The ketogenic diet is a restrictive therapeutic diet with high-fat concentration, normoproteic, and hypoglucidic, and is considered a pharmaconutritional therapy (13). Since the 1920s, the ketogenic diet has been used successfully to treat patients with intractable epilepsy.

The term “ketogenic diet” was coined in 1921 by Dr. Russell Morse Wilder, of the Mayo Clinic, to describe a diet that produced a high level of ketone bodies, three water-soluble compounds, β-hydroxybutyrate, acetoacetate, and acetone, produced by the liver in otherwise healthy people when they were starved or on a very low-carbohydrate, high-fat diet (14, 15). The advent of phenytoin and other modern pharmaceutical anticonvulsants in the 1940s, and the subsequent lack of publications by epilepsy centers, meant that KD confined as an “alternative” medicine was largely ignored by mainstream neurologists. For many decades it was used only as a last recourse in children with intractable epilepsy, until 1993, when a refractory case sparked renewed interest in ketogenic diets. Hollywood producer Jim Abrahams took his 2-year-old son Charlie to Johns Hopkins Hospital, where Charlie started a KD and achieved a seizure control after few days. Abrahams created the Charlie Foundation for Ketogenic Therapies to further promote the dietary therapy (16).

Although glycemic modulation by carbohydrate-restricted diets is currently object of study in cancer research, clinical data are still lacking. Preclinical studies and early available clinical trials suggest positive effects on survival in patients with malignant gliomas (17). First studies of glucose utilization in cancer go back to the 1980s (18). The first case report published in 1995, illustrated the treatment of two female pediatric patients, with advanced stage malignant brain tumors (19), with KD based on MCT oil. They demonstrated that KD decreased the availability of glucose to the tumor without alter overall nutritional status. In summary, KD could have several beneficial effects in cancer treatment, but further supporting clinical data are needed (20).

In addition to the Classic Ketogenic Diet Protocol, alternative protocols have been released with the aim of increasing the percentage of carbohydrates to improve the patient’s compliance. These alternative protocols are proposed in a targeted manner as an adjuvant therapy to anticancer therapy. The ketogenic dietary therapies (KDTs) are the Classic Ketogenic Diet (CKD), the Ketogenic Diet with Medium-Chain Triglycerides (MCT), the Modified Atkins Diet (MAD), and the Low Glycemic Index Treatment (LGIT) (Figure 1).

The expectation of these diets is to induce a state of systemic ketosis.

The Modified Atkins Diet (MAD), involves a high lipid intake and significant carbohydrate restriction, while there is no specific range for proteins. Carbohydrates are limited to 10 g/day in the induction phase, and increased to 20–30 g/day after 1 month of treatment. Carbohydrates can be distributed between meals or included in a single meal. Dietary fiber is not calculated, and can be included in the dietary plan. Unlike the classical protocol, which involves precise calculation of macronutrients and calories, MAD allows greater freedom of food choice for the patient, making the diet very palatable and varied (21).

The Low Glycemic Index Treatment (LGIT) was developed in 2002 as an alternative to the CKD protocol. LGIT involves a daily carbohydrate intake equal to 10% of requirements, which corresponds to about 40–60 g/day, 60% of fat, and a protein intake according to individual needs. The LGIT protocol mandates that the glycemic index of foods should not exceed 50. This dietary approach produces comparable efficacy to the classical KD, but tolerability is Pfeifer & Thiele (22).

The ketogenic diet with Medium-Chain Triglycerides (MCT) was set up in 1971 by Dr. Huttenlocher and colleagues to replace the LCTs of CDK. The use of MCTs resulted in increased levels of ketosis, while maintaining good percentages of carbohydrates and proteins. The ketogenic diet with MCTs consists of 70–75% lipids, of which 60% is MCTs. The amount of MCTs may contribute to gastrointestinal disorders, including diarrhea, vomiting, bloating and abdominal cramps. For this reason, a gradual introduction of MCTs is recommended, starting at 30% of the total (23).

The ketogenic diet emulate the body’s response to starvation or fasting by reducing carbohydrates. Liver hepatocytes convert fat from the diet into fatty acids and ketone bodies, allowing them to be used as alternative sources of energy. Ketone bodies, such as acetone, acetoacetate and β-hydroxybutyrate, cross the blood–brain barrier and are used as an alternative energy source to glucose (13). Because of the constant maintenance of blood sugar levels to avoid hypoglycemic coma, hepatocytes must use all available oxalate for gluconeogenesis, causing a reduction of the TCA cycle; nevertheless, acetyl-CoA will continue to be abundantly produced through beta-oxidation of fatty acids, causing its cytoplasmic accumulation. The excess acetyl-CoA is then used for the production of acetoacetate, which will be spontaneously converted to acetone (a volatile compound that is eliminated with exhalation) and 3-β-hydroxybutyrate, the predominant ketone body, by 3-β-hydroxybutyrate dehydrogenase (24). Once produced, the ketone bodies are released into the bloodstream and reach peripheral cells where they are converted back to acetyl-CoA. Peripheral tissues, that are unable to do gluconeogenesis and have availability of oxaloacetate, use acetyl-CoA in the Krebs cycle (Figure 2).

Warburg and his co-workers demonstrated in the 1920s that, cultured tumor tissues, despite the availability of oxygen, uptake high amounts of glucose metabolized to lactate, a phenomenon known as aerobic glycolysis. These metabolic properties constitute the Warburg effect. Long before Warburg, Pasteur demonstrated that sugar fermentation is suppressed in the presence of oxygen, thus identifying the conversion of glucose to lactate as an expected response to hypoxia (28). However, cancer cells, before exposure to hypoxic conditions, appear to use glycolytic metabolism (29). Warburg thought aerobic glycolysis was a symptom of impaired oxidative metabolism, believing cancer to be a mitochondrial metabolic disease (30). Despite the extensive scientific evidences describing mitochondrial anomalies in various cancers, many investigators have demonstrate that mitochondria and OXPHOS are not seriously compromised or even intact in some cancer cells (31). In fact, also Warburg’s own experiments revealed persistent oxygen consumption in tumor tissue. Besides OXPHOS, approximately 11% (4/36 total ATP molecules) of total cellular energy is produced through substrate-level phosphorylation (SLP) reactions in the mitochondria (2 ATP) and cytoplasm (2 ATP). For this reason, the mol of ATP synthesized per mol of oxygen used (P/O ratio) is not showing the mitochondrial function in particular in cancer cells (32). One possibility is that the Warburg effect allows cells to maintain large pools of glycolytic intermediates that promote utilization of the pentose-phosphate pathway and other biosynthetic pathways branching off from glycolysis. Respiration promote cell proliferation by allowing the production of tricarboxylic acid (TCA) cycle intermediates necessary for anabolism. Mitochondrial DNA sequencing of 1,675 tumor biopsies from over 30 different types of cancer revealed a number of mutations that have deleterious effects on respiration (33). Other evidence of mitochondrial activity in human tumors revealed significant fuel oxidation in human brain and lung tumors (34, 35). Impaired oxidative metabolism may stimulate aerobic glycolysis in cancer, but, conversely, aerobic glycolysis does not predict loss of oxidative metabolism. This crucial conceptual framework of bioenergetic balance suggests that cancer cells are as dependent on the Warburg effect as are undifferentiated cells, the stem cells and multipotent cells, that can obtain mainly ATP from glycolysis rather than from OXPHOS, even in the presence of sufficient oxygen (29). This choice in metabolic control thus results from a return to embryonic/stem epigenetic programming and the ability to program cellular metabolism in a way that sustains cell growth, activating proliferative versus differentiative programs (36). A better understanding of the mechanical links between cell metabolism and growth control may ultimately lead to improvements in current cancer treatments. Pharmaconutrition, e.g., ketogenic diets, should be used early, especially immediately after surgery to reduce the glucose inflow and prevent triggering the Warburg effect.

Epigenetic modifications are biochemical alterations in DNA or chromatin structure that do not alter the sequence of DNA, but affect the regulation of gene transcription. The epigenome can thus be defined as the organism’s ability to respond and adapt to environmental changes (37). Nutritional assessment and his factors, metabolites, nutrients, demonstrate to be a promising and important epigenetic modulators; growing scientific literature suggests that epigenetic changes have a connection with the ketogenic diet (38). The effect of KD on the genome and gene expression is still largely to be investigated. KD might be able to modulate gene expression directly or indirectly through various mechanisms such as the regulation of DNA methylation, modification of histone conformation, by acetylation, methylation, phosphorylation, ubiquitylation and beta-hydroxybutyrylation of lysine. Beta-hydroxybutyrylation of histones would seem to be exclusive to ketone bodies. These epigenetic modifications could underlie the ability of KD to modify the expression of oncogenes and tumor suppressors (39). Indeed, it has been observed that DNA methylation, miRNAs, and histone modifications occur during the early stages of cancer progression and metastasis (40). In glioblastoma (GBM), epigenetic modifications could modulate the inflammatory gene-oncogene loop and associated immunosuppression (41), modifying the inflammatory tumor microenvironment and playing a key role in the mechanisms of carcinogenesis. Increasingly accurate knowledge of these mechanisms could pave the way for new therapeutic strategies targeting the inflammatory tumor microenvironment.

Many studies have established that one of the phenotypes of cancers is increased production of reactive oxygen species (ROS) and oxidative stress (91). ROS are highly reactive oxygen-containing molecules. They are produced in mitochondria, peroxisomes and the endoplasmic reticulum during various biochemical processes. ROS are also generated after exposure to physical agents (ultraviolet radiation and heat) and after chemotherapy and radiotherapy in cancer. Increased ROS levels constitute one of the many hallmarks of a cancer cell. These changes in cancer cells are characterized by respiratory dysfunction, low coupling efficiency of the mitochondrial electron transport chain, increased electron leakage and amplified ROS generation (92). These mechanisms underlie cancer development, including initiation, promotion and progression. Indeed, increased intracellular levels of ROS can result in the activation of oncogenes and oncogenic signals involved in apoptosis, angiogenesis, cell survival and cancer proliferation and metastasis (50) (Figure 7). However, this increase in oxidative stress is maintained at borderline levels by up-regulation of endogenous antioxidant systems (93), allowing them to survive and proliferate in a redox state unlike healthy cells.

There is, however, a threshold points beyond which ROS irreversibly damage and induce apoptosis in cancer cells. This phenomenon has stimulated the of pro-oxidant therapies, such as radiation and chemotherapy, with the intention of disrupting the oxidative balance to induce cell death (94). Many studies demonstrated that the ketogenic diet also elevates oxidative stress in tumor cells, while simultaneously being neuroprotective the production of reactive oxygen species via the mitochondrial Co-enzyme Q pair in normal cells (95). Several studies have shown that a state of ketosis protects healthy cells from oxidative stress by simultaneously decreasing the production of ROS and increasing that of antioxidants. In normal cells, KD may increase the expression of mitochondrial uncoupling proteins (UCPs) through the regulation of Sirtuin 1 activity (96), potentially by altering NAD⁺ levels, although the precise mechanisms are not yet known. Increased UCP activity reduces the mitochondrial membrane potential and lowers the production of ROS and reactive nitrogen species. In addition, BHB inhibits HDAC, leading to an increase in histone acetylation on the FOXO3a transcription factor promoter (97), which induces the expression of the antioxidant proteins superoxide dismutase (SOD) and catalase, resulting in a reduction of ROS. Conversely, intracellular accumulation and active KB signaling could further disrupt cancer cell homeostasis, by affecting ROS levels and glycemic stress-inducing reduction in metabolically maladaptive brain cancer cells (Figure 8) (98).

The reduction of glycolysis process (↓ATP), decreased redox potential (↓NADPH), and reduced availability of ribose-5-phosphate will lead cancer cells toward a an apoptotic process and confer vulnerability to cytotoxic chemo-and/or radiotherapy (99). In addition, a reduction in GH, IGF-1 and insulin levels further reduces the processes of proliferation, angiogenesis and invasion of tumor cells, reducing the pro-survival stimuli mediated by Akt/mTOR and Ras/MAPK signaling (100). In one study, researchers examined the effects of KD, radiation, and chemotherapy on oxidative stress in cancer cells (101). They demonstrated that animals fed KD in combination with radiotherapy showed an increase in lipid peroxidation and oxidative damage that could be related to a reduction in metabolic substrates with antioxidant action.

Inflammasome activation is essential for the host to counteract external pathogens or tissue damage, while abnormal inflammasome activation can cause runaway responses that can contribute to autoinflammatory disorders, cardiometabolic diseases, cancer, and neurodegenerative diseases. When innate immune receptors are activated by various stimuli, transcription of pro-inflammatory genes is induced, including those encoding NLRP3 and pro-IL1β, but other pathways are now also known. The transcribed pro-inflammatory genes are components of the NLRP3 inflammasome complex. The NLRP3 inflammasome is involved in the molecular mechanism of many autoinflammatory diseases.

Inflammasomes can be activated by different stimuli, including microbiome-derived and host-derived signals. Upon activation, inflammasome sensors NAIP/NLRC4, NLRP3/6/7, AIM2/IHI16 initiate inflammasome assembly by recruiting pro-caspase-1 filaments. It is mandatory to maintain a fine balance between inflammasome activation and inhibition, which requires a refine of inflammasome assembly and effector function. What do we now know about the therapeutic role of the ketogenic diet in regulating the fine balance of the inflammasome?

The identification of β-hydroxybutyrate as an NLRP3 inhibitor provides a rationale for investigating the efficacy of an anti-inflammatory ketogenic diet for the treatment of NLRP3-dependent autoinflammatory diseases. Several mechanisms by which ketogenic inhibits or regulates the inflammasome have been identified in many papers, such as:

Rearrangement of tumor microenvironment (TME) components through mutual and dynamic crosstalk is an important factor influencing the process of tumor formation and progression (105). The composition of the TME differs among tumor types, but include cells of the immune system, a convolute network of stromal cells, fibroblasts, blood vessels, and extracellular matrix (106). GBM cells can manipulate the microenvironment around them by developing a niche that supports tumor growth and development (107), and in this tumor TME is is likely to be the key factor compromising treatment efficacy. Non-tumor glial cells and endothelial cells can directly interact with GBM cells and/or modulate GBM metabolism, release soluble molecules such as inflammatory factors, triggering signaling pathways that promote cell invasion, tumor progression and development (108). In one study, researchers demonstrated that a Classical Ketogenic Diet can potentially be used to alter the glycolytic phenotype of GBM’s cells, differently in a murine models the high-fat diet has been show to modify the polarization of macrophages toward a paradoxical “pro-tumor” immune-suppressive response. Mitigation of this immune response, the well known anti-inflammatory properties of this diet appear to be a consistent perspective themes of investigation. Mitochondrial dysfunction (Warburg effect) is not a general feature of all cancer cells within the tumor mass, as has been demonstrated by the existence of oxidative tumor cells. The numerous reports suggesting that OXPHOS is normal or not severely impaired in GBM cells is not consistent with this basic principle.

Several preclinical studies have shown that KD reduces tumor symptoms (oedema, hemorrhage) and tumor volume and prolongs the survival rate (Table 1). Thus, KD modulates gene expression by reducing ROS and oxidative stress molecules, blocks the energy metabolism of cancer cells and enhances cell death through apoptosis. Although KD alone does not achieve the same results, it enhances the effect of chemotherapy on tumor cells. The treatment is safe and feasible: only a few patients have reported mild side effects, such as borderline glycaemia and/or hyperketonemia. Stafford et al. showed that KD treatment reduces the tumor growth rate and prolongs survival. KD modifies the gene expression profile and specifically modulates oxidative stress genes (cyclooxygenase 2, glutathione peroxidases 3 and 7, and peroxiredoxin 4) and reduces the production of reactive oxygen species (ROS) in tumor cells (109). Maurer et al. (26) showed that glioma cells were not able to metabolize ketone bodies. By reducing blood glucose concentration and increasing ketonemia, KD represents a metabolic disadvantage for tumor cells. KD was used in a xenograft model as monotherapy, but without success. A therapeutic combination of KD and a glutamine antagonist, 6-diazo-5-oxo-L-norleucine (DON), was administered in two mouse models to treat orthotopic growth of advanced GBM. The effect of this pharmaconutritional approach was to reduce tumor volume, oedema, hemorrhage, inflammation and disease symptoms, improving the overall survival of the mice (110). Vallejo et al. (111) showed that acetoacetate regulates mitochondrial uncoupling protein 2 (UCP2) and induces apoptosis through caspase activation and PARP cleavage. The study also showed the synergistic effect of the low-dose (50 μM) glycolytic inhibitor 2-deoxy-D-glucose and acetoacetate, which resulted in increased cell death compared to single treatments.

Another study demonstrating the synergistic role of KD was conducted by Maeyama et al. in 2021. They compared the effects of single treatment with bevacizumab and the combination of the drug with KD in a mouse model of glioblastoma. Diet alone showed no significant benefit, but the diet-therapy combination instead showed an enhanced effect of the drug, with reduced tumor growth and increased survival (112).

Only two studies have been performed to evaluate the outcome of KD in pediatric patients. Nebeling et al. described the cases of two children with high-grade astrocytoma. On PET scans, an average 21.8% decrease in glucose uptake in the tumor region was observed. One patient, in particular, showed improvements in mood and developed new abilities during treatment with KD. During the 12 months of treatment, the patient showed no disease progression (19). Perez et al. (113) conducted a study of 5 pediatric patients with DIPG treated with KD for at least 3 months: they showed slightly better survival than reported in the literature (18.7 months vs. 11 months).

Recently, a study performed on 13 glioma patients evaluated a combined treatment (KD, radiotherapy and metformin): it resulted in an increased survival rate and synergistic action with radiotherapy. In vitro, metformin activates AMPK by inhibiting mTOR complexes, with antiproliferative activity (114). One study explored the expression of the enzymes involved in ketone body metabolism in tumor cells compared to healthy controls by immunohistochemical method. Mitochondrial ketone-limiting enzymes (OXCT1 and BDH1) were found to be low in most tumors, which could justify a better response to KD in these patients (27). Confirming this, another study using KD as monotherapy showed tumor progression in two patients with tissue expression of at least one of the two critical mitochondrial ketolytic enzymes (succinyl CoA: 3-oxoacid CoA transferase, beta-3-hydroxybutyrate dehydrogenase 1) (115). One study explored the outcome of KD in a woman with GBM. After 2 months of treatment, a 20% weight loss was observed and no tumor tissue was detected on both FDG-PET and MRI. At 10 weeks after discontinuation of KD, MRI showed tumor recurrence (116). Rieger et al. in 2014 compared the median PFS after a combined treatment with KD and bevacizumab with the PFS of standard treatment. Six of 7 (86%) patients treated with the combination protocol had a median PFS of 20.1 weeks compared to the median normal PFS of 5 weeks. A 6-month PFS of 43% was observed (117). A retrospective study investigated the effects of KD in 6 patients with GBM compared to the standard diet. KD showed a reduction in mean blood glucose in KD-treated patients in the absence of significant side effects. Four patients showed an improvement in survival (14 months) (118). A study by Strowd et al. (119) reported an increase in seizure control and survival rate in eight patients with glioma undergoing MAD. A recent study on 25 patients on an 8-week Glioma Atkins-Based Diet (GLAD) observed the accumulation of ketones in brain tissue and glioma by magnetic resonance spectroscopy. The increase in brain ketone concentration was correlated with ketone concentrations in blood and urine. Magnetic resonance spectroscopy showed an accumulation of ketones in the lesion but also in the remaining brain parenchyma (120). A study of 9 glioma patients used proton magnetic resonance imaging (1H-MRS) to assess metabolic changes during treatment with KD. Only in patients with high KD adherence were ketone bodies detected in the brain tissue, even in the area of the lesion (121). Berrington et al. stated the feasibility of 3 T magnetic resonance spectroscopy to monitor brain ketone levels in patients with high-grade gliomas. In his study, patients followed a modified Atkins diet for 8 weeks. The results showed an increase in acetone at week 8 in both the lesion and the contralateral brain and an increase in β-hydroxybutyrate at the lesion site. Acetone concentrations in the contralateral brain were strongly correlated with ketonuria and reduced fasting glucose (122).

A case report by Seyfried et al. (123) showed the effect of combined surgery and KD in treating a right frontal lobe glioma grade 4 IDH mutated. This is an example of a targeted therapy, as both KD and IDH mutation limit glioma growth acting negatively on glycolysis and glutaminolysis.

A randomized clinical trial [NCT01754350] of 50 patients evaluated the combination of radiotherapy with standard diet or caloric restriction-KD. Results published at the time showed a significantly longer PFS and OS in the restricted calories-KD group (124). Martin-McGill et al. (125) studied a larger sample of 172 glioma patients who underwent modified KD (70% kcal fat, 20 g CHO/day): KD was well tolerated, only two patients reported constipation. Another study by van der Louw et al. confirmed that the adverse effects of KD were mild and transient (126). A retrospective study of 29 patients with grade II-IV astrocytoma evaluated the use of MAD during radiotherapy and standard chemotherapy. Pre-and post-radiotherapy MRI images were evaluated for tumor pseudo progression as a marker of radiation sensitization. Ketosis was achieved in 100% of cases and no serious adverse events secondary to MAD were observed. 58% of GBM patients showed pseudo progression after MAD and radiation and temozolomide therapy (127). A particular case report described by Elsakka et al. (128) reported the effect of combined standard of care treatment along with metformin, hyperbaric treatment and KD in a patient with GBM. This innovative protocol showed a reduction in tumor invasion and symptoms of GBM, indicating that the successful outcome could be attributed to, among other things, KD. As evidence of the feasibility of KD, the KEATING study was performed to investigate the use of KD as an adjuvant therapy for 12 patients with GBM. A modified ketogenic diet (MKD) or a ketogenic diet with medium-chain triglycerides (MCTKD) were used. Recruitment and maintenance rates were low: only four patients completed the three-month diet (MCTKD n = 3; MKD n = 1). The outcome of these patients needs to be further evaluated (129).