Cheng 

et al. BMC Medicine (2022) 20:411

https://doi.org/10.1186/s12916-022-02614-8

RESEARCH ARTICLE Open Access

Combination of an autophagy inhibitor

with immunoadjuvants and an anti‑PD‑L1
antibody in multifunctional nanoparticles
for enhanced breast cancer immunotherapy
Yibin Cheng1†, Caixia Wang2†, Huihui Wang1, Zhiwei Zhang1, Xiaopeng Yang1, Yanming Dong1, Lixin Ma1* and
Jingwen Luo1*    

Abstract 
Background:  The application of combination therapy for cancer treatment is limited due to poor tumor-specific
drug delivery and the abscopal effect.
Methods:  Here, PD-L1- and CD44-responsive multifunctional nanoparticles were developed using a polymer
complex of polyethyleneimine and oleic acid (PEI-OA) and loaded with two chemotherapeutic drugs (paclitaxel and
chloroquine), an antigen (ovalbumin), an immunopotentiator (CpG), and an immune checkpoint inhibitor (anti-PD-L1
antibody).
Results:  PEI-OA greatly improved the drug loading capacity and encapsulation efficiency of the nanoplatform, while
the anti-PD-L1 antibody significantly increased its cellular uptake compared to other treatment formulations. Phar-
macodynamic experiments confirmed that the anti-PD-L1 antibody can strongly inhibit primary breast cancer and
increase levels of CD4+ and CD8+ T cell at the tumor site. In addition, chloroquine reversed the “immune-cold” envi-
ronment and improved the anti-tumor effect of both chemotherapeutics and immune checkpoint inhibitors, while it
induced strong immune memory and prevented lung metastasis.
Conclusions:  Our strategy serves as a promising approach to the rational design of nanodelivery systems for simul-
taneous active targeting, autophagy inhibition, and chemotherapy that can be combined with immune-checkpoint
inhibitors for enhanced breast cancer treatment.
Keywords:  Immuno-chemotherapy, Anti-PD-L1 antibody, Multifunctional nanoparticles, Autophagy response

Background
Chemo-immunotherapy is considered a major strategy
for the treatment of established solid tumors. In recent
years, several methods have been developed for treat-
†
Yibin Cheng and Caixia Wang contributed equally to this work. ing post-surgery residual tumors and preventing tumor
*Correspondence: [email protected]; [email protected] recurrence and metastasis, including a combination of
1
State Key Laboratory of Biocatalysis and Enzyme Engineering, Hubei Key chemotherapeutics and cancer vaccines, checkpoint
Laboratory of Industrial Biotechnology, Hubei Collaborative Innovation therapies, and adoptive T-cell transfer [1–4]. Although
Center for Green Transformation of Bio‑Resources, School of Life Sciences, combination therapy is more effective than individual
Hubei University, Wuhan 430062, P. R. China
Full list of author information is available at the end of the article therapies, it cannot meet current clinical needs due to

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​
mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Cheng et al. BMC Medicine (2022) 20:411 Page 2 of 18

the presence of few T cells (“immune-cold” tumors) and initial burst release and promote controlled release; (2)
a high number of immunosuppressive cells in the tumor enhanced targeted delivery to antigen-presenting and
microenvironment (TME), including myeloid-derived tumor cells; (3) ability to induce effective anti-tumor
suppressor cells, tumor-associated macrophages, and T-cell responses; and (4) good biocompatibility.
regulatory T cells [5]. In this study, we describe an ultrafast, convenient, and
Recent studies have shown that autophagy in the TME universal self-assembly route for the preparation of an
plays a dual role [6, 7]: it can inhibit as well as promote integrative multilayered nanoplatform using immuno-
tumor growth by regulating the immune response [8] adjuvants for targeted tumor and lymph node delivery
and the survival, apoptosis, differentiation, activation, and durable antitumor immunity. The developed MNPs
effector function, and metastasis of immune cells [9]. For were loaded with two chemotherapeutic drugs [pacli-
example, autophagy promotes the survival and differ- taxel (PTX) and CQ], an antigen [ovalbumin (OVA)],
entiation of T cells in the TME [10], while autophagy of an immunopotentiator (CpG), and an immune check-
naive T cells protects them from mitochondrial apoptosis point inhibitor (anti-PD-L1 antibody, also known as
induced by reactive oxygen [11]. High levels of lactic acid atezolizumab). We found that atezolizumab neutral-
in tumors may interfere with naive T cell autophagy and ized PD-L1 and activate immune responses, while act-
weaken anti-tumor response [12]. Moreover, increased ing synergistically with chondroitin sulfate (CS) on the
apoptosis and functional defects of autophagy-deficient outmost layer of the nanoparticles to target PD-L1 and
regulatory T cells can enhance tumor resistance [13]. CD44 for combined autophagy modulation and cancer
Therefore, targeting autophagy may reshape the TME immunotherapy. The described nanoplatform also effi-
and reverse its “immune-cold” character. For instance, ciently delivered PTX and CQ to tumor cells as well as
autophagy inhibition can upregulate Th1 chemokines OVA and CpG to draining lymph nodes in a targeted and
and promote the infiltration of effector immune cells controlled manner, thus enhancing their availability and
at the tumor site [14]. In addition, it can enhance anti- limiting off-target effects. Furthermore, the combination
tumor T cell responses by improving antigen presenta- of CQ with atezolizumab reversed the immunosuppres-
tion and processing [15], while it can upregulate major sive TME, elicited strong tumor antigen-specific immune
histocompatibility complex (MHC) class I molecules on responses, and induced sustained tumor suppression in
the cell surface and prevent the autophagic degradation tumor-bearing mice (Fig. 1).
of granzyme B, thereby allowing cell lysis [16, 17].
Chloroquine (CQ) is a typical autophagy inhibitor that Methods
has been used along with various treatments in multiple Materials
clinical trials for anticancer treatment [18–20]. A recent OVA was purchased from Sigma-Aldrich (St Louis, MO,
study on a mouse model of pancreatic cancer has shown USA). Polyethyleneimine (PEI, molecular weight 1800
that treatment with CQ combined with dual immune- Da) and oleic acid (OA) were obtained from Shanghai
checkpoint blockade therapy using anti-PD-L1 and Aladdin Biochemical Technology Co., Ltd (Shanghai,
anti-CTLA4 antibodies can effectively inhibit autophagy- China). PTX and CQ were purchased from Dalian Mei-
mediated MHC-I degradation and enhance antitu- lun Biotech Co., Ltd. (Dalian, China). 4-Chlorobenzene-
mor immune responses [21]. These results suggest that sulfonate salt (DiD) was provided by Biotium (Hayward,
autophagy targeting combined with chemo-immuno- CA, USA). Fluorescein (FITC)-conjugated rat anti-mouse
therapy using integrative multifunctional nanoparticles CD4 antibody, PE-conjugated rat anti-mouse CD8a anti-
(MNPs), which are generally known for their excellent body, and the Annexin V-FITC/PI apoptosis detection kit
stability, biocompatibility, and encapsulation efficiency were purchased from BioLegend, Inc. (San Diego, CA,
[22], may improve the therapeutic potential of individual USA). Anti-GAPDH was obtained from CWBIO (Bei-
treatments. Nevertheless, the successful design of MNPs jing, China), and monoclonal anti-LC3B antibodies were
requires an appropriate way to add immunoadjuvants, purchased from Sigma-Aldrich. Anti-p62 (SQSTM1)
since the separate administration of chemotherapeutics, antibodies were from Proteintech Group (Rosemont, IL,
cancer antibodies, antigens, and immunopotentiators USA), and horseradish peroxidase (HRP)-conjugated-
may lead to severe side effects and non-specific, systemic goat anti-rabbit IgG (H + L) secondary antibody was
immune responses [23]. Given also that most cancer vac- from Invitrogen (catalog no. 31460, Carlsbad, CA, USA).
cines consisting of cancer antigens and immunopotentia- FITC-conjugated ImmunoPure goat anti-rabbit IgG (H
tors have low efficiency due to initial burst and off-target + L) antibodies were purchased from Feiyi Technology
release, it is clear that adjuvants should show (1) high (Wuhan, China), while anti-PD-L1 antibody (atezoli-
encapsulation efficiency for cancer chemotherapeutics, zumab) was from Selleck Chemicals (Houston, TX, USA).
antigens, and immunopotentiators in order to prevent Mouse mammary breast tumor cells (4T1) and Hela cells
 Cheng et al. BMC Medicine (2022) 20:411 Page 3 of 18

Fig. 1  Schematic illustration of the combined application of chemotherapy, immunotherapy, and PD-L1 blockade therapy in tumor-bearing mice.
CpG, immunopotentiator; CQ, chloroquine, DC, dendritic cell; HS15, Solutol HS15®; OVA, ovalbumin; PEI-OA, polyethyleneimine-oleic acid polymer;
PTX, paclitaxel

were obtained from the Chinese Academy of Science Cell Synthesis and characterization of polymers
Bank (Shanghai, China). PEI (0.05 mmol, 1800 Da) was dissolved in 10 mL of
methanol, and 0.8 mmol of 1-ethyl-3-(3-dimethyl-
aminopropyl) carbodiimide hydrochloride was added.
Animals After stirring for 10 min, a solution of 0.8 mmol OA
Healthy female BALB/c mice (6–8 weeks) were obtained in 10 mL methanol was slowly added using a constant-
from Wuhan Institute of Virology (Wuhan, China). All pressure dropping funnel, and the reaction mixture was
animal experiments were approved by the Ethics Com- stirred at room temperature overnight under nitrogen.
mittee of Wuhan Institute of Virology (Approval number: The mixture was then transferred to a dialysis bag with
WIVA04202102). a molecular-weight cutoff (MWCO) of 3500 Da and
Cheng et al. BMC Medicine (2022) 20:411 Page 4 of 18

dialyzed against a 1:1 water/methanol mixture for 24 h, Encapsulation efficiency (%) =

amount of drug in nanoparticles
× 100
then against purified water for another two days. The total drug dose
obtained oleic acid-modified polyethyleneimine (PEI- (2)
OA) material was ultimately freeze-dried, weighed, and OVA- and CpG-loaded MNPs were also purified by
stored at −20 °C. 1H NMR spectra were recorded in ultrafiltration, and their content was determined using
dimethyl sulfoxide on a Varian UNITY INOVA400 NMR the bicinchoninic acid (BCA) protein assay kit (Pierce,
spectrometer (Palo Alto, CA, USA). Rockford, IL, USA) and electrophoresis on agarose gel
(1%) stained with GoldView (Generay Biotech, Shanghai,
Preparation and characterization of MNPs China).

tion method. PEI-OA, Soluplus and Solutol HS15®

Nanoparticles 1 were prepared by the thin-film hydra-
Chromatographic conditions
(HS15) were first dissolved in methanol at a molar ratio Chromatographic analysis was performed on an Agilent
of 1:6:2. Soluplus and HS15 were the basic materials for 1290 liquid chromatography system with an Agilent C18
nanoparticles preparation. They were combined to make column (250 mm × 4.6 mm, 5 μm). PTX concentrations
nanoparticles possess uniform and suitable particle size. were measured at 227 nm using methanol/water (65:35)
PEI-OA was added to give nanoparticle-positive charge. as the mobile phase. CQ concentrations were measured
After removing the organic solvent by rotary evapora- at 343 nm using a mixture of acetonitrile, methanol,
tion at 37 °C, the obtained thin film was hydrated with water, and ammonium hydroxide (78:20:35:1.65) as the
5% glucose solution at 60 °C to form nanoparticles 1. mobile phase. The flow rate was 1.0 mL/min and the col-
Next, nanoparticles 1 were added to a solution of OVA umn temperature was 40 °C.
(150 μg/mL), CpG (50 μg/mL), and CS (16 mg/mL) in
5% glucose and stirred gently to give nanoparticles 2, Cytotoxicity
which were then rapidly mixed with a solution of ate- 4T1 cells were seeded in 96-well plates (5 × ­103 cells per
zolizumab (2 mg/mL) using a vortex shaker for 1 min at well in 100 μL). After 12 h, cells were treated with fresh
room temperature to afford the desired MNPs (3). Unen- medium (control) or culture medium containing differ-
capsulated drug was removed by ultrafiltration over a ent concentrations of blank nanoparticles (N), PTX-S,
membrane (MWCO 1000 kDa, Spectrum Laboratories, CpG+OVA-S, CpG+OVA+PTX-S, PTX-N, CpG+OVA-
CA, USA), and nanoparticles were diluted to 1:100 in 5% N, CpG+OVA+PTX-N, CpG+OVA+PTX+CQ-N, or
glucose solution, transferred to disposable polystyrene CpG+OVA+PTX+CQ-N/A (PTX, 0.25–40 μg/mL; CQ,
cuvettes, and analyzed using a Zetasizer Nano ZSP (Mal- 0.05–8 μg/mL). Cell proliferation after incubation with
vern Instruments, Worcestershire, UK). In general, the different formulations for 24 h was determined using
ratio of PTX: CQ: PEI-OA: Soluplus: HS15: OVA: CpG: the 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetra-
CS:atezolizumab was 200:40:300:1800:600:3:1:1600:100. zolium bromide (MTT) assay (Sigma-Aldrich, St Louis,
To confirm that the nanoparticles could be loaded MO, USA).
simultaneously with chemotherapeutic drugs and bio-
molecules, MNPs were incubated with DiD- and/or
Apoptosis
FITC-labeled OVA, then analyzed with a high-sensitiv-
ity flow cytometer (Beckman Coulter, Brea, CA, USA) 4T1 cells were seeded in 12-well plates ­(105 cells/well). After
equipped with three single-photon-counting avalanche 48 h, cells were incubated with blank nanoparticles (N), PTX-
photodiode detectors for the simultaneous detection of S, CpG+OVA-S, CpG+OVA+PTX-S, PTX-N, CpG+OVA-
N, CpG+OVA+PTX-N, CpG+OVA+PTX+CQ-N, or
metric data were analyzed using FlowJo™ v10.0 (Tree
side scattering and two-color fluorescence. Flow cyto-
CpG+OVA+PTX+CQ-N/A for 48 h. The final concentra-
Star, Inc., Ashland, OR, USA). tion of PTX was 1 μg/mL and that of CQ was 0.2 μg/mL. At
48 h post-incubation, the treated cells were stained using the
Drug loading capacity and encapsulation efficiency
Annexin V-FITC/PI apoptosis detection kit and analyzed by
Isolated, unencapsulated drugs were redissolved in meth- flow cytometry (Becton, Dickinson and Company, Franklin
anol and detected by HPLC (Agilent, Santa Clara, CA, Lakes, NJ, USA).
USA). Loading capacity and encapsulation efficiency
were calculated as follows: Cellular uptake
4T1 cells were seeded in six-well plates (3 × ­105 cells/
amount of drug in nanoparticles
Loading capacity (%) = × 100 well). After 24 h, DiD+FAM-CpG-S or DiD+FAM-CpG-
total mass of nanoparticles
N/A (both with a DiD concentration of 1 μg/mL) was
(1)
added and cells were incubated for 2 or 4 h, respectively.
 Cheng et al. BMC Medicine (2022) 20:411 Page 5 of 18

Afterward, the cells were digested, collected, and ana- secondary antibody at room temperature for 1 h, while
lyzed by flow cytometry (Beckman Coulter). cell nuclei were stained with DAPI for 5 min in the
4T1 cells were seeded in six-well plates with a cover- dark. Images were acquired with a confocal fluores-
slip in each well and incubated with DiD+FAM-CpG-S cence microscope (ZEISS, LSM 980 with Airyscan 2,
or DiD+FAM-CpG-N/A as above. At 2 or 4 h post-treat- Oberkochen, Germany). For autophagy flux analysis,
ment, cells were fixed with 4% polyformaldehyde, washed HeLa cells were transfected with mCherry-GFP-LC3B
twice with phosphate-buffered saline (PBS), and stained tandem reporter and imaged at 48 h post-transfection by
with DAPI (0.5 μg/mL; Biofroxx, Einhausen, Germany) confocal fluorescence microscopy.
for 5 min. The fluorescence intensity was observed with
a Nikon A1R+ confocal microscope (Tokyo Metropolis,
Tumor targeting ability
Japan).
Female BALB/c mice were injected subcutaneously
with 4T1 cells (100 μL; 2 × ­105 cells/mouse). On day 7,
Cell culture and transfection mice with a tumor volume of ~100 ­mm3 were randomly
4T1 and HeLa cells were cultured in RPMI-1640/ Dul- divided into two groups to receive DiD-S or DiD-N/A via
becco’s modified Eagle medium (DMEM) (HyClone, intravenous injection (DiD, 1 mg/kg). The distribution of
Logan, UT, USA) with 10% fetal bovine serum (PAN-Bio- nanoparticles at 2, 4, 8, 24, and 48 h post-injection and
tech, Aidenbach, Germany). HeLa cells were transfected the fluorescence intensity of major organs ex  vivo were
with mCherry-GFP-LC3B using Lipofectamine 2000 examined using a Lumina III Imaging System (Perki-
(Invitrogen) and cultured with different formulations in nElmer, MA, USA). Frozen tumor sections incubated
RPMI-1640/DMEM for 4 h. with rabbit anti-mouse CD44 and PD-L1 antibodies
and subsequently with FITC-labeled donkey anti-rabbit
Western blot analysis secondary antibody were also stained with DAPI and
Western blot analysis was performed as described [24]. observed by confocal microscopy.
Proteins were first extracted from HeLa cells or tumor
samples. Whole extracts were then analyzed by sodium
dodecyl sulfate-polyacrylamide gel electrophoresis and In vivo antitumor effects
transferred to a 0.45-μm polyvinylidene fluoride mem- A model of 4T1 tumor-bearing BALB/c mice was
brane (Roche Diagnostics, IN, USA). Membranes were established as described above. On day 5, mice were
blocked with 5% bovine serum albumin (BSA) in TBST randomly divided into nine groups (n = 5) and
(20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween- injected intravenously twice with CpG+OVA+PTX-
20), incubated with primary antibodies at 4 °C overnight S, PTX-N, CpG+OVA-N, CpG+OVA+PTX-N,
and then with the HPR-labeled secondary antibody. Pro- CpG+OVA+PTX-N/A, CpG+OVA+PTX+CQ-N, or
teins were detected using an enhanced chemiluminescent CpG+OVA+PTX+CQ-N/A every 5 days. A 5% glucose
reagent kit (Millipore, Billerica, MA, USA) and densi- solution was used as control. The PTX dose was 10 mg/
tometry images were obtained using ImageJ software kg, the CQ dose was 2 mg/kg, and the atezolizumab dose
(vJ2, NIH, Maryland, US). Data were processed using was 50 μg/mouse. The tumor size and body weight were
Microsoft Excel. measured every 2 days. At 5 days after the final injection,
all mice were sacrificed and the tumors of each group
Immunofluorescence analysis were weighed, photographed, and sampled along with
Immunofluorescence analysis was performed according major organs for H&E staining.
to our previous study [25]. First, the tumor and/or major
organ tissues were embedded in paraffin and cut into Immune cell activation and cytokine secretion
2-μm sections using a slicer (Leica, Bensheim, Germany). Freshly isolated tumor tissues were cut into pieces,
4T1 and HeLa cells cultured on glass coverslips were also digested with collagenase IV for 1 h, and filtered with
treated with different preparations for 4 h. Then, the cells cell strainers to prepare a single-cell suspension. About
and tissue sections were fixed with 4% paraformaldehyde 2 × ­107 cells/sample were then incubated with different
at 4 °C for 30 min and antigens were retrieved with an antibodies and the number of the corresponding T cells
AR buffer (Leica, AR9961) and permeabilized with 0.1% was determined by flow cytometry. T cells were detected
Triton X-100 in PBS for 20 min. After blocking with 5% using PE rat anti-mouse CD8α and FITC rat anti-mouse
BSA at room temperature for at least 30 min, the sam- CD4 antibodies. Part of the tumor and spleen tissues
ples were incubated with primary antibodies at 4 °C over- were also fixed in paraffin for immunohistochemical
night. After washing three times with PBS, the samples staining. The levels of IFN-γ and TNF-α in the animal
were incubated with an indirect immunofluorescence serum were analyzed with an ELISA kit (Sigma-Aldrich).
Cheng et al. BMC Medicine (2022) 20:411 Page 6 of 18

Statistical analysis protein, serving as a universal platform for chemo-immu-

Differences between multiple groups were analyzed for notherapy-based combination therapy.
significance using one-way analysis of variance (ANOVA) Next, we examined whether MNPs could serve as
followed by Tukey’s multiple comparisons test, while the a versatile platform for the immobilization of atezoli-
two-tailed Student’s t-test was used for two groups. The zumab. Dynamic light scattering measurements showed
results were presented as mean ± standard deviation that the average hydrodynamic diameter of nanopar-
(SD). Differences associated with *P < 0.05 were consid- ticles increased by approximately 11 nm after incuba-
ered statically significant. tion with atezolizumab at 4 °C for 1 h, suggesting the
successful immobilization of the antibody (Fig.  2h). The
Results nanoparticle size also increased after incubation, with
Design and characterization of MNPs nanoparticles 3 having the largest size, followed by nano-
In order to construct a multifunctional treatment plat- particles 2 and 1. In addition, nanoparticles 1 showed a
form able to co-encapsulate small chemical drugs, nucleic positive surface charge, while nanoparticles 2 and 3 had
acid, peptide and/or protein, we synthesized a polyethyl- a negative surface zeta potential due to the anionic CS
eneimine-oleic acid polymer complex (PEI-OA) by con- polymer. The encapsulation efficiency of atezolizumab-
jugating OA to the carboxyl groups of PEI (molecular coated MNPs co-loaded with PTX, CQ, CpG, and OVA
weight 1800 Da) (Fig. 2a). Nanoparticles 1 were then pre- (CpG+OVA+PTX+CQ-N/A) was 98.75 ± 1.6% for PTX
pared by the thin-film hydration method using PEI-OA, and 96.8 ± 2.1% for CQ, while the total drug loading effi-
HS15, and Soluplus (Fig. 1a). Next, nanoparticles 1 were ciency was around 4.30% for PTX and 0.86% for CQ.
added to a solution of CpG and OVA in 5% glucose to
prepare nanoparticles 2. The positive charge of nanopar- Cellular uptake, cytotoxicity, and apoptosis of MNP‑treated
ticles 1 was partially neutralized by the negative charge tumor cells
of CpG and OVA (Fig. 2b). The nanoparticle solution was To examine the effect of the developed MNPs on the
then added to a CS solution to obtain completely nega- cellular uptake of small and macromolecular drugs, we
tively charged nanoparticles 2, which gave nanoparticles used the 4T1 cell line that overexpresses both the CD44
3 after reaction with atezolizumab. and PD-L1 receptors [26]. Confocal microscopy showed
Scanning electron microscopy indicated that nano- that the fluorescence intensity of DiD+FAM-CpG-N/A
particles 2 and 3 were homogeneously spherical, but was stronger than that of a solution containing free
nanoparticles 3 had a rougher surface and a remark- DiD and FAM-CpG (DiD+FAM-CpG-S) (Fig.  3a), and
ably different size distribution and morphology (Fig. 2c), that the intensity increased with incubation time from
confirming that a layer of antibodies had adsorbed to 2 to 4 h. Moreover, the number of DiD-positive cells in
the MNP surface. To verify that MNPs were generated the DiD+FITC-OVA-N/A group at 2 h post-incubation
through electrostatic interactions between nanoparticles was 18 times higher than the number in the DiD+FITC-
1 and CpG, nanoparticles 1 were incubated with fluores- OVA-S group, and the number increased with incubation
cein amidite-labeled CpG (FAM-CpG) and analyzed by time (Fig.  3b). Similar results were observed for FITC-
confocal microscopy imaging. DiD was encapsulated in OVA-positive cells in both groups, while the intensity of
MNPs (DiD-N/A) and showed red fluorescence. FAM- DiD+FITC-OVA-N/A was found to be nine times higher
CpG was loaded in MNPs and showed green fluores- than that of DiD+FITC-OVA-S (Fig.  3c). These findings
cence. The particles co-loaded with DiD and FAM-CpG indicate that cellular uptake was time-dependent and
(DiD+FAM-CpG-N/A) showed obvious green and red that MNPs promoted drug penetration, consistent with
fluorescence as well as obvious colocalization of the two our previous study [22].
types of fluorescence (Fig.  2d), implying the successful The cytotoxicity of the developed nanoplatform was
integration of nanoparticles 1 and CpG. These results measured in  vitro with MTT assay. Among the tested
were confirmed by agarose gel electrophoresis (Fig.  2e), groups, cells treated with blank nanoparticles showed the
which indicated the complete encapsulation of CpG into highest viability, suggesting that the carrier was non-toxic
the nanoparticles. and caused negligible side effects (Fig. 3d). Moreover, the
To examine whether the described nanoplatform ­IC50 value was 33.21 μg/mL for free PTX and 22.85 μg/
could immobilize large-molecular-weight proteins, mL for a solution containing free CpG, OVA, and PTX
MNPs were incubated with FITC-labeled OVA and ana- (CpG+OVA+PTX-S). However, after loading onto nano-
lyzed by Western blotting and flow cytometry (Fig.  2f, particles, the ­IC50 value of PTX-loaded MNPs (PTX-N)
g). The appearance of both FITC and DiD signals in the and CpG/OVA/PTX-loaded MNPs (CpG+OVA+PTX-
DiD+FITC-OVA-N/A group confirmed that MNPs N) decreased to 22.18 and 18.12 μg/mL, respectively,
could immobilize simultaneously a nuclei acid and a reflecting the beneficial effect of the nanoplatform on
 Cheng et al. BMC Medicine (2022) 20:411 Page 7 of 18

Fig. 2  Construction and characterization of multifunctional nanoparticles (MNPs). a Chemical structure of the polyethyleneimine-oleic
acid (PEI-OA) polymer. b Nanoparticles 1 were prepared by the thin-film hydration method. Encapsulation of an immunopotentiator (CpG)
and ovalbumin (OVA) into 1 by electrostatic interactions, followed by surface coating with atezolizumab and chondroitin sulfate generated
nanoparticles 3. c Transmission electron micrographs (TEM) of nanoparticles 1–3. Scale bar, 200 nm. d Confocal microscopic images of DiD-N/A,
FAM-CpG-N/A, and DiD+FAM-CpG-N/A. Scale bar, 200 nm. e Agarose gel electrophoresis of DiD+FAM-CpG-N/A. f Western blot analysis of
DiD-N/A, FITC-OVA-N/A, and DiD+FITC-OVA-N/A. Free FITC-OVA solution was used as control. g Flow cytometry of DiD-N/A, FITC-OVA-N/A, and
DiD+FITC-OVA-N/A. h Representative size distribution of nanoparticles 3, as determined by dynamic light scattering. i Size, polydispersity index
(PDI), and zeta potential of nanoparticles 1–3. Data are shown as mean ± SD (n = 4). DiD, 4-chlorobenzenesulfonate salt; FAM, fluorescein amidite;
FITC, fluorescein; N, nanoparticles; S, solution

cytotoxicity. Further loading of CQ reduced the I­ C50 value Cells treated with the same formulations as in the
of CpG+OVA+PTX+CQ-N to 5.54 μg/mL, while ate- cell viability study were then stained with FITC-
zolizumab immobilization (CpG+OVA+PTX+CQ-N/A) Annexin V and propidium iodide (PI). Flow cytom-
resulted in an ­IC50 value of 5.23 μg/mL (Fig.  3d). These etry showed that most cells were in early-stage
results indicate that the synergistic effect of CQ and PTX apoptosis or were already dead (Fig.  3e, f ). Compared
along with atezolizumab can significantly enhance the to the PTX-S and CpG+OVA+PTX-S groups, PTX-N
killing effect of individual drugs in cancer cells. and CpG+OVA+PTX-N showed 1.29 and 1.55 times
Cheng et al. BMC Medicine (2022) 20:411 Page 8 of 18

higher ability to induce apoptosis, respectively, imply- significantly accumulated in the CpG+OVA+PTX+CQ-
ing that MNPs can greatly promote cellular uptake. N/A group compared to CpG+OVA+PTX-N/A, suggest-
CpG+OVA+PTX+CQ-N led to higher apoptosis or ing that CQ was effective in reducing the flux as evidenced
necrosis rates than CpG+OVA+PTX-N, indicating that by accumulation of LC3B-II and SQSTM1 (Fig. 4f, g).
CQ can strongly promote cell apoptosis. Furthermore,
67.06% of cells in the CpG+OVA+PTX+CQ-N/A group In vivo targeting ability of MNPs
were in the apoptotic or necrotic stage, which was 1.1 The in  vivo targeting ability of CpG+OVA+PTX+CQ-
times higher than in the CpG+OVA+PTX+CQ-N group N/A was examined by living imaging. At 2 h post-injection,
(60.87%) (Fig. 3e, f ), suggesting that atezolizumab favors DiD-loaded CpG+OVA+PTX+CQ-N/A (N) showed
MNP uptake, leading to increased apoptosis. a stronger fluorescence intensity at the tumor site than a
solution of free DiD (S), suggesting that atezolizumab and
Effect of MNPs on autophagosome formation CS greatly promoted nanoparticle accumulation (Fig. 5a).
The effect of MNPs on autophagosome formation in 4T1 However, at 24 and 48 h post-injection, the fluorescence
cell lines was evaluated using an immunofluorescence intensity decreased in both groups due to body scaveng-
assay with an anti-LC3B antibody, followed by confocal ing, although the residual fluorescence intensity of the N
microscopy. Compared to the control group, the number group remained higher than that of the S group. Ex  vivo
of LC3B-positive puncta increased after treatment with imaging of major organs also showed that fluorescence was
CpG+OVA+PTX-N/A (Fig.  4a, b), indicating that this mainly distributed in the liver, lung, and spleen (Fig.  5b),
formulation promoted autophagosome formation. This as the liver and spleen are rich in reticuloendothelial cells,
result was confirmed by transmission electron micros- such as macrophages, and express the CD44 receptor,
copy of HeLa cells treated with CpG+OVA+PTX-N/A favoring the engulfment of the CD44-responsive MNPs. In
(Fig.  4c). Given that nanomaterial-induced autophagy addition, the fluorescence intensity at 48 h post-injection
may promote cancer cell survival [27], cells were then was 3.05 times higher in the N group than in the S group,
treated with CpG+OVA+PTX+CQ-N/A, which confirming that atezolizumab and CS synergistically pro-
resulted in significant autophagosome puncta accumula- moted the targeting ability of nanoparticles toward 4T1
tion due to the inhibitory effect of CQ (Fig. 4b). tumor cells (Fig. 5c). Further staining of tumor slices with
To investigate the mechanism of nanoparticle-asso- CD44 and PD-L1 showed that CpG+OVA+PTX+CQ-
ciated autophagy, we examined autophagy flux in HeLa N/A emitted stronger fluorescence than the free DiD solu-
cells treated with different formulations using a tandem tion (Fig. 5d), consistent with the in vivo imaging results.
fluorescent indicator, mCherry-GFP-LC3B (GFP, green These results suggest that the enhanced targeting and pen-
fluorescent protein) [28]. Green fluorescence is very sen- etration ability of the developed MNPs was due mainly to
sitive to the acidic environment of lysosomes and quickly the tumor-homing effect of atezolizumab and CS.
quenched in autolysosomes, so red fluorescence was
attributed to autolysosomes. CpG+OVA+PTX-N/A sig- Anti‑tumor effect of MNPs
nificantly promoted autophagosome formation, but even To evaluate the in  vivo antitumor effect of the different
greater autophagosome accumulation was observed in preparations, we used a subcutaneous 4T1 breast cancer
the CpG+OVA+PTX+CQ-N/A group, as CQ inhibited model of Balb/c mice that usually show decreased long-
autophagosome–lysosome fusion (Fig. 4d, e). term survival, even after various treatments [5, 29]. On
HeLa cells in DMEM were treated with different formu- day 0, mice bearing a large breast tumor were subcutane-
lations to test their ability to induce autophagy. Western ously injected with 4T1 cells into the right flank. At 5 days
blot analysis showed that LC3B-II level was significantly post-injection, the tumor volume reached ~50 m ­ m3. On
upregulated in the CpG+OVA+PTX-N/A group com- days 5 and 10, mice were injected intravenously with dif-
pared to controls, while its downstream substrate ferent formulations and the antitumor efficacy, immune
SQSTM1 was downregulated. LC3B-II and SQSTM1 were response, and autophagy response were analyzed (Fig. 6a).

(See figure on next page.)

Fig. 3  In vitro characterization of multifunctional nanoparticles (MNPs). a Confocal microscopy images of 4T1 cells incubated with various
formulations for 2 and 4 h. Scale bar, 100 nm. b, c Number of b DiD-positive and c FITC-positive 4T1 cells after incubation with DiD+FITC-OVA-N/A
and DiD+FITC-OVA-S for 2 and 4 h (DiD, 1 μg/mL). Student’s t-test was performed. *P < 0.05. d Cell viability of 4T1 cells after incubation with
different formulations for 24 h, as determined by the MTT assay. Blank nanoparticles were used as control. e Apoptosis of 4T1 cells stained with
FITC-Annexin V and propidium iodide (PI) after incubation with different formulations for 24 h. f Percentage of cells in the apoptotic or necrotic
stage after treatment with different formulations. Data are shown as mean ± SD (n = 3). Cells without any treatment were set as control group.
CpG, immunopotentiator; CQ, chloroquine; DAPI, 4′,6-diamidino-2-phenylindole; DiD, 4-chlorobenzenesulfonate salt; FAM, fluorescein amidite; FITC,
fluorescein; N, nanoparticles, N/A, nanoparticles coated with atezolizumab; OVA, ovalbumin; PTX, paclitaxel; S, solution
 Cheng et al. BMC Medicine (2022) 20:411 Page 9 of 18

Fig. 3  (See legend on previous page.)

Cheng et al. BMC Medicine (2022) 20:411 Page 10 of 18

Fig. 4  Effect of multifunctional nanoparticles on autophagosome formation. a Signals of LC3B puncta detected in 4T1 cells cultured in RPMI-1640
medium with different formulations for 4 h (a 5% glucose solution was used as control), as determined by immunofluorescence assay with
an anti-LC3B antibody, followed by confocal microscopy. LC3B-positive puncta were detected in the cytoplasm by fluorescein-conjugated
ImmunoPure goat anti-rabbit IgG (green). Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). Enlarged boxes highlight LC3B
signals. Scale bar: 5 μm. b Number of LC3B-positive puncta per cell quantified from ~20 cells treated with different formulations. The means ± SD
are from 3 independent experiments. One-way ANOVA was performed. *P < 0.05; **P < 0.01. (n = 3 independent experiments.) c Transmission
electron micrographs of autophagosomes in the cytoplasm of HeLa cells treated with CpG+OVA+PTX-N/A. Scale bar, 500 nm. d Confocal
microscopy images of HeLa cells transfected with the autophagy dual fluorescent reporter mCherry-GFP-LC3B and cultured with different
formulations in Dulbecco’s modified Eagle medium for 4 h. Scale bar: 5 μm. e Number of autophagosomes and autolysosomes in HeLa cells treated
with different formulations (>15 cells/experiment). Co-localized dots were counted. Data are presented as means ± SD. * P < 0.05, ** P<0.01 (n =
3 independent experiments). f Representative Western blots for LC3B-II and SQSTM1. g Relative protein levels of LC3B-II and SQSTM1, normalized
to levels of GAPDH. Cells without any treatment were set as the control group. Data are shown as mean ± SD (n = 3). *P < 0.05, **P < 0.01. CpG,
immunopotentiator; CQ, chloroquine; N, nanoparticles; N/A, nanoparticles coated with atezolizumab; OVA, ovalbumin; PTX, paclitaxel; S, solution
 Cheng et al. BMC Medicine (2022) 20:411 Page 11 of 18

Fig. 5  In vivo targeted delivery of multifunctional nanoparticles (MNPs). a Fluorescence imaging of 4T1 breast tumor-bearing mice at 2, 4, 8, 24,
and 48 h post-administration of DiD-loaded MNPs (N) or a solution of free DiD (S). b Nanoparticle distribution in tumors and major organs at 48
h post-injection of DiD-loaded MNPs (N) or a solution of free DiD (S). c Semiquantitation of total radiant efficiency in isolated tumors and major
organs at 48 h post-injection of DiD-loaded MNPs (N) or a solution of free DiD (S). Data are shown as mean ± SD (n = 3). Student’s t-test was
performed. *P < 0.05. d, e Confocal imaging of frozen tumor sections. Tumor cells were stained with d anti-CD44 antibody (green) or e anti-PD-L1
antibody (green). Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Red fluorescence (DiD) indicated MNPs. Mice injected
with 5% glucose solution were set as control group. CQ, chloroquine; DiD, 4-chlorobenzenesulfonate salt

Estimation of the tumor volume in the different treat- delayed tumor progression, indicating the beneficial syn-
ment groups indicated that CpG+OVA+PTX-S could not ergistic effect of chemotherapy and immune-checkpoint
suppress tumor growth due to the poor targeting capacity blockade therapy.
of free drugs (Fig.  6b). In contrast, a stronger suppressive The strongest suppressive effect of
effect was observed in the CpG+OVA+PTX-N group, CpG+OVA+CQ+PTX-N/A compared to the other
suggesting that MNPs effectively targeted the tumor site. nanoformulations was also confirmed in measure-
Although further loading of the nanoplatform with CQ ments of tumor weight, which was < 0.3 g only for
enhanced the tumor-suppressive effect and inhibited the CpG+OVA+CQ+PTX-N/A group (Fig.  6c).
autophagy, CpG+OVA+PTX+CQ-N did not completely Based on these results, we also calculated the inhi-
block tumor growth when compared to CpG+OVA+PTX- bition rates of the different preparations, which
N. Conversely, CpG+OVA+CQ+PTX-N/A significantly revealed several treatments that were ineffective
Cheng et al. BMC Medicine (2022) 20:411 Page 12 of 18

Fig. 6  In vivo antitumor efficacy of multifunctional nanoparticles in 4T1 breast tumor-bearing Balb/c mice. a Establishment of a subcutaneous 4T1
breast cancer model of Balb/c mice. A 5% glucose solution was used as control. b Change in tumor volume over time after treatment with different
formulations. Data are shown as mean ± SD (n = 5). c Weight of excised tumors after the completion of the experiment. Data are shown as mean
± SD (n = 5). d Body weight of 4T1-bearing mice treated with different formulations for up to 15 days. Data are shown as mean ± SD (n = 5). e
Survival rates of mice over time after treatment with different formulations. Data are shown as mean ± SD (n = 10). f Tumor sections collected on
day 15 after the indicated treatments and visualized by TUNEL labeling. Scale bar, 50 μm. Student’s t-test was performed. *P < 0.05 vs control. Mice
injected with 5% glucose solution were set as control group. CpG, immunopotentiator; CQ, chloroquine; DAPI, 4′,6-diamidino-2-phenylindole;
N, nanoparticles; N/A, nanoparticles coated with atezolizumab; OVA, ovalbumin; PTX, paclitaxel; S, solution; TUNEL, terminal deoxynucleotidyl
transferase dUTP nick end-labeling

based on poor tumor inhibition rate: PTX-N (16.03%), atezolizumab immobilization (CpG+OVA+PTX+CQ-
CpG+OVA-N (35.72%), CpG+OVA+PTX-S (21.71%), N/A), the inhibition rate increased further to 79.45%, i.e.,
and CpG+OVA+PTX-N (30.24%). In contrast, 1.15 times higher than CpG+OVA+PTX+CQ-N, high-
CpG+OVA+PTX+CQ-N reduced tumor volume by lighting the therapeutic effect of the anti-PD-L1 antibody.
69.45%, which was 2.3 times higher than the reduction This conclusion was confirmed by the tumor inhibition
achieved by CpG+OVA+PTX-N, indicating that CQ rate of CpG+OVA+PTX-N/A (59.97%), which was 1.98
greatly enhanced the therapeutic effect of MNPs. After times higher than that of CpG+OVA+PTX-N (30.24%).
 Cheng et al. BMC Medicine (2022) 20:411 Page 13 of 18

Consistent with these results, analysis of the tumor mor- terminal deoxynucleotidyl transferase dUTP nick end-
phology suggested that CpG+OVA+PTX+CQ-N/A had labeling (TUNEL) and with 4′,6-diamidino-2-phe-
the strongest suppressive effect (Additional file 1: Fig. S1). nylindole (DAPI) (Fig.  6f ). Thus, the administration
The body weight of mice in different groups did of an autophagy inhibitor along with immunoadju-
not change significantly during treatment (Fig.  6d), vants and an immune-checkpoint inhibitor is a more
suggesting that the developed MNPs were safe and effective cancer treatment strategy than the individual
biocompatible. Staining of all major tissues (heart, therapies.
liver, spleen, lung, kidney) with hematoxylin and
eosin (H&E) indicated no obvious toxicity for MNPs
in Balb/c mice (Additional file  1: Fig. S2). In addi- Effect of autophagy on the anticancer activity of MNPs
tion, hemolysis assay (Additional file  1: Fig. S3), rou- The effect of autophagy on the anticancer activity of
tine blood examination (Additional file  1: Fig. S4), MNPs in vivo was assessed by immunofluorescence anal-
apoptosis (Additional file  1: Fig. S5), and autophagy ysis. Compared to the control group, CpG+OVA+PTX-
(Additional file  1: Fig. S6) characterization in the N/A significantly promoted autophagosome formation,
liver and spleen of mice further proved the safety as indicated by the increased number of LC3B-positive
of CpG+OVA+PTX+CQ-N/A for normal organs. puncta. In addition, autophagy inhibitor CQ was added
Among the treatment groups, CpG+OVA+PTX+CQ- into MNP created as CpG+OVA+PTX+CQ-N/A
N/A achieved the highest survival rate (Fig.  6e) and which led to significantly higher autophagosome accu-
caused severe nuclei damage and cytosol degrada- mulation than CpG+OVA+PTX-N/A, indicating that
tion in tumor cells, as determined by co-staining with inhibition of autophagy flux in the tumor greatly favors

Fig. 7  Autophagy inhibition in the tumor enhances the anticancer activity of MNPs. a LC3B-positive puncta, as detected by immunofluorescence
assay with an anti-LC3B antibody, followed by confocal microscopy. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; blue).
Scale bar, 20 μm. b Number of LC3B-positive puncta per cell quantified from ~20 cells treated with different formulations. c Representative
Western blots for SQSTM1 and LC3B-II. d Relative protein levels of LC3B-II and SQSTM1, normalized to levels of GAPDH. Data are shown as mean
± SD (n = 3). One-way ANOVA was performed. *P < 0.05, **P < 0.01. Mice injected with 5% glucose solution were set as the control group. CpG,
immunopotentiator; CQ, chloroquine; N, nanoparticles; N/A, nanoparticles coated with atezolizumab; OVA, ovalbumin; PTX, paclitaxel; S, solution
Cheng et al. BMC Medicine (2022) 20:411 Page 14 of 18

autophagosome accumulation (Fig.  7a, b). Western blot immunotherapy of advanced breast cancer, delay tumor
analysis also confirmed that CpG+OVA+PTX-N/A sig- growth, and prevent recurrence. Our formulations
nificantly upregulated LC3B-II compared to the control appear to exert these effects by inhibiting immune eva-
group, while CpG+OVA+PTX+CQ-N/A promoted the sion of tumor cells via immunosuppression reversal and
accumulation of LC3B-II at the tumor site (Fig.  7c, d), inducing a T cell-mediated antitumor immune response
confirming that autophagy inhibition enhances the anti- through dendritic cell maturation, upregulation of
cancer activity of MNPs. tumor-associated antigens and cytokines, and generation
of T and natural killer cells.
In vivo immune response
Cytotoxic T lymphocytes can directly kill cancer cells Immune‑memory effect and metastasis inhibition
by releasing perforin, granzymes, and granulysin, while During T cell proliferation and differentiation, some
helper T lymphocytes act by regulating adaptive immu- memory cells are differentiated to exert long-term anti-
nity [30]. Here, we determined the levels of CD8+ and tumor effects [33]. To verify the occurrence of long-
CD4+ T cells by flow cytometry and performed immu- lasting immune memory in this study, we established
nohistochemical staining to assess the potential immune a mouse lung metastasis model after removing the
responses induced by the developed nanoparticles. Com- primary tumor. On the 10th day of follow-up, severe
pared to the control group, CpG+OVA+PTX+CQ- lung metastasis was observed in the control group,
N/A significantly increased the levels of both cell types while Ki67 staining confirmed the rapid proliferation
(Fig.  8a, b), while immunohistochemistry on tumor of tumor cells in the lungs, which could then trigger
biopsies revealed more extensive brown areas in systemic spread and lead to death (Additional file  1:
CpG+OVA+PTX+CQ-N/A-treated mice compared Fig. S10a-b). Interestingly, CpG+OVA+PTX+CQ-
to the other treatment groups (Fig.  8c, d, Additional N/A partially inhibited lung metastasis, reduced lung
file  1: Fig. S7). Moreover, CpG+OVA+PTX+CQ-N/A weight (Additional file 1: Fig. S10c) and the number of
increased the number of CD8+ and CD4+ T cells in the lung metastases (Additional file 1: Fig. S10d), and pro-
spleen by 1.48 and 4.63 times, respectively, compared to longed survival time compared to control. These results
the control group (Fig. 8e–g), suggesting that CD8+ and strongly suggest that the described MNPs are able
CD4+ T cells are the main effector cells of antitumor to treat primary breast cancer as well as inhibit lung
response in our animal model. CD3+ and CD8+ T-cell metastasis.
stimulation was confirmed by immunohistochemistry in
the spleen (Additional file 1: Fig. S8, Fig. 8h). In contrast, Discussion
CpG+OVA-N showed reduced immune responses, sug- Cancer is a complex and adaptive ecosystem that can-
gesting that PTX and CQ can lyse tumor cells upon irra- not be easily treated with individual therapies. Although
diation, serving as tumor-associated antigens. combination therapy based on nanomaterials has recently
Our combination therapy not only activated immune attracted particular attention [34–36], the effective deliv-
cells such as T and natural killer cells, but it also pro- ery of nanomedicines with sufficient concentration in
moted the maturation of dendritic cells and regu- malignant cells depends strongly on their transport prop-
lated the secretion of cytokines. CpG+OVA+PTX-N, erties under hypoxia and avascular conditions [37]. There-
CpG+OVA+PTX+CQ-N, and CpG+OVA+PTX-N/A fore, recent studies on anticancer therapy have focused
upregulated the serum levels of tumor necrosis factor-α on the development of biomaterials for targeted drug
(TNF-α), which plays an important role in host defense delivery [38]. In this study, we used a simple, environmen-
and cellular immunity [31], as well as the levels of tally friendly method and a PEI-OA polymer complex to
interferon-γ (IFN-γ), a mediator of Th1 cells that regu- construct the first PD-L1- and CD44-responsive multi-
lates cell-mediated immune responses [32]. Among these functional nanoplatform that can be loaded simultane-
formulations, CpG+OVA+PTX+CQ-N/A nanoparticles ously with two chemotherapeutic drugs, an antigen, an
led to the greatest increase of TNF-α and IFN-γ (Fig. 8i, immunopotentiator, and an immune checkpoint inhibitor.
j), while they could also considerably reduce the number In vitro and in vivo studies of the developed MNPs indi-
of PD-L1+ cells compared to control (Additional file  1: cate that this combination treatment strategy can strongly
Fig. S9), suggesting that they exerted their therapeutic inhibit tumor growth and prevent relapse and lung metas-
effect by regulating cytokine expression and blocking tasis of advanced breast cancer.
PD-L1 in T lymphocytes. Recent preclinical studies at the interface of biomaterial
Taken together, our experiments suggest that science, drug delivery, and cancer vaccines have highlighted
autophagy inhibition in tumors combined with immu- the promising potential of combining different types of
noadjuvants and an anti-PD-L1 antibody can enhance cancer vaccines (e.g., DNA, mRNA, peptide/protein, or
 Cheng et al. BMC Medicine (2022) 20:411 Page 15 of 18

Fig. 8  CpG+OVA+PTX+CQ-N/A induce antitumor immune responses in vivo. a Levels of CD8+ and CD4+ T cells, as determined by flow
cytometry on day 15 after the indicated treatments. b Representative flow cytometric plots of CD8+ and CD4+ T cells in tumors. c, d
Immunohistochemistry on tumor biopsies. Brown regions indicate the presence of c CD8+ T cells and d CD4+ T cells. Scale bar, 100 μm. e, f
Levels of e CD8+ T cells and f CD4+ T cells in the spleen of tumor-bearing mice, as determined by flow cytometry on day 15 after the indicated
treatments. g Representative flow cytometric plots of CD8+ and CD4+ T cells in the spleen. h Immunohistochemical staining of the spleen. Brown
regions indicate the presence of CD8+ T cells. Scale bar, 100 μm. i, j Serum levels of i tumor necrosis factor-α (TNF-α) and j interferon-γ (IFN-γ) on
day 15 after the indicated treatments. Data are shown as ± SD (n = 5). One-way ANOVA was performed. *P < 0.05. CpG, immunopotentiator; CQ,
chloroquine; N, nanoparticles; N/A, nanoparticles coated with atezolizumab; OVA, ovalbumin; PTX, paclitaxel; S, solution

cell-based) with various delivery systems, such as nanopar- in a controlled manner, while tuning of their targeting moi-
ticles, microparticles, self-assembled materials, and bioma- eties and physical properties, such as size, shape, charge, or
terial scaffolds [39–41]. Biomaterial-based cancer vaccines porosity, can achieve selective delivery to target cells and
have also been found to be more effective than conven- tissues with desirable drug release kinetics [44].
tional vaccines [42, 43], as they can be delivered to the body
Cheng et al. BMC Medicine (2022) 20:411 Page 16 of 18

In the present work, we selected PEI-OA as the nano- nodes and spleens also indicated stronger CD8+ T cell
platform core due to its amphiphilicity, positive charge, responses.
and good biocompatibility [22]. PEI derivatives have also Our results also suggested that, in the early stages of
been shown to stimulate multiple damage-associated treatment, PTX lysed most of the tumor cells, which then
molecular pattern receptors and exert potent adjuvantic- acted as tumor-associated antigens capable of stimulating
ity, thus promoting dendritic cell activation and cross- a strong antigen-specific immune response. PTX medi-
presentation [45]. To avoid off-target toxicity and increase ated enhancement of antitumor immunity results, at least
the in  vivo safety of the described MNPs, the inner core in part, from the stimulation of dendritic cells by dead or
was further modified with atezolizumab and CS. Using a dying tumor cells induced by PTX. Recent advances have
model of mice bearing large breast tumors, we found that demonstrated that some chemotherapeutics including
two doses of MNPs loaded with OVA and CpG inhibited PTX could act as immunogenic cell death inducers by
35.72% of the established tumors. These results suggest arousing calreticulin exposure on tumor cell surface and
that our approach may be effective against large tumors, the release of high mobility group box 1 from tumor cells
which are generally not easily permeable to drugs or [54, 55]. In addition, concurrent inhibition of autophagy
delivery materials [46]. Although immunotherapy may with CQ prevented tumor relapse for a longer time
be a promising treatment strategy against breast cancer, than free chemotherapeutic drugs did. The described
most patients respond poorly due to T-cell anergy [47]. killing effect was achieved with only two injections of
Therefore, most recent studies have employed immune- CpG+OVA+PTX+CQ-N/A and resulted in the down-
checkpoint inhibitors, such as PD-1/PD-L1 antibodies, regulation of PD-L1 in T cells and the induction of CD8+
combined with other therapeutic treatments, to improve and CD4+ T cell generation, which are probably the
therapeutic effects [48]. Based on this, our MNPs were main effector cells for tumor regression and recurrence.
further loaded with an anti-PD-L1-antibody to block PD- The present study clearly shows that the developed
L1-mediated immune responses, offering a long-term MNPs not only accumulate selectively at the tumor site,
therapeutic effect in the immunosuppressive TME, which but can also penetrate deep into large tumors, where blood
is a less expensive and potentially safer approach than chi- vessels are sparse and oxygen levels are low. In addition,
meric antigen receptor T-cell immunotherapy [49]. our results on immune responses after combination ther-
TNF-α and IFN-γ are cytokines that play an active role apy provide important insights into the crosstalk between
in the anti-tumor immune responses. TNF-α, a mediator biomaterial-mediated therapy and cancer immunotherapy.
of cellular immunity, can kill tumor cells directly without
harming normal cells [50]. TNF-α is found in the TME
that is involved in all stages of breast cancer development, Conclusions
affecting tumor cell proliferation and survival, epithelial-to- We used PEI-OA to construct an autophagy-responsive
mesenchymal transition (EMT), metastasis, and recurrence nanoplatform that efficiently encapsulates two hydro-
[51]. IFN-γ is a critical effector molecule to tumor rejection phobic chemotherapeutic drugs for enhanced breast
[52]. IFN-γ can impede tumor growth by acting directly on cancer treatment. Co-loading of the anti-PD-L1 anti-
cancer cells; or induced regression of the tumor vascula- body provides MNPs with favorable tumor targeting
ture, resulting in arrest of blood flow and subsequent col- and permeating properties and enhances their ability to
lapse of tumors; or drives the fragility of surrounding Tregs, inhibit the growth of primary breast cancer, while CS
boosts antitumor immunity, and facilitates tumor clearance modification promotes their accumulation at the tumor
[53]. The high production of TNF-α and IFN-γ induced site. CpG+OVA+PTX+CQ-N/A showed high cyto-
by CpG+OVA+PTX+CQ-N/A (Fig.  4e) enlightened us toxicity against 4T1 cells and significantly improved the
that CpG+OVA+PTX+CQ-N/A could be a valuable vec- levels of CD8+ and CD4+ T cells at the tumor site. In
tor for antitumor vaccines. We next assessed the CD8+ addition, the combination therapy showed long-lasting
T cell response induced by CpG+OVA+PTX+CQ-N/A immune memory and effectively inhibited lung metas-
nanoparticles. Tumors and spleens of vaccinated mice were tasis. CpG+OVA+PTX+CQ-N/A show promise for
harvested on day 15 and homogenized into a cell suspen- breast cancer immunotherapy and may serve as a guide
sion; proportions of antigen-specific, CD4+ and CD8+ T for the development of novel drug delivery systems for
cells were determined. CpG+OVA+PTX+CQ-N/A nano- combined active targeting, autophagy inhibition, and
particles induced a stronger Th1 cell (CD4+) and cytotoxic chemotherapy, as well as for the co-delivery of biomacro-
CD8+ T cell response than other groups in tumors, higher molecules and small-molecule drugs.
proportion of CD8+ T cells in spleens (Fig.  8g, h). The
production of IFN-γ (Fig. 8i) and TNF-α (Fig. 8j) in lymph
 Cheng et al. BMC Medicine (2022) 20:411 Page 17 of 18

Abbreviations References
A: Atezolizumab; BCA: Bicinchoninic acid; BSA: Bovine serum albumin; CpG: 1. Kuang J, Song W, Yin J, Zeng X, Han S, Zhao YP, et al. iRGD modified
Immunopotentiator; CQ: Chloroquine; CS: Chondroitin sulfate; DAPI: 4′,6-Diami- chemo-immunotherapeutic nanoparticles for enhanced immunotherapy
dino-2-phenylindole; DC: Dendritic cell; DiD: 4-Chlorobenzenesulfonate salt; against glioblastoma. Adv Func Mater. 2018;28:1800025.

orescein; H&E: Hematoxylin and eosin; HS15: Solutol HS15®; IFN-γ: Interferon-γ;
DMEM: Dulbecco’s modified Eagle medium; FAM: Fluorescein amidite; FITC: Flu- 2. Du H, Zhao S, Wang Y, Wang Z, Chen B, Yan Y, et al. pH/Cathepsin B
hierarchical-responsive nanoconjugates for enhanced tumor penetration
MHC: Major histocompatibility complex; MNP: Multifunctional nanoparticles; and chemo-immunotherapy. Adv Funct Mater. 2020;30:2003757.
N: Nanoparticles; N/A: Nanoparticles coated with atezolizumab; OA: Oleic acid; 3. Wang H, Liang Y, Yin Y, Zhang J, Su W, White AM, et al. Carbon nano-
OVA: Ovalbumin; PEI: Polyethyleneimine; PEI-OA: Polyethyleneimine-oleic acid onion-mediated dual targeting of P-selectin and P-glycoprotein to
polymer; PI: Propidium iodide; PTX: Paclitaxel; S: Solution; SD: Standard devia- overcome cancer drug resistance. Nat Commun. 2021;12:312.
tion; TME: Tumor microenvironment; TNF-α: Tumor necrosis factor-α; TUNEL: 4. Song C, Phuengkham H, Kim YS, Dinh VV, Lee I, Shin IW, et al. Syringeable
Terminal deoxynucleotidyl transferase dUTP nick end-labeling. immunotherapeutic nanogel reshapes tumor microenvironment and
prevents tumor metastasis and recurrence. Nat Commun. 2019;10:3745.
5. Deng C, Zhang Q, Jia M, Zhao J, Sun X, Gong T, et al. Tumors and their
Supplementary Information microenvironment dual-targeting chemotherapy with local immune
The online version contains supplementary material available at https://​doi.​ adjuvant therapy for effective antitumor immunity against breast cancer.
org/​10.​1186/​s12916-​022-​02614-8. Adv Sci. 2019;6:1801868.
6. Hu F, Song D, Yan Y, Huang C, Shen C, Lan J, et al. IL-6 regulates autophagy
Additional file 1: Figure S1. Photographs of tumors. Figure S2. and chemotherapy resistance by promoting BECN1 phosphorylation. Nat
Hematoxylin and eosin staining of sectioned heart, liver, spleen, lung, Commun. 2021;12:3651.
and kidney tissues. Figure S3. Hemolysis assay. Figure S4. Routine blood 7. Feng J, Zhang Y, She X, Sun Y, Fan L, Ren X, et al. Hypermethylated gene
examination. Figure S5. Apoptosis assay in liver and spleen by western ANKDD1A is a candidate tumor suppressor that interacts with FIH1 and
blot. Figure S6. Autophagy assay in liver and spleen by western blot. decreases HIF1α stability to inhibit cell autophagy in the glioblastoma
Figure S7. CD3+ T cells in mouse tumors by immunohistochemical stain- multiforme hypoxia microenvironment. Oncogene. 2019;38:103–19.
ing. Figure S8. CD3+ T cells in mouse spleen by immunohistochemical 8. Haq S, Wang H, Grondin J, Banskota S, Marshall JK, Khan II, et al. Disrup-
staining. Figure S9. PD-L1 immunohistochemistry. Figure S10. Immune- tion of autophagy by increased 5-HT alters gut microbiota and enhances
memory effect and metastasis inhibition. susceptibility to experimental colitis and Crohn’s disease. Sci Adv.
2021;7:eabi6442.
9. Young TM, Reyes C, Pasnikowski E, Castanaro C, Wong C, Decker CE, et al.
Acknowledgements Autophagy protects tumors from T cell–mediated cytotoxicity via inhibi-
Not applicable. tion of TNFα-induced apoptosis. Sci Immunol. 2020;5:eabb9561.
10. Qiao Y, Choi JE, Tien JC, Simko SA, Rajendiran T, Vo JN, et al. Autophagy
Authors’ contributions inhibition by targeting PIKfyve potentiates response to immune check-
JL and YC designed and performed the experiments, analyzed the data, and point blockade in prostate cancer. Nat Cancer. 2021;2:978–93.
wrote the manuscript. CW, HW, ZZ, XY, and YD performed the experiments and 11. Xia H, Green DR, Zou W. Autophagy in tumour immunity and therapy. Nat
edited the manuscript. LM designed the experiments, analyzed the data, and Rev Cancer. 2021;21:281–97.
edited the manuscript. All authors read and approved the final manuscript. 12. Chen Y, Zhao H, Liang W, Jiang E, Zhou X, Shao Z, et al. Autophagy
regulates the cancer stem cell phenotype of head and neck squamous
Funding cell carcinoma through the noncanonical FOXO3/SOX2 axis. Oncogene.
The authors acknowledge financial support from the National Natural Science 2022;41:634–46.
Foundation of China (82003688 and 82101687) and the China National Key 13. Li ZL, Zhang HL, Huang Y, Huang JH, Sun P, Zhou NN, et al. Autophagy
Research and Development (R&D) Program (2021YFC2100100). deficiency promotes triple-negative breast cancer resistance to T cell-
mediated cytotoxicity by blocking tenascin-C degradation. Nat Commun.
Availability of data and materials 2020;11:3806.
The data supporting the findings of this study are available from the cor- 14. Poillet PL, Sharp DW, Yang Y, Laddha SV, Ibrahim M, Bommareddy PK, et al.
responding author upon reasonable request. Autophagy promotes growth of tumors with high mutational burden by
inhibiting a T-cell immune response. Nat Cancer. 2020;1:923–34.
15. Xu X, Araki K, Li S, Han JH, Ye L, Tan WG, et al. Autophagy is essential for
Declarations effector CD8+ T cell survival and memory formation. Nat Immunol.
2014;15:1152–61.
Ethics approval and consent to participate 16. Amaravadi RK, Kimmelman AC, Debnath J. Targeting autophagy
All animal experiments were approved by the Ethics Committee of Wuhan in cancer: recent advances and future directions. Cancer Discov.
Institute of Virology (Approval number: WIVA04202102). 2019;9:1167–81.
17. Ligeon LA, Pena FM, Vanoaica LD, Núñez NG, Talwar D, Dick TP, et al.
Consent for publication Oxidation inhibits autophagy protein deconjugation from phagosomes
Not applicable. to sustain MHC class II restricted antigen presentation. Nat Commun.
2021;12:1508.
Competing interests 18. Zhang Y, Sha R, Zhang L, Zhang W, Jin P, Xu W, et al. Harnessing copper-
The authors declare that they have no competing interests. palladium alloy tetrapod nanoparticle-induced pro-survival autophagy
for optimized photothermal therapy of drug-resistant cancer. Nat Com-
Author details mun. 2018;9:4236.
1
 State Key Laboratory of Biocatalysis and Enzyme Engineering, Hubei Key 19. Maycotte P, Aryal S, Cummings CT, Thorburn J, Morgan MJ, Thorburn A.
Laboratory of Industrial Biotechnology, Hubei Collaborative Innovation Center Chloroquine sensitizes breast cancer cells to chemotherapy independent
for Green Transformation of Bio‑Resources, School of Life Sciences, Hubei Uni- of autophagy. Autophagy. 2012;8:200–12.
versity, Wuhan 430062, P. R. China. 2 Key Laboratory for the Synthesis and Appli- 20. Paola P, Angela S, Chiara Z, et al. Acidic extracellular pH neutral-
cation of Organic Functional Molecules, College of Chemistry and Chemical izes the autophagy-inhibiting activity of chloroquine. Autophagy.
Engineering, Hubei University, Wuhan 430062, P. R. China. 2014;10:562–71.
21. Yamamoto K, Venida A, Yano J, Biancur DE, Kakiuchi M, Gupta S, et al.
Received: 30 May 2022 Accepted: 17 October 2022 Autophagy promotes immune evasion of pancreatic cancer by degrad-
ing MHC-I. Nature. 2020;581:100–5.
Cheng et al. BMC Medicine (2022) 20:411 Page 18 of 18

22. Luo J, Zhang Z, Zeng Y, Dong Y, Ma L. Co-encapsulation of collagenase delivery: synthesis, complexation, and transfection. Adv Funct Mater.
type I and silibinin in chondroitin sulfate coated multilayered nano- 2009;19:2244–51.
particles for targeted treatment of liver fibrosis. Carbohydr Polym. 46. Raaijmakers MIG, Rozati S, Goldinger SM, Widmer DS, Dummer R,
2021;263:117964. Levesque MP. Melanoma immunotherapy: historical precedents, recent
23. Dobrovolskaia MA, Afonin KA. Use of human peripheral blood mononu- successes and future prospects. Immunotherapy. 2013;5:169–82.
clear cells to define immunological properties of nucleic acid nanoparti- 47. Pulluri B, Kumar A, Shaheen M, Jeter J, Sundararajan S. Tumor microenvi-
cles. Nat Protoc. 2020;15:3678–98. ronment changes leading to resistance of immune checkpoint inhibitors
24. Zhang J, Li Y, Liu Q, Huang Y, Li R, Wu T, et al. Sirt6 alleviated liver fibrosis in metastatic melanoma and strategies to overcome resistance. Pharma-
by deacetylating conserved lysine 54 on Smad2 in hepatic stellate cells. col Res. 2017;123:95–102.
Hepatology. 2021;73:1140–57. 48. Zitvogel L, Kroemer G. Targeting PD-1/PD-L1 interactions for cancer
25. Luo J, Zhang P, Zhao T, Jia M, Yin P, Li W, et al. Golgi apparatus-targeted immunotherapy. Oncoimmunology. 2012;1:1223–5.
chondroitin-modified nanomicelles suppress hepatic stellate cell activa- 49. Kataoka N, Kunimatsu Y, Tachibana Y, Sugimoto T, Sato I, Tani N, et al.
tion for the management of liver fibrosis. ACS Nano. 2019;13:3910–23. Atezolizumab in combination with carboplatin and etoposide for heavily
26. Chen C, Guo Q, Fu H, Yu J, Wang L, Sun Y, et al. Asynchronous blockade treated small cell lung cancer. Thorac Cancer. 2020;11:2740–2.
of PD-L1 and CD155 by polymeric nanoparticles inhibits triple-negative 50. Cruceriu D, Baldasici O, Balacescu O, Berindan-Neagoe I. The dual role
breast cancer progression and metastasis. Biomaterials. 2021;275:120988. of tumor necrosis factor-alpha (TNF-alpha) in breast cancer: molecular
27. Zhang Q, Yang W, Man N, Zheng F, Shen Y, Sun K, et al. Autophagy-medi- insights and therapeutic approaches. Cell Oncol. 2020;43(1):1–18.
ated chemosensitization in cancer cells by fullerene C60 nanocrystal. 51. Ma Y, Ren Y, Dai ZJ, Wu CJ, Ji YH, Xu J. IL-6, IL-8 and TNF-α levels cor-
Autophagy. 2009;5:1107–17. relate with disease stage in breast cancer patients. Adv Clin Exp Med.
28. Cheng Y, Lai F, Wang X, Shang D, Zou J, Luo M, et al. Srag regulates 2017;26(3):421–6.
autophagy via integrating into a preexisting autophagy pathway in testis. 52. Gao Y, Yang J, Cai Y, Fu S, Zhang N, Fu X, et al. IFN-gamma-mediated inhi-
Mol Biol Evol. 2021;38:128–41. bition of lung cancer correlates with PD-L1 expression and is regulated
29. Shi X, Yang X, Liu M, Wang R, Qiu N, Liu Y, et al. Chondroitin sulfate-based by PI3K-AKT signaling. Int J Cancer. 2018;143(4):931–43.
nanoparticles for enhanced chemo-photodynamic therapy overcoming 53. Li Q, Zhou Y, He W, Ren X, Zhang M, Jiang Y, et al. Platelet-armored nano-
multidrug resistance and lung metastasis of breast cancer. Carbohydr platform to harmonize janus-faced IFN-gamma against tumor recurrence
Polym. 2021;254:117459. and metastasis. J Control Release. 2021;338:33–45.
30. Janjigian YY, Kawazoe A, Yañez P, Li N, Lonardi S, Kolesnik O, et al. The KEY- 54. Duan X, Chan C, Lin W. Nanoparticle-mediated immunogenic cell death
NOTE-811 trial of dual PD-1 and HER2 blockade in HER2-positive gastric enables and potentiates cancer immunotherapy. Angew Chem Int Ed
cancer. Nature. 2021;600:727–30. Engl. 2019;58(3):670–80.
31. Dumitru CD, Ceci JD, Tsatsanis C, Kontoyiannis D, Stamatakis K, Lin JH, 55. Fucikova J, Kepp O. Detection of immunogenic cell death and its rel-
et al. TNF-α induction by LPS is regulated posttranscriptionally via a Tpl2/ evance for cancer therapy. Cell Death Dis. 2020;11(11):1013.
ERK-dependent pathway. Cell. 2000;103:1071–83.
32. Fuchs A, Vermi W, Lee J, Lonardi S, Gilfillan S, Newberry R, et al. Intraepi-
thelial type 1 innate lymphoid cells are a unique subset of IL-12- and Publisher’s Note
IL-15-responsive IFN-γ-producing cells. Immunity. 2013;38:769–81. Springer Nature remains neutral with regard to jurisdictional claims in pub-
33. Kaech SM, Tan JT, Wherry EJ, Konieczny BT, Surh CD, Ahmed R. Selective lished maps and institutional affiliations.
expression of the interleukin 7 receptor identifies effector CD8 T cells that
give rise to long-lived memory cells. Nat Immunol. 2003;4:1191–8.
34. Uson L, Yus C, Mendoza G, Leroy E, Irusta S, Alejo T, et al. Nanoengineer-
ing palladium plasmonic nanosheets inside polymer nanospheres for
photothermal therapy and targeted drug delivery. Adv Funct Mater.
2021;32:2106932.
35. Fang H, Guo Z, Chen J, Lin L, Hu Y, Li Y, et al. Combination of epigenetic
regulation with gene therapy-mediated immune checkpoint blockade
induces anti-tumour effects and immune response in vivo. Nat Commun.
2021;12:6742.
36. Bosc C, Saland E, Bousard A, Gadaud N, Sabatier M, Cognet G, et al. Mito-
chondrial inhibitors circumvent adaptive resistance to venetoclax and
cytarabine combination therapy in acute myeloid leukemia. Nat Cancer.
2021;2:1204–23.
37. Zhang Q, Zhang J, Song J, Liu Y, Ren X, Zhao Y. Protein-based nanomedi-
cine for therapeutic benefits of cancer. ACS Nano. 2021;15:8001–38.
38. Lu M, Huang Y. Bioinspired exosome-like therapeutics and delivery nano-
platforms. Biomaterials. 2020;242:119925.
39. Kalelkar PP, Riddick M, García AJ. Biomaterial-based antimicrobial thera-
pies for the treatment of bacterial infections. Nat Rev Mater. 2021;7:39–54.
40. Balu R, Dutta NK, Dutta AK, Choudhury NR. Resilin-mimetics as a
Ready to submit your research ? Choose BMC and benefit from:
smart biomaterial platform for biomedical applications. Nat Commun.
2021;12:149.
• fast, convenient online submission
41. Wang H, Mooney DJ. Biomaterial-assisted targeted modulation of
immune cells in cancer treatment. Nat Mater. 2018;17:761–72. • thorough peer review by experienced researchers in your field
42. Wang H, Najibi AJ, Sobral MC, Seo BR, Lee JY, Wu D, et al. Biomaterial- • rapid publication on acceptance
based scaffold for in situ chemo-immunotherapy to treat poorly immu-
• support for research data, including large and complex data types
nogenic tumors. Nat Commun. 2020;11:5696.
43. Dolgin E. Cancer vaccines: Material breach. Nature. 2013;504:S16–7. • gold Open Access which fosters wider collaboration and increased citations
44. Roth GA, Picece VCTM, Ou BS, Luo W, Pulendran B, Appel EA. Design- • maximum visibility for your research: over 100M website views per year
ing spatial and temporal control of vaccine responses. Nat Rev Mater.
2021;7:174–95. At BMC, research is always in progress.
45. Kievit FM, Veiseh O, Bhattarai N, Fang C, Gunn JW, Lee D, et al. PEI–PEG–
chitosan-copolymer-coated iron oxide nanoparticles for safe gene Learn more biomedcentral.com/submissions