There's not a whole lot of physics here. For given $\ell$, the radial part of the Schrödinger equation is equivalent to that of a one-dimensional problem, and for such a problem it is known from the properties of the solutions to this type of differential equation that solutions with higher energies will have a higher number of nodes (irrespective of the potential, provided it is confining). The argument is qualitatively that, with increasing energy, you need to "fit" an increasing number of wavelength in the confining region so the solutions must have increasing curvatures and thus increasing number of nodes.
It is only for the hydrogen atom that the number of nodes is given by $n-\ell-1$. For the 3d harmonic oscillator, the number of nodes $n_r=\frac12(n-\ell)$, where $n=0,1,2\ldots$ is related to the energy by $E_n=(n+\frac32)\hbar\omega$ and $\ell=n,n-2,n-4,\ldots$ so that $n_r\ge 0$. For the infinite spherical well, the situation is even more obscure.