See comment by Mary Reid

Schizophrenia Risk and the Paternal Germ Line
By Dolores Malaspina

Dolores Malaspina

Paternal age at conception is a robust risk factor for schizophrenia. Possible mechanisms include de novo point mutations or defective epigenetic regulation of paternal genes. The predisposing genetic events appear to occur probabilistically (stochastically) in proportion to advancing paternal age, but might also be induced by toxic exposures, nutritional deficiencies, suboptimal DNA repair enzymes, or other factors that influence the

fidelity of genetic information in the constantly replicating male germ line. We propose that de novo genetic alterations in the paternal germ line cause an independent and common variant of schizophrenia.

Seminal findings
We initially examined the relationship between paternal age and the risk for schizophrenia because it is well established that paternal age is the major source of de novo mutations in the human population, and most schizophrenia cases have no family history of psychosis. In 2001, we demonstrated a monotonic increase in the risk of schizophrenia as paternal age advanced in the rich database of the Jerusalem Perinatal Cohort. Compared with the offspring of fathers aged 20-24 years, in well-controlled analyses, each decade of paternal age multiplied the risk for schizophrenia by 1.4 (95 percent confidence interval: 1.2-1.7), so that the relative risk (RR) for offspring of fathers aged 45+ was 3.0 (1.6-5.5), with 1/46 of these offspring developing schizophrenia. There were no comparable maternal age effects (Malaspina et al., 2001).

Epidemiological evidence
This finding has now been replicated in numerous cohorts from diverse populations (Sipos et al., 2004; El-Saadi et al., 2004; Zammit et al., 2003; Byrne et al., 2003; Dalman and Allenbeck, 2002; Brown et al., 2002; Tsuchiya et al., 2005). By and large, each study shows a tripling of the risk for schizophrenia for the offspring of the oldest group of fathers, in comparison to the risk in a reference group of younger fathers. There is also a "dosage effect" of increasing paternal age; risk is roughly doubled for the offspring of men in their forties and is tripled for paternal age >50 years. These studies are methodologically sound, and most of them have employed prospective exposure data and validated psychiatric diagnoses. Together they demonstrate that the paternal age effect is not explained by other factors, including family history, maternal age, parental education and social ability, family social integration, social class, birth order, birth weight, and birth complications. Furthermore, the paternal age effect is specific for schizophrenia versus other adult onset psychiatric disorders. This is not the case for any other known schizophrenia risk factor, including many of the putative susceptibility genes (Craddock et al., 2006).

There have been no failures to replicate the paternal age effect, nor its approximate magnitude, in any adequately powered study. The data support the hypothesis that paternal age increases schizophrenia risk through a de novo genetic mechanism. The remarkable uniformity of the results across different cultures lends further coherence to the conclusion that this robust relationship is likely to reflect an innate human biological phenomenon that progresses over aging in the male germ line, which is independent of regional environmental, infectious, or other routes.

Indeed, the consistency of these data is unparalleled in schizophrenia research, with the exception of the increase in risk to the relatives of schizophrenia probands (i.e., 10 percent for a sibling). Yet, while having an affected first-degree relative confers a relatively higher risk for illness than having a father >50 years (~10 percent versus ~2 percent), paternal age explains a far greater portion of the population attributable risk for schizophrenia. This is because a family history is infrequent among schizophrenia cases, whereas paternal age explained 26.6 percent of the schizophrenia cases in our Jerusalem cohort. If we had only considered the risk in the cases with paternal age >30 years, our risk would be equivalent to that reported by Sipos et al. (2004) in the Swedish study (15.5 percent). When paternal ages >25 years are considered, the calculated risk is much higher. Although the increment in risk for fathers age 26 through 30 years is small (~14 percent), this group is very large, which accounts for the magnitude of their contribution to the overall risk. The actual percentage of cases with paternal germ line-derived schizophrenia in a given population will depend on the demographics of paternal childbearing age, among other factors. With an upswing in paternal age, these cases would be expected to become more prevalent.

Biological plausibility
We used several approaches to examine the biological plausibility of paternal age as a risk factor for schizophrenia. First, we established a translational animal model using inbred mice. Previously it had been reported that the offspring of aged male rodents had less spontaneous activity and worse learning capacity than those of mature rodents, despite having no noticeable physical anomalies (Auroux et al., 1983). Our model carefully compared behavioral performance between the progeny of 18-24-month-old sires with that of 4-month-old sires. We replicated Auroux's findings, demonstrating significantly decreased learning in an active avoidance test, less exploration in the open field, and a number of other behavioral decrements in the offspring of older sires (Bradley-Moore et al., 2002).

Next, we examined if parental age was related to intelligence in healthy adolescents. We reasoned that if de novo genetic changes can cause schizophrenia, there might be effects of later paternal age on cognitive function, since cognitive problems are intertwined with core aspects of schizophrenia. For this study, we cross-linked data from the Jerusalem birth cohort with the neuropsychological data from the Israeli draft board (Malaspina et al., 2005a). We found that maternal and paternal age had independent effects on IQ scores, each accounting for ~2 percent of the total variance. Older paternal age was exclusively associated with a decrement in nonverbal (performance) intelligence IQ, without effects on verbal ability, suggestive of a specific effect on cognitive processing. In controlled analyses, maternal age showed an inverted U-shaped association with both verbal and performance IQ, suggestive of a generalized effect.

Finally, we examined if paternal age was related to the risk for autism in our cohort. We found very strong effects of advancing paternal age on the risk for autism and related pervasive developmental disorders (Reichenberg et al., in press). Compared to the offspring of fathers aged 30 years or younger, the risk was tripled for offspring of fathers in their forties and was increased fivefold when paternal age was >50 years. Together, these studies provide strong and convergent support for the hypothesis that later paternal age can influence neural functioning. The translational animal model offers the opportunity to identify candidate genes and epigenetic mechanisms that may explain the association of cognitive functioning with advancing paternal age.

A variant of schizophrenia
A persistent question is whether the association of paternal age and schizophrenia could be explained by psychiatric problems in the parents that could both hinder their childbearing and be inherited by their offspring. If this were so, then cases with affected parents would have older paternal ages. This has not been demonstrated. To the contrary, we found that paternal age was 4.7 years older for sporadic than familial cases from our research unit at New York State Psychiatric Institute (Malaspina et al., 2002). In addition, epidemiological studies show that advancing paternal age is unrelated to the risk for familial schizophrenia (Byrne et al., 2003; Sipos et al., 2004). For example, Sipos found that each subsequent decade of paternal age increased the RR for sporadic schizophrenia by 1.60 (1.32 to 1.92), with no significant effect for familial cases (RR = 0.91, 0.44 to 1.89). The effect of late paternal age in sporadic cases was impressive. The offspring of the oldest fathers had a 5.85-fold risk for sporadic schizophrenia (Sipos et al., 2004); relative risks over 5.0 are very likely to reflect a true causal relationship (Breslow and Day, 1980).

It is possible that the genetic events that occur in the paternal germ line are affecting the same genes that influence the risk in familial cases. However, there is evidence that this is not the case. First, a number of the loci linked to familial schizophrenia are also associated with bipolar disorder (Craddock et al., 2006), whereas advancing paternal age is specific for schizophrenia (Malaspina et al., 2001). Next, a few genetic studies that separately examined familial and sporadic cases found that the "at-risk haplotypes" linked to familial schizophrenia were unassociated with sporadic cases, including dystrobrevin-binding protein (Van Den Bogaert et al., 2003) and neuregulin (Williams et al., 2003). Segregating sporadic cases from the analyses actually strengthened the magnitude of the genetic association in the familial cases, consistent with etiological heterogeneity between familial and sporadic groups.

Finally, the phenotype of cases with no family history and later paternal age are distinct from familial cases in many studies. For example, only sporadic cases showed a significant improvement in negative symptoms between a "medication-free" and an "antipsychotic treatment" condition (Malaspina et al., 2000), and sporadic cases have significantly more disruptions in their smooth pursuit eye movement quality than familial cases (Malaspina et al., 1998). A recent study also showed differences between the groups in resting regional cerebral blood flow (rCBF) patterns, in comparison with healthy subjects. The sporadic group of cases had greater hypofrontality, with increased medial temporal lobe activity (frontotemporal imbalance), while the familial group evidenced left lateralized temperoparietal hypoperfusion along with widespread rCBF changes in cortico-striato-thalamo-cortical regions (Malaspina et al., 2005b). Other data linking paternal age with frontal pathology in schizophrenia include a proton magnetic resonance spectroscopy study that demonstrated a significant association between prefrontal cortex neuronal integrity (NAA) and paternal age in sporadic cases only, with no significant NAA decrement in the familial schizophrenia group (Kegeles et al., 2005). These findings support the hypothesis that schizophrenia subgroups may have distinct neural underpinnings and that the important changes in some sporadic (paternal germ line) cases may particularly impact on prefrontal cortical functioning.

Genetic mechanism
Several genetic mechanisms might explain the relationship between paternal age and the risk for schizophrenia (see Malaspina, 2001). It could be due to de novo point mutations arising in one or several schizophrenia susceptibility loci. Paternal age is known to be the principal source of new mutations in mammals, likely explained by the constant cell replication cycles that occur in spermatogenesis (James Crow, 2000). Following puberty, spermatogonia undergo some 23 divisions per year. At ages 20 and 40, a man's germ cell precursors will have undergone about 200 and 660 such divisions, respectively. During a man's life, the spermatogonia are vulnerable to DNA damage, and mutations may accumulate in clones of spermatogonia as men age. In contrast, the numbers of such divisions in female germ cells is usually 24, all but the last occurring during fetal life.

Trinucleotide repeat expansions could also underlie the paternal age effect. Repeat expansions have been demonstrated in several neuropsychiatric disorders, including myotonic dystrophy, fragile X syndrome, spinocerebellar ataxias, and Huntington disease. The sex of the transmitting parent is frequently a major factor influencing anticipation, with many disorders showing greater trinucleotide repeat expansion with paternal inheritance (Lindblad and Schalling, 1999; Schols et al., 2004; Duyao et al., 1993). Larger numbers of repeat expansions could be related to chance molecular events during the many cell divisions that occur during spermatogenesis.

Later paternal age might confer a risk for schizophrenia if it was associated with errors in the "imprinting" patterns of paternally inherited alleles. Imprinting is a form of gene regulation in which gene expression in the offspring depends on whether the allele was inherited from the male or female parent. Imprinted genes that are only expressed if paternally inherited alleles are reciprocally silenced at the maternal allele, and vice versa. Imprinting occurs during gametogenesis after the methylation patterns from the previous generation are "erased" and new parent of origin specific methylation patterns are established. Errors in erasure or reestablishment of these imprint patterns may lead to defective gene expression profiles in the offspring. The enzymes responsible for methylating DNA are the DNA methyltransferases, or DNMTs. These enzymes methylate cytosine residues in CpG dinucleotides, usually in the promoter region of genes, typically to reduce the expression of the mRNA. The methylation may become inefficient for a variety of reasons; one possibility is reduced DNA methylation activity in spermatogenesis, since DNMT levels diminish as paternal age increases (Benoit and Trasler, 1994; La Salle et al., 2004). Another possible mechanism is that this declining DNMT activity could be epigenetically transmitted to the offspring of older fathers. There are a number of different DNMTs that differ in whether they initiate or sustain methylation, and which are active at different ages and in different tissues.

Human imprinted genes have a critical role in the growth of the placenta, fetus, and central nervous system, in behavioral development, and in adult body size. It is an appealing hypothesis that loss of normal imprinting of genes critical to neurodevelopment may play a role in schizophrenia. Indeed, one of the most consistently identified molecular abnormalities in schizophrenia has been theorized to result from abnormal epigenetic mechanisms (Veldic et al., 2004), that is, the reduced GABA and reelin expression in prefrontal GABAergic interneurons. An overexpression of DNMT in these GABAergic interneurons, hypermethylating the reelin and GAD67 promoter regions, might be responsible for reducing their mRNA transcripts and expression levels. These decrements could functionally impair the role of GABAergic interneurons in regulating the activity and firing of pyramidal neurons, thereby causing cognitive dysfunction. Later paternal age could be related to the abnormal regulation or expression of DNMT activity in specific cells.

Conclusion
These findings suggest exciting new directions for research into the etiology of schizophrenia. If there is a unitary etiopathology for paternal age-related schizophrenia, then it is likely to be the most common form of the condition in the population and in treatment settings, since genetic linkage and association studies indicate that familial cases are likely to demonstrate significant allelic heterogeneity and varying epistatic effects. Schizophrenia is commonly considered to result from the interplay between genetic susceptibility and environmental exposures, particularly those that occur during fetal development and in adolescence. The data linking paternal age to the risk for schizophrenia indicate that we should expand this event horizon to consider the effects of environmental exposures over the lifespan of the father. The mutational stigmata of an exposure may remain in a spermatogonial cell, and be manifest in the clones of spermatozoa that it will subsequently generate over a man's reproductive life.

References:
Auroux M. Decrease of learning capacity in offspring with increasing paternal age in the rat. Teratology. 1983 Apr;27(2):141-8. Abstract

Benoit G, Trasler JM. Developmental expression of DNA methyltransferase messenger ribonucleic acid, protein, and enzyme activity in the mouse testis. Biol Reprod. 1994 50:1312-9. Abstract

Bradley-Moore M, Abner R, Edwards T, Lira J, Lira A, Mullen T, Paul S, Malaspina D, Brunner D, Gingrich JA. Modeling The Effect Of Advanced Paternal Age On Progeny Behavior In Mice. Developmental Psychobiology, abstract, 2002; (41)3, 230.

Breslow, N. E. and Day, N. E. (1980). The analysis of case-control data. In Statistical Methods in Cancer Research , Volume 1. Lyon: World Health Organization.

Brown AS, Schaefer CA, Wyatt RJ, Begg MD, Goetz R, Bresnahan MA, Harkavy-Friedman J, Gorman JM, Malaspina D, Susser ES. Paternal age and risk of schizophrenia in adult offspring. Am J Psychiatry. 2002 Sep;159(9):1528-33. Abstract

Byrne M, Agerbo E, Ewald H, Eaton WW, Mortensen PB. Parental age and risk of schizophrenia: a case-control study. Arch Gen Psychiatry. 2003 Jul;60(7):673-8. Abstract

Crow JF (1997). The high spontaneous mutation rate: is it a health risk? Proc Natl Acad Sci USA 94:8380-8386.

Craddock N, O'Donovan MC, Owen MJ. Genes for schizophrenia and bipolar disorder? Implications for psychiatric nosology. Schizophr Bull. 2006 Jan;32(1):9-16. Abstract

Dalman C, Allebeck P. Paternal age and schizophrenia: further support for an association. Am J Psychiatry. 2002 Sep;159(9):1591-2. Abstract

Duyao M, Ambrose C, Myers R, Novelletto A, Persichetti F, Frontali M, Folstein S, Ross C, Franz M, Abbott M, et al. Trinucleotide repeat length instability and age of onset in Huntington's disease. Nat Genet. 1993 Aug;4(4):387-92. Abstract

El-Saadi O, Pedersen CB, McNeil TF, Saha S, Welham J, O'Callaghan E, Cantor-Graae E, Chant D, Mortensen PB, McGrath J. Paternal and maternal age as risk factors for psychosis: findings from Denmark, Sweden and Australia.Schizophr Res. 2004 Apr 1;67(2-3):227-36. Abstract

Kegeles LS, Shungu DC, Mao X, Goetz R, Mikell CB, Abi-Dargham A, Laurelle M, Malaspina D. Relationship of age and paternal age to neuronal functional integrity in the prefrontal cortex in schizophrenia determined by proton magnetic resonance spectroscopy. Schizophrenia Bulletin, 31:443; 2005.

La Salle S, Mertineit C, Taketo T, Moens PB, Bestor TH, Trasler JM. Windows for sex-specific methylation marked by DNA methyltransferase expression profiles in mouse germ cells. Dev Biol. 2004 268:403-15. Abstract

Lindblad K, Schalling M. Expanded repeat sequences and disease. Semin Neurol. 1999;19(3):289-99. Abstract

Malaspina D, Friedman JH, Kaufmann C, Bruder G, Amador X, Strauss D, Clark S, Yale S, Lukens E, Thorning H, Goetz R, Gorman J. Psychobiological heterogeneity of familial and sporadic schizophrenia. Biol Psychiatry. 1998 Apr 1;43(7):489-96. Abstract

Malaspina D, Goetz RR, Yale S, Berman A, Friedman JH, Tremeau F, Printz D, Amador X, Johnson J, Brown A, Gorman JM. Relation of familial schizophrenia to negative symptoms but not to the deficit syndrome. Am J Psychiatry. 2000 Jun;157(6):994-1003. Abstract

Malaspina D, Harlap S, Fennig S, Heiman D, Nahon D, Feldman D, Susser ES. Advancing paternal age and the risk of schizophrenia. Arch Gen Psychiatry. 2001 Apr;58(4):361-7. Abstract

Malaspina D. Paternal factors and schizophrenia risk: de novo mutations and imprinting. Schizophr Bull. 2001;27(3):379-93. Review. Abstract

Malaspina D, Corcoran C, Fahim C, Berman A, Harkavy-Friedman J, Yale S, Goetz D, Goetz R, Harlap S, Gorman J. Paternal age and sporadic schizophrenia: evidence for de novo mutations. Am J Med Genet. 2002 Apr 8;114(3):299-303. Abstract

Malaspina D, Harkavy-Friedman J, Corcoran C, Mujica-Parodi L, Printz D, Gorman JM, Van Heertum R. Resting neural activity distinguishes subgroups of schizophrenia patients. Biol Psychiatry. 2005 (a) Dec 15;56(12):931-7. Abstract

Malaspina D, Reichenberg A, Weiser M, Fennig S, Davidson M, Harlap S, Wolitzky R, Rabinowitz J, Susser E, Knobler HY. Paternal age and intelligence: implications for age-related genomic changes in male germ cells. Psychiatr Genet. 2005 (b) Jun;15(2):117-25. Abstract

Reichenberg A, Gross R, Weiser M, Bresnahan M, Silverman J, Harlap, Rabinowitz J, Shulman L, Malaspina D, Lubin G, Knobler HY, Davidson M, Susser E: Advancing paternal age and Autism. Archives of General Psychiatry.

Schols L, Bauer P, Schmidt T, Schulte T, Riess O. Autosomal dominant cerebellar ataxias: clinical features, genetics, and pathogenesis. Lancet Neurol. 2004 May;3(5):291-304. Abstract

Sipos A, Rasmussen F, Harrison G, Tynelius P, Lewis G, Leon DA, Gunnell D. Paternal age and schizophrenia: a population based cohort study. BMJ. 2004 Nov 6;329(7474):1070. Epub 2004 Oct 22. Abstract

Tsuchiya KJ, Takagai S, Kawai M, Matsumoto H, Nakamura K, Minabe Y, Mori N, Takei N. Advanced paternal age associated with an elevated risk for schizophrenia in offspring in a Japanese population. Schizophr Res. 2005 Jul 15;76(2-3):337-42. Epub 2005 Apr 21. Abstract

Van Den Bogaert A, Schumacher J, Schulze TG, Otte AC, Ohlraun S, Kovalenko S, Becker T, Freudenberg J, Jonsson EG, Mattila-Evenden M, Sedvall GC, Czerski PM, Kapelski P, Hauser J, Maier W, Rietschel M, Propping P, Nothen MM, Cichon S. The DTNBP1 (dysbindin) gene contributes to schizophrenia, depending on family history of the disease. Am J Hum Genet. 2003 Dec;73(6):1438-43. Abstract

Veldic M, Caruncho HJ, Liu WS, Davis J, Satta R, Grayson DR, Guidotti A, Costa E. DNA-methyltransferase 1 mRNA is selectively overexpressed in telencephalic GABAergic interneurons of schizophrenia brains. Proc Natl Acad Sci U S A. 2004 Jan 6;101(1):348-53. Abstract

Williams NM, Preece A, Spurlock G, Norton N, Williams HJ, Zammit S, O'Donovan MC, Owen MJ. Support for genetic variation in neuregulin 1 and susceptibility to schizophrenia. Mol Psychiatry. 2003 May;8(5):485-7. Abstract

Zammit S, Allebeck P, Dalman C, Lundberg I, Hemmingson T, Owen MJ, Lewis G. Paternal age and risk for schizophrenia. Br J Psychiatry. 2003 Nov;183:405-8. Abstract

I look forward to reading the comments to this very interesting hypothesis which describes potential mechanisms that may explain why advanced paternal age may contribute to the development of schizophrenia. I wonder whether we may also consider telomere length as a factor in the increased incidence of schizophrenia due to paternal age? Unryn and colleagues report that paternal age is positively linked to telomere length in children, and Yu finds that the shortest TRF length was found in the poor-response schizophrenic group. Is there a connection between paternal age and response to treatment? Are subtelomeric deletions more common as telomere length decreases, and might this explain the decreased learning capacity in the offspring as paternal age increases?

It's also of interest that the Malaspina team reported that increasing duration of marriage had the opposite effect, possibly due to the fact that psychiatric illness reduces schizophrenics' capacity to sustain a marriage. Might we also suspect that paternal ill health due to age-related diseases associated with shorter telomeres such as heart disease may also place stress on the marriage?

References:
Unryn BM, Cook LS, Riabowol KT. Paternal age is positively linked to telomere length of children. Aging Cell. 2005 Apr;4(2):97-101. Abstract

Yu W-Y. Analysis of event-related potentials and telomere length in schizophrenic patients. Master's Thesis. 2001. Unpublished. Abstract