Genomically informed small-molecule drugs overcome resistance to a sustained-release formulation of an engineered death receptor agonist in patient-derived tumor models

Abstract

Extrinsic pathway agonists have failed repeatedly in the clinic for three core reasons: Inefficient ligand-induced receptor multimerization, poor pharmacokinetic properties, and tumor intrinsic resistance. Here, we address these factors by (i) using a highly potent death receptor agonist (DRA), (ii) developing an injectable depot for sustained DRA delivery, and (iii) leveraging a CRISPR-Cas9 knockout screen in DRA-resistant colorectal cancer (CRC) cells to identify functional drivers of resistance. Pharmacological blockade of XIAP and BCL-X_L by targeted small-molecule drugs strongly enhanced the antitumor activity of DRA in CRC cell lines. Recombinant fusion of the DRA to a thermally responsive elastin-like polypeptide (ELP) creates a gel-like depot upon subcutaneous injection that abolishes tumors in DRA-sensitive Colo205 mouse xenografts. Combination of ELP_depot-DRA with BCL-X_L and/or XIAP inhibitors led to tumor growth inhibition and extended survival in DRA-resistant patient-derived xenografts. This strategy provides a precision medicine approach to overcome similar challenges with other protein-based cancer therapies.

Fig. 1. Multivalent proapoptotic DRAs can induce cell death in TRAIL-sensitive and TRAIL-resistant human CRC cell lines.

(A) The DRA is composed of oligomers of the third type III fibronectin domain of tenascin engineered to bind DR5 with a K_D (dissociation constant) = 43 nM for the DRA monomer, linked by flexible glycine-serine linkers (G₄S)₃ and expressed recombinantly in E. coli (10). Monomers and dimers do not induce cell death (top), whereas the hexameric DRA increases apoptotic signaling in sensitive cell lines (bottom). (B and C) Human CRC cell lines’ responses to DRA treatment are superior to TRAIL in TRAIL-sensitive cell lines HCT116 and Colo205. (D) DRAs can induce cell death in TRAIL-resistant cell line HT29 cells. (E) RKO cells are resistant to both TRAIL and DRA treatment.

Fig. 2. DRA-resistant human CRC cells are capable of undergoing apoptosis.

(A) mRNA expression data were obtained from the CCLE for three DRA-resistant and six DRA-sensitive CRC cell lines. A heatmap was created to visualize the data for the expression of nine proapoptotic (bold italic) and six antiapoptotic genes (italic) in each cell line. (B) Percent RKO cells positively stained for annexin V after treatment with etoposide for 48 hours. Annexin V binding to dimethyl sulfoxide (DMSO) control–treated cells is statistically significantly different from treatment with 25 or 50 μM etoposide. (C) Percent CRC247 cells positively stained for annexin V after treatment with etoposide for 48 hours. Annexin V binding to DMSO control–treated cells is statistically significantly different from treatment with 5, 25, or 50 μM etoposide. **P = 0.001 and ***P = 0.0001 as analyzed by one-way analysis of variance (ANOVA), followed by Tukey’s post hoc test.

Fig. 3. CRISPR-Cas9 knockout screen reveals the genetic drivers of resistance to DRA, confirmed by in vitro cytotoxicity testing of screen-informed drug combinations.

(A) Overview of LOF CRISPR screen experimental setup. First, sgRNA and Cas9 expression, packaging, and envelope plasmids were transfected into 293T cells, and lentiviral particles were harvested. The RKO cell line was then infected with the pooled lentiviral library, and puromycin selection was completed. The transfected cell line was subcultured for each desired condition (in duplicate), and treatment continued for 2 to 3 weeks, after which DNA was extracted from all samples. Constructs were barcoded by polymerase chain reaction (PCR) and sent for Illumina sequencing. (B) Results of TRAIL LOF CRISPR screen in RKO cells. (C) Results of DRA LOF CRISPR screen in RKO cells. Gray box indicates genes with depletion metric scores <0.8. Each dot represents a gene and is plotted on the depletion metric of each of its two replicates (replicate 1 on the x axis and replicate 2 on the y axis). Red dots indicate common hits between TRAIL and DRA screens. Blue dots indicate hits uniquely generated in the DRA screen. (D) Cell viability assay results of combination treatment with the CDK4/6 inhibitor Palbociclib, XIAP inhibitor BV6, BCL-X_L inhibitor WEHI-539, and DRA in RKO cells and three human patient-derived cell lines (DRA concentration on the x axis and cell viability on the y axis). (E) Flow cytometry data show increased cytotoxicity (positive annexin V staining) for combination treatment conditions in RKO cells. A-1155463 (A-11) is a BCL-X_L inhibitor (23). One-way ANOVA followed by Bonferroni multiple comparisons test was used to establish significance between A-11/BV6 (dark gray) and A-11/BV6/DRA for both DRA concentrations (red). (F) Clonogenic 2D growth assay experiments of >7 days in CRC247 cells indicate markedly slower cell growth upon treatment with DRA in combination with increasing doses of A-1155463 and BV6 each at 0, 50, and 150 nM (left to right) and 0, 3.7, or 7.5 pM DRA (top to bottom). Percent colony area has been graphed for visual purposes below the primary result. For all panels, error bars show SEM. ****P < 0.0001.

Fig. 4. ELP_depot-DRA fusions form gel-like depots at body temperature and abolish tumors in vivo.

(A) A hydrophobic ELP (35) was fused to DRA (blue) for a depot-forming formulation (ELP_depot-DRA), and a hydrophilic ELP (purple) was fused to DRA as a soluble, non–depot-forming molecular weight–matched control (ELP_soluble-DRA). (B) Optical turbidity, measured at 350 nm (OD₃₅₀), demonstrates phase transition of ELP_depot-DRA at 27.9°C, while ELP_soluble-DRA remains soluble up to ~60°C. (C) Cell viability assay data for DRA-sensitive Colo205 cells show that ELP_soluble-DRA, ELP_depot-DRA, and DRA lead to similar in vitro cytotoxicity. (D to F) Colo205 subcutaneous xenografts were injected once on day 0 with ELP_depot-DRA, ELP_soluble-DRA, soluble DRA, TRAIL, or vehicle (n = 8 per group). All drugs were injected intratumorally. (E) Tumor growth data, shown as tumor volume versus time. Data were analyzed using two-way ANOVA of matched values, followed by Fisher’s least significant difference (LSD) multiple comparisons test to establish significance (P < 0.05) of the difference between groups at each day of treatment. Results indicate statistically significant differences in tumor volumes between and including days 9 and 18 for depot-forming ELP_depot-DRA compared with other groups, including soluble ELP_soluble-DRA. (F) Kaplan-Meier survival results demonstrate prolonged survival for mice treated with depot-forming ELP_depot-DRA formulation. Evaluation of survival data with log-rank test suggests significant differences (*P < 0.05) between ELP_depot-DRA and other treatment groups, with approximately 16 days increased median survival for the slow-release formulation compared with the soluble version.

Fig. 5. Rationally designed drug combinations overcome ELP_depot-DRA resistance in PDX.

(A) Cell viability data for CRC247 show efficacy of triple drug treatment with A-1331852 (A in figure legend), BV6 (B in figure legend), and DRA (A + B + DRA; green) compared with double drug treatments (red). Data were analyzed using two-way ANOVA of matched values to establish significance (*P < 0.05) of the difference between A + B + DRA and DRA. (B to D) CRC247 PDX data demonstrate in vivo efficacy of the A + B + ELP_depot-DRA compared with other treatment groups. DRA formulation used in these plots was ELP_depot-DRA and abbreviated to “DRA” in figure legends in (B) to (D). (B and C) Tumor growth inhibition data; mice (n = 7 per group) were treated with A-1331852 (25 mg/kg, daily, po) and/or BV6 (5 mg/kg, q4d) and/or ELP_depot-DRA (30 mg/kg, weekly, sc). Data were analyzed using two-way ANOVA of matched values, followed by Fisher’s LSD multiple comparisons test to establish significance (*P < 0.05) of the difference between groups at each day of treatment. Results indicate statistically significant tumor volumes between and including days 5 and 13 for triple combination of A-1331852 + BV6 + ELP_depot-DRA compared with every other group. According to P values obtained from a two-way ANOVA followed by Fisher’s LSD test, the tumor sizes of the mice in the ELP_depot-DRA + A + B group are statistically significantly different from those of the mice in the ELP_depot-DRA group (*P < 0.05) from days 2 to 13. (D) Kaplan-Meier survival analysis comparing key treatment groups indicates that median survival increases from 29 to 38 days when BV6 is added to the A-1331852 + ELP_depot-DRA combination. A Gehan-Beslow-Wilcoxon test demonstrated statistically significant difference in survival between all single-drug treatment groups and triple drug combination A-1331852 + BV6 + ELP_depot-DRA (*P < 0.05).