Blocking the binding of bacterial superantigens to the co-stimulatory CD28 ligand on T cells is sufficient to stop the toxins triggering the inflammatory cytokine storm that can lead to lethal toxic shock, researchers claim. A team at the Hebrew University of Jerusalem has found that superantigens must directly interact with the homodimer interface of CD28 in order to trigger the hugely overexaggerated immune reaction that can kill infected patients.
Their studies in mice showed that blocking superantigen-CD28 binding by administering peptides that compete for the binding site on CD28 prevents induction of the cytokine storm. Led by Raymond Kaempfer, Ph.D., at the Institute of Medical Research Israel Canada (IMRIC) at the Hebrew University Faculty of Medicine, the scientists report their findings in PLoS Biology in a paper titled “Binding of Superantigen Toxins into the CD28 Homodimer Interface Is Essential for Induction of Cytokine Genes That Mediate Lethal Shock.”
As the principal co-stimulatory receptor in every immune response, CD28 is a homodimer that interacts with B7 co-ligands to mediate T-cell activity. However, apart from the B7 co-receptors, no other ligands have been identified for CD28. Bacterial superantigens, meanwhile, are known to bypass the immune system’s failsafe antigen presentation mechanisms that usually ensure a correct level and type of immune response is mounted against a specific antigen. Instead, superantigens bind directly as intact proteins to most MHC class II (MHC-II) and T-cell receptor molecules. This results in massive induction of the T helper 1 (Th1) cytokines that mediate toxic shock.
Prior research has, however, indicated that superantigen activity also requires CD28 co-stimulation, while previous work by Dr. Kaempfer’s team has shown that a peptide mimicking a conserved region of a staphylococcal enterotoxin B (SEB) superantigen that is nowhere near the MHC-II and TCR binding sites, can protect mice from superantigen-related death. This finding strongly suggested that this region of the superantigen engages a third receptor.
The mimetic peptide, termed p12B, appears to act early in the immediate response to superantigen and blocks relevant mRNA induction within hours, the Hebrew University team continues. Their new work found that IL2 and IFN-γ mRNA induction by SEB is inhibited selectively by soluble CD28 comprising its extracellular domain fused to IgG1-Fc dimer.
This led to the notion that the inhibitory effect may be due to CD28-Fc competing directly with cell-surface CD28 for SEB binding. To provide more evidence for this, the team showed that induction of IL2 and IFN-γ mRNA using an activating antibody αCD28 was also be blocked by p12B. Significantly, αCD28 itself failed to bind to p12B, suggesting that the peptide competes with αCD28 in binding to cell surface CD28.
They then generated a phage library to select peptides that bound to both CD28-Fc and SEB. The investigators found that these peptides were capable of preventing SEB-related death in the majority of experimental mice tested.
“Thus, superantigen antagonists effective in vivo were selected from random peptide sequences solely by affinity for the SEB binding site in CD28,” they state. Demonstration that SEB binds to cell-surface CD28 was achieved in CD28-expressing cells that were transfected with GFP-tagged SEB. Administering an anti-CD28 antibody to these cells blocked subsequent binding of the labeled superantigen.
Molecular studies further showed that SEB uses its β-strand(8)/hinge/α-helix(4) domain to bind CD28 and that this binding is similar in affinity to the antigen’s known binding to the MHC-II and TCR ligands. Epitope mapping and studies with peptide mimetic confirmed that this SEB domain binds directly to the CD28 dimer interface.
The researchers’ previous work had shown that peptides capable of blocking Th1 cytokine induction in human PBMCs and lethal shock in mice, demonstrating that SEB must bind to CD28 to exert its toxicity, they note. In a new set of studies they evaluated the effects of a mutation that effects amino acid substitutions in the β-strand(8)/hinge/α-helix(4) domain of the superantigen . This variant, designated tk2, severely impaired the ability of SEB to induce IL2, IFN-γ, and TNF-α mRNA and protein. It didn't, however, affect induction of IL1, which is induced by MHC-II and TCR binding.
Importantly, all mice challenged with the tk2 variant of SEB survived, whereas all those challenged with wild-type SEB all died. “Thus, an intact β-strand(8)/hinge/α-helix(4) domain is essential for SEB toxicity,” the authors state.
Given that tk2 is a nonlethal variant of SEB, they were surprised to find that it displayed enhanced binding to CD28-Fc, compared with wild-type SEB. This, the researchers suggest, indicated that the interaction of tk2 with CD28 is nonproductive and that the higher affinity for CD28 might render it dominant-negative.
In fact, they showed that when both tk2 and wild-type SEB were administered to experimental mice, induction of IL2, IFN-γ, and TNF-α was markedly reduced, even when tk2 was present at a 10-fold lower concentration than SEB. Conversely, tk2 induced IL10 additively with wild-type SEB, “confirming the selectivity of the dominant-negative phenotype.”
“These experiments identify the CD28 dimer interface as a receptor target for superantigens,” the authors conclude. “We show that to deliver the signal for Th1 cytokine induction, a superantigen must co-opt CD28 as third receptor, in addition to its full reliance on co-stimulatory signaling through B7-2/CD28.
"Thus, superantigens make unconventional use not only of MHC-II and TCR through direct binding but also of CD28. Given the similarly moderate affinity of SEB for CD28, TCR, and MHC-II, we propose that it is the concerted interaction of the superantigen with all three receptors that allows for stable synapse formation resulting in exceptionally strong Th1 cytokine induction and lethality.”