Abstract
Discogenic pain, caused by intervertebral disc degeneration (IVDD), is
a prevalent and challenging condition to treat effectively. Macrophage
infiltration with neural ectopic in‐growth resulting from structural
disturbances within the intervertebral disc (IVD) is a major cause of
discogenic pain. This work systematically reveals how nanoparticles can
synergistically regulate the immune microenvironment and mitochondrial
communication to attenuate discogenic pain. The antioxidant
metal‐polyphenol nanoparticle system can sequentially regulate
macrophage phenotype and mitochondrial delivery efficiency. This
strategy circumvents the necessity for mitochondrial isolation and
preservation techniques that are typically required in conventional
mitochondrial transplantation procedures. Furthermore, it facilitates
the effective and sustained delivery of mitochondria to damaged cells.
In vivo, this nanoparticle formulation effectively preserves the IVD
height, maintains the structural integrity of the nucleus pulposus
(NP), and restores pain thresholds. Thus, this nanoplatform offers an
effective approach to traditional surgical treatments for discogenic
pain, with significant potential for clinical application.
Keywords: discogenic pain, macrophage, mitochondrial transfer,
nanoparticles
__________________________________________________________________
Gallic acid (GA) and copper ions self‐assemble to form nanoparticles,
which are then modified with mitochondrial targeting peptides and gap
junction modulator. These nanoparticles scavenge mitochondrial reactive
oxygen species to induce M2 polarization and enhance intercellular
mitochondrial transfer. Within the intervertebral disc,
macrophage‐derived healthy mitochondria efficiently maintain
mitochondrial function in damaged cells to alleviate discogenic pain.
graphic file with name ADVS-12-2500128-g005.jpg
1. Introduction
According to the Global Burden of Disease study, low back pain (LBP) is
the leading cause of disability and limitation of daily living in
adults worldwide.^[ [50]^1 ^] It is predicted that more than 800
million people worldwide will suffer from LBP by 2050.^[ [51]^2 ^] As
such, LBP represents a significant medical burden and a major global
challenge to public health infrastructure.^[ [52]^3 ^] Although LBP is
often considered non‐specific, the most common specific cause is IVDD,
also known as discogenic pain.^[ [53]^4 ^] This pain is primarily the
result of a disruption in the internal structure of the IVD. As IVDD
progresses, the collagen fibers in the annulus fibrosus (AF) become
disorganized, and the cartilage endplates (CE) undergo severe
destruction and hardening.^[ [54]^5 ^] These structural changes
compromise the physiological barrier of the NP, resulting in exposure
to the immune system and macrophage infiltration, which tend to become
M1‐polarized.^[ [55]^6 ^] Additionally, nerve fibers that were
initially located outside the AF invade the NP.^[ [56]^7 ^] The
inflammatory and neurotrophic factors secreted by M1 macrophages
stimulate the expression of pain‐associated cation channels and the
production of pain transmitters in the dorsal root ganglion, thereby
developing discogenic pain.^[ [57]^8 ^] Consequently, a complex
pathological microenvironment of multicellular coexistence is formed in
the nucleus pulposus in the IVDD state.
Most previous studies have concentrated on the function of individual
cell types in the context of IVDD,^[ [58]^9 ^] with a notable absence
of research examining the collective influence of intercellular
interactions on disc microenvironmental homeostasis. It is crucial to
acknowledge that IVD is a highly organized and dynamic tissue. The
disruption of one cell type can have a cascading effect on other cell
types, leading to the onset of multifaceted degenerative processes. In
particular, the interaction of immune cells with histiocytes is more
complex than that of other cell types, and this area remains
under‐explored. In IVDD, an imbalance in macrophage polarization
results in significant oxidative stress on NP cells and neurons.^[
[59]^10 ^] Oxidative stress plays a pivotal role in mitochondrial
damage, which is the site of cellular redox reactions. Impaired
mitochondrial function results in cellular dysfunction and reduced
activity.^[ [60]^11 ^] Therefore, it is imperative to investigate the
potential of structured therapeutic approaches within a disrupted
pathological environment. Mitochondrial transfer represents an emerging
therapeutic strategy for mitochondrial complementation. This approach
involves mitochondria transfer between cells in a non‐vertical genetic
manner. It has been demonstrated to be an effective means of restoring
mitochondrial homeostasis.^[ [61]^12 ^] In contrast to direct
replenishment, mitochondrial transfer leverages cells' intrinsic
capacity to transfer mitochondria spontaneously, obviating the
necessity for isolation and preservation.^[ [62]^13 ^] Multiple
investigations have substantiated that macrophages can serve as donor
cells to furnish healthy mitochondria to injured cells.^[ [63]^14 ^]
Macrophages are pervasively distributed, communicate extensively with
neighboring cells, and can migrate to lesions, thus facilitating
mitochondrial delivery.^[ [64]^15 ^] Given the dual role of macrophages
in IVDD, we propose a potential therapeutic strategy to convert
macrophage polarization ratio from M1‐type to M2‐type. Subsequently, M2
macrophages could rescue damaged cells via mitochondrial transfer,
thereby alleviating discogenic pain through mechanisms beyond purely
anti‐inflammatory effects.
Nevertheless, the sustained and effective delivery of mitochondria to
damaged cells represents a significant challenge. One of the primary
limitations to the efficacy of mitochondrial transfer is the quality of
mitochondria in donor cells.^[ [65]^16 ^] Reactive oxygen species
(ROS), a byproduct of mitochondrial metabolism, significantly
contributes to mitochondrial dysfunction. Additionally, ROS, especially
mitochondria‐derived reactive oxygen species (mtROS) plays a pivotal
role in macrophage polarization, inducing M1‐type and inhibiting
M2‐type polarization.^[ [66]^17 ^] Consequently, removing mtROS from
macrophage can restore mitochondrial function and shift macrophage
polarization from M1‐type to M2‐type, thereby improving the quality of
mitochondrial delivery from macrophages. Moreover, the efficacy of
mitochondrial transfer is constrained by its transfer efficiency.^[
[67]^18 ^] Tunneling nanotubes (TNTs) represent a pivotal pathway for
mitochondrial transfer. Functional gap junction channels (GJCs)
constitute TNTs extension terminals that regulate TNTs formation and
stability.^[ [68]^19 ^] The up‐regulation of gap junction proteins
(e.g., Cx43) effectively enhances transfer efficiency.
This study aims to investigate a therapeutic strategy for IVDD by
capitalizing on the potential of macrophage‐mediated mitochondrial
transfer. It is hypothesized that the modulation of macrophage
polarization to favor M2 macrophages, in conjunction with the
enhancement of mitochondrial transfer efficiency, may facilitate the
restoration of mitochondrial function in damaged NP cells and neurons.
To address the challenges of delivering mitochondria, we propose using
nanoparticles to clear excessive ROS in macrophages, thereby improving
the quality of donor mitochondria and promoting mitochondrial transfer
through gap junction modulation. By focusing on these key mechanisms,
we aim to provide insights into a combined immunotherapy and
mitochondrial transfer approach that may offer new avenues for treating
IVDD (Scheme [69]1 ).
Scheme 1.
Scheme 1
[70]Open in a new tab
Synergistic Modulating of Mitochondrial Transfer and Immune
Microenvironment to Attenuate Discogenic Pain.
2. Results
2.1. Macrophage Infiltration and Polarization Imbalance Exacerbate IVDD and
Discogenic Pain
In order to investigate the potential correlation between macrophage
infiltration and IVDD, we employed Uniform Manifold Approximation and
Projection for the visualization of immune cells present in the human
IVD, utilizing a single‐cell RNA‐seq dataset ([71]GSE165722) from the
Gene Expression Omnibus database.^[ [72]^20 ^] The results present
multiple types of immune cells in degenerated intervertebral discs
(Figure [73]1A). Furthermore, the prediction of interaction networks
between different cell types identified macrophages as being involved
in numerous interactions with other cell types, particularly with NP
cells (Figure [74]1B). To validate these findings, human nucleus
pulposus tissue samples with varying Pfirrmann grades (Grade II and
Grade V) were collected and the expression of macrophage marker CD11b
and pain neurotransmitter calcitonin gene‐related peptide (CGRP) was
assessed (Figure [75]1C). The results indicated a significant
upregulation of CD11b and CGRP expression in Grade V samples in
comparison to Grade II samples (Figure [76]1D and Figure [77]S1,
Supporting Information). Furthermore, immunohistochemistry revealed a
positive correlation between CD11b and CGRP expression levels and
Pfirrmann grades (Figure [78]1E). These findings indicate that with
intervertebral disc degeneration, there is an increase in the
infiltration of macrophages and neurons within the IVD.
Figure 1.
Figure 1
[79]Open in a new tab
Macrophage infiltration and polarization imbalance exacerbate IVDD and
discogenic pain. A) Uniform manifold approximation and projection
visualization showing cells inside NP tissues. B) Heatmap showing the
number of potential ligand‐receptor pairs between cell groups. C)
T2‐weighted magnetic resonance imaging (MRI) was utilized to procure
human nucleus pulposus tissues with different degeneration degrees
(Grade II and Grade V). D) Western blot analysis of CD11b and CGRP (n =
5). E) Immunohistochemical staining was performed to examine the
expression of CD11b and CGRP in different degenerating tissues (n = 5).
Scale bar: 100 µm. F) Experimental design, including establishment of a
rat discogenic pain model and subsequent histological and behavioral
pain assessments. G) Histological staining images of rat caudal spines
(n = 5). Scale bar: 1 mm. H) Western blot analysis of CGRP, CD86, and
CD206 (n = 5). I) Hargreaves tests detecting painful behavior in
response to heat stimulation of different groups (n = 5). J) Von Frey
tests detecting painful behavior in response to mechanical stimulation
of different groups (n = 5). Data are expressed as mean ± standard
deviation.
The findings were additionally validated in the Sprague‐Dawley rat
model. A disc puncture at the Co5‐Co6 level was performed in rats to
induce IVDD and associated pain behavior (Figure [80]1F).^[ [81]^21 ^]
The histological staining of the postoperative specimens demonstrated
the successful induction of the discogenic pain model (Figure [82]1G).
Subsequently, protein expression in the IVD was assessed, revealing
elevated expression of macrophage markers CD86, CD206, and pain
neurotransmitter CGRP in the puncture group relative to the Control and
sham groups. It is noteworthy that the M1 macrophage marker CD86
exhibited a notable increase in late‐stage IVDD, whereas changes in the
M2 macrophage marker CD206 were not statistically significant in the
early and late degeneration stages (Figure [83]1H and Figure [84]S2,
Supporting Information). Furthermore, the pain phenotype was evaluated
by administering Hargreaves (thermal pain threshold) and Von Frey
(mechanical pain threshold) tests. The findings indicated that on
postoperative day 3, the pain threshold slightly declined in the
puncture group relative to the sham‐operated group. 1 week following
surgery, the puncture group showed significantly reduced pain
thresholds compared to the other groups, which persisted for at least
28 d. This suggests that the puncture injury may have induced
temperature and mechanical hypersensitivity responses
(Figure [85]1I,J).
It has been demonstrated that activated M1 macrophages secrete
proinflammatory cytokines, neurotrophic factors, and matrix
metalloproteinases, which exacerbate neuroinflammatory pain and
extracellular matrix (ECM) degradation, thereby accelerating IVDD and
its associated symptoms.^[ [86]^8a ^] The aforementioned experimental
results substantiate the pivotal role of macrophages in the
pathogenesis of discogenic pain and underscore their potential as
therapeutic targets.
2.2. M2 Macrophages Transfer Mitochondria via Nanotubes to NP Cells and
Neurons
The ratio of M1 to M2 macrophages determines the fate of tissues during
inflammatory or injury processes. Anti‐inflammatory M2 macrophages have
a high capacity for tissue repair.^[ [87]^22 ^] To investigate the
therapeutic mechanism of M2 macrophages, we first induced macrophage M2
polarization by interleukin‐4 (Figure [88]2A). Subsequently, NP cells
that had been pre‐exposed to 100 µΜ H₂O₂ were directly cocultured with
M2 macrophages. To serve as a control, we established an indirect
coculture system utilizing Transwell chambers, which permitted the
passage of secreted signaling molecules while preventing direct
cell‐to‐cell communication (Figure [89]2B). Following a 16‐h coculture
period, NP cells were isolated using a fluorescence‐activated sorting
process and their status examined. The results demonstrated that both
coculture systems partially restored the H[2]O[2]‐induced decrease in
ATP levels and cell viability. Notably, the coculture group exhibited
superior recovery compared to the Transwell group. Similar outcomes
were also observed in neurons (Figure [90]S3, Supporting Information).
Figure 2.
Figure 2
[91]Open in a new tab
M2 macrophages transfer mitochondria via nanotubes to NP cells and
neurons. A) Flow cytometry analysis of F4/80 and CD206 on macrophages.
B) Experimental design of two different coculture systems.
F4/80‐labeled macrophages separated via fluorescence‐activated cell
sorting for subsequent coculture. C) Mitochondrial morphology changes
as observed by TEM. Scale bar: 500 nm. D) Quantitative analysis of
mitochondrial length in NP cells (n = 3). E) Quantitative analysis of
mitochondrial length in neurons (n = 3). F) SEM images showing
nanotubes (white dashed box) between Raw 264.7 (purple arrow) and NP
cells (white arrow). Scale bar: 5 µm. G) SEM images showing nanotubes
(white dashed box) between Raw 264.7 (purple arrow) and neurons (white
arrow). Scale bar: 10 µm. H) Confocal image showing nanotube‐mediated
mitochondrial communication between macrophages and NP cells. Scale
bar: 25 µm. I) Confocal image showing nanotube‐mediated mitochondrial
communication between macrophages and neurons. Scale bar: 50 µm. J)
Quantification of nanotube length in SEM images (n = 20). K)
Quantification of nanotube width in SEM images (n = 20). Data are
expressed as mean ± standard deviation.
Mitochondria represent the primary site of cellular ATP production and
ROS generation.^[ [92]^23 ^] Subsequently, the impact of distinct
coculture configurations on mitochondrial functionality was examined.
The fluorescence intensity of intracellular ROS was significantly lower
in the coculture group than in the Transwell group, as detected by the
DCFH‐DA probe (Figure [93]S4, Supporting Information). Furthermore,
transmission electron microscopy (TEM) images of mitochondria
demonstrated that both coculture models partially restored the
H₂O₂‐induced mitochondrial morphological alterations, including
swelling, vacuolization, reduction in cristae, and length shortening
(Figure [94]2C). However, NP cells and neurons in the coculture group
exhibited a healthier mitochondrial morphology with effective
restoration of mitochondrial length compared to the Transwell group
(Figure [95]2D,E).
Based on these experimental results, we hypothesized that there may be
cell contact‐dependent mechanisms for M2 macrophages to improve
mitochondrial morphology and function in tissue cells. Studies have
reported that macrophages can deliver healthy mitochondria to damaged
cells via TNTs, vesicles, and other means,^[ [96]^14a–c ^] so we used
scanning electron microscopy (SEM) to observe cellular interactions.
SEM images revealed nanoscale tubular junctions between macrophages, NP
cells, and neurons (Figure [97]2F,G and Figure [98]S5, Supporting
Information). These nanotubes were predominantly 4 to 11 µm long
(Figure [99]2J) and 110 to 450 nm wide (Figure [100]2K). The observed
number of nanotubes may be lower than the actual number because the
fragile nanotubes are susceptible to damage during sample preparation
and electron microscopy imaging (Figure [101]S6, Supporting
Information). Interestingly, nanotubes oriented toward NP cells were
slightly longer than those oriented toward neurons, while their widths
were reversed. These findings further support the hypothesis that M2
macrophages are involved in cellular communication through direct cell
contact. Mito Tracker Red (+) bone marrow‐derived macrophages (BMDM)
were cocultured with Mito Tracker Red (−) NP cells or neurons. After
16 h, the cytoskeleton was labeled with ghost pen cyclic peptide and
observed under a confocal microscope. Mitochondrial punctate
fluorescence was observed in heterogeneous cellular channels composed
of the cytoskeleton (Figure [102]2H,I). Co‐localization of mitochondria
with nanotubes confirmed nanotube‐mediated mitochondrial communication
between macrophages and NP cells/neurons. These results suggest that
mitochondrial transfer between cells under specific pathological
conditions is a non‐specific process, mainly influenced by changes in
cellular status and environmental conditions, and is not limited by
cell type.
2.3. Mitochondrial Transfer Regulates the Mitochondrial Function of Recipient
Cells
Using the following staining strategy (Mito Tracker Red to label
mitochondria in macrophages, Cell Trance Green to label NP cells and
neurons), we quantified the direction and efficiency of mitochondrial
transfer under different coculture times and modes by flow cytometry
(Figure [103]3A). Initially, M2 macrophages and NP cells/neurons were
labeled in red and green clusters, respectively. Over time, we observed
double‐positive cell clusters' appearance and gradual increase,
indicating mitochondrial transfer from M2 macrophages to NP cells and
neurons (Figure [104]3B,C). After 16 h of coculture, 11.3% of
double‐positive cells were found in NP cells and 17.0% in neurons. The
higher transfer efficiency in neurons can be attributed to their
abundant synapses, which promote the formation of mitochondrial
transfer channels.^[ [105]^24 ^] In contrast, no significant change in
the percentage of double‐positive cells was observed in Transwell
chambers, suggesting that mitochondrial transfer depends on direct
cell‐to‐cell contact. To exclude the possibility of non‐specific
staining due to dye leakage, we added conditioned media from Mito
Tracker Red (+) M2 macrophages and Mito Tracker Red (+) M2 macrophages
to parallel cultures of NP cells. The results showed that the addition
of M2 macrophages, rather than the addition of conditioned medium,
resulted in the appearance of Mito Tracker Red (+) NP cells,
demonstrating that the mitochondrial signals in the recipient cells
were derived from exogenous mitochondria rather than non‐specific dye
leakage (Figure [106]S7, Supporting Information).
Figure 3.
Figure 3
[107]Open in a new tab
Mitochondrial transfer regulates the mitochondrial function of
recipient cells. A) Pseudocolor images showing the proportion of
double‐positive cells in different coculture systems at various time
points (n = 3). B) Quantitative analysis of Mito Tracker (+) NP cell
percentage over time (n = 3). C) Quantitative analysis of Mito Tracker
(+) neuron percentage over time(n = 3). D) Representative images of
distribution and morphology of mitochondria after different treatments
(n = 3). Scale bar: 10 µm. E) ImageJ analysis of the area of
mitochondria. F) Representative fluorescent images of JC‐1 staining
under different treatment conditions (n = 3). Scale bar: 100 µm. G)
Quantitative analysis of JC‐1 levels under different treatment
conditions (n = 3). H) Flow cytometry apoptosis analysis of NP cells
and neurons under different treatment conditions (n = 3). I)
Quantitative analysis of apoptosis rate under different treatment
conditions (n = 3). EB, Ethidium bromide. Data are expressed as mean ±
standard deviation.
To investigate the extent to which different types of macrophages are
involved in mitochondrial transfer, we conducted additional flow
cytometry experiments to evaluate the mitochondrial transfer capacity
of M0, M1, and M2 macrophages. The results demonstrated that M0 and M1
macrophages exhibited minimal or no mitochondrial transfer to damaged
NP cells and neurons. In contrast, M2 macrophages displayed a
significantly higher level of mitochondrial transfer activity (Figure
[108]S8, Supporting Information). This observation suggests that
mitochondrial transfer is a unique feature of M2 macrophages, which
aligns with their well‐established biological role in tissue repair,
regeneration, and anti‐inflammatory responses. The polarization state
of macrophages likely determines their mitochondrial transfer capacity,
as M2 macrophages are associated with enhanced intercellular
communication and metabolic support for damaged cells. By comparison,
M1 macrophages, which are more pro‐inflammatory in nature, may
prioritize immune defense mechanisms over intercellular material
exchange.
Additionally, we performed experiments using the opposite labeling
strategy to investigate whether mitochondria could be transferred back
from NP cells or neurons to macrophages. Specifically, we pre‐labeled
the mitochondria in NP cells or neurons before coculturing them with
macrophages. Flow cytometry was then used to detect the presence of
exogenous mitochondria in macrophages after coculture. The results
showed no significant presence of exogenous mitochondria within
macrophages, indicating that NP cells or neurons do not transfer their
mitochondria back to macrophages (Figure [109]S9, Supporting
Information). These findings suggest that the mitochondrial transfer
observed in this study is unidirectional, from macrophages to recipient
cells, and does not affect macrophages through reverse transfer.
Subsequently, the impact of mitochondrial transfer on mitochondrial
function in recipient cells was investigated. Compared to the Transwell
group, the coculture group showed a healthier mitochondrial morphology
(Figure [110]3D). In particular, mitochondrial area and circumference
were effectively restored. (Figure [111]3E and Figure [112]S10,
Supporting Information). Excess ROS from oxidative stress disrupts the
mitochondrial membrane potential (ΔΨm), leading to mitochondrial
depolarization and dysfunction.^[ [113]^25 ^] Therefore, we examined
the changes in ΔΨm in recipient cells under different treatment
conditions. A higher red/green fluorescence ratio represents a normal
ΔΨm. Both culture modes partially increased the red/green fluorescence
ratio compared to the H[2]O[2] group, while the coculture group showed
a more significant improvement (Figure [114]3F,G). As the reduction in
ΔΨm is an early indicator of apoptosis,^[ [115]^26 ^] we investigated
the apoptosis rate further. The results showed that coculture reduced
the apoptosis rate by 12.3% in H[2]O[2]‐induced NP cells and 13.66% in
neurons (Figure [116]3H,I). However, pre‐disruption of macrophage
mitochondria with ethidium bromide (0.5 µM) blocked the recovery of
mitochondrial function and apoptosis in recipient cells. This suggests
that the additional therapeutic effect mediated by direct contact
largely depends on functional mitochondrial delivery rather than other
cellular mechanisms associated with direct contact. Thus, improving
mitochondrial function in donor cells may contribute to improved
recovery of recipient cells.
We also conducted experiments using the microtubule inhibitor
Nocodazole, which has been shown in previous studies to suppress the
formation of TNTs and thereby inhibit mitochondrial transfer.^[
[117]^27 ^] Our SEM images demonstrated that the Nocodazole‐treated
group's nanotube connections between macrophages and NP cells or
neurons were significantly reduced (Figure [118]S5, Supporting
Information). This provides direct evidence that microtubule inhibitors
effectively inhibit TNTs formation. Additionally, we evaluated the
mitochondrial function of recipient cells. We observed that the
addition of Nocodazole not only inhibited mitochondrial transfer but
also impaired the ability of M2 macrophages to restore mitochondrial
function in damaged cells (Figure [119]S11, Supporting Information).
These findings strongly support the conclusion that TNTs are a key
pathway for mitochondrial transfer and that the transferred
mitochondria are essential in restoring mitochondrial function in
recipient cells.
2.4. Preparation, Characterization, and ROS Scavenging Capability of
PGA‐Cu‐S@G
In an alkaline aqueous solution, multiple adjacent phenolic hydroxyl
groups of gallic acid (GA) were used to synthesize metal polyphenol
nanoparticles (PGA‐Cu) through self‐assembly. Meanwhile, the surface of
the nanoparticles contained multiple aldehyde groups that could form
Schiff bases with abundant amino groups in two functional peptides
(SS05 and GAP134). Finally, PGA‐Cu‐SS05‐GAP134 (PGA‐Cu‐S@G) was formed
for further functionalization (Figure [120]4A). The final binding rate
of SS05 is 14.74%, while the binding rate of GAP134 is 15.97% (Figure
[121]S12, Supporting Information). PGA‐Cu‐S@G is well dispersed in
solution with an average hydrodynamic diameter of about 68.1 nm as
measured by dynamic light scattering. The zeta potential of PGA‐Cu‐S@G
(−19.4 mV) is significantly higher than that of PGA alone (−41.5 mV)
and PGA‐Cu measured at −33.3 mV due to doping with SS05 (2.6 mV) and
GAP134 (3.3 mV) (Figure [122]4B and Figure [123]S13, Supporting
Information). SEM and TEM also confirmed that PGA‐Cu‐S@G is a spherical
particle with a diameter close to 70 nm (Figure [124]4C and Figure
[125]S14, Supporting Information), composed mainly of carbon (C),
oxygen (O), and copper (Cu) elements (Figure [126]4D), which is
attributed to the coordination of Cu^2+ with the abundant oxygen groups
on the PGA surface. X‐ray photoelectron spectroscopy was then performed
to investigate the surface composition. The survey spectra confirmed
that PGA‐Cu‐S@G is mainly composed of carbon, nitrogen, oxygen, and
copper elements (Figure [127]S15, Supporting Information). Gaussian
fitting of the high‐resolution Cu 2p spectra showed that copper is
present in monovalent (Cu I) and divalent (Cu II) forms, with a ratio
of Cu I to Cu II of ≈1:1. However, the intensity of the Cu I peaks
decreased. The Cu II peaks increased after H[2]O[2] treatment. The
ratio of the Cu I/Cu II peaks became 2:3. In contrast, the
characteristic peaks of the other elements did not change
significantly, indicating the importance of the Cu I/Cu II valence
transition in the redox process of H[2]O[2] (Figure [128]4E). Studies
suggest that polyphenolic compounds may enhance the catalytic effect by
promoting the valence transition of copper ions,^[ [129]^28 ^] thus
improving the free radical scavenging efficiency and providing
important theoretical support for PGA‐Cu‐S@G as an effective
antioxidant.
Figure 4.
Figure 4
[130]Open in a new tab
Preparation, characterization, and ROS scavenging capability of
PGA‐Cu‐S@G. A) Schematic representation for the synthesis of
PGA‐Cu‐S@G. B) Size distribution of PGA‐Cu‐S@G and zeta potentials of
different materials. C) SEM image of PGA‐Cu‐S@G. Scale bar: 50 nm. D)
TEM (Scale bar: 25 nm) and corresponding energy dispersive X‐ray
spectroscopy for PGA‐Cu‐S@G. Scale bar: 50 nm. E) High‐resolution
spectra of PGA‐Cu‐S@G for Cu 2p before and after H[2]O[2] treatment. F)
FTIR analysis of characteristic peaks for SS05, GAP134, and PGA‐Cu‐S@G.
G) XRD of PGA‐Cu and PGA‐Cu‐S@G. H‐K) UV−vis absorbance spectra showing
the radical eliminating activities of PGA‐Cu‐S@G for H) DPPH·, I)
H[2]O[2], J) ·OH, and K) ·O[2] ^− in 0.5 h (n = 3). L) DPPH·, M)
H[2]O[2] and N) ·OH temporal scavenging efficiency of PGA‐Cu‐S@G with
various concentrations (n = 3). O) ·O[2] ^− scavenging efficiency of
PGA‐Cu‐S@G with various concentrations (n = 3). Data are expressed as
mean ± standard deviation.
To quantify the copper content, we performed ICP‐MS analysis on the
PGA‐Cu‐S@G samples, which showed that copper accounted for 5.68% of the
total mass of the nanoparticles, confirming the effective doping of
Cu^2⁺ (Table [131]S1, Supporting Information). Fourier transform
infrared spectroscopy (FTIR) analysis identified the characteristic
peaks of SS05 and GAP134, confirming the presence of the main peaks in
PGA‐Cu‐S@G (Figure [132]4F). The characteristic peaks of the C═N
stretching vibration appeared in the range of 1620–1650 cm^−1,
indicating that the peptide amino group reacted with the aldehyde group
of GA to form Schiff bases, which were effectively incorporated into
the nanoparticles. The C═C stretching vibration in poly GA has a
characteristic absorption peak at 1590 cm^−1, and the unreacted
carbonyl group has a C═O stretching vibration peak near 1710 cm^−1. The
characteristic peak at 1550 cm^−1 corresponds to the amide II band of
the peptide, and the C‐N stretching vibration peak may appear in the
region of 1280 cm^−1. X‐ray diffraction (XRD) analysis showed that most
of the characteristic peaks of GA disappeared in PGA‐Cu‐S@G and PGA‐Cu,
probably due to the formation of amorphous structure by oxidative
polymerization of GA (Figure [133]4G and Figure [134]S16, Supporting
Information). Moreover, the diameter of PGA‐Cu‐S@G maintained good
stability over a long period of time (Figure [135]S17, Supporting
Information), providing a reliable basis for its potential biological
applications.
Mitochondrial dysfunction is closely associated with sustained high
levels of ROS. Therefore, the scavenging of excessive ROS is crucial
for PGA‐Cu‐S@G to restore mitochondrial function. To comprehensively
investigate the ROS scavenging capacity of PGA‐Cu‐S@G, we selected
three major members of the ROS family – hydrogen peroxide (H[2]O[2]),
hydroxyl radicals (∙OH), superoxide anions (∙O[2] ^−) and indicators
reflecting total antioxidant capacity DPPH∙ and ABTS^+∙ for detection.
The clearance of DPPH∙ was 96.11% after 1 h at 100 µg mL^−1
concentration (Figure [136]4H,L); in ABTS^+∙ analysis, the clearance
efficiency was 98.44% after 1 h at 25 µg mL^−1 concentration (Figure
[137]S18, Supporting Information). PGA‐Cu‐S@G showed high sensitivity
to ∙OH, with 84.65% clearance after 1 h at 12.5 µg mL^−1 concentration;
at 100 µg mL^−1 concentration, it was almost completely cleared in
30 min (Figure [138]4J,N). Furthermore, the results of the SOD‐like
enzyme activity assay of PGA‐Cu‐S@G showed that the clearance
efficiency of ∙O[2] ^− was 74.03% at 100 µg mL^−1 concentration
(Figure [139]4K,O); while for CAT‐like enzyme activity, 93.72% of
H[2]O[2] removal was achieved at 100 µg mL^−1 concentration for 4 h
(Figure [140]4I,M). These results demonstrate that the coordination
polymerization of PGA with copper significantly enhances its
antioxidant capacity. PGA‐Cu‐S@G shows excellent in vitro performance
in scavenging ROS.
2.5. Cellular Internalization, Lysosomal Escape, and Mitochondrial Targeting
of PGA‐Cu‐S@G
After confirming the in vitro capabilities of the nanoparticles, we
investigated their intracellular properties using cell counting kit‐8
(CCK‐8) to evaluate the toxicity of the nanoparticles on NP cells,
neurons, and macrophages. The results showed that peptide modification
did not increase the cytotoxicity of PGA‐Cu, and there was no
significant effect on cell viability at concentrations of 100 µg mL^−1
and below. To assess biocompatibility, we cocultured 100 µg mL^−1
PGA‐Cu‐S@G with the three cell types for 1, 3, and 5 d. Fluorescence
images showed that the cell growth rate was not affected. The hemolysis
assay also confirmed the safety of PGA‐Cu‐S@G at this concentration
(Figure [141]S19, Supporting Information). Therefore, we chose
100 µg mL^−1 as a safe concentration for further studies.
The affinity of various materials for mitochondria was assessed to
reflect their targeting ability. Fluorescein isothiocyanate
(FITC)‐labeled materials were co‐incubated with free mitochondria for
1 h at 37 °C. The intensity of FITC fluorescence within the
mitochondria was detected by flow cytometry (Figure [142]5A). The
results showed that the mitochondrial affinity of PGA‐Cu‐S@G was second
only to that of SS05, which was attributed to the alternating
arrangement of aromatic and basic amino acids in SS05, which confers
hydrophobicity and lipophilicity. Thus, PGA‐Cu‐S@G effectively targets
mitochondria by electrostatic adsorption independent of the ΔΨm. In
addition, the surface modification of the peptide made the
nanoparticles positively charged surfaces, which increased the affinity
of PGA‐Cu‐S@G for negatively charged cells and promoted cellular
uptake. We then investigated the differences in the uptake of
PGA‐Cu‐S@G by macrophages, NP cells, and neurons. To mimic the
pathological microenvironment of the disc, the three cell types were
cocultured and distinguished using different staining strategies. After
16 h of co‐incubation with FITC‐labeled PGA‐Cu‐S@G (PGA‐Cu‐S@G^FITC),
cellular uptake rates were quantified by flow cytometry
(Figure [143]5B). The results showed that the intensity of FITC
fluorescence within macrophages was more than ten times higher than
that of other cells, indicating an absolute dominance of their uptake
of PGA‐Cu‐S@G (Figure [144]5C,D). This may be attributed to the strong
phagocytic ability of macrophages.^[ [145]^29 ^] As the co‐incubation
time increased, the FITC fluorescence intensity in macrophages rose
rapidly in the first 8 h and then gradually stabilized
(Figure [146]5E,F).
Figure 5.
Figure 5
[147]Open in a new tab
Cellular internalization, lysosomal escape, and mitochondrial targeting
of PGA‐Cu‐S@G. A) Evaluation of the binding ability of different
materials labeled with FITC to mitochondria (n = 3). B) Experimental
design of staining strategies for 3 different cell types. C) Gating
strategies of 3 different cell types. Hoechst (+) and Dil (‐) cells
were defined as macrophages. Hoechst (−) and Dil (+) cells were defined
as neurons. Hoechst (−) and Dil (−) cells were defined as NP cells.
Flow cytometry was used to measure fluorescence intensity in the FITC
channel, reflecting the rate of PGA‐Cu‐S@G internalization by different
cells. D) Quantification of FITC mean fluorescence intensity in 3
different cell types (n = 3). E) Flow cytometry analysis of
PGA‐Cu‐S@G^FITC uptake and retention in macrophages over time. F)
Quantification of the temporal uptake rate of PGA‐Cu‐S@G^FITC by
macrophages (n = 3). G) TEM image showing the localization of
PGA‐Cu‐S@G in macrophage mitochondria. Scale bar: 200 µm. H) Confocal
image showing Co‐localization of PGA‐Cu‐S@G^FITC with Mito
Tracker‐labeled mitochondria. Scale bar: 5 µm. I) Confocal image
showing Co‐localization of PGA‐Cu‐S@G^FITC with Lyso Tracker‐labeled
lysosome. Scale bar: 5 µm. J) Co‐localization of PGA‐Cu‐S@G^FITC with
Mito Tracker‐labeled mitochondria. K) Co‐localization of
PGA‐Cu‐S@G^FITC with Lyso Tracker‐labeled lysosome. Data are expressed
as mean ± standard deviation.
The subcellular localization of nanoparticles has a significant impact
on their therapeutic efficacy.^[ [148]^30 ^] Lysosomes present a
challenge to the intracellular functioning of nanoparticles.^[ [149]^31
^] It has been demonstrated that nanoparticles with pH‐buffering
capacity facilitate the influx of hydrogen ions, chloride ions, and
water into the endosome prior to fusion with lysosomes, which results
in endosome rupture and prevents lysosomal degradation of
nanoparticles.^[ [150]^29 , [151]^32 ^] PGA‐Cu‐S@G exhibits a robust
pH‐buffering capacity, which is primarily attributed to its high
concentration of carboxylate groups that dissociate within the endosome
and enhance the escape ability. To evaluate the lysosomal escape and
mitochondrial targeting capabilities of PGA‐Cu‐S@G, we employed a
dual‐labeling approach, wherein PGA‐Cu‐S@G was labeled with FITC and
mitochondria and lysosomes of BMDM were labeled with Mito Tracker Red
and Lyso Tracker Red, respectively. The degree of co‐localization of
PGA‐Cu‐S@G with organelles was quantified using Pearson's R‐value. The
results demonstrated a notable overlap between the fluorescence signals
of PGA‐Cu‐S@G and mitochondria (Pearson's r = 0.68) in comparison to
lysosomes (Pearson's r = 0.34) (Figure [152]5H,I). Co‐localization
quantification maps further demonstrated that the fluorescence of
PGA‐Cu‐S@G exhibited a strong correlation with mitochondrial
fluorescence, whereas the correlation with lysosomal fluorescence was
relatively weak (Figure [153]5J,K). Additionally, TEM images revealed
that PGA‐Cu‐S@G was situated within the mitochondrial matrix of
macrophages (Figure [154]5G). These findings suggest that PGA‐Cu‐S@G
can effectively escape from lysosomes and target mitochondria.
2.6. PGA‐Cu‐S@G Regulates Macrophage Polarization by Scavenging mtROS
The capacity of PGA‐Cu‐S@G to scavenge mtROS to modulate macrophage
polarization was subsequently evaluated. To simulate a hypoxic IVDD
environment, Raw 264.7 cells were exposed to 100 µM H₂O₂ for 6 h. The
effect was then assessed by applying PGA‐Cu‐S@G. MitoSox was used to
detect the mtROS level, and the results demonstrated that the mtROS
level in the PGA‐Cu‐S@G group was reduced to 60% of that in the H₂O₂
group (Figure [155]6A,B). Furthermore, flow cytometry showed that
PGA‐Cu‐S@G reduced the proportion of H₂O₂–induced MitoSox (+) cells
from 40.7% to 23.8% (Figure [156]S20, Supporting Information). These
results demonstrated that PGA‐Cu‐S@G effectively scavenges
extracellular ROS and mtROS. This capacity was also corroborated in NP
cells and neurons (Figure [157]S20, Supporting Information).
Figure 6.
Figure 6
[158]Open in a new tab
In vitro therapeutic evaluation of PGA‐Cu‐S@G. A) Representative
fluorescent images of Mito‐Sox staining under different treatment
conditions (n = 3). Scale bar: 100 µm. B) Quantitative analysis of
Mito‐Sox fluorescence intensity (n = 3). C) Flow cytometry analysis of
CD86 and CD206 (n = 3). D) Quantitative analysis of mitochondrial
transfer efficiency (n = 3). E) Flow cytometry analysis of JC‐1 levels
under different treatment conditions (n = 3). F) Quantitative analysis
of JC‐1 levels within NP cells under different treatment conditions
(n = 3). G) Quantitative analysis of JC‐1 levels within neurons under
different treatment conditions (n = 3). H) Immunofluorescence staining
of COL2A1 in NP cells (n = 3). Scale bar: 100 µm. I) Quantitative
analysis of COL2A1 fluorescence intensity (n = 3). J)
Immunofluorescence staining of CGRP in neurons (n = 3). Scale bar:
100 µm. K) Quantitative analysis of CGRP fluorescence intensity (n =
3). Data are expressed as mean ± standard deviation.
The inhibition of ROS generation or the removal of excess ROS has been
identified as an effective strategy for modulating macrophage
phenotype.^[ [159]^33 ^] Accordingly, the impact of PGA‐Cu‐S@G on
macrophage polarization was subjected to further examination.
Immunofluorescence was conducted to ascertain the mean fluorescence
intensity of macrophage polarization markers (CD86 and CD206). The
results demonstrated that H₂O₂ treatment resulted in augmented M1
polarization and diminished M2 polarization. In contrast, treatment
with PGA‐Cu‐S@G decreased fluorescence intensity for CD86 and increased
fluorescence intensity for CD206, which was observed in both Raw 264.7
and BMDM (Figure [160]S21, Supporting Information). Macrophage
polarization‐related indexes were detected by flow cytometry
(Figure [161]6C). The data demonstrated that the proportion of F4/80
(+) and CD86 (+) cells in the PGA‐Cu‐S@G group was 13%, which was 19.4%
less than that observed in the H₂O₂ group. Conversely, the proportion
of F4/80 (+) and CD206 (+) cells (12.3%) was higher than that of the
control group (0.096%) and the H₂O₂ group (4.77%). Furthermore, the
results of the qRT‐PCR demonstrated that the nanoparticles markedly
reduced the expression of M1 macrophage markers (CD80, CD86, iNOS)
while concurrently elevating the expression of M2 macrophage markers
(CD163, CD206, Arginase) in comparison to the H₂O₂ group (Figure
[162]S21, Supporting Information). These findings provide further
validation that PGA‐Cu‐S@G has the capacity to modulate macrophage
polarization phenotypes, which is attributed to its effective
scavenging of mtROS.
2.7. PGA‐Cu‐S@G aids M2 Macrophages in Repairing Damaged Cells via Efficient
Mitochondrial Transfer
As previously described, TNT is a tubular membrane protrusion that
facilitates non‐adjacent cell communication.^[ [163]^34 ^] M2
macrophages transfer mitochondria to damaged neurons and NP cells via
TNTs. The efficiency of mitochondrial transfer represents a significant
limiting factor in the therapeutic efficacy of M2 macrophages. GAP134,
a small‐molecule GJC modifier, has been demonstrated to enhance Cx43
expression, thereby promoting GJC formation.^[ [164]^35 ^] This, in
turn, affects the formation and function of TNTs by modulating the
extended terminals that constitute TNTs.^[ [165]^19 , [166]^36 ^] The
capacity of PGA‐Cu‐S@G to stimulate Cx43 expression in macrophages was
initially investigated. Macrophages were co‐incubated with PGA‐Cu‐S@G
and PGA‐Cu@S (PGA‐Cu nanoparticles modified with SS05 only) for 8 h,
and the expression of Cx43 was assessed via western blot. The results
demonstrated that PGA‐Cu‐S@G elevated the expression of Cx43 in
comparison to the control group, whereas PGA‐Cu@S exhibited no notable
impact on Cx43 expression (Figure [167]S22, Supporting Information).
Subsequently, the efficiency of macrophage mitochondrial transfer was
evaluated through a quantitative flow cytometry assessment. Macrophages
that had been pre‐labeled with Mito Tracker Green for mitochondrial
labeling were cocultured with NP cells/neurons for 8 h. Macrophages
were specifically conjugated with an F4/80 antibody (labeled with APC).
The proportion of Mito Tracker Green (+) cells in the APC (‐) cell
population was defined as the efficiency of mitochondrial transfer
through a gating strategy (Figure [168]S23, Supporting Information).
The results demonstrated that the proportion of Mito Tracker Green (+)
cells in the PGA‐Cu‐S@G group was markedly higher than that in the
other groups. In comparison to the control group, there was a 15%
increase in NP cells and an 18.8% increase in neurons, while the
addition of nocodazole significantly reduced the proportion of Mito
Tracker Green (+) cells (Figure [169]6D and Figure [170]S23, Supporting
Information). These findings suggest that PGA‐Cu‐S@G enhances the
efficiency of mitochondrial transfer by promoting TNTs formation
through the induction of Cx43 expression.
The subsequent objective was to ascertain whether macrophages treated
with PGA‐Cu‐S@G (PGA‐Cu‐S@G‐Mac) demonstrated augmented mitochondrial
recuperation (Figure [171]6E). The results of the ΔΨm analysis
indicated that the nanoparticles partially reversed the
H[2]O[2]‐induced decrease in ΔΨm. Notably, the Control‐Mac group
exhibited the most pronounced ΔΨm depolarization, which may be
attributed to the M1‐polarizing tendency of macrophages due to
inflammatory factors and ROS secreted by NP cells and neurons under
oxidative stress.^[ [172]^37 ^] Furthermore, the PGA‐Cu‐S@G‐Mac group
showed superior recovery compared to the PGA‐Cu@S‐Mac group, with ΔΨm
of NP cells recovered to 61.7% of the control group and ΔΨm of neurons
recovered to 52.2% (Figure [173]6F,G). Subsequently, the expression of
functional proteins in NP cells and neurons was assessed. The ECM
synthesis protein COL2A1 reflects the regulation of ECM metabolic
homeostasis in NP cells, whereas the pain neurotransmitter CGRP
reflects the transmission of pain signals in neurons. The results
indicated a notable reduction in the fluorescence intensity of COL2A1
in the H[2]O[2] and Control‐Mac groups, whereas the fluorescence
intensity of CGRP exhibited an increase in both groups when compared to
the control group. In contrast, both nanoparticles increased COL2A1
expression (Figure [174]6H,I) and decreased CGRP expression
(Figure [175]6J,K), suggesting an effective alleviation of oxidative
stress‐induced ECM remodeling and pain hypersensitivity. NP cells and
neurons in the PGA‐Cu‐S@G‐Mac group exhibited the most significant
functional recovery. These findings were corroborated by qRT‐PCR
results (Figure [176]S24, Supporting Information). The PGA‐Cu‐S@G‐Mac
group exhibited the most pronounced restoration of the ECM synthesis
function of NP cells and a marked reduction in the synthesis and
expression of pain mediators and pain‐related cation channels. The
findings indicate that PGA‐Cu‐S@G exerts anti‐inflammatory effects by
inducing M2 polarization and facilitating efficient and sustained
mitochondrial delivery by macrophages, ultimately improving
mitochondrial homeostasis within recipient cells.
2.8. In Vivo Therapeutic Evaluation of PGA‐Cu‐S@G
To provide further in vivo evidence for mitochondrial transfer, we
conducted an experiment in which we injected Tom20‐GFP‐labelled M2
macrophages derived from rats into the IVD of IVDD rats. After a 48 h
incubation period, we extracted the rat IVD and dorsal root ganglion
cells for flow cytometry analysis.^[ [177]^14 , [178]^38 ^] For cell
identification, we used CD86 and CD24 as markers for NP cells and MAP2
and NeuN as markers for neurons. The results showed that
Tom20‐GFP‐positive cells accounted for 16.08% of NP cells and 9.98% of
neurons in the gated double‐positive cell population. Notably, when an
additional microtubule inhibitor, Nocodazole, was injected, the
proportion of Tom20‐GFP‐positive subset in the NP cells and neurons
decreased to 9.31% and 4.19%, respectively (Figure [179]S25, Supporting
Information). These findings provide direct evidence of mitochondrial
transfer from macrophages to damaged cells in vivo and suggest that
TNTs play a crucial role in this mitochondrial transfer process.
In parallel, an IVDD model was established in Sprague‐Dawley rats
(Figure [180]7A), and different nanoparticles were injected in situ
into the diseased segments using a microinjector. The efficacy of the
treatment was then assessed radiologically, histologically, and by pain
behavior (Figure [181]7B). In vivo real‐time imaging analysis
demonstrated that PGA‐Cu‐S@G was predominantly localized within the IVD
and persisted for up to 13 days following a single in situ injection
(Figure [182]S26, Supporting Information).
Figure 7.
Figure 7
[183]Open in a new tab
In vivo therapeutic evaluation of PGA‐Cu‐S@G. A) Intraoperative
procedures for IVDD. B) Schematic illustration of animal experiments.
C) X‐ray images of rat coccygeal vertebrae after different treatments
(n = 5). D) DHI changes in different groups from 4 to 8 weeks after
surgery (n = 5). E) MRI images of rat coccygeal vertebrae after
different treatments (n = 5). F) Pfirrmann grade changes of different
groups from 4 to 8 weeks after surgery (n = 5). G) Hargreaves tests
detecting painful behavior in response to heat stimulation of different
groups (n = 5). H) Von Frey tests detecting painful behavior in
response to mechanical stimulation of different groups (n = 5). I) HE
staining images of rat caudal spines (n = 5). Scale bar: 1 mm. J) SO
staining images of rat caudal spines (n = 5). Scale bar: 1 mm. K)
Immunohistochemistry of COL2A1 and MMP13 at 8 weeks after surgery (n =
5). Scale bar: 1 mm. L) Immunohistochemistry of CD86 and CD206 at 8
weeks after surgery (n = 5). Scale bar: 1 mm. Data are expressed as
mean ± standard deviation.
Analysis of the radiographs showed a significant decrease in the disc
height index (DHI) in the IVDD group. In addition, significant
osteophytes were observed on both sides of the IVD. In contrast, the
PGA‐Cu‐S@G group showed the smallest decrease in DHI, which was close
to the control group (Figure [184]7C). MRI is considered the gold
standard for the diagnosis of IVDD. According to the Pfirrmann
classification, MRI images are graded from grade I to grade V based on
signal intensity (Table [185]S2, Supporting Information). The IVDD
group showed a significant decrease in disc signal intensity with
concomitant bone collapse and fusion. The PGA‐Cu‐S@G group had slightly
higher disc signal intensity than the IVDD group but still lower than
the PGA‐Cu‐S@G group (Figure [186]7E). Overall, the changes in DHI and
MRI grading were consistent (Figure [187]7D,F). The results of the pain
behavior tests demonstrated that the thermal and mechanical pain
thresholds of the IVDD group were significantly lower than the baseline
levels and exhibited a declining trend over time. Both nanoparticles
demonstrated the capacity to impede the decline in pain threshold.
However, the PGA‐Cu‐S@G group showed a more pronounced efficacy in
alleviating nociceptive sensitization in comparison to the PGA‐Cu@S
group, facilitating the recovery of thermal pain thresholds as early as
postoperative day 3 (Figure [188]7G,H).
The results were further corroborated by histologic analysis. In the
tissue bulk map, the morphology of NP tissues in the PGA‐Cu‐S@G group
exhibited the greatest similarity to that of the control group (Figure
[189]S27, Supporting Information). The cellular structure and
morphology of the IVD were observed using hematoxylin and eosin (HE)
staining (Figure [190]7I). Over the course of 4 to 8 weeks, it was
observed that the NP cells in the IVDD group underwent gradual
replacement by fibroblasts, accompanied by disruption of the boundary
between the AF and NP. However, in the PGA‐Cu‐S@G group, the number of
NP cells exhibited only slight variation, and the tissue edges remained
distinct. The collagen content of the IVDs was evaluated through the
use of the Safranin‐O/Fast Green (SO) staining method (Figure [191]7J).
Both nanoparticles demonstrated the capacity to impede the denaturation
of the NP. However, the NPs in the PGA‐Cu‐S@G group demonstrated a
greater capacity for proteoglycan and collagen enrichment. In
conjunction with the histologic grading scale (Table [192]S3,
Supporting Information), the PGA‐Cu‐S@G group exhibited the most
optimal disc morphology and structural repair (Figure [193]S28,
Supporting Information). Immunohistochemistry also demonstrated that
PGA‐Cu‐S@G more effectively regulated ECM metabolism (Figure [194]7K)
and reversed macrophage polarization ratio (Figure [195]7L). Notably,
there was no significant difference in immunomodulation between the two
types of nanoparticles (Figure [196]S29, Supporting Information). This
suggests that the difference in efficacy between the two nanoparticles
is not due to the M2 macrophage‐mediated anti‐inflammatory effect but
rather the difference in mitochondrial transfer efficiency.
To more accurately simulate the progression and therapeutic effects of
IVDD in humans, the Bama pig was selected as a large animal model. The
surgical procedure is described in detail in the Methods section. At 8
weeks post‐surgery, MRI revealed that the PGA‐Cu@S group exhibited
heightened disc signal intensity relative to the IVDD group, though it
was slightly lower than that observed in the PGA‐Cu‐S@G group (Figure
[197]S30, Supporting Information). These results indicate that
PGA‐Cu‐S@G maintains excellent therapeutic efficacy in a large animal
model. Furthermore, HE staining was employed to assess the extent of
inflammatory infiltration and tissue damage in vital organs. The
findings demonstrated that the nanoparticles exhibited favorable
biocompatibility (Figure [198]S31, Supporting Information).
2.9. Therapeutic Mechanism of PGA‐Cu‐S@G
The etiology of IVDD is complex and multifactorial, and further
elucidation of the specific molecular mechanisms of PGA‐Cu‐S@G in the
treatment of discogenic pain is necessary. Transcriptomic sequencing
analysis was conducted on macrophages treated with PGA‐Cu‐S@G. The
treatment resulted in the up‐regulation of 3763 genes and the
down‐regulation of 3806 genes in comparison to the H[2]O[2] group
(Figure [199] 8A). The data were subjected to heat map analysis for
visualization (Figure [200]8B). H[2]O[2] induces cellular oxidative
stress and activates various stress‐related signaling pathways, leading
to significant changes in gene expression. PGA‐Cu‐S@G treatment
reversed most of the changes in gene expression induced by hydrogen
peroxide, suggesting that the nanoparticles may be useful in
alleviating oxidative stress. The Kyoto Encyclopedia of Genes and
Genomes (KEGG) enrichment analysis indicated that the differentially
expressed genes (DEGs) were predominantly associated with metabolic and
inflammatory pathways. With regard to metabolic processes, notable
discrepancies were discerned in oxidative phosphorylation,
Tricarboxylic acid cycling, and fatty acid metabolism between the
H[2]O[2] and PGA‐Cu‐S@G groups. This suggests that nanoparticles may
regulate macrophage phenotypes by modulating these processes. With
regard to the inflammatory response, the inhibition of the NF‐𝜅B
signaling pathway and the up‐regulation of the PI3K‐Akt signaling
pathway were of particular note. These changes may be related to the
increase in the number of M2 macrophages induced by PGA‐Cu‐S@G
(Figure [201]8C). Gene ontology (GO) analysis revealed that the DEGs
between the H[2]O[2] and PGA‐Cu‐S@G groups were primarily associated
with ribosomes and mitochondria. These organelles play a pivotal role
in regulating oxidative stress and protein synthesis, respectively.
Nanoparticles affected the immune phenotype and metabolic activity of
macrophages by modulating the function of these two organelles
(Figure [202]8D). Furthermore, the nanoparticle group exhibited a
notable increase in the function of cell‐cell junction assembly, which
may be associated with GAP134‐mediated promotion of Cx43 expression or
the nanoparticle‐induced expression of Cx43 through the JNK pathway.^[
[203]^16 , [204]^18 ^]
Figure 8.
Figure 8
[205]Open in a new tab
Therapeutic Mechanism of PGA‐Cu‐S@G. A) Volcano plot showing
differential gene expression in H[2]O[2] and PGA‐Cu‐S@G groups. p <
0.05, | log2(Fold Change) | ≥ 0. B) Differential gene enrichment maps
in H[2]O[2] and PGA‐Cu‐S@G groups. C) Differential gene pathway
enrichment analysis (KEGG analysis). D) GO analysis of differential
genes. E–G) GSEA enrichment analysis of oxidative phosphorylation,
NF‐𝜅B signaling pathway, and cell‐cell junction assembly between the
H[2]O[2] and PGA‐Cu‐S@G groups. H–J) qRT‐PCR analysis of genes
associated with oxidative phosphorylation, NF‐𝜅B signaling pathway, and
cell‐cell junction assembly (n = 3). Data are expressed as mean ±
standard deviation.
The gene set enrichment analysis (GSEA) demonstrated that PGA‐Cu‐S@G
was capable of upregulating oxidative phosphorylation (Figure [206]8E)
and downregulating the NF‐𝜅B signaling pathway (Figure [207]8F).
Furthermore, PGA‐Cu‐S@G was observed to enhance the function of
intercellular junction assembly, thereby demonstrating the potential to
promote mitochondrial transfer (Figure [208]8G). The results of the
GSEA analysis were further validated by qRT‐PCR, which demonstrated
that the expression of genes associated with oxidative phosphorylation
and intercellular junction assembly was increased (Figure [209]8H,J),
while the expression of genes linked to the NF‐𝜅B signaling pathway was
decreased following treatment with PGA‐Cu‐S@G (Figure [210]8I). These
findings elucidate the underlying mechanisms of this smart nanoparticle
system to alleviate discogenic pain, demonstrating its potential as an
innovative strategy for clinical translation.
3. Conclusion
In conclusion, this study reveals the use of nanoparticles to
coordinate the synergistic regulation of the immune microenvironment
and mitochondrial transfer, thereby providing an effective approach to
the treatment of discogenic pain. The findings demonstrated that
PGA‐Cu‐S@G nanoparticles with effective mitochondrial targeting
significantly augmented mitochondrial mass and transfer efficiency in
macrophages, thereby ensuring mitochondria's effective and sustained
delivery to damaged cells. This offers a solution to the efficiency and
quality challenges of mitochondrial delivery. Furthermore, the
synergistic application of immunotherapy and mitochondrial transfer
therapy establishes a foundation for the treatment of IVDD, thus
advancing the practical clinical application of this therapeutic
approach.
4. Experimental Section
Ethics Approval
Ethics involved in experimental arrangement and implementation in this
study, including patient medical data consultation, NP tissue sample
collection and application, animal obtainment, animal surgical
operation, and animal sample collection, were approved by Medical
Ethics Committee of the Second Affiliated Hospital of Wenzhou Medical
University (No. LCKY2020‐157) and Laboratory Animal Ethics Committee of
Wenzhou Institute, University of Chinese Academy of Sciences (No.
WIUCAS23120101, No. WIUCAS24040303).
Single‐Cell Transcriptome Sequencing Data Acquisition and Analysis
Single‐cell RNA‐seq data from discs with different degrees of
degeneration were self‐loaded from the Gene Expression Omnibus data
repository ([211]GSE165722). The Seurat package was used for cell
normalization and regression based on the expression table according to
the percentage of mitochondria rate to obtain the scaled data. To
correct the batch effect, the RunHarmony function from the Harmony
package was applied. PCA was constructed based on the scaled data with
the top 2000 highly variable genes, and the top 40 principals were used
for tSNE construction and Uniform Manifold Approximation and Projection
construction. The marker genes of each cluster were determined as
described above. Cell communication analysis was performed using the
CellChat package.
Human NP Tissue Sample Collection
NP tissue samples from patients with lumbar intervertebral disc
herniation who underwent discectomy were evaluated using the Pfirrmann
grading system, which classified the degenerated IVD tissue as Grade II
and Grade V. Additionally, these patients, who did not have any
cardio‐cerebral‐vascular disease, cancers, infection, immune and
endocrine diseases, or organ dysfunction, approved the use of NP tissue
samples in scientific studies. All volunteers provided informed consent
for using their NP tissues in medical research. The study population
consisted of patients aged 35–65 (n = 5).
Animal Model
The Sprague‐Dawley rat was chosen as the experimental animal for this
study. All rats were sourced from the Zhejiang Provincial Laboratory
Animal Center (Hangzhou, China). To establish the discogenic pain
model, the sample size was n = 5 for each of the three groups: 1) no
surgical incision (control), 2) a skin incision without injury to the
disc tissues (sham), and 3) Co5‐Co6 NP injury (puncture). To validate
the therapeutic effects of the nanoparticles in vivo, sample sizes were
n = 5 for each of the four groups (Control, IVDD, PGA‐Cu@S, and
PGA‐Cu‐S@G). In particular, following the administration of 2% (w/v)
pentobarbital (40 mg kg^−1), rats weighing between 250 and 300 grams
were selected for needle puncture surgery. A 22G needle was utilized.
The punctures were performed at the fifth/sixth caudal vertebrae
(Co5‐Co6) with the assistance of radiographic guidance. A needle was
inserted into the intervertebral disc space at a fixed depth of 3 mm.
Following insertion, the needle was rotated 360° and left in place for
30 s. Thereafter, 10 µL of the material (PBS; PGA‐Cu@S (100 µg mL^−1);
PGA‐Cu‐S@G (100 µg mL^−1)) was injected into the IVD with a micro
syringe. The injections were administered weekly. Mechanical and
thermal pain thresholds were assessed at various time points. Tissue
degeneration levels were evaluated using the DHI method and graded
according to the Pfirrmann grading system with the assistance of
radiography and MRI. The grading was reviewed by three experts in the
field of orthopedics.
Bama pigs (n = 3, aged 12 months, skeletally mature, weight 20–30 kg)
underwent partial nucleotomy under general anesthesia with isoflurane.
The five lumbar vertebrae (L1‐L5) were exposed, and partial nucleotomy
was performed on L2‐L3, L3‐L4, and L4‐L5 with the 16G needle.
Subsequently, 0.5 mL of the phosphate buffered saline (PBS), PGA‐Cu@S
(100 µg mL^−1), and PGA‐Cu‐S@G (100 µg mL^−1) were administered into
the L2‐L3, L3‐L4, and L4‐L5 discs, respectively. The L1‐L2 disc, which
did not undergo nucleotomy, was used as a control. The incision was
then closed using 0‐0 nylon sutures.
Macrophage Isolation and In Vitro Differentiation
To obtain bone marrow‐derived monocytes, the tibia and femur bones were
flushed with PBS. To generate monocyte‐derived macrophages, 10⁶ cells
were seeded in a 75 cm^2 non‐treated tissue culture flask (NEST,
Jiangsu, China) for 7 days in high‐glucose Dulbecco's Modified Eagle
medium (DMEM; Gibco 11 965 092) and DMEM/F12 medium (Gibco 12 634 028)
(1:1). The medium was supplemented with 30% L929 cell‐conditioned
medium, 10% fetal bovine serum (FBS; Gibco A5670701), 1%
penicillin‐streptomycin (Gibco 15 140 122), and 1% L‐glutamine (200 mM,
Solarbio G0200). To polarize the macrophages toward M2 macrophages, the
cells were stimulated with 20 ng mL^−1 of IL‐4 (Solarbio [212]P00196)
for 24 h.
Cell Culture
NP cells — NP tissues were isolated from Sprague‐Dawley rats and
incubated in 2 mg mL^−1 collagenase II at 37 °C for 2 h. After
centrifugation to remove the collagenase, NP cells were cultured in
DMEM/F12 medium (Gibco 12 634 028) supplemented with 10% FBS (Gibco
A5670701) and 1% penicillin‐streptomycin (Gibco 15 140 122).
PC‐12 cells — PC‐12 cells (ATCC, CRL‐1721) were seeded at a density of
1 × 10^4 to 5 × 10^4 cells mL^−1 in RPMI‐1640 medium (Gibco 11 875 093)
containing 10% FBS (Gibco A5670701) and 1% penicillin‐streptomycin
(Gibco 15 140 122).
Raw 264.7 cells — The Raw 264.7 cells (ATCC, TIB‐71) pellet was
resuspended in fresh pre‐warmed complete high‐glucose DMEM medium
(Gibco 11 965 092) with 10% FBS (Gibco A5670701) and 1%
penicillin‐streptomycin (Gibco 15 140 122). Cells were then seeded into
a culture flask at a density of ≈1 × 10^5 to 2 × 10^5 cells mL^−1.
All cells were maintained in a 37 °C incubator with 5% CO[2], and the
culture medium was refreshed every 2 days.
Isolation of Cells from the Coculture System
NP cells and PC‐12 cells were isolated using a fluorescence‐activated
sorting cytometer (Beckman, Brea, USA). In brief, the cocultured cells
were harvested through trypsinization and suspended in neutral PBS.
Initially, the mixed cell population was gated, followed by the
selection and collection of F4/80‐negative cells. A schematic
representation of the fluorescence‐activated sorting cytometer process
is shown in Figure [213]2B.
Cell Viability Assessment
Cell viability was evaluated using a CCK‐8 assay (Beyotime, Shanghai,
China). Cells were plated at a density of 5 × 10^3 cells per well in
96‐well plates (NEST, Jiangsu, China). A working solution was then
prepared by mixing the CCK‐8 solution with a serum‐free medium in a 1:9
ratio, following the manufacturer's instructions. This solution was
incubated with the cells in the dark at 37 °C for 2 h, after which
absorbance was measured at 450 nm. On days 1, 3, and 5, the cells were
treated with 250 µL of Calcein‐AM/propidium iodide detection solution
(Beyotime, Shanghai, China) for 30 min and then examined under a
fluorescence microscope. Live cells exhibited green fluorescence, while
those with damaged membranes appeared red.
Measurement of Intracellular ATP Levels
Intracellular ATP levels were determined using an ATP assay kit
(Beyotime, Shanghai, China) following the manufacturer's guidelines.
Adherent cells or tissue samples were lysed with lysis buffer and
centrifuged at 12 000 × g for 5 min at 4 °C. The resulting supernatants
were collected for a photochemical reaction with ATP working solutions.
A luminometer was used to measure the light units from both the samples
and ATP standard solutions. ATP levels were then calculated based on a
calibration curve derived from the ATP standards.
Mitochondrial Analysis
Cells were collected using trypsin and fixed with an electron
microscope fixative (Servicebio G1102). They were then treated with
0.1 M phosphate buffer containing 1% osmium tetroxide (Ted Pella Inc.
18 456) for fixation. Following this, the cells were dehydrated through
a graded ethanol series and embedded. Ultrathin sections, ≈60−80 nm
thick, were produced using an ultrathin sectioning machine. The
sections were poststained with dicumyl acetate and lead citrate for
transmission electron microscopy analysis (HITACHI HT7700, Japan).
Cells were seeded in confocal dishes and co‐incubated with Mito Tracker
Red (Thermo Fisher Scientific) for 30 min. After this, the cells were
observed and imaged with a confocal microscope. JC‐1 (Beyotime,
Shanghai, China) was applied for 30 min before imaging to assess the
membrane potential of live cells. Superoxide production was measured
using MitoSox reagent (40778ES50; Yeasen Biotechnology, Shanghai,
China), which was initially dissolved in dimethyl sulfoxide to create a
5 mM stock solution, then diluted to a 5 µM working solution. This
working solution was added to a 12‐well plate with the cells. All cell
nuclei were stained with DAPI (Thermo Fisher Scientific) and visualized
using a laser confocal microscope (Olympus).
Scanning Electron Microscopy
Cells were cultured on 12‐mm‐diameter glass coverslips for the required
duration as specified by the experiment. The samples were fixed using
2.5% glutaraldehyde (Sigma) in 0.1 M sodium cacodylate buffer (Electron
Microscopy Sciences). After fixation, the cells were washed three times
for 15 min each with 0.1 M sodium cacodylate buffer, then post‐fixed in
0.1% OsO[4] (Sigma) in water for 1 h at room temperature, followed by
two washes of 10 min each with water before dehydration. The
dehydration process included 5 min in 35% ethanol, 5 min in 50%
ethanol, 10 min in 70% ethanol, 10 min in 90% ethanol, and two rounds
of 100% ethanol for 10 min each. Once fixed and dehydrated, the
coverslips were dried and placed on SEM stubs for sputter coating using
an EMS 300T D dual‐head sputter coater with gold or platinum/palladium
(5 nm). Imaging was conducted with a Zeiss FESEM Supra 55‐VP
microscope, and the images were processed with ImageJ software.
Synthesis of Nanomaterials
Step 1: dissolve 0.2 g of F127 in a mixed solvent of water (50 mL) and
ethanol (25 mL). Step 2: add 0.1 g of gallic acid to the solution and
stir until fully dissolved. Then, add 0.25 mL of ammonia solution (25%
w/w) and stir mildly at room temperature for 30 min. Step 3: add 50 µL
of formaldehyde solution and continue the reaction for 30 min. Step 4:
add 1 mL of a 10 mg mL^−1 CuCl[2] solution (37% w/w) to the mixture and
stir for 1 h. Step 5: transfer the mixture to a high‐pressure reactor
(100 mL) and perform hydrothermal treatment at 100 °C for 8 h. Next,
the solution was centrifugated at high speed and washed thrice with
distilled water and ethanol. The solution was vacuum‐dried and dried to
obtain PGA‐Cu powder. Subsequently, PGA‐Cu and SS05/GAP134 were
dissolved in distilled water at a mass ratio of 25:1 and stirred at a
constant speed for 6 h. The uncoupled peptides were removed by washing
three times with distilled water and ethanol. Finally, the solution was
vacuum‐dried and dried to obtain PGA‐Cu‐S@G powder. To determine the
final binding rate of SS05 and GAP134 with PGA‐Cu, the characteristic
absorption peak at 275 nm and 223 nm was measured using a UV
spectrophotometer. As previously described, the peptide (n[1]) was
combined with PGA‐Cu of mass N. Subsequently, the mixture underwent
washing and centrifugation using absolute alcohol. The resulting
supernatant was collected, and the unbound SS05 (n[2]) mass was
calculated by measuring the absorbance at 275 nm in the supernatant.
Similarly, the mass of unbound GAP134 was calculated by measuring the
absorbance at 223 nm in the supernatant.
[MATH: Bindingrate%=n1−n2N<
mo linebreak="goodbreak">×100%
:MATH]
(1)
[MATH: SS05Bindingrate%=4000−2526
100000×100%=14.74% :MATH]
(2)
[MATH: GAP134Bindingrate%=4000−2403
100000×100%=15.97% :MATH]
(3)
In Vitro Mitochondrial Targeting Validation
Following the instructions of the mitochondrial isolation kit,
macrophages were collected and mechanically homogenized. The
mitochondria were then isolated through centrifugation. After
isolation, the mitochondria were co‐incubated with FITC‐labeled
materials for 1 h at 37 °C. The mitochondria were subjected to three
washes in PBS to remove any residual unbound materials. Finally, the
fluorescence of the mitochondrial suspension was assessed using a flow
cytometer.
Fluorescence Microscopy
Cells cultured on 12‐mm‐diameter coverslips were fixed with 4%
paraformaldehyde at room temperature for 2 h. After fixing, the cells
were washed thrice with 1 × PBS for 20 min each. They were then
permeabilized by incubating with 0.5% Triton X‐100 at 4 °C for 10 min,
followed by three washes with 1 × PBS. The cells were blocked with a
10% bovine serum albumin solution at room temperature for 1 h, and then
incubated with the primary antibody overnight at 4 °C. This was
followed by staining with a secondary antibody (1:300 dilution) for 1 h
at room temperature. For actin staining, the cells were treated with
50 µg/ml of Alexa Fluor 488 phalloidin (Thermo Fisher Scientific) for
1 h at room temperature. Mitochondria were stained with Mito Tracker
Red (Thermo Fisher Scientific) for 30 min, and the nuclear were stained
with DAPI (Thermo Fisher Scientific) for 15 min. The cells were washed
three additional times with 1 × PBS. Coverslips were then mounted onto
glass slides, and images were captured using a Nikon Eclipse Ti camera
(Nikon Instruments) with NIS‐Elements imaging software. ImageJ software
was used for post‐processing. The antibodies utilized for
immunostaining included CD86 (#91882T; CST, USA), CD206 (#24595T; CST,
USA), COL2A1 (ab307674; Abcam, UK), and CGRP (ab272713; Abcam, UK).
Cell Apoptosis and Necrosis
Cells were collected following the manufacturer's instructions and
gently pipetted into a single‐cell suspension using 100 µL of 1 ×
Binding Buffer. Next, 5 µL of Annexin V‐FITC and 5 µL of PI Staining
Solution (Beyotime, Shanghai, China) were added. The cells were
incubated in the dark at room temperature for 10 min, after which
400 µL of 1 × Binding Buffer was added and gently mixed. Finally, the
apoptosis and necrosis rates of the cells were analyzed using a
CytoFLEX (Beckman Coulter, USA).
Western Blotting
Cellular proteins were extracted by lysing the cells in a RIPA lysis
and extraction buffer containing protease and phosphatase inhibitors
(Thermo Fisher Scientific). Protein concentrations were then normalized
using a BCA protein assay kit (Beyotime, Shanghai, China). Then,
electrophoresis separated proteins on 8–12% sodium dodecyl
sulfate‐polyacrylamide gels and transferred to polyvinylidene fluoride
membranes. The membranes were sealed with a rapid‐closure solution
(Beyotime, Shanghai, China) at room temperature for 20 min before being
incubated overnight at 4 °C with the primary antibody. After that, the
membranes were incubated for 1 h at room temperature with the
appropriate secondary antibody, and detection was performed using an
Enhanced ECL Chemiluminescence Detection Kit (Vazyme, Nanjing, China).
Detailed information about the antibodies used can be found in Table
[214]S4, Supporting Information.
Quantitative Real‐Time PCR
Cells were grown in 6‐well plates, and after different treatments,
total RNA was extracted using a Total RNA Extraction Kit (Beyotime,
Shanghai, China). According to the manufacturer's instructions, cDNA
was synthesized using a Prime Script RT kit (Takara, Kyoto, Japan). An
RT‐PCR system (LightCycler 96; Roche, Basel, Switzerland) was used for
amplification. The RT‐PCR parameters were as follows: 40 cycles of 30 s
at 95 °C, 5s at 95 °C, and 30 s at 60 °C. Primer sequences are
available in the Table [215]S5, Supporting Information.
X‐Ray and MRI Examinations
X‐ray images of rat tails were taken before surgery and at 4 and 8
weeks post‐surgery using an XPERT.8 X‐ray machine (Kubtec, Stratford,
CT, USA). The DHI was calculated from these images following the
methodology detailed in the appendix (Figure [216]S32, Supporting
Information). Additionally, weighted MRI of the rat tail vertebrae and
lumbar vertebrae of Bama pigs was conducted at various time points
post‐surgery using the Achieva 3.0 T MRI system (Philips, Amsterdam,
Netherlands). Image evaluation concentrated on the
high‐signal‐intensity regions corresponding to the NP in T2‐weighted
sagittal images. A panel of more than three orthopedic fellows assessed
the degree of IVDD, measuring and evaluating it according to predefined
criteria.
Histological Analysis
Intervertebral disc tissue was extracted from the Co5‐Co6 region of rat
tails. The collected caudal vertebral tissues were fixed with PFA and
then decalcified using 10% Ethylenediamine tetraacetic acid for 8
weeks. Following decalcification, the tissues were dehydrated, embedded
in paraffin, and sectioned to a thickness of 5 µm. To examine
morphological changes in the medulla, surrounding fibrous ring, and
cartilage endplates, the sections underwent HE, SO, and
immunohistochemical staining procedures.
RNA‐Seq and Data Analysis
Total RNA was extracted from each thymus sample using the RNAmini kit
(Qiagen, Germany). The quality of the RNA was assessed through gel
electrophoresis and Qubit analysis (Thermo, Waltham, MA, USA). Only
samples with high quality (OD260/280 = 1.8 – 2.2, OD260/230 ≥ 2.0, RIN
≥ 6.5, 28S:18S ≥ 1.0, and > 2 µg) were selected for sequencing library
construction. Strk‐specific libraries were created using the TruSeq RNA
Sample Preparation Kit (Illumina, San Diego, CA, USA) and sequenced on
an Illumina Novaseq 6000 sequencer. The quality of the raw data was
checked using FastQC v0.11.2 software.
Statistics and Reproducibility
All experiments and results were independently performed on more than
three occasions. Continuous data were presented as mean ± standard
deviation and analyzed using GraphPad Prism 9.0 software (GraphPad
Prism, USA). Two groups of counting data with a normal distribution
were tested using Shapiro‐Wilk's test, and Levene's variance test was
used to assess the homogeneity of variance. A two‐tailed t‐test was
employed for significance analysis. More than two groups of counting
data with a normal distribution were tested by Shapiro‐Wilk's test and
homogeneous variance testing was performed by Levene's variance test.
Two‐way analysis of variance (ANOVA) was used for significance
analysis. All samples were randomly assigned and analyzed together in
each experiment. The investigators were blinded to group allocation
during data collection and analysis to ensure impartiality.
Conflict of Interest
The authors declare no conflict of interest.
Author Contributions
X.W., Z.G., L.C., and J.S. contributed equally to this work. A.M.W.,
Y.L.Z., Q.P.Q. provided the essential ideas and designed the
experiments. X.Z.W., Z.Y.G., L.J.C., J.S., and Y.Y.H. performed most of
the experiments and analyzed the data. X.Z.W., Z.Y.G., L.J.C., and J.S.
collected the clinical samples. X.Z.W., K.Y.H.K., M.J., Y.M.L., and
P.M. drafted the manuscript. A.M.W., Z.Y.L., Q.P.Q., X.Y.W., and X.Q.W.
revised the manuscript.
Supporting information
Supporting Information
[217]ADVS-12-2500128-s001.docx^ (7.5MB, docx)
Acknowledgements