Matthias Hammerschmidt - A 4

The multiple facets of the type II transmembrane serine protease ST14 and hypotonic stress during carcinogenesis in fish and mammals

The power of anti-cancer therapies is often compromised by the fact that targeted cells, processes and/or molecules can display a combination of oncogenic and tumor-suppressive effects. Our work has unravelled a similar double face for Matriptase 1, also named ST14, a type II transmembrane serine protease, and its downstream mediators epidermal growth factor receptor (EGFR) and phospholipase D (PLD). In zebrafish embryos lacking the cognate ST14 inhibitor Hai1, the hyper-stimulated ST14 – EGFR – PLD pathway promotes both mTORC1-mediated epidermal hyperplasia (oncogenic) and sphingosine-1-phosphate (S1P)-mediated epidermal cell loss (tumor-suppressive). The latter involves entosis of preneoplastic basal keratinocytes by neighboring surface keratinocytes and the subsequent apical extrusion of these cell complexes. Accordingly, combined treatment with the PLD inhibitor FIPI to block the shared signalling pathway and S1P to specifically reactivate the tumor-suppressive branch leads to a perfect rescue and survival of otherwise lethal zebrafish hai1 null mutants. This treatment regime needs to be tested in mammalian carcinoma systems. In addition, the epigenetic basis of the heterogeneity among basal keratinocytes of zebrafish hai1 mutants, and the genetic control of the cell competition, entosis and apical cell extrusion processes leading to the preferential extrusion of preneoplastic basal cells need to be further dissected and characterized.


ST14 (Suppression of Tumorigenicity-14) is an epithelia-specific cell surface-associated protease that reshapes its microenvironment by degrading ECM proteins, while cleaving and activating signaling molecules like growth factors and their receptors. This pericellular proteolysis is crucial for multiple processes of epithelial development, homeostasis and regeneration. Pericellular proteolysis is also critically involved in tumor growth and invasiveness, and upregulation of ST14 has been implicated in a variety of epithelial cancers, while on the other hand, tumor-suppressive functions of ST14 have been revealed (Tanabe and List, 2017). About a dozen ST14 target proteins have been identified and characterized, mainly using in vitro systems. However, the in vivo relevance of identified targets, in particular in the context of carcinogenesis, remains largely elusive, although better knowledge would greatly facilitate the development of targeted anti-carcinogenic therapies. 

Zebrafish models of increased ST14 activity and epidermal malignancy

We have identified zebrafish mutants in the ST14 inhibitor Hai1, which due to increased ST14 activity display compromised epithelial integrity of the embryonic epidermis, keratinocyte hyper-proliferation (Carney et al., 2007; Figure 1E-I) and partial basement membrane degradation, signs of early stage carcinogenesis. However, keratinocytes usually do not invade the underlying dermis, but are extruded apically, which has a tumor-suppressive effect (Figure 1A-D), possibly underlying the progressive healing of the skin defects of hypomorphic hai1 mutants. In contrast, basal keratinocytes of zebrafish pso mutants in ATP1b1a, a Na/K-ATPase β-subunit, display stronger ST14-dependent epidermal malignancy, with basal keratinocytes leaving basally and undergoing metastasis (Hatzold et al., 2016 and unpublished results, Figure 2). 

Identification and characterization of branched pathways downstream of ST14 mediating oncogenesis versus tumor suppression

To identify effectors of ST14-dependent carcinogenesis, we carried out a systematic search for annotated small compounds that alleviate or worsen the epidermal defects of hypomorphic zebrafish hai1 mutants. Crucial pathway components identified in this manner are the formerly known ST14 targets Par2b and EGFR and, as a novel mediator, phospholipase D (PLD).

Strikingly, the pathway branches at the level of the PLD product phosphatidic acid (PA), which promotes keratinocyte hyperproliferation, invasiveness and thereby oncogenesis via mTORC1 (branch 1), and apical cell extrusion and thereby tumor suppression via sphingosine kinase (Sphk) and its product sphingosine-1-phosphate / S1P (branch 2; Figure 1K). Accordingly, amorphic hai1 mutants, which in contrast to the spontaneously healing hypomorphs are lethal, can only be temporarily rescued by pharmacological inhibition of PLD with FIPI. Later, most likely due to the concomitant blockage of the tumor-suppressive branch, they relapse with epidermal hyperplasia even more severe than in untreated mutant siblings. To circumvent this relapse, we reactivated apical cell extrusion in FIPI-treated mutants via administration of S1P (Figure 1E), resulting in a long-lasting rescue of epidermal hyperplasia and a complete survival of the otherwise lethal mutants (Figure 1F-J) (Armistead et al., in revision). We also discovered how apical cell extrusion, which occurs at the surface of the skin, can be tumor-suppressive, given that it is the basal keratinocytes that hyper-proliferate: before their own extrusion, surface cells selectively engulf underlying preneoplastic basal cells (Figure 1A-C) in a process similar to entosis that was formerly identified in epithelial cancer cell lines but had not been described for carcinoma cells in vivo thus far.


The ultimate goal of the project is to tailor improved human cancer therapies, treating carcinomas with combinations of identified small molecules to specifically block oncogenic pathways, while leaving tumor-suppressive pathways and the according processes (cell-cell competition, entosis, apical cell extrusion) unaffected. 

Selected publications

1. Armistead, J., Hatzold, J., van Roye, A., Fahle, E. and Hammerschmidt M. Entosis and apical cell extrusion constitute a tumour suppressive mechanism downstream of Matriptase. J. Cell Biol., in revision

2. Carney, T.J., von der Hardt, S., Sonntag, C., Amsterdam, A., Topczewski, J., Hopkins, N. and Hammerschmidt, M. (2007). Inactivation of serine protease Matriptase1a by its inhibitor Hai1 is required for epithelial integrity of the zebrafish epidermis. Development 134, 3461-71

3. Hatzold, J., Beleggia, F., Herzig, H., Altmüller, J., Nürnberg, P., Bloch, W., Wollnik, B. and Hammerschmidt, M. (2016). Tumor suppression in basal keratinocytes via dual non-cell autonomous functions of a Na,K-ATPase beta subunit. Elife 5, e14277

4. Tanabe, L.M. and List, K. (2017). The role of type II transmembrane serine protease mediated signaling in cancer. FEBS J. 284, 1421-36 

Prof. Dr. Matthias Hammerschmidt

Zoological Institute

Prof. Dr. Matthias Hammerschmidt

Principal Investigator A 4
Executive Board Member

Work +49 221 470 5665

Fax (Work) +49 221 470 5184

Institute for Developmental Biology
Zülpicher Str. 47b
50674 Cologne

Publications - Matthias Hammerschmidt

Link to PubMed

Group Members

Hans-Martin Pogoda (Senior Scientist)
Julia Hatzold (PostDoc)
Joy Armistead (PostDoc)
Erica Benard (PostDoc)
Daniela Welcker (PostDoc)
Ismail Küçükaylak (PhD Student)
Indra Möllenkotte (PostDoc)
Madhuri Puvvada (PhD Student)
Philip Reinoss (PhD Student)
Heike Wessendorf (Technician)
Iris Quinkertz-Riedl (Technician)
Evelin Fahle (Technician)
Christel Schenkel (Technician) 

Figure 1

CMMC Research Odenthal
Figure 1: A-C: Apical cell extrusion and entosis in hai1 mutant embryos; (A) Single confocal plane at basal level of extruding cell complex as shown in (B), revealing rosette of krt4:GFP-labeled surface cells forming concentric actin ring underneath a p63:dsRed-labeled basal cell; (B) Transverse cryosection of extruding surface cell entosing basal cell; position of actin ring (see A) indicated by arrows; (C) Cell cluster as in (B), after extrusion, recovered from medium. D: Quantification of extruded cell numbers in siblings and hai1 mutant embryos after FIPI or S1P treatments. F-I: BrdU incorporation (red) in sibling and hai1 mutant embryos treated with FIPI and S1P, as outlined in E (hpf, hours post fertilization). J: Survival curves of siblings and hai1 null mutants treated with FIPI and/ or S1P, as outlined in E. K: Schematic of ST14 signaling pathway in hai1 mutant embryos, summarizing all data described in Armistead et al., in revision. Mmp9, Matrix metalloprotease 9.

Figure 2

CMMC Research Odenthal
Figure 2: Basal keratinocytes of zebrafish atp1b1a mutant embryos undergo metastasis. Lineage-traced basal keratinocytes are in red, basement membranes (Nid2a) in green; transverse sections through embryos at 96 hpf. In wild-type siblings, labelled basal cells remain in the epidermis (A), whereas in pso/atp1b1a mutants, they are found in the dermis underneath the fragmented skin basement membrane (B), as well as within deeper tissues (C), including digestive organs (D; magnified view of region boxed in N).