Cholesterol Modulates The Physiological Response To Nanoparticles by Changing The Composition of Protein Corona

nature nanotechnology
Article https://doi.org/10.1038/s41565-023-01455-7
Cholesterol modulates the physiological

response to nanoparticles by changing the
composition of protein corona
Received: 15 January 2023 Huan Tang 1,7, Ying Zhang1,7, Tong Yang1, Chen Wang1, Yinhua Zhu1,
Liangjia Qiu1, Jiahui Liu2, Yang Song1, Lirue Zhou1, Junzhe Zhang1,
Accepted: 12 June 2023
Yin Kwan Wong 3, Yuanfang Liu2,4, Chengchao Xu 1,5 , Haifang Wang4
Published online: 3 August 2023 & Jigang Wang 1,5,6
Check for updates

Nanoparticles (NPs) in biological fluids form a layer of biomolecules known
as the protein corona. The protein corona has been shown to determine the
biological identity and in vivo fate of NPs, but whether and how metabolites,
especially disease-related small molecules, regulate the protein corona and
subsequently impact NP fate in vivo is relatively poorly understood. Here we
report on the effects of cholesterol on the generation of protein corona and
subsequent effects. We find that high levels of cholesterol, as in hyperchol
esterolemia, result in a protein corona with enriched apolipoproteins and
reduced complement proteins by altering the binding affinity of the proteins
to the NPs. The cholesterol-mediated protein corona can induce stronger
inflammatory responses to NPs in macrophages and promote the cellular
uptake of NPs in hepatocytes by enhancing the recognition of lipoprotein
receptors when compared with normal protein corona. The result of in vivo
biodistribution assays shows that, compared with healthy mice, NPs in
hypercholesterolemic mice were more likely to be delivered to the liver,
spleen and brain, and less likely to be delivered to the lungs. Our findings
reveal that the metabolome profile is an unexploited factor impacting the
target efficacy and safety of nanomedicines, providing a way to develop
personalized nanomedicines by harnessing disease-related metabolites.
Despite the potentially wide-ranging applications of nanoparticles performances in clinical trials5,6. This huge gap is in part due to a lack
(NPs) in bioimaging, drug delivery, vaccine and cancer immuno- of understanding of the nano–bio interfaces in patients. When admin-
therapy1–4, the number of nanomedicines that have been approved in istered to the human body, nanomedicines inevitably interact with
the market has been fewer than expected, due to their less-than-ideal biomolecules in bodily fluids, forming a protein crown termed ‘protein
State Key Laboratory for Quality Ensurance and Sustainable Use of Dao-di Herbs, Artemisinin Research Center, and Institute of Chinese Materia Medica,
1
China Academy of Chinese Medical Sciences, Beijing, China. 2Beijing National Laboratory for Molecular Sciences, College of Chemistry and Molecular
Engineering, Peking University, Beijing, China. 3Department of Biological Sciences, National University of Singapore, Singapore, Singapore. 4Institute
of Nanochemistry and Nanobiology, Shanghai University, Shanghai, China. 5Department of Nephrology, Shenzhen Key Laboratory of Kidney Diseases,
and Shenzhen Clinical Research Centre for Geriatrics, Shenzhen People’s Hospital, The First Affiliated Hospital, Southern University of Science and
Technology, Shenzhen, China. 6Guangdong Provincial Key Laboratory of New Drug Screening, School of Pharmaceutical Sciences, and School of
Traditional Chinese Medicine, Southern Medical University, Guangzhou, China. 7These authors contributed equally: Huan Tang, Ying Zhang.
e-mail: [email protected]; [email protected]
Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1067

corona’ (PC) on their surface. There is increasing evidence showing that normal FBS (Nor-PC) or FBS containing a high concentration (15.3 mM)
the PC has an important impact on the biodistribution, bioavailability of cholesterol (Cho-PC) by transmission electron microscopy (TEM)
and biotransformation of nanomedicines in vivo7,8. Thus, understand- (Fig. 1f) and cryo-TEM (Extended Data Fig. 2a). It was apparent that
ing the formation and biological roles of the PC is a vital step towards cholesterol indeed reduced the corona thickness of NPs. This was also
the safe and efficacious clinical application of nanomedicines. supported by the calculated TEM size and hydrodynamic diameter of
The formation of the PC can be affected by many factors, such as naked NPs, NPs-Nor-PC and NPs-Cho-PC for AuNPs and SNPs (Extended
the physicochemical properties of NPs and protein species in biological Data Fig. 2b–d).
fluids8–10. The proteomic fingerprints of the PC provide clues for the Because more cholesterol absorbed on SNPs with increasing cho-
early diagnosis, progression and prognosis of diseases8,11. Previous lesterol concentration (Extended Data Fig. 2e), we next investigated
studies have revealed that human plasma samples obtained from dif- cholesterol-affected PC formation through changed cholesterol–
ferent donors could differentially modulate the compositions of the protein or cholesterol–NP interactions. To this end, we conducted a
PC, thereby affecting the physiological response to NPs12–14. Thus, the competition experiment (Extended Data Fig. 3a). On the one hand,
concept of a personalized PC, which refers to PCs custom-generated in pretreating NPs with cholesterol obviously reduced the amounts of the
patients or healthy individuals, has been proposed for the development proteins adsorbed on the surface of NPs; on the other hand, the inhibi-
of precision nanomedicines12,15–17. However, the key factors that dictate tory effect of a high concentration of cholesterol on PC formation was
PC formation and subsequent in vivo behaviours of nanomedicines partially reversed by pre-occupying the NPs with proteins in normal FBS
remain poorly understood. It is well known that metabolites, especially (Fig. 1g,h). Thus, both cholesterol–protein and cholesterol–NP interac-
lipids, adsorb on the surface of NPs18–23, but the biological significance tions can affect the adsorption of proteins on the NPs in the nano–bio
of such interactions on PC formation is not clear18–24. Here we explore interface. Taken together, our results showed that cholesterol in serum
how metabolites drive the formation of the PC and influence the in vivo is a critical factor for personalized PC formation.
fates of nanomedicines, especially in patients with abnormal levels of
disease-related serum metabolites. Proteomic fingerprints of PC modulated by
cholesterol
Impact of cholesterol on PC formation To obtain information about the protein fingerprints of Nor-PC and
As an essential metabolite, cholesterol plays vital roles in membrane Cho-PC formed on SNPs in FBS, proteomic analysis was performed
organization and signal transduction. Dysregulation in cholesterol (Extended Data Fig. 3b). No significant differences were observed
metabolism is related to various human diseases, such as cardiovascular regarding the physiological function, isoelectric point and molecular
disease, neurodegeneration and cancers25. Adults with a total serum weight of the identified proteins (Extended Data Fig. 4a–c). Most of
cholesterol content exceeding ~5.1 mM are at risk of hypercholester- the identified proteins were shared by the two groups of Nor-PC and
olemia26. A previous study found more cholesterol associated with Cho-PC, indicating that cholesterol in serum modulates the affinity of
Fe3O4 NPs after incubation in hyperlipidemic serum than in normal NP-binding proteins (Fig. 2a). Among those 145 overlapping proteins, 26
serum14. In a study of biomolecular coronas of high-density lipoproteins proteins were upregulated and 79 proteins were downregulated in the
on gold nanoparticles (AuNPs), cholesterol was found to be enriched Cho-PC (Fig. 2b). These differentially attached proteins were involved
in the coronas by 20- to 100-fold, much more than that of other lipids in pathways of complements and cholesterol metabolism (Extended
and apolipoprotein A1 (ref. 27). Some NPs are designed for drug target Data Fig. 4d–g). By comparing the 25 most abundant proteins of the two
delivery by leveraging the cholesterol-rich microenvironment28,29. groups (Extended Data Fig. 4h), we found notably more apolipoproteins
Cholesterol is also used to construct a DNA nanostructure, which could and fewer complement proteins in Cho-PC (Fig. 2c,d). Interestingly,
enrich lipoproteins in serum and form a lipoprotein-associated PC30. cholesterol increased the retention of all identified apolipoproteins
We wondered about the impact of cholesterol on the formation of PCs on SNPs (Extended Data Fig. 4i) while lowering the abundance of most
on NPs, which might cause personalized effects of nanomedicines in complement proteins on SNPs except complement component 9 (C9;
hypercholesterolemic patients. To this end, we selected two promising Extended Data Fig. 4j). Similar changes were also observed for the
nanomaterials—silica nanoparticles (SNPs) and AuNPs, which have wide liposomes (Supplementary Fig. 2). Pre-coating SNPs with cholesterol
applications in nanomedicine31,32. We first tested the roles of cholesterol also enriched more apolipoproteins and fewer complement proteins
in PC formation using a defined system (Fig. 1a). NPs were incubated in PC than bare SNPs (Extended Data Fig. 5a). We performed similar
with cholesterol-supplemented fetal bovine serum (FBS) or normal analyses with three human serums with different cholesterol concentra-
FBS, and PCs formed on NPs were then collected for analyses. Both tions and found similar trends (Extended Data Fig. 5b–e). It should be
the electrophoresis and protein quantification results showed that noted that the human serum is a much more complex system such that
excessive cholesterol reduced the PC content in a dose-dependent the dynamic changes of apolipoproteins and complement proteins in
manner (Fig. 1b,c and Extended Data Fig. 1a–d). Similar observations the serum samples might affect PC formation (Extended Data Fig. 5c,e).
were also made using liposomes (Supplementary Fig. 1). To confirm Taken together, our results suggest that the high level of cholesterol in
our results in a more clinically relevant setting, three human serum serum could enrich apolipoproteins and reduce complement proteins
samples from volunteers with different concentrations of cholesterol in the PC on NPs.
(4.4 mM, 5.6 mM and 7.2 mM) were collected for analysis. Reassuringly, PC formation follows the ‘Vroman effect’; in other words, the
we found similar trends (Fig. 1d,e and Extended Data Fig. 1e). We next earlier adsorbed proteins of high abundance on the surface would
directly visualized naked NPs and NPs coated with a PC originating from be competitively displaced by proteins with higher affinity33. We next
Fig. 1 | Dissecting cholesterol-mediated PC formation. a, Schematic diagram of NPs-Cho-PC. g, Coomassie-stained SDS–PAGE gel of PC-I, PC-II, PC-III and PC-IV.
dissecting cholesterol-mediated PC formation. b, Coomassie-stained SDS–PAGE PC-I, PC formed in FBS; PC-II, PC formed in FBS after SNPs were pretreated with
gel of PC on AuNPs under different concentrations of cholesterol. MW, molecular cholesterol; PC-III, PC formed in cholesterol-supplemented FBS after SNPs were
weight. c, The protein quantification of b. The data are presented as mean ± s.d. pre-incubated with normal FBS; PC-IV, PC formed in cholesterol-supplemented
with n = 3 (biologically independent experiments). Statistical significance FBS. h, Densitometric analysis of g. Squares, protein bands with similar changes
was tested with a two-tailed, unpaired Student’s t-test. d, Coomassie-stained in PC-II and PC-IV; circles, protein bands with similar changes in PC-III and PC-IV;
SDS–PAGE gel of PC on SNPs after incubating with human serum samples. The triangles, protein bands with medium intensities compared with those in PC-II
PC formed in the hypercholesterolemic serum (7.2 mM) had fewer proteins. and PC-III.
e, Densitometric analysis of d. f, TEM images of the bare NPs, NPs-Nor-PC and

examined how the affinity of apolipoproteins and complement pro- examples. Consistent with the proteomics results, our immunoblotting
teins to SNPs was modulated by cholesterol. Apolipoprotein E (APOE) analysis confirmed that with increased cholesterol concentration, the
and complement component 1 (C1) were selected as representative amount of APOE binding to SNPs increased, while that of C1q decreased
SDS–PAGE
Serum + cholesterol Cholesterol-mediated protein corona Eluted protein corona
Intensity
m/z
Serum Normal protein corona Eluted protein corona LC–MS/MS
b c d e
M
M
M
M
M
M
m
m
m
1m
m
m
.3
9
MW (kDa) 320
4
6
1.7
2
15
0.
MW (kDa)
5.
4.
5.
7.
P = 0.0003 4.4 mM 7.2 mM
180 300 P = 0.0009 180 5.6 mM
Normalized grey value

130
130
Proteins/AuNPs (µg mg–1)
P = 0.0053 240
100
100
225
70
70
160
55
150 55
40 80
40
75
35 35
0 0
25 25
0.9 1.7 5.1 15.3 25 35 40 55 70 100 130 180
Cholesterol level (mM) MW (kDa)
f g MW
-IV
I
-II
-II
-I
PC
PC
PC
PC
(kDa)
180
AuNPs AuNPs-Nor-PC AuNPs-Cho-PC 130
100
70
55
40
35
25
50 nm 50 nm
50 nm h
PC-I PC-II PC-III PC-IV
250
SNPs SNPs-Nor-PC SNPs-Cho-PC
200
Normalized grey value
150
100
50
50 nm 50 nm 50 nm
0
25 35 40 55 70 100 130 180
MW (kDa)

a b e
C1QA
APOA1 Cholesterol (mM) 0 1.7 5.1
APOA2 C1QB 28 kDa

6 C1QB
C6 APOD APOC3 APOE 38 kDa
f 1.15
C2
Downregulated Without cholesterol
4
–log10P
35 145 16 APOE Not significant 0.95 With cholesterol
Fraction bound
C4
Upregualted 0.75
2 0.55
Nor-PC Cho-PC 0.35
0.15
0
–0.05
−5 0 5
0
3
–9
–8
–7
–1
–1
–1
–1
Fold change (log2(Cho-PC/Nor-PC))
10
10
10
10
10
10
10
×
×
×
×
×
×
2
2
2
2
2
2
2
SNP concentration (M)
c d C9
g
APOC4
20 10 1.15
APOA2 C7
APOA4 C6 Without cholesterol
Protein percentage
Protein percentage
0.95
Fraction bound
APOA1 8 C4 With cholesterol
15 C3
APON 0.75
in PC (%)
in PC (%)
APOC2 6 C2
10 APOD C1S 0.55
APOC3 C1QB
4
APOE C1QA 0.35
5
2 0.15
0 0 –0.05
0.9 1.7 5.1 15.3 0.9 1.7 5.1 15.3
–9
–8
–7
–6
–1
–1
–1
10
10
10
10
10
10
10
Cholesterol level (mM) Cholesterol level (mM)
1×
1×
1×
1×
1×
1×
1×
SNP concentration (M)
h
100 ns
h
Gln 55
Glu 80 Arg 25 Lyn 69 Glu 49
100 ns
Gln 58
Gln 46 Glu 50
Glu 66
Glu 59 Glu 49
i j k
15 0 15
Hydrogen bonds
Contact surface area (nm2)
With cholesterol
Binding energy (kJ mol–1)
Salt bridges
Without cholesterol 12
12 –400
Average number
9 –800 9
6 –1,200 6
3 –1,600 3
With cholesterol
Without cholesterol
0 –2,000 0
0 20 40 60 80 100 0 20 40 60 80 100 Without With
Time (ns) Time (ns) cholesterol cholesterol

Fig. 2 | Proteomic fingerprints of PC modulated by cholesterol. a, Venn affinity of APOE (f) and C1QB (g) to SNPs in the presence or absence of cholesterol
diagram of overlapped proteins in Nor-PC and Cho-PC on SNPs. b, Volcano plot of was measured by MST. h, Molecular dynamics simulation analyses of the
the protein fold change of (log2(Cho-PC/Nor-PC)) plotted against the statistical interactions between APOE and SNPs in the absence (top) or presence (bottom)
significance (−log10P). Upregulated proteins are shown in red and downregulated of cholesterol. Typical snapshots of the initial (left panels) and final (middle and
proteins are shown in green. The horizontal dashed line represents P = 0.05. right panels) structures of APOE adsorbed on NPs from the side (middle panels)
The vertical left dashed line represents fold change = 0.5, and vertical right and top (right panels) views. i–k, The contact surface area (i), binding energy (j)
dashed line represents fold change = 2. c,d, The percentages of the identified and the average number of hydrogen bonds and salt bridges (k) between APOE
apolipoproteins (c) and complement proteins (d) in Nor-PC and Cho-PC under and SNPs with or without cholesterol. The data are presented as mean ± s.d. with
different concentrations of cholesterol. e, Western blotting of C1QB and APOE n = 5,000 (independent binding conformations within 100 ns simulation).
adsorbed on SNPs under different concentrations of cholesterol. f,g, The binding
(Fig. 2e). We also measured the binding affinity of these two proteins to immune response (Fig. 3e,f), the increased immune response to NPs
SNPs with or without cholesterol by microscale thermophoresis (MST). was mainly influenced by cholesterol-mediated changes in PCs.
In the presence of cholesterol, the affinity of APOE to SNPs increased We then performed proteomic analysis to identify the differ-
by nearly 20-fold (8.69 × 10−11 M versus 1.64 × 10−9 M) (Fig. 2f), while entially expressed proteins in macrophages after incubation with
the affinity of C1q to SNPs decreased by about 9-fold (5.56 × 10−8 M SNPs-Cho-PC and SNPs-Nor-PC. There were 788 downregulated and
versus 6.02 × 10−9 M) (Fig. 2g). Consistently, high cholesterol levels also 785 upregulated proteins identified in the SNPs-Nor-PC treatment
promoted the binding of low-density lipoprotein (LDL) and group versus the control group (Extended Data Fig. 7c). The changes
high-density lipoprotein (HDL) to SNPs (Extended Data Fig. 5f,g and were more dramatic in the SNPs-Cho-PC treatment group with 1,483
Supplementary Fig. 3). Thus, we concluded that cholesterol could downregulated proteins and 1,213 upregulated proteins (Extended Data
remodel the protein abundance in the PC by differentially regulating Fig. 7d,e). As expected, the upregulated proteins in the Cho-PC group
the affinity of binding proteins. were enriched in the inflammatory response-related pathways (Fig. 3g).
We next performed molecular dynamics simulations to explore Indeed, we found higher expression levels of inflammation-related
the interactions between APOE and SNPs in the presence and absence proteins or cytokines in the Cho-PC group (Fig. 3h). For the downregu-
of cholesterol. When cholesterol was added into the system, it could lated proteins, we did not find obvious differences in Gene Ontology
be clearly observed that APOE could associate with several choles- analyses (Extended Data Fig. 7f). Even though both Nor-PC and Cho-PC
terol molecules accompanied by a conformational change, thereby activated the immune response of macrophages, we found that Cho-PC
affecting the interactions between APOE and SNPs (Fig. 2h). The triggered a stronger response. We also calculated the gene set varia-
cholesterol-mediated enhancement of the adsorption ability of APOE tion analysis (GSVA) scores of the inflammatory-related pathways and
on SNPs was supported by changes in the averaged values of the contact found consistent changes (Fig. 3i). These results collectively suggest
surface area (Fig. 2i), binding energy (Fig. 2j), and hydrogen bonds that Cho-PC might increase the risk of immune system activation in
and salt bridges (Fig. 2k) between APOE and SNPs. It was likely that response to NPs to a greater extent than Nor-PC.
cholesterol increased the affinity of APOE to SNPs by changing the
protein structural stability and interacting forces (Supplementary Cho-PC enhanced the cellular uptake of NPs in
Fig. 4). Indeed, circular dichroism analysis revealed the conforma- hepatocytes
tional changes of C1Q, APOE, LDL and HDL proteins in the absence and Because cellular uptake is critical for pharmacokinetics and targeting
presence of cholesterol or SNPs (Extended Data Fig. 6). These results efficiency of nanomedicines and drugs are typically metabolized by
support the conclusion that cholesterol–protein and cholesterol–NP liver cells, we next wondered whether Cho-PC might affect the cel-
interactions help to remodel the Cho-PC composition. lular internalization of NPs in hepatocytes. Using HepG2 cells as an
in vitro model of human hepatocytes, we first evaluated the cytotoxicity
Cho-PC increased the immune response of of AuNPs (Extended Data Fig. 8a) and SNPs (Extended Data Fig. 8b).
macrophages to NPs No significant difference was observed among NPs, NPs-Nor-PC and
As PCs have been widely recognized as biological identities of NPs8,34, NPs-Cho-PC within the concentration range tested. We next analysed
we next interrogated the impact of Cho-PC on the in vitro and in vivo the uptake of NPs by flow cytometry (Extended Data Fig. 8c). In the
behaviours of NPs. No difference in cytotoxicity was observed among serum-free medium, the cellular uptake of SNPs-Cho-PC was higher
NPs, NPs-Nor-PC and NPs-Cho-PC within the testing concentrations than that of SNPs-Nor-PC during the 24 h incubation. Because the pres-
after 24 h incubation with macrophages (RAW264.7) (Extended Data ence of serum generally inhibited the uptake of NPs, we performed
Fig. 7a,b). We evaluated the immune response of the treated cells by the subsequent analysis under serum-free conditions. PC produced
measuring the secreted pro-inflammatory cytokines in serum-free from FBS containing higher cholesterol levels led to more efficient
media, including nitric oxide (NO) (Fig. 3a), tumour necrosis factor SNP internalization (Fig. 4a,b), which could be directly visualized by
receptor (TNFα) (Fig. 3b), interleukin (IL)-6 (Fig. 3c) and IL-1β (Fig. 3d). super-resolution microscopy (Fig. 4c–e). We also performed TEM
For both AuNPs and SNPs, PC formed in the higher concentration of analysis to visualize the internalization of AuNPs. Consistently, we
cholesterol could elicit a stronger immune response. Because neither found more accumulation of AuNPs-Cho-PC inside the HepG2 cells
free cholesterol nor cholesterol-pretreated NPs could increase the (Fig. 4f,g). Inductively coupled plasma mass spectrometry (ICP-MS)
Fig. 3 | Cho-PC elicited a stronger immune response of macrophages to NPs. significance was tested with a two-tailed, unpaired Student’s t-test. g, Gene
a–d, NO (a), TNF-α (b), IL-6 (c) and IL-1β (d) released from Raw264.7 cells after Ontology analyses of inflammation-related pathways in Cho-PC and Nor-PC
24 h incubation with SNPs-PC formed in FBS containing different concentrations groups. h, The expression of inflammation-related proteins in the control,
of cholesterol. The data are presented as mean ± s.d. with n = 3 (biologically Cho-PC and Nor-PC groups. i, The GSVA scores of three enriched inflammatory
independent experiments). Statistical significance was tested with a two- pathways for the control, Cho-PC and Nor-PC groups. The data are presented
tailed, unpaired Student’s t-test. e,f, NO (e) and IL-6 (f) released from Raw264.7 as box-and-whisker plots. The median is shown as a line in the middle of the box
cells after 24 h incubation with cholesterol (40 µg ml−1), NPs, NPs coated with extending from the 25th to the 75th percentile and the whiskers indicate the
cholesterol (NPs-Cho), NPs-Nor-PC and NPs-Cho-PC. The data are presented smallest and largest values. Statistical significance was tested with a two-tailed,
as mean ± s.d. with n = 3 (biologically independent experiments). Statistical unpaired Student’s t-test.

analysis found more Au detected in the Cho-PC group than in the Nor-PC cellular uptake of NPs should probably be ascribed to the specific
group (Fig. 4h). Because the addition of cholesterol to the medium protein components of Cho-PC induced by cholesterol rather than
inhibited NP internalization (Extended Data Fig. 8d–g), the enhanced cholesterol per se.
a b AuNPs
c AuNPs
P < 0.0001 SNPs P < 0.0001
SNPs
16 P = 0.0064 P = 0.0008
P < 0.0001 100
Concentration of NO (µM)
P < 0.0001 500 P = 0.0390
Concentration of TNF-α
P < 0.0001
Concentration of IL-6
P = 0.0284 P = 0.0088
12 P = 0.0013
375 75
AuNPs
(pg ml–1)
(pg ml–1)
SNPs
8 250 50
4 125 25
0 0 0
0.9 1.7 5.1 15.3 0.9 1.7 5.1 15.3 0.9 1.7 5.1 15.3
Cholesterol level (mM) Cholesterol level (mM) Cholesterol level (mM)
d P < 0.0001
e P < 0.0001
f
P = 0.0054
120 P = 0.0068 AuNPs
Concentration of NO (µM)
30 P < 0.0001 150
AuNPs P = 0.0039
P < 0.0001 AuNPs
Concentration of IL-1β
SNPs
Concentration of IL-6
SNPs SNPs
P < 0.0001
90
P = 0.0006 20 100
(pg ml–1)
(pg ml–1)
60
10 50
30
0 0 0
0.9 1.7 5.1 15.3
ol
ol
ol
Ps
ho
ol
Ps
ho
-C C
C
er
tr
er
-P
-P
-P
-P
tr
N
Cholesterol level (mM)
-C
-C
on
on
st
or
or
st
ho
ho
Ps
Ps
le
le
-N
-N
C
-C
N
ho
N
ho
Ps
Ps
Ps
Ps
C
C
N
N
N
g h Igfbp2
Tnf
Agt
Ptgs2
Serpine1
Selenof
Selenos
Response to endoplasmic reticulum stress −log(Padjust)
Hist1h3b
H3c1
Fgfbp1
H4c1
60 Ctsl
Regulation of inflammatory response Canx
Ctsb
40
Bcl2l1
Fgr
20 Hspa5
Production of molecular mediator involved Tapbp
Selenot
in inflammatory response Il1rap
Count Atp5pb
Hsp90b1
Colgalt1
Positive regulation of IL-1 production 20 Mtco2
40 Ctss
PF4
60 Parp1
Prdx4
IL-6 production Ndufb11
Map2k6
Cox7a2
Ndufb8
Bcl2l11
Nor-PC Cho-PC Cd14
Tnfrsf10b
Fcer1g
Tlr7
Vamp2
Hck
Plaur
H2afv
Cox6b1
Cox5b
Traf6
i IL-2-STAT5 IL-6-JAK-STAT3 Inflammatory
2 Mapk7
Tlr2
Tab3
signalling signalling response 1 Map2k7
Cox20
Card11
P = 0.024 Ccl9
P = 0.16 P = 0.014 Tgfb1
0
P = 0.0084 P = 0.25 P = 0.0089 Colec12
Spp1
Cd40
–1 Ikbkb
2.5 Usp25
Cox6a1
–2 Nfkb2
Rela
GSVA score
Map2k3
Row Z-score Vim
0 Tgfbr2
Tnfaip3
Nfkbia
Nfkb1
Lyn
Plau
Tradd
−2.5 Icam1
Cebpb
Mapk9
Map2k1
Csf2ra
Pik3ca
Ripk1
Irak4
ol C C ol C C ol C C
Ikbke
ntr or-P ho-P ntr or-P ho-P ntr or-P ho-P Traf2
Co N C Co N C Co N C Control Nor-PC Cho-PC
Il16

To further investigate the key factors required for the enhanced in the liver, lungs, intestine and testis 1 h after injection (Fig. 5e). Com-
cellular uptake of NPs induced by Cho-PC, we examined the internali- pared with normal mice, more AuNPs accumulated in the liver, intes-
zation of NPs in the presence of different types of endocytosis inhibi- tine and testis of hypercholesterolemic mice. For instance, 37% of the
tor (Fig. 4i–k and Extended Data Fig. 8h–m). Consistent with previous injected dose of AuNPs accumulated in the liver of hypercholester-
findings that the inhibitors of clathrin-mediated endocytosis, such olemic mice compared with only 15% in normal mice. On the contrary,
as chlorpromazine hydrochloride and dynasore, have general inhibi- more AuNPs accumulated in the lungs in the normal group than that
tory effects on NP cellular uptake35,36, they suppressed the uptake of in the hypercholesterolemic group. At 24 h after injection (Fig. 5f), the
both SNPs-Nor-PC and SNPs-Cho-PC. The micropinocytosis inhibitor distributional patterns changed. More AuNPs were also found in the
(5-(N-ethyl-N-isopropyl) amiloride (EIPA)) and the microtubule transport brain, kidneys and spleen in the hypercholesterolemic mice. Because
inhibitor (nocodazole) had no obvious inhibitory effects. Interestingly, metal-based NPs, the related ion release and their nanometallomes
methyl-β-cyclodextrin (MBCD), a sequestrator of cholesterol in the cell can potentially interact with essential elements inside the body and
membrane that suppresses caveolae- or lipid-raft-mediated uptake37, and cause dys-homoeostasis and adverse biological effects40, the essential
cytochalasin D, an inhibitor of actin dynamics, exhibited selective inhibi- element homoeostasis might be disturbed by AuNPs in individuals.
tion only on the uptake of SNPs-Cho-PC, but not SNPs-PC. Consistently, We also measured the levels of essential metal elements, including
super-resolution microscopy analysis showed that more SNPs-Cho-PC zinc and copper (Supplementary Fig. 5). Increased levels of zinc and
were co-localized with actin filaments than SNP-Nor-PC in the cytoplasm copper in the liver were observed at 1 h in the hypercholesterolemic
(Fig. 4c,d). These results suggest that the cholesterol- and actin-driven mice. When moving to 24 h, the differences in the liver disappeared,
mechanisms contribute to the enhanced uptake of corona–NP com- but differences in the testis, intestine, kidneys and spleen emerged,
plexes formed in the high-cholesterol serum. We also checked the cell suggesting that the distribution of certain metal elements may be
surface receptors required for uptake. Due to their established roles affected by the intravenous administration of nanomedicines in a
in the PC-mediated cellular uptake of NPs30,38,39, we tested whether the health-state-dependent manner. Looking our findings forward, we
low-density lipoprotein receptor (LDLR) and scavenger receptor class B envisage that the distinct in vivo fates such as biocompatibility and
type 1 (SR-B1) were involved in this case. In a competition assay, the cells biodistribution of NP-based nanomedicines might occur when intrave-
were pre-incubated with LDL and HDL in advance to block LDLRs and nously administered to healthy and hypercholesterolemic individuals,
SR-B1, and bovine serum albumin (BSA) incubation was used as a control. providing important insights for precision nanomedicine (Fig. 6).
Even though the pre-incubation inhibited the uptake of SNPs-Nor-PC and We also measured the in vivo distribution of SNPs using the fluo-
SNPs-Cho-PC to a certain extent, we found that only LDL and HDL had rescently labelled derivatives. The intravenously administered SNPs
more pronounced inhibitory effects on the uptake of SNPs-Cho-PC than in hypercholesterolemic mice were more likely to be delivered to the
that of SNPs-Nor-PC (Fig. 4l). Thus, LDLRs and SR-B1 might be involved in liver, spleen, kidney, testis and brain, and less likely to be delivered to
the enhanced internalization that is mediated by lipoproteins. the lungs (Extended Data Fig. 10c,f). It is interesting that the lipid-rich
organs seem to accumulate more NPs under hypercholesterolemic
In vivo biodistribution of NPs in conditions, and future work is needed to explore the possible link.
hypercholesterolemic mice Nevertheless, our results clearly showed that the in vivo biodistribution
We next wondered whether Cho-PC could change the in vivo biodistri- of NPs can be affected under hypercholesterolemia.
bution of NPs. A hypercholesterolemia mouse model was established
by feeding mice with a high-cholesterol diet for 15 weeks (Fig. 5a). The Conclusion
total serum cholesterol was 2.15 mM in the control mice and reached In this study, we set up an FBS system to investigate the impact of cho-
4.92 mM in the hypercholesterolemia mice (Fig. 5b). Tissue samples were lesterol on PC formation. We found that a high level of cholesterol
then collected at 1 h and 24 h after intravenous AuNPs administration enhanced the retention of apolipoproteins and reduced complement
and subjected to ICP-MS measurement (Fig. 5a). The mouse serum and proteins in the PC. We were aware of the limitations of this FBS sys-
plasma samples were incubated with AuNPs or SNPs to form PCs in situ. tem and verified the findings using hypercholesterolemic and normal
Detailed analyses found similar changes in the protein composition of serums prepared from mice and humans. Reassuringly, we found similar
the PC compared with the findings in FBS and human serum (Fig. 5c,d and results. But it should be noted that, in a complex environment like
Extended Data Fig. 9). Thus, the modulation of PC formulation by differ- the serum, the dynamic changes of apolipoproteins and complement
ential levels of serum cholesterol can also be recapitulated in the mouse proteins might also affect PC formation.
hypercholesterolemia model, consistent with the human serum results. We revealed the impact of Cho-PC on the in vitro and in vivo behav-
We next compared the in vivo biodistribution of AuNPs in blood iours of NPs. Cho-PC elicited a stronger immune response of macrophages
and tissues collected from normal and hypercholesterolemic mice to NPs. The cellular uptake of NPs was studied using HepG2 cells as an
(Extended Data Fig. 10a,b). Distributional differences were observed in vitro model of human hepatocytes. We found that Cho-PC promoted
Fig. 4 | Cho-PC enhanced cellular internalization of NPs in hepatocytes. AuNPs-Nor-PC and AuNPs-Cho-PC for 6 h. The data are presented as mean ± s.d.
a,b, HepG2 uptake of 100 µg ml−1 of FITC-labelled SNP–corona complexes with n = 4 (biologically independent experiments). Statistical significance was
formed in serum with different concentrations of cholesterol quantified by flow tested with a two-tailed, unpaired Student’s t-test. i,j, The summary of cellular
cytometry. For b, the average and s.d. represent three replicates of the median uptake of SNPs-Nor-PC (i) and SNPs-Cho-PC (j) at 6 h in the presence or absence
cell fluorescence intensities obtained. The data are presented as mean ± s.d. of various inhibitors. The data are presented as mean ± s.d. with n = 3 (biologically
with n = 3 (biologically independent experiments). Statistical significance was independent experiments). Statistical significance was tested with a two-tailed,
tested with a two-tailed, unpaired Student’s t-test. c,d, Super-resolution confocal unpaired Student’s t-test. k, The comparison of SNPs-Nor-PC and SNPs-Cho-PC
microscopy analyses of HepG2 treated with SNPs-Nor-PC (c) and SNPs-Cho-PC cellular uptake in the presence or absence of different inhibitors by normalizing
(d) for 6 h. The nuclei were stained with Hoechst (blue), the actin filaments were to the control groups. Statistical significance was tested with a two-tailed,
stained with phalloidin (red) and SNPs were labelled with FITC (green). e, The unpaired Student’s t-test. l, Cellular uptake of SNPs-Nor-PC and SNPs-Cho-PC in
median fluorescence intensity (MFI) ratio in c and d. The data are presented as the presence of BSA, HDL and LDL. The uptake efficiency was normalized by the
mean ± s.d. with n = 3 (biologically independent samples). Statistical significance corona complexes in a serum-free medium. The data are presented as mean ± s.d.
was tested with a two-tailed, unpaired Student’s t-test. f,g, Representative TEM with n = 3 (biologically independent experiments). Statistical significance was
images of HepG2 cells treated with AuNPs-Nor-PC (f) and AuNPs-Cho-PC (g) for tested with a two-tailed, unpaired Student’s t-test.
12 h. h, The cellular Au content was measured by ICP-MS after treatment with

a b l
P = 0.0205
P = 0.0327
80,000
120
Nor-PC Cho-PC
Fluorescence intensity (a.u.)

P < 0.0001 P = 0.0311
Relative uptake ratio (%)

Control 60,000 P = 0.0497
P = 0.0048
80
0.9 mM
40,000
1.7 mM
40
5.1 mM 20,000
15.3 mM
0 0
3 4 5 6 7
10 10 10 10 10
DL
ol
A
M
L
ol
LD
BS
tr
1m
m
FITC intensity
tr
H
on
on
9
1.7
.3
5.
0.
C
15
C
c d e
Nor-PC Cho-PC
P = 0.0002
3
MFI ratio of red/green

2
C
-P
-P
or
ho
5 µm 5 µm
C
f g h
Nor-PC Cho-PC
Content of Au (ng per 104 cells) 15 P = 0.0015
10
0
C
C
-P
-P
2 µm 2 µm
or
ho
N
C
i j P < 0.0001
k
Nor-PC Cho-PC
P = 0.0006
150 P = 0.0002
80,000 Nor-PC 80,000 P = 0.0001 Cho-PC
P = 0.0123
Relative uptake ratio (%)
P < 0.0001 P = 0.0271

P = 0.0078 P = 0.0237
60,000 P = 0.0352 60,000
100
Intensity
Intensity
40,000 40,000
50
20,000 20,000
0 0 0
e
re
od D
re
od D
Dy CD
re
od D
e
PA
PA
ha PA
om l
om l
om l
pr ntro
pr tro
in
ol
in
ol
in
ol
BC
BC
so
so
so
n
oc sin
tr
EI
EI
EI
az
az
az
B
az
az
az
si
N asi
on
on
na
na
na
M
M
la
la
o
l
C
C
C
ha
ha
Dy
Dy
oc
oc
pr
oc
oc
oc
N
N
or
or
or
yt
yt
yt
hl
hl
hl
C
C
C

Chow diet
Tissue collection ICP-MS
High-cholesterol diet
Treatment Injection Euthanasia Euthanasia
Time
0 15 weeks 1h 24 h
b c d
7.5 APOA2 C3
P < 0.0001 5 2.5
APOC3 CFH
Concentration of cholesterol (mM)
C4B
APOA5
Protein percentage in PC (%)
Protein percentage in PC (%)

4 2.0 CFB
APOA1 C5
5.0 APOA4 CFI
3 APOH 1.5 C2
APOE C8B
C9
APOB
2 1.0 C8G
2.5 APOC1 C1QB
APOC2 C1QA
1 APOF 0.5 C1RA
C1SA
0 0 0
al ia al ia al ia
m em rm em rm em
r
ol No ol No ol
No er er er
st st st
le ole ole
o ch ch
ch er er
p er p p
Hy Hy Hy
e f
1h 24 h
48 48
Normal Hypercholesterolemia Normal Hypercholesterolemia
P < 0.0001
32 32
P < 0.0001
AuNPs (%ID)
AuNPs (%ID)
P = 0.0103 P = 0.0299
P < 0.0001
16 16
P = 0.0110 P = 0.0001
P < 0.0001 P = 0.0002
0 0
en
e
s
n
tis
ne
n
gs
ne
tis
e
gs
s
r
rt
rt
r
d
d
ey
ve
in
ey
ve
ai
ee
ai
in
ea
ea
oo
oo
le
s
n
n
Bo
Bo
Br
st
Br
st
dn
Te
Li
Te
dn
Li
l
Lu
Lu
Sp
Sp
H
Bl
te
Bl
te
Ki
Ki
In
In
Fig. 5 | In vivo biodistribution of NPs in normal and hypercholesterolemic hypercholesterolemic mice. e,f, The percentage of injected dose (%ID) of AuNPs
mice. a, Scheme of the experiments. b, The serum cholesterol content in normal in blood and different tissues from the normal and hyperlipidemic mice at 1 h
and hypercholesterolemic mice. The data are presented as mean ± s.d. with n = 5 (e) and 24 h (f) after intravenous injection of AuNPs at the doses of 0.1 g per kg
(biologically independent mice). Statistical significance was tested with a two- body weight (n = 5). The data are presented as mean ± s.d. with n = 5 (biologically
tailed, unpaired Student’s t-test. c,d, The percentage of apolipoproteins (c) and independent mice). Statistical significance was tested with a two-tailed, unpaired
complement proteins (d) identified in PCs formed in serums from normal and Student’s t-test.

Hepatocytes Hepatocytes
Caveolae Lipid raft Clathrin Macropinocytosis Caveolae Lipid raft Clathrin Macropinocytosis
Cholesterol Scavenger receptor
Nanoparticles LDL receptor
Complement Dynamin
Cytoplasm Apolipoprotein Caveolin

Cytoplasm
Lipid raft Clathrin
Fig. 6 | The proposed fates of NPs in healthy and hypercholesterolemic is driven by enhanced cholesterol and actin-driven endocytosis mediated by
individuals. When NPs are intravenously administered, the personalized lipoproteins. The personalized PCs in hypercholesterolemic individuals further
biomolecular corona formed on the surface of NPs greatly affects the fates of NPs change the in vivo biodistribution of NPs, and more NPs are target-delivered
in vivo. The NPs with Cho-PC elicit a stronger immune response in macrophages to the liver and brain in hypercholesterolemic patients, which has important
and have higher internalization in hepatocytes. The enhanced internalization implications for precision nanomedicine.
cellular uptake of NPs in HepG2 cells, which could be directly visualized by author contributions and competing interests; and statements of
super-resolution microscopy and TEM. Detailed analyses showed that the data and code availability are available at https://doi.org/10.1038/
enhanced internalization was mediated by SR-B1 and LDLR on the cell sur- s41565-023-01455-7.
face and the actin-driven mechanisms inside the cells. Most importantly,
we found that Cho-PC formed in hypercholesterolemic serums indeed References
changed the in vivo biodistribution of NPs. For metal-based NPs such as 1. Quesada-Gonzalez, D. & Merkoci, A. Nanomaterial-based devices
AuNPs, they might interact with essential elements inside the body and for point-of-care diagnostic applications. Chem. Soc. Rev. 47,
cause dys-homoeostasis and adverse biological effects. Thus, our results 4697–4709 (2018).
together suggest that the safety of nanomedicines should be evaluated 2. Irvine, D. J. & Dane, E. L. Enhancing cancer immunotherapy with
in a health-state-dependent manner. It has become increasingly clear nanomedicine. Nat. Rev. Immunol. 20, 321–334 (2020).
that NPs will have wide applications in medicine. Our findings suggest 3. Stater, E. P., Sonay, A. Y., Hart, C. & Grimm, J. The ancillary effects
that the PC formed could also be affected by the surrounding metabolic of nanoparticles and their implications for nanomedicine. Nat.
environment, and we believe that our work will stimulate more careful Nanotechnol. 16, 1180–1194 (2021).
design for developing precision nanomedicine in future studies. 4. Guerrini, G., Magrì, D., Gioria, S., Medaglini, D. & Calzolai, L.
Characterization of nanoparticles-based vaccines for COVID-19.
Online content Nat. Nanotechnol. 17, 570–576 (2022).
Any methods, additional references, Nature Portfolio reporting 5. Gadekar, V. et al. Nanomedicines accessible in the market
summaries, source data, extended data, supplementary infor- for clinical interventions. J. Control. Release 330, 372–397
mation, acknowledgements, peer review information; details of (2021).

6. Thi, T. T. H. et al. Lipid-based nanoparticles in the clinic and 26. Civeira, F., Arca, M., Cenarro, A. & Hegele, R. A. A mechanism-
clinical trials: from cancer nanomedicine to COVID-19 vaccines. based operational definition and classification of hypercholes
Vaccines 9, 359 (2021). terolemia. J. Clin. Lipidol. 16, 813–821 (2022).
7. Monopoli, M. P., Åberg, C., Salvati, A. & Dawson, K. A. 27. Kim, S. H. et al. Understanding the biomolecular coronas of
Biomolecular coronas provide the biological identity of nanosized high-density lipoproteins on pegylated Au nanoparticles:
materials. Nat. Nanotechnol. 7, 779–786 (2012). implication for lipid corona formation in the blood. ACS Appl.
8. Ren, J. et al. Chemical and biophysical signatures of the Nano Mater. 5, 2018–2028 (2022).
protein corona in nanomedicine. J. Am. Chem. Soc. 144, 28. Kim, H., Kumar, S., Kang, D. W., Jo, H. & Park, J. H. Affinity-driven
9184–9205 (2022). design of cargo-switching nanoparticles to leverage a
9. Latreille, P.-L. et al. Scratching the surface of the protein corona: cholesterol-rich microenvironment for atherosclerosis therapy.
challenging measurements and controversies. ACS Nano 16, ACS Nano 14, 6519–6531 (2020).
1689–1707 (2022). 29. Fu, Q., Yu, L., Wang, Y., Li, P. & Song, J. Biomarker-responsive
10. Li, M. et al. Nanoparticle elasticity affects systemic circulation nanosystems for chronic disease theranostics. Adv. Funct. Mater.
lifetime by modulating adsorption of apolipoprotein A-I in corona 33, 2206300 (2023).
formation. Nat. Commun. 13, 4137 (2022). 30. Kim, K. R., Kim, J., Back, J. H., Lee, J. E. & Ahn, D. R. Cholesterol-
11. Kamaly, N. Nanoparticle protein corona evolution: from biological mediated seeding of protein corona on dna nanostructures for
impact to biomarker discovery. Nanoscale 14, 1606–1620 (2022). targeted delivery of oligonucleotide therapeutics to treat liver
12. Ju, Y. et al. Person-specific biomolecular coronas modulate fibrosis. ACS Nano 16, 7331–7343 (2022).
nanoparticle interactions with immune cells in human blood. ACS 31. Dreaden, E. C., Alkilany, A. M., Huang, X., Murphy, C. J. & El-Sayed,
Nano 14, 15723–15737 (2020). M. A. The golden age: gold nanoparticles for biomedicine. Chem.
13. Hajipour, M. J., Laurent, S., Aghaie, A., Rezaee, F. & Mahmoudi, M. Soc. Rev. 41, 2740–2779 (2012).
Personalized protein coronas: a ‘key’ factor at the nano-bio 32. Kim, S. E. et al. Ultrasmall nanoparticles induce ferroptosis in
interface. Biomater. Sci. 2, 1210–1221 (2014). nutrient-deprived cancer cells and suppress tumour growth.
14. Shannahan, J. H. et al. From the cover: disease-induced disparities Nat. Nanotechnol. 11, 977–985 (2016).
in formation of the nanoparticle-biocorona and the toxicological 33. Ke, P. C., Lin, S., Parak, W. J., Davis, T. P. & Caruso, F. A decade of
consequences. Toxicol. Sci. 152, 406–416 (2016). the protein corona. ACS Nano 11, 11773–11776 (2017).
15. Ren, J. et al. Precision nanomedicine development based on 34. Xiao, Q. et al. The effects of protein corona on in vivo fate of
specific opsonization of human cancer patient-personalized nanocarriers. Adv. Drug Deliv. Rev. 186, 114356 (2022).
protein coronas. Nano Lett. 19, 4692–4701 (2019). 35. Blanco, E., Shen, H. & Ferrari, M. Principles of nanoparticle
16. Lazarovits, J. et al. Synthesis of patient-specific nanomaterials. design for overcoming biological barriers to drug delivery.
Nano Lett. 19, 116–123 (2019). Nat. Biotechnol. 33, 941–951 (2015).
17. Di Santo, R. et al. Personalized graphene oxide-protein corona in 36. Macia, E. et al. Dynasore, a cell-permeable inhibitor of dynamin.
the human plasma of pancreatic cancer patients. Front. Bioeng. Dev. Cell 10, 839–850 (2006).
Biotechnol. 8, 491 (2020). 37. Francia, V. et al. Corona composition can affect the
18. Chetwynd, A. J. & Lynch, I. The rise of the nanomaterial mechanisms cells use to internalize nanoparticles.
metabolite corona, and emergence of the complete corona. ACS Nano 13, 11107–11121 (2019).
Environ. Sci. Nano 7, 1041–1060 (2020). 38. Lara, S. et al. Identification of receptor binding to the
19. Raesch, S. S. et al. Proteomic and lipidomic analysis of biomolecular corona of nanoparticles. ACS Nano 11,
nanoparticle corona upon contact with lung surfactant reveals 1884–1893 (2017).
differences in protein, but not lipid composition. ACS Nano 9, 39. Ngo, W. et al. Identifying cell receptors for the nanoparticle
11872–11885 (2015). protein corona using genome screens. Nat. Chem. Biol. 18,
20. Braccia, C. et al. The lipid composition of few layers graphene 1023–1031 (2022).
and graphene oxide biomolecular corona. Carbon 185, 40. Benetti, F., Bregoli, L., Olivato, I. & Sabbioni, E. Effects of metal
591–598 (2021). (loid)-based nanomaterials on essential element homeostasis: the
21. Liu, K., Salvati, A. & Sabirsh, A. Physiology, pathology and the central role of nanometallomics for nanotoxicology. Metallomics
biomolecular corona: the confounding factors in nanomedicine 6, 729–747 (2014).
design. Nanoscale 14, 2136–2154 (2022).
22. Kobos, L. M. et al. An integrative proteomic/lipidomic analysis of Publisher’s note Springer Nature remains neutral with regard to
the gold nanoparticle biocorona in healthy and obese conditions. jurisdictional claims in published maps and institutional affiliations.
Appl. Vitr. Toxicol. 5, 150–166 (2019).
23. Lima, T., Bernfur, K., Vilanova, M. & Cedervall, T. Understanding Springer Nature or its licensor (e.g. a society or other partner) holds
the lipid and protein corona formation on different sized exclusive rights to this article under a publishing agreement with
polymeric nanoparticles. Sci. Rep. 10, 1129 (2020). the author(s) or other rightsholder(s); author self-archiving of the
24. Tavakol, M. et al. Disease-related metabolites affect accepted manuscript version of this article is solely governed by the
protein-nanoparticle interactions. Nanoscale 10, 7108–7115 (2018). terms of such publishing agreement and applicable law.
25. Luo, J., Yang, H. & Song, B. L. Mechanisms and regulation
of cholesterol homeostasis. Nat. Rev. Mol. Cell Biol. 21, © The Author(s), under exclusive licence to Springer Nature Limited
225–245 (2020). 2023

Methods (5 mg ml−1) were mixed with the cholesterol solution (15.3 mM) for 1 h,
Reagents and materials and then centrifuged at 20,000g to remove free cholesterol. After that,
Cholesterol (C8667), dithiothreitol (D0632), BSA (A1933) and gold (III) NPs pre-adsorbed with cholesterol were incubated with FBS for 1 h to
chloride trihydrate (520918) were obtained from Sigma-Aldrich. Chlor- form the PC, termed PC-II. Lastly, PC-II was purified by PBS washing
promazine (HY-12708), MBCD (HY-101461), dynasore (HY-15304), EIPA three times for the following analysis. As for the second condition,
(HY-101840), cytochalasin D (HY-N6682) and nocodazole (HY-13520) the initial PC was prepared by the co-incubation of NPs and FBS and
were purchased from MedChem Express. Trypsin (90057) and other subsequently separated by centrifugation at 20,000g. Next, the pre-
liquid chromatography with tandem mass spectrometry (LC–MS/ pared NP–corona complexes were re-dispersed into the FBS solution
MS)-related reagents were purchased from Thermo Fisher. The specific containing the high concentration of cholesterol (15.3 mM) for 1 h, and
primary antibodies against APOE (ab183597) and C1QB (ab92508) the secondarily formed PC on NPs, termed PC-III, was separated and
were purchased from Abcam. The purified proteins including APOE purified according to the abovementioned process.
(PAA704Hu01), C1QB (RPD208Hu01), LDL (PAB107Hu01) and HDL
(PAB006Hu01) were purchased from Cloud-Clone. The TNFα assay kit Preparation of protein samples for mass spectrometric
(ab208348), IL-6 ELISA kit (ab222503), IL-1β ELISA kit (ab197742) and characterization
cell counting kit (ab228554) were obtained from Abcam. The bicin- The in-gel digestion for mass spectrometric characterization of pro-
choninic acid assay (P0012S) and NO assay kit (S00215) were purchased teins was conducted according to a previous study with slight modi-
from Beyotime Biotechnology. The total cholesterol content assay kit fication43. First, samples were loaded into the 10% SDS–PAGE gel, and
(BC1980) was obtained from Solarbio. The red-fluorescent SNPs (40- the gel was run at 120 V until bromophenol blue reached about 1.5 cm
00-701) and 70 nm green-fluorescent SNPs (42-00-701) were purchased below the stacking gel. Then the gel lanes were cut into approximately
from micromod Partikeltechnologie. Liposomes (Q-0172322) were 1 × 1 mm pieces and transferred into an Eppendorf tube. Next, gel pieces
purchased from Xian Qiyue Biology. AuNPs and SNPs were synthesized were destained by incubation with 200 µl of 40 mM NH4HCO3/50%
in-house and the process is described in Supplementary Information. acetonitrile for 15 min at 37 °C twice. After the destaining solution
was removed, the gel pieces were subsequently rehydrated in 100 µl
Characterization of NPs of 20 mM DTT and 100 mM NH4HCO3 solution for 30 min at 56 °C, and
TEM images of NPs were performed with an HT7800 microscope then washed with 500 µl of acetonitrile. Afterwards, 200 μl of 55 mM
(Hitachi) to investigate the morphology and size of the NPs. A IAA and 100 mM NH4HCO3 were added and the solutions were incu-
drop of aqueous dispersion from NP samples was placed onto a bated at room temperature in the dark for 20 min. After the alkylation
carbon-supported copper grid, and then dried under ambient condi- of proteins was completed, the gel pieces were washed with 100 µl of
tions. For the observation of PCs, the NP samples were negatively stained 100 mM NH4HCO3 and acetonitrile (1:1) twice. For the gel digestion,
using 1% uranyl acetate according to the literature41. For cryo-TEM imag- the gel pieces were hydrated with digestion buffer (200 µl of 40 mM
ing42, each sample was deposited to a glow-discharged holey carbon NH4HCO3 containing 5 ng μl−1 trypsin) and incubated for up to 18 h at
grid, and frozen using the Vitrobot Mark IV (Thermo Fischer Scien- 37 °C. The supernatants of the digestion solutions were collected and
tific). Cryo-TEM images were collected using a Talos F200X G2 200 kV transferred into a 1.5 ml low-binding Eppendorf tube. One hundred
Cryo-S/TEM (Thermo Fischer Scientific). The hydrodynamic size and microlitres of 5% formic acid in 50 mM NH4HCO3 was then added to
zeta potential profiles of NPs were determined by dynamic light scat- cover the gel pieces, and the mixtures were incubated for 15 min at 37 °C
tering using a Zetasizer Nano S90 DLS system (Malvern Instruments). to extract the peptides released in the supernatants. The extraction
The results represent the average of three consecutive measurements. process was repeated once with 150 µl of 50% acetonitrile/5% formic
acid aqueous solution. All the supernatants were combined and dried
Preparation and purification of PCs in a vacuum centrifuge. The dried peptides were stored at −20 °C until
Equal volumes of SNPs (5 mg ml−1) or AuNPs (800 pmol), 10% FBS LC–MS/MS analysis.
(Gibco), and cholesterol or PBS were mixed at the desired concentra-
tions and then incubated at 37 °C for 1 h. The above mixtures were then LC–MS/MS analysis
centrifuged at 20,000g for 10 min at 4 °C to remove redundant FBS All proteomic samples were analysed in our lab by an Orbitrap Fusion
protein as well as cholesterol. Next, the NP–corona complexes were Lumos mass spectrometer (Thermo Fisher Scientific) coupled with an
washed with PBS (20,000g) for 10 min three times and then resus- Ultimate 3000 nano-LC system. As huge variations in the proteomics
pended in PBS to reduce possible contamination from proteins with analysis of PCs across core facilities have been reported41, all the prot-
low affinity to NPs. To further lower the risk of protein contamination eomic samples were prepared and analysed in our lab. To avoid possible
at this speed, lower speeds of 8,000g and 14,000g were also used to variation in the outcomes from these artificial factors, the proteomic
collect corona-coated NPs, and no observed differences were found in study was performed by one person, and the samples (corona or cell
the collected PCs (Supplementary Fig. 6). One hundred microlitres of lysate) from one batch experiment were processed in parallel and
100 mM DTT and 1% SDS were added to the NP–corona complexes with analysed in the same mass spectrometer without interruption. Dried
shaking at 37 °C for 30 min, and then the samples were centrifuged at peptide samples were resolved in 10 µl of 0.1% formic acid and 1% ace-
20,000g for 20 min to fully recover proteins from NP adsorption. For tonitrile and then transferred to a 96-well plate autosampler. Before
human PC formation, SNPs (5 mg ml−1) were incubated with different the sample loading, the peptide concentrations were measured (Pierce
volunteer serums (cholesterol levels at 4.4 mM, 5.6 mM and 7.2 mM; Quantitative Colorimetric Peptide Assay (no. 23275, Thermo Scientific))
these samples were collected following the approved protocols that and adjusted into 1 µg µl−1, and the injection volume was 1 µl. Peptides
complied with all relevant ethical regulations at the First Affiliated were trapped on a 2 cm Acclaim PepMap 100 C18 column (75 μm inner
Hospital of Gannan Medical University) for 1 h at 37 °C, followed by cen- diameter, 3 µm resin, Thermo Fisher Scientific) and washed with 0.5%
trifugation at 20,000g for 10 min to remove unbound serum proteins. formic acid and 2% acetonitrile at a flow rate of 5 µl min−1 for 5 min.
Then the PC pellets were washed with PBS three times. Mouse serum Then the peptides were eluted on an Acclaim PepMap 100 C18 HPLC
and plasma PC formation were prepared according to the same steps. reverse-phase analytical column (130 Å, 2 µm, 75 µm × 250 mm) using
a flow rate of 0.4 ml min−1 and a 60 min gradient from 5% to 90% phase
Competitive assay for protein adsorption on NPs B (phase A, 0.1% formic acid in distilled water; phase B, 0.1% formic acid
Two competition conditions were set up to probe the protein adsorp- and 80% acetonitrile). The mass data were collected by the Xcalibur
tion on NPs after immersion in FBS. For the first condition, NPs software (version 4.2) in the data-dependent acquisition mode.
Nature Nanotechnology
All MS spectra were collected in positive-ion mode with a full scan Molecular dynamics simulation
MS spectra range of 300–1,500 m/z at 60,000 resolutions using an The three-dimensional structure of APOE (Protein Data Bank ID 1LE2)
Orbitrap detector. The top-20 most abundant precursors with charge was obtained from the Research Collaboratory for Structural Bioinfor-
states of 2–6 were subjected to higher-energy collisional dissocia- matics Protein Data Bank (https://www.rcsb.org/). Molecular docking
tion with a normalized collision energy of 38% in the ion trap with a calculations were first performed to obtain the optimized binding
fixed first mass of 110 m/z, and an isolation window of 0.7 m/z. Other conformation of the SiO2–APOE system using the HEX software (version
parameters in the centroid format: automatic gain control target, 6.3). Briefly, the [100] surface structure of SiO2 was built and expanded
standard; maximum injection time mode, auto; dynamic exclusion, to 10 nm, and APOE was placed at ~2 nm above the surface of SiO2. The
60 s; and exclude isotopes, true. The raw data have been deposited to optimized conformation was picked out based on the docking scores.
the ProteomeXchange (https://proteomecentral.proteomexchange. Next, the molecular dynamics simulation was carried out as follows.
org/cgi/GetDataset, identifier PXD041073). First, the obtained configuration of the SiO2–protein complex was
placed in a simulation box with a size of 9.8 nm × 9.7 nm × 8.15 nm
MS data processing (x, y, z). To simulate the two conditions with or without cholesterol, 15
Mass spectra were analysed by the Proteome Discoverer software cholesterol molecules were filled into the simulation box. The molecu-
integrated with the Andromeda search engine (version 2.4, Thermo lar dynamics simulations were carried out with the Gromacs software
Fisher Scientific). The collected MS files were searched against a reverse package (version 2021.5) under constant temperature, pressure and
concatenated and non-redundant variant of bovine and mice proteome periodic boundary conditions. Besides, the Charmm36 full atomic
database downloaded from the website of Uniprot (https://www.uni- force field and TIP3P water model were applied in the system. During
prot.org/). The main searching parameters in the identification process the molecular dynamics simulations, all hydrogen bonds were con-
were set as follows: MS error tolerance of 10 ppm, MS/MS error toler- strained to their equilibrium values by the LINCS algorithm45, and the
ance of 0.6 Da, trypsin as digest enzyme, carbamidomethyl cysteine as integration step was 2 fs. The electrostatic interaction was calculated
a fixed modification, N-terminal acetylation and methionine oxidation by the particle-mesh Ewald method. The V-rescale thermostat was used
as a variable modification. A minimum peptide length of seven amino to control the system temperature at 300 K, and the Parrinello Rahman
acids and a maximum of two missed cleavages were allowed. For pep- algorithm was used to maintain the pressure to 1 bar. In the production
tide and protein identifications, a false discovery rate was determined runs, the energy of the two systems was minimized using the steepest
by automatically searching a decoy database with a setting of less descent after eliminating the close contact between atoms. Then, the
than 1%. At least one unique or razor peptide per protein group was constant-temperature, constant-volume ensemble (NVT) balance
required for protein identification. Label-free quantification based on simulation of 100 ps at 300 K was conducted, followed by 100 ns length
MS1 intensity was performed for the protein quantification with ratio molecular dynamics simulations of different SiO2–protein systems,
calculation of pairwise ratio and an alignment time window of 5 min respectively. The coordinates were saved every 10 ps, and GROMACS
based using the Proteome Discoverer software. and VMD were used to visualize the completed simulation results.
Bioinformatic analysis Super-resolution confocal imaging for cellular uptake

Heat maps of normalized protein abundance were plotted using R (ver- Cells (1 × 104 per well) were seeded in confocal dishes and exposed to
sion 4.0.3). The function of corona composition was classified using the 50 μg ml−1 of fluorescein isothiocyanate (FITC)-conjugated SNPs for
David website (https://david.ncifcrf.gov/). Differential proteins were 6 h. Then cells were washed with PBS three times to remove free SNPs.
discriminated based on a fold change of ≥1.5 and a P value of <0.05, Subsequently, cells were fixed with 4% paraformaldehyde for 20 min
and heat maps and volcano maps were drawn to describe the results. and then permeabilized with 0.2% Triton X-100 for 5 min. The actin
The biological process in the Gene Ontology and Kyoto Encyclopedia filaments and nuclear of the cells were stained in turn by incubation
of Genes and Genomes (KEGG) pathway were performed to visualize with Phalloidin (Thermo Fisher Scientific) for 10 min and Hoechst
functional profiles and pathway enrichment by the ‘clusterProfiler’ R (Thermo Fisher Scientific) for 30 min. Cells were observed using a
package, and the net plot function was used to visualize the enrichment super-resolution microscope (N-SIM, Nikon).
analysis for biological process terms. The hallmark gene sets were
obtained from the Gene Set Enrichment Analysis website (https://www. ICP-MS measurement
gsea-msigdb.org/gsea/msigdb/). The normalized expression matrix of HepG2 cells were seeded in 24-well plates and further cultured for 24 h.
proteins used as the input file was conducted to the GSVA with default Next, the cells were incubated with 20 μg ml−1 AuNPs until the indicated
parameters through the ‘GSVA’ R package. The GSVA score obtained time. The cells were washed three times with PBS and underwent cell
by the above described was used to compare the differences between counting by a hemocytometer after tryptic digestion. The collected cells
pathways of interest in different groups. were digested in aqua regia and diluted with 2% HNO3 into a volume of
5 ml. Finally, the concentration of Au was determined by ICP-MS (Nex-
MST assay for measuring the affinity of APOE and C1QB to SNPs lON 350X, PerkinElmer). A NexION quartz cyclonic spray chamber (part
The equilibrium dissociation constant (KD) of SNPs with C1QB and no. N8145013, PerkinElmer) and a concentric quartz nebulizer (part no.
APOE was measured via MST (Monolith NT.115, Nano Temper)44. SNP 8152375, MEINHARD) were used for ICP-MS measurements. The detailed
solutions with a series of preconcerted concentrations were prepared, operating parameters of ICP-MS are summarized in Supplementary Table 1.
and then mixed with the same volume of the C1QB and APOE solu-
tions (0.5 mg ml−1) in the presence or absence of 5.1 mM cholesterol. Animal experiments
Next, the series of mixtures were separately pulled into capillaries All animal experiments were performed in compliance with the institu-
and detected via MST with 40% power. The direct binding of C1QB tional ethics committee regulations and guidelines on animal welfare
and APOE to SNPs was reflected as a change in the thermophoresis with the approval by the China Academy of Chinese Medical Sciences
of the fluorescently labelled protein upon the complex formation. (approval number 2022B155). Male c57/6J mice (20–25 g) were provided
With increasing concentrations of Cho-NP and Nor-NP, the fluores- by the Vitalriver Experimental Animal Co. The mice were housed under
cence counts of bound complex versus unbound free C1QB and APOE normal laboratory conditions (constant temperature at 22 °C; 50–60%
decreased, which indicated a positive thermophoretic effect and vice ambient humidity with 12 h/12 h light/dark cycle) with free access to
versa44. The data were analysed by MO Affinity Analysis Software water and standard rodent food. Mice were kept on a 12 h light/dark cycle.
(version 2.3). After acclimation, mice were randomized into groups (five mice per
group). The mice in the model group were fed with high-cholesterol 43. Shevchenko, A., Tomas, H., Havlis, J., Olsen, J. V. & Mann, M. In-gel
diet (Dyets, no. ASHF3) for 15 weeks, while the mice in the control group digestion for mass spectrometric characterization of proteins and
were fed with a normal diet. After the hyperlipidemia model was estab- proteomes. Nat. Protoc. 1, 2856–2860 (2006).
lished successfully, the mice were intravenously injected with AuNPs 44. Cao, Z. T. et al. Protein binding affinity of polymeric nanoparticles
(1 μmol ml−1) in saline at a dose of 0.1 g per kg of body weight. The mice as a direct indicator of their pharmacokinetics. ACS Nano 14,
in the two groups were excised at 1 h and 24 h, and blood and major tis- 3563–3575 (2020).
sues including heart, liver, spleen, lungs, kidneys, testis, intestine, brain 45. Hess, B. P-LINCS: a parallel linear constraint solver for molecular
and bone were collected. The serum was separated by centrifugation simulation. J. Chem. Theory Comput. 4, 116–122 (2008).
of standing blood for detection of cholesterol level. The weighted bio-
logical samples were transferred into tetrafluoroethylene tubes with Acknowledgements
2 ml of 70% HNO3. Afterwards, the tight tubes were placed in a microwave The work was supported by grants from the National Key Research
digestion system (MARS) and kept at 240 °C for 25 min. Upon cooling, and Development Program of China (2020YFA0908000 and
the digested solutions were diluted into a volume of 40 ml by ultrapure 2022YFC2303600); the Innovation Team and Talents Cultivation
water for ICP-MS measurements. All the tissue and blood samples were Program of National Administration of Traditional Chinese
stored at −80 °C before ICP-MS measurements. The concentrations of Medicine (no. ZYYCXTD-C-202002); the National Natural Science
standard solutions of gold, zinc and copper elements were diluted as Foundation of China (32201177, 32000026, 82141001, 82274182
1, 5, 20 and 100 ppb. The accumulation levels in organs were expressed and 82074098); the CACMS Innovation Fund (CI2021A05101
as the percent gold of the injected dose and nanograms per gram of and CI2021A05104); the Scientific and Technological Innovation
organ. For the ex vivo imaging, Cy3-labelled SNPs were injected into Project of China Academy of Chinese Medical Sciences
the tail vein of the mice at a dose of 20 mg per kg body weight. Mice (CI2021B014); the Science and Technology Foundation of Shenzhen
were killed at 1 h and 12 h to collect blood and major tissues, including (JCYJ20210324115800001); the Science and Technology Foundation
the heart, liver, spleen, lung, kidney, testis, intestine, brain, bone and of Shenzhen (Shenzhen Clinical Medical Research Center for
stomach. All organs and blood were fluorescently imaged by an in vivo Geriatric Diseases); the National Key R&D Program of China Key
imaging system (IVIS SPEOTRUM, PerkinElmer). Semi-quantitative projects for international cooperation on science, technology and
analysis was performed by ImageJ software (version 1.8.0). innovation (2020YFE0205100); the Fundamental Research Funds
for the Central Public Welfare Research Institutes (ZZ14-YQ-061,
Statistics and reproducibility ZZ16-ND-10–24, ZZ14-YQ-057, ZZ14-YQ-050, ZZ14-YQ-051,
Statistical analysis was performed on GraphPad Prism (version 9.4.0). ZZ14-FL-002, ZZ14-ND-010 and ZZ15-ND-10); GuangDong Basic and
All quantitative data were expressed as mean ± s.d. and evaluated with Applied Basic Research Foundation (2021A1515012164), Shenzhen
unpaired one-tailed Student’s t-test. All quantitative experiments were Key Laboratory of Kidney Diseases (ZDSYS201504301616234);
performed at least three times and the replicate samples are indicated and Shenzhen Governmental Sustainable Development Fund
as n. Differences were considered statistically significant when P val- (KCXFZ20201221173612034).
ues were less than 0.05. The experiments of transmission electron
microscopy, gel electrophoresis and western blot were independently Author contributions
repeated three times with similar results, and representative images J.W., C.X. and H.T. conceived the project and supervised the study.
are shown in the figures. H.T. and Y. Zhang designed and performed all experiments with the
help from the other authors. T.Y., C.W., L.Q. and J.Z. assisted in the
Schematic illustrations collection and analysis of proteomic data. Y. Zhu, C.W. and Y.K.W.
Schematic illustrations were created with Adobe Illustrator and the assisted in the molecular dynamics simulation. J.L., Y.S. and L.Z
online software of BioRender. assisted in the animal experiment and inductively coupled plasma
mass spectrometry measurement. Y.L. and H.W. guided and assisted
Reporting summary in the experiment design. H.T., Y. Zhang, C.X. and J.W. analysed data
Further information on research design is available in the Nature Port- and wrote the paper. J.W., C.X. and Y.K.W. edited the paper. All authors
folio Reporting Summary linked to this article. approved the paper.
Data availability Competing interests

All proteomics raw data were deposited to the ProteomeXchange The authors declare no competing interests.
(https://proteomecentral.proteomexchange.org/cgi/GetDataset) with
the identifier PXD041073. Source data are provided with this paper. Additional information
Additional datasets supporting our findings in this study are available Extended data is available for this paper at
from the corresponding authors upon reasonable request. https://doi.org/10.1038/s41565-023-01455-7.
Code availability Supplementary information The online version

The analysis codes for molecular dynamics simulation and bioinfor- contains supplementary material available at
matic analysis of proteomic data are preserved in the China Academy https://doi.org/10.1038/s41565-023-01455-7.
of Chinese Medical Sciences and are available from the corresponding
authors upon reasonable request. Correspondence and requests for materials should be addressed to
Chengchao Xu or Jigang Wang.
References
41. Ashkarran, A. A. et al. Measurements of heterogeneity in Peer review information Nature Nanotechnology thanks Andy
proteomics analysis of the nanoparticle protein corona across Chetwynd, Morteza Mahmoudi and the other, anonymous, reviewer(s)
core facilities. Nat. Commun. 13, 6610 (2022). for their contribution to the peer review of this work.
42. Sheibani, S. et al. Nanoscale characterization of the biomolecular
corona by cryo-electron microscopy, cryo-electron tomography, Reprints and permissions information is available at
and image simulation. Nat. Commun. 12, 573 (2021). www.nature.com/reprints.
Extended Data Fig. 1 | Characterization of cholesterol-mediated protein Statistical significance was tested with a two-tailed, unpaired Student’s t-test.
corona on NPs. (a) Coomassie-stained SDS-PAGE gel of PC on SNPs under (c) Densitometric analysis of Extended Data Fig. 1a. (d) Densitometric analysis
different concentrations of cholesterol. (b) The total protein amounts of PC on of Fig. 1a. (e) Coomassie-stained SDS-PAGE gel of human serum samples with
SNPs under different concentrations of cholesterol. The data were presented as different concentrations of cholesterol.
mean ± standard deviation with n = 3 (biologically independent experiments).
Extended Data Fig. 2 | See next page for caption.
Extended Data Fig. 2 | Characterization of NPs-Nor-PC and NPs-Cho-PC. SNPs-Cho-PC. (d) Summary of physicochemical parameters including TEM size,
(a) Cryo-TEM images of NPs, NPs-Nor-PC, and NPs-Cho-PC. Cryo-TEM imaging HD, polydispersity index (PDI), and zeta potential of naked NPs, NPs-Nor-PC,
showed that cholesterol did not affect the possible protein contamination in and NPs-Cho-PC for AuNPs and SNPs. (e) The loading amount of cholesterol on
the background. (b) Hydrodynamic (HD) size distribution AuNPs, AuNPs-Nor- SNPs under different concentrations of cholesterol. The data were presented as
PC, and AuNPs-Cho-PC. (c) HD size distribution of SNPs, SNPs-Nor-PC, and mean ± standard deviation with n = 3 (biologically independent experiments).
Extended Data Fig. 3 | Schematic illustration for the mechanistic formation PC-III refers to the final PC formed in FBS containing cholesterol after SNPs were
of Cho-PC and proteomic fingerprints identification. (a) Schematic process pre-incubated with FBS. PC-IV refers to the Cho-PC formed in FBS containing
of PC on SNPs under different conditions for elucidating the mechanism of cholesterol. (b) Schematic diagram of identification and quantification of
Cho-PC formation. PC-I refers to the normal PC formed in FBS. PC-II refers to the Cho-PC and Nor-PC at the proteomic level.
PC formed in FBS after SNPs were pre-incubated with the cholesterol solution.
Extended Data Fig. 4 | Proteomic fingerprints identification and to the GO classification. (g) The enriched KEGG pathway of the significantly
quantification of Cho-PC. (a-c) Comparison of the number of PC identified changed proteins in Cho-PC compared with Nor-PC. (h) Heat map of the top
under different concentrations of cholesterol according to their functional 25 abundance proteins in Nor-PC and Cho-PC with the clustering analysis.
classification (a), isoelectric point (b), and molecular weight (c). (d-f) The (i-j) Relative abundances of the identified apolipoprotein proteins (i) and
biological function (d), cell component (e), and molecular function (f) category complement proteins ( j) in Nor-PC and Cho-PC. The data were presented as
of the significantly changed proteins in Cho-PC compared with Nor-PC according mean ± standard deviation with n = 3 (biologically independent experiments).
Extended Data Fig. 5 | Proteomic analysis of Cho-PC in human serum and (c, e) of apolipoproteins (b, c) and complement proteins (d, e) identified in neat
adsorption of LDL and HDL on NPs. (a)The heatmap of apolipoproteins and serum and protein corona on SNPs in human serum with different concentrations
complement proteins identified in PC when bare SNPs and cholesterol-precoated of cholesterol. (f-g) The amount of LDL (f) and HDL (g) adsorbed on SNPs by
SNPs were incubated with the normal FBS. Cholesterol precoated SNPs were measuring the concentrations in supernatants before and after SNPs addition.
prepared by incubating bare SNPs in 5.1 mM cholesterol solution and subsequent The data are presented as mean ± standard deviation with n = 3 (biologically
three PBS washing. (b-e) The heatmaps (b, d) and relative protein abundance independent experiments).
Extended Data Fig. 6 | Circular dichroism analysis of the interactions of cholesterol itself did not induce conformational changes, it further enhanced the
proteins with SNPs in the absence or presence of cholesterol. (a-h) Circular conformational change of APOE elicited by SNPs. The conformation of LDL and
dichroism (a, c, e, g) and secondary structure (b, d, f, h) of APOE (a-b), C1Q (c-d), HDL could be altered by both cholesterol and SNPs, and the change became more
LDL (e-f) and HDL (h-g) in the absence and presence of cholesterol or SNPs. The distinct in the presence of both cholesterol and SNPs.
conformation of APOE and C1Q changed after interacting with SNPs. Even though
Extended Data Fig. 7 | The effects of Nor-PC and Cho-PC on cytotoxicity and (c-d) Volcano plots of the protein expression change for the Nor-PC group (c) and
immune response of macrophages to NPs. (a-b) Cytotoxicity of AuNPs (a) and Nor-PC group (d) versus the control group. (e) The overlap of the significantly
SNPs (b) with various coatings on Raw264.7 cells at 24 h. The data are presented changed proteins in the Cho-PC group and Nor-PC group versus the control
as mean ± standard deviation with n = 3 (biologically independent experiments). group. (f) The enriched biological pathway of down-regulated proteins in the
Statistical significance was tested with a two-tailed, unpaired Student’s t-test. Nor-PC group and Cho-PC group by GO analysis.
Extended Data Fig. 8 | The effects of Cho-PC on cellular internalization of obtained. The data are presented as mean ± standard deviation with n = 3
NPs in hepatocytes. (a) Cell viability assessment of AuNPs, AuNPs-Nor-PC, (biologically independent experiments). Statistical significance was tested with
and AuNPs-Cho-PC on HepG2 cells. The data are presented as mean ± standard a two-tailed, unpaired Student’s t-test. (f-g) Cellular uptake of 100 µg ml−1 of bare
deviation with n = 3 (biologically independent experiments). Statistical SNPs, SNPs-Nor-PC and SNPs-Cho-PC in HepG2 cells in serum-free medium in
significance was tested with a two-tailed, unpaired Student’s t-test. (b) Cell the presence and absence of cholesterol. The average and standard deviation
viability assessment of SNPs, SNPs-Nor-PC, and SNPs-Cho-PC on HepG2 cells. represented three replicates of the median cell fluorescence intensities obtained.
The data are presented as mean ± standard deviation with n = 3 (biologically The data are presented as mean ± standard deviation with n = 3 (biologically
independent experiments). Statistical significance was tested with a two-tailed, independent experiments). Statistical significance was tested with a two-
unpaired Student’s t-test. (c) Uptake kinetics of SNPs-Nor-PC and SNPs-Cho-PC tailed, unpaired Student’s t-test. (h-m) The uptake kinetics of SNPs-Nor-PC and
in the media with or without FBS. (d-e) Cellular uptake of 100 µg ml−1 of FITC SNPs-Cho-PC in the presence or absence of 10 μg/mL chlorpromazine (h), 2.5
labelled SNPs in HepG2 cells under different concentrations of cholesterol in mg/mL methyl-β-cyclodextrin (i), 25 μg/mL dynasore ( j), 100 μM EIPA (k), 2.5
the medium with FBS quantified by flow cytometry. The average and standard μg/mL cytochalasin D (l), and 5 μM nocodazole (m). The data are presented as
deviation represented three replicates of the median cell fluorescence intensities mean ± standard deviation with n = 3 (biologically independent experiments).
Extended Data Fig. 9 | Characterization of Cho-PC in normal and hypercholesterolemia mice. (e) The heatmap of apolipoproteins and complement
hypercholesterolemia mice. (a-b) SDS-PAGE characterization of PC formed proteins identified in neat serum and protein corona on AuNPs after incubation
on AuNPs immersed in plasmas (a) and neat plasmas (b) from the normal and with the serum from normal and hypercholesterolemic mice. (f) The heatmap
hypercholesterolemia mice. (c-d) SDS-PAGE characterization of PC formed of apolipoproteins and complement proteins identified in PC when SNPs were
on AuNPs immersed in serums (c) and neat serums (d) from the normal and incubated with the serums from normal and hypercholesterolemic mice.
Extended Data Fig. 10 | In vivo biodistribution of NPs in normal and injected with Cy3-labelled SNPs at doses of 20 mg per kg body weight. The data
hypercholesterolemia mice. (a-b) The concentrations of Au in blood and are presented as mean ± standard deviation with n = 4 (biologically independent
tissues from the normal and hyperlipidemia mice at 1 h (a) and 24 h (b) after mice). Statistical significance was tested with a two-tailed, unpaired Student’s t
intravenous injection of AuNPs at doses of 0.1 g per kg body weight. The data are test. (d, f) Ex vivo fluorescence images of major organs showing the distribution
presented as mean ± standard deviation with n = 5 (biologically independent of SNPs in normal and hypercholesterolemic mice at 1 h (d) and 12 h (f). (Bl,
mice). Statistical significance was tested with a two-tailed, unpaired Student’s blood; Lu, lungs; St, stomach; Te, testis; Li, liver; Sp, spleen; Bo, bone; Br, brain;
t-test. (c, e) Tissue distribution estimated by the fluorescence intensity at 1 h Ki, kidneys; He, heart; In, intestinal).
(c) and 12 h (e) after normal and hypercholesterolemic mice were intravenously
nature portfolio | reporting summary
Corresponding author(s): Jigang Wang
Last updated by author(s): Mar 27, 2023
Reporting Summary
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.
Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested

A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient)
AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)
For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.
For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.
Software and code

Policy information about availability of computer code
Data collection 1. Transmission electron microscopy (TEM) images of nanoparticles were obtained with a HT7800 microscope (Hitachi, Japan). For cryo-TEM
images were collected using a Talos F200X G2 200 kV Cryo-S/TEM (Thermo Fischer Scientific, USA).
2. The hydrodynamic size and zeta potential profiles of nanoparticles were determined by dynamic light scattering (DLS) using a Zetasizer
Nano S90 DLS system (Malvern Instruments, UK).
3. SDS-PAGE gels were detected by Bio-Rad Gel Recording System (Azure C400 system, USA).
4. Proteomic data was collected with the Xcalibur software™ (version 4.2) integrated in an Orbitrap Fusion Lumos mass spectrometer coupled
with an Ultimate 3000 nano-LC system (Thermo Fisher Scientific, USA).
5. The equilibrium dissociation constant (KD) of SNPs with C1QB and APOE was examined via microscale thermophoresis (MST; Monolith
NT.115, Nano Temper, Germany).
6. The MD simulations were carried out with the Gromacs software package (version 2021.5) with the Charmm36 full atomic force field and
TIP3P water model in the system.
7. Flow cytometry analysis for cellular uptake was analyzed by a Backman CytoFLEX S (Backman, Germany).
8. The fluorescence images of cellular uptake in HepG2 cells were captured with a super-resolution microscope (N-SIM, Nikon).
9. The concentration of Au was determined by ICP-MS (NexlON 350X, PerkinElmer).
10. For the ex vivo fluorescent images, an in vivo imaging system (IVIS SPEOTRUM, PerkinElmer) was used.
April 2023
Data analysis 1. Statistical analysis were performed on Graphpad Prism (version 9.4.0) .
2. The intravital microscopy images and SDS-PAGE gels were analyzed with ImageJ software (version 1.8.0).
3. Mass spectra were analyzed by Proteome Discoverer software integrated with the Andromeda search engine (version 2.4, Thermo Fisher
Scientific).
1
4. The equilibrium dissociation constant (KD) of SNPs with C1QB and APOE was analyzed by MO Affinity Analysis Software (v2.3).
5. Flow cytometry data were analyzed using Flowjo software (version 10.6.2).

For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and
reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.
Data
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy
1.The collected MS files were searched against a reverse concatenated, a nonredundant variant of bovine and mice proteome database downloaded from the
Uniprot website (https://www.uniprot.org/).
2.The function of corona composition was classified using the David website (https://david.ncifcrf.gov/). Differential proteins were selected based on absolute fold
change ≥ 1.5 and p-value (FDR) < 0.05, and the biological process in gene ontology and KEGG pathway were performed to visualize functional profiles and pathway
enrichment by the “clusterProfiler” R package. The normalized expression matrix of proteins used as the input file was conducted to the GSVA analysis with default
parameters through the “GSVA” R package.
3.All proteomics raw data has been deposited to the PRIDE archive (https://www.ebi.ac.uk/pride/archive) with the identifier PXD041073.
4.The 3D-strucuture of apolipoprotein E (PDB ID: 1LE2) was obtained from RCSB Protein Data Bank (PDB) (https://www.rcsb.org/).
5. The Images in Figures 1a, 5a, 6 and Extended data Figure 3 were created by On-line software Biorender (https://www.biorender.com/).
Research involving human participants, their data, or biological material

Policy information about studies with human participants or human data. See also policy information about sex, gender (identity/presentation),
and sexual orientation and race, ethnicity and racism.
Reporting on sex and gender N.A.
Reporting on race, ethnicity, or N.A.

other socially relevant
groupings
Population characteristics N.A.
Recruitment N.A.
Ethics oversight Three human serum samples with different concentrations of cholesterol were randomly selected and analyzed, following
the approved protocols that complied with all relevant ethical regulations at the First Affiliated Hospital of Gannan Medical
University.
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.
Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf
Life sciences study design

All studies must disclose on these points even when the disclosure is negative.
Sample size No statistical method was used to predetermine the sample size. Sample size for all in vitro experiments and animal experiments was
indicated in each figure legend or Methods for every experiment and is common in the field ( Nature Nanotechnology, 2019, 14, 260-268;
Nature Nanotechnology, 2021, 14, 708-716). We used n=3 as a minimum to obtain statistically meaningful and significant results. For in vivo
studies, n=5 is sufficient to detect significant biological differences with good reproducibility based on our previous experiments (ACS Applied
Materials and Interfaces, 2016, 8, 17859−17869).
Data exclusions No samples were excluded for all analysis

April 2023
Replication For all our analyses, we performed experiments across at least 3 biological replicates, all of which were included in our analyses.All attempts
at replication were successful.
Randomization Throughout all studies, samples and animals were randomized into groups
2
Blinding Blinding was used in animal allocation, data collection and analysis. For the experiments other than those involving animals, such as

proteomics and cell experiments, the investigators were not blinded because most of the experiments contain multiple steps (protein corona
preparation, trypic digestion, electrophoresis, cell preparation and treatment, and so on) and the investigators must keep careful track of the
conditions.
Reporting for specific materials, systems and methods

We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material,
system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.
Materials & experimental systems Methods

n/a Involved in the study n/a Involved in the study
Antibodies ChIP-seq
Eukaryotic cell lines Flow cytometry
Palaeontology and archaeology MRI-based neuroimaging
Animals and other organisms
Clinical data
Dual use research of concern
Plants
Antibodies
Antibodies used The specific primary antibodies against APOE (ab183597), and C1QB (ab92508) were purchased from Abcam (UK). The secondary
antibody goat anti-rabbit IgG (ab172730) was purchased from Abcam (UK). All antibodies were diluted at a ratio of 1: 1000.
Validation We have only used antibodies from commercial sources in this study and these antibodies are utilized in a broad variety of
literature reports. The suppliers have a standard validation system to ensure reproducibility. For APOE, in Western blot, a band was
observed at 38 kDa in wild-type HepG2 cell lysates with no signal observed at this size in APOE knockout cell line. The specific primary
antibodies against APOE can react with: Mouse, Rat and Human, and is suitable for: Flow Cyt (Intra), IP, ICC/IF, IHC-P, and WB. For
C1QB, the specific primary antibodies against C1QB can react with Human, and is suitable for: IHC-P and WB.
Eukaryotic cell lines

Policy information about cell lines and Sex and Gender in Research
Cell line source(s) Raw 264.7 cells and HepG2 cells were obtained from ATCC.
Authentication Cells were bought from ATCC (Beijing, China). The supply has certified cell lines with appropriate authentication methods,
including example morphology and STR profiling. RAW 264.7 (ATCC TIB-71) is a monocyte/macrophage cell line that will have
both loosely adherent cuboidal or spindle-shaped cells and rounded viable cells. It is normal for the cells to be loosely
attached and pile and become rounded, especially when the culture is dense. HepG2 cells have an epithelial like morphology
and initially attach in small patches of cells with many clusters still in suspension. After a few days, growth will extend
outward from the adherent cell colonies. The cells were authenticated by Genetic Testing Biotechnology (Suzhou, China)
using short tandem repeat profiling analysis according to the instructions from the supply.
Mycoplasma contamination Cell lines were tested negative for mycoplasma contamination using MycoSEQ™ Mycoplasma Test Kit every 6 month.
Commonly misidentified lines No cell lines listed by ICLAC were used.

(See ICLAC register)
Animals and other research organisms

Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research, and Sex and Gender in
Research
Laboratory animals Male c57/6J mice (20-25 g, 6–8 weeks old) were provided by the Vitalriver Experimental Animal Co. (Beijing, China). The mice were
housed separately for 4 month in the standard laboratory conditions (constant temperature at 22 °C; 50%-60% ambient humidity
with 12-hour/12-hour light/dark cycle) with free access to standard rodent food and water.
April 2023
Wild animals This study did not involve wild animals.
Reporting on sex The sex of the animals was not considered in this study.
3
Field-collected samples This study did not involve samples collected from the field.

Ethics oversight All animal experiments were performed in compliance with the institutional ethics committee regulations and guidelines on animal
welfare with the approval by the China Academy of Chinese Medical Sciences (Approval Number: 2022B155).
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Flow Cytometry
Plots
Confirm that:
The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers).
All plots are contour plots with outliers or pseudocolor plots.
A numerical value for number of cells or percentage (with statistics) is provided.
Methodology
Sample preparation HepG2 cells (3×105 cells/well) were seeded in 6-well plates and incubated with various FITC labeled SNP-corona
complexes. After treatment for the pre-set times, the plates were washed three times with PBS to remove the excess
nanoparticles before the flow cytometry analysis. For investigating the effects of inhibitors on the internalization of
nanoparticles, 10 μg/mL chlorpromazine, 2.5 mg/mL methyl-β-cyclodextrin (MβCD), 100 μM dynasore, 100 μM 5-(Nethyl-N-
isopropyl) amiloride (EIPA), 2.5 μg/mL cytochalasin D and 5 μM nocodazole used to pretreat cells for 3 h
before NPs was added. As for the competition experiments, the cells in serum-free medium were pre-incubated with
BSA, LDL, and HDL (25 μg/mL) for 12 h before being exposed to SNP-corona complexes.
Instrument Backman CytoFLEX S
Software Flowjo software (version 10.6.2)
Cell population abundance Cells were not sorted.
Gating strategy Cell debris and dead cells was excluded in a forward/side scatter dot plot (FSC vs. SSC) and the gate was applied to all
samples. The median fluorescence intensity (MFI) was determined via a histogram.
Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.
April 2023

Cholesterol Modulates The Physiological Response To Nanoparticles by Changing The Composition of Protein Corona

Uploaded by

Copyright:

Available Formats

Cholesterol Modulates The Physiological Response To Nanoparticles by Changing The Composition of Protein Corona

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cholesterol Modulates The Physiological Response To Nanoparticles by Changing The Composition of Protein Corona

Uploaded by

Copyright:

Available Formats

nature nanotechnology

Cholesterol modulates the physiological

Check for updates

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1067

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1068

Serum + cholesterol Cholesterol-mediated protein corona Eluted protein corona

Serum Normal protein corona Eluted protein corona LC–MS/MS

Normalized grey value

Cholesterol level (mM) MW (kDa)

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1069

APOA2 C1QB 28 kDa

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1070

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1071

P < 0.0001 500 P = 0.0390

Cholesterol level (mM)

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1072

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1073

Fluorescence intensity (a.u.)

Relative uptake ratio (%)

MFI ratio of red/green

Content of Au (ng per 104 cells) 15 P = 0.0015

P < 0.0001 P = 0.0271

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1074

Tissue collection ICP-MS

Treatment Injection Euthanasia Euthanasia

Protein percentage in PC (%)

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1075

Cholesterol Scavenger receptor

Nanoparticles LDL receptor

Cytoplasm Apolipoprotein Caveolin

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1076

Nature Nanotechnology | Volume 18 | September 2023 | 1067–1077 1077

Bioinformatic analysis Super-resolution confocal imaging for cellular uptake

Data availability Competing interests

Code availability Supplementary information The online version

Extended Data Fig. 2 | See next page for caption.

Extended Data Fig. 4 | See next page for caption.

Extended Data Fig. 6 | See next page for caption.

Extended Data Fig. 7 | See next page for caption.

Extended Data Fig. 8 | See next page for caption.

Extended Data Fig. 10 | See next page for caption.

A description of all covariates tested

Software and code

nature portfolio | reporting summary

Research involving human participants, their data, or biological material

Reporting on race, ethnicity, or N.A.

Population characteristics N.A.

Life sciences study design

Data exclusions No samples were excluded for all analysis

nature portfolio | reporting summary

Reporting for specific materials, systems and methods

Materials & experimental systems Methods

Eukaryotic cell lines

Commonly misidentified lines No cell lines listed by ICLAC were used.

Animals and other research organisms

Wild animals This study did not involve wild animals.

nature portfolio | reporting summary

Instrument Backman CytoFLEX S

Software Flowjo software (version 10.6.2)