Co-stimulation and modulation of the ensuing immune response

As a consequence of the central role of dendritic cells (DC) in stimulating primary immune responses any bias in the response introduced by the DC has the potential for having a long-term effect on immunity. Examination and analysis of ruminant afferent lymph dendritic cells derived by cannulation allows studies on the properties of ex vivo DC that is not possible in humans and rodents and information can be derived from ruminants that has implications of generic relevance. Previous studies have identified two major populations of DC in afferent lymph draining the skin of cattle that differ in their capacity to stimulate CD4 and CD8 T cells. Differences in expression of cytokine transcripts have now been shown for the two types of DC. The CD11a(+)/SIRPalpha(-) population synthesised more IL-12, whilst the CD11a(-)/SIRPalpha(+) population produced more IL-10. This is likely to affect the bias of the immune response following presentation of antigen to T cells by one DC sub-population or the other. An inability to synthesise IL-1alpha was the reason for the failure of the CD11a(+)/SIRPalpha(-) DC to stimulate CD8 T cells. This property would potentially affect the induction of CD8 responses. Expression of the co-stimulatory molecules CD80, CD86 and CD40 appeared similar for both DC populations and not to relate to differences in function. A further examination of the SIRPalpha molecule on DC indicated that on cross-linking it was tyrosine phosphorylated and that it recruited the SHP-2 protein phosphatase. Associated with this was a blocking of TNFalpha secretion on exposure to LPS. The interaction of SIRPalpha with its ligand CD47 on T cells appeared to be an early event in the stimulation of T cells as binding of the ligand was reduced on activated T cells. CD26 was identified as another molecule expressed by the SIRPalpha(-) DC sub-population. This is reported to have an enzymatic activity on certain chemokines that could result in the promotion of a Th1 bias.A model is proposed that takes these observations into account in which SIRPalpha(-) DC would be expected to promote a Th1 biased response and the SIRPalpha(+) DC a more balanced one.