Serum Concentrations of IL-2 and TNF-α in Patients with Painful Bone Metastases: Correlation with Responses to ⁸⁹SrCl₂ Therapy

MATERIALS AND METHODS

RESULTS

No immediate adverse reactions were noted after the ⁸⁹SrCl₂ injection. Twenty-one percent of patients experienced a mild, transient worsening of symptoms (pain flare phenomenon) that began 1 d after the treatment, lasted 2−4 d, and did not require additional analgesics. The pain flare phenomenon had no correlation with primary tumor type, age, number of bone metastases, or sex. The palliative effect usually started around the second or third week after the injection. No side effects were reported in tissues other than bone marrow after ⁸⁹SrCl₂ administration, and no patients required a blood transfusion for hematologic depression. A decrease in the initial white blood cell and platelet counts began during the third or fourth week after the injection. Three months after the injection, white blood cells and platelets had decreased in 21% and 27%, respectively, of patients. And later, both white blood cell counts and platelet counts tended to return to their pretherapeutic values, with a gradual partial-to-complete recovery within 6 mo. Grade I toxicity was about 19% for platelets and 16% for white blood cells. Grade II was about 8% for platelets and 5% for white blood cells. No patients experienced severe toxicity (grade III or IV).

The serum IL-2 concentrations of all 42 patients were lower than those of the control subjects, whereas TNF-α concentrations were higher in the patients (Table 2). After ⁸⁹SrCl₂ therapy, serum concentrations of IL-2 increased, peaking at the end of the fourth month and then gradually decreasing over the following 2 mo. In contrast, serum concentrations of TNF-α decreased after the therapy, reaching a minimum at the end of the fourth month and then gradually recovering to baseline values over the following 2 mo. A good inverse correlation was found between IL-2 levels and the number of metastatic lesions and pain score, and a direct correlation was found between TNF-α levels and the number of metastatic lesions and pain score (Table 3).

We categorized each patient as a responder or a nonresponder according to pain score. No significant differences in age, sex, or type of primary tumor were found between responders and nonresponders (Table 4). In responders, serum IL-2 concentrations increased from 2.91 ± 0.56 to 4.58 ± 1.14 (P < 0.001), whereas serum TNF-α concentrations decreased from 2.78 ± 0.37 to 1.24 ± 0.55 (P < 0.001). In contrast, no significant changes in IL-2 and TNF-α concentrations were observed in nonresponders (P = 0.565 and 0.542, respectively). The differences in serum IL-2 and TNF-α concentrations between responders and nonresponders were not obvious before therapy (P = 0.519 and 0.178, respectively) but became statistically significant after therapy (P < 0.001) (Table 5).

Using a fall in TNF-α levels or a rise in IL-2 levels of greater than 25%, compared with the baseline value, we found the sensitivity and specificity of IL-2 increase to be 87.9% and 77.8%, respectively, and the sensitivity and specificity of TNF-α decrease to be 81.8% and 66.7%, respectively (Tables 6 and 7). Patients with an increasing IL-2 level and a decreasing TNF-α level had a better prognosis.

DISCUSSION

IL-2 is a cytokine released from T helper lymphocytes. It promotes the generation, proliferation, and differentiation of T lymphocytes; enhances the activity of natural killer cells; induces the generation of lymphokine-activated killer cells; and promotes the production of antibodies by B lymphocytes. Through these mechanisms, it plays an important role in antitumor immune responses (13). Pretreatment serum IL-2 levels have been shown to be of independent prognostic value in patients with advanced non–small cell lung cancer (14). A study by Fischer et al. also indicated that IL-2 secretion correlates with long-term survival in patients with small cell lung cancer (15).

TNF-α, another important cytokine in anticancer therapy, is produced primarily by mononuclear macrophages. TNF-α can initiate an intensive immunoinflammation response, induce natural killer cells and macrophagocytes, and produce carcinolysis (16). TNF-α also can damage vascular endotheliocytes and cause thrombosis or hemorrhage, resulting in tumor necrosis or resolution (17), and inhibit tumor cell proliferation by inducing cell apoptosis (18). Most patients with malignant tumors express high levels of TNF-α, and serum concentrations of TNF-α have shown a correlation with tumor burden and progression (19). Increases in TNF-α probably originate from autosecretion by tumor cells, tumor-infiltrating lymphocytes stimulated by tumor antigen, and circulating monocytes activated by tumor metastases (20). This possibility would be in line with recent findings that, although appropriate concentrations of TNF-α in serum increase immunologic response and inhibit the development of tumors, high levels of TNF-α can paradoxically promote the ability of a tumor to become aggressive and metastasize. TNF-α has been shown to increase adhesion between tumor cells and endotheliocytes, accelerate maturation of the tumor matrix, and increase the gene expression of stromal metalloproteinase (19,21).

The 79% response rate found for ⁸⁹SrCl₂ by our study is similar to response rates found by other studies (4,22). The measured cytokines fluctuated after ⁸⁹SrCl₂ therapy, presumably because ⁸⁹SrCl₂ therapy improves immunologic function by killing tumor cells. After 4 mo, we found cytokine levels reflecting the pretherapeutic findings; perhaps responding patients should receive a second dose then.

At present, the pain score and the number of·metastatic lesions seen on a whole-body bone scan are commonly used to evaluate the effectiveness of radiopharmaceuticals. Although palliation of pain from bone metastases can reflect the status of tumors to some extent, the pain score is influenced by many other factors, including the patient's age, threshold of pain, mentality, and use of analgesics (23). To be meaningful, the pain score must be modified according to analgesic consumption and daily activities. Bone scintigraphy is not an optimal method of following tumor response, because patients who are in end-stage disease and physically weakened can find it difficult to undergo periodic whole-body scanning. Whether bone metastases worsen and spread or the bone begins to repair, the bone scan may show a greater intensity of focal uptake.

In the present study, we found that serum IL-2 and TNF-α levels correlated well with the number of bone metastatic lesions and pain score. Differences between pretherapeutic and posttherapeutic levels of IL-2 and TNF-α were significant in responders but not in nonresponders. In responders, these variables did not differ before ⁸⁹SrCl₂ therapy but differed significantly after therapy. These results demonstrate that serum concentrations of IL-2 and TNF-α are a useful indicator of the response to ⁸⁹SrCl₂ therapy in patients with bone metastases.

In summary, ⁸⁹SrCl₂ therapy improves levels of measured cytokines, presumably by killing bone metastases. Repeating therapy properly might help maintain relatively normal immunity and increase the chance of survival, although no study has yet shown ⁸⁹SrCl₂ therapy to increase survival. The combination of TNF-α and IL-2 measurement and pain scoring may help in monitoring the therapeutic effects of ⁸⁹SrCl₂.