Targeting Positive Cofactor 4 Induces Autophagic Cell Death In MYC-Expressing Diffuse Large B Cell Lymphoma

Background: The MYC-expressing diffuse large B-cell lymphoma (DLBCL) is one of the refractory lymphomas. The pathogenesis of MYC-expressing DLBCL is still unclear, and there is a lack of effective therapy. In this study, we have explored the clinical signicance and the molecular mechanisms of transcription co-activator 4 (PC4) in MYC-expressing DLBCL. Methods: We investigated PC4 expression in 54 cases of DLBCL patients’ tissues and matched normal specimens, and studied the molecular mechanisms of PC4 in MYC-expressing DLBCL both in vitro and in vivo. Results: We reported for the rst time that targeting c-Myc could induce autophagic cell death in MYC-expressing DLBCL cell lines. We next characterized that PC4 was an upstream regulator of c-Myc, and PC4 was overexpressed in DLBCL and was closely related to clinical staging, prognosis and c-Myc expression. Further, our in vivo and in vitro studies revealed that PC4 knockdown could induce autophagic cell death of MYC-expressing DLBCL. And inhibition of c-Myc mediated aerobic glycolysis and activation of AMPK / mTOR signaling pathway were responsible for the autophagic cell death induced by PC4 knockdown in MYC-expressing DLBCL. Through the DLRTM and EMSA assay, we also found that PC4 exerted its oncogenic functions by directly binding to c-Myc promoters. Conclusions: PC4 exerts its oncogenic functions by directly binding to c-Myc promoters. Inhibition of PC4 can induce autophagic cell death of MYC-expressing DLBCL. Our study provides novel insights into the functions and mechanisms of PC4 in MYC-expressing DLBCL, and suggests that PC4 might be a promising therapeutic target for MYC-expressing DLBCL. to exert MYC-expressing suggested that PC4 might be a promising therapeutic target for MYC-expressing DLBCL.

promising phenomenon has been found in the study of tumor energy metabolism, the rapid reduction of the energy charge below a critical limit can trigger autophagy cell death rather than an adaptive autophagic response 10 . Different from apoptosis (Type I programmed cell death), autophagic cell death (Type II programmed cell death) occurs in various types of cancer 11 . In previous studies, inhibition of MYC could induce autophagic cell death in Burkitt lymphoma cell lines 12 , but the speci c mechanism was unclear. The combination of the antidepressants maputiline and uoxetine induce autophagic cell death in drug-resistant Burkitt's lymphoma 13 . In multiple myeloma, metformin can induce autophagic cell death through the AMPK/mTOR pathway [14][15] . This type of cell death can have a contribution to anticancer e cacy or drug resistance, respectively 16,17 . Therefore, targeting autophagy may provide a new potential therapeutic strategy to overcome drug resistance 18 .
In this study, we rstly revealed that targeting PC4 would induce autophagic cell death in MYC-expressing DLBCL cell lines. Through bioinformatics analysis, we found that PC4 was abnormally high expressed in DLBCL and was positively correlated with MYC expression. Considering the key role of c-Myc in DLBCL, it is crucial to elucidate the potential role and underlying molecular mechanisms of PC4 in DLBCL. Then, we rst reported that PC4 was highly expressed in DLBCL, and positively correlated with c-Myc expression and poor prognosis of patients. Next, our ndings revealed that PC4 knockdown induced autophagic cell death in MYC-expressing DLBCL, which is a new type of cell death. Surprisingly, knockdown of PC4 had no signi cant in uence on the c-Myc low expression cell lines and non-cancerous lymphocytic cell lines.

RNA interference
The shRNA lentivirus vector targeting human PC4(shPC4#1:  Cell apoptosis analysis by ow cytometry Cells were treated with AnnexinV-7-AAD (BD Biosciences) for 30 min at 37°C in the dark for apoptosis analysis, then analyzed by ow cytometry.

Western blotting analysis
The cell lines were harvested, washed, and lysed with RIPA buffer (Beyotime, China) which contain protease inhibitor cocktail (Roche) for 30min on ice. Total protein was collected and quantitated by a BCA kit (Beyotime, China) according to the recommended instruction. The protein samples were separated by electrophoresis in gel, and then transferred onto PVDF membranes (Millipore). Blotted membranes were incubated with primary antibodies overnight at 4°C. The membranes were washed 5min for 3 times with TBST, and then incubated with HRP-linked secondary antibody (Cell Signaling Technology, USA) 1h at room temperature. The band intensities were detected and visualized by an enhanced chemiluminescence detection system (Bio-Rad Laboratories). Primary antibodies against c-Myc, LC3, SQSTM1,ATG7,PARP,CASPASE3,Bcl-2,BAX,PI3K,S6K1,AKT,mTOR,4EBP1,AMPK,P38,P53,HIF-1α,GLUT1,PKM2,HK2,LDHA and β-actin were obtained from Cell Signaling Technology. Primary antibodies against PC4 were obtained from Sigma.

Immuno uorescence staining
Cells were xed in 4.0% formaldehyde for 10 min and permeabilized with ice-cold methanol or 0.5% Triton X-100 for 5 min. The following primary antibodies were used, and subsequently using secondary antibodies to detect Primary antibodies. Conjugating to AlexaFluor 488 or 555(Invitrogen). Cells were counterstained with DAPI and mounted in Vectashield (Vector Laboratories).
In vivo tumor growth model For in vivo tumor growth model, 100 ul PBS containing 1x10 7 PC4 stable knockdown TMD8 cells or negative control cells or controls were injected subcutaneously at one dorsal site of athymic male nude mice. Tumor growth was measured every 2 days, which was calculated by the following formula: volume (mm 3 ) = (width 2 * length)/2. At the endpoint, the mice were sacri ced; Carbon dioxide (CO 2 ) administration was used to euthanize mice. Put the mice into the euthanasia box (Clean, see-through airtight box), do not pour CO 2 into the euthanasia box rst. Attach CO 2 and exhaust pipes to the box and make sure other areas are sealed. 100%CO 2 was poured into the box at the rate of 20% replacement of box contents per minute for 10 minutes to ensure that the mice did not move and breathe (The mice gradually lost consciousness when CO 2 was injected into the box for 2-3 minutes ). CO 2 was turned off, and the animals were observed for 2 minutes. After that, the animals were taken out and con rmed to be completely dead by respiratory arrest, cardiac arrest and dilated pupils. And then xenografts were dissected, weighed and xed in 4% paraformaldehyde. Each group of n = 5. The animal carcasses were packed in opaque plastic bags for infectious substances and delivered to the Experimental Animal Center of Army Military Medical University for uni ed treatment.

Transmission electron microscopy
Cells were harvested and immediately xed in 3% glutaraldehyde overnight at 4°C and post xed with 2% osmium tetroxide for 1 hour at 37°C. And then, cells were embedded and stained using uranylacetate/lead citrate. The samples were imaged using a TEM (JEM-1400PLUS, Japan).

RNA-seq assay
Total RNA was extracted using Trizol reagent (Invitrogen, CA, USA) following the manufacturer's procedure. The total RNA quantity and purity were analysis of Bioanalyzer 2100 and RNA 6000 Nano

EMSA (electrophoretic mobility shift assay)
The DNA binding assays were performed using puri ed GST-PC4 protein and biotin-labelled fragments of the promoters containing the W-boxes, using GST protein as a negative control, and non-labeled fragments were used as competitors. The bands at the upper and lower part of membranes indicate shift (protein-probe complex) and unbound free probes, respectively.

Statistical analysis
Statistical analysis was carried out using SPSS 22.0 software (SPSS Inc., Chicago, USA), and all data were presented as means ± SD. Comparisons between two groups were performed using the Student's ttest. Comparisons among three or more groups were performed using a one-way analysis of variance (ANOVA). The survival data was carried out using the Kaplan-Meier method. Correlation between PC4 expression and clinical parameters was determined using the Pearson's χ 2 method. P <0.05 indicated a statistically signi cant difference.

Results
PC4 is highly expressed in DLBCL and positively correlated with c-Myc expression and poor prognosis of patients.
We began our studies by exploring the potential clinical signi cance of PC4, we rstly analyzed PC4 mRNA expression level in tumor tissues compared with normal tissues (Figure 1A), the mRNA expression of PC4 in tumor was apparently higher than that in normal tissues. We also analyzed GEPIA database and revealed that PC4 was positively correlated with MYC expression (P<0.001,R=0.46) in human whole blood (Supplemental Figure 1A). Moreover, GEPIA database showed that PC4 mRNA expression level was signi cantly higher in DLBCL patients tissues (n=47) than in normal tissues (n=337) (Supplemental Figure 1B). Then, we analyzed PC4 mRNA expression level and found that c-Myc(+) tissues had higher PC4 mRNA expression compared to c-Myc(-) tissues ( Figure 1B). In addition, we analyzed PC4 expression in DLBCL specimens with or without positive expression of c-Myc protein, and found that average staining score of PC4 in c-Myc(+) tissues were signi cantly increased compared to c-Myc(-) tissues ( Figure 1C). The data above suggested a possible positive correlation between PC4 expression levels and c-Myc expression level in DLBCL. Meanwhile, through analyzing 159 and 414 cases of DLBCL from public cancer databases (GSE4475 and GSE10846, respectively) we found that the higher PC4 expression group had poorer overall survival compared with lower PC4 expression group ( Figure 1D and 1E).
Although c-Myc expression demonstrated no signi cant change on the basis of gender, age, subtype, Ki-67 expression, it had poorly differentiated in DLBCL with a higher Ann Arbor stage and poorly event-free survival (Table1). Finally, we detected the protein level of PC4 and c-Myc in non-cancerous lymphocytic cell lines (CCRF-SD) and DLBCL cell Lines (DOHH2, OCL-LY10, HBL-1 and TMD8), our result showed that PC4 protein expression were much higher in TMD8 and HBL-1 cells compared to other cell lines ( Figure  1F). The qPCR ( Figure 1G) and Immuno uorescent staining (Supplemental Figure 2A) assays con rmed that the mRNA and protein level of PC4 was up-regulated in DLBCL cell lines (DOHH2, OCL-LY10, TMD8 and HBL-1 cells) compared with non-cancerous lymphocytic cell lines (CCRF-SD cells). Collectively, these results suggest that PC4 is a potential oncogene and positively correlated with c-Myc in DLBCL. Knockdown of PC4 induces cell apoptosis in MYC-expressing DLBCL in vitro and in vivo.
To investigate the functional signi cance of increased PC4 expression in DLBCL, TMD8 and HBL-1 cells were used for the subsequent loss-of-function study. The stable cell lines with PC4 knockdown were established by speci c shRNA (shPC4#1 and shPC4#2) (Figure 2A). The CCK-8 assays demonstrated that PC4 knockdown inhibited the proliferation ( Figure 2B). Then, the expression of apoptosis protein markers, including poly ADP-ribose polymerase (PARP), cleaved caspase 3, B-cell lymphoma-2 (BCL-2) and B-cell lymphoma-2-Associated X (BAX) were detected by western blotting in the constructed cells ( Figure 2C). As shown in Figure 2D, knockdown of PC4 increased apoptosis in stable PC4 Knockdown cell lines than controls, Gene Set Enrichment Analysis (GSEA) showed that the gene sets of apoptosis enrichment in PC4 low (shPC4) in TMD8 cell lines ( Figure 2E), which means that PC4 silencing can induce cell apoptosis. We observe cell vacuoles and cell debris in the constructed cells through a microscope (Supplemental Figure 3A). Moreover, TEM images showed autophagic vacuole (AV) formation in the constructed cells ( Figure 2F). These nding suggested that knockdown of PC4 was associated with autophagy. In addition, we established a subcutaneous xenograft model to study the biological function of PC4 in vivo. TMD8 cells with or without stable PC4-knockdown were inoculated into athymic male nude mice. During the in vivo experiments, xenograft growth in the sh-PC4#1 group was dramatically inhibited compared to the sh-NC group and control group ( Figure 2H). The average tumor weight and tumors size at the experimental endpoint was reduced by PC4 knockdown (Figure 2G and 2I). Besides, knockdown of PC4 had no signi cant impact on mice body weight ( Figure 2J). Taken together, the above data suggests that PC4 promotes DLBCL cell proliferation, and knockdown of PC4 can induce apoptosis and autophagy.
Knockdown of PC4 induces apoptosis in MYC-expressing DLBCL by inducing excessive autophagy.
Non-selective autophagy occurred due to lack of energy 11 . As envisioned, Silencing of PC4 could induce the expression of the LC3II and down-regulated the SQSTM1 protein in DLBCL cells ( Figure 3A). The associated autophagy proteins (beclin1, ULK1)also up-regulated after silencing of PC4 ( Figure 3B). In addition, TEM images showed AV formation in PC4 knockdown cells, which could be reversed by 3methyladenine (3MA; an autophagy inhibitor) ( Figure 3C). Above-mentioned data suggested that autophagy was induced by PC4-knockdown. PC4 knockdown signi cantly suppressed the proliferation of TMD8 and HBL-1 cells, which was partly reversed by 3MA ( Figure 3D). Similarly, the cell apoptosis was also reversed by 3MA ( Figure 3E). Therefore, our results con rmed that PC4 knockdown could induce autophagic cell death in TMD8 and HBL-1 cells. To assess the effect of knockdown of PC4 in c-Myc low expression cell lines (DOHH2 and OCL-LY10) and CCRF-SD were established with PC4 knockdown by speci c shRNA (Supplemental Figure 4A). Interestingly, we found no effect on the level of LC3II and cleaved PARP and cell proliferation (Supplemental Figure 4B and 4C). Collectively, these results suggested that PC4 inhibition can induce autophagic cell death only in the c-Myc high expressing DLBCL cell lines. Galluzzi reported that a rapid reduction in energy charge below a critical limit is likely to trigger autophagic cell death 11 . This is well established that c-Myc is a key regulator in energy metabolism 9 , suggesting that PC4 may directly regulate c-Myc or metabolism. This indicated that PC4 is a tumorspeci c oncogene and may be a novel therapeutic target for MYC-expressing DLBCL.
We conducted Genome-wide analysis to compare the gene expression pro les in TMD8 with or without stable PC4-knockdown to explore the potential mechanism of PC4 on cell proliferation inhibition and apoptosis in DLBCL. The heat map showed that 4 genes, including c-Myc, reduced their expression, while the expression of 36 genes were increased after PC4 knockdown ( Figure 4A). PC4 affects the expression of 12669 genes in TMD8 stable PC4 knockdown cells, with 12172 genes that overlap between TMD8 stable PC4 knockdown cells and control cells ( Figure 4B). Then, we used KEGG enrichment analysis in TMD8 stable PC4 knockdown cells and controls ( Figure 4C), which revealed that after PC4-knockdown AMPK signaling pathway was most increased and PI3K-Akt signaling pathway was most inhibited. Moreover, we established Gene ontology analysis to understand the PC4 full function, the results demonstrated that PC4 could positive regulate the metabolic process and cellular process in DLBCL ( Figure 4D). Therefore, PC4 promotes proliferation in DLBCL through regulating metabolism. According to the data from Genome-wide analysis, we established mTOR signaling pathway and downstream protein (4EBP1 and S6K1) by western blot, our result showed that the level of phosphorylation was decreased after PC4 knockdown ( Figure 4E). Then we detected mTOR upstream signaling pathway including AMPK, P53, P38 and AKT, which revealed that the level of AMPK phosphorylation was comparatively more increased than the other proteins ( Figure 4F). GSEA showed that the gene sets of mTORC1 were enriched in PC4 high (Control) compared with PC4 low (shPC4) in TMD8 cell lines ( Figure 4G). A previous study showed that alteration of the above pathways was associated with energy metabolism declension 19 .
PC4 regulates c-Myc transcription to perform it's oncogene function.
To further illustrate the underlying mechanisms of PC4 in DLBCL, we conducted metabolism-related experiments. We found that the 2-NBDG uptake and the production of Lactate and ATP were signi cantly inhibited after PC4-knockdown ( Figure 5A, 5B and 5C). Furthermore, Silencing of PC4 inhibited the key enzymes of glycolysis including GLUT1, PKM2, HK2 and LDHA ( Figure 5D). Previous studies showed that glycolysis metabolism was regulated by c-Myc 20 and HIF-1α 21 . Then we performed western blotting and q-PCR on c-Myc and HIF-1α and found that the protein and mRNA level of c-Myc was dramatically decreased in stable PC4 knockdown cell lines ( Figure 5E and 5F). GSEA showed that the gene sets of MYC were enriched in PC4 high (Control) compared with PC4 low (shPC4) in TMD8 cell lines ( Figure 5G). To further demonstrate the relationship between PC4 and c-Myc, we subsequently examined the relationship between PC4 and c-Myc. Schematic presentation of PC4 and c-Myc binding sites on the c-Myc locus are shown in Figure 5H, BS: binding site, BS1: CCAACAAATGCAATGGGAGT and BS2: CAGGAGGGGCGGTATCTG. We conducted luciferase reporter assays, which revealed that PC4 regulated c-Myc transcription through two PC4 binding sites in c-Myc promoters ( Figure 5I). However, EMSA con rmed PC4 as a DNA binding protein, which is associated with the c-Myc binding sequence in BS1, not in BS2 ( Figure 5J). To further verify the BS1 function, we conducted BS1 mutation in TMD8 and HBL-1 cell lines, BS mutation sequence: TTGGTGGGCATGGCAAAGAC. Our result con rmed that PC4 directly activate c-Myc transcription through BS1 by EMSA ( Figure 5K). In addition, the protein and mRNA level of PC4 has no signi cant in uence after c-Myc stable knockdown (Supplemental Figure 5A and 5B). Indiation, we conducted c-Myc overexpression experiment in PC4 knockdown cell lines by the speci c plasmid (Supplemental Figure 6A) to prove that c-Myc was responsible for the oncogenic functions of PC4. As expected, c-Myc overexpression rescued the apoptosis (Supplemental Figure 6B, 6D and 6E) and cell proliferation inhibition (Supplemental Figure 6C). The results of the above experiments indicated that PC4 regulated c-Myc transcription to perform its oncogenic function. The potential mechanism of PC4 in diffuse large B-cell lymphoma was shown in Figure 6 Discussion Owing to the increased recurrence and chemotherapy resistance, current strategies for the treatment of c-Myc (+) DLBCL are still facing with many di culties and challenges 22 . And c-Myc is a natural disordered protein and lacks drug recognition sites that can be utilized 8 . Therefore, it is critical to identify the c-Myc PC4 is a nuclear protein which also known as SUB1. As a multifunctional nuclear protein, PC4 is initially isolated, puri ed and identi ed from the upstream stimulatory activity (USA) in mammalian cell nuclear extracts 23-25 . Apart from its transcriptional co-activation function that facilitates RNA polymerase IIdriven gene transcription 26-32 , PC4 also plays an important role in various cellular process including DNA replication, DNA repair and chromatin organization 33-42 . PC4 also participated in the regulation of autophagy [43][44][45] . During the malignant transformation of normal dermal multipotent broblasts, we reported that PC4 is up-regulated and positively correlated with K-Ras and MAPK pathway, implying the potential role of PC4 in tumorigenesis for the rst time 46 . Our study and previous studies con rmed that PC4 is highly expressed in lung cancer 47  Autophagic cell death is different from apoptosis, known as type II programmed cell death, which is mainly characterized by the appearance of abundant vacuole enveloping cytoplasm and organelles, and the degradation of various components inside the vacuole via lysosome 15,18 . In response to metabolic stress, including hunger and energy de ciency, autophagy is mainly regulated by mTOR kinase 52 . In recent years, autophagic cell death has gained enormous attention. Suzanne found that combination of the antidepressants maputiline and uoxetine can induce autophagic cell death in drug-resistant Burkitt's lymphoma 13 . In multiple myeloma, metformin can induce autophagic cell death through the AMPK/mTOR pathway 14 . Furthermore, mTOR inhibitors have been widely used for the treatment of various cancers, and autophagic cell death is one of the main pathways. However, the knowledge of its regulatory mechanism is still insu cient. Galluzzi reported that a rapid reduction in energy charge below a critical limit is likely to trigger the cell death rather than an adaptive autophagic response 11  informed consent to participate in this study. All the participants had the opportunity to discuss any questions or issues. All the animal experiments in the study was carried out in compliance with the ARRIVE guidelines.

Consent for publication
Not applicable.

Competing interests
The authors declare no con ict of interest and any commercial a liations.  (12)