Cancer cells have an altered metabolism which promotes cell survival and proliferation (De Berardinis and Chandel, 2016; Pavlova, Zhu and Thompson, 2022). Such alteration is due to an increased glucose uptake, an exacerbated production of lactate and an inhibition of oxidative phosphorylation system (OXPHOS) that occurs even in the presence of oxygen, thus leading to a state of aerobic glycolysis (Heiden, Cantley and Thompson, 2009). Indeed, tumor cells reprogram their glucose metabolism, restricting it mainly to glycolysis, a phenomenon named as the “Warburg effect” (Geschickter and Warburg, 1930; Warburg, 1956; DeBerardinis and Chandel, 2020). This complex metabolic rewiring is controlled by mitochondria that finely integrate a variety of intracellular signaling pathways, within the mitochondria or with other cellular compartments, to meet bioenergetics needs and facilitate the uncontrolled proliferation of tumor cells (Ward and Thompson, 2012; Epstein, Gatenby and Brown, 2017; Pavlova, Zhu and Thompson, 2022). Glycolytic fueling has been shown to be associated with activating mutations or copy number alterations in genes, such as RAS, MYC, or inactivation or repression of genes, such as the tumor suppressor TP53 (DeBerardinis et al., 2008; Jones and Thompson, 2009). Besides metabolic alterations, these mutations orchestrate several other hallmarks of cancer, including deregulation of cell cycle and cell death pathways, immunosuppression and enhanced migration (Cannino et al., 2018).

Colorectal cancer (CRC) has one of the highest morbidities and mortality rates among solid cancers, accounting for almost 10% of the global incidence of cancer (Sung et al., 2021). CRC is highly heterogeneous at the molecular level and patients display the presence of distinct metabolic alterations that require effective molecular subtyping strategies for therapeutic intervention. Up to 40% of CRC cases contain RAS mutations (Lowy and Willumsen, 1993; Andreyev et al., 2001; Scott et al., 2020). The majority of these mutations are KRAS, highlighting the importance of studying these defects (Downward, 1998; Matallanas et al., 2011; Slattery et al., 2018). Similarly, dysregulation in TP53 gene is one of the most frequent events that occurs in approximately 60% of CRC patients and TP53 mutational status is highly associated with CRC progression and poorer clinical outcome (Muzny et al., 2012; Robles, Jen and Harris, 2016; Nakayama and Oshima, 2019).

Despite researchers’ efforts in understanding the impact of somatic gene mutations in KRAS and TP53 on bioenergetics, cell proliferation and survival in CRC, the dependence of these mechanisms on the mutational status of the tumor still needs elucidation. In the present study, we investigate, at the single cell level, metabolic signatures in isogenic pairs of human colon cancer HCT-116 cells, which differ by single or double genetic mutations in KRAS and TP53 genes. Employing single cell time-lapse imaging approaches, we here show, how KRAS and TP53 mutations influence bioenergetic parameters, including cellular ATP, NADH and mitochondrial membrane potential dynamics following extracellular nutrients availability.

TP53 deficiency enhances cellular ATP production during metabolic stress and the concomitance of oncogenic KRAS reverses this alteration in colon cancer cells.

In order to investigate how human colon cancer HCT-116 cells, that harbor mutations in KRAS and TP53 genes, respond to nutrients availability from the extracellular microenvironment, we started our experiments by characterizing parental HCT-116 WT (TP53 competent but holding a KRAS mutation on exon two codon G13) and three isogenic derived mutant cell lines covering all four combinations of TP53 and KRAS mutational status (p53 KO, TP53-deficient and KRAS mutated; Hke3, TP53 and KRAS competent and Hke3 p53 KD, TP53-silenced and KRAS competent) as listed in Table 1 [see also Supplementary Figure S1A and materials and methods (Brattain et al., 1984; Shirasawa et al., 1993; Bunz et al., 1998; Sur et al., 2009; Fasterius et al., 2017; Charitou et al., 2019)]. To assess whether TP53 loss and mutated KRAS may affect intracellular and mitochondrial bioenergetics in this colon cancer cell model, we monitored cytosolic and mitochondrial ATP levels at single cell resolution, using ATP-sensitive FRET probes targeted to the cytosol (ATeam) or mitochondria (mitoATeam), respectively (Imamura et al., 2009). Experiments were conducted in all four HCT-116 cell lines by time-lapse confocal microscopy and cells were exposed to extracellular substrates as shown in Figure 1A. Briefly, to stimulate the consumption of intracellular metabolites for survival, HCT-116 cells were first deprived of nutrients for 60 min, after which a baseline was recorded for 20 min as equilibration recording time (still in nutrients deprivation condition). To promote mitochondrial respiration, glycolysis, mitochondrial membrane potential depolarization and disruption of ATP synthesis in this order, cells were exposed to sequential addition of 2 mM lactate (at the 20 min time-point), 5 mM glucose (at 40 min) and 10 µM OXPHOS uncoupler FCCP (at 60 min), and recorded till 80 min (Figure 1A). The addition of FCCP indicates the loss of mitochondrial contribution to cytosolic ATP.

Individual single cell analysis revealed divergent cytosolic ATP concentration across all cell lines during substrates exposure (Figures 1B–D). We first observed that TP53-deficient and KRAS mutated (p53 KO) cells retained the cytosolic ATP depletion observed in the first 20 min of baseline, when cells are starved of nutrients, compared to other mutant cells, indicating that these were consuming reserves of ATP present in the cytosol during that period (Figures 1C,E). Moreover, TP53 and KRAS mutations drove a shift in metabolic signatures enabling lactate to become an efficient metabolite to replenish cytosolic ATP following stress induced by nutrient starvation (Figures 1B–D,F). Of note, TP53-silenced and KRAS competent (Hke3 p53 KD) cells displayed the greatest alterations in cytosolic ATP following substrates addition compared to the other cell lines. In detail, lactate and glucose triggered a rapid and constant cytosolic ATP production while a prompt drop of ATP was observed when Hke3 p53 KD cells were subjected to the mitochondrial protonophore FCCP, as shown by quantification of the FRET/CFP ratio (Figure 1D), analysis of the slope of cytosolic signal (Figure 1F) and the corresponding fold change in fluorescence (Figures 1G,H), suggesting that TP53-silenced Hke3 cells generate more cytosolic ATP after metabolic stress.

Similarly, mitochondrial ATP dynamics were determined at single cell level as above-described and analysis of cells monitored showed that, differently to what was detected with the cytosolic ATP-sensitive FRET probe, exposure to lactate to fuel mitochondrial respiration, exhibited minimal changes in mitochondrial ATP in any of the cells regardless of their mutational status (Figures 2A–C,E), indicating that the majority of produced ATP, following lactate addition, was immediately exported into the cytosol. If anything, cells competent in TP53 and KRAS (Hke3) maintained stable mitochondrial ATP basal levels (Figure 2D) and displayed the highest mitochondrial ATP production in response to lactate (Figure 2F). On the other hand, TP53-silenced Hke3 and KRAS competent cells (Hke3 p53 KD) showed significantly increased mitochondrial ATP levels when glycolysis was promoted (Figures 2B,C,F), and the greatest mitochondrial ATP decrease when mitochondrial ATP synthesis was disrupted, compared to the other mutants and parental cell lines (Figures 2B,C, E-G).

In order to confirm whether the ATP production/consumption observed at the single cell microscopy level reflected population-based studies, a biochemical luciferase assay was performed to measure the overall ATP concentration. Indeed, we found that Hke3 p53 KD cells displayed the highest ATP concentration (29.95 μM/μg) following 80 min of starvation at population level compared to the other cells (Figure 3A). In contrast, the lowest overall ATP levels were observed in p53 KO cells (21.58 μM/μg; Figure 3A), indicating that a varying combination of deficiency in TP53 and mutational KRAS status is required to produce significant differences in cellular ATP concentration following nutrient starvation.

Metabolic flux technology using the Seahorse analyzer has emerged to assess the bioenergetic state of cells in vitro/ex vivo. For this reason, we also performed glycolytic capacity and respiratory control assays in Hke3 and Hke3 p53 KD cells using a Seahorse analyzer. Cells were sequentially treated as follows at the indicated times (see also Materials and Methods): Glucose 10 mM (G), Oligomycin 3 μM (O), FCCP 0.5 μM (F), and 1 μM Antimycin/1 μM Rotenone (A/R; Figure 3B) and ECAR and OCR were measured. We observed an increase of the basal respiration (Figure 3C) and mitochondrial Respiratory Control Ratio (RCR = ATP synthesis/proton leak; Figures 3D,E), showing tightly-coupled mitochondria, in the presence of glucose as substrate in TP53-silenced Hke3 cells compared to TP53 competent Hke3 cells, implying that loss of TP53 promoted enhanced mitochondrial function. However, Mitochondrial Uncoupled Respiration (MUR) was not significantly altered (Figure 3F) in Hke3 p53 KD cells compared to their control.

In parallel, to characterize whether TP53 loss could provide different proliferative properties in Hke3 cells in the absence/presence of challenge, we tested cellular viability using HCS at basal conditions or 24 and 48 h following glucose exposure at normal (5 mM) or high-glucose conditions (11 mM). We observed that TP53-silencing in Hke3 cells did not confer a significant proliferative advantage compared to Hke3 cells either in basal conditions or using glucose as a substrate (Figures 3G,H).

Collectively, these experiments demonstrated that TP53 deficiency enhances ATP production during metabolic stress by nutrients availability and the concomitant presence of oncogenic KRAS reverses these ATP alterations (data summarized in Table 2).

In the present study, we identified how tumor suppressors and/or oncogenic mutations, such as KRAS and TP53 genes, alter cancer metabolism, and specifically how they rewire cellular bioenergetic parameters. Using genetically engineered human colon cancer HCT-116 cells in combination with single cell imaging approaches, we demonstrated that TP53 and KRAS mutations and nutrients availability from the extracellular microenvironment affect cellular ATP, NADH and mitochondrial membrane potential dynamics both at cytosolic and mitochondrial level. First, we showed that, in WT KRAS cells, TP53 deficiency leads to the greatest increase in ATP following a period of nutrient starvation and an enhanced ATP production in the presence of substrates, such as lactate and glucose. Conversely, KRAS mutation in TP53-deficient cells reversed these alterations. Moreover, KRAS mutation, independently of TP53 status, produces minimal mitochondrial membrane potential variations in response to nutrients availability. Finally, we determine that the simultaneous presence of TP53 deficiency and KRAS mutation triggers significant shifts in cytosolic and mitochondrial NADH levels during metabolic stress induced by nutrients deprivation or accessibility.

KRAS and TP53 somatic mutations are two of the most common and well characterized mutations in CRC, a disease that develops in a stepwise fashion in terms of genetic mutations, activation of oncogenes and loss of function of tumor suppressing genes (Vogelstein et al., 1988; Fearon and Vogelstein, 1990; Smith et al., 2002; Hasbullah and Musa, 2021). In the recent years, several studies have been focused in understanding the effects of KRAS mutation and TP53 loss in numerous cell models and systems, largely demonstrating that they play a role in cancer cell energy metabolism through regulation of processes, such as glucose transport, lactate metabolism and mitochondrial function. In detail, it has been shown that oncogenic KRAS alters ATP production, decreases aerobic respiration and increases ROS generation, while TP53 dampens glycolysis and modulates OXPHOS and glutamine metabolism (Chen and Russo, 2012; Tarrado-Castellarnau, de Atauri and Cascante, 2016). Moreover, other research groups revealed that KRAS mutation, on its own, mainly increases cell proliferation, reprograms metabolic pathways and affects mitochondrial metabolism (Weinberg et al., 2010; Gaglio et al., 2011; Pylayeva-Gupta, Grabocka and Bar-Sagi, 2011; Simanshu, Nissley and McCormick, 2017). Similarly, p53 influences metabolic signaling through a number of different mechanisms, including the interaction with substrates and enzymes involved in energy metabolism (Vousden and Prives, 2009; Frezza and Martins, 2012; Berkers et al., 2013; Kruiswijk, Labuschagne and Vousden, 2015), i.e the activation of p53 by malate dehydrogenase (MDH) in response to low glucose (Lee et al., 2009) or increasing flux through the pentose phosphate pathway (PPP), which promotes anabolism needed for cell growth (Buzzai et al., 2007). Our study supports these previously reported activities, but also gives new important insights in KRAS and TP53 role in cancer metabolism. In fact, our data obtained from single cell imaging experiments suggested that TP53 deficiency increases cytosolic, mitochondrial and overall cellular ATP production during metabolic stress induced by nutrients availability, including substrates deprivation, glucose and lactate exposure, in HCT-116 colon cancer cells (Figures 1, 2 and Table 2). It is worth to note that cytosolic ATP depends on the amount of ATP produced and cellular ATP consumption. The kinetics of the change in cytosolic ATP concentration is then an indicator how fast this new equilibrium is reached. However, we have also to consider a change of mitochondria from ATP producer to ATP consumer with the possibility of a reversal of the ATP synthase after mitochondrial membrane potential depolarization. For example, mitochondrial ATP in cells with lactate and glucose as metabolites, treated with FCCP, increased which is an indicator that, at least the uptake of ATP from the cytosol into the mitochondrial matrix took place (a requirement for the ATP synthase reversal). Nevertheless, this did not seem to lead to a recovery of mitochondrial membrane potential which would have been indicated by an increase in TMRM nor did it lead to a recovery of mitochondrial NADH. Moreover, cell population measurements of mitochondrial and glycolytic metabolism in Hke3 and Hke3 p53 KD cells using a Seahorse analyzer demonstrated that TP53-silenced Hke3 cells display an increase of the basal respiration and tightly-coupled mitochondria, in the presence of glucose as substrate, compared to TP53 competent Hke3 cells (Figure 3). It is interesting in this context that CRC cells may rely more on OXPHOS than glycolysis compared with normal colon cells (Kaldma et al., 2014; Chekulayev et al., 2015). However, as all population-based assays, this technology is lacking high-temporal resolution. In addition, measurement of bioenergetics in individual cells, their cell-to-cell heterogeneity, and measurement of bioenergetics in specific cellular compartments (such as cytosol vs mitochondria) are only possible at single-cell resolution. Single cell analyses employing cytosolic and mitochondrial FRET-based ATP probes, fluorescent NADH sensors, or probes such as TMRM also enable dynamic computational modelling of cellular bioenergetics in dependence on their mutational status, allowing the identification of new therapeutic targets.

To date, what is known in literature is that glycolytic HCT-116 TP53-deficient cells generate more ATP when compared to TP53-proficient cells (Matoba et al., 2006) and TP53 deficiency can increase metabolites flux through PPP activity, thereby reducing substrates for ATP production during glycolysis (Flöter, Kaymak and Schulze, 2017). p53 also responds to changes in energy levels in cells (Okorokov and Milner, 1999) and it inhibits the expression of glucose transporters GLUT1 and GLUT4, thus dampening glycolysis (Schwartzenberg-Bar-Yoseph, Armoni and Karnieli, 2004; Kondoh et al., 2005), also under hypoxic conditions by inducing RRAD (RAS-related associated with diabetes), the Ras-related small GTPase (Zhang et al., 2014). This would further support our results showing that TP53-silenced Hke3 cells (Hke3 p53 KD) import glucose faster than TP53-proficient cells (Hke3), leading to a larger increase in fluorescence (Figures 1, 2 and Table 2). This effect is accompanied by the ability of p53 to support mitochondrial fatty acid oxidation (FAO) and drive OXPHOS (Lebedeva, Eaton and Shadel, 2009; Liu et al., 2014). Intriguingly, the concomitant presence of oncogenic KRAS in TP53-deficient cells entirely inverts the alteration of cellular ATP levels in response to substrates in our system (Figures 1, 2 and Table 2). In line with our findings, transformed KRAS mouse fibroblasts displayed lower ATP content, reduced OXPHOS ability and more sensitivity to glucose depletion when compared to their immortalized normal counterparts (F. Chiaradonna et al., 2006a; Ferdinando Chiaradonna et al., 2006b). Oncogenic KRAS has also been shown to have significant effects on the activity and function of the mitochondria, suppressing mitochondrial complex I activity (Hu et al., 2012), leading to reduced respiration (Baracca et al., 2010), possibly decreasing mitochondrial ATP generation. Otherwise, KRAS mutated cells use OXPHOS-independent ROS generation through complex III for growth, and utilize glucose metabolism to fuel the PPP rather than ATP production (Weinberg et al., 2010). It has also been shown that the activity of NF-κB resulted enhanced in TP53-deficient primary mouse embryonic fibroblasts and, in these cells, the oncogenic RAS-induced cell transformation and acceleration of aerobic glycolysis were blocked in the absence of NF-κB. However, this was restored by GLUT3 expression, indicating that TP53 loss can facilitate the glycolysis by upregulating the expression of GLUT3 through NF-κB pathway (Kawauchi et al., 2008). Furthermore, it is known that the glycolytic pathway is regulated by p53 through the expression of Hexokinase II (HK2), which controls the production of Glucose-6-Phosphate (G6P) (Mathupala, Heese and Pedersen, 1997; Flöter, Kaymak and Schulze, 2017). Mutations or deletions in TP53 in cancers result in the upregulation of both glucose transporters and HK2 and glycolysis by expression of glycolytic enzymes, like PGM, and inhibition of Tumor Protein 53-Induced Glycolysis and Apoptosis Regulator (TIGAR) (DeBerardinis, 2008). Cancer-associated mutations in TP53 have been shown to result in loss of the ability to block G6PD activity, resulting in an increased PPP flux and glycolysis (Jiang et al., 2011). Another target of TP53 is Glutaminase 2 (GLS2), which is localized in the mitochondria and hydrolyzes glutamine into glutamate, thereby promoting ATP production through the mitochondrial oxidative phosphorylation (Hu et al., 2010; Liu et al., 2019; Yu et al., 2022). In addition, KRAS has also been demonstrated to lead to an increase of glycolytic enzyme expression (Ying et al., 2012; Pupo et al., 2019).

Notably, we also demonstrate that KRAS mutation leads to minimal mitochondrial membrane potential changes, detected in terms of alterations in TMRM fluorescence levels, in response to metabolic substrates and this event occurs independently of TP53 status (Figure 4 and Table 3). Indeed, oncogenic KRAS inhibits mitochondrial function by inducing Hypoxia Inducing Factors 1α (HIF-1 α), which causes increased expression of pyruvate dehydrogenase kinase 1 (PDK1) that, in turn, prevents the activity of pyruvate dehydrogenase (PDH), decreasing the amount of pyruvate that enters the TCA cycle and suppressing mitochondrial O₂ consumption (Kim et al., 2006; Chun et al., 2010; Kaplon et al., 2013; Kierans and Taylor, 2021).

Additionally, our results evidence that, in this human colon cancer cell model, co-operation between KRAS and p53 is crucial in explicating a correct cellular NADH oxidation. In fact, the double concomitant deficiency of TP53 and mutation in KRAS produced significantly altered cytosolic and mitochondrial NADH levels in response to substrates, indicating a clear interplay between them during metabolic stress (Figures 5, 6 and Table 4). In depth, while cytosolic NADH oxidation during lactate exposure occurred rapidly in TP53 and KRAS competent cells (Hke3), thus exhibiting a greater capacity for OXPHOS and superior mitochondrial function, and more slowly during glucose addition in TP53-deficient and KRAS mutated cells (p53 KO; Figure 5), mitochondrial NADH oxidation resulted faster in TP53 and KRAS competent cells (Hke3) during lactate and glucose exposure (Figure 6), implying that both TP53 and KRAS are needed for a proper mitochondrial function. This could indicate higher lactate dehydrogenase (LDH) activity in KRAS-mutated cells or a higher rate of NADH oxidation in WT KRAS cells. For instance, some reports have shown that LDH expression and activity are increased in KRAS mutated lung cancers (McCleland et al., 2013; Xie et al., 2014) and increased GLUT expression in KRAS mutated cells that amplified glycolysis rate in pancreatic cancer (Bryant et al., 2014). Of note, a superior mitochondrial function may lead to increased cytosolic NADH oxidation by the malate-aspartate shuttle, responsible for conversion of oxaloacetate to malate in the inner mitochondrial membrane space and aspartate in the mitochondrial matrix (Birsoy et al., 2015; Sullivan et al., 2015). Oncogenic KRAS produces increased expression of the malate dehydrogenase (MDH) and aspartate aminotransferase (GOT), thus regulating the malate-aspartate shuttle activity (Lyssiotis et al., 2013; Hanse et al., 2017). As TP53-deficient cells displayed lower increases in NADH fluorescence compared to TP53-proficient cells (Figures 5, 6), this may also indicate reduced activity of the malate-aspartate shuttle in cells where TP53 is absent. Several links between TP53-related cancer and cell metabolism and MDH1 have been found (Bensaad et al., 2006; Matoba et al., 2006; Kawauchi et al., 2008; Lee et al., 2009).

In our study, the interplay between different KRAS and TP53 mutational status is obvious and leads to an impairment of both cytosolic and mitochondrial bioenergetics. However, it is worth noting that all our experiments have been performed in atmospheric normoxic oxygen conditions (21%) and we here explore only loss of function of TP53. Indeed, we are aware that cancer tissue oxygenation, and especially in the core of the tumor, would be a hypoxic environment (1–4%) and some p53 missense mutant proteins demonstrate an ability to gain new functions.

Long ago, interactions between RAS and TP53 have been demonstrated to contribute to transformation of normal cells (Eliyahu et al., 1984; Parada et al., 1984). Mutant TP53 has been shown to co-operate with oncogenic KRAS in the formation of PDAC tumors in mouse models (Bailey et al., 2016) and in pancreatic cancer metastasis (Weissmueller et al., 2014). Similarly, combination of oncogenic KRAS and TP53 loss triggered lung tumorigenesis and reliance on catabolism of branched-chain amino acids (BCAAs) as nutrients (Mayers et al., 2016) and KRAS has been shown to suppress p53 by activating the NRF2-regulated antioxidant defense system in human lung cancer cells (Yang et al., 2020). Studies on CRC patients have found that co-mutations of both TP53 and KRAS were rare, suggesting alternate pathways for them in the development of CRC (Smith et al., 2002). Conversely, other reports detected co-existing mutations in 26 out of 140 patients (Tortola et al., 1999), and more recently 26% of 229 patients presented mutations of both KRAS and TP53 (Chow et al., 2016) and a mild correlation was observed between the expression of the KRAS gene and TP53 immunoexpression in colorectal adenocarcinoma (Rachmawati et al., 2019).

Rewiring of cellular metabolism represents a fundamental trait of most cancer cells. Beyond its role in providing energy substrates and biomolecules for cell proliferation, cellular metabolism has also been found to be tightly associated with cancer cell fate and phenotype, and to contribute to the complex tumor heterogeneity that exists within the tumor mass (Lee, 2015). Differences in cellular bioenergetics influences how tumor cells interact with immune cells and other cells of the surrounding tumor microenvironment (TME) (Fu et al., 2017). It has been shown that the presence of TP53 and KRAS mutations has strong implications for the function of the surrounding TME (Michel et al., 2021; Nenkov et al., 2021; Pereira et al., 2022), immune signaling and escape (Mantovani et al., 2008; Hamarsheh et al., 2020). Both immune and stromal cells have been shown to be compromised by TP53- and KRAS-mediated tumor signaling, thereby enhancing tumor development and growth. In CRC, a high prevalence of KRAS mutations correlates with chronic inflammatory diseases, and TP53 mutations contribute to immune surveillance, angiogenesis and remodeling of the extracellular matrix (Michel et al., 2021; Nenkov et al., 2021; Pereira et al., 2022).

In conclusion, our results identify how, in colon cancer cells, the oncogenic mutation in KRAS and deletion of tumor suppressor TP53 rewire cancer metabolism and, in particular, how they co-operate in this system, to alter cytosolic and mitochondrial bioenergetics.