Obesity and Breast Cancer Risk: The Oncogenic Implications of Metabolic Dysregulation

Abstract

Context

Breast cancer is increasing in prevalence in parallel with rising rates of obesity worldwide. Obesity is recognized as a leading modifiable risk factor for the development of breast cancer; however, this association varies considerably by clinicopathologic features, and the underlying mechanisms are complex.

Evidence Acquisition

Pubmed literature search using combinations of “obesity,” “breast cancer risk,” “diet,” “exercise,” “weight gain,” “weight loss,” “adipose tissue inflammation,” “crown-like structure,” “immune markers,” “metformin,” “gliflozins,” “SGLT-2i,” “GLP1-RA,” and related terms.

Evidence Synthesis

Elevated body mass index and weight gain are associated with increased risk of postmenopausal, hormone receptor–positive breast cancer. Emerging evidence suggests that adverse measures of body composition in individuals of any weight can also confer increased breast cancer risk. Mechanistically, various factors including altered adipokine balance, dysfunctional adipose tissue, dysregulated insulin signaling, and chronic inflammation contribute to tumorigenesis. Weight loss and more specifically fat mass loss through lifestyle and pharmacologic interventions improve serum metabolic and inflammatory markers, sex hormone levels, and measures of breast density, suggesting a link to decreased breast cancer risk.

Conclusion

Incorporating markers of metabolic health and body composition measures with body mass index can capture breast cancer risk more comprehensively. Further studies of interventions targeting body fat levels are needed to curb the growing prevalence of obesity-related cancer.

Breast cancer (BC) is the most common malignancy globally, accounting for over 2.3 million new diagnoses yearly (1). Simultaneously, the prevalence of obesity is rising at a dramatic rate projected to affect nearly 50% of adults in the United States by 2030 (2). Obesity is a well-established risk factor for postmenopausal, hormone receptor–positive BC incidence (3-5). More recently, weight gain and high body fat levels, even among individuals with normal body mass index (BMI), have been identified as additional BC risk factors. In this review, we present epidemiologic data linking obesity, weight gain, and visceral adiposity to BC risk, discuss the pathophysiologic mechanisms of body fat as a chronic carcinogen, and provide an overview of the effect of lifestyle and pharmacologic interventions on body composition and BC risk. These data support a shift in paradigm from focusing on weight-based strategies to targeting body composition in reducing BC risk.

Materials and Methods

We conducted a PubMed search using variants of the following keywords: “obesity,” “breast cancer risk,” “weight gain,” “weight loss,” “adipose tissue inflammation,” “crown-like structure,” “immune markers,” “diet,” “exercise,” “metformin,” “gliflozins,” “SGLT-2i,” and “GLP1-RA.” We focused on observational studies and clinical trials published from 1980 to present. Given the large amount of material retrieved, we hand-selected the articles that were most relevant from the literature, with priority given to more recent articles and studies with large sample sizes. We reviewed reference lists of identified articles to find additional publications. For the summary Tables 1 to 3, we included only studies with more than 10 000 subjects. We excluded case reports and case series.

Obesity and Breast Cancer

Relation Between Body Mass Index and Breast Cancer Risk

A clear and consistent relationship exists between elevated BMI and increased postmenopausal BC risk. In several large cohort studies inclusive of various cancer types, high BMI confers an elevated BC risk among postmenopausal women with risk ratios (RRs) or hazard ratios (HRs) ranging from 1.08 to 1.41 (Table 1) (6-8). Among the numerous studies that have focused on incident postmenopausal BC risk regardless of subtype, HRs ranged from 1.31 to 2.13 when comparing BMI of ≥30 kg/m² with normal BMI (Table 2) (3, 9-15). Duration of obesity has also been examined. In an observational study of the Women’s Health Initiative, increased obesity duration per 10 years was associated with increased postmenopausal BC risk in a multivariable model (HR 1.07, 95% CI 1.04-1.10) (16). In an Icelandic study of 88 809 women, each 0.1 kg/m² BMI increase per year was associated with increased postmenopausal BC risk (HR 1.09, 95% CI 1.04-1.13) and cumulative obesity exposure was associated with increased BC risk (17). A “critical window” of elevated BMI that confers the greatest postmenopausal BC risk has not been identified, although studies have shown that obesity both 2 and 6 years prior to diagnosis are associated with increased BC risk, and the association becomes stronger with increasing age following menopause (18, 19).

Table 1.

Selected large observational studies of body mass index and risk of incident obesity-related cancers, including postmenopausal breast cancer

First author (ref. year)	N	BMI (kg/m2)	RR or HR (95% CI)
Parra-Soto (6) (2021)	47 882	Per 1 SD increment	RR 1.08 (1.05-1.11)
Cao (7) (2020)	380 575	≥30 vs 18.5-25	HR 1.29 (1.16-1.43)
Kabat (8) (2015(20))	143 901	Highest vs lowest quintile	HR 1.41 (1.31-1.53)
Reeves (20) (2007)	1 222 630	≥30 vs 22.5-24.9	RR 1.40 (1.31-1.49)

Table 2.

Selected large observational studies of body mass index and risk of postmenopausal breast cancer

First author (ref., year)	N	BMI (kg/m2)	OR, RR or HR (95% CI)
Park (9) (2021)	6 467 388	≥30 vs 18.5-23	HR 1.54 (1.47-1.62)
Moy (10) (2018)	17 781	Linear per kg/m2	HR 1.04 (1.02-1.07)
Lee (11) (2017)	11 227 948	≥30 vs 18.5-22.9	HR 1.31 (1.26-1.39)
Kabat (12) (2017)	20 944	≥30 vs 18.5-24.9	HR 1.51 (1.28-1.78)
Engmann (13) (2017)	18 437	≥35 vs 18.5-24.9	OR 1.54 (1.45-1.64)
Neuhouser (3) (2015)	67 142	≥35 vs <25	HR 1.58 (1.40-1.79)
Suzuki (14) (2013)	36 164	≥29 vs <24	HR 2.13 (1.09-4.16)

The relationship between obesity and premenopausal BC risk is less clear; several studies have shown a protective effect. Four large cohort studies have shown that premenopausal women with BMI ≥30 kg/m² have decreased BC risk compared with women with normal BMI, with RR or HR ranging from 0.62 to 0.91 (9, 11, 20, 21). Median age at the time of enrollment and BMI measurement in these studies ranged from 37.4 to 43.7 years; women were then followed over time for new BC diagnoses that occurred in the premenopausal period. Two meta-analyses confirmed these results. A meta-analysis of the effect of BMI and risk of 23 cancer types showed that for every 5 kg/m² of increased BMI premenopausal BC risk decreased (RR 0.95, 95% CI 0.92-0.98) (22). Another meta-analysis of all studies on obesity and premenopausal BC from 2000 to 2010 found that each 5 kg/m² of increased BMI was inversely associated with BC risk (RR 0.95, 95% CI 0.94-0.97) (23). Increased body fatness in childhood (ages 5-10 years) and adolescence (ages 10-20 years) are also inversely associated with premenopausal BC risk (RR 0.48, 95% CI 0.35-0.55 and RR 0.57, 95% CI 0.39-0.83 respectively) (24, 25). On the other hand, an analysis of 12 243 women in the Breast Cancer Prevention Trial demonstrated increased BC risk for premenopausal women with BMI ≥30 kg/m² compared with BMI <25 kg/m² (HR 1.70, 95% CI 1.10-2.63) (26).

Several hypotheses for the association between obesity and lower premenopausal BC risk have been proposed, although evidence remains limited. Obese premenopausal women have increased frequency of anovulatory menstrual cycles, which has been suggested to lower premenopausal BC risk due to decreased cumulative luteal phase exposure (27). Estrogen and progesterone levels are both higher during the luteal phase, when breast epithelial cell proliferation is also accelerated (28, 29). The combination of exogenous estrogen plus progesterone is associated with increased postmenopausal breast cancer risk compared with estrogen alone (30-33). These observations have been used to support the hypothesis that less exposure to estrogen plus progesterone due to anovulation contributes to decreased premenopausal breast cancer risk. However, observational cohort studies have not confirmed an association between anovulation or irregular menstrual cycles and premenopausal breast cancer risk (34, 35).

Current hormone replacement therapy (HRT) use may modify the association between BMI and BC. In a study of 110 698 women participating in the European EPIC study, those with BMI in the highest vs lowest quintile had increased BC risk only if they were not currently using HRT (RR 1.36, 95% CI 1.06-1.75), whereas there was a trend toward decreased risk in current HRT users (RR 0.71, 95% CI 0.50-1.01) (36). Among 85 917 postmenopausal women in the Women’s Health Initiative Observational study, HRT nonusers with BMI >31.1 vs <22.6 had higher risk of BC (RR 2.52, 95% CI 1.62-3.93), but this association was not seen in women who had ever used HRT (37). In a cohort of 95 256 nurses in the United States, postmenopausal women who had never used hormone replacement therapy had a higher risk of BC for BMI >31 vs ≤20 (RR 1.59, 95% CI 1.09-2.32) (38). The use of HRT artificially elevates circulating estrogen levels regardless of BMI, such that the attributable increased risk of BC related to excess adipose tissue and peripheral conversion of androgens to estrogens in obese women is likely abolished; thus HRT use significantly attenuates the relationship between BMI and BC risk (39).

There are differences in the degree of risk conferred by obesity depending on BC subtype. Histopathologic analyses of various immunohistochemical characteristics have suggested that elevated BMI is associated with risk of less aggressive tumor types, namely hormone receptor–positive, HER2-negative, and low Ki67 (HR per 5 kg/m² 1.44, 95% CI 1.10-1.90) (4). Among pre/perimenopausal women, elevated BMI is more strongly associated with estrogen receptor (ER)–negative cancer than ER-positive cancer (40, 41). Among postmenopausal women, obesity is more strongly associated with hormone receptor–positive cancers (3, 5). Whether obesity in postmenopausal women affects risk of triple negative BC (TNBC) is unclear as studies have yielded conflicting results. An analysis of the Women’s Health Initiative trial and a small cohort study of 176 patients showed no association (3, 42). A case–case study of 2659 postmenopausal women showed a decreased risk of TNBC (OR 0.74, 95% CI 0.54-1.00), and in a large Swedish cohort of 51 823 postmenopausal women BMI was inversely associated with development of progesterone receptor–negative tumors (RR 0.68, 95% CI 0.47-0.98) (5, 41). In a pooled analysis of 2 population-based case–control studies of postmenopausal women, there was a trend toward increased risk of TNBC (OR 2.7, 95% CI 1.0-7.5) in women in the highest quartile of BMI (43). Thus, there is a clear relationship between obesity and postmenopausal hormone receptor–positive BC, but the effect of obesity on risk of TNBC is unclear; heterogeneity between study populations, sample sizes, and tumor subtype classifications may explain some of these discrepant results.

Weight Gain as a Risk Factor for Breast Cancer

There is a clear relationship between weight gain in adulthood and increased risk of postmenopausal BC (Table 3), but the timing during which weight gain exerts this effect is unclear (44-51). In a Chinese population, weight gain after menopause appeared to have a greater effect on BC risk than weight gain from age 18 to menopause (52). In a cohort of women in Norway the opposite trend was observed: weight gain prior to or around menopause was associated with increased risk of postmenopausal BC but postmenopausal weight gain was not (49). In the United States Framingham cohort, there was no significant association between weight gain during specific periods of adulthood (25-44 years, 45-55 years, or after 55 years) with BC risk (51).

Table 3.

Selected large observational studies of weight gain and risk of breast cancer

First author (ref., year)	N	Weight change	RR or HR (95% CI)
Han (44) (2014)	13 901	5% increase from age 25 to enrollment	HR 1.05 (1.02-1.07)
Catsburg (45) (2014)	39 532	5 kg increase since age 20	HR 1.06 (1.01-1.11)
Suzuki (46) (2017)	30 109	BMI increase ≥5 vs stable BMI	HR 1.90 (1.20-3.01)
Welti (47) (2017)	80 943	Weight gain vs stable weight	HR 1.11 (1.03-1.20)
Rosner (48) (2015)	77 232	4-years ≥6.8 kg weight gain vs no change (≤2.3 kg)	RR 1.20 (1.09-1.33)
Alsaker (49) (2013)	28 153	Per kg per year	HR 1.38 (1.09-1.75)

Mirroring the protective effect of obesity for premenopausal BC, weight gain does not seem to significantly increase the risk of premenopausal BC and may even be protective as well. No significant association between premenopausal weight gain and premenopausal BC risk was seen in a large prospective observational study among women in the Nurses’ Health Study; this was corroborated by a meta-analysis of 50 studies (53, 54). However, an analysis of premenopausal patients in the Nurses’ Health Study and Nurses’ Health Study II reported that both weight gain and weight loss in adulthood may decrease risk of premenopausal BC but did not reach statistical significance (≥5 kg loss, HR 0.75 95% CI 0.52-1.09; ≥25 kg gain, HR 0.78, 95% CI 0.55-1.11) with P_trend = .08 (55). A case–control study showed a relation between significant weight gain with BMI increase of ≥10 kg/m² and increased risk of TNBC (56).

Weight gain in early adulthood is associated with earlier BC onset. In a study of 660 BC survivors, the highest amount of BMI gain (0.75-0.99 mg/m² BMI increase/year) was associated with breast cancer diagnosis 5 years earlier compared with the lowest BMI change (0.14-0.16 mg/m² BMI increase/year); this association was significant for both premenopausal and postmenopausal BC diagnoses (57). Weight cycling during early to mid-adulthood has an unclear effect; there was a trend toward increased risk with cycling 1 to 3 times, but a statistically significant protective effect with weight cycling >10 times after adjustment for BMI in addition to various other factors (HR 0.82, 95% CI 0.69-0.97) (47). A case–control study did not show effect of temporary weight cycling on risk (58).

HRT use may influence the association between weight gain and BC risk as well. Among 62 756 postmenopausal women in the Cancer Prevention Study-II Nutrition Cohort, weight gain of 9.5 to 13.6 kg was associated with increased risk of BC (RR 1.4, 95% CI 1.1-1.8) among former or never HRT users, but no association was seen among current HRT users (59). Among 95 256 postmenopausal female nurses in the United States, weight gain of more than 20 kg compared with unchanged weight increased the risk of BC among HRT never-users (RR 1.99, 95% CI 1.43-2.76) (38). In a cohort of 180 885 women age ≥50 in the Pooling Project of Prospective Studies of Diet and Cancer, the risk reduction of weight loss was specific to women who were not using HRT (≥9 kg lost vs stable weight, HR 0.68, 95% CI 0.50-0.93) (60). A meta-analysis demonstrated that for every 5 kg of weight gain during adulthood, the risk of postmenopausal BC was RR 1.11, 95% CI 1.08-1.13 among no- or low-HRT users; no relation was seen in HRT users (RR 1.01, 95% CI 0.99-1.02) (54).

Data regarding weight change and risk of subtypes of BC are conflicting. Two large cohort studies have shown an association between weight gain and TNBC risk: a study of 61 335 postmenopausal women demonstrated that weight gain of ≥5% of total body weight increased TNBC incidence (HR 1.54, 95% CI 1.16-2.05), and a study of 77 232 women demonstrated that premenopausal short-term weight gain was more strongly associated with TNBC than ER-positive/progesterone receptor–positive BC (RR per 11.3 kg weight gain: 1.51, 95% CI 1.09-2.38; 1.13, 95% CI 0.89-1.43 respectively) (48, 61). The same authors demonstrated in a separate study of 74 177 women from the Nurses’ Health Study that weight gain since age 18 is associated with hormone receptor–positive postmenopausal BC (per 30 kg, HR 1.50, 95% CI 1.36-1.65) but not hormone receptor–negative (HR 1.16, 95% CI 0.95-1.42) (53). A meta-analysis showed that for the highest vs lowest categories of weight gain in adulthood, there was an increased risk of TNBC (risk estimate 1.34, 95% CI 1.06-1.63) (62). In summary, weight gain in adulthood is associated with increased postmenopausal BC risk; however, further studies that better elucidate the effect of weight cycling and timing of weight gain on BC risk are needed. Such data could identify optimal windows for risk-reducing interventions.

Body Composition in Normal-weight Individuals

More recently, several studies have examined measures beyond weight and BMI, incorporating metabolic health and body composition to capture BC risk more comprehensively. This has revealed that a substantial proportion of women with normal weight can have increased BC risk due to adverse cardiometabolic health or excess adiposity. Park et al categorized 50 884 individuals from the Sister Study by metabolic health and obesity phenotype using the components associated with metabolic syndrome (central obesity, elevated blood pressure, type 2 diabetes, and dyslipidemia); women were considered metabolically unhealthy if they had any 1 of these cardiometabolic abnormalities, regardless of weight (63). Postmenopausal metabolically unhealthy but normal-weight women had higher BC risk (HR 1.26, 95% CI 1.01-1.56) compared with metabolically healthy normal-weight women. In a study of 3460 postmenopausal women with normal BMI enrolled in the Women’s Health Initiative, those in the highest 2 quartiles of whole body and visceral adiposity had significantly increased BC risk (whole body fat mass ≥22.1 kg: HR 1.43, 95% CI 1.06-1.93; whole body fat percentage ≥38.0: HR 1.45, 95% CI 1.07-1.95; trunk fat mass ≥9.4 kg: HR 1.50, 95% CI 1.12-2.03) (64). Among normal-weight postmenopausal women in the UK Biobank, those in the highest vs lowest quartile of body fat mass index, percent body fat, trunk fat mass index, or trunk fat mass percent had increased risk of BC (65). Together, these studies demonstrate that use of BMI alone to estimate BC risk results in exclusion of a substantial proportion of women with normal BMI but excess adiposity who would benefit from intervention to reduce their elevated risk. Future studies should incorporate metabolic health and body composition measures.

Mechanisms Underlying Adiposity and Carcinogenesis

Adipose Dysfunction

Reframing adiposity from an inert lipid reservoir to an active endocrine organ capable of exerting pleiotropic biologic effects (66) has expanded our understanding of the complex etiologic role obesity plays in BC. Adipose tissue can be classified into 2 functionally distinct compartments, brown adipose tissue and white adipose tissue, based on location, composition and endocrine function (67). Regulated by insulin, white adipose tissue increases glucose uptake transforming it into fat storage. Visceral adipose tissue, in comparison to subcutaneous adipose tissue is more lipolytically active (68) and contributes to higher levels of free fatty acids (FFAs) (69, 70)—a common finding in obesity (69)—which predisposes to hepatic and peripheral insulin resistance (71). Chronically elevated FFAs activate Inhibitor of NF-kB (IkB)/Nuclear Factor-kappa B (NFkB) pathways (72), an important molecular lynchpin that links chronic inflammation with increased cancer risk (73).

Metabolic health entails a balance between insulin/insulin-like growth factor (IGF)-1 and growth hormone (GH) (74). GH stimulates IGF-1 secretion from the liver; in turn, IGF-1 and FFAs, suppress GH secretion via negative feedback to the hypothalamic–pituitary axis. GH exerts a balance of direct and indirect effects on glucose metabolism, including increasing hepatic glucose output, reducing insulin sensitivity, upregulating lipolysis (and increasing FFAs), and reducing visceral adiposity (75, 76). GH exerts metabolic effects that are antagonistic to insulin thus leading to insulin resistance (76). States of energy abundance, as in overweight and obesity, are, furthermore, associated with increased adipocyte lipolysis, which leads to high circulating levels of FFAs, ectopic fat deposition, and hyperinsulinemia, which promote further energy storage (77, 78) (Fig. 1). To accommodate increased energy availability, adipose tissue responds by adipocyte cell hypertrophy or recruitment of new adipocytes from a progenitor reservoir pool (70). Notably, in obesity, adipocyte hypertrophy correlates with dyslipidemia, hyperinsulinemia, and higher inflammatory markers (79). The capacity for adipose tissue to hypertrophy is finite and varies between individuals (80), such that once adipocyte tissue expansion has reached a ceiling, lipids can no longer be safely cleared from the systemic circulation resulting in unchecked high FFA levels exerting myriad lipotoxic effects (81) (Fig. 2).

Figure 1.

Metabolic changes in obesity. High circulating levels of free fatty acids and glucose contribute to peripheral insulin resistance resulting in hyperinsulinemia and high IGF-1, both acting as growth factors, and decreased SHBG, which increases free estrogen.

Figure 2.

Inflammatory changes associated with weight gain. (A) In obesity, adipocyte hyperplasia and hypertrophy occur. (B) States of energy balance are associated with an anti-inflammatory cytokine milieu with higher adiponectin to leptin ratios. In energy abundance, adipocyte hypertrophy eventuates cell stress leading to infiltration of CD-68 + macrophages creating a proinflammatory microenvironment and upregulation of androgen aromatization

Adipokine Signaling

The fat pad synthesizes adipokines; these cell-signaling molecules regulate integration of systemic metabolism and immune function. Adipokines may originate from general body fat, within a tumor capsule, and BC cells themselves, capable of cross-talk. Adipokines exert both proinflammatory and anti-inflammatory effects; the balance determines the overall level of local and systemic inflammation. The 2 most studied adipokines implicated in obesity-related BC are adiponectin and leptin, which exert opposite effects on subclinical inflammation.

Adiponectin

Adiponectin, a generally anti-inflammatory adipokine, plays a key role in glucose and lipid metabolism by enhancing insulin sensitization (66, 82, 83). It also downregulates the expression and release of proinflammatory mediators and regulates key immune responses critical in breast tumorigenesis, including suppression of aberrant cell growth and inhibition of cell invasion (84-86). Adiponectin levels are negatively correlated with obesity and hyperinsulinemia (87), and reductions in body fat mass are accompanied by elevations in adiponectin levels. Epidemiologic studies consistently report a significant inverse relationship between adiponectin levels and postmenopause BC risk (88-90). A similar finding emerged in a study of premenopausal women where baseline adiponectin levels predicted new BC events (91). Moreover, a meta-analysis of 31 studies concluded that low adiponectin levels may be associated with higher BC risk, irrespective of age (92).

Adiponectin exerts its protective antineoplastic effects through 2 main pathways (93). The first is a direct pathway by activating adenosine monophosphate–activated protein kinase (AMPK), a crucial protein regulator responsible for cellular adaptation during inflammatory states and oxidative stress. During stress, AMPK downregulates anabolic and proliferative mechanisms (94), and specifically halts cancer development by inducing key molecules such as P53 and P21 (95), important in cell cycle arrest and apoptosis, respectively. Treatment with adiponectin was shown to inhibit breast tumor growth via stimulation of AMPK (84, 95), as well other hormonally sensitive cancers (96). Adiponectin also exert paracrine effects between adipocytes and adjacent BC cells where it can downregulate aromatase activity (97), lowering estrogen production and attenuating ER stimulation of nearby BC cells, constituting a small part of a complex system of bidirectional crosstalk between adipokines and numerous inflammatory cytokines. The second antineoplastic pathway involves phosphatidylinositol 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) inhibition (96, 98), thereby reducing tumor cell growth. Through anti-inflammatory action, adiponectin may further play a role in mitigating the increased risk of cancer associated with obesity-induced low-grade inflammation (99).

Leptin

A pleiotropic adipokine that upregulates proinflammatory cytokines including tumor necrosis factor-α and interleukin (IL)-6, leptin is synthesized by adipose tissue and breast cells (100, 101). Leptin regulates energy homeostasis and adipose tissue growth (102), and contributes to the chronic low-grade inflammation associated with obesity (103). Plasma leptin levels are abnormally elevated in obese women (104) and are positively correlated with insulin resistance, independent of weight and BMI (105, 106). Elevated leptin and increased risk of BC have been reported even after adjustment for BMI (107). High levels of leptin are associated with more aggressive BC in some epidemiologic studies (108), where its use as a potential prognostic biomarker has been suggested (109).

Leptin receptors have been found on BC cells (110, 111). The pro-oncogenic effects of leptin in BC have been attributed to its mitogenic stimulatory effects (112) and downregulation of apoptosis (113, 114). Leptin further mediates interactions between BC cells and the surrounding microenvironment by engaging in cross-talk (115) and stimulating production of IL-8 and IL-18, which support tumor cell survival and proliferation (116).

Insulin regulates new leptin synthesis from intracellular storage pools through post-transcriptional mechanisms mediated by Janus-activated kinases, JAK/STAT3, mitogen-activated protein kinase (MAPK), and PI3K/Akt signaling pathways (117, 118). Insulin can also induce overexpression of leptin and its receptor on BC cells via the same pathways (110, 119). The JAK/STAT signaling pathway is a key regulatory player in proliferation, survival, and apoptosis across different tissues (120) and plays a critical role in obesity-related BC (121). Leptin has been shown to promote development and progression of BC cells (122) via JAK2/STAT3 pathway in addition to upregulating aromatase activity (118).

Dysregulated Insulin Signaling Pathways

Insulin exerts its effects by binding to 1 of 2 ubiquitously expressed receptor isoforms, IR-A and IR-B, with IR-A exhibiting particularly high mitogenic activity (123, 124).

Insulin signaling activates 2 major pathways, the PI3K pathway and MAPK. Glucose metabolism is mediated via PI3K, which results in downstream activation of AKT/mTOR, an intracellular signaling pathway that plays a pivotal role in regulation of cell survival, growth, and proliferation. In hyperinsulinemia, the IRs are chronically activated, which in the setting of genetic mutations leads to increased cellular glucose uptake, reduction of apoptosis, and increased cellular proliferation.

The MAPK, the second major channel of insulin signaling, is activated independently of the PI3K pathway, which mediates the mitogenic effects of insulin via Ras, a group of intracellular proteins that govern cell growth, angiogenesis, and immune escape. There is growing evidence that Ras hyper-signaling plays a role in BC progression (125) and reduced patient survival (126). In hyperinsulinemia, the stimulatory effect of insulin on the PI3K pathway is blunted, but an enhanced activation of MAPK pathway is observed with an increase in insulin-induced activation of Ras protein (127, 128).

Chronic Inflammation

Hyperinsulinemia induces adipocyte hypertrophy. Adipocyte pools eventually outgrow their blood supply creating hypoxic conditions leading to cell death (129). Hypoxia is one of the initial changes that takes place in diseased adiposity where Hypoxia-inducible factor-1) is activated (130), upregulating leptin and vascular endothelial growth factor favoring a protumorigenic environment. Dying adipocytes are scavenged by macrophages in a histologically recognizable pattern known as crown-like structures (CLS) (131, 132). These CLS-associated macrophages produce a cytokine profile that polarizes the immune microenvironment toward a proinflammatory state, which further contributes to a vicious cycle of deranged adipose gene expression and systemic insulin resistance, as demonstrated by inflammatory cytokine overproduction in CLS present in abdominal adipose beds.

As with other adipose reservoirs, CLS have been isolated in breast tissue and used histologically to assess the severity of local inflammation; with studies consistently showing that CLS favor a proinflammatory breast and increased aromatization activity. Following menopause, the main site of estrogen production shifts to adipose tissue via conversion of circulating steroids. Hyperinsulinemia decreases sex hormone–binding globulin (SHBG), increasing bioavailable sex steroids in the circulation. Adipose tissue houses high levels of aromatase (CYP19A1), which through androgen aromatization synthesize estrogens that can act locally in a paracrine fashion or be released into the circulation. The relative contribution of adipose tissue to steroid metabolism is not trivial and a positive association has been found between BMI and tissue exposure levels to estrogen. Particularly prominent in menopause is the positive relationship observed between fat mass, aromatase expression, and overall higher estrogen levels.

Cancer is a complex and multistep process. The prevailing model linking obesity to hormonally driven BC implicates chronic adipose dysfunction where metabolic (hyperinsulinemia) and endocrine (increased ratio of leptin to adiponectin; upregulated aromatase activity) factors intersect with dysregulated signaling pathways creating an inflammatory state that favors a tumorigenic environment.

Clinical Implications and Targets for Interventions

Weight Loss

Weight loss is associated with BC risk reduction. The Iowa Women’s Health study included 33 660 postmenopausal women which assessed weight change at 4 periods of adulthood showed that weight loss from age 30 to menopause or after menopause led to risk reduction for postmenopausal BC (RR 0.36, 95% CI 0.22-0.60; RR 0.48, 95% CI 0.22-0.65, respectively) (133). Among 180 885 women in the Pooling Project of Prospective Studies of Diet and Cancer, weight loss was linearly associated with reduced BC risk (>2-4.5 kg lost: HR 0.82, 95% CI 0.70-0.96; >4.5-<9 kg lost: HR 0.75, 95% CI 0.63-0.90; ≥9 kg lost: HR 0.68, 95% CI 0.50-0.93) (60). Finally, an analysis of 61 335 women in the Women’s Health Initiative Observational Study showed that women with weight loss had lower BC risk with no interaction by BMI (HR 0.88, 95% CI 0.78-0.98) (61). In women who carry mutations in BRCA1 or BRCA2, weight loss of at least 4.5 kg between the ages of 18 and 30 appears to be protective against early-onset BRCA-associated breast cancers (134). Weight loss via bariatric surgery has been associated with decreased BC risk in both premenopausal and postmenopausal women (135-137).

Lifestyle Modifications: Diet and Exercise

Various studies have assessed the effect of diet and exercise on serum metabolic and inflammatory markers, body fat levels, and mammographic breast density. The SHAPE-2 study was a 4-month, 3-arm randomized controlled trial (RCT) of diet, exercise plus diet, or standard care in postmenopausal women (138). The group receiving combined diet and exercise intervention experienced significant reductions in body weight, and intra-abdominal and subcutaneous fat (139). Exercise, more so than diet, resulted in decreased C-reactive protein, leptin, and sex hormone levels, but this effect was attenuated when adjusting for change in body fat percentage (138, 140). The NEW trial randomly assigned postmenopausal women to a reduced-calorie diet, exercise, or combination, and found that intervention-induced weight loss resulted in lower serum estrogens, free testosterone, and increased SHBG (141). The Italian DAMA study randomized postmenopausal women to a diet that was either plant based, low glycemic load, low in saturated fats and alcohol, a physical activity intervention, both, or neither. A decreased volumetric percent density of breast tissue on mammography was seen in the dietary and physical activity arms compared with the control arm (142). Several studies have shown that diet, exercise, or the combination produces clinically meaningful weight loss and reduction in body fat levels (143).

The Alberta Physical Activity and Breast Cancer Prevention (ALPHA) Trial was an RCT of 320 postmenopausal women who after a 12-month exercise intervention had decreased estradiol, increased SHBG levels (144), improved insulin levels, increased adiponectin:leptin ratio (145), and decreased nondense breast volume which was accounted for by changes in total body fat (146). The Breast Cancer and Exercise Trial in Alberta (BETA) was a 12-month 2-arm RCT of moderate-vigorous aerobic exercise for 5 days/week for either 30 minutes/session or 60 minutes/session in 400 postmenopausal women. In the exercise intervention groups, there were significantly greater reductions in total fat, subcutaneous abdominal fat, and waist-to-hip ratio, with dose–response effects being stronger in obese women (147). At 24 months (12 months after the end of the study intervention), patients who had fat loss greater than the median had more favorable metabolic biomarker profiles (148). Similarly, inflammatory markers decreased with increasing exercise adherence but this finding was mediated by fat loss (149). The WISER Sister study randomized 139 premenopausal women at high BC risk to 150 minutes or 300 minutes of weekly exercise over 5 months. Exercise-induced decreases in body fat levels were associated with improved adiponectin and leptin levels (150), lower follicular phase estrogen (151) and decreased breast volume and intensity measurements on magnetic resonance imaging (152).

Taken together, these studies show that diet and exercise interventions can meaningfully decrease body composition measures of fat mass, and that this decrease in body fat level improves metabolic and sex hormone levels associated with BC risk.

Pharmacologic Interventions

Metformin

Metformin is a biguanide used as first-line therapy for diabetes and occasionally off-label in weight management. Epidemiologic data addressing the association between metformin use and BC risk reduction in women with and without diabetes have been inconclusive (153, 154) due to heterogeneity in the variables controlled across studies, including reproductive status, BC receptor status, duration, and optimal drug dose (155-157) highlighting the need for more research. Nonetheless, a large body of evidence substantiates the mechanistic pathways by which metformin exerts its anticancer effects, which can be broadly classified as indirect and direct. The indirect effects entail systemic changes in host metabolism mediated by reduction in hyperglycemia, hyperinsulinemia, weight, and BMI (158). Direct effects are broader targeting the tumor microenvironment via several pathways relevant to BC metabolism (159, 160), effects which were independent of baseline BMI or insulin levels (161). By inhibiting gluconeogenesis, metformin indirectly activates AMPK, which limits the nutritional substrates requisite for BC cell proliferation (162) leading to downstream inhibition of mTOR, which supplies growth factors to cancer cells (162). Metformin also induces apoptosis by downregulating p53 resulting in cancer cell cycle arrest (163, 164).

The centrality of inflammation to carcinogenesis has shifted the focus on development of anti-inflammatory drugs targeting the NFkB pathway, which supports BC progression, among others. By reducing inflammatory cytokine IL-6 and tumor necrosis factor-α levels, metformin exerts its inhibitory effect on NFkB, thereby suppressing the early stages of inflammation associated with cancer development (165-167). Reduced adipose inflammation was also associated with a decrease in the number of aromatase-expressing CD68+ macrophages; these CLS essentially provide fuel for ER-positive tumors and are abundant in breast tissue in the setting of post-menopausal weight gain (168). Using a mouse model of obesity-induced metabolic dysfunction in menopause, metformin was further shown to decrease tumor size and prevent formation of new tumors (168). A phase II RCT is underway (NCT02028221) evaluating metformin for risk reduction of obesity-associated BC (169).

SGLT-2-inhibitors

Much research has been devoted to the effects of SGLT-2-inhibitors (SGLT2is) on body weight and adiposity, but few studies have tested the efficacy of SGLT2i on weight loss in obese patients without diabetes, much less in BC prevention. Up to 2% reduction in weight has been reported in SGLT2i users (170) independent of baseline BMI (171), which was maintained up to 4 years (172-174). Notably, 75% of the weight decline was from loss of fat mass, including visceral and subcutaneous adipose tissue (175), while sparing lean muscle mass (176). The metabolic shift in substrate utilization accounts for the weight loss. By mimicking a prolonged fasting state, glucosuria increases peripheral fat oxidation via lipolysis, and enhanced clearance of FFAs from the circulation (177), which are then used by the liver for ketogenesis. This metabolic reprograming was associated with augmented fat browning and reduced inflammation (178, 179). In a systematic review of 10 RCTs evaluating the effect of SGLT2is on adipocytokine profile, compared with placebo, treatment with an SGLT2i was associated with leptin inhibition and increased adiponectin expression (180). SGLT2is are currently not approved for weight loss; nonetheless, their underlying mechanisms of action in possibly tempering obesity-induced chronic inflammation and restoring adipose dysfunction independent of weight loss (181) merit further study in BC prevention.

Glucagon-like-protein 1 receptor agonists

Glucagon-like-protein 1 receptor agonists (GLP1-RAs) have emerged as effective and safe weight-reducing agents used for weight management. GLP-1, an incretin hormone, increases pancreatic insulin secretion in a glucose-dependent manner, reduces food intake, and decreases gastric emptying, and has been demonstrated to be effective in the treatment of overweight and obesity in people with and without type 2 diabetes (182-184). In a landmark phase III RCT of adults with overweight and obesity without type 2 diabetes, semaglutide demonstrated clinically remarkable weight reduction of 14.9% vs 2.5% in placebo (185). Phase III RCTs of tirzepatide, a dual-incretin therapy for weight loss, in patients with overweight and obesity without type 2 diabetes are ongoing. Additionally, several clinical studies have demonstrated the effects of liraglutide and exenatide in lowering host inflammatory cytokines, including tumor necrosis factor-α and interleukins 1β and IL-6 independent of weight loss (186-188). There is experimental evidence that GLP1-RA may exert direct effects on adipocyte function by inducing adiponectin secretion (189). Similarly, a meta-analysis of 20 RCTs demonstrated that liraglutide, but not exenatide, had a significant effect on raising adiponectin levels (190). Further study is needed to determine the translational value of these altered host immune mechanisms in BC prevention.

Conclusions

Obesity and weight gain are associated with increased risk of postmenopausal BC; however, BMI does not fully capture BC risk, as dysregulated metabolic pathways and dysfunctional adiposity can occur in women irrespective of BMI. This may underlie recent findings that metabolic markers and body composition more comprehensively capture individuals at higher BC risk, including women with normal BMI. Several interventions including dietary modification, exercise, and oral hypoglycemics can meaningfully improve body composition measures and metabolic and sex hormone levels associated with BC risk. Several important questions remain. First, further research is needed to correlate serum markers of metabolic dysfunction and chronic inflammation with changes that occur in breast tissue in real time in order to individualize predicted BC risk similar to cardiovascular risk prediction models. Second, we need to identify the optimal timing of intervention for BC risk reduction. It appears that the perimenopausal transition, when many women experience weight gain and increased visceral adiposity, is a particularly vulnerable time period (191), but is it not clear if weight gain in early adulthood already sets in motion chronic inflammation that contributes to weight gain in menopause. Finally, intervention studies targeting body fat levels are needed to curb the growing prevalence of obesity-related cancer.

Glossary

Abbreviations

AMPK

adenosine monophosphate–activated protein kinase

BC

breast cancer

BMI

body mass index

CLS

crown-like structure

ER

estrogen receptor

FFA

free fatty acid

GH

growth hormone

GLP1-RA

glucagon-like-protein 1 receptor agonist

HR

hazard ratio

HRT

hormone replacement therapy

IGF

insulin-like growth factor

IL

interleukin

MAPK

mitogen-activated protein kinase

mTOR

mammalian target of rapamycin

PI3K

phosphatidylinositol 3-kinase

RCT

randomized controlled trial

RR

risk ratio

SHBG

sex hormone–binding globulin

TNBC

triple negative breast cancer

Funding

This work was supported in part through the National Institutes of Health/National Cancer Institute (NIH/NCI) Cancer Center Support Grant P30 CA008748. S.S. is supported by the Clinical and Translational Science Center at Weill Cornell Medical Center and Memorial Sloan Kettering Cancer Center CTSA UL1TR00457 grant. N.M.I. is supported through the NIH 1R01CA235711 grant, the Breast Cancer Research Foundation, American Cancer Society, and the Kat’s Ribbon of Hope Foundation.

Disclosures

S.S. has received honoraria from MJH Life Sciences. M.Z. is an employee and stockholder of Eli Lilly and Company. During the initiation of this review, M.Z. was still employed at University of Chicago Medicine. The work completed in this review is not affiliated with Eli Lilly and was conducted entirely independently by M.Z. N.M.I. receives consultant fees from Novartis and Seattle Genetics and research funding (to institution) from Novartis.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.