Letter : Human Hippocampal Neurogenesis Drops Sharply in Children to Undetectable Levels in Adults

To the Editor: Previous studies have concluded that the adult human temporal lobe displays neurogenesis,1 but a manuscript published recently in Nature by Sorrells et al2 showed evidence that the subgranular zone (SGZ) of the dentate gyrus is not a primary neurogenic niche in humans, contrary to what is observed in other mammals. Part of the explanation for this apparent contradiction is that in the adult human temporal lobe neurogenesis may occur mostly in the subiculum,3 which is located at a boundary of the hippocampal formation.4,5 Sorrells et al2 showed that the expression of neurogenesisrelated markers in the SGZ takes place during early infancy, but vanishes in a few years. We have previously shown a similar pattern with newborn and adult samples (Figure 1), but Sorrells et al2 were able to characterize the dynamics of this shift analyzing a large number of specimens with different ages. Nonetheless, the most challenging conclusion by Sorrells et al2 is that the concept stating that the SGZ is a primary adult mammalian neurogenic niche does not apply for humans. This conclusion is in a certain degree complemented by our findings indicating that neurogenesis in the adult human hippocampal formation may have shifted in a great percentage from the SGZ to the subiculum3 across the evolutionary tree. We showed that in the human temporal lobe there is a continuous layer of cells expressing neurogenesis-related markers3 (Figures 1A, 2, and 3). This layer begins (Figure 2A) with the choroid plexus ependymal cells, which express the neural stem cell marker CD133 (Figure 2B). Next, we detected the expression of the neural stem cell marker nestin in cells with long processes located at the taenia fimbria (Figure 2C), which represents the attachment of the choroid plexus to the fimbria. This layer continues through the subpial zone of the medial temporal lobe (Figure 2D) and divides into a minor branch ending at the SGZ and a major branch ending at the Cornu Ammonis 1 (CA1) zone (Figures 2E and 2F). The finding of a minor branch of the nestin positive cells ending at the SGZ is midway between the findings of Sorrells et al2 and a recent study showing expression of neurogenesisrelated markers in the adult human SGZ.6 The analysis of the subiculum has not been reported in these 2 studies.2,6 As in the study by Sorrells et al,2 we3 and Crespel et al7 found nestin positive cells in the molecular layers of the hippocampus proper, located superiorly to the granule cell layer of the dentate gyrus (in opposition to the location of the SGZ, inferior to the granule cell layer). Indeed, the molecular layers are located adjacent to the hippocampal fissure, in a zone that is a remnant of the subpial zone of the hippocampal sulcus during fetal development.3-5,7 The major branch of the nestin positive cell layer in the subpial zone of the medial temporal lobe ends up at the CA1 zone and is adjacent to the subiculum, where we found expression of the immature neuron marker doublecortin (DCX) in neuronal cell bodies and processes (Figure 3A). Next, the DCX positive neuronal processes depict the expected projection from the subiculum to the fimbria and the mammillary body in the hypothalamus (Papez circuit; Figures 3B-3D). Accordingly, the trajectory of DCX positive fibers follows medially through the subiculum (contrary to the CA1 direction), displays a curve that runs inferiorly and laterally up to the subventricular zone of the hippocampus proper, and reaches the fimbria (Figures 1A and 3A-3D). Therefore, as in the study by Sorrells et al,2 no DCX (as well as the immature and mature neuron marker βIII-tubulin) was detected in the SGZ or in the granule cell layer (Figures 3E3H), contrary to what was observed regarding the mature neuron marker microtubule-associated protein-2 (Figures 3I and 3J). With the evidence provided by Sorrells et al2 that the SGZ is not a primary neurogenic niche in adult humans, a next step could be searching for evidence of neurogenesis also in zones that were beyond the scope of their study. The subiculum is a candidate in this regard because indications of neurogenesis were observed in samples of the medial temporal lobe,1,3,8 and we found DCX expression in the subiculum.3 Moreover, the subiculum displays a major enlargement from rodents to primates4 and is the structure where the 3-layer cortex becomes the 6-layer cortex.4 Sorrells et al2 concluded that neurogenesis, if any, is minimal in the adult human SGZ. This result is at least partially in disagreement with other reports.1,3,6-8 Our most plausible explanation for this controversy is that, taking into account the adult human hippocampal formation as a whole (ie, the hippocampus proper [CA1-CA4], dentate gyrus, subicular complex, and fimbria), the SGZ is indeed a minor neurogenic niche in comparison to the subiculum. Anyway, the study by Sorrells et al2 shows that adult human neurogenesis is far from being an issue with consensus. Perhaps, it is time to standardize minimal methodological requirements—which include (but are not restricted to) the maximum premortem agonal period, the maximum time elapsed from death to tissue fixation,3 and a demonstration of the location of expression of neurogenesis markers in the brain cytoarchitecture3,9—and organize an open data repository of “human neurogenomics.”


Letter: Human Hippocampal Neurogenesis Drops Sharply in Children to Undetectable Levels in Adults
To the Editor: Previous studies have concluded that the adult human temporal lobe displays neurogenesis, 1 but a manuscript published recently in Nature by Sorrells et al 2 showed evidence that the subgranular zone (SGZ) of the dentate gyrus is not a primary neurogenic niche in humans, contrary to what is observed in other mammals.Part of the explanation for this apparent contradiction is that in the adult human temporal lobe neurogenesis may occur mostly in the subiculum, 3 which is located at a boundary of the hippocampal formation. 4,5orrells et al 2 showed that the expression of neurogenesisrelated markers in the SGZ takes place during early infancy, but vanishes in a few years.We have previously shown a similar pattern with newborn and adult samples (Figure 1), but Sorrells et al 2 were able to characterize the dynamics of this shift analyzing a large number of specimens with different ages.
Nonetheless, the most challenging conclusion by Sorrells et al 2 is that the concept stating that the SGZ is a primary adult mammalian neurogenic niche does not apply for humans.This conclusion is in a certain degree complemented by our findings indicating that neurogenesis in the adult human hippocampal formation may have shifted in a great percentage from the SGZ to the subiculum 3 across the evolutionary tree.
We showed that in the human temporal lobe there is a continuous layer of cells expressing neurogenesis-related markers 3 (Figures 1A, 2, and 3).This layer begins (Figure 2A) with the choroid plexus ependymal cells, which express the neural stem cell marker CD133 (Figure 2B).Next, we detected the expression of the neural stem cell marker nestin in cells with long processes located at the taenia fimbria (Figure 2C), which represents the attachment of the choroid plexus to the fimbria.This layer continues through the subpial zone of the medial temporal lobe (Figure 2D) and divides into a minor branch ending at the SGZ and a major branch ending at the Cornu Ammonis 1 (CA1) zone (Figures 2E and 2F).
The finding of a minor branch of the nestin positive cells ending at the SGZ is midway between the findings of Sorrells et al 2 and a recent study showing expression of neurogenesisrelated markers in the adult human SGZ. 6The analysis of the subiculum has not been reported in these 2 studies. 2,6s in the study by Sorrells et al, 2 we 3 and Crespel et al 7 found nestin positive cells in the molecular layers of the hippocampus proper, located superiorly to the granule cell layer of the dentate gyrus (in opposition to the location of the SGZ, inferior to the granule cell layer).5]7 The major branch of the nestin positive cell layer in the subpial zone of the medial temporal lobe ends up at the CA1 zone and is adjacent to the subiculum, where we found expression of the immature neuron marker doublecortin (DCX) in neuronal cell bodies and processes (Figure 3A).Next, the DCX positive neuronal processes depict the expected projection from the subiculum to the fimbria and the mammillary body in the hypothalamus (Papez circuit; Figures 3B-3D).Accordingly, the trajectory of DCX positive fibers follows medially through the subiculum (contrary to the CA1 direction), displays a curve that runs inferiorly and laterally up to the subventricular zone of the hippocampus proper, and reaches the fimbria (Figures 1A and  3A-3D).Therefore, as in the study by Sorrells et al, 2 no DCX (as well as the immature and mature neuron marker βIII-tubulin) was detected in the SGZ or in the granule cell layer (Figures 3E-3H), contrary to what was observed regarding the mature neuron marker microtubule-associated protein-2 (Figures 3I and 3J).
With the evidence provided by Sorrells et al 2 that the SGZ is not a primary neurogenic niche in adult humans, a next step could be searching for evidence of neurogenesis also in zones that were beyond the scope of their study.The subiculum is a candidate in this regard because indications of neurogenesis were observed in samples of the medial temporal lobe, 1,3,8 and we found DCX expression in the subiculum. 3Moreover, the subiculum displays a major enlargement from rodents to primates 4 and is the structure where the 3-layer cortex becomes the 6-layer cortex. 4orrells et al 2 concluded that neurogenesis, if any, is minimal in the adult human SGZ.7][8] Our most plausible explanation for this controversy is that, taking into account the adult human hippocampal formation as a whole (ie, the hippocampus proper [CA1-CA4], dentate gyrus, subicular complex, and fimbria), the SGZ is indeed a minor neurogenic niche in comparison to the subiculum.Anyway, the study by Sorrells et al 2 shows that adult human neurogenesis is far from being an issue with consensus.Perhaps, it is time to standardize minimal methodological requirements-which include (but are not restricted to) the maximum premortem agonal period, the maximum time elapsed from death to tissue fixation, 3 and a demonstration of the location of expression of neurogenesis markers in the brain cytoarchitecture 3,9

FIGURE 2. Continuous layer of potential neural stem cells in the human temporal lobe is not centered at the SGZ. A, Hematoxylin-eosin (HE) staining of a coronal cut at the level of the body of the hippocampus. Arrowhead in
-and organize an open data repository of "human neurogenomics."NEUROSURGERY VOLUME 83 | NUMBER 3 | SEPTEMBER 2018 | E133 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/83/3/E133/5033263 by guest on 02 April 2019

FIGURE 1
FIGURE 1.Comparison of results on location of expression of neurogenesis markers.A, Anteromedial view of a coronal cut of the medial temporal lobe at the junction of the head and body of the hippocampus.Black lines (subventricular zone in the hippocampus (a), collateral eminence (b), and SGZ (c)) depict the prevailing view on neurogenic niches.Sorrells et al 2 did not find neurogenesis markers in the SGZ.We found a continuous nestin positive cell layer (dotted blue lines) beginning in the taenia fimbria (arrowhead) and ending in the SGZ and principally in the Cornu Ammonis (CA1)-subiculum boundary, from where DCX expression depicts a continuous layer (red dotted line) up to the fimbria.The line between arrows in the inset corresponds to a zone where we discovered an intraparenchymal ependymal cell layer 3 that seems to underpin part of the trajectory of the DCX positive fibers.B, Image of a coronal cut of a newborn brain at the level of the head of the hippocampus (HH; righthand rectangle) and hypothalamus (H; left-hand rectangle) stained by hematoxylin-eosin (HE).C and D, Lower and higher magnification, respectively.Green (see also arrowhead and arrows, which are closer to the subpial zone): an immunofluorescence image of nestin positive cells in the SGZin the head of the hippocampus of the newborn brain.Cell nuclei are stained in blue by DAPI.Legend: 1, collateral eminence; 2, body of the hippocampus; 3, fimbria; 4, fimbriodentate sulcus; 5, dentate gyrus (margo denticulatus); 6, hippocampal sulcus; 7, parahippocampal gyrus (medial surface) (macroscopic anatomy); 8, subiculum; 9, entorhinal cortex (the presubiculum and parasubiculum are located between 8 and 9); 10, parahippocampus (cytoarchitecture); 11, parahippocampal gyrus (coronal cut) (macroscopic anatomy); 12, uncus; A, amygdala; EC, entorhinal cortex; ES, endorhinal sulcus; GCL, granule cell layer; HS, hypothalamic sulcus; I, insula; IC, internal capsule; III, third ventricle; LV, lateral ventricle; MB, mammillary body; NC, temporal neocortex; SC, subicular complex; T, thalamus; US, uncal sulcus.Scale bars: B = 10,000 μm; C = 500 μm; D = 50 μm.Figure adapted from Nogueira et al 3 under CC BY 2.0 license.
FIGURE 1.Comparison of results on location of expression of neurogenesis markers.A, Anteromedial view of a coronal cut of the medial temporal lobe at the junction of the head and body of the hippocampus.Black lines (subventricular zone in the hippocampus (a), collateral eminence (b), and SGZ (c)) depict the prevailing view on neurogenic niches.Sorrells et al 2 did not find neurogenesis markers in the SGZ.We found a continuous nestin positive cell layer (dotted blue lines) beginning in the taenia fimbria (arrowhead) and ending in the SGZ and principally in the Cornu Ammonis (CA1)-subiculum boundary, from where DCX expression depicts a continuous layer (red dotted line) up to the fimbria.The line between arrows in the inset corresponds to a zone where we discovered an intraparenchymal ependymal cell layer 3 that seems to underpin part of the trajectory of the DCX positive fibers.B, Image of a coronal cut of a newborn brain at the level of the head of the hippocampus (HH; righthand rectangle) and hypothalamus (H; left-hand rectangle) stained by hematoxylin-eosin (HE).C and D, Lower and higher magnification, respectively.Green (see also arrowhead and arrows, which are closer to the subpial zone): an immunofluorescence image of nestin positive cells in the SGZin the head of the hippocampus of the newborn brain.Cell nuclei are stained in blue by DAPI.Legend: 1, collateral eminence; 2, body of the hippocampus; 3, fimbria; 4, fimbriodentate sulcus; 5, dentate gyrus (margo denticulatus); 6, hippocampal sulcus; 7, parahippocampal gyrus (medial surface) (macroscopic anatomy); 8, subiculum; 9, entorhinal cortex (the presubiculum and parasubiculum are located between 8 and 9); 10, parahippocampus (cytoarchitecture); 11, parahippocampal gyrus (coronal cut) (macroscopic anatomy); 12, uncus; A, amygdala; EC, entorhinal cortex; ES, endorhinal sulcus; GCL, granule cell layer; HS, hypothalamic sulcus; I, insula; IC, internal capsule; III, third ventricle; LV, lateral ventricle; MB, mammillary body; NC, temporal neocortex; SC, subicular complex; T, thalamus; US, uncal sulcus.Scale bars: B = 10,000 μm; C = 500 μm; D = 50 μm.Figure adapted from Nogueira et al 3 under CC BY 2.0 license.
FIGURE 2. Continuous layer of potential neural stem cells in the human temporal lobe is not centered at the SGZ.A, Hematoxylin-eosin (HE) staining of a coronal cut at the level of the body of the hippocampus.Arrowhead in A points to the location amplified in the inset, where an intraparenchymal ependymal cell layer 3 is indicated by arrowheads.B, CD133 positive choroid plexus ependymal cells revealed in green by immunostaining and pointed by arrows.C to F, (F is a larger view of the rectangle in E) continuous layer of cells expressing nestin (shown in red by immunofluorescence in C and D and in brown by immunoenzymatic method in E and F) that begins at the taenia fimbria (tf ), runs though the subpial zone (spz; D and arrowheads in E), emits a minor branch to the SGZ (zone inside the dotted lines and pointed by arrow in F) and a major branch that ends at the CA1-subiculum boundary (asterisk in E).Legend: 1 to 4, CA1 to CA4; BdHc, body of the hippocampus; BV, blood vessel; cn, tail of the caudate nucleus; Cp, cerebral peduncle; cp, choroid plexus; cs, collateral sulcus; DG, dentate gyrus; e, ependymal cell layer; ec, entorhinal cortex; f, fimbria; FDS, fimbriodentate sulcus; gcl, granule cell layer; HS, hippocampal sulcus; MG, margo denticulatus; iv, intravascular; lgn, lateral geniculate nucleus; lv, lateral ventricule; phg, parahippocampal gyrus (cytoarchitecture definition); rn, red nucleus; sas, subarachnoid space; SC, subicular complex; sn, substantia nigra; st, stria terminalis; th, thalamus; wm, white matter of the parahippocampal gyrus (gross anatomy definition).Scale bars: A = 5000 μm (inset = 50 μm); B = 20 μm; C and D = 50 μm; E = 500 μm; F = 200 μm. Figure adapted from Nogueira et al 3 under CC BY 2.0 license.

FIGURE 3 .
FIGURE 3. Markers of immature neurons in the hippocampal formation are not located in the SGZ.A to D, DCX (immature neuron) expression revealed in green by immunostaining and pointed by arrows and arrowheads in the subiculum, white matter at the level of the subiculum, fimbria, tracts (fornix and mammillothalamic tract) superiorly (asterisks) to the mammillary body (mb), and mammillary body.E to F, large sagittal sections of the hippocampal formation with glial fibrillary acid protein (GFAP, astrocyte marker) stained in red and a counterpart image of βIII-tubulin (immature and mature neurons) revealed in brown by immunoenzymatic method (horseradish peroxidase (HRP)).Note that βIII-tubulin is expressed in structures related to the projection from the subiculum to the fimbria but not in the SGZ in CA4 (G) and in the entorhinal cortex (EC; H).Contrarywise, the mature neuron marker MAP-2 is expressed in pyramidal neurons (arrows in I), mossy fibers (asterisk in I), and granule cells of the dentate gyrus (DG) (arrows in J),but not in the fimbria (f ).Rectangles 1 to 9 in E show zones analyzed elsewhere3 regarding expression of neurogenesis markers having as reference this large section.Legend: ag, anuclear gap; cp, choroid plexus; dg, dentate gyrus; ec, entorhinal cortex; fds, fimbriodentate sulcus; hs, hippocampal sulcus; wm, white matter.Scale bars: A = 20 μm; B = 100 μm; C and J = 50 μm; D and I = 500 μm; E and F = 5000 μm; G and H = 200 μm.Figure adapted from Nogueira et al 3 under CC BY 2.0 license.
FIGURE 3. Markers of immature neurons in the hippocampal formation are not located in the SGZ.A to D, DCX (immature neuron) expression revealed in green by immunostaining and pointed by arrows and arrowheads in the subiculum, white matter at the level of the subiculum, fimbria, tracts (fornix and mammillothalamic tract) superiorly (asterisks) to the mammillary body (mb), and mammillary body.E to F, large sagittal sections of the hippocampal formation with glial fibrillary acid protein (GFAP, astrocyte marker) stained in red and a counterpart image of βIII-tubulin (immature and mature neurons) revealed in brown by immunoenzymatic method (horseradish peroxidase (HRP)).Note that βIII-tubulin is expressed in structures related to the projection from the subiculum to the fimbria but not in the SGZ in CA4 (G) and in the entorhinal cortex (EC; H).Contrarywise, the mature neuron marker MAP-2 is expressed in pyramidal neurons (arrows in I), mossy fibers (asterisk in I), and granule cells of the dentate gyrus (DG) (arrows in J),but not in the fimbria (f ).Rectangles 1 to 9 in E show zones analyzed elsewhere3 regarding expression of neurogenesis markers having as reference this large section.Legend: ag, anuclear gap; cp, choroid plexus; dg, dentate gyrus; ec, entorhinal cortex; fds, fimbriodentate sulcus; hs, hippocampal sulcus; wm, white matter.Scale bars: A = 20 μm; B = 100 μm; C and J = 50 μm; D and I = 500 μm; E and F = 5000 μm; G and H = 200 μm.Figure adapted from Nogueira et al 3 under CC BY 2.0 license.