Background
Cancer cells are characterized by altered metabolic programs in a challenging microenvironment, which can lead to excessive oxidative stress that can ultimately result in cancer cell death [
1‐
3]. Therefore, cancer cells upregulate antioxidant mechanisms to maintain redox homeostasis [
4].
Solute Carrier Family 7 member 11 (
SLC7A11) is a key gene involved in redox regulation in cancer cells and codes for the multipass transmembrane protein xCT. xCT dimerizes with the transmembrane chaperone protein SLC3A2 to form the amino acid transport system xc
− [
5]. While SLC3A2 allows localization of the transporter to the plasma membrane and acts as a subunit for several other transporter systems, xCT provides the substrate specificity [
5].
System xc
− acts as a Na
+-independent cystine/glutamate antiporter by importing cystine (but not cysteine) and exporting glutamate at a 1:1 ratio [
5]. Once inside the cell, cystine is reduced to cysteine, which is the rate-limiting precursor for the biosynthesis of the antioxidant molecule glutathione (GSH) [
5]. Therefore, xCT is responsible for scavenging of reactive oxygen species (ROS) by mediating the import of cystine into the cells, thereby preventing excessive oxidative stress and allowing cancer cell survival [
6]. Other transporters import cysteine; however, cystine is more abundant than cysteine because of the strongly oxidizing extracellular environment [
7]. Hence, cells rely on xCT to fulfill their cysteine needs by importing cystine.
Different tumor types, including breast cancer, overexpress xCT [
8,
9] to sustain proliferation and resistance to ROS-inducing chemotherapy and radiotherapy [
10,
11], and high xCT expression is associated with poor prognosis in various tumor types, including breast cancer [
9,
12]. In addition, xCT targeting has detrimental effects on cancer cells, resulting in reduced tumor growth and metastasis in mice, and sensitization to chemotherapy [
13]. Therefore, xCT represents a potential target for breast cancer treatment [
8]. Sulfasalazine (SAS), Erastin, and Sorafenib [
6] are pharmacological inhibitors of system xc
-, and we recently developed several anti-xCT vaccine formulations [
14].
However, an xCT-targeting strategy may affect not only cancer cells, but also cells of the immune system. Indeed, xCT is expressed by activated macrophages, granulocytes, and T cells [
5,
15], where it supports their physiological functions. xCT was also found to be expressed by immunosuppressive cell populations, such as Myeloid-Derived Suppressor Cells (MDSC) [
16] and regulatory T cells (Treg) [
17]. It was recently shown that a systemic lack of xCT preserves adaptive antitumor immunity [
18], but a full understanding of the effects of xCT modulation in the immune system is still lacking. In addition, many studies to date have focused on the cell-autonomous functions of xCT in immunocompromised mouse models, thus neglecting the role of xCT in the interaction of cancer cells with the surrounding microenvironment and in tumor-related immune responses.
The expression of xCT in cancer and immune cells could have non-cell-autonomous effects, in addition to its cell-autonomous role in redox balance maintenance. xCT-mediated secretion of glutamate by tumor cells promotes Treg activation and their immunosuppressive functions [
19], while MDSC compete with antigen-presenting cells for importing cystine through xCT, hindering T cell activation [
16]. Moreover, xCT is involved in the release of extracellular vesicles (EV) by tumor cells [
20‐
22], which play a role in communication with various components of the microenvironment [
23].
Finally, although the contribution of xCT to the malignant features of mammary cancer has been reported in the literature, no prior research has investigated its role in tumor initiation and progression in mammary-cancer-prone models.
Hence, the goal of this study was to address these currently neglected aspects of xCT contribution to the development of mammary cancer. On one hand, we investigated for the first time the role of xCT in mammary tumor initiation using a mouse model that lacks xCT and is predisposed to developing mammary cancer (BALB-neuT/xCTnull). On the other hand, we elucidated the differential effects of xCT on both the malignancy of tumor cells and the immune response to tumors by utilizing xCT-proficient and xCT-deficient mammary cancer cells and host mice, respectively. Moreover, we here provide evidence supporting the notion that xCT has non-cell-autonomous functions. These findings may broaden the scope of the currently tested therapeutic approaches against xCT.
Methods
Cell lines
4T1 (ATCC Cat# CRL-2539) cells were purchased from ATCC and cultured in RPMI-1640 medium (Sigma-Aldrich) supplemented with 10% Fetal Bovine Serum (FBS, Sigma-Aldrich). The growth medium of the 4T1 xCT
wt and xCT
KO clones was further supplemented with 100 µM β-Mercaptoethanol (β-ME; Sigma-Aldrich), while the growth medium of 4T1-pLVX and of xCT-overexpressing 4T1 cells (either derived from parental cells or xCT
KO pool) was supplemented with 3 µg/mL puromycin (Sigma-Aldrich). SUT32-2H9 and WT27 cells were cultured in DMEM-F12 medium (Sigma-Aldrich) containing 20% FBS and 100 µM β-ME. The growth medium of SUT32-2H9-pLVX and SUT32-2H9-xCT was further supplemented with 1.5 µg/mL puromycin. Where indicated, SUT32-2H9 cells were cultured in Human Plasma-Like Medium (HPLM, Gibco) supplemented with 20% dialyzed FBS. All growth media were supplemented with penicillin/streptomycin (P/S) solution (Sigma-Aldrich). Cells were maintained in humidified incubators at 37 °C with 5% CO
2. The cells were periodically tested for mycoplasma contamination using the MycoAlert™ Mycoplasma Detection Kit (Lonza Cat# LT07-318) according to the manufacturer’s instructions. All cells used were free of mycoplasma. All cells were used within 10 passages from thawing and were kept in culture cumulatively for no more than six months. All cell lines used originated from female mice. Further details on the establishment of SUT32-2H9 and WT27 cell lines from tumors of BALB-neuT/xCT
null and xCT
wt mice, as well as on selenocystine uptake assay, MTT assay, colony forming efficiency assay, ROS and lipid peroxidation detection, and migration assays can be found in the Additional file
2, Supplementary Methods section.
Generation of xCTKO 4T1 cells and of xCT-overexpressing cells
4T1 cells were transfected using Lipofectamine 2000 (Thermo Fisher Scientific Cat# 11668-019) according to the manufacturer’s instructions, with either an empty (px459) or
Slc7a11-targeting (px459-xCT) CRISPR/Cas9 vector (details on vector design and production are reported in the Additional file
2, Supplementary Methods section). The following day, the transfection medium was replaced with RPMI-1640 medium supplemented with 10% FBS and 100 µM β-ME. Puromycin was added at a final concentration of 3.5 µg/mL. After 48 h selection, cells were plated at 0.3 cells/well in 96-well plates, and single-cell clones were expanded further. The lack of xCT expression was confirmed by western blot analysis. To generate xCT-overexpressing cell lines, SUT32-2H9 and 4T1 target cells were stably transduced with a lentiviral vector expressing the coding sequence of murine xCT under the control of the CMV promoter. Further details are provided in the Additional file
2, Supplementary Methods section.
Western blot
Cells were incubated with RIPA lysis buffer supplemented with 1 mM phenylmethylsulfonyl fluoride, 1 mM NaVO4, 1 mM NaF, and a protease inhibitor cocktail (Sigma-Aldrich) on ice. Proteins (20–30 µg) were added to Laemmli Buffer (Bio-Rad Laboratories) supplemented with 2.5% β-ME and left at room temperature for 30 min. Proteins were separated using a 4–15% polyacrylamide Precast Gel (Bio-Rad Laboratories) and transferred to a PVDF membrane (Immobilion-P, 0.45 μm pore size, Merck Millipore). In some instances, stain-free gels (Bio-Rad Laboratoires) were used. Non-specific binding sites were blocked using 5% non-fat milk (Santa Cruz Biotechnology) in T-TBS 1X (Tris Buffered Saline, 0.1% Tween 20). Membranes were incubated with mouse anti-vinculin (produced in-house, 1:8000) or rabbit anti-mouse xCT (#98051, Cell Signaling Technology; 1:1000), followed by incubation with HRP-linked goat anti-mouse (Sigma-Aldrich, Cat# A0545; 1:2000) or HRP-linked goat anti-rabbit IgG (Sigma-Aldrich, Cat# A4416; 1:2000), respectively. ECL substrate (Cyanagen Cat# XLS142,0250) was used for signal detection, and images were acquired using a Chemidoc™ Touch Imaging System (Bio-Rad Laboratories).
Mouse models
BALB/c and BALB-neuT mice [
24] were obtained from an internal breeding. C3H/HeSnJ-
Slc7a11sut/J mice [
25] were purchased from The Jackson Laboratory and crossed with BALB/c and BALB-neuT mice to obtain BALB/c-xCT
null and BALB-neuT/xCT
null mice, respectively. The details of the crossing and genotyping are reported in the Additional file
2, Supplementary Methods section. Animals were group housed, and food and water were provided
ad libitum. All mice were maintained at the Molecular Biotechnology Center, University of Turin.
Tumor monitoring
10-12-week-old BALB/c-xCTwt or BALB/c-xCTnull females were challenged subcutaneously (s.c.) with cancer cells in the left flank region, corresponding to the 4th mammary gland. Number of cells injected is reported in the corresponding figure legend. Tumor growth was monitored twice a week using a caliper, and the experimental endpoint was set at 28–33 days after injection (or at days 8 and 15 after injection, where specified). In BALB-neuT and BALB-neuT/xCTnull mice, tumor onset and growth were monitored weekly, and the endpoint was set according to ethical criteria. For intravenous (i.v.) cancer cell injection, 10.000 cells (parental 4T1 and pool of 4T1 xCTKO clones) were suspended in 100 µL PBS and injected into the caudal vein of female BALB/c mice, which were culled 28 days post-injection.
Immunophenotyping
At the experimental endpoint, the mice were anesthetized, and peripheral blood was collected through intracardiac puncture and supplemented with heparin solution. The mice were euthanized, and the lungs, tumor, and spleen were collected. The left lung and the tumor were finely minced and enzymatically digested with 100 µg/mL collagenase in DMEM at 37 °C for 30 min and 45 min, respectively, under shaking. The digested tissues were passed through a 70 μm pore cell strainer to obtain a single-cell suspension. Erythrocytes were lysed using an erythrocyte lysis buffer. An Fc Blocking antibody (anti-CD16/CD32 antibody, BioLegend Cat# 101,320) was added, and the cells were stained at 4 °C for 30 min with fluorochrome-labeled antibodies. Samples were acquired using a BD FACSVerse™ instrument, and data were analyzed using FlowJo V10 software. Details of the staining procedure and antibodies used, as well as the phenotyping protocol of SUT32-2H9 and WT27 cell lines, can be found in the Additional file
2, Supplementary Methods section.
In vitro polarization of bone marrow-derived cells
Mice were euthanized and femurs and tibias were collected and stored in ice-cold RPMI-1640 supplemented with 10% FBS and 1% P/S. Bone marrow (BM) was extracted as previously described [
26]. BM cells were incubated in erythrocyte lysis buffer, rinsed with PBS, pelleted, and seeded in a 6 well-plate in RPMI-1640 with 10% FBS or supernatant from 72 h-cultured 4T1 cells. The cells were incubated at 37 °C and replenished with fresh or conditioned medium (CM) every second day. On day 7, the suspension and adherent cells were pooled and stained with fluorochrome-labeled antibody combinations for macrophages/myeloid cells, DC, and B cells, as described in the extended immunophenotyping section (Additional file
2, Supplementary Methods) and were acquired and analyzed as described above.
DNA-based vaccination and SAS treatment
Mice were anesthetized and vaccinated with two intramuscular injections of 20 µL of physiological solution containing 50 µg of pVAX1 (Thermo Fisher Scientific Cat# V26020) or RHuT [
27] plasmids followed by low-voltage electroporation as previously described [
27], at a 12–14 day interval. When combined with SAS treatment, the first vaccination was performed 4 days after the beginning of the treatment, while the second vaccination preceded of 4 days the end of the treatment (Supplementary Fig. S
2A). Each mouse received an intraperitoneal injection of 4 mg of SAS (Sigma-Aldrich, Cat# S0883) in 400 µL of saline solution, twice daily, for 22 consecutive days (Supplementary Fig. S
2A). This resulted in a total daily dose of 8 mg of SAS per mouse, approximately equivalent to 400 mg/kg of body weight. The SAS solution was freshly prepared prior to administration. Initially, SAS was suspended in a small volume of 1 N NaOH (200 mg/mL), then diluted with saline solution and adjusted to a pH of around 8 using 1 N HCl. Subsequently, the SAS solution was further adjusted to a final concentration of 10 mg/mL using saline solution.
In vivo cytotoxicity assay
Spleens from donor mice were disaggregated using a syringe plunger over a 70 μm cell strainer, incubated in erythrocyte lysis buffer, rinsed in PBS, and centrifuged. Equal amounts of cells were incubated at a final concentration of either 5 (CFSEhigh) or 0.5 (CFSElow) µM carboxyfluorescein succinimidyl ester (CFSE; Thermo Fisher Scientific Cat# V12883). CFSEhigh splenocytes were then incubated with a final concentration of 15 µg/mL immunodominant rat ERBB2 (p185neu [63–71] 9-mer) peptide with H-2Kd restriction. Equal amounts of CFSEhigh and CFSElow splenocytes were mixed at a 1:1 ratio and injected intravenously into pVAX1- or RHuT-vaccinated mice. After 48 h, the recipient mice were euthanized, spleens were collected and disaggregated as described above, and the proportion of CFSEhigh and CFSElow splenocytes was assessed using FACS.
Rat and human ERBB2 ELISA
96-well plates (Costar) were coated with 100 ng/well of recombinant extracellular portions of rat (Sino Biological Cat# 80,079-R08H, His Tag) or human (Sino Biological Cat# 10,004-H08H, His Tag) ERBB2 proteins. ELISA of sera from vaccinated mice to detect anti-ERBB2 IgG was then carried out as previously described [
28].
Statistical analysis
Unless otherwise specified, an unpaired t test was used to assess statistically significant differences between the groups. Ratio paired t test was used to assess consistency in the ratios of paired values (where values of experimental groups were normalized on values of control groups), Fisher’s test was used to assess significant differences in the incidence of lung metastases, and Log-Rank (Mantel-Cox) test was used to assess differences in the disease-free survival of BALB-neuT and BALB-neuT/xCTnull mice. Statistical analysis was performed using the GraphPad Prism v8 software. p < 0.05 was considered significant. Definition of center and of dispersion and precision measures (e.g., mean and SD), as well as the number of technical or biological replicates of the experiments described and the specific statistical test used, are reported in the corresponding figure legends. For in vivo experiments, mice were assigned to a given experimental group via simple randomization. Experimenters were not blind to group assignment and outcome assessment.
Discussion
Although previous reports have indicated that xCT depletion in tumor cells leads to impaired tumor growth in vivo [
18], our data obtained in mammary cancer-prone BALB-neuT mice show that congenital, total body xCT deficiency does not affect the time of onset or the multiplicity of tumors, but reduces the incidence of lung metastases. To date, only two studies have evaluated the consequences of xCT deficiency in tumor progression in cancer-prone mice, both using the KPC pancreatic tumor model. In the first study [
37], tamoxifen-induced whole-body
Slc7a11 KO after tumor appearance led to increased survival. In the second study [
38], congenital epithelial-specific
Slc7a11 KO did not alter the incidence, onset, and progression of pancreatic cancer. These differences suggest a pivotal role for xCT in the tumor stroma during tumor progression, or, alternatively, that the timing of xCT deletion may be crucial for the tumor outcome. However, the striking impairment in proliferation, increased ROS content, and lipid peroxidation observed in vitro under basal conditions in tumor cells derived from BALB-neuT/xCT
null mice, and the finding that the addition of β-ME or re-expression of
Slc7a11 is sufficient for the reversion of these phenotypes reveals indeed essential functions of xCT in the biology of these epithelial tumor cells under specific conditions. Thus, it seems likely that xCT deficiency obtained prior to the onset of neoplastic transformation promotes the occurrence of compensatory mechanisms that allow cells to bypass the defective cystine/glutamate exchange in vivo. Although the nature of such compensatory mechanisms remains to be identified, the different environmental conditions of cultured cells, such as supraphysiological oxygen [
39], likely represent a challenging environment that may increase the oxidative stress that cells must face compared to an in vivo environment. Moreover, plasma and culture media contain different concentrations of cystine, cysteine, and GSH. Whereas cysteine is absent in the culture medium, it is present in the plasma [
40], and can be imported by cells via alternative transporters (e.g., ASCT1, ASCT2 [
41], LAT1 [
5]), thus circumventing the requirement of system xc
−. Hence, it is possible that cysteine in plasma could provide a sufficient support in vivo to immune system functionality on one hand, and to tumor initiation and growth on the other hand.
In sharp contrast to the results obtained with tumor cells derived from xCT
null mice, we did not observe significant impairments in xCT
KO 4T1 proliferation in vitro under standard confluence culture conditions. Increased oxidative stress in xCT
KO cells, but not in xCT
wt 4T1 cells, was observed only upon the administration of tBHP, and this was rescued by the addition of β-ME. This indicates that xCT
KO 4T1 cells have indeed a latent imbalance in redox homeostasis, which becomes overt only under stress conditions. 4T1 cells are thus endowed with compensatory mechanisms that allow them to survive and proliferate under normal culture conditions, but that are not sufficient to overcome further increases in oxidative stress. These processes may include the
de novo synthesis of cysteine via the transsulfuration pathway [
42] or the expression of other cystine transporters. These may include the heterodimers SLC3A1/SLC7A9 and SLC3A1/SLC7A13, and the excitatory amino acid transporters (EEAT) [
43]. However, the relevance of such processes in tumor biology is currently unclear, and the specific mechanisms that allow 4T1 cells to bypass xCT deficiency under basal growth conditions in vitro still need to be identified.
The 4T1 model is highly metastasis-prone and thus represents a better experimental context to assess the contribution of xCT to the metastatic process than the poorly metastatic BALB-neuT model [
29], which nevertheless displays a trend of reduced incidence of metastasis within the ethical endpoint when xCT is absent. Importantly, xCT depletion in 4T1 cells resulted in reduced migration in vitro and in a strong impairment of cell metastatic ability in vivo, which is consistent with previous reports [
44,
45]. Notably, xCT
KO 4T1 cells showed a dramatic impairment in their clonogenic ability in vitro, which may play a key role in the conversion of seeded cancer cells into overt metastases in vivo. As shown recently, ferroptosis limits the survival of cancer cells in the bloodstream owing to the high free iron concentration [
46], thus providing a rational explanation for the drastic impairment of the metastatic ability of xCT
KO cells, in spite of the lack of significant differences between primary tumors generated by them and those of xCT
wt cell counterparts. Moreover, xCT-dependent glutamate secretion by breast cancer cells was found to induce autocrine/paracrine activation of metabotropic glutamate receptor (GRM)3/Rab27a-dependent membrane trafficking [
20]. This leads to the relocation and secretion of invasion-promoting proteins that allow breast cancer cells to metastasize [
20]. Thus, the overall impairment of the metastatic ability of xCT
KO 4T1 cells may result from both an increased propensity to undergo ferroptosis in the bloodstream owing to a reduced ROS-buffering capacity, and an impaired cell migration secondary to a blunted glutamate secretion.
In addition to this cell-autonomous mechanism, Rab27a activity in 4T1 cells mediates secretion of EV and cytokines that recruit tumor-promoting neutrophils [
47]. Interestingly, xCT depletion in 4T1 tumor cells resulted in a significant increase in CD4
+ T, CD8
+ T, B, and NK cells, which are important players in the antitumor immune response [
48], and in fewer PMN-MDSC, in the lungs. This was not obvious in xCT
wt and xCT
null BALB-neuT mice [
29], where the presence of PMN-MDSC in the lungs was much lower than that induced in 4T1 tumor-bearing mice, in line with BALB-neuT reduced propensity to develop lung metastases. PMN-MDSC are reportedly involved in the promotion of metastasis by participating in the formation of pre-metastatic niches [
49]; their expansion is induced by different stimuli including G-CSF and GM-CSF [
50]. We indeed observed an increase in the amount of G-CSF in the plasma of tumor-bearing compared to that of healthy mice, but not in mice challenged with xCT
wt 4T1 cells versus xCT
KO cells. Besides cytokines, other in vivo mechanisms may thus be responsible for differential MDSC expansion, such as EV released in an xCT-dependent manner [
21]. EV mediate communication between primary tumors and distant organs, and it has already been reported that breast cancer-derived EV stimulate MDSC expansion [
51], and that xCT is involved in EV release from transformed cells [
20‐
22]. Further experiments are warranted to elucidate the possible mechanisms underlying the observed effects on the metastatic ability.
Given that high xCT expression in tumors is associated with a poor prognosis in oncological patients [
9,
12], its targeting has been investigated by us and others [
6,
14,
52] as a therapeutic strategy. xCT is also expressed in several types of activated immune cells [
5,
15,
17]. Despite this, our results demonstrate that the proportions of different immune cell populations were not altered in healthy xCT
null mice. In addition, cellular and humoral immune responses against non-self-antigens were preserved in the vaccinated xCT
null mice and in mice treated with the xCT inhibitor SAS. Accordingly, systemic xCT deficiency does not alter the composition of the immune infiltrate in the primary tumor in vivo; thus, xCT is dispensable for immune system function. This is in line with a previous report by Arensman et al. [
18], although they focused exclusively on the T-cell response. These data indicate that xCT targeting would be effective in tumor cells, while sparing cells of the immune system. Although pharmacological or immune-mediated inhibition of a protein does not necessarily recapitulate the disruption of its coding gene, our results corroborate our previous observations that anti-xCT vaccination efficiently impairs tumor metastasization while preserving antitumor immunity [
13,
14].
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.