The mechanism of action of antidepressant drugs is far from well understood. The "catecholamine deficiency hypothesis" stated that affective disorders like depression originate in pathologic reduction of catecholamine levels in the brain 1.

The beneficial effects of antidepression treatment on serotonin and dopamine turnover and the associated receptors start a few weeks after onset of therapy, casting doubt on a simple mode of action, which should be effective much faster. Other classes of psychotrophic drugs, like benzodiazepine, which interact with neurotransmitter receptors too, become effective within minutes after drug administration. The mechanism of therapeutic latency of antidepressant and neuroleptic drugs is not clearly understood. Current hypotheses include slow adaptive processes after fast access to primary drug targets2.

Alternatively, it was proposed that antidepressant drugs alter gene expression or induce nerve plasticity by rearranging neuronal connections 3. Because the formation and modification of neuronal connections depend on nerve activity 4, it appears very likely that drugs trigger formation of new connections and strengthens already existing by enhancing exocytosis 5. Recent work on exo- and endocytosis revealed profound effects of antidepressant drugs on neurotransmitter secretion and receptor cycling 67. TGFβ2 occurs in astroglia and various neurons and is known to promote the effects of growth factors such as NT-3 and 4, GDNF, CNTF and bFGF 8. Therefore up-regulation of TGFβ2 and it's secretion via regulated pathways by PC-12 cells and hippocampus neurons point towards indirect regulatory roles 9.

In addition to pharmacological treatments, rTMS was first introduced in 1985 10 as a non-invasive tool of brain stimulation. It was proposed that rTMS might be a promising technique for the therapy of depression 1112. However, recent meta analyses and multi-center trials revealed that rTMS was not as effective as ECT, specifically for the short term treatment 1314.

This study focuses mainly on the analysis of growth factor transcription by real time PCR analysis in a cell culture system, after administration of antidepressant drugs. Additionally, morphological effects of antidepressant treatment were analysed by microscopy and effects of magnetic stimulation on growth factors were also investigated in preliminary pilot experiments.

We hypothesize that antidepressant drugs rather modulate the expression of growth factors than just simply increase their expression, which in turn stabilize, enhance or even create neural connections between specific brain areas, involved in mood regulation.

Because of increasing evidence against the "catecholamine deficiency hypothesis" to explain patho-mechanism of depression, an alternative hypothesis on the mechanism of antidepressants, the "plasticity hypothesis", was proposed 171819.

In this study we analyzed the transcription of growth factors in cell culture systems. SY5Y cells were pre-treated by the phorbolester PMA to induce differentiation into neuronal cells 1520. PMA changed transcription levels of several growth factors massively, but we found that antidepressant drugs counteracted these PMA-induced changes.

Recent post mortem studies in suicide subjects revealed significantly lower levels of BDNF in the prefrontal cortex and hippocampus. Authors of several studies provided evidence that BDNF could be involved in depressive behaviour 21. Because growth factors induce the formation of neurite-like processes 2223, we concluded from our microscopic images that the growth factors were also expressed in SY5Y cells. However, we couldn't confirm the expression for every factor (data not shown). The morphological changes provide some evidence that the modulation (up and down regulation) of several growth factors is responsible for the observed effect. A direct correlation with a single growth factor, like BDNF, can not be causal 24. All four classes of antidepressants, tested in our study, antagonistically reversed strong up or down regulation of growth factors, induced by PMA in cultured SY5Y cells 1520.

The up-regulation of BDNF and Creb in vivo seem to be a central part of the antidepressant properties of antidepressants 171925, however, there is few evidence for the variation of other growth factors so far. Different modes of stress can induce either selective up and down regulation of BDNF splice variants 26 Earlier gene transcription studies in in vivo systems had shown that several genes were either up- or down-regulated by antidepressant treatment 2728. These and our observations assign a role of modulators to antidepressants, rather than pure stimulators of growth factor transcription.

We found that if transcription of a factor increased during the cultivation period, this increase was attenuated by antidepressants. On the other hand, antidepressants increased the transcription of factors that became down regulated upon PMA-treatment. The antagonistic regulation of growth factors was a common property of all antidepressant classes; however, there were differences in kinetics and specificity. Amitriptyline, for example, did not prevent the PMA-induced decrease of bFGF and CNTF, but it worked antagonistically for all other factors. In general, it favoured more down regulation than other drugs. Reboxetine worked in an exactly opposite way on the transcription of growth factors, by strongly increasing down regulated factors and modestly reducing up regulated factors. Fluoxetine and tranylcypromine represented classes with similar modulator properties, but with different kinetics. Fluoxetine has been shown to increase the expression of bFGF in rats 29. This was observed in our experiments too.

In a pilot experiment PC-12 cells were treated with rTMS. These cells were chosen, because they are electrical excitable, which made them potentially susceptible for repetitive magnetic stimulation. These cells showed an increase in BDNF, TGFβ2 and tyrosine hydroxylase (figure 3A). In contrast to PC-12 cells, SY5Y cells showed a decrease of BDNF (figure 3B). BDNF mRNA was elevated in all treated samples, compared to controls; however, statistical significance was not reached, due to small sample numbers. It has been shown previously that rTMS and antidepressants acted in a similar way in rat hippocampus 30 and humans 3132. We could only partly support this hypothesis.

In summary, we found that antidepressants normalized PMA-induced growth factor variations and specifically modulated their expression, depending on the type of the drug. We conclude from our experiments that antidepressant drugs rather modulate PMA-induced changes of growth factor transcription, than induce an overall transcription increase. This suggests that a common basic mechanism of antidepressants normalizes pathologic deregulation of growth factors, which without treatment would affect neuronal plasticity.

AWH: Experimental design, writing of the manuscript and supervision of all experiments. WS: Design, writing of the manuscript and supervision of the TMS experiments. AR: Conducted the experiments on antidepression drugs and analyzed the data. UR: Statistical analysis. CR: Conducted the experiments on repetitive magnetic stimulation and analyzed the data. DB: Supervision of PCR and construction of primer pairs. JMM: Supervision of cell culture. JK: Organization, literature, concept and interpretation. JW: Organization, interpretation and supervision. All authors read and approved the final manuscript.