Increased expression of the androgen receptor with p300 and interleukin-6 coactivators compensate for oligonucleotide suppression of bcl-2: no increased CREB binding protein or interleukin-4 expression

Abstract

Background:

Antisense oligonucleotides (oligos) have been employed against in vivo and in vitro prostate cancer models targeting growth regulatory proteins. While most oligos have targeted growth factors or their receptors, others have been directed against inhibitors of apoptosis and mediators of androgen action. We previously evaluated a set of oligos which targeted and comparably suppressed the expression of the apoptosis inhibitor protein bcl-2. LNCaP cells adapted to this restoration of apoptosis with suppression of caspase 3 (an apoptosis promoter) and an enhanced expression of the androgen receptor (AR), suggesting an increased sensitivity to androgens.

Methods and results:

In a continuation of this study, we evaluated the expression of AR coactivators p300, its homolog CREB binding protein (CREBBP) and cytokines interleukin (IL)-4 and IL-6, finding p300 and IL-6 similarly enhanced.

Conclusions:

LNCaP cells are hormone sensitive and untreated cells express minimal p300 activity. Therefore, the enhanced expression which followed oligo treatment makes its induction more impressive and implies a pattern of gene expression more associated with later stage (androgen insensitive) disease. This suggests that oligo treatment directed against bcl-2 can be evaded through compensatory changes in AR expression and some coactivators, promoting tumor growth, and may promote transformation of the tumor to a more aggressive phenotype.

Introduction

Gene therapy has now reached clinical trials for the treatment of human prostate tumors and antisense oligonucleotides (oligos) have targeted bcl-2 and clusterin in efforts to restore apoptosis following radiotherapy [Mu et al. 2005] or chemotherapy [Yamanaka et al. 2006] of this disease. If such therapy is to be successful, it is important to examine mechanisms by which tumors compensate to evade it and become resistant.

We recently reported that oligo-mediated inhibition of bcl-2 suppressed the expression of the apoptotic promoter caspase-3 in the LNCaP tumor model [Rubenstein et al. 2011a] and androgen receptor (AR) expression was significantly increased [Rubenstein et al. 2011b]. We now evaluate effects upon AR transcriptional coactivators p300, CREB binding protein (CREBBP), and interleukins 4 and 6 (IL-4, IL-6), often associated with and highly expressed by advanced prostate tumors [Bouchai et al. 2011]. Enhanced AR and p300 expression could not only select cells which again evade apoptosis and enhance tumor progression (even with bcl-2 suppression), but could also promote the emergence of a more aggressive (hormone- insensitive) phenotype. For gene therapy to be successful it must be more specific and mechanisms of compensation identified and suppressed.

Effective therapeutics target unique characteristics of etiologic agents, including bacterial cell walls and ribosomes or viral-encoded proteolytic enzymes. The development of tumor resistance is less specific or distinct since cancer cells are not substantially different from noncancerous (differentiated cells). They use the same biochemical pathways, and unless virally induced, most are (even antigenically) similar to normal cells. The effectiveness of chemotherapy capitalizes on the fact that within a tumor mass a greater proportion of cells are in the process of replicating. Therefore, anticancer drugs frequently target some aspect of DNA synthesis. In prostate and breast cancers, growth factor (either hormone or protein targeted) deprivation provides another type of therapy. In an extension of this approach transcriptional activity initiated by DNA hormone response elements recognized by ARs, estrogen receptors (ERs) or coactivators (p300, CREBBP, IL-4 and IL-6), which are more prevalent in advanced, hormone insensitive, disease [Bouchai et al. 2011], could also be targeted. However, for most chemotherapy tumor cell specificity is relative, often lacking, and most agents administered (such as Taxol [Bristol-Meyers Squibb Company, Princeton, NJ]) have significant toxicity towards other replicating cells.

Gene therapy is based on a similar premise and while effective protocols employ either translational suppression (oligo-mediated) or replacement [of inactivated, mutated or deleted suppressor genes like phosphatase and tensin homolog (PTEN)] technology [Huang et al. 2001], both tumor and normal cells express the same genes. Targets for gene therapy are found in many pathways and it is likely that hundreds (or thousands) of genes can ultimately become involved in the malignant transformational process. Although tumors can express an overall altered pattern of gene expression, the levels of many growth regulatory genes are often similar to those of normal cells. Resistance develops because the biochemical pathways involved are complex and highly regulated by many stimulatory and inhibitory factors, each altered by therapy, therefore it has been suggested that tumors can alter their dependence on single influences by reliance on others through compensation [Rubenstein et al. 2011a].

Tumors are essentially heterogeneous masses of rapidly growing and selectively adapted cells whose sole purpose is to survive, replicate, and while doing so, evade therapeutic interventions. The best example is the emergence of hormone insensitive prostate cancer cells following androgen deprivation therapy, resulting in the increased expression of the autocrine loop consisting of transforming growth factor α (TGFα) and its binding site the epidermal growth factor receptor (EGFR) in prostate and breast tumors [Rubenstein et al. 1994].

As bacteria and viruses mutate to evade antibiotic and antiviral agents, tumor cells are under similar selective pressure to evade chemotherapy. Although newly developed forms of gene therapy provide specific ways to inhibit uncontrolled growth or promote (re-establish) apoptosis, the unintended consequences of intervention are poorly understood, and some may compensate for the originally intended effect.

Methods

Oligonucleotides

Oligos (mono- or bispecific) were purchased from Eurofins MWG Operon (Huntsville, AL, USA). Each was phosphorothioated on three terminal bases at 5’ and 3’ positions. Stock solutions were made to a final concentration of 625 μM in sterile Dulbecco phosphate-buffered saline.

Base sequences

Each oligo contained at least one CAT sequence and targeted the area adjacent to the mRNA AUG initiation codon for the respective targeted protein (EGFR or bcl-2). These bispecifics were employed because previous experiments conducted in LNCaP cells with MR24 and MR₄₂ produced significant growth inhibitions of 50% and 59% respectively [Rubenstein et al. 2009].

MR₄ (monospecific targeting bcl-2) T-C-T-C-C-C- A-G-C-G-T-G-C-G-C-C-A-T

MR₂₄ (bispecific targeting EGFR/bcl-2) G-A-G- G-G-T-C-G-C-A-T-C-G-C-T-G-C-T-C- T-C-T-C-C-C-A-G-C-G-T-G-C-G-C-C-A-T

MR₄₂ (bispecific targeting bcl-2/EGFR) T-C-T-C- C-C-A-G-C-G-T-G-C-G-C-C-A-T-G-A-G-G-G-T-C-G-C-A-T-C-G-C-T-G-C-T-C

Cell culture

LNCaP cells were grown in RPMI 1640 supplemented with 10% bovine serum, 1% L-glutamine and 1% penicillin/streptomycin in a 5% CO₂ incubator. Log phase cells were harvested using ethylenediaminetetraacetic acid/trypsin and equally distributed into 75 cm² flasks (Corning, NY, USA). At intervals media were either supplemented or replaced with fresh.

Oligonucleotide treatment prior to polymerase chain reaction

Four days prior to oligo addition, when cell density approached 75% confluence, 10 ml of fresh media was added. Cells were incubated for an additional 3 days before 5 ml of media was replaced with fresh the day before oligos were added. Stock oligos (100 μl) were added to bring the final concentration to 6.25 μM. Incubation proceeded for an additional 24 h in the presence or absence of monospecific MR₄, or the MR₂₄ and MR₄₂ bispecifics.

RNA extraction

Following treatment, media was removed, a single milliliter of cold (4°C) RNAzol B was added to each 75 cm² culture flask and the monolayer lysed by repeated passage through a pipette. All procedures were performed at 4°C. The lysate was removed, placed in a centrifuge tube to which 0.2 ml of chloroform was added, and shaken. The mixture stayed on ice for 5 min, was spun at 12,000 g for 15 min, and the upper aqueous volume removed and placed in a fresh tube. An equal volume of isopropanol was added, the tube shaken, and allowed to stay at 4°C for 15 min before similar centrifugation to pellet the RNA. The supernatant was removed, the pellet washed in a single milliliter of 75% ethanol, then spun for 8 min at 7500 g. The ethanol was pipetted off and the formed pellet air dried at –20°C.

RNA quantitation

RNA was resuspended in 250 μl of diethyl pyrocarbonate (DEPC)-treated water, and quantitated using a Qubit florometer and Quant-iT RNA assay kit (Invitrogen, Carlsbad, CA). DEPC is an inhibitor of RNase activity.

Reverse transcriptase polymerase chain reaction

Extracted RNA was diluted in DEPC treated water to 40 μg/μl. 1-4 μl of this RNA was added to1 μl of both sense and antisense primers (forward and reverse sequences) for bcl-2, AR, p300, CREBBP, IL-4 and IL-6. From a kit purchased from Invitrogen the following reactants were added for reverse transcriptase polymerase chain reaction (RT-PCR): 25 μl of 2X reaction mixture, 2 μl SuperScript III RT/platinum Taq mix, tracking dye and 3 μl MgSO₄ (of a 5mM stock concentration). DEPC-treated water was added to yield a final volume of 50 μl. RT-PCR was performed for 2 × 25 cycles using the F54 program in a Sprint PCR Thermocycler (Bio-Rad, Berkeley, CA). As a control for RT-PCR product production, human actin expression was tested in RNA extracted from HeLa cells which was provided in a kit purchased from Invitrogen (in the reaction mixture, no MgSO₄ was included, the difference compensated for by 3 μl of DEPC-treated water).

Primers

Actin

Forward primer sequence: 5’ CAA ACA TGA TCT GGG TCA TCT TCT C 3’.

Reverse primer sequence: 5’ GCT CGT CGT CGA CAA CGG CTC.

PCR product produced was 353 base pairs in length.

Bcl-2

Forward primer sequence: 5’ GAG ACA GCC AGG AGA AAT CA 3’.

Reverse primer sequence: 5’ CCT GTG GAT GAC TGA GTA CC 3’.

PCR product produced was 127 base pairs in length.

Androgen receptor

Forward primer sequence: 5’ CGG AAG CTG AAG AAA CTT GG 3’.

Reverse primer sequence: 5’ ATG GCT TCC AGG ACA TTC AG 3’.

PCR product produced was 155 base pairs in length.

300

Forward primer sequence: 5’ CGC TTT GTC TAC ACC TGC AA 3’.

Reverse primer sequence: 5’ TGC TGG TTG TTG CTC TCA TC 3’.

PCR product produced was 167 base pairs in length.

CREB binding protein

Forward primer sequence: 5’ CAC CAG CAG ATG AGG ACT CT 3’.

Reverse primer sequence: 5’ TAC ACC GGT GCT AGG AGG AG 3’.

PCR product produced was 222 base pairs in length.

Interleukin 4

Forward primer sequence: 5’ CAG CCT CAC AGA GCA GAA GA 3’.

Reverse primer sequence: 5’ GTT TCA GGA ATC GGA TCA GC 3’.

PCR product produced was 209 base pairs in length.

Interleukin 6

Forward primer sequence: 5’ ATG CAA TAA CCA CCC CTG AC 3’.

Reverse primer sequence: 5’ GAG GTG CCC ATG CTA CAT TT 3’.

PCR product produced was 167 base pairs in length.

Detection and quantitation of product

Agarose gel electrophoresis

Preparation of 1.5% agarose gels was in a 50 ml volume of TBE buffer (1X solution: 0.089 M Tris borate and 0.002M EDTA, pH 8.3), containing 3 μl of ethidium bromide in a Fisher Biotest electrophoresis system (Pittsburg, PA). Samples were run for 2 h at a constant voltage of 70 using a BioRad 1000/500 power supply source. To locate the amplified PCR product, 3 μl of a molecular marker (Invitrogen) which contained a sequence of bases in 100 base pair increments (Invitrogen) as well as 2 μl of a sucrose-based bromphenol blue tracking dye were run in each gel.

Quantitation

Gels were visualized under ultraviolet light and photographed using a Canon 800 digital camera. Photos were converted to black and white format and bands quantitated using Mipav software provided by the National Institute of Health. Means and standard deviations were compared using Student’s t tests to determine significance.

Results

Bcl-2 expression

As a control (data not shown) for RT-PCR product production, human actin expression was tested in RNA extracted from HeLa cells [Rubenstein and Guinan, 2010].

LNCaP cells incubated for 24 h in the presence of 6.25 μM of oligos suppressed bcl-2 expression, and support the finding of comparable biologic activity in both mono- and bispecific oligos measured in the in vitro cell growth inhibition experiments [Rubenstein and Guinan, 2010]. When photographs of the identified product bands were scanned on agarose gels and quantitated using Mipav software, in a series of runs, the greatest expression of bcl-2 was always found in untreated LNCaP cells. Those treated with oligos, whether mono- or bispecific, produced bands which indicated obvious (to the naked eye) suppression. For each oligo evaluated, the greatest amount of suppression measured approached 100% for the monospecific MR₄; and for the bispecifics MR₂₄ and MR₄₂, 86% and 100% respectively. Suppression was found in both repeat PCR runs with bcl-2 primers and in repetitive agarose gel quantifications. Figure 1 presents a bcl-2 product band in the expected 127 base pair region which in this run was inhibited 23% by treatment with the monospecific MR₄, and 86 and 74% respectively by bispecifics MR₂₄ and MR₄₂, as measured by Mipav software.

Figure 1.

Bcl-2 expression is suppressed by oligonucleotides as indicated in a representative agarose gel.

EGFR expression was not evaluated because in previous experiments we found that although oligos directed against EGFR significantly inhibited growth, the overall content of mRNA was not altered [Rubenstein et al. 2002].

Androgen receptor expression

Comparable amounts of extracted RNA from LNCaP cells treated with either mono- or bispecific oligos directed against bcl-2 (and EGFR in the bispecifics) were then evaluated by RT-PCR using primers directed against AR. When background intensity was subtracted, the relative intensity of all bands corresponding to AR representing cells treated with MR₄, MR₂₄ and MR₄₂ compared with controls was enhanced 31.2% ± 26.0 (p = 0.015), 58.5% ± 51.4 (p = 0.019) and 53.1% ± 45.9 (p = 0.019). These results were pooled from both duplicate PCR runs and multiple gels (a total of six gels were evaluated), and indicate similar (significant) enhancement of AR activity is produced by each oligo type. A representative band is depicted in Figure 2.

Figure 2.

Androgen receptor expression is enhanced by oligonucleotides as indicated in a representative agarose gel.

300 expression

Comparable amounts of extracted RNA from LNCaP cells treated with either mono- or bispecific oligos directed against bcl-2 (and EGFR in the bispecifics) were then evaluated by RT-PCR using primers directed against p300. When background intensity was subtracted, the relative intensity of all bands corresponding to p300 representing cells treated with MR₄, MR₂₄ and MR₄₂ compared with controls was increased 82.9% ± 51.9 (p = 0.006), 93.0% ± 87.3 (p = 0.044) and 105.4% ± 65.9 (p = 0.007). These results were pooled from both duplicate PCR runs and multiple gels (a total of six gels were evaluated), and indicate similar (significant) enhancement of p300 activity is produced by each oligo type. A representative band is depicted in Figure 3.

Figure 3.

p300 Expression is enhanced by oligonucleotides as indicated in a representative agarose gel.

CREB binding protein expression

Comparable amounts of extracted RNA from LNCaP cells treated with either mono- or bispecific oligos directed against bcl-2 (and EGFR in the bispecifics) were then evaluated by RT-PCR using primers directed against CREBBP. When background intensity was subtracted, the relative intensity of all bands corresponding to CREBBP representing cells treated with MR₄, MR₂₄ and MR₄₂ compared with controls was varied. Although significant decreases were found in both the monospecific MR₄ (–32.1% ± 5.7; p < 0.05) and bispecific MR₄₂ (–26.8% ± 32.3; p < 0.05) there was no significant change in MR₂₄. This pattern (when similar changes are not seen in both bispecifics) is unique in the many proteins we examined, and we conclude that unlike the increased expression of AR and p300, there is no enhancement of this transcription coactivator. These results were pooled from both duplicate PCR runs and gels. A representative band is depicted in Figure 4.

Figure 4.

CREB binding protein (CREBBP) expression is not enhanced by oligonucleotides as indicated in a representative agarose gel.

Interleukin 4 expression

Comparable amounts of extracted RNA from LNCaP cells treated with either mono- or bispecific oligos directed against bcl-2 (and EGFR in the bispecifics) were then evaluated by RT-PCR using primers directed against IL-4. When background intensity was subtracted no increase in expression of IL-4 was identified, although a 31.1% decrease in the expression in the MR₄₂-treated cells was significant (p = 0.003). It does not appear that compensation for bcl-2 suppression involves IL-4 in the development of greater androgen sensitivity (in contrast to increases in AR, p300 and IL-6). A representative band is depicted in Figure 5.

Figure 5.

Interleukin 4 (IL-4) expression is not enhanced by oligonucleotides as indicated in a representative agarose gel.

Interleukin 6 expression

Comparable amounts of extracted RNA from LNCaP cells treated with either mono- or bispecific oligos directed against bcl-2 (and EGFR in the bispecifics) were then evaluated by RT-PCR using primers directed against IL-6. When background intensity was subtracted, the relative intensity of all bands corresponding to IL-6 representing cells treated with MR₄, MR₂₄ and MR₄₂ compared with controls was increased 236.9% ± 154.4 (p = 0.001585), 219.3% ± 170.4 (p = 0.005231) and 139.2% ± 88.8 (p = 0.001537). These results were pooled from both duplicate PCR runs and multiple gels (a total of seven gels were evaluated), and indicate similar (significant) enhancement of IL-6 activity is produced by each oligo type. A representative band is depicted in Figure 6.

Figure 6.

Interleukin 6 (IL-6) expression is enhanced by oligonucleotides as indicated in a representative agarose gel.

Discussion

In prostate cancers, androgen deprivation provides the first tier of therapy; however, most tumors soon recur and metastasize as hormone-insensitive variants. The AR (also known as NR3C4; nuclear receptor subfamily 3, member 4) plays a principle role in male sexual development, prostate function, cancer progression and various treatment strategies. Following the cytoplasmic binding of the AR to testosterone or its metabolite, dihydrotestosterone, it undergoes a conformational change which is accompanied by dissociation of heat shock proteins and translocation into the cell nucleus. The AR dimerizes, binds to hormone response elements of the DNA, and acts as a transcription factor to enhance the synthesis of growth-stimulating proteins, including insulin-like growth factor (ILGF) [Pandini et al. 2005]. In addition to ILGF, other growth factors such as TGFα, acting through their respective receptors (such as EGFR, which binds TGFα), contribute to unregulated prostate cancer growth. These protein factors and their receptors are also considered targets for suppressive gene therapy with antisense oligos [Rubenstein et al. 1994].

The American Cancer Society estimates that in spite of early detection, screening for prostate-specific antigen (PSA) and effective treatments for localized disease, in the USA there will be 28,170 deaths from prostate cancer with 241,740 newly diagnosed cases [American Cancer Society, 2012]. New types of treatment, including gene therapy and translational inhibition, must be developed and employed (probably in combination with traditional androgen ablation). Additional approaches could target transcriptional activity initiated by DNA hormone response elements recognized by AR and ER or their coactivators (p300, CREBBP, IL-4 and IL-6), more prevalent in advanced, hormone-insensitive disease [Bouchai et al. 2011] in which they are essential for cell growth and (p300) govern cyclin expression, regulating the transition between G1, S, G2 and M phases of mitosis [Heemers et al. 2007]. Acting with IL-6, p300/CREBBP plays a role in the androgen-independent expression of PSA [Debes et al. 2005]. Treatment of prostate cancer cells with siRNA directed against p300 reduces cancer cell growth [Debes et al. 2005] and eliminates the ability of IL-6 to induce PSA [Klocker et al. 2007]. IL-6 is increased in the blood of patients with advanced and metastatic disease [Jehn, CF, 2006], and anti-IL-6 therapy employing tocilizumab and ALD518 is now in clinical trials [Barton, 2005; Smolen and Maini, 2006].

In the LNCaP model, administration of R1881 reduces both CREBBP mRNA and the encoded CREBBP protein, suggesting that following androgen ablation the expression of some coactivators increases and contributes to a state of AR hypersensitivity [Klocker et al. 2007]. Since both transcriptional coactivator proteins p300 and CREBBP are expressed to a greater extent in advanced prostate cancer [Heemers et al. 2007], well differentiated, androgen-sensitive LNCaP cells would be expected to have relatively low expression of p300. This is evident in Figure 3, in which the untreated group p300 expression is barely detectable. The enhanced expression seen following oligo treatment makes its induction appear more impressive, and might indicate a possible transition to a pattern of gene expression more associated with later stage (androgen-insensitive) disease. This suggests that oligo treatment directed against bcl-2 can not only be evaded through compensatory changes in expression which encourage tumor growth, but could also promote further dedifferentiation and hormone insensitivity. These changes occur rapidly following treatment, and in these experiments were detected 24 h after addition of the oligos which suppressed bcl-2 and presumably altered the apoptotic process, leading to compensation.

Innovative protocols to disrupt androgen-driven tumor progression have also employed antisense oligos directed against the enzyme for conversion of testosterone to dihydrotestosterone (5α reductase), heat shock proteins, p300 and the AR itself. Although LNCaP cells express an AR which is mutated in the binding domain, the Eder and Rubenstein groups have separately demonstrated growth inhibition in this in vitro model employing oligos [Eder et al. 2000; Rubenstein et al. 2010].

Gene therapy is a complex process requiring multiple pathways (and the regulatory proteins) to be simultaneously regulated. Effective therapy should target only the ‘driver genes’ (usually kinases) which greatly influence tumor growth and not similarly mutated ‘passenger genes’.

Oligos (produced by Oncogenex Pharmaceuticals, Bothell, WA, USA) have reached clinical trials for the treatment of prostate cancer (OGX-011), while others remain in preclinical development (OGX-225). Often administered in combination with traditional chemotherapy, these oligos target bcl-2, clusterin (OGX-011 in phase II testing), heat shock protein 27 (OGX-427) or insulin growth factor binding proteins (OGX-225) (www.oncogenex.com). Many represent efforts to restore tumor apoptosis by eliminating suppressive bcl-2 [Mu et al. 2005; Yamanaka et al. 2006] associated with treatment resistance. Similar approaches are directed at clusterin. For (tumor suppressor) genes which are either diminished or lacking in expression, gene transfection has been attempted in prostate cells which contain a mutated PTEN [Huang et al. 2001]. Although antisense oligos are specifically directed through complementary base pairing to inhibit mRNA translation of genes, there can be nonspecific effects on nontargeted genes following oligo-mediated bcl-2 suppression. The effectiveness of bcl-2 and overall apoptosis activity is highly regulated and dependent upon the expression of many stimulatory, inhibitory stabilizing factors, as well as the ratio between these proteins. As demonstrated, the specific suppression of one apoptosis inhibitory protein (bcl-2) is nonspecifically compensated by the suppression of a nontargeted promoter, caspase-3 [Rubenstein et al. 2011a]. Clinically these types of experiments are important because they suggest that for oligo-mediated bcl-2 suppression to be effective caspase-3 activity should be either maintained or enhanced [Rubenstein et al. 2011a], and perhaps androgen sensitivity suppressed. In these experiments bispecific oligos also targeted the EGFR. However, in previous experiments we found that growth inhibition produced by monospecific oligos directed against EGFR was not accompanied by decreased mRNA expression [Rubenstein et al. 2002]. Currently Genta is conducting a phase III test using oligos (Genasense [Genta, Berkeley Heights, NJ]; oblimersen) directed against bcl-2 for treating melanoma, chronic lymphocytic leukemia and various solid tumors (www.genta.com/Products_and_Pipeline/Genasense/Genasense.html), but compensatory effects produced by this agent have not (yet) been reported.

Tumors are resilient in their efforts to overcome (even newly developed) therapeutics and become resistant. If gene therapy is to be effective, we must understand how primary effects evoke compensatory changes and it would also be significant to see whether these changes are replicated in an in vivo model. Furthermore, it would be important to measure alterations in the cellular concentration of the affected proteins. If suppressive treatment leads to enhanced expression of undesired proteins, the oligo approach can again be applied, employing other bispecific or even (proposed) multivalent forms [Rubenstein et al. 2006] simultaneously targeting many regulatory proteins.