Ikeda et al. remark that platelets are a main source of lysophosphatidic acid (LPA) and therefore the interpretation of LPA serum concentrations deserves careful attention. However, the same authors previously reported1 an inverse correlation between plasma LPA concentrations and the number of platelets in patients with chronic C hepatitis. Therefore, it is possible that in physiologic conditions platelets remain the main source of LPA, while in chronic inflammation such as hepatitis C, liver cirrhosis, or hepatocellular carcinoma (HCC), the platelet contribution to LPA production may likely become less relevant. In our study we analyzed sera for LPA detection in healthy donors, liver cirrhosis, and HCC patients, performing well-standardized procedures of collection for each sample. Thus, the contribution of platelets to the LPA concentration was, in reality, normalized. On the contrary, the authors should consider that even in plasma or whole blood, platelet activation is an extremely difficult problem to deal with and control. For example, prolonged tourniquet application, or twisting of the needle in the vein, are major factors interfering with the function of platelets during blood withdrawal, as reviewed by Ruggeri.2 Unfortunately, these limitations are common for a number of molecules involved both in cancer and in blood cell biology.3
Moreover, Ikeda et al. investigated patients with chronic hepatitis C, in whom the inflammatory response is a key component of the tissue microenvironment. In their study, the fibrotic status was also questionable, due to their choice of statistical method (comparison among groups should be done with Kruskal-Wallis tests), and because of the very limited number of patients (14), further stratified into four different groups, which means the conclusions were affected by low power.1 In our study,4 we compared liver cirrhosis versus HCC. In the former case, the inflammation is reduced while the fibrotic response is increased, consequently inducing a different microenvironment response.5 This could explain why patients with liver cirrhosis display relatively low levels of LPA. In addition, it is conceivable that when HCC develops in cirrhotic liver, LPA levels rise once more, as in cases of active inflammatory states (i.e., viral hepatitis). Another key point is patient selection. Ikeda et al. do not provide any information with regard to the clinical features of the patients, i.e., etiology, BCLC stage, previous therapy, etc., as well as how they calculated the size of the tumor in patients with multifocal disease, for instance. Finally, some differences between Caucasian and Asian patients with HCC are to be expected, since the natural history is completely different in Western and Southeast Asian countries.6 In our study,4 we demonstrated that LPA has a role in promoting tumor progression and we did not attempt to speculate about the use of LPA as a clinical biomarker. To validate LPA as a potential biomarker for HCC a different study design is required, as well as first considering the power of the study. The enhancement of serum LPA levels reported by Watanabe et al.1 referred to a relatively small number of patients with chronic hepatitis C. In addition, the enzymatic cycling method to detect LPA used in their study is very different from the enzyme-linked immunosorbent assay (ELISA)-based method used in our study. With regard to our data on LPA levels in healthy donors, we effectively detected serum levels 10-fold those detected in plasma. Nevertheless, the values of undiluted human serum samples are perfectly in line with the values reported in the datasheet provided by the manufacturer of the LPA assay kit (Echelon Biosciences) that we used in our study. Moreover, increased LPA concentrations have previously been reported in the serum of patients with HCC.7
In addition, the cholestatic pruritus mentioned by Ikeda et al. occurs in patients with cholestatic disorders, which are quite different from HCC and therefore not comparable. Finally, we investigated 60 patients with HCC and 50 with liver cirrhosis, whereas the number of patients investigated by Ikeda et al. in their letter (21 with HCC and 15 with chronic hepatitis C) is too limited to draw reliable conclusions. In conclusion, we do not see a conflict between our data and those presented in the letter, and we also stand by our hypothesis, supported by other reports in the literature.