The AKT/BCL-2 Axis Mediates Survival of Uterine Leiomyoma in a Novel 3D Spheroid Model

Abstract

A deeper understanding of the pathways that drive uterine leiomyoma (ULM) growth and survival requires model systems that more closely mimic the in vivo tumors. This would provide new insights into developing effective therapeutic strategies for these common benign tumors of childbearing-aged women. In this study, we examined the role of BCL-2 in mediating ULM survival in the context of increased protein kinase B (AKT) and oxidative stress using a three-dimensional (3D), spheroid-based model that more closely resembles the native ULM tumor microenvironment. Human primary cells from matched myometrium (MM) and ULM tissues were used to establish spheroid cultures in vitro. Histological and immunohistochemical methods were used to assess the spheroid architecture and characteristics. Viability assays for 3D cultures were used to evaluate their response to BH3 mimetics and the superoxide inducer, paraquat (PQ). Primary MM and ULM cells formed spheroids in culture. Notably, ULM spheroids exhibited low proliferation, increased oxidative stress, and secretion of interstitial collagen. Knockdown studies revealed that AKT sustained BCL-2 expression in ULM. The targeting of BCL-2 with BH3 mimetics effectively reduced viability and induced apoptosis in a subset of ULM spheroids. ULM spheroids that did not respond to BH3 mimetics alone responded to combination treatment with PQ. In conclusion, BCL-2 mediates AKT survival of ULM, providing compelling evidence for further evaluation of BH3 mimetics for ULM treatment. ULM spheroids recapitulated intrinsic features of the native ULM tumor microenvironment and can be used as a model for preclinical testing of potential therapeutic options for ULM.

BCL-2 mediates survival of primary uterine leiomyoma spheroids amidst increased oxidative stress.

By the age of 50, between 70% and 80% of childbearing-age women have developed uterine leiomyomas (ULMs), the most common benign tumors of the female reproductive system. Despite their benign nature, symptomatic ULMs cause heavy vaginal bleeding, abdominal pain, and reproductive abnormalities in ∼25% of subjects (1). No effective, long-term pharmacological treatments are currently available for ULM, and hysterectomy is still the only definitive therapy (2).

The biology underlying these highly fibrotic, collagen-rich tumors still remains fragmentary. Model systems for studying this disease are mostly limited to primary cell cultures in monolayers, xenografts, and transgenic mouse models. Each of these approaches has provided valuable insights into ULM pathophysiology, but all have their own limitations. In the past decade, there has been a growing interest in three-dimensional (3D) culture systems, as they better mimic the complexity and heterogeneity of in vivo tumors compared with two-dimensional (2D) cell culture models (3). In the ULM research field, a previous study has described a method to generate 3D cultures of ULM by embedding immortalized human ULM cells in collagen gels. Authors showed that these structures recapitulate the molecular phenotype of original tissues and produce extracellular matrix (ECM) (4).

Among all 3D systems under development, spheroid cultures have become increasingly popular in cancer research because of their ease to use and reproducibility properties. Spheroids are spherical aggregates of cells that display chemical gradients of nutrients, catabolites, and oxygen that are similar to those found in tumors (5–7). These elements have made spheroids valuable platforms for in vitro screening of potential anticancer drugs and essential tools that can fill the gap between 2D cultures and animal models in preclinical drug discovery studies (3). Here, we have established a spheroid-based 3D model of ULM that recapitulates important clinical features of the tumors with the ultimate goal of translating molecular mechanisms into the development of therapeutics for ULM treatment.

Our group has recently demonstrated that ULMs feature a dysregulated antioxidant system that increases oxidative stress, leading to protein kinase B (AKT)–driven ULM survival (8). We now hypothesize that the BCL-2 family mediates AKT-driven survival in ULM. It has been previously shown that ULMs express high levels of anti-apoptotic BCL-2 protein, suggesting a role for BCL-2 in ULM survival (9–11). Upregulation of the BCL-2 intrinsic apoptotic pathway is one of the preferred strategies deployed by tumors to delay or escape apoptosis (12). By binding the BH3 domain of prodeath BCL-2 family proteins (e.g., BAX, BAK), BCL-2 prevents permeabilization of the outer mitochondrial membrane (OMM), thus inhibiting apoptosis (13). In the current study, with the use of our newly established spheroid cultures, we demonstrate that BCL-2 can mediate survival of ULM and that BCL-2 is regulated by AKT. Moreover, we found that inhibitors of the BCL-2 pathway affect ULM survival in a patient-dependent manner and that the efficacy of BCL-2 inhibitors is enhanced by inducing superoxide anions with the redox cycling agent, paraquat (PQ).

Materials and Methods

Human tissue collection and primary cell isolation

Human matched myometrium (MM) and ULM tissues were obtained from premenopausal women undergoing hysterectomy at the Northwestern University Prentice Women’s Hospital (Chicago, IL). Samples were collected with informed consent, according to an International Review Board–approved protocol. Women included in the study did not receive hormonal therapies in the last 3 months before surgery. Tissues were digested, and primary MM and ULM cells were isolated and cultured, as previously described (14).

Spheroid cultures and drug treatments

To generate spheroids, primary MM and ULM cells at Passage One were trypsinized, filtered through a 70-µm pore-size cell strainer (Corning, Corning, NY), and cultured in 96-well ultra-low attachment plates (Corning) in mesenchymal stem cell growth medium (MSCGM; Lonza, Walkersville, MD) at the indicated density in a 5% CO₂ atmosphere at 37°C. Two days after cell seeding, 50% of medium was replaced with fresh medium, leaving the spheroids undisturbed. After 2 additional days, spheroids were treated with ABT-199 or ABT-263 (provided by AbbVie Inc., North Chicago, IL) or PQ (Sigma-Aldrich, St. Louis, MO) for 72 hours at the indicated concentrations. Control spheroids were treated with an equivalent amount of the vehicle dimethyl sulfoxide.

Histological and immunohistochemical spheroid analysis and tissue microarray staining

Spheroids were fixed in 4% paraformaldehyde for 2 hours, washed in 50% and 70% ethanol, and transferred into the cap of an Eppendorf tube that was used as an embedding mold. A 0.5% agarose solution was then poured over the mold and allowed to solidify. The spheroid/agarose blocks were then processed, paraffin embedded, and sectioned to 4 µm-thick slices. Hematoxylin and eosin (H&E) and immunohistochemical stainings were performed at the Robert H. Lurie Comprehensive Cancer Center’s Pathology Core Facility. Anti-Ki-67 [M724029-2; Agilent, Santa Clara, CA; Research Resource Identifier (RRID): AB_2687528], anti–3-nitrotyrosine (3-NO; 06-284; EMD Millipore, Billerica, MA; RRID: AB_310089), anti–cleaved caspase 3 (anti-CC3; 9661; Cell Signaling Technology, Danvers, MA; RRID: AB_2341188), and anti–BCL-2 (PA5-20068; Thermo Fisher Scientific, Waltham, MA; RRID: AB_11152761) antibodies were used at 1:100, 1:1500, 1:200, and 1:500 dilutions, respectively. Tissue microarray (TMA) immunostaining was performed, as previously described (8), using anti–BCL-2 (1:500 dilution; PA5-20068; Thermo Fisher Scientific; RRID: AB_11152761) and anti-BAX (1:100 dilution; 5023S; Cell Signaling Technology; RRID: AB_10557411) antibodies. Intensity quantitation of the CC3 signal from immunostained samples was performed using the Color Deconvolution plug-in of ImageJ (Fiji version; National Institutes of Health). Average pixel intensity was measured in the 3,3′-diaminobenzidine channel of immunohistochemistry (IHC) images and then converted into optical density (OD) values using the formula OD = log(Max Intensity/Mean Intensity), where Max Intensity is equal to 255 for eight-bit images. OD values were expressed as relative OD.

Spheroid immunofluorescence staining

Spheroids were grown in 96-well black/clear-bottom, ultra-low attachment plates. Per each well, 37% formaldehyde solution was directly added to one-tenth volume of medium, and the plate was incubated for 30 minutes at room temperature (RT). Fixative was discarded and spheroids washed twice with phosphate-buffered saline (PBS) and incubated with 3% bovine serum albumin in 0.1% Triton X-100/PBS for 30 minutes at RT. Wells were then washed twice with 0.1% Triton X-100/PBS for 5 minutes each. Primary α--smooth muscle actin (α-SMA) antibody (1:200 dilution; NB600531; Novus Biologicals, Littleton, CO) in 1% bovine serum albumin/PBS was added to each well and incubated at RT for 2 hours. Next, wells were washed twice with 0.1% Triton X-100/PBS for 5 minutes each, and secondary Alexa Fluor 488 antibody (1:500 dilution; Thermo Fisher Scientific) was added. After 60 minutes at RT, wells were washed twice with 0.1% Triton X-100/PBS for 5 minutes each, and 1 µg/ml 4′,6-diamidino-2-phenylindole was added to each well for 5 minutes to stain the nuclei. Spheroids were imaged using a Leica DM500 B fluorescence microscope.

Spheroid live/dead cell imaging and viability assay

Live and dead cells within spheroids were assessed via fluorescence microscopy using the Live/Dead Viability/Cytotoxicity Kit (Invitrogen, Carlsbad, CA), according to the manufacturer’s instructions. Green fluorescent calcein acetoxymethyl (AM) stains live cells, whereas red fluorescent ethidium homodimer 1 (EthD-1) stains dead cells. Intensity plot profiles were obtained by drawing a line across the spheroid diameter using the line tool in ImageJ for both red and green channels, and the resulting signal intensities were plotted against the distance in pixels. Following drug treatments, viability of spheroids cultured in 96-well black/clear-bottom, ultra-low attachment plates was measured using the CellTiter-Glo 3D Cell Viability Assay (Promega, Madison, WI), as described in the user’s manual. Luminescence was read using a SpectraMax i3 plate reader (Molecular Devices, San Jose, CA).

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Western blotting

Sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis and Western blotting analysis were performed, as previously described (8). Protein extracts from frozen tissues were prepared by grinding two to three small tissue pieces to a fine powder using mortar and pestle in the presence of liquid nitrogen. The tissue powder, or multiple spheroids, was then transferred to a microcentrifuge tube containing radioimmunoprecipitation assay buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1% Nonidet P-40, 1% Na-deoxycholate, 0.1% SDS), supplemented with 1× protease and phosphatase inhibitors and 1 mM phenylmethylsulfonyl fluoride. Samples were sonicated on ice three times, 5 seconds each time, incubated on ice for 30 minutes, and centrifuged at 13,500 rpm for 15 minutes at 4°C. Supernatants were collected and protein contents measured using the Bradford reagent (Sigma-Aldrich). Anti–BCL-2 (PA5-20068; Thermo Fisher Scientific), anti–phosphorylated AKT (pAKT; S473; 9271; Cell Signaling Technology), anti–pan-AKT (4691; Cell Signaling Technology), anti-AKT1 (2932; Cell Signaling Technology), anti-AKT2 (2964; Cell Signaling Technology), and anti-BAX (2772; Cell Signaling Technology) antibodies were used at 1:1000 dilution and anti-actin (A1972; Sigma-Aldrich) at 1:2000 dilution.

AKT silencing and reverse transcriptase-polymerase chain reaction

Silencing of AKT1, AKT2, and AKT3, using small interfering RNA (siRNA), RNA isolation, and reverse transcriptase-polymerase chain reaction (RT-PCR), was performed as previously described (8).

Statistical analysis

Statistical analysis was performed using Prism software (GraphPad Software). Data are represented as means ± standard deviation (SD). Paired t test was performed when comparing MM with ULM from the same subject; otherwise, one-way analysis of variance or unpaired t test was used. Data from each patient were considered as an independent experiment.

Results

Primary ULM cells form viable spheroids whose architecture is reminiscent of clinical tumors

To generate 3D spheroid cultures of primary MM and ULM cells, increasing numbers of patient-derived MM and ULM cells (from 1000 to 50,000 cells/well) were seeded in ultra-low attachment 96-well plates in MSCGM. Both cell types spontaneously self assembled within a 24-hour period and formed spheroid structures with smooth and continuous borders by day 3, regardless of the initial cell density [Fig. 1(a)]. Based on these results, we chose 50,000 cells/well as optimal starting cell density to obtain spheroids of ∼300 to 500 µm in diameter for further experiments. In Fig. 1(b), spheroids cultured for 4 days were stained for live (calcein AM; green) and dead (EthD-1; red) cells. Fluorescent images and signal intensity plot profiles of calcein AM and EthD-1 stainings indicated high cell viability within both MM and ULM spheroids. Next, H&E staining was performed on spheroid sections to assess general histological features. H&E images showed that ULM spheroids were more cellular than MM spheroids, with smooth muscle cell bundles forming storiform and a short fusiform growth pattern with less organized cells. Moreover, unlike ULM, MM spheroids were encompassed by a well-defined and organized smooth outer layer of cells so as to encapsulate the spheroid core [Fig. 1(c) and Supplemental Fig. 1]. Next, cell proliferation by immunostaining with the proliferation marker Ki-67 revealed that both MM and ULM spheroids contained a low number of Ki-67-positive cells, with ULM spheroids as the least active under the tested conditions [Fig. 1(c)]. Together, these results indicate that primary MM and ULM cells are able to form spheroids whose proliferation rate, cellular architecture, and morphology resemble those of clinical tissues.

Figure 1.

Establishment of ULM cells in spheroid cultures. (a) Primary ULM cells of increasing cell numbers were placed in low adherent plates and cultured in MSCGM. Bright-field (BF) images of MM and ULM spheroids were taken after 4 days of culture. (b) MM and ULM spheroids of 50,000 cells were stained with calcein AM to detect live cells (green fluorescence), as well as EthD-1 (red fluorescence) to detect dead cells. Corresponding intensity plot profiles of the staining are shown. (c) H&E and Ki-67 immunostaining of paraformaldehyde-fixed, paraffin-embedded MM and ULM spheroid sections was done. Black arrowheads indicate positive Ki-67 cells. Original scale bars, 100 µm, unless specified otherwise. (a–c) Representative of at least three patients each.

ULM spheroids share key histological features with the native tissues

α-SMA is a marker of differentiation of smooth muscle cells and a marker for ULM cells (15). To confirm the presence of this cell type in the MM and ULM spheroids, α-SMA staining was performed directly in the ultra-low attachment 96-well plate. Figure 2(a) shows that both MM and ULM spheroids stained strongly positive for α-SMA. Unlike MM, ULMs are highly fibrotic solid masses, characterized by excessive deposition of ECM components, in particular, collagen (16). To assess collagen deposition in ULM spheroids, we performed Masson trichrome staining on 4 mm-thick paraffin-embedded sections from ULM spheroids, cultured for 1 week. Masson trichrome stains red for muscle, blue for collagen, and pink for fibrin. High magnification images of trichrome-stained ULM spheroid sections showed the presence of thin interstitial collagen (blue staining) between cells, suggesting that collagen is being newly synthesized by the ULM cells in situ [Fig. 2(b)]. Aside from ECM deposition, fibrotic diseases are often characterized by increased oxidative stress (17). We recently demonstrated that ULMs feature a dysregulated redox system that results in high levels of reactive oxygen species (ROS) compared with MM (8). To determine whether this property was maintained in 3D structures, ULM spheroid sections were immunostained with an antibody to 3-NO, a well-known marker of oxidative/nitrosative stress (18). ULM spheroids expressed higher levels of 3-NO compared with MM, as shown by increased brown 3,3′-diaminobenzidine chromogen staining, indicating that oxidative stress levels remain persistently elevated in ULM spheroids [Fig. 2(c)]. Altogether, these results demonstrate that ULM spheroids share signature features of the native tissues, including expression of α-SMA, collagen deposition, and enhanced oxidative stress.

Figure 2.

Clinical features of ULM in spheroids. ULM and MM spheroids were cultured for 4 days and were stained with (a) immunofluorescence for α-SMA (green) and 4′,6-diamidino-2-phenylindole (DAPI; blue) to visualize the nuclei; (b) Masson trichrome stain of paraffin-embedded ULM spheroid sections for assessment of newly synthesized collagen (blue stain represents collagen, red stains muscle, and pink stains fibrin); and (c) oxidative/nitrosative stress biomarker 3-NO by IHC. Original scale bars, 100 µm, unless specified otherwise. Bar plot represents immunostaining quantification of 3-NO (n = 3, *P = 0.01, unpaired t test). (a–c) Representative of at least three patients. BF, bright field.

BCL-2 mediates AKT-driven survival in ULM

Tumors with high immunoreactivity for BCL-2 are often resistant to apoptosis and continue to grow and divide (12). Previous studies have shown that ULMs express higher levels of BCL-2 and BAX compared with normal MM but in a small number of specimens (9, 10). Here, we used a TMA containing >200 tissue cores from 60 matched patient-derived MM and ULM tissues (for each patient, three tissue cores were derived from ULM and two from MM). The TMA was stained for the anti–apoptotic BCL-2, as well as proapoptotic BAX proteins [Fig. 3(a)], and scored using a semiquantitative method, as previously described (8). Scoring results revealed significantly higher BCL-2 (P < 0.0001) and BAX (P < 0.001) expression in ULM compared with MM tissues [Fig. 3(b)]. ULMs are also characterized by aberrant activation of AKT signaling (10, 19). Therefore, we next assessed expression of BCL-2, BAX, and pAKT proteins in tissue extracts from 21 matched MM and ULM specimens. A representative Western blot and densitometric analysis of all 21 experiments showed that BCL-2 and BAX levels were higher in ULM and accompanied by increased pAKT expression in most of the samples [Fig. 3(c) and 3(d) and Supplemental Fig. 2a]. As a consequence, the BAX/BCL-2 ratio was augmented in ULM compared with MM, suggesting that ULMs are poised in a state that may render them vulnerable to insults [Fig. 3(e)]. Next, to determine whether there was a correlation between upregulation of the BCL-2/AKT axis and the mediator complex subunit 12 (MED12) mutation status of ULMs, we performed Sanger sequencing of DNA isolated from all 21 pairs of matched MM/ULM specimens. We found that 76% of ULMs harbored mutations in MED12 exon 2; however, there was no correlation with BCL-2 or pAKT expression levels (Supplemental Table 1). Moreover, IHC and Western blotting results from two MM/ULM specimens showed that higher BCL-2 protein levels in the original tissues tracked with higher levels of BCL-2 in both spheroid sections and spheroid protein extracts (Supplemental Fig. 3). Finally, to determine whether higher BCL-2 levels were associated with increased AKT signaling in ULM, AKT was silenced using siRNA against AKT1, AKT2, and AKT3, and BCL-2 levels were assessed by Western blot. As shown in Fig. 3(f) and Supplemental Fig. 2b, silencing of AKT caused a reduction in downstream BCL-2 levels, suggesting that BCL-2 mediates AKT-driven survival in ULM.

Figure 3.

Levels of BCL-2, BAX, and pAKT in ULM primary tumors. (a) Representative images of matched MM/ULM tissues from a TMA section, immunostained with BCL-2 and BAX antibodies, are shown. (b) Intensity of each TMA tissue core was scored using a semiquantitative method, as previously described (8). Samples were categorized into either the moderate/strong (≥2) or the negative/weak (<2) BCL-2 or BAX groups, as shown by the frequency distribution plots (***P = 0.0009; ****P < 0.0001, χ² test). (c) Representative Western blot showing the levels of BCL-2, BAX, and pAKT in protein extracts of matched MM/ULM tissues from a patient, and (d) corresponding densitometric analysis from 21 patients are shown. Total AKT was used as a loading control (*P = 0.03; **P = 0.0011; ***P = 0.0002, paired t test). (e) The BAX/BCL-2 ratio was calculated from the densitometric quantitation of BCL-2 and BAX expression in 21 patients (**P = 0.0055, paired t test). (f) AKT1, AKT2, and AKT3 were silenced in ULM cells using siRNA against each isoform (siAKT1–3), and expression of BCL-2 was assessed by Western blot. A nontargeting siRNA (siCTR) was used as a control. Silencing efficiency was verified by Western blot using anti-AKT1, anti-AKT2, and anti–total AKT antibodies, whereas AKT3 knockdown was determined by RT-PCR using 18S as a reference gene. A representative blot and corresponding densitometric analysis of three independent experiments are shown as means ± SD (n = 3). Anti-actin antibody was used as a loading control (*P < 0.05; ****P < 0.0001, unpaired t test).

BH3 mimetics are effective in decreasing viability of a subgroup of ULMs, whereas the combination of BH3 mimetics and PQ is effective in resistant ULMs

As ULMs feature high levels of anti–apoptotic BCL-2 protein, we investigated how the targeting of BCL-2, using the BH3 mimetics ABT-199 and ABT-263, would affect spheroid viability. Matched MM and ULM spheroids from 14 patients were prepared, as described in Materials and Methods, and treated with increasing doses of ABT-199 and ABT-263 for 72 hours. As a result of natural variation in patient response to treatments, the viability ratio was calculated by dividing the mean values from ULM spheroids by the mean values from the matched MM spheroids. Based on the viability ratio data, responses to BH3 mimetics were stratified into three groups: good response, no response, and toxic response. Viability curves with a ULM/MM ratio close to 1 (0.7 ≤ x ≤ 1.3) were indicative of no response (black) to treatments for either MM or ULM. Curves with a decreased ULM/MM viability ratio (x < 0.7) denoted a good response (blue) to treatments, indicating efficacy in decreasing viability in ULM spheroids and not MM. Curves of an increased ULM/MM viability ratio (x > 1.3) corresponded to a toxic response (red) to treatments, indicating that treatments decreased viability in ULM and also in MM spheroids [Fig. 4(a)]. For ULM and MM spheroids treated with ABT-199, responses were minimal, with the exception of two cases with a toxic response and a partial good response [Fig. 4(b)]. ABT-263 treatment, on the other hand, resulted in a decrease in viability of ULM in eight out of 12 patients, with no response in two cases. Additionally, four cases exhibited increased toxicity (x > 1.3) to ABT-263 [Fig. 4(c)].

Figure 4.

Response of ULM and MM spheroids to BH3 mimetics. (a) Examples of responses to ABT-199 and ABT-263 in MM and ULM spheroids were categorized into three groups. Viability ratio curves (ULM/MM) were calculated, as described in the Materials and Methods section. A good response (GR; blue) to treatments corresponded to decreased ULM/MM ratios (x < 0.7), no response (NR; black) is an ULM/MM ratio close to 1 (0.7 ≤ x ≤1.3), and a toxic response (TR; red) corresponded to ULM/MM viability ratio curves >1.3 (x > 1.3). (b and c) ULM/MM viability ratio curves of spheroids treated for 72 hours with increasing doses of ABT-199 or ABT-263 in 14 patients are shown. Solid black lines were treated with increasing doses of ABT-199 or ABT-263 in combination with 100 µM PQ for 72 hours. (d and e) Spheroids that did not respond to ABT-199 or ABT-263 alone. Dashed lines in (d) and (e) represent the same patients.

We previously showed that ULMs have a dysregulated redox system that renders ULM more sensitive than MM when treated with oxidants (8). The addition of PQ, a well-known redox cycling agent that generates ROS (20), decreased viability preferentially in ULM (8). To determine whether the lack of response to ABT-199 was related to the redox system, spheroids were treated with PQ alone (Supplemental Fig. 2) or with the combination ABT-199 and PQ [Fig. 4(d)]. Viability ratio curves showed that many of the cases that were not sensitive to ABT-199 alone responded to PQ, as well as the combination of ABT-199 with PQ [Fig. 4(d)]. Likewise, the two nonresponsive cases to ABT-263 alone showed sensitivity to ABT-263 + PQ [Fig. 4(e), dashed lines]. Interestingly, these two cases did not respond to either ABT-199 alone or ABT-199 + PQ [Fig. 4(b) and 4(d)] but did respond to ABT-263 + PQ [Fig. 4(e), dashed lines]. We did not find a significant correlation between expression of BCL-2, BAX, pAKT, or MED12 mutation status and the response status to treatments among patients, except that toxic responses to ABT-263 occurred when BCL-2 levels were low in ULM (Supplemental Table 2). Altogether, these results indicate that BH3 mimetics are effective in a subgroup of ULMs and that most of the ULMs that were resistant to BH3 mimetics alone acquired sensitivity to those drugs when challenged with a redox-modulating agent.

Combination of BH3 mimetics and PQ increases CC3 expression in ULM spheroids

Cells undergoing apoptosis are characterized by increased levels of CC3 that serve as a reliable marker of dying cells (21). We have shown that BH3 mimetics alone or in combination with PQ impaired viability of ULM spheroids. To determine whether decreased ULM spheroid viability was a consequence of apoptotic cell death, sections of ULM spheroids treated for 72 hours with BH3 mimetics, alone or in combination with PQ, were immunostained with an antibody against CC3. We observed that ABT-199 and ABT-263 increased levels of CC3 staining compared with vehicle-treated cells. Whereas PQ alone did not significantly enhance the number of CC3-positive cells, the combination of ABT-199 and PQ, and more so with ABT-263 and PQ, increased apoptotic cell death, as shown by high immunoreactivity for CC3 and corresponding IHC intensity quantitation [Fig. 5(a) and 5(b), n = 3].

Figure 5.

Effect of BH3 mimetics and PQ on apoptosis of ULM spheroids. (a) ULM spheroids were treated with 10 µM ABT-199 or ABT-263, alone or in combination with 100 µM PQ for 72 hours. Spheroid sections were immunostained with an anti-CC3 antibody. Representative images are shown. Original scale bar, 100 µm. (b) Quantitation of IHC staining for CC3 in ULM spheroids from three patients was done using the Color Deconvolution plug-in of ImageJ (Fiji version). Results are expressed as means ± SD from three independent experiments (n = 3, *P > 0.05; **** P < 0.0001, one-way analysis of variance).

Discussion

To date, only few pharmacological options are available for women experiencing severe ULM-associated symptoms, with ulipristal acetate being the most promising alternative to invasive surgery (22). However, the development of novel therapeutics for ULM treatment remains an important priority. Currently, laboratory experiments are primarily carried out using adherent, 2D cell monolayers. Although 2D systems have proven to be indispensable tools for assessing the effect of potential therapeutic agents in vitro, their limitations have been increasingly recognized, as they do not fully recreate the in vivo microenvironment (23). To improve on 2D cultures, 3D spheroid systems have been developed and used in drug discovery studies, as they provide more physiologically relevant and more predictive power before in vivo experiments (23). Here, we establish 3D spheroids using primary ULM and MM cells that maintain important intrinsic features of the clinical tumors. These spheroids express α-SMA, estrogen receptor, and progesterone receptor and deposit interstitial collagen. Importantly, as ULM tumors are well-defined solid masses featuring abnormal vasculature, low oxygen content, and higher oxidative stress levels compared with parental healthy MM (8, 24, 25), these characteristics can be better recapitulated in spheroid structures. Indeed, the higher levels of the oxidative/nitrosative stress biomarker 3-NO indicate that high levels of ROS are maintained within ULM spheroids, unlike MM spheroids, which have more efficient mechanisms to clear ROS (8). This feature is important for understanding the survival mechanisms of ULM, as we have demonstrated with the BH3 mimetics and PQ.

Following a diverse array of cytotoxic stimuli, normal cells engage the intrinsic mitochondrial pathway that initiates apoptosis and ensures tissue homeostasis and proper disposal of unwanted cells. Defects in this pathway, which is controlled by the BCL-2 family of proteins, have been associated with multiple diseases, including cancer (26, 27). Upregulation of anti–apoptotic BCL-2 results in abnormal cell survival and apoptosis evasion in a large variety of human cancers, including ULM (9–11). The exact reason for BCL-2 being upregulated in ULM is yet to be determined. We have recently demonstrated that ULM features mitochondrial redox dysregulation that triggers activation of prosurvival AKT (8). Besides being the powerhouse of the cell and the major site of ROS generation, mitochondria are centrally involved in apoptosis (28). Indeed, the OMM is the primary site of action of BCL-2, where it inhibits apoptosis by preventing the dimerization of proapoptotic BAX and Bcl-2 homologous antagonist/killer and consequent formation of pores within the OMM (29). AKT is also known to prevent apoptosis, and this has been ascribed to its ability to phosphorylate proapoptotic BAD (30). Here, we demonstrate a causal relationship between overactive AKT and upregulated BCL-2 expression, as AKT silencing reduced levels of anti–apoptotic BCL-2. Moreover, we found that the BAX/BCL-2 ratio, a well-known determinant of cell susceptibility to apoptosis, was increased in ULM compared with MM as a result of increased BAX levels. An excess of BAX expression over BCL-2 has been shown to increase the formation of BAX–BAX homodimers evoking apoptotic signals for the cells (31). Therefore, although BCL-2 ensures survival in the microenvironment of ULM, higher BAX levels prime ULM for cell death upon stimulation with an apoptotic agent. This is supported by the fact that most ULM spheroids were more sensitive to the BCL-2 inhibitor ABT-263 or PQ compared with MM. BCL-XL is another BCL-2 family member that is capable of inhibiting apoptosis and could also play a role in mediating survival of ULM.

BH3 mimetics are small molecules that mimic the BH3 domain of prodeath, BH3-only proteins (e.g., BID, BIM, BAD) and selectively bind the BH3-binding sites of BCL-2 prosurvival proteins, such as BCL-2, thus neutralizing their anti-apoptotic action (13). Notably, the BH3 mimetic ABT-199 (venetoclax) was approved by the US Food and Drug Administration in 2016 for the treatment of chronic lymphocytic leukemia (32, 33), and ABT-263 (Navitoclax) is currently in a Phase II clinical trial for patients with relapsed small cell lung cancer (34). In our study, the sensitivity of ULM spheroids to ABT-199 was minimal. Interestingly, these spheroids responded to either PQ alone or in combination with ABT-199, suggesting that an alteration of the ULM intracellular redox state influences the response to ABT-199. Likewise, the two nonresponsive cases to ABT-263 alone acquired sensitivity to the combination ABT-263 + PQ. These results highlight the close crosstalk between the intracellular redox state and BCL-2 in ULM. Further investigation of BH3 mimetics, with or without PQ, is warranted, as these compounds are promising candidates for decreasing ULM viability.

Another important feature of ULMs is the naturally occurring senescence, an irreversible arrest of cell proliferation that is often associated with elevated levels of oxidative stress and sustained neoplastic growth if persistently retained in tissues (35, 36). Senescent cells are often accompanied by increased BCL-2 levels and are relatively resistant to apoptotic stressors (26, 37, 38). Interestingly, recent studies have shown that BH3 mimetics can act as senolytic agents capable of preferentially clearing senescent cells from tissues (39, 40). These evidences, together with our data, provide robust support for further development of BH3 mimetics for ULM treatment.

In summary, we have demonstrated that primary ULM cells spontaneously form 3D spheroids that recapitulate innate features of the native ULM microenvironment. We also demonstrated the key role of BCL-2 as a survival factor in ULMs that have an altered redox system and increased AKT signaling.

Glossary

Abbreviations:

2D

two-dimensional

3D

three-dimensional

3-NO

3-nitrotyrosine

AKT

protein kinase B

AM

acetoxymethyl

CC3

cleaved caspase 3

ECM

extracellular matrix

EthD-1

ethidium homodimer 1

H&E

hematoxylin and eosin

IHC

immunohistochemistry

MED12

mediator complex subunit 12

MM

myometrium

MSCGM

mesenchymal stem cell growth medium

OD

optical density

OMM

outer mitochondrial membrane

pAKT

phosphorylated protein kinase B

PBS

phosphate-buffered saline

PQ

paraquat

ROS

reactive oxygen species

RRID

Research Resource Identifier

RT

room temperature

RT-PCR

reverse transcriptase-polymerase chain reaction

SD

standard deviation

SDS

sodium dodecyl sulfate

siRNA

small interfering RNA

TMA

tissue microarray

ULM

uterine leiomyoma

α-SMA

α–smooth muscle actin