Abstract
Background/Aim: The ketogenic diet has recently gained interest as potential adjuvant therapy for cancer. Many researchers have endeavored to support this claim in vitro. One common model utilizes treatment with exogenous acetoacetate in lithium salt form (LiAcAc). We aimed to determine whether the effects of treatment with LiAcAc on cell viability, as reported in the literature, accurately reflect the influence of acetoacetate. Materials and Methods: Breast cancer and normal cell lines were treated with acetoacetate, in lithium and sodium salt forms, and cell viability was assessed. Results: The effect of LiAcAc on cells was mediated by Li ions. Our results showed that the cytotoxic effects of LiAcAc treatment were significantly similar to those caused by LiCl, and also treatment with NaAcAc did not cause any significant cytotoxic effect. Conclusion: Treatment of cells with LiAcAc is not a convincing in vitro model for studying ketogenic diet. These findings are highly important for interpreting previously published results, and for designing new experiments to study the ketogenic diet in vitro.
The ketogenic diet (KD), characterized by high fat, low carbohydrate, and adequate protein consumption, was originally identified for its ability to mimic starvation-induced ketosis, and thus to benefit patients with epilepsy (1). Recently, the KD has been the subject of research seeking to assess its potential as adjuvant treatment for cancer (2). Given the hallmark of the altered metabolism of cancer cells, known as the Warburg effect, scientists hope to hinder tumor growth or increase sensitivity to common therapies by restricting access of tumor cells to nutrients, specifically glucose. In effect, the physiological consequences of KD include both reduction and stabilization of baseline blood glucose and insulin, as well as ketone body production, both thought to contribute to its anti-cancer activity (3). Ketone body production, or ketogenesis, by liver tissues compensates for low glucose intake and provides an alternative circulating energy source.
In vivo experiments show promising results, with elevated ketone body levels correlating with hindered tumor growth (4). Information on the effects of KD on patients and on the efficacy of cancer treatment is just now emerging from preliminary clinical trials. Initial data from these trials indicate a selective positive effect (5-7). Many of the trials focus on breast cancer given the previously described links between nutrition and breast cancer (8, 9).
In order to study the mechanism by which KD may affect cancer cells, many groups have designed experiments to replicate these effects in vitro. The primary component of these in vitro systems is the addition of ketone bodies, which include β-hydroxybutyrate, acetoacetate, and acetone, to the culture medium. The most commonly used treatments include acetoacetate in lithium salt form (LiAcAc) and β-hydroxybutyrate in sodium salt form (NaBHB). Among the experiments utilizing LiAcAc, in vitro results include reduced growth and ATP production of cancer cells (10, 11), induction of oxidative stress (12), and altered metabolic profile (11, 13).
Recently, Vidali, et al. (14) published research demonstrating that the effects of LiAcAc on cell growth consistently correlate with the effects of lithium, and are likely not due to the acetoacetate component of the treatment. These results are proven by comparing the effects of LiAcAc, LiCl, and NaCl on several cancer and normal cell lines, including HEK293 (non-tumor human embryonic kidney), HDFn (normal human dermal fibroblasts), SK-N-BE(2) and SH-SY5Y (neuroblastoma), and 786-0 and CAKI-2 (renal cell carcinoma). There, cell growth was measured by crystal violet assay, and immunofluorescent staining of cleaved caspase 3 was used to assess differences in cell death.
Here, we present our data, which include a panel of breast cancer and non-cancer cell lines, from readouts, which include crystal violet and resazurin cell viability assays, as well as ATP assay. Furthermore, our results show the independent effects of acetoacetate in sodium salt form (NaAcAc) in vitro. Our work, together with the recently published work by Vidali, et al. (14), demonstrate an urgent need to develop new models to study KD in vitro and to refine the interpretation of previously published results.
Materials and Methods
Cell culture. MCF7, MDA-MB-231, Hs578T cell lines were purchased from ATCC (Manassas, VA, USA). HSF064 were a kind gift from Prof. Richard Feinman, SUNY Downstate, USA. HB2 cells were kindly given to us by Prof. Ido Wolf (Tel Aviv Sourasky Medical Center, Tel Aviv, Israel).
All cells were cultured in Dulbecco's modified Eagle medium (DMEM, Gibco) containing 10% fetal calf serum (FCS, Biological Industries), 2 mM L-glutamine, 100 U/ml penicillin and 0.1 mg/ml streptomycin (all Biological Industries) at 37°C in 5% CO2. Cells were passed every 3-4 days using trypsin-EDTA (Biological Industries).
For experiments, cells were seeded in 96-well plates in DMEM containing 5.5 mM glucose. After 24 h, cells were treated with 3 mM, 10 mM, or 30 mM lithium acetoacetate (LiAcAc, Sigma), lithium chloride (LiCl, Sigma), sodium chloride (NaCl, Sigma), or NaAcAc (synthesized in-house). After 48 h the media was removed and fresh media including treatment was added gently, and after an additional 48 h, readout assays were performed as follows.
Crystal violet assay. Cells were washed once in 1× PBS. Next, 50 μl of 2% crystal violet solution prepared in absolute ethanol (Mercury, Rosh Ha'ayin, Israel) was added to each well, and the cells were incubated with shaking at room temperature for 20 min. Next, crystal violet solution was removed, the plate was washed three times with ddH2O and air-dried overnight. Then, 100 μl of 1% SDS solution was added per well and incubated with shaking for 1-2 h at room temperature until the crystal violet stain was completely dissolved. The absorbance at 570 nm was determined using the ClarioStar plate reader (BMG Labtech, Ortenberg, Germany).
Luminescent ATP detection assay kit. Luminescent ATP Detection Assay Kit (Abcam, Cambridge, UK) was used according to the manufacturer's instructions. Briefly, media was removed from each well, and the cells were lysed with the detergent solution. Substrate solution was added to each sample, incubated with shaking for 10 min at room temperature, and luminescence was measured by the ClarioStar plate reader.
Resazurin-based cytotoxicity assay. Resazurin sodium salt (Sigma, St. Louis, MO, USA) was prepared as a 6× (2.5 mM) stock in PBS. At the time of assay, stock solution was added to the cell media to obtain a 1× concentration. Cells were incubated at 37°C for 4 h and fluorescence (excitation 560 nm, emission 590 nm) was measured by the ClarioStar plate reader.
Statistical analysis. For all assays, the average signal from a minimum of 8 technical replicates in each of 3 biological replicates was calculated and normalized to untreated values. Statistical analysis of the data was performed using Python NumPy library and t-test was used for data comparison. A p-value of <0.05 was considered statistically significant.
Synthesis of NaAcAc. The amount of 25.5 ml (50 mmol) of 2 N sodium hydroxide was added to 6.5 ml (50 mmol) of redistilled ethyl acetoacetate and the volume reached 50 ml with the addition of distilled water. The mixture was incubated at 40°C for 1 h and then neutralized to pH 7 with 1 N hydrochloric acid. Hydrolysis was about 98% completed on the basis of titration. Distilled water was added to reach a volume of 100 ml, and the solution was then distilled by a vacuum on a rotary evaporator at 40°C until the volume reached 25 ml, in order to remove ethanol formed during the hydrolysis. Finally, the reaction mixture was lyophilized (15, 16).
Results
In our efforts to explore the effects of ketogenic diet conditions on breast cancer, we sought to calibrate an in vitro system to mimic KD conditions. We utilized a variety of ketone body treatments to optimize the effects observed on human breast cell lines. Here, we present data on the effects of acetoacetate, both in lithium and sodium salt forms, on breast cancer and non-cancer cell lines.
LiCl and LiAcAc treatments inhibit cell growth in a similar manner in breast cancer cells (MCF7, MDA-MB-231, Hs578T), non-cancer breast cells (HB2), and normal fibroblasts (HSF064). Previous studies report that LiAcAc inhibits cell growth in breast cancer cell lines in vitro (10), with a selective effect on cancer versus normal cell lines. These studies, performed in the context of research on the ketogenic diet, assume that the results are mediated by AcAc. In order to assess the contribution of lithium to the effect of LiAcAc on cells, we cultured a panel of commonly used breast cancer cell lines (MCF7, MDA-MB-231, Hs578T), as well a non-cancerous breast cell line (HB2) and a normal fibroblast line (HSF064) in medium containing low levels of glucose, comparable to physiological levels, and treated them with either LiAcAc or LiCl for four days. Previous studies utilizing these compounds included doses ranging from 3-10 mM, with 3 mM being the closest approximation of physiological blood ketone levels during ketosis (17). Cell viability was quantified using the crystal violet assay, with resulting values of treated cultures presented as a percentage of untreated cultures. All cell lines showed reduced viability under conditions of equimolar (10 mM) treatment of both LiAcAc and LiCl, with no significant differences between the two treatments (Figure 1). Thus, we concluded that cell growth inhibition observed under LiAcAc treatment conditions can be attributed to the presence of lithium ions, and that the treatment affects both cancerous and non-cancer cell lines and is not specific to cancer cells.
LiCl and LiAcAc affect cell viability in a similar manner. Five cell lines, including three breast cancer lines (MCF7, MDA-MB-231, Hs578T), one non-cancer line (HB2), and one normal fibroblast line (HSF064) were treated with either 10 mM LiCl or 10 mM LiAcAc. Control samples were untreated. Cell viability was measured after 4 days by crystal violet assay. Both LiCl and LiAcAc treatment significantly reduced cell viability in all cell lines (stars indicate significant difference as compared to control), but with no significant difference between treatments with regards to the degree of effect. Values represent an average of at least 3 independent biological experiments with at least 8 technical repeats within each experiment. Data are presented as the mean value, with 95% confidence interval. Acetoacetate treatment groups were compared to control by t-test (p<0.05).
LiCl and LiAcAc treatments lead to a comparable decrease in ATP levels in MCF7 and HB2. To further support these observations, we assayed cultures treated for four days with LiAcAc or LiCl for ATP production. Here, too, treatment with equimolar levels of LiCl led to significantly lower ATP levels, comparable to those seen in LiAcAc treated cultures. Figure 2 shows results from one representative cancer (MCF7) and one non-cancer cell line (HB2). These results correlate with the effects seen with the crystal violet assay.
LiCl and LiAcAc comparably affect ATP levels. Cells were treated with either 10 mM LiCl or 10 mM LiAcAc. Control samples were untreated. ATP levels were measured after 4 days. Both LiCl and LiAcAc treatment significantly reduced ATP levels in representative non-cancer (HB2) and cancer (MCF7) cells (stars indicate significant difference as compared to control), but with no significant difference between treatments with regards to the degree of effect. Values are an average of at least 3 independent biological experiments with at least 8 technical repeats within each experiment. Data are presented as the mean value, with 95% confidence interval. Acetoacetate treatment groups were compared to control by t-test (p<0.05).
LiAcAc treatment, but not treatment with NaAcAc, causes cytotoxicity in breast cancer and non-cancer cell lines in a concentration-dependent manner. Finally, in order to assess the independent effects of acetoacetate on cell viability, we synthesized NaAcAc and treated cells with varying concentrations. The resazurin cell viability assay, namely, was performed on cultures treated with NaAcAc or LiAcAc at concentrations of 3 mM, 10 mM, or 30 mM for four days. The viability of treated cultures, expressed as a percentage of untreated controls, is presented in Figure 3. Treatment with LiCl led to a dose-dependent reduction in cell viability, while the viability of NaAcAc treated cultures did not differ significantly from untreated cultures and remained significantly different from LiAcAc-treated cultures.
Taken together, these results convincingly indicate that the effects on cell viability which have previously been described under conditions of acetoacetate (LiAcAc) treatment, can be attributed to the Li ions present, and are independent of the AcAc component. These treatments are non-specific, affecting both cancer and non-cancer cell lines. Furthermore, this work is consistent with the work of Vidali et al. (14), and indicates that the phenomenon previously described on neuroblastoma, renal cell carcinoma, and human embryonic kidney cells is maintained in breast cancer cell lines.
Discussion
Breast cancer is among the most common cancers in women and accounts for 30% of all new female cancer diagnoses (18, 19). Currently, treatment strategies of breast cancer include surgical removal of the tumor, chemotherapy, radiation, and hormone therapy in different combinations (20). There is a growing interest in adjuvant therapies, such as diets and supplements, to improve patient outcome. In light of this, and given the known shift in the utilization of glucose by cancer cells, the ketogenic diet has been identified as a potentially relevant additive anti-cancer treatment. To this end, many groups have explored the effects of ketogenic diet on cancer in vitro, in vivo, and in clinical trials (9, 21, 22).
As part of our own efforts to study the effects of ketogenic diet conditions on breast cancer and assess its potential as adjuvant cancer therapy, we aimed to develop an in vitro culture model. Past work in this field has utilized the treatment of cultures with exogenous ketone bodies, acetoacetate (AcAc) and β-hydroxybutyrate (BHB), to this end (12, 23). In particular, several studies utilize the lithium salt of AcAc and demonstrate effects on cell viability among others [reviewed by Vidali, et al. (14)]. Here, we present our work which includes breast cancer and non-cancer cells lines, demonstrating the comparable effects of LiCl and LiAcAc on cell viability and ATP levels, indicating that the observed effects are mediated by the Li ion in a non-specific manner. Furthermore, our work convincingly demonstrates that acetoacetate treatment alone has no significant effect on cell viability, as demonstrated by the treatment of cells with NaAcAc. These findings are consistent with, and go beyond, the recently published work of Vidali et al. (14), which presented data on the effects of LiAcAc on neuroblastoma, renal cell carcinoma, and human embryonic kidney cells.
NaAcAc has no significant effect on cell viability. Cells were treated with either 3 mM, 10 mM, or 30 mM LiAcAc or NaAcAc. Control samples were untreated. Cell viability was measured after 4 days by resazurin assay. LiCl treatment significantly reduced cell viability in a dose-dependent manner (stars indicate significant difference as compared to control), but NaAcAc treatment had no significant effect. Values are an average of at least 3 independent biological experiments with at least 8 technical repeats within each experiment. Data are presented as the mean value, with 95% confidence interval. Treatment groups were compared and tested for significance using the one-way ANOVA test with 2 planned comparisons. Multiplicity adjustment by the Bonferroni method was applied to the p-values (p<0.05).
The focus of this work is to advance the state of cell culture work, which can be used to study the effects of ketogenic diet conditions on cancer. Although some questions arise regarding the mechanism by which Li affects the viability of cells, they are beyond the scope of this study. A well-crafted review of the potential mechanism can be found in the discussion by Vidali, et al. (14).
In order to proceed with the research on the ketogenic diet in vitro, we take into account the following considerations. Under ketosis, the liver produces ketone bodies, acetoacetate and β-hydroxybutyrate, which circulate throughout the body. Initially, we focused our model on treatment with acetoacetate due to the more dramatic effects seen in the literature. Acetoacetate is known to be an unstable molecule, which rapidly undergoes decarboxylation when heated. Thus, the half-life of acetoacetate is mere hours at the standard 37° incubation temperature (24). However, the half-life of the anionic form can be as long as 4 days and thus, we, along with other researchers, chose to work with AcAc in lithium salt form which imparts significant stability. Our current work and that of Vidali, et al. (14) have shown that the effect of LiAcAc is due to Li ions. Thus, future research should reconsider focusing on BHB to study the effects of KD.
In the clinical setting, BHB level in blood and urine is generally used as a measure of ketosis (25), and some even consider it “the principal ketone body” (23). There are a variety of seemingly conflicting results reported in experiments which explore the effect of BHB on cancer cells in vitro and on experimental tumors in mice in vivo. In some cases, an anti-proliferative effect of BHB was shown for multiple cancer cells, such as glioblastoma and tumor stem cells (26), melanoma, cervical carcinoma, or neuroblastoma (27-29). While in other cases, BHB was identified as the factor inducing enhanced proliferation (30). Several hypotheses have been suggested for this “β-hydroxybutyrate paradox” relating to both the metabolic and signaling properties of BHB (23, 31). While Bartmann et al. report that BHB does not influence proliferation of breast cancer cells or their response to chemotherapy or radiation (23), further experimental scenarios may be suggested to test the effects of NaAcA and BHB on the metabolism and aggressiveness of cancer cells in vitro.
Our aim here is not to undermine the potential of a ketogenic diet as a dietary intervention for cancer patients but rather to highlight the limitations of using synthetic ketone bodies in cell culture to model the complex effects of KD on tumors in vivo.
Acknowledgements
The Authors would like to thank Sarah Winkler, Dr. Shiri Barak, and Nili Turian for their dedicated technical assistance with experiments, as well as Dr. Yael Maizels and Dr. Tamar Rubinek for their valuable consultation and advice.
Footnotes
↵* These Authors contributed equally to this work.
Authors' Contributions
Conceptualization, RCH and SH; methodology, RCH and SH; validation, HL, OS, and KA; formal analysis, VK; investigation, HL, OS, and KA; resources, TB; data curation, VK; writing—original draft preparation, RCH; writing—review and editing, SH; visualization, RCH, SH, and VK; supervision, IK; project administration, RCH and SH; funding acquisition, IK.
Conflicts of Interest
The Authors declare no conflicts of interest in relation to this study.
- Received May 21, 2020.
- Revision received June 4, 2020.
- Accepted June 10, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved








