PD-L1 Blockade with Biodegradable Envafolimab-Loaded Microspheres Synergizes with Transarterial Chemoembolization to Overcome Myeloid-Derived Suppressor Cell-Driven Immune Escape in Hepatocellular Carcinoma.

1
Introduction
Hepatocellular carcinoma
(HCC) is one of the leading cancers worldwide.
Classically, HCC develops in genetically susceptible individuals who
are exposed to risk factors, especially in the presence of liver cirrhosis.
Significant temporal and geographic variations exist for HCC and its
etiologies. Geographically, the hepatitis viruses predominate as the
causes of HCC in Asia and Africa. Despite
the application of surgical resection or liver transplant for HCC
treatment, both surgery and transplant present low-ratio eligibility
and high recurrence rate because of the late-stage diagnosis of HCC. According to the Barcelona Clinic Liver Cancer
(BCLC) staging system, which has been commonly used in clinical practice
and endorsed by international guidelines, transarterial chemoembolization
(TACE) is the treatment of choice for intermediate-stage HCC, including
unresectable multinodular HCC without extrahepatic spread. The BCLC
system additionally recommends that TACE should be used when other
recommended treatments are not feasible or unsuccessful in the early
stages of HCC.

TACE exerts dual
antitumor effects through localized chemotherapy
and ischemic necrosis, yet paradoxically modifies the tumor microenvironment
(TME) in ways that may foster therapeutic resistance. Present research
mainly focuses on T lymphocytes and neutrophils. Results from Pinato
et al. showed that TACE significantly
reduces peripheral CD8+ T cells involved in the inhibitory
immune pathway, effectively improving immunotherapy resistance. Concurrently,
TACE also modulates immunosuppressive regulatory T cells (Tregs) in
peripheral blood. Ren et al. studied 33
HCC patients who underwent gelatin sponge microparticles TACE treatment
and observed a decrease in the proportion of Tregs in peripheral blood
from 11.74% ± 1.67% before surgery to 7.59% ± 1.27% after
3–5 weeks after TACE. The inflammatory environment after TACE
is another significant factor that negatively influences the TME. The increased levels of tumor-associated neutrophils
(TANs) and tumor-associated macrophages
(TAMs) may suppress normal immune cells,
fostering tumor growth and metastasis. Tan et al. reported that a reduced number of CD8+ T cells
and an increased number of TAMs were observed in the post-TACE microenvironment.
TREM2 was found to be highly expressed in TAMs following TACE, which
was associated with poor prognosis. Nevertheless, the multifaceted
physiological impact of TACE on the TME is undeniable. This encourages
the exploration of potential synergies with appropriate medications
to expand its applications and improve patient outcomes.
In
the present study, based on the single-cell RNA sequencing (scRNA-seq)
data, we found that TACE therapy results in an increased number of
PD-L1+ myeloid-derived suppressor cells (MDSCs) in HCC.
MDSCs are a heterogeneous population of myeloid cells that develop
in association with chronic inflammation, which is a hallmark of cancer.
Indeed, MDSCs accumulate in the tumor microenvironment, where they
strongly inhibit anticancer functions of T cells and natural killer
cells and exert a variety of other tumor-promoting effects. To counterbalance these paradoxical effects
driven by TACE, we engineered a biodegradable hyaluronic acid-gelatin
microsphere (MS) designed for sustained intra-arterial delivery of
the anti-PD-L1 antibody Envafolimab (KN035). Unlike conventional systemic
immunotherapy, KN035-MS achieves a localized PD-L1 blockade while
mitigating rapid drug clearance, thereby addressing two critical limitations
of post-TACE adjuvant therapy: inefficient drug penetration and MDSC-mediated
immunosuppression. Our mechanistic studies demonstrate that transarterial
KN035-MS administration significantly attenuates TACE-induced PD-L1+ MDSC infiltration and reprograms the TME toward an immunostimulatory
phenotype, marked by CD8+ T-cell activation and M1 macrophage
polarization. This dual-action strategy not only enhances tumor-specific
immunity but also disrupts the feedforward loop of postembolization
immune escape, offering a promising solution to HCC recurrence.

2
Results
2.1
TACE Therapy Results in
an Increased Number
of PD-L1+ MDSCs in HCC
To delineate TACE-induced
remodeling of the tumor immune microenvironment, we performed single-cell
RNA sequencing (scRNA-seq) on CD45+ leukocytes isolated
from paired specimens of five treatment-naive HCC patients and five
post-TACE patients (PRJNA793914). Uniform Manifold Approximation and
Projection (UMAP) analysis resolved eight conserved immune cell populations,
systematically annotated through marker gene expression (Figures
A and S1A). For example, the B cell cluster mainly
expressed CD79A, MS4A1, and the MDSC cell cluster mainly expressed
LYZ and C1QB (Figure S1B). As displayed
in Figure
B, histograms
indicate the specific expression of each cell cluster in different
samples. For example, the expression of T and NK cells in the TACE
group was lower than that in the control group. The expression of
myeloid cells was higher than that in the control group. Most importantly,
we found that MDSCs increased significantly after TACE therapy (Figure
C,D). Pathway enrichment
analysis suggests genes of MDSCs were mainly enriched in positive
regulation of cytokine production and immune response-regulating signal
pathways (Figures
E and S1C). These findings establish TACE-induced
PD-L1+ MDSC expansion as a previously unrecognized mechanism
of immune evasion, positioning localized MDSC targeting as a rational
strategy to prevent HCC recurrence in postembolization patients.

2.2
Clinical Data Reveals That TACE Combined with
the PD-L1 Antibody Effectively Inhibits HCC
To explore the
clinical effect of combined PD-L1 antibody after TACE treatment, we
screened 15 HCC patients treated with TACE, of which 4 were treated
in conjunction with PD-L1 blockade, after excluding noncompliant cases
in accordance with the exclusion criteria from a retrospective analysis
of data from 187 patients who received TACE as their primary treatment
at our center. The specific screening process is shown in Figure S1D. To determine the efficacy of TACE
treatment, enhanced computed tomography (CT) was performed. The rate
of tumor necrosis is proportional to the density and homogeneity of
the iodine oil deposition in the lesion after TACE; regions devoid
of iodine oil deposition are regarded as tumor survivors. Patients
in both the TACE subgroup and the TACE-conjugated anti-PD-L1 subgroup
exhibited substantial iodine-oil deposits in the dynamic-enhanced
CT scan after TACE, as illustrated in Figure
F. While certain iodine-oil deposits were
detected in the tumor foci located in the right lobe of the liver,
the portion of the dynamic-enhanced CT scan devoid of iodine-oil deposits
continued to exhibit enhancement (Figure
F). Then, we analyzed the 15 patients’
detailed data: 2 of the 15 patients attained PR, and the remaining
13 attained Stable Disease (SD) as per the Recist 1.1 criteria; 8
of the 15 patients attained SD as per the mRecist criteria, while
7 attained Partial Response (PR) (Figure
G). Meanwhile, we analyzed alpha-fetoprotein
(AFP) levels, a dependable indicator for assessing the efficacy of
HCC TACE treatment, in 15 patients with HCC at the study baseline,
before and after TACE treatment. Nine patients (60%, 9/15) had AFP
levels <400 μg/L, and 6 patients (40%, 6/15) had AFP levels
≥400 μg/L. Of the six patients with abnormal AFP levels
at the baseline, exactly three were treated with combined anti-PD-L1
therapy, and three patients did not. AFP decreased by 50–100%
in the three patients treated with MS-IDA (Idarubicin); AFP decreased
by 90–100% in patients receiving combination therapy, except
for one patient who experienced a transient resurgence in AFP during
the examination period (Figure
H). These clinical data show that TACE combined with the PD-L1
antibody can effectively inhibit HCC.

2.3
Engineered
Biodegradable Embolized Microspheres
for Sustained PD-L1 Blockade with KN035 via Hepatic Artery Administration
The above scRNA-seq data and clinical data all indicate that the
combination of PD-L1 inhibitors after TACE treatment has a good anticancer
effect, but the drug delivery efficiency and safety of the PD-L1 antibody
need to be further improved. The PD-L1 antibody has shown great potential
as an immunotherapy drug in the treatment of cancer, but its intravenous
injection has some limitations, especially the problem of low drug
utilization. Intravenous injection causes
the drug to quickly enter the bloodstream and be widely distributed
throughout the body, and not all drugs can accurately reach the tumor
site. This not only reduces the concentration of the drug locally
to the tumor but may also increase the potential side effects on healthy
tissue such as immune-related adverse reactions.
To improve
the drug availability and anticancer efficacy and safety of the PD-L1
antibody, we synthesized a novel sustained-release biodegradable embolized
microsphere loading Envafolimab. The drug loading mechanism of KN035
in hyaluronic acid-gelatin microspheres is primarily through physical
adsorption, where electrostatic interactions between the carboxyl
groups of hyaluronic acid and the amino groups of KN035 enable drug
loading. The final form of our biodegradable embolized microspheres
is a dry spherical particle powder, which is convenient for long-term
transportation and storage after sterilization. An optical microscope
was used to observe the morphology and measure the average diameter
of microspheres before and after fully swelling in the drug solution
at 37 °C. The results were as shown in Figure
A–D; after swelling, the diameter
of the microsphere increased, showing good swelling property. Two
hundred microspheres are selected in the field of vision to calculate
the particle size distribution. The microspheres exhibited a relatively
uniform size distribution. The diameter of the dry microspheres ranged
from 100 to 180 μm, with a mean value of 139.6 ± 13.1 μm.
After water absorption and swelling, the particle size distribution
shifted to 220–300 μm, with an average diameter of 254.4
± 23.8 μm. The microspheres were swollen with pure water,
and the particle size was recorded at intervals. The results showed
that the maximum swelling of the microspheres was 271 ± 6%, the
rapid swelling was 215 ± 5% within 1 h, and then the swelling
to 267 ± 9% in 5 h reached a plateau, after which the swelling
rate did not change much (Figure
E). We also tested the mechanical properties of the
microspheres and found that when the microspheres were compressed
by 30% and 50%, they were intact, stress relaxation occurred during
the maintenance process, and the microspheres could be restored to
their original state. The microspheres broke during the 95% compression
process (Figure
F). Figure
G presents the representative
images of the recovery of microspheres after compression and the rupture
of microspheres after compression. The above results showed that our
newly synthesized hyaluronic acid–gelatin microspheres had
good micromorphology, particle size distribution, and good resistance
to compression.
We took 100–300
μm particle size microspheres and
used an emission scanning electron microscope to photograph the structure
before and after KN035 loading. It was found that the surface of the
freeze-dried microspheres showed a highly porous network and absorbed
fine crystalline particles after drug loading compared with that before
KN035 loading (Figure
H,I). The standard curve of the concentration of KN035 in water and
the OD at 278 nm is shown in Figure
J. Figure
K shows the drug release of the microsphere loaded with KN035
in normal saline. The bursting release of KN035 was observed at 1
h, with a release rate of 25.2 ± 5.5%. The rate reached 69.3
± 4.7% after 12 h, and it was nearly completely released at 60
h. Considering the physiologically water-limited environment of embolism,
the drug release rate may be slower than the simulated drug release
test. The above results showed that our newly synthesized hyaluronic
acid–gelatin microspheres had excellent behavior of KN035 loading
and releasing.
To assess whether the microspheres are suitable
for TACE, we also
examined several physical properties. The microsphere catheter had
good passability. During the injection process with the syringe, the
reaction proceeded smoothly without obvious blockage. The microspheres
were not broken, and there were no microsphere residues left in the
catheter or the syringe (Figure
L). The microspheres have the best suspensibility when
the ratio of normal saline/iodixanol is 10:13. Under this condition,
the microspheres can remain suspended for 5 min without sedimentation,
which is suitable for TACE operations. It can effectively deliver
the microspheres to the distal end and enable angiography with contrast
agents at the same time (Figure
M). The microspheres swell after absorbing water, and
their particle size increases rapidly. At 60 s, the swelling rate
of the microspheres reaches the maximum, and the microspheres are
fully swollen. The swelling rate is 178.3 ± 5.4%, and the time
for the microspheres to be fully swollen is 60 s, which has no impact
on the clinical preparation time (Figure S2A,B).

2.4
KN035-Loaded Microspheres Demonstrate Superior
Tumor Growth Inhibition and a Favorable Safety Profile in Humanized
BALB/c-hPD-L1 Mouse HCC Models
To verify the efficacy of
KN035-MS in vivo against HCC, we injected H22 cells with KN035 and
KN035-MS groups into BALB/c-hPD-L1 mice by intratumoral injection
for assessing the antitumor capability (Figure
A). The extracellular domain of BALB/c mice
PD-L1 was replaced with the corresponding human fragment by genome
editing technology, and the intracellular domain of PD-L1 mice was
completely preserved. The humanized BALB/c-hPD-L1 mouse model was
independently developed (Figure
B). We used direct drug injections into the tumor.
According to the results, compared with the KN035 group, the volume
and weight of the tumor decreased significantly after KN035-MS injection
(Figure
C–E).
The indicator Ki67, which measures the degree of cell proliferation
activity, also suggested that the expression in the KN035-MS group
was lower than that in the control group (Figure
F,G). These results indicate that KN035-MS
can effectively inhibit tumor growth compared to KN035 in BALB/c-hPD-L1
mouse HCC models. We also established a rabbit model for TACE. As
shown in Figure S2C, digital subtraction
angiography (DSA) images obtained after transhepatic arterial embolization
with KN035-MS revealed that most terminal branches of the hepatic
artery were rapidly occluded shortly after injection, compared with
the preinjection images. To further validate the efficacy of KN035-MS,
we developed a rat HCC model. DSA performed after transarterial administration
of KN035-MS similarly demonstrated rapid occlusion of most terminal
hepatic artery branches postinjection relative to the baseline (Figure
H). Furthermore,
we used PET-CT to evaluate changes in liver tumors (Figure
I). Representative PET-CT images
in axial, coronal, and midsagittal planes confirmed that KN035-MS
significantly suppressed the growth of hepatic tumors in rats.
To further explore the biosafety of KN035-MS use
in vivo, HE staining
was performed on the heart, spleen, lungs, and kidney in the subcutaneous
tumor BALB/c-hPD-L1 mice model, and it was found that there was no
significant difference between the KN035-MS group and the control
group, and KN035-MS had no significant damage to the heart, spleen,
lungs, and kidneys (Figure
J). The heart, spleen, lungs, and kidneys in the mouse PBMC
model also had no significant changes. These results suggest that
KN035-MS via hepatic artery administration is safe in vivo.

2.5
Mass Cytometry Reveals KN035-Loaded Microspheres
Achieve Superior Suppression of PD-L1+ MDSCs in HCC
To further evaluate the effect of local injection of microspheres
loaded with KN035 on the TME, mass cytometry served to measure the
respective immune cell clusters’ expressions. A total of 35
cell clusters were identified, which were defined considering respective
cell type’s specific markers (Figures
A,B and S3). As
shown in Figure
C,D,
the distribution and proportion of each cell population in different
samples were presented. We calculated the percentage of each cell
population relative to the sample. Results showed that relative to
the KN035 group, KN035-MS injection led to increased CD4+ T, gdT, ILC cells, and M1 macrophages but decreased M2 macrophages,
MDSCs (Figure
E).
In addition, we also analyzed the changes in the proportion of different
cell populations compared to the total cell population, and the results
showed that relative to the KN035 group, KN035-MS injection led to
increased dendritic cells (DC), gdT, DNT, innate lymphocytes (ILC),
CD4+ T cells, and M1 macrophages but decreased M2 macrophages
and MDSCs (Figure
F). In general, compared with the nonencapsulated microspheres, the
injection of KN035-MS activated the TME and reduced the generation
of immunosuppressive cells, which was conducive to the better play
of the PD-L1 antibody in HCC. We synchronously analyzed the detailed
expression of some signature proteins in different groups. Results
revealed that KN035-MS injection led to increased TCR and CD4 expression
but decreased Ly6G and CD206 expression, which was assessed from the
expression of the population sample (Figure
G,H). These results revealed that KN035-MS
inhibited PD-L1+MDSCs cells and reshaped the TME better
than the KN035 group.
To further verify the
effect of KN035-MS treatment on the TME,
we further removed the subcutaneous tumor tissue from BALB/c-hPD-L1
mice and verified it by immunohistochemistry and fluorescence. Although
mass cytometry analyses showed no significant difference in CD8+ T cells, we further validated their expression using immunohistochemistry
and were pleasantly surprised to find that the expression of CD4 and
CD8 in the KN035-MS group was significantly higher than that in the
control group (Figure
A–D). We detected PD1+ CD8+ T cells
by immunofluorescence and found that their expression in the KN035-MS
group was lower than that in the control group, but CD8+ T cells were up-regulated after KN035-MS injection (Figure
E,F). Immunofluorescence also
revealed the higher CD86 expression (M1 macrophage marker) in KN035-MS
relative to the KN035 group, whereas the expression of CD163 (M2 macrophage
marker) was lower in the KN035-MS group (Figure
G,H), which was consistent with mass cytometry.
MDSCs, whose markers are CD11b and Ly6G, also suggested that the expression
in the KN035-MS group was lower than that in the control group (Figure
I,J). All the above
results suggest that KN035-MS can better induce an immunostimulant
tumor microenvironment in HCC than KN035.

2.6
Synergistic Efficacy of TACE with KN035-Loaded
Microspheres in Preventing HCC Recurrence
The above experiments
were carried out in the intratumoral injection model. To further compare
the effect of subcutaneous injection and local injection of the PD-L1
antibody, the tumor-bearing mice were divided into three groups, namely,
the IDA intratumoral injection group, the IDA intratumoral injection
+ KN035 subcutaneous injection group, and the IDA intratumoral injection
+ KN035-MS intratumoral injection group. The results indicated that
the volume and tumor of the IDA intratumoral injection + KN035 subcutaneous
injection group were significantly smaller than those of the IDA intratumoral
injection group, and those of the IDA intratumoral injection + KN035-MS
intratumoral injection group were significantly lower than those of
the other two groups (Figure
B–D). The indicator Ki67 also suggested that the expression
in the IDA intratumoral injection + KN035-MS group was lower than
that in the other two groups (Figure
E). The expression of CD4 and CD8 in the IDA intratumoral
injection + KN035-MS intratumoral injection group was significantly
higher than that in the other two groups (Figure
F,G). We detected PD1+ CD8+ T cells by immunofluorescence and found that their expression
in the IDA intratumoral injection + KN035-MS intratumoral injection
group was lower than that in the other two groups, but CD8+ T cells were up-regulated after KN035-MS injection (Figure
H–J). MDSCs in the IDA
intratumoral injection + KN035-MS intratumoral injection group were
lower than those in the other two groups (Figure
K,L). These results indicate that IDA intratumoral
injection combined with KN035-MS intratumoral injection is beneficial
for further anti-HCC and solves the problem of tumor recurrence after
traditional TACE therapy.

3
Discussion
While TACE remains a cornerstone
in the management of intermediate-stage
HCC, its profound impact on the tumor immune microenvironment necessitates
strategic therapeutic combinations. Our findings demonstrated that
TACE enhances immunosuppressive cell populations, particularly PD-L1+ MDSCs. This biological paradox underscores the critical need
for adjuvant immunotherapies that can counteract TACE-mediated immunosuppression
while capitalizing on locoregional tumor control. The emerging paradigm
of combining immune checkpoint inhibitors (ICIs) with TACE has shown
promise. The Phase II trial by Liang et al. reveals compelling evidence for multimodal therapy superiority.
The combination of Envafolimab (a novel subcutaneous PD-L1 inhibitor)
with targeted agents and TACE achieved unprecedented response rates
(50% by RECIST 1.1 and 83.3% by mRECIST), significantly outperforming
historical controls across BCLC stages B and C populations. Notably,
this regimen demonstrated an improvement in BCLC-C response rates
compared to the IMbrave150 trial’s atezolizumab–bevacizumab
combination, potentially attributable
to the unique pharmacokinetic profile of subcutaneous administration
and locoregional therapy synergy. The pharmacological innovation of
KN035 (a single-domain PD-L1 antibody-Fc fusion protein) presents
distinct advantages in tumor penetration and sustained release kinetics.
With a molecular weight approximately 50% lower than that of conventional
monoclonal antibodies, while retaining full antigen-binding capacity,
KN035’s enhanced tissue penetration addresses a critical limitation
of current ICIs.

Although these
systemic immunotherapies demonstrate improved therapeutic
outcomes, their inherent limitations as whole-body interventions warrant
a critical evaluation. The inherent limitations of traditional diagnosis
and treatment have driven the development and application of emerging
nanotechnologies, and there are already many immunotherapy materials
developed based on the principles of nanotechnology.
,
Current subcutaneous or intravenous administration of PD-L1 inhibitors,
including KN035, inevitably leads to systemic drug distribution, which
may paradoxically amplify off-target immune activation while achieving
suboptimal intratumoral drug concentrations. This systemic exposure
not only increases the risk of immune-related adverse events (irAEs)
but also fails to sufficiently counteract the concentrated immunosuppressive
niche created by TACE. Therefore, we engineered a biodegradable hyaluronic
acid-gelatin MS designed for sustained intra-arterial delivery of
the anti-PD-L1 antibody KN035. Unlike conventional systemic immunotherapy,
KN035-MS achieves a localized PD-L1 blockade while mitigating rapid
drug clearance. While nanomaterial-based drug delivery systems have
dominated recent advances in encapsulated immunotherapy research,
their clinical translation faces inherent limitations.
,
The groundbreaking work by Wang et al. exemplifies both the promise and constraints of nanotechnology:
using a nanoscale microneedle-array delivery platform for PD1 monoclonal
antibodies in subcutaneous melanoma models, they achieved enhanced
tumor-specific lymphocyte infiltration and metastasis suppression
compared with systemic administration. Our microsphere platform represents
a paradigm shift in localized immunotherapy delivery. This therapeutic
triad, which includes sustained release, vascular-targeted localization,
and microenvironment-responsive degradation, overcomes the “passive
targeting dilemma” inherent to nanoscale systems. While nanomaterials
rely on enhanced permeability and retention effects (EPR) for tumor
accumulation, our platform actively exploits tumor vasculature abnormalities
through embolization-guided deposition. Furthermore, the microsphere’s
size-tunable architecture avoids nanoparticle-associated risks of
extrahepatic dissemination. Critically, this delivery paradigm transforms
KN035 from a systemic immune modulator into a tumor-centric PD-L1
neutralizer while limiting systemic exposure to conventional doses.
Based on the scRNA-seq data, we found that TACE therapy results
in an increased number of PD-L1+ MDSCs in HCC. MDSCs are
a heterogeneous group of cells that play a key role in tumor-related
immune suppression. Through their immunosuppressive function, MDSCs
enable tumors to evade immune surveillance. The infiltration of MDSCs
in tumor tissues is closely related to poor prognosis in patients
and resistance to treatment. In tumor
tissues, elevated HIF-1α induced by hypoxia promotes the differentiation
of MDSCs into TAMs and promotes the inhibitory activity of MDSCs through
the upregulation of inducible nitric oxide synthase (iNOS) and arginase.
HIF-1α also binds to the promoter of PD-L1, which MDSCs express
to inhibit antitumor T cell responses. Wang et al. reported that ionizing radiation induces MDSC expansion
and YTHDF2 expression in both murine models and humans. Pharmacological
inhibition of YTHDF2 overcomes MDSC-induced immunosuppression and
improves combined ionizing radiation or anti-PD-L1 treatment. Hou
et al. reported that radiotherapy enhances
metastasis through immune suppression by inducing PD-L1 and MDSC in
distal sites. Blockade of the PD-L1/CXCL10 axis or MDSC infiltration
during irradiation can enhance abscopal tumor control and reduce metastasis.
These studies indicate that local treatment can not only combat cancer
but also reshape the tumor microenvironment, especially leading to
an increase in MDSC. Our findings establish KN035-MS as a multimodal
modulator of the immunosuppressive TME. Localized KN035-MS administration
outperformed systemic KN035 therapy by achieving three synergistic
effects: (1) attenuating TACE-driven PD-L1+ MDSC infiltration;
(2) reprogramming immunosuppressive myeloid populations via M1 macrophage
polarization and ILC recruitment; (3) enhancing CD8+/CD4+ T-cell effector function through prolonged intratumoral drug
retention. These results position KN035-MS not merely as an improved
delivery method but as an immunotherapeutic that redefines the rules
of cancer-immune interactions in HCC management.
While our findings
demonstrate the therapeutic potential of KN035-MS
in reshaping the post-TACE immunosuppressive niche, several limitations
warrant consideration. First, the preclinical models employed, though
comprehensive, cannot fully recapitulate the heterogeneity of human
HCC, particularly in terms of tumor stroma complexity and variations
in hepatic arterial anatomy that may influence microsphere distribution.
Second, while sustained drug release over 20 days was achieved in
murine models, the translational relevance of these kinetics requires
validation in larger animals with human-scale liver volumes and blood
flow dynamics. Third, the safety profile of repeated KN035-MS administrations,
particularly regarding hepatic arterial patency and systemic autoimmunity
risks, needs rigorous evaluation in chronic toxicity studies. Addressing
these limitations in future studies will be essential to advance KN035-MS
toward clinical implementation.

4
Conclusion
Our integrated analysis
of single-cell transcriptomics and preclinical
models uncovers a critical limitation of conventional TACE in HCC:
therapy-induced expansion of PD-L1+ MDSCs, which establishes
an immunosuppressive niche conducive to tumor recurrence. To address
this, we engineered KN035-MS, a biodegradable hyaluronic acid-gelatin
microsphere platform optimized for sustained intra-arterial delivery
of the anti-PD-L1 antibody Envafolimab (KN035). Mechanistically, localized
KN035-MS administration outperformed systemic KN035 therapy by achieving
three synergistic effects: (1) attenuating TACE-driven PD-L1+ MDSC infiltration; (2) reprogramming immunosuppressive myeloid populations
via M1 macrophage polarization and ILC recruitment; (3) enhancing
CD8+/CD4+ T-cell effector function through prolonged
intratumoral drug retention (Figure
). The KN035-MS system overcomes two fundamental barriers
in HCC management: rapid systemic clearance of immunotherapeutics
and post-TACE adaptive immune evasion. Future clinical trials should
validate these preclinical advantages while exploring combinatorial
regimens with other checkpoint inhibitors or angiogenesis-targeting
agents to further amplify the therapeutic efficacy.
BDDE-modified hyaluronic acid and gelatin were
used to synthesize
microspheres. In the rabbit model, KN035-MS was used for hepatic artery
embolization, and KN035-MS has a significant occlusive effect on blood
vessels. Intratumoral injection of KN035-MS in mouse models activated
antitumor immunity, promoted the elevation of CD8+T cells,
CD4+T cells, ILC cells, and M1 macrophages, and decreased
the expression of CD8+ T cell exhaustion markers. At the
same time, KN035-MS also reduced some immunosuppressive cells, such
as M2 macrophages and MDSCs. IDA combined with KN035-MS is beneficial
for further anti-HCC and solves the problem of tumor recurrence after
traditional TACE therapy.

5
Materials
and Methods
5.1
Data Availability Statement
The data
sets generated in this study are available on the public NCBI network
for PRJNA793914 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA793914).

5.2
Clinical Data Retrospective Study Design
The study received approval from the Ethics Committee of Nanjing
Medical University (2019-SRFA-238). Informed consent was not obtained
due to the retrospective nature of the study. We conducted a small
real-world retrospective study using clinical data from our institution
to validate the therapeutic efficacy of MS-IDA and to understand the
benefits of combining CalliSpheres loading of Idarbicin (MS-IDA) with
anti-PD-L1 treatment for patients. A total of 187 medical records
of HCC patients who underwent DEB-TACE treatment at our institution
between January 1, 2021, and November 30, 2023, were meticulously
reviewed by two independent researchers from the Hepatobiliary Center
of the First Affiliated Hospital of Nanjing Medical University.
The following requirements were necessary for inclusion in the study:
(1) participants had to be between the ages of 18 and 80 years; (2)
they must have received a definitive diagnosis of HCC according to
the European Association for the Study of the Liver; (3) participants
should not have a history of prior HCC treatment; (4) they must have
at least one measurable lesion present; (5) participants had to be
undergoing treatment at our center using a DEB-TACE-based regimen;
(6) participants should have an Eastern Cooperative Oncology Group
Performance Status (ECOG PS) score of 0 to 1. Exclusion criteria for
the study were as follows: (1) participants with extrahepatic metastases
from HCC or concurrent tumors other than HCC were not eligible; (2)
individuals who had used other types of microspheres or chemotherapeutic
agents were excluded; (3) participants with proven presence of mixed
liver cancer rather than pure HCC were not included; (4) incomplete
data from the study and lack of detailed full-length case data were
also exclusion criteria.

5.3
The Patient’s Interventional
Procedure
Using CalliSpheres Loading IDA
Under digital angiography,
two skilled interventional liver surgeons carried out the HCC TACE
therapy technique. The right femoral artery was first percutaneously
punctured by using a modified Seldinger technique, after which the
catheter sheath was positioned and fastened. An angiographic image
acquisition process was subsequently used to acquire images of the
abdominal artery or common hepatic artery, including the arterial,
parenchymal, and venous phases. The microcatheter (2.7 Fr. Outer diameter,
Progreat, Tokyo, Japan) was then advanced coaxially to cannulate the
artery supplying the tumor as superselectively as possible, considering
the preoperative magnetic resonance/computed tomography (MR/CT) angiography
and arteriography findings regarding the tumor lesion and the supplying
artery. 10 mg of IDA was placed into CalliSpheres for TACE embolization
after an additional round of arteriography to verify the proper superselection.
The microspheres were sized between 300 and 500 μm and were
left to stand for 30 min, and a subsequent microsphere push was performed
at 1 mL/min. This prevented the microspheres from refluxing by maintaining
a uniform suspension throughout the injection procedure and preventing
the head end of the microcatheter from being wedged. The goal of embolization
was to almost completely stop blood flow in the artery feeding the
tumor (i.e., to empty the drug-carrying microspheres and contrast
suspension from the artery within two to five cardiac cycles). Iodinated
oil was subsequently injected into the supplying artery for supplementary
embolization until the artery supplying the tumor was completely embolized.
Following embolization, patients whose femoral artery was the point
of puncture had manual compression applied to the puncture site for
5–15 min after the catheter sheath was removed. This was followed
by a compression bandage, which ended when the dorsalis pedis artery
at the puncture site palpably pulsed.

5.4
Preparation
of Biodegradable Microspheres
The microspheres were synthesized
with two biodegradable materials,
hyaluronic acid and gelatin. Hyaluronic acid gave the microspheres
excellent hydrophilicity. Gelatin provided the microspheres with a
porous network structure to absorb drugs in the swelling process.
The compositing of these two materials endowed the microspheres with
enhanced mechanical performance and prolonged biodegradation. Two-step
synthesis of the microsphere: (1) hyaluronic acid was precross-linked
by BDDE to solve the problem of the short half-life of hyaluronic
acid in vivo. (2) Then, the cross-linking-emulsion method was used,
and the BDDE precross-linked hyaluronic acid-gelatin microspheres
with an amide-based bond-based interlocking structure were obtained.
For more experimental details about the cross-linking-emulsion method,
refer to our previous similar research work.

5.5
Mechanical Property Analysis of Biodegradable
Microspheres
The microspheres with a particle size of 500
μm were compressed by 30%, 50%, and 95%, respectively, and then
maintained for 10 s, slowly restoring to the original state. A texture
analyzer (TA.XTC-20, Shanghai Bosin Technology) was used to detect
mechanical changes.

5.6
Swelling Rate of Biodegradable
Microspheres
The microspheres were swollen with pure water,
the particle size
of the microspheres was recorded at intervals using an optical microscope,
and the particle size of 50 microspheres was counted each time; then
the change of the swelling of the microspheres with time was calculated.

5.7
Microscopic Morphology of Biodegradable Microspheres
Microspheres with a 100–300 μm diameter were photographed
before and after PD-L1 drug loading. The microspheres were freeze-dried
and photographed by using a field emission scanning electron microscope
(Ultra Plus).

5.8
KN035 Loading Efficiency
According
to the method introduced in the previous article, we calculated that 1 mL of microspheres with swelling volume
required 450 mg of dried microspheres. Based on the above weight–volume
data, we studied the drug loading capacity of microspheres by the
following process: 100 μL, 200 mg/mL drug KN035 (Simcere Pharmaceutical
Group Limited, China) was added to 45 mg of microspheres; after 30
min of drug loading, the free drug was washed off with pure water
three times and collected. The drug KN035 was quantified by an ultraviolet
spectrophotometer, and the drug KN035 loading capacity of 1 mL of
microspheres was calculated to be 182.2 ± 9.9 mg; then, we tested
the drug KN035 release capacity of microspheres.

5.9
Cell Cultures
Mouse HCC cells (H22),
human liver cells (THLE-2), and human HCC cells (Hep-3B) were supplied
by the Cell Bank of Type Culture Collection. H22 cells were cultured
with RPMI 1640 medium (Gibco, USA), Hep-3B cells were cultured with
DMEM medium (Gibco, USA), and THLE-2 cells were cultured with BEGM
medium (Lonza, USA). All cells were cultured in a medium containing
10% fetal bovine serum (Gibco, USA) and 1% penicillin/streptomycin
(Gibco, USA). All cells were cultured under controlled conditions
of 37 °C and 5% CO2 within an incubator.

5.10
Cell and Blood Compatibility of Microspheres
The cytotoxicity
of the microspheres was studied by a CCK-8 assay.
H22, Hep-3B, and THLE-2 cells were cultured in 96-well plates (3 ×
103 cells, 90 μL per well) for 12 h at 37 °C.
Then, microspheres at different concentrations were added, and the
cells were incubated at 37 °C for 48 h. 10 μL of CCK-8
solution was added to each well and incubated for 2 h. The absorbance
of all of the samples was measured using a microplate reader at 450
nm.
The hemolytic effect of the microspheres was evaluated using
blood collected from healthy mice. The whole blood was separated by
centrifugation (500g, 4 °C, 10 min). After the
upper plasma was removed, the remaining cells were repeatedly washed
with PBS buffer until the supernatant became colorless, thereby obtaining
red blood cells. These red blood cells were then mixed with microspheres
at concentrations of 50 or 100 μg/mL in PBS and incubated at
37 °C for 20 min. Following incubation, the absorbance of the
supernatant was measured at a wavelength of 545 nm by using a UV spectrophotometer.
PBS and 0.1% Triton X-100 were used as the negative and positive controls,
respectively.

5.11
Catheter Passability of
Microspheres
To evaluate the catheter passability of microspheres,
based on the
animal experimental model and clinical requirements, we selected a
4.0 Fr catheter (Terumo, Tokyo, Japan). The experimental procedures
were as follows: first, the swelling Ms were suspended in a mixture
of iodixanol and normal saline at the optimal ratio. Next, the microsphere
suspension was injected into the catheter by using a syringe, and
an optical microscope was employed to observe whether the microspheres
could pass smoothly through the outlet of the catheter and whether
there was any damage. Meanwhile, we also checked whether there were
any microsphere residues inside the microcatheter.

5.12
Suspensibility of Microspheres
The
suspensibility of microspheres is a key parameter during the process
of TACE, which directly affects the embolization effect. For this
reason, we designed an experiment using a 2 mL EP tube as the test
container. The experimental steps were as follows: first, different
proportions of normal saline/iodixanol were prepared. In each proportion,
we observed and recorded the sedimentation of microspheres within
5 min. This helps to evaluate the impact of different suspension ratios
on the sedimentation behavior of microspheres.

5.13
Determination of the Swelling Rate of Microspheres
To evaluate the swelling behavior of microspheres, we adopted a
simple measurement method. With the help of an optical microscope,
we recorded in real time the changes in the particle size of dry-Ms
during the water absorption and swelling processes after adding normal
saline. Moreover, we observed the dynamic curves of the swelling rates
of 20 microspheres within the field of view changing over time and
drew a line graph of their swelling rates varying with time.

5.14
In Vivo Tumor Mouse Model and Treatment
The Animal
Management Committee of Nanjing Medical University approved
the animal experiments (IACUC-2404099), and all experimental techniques
and animal care followed the institutional ethical standards for animal-related
research. The Experimental Animal Centre of Nanjing Medical University
uses a specialized specific pathogen-free (SPF) breeding method to
raise all mice, rats, and New Zealand white rabbits. Cervical dislocation
is used to euthanize mice; carbon dioxide inhalation is used to euthanize
rats, and air injection into the marginal ear vein is used to euthanize
rabbits. All animal experiments were conducted under anesthesia with
isoflurane, which is specifically for animal experiments, for anesthesia
during operations.
BALB/c-hPD-L1 mice (5 weeks old) were purchased
from GemPharmatech Co., Ltd. (China). For the establishment of a subcutaneous
tumor-bearing mouse model, 2 × 106 H22 cells, suspended
in 100 μL of PBS, were subcutaneously inoculated into the right
axillary region of BALB/c-hPD-L1 mice. Two groups were as follows:
KN035 and KN035-MS. When the tumors grew to 100–200 mm3 in volume, KN035 and KN035-MS (100 μg, every 5 days)
were intratumorally injected into the mice in two groups. Tumor dimensions
were assessed at 4 day intervals. The calculation of tumor volume
employed the formula: volume (mm3) = width/2 × length/2.
A subcutaneous tumor model was first established in mice. Two ×
106 H22 cells were suspended in 100 μL PBS and inoculated
subcutaneously in the right axilla of BALB/c-hPD-L1 mice. After 12
days, IDA intratumoral injection (100 μg per mouse) was performed
on mice, and the mice were divided into 3 groups: IDA intratumoral
injection (100 μg, every 5 days), the KN035-MS (100 μg,
every 5 days) intratumorally injection group, and the KN035 (100 μg,
every 5 days) subcutaneous injection group. The mice were euthanized
on the 24th day, and the tumor tissues were removed.

5.15
Rabbit and Rat TACE Treatment Model
In the rabbit TACE
model, a 4-French catheter (VER angiographic catheter,
Cordis, USA) was inserted into the right central artery of the rabbit
for arteriography. To enter the correct hepatic artery, a 2.7-French
microcatheter (Progreat Micro Catheter system, Terumo, Japan) was
passed through the catheter into the descending aorta. After the correct
hepatic artery was displayed, the microcatheter was selectively catheterized
until the tip of the catheter was gently placed near the proper hepatic
artery. Then, under the guidance of digital subtraction angiography
(DSA), a mixture of KN035-MS and a contrast agent was injected. After
the mixture was fully injected, the microcatheter was removed, and
the puncture site was carefully pressed to stop bleeding. The DSA
images and videos of each group of rabbits before and after TACE treatment
were recorded.
To establish a rat liver cancer model, two-week-old
male Sprague–Dawley (SD) rats received a single intraperitoneal
injection of diethylnitrosamine (DEN; 25 mg/kg) to initiate tumorigenesis.
This was followed by intraperitoneal injections of carbon tetrachloride
(CCl4; 5 mg/kg) twice a week for 24 weeks to promote carcinogenic
progression. Liver lesions were dynamically monitored via color Doppler
ultrasound to track tumor size. The rats were anesthetized with isoflurane
and placed in a supine position on the surgical table. The tail was
disinfected with iodophor, and a puncture needle was inserted into
the tail artery at a 30° angle at a site with palpable arterial
pulsation. After blood reflux was observed, a guidewire was introduced.
A catheter sheath was then inserted over the guidewire. The contrast
agent was injected through the catheter, and DSA was used to confirm
the position of the catheter tip within the target vessel before KN035-MS
was administered. After catheter removal, the puncture site was compressed
for 5 to 10 min to ensure hemostasis. Postoperative tumor status was
assessed using positron emission tomography–computed tomography
(PET-CT).

5.16
Biosafety Assessment
The organs
(heart, spleen, lungs, kidneys) of each group of mice were dissected
and hematoxylin–eosin (HE)-stained to evaluate the morphological
structure of the organs.

5.17
Immunofluorescence and
Immunohistochemistry
The paraffin-embedded sections underwent
a process of removing
wax and rehydration in preparation for immunohistochemical analysis.
The activity of peroxidase was effectively blocked by a 3% hydrogen
peroxide solution (Beyotime, China). Following an overnight incubation
at 4 °C with primary antibodies (including KI67, CD4, and CD8),
the tissue sections were treated with a biotinylated secondary antibody.
Subsequently, the sections were exposed to a combination of streptavidin-horseradish
peroxidase (Beyotime, China).
In conducting fluorescent immunohistochemistry,
the first step involved fixing the sample at room temperature using
a 4% solution of paraformaldehyde for a duration of 20 min. Subsequently,
the sample was subjected to immersion in a 0.05% solution of Triton
X-100 for a period of 5 min. Following these treatments, the samples
were sealed overnight in a solution containing 2% BSA PBS and were
exposed to primary antibodies (CD8, PD1, CD11b, CD86, CD163, Ly6G)
at a temperature of 4 °C. The treated samples were then linked
with either Alexa Fluor or HRP (Beyotime, China) at room temperature.
A combined secondary antibody was applied and left to incubate for
1 h. To maintain the integrity of the nucleus, DAPI staining (Beyotime,
China) was used. After drying, images were captured by using a laser
scanning confocal microscope from Zeiss, Germany. A detailed list
of the antibodies employed in this investigation can be found in Table S1.

5.18
Mass
Cytometry
We obtained tissue
samples from the KN035 and KN035-MS tumor-bearing groups. We then
processed mouse tumor tissue using the Miltenyi Mouse Tumor Isolation
Kit (Miltenyi Biotec, Germany), in which Percoll removed debris and
divided red blood cells. For more experimental details, refer to our
previous research work.

5.19
Statistical Analysis
Most of the
analyses in this study were performed using GraphPad Prism 10.0 with
a p value of 0.05 for statistical significance. In all statistical
plots, data are expressed as the mean ± SD. The t-test was used to analyze the differences between the two sample
groups. One-way ANOVA was used to analyze the differences between
the three or more sample groups.

Supplementary Material