I am a researcher studying diseases. You have no idea how many mice get killed without generating any data. There's a rule in place whenever you want to work with animals that you need to plan ahead and only use as few animals as you need to get the data that you're looking for. But things in research basically never happen according to plan. It could be due to a variety of factors: unexpected failures, overlooked factors, technical errors, or just simple negligence when performing an experiment. A lot of data and samples obtained from killed mice are discarded for one or more of the above reasons.
I get that mouse experiments are important to prove that our findings can translate to actual living animals, but I personally will not touch a mouse because, frankly, the "useful data per mouse" ratio is way too low for me to justify using mice.
I was in the field for years. A lot of the mice we had were maintained with one copy of the gene of interest and crossbred to produce experimental litters (there are a lot of reasons for it, some technical, some practical). But OMG the numbers of mice we went through just to maintain the lines. Forget about failed experiments etc.
While you didn't get the data you were looking for, at least in many of those cases you mentioned you did identify a flaw or failure and learned how to design an experiment that does.
I wouldn't consider those mice as dieing without teaching you something. It might be a failed experiment, but you learned something.
Like when we were doing "oral" vaccinations with a oral gavage needle (ball tip) and going through the mouth and dosing in the stomach. We had a vial of 70% alcohol to clean the tip. Accidentally drew the alcohol up instead of the vaccine. By the time we finished the cage (6 mice, I think) the first one fell over.
Good question. You may be surprised to hear that my stance isn't that uncommon in research. If I recall correctly, somewhere around 50% of researchers personally will not use mice in their experiments. In these cases, we would either use a lower lifeform (fish or fruit flies), or use immortalized cells. Immortalized cells are aggressive cancer cells that happen to retain some of their cell properties. For instance, immortalized lung cells tend to act somewhat like actual lung cells. It's not a perfect model, since you're experimenting on cancer cells instead of actual cells, but the ease and low cost of growing and using them makes them extremely valuable for a lot of grindwork experiments, where you just need to burn through tons of different hypotheses quickly.
For me, I prefer to use immortalized cells. It works out for me anyways, since I prefer to focus on the mechanism of disease (which tends to be easier on immortalized cells) rather than practical effects of disease (which tends to require animals).