Neurogeneses Hipocampo 2013
Neurogeneses Hipocampo 2013
Neurogeneses Hipocampo 2013
in Adult Humans
Kirsty L. Spalding,1,8 Olaf Bergmann,1,8 Kanar Alkass,1,2 Samuel Bernard,3 Mehran Salehpour,4 Hagen B. Huttner,1,5
Emil Boström,1 Isabelle Westerlund,1 Céline Vial,3 Bruce A. Buchholz,6 Göran Possnert,4 Deborah C. Mash,7
Henrik Druid,2 and Jonas Frisén1,*
1Department of Cell and Molecular Biology
2Department of Oncology-Pathology
Karolinska Institutet, 171 77 Stockholm, Sweden
3Institut Camille Jordan, CNRS UMR 5208, University of Lyon, 69622 Villeurbanne, France
4Department of Physics and Astronomy, Ion Physics, Uppsala University, 751 20 Sweden
5Department of Neurology, University of Erlangen-Nuremberg, Schwabachanlage 6, 91054 Erlangen, Germany
6Center for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory, 7000 East Avenue L-397, Livermore, CA 94550, USA
7Department of Neurology, Miller School of Medicine, University of Miami, Miami, FL 33136, USA
8These authors contributed equally to this work
*Correspondence: [email protected]
http://dx.doi.org/10.1016/j.cell.2013.05.002
SUMMARY 1985). A seminal study by Eriksson et al., 1998, provided the only
direct evidence to date for adult neurogenesis in humans,
Adult-born hippocampal neurons are important for although it did not enable researchers to assess the number of
cognitive plasticity in rodents. There is evidence new neurons generated or the dynamics of this process.
for hippocampal neurogenesis in adult humans, To estimate the extent of adult neurogenesis in humans,
although whether its extent is sufficient to have func- recent studies have quantified the number of cells expressing
tional significance has been questioned. We have the neuronal precursor (neuroblast) marker doublecortin (DCX)
in the subventricular zone, which gives rise to olfactory bulb neu-
assessed the generation of hippocampal cells in
rons, and in the dentate gyrus of the hippocampus (Knoth et al.,
humans by measuring the concentration of nuclear- 2010; Sanai et al., 2011; Wang et al., 2011). Very similar dy-
bomb-test-derived 14C in genomic DNA, and we namics have been revealed in these two regions, which contain
present an integrated model of the cell turnover dy- a large number of neuroblasts shortly after birth that decreases
namics. We found that a large subpopulation of hip- sharply during the first postnatal year and then declines more
pocampal neurons constituting one-third of the neu- moderately through childhood and adult life (Göritz and Frisén,
rons is subject to exchange. In adult humans, 700 2012; Knoth et al., 2010; Sanai et al., 2011; Wang et al., 2011).
new neurons are added in each hippocampus per The decrease in neuroblast numbers in the subventricular zone
day, corresponding to an annual turnover of 1.75% and their migratory path suggested that there is negligible, if
of the neurons within the renewing fraction, with a any, adult olfactory bulb neurogenesis in humans (Arellano and
modest decline during aging. We conclude that neu- Rakic, 2011; Sanai et al., 2011; Wang et al., 2011). Retrospective
birth dating established that olfactory bulb neurons are as old as
rons are generated throughout adulthood and that
the individual, and, if there is any addition of neurons in the adult
the rates are comparable in middle-aged humans human olfactory bulb, less than 1% of the neurons are
and mice, suggesting that adult hippocampal neuro- exchanged over a century (Bergmann et al., 2012). It appears un-
genesis may contribute to human brain function. likely that adult olfactory bulb neurogenesis has any functional
significance in humans. The similar decline in neuroblast
INTRODUCTION numbers in the subventricular zone and the hippocampus poses
the question of whether there is postnatal hippocampal neuro-
New neurons integrate throughout life in the hippocampus and genesis in humans to an extent that may have an impact on brain
olfactory bulb of most mammals. The newborn neurons have function.
enhanced synaptic plasticity for a limited time after their diff- Analysis of the number of neuronal progenitor cells gives an in-
erentiation (Ge et al., 2007; Schmidt-Hieber et al., 2004), which direct indication of the possible extent of neurogenesis. How-
is critical for their role in mediating pattern separation in memory ever, it does not provide information as to whether neuroblasts
formation and cognition in rodents (Clelland et al., 2009; Naka- differentiate and integrate as mature neurons. This is evident
shiba et al., 2012; Sahay et al., 2011). It has long been debated from the studies of the subventricular zone and olfactory bulb,
whether adult neurogenesis decreased during primate evolution where the generation of neuroblasts does not result in the
and whether there is sufficient generation of neurons in adult hu- detectable integration of new neurons in the olfactory bulb (Berg-
mans to contribute to brain function (Kempermann, 2012; Rakic, mann et al., 2012). The strategies used to study the generation of
detail. For scenario A (constant turnover) and scenario 2POP other is not. In this scenario, cells within the renewing fraction
(constant turnover in a fraction of cells), individual turnover rates are set to turn over at a constant rate throughout life. This sce-
could also be estimated. nario indicated that a large proportion of the nonneuronal cells
(51%, 95% confidence interval [CI] [22%–88%]) are continu-
Turnover of Nonneuronal Cells in the Adult Human ously exchanged. The median turnover rate within the subpopu-
Hippocampus lation of nonneuronal cells undergoing exchange is 3.5% per
First, we assessed the turnover dynamics of nonneuronal year (Figure 2C and Table S2). Individual turnover estimates sug-
(NeuN) cells in the human hippocampus. The 14C concentra- gest that there is a decline in the turnover of nonneuronal cells
tion in genomic DNA corresponded to time points after the birth during aging (r = 0.35, p = 0.04). The average age of nonneuro-
of the individuals (Figures 2A and 2B), establishing the turnover nal cells within the renewing fraction at different ages of an indi-
of nonneuronal cells in the human hippocampus. Mathematical vidual is shown in Figure 2D.
modeling of 14C data allowed a detailed analysis of the dynamics
of cell turnover (Bergmann et al., 2009; Bergmann et al., 2012; Hippocampal Neurogenesis in Adult Humans
Spalding et al., 2008). By fitting the models to the data, we can Next, we analyzed the 14C concentration in neuronal genomic
infer how much cell renewal is needed to reproduce the DNA. One can draw several conclusions regarding hippocampal
observed 14C levels and whether the renewal is restricted to a neurogenesis from the raw data (Figure 3). First, the 14C concen-
subpopulation (see Figure S2 and the Extended Experimental tration in genomic DNA of hippocampal neurons corresponds to
Procedures). The best model, based on AIC, was scenario the concentration in the atmosphere after the birth of the individ-
2POP, in which a fraction of the population is renewing and the ual, confirming the postnatal generation of hippocampal neurons
(p < 2 3 105, Wilcoxon signed-rank test, scenario A in Extended different times throughout life. Stereological quantifications
Experimental Procedures). This finding is largely explained by a have revealed a decrease in the number of hippocampal neurons
larger subset of cells turning over within the nonneuronal popu- during aging in humans, the dentate gyrus being least affected
lation than within the neuronal population, and, when comparing (Figure S4). A relative increase in the proportion of neurons in
the turnover rates specifically within the respective sub- the renewing fraction with age fits the 14C data well.
populations that are subject to cellular exchange, there was no The most detailed model, scenario 2POPEd, provides a global
significant difference in turnover rates between the neuronal picture of the dynamics of neuronal turnover. Nonrenewing neu-
and nonneuronal populations (p = 0.054, Wilcoxon signed-rank rons die without being replaced, resulting in a slow decrease in
test, scenario 2POP in Extended Experimental Procedures). neuron level throughout life. Within the renewing neuron popula-
However, given that nonneuronal cells are more abundant than tion, young cells die faster, leading to a neuron age distribution
neurons in the human hippocampus (Figure 1A), a larger number with fewer middle-aged cells than would be expected if all neu-
of nonneuronal cells in absolute numbers are generated. rons were as likely to be replaced. One observation from the
There was no correlation between the neuronal and nonneuro- modeling is that adult-born neurons are preferentially lost and
nal turnover rates within individuals older than 50 years (r = do not survive as long as the neurons generated during develop-
0.14, p = 0.58, scenario 2POP in Extended Experimental Pro- ment. The half-life of a neuron in the renewing fraction is 7.1
cedures), suggesting that the generation of these different cell years—103 shorter than in the nonrenewing fraction. Although
types is regulated independently, as in the mouse (Steiner it is known that adult-born neurons integrate long term in ro-
et al., 2004). However, there was a correlation in young individ- dents, whether they last for the remainder of the animal’s life
uals (<50 years, r = 0.62, p = 0.003). The interindividual varia- has not been studied, although the available data are compatible
tion in the turnover rate of neurons and nonneuronal cells in with a preferential loss of adult-born neurons (Imayoshi et al.,
the hippocampus is similar, having a median absolute deviation 2008; Kempermann et al., 2003; Ninkovic et al., 2007). The inte-
of 0.0226 and 0.0158 per year, respectively. The interindividual grated model of the dynamics of hippocampal neuron numbers
variation may appear largest in the younger subjects, but this and exchange in humans is shown in Figure 6.
is a consequence of the shallow slope of the atmospheric 14C
levels in recent times, which provides less resolution and, there- DISCUSSION
fore, introduces higher variability.
Newborn neurons in the adult hippocampus have distinct
An Integrated Model of Neuronal Dynamics in the Human features for a limited period after their differentiation that give
Hippocampus them a key role in pattern separation and cognitive adaptability
The determination of the fraction of neurons that is subject to ex- in rodents. We have birth dated hippocampal cells in order
change in the human hippocampus and their turnover rate to assess whether adult neurogenesis occurs to a significant
makes it possible to infer the age of the full complement of neu- extent in adult humans, and we provide a detailed view of the
rons in individuals of different ages. The hippocampus is a cell turnover dynamics. There is substantial neurogenesis
mosaic of neurons of different ages, given the large fraction of throughout life in the human hippocampus, and only a modest
cells remaining from development and neurons generated at decline in neurogenesis occurs during aging. There is a
preferential loss of adult-born neurons, and a larger proportion of integration of new neurons in the olfactory bulb rather than a
hippocampal neurons are subject to exchange in humans in lack of generation of neuroblasts.
comparison to the mouse. Nonneuronal cells have more hetero- There are some distinct differences in the pattern of adult hip-
geneous turnover dynamics than hippocampal neurons. pocampal neurogenesis in humans compared to that in rodents,
It is important to consider whether DNA repair may contribute in which this process has been most extensively characterized.
to 14C integration in hippocampal cells. DNA damage and repair First, a much larger proportion of hippocampal neurons are
are largely restricted to proliferating cells and are believed to be subject to exchange in humans. In mice, 10% of the neurons in
several orders of magnitude below the level that is detectable by the dentate gyrus are added in adulthood and subject to ex-
14
C dating in postmitotic cells (Spalding et al., 2005a). DNA change (Imayoshi et al., 2008; Ninkovic et al., 2007). In humans,
repair during cell proliferation will not affect the assessment of approximately one-third of the hippocampal neurons turn over,
cell generation, given that 14C integrates in DNA at a concentra- corresponding to the vast majority of the dentate gyrus neurons.
tion corresponding to that in the atmosphere during mitosis. We Second, although hippocampal neurogenesis declines with age
have not found any measurable 14C integration in the DNA of in both rodents and humans, the relative decline during adult-
cortical, cerebellar, or olfactory bulb neurons over many de- hood appears smaller in humans in comparison to mice. Com-
cades in humans (Bergmann et al., 2012; Bhardwaj et al., parisons of the kinetics of the age-dependent decline in hippo-
2006; Spalding et al., 2005a). Not even neurons surviving at campal neurogenesis between different species have revealed
the perimeter of an ischemic cortical stroke, a situation where a similar chronology rather than correlating to developmental
there is substantial DNA damage and repair, incorporate suffi- milestones (Amrein et al., 2011). In line with this, the most dra-
cient 14C to be detected (H.B.H., O.B., M.S., G.P., and J.F., matic decrease in the number of neuroblasts in the dentate gyrus
our unpublished data). The dynamics of 14C integration in the occurs during the first postnatal months in both mice and hu-
DNA of hippocampal neurons does not appear to be compatible mans (Ben Abdallah et al., 2010; Knoth et al., 2010). An effect
with any pattern of DNA repair previously described; a large frac- of this is that young adult mice are still in the most steeply
tion of hippocampal neurons (35%) would have to exchange declining phase of neurogenesis, making the relative decrease
their entire genome by DNA repair during the lifetime of an indi- in neurogenesis during adult life much larger in mice than in hu-
vidual, whereas there would be no detectable DNA repair in the mans. Although there is an approximately 10-fold decrease in
remaining hippocampal neurons. In contrast, the number of neu- neurogenesis from 2 to 9 months of age in mice (Ben Abdallah
roblasts reported in the adult human dentate gyrus (Knoth et al., et al., 2010), there is an approximate 4-fold decline during the
2010) is sufficient to give rise to the number of new neurons indi- entire adult lifespan in humans (Figure 5A). Third, the impact of
cated by the 14C analysis, and the decline in neurogenesis adult neurogenesis on the total number of neurons in the dentate
closely parallels the decrease in the number of neuroblasts (Fig- gyrus differs between rodents and humans. Hippocampal neuro-
ure 5A). Thus, the 14C concentration in genomic DNA of hippo- genesis in mice and rats is additive and results in a net increase in
campal neurons is likely to accurately reflect neurogenesis. the number of dentate gyrus neurons with age (Bayer, 1985; Im-
Retrospective birth dating reveals that what appears as small ayoshi et al., 2008; Kempermann et al., 2003; Ninkovic et al.,
numbers of neuroblasts present in adulthood (Knoth et al., 2010) 2007). This is not the case in humans, where there is a net loss
give rise to a substantial number of new neurons over time in the of dentate gyrus neurons during adult life. Although the decrease
hippocampus. In this context, it is interesting that the similar den- in neuronal numbers is less pronounced in the dentate gyrus
sity of neuroblasts in the subventricular zone to that in the hippo- than in other subdivisions of the human hippocampus, the gen-
campal dentate gyrus does not result in any detectable addition eration of new neurons does not keep up with neuronal loss (Fig-
of new neurons to the olfactory bulb (Bergmann et al., 2012; Gör- ure 6). Computational models have indicated that the addition of
itz and Frisén, 2012; Knoth et al., 2010; Sanai et al., 2011; Wang new neurons to the circuitry, along with loss of older redundant
et al., 2011). Thus, the lack of olfactory bulb neurogenesis ap- cells and enhanced synaptic plasticity, can maximize the effect
pears to be a consequence of an absence of migration and/or of the new neurons, whereas an isolated exchange of neurons