Investigating early functional alteration in a human iPSC-based model of Parkinson’s Disease

Abstract

Dopaminergic neurons (DAn) were efficiently differentiated from 6 different iPSC lines derived from a 2 control, 3 PD lines, and an isogenic PD line. We characterize the culture at different time point in order to verify the composition of the culture. At day 35 almost 50% of the cells expressed the neuronal marker TUJ1, of which 20% were of DA lineage as judged by TH expression. No astrocytes were found in the culture. Interestingly at day 50 of differentiation the amount of DAn increased up to 25-35% of which 35% were expressing Girk2, a midbrain DA neuronal marker that was not expressed at day 35 and 45-60% of the DA neurons express FOXA2, another important transcription factor that confirm the midbrain fate. Even more importantly, at day 80 of the differentiation process, the amount of DAn that expressed Girk2 increased up to 50%, while the expression of FOXA2 in DA remain stable, confirming the ventral midbrain phenotype. We then test for each line the neuronal activity by calcium imaging assay. Comparing the two groups of controls and isogenic PD line versus PD lines, interestingly we identify two distinct patterns of activity: controls lines display a mixed mode, oscillatory activity normally associated with healthy networks while PD lines display a two-state dynamics, with strong bursting combined with intervals of almost no activity that suggest an impairment in the communication between the neurons. These dynamic differences suggest a more in depth analysis of the functional network that controls and PD create. Using a custom algorithm we demonstrate that just the PD lines were functionally impaired because their neurons, especially at day 80, were not able to form a homogeneous network like the controls one that can be described as a scale-free like systems. Analyzing separately TH and non TH neurons we were able to conclude that the functional connectivity of PD1 lines shows a higher departure from the controls lines along maturation, which indicates poor information flow efficiency. Particularly, PD TH+ neurons connectomes display abnormal network organization that occurs primarily before the general alteration of the network, suggesting that TH+ are leading to general neural connectome alteration. Due to the biophysical simulation analysis we were able to identify as a cause of the functional impairment the reduction in the neurite arborization TH specific just in the PD lines and confirm this phenotype in our biological samples. We then used our in vitro model to take a step backwards and examine the biological and molecular behavior before the functional alteration manifests. We analyzed the culture at D50, when the data suggests that the functional alteration has not yet developed fully, using gene expression profile analysis to identify possible deregulations in pathways that can be connected to the altered functionality. Bioinformatics analysis focused on differentially expressed genes, selected with a pAdjValue of 0.05 and a fold change ≤-2 and ≥ 2. Within these strict selection criteria, we were unable to highlight any gene related to LRRK2 PD and isogenic PD. This confirms the validity of the in vitro model and the robustness of the differentiation protocol and shows that the functional phenotype is not due to macroscopic neurodegenerative conditions.

La enfermedad de Parkinson (EP) es una enfermedad incurable, crónica y progresiva que conduce a la invalidez prematura y la muerte. Se espera que el diagnóstico temprano de la EP mejore dramáticamente el resultado de las terapias actuales. Para ello, empleamos un modelo de EP basado en células neuronales humanas, para detectar alteraciones funcionales tempranas que nos den un diagnostico de la enfermedad antes de la aparición de síntomas motores. Utilizando neuronas derivadas de células madre pluripotentes inducidas (iPSC) de individuos sanos y pacientes con EP asociados con la mutación familiar en el gen LRRK2, hemos comprobado que los dos grupos forman redes complejas y muestran signos evidentes de maduración funcional a lo largo del tiempo. Sin embargo, las redes neuronales de la EP desarrollaron una híper-sincronía anormal, en comparación con las redes de los controles y de la línea isogénica de LRRK2. En esto estudio combinamos análisis de la actividad neuronal a lo largo de tiempo utilizando la técnica del “calcium imaging”, un modelo in silico de “network”, líneas reporteras de neuronas dopaminérgicas, y el análisis del perfil de expresión génica. Con estos experimentos, encontramos que una disminución en la longitud de la neurita de neuronas dopaminérgicas es, entre otras causas, una de las primeras alteraciones funcionales presentes en la red de neuronas derivada de EP. Por lo tanto, nuestros resultados identifican alteraciones tempranas en la función neuronal de la EP que son anteriores al inicio de la degeneración neuronal, resaltando la extraordinaria ventaja que ofrece este modelo de iPSC en la evaluación pre sintomática de las enfermedades degenerativas crónicas.