Research on stem cells has developed as one of the most promising areas of neurobiology. In the beginning of the 1990s, neurogenesis in the adult brain was indisputably accepted, eliciting great research efforts. Neural stem cells in the adult mammalian brain are located in the ‘neurogenic’ areas of the subventricular and subgranular zones. Nevertheless, many reports indicate that they subsist in other regions of the adult brain. Adult neural stem cells have arisen considerable interest as these studies can be useful to develop new methods to replace damaged neurons and treat severe neurological diseases such as neurodegeneration, stroke or spinal cord lesions. In particular, a promising field is aimed at stimulating or trigger a self-repair system in the diseased brain driven by its own stem cell population. Here, we will revise the latest findings on the characterization of active and quiescent adult neural stem cells in the main regions of neurogenesis and the factors necessary to maintain their active and resting states, stimulate migration and homing in diseased areas, hoping to outline the emerging knowledge for the promotion of regeneration in the brain based on endogenous stem cells.
Stem cells are no longer just separate entities or an operational definition of toti-, pluri-, multipotent elements, instead in principle every cell can become a stem cell in vitro and replace endogenous cells in the body.
The discovery of stem cells and their innovation, the ability to generate induced pluripotent stem cells (iPS), together with the advancements in developmental and cellular biology, have laid the foundation to unify biology and medicine as never before. The easiness of laboratory protocols to generate iPS, which by many aspects are similar to embryonic stem cells (ES), from animal and human somatic cells (Takahashi and Yamanaka 2006; Yamanaka and Blau 2010) can change our way of studying development and diseases. The human iPS make it possible to study neurons, previously inaccessible, bearing the genetic information of patients with a specific mutation in a diseased cellular context, which would be difficult to replicate in genetically modified animals.
Thus, the new frontier now opened by the technological advancements and the discovery of adult neurogenesis can help us to generate neurons directly from patient's somatic cells or redirect endogenous neurogenesis. However, it is not yet clear whether iPS could recapitulate all the hallmarks of a disease in vitro and be a physiological ‘test tube’ for drug discovery (Han et al. 2011). Nonetheless, this technology may one day make it possible to generate individual cell therapy for the patients. However, major problems persist regarding the possible use of these ‘new’ stem cells, such as chromosomal instability of iPS or the low efficiency of reprogramming (Hussein et al. 2011). On the other hand, other potential sources of stem cells, as those of human embryonic origin, raise ethical concerns and are still highly controversial (Cattaneo 2011). Therefore, endogenous adult neurogenesis is regarded as a pivotal avenue to heal the brain, which could also bypass the ethical debate.
To develop therapeutical tools from adult neurogenesis, we need to deepen our knowledge on the biology of neural stem cells, to gain insights into the etiology of major brain disorders and to solve the problems related to restoration of brain connectivity.
Stem cells were discovered after the bombing of Hiroshima and Nagasaki in 1945, when more than 140.000 people died immediately but many others developed severe leukemia and total loss of the bone marrow because of radiation (Jacobson et al. 1951). Not long later hematologists discovered that the bone marrow contains a population of cells capable of replacing all the blood cells (Ford et al. 1956; Nowell et al. 1956).
During development, stem cells are the founder cells of every tissue, organ and cell in the body of animals and plants. They are undifferentiated cells, which do not yet have a specific function other than that of being toti-, pluri-or multipotent. In the adult organisms, they constitute a sort of repair system of the body and can divide without limit to replenish other cells as long as the animal or plant is alive. In their continuous dividing, stem cells maintain stable genetic information, as well as stable epigenetic modifications of the chromatin structure.
Until the discovery that differentiated cells can be reprogrammed to pluripotency, a stem cell was defined as a cell with two fundamental properties: the ability to self-renewal through cell division (asymmetric division) and the competence to generate specialized cell type(s) through differentiation. The asymmetric division is a distinctive biological characteristic, which distinguishes the stem cells (or multipotent progenitors) from all other cell types where the mitotic division is symmetrical.
With development, their numbers in most tissues decrease. However, in the adult various organs retain a stem cell compartment, located in a specific microenvironment, called niche, that regulates their behavior (Doetsch 2003b). Stem cells may be present in different proliferative states in different organs. For example, niches show active proliferation in tissues that require a high cell turnover, such as blood, the epidermis, the intestinal epithelium and male gonads, allowing an effective balance between the loss of cells and their replacement. Instead, in tissues where cell loss is limited, such as the liver, the teeth and the brain, the stem cells are present in a quiescent or low proliferation state, which can be activated by physiological and/or pathological stimuli.
Operationally, embryonic or adult stem cells can be also identified by their ability to form cell aggregates when grown in vitro in suspension in the presence of specific factors. This is true also for neural stem cells (NSC), which form the so-called neurospheres (Reynolds and Weiss 1992). These structures are heterogeneous and are constituted by NSC and neural progenitors at different stages of differentiation. When dissociated in single cells, NSC can give rise to all three cellular lineage of the CNS: neurons, astrocytes and oligodendrocytes. Specific mitogens, morphogens and growth factors are necessary for their proliferation and survival in culture and their differentiation, according to defined protocols (Chojnacki and Weiss 2008).
Adult neural stem cells
Adult neurogenesis recapitulates the complete process of neuronal development in embryonic stages. One hallmark of adult neurogenesis is its sensitivity to physiological and pathological stimuli at almost every stage, from proliferation of neural precursors, to development, maturation, integration and survival of newborn neurons (Zhao and Overstreet-Wadiche 2008). Until a few decades ago, it was thought that the adult brain had no stem cells, based on the statement by Ramon y Cajal ‘Once development was ended, the fonts of growth and regeneration of the axons and dendrites dried up irrevocably’ that strengthened the idea that new neurons could not be formed and integrate in an adult brain. Now instead the notion of adult neurogenesis is widely accepted, because of the overwhelming data accumulated in the last 20 years or more (Richards et al. 1992; Colucci-D'Amato and di Porzio 2008). Briefly, the first evidence of neurogenesis in the adult brain goes back to the 1960s when Joseph Altman showed that cells in the dentate gyrus of the hippocampus could incorporate radioactive thymidine (Altman 1962), while Kaplan and Hinds (1977) coupling the autoradiographic technique to electron microscopy demonstrated that the thymidine-labeled cells were neurons. Subsequently, the pivotal work of Nottebohm and collaborators showed that neurogenesis occurs in adult songbird brains (Paton and Nottebohm 1984). Together with the discovery of neurospheres these data put an end to the debate (Reynolds and Weiss 1992). In mammals, the adult NSC (AdNSC) are mainly confined in the ‘neurogenic’ areas, that is, the subgranular zone (SGZ) in the dentate gyrus of the hippocampus, where new granular neurons are generated, and the subventricular zone (SVZ) of the lateral ventricles where new neuroblasts are continuously produced and migrate through the rostral migratory stream (RMS) to the olfactory bulb (Fig. 1; Gage 2000; Curtis et al. 2007). The newly formed neurons integrate into existing circuits and are essential for specific brain functions, in mammals as in birds although the two species diverged during evolution more than 300 million years ago (Toni et al. 2008; Lazarini and Lledo 2011). In adult birds, the new neurons are produced in the lateral ventricle, migrate through the telencephalon, differentiate and integrate into functional circuits (Alvarez-Buylla and Nottebohm 1988). In adult rodents' olfactory bulbs, up to thirty thousand new neurons are born each day and are indispensable for odor learning and mate recognition (Oboti et al. 2011). New neurons in the hippocampus affect some, but not all, hippocampus-dependent learning and memory tasks (Deng et al. 2010). Interestingly, the maturation of newly generated hippocampal neurons in the adult follows a precise temporal sequence that recapitulates neonatal hippocampal development (Espósito et al. 2005).
Strikingly, the adult stem cells in the brain do not appear as simply undifferentiated cells, but express a marker thought to be specific for astrocytes (Doetsch 2003a). In the adult SVZ and SGZ reside cells with characteristics and markers of embryonic radial glia, absent in other CNS regions (Kriegstein and Alvarez-Buylla 2009). Thus, radial glia, thought to be a transient embryonic cell type that function as scaffold for neuronal migration, can function as stem cells in the adult brain. As mentioned, also in the brain multipotent stem cells reside and participate in specialized niches. As pointed out by Alvarez-Buylla and Lim (2004), brain niches have peculiar characteristics: ‘astrocytes serve as both neural stem cells and niche cell, a basal lamina vasculogenesis may be essential components of the niche, and “embryonic” molecular morphogens and signals, including Notch, Eph/ephrins, sonic hedgehog and bone morphogenetic proteins (BMPs), persist in these niches and play critical roles for adult neurogenesis’.
A new population of glial precursors, named NG2 glia, is present in the mammalian adult brain outside the niches. These cells seem to be able to generate NG2 glia, oligodendrocytes, astrocytes and scattered forebrain neurons (Rivers et al. 2008). In addition, a population of reactive astrocytes proliferates in the brain after injury. These cells share hallmarks of embryonic radial glia and neural precursors, are able to grow in vitro as neurospheres and show self-renewal and multipotency (Buffo et al. 2008). The major question is whether these ‘reactive astrocytes’ derive from the mature astrocytes (‘protoplasmic’) or the ependymal cells that are positioned along the ventricles, or both (Robel et al. 2011). These observations can be exploited to ‘awaken’ AdNSC to repair neuronal damage, as they would be present where they would be needed.
Fine tuning of adult neurogenesis
Activation of neurogenesis in the niches requires the intervention of a number of key regulators, which include signaling networks that involve small non-coding RNAs and microRNAs (Schouten et al. 2012), trophic factors as brain-derived neurotrophic factor (BDNF, Sairanen et al. 2005) or glial derived neurotrophic factor (GDNF, Kobayashi et al. 2006). These factors, produced and released by the niche cells, activate intracellular signaling pathways that induce proliferation in the adult lateral ventricle niches as well as axonal growth and synaptogenesis (Colucci-D'Amato et al. 2003; Nonaka 2009). BDNF favors adult SGZ neurogenesis but does not alter SVZ neurogenesis, although it is necessary for SVZ neuroblast migration along blood vessels in the RMS (Snapyan et al. 2009). Interestingly, oxygen is a key element in regulating NSC in niches (Panchision 2009; Mazumdar et al. 2010; De Filippis and Delia 2011). For example, low O2 increases the hypoxia-inducible factor 1alpha, which promotes self-renewal (by facilitation of the Notch pathway) and inhibits NSC differentiation or apoptosis (by blocking the bone morphogenetic protein pathways). Increased O2 tension has opposite effects. Similarly, perinatal hypoxia in mice expressing GFP under the control of the glia promoter GFAP show that higher number of astroglial cells attain a neuronal fate that lasted to adulthood (Salmaso et al. 2012). Thus, manipulation of oxygen contribution could be used to regulate endogenous NSC (Mazumdar et al. 2010).
It is now well established that adult neurogenesis can be increased by external stimuli, such as enriched environment or voluntary physical exercise (Fig. 2; Van Praag et al. 1999) as well as by dietary modulation (Park and Lee 2011). These effects are also regulated by trophic factors. For instance, environment-induced increased neurogenesis is altered in mice with reduced levels of BDNF (Rossi et al. 2006).
In addition, adult neurogenesis is under finely tuned control of neuronal activity mediated by various neurotransmitters and their receptors (Hagg 2009; Doze and Perez 2012). Accordingly, it has been suggested that targeting neurotransmitter receptors with selective agonists or antagonists could modulate adult neurogenesis. Most of the classic neurotransmitters (monoamines, acetylcholine, GABA, glutamate) regulate migration, maturation, integration and survival of newborn neurons. In the adult SVZ, GABA released from neuroblasts promotes their migration and inhibits precursor proliferation (Liu et al. 2005). Moreover, GABA released from hippocampal interneurons promotes cellular maturation, synapse formation and survival of newborn neurons (Tozuka et al. 2005; Ge et al. 2006) through CREB signaling (Jagasia et al. 2009). On the other hand, glutamate regulates survival of neuroblasts in the adult SVZ and of newborn neurons in the adult SGZ, through NMDA receptor signaling (Tashiro et al. 2006; Platel et al. 2010). Similarly to what happens during development, the new hippocampal neurons receive input from local GABAergic cells during the early stages of integration. This GABAergic activity is transiently excitatory and capable of depolarizing the new neurons (Ge et al. 2006). With maturation, the GABAergic activity becomes inhibitory while excitatory inputs derive from new connections with the entorhinal cortex. The survival of the newborn neurons in the dentate gyrus and olfactory bulbs seems to be activity dependent (Van Praag et al. 1999; Petreanu and Alvarez-Buylla 2002). Thus, early depolarization could be used through different means to stabilize the new wiring of the brain in therapeutical interventions aimed at stimulating endogenous neurogenesis. Recent studies indicate that dopamine is a potent regulator of proliferation of adult neural progenitors through the release of EGF in the SVZ (O'Keeffe et al. 2009; O'Keeffe and Barker 2011), exerting variable effects (Park and Enikolopov 2010) possibly because of the activation of distinct subtypes of dopamine receptors (Mu et al. 2011). Similarly, also serotonin increases adult neurogenesis through various 5-HT receptor subtypes (Hagg 2009).
Adult neurogenesis is not restricted to SVZ and SGZ
Whether adult neurogenesis is constrained only in the two canonical zones, SVZ and SGZ, or could occur also in other CNS areas is still an open question (Gould 2007). Until a few years ago, it was largely accepted that the cerebral cortex of normal adults is non-neurogenic (Kornack and Rakic 2001; Rakic 2004). Nevertheless, cortical neurogenesis has been shown in the adult macaque and was thought to derive from neuroblasts generated in the SVZ that migrated in three different regions of the adult neocortex (Gould et al. 1999). Dividing cells that differentiate into neurons have been found within the cortex in adult mice following lesion, using markers for DNA replication and progressive neuronal differentiation (Magavi et al. 2000; Gould 2007; Zhang et al. 2011). According to Magavi et al. (2000), the selective degeneration of layer VI cortical neurons resulted in a proliferating stimulus. The new neurons originated from precursors present in the cortex and from the nearby SVZ as blocking SVZ cell proliferation resulted in the partial reduction but not absence of newborn neurons in the lesioned area (Leker et al. 2007). A ‘local’ neurogenic potential has also been suggested in the rat visual cortex following lesion (Sirko et al. 2009).
Indeed, forebrain neurogenesis is very plastic during development and possibly in adulthood. By in vitro over-expression of a single gene encoding for a neurogenic transcription factor, neurogenin-2 (NGN2) or distal-less homeobox 2 (DLX2), it has been possible to generate glutamatergic or GABAergic neurons, respectively, from cortical mouse post-natal astroglia. These neurons were able to elicit axon potential and form synapses; however, it remains unclear whether they could integrate in functional circuits (Heinrich et al. 2011).
Similarly, the transcription factor FEZ family zinc finger 2 (FEZF2) was sufficient to make corticofugal glutamatergic neurons from GABAergic progenitors fated to become striatal medium spiny neurons in early embryonic mice (Rouaux and Arlotta 2010).
Taken together, these reports demonstrate that neural precursors can migrate to or be locally generated in non-canonical neurogenic regions and differentiate into region specific CNS neurons. They could be directed to differentiate ectopically by cell-autonomous signaling.
In summary, it is plausible that adult cortical neurogenesis occurs throughout life, although these events might be rather rare rendering thus difficult to monitor and demonstrate their incidence.
Besides to the well-characterized adult SVZ and SGZ neural progenitors, in the brain, as in other adult tissues, are present perivascular niches, which harbor cells that express mesenchymal cell markers. These cells, recently isolated also from the neocortex of the adult human brains, have a high density in the CNS and play a role in the establishment and maintenance of the blood–brain barrier (Armulik et al. 2010; Daneman et al. 2010). They are called pericytes, which show various functions that include immune and phagocytic, as well as contractile, hemostatic and angiogenetic activities and have multilineage potency toward either mesodermal or neuroectodermal phenotypes. Perycites can form neurospheres, can generate astrocytes, neurons or oligodendrocytes and have been shown to differentiate into neurons and glial cells after ischemic injury in the monkey hippocampus (Yamashima et al. 2004). As perycites reside in the walls of the microvasculature throughout the entire CNS, they may represent an additional reservoir that can be exploited to heal diseased brains and for these reasons have raised an enormous interest in their potential use in therapy (Paul et al. 2012).
It has recently been discovered that also meninges harbor potential niches that can be activated after injury. These cells are nestin- and doublecortin-positive, migrate into the spinal cord forming glial scars in vivo and can give rise to neurons and oligodendrocytes in vitro (Decimo et al. 2011).
Human adult neurogenesis
Most of our current knowledge on adult neurogenesis has emerged from basic research on laboratory animals (Eriksson et al. 1998). However, cells capable of long-term expansion and differentiation into neurons and glia have been derived from adult human brain, supporting the knowledge that neural precursors are present also in humans (Roy et al. 2000). Initial studies have been restricted to autoptic brain samples (Eriksson et al. 1998; Spalding et al. 2005), while recent advances in brain imaging (Couillard-Despres and Aigner 2011) are bursting our possibilities to monitor human adult neurogenesis. Converging evidence show that SGZ and SVZ are the active neurogenic zones in humans, as in laboratory animals (Eriksson et al. 1998; Curtis et al. 2007). However, a series of differences have emerged showing a lower progenitor proliferation rate in the SGZ and extensive rate in the SVZ when compared with rodents (Curtis et al. 2012). A developmental study of the human SVZ suggests that neurogenesis and neuronal migration extend into post-natal life, but is limited to early childhood. Surprisingly, during this window of neurogenesis, a major migratory pathway of SVZ new neurons targets the pre-frontal cortex, in addition to the olfactory bulb (Sanai et al. 2011). These results are consistent with previous evidence showing that cortical neurogenesis occurs only perinatally in the human brain (Bhardwaj et al. 2006). The latter observations have been questioned since most of the studies on cortical neuronal proliferation employing 14C as detector of DNA duplication, failed. It cannot be excluded that these data are false negative because of the limited sensitivity of the technique, especially if the numbers of the new neurons is very low (Gould 2007). Overall, although evidence suggests the existence of neurogenesis in the human neocortex, it remains still not adequately established.
Neurogenesis in pathological conditions
Enhanced neurogenesis has been shown in several pathological conditions. SVZ and SGZ adult neurogenesis has been described in epilepsy (Zhao and Overstreet-Wadiche 2008). However, the possible role of these new neurons in participating to the epileptogenic network or reestablish the normal homeostasis as well as their integration in the pre-existing circuits remains controversial (Kokaia 2011).
Other pathological conditions also seem to enhance adult neurogenesis. For example, inflammatory processes can modulate NSC activation as in the case of an experimental model of multiple sclerosis (MS) where, blocking the activity of microglial cells, determines a striking increase in the SVZ production of progenitor cells and oligodendrocytes accompanied by reduction of MS symptoms (Rasmussen et al. 2011).
Also in Alzheimer's disease (AD), characterized by extensive loss of neurons in the forebrain, it has been shown a positive effect on AdNSC, albeit glial over neuronal differentiation is promoted (Baratchi et al. 2012). In contrast to findings in AD animal models, where neurogenesis appears impaired, in post-mortem human AD brains there is increased hippocampal neurogenesis with increased expression of immature neuronal marker proteins, such as the microtubule-associated protein Doublecortin, the cell adhesion glycoprotein PSA-NCAM expressed by neuronal progenitors and TUC-4 (turned on after division), a protein expressed in axonal growth (Jin et al. 2004).
Similarly, stress and depression decrease neurogenesis while antidepressant treatment reverses these effects and induces hippocampal adult neurogenesis (Castrén and Rantamäki 2010; Danzer 2012). These effects are related to BDNF expression. Antidepressant-regulated SGZ neurogenesis is also associated to the expression of vascular endothelial growth factor, VEGF, and insulin-like growth factor 1, IGF1, (Warner-Schmidt and Duman 2007). Both VEGF and IGF1 administration increase SGZ neurogenesis and mimic the action of antidepressants.
As mentioned, CNS lesions, such as ischemia, also induce neurogenesis and, in most cases, migration of the new neurons from the neurogenic areas to the damaged sites (Arvidsson et al. 2002; Zhang et al. 2007). Thus, the adult brain has the competence for self-repair after insults. It remains questionable for how long the new neurons can survive ectopically (Arvidsson et al. 2002).
The examples highlighted above indicate that at least in theory it is possible to manipulate endogenous adult neurogenesis. Once understood how to drive AdNSC into an active state, several questions must be addressed before the endogenous adult neurogenesis could be conceived as a putative therapeutical approach. First, it is necessary to learn how to direct newly generated neuroblasts to the sought lesioned area, secondly how to prime undifferentiated neuroblasts toward specific neuronal fates, thirdly how to induce functional synaptogenesis and long-term survival.
We start to recognize some of the molecules involved in these steps. It remains known that normally part of SVZ and SGZ newborn adult neurons survive only few weeks and their survival is modulated by the animal experience and the establishment of functional active synaptic integration. New findings indicate that the latter depends also on the development of proper arborization, which is timely regulated among others factors by microRNAs. Indeed in olfactory bulb and hippocampus, the microRNA (miR) 132 plays such a role (Pathania et al. 2012; Schouten et al. 2012). Similarly, the brain-specific miR-9, abundant in neurogenic regions in embryos and adults, has been implicated in adult neural progenitor differentiation (Zhao et al. 2009). MiR-9 coordinates the proliferation and migration of neural precursors by blocking the mRNA encoding stathmin, a protein that increases microtubule instability: its loss suppresses proliferation but promotes migration in vitro and enhances migration of progenitor cells when transplanted into mouse embryonic or adult brains after stroke (Delaloy et al. 2010).
Combinatorial labeling with exogenous and endogenous markers has provided critical information about the morphological and functional development of progenitors in the adult brain, as well as regarding the identification of the players involved in SVZ neuroblast migration. Nevertheless, most factors guiding migration remain still elusive and it remains unknown whether factors allowing migration of SVZ or SGZ neuroblasts are the same in other areas of the adult brain (Leong and Turnley 2011). As pointed out, cytoskeletal rearrangement, cell adhesion molecules, growth factors, repulsive or attractive axon guidance molecules and regulatory cues are involved (Hagg 2005; Leong and Turnley 2011).
Caveats in neuronal replacement
A major question casts a shadow for the future of cell replacement therapies in neurodegeneration, that is, the diseased environment could have possible negative effects on the exogenous or endogenous new neurons that should replace the lost or diseased brain cells. In clinical transplants, indirect evidence indicates a long-term survival (nine to sixteen years) of fetal midbrain neurons grafted in the caudate/putamen of patients with Parkinson's disease. Nevertheless, in a number of these patients (three of six), at autopsy, grafted neurons appeared altered and showed Parkinson's pathogenic hallmark, namely alpha-synuclein inclusions (Lewy's bodies). Thus, disease factors in the host have affected transplanted nerve cells (Li et al. 2008). The spread of alpha-synuclein to the grafted exogenous neurons remains enigmatic. Prion-like mechanisms have been suggested as causative of the new pathology (Desplats et al. 2009), as well as inflammation, oxidative stress, reduction of neurotrophic support (Brundin et al. 2008), all causes invoked also for the original pathology. Similarly, embryonic striatal neurons, transplanted into two patients with Huntington's disease 10 years earlier, exhibit neurodegeneration as neurons of the patient. Thus, indicating that on the long-term the ‘affected’ brain environment may damage healthy neurons (Cicchetti et al. 2009). If this is possible for new neurons unrelated to the genetic background of the patient, it could be even more plausible for new neurons generated endogenously by the patients themselves.
However, it has to be bore in mind that in the cited cases of parkinsonian patients, only 5–8% of transplanted neurons developed Lewy's bodies. Moreover, the transplants survived for at least a decade. These relatively poor results must be a spur to improve studies on the mechanisms of protection necessary to create an environment that is anti-inflammatory, neuroprotective and neurotrophic, which is still needed, prior and/or during exogenous or endogenous neuronal replacement.
In conclusion, the discoveries reported in this review may pave the way for future therapeutic interventions that involve enhancing adult neural stem cell production and the functional incorporation of new neurons into affected neural circuits, provided that we can learn how to identify and solve the diseased environment negative influences. The biological rationale for the potential therapeutic use of endogenous stem cells is strengthened by their ability to migrate to sites of neuroinflammation or injury in the adult brain, different from those of normal neurogenesis. In addition, endogenous NSC could repair the brain and restore functions without problems of rejection, as in heterologous transplants (Fig. 2).
As also pointed out by others, the final teaching from the studies described above is that an integrated view of developmental and adult neurogenesis is necessary to manipulate adult brain plasticity.
We apologize to those authors whose excellent work in the field could not be cited here because of space limitations. We thank Dr. Davide Viggiano, University of Molise, for critical reading of the manuscript; FIRB-RBIN062YH4, MERIT-RBNE08LN4P-002, PRIN-2009TBCZJB_003, and Ministero della Salute Under-40 2007 for financial support. The authors have no conflicts of interest to declare.