Let us consider the projective line $\mathbb{P}^1$ over a field $k$ and take the following open embeddings $$j_{\epsilon}\colon \mathbb{P}^1\setminus\{0,\infty\} \to \mathbb{P}^1\setminus\{\epsilon\}$$ for $\epsilon \in \{0,\infty\}$. Since $\mathbb{P}^1\setminus\{0,\infty\}$ is isomorphic to $\mathbb{A^1}\setminus \{0\}$ we can fix a closed embedding $$i\colon \mathbb{P}^1\setminus\{0, \infty\}\to \mathbb{A}^2$$ and consider the induced maps: $$(i,j_{\epsilon})\colon \mathbb{P}^1\setminus\{0,\infty\}\to\mathbb{P}^1\setminus\{\epsilon\}\times \mathbb{A}^2.$$ In ``$\mathbb{A}^1$-homotopy theory of schemes'' by Morel and Voevodsky I've read that the pushout of the diagram $$\mathbb{P}^1\setminus\{0\}\times \mathbb{A}^2\leftarrow\mathbb{P}^1\setminus\{0,\infty\}\to\mathbb{P}^1\setminus\{\infty\}\times \mathbb{A}^2$$ in the category of sheaves over the smooth Nisnevich site is represented by an affine scheme but I'm not quite sure why is that true.
I know that the category of affine schemes has pushouts but these are in general not preserved in the category of schemes (see for example this article by Karl Schwede). Even more, Yoneda embedding usually destroys colimits even though some can be recovered after sheafification, like for a cartesian and cocartesian square of open embeddings.
Therefore, a more general question would be: is there some partial result of the form "take a pushout square in the category of affine scheme satisfying some properties, then it stays a pushout square in the category of Nisnevich sheaves" that contains the previous example as a particular case?
Thank you very much for any possible help.
