2.3. Gelatinases in hematology

2.3.1. Hematopoietic stem cells

Resting hematopoietic stem cells (CD34+ cells) derived directly from bone marrow do not express metalloproteinases and show a low rate of migration through matrigel, an artificial basement membrane model. If the same stem cell population is derived from peripheral blood after stem cell mobilization, they express both MMP-2 and MMP-9 and their matrigel transmission is enhanced. The gelatinase expression of bone marrow stem cells may also be activated by various cytokines, including the granulocyte –colony-stimulating factor, the granulocyte-macrophage colony-stimulating factor, and the interleukins -3, -6, and -8. (Janowska-Wieczorek et al. 1999).

These results imply that metalloproteinase expression is mandatory for the mobilization of stem cells from bone marrow into circulation, which is also supported by the fact that MMP-9 accumulates in serum during the administration of colony-stimulating factor (Takamatsu et al. 1999). However, although the migration through matrigel can be blocked by MMP inhibitors (Janowska-Wieczorek et al. 1999), the stem cell mobilization induced by granulocyte colony-stimulating factor can not (Takamatsu et al. 1999).

2.3.2. Granulocytes

Granulocytes are a subset of leukocytes integral to many functions of the immune system, but specifically to protection against bacteria. Upon contact with inflammatory mediators, such as IL-8, the neutrophil changes from a passive state into a highly responsive cell capable of firm adhesion to the endothelium and mobilizes its secretory vesicles (Opdenakker et al. 2001). At the next step of the inflammatory response, the cell secretes MMP-9 from peroxidase-negative gelatinase granules, which leads to basement membrane disruption and transmigration of the neutrophil through the basement membrane into the tissue. At the last stage of the reaction, the mobilization of specific granules enhances the phagocytic potential of the neutrophil, and the fusion of the azurophilic and specific granules with the phagosome results in bactericidal activity (Kjeldsen et al. 1993, Borregaard & Cowland 1997).

This complex reaction is possibly due to the presence of highly differentiated granules containing different enzymes in the neutrophil cytoplasm. There are four types of granules: primary peroxidase-positive, defensive negative and positive, and peroxidase-positive secondary and tertiary granules.

The synthesis of proteins in these granules starts at the promyelocyte stage with the synthesis of peroxidase-positive granules. Peroxidase-negative granules, including gelatinase-containing tertiary granules, are synthesized at the myelocyte stage of development. (Graubert et al. 1993, Borregaard et al. 1995)

In tertiary granules, MMP-9 is stored in an inactive proenzyme form as monomers, dimers, a mono-dimer/lipocalin complex or a monomer/lipocalin/TIMP-1 complex (Kolkenbrock et al. 1996).

Apart from its function as an invasion enzyme, MMP-9 also has a role as a regulator of leukocyte biology. It truncates IL-8 at the amino terminus into a chemokine of tenfold potency, resulting in a positive feedback loop for neutrophil activation and chemotaxis. The CXC chemokines GRO-alpha, CTAP-III, and PF-4 are also degraded by MMP-9. (Opdenakker et al. 2001). Moreover, gelatinases take part in general immune activation by facilitating the secretion of TNF-α and releasing the stored growth factors from the matrix (Goetzl et al. 1996).

2.3.3. Monocyte macrophage lineage

Monocytes express small quantities of both MMP-2 and MMP-9. MMP-9 expression can be induced with TNF-α and interaction with collagen it can be down-regulated with transforming growth factor- β and IL-10 (Galt et al. 2001, Mostafa et al. 2001, Vaday et al. 2001). When monocytes are induced to differentiate into mature macrophages by, for example, TNF-α, the MMP-9 expression enhances. At the macrophage stage of differentiation, cells express continuously small amounts of MMP-2, MMP-9 expression can be induced to increase up to 15-fold over the baseline activity by, for example, TNF-α, IL-1β and interaction with human endothelial cells (Saren et al. 1996, Amorino & Hoover 1997, Xie et al. 1998).

This altered MMP expression probably enables monocyte and macrophage migration to tissues and the inflammatory focus, but may also have deleterious effects in the form of tissue degradation during inflammation. Dendritic cells differentiated from monocytes produce only MMP-9 (Bartholome et al. 2001). It may also be that the macrophages surrounding the malignant tissue serve as a MMP reservoir for the tumor cells not capable of MMP synthesis.

2.3.4. Lymphocytes

In the absence of an exogenous inflammatory stimulus, resting CD4+ T-lymphocytes (helper cells) express small amounts of MMP-9 and TIMP-1. The CD8+ (cytotoxic T-cells) and CD3+ populations exhibit these same activities, albeit in smaller quantities. In comparison, CD56+ T-cells (natural killer cells) express barely detectable levels of these enzymes.

Macrophage inflammatory proteins, which are also known to participate in chemotaxis, (β chemokines) up-regulate the secretion of proMMP-9 from all of these cells, and TNF-α and IL-1 up-regulate the secretion of CD3+ and CD4+ cells, while IFN-α down-regulate it in all T-cell sublines (Johnatty et al. 1997, Hausenberger et al. 1999). In fact, it seems that a given cytokine can transmit different effects in different situations, depending on the concentration and net balance of cytokines around the cell (Vaday et al. 2001). There are also reports about MMP-2 expression in T-lymphocytes (Leppert et al. 1995), but it may be that these findings are due to contamination of lymphocytes with other leukocyte subsets.

Augmentation of MMP secretion stimulates the migration of T-lymphocytes through a reconstituted basement membrane, which, on the other hand, can be inhibited with a MMP inhibitor. These results suggest that MMPs have an integral role in the invasion of T-lymphocytes into the inflammatory focus, and this phenomenon is regulated by the same cytokines and chemokines that also regulate the other parts of the inflammatory response. (Leppert et al. 1995)

Along with its role in inflammatory activation, MMP-9 is assumed also to be involved in cancer-related immunosuppression. MMP-9 is capable of cleaving IL-2R receptor from tumor-infiltrating T-lymphocytes. ( Sheu et al. 2001)

There are fewer reports on MMP-2 and MMP-9 expression in B-lymphocytes. Stetler-Stevenson et al. have shown faint MMP-9 activity by gelatinase zymography but not by Northern blot analysis in normal tonsillar cell suspension containing 62% B-cells and 19% T-cells, with the rest of the cells representing monocytic lineage (Stetler-Stevenson et al. 1997).

TIMP-1 has specific important functions during the development and maturation of normal B-cells. Naïve B-cells are produced in bone marrow and travel to lymph nodes to either undergo apoptosis or to further differentiate into highly proliferative centroblasts. These processes are stimulated by T-cells and germinal center follicular dendritic cells (Liu et al. 1991). It seems that the germinal center microenvironment induces maturing B-cells to express TIMP-1, which is also expressed by the surrounding B-cells inducing differentiation into the centrocytic stage (Li et al.1999, Guedez et al. 2001). TIMP-1 also induces IL-10 expression and up-regulates the expression of CD40, a pro-proliferative factor, and down-regulates that of CD77, a pro-apoptotic factor (Guedez et al. 1998). TIMP-1-induced IL-10 expression further stimulates cells to differentiate into mature plasma cells. While it is known that IL-10 acts as a potent autocrine growth factor and its expression is associated with poor prognosis in non-Hodgkin´s lymphomas, these findings may have importance in the biology of non-Hodgkin´s lymphomas as well (Guedez et al. 2001).

2.3.5. Platelets

The migration of polypoid mature megakaryocytes from the bone marrow microenvironment into the perivascular extracellular space is an integral phenomenon at the beginning of thrombopoiesis. For migration, megakaryocytes need MMP-9, and thrombopoiesis can actually be blocked by an MMP inhibitor (Lane et al. 2000).

While MMP-9 is integral for thrombopoiesis, MMP-2 seems to have a role in platelet aggregation. Two pathways are known to activate platelet aggregation namely the release of endoperoxide/thromboxane A2 and ADP. Sawicki et al. have recently demonstrated that MMP-2 is released during aggregation and that this release correlates with the magnitude of activation. Aggregation inhibitors decrease MMP-2 secretion, just as MMP-2 inhibitors decrease platelet aggregation. These results suggest that MMP-2 could also have a role in regulating platelet aggregation (Sawicki et al. 1997). It has been shown that the ability of tumor cells to induce platelet aggregation correlates with its MMP-2 expression and can be reversed with a synthetic MMP inhibitor or anti-MMP-2 antibodies (Juraz et al. 2001). On the other hand, thrombin stimulates MMP-9 expression in human mesangial cells (Liu et al. 2000).

2.3.6. Bone marrow stromal cells and dendritic cells

On resting stage no bone marrow stromal cells show nor MMP-2 neither MMP-9 expression when studied with immunohistochemical staining method. TIMP-2 positivity can be detected in stromal histiocytes only. (Ogawa et al. 2000) However during migration into a site of bone resorption osteoclasts produce MMP-9, which expression is integral to the migration and bone resorption phenomen (Spessotto et al. 2002).

Both mature and immature monocyte-derivied dendritic cells express both MMP-2, MMP-9 and TIMP-1 (Kouwenhoven et al. 2002). MMP-9 expression can be inhibited by IFN-β (Bartholome et a.l 2001).

2.3.7. Gelatinases in AML

For years, there have been only anecdotal reports of MMP-2 and -9-expressing myeloid cell lines (Kubota et al. 1996, Matsuzaki & Janowska-Wieczorek 1997). However, the ultimate role of these enzymes in the biology and clinical course of leukemia is poorly defined. Both Janowska-Wieczorek et al. and Ries et al. have recently published studies dealing with the MMP and TIMP expression of freshly isolated human blasts. They have shown that all myeloid leukemic blasts secrete MMP-2 and/or MMP-9 as well as TIMP-1 and TIMP-2 when measured at the protein or mRNA levels. (Janowska-Wieczorek et al. 1999, Ries et al. 1999) Because TIMPs have been shown to have growth-potentiating activity in several cell lines, the authors suggest that elevated TIMP activity, apart from its role in regulating MMP activity, could also serve as an autocrine growth stimulus.

MMPs also seem to be integral to AML cell invasion but cannot explain the whole invasion process. All-trans-retinoic acid induces differentiation of the NB4 promyelocytic cell line and, at the same time, increases its in vitro invasiveness through matrigel. NB4 cells express both MMP-2 and MMP-9and this increased invasion can be blocked with a synthetic MMP inhibitor. (Zang et al. 2000).

The results of Sawicki et al. suggest that it is specifically MMP-2, not MMP-9, that is important for AML blast invasion (Sawicki et al. 1998).

2.3.8. Gelatinases in ALL

Hendrix et al have shown with lymphoblastoid cell lines that the ability of these cells to invade through matrigel correlates with the MMP-2 expression and the propensity of cells to invade and metastasize in a SCID mouse model, while Ivanoff et al. found MMP-9 expression to be more important (Hendrix et al. 1992, Ivanoff et al. 1999). On the other hand, this may be an oversimplification of the problem. Kossakowska et al. implanted human Burkitt´s lymphoblastoid cells into a SCID mouse. Some of these cells invaded into the central nervous system, and these cells were shown to contain MMP-9 of mouse origin. Co-culture of malignant cells with mouse astrocytes led to increased MMP-9 expression by astrocytes. These results imply more complicated interactions between tumor cells and the matrix in the invasion of lymphoblasts (Kossakowska et al. 1999). The role of MMPs has not been studied in clinical materials.

2.3.9. Gelatinases in Hodgkin´s lymphoma

There are only two publications dealing with the role of gelatinases in Hodgkin´s lymphoma. Flavell et al. studied the expression of MMP-9 and its correlation with EBV status and survival in Hodgkin´s lymphoma. The authors found that all the samples were positive with MMP-9 staining, but the intensity of staining varied. No statistically significant association between MMP-9 expression and survival was found, although a trend toward a more favorable prognosis for patients with weak staining was found in Kaplan-Meier analyses. This finding was ignored by the authors. The MMP-9 expression did not correlate with the presence of EBV genome (Flavell et al. 2000). Oelman et al. discovered expression of mRNA for MMP-2, TIMP-1, and TIMP-2 in a series of 15 cases of Hodgkin´s lymphoma. They found MMP-2 expression in reactive lymphocytes, but not in malignant cells (Oelman et al. 2002).

2.3.10. Gelatinases in Non-Hodgkin´s lymphomas

Kossakowska et al. have published two reports dealing with MMP and TIMP expression in clinical non-Hodgkin´s lymphoma materials consisting of a wide range of different lymphoma subtypes. They found that low-grade non-Hodgkin´s lymphomas express relatively constant small amounts of MMP-2, MMP-9, TIMP-1, and TIMP-2, while high-grade tumors show more massive and variable, up to sevenfold expression of MMP-9 and TIMP-1. MMP-2 expression was present in less than half of the cases with high-grade histology. (Kossakowska et al. 1991, 1992) Interestingly, in situ hybridization showed TIMP synthesis to be localized in stromal cells surrounding the tumor, while MMP mRNA was localized in macrophages within the tumor and in large malignant lymphoma cells. TIMP-1 expression correlated with high MMP-9 expression, though there were also cases showing only either MMP-9 or TIMP-1 expression. Elevated TIMP expression seemed to correlate with more extensive disease stage in high-grade tumors, but the material was too small and heterogenous to allow definitive clinical correlations (Kossakowska et al. 1991). In their next study, Kosskowska et al applied northern blot analysis to samples from 18 patients with high-grade immunoblastic lymphoma. They found tumor stage and the patient’s age to be the most important prognostic factor, but after controlling for age, MMP-9 expression also seemed to be linkes to a poor prognosis (Kossakowska et al. 1992). Lately, Vacca et al. have also shown, using in situ hybridization and zymography techniques in multiple myeloma and mycosis fungoides, that enhancement of both MMP-2 and MMP-9 expression is linked to more aggressive and advanced disease and more extensive neovascularization (Vacca et al. 1997, 1999, 2000).

The findings of Lalancette et al. also support the idea that MMP-9 may contribute to the clinical aggressiveness of lymphomas. Gelatinase expression in lymphocytes require contacts with ICAM-1, fibronectin, and vitronectin. (Aoudjit et al. 1998, Vacca et al. 2001). Intercellular adhesion molecule-1 (ICAM-1)-deficient mice are usually resistant to lymphoma cell dissemination. However, with repeated in vivo passages, a highly aggressive cell line with a capability to invade in ICAM-1-deficient mice can be developed. These cell lines show high and constant MMP-9 expression. (Lalancette et al. 2000). Aoudjit et al studied a mouse thymus lymphoma model. They found that lymphoma cells induce the host’s stromal cells to synthesize MMP-9, which is needed for tumor progression (Aoudjit et al. 1997). This is in contrast to the results of Kossakowska et al, which showed MMP-9 transcripts to be localized in malignant cells (Kossakowska et al. 1991).

MMP-9 expression in non-Hodgkin´s lymphomas correlates with IL-6 expression, which, in turn, is an adverse prognostic indicator in NHL. IL-6 has been shown to be able to stimulate the MMP-9 expression and invasion of lymphoma cells in vitro. The expression of IL-6, in turn, correlates with TIMP-1 expression and an adverse prognosis (Kossakowska et al. 2000).

The expressions of MMP-9 and TIMP-1 are interconnected processes, and both are associated with an adverse prognosis in several tumor models. To clarify which one is more essential for lymphoma growth, Aoudjit et al generated lymphoma cell lines showing elevated expression of MMP-9 or TIMP-1 or both. After implantation into mouse thymus, they found the tumor’s growth rate to correlate with MMP-9 expression but not with TIMP-1 expression (Aoudjit et al. 1999).

Based on all of the above-mentioned studies, it seems evident that MMP-9 expression is somehow integral to the progression of malignant lymphomas.