GDC-0349 inhibits non-small cell lung cancer cell growth

Non-small cell lung cancer (NSCLC) is a leading cause of cancer-related human mortality with a clear need for new therapeutic intervention. GDC-0349 is a potent and selective ATP-competitive mTOR inhibitor. In A549 cells and primary human NSCLC cells, GDC-0349 inhibited cell growth, proliferation, cell cycle progression, migration and invasion, while inducing significant apoptosis activation. Although GDC-0349 blocked Akt-mTORC1/2 activation in NSCLC cells, it also exerted cytotoxicity in Akt1-knockout A549 cells. Furthermore, restoring Akt-mTOR activation by a constitutively-active Akt1 only partially attenuated GDC-0349-induced A549 cell apoptosis, indicating the existence of Akt-mTOR-independent mechanisms. In NSCLC cells GDC-0349 induced sphingosine kinase 1 (SphK1) inhibition, ceramide accumulation, JNK activation and oxidative injury. Conversely, N-acetylcysteine, the JNK inhibitor and sphingosine 1-phosphate alleviated GDC-0349-induced NSCLC cell apoptosis. In vivo, daily oral administration of GDC-0349 potently inhibited NSCLC xenograft growth in mice. Akt-mTOR in-activation, SphK1 inhibition, JNK activation and oxidative stress were detected in NSCLC xenograft tissues with GDC-0349 administration. In summary, GDC-0349 inhibits NSCLC cell growth via Akt-mTOR-dependent and Akt-mTOR-independent mechanisms.


Introduction
Over 80% of lung cancer is non-small-cell lung cancer (NSCLC), causing over 1.6 million human mortalities worldwide each year 1,2 . Despite the latest developments in early screening, as well as the adjuvant and neoadjuvant therapies, the average survival for patients with advanced NSCLC is only 8-10 months 1,2 . Personalized moleculartargeted therapies are needed for NSCLC patients 3 .
Due to various gene depletion or mutation (PTEN, PI3KCA, and RTK etc), dysregulation and overactivation of phosphatidylinositol 3-kinase (PI3K)-Akt-mammalian target of rapamycin (mTOR) cascade is detected in NSCLC, which is associated with tumorigenesis and cancer progression 3,4 . Activation of PI3K-Akt-mTOR is vital for cancer cell growth, survival, proliferation, migration, and metabolism, as well as angiogenesis and therapyresistance. It thus has become an important therapeutic target of NSCLC 3,4 . Recent have tested the anti-NSCLC efficacy of PI3K-Akt-mTOR inhibitors as mono-therapy or in combination with other anti-cancer drugs 4 . mTOR lies in the central position of PI3K-Akt-mTOR cascade. It is in two multi-protein complexes: mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2) 5,6 . mTORC1 is rapamycin-sensitive and composed of mTOR, Raptor, mLST8, PRAS40, DEPTOR, and several others. mTORC1 phosphorylates p70S6K1 (S6K1) and 4E-binding protein 1 (4E-BP1) 5,6 . mTORC2 has several key components, including mTOR, Rictor, Sin1 and mLST8. It serves as the kinase for Akt phosphorylation (at Ser-473) and several other AGC kinases 5,6 . The two complexes are overactivated in NSCLC, emerging as key therapeutic targets.
Conventional mTORC1 inhibitors, including rapamycin and its analogs, only partially inhibit mTORC1 activity without directly affecting mTORC2 7 . mTORC1 inhibition will lead to feedback activation of oncogenic cascades, including PI3K-Akt and ERK-MAPK 8,9 . The second generation of mTOR kinase inhibitors block both mTORC1 and mTORC2, as well as PI3K 7,8 . These agents can completely shut down the whole PI3K-Akt-mTOR pathway, resulting in better anti-cancer activity 7,8 .
GDC-0349 is a potent and selective ATP-competitive mTOR inhibitor 10 . It blocks both mTORC1 and mTORC2 10 . Zhou et al., has shown that targeting mTOR by GDC-0349 potently inhibited head and neck squamous cell carcinoma cell growth 11 . Its potential effect on NSCLC cells, and the underlying mechanisms, have not been studied thus far. Here, we found that GDC-0349 inhibited NSCLC cell growth via Akt-mTOR-dependent and Akt-mTOR-independent mechanisms.
Cell culture A549 NSCLC cell line and BEAS-2B lung epithelial cells, both from Dr. Jiang 12 , were cultured as described 13 . Primary human NSCLC cells, derived from three NSCLC patients, "NSCLC-1/-2/-3", were described in our previous study 13 . The primary human lung epithelial cells were also provided by Dr. Jiang 12,14 . The primary human cells were cultured as describe early 12,14 . Mycoplasmamicrobial contamination examination, STR profiling, population doubling time and morphology were checked every 3-4 months to confirm the genotype. The writteninformed consent was obtained from each enrolled patient. The protocols of this study were approved by the Ethics Committee of Wenzhou Medical University, in accordance with Declaration of Helsinki.

Cell viability
Cells were seeded into 96-well plates at 3000 cells per well. Following the applied treatment, cell counting kit-8 (CCK8, Dojindo Laboratories, Kumamoto, Japan) was utilized to test cell viability 15 , and the optical density (OD) absorbance tested at the wavelength of 450 nm.

Colony formation assay
As reported 13 , A549 cells (at 6 × 10 5 cells per dish) were re-suspended in 0.5% agar-containing complete medium and added on top of a pre-solidified cell culture dishes. GDC-0349-containing medium was replenished every two days (total five rounds). Cell colonies were counted manually.

Trypan blue staining
Cells were seeded into six-well plates (8 × 10,000 cells per well). Following GDC-0349 treatment, trypan blue dye was added to stain the "dead" cells, and its ratio was calculated by an automated cell counter (Merck Millipore).

EdU (5-ethynyl-20-deoxyuridine) assay
Cells were seeded into six-well plates (8 × 10,000 cells per well) and treated with GDC-0349. An EdU Apollo-567 Kit (RiboBio) was applied to examine and quantify cell proliferation. EdU ratio (% vs. DAPI) was calculated from at least 500 cells from five random views under a fluorescent microscope.

Cell cycle analyses
NSCLC cells were seeded into six-well plates (1 × 100,000 cells per well). Following GDC-0349 treatment, cells were stained with propidium iodide (PI, 10 μg/mL) for 30 min under the dark. FACS was performed to test cell cycle distribution.

Cell migration and invasion assays
As reported 13,16 , NSCLC cells (in serum free medium, 4 × 10,000 cells per chamber) were seeded at the upper surfaces of "Transwell" chambers (BD Biosciences, Heidelberg, Germany). The lower compartments were filled with FBS-containing complete medium. After 20 h NSCLC cells migrating to the lower surface were fixed and stained. Matrigel (Sigma) was added in the chamber surfaces for in vitro cell invasion assays.

JC-1 assay
In cells with mitochondrial depolarization, the fluorescence dye JC-1 shall aggregate in mitochondria, forming green monomers 19 . NSCLC cells were seeded into six-well tissue-culturing plates at 60% confluence. Following GDC-0349 treatment cells were incubated with JC-1 (10 μg/mL) for 20 min under the dark, washed and tested immediately under a fluorescence spectrofluorometer at 488 nm (Molecular Devices, San Jose, CA).

Ceramide assay
NSCLC cells were seeded into six-well plates (8 × 10,000 cells per well) and treated with GDC-0349. Cellular proapoptotic ceramide contents, in fmol by nmol of phospholipids, were examined by the protocol reported elsewhere 21 .
Reactive oxygen species (ROS) assay NSCLC cells were seeded into six-well plates and treated with GDC-0349. Cells were then incubated with CellROX fluorescence dye (10 μM, Invitrogen) for 30 min under the dark at room temperature. CellROX intensity was tested under a fluorescence microplate reader. The representative CellROX fluorescence images were taken as well.

Glutathione content assay
As reported 20 , NSCLC cells were seeded into six-well plates and treated with GDC-0349. Reduced glutathione (GSH) and oxidized disulfide form glutathione (GSSG) in total cell lysates were tested by a GSH-GSSG assay kit (Beyotime, Wuxi, China). The ratio of reduced to oxidized glutathione (GSH/GSSG×100%) was calculated.

Western blotting
As described 13,16,17 , an equal amount protein lysates (30-40 μg per treatment in each lane) were separated by 10-12% SDSPAGE gels and thereafter transferred onto PVDF blots. Membranes were blocked and incubated with indicated primary antibodies, and subsequently incubated with corresponding secondary antibodies. The ECL detection reagents (Bio-Rad, Shanghai, China) were then applied to test signals.

Akt1 knockout
The CRISPR/Cas9-Akt1-KO-GFP construct was provided by Dr. Chen at Jiangsu University 20 . It was transfected to A549 cells by Lipofectamine 2000. GFP-positive A549 cells were sorted by FACS, and monoclonal cells distributed into 96-well plates. Single cells were further cultured in puromycin-containing medium and stable cells established. Akt1 knockout was verified by Western blotting.

Constitutively-active mutant Akt1
A recombinant adenovirus construct expressing the constitutively-active Akt1 (caAkt1, S473D) was from Dr. Zhang 22 . caAkt1 was transduced to A549 cells. GFPpositive A549 cells were sorted by FACS and monoclonal cells distributed into 96-well plates. The caAkt1 expression in single stable cells was verified by Western blotting.

Mice xenograft assay
As reported 13 , the severe combined immunodeficient (SCID) mice (18-20 g, all female) were provided by Soochow University Animal Center. A549 cells or NSCLC-1 primary cells (five-million cells per mice, in 200 µL Matrigel-serum free medium) were inoculated subcutaneously to the flanks of SCID mice. Within three weeks xenograft tumors were established with tumor volume close to 100 mm 3 . Mice were then received GDC-0349 administration or vehicle control. The mice body weights and bi-dimensional tumor measurements 23 were recorded every seven days. The animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) and Ethics Review Board of Wenzhou Medical University.

Statistical analysis
The quantitative data in this study were presented as mean ± standard deviation (S.D.). One-way ANOVA plus a Scheffe' and Tukey Test (SPSS 23.0) were utilized for statistical analyses between different groups. For comparing significance between two treatment groups, a twotailed unpaired T test (Excel 2007) was carried out. p < 0.05 was considered as a significant difference.

GDC-0349 induces NSCLC cell death and apoptosis
In NSCLC cells proliferation inhibition and cell cycle arrest will result in cell death and apoptosis. Trypan blue staining assay results, Fig. 2a, confirmed that GDC-0349 (25/100 nM, for 72 h) treatment induced A549 cell death, with the number of trypan blue-positive staining cells significantly increased. Activities of caspase-3 (Fig. 2b) and caspase-9 (Fig. 2c) were significantly elevated in GDC-0349 (25/100 nM)-treated A549 cells, where cleavages of caspase-3, caspase-9, and PARP were detected (Fig. 2d). These results implied activation of mitochondrion-dependent cell apoptosis pathway 24,25 . It was further supported by the fact that GDC-0349 induced mitochondrial depolarization in A549 cells, causing JC-1 green monomers accumulation in mitochondria (JC-1 green intensity increase, Fig. 2e). Additional studies confirmed that GDC-0349 induced apoptosis activation in A549 cells, increasing TUNEL-positive cell nuclei (results were quantified in Fig. 2f) and the ratio of Annexin Vgated cells (results were quantified in Fig. 2g). As shown 100 nM of GDC-0349 was again more potent than 25 nM in inducing A549 cell death and apoptosis ( Fig. 2a-g).

GDC-0349 blocks Akt-mTOR activation in NSCLC cells
GDC-0349 is a potent and selective mTOR inhibitor 10 . We tested its effect on Akt-mTOR cascade in NSCLC cells. In A549 cells GDC-0349 dose-dependently inhibited phosphorylation of Akt (at Ser-473 and Thr-308) and S6K1 (Thr-389), suggesting blockage of the whole Akt-mTORC1/2 cascade 5,6,26,27 . GDC-0349 treatment (see figure on previous page) Fig. 1 GDC-0349 potently inhibits NSCLC cell viability, proliferation, cell cycle progression, migration and invasion. A549 cells (a-f) or the primary human NSCLC cells (NSCLC-1/-2/-3) (g-i) were treated with applied concentrations of GDC-0349 (5-500 nM) or the vehicle control ("Ctrl", same for all Figures), cells were further cultured for applied time periods, cell viability (CCK-8 assay, a and g), colony formation (b), proliferation (nuclear EdU incorporation, c and h), cell cycle progression (PI-FACS, d), migration ("Transwell" assay, e and i) and invasion ("Matrigel Transwell" assay, f) were tested by assays mentioned in the text, and results were quantified. For nuclear EdU staining assays, five random views (1 × 100 magnification) with total 500 cells (for each condition) were included to calculate average EdU ratios (same for all figures). For "Transwell" and "Matrigel Transwell" assays, five random views for each condition were included to calculate the average number of migrated/invaded cells (same for all figures). The data are presented as mean ± standard deviation (SD, n = 5). *p < 0.05 vs. "Ctrl" cells. The experiments were repeated five times with similar results obtained. Bar = 100 μm (c, e, and f).
A constitutively-active Akt1 (ca-Akt1, S473D 22,31,32 ) was transduced to A549 cells. Stable A549 cells were established with FACS sorting and puromycin selection. Western blotting assay results, Fig. 3j, confirmed the expression of ca-Akt1 (indicated by the green star). In caAkt1-expressing A549 cells, phosphorylation of Akt and S6K1 was completely restored even with GDC-0349 (100 nM, 2 h) treatment (Fig. 3j). However, caAkt1 only  -g), the primary human NSCLC cells (NSCLC-1/-2/-3) (j-l), BEAS-2B epithelial cells or the primary human epithelial cells ("Pri-lung epi") (m and n) were treated with applied concentrations of GDC-0349 (25/100 nM), cells were further cultured for applied time periods, cell death was analyzed and quantified by Trypan blue staining assay (a and j), and caspase activation and apoptosis tested by mentioned assays (b-g, k, l, and n) with cell viability tested by CCK-8 assay (m). A549 cells were pre-treated for 1 h with applied caspase inhibitors (all at 50 μM), followed by GDC-0349 (100 nM) stimulation, cells were further cultured for 48-72 h, when cell viability and apoptosis were examined by CCK-8 (h) and nuclear TUNEL staining (i) assays, respectively. The data are presented as mean ± standard deviation (SD, n = 5). *p < 0.05 vs. "Ctrl" cells. # p < 0.05 vs. GDC-0349 only treatment (h and i). "n.s." stands for no statistical difference (m and n). The experiments were repeated five times with similar results obtained.

GDC-0349 oral administration inhibits NSCLC xenograft growth in SCID mice
To study the potential anti-NSCLC activity of GDC-0349 in vivo, A549 cells were subcutaneously (s.c.) injected to the flanks of SCID mice. Within three weeks A549 xenograft tumors were established and tumor volumes close to 100 mm 3 (labeled as Day-0). A549 xenograftbearing SCID mice were then randomly assigned into three groups, receiving GDC-0349 (10 or 30 mg/kg, daily, oral administration) or the vehicle control.
from the mice and weighted individually. Results showed that A549 xenografts in GDC-0349-treated mice were much lighter than those in vehicle control mice (Fig. 5c). The mice body weights were not significantly different between the three groups (Fig. 5d). We failed to identify any apparent toxicities in experimental animals, indicating that mice were well-tolerated with the applied GDC-0349 treatment.

Fig
. 5 GDC-0349 oral administration inhibits NSCLC xenograft growth in SCID mice. A549 xenografts-bearing SCID mice were randomly assigned into three groups (10 mice per group); Mice received GDC-0349 treatment (oral administration, 10 or 30 mg/kg body weight, daily for 21 days) or vehicle control treatment ("Veh"); Tumor volumes (a) and mice body weights (d) were recorded every seven days. Estimated daily tumor growth was calculated using the described formula (b); Tumors of all three groups were isolated and weighted at Day-35 (c). At Day-7 and Day-14, one tumor from each group was isolated (total six tumors), and tumor lysates achieved; Expression of listed proteins was shown (e and g); Relative SphK1 activity was tested (f),with GSH/GSSG ratio examined as well (h). NSCLC-1 xenografts-bearing SCID mice were treated with GDC-0349 (oral administration, 30 mg/kg body weight, daily for 21 days) or vehicle control; Tumor volumes (i) and mice body weights (k) were recorded every seven days. Estimated daily tumor growth was calculated by the formula: (tumor volume at Day-42-tumor volume at Day-0)/42 (j). The data are presented as mean ± standard deviation (SD). *p < 0.05 vs. "Veh" group.
were homogenized and tumor tissue lysates were achieved. Western blotting assay results, Fig. 5e, demonstrated that phosphorylation of Akt and S6K1 was largely inhibited in GDC-0349-treated A549 xenografts (Fig. 5e). Furthermore, GDC-0349 administration inhibited SphK1 activity in A549 xenografts (Fig. 5f) but provoked JNK activation (Fig. 5g). Total SphK1 expression was unaffected (Fig. 5g). Further analyses demonstrated that the GSH/GSSG ratio was deceased in A549 xenograft tumor tissues with GDC-0349 administration (Fig. 5h), suggesting that GDC-0349 induced oxidative injury in vivo. Therefore, the in vivo signaling changes were in line with the in vitro findings. While blocked Akt-mTOR activation, GDC-0349 also induced SphK1 inhibition, JNK1 activation and oxidative injury in A549 xenografts.
Alternatively, the primary NSCLC-1 cells were injected s.c. to the flanks of SCID mice. Xenograft tumors were established within three weeks. As shown oral administration of GDC-0349 (30 mg/kg, daily for 21 days) potently inhibited NSCLC-1 xenograft growth in SCID mice (Fig. 5i). Calculating estimated daily tumor growth demonstrated that NSCLC-1 xenograft growth was again slowed by GDC-0349 administration (Fig. 5j). The mice body weights were not significantly different between the two groups (Fig. 5k). Therefore, GDC-0349 oral administration inhibited NSCLC xenograft growth in SCID mice.

Discussion
PI3K-Akt-mTOR cascade is vital for cell viability, growth, proliferation, metabolism, and migration [48][49][50] . This cascade is commonly dysregulated and overactivated in NSCLC 4,51 . The pharmacological inhibitors of this cascade have been intensively tested for anticancer therapy 4,51 . mTORC1 inhibitors, including everolimus and temsirolimus, have been approved by FDA for the treatment of certain advanced renal cell carcinoma 52,53 and breast cancer 54,55 . However, mTORC1 inhibitors can only partially inhibit mTORC1 activation without directly suppressing mTORC2 54,55 . Rapamycin and its analogs could induce feedback loop activation of oncogenic cascades, i.e., PI3K-Akt and Erk-MAPK 54,55 . Studies have shown that inhibiting PI3K or Akt alone will also result in activation of upstream receptor tyrosine kinases 48,49,56 . It is therefore necessity of simultaneous targeting of multiple signaling cascades for efficient anticancer therapy for NSCLC 4,51 .
GDC-0349 is a potent and selective ATP-competitive mTOR inhibitor. In this study, we reported that GDC-0349 blocked Akt-mTORC1/2 activation in NSCLC cells. It potently inhibited NSCLC cell growth, proliferation, cell cycle progression, migration and invasion, and simultaneously inducing apoptosis activation. It is however noncytotoxic in lung epithelial cells. Daily oral administration of GDC-0349, at well-tolerated doses, potently inhibited NSCLC xenograft tumor growth in SCID mice. Akt-mTOR inhibition was detected in GDC-0349-treated xenograft tumor tissues.
Combined targeting of PI3K-Akt-mTOR pathway along with other signaling cascades in NSCLC cells have provided promising preclinical results, which should be much better than single PI3K-Akt-mTOR blockage 4 . Although GDC-0349 blocked Akt-mTORC1/2 activation, we suggested that GDC-0349-induced anti-NSCLC cell activity was not solely dependent on Akt-mTOR blockage. First, GDC-0349 was significantly more potent than other known Akt-mTOR inhibitors (rapamycin, perifosine and AZD-2014) in inducing NSCLC cell death. Second, restoring Akt-mTOR activation by caAkt1 only partially attenuated GDC-0349-induced NSCLC cell death and apoptosis. Third, GDC-0349 was able to exert further cytotoxicity in Akt1-KO A549 cells, where Akt-mTOR activation was completely blocked. Indeed our results showed that GDC-0349 induced SphK1 inhibition, ceramide production, JNK activation and oxidative injury in NSCLC cells. SphK1 inhibition, JNK activation and oxidative stress were also detected in A549 xenograft tumors with GDC-0349 administration. These actions are clearly not the result of Akt-mTOR inhibition, as the mTOR inhibitor AZD-2014 or CRISPR/Cas9-induced Akt1 KO failed to affect SphK1 activity, JNK and ROS in NSCLC cells. These Akt-mTOR-independent mechanisms could explain the superior anti-NSCLC cell activity by this compound.
Sphingolipid metabolites, including S1P, ceramide, and sphingosine, are important players in the progression of human cancer 35 . S1P will promote cell survival and proliferation, while ceramide and sphingosine could induce cell growth arrest and apoptosis. SphK1, which catalyzes ceramide and sphingosine to S1P, is often overexpressed in NSCLC 33,57-59 and other human cancers 35 , associated with tumorigenesis, cancer progression and resistance to therapies 33,35,[57][58][59] . SphK1 inhibition will lead to ceramide (and sphingosine) accumulation, mediating cancer cell apoptosis 35 . In this study we show that GDC-0349 potently inhibited SphK1 activation, causing ceramide accumulation and JNK activation in NSCLC cells. Exogenously adding S1P alleviated GDC-0349-induced NSCLC cells death and apoptosis. These results indicating that SphK1 inhibition accounted, at least in part, for GDC-0349-induced anti-NSCLC activity. Although the underlying mechanisms may warrant further characterizations.

Conclusion
Despite the latest developments in screening and early diagnosis of NSCLC, as well as the emerging of novel targeted therapies, the five-year overall survival for advanced NSCLC patients has only moderately improved over the past three decades 1,2 . Therefore, development of new therapeutic interventions for this devastating malignancy is urgently needed 3,60 . The results of this study demonstrated that GDC-0349 targeted multiple signaling cascades (Akt-mTOR and beyond) and potently inhibited NSCLC cell growth in vitro and in vitro. Therefore, it would be interesting to further testing this compound for NSCLC in future studies.