Influence of the molecular weight of immobilized hyaluronic acid on the behavior of CD44-overexpressing cells

INTRODUCTION

Paxillin, a focal adhesion (FA) adaptor protein, associated with tumor invasiveness and progression, is identified as a key regulator of cancer cell migration.¹ Hyaluronic Acid (HA), a negatively charged polysaccharide is a naturally occurring glycosaminoglycan. It is known to have affinity to a surface cell receptor, CD44, overexpressed on gastric cancer cell lines, MKN-45.² This transmembrane protein is related to paxillin expression, since it is associated to the activation of Src-kinase by CD44, a downstream effector of which regulates the phosphorilation of paxillin and consequently, the formation of invadopodia.³ We applied the Layer-by-Layer (LbL) approach to deposit HA of different Mw on gold surface. This approach allows surface immobilization of biomolecules with minimal or no changes of their chemistry, i.e. with preserved bioactivity. Moreover, the immobilized molecules are presented in an ECM fashion: they are spatially restricted but enough flexible to reorganize and bind to CD44.

EXPERIMENTAL METHODS

LbL build-up: the sequential adsorption (10 layers) of HA (6.4 kDa, 752 kDa or 1500 kDa) and PLL (30-70 kDa), was performed on gold quartz crystal sensors). HA was the last layer of all LbL constructs. QCM-D analysis: a suspension of MKN45 was introduced into the QCM-D chambers and allow the cell adhesion onto the surfaces. The process was monitored in situ by the QCM-D: changes in frequency and dissipation. Immunostaining: The adherent cells were fixed and stained for paxillin, actin (phaloidin) and nuclei (DAPI). Statistical Analysis: Shapiro-Wilk normality tests, Kruskal-Wallis and Mann−Whitney test. ***Significant differences (p<0.001).

RESULTS AND DISCUSSION

MKN45 adhered to all studied substrates. However, a significantly less adherent cells were observed for the LbL with HA of low Mw, i.e. the substrate (PLL-HA_6.5)₅ DD/DF plots show similar changes in the DF for all studied Mw but very high DD for the (PLL-HA₇₅₂)₅ and (PLL-HA₁₅₀₀)₅. These results can be explained with the activity of the adherent MKN45 that bind and remodel the HA of high Mw (752 kDa and 1500 kDa). The different signal obtained for the MKN45 on (PLL-HA_6.5)₅ shows inability of the MKN45 to reorganize the HA in the LbL system and is due either to lower adhesion strength between the cells and the substrate or to the very tight binding between the PLL and HA. The paxillin staining, confirmed the first hypothesis and showed more contact sites between MKN45 and (PLL-HA₁₅₀₀)₅ as compared to (PLL-HA_6.5)₅ and (PLL-HA₇₅₂)₅

 CONCLUSION

We demonstrate the LbL deposition is a feasible approach for presenting HA in a ECM relevant way. Adhesion of MKN45 is affected by the Mw of the immobilized HA.: more cells adhered to substrates with HA of high Mw (752 and 1500 kDa). Moreover, the adhesion on these surfaces is also stronger as demonstrated by the ability of MKN45 to reorganize the LbL built up with these HAs. Finally, the substrate (PLL-HA₁₅₀₀)₅ is also the one that induces a higher expression of paxillin, one of the major components of FAs.