Neutrophil tissue inhibitor of matrix metalloproteinases-1 (TIMP- 1) : novel localisation, mobilisation and possible role.

Abstract

At the beginning of this study, the granule localisation and regulation of release of human neutrophil (PMNL) precursor collagenases, proMMP-8 and -9 (type I and type TV/V collagenases, respectively), enzymes highly active against the extracellular matrix (ECM) and thought to be relevant in invasion and inflammation, had been established while that of their inhibitor, tissue inhibitor of matrix metalloproteinases-I (TIMP-1), had not. Electron microscopy immunogold labelling of cryoultramicrotomy sections for granule marker proteins, lysosome-associated membrane proteins (LAMPs) and endocytosed bovine serum albumin-coated gold probes, followed by stereology, established that TIMP-1 was mainly located in a distinct oval, electron translucent organelle, a little larger than azurophil granules. A lack of labelling for endocytic markers and for glycosylphosphatidylinositol-anchored proteins, established using granule fractionation and immunolabelling to be markers for the secretory vesicles, and LAMPs-1 and -2, indicated the non-endosomal, non-secretory and nonlysosomal nature of this organelle. Density gradient cofractionation with the least dense secretory vesicle population and some pleiomorphism of the organelle suggested that it is a "vesicle" rather than a "granule" population. Colocalisation with proMMP-9 in minor subpopulations suggests that TIMP-1 vesicle biogenesis occurs between metamyelocytic and termination differentiation, but before secretory vesicle synthesis. Immunolabelling of phagocytosed and pulse-chased IgG-opsonised latex beads showed that specific and azurophil granules and a small number of proMMP-8-containing granules (a specific granule subpopulation) fuse with the phagosome whereas the TIMP-1 vesicle and proMMP-9-containing granules do not, suggesting that the latter play no role in phagosomal destruction of IgG-opsonised bacteria and that their phagosomal release is not calcium regulated. However, studies using the calcium ionophore, ionomycin, and monitoring extracellular granule marker protein release upon addition of increasing levels of extracellular calcium, showed that all granules, except the TIMP-1 vesicle, appeared to be calcium regulated. This suggests that the regulation of proMMP-9 release is not exclusively via calcium and that TIMP-1 vesicle release is not calcium regulated. Whereas most granules were shown to be associated with microtubule-like structures, the TIMP-1 vesicle and proMMP-9-containing granules were shown to associate with two morphologically different cytoskeletal elements, neither resembling actin nor tubulin. These elements, and the release of the TIMP-1 vesicle and proMMP-9-containing granules, need to be studied further, but results achieved to date may explain the observed differential mode of release of TIMP-1 relative to proMMP-9. The proMMP-9-binding and inhibitory capacity of a 66 kDa high molecular mass form of TIMP-1 was demonstrated in PMNL homogenates and plasma using western ligand blots and a novel reverse zymography method. The role and relevance of this form remains unknown as does the relevance and potential role of proMMP-9ffIMP-1 complexes seen during isolation procedures. The proMMP-9ffIMP-1 complex may occur in vivo, as evidenced by immunolocalisation studies, and, together with TIMP-1 released from its own discrete vesicle population, may be responsible for the fine regulation of extracellular proteolysis.