We have previously demonstrated that the protein expression and activation profile of apoptosis and signal transduction pathway (STP) proteins in acute myelogenous leukemia (AML) cells that survive in vitro chemotherapy exposure (Survivor Cells) differ markedly from that seen in the bulk population of AML blasts present at diagnosis. This methodology, although cumbersome, could provide insight into the adaptive changes utilized by the small fraction of cells that survive chemotherapy but which ultimately lead to relapse and death in the majority of AML patients. Unfortunately, the small number of available survivor cells (<50K at day 4 from a starting number of 1x106) limited the number of time points that could be evaluated and constrained the number of proteins that could be assessed. The development of RPPA, where small amounts of protein (~ 250 cell equivalents/sample/slide) are printed in serial dilution onto nitrocellulose membranes affixed to a glass slide, and the slides are then probed with a single antibody of interest, could resolve these limitations. To test this, we generated survivor cells from two AML cell lines, U937 and NB4, by exposure to 1, 5 or 10μM of ara-c, 0.05 μM of Idarubicin (IDA), or both, and collected cells at various 0.5, 1, 2, 4, 24, 48, 72 and 96 hours after exposure. AnnexinV positive cells were depleted to yield a population of currently viable cells from which whole cell lysates were generated. RPPA were generated from these samples using 8 serial dilutions and then probed with antibodies for 19 apoptosis (BCL-XL, BCL2, Bad, pBAD(ser112, ser136 ser155), Bax, Bak, XIAP) or STP proteins (total and phospho forms of AKT(ser473, Thr308), ERK2, PKCα, MEK,) or actin. Expression was related to the expression in cells cultured for similar duration in the absence of chemotherapy.

In response to different doses of ara-C, the patterns of expression and dynamics over time differed for the two cell lines. For U937 the magnitude and timing of changes were identical for 1μM and 10 μM of ara-C with the exception that increases in PKCα, pPKCα, and MEK, and decreases in pMEK and pERK were greater with 10 than 1 μM of ara-C. In contrast, NB4 survivors of 1μM ara-C showed late (>24 hr) increases in BCL-XL, BCL2, XIAP, Bax, Bad, pBad112, pBad 136, pBad155, PKCα and pPKCα that did not develop in survivors of 10μM ara-C.

When NB4 cells were treated with IDA alone, similar levels and patterns of expression were seen for all proteins, except pBAD-136, when compared to ara-C alone. In contrast, levels and patterns in U937 cells treated with IDA showed lower total AKT and BAD and higher BCL-XL, pERK, XIAP compared to ara-C treated cells. When U937 cells were treated with both agents, the timing and expression levels mirrored that seen in cells treated with IDA alone. NB4 cells treated with both agents showed transient increases in pERK from 30 min to 4 hours and a late increase in pPKCα but were otherwise similar in comparison to cells treated with IDA or ara-C alone.

We conclude that the pattern of STP and apoptosis regulatory protein expression and activation in survivor cells does differ over time and varies by the dose and combinations of chemotherapy agents used. Studies in patient-derived samples will be studied to see if these patterns might be utilized to predict sensitivity or resistance to selected therapies.

Author notes

Corresponding author