Allorecognition Pathway: Transplant Rejection, DSA and GvHD

Transplant rejection is the fastest, most violent adaptive immune response a body can mount — and it is a response to a molecule rather than to a pathogen. Somewhere between one and ten per cent of a recipient’s naive T-cell repertoire will react to a mismatched MHC molecule, which is orders of magnitude above the frequency for a conventional antigen. That single number explains why an untreated allograft fails in days, why the immunology of transplantation has its own vocabulary, and why the antibody toolkit for studying it is unusually deep. This article walks the pathway from the mismatched MHC molecule itself through the three recognition routes, the synapse they share, the cellular and humoral arms they split into, and the regulatory arm that occasionally wins.

The alloantigen is the MHC molecule itself

In a normal immune response the MHC molecule is the frame and the peptide is the picture. In allorecognition the frame is the target. A recipient T cell reacting to H-2K^b or HLA-A, -B and -C is reading polymorphic residues on the helices of the MHC molecule, in combination with whatever peptide happens to be sitting in the groove. Because the recipient repertoire was never negatively selected against those residues, the precursor frequency is extraordinary — commonly quoted at 1–10%, against something closer to one in 10⁵ or 10⁶ for a conventional foreign peptide.

That is why the reagents on the antigen side of this pathway are haplotype-specific rather than pathogen-specific. Mouse work turns on H-2K^b and H-2D^b for C57BL/6, H-2K^d for BALB/c, and the haplotype-level H-2^b reagent when the whole graft is the question. Human work turns on HLA-A/B/C and the allele-restricted HLA-A2/B7, with HLA-DR, HLA-DQ and HLA-DP on the class II side.

Two housekeeping molecules decide how much of that alloantigen ever reaches the surface. β2-microglobulin is the obligate light chain of every class I molecule — without it, class I does not fold or traffic. NLRC5 is the class I transactivator that sets expression level, and CD74, the invariant chain, chaperones class II and blocks its groove until peptide exchange. If a graft downregulates class I to evade CD8 T cells, it walks straight into the NK problem described below.

Route one: direct allorecognition

Donor organs arrive carrying passenger leukocytes — dendritic cells, macrophages and lymphocytes resident in the tissue at the time of retrieval. These are professional antigen-presenting cells expressing intact donor MHC at high density together with CD80, CD86 and CD40. A recipient CD8α T cell binds that intact donor class I through its TCR–CD3 complex with no processing step at all, and receives full co-stimulation from the same cell.

This is the arm that produces classical acute cellular rejection in the first weeks. It is also self-limiting in an important sense: as the donor passenger leukocytes die or migrate out, the supply of intact donor MHC on professional APCs falls away, and the direct response contracts. A model that only measures the first fortnight is measuring mostly this route.

Route two: indirect allorecognition

The indirect route is slower, quieter, and does not stop. Recipient antigen-presenting cells take up shed donor material — membrane fragments, soluble MHC, apoptotic debris — process it down the conventional class II pathway, and present donor-derived allopeptides on their own HLA-DR, HLA-DQ or, in mouse, I-A/I-E to recipient CD4 T cells. Because the recipient supplies both the APC and the MHC, the route persists for as long as the graft sheds anything at all.

Two consequences follow, and both matter more than the mechanism sounds. First, indirect recognition is the engine of chronic rejection and transplant vasculopathy, on a timescale of months to years. Second, an indirect-pathway CD4 cell is precisely the cell that can deliver cognate help to an alloreactive B cell, because the B cell also internalises donor MHC, processes it, and presents the same allopeptides. That is the link between cellular immunity and donor-specific antibody — it runs through the indirect route, not the direct one.

Route three: semi-direct, and the cross-dressed APC

The third route was the awkward one to fit into textbooks and is now the one that explains the timing. Intact donor MHC–peptide complexes are transferred whole to recipient APCs, on exosomes or by trogocytosis — the recipient APC ends up cross-dressed in donor MHC. That single cell can now stimulate a direct-pathway CD8 cell through the acquired donor molecule and an indirect-pathway CD4 cell through processed allopeptide on its own class II, in the same lymph node, at the same time. It resolves the old problem of how CD4 help reaches a CD8 cell that recognises a different molecule altogether.

The synapse all three routes share

However the alloantigen is presented, the T cell still requires the same second signal. CD80 and CD86 engage CD28; CD40L on the T cell engages CD40 on the APC and licenses it; ICAM-1 on graft endothelium and APCs holds CD11a/LFA-1 long enough for the synapse to mature. CTLA-4 is the counterweight, stripping CD80 and CD86 off the APC by trans-endocytosis — the mechanism that CTLA-4-Ig fusion proteins exploit.

This is the layer clinical co-stimulation blockade targets, and it is why blockade timing is so awkward in practice: the same molecules that drive alloreactive priming are the ones regulatory T cells need. CD62L matters here too, since it controls whether naive alloreactive T cells enter the draining lymph node where priming actually happens.

Missing self: the NK arm that MHC downregulation walks into

NK cells read the same molecules with the opposite sign. Ly49C in mouse and NKG2A/C/E in both species are inhibitory receptors for self MHC class I: engagement means “this cell is one of ours” and killing is withheld. A graft that mismatches at class I, or that downregulates it to hide from CD8 T cells, fails to deliver that inhibitory signal and is killed by missing-self recognition instead. The same logic underpins KIR/Ly49 mismatch effects in haematopoietic transplant, where donor NK alloreactivity can be beneficial. NK1.1 is the lineage marker used to remove the arm entirely and ask whether it mattered.

Where it ends: three outcomes, three read-outs

Acute cellular rejection and GvHD

Alloreactive CTLs kill graft parenchyma with perforin and granzyme B, and through FasL — the death-ligand arm that dominates in hepatic GvHD. IFN-γ amplifies everything by upregulating MHC on graft tissue, and induces CXCL9, the chemokine that recruits the next wave. In graft-versus-host disease the direction reverses — donor T cells within the graft attack recipient tissue — but every molecule on this map behaves the same way.

Antibody-mediated rejection

The humoral arm begins with indirect CD4 help to alloreactive B cells (CD20), which mature into plasma cells (CD38) secreting donor-specific antibody. DSA binds graft endothelium, clusters, and fixes C1q; C1s then cleaves C4, and the split product binds covalently to the endothelial surface. That covalent bond is why C4d is the diagnostic marker of antibody-mediated rejection on biopsy: it stays put long after the antibody has washed away. C3 amplifies, C5 completes the terminal pathway, and in rodent models anti-rat κ light chain is how the donor-specific antibody itself is tracked.

Operational tolerance

The third outcome is the one everybody is trying to engineer. IL-2 captured through high-affinity CD25 sustains FoxP3 regulatory T cells, and a Treg-dominated response can hold a graft indefinitely. This is also where the pharmacology becomes genuinely counter-intuitive: anti-CD25 induction removes effector IL-2 signalling but also strips CD25-high Tregs, while low-dose IL-2 favours Tregs precisely because their high-affinity receptor captures cytokine at concentrations effector cells cannot use. Dose and timing decide the direction, not the target.

The three routes side by side

The practical reading is in the last column. An experiment that ends at day 14 is largely a direct-pathway experiment; one that asks about vasculopathy or de novo DSA at three months is an indirect-pathway experiment. The same graft, the same mismatch, two different questions.

Designing the intervention

This pathway has an unusually direct translation into reagents, because the clinical protocol has the same shape as the diagram. Induction depletes: anti-CD4 (GK1.5), anti-CD8α, anti-CD90/Thy-1 for pan-T depletion, anti-NK1.1 for the NK arm, and the anti-CD52 alemtuzumab biosimilar as the clinical lymphodepleting equivalent. Antigen-side blockade masks the alloantigen itself with low-endotoxin antibodies against the relevant haplotype. Co-stimulation and adhesion blockade targets CD28, CD80/CD86, CD40 and CD11a/ICAM-1. The humoral arm is addressed at the cell (CD20, CD38) or at the cascade (anti-C5).

Two experimental cautions. First, endotoxin: lipopolysaccharide contamination in a blocking antibody is a TLR4 agonist, and in an alloimmune model the read-out is an inflammatory state — contamination does not merely add noise, it manufactures the phenotype. Low- and ultra-low-endotoxin functional-grade material is not optional here. Second, isotype: depleting antibodies work through Fc-mediated clearance, so the isotype is part of the mechanism. A non-depleting anti-CD4 and a depleting anti-CD4 answer different questions, and tolerance induction protocols have historically depended on exactly that distinction.

Frequently asked questions

Why is alloreactivity so much stronger than a normal antigen response?

Because the recipient repertoire was never selected against foreign MHC. Between 1 and 10% of naive T cells can react to a mismatched MHC molecule, compared with roughly one in 10⁵–10⁶ for a conventional peptide antigen. It is a frequency effect, not an affinity effect, and it is why rejection is fast.

What is the difference between direct and indirect allorecognition in practice?

Direct recognition uses intact donor MHC on donor cells and drives early acute rejection, then fades with the donor passenger leukocytes. Indirect recognition uses processed donor peptide on recipient class II, never runs out of substrate, and drives chronic rejection and B-cell help. If you care about de novo donor-specific antibody, you care about the indirect route.

Why is C4d the marker of antibody-mediated rejection?

C4d is the split product of C4 that binds covalently to graft endothelium during classical pathway activation. Unlike the antibody or the earlier components, it does not wash off, so it acts as a durable footprint on biopsy that complement-fixing donor-specific antibody was present.

Is GvHD the same pathway as rejection?

Mechanistically, yes — read in reverse. In rejection, recipient T cells attack donor MHC; in graft-versus-host disease, donor T cells within the graft attack recipient MHC. The same recognition routes, the same synapse (CD28, CD40, LFA-1) and the same effectors (perforin, FasL) apply, with tissue tropism deciding the clinical picture.

Why does downregulating MHC not help the graft?

It trades one killer for another. Lower class I means fewer targets for direct-pathway CD8 T cells, but it also means less engagement of the inhibitory NK receptors Ly49C and NKG2A — and NK cells kill what fails to display self MHC. That is missing-self recognition.

Which read-outs track rejection best in a rodent model?

Granzyme B and perforin for cellular killing, IFN-γ and CXCL9 for the amplification loop, C4d and C1q for the humoral arm, and FoxP3 with CD25 if you are asking whether regulation is holding. Urinary and biofluid CXCL9 in particular has been studied heavily as a non-invasive rejection signal.

Featured products for this pathway

For research use only. Not for use in diagnostic or therapeutic procedures.