The Pharmacology and Immunotherapy of IgAN

Why a drug can suppress B cells and still not work here. Three agents in this list have real, measurable effects on B-cell biology and still fail — or only partly succeed — in IgA nephropathy, and each fails for a structurally different reason.

Rituximab depletes CD20+ B cells completely and never reaches the two places that matter: the GALT source, and the plasma cell, which has shed CD20. In the randomized trial, Gd-IgA1 and its autoantibodies did not change at all.

Mycophenolate blocks IMPDH-II and so blocks lymphocytes that are dividing. The long-lived plasma cell is post-mitotic — it is not synthesising DNA and does not need the de novo purine pathway — so the established antibody-secreting pool is untouched. Its trial record fits that ceiling: a clear benefit in Chinese cohorts (composite outcome aHR 0.23) that has not reproduced in Western populations, which is why KDIGO 2025 restricts the suggestion to Chinese patients. Tellingly, eGFR loss accelerated from 2.9 to 6.1 mL/min/1.73m²/year after patients stopped it.

Abatacept and basiliximab remove T-cell help and IL-2-driven expansion — Signal 2 and Signal 3. But IgA class switching in GALT runs through the T-independent TACI route, which needs no T cell at all. Removing T help leaves the mucosal route open.

Nefecon is the control experiment. It is also a B-cell-killing drug — budesonide is a glucocorticoid and induces apoptosis in the Peyer's patch B cells directly. It works where rituximab fails not because it targets something different, but because it is delivered where the cells actually are. Roughly 90% first-pass hepatic metabolism keeps systemic exposure low, which is the whole design. Same target, opposite outcome, on compartment access alone.

The common lesson: B-cell activity is not the target. The pathogenic protein is. An agent has to reach the cell that is secreting it, in the compartment where that cell lives.

Try belimumab and anti-APRIL together. Neither alone brings down the long-lived plasma cell, because either ligand on its own is sufficient to keep it alive — the pair does what a TACI-Fc decoy does in one molecule. That redundancy is the mechanistic argument for dual blockade, and it is why an anti-APRIL agent leaves the marrow niche supplied.

Then look at panel 2. Class switching also has two sufficient routes. Blocking T-cell help leaves the TACI route running; blocking both ligands leaves CD40 running. Only removing both routes stops AID — which is why no agent here abolishes antibody production, and why serum IgG still falls 35% under APRIL blockade even though APRIL is called an IgA pathway. The switch signal is shared; the cytokine context, not the ligand, picks the isotype.

Attenuated is not closed. Remove APRIL alone and BAFF still engages TACI, so the route does not shut — but APRIL is the dominant mucosal IgA-switching ligand, and output collapses anyway: IgA −68.8%, Gd-IgA1 −67.1%. That gap between "the route is still open" and "the output fell by two thirds" is the honest shape of this pharmacology, and it is why the IgA nodes here fade rather than disappear.

Which vaccine responses survive follows the same split. In the phase 1 study, anti-APRIL suppressed every immunoglobulin class dose-dependently and reversibly, yet tetanus and diphtheria recall titres still rose after vaccination — protein antigens, T-dependent, answered by the memory compartment the drug does not touch. The polysaccharide axis is the one that fails: under a TACI-Fc agent, pneumococcal titres fell 30% and protective status was lost three times as often as placebo. Preserved recall to a protein antigen is not evidence that the T-independent response is intact.