Novel mitochondria-based targeting restores responsiveness in therapeutically resistant human lung cancer cells

Abstract

Cisplatin and tyrosine kinase inhibitors (TKIs) are recommended to treat non-small-cell lung cancer (NSCLC). However, ubiquitously acquired drug resistance in NSCLC patients diminishes their therapeutic efficacy. Strategies for overcoming cisplatin and TKI resistance are an unmet medical need. We previously described a group of near-infrared heptamethine carbocyanine fluorescent dyes, referred to as DZ, with tumor-homing properties via differentially expressed organic anion-transporting polypeptides on cancer cells. This group of organic dyes can deliver therapeutic payloads specifically to tumor cells in the form of a chemical conjugate. We synthesized DZ-SIMvastatin (DZ-SIM) initially to target cell membrane cholesterol biosynthesis in lung cancer cells. DZ-SIM induced apoptosis in both cisplatin-sensitive and resistant as well as EGFR TKI-sensitive and resistant lung cancer cells. This conjugate specifically accumulated in and effectively inhibited the growth of xenograft tumors formed by NSCLC cells resistant to first (H1650) and third (PC9AR) generation EGFR TKIs. DZ-SIM induced cell death by targeting mitochondrial structure and function. We concluded that DZ-SIM could be a promising novel therapy for overcoming drug resistance in NSCLC patients.

Introduction

Lung cancer is the leading cause of cancer-related death in the US and China (1,2). Non-small cell lung cancer (NSCLC) is the most common type of lung cancer. Chemotherapy and targeted therapy are the main strategies for advanced NSCLC (3). For patients without drug-targetable gene mutations, the platinum-containing dual-drug regimen is recommended (4,5). Pemetrexed combined with platinum is preferable for adenocarcinoma, while gemcitabine combined with platinum is preferable for squamous carcinoma (6). Cisplatin is an alkylating agent inducing DNA damage and interfering with DNA repair (7,8), widely used in chemotherapy for lung cancer patients. Unfortunately, acquired resistance and severe toxicity lead to frequent therapy failure, metastasis and high rates of recurrence (9,10). The discovery of epidermal growth factor receptor (EGFR) mutation as a therapeutic target is an important advance in the treatment of NSCLC (11,12). For patients with EGFR exon 19 deletion, or L585R substitution in exon 21, the first-generation EGFR tyrosine kinase inhibitor (EGFR-TKI) has great clinical benefit (11,13–15). However, despite good initial clinical response, the inevitable drug resistance still limits long-term response (16–18). Studies showed that the main mechanisms involved in resistance to EGFR-TKIs include the presence of T790M mutation (49%) or amplification (8%), MET amplification (5%), PIK3CA mutation (5%) and conversion from NSCLC to small cell lung cancer (14%) (19). AZD9291 (osimertinib), an FDA approved third generation EGFR-TKI, can selectively and irreversibly inhibit EGFR with the activating mutations as well as the resistant T790M mutation (20). This agent has shown clinical benefit for patients with T790M mutation following disease progression after 1^st and 2^nd generation EGFR-TKIs (21). On the other hand, acquired resistance to AZD9291 has been described in the clinic and confirmed in research (22,23). Novel treatments to overcome chemotherapy resistance are urgently needed.

Mitochondria play an important role both in regulating cell bioenergy production and in therapeutic resistance. Responsible for producing 95% of total cellular ATP through oxidative phosphorylation (OXPHOS) as well as for controlling cell death and survival, mitochondria are critical for cellular function. Cellular bioenergy consists of OXPHOS and aerobic glycolysis essential to cell growth and regulation (24). Metabolic activity in tumor cells is reprogrammed, resulting in the switch from OXPHOS to aerobic glycolysis, known as the Warburg effect (25,26). Though much research has focused on targeting cellular metabolism to inhibit tumor cell growth, improvements in clinical treatments have been elusive (27).

We previously identified a group of near-infrared (NIR) heptamethine carbocyanine organic dyes, referred to as DZ, as effective tumor imaging agents, since these organic dyes are taken up preferentially by cancerous but not normal cells (28). We determined that the specific uptake of DZ was facilitated by the aberrant expression of organic anion-transporting polypeptides (OATPs) in tumor cells, and further enhanced by intratumoral hypoxia through the activation of hypoxia-inducible factor 1α (HIF-1α) (29,30). We further characterized DZ as a dual tumor imaging and tumor targeting agent when chemically conjugated to therapeutic payloads such as docetaxel, gemcitabine, cisplatin or the cholesterol-lowering agent simvastatin (SIM) (28). Critically, we found that these conjugates were mostly cytocidal, inducing rapid apoptosis irrespective of the therapeutic resistance status of the cancer cell to conventional anti-tumor drugs. In this report, we evaluated the anti-tumor effects of DZ-SIM on both cisplatin-resistant and EGFR-TKI resistant cells and AZD9291-resistant xenograft tumors. Our mechanistic investigation indicated that DZ-SIM targeted primarily mitochondria and inhibits mitochondrial respiration, leading to cell death in vitro and in vivo.

Materials and Methods

Cell culture.

NSCLC adenocarcinoma cell lines H1650, H1975, A549, A549DDP, PC9 and PC9AR were used. Cisplatin-sensitive (A549) and resistant (A549DDP) cell lines were kindly provided by the Tianjin Lung Cancer Institute (Tianjin, China). A549DDP was derived by exposing A549 to increased concentrations of cisplatin and then selecting the surviving cells. The H1650, H1975, PC9 and PC9AR cell lines were kindly provided by Dr. Shiyong Sun (Winship Cancer Institute, Emory University School of Medicine, Atlanta, GA). H1650 cells harbor delE746-A750 on exon 19 of the EGFR gene together with PTEN loss and are resistant to EGFR-TKI. H1975 cells with EGFR L585R and T790M acquisition are also resistant to EGFR-TKI. PC9 harbors EGFR delE746-A750 but is sensitive to EGFR-TKI. The AZD9291-resistant cell line, PC9AR, was established by Dr. Sun’s laboratory by exposing PC9 cells to gradually increasing concentrations of AZD9291 for six months (31). The YTMLC-90 lung squamous cancer (LUSC) and H446 small cell lung cancer (SCLC) cell lines were provided by Tianjin Lung Cancer Institute (Tianjin, China). Human peripheral blood mononuclear cells (PBMC) from three healthy donors were isolated by the method of ammonium chloride hemolysis from 7.5 ml anticoagulated (potassium ethylenediaminetetraacetic acid) samples. The use of human samples in research was approved by institutional review board (IRB# Pro00054328); and was in accordance with International Ethical Guidelines for Biomedical Research Involving Human Subject (CIOMS). Informed written consent was obtained from the subjects. These cells were cultured in complete medium (RPMI 1640, Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum (Atlanta Biologicals, Flowery Branch, GA), 100 IU/ml penicillin, and 100 μg/ml streptomycin (ThermoFisher Scientific, Waltham, MA) at 37°C in a humidified incubator supplemented with 5% CO₂. A549DDP was cultured in the presence of 1 µM cisplatin, while PC9AR was grown with 1 µM AZD9291.

Reagents and Instruments.

DZ and DZ-SIM were synthesized in our laboratory and dissolved in dimethylsulfoxide (DMSO) at a concentration of 20 mM and stored at −20°C. Cisplatin, SIM, Gefitinib, Icotinib, and AZD9291 were purchased from MedKoo Biosciences (Morrisville, NC). Annexin V and propidium iodide (PI) were purchased from Biolegend (San Diego, CA). MitoSOX, MitoTracker, LysoTracker, and JC-1 were purchased from ThermoFisher. For the work of chemical synthesis, all chemicals and reagents were purchased from Sigma-Aldrich (St. Louis, MO) and/or ThermoFisher Scientific and used as is since the products derived them were purified and analyzed afterwards. Deionized distilled ultrapure water with resistivity of 18.2 MΩ cm used for making solutions was obtained from Milli-Q Direct Ultrapure Water System from Millipore (Billerica, MA, USA). 1H-NMR data were collected on Bruker 400 MHz spectrometers using standard parameters, chemical shifts are reported in ppm (δ) in reference to residual non-deuterated solvent. Electrospray ionization mass spectrometry (ESI-MS) analysis was performed on new compounds at Cedar-Sinai’s Proteomics & Metabolomics Core using an LTQ Orbitrap Elite mass spectrometer.

Synthesis of DZ-SIM conjugate.

We previously reported experimental setup for the synthesis of conjugates between heptamethine carbocyanine and other molecules (32). The synthesis of DZ is depicted in Supplementary Figure S1. The synthesis began with the preparation of the indolinium salts. Alkylation of 2, 3, 3-trimethylindolenine with 6-bromohexanoic acid or 1,4-butane sultone yielded substituted indolinium salts. The asymmetric cyanine dye DZ was synthesized using a two-step condensation reaction of the appropriate indoleninium salt and Vilsmeier–Haack reagent 2. Briefly, indolinium salt 1a reacted with one equiv. of Vilsmeier-Haack reagent 2 gave a half dye 3 as a major product, which further condensed with compound 1b yielded DZ dye 4. The reaction was catalyzed by sodium acetate using ethanol as the solvent.

The DZ-SIM 7 was synthesized from the modification of SIM 5. The SIM 5 reacted with propane-1,3-diamine in acetonitrile to afford the N-(3-aminopropyl)amide derivative of simvastatin 6 which was conjugated with DZ dye 4 at the side chain terminal carboxylic group via amide bond formation to afford DZ-SIM 7, which was purified by C18 reversed phase chromatography to high purity, and was characterized thoroughly by 1H NMR and mass spectroscopy.

Antibodies.

Primary antibodies to PARP-1 (sc-8007), c-PARP-1 (sc-56196), pH2AX (sc-517348), and pATM (sc-47739) were purchased from Santa Cruz Biotechnology (Dallas, TX). Antibodies to MCL1 (#D2WE), survivin (#71G4B7), and caspase 3 (#9662) were purchased from Cell Signaling Technology (Danvers, MA). HRP-conjugated antibodies were purchased from Santa Cruz Biotechnology.

Cell viability assay.

Cell viability was determined by MTT assay. Briefly, cells were pre-seeded on 96-well plates at concentrations of 8 × 10⁴ cells/ml. After 24 hours of culture, cells in triplicate wells were treated with increasing concentrations of drugs for 24 hours before MTT assay performed using the manufacturer’s recommended protocol (Sigma-Aldrich, St. Louis, MO). The absorbance at 490 nm was detected with a microplate reader (Biorad, USA). The half maximal inhibitory concentration (IC₅₀) was calculated by regression curve based on the cell inhibition rate and corresponding drug concentration.

Colony formation assay.

Cells were seeded in 12-well plates at a density of 200 cells/well and then treated with drug doses 24 hours later. The medium was replaced with fresh medium containing the corresponding drug every 3 days. After a week, the medium was removed and cell colonies were stained with crystal violet (0.1% w/v in 20% methanol), counted under a microscope and photographed.

Apoptosis detection.

Cells were treated with corresponding agents for 24 hours and then stained with Annexin V/PI followed by flow cytometry detection. The results were analyzed by LSR II and FlowJo software. To detect apoptosis with the PI staining method, cells were incubated with individual agents for 6 hours before PI was added to a final concentration of 1 µg/ml. After incubation for 15 minutes, cells in the culture were imaged with a Nikon Ti inverted fluorescence microscope. Values of Q2 and Q3 quadrants were combined as total of apoptotic cells. For each treatment, three repeated flow cytometric detection of the same sample were used for statistical comparison.

Detection and quantification of cytochrome C release.

Cells on poly-L-lysine-coated coverslips were treated with 8 µM of the corresponding agents for 6 hours and then fixed in 4% paraformaldehyde for 15 minutes. After washing in PBS, permeabilization in 0.1% Triton X-100 and staining with cytochrome C antibody (Santa Cruz Biotechnology), the cells were counterstained with fluorescence-labeled secondary antibody (Santa Cruz Biotechnology). Immunofluorescence microscopy was used to obtain from each treatment group 15 random views (600×), from which 100 stained cells were quantified for released cytochrome C intensity by the Image J software.

Determination of mitochondrial transmembrane potential and reactive oxygen species (ROS).

After being treated with corresponding agents for 12 hours, cells were detached by trypsin treatment and collected. After washing twice in fresh complete medium, resuspended cells in the medium were examined. To detect changes in mitochondrial transmembrane potential, cells were incubated with 5 µM JC-1 at 37°C in the dark for 30 minutes. To detect mitochondrial ROS production, cells were incubated with MitoSOX by the manufacturer’s recommended protocol. Stained cells were washed twice in phosphate buffered saline (PBS) before analysis with flow cytometry.

Western blot analysis.

The protocol for western blot analyses was previously reported (33). After treatment for 16 hours, whole cell lysates were prepared in RIPA buffer supplemented with 10 mM ß-glycerophosphate and 1 × Roche’s cOmplete Mini Protease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO). The source and dilution of the antibodies used in this study is provided in Supplementary Table S1.

Seahorse assay.

Metabolic analysis of live cells was conducted with the Seahorse XF24^® extracellular flux analyzer (Seahorse Bioscience, North Billerica, MA). Cells were pre-seeded in XF24 multi-well plates for 24 hours, and then exposed to drugs for 2 hours prior to Seahorse assay. One hour before recording mitochondrial activity, the medium was replaced with XF base minimal phenol red-free DMEM medium supplemented with 10 mM glucose and 1 mM sodium pyruvate. To determine basal oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) under different drug stimulations, the cells were plated in XF24 multi-well plates overnight and then exposed to different drugs in XF base minimal DMEM medium without phenol red supplemented with 10 mM glucose and 1 mM sodium pyruvate overnight while OCR and ECAR were analyzed at same time. To determine spare capacity, OCR and ECAR were recorded before and after sequential injections of oligomycin (1 µM), the electron transport chain uncoupler carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP, 1 µM) and the specific inhibitors of the mitochondrial respiratory chain, antimycin A/rotenone (0.5 µM). Spare capacity is defined as the difference between the maximal respiration and basal respiration.

Tumor xenograft experiments.

Animal studies were conducted in compliance with a protocol approved by the Institutional Animal Care and Use Committee (IACUC#0006403). Female 6-week-old nu/nu BALB/c mice (Jackson Laboratory, Bar Harbor, ME) were used. PC9AR and H1650 cells (2 × 10⁶/site) were inoculated subcutaneously (s.c) in the flanks of the mice. After 7 days, when tumor sizes reached 50 mm³, the mice were randomized into groups (n = 6). Individual groups were treated intraperitoneally (i.p) with vehicle (1% Tween 80), DZ (5 mg/kg), SIM (5 mg/kg) or DZ-SIM (5 mg/kg), or orally with AZD9291 (10 mg/kg, gavage). Treatments were carried out three times a week. During this period, animals treated with DZ or DZ-SIM were subjected to whole body NIR imaging with an IVIS Lumina XR Imaging System (PerkinElmer, Waltham, MA) with a fluorescent filter set with excitation at 783 nm and emission at 840 nm. Background fluorescence was automatically subtracted. Tumor volumes were calculated with the formula a² × b × 0.5236 where a was the smallest diameter and b was the oppositing diameter. Body weight, feeding behavior and motor activity of the animals were monitored as indicators of general heath. Mice treated for 16 or 28 days were euthanized. The tumor tissues and host organs were weighed and rinsed in phosphate buffered saline before NIR imaging again with the same experimental setting as the wholebody imaging.

Immunohistochemical staining (IHC).

Formalin-fixed and paraffin-embedded tumor sections were stained for Ki67 and pH2AX detection following our published protocol (33). The color was developed with 0.05% diaminobenzidine and 0.03% H₂O₂ in 50 mM Tris-HCl and then counterstained with hematoxylin and eosin (H&E). Negative controls for every antibody were also included.

Statistical analysis.

All experiments were performed independently at least three times. Results are presented as the mean ± SD. GraphPad Prism 7.0 was used for statistical analysis and p < 0.05 was considered as statistically significant.

Results

We previously reported that DZ conjugates with chemotherapeutic agents such as cisplatin could target cancer cell organelles, including mitochondria (32). Whereas accumulation of DZ in free form did not affect cancer cell viability, DZ-cisplatin conjugate at low µM concentration induced rapid and complete Burkitt’s lymphoma cell death. Since DZ conjugates killed cancer cells via subcellular organelle targeting, circumventing the known anti-tumor mechanisms of conventional anti-tumor drugs, we tested whether the DZ-SIM conjugate could similarly kill representative HSCLC cells with drug resistance.

1. Synthesis of DZ-SIM conjugate.

The steps of DZ-SIM synthesis are outlined in Supplementary Figure S1. General synthesis protocol of DZ compound and its amide conjugate can be found in our previous publication (28).

Identity of DZ-SIM amide was confirmed by Mass Spectrometry as follows: 1179.65 [M+H]+. 1H NMR (400 MHz, DMSO-d6) δ 8.84 (s, 1H), 8.21 (t, 2H), 7.77 (t, 1H), 7.73 (t, 1H), 7.60–7.54 (m, 3H), 7.48 (d, 1H), 7.41–7.37 (m, 3H), 7.27–7.21 (m, 2H), 6.38 (d, 1H), 6.25 (d, 1H), 5.89 (d, 1H), 5.71 (m, 1H), 5.42 (s, 1H), 5.11(m, 1H), 4.18 (m, 4H), 3.89 (m, 1H), 3.42 (m, 1H), 2.98 (m, 4H), 2.68 (m, 4H), 2.32–2.22 (m, 2H), 2.18 (m, 1H), 2.11 (m, 2H), 2.03 (s, 1H), 2.01 (d, 1H), 1.89 (m, 1H), 1.80 (m, 5H), 1.70 (m, 4H), 1.63 (S, 6H), 1.62 (s, 6H), 1.52 (m, 4H), 1.41 (m, 5H), 1.32 (m, 5H), 1.19 (m, 3H), 0.99 (s, 3H), 0.98 (s, 3H), 0.95 (d, 3H), 0.76 (d, 3H), 0.70 (t, 3H).

Using the newly synthesized compounds, we verified the tumor-cell specificity of the DZ-SIM conjugate with the xenograft tumor models. In validation studies, mice bearing s.c H1650 xenograft tumors were treated with DZ-SIM (5 mg/kg, three times a week) for 4 weeks. The animals were subjected to NIR tumor imaging after the first treatment. Comparative analysis indicated that DZ-SIM displayed similar tumor specificity (Figure 1A), similar to other dye drug conjugates that we have previously reported (28,34). Following the last treatment, the animals were euthanized after 48 hours and tumor tissues and host organs (liver, kidney, spleen) were resected for ex vivo NIR imaging, which confirmed that DZ-SIM accumulated in tumor tissue specifically (Figure 1B). Furthermore, DZ-SIM treatment significantly inhibited tumor growth, as DZ-SIM sharply reduced tumor volume (Figure 1C) and weight (Figure 1D) and Ki67 stain (Figure 1E). The stronger stain of pH2AX indicated increased DNA damage, suggesting DZ-SIM induction of tumor cell death compared to both DZ and SIM control groups (Figure 1E). Interestingly, mice treated with the DZ-SIM regimen did not show significant body weight changes (Supplementary Figure S2) or visible discomfort, suggesting that therapeutic dose of the conjugate has little toxicity to normal tissues and organs of the mouse host.

Figure 1. Tumor specificity of the DZ-SIM conjugate.

H1650 xenograft tumor model was used. A, tumor-bearing mice were subjected to whole body NIR imaging 24 hours after a treatment with DZ-SIM. A representative whole body NIR imaging is shown. B, after treatment with DZ-SIM for 4 weeks, xenograft tumors were subjected to ex vivo NIR imaging, 48 hours after the last DZ-SIM administration. Specific DZ-SIM accumulation in xenograft tumors (T) was detected, while liver (L), spleen (S), and kidney (K) of the corresponding mouse host showed little conjugate retention. C, DZ-SIM inhibited the xenograft tumor growth. In this study, the treatment was started 14 days after tumor cell inoculation (arrow). D, tumor weight was determined at the end of the study. E, histopathologic (H&E) stain revealed death of the tumor cells (arrows). IHC staining of tumor tissues showed reduced growth (Ki67 stain) and elevated cell death (pH2AX stain), compared with either DZ or SIM control groups (200×). For all the presentations, * p ≤ 0.05, ** p≤ 0.01, and *** p≤0.001.

2. DZ-SIM inhibits proliferation and colony formation of TKI-resistant or cisplatin-resistant cancer cells.

We employed a panel of representative NSCLC cell lines and their therapeutic resistant derivative sublines to further investigate the anti-tumor effect of DZ-SIM. These individual cell lines were known to harbor distinct insensitivities to chemotherapeutics or targeting therapies. In our assessment, A549 was relatively sensitive to cisplatin with an IC₅₀ value around 34 µM, while A549DDP was resistant to cisplatin with an IC₅₀ around 90 µM (Supplementary Figure S3). As for the first-generation EGFR-TKI gefitinib, PC9 cells were sensitive with an IC₅₀ around 4 nM, whereas H1650 and H1975 cells were insensitive with IC₅₀ around 10 µM. When the third-generation EGFR-TKI AZD9219 was tested, PC9 cells were sensitive with an IC₅₀ of 96 nM, compared to lineage related PC9AR cells that were insensitive, with an IC₅₀ around 4 µM. These cells formed useful pairs for the comparative study of therapeutic resistance.

To assess their sensitivity to DZ-SIM, all these cancer cell lines were treated with increasing concentrations of DZ-SIM for 24 hours. The treated cells were then subjected to cell viability assay. As anticipated, DZ-SIM completely inhibited the viability of all these cancer cell lines (Figure 2A), most with an IC₅₀ value below 10 µM. The YTMLC-90 lung squamous cancer cells and H446 small cell lung cancer cells were all sensitive to DZ-SIM treatment (Table 1). DZ-SIM thus acted on all the cancer cells tested, disregard of their histological classification or sensitivity to other anti-tumor therapies. Similarly, the action of DZ-SIM seemed not related to genetic and expressional status of EGFR and other signature proteins (Supplementary Table S2). These results defined DZ-SIM as a unique anti-tumor agent competent for inhibiting lung cancer cell growth, Importantly, DS-SIM seemed to be less effective in killing normal cells in culture, as it caused much less death in healthy donor PBMC samples (Supplementary Figure S4). The precursors of DZ and SIM, individually or in combination, displayed little inhibitory effect at comparable concentrations.

Figure 2. DZ-SIM inhibited the growth of cisplatin resistant and EGFR-TKI resistant lung cancer cells.

Representative results are shown. A, DZ-SIM inhibited cell viability in a dose-dependent manner. Lung cancer cells cultured in 96-well plates were treated with the indicated doses of DZ-SIM for 24 hours. MTT assays were used to determine changes in cell viability. Data are presented as percentage changes relative to the control group. B, the effect of a 7-day DZ-SIM treatment on colony formation by gefitinib-resistant H1650 and H1975 cells and by AZD9219-sensitive PC9 and -resistant PC9AR cells. Whereas long-term DZ or SIM treatment displayed inhibitory effect on colony growth, these agents could not kill cancer cells to eliminate colony formation. For all the presentations, * p ≤ 0.05, ** p≤ 0.01, and *** p≤0.001.

Table 1.

IC50 values (µM) of anti-tumor agents on human lung cancer cell lines^*.

	A549	A549DDP	H1650	H1975	PC9	PC9AR	H446	YTMLC-90
	Cisplatin sensitive	Cisplatin resistant	Gefitinib resistant		Gefitinib and ZD9291 sensitive	AZD9291 resistant	SCLC	LUSC
Cisplatin	34	90	>100	>100	4.8	37.6	18.5	10
Gefitinib	71.2	57.9	10.6	10	0.004	53	28.8	>100
AZD9291	9.7	6.9	7.4	7.4	0.096	4	1.8	2.2
DZ	>50	>50	>50	>50	>50	>50	>50	>50
SIM	>50	30	33.3	>50	33	37	39	>50
DZ-SIM	5.7	6.1	3.6	2.9	9.6	11	3.7	4.6

^*All the values were calculated based on triplicate MTT assay results from cells treated for 24 hours.

By closely inspecting cell responses, we determined that DZ-SIM caused a time-dependent reduction in cancer cell viability (Supplementary Figure S5) within a 24-hour time frame. In all the cell lines tested, a 10 µM DZ-SIM treatment reduced their viability significantly at the 8-hour time point and by 24 hours virtually no viable cancer cells could be found, and none could be rescued by re-incubation in fresh culture medium. This was in sharp contrast to conventional cisplatin therapy, which at this concentration failed to inhibit either A549 or A549DDP cells in 24 hours. Given that H1975 and H1650 cells showed sensitivity to the 10 µM SIM treatment, EGFR-TKIs gefitinib and icotinib at the same concentration barely suppressed H1975 or H1650 cells in the same time period. DZ-SIM thus appeared as a potent and fast acting anti-tumor agent.

The inhibitory effect of DZ-SIM could be further demonstrated through colony formation assays (Figure 2B). A week of DZ-SIM treatment at 8 µM concentration abrogated colony formation in cancer cells resistant to both first-generation (H1650 and H1975) and third-generation EGFR-TKIs (PC9AR). In the 7-day treatment, DZ, SIM or their combination showed inhibiting effect but was unable to eliminate colony formation. Comparative analyses indicated DZ, SIM, or their combination with cytostatic activity, and identified DZ-SIM with a cytocidal nature.

3. DZ-SIM induces cancer cell apoptosis.

Through microscopic inspection, we confirmed that 24-hour DZ-SIM treatment resulted in complete death of the treated cells, forming debris impossible to be used for further molecular examination. Using PI staining, we determined that the apoptosis process was activated early in DZ-SIM treatment, as dying H1975 cells became prevalent after 6 hours of incubation (Figure 3A). After 16 hours, most of the cells were dead in a dose-dependent manner as detected by flow cytometry following annexin V-FITC and PI staining of A549DDP cells. In most tests, DZ-SIM treatment at doses from 4 to 8 µM resulted in significant death of lung cancer cells resistant to either cisplatin or EGFR-TKIs, while control treatment by DZ or SIM showed little growth inhibition effect (Figure 3B).

Figure 3. DZ-SIM induced lung cancer cell apoptosis.

Representative results are shown. A, apoptosis detected by PI staining. H1975 cells on 24-well plates treated with 8 µM DZ-SIM were monitored for the commencement of cell death with PI staining and fluorescence microscope imaging. Apoptosis was initiated as early as 6 hours into the DZ-SIM treatment (200×). B, cisplatin resistant A549DDP cells treated for 12 hours were stained with Annexin V/PI for flow cytometric analysis. DZ-SIM induced partial cell death at 4 µM and complete cell death at 8 µM. C, summary of the flow cytometric results with Annexin V/PI stained cells. D, western blot results showed PARP-1 and caspase 3 cleavage in paired A549 and A549DDP cells activated by a 16-hour DZ-SIM treatment, while anti-apoptotic proteins survivin and MCL-1 were decreased. Moreover, DZ-SIM induced double-strand breaks in genomic DNA as indicated by pH2AX and pATM induction.

We next examined the state of the apoptosis-related proteins of PARP-1 and caspase 3, as well as survival-related survivin and MCL-1, following 16-hour DZ-SIM treatment. Western blotting revealed that DZ-SIM induced a dose- (Figure 3C) and time-dependent (Supplementary Figure S6) cleavage of PARP-1 and caspase 3, while the level of survivin and MCL1 proteins decreased concomitant to the appearance of phosphorylated H2AX and ATM, markers of double-strand breaks of the genomic DNA. In these analyses, A549 and its cisplatin-resistant A549DDP derivative were found with similar expression patterns, probably suggesting that the drug resistance is not due to altered expression of cell survival-related genes. Mechanism of the cisplatin resistance remains to be identified.

4. Mitochondrial dysfunction is a critical mediator in DZ-SIM induced apoptosis.

Our previous studies showed that DZ and its conjugate accumulate and enrich in subcellular organelles including mitochondria (28,34,35). As expected, DZ-SIM was found to be co-localized within mitochondria in A549 cells since DZ-SIM NIR signal was superimposed with the MitoTracker stain (Figure 4A). Whereas both DZ and DZ-SIM targeted mitochondria, only DZ-SIM was cytotoxic because it impaired the structural and functional integrity of the organelle. With JC-1 as an indicator of mitochondrial integrity, flow cytometric analysis revealed that DZ-SIM treatment rapidly reduced mitochondria membrane potential in A549 and A549DDP cells (Figure 4B), regardless of their sensitivity or resistance to cisplatin.

Figure 4. DZ-SIM targeted mitochondria.

Representative results are shown. A, co-localization of DZ-SIM with mitochondria. A549 cells were treated with 8 µM DZ-SIM for 15 minutes and then incubated with 100 nM MitoTracker for 30 minutes. The cells were fixed and imaged (600×). B, DZ-SIM reduced mitochondrial membrane potential. Cisplatin-sensitive A549 and cisplatin-resistant A549DDP cells were treated with the indicated agents for 6 hours and stained with mitochondrial probe JC-1 for flow cytometry. Reduced mitochondrial membrane potential is indicated by increased JC-1 monomer (green fluorescence). C, DZ-SIM triggered mitochondrial cytochrome C release. A549DDP cells treated with 6 µM DZ-SIM for 6 hours were subjected to immunofluorescence staining for cytochrome C (600×). D, quantitation of the cytochrome C releasing cells. For all the presentations, * p ≤ 0.05, ** p≤ 0.01, *** p≤0.001 and **** p≤0.0001.

Mitochondria have a determinant function in both cell survival and cell death, since mitochondria generate energy for cell survival but can also release cytochrome C to induce cell death. We used immunofluorescence staining to demonstrate that a 6-hour treatment with 8 µM DZ-SIM resulted in cytochrome C release from the mitochondria in A549DDP cells (Figure 4C), detected as substantial numbers of treated cells (Figure 4D) where the sharp granulated cytochrome C in mitochondria was replaced with a homogeneous stain (Figure 4C).

DZ-SIM also compromised mitochondrial oxidative phosphorylation function. We detected the effect on mitochondrial respiration and aerobic glycolysis using the Seahorse 24XF Extracellular Flux Analyzer. DZ-SIM treatment effectively reduced basal OCR and ECAR in all the tumor cell lines tested (Figure 5A), suggesting that DZ-SIM cut down most of the bioenergy source for cells. In mitochondria stress experiments, 2-hour DZ-SIM treatment inhibited maximal mitochondrial reparation and spare capacity (Figure 5B), while DZ-SIM decreased the glycolytic capacity in the same cells. Marked ROS production due to mitochondrial damage was detected by MitoSOX stain. For instance, a 6-hour DZ-SIM treatment resulted in ROS formation (Figure 5C), significantly higher than in the control groups (Figure 5D). These data indicated that DZ-SIM completely suppressed mitochondrial function resulting in reduced bioenergy production in cancer cells.

Figure 5. DZ-SIM extinguished mitochondria energy production.

Representative results with A549DDP cells are shown. A, DZ-SIM lowered mitochondria energy production. OCR and ECAR were determined during a 16-hour 8 µM DZ-SIM treatment. B, upper panels, representative results from three repeated mitochondrial stress assays. Lower panels: both mitochondria spare capacity and glycolysis capacity were decreased after treatment with 8 µM DZ-SIM. C, DZ-SIM induced mitochondria ROS production. Cells were treated with DZ-SIM for 6 hours and stained with MitoSOX for detection of mitochondrial ROS by flow cytometry. D, Quantification of the triplicate MitoSOX stain.

5. DZ-SIM inhibits tumor development in vivo.

We conducted a second in vivo study to confirm DZ-SIM’s anti-tumor activity using the PC9AR cells in subcutaneous inoculation to nude mice. When the size of tumor reached 50 mm³, the mice were randomized to groups (n = 6) to receive vehicle, DZ (5 mg/kg), Sim (5 mg/kg), DZ-SIM (5 mg/kg), or AZD9291 (10 mg/kg) treatment for 2 weeks. Markedly slowed tumor growth was noted in the group treated with DZ-SIM (Figure 6A), while the body weight remained relatively stable (Figure 6B). Resected tumors during necropsy revealed smaller tumor sizes and weight from DZ-SIM treated mice compared to those receiving other treatments (Figures 6C and 6D). Histopathologic analysis (H&E staining) again showed tumor necrosis specifically in the DZ-SIM-treated group (Figure 6E), and IHC analysis showed suppressed Ki67 levels in the DZ-SIM-treated group (Figure 6E). Together, these data verified the anti-tumor effects of DZ-SIM in vivo.

Figure 6. The effect of DZ-SIM on third generation EGFR-TKI resistant PC9AR tumors.

A, DZ-SIM significantly inhibited PC9AR tumor growth as determined by tumor volume measurement. B, DZ-SIM showed little effect on whole body weight of the treated animals. C, sizes of the tumors resected during necropsy. D, weight of the tumors resected during necropsy. E, upper row: H&E staining revealed specific tumor cell death in the DZ-SIM-treated group (arrows). Lower row: IHC staining shows decreased Ki-67 levels in the DZ-SIM treated group. For all the presentations, *P<0.05, **P<0.01, ***P<0.0005, and **** P<0.0001.

Discussion

Tumor cell resistance to chemotherapy and targeted therapy is the major limiting factor to effective treatment of NSCLC. One approach to tackling this challenge is to find novel drug targets and therapeutic means to kill the resistant tumor cells. In this study, we synthesized (Supplementary Figure S1) and characterized the novel agent DZ-SIM for its tumor-targeting and cancer cell-killing activity (Figures 1 and 6). The results of this study showed that DZ-SIM targets tumor cells and induces apoptosis in both chemotherapy-resistant and EGFR TKI-resistant lung cancer cells in vitro and in vivo by disrupting mitochondrial structure and function.

To critically assess the effect of DZ-SIM on clinical NSCLC, we used representative cancer cell lines with defined resistance to chemotherapy and EGFR-TKIs. Cisplatin is one of the classic chemotherapeutic drugs for lung cancer. It functions by interacting with purine bases to cause DNA damage and growth arrest (36). Derived from cisplatin sensitive A549 cells, A549DDP cells are resistant to cisplatin. EGFR-TKI is an example of a targeted agent for cancer therapy (12), although toxicity and acquired resistance often limit clinical benefits. The development of resistance in clinical lung cancers seems inevitable, even for third generation EGFR-TKI (37). The PC9 cell line used in this study is sensitive to both first and third generation EGFR-TKIs. In comparison, H1650 and H1975 are resistant to first generation EGFR TKI, while PC9AR is resistant to third generation EGFR-TKI. In this study, DZ-SIM effectively killed all these cells regardless of their insensitivity to chemotherapy or targeted therapies (Figure 2). Repeated in vitro tests revealed that DZ-SIM kills a wide spectrum of lung cancer cell lines, regardless of their ability to resist conventional chemotherapy or targeted therapies. These results strongly suggest that DZ-SIM should be further tested as a unique anti-tumor agent for the clinical treatment of NSCLC.

In previous studies, we reported that a DZ-cisplatin conjugate synthesized by our laboratory could target and kill lymphoma cells in vitro and in vivo without obvious side effects (32). Compared to DZ-SIM, the DZ-cisplatin conjugate was found to have lower efficacy in NSCLC cells, especially cisplatin-resistant lung cancer cell lines. Though both conjugates are tumor cell-specific, and both are enriched in subcellular organelles, differences in surface carrier OATP expression or mitochondrial toxicity may contribute to the observed differences in efficacy for lymphoma versus lung cancer cells. Comparative examinations remain to be conducted to elucidate the underlying mechanism.

Our further investigation suggested that DZ-SIM may employ at least two mechanisms to kill lung cancer cells. Following tumor cell-specific uptake, DZ-SIM was enriched in subcellular organelles, including mitochondria and lysosomes. DZ-SIM may kill NSCLC cells by damaging the mitochondrial structure primarily to cause cytochrome C leakage to the cytosol, where it activates the caspase 3-dependent apoptosis cascade (Figure 3). On the other hand, it may sabotage mitochondrial function to stop energy production, and cause ROS accumulation in the cancer cell (Figures 4 and 5). Considering the critical role of mitochondria in sustaining cellular vitality and inducing apoptosis, it may not be surprising that DZ-SIM treatment can induce rapid and complete lung cancer cell death.

DZ-SIM may use additional mechanisms to inhibit the aggressive behavior of NSCLC cells. The SIM moiety, for example, may have a tumor inhibitory effect. SIM specifically targets 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMGCR) in the endoplasmic reticulum to inhibit de novo cholesterol synthesis (38). SIM has been recently reported to have anti-tumor properties (39,40). Lee et al. (41) showed that SIM promoted apoptosis in lung cancer cells. Alizadeh et al. (39) showed that SIM induced A549 cell death by GGPP- and FPP-depletion. The anti-tumor activity of SIM is limited in vivo and in clinical trials, probably due to the non-tumor cell specificity of the compound. In this respect, the tumor cell specificity of DZ-SIM could enrich the SIM moiety inside the cancer cell. In this study, we found that DZ-SIM targeting mitochondrial may be part of the mechanism of DZ-SIM induced cell death. Further investigation is needed to fully illustrate the mechanism of DZ-SIM-induced lung cancer cell death.

In this study, we explored the mechanism of DZ-SIM-induced mitochondrial destruction. In preliminary studies, we found that mitochondrial fusion and fission could affect DZ-SIM-induced apoptosis. Mitochondrial dynamics of fusion and fission play a critical role in cellular homeostasis (42). Several core genes regulating mitochondria fission and fusion have now been identified, including mitofusions (MFN1 and MFN2) for mitochondrial fusion and dynamin-related protein 1 (DRP1), mitochondrial fission 1 protein (Fis1) and mitochondrial fission factor (MFF) for mitochondrial fission (43). Studies have documented that increased mitochondrial fusion suppresses apoptosis, whereas elevation in fission promotes apoptosis (42), which is consistent with our results. The exact genes regulating the balance of mitochondrial fission and fusion which may also be affected by DZ-SIM need to be determined in further research. Collectively, these results demonstrate that DZ-SIM’s anti-tumor properties are derived by targeting subcellular organelles, including mitochondria.

NSCLC is notorious for its therapeutic resistance. In this study, however, we found that DZ-SIM was highly effective in overcoming cisplatin-resistant cells and EGFR-TKI resistant cells both in vitro and in vivo. DZ-SIM concentration- and time-dependently decreased cell viability and colony formation, and induced cell death in both cisplatin-sensitive or resistant cells and EGFR-TKI sensitive or resistant cells. In our in vivo study, cancer cells resistant either to first or third-generation EGFR-TKI were all sensitive to DZ-SIM treatment. DZ-SIM appeared likely to be a potent and efficacious therapeutic drug.

Most chemotherapeutics target the cell division mechanism, either by interfering with DNA replication or blocking chromosome separation. In cancer progression, especially in advanced cases, aberrant cell division makes it difficult to precisely stop cancer cell growth without inflicting side effects on normal cells. EGFR-TKI targeted therapy is often encountered by cancer cell survival mechanisms, which are based on redundant signaling networks due to dynamic crosstalk among individual signaling transduction pathways. In contrast, DZ-SIM employs an alternative tactic by targeting subcellular organelles including mitochondria to effectively undercut energy production and activate the cytochrome C - caspase – apoptosis cascade. DZ-SIM is thus a promising example of subcellular targeted cancer therapy (44,45), one of the most widely used approaches for killing cancer cells by specific destruction of subcellular organelles.

In summary, compared to SIM, which targets HMGCR, DZ-SIM targets the mitochondrial energy-generating system and exhibits anti-tumor effects that overcome cisplatin and EGFR-TKI resistance. Translational research and development of DZ-SIM in lung cancer clinical trial appears warranted.