Fatal Attraction: Immunoglobulin A and the Glomerular Mesangium

• IgA nephropathy
• proteinuria
• Progression
• GFR loss
• Innate immunity
• Affinity maturation

The defining feature of IgA nephropathy (IgAN) is glomerular mesangial deposition predominantly of IgA and here, mostly of dimeric and polymeric IgA₁. It has long been discussed whether the deposited IgA carries any antigenic specificity and whether this might yield clues to the pathogenesis of the disease. In this issue of JASN, Wehbi et al.¹ have new insight into this question by studying affinity maturation of IgA and its relationship to mesangial IgA deposition.

In the process of affinity maturation, B cells, activated by follicular helper T cells, continuously improve the binding of their antibodies after antigen exposure, in particular on repeated exposure.¹ In simple terms, affinity maturation is the consequence of two principles. First, continuous mutation of antigen binding sequences of the IgA genes, which requires the DNA editing enzyme activation-induced cytidine deaminase (AID), generates antibodies with new binding specificity and affinity. Second, another principle then is “selection of the fittest” B cells, in which the many B cell clones must compete for help from the follicular helper T cells. These T cells present antigen to the B cells, and only those clones with high-affinity antibodies to the antigen receive signals that favor their survival, whereas low-affinity B cells are deleted.

In their study, Wehbi et al.¹ used transgenic mice overexpressing human IgA₁. In the presence of AID (i.e., in mice with preserved affinity maturation), mesangial IgA and complement C3 deposits as well as mild mesangioproliferative disease developed with age. No proteinuria, hematuria, or renal failure was noted. In the absence of AID, mice expressed human IgA₁ carrying nonmutated, nonaffinity maturated variable domains. In these mice compared with AID wild-type mice, mesangial IgA and C3 deposits increased as did mesangial expansion. Renal failure developed, but again, it was in the absence of significant proteinuria or hematuria. The authors conclude that IgA-related nephrotoxicity involves IgA produced by innate-like B cells rather than high-affinity maturated IgA antibodies.

Although indeed, polymeric IgA may exhibit different capacities to inflict glomerular damage depending on affinity maturation, the alternative hypothesis is that glomerular damage is a function of the abundance of polymeric IgA. Thus, in the mice with present AID, IgA levels in serum increased four- to sevenfold with age, whereas serum IgA levels increased by about 50-fold in AID-deficient mice.¹ This study thereby extends a number of mouse models of glomerular IgA deposition and variable degrees of renal damage (Table 1). All of these, when examined, exhibited moderately to massively elevated IgA serum levels, much higher than the mild about twofold elevation in serum IgA that is observed in about one half of patients with IgAN.²

Another important observation in the study of Wehbi et al.¹ is that serum IgA, even in the mice with pronounced renal damage, did not exhibit an altered glycosylation, in particular reduced galactosylation. In human primary IgAN, most mesangial IgA₁ exhibits an undergalactosylated hinge region.³ This galactose-deficient IgA is now believed to be a major contributor to the pathogenesis of IgAN, although its serum levels are only mildly elevated and normal in about 50% of patients with IgAN.⁴ The IgA₁ hinge region is unique to humans and higher monkeys, and even small monkeys, like the Brazilian marmoset, no longer have a hinge-containing IgA molecule. In addition to other differences in the IgA system, this has severely hampered the development of good preclinical models of human IgAN. Therefore, the remarkable observation in the study of Wehbi et al.¹ is that human IgA₁ without any obvious glycation defect can induce glomerular damage as long as serum levels are high enough or it is not affinity maturated. Similarly, a second important observation in that study is that human IgA₁ induced glomerular damage independent of CD89, a presumed IgA receptor.⁵ In a previous study, human CD89 overexpression on macrophages/monocytes as opposed to neutrophils in the work by Wehbi et al.¹ induced overt disease with hematuria and proteinuria in mice transgenic for human IgA₁.⁵

Therefore, my conclusion from all of these data is that markedly elevated human or murine polymeric IgA in serum can induce glomerular IgA deposits in mice, a case of “fatal attraction.” If the same is true in humans, it could explain why glomerular IgA deposits have been detected in 5% of random German autopsies⁶ and 15% of Japanese donor kidneys before transplantation,⁷ even in the absence of any known cause of secondary IgAN.⁸ What requires more systematic study is the relative contribution of elevated polymeric serum IgA versus particular features, in particular galactosylation patterns, of circulating IgA in inducing IgAN or an IgAN-like disease. Thus, Table 1 shows that experimental IgA-associated glomerular damage can develop without hematuria, an almost universal finding in human primary IgAN, or proteinuria, and in particular, only rare models seem to lead to ESRD. In addition, in the study of Wehbi et al.¹ as in several other studies, no electron microscopy was performed, and we are left to speculate whether the glomerular localization of IgA or IgA complexes, be it mesangial or capillary, affects disease manifestations, such as hematuria or proteinuria. Indeed, in a previous study,⁵ mice overexpressing human IgA₁ only had endocapillary IgA₁ deposits but did not have mesangial injury or kidney dysfunction, and only sCD89 coinjection or coexpression induced mesangial IgA₁ deposits, hematuria, and proteinuria.

A final good illustration of why we need to better understand the “fatal attraction” of IgA for the mesangium is liver disease: IgAN is detected in 9%–25% of patients receiving a kidney biopsy at the time of a liver transplantation, and patients with liver cirrhosis exhibit two- to fourfold increases in serum IgA levels, including elevated polymeric IgA and undergalactosylated IgA.⁸ However, in contrast to primary IgAN, the mesangial IgA deposits in liver cirrhosis did not stain with an antibody specific for the galactose-deficient IgA₁ hinge region,³ mesangial proliferation is less prominent, and serum IgA₁ from patients with liver cirrhosis was less potent in inducing proliferation of human mesangial cells in vitro.⁸ All of the above are calls for efforts to identify subtypes of what we collectively call IgAN. Better understanding of these different pathomechanisms may allow us to clarify why patients have radically different clinical courses ranging from inert glomerular IgA deposits with little or no clinical manifestations to rapidly progressive IgAN quickly resulting in ESRD.

• Copyright © 2019 by JASN