Dear Sir,

The role of matrix metalloproteases (MMP) in cancer has been widely discussed and recently revised. The idea that these proteolytic enzymes work just as molecular shears to break down the basement membrane tissue boundaries, allowing infiltration of metastatic cancer cells, has now been given up. The general consensus is that MMP are involved in the molecular and structural changes that occur in remodeling tissues, such as those located along the advancing edge of tumors.1, 2, 3, 4 Further evidence, however, has suggested that the tissue inhibitors of MMP (TIMP) need to be investigated to ensure a correct evaluation of the proteolytic balance. Altered levels of MMP or TIMP have been reported in a number of different physiological and pathological situations including wound healing, breast remodeling, tumor growth and spread.5, 6, 7 The fact that MMP or TIMP play an important role in cancer growth and spread is no longer in doubt, although the definition of this role and its clinical significance remains under investigation. Referring to the work by Vihinen et al.,8 Jung et al.9 report that MMP and TIMP may be useful as diagnostic and prognostic factors in a number of diseases. In their study, however, MMP are reported to be important as prognostic markers and potential therapeutic targets rather than diagnostic factors.9 This is also supported by the use of MMP inhibitors in clinical trials to block cancer spread. Unfortunately, the preliminary results of these trials are disappointing, because of both side-effects and lack of therapeutic effectiveness, reviewed by Matrisian et al.10 MMP and TIMP as prognostic markers seem so far to be the most reliable for use by clinicians. In this regard several articles have reported the serum levels of MMP or TIMP in different malignancies, and the common conclusion is that higher levels of MMP or lower levels of TIMP correlate with a worse prognosis and clinical outcome.11, 12, 13 Jung et al.,9 however, argue that blood collection has not been considered carefully, warning clinicians about the reliability of the data. We recognize that they raise important points that deserve to be addressed in side-by-side comparisons. We would happy to perform such comparisons in future studies in a collaborative manner if appropriate protocols for blood drawing are drawn up together, however, the authors should bear in mind that platelet activation is an extremely difficult problem to avoid or carefully control for: e.g., prolonged tourniquet application may cause it many studies have been reported on this issue as recently reviewed by Ruggeri.14 To support their conclusions, the authors show that MMP-2 and MMP-9 measurements depend on the time of handling the sample and the type of sample preparation. Moreover, Jung et al. have used their own enzyme immunoassay, but no details or references are reported regarding the method and the proper quality controls they used. It is also not indicated whether their assay measures the inactive or the active form of MMP, or whether the complexed form with TIMP is also detectable. This is a crucial point because the reliability of the results seems to be closely correlated to the accuracy and the reproducibility of the technique used. In this regard, we have carefully analyzed the measurement of both MMP in breast cancer tissue samples, by gelatin zymography and commercially available ELISA kits. Although the detection of the MMP-2 and MMP-9 inactive forms is quite accurate with the ELISA kit, of the active forms it is not. We agree with Jung et al. that the timing of sample analysis and storage conditions can affect the results.15 On the basis of relatively few data, the authors conclude that serum samples should be excluded from MMP measurement. Therefore, Jung et al. claim that the data we have reported previously “are unreliable surrogate markers of cancer.” We see that Jung et al. have misinterpreted our study, as in fact we reported that the changes of MMP-9 and TIMP-1 levels, as measured before and after surgical eradication of the breast cancer, could be used as a marker of disease activity and not a “marker of cancer” as they report.16

We stand by the conclusions of our study, which are consistent with a wide body of literature on this subject (more than 300 articles listed in PubMed). Additional studies are obviously necessary to clarify the clinical usefulness of TIMP and MMP measurement in cancer patient management, and we hope our article will stimulate such studies.

Yours sincerely,


  1. Top of page

Gianluigi Giannelli, Salvatore Antonaci.