TNF-α,2 a macrophage-derived proinflammatory cytokine, plays a critical role in normal host resistance to infection and to the growth of malignant tumors. Virtually all cell types possess either one or both of TNFR1 and TNFR2, which mediate the vast range of TNF-α effects. Both receptor types shed soluble forms (sTNFR1 and sTNFR2) that are found in blood and urine and that can neutralize or modulate the activities of TNF-α or possibly serve as reservoirs for its controlled release. The soluble receptors are the only natural molecules known to interfere with TNF-α activity, and their shedding from the cell membrane constitutes a mechanism for decreased responsiveness to TNF-α (reviewed in Ref. 1).

The variations in serum levels of the soluble receptors among healthy individuals are likely to be determined genetically and are very stable, with 1-year correlation coefficients (r) of 0.84 and 0.90 for sTNFR1 and sTNFR2, respectively (2). Blood levels of both correlate positively with BMI (3) and age (4). Susceptibility to neoplasia may accompany the rise in sTNFR levels in elderly people by opposing the protective effects of TNF-α. Moreover, Aderka et al.(5) show that the soluble receptors have an inhibitory effect on TNF-α, even at the range of concentrations found in healthy people. We tested the hypothesis that levels of sTNFRs in the high-normal range are associated with increased risk of breast cancer.

A nested, case-control study was conducted on 142 white women who were diagnosed with primary breast cancer and who underwent multiphasic health check-ups >30 years ago when enrolled in the Kaiser Permanente Medical Care Program. For each case, one control who matched by age, date of examination, and length of follow-up (early 1990s) was chosen. A subgroup analysis was conducted on the 81 women who were premenopausal at diagnosis. The mean ages (years) and ranges at multiphasic health check-ups and diagnosis for the entire group were 41.5 (19–73) and 55.1 (31–83), and for the subgroup, were 33.9 (19–46) and 44.2 (31–53). These subjects were described in previous studies (6, 7).

The concentration(s) of sTNFR1 was measured in all of the subjects’ sera that were drawn at least 2 years before diagnosis, and TNF-α, sTNFR1, and sTNFR2 were measured in the subgroup sera, and stored at −40°C; each was assayed in duplicate using ELISA kits from R&D Systems (Minneapolis, MN) with intra- and interassay coefficients of variability of 14.3 and 18.7%, 6.9 and 8.8%, and 2.5 and 5.1%, respectively. (The Quantikine High Sensitivity kit was required for TNF-α measurement.) The means, medians, and ranges of each analyte in these archived samples were comparable with the values obtained in our laboratory in serum stored at −20°C for less than 3 months, which indicated good long-term stability. sTNFR2 was measured in 52 pairs because of limited serum volume.

Baseline comparisons were conducted using the Wilcoxon sign-rank test. Age- and BMI-adjusted Pearson coefficients of correlation (r) were obtained by combining case and control values to examine the cross-sectional relationship between natural log (1n) transformed values for TNF-α and the soluble receptors. For multivariate analysis, conditional logistic regression was used to maintain the case-control matching, and pairs with missing BMI data were dropped. This study had 80% power to detect an OR of 2.1 for the overall analysis and an OR of 2.6 for the premenopausal subgroup at α = 0.05, assuming an exposure prevalence of 25% in the control group. All of the statistical tests were two-sided and were completed using Stata v7.0 (College Station, TX).

There were no statistically significant differences in the baseline levels of sTNFR1 between cases and controls (931 ± 287 versus 1028 ± 490 pg/ml) or in those of TNF-α, sTNFR1, and sTNFR2 between premenopausal cases and controls (2.95 ± 2.56 versus 3.13 ± 2.58 pg/ml, 923 ± 329 versus 1038 ± 604 pg/ml, and 2422 ± 899 versus 2458 ± 980 pg/ml, respectively). The cross-sectional, age- and BMI-adjusted, Pearson correlation coefficients between ln(TNF-α) and ln(sTNFR1) (r = 0.42) and ln(TNF-α) and ln(sTNFR2) (r = 0.64) were significant at a P of <0.005. Table 1 presents crude and adjusted ORs relative to the lowest quartiles for TNF-α and each receptor, and shows neither statistically significant changes in risk for breast cancer with increasing serum levels of each protein nor any significant trends. Univariate analysis comparing cases and controls above the median TNF-α level and below the median sTNFR1 level and those of the reverse complement revealed no association with cancer risk (OR, 1.15; 95% CI, 0.36–3.64). Defining the group in this manner in a logistic setting, controlling for age and BMI, also showed no association with case status (OR, 0.67; 95% CI, 0.16–2.68).

This prospective case-control study found no association between breast cancer risk and serum levels of TNF-α or its soluble receptors. Study strengths include the unbiased selection of control subjects and collection of serum samples at least 2 years before diagnosis because concentrations of both of the soluble receptors increase significantly in cancer patients, including breast cancer patients (5, 8).

To our knowledge, this is the first study to examine whether serum levels of the inhibitory sTNFRs affect breast cancer risk, and the results may not apply to the risk of other solid tumors. The potential for a protective role for TNF-α in breast cancer may be more complicated than for such a hypothesized role in non-hormone-dependent types of cancer because TNF-α coordinately stimulates the activities of the enzymes involved in estrogen synthesis in both normal and malignant breast tissue (9). Also, the role of TNF-α in cancer, in general, is paradoxical with some evidence suggesting that its local production enhances, through inflammatory mechanisms, tumor development and spread. Thus, exogenous TNF-α antagonists, including the analogues of sTNFRs that are used to treat chronic inflammatory diseases, might reduce cancer risk (10). This report, however, indicates neither increased nor decreased risk of breast cancer associated with high endogenous levels of sTNFRs.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

2

The abbreviations used are: TNF, tumor necrosis factor; sTNFR, soluble TNF receptor; BMI, body mass index; OR, odds ratio; CI, confidence interval.

Table 1

ORs for breast cancer by quartiles of serum protein measurements

QuartileQuartile 4 95% CIP for trend
1234
Entire group       
 sTNFR1, pg/ml (na = 123)       
  Median 630 846 983 1359   
  Range 340–729 730–910 919–1096 1110–3790   
  Cases/Controls 39/32 35/37 32/38 36/35   
  OR, crude 1.0 0.78 0.69 0.84 0.44–1.64 0.68 
  OR, adjustedb 1.0 0.90 0.57 0.72 0.28–1.82 0.37 
Premenopausal subgroup       
 TNF-α, pg/ml (na = 71)       
  Median 1.30 2.15 3.00 4.43   
  Range 0.53–1.76 1.80–2.49 2.51–3.46 3.49–20.8   
  Cases/Controls 21/20 20/20 19/22 21/19   
  OR, crude 1.0 0.95 0.82 1.05 0.44–2.53 0.94 
  OR, adjustedb 1.0 0.75 0.58 0.60 0.15–2.31 0.45 
 sTNFR1, pg/ml (na = 71)       
  Median 580 757 980 1410   
  Range 340–663 670–869 878–1130 1131–3790   
  Cases/Controls 21/20 18/22 23/18 19/21   
  OR, crude 1.0 0.78 1.22 0.86 0.36–2.07 0.93 
  OR, adjustedb 1.0 0.67 1.09 0.67 0.20–2.25 0.78 
 sTNFR2, pg/ml (na = 46)       
  Median 1455 2070 2830 3660   
  Range 1040–1770 1840–2400 2440–3140 3160–5520   
  Cases/Controls 16/12 12/12 10/16 14/12   
  OR, crude 1.0 0.75 0.47 0.86 0.30–2.59 0.65 
  OR, adjustedb 1.0 0.43 0.23 0.46 0.06–3.50 0.63 
QuartileQuartile 4 95% CIP for trend
1234
Entire group       
 sTNFR1, pg/ml (na = 123)       
  Median 630 846 983 1359   
  Range 340–729 730–910 919–1096 1110–3790   
  Cases/Controls 39/32 35/37 32/38 36/35   
  OR, crude 1.0 0.78 0.69 0.84 0.44–1.64 0.68 
  OR, adjustedb 1.0 0.90 0.57 0.72 0.28–1.82 0.37 
Premenopausal subgroup       
 TNF-α, pg/ml (na = 71)       
  Median 1.30 2.15 3.00 4.43   
  Range 0.53–1.76 1.80–2.49 2.51–3.46 3.49–20.8   
  Cases/Controls 21/20 20/20 19/22 21/19   
  OR, crude 1.0 0.95 0.82 1.05 0.44–2.53 0.94 
  OR, adjustedb 1.0 0.75 0.58 0.60 0.15–2.31 0.45 
 sTNFR1, pg/ml (na = 71)       
  Median 580 757 980 1410   
  Range 340–663 670–869 878–1130 1131–3790   
  Cases/Controls 21/20 18/22 23/18 19/21   
  OR, crude 1.0 0.78 1.22 0.86 0.36–2.07 0.93 
  OR, adjustedb 1.0 0.67 1.09 0.67 0.20–2.25 0.78 
 sTNFR2, pg/ml (na = 46)       
  Median 1455 2070 2830 3660   
  Range 1040–1770 1840–2400 2440–3140 3160–5520   
  Cases/Controls 16/12 12/12 10/16 14/12   
  OR, crude 1.0 0.75 0.47 0.86 0.30–2.59 0.65 
  OR, adjustedb 1.0 0.43 0.23 0.46 0.06–3.50 0.63 
a

n, number of pairs.

b

Adjusted ORs derived from logistic regression analysis, controlling for age and BMI.

We thank Dr. Paul Visintainer for statistical analysis, Dr. Patricia Moorman for the study subjects, Sylvia Duffy for manuscript preparation, and Nancy P. Durr for editing.

1
Bazzoni F., Beutler B. The tumor necrosis factor ligand and receptor families.
N. Engl. J. Med.
,
334
:
1717
-1725,  
1996
.
2
Aderka D., Engelmann H., Shemer-Avni Y., Hornik V., Galil A., Sarov B., Wallach D. Variation in serum levels of the soluble TNF receptors among healthy individuals.
Lymphokine Cytokine Res.
,
11
:
157
-159,  
1992
.
3
Mohamed-Ali V., Goodrick S., Bulmer K., Holly J. M. P., Yudkin J. S., Coppack S. W. Production of soluble tumor necrosis factor receptors by human subcutaneous adipose tissue in vivo.
Am. J. Physiol.
,
277
:
E971
-E975,  
1999
.
4
Hasegawa Y., Sawada M., Ozaki N., Inagaki T., Suzumura A. Increased soluble tumor necrosis factor receptor levels in the serum of elderly people.
Gerontology
,
46
:
185
-188,  
2000
.
5
Aderka D., Engelmann H., Hornik V., Skornick Y., Levo Y., Wallach D., Kushtai G. Increased serum levels of soluble receptors for tumor necrosis factor in cancer patients.
Cancer Res.
,
51
:
5602
-5607,  
1991
.
6
Moorman P. G., Hulka B. S., Hiatt R. A., Krieger N., Newman B., Vogelman J. H., Orentreich N. Association between high-density lipoprotein cholesterol and breast cancer varies by menopausal status.
Cancer Epidemiol. Biomark. Prev.
,
7
:
483
-488,  
1998
.
7
Krajcik R. A., Borofsky N. D., Massardo S., Orentreich N. Insulin-like growth factor I (IGF-I), IGF-binding proteins, and breast cancer.
Cancer Epidemiol. Biomark. Prev.
,
11
:
1566
-1573,  
2002
.
8
Tesarova P., Kvasnicka J., Umlaufova A., Homolkova H., Jirsa M., Tesar V. Soluble TNF and IL-2 receptors in patients with breast cancer.
Med. Sci. Monit.
,
6
:
661
-667,  
2000
.
9
Purohit A., Newman S. P., Reed M. J. The role of cytokines in regulating estrogen synthesis: implications for the etiology of breast cancer.
Breast Cancer Res.
,
4
:
65
-69,  
2002
.
10
Balkwill F. Tumor necrosis factor or tumor promoting factor?.
Cytokine Growth Factor Rev.
,
13
:
135
-141,  
2002
.