Due to the negatively charged phosphoryl and carboxl groups in endotoxins, anion exchange chromatography was the common depyrogenation method for proteins
https://www.ncbi.nlm.nih.gov/pubmed/18996203
However the success rate of this approach depends from the pI of your protein (because if it as low pI can bind the coloumn with the LPS and the separation is not efficient) and from the stability of your protein in the pH condition required for optimal binding of the endotoxin to the coloumn.
An interesting alternative in lab scale, that i would like to suggest to you are the Endotrap coloumn (http://www.hyglos.de/en/products-services/products/endotoxin-removal/endotrapr-hd.html) that are based on affinity.
There are several kind of this coloumn that allow to you to operate with common buffers used for protein storage and purification and in my experience work very well.
3. Size based : If your desired protein size less then 5kda then you can try TFF (10kda) in flow-through (permeate mode), Endotoxin have higher size so endotoxin will retain in retenate and your product will come in flow-through.