Actinomycin D and Telmisartan Combination Targets Lung Cancer Stem Cells Through the Wnt/Beta Catenin Pathway

Abstract

The failure of lung cancer treatments has been attributed mostly to the development of drug resistance, however the underlying cellular and molecular mechanisms are poorly understood. Cancer initiating stem cells (CSCs), present in tumors in a small percentage, play critical roles in the development of drug resistance, metastasis, and cancer relapse. Hence, novel treatments targeting both bulk cancer cells and CSCs are under intense investigation. Herein, we report that lung cancer cells grown on a 3D fibrous scaffold form tumoroids that resemble in vivo tumors, expand CSCs, and provide a platform to identify anti-CSC drugs. The screening of an NCI library of FDA-approved drugs using tumoroid cultures led to identification of Actinomycin D (AD) as a top CSC inhibitor. Since CSCs are mostly resident in the tumor's inner core, AD was combined with an angiotensin receptor antagonist, Telmisartan (TS), which is known to increase drug permeability in tumors and was shown to have anti-CSC activity. Our results showed that AD + TS administered intra-tumorally was significantly more effective than either drug alone in both syngeneic and xenograft mouse models. The results of mechanistic studies revealed that CSC expansion in tumoroids was associated with activation of β catenin signaling and that AD + TS treatment reduced active β catenin levels in tumors. Together, these results establish the utility of the tumoroid culture system to expand CSCs ex vivo for targeted drug screening, to identify promising novel treatments with both anti-CSC and anti-cancer effects, and to individualize treatments for metastatic drug resistant lung cancer patients.

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1

(A) Fluorescence micrographs of lung cancer cell lines (LLC1 and A549) cultured on 3D scaffold. Cells (5,000/well) were seeded onto scaffold matrix in 96 well plates and NucBlue stained cells visualized by fluorescent microscopy on day 6. (B–F) Cells were cultured on scaffold as described in (A). On day 6, RNA was isolated from cell pellets, and CSC marker gene expression in monolayer vs 3D scaffold cultures was assayed by qPCR. ANOVA (Dunnett) N = 3, *p ≤ 0.05. LLC1 (B) and A549 (C). ALDH enzyme activity using the ALDE Fluor kit (Stem Cell Technologies) (D). Average ALDH activity in LLC1 cells cultured on 3D scaffold over multiple 6 day generations. ANOVA (Dunnett) N = 3, *p ≤ 0.05 (E). ALDH high expressing LLC1 cells were isolated from the parent scaffold population by FACS (BD FACS Aria) and cultured in low attachment conditions at a concentration of 2 cells/µL media for 6 days. Sphere efficiency is shown. One way ANOVA (Tukey) was used to calculate significance. N = 3, *p ≤ 0.05 (F). (G–I) Tumor initiation ability of CSC population isolated from LLC1 or A549 cells. ALDH high LLC cells were injected subcutaneously into both flanks of C57/BL6 mice using 10,000 cells per flank and the size of the resulting tumors was measured by caliper (N = 2) (G). 100,000 LLC1 cells obtained from 3rd generation tumoroid culture were injected subcutaneously into both flanks of C57/BL6 mice and tumor growth was compared to 1 million LLC1 cells obtained from monolayer culture. The best fit growth curves for monolayer and 3rd generation tumoroid derived tumors were determined using the exponential growth equation Y = Y0 * exp(k * X). The growth rate constants for each curve (K) were compared using the Extra sum of squares F test. N = 3 (H). CD44+/24− A549 cells were injected subcutaneously into the flanks of Nod/SCIID immunocompromised mice using the cell numbers indicated for each group (N = 3) (I). The number above each star represents the p value obtained for that comparison.

Figure 2

(A) A qPCR array was used to identify changes in RNA expression for CSC related genes in first, second, and third generation LLC1 scaffold culture. Data was normalized to LLC1 monolayer and fold change values are presented as a table. (B,D) Nos2 (B) and CSC related gene expression in H1299, A549, H460 and LLC1 cultured on scaffold for 6 days normalized to monolayer by qPCR (D). Data represents increase in gene expression in tumoroids as compared to monolayer. One Way ANOVA (Dunnett) was used to calculate significance (Prism) N = 3, *p ≤ 0.05. (C) ICC for Nos2 protein in 1st Gen LLC1 tumoroids fixed on day 6 of culture (600X magnification). The number above each star represents the p value obtained for that comparison.

Figure 3

(A) Dose response curves for mouse and human lung cancer cell lines cultured on monolayer vs 3D scaffold. Cells were plated on scaffold in 96 well plates at a density of 5,000 cells per well. Serial dilutions of AD were added on day 4 and cell viability was assayed on day 6 by CellTiter-Glo. (Promega) N = 3. (B) Sphere formation efficiency of sorted ALDH high LLC1 cells with or without AD treatment (0.5 nM) Student’s t-test N = 3, *p ≤ 0.05. (C,D). ALDH activity of LLC1 cells cultured as spheres on low attachment plates with or without 3 nM AD treatment (C) or as tumoroids on 3D scaffold after treatment with escalating doses of AD (D), as assayed by flow cytometry using ALDE Fluor kit. (E) ALDH activity of LLC1 cells cultured as tumoroids was assayed after 48 hr treatment with cisplatin (20 µM) or paclitaxel (50 µM). (F) Dose response curves for the chemotherapeutic drugs etoposide, cisplatin, and paclitaxel compared with AD in the LLC1 and A549 cell lines cultured either as monolayer (M) or on 3D scaffold (S) N = 3. The number above each star represents the p value obtained for that comparison.

Figure 4

(A) IC50 values were determined for cell lines cultured as tumoroids on scaffold treated with varying concentrations of AD in the presence or absence of TS (10 µM). Drugs were added on day 4 of culture and viability was assayed on day 6 using CellTiter-Glo. Combination index was calculated using CompuSyn Software by comparing the effects of AD and TS alone vs in combination. Dose response to TS alone in monolayer cultures is shown along with the comparison between AD and AD + TS treatment response in LLC1 monolayer culture. IC50 and combination index values are provided. (B) LLC1 cells were cultured as tumoroids on scaffold. On day 4 wells were treated with either AD, TS, or both in combination. On day 6 cells on scaffold were stained with NucBlue dye and examined by fluorescent microscopy. (C) LLC1 cells cultured and treated as in (B) were collected on day 6 of culture and ALDH activity was assayed by ALDE Fluor kit. (D,E) LLC1 cells were cultured and treated as in (A) and as monolayer culture. Cells were treated on day 4 and collected on day 6 (48 hr treatment), and expression of CSC related genes were assayed by qPCR. ANOVA(Dunnett) N = 3, *p ≤ 0.05 (D). Expression of PARP and Caspase3 (both full length and cleaved) was assayed using Wes (E). (F) Production of ROS was assayed in LLC1 monolayer cultures treated with 0.4 nM AD, 1 µM TS, or the combination for 48 hr. CM-H2DCFDA staining was assessed using fluorescent microscopy (Fig. S11) (while brightfield images were also collected for each field as a reference) and by flow cytometry using the FITC channel. The number above each star represents the p value obtained for that comparison.

Figure 5

(A–C)In vivo tumor treatment. Cells were injected subcutaneously into the flanks of mice, 1 million LLC1 monolayer cells into C57BL/6 (A), 5 million A549 monolayer cells into Nod/SCIID mice (B) or 100,000 LLC1 3rd generation tumoroids into C57BL/6 (C). Mice were treated with 50 µg/kg AD, 1 mg/kg TS, or combination by intratumoral injection every 3 days starting when tumors reached 3 mm in diameter. Tumor growth was monitored by caliper measurement and mice were sacrificed when tumors reached 10 mm in diameter. For all experiments N = 3 control, 4 AD, 4 TS, 7 combination. ANOVA (Tukey), *p ≤ 0.05. (D) Tumors initiated with LLC1 3rd generation scaffold cells were assayed for ALDH activity using ALDE Fluor kit following treatment described in (C). The number above each star represents the p value obtained for that comparison.

Figure 6

(A) LLC1 cells were cultured as monolayer or on scaffold. On day 2 of culture cells were transfected with a β catenin reporter plasmid and luminescence was assayed at the indicated timepoints (top panel). LLC1 monolayer were also transfected with β catenin reporter plasmid and treated with TS (1 µM), AD (0.3 nM) or combination after 24 hr (bottom panel). Luminescence was assayed at 72 hr. ANOVA (Dunnett) N = 3, *p ≤ 0.05. (B) Relative protein abundance and phosphorylation for Wnt pathway proteins in drug treated LLC1 tumors. Protein was collected from drug treated LLC1 tumors (Fig. 5C) and a phosphor-antibody array assay was performed to determine changes occurring in the Wnt pathway (Full Moon Biosystems). Cy-3 developed array chips were scanned, and images were analyzed using ImageJ to determine relative normalized staining intensity for each antibody. Fold change for select proteins is shown. ANOVA(Dunnett) N = 3, *p ≤ 0.05. (C) LLC1 cells were cultured on scaffold. On day 4 of culture cells were treated with the Wnt pathway inhibitors, PRI-724 or XAV-939 at the indicated concentrations. ALDH activity was assayed after 48 hr treatment. (D) Sphere formation efficiency was determined for LLC1 cells cultured and treated as in (C). ANOVA (Dunnett) N = 3, *p ≤ 0.05. The number above each star represents the p value obtained for that comparison.

Figure 7

Summary of changes to the Wnt pathway as a result of AD + TS treatment.