Paired test?


I have the following problem. In my experimental setup I have four tissue samples from the same animal in each experiment, two from one side and two from the other. The original plan is to apply a substance alone on a sample from one side and the substance mixed with the inhibitor on another sample from the same side, to see if the inhibitor is able to block the effect of the substance. The plan would be to perform a t-test, where the tissues from the same side are paired together. However, because the tissue preparation is a fine process, it happens in some cases that one of the tissues is not alive and I cannot perform an experiment on that. So I have one sample with the substance and two with substance + inhibitor. Previously it has been shown that there are no differences in the effect between the two sides, so it should actually not matter. The variation depends on the animal. But when I obtain 3 numeric results from such an experiment, one with the substance and two with substance + inhibitor, there is no way to run a paired test for that. I can only use a two sample t-test.

Am I right?