doi.org/10.26434/chemrxiv.11894904.v2

Triptolide Self-Assembling Nanoparticles Engineering with ModifiedErythrocyte Membranes for Targeting and Remodeling InflammatoryMicroenvironment in ArthritisJing Li, Sanpeng Li, Chunbin Li, Hongfeng Li, Chuangjun Liu, qi zhao, Pengfei Zhang, Ping Gong, Lintao Cai

Submitted date: 04/08/2020 • Posted date: 04/08/2020Licence: CC BY-NC-ND 4.0Citation information: Li, Jing; Li, Sanpeng; Li, Chunbin; Li, Hongfeng; Liu, Chuangjun; zhao, qi; et al. (2020):Triptolide Self-Assembling Nanoparticles Engineering with Modified Erythrocyte Membranes for Targetingand Remodeling Inflammatory Microenvironment in Arthritis. ChemRxiv. Preprint.https://doi.org/10.26434/chemrxiv.11894904.v2

Rheumatoid arthritis (RA) is an autoimmune disease causing severe joint damage, disability and decreasedquality of life. Pathologically, numerous blood-derived cells infiltrating in synovium and cytokine secretnecessitating formation of new blood vessels to generate pannus together form an inflammatorymicroenvironment. Triptolide with immunosuppressive activities is a potential drug to treat RA. However, it isstill lack of an effective targeting system to deliver triptolide to RA site safely. Herein an inflammatorymicroenvironment targeting and remodeling nanoplatform is developed to achieve significantly effective RAtreatment. In this system we synthesized a self-assembling triptolide nanoparticles (TPNs) mediated bydipeptide diphenylalanine which is the simplest self-assembly building block, then TPNs were entrapped bymannose-modified erythrocyte membranes to form engineering manRTPNs. For targeting, the immunologicalmolecule of erythrocytes was firstly introduced to target T cells by ligand binding of LFA-3/LFA-2, and thecoated mannose modified erythrocyte membrane also conferred the capacity of targeting to macrophages bymannose and its receptor CD206; for remodeling inflammatory microenvironment, TPNs could selectivelyexert its suppressive effects on different cells of RA including lymphocytes and synovial fibroblasts. Incollageninduced arthritis mice, manRTPNs showed excellent targeting effect and prolonged accumulation at inflamedjoint. After manRTPNs treatment, swollen paws of CIA considerably shrunk to normal, boss loss evenrecovered healthy level and cartilage preserved at synovium cavity, because of systemically conventionalcytokine reduction and expression shift of core genes in networks of RA microenvironment. Therefore, thiswell-defined manRTPNs might be a well promising systematic therapeutic agent for RA.

File list (2)

download fileview on ChemRxivmanuscript-update 20200804.docx (7.30 MiB)

download fileview on ChemRxivmanuscript-update 20200804.pdf (2.67 MiB)

Triptolide self-assembling nanoparticles engineering with

modified erythrocyte membranes for targeting and remodeling

inflammatory microenvironment in arthritis

Jing Li1, Sanpeng Li1, Chunbin Li1, Hongfeng Li1, Chuangjun Liu1, Qi Zhao3, Pengfei

Zhang1*, Ping Gong1,2*, Lintao Cai1*

1Guangdong Key Laboratory of Nanomedicine, CAS-HK Joint Lab for Biomaterials,

Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen

518055, China

2Guangdong Key Laboratory for Research and Development of Natural Drugs,

Guangdong Medical University, Dongguan, 523808, China

3 Faculty of Health Sciences, University of Macau, Taipa, Macau, China.

*Corresponding author: Lintao Cai: E-mail: lt.cai@siat.ac.cn;

Ping Gong: E-mail: ping.gong@siat.ac.cn

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

1

Abstract

Overview: Rheumatoid arthritis (RA) is an autoimmune disease characterized by persistent

synovitis, systemic inflammation and causing severe joint damage. Inflammation and influx of

immune cells is a hallmark of the disease. Triptolide with great immunosuppressive activities

is a potential drug to treat RA. However, its severe toxicity is still an unresolved bottleneck

and largely limits its clinical use. Recently, nanoparticle (NP)-based drug delivery systems

have been developed for detoxification of toxic drugs and diseases management.

Aim: To synthesize triptolide nanoparticles to decline its toxicity and improve water-solubility

and develop a dual targeted biomimicking delivery system which targets to inflammation

microenvironment. Finally, to achieve excellent therapeutic effect in rheumatoid arthritis.

Methods: We synthesized an amphiphilic molecule prodrug with the addition of

diphenylalanine peptide (FF) to triptolide. These prodrugs were self-assembled into

nanoparticles (TPNs) in water, then TPNs were coated with mannose-modified erythrocyte

membranes to form a dual targeting manRTPNs. Specifically, mannose and natural molecule

CD58 on RBC membrane guide it towards activated macrophages and T cells in inflamed

joint through innate and specific immune recognition, respectively. Different cell lines

including RAW264.7, CTLL-2, RSC364 and hFLS-RA and a collagen induced arthritis (CIA)

mice model were used to detect the targeting and therapeutic effect in vitro and vivo.

2

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

2

Results: manRTPNs showed an excellent dual targeting ability to activated macrophages and

T cells in vitro. We found that manRTPNs selectively induced anomalous activated cells death

and indeed altered genes expression and cytokines production of inflamed cells such as

macrophages, T cells and synoviocytes. In CIA mice, manRTPNs exhibited significantly

targeting effect and prolonged drug accumulation at inflamed joints, whereby increased drug

achieved significant therapeutic effect without adverse effect. In detail, manRTPNs

ameliorated joint destruction, swollen paws considerably shrunk to normal, bone and cartilage

destruction were suppressed and bone density was even recovered to healthy level. Basically,

several core genes like NF-κB, cox-2, rank, rankl, vegf, vegfr2, mmps, and cytokines such as

IL-6, TNF-α involved in inflammation, cartilage/bone destruction and angiogenesis were

regulated and altered expression and production. Thus, the inflammation microenvironment

was remodeled by TPNs.

Conclusion: This study demonstrated phenylalanine dipeptide mediating triptolide self-

assembled nanoparticles to solve current problems and the potential of engineered RBC

membranes for inflammation-targeted drug delivery. Therefore, our work is a good example

of triptolide reconstruction, and this agent based on immune recognition may provide an

immune cell targeted strategy for rheumatoid arthritis therapy.

Keywords: triptolide, nanodrug, dual targeting, inflammatory microenvironment, rheumatoid

arthritis

3

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

3

Graphical Abstract

A novel triptolide nanodrug based therapeutic agent for RA management that modified

erythrocyte membranes as a develiry system capable of dual targeted and triptolide

remodeling inflammatory microenvironment in arthritis.

4

54

55

56

57

58

59

4

Introduction

Rheumatoid arthritis (RA) is a serious long-term disease characterized by persistent

synovitis, systemic inflammation and autoantibodies and causing severe joint damage,

disability and decreased quality of life. The etiopathogenesis remains elusive, and viable and

effective treatment options are still limited[1]. Over the last two decades, long-term

management of RA has involved disease-modifying anti-rheumatic drugs (DMARDs),

biological agents such as tumor necrosis factor (TNF) inhibitors, and glucocorticoids[2].

Although the highest clinical remission rate achieved within 50% of control, not all patients

attain desirable levels of clinical remission[3] and clinical use of these therapies is limited

because of their high cost and frequency of adverse effects. The latter includes teratogenicity

and hepatotoxicity of DMARDs like methotrexate[4, 5], risks of infections such as

tuberculosis and osteoporosis caused by biological agents[6, 7] and long-term adverse

reactions to glucocorticoid therapy[8, 9]. Lately, new therapies based on inflammatory cells

such as macrophages and neutrophil or their products such as cell-free DNA in inflamed joint

have been reported[10-12], however inhibiting a single factor may not be enough to

sustainably halt or reverse disease progression due to RA’s complexity and heterogeneity.

Therefore, a new systemic design for drug candidate based on nano-medicine is urgently

needed.

Triptolide is a trace natural triepoxide diterpene from Tripterygium wilfordii (or Léi

Gōng Téng) and has been widely used to treat for RA patients since the 1970s in China[13].

Despite its various biological activities, low water solubility and severe toxicity, especially to

5

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

5

the liver and kidney, remarkably impedes the translation to clinical applications[14].

Considerable efforts in structural modification[15-18] and nanoparticle(NP) -based drug

delivery system [19-21] have been applied to reduce its toxicity. Unfortunately, common

practices that direct conjugations attached to the C14-hydroxyl group of triptolide via a

hydrolytic ester bond have proven biologically unstable; meanwhile, current targeted drug-

loading delivery systems are limited by drug loading or encapsulation ratio and they mostly

target to kidney or tumor[22]. Therefore, an effective and stable system that can deliver

triptolide to other organ or tissue such as joint is necessary. But until now, triptolide-formed

nanoparticles have been rarely reported before. From this point, inspired from

biomacromolecule self-assembly process, various peptide building blocks have been

performed for the creation of biomimetic or bioinspired nanostructured materials[23, 24].

Recently, diphenylalanine (FF) is the simplest peptide building block for self-assembly and

exhibits remarkable advantages including biocompatibility and non-immunogenicity and so

on. Many researchers have developed FF-based nanomaterials such as nanotubes, spherical

vesicles, nanofibrils and nanowires[25]. However, it is barely to link it with toxic drug for

nanodrug formation. Therefore, diphenylalanine (Phe-Phe, FF) mediating self-assembling

nanoparticles is a new strategy to reconstruct triptolide to be nanoparticles whose spherical

nanostructure was more stable and biocompatible than before.

In RA pathogenesis, numerous blood-derived cells infiltrate in synovium and cytokine

secret necessitates formation of new blood vessels to generate pannus, that all of these form an

inflammatory microenvironment[26]. Peripheral-blood monocytes are recruited to joint, and

functionally diverse macrophages distribute in the synovial sublining and lining layers. It is6

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

6

mentioned that Cluster of Differentiation 206 (CD206) on macrophages surface has been

induced to express more [27]. Besides, prominent T cells being activated, especially the CD4+

subset, also infiltrate the synovium. As reported, certain signal proteins on the T cell surface

such as lymphocyte function antigen 2 (LFA-2/CD2) have increased in both RA patients and

animal models[28, 29]. It was reported that CD2 can be effectively targeted by peptides

derived from LFA-3/CD58 as well as fusion protein constructs. Interestingly, the ligand of

CD2 namely CD58 widely distributes on erythrocyte membranes and CD58-CD2 interaction

have mediated process of T cells activation[30]. These immunological characteristics of

erythrocytes scarcely ever attracts attention in application, so it is a valuable attempt to apply

in drug targeting. Then, mannose, the ligand of CD206 was introduced in erythrocyte

membrane for ligand-binding targeting. In addition, the long circulation and flexible

functionalization of erythrocyte membrane also make it be a promising platform and have

wide application[31, 32]. In this study, we prepared mannose-labeled RBC membrane to

entrap newly synthesized triptolide nanoparticles (TPNs) for elevating its vast retention in

inflammation and achieving remarkably anti-inflammatory therapy. Triptolide, a hydrophobic

drug, was linked with diphenylalanine peptide to self-assembly into size-controlled

nanoparticles (TPNs) in water, and this reconstruction has successfully decreased its toxicity

and improved water-solubility. As shown in scheme 1, the nanodrug core was coated with

decorated RBC membranes to obtain mannose-modified RBC cell membrane-coated triptolide

nanoparticles (manRTPNs). By translocating RBC membranes to the nanoparticles, the

biological targeting properties of RBC cell can be conferred to manRTPNs as well, allowing

them to target activated T cells through the CD58-CD2 affinity. Further, the binding

7

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

7

interaction between the mannose and its receptor CD206 enhanced anchoring of manRTPNs

to activated macrophages. According to manRTPNs, the dual-targeting strategy facilitated the

accumulation of nanodrug in inflammation, and the accumulated TPNs efficiently achieved

systemic administration of rheumatoid arthritis by specifically inducing cell death and

remodeling the inflammatory microenvironment through affecting genes network,

downregulating cytokines and pro-angiogenic factors, regulating osteoclastogenesis, and

returning balance of metalloproteinases and their inhibitors. Overall, our study expanded the

RBC cell membrane biological application in nanocarrier for targeting inflammation through

innate and specific immune recognition, and the triptolide nanodrug showed high drug

content, low side effect, good biocompatibility which showed great potential for further

inflammation diseases therapy.

Results and Discussion

Preparation and characterization of manRTPNs

Triptolide is a diterpene triepoxide with several active sites that can be modified,

especially C14 position[33]. Because of its hydrophobicity, triptolide molecule acts as self-

assembly inducer with the addition of diphenylalanine peptide (FF), which is the key moiety

of the self-assembling peptide-drug conjugate. Initially, we added a glucosamine (AG) to the

FF terminal to lengthen the hydrophilic chain, and then linked triptolide to FF-AG by an ester

bond whose cleavage led to drug release in the cell; thus, the self-assembling monomer (TP-

FF-AG) as a triptolide prodrug was synthesized (Figure 1A). The chemical structures of these

compounds were confirmed via mass spectrometry (MS) and nuclear magnetic resonance

8

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

8

(NMR; Figure S1-S16). Due to hydrophobic interactions, triptolide nanoparticles (TPNs)

were formed via self-assembly of synthesized monomer namely TP-FF-AG, which is

uniformly easy to control TPNs size in water solution. The particles morphology was

observed via transmission electron microscopy (TEM; Figure 1B). To synthesize manRTPNs,

RBC membranes were extracted by hypotonic hemolysis[34], and mannose inserted by

PEG2000-DSPE. An extruder was using to camouflage mannose-modified erythrocyte

membranes onto TPNs then form a core-shell structure (Figure 1B). Dynamic light scattering

(DLS) measurements revealed that the TPNs were 135 nm in hydrodynamic diameter and

manRTPNs were about 10 nm larger with PDIs of 0.15-0.2. The surface zeta potential of the

shell was not as negative as that of the cores but was comparable to that of erythrocyte

membrane-derived vesicles (RBC vesicles) (Figure 1C). All these results demonstrated

successful cloaking of TPNs with erythrocyte membranes.

Next, TPNs and manRTPNs were stored in phosphate-buffered saline (PBS) for 1 month,

where they demonstrated a negligible change in size and showed an excellent stability (Figure

1D). Maintaining protein compounds of erythrocyte membranes on nanoparticles’ surfaces is

vital for harnessing their biofunctionality. Therefore, completeness of the membrane coating

was verified via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE;

Figure S17), which indicated the manRTPNs maintained the whole effective biocomponent of

erythrocyte on the surface. Translocation of erythrocyte membranes might improve the

biocompatibility and circulating period of TPNs in vivo, which also sustains systemic delivery

and better targeting through both passive and active uptakes[35]. As reported, the elimination

T1/2 of triptolide molecules in rats following i.v. administration was 15-21 minutes, which

9

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

9

showed that drugs were rapidly eliminated[36]. However, TPNs and manRTPNs have

maintained higher plasma concentration after tail intravenous injection of mice in our work.

As the relative signal of drug down to 50%, TPNs lasted 4 h, and manRTPNs even had a

much longer elimination half-life of over 48 hrs (Figure 1E), which suggested that manRTPNs

significantly expanded blood circle compared to triptolide. Therefore, the self-assembling

TPNs nanostructure was more stable than original drug, and manRTPNs inherited biological

properties of erythrocyte membrane proteins for following efficacy.

Dual-Targeting capacity and immunosuppressive effect of manRTPNs

To determine each targeting capacity of manRTPNs to macrophage and T cell in vitro,

nanoparticles were Cy5.5-labelled and incubated with RAW264.7 and CTLL-2 treated with

IL-4 (100 ng/mL) or TNF-α (10 ng/mL), untreated cells and RSC364 as control. Flow

cytometry (FCM) analysis observed gradual increased uptake of nanoparticles by RAW264.7

and CTLL-2 for 6 hrs (Figure S18-20). An uptake curve was built using mean fluorescence

intensity (MFI) of cells at six time points, and cellular uptake rate was defined as k indicated

in figure 2A-C, which was the slope of the uptake curve. In detail, ks of nanoparticles in

activated RAW264.7 and CTLL-2 were significantly higher than that in untreated cells

(Figure 2A-C; Figure S21A & S22), and ks of manRTPNs were significantly higher than that

to TPNs both in RAW264.7 and CTLL-2 (Figure 2A-B), but no difference of ks existed

between TPNs and manRTPNs in RSC364 that no targeting ligands on it (Figure 2C). Further,

the k value in RAW264.7 was higher than that in CTLL-2 (Figure 2A-B), indicating the

affinity difference of interacted molecule pairs. To further confirm ligand-binding specificity,

10

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

10

Cy5.5-labelled TPNs (without mannose and CD58), RTPNs (with CD58) and manRTPNs

(with mannose and CD58) were incubated with treated macrophages and T cells for 3 hrs as

well, RSC364 and hFLS-RA as control. After treatment, CLSM images showed that CD206

and CD2 (Figure 2D-E, green) respectively overexpressed on surface of RAW264.7 and

CTLL-2. Significant red fluorescence was observed on RAW264.7 incubated with

manRTPNs, but not with TPNs and RTPNs (Figure 2D), while red fluorescence was observed

on CTLL-2 incubated with RTPNs and manRTPNs but not with TPNs (Figure 2E), which was

consistent with corresponding ligands distribution on the nanoparticles. When co-incubated

with RSC364, they also entered cells (Figure 2F; Figure S21b). Similar results were observed

in flow cytometry analysis (Figure 2G-I; Figure S21c). These results demonstrated the ability

of manRTPNs to target inflamed cells conferred by their mannose modified erythrocyte

membrane coating, which is probably attributed to specific affinity of mannose-CD206 and

CD58-CD2.

Except for migrated immune cells, synovial fibroblast is main resident effector of

inflammation. To determine immunosuppressive efficacy of manRTPNs in joint inflammatory

cells, LPS-treated RAW264.7, TNF-α-treated CTLL-2 and RSC364, and hFLS-RA were

incubated with manRTPNs in 96-well plates for 24 hrs. Firstly, cell viability was measured via

MTT assay. TPNs and manRTPNs didn’t significantly reduce RAW264.7 viability even at a

concentration of 500 nM (Figure 3A). However, at the same concentration CTLL-2, RSC364

and hFLS-RA showed an obvious cytotoxicity (Figure 3B-D). Specially, manRTPNs had an

IC50 (half maximal inhibitory concentration) values of 219 nM for activated CTLL-2

cytoxicity, 88.9 nM for activated RSC364 cytoxicity and 95.8 nM for hFLS-RA cytoxicity

11

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

11

(Figure 3B-D). The lower IC50 values of activated RSC364 and hFLS-RA indicated that

triptolide-induced cell death was mainly against anomalous proliferative synoviocytes in

inflamed joints, rather than lymphocytes like macrophages. Cells were following collected

and lysed for mRNA extraction to check certain inflammation associated genes variation by

qPCR. For LPS-treated RAW264.7, transcription factor NF-κB associated with most cytokine

transcription[37] and cyclooxygenase-2 (cox-2) were induced to decrease expression level, but

tissue inhibitor of metalloproteinase-1 (timp-1) showed opposite variation under manRTPNs

treatment. Receptor activator of NF-κB ligand (rankl) over-expressed on TNF-α-treated

CTLL-2 was also downregulated by manRTPNs. After activated with cytokines and

chemokines, synovial fibroblasts release matrix-degrading enzymes, including matrix

metalloproteinases (MMPs) and cathepsins causing cartilage and bone destruction[38]. In

RSC364, the transcripts of metalloproteases such as MMP-2 and MMP-9 and vascular

endothelial growth factor (VEGF) being critical for angiogenesis were lessen by manRTPNs

as well (Figure 3E). Thirdly, medium supernatants were collected to measured cytokines by

ELISA assay. supernatant IL-1β and IL-6 of LPS-treated RAW264.7 and supernatant IFN-� of

TNF-α-treated CTLL-2 lessened more quickly that treated with manRTPNs than TPNs

(Figure 3F-H), while supernatant IL-1β and IL-6 of TNF-α-treated RSC364 had a similar

decreasing curve between TPNs and manRTPNs groups (Figure 3I-J), which indicated that

manRTPNs had superior identification of immune cells than resident synoviocytes at least in

cytokine secretion. Consequently, manRTPNs selectively exerts its immunosuppressive

effects on selectively inducing lymphocytes and synovial fibroblasts cell death, but also

regulating production of proinflammatory cytokines, proinflammatory mediators and matrix

12

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

12

metalloproteinases, which suggested manRTPNs may have an ability to remodel

inflammatory microenvironment through its cell-specific anti-inflammation efficacy.

Targeting capacity and accumulation of manRTPNs in vivo

Upon the good capacity of manRTPNs to target activated macrophages and T cells in

vitro, we further verified their ability to target RA site in vivo. After CIA mouse model being

induced, all four paws swelled in varying degrees[39]. Hind paws were more heavily swollen,

and micro-computed tomography (Micro-CT) imaging showed distinct bone loss of joint

therein (Figure 4A). Then these mice received intravenous injections of free cyanine 5.5

(Cy5.5) solution and Cy5.5-labelled TPNs and manRTPNs for targeting imaging in vivo. As

shown in figure 4A, almost no fluorescence signal was detected in paws of the free-Cy5.5

group, indicating that the dye barely accumulated at local inflammatory sites. In the TPNs

group, we detected a weak signal, which peaked after 4 h (Figure 4A). While manRTPNs

obviously reached the RA sites and lasted there for ≥12 hrs, which would be attributed to their

long circulation and the targeting effect of the modified erythrocyte membranes. Thus,

significant fluorescence of manRTPNs lasted longer and showed greater accumulation than

that of TPNs. 24 hrs after intravenous injections, we dissected mouse and got froze sections of

left paws to observe nanoparticles’ retention in the inflamed joints. Contrary to free-Cy5.5 and

TPNs group, fluorescence microscopy indicated a very high fluorescence intensity of

manRTPNs in the synovial sublining layer, resulting in a reduced accumulation at the

inflamed joint due to the lack of permeability (Figure 4B). Tissue biodistribution further

revealed manRTPNs’ outstanding capacity to target the paws of RA mice, rather than other

13

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

13

groups (Figure 4C). Additionally, higher accumulation was observed in liver, suggesting that

this organ might be passive targeting of nanoparticles and the main organ metabolizing

triptolide (Figure 4C). The results confirmed that the effective accumulation of manRTPNs

provided a promising precondition of RA treatment.

Therapeutic efficacy of manRTPNs

The efficacy of ameliorating joint damage of manRTPNs was evaluated in a CIA mouse

model. Following arthritis induction, CIA mice presented symptoms of swelling in ankle and

paw joints and a clinical scoring was used to estimate the swelling degree (Figure S23).

According to protocol timeline, CIA mice were treated with intravenous injections of

triptolide at 15 mg/kg, TPNs and manRTPNs at doses of 25 mg/kg for 11 days (Figure 5A).

Surprisingly, swollen hindpaws and forepaws both shrunk considerably and even recovered to

the same extent as healthy ones with manRTPNs treatment rather than slightly shrunk in

TPNs and triptolide group, but swelling symptoms of other groups were increasingly severe

(Figure 5B), implying that inflammation process was exactly controlled by manRTPNs. As

paw swelling is only a surface phenomenon of RA, bone loss was further determined using

Micro-CT analysis. Reconstructed three dimensional (3D) bones images and calculated bone

mean density (BMD) showed that serious bone erosion was suppressed and bone density even

recovered to healthy mice level of manRTPNs treatment (Figure 5C-D), suggesting a

protective role of manRTPNs on cartilage and bone damage. Haemotoxylin and eosin (H&E)-

stained sections showed an intense anomalous proliferative cells infiltration in the joints and

synovium of control groups. By contrast, remarkable reduction of immune infiltration and

14

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

14

preservation of cartilage were observed in manRTPNs group (Figure 5E). Further, cytokines

including TNF-α and IL-6 are known to increase with the onset of arthritis and correlate

strongly with disease severity[40]. Immunohistochemical analysis showed that TNF-α and IL-

6 in synovium and articular cavity significantly declined in manRTPNs groups (Figure 7A-B)

and serum levels of them also observed to alleviate similarly (Figure 6C), indicating effective

reduction of arthritis at cytokines level. Moreover, gene heatmap showed that manRTPNs

suppressed several genes expression including rank, rankl, vegf, vegfr2, NF-κB, cox-2, IL-1β,

mmp-2, mmp-3, mmp-9 and mmp-13, that these transcripts are dramatically increased and

contribute in joints damage. Some negative regulators such as opg, timp1 and timp2 were

upregulated their gene expression by manRTPNs (Figure 6D), indicating manRTPNs preserve

appropriate and balanced concentration of cytokines and inflammatory factors through two-

way regulation. RANKL/RANK/ORG signal pathway is key of bone metabolism[41]. The

up-/down-regulation of manRTPNs also recovered the balance of RANKL/RANK/ORG and

metalloproteinases and their inhibitors (TIMP) to normal. Thus, manRTPNs ameliorated

destroyed through aiming to RA pathological factors including cytokines, extracellular matrix

(ECM), bone metabolism and angiogenesis.

Biosafety of manRTPNs nanoparticles

To investigate safety and biocompatibility of TPNs and manRTPNs in vivo, we

administered PBS, TPNs or manRTPNs at a single dose of 25 mg/kg to healthy mice via tail

vein, while to other mice we administered triptolide (original drug of TPNs and manRTPNs)

at 15 mg/kg by intraperitoneal injection. It is worth mentioning that most mice in the

15

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

15

triptolide group died soon (within 24 hrs) after 25 mg/kg administration, indicating the high

toxicity of the original drug (Figure S24). After 1-week treatment, mice treated with triptolide

showed a significant decrease in count of white blood cells (WBC), RBCs and platelets

(Figure 7A), suggesting the occurrence of bone marrow suppression, which is one of the main

side effects of triptolide. The main three type cells comprised WBC including neutrophils,

lymphocytes, and monocytes showed significant count differences between triptolide group

and other groups, which explained the WBC count variation conferred from triptolide (Figure

7B). The hemoglobin concentration did not differ significantly among all the groups in the

blood (HGB), mean corpuscular hemoglobin (MCHC) level, and hemoglobin distribution

width (HDW) (Figure 7C), indicating that triptolide didn’t influence the physiological

function of RBCs. Blood biochemistry analysis showed that blood urea nitrogen (BUN),

aspartate aminotransferase (AST), creatinine (Cr) and lactate dehydrogenase (LDH) levels of

mice treated with triptolide were significantly higher than those of mice in the TPNs and

manRTPNs groups (Figure 7D-G). These variations demonstrated that the hepatotoxicity and

nephrotoxicity of triptolide were greatly reduced in TPNs, with RBC membrane coating

yielding even better safety. Hematoxylin and eosin (H&E) staining of major organs showed

that triptolide treatment caused inflammatory infiltration of the liver and severe renal injury

characterized by loss of the brush border, tubular epithelial cells detachment from the

basement membrane and tubular obstruction. Meanwhile, the TPNs and manRTPNs groups

showed no significant variation in cellular morphology and negligible organ damage, in line

with the PBS group (Figure 7H). However, TPNs and manRTPNs significant reduced

triptolide toxicity.

16

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

16

Conclusion

Our study expanded the RBC cell membrane biological application in nanocarrier for

targeting inflammation through innate and specific immune recognition, and the triptolide

nanodrug showed high drug content, low side effect, good biocompatibility which showed

great potential for further inflammation diseases therapy. Generally, there are two

improvement of manRTPNs in cell-membrane-coated nanoparticle delivery system, namely

triptolide nanodrug core and RBC membrane shell. RBC membrane-camouflaging

nanoparticles has been extensive established platform, but there are still many deposits from

RBC cell itself to further develop disease-targeting and therapy platform except for tumor.

The natural immunological function of RBC cells has been always neglected in targeted drug

delivery construction. So, taking advantage of adhesion molecules affinity between RBC cells

and immunocytes in building nanocarrier for transporting drug to inflammation deeply

expands application of established RBC membrane platform. On the other side, triptolide has

excellent anti-inflammation and immunosuppressive effect and solving its low water-

solubility and toxicity is premise of using it in various diseases management. Our

reconstruction of triptolide isn’t content-limited loading in micelles, but drug-based self-

assembled size-controlled nanoparticles. This strategy takes advantages of drugs’

hydrophobicity to solve its bad water-solubility, which is magic of nanotechnology. While

inflammation is prevalent related with many diseases like cardiovascular diseases,

gastrointestinal tract, cancer and pathogen infection[42]. Therefore, manRTPNs as a drug

delivery system of inflammation targeting and treating may be a promising agent in other

inflammation-associated diseases therapy.

17

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

17

Materials and methods

Animal care

Mice were housed in an animal facility at Shenzhen Institutes of Advanced Technology,

Chinese Academy of Sciences (SIAT). All animal experiments were performed in accordance

with Guidance Suggestions for the Care and Use of Laboratory Animals and approved by

Sciences Animal Care and Use Committee of SIAT.

Erythrocyte membranes derivation

Erythrocyte membranes were collected according to a previously reported method with

modification. Briefly, fresh heparinized whole blood was collected from male mice (20-22 g)

and subsequently centrifuged at 2000 rpm for 15 min at 4 °C to remove the plasma and the

white buffy coat. The collected red blood cells (RBCs) were washed with 1× PBS three times

and suspended in 0.25× PBS for 2 hours at 4 °C, and then the hemoglobin was removed by

centrifugation at 9,000 rpm for 45 min. The resulting pink pellet was purified with 1× PBS,

and the collected erythrocyte membrane was suspended and stored in distilled water. To form

ligand-inserted erythrocyte ghosts, the light pink solution was incubated with DSPE-PEG-

mannose for 30 min, 37°C water bath. Erythrocyte membranes with PEG-mannose or not was

further prepared by extruding through a 200 nm polycarbonate porous membrane with an

Avanti mini-extruder (Avanti Polar Lipids).

Synthesis of TPNs

The TPNs were synthesized by a prodrug self-assembly method. Briefly, after dissolving

5mmol Boc-Phe-Phe-OH (BOC-FF, J&K scientific Ltd., China), 6 mmol 1-Ethyl-3-(3-

18

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

18

dimethylaminopropyl) carbodiimide (EDC, J&K scientific Ltd., China) and 6 mmol N-

Hydroxysuccinimide (NHS, J&K scientific Ltd.,China) in 3 mL anhydrous DMF, another

anhydrous DMF solution containing 5.5 mmol glucosamine (AG, Sigma-Aldrich, USA) and 5

mmol TEA (J&K scientific Ltd., China) was added drop-wise into it. The mixture was

evaporated under reduced pressure for 24 hours to get crude product BOC-FF-AG. Then,

crude BOC-FF-AG was dissolved in 30 mL equal volume mixture of anhydrous

dichloromethane and trifluoroacetic acid (TFA, J&K scientific Ltd., China) and stirred for 4

hours at room temperature. Afterward, pure NH2-FF-AG was obtained after evaporated under

reduced pressure and purified by silica gel column chromatography with dichloromethane and

methanol. Secondly, 1 mmol triptolide (TP, Shanghai Yuanye Bio-Technology Co., Ltd.,

China), 1.2 mmol succinic anhydride (SAA, Sigma-Aldrich, USA) and 0.2 mmol dimethyl

aminopyridine (DMAP, Sigma-Aldrich, USA) were dissolved in 3 mL pyridine and stirred

overnight. After then, the mixture was diluted with ethyl acetate and washed it with solution

containing saturated copper sulfate (J&K scientific Ltd.), water and brine. After reaction,

residue TP-COOH was obtained through drying with anhydrous Na2SO4 and concentration

with rotary evaporator. In the third step, TP-COOH, dicyclohexylcarbodiimide (DCC, Sigma-

Aldrich, USA) and 1-hydroxy-5-pyrrolidinedione (NHS, J&K scientific Ltd., China) of equal

molar ratio were dissolved into 3 mL anhydrous dichloromethane and stirred at room

temperature for overnight. The final product TP-FF-AG was gained by reaction of activated

TP-COOH and NH2-FF-AG for overnight at room temperature and purified by analytical RP-

HPLC and semi-preparative RP-HPLC with simultaneous detection at 265 nm. Analytical RP-

HPLC was performed at room temperature on the Shimadzu LC 20 with UV detector SPD-

19

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

19

20A using Inertsil ODS-SP column (4.6 x 250 mm, 5 µm, 100Å). The RP-HPLC gradient was

started at 10% of B (CH3CN, J&K scientific Ltd., China), then increased to 100% of B over

30 min (A: 0.1% TFA in water). Semi-preparative RP-HPLC was performed on the ULTIMAT

3000 Instrument (DIONEX). For fluorescence imaging experiments, cyanine5.5 (Cy5.5,

Lumiprobe, USA) was linked to intermediate product of NH2-FF-AG after esterification

reaction to obtain Cy5.5-FF-AG in the same way. Ultimately, nanoparticles were self-

assembled by added TP/Cy5.5-FF-AG dissolved in DMSO dropwise into PBS while vigorous

stirring and dialysis for 2 days. 1H NMR and 13C NMR spectra were recorded with a Bruker

VANCE III400 spectrometer (400 MHz). The high-resolution mass spectra (HR-MS) were

measured on a Bruker Micro TOF II 10257 instrument. For membrane coating, nanoparticles

were mixed with prepared membranes in ratio of 1:1. The resultant mixture was subsequently

extruded nine times through a 200 nm polycarbonate porous membrane using an Avanti mini-

extruder to yield the erythrocyte membranes-coating TPNs or mannose inserted erythrocyte

membrane-coating TPNs (namely RTPNs or manRTPNs).

Nanoparticles Characterization

The hydrodynamic diameter and zeta potential of NPs suspended in 1 × PBS were

measured by dynamic light scattering (DLS) (Zeta Plus, Brookhaven Instruments, USA). The

morphologies of TPNs and manRTPNs were observed by transmission electron microscope

(TEM, FEI spirit T12) at an accelerating voltage of 120 keV. The stability experiment of

TPNs and manRTPNs in PBS was monitored by DLS over five weeks and stored at 4°C. The

proteins retained on manRTPNs compared with those on the natural erythrocyte membrane

20

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

20

were observed by SDS-PAGE. Briefly, 2 mg of erythrocyte membrane vesicles and

manRTPNs were collected by centrifugation at 15,000g for 30 min, mixed in SDS sample

buffer (Invitrogen, USA), and heated at 90 °C for 5 min. Then, 20 μL of each sample was run

on a 10% SDS-polyacrylamide gel (Bio-Rad, SDS-PAGE Gel Preparation Kit, USA) at 120 V

for 1 h, followed by Coomassie blue staining and imaging.

Pharmacokinetics

To evaluate the circulation half-life of TPNs and manRTPNs in vivo, 150 μL Cy5.5-

labeled nanoparticles were injected into the tail vein of the mice (n = 3). Twenty microliters of

blood were collected at 1, 5, 15, 30 min, and 1, 2, 4, 8, 24, 48, and 72 h following the

injection. Each particle group contained 3-4 mice. The collected blood samples were diluted

with 30 μL PBS in a 96-well plate before fluorescence measurement. Pharmacokinetics

parameters were calculated to fit a two-compartment model and a one-way nonlinear model.

Drug toxicity analysis in vivo

Purchased from Vital River Laboratory Animal Technology Co. Ltd (Beijing, China),

males BALB/c mice (6∼8 weeks old) were randomly divided into four groups and

intravenously injected with PBS, TPNs and manRTPNs at a single dosage of 25 mg/kg and

triptolide of 15 mg/kg 4 times for eight days. A total of 7 days after last administration, all

mice were euthanized, and blood was collected for and blood routine examination and

biochemical parameters measurement, while main organs including heart, liver, spleen, lung,

and kidney were excised and stained by H&E for histological analysis.

Cell culture

21

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

21

RAW264.7 cells, CTLL-2, RSC364 and hFLS-RA were purchased from the American

Type Culture Collection (ATCC, USA) and cultured in Dulbecco’s minimum essential

medium (DMEM, Corning, USA) supplemented with 10% fetal bovine serum (FBS, Gibco,

USA), 1% penicillin (100 IU/mL, Corning, USA), and streptomycin (100 μg/mL, Corning,

USA) and placed at 37 °C in a 5% CO2 humidified atmosphere. Cell lines were purchased

with certification of authentication and free from Mycoplasma.

Quantification of cell targeting in vitro

A 6-well culture plate was prepared and RAW264.7, CTLL-2, RSC364 and hFLS-RA

cells were implanted at a density of 5 × 105 cells/well. 12-16 hrs later, 80% of cell coverage

was confirmed. 50 �L Cy5.5-labelled TPNs or manRTPNs was adding into prepared cells

every hour for six hours, incubated with covering foil paper and collected them in the sixth

hour. After digestion with trypsin (Gibco, USA), cells were put together and washed three

times with ice-cold PBS, then underwent flow cytometry by FACS Canto II (BD

Biosciences). The data was analyzed using FlowJo software. Mean fluorescence intensity

(MFI) reflected amount of intake nanoparticles amount. An uptake curve was constructed

with MFI and time, and a k value being the slope of uptake curve was defined as uptake rate

to measure targeting capacity of nanoparticles.

Immunofluorescence assay of macrophages and T cells with nanoparticles

Cells were cultured with medium added recombinant mouse IL-4 (100 ng/mL, R&D

System, USA) or recombinant TNF-α (10 ng/mL, R&D System, USA) for six hours and

incubated with TPNs, RTPNs or manRTPNs for 3 hours. Afterward, cells were fixed with 4%

22

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

22

paraformaldehyde (Sigma-Aldrich, USA) for 10 minutes, permeabilized with 0.1% Triton™

X-100 (Sigma-Aldrich, USA) for 10 minutes, blocked with 1% BSA for 1 hour and labeled

with 2 µg/mL anti-mannose receptor/AF488 conjugated antibody (ab195191, Abcam) and

rabbit anti-CD2/AF488 conjugated antibody (bs-2899R, Bioss) for 3 hours at room

temperature. Nuclei were stained with DAPI. All manipulation must be away from light. The

images were captured under confocal microscopy (Leica, TCS SP5).

Cellular cytotoxicity study

The cellular toxicities of TPNs and manRTPNs on macrophages, T cells, FLSs and

hFLSs were determined by MTT assay using CCK-8 Kit. Cells were implanted in a 96-well

culture plate at a density of 8 × 103 cells/well (5 replicates each treated group). After 12 hrs,

PBS, TPNs and manRTPNs with concentrations of 10 nM, 50 nM, 100 nM, 200 nM, 500 nM

were added into medium for further 24 hrs co-culture. Then the cells were mildly washed by

PBS before added CCK-8 mixed solution. Then, the optical density in 450 nm was measured

by Multiskan GO (Thermo Fisher Scientific).

Cytokines concentration analysis

The TNF-α, IL-6 and IFN-� concentrations of cells culture supernatants were determined

with ELISA Kit using mouse ELISA Kits (Biolegend, USA) as manipulation. Also, the TNF-α

and IL-6 concentrations of mouse serum samples were determined with TNF-α and IL-6

ELISA Kits (Biolegend, USA). Briefly, Serum samples of treated mice were collected at the

end of experiment and concentrations of TNF-α and IL-6 were quantified with ELISA kits.

Specifically, the whole blood of mice was collected by submandibular bleeding into

23

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

23

microtubes and allowed to clot at room temperature for 30 min. Samples were then

centrifuged at 2,000g for 6 min to collect supernatant serum. Serum samples were

immediately frozen at -20 °C until analysis by using mouse TNF-α and mouse IL-6 ELISA

kits (Biolegend) within 3 days of collection.

Genes expression profiles analysis of cells

After treated with nanoparticles of serious concentrations, cells were harvested to isolate

RNA. RNA was extracted with TRIzol reagent (Invitrogen, USA) according to the

manufacturer’s instructions. Complementary DNAs (cDNAs) were synthesized with

QuantiTect Reverse Transcription Kit (QIAGEN K.K., Tokyo, Japan) and the specific gene

transcripts were quantified by quantitative real-time PCR using QuantiTect SYBR Green PCR

Kit (QIAGEN K.K., Tokyo, Japan) and analyzed with ABI 7500 real-time PCR system

(Applied Biosystems, USA). The gene expression patterns along with concentration were

highlight with heatmap by using TBtools package.

Collagen induced arthritis (CIA) mice model induction

Seventy-two female DBA/1 mice (6∼8 weeks old) were purchased from Vital River

Laboratory Animal Technology Co. Ltd (Beijing, China). All mice were maintained in a SF

system which is room equipped with an air-filtering system, and the cages and water were

sterilized. CIA was induced as previously reported study[39]. Briefly, bovine type II collagen

(Chondrex, Redmond, WA, USA) was dissolved in 0.1M acetic acid overnight at 4°C. This

was emulsified in an equal volume of complete Freund’s adjuvant (Chondrex, Redmond, WA,

USA). The mice were immunized intradermal at the base of the tail with 100 �L of emulsion

24

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

24

containing 100 �g of type II collagen. On day 21, mice were boosted intraperitoneal with 100

�g type II collagen dissolved in phosphate buffered saline (PBS). Clinical scores of each

hindpaws and forepaws of mice were obtained following the standard evaluation process[39].

In detail, no evidence of erythema and swelling occurred is score 0, erythema and mild

swelling appeared is score 1, erythema and mild swelling extended from the ankle to the

tarsals is score 2, erythema and moderate swelling extended from the ankle to metatarsal

joints is score 3 and erythema and severe swelling encompassed the ankle, paws, and digits or

ankyloses of the limb is score 4. Then total clinical scores of hindpaws and forepaws were

calculated. At least six mice each group.

Nanoparticles targeting to inflamed joints

After CIA model was established, free Cy5.5, Cy5.5-labelled TPNs and manRTPNs were

intravenously injected into the CIA mice (20 mg/kg, n = 4). Fluorescence images were

obtained after 1, 2, 4, 8, 12, and 24 hours using an in vivo imaging system (IVIS, Caliper,

USA). Then, the mice were sacrificed for ex vivo tissue/organs distribution analysis at the end

of experiment. Meanwhile, the synovial tissue was dissected and dehydrated in 30% sucrose

solution overnight and then cry-sectioned at 8 �m. The sections were stained with DAPI and

washed twice. Fluorescence images were obtained using a confocal microscope (Leica, TCS

SP5). The fluorescence intensity of sections in different groups was calculated using software

Image J.

manRTPNs treatment assay

25

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

25

According to CIA mice model conduction, starting at day 21, the clinical scores were

calculated to make sure that CIA mice were almost equably and randomly divided into four

groups, while healthy DBA/1 mice were set as control (n = 5). Five different agents namely

PBS, RBC vesicle, triptolide, TPNs and manRTPNs were intravenously injected every two

days. The timeline of whole therapy experiment from model making to the end was according

to schematic diagram shown in figure 5A. The arthritis index of each group was recorded over

time to evaluate treatment efficacy. At the end, the mice were sacrificed for further analysis.

Micro-CT imaging of mice paws

The knee and ankle joints of each experimental mouse hindpaws were scanned using

micro-CT system (Bruker's SkyScan 1278, UK) to observe joint bone situation. Images were

acquired at 55 kV, 72 μA and 300 ms/frame, with 360 views. The 3D structures of knee joint

and ankle joint were reconstructed from 360 views and BDM was evaluated by corresponding

SkyScan NRecon package. Analysis typically took around 15 min per joint for an experienced

operator.

Histological analysis and immunohistochemical staining assay

At study endpoints, mice were euthanized and hind knee joints were collected for H&E

staining and immunohistochemical staining. Briefly, the ankle and knee joints of scarified

animals as well as their heart, liver, spleen, lung and kidney were fixed in 10% buffered

formalin and then joints were incubated in decalcifying solution (4% hydrochloric acid in 4%

formaldehyde) at room temperature for 7 days for decalcification. After paraffinization,

microtome (Leica) slices of 8 μm were prepared and stained with haematoxylin and eosin,

26

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

26

images were taken by using inverted microscope (Olympus, IX71, Japan) and the

inflammatory cell infiltration in synovial tissues, bone and cartilage was evaluated by Image J

software. After deparaffinization the slices were subjected to antigen recovery in 0.01 M

sodium citrate buffer at 125 °C for 30 s, followed by 10 s at 90 °C, and then subjected to the

endogenous peroxidase inactivation by covering tissue with 3% hydrogen peroxide for 5 min.

After blocking non-specific binding sites with 10% goat serum in PBS, the slices were

incubated with TNF-α/IL-6 monoclonal antibody (Invitrogen) with a dilution rate of 1:100,

respectively, at 4 °C for 24 h. Then the slices were incubated with horseradish peroxidase

(HRP)-conjugated secondary antibody with a dilution rate of 1:800 at 37 °C for 1 hour.

Sections were developed using the DAB substrate and then counterstained with haematoxylin.

The images were captured and analyzed by inverted microscope (Olympus, IX71, Japan) and

Image J software.

Gene expression profiles analysis of tissues

The paws and ankles were dissected from mice on day 22 of arthritis, snap-frozen in

liquid nitrogen, ground into powder, and homogenized. All procedure must be under RNase-

free conditions. The RNA isolation and real-time PCR assay were carried out following the

protocol following. Briefly, total RNA was extracted with TRIzol reagent (Invitrogen,

Carlsbad, CA, USA) from the tissue homogenates according to the manufacturer’s

instructions. The total RNA (1 �g) was reverse transcribed to cDNA using the QuantiTect

Reverse Transcription Kit (QIAGEN, Japan) and the specific gene transcripts were quantified

by quantitative real-time PCR using QuantiTect SYBR Green PCR Kit (QIAGEN K.K.,

27

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

27

Tokyo, Japan) and analyzed with ABI 7500 real-time PCR system (Applied Biosystems,

USA). PCR was performed as 40 cycles at 94°C for 15 s, 55°C for 30 s, and 72°C for 30 s.

The relative RNA expression was calculated with comparative �� method; �-actin as internal

ontrol. The gene expression patterns along with concentration were highlight with heatmap by

using TBtools package. Gene-specific primers were synthesized by Sangon Biotech

(Shanghai) Co., Ltd. and list in the TableS1.

Statistical Analysis

All the results are reported as mean ± SD. The differences among groups were analyzed

using one-way ANOVA analysis and Student’s t-test; *represents P < 0.05, ** represents P <

0.01.

Acknowledgments

This work was supported by National Natural Science Foundation of China (81801838,

31571013, 81701816 and 81601552), Guangdong Natural Science Foundation of Research

Team (2016A030312006), K.C. Wong Education Foundation (GJTD-2018-14), Natural

Science Foundation of Guangdong Province (2018A030313013), Shenzhen Science and

Technology Program (JCYJ20180302145912832, JCYJ20160429191503002,

JCYJ20170818162259843 and JCYJ20170818163739458). We thank Pro. Hongchang Li and

Dr. Yifan Ma of SIAT for their useful and meaningful recommendations to improve this paper,

and thank Mr. Danhui Quan for his help in some drawing.

Competing Interests

The authors have declared that no competing interest exists.

28

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

28

29

571

29

Scheme 1. Schematic illustration of mannose modified erythrocyte membranes-coated

triptolide nanoparticle (manRTPNs) targeting to inflammatory sites of rheumatoid arthritis

and significant therapeutic effect. Targeting capacity to macrophages and T cells of

manRTPNs is conferred from the mannose modified erythrocyte membrane. Specifically,

target activated T cells was through the CD58-CD2 affinity, and the binding interaction

between the mannose and its receptor CD206 enhanced anchoring of manRTPNs to activated

macrophages. Effective management was achieved through remodeling inflammation

microenvironment. Basically, TPNs at RA sites shift core genes expression in three pathways

including angiogenesis, bone resorption and cytokine secretion.

30

572

573

574

575

576

577

578

579

580

581

582

30

Figure 1. Preparation and Characterization of manRTPNs.

(A) Synthetic route illustration of triptolide self-assembling nanoparticles (TPNs). (B)

Representative image of TPNs and manRTPNs examined with transmission electron

microscopy (TEM). manRTPNs samples were stained with uranyl acetate. Scale bar, 200 nm.

(C) Size and ζ potential distribution of TPNs, RBC vesicles and manRTPNs. Each experiment

31

583584

585

586

587

588

31

had three replicates. (D) Stability of TPNs and manRTPNs in PBS over five weeks. Each

experiment had three replicates. (E) Pharmacokinetics of TPNs and manRTPNs in vivo. Each

experiment point had three replicates.

32

589

590

591

592

32

Figure 2. Targeting capacity to immune cells of manRTPNs.

(A-C) Cellular uptake capacity of TPNs and manRTPNs in three cell lines. Flow cytometry

analysis of TPNs and manRTPNs were conducted in RAW246.7 (A), CTLL-2 (B) and

RSC364 (C). k was slope of uptake curve built with mean fluorescence intensity (MFI) which

indicated cell uptake rate. (D-F) Representative fluorescence images by confocal analysis.

RAW246.7 (D), CTLL-2 (E), and RSC364 (F) were co-incubated with TPNs, RTPNs, and

manRTPNs, respectively. Nanoparticles were labeled with Cy5.5 (red), CD206/CD2 antibody

(green) expressed on the IL-4 -treated RAW264.7 and TNF-α treated CTLL-2 and cellular

33

593594

595

596

597

598

599

600

601

33

nuclear was labeled with DAPI (blue). Scale bars, 100 μm. (G-I) Flow cytometry analysis of

RAW246.7 (G), CTLL-2 (H) and RSC364 (I) incubated with various nanoparticles. **

indicated significant difference and p < 0.05.

34

602

603

604

605

34

Figure 3. Efficacy of manRTPNs in inflammatory cells.

(A-D) Cytotoxicity of TPNs and manRTPNs in RAW264.7 (A), CTLL-2 (B), RSC364 (C),

and hFLS-RA (D). Each experiment point had three replicates. (E) Heatmap of inflammatory

factors genes expression. Briefly, genes cox-2, timp-1 and NF-κB were checked in

RAW264.7, gene rankl was checked in CTLL-2 and genes vegf, mmp-9 and mmp-2 were

checked in RSC364, respectively. (F-J) Cytokines concentration variation of inflammatory

cells treated with TPNs and manRTPNs. Supernatant IL-1� was measured in RWA264.7 (F)

and RSC364 (I), supernatant IL-6 in RWA264.7 (G) and RSC364 (J), supernatant IFN-� in

CTLL-2 (H), each experiment had three replicates.

35

606

607

608

609

610

611

612

613

614

615

616

617

35

Figure 4. Targeting effect and accumulation at inflamed site of RA of manRTPNs.

(A) Representative images of free Cy5.5, TPNs and manRTPNs accumulating in arthritic

paws. (B) Fluorescence images of free Cy5.5, TPNs and manRTPNs in arthritic paws fixed

after 24 hours of intravenous injection. Scale bars, 500 μm. Fluorescence of free Cy5.5,

TPNs and manRTPNs was quantified in arthritic paws sections. (C) Distribution of free

Cy5.5, TPNs and manRTPNs in main organs and arthritic paws.

36

618

619

620

621

622

623

624

625

36

37

626

37

Figure 5. Therapeutic effect of manRTPNs in a mouse model of collagen-induced

arthritis.

(A) The study protocol of a therapeutic regimen with CIA mice model. (B) Arthritis indexes

of different treatment groups over 14 days; each experiment had five mice replicates. (c) Bone

mean density (BMD) variation of the mice paw joints before and after different treatment. (D)

Representative 3D reconstructed bone images of arthritic paws in different treatment groups.

(E) Representative images of H&E staining on knee sections from healthy mice and CIA mice

treated with PBS, RBC vesicle, triptolide, TPNs or manRTPNs. Scale bars, 100 μm.

38

627

628

629

630

631

632

633

634

635

636

38

Figure 6. Remodeling inflammatory factor of RA treatment by manRTPNs.

(a) Representative images of TNF-α and IL-6 immunohistochemical staining on knee

sections. Scale bars, 100 μm. (B) Immunohistochemical quantification of TNF-α and IL-6.

(C) Serum concentration of TNF-α and IL-6 in healthy mice and CIA mice treated with

39

637

638

639

640

641

39

different groups. Each experiment point had three replicates. (D) Heatmap of genes

expression profiles in different treatment.

40

642

643

644

40

Figure 7. Biocompatibility and safety of manRTPNs.

(A) Complete cell counts (white blood cells, red blood cells, and platelets) in PBS and

triptolide-, TPNs-, manRTPNs-treated groups. Experiment replicate, three. (B) Complete cell

counts (neutrophil, lymphocyte, and monocyte) in PBS and triptolide-, TPNs-, manRTPNs-

41

645

646

647

648

649

650

41

treated groups. Experiment replicate, three. (C) Concentration of HGB and MCHC in four

groups indicated before. Experiment replicate, three. (D-G) Blood routine examination

parameters including AST(D), LDH-L(E), BUN(F) and Cre(G) after different treatment.

Experiment replicate, three. (H) H&E staining of main organs in four different groups. Scale

bars, 100 μm.

42

651

652

653

654

655

656

42

1. Scott DL, Wolfe F, Huizinga TWJ. Rheumatoid arthritis. Lancet. 2010; 376: 1094-1108.

2. Van Vollenhoven RF. Treatment of rheumatoid arthritis: state of the art 2009. Nat Rev

Rheumatol. 2009; 5: 531-541.

3. Bykerk V. Unmet needs in rheumatoid arthritis. J Rheumatol. 2009; 36: 42-46.

4. Salliot C, Van Der Heijde D. Long-term safety of methotrexate monotherapy in patients

with rheumatoid arthritis: a systematic literature research. Ann Rheum Dis. 2009; 68:

1100-1104.

5. Alcorn N, Saunders S, Madhok R. Benefit-risk assessment of leflunomide an appraisal of

leflunomide in rheumatoid arthritis 10 years after licensing. Drug Safety. 2009; 32: 1123-

1134.

6. Leombruno JP, Einarson TR, Keystone EC. The safety of anti-tumour necrosis factor

treatments in rheumatoid arthritis: meta and exposure-adjusted pooled analyses of serious

adverse events. Ann Rheum Dis. 2009; 68: 1136-1145.

7. Hyrich KL, Watson KD, Isenberg DA, Symmons DPM, Register BB. The british society

for rheumatology biologics register: 6 years on. Rheumatology. 2008; 47: 1441-1443.

8. Kirwan JR, Bijlsma JWJ, Boers M, Shea BJ. Effects of glucocorticoids on radiological

progression in rheumatoid arthritis. Cochrane Db Syst Rev. 2007.

9. Ravindran V, Rachapalli S, Choy EH. Safety of medium- to long-term glucocorticoid

therapy in rheumatoid arthritis: a meta-analysis. Rheumatology. 2009; 48: 807-811.

10. Zhang QZ, Dehaini D, Zhang Y, Zhou JL, Chen XY, Zhang LF, et al. Neutrophil

membrane-coated nanoparticles inhibit synovial inflammation and alleviate joint damage

in inflammatory arthritis. Nat Nanotechnol. 2018; 13: 1182-1190.

43

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

43

11. Li RX, He YW, Zhu Y, Jiang LX, Zhang SY, Qin J, et al. Route to rheumatoid arthritis by

macrophage-derived microvesicle-coated nanoparticles. Nano Letters. 2019; 19: 124-134.

12. Liang HY, Peng B, Dong C, Liu LX, Mao JJ, Wei S, et al. Cationic nanoparticle as an

inhibitor of cell-free DNA-induced inflammation. Nat Commun. 2018; 9: 4291.

13. Tao X, Cush J, Garret M, Pe. L. A phase I study of ethylacetate extract of the Chinese

antirheumatic herb Tripterygium wilfordii Hook F in rheumatoid arthritis. J Rheumatol.

2001; 28: 2167.

14. Fan DP, Guo QQ, Shen JW, Zheng K, Lu C, Zhang G, et al. The effect of triptolide in

rheumatoid arthritis: from basic research towards clinical translation. Int J Mol Sci. 2018;

19.

15. Pao HP, Liao WI, Wu SY, Hung KY, Huang KL, Chu SJ. PG490-88, a derivative of

triptolide, suppresses ischemia/reperfusion-induced lung damage by maintaining tight

junction barriers and targeting multiple signaling pathways. Int Immunopharmacol. 2019;

68: 17-29.

16. Chen YK, Zhang L, Ni JS, Wang XY, Cheng J, Li YC, et al. LLDT-8 protects against

cerebral ischemia/reperfusion injury by suppressing post-stroke inflammation. J

Pharmacol Sci. 2016; 131: 131-137.

17. He QL, Minn I, Wang QL, Xu P, Head SA, Datan E, et al. Targeted delivery and sustained

antitumor activity of triptolide through glucose conjugation. Angew Chem Int Edit. 2016;

55: 12035-12039.

44

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

44

18. Yuan ZX, Wu XJ, Mo JX, Wang YL, Xu CQ, Lim LY. Renal targeted delivery of

triptolide by conjugation to the fragment peptide of human serum albumin. Eur J Pharm

Biopharm. 2015; 94: 363-371.

19. Mei ZN, Chen HB, Weng T, Yang YJ, Yang XL. Solid lipid nanoparticle and

microemulsion for topical delivery of triptolide. Eur J Pharm Biopharm. 2003; 56: 189-

196.

20. Huang CL, Zeng T, Li JW, Tan LS, Deng XL, Pan YC, et al. Folate receptor-mediated

renal-targeting nanoplatform for the specific delivery of triptolide to treat renal

ischemia/reperfusion injury. ACS Biomater Sci Eng. 2019; 5: 2877-2886.

21. Ling D, Xia H, Park W, Hackett MJ, Song C, Na K, et al. pH-sensitive nanoformulated

triptolide as a targeted therapeutic strategy for hepatocellular carcinoma. ACS Nano.

2014; 8: 8027-8039.

22. Xu HT, Liu B. Triptolide-targeted delivery methods. Eur J Med Chem. 2019; 164: 342-

351.

23. Abbas M, Zou QL, Li SK, Yan XH. Self-assembled peptide- and protein-based

nanomaterials for antitumor photodynamic and photothermal therapy. Adv Mater. 2017;

29.

24. Zhu PL, Yan XH, Su Y, Yang Y, Li JB. Solvent-induced structural transition of self-

assembled dipeptide: from organogels to microcrystals. Chem-Eur J. 2010; 16: 3176-

3183.

25. Yan XH, Zhu PL, Li JB. Self-assembly and application of diphenylalanine-based

nanostructures. Chem Soc Rev. 2010; 39: 1877-1890.

45

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

45

26. Elshabrawy HA, Chen ZL, Volin MV, Ravella S, Virupannavar S, Shahrara S. The

pathogenic role of angiogenesis in rheumatoid arthritis. Angiogenesis. 2015; 18: 433-

448.

27. Kinne RW, Brauer R, Stuhlmuller B, Palombo-Kinne E, Burmester GR. Macrophages in

rheumatoid arthritis. Arthritis Res. 2000; 2: 189-202.

28. Davis SJ, Ikemizu S, Evans EJ, Fugger L, Bakker TR, Van Der Merwe PA. The nature of

molecular recognition by T cells. Nat Immunol. 2003; 4: 217-224.

29. Raychaudhuri S, Thomson BP, Remmers EF, Eyre S, Hinks A, Guiducci C, et al. Genetic

variants at CD28, PRDM1 and CD2/CD58 are associated with rheumatoid arthritis risk.

Nat Genet. 2009; 41: 1313-1318.

30. Sable R, Durek T, Taneja V, Craik DJ, Pallerla S, Gauthier T, et al. Constrained cyclic

peptides as immunomodulatory inhibitors of the CD2:CD58 protein-protein interaction.

ACS Chemical Biology. 2016; 11: 2366-2374.

31. Hu CMJ, Fang RH, Copp J, Luk BT, Zhang LF. A biomimetic nanosponge that absorbs

pore-forming toxins. Nat Nanotechnol. 2013; 8: 336-340.

32. Rossi L, Fraternale A, Bianchi M, Magnani M. Red blood cell membrane processing for

biomedical applications. Front Physiol. 2019; 10.

33. He QL, Titov DV, Li J, Tan MJ, Ye ZH, Zhao YM, et al. Covalent modification of a

cysteine residue in the XPB subunit of the general transcription factor TFIIH through

single epoxide cleavage of the transcription inhibitor triptolide. Angew Chem Int Edit.

2015; 54: 1859-1863.

46

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

46

34. Hu CMJ, Zhang L, Aryal S, Cheung C, Fang RH, Zhang LF. Erythrocyte membrane-

camouflaged polymeric nanoparticles as a biomimetic delivery platform. P Natl Acad Sci

USA. 2011; 108: 10980-10985.

35. Fang RNH, Hu CMJ, Chen KNH, Luk BT, Carpenter CW, Gao WW, et al. Lipid-insertion

enables targeting functionalization of erythrocyte membrane-cloaked nanoparticles.

Nanoscale. 2013; 5: 8884-8888.

36. Shao F, Wang GJ, Xie HT, Zhu XY, Sun JG, A JY. Pharmacokinetic study of triptolide, a

constituent of immunosuppressive chinese herb medicine, in rats. Biol Pharm Bull. 2007;

30: 702-707.

37. Ghosh S, Hayden MS. New regulators of NF-kappaB in inflammation. Nat Rev Immunol.

2008; 8: 837-848.

38. Niedermeier M PT, Korb A. Therapeutic opportunities in fibroblasts in inflammatory

arthritis. Best Pract Res Clin Rheumatol. 2010; 24: 527-540.

39. Brand DD, Latham KA, Rosloniec EF. Collagen-induced arthritis. Nat Protoc. 2007; 2:

1269-1275.

40. Noack M, Miossec P. Selected cytokine pathways in rheumatoid arthritis. Seminars in

Immunopathology. 2017; 39: 365-383.

41. Geusens P. The role of RANK ligand/osteoprotegerin in rheumatoid arthritis. Ther Adv

Musculoskelet Dis. 2012; 4: 225-233.

42. Chen L, Deng H, Cui H, Fang J, Zuo Z, Deng J, et al. Inflammatory responses and

inflammation-associated diseases in organs. Oncotarget. 2018; 9: 7204-7218.

47

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761

762

763

47

download fileview on ChemRxivmanuscript-update 20200804.docx (7.30 MiB)

1

Triptolide self-assembling nanoparticles engineering with 1

modified erythrocyte membranes for targeting and remodeling 2

inflammatory microenvironment in arthritis 3

Jing Li1, Sanpeng Li1, Chunbin Li1, Hongfeng Li1, Chuangjun Liu1, Qi Zhao3, Pengfei 4

Zhang1*, Ping Gong1,2*, Lintao Cai1* 5

1Guangdong Key Laboratory of Nanomedicine, CAS-HK Joint Lab for Biomaterials, 6

Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 7

518055, China 8

2Guangdong Key Laboratory for Research and Development of Natural Drugs, Guangdong 9

Medical University, Dongguan, 523808, China 10

3 Faculty of Health Sciences, University of Macau, Taipa, Macau, China. 11

12

*Corresponding author: Lintao Cai: E-mail: lt.cai@siat.ac.cn; 13

Ping Gong: E-mail: ping.gong@siat.ac.cn 14

15

2

Abstract 1

Overview: Rheumatoid arthritis (RA) is an autoimmune disease characterized by persistent 2

synovitis, systemic inflammation and causing severe joint damage. Inflammation and influx of 3

immune cells is a hallmark of the disease. Triptolide with great immunosuppressive activities 4

is a potential drug to treat RA. However, its severe toxicity is still an unresolved bottleneck and 5

largely limits its clinical use. Recently, nanoparticle (NP)-based drug delivery systems have 6

been developed for detoxification of toxic drugs and diseases management. 7

Aim: To synthesize triptolide nanoparticles to decline its toxicity and improve water-solubility 8

and develop a dual targeted biomimicking delivery system which targets to inflammation 9

microenvironment. Finally, to achieve excellent therapeutic effect in rheumatoid arthritis. 10

Methods: We synthesized an amphiphilic molecule prodrug with the addition of 11

diphenylalanine peptide (FF) to triptolide. These prodrugs were self-assembled into 12

nanoparticles (TPNs) in water, then TPNs were coated with mannose-modified erythrocyte 13

membranes to form a dual targeting manRTPNs. Specifically, mannose and natural molecule 14

CD58 on RBC membrane guide it towards activated macrophages and T cells in inflamed joint 15

through innate and specific immune recognition, respectively. Different cell lines including 16

RAW264.7, CTLL-2, RSC364 and hFLS-RA and a collagen induced arthritis (CIA) mice 17

model were used to detect the targeting and therapeutic effect in vitro and vivo. 18

Results: manRTPNs showed an excellent dual targeting ability to activated macrophages and 19

T cells in vitro. We found that manRTPNs selectively induced anomalous activated cells death 20

and indeed altered genes expression and cytokines production of inflamed cells such as 21

3

macrophages, T cells and synoviocytes. In CIA mice, manRTPNs exhibited significantly 1

targeting effect and prolonged drug accumulation at inflamed joints, whereby increased drug 2

achieved significant therapeutic effect without adverse effect. In detail, manRTPNs ameliorated 3

joint destruction, swollen paws considerably shrunk to normal, bone and cartilage destruction 4

were suppressed and bone density was even recovered to healthy level. Basically, several core 5

genes like NF-κB, cox-2, rank, rankl, vegf, vegfr2, mmps, and cytokines such as IL-6, TNF-α 6

involved in inflammation, cartilage/bone destruction and angiogenesis were regulated and 7

altered expression and production. Thus, the inflammation microenvironment was remodeled 8

by TPNs. 9

Conclusion: This study demonstrated phenylalanine dipeptide mediating triptolide self-10

assembled nanoparticles to solve current problems and the potential of engineered RBC 11

membranes for inflammation-targeted drug delivery. Therefore, our work is a good example of 12

triptolide reconstruction, and this agent based on immune recognition may provide an immune 13

cell targeted strategy for rheumatoid arthritis therapy. 14

Keywords: triptolide, nanodrug, dual targeting, inflammatory microenvironment, rheumatoid 15

arthritis 16

17

4

Graphical Abstract 1

2

A novel triptolide nanodrug based therapeutic agent for RA management that modified 3

erythrocyte membranes as a develiry system capable of dual targeted and triptolide remodeling 4

inflammatory microenvironment in arthritis. 5

6

5

Introduction 1

Rheumatoid arthritis (RA) is a serious long-term disease characterized by persistent 2

synovitis, systemic inflammation and autoantibodies and causing severe joint damage, 3

disability and decreased quality of life. The etiopathogenesis remains elusive, and viable and 4

effective treatment options are still limited[1]. Over the last two decades, long-term 5

management of RA has involved disease-modifying anti-rheumatic drugs (DMARDs), 6

biological agents such as tumor necrosis factor (TNF) inhibitors, and glucocorticoids[2]. 7

Although the highest clinical remission rate achieved within 50% of control, not all patients 8

attain desirable levels of clinical remission[3] and clinical use of these therapies is limited 9

because of their high cost and frequency of adverse effects. The latter includes teratogenicity 10

and hepatotoxicity of DMARDs like methotrexate[4, 5], risks of infections such as tuberculosis 11

and osteoporosis caused by biological agents[6, 7] and long-term adverse reactions to 12

glucocorticoid therapy[8, 9]. Lately, new therapies based on inflammatory cells such as 13

macrophages and neutrophil or their products such as cell-free DNA in inflamed joint have 14

been reported[10-12], however inhibiting a single factor may not be enough to sustainably halt 15

or reverse disease progression due to RA’s complexity and heterogeneity. Therefore, a new 16

systemic design for drug candidate based on nano-medicine is urgently needed. 17

Triptolide is a trace natural triepoxide diterpene from Tripterygium wilfordii (or Léi Gōng 18

Téng) and has been widely used to treat for RA patients since the 1970s in China[13]. Despite 19

its various biological activities, low water solubility and severe toxicity, especially to the liver 20

and kidney, remarkably impedes the translation to clinical applications[14]. Considerable 21

6

efforts in structural modification[15-18] and nanoparticle(NP) -based drug delivery system [19-1

21] have been applied to reduce its toxicity. Unfortunately, common practices that direct 2

conjugations attached to the C14-hydroxyl group of triptolide via a hydrolytic ester bond have 3

proven biologically unstable; meanwhile, current targeted drug-loading delivery systems are 4

limited by drug loading or encapsulation ratio and they mostly target to kidney or tumor[22]. 5

Therefore, an effective and stable system that can deliver triptolide to other organ or tissue such 6

as joint is necessary. But until now, triptolide-formed nanoparticles have been rarely reported 7

before. From this point, inspired from biomacromolecule self-assembly process, various 8

peptide building blocks have been performed for the creation of biomimetic or bioinspired 9

nanostructured materials[23, 24]. Recently, diphenylalanine (FF) is the simplest peptide 10

building block for self-assembly and exhibits remarkable advantages including 11

biocompatibility and non-immunogenicity and so on. Many researchers have developed FF-12

based nanomaterials such as nanotubes, spherical vesicles, nanofibrils and nanowires[25]. 13

However, it is barely to link it with toxic drug for nanodrug formation. Therefore, 14

diphenylalanine (Phe-Phe, FF) mediating self-assembling nanoparticles is a new strategy to 15

reconstruct triptolide to be nanoparticles whose spherical nanostructure was more stable and 16

biocompatible than before. 17

In RA pathogenesis, numerous blood-derived cells infiltrate in synovium and cytokine 18

secret necessitates formation of new blood vessels to generate pannus, that all of these form an 19

inflammatory microenvironment[26]. Peripheral-blood monocytes are recruited to joint, and 20

functionally diverse macrophages distribute in the synovial sublining and lining layers. It is 21

7

mentioned that Cluster of Differentiation 206 (CD206) on macrophages surface has been 1

induced to express more [27]. Besides, prominent T cells being activated, especially the CD4+ 2

subset, also infiltrate the synovium. As reported, certain signal proteins on the T cell surface 3

such as lymphocyte function antigen 2 (LFA-2/CD2) have increased in both RA patients and 4

animal models[28, 29]. It was reported that CD2 can be effectively targeted by peptides derived 5

from LFA-3/CD58 as well as fusion protein constructs. Interestingly, the ligand of CD2 namely 6

CD58 widely distributes on erythrocyte membranes and CD58-CD2 interaction have mediated 7

process of T cells activation[30]. These immunological characteristics of erythrocytes scarcely 8

ever attracts attention in application, so it is a valuable attempt to apply in drug targeting. Then, 9

mannose, the ligand of CD206 was introduced in erythrocyte membrane for ligand-binding 10

targeting. In addition, the long circulation and flexible functionalization of erythrocyte 11

membrane also make it be a promising platform and have wide application[31, 32]. In this 12

study, we prepared mannose-labeled RBC membrane to entrap newly synthesized triptolide 13

nanoparticles (TPNs) for elevating its vast retention in inflammation and achieving remarkably 14

anti-inflammatory therapy. Triptolide, a hydrophobic drug, was linked with diphenylalanine 15

peptide to self-assembly into size-controlled nanoparticles (TPNs) in water, and this 16

reconstruction has successfully decreased its toxicity and improved water-solubility. As shown 17

in scheme 1, the nanodrug core was coated with decorated RBC membranes to obtain mannose-18

modified RBC cell membrane-coated triptolide nanoparticles (manRTPNs). By translocating 19

RBC membranes to the nanoparticles, the biological targeting properties of RBC cell can be 20

conferred to manRTPNs as well, allowing them to target activated T cells through the CD58-21

8

CD2 affinity. Further, the binding interaction between the mannose and its receptor CD206 1

enhanced anchoring of manRTPNs to activated macrophages. According to manRTPNs, the 2

dual-targeting strategy facilitated the accumulation of nanodrug in inflammation, and the 3

accumulated TPNs efficiently achieved systemic administration of rheumatoid arthritis by 4

specifically inducing cell death and remodeling the inflammatory microenvironment through 5

affecting genes network, downregulating cytokines and pro-angiogenic factors, regulating 6

osteoclastogenesis, and returning balance of metalloproteinases and their inhibitors. Overall, 7

our study expanded the RBC cell membrane biological application in nanocarrier for targeting 8

inflammation through innate and specific immune recognition, and the triptolide nanodrug 9

showed high drug content, low side effect, good biocompatibility which showed great potential 10

for further inflammation diseases therapy. 11

Results and Discussion 12

Preparation and characterization of manRTPNs 13

Triptolide is a diterpene triepoxide with several active sites that can be modified, 14

especially C14 position[33]. Because of its hydrophobicity, triptolide molecule acts as self-15

assembly inducer with the addition of diphenylalanine peptide (FF), which is the key moiety of 16

the self-assembling peptide-drug conjugate. Initially, we added a glucosamine (AG) to the FF 17

terminal to lengthen the hydrophilic chain, and then linked triptolide to FF-AG by an ester bond 18

whose cleavage led to drug release in the cell; thus, the self-assembling monomer (TP-FF-AG) 19

as a triptolide prodrug was synthesized (Figure 1A). The chemical structures of these 20

compounds were confirmed via mass spectrometry (MS) and nuclear magnetic resonance 21

9

(NMR; Figure S1-S16). Due to hydrophobic interactions, triptolide nanoparticles (TPNs) were 1

formed via self-assembly of synthesized monomer namely TP-FF-AG, which is uniformly easy 2

to control TPNs size in water solution. The particles morphology was observed via transmission 3

electron microscopy (TEM; Figure 1B). To synthesize manRTPNs, RBC membranes were 4

extracted by hypotonic hemolysis[34], and mannose inserted by PEG2000-DSPE. An extruder 5

was using to camouflage mannose-modified erythrocyte membranes onto TPNs then form a 6

core-shell structure (Figure 1B). Dynamic light scattering (DLS) measurements revealed that 7

the TPNs were 135 nm in hydrodynamic diameter and manRTPNs were about 10 nm larger 8

with PDIs of 0.15-0.2. The surface zeta potential of the shell was not as negative as that of the 9

cores but was comparable to that of erythrocyte membrane-derived vesicles (RBC vesicles) 10

(Figure 1C). All these results demonstrated successful cloaking of TPNs with erythrocyte 11

membranes. 12

Next, TPNs and manRTPNs were stored in phosphate-buffered saline (PBS) for 1 month, 13

where they demonstrated a negligible change in size and showed an excellent stability (Figure 14

1D). Maintaining protein compounds of erythrocyte membranes on nanoparticles’ surfaces is 15

vital for harnessing their biofunctionality. Therefore, completeness of the membrane coating 16

was verified via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE; 17

Figure S17), which indicated the manRTPNs maintained the whole effective biocomponent of 18

erythrocyte on the surface. Translocation of erythrocyte membranes might improve the 19

biocompatibility and circulating period of TPNs in vivo, which also sustains systemic delivery 20

and better targeting through both passive and active uptakes[35]. As reported, the elimination 21

10

T1/2 of triptolide molecules in rats following i.v. administration was 15-21 minutes, which 1

showed that drugs were rapidly eliminated[36]. However, TPNs and manRTPNs have 2

maintained higher plasma concentration after tail intravenous injection of mice in our work. As 3

the relative signal of drug down to 50%, TPNs lasted 4 h, and manRTPNs even had a much 4

longer elimination half-life of over 48 hrs (Figure 1E), which suggested that manRTPNs 5

significantly expanded blood circle compared to triptolide. Therefore, the self-assembling 6

TPNs nanostructure was more stable than original drug, and manRTPNs inherited biological 7

properties of erythrocyte membrane proteins for following efficacy. 8

Dual-Targeting capacity and immunosuppressive effect of manRTPNs 9

To determine each targeting capacity of manRTPNs to macrophage and T cell in vitro, 10

nanoparticles were Cy5.5-labelled and incubated with RAW264.7 and CTLL-2 treated with IL-11

4 (100 ng/mL) or TNF-α (10 ng/mL), untreated cells and RSC364 as control. Flow cytometry 12

(FCM) analysis observed gradual increased uptake of nanoparticles by RAW264.7 and CTLL-13

2 for 6 hrs (Figure S18-20). An uptake curve was built using mean fluorescence intensity (MFI) 14

of cells at six time points, and cellular uptake rate was defined as k indicated in figure 2A-C, 15

which was the slope of the uptake curve. In detail, ks of nanoparticles in activated RAW264.7 16

and CTLL-2 were significantly higher than that in untreated cells (Figure 2A-C; Figure S21A 17

& S22), and ks of manRTPNs were significantly higher than that to TPNs both in RAW264.7 18

and CTLL-2 (Figure 2A-B), but no difference of ks existed between TPNs and manRTPNs in 19

RSC364 that no targeting ligands on it (Figure 2C). Further, the k value in RAW264.7 was 20

higher than that in CTLL-2 (Figure 2A-B), indicating the affinity difference of interacted 21

11

molecule pairs. To further confirm ligand-binding specificity, Cy5.5-labelled TPNs (without 1

mannose and CD58), RTPNs (with CD58) and manRTPNs (with mannose and CD58) were 2

incubated with treated macrophages and T cells for 3 hrs as well, RSC364 and hFLS-RA as 3

control. After treatment, CLSM images showed that CD206 and CD2 (Figure 2D-E, green) 4

respectively overexpressed on surface of RAW264.7 and CTLL-2. Significant red fluorescence 5

was observed on RAW264.7 incubated with manRTPNs, but not with TPNs and RTPNs (Figure 6

2D), while red fluorescence was observed on CTLL-2 incubated with RTPNs and manRTPNs 7

but not with TPNs (Figure 2E), which was consistent with corresponding ligands distribution 8

on the nanoparticles. When co-incubated with RSC364, they also entered cells (Figure 2F; 9

Figure S21b). Similar results were observed in flow cytometry analysis (Figure 2G-I; Figure 10

S21c). These results demonstrated the ability of manRTPNs to target inflamed cells conferred 11

by their mannose modified erythrocyte membrane coating, which is probably attributed to 12

specific affinity of mannose-CD206 and CD58-CD2. 13

Except for migrated immune cells, synovial fibroblast is main resident effector of 14

inflammation. To determine immunosuppressive efficacy of manRTPNs in joint inflammatory 15

cells, LPS-treated RAW264.7, TNF-α-treated CTLL-2 and RSC364, and hFLS-RA were 16

incubated with manRTPNs in 96-well plates for 24 hrs. Firstly, cell viability was measured via 17

MTT assay. TPNs and manRTPNs didn’t significantly reduce RAW264.7 viability even at a 18

concentration of 500 nM (Figure 3A). However, at the same concentration CTLL-2, RSC364 19

and hFLS-RA showed an obvious cytotoxicity (Figure 3B-D). Specially, manRTPNs had an 20

IC50 (half maximal inhibitory concentration) values of 219 nM for activated CTLL-2 cytoxicity, 21

12

88.9 nM for activated RSC364 cytoxicity and 95.8 nM for hFLS-RA cytoxicity (Figure 3B-D). 1

The lower IC50 values of activated RSC364 and hFLS-RA indicated that triptolide-induced 2

cell death was mainly against anomalous proliferative synoviocytes in inflamed joints, rather 3

than lymphocytes like macrophages. Cells were following collected and lysed for mRNA 4

extraction to check certain inflammation associated genes variation by qPCR. For LPS-treated 5

RAW264.7, transcription factor NF-κB associated with most cytokine transcription[37] and 6

cyclooxygenase-2 (cox-2) were induced to decrease expression level, but tissue inhibitor of 7

metalloproteinase-1 (timp-1) showed opposite variation under manRTPNs treatment. Receptor 8

activator of NF-κB ligand (rankl) over-expressed on TNF-α-treated CTLL-2 was also 9

downregulated by manRTPNs. After activated with cytokines and chemokines, synovial 10

fibroblasts release matrix-degrading enzymes, including matrix metalloproteinases (MMPs) 11

and cathepsins causing cartilage and bone destruction[38]. In RSC364, the transcripts of 12

metalloproteases such as MMP-2 and MMP-9 and vascular endothelial growth factor (VEGF) 13

being critical for angiogenesis were lessen by manRTPNs as well (Figure 3E). Thirdly, medium 14

supernatants were collected to measured cytokines by ELISA assay. supernatant IL-1β and IL-15

6 of LPS-treated RAW264.7 and supernatant IFN-𝛾 of TNF-α-treated CTLL-2 lessened more 16

quickly that treated with manRTPNs than TPNs (Figure 3F-H), while supernatant IL-1β and 17

IL-6 of TNF-α-treated RSC364 had a similar decreasing curve between TPNs and manRTPNs 18

groups (Figure 3I-J), which indicated that manRTPNs had superior identification of immune 19

cells than resident synoviocytes at least in cytokine secretion. Consequently, manRTPNs 20

selectively exerts its immunosuppressive effects on selectively inducing lymphocytes and 21

13

synovial fibroblasts cell death, but also regulating production of proinflammatory cytokines, 1

proinflammatory mediators and matrix metalloproteinases, which suggested manRTPNs may 2

have an ability to remodel inflammatory microenvironment through its cell-specific anti-3

inflammation efficacy. 4

Targeting capacity and accumulation of manRTPNs in vivo 5

Upon the good capacity of manRTPNs to target activated macrophages and T cells in vitro, 6

we further verified their ability to target RA site in vivo. After CIA mouse model being induced, 7

all four paws swelled in varying degrees[39]. Hind paws were more heavily swollen, and micro-8

computed tomography (Micro-CT) imaging showed distinct bone loss of joint therein (Figure 9

4A). Then these mice received intravenous injections of free cyanine 5.5 (Cy5.5) solution and 10

Cy5.5-labelled TPNs and manRTPNs for targeting imaging in vivo. As shown in figure 4A, 11

almost no fluorescence signal was detected in paws of the free-Cy5.5 group, indicating that the 12

dye barely accumulated at local inflammatory sites. In the TPNs group, we detected a weak 13

signal, which peaked after 4 h (Figure 4A). While manRTPNs obviously reached the RA sites 14

and lasted there for ≥12 hrs, which would be attributed to their long circulation and the targeting 15

effect of the modified erythrocyte membranes. Thus, significant fluorescence of manRTPNs 16

lasted longer and showed greater accumulation than that of TPNs. 24 hrs after intravenous 17

injections, we dissected mouse and got froze sections of left paws to observe nanoparticles’ 18

retention in the inflamed joints. Contrary to free-Cy5.5 and TPNs group, fluorescence 19

microscopy indicated a very high fluorescence intensity of manRTPNs in the synovial sublining 20

layer, resulting in a reduced accumulation at the inflamed joint due to the lack of permeability 21

14

(Figure 4B). Tissue biodistribution further revealed manRTPNs’ outstanding capacity to target 1

the paws of RA mice, rather than other groups (Figure 4C). Additionally, higher accumulation 2

was observed in liver, suggesting that this organ might be passive targeting of nanoparticles 3

and the main organ metabolizing triptolide (Figure 4C). The results confirmed that the effective 4

accumulation of manRTPNs provided a promising precondition of RA treatment. 5

Therapeutic efficacy of manRTPNs 6

The efficacy of ameliorating joint damage of manRTPNs was evaluated in a CIA mouse 7

model. Following arthritis induction, CIA mice presented symptoms of swelling in ankle and 8

paw joints and a clinical scoring was used to estimate the swelling degree (Figure S23). 9

According to protocol timeline, CIA mice were treated with intravenous injections of triptolide 10

at 15 mg/kg, TPNs and manRTPNs at doses of 25 mg/kg for 11 days (Figure 5A). Surprisingly, 11

swollen hindpaws and forepaws both shrunk considerably and even recovered to the same 12

extent as healthy ones with manRTPNs treatment rather than slightly shrunk in TPNs and 13

triptolide group, but swelling symptoms of other groups were increasingly severe (Figure 5B), 14

implying that inflammation process was exactly controlled by manRTPNs. As paw swelling is 15

only a surface phenomenon of RA, bone loss was further determined using Micro-CT analysis. 16

Reconstructed three dimensional (3D) bones images and calculated bone mean density (BMD) 17

showed that serious bone erosion was suppressed and bone density even recovered to healthy 18

mice level of manRTPNs treatment (Figure 5C-D), suggesting a protective role of manRTPNs 19

on cartilage and bone damage. Haemotoxylin and eosin (H&E)-stained sections showed an 20

intense anomalous proliferative cells infiltration in the joints and synovium of control groups. 21

15

By contrast, remarkable reduction of immune infiltration and preservation of cartilage were 1

observed in manRTPNs group (Figure 5E). Further, cytokines including TNF-α and IL-6 are 2

known to increase with the onset of arthritis and correlate strongly with disease severity[40]. 3

Immunohistochemical analysis showed that TNF-α and IL-6 in synovium and articular cavity 4

significantly declined in manRTPNs groups (Figure 7A-B) and serum levels of them also 5

observed to alleviate similarly (Figure 6C), indicating effective reduction of arthritis at 6

cytokines level. Moreover, gene heatmap showed that manRTPNs suppressed several genes 7

expression including rank, rankl, vegf, vegfr2, NF-κB, cox-2, IL-1β, mmp-2, mmp-3, mmp-9 8

and mmp-13, that these transcripts are dramatically increased and contribute in joints damage. 9

Some negative regulators such as opg, timp1 and timp2 were upregulated their gene expression 10

by manRTPNs (Figure 6D), indicating manRTPNs preserve appropriate and balanced 11

concentration of cytokines and inflammatory factors through two-way regulation. 12

RANKL/RANK/ORG signal pathway is key of bone metabolism[41]. The up-/down-regulation 13

of manRTPNs also recovered the balance of RANKL/RANK/ORG and metalloproteinases and 14

their inhibitors (TIMP) to normal. Thus, manRTPNs ameliorated destroyed through aiming to 15

RA pathological factors including cytokines, extracellular matrix (ECM), bone metabolism and 16

angiogenesis. 17

Biosafety of manRTPNs nanoparticles 18

To investigate safety and biocompatibility of TPNs and manRTPNs in vivo, we 19

administered PBS, TPNs or manRTPNs at a single dose of 25 mg/kg to healthy mice via tail 20

vein, while to other mice we administered triptolide (original drug of TPNs and manRTPNs) at 21

16

15 mg/kg by intraperitoneal injection. It is worth mentioning that most mice in the triptolide 1

group died soon (within 24 hrs) after 25 mg/kg administration, indicating the high toxicity of 2

the original drug (Figure S24). After 1-week treatment, mice treated with triptolide showed a 3

significant decrease in count of white blood cells (WBC), RBCs and platelets (Figure 7A), 4

suggesting the occurrence of bone marrow suppression, which is one of the main side effects 5

of triptolide. The main three type cells comprised WBC including neutrophils, lymphocytes, 6

and monocytes showed significant count differences between triptolide group and other groups, 7

which explained the WBC count variation conferred from triptolide (Figure 7B). The 8

hemoglobin concentration did not differ significantly among all the groups in the blood (HGB), 9

mean corpuscular hemoglobin (MCHC) level, and hemoglobin distribution width (HDW) 10

(Figure 7C), indicating that triptolide didn’t influence the physiological function of RBCs. 11

Blood biochemistry analysis showed that blood urea nitrogen (BUN), aspartate 12

aminotransferase (AST), creatinine (Cr) and lactate dehydrogenase (LDH) levels of mice 13

treated with triptolide were significantly higher than those of mice in the TPNs and manRTPNs 14

groups (Figure 7D-G). These variations demonstrated that the hepatotoxicity and 15

nephrotoxicity of triptolide were greatly reduced in TPNs, with RBC membrane coating 16

yielding even better safety. Hematoxylin and eosin (H&E) staining of major organs showed 17

that triptolide treatment caused inflammatory infiltration of the liver and severe renal injury 18

characterized by loss of the brush border, tubular epithelial cells detachment from the basement 19

membrane and tubular obstruction. Meanwhile, the TPNs and manRTPNs groups showed no 20

17

significant variation in cellular morphology and negligible organ damage, in line with the PBS 1

group (Figure 7H). However, TPNs and manRTPNs significant reduced triptolide toxicity. 2

Conclusion 3

Our study expanded the RBC cell membrane biological application in nanocarrier for 4

targeting inflammation through innate and specific immune recognition, and the triptolide 5

nanodrug showed high drug content, low side effect, good biocompatibility which showed great 6

potential for further inflammation diseases therapy. Generally, there are two improvement of 7

manRTPNs in cell-membrane-coated nanoparticle delivery system, namely triptolide nanodrug 8

core and RBC membrane shell. RBC membrane-camouflaging nanoparticles has been 9

extensive established platform, but there are still many deposits from RBC cell itself to further 10

develop disease-targeting and therapy platform except for tumor. The natural immunological 11

function of RBC cells has been always neglected in targeted drug delivery construction. So, 12

taking advantage of adhesion molecules affinity between RBC cells and immunocytes in 13

building nanocarrier for transporting drug to inflammation deeply expands application of 14

established RBC membrane platform. On the other side, triptolide has excellent anti-15

inflammation and immunosuppressive effect and solving its low water-solubility and toxicity 16

is premise of using it in various diseases management. Our reconstruction of triptolide isn’t 17

content-limited loading in micelles, but drug-based self-assembled size-controlled 18

nanoparticles. This strategy takes advantages of drugs’ hydrophobicity to solve its bad water-19

solubility, which is magic of nanotechnology. While inflammation is prevalent related with 20

many diseases like cardiovascular diseases, gastrointestinal tract, cancer and pathogen 21

18

infection[42]. Therefore, manRTPNs as a drug delivery system of inflammation targeting and 1

treating may be a promising agent in other inflammation-associated diseases therapy. 2

Materials and methods 3

Animal care 4

Mice were housed in an animal facility at Shenzhen Institutes of Advanced Technology, 5

Chinese Academy of Sciences (SIAT). All animal experiments were performed in accordance 6

with Guidance Suggestions for the Care and Use of Laboratory Animals and approved by 7

Sciences Animal Care and Use Committee of SIAT. 8

Erythrocyte membranes derivation 9

Erythrocyte membranes were collected according to a previously reported method with 10

modification. Briefly, fresh heparinized whole blood was collected from male mice (20-22 g) 11

and subsequently centrifuged at 2000 rpm for 15 min at 4 °C to remove the plasma and the 12

white buffy coat. The collected red blood cells (RBCs) were washed with 1× PBS three times 13

and suspended in 0.25× PBS for 2 hours at 4 °C, and then the hemoglobin was removed by 14

centrifugation at 9,000 rpm for 45 min. The resulting pink pellet was purified with 1× PBS, and 15

the collected erythrocyte membrane was suspended and stored in distilled water. To form 16

ligand-inserted erythrocyte ghosts, the light pink solution was incubated with DSPE-PEG-17

mannose for 30 min, 37°C water bath. Erythrocyte membranes with PEG-mannose or not was 18

further prepared by extruding through a 200 nm polycarbonate porous membrane with an 19

Avanti mini-extruder (Avanti Polar Lipids). 20

19

Synthesis of TPNs 1

The TPNs were synthesized by a prodrug self-assembly method. Briefly, after dissolving 2

5mmol Boc-Phe-Phe-OH (BOC-FF, J&K scientific Ltd., China), 6 mmol 1-Ethyl-3-(3-3

dimethylaminopropyl) carbodiimide (EDC, J&K scientific Ltd., China) and 6 mmol N-4

Hydroxysuccinimide (NHS, J&K scientific Ltd.,China) in 3 mL anhydrous DMF, another 5

anhydrous DMF solution containing 5.5 mmol glucosamine (AG, Sigma-Aldrich, USA) and 5 6

mmol TEA (J&K scientific Ltd., China) was added drop-wise into it. The mixture was 7

evaporated under reduced pressure for 24 hours to get crude product BOC-FF-AG. Then, crude 8

BOC-FF-AG was dissolved in 30 mL equal volume mixture of anhydrous dichloromethane and 9

trifluoroacetic acid (TFA, J&K scientific Ltd., China) and stirred for 4 hours at room 10

temperature. Afterward, pure NH2-FF-AG was obtained after evaporated under reduced 11

pressure and purified by silica gel column chromatography with dichloromethane and 12

methanol. Secondly, 1 mmol triptolide (TP, Shanghai Yuanye Bio-Technology Co., Ltd., 13

China), 1.2 mmol succinic anhydride (SAA, Sigma-Aldrich, USA) and 0.2 mmol dimethyl 14

aminopyridine (DMAP, Sigma-Aldrich, USA) were dissolved in 3 mL pyridine and stirred 15

overnight. After then, the mixture was diluted with ethyl acetate and washed it with solution 16

containing saturated copper sulfate (J&K scientific Ltd.), water and brine. After reaction, 17

residue TP-COOH was obtained through drying with anhydrous Na2SO4 and concentration with 18

rotary evaporator. In the third step, TP-COOH, dicyclohexylcarbodiimide (DCC, Sigma-19

Aldrich, USA) and 1-hydroxy-5-pyrrolidinedione (NHS, J&K scientific Ltd., China) of equal 20

molar ratio were dissolved into 3 mL anhydrous dichloromethane and stirred at room 21

20

temperature for overnight. The final product TP-FF-AG was gained by reaction of activated 1

TP-COOH and NH2-FF-AG for overnight at room temperature and purified by analytical RP-2

HPLC and semi-preparative RP-HPLC with simultaneous detection at 265 nm. Analytical RP-3

HPLC was performed at room temperature on the Shimadzu LC 20 with UV detector SPD-20A 4

using Inertsil ODS-SP column (4.6 x 250 mm, 5 µm, 100Å). The RP-HPLC gradient was started 5

at 10% of B (CH3CN, J&K scientific Ltd., China), then increased to 100% of B over 30 min 6

(A: 0.1% TFA in water). Semi-preparative RP-HPLC was performed on the ULTIMAT 3000 7

Instrument (DIONEX). For fluorescence imaging experiments, cyanine5.5 (Cy5.5, Lumiprobe, 8

USA) was linked to intermediate product of NH2-FF-AG after esterification reaction to obtain 9

Cy5.5-FF-AG in the same way. Ultimately, nanoparticles were self-assembled by added 10

TP/Cy5.5-FF-AG dissolved in DMSO dropwise into PBS while vigorous stirring and dialysis 11

for 2 days. 1H NMR and 13C NMR spectra were recorded with a Bruker VANCE III400 12

spectrometer (400 MHz). The high-resolution mass spectra (HR-MS) were measured on a 13

Bruker Micro TOF II 10257 instrument. For membrane coating, nanoparticles were mixed with 14

prepared membranes in ratio of 1:1. The resultant mixture was subsequently extruded nine 15

times through a 200 nm polycarbonate porous membrane using an Avanti mini-extruder to yield 16

the erythrocyte membranes-coating TPNs or mannose inserted erythrocyte membrane-coating 17

TPNs (namely RTPNs or manRTPNs). 18

Nanoparticles Characterization 19

The hydrodynamic diameter and zeta potential of NPs suspended in 1 × PBS were 20

measured by dynamic light scattering (DLS) (Zeta Plus, Brookhaven Instruments, USA). The 21

21

morphologies of TPNs and manRTPNs were observed by transmission electron microscope 1

(TEM, FEI spirit T12) at an accelerating voltage of 120 keV. The stability experiment of TPNs 2

and manRTPNs in PBS was monitored by DLS over five weeks and stored at 4°C. The proteins 3

retained on manRTPNs compared with those on the natural erythrocyte membrane were 4

observed by SDS-PAGE. Briefly, 2 mg of erythrocyte membrane vesicles and manRTPNs were 5

collected by centrifugation at 15,000g for 30 min, mixed in SDS sample buffer (Invitrogen, 6

USA), and heated at 90 °C for 5 min. Then, 20 μL of each sample was run on a 10% SDS-7

polyacrylamide gel (Bio-Rad, SDS-PAGE Gel Preparation Kit, USA) at 120 V for 1 h, followed 8

by Coomassie blue staining and imaging. 9

Pharmacokinetics 10

To evaluate the circulation half-life of TPNs and manRTPNs in vivo, 150 μL Cy5.5-11

labeled nanoparticles were injected into the tail vein of the mice (n = 3). Twenty microliters of 12

blood were collected at 1, 5, 15, 30 min, and 1, 2, 4, 8, 24, 48, and 72 h following the injection. 13

Each particle group contained 3-4 mice. The collected blood samples were diluted with 30 μL 14

PBS in a 96-well plate before fluorescence measurement. Pharmacokinetics parameters were 15

calculated to fit a two-compartment model and a one-way nonlinear model. 16

Drug toxicity analysis in vivo 17

Purchased from Vital River Laboratory Animal Technology Co. Ltd (Beijing, China), 18

males BALB/c mice (6∼8 weeks old) were randomly divided into four groups and 19

intravenously injected with PBS, TPNs and manRTPNs at a single dosage of 25 mg/kg and 20

triptolide of 15 mg/kg 4 times for eight days. A total of 7 days after last administration, all mice 21

22

were euthanized, and blood was collected for and blood routine examination and biochemical 1

parameters measurement, while main organs including heart, liver, spleen, lung, and kidney 2

were excised and stained by H&E for histological analysis. 3

Cell culture 4

RAW264.7 cells, CTLL-2, RSC364 and hFLS-RA were purchased from the American 5

Type Culture Collection (ATCC, USA) and cultured in Dulbecco’s minimum essential medium 6

(DMEM, Corning, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA), 1% 7

penicillin (100 IU/mL, Corning, USA), and streptomycin (100 μg/mL, Corning, USA) and 8

placed at 37 °C in a 5% CO2 humidified atmosphere. Cell lines were purchased with 9

certification of authentication and free from Mycoplasma. 10

Quantification of cell targeting in vitro 11

A 6-well culture plate was prepared and RAW264.7, CTLL-2, RSC364 and hFLS-RA 12

cells were implanted at a density of 5 × 105 cells/well. 12-16 hrs later, 80% of cell coverage was 13

confirmed. 50 𝜇L Cy5.5-labelled TPNs or manRTPNs was adding into prepared cells every 14

hour for six hours, incubated with covering foil paper and collected them in the sixth hour. 15

After digestion with trypsin (Gibco, USA), cells were put together and washed three times with 16

ice-cold PBS, then underwent flow cytometry by FACS Canto II (BD Biosciences). The data 17

was analyzed using FlowJo software. Mean fluorescence intensity (MFI) reflected amount of 18

intake nanoparticles amount. An uptake curve was constructed with MFI and time, and a k value 19

being the slope of uptake curve was defined as uptake rate to measure targeting capacity of 20

nanoparticles. 21

23

Immunofluorescence assay of macrophages and T cells with nanoparticles 1

Cells were cultured with medium added recombinant mouse IL-4 (100 ng/mL, R&D 2

System, USA) or recombinant TNF-α (10 ng/mL, R&D System, USA) for six hours and 3

incubated with TPNs, RTPNs or manRTPNs for 3 hours. Afterward, cells were fixed with 4% 4

paraformaldehyde (Sigma-Aldrich, USA) for 10 minutes, permeabilized with 0.1% Triton™ 5

X-100 (Sigma-Aldrich, USA) for 10 minutes, blocked with 1% BSA for 1 hour and labeled 6

with 2 µg/mL anti-mannose receptor/AF488 conjugated antibody (ab195191, Abcam) and 7

rabbit anti-CD2/AF488 conjugated antibody (bs-2899R, Bioss) for 3 hours at room 8

temperature. Nuclei were stained with DAPI. All manipulation must be away from light. The 9

images were captured under confocal microscopy (Leica, TCS SP5). 10

Cellular cytotoxicity study 11

The cellular toxicities of TPNs and manRTPNs on macrophages, T cells, FLSs and hFLSs 12

were determined by MTT assay using CCK-8 Kit. Cells were implanted in a 96-well culture 13

plate at a density of 8 × 103 cells/well (5 replicates each treated group). After 12 hrs, PBS, TPNs 14

and manRTPNs with concentrations of 10 nM, 50 nM, 100 nM, 200 nM, 500 nM were added 15

into medium for further 24 hrs co-culture. Then the cells were mildly washed by PBS before 16

added CCK-8 mixed solution. Then, the optical density in 450 nm was measured by Multiskan 17

GO (Thermo Fisher Scientific). 18

Cytokines concentration analysis 19

The TNF-α, IL-6 and IFN-𝛾 concentrations of cells culture supernatants were determined 20

with ELISA Kit using mouse ELISA Kits (Biolegend, USA) as manipulation. Also, the TNF-α 21

24

and IL-6 concentrations of mouse serum samples were determined with TNF-α and IL-6 ELISA 1

Kits (Biolegend, USA). Briefly, Serum samples of treated mice were collected at the end of 2

experiment and concentrations of TNF-α and IL-6 were quantified with ELISA kits. 3

Specifically, the whole blood of mice was collected by submandibular bleeding into microtubes 4

and allowed to clot at room temperature for 30 min. Samples were then centrifuged at 2,000g 5

for 6 min to collect supernatant serum. Serum samples were immediately frozen at -20 °C until 6

analysis by using mouse TNF-α and mouse IL-6 ELISA kits (Biolegend) within 3 days of 7

collection. 8

Genes expression profiles analysis of cells 9

After treated with nanoparticles of serious concentrations, cells were harvested to isolate 10

RNA. RNA was extracted with TRIzol reagent (Invitrogen, USA) according to the 11

manufacturer’s instructions. Complementary DNAs (cDNAs) were synthesized with 12

QuantiTect Reverse Transcription Kit (QIAGEN K.K., Tokyo, Japan) and the specific gene 13

transcripts were quantified by quantitative real-time PCR using QuantiTect SYBR Green PCR 14

Kit (QIAGEN K.K., Tokyo, Japan) and analyzed with ABI 7500 real-time PCR system 15

(Applied Biosystems, USA). The gene expression patterns along with concentration were 16

highlight with heatmap by using TBtools package. 17

Collagen induced arthritis (CIA) mice model induction 18

Seventy-two female DBA/1 mice (6∼8 weeks old) were purchased from Vital River 19

Laboratory Animal Technology Co. Ltd (Beijing, China). All mice were maintained in a SF 20

system which is room equipped with an air-filtering system, and the cages and water were 21

25

sterilized. CIA was induced as previously reported study[39]. Briefly, bovine type II collagen 1

(Chondrex, Redmond, WA, USA) was dissolved in 0.1M acetic acid overnight at 4°C. This was 2

emulsified in an equal volume of complete Freund’s adjuvant (Chondrex, Redmond, WA, 3

USA). The mice were immunized intradermal at the base of the tail with 100 𝜇L of emulsion 4

containing 100 𝜇g of type II collagen. On day 21, mice were boosted intraperitoneal with 100 5

𝜇g type II collagen dissolved in phosphate buffered saline (PBS). Clinical scores of each 6

hindpaws and forepaws of mice were obtained following the standard evaluation process[39]. 7

In detail, no evidence of erythema and swelling occurred is score 0, erythema and mild swelling 8

appeared is score 1, erythema and mild swelling extended from the ankle to the tarsals is score 9

2, erythema and moderate swelling extended from the ankle to metatarsal joints is score 3 and 10

erythema and severe swelling encompassed the ankle, paws, and digits or ankyloses of the limb 11

is score 4. Then total clinical scores of hindpaws and forepaws were calculated. At least six 12

mice each group. 13

Nanoparticles targeting to inflamed joints 14

After CIA model was established, free Cy5.5, Cy5.5-labelled TPNs and manRTPNs were 15

intravenously injected into the CIA mice (20 mg/kg, n = 4). Fluorescence images were obtained 16

after 1, 2, 4, 8, 12, and 24 hours using an in vivo imaging system (IVIS, Caliper, USA). Then, 17

the mice were sacrificed for ex vivo tissue/organs distribution analysis at the end of experiment. 18

Meanwhile, the synovial tissue was dissected and dehydrated in 30% sucrose solution overnight 19

and then cry-sectioned at 8 𝜇m. The sections were stained with DAPI and washed twice. 20

26

Fluorescence images were obtained using a confocal microscope (Leica, TCS SP5). The 1

fluorescence intensity of sections in different groups was calculated using software Image J. 2

manRTPNs treatment assay 3

According to CIA mice model conduction, starting at day 21, the clinical scores were 4

calculated to make sure that CIA mice were almost equably and randomly divided into four 5

groups, while healthy DBA/1 mice were set as control (n = 5). Five different agents namely 6

PBS, RBC vesicle, triptolide, TPNs and manRTPNs were intravenously injected every two 7

days. The timeline of whole therapy experiment from model making to the end was according 8

to schematic diagram shown in figure 5A. The arthritis index of each group was recorded over 9

time to evaluate treatment efficacy. At the end, the mice were sacrificed for further analysis. 10

Micro-CT imaging of mice paws 11

The knee and ankle joints of each experimental mouse hindpaws were scanned using 12

micro-CT system (Bruker's SkyScan 1278, UK) to observe joint bone situation. Images were 13

acquired at 55 kV, 72 μA and 300 ms/frame, with 360 views. The 3D structures of knee joint 14

and ankle joint were reconstructed from 360 views and BDM was evaluated by corresponding 15

SkyScan NRecon package. Analysis typically took around 15 min per joint for an experienced 16

operator. 17

Histological analysis and immunohistochemical staining assay 18

At study endpoints, mice were euthanized and hind knee joints were collected for H&E 19

staining and immunohistochemical staining. Briefly, the ankle and knee joints of scarified 20

animals as well as their heart, liver, spleen, lung and kidney were fixed in 10% buffered 21

27

formalin and then joints were incubated in decalcifying solution (4% hydrochloric acid in 4% 1

formaldehyde) at room temperature for 7 days for decalcification. After paraffinization, 2

microtome (Leica) slices of 8 μm were prepared and stained with haematoxylin and eosin, 3

images were taken by using inverted microscope (Olympus, IX71, Japan) and the inflammatory 4

cell infiltration in synovial tissues, bone and cartilage was evaluated by Image J software. After 5

deparaffinization the slices were subjected to antigen recovery in 0.01 M sodium citrate buffer 6

at 125 °C for 30 s, followed by 10 s at 90 °C, and then subjected to the endogenous peroxidase 7

inactivation by covering tissue with 3% hydrogen peroxide for 5 min. After blocking non-8

specific binding sites with 10% goat serum in PBS, the slices were incubated with TNF-α/IL-6 9

monoclonal antibody (Invitrogen) with a dilution rate of 1:100, respectively, at 4 °C for 24 h. 10

Then the slices were incubated with horseradish peroxidase (HRP)-conjugated secondary 11

antibody with a dilution rate of 1:800 at 37 °C for 1 hour. Sections were developed using the 12

DAB substrate and then counterstained with haematoxylin. The images were captured and 13

analyzed by inverted microscope (Olympus, IX71, Japan) and Image J software. 14

Gene expression profiles analysis of tissues 15

The paws and ankles were dissected from mice on day 22 of arthritis, snap-frozen in liquid 16

nitrogen, ground into powder, and homogenized. All procedure must be under RNase-free 17

conditions. The RNA isolation and real-time PCR assay were carried out following the protocol 18

following. Briefly, total RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA, 19

USA) from the tissue homogenates according to the manufacturer’s instructions. The total RNA 20

(1 𝜇g) was reverse transcribed to cDNA using the QuantiTect Reverse Transcription Kit 21

28

(QIAGEN, Japan) and the specific gene transcripts were quantified by quantitative real-time 1

PCR using QuantiTect SYBR Green PCR Kit (QIAGEN K.K., Tokyo, Japan) and analyzed 2

with ABI 7500 real-time PCR system (Applied Biosystems, USA). PCR was performed as 40 3

cycles at 94°C for 15 s, 55°C for 30 s, and 72°C for 30 s. The relative RNA expression was 4

calculated with comparative 𝐶𝑇 method; 𝛽-actin as internal ontrol. The gene expression 5

patterns along with concentration were highlight with heatmap by using TBtools package. 6

Gene-specific primers were synthesized by Sangon Biotech (Shanghai) Co., Ltd. and list in the 7

TableS1. 8

Statistical Analysis 9

All the results are reported as mean ± SD. The differences among groups were analyzed 10

using one-way ANOVA analysis and Student’s t-test; *represents P < 0.05, ** represents P < 11

0.01. 12

Acknowledgments 13

This work was supported by National Natural Science Foundation of China (81801838, 14

31571013, 81701816 and 81601552), Guangdong Natural Science Foundation of Research 15

Team (2016A030312006), K.C. Wong Education Foundation (GJTD-2018-14), Natural 16

Science Foundation of Guangdong Province (2018A030313013), Shenzhen Science and 17

Technology Program (JCYJ20180302145912832, JCYJ20160429191503002, 18

JCYJ20170818162259843 and JCYJ20170818163739458). We thank Pro. Hongchang Li and 19

Dr. Yifan Ma of SIAT for their useful and meaningful recommendations to improve this paper, 20

and thank Mr. Danhui Quan for his help in some drawing. 21

29

Competing Interests 1

The authors have declared that no competing interest exists. 2

3

30

1

Scheme 1. Schematic illustration of mannose modified erythrocyte membranes-coated 2

triptolide nanoparticle (manRTPNs) targeting to inflammatory sites of rheumatoid arthritis and 3

significant therapeutic effect. Targeting capacity to macrophages and T cells of manRTPNs is 4

conferred from the mannose modified erythrocyte membrane. Specifically, target activated T 5

cells was through the CD58-CD2 affinity, and the binding interaction between the mannose and 6

its receptor CD206 enhanced anchoring of manRTPNs to activated macrophages. Effective 7

management was achieved through remodeling inflammation microenvironment. Basically, 8

TPNs at RA sites shift core genes expression in three pathways including angiogenesis, bone 9

resorption and cytokine secretion. 10

11

31

1

Figure 1. Preparation and Characterization of manRTPNs. 2

(A) Synthetic route illustration of triptolide self-assembling nanoparticles (TPNs). (B) 3

Representative image of TPNs and manRTPNs examined with transmission electron 4

microscopy (TEM). manRTPNs samples were stained with uranyl acetate. Scale bar, 200 nm. 5

32

(C) Size and ζ potential distribution of TPNs, RBC vesicles and manRTPNs. Each experiment 1

had three replicates. (D) Stability of TPNs and manRTPNs in PBS over five weeks. Each 2

experiment had three replicates. (E) Pharmacokinetics of TPNs and manRTPNs in vivo. Each 3

experiment point had three replicates. 4

5

33

1

Figure 2. Targeting capacity to immune cells of manRTPNs. 2

(A-C) Cellular uptake capacity of TPNs and manRTPNs in three cell lines. Flow cytometry 3

analysis of TPNs and manRTPNs were conducted in RAW246.7 (A), CTLL-2 (B) and RSC364 4

(C). k was slope of uptake curve built with mean fluorescence intensity (MFI) which indicated 5

cell uptake rate. (D-F) Representative fluorescence images by confocal analysis. RAW246.7 6

(D), CTLL-2 (E), and RSC364 (F) were co-incubated with TPNs, RTPNs, and manRTPNs, 7

respectively. Nanoparticles were labeled with Cy5.5 (red), CD206/CD2 antibody (green) 8

34

expressed on the IL-4 -treated RAW264.7 and TNF-α treated CTLL-2 and cellular nuclear was 1

labeled with DAPI (blue). Scale bars, 100 μm. (G-I) Flow cytometry analysis of RAW246.7 2

(G), CTLL-2 (H) and RSC364 (I) incubated with various nanoparticles. ** indicated significant 3

difference and p < 0.05. 4

5

35

1

2

Figure 3. Efficacy of manRTPNs in inflammatory cells. 3

(A-D) Cytotoxicity of TPNs and manRTPNs in RAW264.7 (A), CTLL-2 (B), RSC364 (C), and 4

hFLS-RA (D). Each experiment point had three replicates. (E) Heatmap of inflammatory 5

factors genes expression. Briefly, genes cox-2, timp-1 and NF-κB were checked in RAW264.7, 6

gene rankl was checked in CTLL-2 and genes vegf, mmp-9 and mmp-2 were checked in 7

RSC364, respectively. (F-J) Cytokines concentration variation of inflammatory cells treated 8

with TPNs and manRTPNs. Supernatant IL-1𝛽 was measured in RWA264.7 (F) and RSC364 9

(I), supernatant IL-6 in RWA264.7 (G) and RSC364 (J), supernatant IFN-𝛾 in CTLL-2 (H), 10

each experiment had three replicates. 11

12

36

1

2

Figure 4. Targeting effect and accumulation at inflamed site of RA of manRTPNs. 3

(A) Representative images of free Cy5.5, TPNs and manRTPNs accumulating in arthritic 4

paws. (B) Fluorescence images of free Cy5.5, TPNs and manRTPNs in arthritic paws fixed 5

after 24 hours of intravenous injection. Scale bars, 500 μm. Fluorescence of free Cy5.5, TPNs 6

and manRTPNs was quantified in arthritic paws sections. (C) Distribution of free Cy5.5, 7

TPNs and manRTPNs in main organs and arthritic paws. 8

37

1

38

1

Figure 5. Therapeutic effect of manRTPNs in a mouse model of collagen-induced arthritis. 2

(A) The study protocol of a therapeutic regimen with CIA mice model. (B) Arthritis indexes of 3

different treatment groups over 14 days; each experiment had five mice replicates. (c) Bone 4

mean density (BMD) variation of the mice paw joints before and after different treatment. (D) 5

Representative 3D reconstructed bone images of arthritic paws in different treatment groups. 6

(E) Representative images of H&E staining on knee sections from healthy mice and CIA mice 7

treated with PBS, RBC vesicle, triptolide, TPNs or manRTPNs. Scale bars, 100 μm. 8

9

39

1

Figure 6. Remodeling inflammatory factor of RA treatment by manRTPNs. 2

(a) Representative images of TNF-α and IL-6 immunohistochemical staining on knee sections. 3

Scale bars, 100 μm. (B) Immunohistochemical quantification of TNF-α and IL-6. (C) Serum 4

concentration of TNF-α and IL-6 in healthy mice and CIA mice treated with different groups. 5

40

Each experiment point had three replicates. (D) Heatmap of genes expression profiles in 1

different treatment. 2

3

41

1

2

Figure 7. Biocompatibility and safety of manRTPNs. 3

(A) Complete cell counts (white blood cells, red blood cells, and platelets) in PBS and 4

triptolide-, TPNs-, manRTPNs-treated groups. Experiment replicate, three. (B) Complete cell 5

counts (neutrophil, lymphocyte, and monocyte) in PBS and triptolide-, TPNs-, manRTPNs-6

treated groups. Experiment replicate, three. (C) Concentration of HGB and MCHC in four 7

42

groups indicated before. Experiment replicate, three. (D-G) Blood routine examination 1

parameters including AST(D), LDH-L(E), BUN(F) and Cre(G) after different treatment. 2

Experiment replicate, three. (H) H&E staining of main organs in four different groups. Scale 3

bars, 100 μm. 4

5

43

1. Scott DL, Wolfe F, Huizinga TWJ. Rheumatoid arthritis. Lancet. 2010; 376: 1094-1108. 1

2. Van Vollenhoven RF. Treatment of rheumatoid arthritis: state of the art 2009. Nat Rev 2

Rheumatol. 2009; 5: 531-541. 3

3. Bykerk V. Unmet needs in rheumatoid arthritis. J Rheumatol. 2009; 36: 42-46. 4

4. Salliot C, Van Der Heijde D. Long-term safety of methotrexate monotherapy in patients 5

with rheumatoid arthritis: a systematic literature research. Ann Rheum Dis. 2009; 68: 1100-6

1104. 7

5. Alcorn N, Saunders S, Madhok R. Benefit-risk assessment of leflunomide an appraisal of 8

leflunomide in rheumatoid arthritis 10 years after licensing. Drug Safety. 2009; 32: 1123-9

1134. 10

6. Leombruno JP, Einarson TR, Keystone EC. The safety of anti-tumour necrosis factor 11

treatments in rheumatoid arthritis: meta and exposure-adjusted pooled analyses of serious 12

adverse events. Ann Rheum Dis. 2009; 68: 1136-1145. 13

7. Hyrich KL, Watson KD, Isenberg DA, Symmons DPM, Register BB. The british society 14

for rheumatology biologics register: 6 years on. Rheumatology. 2008; 47: 1441-1443. 15

8. Kirwan JR, Bijlsma JWJ, Boers M, Shea BJ. Effects of glucocorticoids on radiological 16

progression in rheumatoid arthritis. Cochrane Db Syst Rev. 2007. 17

9. Ravindran V, Rachapalli S, Choy EH. Safety of medium- to long-term glucocorticoid 18

therapy in rheumatoid arthritis: a meta-analysis. Rheumatology. 2009; 48: 807-811. 19

44

10. Zhang QZ, Dehaini D, Zhang Y, Zhou JL, Chen XY, Zhang LF, et al. Neutrophil 1

membrane-coated nanoparticles inhibit synovial inflammation and alleviate joint damage 2

in inflammatory arthritis. Nat Nanotechnol. 2018; 13: 1182-1190. 3

11. Li RX, He YW, Zhu Y, Jiang LX, Zhang SY, Qin J, et al. Route to rheumatoid arthritis by 4

macrophage-derived microvesicle-coated nanoparticles. Nano Letters. 2019; 19: 124-134. 5

12. Liang HY, Peng B, Dong C, Liu LX, Mao JJ, Wei S, et al. Cationic nanoparticle as an 6

inhibitor of cell-free DNA-induced inflammation. Nat Commun. 2018; 9: 4291. 7

13. Tao X, Cush J, Garret M, Pe. L. A phase I study of ethylacetate extract of the Chinese 8

antirheumatic herb Tripterygium wilfordii Hook F in rheumatoid arthritis. J Rheumatol. 9

2001; 28: 2167. 10

14. Fan DP, Guo QQ, Shen JW, Zheng K, Lu C, Zhang G, et al. The effect of triptolide in 11

rheumatoid arthritis: from basic research towards clinical translation. Int J Mol Sci. 2018; 12

19. 13

15. Pao HP, Liao WI, Wu SY, Hung KY, Huang KL, Chu SJ. PG490-88, a derivative of 14

triptolide, suppresses ischemia/reperfusion-induced lung damage by maintaining tight 15

junction barriers and targeting multiple signaling pathways. Int Immunopharmacol. 2019; 16

68: 17-29. 17

16. Chen YK, Zhang L, Ni JS, Wang XY, Cheng J, Li YC, et al. LLDT-8 protects against 18

cerebral ischemia/reperfusion injury by suppressing post-stroke inflammation. J Pharmacol 19

Sci. 2016; 131: 131-137. 20

45

17. He QL, Minn I, Wang QL, Xu P, Head SA, Datan E, et al. Targeted delivery and sustained 1

antitumor activity of triptolide through glucose conjugation. Angew Chem Int Edit. 2016; 2

55: 12035-12039. 3

18. Yuan ZX, Wu XJ, Mo JX, Wang YL, Xu CQ, Lim LY. Renal targeted delivery of triptolide 4

by conjugation to the fragment peptide of human serum albumin. Eur J Pharm Biopharm. 5

2015; 94: 363-371. 6

19. Mei ZN, Chen HB, Weng T, Yang YJ, Yang XL. Solid lipid nanoparticle and 7

microemulsion for topical delivery of triptolide. Eur J Pharm Biopharm. 2003; 56: 189-8

196. 9

20. Huang CL, Zeng T, Li JW, Tan LS, Deng XL, Pan YC, et al. Folate receptor-mediated 10

renal-targeting nanoplatform for the specific delivery of triptolide to treat renal 11

ischemia/reperfusion injury. ACS Biomater Sci Eng. 2019; 5: 2877-2886. 12

21. Ling D, Xia H, Park W, Hackett MJ, Song C, Na K, et al. pH-sensitive nanoformulated 13

triptolide as a targeted therapeutic strategy for hepatocellular carcinoma. ACS Nano. 2014; 14

8: 8027-8039. 15

22. Xu HT, Liu B. Triptolide-targeted delivery methods. Eur J Med Chem. 2019; 164: 342-16

351. 17

23. Abbas M, Zou QL, Li SK, Yan XH. Self-assembled peptide- and protein-based 18

nanomaterials for antitumor photodynamic and photothermal therapy. Adv Mater. 2017; 19

29. 20

46

24. Zhu PL, Yan XH, Su Y, Yang Y, Li JB. Solvent-induced structural transition of self-1

assembled dipeptide: from organogels to microcrystals. Chem-Eur J. 2010; 16: 3176-3183. 2

25. Yan XH, Zhu PL, Li JB. Self-assembly and application of diphenylalanine-based 3

nanostructures. Chem Soc Rev. 2010; 39: 1877-1890. 4

26. Elshabrawy HA, Chen ZL, Volin MV, Ravella S, Virupannavar S, Shahrara S. The 5

pathogenic role of angiogenesis in rheumatoid arthritis. Angiogenesis. 2015; 18: 433-448. 6

27. Kinne RW, Brauer R, Stuhlmuller B, Palombo-Kinne E, Burmester GR. Macrophages in 7

rheumatoid arthritis. Arthritis Res. 2000; 2: 189-202. 8

28. Davis SJ, Ikemizu S, Evans EJ, Fugger L, Bakker TR, Van Der Merwe PA. The nature of 9

molecular recognition by T cells. Nat Immunol. 2003; 4: 217-224. 10

29. Raychaudhuri S, Thomson BP, Remmers EF, Eyre S, Hinks A, Guiducci C, et al. Genetic 11

variants at CD28, PRDM1 and CD2/CD58 are associated with rheumatoid arthritis risk. 12

Nat Genet. 2009; 41: 1313-1318. 13

30. Sable R, Durek T, Taneja V, Craik DJ, Pallerla S, Gauthier T, et al. Constrained cyclic 14

peptides as immunomodulatory inhibitors of the CD2:CD58 protein-protein interaction. 15

ACS Chemical Biology. 2016; 11: 2366-2374. 16

31. Hu CMJ, Fang RH, Copp J, Luk BT, Zhang LF. A biomimetic nanosponge that absorbs 17

pore-forming toxins. Nat Nanotechnol. 2013; 8: 336-340. 18

32. Rossi L, Fraternale A, Bianchi M, Magnani M. Red blood cell membrane processing for 19

biomedical applications. Front Physiol. 2019; 10. 20

47

33. He QL, Titov DV, Li J, Tan MJ, Ye ZH, Zhao YM, et al. Covalent modification of a 1

cysteine residue in the XPB subunit of the general transcription factor TFIIH through single 2

epoxide cleavage of the transcription inhibitor triptolide. Angew Chem Int Edit. 2015; 54: 3

1859-1863. 4

34. Hu CMJ, Zhang L, Aryal S, Cheung C, Fang RH, Zhang LF. Erythrocyte membrane-5

camouflaged polymeric nanoparticles as a biomimetic delivery platform. P Natl Acad Sci 6

USA. 2011; 108: 10980-10985. 7

35. Fang RNH, Hu CMJ, Chen KNH, Luk BT, Carpenter CW, Gao WW, et al. Lipid-insertion 8

enables targeting functionalization of erythrocyte membrane-cloaked nanoparticles. 9

Nanoscale. 2013; 5: 8884-8888. 10

36. Shao F, Wang GJ, Xie HT, Zhu XY, Sun JG, A JY. Pharmacokinetic study of triptolide, a 11

constituent of immunosuppressive chinese herb medicine, in rats. Biol Pharm Bull. 2007; 12

30: 702-707. 13

37. Ghosh S, Hayden MS. New regulators of NF-kappaB in inflammation. Nat Rev Immunol. 14

2008; 8: 837-848. 15

38. Niedermeier M PT, Korb A. Therapeutic opportunities in fibroblasts in inflammatory 16

arthritis. Best Pract Res Clin Rheumatol. 2010; 24: 527-540. 17

39. Brand DD, Latham KA, Rosloniec EF. Collagen-induced arthritis. Nat Protoc. 2007; 2: 18

1269-1275. 19

40. Noack M, Miossec P. Selected cytokine pathways in rheumatoid arthritis. Seminars in 20

Immunopathology. 2017; 39: 365-383. 21

48

41. Geusens P. The role of RANK ligand/osteoprotegerin in rheumatoid arthritis. Ther Adv 1

Musculoskelet Dis. 2012; 4: 225-233. 2

42. Chen L, Deng H, Cui H, Fang J, Zuo Z, Deng J, et al. Inflammatory responses and 3

inflammation-associated diseases in organs. Oncotarget. 2018; 9: 7204-7218. 4

5

download fileview on ChemRxivmanuscript-update 20200804.pdf (2.67 MiB)