Obesity and gynecological cancers: A toxic relationship

Abstract Despite the evidence supporting the relevance of obesity and obesity‐associated disorders in the development, management, and prognosis of various cancers, obesity rates continue to increase worldwide. Growing evidence supports the involvement of obesity in the development of gynecologic malignancies. This article explores the molecular basis governing the alteration of hallmarks of cancer in the development of obesity‐related gynecologic malignancies encompassing cervical, endometrial, and ovarian cancers. We highlight specific examples of how development, management, and prognosis are affected for each cancer, incorporate current knowledge on complementary approaches including lifestyle interventions to improve patient outcomes, and highlight how new technologies are helping us better understand the biology underlying this neglected pandemic.

made in preventive and diagnostic medicine, surgery, radiotherapy, and chemotherapy have allowed us to tailor and test targeted therapies (i.e. poly ADP-ribose polymerase [PARP] inhibitors), immunotherapy, and to evolve towards precision medicine approaches. [14][15][16] Regardless of all these improvements, incidence and mortality rates of obesity-related cancers have not improved significantly in the last 30 years. As shown in Figure 2, the average incidence rate for obesity-related cancers is higher and shows an increasing trend among American women.
Specific analysis of incidence and mortality rates and trends in 20 OECD countries (10 with current obesity prevalence under 25% and 10 over 25%) spotlights some relevant issues in gynecologic cancers. For ovarian cancer, both rates remain relatively stable or increase. This is evident for the mortality rates among countries with F I G U R E 1 Trends in obesity rates among women from 20 OECD countries between 1975 and 2016. On the left, trends in countries that currently present obesity prevalence less than 25%. On the right, countries where current prevalence exceeds 25%. The continuous black line (red dots) indicates the average rate trend for 10 countries analyzed. Graphs were built after downloading data from OurWorldInData. org 7 .
F I G U R E 2 Trends in age-standardized incidence and mortality rates for obesity-related and nonrelated cancers among women between 1978 and 2016 in the USA (according to nine SEER registers). Top graphs summarize trends in incidence rates. Bottom graphs show trends in mortality rates. The continuous black line (red dots) indicates the average rate trend for all cancers analyzed. Graphs were built after downloading data from the Global Cancer Observatory [17][18][19] .
higher prevalence of obesity ( Figure 3). For uterine cervix cancer, rates and trends are worse among countries with higher obesity rates ( Figure 4). For uterine corpus cancer, rates and trends are even worse for countries with higher obesity rates and there exists an evident increasing trend in incidence and mortality ( Figure 5).
To address how the progressive increase in BMI impacts on the incidence and mortality rates of gynecologic cancers, we carried out regression analyses using data from two countries with good longterm registers and different obesity prevalence rates: Australia (over 25%) and Norway (less than 25%). Figures 6 and 7 depict the dramatic effects of obesity for uterine corpus cancer, with increasing incidence rates over time as obesity continues to rise.
These reports spotlight the association between obesity and increased cancer risk in gynecologic cancers. 6 The main goal of this review is to summarize the main molecular mechanisms through which obesity contributes to development and affects therapeutic response and patient outcomes in gynecologic cancers. In addition, we identify and expose unanswered questions that warrant further research to modify the current scenario.

| OB E S IT Y AND HALLMARK S OF C AN CER
Obesity impacts cancer hallmarks through different mechanisms ( Figure 8). In gynecological cancers, this occurs mainly through alterations in hormonal, inflammatory, and metabolic pathways. 20 Increased estrogen signaling, obesity-related insulin resistance, and chronic low-grade inflammation contribute in concert to stimulate anabolic processes, inhibit apoptosis, and stimulate cell proliferation, in part, by altering the mitogenic PI3K/AKT/mTOR pathway.
Additionally, obesity leads to major alterations in the lipidic composition of organelles and membrane dynamics which may further contribute to the alteration of cancer hallmarks. 21 Some of the key F I G U R E 3 Trends in age-standardized incidence (1993-2012) and mortality (1990-2016) rates for ovarian cancer in 20 OECD countries. On the left, trends in countries that currently present obesity prevalence less than 25%. On the right, those countries where current prevalence exceeds 25%. The continuous black line (red dots) indicates the average rate trend for 10 countries analyzed. Graphs were built after downloading data from the Global Cancer Observatory [17][18][19] . mechanisms orchestrating alterations in cancer hallmarks are summarized in this section.
1. Sustaining proliferative signaling. One of the principal traits of cancer cells is sustaining proliferative signalling. 14 Adipocyte hypertrophy determines increased expression of aromatase which converts androgens to estrone and estradiol. 20 Additionally, sex hormone-binding globulin levels decrease in obesity, resulting in increased availability of bioactive estrogens. These conditions favor signal transduction by the estrogen α and β receptors leading to activation of the estrogen signaling cascade and downstream activation of the mitogenic PI3K/AKT/mTOR signaling pathway. 20 A central pathway for cell proliferation, this pathway is frequently hyperactivated in many cancers and obesity. Estrogen, insulin, and proliferative inflammatory signaling activate this pathway in obesity. 22 Additionally, chronic low-grade inflammation results in activation of intracellular signaling pathways encompassing nuclear factor-ĸB (NF-ĸB), which regulates interleukin 6 (IL-6). In turn IL-6, via its receptor and intracellular cascade mediated by Janus kinase (JAK) proteins, activates signal transducer and activator of transcription 3 (STAT3), which results in expression of genes that include cyclins, thus inducing cell proliferation. 23 Obesity is also associated with reticular stress and remodeling, and its lipidic composition is altered in obesity. These conditions play a relevant role in proliferative signaling in obesity. 20 In aggregate, these observations spotlight the functional interplay linking obesity, insulin signaling, increased estrogens, and chronic inflammation with sustaining proliferative signaling in gynecological cancers. 20 2. Evading growth suppression. Raft-mediated transforming growth factorβ (TGFβ) signaling negatively regulates cell proliferation. TGFβ signaling is impaired by raft composition alterations, leading to rapid differentiation and cell proliferation.
Interestingly, cholesterol can decrease binding of TGFβ to its receptor and impair signaling by regulating endocytosis and degradation. 21 Therefore, intervention of cholesterol homeostasis by cholesterol-lowering drugs such as statins can be of F I G U R E 4 Trends in age-standardized incidence (1993-2012) and mortality (1990-2016) rates for uterine cervix cancer in 20 OECD countries. On the left, trends in countries that currently present obesity prevalence less than 25%. On the right, those countries where current prevalence exceeds 25%. The continuous black line (red dots) indicates the average rate trend for 10 countries analyzed. Graphs were built after downloading data from the Global Cancer Observatory [17][18][19] . particular interest in patient management in obesity-associated cancers.
3. Avoiding immune destruction. The capacity of the tumor cell to evade the immune antitumor response mechanisms is regarded as one of the key features and hallmarks of cancer cells. 14 Obesity reduces the diversity of T cell receptors (TCRs) on circulating T cells, impacting the number of recognizable antigens. Lymph node size is reduced, as well as the migration of dendritic cells to the lymph nodes and the number of T cells in the lymph nodes. One additional mechanism that may be implicated in the impaired maturation of T cells and their reduced capacity to recognize antigens in obesity is altered composition of membrane rafts. Raft-dependent pathway maturation of T cells is altered in obese patients. 21 Furthermore, suppression of anti-inflammatory pathways enables activation of dendric cells within the white adipose tissue and persistent antigen presentation by dendric cells to T cells. This may lead to T cell exhaustion and reduced T cell effector response. 24 Accordingly, a recent study by Porsche et al. 25 has shown that obesity causes T cell exhaustion within the adipose tissue that is dependent on localized soluble factors and cell-to-cell interactions. Hypercholesterolemia in obese patients hinders the differentiation of hematopoietic stem cells, resulting in negative regulation of natural killer cell production. Additionally, lipid storages in the form of lipid droplets in dendritic cells impair antigen presentation and immunity. This is because lipid droplets act as major eicosanoid reservoirs in the cell, leading to alterations of the immune response. In fact, lipid droplet content alters antigen presentation in a cell type-specific fashion, as well as chemotaxis and phagocytosis. 21 4. Enabling replicative immortality. Increased telomerase activity in cancer cells grants them the ability of unlimited replication. 14 Activated hypoxia-inducible factor 1α (HIF-1α) and phosphoinositide 3-kinase (PI3K) signaling contribute to upregulation of human telomerase reverse transcriptase (hTERT) in cancer cells. 21 These pathways are upregulated in obesity. 26,27 5. Tumor-promoting inflammation. Together with genome instability and mutation, tumor-promoting inflammation is one of the most F I G U R E 5 Trends in age-standardized incidence (1993-2012) and mortality (1990-2016) rates for uterine cancer (endometrial) in 20 OECD countries. On the left, trends in countries that currently present obesity prevalence less than 25%. On the right, those countries where current prevalence exceeds 25%. The continuous black line (red dots) indicates the average rate trend for 10 countries analyzed. Graphs were built after downloading data from Global Cancer Observatory [17][18][19] .
in the activation of glycoprotein 130 (gp130) and downstream activation of JAK proteins and signal transducer and activator of STAT3. The IL-6/JAK/STAT3 pathway has marked effects on the antitumor immune response, exerting an inhibitory effect on neutrophils, natural killer cells, effector T lymphocytes, and antigen-presenting dendritic cells. Therefore, hyperactivation of this pathway seems to have a relevant role in the reduction of the antitumor immune response. 23 Aside from these mechanisms, lipid droplets also affect major inflammatory pathways. Lipid droplets are a major site of eicosanoid synthesis, from which a plethora of cytokine amplifiers originate.
The composition of anti-or proinflammatory molecules determines the type of inflammatory response. Therefore, the altered composition of these molecules in obesity triggers a proinflammatory response that favors tumor formation. 21 Finally, inflammation can impair insulin signaling generating insulin resistance. Increased insulin secretion promotes cell proliferation via stimulation of the PI3K/AKT/mTOR pathway. 22 6. Activating invasion and metastasis. NF-ĸB and STAT3 have been implicated in epithelial-to-mesenchymal transition (EMT), the process by which cancer cells become more invasive and acquire metastatic potential and genomic instability due to impaired DNA repair mechanisms and increased DNA damage. 28 The Wnt signaling pathway is also of pivotal importance in cancer development, as it is both UPR transcription factors in obesity also leads to alteration of antiapoptotic proteins XBP1, ATF6f, and ATF4. 17,25 STAT3 increases survival via upregulating antiapoptotic proteins. 28 10. Deregulating cellular energetics. Cancer cells are known to acquire altered metabolic states favoring aerobic glycolysis over oxidative phosphorylation, a phenomenon termed the Warburg effect. Together with increased glutaminolysis and altered lipid metabolism, they make up the key components in cancer metabolic reprogramming, one of the hallmarks of cancer. 10,14 Metabolic reprogramming is a complex phenomenon orchestrated by several alterations including activation of KRAS, mTOR, MYC, p53. Mutations in mitochondrial DNA and hypoxic conditions that activate HIF-1α can cause mitochondrial dysfunction or inhibition of the mitochondrial respiratory chain, also favoring aerobic glycolysis. 20 Another relevant aspect of this hallmark in obesity is the association it has with insulin resistance. Given that insulin regulates clearance of glucose from the blood, insulin resistance determines postprandial hyperglycemia and hyperinsulinemia. 22 Aside from its effects on glucose homeostasis, excess insulin activates the PI3K pathway, contributing to progression of the disease via increased cell proliferation and inhibition of apoptosis. 30

| OB E S IT Y AND DE VELOPMENT OF GYNECOLOG IC C AN CER S
So far, there is no doubt on the independent and positive correlation between increase in BMI and the risk of developing endometrial adenocarcinoma, particularly the type 1 or endometrioid variant. 31 These tumors are estrogen respondent and usually develop within a hyperplastic epithelium. Conversely, type 2 tumors are less responsive to estrogens and develop within an atrophic background.
Obesity increases the risk of type 1 tumors by roughly three-fold and almost two-fold for type 2 tumors. Metabolic syndrome also doubles the risk of developing endometrial cancer in both pre-and postmenopausal women, most likely due to estrogen independent activation of the PI3K pathway. Notably, obesity associations have not been clearly proven for cervical, ovarian, vaginal, or vulvar cancers. 32,33 Possibly, the inclusion of some histologies in which obesity did not exert a role during earlier stages of carcinogenesis may mask this relationship. More recent studies analyzing specific histologies in these cancers have begun to establish this relationship with obesity. In cervical cancer, an increase in BMI has been associated with a higher risk of developing cervical adenocarcinoma. 34,35 Additionally, persistent cervical infection by high-risk HPV strains is favored in obese women presenting vaginal dysbiosis, which is characterized by an increase in microbial diversity that prompts malignant transformation of the cervical epithelium. [36][37][38][39] For epithelial ovarian cancer, a disease including at least five histological subtypes and arising from fallopian tube fimbria, ovarian, and peritoneal surfaces, most recent studies have proven a relationship between increasing BMI and nonserous histologies. 40,41 For serous histologies, obesity would increase the risk of those arising from the peritoneum. 42 Summarizing the effects of weight gain on the development of cancer, a recent meta-analysis addressed the impact of 5 kg weight gain on the relative cancer risk in adults. 43 Results showed that the overall relative risk of 5 kg weight gain was 1.11. Specifically for gynecologic cancers in postmenopausal women, the relative risks were 1.

| EFFEC TS OF OB E S IT Y ON THE D IAG NOS IS OF GYNECOLOG IC C AN CER S
Obesity is related to diagnosis at earlier ages only in endometrial cancer. 45 Obesity seems to affect clinical presentation of symptoms, contributing to a delayed diagnosis in ovarian cancer. 46 Roughly three-quarters of all ovarian cancer patients are diagnosed at late stages of the disease, with a 5-year survival rate of 50%. 20 For cervical cancer, obese and morbidly obese status entail a higher rate of defective screening, inadequate clinical assessment, and higher risk of missing hidden or partially visible lesions. 47 In fact, obese patients are often diagnosed at advanced stages of the disease despite regular examination and are twice more likely to develop cervical cancer than lean patients. 20 Beyond tumor biology, obesity constitutes a morbid condition with sociocultural implications which is more prevalent among underprivileged (low income and less educated) communities. Lack of knowledge and limitations in healthcare access undoubtably contribute to delayed diagnosis and subsequent management of gynecologic cancers.

| MANAG EMENT OF GYNECOLOG IC C AN CER S AMONG OB E S E AND MORB IDLY OB E S E WOMEN
Obesity does not commonly make it difficult to obtain an adequate biopsy or perform image staging (i.e. MRI, CT, or PET/CT scan) of gynecologic cancers. Once confirmed and if the extension of the disease allows it, the first option will be the complete surgical removal of the tumor coupled with adequate staging. However, in some cases, the presence of obesity may cause the medical team to opt for nonsurgical management or may determine suboptimal surgery or a more complex perioperative management scenario (i.e. high-cost instrumentation/appliances; use of robotic rather than laparoscopic surgery or laparotomy; higher rates of intraoperative, immediate postoperative and 30-day complications; and hospital readmissions), delaying adjuvant therapies, among other possibilities affecting the prognosis. In this respect, Inci et al. 48 have recently identified overweight and obesity as significant predictors of postoperative complications. Similarly, Pyrzak et al. 49 identified a BMI of 30 or higher as a risk factor for complication-related 30-day hospital readmission. It is well known that obesity entails a higher risk of thromboembolism and wound infection, particularly for open, staging, or cytoreductive and time-extended surgeries. [50][51][52][53] With respect to high-grade serous ovarian cancer (HGSOC), achieving optimal debulking, primarily or after neoadjuvant chemotherapy, constitutes an independent and relevant prognosis factor. 54 Recently, Wang et al. 55 have identified five molecular subtypes of HGSOC. One of them, the mesenchymal subtype, is the less optimally debulked and exhibits the poorest outcome. Our group has recently found that high leptin levels as seen among obese women induce EMT in HGSOC cell lines. 56 In addition, we have demonstrated that HGSOCs overexpressing obesity and lipid metabolismrelated genes share significant lower chances of achieving optimal debulking and having positive outcomes. 57 This group is enriched for the mesenchymal subtype. 58 Radiotherapy either alone or in combination with chemotherapy constitutes the primary treatment for those gynecologic cancers not suitable for surgery or those locally advanced (i.e. uterine cervix cancer). It is also indicated as adjuvant therapy in cases harboring risk factors for local recurrence and as part of palliative care. Obesity, particularly extremely morbid obesity, poses a major hindrance to treatment planning, limits the use of common radiotherapy equipment (commonly designed to support certain BMI range and physical characteristics of individuals), and increases the risk of gynecologic and cutaneous radiation-related toxicities. 59,60 Chemotherapy administration and efficacy are also challenged by obesity. Obese cancer patients receiving chemotherapy have worse clinical outcomes. Potential explanations for these adverse results include differences in pharmacokinetics, metabolic dysregulation, induction of chemoresistance, or clinicians' decisions to reduce dose intensity during treatment to minimize toxicities. 61,62 Since 2012, American Society of Clinical Oncology (ASCO) guidelines recommend using actual body weight for dosing in all patients treated with curative intent, irrespective of obesity, to avoid compromising clinical outcomes. Outcomes in obese patients are no different to lean patients when the correct dose is administered. 63

| CLINIC AL OUTCOME S AMONG OB E S E GYNECOLOG I C AL C AN CER PATIENTS
Obesity appears to have a negative impact for all gynecologic cancers regarding prognosis and treatment outcomes. 32 In endometrial cancer, Donkers et al. 64   Our group demonstrated that obese women with HGSOC have poorer progression-free and overall survival compared with the lean counterpart. 56 We also demonstrated in two international cohorts (The Cancer Genome Atlas and Australian Ovarian Cancer Study) that HGSOC overexpressing obesity and lipid metabolism-related genes have poorer oncologic outcomes. 57 Obesity also impacts on secondary cytoreductive surgery and overall survival in women with recurrent disease. 68 In relation to vulvar cancer, obesity was associated with a shorter time to recurrence in the AGO-CaRE-1 study and this was mainly attributed to a higher risk of local recurrence. 69

| NEG ATIVE EFFEC TS OF OB E S IT Y ON OVER ALL SURVIVAL OF GYNECOLOG IC C AN CER SURVIVOR S: RELE VAN CE OF LIFE S T YLE CHANG E S
Perhaps the greatest health threat among gynecologic cancer survivors is weight gain over time or persistence of obesity after treatment completion. 70 Studies demonstrated that women with endometrial cancer have significantly higher risk of mortality from other obesitydriven diseases, such as heart disease or type 2 diabetes, compared with women without cancer. 70,71 A prospective report demonstrated that morbid obesity is associated with a significantly increased risk of death from several women's cancers. For women with a BMI of 40 or higher, the relative risk (RR) is 1.62 (95% CI, 1.40-1.87) and for BMI 35-39.9, the RR is 1.32 (95% CI, 1.20-1.44). 13 Evidence suggests that weight management and physical activity improve overall health and well-being and reduce the risk of morbidity and mortality among cancer survivors. 72 Weight loss after bariatric surgery is more sustained than after other interventions and is protective against endometrial cancer. 6 ASCO is highly committed to reducing the impact of obesity on cancer and the establishment of a multipronged initiative to accomplish this goal. Such an initiative considers: (1) increasing education and awareness of the evidence linking obesity and cancer; (2) providing tools and resources to help oncology providers address obesity with their patients; (3) building and fostering a robust research agenda to better understand the pathophysiology of energetic balance alterations, evaluate the impact of behavioral changes on cancer outcomes, and determine the best methods to aid cancer survivors in the implementation of effective and useful modifications to lifestyle and behavior; and (4) advocating for policy and systems change to address societal factors contributing to obesity and improve access to weight management services for cancer patients. 73

| ROLE OF OB E S IT Y IN COND ITI ONING RE S P ON S E OF FUTURE THER APEUTIC VEN UE S IN GYNECOLOG I C C AN CER S
The current therapeutic scenario is moving to the precision medicine. Beyond choosing the most effective surgery, radiotherapy, and/or chemotherapy scheme for any cancer, hopes are pinned on identifying cancer weaknesses, designing targeted therapies, and enhancing the host's antitumor immune response to improve clinical outcomes. However, the promising responses observed in preclinical models with some targeted therapies (i.e. immunotherapies) have not translated in the same results when challenged in gynecologic cancer patients. 74 Factors contributing to impair their efficacy are aging, the composition of gut microbiome, and obesity. 75 The additive effects of increased conversion of androgens into estradiol and estrone by peripheral hypertrophic adipocytes, increased bioactive estrogens, and increased insulin signaling in insulin-resistant obese patients, which converge into the mitogenicPI3K/AKT/mTOR signaling pathway, are also to be taken into account. As such, addition of metformin and cholesterol-lowering statins 76  A recent study on single-cell RNA sequencing and cell-cell ligandreceptor interactome revealed that mature natural killer cells are depleted in the adipose tissue of obese compared with lean patients, and negatively correlated with patient BMI, with a relative increase of immature natural killer and tissue-resident natural killer cells. 77 These and other developing technologies, including lipidomics approaches, 21 will continue to provide a detailed and unbiased cellular landscape of homeostatic and dysregulated circuits to further our understanding of health and disease, including obesity-related disorders.

AUTH O R CO NTR I B UTI O N S
IW and MC shared the concept design, literature review, and writing of the manuscript.

CO N FLI C T S O F I NTE R E S T
Relating to the submitted work, MC received a grant from Fondecyt nº 1201083. IW has no conflicts of interest to declare.