You can prove it by induction on $n$.
The case $n=1$ follows by the $\aleph_0$-saturation of A. For the inductive case, given an $(n+1)$-type $\Gamma(x_1,\ldots,x_{n+1})$, you can write it as the union of the $n$-type given by all formulas in $\Gamma$ using only variables $x_1,\ldots,x_{n}$ and the rest of the formulas in $\Gamma$ (which are the formulas using also the variable $x_{n+1}$).
By induction, there is a realization $c_1,…,c_n$ in $A$ of the $n$-type, and by hypothesis, the new type $\pi(x_{n+1}):=\Gamma(c_1,\ldots,c_n,x_{n+1})$ is a $1$-type consistent with $Th(A_{c_1,\ldots,c_n})$, so there is $d\in A$ realizing $\pi$.
Notice that the tuple $(c_1,\ldots,c_n,d)$ is a realization in $A$ of the original $(n+1)$-type $\Gamma(x_1,\ldots,x_{n+1})$.