Background
Malignant brain cancer persists as a major disease of morbidity and mortality in adults and is the second leading cause of cancer death in children [
1‐
4]. Many current therapies for malignant brain tumors are ineffective in providing long-term management because they focus on the defects of the tumor cells at the expense of the health and vitality of normal brain cells [
5‐
7]. We previously showed that caloric restriction (CR) is anti-angiogenic, anti-inflammatory, and pro-apoptotic in the experimental mouse (CT-2A astrocytoma) and the human (U87-MG malignant glioma) brain tumors [
8‐
11]. CR targets tumor cells by reducing circulating glucose levels and glycolysis, which tumor cells need for survival, and by elevating ketone bodies, which provide normal brain cells with an alternative fuel to glucose [
5,
9,
11]. We previously used linear regression analysis to show that blood glucose levels could predict CT-2A growth as well as insulin-like growth factor 1 (IGF-1) levels, which influences tumor angiogenesis [
8,
9]. In contrast to glucose, ketone bodies (β-hydroxybutyrate and acetoacetate) bypass glycolysis and directly enter the mitochondria for oxidation [
12,
13]. By bypassing glycolysis, ketone bodies are also effective for treatment of inherited defects in glucose transporters and pyruvate dehydrogenase, which connects glycolysis with respiration [
14‐
16]. Ketone bodies are more energetically efficient than either pyruvate or fatty acids because they have a greater hydrogen/carbon ratio (more reduced) than pyruvate and, unlike fatty acids, do not uncouple mitochondria [
5,
17]. The transition from glucose to ketone bodies for brain energy metabolism is best under the natural conditions of CR [
5,
8,
18].
The metabolism of β-hydroxybutyrate (β-OHB), the major circulating ketone body, for energy depends on the expression of two key mitochondrial enzymes: β-hydroxybutyrate dehydrogenase (β-OHBDH), and succinyl-CoA: 3-ketoacid CoA transferase (SCOT) [
13,
19‐
21]. These enzymes become critical when neurons and glia transition to ketone bodies in order to maintain energy balance under conditions of reduced glucose availability. In addition to serving as a more efficient metabolic fuel than glucose, ketone bodies also possess anti-inflammatory potential through reduction of reactive oxygen species and increase of glutathione peroxidase activity [
5,
17,
22]. Brain tumors, like most malignant tumors, are largely dependent on glucose and glycolysis for their growth and survival due to abnormalities in the number and function of their mitochondria [
5,
8,
23‐
27]. The transition from glucose to ketone bodies as the primary energy source of the brain under calorically restricted conditions exploits the metabolic deficiencies of brain tumor cells while enhancing the health and vitality of normal neurons and glia according to principles of evolutionary biology and metabolic control theory [
5,
28].
Nebeling and co-workers previously found that a high fat/low carbohydrate ketogenic diet (KD), consisting of medium chain triglycerides, provided long-term management of pediatric astrocytoma while enhancing the nutritional status of the patients [
29]. The findings in human pediatric astrocytoma were confirmed in the experimental mouse CT-2A astrocytoma using a lard-based rodent ketogenic diet [
8,
9]. CR and some KDs, however, are not standardized diets and may be difficult to implement in the clinic due to issues of compliance. For example, CR in mice mimics therapeutic fasting in humans, involving water-only dieting, whereas medium chain triglyceride or lard-based ketogenic diets can cause gastrointestinal and kidney problems in both children and adults [
30‐
33]. Our goal was to develop a more effective alternative diet therapy for brain cancer that could extend survival without compromising the health and vitality of normal cells.
In this study we evaluated the therapeutic efficacy of KetoCal
®, a new nutritionally balanced soybean oil ketogenic diet that was formulated specifically for managing refractory epilepsy in children [
34]. No prior studies have evaluated the therapeutic efficacy of KetoCal
® for brain cancer management. Here we show that KetoCal
®, given in calorically restricted amounts significantly reduced circulating plasma glucose levels while significantly elevating ketone body levels in mice bearing orthotopic CT-2A and U87-MG brain tumors. Moreover, the restricted KetoCal
® diet reduced brain tumor growth and microvessel density, while extending mouse survival. Gene expression for β-OHBDH and SCOT was lower in the tumors than in contralateral normal brain suggesting that the brain tumors have reduced ability to metabolize ketone bodies for energy. This preclinical study indicates that KetoCal
® is a safe and effective diet therapy for malignant brain cancer and can be considered as an alternative or adjuvant therapeutic option. A preliminary report on this work has appeared [
35].
Discussion
We found that KetoCal
®, a nutritionally balanced and commercially available ketogenic diet for children with epilepsy, significantly reduced the orthotopic growth and the vascularity of the mouse astrocytoma (CT-2A) and the human glioma (U87). Moreover, KetoCal
® significantly prolonged survival in the tumor-bearing mice. It is important to mention that the anti-angiogenic and growth inhibitory effects of KetoCal
® were observed only when the diet was administered in restricted amounts but were not seen when the diet was administered
ad libitum, or in unrestricted amounts. These findings support previous observations that restriction of dietary calories has powerful anti-angiogenic and anti-inflammatory effects against cancer, including brain cancer [
9‐
11,
51,
57,
58]. Reduced caloric content lowers circulating glucose levels as we found in this study and in our previous studies [
10,
11]. Indeed, tumor growth is more strongly correlated with circulating glucose levels than with circulating ketone body levels [
8]. The reduction in glucose levels following restriction of dietary calories largely accounts for why tumors grow minimally on either restricted ketogenic diets or on restricted high carbohydrate standard diets. Restriction of calories in humans may be difficult to achieve, however, due to issues of compliance. Compliance may be better with KetoCal
® as this diet was designed for managing refractory human epilepsy under calorically restricted conditions. CR, however, is not directly comparable in mice and humans. For example, a 40% CR diet in mice is comparable to therapeutic fasting in humans, which can be difficult for many people [
30]. In addition to reducing circulating glucose levels, a restriction of total calories also reduces potential adverse effects of the high fat content of the diet since energy homeostasis is maximized under CR regardless of caloric origin [
8,
30]. The restricted KetoCal
® diet should therefore be easier to implement than therapeutic fasting for brain cancer patients.
Although we previously showed that CR of a high carbohydrate standard diet or of a rodent ketogenic diet similarly reduce blood glucose levels, which tumor cells depend upon for survival [
10,
11], our findings in this study showed that administration of KetoCal
® under restricted conditions was more effective in elevating circulating ketone bodies than was administration under unrestricted conditions. This is important since mild ketosis, under conditions of reduced glucose availability, is essential for enhancing the bioenergetic potential of normal brain cells [
5,
17,
18]. Additionally, ketone bodies may directly protect normal neurons and glia from damage associated with aggressive tumor growth through a variety of neuroprotective mechanisms [
59‐
64]. In contrast to most conventional brain tumor therapies, which indiscriminately target both normal cells and tumor cells, CR and particularly restricted ketogenic diets such as KetoCal
® are the only known therapies that can target brain tumor cells while enhancing the health and vitality of normal brain cells [
5,
29]. In this regard, the calorically restricted ketogenic diet for brain cancer management stands apart from all conventional therapeutic approaches.
Previous studies indicate that many tumors including brain tumors are largely unable to metabolize ketone bodies for energy due to various deficiencies in one or both of the key mitochondrial enzymes, β-OHBDH and SCOT [
19,
21,
65‐
68]. Our gene expression results in the mouse CT-2A and the human U87 brain tumors are consistent with these previous findings in other tumors and also support the mitochondrial defect theory of cancer [
5,
23,
24,
69]. The deficiencies in these enzymes, however, are important for tumor management only under calorically restricted conditions when glucose levels are reduced and when cells would require ketone bodies for energy. This is most evident from the analysis of tumor growth in the unrestricted KetoCal
® groups where growth was rapid despite mild ketosis. This is attributed to the maintenance of high glucose levels, which the tumor cells will use for energy in preference to ketone bodies due to their dependency on glycolysis. However, when glucose becomes limited, as occurs under CR, the SCOT and β-OHBDH deficiencies would prevent tumor cells from using ketones as an alternative fuel thus metabolically isolating the tumor cells from the normal cells. We suggest that the genes for these enzymes could be useful markers for screening brain tumors and other tumor types that may be responsive to therapy using restricted KetoCal
® or other restricted ketogenic diets. Further studies will be necessary to test this interesting possibility.
Long-term management of malignant brain cancer has been difficult in both children and adults. This has been due in large part to the unique anatomical and metabolic environment of the brain that prevents the large-scale resection of tumor tissue and impedes the delivery of chemotherapeutic drugs. Further, significant neurological abnormalities are often observed in long-term brain tumor survivors [
70‐
72]. In light of the differences in energy metabolism between normal brain cells and brain tumor cells, we recently proposed an alternative approach to brain cancer management based on principles of evolutionary biology and metabolic control theory [
5]. Specifically, the genomic and metabolic flexibility of normal cells, which evolved to survive under physiological extremes, can be used to target indirectly the genetically defective and less fit tumor cells. The results of this study using restricted KetoCal
® as a therapy for experimental brain cancer provide direct support for this alternative approach to brain cancer management. It should be recognized that this alternative therapeutic approach may not be restricted only to brain cancer, but could also be effective for any cancer types containing genetically compromised and metabolically challenged tumor cells. Moreover, KetoCal
® will be more effective in managing brain tumors in humans than in managing brain tumors in mice since prolonged caloric restriction can be tolerated better in humans than in mice due to intrinsic differences in basal metabolic rate [
30].
Although Nebeling and co-workers were successful in managing childhood astrocytoma with a medium-chain triglyceride ketogenic diet [
29,
73], this KD formulation can be difficult to implement, is not standardized, and can produce some adverse effects as previously observed in children taking the diet for epilepsy management [
32‐
34]. It is noteworthy that the children treated in the Nebeling study also expressed reduced blood glucose levels [
29]. When administered in restricted amounts, KetoCal
® may have greater therapeutic efficacy with fewer side effects than medium-chain triglyceride or lard-based KDs. Additionally, KetoCal
® would eliminate or reduce the need for antiepileptic drugs for brain tumor patients since KetoCal
® was designed originally to manage refractory epileptic seizures. Adjuvant steroidal medications, which are often prescribed to brain tumor patients and which can produce severe adverse effects, might also be reduced under the restricted KetoCal
® diet since glucocorticoid levels increase naturally under calorically restricted conditions [
74‐
76]. Ketogenic diets and calorically restricted diets can also antagonize cancer cachexia [
9,
11,
77]. These observations, considered together with the anti-angiogenic effects of the diet, suggest that the restricted KetoCal
® diet can manage brain tumors through multiple integrated systems.
Guidelines for the implementation of KetoCal
® and other calorie restricted KDs in younger and older patients have been described previously [
5,
29,
34,
73]. KetoCal
® could be administered to patients with brain tumors at medical centers or clinics currently using the ketogenic diet for managing epilepsy. Based on our findings in mice and those of Nebeling and co-workers in humans, initiation of randomized clinical trials are warranted to determine whether KetoCal
® is effective for the long-term management of malignant brain cancer and possibly other glycolysis dependent cancers [
78]. This is especially pertinent to patients with glioblastoma multiforme, an aggressive brain tumor type for which few effective therapeutic options are available. While KetoCal
® was formulated for managing childhood seizures, it is likely that new KD formulations can be designed with nutritional and caloric compositions more appropriate for managing brain tumors [
5,
78]. This could also involve the use of low glycemic diets, which are effective in maintaining low circulating glucose levels and are easier to implement than some ketogenic diets [
79,
80]. Additionally, the diets could be combined with specific drugs to further enhance therapeutic efficacy. As a cautionary note, the calorically restricted KD would not be recommended for those few individuals with fasting intolerance due to defects, either inherited or drug-induced, in carnitine or fatty acid metabolism [
49]. Our results in mouse and human brain tumor models suggest that the restricted KetoCal
® diet will be an effective alternative therapy for managing malignant brain cancer in humans and should be considered as either a first line or adjuvant therapeutic option.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
WZ carried out all described methods and drafted the manuscript. PM helped with the design of the study and data analysis. MAK participated in the design of the primers, sequence and data analysis. WTM participated in the feeding of the mice, tumor implantation, and data analysis. JGM carried out the ketone assays, and participated in data and statistical analysis. TNS conceived the study, participated in its design and coordination and helped prepare the manuscript. All authors read and approved the final manuscript.