HRD1 Elicitation of An Anti-Warburg Effect by Targeting PFKP to Inhibit Breast Cancer Growth and Progression

Background: Our previous studies have shown that the E3 ubiquitin ligase of HMG-CoA reductase degradation 1 (HRD1) functions as a tumor suppressor, as overexpression of HRD1 suppressed breast cancer proliferation and invasion. However, its role in breast cancer cell glucose metabolism was unclear. Here, our aim was to uncover the role and molecular mechanisms of HRD1 in regulating aerobic glycolysis in breast cancer. Methods: The effect of HRD1 on robic glycolysis in breast cancer cells were assessed. Then the proliferation, colony formation ability, invasion and migration of breast cancer cells were evaluated. The relationship between HRD1 and PFKP was validated by Mass spectrometry analysis, immuno�uorescence and co-immunoprecipitation. The level of PFKP ubiquitination was measured using ubiquitylation assay. Furthermore, the tumor growth in vivo were observed by subcutaneous tumorigenesis in nude mice. Results: We found that upregulation of HRD1 clearly decreased aerobic glycolysis, and subsequently inhibited breast cancer proliferation and invasion. Mass spectrometry analysis results revealed a large HRD1 interactome, which included PFKP (platelet isoform of phosphofructokinase), a critical enzyme involved in the Warburg Effect in breast cancer. Mechanistically, HRD1 interacted and colocalized with PFKP in the cytoplasm, targeted PFKP for ubiquitination and degradation, and ultimately reduced PFKP expression and activity in breast cancer cells. Conclusions: Our �ndings reveal a new regulatory role of HRD1 in Warburg effect and provide a key contributor in breast cancer metabolism.


Background
Breast cancer is the leading cause of cancer-related mortality among women worldwide due to metastasis, recurrence, or chemoresistance [1].Breast cancer cell proliferation, survival, migration, and drug resistance depend on aerobic glycolysis (termed the Warburg Effect), a key feature of energy metabolism that converts glucose into lactate even in the presence of su cient oxygen [2][3].Growing evidence now demonstrates that relieving the Warburg effect can be e cacious in inhibiting breast cancer proliferation and invasion [4][5].Therefore, a better understanding the regulatory mechanism of the Warburg Effect, and especially of the key glycolytic enzymes involved in aerobic glycolysis, has the potential to identify glycolysis-based therapies for breast cancer.
One glycolytic enzyme, the platelet isoform of phosphofructokinase (PFKP), catalyzes the irreversible conversion of fructose-6-phosphate to fructose-1,6-bisphosphate and is the rst rate-limiting step in glycolysis.Breast cancer is characterized by high PFKP expression that, in turn, has been associated with decreased patient survival [6].Accumulating evidence has demonstrated that PFKP plays a crucial role in promoting aerobic glycolysis and lactate production in breast cancer cells, thereby stimulating cancer cell proliferation and metastasis [7][8].Deletion of PFKP can signi cantly inhibit lactate production to impair invasion and migration by breast cancer cells [9][10].Therefore, suppression of PFKP expression and activity represents a viable strategy for breast cancer treatment.
The expression and activity of PFKP are mainly regulated through post-translational modi cations such as acetylation, phosphorylation, and ubiquitination [11].The PFKP molecule contains 12 acetylation sites and 18 phosphorylation sites, and strong interactions have been demonstrated between acetylation and phosphorylation.
Acetylation of PFKP at K395 by lysine acetyltransferase 5 triggers its transport to the plasma membrane and phosphorylation at Y64.This, in turn, leads to the activation of AKT, which increases PFKP activity by the production of fructose-2,6-bis-phosphate and elevates PFKP expression by enhancing PFKP stability [12].In glioblastomas, PFKP S386 phosphorylation prevents the TRIM21-mediated ubiquitylation and degradation of PFKP, leading to an increase in aerobic glycolysis [13].However, the mechanism underlying the regulation of PFKP stability in breast cancer is unknown.
One possible regulator is HMG-CoA reductase degradation protein 1 (HRD1), also known as synoviolin.This is a RING nger domain-containing E3 ubiquitin ligase that plays an important role in tumorigenesis and tumor progression [14][15][16].HRD1 acts an E3 ligase, with biological roles closely related to its substrates.Our previous study demonstrated that HRD1 inhibited the growth and metastasis of breast cancer cells by targeting IGF-1R and that it sensitized breast cancer cells to tamoxifen by targeting S100A8 for ubiquitylation and degradation [17][18][19].In the current study, we used mass spectrometry analysis to identify an interaction between PFKP and HRD1 in MDA-MB-231 cells.Further study showed that HRD1 catalyzed PFKP ubiquitination and promoted PFKP degradation, indicating a potential involvement of HRD1 in aerobic glycolysis in breast cancer cells.

Cell lines and antibodies
Human breast cancer (MCF-7, MDA-MB-231) and embryonic kidney 293T (HEK293T) cell lines were cultured as previously reported [17].The HRD1 antibody for western blotting was obtained from Sigma Aldrich (St. Louis, MO, USA).PFKP antibody was obtained from Cell Signaling Technology.HRD1 antibody for IHC was purchased from Abgent (San Diego, California, USA).Antibodies against β-Actin, myc-Tag, and HA-Tag were acquired from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

Glucose assays
(1) Cellular glucose uptake was determined by ow cytometry analysis of uptake of 2-[N-(7-nitrobenz-2oxa-1, 3-diazo-l-4-yl) amino] -2-deoxy-D-glucose (2-NBDG).Breast cancer cells were incubated with 50μmol/L of 2-NBDG for 30 min at 37 °C in a humidi ed atmosphere containing 5% CO 2 /95% air.The uorescence intensity of 2-NBDG taken up by the cells was analyzed on an EasyCyte Plus Flow Cytometry System (Millipore, Chicago, IL, USA).(2) The glucose concentration in the medium was measured using an Amplex Red Glucose/Glucose oxidase assay kit (Invitrogen).The absorbance of the samples was measured using a Varioskan multimode microplate spectrophotometer (Thermo, MA) and used to calculate the concentration of glucose in the medium.

Measurement of lactate and ATP levels
Lactate levels in the extracellular medium and the intracellular lactate levels (determined from cell lysates) were measured using a lactate assay kit (Abcam, Cambridge, UK).Data were normalized to the nal cell counts.ATP content was determined by collecting the cells and lysing them in ATP lysis buffer, followed by ATP concentration determinations using an ATP assay kit (Beyotime Biotechnology, Shanghai, China) and following the manufacturer's protocol.
The in vivo ubiquitylation assay was run by transfecting MDA-MB-231 cells with ubiquitin-HA for 48 h, followed lysis in the denaturing buffer (6 mol/L guanidine-HCl [pH=8.0],0.1 mol/L Na 2 HPO 4 /NaH 2 PO 4 , and 10 mmol/L imidazole) containing 5 mmol/L N-ethylmaleimide to prevent de-ubiquitylation.The cell lysates were immunoprecipitated using the appropriate antibodies and the precipitates were washed and subjected to immunoblotting analysis.
Co-immunoprecipitation (Co-IP) and immuno uorescence (IF) staining MDA-MB-231 cells were grown to con uence and processed for Co-IP by standard procedures.Brie y, the cells (5 × 10 7 ) were lysed in cold lysis buffer, centrifuged, and 10% of the supernatants were used as inputs.The remaining supernatants were immunoprecipitated with protein A/G agarose beads, IgG (negative controls), or antibody for target proteins at 4°C overnight.After washing, the pellets were suspended in 2 × SDS and subjected to western blotting.

Migration, invasion, cell proliferation, and colony-formation assays
Stable cell lines of MCF-7 and MB231 cells overexpressing HRD1 were constructed and cultured as previously reported [17].Stable cell lines were treated with 5mmol/L 2-deoxy-D-glucose (2-DG) for 72 h and then cell proliferation was measured using CCK-8 assays.The migration and invasion assays were conducted on stable cell lines treated with 5mmol/L 2-DG for 48 h and then placed in serum-free media.
The colony-formation assays were performed by seeding stable cell lines in 6-well plates at a density of 200 cells per well and treating them with 5mmol/L 2-DG for 3 weeks.

Western blotting
Cells were washed twice in ice-cold PBS and then solubilized in radioimmunoprecipitation assay (RIPA) lysis buffer (Vazyme, Nanjing, China).Samples containing equal amounts of protein were analyzed by western blotting as previously reported [18].

Real-time PCR assay
Total RNA was isolated using TRIzol reagent (Invitrogen) following the manufacturer's protocol.The mRNA was quanti ed by real-time PCR using a LightCycler480 II Sequence Detection System (Roche, Basel, Switzerland).The following primers were used to identify PFKP: forward, 5′-CGG AAG TTC CTG GAG CAC CTC TC-3′ and reverse, 5′-AAG TAC ACC TTG GCC CCC ACG TA-3′.GAPDH was used as an internal control: forward, 5′-CCC CTT CAT TGA CCT TCA ACT A-3′ and reverse, 5′-GAG TCC TTC CAC GAT ACC AAA G-3′.All reactions were performed in triplicate, and the relative PFKP mRNA level was calculated using the 2 -ΔΔCT method.

Animal tumor model
All mouse experiments were approved by the Committee on the Ethics of Animal Experiments of Nanjing Medical University.MDA-MB-231 cells (1× 10 7 in 100 µL PBS) stably expressing HRD1 with or without overexpression of PFKP were injected into the ank region of 6-to 8-week-old female BALB/c nude mice.The tumor size was measured every 4 days, beginning one week after the implantation, and the tumor volume was analyzed using the formula V = 0.5 ×length ×width 2 .
Cells were injected into tail vein of the mice to generate the lung metastasis model.The mice (ten mice per group) were sacri ced 6 weeks after the injection.The size and weight of the lungs were assessed, and visible tumors on the lung surface were counted.

Statistical analysis
Statistical analyses were performed using SPSS 19.0 statistical analysis software.Data were expressed as the mean ± SD. Analysis of variance (ANOVA) was used to determine the statistical differences among the groups.A multivariate analysis of the independent prognostic factors was conducted using the Cox proportional hazards model.A P value < 0.05 was considered statistically signi cant.

Upregulation of HRD1 led to an obvious decrease in aerobic glycolysis in breast cancer cells
The role of HRD1 in breast cancer metabolism was examined by upregulating, HRD1 expression in MCF7 and MDA-MB-231 cells using lentivirus that over-expressed HRD1, as described previously [17].The resulting breast cancer cells that overexpressed HRD1 exhibited a signi cant decrease in glucose uptake, while the glucose level was signi cantly higher in the medium surrounding these cells (Fig. 1A and B).Consistently, HRD1 overexpression reduced lactate production and cellular ATP levels (Fig. 1C and D).The potential regulatory role of HRD1 in breast cancer glucose metabolism was further explored by ECAR and OCR assays to characterize the metabolic alterations in glycolysis due to HRD1 overexpression.HRD1 overexpression reduced ECAR (Fig. 1E), but had no effect on OCR (Fig. 1F).These data clearly supported a role for HRD1 in inhibiting glycolysis but not the tricarboxylic acid (TCA) cycle.
HRD1 elicited an anti-Warburg effect that inhibited growth, migration, and invasion of breast cancer cells Our previous study demonstrated that upregulation of HRD1 inhibited the growth and proliferation of breast cancer cells [17].The importance of the Warburg effect in HRD1-mediated anti-oncogenesis was investigated by treating breast cancer cells with 2-DG to inhibit glycolysis.CCK-8 assays con rmed that 2-DG inhibited the growth of MCF-7 and MDA-MB-231 cells (Fig. 2A), which was similar to previous study [20].Notably, 2-DG also abolished the suppressive effect of HRD1 overexpression on the growth of breast cancer cells (Fig. 2A) as well as the inhibitory effect of HRD1 overexpression on colony-formation by MCF-7 and MDA-MB-231 cells (Fig. 2B).The suppression of migration and invasion induced by HRD1 overexpression was also abolished by 2-DG (Fig. 2C and D).

HRD1 decreased the protein stability of PFKP
The molecular mechanism of HRD1 inhibition of growth of breast cancer cells was explored by mass spectrometry to investigate the proteins that bind to HRD1 in MDA-MB-231 cells.These experiments identi ed PFKP as one of the candidate substrates of HRD1 in breast cancer cells (Supplementary Table S1).Co-IP and IF staining assays con rmed a direct interaction between HRD1 and PFKP (Fig. 3A and B), and the interaction was intensi ed by overexpression of HRD1 (Fig. 3B).The Hrd1 and PFKP protein levels were negatively correlated in breast cancer tissues (P<0.001 by Spearman's correlation test; Fig. 3C).
The expression and activity of PFKP is mainly regulated through post-translational modi cations [11].HRD1 overexpression reduced the PFKP protein level and PFKP activity in MDA-MB-231 cells but had no effect on the level of PFKP mRNA (Fig. 3D-F).Treatment of MDA-MB-231 cells with cycloheximide (CHX) increased the rate of PFKP degradation in HRD1-overexpressing cells (Fig. 3G).However, overexpression of a HRD1 ligase-dead mutant (C291S) had no effect on PFKP expression, activity or protein stability in MB231 cells (supplemental Fig. 1A-C).These data indicated that HRD1-mediated PFKP degradation was dependent on HRD1 enzyme activity.

HRD1 acted as an E3 ubiquitin ligase for PFKP degradation
The molecular mechanism underlying the HRD1-associated decrease in PFKP protein stability was further evaluated using a proteasome inhibitor (MG132) and an autophagy inhibitor (bafilomycin A1, BafA1).Treatment with MG132, but not with BafA1, dramatically blocked the PFKP degradation induced by HRD1 overexpression (Fig. 4A), indicating that PFKP degradation mainly occurred through the proteasomal pathway.In vitro and in vivo ubiquitination assays con rmed that PFKP was a direct substrate of HRD1 (Fig. 4B and C).
The relationship between PFKP and HRD1 was further examined using the PFKP K10R mutant (K10R-PFKP).K10 is a ubiquitylation residue of PFKP [13], and the PFKP K10R mutant was resistant to HRD1mediated ubiquitylation when compared with the wild-type PFPK (Fig. 4D).This nding supported K10 as the site for HRD1-induced PFKP ubiquitination, and further study showed that K10R did not affect the PFKP/HRD1 interaction (Fig. 4E).Therefore, the resistance to HRD1-mediated ubiquitination ubiquitination was not due to effects on the interaction between HRD1 and PFKP by K10R.
Phosphorylation of PFKP at Ser386 also blocks its ubiquitylation and degradation.We investigate the importance of this Ser phosphorylation involved in HRD1-mediated PFKP degradation, by mutating this Ser to Asp (S386D) to mimic a constitutively phosphorylated form and by mutating the Ser to Ala (S386A) to prevent phosphorylation.HRD1 could bind to WT-PFKP and S386A-PFKP, but not to S386D-PFKP (Fig. 4F).In addition, the S386A-PFKP had a higher ubiquitylation level in MDA-MB-231 cells expressing HRD1 than in MDA-MB-231 cells transfected with a control vector.By contrast, S386D-PFKP was hardly ubiquitinated in either cell type (Fig. 4G).Similarly, overexpression of HRD1 promoted S386A-PFKP degradation but had no effect on S386D-PFKP stability (Fig. 4H).

HRD1 inhibited aerobic glycolysis, growth, migration, and invasion of breast cancer cells via PFKP downregulation
The role of PFKP in HRD1-mediated breast cancer inhibition was investigated by overexpressing PFKP in MDA-MB-231 cells stably expressing exogenous HRD1 (Fig. 5A).PFKP was simultaneously upregulated in the HRD1-overexpressing cells and this completely reversed, the decrease in glucose uptake, lactate production, cellular ATP levels, and ECAR caused by overexpression of HRD1 (Fig. 5B-F).The forced expression of PFKP also restored the growth, migration, and invasion that had been reduced by HRD1 overexpression in MDA-MB-231 cells (Fig. 5G-J).Similar results were also observed in MCF-7 cells (supplemental Fig. 2).Therefore, HRD1 appeared to inhibit aerobic glycolysis, growth, migration, and invasion of breast cancer cells by downregulating PFKP.

PFKP was responsible for the in vivo HRD1-mediated inhibition of breast cancer growth and metastasis
The role of PFKP in breast cancer tumorigenesis was investigated by intracranial injection of MDA-MB-231 cells stably expressing HRD1, with or without overexpression of PFKP, into athymic nude mice.HRD1 overexpression inhibited the tumor growth, whereas ectopic expression of PFKP reversed this inhibition (Fig. 6A-B).Western blotting and immunohistochemistry analyses revealed that overexpression of PFKP could restore PFKP levels in the tumor tissues expressing HRD1 (Fig. 6C).Inoculation of MDA-MB-231 cells stably expressing HRD1 and overexpressing PFKP into nude mice increased the number of metastatic nodules when compared with a group injected with cells only overexpressing HRD1 (Fig. 6D).

Discussion
Growing evidence indicates that aerobic glycolysis is the driving force of tumorigenesis in breast cancers [21][22].Here, we identi ed that elevated expression of HRD1 elicited an anti-Warburg effect, and inhibited breast cancer growth and metastasis both in vitro and in vivo.Mechanistically, HRD1 functions as an E3 ubiquitin ligase that catalyzed PFKP ubiquitination and promoted PFKP degradation.A model to depict the role of HRD1 in an anti-Warburg effect in breast cancer is presented (Figure -7).
In this study, HRD1 inhibited aerobic glycolysis in breast cancer cells, providing new evidence for a role as a regulator of tumor metabolism.Our previous studies had revealed that HRD1 had a tumor suppressive effect that involved suppression of breast cancer cell proliferation, migration and invasion [17][18][19].
Aerobic glycolysis is well known to be critical in driving breast cancer tumorigenesis and maintaining cancer cell proliferation [23][24].Hence, the anti-Warburg effect of HRD1 in breast cancer cells could contribute to its suppression of tumor development and progression.This possibility was supported by the nding that 2-DG, a nonmetabolisable glucose competitor, abolished the anti-tumor activity of HRD1 (Figure -2).
As an E3 ligase, HRD1 plays a biological roles that is closely related to its substrates.Here, our mass spectrometry analyses identi ed an interaction between HRD1 and PFKP, the rate-limiting enzyme in glycolysis, in breast cancer cells.The clinical signi cance of HRD-mediated degradation and decreased stability of PFKP was also indicated by the negative correlation between HRD and PFKP expression in human breast cancer tissues.The ubiquitination assays further con rmed that PFKP was a direct substrate of HRD1 and that HRD1 ubiquitinated PFKP at lysine 10.More importantly, restoration of PFKP expression abolished the tumor suppressive effects of HRD1 both in vitro and in vivo.
PFKP is overexpressed in breast cancer and promotes the Warburg effect [6,12], and PFKP expression and activity are mainly regulated by post-translational modi cations [11].Phosphorylation of PFKP at S386 is known to increase stability and pyruvate kinase activity [13], and our PFKP mutation plasmid with this Ser site mutated Asp (S386D-PFKP) to mimic constitutive phosphorylation prevented HRD1mediated the ubiquitylation and degradation.However, the nonphosophorylatable PFKP mutant (S386A) enhanced the binding of HRD1 to PFKP and had a higher ubiquitylation level than the WT-PFKP.This observation suggested that the non-phosphorylated state at S386 was indispensable for maintaining the PFKP ubiquitination induced by HRD1.

Conclusion
In summary, this study revealed a novel role for HRD1 in tumor metabolism.HRD1 inhibited glycolysis and tumor growth and metastasis through direct inhibition of PFKP expression in breast cancer cells.