19 February 2008. Emil Kraepelin considered cognitive deficits central to what came to be called schizophrenia, but they appear nowhere in DSM-IV’s list of key criteria for the disorder. In a recent review article in the British Journal of Psychiatry, Michael Kraus and Richard Keefe of Duke University, Durham, North Carolina, argue that these deficits have not received their fair share of attention (Kraus and Keefe, 2007). “However, a renewed interest in cognition has been evident recently, spurred in part by the strong empirical relationship between cognition and real-world functioning,” they write. Cognition’s comeback also reflects its import as a wellspring of potential endophenotypes for schizophrenia. In fact, Timothea Toulopoulou of the Institute of Psychiatry, King's College London, tells SRF, “Neurocognitive deficits have, for a long time now, been among the most promising indicators of increased risk for schizophrenia.”
Endophenotypes, a strategy for discovering the genetic bases of illness, offer a potential way around the heterogeneity of schizophrenia (see SRF live discussion). These quantifiable, biological markers for a propensity to develop an illness may arise from simpler genetic foundations than does the disease itself. Researchers who advocate treating cognitive deficits as schizophrenia endophenotypes note that they appear before the disease becomes clinically apparent, stay stable throughout its course, and do a good job of predicting functional outcomes. Reports that healthy relatives of subjects with schizophrenia show, to a lesser degree, at least some of the cognitive deficits that accompany schizophrenia bolster the case. For example, in a 2007 study in the American Journal of Medical Genetics, Tao Li, first author Xiaohong Ma, both of Sichuan University, Chengdu, China, and their colleagues found impaired attention, speed of information processing, and executive functioning in well relatives from schizophrenia-affected families, whereas subjects with recent-onset schizophrenia showed broad deficits.
With cognition back on the schizophrenia research runway, grab a seat for SRF’s showcase of recent studies on cognitive endophenotypes. First, we will feature two papers in the Archives of General Psychiatry that uphold treating neurocognitive deficits as schizophrenia endophenotypes. In the December issue, a twin study by Toulopoulou and associates points to intelligence and working memory as prime endophenotypes. The November issue contains a family study conducted by the Consortium on the Genetics of Schizophrenia (COGS) that suggests that many neurocognitive measures share genetic underpinnings. After describing these two studies, SRF will spotlight other uses of neurocognitive endophenotypes in schizophrenia research.
The advantages of intelligence and working memory
Toulopoulou and colleagues note that only a few published studies have examined neurocognition in twins with schizophrenia. They write, “None of these have used genetic modeling to explore whether the well-recognized correlation between schizophrenia and cognitive impairment is in fact due to shared genetic factors.” Consequently, their study of 267 twins (from the Maudsley Schizophrenia Twin Study) used genetic modeling to partition the covariance between schizophrenia and cognition into parts that reflect genetic influences, environmental influences shared by co-twins, and unique environmental influences.
Probands with schizophrenia and control twins without psychosis or schizophrenia spectrum disorder underwent testing on the Wechsler Adult Intelligence Scale, yielding measures of IQ and scores for working memory, perceptual organization, verbal comprehension, and processing speed. Analyses compared monozygotic and dizygotic twins who were either concordant or discordant for schizophrenia with healthy monozygotic and dizygotic control twins.
Subjects with schizophrenia, regardless of zygosity or concordance, and nonpsychotic twins from monozygotic discordant pairs performed worse than control subjects on every test. Dizygotic twins lagged control subjects only in perceptual organization. “This is consistent with the idea that the greater the genetic loading is for schizophrenia, the larger and more dispersed is the cognitive impairment,” Toulopoulou and colleagues explain.
For intelligence, genes contributed 70 percent of the variance, versus 65 percent for working memory and 60 percent for perceptual organization. In contrast, they explained less than half of the variance in verbal comprehension and processing speed. Unshared, but not shared, environmental influences accounted for a significant part of the variance in every measure.
Overlapping genetic influences explained 92 percent of the correlation between intelligence and schizophrenia, hinting at one or more shared genes. As for specific cognitive domains, working memory showed the highest genetic correlation with schizophrenia. Other domains showed weaker, yet still significant genetic correlations, with perceptual organization leading the pack. Environmental factors correlated with schizophrenia and cognition separately but failed to explain their covariance. Toulopoulou says, “In a very simple way, what this study shows us is that intelligence and working memory may be the keys to identifying at least some of the genes that contribute to the development of schizophrenia; so, rather than looking for the genes for schizophrenia, we can look to identify the genes that are linked to these neurocognitive deficits.”
Not so discrete after all
The other Archives study comes from the Consortium on the Genetics of Schizophrenia (COGS). David Braff of the University of California, San Diego, leads this seven-site effort to discern the genetic architecture of endophenotypes in families with schizophrenia. It used kinship information on 183 nuclear families that had at least one member with, and a sibling without, schizophrenia.
Subjects underwent neurocognitive and neurophysiological testing. The main neurocognitive measures assessed attention, verbal learning and memory, and working memory. Secondary cognitive tests evaluated abstraction and mental flexibility, memory for faces, emotion recognition, spatial memory, spatial processing, and sensorimotor dexterity. The neurophysiological tests examined prepulse inhibition of the startle response and P50 event-related potential suppression, both of which reflect sensory filtering, as well as the antisaccade task for eye movements, which involves looking in the opposite direction of a suddenly appearing stimulus.
On behalf of COGS, corresponding author Nicholas Schork and first author Tiffany Greenwood of UCSD, La Jolla, write that analyses found all of the neurocognitive and neurophysiologic endophenotypes, with the possible exception of P50 suppression, to be heritable. The portion of variance explained by genes ranged from 24 percent for spatial memory to 55 percent for spatial processing. Toulopoulou thinks that sampling differences may explain why this study found lower heritability estimates than hers did.
Many of the neurocognitive measures seemed to have a common genetic basis. In particular, spatial processing seemed to share genetic influences with all of the other endophenotypes except sensorimotor dexterity, which appeared to be inherited independently of the other measures. A lot of endophenotypes shared environmental influences, too. According to Greenwood and colleagues, “The fact that the endophenotypes are not 100 percent coheritable suggests that there are subtypes of schizophrenia with different endophenotypic profiles.”
So many domains, so little time
While the Greenwood research looked for commonalities between pairs of endophenotypes, a study led by David Glahn, of the University of Texas Health Science Center at San Antonio, sought the best independent predictors of schizophrenia liability. It varied liability by enrolling families who had at least two siblings with schizophrenia or schizoaffective disorder and comparing them to unrelated healthy control subjects. The 269 Latino subjects completed a battery of neurocognitive tests.
As reported in the March 5, 2007, American Journal of Medical Genetics, Part B, only five measures contributed to the unique variance in schizophrenia proneness; three evaluated processing speed, one working memory, and one verbal episodic memory. Digit-Symbol Coding, a test of processing speed, best differentiated subjects according to their presumed genetic vulnerability. However, it had trouble distinguishing between subjects with schizophrenia and their relatives with bipolar disorder. As Glahn and colleagues note, “Such findings could suggest overlapping etiology in schizophrenia and psychotic bipolar disorder.”
Since people who are struggling with a task may do it more slowly, Raquel Gur of the University of Pennsylvania School of Medicine and her collaborators studied both speed and accuracy of performance. In the May American Journal of Psychiatry, they write, “Examining accuracy and speed separately as endophenotypic markers should improve the specificity of detecting and interpreting genetic effects.”
The study of European Americans compared 154 healthy subjects with 349 members of extended multigenerational families that included at least two immediate relatives with schizophrenia. Results suggest that presumed genetic risk for schizophrenia affects accuracy and speed differently in certain domains. For instance, relatives showed poor accuracy but normal speed on the abstraction and flexibility test; for attention, they had normal accuracy but slow speed. The importance of genes also differed for accuracy and speed. Specifically, the accuracy of verbal, face, and spatial memory, as well as emotion processing, seemed to reflect genetic influences, as did the speed of performing tests of abstraction or mental flexibility and attention.
Studies of episodic memory in relatives of individuals with schizophrenia have typically used verbal tasks, so Olalla Robles, of the Hospital General Universitario Gregorio Marañón in Madrid, and associates tried nonverbal ones instead. As reported in the March 1, 2008, Biological Psychiatry, they tested 162 subjects, including patients with schizophrenia or schizophreniform disorder, first-degree relatives of patients with schizophrenia, and healthy control subjects. They further divided relatives and control subjects into those with and without schizophrenia spectrum personality (SSP) traits in order to learn whether cognitive deficits associated with SSP reflect the effects of genes, symptoms, or both.
Subjects who had a family member with schizophrenia showed impaired ability to recognize melodies and line diagrams that had been previously presented. In relatives but not control subjects, the greater the number of SSP traits, the worse the deficits. “With the caveat that a family study design can suggest but not confirm a genetic relationship, we argue that our data suggest that SSP symptomatology in the presence of familial relationship is likely to be genetically informative,” Robles and colleagues write.
These studies point to various candidate endophenotypes, but as Waddington and colleagues (2007) note, the sheer number of deficits related to schizophrenia makes narrowing the field a challenge. To address this problem, Glahn and colleagues advocate doing the two-step: first, show that a measure is heritable; second, use multivariate statistics to identify traits that independently predict schizophrenia risk. This approach should lessen the number of chance findings in linkage studies.
Endophenotypes at work
Meanwhile, other investigators have used endophenotypes to probe the genes and brain changes behind schizophrenia. Some, including Patrick Sullivan of the University of North Carolina, Chapel Hill, and his associates have done so in subjects with schizophrenia. Their study of 641 subjects found that five adjacent single-nucleotide polymorphisms in neural cell adhesion molecule 1 (NCAM1) explained individual differences in composite neurocognitive scores, but only in those of European descent.
Other research groups have compared subjects with and without schizophrenia or related traits in an effort to connect genes to schizophrenia risk or endophenotypes. They include a group of Helsinki scientists, led by Juho Wedenoja of the National Public Health Institute, who found evidence of linked loci on chromosome 7q22 in families with schizophrenia. Efforts to pinpoint the genes involved could not tie reelin; semaphorin 3A; glutamate receptor, metabotropic 3; or VGF nerve growth factor inducible to schizophrenia in subjects from 245 nuclear families. They did, however, relate reelin variants to memory, including working memory, and executive functioning.
A team led by Nicholas Stefanis of the National and Kapodistrian University of Athens, Greece, used not only neurocognitive measures but also schizotypal features as potential endophenotypes. After studying over 2,000 male military conscripts, the researchers concluded that dysbindin (DTNBP1), and to a lesser extent, neuregulin (NRG1) and D-amino-acid oxidase (DAAO) variants “might exert gene-specific modulating effects on schizophrenia endophenotypes at the population level.”
Neurocognitive endophenotypes can also aid structure-function studies, such as one conducted by Philip Szeszko of Zucker Hillside Hospital in Glen Oaks, New York, and his colleagues. It used magnetic resonance imaging to compare the brains of 33 patients with recent-onset schizophrenia and 30 healthy subjects. Fractional anisotropy revealed white matter abnormalities in the temporal lobes of subjects with schizophrenia. Analyses linked disturbances in the bilateral uncinate fasciculus to impaired verbal learning and memory as well as to more severe negative symptoms.
Whether neurocognitive endophenotypes will lead to those twin grails of schizophrenia research—an understanding of the disorder’s etiology and the identification of molecular targets for treatment—remains to be seen, but these studies offer a glimpse of the possibilities and challenges inherent in using them. As for what’s next, Toulopoulou says, “Whole-genome searches using comprehensive factorial designs stratifying for levels, say, of intelligence, should be able to help to characterize genes that are specific to schizophrenia, are specific to IQ, or influence both.”—Victoria L. Wilcox.
Greenwood TA, Braff DL, Light GA, Cadenhead KS, Calkins ME, Dobie DJ, Freedman R, Green MF, Gur RE, Gur RC, Mintz J, Nuechterlein KH, Olincy A, Radant AD, Seidman LJ, Siever LJ, Silverman JM, Stone WS, Swerdlow NR, Tsuang DW, Tsuang MT, Turetsky BI, Schork NJ. Initial heritability analyses of endophenotypic measures for schizophrenia. Arch Gen Psychiatry. 2007 Nov;64(11):1242-1250. Abstract
Toulopoulou T, Picchioni M, Rijsdijk F, Hua-Hall M, Ettinger U, Sham P, Murray R. Substantial genetic overlap between neurocognition and schizophrenia. Arch Gen Psychiatry. 2007 Dec;4(12):1348-1355. Abstract
Sullivan PF, Keefe RSE, Lange LA, Lange EM, Stroup TS, Lieberman J, Maness PF. NCAM1 and neurocognition in schizophrenia. Biol Psychiatry. 2007 Apr 1;61(7):902-910. Abstract
Gur RE, Nimgaonkar VL, Almasy L, Calkins ME, Ragland JD, Pogue-Geile MF, Kanes S, Blangero J, Gur RC. Neurocognitive endophenotypes in a multiplex multigenerational family study of schizophrenia. Am J Psychiatry. 2007 May;164(5):813-819.
Stefanis NC, Trikalinos TA, Avramopoulos D, Smyrnis N, Evdokimidis I, Ntzani EE, Ioannidis JP, Stefanis CN. Impact of schizophrenia candidate genes on schizotypy and cognitive endophenotypes at the population level. Biol Psychiatry. 2007 Oct 1;62(7):784-792. Abstract
Wedenoja J, Loukola A, Tuulio-Henriksson A, Paunio T, Ekelund J, Silander K, Varilo T, Heikkilä K, Suvisaari J, Partonen T, Lönnqvist J, Peltonen L. Replication of linkage on chromosome 7q22 and association of the regional Reelin gene with working memory in schizophrenia families. Mol Psychiatry. 2007 Aug 7 (Epub ahead of print). Abstract
Szeszko PR, Robinson DG, Ashtari M, Vogel J, Betensky J, Sevy S, Ardekani BA, Lencz T, Malhotra AK, McCormack J, Miller R, Lim KO, Gunduz-Bruce H, Kane JM, Bilder RM. Clinical and neuropsychological correlates of white matter abnormalities in recent onset schizophrenia. Neuropsychopharmacology. 2007 June 20 (Epub ahead of print).
Ma X, Wang Q, Sham PC, Liu X, Rabe-Hesketh S, Sun X, Hu J, Meng H, Chen W, Chen EYH, Deng W, Chan RCK, Murray RM, Collier DA, Li T. Neurocognitive deficits in first-episode schizophrenic patients and their first-degree relatives. Am J Med Genet B Neuropsychiatr Genet. 2007 Jun 5;144(4):407-416.
Glahn DC, Almasy L, Blangero J, Burk GM, Estrada J, Peralta JM, Meyenberg N, Castro MP, Barrett J, Nicolini H, Raventós H, Escamilla MA. Adjudicating neurocognitive endophenotypes for schizophrenia. Am J Med Genet B Neuropsychiatr Genet. 2007 Mar 5;144(2):242-249. Abstract
Robles O, Blaxton T, Adami H, Arango C, Thaker G, Gold J. Nonverbal delayed recognition in the relatives of schizophrenia patients with or without schizophrenia spectrum. Biol Psychiatry. 2008 March 1;63(5):498-504. Abstract