- Email: [email protected]
Ketogenic diet combined with antioxidant N-acetylcysteine inhibits tumor growth in a mouse model of anaplastic thyroid cancer
Surgery xxx (2019) 1e7
Contents lists available at ScienceDirect
Surgery journal homepage: www.elsevier.com/locate/surg
Ketogenic diet combined with antioxidant N-acetylcysteine inhibits tumor growth in a mouse model of anaplastic thyroid cancer Abha Aggarwal, PhDa, Zuliang Yuan, MSa, Justine A. Barletta, MDb, Jochen H. Lorch, MDc, Matthew A. Nehs, MDa,* a
Department of Surgery, Brigham and Women’s Hospital, Boston, MA Department of Pathology, Brigham and Women’s Hospital, Boston, MA c Head and Neck Center, Dana Farber Cancer Institute, Boston, MA b
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 2 June 2019 Available online xxx
Background: Anaplastic thyroid cancer is an aggressive and fatal malignancy. Many advanced cancers are characterized by glucose dependency, leading to oxidative stress and cellular proliferation. Therefore, we sought to determine if a low glucose environment (in vitro) or a ketogenic diet (in vivo) could inhibit anaplastic thyroid cancer tumor growth when combined with the antioxidant N-acetylcysteine. Methods: In vivo, nude mice were injected with the anaplastic thyroid cancer cell line 8505C (n ¼ 6/ group). Group 1 was fed a standard diet; Group 2 was fed a ketogenic diet; Group 3 was given standard diet with N-acetylcysteine (40 mM in the drinking water); and Group 4 was fed ketogenic diet with Nacetylcysteine. Tumor volumes, ketones, and glucose were measured. H&E stains and immunohistochemistry for Ki-67 and Caspase 3 were performed on the tumors. In vitro, 8505C cells were cultured in high glucose (25 mM), low glucose (3 mM), high glucose plus N-acetylcysteine (200 uM), or low glucose plus N-acetylcysteine for 96 hours. We performed CyQUANT proliferation (Thermo Fisher Scientific, Waltham, MA), Seahorse glycolytic stress (Agilent, Santa Clara, CA), and reactive oxidative stress assays. Results: Ketogenic diet plus N-acetylcysteine decreased in vivo tumor volume compared to standard diet (22.5 ± 12.4 mm3 vs 147 ± 54.4 mm3, P < .05) and standard diet plus N-acetylcysteine (P < .05). Blood ketone levels were significantly higher for the mice in the ketogenic diet group compared to standard diet (1.74 mmol/L vs 0.38 mmol/L at week 5, P < .001). However, blood glucose levels were not significantly different between ketogenic diet and standard diet groups. Cells cultured in low glucose plus Nacetylcysteine had significantly reduced proliferation compared to high glucose (98.1 ± 5.0 relative fluorescence units vs 157.8 ± 2.1 relative fluorescence units, P < .001). Addition of N-acetylcysteine to low glucose lowered glycolysis function compared to high glucose (39.0 ± 2.2 mpH/min/cell vs 89.1 ± 13.2 mpH/min/cell, P < .001) and high glucose plus N-acetylcysteine (37.4 ± 2.5 mpH/min/cell vs 70.3 ± 3.3 mpH/min/cell, P < .001). Low glucose plus N-acetylcysteine decreased reactive oxidative stress compared to high glucose (119 ± 34.7 relative fluorescence units vs 277 ± 16.0 relative fluorescence units, P ¼ .014). Conclusion: The combination of a ketogenic diet or glucose restriction with the antioxidantN-acetylcysteine significantly reduced tumor growth in vivo and in vitro. Further studies are warranted to explore these metabolic therapies in anaplastic thyroid cancer. © 2019 Elsevier Inc. All rights reserved.
Introduction
Presented at the American Association of Endocrine Surgeons, Los Angeles CA, April 7-9, 2019. * Reprint requests: Matthew Nehs, MD, Assistant Professor of Surgery, Harvard Medical School, Associate Surgeon, Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115. E-mail address: [email protected] (M.A. Nehs). https://doi.org/10.1016/j.surg.2019.06.042 0039-6060/© 2019 Elsevier Inc. All rights reserved.
Anaplastic thyroid cancer (ATC) is the most aggressive and fatal type of thyroid cancer, accounting for up to 40% of thyroid carcinoma deaths.1,2 Surgical resection is often not possible at presentation because of local invasion into the trachea, larynx, or because it surrounds the carotid artery. Current, nonoperative therapies (chemotherapy, targeted therapy, and radiation) become ineffective due to evolved tumor resistance.3,4 Despite multimodal
2
A. Aggarwal et al. / Surgery xxx (2019) 1e7
interventions, patients with ATC continue to have a median survival of only 5 to 6 months upon diagnosis.5 Thus, there is an urgent need to apply new treatment paradigms in patients with ATC. One such paradigm views the cancer cell in part as a disease of cellular metabolism. In many aggressive malignancies, cancer cells have a metabolic phenotype that is characterized by dysfunctional mitochondrial and a shift to glycolysis as the principal method of energy production (known as the Warburg effect).6 The increased glucose utilization by cancer cells is demonstrated clinically through positron emission tomography scans where patients are injected with radiolabeled fluorodeoxyglucose, which is readily absorbed by the tumor cells.7 Although glucose uptake and glycolysis is dramatically increased in many advanced malignancies, very little of the end product of glycolysis (pyruvate) is channelized into the tricarboxylic acid cycle in the mitochondria. Instead, pyruvate is reduced to form lactate, which also generates NADþ. This allows glycolysis to continue even in the presence of normal oxygen levels (normoxia).5 Since many cancer types exhibit glucose dependency by the Warburg effect,6,8 we hypothesized that glucose restriction in vitro and carbohydrate restriction in vivo through a ketogenic diet (KD) may slow cancer growth and progression. The KD is characterized by a high fat (~90%), moderate protein (~10%), and very low carbohydrate (<1%) macronutrient composition. This diet depletes hepatic and muscular glycogen and shifts the body’s metabolism to fatty acid oxidation.9 This suppresses insulin levels and blood glucose levels while increasing blood ketone levels (eg, acetoacetate and ß-hydroxybutyrate).9 This diet has been used in other preclinical models of advanced malignancy, including glioblastoma multiforme9; however, it has not yet been tested in a mouse model of ATC. In addition to the alteration of the Warburg phenotype seen in ATC, many advanced malignancies generate high levels of reactive oxygen species (ROS) that can drive cellular signaling.10,11 For example, ROS are involved in signaling pathways in proliferation, metabolism, angiogenesis, growth, and survival.11 Thus, decreasing ROS-oxidative stress levels in cancer cells by administering exogenous antioxidant may yield an anti-proliferative effect by diminishing the proliferative signaling mediated by ROS.12 Therefore in this study, we tested the hypothesis that a KD combined with an antioxidant (N-acetylcysteine[NAC]) could inhibit tumor growth in a mouse model of ATC. Materials and methods Cell line and culturing We used the ATC line 8505C (Sigma Aldrich, St. Louis, MO) for all the in vivo and in vitro experiments. This cell line was derived from a 78-year-old female patient. The cells were cultured using high glucose (HG) (25 mM) or low glucose (LG) (3 mM) in DMEM/F12 with L-Glutamine media that was completed with 10% fetal calf serum and 100 units/mL penicillin-streptomycin at 37 C and 5% CO2. All reagents used for culturing were purchased from Thermo Fisher Scientific. Both HG and LG cultures were treated with NAC (200 uM) (Thermo Fisher Scientific, Waltham, MA). In vivo mouse modeling Mice and tumor injections Nude mice (6e8-week-old females) were purchased from the Jackson Laboratory, Bar Harbor, ME. All mice protocols were followed in accordance with the National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee. Through the entire study, the mice were housed (3 mice/cage) in a biosafety level 2 facility in an institutional
vivarium. All mice (N ¼ 24) were acclimatized for a week and then injected with the tumor line 8505C into the flank. Weight, tumor volume, blood glucose, and ketones were measured each week. 8505C cells were seeded in 4-T75 flasks and cultured until just before confluency (4 days). Then, the cells were counted used Trypan Blue and 1 x 106 cells were suspended in 0.3 cc saline solution and injected subcutaneously in the right flank of each mouse, using a 1 cc tuberculin syringe. Mice were examined biweekly by a veterinarian for signs of distress and toxicity. Treatments and toxicity assessment The mice were randomly divided into 4 groups (6 mice/group). The tumors were allowed to grow until they were palpable before initiating each treatment. All treatments were initiated on the ninth day after injection. The mice in the KD groups were fasted overnight to initiate ketosis on the eighth day post-injection.13 Group 1 was fed the standard diet (SD); Group 2 was fed a KD; Group 3 was given SD with antioxidant NAC (40 mM) supplemented in the drinking water13; and Group 4 was given KD supplemented with NAC. Macronutrients in the SD (PicoLab Mouse Diet 20 [LabDiet, St. Louis, MO]) consisted of 55% carbohydrates, 22% fat, and 23% protein, whereas the KD had 89.9% fat,10% protein, and 0.1% carbohydrate. The KD was purchased from Research Diets Inc. (New Brunswick, NJ, USA). Tumor volumes on all 24 mice were measured using digital calipers and calculated as ½ length x width x height. Blood glucose and ketones were measured using glucose and ketone strips (Thermo Fisher Scientific, Waltham, MA) for the SD and KD treated mice.13 Mouse tumor immunohistochemistry (IHC) Tumor samples were procured from all the mice, once euthanized, for each of the 4 treatment groups. Slides were stained with H&E, Ki-67 for proliferation, and Caspase-3 to evaluate for apoptosis. All the slides were made and stained at the IHC Core Laboratory, Brigham and Women’s Hospital. In vitro assays Seahorse glycolytic stress assay The glycolysis function and capacity of the cells under different conditions were measured through extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) using Seahorse XF Glycolysis Stress Kit and Analyzer (Agilent, Santa Clara, CA). Changes in ECAR and OCR after the sequential injection of glucose (triggering glycolysis), oligomycin (inhibiting mitochondrial adenosine triphosphate synthase), and 2-deoxy-D-glucose (inhibiting glycolysis) represents the cellular glycolysis and its maximum capacity, respectively.14 Seahorse XF96 microplates (Agilent, Santa Clara, CA) were coated with 0.1mg/ml poly-D-lysine. 8505C cells were cultured with high or LG medium with or without NAC for 96 hours and seeded (4 x 104) in triplicates. XF assay medium was freshly prepared with XF base medium supplemented with 1 mM glutamine. Cells were washed twice using assay medium and incubated in a non-CO2 incubator for 1 hour. After calibration and initial measurements at baseline, a final concentration of 10 mM glucose, 1mM oligomycin, and 50 mM 2-deoxy-D-glucose were injected sequentially. Three measurements were taken after each injection. Cell number was counted using C-chip (Thermo Fisher Scientific, Waltham, MA) after the assay and was used to normalize the raw ECAR dataset. Intracellular ROS assay Intracellular ROS Oxidative Assay (Cell Biolabs, Inc, San Diego, CA) measures the antioxidant or ROS in a cell.15,16 The assay uses a cell-permeable fluorogenic probe 2’, 7’ -Dichlorodihydrofluorescein diacetate (DCFH-DA), which diffuses into the cell. DCFH-DA is then
A. Aggarwal et al. / Surgery xxx (2019) 1e7
3
Fig 1. Mouse tumor volume after 5 weeks. (A) The average tumor volume (mm3) for each group throughout the experiment period. (B) The endpoint (week 5) tumor volumes and each dot represents an individual mouse for that treatment group. N ¼ 6 for each group except for KD group n ¼ 5 (one died during week 2). *P < .05. (C) The blood glucose for each mouse for the 5-week period (P > .05). (D) The blood ketone for each mouse for the 5-week period (*P < .01). The error bar represents the SEM of each group. (E) Representative tumor images comparing a SD mouse to a KD þ NAC mouse.
deacetylated by cellular esterases to 2’, 7’-Dichlorodihydrofluorescein. When the Dichlorodihydrofluorescein molecule encounters ROS, it is rapidly oxidized to the fluorescent compound 2’, 7’ Dichlorofluorescein, and the fluorescence is measured using a plate-reader.17 8505C cells were seeded in a 96-welled plate (5,000 cells/well) overnight. Then, the cells were pretreated with DFCH-DA (1X), incubated at 37 C for 60 minutes, and washed with Hanks Balanced Salt Solution to remove excess DFCH-DA. After removing DFCH-DA, the cells were treated with HG, LG, HG þ NAC, or LG þ NAC for 96 hours. Fluorescence was measured using a SoftMax Pro plate reader (Molecular Devices Ltd, San Jose, California, USA) at 485 nM (excitation) and 538 nM (emission) wavelengths. Each condition was performed in triplicates. Proliferation assay We tested for cell proliferation using CyQUANT GR proliferation assay. This assay uses a fluorescent dye that binds to the cell’s nucleic acids.18 We performed the CyQUANT GR Proliferation assay in triplicates for each treatment condition as per the manufacturer’s guidelines. Cells were seeded in a 96-welled plate (5,000 cells/well) in triplicates and then exposed to HG, LG, HG þ NAC, or LG þ NAC for 24, 72, and 96 hours. Culture media was removed and 200 uL of the CyQUANT GR working solution was applied to the cells and incubated for 5 minutes at room temperature. Fluorescence was measured at 480 nM (excitation) and 520 nM (emission) with the SoftMax Pro software. Statistical analysis Statistical analysis was performed using Students t-tests (unpaired) on GraphPad Prism 6 (GraphPad Software, Inc, San Diego, CA). The Wave software was used for the Seahorse Glycolytic Stress Test. All independent treatment conditions were performed in triplicates and were computed as means and ± standard error of the mean (SEM).
Results In vivo - KD with NAC significantly slowed tumor growth in a mouse model Mice were fed SD, KD, SD þ NAC, or KD þ NAC and humanely sacrificed 5 weeks after treatment. Tumor volumes were markedly reduced in the group that was fed the KD þ NAC, compared to the SD group (21.5 ± 12.6 mm3 vs 147.2 ± 54.4 mm3, P < .05). KD þ NAC mice also had significantly smaller tumors compared to the SD þ NAC treatment group (21.5 ± 12.6 mm3 vs 321.9 ± 128.6 mm3, P < .05). Two of the 6 mice (33%) in the KD þ NAC group showed no viable tumor at all by the fifth week on final histopathology, as reviewed by our endocrine pathologist. The KD treatment group had a mean tumor volume of 67.3 ± 44.2 mm3. The SD þ NAC group showed the highest tumor volume by the fifth week (321.9 ± 128.6 mm3) but also had the largest variance in the data (Fig 1, A, B). Blood glucose and ketone measurements were taken each week for the SD and KD groups, and mean values calculated on the fifth week. Interestingly, the blood glucose levels between the groups did not show a significant difference between the KD and SD groups (Fig 1, C). However, the blood ketone levels were significantly higher for the mice in the KD group compared to SD (1.74 vs 0.38 mmol/L at week 5, P < .001) (Fig 1, D), thus confirming a state of ketosis. We found no significant differences in the weight of each mouse among the 4 groups of mice with various treatments during the 5 weeks of the study. No apparent signs of NAC toxicity were evident, such as cachexia due to weight loss, lethargy, urine discoloration or anuria, hypothermia, or general appearance (Grimace scale).19
IHC All our IHC slides were assessed by an endocrine pathologist. No significant differences were noted for H&E, Ki-67, and Caspase-3 stains between any of the treatment groups (Figs 2, AeF).
4
A. Aggarwal et al. / Surgery xxx (2019) 1e7
Fig 2. Immunohistochemical analysis for H&E, Ki67, and Caspase-3. (A, C, and E) The IHC staining for standard diet (SD) group. (B, D, and F) The IHC staining for ketogenic diet plus NAC (KD þ NAC) group. All pictures were taken under 20X magnification.
Interestingly, the small tumors in the KD þ NAC group showed large areas of central necrosis that did not stain for Caspase-3 (Fig 2). NAC decreases glycolytic stress We used the Seahorse Glycolysis Stress assay to measure the ECAR and OCR to determine 8505C glycolysis function and capacity for each group (HG, LG, HG þ NAC, and LG þ NAC) (Figs 3,
A, B). Glycolytic function was reduced significantly with the addition of NAC to HG (24.0 ± 2.4 mpH/min/cell) and NAC to LG (23.7 ± 1.8 mpH/min/cell) (P < .001) compared to HG alone (60.4 ± 1.7 mpH/min/cell) or LG alone (54.9 ± 11 mpH/min/cell). Similarly, the glycolysis capacity was significantly lowered with the addition of NAC to HG (37.4 ± 2.5 mpH/min/cell) and NAC to LG (39.0 ± 2.2 mpH/min/cell) compared to either the HG (70.3 ± 3.3 mpH/min/cell) or LG (89.1 ± 13.2 mpH/min/cell)
Fig 3. Seahorse glycolytic stress assay. (A) The extracellular acidification rate (ECAR) and (B) The oxygen consumption rate (OCR) of 8505C cells treated with high glucose (HG, (25mM), low glucose (LG) (3mM), high glucose plus 200mM NAC (HG þ NAC), low glucose plus 200mM NAC (LG þ NAC) in triplicates. All cells were treated for 96 hours. (A and B) The overall ECAR and OCR change throughout the assay, respectively. (C) The average glycolysis rate and glycolytic capacity for each group. The error bar represents the SEM for each group. **P < .001.
A. Aggarwal et al. / Surgery xxx (2019) 1e7
5
LG þ NAC compared to LG treatment alone at 96 hours of exposure (98.1 ± 5.0 RFU vs 146.9 ± 6.9 RFU, P < .01) (Fig 5, B). Discussion
Fig 4. ROS assay. Fig 4 shows the relative fluorescent units (RFUs) for 8505C cells treated with high glucose (HG) (25mM), low glucose (LG) (3mM), high glucose plus 200mM NAC (HG þ NAC), and low glucose plus 200mM NAC (LG þ NAC) in triplicates. Treatment duration was 96 hours. The error bar represents SEM for each group. *P < .05.
groups (P < .001) (Fig, 3, C). LG treatments were not significantly different from HG. Low glucose and NAC reduce ROS-oxidative stress Examining the effect of NAC in cellular oxidative stress, we found that only LG with NAC treatment significantly reduced ROSoxidative stress levels when compared to the HG alone (119 ± 34.7 relative fluorescence units [RFU] vs 277.4 ± 16.0 RFU, P < .02). ROS levels were found to be intermediate values for the LG (226 ± 44 RFU) and HG þ NAC (195 ± 60 RFU) treatment groups (Fig 4). Low glucose and NAC attenuates cellular proliferation We used HG and LG media to mimic the effect of SD and KD in vitro and added 200 mM NAC to the media to compare the proliferation rates after 24-, 72-, and 96-hour treatments of 8505C cells. A significant reduction in proliferation was seen in cells treated with LG þ NAC when compared to HG at 72 hours (94.5 ± 7.6 RFU vs 134.4 ± 9.9 RFU, P < .04) (Fig 5, A), which was further heightened at 96 hours of exposure (98.1 ± 5.1 RFU vs 157.8 ± 2.1 RFU, P < .001) (Fig 5, B). Proliferation reduction was also noted with
Many advanced malignancies are characterized by a reprogramming of energy metabolism leading to increased glucose consumption.20 Cancer cells that exhibit this phenotype have a very high rate of glycolysis that ends in the production of lactate, even in the presence of sufficient oxygen (aerobic glycolysis). This phenomenon is referred to as the Warburg effect, named after Otto Warburg, who first described this in 1925.21 This process involves a high rate of glycolysis followed by the conversion of pyruvate to lactate, which regenerates NADþ and allows glycolysis to continue. Additionally, this increase in glycolysis drives the production of ROS, which have been shown to play an important role in cellular signaling pathways for proliferation, metabolism, angiogenesis, growth, and survival in several advanced malignancies including melanoma22 and pancreatic adenocarcinoma.23 Additionally, the antioxidant NAC has been shown to reduce ROS levels thereby curtailing oxidative stress in cancer cells and limiting their rate of cellular proliferation.12 Given that ATC is one of the fastest growing and most lethal malignancies known, we sought to determine the effect of NAC and glucose restriction on growth of the ATC cell line 8505C. We chose to apply this metabolic targeting to ATC since it has a very high rate of proliferation and most current therapies remain ineffective.1e4 In order to simulate a LG environment in vivo, we utilized the ketogenic diet, which is characterized by a high fat (~90%), moderate protein (~10%), and very low carbohydrate (<0.1%) macronutrient composition. The KD pushes the body’s metabolism from using glucose to fat and ketones as energy sources, and it has previously been shown to slow the progression of several types of malignancies including glioblastoma multiforme.9 Our study found that ATC tumor growth was dramatically inhibited in mice that were fed the KD supplemented with antioxidant NAC in the drinking water. Additionally, we found no histologic evidence of viable tumor in 2 mice in the KD þ NAC group. Interestingly, we found the largest average tumor volumes in the SD þ NAC group, although the variance in the data was large and may reflect random variation in a small sample size. As expected, the mice in the KD group had higher average ketone levels than the mice fed the standard chow (SD), but interestingly, there were no significant differences in blood glucose concentrations between groups. This suggests that the KD may exert an anticancer effect that is independent of blood glucose levels, which has been previously demonstrated.24 We did not observe any IHC differences in
Fig 5. Proliferation assay results of 8505C after 96-hour treatment. (A) The relative fluorescent units (RFUs) for 8505C cells treated with high glucose (HG) (25mM), low glucose (LG) (3mM), high glucose plus 200mM NAC (HG þ NAC), and low glucose plus 200mM NAC (LG þ NAC) in triplicates for 24, 72, and 96 hours. (B) The 96 hours bar graph of the relative fluorescent units (RFUs) for each group. The error bar represents SEM for each group. **P < .01, ***P < .001.
6
A. Aggarwal et al. / Surgery xxx (2019) 1e7
the Caspase-3 and Ki-67 activity between the various treatment groups used in the study. However, the KD þ NAC IHC revealed large areas of central necrosis in very small tumors, which suggests that the cells were dying due to non-apoptotic mechanisms. Our in vitro data support the in vivo findings: the cells cultured in LG supplemented with NAC showed a significantly lower cellular proliferation, reactive oxygen species-mediated oxidative stress, and glycolytic stress when compared to the cells cultured in HG media. Mechanistically, our study showed that restricting glucose in combination with antioxidant NAC results in decreased glycolytic stress and inhibition of ROS signaling, thereby reducing tumor cell proliferation. Our study has several important limitations. First, we used a single cell line (8505C) as the basis for our animal experiment as a proof of principle experiment for the study design. Further studies should expand the cell lines used and vary the mutational profile of the tumors to look for consistent or disparate effects. The heterotopic flank tumor injection is also a limitation of the study compared to an orthotopic model, which has been previously described.25 Future studies will assess the role of the KD on an orthotopic mouse model with lung metastases. Despite these limitations, we believe that this study is the first to demonstrate an effect of the KD/metabolic therapy in a mouse model of ATC. In conclusion, our study demonstrates that a KD combined with the antioxidant NAC decreased tumor growth in a mouse model of ATC. Further studies are needed to determine if targeting cancer cell glucose dependency and ROS-mediated cellular proliferation is an effective adjuvant treatment in patients with ATC. Funding/Support This study was funded by the Robert T. Osteen Fellowship Award and the Department of Surgery, Brigham and Women's Hospital, Boston, MA. Conflict of interest/Disclosure Authors Abha Aggarwal, Zuliang Yuan, Justine A. Barletta, and Matthew A. Nehs do not have any conflicts of interest. Dr. Jochen H. Lorch has research that is supported by Novartis, Bayer, BMG, and Millennium. Dr. Lorch also receives consulting honoraria from Bayer and Genentech. References 1. Hundahl SA, Fleming ID, Fremgen AM, Menck HR. A national cancer data base report on 53,856 cases of thyroid carcinoma treated in the U.S., 1985e1995. Cancer. 1998;83:2638e2648. 2. Chiacchio S, Lorenzoni A, Boni G, Rubello D, Elisei R, Marianni G. Anaplastic thyroid cancer: Prevalence, diagnosis and treatment. Minerva Endocrinol. 2008;33:341e357.
3. Pinchot SN, Sippel RS, Chen H. Multi-targeted approach in the treatment of thyroid cancer. Ther Clin Risk Manag. 2008;4:935e947. 4. Wagle N, Grabiner BC, Van Allen EM, et al. Response and acquired resistance to everolimus in anaplastic thyroid cancer. N Engl J Med. 2014;371:1426e1433. 5. Smallridge RC, Copland JA. Anaplastic thyroid carcinoma: pathogenesis and emerging therapies. Clin Oncol. 2010;22:486e497. 6. Lu Z, Guo Y, Zhang X, et al. ORY-1001 suppresses cell growth and induces apoptosis in lung cancer through triggering HK2 mediated Warburg effect. Front Pharmacol. 2018;9:1411. 7. Chang HT, Hu C, Chiu YL, Peng NJ, Liu RS. Role of 2-[18F] fluoro-2-deoxy-Dglucose-positron emission tomography/computed tomography in the posttherapy surveillance of breast cancer. PLoS One. 2014;9:e115127. 8. Coelho RG, Fortunato RS, Carvalho DP. Metabolic reprogramming in thyroid carcinoma. Front Oncol. 2018;8:82. 9. Poff AM, Ward N, Seyfried TN, Arnold P, D’Agostino PD. Non-toxic metabolic management of metastatic cancer in VM mice: Novel combination of ketogenic diet, ketone supplementation, and hyperbaric oxygen therapy. PLoS One. 2015;10:e0127407. 10. Stafford P, Abdelwahab MG, Kim Y, Preul MC, Rho JM, Scheck AC. The ketogenic diet reverses gene expression patterns and reduces reactive oxygen species levels when used as an adjuvant therapy for glioma. Nutri Metabol (Lond). 2010;7:74. 11. Schieber M, Chandel NS. ROS Function in redox signaling and oxidative stress. Curr Biol. 2014;24:R453eR462. 12. Monti D, Sotgia F, Whitaker-Menezes D, et al. Pilot study demonstrating metabolic and anti-proliferative effects of in vivo anti-oxidant supplementation with N-Acetylcysteine in breast cancer. Semin Oncol. 2017;44: 226e232. 13. Gao P, Zhang H, Dinavahi R, et al. HIF-dependent antitumorigenic effect of antioxidants in vivo. Cancer Cell. 2007;12:230e238. 14. Wu M, Neilson A, Swift AL, et al. Multiparameter metabolic analysis reveals a close link between attenuated mitochondrial bioenergetic function and enhanced glycolysis dependency in human tumor cells. Am J Physiol Cell Physiol. 2007;292:C125eC136. 15. Kim J, Chen C, Yang J, Mochly-Rosen D. Aldehyde dehydrogenase 2*2 knock-in mice show increased reactive oxygen species production in response to cisplatin treatment. J Biomed Sci. 2017;24:33e40. 16. Stebbins KJ, Broadhead AR, Cabrera G, et al. In vitro and in vivo pharmacology of NXT629, a novel and selective PPARa antagonist. Eur J Pharmacol. 2017;809: 130e140. 17. Azimi I, Peterson RM, Thompson EW, Roberts-Thomson SJ, Monteith GR. Hypoxia-induced reactive oxygen species mediate N-cadherin and SERPINE1 expression, EGFR signaling and motility in MDA-MB-468 breast cancer cells. Sci Rep. 2017;7:15140. 18. Jones LJ, Gary M, Yue ST, Haugland P, Singer VL. Sensitive determination of cell number using the CyQUANT cell proliferation assay. J Immunol Methods. 2001;254:85e98. 19. Burkholder T, Foltz C, Karlsson E, Linton CG, Smith JM. Health evaluation of experimental lab mice. Curr Protoc of Mouse Biol. 2012;2:145e165. 20. Vander Heiden MG, DeBerardinis RJ. Understanding the intersections between metabolism and cancer biology. Cell. 2017;168:657e669. 21. Warburg O. The metabolism of carcinoma cells. The Journal of Cancer Research. 1925;9:148e163. 22. Bisevac JP, Djukic M, Stanojevic I, et al. Association between oxidative stress and melanoma progression. J Med Biochem. 2018;37:12e20. 23. Du J, Liu J, Smith BJ, Tsao MS, Cullen JJ. Role of Rac1-dependent NADPH oxidase in the growth of pancreatic cancer. Cancer Gene Ther. 2011;18: 135e143. 24. Fine EJ, Segal-Isaacson CJ, Feinman RD, et al. Targeting insulin inhibition as a metabolic therapy in advanced cancer: A pilot safety and feasibility dietary trial in 10 patients. Nutrition. 2012;28:1028e1035. 25. Nehs MA, Nucera C, Nagarkatti SS, et al. Late intervention with antiBRAF(V600E) therapy induces tumor regression in an orthotopic mouse model of human anaplastic thyroid cancer. Endocrinology. 2012;153: 985e994.
Discussion Dr Eren Berber (Cleveland, OH): I think that as surgeons dealing with cancer outside the thyroid as well, we are seeing that patients are putting themselves on these ketogenic diets even before they come to see us. In addition to the ketogenic diet, we are also seeing that some patients are actually taking Metformin as well, especially for differentiated thyroid cancer. We had a really interesting patient
who was in her 30s and had a 1.5-centimeter papillary thyroid cancer. She refused all surgical therapy and put herself on a ketogenic diet and Metformin. When we saw the patient two years later, the cancer had not advanced on ultrasound and had actually decreased in size to 8 millimeters. I wonder if you have had any experience with Metformin in this context.
A. Aggarwal et al. / Surgery xxx (2019) 1e7
Dr Abha Aggarwal: Thank you for the question, but we have not tested Metformin in any of our experiments. Dr Mahsa Javid (Charleston, SC): Very impressive study. Thank you. I have some related questions. What is your hypothesis about the non-apoptotic causes of reduction in tumor volume? Why do you think there was a difference between the in vitro and in vivo cellular proliferation rate? And did you check for invasion and migration? Dr Abha Aggarwal: It was interesting to us as well. One thing that we can say is that we don't understand the mechanism of how the ketogenic diet works. These were very surprising results for us as well because we thought there would be some kind of significant difference in proliferation since we were seeing that in vitro, but we didn't observe that. We don't know the mechanism of it, basically, but those are great questions.
7
Dr Mark Cohen (Ann Arbor, MI): Two quick questions. Fat storage in mice is a little different than fat storage in humans, and a ketogenic diet has a slightly different effect. When you showed your in vivo data, actually the ketogenic diet by itself had a significant impact on tumor growth compared to the standard diet. Do you think that a lot of this effect is just from the diet, or is the N-acetylcysteine really adding a lot? Dr Abha Aggarwal: I think the most significant effects we were seeing were in combination of N-acetylcysteine and the low glucose or the ketogenic diet. We did not see significant effects just with the ketogenic diet alone. So it was just more of an interaction. We don't know what the ketogenic diet is doing as far as mechanism goes. But we hypothesized that it is increasing the ketosis, so it is promoting fat metabolism, and we know that it has 85% to 90% fat in the diet already, and somehow that's interacting with the N-acetylcysteine (or NAC). But we don't know the exact mechanism of it. That's still a future study for us to do.
Contents lists available at ScienceDirect
Surgery journal homepage: www.elsevier.com/locate/surg
Ketogenic diet combined with antioxidant N-acetylcysteine inhibits tumor growth in a mouse model of anaplastic thyroid cancer Abha Aggarwal, PhDa, Zuliang Yuan, MSa, Justine A. Barletta, MDb, Jochen H. Lorch, MDc, Matthew A. Nehs, MDa,* a
Department of Surgery, Brigham and Women’s Hospital, Boston, MA Department of Pathology, Brigham and Women’s Hospital, Boston, MA c Head and Neck Center, Dana Farber Cancer Institute, Boston, MA b
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 2 June 2019 Available online xxx
Background: Anaplastic thyroid cancer is an aggressive and fatal malignancy. Many advanced cancers are characterized by glucose dependency, leading to oxidative stress and cellular proliferation. Therefore, we sought to determine if a low glucose environment (in vitro) or a ketogenic diet (in vivo) could inhibit anaplastic thyroid cancer tumor growth when combined with the antioxidant N-acetylcysteine. Methods: In vivo, nude mice were injected with the anaplastic thyroid cancer cell line 8505C (n ¼ 6/ group). Group 1 was fed a standard diet; Group 2 was fed a ketogenic diet; Group 3 was given standard diet with N-acetylcysteine (40 mM in the drinking water); and Group 4 was fed ketogenic diet with Nacetylcysteine. Tumor volumes, ketones, and glucose were measured. H&E stains and immunohistochemistry for Ki-67 and Caspase 3 were performed on the tumors. In vitro, 8505C cells were cultured in high glucose (25 mM), low glucose (3 mM), high glucose plus N-acetylcysteine (200 uM), or low glucose plus N-acetylcysteine for 96 hours. We performed CyQUANT proliferation (Thermo Fisher Scientific, Waltham, MA), Seahorse glycolytic stress (Agilent, Santa Clara, CA), and reactive oxidative stress assays. Results: Ketogenic diet plus N-acetylcysteine decreased in vivo tumor volume compared to standard diet (22.5 ± 12.4 mm3 vs 147 ± 54.4 mm3, P < .05) and standard diet plus N-acetylcysteine (P < .05). Blood ketone levels were significantly higher for the mice in the ketogenic diet group compared to standard diet (1.74 mmol/L vs 0.38 mmol/L at week 5, P < .001). However, blood glucose levels were not significantly different between ketogenic diet and standard diet groups. Cells cultured in low glucose plus Nacetylcysteine had significantly reduced proliferation compared to high glucose (98.1 ± 5.0 relative fluorescence units vs 157.8 ± 2.1 relative fluorescence units, P < .001). Addition of N-acetylcysteine to low glucose lowered glycolysis function compared to high glucose (39.0 ± 2.2 mpH/min/cell vs 89.1 ± 13.2 mpH/min/cell, P < .001) and high glucose plus N-acetylcysteine (37.4 ± 2.5 mpH/min/cell vs 70.3 ± 3.3 mpH/min/cell, P < .001). Low glucose plus N-acetylcysteine decreased reactive oxidative stress compared to high glucose (119 ± 34.7 relative fluorescence units vs 277 ± 16.0 relative fluorescence units, P ¼ .014). Conclusion: The combination of a ketogenic diet or glucose restriction with the antioxidantN-acetylcysteine significantly reduced tumor growth in vivo and in vitro. Further studies are warranted to explore these metabolic therapies in anaplastic thyroid cancer. © 2019 Elsevier Inc. All rights reserved.
Introduction
Presented at the American Association of Endocrine Surgeons, Los Angeles CA, April 7-9, 2019. * Reprint requests: Matthew Nehs, MD, Assistant Professor of Surgery, Harvard Medical School, Associate Surgeon, Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115. E-mail address: [email protected] (M.A. Nehs). https://doi.org/10.1016/j.surg.2019.06.042 0039-6060/© 2019 Elsevier Inc. All rights reserved.
Anaplastic thyroid cancer (ATC) is the most aggressive and fatal type of thyroid cancer, accounting for up to 40% of thyroid carcinoma deaths.1,2 Surgical resection is often not possible at presentation because of local invasion into the trachea, larynx, or because it surrounds the carotid artery. Current, nonoperative therapies (chemotherapy, targeted therapy, and radiation) become ineffective due to evolved tumor resistance.3,4 Despite multimodal
2
A. Aggarwal et al. / Surgery xxx (2019) 1e7
interventions, patients with ATC continue to have a median survival of only 5 to 6 months upon diagnosis.5 Thus, there is an urgent need to apply new treatment paradigms in patients with ATC. One such paradigm views the cancer cell in part as a disease of cellular metabolism. In many aggressive malignancies, cancer cells have a metabolic phenotype that is characterized by dysfunctional mitochondrial and a shift to glycolysis as the principal method of energy production (known as the Warburg effect).6 The increased glucose utilization by cancer cells is demonstrated clinically through positron emission tomography scans where patients are injected with radiolabeled fluorodeoxyglucose, which is readily absorbed by the tumor cells.7 Although glucose uptake and glycolysis is dramatically increased in many advanced malignancies, very little of the end product of glycolysis (pyruvate) is channelized into the tricarboxylic acid cycle in the mitochondria. Instead, pyruvate is reduced to form lactate, which also generates NADþ. This allows glycolysis to continue even in the presence of normal oxygen levels (normoxia).5 Since many cancer types exhibit glucose dependency by the Warburg effect,6,8 we hypothesized that glucose restriction in vitro and carbohydrate restriction in vivo through a ketogenic diet (KD) may slow cancer growth and progression. The KD is characterized by a high fat (~90%), moderate protein (~10%), and very low carbohydrate (<1%) macronutrient composition. This diet depletes hepatic and muscular glycogen and shifts the body’s metabolism to fatty acid oxidation.9 This suppresses insulin levels and blood glucose levels while increasing blood ketone levels (eg, acetoacetate and ß-hydroxybutyrate).9 This diet has been used in other preclinical models of advanced malignancy, including glioblastoma multiforme9; however, it has not yet been tested in a mouse model of ATC. In addition to the alteration of the Warburg phenotype seen in ATC, many advanced malignancies generate high levels of reactive oxygen species (ROS) that can drive cellular signaling.10,11 For example, ROS are involved in signaling pathways in proliferation, metabolism, angiogenesis, growth, and survival.11 Thus, decreasing ROS-oxidative stress levels in cancer cells by administering exogenous antioxidant may yield an anti-proliferative effect by diminishing the proliferative signaling mediated by ROS.12 Therefore in this study, we tested the hypothesis that a KD combined with an antioxidant (N-acetylcysteine[NAC]) could inhibit tumor growth in a mouse model of ATC. Materials and methods Cell line and culturing We used the ATC line 8505C (Sigma Aldrich, St. Louis, MO) for all the in vivo and in vitro experiments. This cell line was derived from a 78-year-old female patient. The cells were cultured using high glucose (HG) (25 mM) or low glucose (LG) (3 mM) in DMEM/F12 with L-Glutamine media that was completed with 10% fetal calf serum and 100 units/mL penicillin-streptomycin at 37 C and 5% CO2. All reagents used for culturing were purchased from Thermo Fisher Scientific. Both HG and LG cultures were treated with NAC (200 uM) (Thermo Fisher Scientific, Waltham, MA). In vivo mouse modeling Mice and tumor injections Nude mice (6e8-week-old females) were purchased from the Jackson Laboratory, Bar Harbor, ME. All mice protocols were followed in accordance with the National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee. Through the entire study, the mice were housed (3 mice/cage) in a biosafety level 2 facility in an institutional
vivarium. All mice (N ¼ 24) were acclimatized for a week and then injected with the tumor line 8505C into the flank. Weight, tumor volume, blood glucose, and ketones were measured each week. 8505C cells were seeded in 4-T75 flasks and cultured until just before confluency (4 days). Then, the cells were counted used Trypan Blue and 1 x 106 cells were suspended in 0.3 cc saline solution and injected subcutaneously in the right flank of each mouse, using a 1 cc tuberculin syringe. Mice were examined biweekly by a veterinarian for signs of distress and toxicity. Treatments and toxicity assessment The mice were randomly divided into 4 groups (6 mice/group). The tumors were allowed to grow until they were palpable before initiating each treatment. All treatments were initiated on the ninth day after injection. The mice in the KD groups were fasted overnight to initiate ketosis on the eighth day post-injection.13 Group 1 was fed the standard diet (SD); Group 2 was fed a KD; Group 3 was given SD with antioxidant NAC (40 mM) supplemented in the drinking water13; and Group 4 was given KD supplemented with NAC. Macronutrients in the SD (PicoLab Mouse Diet 20 [LabDiet, St. Louis, MO]) consisted of 55% carbohydrates, 22% fat, and 23% protein, whereas the KD had 89.9% fat,10% protein, and 0.1% carbohydrate. The KD was purchased from Research Diets Inc. (New Brunswick, NJ, USA). Tumor volumes on all 24 mice were measured using digital calipers and calculated as ½ length x width x height. Blood glucose and ketones were measured using glucose and ketone strips (Thermo Fisher Scientific, Waltham, MA) for the SD and KD treated mice.13 Mouse tumor immunohistochemistry (IHC) Tumor samples were procured from all the mice, once euthanized, for each of the 4 treatment groups. Slides were stained with H&E, Ki-67 for proliferation, and Caspase-3 to evaluate for apoptosis. All the slides were made and stained at the IHC Core Laboratory, Brigham and Women’s Hospital. In vitro assays Seahorse glycolytic stress assay The glycolysis function and capacity of the cells under different conditions were measured through extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) using Seahorse XF Glycolysis Stress Kit and Analyzer (Agilent, Santa Clara, CA). Changes in ECAR and OCR after the sequential injection of glucose (triggering glycolysis), oligomycin (inhibiting mitochondrial adenosine triphosphate synthase), and 2-deoxy-D-glucose (inhibiting glycolysis) represents the cellular glycolysis and its maximum capacity, respectively.14 Seahorse XF96 microplates (Agilent, Santa Clara, CA) were coated with 0.1mg/ml poly-D-lysine. 8505C cells were cultured with high or LG medium with or without NAC for 96 hours and seeded (4 x 104) in triplicates. XF assay medium was freshly prepared with XF base medium supplemented with 1 mM glutamine. Cells were washed twice using assay medium and incubated in a non-CO2 incubator for 1 hour. After calibration and initial measurements at baseline, a final concentration of 10 mM glucose, 1mM oligomycin, and 50 mM 2-deoxy-D-glucose were injected sequentially. Three measurements were taken after each injection. Cell number was counted using C-chip (Thermo Fisher Scientific, Waltham, MA) after the assay and was used to normalize the raw ECAR dataset. Intracellular ROS assay Intracellular ROS Oxidative Assay (Cell Biolabs, Inc, San Diego, CA) measures the antioxidant or ROS in a cell.15,16 The assay uses a cell-permeable fluorogenic probe 2’, 7’ -Dichlorodihydrofluorescein diacetate (DCFH-DA), which diffuses into the cell. DCFH-DA is then
A. Aggarwal et al. / Surgery xxx (2019) 1e7
3
Fig 1. Mouse tumor volume after 5 weeks. (A) The average tumor volume (mm3) for each group throughout the experiment period. (B) The endpoint (week 5) tumor volumes and each dot represents an individual mouse for that treatment group. N ¼ 6 for each group except for KD group n ¼ 5 (one died during week 2). *P < .05. (C) The blood glucose for each mouse for the 5-week period (P > .05). (D) The blood ketone for each mouse for the 5-week period (*P < .01). The error bar represents the SEM of each group. (E) Representative tumor images comparing a SD mouse to a KD þ NAC mouse.
deacetylated by cellular esterases to 2’, 7’-Dichlorodihydrofluorescein. When the Dichlorodihydrofluorescein molecule encounters ROS, it is rapidly oxidized to the fluorescent compound 2’, 7’ Dichlorofluorescein, and the fluorescence is measured using a plate-reader.17 8505C cells were seeded in a 96-welled plate (5,000 cells/well) overnight. Then, the cells were pretreated with DFCH-DA (1X), incubated at 37 C for 60 minutes, and washed with Hanks Balanced Salt Solution to remove excess DFCH-DA. After removing DFCH-DA, the cells were treated with HG, LG, HG þ NAC, or LG þ NAC for 96 hours. Fluorescence was measured using a SoftMax Pro plate reader (Molecular Devices Ltd, San Jose, California, USA) at 485 nM (excitation) and 538 nM (emission) wavelengths. Each condition was performed in triplicates. Proliferation assay We tested for cell proliferation using CyQUANT GR proliferation assay. This assay uses a fluorescent dye that binds to the cell’s nucleic acids.18 We performed the CyQUANT GR Proliferation assay in triplicates for each treatment condition as per the manufacturer’s guidelines. Cells were seeded in a 96-welled plate (5,000 cells/well) in triplicates and then exposed to HG, LG, HG þ NAC, or LG þ NAC for 24, 72, and 96 hours. Culture media was removed and 200 uL of the CyQUANT GR working solution was applied to the cells and incubated for 5 minutes at room temperature. Fluorescence was measured at 480 nM (excitation) and 520 nM (emission) with the SoftMax Pro software. Statistical analysis Statistical analysis was performed using Students t-tests (unpaired) on GraphPad Prism 6 (GraphPad Software, Inc, San Diego, CA). The Wave software was used for the Seahorse Glycolytic Stress Test. All independent treatment conditions were performed in triplicates and were computed as means and ± standard error of the mean (SEM).
Results In vivo - KD with NAC significantly slowed tumor growth in a mouse model Mice were fed SD, KD, SD þ NAC, or KD þ NAC and humanely sacrificed 5 weeks after treatment. Tumor volumes were markedly reduced in the group that was fed the KD þ NAC, compared to the SD group (21.5 ± 12.6 mm3 vs 147.2 ± 54.4 mm3, P < .05). KD þ NAC mice also had significantly smaller tumors compared to the SD þ NAC treatment group (21.5 ± 12.6 mm3 vs 321.9 ± 128.6 mm3, P < .05). Two of the 6 mice (33%) in the KD þ NAC group showed no viable tumor at all by the fifth week on final histopathology, as reviewed by our endocrine pathologist. The KD treatment group had a mean tumor volume of 67.3 ± 44.2 mm3. The SD þ NAC group showed the highest tumor volume by the fifth week (321.9 ± 128.6 mm3) but also had the largest variance in the data (Fig 1, A, B). Blood glucose and ketone measurements were taken each week for the SD and KD groups, and mean values calculated on the fifth week. Interestingly, the blood glucose levels between the groups did not show a significant difference between the KD and SD groups (Fig 1, C). However, the blood ketone levels were significantly higher for the mice in the KD group compared to SD (1.74 vs 0.38 mmol/L at week 5, P < .001) (Fig 1, D), thus confirming a state of ketosis. We found no significant differences in the weight of each mouse among the 4 groups of mice with various treatments during the 5 weeks of the study. No apparent signs of NAC toxicity were evident, such as cachexia due to weight loss, lethargy, urine discoloration or anuria, hypothermia, or general appearance (Grimace scale).19
IHC All our IHC slides were assessed by an endocrine pathologist. No significant differences were noted for H&E, Ki-67, and Caspase-3 stains between any of the treatment groups (Figs 2, AeF).
4
A. Aggarwal et al. / Surgery xxx (2019) 1e7
Fig 2. Immunohistochemical analysis for H&E, Ki67, and Caspase-3. (A, C, and E) The IHC staining for standard diet (SD) group. (B, D, and F) The IHC staining for ketogenic diet plus NAC (KD þ NAC) group. All pictures were taken under 20X magnification.
Interestingly, the small tumors in the KD þ NAC group showed large areas of central necrosis that did not stain for Caspase-3 (Fig 2). NAC decreases glycolytic stress We used the Seahorse Glycolysis Stress assay to measure the ECAR and OCR to determine 8505C glycolysis function and capacity for each group (HG, LG, HG þ NAC, and LG þ NAC) (Figs 3,
A, B). Glycolytic function was reduced significantly with the addition of NAC to HG (24.0 ± 2.4 mpH/min/cell) and NAC to LG (23.7 ± 1.8 mpH/min/cell) (P < .001) compared to HG alone (60.4 ± 1.7 mpH/min/cell) or LG alone (54.9 ± 11 mpH/min/cell). Similarly, the glycolysis capacity was significantly lowered with the addition of NAC to HG (37.4 ± 2.5 mpH/min/cell) and NAC to LG (39.0 ± 2.2 mpH/min/cell) compared to either the HG (70.3 ± 3.3 mpH/min/cell) or LG (89.1 ± 13.2 mpH/min/cell)
Fig 3. Seahorse glycolytic stress assay. (A) The extracellular acidification rate (ECAR) and (B) The oxygen consumption rate (OCR) of 8505C cells treated with high glucose (HG, (25mM), low glucose (LG) (3mM), high glucose plus 200mM NAC (HG þ NAC), low glucose plus 200mM NAC (LG þ NAC) in triplicates. All cells were treated for 96 hours. (A and B) The overall ECAR and OCR change throughout the assay, respectively. (C) The average glycolysis rate and glycolytic capacity for each group. The error bar represents the SEM for each group. **P < .001.
A. Aggarwal et al. / Surgery xxx (2019) 1e7
5
LG þ NAC compared to LG treatment alone at 96 hours of exposure (98.1 ± 5.0 RFU vs 146.9 ± 6.9 RFU, P < .01) (Fig 5, B). Discussion
Fig 4. ROS assay. Fig 4 shows the relative fluorescent units (RFUs) for 8505C cells treated with high glucose (HG) (25mM), low glucose (LG) (3mM), high glucose plus 200mM NAC (HG þ NAC), and low glucose plus 200mM NAC (LG þ NAC) in triplicates. Treatment duration was 96 hours. The error bar represents SEM for each group. *P < .05.
groups (P < .001) (Fig, 3, C). LG treatments were not significantly different from HG. Low glucose and NAC reduce ROS-oxidative stress Examining the effect of NAC in cellular oxidative stress, we found that only LG with NAC treatment significantly reduced ROSoxidative stress levels when compared to the HG alone (119 ± 34.7 relative fluorescence units [RFU] vs 277.4 ± 16.0 RFU, P < .02). ROS levels were found to be intermediate values for the LG (226 ± 44 RFU) and HG þ NAC (195 ± 60 RFU) treatment groups (Fig 4). Low glucose and NAC attenuates cellular proliferation We used HG and LG media to mimic the effect of SD and KD in vitro and added 200 mM NAC to the media to compare the proliferation rates after 24-, 72-, and 96-hour treatments of 8505C cells. A significant reduction in proliferation was seen in cells treated with LG þ NAC when compared to HG at 72 hours (94.5 ± 7.6 RFU vs 134.4 ± 9.9 RFU, P < .04) (Fig 5, A), which was further heightened at 96 hours of exposure (98.1 ± 5.1 RFU vs 157.8 ± 2.1 RFU, P < .001) (Fig 5, B). Proliferation reduction was also noted with
Many advanced malignancies are characterized by a reprogramming of energy metabolism leading to increased glucose consumption.20 Cancer cells that exhibit this phenotype have a very high rate of glycolysis that ends in the production of lactate, even in the presence of sufficient oxygen (aerobic glycolysis). This phenomenon is referred to as the Warburg effect, named after Otto Warburg, who first described this in 1925.21 This process involves a high rate of glycolysis followed by the conversion of pyruvate to lactate, which regenerates NADþ and allows glycolysis to continue. Additionally, this increase in glycolysis drives the production of ROS, which have been shown to play an important role in cellular signaling pathways for proliferation, metabolism, angiogenesis, growth, and survival in several advanced malignancies including melanoma22 and pancreatic adenocarcinoma.23 Additionally, the antioxidant NAC has been shown to reduce ROS levels thereby curtailing oxidative stress in cancer cells and limiting their rate of cellular proliferation.12 Given that ATC is one of the fastest growing and most lethal malignancies known, we sought to determine the effect of NAC and glucose restriction on growth of the ATC cell line 8505C. We chose to apply this metabolic targeting to ATC since it has a very high rate of proliferation and most current therapies remain ineffective.1e4 In order to simulate a LG environment in vivo, we utilized the ketogenic diet, which is characterized by a high fat (~90%), moderate protein (~10%), and very low carbohydrate (<0.1%) macronutrient composition. The KD pushes the body’s metabolism from using glucose to fat and ketones as energy sources, and it has previously been shown to slow the progression of several types of malignancies including glioblastoma multiforme.9 Our study found that ATC tumor growth was dramatically inhibited in mice that were fed the KD supplemented with antioxidant NAC in the drinking water. Additionally, we found no histologic evidence of viable tumor in 2 mice in the KD þ NAC group. Interestingly, we found the largest average tumor volumes in the SD þ NAC group, although the variance in the data was large and may reflect random variation in a small sample size. As expected, the mice in the KD group had higher average ketone levels than the mice fed the standard chow (SD), but interestingly, there were no significant differences in blood glucose concentrations between groups. This suggests that the KD may exert an anticancer effect that is independent of blood glucose levels, which has been previously demonstrated.24 We did not observe any IHC differences in
Fig 5. Proliferation assay results of 8505C after 96-hour treatment. (A) The relative fluorescent units (RFUs) for 8505C cells treated with high glucose (HG) (25mM), low glucose (LG) (3mM), high glucose plus 200mM NAC (HG þ NAC), and low glucose plus 200mM NAC (LG þ NAC) in triplicates for 24, 72, and 96 hours. (B) The 96 hours bar graph of the relative fluorescent units (RFUs) for each group. The error bar represents SEM for each group. **P < .01, ***P < .001.
6
A. Aggarwal et al. / Surgery xxx (2019) 1e7
the Caspase-3 and Ki-67 activity between the various treatment groups used in the study. However, the KD þ NAC IHC revealed large areas of central necrosis in very small tumors, which suggests that the cells were dying due to non-apoptotic mechanisms. Our in vitro data support the in vivo findings: the cells cultured in LG supplemented with NAC showed a significantly lower cellular proliferation, reactive oxygen species-mediated oxidative stress, and glycolytic stress when compared to the cells cultured in HG media. Mechanistically, our study showed that restricting glucose in combination with antioxidant NAC results in decreased glycolytic stress and inhibition of ROS signaling, thereby reducing tumor cell proliferation. Our study has several important limitations. First, we used a single cell line (8505C) as the basis for our animal experiment as a proof of principle experiment for the study design. Further studies should expand the cell lines used and vary the mutational profile of the tumors to look for consistent or disparate effects. The heterotopic flank tumor injection is also a limitation of the study compared to an orthotopic model, which has been previously described.25 Future studies will assess the role of the KD on an orthotopic mouse model with lung metastases. Despite these limitations, we believe that this study is the first to demonstrate an effect of the KD/metabolic therapy in a mouse model of ATC. In conclusion, our study demonstrates that a KD combined with the antioxidant NAC decreased tumor growth in a mouse model of ATC. Further studies are needed to determine if targeting cancer cell glucose dependency and ROS-mediated cellular proliferation is an effective adjuvant treatment in patients with ATC. Funding/Support This study was funded by the Robert T. Osteen Fellowship Award and the Department of Surgery, Brigham and Women's Hospital, Boston, MA. Conflict of interest/Disclosure Authors Abha Aggarwal, Zuliang Yuan, Justine A. Barletta, and Matthew A. Nehs do not have any conflicts of interest. Dr. Jochen H. Lorch has research that is supported by Novartis, Bayer, BMG, and Millennium. Dr. Lorch also receives consulting honoraria from Bayer and Genentech. References 1. Hundahl SA, Fleming ID, Fremgen AM, Menck HR. A national cancer data base report on 53,856 cases of thyroid carcinoma treated in the U.S., 1985e1995. Cancer. 1998;83:2638e2648. 2. Chiacchio S, Lorenzoni A, Boni G, Rubello D, Elisei R, Marianni G. Anaplastic thyroid cancer: Prevalence, diagnosis and treatment. Minerva Endocrinol. 2008;33:341e357.
3. Pinchot SN, Sippel RS, Chen H. Multi-targeted approach in the treatment of thyroid cancer. Ther Clin Risk Manag. 2008;4:935e947. 4. Wagle N, Grabiner BC, Van Allen EM, et al. Response and acquired resistance to everolimus in anaplastic thyroid cancer. N Engl J Med. 2014;371:1426e1433. 5. Smallridge RC, Copland JA. Anaplastic thyroid carcinoma: pathogenesis and emerging therapies. Clin Oncol. 2010;22:486e497. 6. Lu Z, Guo Y, Zhang X, et al. ORY-1001 suppresses cell growth and induces apoptosis in lung cancer through triggering HK2 mediated Warburg effect. Front Pharmacol. 2018;9:1411. 7. Chang HT, Hu C, Chiu YL, Peng NJ, Liu RS. Role of 2-[18F] fluoro-2-deoxy-Dglucose-positron emission tomography/computed tomography in the posttherapy surveillance of breast cancer. PLoS One. 2014;9:e115127. 8. Coelho RG, Fortunato RS, Carvalho DP. Metabolic reprogramming in thyroid carcinoma. Front Oncol. 2018;8:82. 9. Poff AM, Ward N, Seyfried TN, Arnold P, D’Agostino PD. Non-toxic metabolic management of metastatic cancer in VM mice: Novel combination of ketogenic diet, ketone supplementation, and hyperbaric oxygen therapy. PLoS One. 2015;10:e0127407. 10. Stafford P, Abdelwahab MG, Kim Y, Preul MC, Rho JM, Scheck AC. The ketogenic diet reverses gene expression patterns and reduces reactive oxygen species levels when used as an adjuvant therapy for glioma. Nutri Metabol (Lond). 2010;7:74. 11. Schieber M, Chandel NS. ROS Function in redox signaling and oxidative stress. Curr Biol. 2014;24:R453eR462. 12. Monti D, Sotgia F, Whitaker-Menezes D, et al. Pilot study demonstrating metabolic and anti-proliferative effects of in vivo anti-oxidant supplementation with N-Acetylcysteine in breast cancer. Semin Oncol. 2017;44: 226e232. 13. Gao P, Zhang H, Dinavahi R, et al. HIF-dependent antitumorigenic effect of antioxidants in vivo. Cancer Cell. 2007;12:230e238. 14. Wu M, Neilson A, Swift AL, et al. Multiparameter metabolic analysis reveals a close link between attenuated mitochondrial bioenergetic function and enhanced glycolysis dependency in human tumor cells. Am J Physiol Cell Physiol. 2007;292:C125eC136. 15. Kim J, Chen C, Yang J, Mochly-Rosen D. Aldehyde dehydrogenase 2*2 knock-in mice show increased reactive oxygen species production in response to cisplatin treatment. J Biomed Sci. 2017;24:33e40. 16. Stebbins KJ, Broadhead AR, Cabrera G, et al. In vitro and in vivo pharmacology of NXT629, a novel and selective PPARa antagonist. Eur J Pharmacol. 2017;809: 130e140. 17. Azimi I, Peterson RM, Thompson EW, Roberts-Thomson SJ, Monteith GR. Hypoxia-induced reactive oxygen species mediate N-cadherin and SERPINE1 expression, EGFR signaling and motility in MDA-MB-468 breast cancer cells. Sci Rep. 2017;7:15140. 18. Jones LJ, Gary M, Yue ST, Haugland P, Singer VL. Sensitive determination of cell number using the CyQUANT cell proliferation assay. J Immunol Methods. 2001;254:85e98. 19. Burkholder T, Foltz C, Karlsson E, Linton CG, Smith JM. Health evaluation of experimental lab mice. Curr Protoc of Mouse Biol. 2012;2:145e165. 20. Vander Heiden MG, DeBerardinis RJ. Understanding the intersections between metabolism and cancer biology. Cell. 2017;168:657e669. 21. Warburg O. The metabolism of carcinoma cells. The Journal of Cancer Research. 1925;9:148e163. 22. Bisevac JP, Djukic M, Stanojevic I, et al. Association between oxidative stress and melanoma progression. J Med Biochem. 2018;37:12e20. 23. Du J, Liu J, Smith BJ, Tsao MS, Cullen JJ. Role of Rac1-dependent NADPH oxidase in the growth of pancreatic cancer. Cancer Gene Ther. 2011;18: 135e143. 24. Fine EJ, Segal-Isaacson CJ, Feinman RD, et al. Targeting insulin inhibition as a metabolic therapy in advanced cancer: A pilot safety and feasibility dietary trial in 10 patients. Nutrition. 2012;28:1028e1035. 25. Nehs MA, Nucera C, Nagarkatti SS, et al. Late intervention with antiBRAF(V600E) therapy induces tumor regression in an orthotopic mouse model of human anaplastic thyroid cancer. Endocrinology. 2012;153: 985e994.
Discussion Dr Eren Berber (Cleveland, OH): I think that as surgeons dealing with cancer outside the thyroid as well, we are seeing that patients are putting themselves on these ketogenic diets even before they come to see us. In addition to the ketogenic diet, we are also seeing that some patients are actually taking Metformin as well, especially for differentiated thyroid cancer. We had a really interesting patient
who was in her 30s and had a 1.5-centimeter papillary thyroid cancer. She refused all surgical therapy and put herself on a ketogenic diet and Metformin. When we saw the patient two years later, the cancer had not advanced on ultrasound and had actually decreased in size to 8 millimeters. I wonder if you have had any experience with Metformin in this context.
A. Aggarwal et al. / Surgery xxx (2019) 1e7
Dr Abha Aggarwal: Thank you for the question, but we have not tested Metformin in any of our experiments. Dr Mahsa Javid (Charleston, SC): Very impressive study. Thank you. I have some related questions. What is your hypothesis about the non-apoptotic causes of reduction in tumor volume? Why do you think there was a difference between the in vitro and in vivo cellular proliferation rate? And did you check for invasion and migration? Dr Abha Aggarwal: It was interesting to us as well. One thing that we can say is that we don't understand the mechanism of how the ketogenic diet works. These were very surprising results for us as well because we thought there would be some kind of significant difference in proliferation since we were seeing that in vitro, but we didn't observe that. We don't know the mechanism of it, basically, but those are great questions.
7
Dr Mark Cohen (Ann Arbor, MI): Two quick questions. Fat storage in mice is a little different than fat storage in humans, and a ketogenic diet has a slightly different effect. When you showed your in vivo data, actually the ketogenic diet by itself had a significant impact on tumor growth compared to the standard diet. Do you think that a lot of this effect is just from the diet, or is the N-acetylcysteine really adding a lot? Dr Abha Aggarwal: I think the most significant effects we were seeing were in combination of N-acetylcysteine and the low glucose or the ketogenic diet. We did not see significant effects just with the ketogenic diet alone. So it was just more of an interaction. We don't know what the ketogenic diet is doing as far as mechanism goes. But we hypothesized that it is increasing the ketosis, so it is promoting fat metabolism, and we know that it has 85% to 90% fat in the diet already, and somehow that's interacting with the N-acetylcysteine (or NAC). But we don't know the exact mechanism of it. That's still a future study for us to do.