Here, we show that imatinib prevents intrinsic and acquired resistance to doxorubicin by: 1) inhibiting c-Abl/Arg activation; 2) promoting doxorubicin-mediated cell cycle arrest at G2/M; 3) inhibiting activation of a STAT3-dependent HSP27/p38/Akt survival pathway; 4) promoting NF-κB-mediated inhibition of anti-apoptotic protein expression in a STAT3-dependent manner; and 5) inhibiting upregulation of the drug transporter, ABCB1, and directly inhibiting ABCB1 function (acquired resistance). These data are novel and significant because the upstream mechanisms that govern NF-κB-mediated transcriptional repression have not previously been identified. Furthermore, this is the first demonstration that HSP27/p38/Akt promote doxorubicin resistance in melanoma cells, and we are the first to show that STAT3 is involved in activation of this pathway.
The role of NF-κB in doxorubicin-induced cell death is controversial as doxorubicin-mediated activation of NF-κB prevents cell death in some cell types, while in other cells, doxorubicin-mediated activation of NF-κB promotes apoptosis by repressing expression of anti-apoptotic genes 
. In addition, the mechanism by which anthracyclines convert NF-κB into a repressor also is under debate. Barker and colleagues showed that doxorubicin induces p65 nuclear localization and DNA binding of a non-acetylated/non-phosphorylated form of p65, which inhibits NF-κB transcriptional activity in a histone deacetylase (HDAC)-independent manner 
. In contrast, Perkins and colleagues demonstrated that anthracyclines induce phosphorylation/acetylation and nuclear translocation of p65 in mouse embryo fibroblasts, and p65 represses gene expression by recruiting HDACs to gene targets 
. In addition, Yu and colleagues showed that p65 acetylation is required for its nuclear retention, which is inconsistent with data from Barker and colleagues who demonstrate that non-phosphorylated/non-acetylated p65 binds DNA, and thus, is in the nucleus 
. Here, we show that doxorubicin induces p65 phosphorylation and nuclear translocation, which is enhanced by imatinib treatment or silencing STAT3, and correlates with decreased NF-κB transcriptional activity and downregulation of NF-κB targets. Thus, STAT3 activation inhibits doxorubicin-mediated p65 nuclear localization, which is contrary to data obtained in untreated cancer cells indicating that STAT3 promotes p65 nuclear retention 
. Thus, our data indicate that STAT3 likely has an opposite role in regulating p65 nuclear localization in response to stimuli that convert NF-κB into a repressor 
Our data are consistent with Perkins and colleagues who demonstrate that doxorubicin increases NF-κB phosphorylation/acetylation/DNA binding but this activated NF-κB represses rather than activates transcription 
. We provide additional data aimed at elucidating the mechanism by demonstrating that imatinib targets (likely c-Abl/Arg) and STAT3 regulate NF-κB function following doxorubicin treatment. It is possible that c-Abl/Arg and STAT3 prevent conversion of NF-κB into a repressor by promoting recruitment of histone acetyltransferases (HATs) to p65/p50-DNA complexes whereas inhibition of c-Abl/Arg or STAT3 promotes recruitment of histone deacetylases (HDACs) 
. Interestingly, in cells that have acquired high-level resistance, doxorubicin increases NF-κB transcriptional activity, and NF-κB does not appear to function as a repressor. This may be because doxorubicin only modestly inhibits STAT3 phosphorylation, and thus, HAT recruitment may be dominant over HDAC recruitment in these cells. Addition of imatinib dramatically inhibits STAT3 and leads to repression of NF-κB targets (cIAP1, XIAP), perhaps by facilitating HDAC/p65 rather than HAT/p65 complexes. In either event, our data are highly significant because they demonstrate that imatinib converts a master survival regulator, NF-κB, from a pro-survival into a pro-apoptotic factor, thereby rendering a conventional chemotherapeutic agent more effective for treating metastatic disease. These data are extremely important because they indicate that NF-κB inhibitors may be ineffective in sensitizing tumors containing activated c-Abl/Arg to anthracyclines, and instead may antagonize anthracycline-induced apoptosis.
In addition to inhibiting survival signaling, we also demonstrate that c-Abl/Arg inhibitors partially prevent ABCB1 upregulation in cells that acquire doxorubicin resistance, and also directly inhibit ABCB1 function. c-Abl/Arg inhibitors have been identified as substrates of drug transporters in other cell types 
; however, this is the first demonstration that they inhibit ABCB1 in melanoma cells. In addition, we are the first to show that c-Abl promotes expression
of an ABC transporter. ABC transporter upregulation has been shown to occur via a variety of pathways including FOXO3a and PI3K/Akt, NF-κB, HSP27, and ERK pathways 
; ongoing experiments are aimed at identifying the mechanism of ABCB1 upregulation. ABCB1 also is a transporter for other chemotherapeutic agents, such as paclitaxel, vinblastine, vincristine, etoposide as well as for the selective estrogen receptor modulator (SERM), tamoxifen 
. Thus, c-Abl/Arg inhibitors are likely to reverse resistance to many of these agents as well. In support of this hypothesis, we demonstrate that c-Abl/Arg inhibitors sensitize melanoma cells to paclitaxel, and Wang and colleagues showed that c-Abl/Arg inhibition sensitizes breast cancer cells to tamoxifen 
Doxorubicin is utilized to treat many cancers including triple-negative breast cancer; however, resistance and toxicity limit its effectiveness 
. Anthracyclines are not routinely used to treat metastatic melanoma due to intrinsic resistance; however, current treatment regimens (dacarbazine-DTIC or temozolamide-TMZ) are also ineffective (<5% success rate), and newer biological agents only extend survival by 4 months 
. Our data demonstrate that c-Abl/Arg inhibitors (imatinib, nilotinib) are effective in reversing intrinsic and acquired doxorubicin resistance, thereby increasing doxorubicin effectiveness at lower doses. These data have important clinical ramifications as they indicate that Abl inhibitor/doxorubicin combinations may be effective for treating cancers driven by activated c-Abl and Arg.