Address correspondence and reprint requests to Eero Castrén, A.I. Virtanen Institute, University of Kuopio, PO Box 1627, 70211 Kuopio, Finland. E-mail: firstname.lastname@example.org
Antipsychotic drug treatment is known to modulate gene expression in experimental animals. In this study, candidate target genes for antipsychotic drug action were searched using microarrays after acute clozapine treatment (1, 6 and 24 h) in the rat prefrontal cortex. Microarray data clustering with a self-organizing map algorithm revealed differential expression of genes involved in presynaptic function following acute clozapine treatment. The differential expression of 35 genes most profoundly regulated in expression arrays was further examined using in situ hybridization following acute clozapine, and chronic clozapine and haloperidol treatments. Acute administration of clozapine regulated the expression of chromogranin A, synaptotagmin V and calcineurin A mRNAs in the cortex. Chronic clozapine treatment induced differential cortical expression of chromogranin A, son of sevenless (SoS) and Sec-1. Chronic treatment with haloperidol regulated the mRNA expression of inhibitor of DNA-binding 2 (ID-2) and Rab-12. Furthermore, the expression of visinin-like proteins-1, -2 and -3 was regulated by chronic drug treatments in various brain regions. Our data suggest that acute and chronic treatments with haloperidol and clozapine modulate the expression of genes involved in synaptic function and in regulation of intracellular Ca2+ in cortex.
Here, we have investigated alterations in gene expression patterns produced by antipsychotic drug treatment in the rat cortex. cDNA expression arrays of more than 1000 genes were used to screen for changes in gene expression in the PFC at 1, 6 and 24 h after acute clozapine administration. Cluster analysis indicated that genes related to presynaptic function were regulated by acute clozapine in the PFC. In situ hybridization was used to examine the regulation of 35 genes altered in the array experiments and the expression of the same set of genes were also investigated after chronic clozapine and haloperidol administration. Several genes related to synaptic function and intracellular Ca2+ metabolism were found to be altered in response to clozapine and haloperidol treatments.
Animal care and drug administration
All animal experiments were performed in accordance to the guidelines of the Society for Neuroscience and were accepted by the experimental animal ethics committee of the University of Kuopio. Male Wistar rats (∼200 g, National Laboratory Animal Center, University of Kuopio, Kuopio, Finland) were kept under standardized temperature, humidity and lighting conditions with free access to food and water. For acute treatment, rats were given a single intraperitoneal injection either with saline (1 mL/kg, control treatment) or clozapine (25 mg/kg, Leponex, Wander Pharma, Novartis Pharma GmbH, Nürnberg, Germany), and were killed 1, 6, or 24 h after the drug administration by decapitation under CO2 anaesthesia. For expression array studies after acute clozapine treatment, three animals were treated per treatment group. The reference group of saline treated control animals for the expression study was killed 24 h following a single injection. Chronic treatments with haloperidol (1 mg/kg, Serenase, Orion, Finland) or clozapine were performed by daily drug injections for 17d. After the last of chronic drug injections the animals were killed following washout periods of 2 h, 24 h, or 6 days. In situ hybridization studies were carried through with 3–6 animals per treatment group, and saline-treated control animals were matched for each time period of antipsychotic drug treatment.
Rat brain prefrontal and anterior cingulate cortices were dissected, flash frozen and stored at −75°C. RNA extraction was performed using AtlasTM Pure Total RNA Labeling System (Clontech, Palo Alto, CA, USA) according to the manufacturer's recommendations. Three animals per group were pooled and homogenized in denaturing solution with a Polytron. Total RNA was isolated from tissue homogenates with phenol–chloroform extraction and dissolved in RNase-free water. DNase I-treated total RNA was precipitated by a second round of phenol–chloroform extraction and the quality of RNA was assessed by denaturing gel analysis and absorbance measurements at A260/A280. Total RNA samples exhibited A260/A280 ratios of 1.8 or higher with no degradation visible by denaturing gel analysis. Poly A+ RNA was enriched by oligo(dT) separation followed by the cDNA probe synthesis using a gene specific cDNA Synthesis Primer Mix (AtlasTM Pure Total RNA Labeling System) and incorporation of α-[33P]dATP (NEN Life Science Products, Inc., Boston, MA, USA) by MMLV reverse transcriptase. Unincorporated nucleotides were removed by column chromatography according to the manufacturer's protocol.
Filter array hybridization
Each cDNA probe was used to hybridize an Atlas Rat 1.2 Array filter (Clontech) of 1176 sequence verified cDNAs. Each experiment compared the expression profiles of 1-, 6-, or 24-h clozapine treated sample to a 24-h saline-treated control sample. The hybridization was performed according to the manufacturer's recommendations and filters were exposed onto phosphorscreen and scanned with STORM 860 PhosphorImager (Molecular Dynamics Inc., Sunnyvale, CA, USA) at 50 µm resolution. The filter arrays were stripped and rehybridized up to three times according to recommended protocols in order to compare the expression profiles of all samples under study. In our experience, stripping and rehybridization of the filters up to three times did not affect hybridization efficiencies.
Filter array data analysis and visualization with self-organizing maps
Digitized filter array images were imported into a data analysis program for image quantitation. The quantitation was performed with ArrayVision software (Imaging Research Inc., St. Catharine's, Canada). Local background was subtracted from spot intensity values. Data were globally normalized by scaling the average intensity of sample and control measurement groups to the same value. In order to compensate for increased variance among ratios calculated from small values, we used a ratio with a correcting factor:
where x is the normalized measurement for a particular gene and k is a constant value of 0.2 multiplied by the average intensity. The size of k was defined by plotting the log-base 2 (ratio) against average with different k-values (Dudoit et al. 2000).
Expression patterns formed by the log-base 2-transformed ratios of three time points (1, 6 and 24 h) of clozapine-treated samples over control treatment were further analysed using the self-organizing map (SOM) algorithm (Kohonen 1997) which has previously been shown to be applicable for analysis of gene expression data (Tamayo et al. 1999; Törönen et al. 1999). Genes were placed into 64 nodes of a rectangular 8 × 8 grid map. Each node contains genes with the most similar expression pattern over the first 24 h of clozapine treatment. Nodes with similar average expression patterns are located next to each other while the least similar groups are placed further apart on the map. This neighbourhood function of SOM facilitates the visualization of the data and allows the recognition of larger clusters of similarly regulated genes formed by neighbouring nodes.
The distribution of genes belonging to defined functional classes in different nodes by SOM was then evaluated. It is important to note that these classifications were not used in the clustering of the expression patterns by SOM but only in the visualization of the results. A classical statistical method, sampling without replacement (Arnold 1990), was used to estimate the probability for finding the observed or stronger correlation between a cluster and a functional class as a result of random sampling. p-value was calculated using a cumulative hypergeometric distribution function similarly to the analysis of contingency tables using Fisher's exact test (Agresti 1992). To enhance reliability, we took advantage of the neighbourhood function of SOM and focused on clusters of nodes, where a low probability (indicating enrichment of genes belonging to a given functional class) was found in several neighbouring nodes.
Functional gene classes
Functionally related genes were grouped into classes for visualization of the clustered data. The gene groups that were studied in this paper were γ-amino butyric acid receptors (GABA) and genes related to GABA neurotransmission, glutamate receptors and genes related to glutamate-mediated neurotransmission, G-protein coupled receptors, genes involved in lipid metabolism and genes related to presynaptic function (PSYN). Gene groups were based on the gene classification provided by the array manufacturer, or on previous literature.
The PSYN gene group was a modification of the selection of presynaptic gene class recently published by Mirnics et al. (2000). The filter array contains 55 genes that were categorized into the PSYN class genes and were identified as follows: neurexins (neurexin I beta, non-processed neurexin II alpha/beta, non-processed neurexin III alpha), rab-family of genes (1 A, 2, 3 A, 3B, 4 A, 4B, 8, 11 A, 12, 13, 14, 15, 16, 26), RIM, RAB GDP dissociation inhibitor (alpha and beta), RAB-related GTP-binding protein, syntaxins (1 A, 1B, 2, 3, 4, 5, 6), Sec1 (syntaxin binding protein 1), Munc18–2 (syntaxin binding protein), SNAP-25, synaptotagmins (II, III, IV, V, XI), synapsins (1 and 2), synaptophysin, synaptobrevins (1 and 2), synaptic vesicle protein (2 and 2B), synaptic vesicle amine transporter 2, chromogranin A, secretogranins (II, III, V), annexins (I, II, III, IV, V), voltage-dependent P/Q-type calcium channel alpha-1 A subunit and neurophilin. The PSYN class genes cannot be strictly assigned only for presynaptic function, but may also be involved in membrane trafficking.
In situ hybridization
A selection of candidate genes from the array data analysis was chosen for further investigation with in situ hybridization based on the magnitude of their expression ratios or their functional similarity to highly regulated genes. In situ hybridization was performed as described previously (Kontkanen et al. 2002). Briefly, the brains were collected after appropriate saline or antipsychotic drug treatment periods (acute treatments for 1, 6 or 24 h, and chronic treatments for 17 day with 2-h, 24-h or 6-day washout periods), frozen on dry ice, and stored at −75°C prior to cutting into 14-µm thick sections in coronal orientation with a cryostat. One control and one drug treated section were mounted onto the same SuperFrost® microscope slide (Menzel-Gläser, Germany), postfixed, dehydrated, and stored at +4°C until used. The sections were hybridized with oligonucleotide probes that had been labelled with [α-33P]dATP by terminal deoxynucleotidyl transferase (MBI Fermentas, Vilnus, Lithuania). Hybridization was performed with 1–3 × 103 cpm/µL of labelled probe at +42°C overnight. After incubation the sections were washed and opposed onto Hyperfilm-βmax films (Amersham, Buckinghamshire, UK) from 3 days up to 2 weeks and developed for 5 min in D-19 (Kodak, Rochester, NY, USA) (Wisden and Morris 1994).
Hybridized sections were quantified with MCIDTM image analysis software (Imaging Research Inc.) and optical densities were converted to specific binding using 14C scaling. The quantified brain regions are schematically shown in Figs 1(a and b). Superficial and deep cortical layers were quantified separately. All brain regions were defined according to rat brain atlas by Paxinos and Watson (1986). Hybridization intensities were quantified from matching brain regions of drug-treated and identically saline-treated control sections. The data are mean values of specific binding of all sections from the treatment group, and are presented as percentage of the mean specific binding of the corresponding control sections. Unless otherwise stated, analysis of variance (anova) followed by two-tailed Student's t-test was used to analyse the differences after chronic antipsychotic drug treatments, and Student's t-test was applied for the acute treatment effects.
We have performed an expression analysis using microarrays of 1176 genes to search for genes regulated in the PFC by acute clozapine treatment. Two different strategies were employed in the data analysis: First, cluster analysis of gene expression patterns using SOM algorithm was performed to identify coregulated genes within functional gene groups. Second, the altered expression of 35 genes with the most substantial regulation in the filter array experiments or their close functional relatives was examined using in situ hybridization. Furthermore, the potential changes in the expression of same set of 35 genes were also evaluated in rat brain sections treated chronically with either clozapine or haloperidol.
Expression array data analysis
Gene expression patterns were produced after three acute treatment periods with clozapine using expression arrays in the rat brain anterior cingulate and prefrontal cortices (Table S1; supplementary material). The expression profiles of 1176 genes over 1-, 6- and 24-h clozapine treatment periods were submitted to an unsupervised cluster analysis using the SOM algorithm. The data were clustered into a map of 64 nodes (Fig. 2a). The number of genes assigned to different nodes varied between 10 and 20 genes, and the most populated node represented genes that showed no change in their expression level after any of the treatment time points (dotted node in the middle of the map, Fig. 2a).
Table 1. Presynaptic (PSYN) genes present in clusters A and B by data clustering with a self-organizing map algorithm
PSYN genes localized to cluster A
PSYN genes localized to cluster B
Expression ratios indicate normalized, non-log2–transformed ratios of clozapine treated animals over control treated animals. The highest and lowest expression ratios were at 24-h clozapine treatment, except the ratios marked with (*) were at 6-h clozapine treatment. Note that genes were clustered to this group based on the shape of the expression profile, not by the height of the maximum ratio. RAB, Ras-related protein.
Visualization of the gene expression patterns of genes related to presynaptic function (PSYN) revealed specific clustering of this class of gene expression profiles. The PSYN class genes were preferentially clustered into two clearly distinct node clusters (A and B) with opposite expression profiles (Fig. 2a). Cluster A consists of a total of 154 genes of which 14 genes belong to the class of PSYN genes (Table 1). Statistical analysis indicated that the number of PSYN genes found in cluster A was significantly higher than what would be expected by chance (p = 0.0083). The PSYN genes of cluster A are up-regulated at 6 h after clozapine injection, and either remain increased at 24 h or return towards baseline (Fig. 2b).
Enrichment of the PSYN class genes in cluster B was also highly significant (p = 0.0071). Cluster B consists of a total of 63 genes of which eight genes are classified as PSYN genes (Table 1, section B). In addition, PSYN genes were enriched in nodes surrounding the cluster B region. Due to the neighbouring function of the SOM, enlargement of the selected area for cluster B still resulted in statistically significant PSYN gene enrichment (data not shown). Genes that were grouped in cluster B were characterized by a decreased expression pattern following a single clozapine injection (Fig. 2c).
We also evaluated clustering in gene groups of neurotransmitter receptors (GABA and glutamate receptors), G-protein coupled receptors, or in genes involved in lipid metabolism. No significant clustering within these functional gene groups was observed (data not shown).
In situ hybridization analysis of acute clozapine treatment
As a second, independent analysis of gene expression changes produced by acute clozapine treatment, we used in situ hybridization to validate the regulation of 35 genes with the most pronounced changes in the filter array experiments or their close functional relatives (Table S1). These genes belonged to many different functional classes and their membership in a particular functional class was not used as a criterion for their selection. In situ hybridization also allowed us to evaluate potential changes in other brain areas outside the PFC. A modest but consistent increase (∼15%) of chromogranin A (PSYN cluster A, ratio difference 2.23 at 6 h) mRNA was observed in the outer (p < 0.05) and inner (p < 0.05) layers of the PFC at 6 h after a single clozapine injection that was in agreement with the filter array results (Fig. 3a). Synaptotagmin V (PSYN cluster B, ratio difference 0.46 at 24 h) mRNA was down-regulated after 24-h clozapine treatment in the inner layers of the frontal cortex (∼30%, p < 0.001), and in the outer and inner layers of the parietal cortex (∼15–20% reduction, p < 0.05 and p < 0.01, respectively) (Fig. 3b). These results were consistent with the filter array results. Array analysis suggested that the expression of calcineurin A might be induced by clozapine (ratio difference of 2.17 at 6 h). In situ hybridization analysis confirmed the up-regulation (∼20%) of the mRNA in response to 6-h clozapine treatment in the inner layers of the infralimbic cortex (p < 0.05) (data not shown). The increase was more pronounced (∼70%) after 24-h clozapine treatment in the inner layers of the frontal cortex (p < 0.001) (Fig. 3c).
In situ hybridization analysis of chronic clozapine and haloperidol treatments
Next, we used in situ hybridization to investigate the expression of the selected 35 candidate genes in brain sections of rats chronically (17 day) treated with clozapine or haloperidol with various washout periods (2 h, 24 h, or 6 day). The long (6-day) washout period was included based on our earlier observations that fos and jun family genes remain up-regulated up to day 6 following chronic antipsychotic drug treatment (Kontkanen et al. 2002). Altogether, six genes were confirmed as differentially expressed by the chronic drug treatments. Chronic treatment with clozapine regulated the expression of chromogranin A (Chga), son of sevenless (SoS) and Sec-1. Chromogranin A mRNA was significantly down-regulated in the parietal cortex by chronic treatment with a 2-h washout period (∼25%, p < 0.01) (Table 2 and Fig. 4a). Expression of SoS (ratio difference of 0.50 at 24-h clozapine) was increased after a 2-h washout period in the anterior cingulate and frontal cortices (∼10% and ∼20%, p < 0.05 and 0.001, respectively), and decreased after a 6-day washout period in the PFC (∼15%, p < 0.05) (Table 2). Sec-1 mRNA was increased in the parietal cortex (∼20%, p < 0.05) and decreased in the PFC (∼20%, p < 0.05) after a 2-h washout period following chronic clozapine treatment (Table 2).
Table 2. The effects of chronic clozapine and haloperidol treatments on the candidate gene expression
n.d., not determined; PFC, prefrontal cortex; SoS, son of sevenless; VSNL-2, visinin-like protein-2. Ant cg, anterior cingulate cortex; Chg A, chromogranin A; dw, day washout; Fr cx, frontal cortex; hw, hour wash-out; IL cx, infralimbic cortex; Par cx, parietal cortex; RAB-12, Ras-related protein 12; Sec-1, syntaxin binding protein 1. All values (± SEM) are expressed as percentage of the saline-treated treatment period matched control sections (n = 3–6, anova followed by two-tailed Student's t-test). *p < 0.05, **p < 0.01, ***p < 0.001.
Chronic administration of haloperidol regulated the expression of inhibitor of DNA-binding 2 (ID2), Rab-12 and visinin-like protein 2 (VSNL-2). Expression of ID-2 (expression ratio of 1.73 at 24-h clozapine treatment) was increased after a 2-h washout period in the frontal cortex (∼25%, p < 0.05) and decreased after a 24-h washout period in the parietal cortex (∼20%, p < 0.05) (Table 2). The expression of Rab-12 mRNA (expression ratio of 0.59 at 24 clozapine treatment) was increased by chronic haloperidol both after a 2-h washout period in the PFC (∼25%, p < 0.05) and after a 24-h washout period in the anterior cingulate and frontal cortices (∼25% and ∼30%, p < 0.01 and 0.05, respectively) (Table 2). Furthermore, chronic administration of haloperidol [17 days 2 hour wash-out (hw)] down-regulated the expression of VSNL-2 mRNA (expression ratio of 0.83 at 24-h clozapine treatment) in all cortical areas examined (Table 2 and Fig. 4c).
We also performed additional in situ hybridization analyses to examine the expression of two other visinin-like protein family members, visinin-like proteins 1 and 3 (VSNL-1 and -3) after the chronic drug treatments. Indeed, the expression of VSNL-1 mRNA (expression ratio of 0.76 at 24-h clozapine treatment) was decreased after chronic haloperidol with a 2-h washout period in the core of nucleus accumbens (data not shown, p < 0.05, Student's t-test). In addition, chronic clozapine with a 6-day washout period reduced the expression of VSNL-3 (expression ratio of 0.56 at 24-h clozapine treatment) in the prefrontal and infralimbic cortices (∼20%, p < 0.05 and ∼30%, p < 0.01 decrease, respectively, Student's t-test) (Fig. 4b).
A number of genes have been reported to be regulated in response to antipsychotic drug treatment. Particularly well studied in this respect are members of Fos and Jun family of transcription factors, which are regulated in many brain areas and bind to their cognate AP-1 DNA binding site in response to typical and atypical antipsychotic drugs (Miller 1990; Merchant and Dorsa 1993; Deutch and Duman 1996; Herdegen and Leah 1998). In this study, we have applied expression arrays to search for genes which are regulated by the activation of these transcription factors within 24 h of antipsychotic drug treatment. Clozapine was selected because it has been shown to regulate the expression of immediate early genes in the PFC (Deutch and Duman 1996; Robertson and Fibiger 1992; Merchant and Dorsa 1993; Robertson et al. 1994), the brain region which has been implicated in the pathophysiology of schizophrenia (Benes 2000; Bunney and Bunney 2000). Cluster analysis identified several genes involved in the presynaptic function to be regulated by antipsychotics, and the regulation of two of these genes, chromogranin A and synaptotagmin V, was confirmed using in situ hybridization. Further in situ hybridization experiments revealed that several genes identified in the array experiments and involved in synaptic function and intracellular Ca2+ metabolism were also regulated after chronic treatment of clozapine or haloperidol. These results are of interest in the light of recent observations which have identified altered expression of genes involved in synaptic function in the brain of schizophrenic patients.
Examination of changes in gene expression patterns over several experimental conditions gives a more comprehensive view on biologically meaningful differences and is less prone to experimental error than concentration on regulation at a single time point (Brown and Botstein 1999). Observation of clusters produced by our data revealed a specific regulation of genes related to presynaptic signalling. The expression profiles of PSYN class genes were enriched in both up-regulated and down-regulated patterns on the SOM. It makes biologically sense that changes in a functional system would likely not induce differential gene expression only to one direction. The change in the expression of PSYN genes in both clusters was the most profound after 24-h treatment period. This suggests that genes involved in presynaptic function are affected after a short delay following clozapine administration.
We further used in situ hybridization to examine the expression patterns of the candidate genes suggested by the array experiments after various periods of acute and chronic treatment with haloperidol and clozapine. A total of 35 candidate genes were chosen for in situ hybridization studies. Differentially regulated mRNA expression patterns were observed from several of the studied genes. Chromogranin A mRNA expression was increased acutely in the PFC whereas long-term treatment resulted in reduced expression in the parietal cortex. Chromogranin A is a calcium-binding protein enriched in large dense core vesicles (Yoo 2000) and it controls vesicle formation and secretion (Kim et al. 2001). Chronic administration of clozapine has been previously reported to induce chromogranin A mRNA in the striatum, nucleus accumbens and dorsal raphe nucleus (Kroesen et al. 1995), whereas a decreased expression has been observed in the PFC (Bauer et al. 2000). Synaptotagmin V mRNA was found to be reduced at 24 h after acute clozapine treatment. Synaptotagmin V, a member of synaptotagmins that are involved in membrane trafficking, has been suggested to participate in exocytosis of synaptic vesicles (Südhof and Rizo 1996). Recently, expression of synaptotagmin V has been reported as decreased in the PFC of schizophrenic subjects (Mirnics et al. 2000). Sec-1 (Munc-18) is a syntaxin binding protein that negatively regulates the formation of the synaptic SNARE membrane fusion complexes (Pevsner et al. 1994), and thereby inhibits synaptic vesicle exocytosis (Dresbach et al. 1998). Rab-12, which belongs to the family of small Ras-related GTPases that participate in the regulation of specificity of vesicular transport and SNARE-mediated membrane fusion (Zerial and McBride 2001), was increased by chronic haloperidol treatment, but localized in the node cluster B with reduced expression patterns after acute clozapine treatment. All the genes listed above are involved in the vesicular transport and presynaptic function. Therefore, although the regulation of these genes was variable depending on the drug and time of examination, these data, together with the results of the cluster analysis suggest that antipsychotic drugs may modify neurotransmitter vesicle release and presynaptic organization.
Recent microarray study revealed significant differences in the expression of a number of genes involved in presynaptic function in the prefrontal cortex of schizophrenic brain (Mirnics et al. 2000). Although synaptotagmin V and calcineurin A were the only genes found to be altered both in our study and in those examining gene expression in the PFC of schizophrenic brain (Mirnics et al. 2000; Hakak et al. 2001), the observation that the same functional classes are regulated in the same brain area both by the disease and by an experimental model of treatment is intriguing. It is possible that the chronic antipsychotic treatment that many of the patients used in postmortem expression studies have been exposed to may at least partially explain this coregulation, although the genes regulated in the brains of schizophrenic patients were not changed in the brain of a monkey chronically treated with haloperidol (Mirnics et al. 2000). A more provocative explanation for the observed coregulation is that antipsychotic drugs may be efficacious in schizophrenia because their effect on presynaptic function partially improves pathological changes in synaptic transmission in the prefrontal cortex of schizophrenic patients.
Ca2+-mediated intracellular signalling affects numerous elementary processes in neurones. Interestingly, several genes involved in the regulation of intracellular Ca2+ stores were found as regulated by antipsychotic drug treatments. In this study, we observed differential mRNA expression of three known visinin-like proteins by typical and atypical antipsychotic drug administration. Visinin-like proteins belong to the family of intracellular EF-hand calcium sensor proteins (Braunewell and Gundelfinger 1999), which regulate effector proteins, such as certain G-protein coupled receptor kinases, at the cell membrane in response to alterations in Ca2+ levels (Braunewell and Gundelfinger 1999; Iacovelli et al. 1999). A member of visinin-like protein family, VSNL-1, has been suggested as abnormally localized in the hippocampus of schizophrenic patients (Bernstein et al. 2002), and alterations in VSNL-1 expression have been observed in phencyclidine model of schizophrenia (Kajimoto et al. 1995). Furthermore, calcineurin A mRNA was regulated by clozapine treatment. Calcineurin A is a Ca2+ and calmodulin-dependent serine-threonine protein phosphatase (Rusnak and Mertz 2000). Clozapine treatment has been suggested to inhibit the activity of calcineurin (Gong et al. 1996). Interestingly, microarray analysis of gene expression in schizophrenic brain has identified an increased expression of calcineurin A in the PFC of neuroleptic-medicated patients (Hakak et al. 2001). Altogether, our results on altered mRNA expression of visinin-like proteins and calcineurin A suggest that intracellular levels of Ca2+ are regulated by antipsychotic drug treatment.
Among the other genes regulated by antipsychotics, SoS is interesting. SoS is a key regulator of the ras-MAP-kinase pathway, which suggests that this pathway which is activated by growth factors and certain G-protein coupled receptors (Cullen and Lockyer 2002), may also be a target of antipsychotic drug action. More detailed studies on the role of ras-MAP-kinase pathway in this context are warranted. Furthermore, microarray results suggested that c-jun may be regulated by acute clozapine (ratio of 2.25 at 1-h clozapine treatment) in the PFC, however, immediate early genes were not studied further by in situ hybridization, since their expression in response to antipsychotic drug treatments has been well documented by several investigators (Dragunow et al. 1990; Robertson and Fibiger 1992; Sebens et al. 1995; Deutch and Duman 1996; Merchant et al. 1996).
Verification of microarray results with in situ hybridization confirmed the regulation of only 10 out of the 35 genes investigated. Conversely, it is probable that there are a number of genes that were not observed as altered in microarray experiments, although they, in fact, are regulated by antipsychotic drugs. One such gene is apparently c-fos, which did not appear to be significantly regulated in response to acute clozapine administration in microarray experiments, although the literature consistently describes induced c-fos expression by clozapine in the PFC of rodent brain. These observations indicate that microarrays produce false positive and negative results at a high rate and underline the need for an independent confirmation of the results, by either biochemical or bioinformatic methods.
Examination of complex gene expression patterns induced by antipsychotic drugs with expression array technology may open avenues for more detailed understanding of the biological and neurochemical basis of antipsychotic drug treatment. Furthermore, expression profiling of genes affected by antipsychotic drug treatment in model systems will help to separate the effects of these drugs from the effects of the disease in medicated patients.
The authors would like to thank Marketa Marvanova, Markus Storvik, Anne-Mari Friis, Anne Lehtelä, and Laila Kaskela for technical assistance. Mikko Kolehmainen, Visipoint Oy is acknowledged for providing the SOM software and Jari-Pekka Ikonen, Comsol Oy, for assistance with statistical analysis.