Project Topics
Get Complete Project Material File(s) Now! »
Microglial precursors differentiation
CSF1R signaling is known as essential for microglial development and maintenance (Chitu et al. 2016; Chitu and Stanley 2017). Indeed, during embryogenesis, YS macrophages depend on CSF1R signaling for their proliferation, differentiation and survival. In CSF1R-mutant mice, YS macrophages are absent and the subsequent tissues colonization by macrophages fails to occur (Ginhoux et al. 2010). Similarly, transient inhibition of CSF1R signaling pathway with a blocking anti-CSF1R antibody efficiently eliminates fetal macrophages, including developing microglia (Hoeffel et al. 2015). CSF1R signaling may induce myeloid cell differentiation by activating the myeloid lineage-determining transcription factor PU.1, which is broadly expressed by hematopoietic cells. Indeed, PU.1-mutant mice have reduced expression of CSF1R (Rosenbauer et al. 2004), suggesting that there is a feedback loop between CSF1R signalling and PU.1 regulation. Interestingly, PU.1-deficient mice lack microglia and all non-parenchymal macrophages (Goldmann et al. 2016; Schulz et al. 2012).
As soon as the precursors emerge in the YS, once the myeloid lineage is determined by expression of PU.1 and CSF1R, microglial progenitors (EMPs) start to differentiate into YS macrophages. They first become CD45+ ckitlow CX3CR1- immature YS macrophages A1 and then become the more mature CD45+ ckit – CX3CR1+ mature YS macrophages A2 (Kierdorf et al. 2013). Of note, the transcription factor interferon regulatory factor 8 (IRF8) is required for maturation of the YS macrophage precursors Indeed, these precursors undergo apoptosis in Irf8−/− mice, resulting in markedly reduced numbers of embryonic and adult microglia (Kierdorf et al. 2013). At GD10.5, macrophages in the brain will have further matured in comparison to the YS as evidenced by down regulation of ckit in immature A1 and mature A2 microglia cells (Kierdorf et al. 2013). A transcriptomic study further dissected microglial development from GD10.5 onwards into three discrete steps based on differential genes regulation of microglia cells in the developing brain to acquire different functions (early-, pre- and adult microglia), with abrupt shifts occurring at approximately GD14.5 and PND14 (Bennett et al. 2016; Matcovitch-Natan et al. 2016) (Figure I.2).
Wiring the developing CNS
In the postnatal brain, synaptic connections are formed in excess and must be remodeled to obtain an organized and efficient synaptic network, characteristic of mature organisms. Developmental synapse elimination (synaptic pruning) is a critical part of this remodeling.
During development, neurons upregulate the expression of the chemokine fractalkine, CX3CL1 (Mody et al. 2001), and its receptor is exclusively expressed by microglia in the CNS (Harrison et al. 1998; Jung et al. 2000). This evidence led researchers to investigate the ability of microglia to prune synapses during development, to control synapse numbers. In 2011, a seminal study by Paolicelli and colleagues showed that microglia engulf synaptic material, thus play a major role in synaptic pruning during postnatal development (Paolicelli et al. 2011). Moreover, in mice lacking CX3CR1, microglia numbers were transiently reduced in the developing brain and synaptic pruning was delayed resulting in an excess of dendritic spines at PND15 (Paolicelli et al. 2011). Parallel studies on the postnatal retinogeniculate system identified the classic complement cascade as another underlying mechanism in the process of synaptic pruning (Schafer et al. 2012). Indeed, complement component 3 (C3) is present at synapses during the early postnatal period and is required for normal developmental remodelling of retinogeniculate axons (Stevens et al. 2007). Given that microglia are the only known resident brain cells expressing the C3 receptor (CR3), the authors showed that C3-CR3 interactions allow the recruitment of microglia to retinogeniculate axons (Figure I.4a & I.5).
Mounting evidence has now shown that, during development, microglia are actively involved in synapse remodeling and maturation (Hoshiko et al. 2012; Paolicelli et al. 2011; Schafer et al. 2012). In CX3CR1 mutant animals, the dendritic spine density of hippocampal CA1 neurons transiently increases at PND15. This phenotype is associated with a temporary reduction in microglial cell numbers and with an accumulation of immature synapses, leading to decreased functional connectivity across brain regions and an autism-like behavioural phenotype (Paolicelli et al. 2011; Zhan et al. 2014). Moreover, it has been shown that microglial signalling is required for synaptic plasticity during the critical period of visual development (Sipe et al. 2016; Tremblay, Lowery, and Majewska 2010) (Figure I.4b & I.5).
In addition, microglia are closely associated with white-matter tracts during development, where they help to control axon fasciculation in the dorsal corpus callosum at GD17.5 (Pont-Lezica et al. 2014), thus promoting axons that travel in the same direction to adhere together to form a tight bundle (Figure I.4c & I.5).
Microglia are also key players in the control of the wiring of forebrain circuits since they limit axon outgrowth. Microglial cells associated with axon fibers acquire a very distinct morphology and will line up parallel to the fibers, suggesting specific function of microglia in neurite development (Cuadros et al. 1993; Dalmau et al. 1997, 1998). This role of microglia is also supported by work performed with microglia-conditioned medium and primary neuronal cultures. Indeed, microglia-conditioned medium increases neurite growth and complexity (Chamak, Morandi, and Mallat 1994; Zhang and Fedoroff 1996). Until recently, there was no in vivo evidence of a definitive role for microglia in the various aspects of axonal growth and guidance. However, Squarzoni and colleagues carefully examined the positioning of microglial cells within the brain during early prenatal period (Squarzoni et al. 2014). They showed that at GD14.5, microglia transiently associate with the extremities of dopaminergic axons when they enter the subpallium, as opposed to adjacent serotoninergic or internal capsule fibers. Furthermore, microglia are largely excluded from the cortical plate until GD16.5, at which point the cortical plate is then progressively colonized (Squarzoni et al. 2014). Using multiple mouse models, including cell-depletion approaches and CX3CR1−/− or CR3−/− mutants, they found that perturbing microglial activity affects the outgrowth of dopaminergic axons in the forebrain and the positioning of subsets of neocortical interneurons (Squarzoni et al. 2014) (Figure I.4d & I.5).
RNA extraction and gene expression Reverse Transcription-PCR
After behavioural testing, animals were killed by terminal anesthesia [rodent cocktail (Ketamine 100mg/mL, Xylazine 20mg/mL and Aceprozamine 10mg/mL; 0.1mL/100g)]. The brains were quickly removed and hippocampi were dissected, frozen on powdered dry ice and kept at -80 C until analysis. Tissues were homogenized in TRIzol (Invitrogen) for RNA extraction according to the manufacturer’s instructions. RT reaction was performed as previously described (Maric, Woodside, and Luheshi 2014). Briefly, 1 μg of total RNA was transcribed by heating in a Gene Amp PCR System 9700 Thermocycler (Applied Biosystems) with 5 μL DEPC water, 1 μL random hexamers (Applied Bioscience), and 1 μL dNTPs (Sigma-Aldrich) at 65 °C for 10 min. cDNA was synthesized by incubating with primer mix [with 2 μL 0.1 dithiothreitol (Invitrogen), 1 μL murine myeloleukemia virus reverse transcriptase (Invitrogen), 4 μL first-strand buffer (Invitrogen) and 2 μL distilled water)] in the thermocycler at 37 °C for 1 h and then at 90 °C for 5 min in a total volume of 20 μL. All RNA samples were reverse transcribed simultaneously to minimize inter-assay variation. Following reverse transcription, the product was stored at -20 °C. Real-time RT-PCR was performed using Taqman gene expression assays for IDO (cat #Mm00492586_m1), Arg1 (cat #Mm00475988_m1), Ptgs2 (cat #Mm00478374_m1), IL-10 (cat #Mm00439614_m1), IL-1 (cat #Mm00446190_m1), TNF (cat #Mm00443258_m1) and 18S (cat #4352930E) as the endogenous control, purchased from Applied Biosystems. In a total volume of 20 μL, each sample-contained 2 μL of prepared cDNA described above, 1 μL primer/probe mixture (Applied Biosystems), and 10 μL taqMan PCR master mix (Applied Biosystems). cDNA was amplified in duplicates by quantitative real-time PCR (7500 Real-Time PCR System, Applied Biosystems). Relative quantitative measurement of target gene levels was performed using the Ct method, where Ct is the threshold concentration. Differences in expression were calculated using the formula fold change = 2(-CT) and expressed as a fold increase relative to the control condition.
Plant production strategies among pulse species
Collected values for the six agro-ecosystem properties revealed three main axes of variation. The first axis sorted pulses based on their biomass and grain production level. The second axis was mostly related to response to competition, especially against weeds, while the third axis sorted pulses based on their ability to fix nitrogen. Despite its wide variation among pulses, water use efficiency was poorly represented and occupied an intermediate position in this space of variation. These results might highlight segregation between species that are adapted to productive versus unproductive or competitive environments. This is consistent with Grime’s CSR triangle theory that predicts plant strategies are selected along environmental gradients of resource availability, stresses and perturbation(Grime et al., 1997a; Grime, 2009a). Grime’s theory highlighted that these gradients might have selected optimized trade-offs between phenology or development and resource acquisition in plants. This results in a CSR classification based on a combination of three extreme plant strategies: competitors (C), stress-tolerant (S), and ruderals (R) (Pierce et al., 2017). When applied to Mediterranean pasture legumes such as Medicago sp.(Yousfi et al., 2010), Trifolium sp.(Norman, Cocks and Galwey, 2002) or Lupinus sp. (Berger, Shrestha and Ludwig, 2017), this theory was in accordance with the variation of traits related to water acquisition and use, and highlighted contrasted plant adaptations to drought (Berger and Ludwig, 2014).
In the species set surveyed here, the third axis might partially represent species adaptation to stressful environments with low levels of nitrogen and water. The first axis of the PCA sorted the species on their productivity ability selected in potentially highly productive environments such as fertilized and irrigated crop systems associated with limited competition due to recurrent perturbations (mechanical or chemical destructions). This axis could be associated to R-strategies. However, productive species could be more or less sensitive to competition mostly depending on their ability to intercept light as suggested by the second axis of the PCA. This competitive ability might have been selected in highly productive and crowded environments and be related to C-strategy. However, agro-ecosystem are ruderal environments: pulses as crop species have a peculiar evolutionary history compared to uncultivated species due to their breeding background. They played an important role in human nutrition and thus might have been more often managed under non-limiting conditions (irrigation and fertilization) and would exhibit high grain production levels compared to species with less importance and that were predominantly used for feed. Thus, it is more likely that axes 1 and 2 show classification of competitive ability inside ruderal strategies. Finally, species that were considered in this study came from a wide range of geographic origins and have therefore been subject to a great diversity of climates, farming practices and resources along their history, leading to contrasted adaptations to stressful conditions, where nitrogen fixation may be determining in competitive capacity. In the current context of climate change and reduction of resource availability, species that are able to maintain relatively high productivity under stressful conditions could be valuable for the provision of food.
Patterns of trait covariations and plant production strategies
Our results showed trait covariation patterns consistent with those predicted and observed by common Plant Ecological Strategies Schemes (PESS) previously described in wild species (Grime et al., 1997a; Westoby, 1998; Saugier, 2001; Moles et al., 2002; Lavergne, Garnier and Debussche, 2003; Quétier et al., 2007). Leaf, seed or growth strategy are major components of CSR strategy variation. Unfortunately, we were unable to position pulse crops in the Grime’s CSR triangle (e.g. using the computation (Pierce et al., 2017) due to the lack of trait values.
We found that seed and leaf traits exhibited the greatest variability among the 43 pulse species, in accordance with Westoby’s LHS (leaf-height-seed) scheme (Westoby, 1998). The first PC axis was best explained by leaf size, leaf area and, especially, plant height, which is recognized as a good indicator of plant capacity for competitive dominance (Gaudet and Keddy, 1988; Westoby, 1998; Hodgson et al., 1999). Variation of seed size and TGW was mostly associated to the second axis of our PCA but had an intermediate position in the first PC plan and indicated that the seed dimension axis defined by Westoby (1998), as an expression of the potential of dispersal and successful establishment of a species (Cipollini and Stiles, 1991; Willson, 1993; Askew et al., 1997), was marginally correlated to plant and leaf size dimensions in the studied pulses; this result was consistent with previous findings by Leishman (Leishman et al., 1985). The third dimension of the LHS scheme relates to the capacity of plants to exploit resource-rich and resource-poor environments. This dimension reflects the trade-off between “fast” traits that promote rapid resource acquisition and “slow” traits that promote resource conservation (Wright et al., 2004a; Díaz et al., 2016; Volaire, 2018). In the search for a single trait that captures the core of this axis, specific leaf area (SLA) is a leading contender (Poorter and Remkes, 1990; Lambers and Poorter, 1992; Poorter et al., 2009). However, our study shows that this trait contributed only marginally to the total trait variation among pulses through its contribution to the third PC axis. Part of pulse variability represented on the third PC axis was best represented by days to maturity and flowering, which can also be negatively related to plant capacity to rapidly acquire resources.
As previously observed by Tribouillois (Tribouillois, Cruz, et al., 2015) for herbaceous Fabaceae crops, the relationship between SLA and LNC was similar to that observed across a large diversity of wild species (Wright et al., 2004a; Díaz et al., 2016) although LNC values of N-fixer species were on average higher than non N-fixer species (see Supplementary figure S2). The total percentage of variability explained by the first three PCs (61%) was relatively low compared to what is usually observed in wild and cultivated species(Wright et al., 2004a; Díaz et al., 2016). This could be attributed to the collection method of data and/or the peculiarities of these cultivated Fabaceae species.
From trait profiles to services
Functional traits are directly or indirectly linked to ecological processes leading to agro-ecosystem properties. Several traits can be involved in one process and one trait can participate in several processes. Thus, a given trait can only predict ecosystem properties value as long as every other trait have fixed values and processes occur at a fixed rate. CART regressions were particularly relevant to overcome this issue. For example, our results show that high WUE efficiency could be achieved by plants with LNC over 50 mg g-1 or, alternatively, with low LNC if total leaf area was above 500 cm2. Similarly, late maturing species matched high yield, but early maturing species were still able to achieve GY up to 1.6 t ha-1 if they had high TSW (>= 164 g) and LNC above 40 mg g-1. In addition, plant traits which are determinant for good performance in one property could be disadvantageous for another when they are combined with other traits. For example, long crop duration is in favor of GY and %Ndfa but could end up to 80% YR under weed infestation among species with large leaflets. It is therefore difficult to define one ideal combination of plant functional traits which would maximize all ecosystem properties. TSW participated in the prediction of most agro-ecosystem properties. However, we found that prediction of ecosystem properties did not rely on the central traits used for the quantification of each axis alone. Leaflet length was found more relevant than plant height to predict competitive ability against weeds and days to maturity a better predictor than SLA for prediction of pulse productivity as indicated by CART results for prediction of biomass and grain yield performances.
The first targeted ecosystem service was food production under dry conditions. The objective was not to differentiate pulse species by their drought resistance but rather by their ability to maintain high production rate under drought, which is a more relevant trait for food production. WUE is a good candidate for that purpose even though it is not the only driver of effective use of available water or drought resistance. Over the 28 species considered here, high WUE was observed mainly for species with high LNC. LNC at anthesis and grain yield under drought have been previously found to be strongly linked (Borrell and Hammer, 2000). Indeed, when nutrient uptake is limited by water availability (Chapin, 1993), remobilization of N from vegetative tissues becomes particularly important for grain growth (Ta and Weiland, 2010). Moreover, high LNC is usually associated to the “stay green” type. More especially in case of terminal drought, it has been shown that if LNC declines to a critical threshold, leaf senescence will set up (Borrell and Hammer, 2000; Borrell, Hammer and Van Oosterom, 2001). It is therefore not surprising that N leaf status is very closely related to the longevity of photosynthetic organs (Kikuzawa, 1991; Reich, Walters and Ellsworth, 1997). In case of terminal drought, high LNC maintains photosynthetic capacity for longer, sometimes leading to higher grain yield, and allows greater N remobilization. Furthermore, species with low LNC were more likely to have high WUE if associated with large leaf area. Although large leaf area may cost more water loss, it is also possible that it would result in more remaining photosynthesis area at the end of the drought stress period, thus allowing to achieve higher yields. WUE was poorly represented in the PCA. Because of this positioning, it could hardly be associated with other properties. Understanding its interaction with grain yield could have been particularly interesting for breeding purposes. Grain yield was not predicted with the same set of traits than WUE except for LNC (LNC is important for pod filling whatever water conditions are). Seed weight (TSW) was a good predictor of grain yield. Indeed, seed size is expected to be positively correlated with seedling biomass (Tamet et al., 1996; Fayaud et al., 2014), plant height and reproductive effort (Chapin, Autumn and Pugnaire, 2002). Seed size might also have been a result of agronomic selection such as erect habit, characteristic of most high yielding species and a trait highly related to domestication (Smartt, 1978a).
Cropping systems that incorporate grain legumes have been shown to strongly decrease N fertilizer rates (by 13–30% for wheat and 49–61% at the rotation level) through nitrogen fixation(Plaza-Bonilla et al., 2017). However, legume species are not equivalent in their ability to supply exogenous nitrogen to the system. In our study, the percentage of nitrogen fixed varied substantially across species (from 23.75 ± 15.34 % to 93.33 ± 3.51 %). In addition to species inherent capacity to fix N, this variation could be attributed to nutritional factors, environmental conditions, rhizobia strains or host characteristics (Reichardt et al., 1987; Nambiar, Rupela and Kumar Rao, 1988; Hardarson et al., 1993). Since the data collected focused on plant traits, an incomplete picture of what drives nitrogen supply might have been obtained. CART regression showed that high nitrogen supply was mostly achieved by late maturing species. This result is in accordance with previous studies (Kumar Rao and Dart, 1987; Piha and Munns, 1987). Nitrogen fixation is more favorable for grain yield during the latter part of the growth cycle of a legume than it is during early growth (Piha and Munns, 1987; Vance, 2011). These factors account for the superior symbiotic performance of late maturing bean cultivars. In addition, any process that increases growth rate also increases tissue turn-over and loss of carbon, nutrients and water, alongside with decreasing allocation to storage (Grime et al., 1997a) and, thus, possible allocation to nodules. Our results highlighted small leaflets as a secondary trait involved in high nitrogen supply. The presence of compound leaves is a widespread trait of legume species and leaves divided into small leaflets appears to be a frequent component of ecological strategies emphasizing a productive photosynthetic apparatus (Leavitt, Dobrenz and Stone, 1979; McKey, 1994a). In theory, nitrogen fixation may require/mobilize 10 to 20% of the total plant photosynthesis(Schubert, 1981; Vance, 2011). Thus, high photosynthetic productivity might allow a greater allocation of photosynthetic compounds to nodules. N2 fixation has been shown to be closely synchronized with the rate of supply of translocate from the shoot to nodules (Tissue, Megonigal and Thomas, 1997; Herridge and Pate, 2008). Here we considered the proportion of nitrogen fixed by a plant (%Ndfa), independently of biomass production aspect, as an indicator of nitrogen supply. This trait is related to plant N fixation efficiency andits ability to grow in nitrogen-poor environments. However, biomass yield is hardly dissociable from nitrogen fixation as it is a process driven by N demand (Hartwig, 1998; Hartwig et al., 2002). CART regression showed that high biomass species had large SLA. High SLA might allow (given favorable growth conditions) a shorter payback time on a gram of dry matter invested in a leaf (Poorter, 1994), therefore improving production rate.
Table of contents :
CHAPTER I: COMPREHENSIVE REVIEW OF THE LITERATURE
I.1. The unique embryonic origin of microglia
I.1.1. Ontogeny of microglia
I.1.2. Development of microglia
I.2. The role of microglia in CNS development
I.2.1. Patterning the developing CNS
I.2.2. Wiring the developing CNS
I.3. The role of microglia in the mature CNS
I.3.1. Regulating mature CNS circuits
I.3.2. Surveilling the CNS
I.3.3. Mediating CNS immune response
I.4. Heterogeneity of microglia
I.4.1. Heterogeneity of microglia throughout the brain
I.4.2. Heterogeneity of microglia across sex
I.5. Maternal Immune Activation & neurodevelopmental disorders
I.6. Maternal Immune Activation & microglia
I.6.1. Effects of Maternal Immune Activation on embryonic microglia
I.6.2. Effects of Maternal Immune Activation on postnatal microglia
I.7. Hypothesis
I.8. Objectives
C. Lacabanne – PhD thesis Table of contents
CHAPTER II: HIPPOCAMPAL MICROGLIA DEPLETION IMPROVES LPS-INDUCED DEPRESSIVE-LIKE BEHAVIOUR AND IDO EXPRESSION
II.1. Manuscript
Abstract
Introduction
Materials and Methods
Results
Discussion
Acknowledgement
Figures
Reference List
CHAPTER III: EARLY MATERNAL IMMUNE ACTIVATION ALTERS MICROGLIA DEVELOPMENTAL TRAJECTORY AND INDUCES LONG-TERM BEHAVIOURAL ALTERATIONS IN THE OFFSPRING IN A SEXUALLY DIMORPHIC MANNER
III.1. Preface
III.2. Manuscript
Abstract
Introduction
Materials and Methods
Results
Discussion
Acknowledgements
Figures
Reference List
CHAPTER IV: INTERLEUKIN-1 MEDIATES THE EFFECTS OF PRENATAL LPS INDUCED BEHAVIOURAL ANOMALIES IN THE OFFSPRING THROUGH MODULATING MICROGLIA
IV.1. Preface
IV.2. Manuscript
Abstract
Introduction
Materials and Methods
Results
C. Lacabanne – PhD thesis Table of contents
Discussion
Acknowledgements
Figures and Tables
Reference List
GENERAL DISCUSSION AND CONCLUSION
BIBLIOGRAPHY
