Prolonged activation of the complement system of blood proteins in response to infection or tissue injury can lead to inflammatory diseases. This thesis investigates the potency, selectivity and structures of reported and rationally designed new ligands for human complement receptors C3aR and C5aR, towards development of new therapeutic agents. Chapter 1 surveys the importance of C3a and C5a in human physiology, the potency of small molecule ligands reported as agonists or antagonists for hC3aR and hC5aR, and includes an analysis of nonpeptidic ligands for GPCRs in inflammation (Blakeney, J. et al. Chem Rev, 2007, 107, 2960; Blakeney, J. & Fairlie, D. Curr Med Chem, 2005, 12, 3027). Chapter 2 reinvestigates reported peptide agonists of C3aR. A C3aR ‘superagonist’ WWGKKYRASKLGLAR 12 was found NOT to be a selective or potent agonist for C3aR, as measured by intracellular calcium release induced by C3a on human differentiated U937 cells. By this criterion, four decapeptides reported as selective hC5aR agonists were shown to activate both hC5aR and hC3aR and to have unusual solution structures. While new hexapeptides FIPLAR 24 (EC50 0.8 μM) and FWPLAR 28 (EC50 0.5 μM) were potent and selective agonists for hC3aR, with no activity at hC5aR and only 10-fold lower potency than hC3a. Chapter 3 compares conformationally restricted cyclic peptides as ligands for C3aR, cyclisation via sidechain-to-C-terminus (i to i+4), sidechain-to-sidechain (i to i+4) or Nterminus- to-sidechain (i to i+3). YSFK[OPLaR] (38, first category) and H-[KKLLD]R-OH (40, second category) inhibited intracellular Ca2+ mobilisation (IC50 35 μM, 4 μM respectively) induced by 100 nM C3a on differentiated U937 cells and are the first structurally characterised cyclic antagonists of C3aR. NMR structures suggest that C3aR recognizes a type II β-turn. Chapter 4 examined structure-activity relationships for a nonpeptidic C3aR antagonist, SB290157. Substitution was tolerated between the bulky diaryl and arginine fragments, the flexible ether could be replaced by a more rigid amide or removed without loss of antagonism, and a 2,2’-diphenylmethyl substitutent was present in all active compounds (SB290157 (8), 56, 60, 93, 94 and 100) except for 83 (2-(3-(1H-indol-3-yl)propanamido)-5- guanidinopentanoic acid) which had comparable activity to SB290157 on THP-1 cells. Chapter 5 explores rational design of nonpeptide antagonists for C5aR using structureactivity relationships, pharmacophore generation, and matching of Cα-Cβ vector constraints from the solution structure of a peptide antagonist to nonpeptidic compounds in virtuallibraries. Even subtle changes to a competitive C5aR antagonist, W54011 (5), greatly reduced potency. A pharmacophore generated by superimposing structures of active and inactive nonpeptidic C5aR antagonists led to novel 2-mercaptobenzoimidazole C5aR antagonists, 179 and 184 (IC50 1.5 and 0.94 μM respectively; intracellular Ca2+, 10 nM C5a ). These studies record some valuable steps towards better understanding of the selective interactions of small molecules with GPCRs that bind to human C3a and C5a.