It has become increasingly evident that heterogeneity between hiPSC lines constitutes an enormous source of inter-experimental variability and lack of robustness. Genetically matched lines solve this problem and allow for “personalized models” to study NDs. Using CRISPR/Cas9, we have recently generated two cohorts of isogenic hiPSC lines to model SMA and familial forms of ALS (fALS). A mild type III, an intermediate type II and a severe type I SMA lines were targeted and the mutation in exon7 of SMN2 gene (responsible for exon7 skipping and the drastic reduction in functional SMN protein production) was corrected (Figure 1A). Further, we have generated a second set of isogenic SMA lines carrying the fluorescent protein clover (stronger than GFP) tagging the C- of full-length SMN protein. For the generation of the fALS isogenic cohort, highly penetrant mutations in SOD1 and TARDBP (coding for TDP43) were introduced in a healthy control hiPSC line (Figure 1B). Importantly, our group also has patient iPSC lines carrying these same fALS mutations. The comparison between the two types of lines will enable us to determine whether the potential alternations found in the patient-derived cultures are due to the disease-causing mutation or to additional and unnoticed genomic abnormalities of that specific patient line.

While the main focus of our group is to understand selective neuronal vulnerability in MNDs, this compelling topic is also observed in the other most common NDs, such as AD, that we also study. For that reason, we have also generated an isogenic familial AD model. By introducing single nucleotide mutations in the APP locus in a healthy control hiPSC line, we have produced a cohort of four lines, two carrying different negative mutations that increase the susceptibility of developing familial AD (called “AV” and “SWE”), one carrying a protective mutation against the disease (named “AT”) and a APP KO line (Figure 1C).

To live-track and isolate the specific neuronal types most highly affected in MNDs or AD, we have generated additional SMA, fALS and fAD lines expressing the fluorescent reporter mNeonGreen under the control of the main MN transcription factor ISLET1 or the cortical neuron transcription factor CTIP2. These tools are now enabling us to study live MN and cortical neuron specification, maturation and survival in a ND context.

Recent studies using direct programming of pluripotent stem cells (PSCs) into different neuronal types have started to elucidate neuronal distinct sensitivities to disease. However, this approach drives the differentiation of PSCs into postmitotic neurons in an extremely fast manner, skipping the different progenitor stages that normally arise during development and which might have a say in the appearance of the disease. Therefore, the direct differentiation of PSCs into neurons, skipping those neurodevelopmental intermediates, could hamper the discovery of the mechanisms that trigger the pathology. Additionally, this direct programming results in a homogenous population of cells. It is currently evident how useful human 3D in vitro models have become to learn about brain development and disease, and their potential as tools to develop precision medicine. However, extensive work still needs to be done to propel SCO technology, since, to date, they have been notably less pursued than cortical organoids. In our team, we also use SCO protocols modified from previous studies that have been used in the field for years. These yield a variety of MNs that suffices to tackle some of the questions we aim to answer. However, those protocols are ineffective at generating the wide rostro-caudal MN diversity that can be found in vivo.

To overcome these limitations, we have developed a more physiologically relevant, human SCO protocol that represents an important methodological development in terms of both reproducibility and developmental patterning. Following developmental cues that drive the patterning of the neural tube along the rostro-caudal axis and the specification of the spinal cord neurons, we have performed a pilot screen of twelve different differentiation paradigms on a healthy control hiPSC line. Using gene expression analysis and immunohistochemistry as readouts, we have identified a protocol condition that enabled the simultaneous generation of neuroectodermal and neuromesodermal progenitor pools that have the potential to differentiate into MNs of the entire rostro-caudal spinal cord axis as well as into skeletal and smooth muscle cells. Thus, using this protocol, our SCOs are composed by different MN and neuronal types and muscle cells, developed intertwined and connected by neuromuscular junctions (NMJs).