Combination with cooling-heat Chinese medicine Qing-kai-ling, not warming-yang Shen-fu, enhances the antitumor effect of getinib in resistant non-small cell lung cancer models in vitro and in vivo

Background: Traditional Chinese Medicine (TCM) prescriptions should be decided according to the TCM treatment principle, and the warming-yang or cooling-heat should be the guide of treatment principle outline. Methods: In order to identify which treatment principle, warming-yang or cooling-heat should be combined with epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) in non-small cell lung cancer (NSCLC), we combined the typical warming-yang drug, Shen-fu (SF), and the typical cooling-heat drug, Qing-kai-ling (QKL) with getinib in resistant NSCLC models. Results: The results demonstrated that QKL combined with getinib induced signicantly increased cell viability inhibition and apoptosis in A549 cell line and signicantly smaller tumor volume and lower tumor weight in H1975 xenograft transplanted nude mice. On the contrary, SF combined with getinib had signicant antagonism effect on both cell viability inhibition and apoptosis in vitro , and on tumor weight in vivo . EGFR phosphorylation inhibition and the downstream PI3K/AKT and RAS/RAF/ERK pathway inhibition served an important role in the synergism effect between QKL and getinib in H1975 xenograft transplanted nude mice. Conclusions: The present study indicated that cooling-heat TCM treatment principle may reverse or delay NSCLC resistance to EGFR-TKIs, and combination of them warrants further study.


Background
Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) are the preferred treatment for intermediate to advanced stage non-small cell lung cancer (NSCLC) in patients with EGFR gene mutation. Despite the demonstrated bene ts of EGFR-TKIs, not all patients respond to treatment. Moreover, even for the exclusively EGFR-mutant advanced NSCLC patients, the median progression free survival (PFS) was only 9-13 months for those who accepted rst-line EGFR-TKI therapy 1 . Approximately 50% of patients who respond well initially to TKIs develop resistance due to the occurrence of secondary mutation T790M in exon20 of the EGFR gene 1 , and the third generation EGFR-TKIs have shown e cacy in these patients 2,3 . For the other half of the patients resistant to EGFR-TKIs, no optimal therapy was available. Even for the patients with T790M mutation, most of them are not taking third generation EGFR-TKIs in the real world Chinese clinic, because of the high price. Therefore, effective therapies to delay the resistance of EGFR-TKIs are therefore required.
TCM therapies have been widely used in cancer, including NSCLC. Although some clinical trials have reported the e cacy of some TCM decoction or patent prescription in combination with EGFR-TKIs, most of them were with small samples and low quality 4 . Besides, the TCM treatment principles in these studies were even different. As a result, most of the TCM therapies in clinical practice are still based on doctors' experience, rather than evidence from clinical trials. A typical TCM treatment process should be programmed as the following sequential steps, rst diagnosing TCM syndrome types, then deciding the TCM treatment principle and nally writing out a TCM prescription. Although the syndrome types in the TCM theory are complicated, Yin-cold (YC) or Yang-heat (YH) type differentiation is believed as the outline of TCM syndrome type diagnosis according to the traditional book Huangdi Neijing. Therefore, the warming-yang or cooling-heat should be the guide of treatment principle outline 5 . We have demonstrated that EGFR gene mutated NSCLC patients are more likely with YC syndrome type 6 . Besides, the most common side effect of EGFR-TKIs was red acneiform rashes [7][8][9][10][11] , with thirsty, red and dry tongue and yellow tongue coating, which are typical symptoms and signs for YH syndrome type. Therefore, EGFR-TKIs may affect with warming-yang in uence according to TCM theory, and we do not know whether the warming-yang in uence of EGFR-TKIs is a treatment effect or just a kind of side effect. If the warmingyang in uence is a therapeutic effect, then TCM therapies with warming-yang principles may help to improve the e cacy of EGFR-TKIs, and vice versa.
In this study, we attempted to know which treatment principle, cooling-heat or warming-yang, should be combined with the EGFR-TKIs. We used the typical warming-yang drug, Shen-fu (SF), and the typical cooling-heat drug, Qing-kai-ling (QKL), because they are without known anticancer effect. SF injection has been used for nearly 30 years in China for patients with YC-syndrome type. The active components of SF injection are extracted from Radix Ginseng and Radix Aconiti Carmichaeli. QKL injection or oral solution has been widely used for the treatment of high fever or acute infection in clinical practice. It is prepared from cholic acid, hyodeoxycholic acid, baicalin, and active materials extracted from Gardeniae Fructus, Bubali Cornu, Margaritifera Concha, Isatidis Radix, and Lonicerae Japonicae Flos. The aim of the present study was to investigate the effect of SF or QKL when combining with ge tinib on resistant NSCLC models in vitro and in vivo.

Preparation of drugs
Ge tinib was kindly provided by AstraZeneca China. SF injection was purchased from Ya-an San-jiu pharmaceutical co., LTD (Sichuan, China), and QKL injection and QKL oral solution were purchased from Ming-xing pharmaceutical co., LTD (Guangzhou, China). The component herbs of SF decoction used in the in vivo experiments, prepared Radix Aconiti Carmichaeli and Radix Ginseng, were purchased from Kang-mei pharmaceutical co., LTD (Guangzhou, China). Seventy-ve gram of Radix Ginseng and 150 g prepared Radix Aconiti Carmichaeli were mixed and rst soaked in 1000 ml water for 30 minutes, and then boiled for 90 minutes. The liquid was ltered through a piece of medical gauze and the drugs were boiled once more with 800 ml water for 90 minutes. The liquid was ltered again and mixed with that from the rst boiling. The solution was concentrated into 250 ml in the rotary evaporator (IKA®RV 10 Basic), with a concentration of 0.9 g/ml crude drug, and then stored at -80˚C until use. 4,5-dimethylthiazol-2, 5-diphenyl-2-H-tetrazolium bromide (MTT) was purchased from MP Biomedicals (California, USA). Annexin V/ propidium iodide (PI) apoptosis Kit was purchased from MultiSciences(Lianke)Biotech Co., Ltd (Hangzhou, China). BCA protein assay kit was purchased from Thermo Fisher Scienti c, Inc ( ML, USA). Rabbit anti-human EGFR, phospho-EGFR (p-EGFR), AKT, p-AKT, ERK and p-ERK monoclonal antibodies (mAb) and horseradish peroxidase HRP conjugated anti-rabbit antibody were purchased from Cell Signaling Technology, Inc (MA, USA). Electro-Chemi-Luminescence (ECL) reagent was purchased from Millipore Corporation (MA, USA).

Cell Culture
Human A549 NSCLC cells were obtained from the Cell Line Bank at the Laboratory Animal Center of Sun Yat-sen University (Guangzhou, China). H1975 cells were obtained from the Cell Line Bank of the Macao University of Science and Technology (Macao, China). Cells were maintained in RPMI-1640 medium (Gibco; ThermoFisher Scienti c, Inc., MA, USA) supplemented with 10% FBS (Gibco; Thermo Fisher Scienti c, Inc.) and 0.5% penicillin-streptomycin sulfate (Gibco; Thermo Fisher Scienti c, Inc.), and incubated at 37 C with 5% CO 2 .

Cell Viability Assay
Human A549 cell was used in the in vitro experiments. MTT assay was used to measure cell viability.
Brie y, cells were plated in 96-well culture plates at the density of 5 × 10 3 per well in complete medium. After 24 hours of incubation, cells were treated with ge tinib (20-100 µM), QKL injection and SF injection (0.6%-1.0%) alone or combination for 48 and 72 hours. Then the cells were incubated with 110 µl of 5 mg/ml MTT at 37˚C for 4 h. After removing the medium, 150 µl dimethylsulfoxide (DMSO, Sigma, MO, USA) was added to each well. After shaking for 10 min, the absorbance at 570 nm was measured spectrophotometrically using a microplate reader (VICTOR X5,PerkinElmer). Each condition was duplicated in 6 wells. After removing the maximum and minimum absorbance to calculate the average, the cellular viability in each condition was expressed as the percentage of the average absorbance to that of control cells.

Apoptosis Assay
Apoptotic cell death was determined by the Annexin V/PI apoptosis Kit according to the manufacturer's instructions. Brie y, cells were plated in 6-well culture plates at a density of 1.5 × 10 6 cells/mL in complete medium. After 24 hours of incubation, cells were treated with ge tinib, QKL and SF injection alone or combination, with the concentration and exposure time according to the MTT assay results. After removing the medium, cells were trypsinized with EDTA free trypsin solution (Gibco; Thermo Fisher Scienti c, Inc.), harvested and then resuspended in 500 µl Binding Buffer (1×) with 5 µl Annexin V-FITC and 10 µl PI. After incubation for 5 min at room temperature in the dark, the samples were analyzed with ow cytometer (Beckman Coulter FC500).

Western Blot Analysis
The cells were plated in 60 mm dishes at a density of 3.5 × 10 6 cells/ml in complete medium. After 24 hours of incubation, cells were treated with ge tinib, QKL and SF injection alone or combination, with the concentration and exposure time according to the MTT assay results. After the treatment, the cells were washed in PBS and lysed in lysis buffer. Protein concentrations were determined using the BCA protein assay kit. The samples corresponding to 30 µg of protein were boiled for 8-10 min, resolved on an 8% denatured SDS-polyacrylamide gel, and transferred onto a PVDF membrane (Millipore, MA, USA). After blocking non-speci c binding sites for 30 minutes using 5% skim milk, the membranes were incubated with rabbit anti-human EGFR, p-EGFR, AKT, p-AKT, ERK and p-ERK monoclonal antibodies for 2 hours at room temperature. After washed with TBST for 3 times, the membranes were incubated with HRP conjugated anti-rabbit antibody for 1 hour at room temperature. After washed with TBST for 3 times, visualization of the protein bands was accomplished using ECL reagent. ImageLab software (version 4.0) was used to calculate the expression of each protein, which was normalized by GAPDH.
Determination of antitumor effect in nude mice.

Immunohistochemistry
For immunohistochemical staining, para n-embedded tumor tissues were applied. The sections were depara nized in xylene and rehydrated with graded alcohol. Hydrogen peroxide (3%) was applied to block endogenous peroxide activity and then boiled in 0.01M citrate buffer (pH 6.0) twice with an autoclave. After blocking non-speci c binding sites using normal goat serum (Boster Biological Technology co.ltd, California, USA), tissue sections were incubated with rabbit anti-human EGFR (1:100), p-EGFR (1:100), AKT (1:200), p-AKT (1:50), ERK (1:200) or p-ERK (1:100) monoclonal antibodies at 4℃ overnight. After washed with phosphate buffer saline (PBS) for 3 times, the sections were incubated with HRP conjugated anti-rabbit antibody for 30 minutes at 37℃, and the peroxidase reaction was developed with diaminobenzidine substrate kit (Zhongshan Golden Bridge-Bio, Beijing, China). Hematoxylin (Dingguo Changsheng Biotechnology CO., Ltd, Beijing, China) was then used for nucleus staining. Image-ProPlus 5.0 software was used to calculate the ratio of integrated optical density (IOD) to area (IOD/Area).

Statistics
Statistical analysis was performed using SPSS 19.0 statistical software (SPSS, Inc., Chicago, USA). The in vitro experiments were performed three times, independently. All data were presented as the mean ± standard deviation (SD). Differences between groups were assessed by two-tailed t test, one-way analysis of variance or analysis of variance for repeated measuring data and least signi cant difference (LSD)-t test was used for multiple comparisons. P < 0.05 was considered to indicate a statistically signi cant difference. q value method was used to evaluate the combination effect of ge tinib and QKL/SF, and it was calculated using the equation: q = EAB/ (EA + EB-EA × EB), where EA and EB were the inhibition effect of ge tinib and QKL/SF, respectively. EAB represented an observed value of combined effect, and (EA + EB-EA × EB) represented an expected value of combine effect. A q value of 1.15 or more is considered synergism, q < 0.85 as antagonism and the value between 0.85 and 1.1.5 is considered as additive effect 13 .

Cell viability in vitro
We rst screened the concentration of ge tinib. According to the MTT assay, 45 µM of ge tinib for A549 was used in the experiments with 48 hours of drug exposure, and 35 µM for A549 was used in the cells exposed to the drugs for 72 hours. As shown in Fig. 1, in A549 cell lines, the 0.9% and 1.0% of QKL injection combined with ge tinib induced signi cantly increased viability inhibition after 72 hours treatment compared with ge tinib alone. However, the q value indicated only additive not synergism effect (0.85 < q < 1.15). There were no signi cant viability differences between ge tinib alone and SF + ge tinib treated A549 cells. According to the maximum combination effect and minimum effect of QKL or SF injection alone in MTT assay, 35 µM of ge tinib and 0.9% of SF or QKL injection with 72 hours in A549 was used in the apoptosis and western blot experiments.

Apoptosis in vitro
Synergism effect of QKL injection and ge tinib on apoptosis were seen in the A549 (q = 1.40) cell line, with signi cantly increased apoptosis rate in the QKL + ge tinib treated cells compared with the ge tinib treated cells (Fig. 2). On the contrary, although the differences in apoptosis rates were not signi cant between the SF + ge tinib and ge tinib alone treated cells, q value method showed antagonism effect of SF injection and ge tinib in A549 (q = 0.75).

Antitumor tumor activity in vivo
In H1975 xenograft transplanted nude mice, QKL oral solution, ge tinib, and QKL + ge tinib inhibited tumor growth, with signi cantly smaller tumor volume and lower tumor weight compared with those in control group (p < 0.05, Fig. 3). Differences in tumor volume and weight among SF decoction, SF + ge tinib and control group were not signi cant. Tumor volume and weight in QKL + ge tinib group were even signi cantly lower than in ge tinib group (p < 0.05). Q value according to the inhibition rate in tumor weight indicated synergism effect between QKL and ge tinib (q = 1.19). Although the differences in tumor volume and weight between SF + ge tinib and ge tinib groups were not signi cant, q value shown antagonism effect between SF and ge tinib (q = 0.50) .

EGFR Pathway Protein Expression
In the experiments in vitro, the protein expression levels of p-EGFR were signi cantly lower in ge tinib, ge tinib + SF, and ge tinib + QKL groups in A549 cell line, comparing with control groups (Fig. 4A). The differences among the three groups were not signi cant. No signi cant differences were seen in EGFR, AKT/p-AKT, or ERK/p-ERK protein expression among A549 groups.
In H1975 xenograft transplanted nude mice, p-EGFR, AKT/p-AKT and p-ERK protein levels were signi cantly lower in ge tinib + QKL treated group, comparing with control group (Fig. 4B). Differences in EGFR and AKT/p-AKT protein levels were also signi cant between ge tinib alone and ge tinib + QKL treated groups. These data suggested that EGFR phosphorylation inhibition and the downstream PI3K/AKT and RAS/RAF/ERK pathway inhibition served an important role in the synergism effect between QKL oral solution and ge tinib in H1975 xenograft transplanted nude mice. Although EGFR protein level in ge tinib + SF group was signi cantly increased, comparing with control group, there were no signi cant differences in p-EGFR or downstream protein levels. Therefore, mechanism of antagonism effect between SF decoction and ge tinib was not clear.

Discussion
EGFR-TKIs are the preferred treatment for intermediate to advanced stage NSCLC in patients with EGFR gene mutation. However, the median PFS was only 9-13 months even for the exclusively EGFR-mutant advanced NSCLC patients. More effective therapies are therefore required. Traditional Chinese medicine, which is popular in Chinese and East Asian societies, may be a potential effective strategy to delay the resistance to EGFR-TKIs.
The present study demonstrated that in ge tinib resistant in vitro and in vivo models, cooling-heat TCM prescription and ge tinib had synergism effect, while warming-yang prescription was antagonistic to ge tinib. Some clinical studies also indicated that TCM prescriptions or decoction with cooling-heat treatment principle may increase the e cacy of rst-generation EGFR-TKIs. In Yang X's study, the TCM treatment principle of the decoction was strengthening vital qi and cooling cancer toxicity, especially cooling heat, and TCM decoction combining with ge tinib signi cantly prolonged progression free survival and overall survival 14 . In He W's meta-analysis, which analyzed the Chinese studies published from 2000 to 2016, only 6 studies reported increased e cacy from TCM when combining with EGFR-TKIs. Four of these 6 studies used speci c TCM patent prescription or decoction, in which 3 were with cooling-heat treatment principle 4 . On the other hand, a case report by Sung-Wook Hwang et al. reported that inappropriate TCM herbs, mainly Ginseng with warming-yang effect, induced resistance to ge tinib whereas withdrawing of the herbs caused sensitivity again to ge tinib 15 .All these studies, together with our ndings, suggested that the observed warming-yang effect of EGFR-TKIs may be a kind of side-effect even in TCM theory, and should be combined with cooling-heat treatment principle TCM therapies.
We also tried to nd out the mechanism of the synergism or antagonism effect of QKL or SF to ge tinib.
However, we only demonstrated that EGFR phosphorylation inhibition served an important role in the synergism effect between QKL oral solution and ge tinib in H1975 xenograft nude mice. Interestingly, H1975 cells are refractory to EGFR-TKIs due to the presence of the T790M mutation in EGFR 16 , which is the mechanism of 50% of the EGFR-TKIs resistant patients 1 . The T790M mutation results in steric hindrance of binding of ge tinib to the ATP-kinase-binding pocket 16 . Our data suggested that QKL induced the EGFR phosphorylation re-inhibition and downstream PI3K/AKT and RAS/RAF/ERK pathway re-inhibition of ge tinib. This re-inhibition in T790M mutant H1975 may resulted from steric re-binding of ge tinib to the ATP-kinase-binding pocket, because downstream pathway re-inhibition was not observed in A549 (with K-ras not EGFR mutation 17 ). Therefore, protein domain structure analyses warrant further researches.
The limitation of this present study was that we used only one drug to present each of the treatment principle. Expansion of the observed synergism effect of QKL and antagonism effect of SF with ge tinib to the effect of cooling-heat or warming-yang treatment principle should be very careful. Based on our initial nding, there are some TCM issues what warrant further research. First, we can further use other TCM patent prescriptions or decoction to present the cooling-heat or warming-yang treatment principle. If similar phenomena would be observed again, our hypothesis would be con rmed. Secondly, since synergism effect of QKL with ge tinib in tumor growth and apoptosis has been demonstrated in vitro and in vivo, reversing the resistance to ge tinib, effect on epithelial-mesenchymal transition (EMT) and metastasis warrants further researches, which should also been transferred to clinical validation. All these researches will be meaningful for guiding the principle of TCM therapies in combination with EGFR-TKIs, or for nding new ways to delay the resistance to EGFR-TKIs.

Conclusions
In conclusion, we found that QKL, with cooling-heat TCM treatment principle, increased e cacy of ge tinib in refractory models, while warming-yang SF acted as antagonist to ge tinib. The in uence of other drugs with cooling-heat or warming-yang TCM treatment principle warrant further study.