Biomedical Engineering Reference
In-Depth Information
created a gold micropattern by shadow-evaporation onto a poly(ethylene terephthalate) (PET)
sheet, and a SH-PVA micropattern formed only on the gold areas. Ater physisorbing ECM pro-
tein onto the PET background, bovine endothelial cells seeded in 15% serum were observed to
attach and spread for at least 2 days on the PET areas only.
2.6.2 Micropatterning Cell-Substrate Adhesiveness
We have seen the previous methods to immobilize proteins, polymers, and SAMs, including
techniques to prevent the physisorption of proteins and to deter cell attachment. Here, we cover
the most prominent methods for designing which areas of a substrate the cells will adhere to. We
note again that glass and most polymers are covered immediately by a physisorbed protein layer
when placed inside a protein-containing biological luid. Combining the fact that it is surpris-
ingly diicult to completely remove this layer and that cell anchorage is exquisitely sensitive to
trace amounts of ECM protein, several researchers have been able to produce cellular patterns
on microfabricated templates of physisorbed (or dried) protein or by focal delivery of the cell
suspensions (using microluidic devices), avoiding chemical immobilization methods altogether.
here are now hundreds of published cell patterning techniques, of which we cover the most
successful and historically relevant. Some of these methods rely on sophisticated chemical sur-
face modiication, some require microluidic devices, yet others only involve a simple stencil—
but not all deliver the same results, resolution, or throughput. he researcher's needs should
determine which technique is best suitable for his or her application.
2.6.2.1 Physical Masking of Background with Photoresist
In 1998, Mehmet Toner's group at Harvard Medical School introduced a method to micropat-
tern cocultures of two cell types that is now out of vogue (thanks to sot lithography) but it is
explained in detail here because it still has a lot of pedagogical value. he process, outlined in
Figure 2.33a
, takes advantage of diferences in adhesiveness between each cell type and consists
of chemisorbing a collagen pattern and then illing the background of the pattern by physi-
sorbing albumin (which does not support cell attachment)—the hepatocytes' integrin recep-
tors recognize RGD-like sequences on the collagen (and not on the albumin areas), so selective
attachment of the hepatocytes on the collagen occurs.
What is surprising is that it works because the collagen is patterned using a traditional lit-
of process whereby the photoresist is removed with acetone, a strong solvent that is a strong
denaturant. In other words, the cells do not really care, for the purposes of cell
attachment
, that
collagen has been denatured. It makes sense: the cells do not really attach to the whole colla-
gen molecule—only to a small sequence of peptides that has not been altered. We note that the
chemisorption reaction chosen was an aminosilane SAM linkage followed by a glutaraldehyde
cross-linker, which presumably chemically immobilizes the protein. As it turns out, if one skips
the aminosilane and glutaraldehyde steps, collagen patterns are also formed (collagen physi-
sorbs so strongly to the surface that acetone denaturation is not enough to remove it from there)
and, most important, the hepatocytes seem to be equally functional. his
also
makes sense—the
cells only see the protein, not the surface under it. his whole process teaches us that there are
more practical strategies; namely, ones based on physisorption and direct delivery of proteins,
to achieve selective cell attachment.
How about the ibroblasts? hey attach to the albumin areas because they are seeded in a serum-
containing medium that contains a lot of other proteins that promote cell adhesion, even though
albumin does not. Nobody knows if these proteins are adhering on top of the albumin, or the albu-
min desorbs, or the ibroblasts help degrade the albumin irst, or all of the above. (Interestingly, the
ibroblasts that settle on top of hepatocytes display a remarkable capacity to “ind” the substrate
even when it seems that the hepatocytes are occupying all of it—either the ibroblasts win some
ierce battle or the hepatocytes were not really occupying it: in a few hours, if properly labeled,
ibroblasts can also be seen tightly packed in between the hepatocytes, although in low numbers.)
