So, every time you calibrate and profile your display do you then tweak the values as suggested by WHCC? If not, why was your methodology working previously (i.e., not needing you to tweak your display to match the test prints)?
Am I missing something? It sounds like you have a standard target to which you calibrate and then profile your display. It sounds like this has always worked for you and your WHCC prints, without ever having to tweak your display to get a match to the WHCC test prints. Now, suddenly, this methodology is not working - so the question is "why?" What is different from the previous successful prints?
WHCC's website suggests target calibration values for white point (6500°K) and gamma (2.2), with a suggestion that if your viewing environment lighting is warmer, to try a white point of 5000°K. Fine. They cannot suggest a luminance target because they know nothing about your working environment (80, 90, 100, 110 cd/m2?) and viewing light intensity.
Did you change the luminance target, even inadvertently, prior to the calibration you did before sending the images off that came back dark?
WHCC suggests AdobeRGB or sRGB as the file color space that you should submit - they stress that the file needs to have a color space tag embedded, otherwise their machines will assume sRGB. So, it sounds like as long as your AdobeRGB files had the color space tag embedded, that should not have caused a problem.
If the color in the prints is still accurate, then the "tweak" would be the luminance target. Why this suddenly changed for your workflow with WHCC, I have no idea.