Back to 2024 Abstracts
Circulating microRNA as a Potential Biomarker for Skeletal Disease in Primary Hyperparathyroidism: A Prospective Case-Control Study
*
Heather Wachtel1, *Jae Ermer
1, Douglas Fraker
1, Rachel Kelz
1, *Thomas Kelly
2, *Matthias Hackl
3, *Michael Levine
41Surgery, University of Pennsylvania, Philadelphia, PA; 2Hologic, Inc, Marlborough, MA; 3TAmiRNA, Vienna, Austria; 4Children's Hospital of Philadelphia, Philadelphia, PA
Background: Primary hyperparathyroidism (PHPT) is associated with increased bone turnover and decreased bone mineral density (BMD), leading to increased risk of fracture. Parathyroidectomy improves BMD and reduces fracture risk. However, conventional markers of BMD imperfectly predict fracture risk in PHPT, and therefore patients who do not meet BMD criteria may still benefit from parathyroidectomy. Because many patients with PHPT have concomitant idiopathic or post-menopausal osteoporosis, biomarkers are needed to distinguish between these causes of low BMD and to predict which patients are best treated by parathyroidectomy. The goals of this study were therefore to: 1) characterize miRNA expression signatures in patients with PHPT; and 2) identify miRNA biomarkers that correlated with bone homeostasis in PHPT.
Methods: We performed a prospective case-control study of post-menopausal females with PHPT undergoing parathyroidectomy and control subjects matched for race, age, and BMD. Target discovery group size was 42 subjects/42 controls (99.5% probability of detecting a 0.5 standard deviations in means, two-sided alpha 0.05). Groups were over-enrolled to allow for attrition. Findings were assessed in an independent validation cohort of subjects with PHPT. We collected clinical and biochemical data, assessed BMD by DXA scan, and measured 27 serum miRNAs related to bone remodeling. We used linear regression to assess the correlation between serum levels of a panel of bone-related miRNAs, conventional biochemical markers of bone homeostasis, and BMD. Principle component analysis and unsupervised cluster analysis were performed.
Results: A total of 135 subjects were evaluated, including 49 with PHPT (discovery group), 47 control patients without PHPT, and an independent validation cohort of 39 PHPT patients. Of 27 miRNAs assessed, 9 (miR-335-5p, miR-130b-3p, miR-125b-5p, miR-23a-3p, miR-152-3p, miR-582-5p, miR-144-5p, miR-320a and miR-19b-3p) were differentially expressed in PHPT compared to matched control subjects in both the discovery and validation groups. All 9 differentially expressed miRNAs significantly correlated with levels of parathyroid hormone (PTH), and 8 of the 9 correlated with calcium levels. No differentially expressed miRNAs were consistently correlated with markers of BMD. On unsupervised cluster analysis, subjects with PHPT demonstrate similar clustering based on expression levels of the nine miRNAs. Subjects with PHPT segregate from controls based on the signature of these 9 miRNAs on principle component analysis.
Conclusions: PHPT is characterized by a unique miRNA signature that is distinct from post-menopausal and idiopathic osteoporosis. Levels of specific miRNAs correlate with PTH, suggesting that bone remodeling in PHPT may be mediated in part by PTH-induced changes in miRNA. miRNA profiling offers a potential tool to guide clinical management and identify patients who achieve maximum skeletal benefit from parathyroidectomy.
Clinical data comparisons between subjects with primary hyperparathyroidism (PHPT) in the discovery group, and control subjects.
| | PHPT
(n=49) | Control
(n=47) | P-value
(PHPT vs. Control) |
| Age, years* | 67 (62, 70) | 67 (63, 70) | 0.97 |
| Race | | | 0.94 |
| African-American (%) | 6 (12.2%) | 6 (12.8%) | |
| Caucasian (%) | 43 (87.8%) | 41 (87.2%) | |
| BMI, kg/m2* | 27.7 (22.0, 31.2) | 26.7 (22.9, 30.7) | 0.63 |
| Biochemical indices | | | |
| Alkaline phosphatase (U/L)** | 83.9 (3.1) | 75.6 (3.1) | 0.061 |
| Bone specific alkaline phosphatase, mcg/L* | 22.8 (19.2, 28.7) | 19.6 (16.1, 22.1) | 0.0040 |
| Calcium, mg/dL* | 10.7 (10.5, 11.0) | 9.4 (9.0, 9.7) | <0.0001 |
| Carboxy-terminal collagen crosslinks, pg/mL* | 0.81 (0.57, 1.08) | 0.43 (0.24, 0.55) | <0.0001 |
| Osteocalcin, ng/mL* | 29.5 (22.8, 39.3) | 17.8 (14.2, 22.5) | <0.0001 |
| PTH, pg/mL* | 77.7 (63.9, 103.6 | 36.7 (29.1, 46.1) | <0.0001 |
| Phosphate, mg/dL** | 3.1 (0.07) | 3.6 (0.07) | <0.0001 |
| Procollagen type 1 Intact N terminal polypeptide, mcg/L* | 47.3 (20.0, 76.8) | 61.3 (46.8, 76.7) | 0.026 |
| Thyroid stimulating hormone, mU/L* | 1.48 (1.09, 2.39) | 1.99 (1.31, 2.68) | 0.31 |
| 25 (OH) Vitamin D, ng/mL* | 35.5 (29.9, 40.2) | 34.7 (28.0, 43.4) | 0.73 |
| 1,25 (OH)2 Vitamin D, pg/mL** | 33.0 (11.3) | 38.4 (9.9) | 0.018 |
| Bone mineral density on DXA | | | 0.18 |
| Normal (%) | 3 (6%) | 6 (13%) | |
| Osteopenia (%) | 28 (57%) | 31 (66%) | |
| Osteoporosis (%) | 18 (37%) | 10 (21%) | |
| T-score on DXA** | | | |
| Femoral neck (n=81) | -1.8 (±0.7) | -1.5 (±0.8) | 0.092 |
| Femur (n=35) | -1.0 (±1.5) | -1.2 (±1.0) | 0.70 |
| Total hip (n=46) | -1.5 (±0.7) | -1.4 (±1.1) | 0.66 |
| Lumbar spine (n=93) | -1.0 (±1.4) | -0.9 (±1.2) | 0.67 |
| Lumbar spine (n=93) | -1.0 (±1.4) | -0.9 (±1.2) | 0.67 |
| Distal radius (n=27) | -1.8 (±1.6) | -1.5(±***) | *** |
| FRAX score for major fracture, % (n=59)* | 11.2% (8.1, 15.5) | 12.4% (8.8, 14.9) | 0.44 |
* Median with interquartile range (IQR)** Mean ± standard deviation (SD)*** Unable to evaluate – only one observation in control group
Unsupervised cluster analysis was performed for the nine miRNAs associated with PHPT on external validation. Subjects with PHPT demonstrate similar clustering based on expression levels of the nine miRNAs. Rows are centered; unit variance scaling is applied to rows. Imputation is used for missing value estimation. Both rows and columns are clustered using correlation distance and average linkage.
Back to 2024 Abstracts
