Identification of an ataxia telangiectasia-mutated protein mediated surveillance system to regulate Bcl-2 overexpression.

Bcl-2 can both promote and attenuate tumorigenesis. Although the former function is relatively well characterized, the mechanism of the latter remains elusive. We report here that enforced Bcl-2 expression in MCF7 cells stabilizes p53, induces phosphorylation of p53 serine 15 (p53pSer15) and inhibits MCF7 cell growth. Consistent with p53 Ser15 being a target of ataxia telangiectasia mutated protein(ATM)/ATR (ATM- and rad3-related) in the DNA damage response, Bcl-2 activates ATM by inducing ATM Ser1981 phosphorylation, which is accompanied with the phosphorylaton of two additional ATM substrates, Chk2 Thr68 and H2AX Ser139. Downregulation of ATM using a specific small interference RNA fragment (ATMRNAi) abolished Bcl-2-induced p53pSer15 and Bcl-2-mediated growth inhibition of MCF7 cells. Ectopic expression of a dominant-negative p53 mutant, p53175H, partially rescued this growth inhibition. Taken together, these observations demonstrate the contribution of ATM-p53 function to Bcl-2-mediated inhibition of MCF7 cell growth, indicating an ATM-mediated surveillance system for regulating Bcl-2 overexpression. Consistent with this concept, we found that MCF7 cells express Bcl-2 heterogeneously with 34.5% of cells being Bcl-2 negative. In general, Bcl-2-positive MCF7 cells proliferate slower than those of Bcl-2 negative. Thus, we provide evidence suggesting that activation of ATM suppresses Bcl-2-induced tumorigenesis, and that attenuation of ATM function may be an important event in breast cancer progression.

Aggressive childhood neuroblastomas do not express caspase-8: an important component of programmed cell death.

Neuroblastomas that overexpress N-Myc due to amplification of the MYCN oncogene are aggressive tumors that become very resistant to treatment by chemotherapy and irradiation. to identify tumor suppressor genes in this group of neuroblastomas we analyzed the expression and function of both apoptosis-related cell cycle regulatory genes in cell lines and patient tumor samples. We found that in a high percentage of neuroblastoma cell lines and patient samples with amplified MYCN, caspase-8 mRNA is not expressed. The caspase-8 gene, CASP8, was deleted or silenced by methylation in the neuroblastoma cell lines while methylation of its promoter region was the predominant mechanism for its inactivation in the patient tumor samples. Reintroduction of caspase-8 into the neuroblastoma cell lines resensitized these cells to drug-induced and survival factor dependent apoptosis. Subsequently others have also shown that caspase-8 is silenced by methylation in neuroblastoma and peripheral neural ectodermal tumors, and that the caspase-9 regulator Apaf-1 is silenced by methylation in melanoma cell lines and patient samples. We conclude that caspase-8 acts as a tumor suppressor gene in neuroblastomas, that its silencing provides a permissive environment for MYCN gene amplification once the tumors are treated with chemotherapeutic drugs/irradiation, and that expression of this gene in these tumor cells may be of clinical benefit. We also discuss the possible significance of the neural crest cell progenitor cell origin and the silencing of important apoptotic regulators via methylation in both neuroblastoma and melanoma tumors.

A Fas-associated death domain protein-dependent mechanism mediates the apoptotic action of non-steroidal anti-inflammatory drugs in the human leukemic Jurkat cell line.

Non-steroidal anti-inflammatory drugs (NSAIDs) are inhibitors of cyclooxygenase-1 and -2 and are useful for prevention and cure of cancers, especially colon and rectal cancers. The NSAIDs indomethacin and sulindac sulfide have been shown to induce apoptosis of colon epithelial cancer cells by a Bax-dependent mechanism that involves mitochondria-mediated activation of a caspase-9-dependent pathway. In this report, we demonstrate that indomethacin and sulindac sulfide induce apoptosis of human leukemic Jurkat cells by a mechanism that requires the Fas-associated Death Domain Protein-mediated activation of a caspase-8-dependent pathway. Therefore, NSAIDs induce apoptosis by different mechanisms depending on the cell type.

Mutant p53 cooperates with ETS and selectively up-regulates human MDR1 not MRP1.

The most frequently expressed drug resistance genes, MDR1 and MRP1, occur in human tumors with mutant p53. However, it was unknown if mutant p53 transcriptionally regulated both MDR1 and MRP1. We demonstrated that mutant p53 did not activate either the MRP1 promoter or the endogenous gene. In contrast, mutant p53 strongly up-regulated the MDR1 promoter and expression of the endogenous MDR1 gene. Notably, cells that expressed either a transcriptionally inactive mutant p53 or the empty vector showed no endogenous MDR1 up-regulation. Transcriptional activation of the MDR1 promoter by mutant p53 required an Ets binding site, and mutant p53 and Ets-1 synergistically activated MDR1 transcription. Biochemical analysis revealed that Ets-1 interacted exclusively with mutant p53s in vivo but not with wild-type p53. These findings are the first to demonstrate the induction of endogenous MDR1 by mutant p53 and provide insight into the mechanism.

Allelic imbalance on chromosome 2q and alterations of the caspase 8 gene in neuroblastoma.

We previously reported a high incidence of loss of heterozygosity (LOH) on chromosome 2q33 in neuroblastoma (NB), observed in various types of human cancers including lung cancer, head and neck cancer and follicular thyroid carcinoma. To better elucidate the role of chromosome 2q aberrations in NB, we examined common allelic imbalance (AI) regions on chromosome 2q in 82 NB patients using 10 polymorphic microsatellite markers. AI on 2q was detected in 26 (32%) of 82 NB cases. There was a distinct common AI region between the D2S115 and D2S307 markers on 2q33. The distance between these markers was about 2.0 cM. Recently, the caspase 8 and caspase 10 genes, both of which encode cystein protease, were mapped to chromosome 2q33. Since the common AI region on 2q33 includes the caspase 8 and caspase 10 genes, the alterations of these genes were examined further. Absent or reduced expression of caspase 8 and caspase 10 were found in 19 (70%) of 27 and two (7%) of 27 NB cell lines by reverse transcription-polymerase chain reaction, respectively. A missense mutation was detected at codon 96, GCT (Alanine) to GTT (Valine), of the caspase 8 gene in one of the NB cell lines lacking caspase 8 expression. Thirteen (68%) of 19 cell lines lacking caspase 8 expression displayed methylation of the CpG island of the caspase 8 gene, whereas only one (13%) of eight cell lines with caspase 8 expression showed caspase 8 methylation (P=0.031). Furthermore, there was a significant association between AI at 2q33 and loss of caspase 8 expression (P=0.026). These results indicated that there was a tumor suppressor gene in the common AI region on chromosome 2q33 involved in the pathogenesis of a subset of NB. It is possible that the caspase 8 gene is one of the candidate tumor suppressor genes for NB and inactivation of this gene plays an important role in the tumorigenesis of NB through mainly its methylation.

Dopaminergic cell death induced by MPP(+), oxidant and specific neurotoxicants shares the common molecular mechanism.

Recent etiological study in twins (Tanner et al. 1999) strongly suggests that environmental factors play an important role in typical, non-familial Parkinson's disease (PD), beginning after age 50. Epidemiological risk factor analyses of typical PD cases have identified several neurotoxicants, including MPP(+) (the active metabolite of MPTP), paraquat, dieldrin, manganese and salsolinol. Here, we tested the hypothesis that these neurotoxic agents might induce cell death in our nigral dopaminergic cell line, SN4741 (Son et al. 1999) through a common molecular mechanism. Our initial experiments revealed that treatment with both MPP(+) and the other PD-related neurotoxicants induced apoptotic cell death in SN4741 cells, following initial increases of H(2)O(2)-related ROS activity and subsequent activation of JNK1/2 MAP kinases. Moreover, we have demonstrated that during dopaminergic cell death cascades, MPP(+), the neurotoxicants and an oxidant, H(2)O(2) equally induce the ROS-dependent events. Remarkably, the oxidant treatment alone induced similar sequential molecular events: ROS increase, activation of JNK MAP kinases, activation of the PITSLRE kinase, p110, by both Caspase-1 and Caspase-3-like activities and apoptotic cell death. Pharmacological intervention using the combination of the antioxidant Trolox and a pan-caspase inhibitor Boc-(Asp)-fmk (BAF) exerted significant neuroprotection against ROS-induced dopaminergic cell death. Finally, the high throughput cDNA microarray screening using the current model identified downstream response genes, such as heme oxygenase-1, a constituent of Lewy bodies, that can be the useful biomarkers to monitor the pathological conditions of dopaminergic neurons under neurotoxic insult.

Absent or reduced expression of the caspase 8 gene occurs frequently in neuroblastoma, but not commonly in Ewing sarcoma or rhabdomyosarcoma.

PROCEDURE: To clarify whether the caspase 8 gene is involved in the pathogenesis of neuroblastoma (NB), we examined alterations of the caspase 8 gene in 15 NB, seven Ewing sarcoma (ES), and eight rhabdomyosarcoma (RMS) cell lines, using reverse transcription-polymerase chain reaction (RT-PCR) and RT-PCR single-strand conformation polymorphism (SSCP) analyses. RESULTS: The caspase 8 gene was not expressed in 11 (73%) of 15 NB cell lines, it was absent in only one of seven ES cell lines, but was present in all eight RMS cell lines examined. No mutations were detected in any cell lines examined. CONCLUSIONS: Inactivation of the caspase 8 gene is considered to be involved in the pathogenesis of NB, but not ES or RMS.

Proteolytic regulation of apoptosis.

Much of the proteolysis that occurs during apoptosis is directed by caspases, a family of related cysteinyl proteases. A relatively small number of cellular proteins are targeted by caspases, yet their function is dramatically affected and apoptosis is triggered. Other proteases, such as granzymes and calpain, are also involved in the apoptotic signaling process, but in a much more cell type- and/or stimulus type-specific manner. At least three distinct caspase-signaling pathways exist; one activated through ligand-dependent death receptor oligomerization, the second through mitochondrial disruption, and the third through stress-mediated events involving the endoplasmic reticulum. These pathways also appear to interact to amplify weak apoptotic signals and shorten cellular execution time. Finally, defects in caspases contribute to autoimmune disease, cancer and certain neurological disorders.

Caspase 8 is deleted or silenced preferentially in childhood neuroblastomas with amplification of MYCN.

Caspase 8 is a cysteine protease regulated in both a death-receptor-dependent and -independent manner during apoptosis. Here, we report that the gene for caspase 8 is frequently inactivated in neuroblastoma, a childhood tumor of the peripheral nervous system. The gene is silenced through DNA methylation as well as through gene deletion. Complete inactivation of CASP8 occurred almost exclusively in neuroblastomas with amplification of the oncogene MYCN. Caspase 8-null neuroblastoma cells were resistant to death receptor- and doxorubicin-mediated apoptosis, deficits that were corrected by programmed expression of the enzyme. Thus, caspase 8 acts as a tumor suppressor in neuroblastomas with amplification of MYCN.

Caspase-8 activation and bid cleavage contribute to MCF7 cellular execution in a caspase-3-dependent manner during staurosporine-mediated apoptosis.

There are at least two distinct classes of caspases, initiators (e.g. caspases-8, -9, and -10) and effectors (e.g. caspase-3). Furthermore, it is believed that there are two distinct primary apoptotic signaling pathways, one of which is mediated by death receptors controlled by caspases-8/10, and the other by the release of cytochrome c and activation of a caspase-9/Apaf1/cytochrome c apoptosome. However, several recent reports have demonstrated that caspase-8, and its substrate Bid, are frequently activated in response to certain apoptotic stimuli in a death receptor-independent manner. These results suggest that significant cross-talk may exist between these two distinct signaling arms, allowing each to take advantage of elements unique to the other. Here we provide evidence that activation of caspase-8, and subsequent Bid cleavage, does indeed participate in cytochrome c-mediated apoptosis, at least in certain circumstances and cell types. Furthermore, the participation of activated caspase-3 is essential for activation of caspase-8 and Bid processing to occur. Although caspase-8 activation is not required for the execution of a cytochrome c-mediated death signal, we found that it greatly shortens the execution time. Thus, caspase-8 involvement in cytochrome c-mediated cell death may help to amplify weaker death signals and ensure that apoptosis occurs within a certain time frame.

Cell cycle regulation and RNA polymerase II.

The cell cycle and transcription by RNA polymerase II (RNAP II) are closely related. They utilize shared components. RNAP II transcriptional activity is modulated during the cell cycle. Cell cycle dependent changes in the phosphorylation status of the carboxyl-terminal domain (CTD) of the largest subunit of RNAP II (RNAP II-LS) alter transcription. Several CTD kinases are members of the cyclin-dependent kinase (cdk) superfamily, including p34cdc2 (cdk1), cdk7, cdk8, and cdk9. Each of these cdks, with their respective cyclin partners, have been linked to cell cycle regulatory events. Other CTD kinases such as casein kinase II (CKII) and c-abl have also been implicated in cell cycle dependent modifications of the CTD. In addition, the stalling of RNAP II complexes at DNA lesions helps stimulate p53 accumulation which largely determines the cell's DNA damage response, including cell cycle arrest. Alzheimer's disease pathology results partially from activation of mitotic cdks in postmitotic neurons which can phosphorylate RNAP II-LS and other targets.

Fas-induced apoptosis in human malignant melanoma cell lines is associated with the activation of the p34(cdc2)-related PITSLRE protein kinases.

The Cdc2L locus encoding the PITSLRE protein kinases maps to chromosome band 1p36 and consists of two duplicated and tandemly linked genes. The purpose of the present study was to determine whether diminution of PITSLRE kinases leads to deregulation of apoptosis. The human melanoma cell lines A375 (Cdc2L wild-type alleles) and UACC 1227 (mutant Cdc2L alleles) were tested with agonist anti-Fas monoclonal antibody. We found that exposure of these cells to anti-Fas for 24, 48, or 72 h resulted in differential sensitivity to Fas-induced apoptosis. In A375, cell death started at 24-48 h post-treatment, and it was maximal by 72 h. Conversely, UACC 1227 cells were resistant to Fas-mediated apoptosis. Induction of PITSLRE histone H1 kinase activity was observed in A375 anti-Fas treated but not in UACC 1227 cells. Also, the PITSLRE protein kinase activity in A375 anti-Fas-treated cells preceded maximal levels of apoptosis. Finally, fluorescence confocal microscopy revealed a nuclear localization of PITSLRE proteins in normal melanocytes and A375 cells but a cytoplasmic localization in UACC 1227 cells. The differences in PITSLRE protein and cellular localization between A375 and UACC 1227 cells appear to account for the differences in sensitivity of the two cells lines to anti-Fas and staurosporine. These observations suggest that alterations in PITSLRE gene expression and protein localization may result in the loss of apoptotic signaling.

Database search strategies used to isolate cell death proteins.

The identification of proteins involved in the early phases of cell death has relied primarily on the modular organization of shared sequences and structural motifs of previously identified proteins in the apoptotic machinery. This property has facilitated the isolation of proteins that interact with each other through structural domains using yeast two-hybrid cloning. Likewise, the conservation in primary sequence of the various shared domains has promoted the use of polymerase chain reaction and database search strategies to isolate additional family members. Here, we discuss the use of database search strategies in the isolation of novel death proteins, as well as how similar strategies may be extended to discover additional, novel cell death proteins.

Cycloheximide-induced T-cell death is mediated by a Fas-associated death domain-dependent mechanism.

Cycloheximide (CHX) can contribute to apoptotic processes, either in conjunction with another agent (e.g. tumor necrosis factor-alpha) or on its own. However, the basis of this CHX-induced apoptosis has not been clearly established. In this study, the molecular mechanisms of CHX-induced cell death were examined in two different human T-cell lines. In T-cells undergoing CHX-induced apoptosis (Jurkat), but not in T-cells resistant to the effects of CHX (CEM C7), caspase-8 and caspase-3 were activated. However, the Fas ligand was not expressed in Jurkat cells either before or after treatment with CHX, suggesting that the activation of these caspases does not involve the Fas receptor. To determine whether CHX-induced apoptosis was mediated by a Fas-associated death domain (FADD)-dependent mechanism, a FADD-DN protein was expressed in cells prior to CHX treatment. Its expression effectively inhibited CHX-induced cell death, suggesting that CHX-mediated apoptosis primarily involves a FADD-dependent mechanism. Since CHX treatment did not result in the induction of Fas or FasL, and neutralizing anti-Fas and anti-tumor necrosis factor receptor-1 antibodies did not block CHX-mediated apoptosis, these results may also indicate that FADD functions in a receptor-independent manner. Surprisingly, death effector filaments containing FADD and caspase-8 were observed during CHX treatment of Jurkat, Jurkat-FADD-DN, and CEM C7 cells, suggesting that their formation may be necessary, but not sufficient, for cell death.

Structure and chromosome localization of the human CASP8 gene.

The human CASP8 gene, whose product is also known as caspase 8 and FLICE, encodes an interleukin-1beta converting enzyme (ICE)-related cysteine protease that is activated by the engagement of several different death receptors. Caspase 8 is immediately recruited to the Fas receptor once it oligomerizes, and its protease activity is crucial for the apoptotic response generated by the resulting death-inducing signaling complex (DISC). We report here that the CASP8 gene contains at least 11 exons spanning approximately 30kb on human chromosome band 2q33-34. This region of human chromosome 2 was previously reported as the location of the CASP10 gene, whose product is closely related to caspase 8. Chromosome 2 band q33-34 is also involved in tumorigenesis, with loss of heterogeneity (LOH) being reported in a number of tumors. We also report EcoRI and HindIII polymorphisms that may prove to be useful in disease analysis. Both caspases 8 and 10 contain long pro-domains with duplicated death effector domains (DEDs), as well as their corresponding cysteine protease catalytic domains. Thus, it appears that CASP8 and CASP10 have evolved by tandem gene duplication, much like the CASP1, CASP4 and CASP5 gene cluster on human chromosome 11q22.2-22.3.

Abnormalities in the p34cdc2-related PITSLRE protein kinase gene complex (CDC2L) on chromosome band 1p36 in melanoma.

The two genes encoding the PITSLRE protein kinase isoforms, CDC2L1 and CDC2L2, are localized to human chromosome band 1p36. The PITSLRE protein kinases are a part of the p34cdc2 supergene family. Several protein products of the CDC2L locus may be effector(s) in apoptotic signaling. The larger PITSLRE p110 isoforms appear to regulate some aspect of RNA splicing/transcription during the cell cycle. One or more of these genes may function as tumor suppressor genes in melanoma. Using fluorescence in situ hybridization, one allele of the CDC2L gene complex on chromosome 1 was either deleted or translocated in 8 of 14 different melanoma cell lines. We also observed mutations in the 5' promoter region of the CDC2L1 gene in four different cell lines relative to normal melanocytes using PCR-SSCP analysis and direct DNA sequencing. Western blot analysis revealed decreased level of PITSLRE protein expression in several cell lines, as well as in four surgical malignant melanoma specimens relative to normal melanocytes. Thus, the decreased PITSLRE protein expression appears to result from deletion of the CDC2L alleles and possibly by mutations within the 5' promoter region. We propose that aberrations in the CDC2L genes may contribute to the pathogenesis or progression of melanoma.

Fluorescence in situ hybridization analysis of chromosome 1p36 deletions in human MYCN amplified neuroblastoma.

BACKGROUND/PURPOSE: Deletion of the short arm of chromosome 1 (1p) is one of the poor prognostic factors in human neuroblastomas. Recent studies have suggested that one or more of the neuroblastoma tumor suppressor genes reside in this region and have identified the shortest region of overlap (SRO) on 1p36. The purpose of this study was to examine deletions of 1p in human neuroblastomas by fluorescence in situ hybridization (FISH). METHODS: Two-color FISH analysis was performed to detect chromosome 1p36 abnormalities in 42 MYCN-amplified neuroblastomas. Four different probes from the 1p36 region, the E2F2, NPPA, D1S160, and CDC2L1 loci were used for detection of 1p abnormalities. A repeat sequence probe, which is specific for the heterochromatic region of chromosome 1 (pUC1.77), was used as a control. RESULTS: Large deletions of 1p36 were observed in 31 (73.8%) of 42 tumors, whereas the remaining 11 (26.2%) showed no deletion. In these 11 tumors, a translocation of 1p was found in one and a duplication of 1p was detected in another. CONCLUSIONS: A strong correlation between 1p abnormalities and MYCN amplification was found in this study. MYCN-amplified neuroblastomas were found to show large deletions of 1p encompassing the SRO. FISH provided a rapid and reliable method to detect hemizygous deletions of 1p.

Cleavage of DFF-45/ICAD by multiple caspases is essential for its function during apoptosis.

Apoptosis involves the proteolysis of specific cellular proteins by a group of cysteine proteases known as caspases. Many of these cellular targets are either functionally inactivated (e.g. poly(ADP-ribose) polymerase) or activated (e.g. other caspases, gelsolin) by such processing, thereby facilitating the cell death process. Caspase 3 is involved in the processing of many of these proteins. Recently, however, it was reported that caspase 3 is dispensable for the cleavage of a large number of cellular caspase substrates during apoptosis. Among these substrates is DFF-45/ICAD, a subunit of the heterodimeric DNA fragmentation factor (DFF), otherwise known as caspase-activated DNase (CAD), that mediates genomic DNA degradation during apoptosis. Conversely, others have reported that caspase 3 is essential for the cleavage and activation of DFF-45/ICAD. To resolve this controversy we examined DFF-45/ICAD processing during apoptosis in MCF-7 breast carcinoma cells that lack functional caspase 3 and in MCF-7 cells expressing caspase 3. We found that DFF-45/ICAD is cleaved by two distinct caspases, one of which is caspase 3. Furthermore, cleavage of the carboxyl-terminal region of DFF-45/ICAD, which is necessary for activation of the enzyme, requires functional caspase 3. In the absence of caspase 3 cleavage of the amino-terminal region of DFF-45/ICAD by another caspase occurs, but the DFF-45 enzyme remains inactive.

Duplication of a genomic region containing the Cdc2L1-2 and MMP21-22 genes on human chromosome 1p36.3 and their linkage to D1Z2.

Cdc2L1 and Cdc2L2 span approximately 140 kb on human chromosome 1p36.3. The products of the Cdc2L genes encode almost identical protein kinases, the PITSLRE kinases, which have functions that may be relevant to the regulation of transcription/splicing and apoptotic signaling. These genes are deleted/translocated in neuroblastomas with MYCN gene amplification, a subset of malignant melanomas, and in a newly delineated deletion syndrome. Here we report that the p36.3 region of human chromosome 1 consists of two identical genomic regions, each of which contain a Cdc2L gene linked to a metalloprotease (MMP) gene in a tail-to-tail configuration. This duplicated genomic region is also linked tightly to D1Z2, a genetic marker containing a highly polymorphic VNTR (variable number tandem repeat) consisting of an unusual 40-bp reiterated sequence. Thus, these genes and the polymorphic marker D1Z2 are organized as follows: telomere-D1Z2-5'-MMP22-3'-3'-Cdc2L2-5'-5'-Cdc2L1 -3'- 3'-MMP21-5'-centromere. Remarkably, the introns and exons of Cdc2L1 and Cdc2L2, as well as their flanking regions, are essentially identical. A total of 15 amino acid differences, 12 nonconservative and 3 conservative, can be found in the 773-786 amino acids specified by the various products of the Cdc2L genes. Two separate promoter/5' untranslated (UT) regions, CpG1 and CpG2, are identical to a reported previously methylated genomic CpG sequence and are used to express >20 different Cdc2L transcripts from the two genes. The expression of CpG2 transcripts from Cdc2L1 and Cdc2L2 is tissue/cell-line specific. CpG1 transcripts are expressed ubiquitously from both genes, with perhaps some bias towards the expression of CpG1 Cdc2L1 mRNAs in certain hematopoietic cells.

Isolation and characterization of two novel metalloproteinase genes linked to the Cdc2L locus on human chromosome 1p36.3.

The terminal end of the short arm of human chromosome 1, 1p36.3, is frequently deleted in a number of tumors and is believed to be the location of multiple tumor suppressor genes. Thus far, a bona fide tumor suppressor gene from this region has not been identified. The isolation and characterization of new 1p36 genes is, therefore, of some interest. Two novel matrix metalloproteinase genes, MMP21 and MMP22, have been identified in the Cdc2L1-2 locus, which spans approximately 120 kb on 1p36.3. These genes encode novel metalloproteinases that contain prepro, catalytic, cysteine-rich, interleukin-1 receptor-related, and proline-rich domains. Their catalytic domains are most closely related to stromelysin-3 and contain the consensus HEXXH zinc-binding region required for enzyme activation, while their cysteine-rich domains appear to be related to a number of human, mouse, and Caenorhabditis elegans metalloproteinase sequences. Of some possible interest is the absence of a highly conserved cysteine residue in the proenzyme domain, the so-called "cysteine switch," which has been shown to be involved in the autocatalytic activation of many metalloproteinases. The MMP genes are located less than 1 kb from the 3' regions of Cdc2L1 and Cdc2L2, suggesting that the MMP and Cdc2L genes are part of a larger region that has been duplicated. Finally, the MMP21/22 genes express multiple mRNAs, some of which are derived by alternative splicing, in a tissue-specific manner.