I would either stain for CD45 to gate on leukocytes or do like a 40% percoll gradient.  

For percoll gradient, we resuspend liver single cell suspensions in pre-warmed media with percoll (5 mL percoll, 8 mL 2% FBS-RPMI, 1:100 heparin).  Centrifuge at 2000 rpm for 20 min at room temperature.  Remove supernatant (which should contain band of hepatocytes) and lyse red blood cells in bottom pellet.

We resuspend the pre-digested tissue in 25mL of 33% Percoll then centrifuge at 690g for 12min at room temp. using minimal acceleration and deceleration. (Both centrifuge and Percoll should be at RT).

After spin remove the layer of hepatocytes at the top of the tube and all of the Percoll leaving only the pellet behind.

Any other depletion/enrichment steps are done after this point.

Theresa - I agree with Cory and Jo regarding a density gradient as the first pass.  I would expand a bit on Cory's mention of performing a chemical lysis of the RBCs that co-pellet with the CD45's you're looking for.  We've increasingly been hearing back from folks working on dispersed liver preps about concerns regarding chemical lysis having a pretty profound effect on the CD45+ cells in that pellet.  In particular, one issue has been that ammonium chloride 'lysis' disrupts cells but leaves large sheets of RBC membrane in the sample.  These membrane sheets can engage CD45+ cells via scavenger or annexin-V receptors and lead to unanticipated activation.

You might consider using a particle-based clean-up targeting the murine erythroid antigen Ter-119 (believed to be a glycophorin-like surface protein).  We are currently performing pilots with folks needing to remove Ter-119 (as well as CD235a in analogous human applications).  If you'd like to discuss this more, I'm happy to think through your application.

Best,  John

www.akadeum.com