Adult human neurogenesis: early studies clarify recent controversies and go further

Evidence on adult mammalian neurogenesis and scarce studies with human brains led to the idea that adult human neurogenesis occurs in the subgranular zone (SGZ) of the dentate gyrus and in the subventricular zone (SVZ). However, findings published from 2018 rekindled controversies on adult human SGZ neurogenesis. We systematically reviewed studies published during the first decade of characterization of adult human neurogenesis (1994–2004) – when the two-neurogenic-niche concept in humans was consolidated – and compared with further studies. The synthesis of both periods is that adult human neurogenesis occurs in an intensity ranging from practically zero to a level comparable to adult mammalian neurogenesis in general, which is the prevailing conclusion. Nonetheless, Bernier and colleagues showed in 2000 intriguing indications of adult human neurogenesis in a broad area including the limbic system. Likewise, we later showed evidence that limbic and hypothalamic structures surrounding the circumventricular organs form a continuous zone expressing neurogenesis markers encompassing the SGZ and SVZ. The conclusion is that publications from 2018 on adult human neurogenesis did not bring novel findings on location of neurogenic niches. Rather, we expect that the search of neurogenesis beyond the canonical adult mammalian neurogenic niches will confirm our indications that adult human neurogenesis is orchestrated in a broad brain area. We predict that this approach may, for example, clarify that human hippocampal neurogenesis occurs mostly in the CA1-subiculum zone and that the previously identified human rostral migratory stream arising from the SVZ is indeed the column of the fornix expressing neurogenesis markers.


Introduction
The discovery of adult mammalian neurogenesis was a paradigm shift in neuroscience (Altman 1962). The long-lasting idea of "no neuron formation after birth" was replaced by an idea of "two post-natal neurogenic niches" (Ming and Song 2011). Currently, adult mammalian neurogenesis is thought to occur primarily only in the subgranular zone (SGZ) of the dentate gyrus and in the subventricular zone (SVZ). Accordingly, neural stem cells (NSCs) in the SGZ give rise to neurons that integrate into the dentate gyrus circuitry (Ming and Song 2011) and NSCs in the SVZ give rise to neuroblasts that migrate through the rostral migratory stream (RMS) and differentiate into neurons in the olfactory bulb (Ming and Song 2011).
The discovery of adult mammalian neurogenesis triggered the search for adult human neurogenesis (Nogueira et al. 2014b). Concepts on adult human neurogenesis derive principally from rodent studies due to technical difficulties to investigate the human brain, which was the source of a relatively small number of studies (Nogueira et al. 2014b). Adult human neurogenesis was thought to be basically a more discrete adult mammalian neurogenesis (Nogueira et al. 2014b). However, recent studies rekindled the debate on the existence of adult human SGZ neurogenesis and its possible behavior in Alzheimer's disease (Boldrini et al. 2018;Sorrels et al. 2018;Moreno-Jiménez et al. 2019;Tobin et al. 2019).
To discuss the return of initial controversies on adult human neurogenesis, we established an operational classification regarding the eras of development of knowledge on adult human neurogenesis, which is the following: 1) discovery and consolidation (1994 -2004): from the discovery of adult human neurogenesis (Kirschenbaum et al. 1994) to the consolidation of the "two-neurogenicniche" concept in humans (Ming and Song 2011); 2) corroboration of the "two-neurogenic-niche" concept (Kumar et al. 2019) and controversy on the existence of the RMS (2005 -2017) (Nogueira et al. 2014b): increasing number of manuscripts that corroborated the "twoneurogenic-niche" concept, although with controversies regarding the existence of the RMS; 3) controversy on SGZ neurogenesis (2018 -present): novel studies with contradictory results on adult human neurogenesis in normal and Alzheimer's disease brains (Boldrini et al. 2018;Sorrels et al. 2018;Moreno-Jiménez et al. 2019;Tobin et al. 2019).
Here we carry out a systematic review of the first-era studies on adult human neurogenesis to show that the recent publications did not bring novelty on location of human neurogenic niches.
More important, we show that the work by Bernier et al. (2000) in the first era reporting indications of neurogenesis in the limbic system, hypothalamus, and striatum and our work in the second era (Nogueira et al. 2014b) reporting indications of neurogenesis in limbic and hypothalamic structures surrounding the circumventricular organs are alternatives to the prevailing view on adult human neurogenesis that cannot be ruled out by the third-era studies because they did not explore areas beyond the canonical mammalian neurogenic niches (Boldrini et al. 2018;Sorrels et al. 2018;Moreno-Jiménez et al. 2019;Tobin et al. 2019).
Indeed, we have been emphasizing that the analysis of larger brain zones may unveil a bigger picture of adult human neurogenesis (Figs. 1 and 2) (Nogueira et al. 2014b(Nogueira et al. , 2017c(Nogueira et al. , 2018. This bigger picture may confirm our findings showing that in humans a continuous layer expressing neurogenesis markers arising from the choroid plexus reaches the SGZ via a minor branch and the Cornu Ammonis (CA)1-subiculum via a major branch (Nogueira et al. 2014b(Nogueira et al. , 2018 and that the zone previously described as the location of the human RMS is actually the column of the fornix expressing neurogenesis markers or, alternatively, the zone encompassing the anterior hypothalamus or septal area (Nogueira et al. 2014b(Nogueira et al. , 2017c(Nogueira et al. , 2018.

Fig. 1
Structures and flux of the potential neurogenic system in the temporal lobe. Figure A is a scheme of the neurogenesis flux based on our findings regarding the temporal lobe (Nogueira et al. 2014b). Figure B is an anteromedial view of the macroscopic anatomy of the medial temporal lobe sectioned at the level of the body of the hippocampus depicted with lines comparing the canonical neurogenic niches with our results (Nogueira et al. 2014b). Figures C and D are respectively stained by hematoxylin-eosin (HE) and the astrocyte marker glial fibrillary acidic protein (GFAP) and show a large coronal view of the structures of the cytoarchitecture of the medial temporal lobe captured by a slide scanner. The choroid plexus (cp.) is regarded a circumventricular organ. Choroid plexus ependymal cells express the neural stem cell marker CD133. The choroid plexus is attached to the fimbria (f) by the taenia fimbria (arrowhead in B and arrows in C and D), which harbors nestin+ cells with long processes. From the fimbria, a continuous nestin+ cell layer (represented by blue dotted lines in B) follows through the SVZ of the temporal horn of the lateral ventricle and through the subpial layer of the medial temporal lobe. This layer forks as a minor branch toward the SGZ and as a major branch ending at the CA1 zone. Adjacent to this zone (dotted red lines in B) -at the subiculum -DCX+ cell bodies send DCX+ cell fibers that contour the subicular complex, follow across the SVZ covering the hippocampus and next across the fimbria up to the bor-der of the brain slices analyzed by Nogueira et al. (2014b). Black lines identify the canonical neurogenic niches, namely, a, hippocampal SVZ; b, SVZ (collateral eminence); c, SGZ. 1 -4 in C and D refers to CA1 -CA4. Arrowheads in C and its inset point to an intraparenchymal ependymal cell layer (Nogueira et al. 2014b). Rectangle D with two smaller rectangles refer to zone with images amplified in our original paper (Nogueira et al. 2014b). Legend: in B, 1, collateral eminence; 2, body of the hippocampus; 3 (and arrowhead), fimbria; 4, fimbriodentate sulcus; 5, dentate gyrus (margo denticulatus); 6, hippocampal sulcus; 7, parahippocampal gyrus (medial surface) (macroscopic anatomy); 8, subiculum; 9, entorhinal cortex (presubiculum and parasubiculum are located between 8 and 9); 10, parahippocampus (cytoarchitecture); 11, parahippocampal gyrus (coronal cut) (macroscopic anatomy); 12, uncus. In C and D, cn, tail of the caudate nucleus; Cp, cerebral peduncle; cs, collateral sulcus; ec, entorhinal cortex; gcl, granule cell layer of the dentate gyrus; lgn, lateral geniculate nucleus; lv, lateral ventricle; p, pons; pas, parasubiculum; phg, parahippocampal gyrus (cytoarchitecture definition); ps, presubiculum; rn, red nucleus; s, subiculum; sas, subarachnoid space; sn, substantia nigra; st, stria terminalis; th, thalamus; wm, white matter of the parahippocampal gyrus (gross anatomy definition). Scale bars: 5,000 μm.

Methods used in the original articles on adult human neurogenesis
The studies carried out in vivo and retrospective and prospective in vitro methods. In vivo methods corresponded principally to immunostaining for neurogenesis markers performed on brain slices (Tables 2 and 3). Retrospective in vitro methods were used in 11 papers (Tables 4 and 5) (Kirschenbaum et al. 1994;Pincus et al. 1997;Pincus et al. 1998;Johansson et al. 1999;Kukekov et al. 1999;Pagano et al. 2000;Akiyama et al. 2001;Arsenijevic et al. 2001;Palmer et al. 2001;Westerlund et al. 2003;Sanai et al. 2004) and consisted basically in neurosphere assay. Prospective in vitro study was applied in three articles (Table 6) (Roy et al. 2000a, b;Nunes et al., 2003). In these articles, cell sorting was carried out by transfection of plasmid DNA encoding the gene for human green fluorescent protein placed under the control of regulatory sequences of genes encoding the immature neuron marker Tα1-tubulin, the NSC marker nestin (Roy et al. 2000a, b), and CNP2 (Nunes et al. 2003), which selects white matter progenitor cells, also selected by fluorescent or immunomagnetic methods of separation using A2B5 as marker.  (Nogueira et al. 2014b) showing that broadly the expression of markers of degree of differentiation of neuronal cells parallels an anatomical and functional flux. We found indications of this flux related to 1st ) homeostasis in the circumventricular organs such as the median eminence (me), the organum vasculosum laminae terminalis (ovlt), and the subfornical organ; 2nd ) autonomic regulation at, for example, the anterior hypothalamus and paraventricular nucleus (pv); 3rd ) emotion related to limbic structures such as the fornix, mammillary body (mb), and septal area (paraterminal gyrus (pt) plus paraolfactory gyrus (po)); and 4th ) arousal influenced by the nucleus reticularis of the thalamus. Figure B is a macroscopic view of the core of the potential neurogenic system, and the inset shows the sample dissected by cuts delimited by the rectangle in the large figure and studied in Figures C -E. Vertical lines in the inset delimitates the anterior (a), middle (mi), and posterior (p) hypothalamus parts. Rectangles labeled in C and D correspond to zones that we presented somewhere else (Nogueira et al. 2014b) with larger magnification images. Note that the column of the fornix (c and yellow arrows in B, f in D, and Figure E) displays a dense expression of DCX that could have been misinterpreted as the location of the RMS principally if the surrounding structures are not analyzed on the same slice. Legend: A, anterior; ac, location of the anterior commissure; b, body of the corpus callosum; bd, body of the fornix; cf., choroidal fissure; cg, cingulate gyrus; db, diagonal band; dm, dorsomedial nucleus zone; fm, forame of Monro; g, genu of the corpus callosum; h, hypothalamus; HE, hematoxylin and eosin; hs, hypothalamic sulcus; I, inferior; ir, infundibular recess; l, lateral hypothalamic zone; lt, lamina terminalis; m, medial hypothalamic zone; mt, mammillothalamic tract; oc, optic chiasm; or, optic recess; P, posterior; pc, postcommissural fornix; po, pre-optic area; pv, paraventricular nucleus zone; r, rostrum of the corpus callosum; S, superior; s, splenium of the corpus callosum; sc, supra-chiasmatic nucleus zone; sg, straight gyrus; sp, septum pellucidus; sz, stratum zonale; t, thalamus; tc, tuber cinereo; tms, thalamic medial striae; vm, ventromedial nucleus zone. Asterisks identify artifacts. Scale bars: B and C: 5,000 μm; D: 200 μm. Figure  Overall, the most explored zones were the temporal and extra-temporal SVZ and the hippocampus. Less explored zones included the temporal, parietal, and frontal cortices, white matter, amygdala, septum, striatum, and corpus callosum.

Location of possible adult human neurogenesis assessed by in vivo methods
Presence of staining of neurogenesis markers in the SGZ Mitosis activity and the presence of NSCs were demonstrated in the hippocampus (Table 11). A heterogeneous neurogenesis in different hippocampus zones was demonstrated in a study that examined the autopsy brains of patients submitted to bromodeoxyuridine (BrdU) injection for detection of newly formed cells as part of cancer management (Eriksson et al. 1998). BrdU+ cells were present, in decreasing order, in the subgranular and granular layer, and in the hilus. Generally, there was a co-localization of BrdU  Pincus et al. (1998) Committed (Hu protein) and uncommitted (musashi protein) progenitor cells Eriksson et al. (1998) Cell Mt/Pr (BrdU); Nr (NeuN, NSE, calbindin); At (GFAP) Kukekov et al. (1999) Cell Mt/Pr (Ki-67); extracellular matrix protein (tenascin-C); At (GFAP) Bernier et al. (2000) Cell Mt/Pr (PCNA); anti-apoptotic protein (Bcl-2); NSC (nestin); Nr (βIIIt, PSA-NCAM); At (GFAP) Akiyama et al. (2001) NSC ( with neuronal markers and sometimes with glial markers (Eriksson et al. 1998). Jin et al. (2004) (Table 11) studied the brain of normal individuals and Alzheimer's disease patients. In both cases, cells with early neuronal markers were found in the subgranular layer. However, only the hippocampus of Alzheimer's disease patients had stained cells in the granular cell layer and in the CA1, demonstrating more significant neurogenesis in these cases. Furthermore, Western Blot analysis revealed that the hippocampus of Alzheimer's disease patients in comparison to controls displayed more significant expression of immature neuron markers (Jin et al. 2004). Blümcke et al. (2001) carried out another study using nestin as NSC marker, suggesting the existence of NSCs at the hippocampus mostly in the molecular layer and in the hilus (Table 11). Nonetheless, only the hippocampus of Table 3 In vivo methods (phenotypes X markers X locations) *development-related extracellular matrix protein. a, autopsy; AD, Alzheimer's disease (Aβ 1−17 , 5A3, 1G7, phospho-τ); b, biopsy; Bcl-2, product of the antiapoptotic gene Bcl-2; BrdU, bromodeoxyuridine; CR, calretinin; Cx, cortex; DCX, doublecortin; FL, frontal lobe; GAD, glutamic acid decarboxylase; GFAP, glial fibrillary acid protein; Hc, hippocampus; MAP, microtubule-associated protein; M∅, macrophage; NeuN, neuron-specific nuclear protein; NF-L, neurofilament; NSE, neuron specific enolase; OB, olfactory bulb; PCNA, proliferating cell nuclear antigen; PSA-NCAM, polysialylated form of neural cell adhesion molecule; PV, parvalbumin; SVZ, subventricular zone; TH, tyrosine hydroxylase ; TUC-4, turned on / Unc-33- epileptic children younger than two years of age presented a significant number of nestin+ cells. Non-epileptic adult hippocampi obtained in autopsies did not stained for nestin.

Diffuse staining of neurogenesis markers in the SVZ of structures of the limbic system, hypothalamus, and basal ganglia
The entire SVZ was immunoreactive to the neurogenesisrelated markers applied by Bernier et al. (2000) (Table 11) including the hypothalamic portion of the third ventricle. The thickest immunoreactive area in the SVZ was near the anterior commissure and the ventral striatum. The septum also displayed significant immunostaining for neurogenesis markers. Sanai et al. (2004) found cell division markers in the SVZ of the lateral ventricles, in vivo (Table 11). These authors calculated that 0.77± 0.29 % of glial fibrillary acidic protein (GFAP)+ NSCs in the SVZ were in division (Ki-67+). On the other hand, the same authors did not detect cell division in the SVZ near the septum or surrounding the III and IV ventricles. Retrospective in vitro methods used to detect adult human neural stem cells *developmentally regulated transcription factor; Bcl-2, product of the antiapoptotic gene Bcl-2; BDNF, brain derived neurotrophic factor; BrdU, Bromodeoxyuridine; Ctrl, control group; DMEM, Dulbecco's modified Eagle's medium; EGF, epithelial growth factor; FGF-2, basic fibroblast growth factor ; GFAP, glial fibrillary acid protein; GFAPp, GFAP promoter; GFP, green fluorescent protein; GPC, glial progenitor cells; IG, immature glia; INr/At, immature neurons and astrocytes; lac-Z, β-galactosidase; LIF-1, leukemia inhibitory factor; L1, adhesion molecule L1; MAP, microtubule-associated protein; MGl, microglia; Mt, mitosis; N-CAM, neural cell adhesion molecule; Neu-N, neuron-specific nuclear protein NF, neurofilament; NPMM, neural progenitor cell maintenance medium; Nr, neuron; NS d/r, neurosphere division and replating; NSC, neural stem cells; NSE, neuron specific enolase; O, oligodendrocyte; PDGF, platelet derived growth factor; PCR, polymerase chain reaction; Sch, Schwann cell; SVZ, subventricular zone; βIIIt, βIII-tubulin  (Curtis et al. 2007;Sanai et al. 2011;Wang et al. 2011). The conclusions were either that the RMS exists only during a short period of the infancy in a trajectory different from the expected (Sanai et al. 2011) or that the RMS exists in adults in a trajectory different from rodents (Curtis et al. 2007;Wang et al. 2011). We discuss below a possible explanation for this contradiction with basis on our results (Nogueira et al. 2014b), which showed a pattern in which the expression of neurogenesis markers follows neural circuits instead of consisting of a separate zone embedded into the parenchyma.

Third-era studies: contradictory results on SGZ as a neurogenic niche
The lack of signs of neurogenesis at the SGZ showed by Sorrells et al. (2018)

Human brain cytoarchitecture: the necessity to encompass the zone from the choroid plexus to the subiculum in SGZ studies
We put forth that the explanation for contradictory results regarding SGZ neurogenesis in the third-era studies is that the SGZ is not the primary neurogenic niche in the human hippocampal formation (Nogueira et al. 2014b(Nogueira et al. , 2018. Although the negative results by Sorrells et al. (2018) are the opposite of the results of the vast majority of investigations on neurogenesis at the human SGZ, they are not unprecedented. The first-era article by Blümcke et al. (2001) contains high-quality images of nestin staining in children SGZ that was practically absent in normal adults.
Our results are a midterm among the results of the thirdera manuscripts because they revealed that human SGZ neurogenesis exists but is a less significant part of a larger zone expressing neurogenesis markers (Fig. 1) (Nogueira et al. 2014b(Nogueira et al. , 2018. In the temporal lobe, this zone begins at the choroid plexus ependymal cell layer (which expresses the NSC marker CD133) (Nogueira et al. 2014b), encompasses Contradictory results on the existence of the RMS Bédard and Parent (2004) found distribution of cells expressing the immature neuronal markers polysialylatedneural cell adhesion molecule (PSA-NCAM) and doublecortin (DCX) from the olfactory peduncle to the olfactory bulb (Table 11). On the other hand, Sanai et al. (2004) did not find cells marked with immature neuronal markers (PSA-NCAM and βIII-tubulin), between the SVZ at the floor of the anterior horn of the lateral ventricle and the olfactory trigone (the origin of the olfactory peduncle) (

Influence of tissue fixation delay and agonal period has not been sufficiently addressed
The second most important factor for contradictory results on human neurogenesis is probably the time elapsed between death and tissue fixation. For example, we showed that nes-tin+ cells are detected when the time elapsed between death  AD, Alzheimer's disease; APP, amyloid precursor protein; Aβ, β amyloid peptide; Bcl-2, anti-apoptotic protein Bcl-2; BrdU, bromodeoxyuridine; CN, caudate nucleus; CR, calretinin; DCX, doublecortin; DG, dentate gyrus; GAD, glutamic acid decarboxylase; GCL, granule cell layer; GFAP, glial fibrillary acid protein; GlL, glomerular layer; GrL, granular layer; Hc, hippocampus; Neu-N, neuron-specific nuclear protein; NSE, neuron specific enolase; OB, olfactory bulb; PCNA, proliferating cell nuclear antigen; PSA-NCAM, polysialylated form of neural cell adhesion molecule; PV, parvalbumin; RMS, rostral migratory stream; SVZ, subventricular zone; SGZ, subgranular zone; TH, tyrosine hydroxylase; TUC-4, turned on / Unc-33-like phosphoprotein-1 / CRMP-4 family; TUNEL, terminal deoxynucleotidyltransferase-mediated dATP nick and labeling; βIIIt, βIII-tubulin and tissue fixation is less than 16 h on average (Nogueira et al. 2014b). Sorrells et al. (2018) fixed brains after a time span larger than 16 h and obtained no neurogenesis marker staining at the SGZ (although they also have not found this staining in hippocampus samples fixed immediately after surgical resection in epilepsy cases). Agonal period is rarely specified in articles on adult human neurogenesis. None of the third-era studies mentioned this period (Boldrini et al. 2018;Sorrels et al. 2018;Moreno-Jiménez et al. 2019;Tobin et al. 2019). We specified the agonal period of individuals whose brains were studied as a whole (Nogueira et al. 2014b), which should be mandatory in studies with negative results: long agonal period may lead for example to the "respirator brain", with degradation of neurogenesis-related proteins.

Theoretical assumption that neurogenesis must be more significant in rodents than in humans could influence the interpretation of results
The dogma of "no neuron formation after birth" has been eliminated but it seems that it was replaced by the dogma stating that adult human neurogenesis must be proportionally less significant than adult mammalian neurogenesis in general. A curious fact that may be underneath this unmentioned assumption is the publication in a paper regarded seminal of an image of a hippocampus that would be displayed correctly for rodents but is inverted in the craniocaudal direction because it indeed shows a human hippocampus (Eriksson et al. 1998). This disposition perhaps indicates an analysis of results influenced by the familiarity of the authors with rodent brain neurogenesis and anatomy. In other words, perhaps researchers in the field of adult mammalian neurogenesis assume a priori that experiments showing staining of neurogenesis markers in humans proportionally more significant than in rodents are likely a technical error.
Contrarywise, Arsenijevic et al. (2001) and Bernier et al. (2000) found indications that adult human neurogenesis may be not restricted to two discrete and separate neurogenic niches. Likewise, we showed indications that hippocampal formation neurogenesis may have been shifted from the SGZ-granule cell layer zone in rodents to the CA1-subiculum zone in humans (Nogueira et al. 2014b(Nogueira et al. , 2018. This idea is coherent with the fact that the cortical structure where the three-layer cortex becomes the six-layer cortex, i.e. the subiculum, displays one of the major increases in primate brain structures in comparison to rodents (Nogueira et al. 2014b(Nogueira et al. , 2018.

Recent studies brought no novelty on location of adult human neurogenesis and disagree on neurogenesis in Alzheimer's disease
Three recent studies replicated findings of neurogenesis in the SGZ (Table 12) (Boldrini et al. 2018;Moreno-Jimenéz et al. 2019;Tobin et al. 2019). However, their findings disagree between each other and between a first-era study regarding Alzheimer's disease (Jin et al. 2004). Neurogenesis in Alzheimer's disease was increased in the first-era study (Jin et al. 2004), normal in a third-era study but with diminishment in mild cognitive impairment (Tobin et al. 2019) and decreased in another study (Moreno-Jimenéz et al. (2019). Moreover, aging led to a decrease in brain plasticity and angiogenesis although neurogenesis remained at the same level in one of these studies (Boldrini et al. 2018). Therefore, the behavior of the SGZ as a neurogenic niche in Alzheimer's disease remains to be determined.

Current prevailing approach of adult human neurogenesis studies may rekindle possible misinterpretations on human RMS
Our previous findings (Nogueira et al. 2014b) lead to the prediction that if the lack of investigation of expression of neurogenesis markers beyond the boundaries of canonical neurogenic niches is repeated in further studies on the SVZ and RMS, than the probable result will be the return of controversies on the RMS because of the same technical aspects that rekindled controversies on the human SGZ (Nogueira et al. 2014b(Nogueira et al. , 2017c(Nogueira et al. , 2018. Previous studies concluded that the human RMS does not exist (Sanai et al. 2004) or exists in a different trajectory in comparison to other mammals during a short postnatal period (Sanai et al. 2011) or during adulthood (Curtis et al. 2007;Wang et al. 2011). We detailed somewhere else (Nogueira et al. 2014b) that the analysis of a larger brain zone allowed the conclusion that the likely explanation for this controversy is that the zone thought to be the human RMS is indeed the column of the fornix expressing neurogenesis markers (Nogueira et al. 2014b).
Alternatively, the most recent second-era paper assessing the human RMS performed analysis in a more anterior zone and published high-quality images with large brain sections (Wang et al. 2011). In this case, the zone analyzed may be related to the neurogenesis marker staining at the anterior hypothalamus or septal area (Nogueira et al. 2014b;Wang et al. 2011).

Bigger picture analysis may enhance alternative conclusions of an orchestrated adult human neurogenesis
This review showed that recent studies (Boldrini et al. 2018;Sorrels et al. 2018;Moreno-Jiménez et al. 2019;Tobin et al. 2019) reported no novelty on location of neurogenic niches at the human temporal lobe. Moreover, some of these recent studies did not take advantage of previous experimental issues involving premortem agonal state and time elapsed between death and tissue fixation (Nogueira et al. 2014b Remarkably, the possibility of existence of neurogenesis in a broad brain area was identified by Bernier et al. (2000) in a first-era study that remains practically unnoticed. These authors showed anti-apoptotic and NSC markers distributed across the SVZ of the limbic system, hypothalamus, and ventral striatum (Bernier et al. 2000) Moreover, these results were reinforced by findings of the same team in non-human primates (Bernier and Parent 1998), which also displayed neurogenesis in a broad area in another study (Gould et al. 1999) that was received with criticism by the time of its publication (Rakic 2002).
Our previous findings follow the line of evidence revealed principally by Bernier et al. (2000), together with in vitro indications of neurogenesis in non-canonical neurogenic niches such as the amygdala (Arsenijevic et al. 2001). We added novel structures displaying staining of neurogenesis markers in such a way that they seem to reveal a potential neurogenic system that encompasses the SGZ and SVZ (Figs. 1 and 2) (Nogueira et al. 2014b(Nogueira et al. , 2017c(Nogueira et al. , 2018. In short, the potential neurogenic system begins at the circumventricular organs, which are the structures without bloodbrain barrier, and follows sequentially across the hypothalamus and limbic system, with gradual vanishing of staining of neurogenesis markers reaching structures of the reticular activating system (Nogueira et al. 2014b).
A caveat on the existence of a neurogenic system is that there is no explanation for its mechanisms. Joseph Altman's indications of adult mammalian neurogenesis was thought to be unlikely because neurons display no figure of mitosis (Altman 1962). The discovery that the adult mammalian brain harbors NSCs that form new neurons solved this issue. Perhaps neurogenesis could also involve neuron DNA replication and the formation of heterokaryons (Giordano-Santini et al. 2016) (curiously, Altman showed an image of binucleated cortical neurons of rats with at least one nucleus stained for newly synthesized DNA) (Altman 1963). Moreover, our identification of a massive DCX staining in trajectories of certain neural circuits disproportional to the expected neurogenesis rate may have identified neurons in a "stand-by" mode (Marichal et al. 2009). Furthermore, NSCs transplanted into a model of spinal cord injury displayed axonal growth through a long distance (Lu et al. 2012). In short, mechanisms of a potential adult human neurogenic system can only be imagined for now and the methods to assess them remain to be developed.

Lack of specific marker of newly formed neuron
The lack of a specific marker that proves that a neuron is newly formed or formed in the post-natal period is another major limitation in adult human neurogenesis studies (Nogueira et al. 2014b). Markers currently used to identify newly formed neurons are expressed in neurons generated in the fetus that remained immature after birth (Piumatti et al. 2018;Sorrells et al. 2019;La Rosa et al. 2020a;La Rosa et al. 2020b;Seki 2020) or re-expressed in neurons as a response to conditions such as aging and inflammation (Hagihara et al. 2019).
Nonetheless, as we have previously discussed (Nogueira et al. 2014b), one possible explanation for the detection of expression of neurogenesis-related markers in the adult human brain is that it is indeed identifying cells involved in post-natal neurogenesis, such as neural stem cells and newly formed neurons. Accordingly, this possible explanation and its caveats are normally mentioned and taken into consideration in the original studies analyzed in this review. Anyway, the findings of these original studies could serve as basis to figure out the puzzle on locations of adult human neurogenesis, principally when a method that proves the existence of adult human neurogenesis will be revealed.

Location of expression of putative neurogenesis markers may contribute to the revelation of adult human neurogenesis
The locations where cells express different markers of the process of neurogenesis do not prove but are indications compatible with the existence of adult human neurogenesis. For example, hypothalamic neurogenesis is controversial in mammals in general. However, our results showing the expression of nestin in the circumventricular organs (which may be reached by systemic, cerebrospinal fluid, and neural factors) (Nogueira et al. 2014b), surrounded by DCXexpressing brain nuclei, indicate that hypothalamic neurogenesis is at least plausible in adult humans.
Furthermore, our demonstration that the status of hypothalamus-related functions may predict the occurrence of neurologic signs (Kim et al. 2015;Nogueira and Loddenkemper 2017;Nogueira et al. 2017a, b) is coherent with a possible existence of an endogenous mechanism of brain plasticity, perhaps involving neurogenesis. In this regard, we showed that the pattern of circadian oscillation of temperature correlates with the likelihood of occurrence in the following 24 h of seizure in epileptic patients (Kim et al. 2015;Nogueira and Loddenkemper 2017) and of intracranial hypertension in cases of severe stroke (Nogueira et al. 2017a, b). We speculate that normal hypothalamus-related functions clinically detectable reflect a normal hypothalamic plasticity and neurogenesis that serve as an endogenous protection against the installation of neurologic signs (Nogueira and Loddenkemper 2017;Nogueira et al. 2017a, b).
Indeed, we hypothesized that the circadian rhythm monitoring may also determine features regarding other life cycles, including the lifespan (Nogueira and Teixeira 2021). Aging is another hypothalamus-related function (Kim and Choe 2019), and the biological age does not necessarily coincide with the chronological age (Nogueira and Teixeira 2021). For example, conditions such as obesity leads to chronic increase of inflammatory activity (Sandu et al. 2015) and acceleration of biological age (Nogueira and Teixeira 2021). Aging itself leads to a decrease in constitutive and injury-induced neurogenesis (Popa-Wagner et al. 2014). The combination of aging and comorbidities (e.g., obesity, type II diabetes, hyperlipidemia) has been underestimated in preclinical models of stem cell therapies for common diseases such as stroke (Popa-Wagner et al. 2014;Sandu et al. 2015). However, stroke may trigger potential endogenous neurogenesis in canonical (e.g., the SVZ) and non-canonical (e.g., the ependymal cell layer) neurogenic niches (Popa-Wagner et al. 2014). Furthermore, stroke models in aged animals revealed that certain stem cell therapies decrease microgliarelated inflammation and increase endogenous neurogenesis (Popa-Wagner et al. 2014). On the other hand, the addition of a comorbidity led to a deleterious effect with the same type of therapy in aged animals, with rupture of the blood-brain barrier and brain hemorrhage (Popa-Wagner et al. 2014). In short, the translation of neurogenesis-based therapies to clinical practice is more likely to succeed from experimental models that mimic the real clinical situation (Sandu et al. 2015). Perhaps the neurogenesis detected in stem cell therapies in experimental animals could, in a certain extension, be monitored in further clinical trials by the alteration of circadian rhythms in function of the biological age (Nogueira and Teixeira 2021).

Hypothesis that a panel of expression of globins could determine time of birth and degree of maturity of neural cells
We hypothesize that the lack of prove of adult human neurogenesis can be overcome using a panel of globins (Schelshorn et al. 2009;Haines et al. 2013;Emara et al. 2014). Our approach would begin with the analysis of the expression of the relatively recently described neuroglobin as a marker of immature neuron (Haines et al. 2013). Second, to detect whether an immature neuron was born before or after birth we propose to carry out immuno-histochemistry experiments to reveal hemoglobin staining in neurons (Schelshorn et al. 2009;Emara et al. 2014). We wonder whether the expression of fetal hemoglobin shifts to the expression of adult hemoglobin in neurons in the same way it occurs in red blood cells. The rationale for this idea is that the expression of adult hemoglobin in adult neurons has already been shown (Schelshorn et al. 2009). To the best of our knowledge, the expression of fetal hemoglobin in fetal neurons has not been studied, however, such expression would follow the same physiological pattern of red blood cells; furthermore, it has already been shown that glioblastoma samples, which contains glioblastoma stem cells, display expression of fetal hemoglobin (Emara et al. 2014).
The primary expected result would be the co-expression of neuroglobin and adult hemoglobin in immature neurons formed after birth. Importantly, this hypothesis could be validated using, for example, the BrdU paradigm in rodents. The results could be reproduced easily, especially in comparison to other methods used to detect the formation of neurons in the adult human brain that in some researchers' view actually may be detecting DNA repair instead of DNA replication during the formation of a neuron (Sorrells et al. 2018).

Conclusions
We conclude that the prevailing view on adult human neurogenesis has been developed by minor increments as from the period when the "two-neurogenic-niche concept" in humans was consolidated. We propose that the standardization of minimal requirements for research on adult human neurogenesis could foster research in this field (Nogueira et al. 2014b(Nogueira et al. , 2017c(Nogueira et al. , 2018. We suggest that further studies should bear in mind the possibility that adult human neurogenesis is organized across the brain circuitry, involves the brain as a whole (Nogueira et al. 2014b(Nogueira et al. , 2017c(Nogueira et al. , 2018 and is a primary factor of brain plasticity (Nogueira et al., 2014a, b;Kim et al. 2015;Nogueira and Loddenkemper 2017;Nogueira et al. 2017a, c).