Effects of Age and Estrogen on Skeletal Gene Expression in Humans as Assessed by RNA Sequencing

Abstract

Precise delineation of the specific genes and pathways altered with aging and estrogen (E) therapy may lead to new skeletal biomarkers and the development of novel bone therapeutics. Previous human bone studies, however, have been limited by only examining pre-specified genes and pathways. High-throughput RNA sequencing (RNAseq), on the other hand, offers an unbiased approach to examine the entire transcriptome. Here we present an RNAseq analysis of human bone samples, obtained from iliac crest needle biopsies, to yield the first in vivo interrogation of all genes and pathways that may be altered in bone with aging and E therapy in humans. 58 healthy women were studied, including 19 young women (mean age ± SD, 30.3 ± 5.4 years), 19 old women (73.1 ± 6.6 years), and 20 old women treated with 3 weeks of E therapy (70.5 ± 5.2 years). Using generally accepted criteria (false discovery rate [q] < 0.10), aging altered a total of 678 genes and 12 pathways, including a subset known to regulate bone metabolism (e.g., Notch). Interestingly, the LEF1 transcription factor, which is a classical downstream target of the Wnt/β-catenin signaling pathway, was significantly downregulated in the bones from the old versus young women; consistent with this, LEF1 binding sites were significantly enriched in the promoter regions of the differentially expressed genes in the old versus young women, suggesting that aging was associated with alterations in Wnt signaling in bone. Further, of the 21 unique genes altered in bone by E therapy, the expression of INHBB (encoding for the inhibin, beta B polypeptide), which decreased with aging (by 0.6-fold), was restored to young adult levels in response to E therapy. In conclusion, our data demonstrate that aging alters a substantial portion of the skeletal transcriptome, whereas E therapy appears to have significant, albeit less wide-ranging effects. These data provide a valuable resource for the potential identification of novel biomarkers associated with age-related bone loss and also highlight potential pathways that could be targeted to treat osteoporosis.

Trial registration: ClinicalTrials.gov NCT02349113.

Conflict of interest statement

Competing Interests: This study was supported, in part, by an investigator-initiated grant from Merck. There are no other relevant declarations. This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.

Figures

Fig 1. Flow diagram for enrollment and allocation of the subjects in the E arm of the study.

Fig 2. (A) Notch pathway and (B) Wnt/β-catenin pathway genes are altered with aging in human bone biopsies.

Genes in the Notch pathway and Wnt/β-catenin pathway are significantly altered in old relative to young women revealed by RNAseq based on pathway analysis using the Ingenuity Pathway Analysis software (see Statistical analyses). Values are presented as median fold changes (95% CIs) for old relative to young subjects. †p < 0.01; ‡p < 0.001.

Fig 3. Networks derived from age-related differentially expressed genes (DEGs) in bone.

(A) Top 10 scoring networks and associated network functions derived from the 678 DEGs altered (median count ≥ 10, q < 0.10) in the young versus old dataset determined using the Ingenuity Pathway Analysis software (see Statistical Analysis). (B) Network of cell morphology, hematological system development and function, and protein synthesis (score = 33); key molecules implicated in bone metabolism include LEF1 and CDKN1A (p21). (C) Network of cardiovascular system development and function, cellular development, and organismal development (score = 31); key molecules implicated in bone metabolism include Notch signaling components (NOTCH3, NOTCH4, HEY1, JAG2, and DLL4) and SEMA3A. Green indicates downregulated genes, and red indicates upregulated genes.

Fig 4. Effects of age and estrogen (E) on serum sclerostin levels.

Serum sclerostin levels in the young versus old subjects (A) and in the old versus E-treated subjects (B) by the Biomedica and Meso Scale Discovery (MSD) assays. Data are mean ± SEM; note the difference in scales for the two sclerostin assays. *p < 0.05; †p < 0.01.

Fig 5. Effects of age and estrogen (E) on gene expression of INHBA and INHBB.

Bone INHBB and INHBA gene expression levels by RNAseq in the old relative to young women (A) and in the E-treated relative to untreated old women (B). Data are shown as median fold changes (95% CIs). †p < 0.01; ‡p < 0.001.