Next generation assays will need to be robust and standardized in order to make the transition from a research procedure to a routine clinical assay. Flow cytometry provides a unique and sensitive method to accomplish these requirements.
Signal Transduction pathways are essential elements in communications from the cell’s environment, and play a central role in regulating cell proliferation, differentiation, metabolism ,and death. Signaling pathways are frequently altered in human leukemias, and likely play important roles in abnormal growth and death patterns associated with disease.
Acute Myeloid Leukemia (AML) is a common form of leukemia. Biologically, AMLs are heterogeneous, and finding the optimal treatment for patients represents a clinical challenge.
Figure 1. Signal transduction pathways commonly implicated in adult AML include ERK MAP kinase, PI3 kinase, and JAK/STAT pathways. Both ERK and PI3/Akt pathways can contribute to activation of ribosomal S6 protein, a useful monitor of cellular protein translation. Simultaneous measurement of four signaling proteins is accomplished using antibodies conjugated with the indicated fluorophores. Activation or inhibition of specific pathways is accomplished by the in vitro addition of specific agonists (e.g. SCF, Flt3 ligand, or GM-CSF), or in the presence of pathway specific inhibitors (e.g. Ly294002, rapamycin, UO-126, or Sorafenib), or using combinations of agonists and inhibitors. |
AMLs demonstrate a variety of genetic abnormalities1, frequently involving changes in receptor tyrosine kinases and their downstream signaling pathways.2 Changes in a number of signaling pathways have been implicated in AML, including PI3 kinase, ERK MAP kinase, and Jak/STAT pathways. As shown in Figure 1, we have developed and validated a panel of four different monoclonal antibody-conjugates that can be used simultaneously to monitor signaling through these pathways using flow cytometry. With this approach, we are able to simultaneously monitor three signaling pathways (PI3 kinase, ERK, and STAT5) implicated as critical for the survival or proliferation of AML cells in vivo. In addition, monitoring phosphorylation of the ribosomal S6 protein can provide an indication of upstream signaling through the PI3K or ERK pathways. Used in conjunction with phenotypic markers to identify specific leukemic or normal cell populations, flow cytometry provides a sensitive method to identify signaling pathways in AML, and has been used to measure signaling pathways in leukemic patient samples.3-5 Highly sensitive and specific analytical methods are needed that can be used with clinical samples to identify specific drug targets expressed in individual patients, as well as allowing pharmacodynamic monitoring to determine if drug treatment results in inhibition of the signaling target and the appropriate downstream elements of the pathway. To impact on patient care, next generation assays will need to be robust and standardized to make the transition from a research procedure to a routine clinical assay. As we demonstrate here, flow cytometry provides a unique and sensitive method to accomplish these requirements.
Materials and methods
One hundred microliters of heparin anticoagulated peripheral blood was placed into individual 12 × 75 cm polystyrene tubes. For inhibition of signaling pathways, the following agents were used (Figure 1): UO126 (100 µM, Cell Signaling Technologies), rapamycin (1 µg/ml, LC Laboratories), LY294002 (500 µM, LC Laboratories), or Sorafenib (1 mM, Toronto Research Chemicals) were added 30 min prior to stimulation or fixation. Stimulation of signaling pathways was performed by the addition of Stem Cell Factor (SCF: 100 ng/ml, R&D Systems), FLT3 ligand (25 ng/ml, ORF Genetics), or GM-CSF (25 ng/ml, BioLegend) for at least 5 min. All tubes were placed into a 37°C water bath immediately after the addition of inhibitors or activators. Cell fixation, RBC lysis, and WBC permeabilization were performed as previously described (Shankey, et al. Cytometry 2005). Whole blood was fixed by adding formaldehyde (EM grade, Polysciences, Inc) to a final concentration of 4%. After 10 min incubation at room temperature, TX-100 (Thermo Fisher Scientific) was added to a final concentration of 0.1%, and incubated at 37°C for 15 min.
After treatment with 80% methanol and washing, cell pellets were incubated with a cocktail of signaling antibodies, composed of Phospho-ERK1/2-Alexa 488 (P-ERK, Thr202/Tyr204, clone 20A, BD Biosciences), Phospho-STAT5-PE (P-STAT5, Tyr694, clone C71E5 (CST), PE conjugate from Custom Design Services, Beckman Coulter), Phospho-Akt-Alexa 647 (P-Akt/PKB, Ser473, clone 193H12, Beckman Coulter), Phospho-S6-Pacific Blue (P-S6 Ser235/236, clone D57.2.2E [CST], Pacific Blue conjugate from Custom Design Services, Beckman Coulter), plus one or more antibodies to identify the AML blast cell population (typically included CD34, CD117, or CD45, Beckman Coulter). Fixed cells were incubated with antibody cocktails for 30 minutes on ice protected from light, washed, and analyzed using a Gallios flow cytometer (standard 3-laser configuration, Beckman Coulter). A minimum of 3000 blast cells were acquired for each condition. Analysis of flow cytometry data was accomplished using the Kaluza software program (Beckman Coulter)
Results and discussion
More than 100 different AML patient’s have been analyzed using the approach described here. Some of these results have been published,6-8 and demonstrate heterogeneity in the signaling pathways that are activated in individual patient’s peripheral blood leukemic cells. These results demonstrate the ability of flow cytometry to monitor the impact of targeted therapeutic agents (in vitro or in vivo exposure), even when leukemic blast populations represent less than 1% of the cells present in the sample.
Figure 2. Flow cytometric analysis of a peripheral blood sample from a single AML patient containing 25% CD34+ events (blasts colored in red). The levels of four different signaling phospho-proteins are shown as correlated 2D histograms. in vitro treatment of individual samples (from top to bottom) included: no treatment (basal), Stem Cell Factor (SCF), SCF plus Ly294002, SCF plus Sorafenib, or GM-CSF. |
The results shown in Figure 2 demonstrate the analysis results of the four signaling pathways employed by peripheral blast (CD34+) cells from a single patient (here, 25% of peripheral blood cells were CD34+; colored red in in Figure 2). Constitutive or basal signaling levels from this individual (top row) are likely significantly above background levels for all four of the signaling pathways monitored here. Treatment of samples from this patient in vitro with the MEK inhibitor UO126 specifically decreased P-ERK and P-S6 levels, while treatment with the PI3 kinase inhibitor LY294002 specifically decreased P-Akt and P-S6 levels.
As previously reported,7 these results indicated that, in some AML patients, phosphorylation of the S6 protein can result from activation by upstream ERK and/or PI3 kinase pathways, or by pathway interactions illustrated in Figure 1. Addition of SCF in vitro results in increased signaling through ERK, Akt, and in this patient, increased signaling through STAT5 (Figure 2, second row); SCF stimulation also results in most of the AML blast (CD34+) cells expressing high levels of P-S6. Qualitatively, this increase is seen by comparing the boxes (B, C, and D, Figure 2) for the SCF stimulated sample with the identical boxes in the basal sample (boxes were set to include ~95% of the blast, or CD34+ population in the basal sample, with the exception of box D, where ~12% of the CD34+ events were P-S6 positive in the basal sample). Stimulation with SCF in the presence of the PI3 kinase inhibitor LY294002 (Fig 2, third row) has minimal impact on P-ERK or P-S6, but does show inhibition of phosphorylation of Akt as expected based on the pathways illustrated in Figure 1 (box C, Fig 2 third row).
The data suggest the presence of two distinct subpopulations of CD34+ cells after LY294002 inhibition, either alone (not shown), or in the presence of SCF stimulation (third row). The subpopulation with lower (and likely background) levels of expression of all four signaling markers in the sample treated with SCF and LY294002 implies the possible existence of a complex set of interactions between the PI3K/Akt pathway and other pathways (note population in box B, Figure 2 third row, expressing lower than basal or constitutive levels of P-ERK). in vitro treatment with Sorafenib lowered the levels of all four signaling phospho-proteins in the presence of SCF (fourth row) to levels significantly lower than those seen in the untreated sample (compare boxes B, C, and D, top row with same boxes in fourth row). Finally, in vitro stimulation with GM-CSF demonstrated significant increases in the level of P-STAT5 and P-S6, with minimal change from the basal level of P-ERK and P-Akt expression (bottom row).
Previously published studies have used semi-quantitative analytical methods to measure increases or decreases in single phospho-protein expression levels, including using “heat-maps.”4 While this approach is highly useful to compare multiple samples and/or treatments, it fails to measure or quantify interactions between pathways.
For example, in vitro treatment of this sample with SCF results in increase in P-ERK, P-Akt and P-S6 signals in a coordinated fashion, and treatment with LY294002 decreases P-Akt, but also appears to decrease P-ERK expression in a subpopulation of CD34+ cells.
Figure 3. Results from same AML patient presented in figure two, here shown as simultaneous visualization of all four signaling phospho-proteins as “radar plots.” in vitro treatment of individual samples (from left to right) included: no treatment (basal), SCF plus Ly294002, or GM-CSF. |
One approach to visualizing complex interactions by multiple signaling pathways is demonstrated in the “radar plots” shown in Figure 3. A radar plot displays events in a manner similar to a dot plot, but with the ability to use more than two parameters to determine where the events are displayed. Each event is plotted according to the length and orientation of each axis, proportional to the event’s strength along each axis. The resulting display is similar to the biplot used when doing principal component analysis.
The expression of four different markers (restricted here to only the CD34+ cells) is shown for basal, SCF plus LY294002, and GM-CSF treated samples from the same AML. The distinct advantage of this approach is visualization of interactions between two or more pathways.
As demonstrated in Figure 3, in the GM-CSF treated sample, there is a strong positive correlation between P-STAT5 expression level and P-S6 expression. While these types of tools are useful to study complex interactions, they require careful analysis by a observers. What is needed are tools to measure the signaling responses seen, comparing treatments or patients in a manner that is quantitative, at least semi-automated, and observer independent.
About the authors
David W. Hedley, MD, senior scientist, and Sue Chow are with the division of applied molecular oncology, Ontario Cancer Institute/Princess Margaret Hospital, University of Toronto. T. Vincent Shankey, PhD, is principal staff advanced research scientist with the Advanced Technology/Cellular Analysis Business Group, Beckman Coulter.
References
1. Mrózek, K., Radmacher, M.D., Bloomfield, C.D., and Marcucci, G. Molecular signatures in acute myeloid leukemia. Curr Opin Hematol. 16:64-69 (2009).
2. Matsumara, I., Mizuki, M., and Kanakura, Y. Roles for deregulated receptor tyrosine kinases and their downstream signaling molecules in hematological malignancies. Cancer Sci 99:479-485 (2008).
3. Pallis, M., Seedhouse, C., Grundy, M., and Russell, N. Flow cytometric measurement of phosphorylated STAT5 in AML: lack of specific association with FLT3 internal tandem duplications. Leuk. Res. 27:803-805 (2003).
4. Irish, J.M., Hovland, R., Krutzik, P.O., Perez, O.D., Bruserud, O., Gjertsen, B.T., and Nolan, G.P. Single cell profiling of potentiated phosphor-protein networks in cancer cells. Cell 118:217-228 (2004).
5. Ricciardi, M.R., McQueen, T., Chism, D., Milella, M., Estey, E., Kaldjian, E., Sebolt-Leopold, J., Konopleva, M, and Andreeff, M. Quantitative single cell determination of ERK phosphorylation and regulation in relapsed and refractory primary acute myeloid leukemia. Leukemia. 19:1543-1549 (2005).
6. Chow, S., Hedley, D., Grom, P., Magari, R., Jacobberger, J.W., and Shankey, T.V. Whole blood fixation and permeabilization protocol with red blood cell lysis for flow cytometry of intracellular phosphorylated epitopes in leukocyte subpopulations. Cytometry 67A:4-17 (2005).
7. Chow, S., Minden, M.D., and Hedley, D.W. Constitutive phosphorylation of the S6 ribosomal protein via mTOR and ERK signaling in the peripheral blasts of acute leukemia patients. Exptl. Haematol. 34:1182-1190 (2006).
8. Tong, F.K., Chow, S., and Hedley, D. Pharmacodynamic monitoring of zBAY 43-9006 (Sorafenib) in phase I clinical trials involving solid tumor and AML/MDS patients, using flow cytometry to monitor activation of the ERK pathway in peripheral blood cells. Cytometry 70B:107-114 (2006).
9. Hedley, D.W., Chow, S., Goolsby, C., and Shankey, T.V. Pharmacodynamic monitoring of molecular-targeted agents in the peripheral blood of leukemia patients using flow cytometry. Toxocol. Pathol. 36:133-139 (2008).