Glial cells can grow as multilayers upside-down

In my 1st post detailing glial cells on pnc-Si transwells, I saw lots of cell clumps above the membrane window (see here).  Since these cells were grown in the wells, I thought they could be piling on top of each other in the wells and getting trapped.  Here, I decided to flip the sample so that the glial cells grew in the wells, but upside down to see if this cell clumping still occurred.  I suspected that cells would fall to the bottom of the 24-well plate instead of forming monolayers.

Here, NG108 cells (P13) were seeded at 50,000 cells/cm2 on the well-side of pnc-Si and then inverted so that they were upside-down in the 24-well plate for growth.  TEER was measured and then the cells were stained with Live/Dead on day 7.

PET transwells showed a largely viable, nearly confluent cell layer with some multi-layer clumps.  Even though these cells are hanging upside-down, their intercellular adhesion is strong enough to maintain multiple layers away from the membrane surface.

TEER:

TEER of control samples (PET) didn’t change.  The pnc-Si results were confusing.  Most of the samples stayed right around their day 0 values (that is, there was no increase in barrier function).  However, a couple of samples showed ~50% increase in TEER over 1 week.  Since there was a fairly confluent monolayer of cells on PET (see above) that didn’t cause an increase in TEER, I didn’t know what to make of this.  So I looked at the images:

Sample 1 (same sample as in the graph) with the phase contrast channel shown in blue.  There is clearly a huge clump of cells away from the membrane, but the cells don’t cover the entire free-standing membrane area.

Sample 2 (same as in graph) is below.  There are so many cells that the entire image is blurry.  It appears that there is a layer several cells deep that covers the free-standing pnc-Si area.

These drastically different morphologies likely explain the difference in TEER values for these 2 samples.  The other samples (4, 5 and 7) exhibited similar coverage differences.

How are these cells clumping together when they are growing upside down?  Here’s an image to show my working model:

On the left is an image with the focal plane at the supported membrane and on the left is an image with the focal plane at the free-standing membrane.  On the left image, cells on the supported membrane and the interface between the well walls and the supported pnc-Si area are both in focus.  Clearly, there are cells growing on the sloped well walls.  It looks like cells growing on supported pnc-Si form an almost continuous layer with cells growing on the well walls and then down onto the free-standing membrane.  Now, this image isn’t perfect since there are very few cells on the upper-right area of supported pnc-Si.  BUT, if the cells on either side of the wells (on the supported pnc-Si) form a somewhat continuous layer with cells on the well walls and the free-standing membrane, then dividing cells can stay attached to this cell layer within the wells – thus forming clumps.

This phenomenon would certainly be time- and ‘initial cell seeding density’ – dependent.  I’m not sure why there are differences in clumping in this experiment.  It could be differences in handling since these cells are not very adherent.  I need to minimize this clustering in order to minimize the TEER of the glial cells.