Differential Sensitivity to xCT-inhibition by ME1 expression

ACXT-3102 Generation

Sigma-2 ligand SV119 and Erastin analogue (dm-Erastin) were synthesized according to published methods (39).ACXT-3102 was previously named SW-V 49. The method for synthesis of the sigma-2/dm-Erastin conjugate has been previously described and published by our laboratory(18,21). Briefly, a dm-Erastin precursor with a terminal aldehyde group is created through a series of condensation, cyclization, and substitution reactions. This intermediate is covalently coupled to the primary amine of the sigma-2 ligand SV119 via reductive amination using sodium borohydride, yielding the final target conjugate (SW V49/ACXT-3102) as an oxalate salt. To increase the solubility, the free amine is treated with hydrochloric acid followed by precipitation. For full chemical synthesis, please see previously published methods from our laboratory (18, 21, 39). The HCl salt formulation of ACXT-3102 was used in all in vitro and in vivo experimentation described in this paper. ACXT-3102 vehicle is dissolved in isotonic saline – 0.9% NaCl without the need for additional solvent. Equivalent volume of isotonic saline – 0.9% NaCl was used as control for all in vitro and in vivo experimentation.

Patient-derived xenograft mRNA expression analysis

Clinical and sequence data were obtained from the Washington University Development and Trial Centers (WU-PDTC). Samples were validated using various QC metrics, including overall coverage and pedigree consistency. Cancer groups were defined based on Cancer Therapy Evaluation Program (CTEP) diagnoses. RNA-seq data in fastq format were prepared for analysis by trimming adapters and filtering to remove reads mapping uniquely to the mouse. The remaining reads were analyzed using Kallisto(40) to estimate gene-level expression as TPM (transcripts per million) values.

Cell Culture Conditions

Cancer cell lines ASPC1 (RRID: CVCL_0152), HCT-116 (CVCL_0291), RKO (CVCL_0504), SKUT1 (RRID: CVCL_0533), SK-LMS-1 (RRID: CVCL_0628), SHP-77(RRID: CVCL_1693), and NCI-H69 (RRID: CVCL_1579) were obtained in 2018 from the American Type Culture Collection (ATCC, Manassas, VA, Cat. #: CRL-1682, CCL-247, CRL-2577, HTB-115, HTB-88, CRL-2195, and HTB-119). All cells were authenticated by ATCC prior to receipt. All cell lines were cultured in a humidified incubator at 37°C in 5% CO₂. ASPC1 and NCI-H69 cell lines were cultured in RPMI-1640 media (Sigma Aldrich, St. Louis, MO, #30-2011) supplemented with 10% FBS (Atlanta Biologicals, Flowery Branch, GA, #S11150). RKO, SKUT, and SK-LMS-1 cell lines were cultured in Dulbecco Modified Eagle Medium (Sigma Aldrich, St. Louis, MO, #30-2002) supplemented with 10% FBS. SHP-77 cell lines were cultured in RPMI-1640 supplemented with 10% heat-inactivated FBS. All media was treated with a combination of penicillin (100 mg/mL) and streptomycin (100 mg/mL) (Sigma Aldrich, St. Louis, MO, #30-2300). All cell lines were confirmed to be mycoplasma negative with the Lonza MycoAlert Mycoplasma Detection Kit (Lonza, #LT07-118) and tested routinely at minimum three-month intervals according to standard laboratory protocols. All cell lines were cultured for less than 6 months after thawing.

Generation of ME1 knockdown and overexpressing cell lines

The RKO ME1 overexpressing (RKO ME1 OE) cell line was kindly provided by Dr. Van Tine (Washington University). ASPC1 ME1 knockdown (ASPC1 ME1 KD) and HCT-116 ME1 knockdown (HCT-116 ME1 KD) were generated by incubation of cell lines with shRNA ME1 lentiviral transduction particles (Sigma Aldrich, St. Louis, MO), (Mission shRNA–TRCN0000064730) sequence CCGGGCTGAGGTTATAGCTCAGCAACTCGA GTTGCTGAGCTATAACCTCAGCTTTTTG. Cells were tested for sensitivity to puromycin (Sigma Aldrich, St. Louis, MO), and concentrations of puromycin, which resulted in 100% cell death, were identified. Cells were incubated with shRNA ME1 transduction particles according to manufacturer instructions. After incubation with the shRNA vector, cells underwent puromycin selection to confirm appropriate knockout. Knockout was further verified by ProteinSimple Wes^™ as described below.

RKO ME1 overexpressing (OE) cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) with 10% FBS. ASPC1 ME1 KD cells were cultured in RPMI-1640 with 10% FBS. HCT-116 ME1 KD cells were cultured in McCoy’s Media with 10% FBS. Penicillin (100 mg/mL) and streptomycin (100 mg/mL) were added to the media; cells were maintained in a humidified incubator at 37°C with 5% CO₂. KD and OE cell lines underwent regular selection by treatment with puromycin as well as confirmatory Western blot testing for the presence or absence of ME1. Puromycin was withdrawn 24 hours before any metabolic assay was performed. All cell lines were cultured for less than six months and were viable in culture with stable ME1 expression status over multiple passages.

Drug Treatment and Cell Death Assays

Cancer cell lines were plated at a density of 2.5 ×10⁴ cells/well in a clear-bottom 96-well plate for 24 hours before treatment. Cells were stained with IncuCyte NucLight Rapid Red Dye for Live Cell Cytoplasmic Labeling (#4717, Sartorius, Ann Arbor, MI) according to standard manufacturer protocol. The cells were then treated for a designated time (4–24 hours, depending on the experimental endpoint) with ACXT-3102. Untreated cells on each plate served as controls. Cell death assays were conducted with treatment media including 50 nM YOYO-1 Iodide (Cat # Y3601, Invitrogen, Carlsbad, CA). Green YOYO-1 Iodide fluorescence and red nuclear fluorescence were monitored on the IncuCyte S3 live cell imaging system and analyzed using IncuCyte image analysis software. Cell death was quantified by normalizing total green area/mm² to total red area/mm² to account for differences in cell density and then normalized to the starting time point. Percent cell death was calculated using a 100% death endpoint measurement. Quantification of analyses is summarized as mean ± SD, with variance across groups compared for statistical significance by Student’s t-test on Prism 8 software. All assays were completed in triplicate and repeated at three independent time points.

Cell Proliferation Assessment

Cells were plated at a density of 5.0 ×10⁴ cells/well in a clear-bottom 96-well plate for 24 hours before proliferation assessment. Cells were stained with IncuCyte NucLight Rapid Red Dye for Live Cell Nuclear Labeling (Cat #4717, Sartorius, Ann Arbor, MI) according to standard manufacturer protocol. Cells were then treated with relevant compounds, and red nuclear fluorescence was monitored on the IncuCyte S3 live cell imaging system and analyzed using IncuCyte image analysis software. Proliferation was quantified by normalizing the total red area/mm² to the starting time point. All assays were completed in triplicate and repeated at three independent time points.

Immuno Assays

For analysis of protein expression (SantaCruz Biotech, Santa Cruz, CA), cells were seeded at 3 ×10⁵ in a 6-well plate one day before collection. On the day of collection, media was removed, and cells were washed 3x with phosphate-buffered saline (PBS). Cells were collected and lysed per reagent protocol with 1X RIPA buffer (cell lysis buffer) (Cell Signaling Technology, Danvers, MA) with protease inhibitor. Lysates were run on a ProteinSimple Wes Capillary electrophoresis instrument (ProteinSimple, Bio-Techne, Minneapolis, MN), using instrument default settings and manufacturer’s standard protocols for immunoassay and total protein analysis through a CE-SDS protein analysis system thus no electrophoresis bands are generated or demonstrated based on standard protocol for this assay. Anti-ME1 antibody (RRID: AB_10679994) (malic enzyme 1, ab 97445, Abcam, Cambridge, UK) was used at an antibody concentration of 1:50, Protein levels were visualized and quantified using Compass for SimpleWestern software (ProteinSimple, Bio-Techne, Minneapolis, MN). Band density differences are expressed as relative densitometry normalized to total protein. Quantification of analyses is summarized as mean ± SD, with variance across groups compared for statistical significance by Student’s t-test. For tumor-derived lysates, tumors were collected fresh after mouse takedown. Tumors were isolated by dissection and immediately placed on ice. Tumors were lysed with 1X RIPA buffer with protease inhibitor. Lysates were run on a ProteinSimple Wes Capillary electrophoresis instrument using instrument default settings and the manufacturer’s standard protocols. Antibodies were used in the same manner as for cell-derived lysates at equal concentrations. Band density differences are expressed as relative densitometry normalized to total protein. Quantification of analyses is summarized as mean ± SD, with variance across groups compared for statistical significance by Student’s t-test.

Formalin Fixation and Immunohistochemistry Analysis

Tumors collected from orthotopic models were fixed overnight in 10% Neutral Buffer Formalin before being stored in 70% EtOH prior to paraffin embedding. Cleaved Cascade 3 Antibody (Cell Signaling Technology, Cat No. 9664) was used.

In vivo assessment of ACXT-3102 in murine and rat models of cancer.

Mouse xenograft experiments were performed in adherence with the animal studies protocol approved by the Washington University Institutional Animal Care and Use Committee (IACUC). Animals were maintained in pathogen-free conditions following IACUC guidelines (St. Louis, MO). Homozygous athymic nude mice (RRID: IMSR_CRL:490)(6 weeks old, females, Charles River Laboratories, strain #490) were injected in the right flank with 100uL of a single-cell suspension of SHP77, SYO-1, HCT116, ASPC1, ASPC1 ME1 KD, RKO or RKO ME1 OE cells suspended in DMEM media (1.5 × 10⁵ cells per mouse) with 50% matrigel (Corning, Corning, NY). Wild-type C57BL/6NTac (RRID: IMSR_TAC:B6) (6-week-old, females, Taconic Biosciences, Model No. B6-F) were injected in the right flank after hair removal with 100uL of single-cell suspension of KP2 cells suspended in DMEM media (1.5 × 10⁵ cells per mouse) with 50% matrigel (Corning, Corning, NY). Similarly, RNU Nude Rats (RRID: RGD_2312499) (8-week-old females, Charles River Laboratories, strain #316) were injected in the right flank with cellular suspensions of PAUT cells. For orthotopic models, tumors were established by injecting 300,000 KP2 cells into the pancreas of wild-type C57BL/6NTac (RRID: IMSR_TAC:B6) (6-week-old, females, Taconic Biosciences, Model No. B6-F). Six days after surgery, mice were randomized to receive vehicle control or ACXT-3102 (60mg/kg) by oral gavage. After two weeks of continuous treatment, tumors were harvested for downstream analysis.

For subcutaneous models, animals were monitored daily following injection for tumor development. Treatment was begun when the mean tumor volume was 100mm³ as measured by caliper for murine models and 1000 mm³ for rat models. Animals were randomized into treatment groups by sequential occurrence of tumors. Mice and rats were weighed daily and received daily oral gavage with ACXT-3102 (30 mg/kg, 45mg/kg, or 60 mg/kg by experimental group) or equal volume vehicle only (sterile isotonic saline −0.9% NaCl) for control groups. All experimental concentrations of the drug were confirmed by in vitro efficacy studies as described above before use in in vivo models.

Tumors were measured in two dimensions daily with a digital caliper, and tumor volumes were calculated by the standard formula of tumor volume = (length × width²)/2. Graphical representation of data for periods>10 days demonstrates measurements only every third day for clarity. Graphical representation of data for orthotopic models is presented as endpoint data only.

Animals were treated for 10 days for murine dose-response studies or 30 days for efficacy studies and followed to the experimental endpoint. Animals were euthanized by CO₂ chamber according to animal studies protocols when the experimental endpoint, or humane endpoint, was reached (tumor size >2 cm³ in murine models and >2.5cm³ in rat models, tumor ulceration, or weight loss greater than 20% of starting weight).

Statistical analyses

Statistical analyses and data plotting were performed using GraphPad Prism software version 8 (San Diego, CA) and SPSS software version 26 (IBM Corp, Armonk, NY). Results for all in vitro studies are derived from at least three replicates of each data point repeated at three time points. IC50 values were calculated by curve fitting normalized viability versus drug concentration.

In vitro, protein expression and cell death are expressed as mean +/− standard deviation, and a two-tailed student’s t-test was performed for individual group comparison at a single time point and two-way ANOVA testing was performed when multiple time points are analyzed.

Statistical analysis for in vivo study data was performed by analysis of variance for multiple-group comparison or 2-way ANOVA for individual group comparison. Kaplan-Meier survival analysis was used to assess differences in overall survival between treatment groups, and log-rank testing was used to determine the significance. Data for in vivo tumor growth curve analysis is expressed as mean +/− standard error of the mean. The number of mice per group for all in vivo experiments is presented in the results section and figure legends. * p < 0.05, ** p < 0.01, *** p < 0.001 were considered to represent statistically significant group differences.

Research Oversight and Ethics

Animal care approval was provided by the Washington University IACUC and all animal experiments were carried out in compliance with IACUC animal care policies. Use of de-identified patient-derived xenograft samples were obtained from the Washington University in Saint Louis Tissue Bank and were fully de-identified and determined by the IRB to be exempt from written informed consent requirements in accordance with the Declaration of Helsinki and relevant institutional ethical guidelines.

Data Availability

The analyzed TCGA (https://portal.gdc.cancer.gov/), NCI-60 gene expression (https://dtp.cancer.gov/mtweb/browse.jsp), and CCLE data (https://sites.broadinstitute.org/ccle/datasets) within this article ainstire publicly available through respective online resources. For access to the other raw data files associated with figures within this paper, please contact the corresponding author. For access to the xCT inhibitor ACXT-3102, contact Accuronix Therapeutics (bkeller@accuronix.com) or by contacting Dr. William Hawkins (hawkinwi@musc.edu).

ACXT-3102 is Effective Against a Variety of Solid Tumors

As has been demonstrated, ACXT-3102 is a potential chemotherapeutic that shows efficacy in a variety of tumor models (21,36). To determine the potential efficacy of ACXT-3102 in multiple models and establish a minimal therapeutic efficacious dose, ACXT-3102 was tested across multiple solid tumor models. Following tumor growth to 100mm³ for mouse or 1000mm³ for rat studies, all animals were randomized into treatment groups and were dosed according to standard laboratory protocol with once daily oral gavage doses of ACXT-3102 using low dose (LD, 30mg/kg), mid-dose (MD, 40 or 45mg/kg), or high dose (HD, 60mg/kg) protocols for 10 days. Across all experiments, the drug vehicle (isotonic saline −0.9% NaCl) was used as a control. ACXT-3102 did not result in significant weight loss in any xenografted model when compared to vehicle control and was well tolerated for the duration of the study period for all mice. All mice were treated with LD, MD, and HD dosing regimens. Rat lines were dosed with only LD and MD dosing strategies.

In an immunocompetent KP2 mouse model of pancreatic cancer grafted into B6 mice, LD and MD ACXT-3102 failed to show superiority to the vehicle. However, HD ACXT-3102 demonstrated significant tumor control (Fig 1a) after 10 days of treatment. This was additionally demonstrated in a human xenografted model of pancreatic cancer using the ASPC1 cell line (Fig 1b). In this model, LD ACXT-3102 was ineffective; MD and HD ACXT-3102 were able to achieve tumor control with no significant difference between MD and HD strategies. We further tested the efficacy in a non-murine model using PAUT human xenografted pancreatic cancer cells onto a rat model (Fig 1c). We demonstrated similar results, in which tumor control was achieved, but required MD ACXT-3102 without tumor control able to be achieved with LD ACXT-3102. We found similar results in xenografted lung cancer (SHP-77) (Fig 1d) and xenografted colorectal cancer (HCT116) models (Fig 1e). In both cases, only HD regimens were able to achieve tumor control after 10 days of treatment with no significant difference between LD and MD strategies.

One cell line, RKO was identified to demonstrate improved sensitivity to ACXT-3102 in xenografted models. RKO colon cancer xenografted models showed similar durable tumor control with LD strategies without improved control by MD or HD ACXT-3102. (Fig 1g) SYO-1 sarcoma models have previously demonstrated substantial tumor control with LD ACXT-3102, with no additional advantage achieved by HD strategies, which has been previously published by our laboratory (36).

Additional experimentation by an orthotopic model was performed by surgical injection of KP2 cells into the pancreas of an immunocompetent mouse model. After two weeks of continuous treatment, tumors were harvested and endpoint tumor weight (Fig 1g) and volume (Fig 1h) were measured showing statistically significant tumor control using HD ACXT-3102. Representative tumor images are found in Supplemental Figure 1. ME1 expression of KP2 cells was confirmed and histopathological analysis was performed on tumor tissue.

ME1 Deficiency Is Present Within Cell Lines and Correlates to ACXT-3102 Sensitivity within the NCI60

To better understand the demonstrated improved efficacy of ACXT-3102 in these models, RNA sequencing data from the US National Cancer Institute (NCI) Human Tumor Cell Lines Panel (NCI-60) was used as a screening platform to identify expression profiles that sensitize to ACXT-3102 treatment. Due to the previously demonstrated mechanism of ROS-mediated cell death in response to Erastin treatment, we investigated a link between ACXT-3102 and over 100 genes associated with ROS homeostasis (Supplemental Table 1). ACXT-3102 doses required to induce half-maximal cancer cell death in vitro (IC50) were correlated with gene expression across multiple cancer subtypes. A total of seven genes demonstrated a correlation between gene expression and ACXT-3102 sensitivity. Of these, malic enzyme 1 (ME1) demonstrated the strongest significant correlation between drug dose and gene expression level (Fig 2a). To determine the translation of our findings into human tumors, we investigated the expression of ME1 in multiple human tumor models. We quantified the expression of ME1 using mRNA from patient-derived xenograft samples (Fig 2b). As we hypothesized that ME1 expression was the potential explanation for our differential sensitivity demonstrated in vivo, we examined the Cancer Cell Line Encyclopedia (CCLE) to investigate the commonality of ME1-low and non-expressing cell lines. We found a small population of cell lines across multiple solid tumors (pancreas, lung, and colorectal) as well as sarcoma lines, which demonstrated low or non-expression of ME1 (Fig 2c–f). Notably, both cell lines that had demonstrated increased sensitivity to ACXT-3102 in animal models, SYO-1 and RKO, were identified in this search as ME1 non-expressing cell lines.

ME1 Deficiency Defines a Small Population of Solid Tumors and Sarcoma

To further investigate the commonality of this subset of tumors, The Cancer Genome Atlas Program (TCGA) data across several solid tumor and sarcoma samples was examined. Multiple malignancies demonstrated alterations in ME1 expression (Supplemental Figure 2). A small population of pancreatic, colorectal, lung cancer, and sarcoma samples demonstrated deep deletions (3a-d), with multiple tumors also demonstrating shallow deletions. Taken together, these data identify a subset of human tumors with naturally occurring low or absent ME1 expression.

The effects of ME1 on disease-specific mortality was investigated next. Within pancreatic, colorectal, lung, and sarcoma, the 10% of tumor samples expressing the lowest percentage of ME1 expression did not demonstrate significantly increased disease-specific risk as compared with tumor samples within the highest decile of ME1 expression (3e-h).

ME1 Status Determines Sensitivity in vitro to xCT-inhibition mediated cell death

To further investigate this observation, the direct effect of ACXT-3102 on tumor cell death in vitro was determined. the ME1 status of multiple cell lines was measured by WES (Fig 4a). Endogenous ME1 expression was reduced by small hairpin-induced RNA interference (shRNA knockdown, KD) in ME1-expressing ASPC1 (pancreatic cancer) and HCT116 (colon cancer) cell lines. ME1 expression was confirmed to be decreased to 50% and 10% of endogenous protein levels, respectively (Fig 4b). ME1 was re-expressed by stable, ectopic overexpression (OE) via standard DNA transfection in ME1-deficient colon cancer cells, RKO (Fig 4b). Therefore, the functional implications of reduced ME1 expression on ACXT-3102 sensitivity in isogenic cell line pairs were examined by cell death assays on the IncuCyte S3. These findings mirrored the in vivo results and demonstrated increased sensitivity to ACXT-3102-induced cell death mediated by ME1 expression. There was a significant increase in sensitivity among ASPC1 ME1 KD (Fig 4c) and HCT116 ME1 KD (Fig 4d) cells as compared to parent ME1 wild-type lines. Additionally, the reintroduction of ME1 expression into RKO cells demonstrated increased resistance to ACXT-3102-induced cell death (Fig 4e). These findings demonstrate that ME1 expression negatively correlates with sensitivity to ACXT-3102 in vitro. xCT inhibition does not seem to result in increased ME1 expression among ME1-expressing cells as a compensatory mechanism of resistance to ACXT-3102 (Supplemental Figure 3).

Induced ME1 Overexpression Abrogates xCT-inhibition sensitivity in vivo

Paired ME1-expressing and non-expressing cell lines were injected subcutaneously into nude mice. Daily ACXT-3102 treatment via oral gavage was initiated when tumors reached a volume of 100mm³. The vehicle (isotonic saline - 0.9% NaCl) was used as a control. Tumors were dosed using laboratory standard HD (60mg/kg) or LD (30mg/kg) dosing, as had been proven effective previously in xenografted tumor models.

ASPC1 ME1 KD tumors demonstrated significantly improved control compared to ASPC1 WT tumors (Fig 5a–b). additionally investigated this finding using RKO WT and RKO ME1 OE tumors. RKO WT tumors demonstrated significant sensitivity to ACXT-3102 with durable treatment response over 30 days of treatment using LD ACXT-3102 (30mg/kg) (Fig 5c). Similar 30mg/kg dosing of the RKO ME1 OE tumors demonstrated no tumor control (Fig 5d). Tumor growth reduction was demonstrated with HD, 60mg/kg, ACXT-3102 (Fig 5e). The persistence of the ME1 expression phenotype in the RKO ME1 OE tumors was confirmed after in vivo implantation, indicating the stability of the over-expression model (Fig 5f). Following 30 days of treatment, all mice were followed to their survival endpoint. A substantial extension of median survival of nearly three-fold (21 days v. 59 days) was seen in the treatment group compared to the control group among ME1-deficient RKO-WT tumors using LD ACXT-3102 (Fig 5g). RKO ME1 OE mice dosed with LD ACXT-3102 strategies experienced no improvement in survival compared to control mice (17 days v. 17 days). However, HD ACXT-3102 strategies were able to demonstrate improved survival in the ME1-expressing tumors, though with less robust results than the survival seen in ME1-null tumors (17 days v. 35 days) (Fig 5h). Overall, these findings demonstrate that ME1-null tumors demonstrate improved tumor control and survival after ACXT-3102 with lower dosing strategies as compared to ME1-expressing lines in vivo.