PBMC Isolation vs Whole Blood Lysis for Flow Cytometry Phenotyping

Your collaborator at the clinical site stains lysed whole blood. Your lab isolates PBMCs on a density gradient. For the same donors, the two datasets disagree on the CD4 percentage, and the reviewer wants to know which one is right. Both can be, because PBMC isolation and whole blood lysis start flow cytometry phenotyping from different cell populations with different denominators. This post compares the two preps: what each keeps, how each looks on FSC versus SSC, which frequencies shift, and which to pick for a given assay.

What each prep keeps and removes

Whole blood lysis destroys red cells and keeps every leukocyte: lymphocytes, monocytes, neutrophils, eosinophils and basophils. Ammonium chloride lysis does not fix the cells. Some commercial lysing solutions contain fixative, so check the formulation before planning any later step that needs live cells. Lyse/no-wash keeps every cell in the tube. Lyse/wash removes debris and unbound antibody but loses some cells at each spin.

PBMC isolation uses density-gradient centrifugation (Ficoll-type media or cell preparation tubes). Mononuclear cells, meaning lymphocytes and monocytes, band at the interface, while granulocytes and red cells pellet. Every isolation loses some cells, and the losses are not proportional across subsets. You cannot calculate cells per microlitre of blood from a PBMC prep.

What the preps look like on FSC vs SSC

The scatter plot is the fastest way to see which prep you are looking at, and the first place they diverge in analysis.

• Lysed whole blood shows three clouds. Lymphocytes sit low on FSC and SSC, monocytes higher on FSC with intermediate SSC, and granulocytes at high SSC. In a healthy adult, neutrophils make up the majority of leukocytes, so the granulocyte cloud dominates the plot.
• PBMCs show lymphocytes and monocytes, with the high-SSC granulocyte region close to empty. Events there are a QC signal, not a population to ignore.

Granulocytes in a PBMC prep usually point to processing delay. In McKenna et al. (J Immunol Methods, 2009), granulocyte contamination of PBMCs rose 2.3-fold within 6–8 hours of room-temperature storage and 11.3-fold by 24–26 hours. Refrigerating the blood for 22–26 hours made it worse, at 84-fold. The contaminating cells were activated CD11b+CD15+ granulocytes, and they inhibited T cell function.

A 2024 CyTOF study of seven donors found that two or more days of delay produced extensive red cell contamination, and four days reduced monocyte frequencies. The authors supported the existing recommendation to process within 36 hours (Higdon et al., Clin Exp Immunol, 2024).

How the two preps shift subset frequencies

The clearest recent head-to-head is Tompa et al. (Biomedicines, 2024). The authors ran the same multicolour panels on whole blood, fresh PBMCs and cryopreserved PBMCs from 10 healthy donors. The differences were sharpest against cryopreserved PBMCs:

• against both PBMC preps: lower CD4+ percentages and lower perforin in CD8+ T cells
• against cryopreserved PBMCs: higher CD8+, Th1, Th17, Th1Th17, CD25+CD127^low Treg and B cell fractions, and fewer NK cells
• against fresh PBMCs: a higher CD8+ proportion, with weaker trends in the same direction for the other subsets
• very low FoxP3 intensity in whole blood with the same antibody lot, which the authors suspect reflects fixation and permeabilisation volumes not adjusted for whole blood

The authors also note that whole blood carries other cells, proteins and lipids that can affect staining, including autofluorescence from haemoglobin porphyrins. The FoxP3 result matters if you gate Tregs by intracellular staining. The CD25, CD127 and FoxP3 Treg gating strategy assumes a FoxP3 signal you may not get in whole blood unless the fix/perm step is re-optimised for it.

The denominator matters too. "Percent of leukocytes" in lysed blood includes granulocytes; "percent of PBMC" does not. Even "percent of lymphocytes" is comparable only if the upstream gates were drawn the same way.

Whole blood lysis vs PBMC isolation, side by side

Which prep to use, by assay

• Absolute counts or clinical monitoring (CD4 counts, lymphocyte subset counts): use lysed whole blood with lyse/no-wash and beads. The bead arithmetic is in absolute cell counts with counting beads.
• Neutrophil or eosinophil work: use whole blood. A gradient removes the cells you want.
• Multi-timepoint cohorts and functional assays (intracellular cytokine staining, proliferation): use PBMCs, processed and frozen the same way at every timepoint, so all samples can be thawed and stained together.
• Rare populations: PBMCs let you load more mononuclear cells per tube. Run the event-count arithmetic before choosing.
• Tandem-heavy panels on whole blood: keep stain-to-acquisition time short. For a critical marker, consider a non-tandem fluorochrome. The fluorophore spectrum viewer shows which detectors a replacement dye would spill into.

In Cytomaton, save one gating template per prep type rather than one shared template. Scatter gates for lysed whole blood and PBMCs should not come from the same file, and a template applied across the wrong prep looks correct until someone checks the granulocyte region.