Why is PCA sometimes used as a pre-processing step before regression?

A) To reduce dimensionality and multicollinearity.
B) To increase the number of features.
C) To introduce noise into the dataset.
D) To eliminate outliers.