Transcriptomic signatures of treatment response to the combination of escitalopram and memantine or placebo in late-life depression

Abstract

Drugs that target glutamate neuronal transmission, such as memantine, offer a novel approach to the treatment of late-life depression, which is frequently comorbid with cognitive impairment. The results of our recently published double-blind, randomized, placebo-controlled trial of escitalopram or escitalopram/memantine in late-life depression with subjective memory complaints (NCT01902004) indicated no differences between treatments in depression remission, but additional benefits in cognition at 12-month follow-up with combination treatment. To identify pathways and biological functions uniquely induced by combination treatment that may explain cognitive improvements, we generated transcriptional profiles of remission compared with non-remission from whole blood samples. Remitters to escitalopram compared with escitalopram/memantine combination treatment display unique patterns of gene expression at baseline and 6 months after treatment initiation. Functional enrichment analysis demonstrates that escitalopram-based remission associates to functions related to cellular proliferation, apoptosis, and inflammatory response. Escitalopram/memantine-based remission, however, is characterized by processes related to cellular clearance, metabolism, and cytoskeletal dynamics. Both treatments modulate inflammatory responses, albeit via different effector pathways. Additional research is needed to understand the implications of these results in explaining the observed superior effects of combination treatment on cognition observed with prolonged treatment.

Conflict of interest statement

Compliance with ethical standards

Conflict of interest HL received research support from Allergan/Forest Laboratories. All other authors report no financial relationships with commercial interests.

© 2020. The Author(s), under exclusive licence to Springer Nature Limited.

Figures

Fig. 1. ESC/PBO and ESC/MEM remission profiles differ pre- and post-treatment.

a Agglomerative hierarchical clustering and Venn diagram of differentially expressed genes in at least one of the tested remission profiles at an FDR adjusted p value < 0.05. Red: positive log2 fold change. Blue: negative log2 fold change. b Rank-rank hypergeometric overlap (RRHO) heatmap and Venn diagram of concordant and discordant genes in ESC/PBO and ESC/MEM remitter compared with non-remitter profiles at baseline. Overlapping genes that change significantly in the same direction are located in the bottom left and upper right quadrants. Overlapping genes that change significantly in different directions between conditions are located in the upper left and bottom right quadrants. c RRHO analysis of ESC/PBO and ESC/MEM remitter compared with non-remitter profiles at 6-month follow-up (controlled for pretreatment expression). ESC/PBO escitalopram/placebo, ESC/MEM escitalopram/memantine, R remitter, NR non-remitter, Pre baseline, Post 6-month follow-up, D downregulated, U upregulated.

Fig. 2. Signaling pathways uniquely enriched by ESC/PBO and ESC/MEM remitters compared with non-remitters.

Gene set enrichment analysis of a ESC/PBO and b ESC/MEM pretreatment, remitter, and non-remitter transcriptional profiles. Analyzed gene sets include hallmark and canonical pathways. Red: positive normalized enrichment score, indicating activation post-treatment. Blue: negative normalized enrichment score, indicating repression post-treatment. Triangle symbol size is proportional to −log(FDR adjusted p value). ESC/PBO escitalopram/placebo, ESC/MEM escitalopram/memantine, R remitter, NR non-remitter, Pre baseline, Post 6-month follow-up.

Fig. 3. Leading edge analysis of ESC/PBO and ESC/MEM enriched gene sets.

Frequency count of genes that appears in the leading edge of significantly enriched hallmark pathway, canonical pathway, and gene ontological gene sets in a ESC/PBO and b ESC/MEM remission (FDR adjusted p value < 0.05). The Jaccard similarity index was calculated between the leading edges of all significantly enriched gene sets. The most highly correlated gene sets are shown by correlogram for c ESC/PBO and d ESC/MEM remission where intensity of hue signifies a higher Jaccard similarity index.

Fig. 4. Network analysis of enriched gene sets in ESC/PBO and ESC/MEM remitters compared with non-remitters.

Clustering of gene set enrichment analysis results of a ESC/PBO and b ESC/MEM remitter profiles. Tested gene sets included hallmark pathways, canonical pathways, and gene ontology gene sets. Only gene sets with an FDR adjusted p value < 0.05 were included. Nodes represent gene sets and lines demonstrate their connectivity based the degree or overlap between set genes. Highly related terms are enclosed in a circle with a label automatically summarized from the word-frequency of the contained terms. c Immune response subnetwork in ESC/PBO remission. d Proteasomal degradation subnetwork in ESC/MEM remission. Red: positive normalized enrichment score, indicating activation posttreatment. Blue: negative normalized enrichment score, indicating repression post-treatment.