The HDAC inhibitor AR42 interacts with pazopanib to kill trametinib/dabrafenib-resistant melanoma cells in vitro and in vivo

Abstract

Studies focused on the killing of activated B-RAF melanoma cells by the histone deacetylase (HDAC) inhibitor AR42. Compared to other tumor cell lines, PDX melanoma isolates were significantly more sensitive to AR42-induced killing. AR42 and the multi-kinase inhibitor pazopanib interacted to activate: an eIF2α-Beclin1 pathway causing autophagosome formation; an eIF2α-DR4/DR5/CD95 pathway; and an eIF2α-dependent reduction in the expression of c-FLIP-s, MCL-1 and BCL-XL. AR42 did not alter basal chaperone activity but increased the ability of pazopanib to inhibit HSP90, HSP70 and GRP78. AR42 and pazopanib caused HSP90/HSP70 dissociation from RAF-1 and B-RAF that resulted in reduced 'RAF' expression. The drug combination activated a DNA-damage-ATM-AMPK pathway that was associated with: NFκB activation; reduced mTOR S2448 and ULK-1 S757 phosphorylation; and increased ULK-1 S317 and ATG13 S318 phosphorylation. Knock down of PERK, eIF2α, Beclin1, ATG5 or AMPKα, or expression of IκB S32A S36A, ca-mTOR or TRX, reduced cell killing. AR42, via lysosomal degradation, reduced the protein expression of HDACs 2/5/6/10/11. In vivo, a 3-day exposure of dabrafenib/trametinib resistant melanoma cells to the AR42 pazopanib combination reduced tumor growth and enhanced survival from ~25 to ~40 days. Tumor cells that had adapted through therapy exhibited elevated HGF expression and the c-MET inhibitor crizotinib enhanced AR42 pazopanib lethality in this evolved drug-resistant population.

Keywords: ER stress; autophagy; chaperone; death receptor.

Figure 1. Pazopanib and AR42 interact to kill PDX isolates of mutant B-RAF melanoma including vemurafenib resistant cells

(A–C) Sarcoma cells (HT1080, MES-SA, SKE-S1), renal carcinoma cells (A498, UOK121LN) and B-RAF melanoma cells (SKMEL24, SKMEL28, TPF-12-293, TPF-12-510, TPF-12-542, TPF-8-196, TPF-11-1081, TPF-12-198) were treated with drugs as indicated for 24 h to determine viability (n = 3 +/– SEM). ^#p < 0.05 greater than all values in sarcoma and RCCs. (D) Melanoma cells were treated with drugs as indicated for 24 h to determined viability. (n = 3 +/– SEM). ^#p < 0.05 greater than value in trametinib/dabrafenib treated cells.

Figure 2. [Pazopanib + AR42] kills trametinib/dabrafenib resistant melanoma cells and re-sensitizes resistant cells to the MEK/B-RAF inhibitor drug combination

(A) Wild type parental MEL28 and MEL28-R cells were treated with drugs for 24 h to determine viability. (n = 3 +/– SEM). ^¶p < 0.05 greater than corresponding value in parental wild type MEL28 cells. (B) Cells were treated as indicated with drugs, alone or in in combination, for 24 h to determine viability. (n = 3 +/–SEM) ^#p < 0.05 greater than value in T/D treated cells. (C) Wild type parental MEL28 cells and MEL28-R cells were treated as indicated with drugs, alone or in combination, for 24 h to determine viability. (n = 3 +/– SEM). *p < 0.05 less than corresponding value in parental wild type cells; ^¶p < 0.05 greater than corresponding value in parental wild type cells.

Figure 3. The regulation of protein phosphorylation by [pazopanib + AR42] in mutant B-RAF melanoma cells

(A) TPF-12-293 vemurafenib resistant cells were treated with drugs for 6 h. Immuno-fluorescence was performed (n = 3 +/– SEM). ^#p < 0.05 greater than vehicle control; *p < 0.05 less than vehicle control. (B) TPF-12-293 cells were either: transfected with an empty vector plasmid (CMV) or with plasmids to express dominant negative caspase 9, BCL-XL or c-FLIP-s; or with a scrambled siRNA (siSCR) or with siRNA molecules to knock down the indicated proteins. Twenty-four h after transfection cells were treated drugs for 24 h to determine viability. (n = 3 +/– SEM). ^¶p < 0.05 greater than corresponding value in siSCR transfected cells; *p < 0.05 less than corresponding value in siSCR/CMV cells **p < 0.05 less than corresponding value in siCD95 cells.

Figure 4. Multiple toxic BH3 domain proteins are required to mediate the death response to [pazopanib + AR42]

(A) TPF-12-293 and TPF-08-196 cells were transfected to knock down the indicated proteins. Twenty-four h after transfection cells were treated with drugs for 24 h to determine viability. (n = 3 +/– SEM). *p < 0.05 less than corresponding value in siSCR cells. (B) TPF-12-293 and TPF-08-196 cells were treated with drugs for 6 h. Immuno-fluorescence was performed (n = 3 +/– SEM) *p < 0.05 less than corresponding value in vehicle control; ^#p < 0.05 greater than corresponding value in vehicle control. (C) TPF-12-293 cells were transfected to knock down the expression of DR4 or DR5. Twenty-four h after transfection cells were treated with drugs for 24 h to determine viability. (n = 3 +/– SEM). *p < 0.05 less than corresponding value in siSCR cells. (D) TPF-12-293 and TPF-08-196 cells were transfected with an empty vector plasmid (CMV) or with plasmids to express: activated AKT; activated MEK1; activated mTOR; activated STAT3; dominant negative IκB S32A S36A. Twenty-four h after transfection cells were treated with drugs for 24 h to determine viability. (n = 3 +/– SEM). *p < 0.05 less than corresponding value in CMV transfected cells.

Figure 5. ATM-AMPK signaling and ER stress signaling are required for [pazopanib + AR42] to kill

(A) TPF-12-293 cells were transfected to knock down the expression of ATM. Twenty-four h after transfection cells were treated with drugs for 6 h. Immuno-fluorescence was performed. (n = 3 +/– SEM) ^#p < 0.05 greater than corresponding vehicle control; *p < 0.05 less that corresponding vehicle control. (B) and (C) TPF-12-293 cells were transfected to knock down the expression of AMPKα. Twenty-four h after transfection cells were treated with drugs for 6 h. Immuno-fluorescence was performed. ^#p < 0.05 greater than corresponding vehicle control; *p < 0.05 less that corresponding vehicle control. (D) TPF-12-293 cells were transfected to knock down the expression of eIF2α. Twenty-four h after transfection cells were treated with drugs for 6 h. Immuno-fluorescence was performed. ^#p < 0.05 greater than corresponding vehicle control; *p < 0.05 less that corresponding vehicle control. (E) TPF-12-293 cells were transfected with a plasmid to express LC3-GFP and in parallel transfected to knock down expression of AMPKα, ULK1, eIF2α or Beclin1. Twenty-four h after transfection cells were treated with drugs for 6 h or 12 h as indicated. (n = 3 +/– SEM). *p < 0.05 less that corresponding value in siSCR cells. (F) TPF-12-293 cells were transfected to knock down the expression of AMPKα, cathepsin B or eIF2α. Twenty-four h after transfection cells were treated with drugs for 24 h to determine viability. (n = 3 +/– SEM). *p < 0.05 less than corresponding value in siSCR cells.

Figure 6. [Pazopanib + AR42] regulates chaperone function

(A) TPF-12-293 cells were treated with drugs and the dose-response of chaperone ATPase activity in response to increasing concentrations of pazopanib in vitro is plotted (n = 3 +/– SEM) ^#p < 0.05 greater inhibition than vehicle control. (B) TPF-12-293 cells were transfected to knock down ATM expression. Twenty-four h after transfection cells were treated with drugs and the dose-response of chaperone ATPase activity in response to increasing concentrations of pazopanib in vitro is plotted (n = 3 +/– SEM) *p < 0.05 less inhibition than siSCR control. (C) TPF-12-293 cells were treated drugs and the dose-response of chaperone ATPase activity in response to increasing concentrations of pazopanib in vitro is plotted (n = 3 +/– SEM) ^#p < 0.05 greater inhibition than vehicle control. (D) and (E) TPF-12-293 cells were either not transfected or transfected with plasmids to express either FLAG-HSP90 or HA-HSP70. Twenty-four h after transfection cells were treated with drugs as indicated. For transfected cells after 1 h, HSP90 and HSP70 were immunoprecipitated and the amount of RAF-1 and B-RAF co-precipitating with the chaperones determined by SDS PAGE and western blotting. For non-transfected TPF-12-293 and TPF-08-196 cells after 6 h, cells were fixed in place and immuno-fluorescence performed to detect the total expression of RAF-1 and of B-RAF (n = 3 +/– SEM) *p < 0.05 less than vehicle control value.

Figure 7. HDAC6 regulates autophagosome formation

(A) Upper: The percentage catalytic inhibition of each HDAC by AR42 at a concentration of 1 μM; Lower: Melanoma cells were transfected to knock down the expression of HDAC6, combined with molecules to knock down the expression of HDAC1 or HDAC2 or HDAC3 or HDAC8 or HDAC10. Twenty-four h after transfection the cells were treated with pazopanib to determine viability (n = 3 +/– SEM). ^¶p < 0.001 greater than corresponding values in siSCR; [siHDAC6 + siHDAC1]; [siHDAC6 + siHDAC3]; [siHDAC6 + siHDAC8] cells. (B) Melanoma cells were transfected with: left graph: an empty vector plasmid, a plasmid to express wild type HDAC6 or a plasmid to express dominant negative HDAC6; right graph: to knock down expression of HDAC6. In parallel all cells were transfected with a plasmid to express LC3-GFP. Twenty-four h after transfection cells were treated with drugs for 6 h. The numbers of punctate LC3-GFP staining vesicles were counted (n = 3 +/– SEM) ^#p < 0.05 greater than corresponding value in siSCR/CMV cells; *p < 0.05 less than corresponding value in siSCR/CMV cells. (C) Melanoma cells were transfected to knock down expression of HDAC6. Twenty-four h after transfection cells were treated with drugs for 6 h. Staining was performed to determine the levels of P-γH2AX; p62/SQSMT1; LAMP2. (n = 3 +/– SEM) *p < 0.05 greater than vehicle control in siSCR cells; **p < 0.05 greater than corresponding value in siSCR cells.

Figure 8. AR42 down-regulates HDAC6 expression through AMPK/eIF2α-autophagy signaling

(A) TPF-12-293 and TPF-08-196 cells were transfected to knock down the expression of Beclin1 (B1) or ATG5 (A5). Twenty-four h after transfection cells were treated for 6 h with drugs and the expression of HDAC6 was determined (n = 3 +/– SEM) *p < 0.05 lower staining intensity than corresponding intensity in vehicle control treated cells; ^#p < 0.05 greater staining intensity than corresponding intensity in vehicle control treated cells. (B) Melanoma cells were transfected to knock down expression of AMPKα, eIF2α or Beclin1. Twenty-four h after transfection cells were treated with drugs for 6 h and the expression of LAMP2, LC3 and p62/SQSMT1 determined (n = 3 +/– SEM) *p < 0.05 lower staining intensity than corresponding intensity in vehicle control treated cells; ^#p < 0.05 greater staining intensity than corresponding intensity in vehicle control treated cells. (C) TPF-12-293 and TPF-08-196 cells were treated for 6 h with drugs in the presence or absence of Velcade (10 nM). The expression of HDAC6 was determined. There is no statistically significant difference in the reduction of HDAC6 regardless of the presence of bortezomib. (D) Melanoma cells were treated with drugs for 6 h and the co-localization of HDAC6 with p62/SQSMT1 and with LAMP2 determined.

Figure 9. HDAC6 expression regulates the production of reactive oxygen species and tumor cell killing

(A) Melanoma cells were transfected to knock down expression of HDAC6. In parallel they were transfected as indicated with an empty vector plasmid (CMV) or with plasmids to express active wild type HDAC6 (WT6) or a catalytically inactive HDAC6 (DN6). Twenty-four h after transfection cells were treated with drugs for 3 h. The level of reactive oxygen species (ROS) was determined (n = 3 +/– SEM) ^#p < 0.05 greater than corresponding value in siSCR/CMV cells; *p < 0.05 less than corresponding value in CMV cells. (B) Melanoma cells were transfected with an empty vector control plasmid (CMV), or with plasmids to express wild type HDAC6 or dominant negative HDAC6. Twenty-four h after transfection, cells were treated with drugs for 6 h and the expression of p62/SQSMT1 and LAMP2 determined. (C) Melanoma cells were transfected with an empty vector control plasmid (CMV), or with plasmids to express wild type HDAC6 or dominant negative HDAC6. Twenty-four h after transfection, cells were treated with drugs for 24 h to determine viability (n = 3 +/– SEM). ^#p < 0.05 greater than corresponding value in CMV cells; *p < 0.05 less than corresponding value in CMV cells.

Figure 10. [Pazopanib + AR42] suppresses melanoma tumor growth in vivo

(A) Tumor volumes were measured at the start of treatment (Day 0) and on the indicated days (n = 10, +/– SEM). *p < 0.05 less than Trametinib/Dabrafenib treated tumors; ^¶p < 0.05 less than AR42 alone or pazopanib alone treated tumors. (B) Animals were sacrificed when the tumor volume reached 250 mm³. Survival data are plotted as a Kaplan Meier survival curve. *p < 0.05 greater survival compared to vehicle control; ^¶p < 0.05 greater survival compared to AR42 alone. (C–E) Four micron slides of vehicle treated; AR42 treated; and [pazopanib + AR42] tumors were made and immunohistochemistry performed to determine the expression and co-localization of: C. F4/80 and CD11b (macrophages); D. GR-1 and CD11b. Note that GR-1 staining used a conjugated antibody that was detected in the Far Red range. The hue of the image was converted to green to permit co-localization analyses with CD11b (neutrophil); E. CD335 and CD69 (natural killer cell).

Figure 11. AR42 and [pazopanib + AR42] reduce plasma levels of MMP1, MMP2, MMP3 and IL-10 and increase plasma IL-8 levels

As tumor volumes reached 250 mm³, animals were sacrificed and blood and tumor obtained. Clarified plasma and tumor cell lysates were then subjected to multiplex assays as described in the Methods to detect the plasma levels of the indicated cytokines and phosphorylation status of signal transduction proteins using a Bio-Rad MAGPIX multiplex instrument (total 3 animals per condition, +/– SEM). ^#p < 0.05 less than corresponding value from vehicle control treated animals.

Figure 12. AR42 and [pazopanib + AR42] induce compensatory survival through expression of PDGF, bFGF, HGF and prolactin

(A–C) As tumor volumes reached 250 mm³, animals were sacrificed and blood and tumor obtained. Clarified plasma and tumor cell lysates were then subjected to multiplex assays as described in the Methods to detect the plasma levels of the indicated cytokines and phosphorylation status of signal transduction proteins using a Bio-Rad MAGPIX multiplex instrument (total 3 animals per condition, +/– SEM). ^#p < 0.05 less than corresponding value from vehicle control treated animals; *p < 0.05 greater than corresponding value from vehicle control treated animals.

Figure 13. AR42 and [pazopanib + AR42] reduce signaling by ERBB1 and induce signaling through the PDGFR

(A–C) As tumor volumes reached 250 mm³, animals were sacrificed and blood and tumor obtained. Clarified plasma and tumor cell lysates were then subjected to multiplex assays as described in the Methods to detect the plasma levels of the indicated cytokines and phosphorylation status of signal transduction proteins using a Bio-Rad MAGPIX multiplex instrument (total 3 animals per condition, +/– SEM). ^#p < 0.05 less than corresponding value from vehicle control treated animals; *p < 0.05 greater than corresponding value from vehicle control treated animals. (D) MEL28-R tumor cells were isolated from tumors previously treated with [pazopanib + AR42]. Cells were treated with drugs in the presence or absence of: afatinib (0.5 μM); sunitinib (0.5 μM); dasatinib (0.5 μM); ruxolitinib (0.5 μM); or crizotinib (0.5 μM). Tumor cells were isolated 12 h after treatment and cell viability determined (n = 3 +/– SEM) *p < 0.05 greater than corresponding value in [pazopanib + AR42] cells treated with vehicle control; ^#p < 0.05 less than corresponding value in [pazopanib + AR42] cells treated with vehicle control.

Figure 14. A simplified model of the molecular pathways by which pazopanib and AR42 combine to kill cancer cells

As an HDAC inhibitor, AR42 causes DNA damage and causes inhibitory chaperone acetylation. Pazopanib, as a multi-kinase and chaperone inhibitor, also inhibits chaperone activities as well as many class III receptor tyrosine kinases. DNA damage causes activation of ATM. ATM signals to activate the AMPK. AMPK signaling inactivates RAPTOR and TSC2 resulting in the inactivation of mTORC1 and mTORC2. Downstream of mTOR is the kinase ULK-1; the drug combination via AMPK promotes ULK-1 S317 phosphorylation which activates the kinase; the drug combination via mTOR inactivation reduces ULK-1 S757 phosphorylation which also activates the kinase. Activated ULK-1 phosphorylates ATG13 which is the key gate-keeper step in permitting autophagosome formation. AR42-induced ATM signaling also acts to reduce the activities of multiple chaperone proteins. Reduced HSP90 and HSP70 function lowers the expression of all receptor tyrosine kinases and the activities of STAT3, STAT5, ERK1/2 and AKT that results in lower expression of ROS/RNS detoxifying enzymes such as TRX and SOD2. Reduced GRP78 function causes activation of PERK and subsequently eIF2α. Enhanced eIF2α signaling reduces the transcription of proteins with short half-lives such as c-FLIP-s, MCL-1 and BCL-XL, and enhances expression of Beclin1, DR4 and DR5. Thus the convergent actions of reduced HSP90 and HSP70 chaperone activity and eIF2α signaling lead to a profound reduction in the protein levels of c-FLIP-s, MCL-1 and BCL-XL which facilitates death receptor signaling through CD95, DR4 and DR5 to activate the extrinsic apoptosis pathway. Enhanced Beclin1 expression converges with elevated ATG13 phosphorylation to produce high levels of autophagosome formation that acts to reduce HDAC2/5/6/10/11 expression but also to stall autophagosome fusion with lysosomes and to stall autolysosome maturation, which likely through cytosolic cathepsin proteases converges with the extrinsic apoptosis pathway to cleave BID and cause mitochondrial dysfunction. Tumor cell killing downstream of the mitochondrion was mediated by AIF and not caspases 3/7. The tumoricidal actions of AIF were facilitated by reduced HSP70 functionality as this chaperone can sequester AIF in the cytosol and prevent its translocation to the nucleus.