Reply

In Response: We thank Farace and coauthors for their letter and comments on our recent article discussing the potential role of vascular endothelial growth factor receptor-3–positive lymphatic/vascular endothelial progenitor cells (LVEPC) in patients with small cell lung cancer (SCLC; ref. 1).

Regarding the methodology of the quantification and characterization of LVEPCs, 900 μL of venous blood (using EDTA as an anticoagulant) was prepared by lysing erythrocytes with 5 mL of nonfixative–containing fluorescence-activated cell sorting lysis solution (lysing of RBC as described was determined not to affect the antigen staining process by the validation work of our group and others; ref. 2). Mononuclear cells were gated based on forward-scatter and side-scatter characteristics. Fluorochrome-matched isotype controls (isotype-identical antibodies were purchased from the same manufacturer and applied in the same concentration as the relevant primary antibodies) were used to set the appropriate gate parameters and served as negative controls. For each sample, a minimum of 100,000 events was acquired. The absolute numbers of CD34+/VEGFR3+ LVEPCs/mL of peripheral blood (PB) were calculated, subtracting the values of the appropriate isotype controls and then multiplying the result by a dilution factor of 10 (i.e., the raw data shown in Fig. 1A. do not represent absolute numbers). In accordance with several other studies on different circulating endothelial progenitors in various malignancies [refs. 2, 3–7; including the recent article by Farace's group in this journal (ref. 8)], no attempts were made to characterize LVEPCs morphologically in our study.

Farace and colleagues refer to the LVEPC numbers of six healthy blood donors reported in the initial study of Salven et al. (9) and compare these levels to the median LVEPC value of our cancer patients (1,625/mL of PB). In our control group, the median value of CD34+/VEGFR3+ LVEPCs was significantly lower (455/mL of PB). However, because this number is still higher than could have been predicted from the results of the original article by Salven et al. (9), it seems obvious that additional studies are needed to reach a consensus on the levels of VEGFR3+ progenitors in normal as well as in different pathologic conditions. Nevertheless, this is also true in the field of hemangiogenic progenitors (i.e., VEGFR2+ circulating endothelial progenitors), where even healthy individual's PB level of CD34+/VEGFR2+ circulating endothelial progenitors varies from ∼100/mL (3, 4, 10) to ∼1000/mL (11) or where the percentages of healthy subject's PB CD34+ cells expressing VEGFR2 range between ∼2% (12) and 27% (13).

Farace et al. argue in their letter that not including performance status (PS) and response to therapy in our multivariate analysis precludes considering the pretreatment LVEPC level in SCLC patients as a prognostic factor. We respectfully disagree on this issue but share their view on the importance of prophylactic cranial irradiation and the timing of thoracic radiotherapy in SCLC. Our study included 88 PS 0-2 patients with limited stage SCLC. Prophylactic cranial irradiation was given to all patients who showed a complete response. In most of our cases, thoracic irradiation and chemotherapy were done concurrently. As shown in the article (Table 2), in our univariate analysis, no significant associations between LVEPC levels and age, smoking history, gender, or tumor stage were detected. More importantly, the difference in pretreatment LVEPC levels between the groups of therapy-sensitive (progression of >3 months after first-line therapy) and therapy-resistant (progression during or within 3 months after first-line therapy; Table 2) cases was not statistically significant. Therefore, response to therapy was not included into our multivariate analysis. Furthermore, although we did not analyze the influence of PS on the survival of our patients, according to a recent study of prognostic factors for limited stage SCLC (14) and an analysis of 14 SCLC trials (15), poorer PS led to worse survival only in extensive stage but not in limited stage SCLC. Thus, we believe that our multivariate analysis of age, gender, T and N stages, and VEGF-C and LVEPC levels is correct and supports the key findings of our study.

Finally, although we are aware that targeted antivascular drugs (including thalidomide; ref. 16) have failed to improve clinical outcomes substantially in SCLC, observations of high capillary densities and increased levels of various (lymph)angiogenic cytokines (17–24) justify further research on the clinical value of such agents in patients with this malignancy. The use of different subtypes of circulating endothelial progenitors as surrogate biomarkers will hopefully provide valuable data in these studies.