TWINGEN: protocol for an observational clinical biobank recall and biomarker cohort study to identify Finnish individuals with high risk of Alzheimer's disease.

INTRODUCTION: A better understanding of the earliest stages of Alzheimer's disease (AD) could expedite the development or administration of treatments. Large population biobanks hold the promise to identify individuals at an elevated risk of AD and related dementias based on health registry information. Here, we establish the protocol for an observational clinical recall and biomarker study called TWINGEN with the aim to identify individuals at high risk of AD by assessing cognition, health and AD-related biomarkers. Suitable candidates were identified and invited to participate in the new study among THL Biobank donors according to TWINGEN study criteria. METHODS AND ANALYSIS: A multi-centre study (n=800) to obtain blood-based biomarkers, telephone-administered and web-based memory and cognitive parameters, questionnaire information on lifestyle, health and psychological factors, and accelerometer data for measures of physical activity, sedentary behaviour and sleep. A subcohort is being asked to participate in an in-person neuropsychological assessment (n=200) and wear an Oura ring (n=50). All participants in the TWINGEN study have genome-wide genotyping data and up to 48 years of follow-up data from the population-based older Finnish Twin Cohort (FTC) study of the University of Helsinki. The data collected in TWINGEN will be returned to THL Biobank from where it can later be requested for other biobank studies such as FinnGen that supported TWINGEN. ETHICS AND DISSEMINATION: This recall study consists of FTC/THL Biobank/FinnGen participants whose data were acquired in accordance with the Finnish Biobank Act. The recruitment protocols followed the biobank protocols approved by Finnish Medicines Agency. The TWINGEN study plan was approved by the Ethics Committee of Hospital District of Helsinki and Uusimaa (number 16831/2022). THL Biobank approved the research plan with the permission no: THLBB2022_83.

Connecting cohorts of Finnish biobanks creates a research resource for the study of healthy ageing.

AIMS: Connecting cohorts with biobanks is a Finnish biobank collaboration, creating an infrastructure for the study of healthy ageing. We aimed to develop a model for data integration and harmonisation between different biobanks with procedures for joint access. METHODS: The heart of the collaboration is the integrated datasets formed by using data from three biobanks: (a) Arctic Biobank, hosting regional birth cohorts and cohorts of elderly; (b) hospital-affiliated Borealis Biobank of Northern Finland; and (c) THL Biobank, hosting population-based cohorts. The datasets were created by developing a data dictionary, harmonising cohort data and with a joint pseudonymisation process. RESULTS: The connecting cohorts with biobanks resource at its widest consists altogether of almost 1.4 million individuals from collaborating biobanks. Utilising data from 107,000 cohort participants, we created harmonised datasets that contain attributes describing metabolic risk and frailty for studies of healthy ageing. These data can be complemented with medical data available from Biobank Borealis and with samples taken at hospital settings for approximately 38,000 cohort participants. In addition, the harmonised connecting cohorts with biobanks datasets can be expanded with supplementary data and samples from the collaborating biobanks. CONCLUSIONS: The connecting cohorts with biobanks datasets provide a unique resource for research on ageing-related personalised healthcare and for real-world evidence studies. Following the FAIR principles on findability, accessibility, interoperability, and reusability, the reused and harmonised datasets are findable and made accessible for researchers. The same approach can be further utilised to develop additional datasets for other research topics.

TWINGEN - protocol for an observational clinical biobank recall and biomarker study to identify individuals with high risk of Alzheimer's disease.

Introduction: A better understanding of the earliest stages of Alzheimer's disease (AD) could expedite the development or administration of treatments. Large population biobanks hold the promise to identify individuals at an elevated risk of AD and related dementias based on health registry information. Here, we establish the protocol for an observational clinical recall and biomarker study called TWINGEN with the aim to identify individuals at high risk of AD by assessing cognition, health and AD-related biomarkers. Suitable candidates were identified and invited to participate in the new study among Finnish biobank donors according to TWINGEN study criteria. Methods and analysis: A multi-center study (n=800) to obtain blood-based biomarkers, telephone-administered and web-based memory and cognitive parameters, questionnaire information on lifestyle, health and psychological factors, and accelerometer data for measures of physical activity, sedentary behavior and sleep. A sub-cohort are being asked to participate in an in-person neuropsychological assessment (n=200) and wear an Oura ring (n=50). All participants in the TWINGEN study have genome-wide genotyping data and up to 48 years of follow-up data from the population-based older Finnish Twin Cohort (FTC) study of the University of Helsinki. TWINGEN data will be transferred to Finnish Institute of Health and Welfare (THL) biobank and we aim to further to transfer it to the FinnGen study where it will be combined with health registry data for prediction of AD. Ethics and dissemination: This recall study consists of FTC/THL/FinnGen participants whose data were acquired in accordance with the Finnish Biobank Act. The recruitment protocols followed the biobank protocols approved by Finnish Medicines Agency. The TWINGEN study plan was approved by the Ethics Committee of Hospital District of Helsinki and Uusimaa (number 16831/2022). THL Biobank approved the research plan with the permission no: THLBB2022_83.

Invasion depth estimation of carcinoma cells using adaptive stain normalization to improve epidermis segmentation accuracy.

Submucosal invasion depth is a significant prognostic factor when assessing lymph node metastasis and cancer itself to plan proper treatment for the patient. Conventionally, oncologists measure the invasion depth by hand which is a laborious, subjective, and time-consuming process. The manual pathological examination by measuring accurate carcinoma cell invasion with considerable inter-observer and intra-observer variations is still challenging. The increasing use of medical imaging and artificial intelligence reveals a significant role in clinical medicine and pathology. In this paper, we propose an approach to study invasive behavior and measure the invasion depth of carcinoma from stained histopathology images. Specifically, our model includes adaptive stain normalization, color decomposition, and morphological reconstruction with adaptive thresholding to separate the epithelium with blue ratio image. Our method splits the image into multiple non-overlapping meaningful segments and successfully finds the homogeneous segments to measure accurate invasion depth. The invasion depths are measured from the inner epithelium edge to outermost pixels of the deepest part of particles in image. We conduct our experiments on skin melanoma tissue samples as well as on organotypic invasion model utilizing myoma tissue and oral squamous cell carcinoma. The performance is experimentally compared to three closely related reference methods and our method provides a superior result in measuring invasion depth. This computational technique will be beneficial for the segmentation of epithelium and other particles for the development of novel computer-aided diagnostic tools in biobank applications.

Whole slide image registration via multi-stained feature matching.

In the recent decade, medical image registration and fusion process has emerged as an effective application to follow up diseases and decide the necessary therapies based on the conditions of patient. For many of the considerable diagnostic analyses, it is common practice to assess two or more different histological slides or images from one tissue sample. A specific area analysis of two image modalities requires an overlay of the images to distinguish positions in the sample that are organized at a similar coordinate in both images. In particular cases, there are two common challenges in digital pathology: first, dissimilar appearances of images resulting due to staining variances and artifacts; second, large image size. In this paper, we develop algorithm to overcome the fact that scanners from different manufacturers have variations in the images. We propose whole slide image registration algorithm where adaptive smoothing is employed to smooth the stained image. A modified scale-invariant feature transform is applied to extract common information and a joint distance helps to match keypoints correctly by eliminating position transformation error. Finally, the registered image is obtained by utilizing correct correspondences and the interpolation of color intensities. We validate our proposal using different images acquired from surgical resection samples of lung cancer (adenocarcinoma). Extensive feature matching with apparently increasing correct correspondences and registration performance on several images demonstrate the superiority of our method over state-of-the-art methods. Our method potentially improves the matching accuracy that might be beneficial for computer-aided diagnosis in biobank applications.

Retinex model based stain normalization technique for whole slide image analysis.

Medical imaging provides the means for diagnosing many of the medical phenomena currently studied in clinical medicine and pathology. The variations of color and intensity in stained histological slides affect the quantitative analysis of the histopathological images. Moreover, stain normalization utilizing color for the classification of pixels into different stain components is challenging. The staining also suffers from variability, which complicates the automatization of tissue area segmentation with different staining and the analysis of whole slide images. We have developed a Retinex model based stain normalization technique in terms of area segmentation from stained tissue images to quantify the individual stain components of the histochemical stains for the ideal removal of variability. The performance was experimentally compared to reference methods and tested on organotypic carcinoma model based on myoma tissue and our method consistently has the smallest standard deviation, skewness value, and coefficient of variation in normalized median intensity measurements. Our method also achieved better quality performance in terms of Quaternion Structure Similarity Index Metric (QSSIM), Structural Similarity Index Metric (SSIM), and Pearson Correlation Coefficient (PCC) by improving robustness against variability and reproducibility. The proposed method could potentially be used in the development of novel research as well as diagnostic tools with the potential improvement of accuracy and consistency in computer aided diagnosis in biobank applications.

The Role of MMP8 in Cancer: A Systematic Review.

Matrix metalloproteinases (MMPs) have traditionally been considered as tumor promoting enzymes as they degrade extracellular matrix components, thus increasing the invasion of cancer cells. It has become evident, however, that MMPs can also cleave and alter the function of various non-matrix bioactive molecules, leading to both tumor promoting and suppressive effects. We applied systematic review guidelines to study MMP8 in cancer including the use of MMP8 as a prognostic factor or as a target/anti-target in cancer treatment, and its molecular mechanisms. A total of 171 articles met the inclusion criteria. The collective evidence reveals that in breast, skin and oral tongue cancer, MMP8 inhibits cancer cell invasion and proliferation, and protects patients from metastasis via cleavage of non-structural substrates. Conversely, in liver and gastric cancers, high levels of MMP8 worsen the prognosis. Expression and genetic alterations of MMP8 can be used as a prognostic factor by examination of the tumor and serum/plasma. We conclude, that MMP8 has differing effects on cancers depending on their tissue of origin. The use of MMP8 as a prognostic factor alone, or with other factors, seems to have potential. The molecular mechanisms of MMP8 in cancer further emphasize its role as an important regulator of bioactive molecules.

Matrix metalloproteinase 9 inhibits the motility of highly aggressive HSC-3 oral squamous cell carcinoma cells.

Pro-tumorigenic activities of matrix metalloproteinase (MMP) 9 have been linked to many cancers, but recently the tumour-suppressing role of MMP9 has also been elucidated. The multifaceted evidence on this subject prompted us to examine the role of MMP9 in the behaviour of oral tongue squamous cell carcinoma (OTSCC) cells. We used gelatinase-specific inhibitor, CTT2, and short hairpin (sh) RNA gene silencing to study the effects of MMP9 on proliferation, motility and invasion of an aggressive OTSCC cell line, HSC-3. We found that the migration and invasion of HSC-3 cells were increased by CTT2 and shRNA silencing of MMP9. Proliferation, in turn, was decreased by MMP9 inhibition. Furthermore, arresten-overexpressing HSC-3 cells expressed increased levels of MMP9, but exhibited decreased motility compared with controls. Interestingly, these cells restored their migratory capabilities by CTT2 inhibition of MMP9. Hence, although higher MMP9 expression could give rise to an increased tumour growth in vivo due to increased proliferation, in some circumstances, it may participate in yet unidentified molecular mechanisms that reduce the cell movement in OTSCC.

Human Tumor Tissue-Based 3D In Vitro Invasion Assays.

Here we describe a protocol to utilize human benign leiomyoma tissue in in vitro 3D model that enables an assessment of cell invasion. The chapter also describes detailed instructions for image analysis to quantify the results. Leiomyoma is a benign tumor of the uterus which mimics authentic components of the tumor microenvironment including fibroblasts, vessels, collagen fibers, and extracellular protein composition. The leiomyoma invasion model represents a superior 3D model for cell invasion studies compared to the other non-human organotypic models.

Neoplastic extracellular matrix environment promotes cancer invasion in vitro.

The invasion of carcinoma cells is a crucial feature in carcinogenesis. The penetration efficiency not only depends on the cancer cells, but also on the composition of the tumor microenvironment. Our group has developed a 3D invasion assay based on human uterine leiomyoma tissue. Here we tested whether human, porcine, mouse or rat hearts as well as porcine tongue tissues could be similarly used to study carcinoma cell invasion in vitro. Three invasive human oral tongue squamous cell carcinoma (HSC-3, SCC-25 and SCC-15), melanoma (G-361) and ductal breast adenocarcinoma (MDA-MB-231) cell lines, and co-cultures of HSC-3 and carcinoma-associated or normal oral fibroblasts were assayed. Myoma tissue, both native and lyophilized, promoted invasion and growth of the cancer cells. However, the healthy heart or tongue matrices were unable to induce the invasion of any type of cancer cells tested. Moreover, when studied in more detail, small molecular weight fragments derived from heart tissue rinsing media inhibited HSC-3 horizontal migration. Proteome analysis of myoma rinsing media, on the other hand, revealed migration enhancing factors. These results highlight the important role of matrix composition for cancer invasion studies in vitro and further demonstrate the unique properties of human myoma organotypic model.

Endostatin induces proliferation of oral carcinoma cells but its effect on invasion is modified by the tumor microenvironment.

The turnover of extracellular matrix liberates various cryptic molecules with novel biological activities. Endostatin is an endogenous angiogenesis inhibitor that is derived from the non-collagenous domain of collagen XVIII. Although there are a large number of studies on its anti-tumor effects, the molecular mechanisms are not yet completely understood, and the reasons why endostatin has not been successful in clinical trials are unclear. Research has mostly focused on its anti-angiogenic effect in tumors. Here, we aimed to elucidate how endostatin affects the behavior of aggressive tongue HSC-3 carcinoma cells that were transfected to overproduce endostatin. Endostatin inhibited the invasion of HSC-3 cells in a 3D collagen-fibroblast model. However, it had no effect on invasion in a human myoma organotypic model, which lacks vital fibroblasts. Recombinant endostatin was able to reduce the Transwell migration of normal fibroblasts, but had no effect on carcinoma associated fibroblasts. Surprisingly, endostatin increased the proliferation and decreased the apoptosis of cancer cells in organotypic models. Also subcutaneous tumors overproducing endostatin grew bigger, but showed less local invasion in nude mice xenografts. We conclude that endostatin affects directly to HSC-3 cells increasing their proliferation, but its net effect on cancer invasion seem to depend on the cellular composition and interactions of tumor microenvironment.

Macrophages modulate migration and invasion of human tongue squamous cell carcinoma.

Oral tongue squamous cell carcinoma (OTSCC) has a high mortality rate and the incidence is rising worldwide. Despite advances in treatment, the disease lacks specific prognostic markers and treatment modality. The spreading of OTSCC is dependent on the tumor microenvironment and involves tumor-associated macrophages (TAMs). Although the presence of TAMs is associated with poor prognosis in OTSCC, the specific mechanisms underlying this are still unknown. The aim here was to investigate the effect of macrophages (Mfs) on HSC-3 tongue carcinoma cells and NF-kappaB activity. We polarized THP-1 cells to M1 (inflammatory), M2 (TAM-like) and R848 (imidazoquinoline-treated) type Mfs. We then investigated the effect of Mfs on HSC-3 cell migration and NF-kappaB activity, cytokine production and invasion using several different in vitro migration models, a human 3D tissue invasion model, antibody arrays, confocal microscopy, immunohistochemistry and a mouse invasion model. We found that in co-culture studies all types of Mfs fused with HSC-3 cells, a process which was partially due to efferocytosis. HSC-3 cells induced expression of epidermal growth factor and transforming growth factor-beta in co-cultures with M2 Mfs. Direct cell-cell contact between M2 Mfs and HSC-3 cells induced migration and invasion of HSC-3 cells while M1 Mfs reduced HSC-3 cell invasion. M2 Mfs had an excess of NF-kappaB p50 subunit and a lack of p65 subunits both in the presence and absence of HSC-3 cells, indicating dysregulation and pro-tumorigenic NF-kappaB activation. TAM-like cells were abundantly present in close vicinity to carcinoma cells in OTSCC patient samples. We conclude that M2 Mfs/TAMs have an important role in OTSCC regulating adhesion, migration, invasion and cytokine production of carcinoma cells favouring tumor growth. These results demonstrate that OTSCC patients could benefit from therapies targeting TAMs, polarizing TAM-like M2 Mfs to inflammatory macrophages and modulating NF-kappaB activity.

Caveolin-1 accumulation in the tongue cancer tumor microenvironment is significantly associated with poor prognosis: an in-vivo and in-vitro study.

BACKGROUND: Caveolin-1 (CAV1) may be upregulated by hypoxia and acts in a tumor-dependent manner. We investigated CAV1 in tongue squamous cell carcinoma (TSCC) and its association with clinical outcomes, and studied in vitro possible ways for CAV1 accumulation in the tumor microenvironment (TME). METHODS: TSCC cases (N = 64) were immunohistochemically stained for CAV1. Scores were separately assessed in the tumor and TME and plotted for association with recurrence and survival (univariate analysis with log-rank test). In vitro studies were performed on a 3D myoma organotypic model, a mimicker of TME. Prior to monoculturing HSC-3 tongue cancer cells, the model underwent modifications in oxygenation level (1%O2 hypoxia to upregulate CAV1) and/or in the amount of natural soluble factors [deleted by 14-day rinsing (rinsed myoma, RM), to allow only HSC-3-derived factors to act]. Controls included normoxia (21%O2) and naturally occurring soluble factors (intact myoma, IM). HSC-3 cells were also co-cultured with CaDEC12 cells (fibroblasts exposed to human tongue cancer). CAV1 expression and cellular distribution were examined in different cellular components in hypoxic and rinsed myoma assays. Twist served as a marker for the process of epithelial-mesenchymal transition (EMT). Exosomes isolated from HSC-3 media were investigated for containing CAV1. RESULTS: Expression of CAV1 in TSCC had a higher score in TME than in the tumor cells and a negative impact on recurrence (p = 0.01) and survival (p = 0.003). Monocultures of HSC-3 revealed expression of CAV1 mainly in the TME-like myoma assay, similar to TSCC. CAV1+, alpha-smooth muscle actin (αSMA) + and Twist + CAF-like cells were observed surrounding the invading HSC-3, possibly reflecting EMT. RM findings were similar to IM, inferring action of HSC-3 derived factors, and no differences were seen when hypoxia was induced. HSC-3-CaDEC12 co-cultures revealed CAV1+, αSMA+ and cytokeratin-negative CAF-like cells, raising the possibility of CaDEC12 cells gaining a CAF phenotype. HSC-3-derived exosomes were loaded with CAV1. CONCLUSIONS: Accumulation of CAV1-TME in TSCC had a negative prognostic value. In vitro studies showed the presence of CAV1 in cancer cells undergoing EMT and in fibroblasts undergoing trans-differentiation to CAFs. CAV1 delivery to the TME involved cancer cell-derived exosomes.

Anti-heparanase aptamers as potential diagnostic and therapeutic agents for oral cancer.

Heparanase is an endoglycosidase enzyme present in activated leucocytes, mast cells, placental tissue, neutrophils and macrophages, and is involved in tumour metastasis and tissue invasion. It presents a potential target for cancer therapies and various molecules have been developed in an attempt to inhibit the enzymatic action of heparanase. In an attempt to develop a novel therapeutic with an associated diagnostic assay, we have previously described high affinity aptamers selected against heparanase. In this work, we demonstrated that these anti-heparanase aptamers are capable of inhibiting tissue invasion of tumour cells associated with oral cancer and verified that such inhibition is due to inhibition of the enzyme and not due to other potentially cytotoxic effects of the aptamers. Furthermore, we have identified a short 30 bases aptamer as a potential candidate for further studies, as this showed a higher ability to inhibit tissue invasion than its longer counterpart, as well as a reduced potential for complex formation with other non-specific serum proteins. Finally, the aptamer was found to be stable and therefore suitable for use in human models, as it showed no degradation in the presence of human serum, making it a potential candidate for both diagnostic and therapeutic use.

Insights into the role of components of the tumor microenvironment in oral carcinoma call for new therapeutic approaches.

The research on oral cancer has focused mainly on the cancer cells, their genetic changes and consequent phenotypic modifications. However, it is increasingly clear that the tumor microenvironment (TME) has been shown to be in a dynamic state of inter-relations with the cancer cells. The TME contains a variety of components including the non-cancerous cells (i.e., immune cells, resident fibroblasts and angiogenic vascular cells) and the ECM milieu [including fibers (mainly collagen and fibronectin) and soluble factors (i.e., enzymes, growth factors, cytokines and chemokines)]. Thus, it is currently assumed that TME is considered a part of the cancerous tissue and the functionality of its key components constitutes the setting on which the hallmarks of the cancer cells can evolve. Therefore, in terms of controlling a malignancy, one should control the growth, invasion and spread of the cancer cells through modifications in the TME components. This mini review focuses on the TME as a diagnostic approach and reports the recent insights into the role of different TME key components [such as carcinoma-associated fibroblasts (CAFs) and inflammation (CAI) cells, angiogenesis, stromal matrix molecules and proteases] in the molecular biology of oral carcinoma. Furthermore, the impact of TME components on clinical outcomes and the concomitant need for development of new therapeutic approaches will be discussed.

Human bone marrow mesenchymal stem cells induce collagen production and tongue cancer invasion.

Tumor microenvironment (TME) is an active player in carcinogenesis and changes in its composition modify cancer growth. Carcinoma-associated fibroblasts, bone marrow-derived multipotent mesenchymal stem cells (BMMSCs), and inflammatory cells can all affect the composition of TME leading to changes in proliferation, invasion and metastasis formation of carcinoma cells. In this study, we confirmed an interaction between BMMSCs and oral tongue squamous cell carcinoma (OTSCC) cells by analyzing the invasion progression and gene expression pattern. In a 3-dimensional myoma organotypic invasion model the presence of BMMSCs inhibited the proliferation but increased the invasion of OTSCC cells. Furthermore, the signals originating from OTSCC cells up-regulated the expression of inflammatory chemokines by BMMSCs, whereas BMMSC products induced the expression of known invasion linked molecules by carcinoma cells. Particularly, after the cell-cell interactions, the chemokine CCL5 was abundantly secreted from BMMSCs and a function blocking antibody against CCL5 inhibited BMMSC enhanced cancer invasion area. However, CCL5 blocking antibody did not inhibit the depth of invasion. Additionally, after exposure to BMMSCs, the expression of type I collagen mRNA in OTSCC cells was markedly up-regulated. Interestingly, also high expression of type I collagen N-terminal propeptide (PINP) in vivo correlated with the cancer-specific mortality of OTSCC patients, whereas there was no association between cancer tissue CCL5 levels and the clinical parameters. In conclusion, our results suggest that the interaction between BMMSC and carcinoma cells induce cytokine and matrix molecule expression, of which high level of type I collagen production correlates with the prognosis of OTSCC patients.

Cathepsin K is present in invasive oral tongue squamous cell carcinoma in vivo and in vitro.

OBJECTIVES: Cathepsin K, a lysosomal cysteine protease, is expressed in the tumor microenvironment (TME) of skin carcinoma, but nothing is known about cathepsin K in oral tongue squamous cell carcinoma (OTSCC). Our aim was to describe the expression of cathepsin K in invasive OTSCC in vitro and in a series of clinical cancer specimens. MATERIALS AND METHODS: OTSCC invasion in vitro was studied using invasive HSC-3 tongue carcinoma cells in 3D organotypic models. In total, 121 mobile tongue OTSCCs and 10 lymph node metastases were analyzed for cathepsin K expression. The association between cathepsin K expression and clinicopathological factors was evaluated. RESULTS: Cysteine protease inhibitor E64 and cathepsin K silencing significantly (p<0.0001) reduced HSC-3 cell invasion in the 3D models. Cathepsin K was expressed in a majority of carcinoma and metastatic cells, but the expression pattern in carcinoma cells did not correlate with clinical parameters. Instead, the weak expression of cathepsin K in the invasive TME front correlated with increased overall recurrence (p<0.05), and in early-stage tumors this pattern predicted both cancer recurrence and cancer-specific mortality (p<0.05 and p<0.005, respectively). CONCLUSIONS: Cathepsin K is expressed in OTSCC tissue in both carcinoma and TME cells. Although the diminished activity and expression in aggressive tongue HSC-3 cells reduced 3D invasion in vitro, the amount of cathepsin K in carcinoma cells was not associated with the outcome of cancer patients. Instead, cathepsin K in the invasive TME front seems to have a protective role in the complex progression of tongue cancer.

Fluctuating roles of matrix metalloproteinase-9 in oral squamous cell carcinoma.

One hallmark of cancer is the degradation of the extracellular matrix (ECM), which is caused by proteinases. In oral cancers, matrix metalloproteinases (MMPs), especially MMP-9, are associated with this degradation. MMPs break down the ECM allowing cancer to spread; they also release various factors from their cryptic sites, including cytokines. These factors modulate cell behavior and enhance cancer progression by regulating angiogenesis, migration, proliferation, and invasion. The development of early metastases is typical for oral cancer, and increased MMP-9 expression is associated with a poor disease prognosis. However, many studies fail to relate MMP-9 expression with metastasis formation. Contrary to earlier models, recent studies show that MMP-9 plays a protective role in oral cancers. Therefore, the role of MMP-9 is complicated and may fluctuate throughout the different types and stages of oral cancers.