Eric M. Tam

Multi-step pericellular proteolysis controls the transition from individual to collective cancer cell invasion.

Invasive cell migration through tissue barriers requires pericellular remodelling of extracellular matrix (ECM) executed by cell-surface proteases, particularly membrane-type-1 matrix metalloproteinase (MT1-MMP/MMP-14). Using time-resolved multimodal microscopy, we show how invasive HT-1080 fibrosarcoma and MDA-MB-231 breast cancer cells coordinate mechanotransduction and fibrillar collagen remodelling by segregating the anterior force-generating leading edge containing β1 integrin, MT1-MMP and F-actin from a posterior proteolytic zone executing fibre breakdown. During forward movement, sterically impeding fibres are selectively realigned into microtracks of single-cell calibre. Microtracks become expanded by multiple following cells by means of the large-scale degradation of lateral ECM interfaces, ultimately prompting transition towards collective invasion similar to that in vivo. Both ECM track widening and transition to multicellular invasion are dependent on MT1-MMP-mediated collagenolysis, shown by broad-spectrum protease inhibition and RNA interference. Thus, invasive migration and proteolytic ECM remodelling are interdependent processes that control tissue micropatterning and macropatterning and, consequently, individual and collective cell migration.

Characterization of the Distinct Collagen Binding, Helicase and Cleavage Mechanisms of Matrix Metalloproteinase 2 and 14 (Gelatinase A and MT1-MMP) THE DIFFERENTIAL ROLES OF THE MMP HEMOPEXIN C DOMAINS AND THE MMP-2 FIBRONECTIN TYPE II MODULES IN COLLAGEN TRIPLE HELICASE ACTIVITIES

Matrix metalloproteinase-2 (MMP-2, gelatinase A) and membrane type (MT)1-MMP (MMP-14) are cooperative dynamic components of a cell surface proteolytic axis involved in regulating the cellular signaling environment and pericellular collagen homeostasis. Although MT1-MMP exhibits type I collagenolytic but poor gelatinolytic activities, MMP-2 is a potent gelatinase with weak type I collagenolytic behavior. Recombinant linker/hemopexin C domain (LCD) of MT1-MMP binds native type I collagen, blocks MT1-MMP collagenolytic activity in trans, and by circular dichroism spectroscopy, induces localized structural perturbation in the collagen. These changes were reflected by enhanced cleavage of the MT1-LCD-bound collagen by the collagenases MMP-1 and MMP-8 but not by trypsin or MMP-7. Thus, the MT1-LCD alone can initiate triple helicase activity. In contrast, the native and denatured collagen binding properties of MMP-2 reside in the fibronectin type II modules, accordingly termed the collagen binding domain (CBD). Recombinant CBD (but not the MMP-2 LCD) also changed the circular dichroism spectra leading to increased MMP-1 and -8 cleavage of native collagen. However, recombinant CBD reduced gelatin and collagen cleavage by MMP-2 in trans as did CBD23, which comprises the second and third fibronectin type II modules, but not the CBD23 mutant W316A/W374A, which neither binds gelatin nor collagen. This indicates that MMP-2 and MT1-MMP bind collagen at a different site than MMP-1 and MMP-8. Thus, MMP-2 utilizes the CBD in cis for collagen binding and triple helicase activity, which compensates for the lack of collagen binding by the MMP-2 LCD. Hence, the MMP family has evolved two distinct mechanisms for collagen triple helicase activity using two structurally distinct domains, with triple helicase activity occurring independent of α-chain hydrolysis.

Pharmacoproteomics of a Metalloproteinase Hydroxamate Inhibitor in Breast Cancer Cells: Dynamics of Membrane Type 1 Matrix Metalloproteinase-Mediated Membrane Protein Shedding

ABSTRACT Broad-spectrum matrix metalloproteinase (MMP) inhibitors (MMPI) were unsuccessful in cancer clinical trials, partly due to side effects resulting from limited knowledge of the full repertoire of MMP substrates, termed the substrate degradome, and hence the in vivo functions of MMPs. To gain further insight into the degradome of MMP-14 (membrane type 1 MMP) an MMPI, prinomastat (drug code AG3340), was used to reduce proteolytic processing and ectodomain shedding in human MDA-MB-231 breast cancer cells transfected with MMP-14. We report a quantitative proteomic evaluation of the targets and effects of the inhibitor in this cell-based system. Proteins in cell-conditioned medium (the secretome) and membrane fractions with levels that were modulated by the MMPI were identified by isotope-coded affinity tag (ICAT) labeling and tandem mass spectrometry. Comparisons of the expression of MMP-14 with that of a vector control resulted in increased MMP-14/vector ICAT ratios for many proteins in conditioned medium, indicating MMP-14-mediated ectodomain shedding. Following MMPI treatment, the MMPI/vehicle ICAT ratio was reversed, suggesting that MMP-14-mediated shedding of these proteins was blocked by the inhibitor. The reduction in shedding or the release of substrates from pericellular sites in the presence of the MMPI was frequently accompanied by the accumulation of the protein in the plasma membrane, as indicated by high MMPI/vehicle ICAT ratios. Considered together, this is a strong predictor of biologically relevant substrates cleaved in the cellular context that led to the identification of many undescribed MMP-14 substrates, 20 of which we validated biochemically, including DJ-1, galectin-1, Hsp90α, pentraxin 3, progranulin, Cyr61, peptidyl-prolyl cis-trans isomerase A, and dickkopf-1. Other proteins with altered levels, such as Kunitz-type protease inhibitor 1 and beta-2-microglobulin, were not substrates in biochemical assays, suggesting an indirect affect of the MMPI, which might be important in drug development as biomarkers or, in preclinical phases, to predict systemic drug actions and adverse side effects. Hence, this approach describes the dynamic pattern of cell membrane ectodomain shedding and its perturbation upon metalloproteinase drug treatment.

Protease degradomics: mass spectrometry discovery of protease substrates and the CLIP-CHIP, a dedicated DNA microarray of all human proteases and inhibitors.

Abstract The biological role of most proteases in vivo is largely unknown. Therefore, to develop robust techniques to analyze the protease degradome in cells and tissues and to elucidate their substrate degradomes we have developed a dedicated and complete human protease and inhibitor microarray that we have called the CLIP-CHIP. Oligonucleotides (70-mers) for identifying all 715 human proteases, inactive homologs and inhibitors were spotted in triplicate onto glass slides with a dedicated subarray containing oligonucleotides for specific human breast carcinoma genes. Initial analyses revealed the elevated expression of a number of proteases in invasive ductal cell carcinoma including ADAMTS17, carboxypeptidases A5 and M, tryptasegamma and matriptase-2. Matrix metalloproteinases (MMPs) showed a restricted expression pattern in both normal and cancerous breast tissues with most expressed at low levels. However, of the several MMPs expressed in significant quantities, the carcinoma samples showed only slightly elevated amounts other than for MMP-28 which was strongly elevated. To discover new protease substrates we developed a novel yeast twohybrid approach we term inactive catalytic domain capture (ICDC). Here, an inactive mutant protease catalytic domain lacking the propeptide was used as a yeast two hybrid bait to screen a human fibroblast cDNA library for interactor proteins as a substrate trap. Wntinduced signaling protein-2 (WISP-2) was identified by ICDC and was biochemically confirmed as a new MMP substrate. In another approach we used isotopecoded affinity tag (ICAT) labeling with tandem mass spectrometry to quantitate the levels of secreted or shed extracellular proteins in MDAMB-231 breast carcinoma cell cultures in the presence or absence of membrane type 1-MMP (MT1-MMP) overexpression. By this proteomic approach we identified and biochemically confirmed that IL-8, the serine protease inhibitor SLPI, the death receptor-6, proTNFα and CTGF are novel substrates of MT1-MMP. The utility and quantitative nature of ICAT with MS/MS analysis as a new screen for protease substrate discovery based on detection of cleaved or shed substrate products should be readily adaptable to other classes of protease for assessing proteolytic function in a cellular context.

Domain Interactions in the Gelatinase A·TIMP-2·MT1-MMP Activation Complex THE ECTODOMAIN OF THE 44-kDa FORM OF MEMBRANE TYPE-1 MATRIX METALLOPROTEINASE DOES NOT MODULATE GELATINASE A ACTIVATION

On the cell surface, the 59-kDa membrane type 1-matrix metalloproteinase (MT1-MMP) activates the 72-kDa progelatinase A (MMP-2) after binding the tissue inhibitor of metalloproteinases (TIMP)-2. A 44-kDa remnant of MT1-MMP, with an N terminus at Gly285, is also present on the cell after autolytic shedding of the catalytic domain from the hemopexin carboxyl (C) domain, but its role in gelatinase A activation is unknown. We investigated intermolecular interactions in the gelatinase A activation complex using recombinant proteins, domains, and peptides, yeast two-hybrid analysis, solid- and solution-phase assays, cell culture, and immunocytochemistry. A strong interaction between the TIMP-2 C domain (Glu153-Pro221) and the gelatinase A hemopexin C domain (Gly446-Cys660) was demonstrated by the yeast two-hybrid system. Epitope masking studies showed that the anionic TIMP-2 C tail lost immunoreactivity after binding, indicating that the tail was buried in the complex. Using recombinant MT1-MMP hemopexin C domain (Gly285-Cys508), no direct role for the 44-kDa form of MT1-MMP in cell surface activation of progelatinase A was found. Exogenous hemopexin C domain of gelatinase A, but not that of MT1-MMP, blocked the cleavage of the 68-kDa gelatinase A activation intermediate to the fully active 66-kDa enzyme by concanavalin A-stimulated cells. The MT1-MMP hemopexin C domain did not form homodimers nor did it bind the gelatinase A hemopexin C domain, the C tail of TIMP-2, or full-length TIMP-2. Hence, the ectodomain of the remnant 44-kDa form of MT1-MMP appears to play little if any role in the activation of gelatinase A favoring the hypothesis that it accumulates on the cell surface as an inactive, stable degradation product.

Cell surface chondroitin sulfate glycosaminoglycan in melanoma: role in the activation of pro-MMP-2 (pro-gelatinase A)

We previously reported that CS (chondroitin sulfate) GAG (glycosaminoglycan), expressed on MCSP (melanoma-specific CS proteoglycan), is important for regulating MT3-MMP [membrane-type 3 MMP (matrix metalloproteinase)]-mediated human melanoma invasion and gelatinolytic activity in vitro. In the present study, we sought to determine if CS can directly enhance MT3-MMP-mediated activation of pro-MMP-2. Co-immunoprecipitation studies suggest that MCSP forms a complex with MT3-MMP and MMP-2 on melanoma cell surface. When melanoma cells were treated with betaDX (p-nitro-beta-D-xylopyranoside) to inhibit coupling of CS on the core protein, both active form and proform of MMP-2 were no longer co-immunoprecipitated with either MCSP or MT3-MMP, suggesting a model in which CS directly binds to MMP-2 and presents the gelatinase to MT3-MMP to be activated. By using recombinant proteins, we determined that MT3-MMP directly activates pro-MMP-2 and that this activation requires the interaction of the C-terminal domain of pro-MMP-2 with MT3-MMP. Activation of pro-MMP-2 by suboptimal concentrations of MT3-MMP is also significantly enhanced in the presence of excess C4S (chondroitin 4-sulfate), whereas C6S (chondroitin 6-sulfate) or low-molecular-mass hyaluronan was ineffective. Affinity chromatography studies using CS isolated from aggrecan indicate that the catalytic domain of MT3-MMP and the C-terminal domain of MMP-2 directly bind to the GAG. Thus the direct binding of pro-MMP-2 with CS through the C-domain would present the catalytic domain of pro-MMP-2 to MT3-MMP, which facilitates the generation of the active form of MMP-2. These results suggest that C4S, which is expressed on tumour cell surface, can function to bind to pro-MMP-2 and facilitate its activation by MT3-MMP-expressing tumour cells to enhance invasion and metastasis.

Identification of the TIMP‐2 Binding Site on the Gelatinase A Hemopexin C‐Domain by Site‐Directed Mutagenesis and the Yeast Two‐Hybrid System

Concanavalin A, 1 PMA, 2 and collagen 3 can induce the cellular activation of the matrix metalloproteinase (MMP) gelatinase A (MMP-2). In this process, gelatinase A utilizes tissue inhibitor of metalloproteinases-2 (TIMP-2) as an adapter to dock with membrane-type MMP (MT-MMP) on the cell membrane. 4 Here, the NH 2 -domain of TIMP-2 binds to the active site of the MT-MMP with the TIMP-2 COOHdomain thought to bind to the gelatinase A COOH terminal hemopexin domain (Cdomain). A second, uncomplexed, furin-activated MT-MMP then initiates gelatinase A activation by cleavage of the prodomain. Activation is completed by autolytic intraand intermolecular gelatinase A cleavages (reviewed in Ref. 4). These interactions in the quaternary activation complex are specifically required since progelatinase A cell surface binding to β 1 integrin-bound collagen via the triple fibronectin type II repeats of the enzyme is not only insufficient for activation, but in fact protects progelatinase A from activation. 5 In this study we sought to identify the TIMP-2 binding site on the gelatinase A C-domain in molecular detail. First, we confirmed binding of the recombinant (r) gelatinase A C-domain protein 6 with full-length TIMP-2 (F IG . 1A). Notably, TIMP-2 did not bind to recombinant C-domain of MT1-MMP. Next, the binding of recombinant TIMP-2 NH 2 -domain protein 7 with the gelatinase A rC-domain protein 6 was studied. By both solid-phase assays and affinity chromatography, a weak binding interaction was demonstrated between these two domains that was disrupted by 0.3–0.5 M NaCl. In contrast, TIMP-1 NH 2 -domain protein did not bind. However, it is known that gelatinase A activation does not occur using TIMP-2 NH 2 domain protein in place of full-length TIMP-2. 8 This indicates that the TIMP-2 C-

Noncompetitive Inhibition of Hepatocyte Growth Factor-dependent Met Signaling by a Phage-derived Peptide

Dysregulation of hepatocyte growth factor (HGF)-induced signaling via its receptor tyrosine kinase Met results in tumor progression and metastasis. To initiate signaling, pro-HGF must be proteolytically activated to reveal a secondary Met binding site within the serine protease-like beta-chain of HGF. Although HGF/Met is a large complex, we sought to discover relatively small antagonists that might interfere with this critical Met binding region. Pools of disulfide-constrained random peptide libraries displayed on phage were selected for binding to HGF, ultimately resulting in a disulfide-constrained 15-mer peptide (VNWVCFRDVGCDWVL) termed HB10, which bound to the recombinant human HGF beta-chain (HGF beta) and competitively inhibited binding to Met with an IC(50) of 450 nM. In MDA-MB435 cells, HB10 reduced HGF-dependent Met phosphorylation by 70%, and phosphorylation of downstream kinases AKT and ERK1/ERK2 by 74% and 69%, respectively. Addition of HB10 also inhibited HGF-dependent migration of these cells with an IC(50) of approximately 20 microM. The 2D (1)H-NMR structure of HB10 revealed a beta-hairpin loop stabilized by the disulfide bond and cross-strand pairing of Trp3 and Trp13. HGF beta mutants deficient in Met binding also have reduced HB10 binding, suggesting an overlapping binding site. Notably HB10 did not inhibit full length HGF binding to Met. Thus steric hindrance of the interaction between HGF beta domain binding to Met is sufficient for inhibiting full-length HGF-dependent Met signaling and cell migration that is consistent with a noncompetitive inhibitory mechanism of Met signal transduction.

Abstract 3842: Design and engineering of TRAIL fusion proteins for cancer therapy

Protein-based agonists of apoptotic death receptors have shown remarkable preclinical efficacy but limited clinical response. The short circulating half-life of recombinant human TRAIL and the necessity of Fc-mediated clustering for potentiating agonistic antibodies against DR4 and DR5 have been proposed to be major impediments to the clinical success of this class. To address these limitations we have created Fc-scTRAIL, a single fusion polypeptide consisting of an IgG1 Fc region followed by three successive TRAIL monomers connected by two fifteen-amino acid linkers. While Fc-scTRAIL showed potent activity in vitro, we observed a low TM (48 °C) and rapid inactivation in serum indicating protein instability. Subsequently, we applied a directed evolution approach using yeast surface display to identify mutations that would stabilize the TRAIL trimer. When individual mutations were transferred to the Fc-scTRAIL format, we observed a dramatic increase in the TM (66-70 °C) while the combination of three mutations improved serum stability by ten-fold. Stabilized Fc-scTRAIL shows greater pro-apoptotic activity across a panel of cancer cell lines when compared to mapatumumab (anti-DR4) and drozitumab (anti-DR5), or the combination of antibodies even in the presence of anti-Fc cross-linking. Moreover, anti-Fc did not improve Fc-scTRAIL activity suggesting that the hexavalent design of the molecule maximizes death receptor activation. Currently, in vivo evaluation of Fc-scTRAIL for pharmacokinetic properties and activity is underway. We believe this format, when combined with an appropriate patient selection strategy, will result in improved clinical outcomes. Citation Format: Eric M. Tam, Hannah Hudson, Tamara Dake, Sara Ghassemifar, Andreas Raue, Yasmin Hashambhoy-Ramsay, Stephen L. Sazinsky, Anahita Daruwalla, Neeraj Kohli, Lihui Xu, Charlotte F. Mc Donagh, Birgit Schoeberl, Diana H. Chai. Design and engineering of TRAIL fusion proteins for cancer therapy. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3842.

Abstract B1-29: The effect of higher-order receptor clusters on TRAIL induced apoptotic signaling

The trimeric TNF-related apoptosis-inducing ligand (TRAIL) is an endogenous ligand that binds to trimeric death receptors (DR4/5). TRAIL is known to induce apoptosis mainly in malignant cells while normal cells remain unaffected. This makes TRAIL a most interesting target for cancer therapy. We investigate the effects of ligand valency on the degree of Caspase-8 activation and apoptosis induction. As a model system we use the semi-sensitive cell line DU145 and multivalent anti-DR5 fibronectin domains that can potentially form higher-order receptor clusters. In order to disentangle the effects of receptor clusters of different size, a dynamic model of ligand-receptor binding and Caspase-8 activation is coupled to a model of cell growth and cell death. The dynamic model is calibrated using quantitative experimental data of ligand on cell binding, Caspase-8 activity and cell viability. Our results show that higher-order receptor clusters do amplify apoptotic signaling. The model predicts that receptor clusters of size three and six are more abundant than others, indicating the special role of preformed receptor trimers. However, Caspase-8 activation only increases after receptor clusters of size larger than three. On the cell population level, the model can quantitatively predict the induction of apoptosis in this cell line. This work is an important step towards an integrative model of TRAIL induced apoptosis that finally aims at the understanding of heterogeneity both on a single cell level across cell lines. Citation Format: Andreas Raue, Diana Chai, Hannah Hudson, Eric Tam, Birgit Schoeberl. The effect of higher-order receptor clusters on TRAIL induced apoptotic signaling. [abstract]. In: Proceedings of the AACR Special Conference on Computational and Systems Biology of Cancer; Feb 8-11 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 2):Abstract nr B1-29.