Skip to main content

European stem cell consortium for neural cell replacement, reprogramming and functional brain repair

Final Report Summary - NEUROSTEMCELLREPAIR (European stem cell consortium for neural cell replacement, reprogramming and functional brain repair)

Executive Summary:
In the context of Neurostemcellrepair (www.neurostemcellrepair.org) we have worked in order to improve the quality and specification of transplantable mesencephalic dopaminergic (mesDA) cell preparations obtained from human embryonic stem cells (hESC) and long term neuroepithelial stem (Lt-NES) cells. By combining different approaches (modulation of Wnt signaling, exposure to laminins, monitoring of transcription factors expression etc.) we were able to improve mesDA specification, survival and differentiation. We also described an alternative strategy for Parkinson’s disease (PD) treatment, in which dopaminergic (DA) neurons are generated by direct reprogramming of mouse and human astrocytes into induced DA (iDA) neurons in vitro and in vivo using three transcription factors. We further analysed the molecular diversity of endogenous and human pluripotent stem cell (hPSC)-derived mesDA cells, by performing single-cell RNA-sequencing to examine ventral midbrain development in human and mouse. Cell types “Rgl1” and “Rgl3” were identified as two main pillars of mesDA neurogenesis in vivo, suggesting that recapitulating mesDA neuron development requires focussing on the generation of these cell types.
In order to be clinically relevant, hESC-derived DA neurons should be able to integrate and function upon transplantation into the adult brain. With trans-synaptic rabies-based tracing technology, we have shown that functional connectivity and circuit integration is established between grafted hESC-derived DAergic neurons that are stably maintained for at least 6 months, and host neural circuits within a rat model of PD. This work represents a significant milestone in the pre-clinical development of human DA neurons derived from hESC.
We reported further advance in the direct conversion of human fibroblasts into mature and functional neurons, termed induced neurons (iNs) by developing a new protocol that allows for highly efficient conversion of adult fibroblasts into iNs. Initial studies using this method have demonstrated that these iNs are functional and they also can integrate in the host circuitry.
We started to target the issue of final composition of cell transplants and to compare hESC and fetal-derived grafts for their immunoreactivity towards antibodies, which is crucial to refine future quality control strategies for cell therapy products entering the clinic for treatment of PD. We also implemented approaches to standardize the cell production procedures and assure that the cell product is not contaminated with unwanted and potentially harmful cell types.

We have successfully developed and optimized a reliable, reproducible and robust in vitro protocol for differentiation of hESC into striatal projection medium spiny neurons (MSNs). At the present time, this protocol is our best candidate for future migration to GMP-compliant production and represents the starting point for the development of experimental transplantation in Huntington’s disease (HD). Striatal progenitors generated using this approach and transplanted in HD animal models were able to survive, mature and integrate in the host tissue. Upon their transplantation, behavioural analysis showed partial recovery of lesion-dependent motor phenotypes as early as one month after transplant. Furthermore, we have identified that the GMP hESC lines (RC9 and RC17) are able to reach full maturation in a proportion of cell similar to that seen in validated lines (H9). To the aim of selecting better in vitro substrate for differentiation, we identified extra cellular matrix (ECM) molecules that offer potential for modulation of the innate and adaptive immune response as a possible therapeutic strategy. To overcome the limited axonal outgrowth and integration of donor cells our preliminary results show that integrin expression is correlated with the regenerative potential of axons.
In order to increase the PSA-NCAM levels in the transplant region we have shown that overexpression of PST leads to enhanced neurite outgrowth in lt-NES derived cells. We have also successfully applied a MACS sorting strategy to enrich for PSA-NCAM expressing progenitors. Finally, we sought to identify new inflammation-related genetic variants contributing to the disease process and clinical expression in HD in order to refine our understanding of the neurodegeneration processes and thus of the host environment, for future development of better strategies for graft integration. In addition, we have undertaken a tissue microarray (TMA) high-throughput study to look at the changes in protein expression of the inflammatory markers identified with the genetic analysis as well as standard inflammatory markers in the striatum of HD patients.

In order to optimise the survival, integration and functionality of novel cell therapies for both PD and HD we successfully defined key aspects for cell replacement approaches that include: the optimal cell placement and potency of current cell preparations, the establishment of behavioural tasks to reveal changes in relevant domains, the use of targeted training protocols to enhance recovery and the development of novel imaging technologies to visualise the integration and connectivity of neurons. We have collected a large data set addressing each of these themes and have successfully achieved our aims of demonstrating the survival, integration, safety and functional efficacy of hESC-derived DA grafts for PD. Previous clinical trials using human ventral midbrain (hVM) tissue have demonstrated proof-of-principle that DA grafts can alleviate motor dysfunctions long-term in patients, but much of the recent preclinical work has aimed to optimise the treatment to avoid side-effects and to develop novel stem cell derived cell therapies that are suitable for widespread clinical application. Here we have reached consistency in the survival and functionality of DA neurons based cell therapy products and have laid the foundation for imminent implementation of this therapy to patients with PD. For HD, progress has been made in the development of new protocols to generate authentic MSNs and optimisation of the transplantation of these cells has been undertaken. Improvements in phenotype and survival have been observed, together with promising signs of disease alleviation detected during striatum-dependent motor tasks in live animals. Some baseline criteria for the expected functionality of the grafts has been established using human fetal WGE grafts. Further work needs to undertake the final stage of demonstrating the functional efficacy of these cell therapy products, before consideration of the clinical application of these cells.

Significant advances have been made in the translational aspects and good manufacturing practice (GMP) compliance of our project as a consequence of a successful collaboration between academia and the manufacturing sites.
After protocol refinement, three successful batches of mesDA progenitor cells (cell therapy product - CTP) were manufactured using a standardised differentiation protocol and GMP risk assessed reagents. This protocol was performed concurrently with optimisation of cryopreservation and thaw strategies and with standardised quality control (QC) sampling points for in-process controls and final product characterisation. Manufacturing in accordance with validated protocols will ensure robust and reliable processing and streamline the route to the clinic. A QC strategy to characterise the CTP throughout the differentiation process (in-process controls) and to define the final CTP for release has been formulated for the mesDA differentiation protocol. Using panels of protein markers for hESC pluripotency and mesDA differentiation, we have designed a robust flow cytometry assay that can reliably determine the identity and purity of the CTP at different time points throughout the differentiation process. In addition, this assay is highly sensitive in the detection of contaminating pluripotent cells within the final CTP.
We performed the mesDA neuronal progenitor differentiation procedure also within the CliniMacs prodigy automated, closed system to allow integrated cell processing from starting material (hESC) to final CTP. Although purification was successfully performed, further investigation is required to evaluate the fully enclosed differentiation process and to investigate mesDA CTP more fully.
In parallel with MesDA differentiation GMP protocol development, we conducted process gap and risk analysis and develop a target product profile (TPP) to allow for the optimisation of GMP compatibility. The TPP defines critical areas of the development process including non-clinical and clinical development strategy, regulatory compliance, manufacturing process description, final formulation and administration, QC regime (Intermediates and In-Process Control), final product characterisation profile, final product safety testing requirements, raw materials and consumables and banking of RC-17 hESC line for use within Neurostemcellrepair Project.
The harvested CTP is cryopreserved and a number of tests demonstrated the robustness of the cryoprotocol and cryopreservation strategy.Conditions at which the CTP can be held until transplantation have been identified.

Altogether this project has highlighted solutions to several obstacles that were holding back the stem cells field from clinical application in neurodegenerative diseases and contributed to the identification and definition of a set of standard procedures and donor DA neurons preparations which can be safely and effectively applied in PD patients. Neurostemcellrepair has also contributed to the genesis and development of a worldwide effort named “G-force PD” (http://www.gforce-pd.com) an initiative that brings together major academic networks in Europe, the US, and Japan that share common therapeutic ambitions regarding human pluripotent stem cell-derived DA neurons for PD, with the goal to discuss how to use stem cell derived DA neurons in first-in-human clinical and highlight key clinical translation considerations.
Project Context and Objectives:
Neurostemcellrepair has been working (2013-17) as a world-leading Consortium, established to take stem cells through the final pre-clinical steps necessary before their clinical use in trials for Parkinson’s Disease (PD), as well as to achieve substantial advancements in their adoption as cell replacement strategies for Huntington’s Disease (HD). These two disorders are at different stages in their clinical translation for stem cell treatment, yet they have common paths and requirements that need to be fulfilled. Addressing one condition benefits the other and is also useful for establishing a platform by which we can treat a wider spectrum of neurodegenerative disorders. PD was taken as the prototypical disease because it has been successfully treated with dopamine cell therapies and, in addition, earlier developments had provided cells and protocols ready for optimization for GMP compatibility and translation to the clinic. At the same time, we further advanced the field by implementing emerging stem cell programming tools and methods for enhanced tissue integration of grafted cells in order to develop new approaches, validated at pre-clinical stages, for the treatment of PD and HD.

Experience gained from clinical trials using grafts of fetal mesencephalic dopaminergic (DA) and striatal GABAergic progenitors have shown that effective repair can be achieved by neural transplantation. Notably, transplanted DA neurons, derived from the ventral mesencephalon (VM), can functionally reinnervate the denervated striatum, restore dopamine release and, at least in some PD patients, induce substantial long-term clinical improvement (Politis, Sci Transl Med, 2010; Barker, Lancet Neurology, 2013). Similarly, fetal striatal tissue grafted into mild HD patients has, in some cases, produced long-term amelioration of motor, behavioural and cognitive dysfunction (Reuter, J Neurol Neurosur Psychiatry, 2008). On the basis of these observations a new clinical trial was initiated (Transeuro, http://www.transeuro.org.uk) aimed at developing efficacious and safe conditions for fetal VM-based therapies for PD. However, to move to large-scale applications, readily available, renewable, and bankable cells are needed.

Discoveries stemming from members of this Consortium had identified morphogens and transcription factors (TF) critical for successful phenotype specification and differentiation of therapeutically relevant DA neurons from human pluripotent stem cells. These findings led to cell differentiation protocols that currently represent the gold standard for cell therapy with ventral mesencephalic (mes) DA neurons (Kriks, Nature, 2011; Kirkeby, Cell Reports, 2012) and to novel ontogenetic factors that promote the dopaminergic phenotype (Theofilopoulos, Nat Chem Biol, 2012; Andersson, PNAS, 2013). Inspired by this work, a protocol to generate authentic striatal GABAergic medium spiny neurons (MSNs) was also published by other members of this Consortium (Delli Carri, Development, 2013). Furthermore, partners in this Consortium have identified a scalable source of neural progenitors able to mature into functional neurons (Koch, PNAS, 2009). With the advent of induced pluripotent stem (iPS) cell technologies and direct reprogramming of somatic cells, further sources of transplantable cells have become available. One partner in this Consortium was one of the first groups to report on the reprogramming of fibroblasts into DA neurons (Pfisterer, PNAS, 2011), while work by another partner identified small molecules that increase efficiency of neuronal conversion (Ladewig, Nat Methods, 2012). Hence, this Consortium brought together several groups that are leaders in the world and are capable of moving this field forward, towards a safe clinical translation of stem cell-based interventions, especially for PD.

In particular, the project aimed at fulfilling the final requirements for progressing stem cells towards clinical translation for dopamine cell replacement in PD, while addressing crucial issues on the way to future applications also in HD. Thus, our main goal has been to optimise and standardise the current cell-generating protocols and incorporate tools, methods and technologies to ensure efficient, accurate and safe integration and functional efficacy of human stem cell-derived neurons, to promote brain repair and functional recovery in PD and HD.

Therefore, Neurostemcellrepair has identified seven key areas of implementation, where further knowledge and technological innovation were required to allow for the move towards clinical application. First, All activities in the project were focussed on cells of human origin only, as these are be the cells that are going to be used for clinical applications. Secondly, we addressed the issues of cell survival, proliferation and differentiation after transplantation in animal models through the iterative refinement of protocols to enhance quality, efficiency and safety of the cell product. Thirdly, the biological potency of the selected cell preparations was scored functionally in short- and long-term animal experiments, against the in vivo behaviour of the gold standard preparation, i.e. human fetal derived mesDA and MSN progenitors. Fourth, to improve adaptive cell replacement and reconnection of neural circuits, novel approaches were used to enhance the integration of transplanted cells into the recipient tissue and the pattern of donor-host interactions was evaluated through a full repertoire of anatomical, functional and behavioural studies in animals. Fifth, a key area of implementation related to the fulfilment of regulatory standards that was addressed through manufacturing and banking of GMP-compatible DA neurons. Finally, for relevant technological developments, we incorporated strategies for cell purification, imaging of functional connectivity patterns of engrafted cells through development of high-speed 3D light sheet microscopy methods and the transfer of a DA differentiation protocol to an automated closed system instrument.

To address these points, Neurostemcellrepair brought together 12 teams: 8 European research groups, 1 industrial partner and 3 SMEs from 4 European countries. These teams represent the wide range of competences necessary to tackle the task, including human stem cell specialists, developmental neurobiologists, experts in animal models of neurodegenerative diseases as well as scientists with strong links to the clinic and to stem cell manufacturing. Neurostemcellrepair represented an additional step towards clinical development of the solid work carried out within the framework of the FP7 funded Project Neurostemcell (http://www.neurostemcell.org) which, between 2011 and 2013 had lead to the identification of the best protocols for the generation of authentic and transplantable mesDA progenitors and neurons from hES cells (Kriks, Nature, 2011; Kirkeby, Cell Reports, 2012) and also of authentic GABAergic MSNs (Delli Carri, Development, 2013).
Project Results:
The S & T results/foregrounds have been detailed in the final deliverables of each workpackage (D.1.10; D2.17; D3.12;D4.8; D5.12) which also contain additional information such as figures and tables, references, and a list of the connected publications. The following text is an excerpt of the mentioned documents.

WP1 - IDENTIFICATION OF OPTIMAL PROTOCOLS AND PROCEDURES FOR THE EFFICIENT GENERATION OF AUTHENTIC FUNCTIONAL MESDA NEURONS AND MSNS TO BE USED FOR CELL TRANSPLANTATION AND GMP PRODUCTION

1.1 Development of novel cell platforms and tools for human MesDA A9 and MSN DARP32+ neuron differentiation and reprogramming.
An important effort in this project has been placed to improve and develop novel approaches to control gene expression during neural differentiation. These tools include (1) non-integrating self-regulated viral vectors, (2) small molecules, (3) synthetic RNAs (sRNAs) and (4) long non-coding RNAs (LncRNAs). We have also examined established a new cellular platform designated as human induded neural progenitor cells (iNPCs; 5) from cord-bllod cells that could be used as novel source for human neurons.

(1) Lentiviral vectors are integrating retro-elements that have the capacity to affect gene expression levels of endogenous genes that can lead to transformation events. These vectors are often used in combination with doxycycline-regulated systems that are not optimal for clinical use since they contain elements of bacterial origin. We now developed a vector system that circumvent these issues and allow for generation of mature and functional iN cells using non-integrating, self-regulating lentiviral vectors (MP). These modifications make the system suitable for clinical translation and therefore represent a major step forward in the development of induced neurons for clinical applications. These vectors consist on the incorporation of 4 complementary binding sites of miR-124 in expression vectors containing transcription factors. Since miR-124 is expressed exclusively in neurons, the expression of transcription factors is not inhibited or degraded resulting in high levels of expression. When a stable neuronal fate is reached endogenous miR-124 is expressed, binds to the miR-target sequence in the expression vector and the expression of transcription factors is inhibited. Thus, the system is self-regulating alleviating the need for supply of drugs or chemicals for gene regulation.
(2) Small molecules were examined for their capacity to improve direct neuronal conversion (MP). Six out of 300 compounds increased the generation of neurons in a dose- dependent manner. These compounds were GSK3b inhibitors, cAMP/PKA, Adenyl cyclase activators, HDAC inhibitor and acted on Src, respectively. Moreover, when combined these six compounds gave a neuronal purity of 50% and a conversion efficiency of almost 500%.
(3) Current methods for transient or persistent overexpression of genes are mainly based on transfer of DNA and/or the use of viral vectors. DNA and viral vectors have their caveats as the delivery of DNA is limited by the defence mechanisms of cells and the use of viral vectors is potentially genotoxic. In our project, we have explored the versatility of synthetic RNA and antisense-oligonucleotides (ASO) as a mean to regulate gene expression, which might bear less risks for clinical use. While the efficiency of ASO was poor, we found that sRNAs allow the conversion of fibroblasts or the differentiation of pluripotent stem cells into mesDA or MSN neurons.
(4) We also examined lncRNAs (EC), which are involved in a broad range of important cellular processes, including chromatin modification, RNA processing, and regulation of gene transcription through interaction with DNA and proteins. Notably, lncRNAs take upstream control of gene expression programs, while miRNAs act downstream. In this project, we identified human lncRNAs that will be used to guide the differentiation of human ES cells and the reprogramming of somatic cells into MSN neurons.
(5) Finally, we also explored the possibility of using neonatal human cord blood (CB)-derived CD34+ cells available in cell banks to generate expandable transgene-free human induced neural progenitor cells (hiNPCs) that could be used to generate haplotype-matched mesDA or MSN neurons (OB). Upon transduction of CD34+ cells with non-integrating Sendai viruses coding for SOX2 and c-MYC we were able to obtain primary hiNPC colonies within 20 days. After serial passaging clonal hiNPC lines were found to be free of viral replicons and to express typical neural progenitor markers such as PAX6, SOX2, NESTIN and PLZF. hiNPCs could be stably expanded for more than 30 passages and were able to differentiate into neuronal and glial subtypes. Neurons derived from these hiNPCs were able to fire action potentials upon current injection and exhibited spontaneous postsynaptic currents indicating formation of neuronal circuits.
Results in this section open the door to the development of novel applications of readily available banked cord cells and to the application of novel tools to different protocols for the improved generation of MSN and MesDA neurons.

1.2 Generation of stable human mesDA A9 subtype neurons.
Pioneering work by the groups of Anders Björklund and Olle Lindvall as well as clinical transplantation studies have provided proof of concept that cell replacement therapy with human fetal ventral midbrain (fVM) tissue (containing mesDA precursors) can change the course of Parkinson’s disease (PD) (reviewed in Lindvall et al., 1988). In few cases, the therapeutic benefit of fVM tissue transplantation has been extraordinary and patients have been able to withdraw their L-DOPA medication and remained asymptomatic for up to 15 years (Kefalopoulou et al., 2014). However, ethical issues associated to the use of fetal tissue and difficulties in obtaining sufficient tissue for transplantation (6-7 embryos are needed to treat one patient), and standardizing its quality have led to the search for alternative cell sources such as human pluripotent stem cells (hPSCs). Both embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) are currently considered as promising sources for cell replacement therapy. With the advent of stem cell technologies and the growing knowledge of midbrain DA neuron development, it has become apparent that developmentally-based strategies can harness all the potential of stem cells for cell replacement therapy (Arenas et al., 2015). Several groups have recently developed protocols for the induction of mesDA neurons from hPSCs that upon transplantation in 6-hydroxydopamine (6-OHDA)-lesioned hemi-parkinsonian mice and rats improved their motor behavior without tumor formation (Kriks et al., 2011; Kirkeby et al., 2012; Doi et al., 2014). Moreover, when compared to human fVM tissue, these preparations have been recently found to be equipotent (same activity per transplanted number of cells) (Grealish et al., 2014). This has been a major achievement in the field and has triggered the preparation of new clinical trials using hPSC-derived mesDA cell preparations (Barker et al., 2015). However, despite the success achieved so far, high quality hPSC-derived cell preparations enriched in only mesDA neurons of the subtype predominantly affected in PD, A9/substantia nigra pars compacta (SNc) neurons, have not yet been achieved. The objective of this section has been to explore developmental signaling molecules and transcription factors that could bias the differentiation of hPSCs or neural stem cells towards A9/SNc neurons and allow the development of cell differentiation protocols to be used in cell replacement strategies for PD.
Our work has focus on two aspects: (1) Examine the possible application of midbrain developmental factors to promote the differentiation of human ES and long term neuroepithelial stem (LtNES) cells into mesDA neurons of the A9/substantia nigra subtype; (2) Further develop existing protocols for the differentiation of human ES into mesDA neurons to improve the yield of A9/SNc neurons; (3) Determine the quality of hPSC-derived mesDA cells as compared by single cell RNA-sequencing to endogenous human midbrain tissue.
(1) Exploring midbrain developmental factors. Here we explored the function and possible application of multiple candidate molecules to improve protocols for the differentiation of human ES and long term neuroepithelial stem (LtNES) cells into mesDA neurons. These include transcription factors, microRNAs, morphogens, ligands of the nuclear receptor, as well as extracellular matrix proteins. This work has been structured and performed in the context of four distinct work sets aiming at:
i) Modulating the levels of Otx2 and other transcription factors in order to increase A9 vs A10 ratio
In hPS cells (MP).
ii) Develop an application of LXR ligands to promote mesDA differentiation of hPS and LtNES (EA).
iii) Use of Wnt signaling tools, such as GSK3β inhibitors, spondins and Wnt5a to increase A9 vs A10 ratio in LtNES/hPS cells (EA).
iv) Examine the usefulness of microRNAs and transcription factors, such as PBX1 and miR-181a, to promote the mesDA differentiation of LtNES (EA and OB).
v) Apply laminins in mesDA differentiation of LtNES and hPS cells, such as 111, 521 and 511 (EA and MP).
Amongst these strategies, we found that factors such as Wnt5a or laminins can be readily applied to improve the differentiation of hEs of hLtNES into mesDA neurons.

(2) Further development of existing protocols. Here we refined our previous work of patterning mesencephalic dopaminergic (mesDA) neurons from hPSCs into an extremely efficient, GMP-compatible, protocol that allows production of billions of DA progenitors from just a single differentiation batch (MP).
Our new optimized protocol (Figure 2) overcomes the critical issues of batch-to-batch variability, as this protocol is specifically designed to generate only caudalized VM progenitors, which we have shown through a large number of transplantations in adult rats to give rise to mature and functional mesDA neurons. In this protocol, we fine-tune the caudal patterning of the FOXA2+/LMX1+-expressing VM progenitors through carefully timed addition of fibroblast growth factor 8b (FGF8b) to the cultures. We further show how the VM-patterned progenitors can be cryopreserved for future in vitro maturation or transplantation studies directly from the frozen vials without the need for intermediate culturing. This differentiation protocol can be used to produce high-purity cultures of mesDA progenitors to be used for transplantation studies, as well as for in vitro disease modeling studies such as drug screens, transcriptomic and proteomic analyses, gene therapy testing and other biomedical applications.
(3) Determine the quality of hPSC-derived mesDA cells. Understanding human embryonic ventral midbrain is of major interest to develop cell therapies for Parkinson’s disease. However, the cell types, their gene expression dynamics and their relationship to commonly used rodent models remain to be defined. We performed single-cell RNA-sequencing to examine ventral midbrain development in human and mouse. We found 25 molecularly-defined human cell types, including five subtypes of radial glia-like cells and four progenitors. In the mouse, two mature fetal dopaminergic neuron subtypes diversified into five adult classes during postnatal development. Cell types and gene expression were generally conserved across species, but with clear differences in cell proliferation, developmental timing and dopaminergic neuron development. Additionally, we developed a method to quantitatively assess the fidelity of dopaminergic neurons derived from human pluripotent stem cells, at a single-cell level. Thus, our study provides insight into the molecular programs controlling human midbrain development and provides a foundation for the development of cell replacement therapies.

1.3 Generation of functional inducible DA neurons in vitro and in vivo.
Cell replacement therapy for Parkinson’s disease has mainly focused on transplanting the precursors of cell types lost by disease. Here we describe an alternative strategy in which induced dopaminergic neurons (iDANs) can be generated by direct conversion of human astrocytes in vitro or by direct reprogramming of mouse striatal astrocytes in vivo (EA). This process requires three transcription factors (Ascl1, Lmx1a, Neurod1) and one microRNA, miR-218, collectively referred to as ‘NeAL218’. Additionally, a cocktail of small molecules promoting chromatin remodeling and activating Shh and Wnt signaling increased the efficiency of human astrocytes reprogramming to 16.48 ± 8.6%. Reprogrammed iDANs expressed mesDA-specific TFs (FOXA2, EN1, NURR1 and PITX3), DA neuron markers (DAT, VMAT2) and SNC/A9 neuron markers (GIRK2 and ALDH1A1), while the VTA/A10 neuron marker, Calbindin, decreased. Moreover, mature TH+ neurons exhibited long processes and co-expressed MAP2+, SYN1+, DDC+, DAT+ and GIRK2+ cells. These cells also exhibited inward and outward voltage-gated currents, rapid spontaneous activity, SAG rectification, calcium responses to KCl and capacity to generate multiple action potentials by day 14 in vitro, indicative of successful reprogramming of human astrocytes into functional iDANs. Moreover, in a 6-OHDA-mouse model of PD, adult striatal astrocytes in GFAP-tTA mice were reprogrammed into iDANs (Figure 2) capable of producing action potentials and receiving synaptic afferents by intrastriatal injection of NeAL218, but not by GFP. At a behavioral level we found that amphetamine-induced circling was not improved. However, apomorphine-induced or spontaneous circling behavior improved by NeAL218, but not by control GFP. Notably, a more detailed analysis of gait revealed that mice treated with NeAL218, but not GFP, did less gait cycle errors (wrong limb order), and other parameters found altered in 6-OHDA mice injected with GFP, such as reduced number of gait cycles, or impaired sagital symmetry were rescued. These findings suggest direct in vivo reprogramming of endogenous striatal astrocytes into iDANs may provide an alternative to cell replacement therapy for PD. Such alternative would not require cell transplantation or immunosuppression. We suggest that further optimization of this approach may enable clinical therapies for Parkinson’s disease by delivery of genes rather than cells.

1.4 Generation of stable human MSN DARP32+ neurons
Here, we describe the final ontogeny-recapitulating protocol developed in Elena Cattaneo’s lab for the direct differentiation of human pluripotent stem cells into striatal projection medium spiny neurons (MSNs). At the present time, this procedure is our best candidate for future migration to GMP-compliant production.
The protocol consists of three phases where cells in monolayer are exposed to different sets of morphogens (Figure 3). At first, neuronal induction is obtained by simultaneous repression of both BMP and TGF-β signalling according to the dual SMAD inhibition strategy developed by Lorenz Studer’s lab (Chambers S. et al., Nat. Biotech., 2009). In a second phase, the exposure to appropriate concentration of the developmental factors Sonic Hedgehog (SHH) and the Wnt inhibitor Dickkopf-1 (DKK-1) triggers ventral telencephalic specification. Terminal differentiation and cell cycle exit are then facilitating by blockage of Notch signalling mediated by the γ-secretase inhibitor DAPT.
This procedure has been extensively validated in different cell type including H9, RC9 and RC17 and HD-iPSCs and leads in 50 days to electrophysiologically active GABA+/DARPP32+/CTIP2+ positive striatal neurons with a relative abundance of 10-20% of the final cell culture.
The following protocol has been applied successfully in different lines of human embryonic stem cell (H9, H9-GFP, RC9 and RC17) and in three different lines of iPS derived from healthy donors and generated by the CHDI Foundation (HD-28CAG clone 6 (28#6), HD-33CAG clone 1 (HD-33#1), HD-21CAG clone 1 (HD-21#1)).
The protocol described here is the result of several years of optimization and testing and represents the starting point for the development of an effective cell therapy for HD. We found that striatal progenitors generated using this approach and transplanted in HD animal models show the ability to survive, mature and integrate in the host tissue showing promising potential for future approach aimed at reestablishing damaged neuronal circuits. We are now in the process of assessing functional efficacy of this approach before engaging in the GMP transfer of this procedure.
Thus, the results obtained so far represent an important step forward towards the development of human PS cells for cell replacement therapy in Huntington’s disease

WP2 - IDENTIFICATION OF OPTIMAL PROTOCOLS AND PROCEDURES FOR EFFICIENT GENERATION OF TRANSPLANTABLE DA NEURONS THAT MATCH THE PERFORMANCE OF AUTHENTIC HUMAN FETAL MIDBRAIN DA NEURONS, IN THE ABSENCE OF ANY PROLIFERATING CONTAMINANTS, AFTER TRANSPLANTATION IN RODENT MODELS OF PD

The Parmar group (LU) has worked in close collaboration with Roslin Cells and Miltenyi in order to optimise our research-grade DA differentiation protocol (Kirkeby et al 2012) to GMP standards in the best possible way. Concomitantly, we recently discovered that in vivo DA yield of our grafted ventral midbrain (VM)-patterned cells correlates specifically with a caudalised phenotype of the VM progenitors, whereas a more rostral patterning of the cells gives rise to DA-poor grafts (Kirkeby et al., 2017). Importantly, these two different phenotypes cannot be distinguished by the markers we and others commonly used to assess the VM patterning (LMX1A, FOXA2 and OTX2). Instead, the caudal phenotype must be assessed by qRT-PCR for a set of genes which we have recently identified as predictive of good graft outcome (EN1, SPRY1, CNPY1 etc, Kirkeby et al., 2016). We have therefore applied screening for these markers in our VM differentiation protocol and we have succeeded in optimising the protocol specifically towards generating VM progenitors of a caudal phenotype (Kirkeby et al. 2017). Over the course of the last 7 years, we have tested the in vivo performance of our hESC-derived cells in a large number of rats (including cells from our research-grade protocol as well as from our GMP-adapted protocol), thereby providing us an extensive dataset of graft outcomes.
All of these GMP adaptations were tested on the Roslin GMP cell line RC17, as this is a line with clinical potential, and also on H9 and the Karolinska GMP-like hESCs HS980, HS983a, HS999 and HS1001 (the KI lines has not been funded by NSCR). In addition, the in vivo graft outcome and functionality of the cells produced by the GMP protocol has been validated in long-term rat studies for three separate batches of RC17, for one batch of H9 and for one batch of HS980

2.1 Optimal protocol
During the course of NSCR, we have developed a highly efficient GMP-compatible hESC differentiation protocol for production of pure cultures of caudal VM progenitors which give rise to DA rich and functional graft in vivo (Kirkeby et al., 2017; Nolbrant et al., 2017, Fig. 1). This protocol produces a high yield of DA progenitors, thereby enabling manual production of a large number of cells for clinical trial. The protocol has further been tested for its reproducibility on various different GMP-cell lines (Fig. 2). The protocol has been transferred to Roslin Cells and the Roslin personnel has been trained in the procedure and have already produced the 3 medium-size test batches of differentiation, which have been transferred to Cardiff, Cambridge and Lund for transplantation into a rat model of PD. In addition, the protocol has been transferred to Miltenyi, who is now testing the protocol on H9 in the Prodigy system for enabling closed-system production for future clinical upscaling

2.2. In vivo performance of hESC-derived mesDA cells

Previous publications have demonstrated that efficient functional recovery is achieved with 15,000 – 18,000 TH+ hESC-derived DA neurons surviving in the rat 6-OHDA lesion model of PD (Kirkeby et al. 2012, Kriks et al. 2011). With such large cell numbers, it is difficult to interpret the efficacy of hESC-derived DA neurons. Thus, we aimed to address the question - what was the minimum number of hESC-derived DA neurons that are sufficient for functional recovery in a rat model of PD?

To this end, we performed a potency assay and transplanted unilaterally lesioned 6-OHDA rats with 10-fold fewer cells than used as standard in previous publications (Kirkeby et al. 2012, Kriks et al. 2011) and monitored them for functional recovery as assessed using amphetamine-induced rotations. At 16-weeks post-transplantation the grafts of hESC-derived DA neurons demonstrated a statistically significant normalisation of amphetamine-induced rotations as compared to lesion controls (Fig. 3A, t4 = 6.76 p < 0.01; n = 5). Histological analysis revealed that TH+ DAergic neurons with a mesencephalic morphology survived (Fig. 3C). Upon quantification we observed 986 ± 333 TH+ neurons per rat (n = 5), with 2 rats with fewer than 500 surviving TH+ neurons demonstrating complete functional recovery (Fig. 3B).
When comparing this data with the most recent data on hVM rodent transplantation from the ongoing EU-funded clinical trial – TRANSEURO – it is reported that as few 657 ± 199 surviving TH+ neurons sourced from hVM are sufficient to provide complete functional recovery (Rath et al. 2013). This is in-keeping with historical pre-clinical data that supported the first hVM clinical transplantation trials, that reported normalisation of amphetamine-induced rotations was achieved with 1,200 surviving TH+ neurons sourced from hVM (Brundin et al. 1986).

This data demonstrates that the minimum number of hESC-derived DA neurons that are sufficient to normalise amphetamine-induced rotational bias is similar to that of hVM-derived DA neurons – indicating that they are equipotent. Thus greatly supporting the hypothesis that hESC-derived DA neurons generated using our protocol (Kirkeby et al. 2012) match the in vivo performance of hVM-derived DA neurons.

2.3. Absence of proliferation post-transplantation

We previously characterised that with an initial phase of proliferation of the transplanted progenitors occurring around 6 weeks post-transplantation, this was primarily restricted to Nestin+ neural progenitors that subsequently terminally differentiated some time before 18 weeks post-transplantation (Kirkeby et al. 2012). With collaboration with MirCen in France, we performed a set of long-term transplantations in athymic rats with a survival time of 4 weeks (Grealish et al. 2014). Thus providing the longest-described survival of a hESC-derived grafts in vivo, and allowing sufficient time for any possible proliferative contaminant within the graft to reveal itself. We employed clinically relevant imaging techniques, T2-weighed MRI to monitor graft survival and growth. This was also combined with MR spectroscopy (MRS), to identify the cell populations within the graft in a non-invasive manner.

T2-weighed MRI imaging showed surviving transplants that increased in volume from 1 to 5 months post-transplantation, indicating initial proliferation and subsequent maturation of the transplanted cells in all animals (Fig. 4C). No physical malformations were observed within the grafted striatum, although the transplant is clearly visualized. MRS investigation 5 months post-grafting, revealed a high NAA content within the transplanted area indicative of a neuron rich graft (Fig. 4A and 4B), that is described in detail in the figure legend below.

Thus using non-invasive, clinically compatible imaging techniques we have demonstrated that transplants of hESC-derived mesDA neurons do not proliferate at high rate, nor do they disturb the host parenchyma and display a highly neuronal phenotype

2.4 Assessment of in vivo reproducibility of optimised GMP-compatible protocol
During the process of our in vivo assessments of graft outcome, we realised some significant variation of graft outcome when applying our research grade protocol (Kirkeby et al., 2017). We have therefore assessed thoroughly the reproducibility of the in vivo outcome from grafts produced with our optimised GMP-compatible protocol. We have found that implementation of the GMP-compatible protocol has eliminated the occurrence of batches with very poor dopaminergic yield, which would previously appear intermittently when using our research-grade protocol (Fig. 5A). In addition, we have validated that all transplanted batches of cells from our optimised GMP protocol (6 batches so far) have produced complete functional recovery in a rat model of PD (Fig. 5B). Finally, we have also tested a cryopreservation protocol on our cells and subsequently transplanted the cells directly from the freezer after only a rapid washing procedure. Also these cells (Fig. 5B, RC17 CRYO) produced functional recovery in our animal model of PD, meaning that we have succeeded in generating a reproducible, high-yield GMP protocol as well as a cryopreservation and transplantation strategy which all in all enable us in proceeding towards clinical translation of our hESC-derived mesDA stem cel product for PD.

WP3 - IDENTIFICATION OF OPTIMAL PROTOCOLS AND PROCEDURES FOR EFFICIENT GENERATION OF TRANSPLANTABLE MSNS THAT MATCH THE PERFORMANCE OF AUTHENTIC HUMAN FETAL STRIATAL MSNS IN THE ABSENCE OF ANY PROLIFERATING CONTAMINANTS, AFTER TRANSPLANTATION IN RODENT MODELS OF HD

Huntington’s Disease (HD) is a neurodegenerative disorder characterized by the prominent loss of medium spiny neurons (MSNs). With the aim to optimize the differentiation and integration of grafted human striatal progenitors in view of promoting functional recovery, we grafted GFP-expressing H9 cells into the striatum of athymic rats after monolateral quinolinic acid (QA) lesion, and we examined transplanted cell maturation and organizations two months (8w) after transplantation according to the design showed in Figure 1. Starting from 8 weeks post transplant (wpt), grafted cells show good survival rates with host tissue showing normal astrogliosis regardless whether it received cells or only sham surgery (Figure 2), and start to express striatal markers such as Ebf1, Ctip2, Darpp32 (Figure 3). Proliferating cells are still detectable by Ki67 labeling but represent only around 8% of the total human cells. Few Ctip2/Darpp32-positive cells are visible at this early stage of differentiation, suggesting that longer time may be necessary to increase cell maturation into MSN phenotype.

By combining GFP+ cell and hNCAM-labelling, we were able to trace graft integration within the host tissue. Positive processes are visible into the rat striatum, corpus callosum, globus pallidus, thalamus and substantia nigra, confirming that donor cells can send axonal projections to the proper brain targets at 8 wpt (Figure 4).

We took advantage of a rabies virus-based molecular tool developed by the group of Malin Parmar to trace graft-to-host connection up to 2 months after surgery. This system allows to specifically mark graft cells (GFP) making them permissive for rabies virus infection (RFP). The genome of the rabies carries a coding sequence for RFP that can be retrogradely transported from the started cells (GFP+/RFP+) to the first level input neurons (RFP only). We showed that 4 and 8 weeks after transplantation, starter cells and mCherry+ traced neurons are organized in nets, with a progressive increase in connectivity (Figure 5). Starter neurons make connections with both cells from the graft (HuNu+) and from the host (HuNu-). Only intra-striatal networks are visible.
To monitor the effect of the striatal lesion and assess possible functional improvements in relation to the histological findings, rats underwent behavioral tests at different time points both before and after transplantation (2, 4, 8 weeks). QA lesion induced a prominent decrease in the use of the contralateral forelimb, but MSN engraftments appeared to promote the amelioration of striatum-dependent performances in some tests, starting from 4wpt, in comparison to sham control animals (Figure 6).
From 4wpt, engrafted animals start to show better performance in the adjusting step test and vibrissae-evoked hand placing test. Further experiments at longer survival time-points are needed to establish long term functional outcomes and correlate functional improvements to neurological findings.

In conclusion, these preliminary observations suggest that hES-derived MSN progenitors according to Delli Carri protocol can survive, mature and extend processes to area distal to the graft and located outside the striatal nuclei.
Data at 8wpt suggest that grafted cells survive into the host striatum and start to express markers of striatal differentiation. Cells show mature morphology and phenotype, appearing integrated in the host tissue. Indeed, since one month after transplantation, grafted cells start to be organized in nets and to be connected with both transplanted and host cells. They also send axonal projection to proper striatal and extra-striatal targets. Cell differentiation and integration also support some functional recovery observed in striatum-specific behavioral tests.
Due to initial difficulties on establishing the proper in vitro maturation time point for optimal survival of the grafts and in consideration of the challenging logistic necessary to perform a comparison between hES-derived cells and cells from human WGE that need to be available for the same set of animals, we haven’t been able to perform a side-by-side potency assay in this context.

Thus, experiments at longer survival time-points are needed to confirm the capability of these cells to become functionally integrated and improve motor performances. These data need further validation, but are promising for future development of cell replacement therapies for HD.

WP4
IN VIVO FUNCTIONAL EFFICACY OF OPTIMISED DAERGIC AND MSN CELL LINES, IDENTIFYING PARAMETERS INDICATING SUITABILITY FOR PROGRESSION TO CLINICAL IMPLEMENTATION

Aim 1: To undertake cell potency assays to optimize cell number for functional evaluation of each candidate cell therapy/protocol based on cell survival, absence of tumourigenesis and simple functional screening assays.
Objective: To determine optimal cell potency / placement for transplantation of candidate DAergic and MSN cells.

Parkinson’s disease: Using preclinical models of Parkinson’s disease (PD), combined with a previously described protocol (Kirkeby et al. 2012), M. Parmar’s lab sought to investigate two aspects of the clinical translation of hESC-derived dopamine (DA) neurons for cell transplantation: 1) cell potency (i.e. how well do the cells work compared to fetal VM cells?); and 2), phenotypic stability and optimal survival time to monitor the cells after grafting. Cell potency results indicated that the functional potency of grafted hESC-derived DA neurons is similar to that of human DA neurons obtained from fetal ventral mesencephalon (VM). Across over 10 experiments, in which over 100 rats were grafted with heSC-derived dopaminergic grafts, we have never observed any evidence of overgrowth or tumourigenesis. Prof Parmar’s lab has tested this explicitly by spiking in pluripotent cells to the cell preparation. However, in functional assays, neither M Parmar’s lab nor Prof Dunnett’s lab have ever independently observed any tumour growth, even after survival of the graft for more that 1 year. Lastly, Prof Parmar’s and Prof Dunnett’s labs consistently (in more than 10 experiments) observed improvement in motor deficits from ~16 weeks post-graft.

Huntington’s disease: Experiments were undertaken to assess the ability of ES- and iPS-derived medium spiny neurons (MSNs) to survive in vivo, either as an unsorted population or after sorting for PS-NCAM. The Cattaneo lab conducted a series of transplantation experiments using hPS as cell source with the aim of assessing survival, maturation and integration of committed ventral telencephalic progenitors. In 4 different experiments the lab have been able to show that 1) progenitors derived after 30 days of in vitro differentiation protocol poorly survive in vivo even at short time after surgery (2 weeks); 2) Day 20 progenitors are able to survive (up to 2 months), integrate and mature at least in part towards the correct lineage (Ctip2+/DARPP32+). Up to 50% of the graft cells showed an MSN phenotype and only a residual proportion was labelled by interneuronal markers. The nature of the remaining cells has yet to be determined; 3) two months after transplant, day 20 progenitors showed to be able to extend processes outside the nucleus striatum reaching as far as the cortex, the substantia nigra and the other hemisphere through the corpus callosum. Thus, these cell populations have been optimised to produce an improved MSN-rich graft. These data represent a proof-of-principle for potential therapeutic applications of these cell preparations. Further experiments will be needed to define the therapeutic potency of these cells, using sensitive behavioural tasks aimed at assessing the potential to improve motor dysfunctions in rodent models of Huntington’s disease.

Aim 2: To characterise the functional efficacy of novel PS-derived cell lines on simple and complex motor function in comparison to fetal tissue grafts as the ‘gold standard’ positive control
Objective: To evaluate the ability of cell lines to relieve motor dysfunction in animal models of PD and HD.

Parkinson’s disease: In order to standardise the behavioural tasks across laboratories to achieve maximal comparable data from collaborators, we sought to define the optimal behavioural battery for assessment of motor function in rodent models of Parkinson’s and Huntington’s disease. We achieved this by reviewing the multiple behavioural tests available, clarifying which test was appropriate for each of the 2 types of rodent model, and defining the parameters for utilising these tasks in the laboratory. A comprehensive investigation into the functional efficacy of the hESC-derived DA neurons generated by M Parmar has been completed. We reported on the survival and differentiation of a novel lt-NES derived DA cell therapy product in vivo, as well as a systematic histological and behavioural investigation of hVM tissue grafts and hESC-derived dopaminergic cells differentiated according to the protocols published in Jaeger et al., 2011 and Kirkeby et al., 2012. Experiments demonstrated that stem cell-derived dopamine neurons engraft efficiently, survive and integrate into the brain. Indeed, the histological characteristics are very similar to authentic fetal hVM dopamine neurons and they are capable of alleviating several key motor impairments to a similar level as the authentic fetal dopamine neurons. The hESC-derived grafts integrate into the host brain, form synaptic connections and harbour similar portions of non-DAergic cells (5-HT, GABA, GFAP). Moreover, these results have been observed in multiple laboratories now, suggesting that these effects are robust and reproducible. These data contribute to the growing body of evidence that demonstrates the safety and efficacy of hESC-derived dopamine cells and suggests that we are now very near to clinical application of these cells.

Huntington’s disease: Although the development of the protocols to derive efficient ES-derived MSNs for Huntington’s disease is still underway, several protocols do now produce neural cells that express the important markers of MSNs. It is, therefore, essential that we also assess these different ES-derived MSN neurons in rodent models of disease, in order to ensure their safety, survival and functional efficacy. Within this NSCRepair consortium, the decision was been made to undertake the in vivo assessment of ES-derived MSNs in Cambridge and Italian laboratories until functional MSNs were established and well-characterised. By the end of the grant, an efficient protocol for producing MSNs was being utilised in in vivo models in Turin, but further functional analyses will be necessary in the future to determine the functionality of the cells.

It has been important, however, to establish a clear baseline characterisation of the "gold-standard" fetal medium spiny neurons using human WGE grafts, to determine the level of functional improvement that can be expected from a novel ES-derived MSN cell therapy product. Here, we reported a comprehensive assessment of the functional efficacy of rat and human whole ganglionic eminence (WGE) grafts on a range of motor tasks. The data reveal subtle but comparable improvements in behavioural performance in rats grafted with rWGE and hWGE primary tissue, despite significant differences in the total volume of surviving graft tissue.

Aim 3: To establish impairments in key cognitive / psychiatric functions disturbed in PD and HD.
Objective: To develop highly translational behavioural assays to evaluate relevant cognitive functions.

This work aimed to develop tasks sensitive to response conflict, response inhibition and impulsivity in different models of Huntington’s and Parkinson’s disease. There is recent recognition that an array of non-motor symptoms arise early in the manifestation of Huntington’s and Parkinson’s disease. We have described the establishment of three tasks, the Response Conflict task, Stop-Signal task and Delayed Reinforcement task. The main findings from 5 separate experiments reveal: 1) that the ventral striatum, but not the medial or lateral regions, is the primary striatal subregion involved in performing the Response Conflict task; 2) a novel version of the Stop-Signal task can be performed by naïve rats, but the involvement of the striatal subregions has not been observed; 3) impulsivity changes are clearly evident in mutant transgenic models of Huntington’s disease. However, whilst we observed varying degrees of involvement of striatal MSNs and dopamine transmission in these tasks, they have not proven to be more sensitive than the classic Lateralised Choice Reaction Time Task, in detecting striatally-dependent motor and non-motor dysfunctions.

Aim 4: To reveal functional efficacy of novel cell lines on cognitive and neuropsychiatric function using specialised, automated equipment sensitive to subtle neuronal manipulations.
Objective: To utilise novel and established behavioural tasks of cognitive function to determine the ability of the best candidate cells lines to relieve neuropsychiatric and neuropsychological aspects of disorder.

A comprehensive experiment has been undertaken in Cardiff, in which hESC-derived DA neurons from M Parmar’s lab are utilised in the established Lateralised Choice reaction Time task, which assesses attentional function, motivation and visuospatial function. The experiment was run over an 18 month period and rats have recently been culled, brain tissue has been harvested and the immunohistochemical analysis of the tissue is underway. The animals were kept alive for over 1 year post-graft. Thus far, no improvements in any aspects of cognitive function (attention, motivation and visuospatial function) have been observed. Imprtantly, however, we know that survival of the grafts was poor, based on the amphetamine-rotation data. This will be a result of using a more novel (and perhaps more variable) neonatal tolerisation technique, which has the theoretical potential to sustain graft survival for up to 2 years, but is also less well-characterised than the short-term immunosuppressant model. Thus, due to the poor survival of grafts in this experiment, we cannot conclusively determine whether these grafts are capable of alleviating non-motor deficits, leaving only the established knowledge of their ability to improve motor impairments.

Aim 5: To encourage and improve recovery through targeted training and specific protocols designed to enhance anatomical repair by favouring meaningful synaptic connections.
Objective: To utilise targeted training protocols to encourage the development of appropriate synaptic connections, to increase axonal outgrowth and to enhance recovery of motor and cognitive dysfunction.

Several forms of training protocols have been shown to delay the progression of neurodegenerative diseases through plastic modification and/or to preservation of the affected circuits. Moreover, training protocols such as enriched environments and targeted training protocols have been shown to improve the integrative potential and functional efficacy of rodent foetal cell grafts in disease animal models (Mazzocchi-Jones, Neurorehab Neur Rep, 2011; Brasted, PNAS, 1999). Therefore, in view to foster functional repair, the Turin lab aimed at enhancing the integration of human transplanted cells into the damaged recipient tissue and the pattern of donor-host interactions by employing rehabilitation strategies as behavioural ‘enhancement’ techniques.
As a first step toward this aim, here they investigated strategies to promote functional recovery in Huntington disease rodent models. In particular, they examined the effects of two different protocols of training: 1) enriched environment (EE), to provide a global, non-specific stimulation to lesioned rats and 2) constraint induced movement therapy (CIMT) to selectively elicit movement of limbs damaged in the pathology model. Their results show that both global and selective stimulation of motor functions are able to provide amelioration of motor impairment. They demonstrate that both EE and CIMT are able to provide positive effects on motor symptoms related to voluntary movement - tested with staircase - and spontaneous movement – tested with Hand Placing tests and Adjusting Step test Vibrissae Evoked. The ameliorations of motor symptoms is observed at 3 weeks after the beginning of the treatment and lasts for several weeks. Based on the effectiveness of these interventions, it is expected that we will enhance functionality of human stem-cell derived therapies by relevant adaptive circuit rewiring.

Aim 6: Implementation high-speed 3D light sheet microscopy for imaging of functional connectivity patterns of grafted neurons (Life&Brain, OB)
Objective: To develop and implement advanced live imaging technologies for assessing functional integration of grafted neurons.

It is necessary to develop imaging protocols capable of identifying functional synapse formation between grafted cells and the host brain as one means of demonstrating the functional efficacy of a cell therapy. To date, commercially available solutions aim for high resolution of small specimens but are often unsuitable to be used in the context of larger probes. Here, Life & Brain report a proof-of-concept study for tracing transsynaptic connectivity between transplanted long-term neuroepithelial (lt-NES) cells. They show that the light-sheet fluorescence microscope (LSFM, MPI Heidelberg) enables a simultaneous visualization of virtually all host neurons innervating a grafted mRFP1-labeled cell population infected by rabies virus. Based on this exploratory work they developed an optimized LSFM image acquisition platform facilitating the fast and high-resolution imaging of specimens up to the size of a rat brain. In addition, for comprehensive neuroanatomic analysis they developed a set of tools that substantially improve the speed and accuracy of neural circuit reconstructions in the context of LSFM-generated whole brain preparations. First, they develop an optimized tissue clearing technique (FluoClearBABB) suitable for a straightforward preparation of samples up to the size of a rat brain. In order to map host neurons, which project onto and establish synaptic interaction with grafted human neural stem cell-derived neurons they established a rabies virus-based (RABV) mono-transsynaptic tracing system. Thereafter they generated human embryonic stem cell-derived neural stem cells (lt-NES cells) ubiquitously expressing (i) mRFP1 to label all transplanted cells, and (ii) a synapsin promoter-driven combination of the avian TVA receptor, the B19 rabies glycoprotein and a H2B.EGFP fusion protein. The lt-NES cells (~60.000) were stereotaxically delivered to the striatum or the hippocampal dentate gyrus of adult unlesioned immunodeficient Rag2-/- mice to permit long-time survival of the cells. Ten weeks later the graft was infected by stereotaxic injection of recombinant pseudotyped, glycoprotein-deleted EGFP-expressing rabies virus (RABVΔG-EGFP(EnvA) and mice were subsequently subjected to our LSFM set up for image acquisition and analysis. Results from the striatal and hippocampal grafts indicated retrograde labelling of synaptically connected host neurons projecting onto the transplant-derived neurons. To provide quantifiable anatomical data on the functional integration of donor neurons we devised an approach for the quantitative assessment of the afferently connected host neurons by co-registering our LSFM data sets with high-precision three-dimensional (3D) MRI reference data sets.

Collectively, our results suggest that establishment of afferent and efferent connections between host brain and grafted human neural stem cells is largely dictated by the topology of pre-existing, endogenous fibre tracts and circuitries rather than the regional identity of the donor cells (Doerr et al., 2017).
Finally, they set out to explore if the developed technology could be used in the context of compound-based enhancement of host-graft connectivity. After optimisation of the technique, they demonstrated suitable quantification sensitivity to detect differences in transplant-host connectivity between treated and untreated mice.

The overall aim of this workpackage has been to optimise the survival, integration and functionality of novel cell therapies for both Parkinson’s and Huntington’s disease. The experiments aim to address four key themes: 1) defining optimal cell placement and potency, 2) establishing behavioural tasks to reveal changes in relevant psychiatric domains, 3) utilising targeted training protocols to enhance recovery and 4) developing novel imaging technologies to visualise the integration and connectivity of neurons.

We have collected a large data set addressing each of these themes and have successfully achieved our aim of demonstrating the survival, integration, safety and functional efficacy of hESC-derived dopamine grafts for Parkinson’s disease. Previous clinical trials using hVM tissue have demonstrated proof-of-principle that dopaminergic grafts can alleviate motor dysfunctions long-term in patients, but much of the recent preclinical work has aimed to optimise the treatment to avoid side-effects and to develop novel stem cell derived cell therapies that are suitable for widespread clinical application. Here we demonstrate consistency in the survival and functionality of these cell therapy products and have laid the foundation for imminent implementation of this therapy to patients with Parkinson’s disease.

For Huntington’s disease, clear progress has been made in the development of new protocols to generate authentic MSNs and optimisation of the transplantation of these cells has been undertaken. Clear improvements in phenotype and survival have been observed. Baseline criteria for the expected functionality of the grafts has been established using human fetal WGE grafts. Now, further work needs to undertake the final stage of demonstrating the functional efficacy of these cell therapy products, before consideration of the clinical application of these cells.

Preclincally, we have developed new tools (established novel behavioural tasks sensitive to striatal dysfunction and developed highly sensitive imaging platforms for detecting subtle changes in graft-host connectivity). Therefore the preclinical study of these therapies is now more refined and better able to detect optimal vs suboptimal cell therapy products.

Lastly, we have utilised the ‘gold-standard’ human fetal tissue for PD (hVM) and HD (hWGE) to establish optimised baselines against which we can compare our novel cell therapy products. hESC-derived MSNs will be compared functionally to hWGE in future experiments and hESC-derived DA neurons have now consistently been found to exude similar survival and efficacy as authentic hVM grafts.

Thus, overall, clear progress has been made in optimising and developing protocols and cell therapies for both conditions and we are ready to progress the hESC-derived dopaminergic grafts to the clinic to treat people with Parkinson’s disease.


WP5

GMP-TRANSLATED MANUFACTURED DA NEURONS

Sourcing appropriate hESC cellular material.

Testing the lines

To identify a cell line that would respond to the mesDA differentiation protocol, three GMP-grade hESC lines, derived by Roslin Cells (RC9, RC11 and RC17), were distributed to Partners at Lund University. In vitro data from these initial studies showed that RC17 responded optimally to the mesDA differentiation protocol. Specifically RC17;
• Expressed high levels of FoxA2+/Lmx1a+ and Otx2+ (>80%) representative of pure mesDA progenitor cells.
• Expressed the anterior and dorsal marker Pax6 and the lateral marker Nkx6.1 and very low expression of the caudal marker HoxA2
• Appeared resistant to generating hindbrain cells as the majority of the cells assayed were Otx2 positive.
• had an intrinsic tendency to form the ventral midbrain (VM) target population
• Immunochemical staining of RC17 cells at day 60 of differentiation returned cultures which were rich in TH+/Lmx1a+ and Nurr1+/MAP2+ neurons

Therefore, in vitro analysis indicated that RC17 performed optimally in the mesDA progenitor differentiation protocol and differentiated cells derived from this line were progressed for subsequent in vivo engraftment studies. This line was also selected for optimisation of the mesDA progenitor differentiation protocol and was intended for GMP TT and manufacturing procedures at RoslinCT.

Creating a research grade RC17 bank

Following selection of RC17 as the most responsive hESC line to the mesDA differentiation protocol, it was important to create a research grade bank of characterised RC17 cells for distribution to NSR consortium members for subsequent translational research. This bank was generated at RoslinCT and expanded from the parental GMP grade RC17 hESC line (at passage 12). In collaboration with Isenet, RoslinCT performed the following QC analysis on this line;
• Genetic identity by microsatellite PCR
• Viral screen and mycoplasma testing
• Pluripotency Marker analysis
• Cytogenetic (Q-banding), CGH and KaryoliteBOBs
• Cell viability as a function on cryo-reagent
• DNA damage in relation to cryo-reagent and thawing
• Telomere length, hTERT expression and telomerase activity
• Evaluation of stress (ER and mitochondrial) as a function of cryostoring and thawing
• Standardizing storing conditions using trehalose

In addition to receiving the characterised line, NSR consortium members were supplied with a certificate of analysis (CoAs) summarising the QC tests performed (Appendix 1). They were also provided with approved Standard Operating Procedures (SOPs) for all hESC culture, expansion and banking procedures and those for first line quality controls to establish microbial status, chromosome integrity, gene expression and protein analysis.
Development of differentiation protocols using reliable and consistent cell banks reduces variability within and between laboratories or consortium groups. Consistency and standardisation of procedures enables meaningful comparison of returned results between groups and allows streamlining of robust procedures for GMP translation. In addition, understanding the mechanisms of the mesDA progenitors derived from this line is fundamental when assessing post transplantation efficacy in regard to cell proliferation, survival, migration, and differentiation.

Technology Transfer and GMP Translation
Protocol development

The conversion of protocols developed in academic laboratories to those suitable for clinical manufacture is critical for successful GMP translation of any proposed cell therapy. With this in mind, NSR consortium members initiated the translation process early in protocol development and TT was performed in several stages. The first stage identified key manufacturing parameters for mesDA neuronal progenitor differentiation and scale up. The second involved completion of a detailed gap analysis to identify any problems associated with the proposed manufacturing strategy where alternative solutions were provided in order to reach the intended goal.

The third involved a thorough risk assessment of not only the process but also the materials used within the process to ensure that the manufacturing protocol was performed in line with GMP regulations and to enable robust and reproducible processing whilst minimising risk to the intended patient. As part of this risk assessment, high risk and therefore GMP incompatible items were replaced with items that could be satisfactorily risk assessed and were available at the time of protocol development (Appendix 2).

The fourth step involved protocol TT from the academic site to RoslinCT where procedural steps were performed in line with manufacturing parameters but to reduced scale then qualified with GMP compliant materials, reagents and consumables. The final step will involve transfer of the qualified procedure to the GMP facility for process validation and subsequent GMP production.

Concurrently, development of an automated system was performed by Miltenyi Biotec using the CliniMACS Prodigy.

Therapeutic Product Profile (TPP)

In parallel with mesDA differentiation protocol development, RoslinCT compiled a Target Product Profile (TPP) document for final CTP definition. This document forms part of a general tool set which supports a risk based approach to product development. This document takes into consideration the European Marketing Authority (EMA) reflection paper regarding the derivation of stem cell-based medicinal products (EAM/CAT/571134/2009) and was continually populated and updated throughout the development process. Critical product parameters within this document include;

• Non-Clinical and Clinical Development Strategy
• Regulatory Compliance
• Manufacturing Process Description
• Final Formulation and Administration
• Quality Control (QC) Regime (Intermediates and In-Process Control)
• Final Product Characterisation Profile
• Final Product Safety Testing Requirements
• Raw Materials and Consumables
• Banking of RC-17 hESC Line for use within NeuroStemCellRepair Project (P014).

Technology Transfer (TT) – Lund to RoslinCT

TT of the NSR hESC-derived mesDA second generation differentiation protocol was successfully performed under the supervision of partners in Lund to personnel within the Cell Therapy Development (CTD) department of RoslinCT.
In parallel, a flow cytometry assay to assess the quality of the mesDA neuronal progenitor cells produced by this process was developed by collaborators in Lund and Miltenyi Biotec (Del 5.10) then applied to the Cell Therapy Product (CTP) generated at RoslinCT.
After completion of the initial TT run (NSRUST01), the original protocol was modified to include risk assessed GMP-compliant reagents and subsequently used to perform three preliminary up-scale batches (NSRUST03-05) in T75 flasks to evaluate original protocol performance. Upon flow cytometry analysis of the cellular material generated during the up-scaled runs, the original differentiation protocol was modified further to optimise the CTP purity. These modifications included the reduction of Chiron (CHIR99021) from a concentration of 0.9 µM to 0.7 µM and increasing Sonic HedgeHog (SHH) concentration from 200 ng/mL to 300 ng/mL. In addition, the Quality Control (QC) flow cytometry marker panel was reassessed to include monoclonal antibody panels to increase specificity of cell characterisation throughout the manufacturing process. Incorporating these modifications to the differentiation protocol and QC strategy allowed protocol optimisation with defined sampling points for subsequent CTP manufacture.

Refined manufacturing parameters
This refined strategy (described above) was applied during the manufacture of a further three up-scaled batches (NSRUST07-09) at a scale of 2x T75 flasks (per run) within the RoslinCT CTD department. These runs were performed under ‘clean bench’ conditions in accordance with SOP/RDP/112. During the differentiation process, QC samples were collected at three sample points (SP): Day 0 (SP1), Day 11 (SP2) and Day 16 (SP3). These sample points covered both in-process and final manufacturing points and were tested by flow cytometry by RoslinCT and Miltenyi Biotec.

The harvested CTP was cryopreserved in accordance with optimized cryopreservation procedures (CTD/REP/16/21) and the resulting CTP was then distributed to NSR partners in Cardiff, Cambridge and Lund to evaluate CTP engraftment when transplanted into PD animal models (artificial brain lesions to mimic the effect of Parkinson disease). Manufacturing of NSRUST07-09 batches under clean-bench conditions is comprehensively documented and evaluated in-line with development qualification procedures and processing activities and associated manufacturing SOPs and standard forms are shown in Figures 1 and 2 and Appendix 3.

QC strategy
Designing a flow cytometry assay
Protocols for the generation of therapeutically relevant mesDA progenitor cells in the context of Parkinson’s disease (PD) are optimized and well established but determination of contaminating pluripotent cells within in-vitro differentiated cell cultures is unknown. As a result, there is a need for an in vitro assay for comprehensive cell product characterization and in-process control, to determine the identity, purity and engraftment potential of the produced CTP.

Typically mesDA neuronal progenitors are characterized by expression analysis of mesencephalic floor plate markers such as LMX1a, FOXA2 and OTX2 on the mRNA level by PCR or on the protein level by immunofluorescence (IF) microscopy.

To extend this QC analysis, Miltenyi Biotec developed a flow cytometry assay to monitor changes in protein expression throughout the mesDA progenitor differentiation process. Development of this assay investigated several antibody clones that were specific for mesencephalic floor plate markers including FOXA2, OTX2 and LMX1b. Furthermore, antibodies specific for negative markers including PAX6, NKX2.1 NKX6.1 GBX2, BARHL1 were evaluated to detect unwanted cell populations. CHIR99021 (CHIR) and hSHH-C24-II (SHH) concentrations were modified within the differentiation protocol to induce non-midbrain cell populations and to test assay sensitivity. The final assay was performed with 3 antibody panels (Table 1), each containing 3 matched antibody conjugates. Furthermore, read out parameters and signal intensity boundaries correlating to certain cell compositions were defined and detailed protocols as well as a gating strategy guideline for analysis were compiled (Del 5.10).

The NSR partners at Lund and RoslinCT provided Miltenyi with different batches of fixed or frozen RC17 or H9 differentiated cell populations for further validation of the assay, including 3 cell batches produced at RoslinCT for pre-clinical transplantation experiments. Finally, spike-in experiments of pluripotent cells were performed with percentages between 0.01- 5% of Oct3/4 positive cells to analyse the limit of detection when using a MACSQuant Flow Cytometer. This was used to gain a measure of the contaminating and therefore, potentially tumorigenic pluripotent cells remaining within the differentiated CTP. The following thresholds for percentages of different markers were defined for QC release of the cell product:

• FOXA2/OTX2 double positive cells: >85%
• PAX6 positive cells : <5%
• NKX6.1 positive cells : <5%
• SOX1 positive cells : <5%
• NKX2.1 positive cells : 5-90%
• OCT3/4 positive cells: ≤ detection limit of 0.02%

In summary, Miltenyi showed that the flow cytometry based QC assay allows to reliably determine the identity and purity of the cell product at different time points of the differentiation process and allows for highly sensitive detection of contaminating pluripotent cells prior to cell administration to patients suffering from PD in a cell replacement therapy.

Manufacturing QC overview
There are several key QC check points and sampling points throughout the manufacturing process designed to monitor the progression of mesDA neuronal progenitor differentiation. These QC checks are summarised in tables 2, 3 and 4. Visual inspection by microscopy was performed routinely and flow cytometry was performed at seeding and harvesting stage when cellular material was available. Key processing days and sampling points are displayed in Table 2 below. CTP microbial, sterility and mycoplasma testing is intended upon harvest on Day 16 and cell recovery analysed upon thaw.

Optimising Cryopreservation and Thaw Strategies
The ability to cryopreserve a harvested CTP allows greater flexibility when designing a clinical delivery strategy as a cryopreserved product can be fully characterised prior to release and can be shipped to multiple clinical sites at the convenience of the patient. To investigate the possibility of mesDA progenitor cryopreservation, an initial cryopreservation study was performed by partners in Lund. This group compared different cryopreservation strategies and showed that there was no difference in DA neuron content when comparing in vivo grafts from cryopreserved and fresh cells upon transplantation into animal models.

To further investigate these initial cryopreservation strategies, the project “GMP translation of optimized cryopreservation and thawing protocols for hESC derived DA neuronal progenitors” was initiated with the reserve budget (year 2) and represents a collaboration established between several NSR partners. This investigation was performed to address the safety and preparation of the produced mesDA progenitors at cryopreservation and upon thaw within the clinic.

This additional work was performed at RoslinCT (developed with guidance from the University of Loughborough) and is fully documented in the NSR cryo-optimisation study (CTD/EXP/16/21) for mesDA neuronal progenitor cells. During this study, CTP generated during large scale NSRUST03-05 batches were formulated and cryopreserved using two types of cryopreservation media (NIM+5% DMSO or CryoBrew+5% DMSO) to determine a suitable cryopreservation media for optimal CTP recovery and viability. The optimal cryopreservation media and cryopreservation strategy was determined by performing viability cell counts using trypan blue and the Countess II automated cell counting instrument. Specifically, samples were collected during CTP formulation and cell viability was determined for both cryopreservation media types at time points up to 180 min post formulation. A schematic showing the cryopreservation testing strategy is shown in Figure 3. Post thaw, cell viability and recovery was assessed to compare cell survival in each cryopreservation media formulation during the cryopreservation process. In addition, representative samples were re-plated upon thaw and counted after 5 days in culture to determine their survival upon re-plating.

Bed side stability study

This new project also included NSR Bed side stability study (CTD/EXP/17/05) performed within the CTD Laboratory at RoslinCT. This protocol was designed and executed to identify a suitable hold media and suitable incubation condition for thawed CTP prior to transplantation within the clinic. This study, utilised mesDA CTP cryopreserved during the large scale, Clean Bench NSRUST07-09 batches (CTD/REP/17/04). These cells were thawed and formulated using two hold conditions (KnockOut-DMEM/0.05% Pulmozyme or Hank’s Balanced Salt Solution/0.05% Pulmozyme) and incubated at various time points up to 10 hours at either 5±3˚C or 21±3˚C. Determination of this hold step was critical to ensure a balance was reached between the time taken to prepare the cells within the clinic and the time taken to prepare the patient for surgery with minimal impact to the cells or to the patient.

This investigation aimed to establish the DA neuron differentiation procedure within an automated, closed system to allow integrated cell processing from starting material (hESC) to final CTP. Automated procedures would include elements of the full processing workflow including sample preparation, cell washing, MACS cell separation, cell culture, and final formulation in a closed system. The success of this investigation would provide advantages over a static flask method in terms of reduced production costs, reduced clean room classification for use, reduced operator skill, increased reproducibility, increased scalability and ease of TT to multiple manufacturing sites.

This investigation was performed by Miltenyi Biotec who established and technically evaluated a magnetic sorting process for mesencephalic dopaminergic neuron progenitor cells (mesDA) (Report Del 5.7) derived from pluripotent stem cells (PSCs) in an automated, closed system. The enrichment of cells was based on an indirect sorting approach using a previously described surface marker (Report Del 2.9) integrin associated protein (IAP, CD47), and the cell processing device CliniMACS Prodigy. A novel program was developed for the automated processing of the cells in the CliniMACS Prodigy device with an adapted tubing set TS310 to meet processing requirements. Using the protocol described in the schematic below (Figure 4) Miltenyi Biotec could successfully separate CD47 positive cells to high purity and viability without substantial cell loss. This investigation was performed using a surrogate cell population of CD47 positive HEK cells rather than hESC derived mesDA progenitors and although Miltenyi Biotec have not observed performance differences between pluripotent stem cell (PSC) derived CD47 positive mesDA Neurons and HEK cells, the HEK cells may still differ in the degree of CD47 expression and viability. Therefore, future work will investigate further separation runs with PSC derived mesDA neuron progenitors.

Scalability and Automation
Clean Bench manufacture of the mesDA differentiation protocol was performed in standard cell culture plasticware at RoslinCT. Specifically, small and large scale manufacture was performed in T25 cm2 and T75 cm2 tissue culture flasks respectively. However, there are several avenues for performing manufacture to greater scale. For example, larger flask/vessel formats such as multiple flasks of the same surface area or vessels with larger surface area or multiple layers such as CellStack cell chambers. In addition there is scope for scale-up and automation using the CliniMACS Prodigy.
As the clean bench mesDA progenitor manufacture at RoslinCT involved T75 cm2 tissue culture flasks, a proposed expansion strategy using this format is detailed in Table 5. If alternative plasticware (e.g. cell factories) or automated systems (e.g. CliniMACS Prodigy) are used for generating large quantities of CTP, it is advised to assess and compare the potency and identity of the CTP generated with the novel system with the CTP derived from the RoslinCT manufacturing runs to ensure comparability between systems.

5.1 Results
Sourcing appropriate hESC cellular material.
Testing the lines
When testing hESC lines for optimal mesDA neuronal progenitor differentiation, RC17 performed optimally where differentiated cells derived from this line expressed expected mesDA progenitor markers in vitro and engrafted when transplanted to in vivo PD animal models. Following these results, a characterised RC17 bank was generated by RoslinCT in collaboration with Isenet, and vials of this line were distributed to NSR consortium members for further optimisation of the differentiation protocol and initiation of TT procedures. Approved SOPs and CoAs for line culture and identity were provided to NSR consortium members with cell line dispatch. As it was found that optimal growth conditions, prior to initiation of differentiation, required transitioning the RC17 line to laminin 521 and iPS brew, future work must include the generation of a RC17 GMP grade bank grown under these transitioned conditions. These transitioned cells will act as a cellular raw material for the GMP differentiation process.

Technology transfer
Reagents
All reagents used within the hESC to mesDA progenitor differentiation protocol were selected based on a risk assessment of their properties and were xeno-free, human component free or chemically defined as far as possible. Only low risk items were included into the GMP translated protocol to ensure robust and reliable processing and patient safety. Due to the experimental nature of this novel differentiation strategy, many of the items were of research grade rather than manufactured in line with GMP standards. Where this occurred, reagents were selected on the basis of their manufacture being compliant with ISO 9001 or ISO 13485 standards. In addition, suppliers were asked to provide CoAs, Certificates of Origin, Certificate of Conformity, safety data and TSE declarations where appropriate. If a critical regent was not manufactured in a controlled manner but could not be sourced elsewhere, suppliers of these items were asked if they could make these products bespoke and in accordance with GMP or ISO standards.

This specifically relates to the R&D grade items from Miltenyi Biotec. These R&D items were risk assessed in line with the Miltenyi Biotec ISO accreditation / GMP documentation as supplied by Miltenyi Biotec to RoslinCT during the risk assessment process. However, these items are currently available at R&D grade only (manufactured without ISO accreditation). Miltenyi Biotec are in the process of investigating their ability to produce these components in line with ISO/GMP regulations as originally stated by Miltenyi Biotec upon initiation of the NSR project. It is expected that these product will be available in accordance with ISO 9001:2015 by end of 2018. A list of the reagents and consumables employed within this protocol and rationale for their use is described in Appendix 2.

However, the use of research grade items will impact on the final testing strategy for the CTP as a full GMP viral screen (final product) and in-process sterility and mycoplasma testing will be required. Impurities associated with research grade products may be mitigated via in vivo toxicology studies during pre-clinical investigation. In the event that suppliers cannot manufacture their research grade product to ISO or GMP standards, these items must be replaced with alternative products and a small pilot study must be performed to ensure that the differentiation process progresses as expected prior to initiating the GMP manufacturing runs. Alternative products, that were not available at project initiation, have been investigated by RoslinCT for suitability for the next phase of the project (data not shown).

Therapeutic Product Profile (TPP)
This document used to define the CTP for clinical application was compiled in parallel with the development of the mesDA progenitor differentiation protocol and was continually updated throughout the project in line with protocol refinement. There are several items that require completion such as the final product name, product labelling, product packaging, impurity testing and determination of virology testing on the final CTP. Completion of these sections are expected as part of the next phase of investigation.

Protocol refinement
TT of the mesDA differentiation protocol was successfully performed as a collaboration between Lund University and the manufacturing site, RoslinCT. After protocol refinement, three successful batches of mesDA progenitor cells (CTP) were manufactured using a standardised differentiation protocol and GMP risk assessed reagents. This protocol was performed concurrently with optimisation of cryopreservation and thaw strategies and with standardised QC sampling points for in-process controls and final product characterisation. To alleviate the pressures of cleanroom manufacturing, it is recommended that the mesDA progenitor differentiation process (if performed in accordance with SOP/RDP/112) is initiated on a Thursday so that CTP harvest (on Day 16) occurs on a Wednesday. This will ensure that key manufacturing steps are performed during the working week and ensure that supporting infrastructure is available for product manufacture and final CTP release. Manufacturing in accordance with robust manufacturing protocols (refer to Appendix 3) will ensure robust and reliable processing and streamline the route to the clinic.

QC Strategy
A QC strategy to characterise the CTP throughout the differentiation process (in-process controls) and to define the final CTP for release has been formulated for the mesDA differentiation protocol. This strategy is fully described in the TPP that accompanies this document and fundamental to ensure that manufacturing is progressing correctly and to allow standardised characterization of manufactured CTP. Miltenyi Biotec, using panels of protein markers for hESC cell pluripotency and mesDA differentiation, have designed a robust flow cytometry assay that can reliably determine the identity and purity of the CTP at different time points throughout the differentiation process. In addition, this assay is highly sensitive in the detection of contaminating pluripotent cells within the final CTP.



Optimisation of cryopreservation strategies
Two cryo-media formulations were investigated and compared during this study, CryoBrew 5% DMSO and NIM+5% DMSO. The following results were returned;
• Both formulations returned a cell viability of >85% even when harvested CTP was exposed to the cryo-medium for ≥180 min prior to cryopreservation.
• CTP cryopreserved under all described conditions (cryo-media exposure time and cryo-media type) showed >85% viability upon thaw.
• Therefore, harvested CTP can be held in both NIM+5% DMSO and CryoBrew 5% DMSO at 5±3˚C for up to 180 min from harvest prior to cryopreservation.
• There was no difference observed between cells re-plated from sampling points (0/180 and 180/0) from either cryopreservation media until Day 3 post re-plating. This has shown that harvested CTP remains viable and can recover upon thaw after being in formulation for 180 min during the cryopreservation event. Both NIM+5% DMSO and CryoBrew 5% DMSO formulations produced similar results.

These data show that the parameters from harvest to cryopreservation, in a controlled rate freezer (CRF), can be performed up to 180 min with no detrimental effect to cells pre/post cryopreservation using either type of cryopreservation media. As both cryo-medias are comparable, the selection of cryo-media may be based on the logistics of manufacturing and regulatory aspects rather than performance in culture.
Considering all elements of investigation, CryoBrew 5% DMSO is the recommended cryo-medium for CTP cryopreservation providing cryopreservation is performed within 180 min from formulation using a CRF and in accordance with CTD/REP/16/2. At present, this cryo-medium is available at R&D grade only and must be produced bespoke (Miltenyi Biotec) and in accordance with GMP regulations if used in the GMP manufacture of mesDA neuronal progenitors. This will allow ease of risk assessment and comply the regulatory standards for ATMP manufacturing.

Bed side stability Summary
This study was performed to investigate optimal parameters for thaw and CTP preparation prior to CTP transplantation within a clinical setting. Using thawed CTP preparations from NSRUST07-09, two hold media preparations (HBSS/0.05%Pulmozyme and KO-DMEM /0.05% Pulmozyme) and two hold temperatures (5±3˚C and 21±3˚C) were compared in regard to cell viability ≥10 hour incubation period. This incubation period was designed to mimic the expected time from thaw to transplantation within the clinic. After incubation, both hold media conditions showed comparable viability (>80%) where hold time (≤10h) or hold incubation temperature (5±3˚C and 21±3˚C) did not appear to affect cell viability.
In accordance with NSR consortium agreement and due to its simple composition, HBSS/0.05%Pulmozyme was chosen as the preferred hold media to assess the effect of hold time and hold temperature upon CTP survival when plated at low and high density and placed into culture for 5 days. This experiment was designed to provide an in vitro gauge of graft survival. The returned results showed the following;

• The viability of NSRUST07-09 cells, obtained on Days 0 and 5, remained high (>80%) and was consistent across time points and cell batches for all experimental conditions.
• Cell numbers obtained on Day 5, regardless of seeding density (4x105 cells/well and 1x105 cells/well), provided evidence that cell survival reduced with prolonged hold times. This trend was consistent across all batches.
• In addition, cell survival reduced further if cells were incubated for prolonged periods at 21±3˚C.
This study has shown that CTP, thawed within a clinical setting, can be held in HBSS/0.05% Pulmozyme hold media at 5±3˚C until transplantation. However, a time limit must be specified and must not exceed 4h at an incubation temperature of 5±3˚C to ensure graft survival. Other hold media formulations may be investigated further to prolong the hold time prior to the transplantation. However, this is out with the scope of this project.


Automation and scale up

The mesDA neuronal progenitor differentiation procedure within an automated, closed system to allow integrated cell processing from starting material (hESC) to final CTP was performed by Miltenyi Biotec using the CliniMACS prodigy and a bespoke tubing set for purification of CD47 positive cells. Although purification was successfully performed by Miltenyi Biotec, father investigation is required to evaluate the fully enclosed differentiation process and to investigate mesDA CTP rather than the surrogate HEK cell. The success of this investigation would provide advantages over a static flask method in terms of reduced production costs, reduced clean room classification for use, reduced operator skill, increased reproducibility, increased scalability and ease of TT to multiple manufacturing sites.

At present, scale up is expected to commence using static flasks with large surface areas rather than the CliniMACS Prodigy. Although the Prodigy offers an extremely attractive method for scale up, comparability studies must be performed to investigate both strategies and their effects on the returned CTP.
Potential Impact:
Chronic neurodegenerative debilitating diseases, such as Parkinson’s (PD) and Huntington’s (HD) disease constitute a major and increasing health problem and a heavy burden on the health care system in Europe. PD is the second most common neurodegenerative disease (after Alzheimer’s disease) affecting 1–2 percent of all individuals over the age of 50, and HD is the most common monogenetic disease of the nervous system with a prevalence of 3–5 individuals/100 000 in Europe. HD develops slowly over 10–20 years as a progressive devastating disease, eventually leading to severe disability and death. The pharmacological treatment available today is only palliative and of little help to the patients. In the case of PD, current pharmacological and neurosurgical therapies are effective in most patients, but they represent symptomatic treatments, which do not affect the long-term course of the disease. Thus, for both diseases, there is a tremendous need for new therapies that could modify the underlying disease processes and/or restore function in affected individuals.

Stem cell technologies hold the promise of taking cell transplantation all the way from a highly experimental procedure to a clinically applicable therapy for major neurodegenerative diseases. However, neuronal replacement in the adult brain poses a number of challenges with respect to donor cell generation, delivery and functional integration, and not all central nervous system (CNS) diseases may be equally suitable targets for such an approach. The best candidate diseases for neuronal transplantation are thought to be those where the leading cause of disability is linked to a defined, localized degeneration of neurons, such as the loss of midbrain dopaminergic (DA) neurons in PD and the early degeneration of striatal neurons in HD. Stem cell technologies have evolved at a very rapid pace during the last years and it is nowadays possible to produce human cells capable of variable degree of functional recovery in animal models of PD and HD. The main goal of this project was, therefore, to take human stem cells through the final steps towards their clinical application in cell replacement therapy for PD, and to further enhance functional cell replacement for the treatment of both, PD and HD. In that regard, our work also benefited from the Transeuro project (http://www.transeuro.org.uk/) which aims at developing an efficacious and safe cell replacement therapy methodology for PD using mesencephalic dopaminergic (mesDA) neurons from human ventral midbrain (VM) fetal tissue. These developments have facilitated the application of human stem cell derived-DA neurons developed by our Consortium, enabling their rapid and effective translation from the laboratory to clinical trials.

The project covered a significant number of crucial issues of basic, clinical and industrial research, as indicated below:

• The expansion of intermediate progenitors capable of generating fully differentiated and phenotypically stable mesDA neurons was a key goal of Neurostemcellrepair which was paralleled by efforts in developing better protocols for the production of medium spiny neuron (MSN) progenitors. Both goals were fulfilled as more homogenous and efficacious cell preparations have been obtained, while reducing the time necessary to differentiate cells to the desired stages for transplantation, especially for the mesDA neurons.

These advancements were based on the exploitation of new approaches to control gene expression and therefore progenitor cell identity during neural differentiation which include small molecules, the addition of midbrain developmental factors, synthetic RNAs (sRNAs) and long non-coding RNAs (LncRNAs). We have also established a new cellular platform designated as human induced neural progenitor cells (iNPCs) from cord-blood cells that will be exploited as novel source for human neurons.

In particular, with the objective of generating stable human mesDA A9 subtype neurons, we have applied midbrain developmental factors to promote the differentiation of human embryonic stem cells (hESC) and long term neuroepithelial stem (LtNES) cells into mesDA neurons of the A9/substantia nigra subtype. We were also able to determine the quality of human pluripotent stem cell (hPSC) -derived mesDA cells by single cell RNA-sequencing as compared to endogenous human midbrain tissue. The new protocol overcomes the critical issues of batch-to-batch variability, as this protocol is specifically designed to generate only caudalized VM mesDA progenitors, which we have shown through a large number of transplantations in adult rats to give rise to mature and functional mesDA neurons. This differentiation protocol can be used to produce high-purity cultures of mesDA progenitors to be used for transplantation studies, as well as for in vitro disease modeling studies such as drug screens, transcriptomic and proteomic analyses, gene therapy testing and other biomedical applications. We determined the quality of hPSC-derived mesDA cells and we also developed a method to quantitatively assess the fidelity of hPSC-derived DA neurons at a single-cell level. Thus, our study provides insight into the molecular programs controlling human midbrain development and provides a foundation for the development of cell replacement therapies.
As regards the generation of stable human MSN DARP32+ neurons, we found that striatal progenitors generated using our protocol and transplanted in HD animal models were able to survive, mature and integrate in the host tissue. We are now in the process of increasing the homogeneity and purity of the cell preparation and assessing the functional efficacy before engaging in the good manufacturing practises (GMP) transfer of this procedure.

• To develop strategies to control optimal donor cell adaptation in the host milieu. Toward this aim, controlled expression of key molecules involved in differentiation have been used to increase the maturation of the donor cells and enhance their connectivity. The project provided critical advancements in the engraftment and functional integration especially of human stem cell-derived mesDA neurons in animal models of PD and HD.
We generated defined cultures of mesDA progenitors which can survive transplantation to the adult PD brain and form mature and functional DA neurons in vivo. We assessed the in vivo performance of human embryonic stem cell (hESC)-derived mesDA cells and tested the minimum number of hESC-derived DA neurons that are sufficient for functional recovery in a rat model of PD. Potency assays were successfully employed, demonstrating that the minimum number of hESC-derived DA neurons that are sufficient to normalise rotational bias is similar to that of human VM-derived DA neurons thus greatly supporting the hypothesis that hESC-derived DA neurons generated using our protocol match the in vivo performance of human VM-derived DA neurons.
We have applied trans-synaptic rabies-based tracing technology to map connections between host neurons and transplanted human neurons in the grafts. Using two different experimental setups, we have been able to map both host-to-graft connectivity as well as graft-to-host connectivity of our transplanted mesDA cells. With this technology, we have shown that host neural circuitry, namely the natural striatal afferents, form synaptic contacts with grafted hESC-derived neurons that are established at 6 weeks post-transplantation and are stably maintained for at least 6 months. We also confirm that functional connectivity and circuit integration is established between grafted hESC-derived DAergic neurons and host neural circuits within a rat model of PD. This work represents a significant milestone in the pre-clinical development of human DA neurons derived from hESCs. In fact, we demonstrated robust survival of hESC-derived DA neurons for up to 6 months in a pre-clinical rat model of PD, which continue to provide functional recovery and extensive integration and connectivity during this time. Addressing the issues of long-term graft survival and connectivity are a key step in the clinical translation of hESC-derived DA neurons as provide us with a better understanding of graft behaviour post-transplantation in vivo and also address safety concerns.
Finally, using non-invasive, clinically compatible imaging techniques we have demonstrated that transplants of hESC-derived mesDA neurons do not proliferate at high rate, nor do they disturb the host parenchyma and display a highly neuronal phenotype.

Preliminary observations suggest that hESC-derived MSN progenitors can survive, mature and extend processes to area distal to the graft and located outside the striatal nuclei. Data suggest that grafted cells survive into the host striatum and start to express markers of striatal differentiation. Cells show mature morphology and phenotype, appearing integrated in the host tissue. While experiments at longer survival time-points are needed to confirm the capability of these cells to become functionally integrated and improve motor performances, these data are promising for future development of cell replacement therapies for HD.


• To develop strategies to improve the reprogramming of human fibroblasts into mesDA neurons and MSNs and their engraftment in models of PD or HD. In fact, direct cell fate conversion may provide an even faster and potentially in vivo applicable route for the generation of replacement cells. These approaches have been integrated in a comparative manner alongside classic stem cell-based differentiation paradigms.
We tested alternative strategies in which induced dopaminergic neurons (functional inducible DA neurons - iDANs) can be generated by direct conversion of human astrocytes in vitro or by direct reprogramming of mouse striatal astrocytes in vivo. We found that direct in vivo reprogramming of endogenous striatal astrocytes into iDANs may provide an alternative to cell replacement therapy for PD. Such alternative would not require cell transplantation or immunosuppression. We suggest that further optimization of this approach may enable clinical therapies for PD by delivery of reprogramming genes rather than cells.

• To provide novel and urgently needed procedures and tools for detailed analysis of engrafted donor cells in the host tissue. Our project had a strong component of technological development, which also relates to monitoring of grafted stem cells with innovative microscopic imaging technology.
In order to optimise the survival, integration and functionality of novel cell therapies for both PD and HD, we addressed four key themes: 1) defining optimal cell placement and potency, 2) establishing behavioural tasks to reveal changes in relevant domains, 3) utilising targeted training protocols to enhance recovery and 4) developing novel imaging technologies to visualise the integration and connectivity of neurons. We have collected a large data set addressing each of these themes and have successfully achieved our aim of demonstrating the survival, integration, safety and functional efficacy of hESC-derived dopamine grafts for PD.
For Huntington’s disease, we have established novel behavioural tasks sensitive to striatal dysfunction and developed highly sensitive imaging platforms for detecting subtle changes in graft-host connectivity. Therefore the preclinical study of these therapies is now more refined and able to detect optimal vs suboptimal cell therapy products.
Finally, we developed an optimized image acquisition platform facilitating the fast and high-resolution imaging of specimens up to the size of a rat brain. The developed light-sheet fluorescence microscopy (LSFM) system was tailored for long optical working-distances and is thus suitable for the analysis of neural grafts and their connectomes. The acquisition platform permits whole-mount visualization of cleared mouse and rat brains at single cell resolution. We used this approach for quantitative assessment of human transplant innervation in the context of an entire mouse brain. We show that we were able to successfully reconstruct in 3D EGFP/mRFP fluorescently labelled cell grafts. Moreover, we demonstrate a whole-mount 3D reconstruction of the neural graft and its axonal innervation by host neurons. Collectively, our results suggest that establishment of afferent and efferent connections between host brain and grafted human neural stem cells is largely dictated by the topology of pre-existing, endogenous fibre tracts and circuitries rather than the regional identity of the donor cells. We explored if this technology may be used in the context of compound-based enhancement of host-graft connectivity. After optimisation of the technique, we demonstrated suitable quantification sensitivity to detect differences in transplant-host connectivity between treated and untreated mice.

• The development of second-generation and clinical-grade protocols for the differentiation of mesDA neurons from human pluripotent stem cells in GMP conditions was a major objective of the project. We worked toward this objective not only by incorporating new factors and improving the efficiency and quality of the cells we produce, but most importantly by engaging to adopt and standardise best in practice protocols and securing appropriate transfer and implementation of our protocols in GMP in order to deliver GMP-translated manufactured DA neurons by the end of project.

Neurostemcellrepair aimed at closing the gap between development and clinical implementation of cell replacement therapies for PD and to advance also in HD.
During the course of our project we have succeeded in adapting and optimising our research-grade mesDA differentiation protocol into a GMP-compatible protocol with high efficiency, purity and reproducibility. We have validated the clinical value of this protocol through extensive in vivo testing of transplanted progenitors in rodent PD models, and we found that the cells are non-tumorigenic and display predictable functionality in rat models of PD. This protocol produces a high yield of DA progenitors, thereby enabling manual production of a large number of cells for clinical trial. We also assessed the in vivo reproducibility of optimised GMP-compatible protocol and succeeded in generating a reproducible, high-yield GMP protocol as well as a cryopreservation and transplantation strategy which will enable us in proceeding towards clinical translation of our hESC-derived mesDA stem cel product for PD.
We also successfully identified critical stages of the manufacturing process including Quality Control (QC) points (in process and final cell therapy product -CTP) as well as optimal parameters for CTP cryopreservation and stability prior to transplantation within a clinical setting. Concurrent to the GMP translation work, we adapted the differentiation protocol to incorporate use of the CliniMACS Prodigy with an aim to fully enclose and automate the differentiation system. Concurrently, a cryopreservation study was performed to optimise parameters for CTP cryopreservation to assess optimal thaw and formulation prior to transfusion. A bedside stability study was performed to optimise thaw conditions and reconstitution conditions prior to transplantation that could be applied within a clinical setting.

By taking advantage of these discoveries and the acquired expertise in the field of regenerative medicine, Neurostemcellrepair largely contributed to close the gap between development and clinical implementation of cell replacement therapies for PD and HD.
List of Websites:
http://www.neurostemcellrepair.org