TNF-blockade in patients with advanced hormone refractory prostate cancer

Abstract

Interleukin-6 (IL-6) levels are universally elevated in advanced prostate cancer and these levels increase in proportion to tumor burden and decrease with therapeutic response [1–4]. Despite this intriguing relationship, a role for IL-6 in prostate cancer progression is unclear. Several preclinical studies suggest that IL-6 may be an anti-apoptotic agent and a key growth factor in hormone-refractory prostate cancer (HRPC) [5–7]. Furthermore, anti-IL-6 antibodies can induce apoptosis and tumor regression in mice bearing human hormone-refractory prostate cancer xenografts [8–10]. We hypothesized that TNFα-inhibition may block IL-6 production and result in a measurable clinical response in patients with bone metastases and HRPC. A pilot study was initiated to examine the effects of TNFα-blockade (infliximab, Centocor) in patients with treatment-refractory HRPC and painful bone metastases. Six patients with HRPC and painful bone metastases were enrolled and treated with infliximab 5 mg/kg at 0, 2, 6 and 12 weeks. Two of the six patients had a transient but complete response in pain that lasted between 2–5 days. Biochemically, IL-6 levels declined with pain resolution. The clinical or biochemical effect of TNFα-blockade was transient and could not be reproduced after 3–4 weeks of treatment and IL-6 levels increased thereafter. The remaining patients’ were refractory to any clinical benefit in pain from infliximab and all showed an increase in IL-6 throughout the course of treatment. (Table 1) All patients withdrew from the study after there was radiographic evidence of disease progression. No treatment-related adverse events were reported. Table 1 Patient characteristics The observations from this study were unexpected but may allow us to distinguish the source of IL-6 in these patients. It remains unclear whether IL-6 is directly produced by the classic inflammatory cascade (TNFα dependent) within a bone metastasis, or if the tumor cells produce the majority of the IL-6 in HRPC. TNFα-blockade should at least, in part, inhibit IL-6 production and alleviate pain related to such inflammation. In our patients, we only witnessed pain relief in two subjects. Although this clinical benefit was associated with a drop in IL-6 levels, both the clinical and IL-6 responses were transient. In addition, those patients without a clinical benefit demonstrated an increase in IL-6 levels when treated with infliximab. From these results, it is clear that IL-6 expression in HRPC is not TNFα-dependent. (figure 1) Figure 1 Representative subjects and IL-6 (pg/mL) levels in patients with a clinical response (pain resolution) and without a clinical response. The pain response and concomitant IL-6 suppression does suggest that there was an initial, albeit transient, TNFα-dependent component in two patients. This rapid development of resistance in these two patients and the reactive increase in IL-6 levels in the remaining patients suggest an alternative pathway of IL-6 expression that is not TNFα-dependent. Supporting the concept that IL-6 expression in HRPC may be tumor in origin rather than a purely reactive inflammatory process. Its function as an autocrine growth factor in hormone-refractory disease has been postulated [8,11] and blocking the IL-6 signal does lead to apoptotic death and growth suppression in pre-clinical models [8–10]. Addiction to growth factor signals is not a novel concept in Oncology and dependence on IL-6 could clearly play a role in prostate cancer growth, metastasis and progression [12]. In summary, we have observed that IL-6 expression is not TNFα-dependent in patients with painful bone metastases in HRPC. While it appears that pain is associated with fluctuations in IL-6 levels and TNFα-blockade with infliximab is safe in HRPC, it is not a useful therapeutic option. Direct blockade of IL-6 may be needed for therapeutic efficacy.