Iron-catalyzed oxidative stress compromises cancer promotional effect of BRCA2 haploinsufficiency through mitochondria-targeted ferroptosis

Abstract

Pathogenic variants in BRCA2 are hereditary risks for various cancers, including breast, ovary, pancreas and prostate. Genomic instability due to insufficient homologous recombination is thought as responsible for carcinogenesis. Reportedly, endogenous or exogenous aldehydes, including formaldehyde and acetaldehyde, suppress BRCA2 function. However, molecular sequences how BRCA2 insufficiency leads to carcinogenesis remains unelucidated. To assess whether Fenton reaction-based oxidative stress is a promotional risk factor of carcinogenesis in BRCA2 haploinsufficiency, we here applied iron-induced renal carcinogenesis to a newly established rat heterozygous mutation model of Brca2 (mutant, T1942Kfs/+; MUT). Rat MUT model, despite significant increase in spontaneous malignant tumors, showed no promotional effect on renal carcinogenesis induced by ferric nitrilotriacetate (Fe-NTA) in contrast to our previous study using Brca1 mutant rats. Array-based comparative genome hybridization of renal cell carcinoma in MUT revealed significant increase in the frequency of homozygous Cdkn2A deletion. Whereas acute-phase analysis of the kidney after single or 1-week Fe-NTA administration to MUT showed suppressed lipid peroxidation, consistent with ferroptosis-resistance, ferroptosis and regeneration of tubular cells were coexistent with higher cytoplasmic catalytic Fe(II) levels in the subacute phase of MUT after 3-week Fe-NTA administration. Mechanistically, mitochondrial dysfunction with excess iron, promoted by insufficient BRCA2 presumably for maintaining DNA integrity, eventually initiated ferroptotic process. In conclusion, iron-dependent oxidative stress plays double-edged roles either for cell death or proliferation in carcinogenesis and its biological consequences are distinct between BRCA2 and BRCA1 haploinsufficiency. Our results suggest that iron-catalyzed oxidative stress is not a major driving force of carcinogenesis in BRCA2 pathogenic variants.

Graphical abstract

Highlights

• •Brca2(±) rats revealed increased cancer incidence.
• •Fe-driven oxidative stress did not elevate cancer risk in Brca2(±) rats.
• •Brca2(±) normally increases cytoplasmic Fe, leading to initial ferroptosis resistance.
• •Chronic Fe overload induced ferroptosis sensitivity via mitochondrial dysfunction.

Abbreviations

aCGH

array-based comparative genome hybridization

BSA

bovine serum albumin

CDK2/6

cyclin-dependent kinase 2/6

DFP

deferiprone

DMSO

dimethyl sulfoxide

FAC

ferric ammonium citrate

FBS

fetal bovine serum

Fer1

Ferrostatin-1

Fe-NTA

ferric nitrilotriacetate

FFPE

formalin-fixed paraffin-embedded

FPN

ferroportin 1

FtH/L

ferritin heavy chain/light chain

GSEA

gene set enrichment analysis

GPX4

glutathione peroxidase 4

GSK3β

glycogen synthase kinase-3β

γH2AX

γH2A histone family member X

HE

hematoxylin & eosin

HRP

horse radish peroxidase

IHC

immunohistochemistry

ip

intraperitoneal(ly)

IRP1/2

iron regulatory protein 1/2

KEAP1

Kelch-like ECH-associated protein 1

MFN1

mitoferrin 1

MUT

mutant [Brca2(T1942Kfs/+)]

NCOA4

nuclear receptor coactivator 4

Nrf2

nuclear factor erythroid 2-related factor 2

PCBP1/2

poly(repatedCytidine) binding protein 1/2

PBS

phosphate-buffered saline

PLIN3

perilipin 3

PVDF

polyvinylidene fluoride

RCC

renal cell carcinoma

REF

rat embryonic fibroblast

RIPA

radioimmunoprecipitation assay

RT

room temperature

SFXN3

sideroflexin 3

SD

Sprague-Dawley

SDS-PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis

SEM

standard error of means

SRY

Sex-determining region Y protein

TBST

Tris-buffered saline with Tween-20

TfR1

transferrin receptor 1

TPM

transcripts per million

TUNEL

terminal deoxynucleotidyl transferase dUTP nick end labeling

WT

wild-type

xCT

system x_c⁻ cystine/glutamate antiporter (SLC7A11)

1. Introduction

BRCA2 DNA repair associated (BRCA2) was cloned in 1995 [1] as a novel tumor suppressor gene and its pathogenic variants (PVs) are genetic risks for various cancers not only of breast and ovary but also of prostate and pancreas [2]. Frequency of germline PVs in BRCA2 is 0.17 % among the healthy female population in Japan [3]. The main function of BRCA2 is homologous recombination as error-free DNA repair system for DNA double-strand breaks. BRCA2 deficiency leads to genomic instability due to impaired homologous recombination, which is considered a key driver of BRCA2-associated carcinogenesis [4]. Alternatively, recent studies have shown that endogenous and exogenous aldehydes, such as formaldehyde and acetaldehyde, as well as the glycolytic metabolite methylglyoxal, can inhibit BRCA2 function [5,6]. Nevertheless, epidemiological reports indicate that exogenous mutagens, including radiation, smoking, and alcohol intake, show no association with carcinogenesis among BRCA2 pathogenic variant carriers [7,8], implying that the factors responsible for BRCA2-related carcinogenic risk have yet to be elucidated.

Genetically engineered animal models are powerful tools to understand the association between the gene and its biological significance. Whereas several mouse models of Brca2 mutation were generated in the 1990's and 2000's, conventional heterozygous Brca2 mutation mice thus far revealed no definite phenotype [9]. To our knowledge, there is only one carcinogenesis experiment among the animal models of Brca2 mutation, which paradoxically showed a delayed mammary/ovarian carcinogenesis in mice [10] by an oxidative stress inducer, 7,12-dimethylbenz [a]anthracene [11]. Therefore, the role of oxidative stress in BRCA2-deficiency-associated pathology is still controversial and the underlying mechanisms of oncogenesis mediated by BRCA2 dysfunction remain poorly understood.

Iron is an essential element for all the lives on the earth, but excess iron may lead to carcinogenesis via oxidative DNA damage, including double-strand breaks and DNA-protein crosslinks via Fenton reaction [[12], [13], [14], [15]]. Ferroptosis, an iron-dependent regulated necrosis accompanied by lipid peroxidation [[16], [17], [18]], can be regarded as a disease-prevention mechanism in carcinogenesis by evading excess accumulation of oxidative DNA modifications in cells [19]. Recently, BRCA1 deficiency was shown to induce ferroptosis-resistance via an increase in glutathione peroxidase 4 (GPX4) [20]. Furthermore, we have shown that BRCA1 haploinsufficiency promotes Fenton reaction-based carcinogenesis in rats [21]. In the present study we hypothesized that BRCA2-deficiency as well may interfere with susceptibility to ferroptosis.

Recently, we established a novel rat heterozygous Brca2 mutant model ([Brca2(T1942Kfs/+)]; MUT), which exhibited genome instability by increased DNA double-strand breaks in follicular granulosa cells of ovary and spermatocytes in testis, ending in female and male subfertility, respectively [22,23]. Here we have applied ferric nitrilotriacetate (Fe-NTA)-induced renal carcinogenesis protocol to this rat Brca2 MUT model to understand the role of Fenton reaction-based oxidative stress in BRCA2-deficiency-associated carcinogenesis. We unexpectedly found no promotional effect on carcinogenesis in BRCA2 haploinsufficiency and identified the responsible molecular mechanisms by analyzing the early stage of carcinogenesis.

2. Materials and methods

2.1. Materials

The materials used are as follows. Fe(NO₃)₃.12H₂O (Wako, Osaka, Japan), nitrilotriacetic acid disodium salt (Tokyo Chemical Industry, Tokyo, Japan), ammonium iron (III) citrate (Sigma-Aldrich, St. Louis, MO), DMSO (Invitrogen, Carlsbad, CA), RSL3 (Selleck, Houston, TX), Ferostatin-1 (Selleck), and deferiprone (DFP; Selleck). All chemicals used were of analytical grade.

2.2. Rat Brca2(T1942Kfs/+) model

All the protocols of animal experiments were approved by the Animal Experimental Committee of Nagoya University Graduate School of Medicine (M250298-002). A rat heterozygous mutation model of Brca2(T1942Kfs/+) (mutant, MUT; Jcl: SD, CLEA Japan, Inc, Tokyo, Japan) was established by CRISPR-Cas9 system [23]. Genotyping was performed by PCR as previously described [22,23].

2.3. Renal carcinogenesis protocol with ferric nitrilotriacetate (Fe-NTA) administration

Fe(NO₃)₃·12H₂O and disodium nitrilotriacetate were dissolved in deionized water at final concentrations of 300 mM and 600 mM, respectively. These solutions were immediately mixed at a 1:2 (v/v) ratio to prepare Fe-NTA, and the pH was adjusted to 7.4 using sodium carbonate [24]. All the rats used here were maintained under a standard diet (CE-2; CLEA Japan) at the laboratory animal facility of Nagoya University Graduate School of Medicine. Forty-seven male wild-type Sprague-Dawley rats (CLEA Japan; n = 19 for untreated control and n = 28 for carcinogenesis protocol) and 41 male Brca2(T1942Kfs/+) mutation rats (MUT; n = 19 for untreated control and n = 22 for carcinogenesis protocol) were enrolled in this study. As the renal carcinogenesis protocol, we injected intraperitoneally (ip) to rats at 4 or 5 weeks of age with Fe-NTA at a dose of 5 mg iron/kg for the first 2 days, followed by 7 mg iron/kg for the next 3 days for the 1st week, 10 mg iron/kg 5 times a week during the next 9 weeks and 10 mg iron/kg three times for the last week (a total of 11 weeks). However, in case the weight of rats decreased >10 % compared to the previous injection, the next injection was canceled. The rats were weighed every two weeks. The rats that showed tumor or significant weight loss were euthanized, and the carcinogenesis experiments were terminated 600 days after the final Fe-NTA injection. We dissected renal cell carcinoma (RCC), non-tumorous renal tissues and other major organs to evaluate metastasis and invasion. A half of them were fixed in 10 % phosphate-buffered formalin for hematoxylin & eosin (HE) staining or immunohistochemistry and the other half was frozen in −80 °C until analysis. Pathological diagnosis was derived from independent histopathologic reviews by two pathologists (YMa and ST). We determined whether a tumor is benign or malignant based on cellular morphology, growth pattern [25] and immunohistochemistry of Ki-67 (≥2 % for malignant pheochromocytoma [26]).

2.4. Acute and subacute study

As an acute model, male wild-type Sprague-Dawley rat and male Brca2(T1942Kfs/+) mutation rats (MUT) at 4 or 5 weeks of age were injected ip with Fe-NTA at a dose of 15 mg iron/kg and were euthanized at the specified time (3, 6 or 24 h) after the injection. As a subacute model, these rats were injected ip with Fe-NTA at a dose of 5 mg iron/kg for the first two days, followed by 7 mg iron/kg for the next three days for the first week (1-week model), and 10 mg iron/kg 5 times a week for the next 2 weeks (3-week model). The rats were euthanized 48 h after the final injection. The fresh kidney tissues were excised and processed as described above.

2.5. Immunoblot analysis

Immunoblot analysis was performed as described previously [27]. Briefly, tissue was homogenized in RIPA buffer with cOmplete Mini (Roche, Basel, Switzerland) and PhosSTOP (Roche). The lysates were centrifuged at 18,000×g for 30 min at 4 °C. The supernatant was used. Protein Assay Bicinchoninate kit (Nacalai Tesuqe, Kyoto, Japan) was used to measure protein concentration. Cytoplasmic and nuclear fractions were divided using the NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, Waltham, MA). The sample proteins were separated with SDS-PAGE, using 8 % or 12 % polyacrylamide gel and transferred to PVDF membrane, which were incubated in blocking buffer (5 % defatted milk or 1 % BSA/TBST) for 1 h at room temperature (RT). We incubated them with the primary antibody for 1 h to overnight and then HRP-conjugated secondary antibody for 1 h at RT, followed by reaction with Chemi-Lumi One Ultra or Super kit (Nacalai Tesque). Finally, the bands were visualized with LuminoGraph I (ATTO, Tokyo, Japan) and quantified with Fiji software as described [28]. Antibodies used in this study are summarized in Table S1.

2.6. Immunohistochemistry

Immunostaining was performed with BOND MAX/III (Leica, Wetzlar, Germany) using BOND Intense R Detection (DS9263; Leica) for HNEJ-1 or BOND Polymer Refine Detection (DS9800; Leica) for the other antibodies as described [29]. The images were captured with Cell Sens Dimension software (Olympus, Tokyo, Japan) and quantitated using Fiji. Antibodies used are summarized in Table S1.

2.7. Array-based comparative genomic hybridization (aCGH)

Genomic DNA from RCC samples was extracted using DNeasy Blood and Tissue Kit (Qiagen GmbH, Hilden, Germany), which was evaluated for quality and concentration by electrophoresis and NanoDrop2000 (Thermo Fisher Scientific). We labeled the DNA from 16 samples of rat primary RCCs (4 in wild-type [WT] with an early-onset of <300 days, 4 in WT with late-onset after 300 days, 4 in Brca2 MUT rat with an early-onset and 4 in Brca2 MUT rat with a late-onset; diameter ≥20 mm) with Cy-5, and the control DNA from MUT or WT with Cy-3, which were applied to the aCGH microarray slides (SurePrint G3 Rat CGH 4 × 180 k Microarray, G4826A#27064, Agilent Technologies, Santa Clara, CA) according to the Agilent Oligonucleotide Array-Based CGH for Genomic DNA Analysis protocol Ver.7.3. The image data from Agilent scanner was dissected with Agilent Feature Extraction Software 10.7. The original data can be accessed from the GEO database under the accession number GSE288318.

2.8. RNA sequencing and analysis

Total RNA was isolated from kidneys using NucleoSpin RNA Plus (#740984, Takara, Kusatsu, Japan). The concentration and quality of the RNA were evaluated with NanoDrop2000 and Agilent 2100 bioanalyzer. Preparation of RNA library and transcriptome sequencing were conducted by Rhelixa Inc. (Tokyo, Japan), using the HiSeq Illumina platform. The quality of the raw sequence reads was evaluated with FastQC (Version0.11.7; https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Low quality (<20) bases and adapter sequences were trimmed, using Trimmomatic software (Version 0.38) with the following parameters: ILLUMINACLIP: path/to/adapter.fa:2:30:10 LEADING:20 TRAILING:20 SLIDINGWINDOW:4:15 MINLEN:36. Next, the trimmed reads were aligned to the reference genome using RNA-seq aligner HISAT2 (Version 2.1.0) and the data was converted into bam files with Samtools (Version 1.9). The bam files were used to estimate the abundance of uniquely mapped reads by featureCounts (Version 1.6.3), and the raw read counts were normalized with transcripts per million (TPM) and differentially expressed genes were assessed with DESeq2 (Version 1.24.0). Gene set enrichment analysis (GSEA) was performed with GSEA v4.3.2, with the Rat_Ensembl_Gene_ID_Human_Orthologs_MSigDB.v2023.Hs.chip and gene set databases, WikiPathways. Visualization was performed using R Studio with ggplot2 [30,31] The original RNA-seq data can be accessed from the DRA (DDBJ Sequence Read Archive) database under the accession code DRR689908-689925, with BioProject Accession PRJDB20254.

2.9. Detection of formaldehyde

Formaldehyde concentration in the kidneys was measured by DFOR-100 formaldehyde kit (BioAssay Systems, Hayward, CA) according to the manufacturer's instructions.

2.10. Cell lines, media and cell viability

Rat embryonal fibroblasts were prepared from MUT and WT rats in-house as described [32]. They were cultured in DMEM (Wako), containing 10 % fetal bovine serum (FBS; Biowest, Nuaillé, France) and 1 % antibiotic-antimycotic mixed stock solution (100x) (Nacalai Tesuqe) at 37 °C in a humidified incubator with 5 % CO₂. For the evaluation of cell viability, 5000 cells/well were seeded in 96-well plates (Thermo Fisher Scientific) and incubated for 24 h at 37 °C. The cells were then treated as indicated, and the cell count reagent SF (Nacalai Tesque) was used to measure the cell viability.

2.11. Histological examination

Kidneys were removed 3 h after ip administration of 15 mg iron/kg Fe-NTA (n = 17, MUT and WT, respectively). Tubular injury area was evaluated as described [33].

2.12. TUNEL assay

Evaluation of cell death were performed on 4 μm-thick FFPE specimen using TUNEL Assay Kit-HRP-DAB (Abcam, Cambridge, UK) according to the manufacturer's instructions.

2.13. Analysis of catalytic ferrous iron

FerroOrange (RhoNox-4; Dojindo, Kumamoto, Japan) and Mito-FerroGreen (Dojindo) was used to detect cellular and intramitochondrial catalytic Fe(II) on the kidney frozen sections as described [34]. Photographs of the mounted slides were taken using a BZ-X800 microscope (Keyence, Osaka, Japan) and quantified using Fiji.

2.14. Electron-microscopical analyses

For the electron-microscopic analysis, renal cortical lesion was excised as cubes with 1–3 mm edge and immediately fixed in 2 % glutaraldehyde diluted in phosphate-buffered saline (PBS). Transmission electron microscopy was performed with a JEM-1400PLUS (JEOL; Tokyo, Japan) as described [35,36].

2.15. Statistical analysis

Statistical analysis was performed using an unpaired Student's t-test, two-way ANOVA, log-rank test, Pearson's chi-square test with GraphPad Prism 10.4.1 (627) software (San Diego, CA). Statistical significance levels (P values) are indicated as follows: ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001; ∗∗∗∗, P < 0.0001; ns, not significant. The data are shown as the means ± SEM.

3. Results

3.1. Despite high malignant tumor incidence, BRCA2-deficiency does not promote Fenton reaction-based carcinogenesis but favors Cdkn2a deletion

After establishing Brca2 mutant rats by using CRISPR/Cas9 technology [23], we conducted renal carcinogenesis experiments by injecting Fe-NTA to these MUT rats. Genotyping was done at least twice for confirmation (Fig. S1A). We first evaluated whether MUT rats reveal impairment of homologous recombination. MUT rats presented significantly lower induction of BRCA2 in the kidney 3 h after Fe-NTA administration in comparison to WT (Fig. 1A) whereas MUT rats showed significantly higher expression of γH2AX, one of the markers for DNA double-strand breakage (Fig. 1B). Based on these results, we confirmed that our animal model reproduces homologous recombination deficiency.

Fig. 1.

Despite high malignant tumor incidence, BRCA2-deficiency does not promote Fenton reaction-based carcinogenesis but favors Cdkn2a deletion

(A) Immunoblot analysis of BRCA2 expression of kidney samples after Fe-NTA single injection (W/WT, wild-type; M/MUT, mutant; N/NT, non-treated; n = 3 or 4; means ± SEM; ∗, P < 0.05). (B) Representative immunohistochemical images of γH2AX protein in kidney 3 h after Fe-NTA single injection (bar = 100 μm; n = 4; means ± SEM; ∗, P < 0.05). (C) Probability of survival curve under Fe-NTA-induced renal carcinogenesis experiment in the wild-type (WT; n = 20) and Brca2 heterozygous knockout (Mutant [MUT]; n = 17) groups. (D) A case of Fe-NTA-induced RCC in Brca2 MUT rat (left panel, macroscopic appearance; right panel, histology of the RCC; bar = 100 μm). Li, liver; Ca, carcinoma; arrow heads, peritoneal invasion. (E) Summary of aCGH analysis of early-onset RCC in WT and Brca2 MUT (n = 4, respectively). Individual data on all the 16 RCCs is included in Fig. S1C and Table S3. (F) Representative immunohistochemical images of p16^INK4A protein in RCCs (bar = 50 μm). See text for details.

In the untreated control experiments both of WT and MUT rats, we observed eight carcinomas (WT, 5.26 % and MUT, 36.8 %; P = 0.017; Table S2 and Fig. S1B), including hepatocellular carcinoma, adenocarcinoma of lung, histiocytic sarcoma and malignant pheochromocytoma, by day 710 after birth (equal to day 600 after the last injection) with full autopsy evaluation. Given the high spontaneous tumor incidence in MUT rats, we investigated whether iron-catalyzed oxidative stress contributes to the cancer-promoting effect of BRCA2 haploinsufficiency.

WT and MUT groups under Fe-NTA-induced renal carcinogenesis experiment generated RCCs in 20/28 (71.4 %) and 17/22 (77.3 %) of rats, respectively (Table 1). Surprisingly, there was no significant promotional effect of renal carcinogenesis in MUT in comparison to WT (Fig. 1C). RCCs of Brca2 MUT groups showed a lower but statistically not significant pulmonary metastasis (WT, 60.0 % and MUT, 47.1 %; P = 0.431; Table 1 and S3). RCCs of WT and MUT showed an incidence of peritoneal invasion/dissemination at 65.0 % and 70.6 %, respectively (Table 1 and S3). All the renal tumors were diagnosed as adenocarcinoma of renal tubular origin by immunohistochemistry of CD10 or vimentin, renal carcinoma markers. We applied the Fuhrman grade [37] to all tumors in Table S3. The representative case of Fe-NTA-induced RCC in MUT is shown (Fig. 1D). We used aCGH to examine genomic alterations of RCCs in MUT (Fig. 1E and S1C). RCCs from MUT as well as WT groups revealed no deletion of Brca2 with retained BRCA2 expression (data not shown), which was the same as the RCCs in Brca1 mutant rats [21]. Both RCCs revealed a high frequency of c-Met amplification (in chromosome 4) which we previously reported [21,38]. Of note, RCCs in MUT showed no amplification of c-Myc (in chromosome 7) which is the characteristic of the RCCs in the Brca1 mutant rats [21]. Additionally, we found homozygous Cdkn2a deletion in 3 cases (37.5 %) out of the eight MUT RCCs. No case of 8 WT RCCs showed homozygous Cdkn2a deletion. To confirm aCGH results, we performed immunohistochemistry (IHC) of p16^INK4a encoded by Cdkn2a (Fig. 1F and S1D). Indeed, increased deletion of Cdkn2a was confirmed with IHC in MUT RCCs in comparison to WT RCCs (4/8 in MUT RCCs vs 2/11 in WT RCCs; in total, 7/16 [43.8 %] in MUT RCCs vs 2/19 [10.5 %] in WT RCCs; P = 0.0251; Table S4).

Table 1.

Summary of Fe-NTA-induced renal carcinogenesis.

	Non-treated control		Fe-NTA carcinogenesis protocol
	Wild-type	Brca2 MUT	Wild-type	Brca2 MUT
Renal cell carcinoma (RCC)	0/19	0/19	20/28 (71.4 %)	17/22 (77.3 %)
RCC with pulmonary metastasis	0/0	0/0	12/20 (60 %)	8/17 (47.1 %)
RCC with peritoneal invasion/dissemination	0/0	0/0	13/20 (65 %)	12/17 (70.6 %)

Fe-NTA, ferric nitrilotriacetate; Brca2 MUT, Brca2(T1942Kfs/+) rat. Refer to text for details.

3.2. BRCA2-deficiency grants ferroptosis-resistance in the acute-phase of Fe-NTA-induced renal carcinogenesis

RNA-seq analysis in the acute phase indicated that gene sets related to ferroptosis were inversely enriched in the MUT rats (Fig. 2A and B). To ensure the results of comprehensive analysis, we assessed the necrotic area of kidney 3 h after Fe-NTA injection. As predicted, the necrotic area of kidney was significantly smaller in MUT than in WT (Fig. 2C). Further, we found that WT showed significantly more broad and intense immunostaining of HNEJ-1 which detects specific lipid peroxidation of ferroptosis [29] (Fig. 2D). To quantify the level of aldehydes, we examined the concentration of formaldehyde in kidney 3 h after Fe-NTA injection. The concentration of formaldehyde in kidney was significantly lower in MUT than in WT (Fig. S2A).

Fig. 2.

BRCA2-deficiency grants ferroptosis-resistance in the acute phase of Fe-NTA-induced renal carcinogenesis

(A) Top enriched gene sets in WT compared to Brca2 MUT by GSEA on kidney at 1 week in the Fe-NTA-induced renal carcinogenesis protocol using gene set of “WikiPathways” (n = 3). (B) Enrichment plots of the gene sets associated with ferroptosis. (C) Representative hematoxylin & eosin staining of kidney 3 h after Fe-NTA single injection (bar = 100 μm). Tubular injury area (area enclosed by dotted line) was evaluated (n = 17; means ± SEM; ∗, P < 0.05). (D) Representative immunohistochemical images of HNEJ-1 monoclonal antibody in kidney 3 h after Fe-NTA single injection (bar = 100 μm; n = 4; means ± SEM; ∗∗∗, P < 0.001). (E) Immunoblot analysis of ferroptosis-related proteins in kidney after Fe-NTA single injection (n = 3 or 4; means ± SEM; ∗, P < 0.05; ∗∗, P < 0.01; ns, not significant). Note that the same immunoblot membrane was used as in Fig. S2B. (F) Cell viability by WST assay on WT or Brca2 MUT rat embryonic fibroblast treated for 24 h with FAC or RSL3 in the presence of Ferrostatin-1 (1 μM). Unpaired two-tailed t-test or two-way ANOVA analysis; ∗∗∗, P < 0.001; ∗∗∗∗, P < 0.0001; ns, not significant.

Regarding ferroptosis-suppressor pathways, we found significantly higher expression of xCT, GPX4 and Nrf2 in the MUT kidney 3 h after Fe-NTA injection although there was no difference in KEAP1 (Fig. 2E). Nrf2, a master regulator of anti-oxidative responses, is negatively regulated by GSK3β, which is inactivated by AKT-mediated phosphorylation [39]. At the acute-phase, phosphorylation of GSK3β protein was found significantly increased in the MUT kidney (Fig. S2B).

To further confirm results in vivo, we used male rat embryonic fibroblasts (REF) from the corresponding rats (Fig. S2C and S2D). Brca2 MUT REF showed fewer necrotic cell death induced by ferric ammonium citrate or RSL3 treatment, which was canceled by Ferrostatin-1 (Fig. 2F). These results demonstrate that BRCA2-deficiency promotes ferroptosis-resistance in the acute phase of Fe-NTA-induced renal carcinogenesis (Fig. S2E).

3.3. BRCA2-deficiency promotes lipid peroxidation leading to ferroptosis via mitochondrial damage in the subacute phase of Fe-NTA-induced renal carcinogenesis

Next, we studied the kidneys at 3 weeks of carcinogenesis protocol. In this subacute-phase, phosphorylation of BRCA2 protein was significantly increased in MUT rats (Fig. 3A) whereas MUT rats showed significantly higher expression of γH2AX (Fig. 3B). Moreover, we observed that gene sets related to DNA damage response were highly enriched in the MUT kidney after 3 weeks of carcinogenesis protocol by RNA-sequencing (Fig. S3A and S3B). These data suggest that phosphorylated BRCA2 protein expression was upregulated in response to DNA double-strand breakage. We found significantly higher frequency of DNA strand breaks in the MUT kidney after 3 weeks of carcinogenesis protocol by using the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay (Fig. 3C). As an apoptosis marker, we assessed the expression of cleaved caspase-3 in this situation. Cleaved caspase-3 was decreased in the MUT kidney (Fig. S3C). We also found that MUT showed significantly more broad and intense renal tubular immunostaining of HNEJ-1 (Fig. 3D). To evaluate mitochondrial morphology, we performed electron microscopic analysis. Mitochondria became smaller, with significantly increased lysosomal and autophagosomal fractions in MUT kidneys both under basal conditions and after 3 weeks of the carcinogenesis protocol (Fig. 3E–G). Additionally, the number of cristae was significantly decreased whereas roundness of mitochondria was not changed (Fig. 3G). These results provide evidence that Fe-NTA-induced mitochondrial damage was exacerbated and that ferroptosis was executed in MUT kidneys after 3 weeks of the carcinogenesis protocol. Moreover, the expression of GPX4 was decreased in the MUT kidney at 3 weeks (Fig. 3H). On the other hand, xCT expression was higher in the MUT kidney at 3 weeks (Fig. S3D). In total, the balance for ferroptosis was directed promotional at the subacute phase in MUT. We thereafter sought to find the responsible molecular mechanism.

Fig. 3.

BRCA2-deficiency promotes lipid peroxidation leading to ferroptosis via mitochondrial damage in the subacute phase of Fe-NTA-induced renal carcinogenesis

(A) Immunoblot analysis of phospho-BRCA2(Ser2095) and BRCA2 in kidney at 3 weeks of Fe-NTA injection (n = 3; means ± SEM; ∗, P < 0.05). (B) Representative immunohistochemical images of γH2AX protein in kidney 3 weeks of Fe-NTA injection (bar = 100 μm; n = 3; means ± SEM; ∗, P < 0.05). (C) Representative images of TUNEL staining in kidney at 3 weeks of Fe-NTA injection (bar = 50 μm; n = 3 or 4; means ± SEM; ∗, P < 0.05). (D) Representative immunohistochemical images of HNEJ-1 monoclonal antibody in kidney at 3 weeks of Fe-NTA injection (bar = 100 μm; n = 3; means ± SEM; ∗, P < 0.05). (E) Representative electron microscopic images of mitochondria in kidney (yellow arrows, small-sized deformed mitochondria in the MUT in the untreated control; blue arrows, shrunken mitochondria with immature cristae in the MUT at 3 weeks in the Fe-NTA-induced renal carcinogenesis protocol; ∗, lysosome; ∗∗, autophagosome; bar = 2.0 μm in the left panels; 1.0 μm in the right panels). (F) Analysis of the number of lysosome and autophagosome per 20 μm². (n = 18–21; means ± SEM; ∗, P < 0.05; ∗∗∗, P < 0.001). (G) Analysis of mitochondrial area, number of cristae and mitochondrial roundness (n = 24–31; means ± SEM; ∗, P < 0.05; ∗∗∗∗, P < 0.0001; ns., not significant). (H) Representative immunohistochemical images of GPX4 protein in kidney at 3 weeks of Fe-NTA injection (bar = 100 μm; n = 3 or 4; means ± SEM; ∗, P < 0.05).

3.4. BRCA2-deficiency increases catalytic ferrous iron and transport iron in mitochondria

We next investigated catalytic Fe(II) to determine whether the Brca2 mutation affects iron metabolism. Indeed, FerroOrange fluorescence, indicative of catalytic Fe(II), was significantly increased in the MUT kidney under both untreated control conditions and after 3 weeks of the carcinogenesis protocol (Fig. 4A).

Fig. 4.

BRCA2-deficiency increases catalytic ferrous iron and transport iron in mitochondria

(A) FerroOrange staining of the frozen sections of kidney (bar = 50 μm; n = 3; means ± SEM; ∗, P < 0.05). (B) Immunoblot analysis of iron-related proteins in kidney at 3 weeks of Fe-NTA injection (n = 3; means ± SEM; ∗, P < 0.05; ∗∗, P < 0.01; ns, not significant). (C) Immunoblot analysis of MFN1 and SFXN3 in kidney at 3 weeks of Fe-NTA injection (n = 3; means ± SEM; ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001; ns, not significant). (D) Mito-FerroGreen staining of the frozen sections of kidney (bar = 50 μm; n = 3; means ± SEM; ∗, P < 0.05; ns, not significant).

We next examined the expression of iron metabolism-related proteins. In the untreated control condition, the MUT kidney exhibited higher levels of IRP1, IRP2, FPN, NCOA4, DMT1, PCBP1, and PCBP2 compared to the WT kidney (Fig. 4B). No differences were observed in the expression of TfR1, FtL and FtH between WT and MUT kidneys under control condition (Fig. 4B). However, after 3 weeks of the carcinogenesis protocol, the MUT kidney showed decreased levels of IRP1, IRP2, TfR1, FPN, FtL and FtH (Fig. 4B). In contrast, NCOA4, PCBP1, and PCBP2 levels were increased in the MUT kidney at this time point whereas DMT1 expression remained unchanged between WT and MUT kidneys (Fig. 4B).

Given the importance of mitochondria in ferroptosis, we hypothesized that mitochondrial iron overload occurs in the MUT kidney after 3 weeks of the carcinogenesis protocol. Two known mitochondrial iron importers from the cytosol are MFN1 and SFXN3 [40]. SFXN3 was constitutively expressed in the kidney and showed a significant increase in the MUT under untreated condition; its expression was not significantly increased by Fe-NTA treatment for 3 weeks (Fig. 4C). In contrast, MFN1 expression was low under untreated condition but was significantly induced by Fe-NTA in both WT and MUT kidneys, with a more pronounced increase in the MUT (Fig. 4C). Furthermore, Mito-FerroGreen fluorescence intensity, indicative of catalytic Fe(II) in mitochondria, was significantly elevated in the MUT kidney after 3 weeks of the carcinogenesis protocol (Fig. 4D).

To directly ascertain whether mitochondria have a critical role of ferroptosis in MUT, we performed rescue experiments by using deferiprone (mitochondrial iron chelator) [41,42]. DFP inhibited FAC-induced ferroptosis in MUT but not in WT cells in vitro (Fig. S4A).

Taken together, these results indicate that BRCA2 deficiency increases iron demand and promotes mitochondrial iron transport in MUT kidneys. This mitochondrial iron overload was thought sufficient to induce ferroptosis in the MUT kidney at 3 weeks of carcinogenesis protocol (Fig. S4B).

3.5. BRCA2-deficiency causes cell-cycle acceleration in the subacute-phase of Fe-NTA-induced renal carcinogenesis

To better understand why BRCA2-deficiency increases catalytic Fe(II) at 3 weeks of carcinogenesis protocol, we analyzed the results of RNA-seq in the subacute-phase. RNA-seq analysis revealed that gene sets related to mitochondria were highly enriched in the MUT rats (Fig. S5A). It is well known that mitochondrial function is highly associated with cell proliferation. Given that ferroptosis was promoted in the MUT kidney at 3 weeks of carcinogenesis protocol, we predicted an acceleration of cell proliferation in this situation. As expected, gene sets related to cell-cycle were highly enriched in the MUT rats (Fig. 5A). Moreover, Ki-67 cellular proliferation index was significantly higher in the MUT kidney at 3 weeks (Fig. 5B). To confirm these results, we examined the expression levels of proteins associated with cell cycle. We observed that MUT kidney at untreated control and 3 weeks of carcinogenesis protocol showed higher levels of CDK2, CDK6, Cyclin D1 and Cyclin D3 with the high expression of p21 (Fig. 5C).

Fig. 5.

BRCA2-deficiency causes cell-cycle acceleration in the subacute phase of Fe-NTA-induced renal carcinogenesis

(A) Enrichment plots of the gene sets associated with cell-cycle. (B) Representative images of Ki-67 protein in kidney at 3 weeks of Fe-NTA injection (bar = 100 μm; n = 3 or 4; means ± SEM; ∗, P < 0.05). (C) Immunoblot analysis of cell-cycle-related proteins in kidney at 3 weeks of Fe-NTA injection (n = 3; means ± SEM; ∗, P < 0.05; ∗∗, P < 0.01, ∗∗∗, P < 0.001). (D) Immunoblot analysis of p21 in the cytoplasmic or nuclear fraction in kidney at 3 weeks of Fe-NTA injection. Cytoplasmic p21 is normalized by β-actin and nuclear p21 is normalized by LaminB1 (n = 3 or 4; means ± SEM; ∗, P < 0.05). (E) Immunoblot analysis of phospho-AKT (Ser473) and AKT in kidney at 3 weeks of Fe-NTA injection (n = 3; means ± SEM; ∗, P < 0.05).

Because nuclear p21 is established to inhibit CDK2 [43], we analyzed the localization of p21. Notably, nuclear translocation of p21, a cell cycle suppressor, was significantly suppressed in the MUT kidney at 3 weeks of carcinogenesis protocol (Fig. 5D). We also found that phosphorylation of AKT protein was significantly increased (Fig. 5E), which is consistent with the observation that nuclear translocation of p21 was inhibited.

4. Discussion

In the present study, we established a novel Brca2 haploinsufficient rat model in the context of carcinogenesis and demonstrated that BRCA2 haploinsufficiency suppresses the promotion of Fe-NTA-induced renal carcinogenesis. We observed no significant change between WT and Brca2 MUT in the survival curve and RCC incidence in contrast to the results of Brca1 mutants [21], which significantly promoted this renal carcinogenesis. This is consistent with a previous study that Brca2 mutant mice significantly slowly develop chemical carcinogen-induced tumors [10]. Here Brca2 mutant cells suffer increased carcinogen-induced apoptosis, which eliminate premalignant cells [10], supporting our conclusion that higher susceptibility to ferroptosis during carcinogenesis cancels promotion of Fe-NTA-induced carcinogenesis.

In human epidemiological studies, exogenous mutagens revealed no association with carcinogenesis among BRCA2 carriers whereas no such tendency was observed among BRCA1 carriers [7,8]. Of note, a recent study showed that environmental and exogenous aldehydes induce genome instability and transformation in BRCA2 deficient human cells in vitro [5]. Because our carcinogenesis models generate a variety of such aldehydes [44], including formaldehyde (Fig. S2A), no promotional effect of renal carcinogenesis in Brca2 MUT is in contradiction with the previous study [5]. One possible interpretation of this contradiction is that initiating alone, which causes accumulation of mutations in DNA due to dysfunctional BRCA2, is not sufficient for BRCA2-deficiency-associated carcinogenesis. Cells have evolved mechanisms, such as apoptosis and ferroptosis, to eliminate cells harboring excessive DNA damage and thus mutations. As carcinogenesis requires proliferation of mutated cells, BRCA2-deficiency may inhibit such promotional effect.

Our findings provide evidence that iron-catalyzed oxidative stress does not promote BRCA2-deficiency-associated carcinogenesis. Given the evidence that BRCA1-deficiency significantly promotes Fe-NTA-induced renal carcinogenesis [21], there should be different molecular mechanisms between BRCA1-deficiency- and BRCA2-deficiency-associated carcinogenesis. One possible explanation for this difference is that BRCA1 works as both DNA damaging sensor and effector whereas BRCA2 works only as a mediator in homologous recombination [4]. Therefore, it is likely that these different functions of DNA repairs contribute to disparate results of Fe-NTA-induced renal carcinogenesis between BRCA1-and BRCA2-deficiency. Another possibility to explain our results is based on another BRCA2 function other than DNA repair. BRCA2 works for the stabilization of stalled DNA replication forks [45] and spindle assembly checkpoint [46]. Such an unexpected function of BRCA2 may contribute to our present results.

Our findings highlight the importance of the susceptibility of ferroptosis in carcinogenesis. In the context of carcinogenesis, ferroptosis is a suppressive means, such as removing cells that underwent iron-induced oxidative DNA damage whereas acquisition of ferroptosis-resistance, with or without genome alteration, is a major mechanism of Fe-NTA-induced renal carcinogenesis [19,47,48]. Recent work has shown that BRCA2-deficiency leads to an increase in oxidative stress [49]. These data indicate that BRCA2-deficient cells may survive in an oxidative stress environment via activating ferroptosis-suppressors. In agreement with these results, BRCA2-deficiency acquired ferroptosis-resistance compared to WT in vitro and in vivo and showed significantly higher expression of ferroptosis-suppressors; xCT, GPX4 and Nrf2 in the MUT kidney in the acute phase of carcinogenesis protocol. On the other hand, we found that repeated oxidative stress induced vulnerability to ferroptosis. This phenomenon could explain the results that there was no difference between WT and BRCA2-deficiency in Fe-NTA-induced renal carcinogenesis in spite of higher fraction of DNA double-strand breaks. Furthermore, we found a higher incidence of Cdkn2a deletion in MUT RCCs. Cdkn2a deletion is specific genome alteration in iron-induced carcinogenesis, including RCCs [38,50] and mesothelioma [51]. The high frequency of Cdkn2a deletion we observed may reflect genome instability in BRCA2-deficiency.

Iron is a double-edged sword. Excess iron generates oxidative DNA damage and finally induces ferroptosis [14,15]. On the other hand, iron is an essential nutrient for important biological processes, including oxygen transport, DNA replication and DNA repair [52]. Ribonucleotide reductase is the first enzyme in DNA synthesis and Fe(II) is necessary as its cofactor [53]. It is well known that many enzymes associated with DNA repair require iron-sulfur cluster as their cofactor [54]. XPD and FANCJ DNA helicase are representative examples [55]. Even in the normal condition, BRCA2-deficiency showed higher accumulation of catalytic Fe(II). Higher expression of IRP1 and IRP2 indicate that Brca2 mutant cells are relatively iron-deficient, and these cells utilize Fe(II) by upregulating cytosolic Fe(II) chaperones, PCBP1 and PCBP2. At the 3 weeks of carcinogenesis protocol, lower expression of IRP1, IRP2 and TfR1 in MUT kidney suggests iron overload in this situation, consistent with higher accumulation of catalytic Fe(II). In contrast, higher NCOA4 and lower FtH and FtL indicate that ferritinophagy was activated despite iron overload condition. In this condition, BRCA2-deficiency showed accumulation of catalytic Fe(II) in mitochondria via higher expression of MFN1, leading to mitochondrial iron overload in the MUT kidney at 3 weeks of carcinogenesis protocol. We observed gene sets related to iron-sulfur cluster biogenesis were highly enriched in MUT rats at 3 weeks of carcinogenesis protocol (Fig. S5B). Iron can enhance DNA repair, promote cell survival and cause platinum-based chemotherapy resistance [56]. Paradoxically, BRCA2-deficiency under iron-based oxidative stress requires more Fe(II) to compensate for the shortage of homologous recombination.

Kidney is essential for the higher species to excrete urea, a toxic metabolite of proteins [57]. Thus, when epithelial cells in the proximal tubule are damaged, they rapidly proliferate [58]. Renal tubular proliferation was significantly increased in the MUT at 3 weeks of carcinogenesis protocol, when ferroptosis was simultaneously enhanced. Recent study revealed that cell-cycle inhibited lipid droplet formation and increased vulnerability to ferroptosis [59]. However, we found no significant increase in lipid droplet formation in the MUT kidney at 3 weeks with immunohistochemical staining of PLIN3 (a lipid droplet marker) (Fig. S5C). Furthermore, we observed that nuclear translocation of p21 was inhibited via activation of phosphorylation of AKT protein. This association between BRCA2-deficiency and AKT activation may reflect the evidence that PI3K/AKT/PTEN alterations in patients with breast cancers showed lower expression of BRCA1/2 [60]. However, it remains unclear how BRCA2-deficiency overrides cell-cycle checkpoint. One possibility is that BRCA2 has an unknown role in cell-cycle checkpoint like BRCA1 [61]. Further studies are required to clarify the other BRCA2 functions.

In conclusion, our study demonstrated that iron-catalyzed oxidative stress does not promote carcinogenesis in the context of BRCA2 deficiency, where ferroptosis counteracts the oncogenic potential of genomic instability. Importantly, we show for the first time that BRCA2 deficiency under iron-catalyzed oxidative stress alters susceptibility to ferroptosis, leading to increased iron accumulation through modulation of iron metabolism, while concurrently promoting cell proliferation (Fig. 6). These findings provide novel insights into the molecular mechanisms underlying BRCA2 haploinsufficiency-associated carcinogenesis, which may inform the development of new therapeutic strategies for BRCA2-associated cancers.

Fig. 6.

Summary of the present study

Graphical abstract of how BRCA2-deficiency acts on Fe-NTA induced renal carcinogenesis.

Limitations of this study include: (1) the use of a preclinical rat model, and (2) the exclusive use of male animals, due to their higher susceptibility to Fe-NTA-induced renal carcinogenesis.

CRediT authorship contribution statement

Yuki Maeda: Writing – original draft, Methodology, Investigation, Data curation, Conceptualization. Yashiro Motooka: Writing – review & editing, Validation, Data curation. Shinya Akatsuka: Writing – review & editing, Project administration, Data curation. Hideaki Tanaka: Data curation. Tomoji Mashimo: Resources, Data curation. Shinya Toyokuni: Writing – review & editing, Writing – original draft, Supervision, Investigation, Funding acquisition, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.