High-throughput microscopy exposes a pharmacological window in which dual leucine zipper kinase inhibition preserves neuronal network connectivity

Abstract

Therapeutic developments for neurodegenerative disorders are redirecting their focus to the mechanisms that contribute to neuronal connectivity and the loss thereof. Using a high-throughput microscopy pipeline that integrates morphological and functional measurements, we found that inhibition of dual leucine zipper kinase (DLK) increased neuronal connectivity in primary cortical cultures. This neuroprotective effect was not only observed in basal conditions but also in cultures depleted from antioxidants and in cultures in which microtubule stability was genetically perturbed. Based on the morphofunctional connectivity signature, we further showed that the effects were limited to a specific dose and time range. Thus, our results illustrate that profiling microscopy images with deep coverage enables sensitive interrogation of neuronal connectivity and allows exposing a pharmacological window for targeted treatments. In doing so, we revealed a broad-spectrum neuroprotective effect of DLK inhibition, which may have relevance to pathological conditions that ar.e associated with compromised neuronal connectivity.

Keywords: Antioxidant depletion; Calcium imaging; High-content screening; Neurodegeneration; Neuronal connectivity; Neuronal network; Synapse; hTau.P301L.

Fig. 1

General workflow. Workflow of the microscopy-based pipeline. Hippocampi and/or cortices from WT E18 C57Bl6 mouse embryos were dissected, after which cell suspensions were created and seeded in 96-well plates. Cultures that would be used in the functional assay were transfected with an AAVDJ-hSyn1-GCaMP6f-nls-dTomato vector on 0 DIV. At 3 DIV, cultures were treated with arabinosylcytosine (AraC) to suppress the excessive growth of non-neuronal cells. Cultures were treated every three or four days. At fixed time points, cultures were either fixed and subjected to immunostaining, or used for live cell imaging to assess the morphological and functional characteristics, respectively

Fig. 2

Culture age correlates with morpho functional changes. a Representative images of cortical cultures after fixation and immunocytochemistry at 6, 12, 18, 24, 30, 36, 42 and 48 DIV; b Dendrite network density and nodes increased gradually until 36 DIV, after which they decreased. The deterioration of the dendrite network was less pronounced when normalized to the number of neurons, which decreased over time. The presynaptic density and synapse density increased during the first 18 DIV. The density of postsynaptic spots increased gradually. The ratio between neuronal and non-neuronal nuclei remained more or less the same until 36 DIV (n_bio = 2 x n_tech = 6); c Representative immunocytochemistry images and Ca²⁺- traces (of 10 neurons) for a more resolved experiment (3, 7, 10, 14, 18 DIV) of primary hippocampal and cortical cultures; d Similar trends are observed in selected morphological descriptors for hippocampal and cortical cultures (n_bio = 3 x n_tech = 6). Synchronous activity increased from DIV 10 onwards (n_bio = 3 x n_tech = 6)

Fig. 3

Cultures can be clustered and classified based on their morphofunctional signature. a Hierarchical clustering based on z-scores of all descriptors. The correlation of the descriptors with culture age (DIV) is indicated (black/orange color coded) as well as the descriptor set to which they belong; b PCA based on morphological descriptors distinguished hippocampal from cortical cultures. In a biplot of the first two principal components, data points cluster according to culture time and follow distinct temporal trajectories along the direction of decreasing nuclear count and increasing dendrite- and synapse density (n_bio = 3 x n_tech = 6). A biplot of PCA on functional descriptors could only differentiate early (3–10 DIV) from later time points (14–18 DIV) (n_bio = 3 x n_tech = 6) (Ellipses show the 67% confidence interval); c Confusion matrices and classification results using a random forest classifier. The misclassification rate (MCR) of the classification based on morphological descriptors was 5% (n_bio = 3 x n_tech = 6) and much lower than the MCR of the classification based on functional data (45%) (n_bio = 3 x n_tech = 6); d The z-scores of a subset of descriptors were multiplied with their respective correlation with culture age (weighted z-scores) and summed to obtain a connectivity score that allows facile interpretation of the degree of connectivity at a given time point (n_bio = 3 x n_tech = 6)

Fig. 4

Focused compound assay identifies DLK as positive modulator of neuronal connectivity. a Connectivity score map after normalization to culture age-matched controls. Rapamycin is a strong negative modulator of neuronal connectivity and GNE3511 exerts a lasting positive effect on neuronal connectivity. (Morph.: n_bio = 3 x n_tech = 5 - Func.: n_bio = 3 x n_tech = 6, except for GNE3511: Morph.: n_bio = 2 x n_tech = 6 - Func.: n_bio = 2 x n_tech = 9); b Example images of control cultures and cultures treated with 0.1 μM rapamycin or GNE3511 (14 DIV) (I, II and III in panel a); c Z-scores showing the effect of 0.1 μM rapamycin or GNE3511 on dendrite density, synapse density, neuronal count (rapamycin: n_bio = 3 x n_tech = 5, GNE3511: n_bio = 2 x n_tech = 6) and correlation of calcium bursts (rapamycin: n_bio = 3 x n_tech = 6, GNE3511: n_bio = 2 x n_tech = 9)

Fig. 5

DLK inhibition can partially prevent the age-related loss of connectivity in culture. a Connectivity score map showing that chronic treatment with 0.01 μM GNE3511 could prevent the age-related loss of connectivity. Other effects were not lasting or neurotoxic (Morph.: n_bio = 1 x n_tech = 6 - Func.: n_bio = 1 x n_tech = 9); b Z-scores of example descriptors (Morph.: n_bio = 1 x n_tech = 6 - Func.: n_bio = 1 x n_tech = 9)

Fig. 6

DLK inhibition has neuroprotective potential in cultures deprived from antioxidants. a Primary cultures deprived from antioxidants (-AO), display impaired neuronal network connectivity at DIV 14 (Morph.: n_bio = 2 x n_tech = 6 - Func.: n_bio = 2 x n_tech = 9); b Normalized connectivity scores of cultures deprived of antioxidants show that treatment from 0 DIV onwards with DLK inhibitors (GNE3511 and GNE8505) could prevent connectivity loss at 14 DIV (excl. Highest concentration) (Morph.: n_bio = 2 x n_tech = 6 - Func.: n_bio = 2 x n_tech = 9); c Representative images of control cultures, -AO cultures and -AO cultures treated with 1 μM GNE8505 (I, II and III in panel b); d Bar plots showing z-scores normalized to -AO cultures of dendrite density, synapse density and neuronal count (n_bio = 2 x n_tech = 6), as well as the absolute number of correlated calcium bursts (n_bio = 2 x n_tech = 9). Significant differences compared to control cultures or cultures deprived from antioxidants are indicated (p < 0.05, pairwise Wilcoxon test with Bonferroni correction)

Fig. 7

DLK inhibition has both neuro-protective and -restorative potential in cultures overexpressing hTau.P301L. a Cultures overexpressing hTau.P301L, showed impaired neuronal network connectivity (Morph.: n_bio = 2 x n_tech = 6 - Func.: n_bio = 2 x n_tech = 9); b Continuous treatment with DLK inhibitors prevented loss of neuronal connectivity of the hTau.P301L model according to the normalized connectivity scores (Morph.: n_bio = 2 x n_tech = 6 - Func.: n_bio = 2 x n_tech = 9); c Representative images of control cultures, cultures overexpressing hTau.P301L and hTau.P301L cultures treated with 1 μM GNE8505 (I, II and III in panel b); d Bar graphs showing z-scores of morphological descriptors (n_bio = 2 x n_tech = 6) as well as the absolute number of correlated calcium bursts (n_bio = 2 x n_tech = 9). Significant differences compared to control cultures or cultures overexpressing hTau.P301L are indicated (p < 0.05, pairwise Wilcoxon test with Bonferroni correction); e DLK inhibition rescues cultures overexpressing hTau.P301L after connectivity loss had already started (Morph.: n_bio = 1 x n_tech = 6 - Func.: n_bio = 1 x n_tech = 6); f Representative images of control cultures, cultures overexpressing hTau.P301L and hTau.P301L cultures treated with 1 μM GN3511 or 10 μM GNE8505 (IV, V, VI and VII in panel e); g Bar graphs showing z-scores of morphological descriptors (n_bio = 1 x n_tech = 6) as well as the absolute number of correlated calcium bursts (n_bio = 1 x n_tech = 6). Significant differences compared to control cultures or cultures overexpressing hTau.P301L are indicated (p < 0.05, pairwise Wilcoxon test with Bonferroni correction)