Stem cell therapy in autoimmune disease

The name is misleading in a way that matters. Mesenchymal stromal cells infused for autoimmune and inflammatory disease are not expected to engraft, differentiate or replace tissue. Tracking studies show that intravenously administered MSCs are trapped in the pulmonary capillary bed within minutes — they are large, adherent cells in a vessel narrower than they are — and are largely cleared within a day or two. Any account of their clinical effect has to be consistent with that.

Immunomodulation, and the licensing requirement

The mechanism that fits is paracrine and indirect. MSCs secrete indoleamine 2,3-dioxygenase, prostaglandin E2, TGF-β, TSG-6 and nitric oxide, which suppress T-cell proliferation, shift macrophages toward a regulatory phenotype and induce regulatory T cells. Crucially, this activity is not constitutive: it must be licensed. Exposure to interferon-γ, with TNF-α, strongly upregulates IDO and the rest of the suppressive programme. An MSC placed in a quiet environment does relatively little; the same cell in an inflamed one becomes strongly immunosuppressive.

A second strand goes further and treats the cells’ death as part of the mechanism. Work published in 2017 found that MSCs are killed by recipient cytotoxic cells and that the apoptotic bodies are taken up by host monocytes, which then adopt an immunoregulatory phenotype — and that patients whose own cells failed to kill the infused MSCs did not respond. If correct, the active agent is the host response to a dying allogeneic cell, which reframes the product entirely.

Both accounts predict the same clinical pattern: efficacy tracks the recipient’s inflammatory state. That is consistent with approvals being concentrated in acute, strongly inflamed indications — steroid-refractory graft-versus-host disease, local injection for perianal fistulising Crohn’s disease — and with the long record of trials in quieter chronic autoimmunity reading out flat.

Regulatory T cells are a different bet

Treg therapy is not paracrine. The cells are meant to persist and suppress antigen-specifically through CTLA-4-mediated costimulation stripping, IL-2 consumption and IL-10 and TGF-β release. Its two problems are lineage stability — FOXP3 expression can be lost in an inflamed environment, converting a suppressor into an effector — and specificity, since polyclonal Tregs suppress broadly. Chimeric antigen receptor Tregs address the second by redirecting suppression to a chosen tissue antigen.

The unsolved problem is potency

Both classes share a defect that explains much of the field’s inconsistency: there is no validated potency assay that predicts clinical effect. MSC preparations vary by tissue source, donor, passage number and expansion medium, and cells that look identical by the standard surface-marker panel differ substantially in secretory behaviour. Because the product is defined largely by process, a manufacturing change is a change of drug, and trials from different laboratories may not be testing the same thing.

Autologous haematopoietic stem cell transplantation belongs in a separate category. There the intent is a genuine immune reset — lymphoablation followed by reconstitution of a fresh repertoire — and the outcome is bought with the real toxicity and treatment-related mortality of conditioning.

See also the technology article at /technology/stem-cell-therapy-autoimmune/.