Homeostatic iron regulatory protein drives glioblastoma growth via tumor cell-intrinsic and sex-specific responses

Abstract Background Glioblastoma (GBM) displays alterations in iron that drive proliferation and tumor growth. Iron regulation is complex and involves many regulatory mechanisms, including the homeostatic iron regulator (HFE) gene, which encodes the homeostatic iron regulatory protein. While HFE is upregulated in GBM and correlates with poor survival outcomes, the function of HFE in GBM remains unclear. Methods We interrogated the impact of cell-intrinsic Hfe expression on proliferation and survival of intracranially implanted animals through genetic gain- and loss-of-function approaches in syngeneic mouse glioma models, along with in vivo immune assessments. We also determined the expression of iron-associated genes and their relationship to survival in GBM using public data sets and used transcriptional profiling to identify differentially expressed pathways in control compared to Hfe-knockdown cells. Results Overexpression of Hfe accelerated GBM proliferation and reduced animal survival, whereas suppression of Hfe induced apoptotic cell death and extended survival, which was more pronounced in females and associated with attenuation of natural killer cells and CD8+ T cell activity. Analysis of iron gene signatures in Hfe-knockdown cells revealed alterations in the expression of several iron-associated genes, suggesting global disruption of intracellular iron homeostasis. Further analysis of differentially expressed pathways revealed oxidative stress as the top pathway upregulated following Hfe loss. Hfe knockdown indeed resulted in enhanced 55Fe uptake and generation of reactive oxygen species. Conclusions These findings reveal an essential function for HFE in GBM cell growth and survival, as well as a sex-specific interaction with the immune response.

Glioblastoma (GBM) is the most common primary malignant brain tumor and has an exceptionally poor prognosis.When treated with standard-of-care therapy, which includes maximal safe surgical resection, radiation, and chemotherapy, the median survival for patients with GBM is approximately 17-20 months. 1 Multiple factors have been identified that facilitate GBM growth and therapeutic resistance, culminating in disease recurrence.Tumor cellintrinsic factors, including cellular heterogeneity, metabolic, and genetic aberrations, along with extrinsic factors within the neural, endothelial, and immune cell compartments, drive tumor progression and contribute to therapeutic failure. 2 Recently, patient sex has been reported as a determinant of incidence and survival in GBM, with male patients exhibiting greater risk of developing GBM, as well as worse prognosis, when compared to female patients. 35][6][7] As such, the intersection of sex and other tumor cell-intrinsic and cell-extrinsic features is a critical area of investigation.One biological process known to exhibit a large degree of sexual differences and to drive radioresistance and tumor progression in GBM is iron metabolism, an attractive focus for therapeutic development. 8,9ron is an indispensable element required for enzymatic function in a number of normal cellular processes such as ATP production, DNA synthesis, and cell cycle regulation. 10e brain is dependent on iron for normal development and function, playing a role in neurotransmitter synthesis, myelination, and microglia polarization. 11The redox cycling potential of iron contributes to its protumorigenic effects, including the generation of free radicals, which can damage DNA and promote malignant transformation. 10herefore, a complex regulatory mechanism involving iron uptake, storage, and release, mediated by a variety of ironhandling proteins, ensures the tight maintenance of intracellular iron levels to prevent toxicity.Dysregulation of iron metabolism, primarily driven by the aberrant expression of iron-handling proteins, is a hallmark of the tumor state, and increased import accompanied by reduced export is frequently observed in many different cancers. 12Iron is therefore a putative therapeutic target for cancer, although the use of iron chelators in vivo has been limited by their lack of tumor specificity and their side effects. 13GBM tumors exhibit increased iron uptake compared to nontumor brain tissue, a property that has been exploited for specific targeting and imaging of these lesions. 14,15Previous work has demonstrated that cancer stem-like cells within GBM tumors are efficient iron scavengers, upregulating their expression of specific iron-handling proteins to enhance uptake and storage and drive proliferation. 16Conversely, disrupting intracellular iron storage in GBM cells induces cell death by multiple parallel mechanisms, including irondependent ferroptosis. 17Given the pivotal role of iron in tumor initiation and growth, further insight into its regulation in GBM is necessary to improve understanding of this disease and to aid in the development of new therapies.

Importance of the Study
Dysregulation of iron metabolism is an important feature of glioblastoma (GBM) that contributes to tumor growth and negatively impacts survival.Iron regulation is a complex process that involves many pathway members, but the individual function of these genes is not well elucidated.We identify homeostatic iron regulator (HFE), a critical regulator of iron homeostasis, as important for GBM cell growth and survival.Our findings also suggest a sex difference in HFE where females are more sensitive to HFE alterations, which impact antitumor natural killer cell and CD8+ T cell activity.This ultimately results in differential survival outcomes in which females show a greater difference in HFE-dependent survival in preclinical model and human GBM patient outcome.Our findings demonstrate that HFE enables tumor cell proliferation and survival in GBM in a sex-specific manner and may be a viable target for modulating tumor iron flux, altering immune response, and inducing apoptosis in tumor cells.

Troike et al.: Tumor cell-intrinsic HFE drives glioblastoma growth
In a normal cellular state, iron homeostasis is maintained through a tightly regulated balance of iron import, export, and storage.When this balance is disrupted, iron can accumulate and lead to pathologies such as hereditary hemochromatosis (HH), an iron overload disorder.Dysregulated iron absorption in HH causes tissue iron accumulation, culminating in oxidative damage and cell death. 18Additionally, ferroptosis, an iron-dependent form of cell death, has been reported to contribute to tissue damage in mouse models of hemochromatosis. 19Most cases of HH are driven by mutations in the homeostatic iron regulator (HFE) gene, which encodes the transmembrane HFE protein. 20In normal cells, HFE functions as a cellular iron sensor, mediating the uptake and release of iron indirectly through interactions with other iron-associated proteins. 21Prior work in GBM has demonstrated an inverse relationship between HFE expression and survival in female patients. 22However, no study to date has mechanistically clarified how HFE modulates GBM patient survival.Here, we demonstrate that Hfe loss in glioma cells enhances iron uptake and the generation of reactive oxygen species (ROS), which in turn promotes tumor cell death and extends survival.In addition, Hfe manipulation was subject to sex differences in immune response, which likely underscores the sex differences observed in GBM patient outcomes.Taken together, these findings highlight the importance of HFE in iron homeostatic maintenance and its potential as a target for therapeutic management of GBM.

Perl's Prussian Blue Staining
Formalin-fixed paraffin-embedded patient samples were assembled as tissue microarrays and kindly provided by Dr. Bjarne Kristensen with approval by the Regional Committee on Health Research Ethics for Southern Denmark (S-20150148).Staining of tumor sections with Perl's Prussian blue and nuclear fast red counterstain was performed by the Lerner Research Institute Imaging Core.Images were acquired on a Leica Aperio slide scanner using a 20× objective.

Patient mRNA Expression and Survival
Clinical and microarray expression data for the IDH-wildtype subset of the glioblastoma cohort of TCGA was downloaded from the GlioVis portal (http://gliovis.bioinfo.cnio.es). 23Additionally, mutational data were obtained using RTCGAToolbox. 24Wild-type IDH1 status was confirmed by inspecting the somatic mutations in the analyzed cohort.Cox proportional hazard model from the survival R package 25 was used to evaluate effects of HFE expression on overall survival.Previous findings indicated that there is a complex interaction among age, sex, and HFE expression. 22We therefore updated the survival model to address previous statistical analyses challenges.For the full cohort analysis, we used the following covariates: Diagnosis Age, Sex, O 6 -methylguanine-DNA methyltransferase (MGMT) methylation status, and the interaction terms-Diagnosis Age × Sex, MGMT methylation status × Sex, HFE median expression group × Sex.Interaction terms were used to account for the potential presence of sex-specific effects.Similarly, we performed sex-specific analysis of HFE expression effects.In this case, the Cox proportional hazard model was adjusted for diagnosis age, sex, and MGMT methylation status.

Cell Culture
Syngeneic mouse GBM cell lines (CT2A, GL261, and KR158) were grown in adherent conditions in RPMI 1640 media with 10% fetal bovine serum and 1% penicillin-streptomycin. Media was replaced every other day, and cells were passaged with Accutase and phosphate-buffered saline when sub-confluent (70%-80%).Cells were maintained in humidified incubators at 37°C and 5% CO 2 .Mouse astrocytes were derived from 3-day-old C57BL/6 neonatal mice as previously demonstrated. 26

Cell Treatment With DFO and FAC
Deferoxamine (DFO; Sigma; D9533) was reconstituted in DMSO at a concentration of 76 mM and further diluted to a 5 mM stock solution in water.Ferric ammonium citrate (FAC; Sigma; RES20400-A7) was diluted in water at a concentration of 5 mM.Cells were treated with 10 μM DFO or 15 μM FAC for 3 days prior to collection and counting.

Cell Viability
For trypan blue exclusion, cells were washed with PBS, trypsinized, collected, and spun down at 300×g for 5 min.Cells were then resuspended in media, and 20 μL of cells were taken for counting.An equal volume of trypan blue (20 μL) was added to the cells and mixed thoroughly.A total of 10 μL of the mixture was applied to a cell counting slide (Bio-Rad; 1450003) and measured using a Bio-Rad Automated Cell Counter.

Radioactive Iron Uptake
55 Fe uptake was performed as previously described. 27ells were grown to 70%-80% confluence, washed, and incubated in serum-free RPMI 1640 medium for 24 h.The cells were incubated with 55 Fe-NTA (Perkin Elmer; NEZ043001MC) in the same medium for 4 h at 37°C in a 5% CO 2 incubator.The medium was aspirated and the cells were washed twice with 150 μM NaCl 100 μm EDTA to remove excess iron. 55Fe-NTA uptake was measured in triplicate wells by lysis in RIPA buffer followed by liquid scintillation counting.All values were normalized to total protein concentration as determined by Bradford assay.

Quantitative Real-Time PCR
RNA was extracted from cells using an RNeasy kit (Qiagen; 74134), and concentrations were measured using a NanoDrop spectrophotometer.cDNA was synthesized using qScript cDNA SuperMix (Quanta Biosciences; 101414-102).qPCR was performed in Fast SYBR Green Mastermix (Applied Biosystems; 01120793) and an Applied Biosystems QuantStudio 3. Primer sequences are shown in Supplementary Table 3.During qPCR analysis, threshold cycle values were normalized to Gapdh.

MISSION
pLKO.1-puro Non-Mammalian shRNA Control Plasmid (SHC002) and Hfe shRNA plasmids TRCN0000105417 (KD1) and TRCN0000105419 (KD2) were purchased from Sigma.Lentivirus was packaged in 293T cells with psPAX2 and pMD2.G using calcium phosphate transfection.Lentiviral particles were collected from media and concentrated using a PEGit virus precipitation solution.For viral transduction, cells were grown in 10-cm tissue culture plates, and concentrated lentivirus was added to cells with the addition of 1:100 concentration of polybrene (Sigma; TR-1003).After 24 h of incubation, media was changed, and cells were incubated for an additional 24 h prior to selection with puromycin (5 μg/mL).
For human cell model 3832, HFE shRNA plasmids TRCN000060018, TRCN000060019, TRCN000060020, TRCN000060021, and TRCN000060022 were purchased from Sigma.Lentivirus was packaged in 293T cells with psPAX2 and pMD2.G using calcium phosphate transfection.Lentiviral particles were collected from media and concentrated using a PEGit virus precipitation solution.For transfection, cells were grown in 6-well tissue culture plates, and concentrated lentivirus was added to cells.After 24 h of incubation, media was changed and cells were incubated for an additional 24 h prior to selection with puromycin (2 μg/mL).Control MISSION pLKO.1-puroNon-Mammalian shRNA Control Plasmid (SHC002) was also used.

Intracranial Tumor Implantation
Intracranial implantation experiments with syngeneic tumor cell lines were performed as previously described. 28Sixweek-old C57BL/6 male and female mice were anesthetized using inhaled isoflurane, and an insulin syringe attached to a stereotaxic apparatus was used to inject cells into the left hemisphere at a depth of approximately 3.5 mm.Each syringe was prepared with equal cell numbers suspended in 10 μL of null RPMI 1640 media (20 000 KR158 cells transfected with shcontrol or Hfe KD2; 10 000 CT2A or GL261 cells transfected with control vector or Hfe overexpression).Animals were monitored over time for the presentation of neurological and behavioral symptoms associated with endpoint.Investigators were blinded to experimental conditions while monitoring animals.All animal experiments were performed in compliance with institutional guidelines and were approved by the Institutional Animal Care and Use Committee of the Cleveland Clinic (protocol 2019-2195).

Immunophenotyping by Flow Cytometry
At the indicated time points, immune cell profiling was performed on the mice bearing Hfe overexpressing or knockdown glioma cells when endpoint symptoms were present.Animals were euthanized by CO 2 asphyxiation, and tumor-bearing brain, blood, and bone marrow were collected.Single-cell suspensions were prepared from the tumor-bearing hemisphere by enzymatic digestion using collagenase IV (Sigma) and DNase I (Sigma), followed by filtering through a 40-µm cell strainer.Blood was collected in EDTA-containing capillary tubes (RAM Scientific), followed by RBC lysis (BioLegend).Bone marrow was flushed from femur and tibia and filtered through a 40-µm cell strainer.For flow cytometry analysis, cells were stained with LIVE/ DEAD stain (Thermo Fisher) on ice for 15 min in dark.Following washing step with PBS, surface staining was performed on ice for 30 min in the dark with the following anti-mouse monoclonal antibodies diluted in Brilliant Buffer (BD Biosciences): CD45, CD11b, CD11c, B220, NK1.1, CD3, CD4, CD8, Ly6G, Ly6C, I-A/I-E, CD69, PD-1, and LAG3.After incubation, cells were washed and fixed with FOXP3 Fix/Permeabilization buffer (eBioscience) at 4°C overnight.Intracellular staining was performed with the following antibodies diluted in FOXP3 permeabilization buffer: Foxp3, Ki-67, CD206, CD68, CTLA-4, IFN-γ, TNF, and Granzyme B. For cytokine expression, cells were stimulated with Cell Stimulation Cocktail containing protein transport inhibitor (eBiosicence) for 4 h at 37°C and subjected to staining as described above.Stained cells were acquired by a spectral cytometer Aurora (Cytek), and the acquired data were analyzed with FlowJo software (v10, BD Biosciences).

Statistical Analysis
For 2-group comparisons, P values were calculated using unpaired t test.For multiple comparisons within one condition, a one-way ANOVA with Dunnett's multiple comparisons test was used.For multiple comparisons in multiple conditions, a 2-way ANOVA with Tukey's multiple comparisons test was used.Log-rank tests were used for in vivo survival analysis.All statistical analyses were performed using GraphPad Prism 9.All in vitro experiments were carried out in at least technical triplicate for each experimental group.Additional statistical details, including P values and sample size, can be found in figure legends.

HFE is Upregulated in GBM Tumors and is Associated With Worse Survival in Female Patients
Homeostatic iron regulator is a critical iron regulator, and its expression has been reported to correlate with GBM patient survival, although no mechanistic description has Troike et al.: Tumor cell-intrinsic HFE drives glioblastoma growth yet emerged to explain these observations. 22We therefore investigated the role of HFE in GBM by first determining whether HFE levels in GBM patient tumors differ from those in nontumor brain tissue.We utilized the GEPIA database 29 and found significantly elevated HFE expression in GBM compared to nontumor brain tissue (Figure 1A).Expression data from The Cancer Genome Atlas (TCGA) revealed a direct correlation between HFE and tumor grade, with the highest expression seen in GBM (grade IV) (Figure 1B).Using these data sets, we sought to investigate the survival effects of high (greater than median) and low (lower than median) HFE expression on overall survival in TCGA.In isocitrate dehydrogenase 1 (IDH1) wild-type GBM patients, we found a statistically significant difference in survival, with high-HFE patients displaying a poorer prognosis (Figure 1C).While no differences were observed in male GBM patients (Figure 1D), high HFE expression correlated with truncated survival in female patients (Figure 1E), confirming previous findings 22 and likely the cause of the difference observed in overall survival when not stratified by biological sex.Both HFE expression group assignment and the interaction term of HFE expression group with sex significantly affected observed survival of IDH1 wild-type GBM patient groups.Further, we confirmed that this effect was sex-specific (Figure 1F, Supplementary Table 1) and not driven by MGMT methylation status (Supplementary Figure 1).
As HFE primarily regulates iron status through modulation of other proteins, we also analyzed survival based on high and low expression of these iron-associated genes (Supplementary Table 2).Ferritin is an evolutionarily conserved iron storage protein.Similar associations between ferritin expression and survival in GBM have previously been reported, with a reduction of ferritin expression leading to reduced tumor cell growth. 16Importantly, we found that high expression of the ferritin subunits ferritin heavy chain (FTH) and ferritin light chain (FTL) predicted shorter survival in female GBM patients, while no differences were present in male patients.These data suggest an important tumor-intrinsic role for HFE in GBM that may contribute to sex-specific effects on survival.

Hfe Knockdown in Mouse Glioma Cells Induces Apoptosis in Vitro and Inhibits Tumor Initiation in Vivo
While HFE expression has been reported to inform GBM survival, 22 HFE function in tumor cells remains unclear.To directly assess HFE function, we genetically manipulated Hfe levels in tumor cells to determine the effect on cell growth and survival.We first evaluated baseline Hfe mRNA levels in 3 syngeneic mouse GBM models (CT2A, GL261, and KR158) compared to wild-type mouse astrocytes (Figure 2A).CT2A Hfe expression was significantly lower than that of astrocytes, while KR158 had significantly higher expression compared to mouse astrocytes.Based on these results, we utilized 2 separate, nonoverlapping shRNA constructs to perform genetic knockdown of Hfe in KR158 cells.Knockdown 1 (KD1) and knockdown 2 (KD2) reduced Hfe mRNA expression by approximately 60% and 80%, respectively, compared to a nontargeting shRNA (Figure 2B).This result was also confirmed in human GBM model 3832, where cell viability was also decreased with HFE knockdown (Supplementary Figure 3C and D).To determine the impact of Hfe knockdown on tumor cell growth in vitro, we first performed a trypan blue-exclusion assay and cell counts.By day 3 of growth, significantly fewer cells were present in both Hfe knockdown groups compared to the control (Figure 2C), and caspase 3/7 activity, a surrogate of apoptosis, was significantly increased in both knockdown cell lines, suggesting greater induction of cell death with Hfe reduction (Figure 2D).To directly quantify proliferation, we performed a dye-dilution assay in which cells were incubated with carboxyfluorescein succinimidyl ester (CFSE), a fluorescent dye that stains cells, and grown in culture.The cells were then collected, and fluorescence was quantified.Rapidly proliferating cells have a lower concentration of dye as it becomes diluted with each subsequent division, while slowly dividing cells have a higher dye concentration (Supplementary Figure 2A).CFSE dilution was significantly elevated with Hfe knockdown, indicating that the decreased cell number may also be driven by a decrease in the rate of proliferation (Supplementary Figure 2B).Based on the observation that elevated HFE expression in female GBM patients is predictive of poorer survival, we intracranially implanted control and Hfe-knockdown cells into male and female C57BL/6 mice to determine whether Hfe loss alters tumor initiation (Figure 2E).We found increased tumor latency with Hfe knockdown in both male and female recipient mice (Figure 2F and G).Interestingly, 3 of the female mice implanted with KD2 cells failed to present with symptoms after 100 days, at which point exploratory necropsy failed to reveal the presence of a tumor (data not shown).This trend of extended median survival was maintained when Hfe was knocked down in CT2A, which expresses a lower level of Hfe compared to KR158, with a longer tumor latency in KD2 conditions, which was the stronger knockdown (Supplementary Figure 2C and D).Taken together, these data suggest that loss of Hfe induces an apoptotic phenotype in mouse glioma cells and reduces tumor initiation in vivo.

Increased Hfe Expression in Mouse Glioma Cells Drives Proliferation and Tumor Initiation in Vivo
Based on the observation that CT2A and GL261 cells express lower and comparable levels of Hfe, respectively, compared to astrocytes, we performed genetic overexpression of Hfe in these cells using stable lentiviral transduction.Overexpression was validated by RT-qPCR in both models (Figure 3A and B).Increased cell number as assessed by trypan blue exclusion was observed at day 5 (Figure 3C and D).To determine whether enhanced proliferation driven by Hfe overexpression could account for the differences in cell number, we assayed cell division by measuring CFSE dye dilution.Hfe overexpression in both CT2A and GL261 cells significantly increased the proliferation rate compared to controls (Supplementary Figure 3A), with no difference in caspase activity (Supplementary Figure 3B).To determine the impact of this perturbation on in vivo tumor initiation, male and

Troike et al.: Tumor cell-intrinsic HFE drives glioblastoma growth
female mice were transplanted with vector control or Hfe overexpression cells from both CT2A and GL261 models.Female mice implanted with both overexpression cell lines succumbed to disease more quickly than those implanted with control cells, while males exhibited no difference in survival outcomes with overexpression, suggesting that tumor growth is augmented only in females when Hfe levels are high (Figure 3E-H).These findings are consistent with the results observed in Hfeknockdown conditions and reveal a role for Hfe in tumor cell proliferation and tumor initiation capacity in females.

HFE Overexpression Impacts the Immune Activation State in a Sex-Specific Manner
To better understand how Hfe alters survival in the context of sex differences, we performed immune profiling on mice implanted with both Hfe knockdown and Hfe overexpression constructs at an intermediate timepoint prior to the development of neurological symptoms, representing the experimental endpoint.We implanted cells with Hfe overexpression in C57BL/6 mice and performed immune-cell profiling on day 28 posttumor implantation (Figure 4A).We did not observe any clear differences in the frequency of infiltrated immune cells between male and females in these conditions (Supplementary Figure 4G-K).However, higher frequencies of CD69, an activation and cytotoxic functional marker, were observed in the natural killer (NK) cell population in males compared to females.Conversely, female NK cells expressed higher IFN-y compared to male NK cells.There was no difference in TNF production in CD8 T cells (Figure 4B).While some sex differences in cytokine expression were observed in NK cells, there was no clear difference between the control and Hfe overexpression group.
Next, we analyzed the immune landscape of C57BL/6 mice implanted with Hfe-knockdown cells to understand the contribution of immune components in the extended survival observed in Figure 2F and G. Immune profiling was performed at 49 days for males and 56 days for females when endpoint symptoms were present (Figure 4C).Similar to our observed Hfe overexpression results, CD69 frequency in NK cells was higher in male cells than female, but female Hfe-knockdown animals had increased CD69 frequency compared to female control animals (Figure 4D).IFN-γ production in NK cells was higher in control females compared to control males, and we observed a trend of increased IFN-γ production in males but not females implanted with Hfe-knockdown cells.Similar trends were observed in TNF expression in CD8+ T cells, while the sex differences in TNF production in baseline male and female control animals were maintained as previously observed. 7e then implanted Hfe knockdown cells in immunodeficient NOD scid gamma (NSG) mice and assessed for animal survival (Figure 4E).Overall, Hfe knockdown slightly increased survival (Figure 4F) but not to the extent observed in immune-competent recipient mice.This effect was more pronounced in male recipient mice, where median survival increased by 4 days (Figure 4G and H).Taken together with the differences in immunocompetent models (Figure 2F and G), these results suggest that immune system differences contribute to the survival differences observed between males and females with Hfe manipulation.

Hfe Knockdown Increases Iron Uptake and the Production of ROS
As Hfe is also known as an iron sensor and metabolic iron regulator, we tested the functional effect of iron depletion and supplementation on tumor cell growth.We first confirmed that the presence of iron can be detected in human GBM tissue using Perl's stain (Supplementary Figure 5A).We next treated mouse glioma cell models CT2A, GL261, and KR158 with the iron chelator DFO and iron donor FAC, which respectively inhibited and enhanced cell growth (Supplementary Figure 5B).We then investigated GBM iron uptake using radioactively labeled 55 Fe, which revealed significantly increased uptake in Hfe-knockdown cells compared to controls (Figure 5A).This is consistent with the known function of HFE as a competitive inhibitor of iron uptake via transferrin receptor 1 (TFRC) binding on the cell surface.Overexpression of Hfe had no effect on iron uptake (Supplementary Figure 5C and D).As HFE does not directly interact with iron but rather with other iron-handling proteins, we hypothesized that modulating Hfe expression in these cells would in turn disrupt the expression of other iron-associated genes.We investigated mRNA levels and observed a significant reduction in the expression of ferritin heavy chain (Fth1) (Figure 5B) and transferrin receptor 1 (Tfrc) (Figure 5C), indicating that less iron was stored intracellularly.Conversely, we observed a significant upregulation of ferroportin (Slc40a1) in knockdown cells compared to control (Figure 5D), indicating that more iron was being exported out of the cell.This was also visualized when we used Perl's stain to illustrate ferric (Fe 3+ ) iron.Downregulation of transferrin receptor, which is involved in iron import, and upregulation of ferroportin, which is responsible for iron export, may represent an attempt by Hfe-knockdown cells to reduce high intracellular iron levels caused by increased uptake.Furthermore, a reduction in ferritin suggests a limited ability of these cells to store extra iron, which could result in cytotoxicity and induce apoptotic cell death.Overexpression of Hfe did not impact expression of transferrin receptor or ferritin, although ferroportin expression was significantly decreased in Hfe overexpression cells (Supplementary Figure 5E-J).When compared to control, Hfe knockdown conditions showed more iron deposition in the tumor microenvironment (Supplementary Figure 5K-P).RNA extraction from endpoint tumors did not show notable changes in iron regulation genes in Hfe-knockdown mice, with the exception of transferrin receptor in male mice implanted with Hfe-knockdown cells (Supplementary Figure 6).
To investigate the mechanisms through which Hfe loss induces an apoptotic phenotype, we interrogated signaling pathway alterations after Hfe knockdown using the NanoString sequencing platform (Figure 5E).Several pathways were differentially expressed when comparing control cells to Hfe KD1 and KD2 (Figure 5F).Among these, oxidative stress was reported as the top pathway upregulated in both knockdown cell lines compared to control.Given our findings that Hfe loss induces a cell death phenotype, we measured the generation of ROS in these cells and observed a significant induction of hydrogen peroxide (H 2 O 2 ) production with Hfe knockdown compared to control (Figure 5G).H 2 O 2 -mediated apoptosis is well documented and can act through activation of caspases, 30,31 as shown following Hfe knockdown (Figure 2D).Collectively, these findings suggest that loss of Hfe function results in an iron overload phenotype and increased production of ROS, leading to cell death.

Discussion
Tumor cells co-opt iron metabolic programs to satisfy their high iron demand and support rapid proliferation.Targeting iron metabolism has emerged as a popular strategy for the development of clinical trials in a number of different cancers, including brain, prostate, colon, liver, and lung. 123][34][35] Thus, a better understanding of the molecular mechanisms governing aberrant tumor iron metabolism is necessary to improve therapeutic development.7][38] As an iron sensor, HFE is critical for the maintenance of intracellular iron homeostasis but indirectly regulates iron flux by binding and modulating expression of other iron-associated proteins.Previous work has described an association between increased HFE expression and truncated GBM patient survival. 22Accordingly, we find that HFE is upregulated in GBM tumors compared to nontumor brain tissue.However, the mechanisms underlying these observations have yet to be elucidated.One of the possible technical limitations of the cohort-based analysis is the usage of microarray gene expression data.RNA sequencing is considered to be a "gold standard" method, yet it was not available for the entire TCGA cohort, limiting statistical power.Importantly, consistency between microarray and RNA sequencing data in TCGA was not explicitly checked in our analysis, but downstream murine model experiments serve as a confirmation that our findings in TCGA could not be attributed to technical artifacts.
Iron uptake and storage can be exploited by tumor cells to enhance iron sequestration and promote tumor growth.GBM cancer stem-like cells demonstrate increased expression of iron uptake proteins transferrin (TF) and transferrin receptor (TFRC), as well as the iron storage protein ferritin.Knocking down either subunit of ferritin (FTH1 or FTL) in these cells was sufficient to reduce tumor sphere formation in vitro and tumor initiation in vivo. 16We also observed that Hfe perturbations impact survival of animals with intracranial tumors.We did not directly test a tumor initiation versus growth dependence for Hfe, which could be performed with inducible Hfe knockdown/knockout or overexpression constructs, with assessments of in vivo iron and ROS levels, apoptosis, and expression of ironrelated genes including ferritin, ferroportin, and transferrin receptor.These investigations represent the focus of future studies.Our findings demonstrate increased iron uptake with Hfe knockdown and a counterintuitive induction in apoptotic cell death.When we interrogated the impact of Hfe knockdown on iron-associated gene expression, we found a reduction in ferritin heavy chain (Fth1).Ferritin is essential for limiting the accumulation of intracellular ferrous iron (Fe 2+ ), which can be used in the Fenton reaction to catalyze the formation of superoxide and hydroxyl radicals (Fe 2+ + H 2 O 2 → Fe 3+ + HO • + OH − ). 39These free radicals can cause oxidative stress and induce several forms of cell death, including caspase-dependent apoptosis. 40ndeed, we found that Hfe knockdown induced ROS production in the form of H 2 O 2 , which is consistent with previous reports that ferritin degradation in human GBM cells promotes ROS generation and induces ferroptosis, an iron-dependent form of cell death. 17In fact, ferroptotic induction via delivery of iron oxide nanoparticles has been reported as an effective therapeutic strategy in animal models of GBM. 41This work demonstrated that increasing intracellular iron levels while simultaneously augmenting H 2 O 2 production results in potent ROS generation and ferroptosis with minimal off-target toxicity. 42Thus, a shift in focus from iron chelation, which broadly targets iron and other metal ions, to iron storage may represent a less toxic strategy in the treatment of GBM tumors (Figure 6).
Sex is an important determinant of tumor risk and survival outcomes in GBM. 3 Men are more likely to develop GBM at a ratio of 1.6:1 compared to females, and male GBM patients are reported to have worse prognoses, with an estimated median survival of 15.9 months compared to 22.6 months in female patients. 3Our work supports previous findings that tumor-intrinsic HFE exerts a sexspecific effect on overall survival, with high HFE expression ablating the survival advantage normally observed in female patients. 22The basis for these differences is still unclear but may be attributed to sex-mediated differences in iron metabolism.Men typically have larger iron stores than women, as much as 2-to 3-fold higher in some tissues, and women experience iron deficiency and anemia at much higher rates than men. 43,44Historically, this has largely been attributed to blood loss through menstruation and conditions such as pregnancy, but other factors such as hormones likely play a role.For example, there is an inverse relationship between serum ferritin and estrogen levels, which causes iron stores to gradually increase after menopause. 45However, both serum ferritin and iron stores are still reportedly lower in postmenopausal women than in men. 46,47A consequence of these sex differences in iron metabolism is that high serum iron levels are much more prevalent in men than women and have been associated with increased cancer risk. 41We observed truncated survival only in female mice implanted with Hfeoverexpressing tumor cells.This finding is consistent with human GBM clinical data and may be partially explained by sex differences in iron levels.Conversely, Hfe knockdown led to protracted survival in both male and female animals.We suspect that this survival benefit is mediated by the apoptotic phenotype and may therefore not discriminate based on sex despite sex differences in the immune response observed.Taken together, these results suggest a role for sex differences in HFE-mediated tumor iron regulation that ultimately results in the observed differential survival benefits.In addition, our findings indicate that Hfe manipulation impacts both the tumor cells in a cellintrinsic manner as well as the immune response in a sexspecific manner, suggesting that Hfe likely serves multiple roles in GBM.
While these findings are one of the first descriptions of HFE function in GBM, this work has limitations.Due to the complex regulatory nature of HFE, it is difficult to determine the direct target and sequence of events leading to apoptosis upon Hfe depletion.It is possible that these effects are mediated primarily by changes in iron status with HFE modulation.However, HFE also forms cell surface protein complexes that initiate downstream signaling and may directly influence ROS production or apoptosis.Further complicating our understanding of this process is the regulation of iron-associated gene and protein expression by iron availability.The mRNA transcripts of most iron-associated genes possess iron response elements on their 5ʹ untranslated regions, which regulate translation rates.Thus, HFE may modulate expression directly, through intracellular signaling, or indirectly, through iron availability, and more insight into this complex relationship Troike et al.: Tumor cell-intrinsic HFE drives glioblastoma growth is necessary to fully appreciate its role in tumor cells.Future studies will include epistatic rescue studies in Hfemanipulated conditions, including the potential requirement for ferritin heavy chain in this process, as well as the ability to rescue the effects of Hfe knockdown with ROS inhibition.Our studies demonstrate that Hfe changes impact iron regulation and many tumor cell phenotypes in vitro, but the extent to which this effect is maintained in vivo has not been extensively examined and represents a limitation of the current study.Another limitation is the influence of sex chromosomes on our mouse tumor cell lines.Upon karyotyping, we discovered that these lines have only one X chromosome, and therefore any contribution from sex chromosomes is not likely to be appreciated.Thus, our ability to thoroughly interrogate the role of tumor cell-derived sex differences is currently limited and  would rely on other models in which the sex chromosome complement is intact.The development of new syngeneic mouse GBM cell lines in which the cell sex is known would be beneficial to fully appreciate these differences and represents an area for future exploration.Another limitation is that these studies rely on mouse GBM models, and this therefore limits the ability to extrapolate these findings to human GBM and human GBM cells.While we show some indication that human GBM cells are also reliant on Hfe for growth in vitro and there is an association with GBM patient outcome, additional mechanistic studies are required to validate these findings in human models.Despite these limitations, HFE remains an intriguing mechanism for targeting tumor iron flux.In conclusion, our findings demonstrate that HFE drives tumor cell proliferation and may be a viable approach for targeting iron metabolism and inducing cell death.

#Figure 1 .HfeHfeFigure 2 .Figure 3 .
Figure 1.Homeostatic iron regulator (HFE) is expressed in glioblastoma (GBM) and correlates with poorer survival in female patients.(A) HFE mRNA expression from GEPIA in nontumor brain tissue (left, n = 207) compared to GBM tumor tissue (right, n = 163).(B) HFE mRNA expression from TCGA compared among glioma grades II-IV.(C) Overall survival data from IDH1 wild-type GBM patients stratified by high vs low HFE mRNA expression from TCGA microarray data.Male (D) and female (E) survival data from the same data set stratified by high vs low HFE mRNA expression.(F) Forest plot indicates that HFE expression is correlated with age and sex.

Figure 4 .
Figure 4. Sex-specific Hfe alterations in the tumor microenvironment.(A) Schematic of C57BL/6 mice implanted with CT2A control or Hfeoverexpressing cells.Tumor-bearing left hemispheres were taken on day 28 for immune profiling and analyzed via flow cytometry.(B) Frequencies of CD69 from NK cells, IFN-y production from NK cytokines, and TNF production from CD8+ T cells.(C) Schematic of C57BL/6 mice implanted with KR158 control or Hfe-knockdown cells.Tumor-bearing left hemispheres were taken on day 49 for males or day 56 for females for immune profiling and analyzed via flow cytometry.(D) Frequencies of CD69 from NK cells, IFN-y production from NK cytokines, and TNF production from CD8+ T cells.(E) Schematic of NSG mice implanted with KR158 control or Hfe-knockdown cells.(F) Kaplan-Meier survival curves of KR158 control and Hfe-knockdown cells intracranially implanted in male (G) and female (H) mice.n = 5 for each group; median survival is provided on plots.*P < .05;**P < .01;***P < .001one-way ANOVA with Dunnett's multiple comparisons test or log-rank test for survival data.Error bars represent standard deviation.

Figure 5 .
Figure 5. Loss of Hfe enhances iron uptake and induces the formation of reactive oxygen species (ROS).(A) Radioactive 55 Fe uptake normalized to total protein content.Hfe knockdown construct 2 increases 55 Fe uptake compared to control cells.(B-D) Iron-associated gene expression measured by RT-qPCR.Fold change compared to control and normalized to Gapdh is shown.(E) NanoString overall clustering and pathway score comparing Hfe knockdown to control in KR158 cells.(F) Pathway analysis demonstrates that oxidative stress is one of the more highly upregulated pathways.(G) H 2 O 2 production in Hfe-knockdown and control cells was measured by ROS-Glo and normalized to cell number.*P < .05;**P < .01;***P < .001determined by one-way ANOVA with Dunnett's multiple comparisons test.Error bars represent standard deviation.

Figure 6 .
Figure 6.Loss of tumor cell-intrinsic Hfe induces cell death through reactive oxygen species (ROS) generation.Under normal conditions (left panel) homeostatic iron regulator (HFE) associates with transferrin receptor (TFRC) on the cell membrane to competitively inhibit transferrin (TF)-bound iron uptake and maintain homeostasis.Our findings indicate that loss of HFE (right panel) permits greater iron uptake and results in downregulation of ferritin, leading to reduced iron storage, greater iron in the tumor microenvironment, greater production of ROS, and increased cytokines released by NK and CD8+ T cell cells that contribute to the apoptotic phenotype.Survival becomes even more dependent on sex when Hfe is reduced.