Mean and variance estimation in high-dimensional heteroscedastic models with non-convex penalties