Gentlemen, there is a reason why most of Iberia and North Italy and even a bit of Tuscany plot near one another on virtually every PCA I've ever seen.
It's because they share a lot of ancestry in common, most of it dating to the Neolithic and Chalcolithic.
Differing gene flow impacts on Southern Italy/Sicily pull it apart from Spain,but that underlying similar EN/MN ancestry is still shared between the two.
Some commercial tests pick it up and some don't. Perhaps 23andme doesn't show it because they have a lot more Italian samples, and so the alleles in a particular person find a match in other Italians quite easily. Perhaps some other companies include "Northern Italy" in the "Italian" cluster? I don't know. Never having taken any of these tests I lose track of the definition of their clusters.
That doesn't mean that there weren't continuing contacts and admixture to some degree. However, I think whatever there was was much more prevalent in Northwest Italy because of the Beakers and continuing admixture along the north Mediterranean rim from Liguria/Tuscany all the way to Valencia. That's why a lot of Northwest Italians, including me, get chunks of "Iberian" even in 23andme, whereas Southern Italians get almost none, but get quite a bit of Balkan.
If one were to follow Graham and Coop, there's been no significant admixture between Spaniards and Italians for thousands of years. However, I don't think the Popres data they used included many Piemontese, Ligurians, or Northwest Tuscans.
http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1001555
Some soldiers from Aragon are not the kind of contact that substantially changes the genome of an area, in my opinion. There has to be more than that.