Changes in the Peripheral Blood Gene Expression Profile Induced by 3 Months of Valproate Treatment in Patients with Newly Diagnosed Epilepsy

Valproic acid (VPA) is a widely used antiepileptic drug with a broad range of effects and broad clinical efficacy. As a well-known histone deacetylase (HDAC) inhibitor, VPA regulates epigenetic programming by altering the expression of many genes. The aim of study was to analyze differences in gene expression profiles before and after the start of VPA treatment in patients with newly diagnosed epilepsy. RNA sequencing was used to compare whole-genome gene expression patterns of peripheral blood from nine patients with epilepsy before and 3 months after the start of treatment with VPA. Of the 23,099 analyzed genes, only 11 showed statistically significant differential expression with false discovery rate-adjusted p-values below 0.1. Functional annotation and network analyses showed activation of only one genetic network (enrichment score = 30), which included genes for cardiovascular system development and function, cell morphology, and hematological system development and function. The finding of such a small number of differently expressed genes between before and after the start of treatment suggests a lack of HDAC inhibition in these patients, which could be explained by the relatively low doses of VPA that were used. In conclusion, VPA at standard therapeutic dosages modulates the expression of a small number of genes. Therefore, to minimize the potential side effects of HDAC inhibition, it is recommended that the lowest effective dose of VPA be used for treating epilepsy.


Introduction
Valproate [also known as valproic acid (VPA)] is one of the most frequently used antiepileptic drugs worldwide (1) and is widely applied for other conditions, such as migraine prophylaxis and bipolar disorder (2). Usually, VPA is used for treating chronic conditions requiring long-term treatment, which emphasizes the importance of the long-term safety of the drug. VPA has a broad range of Abbreviations: HDAC, histone deacetylase; VPA, valproic acid. effects and broad clinical efficacy, leading to its frequent designation as a "dirty drug" (3). VPA acts directly at the level of gene transcription by inhibiting histone deacetylation (HDAC) and increasing access to transcription sites (4). Recent data have shown that HDAC inhibitors and VPA regulate innate and adaptive immune pathways, with possible anti-inflammatory effects (5). The use of VPA during pregnancy has been associated with an approximately two to threefold increase in the rate of major anomalies, such as spina bifida, as well as cardiac, skeletal, craniofacial, and limb defects. VPA-induced inhibition of HDAC can result in cell-cycle disruption, growth arrest, and apoptosis, which may explain the teratogenic action of the drug (6).
In in vitro and in vivo preclinical studies, VPA showed strong antitumor effects in a wide variety of cancers by modulating multiple pathways, including cell-cycle arrest, angiogenesis, apoptosis, and differentiation (4). VPA inhibits the proliferation and induces the differentiation of malignancies, such as leukemia, lymphoma, teratocarcinoma, and medulloblastoma. Clinically, VPA has been used to treat leukemia and some solid tumors (7). About half of patients treated with VPA report weight gain, associated with important metabolic and endocrine abnormalities (8). Although the molecular mechanisms underlying VPA-induced weight gain are unknown, VPA was shown to suppress adiponectin gene expression in adipocytes (9). Accumulating evidence suggests that endoplasmic reticulum stress plays a role in the pathogenesis of type 2 diabetes, obesity, and insulin resistance (10). VPA modulates this stress response through the increased expression of members of the endoplasmic reticulum stress protein family, such as WFS1 (11), GRP94, and calreticulin (12).
As is evident, VPA is able to modulate the expressions of genes in different molecular pathways. This study was designed to evaluate the effect of 3 months of VPA treatment in patients who were recently diagnosed with epilepsy, by analyzing the differences in gene expression profiles before and after the start of treatment. Gene expression profiling in these patients could provide clues for understanding the molecular mechanisms of various side effects of VPA, such as metabolic changes, teratogenicity, and others.

Subjects
Nine otherwise healthy subjects (age range: 18-38 years) with newly diagnosed epilepsy and for whom VPA treatment was indicated were considered for enrollment in a case-crossover study on the effect of VPA on peripheral blood RNA expression. Patients were excluded if there was history or evidence of non-compliance with medical regimens; treatment with any medication other than VPA, including hormonal contraception; evidence of a cardiovascular, hepatic, renal, oncologic, or progressive neurological disease that could have an impact on blood RNA expression; evidence of progressive lesions on computed tomography or magnetic resonance imaging; or if the patient was pregnant.
All patients were clinically examined and interviewed by the first author. The main subject characteristics, including epileptic syndrome and VPA dose, are summarized in Table 1. Before and 3 months after the start of VPA treatment, the weight and height were measured, and the body mass index (BMI) was calculated as the weight (in kilogram) divided by the square of height (in meter). Seizures and epileptic syndromes were classified by using the criteria proposed by the ILAE (13). No patients had epileptic seizures during the evaluation period.

Sample Collection and RNA Preparation
Blood samples were collected during two visits, before and around 3 months after the start of treatment with VPA. The daily VPA dose was 600-1200 mg (mean dose: 11.4 ± 2.8 mg/kg). The study protocol and consent forms were approved by the Ethics Committee on Human Research at Tartu University. Blood samples were collected into Tempus tubes (Applied Biosystems, Foster City, USA). Blood was frozen and stored until further processing. RNA was extracted from whole blood with an RNA extraction kit, in accordance with the manufacturer's protocol (Applied Biosystems).

Statistical Analysis
Differential gene expression was analyzed by using the EdgeR package with non-normalized raw counts after the quality control of samples. EdgeR is a very flexible tool for RNA sequencing data analysis, which uses model-based scale normalization, dispersion estimates, negative binomial model fitting, and testing procedures to determine the differential expression of genes (14,15). As our sample contains paired samples (pre-and post-VPA), a paired testing approach was used. General linear modeling was applied, with subjects being added to the contrast matrix. The general linear modeling likelihood ratio test was applied to compare preand post-VPA results. Sequencing of the fragment library resulted  in an average of 15 million reads per sample. False discovery rate adjustment was used for multiple testing corrections (16). The threshold for statistical significance was an adjusted false discovery rate of 0.1.

Functional Analysis of Differentially Expressed Genes
Functional network analysis is used to identify the biological functions that are most significantly related to the molecules in a network. To define the functional networks of differentially expressed genes, the data were analyzed by using the Ingenuity Pathway Analysis (Ingenuity Systems, www.ingenuity.com), which calculates a significance (network) score for each network. This score indicates the likelihood that the assembly of a set of focus genes in a network could be explained by random chance alone (e.g., a score of 2 indicates that there is a 1 in 100 chance that the focus genes are together in a network due to random chance). A data set containing the Affymetrix probe-set identifiers and their corresponding fold change (log 2 ) values was uploaded into the Ingenuity Pathway Analysis software. Each gene identifier was mapped to its corresponding gene object in the Ingenuity Pathways Knowledge Base to identify molecules whose expression was significantly differentially regulated. These focus genes (or Network Eligible molecules) were overlaid onto a global molecular network developed from information contained in the Ingenuity Pathways Knowledge Base. Networks of these focus genes were then algorithmically generated on the basis of their connectivity. A network is a graphical representation of the molecular relationships between genes or gene products (represented as nodes). The biological relationship between two nodes is represented as an edge (line). All edges are supported by at least one reference from the literature, or from canonical information stored in the Ingenuity Pathways Knowledge Base.

Results
Comparison of blood RNA samples isolated from patients with epilepsy before and 3 months after the start of VPA treatment revealed distinct gene expression profiles. Altogether, 11 of the 23,099 analyzed genes showed statistically significant differential expression with multiple testing-adjusted false discovery rate values less than 0.1 ( Table 2). Relative differences in the expression signal (fold change or log fold change) between two time points were generally low. After initiation of VPA treatment, only two genes were up-regulated more than 1.1-fold, and no genes were down-regulated more than 1.1-fold, compared to the status before VPA treatment. Functional annotation of expression profiles was subsequently applied to identify functional changes in the context of genetic networks. A dataset containing 11 genes that had significant differential expression between before and after initiation of VPA treatment was uploaded into the Ingenuity Pathway Analysis software program. Only one genetic network was identified (enrichment score = 30), which included genes related to cardiovascular system development and function, cell morphology, as well as hematological system development and function (Table 3; Figure 1).

Discussion
In this study, we examined VPA-induced changes in gene expression in the peripheral blood of patients with newly diagnosed epilepsy who were treated with therapeutic doses of the drug.
More than a decade ago, VPA was identified as a potent promotor of histone acetylation. Histone acetylation leads to relaxation of the nucleosome structure and transcriptional activation (17). In theory, HDAC inhibition could directly or indirectly affect between 2 and 5% of all genes (18). In mouse embryonic stem cells exposed to VPA, the expression levels of 2.4% of genes were altered FIGURE 1 | Annotation enrichment analysis. A network including the gene functions "cardiovascular system development and function," "cell morphology," and "hematological system development and function" was significantly enriched in patients with epilepsy after initiation of treatment with valproic acid [score = -log(p-value) = 30]. Red nodes designate up-regulated genes. Numbers indicate the log 2 fold change (0 is equal expression). Uncolored nodes are genes in this network that were not in our list of differentially expressed genes. by more than 1.5-fold (19). This property of VPA is believed to have an impact on embryogenesis (20), causing increased rate of major congenital malformations among children exposed to this drug in utero (21). Only one study has compared the genomic expression patterns in peripheral blood between epilepsy patients treated with VPA or without any drug. Expression levels of 461 genes were altered in VPA-treated compared to drug-free epilepsy patients; however, no information was provided regarding the VPA doses that were used (22).
Although we did not measure VPA-induced HDAC inhibition, the finding that a surprisingly small number of genes were differently expressed before and after the start of VPA treatment suggests that HDAC inhibition in our patients was very low or even absent. In vitro, HDAC inhibition is achieved at concentrations of 0.3-1.0 mM, corresponding to the serum drug concentration in patients treated with a daily dose of 20-30 mg/kg (17). According to the guidelines of epilepsy treatment in adults (23), we began VPA treatment in our patients at a minimal effective dose, with the mean daily dose of 11.4 ± 2.8 mg/kg. The lack of the HDAC inhibition in our study is probably related to the relatively low daily VPA dose that was used. Indeed, the risk of major congenital malformations in patients treated with VPA increases with doses above 700 mg/d and is pronounced with doses above 1500 mg/d (21), which are clearly higher than the doses used in our study (600-1200 mg/d). On the other hand, higher VPA doses than those used here were applied for the treatment of tumors (20 mg/kg; 1500 mg/d) to modulate gene expression via HDAC inhibition (24,25). Therefore, to avoid the side effects of HDAC inhibition, it is recommended that patients be treated with the lowest effective dose of VPA.
During the evaluation period of about 3 months, no patient reported any epileptic seizures. However, the small sample size and short evaluation period prohibit us from making conclusions about whether seizure freedom was related to the VPA treatment or to the natural course of disease. Most patients had a presumed genetic etiology for their epilepsy, which usually suggests a selflimiting course. The anticonvulsive effect of VPA is mediated mainly through the reduced degradation, increased synthesis, decreased turnover, and, therefore, increased activity of gamma amino butyrate, and is not related to the alteration of gene expression (4). We could speculate that the anticonvulsive effect of VPA in our study was achieved without significant general activation of transcriptional activity.
After exposure to VPA, we observed changes in the expression levels of several genes (TPT1, ARAP3, and KLHDC8B) that are reported to be important in cancer-related pathways (26)(27)(28). Furthermore, the expression levels of two genes related to mitochondrial function, MOSC1 and ALAS2, were significantly changed. The product of MOSC1 localizes to the outer mitochondrial membrane and contributes to the regulation of nitric oxide synthesis (29). ALAS2 is an erythroid-specific, mitochondrially located enzyme that participates in heme biosynthetic pathways (30). Among the antiepileptic drugs, VPA has the highest potential to induce mitochondrial toxicity and could be even fatal in patients with some mitochondrial diseases. Use of VPA could be associated with the inhibition of respiratory chain complexes, impaired structural organization, and altered potential of the inner mitochondrial membrane (31,32). Further investigations are needed to determine whether the differentially expressed genes identified in our study play roles in VPA-related mitochondrial toxicity.
Our study has several limitations. We had a relatively small sample of patients. However, the paired design of the study allowed us to minimize the influence of external factors and to find significant changes with this sample size. Our study does not account for the possible occurrence of other events that could influence gene expression during the evaluation period. The process of epileptogenesis starts long before the first seizure and continues during the disease course. This process could affect gene expression in the peripheral blood. The relatively low VPA dose received by study participants leaves open the possibility that the expressions of many genes were not influenced. On the other hand, the results probably reflect the gene expression profile in real clinical practice, when relatively low doses of VPA are used. Future studies should explore whether the use of higher VPA blood concentrations (e.g., VPA doses >1500 mg/d) significantly changes the gene expression profile. In this case, due to ethical issues, a different study design should be used, probably comparing gene expression in independent samples, such as patients treated with high VPA doses compared to patients treated with other antiepileptic drugs or healthy subjects. Another important limitation of our study was the absence of relevant clinical problems in our patients, besides epilepsy. It was not possible for us to correlate changes in gene expression with the potential side effects of the drug.
In conclusion, the findings of this study suggest that, at standard dosages, VPA modulates the expression of a surprisingly small number of genes. Further studies should be done to evaluate the effect of higher VPA dosages on gene expression. On the other hand, to avoid potential side effects of HDAC inhibition, we recommend that patients with epilepsy who are receiving this drug be treated with the lowest effective dose of VPA.