Transcriptomic and metabolic signatures of neural cells cultured under a physiologic-like environment

Abstract

Cultured brain cells are used conventionally to investigate fundamental neurobiology and identify therapeutic targets against neural diseases. However, standard culture conditions do not simulate the natural cell microenvironment, thus hampering in vivo translational insight. Major weaknesses include atmospheric (21%) O₂ tension and lack of intercellular communication, the two factors likely impacting metabolism and signaling. Here, we addressed this issue in mouse neurons and astrocytes in primary culture. We found that the signs of cellular and mitochondrial integrity were optimal when these cells were acclimated to grow in coculture, to emulate intercellular coupling, under physiologic (5%) O₂ tension. Transcriptomic scrutiny, performed to elucidate the adaptive mechanism involved, revealed that the vast majority of differentially expressed transcripts were downregulated in both astrocytes and neurons. Gene ontology evaluation unveiled that the largest group of altered transcripts was glycolysis, which was experimentally validated by metabolic flux analyses. This protocol and database resource for neural cells grown under in vivo-like microenvironment may move forward the translation of basic into applied neurobiological research.

Keywords: astrocyte; energy metabolism; glycolysis; hypoxia; neuron; transcriptomics.

Figure 1

Incubation of astrocytes and neurons cultured under physiologic-like conditions induce changes in cell viability and mitochondrial fitness.A, strategy used to incubate astrocytes and neurons either in monoculture or coculture at different pO₂. B and C, apoptotic cell death (B) and mitochondrial membrane potential (Δψ_m) (C) in primary astrocytes and neurons grown in monoculture at atmospheric (21%), 6, 5, 4, 3 or 2% pO₂ for 6 days. Data are mean ± SD. p value is indicated; n = 3 to 40 independent culture preparations; two-way ANOVA followed by Tukey. D–F, apoptotic cell death (D), (Δψ_m (E) and mitochondrial ROS (F) in primary neurons grown for 6 days in monoculture, or for 3 days in monoculture plus three further days in coculture with astrocytes, at atmospheric (21%) or 5% pO₂. Data are mean ± SD. p value is indicated; n = 4 to 25 (D), n = 9 to 16 (E) or n = 9 to 13 (F) independent culture preparations; multiunpaired t test, two-sided. G–I, apoptotic cell death (G), (Δψ_m (H) and mitochondrial ROS (i) in primary astrocytes grown in monoculture or coculture with neurons, at atmospheric (21%) or 5% pO₂. Data are mean ± SD. p value is indicated; n = 4 to 14 (G), n = 8 to 17 (H) or n = 8 to 15 (I) independent culture preparations; multiunpaired Student’s t test, two-sided. pO₂, partial oxygen tension; ROS, reactive oxygen species.

Figure 2

Transcriptomic analysis in astrocytes and neurons cultured under physiologic-like conditions suggests glucose energy metabolism and differentiation pathways reprogramming.A–D, PCA analysis of the transcriptome of astrocytes grown for 6 days in monoculture (A), or for 3 days in monoculture plus three further days in coculture with neurons (B); and neurons grown for 6 days in monoculture (C), or for 3 days in monoculture plus three further days in coculture with astrocytes (D), at atmospheric (21%) or 5% pO₂. E–H, transcripts altered in cells grown at 5% versus at 21% pO₂ in mono- (E, G) or coculture (F, H). Data are mean log₂ expression levels and -log₁₀ FDR values from n = 3 independent culture preparation for each condition; Empirical Bayes method followed by Benjamini-Hochberg correction. I, functional annotations of the transcripts most significantly altered in neurons grown at 5% versus at 21% pO₂ in coculture with astrocytes. Fisher’s Exact test followed by Benjamini-Hochberg correction. FDR, false discovery rate; PCA, perchloric acid; pO₂, partial oxygen tension.

Figure 3

Functional analysis reveals glycolysis and differentiation adaptations in astrocytes and neurons cultured under physiologic-like conditions.A and B, rates of glycolytic flux in neurons (A) or astrocytes (B) grown in monoculture for 6 days, or in monoculture for 3 days plus three further days in coculture, at 21 or 5% pO₂. Data are mean ± SD. p value is indicated; n = 10 to 15 (A), n = 7 to 13 (B) independent culture preparations; multiunpaired t test, two-sided. C and D, PFK1 activities in neurons (C) or astrocytes (D) grown in monoculture or coculture at 21 or 5% pO₂. Data are mean ± SD. p value is indicated; n = 8 to 13 (C), n = 12 to 15 (D) independent culture preparations; multiunpaired t test, two-sided. E, lactate production in neurons or astrocytes grown in monoculture or coculture at 21 or 5% pO₂. Data are mean ± SD. p value is indicated; n = 3 to 6 independent culture preparations; multiunpaired t test, two-sided. F, glucose consumption in neurons or astrocytes grown in monoculture or coculture at 21 or 5% pO₂. Data are mean ± SD. p value is indicated; n = 4 to 11 independent culture preparations; multiunpaired t test, two-sided. G, cell cycle analysis in neurons cocultured with astrocytes at 21 or 5% pO₂ for 3 days. Data are mean ± SD. p value is indicated; n = 4 to 5 independent culture preparations; multiunpaired t test, two-sided. H, GFAP staining in astrocytes grown in monoculture or coculture with neurons at 21 or 5% pO₂. Data are mean ± SD. p value is indicated; n = 4 to 5 independent culture preparations; multiunpaired t test, two-sided. I, Tuj1 staining in neurons grown in monoculture or coculture with astrocytes at 21 or 5% pO₂. Data are mean ± SD. p value is indicated; n = 3 to 4 independent culture preparations; multiunpaired t test, two-sided. The scale bar represents 25 μm. GFAP, glial fibrillary acidic protein; PFK1, 6-phosphofructo-1-kinase; pO₂, partial oxygen tension.

Figure S1