Potentiation of ATRA Activity in HL-60 Cells by Targeting Methylation Enzymes

All Trans Retinoic Acid (ATRA) is a Differentiation Inducer (DI) of Acute Promyelocytic Leukemia (APL) with proven clinical utility. Its benefits for other types of Acute Myelocytic Leukemia (AML) have been limited. In APL, ATRA targets the PML-RARA (promyelocytic leukemia/retinoic acid receptor-alpha)/DNA methyltransferase (DNMT)/Histone Deacetylase (HDAC) complex, facilitating its


Introduction
Methylation enzymes play a critical role in the regulation of cellular replication and differentiation, and can be dysregulated in cancer, including AML [1,2]. DNA methylation controls the expression of tissue specific genes, and pre-r RNA ribose methylation controls the production of ribosome's, which in turn dictate the commitment of cells to initiate replication [3][4][5]. Agents that interfere with methylation pathways and DNMT can induce cancer cell differentiation [1].
Biological generation of donor methyl groups that support DNA methylation is mediated by the ternary MMS enzyme complex consisting of MAT-MT-SAHH [6]. In the monomeric state the individual enzymes undergo rapid inactivation, while their engagement in the ternary enzyme complex promotes their stability and function. Monomeric SAHH is the most unstable, followed by MT and then MAT. Stability corresponds to their molecular size [6]. SAHH requires a stabilizing factor to assume a configuration favorable for the formation of a dimeric enzyme complex with MT, which can then form a ternary enzyme complex with MAT. In steroid hormone target tissues, such as prostate and breast, steroid hormones act to stabilize SAHH [6]. SAHH in other tissues also requires stabilizing factors similar to steroid hormones. In normal cells, SAHH is the dominant factor regulating the stability and activity of methylation enzymes. Therapeutic targeting to block stabilizing factor function can disrupt the ternary complex, leading to depletion of methyl group pools and cellular differentiation.
In cancer cells, MMS associates with telomerase (hTERT), altering the regulation and kinetic properties of the ternary enzyme complex [7,8]. Its benefits for other types of Acute Myelocytic Leukemia (AML) have been limited. In APL, ATRA targets the PML-RARA (promyelocytic leukemia/retinoic acid receptor-alpha)/DNA methyltransferase (DNMT)/Histone Deacetylase (HDAC) complex, facilitating its degradation, leading to loss of gene silencing and Terminal Differentiation (TD). In other forms of AML, ATRA targeting of WT RARA as a single agent fails to modulate the epigenetic changes blocking differentiation. However, when combined with agents that inhibit DNMT, such as 5-azacytidine, ATRA shows improved in vitro and clinical activity against AML. We previously demonstrated that targeting the methylation enzyme complex (MMS), consisting of Methionine Adenosyltransferase (MAT), Methyltransferase (MT) and S-Adenosylhomocysteine Hydrolase (SAHH), induced differentiation in the AML M2 HL-60 cell line model. Inhibitors of the ternary methylation enzyme complex act as Differentiation Helper Inducers (DHIs). While they are unable to induce significant terminal differentiation by themselves, they potentiate the action of DI's. We report here that DHIs that destabilize SAHH potentiate the capacity of ATRA to induce terminal differentiation in both sensitive and resistant HL-60 cells in vitro.
We also evaluated Tyrosine Kinase Inhibitors (TKIs) that interfere with the production of the stabilizing factor of SAHH, and steroid analogs that compete with the endogenous steroid stabilizing factor of SAHH. While 72% of early passage (sensitive) HL-60 cells demonstrated induction of Terminal Differentiation (TD) after exposure to 1µM concentrations of ATRA, only 43% of late passage (resistant) cells showed TD. When ATRA was combined with TKI imatinib mesylate or with the steroid analogs resveratrol or β-sitosterol, agents with no innate capacity to induce differentiation, late passage HL-60 cell TD increased significantly to 98%, 99% and 94% of cells, respectively. Only modest improvements in TD percentages were seen for combinations of ATRA with cytotoxic chemotherapy agents: S-adenosylmethionine (AdoMet), which may exercise a positive influence on the stability of ternary methylation enzymes. Binding of AdoMet by β-cystathionase has been shown to protect that enzyme against protease digestion [9]. These findings suggest that the increased K m value for MAT LT in malignant cells may contribute to MMS complex stability and down-stream DNA methylation and gene silencing. Consistent with this model, it was reported that the pool size of AdoMet and S-adenosylhomocysteine (AdoHcy) was diminished in cancer cells undergoing drug-induced terminal differentiation [10]. DNA methylation maintains cell cycle transit, while incomplete methylation diverts replicating cells into terminal differentiation [11]. Therefore, factors affecting the integrity of ternary methylation enzymes are critical for cell-cycle regulation and differentiation.
ATRA, the standard therapy for APL, produces excellent initial therapeutic outcomes, with up to 90% of cases showing complete response [12]. However, remissions can be short-lived and relapse with resistance to further treatment occurs. This shortcoming, due to incomplete induction of terminal differentiation by ATRA alone, can be remedied by its use in combination therapy with drugs such as Arsenic Trioxide (ATO) [13], which is an effective DHI. Thus, a combination of DI and DHI is essential to make a perfect drug for cancer therapy.
We previously found that inhibitors of MAT and MT could potentiate ATRA induced TD of HL-60 cells at dosages not appreciably affecting the growth and differentiation of HL-60 cells [14][15][16]. We report here that Signal Transduction Inhibitors (STIs), polyphenols and steroids used in combination with ATRA were capable of dramatically potentiating TD of both ATRA sensitive and resistant HL-60 cells. These agents may act in part by preventing the production of, or by antagonizing stabilizing factor of SAHH, leading to decreased methyl pool generation.

Chemicals and reagents
Chemicals and cell culture supplies used in this study were purchased from Sigma, St. Louis, MO, unless otherwise indicated. 35x10 mm cell culture dishes were from CytoOne, USA Scientific. Com. Imatinib mesylate crystal was a gift from Dr. H. Nguyen of Novartis Corp., East Hanover, NJ. Sunitinib malate capsules were from Pfizer Labs, NY, NY. Pazopanib tablets were from GlaxoSmithKline, RTP, NC. Metformin tablets were from Zydus Pharm, Pennington, NJ.

Culture of HL-60 cells
HL-60 cells were purchased from ATCC, Manassas, VI, which were initially maintained in ISCOVE's modified medium, supplemented with 10% fetal bovine serum, 2 mM glutamine, 50 units/ ml penicillin-50 µg/ml streptomycin for a few generations, and then transferred to RPMI 1640 medium to replace ISCOVE's modified medium. Cells were subcultured every 3 to 4 days at an initial concentration of 5-10 x 10 4 cells/ml.

Nitroblue Tetrazolium (NBT) assay
NBT assay was conducted according to Blair et al., [17]. Each 35x10 mm cell culture dish contained 2 ml of RPMI culture medium. HL-60 cells at an initial concentration of 5-10 x 10 4 cells were incubated with or without drugs for 5 or 3 days depending on the cell passages. Approximated 2.5 x 10 5 cells were precipitated at 600xg for 5 min. The cell pellet was suspended in 3 drops from a Pasteur pipet of NBT reagent consisting of 1 mg NBT and 5 µg phorbol-12-myristate 13-acetate per ml Hank Balanced Salt Solution (HBSS), and incubated at 37 o C for 30 min. The reaction was terminated by the addition of a drop from a Pasteur pipet of 4% paraformaldehyde in HBSS. NBT+ cells were counted using a hemacytometer.

Determination of potency of DHIs
The potency of DHIs was assessed by the Reductive Index (RI) as previously described [15]. Cell culture dishes were divided into several sets of 5 dishes containing ATRA of different concentrations to induce between 0 to 60% NBT+. One set had ATRA alone as control to yield ED 50 of ATRA. Other sets had different concentrations of DHIs together with ATRA concentrations matching the control set. After incubation at 37 o C for 72 h, cell numbers from each dish were determined, and an aliquot was withdrawn for NBT assay as above described. NBT+ cells in the control dishes without any drug were always below 1%. In the presence of different DHIs alone, NBT+ cells in general were below 10%. The respective control value was subtracted from each experimental value to yield the actual ED value. ED 50 values, defined as the dosages that induced 50% NBT+ cells, were estimated from plots of NBT+ values versus concentrations of ATRA in the absence and presence of DHIs. The reductive index is defined as the ED 50 value in the presence of DHI divided by the ED 50 value of ATRA alone. This value is inversely related to the effectiveness of the DHI agent.

Responsiveness of Early and Late Passage HL-60 Cells to the Induction of TD by ATRA
Early passage HL-60 cells (passaged in vitro for ≤ 3 months) replicated very slowly. They took ~5 days for 2 doublings, which was adequate to complete the differentiation process. In contrast, late passage cells cultured continuously for ≥ one year required only three days of incubation to undergo more than 2 doublings. The NBT assay was conducted on the 5 th day of the early passage cells carried for < 3 months, and on the 3 rd day of the late passage cells. Morphology of control HL-60 cells and TD cells induced with 1 µM ATRA for 72 h is shown in figure 1. The control cells typically show enlarged nuclei with very thin cytoplasm, whereas TD cells have more visible cytoplasm and shrunk nuclei. Approximately 15% of early passage HL-60 cells underwent spontaneous differentiation, i.e. NBT+, as shown in table 1. The early passage cells were more sensitive to induction of TD by ATRA. As the doubling time of HL-60 cells was reduced after multiple in vitro passages, these cells gradually lost the ability to undergo spontaneous differentiation and were less sensitive to the induction of TD by ATRA. The induction of TD of the early passage cells by ATRA was 95% at the peak concentration of 6 µM, which gradually declined as in vitro passages went on, and was reduced to only 33% by continuous in vitro passages for 5 years. Dosages above 6 µM did not improve the extent of TD.

Improved Induction of TD by Imatinib, resveratrol and β-sitosterol
Imatinib mesylate was effective at promoting TD of the early passage cells, but not the late passage cells as shown in table 2. Resveratrol and β-sitosterol promoted TD of the early passage cells to a similar degree as imatinib, and were also ineffective at promoting TD of the late passage cells. It has been reported that resveratrol is an effective inhibitor of growth signals [18,19], and that β-sitosterol is capable of modulating the growth of estrogen-responsive cancer cells [20]. The effect of resveratrol and β-sitosterol to potentiate induction of TD of early passage cells may be attributable to its capacity to potentiate the endogenous DI. These three agents dramatically potentiated the ATRA induced TD for both early and the late passage HL-60 cells, as shown in table 3. The extent of NBT+ cells approached 100%.

Improved Induction of TD by Cytotoxic Drugs
Cytotoxic drugs were partially effective at potentiating induction of TD, as shown in table 4. The potentiation was, however, not as dramatic as growth inhibitors or steroid analogs. In addition, the effectiveness to potentiate TD by cytotoxic drugs fell in narrow dosage ranges. Higher doses caused significant inhibition of cell growth that interfered with the terminal differentiation process.   We previously defined DHIs as inhibitors of individual enzymes of MMS that don't induce terminal differentiation as singe agents, but when applied in combination with a DI such as ATRA, can potentiate induction of terminal differentiation [14,21]. The potency of DHIs was determined by the RI (see materials and methods) [15]. Briefly, ED 50 values for TD in the absence and in the presence of a DHI were obtained from plots of NBT+ values versus ATRA concentrations. RI values were calculated from these data according to the following formula: RI = ED 50 of ATRA in the presence of a DHI/ED 50 of ATRA alone. Dosages of various growth inhibitors needed to achieve a RI of 0.5 are listed in table 5 and table 6.  Cell incubation and NBT assay were conducted as described in table l. Imatinib mesylate, resveratrol, and β-sitosterol were dissolved in methanol. The volume applied was not to exceed 2.5%. Data are the average of at least two determinations expressed as mean ± S.D. STIs are in general small molecule inhibitors of protein tyrosine kinases. Aberrant activation of tyrosine kinases secondary to mutations produce excess growth signals that result in the production of stabilizing factors for SAHH. We tested a variety of TKIs with ATRA, including sunitinib malate, a multikinase inhibitor, berberine, a potent inhibitor of EGFR-MEK-ERK signaling pathway, pazopanib, a potent inhibitor of the vascular endothelial growth factor, and imatinib mesylate, which targets Bcr-Abl in Philadelphia chromosome positive chronic myeloid leukemia and CD117 (cKIT) in gastrointestinal stromal tumors [22][23][24][25]. We also evaluated metformin, a well-known oral hypoglycemic agent, in combination with ATRTA based on its capacity to inhibit mTOR [26,27]. Among TKIs studied, sunitinib malate had the most potent activity as a DHI. More selective TKIs, such as pazopanib and imatinib mesylate, were less active (Table 5). Broad spectrum TKIs may be better candidates as DHIs.

In vitro passages of HL-60 cells
Among non-specific growth inhibitors studied, ATO and CoCl 2 had impressive activity as DHIs (Table 5). ATO, although quite toxic, required very low dosages to function as a DHI. CoCl 2 , an agent that up regulates hypoxia inducible factor, induced HL-60cell attachment. When cells became attached to the culture dish, cell growth was greatly diminished. The induction of cell attachment may enable CoCl 2 to act as a DHI. Sodium selenite is linked to an array of health benefits, including prevention of cancer. However, its cancer fighting potential has never been well characterized. Its activity as a DHI was limited. As listed in table 5, we found that several polyphenols had impressive activity as DHI's. Many of these polyphenols are present in foods that are regularly consumed. The activity of polyphenols may be attributable to inhibition of signaling pathways [18,23,[28][29][30][31].

Steroid Analogs as Differentiation Helper Inducers
Two factors critical for the stability of ternary cancer methylation enzymes are the steroid factors produced in response to growth signals for the stabilization of SAHH, and the association of telomerase with MAT. We do not know the identity of the stabilizing factor of the SAHH of HL-60 cells. It is very likely a steroid derivative, since dihydrotestasterone is the stabilizing factor of prostate SAHH [6]. We therefore examined a series of steroidal compounds to determine their capacities to potentiate ATRA activity (Table 6). Pregnenolone, derived from cholesterol, is the precursor for all steroid hormones, showed moderate DHI activity. Vitamin D can also be synthesized from cholesterol, which has a similar basic ring structure as steroid hormones. Vitamin D3 and dexamethasone were the two most active analogs. This implies that the actual stabilizing factor of SAHH of HL-60 cells may be structurally close to these two agents. It was apparent that biologically active steroids were better DHIs than the inactive pregnenolone precursor, which had the least activity among the steroid analogs studied. Biologically active steroids may not be good candidates for clinical consideration as DHIs simply because their biological activity may complicate cancer therapy. On the other hand, biologically inactive steroids such as pregnenolone, pregnenolone sulfate, guggulsterone, and β-sitosterol are good candidates for drug development as differentiation inducing agents.

Discussion
Our findings suggest that agents that target the methylation enzyme complex given in combination with ATRA can promote HL-60 cell differentiation. While differentiation therapy for solid tumors has not led to wide spread benefits, it has made significant headway for the therapy of hematological cancers [6,8,19,[24][25][26]. Unfortunately, acute leukemia's are made up of rapidly replicating cancer stem cells that develop resistance to differentiation therapy as well as chemotherapy [32,33]. Cancer stem cells are less responsive to cytotoxic chemotherapy because they are driven in part by signal transduction pathways that block apoptosis and they over express drug efflux pumps [34]. The option for the eradication of cancer stem cells is very limited. The ideal therapeutic agents must be small molecules that are relatively non-toxic and that bypass drug efflux pumps to reach  adequate intracellular concentrations to trigger cancer stem cells to undergo differentiation. Induction of differentiation is, therefore, an attractive strategy to eradicate cancer stem cells.
Our study showed dramatic improvement in the induction of TD after combined therapy with ATRA with TKIs or steroid analogs. This is thought to occur as a result of blockade of growth signals and methyl group transfer associated with gene silencing (Figure 2). These approaches may offer clinical utility for the treatment of ATRA-resistant clones. As shown in table 1, HL-60 cells gradually became resistant to ATRA-induced TD during prolonged propagation in vitro.
Gene silencing secondary to abnormal methylation enzyme activity is thought to play a role in the development of ATRA resistance in AML [2]. Our data suggest that a multi agent approach combining agents that disrupt DNMT activity in conjunction with targeting RARA can produce favorable therapeutic results. The use of ATRA alone could stimulate at most 48% of resistant HL-60 cells to undergo TD, whereas ATRA in combination with DHI agents led to TD in almost 100% of the resistant cells.
ATRA is an indirect DI that targets the RARA receptor to promote differentiation. The responsible direct DI is very likely oligoisoadenylate. Activation of oligoisoadenylate synthetase is a common feature of many DIs not targeting abnormal methylation enzymes directly [35]. We speculate that early passage HL-60 cells express low levels of oligoisoadenylate, enabling these cells to undergo  Cell culture and NBT assay were conducted as described in table 1. HL-60 cells used for the determination of reductive indices of various DHIs were maintained in vitro between 3 to 4 years. The N 3 /N 0 ratios were between 9.5 and 11.8. Berberine, imatinib mesylate and all polyphenols were dissolved in methanol to apply to the medium not to exceed 2.5% of the volume. Sunitinib was poured out of the capsule to be extracted with HBSS in a Dounce homogenizer. Pazopanib and metformin was extracted with HBSS in a Dounce homogenizer. Insoluble materials were removed by centrifugation at 1200xg for 10 min. CoCl 2 was dissolved in HBSS. As 2 O 3 and selenite was suspended in HBSS with phenol red indicator. 1 N NaOH was added drop wise until pink color persisted. All aqueous preparations were sterilized by passing through 0.2 µm membrane filters. Reductive indices were obtained as described in Materials and Methods. Dosages of DHIs needed to achieve reductive index of 0.5 were obtained from the plots of reductive indices versus concentrations. Data are the average of ≥ two determinations expressed as mean ± S.D.  Cell culture, NBT assay, and determination of reductive indices were conducted as described in table 5. Data are the average of ≥ two determinations expressed as mean ± S.D. spontaneous differentiation, while expression of this enzyme is totally lost in the late passage cells, preventing their capacity to carry out spontaneous differentiation.

Steroid analogs
Hypoxic growth in vivo is another element that could contribute to ATRA resistance. It has been shown that cancer stem cells able to undergo differentiation in normoxia were unable to undergo differentiation in hypoxia [36,37]. Hypoxia influences the switch between differentiation and stemness [36]. The enhanced expression of telomerase in hypoxia may be responsible for the blockade of differentiation. Cancer stem cells already express an unusually high level of telomerase [38], which has to be down-regulated for differentiation to take place [39]. Hypoxia Inducible Factor (HIF), stabilized under hypoxic conditions, is a transcription cofactor that up regulates telomerase, leading to stabilization of the MMS complex [40]. Resveratrol and topotecan treatment lead to decreased HIF under hypoxic conditions [41,42]. Consequently, resveratrol and topotecan may be candidates for targeting cancer stem cells. They were both found to promote TD in late passage HL-60 cells. Taken together, these data support the notion that ATRA in combination with DHI's may be of value in the treatment of non-M3AML.