Extracellular Matrix Invasion in Metastases and Angiogenesis: Commentary on the Matrigel “Chemoinvasion Assay”

In the late 1970s and early 1980s, the role of extracellular matrix (ECM) and basement membranes (BM) in hindering cellular migration within distant organs became widely accepted. The pivotal work of Lance Liotta at NCI (Rockville, MD) and Josh Fidler at MD Anderson Cancer Center (Houston, TX), describing metastasis as a multistep phenomenon of which crossing ECM compartments was a rate-limiting step, was crucial for further studies on the topic (1, 2). BMs are continuous sheets of ECM underlying most epithelial or endothelial cells. Several leaders in the field over 30 years ago devoted their careers to study the interactions of metastatic tumor cell with BMs. Laminin, its major component, was first isolated by Pam Gehron-Robey and Rupert Timpl in George Martin's laboratory at NIH (Bethesda, MD) and published in 1979. The predominant components of BMs besides laminin are type IV collagen, the proteoglycan perlecan, nidogen/entactin, and several other factors (including collagen VII, collagen XVII, and collagen XVIII).

The network formed by these molecules creates a dense, although thin, barrier that hinders the passage of cells and macromolecules. BMs can become permeable during inflammation, tissue development and repair, and in invasion and metastasis. Isolation and characterization of the molecular components of these matrices was relatively difficult, as compared with that of many other connective tissues. Early assays to study invasion employed tissues that contained BMs as barriers, and they were sort of cumbersome; the tissues included chicken chorioallantoic membrane, bladder wall, chick heart, and the amnion (3).

When I arrived in the United States after three years of postdoctoral work on tumor cell chemotaxis at the Max Planck Institute for Biochemistry in Martinsried (Germany), the laboratory of George R. Martin and Hynda Kleinman at NIH was investigating the Engelbreth–Holm–Swarm (EHS) tumor. This spontaneously arisen transplantable tumor produced multiple layers of BM. As Lance Liotta states in his historical editorial (1): “An explosion of new knowledge about the biochemistry of basement membranes and the generation of reagents for cancer research emanated from the discovery of the Engelbreth-Holm-Swarm (EHS) sarcoma. The EHS sarcoma matrix, for the first time, provided biochemical quantities of basement membrane collagen, proteoglycans, and glycoproteins.”

The major protein and glycan components of BMs were characterized from this tumor and their cDNAs cloned (3). Extracts of the EHS tumor had the ability to form a gel, which was named Matrigel (from matrix and gel). Matrigel provided a BM matrix with several unique characteristics, which became widely known (4); it is liquid at 4°C but can polymerize into a gel when warmed, producing a 3-dimensional matrix. The gel exerts biological activity on numerous cell types, recapitulating several properties of the tumor microenvironment.

Very shortly following my arrival at NIH as a Fogarty Visiting Scientist, I had the idea to use diluted Matrigel to coat the porous filters of the Boyden chambers, the common device for studying cell migration, that I had extensively employed at the Max Planck Institute to investigate the enhanced migratory capabilities of cancer cells (5) and the delayed migration of senescent fibroblasts (6). This latter study obtained the Max-Buerger award from the German Society of Gerontology.

In the assay I developed at NIH, Nucleopore chemotaxis filters with 8 or 12 μm diameter pores were coated with a 1:10 or 1:20 dilution of liquid Matrigel and allowed to dry. The Matrigel was then reconstituted with culture media, creating a thin, compact coating of reconstituted BM, as verified by electron microscopy. The filter layered with rehydrated Matrigel was placed between the two compartments of a Boyden chemotaxis chamber on a porous filter and proved to be a powerful barrier to cell migration of “normal cells,” while metastatic ones were able to penetrate it in response to stimulatory factors, mimicking metastasis to distant organs. The different times of assay, Matrigel concentration, type of chemoattractant, cell number, and cell types were tested and published in the Cancer Research article (7). The Boyden chamber–Matrigel system was proposed as a rapid and innovative assay to quantify the invasive potential of tumor cells and named the “chemoinvasion assay.”

I had the fortune and the privilege in the years from 1985 to 1987, when the article was conceived and then published, to meet and discuss with giants in the field: Josh Fidler at MD Anderson (Houston, TX), Lance Liotta, Stuart Aaronson, Mark Lippman, and Mike Sporn, all at NCI (Bethesda, MD). From those scientists, I received precious advice, insight, and material related to metastasis, invasion, cancer biology, extracellular matrix, and the “tumor microenvironment” (the term was introduced later). I obtained cancer cell cultures and normal or benign counterparts from several laboratories and used a large variety of them, and with the help of the Japanese postdoctoral fellow Yukihide “Rocky” Iwamoto, to validate the method.

The vast international use and the “longevity” of the assay is closely related to its properties: it is a simple method, reproducible, and it can be adapted to commercial chambers, such as the transwell plates, with cross-laboratory consistency. It has allowed and is allowing researchers to: (i) identify factors that can induce/promote invasion; (ii) measure and discriminate different degrees of “malignancy”; (iii) provide evidence about migratory/invasive cellular functions that can then be verified in vivo; and (iv) elucidate the role of molecules or pathways crucial for migratory/invasive properties.

In the Cancer Research article (7), we presented another application of Matrigel in cancer research, besides the chemoinvasion assay with the Boyden chamber. This was the use of a thick layer of Matrigel in culture wells for morphologic studies of tumor cell behavior; the publication was the first showing the branching, invasive morphology adopted by metastatic tumor cells on Matrigel. In that publication, as an example, a different morphology was demonstrated by cells from benign prostate hyperplasia, and benign and malignant prostate cancer cells, the latter forming branching networks in a collective invasion modality. This second application, a simple, fast, and visual assay, was later defined as the “Matrigel morphogenesis assay.” The assays received the Doerenkamp Zbinden/John's Hopkins award as an alternative to animal research.

The chemoinvasion assay with Matrigel and its numerous variations, 30 years after its publication, is still the most employed in vitro system for testing cells with different invasive abilities, has been applied to angiogenesis, and is able to provide insights into numerous biological investigations. This model was optimized to fit the needs of individual laboratories and to run several samples at the same time, and it has been adapted to commercial packages. Screening systems, first employed by Mary Hendrix as the MICS assay (8), are also commercially available.

The chemoinvasion assay was improved for use under sterile conditions, allowing the isolation of cells with a higher invasive potential from mixed populations of cell lines or primary tumors. We proposed the use of the tool for selection of highly migratory cells (3), and it was employed by others, providing isolate cell populations with a greater metastatic power in vivo and formation of colonies in soft agar (9).

Concepts on metastasis have evolved since the introduction of the chemoinvasion assay. Cancer stem cells (CSC) or cancer-initiating cells are able to recapitulate the tumor of origin when transplanted in vivo. They are characterized by a slower proliferative rate or are relatively quiescent and undergo asymmetric division. CSCs have been reported to be resistant to most chemotherapy agents. Various investigations have implicated that metastasis comes from CSCs. Highly invasive CSC subpopulations can be selected in vitro using the chemoinvasion assay (9). They show decreased differentiation, enhanced colony formation in soft agar, and demonstrate the ability to from spheroids when cultured in stemness conditions (serum deprivation in low attachment supports), higher invasive capabilities, increased tumor take, and metastases formation in vivo. Matrigel has been shown by us and others to support the take of even a few tumor cells in vivo (10, 11), providing a microenvironment particularly convenient for CSCs. The chemoinvasion assay can represent an inexpensive cell-sorting apparatus to be applied in stem cell research and to isolate rare cell populations, retaining high viability of the selected cells.

The development of imaging techniques and cell-labeling approaches also provides a new application for the 3-dimensional assay of morphogenesis. Social networks can be studied. Evidence shows that tumors often invade as cords of cells closely interacting (12), a phenomenon termed “collective migration.” These cell societies may be easily studied in the morphogenesis assay in Matrigel.

Novel tools have been envisaged that can be applied in the chemoinvasion assays. The use of siRNA oligonucleotides and the CRISPR/Cas9 system to silence expression of specific genes provides molecular technologies that can elucidate the role of new pathways in cancer, metastases, angiogenesis, and developmental biology. Cancer cells are stimulated to migrate as a function of tumor size, hypoxia, and metabolism, the effects of which can be tested in the chemoinvasion assay. Nanotechnology has been proposed for tracing and treating cells. We are undergoing rapid evolution in technology that, combined with the simplicity of the chemoinvasion assay, still offers a vast range of applications.

The development of the chemoinvasion assay in the NIH laboratory, the “house of Matrigel,” led to life-long friendships and to my wedding to next-door scientist Douglas Noonan that was celebrated in the same year of the Cancer Research article publication and was followed by the birth of our two children, Thomas and Silvia Noonan, now university students.

No potential conflicts of interest were disclosed.

I thank Antonino Bruno, Gabriele D'Uva, and Silvia Noonan for critical reading of the manuscript. I thank George Martin and Hynda Kleinman, formerly at Laboratory of Developmental Biology and Anomalies, NIH (Bethesda, MD), for their encouragement as my mentors, and Douglas Noonan for a life besides Matrigel.

The studies are supported by grants from the AIRC (Associazione Italiana per la Ricerca sul Cancro) and the Health Ministry.