C-reactive protein (CRP) intrigues clinicians from varied specialties. It is readily measured, changes rapidly in response to stress, and correlates with meaningful clinical end points. Serum or plasma CRP concentrations are measured to monitor disease activity and response to therapies in a variety of rheumatologic conditions.[2, 3] Measurement of CRP concentrations using highly sensitive assays can aid in estimation of the risk of cardiovascular disease.
Elevated CRP concentrations have been linked to decreased survival in multiple types of cancer over the past 2 decades. Five years ago, our group published findings linking higher CRP concentrations with decreased overall survival in prostate cancer patients receiving docetaxel-based chemotherapy. The association of CRP with survival was stronger than all other predictors, including serum prostate-specific antigen, Eastern Cooperative Oncology Group performance status, and age. These 2 findings in men with metastatic, castration-resistant prostate cancer were confirmed using two additional retrospective analyses: one by our group and another by Ito et al. One of these analyses also suggested that CRP levels are inversely associated with the probability of response to docetaxel.
Although intriguing, these findings have not been prospectively confirmed, and they lead to several additional questions. First, because CRP concentrations have been associated with prognosis in noncancer conditions, it is unclear whether the associations between elevated CRP and overall survival reflect the impact of cancer-related deaths or could be explained by competing causes of mortality. Second, to the extent that CRP is associated with worse outcomes in cancer patients, it is unclear whether CRP is merely a marker of disease and/or host factors or if CRP itself contributes to the disease process or therapeutic resistance. Finally, because CRP is a marker of inflammation, and inflammation is associated with cancer development and progression, it would be of interest to understand whether CRP reflects a host inflammatory state that promotes the initiation or progression of the neoplastic process and/or mediates resistance to therapy.
In this issue of Cancer, the association between serum CRP concentrations and relapse-free survival is examined in prostate cancer patients who received radiation therapy for localized disease, either in addition to radical prostatectomy or as definitive treatment. This work adds to our body of knowledge because it focuses on clinically localized prostate cancer rather than the more commonly examined metastatic disease and because the end point of interest is cancer-specific. The evaluation of cancer-specific end points has the potential to isolate the relationship between CRP and cancer and eliminate the confounding impact of cardiovascular and other inflammatory conditions.
Hall et al showed a statistically significant association between elevated CRP and shortened biochemical relapse-free survival on both univariate and multivariate analyses in men treated definitively with radiation therapy. They also showed that CRP was associated with prostate-specific antigen in subsets with intermediate and high-risk prostate cancer.
We congratulate Hall et al for their exciting observations and join them in noting some of the limitations of the analysis. There a number of methodological limitations that do not diminish the findings but classify them as preliminary and hypothesis-generating rather than definitive. The retrospective nature of the analysis; the heterogeneity of the patient population, treatments administered, and timing of CRP evaluation; and the reasons for CRP measurement all introduce a measure of uncertainty about the reproducibility of the findings.
One particular issue we would like to mention is the fact that CRP was measured after the initial treatment. We are unsure how one might interpret CRP measurements in this setting. They could represent a measure of systemic inflammation that is long-standing and unrelated to cancer and reflects host factors that could contribute to the behavior of the neoplastic process. One could speculate that, to some extent, these measurements could also reflect the effect of radiation itself. Little is known about the effect of radiation for prostate cancer on circulating CRP levels. Small analyses in other cancers suggest that some inflammatory markers do change after radiation therapy, but CRP may not. Brachytherapy but not external beam therapy increased serum levels of interleukin-6 (IL-6) in a small study of cervical cancer patients. In breast cancer patients embarking on adjuvant radiation therapy, CRP was elevated when compared with healthy controls and remained elevated after completion of radiation therapy. Surfactant-protein D, another inflammatory marker, did increase with adjuvant radiation. Therefore, we do not know if the CRP concentrations measured in this study were affected by radiation.
Finally, we cannot be certain that CRP levels are not elevated as a direct result of persistent but subclinical cancer. If the CRP elevations observed in this study were a reflection of the presence of subclinical quantities of cancer, rather than a measure of unrelated host inflammation, CRP could be acting as a sensitive tumor marker. Although distinguishing between host and tumor-related cause of CRP elevation is a difficult challenge, measurement of CRP prior to any treatment and then serial measurement over time might enable us to gain some insights into these distinctions.
The second question we pose—whether CRP is merely a marker or an actor in the disease process—remains unanswered. CRP does have biologic functions. It has been shown to activate the complement system. A role for CRP in prostate cancer progression or resistance to therapy, if such a role exists, has not been discovered. Using reverse-transcription polymerase chain reaction, both our group (unpublished data) and Elsberger et al have found no evidence of CRP production by the prostate tumors, but both of our groups showed the presence of CRP within the prostate cancer tumor via immunohistochemistry. Elsberger et al examined matched pairs of tumors taken from men at diagnosis and time of the development of castration resistance for levels of intratumoral CRP. They described an association between the presence of cytoplasmic CRP in hormone-naïve tumors and the presence of metastatic disease at diagnosis. They also described an association between nuclear CRP and metastatic disease at the time of relapse. There was a trend toward increased CRP in both cytoplasmic and nuclear compartments during progression to castration resistance. These findings have not been widely confirmed, and a biologic role for intracellular CRP in prostate cancer has not been clarified.
If CRP did play a specific role in prostate cancer progression and/or resistance, direct inhibitors of CRP might make for attractive therapeutic agents. To date, no such agents have been developed; instead, clinical trials have focused on targeting inflammation in other ways.
CNTO328, a monoclonal antibody against IL-6, has been studied in combination with chemotherapy and as a single agent. Although these phase 2 studies were not designed to yield definitive results, neither use in combination with mitoxantrone nor single-agent treatment with CNTO328 appeared to improve clinically important end points, despite the fact that CNTO328 led to expected decreases in IL-6 and CRP. The initial report from the STAMPEDE trial indicates the addition of celecoxib to hormonal therapy in patients with metastatic or high-risk localized prostate cancer receiving hormonal therapy was not beneficial as well. There are other examples of targeting inflammation in clinical trials for prostate cancer, and thus far the results have not been encouraging—with the exception of steroids, which have a modest but reproducible level of activity in advanced prostate cancer. It would be fair to say, however, that efforts to date have been constrained by limitations in our understanding of the role of inflammation in prostate cancer progression and treatment resistance. Absent a deeper understanding of this topic, our approach has been largely to test the immunomodulatory tools that we have rather than to design new tools specifically crafted to address cancer-relevant elements of inflammation.
There is a hypothesis that CRP is a marker of host inflammation that renders that host particularly hospitable to the development of cancer, but the evidence is not strong. Although an inflammatory infiltrate is commonly associated with prostate cancer and many other tumor types, it is not clear whether CRP measurement in the blood can tell us useful things about this process. The Physicians' Health Study examined quartiles of CRP in men at enrollment, at which time participants with known malignancy would have been excluded, and examined the incidence of prostate cancer. For the group as a whole, there was no association between elevated CRP and subsequent prostate cancer development. However, when examining only healthy weight men, there was an association. In a separate analysis, a Swedish group used a large national database to examine the relationship between elevated CRP levels and subsequent development of prostate cancer and found that there was no association. This does not mean that inflammation has no role in prostate cancer development. A history of prostatitis is linked to a higher risk of subsequent prostate cancer development. These results suggest that the localized inflammation that may promote the development of some prostate cancers is not reliably detected through measurement of circulating CRP. These data also suggest that the elevation in CRP observed in studies of cancer patients may occur after prostate cancer development and may reflect processes that occur in the presence of an established tumor.
As the authors point out, all of the studies of CRP in prostate cancer are preliminary and retrospective. In the metastatic disease setting, there is an effort to prospectively validate CRP as a marker of outcome in advanced prostate cancer. CRP has been prospectively built into the AFFINITY trial, which is evaluating the clinical benefit of cabazitaxel alone or in combination with custirsen in metastatic castration-resistant, docetaxel-pretreated patients (NCT01578655). We would welcome confirmatory prospective studies in localized prostate cancer.
In conclusion, we applaud the authors for describing the thought-provoking association between elevated CRP and decreased time to biochemical relapse in men with clinically localized prostate cancer undergoing radiation therapy. A prospective study that includes pretreatment CRP measurement and synchronized posttreatment measurements would help elucidate the relationship between prostate cancer aggressiveness and CRP levels and would confirm the findings of this retrospective analysis.