With great interest we read the article Two distinct expression patterns of urokinase, urokinase receptor and plasminogen activator inhibitor-1 (PAI1) in colon cancer liver metastases’ by Martin Illemann et al. The authors describe 2 distinct patterns of the expression of urokinase-type plasminogen activator (uPA), urokinase-type plasminogen activator receptor (uPAR), and PAI1 in 14 colon cancer liver metastases that correlate closely with 2 morphological growth patterns. In the liver metastases with a predominantly desmoplastic reaction at the periphery, the expression of uPAR, uPA and PAI1 is very similar to that found in the primary colon tumours. These molecules were primarily expressed in stromal cells at the metastases’ periphery. Furthermore, uPAR and uPA were also expressed in a subset of budding cancer cells at the tumour interface. In the liver metastases of which the growth is characterised by direct contact between the cancer cells and the hepatocytes on the contrary, uPAR and uPA-mRNA expression were only seen associated with the presence of necrosis within the liver metastases. In addition, PAI1-immunoreactivity was in all liver metastases seen in hepatocytes at the metastases periphery. We have been studying growth patterns of human colorectal cancer metastases for many years now. As mentioned by Illeman et al., we described 3 different growth patterns with differences in angiogenesis and desmoplasia. In the desmoplastic growth pattern, the metastases are separated from the surrounding liver parenchyma by a rim of desmoplastic stroma in which a dense inflammatory infiltrate and numerous capillaries are present. In the pushing pattern, liver cell plates are pushed aside and run in parallel with the circumference of the metastases at the tumour–liver parenchyma interface. There is no desmoplastic stroma formation. In the nonangiogenic replacement growth pattern, tumour cells replace hepatocytes in the liver plates, at the interface or throughout the metastasis, conserving tissue architecture of the liver, without inflammation or fibrosis. The growth pattern of liver metastases seems not only to vary between tumours of the same histological subtype, but also between tumours of a different subtype. Although liver metastases of patients with colorectal cancer grow according to the desmoplastic , pushing and replacement growth pattern in, respectively, 42, 46 and 12%, liver metastases of patients with breast cancer on the contrary grow according to the replacement growth pattern in virtually all cases. Although we have published data that show important differences in hypoxia and angiogenesis between the different growth patterns, the biology and underlying molecular mechanisms of the different growth pattern of liver metastases remain largely unknown. Therefore, the findings of Illeman et al. regarding different expression of 3 key components of the uPA extracellular protease system in metastases with a desmoplastic and pushing growth pattern are important. In this letter, we would like to add extra data suggesting a role not only in liver metastases of patients with colorectal cancer, but also of patients with breast cancer. In a study of differential expression of hypoxia and (lymph)angiogenesis-related genes at different metastatic sites in breast cancer, uPA (in that paper called PLAU) was among the 49 genes that were differentially expressed between primary tumours and/or lymph node metastases and/or liver metastases. The expression of uPA was higher in liver metastases compared with primary breast tumours and to lymph node metastases. There was no significant difference between the expression of uPA between primary breast tumours and lymph node metastases. uPAR and PAI1 were not investigated in that study. To further explore possible differences in expression of the uPA system between breast cancer localisations in different organs, we now analysed the expression of uPA, uPAR and PAI1 using the data described in that paper. The localisations were grouped in primary tumours (n 5 11), lymph node metastases (n 5 15), liver metastases (n 5 4), lung metastases (n 5 8) and other metastases (n 5 24). For a detailed description of materials and methods, we refer to the previous publication. Figure 1 represents the results. The median mRNA expression for uPA was significantly lower in liver metastases compared with primary tumours (p 5 0.04) and lymph node metastases (p 5 0.04). There was a borderline significant difference in uPA mRNA expression between liver metastases and other metastases (p 5 0.09), and there was no difference between liver metastases and lung metastases. There was a significantly higher mRNA expression of PAI1 in liver metastases than in lymph node metastases (p 5 0.04) and a borderline significant higher expression in liver metastases than in primary tumours (p 5 0.08). There were no significant differences in the expression of uPAR between liver metastases and other tumour sites. Furthermore, there was a significantly higher PAI1 expression in lung metastases, than in primary tumours (p 5 0.009), lymph node metastases (p 5 0.002) and other metastases (p 5 0.03) and there was a