That paper looks at 40 people and has no idea what the clinical correlates of the non-spike responses are because no one has any idea. You cannot take that paper and say that a substantial factor of people are protected due to preexisting responses.