CD8⁺ T cells mediate the antitumor activity of frankincense and myrrh in hepatocellular carcinoma

Human HCC cell line HCCLM3 was obtained from Live Cancer Institute of Zhongshan Hospital (Shanghai, China). Mouse HCC cell line Hepa1-6 was purchased from Shanghai Institutes for Biological Sciences, Chinese Academy of Science (Shanghai, China). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin and 100 μg/ml streptomycin (all from Life Technologies, USA). All cells were kept at 37 °C in a humidified atmosphere of 5% CO₂ incubator.

CD8⁺NKG2D⁺ cells were generated using our group’s protocol [22]. In brief, human peripheral blood lymphocytes were isolated from healthy volunteers. Cells were cultured in GT-T551 (SK551S, TaKaRa, Japan) medium supplemented with 2% autologous serum, 100 units/ml penicillin, 100 μg/ml streptomycin, 2000 IU/ml recombinant human IFN-γ (Chemo Wanbang Biopharma, Shanghai, China) for 24 h. To promote the expression of NKG2D, OK432 (Chugai Pharmaceutical, Chome, Japan), an immunotherapeutic agent prepared from inactivated Streptococcus pyogenes, was added into the culture medium at the concentration of 10 mg/ml [23]. On 2nd day, cells were transferred to a flask pre-coated with anti-CD3 antibody (T&L Biological Technology, Beijing, China). GT-T551 medium supplemented with 700 IU/ml recombinant human IL-2 (BD Biosciences, USA) was used for further culture. After 2 weeks, cells were subjected to flow cytometry to control the phenotype and quality.

Frankincense (#71202000) and myrrh (#71202500) were purchased from Jiangsu Province Hospital of Traditional Chinese Medicine (Nanjing, China). Recombinant human TNF-α (HNAE-10620), mouse TNF-α (MNAE-50349), human IL-6 (HNAE-10395), and mouse IL-6 (MNAE-50136) were all obtained from Sino Biotechnology (Beijing, China).

The extract of frankincense and myrrh was prepared as previously described [24]. Briefly, powered resin of frankincense (1.0 kg) or myrrh (1.0 kg) was extracted with water (2 × 10 L) under reflux for 1 h. The extraction was repeated twice. Then the extracted solution was boiled in a reflux condensation device for 2 h, cooled to room temperature (RT), and centrifuged at 3000 rpm for 15 min to remove the residue. Finally, the supernatant was evaporated in a rotary evaporator for 12 h to obtain the powder of frankincense or myrrh.

To prepare the solution, 100 mg of frankincense powder and 100 mg of myrrh powder were dissolved in 5 ml cell culture medium (for in vitro experiments) or PBS (for in vivo therapy) in water bath at 90 °C overnight. Then the supernatant was obtained after centrifugation at 12,000 rpm for 30 min followed by filtration through a 0.22 μm filter.

MTT (M5655, Sigma, USA) (5 mg/ml, 20 μl) was added into each well in 96-well plate and incubated for 3 h at 37 °C. Then, the supernatant was discarded. 100 μl of isopropyl alcohol (12090611516, Nanjing Chemical Reagent Co., China) was then added and agitated for 20 min to dissolve the crystallization. The absorbance was measured using a Multimode Reader (SMP500-13497-JWYK, Molecular Devices, USA) at 570 nm. Cell viability was calculated as the ratio of the absorbance of treated cells to that of controls.

Harvested cells were stained with 0.2% trypan blue (C3601-2, Beyotime, China). Cell numbers and viability were determined with a Countstar Automated cell counter (Inno-Alliance Biotech, USA).

To test the oncolytic effect of CD8⁺NKG2D⁺ cells, HCCLM3 cells stably expressing luciferase (HCCLM3-luciferase) were used as target cell. The viability of HCCLM3-luciferase cells was determined by luciferase assay according to manufacturer’s instructions (E1500, Promega, USA). After co-culture in a black 96-well plate, cell lysis buffer was added into each well for 15 min. Luciferase assay substrate was added and then the plate was subjected to GloMax 96 microplate luminometer (Promega, USA).

Cells were lysed in radio-immunoprecipitation assay buffer containing a protease inhibitor cocktail (11873580001, Roche, Switzerland). The protein concentration was determined using BCA Kit (P10010, Beyotime, China). Equal amounts of protein were separated by SDS-PAGE and electrophoretically transferred onto polyvinylidene difluoride membranes (03010040001, Roche, Switzerland). The membranes were blocked, incubated with the specific primary antibodies, washed, and followed with incubation with appropriate horseradish peroxidase-conjugated secondary antibodies. Signals were detected using an enhanced chemiluminescence reagent (WBKLS0500, Millipore, USA) and subjected to the Alpha Innotech Fluor Chem-FC2 imaging system (ProteinSimple, USA). The gray value was analyzed with Software Image J (freely available at https://​imagej.​nih.​gov/​ij/​download.​html). Antibodies used were as follows: anti-glyceraldehyde-3-phosphate dehydrogenase (anti-GAPDH) (MB001, Bioworld, USA; 1:5000 diluted), anti-IκBα (4814S, Cell Signaling Technology, USA, 1:1000 diluted), anti-phospho-IκBα (2859S, Cell Signaling Technology, USA; 1:1000 diluted), anti-STAT3 (9139S, Cell Signaling Technology, USA; 1:1000 diluted), and anti-phospho-STAT3 (9145S, Cell Signaling Technology, USA; 1:1000 diluted).

Total cellular RNA was extracted with TRIzol (15596-026, Thermo Scientific, USA), and was reverse-transcribed using the synthesis system (RR036A, TaKaRa, Japan). Then SYBR green PCR master mix (04913914001, Roche, Switzerland) reagent was used according to the manufacturer’s protocol and the PCR was performed using the Real-Time PCR system (Viia7, Thermo Scientific, USA). Gene expression was calculated with the comparative threshold cycle (C _T ) method and normalized to the endogenous levels of GAPDH. Primer sequences used were as follows:

human TNF-α forward 5′- CCCAGGGACCTCTCTCTCTAATC-3′ and reverse 5′- ATGGGCTACAGGCTTGTCACT-3′;

human IL-6 forward 5′- GGTACATCCTCGACGGCATCT-3′ and reverse 5′- GTGCCTCTTTGCTGCTTTCAC-3′;

mouse tnf-α forward 5′- ACGCTCTTCTGTCTACTGA -3′ and reverse 5′- GCCATAGAACTGATGAGAGG-3′;

mouse il-6 forward 5′- CCACCAAGAACGATAGTCA -3′ and reverse 5′- TTGTCACCAGCATCAGTC -3′.

4–6-week-old male C57BL/6 mice and Balb/c nude mice were obtained from Nanjing Biomedical Research Institute of Nanjing University (Nanjing, China). Mice were kept in the conditional microenvironment (temperature 22 ± 2 °C, humidity 60 ± 10%, 12/12 h light/dark cycle).

To establish HCC xenograft models, each mouse was subcutaneously injected with 5 × 10⁶ Hepa1-6 cells or 1 × 10⁷ HCCLM3 cells suspended in 200 μl PBS containing 10% Matrigel Matrix (354234, Corning, USA). Therapy started on day 7 post tumor cell inoculation, when mice showed visible tumor nodules (4–6 mm in diameter). FM was given daily by oral gavage at 60 mg/kg until the therapy stopped at the indicated time point.

For the CD8 and NK cells depletion experiment, 500 μg of anti-CD8α mAb (clone YTS 169.4, BE0017, BioXCell, USA) or anti-NK mAb (clone PK136, BE0036, BioXCell, USA) was injected intraperitoneally every 2 weeks from 1 day before therapy. Depletion of CD8 or NK cells was confirmed by fluorescence-activated cell sorting (FACS) analysis of peripheral blood.

For all the in vivo experiments, tumor volume was monitored every 2–3 days by caliper measurement and calculated by length × width²/2. Mice were euthanized when tumor volume reached 2 cm³, or when mice appeared moribund.

For CD8⁺NKG2D⁺ cells analysis, cells were harvested and stained with anti-CD3-FITC (E10472-1633, eBioscience, USA), anti-CD8-PE (561946, BD, USA), anti-NKG2D-APC (130-099-216, Miltenyi Biotec, Germany), anti-CD27-PE (560985, BD, USA), and anti-CD57-APC (560845, BD, USA) antibodies. To detect the infiltration of CD8 or NK cells in the tumor mass, tumors were removed and crushed into single cell suspension. Then cells were stained with mouse anti-CD8-PerCp (46-0083-80, eBioscience, USA) or anti-NK1.1-FITC (553164, BD, USA) antibodies. To confirm the depletion of CD8 or NK cells, cells were isolated from peripheral blood and stained with anti-CD8-PerCp (46-0083-80) or anti-NK1.1-PerCp antibodies (45-5941-82) (all from eBioscience, USA). The staining was performed on ice for 30 min. Samples were acquired on the FACS Calibur (BD, USA) and data was analyzed using FlowJo software (Version 8.5.3, Tree Star Inc., USA).

Immunity activity in the TME was measured with mouse IFN-γ ELISpot Kit (3321-2A, Mabtech, Sweden) according to the guidelines. In brief, the plate was coated with mouse INF-γ capture antibodies and kept overnight at 4 °C in the dark. Cells isolated from tumor masses were seeded at a density of 1.5 × 10⁵ cells/well. After 24 h incubation at 37 °C in a humidified atmosphere of 5% CO₂ incubator, cells were removed and the wells were rinsed with PBS. The biotinylated detection antibodies against IFN-γ were then added and the plate was incubated for 2 h at RT. Then the wells were rinsed with PBS followed by incubation with Streptoavidin-ALP for 1 h. The substrate BCIP/NBT-plus was then added. The reaction was stopped by washing wells with tap water when spots emerged. The plate was subjected to ELISpot reader system (Autoimmun Diagnostika GmbH, Germany) to detect spots number and activity. The activity of spot was characterized as the average of spot size and intensity for each well.

The results are presented as mean ± SD. Statistical analyses were performed using Student’s t test. Animal survival is presented using Kaplan–Meier survival curves and was statistically analyzed using log rank test (GraphPad Prism version 6). p value < 0.05 was considered to be statistically significant.

To investigate the anti-inflammatory activity of FM, cytotoxicity assay of FM on HCC cells was firstly conducted to obtain non-toxic concentrations (Fig. 1a). We found that FM strongly inhibited the phosphorylation of IκBα both in the absence or presence of TNF-α in a dose-dependent manner (Fig. 1b). Of note, at a non-toxic concentration of 0.5 mg/ml, FM exerted a significant reduction of p-IκBα. In line, the expression of NF-κB-regulated genes including TNF-α and IL-6 was significantly decreased in HCC cells treated with FM in the absence or presence of TNF-α stimulation (Fig. 1c). Meanwhile, the level of phosphorylated STAT3 was also dramatically reduced (Fig. 1d). Moreover, FM also massively inhibited the phosphorylation of STAT3 stimulated by IL-6 in HCC cells (Fig. 1e). Therefore, our data confirm that FM possesses a potent inhibitory effect on NF-κB and STAT3 activation.

Next, we wanted to know whether reduced activation of NF-κB and STAT3 could result in an improved antitumor immunity. To this end, CD8⁺NKG2D⁺ cells were employed due to their NKT character. CD8⁺NKG2D⁺ cells were generated from human peripheral blood and could be activated directly by their cognate ligands such as MICA/B often expressed on the surface of tumor cells. Therefore, no antigen-presenting cells were needed for the activation of T cells in the in vitro settings. Previous study found that naïve-derived effector CD8⁺ T cells, which expressed higher CD27 and lower CD57, exhibited higher proliferative potential with better therapeutic efficacy in clinic [25]. Hence, the differentiation status of CD3⁺ T cells was firstly qualified by CD27 or CD57 antibodies staining. Next, we determined the frequency of CD8 and NKG2D double positive cells gated from CD3⁺ cells (Fig. 2a). At a non-toxic dose (0.5 mg/ml), FM markedly reduced the activation of NF-κB and STAT3 in the mixture of HCC and CD8⁺NKG2D⁺ cells (Fig. 2b), and downregulated the expression of TNF-α and IL-6 (Fig. 2c). We then further determined the oncolysis of CD8⁺NKG2D⁺ cells on HCC cells pre-treated with or without FM. We found that FM pre-treated HCC cells were more sensitive to CD8⁺NKG2D⁺ cell-mediated oncolysis (Fig. 2d). These results show that FM reduces immune cells-triggered inflammation and enhances the sensitivity of HCC cells to the cytotoxicity of immune cells.

It has been shown that the purified extract of frankincense (boswellic acid), essential oil of frankincense, and extract of myrrh (guggulsterone) exhibit antitumor capabilities in immunocompromised mice [18, 26, 27]. However, it is not yet clear if the water-decocting FM, which is commonly used in clinic, also has the similar antitumor efficacy. Then we performed an in vivo experiment in nude mice bearing subcutaneous human HCC. The mice were treated with FM at a dose which previously has been shown to inhibit inflammatory responses in an arthritis rat model [17]. To our surprise, FM failed to inhibit tumor growth, and there was no difference of survival time between treated and untreated groups (Fig. 3a, c). Next, we wanted to know the antitumor efficiency of FM in immune competent mice. Therefore, we performed another in vivo experiment in HCC-bearing mice with intact immunity using the same dose of FM. Indeed, FM significantly inhibited tumor growth and prolonged lifespan of mice from an average of 15 d in control group to an average of 22 d in FM treated group (Fig. 3d, f). In both tumor models, FM had no obvious influence on the body weight (Fig. 3b, e). These data suggest that the water-decocting FM at the anti-inflammatory dose (non-toxic dose) also possesses an antitumor ability, which depends on the host immunity system.

Having shown that the antitumor capability of FM is immune-dependent, we next investigated if the immune activity in the TME could be modulated by FM. In the immune competent C57BL/6 mice, we established a subcutaneous murine HCC model and monitored the tumor growth and body weight during the intervention (Fig. 4a). Then, using ELISpot assay, we found that the amount of antitumor cytokine IFN-γ was robustly upregulated within the TME from the mice treated with oral garage of FM (Fig. 4b). Then we asked if this effect was contributed by increased infiltration of immune cells. However, by determination of CD8⁺ T and NK cells from the cell suspensions isolated from tumor masses, we did not observe the increased infiltration of either CD8⁺ T (upper panel) or NK cells (lower panel) in FM treated mice (Fig. 4c). The results indicate that FM may restore immune responses against HCC in the TME.

Next, we wanted to figure out which subset of lymphocytes contributed to the antitumor activity of FM. In immune competent mice bearing HCC, we depleted either CD8⁺ T cells by anti-CD8 or NK cells by anti-NK neutralizing antibodies 1 day before FM treatment (Fig. 5a). The depletion efficacy was monitored 1 and 13 days after administration of antibodies by flow cytometry (Fig. 5b). Then we monitored the tumor growth, and found that the antitumor effect of FM was completely abrogated by the depletion of CD8⁺ T cells, but not by the depletion of NK cells (Fig. 5c, d). No side effects were observed during the experiment as monitored by body weight (Fig. 5e). In conclusion, our results illustrate CD8⁺ T cells mediate the antitumor activity of FM.