19S Proteasome Subunits as Oncogenes and Prognostic Biomarkers in FLT3-Mutated Acute Myeloid Leukemia (AML)

26S proteasome non-ATPase subunits 1 (PSMD1) and 3 (PSMD3) were recently identified as prognostic biomarkers and potential therapeutic targets in chronic myeloid leukemia (CML) and multiple solid tumors. In the present study, we analyzed the expression of 19S proteasome subunits in acute myeloid leukemia (AML) patients with mutations in the FMS-like tyrosine kinase 3 (FLT3) gene and assessed their impact on overall survival (OS). High levels of PSMD3 but not PSMD1 expression correlated with a worse OS in FLT3-mutated AML. Consistent with an oncogenic role for PSMD3 in AML, shRNA-mediated PSMD3 knockdown impaired colony formation of FLT3+ AML cell lines, which correlated with increased OS in xenograft models. While PSMD3 regulated nuclear factor-kappa B (NF-κB) transcriptional activity in CML, we did not observe similar effects in FLT3+ AML cells. Rather, proteomics analyses suggested a role for PSMD3 in neutrophil degranulation and energy metabolism. Finally, we identified additional PSMD subunits that are upregulated in AML patients with mutated versus wild-type FLT3, which correlated with worse outcomes. These findings suggest that different components of the 19S regulatory complex of the 26S proteasome can have indications for OS and may serve as prognostic biomarkers in AML and other types of cancers.


Introduction
Acute myeloid leukemia (AML) is a heterogeneous leukemic disease characterized by abnormal stem cells in the bone marrow. AML stem cells result in the accumulation of immature myeloid cells in the peripheral blood or bone marrow, which ultimately leads to bone marrow failure [1]. Despite initial high response rates to standard induction chemotherapy regimens, most patients eventually relapse and succumb to their disease [1]. Molecular profiling of AML through cytogenetics and targeted sequencing panels has revealed the biological complexity of this disease and is used for risk stratification and guiding therapeutic options for patients [2]. For example, alterations in the FMS-like tyrosine kinase-3 (FLT3) gene are common driver mutations of AML and are present in over 30% of all cases. Patients with FLT3 internal tandem duplications (FLT3-ITDs) have a poor prognosis, with higher relapse rates and reduced overall survival (OS) [1]. Upon the We found that the expression of PSMD3 but not PSMD1 correlated with worse OS in FLT3+ AML. We further assessed the potential for PSMD3 to act as an oncogene in FLT3+ AML both in vitro and in vivo, and whether other 19S proteasome subunits could serve as novel prognostic biomarkers. Collectively, our findings suggest that different components of the 19S regulatory complex in the 26S proteasome have indications for OS and may serve as prognostic biomarkers and novel therapeutic targets in FLT3-mutated AML and other types of cancers.

High Levels of PSMD3 but Not PSMD1 mRNA Expression Correlated with Worse OS in AML Patients with Mutated versus Wild-Type FLT3
Based on our recent data in CML and other types of cancers [11,21], we hypothesized that PSMD1 and PSMD3 mRNA expression would be upregulated in AML versus normal mononuclear cells (MNCs), and that high expression would correlate with a worse OS. In fact, Dai et al. previously reported that high PSMD3 mRNA levels were associated with worse outcomes in AML [24]. Using data from GEPIA2, we observed no significant difference in PSMD1 expression comparing AML versus normal mononuclear cells ( Figure 1A). Consistent with this report, data from UALCAN revealed no correlation in OS comparing AML patients with high versus low PSMD1 expression (p = 0.1, Figure 1B). Taking these data one step further, we used UALCAN to analyze OS between FLT3-mutated AML patients with high versus low PSMD1 expression, which again had no effect on OS (p = 0.3, Figure 1C). Data from GEPIA2 surprisingly showed that PSMD3 mRNA expression levels were significantly downregulated in AML versus normal controls (p < 0.001, Figure 1D). Despite this observation, TCGA data available at UALCAN showed that high levels of PSMD3 mRNA expression correlated with a significant reduction of OS when all AML subtypes were considered (p = 0.0029, Figure 1E), and this was especially true for AML patients with FLT3 mutations (p = 0.0015, Figure 1F). Altogether, these data implicate PSMD3 as a potential oncogene in FLT3-mutated AML. mutated AML patients with high versus low PSMD1 expression, which again had no effect on OS (p = 0.3, Figure 1C). Data from GEPIA2 surprisingly showed that PSMD3 mRNA expression levels were significantly downregulated in AML versus normal controls (p < 0.001, Figure 1D). Despite this observation, TCGA data available at UALCAN showed that high levels of PSMD3 mRNA expression correlated with a significant reduction of OS when all AML subtypes were considered (p = 0.0029, Figure 1E), and this was especially true for AML patients with FLT3 mutations (p = 0.0015, Figure 1F). Altogether, these data implicate PSMD3 as a potential oncogene in FLT3-mutated AML.

PSMD3 Knockdown Reduced Survival of FLT3-Mutated AML Cell Lines In Vitro and In Vivo
We have already reported that shRNA-mediated knockdown of PSMD3 had no effect on survival or apoptosis of normal cord blood CD34 + progenitor cells, but reduced survival and increased apoptosis of CML cell lines in vitro and in vivo [21]. To assess the role of PSMD3 in AML, we transduced the FLT3-mutated AML cell lines, MOLM-13, MOLM-14, and MV4-11 with doxycycline-inducible shPSMD3, and confirmed PSMD3 knockdown at the mRNA and protein level in the presence of doxycycline (Figure 2A,B, Supplementary Figure S1A). Consistent with an oncogenic role for PSMD3 in FLT3-mutated AML, shRNAmediated knockdown of PSMD3 impaired survival of all three cell lines in colony formation assays in vitro, in the presence and absence of the FLT3 TKI, quizartinib (also known as AC220, Figure 2C,D,left, Supplementary Figure S1B, left). However, while shPSMD3 increased apoptosis of CML cell lines in the presence and absence of therapy [21], the phenotype was different in FLT3-mutated AML cell lines. Rather, in the absence of the FLT3 TKI, quizartinib, shPSMD3 significantly increased apoptosis of MOLM-14 and MV4-11 cells, but not MOLM-13 cells, with no further effects when cells were cultured with TKIs ( Figure 2C,D,right, Supplementary Figure S1B, right).
To confirm these data using an in vivo model, we intravenously injected 3 × 10 6 shNTor shPSMD3-expressing MOLM-13 cells into sublethally irradiated NOD-scid IL2Rgamma null (NSG) recipient mice and assessed for differences in OS in the presence of doxycycline. Consistent with an oncogenic role for PSMD3 in AML, mice receiving shPSMD3-expressing cells had a significant increase in OS compared with mice receiving the shNT-expressing control cells (p = 0.0027, Figure 3). Altogether, these data suggest an oncogenic role for PSMD3 in promoting survival of FLT3-mutated AML cells both in vitro and in vivo. However, the effect of PSMD3 knockdown on apoptosis was variable, with significant effects in the MOLM-14 and MV4-11 cell lines, but no effect in MOLM-13 cells, implying alternative oncogenic functions compared with our previous observations in CML [21]. tated AML, shRNA-mediated knockdown of PSMD3 impaired survival of all three cell lines in colony formation assays in vitro, in the presence and absence of the FLT3 TKI, quizartinib (also known as AC220, Figure 2C,D,left, Supplementary Figure S1B, left). However, while shPSMD3 increased apoptosis of CML cell lines in the presence and absence of therapy [21], the phenotype was different in FLT3-mutated AML cell lines. Rather, in the absence of the FLT3 TKI, quizartinib, shPSMD3 significantly increased apoptosis of MOLM-14 and MV4-11 cells, but not MOLM-13 cells, with no further effects when cells were cultured with TKIs ( Figure 2C,D,right, Supplementary Figure S1B, right). To confirm these data using an in vivo model, we intravenously injected 3 × 10 6 shNTor shPSMD3-expressing MOLM-13 cells into sublethally irradiated NOD-scid IL2Rgamma null (NSG) recipient mice and assessed for differences in OS in the presence of doxycycline. Consistent with an oncogenic role for PSMD3 in AML, mice receiving shPSMD3expressing cells had a significant increase in OS compared with mice receiving the shNTexpressing control cells (p = 0.0027, Figure 3). Altogether, these data suggest an oncogenic role for PSMD3 in promoting survival of FLT3-mutated AML cells both in vitro and in vivo. However, the effect of PSMD3 knockdown on apoptosis was variable, with significant effects in the MOLM-14 and MV4-11 cell lines, but no effect in MOLM-13 cells, implying alternative oncogenic functions compared with our previous observations in CML [21].

PSMD3 Knockdown in FLT3+ AML Resulted in An Increase of Global Ubiquitylated Proteins, but Had Little Effect on NF-κB Luciferase Reporter Activity
In CML, we previously reported that PSMD1 or PSMD3 knockdown decreased NF-κB luciferase reporter activity, which correlated with a global accumulation of ubiquitylated proteins [21]. Similarly, knockdown of PSMD3 in FLT3+ AML cells correlated with an increased level of global ubiquitylated proteins ( Figure 4A). Therefore, we speculated that knockdown of PSMD3 would decrease NK-κB transcriptional activity in FLT3+ AML cells. Surprisingly, luciferase reporter assays demonstrated that, while quizartinib showed the expected reduction of NK-κB luciferase reporter activity, PSMD3 knockdown had no significant effect ( Figure 4B). However, in MOLM-14 cells, shPSMD3 showed a trend for reduced NK-κB transcriptional activity, which was not further reduced by quizartinib therapy ( Figure 4B; p = 0.07). These data further suggest that the role of PSMD3 in FLT3+ AML cells is distinct from its function in BCR-ABL1+ CML cells.
To address the oncogenic role of PSMD3 in FTL3+ AML, we used the correlation feature available at UALCAN to calculate the gene sets upregulated with PSMD3 in TCGA AML data (Supplementary Figure S2). Our analysis implicated a potential role for PSMD3 in the following pathways: metabolism of RNA (R-HSA-8953854), ribonucleoprotein complex biogenesis (GO:0022613), ribonucleoprotein complex subunit

PSMD3 Knockdown in FLT3+ AML Resulted in An Increase of Global Ubiquitylated Proteins, but Had Little Effect on NF-κB Luciferase Reporter Activity
In CML, we previously reported that PSMD1 or PSMD3 knockdown decreased NF-κB luciferase reporter activity, which correlated with a global accumulation of ubiquitylated proteins [21]. Similarly, knockdown of PSMD3 in FLT3+ AML cells correlated with an increased level of global ubiquitylated proteins ( Figure 4A). Therefore, we speculated that knockdown of PSMD3 would decrease NF-κB transcriptional activity in FLT3+ AML cells. Surprisingly, luciferase reporter assays demonstrated that, while quizartinib showed the expected reduction of NF-κB luciferase reporter activity, PSMD3 knockdown had no significant effect ( Figure 4B). However, in MOLM-14 cells, shPSMD3 showed a trend for reduced NF-κB transcriptional activity, which was not further reduced by quizartinib therapy ( Figure 4B; p = 0.07). These data further suggest that the role of PSMD3 in FLT3+ AML cells is distinct from its function in BCR-ABL1+ CML cells.  To address the oncogenic role of PSMD3 in FTL3+ AML, we used the correlation feature available at UALCAN to calculate the gene sets upregulated with PSMD3 in TCGA AML data (Supplementary Figure S2). Our analysis implicated a potential role for PSMD3 in the following pathways: metabolism of RNA (R-HSA-8953854), ribonucleoprotein complex biogenesis (GO:0022613), ribonucleoprotein complex subunit organization (GO:0071826), mitochondrial gene expression (GO:0140053), nucleobase-containing compound biosynthetic process (GO:0034654), and DNA metabolic process (GO:0006259), among others ( Figure 4C). To corroborate these results at the protein level, we performed mass spectrometry-based proteomics analyses in MOLM-13 and MOLM-14 cells after PSMD3 knockdown. In the MOLM-14 AML cell line, PSMD3 knockdown resulted in a significant increase of 149 proteins and a decrease of 265 proteins. In contrast, PSMD3 knockdown in the MOLM-13 cell line resulted in an increase of only 25 proteins and a decrease of only 8 proteins. The leading proteins that were up-or down-regulated upon PSMD3 knockdown in both cell lines are listed in Supplementary Tables S1 and S2. Interestingly, when pathway enrichment analysis was performed on the upregulated proteins, the leading dysregulated pathways were vacuolar transport (G0:0007034) in the MOLM-13 cell line and neutrophil degranulation (R-HSA-6798695) in the MOLM-14 cell line ( Figure 4D,E). In proteomics analyses, additional pathways dysregulated by shPSMD3 included several metabolic pathways, such as the monocarboxylic acid metabolic process, generation of precursor metabolites and energy, glycosyl compound metabolic process, hexose metabolic process, and small molecule biosynthetic process ( Figure 4E).  We next asked whether the expression of other non-ATPase subunits of the 19S proteasome were altered in AML patients with mutated versus wild-type FLT3 and whether they correlate with OS. TCGA data available from UALCAN demonstrated that the genes encoding PSMD2 (p = 0.046), PSMD6 (p = 0.03), PSMD7 (p = 0.046), and PSMD9 (p = 0.048) were found to be significantly upregulated in AML patients with mutated versus wildtype FLT3 ( Figure 5A-D). Similar to PSMD3 (Figure 1C), GEPIA2 mRNA expression data showed that PSMD2 and PSMD7 (p < 0.01), but not PSMD6 or PSMD9, were surprisingly downregulated in AML versus normal MNCs ( Figure 6A-D). When all AML subtypes were considered, high expression of mRNA encoding PSMD2 (p = 0.0001), PSMD6 (p = 0.046), PSMD7 (p = 0.00046), and PSMD9 (p = 0.003) correlated with a worse OS in TCGA AML data ( Figure  6E-H). In FLT3-mutated AML, high levels of PSMD2 (p = 0.00016) and PSMD7 (p = 0.00058), but not PSMD6 (p = 0.17) or PSMD9 (p = 0.052) mRNA expression correlated with a significant reduction in OS ( Figure 6I-L). Additionally, mRNA encoding other non- Similar to PSMD3 (Figure 1C), GEPIA2 mRNA expression data showed that PSMD2 and PSMD7 (p < 0.01), but not PSMD6 or PSMD9, were surprisingly downregulated in AML versus normal MNCs ( Figure 6A-D). When all AML subtypes were considered, high expression of mRNA encoding PSMD2 (p = 0.0001), PSMD6 (p = 0.046), PSMD7 (p = 0.00046), and PSMD9 (p = 0.003) correlated with a worse OS in TCGA AML data 8 of 17 ( Figure 6E-H). In FLT3-mutated AML, high levels of PSMD2 (p = 0.00016) and PSMD7 (p = 0.00058), but not PSMD6 (p = 0.17) or PSMD9 (p = 0.052) mRNA expression correlated with a significant reduction in OS ( Figure 6I-L). Additionally, mRNA encoding other non-ATPase UPS subunits, including PSMD4 (p = 0.044), PSMD8 (p = 0.0032), and PSMD13 (p = 0.00022) also correlated with reduced OS in AML (Supplemental Figure S3A-C). Altogether, while expression of many PSMD subunits is reduced in AML versus normal MNCs, AML patients with elevated expression of several of these 19S subunits correlates with poor outcomes. ATPase UPS subunits, including PSMD4 (p = 0.044), PSMD8 (p = 0.0032), and PSMD13 (p = 0.00022) also correlated with reduced OS in AML (Supplemental Figure S3A-C). Altogether, while expression of many PSMD subunits is reduced in AML versus normal MNCs, AML patients with elevated expression of several of these 19S subunits correlates with poor outcomes.  Thus far, our data suggest an oncogenic role for 19S PSMD subunits in FLT3-mutated AML. Using cBioPortal to analyze multiple TCGA cancers, we previously reported that altered PSMD1 expression was associated with mutations or deep deletions in several different solid tumors, whereas PSMD3 was primarily associated with gene amplifications [11]. Therefore, we hypothesized that genomic alterations could be responsible for the differences in expression of these 19S subunits in AML. In contrast, for the non-ATPase subunits of the 19S proteasome that were reported here to correlate with OS in AML (PSMD2, PSMD3, PSMD4, PSMD6, PSMD7, PSMD8, PSMD9, and PSMD13; Figures 1, 6 and S3), cBioPortal revealed very few reported genomic alterations. The only genes with reported genomic alterations were deep deletions in PSMD7 and PSMD13 ( Figure S4). Otherwise, the remaining 19S proteasome subunit genes had no genomic alterations reported in AML, suggesting alternative methods for regulation of these genes in FLT3+ AML.

Discussion
AML is a complex, heterogeneous disease that initially responds well to standard chemotherapy drug regimens, yet despite initial responses, patients often relapse due to the presence of common driver mutations. Mutations in the FLT3 tyrosine kinase receptor are common in AML, and using FLT3 TKIs has shown promising results for many AML patients [25][26][27][28]. Unfortunately, most patients eventually develop resistance to FLT3 inhibitors, making FLT3 a poor prognostic biomarker in this disease. One novel strategy to circumvent this method of resistance is to selectively target the UPS. The UPS is a vital cellular organelle involved in the regulation of protein homeostasis [20]. In AML cells with FLT3 mutations, proteasome inhibitors such as bortezomib are more effective compared with cells harboring wild-type FLT3, but are still prone to resistance mechanisms [16,17]. Recent work by Mo et al. has shown that knockdown of the proteasome subunit, PSMB5, improved bortezomib sensitivity in patients with multiple myeloma through the increased expression of the pro-apoptotic signaling gene, BAX, and decreased expression of the anti-apoptotic genes, BCL2 and AKT [29].
Another potential strategy to overcome drug resistance in cancer therapy is to target the 19S regulatory complex instead of the 20S core complex (such as PSMB5), which is the target of standard proteasome inhibitors such as bortezomib [14]. For example, several groups have demonstrated the potential for targeting adhesion-regulating molecule 1 (ADRM1/hRPN13) in cancer therapy and bortezomib resistance [30][31][32][33][34][35][36]. This rigorous work has led to the development of the putative RPN13 inhibitor, RA190 [37][38][39][40]. However, more work must be done before we can effectively target the 19S regulatory particle as a therapeutic approach in AML.
We have previously reported that the proteasome subunits, PSMD1 and PSMD3, are upregulated during progression of CML from the indolent chronic phase to the aggressive blast phase of the disease [21]. We also previously showed that PSMD1 and PSMD3 mRNA expression are upregulated in other types of cancers compared with normal tissue, and that this often correlates with a worse prognosis [11]. In the present study, we assessed whether PSMD1, PSMD3, and other non-ATPase subunits of the 19S proteasome could serve as prognostic biomarkers and potential therapeutic targets in FLT3+ AML. Using publicly available databases, we analyzed PSMD1 and PSMD3 expression in FLT3-mutated AML patients and found no significant difference in OS comparing patients with high versus low PSMD1 expression ( Figure 1C). However, while PSMD3 mRNA levels were significantly reduced in AML versus normal cells, patients with higher PSMD3 expression levels demonstrated a significant reduction in OS, especially for AML patients harboring mutated FLT3 (p = 0.0015) ( Figure 1F). These data suggest that PSMD3 could be a novel prognostic biomarker for risk stratification and therapy response prediction in AML. Therefore, we used shRNAs to assess the oncogenic potential of PSMD3 in FLT3+ AML cell lines in vitro and in vivo. Knockdown of PSMD3 resulted in reduced survival of FLT3+ AML cells in colony formation assays ( Figure 2C,D,right), which correlated with better OS when injected into immunocompromised recipient mice (Figure 3). These findings suggest that targeting specific proteasomal subunits could be a potential therapeutic option for patients with AML and other types of cancers.
Indeed, the proteasomal subunit, PSMD2, has prognostic significance in patients with bladder urothelial carcinoma [41] and lung adenocarcinoma [42,43]. Other reports demonstrated a role for PSMD6 in lung adenocarcinoma [20] and esophageal squamous cell carcinoma [43]. PSMD7, a core component of the 26S proteasome, may play a role in the progression of breast cancer and could also serve as a prognostic indicator or therapeutic target [44]. PSMD7 mRNA expression was found to be upregulated in breast cancer tissue, and knockdown led to cell cycle arrest and apoptosis by decreasing the expression of cell cycle proteins and increasing the stability of the p21 and p27 cell cycle regulators [44]. PSMD7 may also play a role in the prognosis of patients with head and neck squamous cell carcinoma. Patients with high PSMD7 expression demonstrated poor clinical outcomes in patients with lung cancer, which also contributed to cisplatin resistance in patients with gastric cancer [45][46][47]. Additionally, recent work by Ud Din Farooqee et al. has implicated PSMD9 as a potential drug target for cancer therapy or ribosome-associated disorders [48]. PSMD9 was shown to affect two critical aspects in patients with breast cancer: (1) trafficking of ribosomal subunits into the nucleus and (2) contributing to decreased breast cancer cell growth and death. These findings implicate a structural role for PSMD9 during ribosome interactions [48]. When we analyzed whether the expression of other PSMD subunits also correlated with OS in AML, patients with elevated levels of PSMD2, PSMD7, and possibly PSMD9 had worse outcomes ( Figure 6I-L). High expression of PSMD4, PSMD8, and PSMD13 also correlated with reduced OS in AML ( Figure S3A-C).
In CML, we observed reduced NF-κB luciferase reporter activity with PSMD3 knockdown [21]. When we assessed the effects of PSMD3 knockdown on NF-κB luciferase activity in FLT3+ AML, we observed minimal changes ( Figure 4B), suggesting that additional signaling pathways are regulated by PSMD3 expression in this particular disease. The divergent phenotypes elicited by PSMD3 knockdown in CML versus AML suggest distinct functional roles for PSMD3 in these myeloid malignancies. Indeed, proteomics analyses revealed vacuolar transport and neutrophil degranulation as the top dysregulated pathways upon knockdown of PSMD3 in FLT3+ AML cells ( Figure 4D,E). This is not entirely surprising, as the PSMD3 gene on chromosome 17 is upstream of the gene encoding colony-stimulating factor 3 (CSF3), a member of the interleukin (IL)-6 superfamily of cytokines. CSF3 controls the production, differentiation, and function of human granulocytes, and mice lacking CSF3 are neutropenic [49]. In humans, common variations in the PSMD3-CSF3 gene were previously associated with neutrophil counts [50]. Additional pathways dysregulated by PSMD3 knockdown in FLT3+ AML included immune cell cytokine signaling and energy metabolism pathways. Future studies will assess the mechanism by which these proteins contribute to survival in FLT3-mutated AML patients and whether they serve as novel targets for therapy.
Many studies have documented the prognostic significance of PSMD subunit expression in multiple cancers [11,[20][21][22][23]44,46,48,51]. Despite their recent progress as prognostic biomarkers, more research is needed before we can effectively target the 19S proteasome as a therapeutic option in cancer. Importantly, since the PSMD subunits of the 19S proteasome are non-ATPase proteins that lack catalytically active target sites, they could be challenging molecular targets that require the blockage of protein-protein interactions. A different approach could be to target the proteasome 26S subunit, ATPase (PSMC) members of the 19S proteasome, which are ATPases that harbor catalytically active target sites. Indeed, PSMC2, PSMC3, PSMC4, PSMC5, and PSMC6 were recently reported as highly expressed in breast cancer, which correlated with worse outcomes [52]. PSMC2 has been implicated as an oncogene in multiple cancer types, including glioma, gastric cancer, squamous cell carcinoma, and skin cutaneous melanoma, to name a few [53][54][55][56]. In colorectal cancer, PSMC5 promoted cancer cell proliferation and metastasis by activating the epithelial-mesenchymal transition and modulating immune cell infiltration [57]. A CRISPR genome-wide screen identified dependence on PSMC6 for bortezomib sensitivity in multiple myeloma [58].
Whether the PSMC and PSMD subunits of the 19S proteasome are viable targets for therapeutic intervention will be an important topic of future studies. Altogether, our data point to subunits of the 19S proteasome as novel prognostic biomarkers and putative targets for therapy in FLT3+ AML, CML, and other types of cancers that are worthy of future investigation.

Analysis of 19S Regulatory Subunit Expression in AML Patients Using Publicly Available Databases
GEPIA2 (http://gepia2.cancer-pku.cn/, accessed on 2 August 2021) integrates data from TCGA and the Genotype-Tissue Expression (GTEx) project to provide large-scale gene expression profiling [12][13][14]. Using GEPIA2, the differential mRNA expression of the 19S proteasomal subunits, PSMD1 and PSMD3, were analyzed in AML versus normal tissue. We also performed gene expression analyses for the other non-ATPase subunits of the 19S proteasome in AML (Table 1). UALCAN (http://ualcan/path.uab.edu/, accessed on 29 July 2021) is another publicly available platform comparing transcriptome or proteomics data from either TCGA or the Clinical Proteomic Tumor Analysis Consortium (CPTAC), respectively [59]. We used UALCAN to compare PSMD subunit expression in AML patients based on clinicopathological features (mutant versus wild-type FLT3). UALCAN was also used to assess the effects of PSMD subunit expression on OS in AML patients.

Analysis of PSMD3 mRNA Expression Using Reverse Transcription-Quantitative Polymerase Chain Reaction (RT-qPCR)
The PureLink RNA Mini Kit by Invitrogen (Thermo Fisher Scientific) was used to extract total RNA from cell lines and then quantified using a NanoDrop TM 2000 (Thermo Fisher Scientific). RT-qPCR was performed using the Luna Universal One-Step qPCR Kit (New England Biolabs, Ipswich, MA, USA) on a StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). GUSB mRNA levels were measured as controls, and the primer sequences are listed in Supplemental Table S4. Each RT-qPCR assay was performed in triplicate and analyzed using the comparative cycle threshold method (2 −∆∆Ct ) to calculate the relative gene expression level of PSMD3.  Table S5.

Apoptosis Assays of FLT3+ AML Cells Expressing shPSMD3
MOLM-13, MOLM-14, and MV4-11 cells expressing shPSMD3 were treated with doxycycline (100 ng/mL, 72 h) with and without the indicated inhibitors. Apoptosis was measured by staining of resulting cells with both APC-conjugated AnnexinV (BioLegend Inc., San Diego, CA, USA) and 7-aminoactinomycin D (7-AAD, eBioscience at Thermo Fisher Scientific), and analyzed on a BD FACSCanto II (BD Biosciences, San Jose, CA, USA) flow cytometer. All flow cytometry data were analyzed using FlowJo (Ashland, OR, USA), and only late apoptotic events are reflected in the bar graphs.

Colony Formation Assays on FLT3+ AML Cells Expressing shPSMD3
Colony formation assays were used to evaluate the effects of shPSMD3 on MOLM-13, MOLM-14, and MV4-11 cell survival. Briefly, 10 3 cells/dish were plated in 0.9% MethoCult (StemCell Technologies, Vancouver, BC, Canada) methylcellulose reagent as previously described [17]. AML cell lines were plated in cytokine-free conditions ± quizartinib (20 nM) to inhibit FLT and/or doxycycline (100 ng/mL) to induce the knockdown. The resulting colonies were then scored and quantified after 14 days of culture in a humidified chamber at 37 • C with 5% CO 2 .

NOD-Scid Il2rgamma null (NSG) AML Xenografts
Using either shPSMD3 or the shNT control, MOLM-13 cells were stably transduced and then placed in puromycin (2 µg/mL, 72 h) for selection. Resulting cells (3 × 10 6 cells/mouse, n = 4-5 mice/group) were then intravenously injected into sublethally irradiated (1 × 200 Rad, RS 2000 X-ray Biological Irradiator, Rad Source Technologies) NSG recipient mice (Jackson Laboratories, Bar Harbor, ME, USA) to produce AML cell line xenografts. To induce in vivo gene expression of the shPSDM3 or shNT vectors, recipient mice were placed on 625 mg/kg doxycycline hyclate chow (Envigo Teklad, Indianapolis, IN, USA #TD.01306). Recipient mice were inspected daily for visual signs of leukemia, and moribund mice were humanely euthanized. All experiments were performed with approval by the Texas Tech University Health Sciences Center at El Paso Institutional Animal Care and Use Committee.

NF-κB Luciferase Reporter Assays
To measure endogenous NF-κB transcriptional activity, AML cell lines expressing sh-PSMD3 or the shNT control vector were lentivirally transduced using a pGreenFire1-NF-κB Lentivector reporter system per the manufacturer's instructions (System Biosciences LLC, Palo Alto, CA, USA; Supplemental Table S3). Cells at 50-70% confluency were transduced using the TransDux solution with media to a 1X final concentration. Stably transduced cells were selected for 10 days in the presence of Geneticin TM (Life Technologies) and resulting cells expressing the pGreenFire1-NF-κB reporter system were seeded at 5000 cells per well using a sterile flat-bottom 96-well plate (Greiner Bio-One, Kremsmunster, Austria) in the presence of the indicated inhibitors. After 72h, luciferase reporter activity was analyzed using the ONE-Glo Luciferase Assay System (Promega Corporation, Madison, WI, USA) on a CLARIOstar Plus (BMG Labtech, Ortenberg, Germany) plate reader.

Proteomics Sample Preparation and Analyses
Proteomics analyses were performed at the University of Texas Southwestern (UTSW) Medical Center Proteomics Core Facility (Dallas, TX, USA). Briefly, 10 6 MOLM-13 and MOLM-14 cells and derivative lines were lysed in 1X RIPA buffer containing protease and phosphate inhibitors (Thermo Fisher Scientific). Samples were digested overnight with trypsin (Pierce TM , Thermo Fisher Scientific) after reduction and alkylation with dithiothreitol (DTT) and iodoacetamide (Sigma-Aldrich, Darmstadt, Germany). Samples then underwent solid-phase extraction cleanup with an Oasis HLB plate (Waters Corporation, Milford, MA, USA), and the resulting samples were injected onto an Orbitrap Fusion Lumos mass spectrometer (MS), Q Exactive TM HF (Thermo Fisher Scientific) coupled to an Ultimate 3000 RSLC-Nano liquid chromatography system (Thermo Fisher Scientific). Samples were injected onto a 75 µm, 75-cm long EASY-Spray column (Thermo Fisher Scientific) and eluted with a gradient from 0-28% buffer B over 90 min. Buffer A contained 2% (v/v) acetonitrile (ACN) and 0.1% formic acid in water, and buffer B contained 80% (v/v) ACN, 10% (v/v) trifluoroethanol, and 0.1% formic acid in water. A positive ion mode was used with a source voltage of 1.8 kV, with an ion transfer tube temperature of 275 • C.
MS scans were acquired at 120,000 resolution in the Orbitrap (Thermo Fisher Scientific). Up to 10 MS/MS spectra were obtained in the ion trap for each full spectrum acquired, using higher-energy collisional dissociation (HCD) for ions with charges 2-7. Dynamic exclusion was set for 25 s after an ion was selected for fragmentation. The raw MS data files were analyzed using Proteome Discoverer v2.4 SP1 (Thermo Fisher Scientific), with peptide identification performed using Sequest HT searching against the human protein database from UniProt, which is a freely accessible resource for protein sequence and functional information. Fragment and precursor tolerances of 10 ppm and 0.6 Da were specified, and up to three missed cleavages were allowed. Carbamidomethylation of cysteine (Cys) was set as a fixed modification, with oxidation of methionine (Met) set as a variable modification. Pathway enrichment analysis of the dysregulated proteins was performed using Metascape [62].

Genomic Alterations of PSMD Subunits in AML
cBioPortal (https://www.cbioportal.org/, accessed on 30 July 2021) is an online cancer genomics database that was used for the analysis of genomic variations across various cancer types [63]. The database was used to analyze data sets that contained point mutations, deep deletions, and DNA amplifications for the different PSMD subunits that correlated with OS in AML, similar to our previous analysis performed in other types of cancers [11]. The data is represented as a bar graph for the various PSMD subunits in AML.

Statistical Analyses
All experiments were performed in triplicate unless otherwise noted. A two-tailed Student's t-test was used to analyze cell line and mouse data demonstrating equivocal variance. Welch's t-test was used to calculate significant differences in expression comparing normal versus AML specimens, or AML specimens based on clinicopathological features (wild-type versus mutated FLT3). For analysis of OS, primary AML samples were divided into high versus low PSMD subunit expression, with sample values higher than or lower than the 3rd quartile value, respectively, with data presented as Kaplan-Meier curves. For proteomics data analysis, each sample was normalized to its internal expression of GUS, and resulting values were analyzed with a Student's t-test with a false-discovery rate (FDR) cutoff of 1% for all peptides. Error bars represent the standard error of the mean (SEM), and a p-value < 0.05 was considered statistically significant.

Institutional Review Board Statement:
The animal study protocol was approved by the Institutional Animal Care and Use Committee at Texas Tech University Health Sciences Center at El Paso (protocol #18006, approved 29 August 2022).