Article Figures & Data
Figures
- Figure 1. Measuring UbL-conjugating activities in acute myeloid leukemias cell extracts using ProtoArrays.
(A) Flowchart followed for UbL signature characterization. Extracts from parental and chemoresistant HL-60 or U937 cells were first supplemented with recombinant UbLs (to avoid rate-limiting amounts of the modifiers) and UbL-vinyl-sulfones (to inhibit UbL deconjugating activities). They were then incubated with protein ProtoArrays. After extensive washes, the arrays were incubated with, first, a primary mouse anti–SUMO-1 antibody and a rabbit anti-Flag antiserum recognizing the Flag-tag present on the recombinant ubiquitin added to the reaction and, then, appropriate fluorescent secondary antibodies. Fluorescence signals were processed using the PAA R package. The statistical analysis was performed to identify a UbL signature of chemoresistance, as described in the Materials and Methods section. Three independent experiments were performed for each cell line. (B) IC50 of chemosensitive and chemoresistant acute myeloid leukemia cell lines. IC50 of chemosensitive parental HL-60 and U937 (wt) cells and of their resistant counterparts (see the Materials and Methods section) (ARA-R and DNR-R) was assayed after 24 h of exposure to drugs. n = 3, Mean ± SEM, paired t test with * corresponding to P < 0.05. (C) Identification of ubiquitylated- and SUMOylated proteins. Normalized Ub and SUMO-1 fluorescence data obtained on all arrays incubated with extracts were compared with averaged signals on control arrays (extracts supplemented with NEM to inhibit UbL conjugation activities) to identify robustly UbL-conjugated proteins. Proteins showing significant differences between the two groups when using both the Welch- and the Wilcoxon–Mann–Whitney statistical tests and having mean fluorescence intensity values higher than 800 (arbitrary threshold) on ProtoArrays were selected for further analysis. The Venn diagram shows the number or proteins identified as modified by SUMO-1 and/or ubiquitin.
- Figure S1. Identification of ubiquitylated- and SUMOylated proteins.
Normalized Ub and SUMO-1 fluorescence data obtained on all arrays incubated with extracts were compared with averaged signals on control arrays (extracts supplemented with NEM to inhibit UbL conjugation activities) to identify robustly UbL-conjugated proteins. The names of selected proteins, in particular those found to be differentially modified (see Fig 2) are indicated.
- Figure 2. UbL-conjugated protein signature of chemoresistance.
(A) Identification of UbL-modified biomarkers of acute myeloid leukemias chemoresistance. Modification levels of the proteins modified by ubiquitin (left panel) or SUMO-1 (central panel) selected in Fig 1C were compared between all parental (U937 and HL-60) and drug-resistant (ARA-R or DNR-R) sublines. Differentially modified proteins with significant P-values in both Wilcoxon signed rank- and one sample t test and with a drug-resistant versus parental cell ratio higher than 1.25 or lower than 0.8 are indicated in red for ubiquitylated proteins and in blue for SUMOylated ones. The Venn diagram shows the overlap between differentially ubiquitylated- and SUMOylated proteins. (B) Identification of UbL-conjugated biomarkers specific for HL-60 and U937 cell resistance to Ara-C or DNR. Statistical analyses between drug-resistant and parental cells were performed separately for U937 and HL-60 cell lines and for each drug resistance. The number of proteins showing a significant P-value in one sample t test and a ratio between drug-resistant and parental cells higher than 1.5, or lower than 0.66, are shown. (C) Ontology analysis of the UbL signature. An ontology analysis of the 122 proteins of the UbL signature was performed using the Panther software.
- Figure S2. Ubiquitin-conjugated protein signature of chemoresistance.
Modification levels of the proteins modified by ubiquitin selected in Fig 1C were compared between all parental (U937 and HL-60) and drug-resistant (ARA-R or DNR-R) sublines. Differentially modified proteins with a drug-resistant versus parental cell ratio higher than 1.25 or lower than 0.8 are named.
- Figure S3. SUMO-conjugated protein signature of chemoresistance.
Modification levels of the proteins modified by SUMO selected in Fig 1C were compared between all parental (U937 and HL-60) and drug-resistant (ARA-R or DNR-R) sublines. Differentially modified proteins with a drug-resistant versus parental cell ratio higher than 1.25 or lower than 0.8 are named.
- Figure 3. Generation of a UbL-predictive score of acute myeloid leukemias response to chemotherapies.
(A) Selection of the best predictive proteins. A list of 30 predictors among the 122 signature proteins was chosen to run a genetic algorithm. These predictors corresponded to 23 proteins modified either by ubiquitin (Ub), SUMO (Su), or both UbLs. The rate of selection of each of these proteins and its associated UbL (Ub or Su) is indicated on the graph. The proteins were separated in five subsets for linear discriminant analysis analysis. (B, C) Probability of acute myeloid leukemias sensitivity/resistance using the subset 2 solution. (B, C) The seven proteins showing the highest rate of selection in the genetic algorithm (subset 2) were used for linear discriminant analysis to predict the probability of resistance for the U937 and HL60 cell lines (B) or patient samples (C). Cells were considered sensitive to chemotherapy if the probability to belong to the group of resistant cells was below 50% and resistant if over 50%.
- Figure 4. Flow cytometry assay for the detection of UbL-conjugated chemoresistance biomarkers in cell lines and patient samples.
(A) Principle of the assay. Recombinant signature proteins are produced in E. coli and coupled to differently colored xMap beads. Coupled beads are then mixed and incubated with cell extracts supplemented with recombinant UbL, in the presence of UbL-vinyl-sulfones to inhibit deconjugating activities. Protein modification is quantified by flow cytometry using a combination of primary antibodies directed to UbLs and fluorescent secondary antibodies. (A, B, C) Analysis of STAM, SQSTM1, UBADC1, and HIP2 ubiquitylation. Extracts from parental, ARA-R, or DNR-R U937 cells were used as described in (A) with beads coupled to STAM, SQSTM1, UBADC1 and HIP2. (B) Representative flow cytometry profiles for their ubiquitylation are shown in (B). (C) Quantification are presented in (C). For quantification, background (cell extracts supplemented with NEM to inhibit UbL conjugation activities) was subtracted and ratio between resistant and parental cell lines was shown (n = 6 for STAM, n = 3 for SQSTM1, UBADC1 and HIP2). Mean ± SEM. Paired t test, *P < 0.05, **P < 0.01, ***P < 0.001. (D) Extracts from bone marrow aspirates from 37 patients at diagnosis were used in a multiplexed flow cytometry analysis using xMap beads coupled to STAM, UBADC1, and SQSTM1. 29 patients responded to induction chemotherapy (<10% of blasts in bone marrow 30 d after the beginning of chemotherapy) and 10 did not (>10% of blasts in bone marrow 30 d after the beginning of chemotherapy). Mean ± SEM. Unpaired t test with Welch’s correction. For the refractory patients, those showing high ubiquitylation for at least one of the biomarkers are colored-coded. MFI, median fluorescence intensity; ns, not significant.
Supplementary Materials
