Leukemia (2001) 15, 1217–1224  2001 Nature Publishing Group All rights reserved 0887-6924/01 $15.00 www.nature.com/leu A new matrix metalloproteinase inhibitor SI-27 induces apoptosis in several human myeloid leukemia cell lines and enhances sensitivity to TNF alpha-induced apoptosis Y Nakamura1, K Sato1, N Wakimoto1, F Kimura1, A Okuyama2 and K Motoyoshi1

1Third Department of Internal Medicine, National Defense Medical College, Tokorozawa, Saitama; and 2Banyu Tsukuba Research Institute, Tsukuba, Ibaragi, Japan

MMP inhibitors are used clinically for the stabilization of tumor The abnormal MMP/TIMP expression in hematopoietic malig- growth, thus prolonging survival in cancer patients. However, nancies may indicate the premature egress of leukemic their role in the treatment of hematopoietic malignancies remains unclear. In the present study, we investigated the blasts from bone marrow and their dissemination into effects of a new MMP inhibitor, SI-27, in hematopoietic malig- peripheral tissues. Consequently, MMP inhibitors should be nancies. SI-27 alone induces apoptosis in several human candidates for agents in the treatment of hematopoietic myeloid leukemia cell lines such as U937, NB4, and HL60 cells malignancies. by activating caspase 8, 9, and 3. Apoptosis was measured with To date, most patients with acute myeloid leukemia achieve annexin V positive staining, a drop in mitochondrial transmem- complete remission with cytotoxic chemotherapy, but, due to brane potential (⌬⌿m), presence of hypodiploid DNA, and cleavage of PARP and I␬B␣. Furthermore, at lowered concen- recurrence, long-term survival is limited. Anti-cancer drugs act trations, which did not directly induce apoptosis, SI-27 acted to induce apoptosis in cancer cells. However, they also acti- to sensitize U937 cells and other cells to tumor necrosis factor vate some anti-apoptotic pathways in cancer cells, such as ␣ (TNF-␣)-mediated apoptosis. The accumulation of membrane NF-␬B, thus blocking apoptosis.11 This is at least one reason Fas, the Fas ligand, and TNFR1 were not apparent due to for the inefficient elimination of leukemia cells by anti-cancer exposure to SI-27, and antagonistic anti-Fas or anti-Fas ligand drugs, leading to recurrence. Therefore, new drugs that cause antibodies did not block SI-27-induced apoptosis. Thus, SI-27- induced apoptosis is not mediated by the Fas pathway. These apoptosis without activating the survival signal pathway are results suggest that MMP inhibitors, alone or in combination required. with other cytotoxic agents, can provide a unique method for Naito et al12 isolated [L-N-(N-hydroxy-2-isobutylsuccinyn- treating acute myeloid leukemia, refractory to classical anti- amoyl)-seryl-L-valine], BE16627B (Banyu Tsukuba Research cancer drugs, and may thus suppress recurrence. Leukemia Institute, Ibaragi, Japan), a new inhibitor of MMPs, from (2001) 15, 1217–1224. microbial cultures. This MMP inhibitor has been shown to Keywords: matrix metalloproteinase inhibitor; SI-27; apoptosis; TNF-␣ effectively suppress growth of human fibrosarcoma HT1080 cells in an in vivo nude mouse model. From BE16627B, we synthesized a new MMP inhibitor, SI-27, that demonstrates significantly greater MMP-inhibiting activity. This MMP Introduction inhibitor suppresses the growth of human fibrosarcoma HT1080 cells in vivo more efficiently than BE16627B Matrix metalloproteinases (MMPs)belong to a family of zinc- (unpublished observation). The structure of SI-27 is shown in containing extracellular endopeptidases that selectively Figure 1. degrade components of the extracellular matrix, such as prote- In the present study, we investigated the effects of the new oglycans, laminin, and fibronectin. Currently, 19 members MMP inhibitor, SI-27. SI-27 alone activates some caspases have been identified in the human MMP family. Under normal and induces apoptosis in several human myeloid leukemia conditions, the activities of MMPs seem to be balanced by cell lines. Furthermore, at lower concentrations that do not tissue inhibitors of metalloproteinases (TIMPs). It has been directly induce apoptosis, SI-27 acts to sensitize U937 cells shown that MMPs, as well as TIMPs, are key players in many and other cells to tumor necrosis factor ␣ (TNF-␣)-mediated physiological and pathological processes such as develop- apoptosis. It is most likely that this sensitization is ment, angiogenesis, connective tissue remodeling, wound mediated by blocking NF-␬B activation through the cleavage healing, immunity, and inflammation.1–3 In cancer patients, of I␬B␣. MMPs also play an important role in disease progression and their activities are reported to associate with tumor pro- motion,4 angiogenesis,5 invasion, and metastasis.6 Much of the work on the role of MMP inhibitors has concentrated on their inhibitory effects in tumor invasion models, both in vivo and in vitro.7,8 Synthetic MMP inhibitors, such as BB-2516 (marimastat), have been designed for clinical application to inhibit metastasis.9 Since MMP inhibitors alone do not induce cell death in most cancer cells in vitro, they are often used in the stabilization of tumor growth in cancer patients, thus improving survival.6 Dysregulated production of MMPs and TIMPs has been reported in hematopoietic malignancies.10

Correspondence: K Motoyoshi, Third Department of Internal Medi- cine, National Defense Medical College, 3–2 Namiki, Tokorozawa, Saitama 359–8513, Japan; Fax: +81–42–996–5202 Figure 1 The molecular structure of SI-27. Received 12 January 2001; accepted 18 April 2001 New MMP inhibitor SI-27 induces apoptosis Y Nakamura et al 1218 Materials and methods Western blots

Enzymes and MMP assays 3 × 105/ml of U937 cells were treated with 200 ␮M SI-27. Cells were then harvested and lysed in an SDS sample buffer (0.04 M Tris-HCl (pH 6.8), 2% SDS, 5% 2-mercaptoethanol, MMP activity was measured as previously reported.12 One 5% glycerol, and 0.5% BPB)and boiled for 5 min. An aliquot unit of MMP activity was defined as that causing 1 ␮gof of the lysate (2 × 105 cells)was then analyzed by Western degraded substrate to become soluble in trichloroacetic acid blotting. Proteins were resolved by sodium dodecyl sulfate- in 1 min at 37°C. All enzymatic assays were carried out in 50 polyacrylamide gel electrophoresis (SDS-PAGE)and electrob- mM Tris-HCl buffer (pH 7.5)containing 10 m M CaCl and 2 lotted on to polyvinylidene difluoride membranes. Mem- 0.05% Brij35. The substrates used were [3H]gelatin (1.0 × 105 branes were blocked for 1 h with 5% non-fat dry milk powder d.p.m./1 ␮g/assay)for MMP-2 and MMP-9 and [ 3H]casein (2.5 in Tris-buffered saline (TBS: 10 mmol/l, Tris-HCl pH 7.4, 100 × 105 d.p.m./20 ␮g/assay)for MMP-3. [ 3H]-labeled substrates mmol/l NaCl)with 0.05% Tween20 (TBST)and then immuno- were prepared by acetylation with [3H] acetic anhydride blotted for 1 h with several antibodies. Antibodies were pur- (Amersham Japan, Tokyo, Japan)according to the method of chased from the following suppliers: against caspase 3 and Cawston and Barrett.13 caspase 2 (Transduction Laboratories, Lexington, KY, USA), against caspase 8 were (Immnotech, Marseille, France), against caspase 9 (Millennium Biotechnology, Ramona, CA, ␬ ␣ Cell lines and culture USA), against Bcl-2, Bcl-XL, Bax, and I B (Santa Cruz Biotechnology, Santa Cruz, CA, USA), against PARP (Pharmingen, San Diego, CA, USA), and against Actin The human cell lines U937, HL-60, NB-4, and THP-1 (Boehringer Mannheim Biochemica). Membranes were (provided from Saitama Cancer Center Research Institute, washed three times with TBST and incubated with peroxidase- Japan)were maintained in culture in RPMI 1640 sup- conjugated affinity-purified rabbit anti-mouse or goat anti-rab- plemented with 10% fetal bovine serum (FBS; GIBCO-BRL, bit IgG for 1 h. After extensive washing, the reactions were Grand Island, NY, USA), 100 unit/ml penicillin G and 100 developed by enhanced chemiluminescent staining (DuPont ␮ g/ml streptomycin. SI-27 was dissolved in dimethylsulfoxide NEN, Boston, MA, USA). Anti-actin antibody staining con- × (DMSO)to produce a 20 m M stock solution. Cell lines (5 firmed the loading and transfer of equal amounts of protein. 104/ml)were treated with a vehicle (DMSO)or several con- centrations of SI-27 for 24 h. 100 ␮M and 200 ␮M SI-27 con- tained 1% DMSO while all others contained 0.1% DMSO. At the end of treatment, surviving cells were identified by a WST- Flow cytometric analysis of Fas, the Fas ligand, and 1 (Wako Pure Chemicals Industries, Osaka, Japan)assay, with TNF receptor 1 membrane expression the untreated controls acting as a 100% reference. The WST- 1 assay was performed to measure the growth of cells with U937 cells were treated with 20 or 200 ␮M SI-27 for 24 h. SI-27 according to the manufacturer’s instructions. Fas, the Fas ligand, and TNF receptor 1 expression were meas- ured by flow cytometry by incubating cells for 30 min at 4°C with mouse anti-Fas (ZB4), hamster anti-Fas ligand (4A5) (MBL, Nagoya, Japan), and mouse anti-TNF receptor 1 (Dako, Determination of apoptotic cell death Glostrup, Denmark)antibodies. After two washes in PBS, cells were incubated for 30 min at 4°C with a fluorescein isothiocy- The leakage of fragmented DNA from apoptotic nuclei was anate-labeled rabbit anti-mouse IgG (Dako)or rabbit anti- measured by the method of Nicoletti et al.14 Briefly, cells were hamster IgG (Cappel, Aurora, OH, USA)and analyzed with prepared in a lysis buffer (phosphate-buffered saline (PBS), FACScalibur. 0.2% Triton X-100, 50 ␮g/ml propidium iodide (PI)) and then subjected to FACS analysis to quantify the population in the sub-G1 region. Phosphatidylserine (PS)externalization was Fas and the Fas ligand blocking assay measured by staining with FITC-conjugated annexin V (Trevigen, Gaithersburg, MD, USA)according to the manufac- For the blocking assay, U937 cells were incubated for 1 h turer’s protocol. A drop in mitochondrial transmembrane with 10 ␮g/ml antagonistic anti-Fas (ZB4)or anti-Fas ligand ⌬⌿ potential ( m)was measured by staining with CMX-Ros (4A5)antibodies before 200 ␮M SI-27 treatment. Then, 24 h (Molecular Probes, Eugene, OR, USA)as previously later, apoptosis was determined as described above. described.15 We used intact cells (not fixed cells)and ana- lyzed them immediately to measure ⌬⌿m via flow cytometry. All flow cytometry analyses were performed on a FACScalibur using CellQuest analysis software (Becton Dickinson, San Results Jose, CA, USA). The broad range caspase inhibitors zVAD-fmk and Z-Asp- Inhibition of MMPs by SI-27 CH2-DCB, as well as specific caspase inhibitors Ac-DEVD- CHO, Ac-IETD-CHO, Ac-YVAD-CHO, and Ac-LEHD-CHO Like BE16627B,12 SI-27 inhibits metalloproteinases, such as were purchased from the Peptide Institute (Osaka, Japan). Cas- MMP-2, MMP-3, MMP-9, thermolysin, bacterial collagenase,

pase inhibitors were added 30 min before SI-27 treatment. and leucine aminopeptidase. Its IC50 values for these enzymes After 24 h, apoptosis was analyzed by flow cytometry. Tumor are shown in Table 1. SI-27 did not inhibit any other protein- necrosis factor-␣ (TNF-␣)was purchased from Boehringer ase tested, including trypsin, chymotrypsin, papain, pepsin, Mannheim Biochemica (Mannheim, Germany). and elastase.

Leukemia New MMP inhibitor SI-27 induces apoptosis Y Nakamura et al 1219 Table 1 The effect of inhibition of MMPs by SI-27 27 treatment inhibits the growth of (at least some types of) leukemia cells but does not affect that of normal hemato- ␮ Enzyme IC50 ( M) poietic cells. Interestingly, SI-27-treated U937 cells displayed morpho- MMP2 0.036 logical signs of apoptosis including cell shrinkage, membrane MMP9 0.090 blebbing, and eventual disintegration into numerous apoptotic MMP3 0.17 Bacterial collagenase 0.90 bodies within 12 h of treatment (data not shown). These obser- Thermolysin 0.27 vations suggest SI-27 can induce apoptosis in some leukemia Aminopeptidase M 15.0 cell lines. To test this hypothesis, we analyzed apoptosis in a Trypsin Ͼ200 number of ways, including ⌬⌿m loss by CMX-Ros staining, Chymotrypsin Ͼ200 phosphatidylserine (PS)externalization by annexin V binding, Ͼ Papain 200 and the presence of hypodiploid DNA in U937 cells before Pepsin Ͼ200 Elastase Ͼ200 and after SI-27 treatment. U937 cells, after 24 h treatment with 200 ␮M SI-27, showed low ⌬⌿m and DNA fragmen- One unit of MMP activity was defined as that causing 1 ␮gof tation, and were annexin V-positive, which typically indicates degraded substrate to become soluble in trichloroacetic acid in 1 apoptosis, much more than those receiving DMSO treatment min at 37°C. alone (Figure 3).

SI-27 induces apoptosis in several myeloid leukemia cell lines SI-27 induces some caspases’ activation, but does not affect the protein levels of Bcl-2, Bcl-XL, Bad and BAX We examined the effect of SI-27 on the survival of several hematopoietic cell lines. A WST-1 assay revealed that SI-27 To elucidate the molecular basis for SI-27-induced apoptosis, had a dose-dependent suppressive effect on the growth of we examined the activation of caspases. Appropriate cleavage some of leukemia cell lines such as U937, NB-4, and HL-60, of procaspase is necessary for the activation of caspase cells, but not THP-1 cells (Figure 2). Normal peripheral blood- enzymes. Figure 4a indicates that SI-27 treatment induced the derived mononuclear cells (PBMCs)were used as a normal cleavage of initiator procaspase 8 and 9, and effector procas- control for the WST-1 assay in Figure 2. After 24 h of culture pase 3, but not procaspase 2. To determine whether cleaved in RPMI1640 containing 10% FBS with or without 1% DMSO, procaspase becomes active, we analyzed the cleavage of vari- 70% or 80% of PBMCs survived, respectively. The survival ous caspase substrates. Figure 4b indicates that SI-27 treat- rate of PBMCs treated with 200 ␮M SI-27 was 70%, and com- ment induced the cleavage of PARP and I␬B␣ both substrates parable with that of 1% DMSO treatment (data not shown). for caspase 3. In contrast with caspases, the levels of Bcl-2

In addition to the data in Figure 2, these data suggest that SI- family proteins, anti-apoptotic Bcl-2, and Bcl-XL, and pro-

Figure 2 SI-27 inhibits the growth of some, but not all, myeloid leukemia cell lines. Cell lines (5 × 104/ml)were treated with a vehicle (DMSO)or SI-27 at various concentrations for 24 h. 100 ␮M and 200 ␮M SI-27 contained 1% DMSO and all other concentrations contained 0.1% DMSO. At the end of treatment, surviving cells were identified by a WST-1 assay. The untreated controls were used as a 100% reference. All standard errors were less than 1% of the reference.

Leukemia New MMP inhibitor SI-27 induces apoptosis Y Nakamura et al 1220

Figure 3 SI-27 induces apoptosis in U937 cells. After treatment with 1% DMSO or 200 ␮M SI-27 for 24 h, cells were stained with PI for hypodiploid DNA, annexin V for phosphatidylserine (PS)externalization and CMX-Ros for ⌬⌿m, then subjected to flow cytometry. Untreated cells were used as a control. Experiments were repeated six times with similar results. Mean values and standard errors of mean (means ± s.e.m.) are shown.

apoptotic BAD and BAX did not change during SI-27 TNF-␣ activates not only death signals, but also survival sig- treatment. nals that are mediated by the activation of NF-␬B transcription factors, leading to limitations in TNF-␣ cancer therapy. How- ever, blocking NF-␬B activation has been reported to induce Pancaspase inhibitors block SI-27-induced apoptosis, apoptosis by TNF-␣.16 As shown in Figure 4b, 200 ␮M of SI- but not by specific caspase inhibitors 27 cleaves I␬B␣. It has been suggested that cleaved I␬B␣ may become a super-repressor form of I␬B␣. It may block the acti- To further demonstrate the activation of caspases during SI- vation of NF-␬B,17 thus improving the cancer fighting abilities 27-induced apoptosis, the effect of pancaspase inhibitors, as of TNF-␣. We therefore investigated whether TNF-␣-induced well as specific caspase inhibitors, on apoptosis was exam- apoptosis was augmented by SI-27. We found that 20 ␮M SI- ined. Several caspase inhibitors were added 30 min before SI- 27, which cannot cleave I␬B␣ nor activate caspases, nor 27 treatment, and then apoptosis was analyzed by flow cyto- directly induce apoptosis on U937 cells, markedly enhanced metry. The pancaspase inhibitors zVAD-fmk and Z-Asp-CH2- the cytotoxicity of TNF-␣ compared to DMSO-treated controls DCB efficiently blocked SI-27-induced apoptosis, but specific (Figure 6). caspase inhibitors for caspase 1 (Ac-YVAD-CHO), 8 (Ac-IETD- SI-27 sensitized HL-60, NB4, as well as U937 cells but not CHO), 3 (Ac-DEVD-CHO), and 9 (Ac-LEHD-CHO, data not THP-1 cells to TNF-␣ (data not shown). As expected, TNF-␣ ⌬⌿ shown), did not. Loss of m was only partially recovered, in the presence of 20 ␮M SI-27 induced cleavage of I␬B␣ in while hypodiploid DNA was completely recovered by the U937 cells but TNF-␣ with DMSO alone did not (data not addition of the pancaspase inhibitor zVAD-fmk or Z-Asp- shown). CH2-DCB (Figure 5). The cleavage of PARP and I␬B␣ by SI- 27 were also blocked by the pancaspase inhibitors, zVAD- fmk and Z-Asp-CH2-DCB, but were not blocked by specific SI-27 does not affect the expression of membrane Fas, caspase inhibitors (data not shown). the Fas ligand or TNF receptor 1 on U937 cells

Since MMP inhibitors block the shedding of several receptors SI-27 sensitizes U937 cells to apoptosis induced by and ligands, including the TNF-␣ receptor (TNFR), one of the TNF-␣ mechanisms by which SI-27 causes increased sensitization to the cytotoxic action of TNF-␣ may be via the up-regulation of The effectiveness of cancer therapy is determined by its ability TNFR.18 To clarify this, we measured the expression of TNFR to trigger apoptosis in malignant cells and its toxicity profile. on the surface of U937 cells, but found that 20 ␮M or 200

Leukemia New MMP inhibitor SI-27 induces apoptosis Y Nakamura et al 1221

Figure 4 SI-27 induces some caspases activation, but does not affect the expression of Bcl-2 family proteins. (a)After treatment with a vehicle (DMSO)or SI-27 for 24 h, cells were harvested for Western blot analysis. Untreated cells were used as a control. Activation of caspases assessed by immunoblotting. Arrows indicate procaspases Figure 5 The effects of caspase inhibitors on SI-27-induced and arrowheads show cleaved caspases. (b)Same cell lysates of (a) apoptosis. After treatment with 200 ␮M of SI-27 alone or with 200 ␮M were used for Western blot analysis of Bcl-2 family proteins and cas- of SI-27 and 100 ␮M of various caspase inhibitors for 24 h, apoptosis pase substrates. Arrows indicate wild-type and arrowheads show was analyzed by flow cytometry. Untreated cells were used as a con- cleaved substrates. Actin was used as a control for equal protein trol. The apoptosis of untreated cells was almost the same as that in loading. Figure 3. Experiments were repeated six times with similar results. Mean values and standard errors of mean (means ± s.e.m.)are shown.

␮M SI-27 did not up-regulate TNFR1 (Figure 7)or R2 (data not MMP inhibitor, SI-27, showed a direct apoptosis-inducing shown). Another possible mechanism of sensitization to TNF- effect on several leukemia cell lines. SI-27 exposure induced ␣-induced apoptosis by SI-27 may be via the blocking a variety of changes in pro-apoptotic proteins including casp- NF-␬B activation. ases. Apoptosis was documented by annexin V-positive stain- Some MMP inhibitors are believed to block the shedding ing, a drop in ⌬⌿m, the presence of hypodiploid DNA, and of the membrane Fas ligand and Fas19 and this receptor up- cleavage of PARP and I␬B␣. These effects do not appear to regulation is sensitized to the ligand signal. Therefore, we ana- be the result of non-specific toxicity, as SI-27 does not induce lyzed the expression of membrane Fas and the Fas ligand on apoptosis in THP-1 cells or other hematopoietic cell lines, U937 cells. Neither 20 ␮M nor 200 ␮M of SI-27 caused the such as Jurkat, and KG-1 (data not shown). accumulation of Fas or the Fas ligand (Figure 7). Furthermore, Several recent studies have shown that some metalloprot- 10 ␮g/ml of antagonistic Fas and the Fas ligand antibodies did einases mediate TNF-␣ and Fas ligand processing, and that not block SI-27-induced apoptosis (Figure 8). metalloproteinase inhibitors could block the shedding of the Fas ligand.19 As U937 cells express the membrane-type Fas ligand and Fas receptor, and because TNF-␣ up-regulates the Discussion membrane-type Fas ligand (unpublished observation), we investigated whether SI-27 treatment increased the expression Synthetic MMP inhibitors, such as BB-2516 (marimastat)have of membrane-type Fas ligand or Fas receptor. However, SI- been designed for use in clinical applications to inhibit met- 27-treated U937 cells showed neither membrane Fas ligand astasis in cancer patients. Also, the low molecular weight accumulation nor protection from apoptosis with anti-Fas and broad-spectrum MMP inhibitor BE16627B suppresses primary anti-Fas ligand neutralizing antibodies. These data suggest that subcutaneous tumor growth, as well as lung colonization of SI-27-induced apoptosis is independent of the Fas pathway human fibrosarcoma cells, in nude mice with no apparent activated by the increased expression of Fas receptor and Fas cytotoxicity.12 Another hydroxamic acid-based MMP inhibi- ligand on U937 cells. Recently, Mitsiades et al21 reported that tor, KB-R7785, also prevented angiogenesis and metastasis of two MMP inhibitors, BB-3103 and A151011, induced C26 murine adenocarcinoma in an in vivo mouse model.20 apoptosis in a Fas-sensitive cell line, which could then be In the present study, we demonstrated that a new synthetic blocked by Fas-neutralizing antibodies. However, these

Leukemia New MMP inhibitor SI-27 induces apoptosis Y Nakamura et al 1222

Figure 6 The synergistic effects of SI-27 with TNF-␣ on apoptosis induction. After treatment with TNF-␣ and 20 ␮M SI-27, or with TNF-␣ and 0.1% DMSO for 24 h, apoptosis was analyzed by flow cytometry. The apoptosis of untreated cells was indistinguishable with that of cells treated with TNF-␣ alone. Experiments were repeated six times with similar results. Mean values and standard errors of mean (means ± s.e.m.) are shown.

Figure 7 SI-27 does not induce accumulation of the membrane Fas receptor, Fas ligand, or TNF receptor 1. U937 cells were treated with 20 or 200 ␮M of SI-27 for 24 h, and Fas, the Fas ligand, and TNF receptor 1 expression were measured by flow cytometry. Cells were incubated for 30 min at 4°C with a mouse anti-Fas (ZB4), hamster anti-Fas ligand (4A5) (MBL), or mouse anti-TNF receptor 1 (Dako) antibodies, respectively. After two washes in PBS, cells were incubated for 30 min at 4°C with a fluorescein isothiocyanate-labeled rabbit anti-mouse IgG (Dako)or rabbit anti-hamster IgG (Cappel)and analyzed on FACScalibur. Untreated U937 cells were used as a control. Expressions of the membrane Fas, Fas ligand, or TNF receptor 1 on untreated cells were indistinguishable to that on U937 cells treated with 1% DMSO.

results could vary with different cell lines and MMP inhibitors cell permeable than ubenimix and the former induces used. Similarly, Fas and the Fas ligand pathway have been apoptosis more efficiently than the latter.25 Therefore, like reported to mediate anti-cancer drug-induced apoptosis, but ubenimix, accessible targets of SI-27 may be intracellular met- this type of apoptosis was dependent on the cell type exam- alloproteinases or other types of proteinases, and SI-27 may ined.22–24 Therefore, it may be that MMP inhibitors behave induce apoptosis by blocking them. The precise mechanism in a comparable manner, suggesting that results may be cell of its action currently remains unknown and further line-dependent. investigation is needed. BE16627B, a hydroxamic acid-based MMP inhibitor, does There are two conflicting pathways downstream of TNFR; not induce apoptosis and does not sensitize to TNF-␣ in U937 one leads to apoptosis mediated by caspases and another acti- cells (data not shown). On the basis of its structure, SI-27, but vates the survival signaling pathway mediated by NF-␬B, not BE16627B, is believed to be cell permeable. For instance, which is anti-apoptotic.16 Inhibition of NF-␬B activation by a the aminopeptidase inhibitor ubenimix-methyl-ester is more non-inducible form of I␬B␣ renders cells more susceptible to

Leukemia New MMP inhibitor SI-27 induces apoptosis Y Nakamura et al 1223 References 1 Massova I, Kotra LP, Fridman R, Mobashery S. Matrix metalloprot- einases: structures, evolution, and diversification. FASEB J 1998; 12: 1075–1095. 2 Werb Z. ECM and cell surface proteolysis: regulating cellular ecol- ogy. Cell 1997; 91: 439–442. 3 Goetzl EJ, Banda MJ, Leppert D. Matrix metalloproteinases in immunity. J Immunol 1996; 156: 1–4. 4 Sternlicht MD, Lochter A, Sympson CJ, Huey B, Rougier JP, Gray JW, Pinkel D, Bissell MJ, Werb Z. The stromal proteinase MMP3/stromelysin-1 promotes mammary carcinogenesis. Cell 1999; 98: 137–146. 5 Stetler-Stevenson WG. Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. J Clin Invest 1999; 103: 1237–1241. 6 Chambers AF, Matrisian LM. Changing views of the role of matrix metalloproteinases in metastasis. J Natl Cancer Inst 1997; 89: 1260–1270. 7 Denis LJ, Verweij J. Matrix metalloproteinase inhibitors: present achievements and future prospects. Invest New Drugs 1997; 15: 175–185. 8 Wojtowicz-Praga SM, Dickson RB, Hawkins MJ. Matrix metallop- roteinase inhibitors. Invest New Drugs 1997; 15: 61–75. 9 Wojtowicz-Praga S, Torri J, Johnson M, Steen V, Marshall J, Ness E, Dickson R, Sale M, Rasmussen HS, Chiodo TA, Hawkins MJ. Phase I trial of Marimastat, a novel matrix metalloproteinase inhibitor, administered orally to patients with advanced lung can- cer. J Clin Oncol 1998; 16: 2150–2156. 10 Janowska-Wieczorek A, Marquez LA, Matsuzaki A, Hashmi HR, Larratt LM, Boshkov LM, Turner AR, Zhang MC, Edwards DR, Kos- sakowska AE. Expression of matrix metalloproteinases (MMP-2 and -9)and tissue inhibitors of metalloproteinases (TIMP-1 and - Figure 8 Antagonistic Fas and the Fas ligand antibodies do not 2)in acute myelogenous leukaemia blasts: comparison with nor- block SI-27-induced apoptosis. U937 cells were incubated for 1 h mal bone marrow cells. Br J Haematol 1999; 105: 402–411. ␮ with 10 g/ml antagonistic Fas (ZB4)or Fas ligand (4A5)antibody 11 Wang CY, Cusack JC Jr, Liu R, Baldwin AS Jr. Control of inducible ␮ before 200 M SI-27 treatment. Twenty-four hours later, apoptosis was chemoresistance: enhanced anti-tumor therapy through increased determined as shown in Figure 5. apoptosis by inhibition of NF-kappaB. Nat Med 1999; 5: 412–417. 12 Naito K, Kanbayashi N, Nakajima S, Murai T, Arakawa K, Nishi- TNF-␣-induced apoptosis. Caspase 3 is known to cleave I␬B␣ mura S, Okuyama A. Inhibition of growth of human tumor cells in vitro.17 Once cleaved, I␬B␣ loses the C-terminal region in nude mice by a metalloproteinase inhibitor. Int J Cancer 1994; responsible for degradation, and becomes a super-repressor 58: 730–735. form of I␬B␣ thus blocking NF-␬B activation. If I␬B␣ cleaved 13 Cawston TE, Barrett AJ. A rapid and reproducible assay for col- lagenase using [1–14C]acetylated collagen. Anal Biochem 1979; by SI-27 is converted to the super-repressor form, SI-27- 99: 340–345. cleaved I␬B␣ also blocks NF-␬B activation, and sensitizes 14 Nicoletti I, Migliorati G, Pagliacci MC, Grignani F, Riccardi C. A cells to apoptosis by TNF-␣. rapid and simple method for measuring thymocyte apoptosis by What determines the sensitivity of the cells to SI-27 treat- propidium iodide staining and flow cytometry. J Immunol Meth ment remains unknown. If I␬B␣ cleaved by SI-27 blocks the 1991; 139: 271–279. activation of the NF-␬B pathway, the cell lines in which sur- 15 Macho A, Decaudin D, Castedo M, Hirsch T, Susin SA, Zamzami ␬ N, Kroemer G. Chloromethyl-X-Rosamine is an aldehyde-fixable vival mainly depends on the NF- B pathway are sensitive to potential-sensitive fluorochrome for the detection of early SI-27. Chemotherapy has been highly effective in inducing apoptosis. Cytometry 1996; 25: 333–340. complete remission in acute myelogenous leukemia.26 Unfor- 16 Beg AA, Baltimore D. An essential role for NF-kappaB in pre- tunately, despite intensive chemotherapy, the vast majority of venting TNF-alpha-induced cell death. Science 1996; 274: 782– patients continue to relapse and succumb to the disease. 784. Therefore, new drugs with alternative mechanisms of action, 17 Barkett M, Xue D, Horvitz HR, Gilmore TD. Phosphorylation of ␬ IkappaB-alpha inhibits its cleavage by caspase CPP32 in vitro. J such as NF- B blockers, need to be developed for use, alone Biol Chem 1997; 272: 29419–29422. or in combination with chemotherapy, to treat this disease. 18 Williams LM, Gibbons DL, Gearing A, Maini RN, Feldmann M, Our findings suggest the possibility that MMP inhibitors can Brennan FM. Paradoxical effects of a synthetic metalloproteinase provide a unique method of treatment for hematopoietic inhibitor that blocks both p55 and p75 TNF receptor shedding and malignancies refractory to classical anti-cancer drugs, and TNF alpha processing in RA synovial membrane cell cultures. J further that MMP inhibitors alone or in combination with Clin Invest 1996; 97: 2833–2841. 19 Tanaka M, Suda T, Haze K, Nakamura N, Sato K, Kimura F, other cytotoxic agents may suppress recurrence of disease. Motoyoshi K, Mizuki M, Tagawa S, Ohga S, Hatake K, Drummond The search for anti-cancer drugs with synergistic effects with AH, Nagata S. The Fas ligand in human serum. Nat Med 1996; SI-27 may prove to be highly beneficial in the fight against 2: 317–322. hematopoietic malignancies. 20 Lozonschi L, Sunamura M, Kobari M, Egawa S, Ding L, Matsuno S. Controlling tumor angiogenesis and metastasis of C26 murine colon adenocarcinoma by a new matrix metalloproteinase inhibi- tor, KB-R7785, in two tumor models. Cancer Res 1999; 59: Acknowledgements 1252–1258. 21 Mitsiades N, Poulaki V, Leone A, Tsokos M. Fas-mediated This work was supported by a grant from the Research Foun- apoptosis in Ewing’s sarcoma cell lines by metalloproteinase dation for Community Medicine. inhibitors. J Natl Cancer Inst 1999; 91: 1678–1684.

Leukemia New MMP inhibitor SI-27 induces apoptosis Y Nakamura et al 1224 22 Friesen C, Herr I, Krammer PH, Debatin KM. Involvement of the not dependent on Fas/the Fas ligand interactions. Blood 1997; 90: CD95 (APO-1/FAS)receptor/ligand system in drug-induced 935–943. apoptosis in leukemia cells. Nat Med 1996; 2: 574–577. 25 Sekine K, Fujii H, Abe F. Induction of apoptosis by bestatin 23 Wesselborg S, Engels IH, Rossmann E, Los M, Schulze-Osthoff K. (ubenimex)in human leukemic cell lines. Leukemia 1999; 13: Anticancer drugs induce caspase-8/FLICE activation and apoptosis 729–734. in the absence of CD95 receptor/ligand interaction. Blood 1999; 26 Mayer RJ, Davis RB, Schiffer CA, Berg DT, Powell BL, Schulman 93: 3053–3063. P, Omura GA, Moore JO, McIntyre OR, Frei E, 3rd. Intensive post- 24 Eischen CM, Kottke TJ, Martins LM, Basi GS, Tung JS, Earnshaw remission chemotherapy in adults with acute myeloid leukemia. WC, Leibson PJ, Kaufmann SH. Comparison of apoptosis in wild- Cancer and Leukemia Group B. N Engl J Med 1994; 331: 896– type and Fas-resistant cells: chemotherapy-induced apoptosis is 903.

Leukemia