Tandem Molecular Self-Assembly Selectively Inhibits Lung Cancer Cells by Inducing Endoplasmic Reticulum Stress

The selective formation of nanomaterials in cancer cells and tumors holds great promise for cancer diagnostics and therapy. Until now, most strategies rely on a single trigger to control the formation of nanomaterials in situ. The combination of two or more triggers may provide for more sophisticated means of manipulation. In this study, we rationally designed a molecule (Comp. 1) capable of responding to two enzymes, alkaline phosphatase (ALP), and reductase. Since the A549 lung cancer cell line showed elevated levels of extracellular ALP and intracellular reductase, we demonstrated that Comp. 1 responded in a stepwise fashion to those two enzymes and displayed a tandem molecular self-assembly behavior. The selective formation of nanofibers in the mitochondria of the lung cancer cells led to the disruption of the mitochondrial membrane, resulting in an increased level of reactive oxygen species (ROS) and the release of cytochrome C (Cyt C). ROS can react with proteins, resulting in endoplasmic reticulum (ER) stress and the unfolded protein response (UPR). This severe ER stress led to disruption of the ER, formation of vacuoles, and ultimately, apoptosis of the A549 cells. Therefore, Comp. 1 could selectively inhibit lung cancer cells in vitro and A549 xenograft tumors in vivo. Our study provides a novel strategy for the selective formation of nanomaterials in lung cancer cells, which is powerful and promising for the diagnosis and treatment of lung cancer.


Introduction
Nanomaterials [1,2] are promising for cancer theranostics [3][4][5][6][7][8][9], but it has been documented that less than 1% of administered nanomedicines accumulate in tumors [10], thus leading to poor therapeutic efficacy of nanomedicines [11]. It remains a challenge to develop novel strategies to boost the therapeutic efficacy of nanomedicines. Recently, the in situ formation of nanomaterials in cancer cells and tumors has emerged as a promising strategy for cancer diagnosis and therapy due to the enhanced selectivity, permeation, and retention of the nanomaterials in tumors [12][13][14][15][16][17][18][19][20]. The successful examples that have been reported now primarily rely on using a single trigger to control the formation of the nanomaterials in situ. For example, nanofibers have been selectively formed in different kinds of cancer cells by the overexpression of enzymes, including alkaline phosphatase (ALP), matrix metalloproteinase (MMP), transglutaminase, and cathepsin B [21][22][23][24][25][26][27][28][29][30][31][32]. The combination of two or more triggers to form nanomaterials in situ may provide for more sophisticated means of control and manipulation, but this strategy has been reported only rarely [13,33]. Lung cancer cells, including A549 cells, show elevated expression levels of both extracellular ALP and intracellular reductase [34][35][36]. Taking advantage of these two overexpressed enzymes in A549 cells, we reported, in this study, a peptide derivative capable of responding to these two enzymes and showing a selective tandem molecular self-assembly in A549 cells.

Molecular Design and Compound
Synthesis. In our pilot study, we reported a tandem molecular self-assembly controlled by ALP and glutathione (GSH), specifically in liver cancer cells [13]. We opted to design molecules capable of selectively self-assembling and forming nanomaterials in other types of cancer cells [22,23,37]. Most cancer cells exhibit high expression levels of extracellular ALP, which has been widely used for the formation of nanofibers around and inside cancer cells. In addition, we realized that lung cancer cells also show high expression levels of intracellular reductase. We therefore designed the molecule NBD-GFFpYG-N=N-ERGD (Comp. 1 in Figure 1(b)) to be capable of responding to both ALP and reductase. We hypothesized that the conversion from Comp. 1 to NBD-GFFYG-N=N-ERGD (Comp. 2) by extracellular ALP might lead to the formation of nanoparticles or short nanofibers, which could be efficiently taken up by cells through endocytosis. The existence of the azo group in Comp. 2 could facilitate lysosomal escape and mitochondrial accumulation of the nanomaterials [38,39]. Following mitochondrial accumulation, the reductase in the mitochondrial membrane could convert Comp. 2 to NBD-GFFYG-aniline (Comp. 3 in Figure 1(b)), which could self-assemble into nanofibers in the mitochondria, leading to the disruption of the mitochondrial membrane and the release of cytochrome C (Cyt C), as well as the induction of oxidative stress, which can produce reactive oxygen species (ROS). The ROS could ultimately increase the ER stress and activate the unfolded protein response (UPR), leading to the selective cell death of the lung cancer cells.
The synthesis of Comp. 1 was simple and straightforward. We first synthesized the Fmoc-protected molecule containing an azobenzene (Comp. S2 in Scheme S1) that could be directly used for standard Fmoc-solid phase peptide synthesis (SPPS). The Comp. 1 was then obtained through standard SPPS using tritylchloride resin and purified by reversed-phase high-performance liquid chromatography. We also synthesized several control compounds via similar procedures, including NBD-GFFYG-N=N-ERGD (only reductase-responsive Comp. 2), NBD-GFFpYGERGD (only ALP-responsive Comp. 4), and NBD-GFFYGERGD (non-ALP-and non-reductase-responsive Comp. 5).

Dual
Enzyme-Triggered Tandem Molecular Self-Assembly. The Comp. 1 could form a clear solution (Figure 2(a)) in phosphate buffer saline (PBS, pH = 7:4) at a concentration of 200 μM (0.03 wt%), which was below its critical aggregation concentration (CAC = 263:6 μM, see the Supporting Information). The transmission electron microscopy (TEM) images revealed amorphous structures in the PBS solution of Comp. 1. These results indicated that Comp. 1 did not self-assemble into regular nanostructures at the concentration of 200 μM. We thereafter added the enzyme ALP (1 U·mL -1 ) to the PBS solution of Comp. 1 at 37°C, and the solution remained clear after 12 h (Figure 2(a)). The LC-MS trace indicated that over 95% of Comp. 1 had been converted to Comp. 2 within 6 h (Figures 2(a) and S11). Accordingly, the nanofibers with a diameter of 6-10 nm formed in the resulting solution of Comp. 2 (Figure 2(c)). With the addition of rat liver microsomes (226 μg·mL -1 ) and NADPH (50 equiv.) to trigger the conversion from azobenzene to aniline, a yellowish precipitate was clearly observed within 24 h (Figure 2(b)). The LC-MS trace indicated that over 70% of Comp. 2 had been converted into Comp. 3 within 24 h (Figures 2(b) and S12). The TEM image exhibited nanofibers with a diameter of 5-7 nm and nanoparticles with a diameter of 30-40 nm in the precipitate. The results obtained by the dodecyl sulfate sodium saltpolyacrylamide gel electrophoresis (SDS-PAGE) indicated that the precipitate consisted of the proteins in the rat liver microsomes and both Comp. 2 and Comp. 3 ( Figure S14). There were hardly any proteins contained in the supernatant, suggesting that the self-assembling peptide derivatives could form tight complexes with the proteins. The above observations strongly indicated the tandem molecular self-assembly behavior of Comp. 1 with the catalysis by ALP and reductase.
2.3. Intracellular Tandem Molecular Self-Assembly. Generally, cancer cells exhibit higher expression levels of extracellular ALP than normal cells. Before testing the tandem molecular self-assembly of Comp. 1 in cancer cells, we synthesized the compound TPE-GFFYEG-N=N-EEEE to measure the expression levels of reductase in different cells. Tetraphenylethylene (TPE) is a fluorescent probe with an aggregation-induced emission (AIE) property [37], and TPE-GFFYEG-N=N-EEEE can be converted to TPE-GFFYEG-aniline by reductase. TPE-GFFYEG-aniline could self-assemble into nanostructures and emit a stronger blue fluorescence than TPE-GFFYEG-N=N-EEEE. Therefore, the intensity of blue fluorescence indicated the expression level of reductase in the cells. As shown in Figures S17 and S18, the A549 cells showed the strongest blue fluorescence in the confocal fluorescence microscopy images of all of the five tested cancer cells (A549, U87, MCF-7, PC-3, and HeLa cells), suggesting that A549 cells exhibited the highest expression level of reductase. The high expression levels of both extracellular ALP and intracellular reductase in A549 cells suggested that the tandem molecular self-assembly of Comp. 1 might work in A549 cells.
To test whether Comp. 1 could form nanofibers in live cells, Bio-TEM was first used. As shown in Figures 2(e) and S16A, the ultrathin sections of A549 cells at 6 h post administration of Comp. 1 (200 μM) displayed nanofibers with the diameters of 4-7 nm in the cytoplasm. These observations clearly indicated the good self-assembly property of our compound in live cells. We then incubated the A549 cells with 200 μM of Comp. 1 and obtained confocal laser scanning microscopy (CLSM) images of the cells at different time points. As shown in Figures 3(a) and 3(b), there were many green fluorescent dots representing the NBD-peptide within the A549 cells at the 1 h time point. The green fluorescence from NBD colocalized well with the red fluorescence from Lyso-Tracker, which indicated the efficient uptake of the NBD-peptide by cells through endocytosis pathways, in which the self-assemblies could bind to cell membranes by interacting with integrins ( Figure S19). However, there was little overlap of the green (NBD) and red (Lyso-Tracker) fluorescence at the 4 h time point, suggesting that the assemblies of NBD-peptide efficiently escaped from the late endosomes/lysosomes. The Comp. 2 containing the azobenzene group but without the phosphate group had a small fraction of lysosomal escape ( Figure S20) at the 4 h time point. Both Comp. 4 and Comp. 5 without the phosphate and the azobenzene groups barely escaped from the lysosomes within 4 h. The mean fluorescence intensity in the cells treated with Comp. 1 and Comp. 2 was similar, which was significantly higher than that in cells treated with Comp. 4 and Comp. 5. Therefore, it was reasonable to hypothesize that the azobenzene group played a vital role in the cellular uptake and the ability of lysosomal escape of the molecules. We increased the incubation time of cells treated with Comp. 1 (200 μM) to 8, 10, and 12 h. The results in Figure 3(c) indicated that the green fluorescence from NBD colocalized well with the red fluorescence from the Mito-Tracker in the cytosol at the 8 h time point, suggesting that the NBD-peptide that escaped from the   We performed a time-dependent Western blot assay to monitor the expression levels of Cyt C at different time points. We prepared the cytosolic fraction from A549 cells treated with 50 μM of Comp. 1 according to an established method for the assay. As shown in Figure 3(f), the concentration of Cyt C in the cytosol, released from mitochondria, increased dramatically during the first 6-8 h and remained at high levels in the following 24 h. JC-1 staining was also used to measure mitochondrial membrane potential. As shown in Figures 3(e) and S25, Comp. 1 caused mitochondrial depolarization at 4 h time point. We also prepared the whole-cell fraction (containing both cytosol and mitochondria) of A549 cells treated with Comp. 1; the timedependent Western blot results indicated that the Cyt C in the whole-cell fraction remained constant during the 24 h (Figures 3(g) and S29b). These observations suggested that the self-assembly of Comp. 1 in the mitochondria led to the disruption of the mitochondrial membrane and the release of Cyt C to the cytosol. Normally, cytoplasmic vacuolization is related to ER stress. We therefore also used a timedependent Western blot assay to examine the protein expression of ER stress-related signaling markers. As shown in Figures 3(h) and S29a, calnexin significantly increased after a 12 h treatment with Comp. 1, which promoted the unfolded protein response (UPR). The ER chaperone protein BiP, whose expression level was representative of ER stress, was obviously upregulated at the 3 h time point, which correlated well with the lysosomal escape of Comp. 2 at this time point. BiP was a short-lived protein, and its expression level decreased at the 6 h time point [22,38]. This protein maintained high expression levels from 8 h to 24 h, representing high levels of ER stress. Since the PERK signaling pathway was activated by BiP, its downstream protein CHOP, a proapoptosis protein, was also upregulated, ultimately leading to cell death [39,40]. Phosphorylated eIF2α (the indicator of PERK-UPR pathway) also was detected by Western blotting, as shown in Figure S28A, phosphorylation of eIF2α began to increase at 6 h, and the expression level increased significantly at 12 h. Taken together, the tandem molecular self-assembly of Comp. 1 selectively induced death of the non-small-cell cancer cell line (A549 cells) via the release of Cyt C from the mitochondria to the cytosol and the excessive activation of ER stress.

Mechanism Study of ER Stress and Evaluation of
Inhibiting Cancer Cells In Vitro. As oxidative stress and ER stress are integrally interconnected, ROS can directly disturb the ER protein-folding environment, reduce proper protein folding, and induce UPR activation. Damaged mitochondria can produce oxidative stress, resulting in the production of intracellular ROS. The resulting ROS can react indiscriminately with proteins [40][41][42]. To understand the mechanism of cytoplasmic vacuolization and cell death, the production of ROS was measured at different time points using H2DCFDA. The A549 cells were incubated with Comp. 1 (200 μM) at different time points. As the incubation time increased, the amount of ROS produced increased (Figure 4(a)), indicating that Comp. 1 could cause oxidative stress in the mitochondria and the production of ROS. Pretreating the cells with the antioxidant N-acetyl cysteine (NAC) before treatment with Comp.
1 could inhibit cytoplasmic vacuolization ( Figure S26) and reduce cytotoxicity ( Figure S27) and the expression of ER stress markers ( Figure S28B). Taken together, the tandem molecular self-assembly of Comp. 1 first disrupted the mitochondrial membrane, subsequently causing oxidative stress and increasing the levels of ROS, ultimately inducing the UPR, endoplasmic reticulum (ER) stress [41,43,44], and cell death. We then investigated the inhibitory capacity of our compounds against different cancer cell lines. As shown in

Discussion
In summary, we introduced a compound with selective tandem molecular self-assembly into the A549 lung cancer cell line. The Comp. 1 responded to extracellular ALP and intracellular reductase in a stepwise manner, showing a tandem molecular self-assembly behavior in the A549 cells. The tandem molecular self-assembly led to the release of Cyt C from the mitochondria to the cytosol, the generation of ROS, and an elevation in ER stress and UPR, successively. Subsequently, the severe ER stress led to disruption of the ER, formation of vacuoles, and ultimately, apoptosis of the A549 cells. We demonstrated that the Comp. 1 selectively inhibited lung cancer cells both in vitro and in vivo. Our study provided a novel strategy for the selective formation of nanomaterials in lung cancer cells, which is powerful and promising for the diagnostics and therapy of lung cancer.  DMSO-d6 as the solvent. HPLC was conducted at the LUM-TECH HPLC (Germany) system using a C18 RP column with methanol (0.05% of TFA) and water (0.05% of TFA) as the eluents. LC-MS was conducted at the LCMS-20AD (Shimadzu) system. HR-MS was performed at the Agilent 6520 Q-TOF LC/MS using ESI-L low concentration tuning mix (Lot No. LB60116 from the Agilent Tech.

Transmission Electron
Microscopy. 10 μL of samples were added to the carbon-coated copper grids; excess samples were removed with filter paper, then uranyl acetate for negative staining. At last, samples were placed in the desiccator overnight and observed with transmission electron microscopy.
A549 cells reach to about 70% confluence in 10 cm culture dish, remove the culture medium, and add the fresh culture medium containing 200 μM compound 1. After 4 or 6 h, remove the culture medium, wash the cells by PBS for three times, and scrape the cells with a cell scraper, then centrifuge it for 5 min. The cells were fixed with a 2.5% glutaraldehyde solution at 4°C overnight.

Reductase Concentration Analysis.
We synthesized reductase-based AIE probe TPE-GFFYEG-N=N-EEEE (TPE) to detect its concentration inside different cell lines. Cells were seeded in a CLSM cell culture plate at a concentration of 1 × 10 5 cell. After incubation for 24 h, upper medium was removed then cell was incubated with TPE (5 μM) for 6 h. The medium was removed, and the cells were washed with DMEM for three times. We then used CLSM to observe fluorescence intensity and used ImageJ to analyze relative fluorescence intensity (λ ex = 405 nm, all test conditions are consistent).

Lysosome and Mitochondrial
Colocalization. A549 cells was seeded in CLSM cultural dish at a density of 1 × 10 5 cells. After incubation for 12 h, a medium was removed. The cells were then incubated with Comp. 1, 2, 4, or 5 (200 μM) for 1 or 4 h (lysosome colocation). Cells were incubated with Comp. 1 (200 μM) for 8, 10, and 12 h (mitochondrial colocation). Next, the medium containing different compound was removed and washed by PBS for three times, and lysosome tracker (1×) was incubated with cells for 45 min. The medium was removed and the cells were washed by PBS three times. 1-2 mL of DMEM was added for imaging by live cell imaging systems (λ exc: = 488 nm, emission = 510-560 nm; λ exc: = 543 nm, emission = 650-750 nm). Mito-Tracker (1×) was incubated with cells for 30 min. The medium was then removed and washed by PBS three times. 1-2 mL of DMEM was added for imaging by live cell imaging systems (λ exc: = 488 nm, emission = 510-560 nm; λ exc: = 633 nm, emission = 650-750 nm, all test conditions are consistent).
A549 cells were seeded in CLSM cultural dish at a density of 1 × 10 5 cells. After incubation for 12 h, a medium was removed. The cells were then incubated with Comp. 1 (200 μM) at different time points. The cells also were incubated with FCCP (100 μM) for 4 h as the positive control group. Then, removing medium, JC-1 staining kit (20 μM) was incubated with A549 cells for 15 min. The medium was removed, and the cells were washed by PBS three times. 1 mL of PBS was added for imaging by live cell imaging systems (green channel (depolarization) λ exc: = 488 nm, emission = 500 − 540 nm; red channel (hyperpolarization) λ exc: = 543 nm, emission = 570-620 nm, all test conditions are consistent).

ROS Detection.
The fluorescent dyes H2DCFDA were used to measure the intracellular ROS, cancer cells were incubated with H2DCFDA for 60 min at 37°C in the dark and then washed with PBS thrice. The oxidation of H2DCFDA was detected using a fluorescence microscope. 4.14. Cell Proliferation. In this experiment, all cell lines were incubated in a 96-well plate at a density of about 6000. According to previous method [45], cell was fixed by cold 10% trichloroacetic acid solution at 4°C overnight after incubated with compound for 48 h. Then, discarding the fixative solution and washed by water for three times, put the 96well plate in a 37°C oven to dry. Next, 0.4% SRB solution was used to stain cell for 20 min at room temperature to form protein-bound dry, then using 1% AcOH solution to clean excessive SRB. Put the 96-well plate in a 37°C oven to dry and add 10 mM Tris base solution of 100 μL every well to solubilize the protein-bound dry for 30 min. Measure the OD at 570 nm in a microplate reader.

4.16.
In Vivo Antitumor Assay. A549 cells were maintained in our lab. The BALB/c nude (6 weeks old, female) mice purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. were used. 5 × 10 7 A549 cells (100 μL) mixed with growth factor-reduced Matrigel (50 μL, BD Biosciences) were subcutaneously injected into the right flank of each mouse. The drug treatment was started when tumor volume reached 80 mm 3 ; the tumor volume was calculated by the formula: length * width 2 /2. Mice (n = 5) received compounds 1, 2, 4, and 5 (5.0 mg/kg) in PBS by i.v. injection, whereas the control group (n = 5) received PBS only. The compound was injected every three days. The mice were weighed, and tumors were measured every two days during the treatment period.
4.17. H&E Staining of Tumor Tissue. The mice were sacrificed to obtain tumor tissues; tumor tissues were fixed by 4% formalin over 48 h, embedded in paraffin, cut into 5-6 μm sections for H&E staining, and evaluated using light microscopy as previously described method.