Significant outcomes
• The Leu7Pro variation of NPY was associated with depression, but not schizophrenia, in a large Danish sample.
• The ‘Pro7 allele' was associated with reduced levels of NPY in cell cultures.
• The study is consistent with a role for NPY in the aetiology of depression.
Limitations
• Swedish and Dutch studies reported lower frequency of the “Pro7 allele” in depressed patients opposite to the finding in the present Danish sample. However, the former studies used different sampling procedures and control populations.
Introduction
Depression is a complex, severe psychiatric disorder with a lifetime prevalence of at least 10% with clear evidence for a genetic component; thus twin studies indicate heritability of 40–50% (Reference Kessing1,Reference Levinson2). Both clinical and animal studies suggest that neuropeptide Y (NPY) plays an important role in depression (Reference Redrobe, Dumont and Quirion3). For instance, the synthesis of NPY is decreased in the hippocampus in rodent models of depression (Reference Husum, Vasquez and Mathe4–Reference Jimenez-Vasquez, Mathé, Thomas, Riley and Ehlers6). Conversely, treatment with antidepressant drugs (Reference Caberlotto, Fuxe, Overstreet, Gerrard and Hurd7–Reference Mathé, Jousisto-Hanson, Stenfors and Theodorsson9) and electroconvulsive stimulation increase hippocampal NPY and NPY receptor expression (Reference Stenfors, Theodorsson and Mathé10–Reference Mikkelsen and Woldbye12) as well as release of NPY (Reference Christensen, Olesen, Kristiansen, Mikkelsen and Woldbye11). In addition, intracerebral administration of NPY has antidepressant-like activity in rodents (Reference Husum, Mikkelsen, Hogg, Mathé and Mørk13,Reference Redrobe, Dumont, Fournier and Quirion14) and depressed patients have decreased concentrations of NPY in the cerebrospinal fluid (CSF) (Reference Heilig, Zachrisson and Thorsell15,Reference Widerlöv, Lindstrom, Wahlestedt and Ekman16), suggesting a possible link between shortage of NPY and predisposition to depression. NPY also causes anxiolytic-like effects in rodents (Reference Bannon, Seda and Carmouche17,Reference Sørensen, Lindberg, Wörtwein, Bolwig and Woldbye18). Consistent with this concept, certain haplotypes of the NPY gene have been implicated in the regulation of emotions and stress responses in humans (Reference Zhou, Zhu and Hariri19).
The most thoroughly investigated polymorphism (rs16139) of several in the NPY gene (Reference Karvonen, Pesonen and Koulu20–Reference Mottagui-Tabar, Prince, Wahlestedt, Zhu, Goldman and Heilig23) is located at position 20 T>C from the transcription start (A in AUG; in the literature sometimes referred to as 1128 T>C) on chromosome 7p15.1. The 20 T>C polymorphism causes a substitution of amino acid residue number 7 leucine with proline in the signal peptide of NPY (Leu7Pro variation) (Reference Karvonen, Pesonen and Koulu20). The Leu7Pro variation has been associated with high serum cholesterol and low-density lipoprotein (LDL) cholesterol levels (Reference Karvonen, Pesonen and Koulu20), carotid atherosclerosis (Reference Niskanen, Karvonen and Valve24), retinopathy in type 2 diabetes (Reference Niskanen, Voutilainen-Kaunisto and Terasvirta25), high alcohol consumption and alcoholism (Reference Lappalainen, Kranzler and Malison26,Reference Kauhanen, Karvonen and Pesonen27). In one study with alcoholism, association with depression was examined, but no association was found (Reference Lappalainen, Kranzler and Malison26). Later, a small study with only 51 patients suggested that the ‘Pro7 allele' protects against depression (Reference Heilig, Zachrisson and Thorsell15), and this was also suggested in two recent papers (Reference Sjöholm, Melas, Forsell and Lavebratt28,Reference Bosker, Hartman and Nolte29).
The present study aimed at replicating the association between the Leu7Pro variation with major depression in a large sample of Danish patients. To determine whether effects were specific for depression, we also studied the occurrence of Leu7Pro in patients diagnosed with schizophrenia (psychiatric disease controls). Moreover, as the functional consequences of carrying the ‘Pro7 allele' are not clear, we also explored the effect on levels of NPY peptide and mRNA in vitro in two different cell lines.
Materials and methods
Subjects
Control samples. Two groups of controls were examined: screened controls and population controls. Screened controls (n = 287) consisted of healthy subjects of whom 51 were employees or students from Rigshospitalet University Hospital, and the rest were control samples collected in earlier studies (Reference Kessing and Jørgensen30–Reference Haugboel, Pinborg and Arfan32). These were all questioned about the occurrence of personal or family history (including grandparent) of depression, anxiety, or other psychiatric disorders, and only persons without such history were included. The mean age of the screened controls was 48 years (range = 19–84 years, SD = 19) for males (n = 109) and 48 years (range = 19–85 years, SD = 18) for females (n = 178). Population controls (n = 2625) consisted of 374 physically healthy blood donors collected in the Copenhagen area in 2005; a random sample of 451 unrelated individuals from the adult Danish population (age unknown) collected during the late 1980s in the Copenhagen area; and a large group of 1804 individuals randomly selected from a random sample of individuals aged 30–60 years living in the Copenhagen area participating in a non-pharmacological intervention study for prevention of ischaemic heart disease (Reference Jørgensen, Borch-Johnsen and Thomsen33). All population controls were anonymous, thus psychiatric clinical state was unknown. The mean age of the population controls was 48 years (range = 30–66 years, SD = 9) for males (n = 1254) and 47 years (range = 19–65 years, SD = 9) for females (n = 1371).
Patients with depression. Blood was sampled from patients diagnosed with major depression independently by two trained psychiatrists (n = 593; ICD-10 diagnoses F32-33) that were admitted to psychiatric departments at hospitals in the Copenhagen area. The sample consisted of patients collected within the period 2002–2008 (n = 389; The Danish Psychiatric Biobank) and 1993–1998 (n = 204; described elsewhere (Reference Kessing and Jørgensen30,Reference Mellerup, Bennike and Bolwig31,Reference Lindberg, Koefoed and Hansen34)). The mean age of the depressed patients was 50 years (range = 19–89 years, SD = 15) for males (n = 210) and 48 years (range = 18–93 years, SD = 19) for females (n = 383), respectively. Additional information on comorbid anxiety disorder (F40 and F41 in the ICD-10 diagnostic criteria) was present for a subset of patients (n = 253; 95 males and 158 females).
Psychiatric disease controls. Blood samples from patients with schizophrenia were also analysed (n = 503; ICD-10 diagnosis F20). They were sampled in the Copenhagen area within the period 2002–2007 (The Danish Psychiatric Biobank; n = 302) and in the Aarhus area in 2002–2005 (Opus Project; n = 201). Of these, 292 were men (mean age = 37, range = 19–77, SD = 11) and 211 were women (mean age = 36, range = 18–84, SD = 12). The patients were described in more details elsewhere (Reference Koefoed, Hansen and Woldbye35).
The subjects of the present study were Danish Caucasians. Data on the origin of the depressed patients were available in 277 cases. The majority of these (90.6%) had both parents born in Denmark. A minor fragment had one grandparent (7.6%) born outside Denmark (mainly Germany or Sweden) while 1.8% had no information on birthplace of grandparents. More details regarding origin were also available for blood donor population controls, with a majority (94%) having both parents born in Denmark (Reference Hansen, Olsen and Lindow36). A minor fraction had one parent (1.5%) or both parents (0.5%) born outside Denmark in other European countries while there was no information in 4% with regard to place of birth of both parents.
The Scientific-Ethical Committees in Denmark as well as the Danish Data Protection Agency approved the present study, and all participants gave informed consent. This study was conducted in accordance with the Declaration of Helsinki.
Genotyping (rs16139)
Genomic DNA was extracted from peripheral blood lymphocytes using QIAamp® DNA Blood Mini kit or FlexiGene DNA kit (Qiagen, Albertslund, Denmark) according to the manufacturer's instructions. For 528 of the samples, the NPY exon 2 was analysed at the denaturing high-performance liquid chromatography (D-HPLC) wave system (Transgenomic, Omaha, NE, USA). Fragments showing mutations were analysed on an ABI PRISM DNA Sequencer (ABI310 or ABI3730) (Applied Biosystems, Foster City, CA, USA). The rest of the samples were analysed by real-time PCR using TaqMan Assay-by-design genotyping assay (Applied Biosystems). Each 96-well plate contained two control samples (one heterozygote (CT) and one homozygote (CC)) as well as one negative control (no template). The primer and probe sequences, and temperatures used at the D-HPLC for each fragment are available on request. The success rate was 99.7%; only subjects with successful samples were included in the study. The genotype for 3.2% of the samples was reanalysed resulting in 100% concordance.
Plasmids, cell cultures and transient DNA transfections
The pGEM-1-NPY plasmid, containing the human NPY gene, including signal peptide, NPY and C-terminal peptide of NPY (CPON), was a kind gift from Professor Thue W. Schwartz, University of Copenhagen. The pGEM-1-NPY plasmid was digested with EcoRI and BamHI and the NPY coding region was cloned into pcDNA3.1 (Invitrogen, Carlsbad, CA, USA), generating pcDNA3.1-NPY-Leu7. Mutagenesis was performed using the QuickChange mutagenesis kit (Stratagene, La Jolla, CA, USA) according to the instructions from the manufacturer, generating pcDNA3.1-NPY-Pro7. Both constructs (the CMV promoter, NPY cDNA and the poly-A sequence) were verified by sequencing.
The Syrian hamster beta-cell line HIT-T15 (LGC Standards AB, Boras, Sweden) and human neuroblastoma SK-N-MC cells (ATCC, Manassas, VA, USA) were cultured at 10% CO2 and 37 °C in Dulbecco's modified Eagle's medium (with Glutamax, 5 mM glucose) supplemented with 10% fetal calf serum, 100 units/ml penicillin and 100 μg/ml streptomycin (Invitrogen). One day before transfection, 8 × 105 cells were seeded in 6-well culture dishes. Cells were transfected with 8 μg pcDNA3.1, pcDNA3.1-NPY-Leu7 or pcDNA3.1-NPY-Pro7 using FuGENE 6 (Roche A/S, Hvidovre, Denmark) according to the manufacturer's instructions. Forty-eight hours later, 5 ml culture medium was recovered, and the cells were harvested in PBS/EDTA, pelleted and resuspended in 1-ml sterile H2O. Transfection series were performed in triplicates and carried out at least three times.
To ensure that transfection efficiency did not vary significantly between cells transfected with pcDNA3. 1-NPY-Leu7 and pcDNA3.1-NPY-Pro7, quantitative PCR was performed as follows: transfected HIT-T15 and SK-N-MC cells were harvested as above in PBS/EDTA, and washed once in PBS. Nucleospin Tissue kit (Macherey-Nagel, Düren, Germany) was subsequently used for DNA purification according to the manufacturer's instructions. The amount of purified plasmid DNA encoding NPY was measured in 96-well plates on an Icycler (Bio-Rad, Grenaa, Denmark). Each well contained a mixture of plasmid DNA, SYBR Green I and 750 nM of forward and reverse primers. The plasmid DNA was quantified by two sets of primers recognising the CMV promoter and ampicillin resistance gene, both located on the transfection plasmid. Relative quantification was achieved by comparing each sample to in-plate reference gene controls (glyceraldehyde-3-phosphate dehydrogenase, GAPDH). Primer sequences are available on request.
NPY radioimmunoassay
HIT-T15 and SK-N-MC cells were transfected as described above. The subsequent procedure for radioimmunoassay was previously described (Reference Rehfeld37). Cells and culture media were boiled for 20 min and cell debris was removed by centrifugation. Extracts were analysed by radioimmunoassay using anti-NPY antibody (#6730-0204, Biogenesis, Kidlington, UK) and 125I-labelled NPY (#IM170, GE Healthcare, Hillerød, DK). The NPY antibody recognised primarily the mid-portion of the molecule, with little or no cross-reactivity for N-terminal NPY1-9 or C-terminal NPY 31-36 (Biogenesis).
NPY mRNA determination
Transfected HIT-T15 and SK-N-MC cells (two individual transfection series run in triplicates) were harvested as above in PBS/EDTA, and washed once in PBS. Total RNA was purified using RNeasy spin columns and on-column DNAse treated (Qiagen), to remove plasmid DNA encoding NPY, which would otherwise contribute to the final measurement of NPY mRNA levels. Total RNA (1 μg) was reverse transcribed using TaqMan RT master mix (Applied Biosystems) and random hexamers in a 100-μl reaction. Relative quantification of the cDNA was performed in 96-well plates on the iCycler (Bio-Rad). Each well contained 2 μl of cDNA mixed with 10 μl of SYBR Green I master mix (Bio-Rad), 0.2 μl of uracil-N-glycosylase (Epicentre Biotechnologies, Madison, WI, USA) and 750 nM forward and reverse primers (MWG Biotech, Ebersberg, Germany). All samples were run in duplicates. Collected data were analysed using the iCycler software, and the cycle threshold (Ct; i.e., the number of cycles necessary for the studied gene to be linearly expanded) for each sample was determined. Relative quantification was achieved by subtracting each Ct sample with the in-plate Ct of a reference control gene (delta Ct). The smaller the delta Ct value the more the gene is expressed compared with in-plate tested reference gene.
A group of reference genes, that is, GAPDH, beta-actin, TATA-box binding protein (TBP) and HPRT, was tested. GAPDH and TBP were found to be the least regulated genes and were consequently used for subsequent quantification of mRNA levels. Water controls and genomic DNA controls were run simultaneously with the samples on all plates and did not differ significantly from background. Primer sequences are available on request.
Data analysis
Genotype and allele frequencies of the 20 T > C polymorphism were analysed using Fisher's exact test (http://www.graphpad.com/quickcalcs/contingency1.cfm) or chi-squared test. OR and 95% confidence interval (CI 95%) were calculated using the website www.hutchon.net/ConfidOR.htm. Logistic regression was used to include age and gender as covariates in the analysis. Other data were analysed using Mann-Whitney U test, when relevant, following significant Kruskal-Wallis analysis of variance by ranks. NPY peptide data and delta Ct values of quantitative PCR are presented as mean ± standard error of the mean (SEM). The level of significance was p < 0.05.
Results
Genotyping
In contrast to previous studies, the Leu7Pro variation was found to be significantly more frequent in patients with major depression as compared to the combined control group (p < 0.0004; Table 1). This effect did not depend on gender. Thus increased frequencies of the Leu7Pro variation were detected in both males (p = 0.02) and females (p = 0.002) (data not shown). Similarly, a significantly higher ‘Pro7 allele' frequency (i.e., MAF, minor allele frequency) was seen in depressed patients as compared to the combined control groups (5.5% vs. 3.8%; p = 0.009; OR = 1.46, 95% CI = 1.10–1.93). This was also seen when we compared the depressed sample with the screened controls (p = 0.01; OR = 2.02, 95% CI = 1.16–3.53) and population controls (p = 0.02; OR = 1.41, 95% CI = 1.06–1.88), respectively. Comparing the combined cases and controls including schizophrenia disease controls also resulted in overall significance (p = 0.033). No significant differences in the frequency of the Leu7Pro variation were found between patients with schizophrenia and control individuals (Table 1). Gender or age did not vary between genotype groups (p > 0.95 and p > 0.22, respectively; data not shown).
Distribution of the Leu7Pro variation (T/C) (rs16139)
OR, odds ratio; CI, confidence interval; ND, not determined.
* Fisher's exact test between cases and controls, significant results are presented in bold.
† MAF: minor allele frequency (C-allele; Pro7 (P7) variant).
‡ Chi-square test between cases and controls, significant results are presented in bold.
Consistent with the association of the Leu7Pro variation with major depression, the genotype frequencies of depressed male or female patients were not in Hardy-Weinberg equilibrium, also when including both sexes (p < 0.005, X 2 = 11.53, df = 2). The genotypic distributions for the control group and psychiatric disease controls were all in Hardy-Weinberg equilibrium.
When cases (n = 253) were divided into groups according to comorbid anxiety (F40 or F41 in the ICD-10 system), no differences were found in the distribution of the Leu7Pro variation between patients with or without comorbid anxiety (MAF 6.4 and 6.3%, respectively; p = 0.98). Significantly more women (56.3%) showed comorbid anxiety as compared to men (30.5%; p < 0.001).
Transfections and measurements of NPY and its mRNA
To study the potential functional implications of Leu7Pro, HIT-T15 and SK-N-MC cells were transfected with plasmids encoding either the NPY ‘Leu7 allele' or the mutant ‘Pro7 allele'. HIT-T15 cells transfected with the ‘Leu7 allele' synthesised and released substantial amounts of NPY peptide as compared to cells transfected with empty cDNA (Fig. 1a). In contrast, NPY peptide concentrations were significantly lower in cell extracts (68% reduction; p = 0.0007) and cell media extracts (48% reduction; p = 0.003) from HIT-T15 cells transfected with cDNA carrying the Pro7 substitution as compared to HIT-T15 cells carrying the Leu7 plasmid. In the SK-N-MC cell line, NPY peptide concentrations appeared substantially lower than in HIT-T15 cells in cell extracts where the NPY concentration was not consistently above the assay detection limit (data not shown). However, as in HIT-T15 cells, the NPY concentrations were found to be significantly lower (34% reduction; p = 0.019) in cell media extracts from SK-N-MC cells transfected with Pro7 mutant as compared to the Leu7 cDNA (Fig. 1c). As shown in Fig. 1a and c both cell types had a low level of endogenous NPY secretion.
Levels of NPY peptide and mRNA in vitro in cells transfected with the Leu7 NPY gene, the Pro7 mutant and cDNA empty of the NPY gene. NPY peptide concentrations as measured with radioimmunoassay (RIA) in cell and/or media extracts from (a) 800 000 HIT-T15 and (c) 800 000 SK-N-MC cells carrying the Pro7 mutant were significantly lower than in those carrying the Leu7 NPY gene. Levels of NPY peptide in SK-N-MC cell extracts are not shown as they were not consistently above the assay detection limit. In contrast, NPY mRNA levels measured using quantitative PCR did not differ between Leu7 and Pro7 transfected cells, neither in HIT-T15 (b) nor in SK-N-MC (d) cells. Data are mean (±SEM) from individual transfection series performed in triplicates (n = 9–12, RIA; n = 6, quantitative PCR). ***p < 0.001, **p < 0.01, †p < 0.02 versus WT, Mann-Whitney U test following significant Kruskal-Wallis analysis of variance by ranks.
In contrast, quantitative PCR using GAPDH as reference gene showed that NPY mRNA levels were similar in Pro7 and Leu7 transfected cells of both studied cell lines (Fig. 1b, d), indicating that the reduction in NPY levels in mutant transfected cells did not result from a decrease in NPY gene expression. Similar results were obtained using TBP as reference gene (data not shown). The fact that NPY mRNA levels were similar in Pro7 and Leu7 transfected cells would suggest that transfection efficiency was also similar. As further evidence that differences in NPY concentrations in Pro7 and Leu7 transfected cells did not result from a decrease in transfection efficiency, we also tested transfection efficiency in the cell lines by measuring levels of two DNA segments located on the transfection plasmid: the CMV promoter and ampicillin resistance gene. Quantitative PCR confirmed that there were, indeed, no significant differences in transfection efficiency in any of the two cell lines as revealed by the following delta Ct values (mean ± SEM; GAPDH as reference gene): CMV promotor wild-type (−9.3 ± 0.1) versus mutant (−9.0 ± 0.2; p = 0.15); ampicillin resistance gene wild-type (−9.0 ± 0.2) versus mutant (−8.6 ± 0.2; p = 0.17).
Discussion
The present study shows that the Leu7Pro variation may be associated with increased risk of major depression. The frequency of Leu7Pro was not significantly different from the control population in psychiatric disease control patients with schizophrenia. We also show in two cell lines in vitro, that the Leu7Pro variation can be associated with decreased levels of NPY protein, apparently without accompanying changes in NPY mRNA levels.
Our results are not consistent with previous studies, reporting either no difference between depressed patients and controls (Reference Lappalainen, Kranzler and Malison26) or decreased frequency of Leu7Pro variation in depression (Reference Heilig, Zachrisson and Thorsell15,Reference Sjöholm, Melas, Forsell and Lavebratt28,Reference Bosker, Hartman and Nolte29). Thus our study suggests Pro7 to be a risk allele as opposed to a protective allele in previous studies. The lack of agreement between the present and previous studies is puzzling. Smaller sample sizes could play a role. For instance, Lappalainen et al. (Reference Lappalainen, Kranzler and Malison26), who reported no association, studied a smaller sample of 122 depressed patients and the frequency of Leu7Pro in this study was low in the control group (2.0%) compared to that of the present study (3.8%). Similarly, Heilig et al. (Reference Heilig, Zachrisson and Thorsell15) reported the opposite of the present study in an even smaller Swedish sample (n = 51) of treatment resistant depression. However, decreased frequency of the ‘Pro7 allele' was also found in a larger depression sample (n = 461) from a Swedish population (Reference Sjöholm, Melas, Forsell and Lavebratt28), and this was confirmed in a recent replication study using a Dutch population (GAIN; n = 1862) (p = 0.034; OR = 0.73) (Reference Bosker, Hartman and Nolte29).
Differences in ethnic origin of the subjects could also contribute to apparent discrepancies between our study and previous studies. Thus the large Swedish population (Reference Sjöholm, Melas, Forsell and Lavebratt28) was taken from a large cohort of Swedish citizens, in which up to 10% were born outside Sweden (Reference Hällström, Damström-Thakker, Forsell, Lundberg and Tinghög38), with no information on place of birth, which might indicate that some of the cases were of non-Caucasian origin. In the large study by Bosker et al. the individuals were of Dutch or Western European origin (Reference Bosker, Hartman and Nolte29). Given the north-to-south gradient of decreasing frequency of the ‘Pro7 allele' with reported frequencies in Dutch (3.5%) and German (3.6%) versus Swedish (3.9–4.6%) and Finnish (5.8–6.9%) populations and, more importantly, the absence of the ‘Pro7 allele' in non-Caucasian populations (Reference Ding39), it is important that ethnicity of patients and controls is similar. All our patients were Caucasians born in Denmark and only a minority of their parents were born outside Denmark in another European country. To clarify this issue further future studies might explore ethnicity in more depth by genetic stratification using ethnically specific marker genes.
Differences in sampling procedures could also be involved. Thus our patients were diagnosed by trained psychiatrists and recruited among patients hospitalised at psychiatric departments. In contrast, the Dutch cases (Reference Bosker, Hartman and Nolte29) were recruited from mental health-care organisations, primary care and community samples, and the Swedish cases (Reference Sjöholm, Melas, Forsell and Lavebratt28) were identified on the basis of self-reported questionnaires. Yet another explanation could be differences in diagnostic criteria in the different studies. We collected patients in our sample based on a diagnosis of major depression in hospitalised patients (F32 or F33 in the ICD-10 system), whereas the Dutch cases were included on the basis of a lifetime diagnosis of major depression (DSM-IV system) as diagnosed through the Composite International Diagnostic Interview (CIDI Version 2.1.12) (Reference Bosker, Hartman and Nolte29). Moreover, the Swedish study (Reference Sjöholm, Melas, Forsell and Lavebratt28), in addition to major depression, also included patients with dysthymia and mixed anxiety depression. We found no difference in genotype and allele frequencies between patients with comorbid anxiety as compared to patients without, suggesting that our association of Leu7Pro with depression is not influenced by comorbid anxiety.
Other more common variations in the NPY gene have been studied (Reference Zhou, Zhu and Hariri19,Reference Itokawa, Arai and Kato22,Reference Lindberg, Koefoed and Hansen34,Reference Domschke, Dannlowski and Hohoff40). The T allele in the promoter polymorphism −399C>T (rs16147) has been associated with decreased expression of NPY (Reference Itokawa, Arai and Kato22) and diplotypes containing the NPY-399C allele resulted in low expression of NPY and predicted emotional regulation in healthy individuals (Reference Zhou, Zhu and Hariri19). The low expression diplotypes were overrepresented in depressed subjects and also found to increase neural responsivity to negative stimuli within the medial prefrontal and anterior cingulate cortices (Reference Mickey, Zhou and Heitzeg41). However, the diplotype analysis did not include the Pro7 variant; therefore the impact of Leu7Pro was not analysed in the latter study.
The signal sequence in prepro-peptides has multiple functions in targeting and subsequent translocation or insertion of proteins into the membrane of the endoplasmatic reticulum (ER) (Reference Hegde and Bernstein42). During transport into the ER lumen, signal sequences are often cleaved from the precursor protein. Secondary and tertiary structure of the signal sequence of preproNPY could be dramatically altered by the Leu7Pro substitution, as proline easily forms brakes and kinks in alpha-helical structures (Reference Fitches, Appleby, Lane, De, Leone and Olds43). Therefore, it could be expected that the Leu7Pro variation could lead to modifications in the formation, storage, and release of mature NPY. Consistent with this view, previous studies have reported that substitution of leucine with proline in the signal peptide may be associated with drastic decreases in expression of the mature protein. For instance, Lanza et al. (Reference Lanza, De La Salle and Baas44) showed that leucine for proline substitution (also in position 7) in the signal peptide of platelet glycoprotein IX (GPIX) was associated with abolished GPIX synthesis. Likewise, leucine with proline substitution in the signal peptide of calcium-sensing receptor (CASR) resulted in markedly reduced levels of CASR that appeared to be due to a failure of insertion into the ER (Reference Pidasheva, Canaff, Simonds, Marx and Hendy45).
Studying the HIT-T15 and SK-N-MC cell lines, the present study shows that the ‘Pro7 allele' could be associated with a decrease in both intra- and extracellular levels of NPY. A previous study using PC12 cells did not find that the Leu7Pro variation affected the site of cleavage and targeting or uptake of NPY into the ER (Reference Ding, Kull and Liu46). In addition, another study found increased synthesis and secretion of NPY in the Pro7 mutant when co-transfected with ‘Leu7 constructs' (Reference Mitchell, Wang, Ramamoorthy and Whim47). The reason for these apparent discrepancies is not clear. Perhaps the effect of the Leu7Pro variation could be cell-type specific. Thus the effect of Leu7Pro appeared to be more pronounced in HIT-T15 as compared to SK-N-MC cells, and different cell types (PC12 cells, ArT-20 cells and mouse chromaffin cells) were used in the other studies (Reference Ding, Kull and Liu46,Reference Mitchell, Wang, Ramamoorthy and Whim47). Nonetheless, Leu7/Pro7 heterozygous individuals had decreased concentrations of NPY in plasma (Reference Kallio, Pesonen, Jaakola, Karvonen, Helenius and Koulu48,Reference Jaakkola, Kuusela and Jartti49) in accordance with the present in vitro findings of decreased intra- and extracellular NPY levels. However, an increase in plasma NPY concentrations has been reported during exercise of Leu7/Pro7 individuals as compared to Leu7/Leu7 controls (Reference Kallio, Pesonen and Kaipio50). Since only heterozygous individuals were tested in the latter study, it is not directly comparable to our in vitro data. Thus the HIT-T15 and SK-N-MC cell lines are basically a ‘homozygous situation', expressing only the Pro7 allele.
It is not known how the Leu7Pro variation might increase the risk of depression. However, the present finding that the Pro7 substitution was associated with impaired NPY synthesis and decreased extracellular levels in vitro is consistent with data from the literature, that central administration of NPY has actions that may antagonise key symptoms of depression, including orexigenic, sleep inducing, anti-stress and anxiolytic effects (Reference Redrobe, Dumont and Quirion3,Reference Heilig51). In addition, central injection of NPY induces antidepressant-like effects in experimental animal models (Reference Redrobe, Dumont, Fournier and Quirion14,Reference Ishida, Shirayama and Iwata52) and depressed patients have significantly lower concentrations of NPY in CSF (Reference Heilig, Zachrisson and Thorsell15,Reference Widerlöv, Lindstrom, Wahlestedt and Ekman16). Conversely, several types of antidepressant treatments cause increases in NPY gene expression in rodents or in NPY concentration in CSF of human patients (Reference Christensen, Olesen, Kristiansen, Mikkelsen and Woldbye11,Reference Mikkelsen and Woldbye12,Reference Nikisch, Agren, Eap, Czernik, Baumann and Mathé53). Moreover, animal and human data indicate that NPY exerts widespread effects on the release of hypothalamic hormones implicated in the pathophysiology of depression, including corticotropin-releasing hormone (CRH), adrenocorticotropic hormone (ACTH) and gonadotropins (Reference Redrobe, Dumont and Quirion3).
In conclusion, the present study shows that carrying the ‘Pro7 allele' in the signal sequence of NPY may be a risk factor for major depression in the Danish population. Our results were based on a large cohort of patients, two types of controls, more narrow diagnostic criteria, which could account for the results that are in the opposite direction from previous studies. In vitro, the Leu7Pro variation showed functional consequences in the form of impaired NPY biosynthesis and reduced extracellular levels consistent with previously described antidepressant effect of NPY in the brain.
Acknowledgements
We thank Merete E. Nielsen, Birgit H. Hansen, Stine Østergaard, and Bente Bennike for excellent technical assistance. We also thank Preben B. Mortensen for access to some of the patient samples. Funding of this study was provided by The Broedrene Hartmanns Foundation, Danish Psychiatric Research Foundation, H:S Research Council, Ivan Nielsen's Foundation, John and Birthe Meyer Foundation, Pool for Support of Psychiatric Research in the County of Aarhus, and Stanley Medical Research Institute, and Center for Pharmacogenomics, University of Copenhagen. All authors declare that they have no conflicts of interest.
