How do I get marginal probability over a multidimensional array of discrete random variables?

Say I have a 3x3 array which represents a joint probability distribution, P(X1, X2), over discrete random variables.

% I'm calling P(X1, X2) 'dist'in matlab

dist = [15/45, 12/45, 1/45; 12/45, 4/45, 0; 1/45,0 ,0] % X1 is the top, X2 is the side

As an example, the probability P(1,3) is the probability that under the distribution X1=1, and X2=3. Now, to find the marginal probability of P(X1 = 1), you simply sum down the column to get it.

P(X1 = 1) = 15/45 + 12/45 + 1/45 = 28/45

Easy enough. And if I want to find the marginal distribution vector for X1, I can use matlab's sum(dist).

My problem is, I have a 4-D double array representing a joint distribution over discrete random variables -- P(X1,X2,X3,X4). How do I find the marginal distribution vector of say P(X1)? I know it's a similar process as the bi-variate example, but is there a simple way to do this in matlab without having to write a whole bunch of for loops?