Introduction: The endoplasmic reticulum (ER) is the major site of protein synthesis and folding in the cell. Three pathways are integrated to maintain ER homeostasis: ER-associated degradation (ERAD), unfolded protein response (UPR), and autophagy. One of the chief elements of ERAD is the highly conserved AAA-ATPase superfamily member valosin-containing protein (VCP, or p97). Various studies have reported an upregulation of VCP in cancer and an association between elevated VCP expression and unfavorable cancer outcome. A promising therapeutic approach relies on targeting the cellular stress response, and systemic functional genomic screens have identified VCP as potential target for inhibition in acute myeloid leukemia (AML). As clinical impact of VCP has not been studied, we wondered whether baseline protein expression levels were predictive of outcome in acute leukemia (AL).

Methods: Reverse Phase Protein Array (RPPA) was performed with strictly validated antibodies, including antibodies against VCP, IRE1 and GRP78, to determine the protein expression levels of diagnostic leukemic cells from 500 pediatric AML, 818 adult AML, 268 pediatric T-ALL and 93 adult T-ALL patient samples. Pediatric patients participated on either the COG AAML1031 or AALL1231 clinical trial comparing standard therapy (ADE or AFBM) to standard therapy plus bortezomib (ADE+B or AFBM+B). Adults were treated under a variety of protocols. Pearson correlation analyses was used to identify significant protein-protein correlations. Estimates of survival was calculated using the Kaplan-Meier method.

Results: VCP protein was expressed in both AML and T-ALL. Although VCP was slightly more highly expressed in AL vs normal CD34+ cells, the majority of patients had VCP expression within the normal range. In pediatric AML, VCP was more highly expressed in younger patients (< 2 y/o), KMT2A (formerly MLL)-rearrangement (p<0.001) and high WBC count (p=0.01) while it was lower in inv(16) (p=0.0003), t(8;21) (p=0.0016), c-Kit (exon 8, p=0.00061; exon 17, p=0.0003), and MYH11 (p=0.0013). FLT3, NPM1, CEBPA, NRAS, KRAS, PTPN11, IDH1/2, and GATA2 mutation status were not associated with VCP expression. Survival analysis identified a significant association between VCP and outcome. In pediatric AML, 5-yr overall survival (OS) was 81% in low-VCP vs 60% and 66% in middle and high-VCP (p<0.001) (Figure 1A). A similar observation was done for event-free survival (5-yr EFS, 59% vs 45-48%, p=0.004) and relapse risk (5-yr RR, 33% vs. 40-47%, p=0.019). This was independent of treatment (with or without bortezomib). High-VCP remained a significant independently unfavorable prognostic variable in multivariate analysis together with high-risk group (per AAML1031 definition) and age < 2 y/o (HR=1.45, HR= 1.75, HR=2.39, respectively). The same observation was seen in adult AML and T-ALL (Figure 1B-C); low-VCP did better than high VCP (AML: 5-yr OS, 32% vs 24%, p=0.029; T-ALL: 4-yr OS, 25% vs 50-60%). In pediatric T-ALL, no association with survival was found (4-yr OS; 89% vs 85%, p=0.45). Protein-protein correlations identified an inverse correlation with VCP and UPR proteins IRE1 and GRP78 in each subtype of AL. Other strong positive correlations were with proteins involved in histone modification, mTor and ribosomal proteins, and strong negative correlations with transcription factors, DNA repair and signal transduction pathways.

Conclusion: Using a proteomics approach we identified low-VCP as favorable prognostic indicator. This prognostic association was independent of treatment with a proteasome inhibitor, suggesting that the prognostic effect was potentially separate from the proteasome, and points toward a potential for VCP drug inhibition. Negative correlation with VCP and IRE1 and GRP78 might imply that cells with higher VCP rely on proteasomal degradation, whereas those with low VCP are more dependent on other pathways for protein degradation. Also, adaptation in VCP levels may not be a prominent feature of the stress response to chemotherapy. More information is needed to understand what is driving VCP expression.

Figure 1
Figure 1.
View largeDownload slide

Disclosures

No relevant conflicts of interest to declare.

© 2021 by The American Society of Hematology
2021
Sign in via your Institution