Cancer is one of the greatest threats to human health and continues to be a central challenge in biomedical research. In recent years, the tumor heterogeneity concept has attracted much attention. Today, tumors are understood as complex systems assembled from different cellular niches, among them, cancer stem cells (CSCs) are a subclass of cells with self-renewal property and multilineage differentiation potential (1). CSCs contribute to several adverse clinical outcomes. They promote metastasis, epithelial–mesenchymal transition (EMT) and resistance to radiotherapy and chemotherapy (2). Understanding the survival mechanisms of CSCs is therefore essential for unraveling the fundamental biology of cancer and developing novel therapeutic strategies (3).

Another outstanding feature of tumor cells is their characteristic metabolic reprogramming that underlies survival and malignant growth. Regardless of oxygen availability, tumor cells generally use glucose to generate ATP to support rapid tumor growth. This phenomenon is called Warburg effect (4). This strategy provided not only energy, but also diverted glycolytic intermediates for biosynthesis such as the pentose phosphate pathway and the serine pathway to provide precursors for the formation of nucleotides, lipids, and amino acids (5, 6). Through metabolic reprogramming, tumor cells enhance their adaptability to dynamic microenvironments and promote malignant proliferation (7, 8).

Initially, CSCs were considered to share similar metabolic features with the bulk of tumor cells. However, in recent years, CSCs were identified as containing unique metabolic programs. Some evidence suggested that CSCs are primarily reliant on oxidative phosphorylation (OXPHOS), whereas others showed the importance of glycolysis as a metabolic energy source in CSCs (7). In addition, CSCs have also shown distinct preferences for lipid metabolism and amino acid metabolism. Recent reviews have emphasized the importance of CSC metabolism in cancer initiation, progression, and metastasis and underscore its critical role as an area deserving more attention to systematic analysis (9, 10).

Ovarian cancer is the deadliest gynecological malignancy with a prognosis of less than 50% 5-year survival rate (11). It is highly prone to chemotherapy resistance and metastasis, imposing a substantial clinical burden. Given the essential role of CSCs in tumor progression, a deeper understanding of ovarian CSC (OCSC) metabolism is important for improving patient prognosis. Building on high-level studies published in the last 5 years, this review sums up the main metabolic pathways that CSCs participate in, the main steps and molecular mechanisms of the OCSC metabolic regulation and the emerging approaches targeting OCSCs, with the aim of revealing the role of metabolism in OCSC survival and its implication in OCSC therapy.

CSCs, also known as tumor-initiating cells (TICs), constitute a distinct tumor subpopulation defined by their capacity for self-renewal and multilineage differentiation. Accumulating evidence shows that these malignant traits of CSCs are closely correlated with CSC plasticity -- particularly their metabolic plasticity. This metabolic flexibility enables CSCs to adapt to diverse microenvironments (12). Therefore, targeting the CSC and controlling its plasticity is a new approach for therapeutic treatments.

Surface markers are widely employed for CSC identification. Researchers have described a variety of CSC markers, with some shared across tumor types. CD133, a transmembrane glycoprotein, is one of the best-established markers and has been identified in multiple solid tumors, including pancreatic carcinoma (13), breast cancer (14), liver cancer (15), ovarian cancer (16), oral squamous cell carcinoma (17), and melanoma (18). Aldehyde dehydrogenase (ALDH), detected through specialized fluorescence-based assays, is a functional CSC marker in breast, ovarian, bladder, head and neck, and lung cancers (19). ALDH activity contributes to reactive oxygen species (ROS) scavenging and maintenance of stem cell homeostasis (20). CD44, the predominant hyaluronic acid receptor, mediates adhesion-related signaling pathways and has been implicated in oral cancer, esophageal carcinoma, and colorectal cancer (17, 21, 22). In addition to surface proteins, core pluripotency transcription factors including SOX2, NANOG, and OCT4 also serve as intracellular markers for CSC identification (23–25).

Although many studies have investigated surface markers of CSCs, these markers are not considered the gold standard for CSC identification. The definition of CSCs is fundamentally based on their functional properties, including self-renewal capacity and tumor-initiating potential. Accordingly, a range of functional assays and culture methods have been developed to isolate and expand CSCs. In vitro, researchers exploit CSC traits such as anchorage-independent growth and adaptation to serum-free culture by employing low-adhesion plates and serum-free media, thereby selectively inducing and enriching the formation of three-dimensional tumor spheres primarily composed of CSCs (26). In vivo, the remarkable tumorigenic capacity of CSCs has been proved through serial transplantation assays. This standardized protocol involves implanting tumor cells into immune-deficient animal models to establish primary tumors. Then isolating cells from these tumors, and re-implanting them into a secondary host (27). This cyclical process generates CSCs in a selective way that enriches the most tumor-initiating powerful sub-population of CSCs. Together, these methods for isolating, identifying, and characterizing CSCs have contributed greatly to research in this field. They further deepen our understanding about the molecular factors involved in the biology of CSCs, especially for OCSCs, and provide the solid groundwork for further studies. The surface markers and functional roles of CSCs in tumor progression is shown in Figure 1.

There has been a large amount of research focused on the molecular dynamics active within OCSCs in recent years, highlighting the key roles that diverse regulatory processes and complex interactions play in shaping OCSC formation. Stemness maintenance within OCSCs is coordinated and regulated by multiple evolutionarily conserved signaling pathways, including the Wnt, NF-κB, Notch, Hedgehog, JAK–STAT, and PI3K/AKT/mTOR signaling axes (23). Moreover, intracellular mechanisms such as epigenetic reprogramming and metabolic plasticity cooperate with extracellular elements of the tumor microenvironment. These include hypoxic conditions and interactions with tumor-associated fibroblasts (CAFs), immune cells, adipocytes, and other stromal elements (28–31). Collectively, these factors identify a multidimensional regulatory network that strongly shapes OCSC self-renewal and therapy-resistant phenotypes. For information on the molecular mechanisms and markers for OCSCs, please refer to our previous review (32).

CSCs display significant metabolic plasticity, including glucose, fatty acid, and amino acid metabolism to sustain stemness and survival. In the case of OCSCs, the metabolic adaptability of OCSCs is characterized by distinct, interconnected programs. These distinct metabolic programs help provide chemoresistance and drive recurrence in them. The interconnected metabolic networks are shown in Figure 2.

Recent research has demonstrated striking metabolic heterogeneity of CSCs, characterized by paradoxical states in which both high-OXPHOS/high-ROS and low-OXPHOS/low-ROS profiles can sustain stemness (16, 52, 56). Such heterogeneity reflects transitions between stem cell quiescence and activated stem cell states (68).

Quiescent stem cells, in oxygen-restricted niches, are generally characterized by low activity, low content of mitochondria and reliance on glycolysis and lipid metabolism. When activated by microenvironment cues or growth factors, these cells toward proliferative self-renewal and differentiation, accompanied by decreased glycolytic dependence and increased mitochondrial OXPHOS (65). These observations show metabolic plasticity, which is a defining CSC characteristic that describes the ability to dynamically reprogram metabolic pathways in response to environmental stressors (69). The active and quiescent duality of the two populations of OCSCs is shown in Figure 3.

Plasticity functions as a common regulatory mechanism that drives CSC metastasis, drug resistance and relapse (70). All proliferative states of CSCs show enhanced OXPHOS, higher lipid metabolism and increased oxygen consumption (56). Metastasis requires many metabolic adaptations (70). Drug resistance arises from distinct metabolic adaptations in different CSC subpopulations. These subpopulations may rely on either highly active OXPHOS or hypoxia-driven metabolic pathways, which are engaged in response to specific therapeutic pressures (71, 72).

During ovarian cancer metastasis—which requires “detachment” followed by “re-adhesion”—the OCSCs exhibit a distinct type of metabolic reprogramming (73). “Detachment” here is driven mostly by EMT and is strongly correlated with high glycolytic activity. For example, overexpression of hexokinase 2 (HK2) increases glycolytic efficiency in OCSCs and increases lactate production, acidifies the tumor microenvironment, and thereby promotes migration, invasion, and metastatic niche formation in vivo (72). By contrast, OCSCs treated with the glycolytic inhibitor lonidamine prevent the progression of EMT and stemness in vitro (74). In contrast, during re-adhesion phase, OCSCs adopt a distinct metabolic profile. Indeed, simulation experiments by Compton et al. concluded that free-floating OCSC spheroids, upon reattaching at secondary sites, shift their metabolic profile from glutamine dependency to metabolism by glucose-mediated OXPHOS, while maintaining steady FAO activity (75). Thus, OCSCs manifest context-dependent metabolic states, such that the detachment phase resembles quiescent stem cells (glycolytic, lipid-dependent, low oxygen consumption), whereas the re-adhesion phase resembles activated stem cells (glucose-fueled OXPHOS). Such adaptability implies metabolic plasticity required for the OCSC survival and metastatic colonization.

Drug resistance of OCSC is also driven by metabolic plasticity. During chemotherapy, OCSCs respond to the stress by reprogramming glucose and lipid metabolism. In ovarian cancer, drug-resistance is marked by increased lipid metabolism. Artibani et al. perform multi-region transcriptomic analyses on paired pre- and post-chemotherapy biopsy samples from 17 patients and detected an adipocyte-like transcriptional signature of minimal residual disease (MRD) cells that overlapped with OCSC markers (76). Further in vitro studies confirmed that MRD-like cells depend on FAO for survival and chemoresistance (76). For example, Huang et al. showed that accumulation of periostin in the microenvironment of recurrent ovarian cancer upregulates fatty acid synthase (FASN) in OCSCs (77). Moreover, aberrant activation of fatty acid desaturases SCD1 and FADS2 in ovarian cancer peritoneal metastases enhance lipid metabolism, facilitating tumor invasion. Subsequently pharmacological inhibition or genetic silencing of SCD1/FADS2 led to delay of tumor growth, suppressed CSC formation, and reversal of platinum resistance. Pre-clinical trials also demonstrated that combining SCD1/FADS2 inhibitors together with cisplatin to blocked tumor dissemination, suggesting a promising therapeutic strategy for peritoneal metastasis in ovarian cancer (78).

Role of glucose metabolism in chemoresistance is less clear. Some studies give a dominant role of aerobic metabolism. For example, CPI-613, a TCA cycle inhibitor, prevents enrichment of OCSCs after chemotherapy (79). Sriramkumar et al. have shown that platinum-based chemotherapy activates SIRT1, which promote OXPHOS in OCSCs, leading to ALDH⁺ stem cell expansion, chemoresistance, and recurrence. Combining OXPHOS inhibitors with SIRT1 inhibitors synergistically promoted prolonged survival of tumor-bearing mice. In contrast, other studies suggest that increase in glycolysis contributes to chemoresistance of OCSCs (80). For example, overexpression of HK2 increases lactate production, acids the microenvironment and induces upregulation of the FAK/ERK1/2 pathway, conferring resistance (69).

There is also evidence of synergistic contributions from both glycolysis and aerobic metabolism. Wang et al. documented that BAG5-mediated resistance is associated with concurrent upregulation of both pathways (71). Taken together, these findings highlight the central role of metabolic plasticity in the treatment resistance of OCSCs. Similar to metastasis, this adaptability allows OCSCs to survive therapeutic stress, which reinforces the need to understand how microenvironmental changes will regulate metabolite reprogramming of OCSCs for the development of powerful targeted therapies.

CSCs provide an essential perspective for understanding disease intractability through their strong capacity for metabolic reprogramming. At the same time, they open promising avenues for therapeutic innovation. In ovarian cancer, however, research specifically focused on OCSC metabolism remains limited, offering a fertile ground for future exploration.

Despite this therapeutic potential, OCSCs research faces substantial challenges, primarily due to the remarkable metabolic plasticity of these cells. Specific surface markers for OCSCs remain poorly defined, and the heterogeneity of OCSC subpopulations marked by different surface antigens, as well as the possibility that these markers change dynamically with cell state, remains unresolved. Moreover, metabolic data derived from in vitro experiments often fail to capture the complexity of the in vivo tumor microenvironment. Constructing reliable in vivo models also remains difficult, and accurately tracking, isolating, and eradicating OCSCs continues to present significant obstacles. These limitations restrict the depth of mechanistic insights and impede clinical translation. Moreover, OCSCs exhibit extensive adaptability across nearly all major metabolic pathways. This poses significant challenges for designing metabolism-targeting interventions. However, it also presents a major opportunity: by disrupting this adaptive network, we may effectively inhibit and eradicate OCSCs.

Looking forward, these challenges also highlight potential directions for advancement. The rapid development of metabolomics and single-cell multi-omics, including the integration of scRNA-seq with metabolic flux analyses, will enable fine-grained profiling of OCSC metabolic signatures and nutrient utilization patterns. This integrated approach will improve our understanding of OCSC heterogeneity, stemness hierarchy, metastatic potential, and therapeutic response by revealing dynamic metabolic adaptations. Similarly, the increasing use of lineage-tracing technologies in vitro and in vivo will offer unprecedented opportunities to dissect OCSC dynamics in real time. Advances in experimental modeling also hold promise, with 3D organoid co-cultures, patient-derived models, organ-on-a-chip systems, and genetically engineered mouse models enabling more physiologically relevant studies. Optimized patient-derived xenograft models, combined with advanced metabolic imaging and omics approaches, may further increase the predictive value of preclinical research. From a therapeutic perspective, repurposing existing metabolism-targeting drugs such as metformin is a practical starting point, while novel drug delivery technologies, including nanoparticles, offer opportunities to improve specificity and efficacy against OCSCs.

In conclusion, targeting OCSC metabolism represents a highly promising direction in cancer therapy. Overcoming ovarian cancer recurrence and treatment resistance will require continued mechanistic investigation, innovative technologies, and optimized preclinical models. With increased research investment and interdisciplinary collaboration, significant breakthroughs are achievable, paving the way for more effective treatments against ovarian cancer.