The prevalence of obesity has markedly increased over the past two decades, especially in the industrialized countries. According to the Centers for Disease Control and Prevention (www.cdc.gov), in the United States alone, approximately 66% of adults are overweight as defined by body mass index (BMI, weight in kilograms divided by the square of the height in meters) in excess of 25, and 31% of adults are obese (BMI > 30) (Hedley et al., 2004). Globally, the overweight population has exceeded 1 billion (www.cdc.gov). Epidemiological data gathered over the past two decades clearly demonstrate that obesity in adults is associated with increased risk of cardiovascular disease, diabetes, some forms of cancer, and numerous other health disorders (Klein et al., 2002). Consequently, the regulation of body weight and obesity-related pathology is rapidly becoming a critical concern for public health experts and medical scientists worldwide (Kopelman, 2000).
While the impact of excess body weight on the development of cardiac disease and diabetes has been well documented, the link between obesity and carcinogenesis is just being recognized. The mechanism of adipose tissue-induced cancer is not known, but several possible scenarios can be envisioned. In the case of hormone-dependent neoplasms such as breast cancer, increased production and secretion of estrogenic compounds, growth factors, and angiogenic stimulators by excess fat tissue could contribute to tumor growth and metastasis (Sierra-Honigmann et al., 1998; Ahima and Flier, 2000; Miyazawa-Hoshimoto et al., 2004).
This review will focus on the link between leptin, a cytokine that is elevated in obese individuals, and cancer development. First, we briefly discuss the biological functions of leptin and the regulation of its production. Next, we summarize the effects of leptin on different cancer types in experimental cellular and animal models. Finally, we analyze epidemiological data on the relationship between circulating leptin levels and the presence of cancer or cancer risk in patients. The potential of targeting leptin in cancer treatment or prevention will be discussed.
LEPTIN AND ITS ACTIVITIES
Leptin, a product of the obese (ob) gene, is a 16 kDa cytokine that was discovered in 1994 as a regulator of body weight and energy balance acting in the hypothalamus (Zhang et al., 1994). Ob/ob mice with mutations in the gene encoding leptin do not produce leptin and become morbidly obese, infertile, hyperphagic, hypothermic, and diabetic due to unopposed appetite (Huang and Li, 2000). In humans, like in animals, leptin is an important regulator of energy balance. However, unlike in mice, human obesity is not related to leptin deficiency but to the development of leptin resistance (Bjorbaek and Kahn, 2004; Correia and Haynes, 2004; Hukshorn and Saris, 2004; Mark et al., 2004). In fact, mutations in the human ob gene are exceptionally rare (Chagnon et al., 2003).
In adult animals, leptin mRNA is primarily detected in white and brown adipose tissue (Masuzaki et al., 1996). In addition, a number of non-adipocyte tissues have been shown to synthesize and secrete leptin. The other leptin sources include gastric mucosa cells (Bado et al., 1998), mammary epithelial cells (Smith-Kirwin et al., 1998), myocytes (Wang et al., 1998), and the placenta (Senaris et al., 1997). Leptin expression has also been reported in the testes, ovary, and hair follicles (Hoggard et al., 1997). Subsequent studies documented that in addition to its primary function as a regulator of food intake, leptin can affect fetal development, sex maturation, lactation, hematopoiesis, and immune responses (Wauters et al., 2000; Bonnet et al., 2002; Brann et al., 2002; Neville et al., 2002; Goumenou et al., 2003).
In humans, the major factor influencing plasma leptin concentrations is adipose tissue mass (Maffei et al., 1995). Circulating leptin levels exhibit a particularly strong positive correlation with total body fat, and to a lesser degree with BMI (Frederich et al., 1995; Ahima et al., 1996; Boden et al., 1996; Sinha et al., 1996; der Merwe et al., 1999; Thomas et al., 2000). Higher concentration of serum leptin in obese individuals is associated with both increased fat mass and increased leptin release from larger adipocytes (Hamilton et al., 1995; Kolaczynski et al., 1996a,b). Importantly, serum leptin levels are significantly higher in women than in men, even after the adjustment for total body fat mass (Havel et al., 1996; Rosenbaum et al., 1996). One explanation for this tendency is differential regulation of leptin expression by sex hormones, with estrogens reported to upregulate (Casabiell et al., 1998; Castracane et al., 1998) and testosterone observed to decrease leptin levels (Blum et al., 1997; Elbers et al., 1997; Jockenhovel et al., 1997).
The synthesis of leptin in adipocytes is influenced by different humoral factors, most notably insulin (Cusin et al., 1995; Leroy et al., 1996), tumor necrosis factor alpha (TNF-α) (Zhang et al., 2000), glucocorticoids (De Vos et al., 1995; Dagogo-Jack et al., 1997), reproductive hormones (Machinal-Quelin et al., 2002), and prostaglandins (Fain et al., 2000). Importantly, many of these factors have been shown to be associated with neoplastic processes.
In the context of cancer, it is noteworthy that leptin expression can be induced under hypoxic conditions, which often occur in solid tumors (Ambrosini et al., 2002; Grosfeld et al., 2002). Hypoxia and chemical inducers of cellular hypoxia are able to activate the leptin gene promoter through the hypoxia-induced factor-1 (HIF-1) in human adipocytes and fibroblasts (Ambrosini et al., 2002; Grosfeld et al., 2002). These data suggest that leptin may play a role in vascular remodeling (Stenmark et al., 2002). Indeed, leptin has been shown to regulate neoangiogenesis by itself and in concert with vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) 2 (Bouloumie et al., 1998; Sierra-Honigmann et al., 1998; Cao et al., 2001). In addition to the proangiogenic function, leptin can enhance endothelial cell growth (Bouloumie et al., 1998; Sierra-Honigmann et al., 1998) and suppress apoptosis through a Bcl-2-dependent mechanism (Artwohl et al., 2002). The role of leptin in neovascularization is supported additionally by the observation that the hormone can increase the levels and activity of enzymes involved in angiogenesis, for example, matrix metalloproteinases (MMPs) 2 and 9 (Park et al., 2001; Kume et al., 2002).
In addition to its involvement in endothelial cell function, leptin has been shown to act as a mitogen, transforming factor, or migration factor for many different cell types, including smooth muscle cells (Oda et al., 2001), normal and neoplastic colon cells (Hardwick et al., 2001; Liu et al., 2001), and normal and malignant mammary epithelial cells (Dieudonne et al., 2002; Laud et al., 2002).
The activities of leptin are mediated through the transmembrane leptin receptor (ObR) (Tartaglia, 1997; White et al., 1997). In human tissues, at least four isoforms of ObR with different COOH-terminal cytoplasmic domains have been described (Barr et al., 1999). The full (long) form of ObR (ObRl) is 1,165 amino acids long (Mr ∼ 150,000–190,000) and contains the extracellular, transmembrane, and intracellular domain (Fig. 1). The extracellular domain binds ligand, whereas the intracellular tail recruits and activates signaling substrates. Only ObRl has a full signaling potential, whereas short ObR isoforms (ObRs) lack major domains recruiting downstream effectors and have diminished or abolished signaling capability (Fig. 1) (Bjorbaek et al., 1997, 2001; Sweeney, 2002; Zabeau et al., 2003). ObRl is highly expressed in the hypothalamus (Bjorbaek et al., 1997). However, lower levels of ObRl have been identified in many peripheral organs (Morton et al., 1999; Frank et al., 2000; Buyse et al., 2001; Goiot et al., 2001; Akerman et al., 2002; Ebenbichler et al., 2002; Lee et al., 2002; Kim et al., 2003).
The signaling pathways activated by ObRl include the classic cytokine JAK2/STAT3 (Janus kinase 2/signal transducer and activator of transcription 3) pathway; the Ras/ERK1/2 (Ras/extracellular signal-regulated kinases 1/2) signaling cascade; and the PI-3K/Akt/GSK3 (phosphoinositide 3 kinase/protein kinase B/glycogen synthase kinase 3) growth/antiapoptotic pathway. In addition, leptin has been found to induce phospholipase C (PLC)-gamma, protein kinase C (PKC), p38 kinase, and nitric oxide (NO) (Bjorbaek et al., 1997; Sweeney, 2002; Zabeau et al., 2003) (Fig. 2). Ultimately, induction of ObRl can activate several genes involved in cell proliferation, including c-fos, c-jun, junB, egr-1, and socs3, and upregulate the expression angiogenic factors, such as VEGF (Sweeney, 2002; Zabeau et al., 2003; Frankenberry et al., 2004) (Fig. 2).
The short forms of ObR are ubiquitously expressed (Fei et al., 1997). Their function is not clear, but there is evidence that ObRs can be involved in intra- and transcellular leptin transport (Hileman et al., 2000). In addition to the cellular isoforms of ObR, a soluble leptin receptor (SLR) that is generated by MMP-mediated shedding of the ObR ectodomain has been identified (Maamra et al., 2001; Ge et al., 2002) (Fig. 1). In human plasma as well as in vitro models, SLR sequesters leptin, preventing the activation of ObR signaling (Ge et al., 2002; Yang et al., 2004). The abundance of SLR is regulated by different physiological and pathological conditions, thus this receptor may serve as a potent modulator of leptin action in various tissues (Chan et al., 2002).
Recent studies suggested that leptin signaling can crosstalk with both polypeptide growth factor signaling and with steroid receptor function. For instance, insulin is known to increase leptin expression (Cusin et al., 1995; Saladin et al., 1995; Hardie et al., 1996; Leroy et al., 1996), but it can also induce leptin resistance by the inhibition of leptin signaling through JAK2 (Kellerer et al., 2001). In addition, the identification of leptin receptor on pancreatic beta-cells suggests the existence of a “adipoinsular” feedback loop whereby leptin may inhibit insulin secretion (Kieffer et al., 1996; Kulkarni et al., 1997).
Our work demonstrated that leptin modulates cellular response to estrogens. Specifically, leptin enhances the stability of the estrogen receptor alpha (ERα), leading to the maintenance of ERα-dependent transcription in breast cancer cells in the presence of antiestrogens (Garofalo et al., 2004). Leptin can also increase estrogen levels through the activation of aromatase expression and activity (Kitawaki et al., 1999; Magoffin et al., 1999; Catalano et al., 2003).
EFFECTS OF LEPTIN ON CANCER CELLS IN VITRO
The effects of leptin on cancer cells in vitro are summarized in Table 1.
Table 1. Effects of leptin on cancer cells in vitro
Type of cancer
Effects of leptin
Increased cell proliferation
Human breast cancer cell lines; T47D, MCF-7, ZR75-1
Dieudonne et al. (2002); Hu et al. (2002); Laud et al. (2002); Okumura et al. (2002); Catalano et al. (2003); Somasundar et al. (2003); Garofalo et al. (2004); Yin et al. (2004)
Leptin is necessary for normal mammary gland development in rodents and humans (Smith-Kirwin et al., 1998; Hu et al., 2002; Neville et al., 2002). However, new data suggest that leptin might be involved in mammary carcinogenesis (O'Brien et al., 1999; Dieudonne et al., 2002; Hu et al., 2002; Laud et al., 2002; Okumura et al., 2002; Cleary et al., 2003). Leptin and ObR have been found in normal and cancer mammary epithelium by immunohistochemistry with specific antibodies (Abs) (Ishikawa et al., 2004). Interestingly, both leptin and ObR appear to be significantly overexpressed in cancer tissue relative to non-cancer epithelium (Ishikawa et al., 2004).
The expression of ObR protein has also been demonstrated in cultured breast cancer cells with Ab against the common domain of ObR (Garofalo et al., 2004), or a specific ObRl region (Dieudonne et al., 2002; Hu et al., 2002; Laud et al., 2002; Garofalo et al., 2004; Yin et al., 2004). Interestingly, we observed that ERα-positive breast cancer cells MCF-7 and T47D express higher levels of ObRl than ERα-negative cell lines MDA-MB-231 and MDA-MB-435 (Garofalo et al., 2004). RT-PCR methods confirmed that human breast cancer cells express detectable levels of ObR mRNA, including ObRl-encoding RNA (Dieudonne et al., 2002).
In ERα-positive breast cancer cell lines, T47D, MCF-7, and ZR75-1, leptin stimulates DNA synthesis and cell growth acting through multiple signaling cascades, including the JAK/STAT, ERK1/2, and PKC-α pathways (Dieudonne et al., 2002; Hu et al., 2002; Laud et al., 2002; Okumura et al., 2002; Catalano et al., 2003; Somasundar et al., 2003; Garofalo et al., 2004; Yin et al., 2004). In addition to the classic cytokine signals, leptin is able to induce the Akt/GSK3 survival pathway in breast cancer cells (Garofalo et al., 2004). Interestingly, the stimulation of Akt/GSK3 occurs several hours after leptin addition, suggesting the involvement of intermediate signaling mechanisms (Garofalo et al., 2004).
Leptin-induced cell cycle progression is accompanied by upregulation of cdk2 and cyclin D1 levels (Okumura et al., 2002) and hyperphosphorylation/inactivation of the cell cycle inhibitor, pRb (Garofalo et al., 2004). Noteworthy, in ERα-positive T47D breast cancer cells, leptin induced not only cell growth but also cellular transformation (anchorage-independent growth). This activity of leptin was not observed in normal breast epithelial cells (Hu et al., 2002).
Of particular interest is the link between leptin activity and ERα. Recent reports demonstrated that ERα and ObR are coexpressed in malignant mammary tissue and breast cancer cell lines (Dieudonne et al., 2002; Hu et al., 2002; Laud et al., 2002). Notably, mitogenic effects of leptin and leptin-dependent activation of STAT3 require SRC-1, a member of the p160 family of steroid receptor modulators (Yin et al., 2004), which might represent crosstalk between steroid receptor- and leptin-induced transcriptional mechanisms. Furthermore, leptin has been found to modulate both estrogen synthesis and ERα activity. For instance, leptin can upregulate the aromatase gene expression and aromatase activity in MCF-7 cells, possibly leading to increased estrogen synthesis (Catalano et al., 2003). We reported that leptin is able to interfere with the action of antiestrogens via post-transcriptional modulation of ERα. Specifically, in MCF-7 cells treated with the pure antiestrogen ICI 182, 780, leptin markedly decreased ERα ubiquitination, increasing its half-life. These effects coincided with elevated nuclear ERα expression, increased ERα recruitment to estradiol-sensitive promoters, and increased estrogen response element-dependent transcription (Garofalo et al., 2004). All these observations suggest that high levels of leptin in obese breast cancer patients might contribute to tumor growth and the development of antiestrogen resistance.
There is accumulating evidence that leptin signaling might be involved in colon cancer. The presence of ObRs and ObRl mRNA has been documented using RT-PCR in colon cancer cell lines, human colon tumors, polyps, and adjacent mucosa (Attoub et al., 2000; Hardwick et al., 2001; Rouet-Benzineb et al., 2004). In support of these data, the expression of ObR in colon cancer tissues and cell lines has been confirmed by immunodetection with Abs against the intracellular domain of ObR (Hardwick et al., 2001).
Several in vitro experiments demonstrated mitogenic activity of leptin in colonic epithelial cells. In this model, leptin can induce proliferation through the activation of the NF-kappaB and ERK1/2-dependent pathways as well as upregulate c-fos expression (Hardwick et al., 2001; Liu et al., 2001; Rouet-Benzineb et al., 2004). Furthermore, leptin-induced NF-kappaB is able to decrease apoptotic effects of butyrate in colon cancer HT-29 cells (Rouet-Benzineb et al., 2004). Moreover, leptin can activate the invasiveness of PC/AA/C1 early stage colon cancer cells through PI-3K, Rho-, and Rac-dependent pathways (Attoub et al., 2000).
There is some evidence suggesting that leptin might play a role in the development of prostate cancer. The presence of ObR has been detected in normal prostate epithelia (Cioffi et al., 1996) and in benign and malignant prostate epithelial cells stained with an Ab recognizing a common domain of the receptor (Stattin et al., 2001). In agreement with this, both ObRs and ObRl mRNAs have been found in malignant prostate cells with RT-PCR-based methods (Stattin et al., 2001; Onuma et al., 2003; Somasundar et al., 2004). Recently, Somasundar et al. (2004) reported that leptin can increase growth and survival of prostate cancer cells through either the PI3K/Akt or ERK1/2 pathways, depending on cell type. In addition, leptin also induced a proliferative response, via JNK activation, in androgen-independent PC-3 and DU145 human prostate cancer cells, but not in androgen-dependent, ObRl expressing LNCaP-FGC cells (Onuma et al., 2003).
In addition to the mitogenic effects, leptin has been found to act as a motility factor well as an inducer of pro-metastatic factors VEGF, transforming growth factor TGF-β1 and FGF-β in prostate cancer cells in vitro (Frankenberry et al., 2004).
Ovarian and endometrial cancer
Epithelial ovarian cancer (EOC) is a leading cause of death from gynecological malignancies. The involvement of leptin in EOC is not clear. One study (Choi et al., 2004) found both ObRs and ObRl mRNAs in immortalized ovarian surface epithelium cell lines and in BG-1, OVCAR-3, and SKOV-3 ovarian cancer cell lines. Despite this ubiquitous expression of ObR, only BG-1 cells responded to leptin with proliferation via the ERK1/2 MAPK pathway (Choi et al., 2004).
Like in ovarian cancer, the involvement of leptin in endometrial carcinogenesis still needs further investigation. In a study on 70 cases, Yuan et al. (2004) demonstrated by RT-PCR and immunoblotting that ObRs and ObRl are expressed in both cancer and non-cancer endometrium. The abundance of ObRl was similar in cancer and normal tissues, but the levels of ObRs were significantly decreased in malignant cells. Moreover, forced expression of ObRs in the endometrial cancer cells RL95-2 suppressed cell proliferation (Yuan et al., 2004), suggesting that loss of ObRs in endometrial cancer might contribute to malignant progression.
Several authors proposed a link between obesity and pancreatic beta cell cancer, but the data on the role of leptin in this disease are controversial, and only a few studies have been completed using human cell models. Okuya et al. (2001) demonstrated that leptin can suppress apoptosis and stimulate the proliferation of rat insulin-secreting tumor cell lines. An induction of c-fos expression and proliferative response was also observed in RINm5F rat-insulinoma-derived cell line (Islam et al., 1997). Similarly, in BRIN-BD11 rat insulinoma cell line, leptin stimulated STAT3 and STAT5b phosphorylation (Briscoe et al., 2001). However, Somasundar et al. (2003) reported that leptin reduced cell growth of Mia-PaCa and PANC-1 human pancreatic cancer cells, suggesting that leptin action might be dependent on cell context.
The evaluation of ObR expression in human cell lines and tissues has not been attempted. ObRs and ObRl mRNAs were detected in rat insulinoma RIN m5F cells, mouse insulinoma cells BetaCT6 (Islam et al., 1997; Kulkarni et al., 1997), and BRIN-BD11 cells (fusion between rat beta-cells and rat insulinoma cells RINm5F cells) (Briscoe et al., 2001).
The involvement of leptin in lung cancer is unknown. Preliminary reports demonstrated the existence of ObRl in human lung tissue and SQ-5 cells derived from human lung squamous cell carcinoma. In this cell line, leptin enhanced cell growth via the ERK1/2 pathway (Tsuchiya et al., 1999).
The data on leptin and adrenal cancer are limited. Immunohistochemical analysis demonstrated a strong expression of ObRl in both benign and malignant adrenal tumors, and a weak ObRl staining in pheochromocytomas (Glasow et al., 1999). Although leptin is known to inhibit adrenal steroidogenesis (Malendowicz et al., 1997; Pralong et al., 1998), it apparently does not affect the growth of human adrenal cells and human adrenocortical carcinoma cell line, NCI-H295 (Glasow et al., 1999).
Preliminary studies of Jin et al. (1999) suggested that leptin, ObRl, and ObRs are commonly expressed in normal and neoplastic anterior pituitary cells. Interestingly, leptin immunoreactivity in adenoma was decreased compared with its abundance in normal pituitaries. In addition, leptin was found to inhibit proliferation of the human HP75 and in the rat pituitary GH3 cell lines, but stimulate pancreastatin secretion in vitro. Thus, in this system, leptin might play a role in the growth and differentiation of anterior pituitary cells, but not in tumorigenesis (Jin et al., 1999).
In addition to solid tumors, ObRl and ObRs mRNAs have been detected in several myeloid and lymphoid leukemic cell lines (Bennett et al., 1996; Cioffi et al., 1996; Nakao et al., 1998). Leptin has been found to stimulate the proliferation of human myeloid leukemia cell lines OCI/AML2 and MO7E, although the induction of postreceptor signaling in these cells did not reflect the levels of ObRl or ObRs (Konopleva et al., 1999). Noteworthy, Konopleva et al. (1999) reported that freshly prepared leukemic cells from some acute myeloid leukemia patients exhibited a proliferative response to leptin alone, and that combinations of leptin with other hematopoietic cytokines (IL-3, G-CSF, and SCF) induced an additive or synergistic mitogenic response in 7 of 14 acute myeloid leukemia cases.
OBESITY, SERUM LEPTIN LEVELS, AND CANCER
The major reports on the association between circulating leptin levels and cancer are summarized in Table 2.
Table 2. Association between serum leptin levels and cancer in vivo
Type of cancer
Association of serum leptin levels with other markers
Positive association with elevated values of ER and PR in patients with breast cancer
Numerous studies have established that obesity is a risk factor for breast cancer development in postmenopausal women (Cleary and Maihle, 1997; Chlebowski et al., 2002; Rose et al., 2002; Stephenson and Rose, 2003). Notably, a large body of evidence suggests that postmenopausal obesity, assessed by BMI, is associated with the increased risk of developing hormone-dependent breast cancer (i.e., ER and progesterone receptor (PR)-positive tumors), but not with the incidence of ER/PR-negative tumors (Potter et al., 1995; Enger et al., 2000; Huang et al., 2000). In support of this notion, increased upper body obesity, defined by waist-to-hip ratio (WHR) above 0.8, has been found to correlate with poor prognosis, but only in postmenopausal women with ER-positive tumors (Borugian et al., 2003). Consistent with these observations, obesity increased tumorigenic potential of ERα-positive, but not of ERα-negative, breast cancer xenografts grown in MMTV-neu animal models (Cleary et al., 2004). The mechanism of obesity-induced hormone-dependent breast cancer is not known, but the possible contributing factors are the increased exposure of mammary epithelial cells to estrogens locally produced by adipose tissue and downregulation of sex hormone-binding globulin occurring in women with high WHR (Soler et al., 1989).
The impact of excess body weight on breast cancer in premenopausal women is still unclear. Several studies have suggested that in the premenopausal population, obesity assessed by BMI might protect against breast cancer development, although the physiological mechanism of this protection remains speculative (London et al., 1989; Vatten and Kvinnsland, 1992; Huang et al., 1997). On the other hand, some new data indicated that excess upper body obesity increases breast cancer risk even in premenopausal women (Harvie et al., 2003). Moreover, high BMI is significantly associated with an increased risk of inflammatory breast cancer, the most lethal form of breast cancer in both premenopausal and postmenopausal populations (Chang et al., 1998).
Despite the unquestionable link between obesity and certain forms of breast cancer and the large body of in vitro data suggesting the role of leptin in breast cancer development, the association between circulating leptin and breast cancer or breast cancer risk is still unclear. Some reports found that serum leptin was associated with breast cancer regardless of the menopausal status (Tessitore et al., 2000; Han et al., 2005), while other data suggested a negative correlation between leptin and breast cancer in the premenopausal, but not postmenopausal, group (Petridou et al., 2000). Furthermore, several authors described that circulating leptin is not significantly associated with breast cancer or breast cancer risk in pre- and postmenopausal women (Mantzoros et al., 1999; Sauter et al., 2004; Stattin et al., 2004).
The inconsistent data obtained with circulating leptin as a breast cancer marker could be, at least in part, explained by differences in detection techniques, or by differences in sample preparation where some influencing factors, such as food intake and circadian rhythm, were not controlled. Clearly, better-controlled studies are needed to unequivocally establish whether serum leptin is associated with breast cancer etiology. Moreover, it is possible that breast carcinogenesis is induced by overabundance of local rather then systemic leptin. This concept could be addressed by direct examination of the leptin system in breast tumors. For instance, one recent study documented overexpression of leptin and ObR in breast cancer tissues versus non-cancer breast epithelium. In addition, there was an increased incidence of hematogenous metastasis or cancer recurrence in distant organs in patients with increased ObRl expression in primary breast tumors, while patients with ObR-negative and low leptin-expressing tumors displayed good outcome (Ishikawa et al., 2004).
Obesity has been consistently associated with higher risk of colorectal cancer in men (relative risks of approximately 1.5–2.0) and women (relative risks of approximately 1.2–1.5), in both case control and cohort studies (Calle and Thun, 2004). Similar relationships have been noticed for colon adenomas, with stronger association for larger adenomas (Giovannucci et al., 1996). The link between colorectal cancer and obesity is more significant in men than in women (Terry et al., 2001, 2002), but the possible reasons for this gender difference remain speculative. One hypothesis is that central adiposity (characteristic for men) is a stronger predictor of colon cancer risk than peripheral adiposity or general overweight. In fact, waist circumference and WHR are related strongly to risk of colorectal cancer and large adenomas in men (Giovannucci et al., 1995). However, the association between WHR and colorectal cancer in women was not stronger than the association between BMI and colorectal cancer in several studies that examined both parameters, implying that body fat distribution alone cannot account for the gender difference (Caan et al., 1998; Giacosa et al., 1999). Another possible explanation is that there may be an offsetting beneficial effect of obesity on colorectal cancer risk in women. Substantial evidence supports the protective role of exogenous estrogens (in the form of postmenopausal hormone therapy) on the risk of colorectal cancer in women (Calle et al., 1995; Rossouw et al., 2002). Thus, high levels of circulating estrogens associated with postmenopausal obesity in women may diminish the obesity-associated risk of colorectal cancer.
The studies on circulating leptin levels and colorectal cancer are not conclusive. For instance, Stattin et al. (2003b) observed that elevated levels of circulating leptin were associated with about a twofold increase in risk of colorectal cancer in men, but not in women. On the other hand, Tessitore et al. (2000) found no difference between serum leptin levels in cancer patients and controls. Other authors noted that serum leptin levels in patients with colon cancer were significantly decreased (Arpaci et al., 2002). In the same study, leptin levels were lower in colon cancer patients despite lack of weight loss and BMI measurements comparable to that of control subjects.
Circulating leptin is increased by high-fat diets, which are also implicated in stimulating colon cell proliferation (Lin et al., 1998; Bahceci et al., 1999; Baile et al., 2000). It has been reported that the presence of dietary fiber can decrease serum leptin levels and reduce colon carcinogenesis through reduced colonocyte cell proliferation (Alberts et al., 1990; Agus et al., 2000). A positive relationship between dietary fat, serum leptin, and colonic epithelial cell proliferation has also been reported in animal models (Hardwick et al., 2001; Liu et al., 2001). Globally, these preliminary data suggest that high fat diet-related colon carcinogenesis is, at least, in part mediated through a mechanism involving higher systemic leptin levels.
Results of studies examining the association between BMI and prostate cancer risk are inconclusive. Although several large studies have found an increased BMI in adulthood to be associated with an increased risk of prostate cancer development, others have shown no such association (Freedland and Aronson, 2005). Recently, Giovannucci et al. (2003) published the analysis of a large cohort of men followed in the Health Professionals Follow-Up Study 1986–2000. In this study, prostate cancer risk was inversely associated with BMI in men younger than 60 years of age, but no association was found in men older than 60 (Giovannucci et al., 2003). The lack of significant impact of obesity on prostate cancer in older men was confirmed by other authors (Porter and Stanford, 2005). Interestingly, the protective effect of obesity was not observed in cases of advanced metastatic prostate cancer, regardless of patient age (Giovannucci et al., 2003).
In contrast, two large prospective studies of the American Cancer Society reported a positive association between obesity and increased risk of dying from prostate cancer (Moyad, 2002; Calle et al., 2003). A similar link has been observed for adolescent obesity and prostate cancer deaths, implicating early life events in prostate cancer development (Kaaks et al., 2000). The reasons for the discrepancies observed in the above studies are not clear and could be related to different methodology of data collection and analysis.
Like with the studies on obesity and prostate cancer, many reports on circulating leptin levels and prostate cancer risk have yielded conflicting results. Several authors noted a positive correlation between serum leptin and prostate cancer risk (Stattin et al., 2001; Saglam et al., 2003). Hsing et al. (2001) found that prostate cancer risk is associated with higher leptin levels and with WHR values above 0.87, which suggested that leptin could interact with humoral factors related to abdominal obesity, such as sex hormones, insulin and IGF-1, to increase the risk of prostate cancer. Interestingly, two studies that examined serum leptin in men with prostate cancer reported that higher leptin levels were associated with more advanced tumors, characterized by larger size and higher grade (Chang et al., 2001; Saglam et al., 2003). However, several other studies concluded that circulating leptin is not associated with prostate cancer (Lagiou et al., 1998; Hsing et al., 2001; Stattin et al., 2003a).
There is convincing and consistent evidence from both case-control and cohort studies that obesity is strongly related to endometrial cancer (Calle and Thun, 2004). In fact, a linear increase in the risk of endometrial cancer with increasing weight or BMI has been observed in most of the relevant studies (Calle et al., 2003; Soliman et al., 2005). It is thought that the key factor involved in endometrial cancer etiology in obese women is excess available estrogen, either insufficiently balanced by progesterone in premenopausal women or synthesized by adipose tissue in postmenopausal women. Furthermore, there is evidence that chronic hyperinsulinemia that is associated with obesity might increase the risk of endometrial cancer (Kaaks et al., 2002).
Like with other obesity-related cancers, the link between circulating leptin and endometrial cancer is unclear. Limited studies by Petridou et al. (2002) indicated that serum leptin levels were positively associated with endometrial cancer. Similarly, Yuan et al. (2004) reported that circulating leptin was significantly higher in endometrial cancer patients than in normal controls, however this association was not observed after BMI normalization.
Numerous epidemiological studies have examined the association between BMI and ovarian cancer. The extensive review of Purdie et al. (2001) concludes that there is a weak association between increased BMI and ovarian cancer based on the results of cohort and population based case-control studies. In addition, the link between ovarian cancer risk and higher WHR has been reported by the Iowa Women's Health Study Cohort (Mink et al., 1996) and an Italian multicenter case-control study (Dal Maso et al., 2002). Furthermore, higher BMI in young adulthood has been found associated with an increased risk of premenopausal ovarian cancer (Fairfield et al., 2002), especially for serous borderline tumors (Kuper et al., 2002). No large studies addressed the correlation between circulating leptin and ovarian cancer.
Several recent studies have suggested that obesity may be associated with increased risk for pancreatic cancer development in man and women (Moller et al., 1994; Silverman et al., 1998; Michaud et al., 2001; Calle et al., 2003). However, other studies found no such association (Stolzenberg-Solomon et al., 2002; Lee et al., 2003), or reported an association in men but not in women (Gapstur et al., 2000).
Serum leptin concentrations were significantly elevated in patients with insulinomas producing chronically high insulin levels. On the other hand, leptin levels return to normal, after surgical treatment and normalization of insulin values (Popovic et al., 1998). In weight-losing patients with pancreatic cancer, low leptin concentration is associated with increased insulin resistance (Barber et al., 2004). However, high plasma leptin levels do not appear to contribute to cachexia in these patients (Brown et al., 2001).
BMI has been reported to be inversely associated with lung cancer in several study populations that did not exclude smokers from the analysis (Calle and Thun, 2004). This negative correlation can be explained by the confounding effects of smoking, which is inversely associated with BMI (Henley et al., 2002). No association between BMI and lung cancer was observed in non-smoking populations (Calle et al., 2003).
The impact of circulating leptin on lung cancer progression has not been studied. Studies that addressed the role of circulating leptin in the development of anorexia and cachexia in lung cancer patients concluded that leptin is not a primary regulator of these processes (Simons et al., 1997).
Epidemiological data suggest that obesity is associated with increased risk of certain types of cancer. In humans, high BMI is directly associated with elevated levels of the obesity hormone, leptin. Leptin, in addition to its neuroendocrine function, can act as a mitogen and an angiogenic factor. Consequently, several recent studies addressed the possible role of leptin in cancer development and progression. The possibility that the hormone might activate cell growth, transformation, or drug resistance has been assessed with different cellular and animal cancer models. The resulting data indicate that many types of cancer cells can respond to leptin as a mitogen/survival factor. To date, the best evidence that the hormone can indeed be involved in neoplastic processes has been provided by studies on breast and colorectal cancer models, while the results for other cancer types are very limited and often inconsistent or inconclusive. In any case, more research is needed to examine molecular mechanisms of leptin involvement in breast and colorectal cancers and to prove or exclude its role in other types of neoplasms.
Epidemiological studies measuring cancer risk in relation to obesity (assessed by BMI or WHR) confirmed that excess body fat can increase the risk of developing postmenopausal breast cancer and endometrial cancer. There is also supporting evidence for the association between obesity and colorectal cancer. Unfortunately, the attempts to correlate serum leptin abundance with cancer incidence or progression were not conclusive, regardless of the disease studied. This perhaps was related to differences in sample preparation and measurement techniques as well as the lack of control for other factors that influence leptin expression, such as food intake.
Taking into consideration data obtained with cultured cells and tumor specimens, one cannot exclude that local, not systemic, leptin concentrations are critical for tumor progression. Leptin abundance in tumor environment can be regulated by surrounding adipose tissue. In addition, tumor cells themselves, as shown in breast cancer specimens, can produce the hormone. This paracrine/autocrine leptin axis could become a target for leptin-inhibiting drugs, which might prove effective in cancer treatment and prevention.
This work was partially supported by the funds from Sbarro Health Research Organization (to E.S.).