Introduction

Access to safe, affordable, and effective chemopreventive agents for the clinical crisis of severe neonatal hyperbilirubinemia (SNH) remains an unmet, urgent, and current global health need. A reliable, single intervention strategy could bridge the ongoing fractured maternal child health services in disrupted and dislocated communities.1,2,3 The burden of persistent, unmonitored, and untreated SNH in resource-constrained communities is well documented.4,5,6 The adverse consequences are well known such that the risk of irreversible life-long brain injury places about 2% of the world's newborn population at jeopardy of both morbidity and mortality.5,6,7,8 Introductions of biologically relevant hemeoxygenase inhibitors, invented at the Rockefeller Institute,9,10,11,12 and supported by US academic (NIH) and corporate sponsors (tin-mesoporphyrin3,10,13) failed to attain United States Food and Drug Administration (US-FDA) approval.14

Recent advances in nano-chemistries, such as applications of nanometer-sized colloidal particles (i.e., nanoparticles), have emerged for their unique tunable electronic, optical, and magnetic properties that are derived from the nanoscale dimensions and compositions.15,16,17,18 In fact, several therapeutic and diagnostic modalities have been reported for a diverse array of nanomaterials.18,19,20 Doxil® (doxorubicin liposomes) is one such agent that has been approved by the US-FDA for treatment of AIDS-associated Kaposi’s sarcoma.21 Similarly, Abraxane®, a 130-nm paclitaxel-functionalized albumin, was approved as a second-line drug treatment for breast cancer.22 Other clinical nano-pharmacotherapeutics include Combidex® (ferumoxtran-10), DexFerrum®, Ferrlecit®, Injectafer®, etc. (comprehensively reviewed in ref. 19). Engineered multivalent manganese oxide (i.e., MnxOy) nanomaterials were recently considered as therapeutic agents in animal models of hepatic fibrosis,23 chronic kidney disease24 as well as neurodegenerative disorders25,26,27 and encouraged our inquiry. Manganese (Mn) is an essential trace mineral nutrient and serves as a co-factor for key metabolic enzymes such as superoxide dismutase (SOD), arginase, and pyruvate carboxylase.28 The bioavailability of Mn is well documented and promoted for its vital role in the metabolism of amino acids, cholesterol, glucose, and carbohydrate; scavenging of reactive oxygen species; bone formation; synthesis of sex hormones, and immune response.29,30,31,32 Additional beneficial properties of nano-doses of manganese include its exciting redox modulatory and catalytic properties for therapeutic applications.27,33,34,35 On the other hand, occupational over-exposure to Mn has known adverse events36,37 and led to designated safe therapeutic thresholds.38 Prior to embarking on this current study, we demonstrated that citrate-functionalized manganese oxide-nanoparticles (C-Mn3O4 NPs) do not manifest characteristic evidence of Mn-neurotoxicity; rather, Mn-induced neurologic disorder (e.g., idiopathic Parkinson’s disorder) was ameliorated through mitochondrial protection and intracellular redox modulation.26 Biocompatibility and safety of C-Mn3O4 NPs have also been reported.39 Our preliminary experiments with these nanoparticles were also suggestive of unprecedented catalytic activity toward bilirubin degradation without any photo- or chemo-activation in a controlled in vitro system.40

For this current study, we designed a cohort of two-phased experiments to examine the processes and mechanisms for bilirubin degradation by C-Mn3O4 NPs (the nanodrug) in a proven rodent model specifically bred to manifest SNH and associated neurotoxicity. Each set of cohorts comprised: (i) primary control of healthy rodents, (ii) control SNH model with phenylhydrazine (PHz)-induced SNH (kernicteric model in the second phase), (iii) SNH model treated with the chosen nanodrug as the specific experimental model (kernicteric model treated with the nanodrug), and (iv) another control cohort of healthy rodents treated with nanodrug alone. Our primary outcome measures were (a) to determine both systemic and neural bilirubin degradation; (b) effect on neuromotor function and neurobehavioral abilities of survivors in the kernicteric model; and (c) the pharmacokinetics and dynamics of the nanodrug.

Methods

Rodent model

To evaluate the in vivo effectiveness of the nanodrug in degradation of bilirubin, we developed a chemical (i.e., PHz) induced mice model that closely resembles the pathophysiological condition of human neonatal jaundice characterized by high systemic unconjugated bilirubin (UCB) due to higher hemoglobin (Hb) turnover, shorter red blood cell lifespan and inability of developing liver to efficiently eliminate or conjugate bilirubin.41,42

All animal studies and experimental procedures were performed at Central Animal Facility, Department of Zoology, Uluberia College, India (Reg. No.: 2057/GO/ReRcBi/S/19/CPCSEA) following the protocol approved by the Institutional Animal Ethics Committee (Ref: 02/S/UC-IAEC/01/2019). Non-diabetic infant C57BL/6j mice of both sexes (age: 6 days, BW: 7.1 ± 0.9 g) were used. For detailed methods see the Supplementary Methods section.

Induction of SNH

PHz-intoxication induced drastic catabolism of the circulating Hb to accumulated bilirubin and lead to SNH by hemolysis without hepatic involvement.42,43,44 To minimize the potential impact of UDP-glucuronosyltransferases (UGT1) maturation on bilirubin clearance, we completed our studies within postnatal day 14 as prior studies suggested the expression of UGT1 in mice commences at postnatal age 14 days.6,45

Randomization and dosing

Mice were randomly divided into four cohorts (Phase I study): (i) healthy control (n = 20): healthy control received a placebo, normal saline (150 µL, po); (ii) SNH model (n = 30): SNH model received three doses of PHz (30 mg kg−1 BW, i.p.) at postnatal ages 6, 8 and 10 days; (iii) nanodrug treated (n = 30): the experimental cohort of SNH model treated with nanodrug, received C-Mn3O4 NPs (0.25 mg kg−1 BW; po) at postnatal age 10 days; (iv) nanodrug control (n = 20): healthy rodent cohort treated with the nanodrug, C-Mn3O4 NPs (0.25 mg kg−1 BW; oral) alone as the non-SNH model at postnatal age 10 days. All primary outcomes were evaluated at age 10 days.

To assess the effect on bilirubin neurotoxicity, a kernicterus-like syndrome, a series of experiments were designed on four rodent cohorts (Phase II study) randomized to: (i) healthy control (n = 20): healthy controls treated with an oral saline placebo (150 µL; po); (ii) SNH model (n = 30): SNH model administered with three serial doses of PHz (30 mg kg−1 BW, i.p.) at postnatal ages 6, 8 and 10 days; (iii) nanodrug treated (n = 30): SNH model serially treated with C-Mn3O4 NPs (0.25 mg kg−1 BW; po) at postnatal ages 6, 8, 10 and 12 days; (iv) nanodrug control (n = 20): healthy cohort solely treated with C-Mn3O4 NPs (0.25 mg kg−1 BW; po) at postnatal ages 8, 10, and 12 days. Brain samples (n = 10) were collected from each group on day 13. All neurobehavioral studies were performed in the remainder cohorts between postnatal days 14 and 18. Supplementary Table 1 outlines the timeline of the phased experiments.

Serum bilirubin assay

Blood samples were collected from retro-orbital-plexus and serum was instantly separated by centrifugation at 4 °C, in dark. Total serum bilirubin (TSB), and UCB were measured following di-azo method using commercially available test kits (Autospan Liquid Gold, Span Diagnostic Ltd., India).

Brain gross anatomy

The gross anatomy of the brain was studied immediately after excision at the postnatal age of 14 days. The brains were photographed using a digital camera in the manual mode keeping all the instrumental parameters the same across groups. The deposition of bilirubin was calculated by counting yellow pixel density in Photoshop 2022 (Adobe Inc.).

Brain histopathology

For histopathological investigation, brains were removed from the sacrificed animals, and fixed in 10% buffered formalin for 7 days. Paraffin-embedded sections (5–8 µm) were stained with hematoxylin and eosin following the standard method.

Neural bilirubin loads (post-mortem)

Gross neural bilirubin loads were measured using a previously published protocol.46 Briefly, the brain sections were homogenized, and bilirubin was extracted using an organic extraction solvent: methanol/chloroform/n-hexane (63:31:6, v/v) and measured using a spectrophotometer (Model UV-Vis 2600, Shimadzu, Japan) at 420 nm.

Brain oxidative status

Extent of lipid peroxidation (i.e., malonyldialdehyde (MDA) level) was determined in the brain homogenates according to Buege methodology.47 The activity of SOD, catalase (CAT), glutathione peroxidase (GPx), glutathione-disulfide (GSSG), and reduced glutathione (GSH) were estimated using commercially available test kits (Sigma-Aldrich, MO).

Distribution in the brain

Mn-contents in the brain (12 h after po administration of nanodrug) were assayed by inductively coupled plasma atomic emission spectroscopy (ICP-AES) (ARCOS-Simultaneous ICP Spectrometer, SPECTRO Analytical Instruments, Germany). Open acid-digestion method was used during the sample preparation.48

Neuromotor function analysis (survivors)

The neuromotor functions were evaluated using four tests: (i) open field test (OFT) for measurements of spontaneous activity; (ii) balance beam test for fine motor control; (iii) pole test for motor coordination; and (iv) gait analysis for movement pattern. See Supplementary Methods for details.

Cognitive function test

To assess hippocampus-dependent learning and memory, we used two variants of well-known Morris water maze (MWM) test: (i) the hidden platform test, and (ii) cue reversal test. Furthermore, we performed novel object recognition (NOR) test to assess recognition memory. See Supplementary Methods for details.

Pharmacokinetics and dynamics (PK-PD) studies

For PK-PD analysis a separate group of 36 animals were induced SNH with three doses of PHz (30 mg kg−1 BW, i.p.) at postnatal ages 6, 8, and 10 days and then treated with the nanodrug (0.25 mg kg−1 BW, po at postnatal day 10) and were sampled at 0, 0.5, 1, 2, 4, 6, 8, 12, and 24 h (n = 4 each time). Mn-contents were assayed by ICP-AES.

Statistical analysis

All quantitative data are expressed as mean ± standard deviation, unless otherwise stated. Unpaired t-test with Welch’s correction was used to compare between two groups. One-way analysis of variance followed by correction of false discovery rate (post hoc FDR: two-stage step-up method of Benjamini et al.49) for multiple comparisons was performed for comparison between multiple groups. Designated sample size always refers to biological replicates (independent animals). GraphPad Prism v8.0 (GraphPad Software), and Sigmaplot v14.0 (Systat Software, Inc.) were used for statistical analysis. For all comparisons, a p value < 0.05 was considered statistically significant.

Results

Bilirubin loads in blood chemistry for the chemically induced C57BL/6j mice model (Fig. 1a) depicts significantly elevated (p < 0.0001) systemic UCB: (8.58 ± 0.37 mg dL−1) and TSB (11.04 ± 1.01 mg dL−1) at postnatal age 10 days compared with the similar aged healthy control littermates (UCB: 0.32 ± 0.02 mg dL−1; TSB: 0.45 ± 0.09 mg dL−1). Single oral nanodrug (0.25 mg kg−1 BW) to the SNH mice at postnatal age 10 days, irreversibly decreased both UCB (0.74 ± 0.19 mg dL−1; t(6.07) = 43.07, p < 0.0001) and TSB (1.69 ± 0.14 mg dL−1; t(4.31) = 28.46, p < 0.0001) to normal range within 12 h (Fig. 1a–c). Logistic regression of the UCB-degradation kinetic data (adj. R2 0.997) revealed the time required for degrading 50% and 80% bilirubin to be 6.07 ± 0.08 h and 7.44 ± 0.17 h, respectively. Similarly for TSB the times required were 6.47 ± 0.18 h and 8.05 ± 0.37 h (adj. R2 0.990). The sequential rates of bilirubin degradation are illustrated in Fig. 1 (insets b and c) that show significant UCB- and TSB-degradation rates consequent to the nanodrug as compared to other control groups during the 4 to 8 h post-administration interval at 1.84 ± 0.26 mg dL−1 h−1 (~20% h−1) and 2.19 ± 0.31 mg dL−1 h−1 (~18% h−1), respectively.

Fig. 1: In vivo bilirubin degradation efficacy of C-Mn3O4 NPs nanodrug in SNH model (postnatal age 10 days).
figure 1

a Tabular representation of UCB and TSB levels before and after nanodrug administration. b Time-dependent UCB-degradation kinetics after a single oral dose of nanodrug (0.25 mg kg−1 BW). Inset shows percent UCB-degradation rate during 4–8 h post administration. c Time-dependent TSB-degradation kinetics after a single oral dose of nanodrug (0.25 mg kg−1 BW). Inset shows percent TSB-degradation rate during 4–8 h post administration. Data are expressed as mean ± SD (n = 6). One-way ANOVA followed by FDR post hoc multiple comparison test was performed for comparison among the groups.

Full size image

Data for the model with kernicterus-like syndrome, induced by repeated and prolonged PHz administration are illustrated in Fig. 2. Here, the “kernicteric” SNH model showed severely elevated bilirubin levels (UCB: 12.3 ± 1.8 mg dL−1, compared to 0.2 ± 0.1 mg dL−1 of control littermates; F (3, 18.3) = 134.2, p < 0.001; TSB: 14.4 ± 2.4 mg dL−1 compared to 0.3 ± 0.1 mg dL−1 of control littermates; F (3, 18.3) = 104.0, p < 0.001) (Fig. 2a). Physical examination revealed signs of central nervous system (CNS) toxicity as indicated by characteristic seizure patterns and inability to maintain balance. Deposition of bilirubin in the brain tissue was measured at 7.4 ± 1.2 nmol g−1 compared to 0.9 ± 0.5 nmol g−1 of control littermates (F (3, 7.7) = 48.6, p < 0.0001) (Fig. 2b). Digital photographs (Fig. 2c) clearly depict the intense yellow pigmentation (of bilirubin) in the brain isolated from SNH model cohort as compared to others. Treatment with the nanodrug in the SNH model resulted in decrease in both systemic and neural bilirubin loads (Fig. 2a–c). These observations were evident in the histopathological studies of isolated brain sections (Fig. 2d). Neural damage as quantified by the number of neural cells per field was noticeably reduced in cerebellum (~35%), Purkinje (~30%), and substantia nigra (~40%) in SNH model. Other pathological features of neurotoxicity in the brain micrographs of SNH model included extensive spongiosis (vacuolation linked to depletion of neurons), shrinkage in cell size, eosinophilic neurons (characteristic of degenerating cells), and gliosis. In contrast, the nanodrug-treated model showed normal cellular architecture with mild gliosis, a residual of SNH-induced brain damage (Fig. 2d). The number of cells per field in both cerebellum and midbrain region was comparable to control animals. No shrinkage of cells was observed. Nanodrug control cohort showed normal brain architecture.

Fig. 2: Effect of the nanodrug on prevention of SNH-associated neurotoxicity.
figure 2

a UCB and TSB levels on postnatal age 14 days. b Brain bilirubin load. c Representative digital micrographs of brains and quantification of yellow color intensity. The intense yellow discoloration of the brain from SNH model cohort in comparison to the other cohorts must be noted. d Micrographs of hematoxylin and eosin (H&E)-stained histopathological sections of different brain regions. The dotted areas indicate the regions of interest where degenerative changes like reduction in neural cells numbers, spongiosis, shrinkage in cell size and eosinophilic neurons took place. The black arrows indicate the Purkinje cells, whose density is greatly reduced in SNH model. In bar plots, data are expressed as mean ± SD (n = 6). Violins depict kernel density estimation of the underlying data distribution. Three lines (from the bottom to the top) in each violin plot show the location of the lower quartile (25th), the median, and the upper quartile (75th), respectively. Individual data points are represented as colored circles or squares (n = 10). One-way analysis of variance (ANOVA) followed by FDR post hoc multiple comparison test was performed for comparison among the groups.

Full size image

Figure 3 illustrates the bilirubin-induced oxidative damages in the cerebellum. In SNH model, unlike the other three cohorts, the cerebellar GSH pool was greatly reduced (7.9 ± 2.4 nmol mg−1 protein, compared to 21.4 ± 3.7 nmol mg−1 protein of healthy control; F (3, 15.1) = 36.3, p < 0.001) (Fig. 3a), and the redox balance was shifted toward oxidative distress (GSH:GSSG ratio: 0.5 ± 0.1, compared to 5.1 ± 0.4 of healthy control; F (3, 12.1) = 100.4, p < 0.001) (Fig. 3b). Consequently, lipid peroxidation was significantly higher (MDA content: 26.0 ± 3.5 nmol g−1 tissue, compared to 10.7 ± 1.6 nmol g−1 tissue of healthy control; F (3, 13.3) = 53.1, p < 0.001) in the SNH model (Fig. 3c). Among the three major antioxidant enzymes that regulate the redox status of cerebellar tissue, activity of CAT and GPx were reduced in the SNH model (Fig. 3d, e). No discernable effect on SOD activity was observed among the groups (Fig. 3f). Brain Mn content determined by the ICP-AES provided direct evidence of the presence of the nanodrug in the brain (2.3 ± 0.5 µg g−1 tissue in nanodrug treated cohort compared to 0.7 ± 0.1 µg g−1 tissue in SNH model; t(5.39) = 7.68, p = 0.0004).

Fig. 3: Effect of the nanodrug on SNH-associated brain (cerebellar) oxidative distress.
figure 3

a Reduced glutathione (GSH) content. b Ratio of reduced and oxidized glutathione (GSH:GSSG). c Lipid peroxidation. d Catalase (CAT) activity. e Glutathione peroxidase (GPx) activity. f Superoxide dismutase (SOD) activity. In bar plots data are expressed as mean ± SD (n = 6). Box plot shows lower quartile (25th), the median, the upper quartile (75th), and range of data points. Individual data points are represented as colored circles (n = 6). One-way analysis of variance (ANOVA) followed by FDR post hoc multiple comparison test was performed for comparison among the groups.

Full size image

Representative movement patterns of the rodents in the OFT are shown in Fig. 4a. Decreased spontaneous activity (i.e., total distance moved; Fig. 4b) and increased anxiety parameters (i.e., time spent in the central zone; Fig. 4c) confirmed “long-term” neuralgic sequelae, i.e., motor function and ataxia in SNH model. Higher time to cross the beam (Fig. 4d) and to descend the pole (Fig. 4e) indicates obvious motor-coordination impairment, a prominent signature of bilirubin neurotoxicity. In the gait analysis the SNH model cohorts showed irregular movement patterns (Fig. 4f) with decreased stride length (Fig. 4g), sway length (Fig. 4h), stance angle (Fig. 4i), and stance length (Fig. 4j). No differences between healthy control and nanodrug treated cohorts were found in any of the tests except time to cross the pole (p = 0.0002). The Hb-levels were within the normal range and similar for both control and experimental models at the time of behavior studies (data not shown) ruling out the possible effect of anemia in the observed motor functions.

Fig. 4: Effect of the nanodrug on SNH-associated neuromotor function.
figure 4

Open field test (OFT): a Movement pattern of the rodents in the open field. b Total distance moved. c Times spent in the central region. Beam Traversal Test: d Time to cross the beam. Pole Test: e Time to descend the pole. Gait Analysis: f General gait pattern of rodents and different parameters used for quality assessment of the movement. g Stride length. h Sway length. i Stance angle. j Stance length. Violins depict kernel density estimation of the underlying data distribution. Three lines (from the bottom to the top) in each violin plot show the location of the lower quartile (25th), the median, and the upper quartile (75th), respectively. Individual data points are represented as white circles (n = 10). One-way analysis of variance (ANOVA) followed by FDR post hoc multiple comparison test was performed for comparison among the groups.

Full size image

Swimming patterns of different cohorts for the MWM as a measure of hippocampus-dependent learning and memory are shown in Fig. 5a and characterized by known adult swimming features of mice. The times taken to escape from the water during four successive trials are listed in Fig. 5b. All animals except the SNH model cohort learned to escape rapidly from water with stable terminal acquisition latencies of <40s (p < 0.0001). The place navigation was further impaired during the platform reversal task (postnatal days 14 and 15). The number of times that SNH model cohort crossed the platform location at postnatal age of 12 days was distinctly different than that of both healthy control and nanodrug-treated group (Fig. 5c). The SNH model cohort spent significantly less time in the target quadrant (Fig. 5d). In the cued platform reversal test (Fig. 5e) all mice except the SNH model cohort showed a preference for the relative position over the absolute position of the platform, which resulted into “resultant heading error” of ~–75° for SNH model and ~–5° for others (p < 0.0001) (Fig. 5f). Retention memory was also impaired in SNH model as they showed a strong preference for neither relative nor absolute platform positioning resulting in the observed scattered directionality across the maze (Fig. 5f). All studied impairments were reversed in the nanodrug-treated cohort.

Fig. 5: “Long-term” effects in learning and memory.
figure 5

a–d Different parameters associated with Morris water maze (MWM) test. The SNH-BIND model showed difficulties in hippocampus-dependent learning. e, f Platform reversal MWM test for assessment of cued learning and retained memory. The SNH-BIND group did not follow the cues or retained memory resulting in greater heading errors. g–j Recognition memory as tested using novel object recognition (NOR) task. The SNH-BIND group was unable to recognize the novel object resulting in lower preference index for the novel object and lower discrimination index. In bar plots, data are expressed as mean ± SD (n = 6). Violins depict kernel density estimation of the underlying data distribution. Three lines (from the bottom to the top) in each violin plot show the location of the lower quartile (25th), the median, and the upper quartile (75th), respectively. Individual data points are represented as colored circles or squares (n = 10). Individual data points are represented as white circles (n = 10). One-way analysis of variance (ANOVA) followed by FDR post hoc multiple comparison test was performed for comparison among the groups.

Full size image

Data for recognition memory assessed by the NOR task (Fig. 5g) are presented in Fig. 5h–j. As expected, the control mice with fully intact recall memory, spent more time with the novel object than the familiar one (Fig. 5h). In contrast, however, SNH model cohort showed no discrimination (p = 0.502) between the novel and familiar objects (Fig. 5h). Thus, the preference index (Fig. 5i) and discrimination ability (Fig. 5j) were both impaired in the SNH model. Treatment with nanodrug reversed these profound cognitive-deficits and the animals behaved like the healthy control cohort (Fig. 5h–j).

Pharmacokinetics and pharmacodynamics are shown in Fig. 6. Time-dependent plasma concentration profiles of the nanodrug after po delivery are presented as plasma Mn-concentration versus time plot. The PK parameters were calculated for a non-compartmental approach and presented in Fig. 6. The mean plasma concentration curve presented three-peak absorption phases at ~1.0, ~6.0, and 12.0 h (Fig. 6a). The deconvoluted spectra show a significant overlap between intestinal absorption and hepato-biliary reabsorption, which resulted in the higher third peak (~12 h). The first peak may be due to the absorption in the upper gastrointestinal tract, as direct delivery of the nanodrug to the stomach by oral gavage resulted in the disappearance of the first peak (Fig. 6c). Urinary Mn levels were below the detectable range. Mn content of feces have maximum concentration (CMAX = 0.68 ± 0.07 µg g−1) at ~18.0 h (Fig. 6b). The effect of the nanodrug on bilirubin degradation correlated with plasma nanodrug-concentration profile (Fig. 6d) with an inverse relationship. Figure 6e describes a dose-dependent correlation (adj. R2 0.995; non-linear logistic-regression model) suggestive of in vivo catalytic degradation of bilirubin by the nanodrug.

Fig. 6: Pharmacokinetics and pharmacodynamics of the nanodrug.
figure 6

a Plasma concentration-time profile following oral administration of the nanodrug. The inset table shows calculated pharmacokinetic parameters. b The elimination profile of C-Mn3O4 NPs through feces. c The plasma concentration-time profile of the nanodrug administered directly to the stomach. The inset table shows calculated pharmacokinetic parameters. d The change in UCB level and plasma nanodrug concentration with time. e Pharmacodynamics of bilirubin degradation. Data are expressed as mean ± SD (n = 4).

Full size image

Discussion

We report a novel chemopreventive nanoceutical agent, C-Mn3O4 NPs (the nanodrug), that has the capacity to selectively degrade bilirubin in vivo and unique properties to reverse acute bilirubin neurotoxicity. We provide direct evidence for a single oral dose of the nanodrug to degrade UCB, both systemic and neural. Significant (p < 0.0001) UCB and TSB-degradation rates were reported within 4–8 h at 1.84 ± 0.26 mg dL−1 h−1 (~20% h−1) and 2.19 ± 0.31 mg dL−1 h−1 (~18% h−1) as compared to control cohorts, respectively. In the long-term SNH with kernicterus-like syndrome model, brain bilirubin load was decreased by 5.6 nmol g−1 (p = 0.0002) along with improved neurobehavior, neuromotor movements as well as hippocampus-dependent recall memory and learning. Histopathological studies indicated that the nanodrug prevented neural cell injury in cerebellar, Purkinje, and substantia nigra region. In addition, the study drug ameliorated spongiosis, cell shrinkage, and eosinophilic neurons as observed in SNH brain parenchyma. Bilirubin-mediated oxidative damages in the brain as indicated by high GSH content, increased lipid peroxidation and low GSH:GSSH ratio, and decreased catalytic activities of antioxidant enzymes (i.e., CAT, GPx) in SNH model were reversed in nanodrug-treated cohort. The PK-PD studies indicate a therapeutic window that is almost immediate and sustainable. Within 24 h, the nanodrug is completely excreted through the fecal gastrointestinal route.

Our current data corroborate previous in vitro studies23,40,50,51 that elucidated the direct destruction of bilirubin by the nanodrug to methyl-vinyl-maleimide. UV-visible spectroscopy, time-correlated single-photon counting, and electrochemical studies revealed the complex mechanism of bilirubin breakdown involving both surface catalysis (Langmuir–Hinshelwood type) and transient free radical-mediated charge transfer.51 The origin of these catalytic mechanisms lies in the unique electronic structure of the nanodrug having closely packed tetrahedral and octahedral sites within a face-centered cubic oxygen framework. The valence band of the structure is dominated by tetrahedral Mn2+, while the octahedral Mn-ions are in a Jahn-Tellar distorted, high-spin Mn3+ state.23,48,52 The high-energy tetrahedral Mn2+ spontaneously oxidizes to Mn3+, followed by the removal of electron from octahedral site, forming Mn4+. When three electrons per formula unit are removed, the valence band becomes dominated by oxygen states and electrons are extracted from O2p orbitals instead of tetrahedral Mn3+, resulting in the formation of oxygen holes on the nanodrug surface.52 The holes degrade bilirubin by direct surface catalysis or creating transient free radicals upon interaction with dissolved molecular oxygen. While the reduction of bilirubin in the systemic circulation is clearly the result of direct interactions as described above (pharmacodynamics also shows a correlation between plasma nanodrug concentration and bilirubin degradation rate), the reduction of the brain bilirubin content may be both the consequence of the reduced blood content, as in phototherapy, as well as to a direct effect (degradation of bilirubin) operated by the nanodrugs that entered in the CNS. Our biodistribution study using ICP-AES further supports this argument by confirming the presence of nanodrug in the brain tissues. The amelioration of the brain oxidative damages could be a causal effect between reduction of brain bilirubin content and redox modulatory nature of the nanodrug. However, additional inquiries are needed to understand the complete molecular mechanism, which is beyond the scope of this article. It is worth mentioning here that, the free radicals generated during the breakdown process are transient and in absence of the inducer (i.e., bilirubin) are readily adsorbed by the nanodrug to replenish its initial electronic configuration.39 Due to such redox buffering nature, the nanodrug showed neither cellular-toxicity nor damage to any bio-macromolecule.24,25,53 A comprehensive repeated-dose sub-chronic (90 days) toxicity study39 according to OECD guideline-40854 further illustrated the biocompatibility and non-toxicity of this nanodrug in adult rodents.

At the outset, our data build the evidence of known properties of this manganese-based class of nanoparticles55,56,57 as potential therapeutic agents. Our new results appear to confirm their crucial role in in vivo redox-mediated catalysis. We have demonstrated that these NPs do not possess discernible neurotoxic effects but appear to be neuroprotective. These findings are consistent with the previously reported amelioration of adult neuralgic conditions like Mn-induced Parkinson’s syndrome26 and Huntington syndrome.25

Although the data are premature for immediate clinical trials, preparation of a comprehensive drug-safety-dossier including studies that would meet FDA endorsement could help overcome this problem. Prior commercial experiences to meet FDA approval have illustrated the limitations of establishing an effective collaboration of academia, industry, public health, regulatory and societal partners. Drug development strategies based on efficacy, safety, and affordability of an intervention are governed by its delivery as a wide-spread preventive, limited therapeutic, or a rare (orphan) application. Our preliminary report of this bilirubin degrading agent is a novel candidate for such a prospective inquiry.

The PHz-induced mice model is a widely used animal model for testing the efficacy of drugs against SNH but differs from human neonates who manifest increased bilirubin production and/or genetic mutation of bilirubin glucuronidation-related elimination. Our approach conducted the rodent studies at postnatal ages less than 14 days to avoid the developmental interference of bilirubin glucuronidation.6 The observed yellow discoloration of brain, and histopathological damages to the midbrain, cerebellum, especially Purkinje cells, and the hippocampus region are well-known hallmarks of kernicterus and its sequelae.58,59 Thus, a potential limitation is whether a chemically induced rodent model of SNH could be improved by future validation studies in naturally occurring animal models of unconjugated hyperbilirubinemia that manifest hemolytic, elimination, or combined disorders of bilirubin loads. Together with increased bilirubin, the experimental SNH model used in this study also presents an increased Fe and CO production. This may better represent a neonate with Rh incompatibility or G6PD deficiency.60 Thus, replicating the experiments with long-term follow-up studies in a model of spontaneous and “pure” hyperbilirubinemia (e.g., the Gunn rat) will be important to confirm the efficacy and safety before suggesting the treatment for clinical trials. In this regard, assessment of inflammation and associated immunogenic parameters would be crucial to evaluate long-term effects of the treatment. We deferred the question about possible interaction with phototherapy since most neonates with hyperbilirubinemia are treated with photon exposure. Whether this nanodrug could augment use or substitute the use phototherapy to reverse acute bilirubin neurotoxicity needs to be resolved prospectively. The road map to ascertain potential clinical application of any chemo-preventive agent has been well charted by series of laboratory and clinical studies undertaken for seeking FDA approval of heme-oxygenase inhibitors. We propose a similar deliberative investigational strategy for a drug that could prevent life-long kernicteric disabilities.

In conclusion, we introduce a novel chemo-preventive nanoceutical agent that selectively degrades bilirubin in vivo and possesses unique properties to reverse acute bilirubin neurotoxicity. We provide direct evidence for a single oral dose of functionalized Mn3O4 nanoparticles (C-Mn3O4 NPs, the nanodrug) that is therapeutically effective in a chemically induced unconjugated SNH rodent model. Future clinical therapeutic options for newborns at risk for incipient bilirubin neurotoxicity merit comprehensive inquiry and ascertainment to define public health benefits and individual risks.