Glutathione peroxidase 4 suppresses manganese-dependent oxidative stress to reduce colorectal tumorigenesis

The role of glutathione peroxidase 4 (GPX4) in ferroptosis and various cancers is well-established; however, its specific contribution to colorectal cancer has been unclear. Surprisingly, in a genetic mouse model of colon tumors, the deletion of GPX4 specifically in colon epithelial cells increased tumor burden but decreased oxidized glutathione. Notably, this specific GPX4 deletion did not enhance susceptibility to dextran sodium sulfate (DSS)-induced colitis in mice with varied iron diets but showed vulnerability in mice with a vitamin E-deficient diet. Additionally, a high manganese diet heightened susceptibility, while a low manganese diet reduced DSS-induced colitis in colon epithelial-specific GPX4-deficient mice. Strikingly, the low manganese diet also significantly reduced colorectal cancer formation in both colon epithelial-specific GPX4-deficient and wildtype mice. Mechanistically, antioxidant proteins, especially manganese-dependent superoxide dismutase (MnSOD or SOD2), correlated with disease severity. Treatment with tempol, a superoxide dismutase mimetic radical scavenger, suppressed GPX4 deficiency-induced colorectal tumors. In conclusion, the study elucidates the critical role of GPX4 in inhibiting colorectal cancer progression by regulating oxidative stress in a manganese-dependent manner. The findings underscore the intricate interactions between GPX4, dietary factors, and their collective influence on colorectal cancer development, providing potential insights for personalized therapeutic strategies.


Introduction
Ferroptosis, a distinctive process among various types of cell death, has garnered signi cant attention in cancer research.It is a mechanism through which cells, particularly cancer cells, undergo elimination due to the accumulation of excessive iron and lipid peroxides (Stockwell et al. 2017).Ferroptosis can be considered an intrinsic defense mechanism against cancer, especially given the heightened vulnerability of many cancer cells to ferroptosis due to their elevated metabolic rate and increased demand for iron and lipids.Consequently, inducing ferroptosis in cancer cells emerges as a promising therapeutic strategy.
While ferroptosis holds promise as a cancer treatment strategy, several challenges persist.For example, colorectal cancer, which accumulates substantial iron, has developed mechanisms to resist ferroptosis.
Notably, the expression of glutathione peroxidase 4 (GPX4), a key protective protein that reduces lipid peroxides and prevents ferroptosis, is upregulated in colorectal cancer patients undergoing neoadjuvant chemoradiotherapy, correlating with a poor prognosis (Zhang et al. 2022).
The resistance of colorectal cancer to ferroptosis poses a challenge in cancer treatment.Researchers are actively exploring ways to exploit ferroptosis as a therapeutic approach.Investigations involve the use of small molecules and drugs that either inhibit GPX4 or induce lipid peroxidation to selectively target cancer cells.Additionally, there is ongoing research into combining ferroptosis-inducing agents with traditional cancer therapies, such as chemotherapy or radiation.
A comprehensive understanding of the precise mechanisms underlying ferroptosis in cancer cells and the development of effective and safe therapies are ongoing areas of research.For example, a study demonstrated that the inhibition of serine/arginine-rich splicing factor 9 (SRSF9) enhances the sensitivity of colorectal cancer to erastin-induced ferroptosis by reducing GPX4 expression (Wang et al., 2021).In a separate study, sodium butyrate was found to promote ferroptosis by inducing lipid reactive oxygen species (ROS) production through the downregulation of GPX4, leading to reduced tumor growth in xenografts and a colitis-associated colorectal tumor model (Wang et al., 2023).Furthermore, the GPX4 inhibitor RSL3 was shown to hinder colorectal cancer cell growth by inducing ferroptosis, an effect reversible by overexpressing GPX4 (Sui et al., 2018).Moreover, the de ciency of GPX4 in myeloid cells has been demonstrated to actively promote the progression of colorectal cancer (Canli et al., 2017).Speci cally, this de ciency leads to an augmented production of ROS, instigating genome-wide DNA mutations within intestinal epithelial cells, ultimately fostering tumor invasion.However, no studies have investigated the in uence of colon epithelial-speci c GPX4 knockout on colorectal cancer progression.
Our ndings reveal that GPX4 depletion in colorectal epithelial cells unexpectedly enhances colorectal tumorigenesis.Mechanistically, under conditions of high oxidative stress, including vitamin E de ciency, elevated manganese (Mn) exposure, and tumor development, superoxide dismutase (SOD2) is upregulated in the colon tissues of colorectal epithelial GPX4 de cient mice.A low Mn diet and the SOD mimetic radical scavenger Tempol mitigate GPX4 de ciency induced SOD2 expression and colorectal tumor formation.These ndings underscore the crucial role of GPX4 in suppressing Mn-dependent oxidative stress to curtail colorectal tumorigenesis.

Depletion of GPX4 speci cally in colon epithelial cells promotes colorectal tumorigenesis.
To investigate the role of GPX4 in colorectal tumorigenesis, we generated mice with colon epithelial cell speci c Gpx4 depletion.GPX4 deletion was con rmed by genotyping (Figure S1A) and qPCR (Figure S1B).Colorectal cancer was induced as described in the methods, with the treatment scheme illustrated in Figure S1C.We observed that, compared to Cdx2 CreERT2 Apc F/+ mice, mice with colon epithelial-speci c GPX4 depletion (Cdx2 CreERT2 Apc F/+ Gpx4 F/F ) showed no difference in body weight change compared with GPX4 wildtype mice (Fig. 1A).However, GPX4 de cient mice exhibited increased colorectal tumor formation under a dissection microscope (Fig. 1B).The number of total tumors (Fig. 1C), tumors with a diameter less than 3mm (Fig. 1D), and tumors with a diameter more than 3mm (Fig. 1E) were all increased in the GPX4 de cient mice.Additionally, there was an increased tumor burden in the GPX4 de cient mice (Fig. 1F).H&E staining of colorectal tumors from GPX4 de cient and wildtype mice (Fig. 1G) indicated a similar disease score in both groups (Fig. 1H).Ki67 staining (Fig. 1I), and quanti cation (Fig. 1J) suggested increased cell proliferation in the GPX4 de cient mice, while the apoptosis marker CC3 was not changed (Fig. 1K and 1L).These results suggest a tumor-inhibiting function of GPX4 in colorectal cancer.
Depletion of GPX4 speci cally in colon epithelial cells results in reduced oxidized glutathione levels in colorectal tumors.
To understand the mechanism behind the enhanced colorectal tumorigenesis in GPX4-de cient mice, we conducted metabolomics analysis on colon tumors from Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice and Cdx2 CreERT2 Apc F/+ mice.Targeted metabolomics revealed two signi cantly changed metabolites (Table S2): Threonine and Homovanillic acid.In untargeted metabolomics, 41 signi cantly changed metabolites were identi ed in the negative mode and 85 in the positive mode (Table S2).Enrichment analysis of these 127 signi cantly changed metabolites using MetaboAnalyst 6.0 highlighted glutathione metabolism and Vitamin B6 metabolism as the top two enriched metabolite sets (Fig. 2A).This enrichment is due to decreased levels of oxidized glutathione (Fig. 2B), glycine (Fig. 2C), and spermidine (Fig. 2D) in the colon tumors of Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice compared to those in Cdx2 CreERT2 Apc F/+ mice.These ndings suggest a redox imbalance status due to the critical role of GPX4 in glutathione oxidation.
GPX4 depletion speci cally in colon epithelial cells increases susceptibility to colitis in mice treated with a vitamin E-de cient diet.
As in ammation is a high-risk factor for colorectal cancer and intestinal epithelial GPX4 restricts enteritis (Mayr et al, 2020), we investigated whether colon epithelial GPX4 de ciency in mice affects their susceptibility to colitis.There was no difference in body weight and colon length between Cdx2 CreERT2 Gpx4 F/F mice and Gpx4 F/F mice after treating mice with tamoxifen and DSS.Despite GPX4 being a wellknown ferroptosis inhibitor, no differences were observed in body weight and colon length between Cdx2 CreERT2 Gpx4 F/F mice and Gpx4 F/F mice with different iron concentrations in their diets (3.5ppm, 40ppm, 1000ppm) (Figure S2A-S2H).Similar results were seen after Salmonella treatment (Figure S2I and S2J).
A vitamin E supplement diet has been demonstrated to reduce GPX4 de ciency-enhanced ferroptosis in hematopoietic stem and progenitor cells (Hu et al. 2021).Interestingly, Cdx2 CreERT2 Gpx4 F/F mice exhibited increased body weight loss (Fig. 3A) and shorter colon lengths compared to Gpx4 F/F mice (Fig. 3B) when induced with DSS and treated with a vitamin E-de cient diet.qPCR analysis indicated that the expression level of Gpx4 was signi cantly decreased in the colon tissues of Cdx2 CreERT2 Gpx4 F/F mice compared to wildtype control (Figure S3A).Interestingly, proin ammatory cytokines including Tnfa (Figure S3B), Il6 (Figure S3C), Cxcl1 (Figure S3D) and Ptgs2 (Figure S3E), as well as stem cell marker Lgr5 (Figure S3F) was not changed.However, the anti-in ammatory cytokines including Il10 (Figure S3G), Il22 (Figure S3H), and transcription factor Foxm1 (Figure S3I) were signi cantly decreased in the colon tissues of Cdx2 CreERT2 Gpx4 F/F mice.qPCR analysis revealed a signi cant decrease in the expression level of Gpx4 in the colon tissues of Cdx2 CreERT2 Gpx4 F/F mice compared to the wildtype control (Figure S3A).Interestingly, proin ammatory cytokines, including Tnfa (Figure S3B), Il6 (Figure S3C), Cxcl1 (Figure S3D), Ptgs2 (Figure S3E), and the stem cell marker Lgr5 (Figure S3F), remained unchanged.However, anti-in ammatory cytokines, including Il10 (Figure S3G), Il22 (Figure S3H), and the transcription factor Foxm1 (Figure S3I), were signi cantly decreased in the colon tissues of Cdx2 CreERT2 Gpx4 F/F mice.The immunoblot analysis of several major antioxidant proteins, including SOD2, NQO1, and HO-1 (Fig. 3C), revealed an increase in these proteins in Cdx2 CreERT2 Gpx4 F/F mice, con rming a state of redox imbalance.H&E staining (Fig. 3D) and pathology scoring (Fig. 3E) indicated increased in ammation and tissue damage.While the proliferation marker Ki67 remained unchanged (Fig. 3F and 3G), an increase in the apoptosis marker CC3 was observed in colon tissues of Cdx2 CreERT2 Gpx4 F/F mice (Fig. 3H and 3I).This suggests that GPX4 de ciency contributes to enhanced apoptosis during the colitis process.Together, in the absence of vitamin E, mice de cient in GPX4 exhibit heightened susceptibility to DSSinduced colitis.
GPX4 depletion speci cally in colon epithelial cells increases intestinal in ammation in mice treated with a high Mn diet.
While GPX4 is a selenium-dependent enzyme (Weaver et al, 2022), SOD2 is one of the rare mitochondrial enzymes evolved to use Mn as a cofactor over the more abundant element iron (Naranuntarat et al, 2017).Compared with iron, Mn can catalyze the Fenton reaction more effectively to induce higher ROS production (Cheng et al, 2021).We found that when treated with a high Mn diet, Cdx2 CreERT2 Gpx4 F/F mice showed more body weight loss (Fig. 4A) and shorter colon lengths (Fig. 4B).Also, there was an increased expression of antioxidant proteins such as SOD2, NQO1, and HO-1 in colon tissues of Cdx2 CreERT2 Gpx4 F/F mice compared to Gpx4 F/F mice (Fig. 4C).However, H&E staining revealed a similar disease score (Fig. 4D and 4E).We found that cell proliferation, as indicated by Ki67 staining, remained unchanged (Fig. 4F and 4G).However, there was an observed increase in apoptosis, demonstrated by CC3 staining in the colons of Cdx2 CreERT2 Gpx4 F/F mice (Fig. 4H and 4I).These results indicate that in the presence of a high concentration of Mn, GPX4 de cient mice are more susceptible to DSS-induced colitis.
GPX4 depletion speci cally in colon epithelial cells reduces intestinal in ammation in mice treated with a low Mn diet.
To test whether Mn is required for enhanced colonic in ammation due to GPX4 de ciency, we treated mice with a low Mn diet.We observed that Cdx2 CreERT2 Gpx4 F/F mice experienced less body weight loss (Fig. 5A) and longer colon length (Fig. 5B) than Gpx4 F/F mice under Mn de cient conditions.Additionally, there was a reduction in the expression of antioxidant proteins such as SOD2, NQO1, and HO-1 in the colon tissue of Cdx2 CreERT2 Gpx4 F/F mice (Fig. 5C).H&E staining and pathological scoring revealed reduced in ammation in Cdx2 CreERT2 Gpx4 F/F mice when treated with a Mn de cient diet (Fig. 5D and  5E).We also noted Ki67 expression remained unchanged (Fig. 5F and 5G), while reduced apoptosis through CC3 staining in the colons of Cdx2 CreERT2 Gpx4 F/F mice (Fig. 5H and 5I), These results indicate that GPX4 depletion reduces intestinal in ammation in mice treated with a low Mn diet.
Mn de ciency reduces colorectal tumor formation in mice.
Mn plays a crucial role in intestinal epithelial barrier formation, and a low Mn diet has been linked to an increase in DSS-induced colitis in mice (Choi et al. 2020).However, whether Mn in uences the progression of colitis-induced colorectal cancer is not clear.To address this question, we induced colorectal tumor formation under three dietary conditions: a control diet, a Mn-de cient diet, and a high Mn diet.Unfortunately, the high Mn diet was associated with reduced mouse survival during DSS treatment (Figure S4).We observed that the Mn-de cient diet did not impact mouse body weight (Fig. 6A) but signi cantly reduced colorectal tumor formation (Fig. 6B).The Mn-de cient diet group exhibited decreased total tumor numbers (Fig. 6C), smaller tumors (sizes less than 3mm; Fig. 6D), larger tumors (sizes more than 3 mm; Fig. 6E), and overall tumor burden (Fig. 6F).H&E staining revealed lower disease scores in the Mn-de cient diet group (Fig. 6G and 6H).Ki67 staining demonstrated reduced tumor cell proliferation (Fig. 6I and 6J), while CC3 staining showed no change in apoptosis (Fig. 6K and 6L).
Western blot analysis indicated reduced levels of the antioxidant protein SOD2 and the Mn e ux protein ZnT10 in the Mn-de cient diet group, while the protein levels of HO-1 and NQO1 remained unchanged (Fig. 6M).Further analysis of metal levels in colon tumors via laser ablation ICP-MS revealed signi cantly lower Mn levels, but not other metals, in the Mn-de cient diet group compared to the control diet group (Figure S5A and S5B).The Mn-de cient diet fed to Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice showed a reduction in total tumor numbers (Figure S6A) and tumor numbers at size 2-3mm (Figure S6B) compared to the control diet.However, it led to no signi cant differences in smaller tumors (sizes less than 1mm; Figure S6C), tumor numbers at size 1-2mm (Figure S6D), and larger tumors (sizes more than 3 mm; Figure S6E), as well as overall tumor burden (Figure S6F).These results suggest that Mn is essential for colorectal tumor formation.
To investigate whether GPX4 de ciency-enhanced colorectal cancer depends on oxidative stress, we treated both Cdx2 CreERT2 Apc F/+ and Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice with the general SOD-mimetic agent tempol.We observed no impact of tempol on mouse body weight (Figure S7A); however, under the microscope, Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice exhibited increased colorectal tumor formation, a phenomenon that was effectively blocked by tempol treatment (Fig. 7A).Following tempol treatment, Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice demonstrated a reduction in total tumor number (Fig. 7B), smaller tumors (sizes less than 3 mm; Fig. 7C), larger tumors (sizes more than 3 mm; Fig. 7D), and overall tumor burden (Fig. 7E).H&E staining further indicated that tempol treatment inhibited the enhanced tumor formation seen in Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice (Fig. 7F).Furthermore, tempol treatment effectively inhibited the increased cell proliferation observed in Cdx2 CreERT2 Apc F/+ Gpx4 F/F mice (Fig. 7G and 7H) and downregulated the expression of antioxidant proteins, including HO-1 and SOD2 (Fig. 7I).However, CC3 staining revealed no signi cant change in apoptosis (Figure S7B and S7C).In conclusion, our results suggest that GPX4 de ciency may promote colorectal tumor formation by increasing ROS production.

Discussion
The major ndings of our study center around the crucial role of GPX4 in colorectal cancer progression.GPX4, a key player in ferroptosis and cancer, has an unclear role in colorectal cancer.An earlier study reported no signi cant overall association with colorectal adenoma risk for GPX4 (Peters et al., 2008).
However, a recent systematic review showed that carriers of the GPX4 (rs173041) T allele were associated with an increased risk of developing colorectal cancer (Barbosa et al., 2022).Another most recent study reports that high expression of GPX4, which suppresses ferroptosis, was associated with poorer 5-year overall survival only in KRAS mutant tumors from male colorectal cancer patients (Yan et al., 2023).Surprisingly, in our mouse model, GPX4 deletion in colon epithelial cells increased tumor burden.Treatment with tempol suppressed GPX4 de ciency-induced colorectal tumors, highlighting GPX4's crucial role in inhibiting oxidative stress in colorectal cancer progression.
Oxidative stress, a common feature in various human diseases, including colorectal cancer, is emerging as a signi cant contributor to colorectal cancer development.In colitis-associated colorectal cancer mice, oxidative stress is heightened in the colon (Lei et al., 2021).GPX4 is the sole enzyme capable of reducing toxic lipid hydroperoxides in biological membranes to the corresponding alcohols using glutathione as the electron donor (Zhang et al., 2021).This aligns with our ndings that the deletion of GPX4 resulted in decreased oxidized glutathione levels in colorectal tumors.Consistently, transgenic mice overexpressing GPX4 are protected against oxidative stress-induced apoptosis (Ran et al., 2004).Interestingly, in our study, apoptosis levels were signi cantly altered in the acute colitis model but not in the colitis-associated colorectal cancer mouse model, suggesting other compensatory apoptotic factors are involved in colorectal tumors.
GPX4 and SOD2 are two of the most important antioxidant defense enzymes that protect cells against oxidative stress (Fan et al., 2011).As a selenoprotein, GPX4 has been reported to play a major role in maintaining the oxidative phosphorylation system and protecting mitochondria from oxidative damage in gut epithelial cells (Cole-Ezea et al., 2012).One study reported signi cant two-loci interactions between rs4880 (SOD2) and rs713041 (GPX4), re ecting functional interactions between the gene products (Meplan et al., 2010).Deletion of SOD2 in skeletal muscle leads to no major impairment in whole-body metabolism, which is likely partly explained by a compensatory response that may exist from other redox enzymes, including GPX4 (Zhuang et al., 2021).Considering GPX4 deletion didn't worsen DSS-induced colitis with varied iron diets but showed vulnerability with a vitamin E-de cient diet and increased expression of Mn-dependent SOD2, we propose GPX4 and SOD2 coordinately regulate redox homeostasis.
ROS levels are higher in the ulcerative colitis region, but ROS scavenging enzyme SOD2 is barely detected in resident macrophages, resulting in distinct ROS vulnerability for in ammatory macrophages and resident macrophages (Du et al., 2023).SOD2 upregulation in cancer cells establishes a steady ow of H 2 O 2 originating from mitochondria that sustains AMP-activated kinase (AMPK) activation and the metabolic shift to glycolysis.Restricting SOD2 expression or inhibiting AMPK suppresses the metabolic switch and dampens the viability of transformed cells, indicating that the SOD2/AMPK axis is critical to support cancer cell bioenergetics (Hart et al., 2015).Also, the absence of SOD2 delays p53-induced tumor formation (Case and Domann 2014).Moreover, overexpression of miR-212 reduces the levels of SOD2 to block the epithelial mesenchymal transition process during colorectal tumor metastasis (Meng et al., 2013).Consistently, we found that antioxidant proteins, especially SOD2, correlated with DSS-induced colitis severity and colorectal tumor numbers in a Mn-dependent manner.
The consumption of antioxidant micronutrients, including Mn, does not modulate the effects of smoking on colorectal cancer risk (Hansen et al., 2013).However, residing in the proximity of industries releasing Mn increases colorectal cancer risk (García-Pérez et al., 2020).In the absence of GPX4, we found a high Mn diet increased susceptibility, while a low Mn diet reduced DSS-induced colitis and signi cantly decreased colorectal cancer formation.These ndings represent conceptual advances in understanding the nuanced impact of GPX4 in colorectal cancer development, shedding light on its role in oxidative stress regulation, particularly in a Mn-dependent manner.The study provides insights into the intricate interplay between GPX4, dietary factors, and their in uence on colorectal cancer progression.
In line with our results, hepatocyte restricted GPX4 loss does not suppress hepatocellular tumorigenesis (Conche et al., 2023).In hepatocellular carcinoma, ferroptosis does not provide a cell-autonomous tumor suppressor function but rather triggers an adaptive immune response, placing ferroptosis upstream of CD8 + T cells.Ferroptosis is a potent anticancer target for the treatment of hepatocellular carcinoma and colorectal cancer liver metastasis in combination with immune checkpoint and myeloid-derived suppressor cell blockade, while primary colorectal cancer is resistant to this combinatorial treatment.Our study found that the low Mn diet signi cantly curtailed colorectal cancer in both GPX4-de cient and wildtype mice.This provides an alternative strategy to treat colorectal cancer.
In conclusion, the study not only unravels the speci c mechanisms by which GPX4 in uences colorectal cancer but also suggests potential therapeutic strategies.The ndings may resonate with researchers and clinicians working in the elds of cancer biology, oxidative stress, and precision medicine.Moreover, the study's consideration of dietary factors adds a practical dimension, addressing potential preventive measures and highlighting the importance of personalized approaches in cancer care.
Mn de cient diet colorectal tumor formation.