Fig 2: Evolution of weight and motor function in injured xCT+/+ and xCT-/- at different timing post-injury. xCT-/- mice recovered pre-SCI weight and gained weight faster than xCT+/+ between 3 and 4 weeks post-injury (a). Use of forelimbs, either ipsilateral or contralateral (or both) to the spinal lesion, was not different between injured xCT+/+ and xCT-/- mice (b). Grip strength test showed a better recovery of the ipsilateral grip strength for xCT-/- mice than for xCT+/+ mice at 2- and 3- weeks post-injury (c). Both groups of mice recovered similarly at later time points. The contralateral grip strength was not impaired by the unilateral spinal cord injury (d). *p < 0.05 and ***p < 0.001, two-way ANOVA followed by Sidak’s multiple comparisons, n = 10–12 mice/group between -1 and 2 weeks post-injury, n = 4–8 mice/group between 3 and 6 weeks post-injury. Data are expressed as mean +/- SEM.

Fig 3: CMP (50 mg/kg, 1 time/day for 28 days) decreased in vivo tumor growth by suppressing NRF2-associated ferroptosis. (A) Representative tumor images. CMP significantly suppressed tumor (B) volume and (C) weight compared with those of the control. CMP significantly increased the accumulation of (D) Fe2+, (E) MDA and (F) SOD but reduced the levels of (G) GSH. (H) Western blot assay showed that Nrf2, HO-1, xCT and GPX4 protein levels were suppressed in nude mice treated with CMP compared with those of the control. **P<0.01, ***P<0.001 vs. con. CMP, carboxymethylated pachyman; Nrf2, nuclear factor erythroid 2-related factor 2; MDA, malondialdehyde; SOD, superoxide dismutase; GSH, glutathione; HO-1, heme oxygenase-1; xCT, cystine/glutamate antiporter system X(c)(-); Con, control.

Fig 4: Quantification of lesion size, motor neuron and oligodendrocyte number in injured xCT+/+ and xCT-/- spinal cords. The diagram depicts how the unilateral lesion extends rostro-caudally from - 1000 to + 1000 µm around the epicenter (a). The epicenter (set at 0 µm) was defined as the spinal level where the area of lesioned tissue is maximal. Representative micrographs of the lesion epicenter is shown for xCT+/+ (left) and xCT-/- (right) mice (b). Quantification of lesioned area (in percentage of ipsilateral hemicord) or lesion rostro-caudal extension did not show any difference between groups at 2 or 6 weeks post-injury (c). Quantification of lesion volume did not show any difference between groups at 2 or 6 weeks post-injury (d). As depicted, the number of spared motor neurons was counted in the inferior quadrants of both contralateral and ipsilateral sides, including the ventral horn and part of the intermediate column (e). For some distances along the lesion, ipsilateral motor neuron number was significantly higher in injured xCT-/- mice than in the xCT+/+ mice, at both 2 and 6 weeks post-injury (f). Data are expressed as mean +/- SEM. *p < 0.05, Mann–Whitney test, n = 3–4 mice/group. Using anti-p25a immunohistochemistry, the number of spared oligodendrocytes was evaluated in the injured xCT+/+ and xCT-/- hemicords, at three distances from the epicenter, 0 µm, + 500 µm and + 1000 µm (g,h). two-way ANOVA followed by Sidak multiple comparisons test, n = 3–4 mice/group. Data are expressed as mean +/- SEM. Scale bars = 50 µm in panel g.

Fig 5: Cellular slc7a11/xCT gene expression in the mouse spinal cord. A probe targeting slc7a11 mRNA was applied onto uninjured and injured spinal cords. The chromogenic signal of the in situ hybridization was revealed as pink deposits, whose size was proportional to the local amount of transcripts (a,b). In the normal spinal cord, slc7a11/xCT mRNA expression was predominantly detected in nervous cells satellite to motor neurons (a, inset 1), in the meningeal layer and along the wall of the perforating blood vessels (a, inset 2), and in the neuropil around the dorsal horn neurons (a, inset 3). A higher abundance of transcripts was detected in the lesion area from spinally-injured mice (b). qRT-PCR confirmed the overall increased expression of slc7a11/xCT mRNA in injured spinal cords, significantly peaking at 4 days post-injury (c) (*p < 0.05, Kruskal–Wallis ANOVA test, Un. n = 3; 4d n = 8; 2w n = 3; 6w n = 3). Immunoblot showing xCT and GAPDH protein expression at different timings post-SCI (d). Full length blot of xCT detection is displayed where xCT specific band was detected at approximately 40 kDa. GAPDH was used as loading control. Uncropped images are available in Supplementary Fig. 2. Compared to uninjured tissue (Un.), xCT was upregulated at the protein level at two weeks post-injury (e) (*p < 0.01, Kruskal–Wallis ANOVA test, Un. n = 4; 1w n = 4; 2w n = 5). Fluorescent slc7a11 in situ hybridization (orange) and immunofluorescent co-labeling (green) of GFAP, Iba1, p25a, MAP2 and vimentin in uninjured spinal cords (f). The majority of slc7a11 + cells were co-labeled with GFAP, vimentin, MAP2 and Iba1 (white arrows) (g). The nuclei were counterstained with DAPI (blue). slc7a11 expression was rarely detected in p25a + oligodendrocytes. Injury did not significantly modify the cellular expression profile of slc7a11 (p = n.s., Mann–Whitney test, n = 3 spinal slices in each group). Data are expressed as mean +/- SEM. Scale bars = 250 µm in a and 10 µm in f. Abbreviations : n.d. = not done, as MAP2 + motor neurons were completely lost at the injury epicenter.