Patch-based 3D U-Net and transfer learning for longitudinal piglet brain segmentation on MRI.

BACKGROUND AND OBJECTIVES: In order to study neural plasticity in immature brain following early brain lesion, large animal model are needed. Because of its morphological similarities with the human developmental brain, piglet is a suitable but little used one. Its study from Magnetic Resonance Imaging (MRI) requires the development of automatic algorithms for the segmentation of the different structures and tissues. A crucial preliminary step consists in automatically segmenting the brain. METHODS: We propose a fully automatic brain segmentation method applied to piglets by combining a 3D patch-based U-Net and a post-processing pipeline for spatial regularization and elimination of false positives. Our approach also integrates a transfer-learning strategy for managing an automated longitudinal monitoring evaluated for four developmental stages (2, 6, 10 and 18 weeks), facing the issue of MRI changes resulting from the rapid brain development. It is compared to a 2D approach and the Brain Extraction Tool (BET) as well as techniques adapted to other animals (rodents, macaques). The influence of training patches size and distribution is studied as well as the benefits of spatial regularization. RESULTS: Results show that our approach is efficient in terms of average Dice score (0.952) and Hausdorff distance (8.51), outperforming the use of a 2D U-Net (Dice: 0.919, Hausdorff distance: 11.06) and BET (Dice: 0.764, Hausdorff distance: 25.91). The transfer-learning strategy achieves a good performance on older piglets (Dice of 0.934 at 6 weeks, 0.956 at 10 weeks and 0.958 at 18 weeks) compared to a standard training strategy with few data (Dice of 0.636 at 6 weeks, 0.907 at 10 weeks, not calculable at 18 weeks because of too few training piglets). CONCLUSIONS: In conclusion, we provide a method for longitudinal MRI piglet brain segmentation based on 3D U-Net and transfer learning which can be used for future morphometric studies and applied to other animals.

A novel ex vivo Huntington's disease model for studying GABAergic neurons and cell grafts by laser microdissection.

Organotypic brain slice cultures have been recently used to study neurodegenerative disorders such as Parkinson's disease and Huntington's disease (HD). They preserve brain three-dimensional architecture, synaptic connectivity and brain cells microenvironment. Here, we developed an innovative model of Huntington's disease from coronal rat brain slices, that include all the areas involved in the pathology. HD-like neurodegeneration was obtained in only one week, in a single step, during organotypic slice preparation, without the use of neurotoxins. HD-like histopathology was analysed and after one week, a reduction of 40% of medium spiny neurons was observed. To analyse new therapeutic approaches in this innovative HD model, we developed a novel protocol of laser microdissection to isolate and analyse by RT-qPCR, grafted cells as well as surrounding tissue of fresh organotypic slices. We determined that laser microdissection could be performed on a 400µm organotypic slice after alcohol dehydration protocol, allowing the analysis of mRNA expression in the rat tissue as well as in grafted cells. In conclusion, we developed a new approach for modeling Huntington's disease ex vivo, and provided a useful innovative method for screening new potential therapies for neurodegenerative diseases especially when associated with laser microdissection.

Nanoprecipitated catestatin released from pharmacologically active microcarriers (PAMs) exerts pro-survival effects on MSC.

Catestatin (CST), a fragment of Chromogranin-A, exerts angiogenic, arteriogenic, vasculogenic and cardioprotective effects. CST is a very promising agent for revascularization purposes, in "NOOPTION" patients. However, peptides have a very short half-life after administration and must be conveniently protected. Fibronectin-coated pharmacologically active microcarriers (FN-PAM), are biodegradable and biocompatible polymeric microspheres that can convey mesenchymal stem cell (MSCs) and therapeutic proteins delivered in a prolonged manner. In this study, we first evaluated whether a small peptide such as CST could be nanoprecipitated and incorporated within FN-PAMs. Subsequently, whether CST may be released in a prolonged manner by functionalized FN-PAMs (FN-PAM-CST). Finally, we assessed the effect of CST released by FN-PAM-CST on the survival of MSCs under stress conditions of hypoxia-reoxygenation. An experimental design, modifying three key parameters (ionic strength, mixing and centrifugation time) of protein nanoprecipitation, was used to define the optimum condition for CST. An optimal nanoprecipitation yield of 76% was obtained allowing encapsulation of solid CST within FN-PAM-CST, which released CST in a prolonged manner. In vitro, MSCs adhered to FN-PAMs, and the controlled release of CST from FN-PAM-CST greatly limited hypoxic MSC-death and enhanced MSC-survival in post-hypoxic environment. These results suggest that FN-PAM-CST are promising tools for cell-therapy.

Characterization and comparison of two novel nanosystems associated with siRNA for cellular therapy.

To direct stem cell fate, a delicate control of gene expression through small interference RNA (siRNA) is emerging as a new and safe promising strategy. In this way, the expression of proteins hindering neuronal commitment may be transiently inhibited thus driving differentiation. Mesenchymal stem cells (MSC), which secrete tissue repair factors, possess immunomodulatory properties and may differentiate towards the neuronal lineage, are a promising cell source for cell therapy studies in the central nervous system. To better drive their neuronal commitment the repressor Element-1 silencing transcription (REST) factor, may be inhibited by siRNA technology. The design of novel nanoparticles (NP) capable of safely delivering nucleic acids is crucial in order to successfully develop this strategy. In this study we developed and characterized two different siRNA NP. On one hand, sorbitan monooleate (Span(®)80) based NP incorporating the cationic components poly-l-arginine or cationized pullulan, thus allowing the association of siRNA were designed. These NP presented a small size (205 nm) and a negative surface charge (-38 mV). On the other hand, lipid nanocapsules (LNC) associating polymers with lipids and allowing encapsulation of siRNA complexed with lipoplexes were also developed. Their size was of 82 nm with a positive surface charge of +7 mV. Both NP could be frozen with appropriate cryoprotectors. Cytotoxicity and transfection efficiency at different siRNA doses were monitored by evaluating REST expression. An inhibition of around 60% of REST expression was observed with both NP when associating 250 ng/mL of siRNA-REST, as recommended for commercial reagents. Span NP were less toxic for human MSCs than LNCs, but although both NP showed a similar inhibition of REST over time and the induction of neuronal commitment, LNC-siREST induced a higher expression of neuronal markers. Therefore, two different tailored siRNA NP offering great potential for human stem cell differentiation have been developed, encouraging the pursuit of further in vitro and in vivo in studies.

Modeling nigrostriatal degeneration in organotypic cultures, a new ex vivo model of Parkinson's disease.

Parkinson's disease (PD) is the second most frequent neurodegenerative disorder afflicting 2% of the population older than 65 years worldwide. Recently, brain organotypic slices have been used to model neurodegenerative disorders, including PD. They conserve brain three-dimensional architecture, synaptic connectivity and its microenvironment. This model has allowed researchers a simple and rapid method to observe cellular interactions and mechanisms. In the present study, we developed an organotypic PD model from rat brains that includes all the areas involved in the nigrostriatal pathway in a single slice preparation, without using neurotoxins to induce the dopaminergic lesion. The mechanical transection of the nigrostriatal pathway obtained during slice preparation induced PD-like histopathology. Progressive nigrostriatal degeneration was monitored combining innovative approaches, such as diffusion tensor magnetic resonance imaging (DT-RMI) to follow fiber degeneration and mass spectrometry to quantify striatal dopamine content, together with bright-field and fluorescence microscopy imaging. A substantia nigra dopaminergic cell number decrease was observed by immunohistochemistry against rat tyrosine hydroxylase (TH) reaching 80% after 2 days in culture associated with a 30% decrease of striatal TH-positive fiber density, a 15% loss of striatal dopamine content quantified by mass spectrometry and a 70% reduction of nigrostriatal fiber fractional anisotropy quantified by DT-RMI. In addition, a significant decline of medium spiny neuron density was observed from days 7 to 16. These sagittal organotypic slices could be used to study the early stage of PD, namely dopaminergic degeneration, and the late stage of the pathology with dopaminergic and GABAergic neuron loss. This novel model might improve the understanding of PD and may represent a promising tool to refine the evaluation of new therapeutic approaches.

Effect of various additives and polymers on lysozyme release from PLGA microspheres prepared by an s/o/w emulsion technique.

Incomplete protein release from PLGA-based microspheres due to protein interactions with the polymer is one of the main issues in the development of PLGA protein-loaded microspheres. In this study, a two-dimensional adsorption model was designed to rapidly assess the anti-adsorption effect of formulation components (additives, additives blended with the polymer or modified polymers). Lysozyme was chosen as a model protein because of its strong, non-specific adsorption on the PLGA surface. This study showed that PEGs, poloxamer 188 and BSA totally inhibited protein adsorption onto the PLGA37.5/25 layer. Similarly, it was emphasised that more hydrophilic polymers were less prone to protein adsorption. The correlation between this model and the in vitro release profile was made by microencapsulating lysozyme with a low loading in the presence of these excipients by a non-denaturing s/o/w encapsulation technique. The precipitation of lysozyme with the amphiphilic poloxamer 188 prior to encapsulation exhibited continuous release of active lysozyme over 3 weeks without any burst effect. To promote lysozyme release in the latter stage of release, a PLGA-PEG-PLGA tribloc copolymer was used; lysozyme was continuously released over 45 days in a biologically active form.

Effective GDNF brain delivery using microspheres--a promising strategy for Parkinson's disease.

Glial cell line-derived neurotrophic factor (GDNF) has shown promise in the treatment of neurodegenerative disorders of basal ganglia origin such us Parkinson's disease (PD). In this study, we investigated the neurorestorative effect of controlled GDNF delivery using biodegradable microspheres in an animal model with partial dopaminergic lesion. Microspheres were loaded with N-glycosylated recombinant GDNF and prepared using the Total Recirculation One-Machine System (TROMS). GDNF-loaded microparticles were unilaterally injected into the rat striatum by stereotaxic surgery two weeks after a unilateral partial 6-OHDA nigrostriatal lesion. Animals were tested for amphetamine-induced rotational asymmetry at different times and were sacrificed two months after microsphere implantation for immunohistochemical analysis. The putative presence of serum IgG antibodies against rat glycosylated GDNF was analyzed for addressing safety issues. The results demonstrated that GDNF-loaded microspheres, improved the rotational behavior induced by amphetamine of the GDNF-treated animals together with an increase in the density of TH positive fibers at the striatal level. The developed GDNF-loaded microparticles proved to be suitable to release biologically active GDNF over up to 5 weeks in vivo. Furthermore, none of the animals developed antibodies against GDNF demonstrating the safety of glycosylated GDNF use.

Combining polymeric devices and stem cells for the treatment of neurological disorders: a promising therapeutic approach.

Cell therapy will probably become a major therapeutic strategy for neuronal disorders in the coming years. Nevertheless, due to poor survival of grafted cells and limited differentiation and integration in the host tissue, certain ameliorations must be envisaged. To address these difficulties, several strategies have been developed and among them, two methods seem particularly promising : in situ controlled drug delivery and implantation of cells adhered on biomaterial-based scaffolds. Indeed, the ability of drugs, such as growth factors, to regulate neuronal survival and/or plasticity infers the use of these molecules to treat neurodegeneration associated with human diseases. Moreover, the synthesis of cell scaffolds which mimic the extra-cellular matrix can help guide morphogenesis and tissue repair. Furthermore, cells can be cultivated on these matrices that may eventually make graft therapy a more practical approach for the treatment of neurological diseases. Nevertheless, for those two encouraging approaches multiple parameters have to be considered, such as the drug targeting strategy, but also the physical and morphological characteristics of the scaffold and the type of cells to be conveyed. This review thus focuses on those two promising strategies and also on their possible association to improve stem cell therapy of neurodegenerative disorders. Indeed, tissue replacement by grafting cells within or adhered onto drug delivering biomaterial-based devices, has recently been reported and seems to be very promising.

Pharmacologically active microcarriers: a tool for cell therapy.

To overcome certain problems encountered in cell therapy, particularly cell survival, lack of cell differentiation and integration in the host tissue, we developed pharmacologically active microcarriers (PAM). These biodegradable particles made with poly(D,L-lactic-co-glycolic acid) (PLGA) and coated with adhesion molecules may serve as a support for cell culture and may be used as cell carriers presenting a controlled delivery of active protein. They can thus support the survival and differentiation of the transported cells as well as their microenvironment. To develop this tool, nerve growth factor (NGF)-releasing PAM, conveying PC12 cells, were produced and characterized. Indeed, these cells have the ability to differentiate into sympathetic-like neurons after adhering to a substrate, in the presence of NGF, and can then release large amounts of dopamine. Certain parameters such as the size of the microcarriers, the conditions enabling the coating of the microparticles and the subsequent adhesion of cells were thus studied to produce optimized PAM.

In vivo evaluation of pharmacologically active microcarriers releasing nerve growth factor and conveying PC12 cells.

Cell therapy will probably become a major therapeutic strategy in the coming years. Nevertheless, few cells survive transplantation when employed as a treatment for neuronal disorders. To address this problem, we have developed a new tool, the pharmacologically active microcarriers (PAM). PAM are biocompatible and biodegradable microparticles coated with cell adhesion molecules, conveying cells on their surface and presenting a controlled delivery of growth factor. Thus, the combined effect of growth factor and coating influences the transported cells by promoting their survival and differentiation and favoring their integration in the host tissue after their complete degradation. Furthermore, the released factor may also influence the microenvironment. In this study, we evaluated their efficacy using nerve growth factor (NGF)-releasing PAM and PC12 cells, in a Parkinson's disease paradigm. After implantation of NGF-releasing or unloaded PAM conveying PC12 cells, or PC12 cells alone, we studied cell survival, differentiation, and apoptosis, as well as behavior of the treated rats. We observed that the NGF-releasing PAM coated with two synthetic peptides (poly-D-lysine and fibronectin-like) induced PC12 cell differentiation and reduced cell death and proliferation. Moreover, the animals receiving this implant presented an improved amphetamine-induced rotational behavior. These findings indicate that PAM could be a promising strategy for cell therapy of neurological diseases and could be employed in other situations with fetal cell transplants or with stem cells.

Biocompatibility of implantable synthetic polymeric drug carriers: focus on brain biocompatibility.

Numerous polymeric biomaterials are implanted each year in human bodies. Among them, drug delivery devices are potent novel powerful therapeutics for diseases which lack efficient treatments. Controlled release systems are in direct and sustained contact with the tissues, and some of them degrade in situ. Thus, both the material itself and its degradation products must be devoid of toxicity. The knowledge and understanding of the criteria and mechanisms determining the biocompatibility of biomaterials are therefore of great importance. The classical tissue response to a foreign material leads to the encapsulation of the implant, which may impair the drug diffusion in the surrounding tissue and/or cause implant failure. This tissue response depends on different factors, especially on the implantation site. Indeed, several organs possess a particular immunological status, which may reduce the inflammatory and immune reactions. Among them, the central nervous system is of particular interest, since many pathologies still need curative treatments. This review describes the classical foreign body reaction and exposes the particularities of the central nervous system response. The recent in vivo biocompatibility studies of implanted synthetic polymeric drug carriers are summarized in order to illustrate the behavior of different classes of polymers and the methodologies used to evaluate their tolerance.

Intraseptal implantation of NGF-releasing microspheres promote the survival of axotomized cholinergic neurons.

Neurotrophic factors therapy requires their precise delivery to the targeted neuronal population. For this purpose, a wide range of strategies have been developed, and among them the stereotaxic implantation of biodegradable microparticles. To assess the in vivo activity of NGF-releasing PLGA microspheres, unloaded and NGF-loaded microparticles were implanted in the rat brain, near the septal cholinergic neurons, axotomized by an unilateral transection of the fornix-fimbria. Histological analysis at two and six weeks after implantation revealed a non-specific astro- and micro-glial reaction around the microspheres, identical for both unloaded and NGF-loaded microspheres. No neuronal toxicity was noticed, and healthy looking neurons were observed in contact with the microspheres. In the non-treated animals, the percentage of axotomized surviving neurons, when compared to the contralateral intact side, was 31 +/- 2 and 27 +/- 1% at two and six weeks, respectively. Unloaded microspheres caused no protective nor neurotoxic effects (40 +/- 9 and 39 +/- 6% of surviving cholinergic neurons at two and six weeks, respectively). In contrast, NGF-loaded microspheres showed a significant effect on the survival of axotomized cholinergic neurons at two and six weeks after implantation (66 +/- 9 and 61 +/- 5% when compared to the contralateral intact side, respectively). These results show that PLGA microparticles present no neurotoxicity and release sufficient amounts of bioactive NGF to significantly limit the lesion-induced disappearance of cholinergic neurons in the septum during at least six weeks. PLGA microparticles can be used in the future to administer neurotrophic factors in central nervous system disorders.

Low affinity NGF receptor expression in the central nervous system during experimental allergic encephalomyelitis.

In the central nervous system (CNS), p75, or low-affinity nerve growth factor receptor (LNGFR), is assumed to play a critical role in mediating the effects of neurotrophins on neuronal survival. Recent studies have shown that nerve growth factor (NGF) can act also on immune cells through its binding to p75. Using immunohistochemistry, we have investigated the expression of the p75 receptor in the CNS during chronic relapsing experimental allergic encephalomyelitis (EAE) of the Lewis rat, an animal model of multiple sclerosis (MS). We report here a sequential expression of p75, first in Purkinje cells during the first attack, and secondly on both endothelial and perivascular cells in the latter stages of the disease. Moreover, starting from the second attack, p75 was also expressed on glial ensheathing cells, likely myelinating cells, located primarily in the dorsal roots. These data suggest that during EAE, LNGFR may play an important role in leukocyte-endothelial cell interactions and in the maintenance of Purkinje cells survival.

Early events of the inflammatory reaction induced in rat brain by lipopolysaccharide intracerebral injection: relative contribution of peripheral monocytes and activated microglia.

We have previously demonstrated that lipopolysaccharide (LPS) intracerebral injection induced only minimal inflammatory reaction in rat brain, apart from an increased number of 'brain macrophages' observed 24 h after LPS administration [Montero-Menei et al., Brain Res., 653 (1994) 101-111]. However, the nature of these 'brain macrophages' in the inflammatory response is still unclear. The present study focused on the early time-points (from 5 h to 24 h) after LPS injection or stab-lesion, and was aimed at the identification of the peripheral (monocytes) or parenchymal (microglia) origin of these 'brain macrophages'. OX42- and ED1-labeling did not clearly discriminate between monocytes/macrophages and reactive microglia, both cell types being immunoreactive. In other experiments, rats were made aplasic by irradiation prior to lesioning. These experiments clearly demonstrated that LPS induces an intense monocyte recruitment and, to a lesser extent, microglial activation since about 80% of the cells present 24 h after LPS injection consisted of recruited monocytes not observed in aplasic rats. Interestingly, our data show that LPS exerts a sequential dual action by first inhibiting the monocyte recruitment observed 5 h after stab lesion and then enhancing it at 15 h and 24 h after injection. A possible involvement of cytokines, chemokines and adhesion molecules in the mechanisms occurring in the early events of brain inflammatory reaction is discussed.

Lipopolysaccharide intracerebral administration induces minimal inflammatory reaction in rat brain.

An inflammatory reaction, essential for defence against infection and for wound repair, may also induce irreversible tissue damage. It appears that the central nervous system has developed its own immunosuppressive strategy in order to limit the destructive effects of inflammation. To clarify this point, we have characterized in one unique model of inflammation induced in the rat by intracerebral lipopolysaccharide injection the kinetics of the inflammatory reaction, the participation of immunitary and glial cells and of three growth factors. Among these molecules, brain-derived neurotrophic factor mRNA expression was found decreased following LPS injection. No striking differences were observed in the brain parenchyma after stab lesion or inflammatory lesion apart from an increase in the number of monocytes/macrophages recruited early to the lesion area. Macrophages were later accumulated around the lesion when astroglia and microglia reactions occurred. Some of the macrophages and microglia expressed major histocompatibility complex class II antigens on their surface whereas no T or B lymphocytes were observed in the brain parenchyma. However, a subpopulation of CD3- and CD4-negative CD8-positive cells, likely natural killer cells, was observed around the lesion site; this recruitment was inhibited by the highest dose of LPS. This study therefore supports the hypothesis of a suppression of some aspects of cell-mediated immunity in the brain, mechanisms which need to be further characterized.

Pure Schwann cell suspension grafts promote regeneration of the lesioned septo-hippocampal cholinergic pathway.

Regeneration of central nervous system (CNS) axons has been studied in the cholinergic septo-hippocampal system using various 'bridges' able to support fiber growth. In this study, a pure Schwann cell (Sc) suspension labeled with bisbenzimide (Hoechst 33342) was grafted in the lesioned septo-hippocampal pathway. At 2 weeks post-grafting, acetyl-cholinesterase (AChE)-positive fibers invaded the graft and grew in association with the Hoechst-labeled Sc, some of which expressed the low-affinity nerve growth factor receptor (NGF-R). At 2 months and 4 months post-grafting, the dorsal hippocampus was reinnervated with an apparently normal innervation pattern. Analysis of fiber growth in the hippocampus at four months post-grafting revealed a significant increase of reinnervation in the grafted animals (2 mm) compared to the non-grafted ones. No difference was observed in the number of cholinergic septal neurons expressing the NGF-R. These results demonstrate that a Sc suspension grafted into the lesioned septo-hippocampal system, integrates well into the host tissue, and supports axonal CNS outgrowth, implying that Sc by themselves provide an adequate environment for regeneration to occur.