Santarelli et al. published their discoveries in 2003. They found that chronic — but not acute — treatment of mice with fluoxetine caused a significant increase in cell proliferation, specifically in neurogenesis. They also found that latency to eat in a novel environment was lowered after 28 days of antidepressant treatment (and importantly, that home cage feeding patterns were not affected). Furthermore, they utilized a 5HT1a receptor knockout mice to assess whether the anxiolytic effects of the antidepressants would persist despite the lack of this receptor, which had been shown to be important for limbic function. Not surprisingly, they found that these mice were not responsive to fluoxetine (an SSRI) but they were responsive to imipramine and desipramine (antidepressants that block norepinephrine reuptake). This demonstrates that this receptor is important for the function of SSRIs, but other antidepressants probably work by other mechanisms.
They also treated WT and KO mice with a direct 5HT1a agonist to show that this anxiolytic effect in WT mice and lack thereof in KO mice was acting directly on the 5HT1a receptor and not by some indirect mechanism. A potential confound of this knockout mouse method is that there may be compensatory mechanisms during development that affect the global circuitry of the brain, and thus its overall function. The authors did mention that the anxious phenotype of these mice results from the lack of the receptor in early postnatal stages, but that does not detract from the possibility of compensatory development.
The authors continued to describe an irradiation technique they used to mitigate cell proliferation in the hippocampus. This involved selectively irradiating the dorsal surface of the brain in only the subgranular zone of the brain (which includes the hippocampus). An issue with this technique is that it irradiates all dorsal-ventral structures in that zone in the rostral-caudal plane. And although there is no evidence for neurogenesis in the other regions of the SGZ, is a reduction in cell proliferation the only effect of irradiation? What effect does irradiation have on, say, synaptic plasticity?
Bessa et al. sought to elucidate the mechanisms in action during antidepressant use; however, they used very different techniques than Santarelli et al. did. It is important to note that Bessa et al. was published 6 years after Santarelli et al. was. In that time, a lot of questions had been asked and a lot of work had been done to address the possible confounds of the original work. For example, in 2008, Holick et al., which included the author Rene Hen (correspondence on Santarelli et al.), showed that the effects of chronic fluoxetine was neurogenesis-independent!
Bessa et al. approached the question by saying: if neurogenesis is blocked, can the effects of antidepressants still be manifested? They used MAM, an antiproliferative agent, to show that after chronic stress, MAM does not have an effect on the ameliorative effects of antidepressants. Interestingly, they found that MAM had no effect on behavior in the forced swim test (FST); however, MAM had a strong effect on behavior in the novelty-suppressed feeding test (NSF) — even in the absence of stress.
The fact that there was a big difference in behavior between groups (MAM+ and MAM-) when comparing FST and NSF highlights the difficulties of finding animal correlates to human behavior, and all of their work highlights the difficulties of finding neural and biologically relevant correlates to behavior. Both the FST and NSF paradigms are supposed to be measures of depression; however, a pharmacological agent that’s supposed to suppress the anxiolytic effects of antidepressants (MAM) had variable effects on the two tests. Does that conclusively address the question of whether neurogenesis is crucial for the therapeutic effects of antidepressants? Maybe neurogenesis is important for manifesting the relieving stress in the context of eating but not in the context of survival swimming — it’s a grey area.
Furthermore, the Golgi analysis of dendritic arborization and qPCR verification of synaptic plasticity markers (Ncam1 and Syn1) established the evidence for neuronal remodeling in healthy mice as well as in those treated with antidepressants. It also evidenced the lack of neuronal remodeling in mice that have undergone chronic stress.
Regarding neuronal remodeling, another interesting finding was that fluoxetine did not increase dendritic arborization in the mPFC, while all the other antidepressants did, and all the antidepressants increased dendritic arborization in the hippocampus. This brings up the issue of studying drugs that are already on the market on the systems level. Is a strong network of dendritic connectivity the strongest combat for depressive behavior? This paper is, of course, does not conclusively say what effects antidepressants have on the systems level, but it is interesting to note that synaptic remodeling in the hippocampus accompanies the amelioration of symptoms caused by chronic antidepressant use.
The definitive effect that antidepressants have on the entire brain to, in turn, affect behavior is still blurry; however, both papers have given important insights into the effects that antidepressants have on different properties of cells (expression of different proteins and presence of morphological changes). The work of one group was able to influence the work of another, which of course, drives science forward.
Note: I was curious about why they stained for both BrdU and Ki-67. What more information does that provide? The patterns of co-localization were the same between BrdU+ and Ki-67+ cells (i.e. co-localization with neuron markers decreased with chronic stress, that effect was reversed with antidepressants, and MAM again decreased co-localization, thereby proving that MAM did indeed lower neurogenesis).
