Modeling anorexia nervosa: transcriptional insights from human iPSC-derived neurons

Abstract

Anorexia nervosa (AN) is a complex and multifactorial disorder occurring predominantly in women. Despite having the highest mortality among psychiatric conditions, it still lacks robust and effective treatment. Disorders such as AN are most likely syndromes with multiple genetic contributions, however, genome-wide studies have been underpowered to reveal associations with this uncommon illness. Here, we generated induced pluripotent stem cells (iPSCs) from adolescent females with AN and unaffected controls. These iPSCs were differentiated into neural cultures and subjected to extensive transcriptome analysis. Within a small cohort of patients who presented for treatment, we identified a novel gene that appears to contribute to AN pathophysiology, TACR1 (tachykinin 1 receptor). The participation of tachykinins in a variety of biological processes and their interactions with other neurotransmitters suggest novel mechanisms for how a disrupted tachykinin system might contribute to AN symptoms. Although TACR1 has been associated with psychiatric conditions, especially anxiety disorders, we believe this report is its first association with AN. Moreover, our human iPSC approach is a proof-of-concept that AN can be modeled in vitro with a full human genetic complement, and represents a new tool for understanding the elusive molecular and cellular mechanisms underlying the disease.

Figure 1

Generation and characterization of iPSCs from AN patients. (a) Patient profile summary. (b) Schematic view and representative images showing the morphological changes observed during the primary fibroblast cell reprogramming, neural induction and differentiation processes. Scale bars represent 200 μm. (c) Representative immunofluorescence images illustrating the expression of pluripotency markers in the generated iPSCs, including OCT4, NANOG, TRA-1–60 and LIN28. Scale bar represents 100 μm. (d) Representative images of G-banding karyotype analysis from cell chromosomes demonstrating the genetic stability of iPSCs; no karyotypic abnormalities were induced by the reprogramming process. (e) Representative images of hematoxylin and eosin staining of teratomas derived from iPSCs showing tissues from the three germ layers. Scale bar represents 100 μm. (f) Expression of pluripotency and three germ layer markers in iPSCs and EBs, respectively, assessed by RT-PCR (OCT4, NANOG and LIN28—pluripotency; AFP—endoderm; MSX1—mesoderm; PAX6—ectorderm). The H9-hESC was used as a control for pluripotency and differentiation capability; B2M was used as reference gene. (g) Cluster analysis showing correlation coefficients of RNA-seq transcripts from iPSCs and hESC, and a distinguished gene expression profile from primary fibroblast cells (FIBRO). A panel of human pluripotency-related genes (isoform level; Supplementary Table S3) was considered. (h) Heatmap and hierarchical clustering-based dendrogram of hESC, iPSCs and fibroblasts for AN and control samples. Considering the entire cellular transcriptome expression profile of evaluated cells, two subgroups were identified: iPSCs, with a molecular signature similar to that exhibited by hESCs, and fibroblasts with a completely different expression profile. In g and h, colors indicate the range of each gene's expression, with least expression shown in red and highest expression shown in green. AN, anorexia nervosa; anxiety, patient showed/treated for anxiety; CTL, control (unaffected individual); famhx-ED, first, second or third degree relative with a history of an eating disorder; fi, fibroblast; iPSC, induced pluripotent stem cell; MDD, major depressive disorder; neu, neurons; NPC, neural progenitor cell; OCD, obsessive compulsive disorder; psychotropic, patient was prescribed at least one psychotropic medication.

Figure 2

Derivation of neural progenitor cells and neurons from AN-iPSCs. (a) Representative images showing that the neural progenitor stage-specific markers NESTIN and SOX2 are expressed by AN iPSC-derived NPCs. Scale bar represents 100 μm. (b) Cluster analysis showing that after 4 weeks of differentiation neurons (NE) exhibit a molecular signature distinct from that of their iPSC counterparts. Colors indicate the range of each gene's expression, with least expression shown in red and highest expression shown in green. (c) Gene expression changes observed during the differentiation process measured by qRT-PCR using stage-specific markers for iPSCs (OCT4), NPCs (NESTIN) and neurons (MAP2). The expression levels of each gene were quantified, normalized to B2M (reference gene), and the results are presented as mean±s.e.m. (n⩾8 for each group). (d) Representative immunofluorescence images of cells after neuronal differentiation. IPSC-derived neural cultures express neuronal (MAP2, NEUN) and glial (GFAP) markers, together with specific cortical proteins (CTIP2). Excitatory (VGLUT1 and SYN1) and inhibitory (GAD65-67 and GABA) neuronal proteins are also observed in the generated neural population. The presence of LMX1A and FOXA2 among the neuronal cells, although low, is an evidence of the dopaminergic neuronal fate. Scale bar represents 100 μm. (e) Quantification of the percentage of MAP2⁺ (neuron), VGLUT1⁺ (glutamatergic), GABA⁺ (GABAergic) and GFAP⁺ (glia) labeled cells is presented as mean±s.e.m. (n⩾8 for each group). (f) Representative western blotting of control and AN-derived neural proteins that were lysed and immunoblot for neuronal (TUJ1, VGLUT1, GAD65-67, PSD95, SYN1 and TH) and glial (GFAP) markers, along with the dopamine transporter (DAT); β-ACTIN was used as housekeeping control (reference). (g) Quantification of proteins in AN and control neural cultures assessed by Western blot analysis; β-ACTIN was used for normalization (n=8 for each group). No differences were observed between control and affected samples (P<0.05, Student's t-test). AN, anorexia nervosa; iPSC, induced pluripotent stem cell; NPC, neural progenitor cells.

Figure 3

Transcriptional analysis of AN iPSC-derived neuronal cultures. (a) Heatmap and the hierarchical clustering-based dendrogram of samples after 4 weeks of differentiation. No significant differences are observed between iPSC-derived neurons (NE) from AN patients and controls. (b) Heatmap and the hierarchical clustering-based dendrogram displaying the transcriptional pattern of genes associated with neural development and differentiation in AN and control neural cultures (list of genes at Supplementary Table S5). (c) Heatmap and hierarchical clustering-based dendrogram of genes with minimum fold-change variation of 2 and false discovery rate (FDR)-adjusted P-value<0.01 between AN and control neurons. (d) Neuronal specific cell type-based clustering analysis of AN and control samples using 13 selected candidate genes most differentially expressed between the two subgroups (P<0.01). In all heatmaps (a–d), colors indicate the range of each gene's expression, with least expression shown in red and highest expression shown in green. (e) Validation of 13 selected candidate genes from the RNA-seq analysis by qRT-PCR. Downregulated (red histogram) or upregulated (green histogram) genes differentially expressed between AN and control neurons. Independent neuronal cultures generated from the same clones used in the RNA-seq analysis were used in this validation; GAPDH was used as reference gene. Error bars are represented by standard deviation. (f) Topological structure of AN-PIN and random-PIN showing that AN-PIN is denser and with smaller shortest path than random-PIN (Kolmogorov–Smirnov test, P<0.01). (g) Enriched GO functional pathways terms found in neurons derived from AN patients (the complete list is found in the Supplementary Table S8). (h) Volcano plot of PCR array analysis for human neurotransmitter receptors. Plot illustrates that although control and AN-derived neurons do not show significant differences in expression for estrogen receptors and dopamine/serotonin neurotransmitter systems constituents, the TACR1 gene is upregulated in AN neurons (2.0-fold differential expression between the groups at P<0.05; Student's t-test). (i) Upper panel: representative western blotting of control and AN-derived neural proteins that were lysed and immunoblot for TACR1; β-ACTIN was used as reference. Bottom panel: quantification of TACR1 protein in AN and control neural cultures assessed by Western blot analysis. Increased levels of TACR1 were observed in AN (P<0.05, Student's t-test). Proteins were detected using Odyssey CLx infrared imaging system. (j) BrainSpan analysis of the TACR1 gene in brain regions from striatal networks. RNA-seq RPKM (reads per kilobase per million) identified during the different stages of human brain development. ACC, anterior cingulate cortex; AN, anorexia nervosa; iPSC, induced pluripotent stem cell; MNT, mediodorsal nucleus of thalamus; OFC, orbital frontal cortex; STR, striatum; VPC, ventrolateral prefrontal cortex.