Figure S5: A correlation plot comparing the protein expression in seven independent cultures of our MSCs derived from a healthy donor. microenvironment, but patients differ both with regard to the number and amount of released proteins. Inhibition of this bidirectional communication through protein release between AML cells and leukemia-supporting normal cells may become a new strategy for cancer treatment. Abstract Extracellular protein release is important both for the formation of extracellular matrix and for communication between cells. We investigated the SBC-110736 extracellular protein release by in vitro cultured normal mesenchymal stem cells (MSCs) and by primary human acute myeloid leukemia (AML) cells derived from 40 consecutive patients. We observed quantifiable levels of 3082 proteins in our study; for the MSCs, we detected 1446 proteins, whereas the number of released proteins for the AML cells showed wide variation between patients (average number 1699, range 557C2380). The proteins were derived from various cellular compartments (e.g., cell membrane, nucleus, and cytoplasms), several organelles (e.g., cytoskeleton, endoplasmatic reticulum, Golgi apparatus, and mitochondria) and had various functions (e.g., extracellular matrix and exosomal proteins, cytokines, soluble adhesion molecules, protein synthesis, post-transcriptional modulation, RNA binding, and ribonuclear proteins). Thus, AML patients were very heterogeneous both regarding the number of proteins and the nature of their extracellularly released proteins. The protein release profiles of MSCs and primary AML cells show a considerable overlap, but a minority of the proteins are released only or mainly by the MSC, including several extracellular matrix molecules. Taken together, our observations suggest that the protein profile of the extracellular bone marrow microenvironment differs between AML patients, these differences are mainly caused by the protein release by the leukemic cells but this leukemia-associated heterogeneity of the overall extracellular protein profile is modulated by the constitutive protein release by normal MSCs. = 40)= 0.0001); patients in the lower brown main cluster generally released higher numbers of proteins (cluster 2; 23 patients, median number 2030 proteins, range 1686C2380) compared with the patients in the upper yellow main cluster (cluster 1; 17 patients, median number SBC-110736 1282, range 557C1864, Wilcoxons test, = 0.0034). A statistical analysis based on the proteins in the two main clusters (Figure 4; main yellow/cluster 1 and brown/cluster 2) resulted in 144 proteins with significantly different protein abundance, using Welchs = 0.0455), the patients in the bright yellow upper subcluster showing reduced release especially of proteins in the middle right brown protein clusters (exosomal/extracellular and nuclear proteins). In contrast, the patients in the lower yellow subcluster showed a generally reduced number of released extracellular proteins, and this patient subcluster also included a significantly increased number of patients with secondary AML (CMML or MDS; Fishers exact test, = 0.0197). As would then be expected this cluster also included Rabbit polyclonal to CREB1 a larger fraction of patients above 70 years of age (7/8 versus 14/18; Fishers SBC-110736 exact test, = 0.0461); younger patients were especially seen in the lower dark brown cluster. Furthermore, the patient clusters did not differ significantly with regard to gender, differentiation (FAB classification/CD34 expression) karyotype or Flt3/NPM1 abnormalities. Finally, the patients in the lowest dark brown subcluster showed high levels of proteins in the middle yellow protein cluster that especially included exosomal/vesicular/extracellular proteins (e.g., collagen trimer and endoplasmic reticulum proteins). Thus, AML patients can be further subclassified into distinct subsets based on the number and nature of constitutively released proteins by their AML cells, and these differences correspond to differences in important functional cell characteristics. Previous studies have shown that AML patients differ with regard to the proliferative capacity and the degree of spontaneous apoptosis during in vitro culture of their leukemic cells [22,23]. However, the proliferative capacity after seven days of in vitro culture and the percentage of viable cells after two days of culture did not differ significantly between the two main patient clusters or between the subclusters (see Figure 4 and Figure 6) identified by the cluster analysis presented in Figure 4. Thus, even though protein release by necrotic/apoptotic cells during in vitro culture may contribute to the extracellular protein release of our cultured AML cells, such differences seem to be relatively small and cannot explain the patient subset classification identified in Figure 4. Open in a separate window Figure 6 Identification of AML patient subsets by hierarchical clustering analysis of constitutive leukemic cell protein release. The AML cells were derived from 40 patients, and the cells were cultured alone for 48 h in serum-free medium before supernatants.