Multiple Regression on Longitudinal AND Cross-sectional data


I'm trying to make sure I'm doing a hierarchical multiple regression analysis correctly (and interpret the results correctly) with regards to the predictor variables.

> I have six predictor variables (A, B, C, D, E, F) and one outcome variable (X).
> Three predictor variables were measured at Time 1 (A, B, C) and three were measured at Time 2 (D, E, F).
> The outcome variable (X) was measured at Time 2 also.

Should I be:

1. Entering T1 predictors into as a seperate block and then entering T2 predictors as the next block in the model?
2. Should the entering of the predictor variables be based purely on previous research and theory? And if so can a predictor from Time 1 be entered into a block with a predictor from Time 2?
3. If say predictor A (T1) and F (T2) explain a significant amount of the variance (say 20% each) is it ok to say they both predict X equally? Even though A longitudinally predicts X while F cross-sectionally predicts X?

I guess I'm a bit confused by the idea that I'm potentially going to be saying that a predictor variable measured at the same time as the outcome variable 'predicts' - or perhaps I'm getting 'prediction' and 'causation' confused here?

Thank you for any help.