Fig 2: LSECs detoxify hemoglobin upon hemolysis.(A–D) Mice were injected with native hemoglobin (Hb, 10 mg/mouse) and analyzed at the indicated time points. (A) Heme iron levels in the liver were determined with the Heme Assy Kit and normalized to protein content. (B) Non-heme iron content in the liver was measured by bathophenanthroline colorimetric assay. (C, D) Hmox1 gene expression measurement by RT-PCR in (C) the liver and (D) FACS-sorted KC and LSEC populations. (E, F) Mice were injected i.v. with high-dose Hb (10 mg). (E) Frozen mouse liver slices were processed and stained for HO-1 (white), the LSEC marker CD146 (orange), and F4/80 identifying KCs (green) and imaged using confocal microscopy. Scale bars, 20 μm. (F) Total iron content in magnetically-sorted cells was determined 20 min after injection by ICP-OES and normalized to protein content. (G, H) Hemolysis was induced in mice by i.p. injection of phenylhydrazine (PHZ, 0.125 mg/g) 6 h before analysis. Hmox1 gene expression in (G) FACS-sorted cell populations of KCs and LSECs, and (H) spleen and aorta. (I) Non-heme iron content in the liver. Data are expressed as mean ± SEM, and each data point represents one biological replicate (individual mouse), n = 4–5 (A–C), 5–6 (D), 3–4 (F), 5 (G), 4 (H, I). One-way ANOVA with Tukey’s Multiple Comparison tests was used in (A–C), and separately for KCs and LSECs in (D); two-way ANOVA with Tukey’s Multiple Comparison tests was used in (F), and Welch’s unpaired t test was used to determine statistical significance in (G–I), separately for each cell type/organ; exact P values are shown on graphs, a value above the bar in (C) indicates comparison with control mice. Source data are available online for this figure.

Fig 3: LSECs constitutively express proteins critical for iron recycling and macropinocytosis.(A) Frozen mouse liver slices were processed and stained for FPN or HO-1 (white) in KCs (F4/80, green) or LSECs (CD146, orange) and imaged using confocal microscopy. Scale bars, 5 μm (upper panel) and 10 μm (bottom panel) (B) FPN and HO-1 presence in F4/80+ (white) KCs and CD31+ endothelial cells (green) imaged by confocal microscopy in murine NPC in vitro cultures. (A, B) Nuclei (blue), scale bars, 20 μm. Arrows indicate ECs positive for FPN or HO-1. (C) Flow cytometry analysis of FPN expression in LSECs and KCs. (D) Flow cytometry analysis of HO-1 expression in LSECs and KCs. (E) Shown are selected proteins that are statistically significantly (P < 0.01) more abundant in mouse FACS-sorted LSECs than in spleen and heart endothelial cells (ECs). Log2 fold change (from three independent cell isolates) in protein abundance in LSECs versus the indicated cell type is presented. Data are retrieved from Dataset EV1, where individual label-free quantification values of protein abundance and calculated P values are included. (F) Total iron content in FACS-sorted cells was determined by ICP-OES and normalized to cell numbers. (G) Cytosolic ferrous iron (Fe2+) levels in the indicated endothelial cells were measured with a FerroOrange probe and flow cytometry. (H) Heme levels were quantified in magnetically-sorted cells using the Heme Assay Kit and normalized to the protein content. (I) Flow cytometry analysis of FPN expression in LSEC and KCs isolated from mice maintained on balanced or iron-deficient diets. (J) Cytosolic ferrous iron (Fe2+) content was measured with a FerroOrange probe and flow cytometry in LSECs and KCs derived from mice injected with mini-hepcidin (PR73). Numerical data are expressed as mean ± SEM, and each data point represents one biological replicate (individual mouse), n = 5 (C, D, G, H), 5–6 (F), 6 (I), 8–9 (J). Welch’s unpaired t test was used to determine statistical significance in (C, D, I, J); one-way ANOVA with Tukey’s Multiple Comparison tests was used in (F, G, H); exact P values are shown on graphs. Source data are available online for this figure.

Fig 4: LSECs represent the major cell type that sequesters Hb.(A, B) Hemoglobin (Hb) distribution in control and macrophage-depleted mice (clodronate) injected with AlexaFluor 750 labeled Hb (Hb-AF750, 10 µg/mouse), imaged with Bruker in vivo Imaging System. (A) The efficiency of macrophage depletion in the liver was examined by the percentage of liver KCs in control and clodronate-injected mice. (B) Representative images of organs isolated from Hb-AF750 (10 µg/mouse, 1 h) i.v.-injected mice. (C) Frozen liver slices from mice injected with Hb-AF647 (red) were processed and stained for Kupffer cells (KCs) (F4/80, green) or LSECs (CD146, green) and nuclei (blue). Merged-channel images for this figure are presented in Fig. 1C. (D) Murine NPCs in vitro cultures were treated with Hb-AF750 (0.5 µg/ml) for 1 h. Normalized Hb-AF750 fluorescence intensity and percentage of Hb-AF750+ LSECs and KCs, measured with flow cytometry. (E) Plots illustrating high mRNA expression of human LCES markers CD32B and CD36 in liver ECs, visualized using the Human Liver Cell Atlas (Guilliams et al, 2022). Data are expressed as mean ± SEM, and each data point represents one biological replicate, n = 10 (A), 4 (D). Welch’s unpaired t test was used to determine statistical significance in (A, D); exact P values are shown on graphs.

Fig 5: LSECs represent the major cell type that sequesters Hb.(A) Hemoglobin (Hb) distribution in control and macrophage-depleted mice (clodronate) injected with AlexaFluor 750-labeled Hb (Hb-AF750, 10 µg/mouse), imaged with Bruker in vivo Imaging System. Results are presented as total fluorescent counts from the isolated organs. (B) AF750 fluorescence from the indicated liver cell populations isolated from Hb-AF750-injected mice was analyzed with flow cytometry. (C) Frozen liver slices from mice injected with Hb-AF647 (red) were processed and stained for Kupffer cells (KCs) (F4/80, green) or LSECs (CD146, green) and nuclei (blue). Single-channel images for this figure are presented in Fig. EV1C. (D) Murine liver primary non-parenchymal cells (NPCs) were treated in vitro with Hb-AF750 (red) (0.5 µg/ml) and stained for F4/80 (blue) and the LSEC marker STAB2 (green). (C, D) Areas in the highlighted rectangles are shown at 5× higher magnification. (E) Flow cytometric analysis of human freshly isolated primary NPCs, exposed in culture to human Hb-AF647 and stained with the human LSEC markers CD36 and CD32B. The intensity of Hb-A647 uptake was measured in two LSEC subpopulations, CD36hiCD32Bhi and CD36midCD32Bmid, as compared to other NPCs CD36-CD32B-. Numbers indicate MFI. Data are representative of three independent donors of human NPCs. (F) Hb-AF647 (green) uptake was imaged in freshly isolated human liver NPCs stained for the LSEC markers CD36 and CD32B (orange). (C, D, F) Arrows indicate Hb presence in LSECs; Scale bars, 20 μm. Numerical data are expressed as mean ± SEM; each data point represents one biological replicate (individual mouse or independent cell-based experiment), n = 4–5 (A), 6 (B). Two-way ANOVA with Tukey’s Multiple Comparison tests was used in (A), and Welch’s unpaired t test was used to determine statistical significance in (B); exact P values are shown on graphs. Source data are available online for this figure.