Medscape: Tell me about your research on paternal age and schizophrenia.

Dolores Malaspina, MD, MPH: I have been compelled by the idea that schizophrenia is not a single disease. The consensus in the field is that schizophrenia is a syndrome, and a syndrome is a collection of different disorders. Yet there is still some controversy over whether or not there are variants of schizophrenia that might have separate causes and respond differently to various medications.

Since beginning my research in the late 1980s, I have focused on this heterogeneity, and one way that I've done that is by examining aspects of the disease in people who come from densely affected families, where 2 or more relatives have schizophrenia, and comparing them with cases of schizophrenia that have no family history of any chronic psychosis.

Now, in genetic research, it's known that for human genetic diseases, when a new case presents itself in a family, the mutation almost always arises during spermatogenesis. We have known for almost 100 years that the late born children in a family have more new genetic diseases. In the 1950s, a scientist named Penrose showed that only the age of the father predicts these genetic diseases. Over the last decade, it was shown that the risk for many complex genetic diseases was also correlated with paternal age. I thought that if schizophrenia cases with no family history were due to new genetic events, maybe they would also be correlated with the father's age.

I have the good fortune to be funded by the National Institutes of Health to study a very special birth cohort in Israel of about 100,000 pregnancies. We have a rich amount of demographic and clinical data on the parents, including the age of the father. The analysis showed what we considered to be a striking effect of the age of the father on the risk for schizophrenia.

Medscape: Could you tell me more about this group of research subjects from Israel?

Dr. Malaspina: The offspring were born between 1964 and 1976, and the original birth cohort was designed to examine the health of women during pregnancy as well as fetal outcomes. Israel maintains a high-quality psychiatric case registry. Working with the people at the Ministry of Health in Israel, my colleagues linked the birth cohort data to the psychiatric case registry data. The results showed that the risk of schizophrenia was tripled for the offspring of the oldest group of fathers.

We found that paternal age explained over a quarter of the risk for schizophrenia in the population. At the time, people were skeptical. But the findings have been replicated many times now, and not a single study has failed to find this strong relationship between father's age and the risk for schizophrenia. And at this point, other explanations for the relationship have been ruled out, including social factors in the family, prenatal care, and parental psychiatric ailments. There simply seems to be a relationship between paternal age and schizophrenia risk.

Medscape: Can you explain why that relationship between paternal age and schizophrenia exists?

Dr. Malaspina: When Penrose found that paternal age predicted new human genetic diseases, he proposed the Copy Error Theory. He said that each time the spermatozoa are copied there's an opportunity for a new mutation. Sperm cells divide every 16 days after puberty, so the DNA in the sperm of a 20-year-old father has been copied 100 times, but sperm DNA from a 50-year-old father has been copied more than 800 times. By comparison, egg cells from the mother only undergo a few dozen cell divisions all together. It is clear that there are many more opportunities for mutations to occur during spermatogenesis and that these increase with the age of the father. That is why new mutations are introduced in mammals in proportion to paternal age.

To further establish that paternal age is associated with schizophrenia risk, we went back to examine if paternal age is related to other factors associated with schizophrenia risk. We looked at intellectual functioning at age 17 in our birth cohort. Those data were available because adolescents in Israel are screened for military service. Working with personnel at the Israeli Defense Force, we examined whether intelligence was related to paternal age. And what we found was a very strong specific effect of paternal age on performance IQ. Very young mothers and very old mothers had offspring with impairments in verbal and performance intelligence. While there was no effect of late fathers' age on verbal IQ, there was a strong effect on performance intelligence, or nonverbal intelligence, which we have published.

In a parallel study, we examined the effect of late paternal age in a mouse model. Working with my colleague, Jay Gingrich, we studied several cohorts of inbred mice to compare offspring with younger and older fathers. The mouse model demonstrated striking effects of paternal age on the behavior of mice.

Those 3 lines of evidence provide converging data that paternal age does influence neural functioning and that paternal age is a plausible risk factor for schizophrenia.

Medscape: Could you describe what is meant by sporadic schizophrenia and how that relates to paternal age?

Dr. Malaspina: This goes along with the issue of whether schizophrenia is one single disease or several different variants, several different diseases. If it is several diseases, we could make much more progress if we knew how to separate individuals who have 1 variant of the disease from individuals who have the other variant, such as for treatment studies.

So, we have this finding that father's age predicts schizophrenia, but we don't know if the genetic changes are in the same genes that cause familial schizophrenia or if they occur at a different place. Some of the birth cohorts have actually looked to see how the risk of schizophrenia with paternal age is related to the family history of schizophrenia. The finding is that father's age is not connected to the risk of schizophrenia when it runs in families, but only for cases with no family history. That is called sporadic schizophrenia.

We have also looked at patients, with the help of funding from the National Alliance for Research on Schizophrenia and Depression, and we have examined whether or not cases with late paternal age and no family history have different symptoms and brain abnormalities from those of other cases. That work is under way.

Medscape: You also looked at the duration of the parents' marriage.

Dr. Malaspina: Yes, and we found that the duration of marriage was protective against the risk for schizophrenia. This goes in the opposite direction of paternal age, but it's an independent factor. Couples that have a very long marriage are less likely to have offspring with schizophrenia. One possibility is that parents who have mental disorders themselves may have shorter marriages. Another possibility is that there is an increased risk of schizophrenia when there is a marital separation.

Medscape: A variety of environmental factors can influence the development of schizophrenia. How do you control for that?

Dr. Malaspina: On the one hand, there may be scores of different intrauterine exposures that increase the risk for schizophrenia through different pathways. Another possibility, though, is that there are only a few final common pathways through which various intrauterine adversities are linked to the risk for schizophrenia.

The Barker hypothesis deals with the area of fetal programming. Research shows that the risk for many adult-onset chronic diseases, such as cardiovascular disease, obesity, diabetes, and hypertension, is related to fetal development. The mechanism may be that an adverse fetal environment compromises the development of organs and tissues and changes lifelong gene expression. The fetus survives, but its health is compromised. Effects on the developing nervous system could contribute to schizophrenia risk. So that's a possible pathway for the risk for schizophrenia, through a variety of prenatal exposures.

The benefit of our study in Israel is that we had such a wealth of obstetric data. The birth cohort involved early pregnancy interviews with the mom. It also involved evaluations of the progress of the pregnancy and records of the delivery. Our study was able to show that other prenatal exposures did not explain the linkage of paternal age to the risk of schizophrenia. Also, there have been many excellent studies after ours was conducted that have looked at numerous fetal exposures and found that those also do not explain the risk of paternal age for schizophrenia.

I do, however, believe that many fetal exposures can increase the risk of schizophrenia. I would suggest that the mechanism of these events may be via changes in lifelong gene expression.

Medscape: What about the influence of environmental factors after birth, during childhood and adolescence?

Dr. Malaspina: I think 3 of the interesting factors that have been linked to the risk of schizophrenia are severe stress in a stress-sensitive person who has underlying genes for schizophrenia, traumatic brain injury in those with underlying genes for schizophrenia, and, very importantly, cannabis exposure in early adolescence.

Medscape: Your research about paternal age became public in 2001. Do you think fewer men over a certain age might choose to have children as a result?

Dr. Malaspina: I haven't heard that. I would personally not discourage anyone from having a child at any age. People weigh their own risks. For the offspring of older fathers, the risk of schizophrenia is about 3%. That means that 97% of the offspring do not have schizophrenia. Other cognitive diseases linked to paternal age include mental retardation of unknown etiology and Alzheimer's disease, and there is a strong relationship between paternal age and autism.

Medscape: What do you expect to be the future of your research in this area?

Dr. Malaspina: The genes for schizophrenia that we have identified lately are very interesting; they explain a large degree of the risk of the disease. Attention probably should turn toward factors that affect the expression of these genes and other genes. This is the area of epigenetics, the code that determines whether or not genes will be expressed.

We're pursuing a gene expression hypothesis for paternal age and schizophrenia. Humans have dozens or hundreds of genes that are expressed, not on the basis of being dominant or recessive, but on the basis of which parent we have inherited them from. So genes that control the growth of the fetus tend to be expressed on the basis of inheritance from the father. Other genes are expressed only on the basis of inheritance from the mother. These are called "parent of origin genes" or "parentally-imprinted genes." In these genes, the father's copy is expressed and the mother's is silenced, or vice versa. We are interested in this mechanism of gene-silencing. For the male parent, the silencing, or the activation/expression of genes from dad, takes place late in spermatogenesis. So our hypothesis and model right now for how paternal age affects the risk for schizophrenia is that it has altered the expression of genes inherited from the father.

Even exposures that interact with genetic susceptibility may act by changing gene expression, such as traumatic brain injury, cannabis, and stress. Maybe we can integrate our understanding of the many exposures tied to schizophrenia and the many genes tied to schizophrenia with the understanding that certain exposures may act by changing gene expression.

Meanwhile, some individuals who develop schizophrenia have a good outcome and stability without much deterioration -- but not as many as we would like. If we can't prevent the disease, perhaps we can learn the risk factors for deterioration and how to prevent it.

Although I see schizophrenia as a syndrome of separate illness variants, I think the field has benefitted from considering it as a single disease. From here forwards, we may be diluting our ability to find risk factors and optimize outcome by considering the disease as a whole. To go forward in schizophrenia, we need to better understand how similar symptoms may arise from abnormalities in different neural circuits; that the set of symptoms we call schizophrenia could reflect a common pathway, but that the underlying biology may differ for groups of people, and that those differences may explain which medications they should receive, or which factors are more adverse for them. I think the field needs to move toward a finer understanding of the variants that exist. The identified genes may be clearly explanatory for some cases but not for others.

This interview is published in collaboration with NARSAD, The Mental Health Research Association, and is supported by an educational grant from Pfizer.

Dolores Malaspina, MD, Professor of Clinical Psychiatry, Columbia University, New York, NY; Director of Clinical Neurobiology, New York State Psychiatric Institute and Columbia University Medical Center, New York, NY

Disclosure: Jessica Gould has disclosed no relevant financial relationships.

Disclosure: Dolores Malaspina, MD, has disclosed no relevant financial relationships.