Primary Hyperparathyroidism and Gut Microbiota

Methods. The project will be organized in three working packages (WP). WP1. Determine if the propensity of patients with PHPT to develop bone loss can be predicted by the composition of stool microbiome (month 0-11). We will screen men and women age 30-80 years diagnosed with PHPT and evaluated by 3-site DXA (lumbar spine, hip, and distal radius- both 1/3 and ultra-distal), and bone turnover markers. Two groups with 45 patients each will be enrolled: 1) Subjects with normal BMD or mild osteopenia (distal radius T-score > -1.5), no fragility fractures and CTX levels that are associated with normal bone turnover (100-250 pg/ml) (not bone looser). 2) Subjects with osteoporosis (distal radius T-score < -2.5) with or without one or more fragility fractures) and CTX levels that are associated with active disease (>500 pg/ml) (bone looser). Blood samples and a faecal sample will be collected by each participant. Stool samples will be stored immediately in a cooler with -20°C freezer packs and stored at -80°C until shipped to COSMOSID for Metagenomic Shotgun sequencing and analysis of the microbiome community structure.

Metagenomics Shotgun sequencing and analysis. The microbiome diversity and composition of each stool sample will be determined by CosmosID Inc., Rockville, MD. Briefly, DNA from human faecal samples will be isolated using the QIAGEN DNeasy PowerSoil Pro Kit. DNA libraries will be prepared using the Illumina Nextera XT library preparation kit and Library quantity assessed with Qubit (ThermoFisher). Libraries will be then sequenced on an Illumina HiSeq platform 2x150bp. Bioinformatic analysis of sequencing reads will be done by CosmosID bioinformatics platform that uses high performance data mining algorithms and the highly curated dynamic comparator database GenBook®.

End points and interpretation: We will determine if PHPT patients loosing bone have a higher frequency of one or more Th17 cell-inducing-SFB equivalent bacteria as compared to control PHPT patients. This analysis will therefore yield evidence pertaining to species and strains that may exacerbate bone loss in PHPT. The DXA endpoints will be BMD, T-scores and Z-scores of the distal and 1/3 radius, lumbar spine and hip.

Power Calculation: Our analysis of the preliminary data set revealed that BL was abundant at mean relative abundance of ~5.8% (SD 0.055). From these data we calculated that enrolling 45 per group will provide 80% statistical power of detecting a difference for BL if the mean relative abundance of BL is > 1.5-fold (8.7%) higher in PHPT patients with low bone density as compared to control PHPT patients.

WP2. Evaluate how GM composition modulates immune system in humans (month 3-10). In order to evaluate Th subsets, T cells will be analysed by real time RT-PCR measuring FOXP3, IL-17A, TNF-alpha and IL-4 gene expression. Briefly, red cells will be lysed in peripheral blood samples and total nucleated cells collected and frozen at -80°C until RNA extraction. RNA will be isolated using TRIzol reagent, chloroform extraction, and subsequent isopropanol precipitation according to standard procedure. Relative cytokine expression will be determined using the 2-ΔΔCT method with normalization to β-Actin. In order to evaluate inflammation we will measure IL-17A (the main species of IL-17 produced by Th17 cells), free RANKL, OPG, TNF by ELISA technique. To assess bone turnover P1NP (a marker of bone formation), CTX (a marker of bone resorption), 25OHvitamin D (25OHD) will be measured by ELISA technique.

End points and interpretation. The data will reveal the existence of correlations between rate of bone resorption, Th cell subset and serum levels of IL-17 and other cytokines. Should we find that the composition of the GM is reflective of the activity of the disease; the data will serve as the foundation for future intervention studies to determine the efficacy of antibiotics or other interventions as, for example, with probiotics in mitigating the skeletal complications of PHPT. We also expect that PHPT patients with skeletal disease will have lower 25OHD levels than those without skeletal disease. Should we find a correlation between 25OHD levels and bacterial species and strains that favours Th17 cell formation; the data will provide novel mechanistic information on how 25OHD levels affect the propensity to bone disease in PHPT. Should this be the case, a very attractive subject for a future intervention study will be the proposition that vitamin D replacement will restore a microbiota composition that is not associated with skeletal disease.

Risk analyses and contingency plan. Oestrogen deficiency expands Th17 cells and their production of IL-17 may contribute to postmenopausal bone loss in humans. This effect might represent an important confounder, which may limit the assessment of the role of IL-17 in the bone loss induced by PHPT in early postmenopausal women. If necessary, we will manage this limitation by statistically adjusting for the effects of sex, age and menopause.

WP3. Evaluate the causal relationship between GM composition and T cells activation (month 6-10). In order to evaluate if Th cells are directly stimulated by proteins from SFB equivalent bacteria we will test the peripheral blood mononuclear cells (PBMCs) directly toward BL ATCC 15707 and evaluate the type of responding T cells by ELISPOT technique. Briefly, PBMCs from 10 patients with increased Th17 cells in peripheral blood will be tested against BL ATCC 15707 and the production of IL-17, TNF alpha and RANKL will be tested by ELISPOT analyses for. These cytokines, within the PBMCs, are produced specifically by T cells following their activation. ELISPOT technique allows testing multiple proteins in a single experiment. TNF alpha and RANKL other than IL-17 have been selected as candidate cytokines as the applicant and others have previously shown an increased level of these cytokines in T cells from patients affected by post-menopausal osteoporosis. Should we find a direct activation of T cells to BL ATCC 15707, the type of T cell is responsible for activation will be further characterized thanks to the use of single-cell RNA sequencing technique (scRNA-seq) with the the Fluidigm C1 system as it allows a more in-depth sequencing per cell (about 200.000 readings). Briefly, PBMC cells from 5 responders patients will be challenged with ATCC 15707 or with a control microbe (B. fragilis) for 6 hours. After incubation T cells will be separated by total blood nucleated cells by negative immunomagnetic cell isolation system, and evaluated with scRNA-seq. The use scRNA-seq will enable us to characterize activated T- cells and their gene expression End points and interpretation. We will determine if ATC C 15707 is able to directly activate T cells and further we will characterise the type of responding T cell. This will allow us to confirm the association data obtained in the first part of the study and provide a mechanistic insight into the relationship between GM, immune activation and bone loss.

Risk analyses and contingency plan. If T cells from patients with increased number of Th17 in the peripheral blood will not positively answer to immune challenge with ATCC 15707, we will test T cells reactivity towards others microbes increased in the GM of PHPT bone losers. GM strain able to stimulate Th17 cells will be chosen according to the results obtained in WP1. In order to define a positive T cells activation we set a threshold of 70% of positive results to immune challenge according with our preliminary data and with previous results published by the PI.