Hypomorphic biliverdin reductase a mutations define bilirubin anti-malarial threshold

Summary

Jaundice, a condition characterized by elevated circulating bilirubin generated by biliverdin reductase A (BVRA), is protective against malaria. Beyond its oxidoreductase activity, BVRA acts as a protein kinase and transcriptional regulator. To probe the protective effect of BVRA oxidoreductase activity, we generated mice harboring G17A and E97A missense mutations in its NAD(P)H-binding and reductase motifs, respectively. BVRA oxidoreductase activity was reduced by ∼95% and ∼80%–90% in Blvra^G17A and Blvra^E97A vs. wild-type (Blvra^WT) mice, respectively. Plasmodium chabaudi chabaudi AS (Pcc) infection was lethal in Blvra^G17A and Blvra^E97A, compared to the non-lethal outcome in control Blvra^WT mice. Quantification of circulating unconjugated bilirubin in Pcc-infected mice revealed dose response effect whereby the mutant strains failed to reach a protective ∼20–30 μM (∼1–2 mg/dL) threshold. These findings establish the antimalarial effect of BVRA oxidoreductase activity and define a threshold of circulating bilirubin required for malaria protection, informing on therapeutic development and biomarker-guided strategies.

Graphical abstract

Highlights

• •Establishment of catalytic deficient BVRA mouse mutants
• •Antimalarial effect of BVRA relies on its catalytic activity
• •A minimal bilirubin threshold for parasite control
• •A minimal bilirubin threshold for malaria resolution

Biological sciences

Introduction

Plasmodium falciparum malaria remains a major global burden to humanity.¹ In the absence of vaccines providing sterile lifelong protection,² anti-malarial drugs remain our first line of defense.³ This, however, is poised by the selection of drug-resistant virulent parasite strains,¹ requiring the development of orthogonal approaches targeting vital pathways deployed by Plasmodium throughout different stages of infection.⁴ These approaches can be inferred from rodent models of malaria, when targeting evolutionarily conserved vital pathways used by the parasite to infect their hosts. Vertebrate, including mice and humans, develop malaria upon the inoculation of Plasmodium sporozoites in the dermis, through the bite of infected Anopheles mosquitoes. Sporozoites reach the liver, via the bloodstream, developing and rupturing infected hepatocytes to release merozoites, which invade and multiply exponentially in red blood cells (RBC), during the symptomatic blood stage of infection.⁵

Plasmodium blood stage is characterized by circadian cycles of intravascular hemolysis with release of Hb tetramers (α₂β₂) into plasma. Extracellular Hb dissociates into dimers (αβ) that undergo (auto) oxidation and release their non-covalently bound prosthetic heme groups.^6,7 This pathologic process produces cytotoxic^8,9,10,11,12 labile heme,^6,7,13,14 an independent risk factor for severe P. falciparum malaria¹³ that promotes disease severity and mortality in mice.^{14,15,16,17,18}

The infected host limits the pathogenic effects of labile heme via its catabolism by heme oxygenase-1.^{14,15,19,20,21} This stress-responsive enzyme cleaves heme’s tetrapyrrole ring to produce biliverdin, releasing equimolar amounts of carbon monoxide and iron.²² Carbon monoxide binds avidly to iron in the heme groups of extracellular Hb, inhibiting further heme release from Hb and conferring protection against experimental cerebral malaria in mice.^14,15 Catalytic iron induces ferritin and ferroportin 1 (SLC40A1), which restrain iron redox activity and promote iron cellular export, preventing severe non-cerebral malaria^12,20,23 including life-threatening malarial anemia,²³ respectively.

We found that the reduction of biliverdin to bilirubin, catalyzed by BVRA,^24,25,26 is protective against malaria.²⁷ Once considered as a lipophilic waste product, bilirubin is now recognized as a cytoprotective^28,29,30,31 radical-trapping antioxidant,^28,30,32 while also regulating immune responses³³ and different aspects of energy metabolism.^34,35,36

BVRA exerts “non-canonical” functions,^{25,37,38,39,40,41,42} attributed to its dual-specificity kinase (Ser/Thr and Tyr) and a DNA/chromatin-binding motifs, at the N-terminal (aa 1–107) and C-terminal (aa 107–296) domains, respectively.^25,39,43 The dual kinase motif of BVRA phosphorylates the insulin receptor (INSR) substrate-1 (IRS-1), thereby regulating insulin signaling and downstream components of this pathway.^34,44 The DNA/chromatin-binding activity regulates the activating transcription factor 2 (ATF-2).^38,45 Moreover, BVRA also regulates the transcription factor nuclear factor erythroid-derived factor-like 2 (NRF2), independently of its catalytic activity.⁴⁶

Circulating bilirubin is lipophilic and is conjugated to glucuronic acid, via a reaction catalyzed in the liver by UDP glucuronosyltransferase family 1 member A1 (UGT1A1),⁴⁷ preventing its cytotoxicity.⁴⁸ When bilirubin conjugation falls below its production rate, unconjugated hyperbilirubinemia develops, eventually leading to jaundice.

Recognized clinically by yellowish discoloration of the skin and sclerae,⁴⁹ jaundice is used clinically as a biomarker of liver dysfunction. While accurate, this led to the perception of jaundice representing a maladaptive response, a notion challenged by the recent finding that unconjugated bilirubin confers resistance to malaria.²⁷

The antimalarial effect of jaundice is attributed to the cytotoxic effect of unconjugated bilirubin on Plasmodium spp. parasites, the causative agents of malaria, suggesting that BVRA enzymatic activity is protective against malaria.²⁷ Here, we tested this hypothesis in mice carrying genetic disruptions in the NAD(P)H-binding or reductase domains. We found that both mutant mouse strains succumbed to malaria, suggesting that the antimalarial effect of BVRA relies strictly on its oxidoreductase activity. Moreover, a comparison between mutant and wild type mouse strains revealed the minimum level of circulating bilirubin required for its antimalarial effect.

Results

CRISPR-Cas9-mediated generation of BVRA catalytic mutant mice

To determine whether the protective effect of BVRA against malaria relies strictly on its oxidoreductase activity, we generated two C57BL/6J mouse strains carrying a G17A mutation targeting the NAD(P)H-binding motif (Figures 1A and 1B) or an E97A mutation targeting the oxidoreductase biliverdin-binding motif of BVRA (Figures 1F and 1G). By targeting these domains, we set out to distinguish between cofactor-[NAD(P)H] and substrate (biliverdin IXα)-dependent mechanisms underlying BVRA’s antimalarial activity.

Figure 1.

Generation and validation of Blvra G17A and Blvra E97A mutant mice

(A) Crystal structure of rat biliverdin reductase A (BVRA, PDB: 1LC3) showing the G17A mutation site within the NADPH-binding motif. The protein backbone is shown as a marine blue cartoon. The NADPH-binding motif (residues 13–19, green sticks) contains the conserved Rossmann fold (GXGXXG) essential for cofactor binding. The G17A mutation site is highlighted in red sticks, a critical position within this cofactor-binding motif.

(B) Schematic representation of CRISPR-Cas9 targeting strategy showing the guide RNA (gRNA) target sequence and protospacer adjacent motif (PAM) for wild-type (WT) and mutant sequences to introduce the G17A mutation in the NAD(P)H-binding motif.

(C) Schematic of genotyping PCR (gPCR) and Sanger sequencing analysis.

(D) gPCR analysis showing gel electrophoresis of amplification products from Blvra^G17A mutant, Blvra^WT control samples and no template control as NTC, together with molecular weight (MW) markers indicated in base pairs (bp).

(E) Sanger sequencing chromatogram confirming the precise insertion of the G17A mutation in Blvra^G17A mice compared to Blvra^WT controls.

(F) BVRA crystal structure of rat biliverdin reductase A (BVRA, PDB: 1LC3) showing the E97A mutation site within the biliverdin-binding (reductase) domain (residues 90–105, green sticks). The protein backbone is shown as a marine blue cartoon. The E97A mutation is highlighted in orange sticks, indicating its position within the substrate-binding region.

(G) Schematic representation of CRISPR-Cas9 targeting strategy showing the gRNA target sequence and PAM for WT and mutant sequences to introduce the E97A mutation in the reductase domain.

(H) Schematic of genotyping gPCR and Sanger sequencing analysis.

(I) gPCR analysis showing gel electrophoresis of amplification products from Blvra^E97A mutant, Blvra^WT control samples and no template control as NTC, together with molecular weight (MW) markers indicated in base pairs (bp).

(J) Sanger sequencing chromatogram confirming the precise insertion of the E97A mutation in Blvra^E97A mice compared to Blvra^WT controls.

The Blvra^G17A strain carries a GGC→GCA missense mutation replacing glycine (G) at position 17 by an alanine (A) (G17A) in the NADPH-binding domain of BVRA (Figures 1A and 1B). This is expected to disrupt the flexibility of the BVRA NADPH-binding loop,⁴⁶ potentially affecting cofactor binding affinity and therefore catalytic efficiency (Figures 1A and 1B). The GCA missense mutation was confirmed by genomic PCR (Figures 1C and 1D) and validated by Sanger sequencing (Figures 1C and 1E).

The Blvra^E97A strain carries a GAA→GCC missense mutation replacing glutamate (E) at position 97 by an alanine (A) (E97A) in the biliverdin-binding (reductase) domain of BVRA (Figures 1F and 1G). This is expected to disrupt BVRA reductase activity, impairing biliverdin IXα reduction into bilirubin IXα (Figures 1F and 1G). The GCC missense mutation was confirmed by genomic PCR and Sanger sequencing (Figures 1H–1J).

BVRA expression in Blvra E97A and Blvra G17A mouse strains

The relative level of Blvra mRNA expression in Blvra^G17A mice was comparable to wild-type (Blvra^WT) controls, as assessed by RT-qPCR from whole spleen, kidney, and liver (Figure 2A). Blvra^−/− mice showed neither Blvra mRNA nor protein expression (Figures 2A and 2B). BVRA protein expression was lower in Blvra^G17A vs. Blvra^WT mice, as assessed by western blot (Figure 2B).

Figure 2.

Expression analysis of BVRA missense mutants G17A and E97A in C57BL/6J mice

(A) RT-qPCR analysis of Blvra mRNA expression in spleen, kidney, and liver from Blvra^WT, Blvra^−/− and Blvra^G17A mice. Data are expressed as Blvra/Rplp0 (2^−ΔCt) and presented as individual data points with bar graphs showing mean ± SD. n = 3 mice per genotype.

(B) Western blot analysis of BVRA protein expression in spleen, kidney, and liver. Representative blots show BVRA (33 kDa) and β-actin (42 kDa) loading control. Quantification of BVRA protein levels normalized to β-actin is shown below each blot as individual data points with bar graphs showing mean ± SD. n = 3–5 per genotype.

(C) RT-qPCR analysis of Blvra mRNA expression in spleen, kidney and liver tissues from Blvra^WT, Blvra^−/−, and Blvra^E97A mice. Data are expressed as Blvra/Rplp0 (2^−ΔCt) and presented as individual data points with bar graphs showing mean ± SD. n = 3 per genotype.

(D) Western blot analysis of BVRA protein expression in spleen, kidney, and liver tissue. Representative blots and quantification of BVRA protein levels normalized to β-actin below each blot, shown as individual data points with bar graphs showing mean ± SD. n = 3–6 mice per genotype. Statistical comparisons were performed using one-way ANOVA with Tukey’s post hoc test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; NS, not significant.

The relative level of Blvra mRNA expression in Blvra^E97A mice was slightly elevated in the kidney and liver but not the spleen, where there was a slight decrease, compared to Blvra^WT controls, as quantified by RT-qPCR (Figure 2C). BVRA protein expression was only modestly reduced in the spleen of Blvra^E97A vs. Blvra^WT controls, as assessed by western blot (Figure 2D). This was not observed in the kidney and liver extracts, which may reflect tissue-specific differences in protein levels (Figure 2D).

Bilirubin production is diminished in Blvra E97A and Blvra G17A mice

To assess the functional consequences of the E97A and G17A on BVRA enzymatic activity, we performed ex vivo activity assays in liver, spleen, and kidney homogenates from Blvra^E97A, and Blvra^G17A vs. Blvra^WT, and Blvra^−/− mice. Splenic BVRA enzymatic activity was barely detectable in Blvra^G17A compared to ∼0.2 nM bilirubin/μg protein/minute in Blvra^WT mice (Figure 3A). Area under the curve (AUC) analysis confirmed a ∼95%–97% reduction, comparable to a 97%–98% reduction in Blvra^−/− mice, with no significant difference between mutant genotypes (Figure 3B), consistent with loss of catalytic activity by purified BVRA G17A mutant protein.^46,50

Figure 3.

BVRA enzymatic activity is repressed in Blvra G17A and Blvra E97A mutant mice

(A and B) Ex-vivo BVRA enzymatic activity in tissue homogenates from Blvra^G17A mice. (A) Time course of bilirubin production measured as bilirubin concentration (nM/μg protein) over 50 min in spleen (left), kidney (middle), and liver (right) from Blvra^WT, Blvra^−/−, and Blvra^G17A mice. (B) Area under the curve (AUC) analysis spleen (left), kidney (middle), and liver (right) comparing enzymatic activity between genotypes.

(C and D) Ex-vivo BVRA enzymatic activity in tissue homogenates from Blvra^E97A mice. (C) Time course of bilirubin production measured as BR concentration (nM/μg protein) over 50 min in spleen (left), kidney (middle), and liver (right) from Blvra^WT, Blvra^−/−, and Blvra^G17A mice. (D) Area under the curve (AUC) analysis for spleen (left), kidney (middle), and liver (right) comparing enzymatic activity between genotypes. Data represent mean ± SD. n = 3 mice per genotype. Statistical comparisons were performed using one-way ANOVA with post hoc test. ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; NS, not significant. (A and B) Ex vivo BVRA activity from Blvra^G17A mice. (C and D) Ex vivo BVRA activity from Blvra^E97A mice. Mean ± SD. n = 3 per genotype. One-way ANOVA. ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; NS, not significant.

Renal BVRA enzymatic activity, estimated to be ∼0.24 nM bilirubin/μg protein/minute in Blvra^WT mice was reduced by ∼95%–97% in Blvra^G17A mice and greater than 97% in Blvra^−/− mice (Figures 3A and 3B). Hepatic BVRA enzymatic activity was relatively lower, compared to spleen or kidney from Blvra^WT mice, corresponding to ∼0.14 nM/μg protein/minute (Figure 3A). Liver from Blvra^G17A mice presented residual BVRA activity while those from Blvra^−/− mice showed no measurable activity (Figure 3A), corresponding to a ∼90%–93% and 95% reduction, compared to Blvra^WT controls, respectively (Figure 3B).

The oxidoreductase activity of BVRA was substantially diminished, yet detectable, in the spleen, kidney and liver of Blvra^E97A vs. Blvra^WT controls (Figure 3C). The spleen from Blvra^E97A mice exhibited lower BVRA enzymatic activity, corresponding to ∼0.15 vs. 0.75 nM/μg protein/minute in Blvra^WT mice in the first 20 min (Figure 3C). AUC analysis estimated a ∼80% reduction in BVRA enzymatic activity, yet higher in Blvra^E97A vs. Blvra^−/− mice (Figure 3D). Renal BVRA activity in Blvra^E97A mice was also decreased to ∼0.2 nM bilirubin/μg protein/min vs. ∼0.85 nM/μg protein/min in Blvra^WT mice (Figure 3C), corresponding to a ∼90%–95% reduction, similar to Blvra^−/− mice (Figure 3D). Liver from Blvra^E97A mice also showed a ∼85%–90% reduction of BVRA activity to ∼0.15 nM bilirubin/μg protein/min vs. 0.6 nM bilirubin/μg protein/min in Blvra^WT mice (Figures 3C and 3D).

These findings suggest that the G17A mutation in the NADPH/NADH binding motif of BVRA results in nearly complete loss of expression and activity. In contrast the E97A mutation in the reductase motif reduces BVRA activity while preserving its expression.

Non-canonical Blvra E97A and Blvra G17A activity

We tested whether silencing the endogenous NADPH/NADH binding (G17A) or reductase (E97A) motifs of BVRA impacts NRF2 activation, as demonstrated recently in neuronal cells.⁴⁶ To test this hypothesis, we exposed mouse bone marrow-derived macrophages (BMDM) to heme (Figure 4A), a potent pro-inflammatory agonist that induces a profound macrophage morphological⁵¹ and transcriptional remodeling,⁵² including the activation of NRF2.⁵³

Figure 4.

BVRA oxidoreductase activity is dispensable for heme-induced NRF2 activation and insulin receptor signaling in BMDM

(A) BMDM from Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice were stimulated with heme and NRF2 activation assessed by immunofluorescence.

(B) Representative immunofluorescence images of NRF2 nuclear translocation in BMDM following heme stimulation. Images show confocal z stack cross-sections with orthogonal views. NRF2 is shown in red, DAPI nuclear stain in blue, and phalloidin (cytoskeleton marker) in white. Scale bars, 5 μm.

(C) Quantification of cytoplasmic (left graph) or nuclear (right graph) NRF2 MFI in BMDM stimulated with (+) or without (−) heme. Each data point represents a technical replicate. n = 5–10 replicates. Data in (C) is representative from two independent experiments with a similar trend.

(D) BMDM from Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice were stimulated with heme and NRF2-regulated gene expression assessed by qRT-PCR.

(E) RT-qPCR analysis of NRF2 target genes (Nqo1, Fth, Gclc, Hmox1, Nrf2, and Blvra) in BMDM stimulated with (+) or without (−) heme. Data expressed as gene/Rplp0 (2^−ΔCt). Each data point represents a technical replicate. n = 3–6 replicates from two independent experiments with a similar trend.

(F) BMDM from Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice were stimulated with insulin and INSR signaling assessed by western blot.

(G) Representative western blots showing IRS-1 phosphorylation at tyrosine 612 (IRS-1^Y612), at serine 307 (IRS-1^S307), total IRS-1, and vinculin as loading control in BMDM stimulated with insulin vs. control vehicle.

(H) Quantification of IRS-1^Y612 normalized to IRS-1^S307 in BMDM stimulated with (+) or without (−) insulin. Each data point represents a technical replicate. n = 3 replicates from one experiment. Data in (C, E, and H) are presented as individual data points with bar graphs showing mean ± SD. Statistical comparisons were performed using two-way ANOVA with Šídák’s post hoc test. ∗p < 0.05, ∗∗p < 0.01; NS, not significant.

Heme induced NRF2 protein expression in the cytoplasm of Blvra^WT but not Blvra^−/− BMDM, as quantified by immunofluorescence (Figures 4B, 4C, and S1E). This was restored in Blvra^G17A and Blvra^E97A BMDM, suggesting that BVRA regulates NRF2 expression via a mechanism independent of its oxidoreductase activity.

Heme induced NRF2 nuclear translocation to the same extent in Blvra^WT vs. Blvra^−/− Blvra^G17A and Blvra^E97A BMDM, as quantified by immunofluorescence (Figures 4B, 4C, and S1E). This suggests that BVRA does not modulate NRF2 nuclear translocation, consistent with prior studies.⁴⁶

The specificity of the antibody used to detect mouse NRF2 protein was confirmed by immunoblotting and confocal microscopy, using Nrf2-deficient (Nrf2^−/−) BMDM exposed to the 26S proteasome inhibitor MG-132, to induce NFR2 nuclear accumulation (S1A-D).

Heme also induced the accumulation of mRNA encoding a number of NRF2-regulated genes, including NAD(P)H:quinone oxidoreductase 1 (Nqo1), ferritin H chain (Fth), glutamate-cysteine ligase catalytic subunit (Gclc) and heme oxygenase-1 (Hmox1), in Blvra^WT BMDM (Figures 4D and 4E). The induction of these NRF2-regulated genes occurred to the same extent in Blvra^−/−, Blvra^G17A and Blvra^E97A, compared to Blvra^WT BMDM (Figures 4D and 4E).

Heme induced the Nrf2 mRNA accumulation in Blvra^WT but not in Blvra^−/−, Blvra^G17A nor in Blvra^E97A BMDM (Figures 4D and 4E). This suggests that the oxidoreductase activity of BVRA interferes with the positive autoregulatory loop, whereby NRF2 regulates its own expression via binding to an antioxidant response element (ARE) in its promoter.⁵⁴

Of note, Blvra mRNA expression in Blvra^WT BMDM was reduced in response to heme (Figures 4D and 4E), suggesting that heme represses Blvra transcription. This repression was also observed in Blvra^G17A and Blvra^E97A BMDM (Figures 4D and 4E), suggesting that this occurs via a mechanism that is not regulated by BVRA oxidoreductase activity.

We then asked whether the oxidoreductase activity of BVRA interferes with INSR signaling,^34,44 via the dual-specificity kinase domain of BVRA.^25,39,43 Insulin induced the phosphorylation of IRS-1 at tyrosine 612 (IRS-1^Y612) and at serine 307 (IRS-1^S307) in Blvra^WT BMDM (Figures 4F, 4G, S2A, and S2B), a marker of INSR activation⁵⁵ and repression,⁵⁶ respectively. The ratio of IRS-1^Y612 over IRS-1^S307 phosphorylation, reflecting a net induction of INSR signaling, was increased in Blvra^WT but not in Blvra^−/− BMDM (Figures 4F, 4G, and S2A, and S2B). The ratio of IRS-1^Y612 over IRS-1^S307 phosphorylation was induced in response to insulin in Blvra^G17A and Blvra^E97A BMDM (Figures 4F, 4G, S2A, and S2B). This suggests that endogenous BVRA regulates INSR activation via a mechanism that does not rely on its oxidoreductase activity.

BVRA enzymatic activity is essential to survive malaria

To investigate the contribution of BVRA oxidoreductase activity to its antimalarial effect,²⁷ we used Plasmodium chabaudi chabaudi (Pcc) AS, a well-established experimental model of malaria non-lethal to C57BL/6J mice.⁵⁷ We then quantified serum unconjugated bilirubin concentration, using an UnaG-based assay,⁵⁸ and compared multiple disease parameters in Blvra^WT vs. Blvra^−/−, Blvra^G17A and Blvra^E97A mice through a 15-day time course of infection.

Blvra^G17A mice presented barely detectable levels of unconjugated bilirubin at steady state, prior to Pcc infection, slightly above Blvra^−/− mice (Figures 5B and 5C), consistent with almost complete loss of BVRA expression (Figure 2B) and enzymatic activity (Figures 3A and 3B). In contrast, Blvra^E97A mice presented ∼1–2 μM of circulating unconjugated bilirubin at steady state, similar to Blvra^WT mice (Figures 5A–5C). This confirms that the BVRA E97A mutation retains protein expression (Figure 2D) but only residual enzymatic function (Figures 3C and 3D).

Figure 5.

BVRA oxidoreductase activity determines malaria survival through dose-dependent bilirubin production

(A) Experimental schematic showing Plasmodium chabaudi chabaudi (Pcc) infection in mice with different Blvra where bilirubin production was evaluated.

(B) Time course of serum unconjugated bilirubin concentration following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype) as measured by the UnaG-based assay. Data are shown as mean ± SD.

(C) Serum unconjugated bilirubin levels before infection (green shaded region from B) in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 3 to 6 per genotype).

(D) Serum unconjugated bilirubin levels at day 7 post-infection (gray shaded region from B) in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype).

(E) Experimental schematic for disease severity and survival analysis.

(F) Disease severity scores (DSS) over time following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype). Data are shown as mean ± SD.

(G) Disease trajectories showing the correlation between DSS and serum bilirubin levels for each genotype. Direction of disease trajectory is shown by arrows in each curve.

(H) Kaplan-Meier survival curves of Pcc-infected Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype). (A–H) Data pooled from n = 3 independent experiments.

Statistical comparisons were performed using two-way ANOVA (B and F); one-way ANOVA with Tukey’s multiple comparison test (C and D); log rank (Mantel-Cox) test (H). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 0.0001; NS, not significant.

Blvra^WT mice developed jaundice in response to Pcc infection (Figures 5A and 5B), consistent with described.^27,59 The concentration of unconjugated bilirubin increased progressively in the first few days, reaching the maximal level of ∼30–40 μM by day 7 post-infection (Figures 5A, 5B, and 5D). In contrast, Blvra^G17A and Blvra^−/− mice presented residual, and no detectable levels of circulating unconjugated bilirubin throughout the course of infection, respectively (Figures 5A, 5B, and 5D). On the other hand, Blvra^E97A developed mild jaundice in response to Pcc infection, reaching a maximal level of ∼10–15 μM by day 7, lower than infected Blvra^WT mice (Figures 5A, 5B, and 5D).

Malaria severity diverged markedly between Blvra^WT vs. Blvra^G17A, Blvra^E97A or Blvra^−/− mice (Figures 5E–5G) with Blvra^WT mice reaching disease severity scores (DSS) below 5 at days 7–9 after Pcc-infection, reducing DSS to 0 thereafter (Figures 5E–5G). In contrast, Blvra^G17A and Blvra^E97A mice reached maximal DSS scores of 10–13 at days 7–9 after Pcc-infection (Figures 5E–5G), similar to Blvra^−/− mice (Figures 5E–5G).

The disease trajectories,^20,60,61 established by the relationship of DSS and circulating unconjugated bilirubin, presented counter clock progressions (Figure 5G). Blvra^WT increased the concentration of circulating unconjugated bilirubin to ∼35–40 μM maxima, concurrent with moderate ∼5–7 DSS scores, decreasing both parameters thereafter toward full disease recovery (Figure 5G). In contrast, Blvra^G17A mice displayed a near-vertical DSS trajectory reaching ∼12–13 disease score, concurrent with a maximal bilirubin concentration below 5 μM, somewhat above Blvra^−/− mice (Figure 5G). Blvra^E97A mice showed an intermediate disease trajectory whereby bilirubin concentration reached a maximal level of ∼15–20 μM, failing however, to restrain an upward progression of DSS toward ∼10–12 (Figure 5G).

Approximately 80% of Blvra^WT mice survived Pcc infection (Figure 5H), while Blvra^G17A and Blvra^E97A mice presented a progressive incidence of mortality starting at 7–8 days and reaching 100% by days 10–12 post-infection, similar to Blvra^−/− mice (Figure 5H).²⁷ This demonstrates that BVRA oxidoreductase activity is required to sustain circulating bilirubin above a 20–30 μM threshold required to mitigate malaria severity, similar to observed in P. falciparum malaria.²⁷

BVRA enzymatic activity exerts antiplasmodial effects

The kinetics of parasitemia (% of infected RBC) showed distinct temporal patterns between genotypes (Figures 6A and 6B). Initially, all genotypes exhibited similar infection dynamics through day 6 post-infection, reaching peak parasitemias of ∼50%–60% (Figures 6A and 6B). Blvra^WT mice presented parasitemias of ∼15%–20% at day 8 post-infection, compared to ∼30%–50% in Blvra^G17A, Blvra^E97A and Blvra^−/− mice (Figure 6C). The disease trajectories established by the relationship of DSS vs. parasitemia, also showed fundamental differences among genotypes (Figure 6D). While Blvra^WT mice progressed toward a maximal DSS of ∼5–6 before transitioning toward full recovery (Figure 6D), Blvra^G17A and Blvra^E97A mice presented an irreversible upward DSS progression to 13 (i.e., death), similar to Blvra^−/− mice (Figure 6D). This is consistent with the BVRA catalytic activity containing parasitemia.²⁷

Figure 6.

Biliverdin reductase A oxidoreductase activity is essential to limit life-threatening Parasitemia

(A) Experimental schematic for disease severity and parasitemia.

(B) Time course parasitemia (% infected RBC; iRBC) following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A and Blvra^E97A mice (n = 5 to 8 per genotype). Data are shown as mean ± SD.

(C) Highlighted iRBC levels at day 8 (gray shaded region from B). Each circle represents an individual mouse.

(D) Correlation between DSS and parasitemia for each genotype, showing inverse relationship in Blvra^WT mice. Direction of disease trajectory is shown by arrows in each curve.

(E) Experimental schematic for total RBC analysis.

(F) Time course of total RBC counts following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype). Data are shown as mean ± SD.

(G) Highlighted total RBC count at day 9 (gray shaded region from F). Each circle represents an individual mouse.

(H) Correlation between DSS and total RBC count for each genotype, showing inverse relationship in Blvra^WT mice. Direction of disease trajectory is shown by arrows in each curve.

(I) Experimental schematic for parasite burden analysis.

(J) Time course of parasite burden (total infected RBC per μL blood; iRBC) following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype). Data are shown as mean ± SD.

(K) Highlighted parasite burden levels at day 9 (gray shaded region from J). Each circle represents an individual mouse.

(L) Correlation between DSS and parasite burden for each genotype, showing inverse relationship most prominently in Blvra^WT mice.

(M) Experimental schematic for parasitemia and serum bilirubin.

(N) Disease trajectories showing the correlation between parasitemia and serum unconjugated bilirubin levels for each genotype. Direction of disease trajectory is shown by arrows in each curve.

(O) Experimental schematic for parasite burden and serum unconjugated bilirubin.

(P) Disease trajectories showing the correlation between parasite burden and serum unconjugated bilirubin levels for each genotype. Direction of disease trajectory is shown by arrows in each curve. (A–P) Data pooled from n = 3 independent experiments.

Statistical comparisons were performed using two-way ANOVA (B, F, and J); two-way ANOVA with linear mixed effects model (C and K) as primary analysis and treating independent experiments as linked experiments; one-way ANOVA with Tukey’s multiple comparison test (G) ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; NS, not significant.

Anemia severity was comparable among Blvra^WT vs. Blvra mutant mice, as monitored by the number of circulating RBC throughout the course of the infection (Figures 6E and 6F). Analysis of disease trajectories established by the relationship of DSS vs. total RBC count presented counterclockwise loop progressions (Figure 6H). While Blvra^WT mice developed maximal DSS of ∼5–6 at ∼2 × 10⁶ RBC/μL, transitioning to recovery from anemia, Blvra mutant mice underwent irreversible DSS progression under the same levels of anemia (Figure 6G). This suggests that severe anemia does not account for the vulnerability of Blvra mutant mice during Pcc infection.

The kinetics of parasite burden (nbr. RBC/μL) were similar to parasitemia (Figures 6I and 6J). Blvra^WT mice decreased parasite burden to approximately 2–3x10⁶ iRBC/μL at day 9 post-infection, while Blvra^G17A and Blvra^E97A mice presented 1.94- and 1.63-fold higher parasite burdens corresponding to ∼2–6x10⁵ iRBC/μL, respectively, similar to Blvra^−/− mice (Figure 6K). This is consistent with BVRA catalytic activity exerting an antimalarial effect²⁷ that contains malaria severity.

Disease trajectories representing the relationship of circulating bilirubin vs. parasitemia or parasite burden confirmed fundamental differences in Blvra^WT vs. Blvra mutant mice (Figures 6M–6P). Blvra^WT mice presented stable bilirubin concentrations at parasitemia below ∼50% (Figures 6M and 6N) and parasite burden of ∼2–3x10⁶ iRBC/μL (Figures 6O and 6P). Above this threshold, however, bilirubin concentration increased to reach a peak concentration of ∼30–35 μM, followed by a sharp decrease, associated with parasite clearance (Figures 6M–6P). In contrast, Blvra^G17A and Blvra^E97A mice failed to increase bilirubin and correspondingly, both parasitemia and parasite burden were sustained, similar to Blvra^−/− mice (Figures 6M–6P). This suggests that the accumulation of bilirubin above a ∼15–20 μM threshold is necessary for effective parasite clearance.

BVRA enzymatic activity regulates the host metabolic response to malaria

Comparison of host physiological parameters as a function of parasitemia and parasite burden revealed fundamentally distinct disease trajectories in Blvra^WT vs. Blvra mutant mice (Figure 7). Body temperature trajectories revealed profound thermoregulatory defects in Blvra mutant mice (Figures 7A–7D). All genotypes displayed clockwise loop trajectories with stable body temperature (∼37°C) until a parasitemia/parasite burden threshold presenting a downward hypothermic inflection thereafter (Figures 7A–7D). While Blvra^WT mice developed sub-lethal hypothermia of ∼33°C–35°C (Figures 7A and 7B), this was not the case for Blvra mutants, which developed irreversible downward trajectories culminating in life-threatening hypothermia (∼25°C–27°C) (Figures 7A–7D).

Figure 7.

Biliverdin reductase A oxidoreductase activity is essential to limit severe disease outcome

(A) Experimental schematic for body temperature analysis during infection.

(B) Body temperature over time following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype). Data are shown as mean ± SD.

(C) Correlation between body temperature and parasitemia for each genotype. Direction of disease trajectory is shown by arrows in each curve.

(D) Correlation between body temperature and parasite burden levels for each genotype. Direction of disease trajectory is shown by arrows in each curve.

(E) Experimental schematic for weight loss analysis during infection.

(F) Weight loss over time following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype). Data are shown as mean ± SD.

(G) Correlation between weight loss and parasitemia for each genotype. Direction of disease trajectory is shown by arrows in each curve.

(H) Correlation between weight loss and parasite burden levels for each genotype. Direction of disease trajectory is shown by arrows in each curve.

(I) Experimental schematic for glucose analysis during infection.

(J) Glucose levels over time following Pcc infection in Blvra^WT, Blvra^−/−, Blvra^G17A, and Blvra^E97A mice (n = 5 to 8 per genotype). Data are shown as mean ± SD.

(K) Correlation between glucose levels and parasitemia for each genotype. Direction of disease trajectory is shown by arrows in each curve.

(L) Correlation between glucose levels and parasite burden levels for each genotype. Direction of disease trajectory is shown by arrows in each curve. (A–L) Data pooled from n = 3 independent experiments. Statistical comparisons were performed using two-way ANOVA (B, F, and J). ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

Body weight trajectories showed differences between Blvra^WT vs. Blvra mutant mice analogous to those observed for thermoregulation (Figures 7E–7H). Blvra^WT mice restrained weight loss to ∼10%–15%, whereas Blvra^G17A, Blvra^E97A, and Blvra^−/− mice underwent irreversible ∼20%–30% weight loss trajectories consistent with severe cachexia (Figures 7E–7H).

Glycemia trajectories confirmed major differences between Blvra^WT mice vs. Blvra mutants (Figures 7I–7L). Blvra^WT mice maintained stable glycemia (∼100–150 mg/dL) throughout infection, whereas Blvra mutants presented upwards glucose trajectories progressing toward severe hyperglycemia (∼400–600 mg/dL) after the peak of parasitemia and parasite burden (Figures 7I–7L). Notably, Blvra^E97A mice exhibited the most pronounced hyperglycemic response, consistent with BVRA oxidoreductase activity controlling glycemia during malaria.

Discussion

Through the generation and characterization of mouse strains carrying hypomorphic Blvra mutations, we demonstrate that BVRA-mediated antimalarial protection correlates with unconjugated bilirubin levels. Both mutations result in reduced BVRA enzymatic activity alongside a reduction of circulating bilirubin levels upon Plasmodium infection, supporting that bilirubin production is central to BVRA’s antimalarial effect. However, we cannot exclude that these mutations also affect protein expression, and as such the relative contributions of BVRA oxidoreductase activity vs. non-canonical activities^{25,37,38,39,40,41,42} to antimalarial protection remain to be definitively established through mutations that selectively abolish catalytic activity while preserving protein expression.

Silencing the endogenous oxidoreductase activity of BVRA in macrophages did not compromise the induction of a number of NRF2-regulated genes in response to heme (Figures 4A–4D). We noticed however, that BVRA oxidoreductase activity was required to sustain NRF2 expression (Figures 4A–4D), presumably via a positive autoregulatory loop enforcing NRF2 expression.⁵⁴ These observations suggest that the protective effect exerted by the oxidoreductase activity BVRA against malaria is not readily explained by a failure to induce the expression of key NRF2-regulated genes, such as Hmox1 and Fth, which exert a major impact on malaria severity.^15,20,21 The impact of the oxidoreductase activity BVRA on the NRF2 positive feedback loop⁵⁴ and how this could affect chronic outcome of malaria remains to be tested experimentally.

Silencing the endogenous oxidoreductase activity of BVRA also did not exert a major impact over INSR signaling in macrophages (Figures 4E–4G, and S2). This suggests that the oxidoreductase activity of BVRA does not interfere with the dual-specificity kinase domain of BVRA,^25,39,43 involved in the regulation of INSR signaling.^34,44 Importantly, regulation of glycemia does not appear to play a major role in the outcome of Pcc infection,⁶¹ further supporting the notion that the protective effect of BVRA against Pcc infection is not readily explained by a putative effect of its dual-specificity kinase domain over INSR signaling.

The observation that the BVRA G17A mutation compromises BVRA protein expression despite physiologic mRNA expression levels in Blvra^G17A mice (Figures 2A and 2B), suggests that the glycine at position 17 within the NAD(P)H-binding domain confers conformational flexibility critical for cofactor binding and maintenance of proper protein folding. The BVRA E97A mutation severely impairs oxidoreductase activity while still allowing for protein expression (Figures 2C and 2D), suggesting that the NAD(P)H-binding domain supports cofactor recognition and structural integrity, whereas the reductase domain primarily governs bilirubin production.

The observation that Blvra^E97A mice maintain physiological bilirubin levels at steady state (Figures 5A and 5B) is critical for mechanistic interpretation of the antimalarial effects of jaundice.²⁷ It demonstrates that E97A mutation is an oxidoreductase hypomorph that allows circulating bilirubin to accumulate after birth, suggesting that the antimalarial effects of jaundice²⁷ are not related to a putative pre-sensitization mechanism due to baseline bilirubin deficiency.

The dose-response relationship across the Blvra mutant lines enables interpolation of the minimum effective concentration of unconjugated bilirubin providing antimalarial protection (Figures 5A–5D). As Blvra^E97A mice succumb to Pcc infection with kinetics nearly identical to Blvra^G17A and Blvra^−/− mice, this indicates that the protective threshold for unconjugated bilirubin lies between 20 and 30 μM. The convergence of a protective 20–40 μM range in mice and 8–50 μM range in asymptomatic humans,²⁷ supports the idea of evolutionary conservation across host and Plasmodium species.

We note that the apparent high BVRA activity in the spleen (Figure 2), may reflect heme catabolism by erythrophagocytic macrophages,⁶² a cell compartment depleted during malaria.²³ To what extent the contribution of BVRA expression in the spleen contributes to the overall antimalarial effect of bilirubin remains, however, to be established.

Mechanistically, bilirubin inhibits P. falciparum growth through a coordinated targeting of mitochondrial function, inhibition of hemozoin crystallization and disruption of the parasite food vacuole.^27,63 Bilirubin disrupts the parasite’s mitochondrial dihydroorotate dehydrogenase (DHODH), a NADPH-dependent oxidoreductase that catalyzes the rate-limiting step in the de novo pyrimidine synthesis, therefore, repressing parasite proliferation.⁶⁴ Moreover, bilirubin interferes with the parasite’s capacity to detoxify heme into hemozoin,²⁷ similar to, but far less potently than quinoline- and artemisinin-based antimalarial drugs.⁶⁵

The antiplasmodial effect of bilirubin probably explains why all Plasmodium spp. lack a heme oxygenase system and instead detoxify heme through hemozoin crystallization. This fundamental asymmetry provides bilirubin with selective antiparasitic activity.²⁷

We have previously shown that Plasmodium parasites exhibit increased virulence when transferred from infected to naive Blvra-deficient mice, establishing that bilirubin reduces parasite virulence.²⁷ Whether parasite virulence also increases when the concentration of bilirubin falls below the protective 20–30 μM threshold has not been directly established.

The disease trajectory of Blvra^WT mice vs. Blvra mutants provides a mechanistic framework for understanding how BVRA oxidoreductase activity determines malaria outcomes. The superimposable trajectories of Blvra mutants across thermoregulation (Figures 7A–7D) and body weight (Figures 7E–7H) suggest that 20–30 μM bilirubin threshold operates through resistance rather than enhanced disease tolerance, consistent with bilirubin’s direct antiplasmodial activity.²⁷

We have previously shown that bilirubin administration restores survival in Blvra-deficient mice,²⁷ providing proof-of-concept for bilirubin-based therapeutics. Improved bilirubin formulations,⁶⁶ such as pegylated bilirubin⁶⁷ and hyaluronic acid coated bilirubin nanoparticles,^68,69 could potentially achieve sustained therapeutic concentrations (20–30 μM), below those associated with kernicterus risk in neonates (>340–425 μM).

In conclusion, converging evidence from genetic, biochemical, and pharmacological approaches, unified by UnaG-based quantification methodology, establishes that unconjugated bilirubin concentrations of 20–30 μM represent the minimum threshold for antimalarial protection, providing a quantitative framework for therapeutic development and biomarker-guided treatment strategies.

Limitations of the study

Our study identifies that BVRA activity is essential to limit malaria severity in mice. However, we cannot exclude that these mutations also reduce BVRA protein expression alongside enzymatic activity, precluding definitive separation of effects attributable to oxidoreductase activity from protein expression. While our data demonstrate that bilirubin levels correlate with antimalarial protection, whether non-canonical activities of BVRA also contribute to this protective effect remains to be established. The definition of a bilirubin threshold limiting malaria severity was inferred from a well-established experimental model, which however, differs from the human disease. Therefore, the precise threshold for bilirubin protection during P. falciparum malaria requires further investigation.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Miguel P. Soares (miguel.soares@gimm.pt).

Materials availability

C57BL/6J Blvra^G17A and C57BL/6J Blvra^E97A mouse strains generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• •All data reported in this paper will be shared by the lead contact upon request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

The authors are indebted to all members of the Inflammation Group (GIMM) for insightful technical and intellectual contributions, to the staff at the GIMM Flow Cytometry and Rodent Facilities.

This work was supported by GIMM-CARE (European Union grant no. 101060102. GIMM-CARE is co-funded by the Portuguese Government, the Foundation for Science and Technology (FCT), ARICA – Investimentos, Participações e Gestão, Jerónimo Martins, the Gulbenkian Institute for Molecular Medicine, and Lisbon Academic Medical Centre (CAML) (https://doi.org/10.3030/101060102), and by national funds through FCT under the Associate Laboratory Programme (LA/P/0082/2020) (https://doi.org/10.54499/LA/P/0082/2020), and R&D Unit Funding Programme (UID/06357/2025) (https://doi.org/10.54499/UID/06357/2025). We acknowledge individual support by: Fundação para a Ciência e a Tecnologia (UI/BD/152257/2021 to M.M., 2020.04797.BD, COVID/BD/153665/2024 to A.F.; FEDER/29411/2017 to S.R.; PTDC/MED-FSL/4681/2020 https://doi.org/10.54499/PTDC/MED-FSL/4681/2020 to S.P.; CEECIND/01589/2017 and https://doi.org/10.54499/2023.11177.PEX to A.R.C.; 2023.09168.CEECIND to E.J.; FEDER/29411/2017, PTDC/MED-FSL/4681/2020 https://doi.org/10.54499/PTDC/MED-FSL/4681/2020, 2022.02426.PTDC https://doi.org/10.54499/2022.02426.PTDC and Congento LISBOA-01-0145-FEDER-022170 to M.P.S. Gulbenkian Foundation (S.R., S.C., M.P.S., and IBB 2021-51/BI-D/2021 to S.T.). GIMM Foundation (GIMM/BI/36-2025 to M.M., GIMM/BI/37-2025 to S.T., GIMM Cross-Site collaborative project to M.P.S.). la Caixa Foundation HR18-00502 (E.J. and M.P.S.). American Heart Association/Paul Allen Frontiers Group (Project 19PABH134580006 to B.D.P.). DFG Cluster of Excellence “Balance of the Microverse” EXC 2051; 390713860 (E.J., M.P.S. as associated member). NIH/NIA (1R21AG073684–01, R01AG071512 to B.D.P.). The Johns Hopkins Catalyst Award (B.D.P.). Solve ME/CFS Initiative (grant 90089823 to B.D.P.). US Public Health Service (grant DA044123 to B.D.P). Oeiras-ERC Frontier Research Incentive Awards (M.P.S.). H2020-WIDESPREAD-2020-5-952537 SymbNET Research Grants (M.P.S.).

Author contributions

Conceptualization, M.P.S. and M.Mesquita; formal analysis, M.Mesquita, A.F., and R.M.; resources, B.D.P.; investigation, A.F., A.N., A.R.C., M.C., M.Mallo, M.Mesquita, R.M., S.T.R., S.R., S.P., S.C.; visualization, M.Mesquita. and M.P.S.; funding acquisition, M.P.S.; project administration, M.P.S.; supervision, M.P.S. and E.J.; writing—original draft, M.P.S.; writing—review and editing, M.Mesquita.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the author(s) used Claude to edit the manuscript body of text. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Polyclonal anti-Biliverdin Reductase A	ThermoFisher Scientific	Cat#PA5-92059; RRID: AB_2806368
Rabbit monoclonal anti- beta-Actin	Cell Signaling Technology	Cat#4970; RRID: AB_2223172
Rabbit monoclonal anti- IRS-1	Cell Signaling Technology	Cat#3407; RRID: AB_2127860
Rabbit polyclonal anti- Phospho-IRS-1 (Tyr612)	GeneTex	Cat#GTX24868; RRID: AB_380452
Rabbit polyclonal anti- Phospho-IRS-1 (Ser307)	Cell Signaling Technology	Cat#2381; RRID: AB_330342
Rabbit polyclonal anti- Vinculin	Cell Signaling Technology	Cat#4650; RRID: AB_10559207
HRP conjugated goat anti-rabbit IgGH+L	Invitrogen	Cat# 31460; RRID: AB_228341
Rabbit monoclonal anti-Nrf2	Cell Signaling Technology	Cat#12721; RRID: AB_2715528
Donkey Anti-Rabbit IgG H&L (Alexa Fluor® 568)	Abcam	Cat# ab175470; RRID: AB_2783823
Phalloidin-conjugated Alexa Fluor 647	ThermoFisher Scientific	Cat#A22287; RRID: AB_2620155
Chemicals, peptides, and recombinant proteins
Standard, L10, 10 mM, Latex Particle	Beckman Coulter	Cat# 6602796
RPMI 1640 w/Glutamax	ThermoFisher Scientific	Cat# 61870036
Hemin	Frontier Scientific	Cat# H651-9
Bilirubin	Frontier Scientific	Cat# B584-9
TripleXtractor	Grisp	Cat# GB23
Xpert cDNA Synthesis Supermix	Grisp	Cat# GK86.0100
iTaq Universal SYBR Green Supermix	Bio-Rad	Cat# 1725124
Paraformaldehyde	ThermoFisher Scientific	Cat# 043368.9M
cOmplete™, Mini, EDTA-free Protease Inhibitor CocktailL	Roche	Cat# 11836170001
Bio-Rad Protein Assay Dye Reagent Concentrate	Bio-Rad	Cat# 5000006
SuperSignal™ West Pico PLUS Chemiluminescent Substrate	ThermoFisher Scientific	Cat# 34580
Insulin solution human	Merck	Cat# I9278
MG-132	Merck	Cat# 474790
Experimental models: Organisms/strains
C57BL/6J	Jackson Laboratory	Cat#000664; RRID: IMSR_JAX:000664
C57BL/6J Blvra−/−	Vasavda et al.30	Provided by Dr. B. Paul
C57BL/6J BlvraG17A	This manuscript	N/A
C57BL/6J BlvraE97A	This manuscript	N/A
C57BL/6J Nrf2−/−	Jackson Laboratory70	Cat#017009; RRID: IMSR_JAX:017009
Plasmodium chabaudi chabaudi strain: PccAS-GFPML	N/A	Kind gift from Dr. J. Thompson71
Oligonucleotides
Mouse gRNA BlvraG17A: GTGGTTGGTGTTGGCAGAGC	This manuscript	N/A
Mouse gRNA BlvraE97A: ATGCCATGGGGTATTCCACG	This manuscript	N/A
Mouse Rep. DNA BlvraG17A: CTCCATTG CATAGCTGGGCTGTTTTTAAA CCCCAC ATCGGTCTTTGATATTTCAGCCAAAGAG GAAATTTGGTGTGGTAGTGGTTGGTGT GGCACGGGCTGGCTCTGTGAGGATA AGGGACTTGAAGGATCCACACTCTTC AGCATTCCTAAACCTGATTGGATATGTGTC	This manuscript	N/A
Mouse Rep. DNA BlvraE97A: TTTGA AGTGTCCTTTCCC TGTCTTTTTGT TTTCAAAGGCAGTTTCTTCAGGCTG GCAAGCATGTCCTCGTCGCCTATC CCATGGCATTGTCATTTGCGGCAGC GCAGGAGCTGTGGGAGCTGGC TGCACAGAAAGGTGATGTT	This manuscript	N/A
Mouse Genotyping primer BlvraWT: Fwd TAGTGGTTGGTGTTGGCAGA; Rev: GGCTTTCCCTTACTCTGGGTC;	This manuscript	N/A
Mouse Genotyping primer BlvraG17A: TAGTGGTTGGTGTTGGCAGA; Rev: GGCTTTCCCTTACTCTGGGTC	This manuscript	N/A
Mouse Genotyping primer BlvraE97A Fwd: AGCATGTCCTCGTGGAATAC, Rev: TGGGAACTCAGTAGCAAAGCC	This manuscript	N/A
Mouse RT- qPCR primer Rplp0, Fwd: 5′-CTTTGGGCATCACCACGAA-3′, Rev: 5′-GCTGGCTCCCACCTTGTCT-3′	Weis et al.72	N/A
Mouse RT- qPCR primer Blvra, Fwd: 5′-AGCCGCTGGTAAGCTCC-3′, Rev: 5′-ACCAACCACTACCACACCAAA-3′	Figueiredo27	N/A
Mouse RT- qPCR primer Fth, Fwd: 5′- CCATCAACCGCCAGATCAAC-3′, Rev: 5′- GCCACATCATCTCGGTCAAA-3	Figueiredo27	N/A
Mouse RT- qPCR primer Hmox1, Fwd: 5′- TGACACCTGAGGTCAAGCAC-3′, Rev: 5′- TCTCTGCAGGGGCAGTATCT-3	Ramos et al.20	N/A
Mouse RT- qPCR primer Nqo1, Fwd: 5′- GGTAGCGGCTCCATGTACTC-3′, Rev: 5′- CGCAGGATGCCACTCTGAAT-3	This manuscript	N/A
Mouse RT- qPCR primer Nrf2, Fwd: 5′- AGGACATGGAGCAAGTTTGG-3′, Rev: 5′- ATCAGCCAGCTGCTTGTTTT-3	This manuscript	N/A
Mouse RT- qPCR primer Gclc, Fwd: 5′- ACATCTACCACGAGTCAAGGACC-3′, Rev: 5′- CTCAAGAACATCGCCTCCATTCAG-3	This manuscript	N/A
Recombinant DNA
Plasmid: UnaG pMAL-6P2-6xHIS	Vasavda et al.30	Kind gift from Dr. B. Paul
Software and algorithms
GraphPad Prism 8 GraphPad Softwa	GraphPad Software	https://www.graphpad.com; RRID: SCR_002798
ImageJ	NIH	https://ImageJ.net/ij/; RRID: SCR_003070
iBright Analysis Software	ThermoFisher Scientific	https://www.thermofisher.com; RRID: SCR_017632
QuantStudio 7 Flex Real-Time PCR System Software	Applied Biosystems	https://www.thermofisher.com; RRID: SCR_020245
SnapGene	Dotmatics	https://.snapgene.com;RRID:SCR_015052
BD FACSDiva FACSDiva software v 8.0.3	BD Biosciences	https://www.bdbiosciences.com; RRID: SCR_001456
Cell Profiler	Broad Institute73	https://cellprofiler.org;RRID:SCR_007358
SpectroFlo	CytekBio	https://cytekbio.com; RRID: SCR_019826
FlowJo	FlowJo	https://www.flowjo.com; RRID: SCR_008520

Experimental model and study participant details

Mice

Mice were bred and maintained under specific pathogen-free (SPF) conditions at the Gulbenkian Institute for Molecular Medicine (GIMM). Mice were housed at standard vivarium temperature (22°C) in a 12-h light/dark cycle with free access to water and standard chow pellets. All experimental protocols were approved in a two-step procedure, by the Animal Welfare Body of the IGC and by the Portuguese National Entity that regulates the use of laboratory animals in research (Direção Geral de Alimentação e Veterinária; DGAV). Experimental procedures followed the Portuguese (Decreto-Lei no 113/2013) and European (Directive 2010/63/EU) legislation. C57BL/6J mice were obtained from the GIMM animal facility. C57BL/6J Blvra^−/− mice were generated at Ozgene (Australia), as described.³⁰

C57BL/6J Blvra^G17A and C57BL/6J Blvra^E97A mouse strains were developed in collaboration with the GIMM’s transgenic unit. Briefly, they were generated by CRISPR/Cas9, using gRNAs targeting the GTGGTTGGTGTTGGCAGAGC or the ATGCCATGGGGTATTCCACG sequence respectively for the C57BL/6J Blvra^G17A and the C57BL/6J Blvra^E97A mice, and the replacement oligonucleotides CTCCATTGCATAGCTGG GCTGTTTTTAAACCCCACATCGGTCTTTGATATTTCAGCCAAAGAGGAA ATTTGGTGTGGTAGTGGTTGGTGTGGCACGGGCTGGCTCTGTGAGGATAAGGGACTTGAAGGATCCACACTCTTCAGCATTCCTAACCTGATTGGATATGTGTC or TTTGAAGTGTCCTTTCCCTGTCTTTTT GTTTTCAAAGGCAGTTTCTTCAGGCTGGCAAGCATGTCCTCGTCGCCTATCCCATGGCATTGTCATTTGCGGCAGCGCAGGAGCTGTGGGAGCTGGCTGCACAGAAAGGTGATGTT respectively for the C57BL/6J Blvra^G17A and C57BL/6J Blvra^E97A mice (all oligos purchased from Integrated DNA Technologies). This resulted in the replacement of glycine 17 for an alanine, in C57BL/6J Blvra^G17A, and glutamate 97 for an alanine, in C57BL/6J Blvra^E97A.

For better discrimination of Blvra^WT vs. Blvra^G17A strains during genotyping we replaced in the Blvra^G17A strain, the 5′ GTT and 3′AGA codons, flanking the mutated GGC->GCA codon, by the synonymous 5′ GTG and 3′CGG codons, respectively. For better discrimination of Blvra^WT vs. Blvra^E97A strains during genotyping we replaced in the Blvra^E97A strain, the 5′GTG and 3′TAC codons, flanking the mutated GAA->GCC codon, by the synonymous 5′GTC and 3′TAT codons, respectively.

Ribonucleoprotein complexes (1 μM gRNA, 100 ng/μL Cas9 protein) were mixed with the replacement oligo (30 ng/μL) and microinjected into the pronucleus of fertilized C57BL/6J mouse oocytes, which were transferred into the uterus of pseudo pregnant females, according to standard procedures. Genotyping was performed by PCR on genomic DNA using primers to amplify WT or mutant sequences. Sanger sequencing revealed 2 C57BL/6J Blvra^G17A founders and C57BL/6J Blvra^E97A 3 founders. These were bred with C57BL/6J mice to obtain heterozygous mice for the desired mutations. Offspring was genotyped using the following primer pairs: Blvra^WT: Fwd TAGTGGTTGGTGTTGGCAGA; Rev: GGCTTTCCCTTACT CTGGGTC; Blvra^G17A: TAGTGGTTGGTGTTGGCAGA; Rev: GGCTTTCCCTTACTCTGGGTC; Blvra^E97A Fwd: AGCATGTCCTCGTGGAATAC, Rev: TGGGAACTCAGTAGCAAAGCC. The progeny of positive mutant lines and C57BL/6J mice was subsequently bred to homozygosity.

Method details

RNA extraction and RT-qPCR

Mice were sacrificed by CO₂ inhalation at steady state, transcardially perfused in toto with ice-cold PBS (1X, 20 mL) and organs were harvested, snap frozen in liquid nitrogen and stored at −80°C. Total RNA was extracted using tripleXtractor reagent (GRISP), chlorophorm, isopropanol and ethanol, according to manufacturer’s instructions. cDNA was synthesized using the Xpert cDNA Synthesis Mastermix (GRiSP), followed by RT-qPCR using the iTaq Universal SYBR Green Supermix (Bio-Rad) on a QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems). Transcript values were calculated from the threshold cycle (Ct) of each gene using the 2^−ΔCT method using using Acidic ribosomal phosphoprotein P0 (Rplp0) as the housekeeping control gene. Primers for qPCR include: Rplp0, Fwd: 5′-CTTTGGGCATCACCACGAA-3′, Rev: 5′-GCTGGCTCCCACCTTGTCT-3’; Blvra, Fwd: 5′-AGCCGCTGGTAAGCTCC-3′, Rev: 5′-ACCAACCACTACCACACCAAA-3′; Fth, Fwd: 5′- CCATCAACCGCCAGATCAAC-3′, Rev: 5′- GCCA CATCATCTCGGTCAAA-3; Hmox1, Fwd: 5′- TGACACCTGAGGTCAAGCAC-3′, Rev: 5′- TCTCTG CAGGGGCAGTATCT-3; Nqo1, Fwd: 5′- GGTAGCGGCTCCATGTACTC-3′, Rev: 5′- CGCAGGATG CCACTCTGAAT-3; Nrf2, Fwd: 5′- AGGACATGGAGCAAGTTTGG-3′, Rev: 5′-ATCAG CCAGCTGCTTGTTTT-3; Gclc, Fwd: 5′- ACATCTACCACGAGTCAAGGACC-3′, Rev: 5′-CTCAAGA ACATCGCCTCCATTCAG-3.

BVRA enzymatic activity assay

BVRA enzymatic activity was measured as described.⁷⁴ Snap-frozen tissues were homogenized in homogenization buffer (10 mM Tris-HCl pH7.5, 250 mM sucrose) and the enzymatic activity was then measured in 50 μL of protein homogenate in assay buffer (100 mM Tris-base, 1 mM EDTA, pH 8.7, 1 mM NADPH, and 3 μM biliverdin), at 37°C in a spectrophotometer with shaking for 1h. The rate of reaction was determined by monitoring the change in absorbance at 450 nm over time. Bilirubin levels in the assay samples were determined using a standard curve of bilirubin IXα (0–300 μM) (Frontier Chemicals) and then normalized to cellular protein levels and presented as bilirubin concentration per μg of total protein. The activity assay was performed in a final volume of 600 μL.

BMDM generation and stimulation

Primary mouse BMDMs from wild-type (WT) and indicated mutant mice (Blvra^−/−, Blvra^G17A, Blvra^E97A) were grown for 6 days in RPMI (GIBCO) supplemented with 10% heat-inactivated fetal bovine serum (GIBCO), 10% L929 conditioned media and 1% penicillin and streptomycin.^75,76 BMDMs were seeded at a density of 2 × 10⁶ cells/well in 6 well plates, 1 × 10⁶ cells/well in 12 well plates or 100 × 10⁴ cells/well in 8-chamber slides (μ-slide, Ibidi) and allowed to adhere overnight. Cells were stimulated with 50 μM heme (Frontier Scientific) for 24h, 500 nM insulin (Merck) for 10 min or 10 μM MG-132 (Merck) for 6h and treated for downstream analysis. For microscopical assays, BMDM were fixed with 4% PFA and stained with DAPI (1 μg/mL) and phalloidin-Alexa Fluor 647 (0.2 μg/mL). Five random fields per well were acquired using a Zeiss LSM900 confocal microscope equipped with a Plan-Apochromat 20×/0.8 M27 or Plan-Apochromat 63x/1.4 Oil DIC M27 objectives Automatic image analysis was performed using the CellProfiler software,⁷³ which was programmed to: a) split images into DAPI, Phalloidin-conjugated to Alexa Fluor 647 (AF647) and Alexa Fluor 568 (AF568) channels; b) load the channels into the pipeline and identify primary objects (nuclei; DAPI) using Global method with minimum cross-entropy thresholding, and shape method to distinguish clumped objects; c) identify secondary objects (whole cell area; Phalloidin-AF647) based on the primary objects (nuclei) identified before using the propagation method with 3-class Otsu adaptive thresholding; d) generate cytoplasm objects by masking detected nuclei from the whole cell area; e) extract fluorescence intensity of NRF2-AF568 in nuclei and cytoplasm. Primary and secondary objects touching the image borders are excluded from analysis. The level of cytoplasmic or nuclear translocation of NRF2 is expressed as the mean fluorescence intensity (MFI) of AF568, per captured microscopy field.

Western blot

Mice were sacrificed, transcardially perfused in toto with ice-cold PBS (1X, 20 mL) and organs were harvested and kept at −80°C until tissue lysates preparation. Tissues were lysed using 2% SDS-PAGE sample buffer (100 mM Tris, pH 6.8, 20% glycerol, 4% SDS, 0.2% bromophenol blue, 100 mM DTT and 1X protease inhibitor cocktail (cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail; Roche)) or NP40 extraction buffer (0.15M NaCl, 1% NP-40, 0.05M Tris, 1X protease inhibitor cocktail (cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail; Roche) and homogenized in a tissue lyser (Qiagen) with tungsten carbide beads (Qiagen). For the SDS-PAGE sample method, supernatants were collected and total protein was quantified at λ₂₈₀nm using the DS-11 FX Spectrophotometer (DeNovix). For NP40 method, supernatant was collected and protein was quantified via Bradford assay (BioRad). Protein was resolved (50 μg) on a 12% SDS-PAGE and transferred to Polyvinylidene fluoride (PVDF) membranes. Membranes were blocked for 1 h at room temperature (5% bovine serum albumin in 1X TBS-T), washed in 1X TBS-T and incubated with primary antibodies, overnight at 4°C. The primary antibodies used were rabbit polyclonal anti-BVRA (Thermofisher scientific, PA5-92059; 1:750), rabbit monoclonal anti- IRS-1 (Cell Signaling Technology, 3407; 1:1000), rabbit polyclonal anti- Phospho-IRS-1 (Tyr612) (GeneTex, GTX24868, 1:1000), rabbit polyclonal anti- Phospho-IRS-1 (Ser307) (Cell Signaling Technology, 2381, 1:1000, rabbit polyclonal anti- Vinculin (Cell Signaling Technology, 4650, 1:1000) and rabbit polyclonal anti-β-actin (Cell Signaling Technology, 4967; 1:1000). Membranes were washed 3 times (1X TBS-T) and incubated (1 h; RT) with the peroxidase-conjugated secondary antibody (HRP conjugated goat anti-rabbit IgGH+L; Invitrogen, 31460; 1:5000). Membranes were washed 3 times (1X TBS-T) and peroxidase activity was detected using SuperSignal West Pico PLUS Chemiluminescent Substrate (ThermoFisher Scientific). Blots were developed using Amersham Imager 680 (GE Healthcare), equipped with a Peltier cooled Fujifilm Super CCD or using an iBright FL1500 (ThermoFisher Scientific). Densitometry analysis was performed using ImageJ software, from images without saturated pixels.

Plasmodium chabaudi chabaudi AS infection and disease assessment

Mice (females and males, 10–16 weeks old) were infected with Plasmodium chabaudi chabaudi AS transgenic GFP-expressing Pcc AS (Pcc AS-GFPML).²⁷ Infections were performed by intraperitoneal (i.p.) administration of freshly isolated blood (Passage 29; 2 × 10⁶ infected RBC diluted in 200 μL PBS) collected from a previously infected C57BL/6J mouse. Mice were monitored daily from day 0 (day of infection) onwards for parasitemia (% infected RBC; iRBC), parasite burden (number of iRBC per μL of blood), body weight (Ohaus CS200 scaler, Sigma Aldrich), core (i.e., rectal) body temperature (Rodent thermometer; BIO-TK8851, Bioset), blood glucose concentration (AccuCHECK Performa glucometer, Roche) and survival, essentially as described.²⁷ Briefly, the number of RBC per μL of blood was quantified by flow cytometry (LSR Fortessa X20 analyzer; BD Bioscience) using a standard concentration of reference latex beads (10 μm; Coulter CC Size Standard L10, Beckman Coulter, 6602796), gating on RBC, based on size and granularity and on bead population.

UnaG protein synthesis and purification

UnaG protein synthesis and purification was done as before.³⁰ Briefly, UnaG was expressed in pMAL-6P2-6xHIS in BL21(DE3) cells.³⁰ Starter cultures were grown to saturation in Luria Broth (LB) (overnight at 37°C), diluted 10-fold in LB and grown to an OD600 of 0.3 at 37°C, after which cultures were moved to 18°C. UnaG expression was induced at OD600 of 0.6, by the addition of 400 μM isopropyl β-D-1thiogalactopyranoside (16 h, 18°C). Cells were harvested by centrifugation (3,800 g, 4°C, 25 min), resuspended in 15 mL of resuspension buffer (50 mM HEPES, 300 mM NaCl, 0.5 mM TCEP, 10% glycerol, 1 mM PMSF, 2.34 μM leupeptin, 1.45 μM pepstatin at pH 7.4) per liter of culture. The sample was further lysed by sonication (QSonica Sonicator) with an amplitude of 5% in a pulse mode of 0.8s ON and 0.5s OFF for total 30 s. Lysate was clarified by centrifugation (26,000 g, 4°C, 30 min) and loaded onto an amylose column (MBPTrap HP Column, GE Healthcare). Protein was eluted with 20 mM maltose in protease buffer (50 mM Tris, 150 mM NaCl, 0.5 mM TCEP, 10% glycerol, 0.01% Triton X-100 at pH 7.4). MBP was removed by the addition (16 h, 4°C) of PreScission Protease (GE Healthcare). UnaG was purified via a nickel column (HisTrap HP Column, GE Healthcare) to remove the cleaved tag and eluted with (30 mM Tris, 1M NaCl, 0.5 mM TCEP, 10% glycerol, 500 mM imidazole, pH 7.4). UnaG was further purified on a gel-filtration column (S-200, GE Healthcare) in gel filtration buffer (50 mM HEPES, 300 mM NaCl, 0.5 mM TCEP, 10% glycerol at pH 7.4), concentrated using 10 kDa amicon ultracentrifugal filter (Merck), and flash frozen in gel filtration buffer at 30% glycerol for storage at −80°C.

UnaG-based assay for quantification of unconjugated bilirubin

Blood from Blvra^WT, Blvra^−/−, Blvra^G17A and Blvra^E97A mice was obtained from the tail vein, before (Day 0) and during Pcc infection. Samples were collected and immediately centrifuged (1000 g, 15 min) and plasma was collected, frozen in liquid nitrogen and stored at −80°C until used for bilirubin quantification. Total plasma unconjugated bilirubin was determined using UnaG.⁷⁷ Plasma samples from mice noninfected or infected with Pcc was diluted 1:100 in PBS. Then, 100 pM of UnaG was added to the diluted plasma sample and incubated for 10 min at RT. After incubation, fluorescence intensity was measured at excitation λ₄₈₀ nm and emission λ₅₃₀ nm using a microplate reader (Promega GloMax). Bilirubin IXα (Frontier Chemicals) was used as standard.

Quantification and statistical analysis

All statistical analyses were performed using GraphPad Prism 10 software. Statistically significant differences between more than two groups were assessed using two-way ANOVA with Tukey’s multiple comparison test or two-way ANOVA with Linear Mixed Effects Model. Survival curves are represented by Kaplan-Meier plots and differences assessed using the log rank test. One-way ANOVA with Tukey’s post hoc test was used for comparisons of protein expression, mRNA expression, bilirubin levels and enzymatic activity. two-way ANOVA with Sidak’s post hoc test was used for NRF2 immunofluorescence and insulin signaling experiments. Differences were considered statistically significant at a p value < 0.05. NS: not-significant, p > 0.05; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. The exact value of n and what n represents is stated in the corresponding figure legends. Data are presented as mean ± SD. All animal experiments were performed using male and female mice and no sex-dependency was observed.

Published: April 30, 2026