Human Cell Viability Assay?
Michael J Conboy
conboymj at leland.Stanford.EDU
Thu Sep 30 02:39:57 EST 1993
Regarding the MTT assay, I have used this to determine the
growth rate of cultured myoblasts and have found the assay to be quite
convenient and reproducible. What kind of error are you experiencing
and where? Is it variability within duplicate wells of the same
sample, day to day variability in each time you reproduce the assay,
or a discrepancy in the results obtained in the MTT assay versus another
assay, such as counting or thymidine incorporation? How big is the
error and how precise do you have to be?
I must admit that I had some trouble with this assay at first.
I had occasional pipetting error using a multichannel pipettor, which
I solved by using a different brand of better fitting tips and watching
how much liquid was drawn into the tips EACH time. My pipetting error
is less than 5%. Feeding the cells daily introduced error until I changed
only half the media each time, and then very gently. I think I was
dislodging cells from the bottom of the wells; your cells may be
different; for my cells it is better not to feed them at all. The
development time with the MTT substrate can affect plate to plate
reproducibility. I developed for 4 hours, then switched to 3 as there
seemed to be less crystal formation, but 4 and 3 will give different
results. Pick a time and do all assays the same. Lastly I found a big
variability in absorbance readings depending on how long I waited after
I added the solubilization solution before reading the plate. My
protocol said "1 hour to overnight", but the reading will increase 0.2 A570
or more during this time. I could not correct for this with null wells,
so it is best to use consistently a set time. I have no problems mixing
with a multichannel pipet.
I find the accuracy of the assay to be the same as direcly
counting the cells, BUT only during the "window" of growth where the
absorbance readings are accurate (approximately 0.1 to 1.0 corrected
A570). Direct counting is more accurate for cell densities below the
detection of the MTT assay, say <5% confluent. I have no experience
with thymidine incorporation assays.
Also note that the MTT assay measures mitochondrial activity,
not cell number directly. Anything which increases this activity without
actually increasing cell number will give error. In my case as myoblasts
differentiate, they increase the number of active mitochondria. Thus
I had to combine MTT with visual observation of the cells to make sure
they were not differentiating. Other variables, temperature, media
age/composition will also affect mitochondrial activity and must be
held constant/consistant. For instance, in a stack of 96 well plates,
the ones in the middle will take the longest to equilibrate to a
temperature. After the 4 hour MTT development, they will show low readings.
I'm not through using this assay, thus all comments and improvements
More information about the Cellbiol