The biological activity of tumor necrosis factor (TNF alpha/beta) and interleukin-1 beta (IL-1 beta) can be blocked by soluble, naturally occurring molecules--TNF alpha/beta binding proteins (BP-55 and BP-75), derived from the extracellular portion of the 55- and 75-kDa TNF alpha/beta membrane receptors, and IL-1 receptor antagonist (IL-1ra), respectively. We examined the levels of these cytokines and their inhibitors in cell-free ascites of 18 patients with advanced ovarian carcinoma by ELISA. Levels of both TNF BP and IL-1ra dramatically exceeded those of TNF and IL-1; thus, it is unlikely that these cytokines are active in ascites from patients with this disease. We then elutriated solid tumor samples from three additional patients, yielding pure populations of tumor cells, macrophages, and lymphocytes. Cells were cultured for up to 48 hr and the spontaneous production of TNF, IL-1, and their inhibitors was measured by ELISA. Tumor cells and macrophages both released inhibitors for TNF and IL-1. Tumor cells released IL-1ra and BP-55, while macrophages released IL-1ra and BP-75. Kinetic studies showed that both tumor cells and macrophages produced an initial burst of TNF alpha and IL-1 beta which was overtaken within 48 hr by a sustained production of TNF BP and IL-1ra. Lymphocytes released no TNF alpha or TNF beta, which alone suggests that tumor associated lymphocytes are locally quiescent in vivo. TNF and IL-1 inhibitors originate from tumor cells and tumor associated macrophages and probably block TNF and IL-1 activity locally and regionally in ovarian carcinoma patients. Whether this phenomenon contributes to the pathogenesis of this disease remains to be determined.

The shed portion of the 55 and 75 kDa membrane receptors for tumor necrosis factor (TNF) and lymphotoxin (LT) have been described in the serum of patients with cancer. This study was designed to determine whether serum levels of the 55 and 75 kDa soluble TNF/LT receptors (sTNFr) had clinical significance in patients with gynecologic malignancies. Serum samples from 79 patients with ovarian, endometrial, or cervical cancer were assayed for CA 125 levels by RIA and the 55 and 75 kDa sTNFr levels by ELISA. Receptor and CA 125 levels were also analyzed with respect to disease status and response to treatment in banked serum samples from 14 patients with epithelial ovarian cancer who had been followed clinically for 1-3 years. Patients resulted were compared to serum samples tested from normal donors. We found that serum levels of both sTNFr's were elevated in the 79 patients with various gynecologic malignancies [55 kDa of 3.07 +/- 3.79 ng/ml (P < 0.02) and 75 kDa of 2.93 +/- 1.27 ng/ml (P < 0.001)] compared to 16 normal controls (55 kDa of 0.65 +/- 0.22 ng/ml and 75 kDa of 1.62 +/- 0.37 ng/ml). Serum levels of 55 and 75 kDa TNF/LT receptors were a more sensitive indicator of active cancer and had greater predictive value for detecting tumor in patients with ovarian cancer than CA 125. The sTNFr's were also more sensitive than CA 125 in detecting persistent or recurrent tumor and measuring response to therapy. These preliminary results suggest that measurement of serum levels of 55 and 75 kDa sTNFr's, even though not tumor specific, may be a uniquely new method for identifying and monitoring patients with gynecologic malignancy.

Background

A prospective study was undertaken within the Gynecologic Oncology Group to determine whether serum levels of soluble tumor necrosis factor receptors I (sTNFR-I) and II (sTNFR-II), alone or in combination with CA 125, were associated with clinicopathologic characteristics or outcome in patients with epithelial ovarian malignancies.

Methods

Quantitative immunoassays were performed on valid pretreatment serum specimens obtained from patients with epithelial ovarian malignancies to assess levels of sTNFR-I, sTNFR-II, and CA 125. The authors then analyzed the results of these immunoassays for potential correlations with clinicopathologic characteristics and outcome.

Results

The median age of the 139 women evaluated was 59 years. Seventy-eight percent had Stage III or IV disease, and 58% had serous carcinomas. sTNFR-II was associated with age (P = 0.013), and CA 125 was associated with histologic subtype (P = 0.0009). In addition, sTNFR-I (P = 0.037) and CA 125 (P < 0.0001) were associated with extent of disease. After adjusting for patient age, histologic subtype, and extent of disease, all three biomarkers were predictive of progression-free survival, but not overall survival, when the combination was included in the model. The authors observed a 51% reduction (hazard ratio [HR], 0.49; 95% confidence interval [CI], 0.24-0.99), a 2.9-fold increase (HR, 2.87; 95% CI, 1.15-7.20), and a 22% increase (HR, 1.22; 95% CI, 0.99-1.51) in the risk of progression for each unit increase in the log-transformed levels of sTNFR-I, sTNFR-II, and CA 125, respectively.

Conclusions

The observations made in the current study-that among patients with low or high CA 125 levels, those with high sTNFR-I levels and low sTNFR-II levels had the lowest risk, that patients with low-low or high-high sTNFR-I and sTNFR-II levels, respectively, had an intermediate risk, and that patients with low sTNFR-I levels and high sTNFR-II levels had the highest risk of progression-suggested the potential value of simultaneous assessment of all three biomarkers in patients with epithelial ovarian malignancies.