MMPs as therapeutic targets
In the 1980s and early 1990s, huge resources were put into the
development of MMP inhibitors (MMPIs) for treatment of
cancer patients. At this time, MMPs were simply thought of as
matrix-degrading enzymes, and MMPIs were believed to be
able to greatly reduce the invasive and metastatic potential of
cancer cells. In clinical trials, however, the drugs generally
offered no survival benefit to patients or sometimes even
decreased survival compared to patients in control groups. In
addition, severe side effects were reported. Later, several
explanations to the failure of these first MMPIs have been
proposed, which have been covered in several excellent re-
views [115–117] and will therefore only be briefly mentioned
here. First of all, the patients enrolled in the clinical trials had
advanced disease. Today, we know that MMPs are involved in
early phases of tumorigenesis, and the drugs might have had a
better effect if they were applied in patients with non-
metastatic cancer. Preclinical mouse experiments were gener-
ally very successful, and in these experiments, the MMPIs
were usually given before metastases had developed. Also, as
these first generations of MMPIs were broad spectrum inhib-
itors, they could inhibit both tumor-promoting and tumor-
repressing MMP activities. Furthermore, the presence of
MMPs in the patients enrolled in the clinical trial was not
verified, which means that the drug targets might actually not
have been expressed. In most of the clinical studies, MMPIs
were the only therapy given. Today, when new cancer drugs
are tested, they are usually given as adjuvant therapy together
with conventional cytostatic drugs or radiation, not as an
alternative. This is also likely to contribute to the poor results
from the clinical trials of the first generations of MMPIs.
Despite the disappointing results from these first clinical
trials, researchers have not given up on the idea of using
MMP inhibitors in the treatment of cancer patients due to
the growing understanding of the dual role of these enzymes
in the disease progression [118]. To succeed, however, les-
sons must be learned from the previous trials, and the
increased knowledge of the complexity of MMP functions
must be implemented in drug and study design. Most im-
portantly, it must be verified which MMPs or MMP func-
tions are harmful and should be targeted and which offer
host-protective effects and should therefore be spared.
Treatment with MMPIs must therefore be adjusted to each
patient according to cancer stage and the expression level of
various MMP targets. This means that new MMPIs, which
are specific for either an MMP or degradation of certain
substrates, must be developed. This is a challenging task
because the active site is very similar between members of
the MMP family. The active site cleft does however contain
subsites or pockets that differ between various MMPs. The
substrate to be cleaved must fit into these pockets, and they
are therefore important in defining substrate specificity
[119]. By targeting such subsites, it is therefore possible to
design more specific inhibitors than by targeting the active
site Zn ion [120]. Targeting of exosites or non-catalytic sites
in MMPs is another possible approach of inhibiting detri-
mental MMP effects. Large protein substrates often require
coordinated interactions between the active site and non-
catalytic domains to be efficiently cleaved [121]. A classic
example of the requirement of an exosite is the cleavage of
triple helical collagen by collagenases, such as MMP-1. The
triple helical cord is too large to fit into the active site cleft
of collagenases. To be processed, the helix must therefore be
unwound and the hemopexin domain of the collagenases
acts as a un-helicase. The unwound α-chains are small
enough to fit into the active site cleft and can be hydrolyzed
[122–124]. By targeting exosites of defined substrates, it
may be possible to inhibit specific MMP functions instead
of the total activity. To be able to design such drugs, how-
ever, detailed knowledge about enzyme–substrate interac-
tions are required as well as thorough knowledge of the
regulation and function of specific MMPs in different
cancers.
Targeting of MMPs at the expressional level is another
possibility, where some promising in vitro and mouse ex-
periments have been reported [125–128]. This approach
also merits knowledge of which MMPs that should be
regarded as targets and non-targets in individual patients to
avoid side effects [117]. Another interesting approach is to
take advantage of the increased MMP expression in tumors
to make targeted cytostatic drugs. As some MMPs have
much higher expression levels in cancer cells than in normal
cells, cytostatic drugs can be produced as pro-drugs that are
processed by a cancer-associated MMP. This would cause