How can I be more accurate in my cell counting?

Question

For some reason, when I use my hemocytometer, my counts and calculations always come out really small. My equation is: (cell count / # of boxes) * dilution factor * 10,000 = cells/ml. (I usually only count the four corner boxes for my fibroblast cultures).

My calculations usually are less than half of what I expect them to be from the microscope estimations and confluency. I have tried various dilutions and have even vortexed the cell mixtures to ensure even distribution. Are there any suggestions as to what I'm possibly doing wrong?

EDIT: I usually use a dilution factor of 10 and then multiply the end number by the original number of mLs that the sample was taken from.

Answer 1

Without further information it is quite hard to say. However, there are a few things that might be playing into this.

1. Statistics - For your counts you need to be counting a decent number of cells on each side (and each square) of the hemocytometer. Counting 4 squares is fine, however you need to be counting at least 100 cells over that space (see step 3 in link from Bitesize Bio). This doesn't mean stop counting if you reach 100, finish the square you are in at least and it means 20+ cells per large square. If you are counting fewer cells, your counts will be inaccurate.

2. Statistics - hemocytometers have about an error of about +/-15% in the hands of an experienced user.

3. Statistics - Low cell density; this plays into the one above. If you resuspend your cells in too large a volume, it is much harder to take a representative sample of your cells, this means you can't rely on the sample to accurately represent your cell population. This also goes for your subsample and dilution - there's no reason you can't do a 1:1 dilution with trypan blue rather than a 1 in 10.

4. Settling - cells settle quite fast in all the tubes you use. You must resuspend the cells immediately prior to sampling and immediately prior to loading the hemocytometer. I've documented an example of how fast they settle in my own counting (out of interest, decided to see if there was an effect): If I take up 100 ul of cells in a micropipette (say P-200) and load both sides of a hemocytometer from that same volume (takes me less than 20 s for both sides), there is a consistent ~20% difference between the sides due to settling of the cells in the tip. If I take separate samples, that error goes away.

5. Cell type - Fibroblasts are large. This means you don't get as many per area as you might see in examples of cell density (e.g. Invitrogen's guide to Useful Numbers for Cell Culture), which plays into 1 and 3 above.

6. Cell type - Fibroblasts attach quite firmly. This means you need to visually inspect the cells to ensure detachment and look to see that you have successfully harvested the majority (there are always a few left behind). Don't use too large a volume for this as it plays into 1 and 3 again.

7. Cell handling - if you over detached your cells (i.e trypsin for too long) or scraped your cells off instead of using trypsin (or similar), then the cells will clump. It is very hard to count clumps of cells accurately and even harder to sample them accurately - they settle much faster than single cells.

8. Cell handling - Making a single cell suspension. You successfully detached your cells, but they have clumped and you didn't make a single cell suspension. As in 7 above, these are very hard to count. You must ensure that the majority of the cells you count are single; you should have <5% clumps.

9. Incorrect usage - you are counting using the large squares right? That's the ones sub-divided into 16 or 25 smaller squares (see red squares in image by step 3 in Bitesize Bio link) and often surrounded by 2-3 lines (Neubauer hemocytometer). If you were using the small subdivisions, you will be out by a factor of 16, but sampling errors will play a big role in how accurate the counts are.