Recently, evidence has been accumulating, which demonstrates that the adipose tissue is not merely a fat-storing tissue, but also produces various cytokines including leptin.1 Leptin is a multifunctional polypeptide hormone mainly secreted from the adipose tissue and its serum level positively correlates with the body mass index (BMI).2, 3, 4 The main role of leptin is to act upon the brain so that it decreases appetite and increases energy expenditure.5, 6 In addition to this preventive effect against obesity, leptin has recently been shown in vitro to have growth-stimulating effects on various malignant cells including breast cancer.7, 8 Cleary et al. have reported that oncogene-induced mammary tumor development significantly decreases in leptin-deficient Lepob/Lepob mice as compared with wild mice.9 These results suggest that leptin has an important role in breast carcinogenesis as well as the progression of breast cancer.
Leptin exerts its effects through the transmembrane leptin receptor (lep-R). This receptor is similar to class I cytokine receptors10 and contains 6 isoforms that differ at the cytoplasmic tail.11, 12 In human breast tumors, expression of the longest isoform (Lep-R(L)) and one of the shorter isoforms (Lep-R(S)) has been demonstrated.13 Very recently, it has been reported that immunohistochemically detected high expression of both leptin and Lep-R in breast cancer tissue is associated with a poor prognosis.14 These results seem to indicate a significant involvement of the leptin and Lep-R pathway in the progression of breast cancer. This immunohistochemical study,14 however, dealt with Lep-R expression, but not Lep-R(L) and Lep-R(S) expression separately, in breast cancer tissue, and the effect of serum leptin level on patient prognosis was not studied either. Since it is possible that serum leptin may affect tumor growth and, subsequently, patient prognosis, we examined in the study presented here both Lep-R(L) and Lep-R(S) mRNA levels in breast cancer tissue, by means of a real-time PCR assay. In addition, the serum leptin level was assessed to investigate its effect on patient prognosis.
Material and methods
Patients and surgical specimens
Ninety-one breast cancer patients who underwent mastectomy or breast conserving surgery during the period from March 2000 to October 2001 at Osaka University Hospital were consecutively recruited for our study. Tumor tissue samples were obtained from surgical specimens (n = 91), with special attention being paid to avoiding contamination by adipose tissue, snap frozen and kept at −80°C until use. Part of each tumor sample was subjected to histological examination to confirm the absence of adipose tissue contamination. All breast cancers were histologically diagnosed as invasive ductal carcinoma (85 cases) or invasive lobular carcinoma (6 cases), and patients with noninvasive carcinoma and patients who had undergone chemotherapy or hormonal therapy before surgery were excluded. Histological grade was determined according to the modified Scarff–Bloom–Richardson criteria.15 Informed consent was obtained from each patient before surgery.
For adjuvant hormone therapy, 54 patients were treated with tamoxifen (20 mg/day; n = 41), tamoxifen+goserelin (n = 12), or anastrozole (1 mg/day; n = 1). For adjuvant chemotherapy, 6 cycles of CMF (cyclophosphamide; 100 mg/day on p.o. days 1–14 + methotrexate 40 mg/m2 i.v. on days 1 and 8 + 5-FU 600 mg/m2 i.v. on days 1 and 8) were administered to 6 patients, 4 cycles of CE (cyclophosphamide; 600 mg/m2 i.v. on day 1 + epirubicin 60 mg/m2 i.v. on day 1) to 1 patient and 4 cycles of docetaxel (600 mg/m2 i.v. on day 1) to 2 patients. Sixteen patients were treated with combination chemotherapy [CMF (n = 2), CE (n = 10), docetaxel (n = 2) and other regimens (n = 2)] and hormone therapy [tamoxifen (n = 14), and goserelin (n = 2)]. Twelve patients received no adjuvant therapy. Indication for adjuvant treatment was decided essentially on the basis of the St. Gallen recommendations.16, 17 Patients were given a physical examination every 3 months for 2 years postoperatively, then every 6 months together with a blood test and chest X-ray examination. The median follow-up period for these 91 patients was 47 months, ranging from 6 to 53 months. Thirteen patients developed recurrences, i.e., 7 developed bone metastases, 4 liver metastases, 2 lung metastases and 3 soft tissue metastases. Ipsilateral breast recurrences after breast conserving surgery were not counted as recurrences.
RNA extraction and reverse transcription
Total RNA was extracted from the frozen tumor specimens by means of TRIZOL reagent (Molecular Research Center, Cincinnati, OH) according to the protocol provided by the manufacturer. Using oligo-(dT)15 primer and Superscript II (Life Technologies, Gaithersburg, MD), 3 μg of total RNA was reverse-transcribed for single-strand cDNA at 42°C for 90 min, followed by heating at 70°C for 10 min.
Real-time PCR assay of Lep-R(L), Lep-R(S), and leptin mRNA levels in tumor tissue
For real-time PCR for Lep-R(L), Lep-R(S) and leptin the ABI Prism 7700 Sequence Detection System (PerkinElmer Applied Biosystems) was used. The primer and probe mixture of Lep-R(L), Lep-R(S) and leptin was purchased from PerkinElmer Applied Biosystems under the following PCR conditions: incubation at 50°C for 2 min and denaturing at 95°C for 10 min were followed by 40 cycles of 96°C for 30 sec, 60°C for 1 min and 72°C for 30 sec. To normalize the transcript content in each of the samples, the β-glucuronidase transcripts were measured and used as quantitative control. The primer and probe mixture for β-glucuronidase was purchased from PerkinElmer Applied Biosystems and used according to the manufacturer's protocol. The standard curves for Lep-R(L), Lep-R(S), leptin and β-glucuronidase mRNA in tumor tissue were generated using serially diluted solutions of each PCR product as templates and target gene expression was calculated from these standard curves with the 10−12 μg PCR product for Lep-R(L), Lep-R(S) and leptin and the 10−8 μg PCR product for β-glucuronidase defined as 1. Finally, intratumoral mRNA expression levels of Lep-R(L), Lep-R(S) and leptin were determined as percentages of those of β-glucuronidase. The samples were submitted to duplicate real-time PCR assays and the resultant mean values were used as the relative expression levels.
Serum leptin assay
Blood samples were obtained from 67 patients immediately before surgery, and the serum was separated by centrifugation and stored at –20°C until use. Serum leptin levels were measured by enzyme-linked immunosorbent assay, using the kit provided by Linco Research (St. Charles, MO).
ER and PR assays
Estrogen receptor (ER) and progesterone receptor (PR) contents of breast cancer tissues were measured by means of enzyme immunoassay, using the kit provided by Abbott Research Laboratories (Chicago, IL). The cut-off value was 13 fmol/mg protein for ER and 10 fmol/mg protein for PR in accordance with the manufacturer's instruction.
The relationships between serum leptin or Lep-R (L)mRNA, Lep-R (S)mRNA and leptin mRNA levels in tumor tissue and clinicopathological factors of tumors were analyzed with the Mann–Whitney test. relapse-free survival (RFS) curves were calculated with the Kaplan–Meier method and the log-rank test was used to determine differences in RFS rates. Cox proportional hazards model was used to calculate the hazard ratio for each factor by means of univariate and multivariate analysis. Correlation of serum leptin level with Lep-R(L) mRNA, Lep-R(S) mRNA or leptin mRNA level in tumor tissue was analyzed with a simple curve fit test. Statistical significance was defined as p < 0.05.
Relationship between Lep-R(L), Lep-R(S) or leptin mRNA levels in tumor tissue and clinicopathological parameters of breast tumors
Lep-R(L), Lep-R(S) and leptin mRNA levels in tumor tissues were determined with a real-time PCR assay, and their relationship with the various clinicopathological parameters of breast tumors was investigated (Table I). Neither Lep-R(L) or Lep-R(S) mRNA levels were significantly associated with any clinicopathological parameters, including menopausal status, BMI, tumor size, histological grade, lymph node status, ER status or PR status. Leptin mRNA levels were significantly (p < 0.05) higher in ER-positive breast tumors than in ER-negative tumors. Postmenopausal status and low histological grade showed a tendency (p < 0.10) to be associated with high Leptin mRNA levels.
Table I. Relationship between Lep-R(L), Lep-R(S) or Leptin mRNA Level in Tumor Tissue and Clinicopathological Parameters of Breast Cancers
Relationship between Lep-R(L), Lep-R(S) or leptin mRNA levels in tumor tissue and patient prognosis
Patients were divided into intratumoral Lep-R(L) mRNA high and low groups, using a median value of 3.95 as the cut-off value, and into intratumoral Lep-R(S) mRNA high and low groups, using a median value of 60.0 as the cut-off value. There were no significant differences in the RFS rates between the intratumoral Lep-R(L) mRNA high and low groups (p = 0.45, Fig. 1a) or between the intratumoral Lep-R(S) mRNA high and low groups (p = 0.14) (Fig. 1b). Since the respective RFS rates tended to be lower in the intratumoral Lep-R(L) and the Lep-R(S) mRNA high groups than the intratumoral Lep-R(L) and the Lep-R(S) mRNA low groups, the patients were divided again into both intratumoral Lep-R(L) and Lep-R(S) mRNA high group and the others. The RFS rates of both intratumoral Lep-R(L) and Lep-R(S) mRNA high group (n = 27) were significantly (p < 0.01) lower than those of the others (n = 64) (Fig. 1c). Univariate analysis showed that tumor size, histological grade, PR status and Lep-R(L)/Lep-R(S) mRNA status in tumor tissue were significant (p < 0.05) prognostic factors, while there was no significant association between RFS rates and ER status, BMI, Lep-R(L) mRNA levels, Lep-R(S) mRNA levels, Leptin mRNA levels or serum leptin levels. Furthermore, multivariate analysis has shown that both intratumoral Lep-R (L) and Lep-R (S) mRNA high level is a significant (p < 0.05) risk factor for relapse, regardless of tumor size, histological grade or PR status (Table II).
Table II. Univariate and Multivariate Analysis of Various Prognostic Factors
Hazard ratio of large tumor size (>2.0 cm) against small tumor size (≤2.0 cm), histological grade III against grade I+II, lymph node positive against lymph node negative, ER-positive against ER-negative, PR-positive against PR-negative, BMI 21.0–23.0 or >23.0 against BMI <21.0, Leptin receptor (L) mRNA high against low, Leptin receptor (S) mRNA high against low, both Leptin receptor (L) and Leptin receptor (S) mRNA high against others, Leptin mRNA high against low, and serum leptin high against low.
Values in these columns indicate 95% confidence interval (CI).
Prognostic significance of Lep-R(L) and Lep-R(S) mRNA levels in tumor tissue in terms of intratumoral leptin mRNA or serum leptin levels
The relationship between Lep-R(L) and Lep-R(S) mRNA levels in tumor tissue and prognosis was studied in terms of the leptin mRNA levels in tumor tissue. The patients were divided into intratumoral leptin mRNA high and low groups using a median value of 0.68 as the cut-off value. The RFS rates of the intratumoral leptin mRNA high and low groups did not differ significantly. However, the RFS rates were significantly lower in both intratumoral Lep-R(L) and Lep-R(S) mRNA high group than the others in the patient subset with the high intratumoral leptin mRNA levels but not in that with the low intratumoral leptin mRNA levels (Figs. 2b and 2c).
We next studied the relationship between Lep-R(L) and Lep-R(S) mRNA levels in tumor tissue and prognosis in terms of serum leptin levels. The patients were divided into serum leptin high and low groups, using a median value of 7.5 ng/ml as the cut-off value. The RFS rates of the serum leptin high and low groups were not significantly different (Fig. 3a), but we found that the RFS rates were significantly lower in both intratumoral Lep-R(L) and Lep-R(S) mRNA high group than the others in the patient subset with the high serum leptin levels but not in that with the low serum leptin levels (Figs. 3b and 3c).
Comparison of BMI in both intratumoral Lep-R(L) and Lep-R(S) mRNA high group and the others
There was no significant difference in BMI between both intratumoral Lep-R(L) and Lep-R(S) mRNA high group and the others when all patients were taken into consideration. (Table III). The patients were then divided into intratumoral leptin mRNA high and low groups or into serum leptin high and low groups, and BMI was compared between both intratumoral Lep-R(L) and Lep-R(S) mRNA high group and the others in each of the subsets (Table III). No significant difference was found in BMI in any subset.
Table III. Comparison of BMI1 between Both Intratumoral Lep-R(L) and Lep-R(S) mRNA High Group and the Others, According to Intratumoral Leptin mRNA Level or Serum Leptin Level
Both intratumoral Lep-R(L)
Number of patients
Number of patients
Body mass index (kg/m2); values in mean ± SE.
22.00 ± 0.59
22.78 ± 0.32
Lep mRNA level in tumor tissue
22.22 ± 0.83
23.21 ± 0.48
21.76 ± 0.88
22.37 ± 0.43
Serum leptin level
23.16 ± 0.77
23.34 ± 0.56
20.21 ± 0.75
21.36 ± 0.46
Correlation of Lep-R(L), Lep-R(S) or leptin mRNA levels in tumor tissue with serum leptin levels
Neither Lep-R(L) mRNA nor Lep-R(S) mRNA levels in tumor tissue correlated with serum leptin levels as shown in Figure 4 (r = 0.02, p = 0.89 and r = 0.20, p = 0.11, respectively) but there was a significant correlation between intratumoral leptin mRNA levels and serum leptin levels (r = 0.40, p < 0.005).
Leptin is thought to exert its effects via 2 transmembrane receptors, i.e., Lep-R(L) and Lep-R(S).11, 12 Lep R(L) activates both the JAK2 (Janus activated kinase2)/signal transducers and activators of transcription 3 (STAT3) pathway and the mitogen-activated protein kinase (MAPK) pathway, while Lep-R(S) mainly activates the MAPK pathway because it lacks the STAT binding region in the cytoplasmic domain.18, 19, 20 We observed that Lep-R(S) mRNA levels in tumor tissue are much higher (∼14 times) than Lep-R(L) mRNA levels in breast tumor tissue. This is consistent with the previously reported finding that Lep-R(S) levels were 8 times higher than Lep-R(L) levels in peripheral blood mononuclear cells,21 and suggests that Lep-R(S) may play a more important role than Lep-R(L) in the signal transduction of leptin. However, since it has been demonstrated that the growth stimulation of leptin in T-47D breast cancer cells is blocked not only by a MAPK inhibitor13 but also by a STAT3 inhibitor,8 it can be hypothesized that not only the Lep-R(S) pathway (mainly the MAPK pathway) but also the Lep-R(L) pathway (both the MAPK and STAT3 pathway) plays an important role in the growth-signal transduction of leptin in breast cancer cells. Our observation that high intratumoral levels of both Lep-R(L) and Lep-R(S) mRNA, but not those of either Lep-R(L) or Lep-R(S) mRNA, are significantly associated with a poor prognosis seems to be consistent with this hypothesis. It is speculated that the Lep-R(L) and Lep-R(S) pathways may work synergistically to stimulate breast tumor growth. The most important finding of our current study is that a significant association of high intratumoral levels of both Lep-R(L) and Lep-R(S) mRNA with a poor prognosis was observed only in the subset of the patients with high intratumoral leptin mRNA levels or high serum leptin levels. These results seem to further support the hypothesis that the leptin and Lep-R(L)/Lep-R(S) pathways are implicated in the growth stimulation of human breast cancer.
It has been reported that obesity is associated with a poor prognosis for breast cancer compared with nonobesity.22, 23 Although multiple factors are thought to be involved in this association, our observation seems to suggest it could be explained, at least in part, by the growth-stimulating effects of leptin. Since there was no significant difference in BMI between both intratumoral Lep-R(L) and Lep-R(S) mRNA high group and the others as shown in Table III, we speculate that leptin influences breast cancer prognosis independently of other factors associated with obesity. Table I shows that intratumoral leptin mRNA levels in ER-positive tumors are significantly higher than in ER-negative tumors. This observation, taken together with the previously reported findings demonstrating that estrogens stimulate leptin promoter activity24 and leptin enhances aromatase expression in MCF-7 cells,25 seems to indicate a possible involvement of leptin in the growth stimulation of ER-positive breast cancers in an autocrine manner. A recent study has disclosed that leptin can promote angiogenesis by enhancing vascular endothelial cell proliferation and migration26, 27 and can stimulate the secretion of both MMP-2 and MMP-9, which play an important role in cell invasion, in cytotrophoblastic cells.28 Thus, in addition to the growth stimulating effect of leptin on cancer cells, leptin may promote tumor progression and metastasis through the enhancement of angiogenesis and invasiveness.
These results indicate the involvement of serum leptin in the progression of established breast tumors. However, we also speculate that a high serum leptin level may be associated with breast carcinogenesis since leptin stimulates the proliferation of normal human breast epithelial cells in vitro8 and mammary tumorigenesis is inhibited in leptin-deficient mice.9 Several reports on this issue have already been published. Except for the study by Tessitore et al., who found a significant association of high serum leptin levels with breast cancer risk,29 however, all other studies have failed to demonstrate a significant correlation between serum leptin levels and breast cancer risk.30, 31, 32 We were also unable to show a significant difference in serum leptin levels between breast cancer patients (9.52 ± 0.90, n = 104) and healthy women (10.86 ± 0.89, n = 104; data not shown). Since all these studies including ours are based on a relatively small sample size, we feel that it is prudent, for the time being, to consider that the influence of serum leptin levels on human breast carcinogenesis remains to be established.
While intratumoral leptin mRNA levels significantly (r = 0.40, p < 0.005) correlated with serum leptin levels, it is highly unlikely that leptin produced in tumors affects the serum leptin level for the following reasons. First, serum leptin levels in breast cancer patients and healthy women are not significantly different.30, 31, 32 Second, serum leptin levels do not show a significant difference between tumor stages.33 Third, leptin mRNA levels in tumor tissue are much lower (1/12th) than those in adipose tissue (data not shown). Thus, the reason for the association between serum leptin levels and intratumoral leptin mRNA levels is currently unknown. It is possible that high serum leptin levels preferentially stimulate the growth of leptin mRNA expressing tumor cells or that patients with high serum leptin levels may have a high expression of leptin mRNA in the normal breast tissue from the outset, so that the resultant breast tumors may show high leptin mRNA levels.
In conclusion, we have demonstrated that breast tumors with high intratumoral levels of both Lep-R(L) and Lep-R(S) mRNA are associated with a poor prognosis only in the subset of patients with high serum leptin or high intratumoral leptin mRNA levels. This suggests that the leptin and Lep-R(L)/Lep-R(S) pathways are implicated in the growth stimulation of breast tumors. The association of obesity with a poor prognosis for breast cancer may be explained, at least in part, by leptin. The leptin and Lep-R(L)/Lep-R(S) pathways could thus become a new target for treatment. However, our preliminary results need to be confirmed by a future study including a larger number of patients with a longer follow-up period.