All three could push the true total higher, but number two easily could push it lower instead. And I'm not a microbiologist but I think lack of antibodies only means that we won't be able to test for past infection. The seronegative patients (determined by a very precise method for measuring antibodies) in the second cited paper were still able to clear their infections.
Additionally, does your third point also indicate that they will never be immune, or just that we won't be able to test for it?