The definition of theta says it all
$c_1 g(n) \leq f(n) \leq c_2 g(n)$
if f(n)=$\theta(g(n))$ for some constants $c_1$ and $c_2$ and for all n$\geq n_o$
Your g(n) must be a function such that by means of some constants c1 and c2, you are able to satisfy both the lower bounds and Upper bounds (i.e.
$c_1g(n)$ for big omega and $c_2g(n)$ for big Oh.)
and if you are able to do so with the help of a function called g(n), then your function f(n) is in $\theta(g(n))$ means in simple language you have f(n)=g(n).
So in your example $n^2 + O(n^2)$ the $O(n^2)$ means this term has upper bound of $n^2$ and this term may be a constant, or a linear function of n or a quadratic function of n but not more than this.
Since the term $n^2$ is already there along with big oh, it must be at least $n^2$ in terms of growth of function and it cannot exceed $n^2$ in terms of growth of this function.