Abstract

   Chronic kidney disease (CKD) ultimately causes renal fibrosis and
   end-stage renal disease, thus seriously threatens human health.
   However, current medications for CKD and fibrosis are inefficient,
   which is often due to poor targeting capability to renal tubule. In
   this study, we discover that biomimetic high-density lipoprotein (bHDL)
   lipid nanoparticles possess excellent targeting ability to injured
   tubular epithelial cells by kidney injury molecule-1(KIM-1) mediated
   internalization. Thus, we co-load anti-inflammatory drug triptolide
   (TP) and anti-fibrotic drug nintedanib (BIBF) on bHDL nanoparticles to
   treat CKD. Based on the targeted delivery and mutual enhancement of the
   efficacy of co-delivered drugs, the bHDL-based system effectively
   reduces kidney injury and alleviates renal fibrosis in different CKD
   mouse models. The mechanistic study shows that BIBF and TP
   synergistically remodel the fibrotic niches by decreasing inflammatory
   cytokines, limiting immune cell infiltration and inhibiting the
   activation of myofibroblasts. The bHDL vehicle also possesses high
   manufacturability, good safety and adequately reduces the toxicity of
   TP. Thus, this system is promising for the treatment of CKD and bHDL
   has good potential for delivering agents to damaged renal tubular
   epithelial cells.

   Subject terms: End-stage renal disease, Drug delivery, Endocytosis
     __________________________________________________________________

   Effectively delivering medications to the renal tubule to delay or halt
   chronic kidney disease progression remains a significant unmet clinical
   challenge. Here, authors introduce an innovative strategy for renal
   tubule targeting using biomimetic high-density lipoprotein (bHDL)
   nanoparticles.

Introduction

   Chronic kidney disease (CKD) is a common syndrome characterized as
   chronic structural and functional dysfunction of the
   kidney^[48]1–[49]3, which can be triggered by various factors,
   including hypertension, diabetes, injury and complex genetic factors.
   CKD often persists for months leading to permanent renal damage, which
   finally develop to renal fibrosis and end-stage renal disease that
   requires dialysis or even kidney transplantation for the
   patients^[50]2,[51]4. As CKD advances, a range of fatal complications
   may occur, such as cardiovascular disorders, metabolic acidosis and
   mineral and bone disorder^[52]5–[53]7. CKD’s rising incidence has
   rendered it a crucial public health issue, posing a significant threat
   to human well-being, as well as a substantial financial
   burden^[54]8,[55]9. Unfortunately, no effective treatments are
   currently available for patients with CKD, especially for renal
   fibrosis^[56]3. Hence, there are unmet clinical needs for new
   interventions that can effectively halt the progression of CKD and
   alleviate renal fibrosis.

   Renal tubular epithelial cells (RTECs) play a vital role in the
   progression of CKD and renal interstitial fibrosis^[57]10. During
   kidney injury, these cells are not merely the victim of injury, but
   they also play a vital role in aggravating it^[58]11, indicating their
   potential as promising target cells for CKD targeted therapy. However,
   efficient and safe delivery of therapeutic drugs to RTECs is
   challenging due to the unique physiological and structural
   characteristics of the kidney, including the size barrier and charge
   barrier of the glomerular filtration apparatus comprising of
   fenestrated endothelium, glomerular basement membrane and
   podocytes^[59]12,[60]13. Various strategies have been tested to achieve
   drug delivery to RTECs such as controlling the charge and particle size
   of nanoparticles^[61]14 and applying targeting ligands^[62]15,[63]16.
   Even though these strategies effectively enhance drug accumulation in
   kidney, the efficiency of these systems is usually compromised by the
   hepatic distribution partly resulting from the formation of protein
   corona^[64]17.

   Kidney injury molecule-1 (KIM-1), a type I transmembrane glycoprotein
   which is highly expressed on the injured proximal TEC, is being
   investigated as a potential renal-targeted delivery
   target^[65]18,[66]19. Initially, KIM-1 was thought to be an early
   marker of renal tubular injury in the proximal area. In recent years,
   research has shown that it can serve as a receptor and mediates the
   uptake of a wide array of substances^[67]20–[68]22. KIM-1 is able to
   serve as a scavenger receptor to clear apoptotic cells to protect
   against acute kidney injury(AKI)^[69]21,[70]23. Therefore, the
   selective expression and functional specificity render KIM-1 an
   exemplary target for renal tubule-specific drug delivery.

   Biomimetic high-density lipoprotein (bHDL) nanoparticles are simple
   delivery system comprised of apolipoproteins and phospholipids^[71]24.
   bHDL nanoparticles exhibit ultra-small particle sizes, a key
   characteristic enabling them to traverse the glomerular filtration
   barrier effectively, enhancing their chances to arrive at RTECs.
   Furthermore, bHDL nanoparticles retain the inherent structure of
   natural HDL, granting them prolonged circulation time with slow
   clearance by reticuloendothelial system. Besides, bHDL nanoparticles
   possess the capability of naturally targeting to scavenger
   receptors^[72]25,[73]26. As previously mentioned, KIM-1 operates as a
   scavenger receptor, binding to phosphatidylserine (PS), oxidized LDL
   (ox-LDL), as well as natural LDL and fatty acids. Ox-LDL is a typical
   scavenger receptor ligand and binds to a variety of scavenger
   receptors, including LOX-1, scavenger receptor class A, CD36 and
   scavenger receptor class B type I (SR-BI)^[74]27. Notably, natural LDL
   does not bind to class A, C, and D scavengers and is a specific ligand
   for SR-BI. SR-BI can also effectively bind PS and fatty acids,
   indicating that KIM-1 possesses more functional similarities with SR-BI
   than others scavenger receptor. HDL is a type of classical receptor of
   SR-BI, implying a possible link between KIM-1 and HDL. However, the
   potential interaction between KIM-1 and HDL remains uncertain. Given
   the facts, it is crucial to explore whether KIM-1 can augment the
   uptake of b-HDL nanoparticles by damaged RTECs, which may imply the
   promising application of bHDL nano-delivery system for precise drug
   delivery within the renal milieu.

   Current treatments for CKD and fibrosis primarily aim to suppress
   typical symptoms and usually target single signaling pathway. However,
   therapeutic outcomes of these treatments are unsatisfactory, which is
   likely due to the intricate nature of CKD and the presence of fibrotic
   niches where inflammation and fibrosis invariably coexist, as well as a
   complex interplay between multiple cells^[75]28. Simultaneous
   regulation of inflammation and fibrosis may be an effective therapeutic
   approach. For this purpose, we proposed that triptolide (TP) and
   nintedanib (BIBF) may be a good combination that could remodel the
   fibrotic niche. Here, TP is the main active ingredient obtained from
   the root bark of Tripterygium wilfordii., which is widely used in the
   treatment of renal diseases in China, showing impressive
   anti-inflammatory and immunosuppressive activities^[76]29. However, its
   clinical use is severely restricted by its low solubility/permeability,
   poor bioavailability, and potential off-target toxicity^[77]30. BIBF is
   a small-molecule tyrosine kinase inhibitor that has been approved for
   the treatment of idiopathic pulmonary fibrosis^[78]31. It impedes the
   proliferation, migration, and transformation of fibroblasts, thereby
   exhibiting a potent antifibrotic effect^[79]32. Despite its efficacy,
   the absolute bioavailability of its oral formulation is only
   4.69%^[80]33. Therefore, we aimed to develop a targeted approach based
   on bHDL nanoparticles that can improve their bioavailability and
   co-deliver TP and BIBF preferentially to the disease site to modify the
   renal malfunctions, while mitigating unintended drug distribution and
   reducing systemic toxicity.

   In this study, we constructed bHDL lipid nanoparticles co-loaded with
   TP and BIBF for the treatment of CKD and explored the renal targeting
   ability of bHDL nanoparticles as well as the targeting mechanism. As
   expected, bHDL exhibited pronounced renal targeting, particularly
   accumulating in impaired RTECs. Further mechanistic studies highlighted
   the pivotal role of KIM-1 in the endocytic uptake of bHDL
   nanoparticles. Meanwhile, the simultaneous delivery of
   anti-inflammatory and antifibrotic drugs to the kidneys via bHDL
   nanoparticles can effectively reduce renal injury and delay fibrosis in
   multiple CKD mouse models.

Results

KIM-1 is markedly upregulated after kidney injury

   Several studies have shown that KIM-1 is not expressed in normal kidney
   tissues, and the expression level rise significantly after kidney
   damage in AKI and diabetic kidney disease (DKD) mouse models, which is
   positively correlated with the degree of kidney
   injury^[81]20,[82]34,[83]35. To explore the KIM-1 expression level in
   CKD, three CKD mouse models with different etiologies, unilateral
   ureteral obstruction (UUO), folic acid-induced and adenine-induced
   mouse model, were constructed and the KIM-1 expression was determined.
   KIM-1 was highly expressed in all kinds of model mice as compared to
   the control mice (Fig. [84]1a–c). Co-staining KIM-1 with lotus
   tetragonolobus lectin (LTL), a marker of proximal tubule, demonstrated
   that KIM-1 is mainly expressed in the proximal tubules (Fig. [85]1a).
   Experimental results based on western blotting as well as qPCR also
   proved that KIM-1 is significantly raised after injury from the protein
   and transcriptional level (Fig. [86]1b–d). Additionally, KIM-1
   expression level on cisplatin-induced RTECs was also assessed and the
   protein level of KIM-1 was dramatically elevated in the HK-2 cells
   after treating with cisplatin for 24 h (Supplementary Fig. [87]6c),
   which is in consistent with previous study^[88]36. In conclusion, KIM-1
   is highly expressed on injured RTECs and could be a specific delivery
   target for injured RTEC-targeted renal therapy.

Fig. 1. KIM-1 is markedly upregulated in UUO model, folic acid model and
adenine model mice.

   [89]Fig. 1
   [90]Open in a new tab

   a Representative immunofluorescence images(left) and the Pearson
   correlation coefficient(right) of colocalization of KIM-1(red) and LTL
   (green) in the kidney of control, UUO model, folic acid model and
   adenine model mice. Scale bars, 100 µm. Data are mean ± SD (n = 3
   independent samples). b Western blot analysis of KIM-1 protein
   expression in the kidney of control, UUO model, folic acid model and
   adenine model mice. The value of 60 and 55 kDa are referred to the
   predicted protein weight of KIM-1 and α-tubulin. α-tubulin was used as
   internal control. n = 3 biologically independent samples. c
   Semi-quantification of KIM-1 protein expression by Image J based the
   western blots shown in b and it was normalized with corresponding
   α-tubulin signal. Data are mean ± SD (n = 3 independent samples,
   two-tailed unpaired t-test). d Real-time qPCR analysis of Havcr1 mRNA
   expression level in kidney of control, UUO model, folic acid model and
   adenine model mice. Data are displayed as normalized fold expressions
   relative to control group, and Acta mRNA was used as internal control.
   Data are mean ± SD (n = 4 independent samples, two-tailed unpaired
   t-test). Control: untreated healthy mice, UUO model: unilateral
   ureteral obstruction model mice, folic acid model: mice treated with
   folic acid to induce kidney injury, adenine model: mice treated with
   adenine to induce kidney injury, LTL lotus tetragonolobus lectin.
   Source data are provided as a Source Data file.

Fabrication and characterization of the kidney targeted nanoparticles

   KIM-1, markedly expressed on impaired RTECs, presents a potential
   target for renal-specific drug delivery^[91]37,[92]38. Considering the
   features of bHDL nanocarrier, we postulated that it holds promise as a
   means of targeting KIM-1. To explore the KIM-1 targeting of bHDL, we
   firstly fabricated blank bHDL lipid nanoparticles, with the
   phospholipid 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) and
   D-4F, a widely used apoA1 mimic peptide. Meanwhile, we co-loaded the
   anti-inflammatory drug TP and the antifibrotic drug BIBF into bHDL to
   obtain co-loaded nanoparticles TP/BIBF-bHDL for CKD treatment.
   Morphological analysis using transmission electron microscopy (TEM) and
   dynamic light scattering (DLS) revealed that blank bHDL possessed a
   predominantly discoidal structure with an average diameter of 10.9 nm
   (Fig. [93]2a, Supplementary Fig. [94]1). Upon drug loading,
   TP/BIBF-bHDL became less discoidal with a slight increase in diameter.
   Cryo-EM images showed that TP/BIBF-bHDL had higher contrast than blank
   bHDL, and displayed particles with both spherical and discoidal shapes
   (Fig. [95]2b). Zeta potential results showed that both blank and
   drug-loaded bHDL nanoparticles exhibited positive charge (Fig. [96]2c),
   which could facilitate their renal accumulation^[97]39. The positive
   charge of the nanoparticles may be attributed to the D-4F peptide,
   which contains multiple positively charged lysine residues. Upon
   intravenous injection, the dense positive charge on the nanoparticles
   may damage red blood cells, leading to potential toxicity. Therefore,
   the hemolytic activity of TP/BIBF-bHDL was investigated to assess its
   biocompatibility. Result showed that TP/BIBF-bHDL did not cause
   significant hemolysis with a hemolysis rate lower than 5%
   (Supplementary Fig. [98]2), confirming its safety.

Fig. 2. Fabrication and characterization of TP/BIBF-bHDL.

   [99]Fig. 2
   [100]Open in a new tab

   a TEM images of bHDL and TP/BIBF-bHDL. Scale bars, 50 nm. b Cryo-EM
   images of bHDL and TP/BIBF-bHDL. Scale bars, 25 nm. c Size and zeta
   potential of bHDL, TP/BIBF-bHDL, TP-bHDL and BIBF-bHDL measured by DLS.
   Data are mean ± SD (n = 3 independent samples). d Changes in particle
   size of TP/BIBF-bHDL under 37 °C and 4 °C. Data are mean ± SD (n = 3
   independent samples). e TEM images of TP/BIBF-bHDL and TP/BIBF-bHDL
   after incubation with 10%FBS or 10% lipoprotein-depleted FBS at 37 °C
   for 2 hours. Scale bars, 50 nm. n = 3 independent experiments. f Size
   of TP/BIBF-bHDL displayed in e and measured by image J. Data are
   mean ± SD (n = 20). g Quantification of serum protein absorbed on the
   bHDL nanoparticles and liposomes by BCA assay. Data are mean ± SD
   (n = 3 independent samples, one-way ANOVA with Dunnett’s multiple
   comparisons test and adjustment applied). h In vitro release profiles
   of TP/BIBF-bHDL in the presence of PBS and FBS. Data are mean ± SD
   (n = 3 independent samples). bHDL blank biomimetic high-density
   lipoprotein (HDL) nanoparticles, TP/BIBF-bHDL triptolide and nintedanib
   co-loaded biomimetic HDL nanoparticles, FBS fetal bovine serum. Source
   data are provided as a Source Data file.

   Furthermore, we investigated the stability of the drug-laden bHDL
   nanoparticles, which exhibited no significant alterations in the
   particle size when they were suspended in phosphate buffer saline
   (PBS) at 37 °C for 12 h and 4 °C for 5 days (Fig. [101]2d). The
   stability of the bHDL nanoparticles in serum was also performed by
   monitoring changes in particle size and serum protein absorption. Given
   the presence of endogenous lipoproteins in serum, both normal fetal
   bovine serum (FBS) and lipoprotein-deproteinized FBS were used. After a
   2-hour incubation at 37 °C with 10% FBS and 10%
   lipoprotein-deproteinized FBS, the TP/BIBF-bHDL exhibited no
   aggregation with only a slight increase in particle size of
   approximately 4 nm (Fig. [102]2e, f), demonstrating excellent stability
   in serum. The serum protein level absorbed on bHDL nanoparticles was
   estimated by Bicinchoninic acid (BCA) assay, with liposomes included as
   a control. The bHDL nanoparticles absorbed only ~25% proteins compared
   to liposomes (Fig. [103]2g). Interestingly, bHDL nanoparticle absorbed
   ~60% less protein when incubated in the lipoprotein-deproteinized FBS,
   indicating that lipoproteins are the primary proteins to be absorbed.
   Drug release profiles of TP/BIBF-bHDL in both PBS and FBS exhibited
   comparable trends, affirming the robust stability of TP/BIBF-bHDL in
   FBS (Fig. [104]2h). Additionally, we quantified the encapsulation
   efficiency (EE%) and drug loading capacity (LC%), with values of 97.7%
   and 6.01% for BIBF and 84.78% and 1.3% for TP, respectively
   (Table [105]S1).

bHDL nanoparticles present excellent capability of targeting injured RTECs

   To investigate the renal targeting efficacy of bHDL nanoparticles, we
   prepared DiD-labelled bHDL (DiD-bHDL) to assess the biodistribution in
   UUO mice, a classical mouse model of renal fibrosis. DiD-liposomes with
   a particle size of 80 nm were prepared (Supplementary Fig. [106]3),
   with the same lipid composition as bHDL but lacking D-4F. Following
   intravenous injection of DiD, DiD-liposomes and DiD-bHDL, vital organs
   were harvested at predetermined time points and placed in an in vivo
   imager to visualize the distribution. As shown in Fig. [107]3a-b,
   DiD-bHDL were rapidly taken up by the kidney, liver and lung after
   injection, and the renal localization of DiD-bHDL is higer than free
   DiD and DiD-liposomes. However, it should be noted that free DiD dye
   may form micelles, which could hinder its ability to pass through the
   kidney filter, thereby reducing its distribution in the kidney. The
   accumulation of bHDL in the kidney displayed a time-dependent increase,
   peaking at 2 hours post-injection. After 24 hours, substantial
   fluorescence was still visible in the injured kidney in the bHDL group.
   The fluorescence intensity was also semi-quantitatively analyzed
   (Fig. [108]3c–e, Supplementary Fig. [109]4a–c). The liver serves as the
   primary site for the uptake and elimination of nanoparticles, with both
   bHDL and liposomes distributing to this organ. It’s worth noting that
   the distribution of bHDL in the liver is lower than that of normal
   liposomes, which are composed of the same phospholipid DMPC,
   cholesterol and lack the D-4F component. Furthermore, we compared the
   biodistribution of bHDL in UUO model mice and healthy mice. The bHDL
   nanoparticles showed greater accumulation in the ligation kidneys,
   suggesting that bHDL can specifically target the damaged kidneys
   (Fig. [110]3b and d, Supplementary [111]4g). The localization of bHDL
   within the injured kidneys was further elucidated through
   immunofluorescence, revealing a preferential accumulation of bHDL in
   the damaged RTECs (Fig. [112]3f).

Fig. 3. bHDL nanoparticles present excellent targeting ability to injured
RTECs.

   [113]Fig. 3
   [114]Open in a new tab

   a Bioluminescence images of major organs in UUO mice after intravenous
   administration of DiD, DiD-liposome or DiD-bHDL at 0.5, 2, 8 and
   24 hours. b Bioluminescence images of kidneys in UUO and healthy mice
   after DiD-bHDL administration at various time points. c Fluorescence
   intensity in organs of UUO mice treated with DiD, DiD-liposome or
   DiD-bHDL at 2 hours. d Fluorescence intensity in kidneys of healthy and
   UUO mice treated with DiD-bHDL. e Fluorescence intensity in the injured
   kidney, liver and lung of UUO mice treated with DiD-bHDL. Data in b–d
   are mean ± SD (n = 3 biologically independent samples for a–d, two-way
   ANOVA with Sidak’s multiple comparisons test and adjustment applied). f
   Immunofluorescence images of DiD-bHDL localization in kidneys of UUO
   mice at 2- and 24-hour post-treatment. LTL (green) marks the renal
   tubules. Scale bars, 50 µm. n = 3 independent samples. g BIBF
   concentration in major organs of UUO mice 2 hours after intravenous
   administration of TP/BIBF or TP/BIBF-bHDL. h BIBF concentration in
   normal and ligated kidneys of UUO mice treated with TP/BIBF-bHDL at
   different time points. i Mean concentration-time curves of BIBF in
   kidneys of UUO mice after intravenous administration of TP/BIBF or
   TP/BIBF -bHDL. Data in g–i are mean ± SD (n = 5 biologically
   independent samples, two-way ANOVA with Sidak’s multiple comparisons
   test). j Uptake of DiD-bHDL in normal and cisplatin-injured HK-2 cells
   for 2 hours. k Uptake of DiD-bHDL in normal and TGF-β1-stimulated
   NRK-49F cells for 2 hours. Data in j–k are mean ± SD (n = 3 independent
   samples, two-tailed unpaired t-test). l Mean concentration-time curves
   of BIBF in the plasma of rats treated with TP/BIBF or TP/BIBF-bHDL.
   Data are mean ± SD (n = 6 biologically independent samples, two-way
   ANOVA with Sidak’s multiple comparisons test and adjustment applied).
   DiD free DiD, DiD-liposome DiD-laden liposomes, DiD-bHDL DiD-loaded
   biomimetic high-density lipoprotein nanoparticles, UUO mice unilateral
   ureteral obstruction model mice, TP/BIBF free triptolide and nintedanib
   solution, TP/BIBF-bHDL triptolide and nintedanib co-loaded biomimetic
   high-density lipoprotein nanoparticles. Source data are provided as a
   Source Data file.

   To more precisely delineate the systemic distribution of the
   therapeutic agents, we further investigated the targeting ability of
   bHDL to the damaged kidney by quantifying the concentration of BIBF in
   tissues. The distribution of the drug in each organ after tail vein
   injection of TP/BIBF or TP/BIBF-bHDL is shown in Fig. [115]3g and
   Supplementary Fig. [116]4d–f. These findings align with the results
   obtained from fluorescence imaging. Encapsulation of the drug in bHDL
   markedly increased its distribution in the kidneys, especially in the
   damaged kidneys (Fig. [117]3h), while concurrently reducing its
   distribution in the spleen. Analysis of the time profile of drug
   concentration in the kidney showed that the renal drug concentration in
   the TP/BIBF-bHDL group was significantly higher than that in the free
   drug group at all time points after drug administration (Fig. [118]3i),
   and the area under curve (AUC[0-∞]) in the kidney of the TP/BIBF-bHDL
   group was 72.85 ± 5.70 µg ml^−1h, approximately double that of the free
   drug group (Table [119]S3).

   Additionally, we conducted in vitro cellular uptake experiments using
   HK-2 cells, with and without cisplatin pretreatment, to elucidate the
   capacity of bHDL to target injured RTECs. Our findings revealed that
   damaged HK-2 cells exhibited a higher uptake of bHDL compared to their
   normal counterparts (Fig. [120]3j). Concurrently, we assessed the
   uptake of bHDL by NRK-49F cells, a type of renal fibroblast cell, which
   revealed notably lower uptake relative to HK-2 cells, as shown in
   Supplementary Fig. [121]4h. Moreover, stimulation of fibroblast cells
   with TGF-β1 did not result in an increased uptake of bHDL
   (Fig. [122]3k). TGF-β1 could induce the fibroblast cells transition to
   myofibroblast cells which have decreased endocytosis ability due to the
   increased contractility and membrane tension^[123]40. Collectively,
   these results provide compelling evidence supporting the enhanced
   targeting capability of bHDL toward injured RETCs.

   In addition, we performed pharmacokinetic studies to elucidate the in
   vivo behaviors of bHDL nanoparticles. After the administration of
   TP/BIBF-bHDL and TP/BIBF through tail vein injection and gavage
   administration of TP/BIBF, the concentration of BIBF in plasma at
   various time points was measured by LC-MS/MS. The key pharmacokinetic
   parameters were calculated using DAS statistical software and displayed
   in Table [124]S2. As shown in Fig. [125]3l, BIBF exhibited low oral
   bioavailability, due to its low solubility and poor absorption in the
   intestinal environment. At all sampling time points, the blood drug
   concentration in the TP/BIBF-bHDL group was significantly higher than
   that in the TP/BIBF group, with 80% increase in AUC. Meanwhile, the
   half-life time (t[1/2]) of TP/BIBF-bHDL is also 30% higher. These
   results indicate that encapsulation of drugs in bHDL significantly
   improved drug bioavailability, reduced drug clearance and prolonged its
   half-life.

KIM-1 mediates bHDL nanoparticles uptake by injured RTECs

   We further explored the cellular uptake mechanism of bHDL
   nanoparticles. To clarify the mechanism, we first investigated the
   endocytic pathways involved in the uptake of bHDL nanoparticles using
   various endocytosis inhibitors. As shown in Fig. [126]4a, the uptake of
   bHDL was significantly impeded by low temperature and cytochalasin D
   (an actin-disrupting agent), as well as O-Phospho-L-serine and BLT-1
   (an inhibitor of SR-BI). The cubilin antibody slightly reduces the
   uptake of bHDL, but this reduction is not statistically significant,
   suggesting that the cubilin receptor is not the primary endocytic
   pathway for bHDL nanoparticles. Conversely, inhibitors like
   chlorpromazine (clathrin-mediated endocytosis inhibitor), amiloride
   (macropinocytosis inhibitor), cyclodextrins (lipid raft inhibitor), and
   gentamicin (megalin-mediated endocytosis inhibitor) did not affect bHDL
   uptake. These findings indicate that the uptake of bHDL is mainly via
   phagocytosis, phosphatidylserine receptor and SR-BI-mediated
   endocytosis. Moreover, pre-incubation of cells with D-4F significantly
   reduced the uptake of bHDL, highlighting the critical role of D-4F in
   RTECs uptake processes (Fig. [127]4b).

Fig. 4. KIM-1 mediates bHDL uptake by injured RTECs.

   [128]Fig. 4
   [129]Open in a new tab

   a Uptake of bHDL under 4°C and with different endocytic inhibitors.
   Data are mean ± SD (n = 3 independent samples, two-tailed unpaired
   t-test). b Effect of D-4F on bHDL uptake in HK-2 cell. Data are
   mean ± SD (n = 3 independent samples, one-way ANOVA with Dunnett’s
   multiple comparisons test and adjustment applied). c Uptake of DiD-bHDL
   in normal and cisplatin-injured HK-2 cells for 2 hours, with or without
   pretreatment using KIM-1 antibody (30 µg/ml, 0.5 hours). Data are
   mean ± SD (n = 3 independent samples, two-tailed unpaired t-test). The
   control refers to untreated cells; KIM-1 antibody pretreated refers to
   cells with antibody pretreatment. d Representative confocal images of
   cisplatin-injured HK-2 cells after 2 hours of exposure. The nuclei were
   stained DAPI (blue), KIM-1 was labeled in green and DiD-bHDL was
   visualized in red. Scale bars, 20 µm. e Representative
   immunofluorescence images of colocalization of KIM-1(green) with
   DiD-bHDL or DiD-liposome (red) in the kidney. Nuclei were stained DAPI
   (blue). Scale bars, 50 µm. f Quantification of the colocalization of
   KIM-1 and DiD-bHDL or DiD-liposome by Pearson’s correlation
   coefficient. The Pearson’s correlation coefficient was measured by
   Colocalization Finder with image J. Data are mean ± SD (n = 3
   biologically independent samples, two-tailed unpaired t-test). g, h
   Uptake of DiD-bHDL in HK-2 cells with KIM-1 protein knockdown(g) or
   overexpression(h) for 2 hours after 24-hour cisplatin
   treatment(2 µg/ml). Data are mean ± SD (n = 3 biologically independent
   samples, two-tailed unpaired t-test). i Western blot analysis of SR-BI
   expression in normal and cisplatin-injured HK-2 cells treated with
   cisplatin(2 µg/ml) for 24 hours. n = 3 independent samples. j
   Semi-quantification of SR-BI protein level (i) using image J, with
   β-actin as the internal control. Predicted protein weights: β-actin
   (43 kDa) and SR-BI (74 kDa). Data are mean ± SD (n = 3 biologically
   independent samples, two-tailed unpaired t-test). DiD-liposome
   DiD-laden liposomes, DiD-bHDL DiD-loaded biomimetic high-density
   lipoprotein nanoparticles, control untreated cells, CDDP-induced
   cisplatin-induced injured cells, KIM-1^-/- HK-2 HK-2 cells with HAVCR1
   knocked down, OE-KIM-1 HK-2 HK-2 cells with HAVCR1 overexpression.
   Source data are listed in the Source Data file.

   Injury resulted in a noteworthy upregulation of KIM-1 protein on RTECs
   (Supplementary Fig. [130]6c), accompanied by a substantial augmentation
   in their uptake of bHDL (Fig. [131]3j). The enhanced cellular uptake of
   bHDL in injured HK-2 cells may be attributed to the interaction between
   bHDL and KIM-1. To substantiate this hypothesis, a competitive
   inhibition experiment was firstly performed to examine the effect of
   KIM-1 antibody on bHDL uptake. HK-2 cells and cisplatin-induced injured
   HK-2 cells were incubated with DiD-bHDL for 2 hours, with or without
   pretreatment using the KIM-1 antibody. As shown in the Fig. [132]4c,
   the KIM-1 antibody effectively inhibited the uptake of bHDL by 30% in
   the damaged cells, while paradoxically augmenting its uptake in normal
   HK-2 cells. The reason for this is that normal HK-2 cells express less
   KIM-1, and the addition of antibody stimulated the damaged cells to
   express KIM-1 to a certain extent and increase their uptake. Whereas in
   cisplatin-induced injured HK-2 cells with high KIM-1 expression, bHDL
   could enter the injured HK-2 cells through KIM-1-mediated endocytosis,
   which was inhibited by the addition of antibody. Then confocal laser
   scanning microscopy was applied to explore whether the uptake of bHDL
   was mediated by KIM-1, and the colocalization of KIM-1 and bHDL in
   cisplatin-induced injured HK-2 cells were observed with a Pearson’s
   correlation coefficient of 0.9, indicating there is an interaction
   between KIM-1 and bHDL (Fig. [133]4d). The colocalization between KIM-1
   and bHDL was also evaluated in vivo. Kidneys from UUO mice were
   harvested 2 hours following the intravenous administration of DiD-bHDL.
   Immunofluorescence staining was subsequently performed to observe the
   co-localization of DiD-bHDL with KIM-1. A DiD-liposome group was added
   as a comparison. As shown in Fig. [134]4e and f, DiD-bHDL co-localized
   well with KIM-1 in kidney tissue with a Pearson’s correlation
   coefficient of 0.94, indicating the direct association of bHDL and
   KIM-1. In contrast, a modest co-localization of KIM-1 with liposomes
   was observed, as evidenced by a Pearson’s correlation coefficient of
   0.64. The reason for this may be that expression of KIM-1 confers a
   phagocytic phenotype on RETCs, leading to their uptake by liposomes.
   All these results reveal that bHDL enters cells through endocytosis
   mediated by KIM-1.

   To further characterize the role of KIM-1 in the endocytosis of bHDL in
   HK-2 cells, we performed a cellular uptake experiment in HK-2 cells
   overexpressing KIM-1. KIM-1 was overexpressed in HK-2 cells using
   lentivirus tool ( ~ 54% increase, supplementary Fig. [135]5). The
   uptake results showed that the overexpression of KIM-1 in HK-2 cells
   resulted in a significant increase ( ~ 60%) of bHDL uptake after injury
   (Fig. [136]4h). This represents a notable increase compared to HK-2
   cells without KIM-1 overexpression, which exhibited only around 20%
   uptake. We also attenuated the expression of KIM-1 in HK-2 cells by
   transfecting them with KIM-1-specific small interfering RNA. The
   downregulation of KIM-1 was confirmed using fluorescence microscopy
   (Supplementary Fig. [137]6a), qPCR (Supplementary Fig. [138]6b), and
   western blot analysis (Supplementary Fig. [139]6c-d). Flow cytometry
   analyses revealed a significant decrease in the uptake of bHDL in the
   KIM-1 knockdown HK-2 cells compared to the normal HK-2 cells after
   injury (Fig. [140]4g).

   To verify the increased uptake of bHDL by damaged cells compared with
   normal cells was caused by increased expression of KIM-1, not the SR-BI
   receptor, a widely known HDL receptor^[141]41, we assessed the
   expression of SR-BI on cisplatin-induced injured HK-2 cells. Western
   blot results showed that cisplatin injury did not increase the
   expression of SR-BI on HK-2 cells but rather slightly decreased it
   (Fig. [142]4i and j). In addition, we also detected the expression of
   SR-BI in the injured kidney of UUO mice by immunofluorescence. The data
   obtained from the fluorescence semi-quantitative analysis demonstrated
   that the injury did not significantly enhance the expression of SR-BI
   in the damaged kidneys (Supplementary Fig. [143]7a, b). It is
   noteworthy that injured kidneys are often accompanied by macrophage
   infiltration, which is highly expressed SR-BI. To gain a more accurate
   understanding of the impact of injury on SR-BI expression in RTECs, we
   conducted a detailed analysis of published single-cell sequencing data
   from UUO-injured kidneys^[144]42. No notable elevation in SR-BI
   expression could be found in RETCs (Supplementary Fig. [145]7c).
   Conversely, in macrophages, SR-BI expression demonstrated a substantial
   increase and diminished with the prolongation of UUO modelling time.
   Thus, all those evidences showed that more bHDL uptake in injured HK-2
   cells was not associated with SR-BI.

TP/BIBF-bHDL relieves renal injury and fibrosis in UUO model, folic acid
model and adenine model mice

   Then we evaluated the efficacy of TP/BIBF-bHDL in alleviating kidney
   injury and fibrosis in UUO mouse model. As shown in Fig. [146]5a,
   following ureteral ligation, the mice received various treatments
   through intravenous administration via the tail vein, repeated bi-daily
   over a 14-day period. On the fifteenth day after ligation, mice were
   sacrificed and samples were collected to assess renal damage and
   fibrosis levels. Protein-to-creatinine ratio (PCR) is used to assess
   the severity of total proteinuria. Compared to the control group, UUO
   mice showed higher urinary protein levels in urine and the protein
   level decreased after drug treatments, with the TP/BIBF-bHDL group
   showing the greatest decrease, approaching the levels observed in the
   control group (Supplementary Fig. [147]8a). Upon UUO, the contralateral
   kidney undergoes adaptive changes to compensate for the loss of
   function in the obstructed kidney. These changes initially provide
   benefit but eventually causes damages to the contralateral kidney,
   leading to increased urinary protein. The anti-proteinuric effect of
   TP/BIBF-bHDL is likely exerted on the contralateral side, as the
   obstructed kidney cannot contribute to urine output. We evaluated the
   level of creatinine (Cre) and urea in serum of mice, which are common
   indicators of renal function, and ureteral ligation causes increased
   urea levels that gradually returned to normal after treatment
   (Supplementary Fig. [148]8b-c). And the differences in blood creatinine
   levels among the groups were not significant (Supplementary
   Fig. [149]8b), which may be caused by the strong compensatory ability
   of the kidneys. Hematoxylin and eosin staining (HE) staining revealed
   severe glomerular atrophy, notable tubular dilatation, and
   proteinaceous tubular casts in UUO model mice, indicative of pronounced
   kidney injury. Post-treatment, these renal tissue alterations were
   ameliorated, with the TP/BIBF-bHDL group showing the highest degree of
   injury relief (Fig. [150]5b–e, Supplementary Fig. [151]8d). In
   addition, apoptosis studies also demonstrated that TP/BIBF-bHDL
   treatment effectively alleviate renal injury (Fig. [152]5d,
   Supplementary Fig. [153]8f). Masson staining identified collagen
   deposition in the renal interstitium of the model mice, which was
   significantly reduced in the group treated with the co-loaded
   preparation (Fig. [154]5c, f, Supplementary Fig. [155]8e). Fibrosis is
   characterized by the deposition of extracellular matrix, which is
   produced and secreted by activated myofibroblasts. To assess the
   anti-fibrotic effects of TP/BIBF-bHDL, we tested its impact on level of
   α-SMA expression. Consistent with the Masson staining results,
   TP/BIBF-bHDL effectively decreased α-SMA expression to the greatest
   extent (Fig. [156]5 g, o). In conclusion, our findings collectively
   demonstrate that TP/BIBF-bHDL significantly alleviates renal injury and
   fibrosis in UUO model mice.

Fig. 5. In vivo treatment efficacy of TP/BIBF-bHDL in UUO model and folic
acid model mice.

   [157]Fig. 5
   [158]Open in a new tab

   a Scheme of UUO model and administration regimen. b–d HE (b), Masson
   staining (c) and TUNEL assay (d) of kidney tissues of UUO mice after
   different treatments. Scale bars, 50 μm. n = 3 biologically independent
   samples. e, f The injury score (e) and fibrosis area (f) of kidney of
   UUO mice after different treatments. Data are mean ± SD (n = 3
   biologically independent samples, one-way ANOVA with Dunnett’s multiple
   comparisons test and adjustment applied). g Western blot analysis of
   α-SMA expression in kidney of UUO mice after different treatments.
   n = 3 biologically independent samples. h Description of folic acid
   model and administration regimen. i–k HE (i), Masson staining (j) and
   TUNEL assay (k) of kidney tissues from folic acid-induced mice after
   different treatments. Scale bars, 50 μm. n = 3 biologically independent
   samples. l, m The injury score (l) and fibrosis area (m) of kidney of
   folic acid model mice after treatment. Data are mean ± SD (n = 3
   biologically independent samples, one-way ANOVA with Dunnett’s multiple
   comparisons test and adjustment applied). n Western blot analysis of
   α-SMA expression in kidney of folic acid-induced mice after different
   treatments. n = 3 biologically independent samples. o, p
   Semi-quantification of α-SMA protein expression in UUO model mice and
   folic acid model mice with different treatments by Image J based the
   western blots shown in g (o) and n (p). It was normalized with
   corresponding α-tubulin signal. α-tubulin was used as internal control.
   The value of 42 and 55 kDa are referred to the predicted protein weight
   of α-SMA and α-tubulin. Data are mean ± SD (n = 3 biologically
   independent samples, one-way ANOVA with Dunnett’s multiple comparisons
   test and adjustment applied). a, h created in BioRender. He, S. (2024)
   [159]https://BioRender.com/k11e324. UUO model unilateral ureteral
   obstruction model mice, control healthy mice without any treatment, PBS
   Phosphate buffered saline, TP/BIBF mixed solution of triptolide and
   nintedanib, TP/BIBF-bHDL triptolide and nintedanib co-loaded biomimetic
   high-density lipoprotein nanoparticles, HE hematoxylin-eosin staining,
   TUNEL Terminal Deoxynucleotidyl Transferase-mediated dUTP Nick-End
   Labeling. Source data are listed in the Source Data file.

   Encouraged by the results, we further investigated the efficacy of
   TP/BIBF-bHDL in folic acid model and adenine model mice. In comparison
   to the control mice, the administration of both folic acid and adenine
   led to severe kidney injury and fibrosis. As expected, TP/BIBF-bHDL was
   most effective in improving kidney function, relieving renal injury and
   reducing collagen deposition caused by folic acid (Fig. [160]5i–n, p,
   Supplementary Fig. [161]8g–k) and adenine (Supplementary Fig. [162]9).
   The combined administration of TP and BIBF demonstrated superior
   antifibrotic effects compare to single treatment, indicating that the
   simultaneous anti-inflammatory and antifibrotic effects were more
   effective for disease remission (Fig. [163]5i and n, Supplementary
   Fig. [164]9e and i).

TP/BIBF-bHDL reverses renal fibrosis by remodeling the fibrosis niches

   To elucidate the underlying therapeutic mechanisms, we further employed
   RNA sequencing for comprehensive analysis. Total RNA was extracted from
   the kidneys of normal mice and UUO model mice treated with PBS or
   TP/BIBF-bHDL, and then subjected to transcriptome sequencing for
   analysis. Comparative gene expression studies among the different mouse
   groups identified 6044 and 2196 differentially expressed genes (DEGs)
   in the kidneys of control versus UUO-PBS mice, and UUO-PBS versus
   UUO-TP/BIBF-bHDL mice, respectively, with 2018 DEGs common to both
   comparisons. A heatmap of these DEGs indicated a significant
   upregulation of genes related to inflammation, immune cell
   infiltration, and extracellular matrix (ECM) generation post-model
   induction, whereas TP/BIBF-bHDL treatment markedly downregulated these
   genes (Fig. [165]6c). Kyoto Encyclopedia of Genes and Genomes (KEGG)
   pathway enrichment analysis revealed that the significantly
   up-regulated differential genes in UUO mice were mainly enriched for
   viral protein-cytokine receptor interaction, ECM-receptor interaction,
   cell adhesion molecules, cytokines-cytokine receptor interaction,
   complement and coagulation cascades and PI3K-Akt signaling pathways
   (Fig. [166]6d). In contrast, the significantly downregulated
   differential genes in TP/BIBF-bHDL-treated UUO mice were mainly
   enriched in ECM-receptor interaction, cytokines-cytokine receptor
   interaction and PI3K-Akt signaling pathways (Fig. [167]6e). Gene
   Ontology (GO) enrichment analysis of DGEs indicated a predominant
   enrichment in inflammatory response, immune response, extracellular
   matrix degradation, and activation of cytokines and chemokines
   (Fig. [168]6f, g). These findings reveal that TP/BIBF-bHDL treatment
   effectively lessen immune cell infiltration, dampen inflammatory factor
   production, and reduce ECM synthesis. Consequently, this implies the
   potential of TP/BIBF-bHDL in remodeling the fibrotic microenvironment
   and impeding the progression of fibrosis.

Fig. 6. Mechanism of TP/BIBF-bHDL treatment in UUO mice.

   [169]Fig. 6
   [170]Open in a new tab

   a, b Volcano of DEGs in kidneys among healthy mice and UUO mice treated
   with PBS and TP/BIBF-bHDL. The genes associated with fibrosis were
   marked. c Heatmap of genes related to ECM production, ECM degradation,
   inflammation and immune infiltration in different groups. d The top 20
   KEGG pathways enrichment up in the UUO mice compared to the control
   mice. e The top 20 KEGG pathways enrichment down in the
   TP/BIBF-bHDL-treated UUO mice compared to PBS-treated UUO mice. f The
   top 30 GO terms up in the UUO mice compared to the control mice. g The
   top 30 GO terms down in the TP/BIBF-bHDL-treated UUO mice compared to
   PBS-treated UUO mice. UUO model unilateral ureteral obstruction model
   mice, control untreated healthy mice, PBS phosphate buffered saline,
   TP/BIBF-bHDL triptolide and nintedanib co-loaded biomimetic
   high-density lipoprotein nanoparticles.

   To explore the capability of TP/BIBF-bHDL in modulating the fibrotic
   niches, we conducted a detailed analysis of its impact on myofibroblast
   activation, ECM deposition, immune cell infiltration, and inflammatory
   factors secretion in the kidney, employing immunohistochemical
   techniques. In line with our transcriptomic findings, significant
   elevations were observed in the kidneys of UUO mice in pivotal markers
   associated with myofibroblast activation (α-smooth muscle actin,
   α-SMA), ECM deposition (collagen type I alpha 1 chain, COL1A1), immune
   cell infiltration (macrophages identified by F4/80, and T-cells by
   CD3), as well as inflammatory factors (tumor necrosis factor-alpha,
   TNF-α, and interleukin-1 beta, IL-1β). These elevations were notably
   reduced in TP/BIBF-bHDL-treated UUO mice (Fig. [171]7a–g). qPCR
   experiments further substantiated the effective amelioration of the
   fibrotic microenvironment and mitigation of kidney injury by
   TP/BIBF-bHDL (Fig. [172]7h). We further investigated the ability of
   TP/BIBF-bHDL to regulate the fibrotic microenvironment in folic acid
   model and adenine model mice. Results indicated that TP/BIBF-bHDL was
   also effective in remodeling the fibrotic microenvironment and
   attenuating renal fibrosis in the folic acid model (Supplementary
   Fig. [173]10) and adenine model mice (Supplementary Fig. [174]11).
   TP/BIBF-bHDL treatment significantly reduced fibroblast activation and
   ECM production in the kidneys of folic acid-induced and adenine-induced
   mice. The levels of inflammatory factors and chemokines were lower in
   the kidneys of folic acid model and adenine model mice treated with
   TP/BIBF-bHDL compared to the PBS group. Collectively, these results
   suggest that TP/BIBF-bHDL, which simultaneously delivers
   anti-inflammatory and antifibrotic drugs to RTECs, mitigates renal
   injury and attenuates fibrosis by improving the fibrotic niches from an
   inflammatory, immune and fibrotic perspective, a mechanism that
   distinguishes it from those antifibrotic strategies that target a
   single pathogenetic process or pathway.

Fig. 7. TP/BIBF-bHDL exerts antifibrotic effects by remodeling the fibrotic
niche in UUO mice.

   [175]Fig. 7
   [176]Open in a new tab

   a Representative immunohistochemical images for
   COL1A1、α-SMA、F4/80、CD3、TNF-α and IL-1β in the kidney of UUO mice with
   different treatments. Scale bars, 40 µm. b–g Semi-quantitation of the
   COL1A1、α-SMA、F4/80、CD3、TNF-α and IL-1β expression level in the kidney
   of UUO mice with different treatments. Data are mean ± SD (n = 5
   biologically independent samples, one-way ANOVA with Dunnett’s multiple
   comparisons test and adjustment applied). h mRNA expression of
   Acta2、Col1a1、Fn1、Tgfb1、Tnf、Il1b and Ccl2 in the kidney of UUO mice with
   different treatments. Data are mean ± SD (n = 4 biologically
   independent samples, one-way ANOVA with Dunnett’s multiple comparisons
   test and adjustment applied). UUO model unilateral ureteral obstruction
   model mice, control untreated healthy mice, PBS phosphate buffered
   saline, TP triptolide solution, BIBF nintedanib solution, TP/BIBF a
   mixed solution of free triptolide and nintedanib, TP-bHDL, BIBF-bHDL
   and TP/BIBF-bHDL triptolide-loaded, nintedanib-loaded and triptolide
   and nintedanib co-loaded biomimetic high-density lipoprotein
   nanoparticles. Source data are listed in the Source Data file.

bHDL possess a good safety profile and ameliorate the toxicity of triptolide

   Then the safety profiles of bHDL and TP/BIBF-bHDL were investigated in
   vitro and in vivo. First, we assessed the cytotoxicity of blank bHDL
   nanoparticles in HK-2 and NRK-49F cells. Blank bHDL nanoparticles has
   good biocompatibility and does not show obvious toxicity even when the
   concentration of bHDL reaches 32 times the therapeutic concentration,
   indicating its excellent biocompatibility (Supplementary Fig. [177]12a,
   b). Next, we conducted the safety evaluation in vivo and the safety of
   bHDL and TP/BIBF-bHDL after intravenous administration for 7 days, 14
   days and 28 days were explored separately. As shown in Supplementary
   Fig. [178]12c, there was no difference in body weight change among
   control, bHDL, TP/BIBF, and TP/BIBF-bHDL groups. Even after the
   long-term administration (injection once every two days for 28 days),
   no obvious injury was found in major organs showed by HE staining
   (Fig. [179]8a). Moreover, blood routine examination demonstrated that
   there were no significant differences in all groups (Fig. [180]8c).
   It’s worth noting that long-term administration of TP/BIBF resulted in
   liver damage, with an increase in the liver function index ALT after 14
   and 28 days of administration (Fig. [181]8b, Supplementary
   Fig. [182]12d and e). Neither the bHDL group nor the TP/BIBF-bHDL group
   exhibited significant differences in the liver function markers ALT and
   AST at 14 and 28 days compare to the control group. These results
   showed that bHDL nanoparticles have good biocompatibility and do not
   bring toxic side effects even when administered for a long period of
   time and reduce the toxic side effects of drugs to some extent.

Fig. 8. bHDL nanoparticles possess a good safety profile and ameliorate the
toxicity of triptolide.

   [183]Fig. 8
   [184]Open in a new tab

   a HE staining of heart, liver, spleen, kidney, brain, testis in mice
   treated with different formulations for 28 days. Scale bars, 50 µm. b,
   c Hematological tests and Blood routine examination of healthy mice on
   day 28 after different treatments. Data are mean ± SD (n = 5
   biologically independent samples, one-way ANOVA with Dunnett’s multiple
   comparisons test and adjustment applied). d HE staining of heart,
   liver, spleen, kidney, brain in mice treated with TP, TP/BIBF, TP-bHDL
   and TP/BIBF-bHDL for 3 times at the concentration of 0.5 mg/mL TP,
   respectively. Scale bars, 100 µm. Control healthy mice, bHDL blank
   biomimetic high-density lipoprotein (HDL) nanoparticles, TP free
   triptolide, TP/BIBF mixed solution of free triptolide and nintedanib,
   TP-bHDL triptolide-loaded biomimetic HDL nanoparticles,
   TP/BIBF-bHDL triptolide and nintedanib co-loaded biomimetic HDL
   nanoparticles, ALT alanine aminotransferase, AST aspartate
   aminotransferase, Cre Creatinine, UREA urea nitrogen, CKMB creatine
   kinase MB, LDHI lactate dehydrogenase I, WBC white blood cell, RBC red
   blood cell, LYM lymphocyte, RDW red cell distribution width, PLT
   platelet, HGB hemoglobin, HE hematoxylin-eosin staining. Source data
   are listed in the Source Data file.

   In the realm of targeted drug delivery systems, a key objective,
   alongside enhancing drug efficacy, is to minimize drug toxicity. To
   evaluate the potential of bHDL in reducing toxicity, we performed acute
   toxicity experiment. Given TP’s high toxicity and narrow therapeutic
   window, we selected it as model drug to measure the median lethal dose
   (LD[50]) of TP and TP-bHDL. As the results shown in Table [185]S4, the
   LD[50] of TP-bHDL was 0.88 mg/kg, significantly greater than TP
   (0.48 mg/kg), underscoring bHDL’s substantial capability in mitigating
   TP’s toxicity. For a more comprehensive assessment of its toxicity
   reduction, various formulations, including TP solution, TP/BIBF
   solution, TP-bHDL, and TP/BIBF-bHDL, were administered at a dose
   equivalent to 0.5 mg/kg of TP (every other day). On the 7th day, all
   mice in the TP group had succumbed, whereas 20% survival was noted in
   the TP/BIBF group, 80% in the TP-bHDL group, and complete survival in
   the TP/BIBF-bHDL group. The major organs of the mice in each group were
   taken for pathological examination, revealing that TP group mice showed
   severe liver damage and cardiac damage (Fig. [186]8d). Additionally,
   significant cardiomyocyte atrophy, cardiac myofibrillar lysis, and
   hepatocellular atrophy could be observed. In the TP/BIBF
   co-administration group, significant cardiac injury was seen, but to a
   lesser extent than in the TP group, suggesting that co-administration
   could alleviate the toxicity of TP to some extent. In contrast, in the
   TP-bHDL as well as TP/BIBF-bHDL groups, no significant cardiac injury
   and liver injury were found. These findings collectively demonstrated
   that bHDL represents an efficient and safe drug delivery system,
   markedly reducing drug toxicity and enhancing overall safety.

Discussion

   As mentioned, developing drug delivery systems for RTECs is challenging
   due to multiple hard-to-pass barriers^[187]14, and for clinical
   translation there are even more practical issues. For example,
   lipid-based liposomes and micelles are often too big to pass through
   the glomerulus filter^[188]14,[189]43; polymeric solid nanoparticles
   usually suffer from fast protein corona formation and liver
   accumulation^[190]44,[191]45; and a large number of more complicated
   designs are hard to manufacture in large quantity and have safety
   concerns. These all limited the potential of current RTEC targeted
   delivery systems.

   In this study, we revealed that the simple and biocompatible bHDL
   nanoparticles have strong natural targeting properties to the injured
   RTECs and KIM-1 mediated the phagocytosis of injured RETCs. The bHDL
   nanoparticles possess several advantages. First, the mediator KIM-1 is
   only expressed on damaged RTECs but not normal RTECs, which grant the
   bHDL NPs high delivery selectivity^[192]35. Though KIM-1, also named
   T-cell immunoglobulin and mucin domain 1 (TIM-1), is extensively
   expressed in activated T and B cells too^[193]18,[194]19, this seems to
   not decisively affect the RTEC targeting capability of bHDL. Second, we
   found that KIM-1 likely initiated an endocytosis process that engulf
   the full bHDL. This is distinct to the endocytosis mediated by
   conventional bHDL binding partner SR-BI, where the endocytosis
   selectively uptake HDL cholesteryl esters and not whole
   nanoparticles^[195]25,[196]46. Third, bHDL, like other HDLs, are small
   biocompatible particles that are not primary objects for blood
   clearance nor prone to protein corona attachment^[197]47. These
   characteristics all help reduce its liver accumulation, enhance
   glomerulus passage, and elongate its circulation time, which ultimately
   lead to better RTEC delivery efficiency. Fourth, these vehicles are
   built from simple phospholipids and apoA-I mimicking peptide D-4F, with
   a straightforward method. The phospholipid DMPC has been applied in the
   Visudyne, which is a marketed product. Additionally, the D-4F has been
   tested in clinical trials for the treatment of cardiovascular
   diseases^[198]48. The simple structure and ingredient improve its
   manufacturability and make it easier to pass safety inspections.
   Nonetheless, several unresolved issues remain that warrant attention in
   future studies. For example, further investigation is needed into the
   distribution of surface charges after the D-4F peptide binds with
   phospholipids. Notably, the L-4F peptide, which exhibits similar
   properties and behavior to the D-4F peptide. Therefore, our future
   research should explore the potential of the L-4F peptide as an
   alternative for constructing a RTECs drug delivery system.

   As for the therapeutic strategy, consensus remains elusive regarding an
   effective treatment modality for fibrosis^[199]49,[200]50. It is known
   that inflammation is a pivotal driver in the pathogenesis of
   fibrosis^[201]49,[202]51, however, simple anti-inflammatory therapies,
   such as corticosteroids, has proven inadequate in mitigating
   fibrosis^[203]50,[204]52. Many attempts have been made to develop drugs
   targeting specific cells or pathways in the fibrosis process^[205]53.
   Presently, Nintedanib and Pirfenidone are the only anti-fibrotic agents
   with clinical approval, yet their therapeutic responses exhibit notable
   variability among patients^[206]54. As understanding of the pathology
   of fibrosis grows, it is increasingly recognized as a complex process
   involving multiple factors, with the fibrotic niches playing a crucial
   role in its progression^[207]55,[208]56. This suggests that
   multifaceted interventions in the fibrotic process might yield improved
   therapeutic outcomes. In this study, we delivered both
   anti-inflammatory and anti-fibrotic drugs to RTECs, with the aim of
   improving the treatment effect. TP not only inhibited the release of
   pro-inflammatory cytokines such as TNF-α and MCP-1 from RETCs, thereby
   reducing the infiltration of inflammatory cells in the renal
   interstitium^[209]57; but also suppressed the antigen-presenting
   capabilities of RETCs, thereby reducing T-cell activation and exerting
   immunosuppressive effects^[210]58. Nintedanib mitigates fibroblast
   proliferation, migration, and transformation by targeting PDGFR, VEGFR,
   and FGFR, thus plays a significant role in antifibrotic
   therapy^[211]59,[212]60. In addition, nintedanib could also suppress
   the inflammation of lung epithelial cell and inhibit the fibroblast-to
   myofibroblast transition by suppressing the TGF-β/Smad signaling
   pathways^[213]61,[214]62. The co-delivery system effectively targets
   both inflammation and fibroblast activation, thereby improving its
   capability to combat fibrosis by remodeling the fibrotic niches. The
   fibrotic niche not only provides an ideal environment for the
   activation and proliferation of fibroblasts but also negatively impacts
   the regenerative capacity and plasticity of surviving parenchymal
   cells, thereby hindering tissue repair and regeneration. Targeting and
   disrupting the fibrotic niches might be an effective therapeutic
   strategy to halt the progression of fibrosis. Our research indicates
   that combination therapies addressing both anti-inflammatory and
   anti-fibrotic pathways may offer a more effective approach to
   modulating the fibrotic niches than strategies focusing solely on a
   single aspect.

   In summary, we introduced a new renal tubule-targeted approach using
   bHDL nanoparticles and loaded both anti-inflammatory agent triptolide
   and antifibrotic compound nintedanib for CKD management (Fig. [215]9).
   We found that biocompatible bHDL is an excellent targeted drug delivery
   system for injured RETCs and revealed the underlying targeting
   mechanism that KIM-1 facilitates the endocytosis of these
   nanoparticles. This insight lays a groundwork for the practical
   application of bHDL in renal therapeutics. Our findings also indicate
   that co-administration of anti-inflammatory and anti-fibrotic drugs
   offers a more effective means of impeding the progression of fibrosis
   in CKD. Our study thus contributes several insights into the
   development of renal tubule-targeted drug delivery systems and the
   therapeutic approach for kidney diseases.

Fig. 9. Schematic illustration of TP/BIBF-bHDL preparation and KIM-1 mediated
targeting of bHDL nanoparticles to injured renal tubules for CKD treatment.

   [216]Fig. 9
   [217]Open in a new tab

   DMPC 1,2-dimyristoyl-sn-glycero-3-phosphocholine, KIM-1 kidney injury
   molecule 1, D-4F D-4F peptide, Injured RTECs injured renal tubular
   epithelial cells. Created in BioRender. He, S. (2024)
   [218]https//BioRender.com/u96q289.

Methods

   All the experiments in this study were conducted in strict accordance
   with the ethical standards and guidelines established by the Animal
   Ethics Committee of Sichuan University. The experimental protocols were
   reviewed and approved by the Ethics committee of Sichuan University to
   ensure compliance with relevant ethical regulations (KS2020022).

Materials

   1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC) were purchased from
   Shochem (Shanghai) Co., Ltd. D-4F (Ac-DWFKAFYDKVAEKFKEAF-NH2) was
   synthesized by GL Biochem (Shanghai) Ltd. Lotus Tetragonolobus Lectin
   (LTL), fluorescein (FITC)([219]L32480), KIM-1 Polyclonal Antibody
   (PA5-79345,Lot35FAFA04), and Protein Ladder (26619) were bought from
   Thermo Fisher Scientific Inc. Triptolide was purchased from chengdu
   Bio-purify phytochemicals Ltd. MTT assay kit, Nystatin, Dextran
   Sulfate(98%,CAS: 9011-18-1), Puromycin (99%, CAS: 58-58-2),
   Chlorpromazine Hydrochloride(98%,CAS: 69-69-0) and Amiloride HCl
   dihydrate (98%,CAS: 17440-83-4) was bought from Beijing Solarbio
   Science & Technology Co., Ltd. α-Smooth Muscle Actin Rabbit mAb
   (19245 T), COL1A1 (E8F4L,72026S) Rabbit mAb and F4/80 Rabbit mAb(D2S9R,
   70076S) were purchased from Cell Signaling Technology, Inc. Beta-actin
   antibody(sc81178, Lot#J1116),Alpha Tubulin antibody(sc-5286,
   Lot#L1522), SR-BI antibody(sc-518140, Lot#12322), mouse TIM-1
   antibody(A-12)(sc-518008, Lot#J2617), TNF alpha antibody(sc-52746) and
   CD3 antibody(sc-20047, Lot#G2821) were bought from Santa Cruz
   Biotechnology, Inc. SR-BI antibody(ab217318) for immunofluorescence was
   bought from Abcam. Rabbit IL-1β antibody (bioss, BS-0812R) was bought
   from Bioss CO., LTD. Lentivirus-RNAi human HAVCR1 (also referred as
   KIM-1) and HAVCR1 overexpression Lentivirus vector were produced by
   GeneChem.

Cell lines and Animals

   HK-2 cells were purchased from National Intra-structure of Cell Line
   Resource of China and NRK-49F were obtained from the Cell Resource
   Center, Peking Union Medical College. HK-2 cells were cultured in the
   DMEM/F12 medium containing 10% FBS, 1%ITS-X, and 1%
   streptomycin/penicillin at 37°C under 5% CO[2] humidified environment.
   NRK-49F cells were cultured in DMEM medium containing 10% FBS, 1%ITS-X,
   and 1% streptomycin/penicillin at 37°C under 5% CO2 humidified
   environment.

   Male Sprague Dawley rats (6-8weeks), male BALB/c mice (6 to 8 weeks)
   and male C57BL/6 mice (4 to 6 weeks) were purchased from Beijing HFK
   bioscience co., LTD. All animals used in the study were fed with a
   standard laboratory chow diet, providing a balanced nutrient profile
   for the animals. The animals were housed in a controlled environment
   with a 12-hour light/dark cycle, maintained at a temperature of
   22 ± 2 °C and humidity of 50 ± 10%. Food and water were provided
   freely. All the animal experiments were completed under the guidance of
   the Animal Ethics Committee of Sichuan University and all experiments
   received ethical approval (KS2020022).

HK-2 cell lentiviral transduction

   Lentivirus-RNAi human HAVCR1 (also named KIM-1) and HAVCR1
   overexpression Lentivirus vector were produced by GeneChem. HK-2 cells
   were seeded onto 6-well plates and when they grew to 50% confluence,
   transfected them with the lentivirus vectors for 48 hours. Puromycin
   was used to select stably transduce KIM-1 knockdown or KIM-1
   overexpression HK-2 cells and the validation was confirmed by qPCR.

Mouse models

UUO mice model

   Male BALB/C mice (4-6 weeks old) were anesthetized by intraperitoneal
   injection with tribromoethanol, restrained on the operating table, and
   incised the left abdomen of the mice, exposing the left ureter. Two
   ligatures were created at the opening of the renal pelvis and the upper
   1/3 of the ureter, and the left ureter was then cut at the midpoint of
   the two ligature points. After ligation, the kidney was repositioned
   and the muscle and skin were sutured layer by layer. In the control
   group, after opening the abdomen, the ureter was isolated without
   ligation and sutured directly to the abdomen. On the day after the
   ligation, UUO mice were treated with PBS, TP, BIBF, TP/BIBF, TP-bHDL,
   BIBF-bHDL and TP/BIBF-bHDL by tail vein injection at an equal dose
   (0.1 mg/kg TP and 2 mg/kg BIBF), every two days for fourteen days. The
   control group were treated with PBS. After the completion of treatment,
   blood samples were obtained for the analysis of blood-related indices.
   The mice were euthanized, and their kidneys were harvested and
   subsequently fixed for further experimentation.

Folic acid mice model

   Male C57BL/6 mice were injected intraperitoneally on day 1 and 14 with
   folic acid solution prepared by dissolving folic acid in 0.3 M sodium
   bicarbonate solution at a dose of 250 mg/kg. From the 15th day of
   modelling, PBS, TP, BIBF, TP/BIBF, TP-bHDL, BIBF-bHDL and TP/BIBF-bHDL
   were administered to the mice via tail vein injection at a dose of
   0.1 mg/kg TP and 2 mg/kg BIBF, every other day for a period of 14 days.
   Healthy male C57BL/6 mice treated with PBS were used as a control
   group. After the administration, blood was collected for blood-related
   indices. Mice were executed and kidneys were removed and fixed for
   subsequent experiments.

Adenine mice model

   Male BALB/C mice (4-6 weeks) were gavaged with adenine at a dose of
   70 mg/ (kg.d) for 21 consecutive days for modelling. Healthy male
   BALB/C mice treated with PBS were used as a control group. The
   treatment for adenine mice was started from the 15th day of modeling
   and treatment protocols were consistent with the folic acid model.

Preparation of bHDL nanoparticles

   bHDL and drug-laden bHDL nanoparticles were prepared as previously
   reported^[220]63, and the detailed steps were as follows:

Blank bHDL nanoparticles

   20 mg of DMPC were dissolved in a 3 mL mixture of chloroform and
   methanol solvent (9:1,v/v), then evaporated under vacuum at 37 °C for
   15 min to remove the solvent and hydrated the formed lipid film through
   sonication with 2 mL of PBS solution. Following this, 10 mg of the D-4F
   mimetic peptide dissolved in 2 mL of water was added to the hydrated
   lipid solution and sonicated for 10 minutes by a probe ultrasonic
   instrument in ice water bath. After completion of ultrasonication, the
   preparation solution underwent 3 cycles of hot and cold immersion, with
   each cycle involving a 60 °C and ice water bath respectively.

Drug-laden bHDL nanoparticles

   Briefly, DMPC (20 mg), TP (0.2 mg) and BIBF (2 mg) were dissolved with
   3 mL mixture solvent (chloroform: methanol = 9:1,v/v) in a 25 ml
   round-bottom flask. Then, the solvent was removed by evaporation and
   2 mL PBS solution was used to hydrate the lipid film. Then, 2 mL of
   D-4F(10 mg) peptide solution was added to the lipid-drug solution and
   sonicated them by probe ultrasound instrument (200 W,15 min). After
   that, the mixture was subjected to three thermal (60°C) and cold cycles
   (10 min for each cycle) to obtain TP/BIBF-bHDL nanoparticles. The
   preparation method of TP-bHDL and BIBF-bHDL were similar to
   TP/BIBF-bHDL, except that they were only encapsulated with TP or BIBF.

DiD-bHDL nanoparticles

   DiD-bHDL nanoparticles are prepared in a similar way to drug-laden
   nanoparticles, except that the drug is replaced with DiD(120 µg).

Characterization of bHDL nanoparticles

   The particle size distribution and zeta potential were evaluated by a
   dynamic light scatter (Zetasizer Nano ZS90, Malvern Instrument, UK).
   Transmission electron microscopy and cryo-electron microscopy were used
   to examine the morphology of bHDL nanoparticles.

   A combination of ultrafiltration centrifugation and HPLC methods were
   employed to measure the EE% and LC% of TP/BIBF-bHDL. Initially, free
   drugs and nanoparticles were separated through ultrafiltration
   centrifugation method. 2 mL of TP/BIBF-bHDL nanoparticle solution was
   added to an Amicon centrifugal filter tube with a 3 kDa molecular
   weight cut-off (MWCO) and centrifuged for 10 minutes (9280 x g) to
   separate free drugs and drug-laden nanoparticles. Subsequently, the
   concentrations of TP and BIBF in the nanoparticles were analyzed using
   HPLC after breaking the emulsion with methanol at three times the
   volume. EE% and LC% were calculated using the following equations:
   [MATH: <mi mathvariant="normal">E</mi><mi
   mathvariant="normal">E</mi><mi>%</mi><mo>=</mo><mfrac><mrow><mi>W</mi><
   mi>e</mi></mrow><mrow><mi
   mathvariant="normal">W</mi><mi>t</mi></mrow></mfrac><mo>×</mo><mn>100</
   mn><mi>%</mi> :MATH]
   1

   and
   [MATH: <mi
   mathvariant="normal">LC</mi><mi>%</mi><mo>=</mo><mfrac><mrow><mi
   mathvariant="normal">W</mi><mi>e</mi></mrow><mrow><mi>W</mi><mi>d</mi><
   /mrow></mfrac><mo>×</mo><mn>100</mn><mi>%</mi> :MATH]
   2

   where W[e] means the amount of TP or BIBF in the nanoparticles, W[t]
   means the total amount of the drugs and W[d] means the total weight of
   the nanoparticles.

   The HPLC analysis was conducted on a high-performance liquid
   chromatography (Agilent 1260, USA) equipped with a C18 column (250 mm×
   4.6 mm, 5 µm). The mobile phase consisted of acetonitrile and 0.05%
   formic acid in water, with a gradient elution at a flow rate of
   1 ml/min. Detection was performed on a UV detector, set at wavelengths
   of 210 nm for TP and 385 nm for BIBF. And the injection volume of each
   sample was 20 µL and the column temperature was set at 35 °C. All
   samples were prepared via protein precipitation using methanol prior to
   analysis.

Preparation of DiD-liposomes

   DiD-liposomes were prepared by thin-film hydration method. Briefly,
   DMPC (20 mg), cholesterol (1 mg) and DiD (120 µg) were dissolved in
   3 mL of chloroform and evaporated the solvent to form a lipid film.
   Then, hydrated the lipid film with 4 ml water and sonicated it for
   20 min to form the DiD-Liposomes.

The stability study of bHDL nanoparticles

   The stability of bHDL nanoparticles was estimated by observing the
   particle size alterations in different conditions, as well as comparing
   the in vitro release behavior of bHDL nanoparticles in PBS and 10% FBS.
   First, TP/BIBF-bHDL nanoparticles were placed at 4 °C and 37 °C for
   various durations, and the changes in particle size were detected by
   DLS. Then, to explore the stability of TP/BIBF-bHDL in serum, TEM was
   used to observe the size and shape of bHDL nanoparticles after
   incubating with 10% FBS for 2 hours at 37 °C. Given the endogenous
   lipoproteins in FBS may affect the observation, we observed the
   particle size of nanoparticles in both normal FBS as well as
   lipoprotein-removed FBS.

   The in vitro drug release of bHDL nanoparticles was evaluated by
   dialysis method. Briefly, 0.5 ml of TP/BIBF-bHDL was added into the
   dialysis bag (MWCO = 3 kDa) and incubated in 30 mL of PBS or 10%FBS
   with gentle stirring. At designated time points, the buffer was
   extracted and replaced with an equal volume of fresh medium. The
   concentration of the drugs in the buffer was determined by HPLC.

Analysis of serum protein absorbed on bHDL nanoparticles

   The protein level absorbed on the bHDL nanoparticles in serum was then
   estimated and liposome group was added as a control. Briefly, 0.5 mL of
   bHDL nanoparticles or liposomes were incubated with 0.5 mL of FBS or
   lipoprotein-removed FBS for 2 hours at room temperature. Subsequently,
   the bHDL-protein complexes were separated by chemical precipitation
   with a mixture of dextran sulfate and magnesium chloride, while the
   liposome-protein complexes were isolated by ultracentrifugation at
   180,000  × g for four hours. Then, the complexes were lysed with 200 µl
   of RIPA, and the amount of adsorbed protein was quantified using the
   BCA assay.

Hemolysis Assay

   Mice blood was collected and centrifuged at 3341  ×  g for 5 minutes to
   obtain red blood cells(RBCs). The collected RBCs were washed three
   times with PBS and then prepared into RBCs suspension with a
   concentration of 8%(v/v). The RBCs suspension was mixed with
   TP/BIBF-bHDL in equal volume and incubated for 1 hour at 37°C. 1%Triton
   X-100 was used as a positive control and PBS solution was used as the
   negative control. After incubation, the supernatant was centrifuged
   (3341 x g, 5 min) and the absorbance was tested at 570 nm via
   microplate Reader. The hemolysis rate was calculated according to the
   equation:
   [MATH: <mi mathvariant="normal">Hemolysis</mi><mspace
   width="0.25em"></mspace><mi>%</mi><mo>=</mo><mfrac><mrow><msub><mrow><m
   i>A</mi></mrow><mrow><mi>s</mi><mi>a</mi><mi>m</mi><mi>p</mi><mi>l</mi>
   <mi>e</mi></mrow></msub><mo>−</mo><msub><mrow><mi>A</mi></mrow><mrow><m
   i>n</mi><mi>e</mi><mi>g</mi><mi>a</mi><mi>t</mi><mi>i</mi><mi>v</mi><mi
   >e</mi></mrow></msub></mrow><mrow><msub><mrow><mi>A</mi></mrow><mrow><m
   i>p</mi><mi>o</mi><mi>s</mi><mi>i</mi><mi>t</mi><mi>i</mi><mi>v</mi><mi
   >e</mi></mrow></msub><mo>−</mo><msub><mrow><mi>A</mi></mrow><mrow><mi>n
   </mi><mi>e</mi><mi>g</mi><mi>a</mi><mi>t</mi><mi>i</mi><mi>v</mi><mi>e<
   /mi></mrow></msub></mrow></mfrac><mo>×</mo><mn>100</mn><mi>%</mi>
   :MATH]
   3

In vivo biodistribution assay

   UUO model mice were applied to conduct the in vivo biodistribution
   experiment. Forty-eight UUO model mice were randomly divided into 3
   groups and injected with DiD solution, DiD-liposome and DiD-bHDL (6 μg
   DiD each mouse) via tail vein. At 0.5, 2, 8 and 12 hours after
   administration, the mice were euthanized, and their organs were
   isolated and analyzed for fluorescence intensity using in vivo imaging.
   To investigate the localization of bHDL in the kidney, isolated kidney
   tissues were promptly fixed and subjected to analysis through
   immunofluorescence staining. To compare the distribution of bHDL
   nanoparticles in healthy and injured kidney, healthy male mice were
   also applied to conduct the biodistribution study.

   To achieve a more precise characterization of the in vivo distribution
   of bHDL, we utilized the Liquid Chromatography Mass Spectrometry
   (LC-MS/MS) method for distribution experiment. Fifty UUO model mice
   were injected with TP/BIBF solution or TP/BIBF-bHDL into the tail vein
   at a dose of 2 mg/kg BIBF. At each predetermined time point, and the
   mice were euthanized, their organs were isolated, weighed, and
   homogenized with twice the tissue weight in saline to prepare a
   homogenate. Subsequently, the quantity of BIBF in each tissue was
   determined through LC-MS/MS. Pharmacokinetics parameters of TP/BIBF or
   TP/BIBF-bHDL in kidney after intravenous administration were calculated
   by Data and Statistics Software (DAS3.0; Shanghai, China). Relative
   uptake efficiency (Re [kidney]) and concentration efficiency (Ce
   [kidney]) were calculated using the following equations:
   [MATH: <msub><mrow><mi
   mathvariant="normal">Re</mi></mrow><mrow><mi>k</mi><mi>i</mi><mi>d</mi>
   <mi>n</mi><mi>e</mi><mi>y</mi></mrow></msub><mo>=</mo><msub><mrow><mrow
   ><mo>(</mo><mrow><msub><mrow><mi
   mathvariant="normal">AUC</mi></mrow><mrow><mn>0</mn><mo>−</mo><mi
   mathvariant="normal">t</mi><mo>,</mo><mi
   mathvariant="normal">kidney</mi></mrow></msub></mrow><mo>)</mo></mrow><
   /mrow><mrow><mi mathvariant="normal">TP</mi><mo>/</mo><mi
   mathvariant="normal">BIBF</mi><mo>−</mo><mi
   mathvariant="normal">bHDL</mi></mrow></msub><mo>/</mo><msub><mrow><mrow
   ><mo>(</mo><mrow><msub><mrow><mi
   mathvariant="normal">AUC</mi></mrow><mrow><mn>0</mn><mo>−</mo><mi
   mathvariant="normal">t</mi><mo>,</mo><mi
   mathvariant="normal">kidney</mi></mrow></msub></mrow><mo>)</mo></mrow><
   /mrow><mrow><mi mathvariant="normal">TP</mi><mo>/</mo><mi
   mathvariant="normal">BIBF</mi></mrow></msub> :MATH]
   4
   [MATH: <msub><mrow><mi mathvariant="normal">C</mi><mi
   mathvariant="normal">e</mi></mrow><mrow><mi
   mathvariant="normal">kidney</mi></mrow></msub><mo>=</mo><msub><mrow><mr
   ow><mo>(</mo><mrow><msub><mrow><mi
   mathvariant="normal">C</mi></mrow><mrow><mi>max</mi><mo>,</mo><mi
   mathvariant="normal">kidney</mi></mrow></msub></mrow><mo>)</mo></mrow><
   /mrow><mrow><mi mathvariant="normal">TP</mi><mo>/</mo><mi
   mathvariant="normal">BIBF</mi><mo>−</mo><mi
   mathvariant="normal">bHDL</mi></mrow></msub><mo>/</mo><msub><mrow><mrow
   ><mo>(</mo><mrow><msub><mrow><mi
   mathvariant="normal">C</mi></mrow><mrow><mi>max</mi><mo>,</mo><mi
   mathvariant="normal">glomeruli</mi></mrow></msub></mrow><mo>)</mo></mro
   w></mrow><mrow><mi mathvariant="normal">TP</mi><mo>/</mo><mi
   mathvariant="normal">BIBF</mi></mrow></msub> :MATH]
   5

Pharmacokinetics experiment

   To investigate the in vivo behavior of TP/BIBF-bHDL, we also performed
   pharmacokinetic study in vivo. Eighteen male SD rats were randomly
   divided into three groups and received equal doses (2 mg/kg BIBF) of
   TP/BIBF or TP/BIBF-bHDL by tail vein injection or administration of
   TP/BIBF (2 mg/kg BIBF) by gavage, and blood was collected from the
   orbital region of the rats in each group at scheduled time points after
   administration. Then the plasma was obtained through centrifugation
   (3341 × g, 10 min) and the drug content in plasma was estimated via the
   LC-MS/MS method.

In vitro cellular uptake assay

   HK-2 cells were seeded onto 6-well plates, cultured at 37 °C, 5% CO[2]
   for 24 hours, then treated with cisplatin for another 24 hours at the
   concentration of 2 µg/mL to induce injury. Then the injured HK-2 cells
   and normal HK-2 cells were treated with DiD-bHDL at a dose of 200 ng/ml
   at 37°C for 2 hours. After incubation, the cells were washed with PBS,
   digested with trypsin and the fluorescence intensity in the cells was
   evaluated by Flow Cytometry. The uptake experiment of bHDL was also
   performed in KIM-1 knockdown and KIM-1 overexpression HK-2 cells with
   or without cisplatin-induced injury.

   NRK-49F cells were also applied to evaluate the uptake of bHDL. NRK-49
   cells were inoculated in 6-well plates and cultured for 24 hours with
   or without 10 ng/mL TGF-β1, which could induce transformation of
   fibroblasts into myofibroblasts. Then DiD-bHDL nanoparticles were added
   to the cells for 2 hours incubation and the fluorescence intensity were
   measured with the flow cytometry.

In vitro competitive inhibition assay

   HK-2 cells were inoculated in 6-well plates and cultured overnight at
   37 °C, 5% CO[2]. HK-2 cells were pretreated with different endocytosis
   inhibitors (Chlorpromazine at a concentration of 25 µM, Nystatin at a
   concentration of 50 µM, Amiloride at a concentration of 50 µM, β-CD and
   Dextran sulfate sodium at a concentration of 50 µg/mL, Gentamicin and
   O-Phospho-L-serine at a dose of 20 µg/ml, Cytochalasin D at a
   concentration of 1.25 µg/mL, cubilin antibody at a concentration of
   30 µg/mL, BLT-1 at a concentration of 25 µM and D-4F with different
   dose) for 0.5 hours, and then administrated with DiD-bHDL at a
   concentration of 200 µg/mL for 2 hours. After incubation, the cells
   were washed with PBS and tested by Flow cytometry. Moreover, HK-2 cells
   were treated with DiD-bHDL at a dose of 200 µg/mL for 2 hours at 4°C
   and the uptake of DiD-bHDL was estimated.

   To explore whether KIM-1 mediated the uptake of bHDL by the injured
   HK-2 cells, the competitive inhibition assay was also conducted with
   the KIM-1 antibody. HK-2 cells were incubated with 2ug/mL cisplatin for
   24 hours to induce injury and then treated the cells with or without
   KIM-1 antibody(30 µg/mL) for 0.5 hours. Then DiD-bHDL nanoparticles
   were added to the cells for another 2 hours and flow cytometry was
   applied to evaluate the uptake of bHDL. Normal HK-2 cells were also
   added as control.

KIM-1 targeting assay

   To further examine whether KIM-1 mediates bHDL’s uptake, HK-2 cells
   were inoculated in laser confocal petri dishes and cultured overnight.
   HK-2 cells were treated with cisplatin at a concentration of 2 µg/mL
   for 24 hours and then administrated with DiD-bHDL for 2 hours. After
   incubation, the cells were washed with PBS, stained with KIM-1 antibody
   and FITC-labeled secondary antibody. DAPI was used to stain the nuclei.
   After staining, CLSM was applied to examine the co-localization of
   KIM-1 and DiD-bHDL.

   Moreover, co-localization of KIM-1 and DiD-bHDL were estimated in vivo.
   UUO mice were administered with DiD-bHDL or DiD-liposome at a dose of
   6 µg by tail vein injection on the seventh day after ureteral ligation.
   Two hours after administration, the mice were euthanized and kidneys
   were extracted, fixed and subjected to immunofluorescence staining,
   whereby KIM-1 protein was labelled using KIM-1 antibody and
   FITC-labeled secondary antibody. The fluorescent sections were
   subjected to CLSM to observe co-localization. Normal mice were added as
   control.

Renal function assessment

   The levels of Cre and urea in the serum were assessed to evaluate the
   renal function of mice after different treatment. After the completion
   of treatments, the blood was collected, centrifuged to take the upper
   serum layer, and the levels of Cre and urea were measured using a Roche
   blood biochemistry test.

Histopathology and Immunohistochemistry staining

   Renal tissues were fixed in 4% paraformaldehyde, dehydrated, hyalinized
   and embedded in paraffin. Then the HE staining and Masson staining were
   conducted with the paraffin-embedded kidney tissues. HE-stained
   sections were examined under the microscope to observe specific lesions
   and kidney damage was scored using a four-point scale: no damage (0
   points), slight (1 point), mild (2 points), moderate (3 points) and
   severe (4 points), respectively. Masson sections were examined using a
   microscope and images were taken. The fibrous tissue area (Area) in the
   images was determined using an Image-Pro Plus 6.0 image analysis system
   (Media Cybernetics, USA). The fibrous tissue expression area percentage
   was calculated by the following formula:
   [MATH: <mi mathvariant="normal">Fibrosis
   area</mi><mrow><mo>(</mo><mrow><mi>%</mi></mrow><mo>)</mo></mrow><mo>=<
   /mo><mfrac><mrow><mi>f</mi><mi>i</mi><mi>b</mi><mi>r</mi><mi>o</mi><mi>
   u</mi><mi>s</mi><mspace
   width="0.16em"></mspace><mi>t</mi><mi>i</mi><mi>s</mi><mi>s</mi><mi>u</
   mi><mi>e</mi><mspace
   width="0.16em"></mspace><mi>a</mi><mi>r</mi><mi>e</mi><mi>a</mi></mrow>
   <mrow><mi>f</mi><mi>i</mi><mi>e</mi><mi>l</mi><mi>d</mi><mspace
   width="0.16em"></mspace><mi>o</mi><mi>f</mi><mspace
   width="0.16em"></mspace><mi>v</mi><mi>i</mi><mi>e</mi><mi>w</mi><mspace
   width="0.16em"></mspace><mi>a</mi><mi>r</mi><mi>e</mi><mi>a</mi><mrow><
   mo>(</mo><mrow><mi>p</mi><mi>i</mi><mi>x</mi><mi>e</mi><mi>l</mi><mspac
   e
   width="0.16em"></mspace><mi>a</mi><mi>r</mi><mi>e</mi><mi>a</mi></mrow>
   <mo>)</mo></mrow></mrow></mfrac> :MATH]
   6

   For immunohistochemistry staining, paraffin-embedded sections were
   obtained, enclosed in serum and then incubated sequentially with
   primary antibodies against α-SMA, COL1A1, F4/80, CD3, IL-1β and TNF-α,
   and secondary antibodies. Finally, staining was conducted using
   3,3’-diaminobenzidine (DAB) and re-stained with hematoxylin.
   Photographs of the sections were taken and the percentage of positive
   area per image (% DAB Positive Tissue) was calculated using the Halo
   data analysis system.

Immunofluorescence staining

   Renal tissues were embedded in the Optimal Cutting Temperature (OCT)
   compound and sectioned. The sections were stained with FITC-labeled
   LTL, and sequentially incubated with primary antibodies against KIM-1
   and fluorescently labeled secondary antibodies. DAPI was applied to
   stain the nuclei. To estimate the expression of SR-BI in kidney,
   sections were stained with primary antibodies against SR-BI and
   subsequently with fluorescently labeled secondary antibodies. The
   nuclei were stained using DAPI.

   For TUNEL staining, Paraffin-embedded kidney sections were stained with
   fluorescent TUNEL incubation solution according to the operation of the
   TUNEL kit (Roche, Swiss). The nuclei were stained with DAPI.

Western blot assay

   The protein expression of KIM-1, SR-BI and α-SMA in kidney and HK-2
   cells were evaluated by Western blot. Renal tissues and HK-2 cells were
   lysed by RIPA Lysis buffer containing 1% protease inhibitor for
   30 minutes. After lysis, the protein concentration of each sample was
   estimated by BCA assay kit. Loading buffer was added to the samples and
   samples were boiled for 10 minutes. Then samples were uploaded into the
   PAGE gel to separate the proteins and transferred them with PVDF films.
   Then these films were closed with 5% milk solution for 1 hour,
   incubated with primary antibodies at 4°C overnight and followed by
   HRP-conjugated secondary antibodies. Proteins were detected using a ECL
   detection kit.

Quantitative PCR assay

   Gene expression in both kidney tissue and HK-2 cells was estimated by
   qPCR. Total RNA was extracted using FastPure cell/Tissue Total RNA
   Isolation Kit provided by Nanjing Vazyme Biotech Co., Ltd. Then total
   RNA was reverse transcribed to obtain cDNA using 5ˣ HiScript III qRT
   SuperMix from Vazyme Biotech Co., Ltd. ChamQ SYBR qPCR Master Mix was
   used in the PCR study and the experimental procedures was conducted
   according to the protocol provided by Vazyme. The primers for
   quantitative real-time PCR were provided by Sangon Biotech and listed
   in Table [221]S5. Relative expression of the target genes was
   normalized to β-actin levels in kidney tissues and GAPDH levels in HK-2
   cells.

RNA-sequencing

   Total RNA was extracted from kidneys of control mice or UUO model mice
   treated with PBS or TP/BIBF-bHDL (0.1 mg/kg TP, 2 mg/kg BIBF). Then the
   RNA sequencing was conducted by Shanghai OE Biotech Co., Ltd and the
   transcriptome sequencing data was collected on an Ilumina Novaseq 6000
   platform. DESeq2 software was used to analyze differentially expressed
   genes.

Safety evaluation assay

   HK-2 cells and NRK-49Fcells were seeded onto 96-well plates and
   cultured overnight. Blank bHDL was diluted with culture medium into its
   multiplicative concentration solution based on the normal administered
   concentration, which ranges from 0.5 to 32 times. The cells were
   treated with the diluted bHDL solutions for 24 hours and then cell
   counting kit-8 (CCK-8) was applied to test the cell viability of the
   cells.

   Sixty Male BALB/C mice aged 4-6 weeks were randomly divided into four
   groups and administrated with PBS, bHDL, TP/BIBF and TP/BIBF-bHDL via
   tail vein injection at a concentration of 0.1 mg/kg TP and 2 mg/kg
   BIBF. The drug was administered every two days for 28 consecutive days,
   and the body weights of mice were recorded. On the 7th, 14th and 28th
   days of administration, blood samples were collected for Routine blood
   tests and blood biochemistry examination, and tissue samples were taken
   for pathological examination by H&E staining.

   To further investigate the toxicity-reducing effect of bHDL, we
   performed acute toxicity experiments to measure the LD[50] of both
   TP/BIBF and TP/BIBF-bHDL. Due to the high toxicity of TP, we selected
   to calculate the LD[50] based on the concentration of TP. Fifty male
   BALB/C mice (4-6 weeks) were randomly divided into ten groups and were
   injected via the tail vein with varying concentrations of TP (0.3,
   0.375, 0.47, 0.58, and 0.74 µg/mL TP) and TP-bHDL (0.18, 0.3, 0.5,
   0.83, and 1.38 µg/mL TP). After administration, the condition of mice
   in each group was monitored and the death of mice at varying doses was
   recorded over 7 days. Moreover, twenty male BALB/C mice (4-6 weeks)
   were randomly assigned to four groups and injected by tail vein with
   TP, TP-bHDL, TP/BIBF, and TP/BIBF-bHDL at a concentration of 0.5 mg/kg
   TP for three times. The condition of the mice was recorded under close
   observation. On the seventh day after administration, mice were
   executed and tissues of the heart, liver, spleen, lungs and kidneys
   were taken for histopathological examination.

Statistical analysis

   All data analyses were performed using GraphPad software (Vision 8).
   All experimental data are presented as mean ± standard deviation. For
   two-group comparison, unpaired T test with two tailed test was used for
   comparison. Ordinary one-way ANOVA with Dunnett’s multiple comparisons
   test and two-way ANOVA with Sidak’s multiple comparisons test were
   conducted for multiple comparisons. Statistical significance was
   defined as a p value of less than 0.05.

Reporting summary

   Further information on research design is available in the [222]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [223]Supplementary Information^ (14.2MB, pdf)
   [224]Reporting Summary^ (3.1MB, pdf)
   [225]Transparent Peer Review file^ (10.3MB, pdf)

Source data

   [226]Source Data^ (2.1MB, xlsx)

Acknowledgements