Cancer-associated adipocytes: key players in breast cancer progression

To meet the extreme needs of energy of dividing cells, alterations happen in the metabolism of all macromolecules, such as proteins, carbohydrates, and lipids, in cancer cells [56]. As described in the “Warburg effect” [57], glycolysis increases in tumor cells despite the presence of oxygen [58]. Aerobic glycolysis provides the advantage of biosynthesis for tumor cells, and high-throughput substrates for glycolysis can efficiently divert carbon to key subsidiary biosynthetic pathways [56]. Furthermore, cancer has heterogeneity in the genetic and microenvironmental parameters that influence cell metabolism. The heterogeneity in tumor metabolism is not only demonstrated in the tumor itself, but also highlight the strong influence of the tumor microenvironment. The extent of tumor heterogeneity among different metabolic stages predicts metastatic specific organs. For example, cancer cells metastasized to the brain usually have flexible metabolic ability to utilize different fuels, such as acetate. And cancer cells metastasized to the liver could maximize the energy resources available to meet the energy demands [59]. However, previous studies on tumor metabolism have paid much attention to cancer cells themselves, ignoring different tumor “compartments” or mimicking of the tumor microenvironment. Recent studies have shown that in addition to cancer cells, different metabolic compartments, such as CAAs, in the tumor microenvironment play essential roles in the process [3, 60‐62]. Glycolytic stromal cells promote tumor growth in breast cancer [63, 64]. The absence of Cav-1 is thought to be a hallmark of metabolic stress and is associated with poor survival in breast cancer [65]. Likewise, our results demonstrated that adipocytes co-cultivated with tumor cells exhibited increased UCP1 and monocarboxylate transporter 4 (MCT4) levels and reduced Cav-1 expression compared with single cultivation [66]. UCP1, which is generated only in beige/brown adipocytes, is also elevated in the adipose tissue close to breast cancer tissue [67]. Furthermore, increased expression of UCP1 leads to increased lipid mobilization and energy expenditure in vivo. Conversely, the activation of UCP1+ cells significantly decreases tumor development [67, 68]. In addition, monocarboxylate transporters (MCTs), which are transmembrane proteins, mediate the transport of diverse monocarboxylates, including lactate, ketones, and pyruvate and are related to poor prognosis in breast cancer. Of the MCTs, MCT1 and MCT2 are responsible for the uptake lactic acid to gain energy, while MCT4 facilitates lactic acid efflux to maintain a stable intracellular pH [69, 70]. Our results indicated that the overexpression of MCT1 is observed in hormone receptor-negative and high-proliferative subtypes and that when MCT4 is overexpressed in adjacent adipose tissue, the association between MCT1 expression and poor prognosis of breast cancer is further confirmed [71]. Moreover, it is clear that high expression of MCT2 in primary mammary-derived adipocytes (MGDAs) promotes the growth of breast cancer cells both in vitro and in vivo [72]. Meanwhile, MCT1 is a marker of mitochondrial metabolism (oxidative phosphorylation, OXPHOS), which is involved in the uptake of monocarboxylates and predominantly transports lactate into cells [69, 70]. Mitochondrial metabolism is more efficient and produces more adenosine triphosphate (ATP) to promote tumor proliferation and metastasis. In addition, the absence of caveolin 1 (Cav-1, caveolae protein) promotes aerobic glycolysis of stromal cells, as well as the secretion of energy-rich metabolites (for example, lactate and pyruvate), which are considered as energy resources for cancer cells and provide fuel for the mitochondria of adjacent cancer cells [73, 74]. Mechanistically, hypoxia induces the loss of Cav-1, which targets the MCT4 gene and upregulates its expression.

Overall, it is probably unsurprising that metabolic coupling has been formed between tumors and CAAs though the existence of a “monocarboxylate shuttle” in tumor-stromal interactions.

Mammary fat consists of large lipid droplets, regulates lipid metabolism, and is a major component of energy storage. Cancer cells often invade the adipose tissue and induce adipocytes to release free fatty acids (FFA), which are absorbed by cancer cells and are used to produce ATP and facilitate tumor growth [3, 75, 76]. It is now widely recognized that FFAs may promote cancer progression through multiple mechanisms [77]. First, the sources of FFAs can act as structural units for newly synthesized membrane phospholipids and determine the phospholipid composition of membranes. Cancer cell membranes become enriched with saturated and/or mono-unsaturated fatty acids by enhancing de novo lipogenesis. In addition, due to the denser packaging of saturated lipids, increased lipogenesis also changes lateral and transverse membrane dynamics, resulting in cancer cells that are more resistant to oxidative stress-induced cell death and a limited uptake of drugs [77]. FFAs may also be used to supply energy for cells. Some specific types of tumors exhibit upregulated fatty acid uptake and elevated dependence on FFA β-oxidation as their main energy source [60‐62, 78, 79]. Ultimately, FFAs can also be applied to the biosynthesis of a series of lipid-signaling molecules that promote tumorigenesis. For example, phosphatidylinositol-3,4,5-trisphosphate [PI(3,4,5)P3], a lipid messenger formed by the activation of PI3K that activates AKT to regulate cell proliferation and survival, plays a particularly important role in cancer progression [80]. Our studies showing that fatty acid transports such as CD36 and fatty acid transport protein-1 (FATP1) are primarily detected in most breast cancer tissues located adjacent to the adipose tissues support this idea; in contrast, the expression of fatty acid transports was elevated in adipocytes that were co-cultivated with breast cancer cells compared to the expression in adipocytes cultivated alone. Furthermore, adipocyte-induced CD36 and FATP1 expression contribute to cancer progression by fueling or signaling [61, 78, 81]. Fatty acid-binding proteins (FABPs) are also a series of proteins that serve as intracellular fatty acid transporters that regulate lipid trafficking and cellular responses [82]. The expression of FABP4 was found in ovarian cancer and acute myeloid leukemia at the adipocyte-tumor cell margin, providing fatty acids for rapid tumor growth [62, 79]. However, FABP4 expression is downregulated in the fat tissue surrounding breast cancer [1], while circulating, FABP4 levels in plasma are increased in breast cancer patients, and exogenous FABP4 accelerates breast cancer cell proliferation [83, 84]. Furthermore, FABP4 is also released from these macrophages, and the expression of fatty acid-binding proteins in tumor-associated macrophages (TAM) promotes breast cancer progression [82, 85], suggesting that macrophages, not adipocytes, may be the main source of FABP4 in patients with breast cancer. Moreover, our results indicated that FABP5 mRNA is highly expressed in adipocytes that were co-cultured with breast cancer cells [66]. FABP5 has also been shown to be overexpressed at the invasive front of the breast tumor and linked to poor survival [86, 87]. Similarly, Wang and colleagues described a novel mechanism in which FFAs were released by adipocytes after lipolysis induced by tumor secretions and were transferred into tumor cells to enhance FFA β-oxidation [3], in which FFA transformation maybe attributed to FABP5 high expression [86]. Thus, fatty acid metabolism patterns may alter based on the tissue-specific characteristics in cancers that are produced in the vicinity of adipose tissues and translated into the expected availability of these nutrients.

Ketone bodies are catabolites produced by fatty acid β-oxidation or aerobic glycolysis. Ketone bodies are produced and released by stromal cells and may upregulate cancer cell invasion by increasing OXPHOS [69, 88]. Moreover, it is well known that tumor cells need ketones as components of the tricarboxylic acid (TCA) cycle. Additionally, ketones are an ideal substrate for ATP production in invasive cancer cells because ketone bodies produce more ATP and consume less oxygen than glucose [89, 90]. For instance, breast cancer cells that were co-cultured with CAAs overexpress MCT2, inducing H3K9 acetylation, upregulating several tumor-promoting genes, and increasing the uptake of β-hydroxybutyrate to promote tumorigenesis [72]. CAAs are related to ketogenesis. The rate-limiting step in ketogenesis is the generation of 3-hydroxy-3-methyl glutaryl coenzyme A (HMG-CoA) by 3-hydroxybutyrate dehydrogenase 1 (BDH1), 3-hydroxymethyl-3-methylglutaryl-CoA lyase (HMGCL), and 3-hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2) [88, 91]. Moreover, our results indicated that adipocytes co-cultivated with MDA-MB-231 cells exhibited increased activity of several rate-limiting enzymes, as exemplified by the strong increase in the mRNA levels of BDH1, HMGCL, and HMGCS2 [66]. Consistently, HMGCS2 was overexpressed in fibroblasts to accelerate breast cancer xenograft growth [88, 92]. In contrast, cancer cells exhibit high ketolytic activity and utilize ketone bodies by increasing acetyl-CoA acetyltransferase (ACAT) and 3-oxoacid CoA-transferase (OXCT) expression. Our results also showed that overexpression of ACAT 1/2 and OXCT 1/2 mRNA was observed in breast cancer cells in the presence of adipocytes. Similarly, high expression of OXCT or ACAT stimulates breast cancer growth and metastasis and is related to worse outcomes in breast cancer patients [88, 92, 93]. Furthermore, UCP1 expression has been observed to be increased in cancer-associated fibroblasts (CAFs) adjacent to breast cancer tissue, resulting in lipid mobilization and increasing ketone body production, which promotes tumor growth by supplying high-energy nutrients in a paracrine manner [94]. In addition, high expression of UCP1 was detected specifically in the resident adipose tissue and in adipocytes that were co-cultivated with tumor cells, demonstrating that CAAs may have a similar mechanism as that of CAFs.

In summary, FFAs and ketones play the role of paracrine high-energy nutrients in the “reverse Warburg effect” and can act as monocarboxylates to provide the energy required for cancer biomass synthesis, migration, and invasion.

Amino acids, such as glutamine, glycine, serine, and proline, play important roles in cancer metabolic synergy [89, 95, 96]. In human breast cancer, asymmetric metabolism of these amino acids is also present between cancer and stromal cells.

Glutamine is an important amino acid for the mitochondrial metabolism of cancer cells and replenishes intermediate catabolites for the TCA cycle. Glutamine can be a pivotal source of the TCA cycle intermediates and ATP in cancer cells [97, 98]. Glutamine is also the substrate of the antioxidant glutathione and contributed to nucleotide synthesis [98, 99]. Recently, it has been recognized that stromal cells generate glutamine to promote cancer growth [100]. More recent studies on amino acids in the tumor microenvironment have demonstrated that glutamine can be obtained by cancer cells from the tumor microenvironment by increasing the activity of glutamate ammonia ligase, a key enzyme in glutamine synthesis [101]. Increased glutamine exposure induces the upregulation of glutamine importers and mitochondrial biogenesis in cancer cells, which results in resistance to cancer therapy [100, 101]. Likewise, glutamine could be provided by cultured astrocytes to brain cancer cells [102]. Furthermore, the secretion of alanine and ammonia accompanies glutamine catabolism. Stroma-associated pancreatic stellate cells could release alanine by relying on autophagy, and then cancer cells take up alanine as an alternative carbon source [103]. Ammonia facilitates autophagy and promotes autolysosomes to diffuse [104]. Cancer cells usually exhibit higher expression levels of glutamine metabolism enzymes such as glutaminase-1 (GLS1) and glutamate dehydrogenase (GDH) than those of stromal cells [105]. Mechanistically, several tumor stromal p62 levels decreased, which resulted in decreased consumption of glutamine through the mTORC1/c-MYC signaling pathway. Similarly, p62 deletion in adipocytes increased tumorigenesis through osteopontin secretion, and inhibition of mTORC1 caused adipocytes to regularly shut down energy utilization, which increased nutrient availability for cancer cells [106].

Adipose stromal cells in the tumor microenvironment have also been shown to metabolize symbiotically with cancer cells through arginine metabolis [107]. Citrulline and nitric oxide are generated by cancer cells by consuming and catabolizing arginine, which is mediated by inducible nitric oxide synthase. Nitric oxide not only suppresses oxidative phosphorylation but also facilitates glycolytic activity and proliferation. The citrulline produced in cancer cells is secreted into the extracellular matrix, taken up by stromal adipocytes, converted back into arginine, and released for cancer cells, thus forming this symbiotic metabolic shuttle. Cysteine, as one of the three amino acids of glutathione, contributes greatly to maintaining the redox balance. Given that cancer cells are often subjected to intense redox stress, especially when exposed to chemotherapies, the need for cysteine increases [108]. In blood, cysteine is mainly found in the form of oxidized dimers. Accordingly, cystine is consumed in cancer-associated bone marrow stromal cells to generate cysteine, which is released and then transported into the chronic lymphoid leukemia cells via neutral amino acid transporters. This cysteine in leukemia cells is used to generate glutathione and protect cancer cells from chemotherapy-induced cell death. Similarly, it was reported that ovarian CAFs secreted cysteine and glutathione to adjacent ovarian cancer cells, promoting chemoresistance [109]. In addition, it is reasonable to propose that the release of amino acids mediated by adipocytes may be a pathogenic factor related to worse survival outcomes.

Adipocytes are the main matrix cell type in breast tissue. Because adipocytes surround breast cancer cells closely, the role of adipocytes in the evolution of breast cancer has recently been the focus of more attention. Most of the theoretical basis for studying the role of adipocytes in facilitating breast cancer evolution is associated with epidemiological evidence, suggesting a lower survival rate in obese breast cancer patients [3]. A key point is that we cannot identify whether the influence of adipocytes on breast cancer in obese individuals is different from that in normal or lean individuals based on the significant differences in adipocyte biology between thin and obese individuals, such as secretory protein spectrum changes, adipocyte size, and insulin sensitivity [4]. A laboratory study found that, compared with normal or thin women, the proliferation ability of MCF-7 cells increased when cultured with the conditioned media of adipocytes from obese women (body mass index (BMI): 30–35 kg/m²) [5]. Another study revealed that the estrogen-dependent cell E0771 in the mammary fat pad of mice fed a high-fat diet (a dietary-induced obesity (DIO) and hyperinsulinemia model) had a larger allograft volume than animals fed a normal diet [6]. These two studies showed the specific effects of obesity on breast cancer cell proliferation, whereas few studies have focused on exploring the mechanism by which the obese environment affects the progression of breast cancer. Although adipocytes may be the mediator between obesity and breast cancer, the specific mechanism remains to be elucidated.

Moreover, the expression of programmed death-ligand 1 (PD-L1) in adipocytes prevents the anti-tumor function of CD8+ T cells. The lipogenesis inhibitor then selectively reduces PD-L1 expression in the adipose tissue and enhances the anti-tumor efficacy against PD-L1 or anti-PD-1.

Adipocytes contribute to immunosuppressive effects to avoid tumor elimination by the host antitumor immunity. Interestingly, the expression of PD-L1 has been shown to be significantly elevated in mature adipocytes compared to the level in preadipocytes. The expression of PD-L1 in adipocytes prevents the anti-tumor functions of CD8+ T cells. Then, the adipogenesis inhibitor selectively decreases the expression of PD-L1 in the adipose tissue but enhances the antitumor efficacy of anti-PD-L1 or anti-PD-1 [132]. Moreover, UCP-1 was overexpressed in cancer-associated fibroblasts resulting in the release of ATP-rich vesicles by activating autophagy [94], and our study verified that UCP-1 was also highly expressed in cancer-associated adipocytes after the uptake of tumor exosomes [115]. In obese models, ATP promotes a Th17 polarizing microenvironment by activating the P2X7 receptor pathway and increases the levels of IL-1β, IL-6, and IL-17 [133]. Moreover, IL-1β and IL-6 are also elevated in the tumor-adipocyte microenvironment [1] and interact with adipocytokines to aggravate tumor metastasis [134]. Meanwhile, crown-like structures (CLS) are characterized by adipocytes around infiltrating macrophages, and the CLS of the breast is found in nearly 50% of patient samples and is associated with lower survival [135]. ATP also increased the synthesis and release of CCL2 through activated P2Y2 receptors and attracted macrophages to the surrounding adipocytes [136]. Taken together, these findings imply that tumors may induce the release of ATP from stromal adipocytes to regulate the immune microenvironment and favor cancer progression.

Inflammatory factors (CCL5, IL-6, etc.) and transporters (MCT1, CD36, etc.) are potential targets for breast cancer treatment. So far, there have been some targeted treatments used in the clinical practice or for clinical and preclinical studies, such as Cenicriviroc (inhibitor of CCR2) inhibiting CCL2 [137, 138], Maraviroc (inhibitor of CCR5) inhibiting CCL5 [139], Tocilizumab (monoclonal antibody of IL-6R) [140], Canakinumab (monoclonal antibody of IL-1β) [141], Infliximab (monoclonal antibody of TNF-α) [142‐144], Bevacizumab (monoclonal antibody of VEGF-A) [145], AZD3695 (inhibitor of MCT1) [146], Nivolumab (monoclonal antibody of PD-1) [147], and Durvalumab and Avelumab (monoclonal antibodies of PD-L1) [148, 149] (Table 2). There are still many targets, such as inhibitor of exosomes, worth to be investigated for breast cancer precise therapy.