Microsurgery 2003; 23(1): 49-55 and Microsurgery 2002; 22(6): 234-41). Oligodendrocytes are known to produce DHEA (Endocrinology 1999; 140(8): 3843-52 and Journal of Neurochemistry 2000; 74(2): 847-59). Schizophrenics exhibit significantly low DHEA (Biological Psychiatry 1973, 6: 23).

In 1985, I first suggested schizophrenia and Alzheimer's disease result from low DHEA (A Theory of the Control of the Ontogeny and Phylogeny of Homo sapiens by the Interaction of Dehydroepiandrosterone and the Amygdala). (At the time, I was not aware of Biological Psychiatry 6: 23 and the first reports (three in the same year) of low DHEA in Alzheimer's disease did not appear until 1989 in Lancet.) Tkachev, et al., also report that other research has "not shown myelin-related gene expression changes" in Alzheimer's disease, among other neuropathies. My prediction of the effects of low DHEA in schizophrenia and Alzheimer's disease are derived from my principal hypothesis that evolution selected DHEA because DHEA optimizes replication and transcription of DNA. (In fact, it is my hypothesis that mammals evolved because of DHEA, "Hormones in Mammalian Evolution," Rivista di Biologia / Biology Forum 2001; 94: 177-184 and is very important to human evolution, "Androgens in Human Evolution. A New Explanation of Human Evolution," Rivista di Biologia / Biology Forum 2001; 94: 345-362.) It followed that all tissues, especially those of the brain, are directly affected by levels of DHEA. Among some other diseases of the nervous system, I attempted to demonstrate a connection of low DHEA with schizophrenia, Alzheimer's disease and depression. (A connection of depression with low DHEA has since been supported.) I was suggesting that these neuropathies represent reduced transcription, or "downregulation," of genes as causes of these diseases. I was pleased to read Tkachev, et al. Now, according to my explanation, different genes that are vulnerable to the effects of low DHEA will fail to function properly according to their vulnerabilities. Therefore, as Tkachev, et al., report, the effect is not the same in all tissues and all diseases. However, DHEA naturally begins to decline around twenty to twenty-five, reaching very low levels in old age. I suggest this loss of DHEA begins to expose genes which do not function well during times of low DHEA and this combination may account for many diseases which appear near or subsequent to age twenty. This may also explain the findings of Tkachev. (Please read my explanation of schizophrenia from 1996.)

Lancet 2003; 362: 798-805

Summary

Background Results of array studies have suggested abnormalities in expression of lipid and myelin-related genes in schizophrenia. Here, we investigated oligodendrocyte-specific and myelination-associated gene expression in schizophrenia and bipolar affective disorder.

Methods We used samples from the Stanley brain collection, consisting of 15 schizophrenia, 15 bipolar affective disorder, and 15 control brains. Indexing-based differential display PCR was done to screen for differences in gene expression in schizophrenia patients versus controls. Results were cross-validated with quantitative PCR, which was also used to investigate expression profiles of 16 other oligodendrocyte and myelin genes in schizophrenia and bipolar disorder. These genes were further investigated with an ongoing microarray analysis.

Findings Results of differential display and quantitative PCR analysis showed a reduction of key oligodendrocyte-related and myelin-related genes in schizophrenia and bipolar patients; expression changes for both disorders showed a high degree of overlap. Microarray results of the same genes investigated by quantitative PCR correlated well overall.

Interpretation Schizophrenia and bipolar brains showed downregulation of key oligodendrocyte and myelination genes, including transcription factors that regulate these genes, compared with control brains. These results lend support to and extend observations from other microarray investigations. Our study also showed similar expression changes to the schizophrenia group in bipolar brains, which thus lends support to the notion that the disorders share common causative and pathophysiological pathways.

Discussion

We believe that our results provide strong evidence for oligodendrocyte and myelin dysfunction in schizophrenia and bipolar affective disorder. Expression profiles of most known oligodendrocyte-related and myelin-related genes were greatly reduced, and several transcription factors known to coordinate myelin gene expression showed corresponding alterations. The high degree of correlation between the expression changes in schizophrenia and bipolar disorder provide compelling evidence for common pathophysiological pathways that may govern the disease phenotypes of schizophrenia and bipolar affective disorder.

Some genes--particularly from the bipolar group--that were significantly changed with quantitative PCR were unchanged by microarray analysis. This difference could be attributable to the different sample set used (only 11 of 15 schizophrenia, 14 of 15 bipolar, and 11 of 15 control brains passed our stringent quality control measures for microarray analysis inclusion), or more importantly, it could be because of differences in the dynamic range between the two techniques.^21-23 The bipolar group microarray results could be affected more than the quantitative PCR data, because this dataset generally showed a wider dispersion as measured by the index coefficient of variation (data not shown). Furthermore, some of the Affymetrix probe sets might not correspond exactly with the transcripts investigated by quantitative PCR.

Genes implicated in pathological processes of complex diseases are difficult to identify. Complexity arises from involvement of multiple genes and environmental factors. These are difficult to disentangle and, together, represent what Gottesman termed the epigenetic puzzle.²⁴ The main advantage of global profiling techniques is that many pieces of the puzzle can be examined within a short period, leading to identification of abnormal pathways of gene expression, signal transduction, or both.

Our data suggest oligodendrocyte dysfunction in schizophrenia and bipolar disorder. Myelin synthesis by oligodendrocytes provides the basis for rapid impulse conduction in the CNS, and there is increasing evidence that oligodendrocytes also play an important part in axonal survival. Myelin production requires coordinated synthesis of myelin-specific lipids and membrane-associated proteins that facilitate complex spiral wrapping, membrane compaction, and fine architecture of multilayered myelin sheaths. Temporal onset of psychotic disorders in late adolescence or early adulthood coincides with the concluding myelination of the prefrontal cortex.²⁵

An important finding in our study was downregulation of most genes in the group of myelin-associated proteins (mature oligodendrocyte marker genes), and it might be of relevance that almost all the downregulated myelin proteins are located in the major dense line of myelin and represent transmembrane-spanning proteins.¹⁵ Downregulation of the transmembrane myelin proteins might lead to a defect in myelin compaction. Transgenic mice without PLP1 or MAG expression show ultrastructural changes in myelin structure and deficits in myelin compaction and impairment in organisation of the periaxonal region.^26,27 These abnormalities might correlate with ultrastructural changes of oligodendroglial cells and myelin sheaths noted in schizophrenia and bipolar disorder. Uranova and colleagues²⁸ reported reduced compaction of myelin sheaths, and abnormal inclusions between myelin lamellae and diminution of oligodendrocyte cell bodies by quantitative electron microscopy in the prefrontal cortex of schizophrenia and bipolar disorder tissue from the Stanley brain collection. Hof and colleagues²⁹ reviewed molecular and cellular evidence for oligodendrocyte dysfunction in schizophrenia, and reported a substantial reduction in oligodendrocytes amounting to more than 20% in layer III of area 9 and white matter of the superior frontal gyrus. Further evidence comes from naturally arising human null mutations of the PLP1 gene. Patients without PLP1 develop axonal degenerations in the absence of demyelination and inflammation.³⁰ Clinical presentation and severity of both human and mouse PLP1 null mutations are variable and often entail motor impairment, cognitive impairment, and ataxia. PLP1 is thought to mediate axonal-glial interactions through an, as yet, unknown mechanism. PLP1-deficient patients show reduced N-acetyl aspartate (NAA) concentrations by proton magnetic resonance spectroscopy analysis. Low NAA amounts in the prefrontal cortex have been noted in many independent studies of patients with schizophrenia and bipolar disorder, including studies of drug-free patients,³¹ indicating that these findings are not secondary to antipsychotic drug treatment. On the contrary, NAA measures in the prefrontal cortex are selectively amplified by treatment with antipsychotic medication.³²

We considered whether the noted abnormalities in oligodendrocyte and myelin-related genes could be secondary to antipsychotic drug treatment. This occurrence was unlikely because most patients with bipolar disorder did not receive high doses or indeed any antipsychotic drugs; nevertheless these patients showed, for most transcripts, great downregulation. Three individuals in the schizophrenia and in the bipolar disorder group had not been treated with any drug before death or indeed ever; however, several of these individuals still had substantial reduction of most of the genes investigated. Additionally, Pongrac and colleagues² investigated PLP1 gene expression in chronically haloperidol-treated monkeys by microarray, but reported no effect on PLP1 transcripts. Analysis of covariance showed no significant relation between gene expression and several measures of drug exposure, namely lifetime drug exposure, number of years of antipsychotic and type of antipsychotic (typical vs atypical antipsychotics) at time of death for either the schizophrenia or the bipolar disorder group. Similarly, Hakak and colleagues also considered drug effects for the downregulation of myelin-related genes.¹ They separately analysed patients who were drug-free before death to test this possibility and reported no evidence for a confounding effect of antipsychotic drugs. There was also no statistical evidence that the inclusion of other potential confounding variables influenced the main factor effect, which is clinical diagnosis.

However, a more relevant concern is whether the noted changes in myelin-related genes are directly related to the disease process in schizophrenia and bipolar disorder. A cDNA array study on the frontal cortex of alcohol abuse patients noted changes in various myelin-related genes in some groups.³³ On the other hand, results of array studies on post-mortem brains of patients with Alzheimer's disease, Rett's syndrome, and multiple sclerosis have not shown myelin-related gene expression changes.^34-37 We examined the effect of heavy alcohol use in our brain collection and noted no consistent or significant correlation between heavy alcohol use and downregulation of myelin-related genes. However, the alcohol misuse patients used in the study by Mayfield and colleagues³³ will have been more severely and chronically alcohol dependent.

Another important question is whether a fall in oligodendrocyte numbers causes the gene expression changes. Several studies have noted a decrease in oligodendrocyte numbers; however, other studies could not confirm this observation.²⁹ In our study, we noted a selective, rather than a global, reduction in oligodendrocyte-related gene expression, which may point more towards a cellular dysfunction. To address the question of death or dysfunction, more refined expression studies are needed, ideally with laser-captured microdissection of oligodendrocytes.

The schizophrenia research field is plagued by non-reproducible results, most probably attributable to several factors, particularly small sample size and inconsistency in study design and result assessment bias with semiquantitative techniques. The great advantage of array studies, especially with commercially available and standardised profiling platforms like the Affymetrix system, is that these technologies now allow direct comparisons of expression profiles of many genes in one experiment, and that findings from different brain collections can easily be correlated to obtain the necessary statistical power to attempt to unravel complex disorders such as schizophrenia and bipolar disorder.

Once we have identified reproducible disease changes and have established whether oligodendrocytes are sick or dead, more questions arise as to whether these findings are disease-specific or a so-called sick-brain event, whether changes are secondary to another disease process, or both of these. Complex neuropsychiatric disorders represent arguably the ultimate challenge for global profiling technologies. At present, no other approach holds as much promise to, eventually, move psychotic illnesses into the realm of biologically understandable disorders.