- Top of page
- Materials and methods
- Supporting Information
SPARC (secreted protein acidic and rich in cysteine) is a matricellular protein highly expressed during development, reorganization and tissue repair. In the central nervous system, glial cells express SPARC during development and in neurogenic regions of the adult brain. Astrocytes control the glutamate receptor levels in the developing hippocampus through SPARC secretion. To further characterize the role of SPARC in the brain, we analyzed the hippocampal-dependent adult behavior of SPARC KO mice. We found that SPARC KO mice show increased levels of anxiety-related behaviors and reduced levels of depression-related behaviors. The antidepressant-like phenotype could be rescued by adenoviral vector-mediated expression of SPARC in the adult hippocampus, but anxiety-related behavior persisted in these mice. To identify the cellular mechanisms underlying these behavioral alterations, we analyzed neuronal activity and neurogenesis in the dentate gyrus (DG). SPARC KO mice have increased levels of neuronal activity, evidenced as more neurons that express c-Fos after a footshock. SPARC also affects cell proliferation in the subgranular zone of the DG, although it does not affect maturation and survival of new neurons. SPARC expression in the adult DG does not revert the proliferation phenotype in KO mice, but our results suggest a role of SPARC in limiting the survival of new neurons in the DG. This work suggests that SPARC could affect anxiety-related behavior by modulating neuronal activity, and that depression-related behavior is dependent upon the adult expression of SPARC, which affects adult brain function by mechanisms that need to be elucidated.
SPARC (secreted protein acidic and rich in cysteine, also known as osteonectin) is a matricellular glycoprotein involved in tumor progression and wound healing. SPARC is highly expressed in different tissues during embryogenesis, but in the adult its expression is restricted to tissues undergoing remodeling, repair and tumorigenesis (Bradshaw & Sage 2001, Sage et al. 1989). In the brain, it is expressed both during development and in adulthood. In the embryonic mouse brain, SPARC is expressed by the radial glia in regions undergoing neurogenesis (Vincent et al. 2008). In the postnatal brain, SPARC is still expressed by the radial glia, but it is also present in the rostral migratory stream, surrounding proliferating cells.
In the adult mouse brain, glial cells express SPARC in different regions of the central nervous system, especially in neurogenic regions such as the subventricular zone and the subgranular zone of the dentate gyrus (SGZ-DG) (Vincent et al. 2008). It was also detected in the molecular layers of the CA1 and the cerebellum, linking its expression to regions enriched in synapses (Mendis et al. 1995). Moreover, SPARC expression is upregulated in the hippocampus upon entorhinal deafferentation (Liu et al. 2005) and after epileptic episodes (Ozbas-Gerceker et al. 2006); further supporting that SPARC expression is linked to regions of increased proliferation, axonal sprouting and synaptogenesis.
Recently, it was shown that postnatal astrocytes secrete SPARC and that this protein is essential to the regulation of AMPA receptors at maturing synapses in the hippocampus (Jones et al. 2011). Neurons grown in the presence of SPARC KO astrocytes exhibit increased synaptic strength and a diminished NMDAR/AMPAR ratio, which reduces the capacity of neurons to develop long-term potentiation. Moreover, another recent work shows that SPARC is a negative regulator of synapse formation (Kucukdereli et al. 2011)
We reasoned that the pattern of expression of SPARC and the postnatal role of this protein in determining neuronal activity could reflect a role of this protein in regulating hippocampal-dependent behavior in the postnatal and/or the adult mouse. The aims of this work were then: (1) to determine whether lack of SPARC could result in hippocampal-dependent behavioral abnormalities, (2) to determine the role of adult hippocampal SPARC in these behavioral phenotypes and (3) to identify possible molecular and cellular mechanisms that mediate this effect. The hippocampus is involved both in cognitive function and in mood regulation, modulating anxiety states (Deacon et al. 2002) and depression (Mcewen & Magarinos 2001). We found that SPARC KO mice exhibit anxiety-related behavior and antidepressant-like behavior, and that these mice show a diminished response to spatial novelty. Moreover, we found evidence of increased neuronal activity and reduced cell proliferation in the adult DG of SPARC KO mice. We next found that the re-expression of SPARC in the adult DG of the hippocampus reverts the antidepressant-like behavior in SPARC KO mice.
- Top of page
- Materials and methods
- Supporting Information
We show here for the first time a role of SPARC in anxiety- and depression-related behaviors, and we show that hippocampal SPARC is determining basal levels of behavioral despair in the mouse. Moreover, we show that the increased neuronal activity that was previously reported in the SPARC KO postnatal hippocampus (Jones et al. 2011) is still observable in the adult hippocampus. Finally, we show that SPARC KO and Het mice have less proliferating cells in the SGZ-DG, but that this does not result in reduced neurogenesis.
SPARC KO mice showed increased anxiety-related behavior in the OF and LD tests, but not in the EPM test. Others and we have previously shown that different behavioral tests can reveal diverse coping strategies in mice (Carola et al. 2006, Lucchina et al. 2010) and that, in particular, behavior in the OF does not correlate with EPM in the C57BL/6J strain (Rogers et al. 1999). This study adds to the previous literature and stresses the importance of using multiple behavioral tests to characterize the effects of a particular treatment. Alternatively, time of testing could be affecting our results, preventing us from detecting differences in behavior in the EPM. For practical reasons, we performed all our behavioral experiments during the light phase, i.e. the resting phase of rodents. Previous reports have shown that behavior in the EPM changes along the day (Andrade et al. 2003), and that testing during the dark phase can improve discrimination between experimental groups (Hossain et al. 2004, Huynh et al. 2011). Further studies will help us to determine whether behavior in the EPM is affected in SPARC KO mice in a phase-dependent manner.
SPARC KO mice showed reduced behavioral despair both in the FS and TS tests, a behavior sensitive to antidepressant drugs (Lucki et al. 2001). Whether other symptoms of depression such as anhedonia are also modulated by hippocampal SPARC needs to be further studied.
Considering the high comorbidity between major depression and generalized anxiety disorder (Moller 2002), and the evidence suggesting that these disorders share genetic risk factors (reviewed by Tanti & Belzung 2010), the behavioral phenotype of SPARC KO mice may be unexpected. However, other mouse models have shown these opposing effects on behavior. Heat shock factor 1 KO mice show reduced anxiety-like behavior and increased depression-related behavior (Uchida et al. 2011), while a phenotype of increased anxiety-related behavior and antidepressant-like behavior was found in serotonin-1A receptor (5HT1AR) KO mice (Heisler et al. 1998, Ramboz et al. 1998), GABA(B) KO mice (Mombereau et al. 2004) and corticotropin-releasing factor (CRF)-overexpressing mice (Van Gaalen et al. 2002). Because both 5HT1AR and GABA(B) are localized both pre- and post-synaptically, it has been suggested that different receptor populations could contribute to the expression of anxiety- or depression-related behaviors.
Characterization of the GABAergic or the serotoninergic systems in the SPARC KO hippocampus is still needed. Preliminary evidence from our group shows that mRNA levels of 5HT1AR in SPARC KO mice are half those of WT mice (WT 100 ± 7.37%; KO 46.96 ± 0.28%; t6 = 7.19; P = 0.0004), suggesting that serotonin transmission is indeed affected in SPARC KO mice. In addition, previous work has shown that SPARC levels can determine glutamatergic synaptic strength in the postnatal hippocampus (Jones et al. 2011). We were not able to detect basal neuronal activity in the adult hippocampus by c-Fos immunohistochemistry. This was probably due to the antibody that we used and the protocol that we followed, as other reports have shown basal c-Fos immunoreactivity in this region. Because of this, we could not determine whether the increase in synaptic strength that is observed early in life persists in adult SPARC KO mice. However, our c-Fos analysis is highly reliable for addressing differences in neuronal activation upon different stimuli such as a footshock. We show, indeed, that a mild footshock results in few cells expressing c-Fos in the DG of WT mice. Interestingly, the number of neurons expressing c-Fos upon the footshock protocol is highly increased in SPARC KO animals, supporting the hypothesis that the KO hippocampus has increased neuronal activity also in adulthood.
Our results also show that the antidepressant-like phenotype can be rescued by SPARC re-expression in the adult DG, while anxiety-related behavior in KO mice is not affected by this treatment. This result is also in agreement with a possible mechanism involving 5HT1AR, as it was shown that adult expression of 5HT1AR in the adult forebrain cannot rescue the anxiety-related phenotype observed in 5HT1AR KO mice (Gross et al. 2002). However, we cannot rule out that we failed to rescue this behavior because adult expression was restricted to the DG, and other structures could actually be regulating this behavior through SPARC (e.g. amygdala or cortex). Therefore, we need further evidence to corroborate this hypothesis and to prove that SPARC KO phenotype is dependent upon the downregulation of serotonin transmission in the postnatal hippocampus.
As an alternative cellular mechanism through which SPARC could modulate mood states, we focused on adult neurogenesis. The pattern of expression of SPARC in the normal adult brain – being it particularly high in neurogenic regions such as the SVZ and the SGZ of the DG (Vincent et al. 2008) – pointed us toward studying this plasticity phenomenon. Moreover, a high number of evidence shows that SPARC inhibits cell proliferation in different cell types, and that it also inhibits the proliferative effect of different growth factors (Brekken & Sage 2000). Our behavioral data further pushed us in that direction. The role of neurogenesis in modulating affective states is today a matter of debate. Initially, hippocampal neurogenesis has been linked to depression (Duman 2004). In particular, it was shown that antidepressant treatment increases neurogenesis in the adult hippocampus (Malberg et al. 2000), but also that neurogenesis is necessary for these drugs to have their effects (Santarelli et al. 2003). Therefore, it has been expected that low levels of neurogenesis would be correlated with depression. But studies in which neurogenesis was blocked actually linked neurogenesis to anxiety (reviewed by Petrik et al. 2012): reducing neurogenesis resulted in increased levels of anxiety-like behavior in mice, while depression-related behaviors were not altered (Revest et al. 2009).
We found that SPARC KO and Het mice have reduced numbers of proliferating cells in the SGZ-DG. However, this reduction did not result in diminished levels of neurogenesis, as the density of new young neurons (DCX-positive) and adult neurons (CB-positive) was similar in all genotypes. Interestingly, our results show that the new cells generated in KO and Het mice are more prone to survive for 1 month in the DG than those proliferating in the WT brain. This could be due to either a compensatory effect elicited by the low levels of proliferation observed in KO and Het mice, or to a direct role of SPARC in limiting the number of cells that persist in the DG. In summary, we show that SPARC has a role in promoting cell proliferation in the DG, but we cannot establish whether SPARC also determines the percentage of these cells that will survive or whether there are other mechanisms in the DG that guarantee a certain number of new neurons and that counteract the effect of reduced proliferation in SPARC KO and Het mice. These alternatives need to be further studied.
Adult neurogenesis in the SPARC KO mice was not altered upon SPARC re-expression in the adult hippocampus with adenoviral vectors, a treatment that reverses the antidepressant-like behavior but does not affect the anxiety-related behavior. SPARC KO mice injected with AdSPARC or Adβgal showed similar levels of proliferating cells in the DG, and a similar number of these cells survived after 28 days. Moreover, the proportion of young and adult neurons was similar in both groups. This suggests that the depression-related behavior, i.e. the behavioral despair, is not modulated by neurogenesis in the adult DG. Neurogenesis could underlie the anxiety-related phenotype, as both proliferation and behavior are not rescued upon adult SPARC re-expression. However further studies are needed to confirm this link.
It is also worth to notice that the injection of adenoviral vectors into the DG increased the number of BrdU-positive cells observed 28 days after labeling (Fig. 5e vs. Fig. 7e; t-test with Welch's correction, P = 0.044). We have previously shown that the injection of adenoviral vectors into the DG results in the activation of astrocytes and microglial cells that lasts at least 28 days (Depino et al. 2011). The results presented here suggest that such activation is not relevant in the short term (KO mice in Fig. 5b show similar numbers of BrdU-positive cells when compare with KO mice injected with Adbgal, Fig. 7b) but that this treatment results in more BrdU-positive cells later (Fig. 7e). As the number of new neurons (BrdU + CB-positive cells) is not altered by these treatments, we conclude that SPARC could have a role in the survival of glial cells. Such a role of SPARC on glial cells activation and function needs to be further studied.
In summary, our results prove a role of SPARC in hippocampal function, and suggest that the correct expression of this matricellular protein is essential for the correct establishment of adult behavior and neuronal function. Further studies would contribute to elucidate whether this protein is involved in human disease, e.g. in anxiety and depression, besides the long known role of this protein in cancer and the more recent link of this protein with obesity.