Variation in PBMC cytokine response to crystal-related stimuli: how much can we explain?

 

Nicholas A Sumpter (1), Brenda Kischkel (1), Suzanne Ruijten (1), Hang-Korng Ea (2), Leo A.B. Joosten (1,3)

 

Affiliation(s):

1. Department of Internal Medicine, Radboud University Medical Center, Nijmegen, Netherlands
2. Department of Rheumatology, Lariboisierre Hospital, Paris, France
3. Department of Medical Genetics, Iuliu Hatieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania

 

 

Objectives: The extent of immune response to crystal deposition likely influences the onset and severity of crystal-associated diseases. We observe a marked variability in cytokine production during ex vivo stimulation assays with crystal-related stimuli. Therefore, we aimed to quantify the contribution of genetic and non-genetic factors to the variability in peripheral blood mononuclear cell (PBMC) cytokine response to crystal-related stimuli seen between individuals.

Methodology: We used data from the 200FG cohort, including a total of 547 healthy participants from the Netherlands. Each year, approximately 200 of these participants have blood collected, then PBMCs are isolated prior to stimulation with a variety of stimuli for 24h or 7 days. Various cytokines are measured in the supernatants from these stimulation assays using ELISA. We also have whole genome genotyping data on 431 of the 547 participants.

For this study, data from the 2017, 2018, and 2023 cohorts were included (N = 152, 201, and 194 respectively). The 2017 cohort included stimulation with C16 fatty acid (a TLR agonist) alone, or in combination with monosodium urate (MSU) crystals. The 2018 cohort included stimulation with lipopolysaccharide (LPS), basic calcium phosphate (BCP) crystals, and the combination. The 2023 cohort included stimulation with LPS, calcium pyrophosphate (mCPPD) crystals, and the combination. All stimulants were prepared in a single batch, all stimulations were 24 hours long, and in all years IL-1β, IL-6, and IL-1RA were measured in the supernatant.

For each cytokine/stimulus pair, we first performed a GWAS, then we measured the percentage of variance explained (R2) by age, gender, PBMC lymphocyte:monocyte ratio, PBMC neutrophil percentage, date of stimulation, and polygenic risk scores comprised of all independent variants associated at P < 10-5 after adjusting for all other variables listed. We excluded any cytokine/stimulus pairs with more than 25% of data at the lower limit of detection.

Findings: We found similar synergistic IL-1β response patterns for all three crystals, though the other cytokine responses were inconsistent (Figure 1). We found that the percentage of variance explained for any factor was highly variable between cytokines and stimuli (Figure 2). Notably, PBMC lymphocyte:monocyte ratio explained up to 42% (median 5%) of variance in cytokine response, and date of stimulation explained up to 63% (median 11%). Date of stimulation was an important confounder for many cytokine responses. The polygenic risk scores explained up to 64% (median 16%) of variance, highlighting the importance of individual genetics in cytokine response.

Significance: This study highlights the importance of genetics, PBMC cell composition, and stimulation batches in determining the extent to which an individual responds to a stimulus. It therefore seems likely that individuals who are predisposed either genetically or environmentally to have higher cell counts of monocytes relative to lymphocytes could be at higher risk for developing diseases such as gout, osteoarthritis, or CPPD disease. Additionally, the genetic associations likely implicate modified cell signalling pathways in response to stimulation. The small explanatory power of gender also suggests that the gender differences seen in crystal arthritis disease frequencies may primarily result from non-inflammatory components of the diseases such as serum urate or phosphate/pyrophosphate concentrations.