Intracellular trafficking of perforin: to thwart a killer.

How do killer cells restrain perforin, the most potent toxin known to biologists, at its point of synthesis in the endoplasmic reticulum, where conditions are ideal for its activation? In this issue of Immunity, Brennan et al. (2011) study its trafficking, offering insights into protective mechanisms.

Human and mouse granzyme A induce a proinflammatory cytokine response.

Granzyme A (GzmA) is considered a major proapoptotic protease. We have discovered that GzmA-induced cell death involves rapid membrane damage that depends on the synergy between micromolar concentrations of GzmA and sublytic perforin (PFN). Ironically, GzmA and GzmB, independent of their catalytic activity, both mediated this swift necrosis. Even without PFN, lower concentrations of human GzmA stimulated monocytic cells to secrete proinflammatory cytokines (interleukin-1beta [IL-1beta], TNFalpha, and IL-6) that were blocked by a caspase-1 inhibitor. Moreover, murine GzmA and GzmA(+) cytotoxic T lymphocytes (CTLs) induce IL-1beta from primary mouse macrophages, and GzmA(-/-) mice resist lipopolysaccharide-induced toxicity. Thus, the granule secretory pathway plays an unexpected role in inflammation, with GzmA acting as an endogenous modulator.

Granzyme A produced by γ(9)δ(2) T cells induces human macrophages to inhibit growth of an intracellular pathogen.

Human γ(9)δ(2) T cells potently inhibit pathogenic microbes, including intracellular mycobacteria, but the key inhibitory mechanism(s) involved have not been identified. We report a novel mechanism involving the inhibition of intracellular mycobacteria by soluble granzyme A. γ(9)δ(2) T cells produced soluble factors that could pass through 0.45 µm membranes and inhibit intracellular mycobacteria in human monocytes cultured below transwell inserts. Neutralization of TNF-α in co-cultures of infected monocytes and γ(9)δ(2) T cells prevented inhibition, suggesting that TNF-α was the critical inhibitory factor produced by γ(9)δ(2) T cells. However, only siRNA- mediated knockdown of TNF-α in infected monocytes, but not in γ(9)δ(2) T cells, prevented mycobacterial growth inhibition. Investigations of other soluble factors produced by γ(9)δ(2) T cells identified a highly significant correlation between the levels of granzyme A produced and intracellular mycobacterial growth inhibition. Furthermore, purified granzyme A alone induced inhibition of intracellular mycobacteria, while knockdown of granzyme A in γ(9)δ(2) T cell clones blocked their inhibitory effects. The inhibitory mechanism was independent of autophagy, apoptosis, nitric oxide production, type I interferons, Fas/FasL and perforin. These results demonstrate a novel microbial defense mechanism involving granzyme A-mediated triggering of TNF-α production by monocytes leading to intracellular mycobacterial growth suppression. This pathway may provide a protective mechanism relevant for the development of new vaccines and/or immunotherapies for macrophage-resident chronic microbial infections.

Interleukin-1R signaling is essential for induction of proapoptotic CD8 T cells, viral clearance, and pathology during lymphocytic choriomeningitis virus infection in mice.

The T cell granule exocytosis pathway is essential to control hepatotropic lymphocytic choriomeningitis virus strain WE (LCMV-WE) but also contributes to the observed pathology in mice. Although effective antiviral T cell immunity and development of viral hepatitis are strictly dependent on perforin and granzymes, the molecular basis underlying induction of functionally competent virus-immune T cells, including participation of the innate immune system, is far from being resolved. We demonstrate here that LCMV-immune T cells of interleukin-1 receptor (IL-1R)-deficient mice readily express transcripts for perforin and granzymes but only translate perforin, resulting in the lack of proapoptotic potential in vitro. LCMV is not cleared in IL-1R-deficient mice, and yet the infected mice develop neither splenomegaly nor hepatitis. These results demonstrate that IL-1R signaling is central to the induction of proapoptotic CD8 T cell immunity, including viral clearance and associated tissue injuries in LCMV infection.

Extracellular Granzyme A Promotes Colorectal Cancer Development by Enhancing Gut Inflammation.

If not properly regulated, the inflammatory immune response can promote carcinogenesis, as evident in colorectal cancer (CRC). Aiming to gain mechanistic insight into the link between inflammation and CRC, we perform transcriptomics analysis of human CRC, identifying a strong correlation between expression of the serine protease granzyme A (GzmA) and inflammation. In a dextran sodium sulfate and azoxymethane (DSS/AOM) mouse model, deficiency and pharmacological inhibition of extracellular GzmA both attenuate gut inflammation and prevent CRC development, including the initial steps of cell transformation and epithelial-to-mesenchymal transition. Mechanistically, extracellular GzmA induces NF-κB-dependent IL-6 production in macrophages, which in turn promotes STAT3 activation in cultured CRC cells. Accordingly, colon tissues from DSS/AOM-treated, GzmA-deficient animals present reduced levels of pSTAT3. By identifying GzmA as a proinflammatory protease that promotes CRC development, these findings provide information on mechanisms that link immune cell infiltration to cancer progression and present GzmA as a therapeutic target for CRC.

Granzyme B activates procaspase-3 which signals a mitochondrial amplification loop for maximal apoptosis.

Granzyme B (GrB), acting similar to an apical caspase, efficiently activates a proteolytic cascade after intracellular delivery by perforin. Studies here were designed to learn whether the physiologic effector, GrB-serglycin, initiates apoptosis primarily through caspase-3 or through BH3-only proteins with subsequent mitochondrial permeabilization and apoptosis. Using four separate cell lines that were either genetically lacking the zymogen or rendered deficient in active caspase-3, we measured apoptotic indices within whole cells (active caspase-3, mitochondrial depolarization [DeltaPsim] and TUNEL). Adhering to these conditions, the following were observed in targets after GrB delivery: (a) procaspase-3-deficient cells fail to display a reduced DeltaPsim and DNA fragmentation; (b) Bax/Bak is required for optimal DeltaPsim reduction, caspase-3 activation, and DNA fragmentation, whereas BID cleavage is undetected by immunoblot; (c) Bcl-2 inhibits GrB-mediated apoptosis (reduced DeltaPsim and TUNEL reactivity) by blocking oligomerization of caspase-3; and (d) in procaspase-3-deficient cells a mitochondrial-independent pathway was identified which involved procaspase-7 activation, PARP cleavage, and nuclear condensation. The data therefore support the existence of a fully implemented apoptotic pathway initiated by GrB, propagated by caspase-3, and perpetuated by a mitochondrial amplification loop but also emphasize the presence of an ancillary caspase-dependent, mitochondria-independent pathway.

Granzyme B translocates across the lipid membrane only in the presence of lytic agents.

Granzyme B (GrB), a component of the cytotoxic cell granule secretion pathway, is designed to kill infected and transformed cells after intracellular delivery by the pore forming protein, perforin. The mechanism of the delivery remains speculative. In this study we tested the hypothesis that GrB possesses capacity to bind and disrupt lipid membranes. Here in comparison to previous studies that show GrB interacts with carbohydrate moieties, the protease does not bind membrane phospholipids nor has intrinsic membranolytic properties. To study the transmembrane movement of GrB, we developed a model membrane system consisting of a high-molecular weight GrB substrate encapsulated in unilamellar vesicles. Intra-vesicle proteolysis clearly requires concentrations of lytic agents (streptolysin O, perforin or Triton X-100) that disrupt unilamellar membranes.

Cytotoxic granule-mediated apoptosis: unraveling the complex mechanism.

The molecular details of cytotoxic granule-mediated apoptosis have been gleaned from the study of the effects of isolated granzymes and perforin on target cells. Recent evidence indicates that the physiological apoptosis-inducing form is a multi-component macro-complex consisting of cationic granule proteins non-covalently linked to the chondroitin-sulfate proteoglycan, serglycin.

Granzyme A Contributes to Inflammatory Arthritis in Mice Through Stimulation of Osteoclastogenesis.

OBJECTIVE: Granzyme A (GzmA) levels are elevated in the plasma and synovium of patients with rheumatoid arthritis (RA), suggesting involvement of this protease in the pathogenesis of the disease. GzmA contributes to sepsis by regulating the production of proinflammatory cytokines. The purpose of this study was to evaluate the contribution of GzmA to the pathogenesis of RA in vivo and to examine the possibility that GzmA acting via tumor necrosis factor (TNF) stimulates osteoclastogenesis. METHODS: Inflammatory arthritis induced by type II collagen was evaluated in wild-type, GzmA-deficient, and perforin-deficient mice. The osteoclastogenic potential of GzmA was examined in vitro using bone marrow cells and colony-forming unit-granulocyte-macrophage (CFU-GM) cells and in vivo using GzmA-deficient mice. RESULTS: Gene deletion of GzmA attenuated collagen-induced arthritis, including serum levels of proinflammatory cytokines, joint damage, and bone erosion in affected mice, suggesting that osteoclast activity is reduced in the absence of GzmA. Accordingly, GzmA-treated bone marrow cells produced multinucleated cells that fulfilled the criteria for mature osteoclasts: tartrate-resistant acid phosphatase (TRAP) activity, ß integrin expression, calcitonin receptor expression, and resorptive activity on dentin slices. GzmA appeared to act without accessory cells, and its activity was not affected by osteoprotegerin, suggesting a minor contribution of RANKL. It also induced the expression and secretion of TNF. Neutralization of TNF or stimulation of CFU-GM cells from TNF-/- mice prevented GzmA-induced osteoclastogenesis. GzmA-deficient mice had reduced osteoclastogenesis in vivo (fewer calcitonin receptor-positive multinucleated cells and fewer transcripts for cathepsin K, matrix metalloproteinase 9, and TRAP in joints) and reduced serum levels of C-terminal telopeptide of type I collagen. CONCLUSION: GzmA contributes to the joint destruction of RA partly by promoting osteoclast differentiation.

Detection of functional cell surface perforin by flow cytometry.

How perforin (PFN) delivers the granzymes during cytotoxic granule mediated apoptosis remains a mystery. A major obstacle has been the inability to visualize PFN in either monomeric or polymeric form after interaction with the target cell surface. An antibody based technique is described which detects cell surface PFN on intact cells by flow cytometry. The methodology requires the presence of calcium (Ca2+) at a concentration which supports binding but not polymerization of PFN. Functionality was ensured by showing the cell surface PFN was able to deliver GrB causing caspase-3 activation and mitochondrial depolarization. The technique demonstrates a role for heparan sulfate proteoglycans in PFN binding. Further, the variable sensitivity of effector versus target cell lines to the permeabilizing effects of PFN could not be attributed to differential binding of PFN.

Detection of human perforin by ELISpot and ELISA: ex vivo identification of virus-specific cells.

The perforin (PFN) protein is essential for the elimination of target cells by cytotoxic T lymphocytes (CTL) and natural killer (NK) cells. The study of cells releasing PFN has been hampered by a lack of sensitive methods. We therefore produced PFN-reactive monoclonal antibodies (mAb) and developed capture enzyme-linked immunosorbent (ELISA) and enzyme-linked immunospot (ELISpot) assays. Three mAbs were generated and shown to react with unique determinants of PFN. All mAbs recognized intracellular PFN in human peripheral blood mononuclear cell (PBMC) as assessed by flow cytometry and immunohistochemistry. Functional PFN capture ELISA and ELISpot assays were developed utilizing two of the mAbs for capture and the third mAb for detection. When examining PFN release by the YT lymphoma cell line, the ELISpot displayed a greater detection sensitivity than the ELISA. Assessment of PFN release by a CTL clone using ELISpot gave results consistent with a parallel (51)Cr-release cytotoxicity assay. Moreover, PFN release by PBMC could be quantified by ELISpot and ELISA after ex vivo stimulation with defined CTL epitopes from common viruses. These novel immunoassays will be valuable for further investigations of the mechanisms underlying granule-mediated apoptosis. In addition, the capture immunoassays could provide tools for studying CTL responses in infectious and tumor diseases as well as for vaccine development.

Perforin rapidly induces plasma membrane phospholipid flip-flop.

The cytotoxic cell granule secretory pathway is essential for host defense. This pathway is fundamentally a form of intracellular protein delivery where granule proteases (granzymes) from cytotoxic lymphocytes are thought to diffuse through barrel stave pores generated in the plasma membrane of the target cell by the pore forming protein perforin (PFN) and mediate apoptotic as well as additional biological effects. While recent electron microscopy and structural analyses indicate that recombinant PFN oligomerizes to form pores containing 20 monomers (20 nm) when applied to liposomal membranes, these pores are not observed by propidium iodide uptake in target cells. Instead, concentrations of human PFN that encourage granzyme-mediated apoptosis are associated with pore structures that unexpectedly favor phosphatidylserine flip-flop measured by Annexin-V and Lactadherin. Efforts that reduce PFN mediated Ca influx in targets did not reduce Annexin-V reactivity. Antigen specific mouse CD8 cells initiate a similar rapid flip-flop in target cells. A lipid that augments plasma membrane curvature as well as cholesterol depletion in target cells enhance flip-flop. Annexin-V staining highly correlated with apoptosis after Granzyme B (GzmB) treatment. We propose the structures that PFN oligomers form in the membrane bilayer may include arcs previously observed by electron microscopy and that these unusual structures represent an incomplete mixture of plasma membrane lipid and PFN oligomers that may act as a flexible gateway for GzmB to translocate across the bilayer to the cytosolic leaflet of target cells.

Human perforin permeabilizing activity, but not binding to lipid membranes, is affected by pH.

The various steps that perforin (PFN), a critical mediator of innate immune response, undertakes to form a transmembrane pore remains poorly understood. We have used surface plasmon resonance (SPR) to dissect mechanism of pore formation. The membrane association of PFN was calcium dependent irrespective of pH. However, PFN does not permeabilize large or giant unilamellar vesicles (GUV) at pH 5.5 even though the monomers bind to the membranes in the presence of calcium. It was possible to activate adsorbed PFN and to induce membrane permeabilization by simply raising pH to a physiological level (pH 7.4). These results were independently confirmed on GUV and Jurkat cells. The conformational state of PFN at either pH was further assessed with monoclonal antibodies Pf-80 and Pf-344. Pf-344 maps to a linear epitope within region 373-388 of epidermal growth factor (EGF)-like domain while the Pf-80 appears to recognize a conformational epitope. Pf-344 interacts with the EGF-like domain after PFN monomers undergo pore formation, the site recognized by Pf-80 is only accessible at acidic but not neutral pH. Thus, the Pf-80 mAb likely interacts with a region of the monomer that participates in oligomerization prior to insertion of the monomer into the lipid bilayer and thus may have therapeutic utility against PFN-mediated immunopathology.

A novel mechanism for protein delivery: granzyme B undergoes electrostatic exchange from serglycin to target cells.

The molecular interaction of secreted granzyme B-serglycin complexes with target cells remains undefined. Targets exposed to double-labeled granzyme B-serglycin complexes show solely the uptake of granzyme B. An in vitro model demonstrates the exchange of the granzyme from serglycin to immobilized, sulfated glycosaminoglycans. Using a combination of cell binding and internalization assays, granzyme B was found to exchange to sulfated glycosaminoglycans and, depending on the cell type, to higher affinity sites. Apoptosis induced by purified granzyme B and cytotoxic T-cells was diminished in targets with reduced cell surface glycosaminoglycan content. A mechanism of delivery is proposed entailing electrostatic transfer of granzyme B from serglycin to cell surface proteins.

Lack of the CD8+ cell anti-HIV factor in CD8+ cell granules.

In HIV infection, CD8+ cells show cytotoxic and noncytotoxic anti-HIV activity. The latter function is mediated, at least in part, by a secreted antiviral protein, the CD8+ cell antiviral factor (CAF). Because antiviral effector molecules, such as perforin and granzymes, reside in the exocytic granules of CD8+ T cells, we examined the possibility that granules contain CAF-like activity. CD8+ cells from HIV-infected individuals showing strong CAF-mediated antiviral activity were induced to release their granule constituents into culture media. Within 1 hour of stimulation, high levels of granzyme B (a primary granule constituent) were found in the culture fluids of previously activated CD8+ cells. The same culture fluids contained no or very low amounts of CAF activity, as measured with HIV-infected CD4+ cells. Maximal levels of CAF activity were not observed until 5 or 7 days after stimulation, consistent with typical CAF production kinetics. In addition, extracts of granules purified from antiviral CD8+ cells did not show any CAF activity, whereas the cytoplasmic fraction of these cells showed substantial levels of antiviral activity. These findings suggest that CAF does not reside at appreciable levels in the exocytic granules of antiviral CD8+ T cells.

The human cytomegalovirus protein UL16 mediates increased resistance to natural killer cell cytotoxicity through resistance to cytolytic proteins.

Several reports have shown that human cytomegalovirus (HCMV)-infected cells are resistant to NK lysis. These studies have focused on receptor-ligand interactions, and different HCMV proteins have been indicated to mediate inhibitory NK signals. Here, we report that the HCMV protein UL16 is of major importance for the ability of HCMV-infected cells to resist NK cell-mediated cytotoxicity. Fibroblasts infected with the UL16 deletion mutant HCMV strain exhibited a 70% increased sensitivity to NK killing at 7 days postinfection compared to AD169-infected cells. Interestingly, HCMV-infected cells did not appear to engage inhibitory molecules on NK cells, since the levels of granzyme B were not reduced in supernatants obtained from NK cell cocultures with infected target cells compared to uninfected target cells. Furthermore, HCMV-infected cells, but not cells infected with the UL16 deletion mutant HCMV strain, exhibited a significantly increased resistance to the action of cytolytic proteins, including perforin, granzyme B, streptolysin O, and porcine NK lysin. In addition, fluorescence-activated cell sorting for UL16-positive transfected cells resulted in protection levels of 90% compared to control cells carrying the green fluorescent protein vector. Thus, the UL16 protein mediates an increased protection against the action of cytolytic proteins released by activated NK cells, possibly by a membrane-stabilizing mechanisms, rather than by delivering negative signals to NK cells.

Cytotoxic cell granule-mediated apoptosis: perforin delivers granzyme B-serglycin complexes into target cells without plasma membrane pore formation.

The mechanism underlying perforin (PFN)-dependent delivery of apoptotic granzymes during cytotoxic cell granule-mediated death remains speculative. Granzyme B (GrB) and perforin were found to coexist as multimeric complexes with the proteoglycan serglycin (SG) in cytotoxic granules, and cytotoxic cells were observed to secrete exclusively macromolecular GrB-SG. Contrary to the view that PFN acts as a gateway for granzymes through the plasma membrane, monomeric PFN and, strikingly, PFN-SG complexes were shown to mediate cytosolic delivery of macromolecular GrB-SG without producing detectable plasma membrane pores. These results indicate that granule-mediated apoptosis represents a phenomenon whereby the target cell perceives granule contents as a multimeric complex consisting of SG, PFN, and granzymes, which are, respectively, the scaffold, translocator, and targeting/informational components of this modular delivery system.

Cytotoxic cell granule-mediated apoptosis. Characterization of the macromolecular complex of granzyme B with serglycin.

We have recently shown that the physiological mediator of granule-mediated apoptosis is a macromolecular complex of granzymes and perforin complexed with the chondroitin-sulfate proteoglycan, serglycin (Metkar, S. S., Wang, B., Aguilar-Santelises, M., Raja, S. M., Uhlin-Hansen, L., Podack, E., Trapani, J. A., and Froelich, C. J. (2002) Immunity 16, 417-428). We now report our biophysical studies establishing the nature of granzyme B-serglycin (GrB.SG) complex. Dynamic laser light scattering studies establish that SG has a hydrodynamic radius of approximately 140 +/- 23 nm, comparable to some viral particles. Agarose mobility shift gels and surface plasmon resonance (SPR), show that SG binds tightly to GrB and has the capacity to hold 30-60 GrB molecules. SPR studies also indicate equivalent binding affinities (K(d) approximately 0.8 microm), under acidic (granule pH) and neutral isotonic conditions (extra-cytoplasmic pH), for GrB.SG interaction. Finally, characterization of GrB.SG interactions within granules revealed complexes of two distinct molecular sizes, one held approximately 4-8 molecules of GrB, whereas the other contained as many as 32 molecules of GrB or other granule proteins. These studies provide a firm biophysical basis for our earlier reported observations that the proapoptotic granzyme is exocytosed predominantly as a macromolecular complex with SG.

NF-kappaB protects from the lysosomal pathway of cell death.

The programme of gene expression induced by RelA/NF-kappaB transcription factors is critical to the control of cell survival. Ligation of 'death receptors' such as tumor necrosis factor receptor 1 (TNF-R1) triggers apoptosis, as well as NF-kappaB, which counteracts this process by activating the transcription of anti-apoptotic genes. In addition to activating caspases, TNF-R1 stimulation causes the release of cathepsins, most notably cathepsin B, from the lysosome into the cytoplasm where they induce apoptosis. Here we report a mechanism by which NF-kappaB protects cells against TNF-alpha-induced apoptosis: inhibition of the lysosomal pathway of apoptosis. NF-kappaB can protect cells from death after TNF-R1 stimulation, by extinguishing cathepsin B activity in the cytosol. This activity of NF-kappaB is mediated, at least in part, by the upregulation of Serine protease inhibitor 2A (Spi2A), a potent inhibitor of cathepsin B. Indeed, Spi2A can substitute for NF-kappaB in suppressing the induction of cathepsin B activity in the cytosol. Thus, inhibition of cathepsin B by Spi2A is a mechanism by which NF-kappaB protects cells from lysosome-mediated apoptosis.