Clinical Translation of a Prognostic Follicular Lymphoma Signaling Profile

Phospho-flow cytometry signaling profiles have great potential to identify and monitor negative prognostic subpopulations of cells within tumors. We recently reported that altered B cell receptor (BCR) signaling identifies a subset of lymphoma negative prognostic (LNP) cells within follicular lymphoma (FL) tumors from patients with shorter overall survival (Irish et al., PNAS 2010). Each 1% increase in LNP cells observed within the tumor at diagnosis increased the patient's annual risk of death by 2.5%. However, before signaling profiles such as this can be used in the clinic, the assay must be simplified and translated from the complex and time-consuming version used in research studies. Here we report a prognostic signaling profile suitable for follicular lymphoma clinical specimens of less than one million cells.

To simplify LNP cell counting in FL patient specimens, detection of six key phospho-protein readouts for BCR signaling was combined on one cytometer channel, creating a single readout for BCR network signaling activity (termed p-BCR network). This combined stain measured phosphorylation of SYK, PLCγ, BLNK, Src family kinases (SFK), STAT5, and ERK using Alexa 488 dye conjugates of each phospho-specific antibody. LNP cells are defined by a lack of BCR signaling response at all of these phospho-protein readouts.

Upon receipt of a patient's tumor sample, 0.25 million cells were stimulated for 4 minutes by polyclonal F(ab')₂ against IgM and IgG (10 μ g/mL each; α-BCR), left unstimulated as a negative control, or exposed to a positive control stimulus (α-BCR + H₂O₂). Samples were stained with a four-color flow cytometry panel detecting p-BCR network, lymphoma B cell markers BCL2 and CD20, and CD3 or CD5 as a marker for T cells. Approximately 50,000 lymphoma B cells were collected and LNP cells quantified as before. The p-BCR network assay typically required 2h to complete. In parallel, LNP cells were quantified in samples from the same specimens using the original method, where phospho-proteins were detected individually using a set of staining panels. An additional control staining panel measured total tyrosine phosphorylation (p-Y) in place of p-BCR network staining in the four-color panel.

LNP cell counts measured by combined p-BCR network stain using 0.75 million FL tumor cells total were equivalent to those obtained by individual measurement of phospho-proteins (R² = 0.98). Critically, in the combined p-BCR network stain the same patterns of signaling in lymphoma cell subpopulations were observed as when individual phospho-proteins were measured. Thus, not only was the same quantity of LNP cells detected by the p-BCR network stain, the same subpopulations of cells were identified by the two techniques. Total phospho-tyrosine following BCR stimulation was examined as a potential surrogate for p-BCR network staining, but high basal p-Y signaling and additional BCR mediated p-Y signaling obscured the LNP cell subset in some patient specimens. Thus, a total p-Y readout could not be used in place of the p-BCR network readout or detection of individual phospho-protein readouts.

A clinical prognostic follicular lymphoma signaling profile can be measured in small specimens of live tumor cells, such as fine needle aspirates, and collected using commonly available flow cytometers. These results create new opportunities to use signaling profiles to identify FL patients who might benefit from clinical trials and to monitor emergence of negative prognostic lymphoma cells over time following therapy.

No relevant conflicts of interest to declare.