Patterning mammalian cells for modeling three types of naturally occurring cell-cell interactions.

Herein we report a method for the simulation of three types of cell–cell interactions in vitro on the same substrate by using selective modification of the surface by microfluidic channels. In vivo interactions between different types of cells include the following types: 1) those between two types of cells that are both immobilized and confined to isolated areas (for brevity, we call this kind of interaction “type I”), for example, epithelial cells and fibroblasts, and epithelial cells and polar cells during ovarian development; 2) those between one cell type that is immobile and another that moves freely (“type II”), such as glial cells and neurons during the development of the nervous system and in various neurodegenerative disorders; and 3) those between two or more types of cells that are both moving freely (“type III”), for example, hepatocytes and fibroblasts in the liver. The development of an organism, the establishment of neural networks, and the formation of tumors involve all aforementioned types of interactions between different types of cells. None of the existing techniques can, however, dynamically control all of these behaviors and model all three types of cell–cell interactions with micrometer-scale precision. Even though we have a large number of available methods to pattern biological molecules (such as proteins) on solid surfaces, two characteristics of the interactions between mammalian cells and solid surfaces make it difficult to pattern multiple types of cells: 1) Most adherent mammalian cells, when cultured in vitro, randomly move and scatter around; and 2) newly seeded cells tend to squeeze into even the most tightly formed monolayers of cells and adhere to the surface, resulting in nonhomogeneous populations of cells in patterns. These features make patterning of multiple types of cells, in other words, the ability to control the adhesion and migration of mammalian cells to allow all three types of cell– cell interactions, still quite difficult in comparison to the patterning of multiple types of proteins. Several research groups have reported techniques for patterning multiple types of cells; some of these methods allow the confinement of two or more types of cells to specific locations on surfaces (to simulate type I), 6] as well as the movement of these different types of cells toward each other (to simulate type III). 7] To our knowledge, however, no published work demonstrates the simulation of type II cell– cell interactions though it is an important type of cell–cell interaction often encountered in normal and pathological physiology. Herein, we present a new method that achieves all three types of cell–cell interactions: “no release” (the confinement and segregation of multiple types of cells, type I), “partial release” (the selective release of one cell type for free motility while keeping the other cell type confined, type II), and “complete release” (the complete release of all types of cells, type III) of multiple types of cells (Figure 1). Our method uses an alkanethiol [(HS(CH2)11(OCH2CH2)6OH, abbreviated as EG6] [9] to generate self-assembled monolayers (SAMs) that resist cell adhesion (“inert” areas), and the extracellular matrix protein fibronectin (FN) physically adsorbed from solution to generate surfaces that promote cell adhesion (“permissive” areas); we employ microchannels both to create areas with different surface properties (either permissive or inert) and to transport different types of cells to designated locations on the same surface. Our strategy is dramatically simpler than previously reported methods. We strongly believe that this strategy will find wide application in fundamental research in cell biology. We first demonstrate the approach to achieving “partial release” of one type (out of two types) of cells (type II, Figure 1). We used a poly(dimethylsiloxane) (PDMS) stamp with embedded microfluidic channels for selective modification of the surface of the gold-coated glass substrate. The PDMS stamp reversibly sealed with the gold substrate through physical contact. To vary the distances between different types of cells on the same substrate, we made one of the microchannels long and serpentine. When we modified the gold-coated substrate with FN and EG6 in PDMS microchannels, we noted that EG6 could diffuse under the PDMS stamp to form inert SAMs, but FN could not. (We investigated the dynamics of this migration; see Figure S1 in [*] Dr. Z. Chen, Y. Li, W. Liu, D. Zhang, Y. Zhao, B. Yuan, Prof. X. Jiang CAS Key Lab for Biological Effects of Nanomaterials and Nanosafety, National Center for NanoScience and Technology 11 Beiyitiao, ZhongGuanCun, Beijing 100190 (China) Fax: (+ 86)10-6265-6765 E-mail: xingyujiang@nanoctr.cn