4 December 2009. While the possibility of regenerative cell therapy for Alzheimer disease remains fraught with hope and hype, two papers in today’s issue of Cell Stem Cell make one thing clear: the growth of new neurons relies on proper inhibitory GABA signaling. In independent, but complementary studies, researchers at the Gladstone Institute of Neurological Disease, San Francisco, report that Aβ peptides and apolipoprotein E4 impair neurogenesis in AD mouse models, and that targeting GABA neurotransmission can rescue these defects. Though the exact relationship between neurogenesis and AD remains unclear, the new findings suggest it may be possible to restore the brain’s ability to pump out new neurons even in highly toxic scenarios.
Given that the hippocampus is both a key site for neurogenesis and a prime target for amyloid deposition and neuronal loss in AD, it is natural to question whether Aβ has impact on the brain’s ability to churn out new cells. Thus far, though, the experimental evidence to support that idea has been controversial. Some studies have found increased neurogenesis in amyloidogenic mouse models (Jin et al., 2004b) and AD patients (Jin et al., 2004a; He and Shen, 2009), whereas others have not
(Boekhoorn et al., 2006; Li et al., 2008).
However, most of those investigations tracked numbers of adult-born neurons without addressing function. “What is new in the [current papers] is they’re using methodology that goes beyond quantifying cells,” said Henriette van Praag of the National Institute on Aging in Baltimore, Maryland, who was not involved with the work. “They're really looking at what happens to newly born neurons in these AD mouse models in terms of function, development, and maturation.” Li Gan led one study that used J20 mice (an AD transgenic strain that overexpresses mutant human amyloid precursor protein [hAPP]), and Gladstone colleague Yadong Huang was principal investigator on the second study, which looked at neurogenesis in various lines of ApoE knockout and knock-in mice. ApoE4 is the strongest genetic risk factor for late-onset AD.
In the first study, joint first authors Binggui Sun and Brian Halabisky and colleagues used confocal microscopy and patch-clamp recordings to analyze morphology and function of newborn neurons in wild-type and J20 mice. The researchers used retroviruses expressing enhanced green fluorescent protein (EGFP) to label newborn granule cells in the dentate gyrus, a site for neurogenesis in the hippocampus. Since only mitotic cells take up virus and express EGFP, the number of days post-infection could be taken to represent the age of the labeled neurons.
This approach revealed accelerated development of J20 granule cells during the first three weeks after their birth. In terms of morphology, newly born neurons in J20 mice had longer dendrites and twice the spine density of wild-type animals. Functionally, the transgenic cells had stronger evoked inhibitory post-synaptic currents (eIPSCs). “And more decisively, if you look at their Cl-reversal potential, [J20 newborn neurons] are more hyperpolarized. That's a very clear indication that they are more mature, or that they mature faster,” Gan told ARF. GABAA-mediated Cl- reversal potentials typically transition from depolarizing to hyperpolarizing during the third week after the birth of granule cells (Ge et al., 2007).
By four weeks, though, the trend had reversed. Despite their early growth spurt, 28-day-old adult-born granule cells in J20 mice had shorter dendrites, fewer dendritic branches, and lower spine density than did wild-type newborn neurons. These differences persisted through 122 days post-infection, when new granule cells have presumably completed their maturation. The functional data supported the morphology: by 28 days, newborn J20 neurons had smaller eIPSCs compared to adult-born granule cells in wild-type mice.
Interestingly, the spine defects largely disappeared when the researchers analyzed a transgenic strain expressing wild-type hAPP at levels comparable to those of mutant hAPP in the J20 mice. Furthermore, the morphological abnormalities in J20 granule cells progressively intensified with age, paralleling an increase in hippocampal Aβ42 levels. These data point the finger more squarely at Aβ, and not APP itself, as the trigger for the impaired neurogenesis. Consistent with this idea, a study published in the November 11 Journal of Neuroscience suggests that Aβ immunotherapy can protect morphology and survival of newborn neurons in APP/PS1 mice (Biscaro et al., 2009).
As for potential mechanisms underlying the impaired neurogenesis in AD mice, the intriguing pattern of early acceleration followed by impaired maturation strongly implicated GABA signaling dysfunction. “That wasn't my hypothesis to begin with. I wasn't thinking of GABA,” Gan told ARF. She initially placed her bets on the cholinergic system, presuming that reversal potentials of adult-born J20 neurons would never switch from depolarizing to hyperpolarizing. Instead, “the switch occurred much faster,” Gan said.
To test whether the accelerated early development of newly born J20 granule cells in fact stemmed from excess GABAergic signaling, the researchers treated the mice with a GABAA receptor antagonist (picrotoxin) to inhibit GABA neurotransmission for the first seven days of development. Sure enough, early treatment with picrotoxin normalized the unusually long and dense spines in J20 adult-born neurons. Even more striking, Gan said, is that early administration of the drug in wild-type mice led to decreased spine density and dendritic length at 28 days. “Those neurons have normal GABA. If you block it, it’s detrimental to their growth,” she said.
In some ways, Gan’s findings run counter to Huang’s observations in ApoE4 knock-in mice. Whereas the hAPP mice had excessive GABAergic transmission, Huang and colleagues report that ApoE4 knock-in mice have reduced GABA signaling due to lower numbers of GABAergic interneurons in the dentate gyrus. However, the abnormalities in ApoE4 mice led to developmental problems that closely resembled those Gan’s team saw in newly born granule cells of J20 hAPP mice. Using bromodeoxyuridine (BrdU) labeling, first author Gang Li and colleagues found that “more newborn cells get generated [in ApoE4 knock-in mice] after one day, but fewer go on to become mature neurons,” Huang said. “They die in the middle stages.” The morphological defects in newborn neurons of ApoE4 knock-in mice were also consistent with those in the J20 hAPP mice. And, in this week’s issue of the Journal of Neuroscience, a team led by Hyang-Sook Hoe of Georgetown University Medical Center, Washington, D.C., reports that ApoE4 knock-in mice have lower spine density and dendritic complexity compared with ApoE2 and ApoE3 knock-in animals.
However, despite similar phenotypes, the directionality of the GABA defect—and hence the treatment for impaired neurogenesis—differed in ApoE4 knock-in and J20-hAPP mice. “In our case, [the ApoE4 mice] start with lower GABA signaling,” Huang said. As such, boosting GABA neurotransmission with a GABAA receptor agonist (pentobarbital) appeared to rescue development of newborn granule cells. “In [Gan’s] case, the baseline GABA inhibition is increased. So she rescues neurogenesis by bringing [GABA signaling] down to normal levels,” Huang said. “You can imagine that in normal conditions, you need a normal range of GABA signaling. Too low is bad. Too high is not good, either.”
The Huang and Gan labs have crossed their mice to create a bigenic strain that expresses ApoE4 and high levels of mutant APP. These mice will enable the scientists to ask, in the presence of both sources of toxicity, which will dominate, and what their net effect on neurogenesis will be.
Like Gan, Huang did not set out to look at GABA. His team’s starting point was the observation of high ApoE expression in the subventricular zone within the dentate gyrus of their EGFP-ApoE reporter mice. These animals have EGFP, inserted into the ApoE gene locus. Huang and colleagues engineered the mice to provide a real-time readout for ApoE expression, in hopes of settling a longstanding debate about whether neurons make ApoE. In the reporter mice, neurons did not make ApoE under normal conditions, but did in response to excitotoxic injury (Xu et al., 2006). They noticed something else in those mice. “There were many very strong green cells in the subgranular zone,” Huang told ARF. “This caught our eye right away, and suggested to us that [the green cells] might be neural stem cells.”
The new paper describes how his team confirmed this hunch. Staining brain sections from EGFP-ApoE reporter mice with antibodies against neural stem cell markers (nestin and Sox2), they saw that fluorescent cells were positive for both. Having shown that neural stem cells make ApoE, Huang figured “there might be some critical function there.” His team BrdU-labeled hippocampal cells in ApoE knockout mice, and found that they produced about the same number of new cells as did wild-type animals, but more of the cells became astrocytes instead of neurons. Revealing a possible mechanism, Huang and coworkers showed that neural stem cells from ApoE-deficient mice had 80 percent reduced levels of Noggin, a protein known to inhibit astrogenesis and promote neurogenesis. In line with this, they found that adding recombinant Noggin to cultures of neural stem cells from ApoE-knockout mice prevented glial differentiation and restored wild-type levels of neurogenesis. Similarly, a study in the October issue of PLoS ONE suggests that infusing Noggin into the brain enhances cognition and boosts neurogenesis in wild-type mice (Gobeske et al., 2009).
In spite of these and other recent advances, uncertainty remains as to how neurogenesis ties in with AD. While some studies (e.g., Winocur et al., 2006) show that blocking neurogenesis impairs learning and memory in rodents, other reports suggest that production of new neurons is not so essential in cognition-boosting scenarios—for example, environmental enrichment—in AD mouse models (Meshi et al., 2006 and ARF related news story).
As for the use of stem cell therapy in AD, “we are very far from that,” Gan said. “Whether using transplanted cells or stimulating endogenous precursors, one thing we need to understand for any regenerative strategy is how those neurons can survive in a highly toxic environment.” The new papers offer hope in this regard, suggesting that proper tweaks to the GABAergic system may be able to relieve neurogenesis impairment in the presence of ApoE4 or pathogenic Aβ peptides.
Van Praag thinks the morphological and functional features of neurogenesis described in these papers could have potential as biomarkers for what is happening in the rest of the brain. Because newborn neurons seem more sensitive to changes in the local environment, they “may be a well-defined population that could give us an interesting readout of disease state,” she said. “An intervention that would ameliorate some of the functional deficits of neurogenesis could probably be promising as a general treatment.”—Esther Landhuis.
Sun B, Halabisky B, Zhou Y, Palop JJ, Yu G, Mucke L, Gan L. Imbalance between GABAergic and Glutamatergic Transmission Impairs Adult Neurogenesis in an Animal Model of Alzheimer’s Disease. Cell Stem Cell. 2009 Dec 4;5:624-633. Abstract
Li G, Bien-Ly N, Andrews-Zwilling Y, Xu Q, Bernardo A, Ring K, Halabisky B, Deng C, Mahley RW, Huang Y. GABAergic Interneuron Dysfunction Impairs Hippocampal Neurogenesis in Adult Apolipoprotein E4 Knockin Mice. Cell Stem Cell. 2009 Dec 4;5:624-633. Abstract
Dumanis SB, Tesoriero JA, Babus LW, Nguyen MT, Trotter JH, Ladu MJ, Weeber EJ, Turner RS, Xu B, Rebeck GW, Hoe H-S. ApoE4 Decreases Spine Density and Dendritic Complexity in Cortical Neurons In Vivo. J. Neurosci. 2009 Dec 2;29(48):15317-15322. Abstract