Skip to main content
NCBI home page
International Journal of Methods in Psychiatric Research logoLink to International Journal of Methods in Psychiatric Research
. 2014 Feb 12;23(2):279–288. doi: 10.1002/mpr.1435

Mood‐stabilizers differentially affect housekeeping gene expression in human cells

Timothy R Powell 1,✉, Georgia Powell‐Smith 1, Kate Haddley 2, Peter Mcguffin 1, John Quinn 2, Leonard C Schalkwyk 1, Anne E Farmer 1, Ursula M D'Souza 1
PMCID: PMC6878232  PMID: 24677680

Abstract

Recent studies have revealed that antidepressants affect the expression of constitutively expressed “housekeeping genes” commonly used as normalizing reference genes in quantitative polymerase chain reaction (qPCR) experiments. There has yet to be an investigation however on the effects of mood‐stabilizers on housekeeping gene stability. The current study utilized lymphoblastoid cell lines (LCLs) derived from patients with mood disorders to investigate the effects of a range of doses of lithium (0, 1, 2 and 5 mM) and sodium valproate (0, 0.06, 0.03 and 0.6 mM) on the stability of 12 housekeeping genes. RNA was extracted from LCLs and qPCR was used to generate cycle threshold (C t) values which were input into RefFinder analyses. The study revealed drug‐specific effects on housekeeping gene stability. The most stable housekeeping genes in LCLs treated: acutely with sodium valproate were ACTB and RPL13A; acutely with lithium were GAPDH and ATP5B; chronically with lithium were ATP5B and CYC1. The stability of GAPDH and B2M were particularly affected by duration of lithium treatment. The study adds to a growing literature that the selection of appropriate housekeeping genes is important for the accurate normalization of target gene expression in experiments investigating the molecular effects of mood disorder pharmacotherapies. Copyright © 2014 John Wiley & Sons, Ltd.

Keywords: reference genes, lithium, sodium valproate, gene expression normalization

Introduction

Mood‐stabilizers such as lithium and sodium valproate are commonly prescribed pharmacotherapies used in the treatment of bipolar disorder. Despite being clinically prescribed, the mechanisms of action of each of these drugs are still relatively unknown (Shaldubina et al., 2001). It is largely accepted however that gene expression changes associated with these drug treatments likely play a role in their therapeutic effects (Sugawara et al., 2010). Numerous studies using patient tissue samples, such as brain and blood, have attempted to unearth the molecular pharmacology of these mood‐stabilizing drugs and how they confer their therapeutic effects (McQuillin et al., 2007; Chetcuti et al., 2006; Tsuang et al., 2005). The quantitative polymerase chain reaction (qPCR) is the standard method of investigating candidate gene expression changes and is used to validate large‐scale microarray expression hits (VanGuilder et al., 2008).

qPCR is a reaction in which a particular gene or region of DNA is amplified and detected in real‐time. The relative quantification method of normalization is often employed for qPCR data and requires the subtraction of relatively stable reference genes' expression from the potentially more dynamic expression of the target gene (Livak and Schmittgen, 2001). As the reference genes used for normalization are subject to the same conditions as the target gene itself, it helps to control for variables such as RNA integrity and reverse transcription efficiency, as well as controlling for differences in the amount of starting material.

Most studies have used so‐called “housekeeping” genes as the reference genes for normalization, as the products of these genes are essential for basal cell metabolism and so are assumed to be expressed at a constant and detectable level in all nucleated cell types during all developmental stages (Thellin et al., 1999). Many of the commonly used housekeeping genes such as beta‐actin (ACTB), glyceraldehyde 3‐phosphate dehydrogenase (GAPDH), 18S rRNA (18S) and beta‐2‐microglobulin (B2M) were first used in traditional qualitative or semi‐quantitative methods because of their high expression levels in all cells (Hendriks‐Balk et al., 2007).

Recent research however suggests that housekeeping genes are not as stable as was originally thought, with evidence demonstrating different levels of stability in different cell types (Vandesompele et al., 2002) and conditions (Hruz et al., 2011). Previous studies have reported that antidepressants cause changes to the expression of housekeeping genes (Sugden et al., 2010; Powell et al., 2012). For instance, Sugden et al. (2010) investigated the effects of antidepressants on housekeeping gene expression in a mouse fibroblast cell line. It was revealed that antidepressants affect housekeeping gene expression stability with some genes (ATP synthase and cytochrome c1) showing greater levels of stability than others [Eukaryotic translation initiation factor 4A2 (Eif4a2)].

Subsequently, the selection of appropriate housekeeping genes is important for the accurate normalization of gene expression experiments investigating the molecular effects of pharmacotherapies for major depressive disorder. However, there is no evidence as to whether mood‐stabilizers used to treat bipolar disorder also affect the expression of housekeeping genes. In this study we aimed to investigate the in vitro effects of the mood‐stabilizers lithium and sodium valproate on the expression of a panel of candidate housekeeping genes using mood disorder patient‐derived lymphoblastoid cell lines (LCLs) treated with these drugs. We aimed to establish: the most appropriate reference genes for LCLs treated with each drug, the most inappropriate reference genes for LCLs treated with each drug, and the effects of drug treatment duration on housekeeping gene stability.

Materials and methods

Participants

Samples in this study were collected as part of the Depression Case Control (DeCC) Study. The clinical methodology used in the DeCC collection was described in detail previously (see Gaysina et al., 2007). Briefly, subjects were identified from psychiatric clinics, hospitals, and general medical practices, and from volunteers responding to media advertisements. Only subjects of White European parentage were included. Subjects were over the age of 18 and had experienced two or more episodes of depression of at least moderate severity separated by at least two months of remission as defined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM‐IV) operational criteria, or the International Classification of Diseases, 10th Revision (ICD‐10) operational criteria. Participants were excluded if they had schizophrenia or bipolar disorder or their depression was caused by a physical illness, medication, alcohol or substance abuse, if they were intravenous drug users or were related to someone already recruited for the study. The study was approved by the Joint South London and Maudsley NHS Trust Institute of Psychiatry Research Ethics Committee and informed written consent was obtained from all the participants at the time of sample collection. Ten milliliters of whole blood from each individual was sent to the Human Genetic Cell Bank at the European Collection of Cell Cultures (ECACC), during which immortalized epstein‐barr virus (EBV)‐transformed LCLs were generated. The current study utilized LCLs from five females diagnosed with major depressive disorder, with a mean age of 53.6 ± 11.7 years, and who had not previously taken any mood‐stabilizing medications.

Cell culture

The LCL samples were supplied by ECACC as frozen ampoules containing 1 ml of cells at a density of approximately 2 × 107 cells/ml, in cell culture freezing medium [10% dimethyl sulphoxide (DMSO)]. Five cell lines were selected and grown in suspension in RPMI‐1640 medium (Sigma‐Aldrich, Poole, UK) supplemented with 10% foetal bovine serum, 2 mM l‐glutamine, 100 units/ml penicillin, 0.1 mg/ml streptomycin and 0.05 mg/ml neomycin at 37°C in a humidified atmosphere containing 5% carbon dioxide (CO2) (see Supplementary I nformation, S1).

Drug treatment of LCLs

The lithium chloride drug (Sigma‐Aldrich) dilution was prepared by diluting the supplied 8 M lithium chloride (LiCl) solution in RNase‐free water to produce a sterile 1 M LiCl stock solution, and then further diluting the stock solution to 1, 2 and 5 mM for the treatment procedure. Sodium valproate (Sigma‐Aldrich) was obtained in powder form and diluted in RNase‐free water to a concentration of 3 mM. This stock solution was then diluted accordingly to concentrations of 0.06, 0.3 and 0.6 mM for the treatment procedure. A vehicle control dilution was also prepared using RNase‐free water in serum‐free supplemented growth medium.

Each of the five cell lines underwent an acute drug treatment as part of a four‐stage culture protocol: (i) 72‐hour growth phase, (ii) 24‐hour serum‐starve phase, (iii) 24‐hour drug administration, and (iv) 24‐hour recovery phase (for details see Supplementary Information, S2). After the final 24‐hour recovery phase, cell pellets were obtained following centrifugation and were immediately stored at −80ºC for RNA extraction.

Each of the five cell lines were also treated with a chronic seven‐day dose of LiCl. This involved the same doses as with the acutely treated cells but instead utilized a two stage culture protocol: (i) 72‐hour growth phase and a (ii) seven‐day lithium administration phase (see Supplementary Information, S3 for details).

RNA extraction and cDNA synthesis

RNA extraction was performed using TRI reagent (Sigma‐Aldrich). The purity and quantity of RNA was subsequently measured using the Nanodrop ND1000 (Thermoscientific, Wilmington, DE), which showed all samples had 260/280 ratios of between 1.8 and 2.1. The RNA integrity numbers (RINs) of samples were assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Berkshire, UK) and all samples showed RINs > 9. The reverse transcription reaction was prepared using 1 µg of total RNA and the QuantiTect Reverse Transcription Kit (Qiagen, Crawley, UK). Complementary DNA (cDNA) synthesis was carried out in two steps: a genomic DNA (gDNA) wipe‐out step, followed by reverse transcription step, as according to the manufacturer's protocol. Briefly, following gDNA removal, the samples were incubated for 15 minutes at 42°C with 1 ul Quantiscript Reverse Transcriptase, 5 × Quantiscript Reverse Transcriptase buffer, and 1 ul Reverse Transcriptase primer mix (oligo‐dT and random primers). The reverse transcriptase enzyme was subsequently inactivated at 95°C for three minutes. The cDNA samples generated were stored at −20°C prior to use in the qPCR experiments.

Quantitative polymerase chain reaction (qPCR)

The qPCR experiments were performed in 384‐well plates using reagents from Primer Design (Southampton, UK). Reagents included Precision‐R MasterMix and the human geNorm Kit (PrimerDesign) which included 12 pre‐designed PerfectProbeTM fluorescent (FAM‐labelled) primer/probe sets for 12 human housekeeping genes. No cycle threshold (C t) values for the housekeeping gene Eif4a2 were detected in our samples. This was due to a technical problem with the primer/probe set that had been supplied for this gene. Therefore the gene Eif4a2 was excluded from the assays and from the final analyses. Details of the remaining 11 housekeeping genes are shown in Table 1 (see Supplementary Information, S4 for more detailed information on primers). The qPCR was set up according to the manufacturer's instructions, whereby each well on our 384‐well plate contained a 20 µl qPCR reaction mixture consisting of: 5 µl cDNA, 4 µl RNase‐free water, 1 µl of primer/probe mix and 10 µl of Precision‐R MasterMix. qPCR assays were performed in duplicate to generate two technical replicates, and a negative control sample was also included for each reference gene.

Table 1.

A list of the 11 housekeeping genes included in the human geNorm kit, accession number and protein function

Accession number Gene name Protein function
NM 001101 Actin, beta (ACTB) Cytoskeletal structural protein
NM 002046 Glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) Glycolytic pathway enzyme
NM 021009 Ubiquitin C (UBC) Protein modifier implicated in numerous functions
NM 004048 Beta‐2‐microglobulin (B2M) Beta chain of major histocompatibility complex class I molecules
NM 003406 Tyrosine 3‐monooxygenase/tryptophan 5‐monoxygenase activation protein, zeta polypeptide (YWHAZ) Adapter protein involved in mediating signal transduction
NM 012423 Ribosomal protein L13a (RPL13A) Component of the 60S ribosomal subunit
NM 022551 18S ribosomal RNA (18S rRNA, 18S) Component of the 40S ribosomal subunit
NM 001916 Cytochrome c‐1 (CYC1) Heme‐containing component of cytochrome b‐c1 complex of the mitochondrial respiratory chain
NM 004168 Succinate dehydrogenase complex, subunit A, flavoprotein (SDHA) Involved in complex II of the mitochondrial electron transport chain
NM 003286 Topoisomerase (DNA) I (TOP1) Enzyme that controls and alters the topology of DNA during transcription
NM 001686 ATP synthase (ATP5B) Subunit of mitochondrial ATP synthase, produces ATP from ADP using a protein gradient across the membrane
Open in a new tab

The qPCR reactions were performed using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, California, USA). Thermal cycling conditions consisted of an enzyme activation stage (95°C for 10 minutes), followed by 50 cycles of a denaturation stage (95°C for 15 seconds) and hybridization and data collection stage (50°C for 30 seconds), and a final extension stage (72°C for 15 seconds). The software program SDS 2.1 (Applied Biosystems) generated C t values from the data collected.

Statistical analysis

Average C t values from our technical replicates were input into The RefFinder statistical analysis web‐based tool (available from http://www.leonxie.com/referencegene.php). This tool was used for evaluating the stability of putative reference genes using an integrative weighted analysis which incorporates results from four already established analyses used for the selection of reference genes. These four already established statistical analyses for the selection of reference genes include: geNorm [as described in Vandesompele et al. (2002)], Normfinder [as described in Andersen et al. (2004)], Bestkeeper [as described in Pfaffl et al. (2004)] and the comparative ΔC t method [as described in Silver et al. (2006)]. Based on the rankings from each method, RefFinder assigns an appropriate weight to an individual gene and calculates the geometric mean of their weights for the overall final ranking.

geNorm analysis was also considered separately as it can additionally be used to calculate the optimal number of reference genes required for normalization purposes. It achieves this by calculating standard deviations (SDs) of log‐transformed expression ratios for all housekeeping genes and then carrying out a pairwise comparison of the SD of a particular gene with each of the remaining other housekeeping genes. One by one, the least stable reference genes are excluded based on high M values (which represent expression stability scores), leaving at least two remaining genes that correspond to the most stably expressed genes. According to Vandesompele et al. (2002), a combination of two or more reference genes producing a V (variation measure) of V < 0.15 is sufficient for optimal normalization.

Results

Number of reference genes required for accurate normalization

The geNorm approach was used to establish the number of reference genes for accurate normalization. The analysis revealed that two reference genes were optimal, as three reference genes would not increase accuracy above that obtained using two reference genes. The variation between the mean of the two most stable genes compared with that of the three most stable was V = 0.05 for acute lithium treated cells, V = 0.04 for chronic lithium treated cells, and V = 0.04 for acute sodium valproate treated cells; all of which are well below the threshold of V < 0.15 proposed by Vandesompele et al. (2002). The pairwise variation was relatively stable across all the comparisons, none of which exceeded the threshold value of 0.15 (see Supplementary Information, S5–S7). Nevertheless, the selection of the most stable housekeeping genes will likely increase the chances of detecting target gene expression differences of smaller magnitudes.

Acute treatment (24 hours) of LCLs with sodium valproate

The C t values of 11 housekeeping genes in LCLs treated with the different concentrations of sodium valproate are shown in Figure 1. The lowest generated mean C t value, and therefore the most highly expressed gene was 18S (16.87 ± 0.19 SD). The highest mean C t value and therefore the lowest expressing gene was SDHA (30.11 ± 0.67 SD). RefFinder analyses revealed that ACTB and RPL13A were the two most stable reference genes producing stability scores of 1.32 and 1.68, respectively. The two most unstable reference genes were SDHA and CYC1 producing stability scores of 9.24 and 10.24, respectively, see Figure 2. Results from each of the individual analyses contributing to RefFinder results (geNorm, Normfinder, Bestkeeper and the comparative ΔC t method) can be found in the Supplementary Information, S8.

Figure 1.

Figure 1

Open in a new tab

qPCR cycle threshold values in 11 reference genes in LCLs treated with 0, 0.06, 0.3 and 0.06 mM sodium valproate for 24 hours. Expression levels are shown as median (lines), 25th percentile to the 75th percentile (boxes), and ranges (whiskers). The mean C t values (white diamonds) and outliers (black triangles) are also indicated.

Figure 2.

Figure 2

Open in a new tab

Bar chart showing RefFinder expression stability values (y‐axis) for 11 reference genes (x‐axis) in LCLs treated with 0, 0.06, 0.3 or 0.6 mM sodium valproate for 24 hours.

Acute treatment (24 hours) of LCLs with lithium

The C t values of 11 housekeeping genes in LCLs treated with different concentrations of lithium for 24 hours are shown in Figure 3. The lowest mean C t value, and therefore most highly expressed gene was 18S (18.79 ± 1.74 SD), whilst the highest mean C t value, and lowest expressed gene was succinate dehydrogenase complex, subunit A, flavoprotein (SDHA) (27.56 ± 1.45 SD). RefFinder analyses revealed that GAPDH and ATP5B were the two most stable reference genes producing stability scores of 2 and 2.63, respectively. The two most unstable reference genes were SDHA and 18S producing stability scores of 9.49 and 11, respectively, see Figure 4. Results from each of the individual analyses contributing to RefFinder results (geNorm, Normfinder, Bestkeeper and the comparative ΔC t method) can be found in the Supplementary Information, S9.

Figure 3.

Figure 3

Open in a new tab

qPCR cycle threshold values in 11 reference genes in LCLs treated with 0, 1, 2 or 5 mM lithium chloride for 24 hours. Expression levels are shown as median (lines), 25th percentile to the 75th percentile (boxes), and ranges (whiskers). The mean C t values (white diamonds) and outliers (black triangles) are also indicated.

Figure 4.

Figure 4

Open in a new tab

Bar chart showing RefFinder expression stability values (y‐axis) for 11 reference genes (x‐axis) in LCLs treated with 0, 1, 2 or 5 mM lithium chloride for 24 hours.

Chronic treatment (seven days) of LCLs with lithium

The C t values of 11 housekeeping genes in LCLs treated with different concentrations of lithium for seven days are shown in Figure 5. The lowest mean C t value, and therefore most highly expressed gene was 18S (18.26 ± 0.60 SD), whilst the highest mean C t value, and lowest expressed gene was SDHA (29.28 ± 0.53 SD). RefFinder analyses revealed that ATP5B and CYC1 were the two most stable reference genes producing stability scores of 1.78 and 2.63, respectively. The two most unstable reference genes were SDHA and ACTB producing stability scores of 10 and 11, respectively, see Figure 6. Results from each of the individual analyses contributing to RefFinder results (geNorm, Normfinder, Bestkeeper and the comparative ΔCt method) can be found in the Supplementary Information, S10. A comparison of the stability values of the housekeeping genes in LCLs following acute and chronic lithium treatment are shown in Figure 7.

Figure 5.

Figure 5

Open in a new tab

qPCR cycle threshold values in 11 reference genes in LCLs treated with 0, 1, 2 or 5 mM lithium chloride for seven days. Expression levels are shown as median (lines), 25th percentile to the 75th percentile (boxes), and ranges (whiskers). The mean C t values (white diamonds) and outliers (black triangles) are also indicated.

Figure 6.

Figure 6

Open in a new tab

Bar chart showing RefFinder expression stability values (y‐axis) for 11 reference genes (x‐axis) in LCLs treated with 0, 1, 2 or 5 mM lithium chloride for seven days.

Figure 7.

Figure 7

Open in a new tab

Comparison of the 11 reference gene expression profiles for LCLs treated with 0, 1, 2 or 5 mM lithium chloride for 24 hours (grey line) or for seven days (black line). Gene names are shown on the x‐axis and RefFinder stability values are marked on the y‐axis.

Discussion

Recent studies investigating the effects of antidepressants in vitro and ex vivo revealed that housekeeping gene expression, which was once believed to be “stable” across conditions, in fact shows expression variability (Sugden et al., 2010; Powell et al., 2012). This in turn has consequences for studies investigating the gene expression effects of mood disorder pharmacotherapies using qPCR experiments and the relative quantification method of normalization. The current study aimed to investigate the effects of the mood‐stabilizers lithium and sodium valproate, which are used to treat bipolar disorder, on housekeeping gene expression. This study investigated the effects of acute drug treatments (sodium valproate, lithium) and a chronic drug treatment (lithium) at a variety of doses on the expression of a panel of 11 housekeeping genes. The drug dose ranges chosen incorporates doses that are believed to be therapeutic. This is based on serum concentrations of lithium found in bipolar disorder patients (approximately 1 mM) (Taylor et al., 2007), and on the dose required for sodium valproate to demonstrate neuroprotective effects (approximately 0.1 mM) (Biermann et al., 2011). The current study used an in vitro experimental design carried out using LCLs derived from mood disorder patients. LCLs have previously been shown to have gene expression profiles which correlate highly with levels of gene expression in non‐transformed lymphocytes and whole blood collected from psychiatric patients (Rollins et al., 2010). Consequently, these cell lines were specifically selected from mood disorder patients with the aim of creating the best proxy model for investigating the gene expression effects of mood‐stabilizers that may occur in vivo in mood disorder patients.

The study revealed that all housekeeping genes were expressed at acceptably detectable levels under drug treatment conditions (i.e. C t < 37) in LCLs producing a broad range of C t values (see Figures 1, 3 and 5). Across all drug groups, 18S was the most highly expressed housekeeping gene, whereas SDHA was the lowest expressing gene. Both mood‐stabilizers were found to affect the stability of housekeeping gene expression and they did so differentially in a drug‐specific manner. geNorm analyses revealed that two housekeeping genes were sufficient for optimal normalization of target genes in all drug treatment groups.

In the LCLs treated acutely with sodium valproate, the two most stable housekeeping genes according to the RefFinder analyses were ACTB and RPL13A (see Figure 2). ACTB, a gene encoding a cytoskeletal structural protein, has previously been used as a normalizing reference gene in mice treated with sodium valproate (Wu et al., 2010) and has been shown to be one of the most stable housekeeping genes in studies investigating the effects of valproate on forskolin‐stimulated human adrenal carcinoma (H295R) cells (von Krogh et al., 2010). RPL13A, a component of the 60S ribosomal subunit, has been shown to be one of the most stable housekeeping genes in a rat model of cerebral ischemia (Tian et al., 2006), but has not been previously used for normalization purposes in cells treated with sodium valproate. The housekeeping genes SDHA and CYC1 showed the highest expression variability according to RefFinder analyses, and thus would be considered unsuitable as reference genes in LCLs exposed acutely to sodium valproate (see Figure 2).

In the LCLs acutely treated with lithium, the two most stable housekeeping genes according to RefFinder were GAPDH and ATP5B (see Figure 4). GAPDH, a gene encoding a glycolytic pathway enzyme has previously been shown to demonstrate stable gene expression in human epithelial cells treated with lithium (Nemeth et al., 2002) and has also previously been used for normalization in human placental cells treated with the drug (Roberts et al., 2007). ATB5B, a gene which encodes a protein involved in catalysing ATP formation has previously been shown to be one of the most stable genes in cells treated with antidepressants (Sugden et al., 2010). ATB5B, was further shown to be the most stable reference gene in LCLs treated with lithium chronically along with CYC1 (see Figure 6). The greatest variation in expression and therefore worst reference genes in LCLs treated acutely with lithium were observed in the genes SDHA and 18S (see Figure 4 and Table 1), and in SDHA and ACTB in LCLs treated chronically with lithium.

Previous research on the effects of antidepressants revealed that duration of drug treatment had little effect on housekeeping gene expression stability (Sugden et al., 2010). Here, we tested whether duration of treatment with a mood‐stabilizer might affect housekeeping gene expression. We observed the effects of a chronic (seven day) administration of lithium on housekeeping gene expression in LCLs, and compared it to an acute 24 hour treatment of lithium. The stability of genes in LCLs treated chronically showed a similar trend to the relative stability of genes in LCLs treated acutely. However, there were two genes which were key exceptions. In the chronically treated cells, GAPDH became less stable and B2M became more stable (see Figure 7). Subsequently, GAPDH may be the most stable reference gene in LCLs treated acutely with lithium, but it is less suitable for LCLs treated chronically with the drug. If comparisons were to be drawn across different drug duration groups, ATP5B and UBC would in fact be considered two of the most stable reference genes (see Figure 7). These results consequently suggest that careful consideration is needed for the selection of reference genes not only for mood‐stabilizer type but also for treatment duration.

To conclude, this is the first study to investigate the differential effects of mood‐stabilizers on housekeeping gene expression in human cells from mood disorder patients. The study revealed that lithium and sodium valproate caused drug‐specific effects on housekeeping gene expression stability, with lithium treatment duration also having some influences on housekeeping gene stability, particularly in the genes GAPDH and B2M. ACTB and RPL13A were the two most stable genes in LCLs treated acutely with sodium valproate, whereas GAPDH and ATP5B were the two most stable genes in LCLs treated acutely with lithium, and ATB5B and CYC1 were the two most stable genes in LCLs treated chronically with lithium. SDHA was amongst the most variably expressed housekeeping genes across all drug groups and as such may be considered an unsuitable reference gene in in vitro experiments investigating the effects of mood‐stabilizers. The use of mood‐disorder patient derived LCLs in this study arguably means results from this in vitro experiment might not only extrapolate to other in vitro studies investigating the effects of mood‐stabilizers but also to ex vivo studies investigating gene expression in lymphocytes extracted from mood disorder patients; although this would need to be confirmed in future studies. The study provides further evidence to a growing literature that the selection of appropriate housekeeping genes is important for the accurate normalization of target gene expression in experiments investigating the molecular effects of mood disorder pharmacotherapies.

Declaration of interest statement

Powell, Powell‐Smith, Haddley, Quinn, Schalkwyk, and D'Souza report no competing interests. McGuffin and Farmer have received consultancy fees and honoraria for participating in expert panels from pharmaceutical companies, including Lundbeck and GlaxoSmithKline.

Supporting information

Supporting info item

Click here for additional data file. (144.2KB, docx)

Acknowledgements

Timothy R Powell and Georgia Powell‐Smith were funded by Medical Research Council PhD studentships. The authors also acknowledge Primer Design Ltd for their generous sponsorship towards this gene expression component of this study, and in particular Dr Jim Wicks who provided technical advice.

References

  1. Andersen C.L., Jensen J.L., Ørntoft T.F. (2004) Normalization of real‐time quantitative reverse transcription‐PCR data: a model‐based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Research, 64(15), 5245–5250. [DOI] [PubMed] [Google Scholar]
  2. Biermann J., Boyle J., Pielen A., Lagrèze W.A. (2011) Histone deacetylase inhibitors sodium butyrate and valproic acid delay spontaneous cell death in purified rat retinal ganglion cells. Molecular Vision, 17L, 395–403. [PMC free article] [PubMed] [Google Scholar]
  3. Chetcuti A., Adams L.J., Mitchell P.B., Schofield P.R. (2006) Altered gene expression in mice treated with the mood stabilizer sodium valproate. International Journal of Neuropsychopharmacology, 9(3), 267–276. [DOI] [PubMed] [Google Scholar]
  4. Gaysina D., Cohen S., Craddock N., Farmer A., Hoda F., Korszun A., Owen M.J., Craig I.W., McGuffin P. (2007) No association with the 5,10‐methylenetetrahydrofolate reductase gene and major depressive disorder: results of the Depression Case Control (DeCC) study and a meta‐analysis. American Journal of Medical Genetics Part B, 147B(6), 699–706. [DOI] [PubMed] [Google Scholar]
  5. Hendriks‐Balk M.C., Michel M.C., Alewijnse A.E. (2007) Pitfalls in the normalization of real‐time polymerase chain reaction data. Basic Research in Cardiology, 102(3), 195–197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Hruz T., Markus W., Docquier M., Pfaffl M.W., Masanetz S., Borghil L., Verbrugghe P., Kalaydjieva L., Bleuler S., Laule O., Descombes P., Gruissem W., Zimmermann P. (2011) RefGenes: identification of reliable and condition specific reference genes for RT‐qPCR data normalization. BMC Genomics, 12, 156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. von Krogh K., Harjen H., Almås C., Zimmer K.E., Dahl E., Olsaker I., Taubøll E., Ropstad E., Verhaegen S. (2010) The effect of valproate and levetiracetam on steroidogenesis in forskolin‐stimulated H295R cells. Epilepsia, 51(11), 2280–2288. [DOI] [PubMed] [Google Scholar]
  8. Livak K.J., Schmittgen T.D. (2001) Analysis of relative gene expression data using real‐time quantitative PCR and the 2(‐delta delta C(T)) method. Methods, 25(4), 402–408. [DOI] [PubMed] [Google Scholar]
  9. McQuillin A., Rizig M., Gurling H.M. (2007) A microarray gene expression study of the molecular pharmacology of lithium carbonate on mouse brain mRNA to understand the neurobiology of mood stabilization and treatment of bipolar affective disorder. Pharmacogenetics and Genomics, 17(8), 605–617. [DOI] [PubMed] [Google Scholar]
  10. Nemeth Z.H., Deitch E.A., Szabo C., Fekete Z., Hauser C.J., Hasko, G. (2002) Lithium induces NF‐κB activation and interleukin‐8 production in human intestinal epithelial cells. Journal of Biological Chemistry, 277, 7713–7719. [DOI] [PubMed] [Google Scholar]
  11. Pfaffl M.W., Tichopad A., Prgomet C., Neuvians T.P. (2004) Determination of stable housekeeping genes, differentially regulated target genes and sample integrity: BestKeeper – Excel‐based tool using pair‐wise correlations. Biotechnology Letters, 26, 509–515. [DOI] [PubMed] [Google Scholar]
  12. Powell T.R., Schalkwyk L.C., Heffernan A.L., Breen G., Lawrence T., Price T., Farmer A.E., Aitchison K.J., Craig I.W., Danese A., Lewis C., McGuffin P., Uher R., Tansey K.E., D'Souza U.M. (2012) Tumor necrosis factor and its targets in the inflammatory cytokine pathway are identified as putative transcriptomic biomarkers for escitalopram response. European Neuropsychopharmacology, 23(9), 1105–1114. [DOI] [PubMed] [Google Scholar]
  13. Roberts J., Scott A.C., Howard M.R., Breen G., Bubb V.J., Klenova E., Quinn J.P. (2007) Differential regulation of the serotonin transporter gene by lithium is mediated by transcription factors, CCCTC binding protein and Y‐box binding protein 1, through polymorphic intron 2 variable number tandem repeat. Journal of Neuroscience, 27(11), 2793–2801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Rollins B., Martin M.V., Morgan L., Vawter M.P. (2010) Analysis of whole genome biomarker expression in blood and brain. American Journal of Medical Genetics Part B, 153B(4), 919–936. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Shaldubina A., Agam G., Belmaker R.H. (2001) The mechanism of lithium action: state of the art, ten years later. Progress in Neuro‐Psychopharmacology and Biological Psychiatry, 25(4), 855–866. [DOI] [PubMed] [Google Scholar]
  16. Silver N., Best S., Jiang J., Thein S.L. (2006) Selection of housekeeping genes for gene expression studies in human reticulocytes using real‐time PCR. BMC Molecular Biology, 7, 33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Sugawara H., Iwamoto K., Bundo M., Ishiwata M., Ueda J., Kakiuchi C., Ishigooka J., Kato T. (2010) Effect of mood stabilizers on gene expression in lymphoblastoid cells. Journal of Neural Transmission, 117(2), 155–164. [DOI] [PubMed] [Google Scholar]
  18. Sugden K., Pariante C.M., McGuffin P., Aitchison K.J., D'Souza U.M. (2010) Housekeeping gene expression is affected by antidepressant treatment in a mouse fibroblast cell line. Journal of Psychopharmacology, 24(8), 1253–1259. [DOI] [PubMed] [Google Scholar]
  19. Taylor D., Kerwin R., Paton C. (2007) The Maudsley Prescribing Guidelines, 9th edition London: Informa Healthcare. [Google Scholar]
  20. Thellin O., Zorzi W., Lakaye B., De Borman B., Coumans B., Hennen G., Grisar T., Igout A., Heinen E. (1999) Housekeeping genes as internal standards: use and limits. Journal of Biotechnology, 75(2‐3), 291–295. [DOI] [PubMed] [Google Scholar]
  21. Tian Y.F., Zhang P.B., Xiao X.L., Zhang J.S., Zhao J.J., Kang Q.Y., Chen X.L., Qiu F., Liu Y. (2006) The quantification of ADAMTS expression in an animal model of cerebral ischemia using real‐time PCR. Acta Anaesthesiologica Scandinavica 51(2), 158–164. [DOI] [PubMed] [Google Scholar]
  22. Tsuang M.T., Nossova N., Yager T., Tsuang M.M., Guo S.C., Shyu K.G., Glatt S.J., Liew C.C. (2005) Assessing the validity of blood‐based gene expression profiles for the classification of schizophrenia and bipolar disorder: a preliminary report. American Journal of Medical Genetics Part B, 133B(1), 1–5. [DOI] [PubMed] [Google Scholar]
  23. Vandesompele J., Preter K.D., Pattyn F., Poppe B., Van Roy N., De Paepe A., Speleman F. (2002) Accurate normalization of real‐time quantitative RT‐PCR data by geometric averaging of multiple internal control genes. Genome Biology, 3, RESEARCH0034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. VanGuilder H.D., Vrana K.E., Freeman W.M. (2008) Twenty‐five years of quantitative PCR for gene expression analysis. BioTechniques, 44(5), 619–626. [DOI] [PubMed] [Google Scholar]
  25. Wu G., Nan C., Rollo J.C., Huang Z., Tien J. (2010) Sodium valproate‐induced congenital cardiac abnormalities in mice are associated with the inhibition of histone deacetylase. Journal of Biomedical Science, 17, 16. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting info item

Click here for additional data file. (144.2KB, docx)

Articles from International Journal of Methods in Psychiatric Research are provided here courtesy of Wiley

RESOURCES