CHIP and gout: trained immunity?

In this issue of Blood, Agrawal et al¹ identified clonal hematopoiesis of indeterminate potential (CHIP) as a risk factor for gout, a common inflammatory arthritis that is defined by an NLRP3-inflammasome and an interleukin-1β (IL-1β)–dependent innate immune system response to monosodium urate (MSU) crystals.

CHIP is defined as the presence of clonally expanded cells with somatic mutations in an individual who has no evidence of hematologic malignancy. But CHIP is a risk factor for hematologic malignancy and also for chronic conditions such as cardiovascular disease² and worse outcomes in chronic kidney disease.³ The somatic mutations are present in a suite of several dozen genes, most prominently mutations in TET2 and DNMT3A, that encode epigenetic regulators. That murine Tet2-deficient immune cells have dysregulated NLRP3 inflammasome activity and elevated release of IL-1β makes the CHIP endophenotype a compelling candidate for an etiological role in gout, an archetypal condition of acute IL-1β dysregulation.

To address the question of whether CHIP is associated with gout, Agrawal et al included 169 805 individuals from the UK Biobank (UKB), a resource with comprehensive data that include exome sequences. They also included 8019 participants from the Mass General Brigham Biobank (MGBB). Agrawal et al excluded people with hematologic malignancy and known germline variants; they then searched for somatic mutations in 71 known CHIP genes with variant allele frequency (VAF) ≥2% and identified 645 CHIP somatic mutations in 11 685 individuals. In the MGBB, the risk of gout was increased by ∼70% in individuals with VAF ≥2%, whereas increased risk of prevalent gout (∼30%) and incident gout (∼70%) was restricted to individuals with VAF ≥10% in the UKB.

A flare of gout requires hyperuricemia (defined as serum urate >7 mg/dL), deposition of MSU crystals in joints, and the IL-1β–driven innate immune response. To genetically test the hypothesis that CHIP was involved in the progression from hyperuricemia to gout, association with prevalent gout was tested in the UKB by using controls with asymptomatic hyperuricemia. There was an ∼30% increased risk of gout in this cohort, consistent with a role in the progression from hyperuricemia to gout. However, there are 2 caveats for this analysis: the first is that the controls were not hyperuricemic by the commonly used clinical definition (the authors defined hyperuricemia as being above the median serum urate level [between 5 and 6 mg/dL]); and second, a true hyperuricemic control cohort should be matched to a gout case cohort for lifetime exposure to elevated urate,⁴ which is not possible using single urate measures. This analysis must be interpreted with much caution. It did not eliminate the possibility that CHIP is associated with gout by having an effect on serum urate levels, a hypothesis supported by the observation that germline variation in some CHIP genes is associated with serum urate levels (eg, TET2; see figure).^5,6 It would be revealing to test for the association of CHIP with urate levels in the UKB and MGBB.

Evidence for the involvement of CHIP in the progression from hyperuricemia to gout must be buttressed by other means. In this case Agrawal et al pivoted to a mouse model of acute gout. They showed that mice with Tet2 that was engineered to be inactive in hematopoietic cells had increased IL-1β response to MSU crystals injected into the peritoneal cavity. In the same model, there was marginally increased swelling in response to MSU crystals injected into the paw. One limitation is that there is no good mouse model of hyperuricemia and gout.⁷ Moreover, MSU crystals dissolve rapidly when injected into mice because of the low urate levels. The authors then exposed isolated macrophages from Tet2 knockout mice to MSU crystals and the toll-like receptor 4 stimulant bacterial lipopolysaccharide (LPS); the Tet2-deficient macrophages had increased IL-1β response that was sensitive to genetic blockade of the Nlrp3 inflammasome. These findings support the hypothesis that CHIP, or at least the predominant CHIP gene (Tet2), is causally associated with the innate immune response to MSU crystals. The obvious extension of this experimental work into humans would be to investigate the effect of knockdown or pharmacological inhibition of TET2 in ex vivo monocytes exposed to MSU crystals and LPS.

Collectively, the results of the study by Agrawal et al implicate the CHIP phenomenon in the pathogenesis of the progression from hyperuricemia to clinical gout. The findings require replication by others. That somatic mutations in CHIP genes are strongly associated with increased age⁸ provides a mechanism (in addition to reduced kidney function and reduced excretion of urate) for the increasing incidence of gout with increasing age. That TET2 (the predominant CHIP gene) is an epigenomic mediator as are other CHIP genes (eg, DNMT3A, EZF2) raises the possibility that the molecular mechanism implicated in gout involves the training phenomenon, whereby external stimuli (in this case, soluble urate) train the innate immune system to be hyperresponsive to MSU crystals through reprogramming of the epigenome.⁹ This hypothesis could be tested by knocking down the function of TET2 and other epigenome-modifying CHIP genes and then testing for an effect on the training phenotype in an ex vivo model of training by soluble urate,¹⁰ including examining for an effect on the epigenome (eg, DNA methylation and chromosomal architecture). We have observed the influence of training on DNA methylation at the DNMT3A locus (see figure).

The association of somatic mutations in CHIP genes with gout raises the question of association of germline variation in these genes with gout. Querying of the Genome Wide Association Studies (GWAS) Catalog (ebi.ac.uk/gwas) for 14 well-established CHIP genes reveals genetic association of 3 of them with urate and/or gout: IDH2 (urate and gout), TET2 (urate), GNAS (urate) (see figure). Notably, 9 of the genes were associated with various white blood cell phenotypes. Definitive testing of the hypothesis that inherited genetic variation in CHIP genes is causal for gout will require expansion of GWAS. The observation that all 3 genes are associated with urate raises the intriguing possibility that regulation of urate levels is a factor. This would be consistent with urate-driven training of the innate immune system as a mechanism linking the CHIP phenomenon to gout. Identifying and characterizing the functional impact of causal germline and somatic mutant variants will be key to identifying the putative mechanism.