Evidence for the neurogenesis hypothesis of depression?
Last week I attended the morning session of the Sackler Institute’s summer program on brain plasticity, or neurogenesis. Rene Hen of Columbia and Eero Castrén of the University of Helsinki, two respected voices in this crowded field, each gave a heroic 90-minute talk on their labs’ investigations of neurogenesis in the hippocampi of adult mice, and its implications for human depression and antidepressant function. Hen made reference to a new paper published in Biological Psychiatry (hereafter Boldrini et al), which provides human-based evidence that is consistent with the neurogenesis hypothesis of depression.
Briefly, the neurogenesis hypothesis states that structural defects in the hippocampus are at the heart of major depressive disorder (MDD). The origin of this hypothesis can be traced back to rodent experiments by Ronald Duman and Eric Nestler in the late 90s, which sometimes elicits a chorus of skeptics decrying “mouse psychiatry!” However, other groups have independently replicated their core observation in subsequent years: modest albeit statistically significant hippocampal neurogenesis in rodent brains chronically exposed to antidepressants or electroconvulsive therapy.
Pivoting to humans from rodents, the first prediction of the neurogenesis hypothesis is that people suffering from MDD would have smaller hippocampi. Examination of postmortem samples has borne out this prediction. However, there’s variance in published studies on human hippocampal volumes. One experimentally validated source of variance, at least in rodents, is the animal’s age; older individuals have lower baseline levels of neurogenesis. Presumably other factors are at play, too.
The second prediction of the neurogenesis hypothesis is that people suffering from MDD but taking antidepressants (ADs) would have normal hippocampi again. Since there’s no way to measure neurogenesis in live humans yet, neurogenesis researchers are again limited to postmortem samples. All the usual caveats apply when examining postmortem samples, and Boldrini et al do a reasonable job of controlling for age, prior substance abuse history, etc. Postmortem samples are clearly not ideal experimental substrates but they shouldn’t be dismissed out of hand given the mountain of rodent-based data.
Okay, with the introduction out of the way, let’s dive into the paper proper. The authors augment and tweak their previous study design by including matched triplets comprised of a non-psychiatric control sample, an untreated MDD sample, and an antidepressant-treated MDD sample. The data set is summarized in Table 1 of the paper:
Not being a neuroanatomist myself, I had to defer to the authors’ prudence in the selection of protein markers of neurogenesis. They chose a protein nestin, which labels neural progenitors cells (NPCs) and the microvasculature, a double whammy of sorts because you also get spatial information on angiogenesis. (Cell types can be distinguished to a first approximation by morphological features, e.g., axons) The authors gained more confidence about which cells are newly born by staining for a nuclear protein called Ki-67, which is an accepted marker of cell division.
The authors spend a fair amount of time explaining the pros and cons of this marker versus that marker because their entire case rests on specificity. For example, glial cells re-express nestin after stress, and so they used a glial specific marker to discriminate between neuronal nestin and non-neuronal nestin. That said, others in the neurogenesis field don’t think nestin is the best marker of neurogenesis, and instead swear by doublecortin.
Moving on, the authors compared the untreated MDD hippocampi to the antidepressant-treated MDD hippocampi, and found evidence for neurogenesis and angiogenesis in the same locations. For example, on average 85% of NPCs (brown/purple blobs) appear to colocalize with capillaries (brown tubes). Quantification aside, the putative neurogenic and angiogenic effects of antidepressant treatment are discernible to the naked eye, as shown here in Figure 4 of the paper:
The coincidence of neurogenesis and angiogenesis makes sense physiologically, and it’s seen in other developing tissues, as well as in tumorigenesis. But as was stated above, cause vs correlation cannot be disentangled using a portmorten approach. There is a contrarian view that neurogenesis may be part of a general stress response, and may not actually constitute the mechanism of the antidepressant response.
Apropos, this paper has a wrinkle that you’d miss if you didn’t read the Discussion section carefully. Close examination of the dentate gyrus, a region of the hippocampus containing “neurogenic niches,” in non-MDD samples vs untreated MDD samples revealed no differences in volume, though I stated above that there is considerable variance in this measurement across independent studies. Here’s what the authors had to say:
“The relationship between MDD, neurogenesis, and angiogenesis is not known. Deficient adult hippocampal neurogenesis is hypothesized as contributing to MDD pathogenesis (10), but we are not finding an effect of MDD on NPC number or vascularization, as we previously reported…it is possible that later stages of NPC maturation or cell survival are compromised in MDD. According to this hypothesis, NPCs would not be fewer inMDDversus control subjects, but neuroblasts or mature granule cells would be. Moreover, a deficit of neuropil could contribute to smaller DG in MDD. We are currently testing these hypotheses.”
It is advisable not to bow at the altar of Serotonin or Neurogenesis. The answer is not something in between, just more complex.