A novel serine esterase expressed by cytotoxic T lymphocytes

Structural organization of the hCTLA-1 gene encoding human granzyme B

Cytotoxic T lymphocytes (CTLs) and natural killer/lymphokine-activated cells produce granzymes, a family of serine esterase proteins located in cytoplasmic granules. These might be involved in different cytotoxic pathways. We report the structural organization of the human gene encoding granzyme B (hCTLA-1). A 4.75-kb genomic DNA fragment containing all the sequences of granzyme B-encoding cDNA clones has been sequenced. The gene is composed of five exons and four introns. A comparison with the genomic organization of murine CCP1/CTLA-1 showed very similar structure and a 76% nucleotide homology in the coding sequences. This suggests that both genes may have a common ancestor. No typical regulatory element was detected in the 1160 bp upstream from the ATG start codon. The detection of a second locus related to hCTLA-1 is also described.

Cloning of an interleukin-4 inducible gene from cytotoxic T lymphocytes and its identification as a lipase

Interleukin-4 (IL-4) has been demonstrated to be an important lymphokine for the generation of cytotoxic T lymphocytes (CTLs). Here we describe an IL-4 inducible gene specifically expressed in CTLs. By sequence homology, this gene is likely to be the mouse homolog of pancreatic lipase. Oocyte translation of in vitro transcribed mRNA results in the expression of a protein with lipase activity, and Northern analysis of various tissues and a large panel of hematopoietic cell types demonstrates that this gene is expressed only in the pancreas and CTLs. Lysates of CTLs grown in IL-4, but not in IL-2, exhibit lipase activity. Furthermore, Northern analysis of CTLs grown in the presence of IL-4 for as little as 5 days demonstrates a marked induction of lipase mRNA, which correlates with enhanced cytolysis by these cells. These results suggest that this lipase may have an important role in CTL effector function.

Genomic organization of the mouse pore-forming protein (perforin) gene and localization to chromosome 10. Similarities to and differences from C9

Genomic clones encompassing the entire coding region of the mouse lymphocyte pore-forming protein gene (Pfp) have been isolated and used to determine its intron-exon organization. In contrast to C9, Pfp has a simple structure, consisting of only three exons (two of which encode polypeptide), a large 5' intron, and a single, smaller intron that is situated approximately one-third of the way through the protein-coding portions of the gene. The regions encoding the homologous domains of PFP and C9 are encoded on exons 7, 8, 9, and 10 of C9, but form only approximately half of the open reading frame of exon III in Pfp. Although encoding polypeptides with related functions, the two genes possess such sharply contrasting structures as to suggest that their analogous regions may have risen independently, by a process of convergent evolution. Using a panel of somatic cell hybrid cell lines, Pfp has been mapped to chromosome 10.

Mechanism of L-leucyl-L-leucine methyl ester-mediated killing of cytotoxic lymphocytes: dependence on a lysosomal thiol protease, dipeptidyl peptidase I, that is enriched in these cells

Exposure of murine or human lymphocytes to L-leucyl-L-leucine methyl ester (Leu-Leu-OMe) results in selective killing of cytotoxic lymphocytes, whereas helper T cells and B cells remain functionally intact. Cytolytic lymphocytes incubated in the presence of toxic concentrations of Leu-Leu-OMe were found to contain membranolytic metabolites of the structure (Leu-Leu)n-OMe, where n greater than or equal to 3. The sensitivity of cytotoxic lymphocytes to Leu-Leu-OMe was found to be dependent upon production of these metabolites by a lysosomal thiol protease, dipeptidyl peptidase I, which is present at far higher levels in cytotoxic lymphocytes than in cells without cytolytic potential or not of bone marrow origin. Thus, this granule enzyme is required for the unique effects of Leu-Leu-OMe and may provide a target for the development of other immunotherapeutic agents designed to delete cytotoxic lymphocyte responses.

CAP 37, a 37 kD human neutrophil granule cationic protein shares homology with inflammatory proteinases

We have previously shown that a major granule-associated cationic protein CAP 37 (Mr = 37 kD) derived from human PMN is a monocyte-specific chemoattractant. The N-terminal amino acid sequence of this novel chemotactic protein shares significant homology with a number of inflammatory molecules with protease activity including elastase and cathepsin G. However, a critical substitution of a serine for a histidine at position 41, results in its lack of serine protease activity.

Cell surface-associated proteinases in NK cell-mediated cytotoxicity: enhancement of enzyme expression is unique to activation with interferon-alpha

The human NK cell-mediated cytotoxicity reaction is sensitive to proteinase inhibitors with specificity for chymotrypsin-like enzymes inhibitable by 1-tosylamide 2-phenylethyl chloromethyl ketone (TPCK). Evidence is presented in support of previous data suggesting that this type of cytotoxicity is attributable to enzymes associated with the surface membrane of the NK cell. Activation of the cells with IFN-alpha results in increased cytolytic activity, the suppression of which requires an almost two- to threefold increase in the concentration of proteinase inhibitors. Treatment of NK cells with IFN-alpha results in increased surface binding of [3H]diisopropyl fluorophosphate ([ 3H]DFP). This effect is not inhibited by cycloheximide (50 micrograms/ml), suggesting translocation of preexisting enzymes to the surface membrane. TPCK can compete with [3H]-DFP for binding to the cell surface and can abrogate the increase in [3H]DFP binding observed after IFN-alpha stimulation of the cells. Treatment with IFN-gamma does not increase cell surface-associated proteolytic activity and stimulation with IL-2 results in much smaller increments. The sensitivity of cytotoxicity to proteinase inhibitors is confined to the initial 2-5 min of the reaction. This suggests that cell surface-associated proteinases play a role in the programming of NK cells for lysis, whereas subsequent events may be dependent on secreted enzyme moieties.

Heterogeneity of granules of murine cytolytic T lymphocytes. Isolation of a homogeneous population of dense granules

A separation procedure based on Percoll gradient centrifugation is described here that allows enrichment for two main populations of granules, with average densities of 1.05 and 1.1 g/ml. The heavy-density granules, which comprise 5-20% of the total amount of membrane-lytic activity, are morphologically and biochemically more homogeneous than the light-density granules, which are contaminated with other subcellular organelles. A trypsin-like serine esterase activity, used in the past as a granule marker, was mainly found in the free cytosol and in the light-density fractions, while a small portion of this activity was associated with the heavy-density fractions. This enzyme marker is therefore not exclusively associated with granules and should not be used as the only granule marker.

Differential sensitivity of anti-IgM-induced and NaF-induced inositol phospholipid metabolism to serine protease inhibitors in BAL17 B lymphoma cells

A BAL17 B lymphoma cell line bearing mu and delta chains on its surface behaves in a similar manner to normal mature B cells in terms of initial biochemical transmembrane signalling [Mizuguchi, Beaven, Ohara & Paul (1986) J. Immunol. 137, 2162-2167; Mizuguchi, Yong-Yong, Nakabayashi, Huang, Beaven, Chused & Paul (1987) J. Immunol. 139, 1054-1059]. Therefore the effects of protease inhibitors on increases in inositol phospholipid metabolism and intracellular free calcium concentration ([Ca2+]i) were examined. We show that the serine protease inhibitors Tos-Phe-CH2Cl (1-chloro-4-phenyl-3-L-tosylamidobutan-2-one-, TPCK) and Tos-Lys-CH2Cl (7-amino-1-chloro-3-L-tosylamidoheptan-2-one; TLCK) inhibit anti-IgM-mediated accumulation of inositol phosphates in a dose-dependent manner. InsP3 production induced by anti-IgM is also inhibited by pretreatment with Tos-Lys-CH2Cl or Tos-Phe-CH2Cl. Tos-Lys-CH2Cl- Tos-Phe-CH2Cl-mediated inhibition is not overcome by high concentrations of anti-IgM. Moreover, anti-IgM-mediated increases in [Ca2+]i are inhibited by pretreatment of the cells with these inhibitors. However, increases in inositol phospholipid metabolism caused by NaF, an activator of guanine-nucleotide-binding proteins (G-proteins), are approx. 10-fold more resistant to Tos-Lys-CH2Cl and Tos-Phe-CH2Cl inhibition compared with anti-IgM-induced changes. Furthermore, NaF-induced increases in [Ca2+]i are not inhibited by Tos-Lys-CH2Cl or Tos-Phe-CH2Cl pretreatment, suggesting that the inhibitors act at a step proximal to phospholipase C activation. The Tos-Lys-CH2Cl or Tos-Phe-CH2Cl treatment does not change the membrane IgM density as measured by flow cytometry, indicating that the active site of the inhibitors is distal to the membrane IgM molecule. These results indicate that serine proteases may be involved in coupling the receptor cross-linkage to G-protein.

Two proteins targeted to the same lytic granule compartment undergo very different posttranslational processing

The granules of natural killer (NK) cells contain cytolysin and serine proteases, proteins that are expressed specifically in cytolytic cells and are released in response to target binding. We have used immunofluorescence and immunoelectron microscopy to localize these proteins with respect to the various morphological compartments of granules in a rat NK cell line. Both cytolysin and the proteases are limited to the core regions of the dense core granules. While these proteins are targeted to the same compartment, they differ markedly in their posttranslational processing. Cytolysin bears N-linked oligosaccharides that are converted to the complex type, while the major trypsin-like protease, granzyme A, bears only high-mannose-type oligosaccharides. The glycans of granzyme A, but not those of cytolysin, are modified with phosphomannose moieties. These results suggest that one possible mechanism for packaging proteins into NK granules is the mannose 6-phosphate-dependent lysosomal targeting system. However, the absence of the mannose 6-phosphate modification from cytolysin suggests the existence of yet another targeting system.

Interaction of lymphokine-activated killer cells with susceptible targets does not induce second messenger generation and cytolytic granule exocytosis

CTL activation by specific targets leads to a rapid rise of inositol phosphates (InsPs) and of cytoplasmic-free Ca2+ concentration ([Ca2+]i). While these events are considered necessary to trigger granule secretion, Ca2+-independent cytolytic mechanisms have been recently proposed in addition or as an alternative to the classical Ca2+-dependent exocytosis model. We observed that lymphokine-activated killer (LAK) cells, obtained after stimulation with supraoptimal concentrations of IL-2 in short- or long-term cultures, kill susceptible targets in the absence of a [Ca2+]i rise and InsP3 formation. Moreover, LAK cell-mediated lysis was not associated with an increase in cytotoxic granule exocytosis, as evaluated by BLT-esterase release into the culture supernatant. Furthermore, using an antigen-specific CTL clone, which acquires LAK-like activity when cultured in medium containing high IL-2 doses, second messenger generation and cytolytic granule content secretion were not detected during lysis of unrelated target cells, while killing of specific targets triggered both these processes. These findings suggest that two lytic pathways may coexist in the same effector cells: a second messenger-dependent pathway involving degranulation, which is activated after TCR interaction with specific targets, and another pathway, independent of any known second messenger generation, responsible for unrelated target cell lysis.

Induction of target cell DNA release by the cytotoxic T lymphocyte granule protease granzyme A

The rapid breakdown of target cell DNA during CTL-mediated lysis has been difficult to explain by the granule exocytosis model of cytotoxicity. The involvement of CTL granule proteases in this process was strongly suggested by experiments in which CTL were pretreated with the serine protease inhibitor PMSF, in combination with agents that raise the pH of acidic intracellular compartments. While PMSF pretreatment alone had little effect on target lysis or DNA breakdown, the combination of PMSF and NH4Cl or monensin profoundly reduced target cell DNA release, while little effect was observed on target lysis, as measured by 51Cr release. CTL granule extracts cause release of 125I-DNA from detergent-permeabilized cells. This nuclear DNA-releasing (NDR) activity is inhibited by serine esterase inhibitors that also inhibit the granule BLT-esterase activity, and is specifically immunoabsorbed by antibodies to the CTL granule protease granzyme A. The NDR activity comigrates with BLT-esterase activity during subcellular fractionation, solubilization, gel filtration, and aprotinin-Sepharose affinity chromatography. SDS-PAGE analysis of the affinity-purified product indicates a molecular mass of 60,000 daltons under non-reducing conditions, which moves to 30,000 daltons upon reduction, consistent with previously reported behavior of granzyme A. When the purified material was reduced and alkylated, both esterase and NDR activities comigrated at 30,000 daltons upon gel filtration. Although fully lytic concentrations of purified LGL granule cytolysin alone failed to induce target cell DNA release, a combination of purified granzyme A and the cytolysin induces substantial DNA release.

Comparison of fluorochrome-labeled and 51Cr-labeled targets for natural killer cytotoxicity assay

An alternative method for measuring in vitro cellular cytotoxicity has been developed utilizing the carboxyfluorescein derivative 2',7'-bis(carboxyethyl)-5,6-carboxyfluorescein (BCECF) as the target cell label. Target cells labeled with the fluorescent dye are incubated with effector cells, if killing of targets occurs, BCEDF is released analogous to 51Cr release. Measurement of specific lysis in this assay is based on the direct measurement of dye retained by the remaining viable target cells using the Pandex FCA. In paired experiments we have compared the fluorochrome assay to the standard 51Cr release assay in measuring porcine natural killer cytotoxicity. The target labeling time with BCECF is 30 min as opposed to 1 h with 51Cr; and there is no significant dye reincorporation after release. The optimal target number per incubation well for the BCECF assay is 5 X 10(3) cells. In both the BCECF and 51Cr release assays, maximum percent specific lysis is reached after 3-4 h incubation. By 2 h incubation, the BCECF assay reaches the maximum seen with 51Cr and in a 4 h assay the maximum NK activity measured with BCECF labeled targets is always higher than that measured with 51Cr-labeled targets. In paired experiments, we have shown the reproducibility of the BCECF assay and that the BCECF assay measures NK enhancement by NK enhancing monoclonal antibody and inhibition by NK inhibiting monoclonal antibody as good as the 51Cr release assay, if not better. In conclusion, the BCECF assay is a reliable and reproducible measure of in vitro cellular cytotoxicity, eliminates the use of radioisotopes and is cost efficient.

A cytotoxic serine proteinase isolated from mouse submandibular gland

We have isolated a novel cytotoxic factor from the submandibular glands of male BALB/c mice by Sephadex G-50 gel filtration chromatography and reverse-phase HPLC. The cytotoxic factor is a serine proteinase, which belongs to the mouse glandular kallikrein (mGK) family, with an Mr of approximately 27,000. The purified serine proteinase showed cytotoxic activity against mouse thymocytes in a dose-dependent manner, and a serine proteinase inhibitor, diisopropyl fluorophosphate, blocked its cytotoxic activity.

Detection of perforin and granzyme A mRNA in infiltrating cells during infection of mice with lymphocytic choriomeningitis virus

The analysis of gene expression in cytotoxic T cells by in situ hybridization of serial liver and brain sections from mice infected with lymphocytic choriomeningitis virus (LCMV) and immunostaining with T cell marker- and virus-specific antibodies revealed a close histological association of infiltrating lymphocytes expressing the perforin and granzyme A genes with virally infected cells. Maximal frequency of perforin and granzyme A mRNA-containing cells on liver sections preceded by about 2 days maximal LCMV-specific cytotoxicity of the lymphoid liver infiltrating cells. These results are most consistent with an involvement of perforin and granzyme A in cell-mediated cytotoxicity in vivo.

N-alpha-benzyloxycarbonyl-L-lysine thiobenzyl ester (BLT) serine esterase in human cytolytic effector cells and cell line targets

The granules of in vitro primed cytotoxic mouse T cells and cytotoxic cell lines have been shown to contain high levels of N-alpha-benzyloxycarbonyl-L-lysine thiobenzyl ester (BLT) esterase. The enzyme activity has been suggested to be associated with the cytotoxic capacity of killer cells. We investigated human leucocytes and found that neutrophils, monocytes, cytotoxic T lymphocytes (CTL), natural killer (NK) cells [large granular lymphocytes (LGL)], and interleukin 2 activated killer (LAK) cells, which all display efficient cytotoxic capacity, show only marginal BLT esterase activity. The low BLT esterase activity in human lymphocytes increases about twofold when cells are stimulated in vitro with interleukin 2 (IL-2), phytohaemagglutinin (PHA), or cultured in mixed lymphocyte culture (MLC). Mouse T lymphocytes have about 20 times more BLT esterase activity than human T lymphocytes. The BLT activity in mouse T cells also increases about twofold in MLC. The human leukaemia cell lines (K562, U937, MOLT-4, Jurkat) and the mouse mastocytoma line (P815), which are frequently used as target cells, contain more BLT esterase activity than human resting or activated lymphocytes. We did not find a direct correlation between the cytotoxic capacity and the BLT esterase activity of killer cells.

Cytolysis by Ca-permeable transmembrane channels. Pore formation causes extensive DNA degradation and cell lysis

This study investigates the effect of the purified membrane pore formers, staphylococcal alpha-toxin and CTL perforin, on target cell lysis as measured by 51Cr release and on nuclear damage as measured by DNA degradation and 125IUdR release. Both pore formers cause dose-dependent cell lysis, which is accompanied by DNA release. The ratio of DNA/Cr release depends on the nature of target cell and shows the same pattern as the ratio of release of the two markers reported for CTL-mediated lysis of the same targets. DNA degradation is dependent on the presence of intracellular Ca in the target cell and is totally blocked if Ca is chelated by Quin 2 intracellularly and EGTA extracellularly. DNA degradation, in addition, is inhibited by the lysosomotropic agents NH4Cl, chloroquine, and monensin. rTNF doubles the degree of DNA degradation mediated by alpha-toxin in 3-h assays. We conclude that pore formers alone can mediate DNA degradation. In addition, they may promote the uptake of other factors and thereby accelerate their time course of action. DNA degradation by pore formers requires active target participation in a pathway that is dependent on intracellular Ca and lysosomes. These aspects of target lysis resemble CTL- and NK cell-mediated cytolysis.

Idiopathic pulmonary fibrosis: cellular and molecular pathogenesis

Diffuse involvement of the pulmonary interstitium with abnormal fibrous tissue is a process that occurs in many settings. There are many possible etiologies for pulmonary fibrosis, but in the majority of individuals, a clear cause cannot be determined and a diagnosis of idiopathic pulmonary fibrosis (IPF) results. Despite limited knowledge concerning the etiology, recent advances in biomedical technology offer great promise for increasing our understanding of IPF. This review will focus on current concepts of the pathogenesis and therapy of IPF.

Interleukin 1 induction of a serine esterase in a murine T cell line is inhibited by fetal calf serum

In PC60 cells a serine esterase associated with the cytotoxic granules (BLT-esterase) is induced by interleukin 1 (IL-1). The induction was markedly reduced by fetal calf serum (FCS). A defined culture medium was therefore developed in which the PC60 cells proliferated equally well as in serum-containing medium. In this new medium, IL-1 induced BLT-esterase to much higher levels of activity than in serum-containing medium. When testing IL-1 induction of BLT-esterase with various batches of FCS, no correlation was found between cell proliferation and the responses to IL-1. It was concluded that FCS contains one (or several) inhibitor(s) of IL-1 action.

Expression of cytoplasmic granules with T cell-associated serine proteinase-1 activity in Ly-2+(CD8+) T lymphocytes responding to lymphocytic choriomeningitis virus in vivo

Monoclonal antibodies (mAb) with specificity for the T cell-associated serine proteinase-1 of the mouse (MTSP-1) were used to study expression and storage of this enzyme in T lymphocytes during lymphocytic choriomeningitis (LCM) virus infection in vivo. Immunohistochemical analysis of splenic tissue at the peak of LCM virus-specific T cell-mediated cytolytic responses, i.e., at day 7 post infection, revealed high numbers of MTSP-1+ T lymphocytes in the interfollicular T cell-dependent area of the spleen. More than 50% of Ly-2+(CD8+) cells but only low numbers of Ly-2-(CD8-) cells, previously enriched by flow cytofluorometry, contained large amounts of cytoplasmic granules which stained both with MTSP-1-specific mAb and the esterase substrate N-alpha-benzyloxycarbonyl-L-lysine thiobenzyl ester. These data demonstrate that in vivo generated LCM virus-induced cytolytic T lymphocytes develop cytoplasmic storage granules containing MTSP-1 and suggest that the mechanism of granule exocytosis is operative in vivo; possibly MTSP-1 is one effector molecule participating in the Ly-2+(CD8+) T lymphocyte-mediated control of virus infection.

Cloning, analysis, and expression of murine perforin 1 cDNA, a component of cytolytic T-cell granules with homology to complement component C9

The nucleotide sequence coding for the cytotoxic T-lymphocyte (CTL) protein perforin 1 (P1) has been determined and the corresponding protein sequence has been derived. Murine CTL cDNA libraries contained in the vector lambda gt11 were screened by using a monospecific antiserum to purified P1. Three recombinant phages were isolated and their cDNA inserts were sequenced. The derived protein sequence contains 554 amino acids and displays, as expected, considerable homology with certain functional domains in the complement components C9, C8 alpha, C8 beta, and C7. The identity of P1 cDNA clones was verified by prokaryotic expression and the reactivities of antisera produced to the expressed proteins. In addition, antisera were produced to two synthetic peptides located in the center and C-terminal portions of P1. All antisera reacted with purified P1. In Northern blot analyses, P1 cDNA probes recognized a 2.9-kilobase mRNA only in CTL. Perforin mRNA was found in all cloned CTL and in all mixed lymphocyte reactions that gave rise to cytotoxic cells. Perforin mRNA was also detected in virus-specific CTL that had been generated in vivo and isolated from liver tissue of mice infected with lymphocytic choriomeningitis virus. The cell-specific expression of perforin is consistent with its postulated role in cytolysis.

Localization, implications for function, and gene expression of chymotrypsin-like proteinases of cytotoxic RNK-16 lymphocytes

Rat RNK-16 leukemia cells kill YAC-1, which are the cells lysed by rodent natural killer lymphocytes. We found chymotrypsin-like proteinase ('chymase') activity in the RNK-16 dense granules that also contain cytolytic activity. The chymase activity hydrolyzed the thiobenzyl peptide substrate Suc-Phe-Leu-Phe-SBzl and, in comparison to RNK-16 tryptase activity, was selectively inhibited by three different types of serine proteinase inhibitors. The selective inhibitors were the fungal aldehyde chymostatin, the chloromethylketone Z-Gly-Leu-Phe-CH2Cl, and the mechanism-based or 'suicide' inhibitor 7-amino-4-chloro-3-(2-phenylethoxy)isocoumarin. These proteinase inhibitors also blocked RNK-16 granule-mediated cytolysis. Chymostatin, a reversible inhibitor, delayed granule-mediated cytolysis, whereas the irreversible chloromethylketone and isocoumarin proteinase inhibitors completely abrogated granule-mediated cytolysis. The two irreversible inhibitors displayed biphasic inhibition of the chymase activity, indicating that at least two chymases are present in the granules. By Northern blot analysis, we found that RNK-16 mRNA hybridized strongly with a cDNA probe of CCPI, a mouse cytotoxic T lymphocyte serine proteinase gene. These data imply that chymase activity in the cytotoxic granules is important for cytolytic function and is likely to belong to a new subfamily of serine proteinases.

Inhibition of lymphocyte protease granzyme A by antithrombin III

T-lymphocytes contain a cytoplasmic granule associated homo-dimeric protease designated granzyme A. Upon T-cell target cell interaction, the granules undergo exocytosis and granzyme A, and other granule constituents, are released. Here we show that granzyme A secreted into plasma is immediately inactivated by antithrombin III. The rate of complex formation is enhanced 400-fold in the presence of heparin. Two different complexes are generated: granzyme A-antithrombin III and granzyme A-(antithrombin III)2, respectively, indicating that both active centers of granzyme A are functional. Thus, the proteolytic activity of lymphocyte protease granzyme A, whose physiologically relevant function is unknown, is well regulated in plasma.

Involvement of a serine protease in tumour-necrosis-factor-mediated cytotoxicity

We investigated the effect of various protease inhibitors on the anti-proliferative and cytotoxic action of tumour necrosis factor (TNF) on mouse L929 fibrosarcoma cells. 1. The following serine-type protease inhibitors led to inhibition of TNF action: phenylmethylsulfonyl fluoride, N alpha-p-tosyl-L-lysine chloromethane, N alpha-p-tosyl-L-phenylalanyl chloromethane, N alpha-p-tosyl-L-arginine methyl ester, L-leucine methyl ester, DL-phenylalanine methyl ester, N-acetyl-DL-phenylalanine-beta-naphthyl ester, p-nitrophenyl p'-guanidino-benzoate and antipain. We could not detect an effect of inhibitors specific for thiol protease on TNF. 2. Inhibition of TNF-mediated cytotoxicity was evident in both the presence and absence of actinomycin D or cycloheximide. 3. TNF itself was not found to be a protease, as it had no proteolytic activity in a sensitive colorimetric assay. [1,3-3H]Diisopropyl fluorophosphate, an effective irreversible inhibitor of serine proteases, did not bind to TNF. Pretreatment of TNF with N alpha-p-tosyl-L-lysine chloromethane did not influence its biological activity. 4. The addition of protease inhibitor to the cells at various times after TNF administration led to a gradual loss of protection, suggesting that the protease acts at a rather late stage. 5. Protease inhibitors did not influence TNF binding, internalization or metabolization. 6. No increase in supernatant protease activity or in cell-associated protease activity could be detected after treatment of L929 cells with TNF. Our results document the involvement of protease activity, acting quite late during the cytolytic and growth inhibiting processes induced by TNF.

Isolation and sequence analysis of serine protease cDNAs from mouse cytolytic T lymphocytes

Three new cDNA clones (designated MCSP-1, MCSP-2, and MCSP-3) encoding mouse serine proteases were isolated from cloned cytolytic T lymphocytes (CTL) by a modified differential screening procedure. The putative mature proteins of MCSP-2 and MCSP-3 are each composed of 228 amino acids with molecular weights of 25,477 and 25,360, respectively. NH2-terminal amino acids of MCSP-2- and MCSP-3-predicted proteins were identical to those reported for granzyme E and F, respectively. The third species, MCSP-1, was closely related to the two other cDNA species but approximately 30 amino acids equivalents of the NH2-terminal portion of the cDNA were not cloned. The amino acids forming the active sites of serine proteases were well conserved among the three predicted proteins. The active site pocket residue positioned six residues before the active-site Ser184 is alanine in MCSP-1, threonine in MCSP-2, and serine in MCSP-3, indicating that both MCSP-2 and MCSP-3 may have chymotrypsin-like specificity. There are three potential asparagine-linked glycosylation sites in MCSP-1 and MCSP-3, and four in MCSP-2-deduced amino acid sequences. Amino acid comparison of MCSP-1 with four other reported serine proteases whose active site pocket residue is alanine revealed that MCSP-1 was substantially different from the other molecules, indicating that MCSP-1 may be a new member of mouse T cell serine protease family. Antibodies made against a MCSP-1 lacZ gene fusion protein stain granules of CTL and react on immunoblots with two distinct granule protein bands of 29 and 35-40 kD. Only the 35-kD species labels with [3H]DFP. Since a protease cascade may play a key role in cytolytic lymphocyte activation, our isolation of cDNAs representative of unique serine esterases should help to investigate such a cascade process.

Distinct patterns of virus-specific T cell-mediated cytolysis of transformed versus primary target cells

Ideas about the mechanism(s) by which cytotoxic T lymphocytes (CTL) lyse appropriate target cells are still controversial. We studied the action of established murine CTL clones as well as of freshly prepared primarily induced CTL against two types of target cells. Transformed tumour cell lines (MC57G and L929) and untransformed cells such as peritoneal exudate cells (PEC) or fibroblasts were examined as target cells by scanning and transmission electron microscopy and with the 51Cr-release assay. We found independent of which CTL used, that in transformed cells the cell membrane seemed to be the first target of CTL attack, whereas in untransformed cells the first noticeable events appeared to take place in the nucleus of the target cells; the membrane of attacked untransformed PEC or fibroblasts was found to be intact at a time when the cellular organelles already were disintegrated. The morphological observations were paralleled by differences in the kinetics of 51Cr-release; untransformed target cells released their label only after a 2 h long period, whereas transformed cells released 51Cr considerably earlier.

Molecular cloning of an inducible serine esterase gene from human cytotoxic lymphocytes

A cDNA clone encoding a human serine esterase gene was isolated from a library constructed from poly(A)+ RNA of allogeneically stimulated, interleukin 2-expanded peripheral blood mononuclear cells. The clone, designated HSE26.1, represents a full-length copy of a 0.9-kilobase mRNA present in human cytotoxic cells but absent from a wide variety of noncytotoxic cell lines. Clone HSE26.1 contains an 892-base-pair sequence, including a single 741-base-pair open reading frame encoding a putative 247-residue polypeptide. The first 20 amino acids of the polypeptide form a leader sequence. The mature protein is predicted to have an unglycosylated Mr of approximately equal to 26,000 and contains a single potential site for N-linked glycosylation. The nucleotide and predicted amino acid sequences of clone HSE26.1 are homologous with all murine and human serine esterases cloned thus far but are most similar to mouse granzyme B (70% nucleotide and 68% amino acid identity). HSE26.1 protein is expressed weakly in unstimulated peripheral blood mononuclear cells but is strongly induced within 6-hr incubation in medium containing phytohemagglutinin. The data suggest that the protein encoded by HSE26.1 plays a role in cell-mediated cytotoxicity.

Homology of perforin to the ninth component of complement (C9)

Perforin is one of the cytolytic factors present in the cytoplasmic granules of mouse cytotoxic T lymphocytes and natural killer cells. We have determined the sequence of the N-terminal amino acids of perforin purified from a mouse natural killer cell line, and, by using oligonucleotide probes corresponding to the amino acid residues, we have identified a complementary DNA encoding perforin from the cDNA library of a mouse cytotoxic T lymphocyte clone. As predicted from the functional similarities between perforin and the ninth component of the serum cytolytic system, complement (C9) (refs 4-8), the deduced primary structure of perforin has homology with C9 at their respective functionally conserved regions. We find that perforin is only expressed in killer cell lines, and not in helper T lymphocytes or other tumour cells tested. Thus we have provided direct molecular evidence that a killer-cell-specific protein evolutionally linked to C9 is involved in cell-mediated cytolysis.

Automated fluorometric assay for T cell cytotoxicity

A fluorometric assay avoiding the use of radioactivity has been developed for detecting cytotoxic T lymphocytes (Tc cells). The method involves labelling targets with Hoechst dye no. 33342 (H33342) which becomes brightly fluorescent on binding to DNA. Lysis of target cells by Tc cells is quantified by measuring the release of fluorescent H33342 into the supernatant of culture wells. The fluorescence is measured using an automated Microfluor reader which allows results to be obtained rapidly. The assay has been used to detect alloreactive Tc cells and H-2 restricted Tc cells against influenza virus in a short-term 6 h assay using P815 and L929 as targets with comparable results to those obtained with 51Cr labelling. In contrast, lymphocyte blasts were found to be less sensitive in 6 h fluorometric assays when compared with the 51Cr assay. In long-term overnight assays (possible because of the low spontaneous release of H33342 from targets) lymphocyte blasts gave high specific lysis and some anti-self reactivity. The cause of the anti-self reactivity may reflect fundamental differences between the H33342 and 51Cr release assays.

Rapid changes in target cell lysosomes induced by cytotoxic T cells: indication of target suicide?

Although many studies have attempted to elucidate how cytotoxic T (Tc) lymphocytes cause the death of target cells, the mechanism is still controversial. In the present study the effect on the integrity of the lysosomes of the target cell has been investigated. We show here that the specific recognition and attachment of cloned type A influenza-specific Tc cells to A/X31 influenza virus-infected target cells caused rapid change in the amount of lysosomal naphthylamidase activity that was bound within the lysosomes, indicating that the lysosomal membranes in the target cells had been totally labilized. Target cells infected with type B influenza virus served as controls. We therefore suggest that the viral specificity of Tc lymphocytes allows for recognition and intimate membrane contact with suitably infected targets. This intimate contact induces sufficient perturbation of the target cell plasma membrane so as to cause total labilization of the target cell lysosomes which could account for intracellular lysis.

Inactivation of a T cell receptor-associated GTP-binding protein by antibody-induced modulation of the T cell receptor/CD3 complex

TCR modulation induced by anti-TCR or anti-CD3 mAbs leads to a transient state of refractoriness of the T cell to all signals given via cell surface structures. To investigate the underlying mechanisms, we have used human CTL permeabilized with the alpha toxin of S. aureus. This method of permeabilization allows manipulation of the interior milieu of the cell, but maintains its functional and structural integrity. Introduction of the G protein activator GTP gamma S into permeabilized CTL leads to triggering of granule exocytosis. The G protein inactivator GDP beta S inhibited exocytosis induced by TCR triggering but not that induced by activation of protein kinase C. This indicates that the G protein that triggers exocytosis is localized after CD3 triggering but before formation of the polyphosphoinositol breakdown product diacylglycerol. In TCR-modulated CTL, GTP gamma S is no longer able to activate exocytosis. Such CTL, however, still respond to PKC activators. This demonstrates that a TCR-associated G protein has been functionally inactivated by TCR modulation.

A T cell- and natural killer cell-specific, trypsin-like serine protease. Implications of a cytolytic cascade

A new trypsin-like serine protease was cloned from both a murine cytotoxic T lymphocyte and a human PHA-stimulated peripheral blood lymphocyte cDNA library. In both the mouse and human system, this transcript had a T cell- and NK-specific distribution, being detected in cytotoxic T lymphocytes (CTL), some T-helper clones, and NK, but not in a variety of normal tissues. T-cell activation with Con A plus IL-2 induced mouse spleen cells to express this gene with kinetics correlating with the acquisition of cytolytic capacity. Both the mouse and human nucleotide sequences of this gene encoded an amino acid sequence with 25-40% identity to members of the serine protease family. The active-site "charge-relay" residues (His-57, Asp-102, and Ser-195 of the chymotrypsin numbering system) are conserved, as well as the trypsin-specific Asp (position 189 in trypsin). We reviewed the evidence of this serine protease's role in lymphocyte lysis and proposed a "lytic cascade." We discussed the biological and clinical implications of a cascade, proposing these enzymes as markers for cytolytic cells and as targets for rational drug therapy. Genetic and acquired deficits in the lethal hit-delivery system are considered as a basis for approaching some immunodeficiency states, including severe EBV infections, T-gamma leukemias, and T8+ lymphocytosis syndromes.

Age-related decline in the lethal hit but not the binding stage of cytotoxic T-cell activity in mice

The activity of allogeneic cytotoxic T lymphocytes (CTL) declines with aging, as measured at the population level in chromium release assays, and is associated with a decreased proliferative response to the inducing alloantigens. At the single cell level, no difference was noted between target binding cells obtained from alloantigen-stimulated spleen cells of young and old mice. However, a consistent, large, and significant difference was observed between CTL generated from the spleens of young and old mice in the percentage of bound effector cells with lysed targets, i.e., the percentage of killer cells. These results show age-related changes in the lethal hit stage and not the binding stage of CTL activity. This change could be used as a probe to identify the important mechanism(s) which mediates CTL activity.

Identification and sequencing of cDNA clones encoding the granule-associated serine proteases granzymes D, E, and F of cytolytic T lymphocytes

Cytoplasmic granules of cytolytic T lymphocytes contain at least six related serine esterases (granzymes) that are released together with perforin, a pore-forming protein related to complement component C9, during target-cell lysis. Polyclonal antibodies were used to isolate a large number of cDNA clones from an expression library derived from cytolytic-T-cell mRNA. Three distinct full-length cDNA clones coding for granzymes D, E, and F were identified by restriction site mapping and nucleotide sequencing. The three deduced amino acid sequences are highly similar to one another (between 72% and 90% amino acid identities) and to the sequences of granzymes B and C, cathepsin G, and rat mast-cell proteases I and II (between 43% and 57% amino acid identities). Cysteine residues capable of forming intramolecular disulfide bonds are conserved, as are the catalytic-site residues characteristic of serine proteases. Comparison of the cDNA-derived protein sequences with the amino termini of the isolated granzymes provides evidence that they are stored in a fully processed, activated form after removal of the signal peptide and two additional residues (propeptide) at the amino terminus. Immunoelectron microscopic studies demonstrated that granzymes D, E, and F are present in the same morphologically distinct cytoplasmic granules in which perforin has been found previously.

Determination of frequency of T cells expressing the T cell-specific serine proteinase 1 (TSP-1) reveals two types of L3T4+ T lymphocytes

TSP-1 is a murine T cell-specific serine proteinase which is exclusively expressed in activated but not in resting T lymphocytes. Among T lymphocyte clones tested so far the enzyme was found to be associated with all Ly-2+ but only with a fraction of L3T4+ lines. Here we have applied a limiting dilution system to determine the frequency of precursor cells of resting L3T4+ and Ly-2+ lymphocytes which can be induced in vitro by antigen/lectin to express TSP-1. T cell subsets were either positively enriched by flow cytofluorometry cell sorting or by negative selection using monoclonal antibodies and complement. Following stimulation of lymphocytes in vitro, individual microwells were tested for growth by visual examination and for the TSP-1 protein/enzyme by analyzing cell lysates using either a specific rabbit anti-TSP-1 antiserum and/or the chromogenic model peptide substrate H-D-Pro-Phe-Arg-p-nitroanilide. Moreover, a large panel of L3T4+ and Ly-2+ T lymphocyte clones generated from primary cultures were similarly investigated. In some cell cultures the presence of TSP-1 was also tested on the mRNA level using a TSP-1-specific oligonucleotide probe. The data show that the majority, if not all, of antigen/lectin-induced-Ly-2+ T cells expressed TSP-1. In contrast, only 12%-27% of the growing lectin or antigen-reactive L3T4+ lymphocytes were positive for the enzyme. Studies performed in parallel with L3T4+ and Ly-2+ lymphocyte populations sensitized in bulk culture showed that under these conditions similar levels of TSP-1-specific mRNA and enzyme activity are detected in both subsets. The finding of primary L3T4+ T lymphocyte clones with distinct patterns of TSP-1 production provides evidence for the existence of two types of L3T4+ effector cells with different functional capacities. The data also suggest a cooperation between distinct L3T4+ lymphocytes for induction of optimal TSP-1 activity in L3T4+ T cells.

Phenylmethylsulfonyl fluoride inhibits chemotactic peptide-induced actin polymerization and oxidative burst activity in human neutrophils by an effect unrelated to its anti-proteinase activity

Stimulation of polymorphonuclear leukocytes with the chemotactic peptide N-formylmethionylleucylphenylalanine (fMet-Leu-Phe) causes conversion of monomeric actin to polymeric actin. We studied the role of proteinase inhibitors phenylmethylsulfonyl fluoride PMSF) and diisopropyl fluorophosphate in fMet-Leu-Phe-induced actin polymerization in polymorphonuclear leukocytes. Pre-incubation of cells with PMSF (2 mM) for 1 min caused inhibition of fMet-Leu-Phe-induced actin polymerization, as studied by 7-nitrobenz-2-oxa-1,3-diazole (NBD) -phallacidin labeling and flow cytometry. PMSF also inhibited fMet-Leu-Phe-induced hydrogen peroxide release, superoxide anion generation and chemiluminescence. In contrast, diisopropyl fluorophosphate (5 mM) was unable to inhibit fMet-Leu-Phe-induced actin polymerization and superoxide generation, but was effective in inhibiting hydrogen peroxide production and chemiluminescence. PMSF did not cause any change in membrane potential by itself and failed to inhibit the membrane potential changes induced by fMet-Leu-Phe, indicating that PMSF does not affect the binding of fMet-Leu-Phe to the receptors. The high concentration of PMSF required coupled with the fact that diisopropyl fluorophosphate was unable to inhibit fMet-Leu-Phe-induced actin polymerization suggested that this activity of PMSF might be unrelated to proteinase inhibitory activity. Polymyxin B, a membrane-active antibiotic, had an effect similar to PMSF on fMet-Leu-Phe-induced actin polymerization. This suggests that PMSF may also be acting via its membrane effect rather than its anti-proteinase effect.

Mechanism of target cell lysis by cytotoxic T lymphocytes (CTL) employing RSV-induced H-2 congenic and recombinant mouse tumor cells: demonstration of soluble mediators released from H-2-restricted CTL clones

The secretion and the specificity of cytotoxic mediators from H-2-restricted cytotoxic T lymphocytes (CTL) were examined using non-virus-producing target tumor cells induced by the Schmidt-Ruppin strain of Rous sarcoma virus (SR-RSV) in B10 congenic and recombinant mice. By using rat concanavalin A supernatant, two H-2-restricted CTL clones were established from cytotoxic effector cells of B10.A(5R) mice primed with SR-RSV-induced syngeneic tumor Cell-free supernatants from the H-2-restricted CTL clones cocultured with syngeneic tumor cells had selectively high cytotoxic activity for syngeneic and H-2-compatible tumor cells, but not for H-2-incompatible tumor cells. YAC-1 cells, and B10.A(5R) blasts as defined in the 5-hr 51Cr-release assay. The cytotoxic activity was detected in the cell-free supernatants from the CTL clones cocultured with the CTL-sensitive syngeneic and H-2-compatible tumor cells, but not with the CTL-insensitive tumor cells and YAC-1 cells. The cytotoxic activity of the cell-free supernatant could be adsorbed by the syngeneic tumor cells, but not by YAC-1 and L(s) cells. Thus, the H-2-restricted CTL clones against SR-RSV-induced tumor cells were capable of releasing cytotoxic mediators by coculturing with syngeneic or H-2-compatible tumor cells, and the cytotoxic mediators showed a certain H-2-restricted manner in killing the target cells. These results suggest that the lysis of RSV-induced tumor cells by H-2-restricted CTL can at least in part be mediated by cytotoxic factors.

Mechanistic, functional and immunopharmacological implications of biochemical studies of antigen receptor-triggered cytolytic T-lymphocyte activation

Biochemical events that follow the engagement of cytotoxic T lymphocytes (CTL) with an Ag-bearing target cell (TC) or triggering by the crosslinking of the Ag-receptor (TcR) by immobilized anti-TcR mAb were studied using cloned CTL and a novel CTL activation assay. The approach described here was undertaken to shed light on the molecular mechanisms of "ON", "STOP" and "OFF" signalling that allow CTL to be activated, kill TC and disengage from the target cell after delivery of the "lethal hit" and then to proceed with the destruction of the next Ag-bearing target encountered. Biochemical studies of TcR-regulated and TcR-triggered constitutive exocytosis in CTL provided a detailed description of the molecular requirements for this important phenomenon in T lymphocytes and provided an alternative CTL activation assay; this assay measures the TcR-dependent response in the absence of a TC. These studies also helped to envision CTLs screening activities as a cycle of engagements-disengagements with the TC, where every surrounding cell is treated by the CTL as a potential Ag-bearing TC. Both constitutive and regulated exocytosis in CTL are triggered through a transmembrane signalling pathway which involves protein kinase C and extracellular Ca2+ that, most likely, is translocated through Ca2+ channels. This is followed by the involvement of calmodulin (CaM)-binding proteins, e.g., calcineurin, a CaM-dependent phosphatase, which was shown to be a major CaM-binding protein in murine lymphocytes. Unexpectedly, these biochemical studies demonstrated that the granule exocytosis model of CTL-mediated cytotoxicity cannot account for the mechanism of target cell lysis by CTL, at least in in vitro conditions in the absence of extracellular Ca2+. These results indicate the existence of an extracellular Ca2+-independent, TcR-regulated CTL response and raise the possibility that second messenger(s) other than Ca2+ and/or products of phosphoinositide turnover are involved in T-cell lysis. Predominance of "non-lethal" engagements between some CTL and TC, revealed during time-lapse cinematographic studies, together with comparative studies of TcR-regulated exocytosis of granules and of constitutive exocytosis of gamma-interferon, suggested that TC destruction by CTL may not be their only or even their most important function in vivo. It is possible that CTL, triggered by Ag recognition to exocytose storage granules and to synthesize and constitutively exocytose macrophage-activating factors, in turn promote tumor destruction by macrophages.(ABSTRACT TRUNCATED AT 400 WORDS)

A high proportion of T lymphocytes that infiltrate H-2-incompatible heart allografts in vivo express genes encoding cytotoxic cell-specific serine proteases, but do not express the MEL-14-defined lymph node homing receptor

The role of cytotoxic cells in in vivo immune functions such as allograft rejection is unknown. To begin to assess the function of cytolytic cells in vivo we have begun with cytolytic cell-specific functional molecules: we have isolated and characterized cytolytic cell-specific cDNA clones from cytolytic T cell clones, both encoding distinct serine esterases. The HF gene encodes a trypsin-like enzyme while the C11 gene encodes an enzyme with likely specificity for acidic residues. Here we demonstrate, using in situ hybridization with RNA probe, that both genes are expressed selectively in a subset of T lymphocytes that have infiltrated cardiac allografts. The phenotype of these cells is consistent with the most frequent phenotype of active CTL raised in vitro: they are predominantly CD4-, CD8+, MEL-14- T cell blasts. Thus the expression of these genes, each of which encodes serine esterase found in killer cell granules in vitro, is a valid marker for these cells in vivo as well. The kinetics of their accumulation is consistent with, but not proof of, a putative role in allograft rejection. It is likely that HF and C11 gene expression will be of diagnostic value.

Granzymes, a family of serine proteases released from granules of cytolytic T lymphocytes upon T cell receptor stimulation

The cytolytic potential of T effector cells appears to be intimately related to the presence of proteins stored in specialized cytoplasmic granules. A striking biological property of isolated granules is their lytic activity for a variety of target cells in a nonrestricted manner. Proteins contained within these granules of CTLs are specifically released upon target cell recognition. We have isolated and characterized six granule-associated proteins in two murine CTL lines in addition to the pore-forming and target membrane-disrupting perforin. Six full length cDNA clones have been identified in a CTL-specific cDNA expression library which code for the granule-associated serine esterases, designated as granzymes A to F. Granzymes A and B represent the genuine proteins encoded by the H factor/CTLA-3 cDNA and the CTLA-1/CCPI cDNA, respectively. The covalent amino acid structures of all six granzymes show the hallmarks for serine proteases and are highly related to that of rat mast cell protease I and II and cathepsin G, which have been found in granules of mast cells and neutrophilic granulocytes, respectively. The primary translation products are processed by removal of a hydrophobic signal peptide and a two residue-long propeptide at the amino-terminus. Immuno-electron microscopy shows that granzymes and perforin are stored together within secretory granules of CTLs. Simultaneous release of at least two of these granzymes has been observed during degranulation of a murine CTL line by anti-T3 antibodies. The biological role, particularly the proteolytic events elicited by granzyme A and other granzymes in the context of target cell recognition, are not known at present. It is unlikely that they form a proteolytic activation cascade together with pore-forming proteins analogous to the complement system. The strictly regulated secretion of granzymes and the lack of measurable enzymatic activity in the case of granzymes B, C, E and F towards a variety of synthetic substrates suggest a highly specific function for each of them.

Perforin-dependent and -independent pathways of cytotoxicity mediated by lymphocytes

There is little doubt at the present time that both perforin-dependent and -independent pathways are important in mediating the cytotoxicity associated with lymphocytes. The cell distribution of perforin, initially thought to include both CTL and NK cells, now must be viewed with caution because all previous biochemical studies on CTL have been conducted with cell lines propagated in long-term cultures in the presence of T cell growth factors (IL-2 and perhaps some still undefined factors). Under these conditions, CTL are known to assume a broader, NK-like specificity in target cell killing and may thus differ significantly from primary CTL generated in the body. Accordingly, perforin does not seem to be present in primary CTL activated directly through mixed lymphocyte reactions. It remains to be shown how primary CTL lyse target cells in vivo. Initial studies conducted in several laboratories have already provided some clues. It now seems that even in cultured, perforin-containing CTL, the perforin pathway is not an obligatory mechanism required for target cell killing. Other pathways, possibly involving TNF/lymphotoxin-like molecules, may play a direct role in this type of cytotoxicity. Other still unidentified factors now also need to be sought, including membrane polypeptides that may develop cytotoxicity directly upon cell contact and binding. Although from the studies reviewed here it is clear now that perforin has a more limited role in cell killing than originally proposed, it is still intriguing that it should share structural and functional homologies with complement proteins, drawing paradoxical analogies between two systems (the cellular and the humoral immune systems) which have evolved to become specialized to carry out separate immunological tasks. The cloning of the genes for perforin and for all the C proteins that comprise the MAC should reveal important information on how these genes originated and then diverged during evolution. The cellular distribution of other granule products, such as serine esterases, also must be viewed with caution. A serine esterase activity was initially thought to be CTL-specific. This information stimulated an intensive research activity in many laboratories that resulted in both the purification of a serine esterase family and the cloning of several serine esterase transcripts. It is becoming clear from recent evidence that this group of enzymes is not truly CTL-specific and therefore would not be expected to develop any function rendered absolutely necessary for cytolysis.(ABSTRACT TRUNCATED AT 400 WORDS)

A differential molecular biology search for genes preferentially expressed in functional T lymphocytes: the CTLA genes

One approach to the isolation of molecules involved in T cell-mediated cytolysis stems from the postulate of a possible correlation between molecular phenotype and molecular functional involvement. Accordingly, CTL-specific molecules have been looked for, using a strategy based on the differential screening of a subtracted cDNA library. This led to the isolation and characterization of the following structures, expressed mostly (but no exclusively) in CTLs and inducible upon lymphocyte activation: CTLA-1 and CTLA-3 (serine-proteases), CTLA-4 (a member of the Ig superfamily) and CTLA-2 alpha and beta (homologues to the proregion of cysteine-proteases). The theoretical and practical limitations and the prospects of this type of approach are discussed.