siRNA delivery with chitosan: influence of chitosan ... - ACS Publications

EGFP knockdown with Pearson product moment correlation coefficient (PPMCC) ...... efficiency of our system to silence a target gene in PTECs, and sugg...
0 downloads 0 Views 2MB Size
Subscriber access provided by READING UNIV

Article

siRNA delivery with chitosan: influence of chitosan molecular weight, degree of deacetylation and amine to phosphate ratio on in vitro silencing efficiency, hemocompatibility, biodistribution and in vivo efficacy Mohamad Gabriel Alameh, Marc Lavertu, Nicolas Tran-Khanh, Chi-Yuan Chang, Frédéric Lesage, Martine Bail, Vincent Darras, Anik Chevrier, and Michael D. Buschmann Biomacromolecules, Just Accepted Manuscript • DOI: 10.1021/acs.biomac.7b01297 • Publication Date (Web): 06 Dec 2017 Downloaded from http://pubs.acs.org on December 7, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biomacromolecules is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

Title: siRNA delivery with chitosan: influence of chitosan molecular weight, degree of deacetylation and amine to phosphate ratio on in vitro silencing efficiency, hemocompatibility, biodistribution and in vivo efficacy Authors: Mohamad Alameh2, Marc Lavertu1,2, Nicolas Tran-Khanh1, Chi-Yuan Chang1, Frederic Lesage3, Martine Bail1, Vincent Darras1, Anik Chevrier1 and Michael D. Buschmann1,2†**. Affiliation: 1 Polytechnique Montreal, Department of Chemical engineering, Montreal, QC, Canada. 2 Polytechnique Montreal, Institute of Biomedical engineering, Montreal, QC, Canada. 3 Polytechnique Montreal, Department of Electrical engineering, Montreal, QC, Canada **Corresponding

author.

Email

address:

[email protected]

or

[email protected] (M.D. Buschmann) Present address

1 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 64

†Michael D. Buschmann, Dept. of Bioengineering, Volgenau School of Engineering, George Mason University, 4400 University Drive, MS 1G5, Fairfax, VA 22030, USA

Abstract: Chitosan (CS) shows in vitro and in vivo efficacy for siRNA delivery, but with contradictory findings for incompletely characterized systems. To understand which parameters produce effective delivery, a library of precisely characterized chitosans was produced at different DDAs and Mn. Encapsulation and transfection efficiencies were characterized in vitro. Formulations were selected to examine the influence of Mn and N:P ratio on nanoparticle uptake, metabolic activity, genotoxicity, and in vitro transfection. Hemocompatibility and in vivo biodistribution were then investigated for different Mn, N:P ratio, and dose. Nanoparticle uptake and gene silencing correlated with increased surface charge, which was obtained at high DDA and high Mn. A minimum polymer length of ~60-70 monomers (~10kDa) was required for stability and knockdown. In vitro knockdown was equivalent to lipid control with no metabolic or genotoxicity. An inhibitory effect of serum on biological performance was dependent on DDA, Mn and N:P. In vivo biodistribution in mice show accumulation of nanoparticles in kidney, with 40-50% functional knockdown.

Key words: Chitosan, siRNA, Degree of deacetylation, Molecular weight, Invivofectamine, Biodistribution.

2 ACS Paragon Plus Environment

Page 3 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

Introduction 1

Oligonucleotide (ON) therapeutics represent a novel class of molecules designed to modulate gene

2

expression through direct interference with ribonucleic acids (RNA) or proteins1. Clinical translation

3

depends on the efficient delivery and cellular uptake of large and negatively charged ON2. Advances in

4

understanding ON biology, mechanism of action, physicochemistry and their interactions with molecular

5

machines (i.e. RNAi Inducing Silencing Complex) and/or sensors (i.e. Toll-Like Receptors) has allowed

6

the introduction of design rules and chemical modifications that improve nuclease resistance and reduce

7

immune activation and sequence dependent off-target effects. Although modifications improve serum

8

half-life (t 1/2 ) and promote siRNA binding to proteins, intracellular translocation to the pharmacological

9

site of action and renal elimination cannot be circumvented by this strategy nor can improvement of

10

sequence design. As a consequence, chemical conjugations with ligands or encapsulation in delivery

11

systems constitute the two predominant strategies used in clinical development of siRNA delivery.

12

Trivalent N-acetylgalactosamine (GalNac) conjugation on the sense strand results in highly potent

13

hepatocyte targeted siRNA able to escape endosomal compartments and achieve meaningful knockdown

14

in phase II/III clinical trials 2. The conjugation approaches continue to demonstrate potency in pre-clinical

15

and clinical studies but face serious challenges with the recent discontinuation of the Revusiran and ARC-

16

520/521 programs 3, 4.

17

Encapsulation of siRNA into delivery systems physically protects the siRNA from serum nucleases,

18

increases bioavailability and allows efficient delivery to certain targeted organs and cells using lipid

19

nanoparticles (LNP) that have reached clinical trials 3, 5, 6. However, LNPs are associated with serious side

20

effects such as immune activation and their clinical administration is preceded by and/or accompanied

21

with prophylactic anti-inflammatory steroids

22

to the above mentioned issues but also include their limited capability to deliver nucleic acid cargos

23

beyond the liver 9. Therefore, delivery systems that meet criteria such as colloidal stability, high

7, 8

. The shortcomings of LNP based vectors are not limited

3 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 64

24

encapsulation efficiency, low toxicity, reduced renal clearance, and deliver siRNA efficiently to

25

extrahepatic organs are critically needed.

26

Chitosan (CS) is a family of cationic bio-copolymers composed of β (1-4) linked N-acetyl glucosamine

27

(GlcNAc) and D-glucosamine (Glc) that has gained considerable attention for ON delivery. It is

28

characterized by low in vitro and in vivo toxicity, ease of production/chemical conjugation, and is

29

generally recognized as safe (GRAS) 10. CS can be fine-tuned to reach specific degrees of deacetylation

30

(DDA), or fractions of protonatable amine (charge), and average molecular weights (Mw or Mn). The

31

high degree of protonation of amino groups (NH 2 ) occurring at a pH below chitosan pKa (~6.5-6.9)

32

favors the spontaneous formation of nanosized polyelectrolyte complexes through electrostatic interaction

33

with polyanionic molecules such as ON.

34

Early work showed that transfection efficiency (TE) of plasmid-containing chitosan nanoparticles

35

depended on a fine equilibrium between chitosan tunable parameters of DDA, Mn and the molar ratio of

36

chitosan amine to plasmid phosphate (N:P) as well as other extrinsic factors such as pH and the presence

37

of serum proteins

38

vitro

39

siRNA delivery were performed using partly deacetylated (DDA ~ 80-85%) chitosan formulated at high

40

N:P ratio (>25)

41

such as premature dissociation, limited dosing, blood incompatibility and non-specific effects due to large

42

quantities of free excess chitosan. Although gene knockdown (KD) has been achieved, experimental

43

discrepancies, differences in chitosan sources, and lack of characterization rendered results inconclusive

44

in identifying optimal parameters for siRNA delivery 10. Reports correlating transfection efficiency (TE)

45

as a function of chitosan DDA, Mn, and N:P ratio have been contradictory 19-22, 24. Therefore a systematic

46

study of siRNA delivery with accurately characterized chitosans that investigates the effect of intrinsic

47

(DDA, Mn and N:P ratio) and extrinsic parameters (serum, pH, ionic strength and mixing conditions) on

48

cell uptake, TE, toxicity, genotoxicity, hemocompatibility and in vivo bio-distribution is urgently needed.

16-27

11-15

. Chitosan has also been used to deliver short double-stranded siRNAs both in

23, 27-32

and in vivo

. Most reports evaluating physicochemical parameters for efficient in vitro

21-26, 28

. Such formulations could pose significant practical problems for in vivo delivery

4 ACS Paragon Plus Environment

Page 5 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

49

To understand the correlation between nanoparticle physicochemical properties and

50

knockdown efficiency (KD), we produced a library of chitosans that were precisely characterized

51

by gel permeation chromatography for molecular weight and 1H NMR for DDA. The effect of

52

DDA, polymer length (Mn), mixing ratio (N:P) and pH was systematically examined and

53

correlated with nanoparticle physicochemical properties including size, surface charge,

54

encapsulation efficiency (EE), and in vitro delivery. Potent formulations were selected for further

55

characterization and tested in the presence of increased serum concentrations and correlated to

56

knockdown efficiency. In addition, off-target effects and nanoparticle mediated toxicity were

57

examined using a metabolic assay coupled with genotoxicity testing. We also show that selected

58

formulations may be intravenously administered at doses up to 14 mg/kg of chitosan, accumulate

59

in the proximal tubule epithelial cells (PTEC) and induce functional target knockdown in the

60

kidney.

61

5 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 64

62

Material and methods.

63

siRNA sequences and chitosan characterization

64

siRNA sequences were custom synthesized by Dharmacon Inc (GE Dharmacon, Lafayette,

65

CO, USA) except for the non-targeting siRNA (siNT) which was purchased as a predesigned

66

product from the same supplier (D-001710-01-50). All siRNA sequences used in vitro were

67

provided by the manufacturer in a lyophilized format following standard desalting. The anti-

68

EGFP siRNA sense sequence was 5’-GAC GUA AAC GGC CAC AAG UUC-3’ and the

69

antisense sequence was 3’-CGC UGC AUU UGC CGG UGU UCA-‘5, duplex Mw 13,360

70

g/mol. The siRNA sequence has been used in two chitosan-siRNA studies

71

siRNA sequence was modified at the 5’ end of the sense strand and purified by HPLC for in vivo

72

administration. For in vivo efficacy studies, the anti-GAPDH siRNA was purchased from Life

73

technologies as a predesigned Ambion® In Vivo GAPDH Positive control siRNA (Life

74

technologies, Burlington, ON, Canada). The anti-GAPDH sense sequence was 5’-GGU CAU

75

CCA UGA CAA CUU UTT-3’, duplex Mw 13400 g/mol. The position of the LNA

76

modifications, as well as other Silencer™ modifications, was not disclosed by the manufacturer.

23, 24

. The DY647

77

Chitosans (Table 1) with different DDAs were obtained from Marinard, (Laval, QC, Canada)

78

and depolymerized in our laboratory using nitrous acid to achieve specific number-average

79

molecular weight targets (Mn) of 5, 10, 40, 80 and 120 kDa (Table 1). Chitosan number and

80

weight-average molecular weights (Mn and Mw) were determined by gel permeation

81

chromatography (GPC) using a Shimadzu LC-20AD isocratic pump coupled with a Dawn

82

HELEOS II multi angle laser light scattering detector (Wyatt Technology Co, Santa Barbara,

83

CA), an Optilab rEX interferometric refractometer (Wyatt Technology Co), and two Tosoh

84

TSKgel (G6000PWxl-CP and G5000PWxl-CP; Tosoh Bioscience LLC, King of Prussia, PA) 6 ACS Paragon Plus Environment

Page 7 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

85

columns. Chitosans were eluted at pH 4.5 using an acetic acid (0.15 M)/sodium acetate (0.1

86

M)/sodium azide (4 mM) buffer. The injection volume was 100µL, the flow rate 0.8 mL min−1

87

and temperature 25°C. The dn/dc values for chitosan with a 92 and 80% DDA were determined

88

at 0.208 and 0.201 using a laser’s wavelength of 658 nm. The degree of deacetylation was

89

determined by 1H NMR as per in house published methods 33.

90 91

Table 1 Characterization of chitosans tested in this study 1.

92 93

Chitosan

DDA (%)

Mn (kDa)

Mw (kDa)

PdI

Dp

72-10

75.4

11.8

17.9

1.60

69

72-120

71.7

140.4

182.5

1.30

811

80-10

84.4

10.8

14.5

1.34

64

80-120

82.9

177.4

290.4

1.64

1,054

92-5

91.7

4.3

6.4

1.51

26

92-10

92.0

9.0

13.7

1.52

55

92-40

92.5

40.7

54.9

1.35

248

92-80

92.7

81.1

267.6

3.30

494

92-120

91.9

137.6

180.7

1.31

836

98-10

98.9

8.8

11.4

1.29

54

98-120

98.5

127.5

188.3

1.47

788

Preparation of nanoparticles by manual mixing 1

Different chitosans are denoted according to their chemical composition using the nomenclature [DDA-Mn] and are represented in the first column of the table. The degree of deacetylation (DDA) was determined by 1H NMR. The number and weight average molecular weight (Mn and Mw) were determined by gel permeation chromatography (GPC). The polydispersity index (PdI) was calculated as Mw/Mn. The degree of polymerization (Dp) or chain 𝐌𝐌 𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜 length was computed using the following equation 𝐃𝐃 = 𝐀𝐀𝐀𝐀𝐀𝐀𝐀 𝐦𝐦𝐦𝐦𝐦𝐦𝐦 𝐦𝐦𝐦𝐦𝐦 𝐦𝐦𝐦𝐦 𝐚𝐚 𝐬𝐬𝐬𝐬𝐬𝐬𝐬𝐬 𝐃𝐃𝐃

7

ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 64

94

Chitosans were dissolved overnight in nuclease free water (Life technologies, Burlington, ON,

95

Canada) and 1N HCl (Sigma-Aldrich, Oakville, ON, Canada), using a glucosamine to HCl ratio

96

of 1:1, to a final concentration of 5 mg/mL. The stock solutions were sterile filtered using a 0.22

97

µm PVDF filter and used to prepare solutions at specific N:P ratio by dilution in nuclease-free

98

water. Before complexation, siRNA was diluted to a working concentration of 0.1 mg/mL.

99

Unless otherwise stated, nanoparticles were formed by simple electrostatic complexation

100

following manual addition of chitosan to siRNA at a 1:1 ratio (v/v). The final volume never

101

exceeded 250 µL and chitosan was pipetted into siRNA. Nanoparticles were kept at room

102

temperature for 20-30 minutes before further use.

103

Determination of nanoparticle size and surface charge

104

Size and surface charge (ζ-potential) of nanoparticles were determined by Dynamic Light

105

Scattering (DLS) and Laser Doppler velocimetry using a ZetaSizer Nano ZS device (Malvern

106

Instruments Ltd, Malvern, UK). The scattering angle of the detector was fixed at 173o and

107

measurements were performed at 25oC using the viscosity of water as sample diluent.

108

Nanoparticles were diluted 1:4 and 1:8 using sterile filtered NaCl solution (20 or 150mM) before

109

determination of size and ζ-potential respectively. For data in Figure 1, size and ζ-potential were

110

measured immediately post incubation in buffer (~2.5 min for size and 5 min for ζ-potential). For

111

data in supplemental information Figure S1, size and ζ-potential were measured at the respective

112

time point post incubation in the buffer. The Smoluchowski equation was used to calculate ζ-

113

potential from the measured electrophoretic mobility. All measurements were done in duplicate

114

and replicated twice (N=3, n=6).

115 116

Encapsulation efficiency and siRNA release 8 ACS Paragon Plus Environment

Page 9 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

117

The encapsulation efficiency (EE) of siRNA was determined using the Quant-iT™ Ribogreen®

118

RNA reagent (Life technologies, Burlington, ON, Canada) to assess siRNA free in solution after

119

mixing with chitosan. Encapsulation efficiency (EE) was calculated as percentage fluorescence

120

intensity relative to non-formulated siRNA.

121

siRNA release at different pH

122

Nanoparticles were incubated in buffers with different pH, 25 mM MES (pH 6.5), 1X PBS (pH

123

7.2) or 1X TAE (pH 8.0) and aliquots were taken at 24 hours post incubation. Aliquots were

124

further diluted 1:200 in respective buffers, mixed with an equal volume of Quant-iT™

125

RiboGreen® reagent to detect free siRNA, incubated for 5 minutes at room temperature and

126

fluorescence measured using the TECAN Infinite® M200 PRO microplate system (Tecan

127

Systems, Mannedorf, Switzerland).

128

siRNA release in the presence of heparin

129

The effect of physiological concentrations of heparin on siRNA release was tested. Heparin

130

(Sigma Aldrich, Oakville, ON, Canada) solution at 1 mg/mL was prepared in nuclease free

131

water, sterile filtered and diluted to 5 µg/mL. Nanoparticles were prepared as mentioned above

132

and diluted 1:2 in heparin, incubated for 1h and then diluted 1:100 in Tris EDTA (TE 1X, pH

133

7.2) and a volume of 100 µL transferred into black plates (Corning, NY, USA). Quant-iT™

134

RiboGreen® reagent (1:200) was prepared in TE 1X (pH 7.2) and an equal volume (100µL) was

135

added to the nanoparticles. Plates were incubated on a rotary mixer for 5 minutes in the dark and

136

fluorescence measured using a TECAN Infinite® M200 PRO microplate system (Tecan Systems,

137

Mannedorf, Switzerland). The excitation and emission wavelengths were 480 and 530 nm

138

respectively. Naked siRNA, heparin alone, TE 1X, chitosan/heparin and siRNA/heparin were

139

used as controls to assess background, assay interference and as subtraction blank. 9 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 64

140

siRNA release in the presence of serum albumin

141

The effect of physiological concentrations of bovine serum albumin (BSA) on siRNA release

142

was tested. BSA stock solution (125 mg/mL) was prepared in nuclease free water, quantified

143

using the micro BCA assay (Life technologies, Burlington, ON, Canada) and diluted to reach 50

144

mg/mL respectively. The Quant-iT™ RiboGreen® assay was identical to the heparin competition

145

assay described above with the only difference that nanoparticles challenged with BSA were

146

diluted 1:2000 in TE 1 X (pH 7.2). This high dilution was used to overcome assay inhibition.

147

Naked siRNA, BSA alone, TE 1X, chitosan-BSA and siRNA-BSA were used as controls to

148

assess background, assay interference, and as subtraction blank.

149

Cell culture

150

The EGFP+ H1299 cell line (human lung carcinoma) was provided by Prof. Jorgen Kjems

151

(Aarhus University, Denmark) and has been used in several chitosan-siRNA studies 21, 23, 24, 26, 28.

152

This cell line was generated by Dr Anne Chauchereaux (Gustave Roussy Institute, Paris, France)

153

by transducing the parental H1299 cell line (ATCC, Manassas, VA, USA) with pd2EGFP-N1

154

(Clonetech, Mountain View, CA, USA). The plasmid encodes a modified version of EGFP with

155

a turnaround rate of 2 h (t 1/2 ~2h). Cells were cultured in RPMI-1640 (ATCC, Manassas, VA,

156

USA) supplemented with 10% heat inactivated FBS (Life technologies, Burlington, ON,

157

Canada), 1% GlutaMAX™ (Life technologies, Burlington, ON, Canada) and 500 µg/mL of

158

G418 antibiotic at 37oC in a 5% CO 2 environment.

159 160 161

In vitro transfection

10 ACS Paragon Plus Environment

Page 11 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

162

For target gene knockdown and siRNA uptake, cells were seeded in 24 well plates at a density

163

of 45,000 cells/well to reach ~75-80% confluence on the day of transfection. Prior to

164

transfection, cells were washed once with 500 µL of warm Ca2+/Mg2+ free PBS and replenished

165

with fresh RPMI-1640 which contained varying amounts of FBS (0-94%). Nanoparticles were

166

prepared as described above and a specific volume added to cells to reach the target siRNA

167

concentration 25-400 nM per well. Cells were incubated for either 4h (toxicity), 8h

168

(genotoxicity) or 48 hours (knockdown, toxicity, and genotoxicity). For the serum challenge

169

experiment, medium containing increasing amounts of FBS (0-94%) was aspirated and replaced

170

with complete RPMI-1640 medium 4 hours post transfection.

171

DharmaFect® 2-siRNA nanoparticles were prepared by diluting siRNA stock solution to 0.025

172

mg/mL (4 µM) in Opti-MEM® serum free media and complexed to the lipid component as per

173

manufacturer recommendation; a volume of 6.4 µL of DharmaFect® 2 in 153.6 µL Opti-MEM®

174

was used for complexation.

175

Assessment of EGFP knockdown and nanoparticle uptake using flow cytometry

176

EGFP knockdown was measured using a MoFLo™ flow cytometer (Beckman Coulter,

177

Mississauga, ON, Canada) equipped with a 488 nm argon laser for excitation (model ENTCII-

178

621, Coherent, Santa Clara, CA, USA) and a 510/20nm (FL1) band pass filter to detect

179

fluorescence. Nanoparticle uptake was measured on a BD FACSAria™ equipped with a 633 laser

180

and a 660/20nm (FL1) band pass filter to detect fluorescence of internalized DY647 labeled

181

siRNA. Before analysis, EGFP+ H1299 cells were washed twice with ice-cold Ca2+/Mg2+ free

182

PBS, trypsinized and suspended in complete RPMI-1640 medium. For uptake, cells were

183

incubated with Streptomyces griseus type III chitosanase as per Alameh et al

184

and trypsinization. A dot plot of forward scatter (FSC) versus side scatter (SSC) was created and

16

before washing

11 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 64

185

a collection gate established using the Summit 3.0 software (Beckman Coulter, Mississauga, ON,

186

Canada) to exclude cell debris, dead and aggregated cells. 20,000 events per sample were

187

recorded and non-transfected cells (untreated) were used to establish baseline EGFP expression

188

in terms of absolute fluorescence intensity (FI). EGFP knockdown in treated samples was

189

calculated as mean FI relative to non-transfected cells (KD (%) =

190

MIQE compliant quantitative real time PCR (qPCR)

mean FI (treated)

mean FI (untreated)

× 100).

191

The secondary structure of endogenous and target genes used in this study was in silico

192

assessed using the mFold freeware 34. The position of the primer and probe sequences relative to

193

the structure was determined prior to qPCR analysis. Before extraction, transfected EGFP+

194

H1299 cells were treated with Streptomyces griseus type III chitosanase (Sigma Aldrich,

195

Oakville, ON, Canada) as described in

196

transfection using the RNeasy® Plus Mini extraction kit (Qiagen, Toronto, ON, Canada) as per

197

manufacturer protocol. Following extraction, samples were digested with 2 µL TURBO™ DNA-

198

free kit (Life technologies, Burlington, ON, Canada) for 30 minutes to remove potential genomic

199

DNA contamination and then inactivated with 5 µL DNase inactivation reagent, spun down for

200

1.5 minutes at 10,000 g and the supernatant collected. Purity of total RNA was determined by

201

measuring A 230 /A 260 , A 260 /A 280 and A 340 using a Jenway spectrophotometer. The integrity of

202

total RNA (RIN) was determined using the Agilent 2100 Bioanalyzer system (Agilent

203

technologies, Mississauga, ON, Canada). Total RNA concentration was determined using the

204

Quant-iT™ RiboGreen® Assay (Life technologies, Burlington, ON, Canada). RiboGreen®

205

reagent and extracted RNA samples were diluted 1:200 and 1:100 respectively using the supplied

206

nuclease free TE 1X buffer (pH 8.0), mixed at 1:1 ratio (v/v) and a volume of 200 µL pipetted

207

into 96 well black plates (Corning, NY, USA). Plates were incubated at room temperature in the

16

. Total RNA extraction was performed 48 hours post

12 ACS Paragon Plus Environment

Page 13 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

208

dark for 5 minutes before measuring fluorescence using a TECAN Infinite® M200 PRO

209

microplate system (Tecan Systems, Mannedorf, Switzerland). The excitation and emission

210

wavelengths were 485 and 530 nm respectively and concentrations (ng/mL) were derived from a

211

ribosomal RNA (rRNA) standard curves prepared on the same plate as per manufacturer

212

recommendations. A total 1 µg of extracted RNA was reverse transcribed at 42oC for 60 min

213

using the SuperScript® VILO™ cDNA synthesis kit (Life technologies, Burlington, ON,

214

Canada). For qPCR, 10 ng cDNA was amplified using TaqMan Fast Advanced Master Mix®

215

(Life technologies, Burlington, ON, Canada) on a ABI HT-7900 real-time PCR system (Applied

216

Biosystems, Mississauga, ON, Canada). All reactions were performed in a 384 well plate in a

217

final volume of 10 µL. Quantitative real-time PCR (qPCR) was performed using the following

218

cycle conditions: 2-minute hold at 55 oC, 10-minute hold at 95 oC followed by 40 cycles at 95 oC

219

for 15 s and 60 oC for 1 min. Primer-probe efficiency was determined using the standard curve

220

method. For each assay tested, a 6 log 10-fold dilution curve was constructed by plotting the

221

quantification cycle (Cq) versus Log cDNA concentration and linearly regressed to fit the data.

222

The PCR reaction efficiency was estimated using E = 10(−slope) . The list of assays (primer-probe

223

1

pairs) used in this study and their respective efficiencies can be found in the supplementary

224

methods section. Reference gene stability was assessed using the geNorm statistical package

225

(Biogazelle NV, Zwijnaarde, Belgium) with a panel of 10 reference genes under 10 experimental

226

conditions including non-treated, lipid-treated, high and low N:P ratio, high and intermediate

227

DDA and high and low Mn). The experiment was replicated once and the M (stability parameter)

228

and V (pair wise variation of normalization factors) scores determined 35.

229 230 13 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 64

231

Assessment of nanoparticle toxicity using the alamarBlue® assay

232

The effect of nanoparticles on metabolic activity was measured using alamarBlue® (Life

233

technologies, Burlington, ON, Canada). Preliminary experiments were performed as per

234

manufacturer protocols to define optimal cell density, incubation time and assess assay

235

interference with chitosan. EGFP+ H1299 cells were seeded in 96 well plates (CellBIND®, Fisher

236

Scientific, Mississauga, ON, Canada) at a density of 5,000 cells/well 20-24 hours prior to

237

transfection. Cells were transfected as previously described at a final anti-EGFP siRNA

238

concentration of 100 nM. Metabolic activity was measured at 4 hours and 44 hours post

239

transfection by replacing medium over cells with 200 µL of complete RPMI-1640 medium

240

supplemented with 10% alamarBlue® reagent. Absorbance was measured at 570 and 600 nm

241

using a TECAN Infinite® M200 PRO microplate system (Tecan Systems, Mannedorf,

242

Switzerland) 4 hours after the addition of assay reagent. Metabolic activity of cells was evaluated

243

as the percentage reduction of alamarBlue® relative to untreated cells.

244 245

Assessment of nanoparticle genotoxicity using the comet or single cell gel electrophoresis assay.

246

The assessment of nanoparticle genotoxicity was performed using the Trevigen alkaline

247

cometAssay® kit (Trevigen Inc, Gaithersburg, MD, USA) at 8 and 48 hours post-transfection.

248

The 8-hour time point was assessed to mimic the alamarBlue® time point (4 h transfection + 4 h

249

incubation with reagent). EGFP+ H1299 cells were transfected as described above, trypsinized,

250

resuspended in complete RPMI-1640 media and counted using the Countess automated cell

251

counter system (Life technologies, Burlington, ON, Canada). Cells were centrifuged at 100 g for

252

3 minutes, and pellet suspended in ice cold PBS (Ca2+/Mg2+ free) at a concentration of 100,000

253

cells/mL. Agarose embedding, lysis, and electrophoresis were performed according to the 14 ACS Paragon Plus Environment

Page 15 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

254

manufacturer protocol. Slides were stained with 100 µL of (1:10,000) SYBR gold® nucleic acid

255

stain (Life technologies, Burlington, ON, Canada), dried at 37oC and imaged following

256

excitation at 488 nm using an Axiovert epi-fluorescent microscope (Carl Zeiss, Toronto, ON,

257

Canada). Images were analyzed using the Open Comet plugin 36 installed in the ImageJ freeware

258

(NIH, Besthada, CA, USA). The experiment was replicated once and at least 100 comets were

259

counted per experiment (N=2, n≥ 200).

260

Assessment of nanoparticle hemocompatibility at doses relevant for in vivo administration

261

Hemolytic and hemaglutination properties of nanoparticles were tested according to ASTM E2524 38

37

262

and Evani et al

263

protocol approval by the University Ethics Committee. Nanoparticles were prepared using an in-house

264

automated in-line mixing system 39 and freeze-dried (FD) in the presence of 0.83% w/v trehalose (Sigma-

265

Aldrich, Oakville, ON, Canada), and 5.8 mM histidine (pH 6.5) (Sigma-Aldrich, Oakville, ON, Canada).

266

FD samples were rehydrated to 12X using nuclease free water to reach the highest tested concentration

267

(or dose) and iso-osmolality then serially diluted using 10% w/v trehalose buffer (300 mOsm) to a final

268

siRNA concentration of 0.1, 0.25, 0.5, and 0.8 mg/mL. The plasma-free hemoglobin (PFH) in the blood

269

was measured at 0.49 mg/mL prior to initiating the assay. Total blood hemoglobin (TBH) was adjusted

270

with PBS to a concentration of 10 ± 1 mg/mL. Nanoparticles were mixed with PBS and diluted TBH

271

blood (TBHd) at a 1:7:1 volumetric ratio, with 100 µL of nanoparticles at the target concentration

272

pipetted into an Eppendorf tube containing 700 µL PBS and 100 µL of blood (TBHd 10 ± 1 mg/mL). For

273

colorimetric determination of hemolysis, samples (700 µL) were incubated for 3 h in a water bath at 37

274

°C and visually inspected every 30 min for nanoparticle flocculation, dispersion, sinking or floating. The

275

supernatant was collected following centrifugation at 800 g for 15 min and absorbance measured at 540

276

nm on a TECAN Infinite® M200 PRO microplate system (Tecan Systems, Mannedorf, Switzerland. A

277

four-parameter regression algorithm (4PL) was used to obtain the calibration curve required to calculate

278

the hemoglobin concentration in the supernatant of each sample (PFH sample). The percentage of

respectively. Human blood was collected on consenting and healthy donors following

15 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 64

280

hemolysis was computed as: Hemolysis (%) = 100 × (PFHsample ⁄TBHd). For hemagglutination, the

281

visualized using an Axiovert light microscope and the area covered by red blood cells (RBCs) estimated

282

and scored.

279

remaining 200 µL of each sample prepared above were pipetted in 96 well assay plates, incubated for 3 h,

283

In vivo biodistribution and efficacy studies

284

All in vivo experiments described in this manuscript were randomized double blinded and

285

approved by the University of Montreal Ethics Committee (CDEA) and the Montreal Heart

286

Institute Research Center Ethics Committee. Mice were purchased from Charles Rivers (Charles

287

River, Quebec, Canada), housed and acclimatized in a specific pathogen free facility with

288

unrestricted access to water and food. Mice had body condition scores (BCS) of 340 and their

289

body weights (BW) were in the range of 20-25 g at the time of injection. All injection volumes

290

were calculated as 10 µL/g of BW and injections performed within 10-15 seconds. Mice were

291

euthanized under anesthesia (mixture of 3% Forane™ and 20-80% oxygen-air vol/vol) by cardiac

292

puncture followed by cervical dislocation.

293 294

Determination of chitosan-siRNA biodistribution using ex-vivo whole organ imaging

295

Balb/c nude female (♀) mice aged 6 weeks and weighing 20-22 g were injected for the

296

biodistribution experiments All test articles i.e. Naked siRNA, Invivofectamine® 2.0 and

297

chitosan based nanoparticles formulated at N:P 5 (Mn 10, 40 and 120 kDa) were intravenously

298

injected (I.V.) at a dose of 0.5 mg/kg DY647 labeled siRNA. The DY647 fluorophore was injected

299

at a dose of 0.5 mg/kg. Mice were euthanized 4 hours post administration and immediately

300

perfused using PBS (1 X 20 mL) and 10% Neutral Buffer Formalin (NBF, 1 X 40 mL). Ex-vivo

301

imaging on collected organs was performed using a whole animal imaging system mounted with 16 ACS Paragon Plus Environment

Page 17 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

302

an EMCCD EM N2 camera (NUVU Cameras, Montreal, QC, Canada). Controls included PBS,

303

naked DY647 labeled siRNA, DY647 alone, and commercially available lipid control

304

Invivofectamine® 2.0 (Life technologies, Burlington, ON, Canada). The latter was prepared as

305

per manufacturer recommendation.

306

Determination of in vivo functional gene knockdown

307

Preparation, lyophilisation, reconstitution, and characterization of injected nanoparticles for in

308

vivo efficacy

309

Low (10kDa) and high (120kDa) molecular weight chitosans with a degree of deacetylation of

310

92 and 98% (Table 1) were dissolved as described above to a final concentration of 5 mg/mL.

311

The stock solutions were sterile filtered using a 0.22 µm PVDF filter (EMD Millipore,

312

Etobicoke, ON, Canada) and used to prepare solutions, containing 1% trehalose and 3.8 mM

313

histidine, at an N:P ratio of 5 by dilution in nuclease-free water, 4% w/v trehalose and 28 mM

314

histidine (pH 6.5). Before complexation, anti-GAPDH siRNA stock solutions (4 mg/mL) were

315

diluted to 0.2 mg/mL in the same buffer as chitosan. Nanoparticles were prepared at a final N:P

316

ratio of 5:1 using simple manual mixing. All nanoparticles were incubated for 30 min at room

317

temperature upon preparation before analyses or freeze-drying. Anti-GAPDH nanoparticles were

318

lyophilized under sterile conditions using a Laboratory Series Freeze-Dryer PC/PLC (Millrock

319

Technology, Kingston, NY, USA). An optimized 1-day cycle comprising the following program:

320

rapid cooling to 5 °C, 30 min hold, rapid cooling to -5 °C, 30 min hold, temperature decrease

321

from -5 to -40 oC, at a rate of 1 °C/min, 2 hours hold, initiation of primary drying for 10 h at -

322

32 oC and 60 mTorr; followed by secondary drying cycle at 60 mTorr, increase in shelf

323

temperature to 30°C, at rate of 0.2°C/min, and a 6 hours hold was used. Nanoparticle volumes of

324

3.84, or 7 mL were freeze-dried in 10 or 20 mL serum vials respectively, using 20 mm butyl17 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 64

325

lyophilisation stoppers. Samples were backfilled with argon, stoppered, crimped, and stored at

326

4°C until reconstitution. All FD samples were reconstituted at the animal facility to 10-times

327

their initial concentration using water for injection and their concentration adjusted by diluting

328

with a nearly-isotonic aqueous solution comprising 10% w/v trehalose, so that the desired dosage

329

(mg siRNA/kg animal body weight) would be reached upon injection of 10 µL of nanoparticles

330

per gram of animal body weight. Immediately after injection, reconstituted nanoparticles were

331

characterized for their size and polydispersity as described above.

332

In vivo efficacy and monitoring of clinical signs and body weight

333

Balb/c male (♂) mice aged 6-7 weeks and weighing 22-25g were used for the efficacy study (3

334

animals/group). Uncoated anti-GAPDH NPs were prepared as described above and administered

335

at 2.5 mg/kg every other day for a total of three injections and mice were sacrificed 72 hours

336

following the last injection. Naked anti-GAPDH siRNA (siGAPDH) was administered at 2.5

337

mg/kg following the same schedule. Clinical signs were determined for a period of 4-hours post-

338

administration of test articles and at euthanasia. The clinical signs were recorded by trained

339

personnel and qualified animal care technicians. Clinical signs were scored for body condition,

340

general aspect, natural behavior, and provoked behavior. Body weight was recorded prior to each

341

injection and at euthanasia using an Avery Berkel scale (Avery Berkel, Fairmont, MN, USA).

342

Body weight was expressed as percent change relative to the previous injection. Mice were

343

euthanized, under anesthesia, using cardiac puncture, followed by cervical dislocation, total

344

circulating blood volume (tCBV) and organs collected. tCBV was serum separated and

345

immediately stored at -80 oC, and organs split into halves and stored in LiqN and fixed in 10%

346

NBF before protein extraction and determination of GAPDH enzymatic activity.

347

In vivo assessment of functional knockdown using the KDalert® assay 18 ACS Paragon Plus Environment

Page 19 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

348

Following collection, organs were snap frozen in liquid nitrogen and stored at -80oC until use.

349

Frozen tissues were cut on dry ice, weighed (~20 mg), and disrupted using the TissueLyzer® II

350

system (Qiagen Inc, Toronto, ON, Canada). Tissues were disrupted using the 5 mm steel beads

351

(Qiagen Inc, Toronto, ON, Canada) under the following conditions: 2 x 30 Hz, 20 seconds per

352

cycle. Homogenized tissues were re-suspended in 750 µL of KDalert™ lysis buffer (Life

353

technologies, Burlington, ON, Canada), and incubated on ice for 30 minutes with inversions

354

every 10 minutes. Lysates were clarified by centrifugation (2270 g, 30 minutes, 4°C),

355

transferred to new tubes, and diluted (1:20), in KDalert™ lysis buffer. The standard curve was

356

prepared by diluting GAPDH stock solution (26 U/mL) with lysis buffer at a 1:100 ratio

357

(GAPDH: Lysis), followed by 2-fold serial dilutions from 1:5 to 1:320. Twenty microliters of

358

diluted samples, and standards were transferred into 96 well plates (Corning, NY, USA), and

359

180µL of the KDalert™ Master Mix (Life technologies, Burlington, ON, Canada) was pipetted

360

into each well. Plates were incubated for 15 minutes at room temperature and absorbance

361

measured at 610 (10) nm using a TECAN Infinite® F-500 microplate system (Tecan Systems,

362

Mannedorf, Switzerland). GAPDH activity in units (U) was computed from the standard curve

363

and normalized to total protein content (mg) of the lysate sample as determined using the BioRad

364

DC Protein Assay kit (Bio-Rad Laboratories, Mississauga, ON, Canada).

365

Assessment of in vitro knockdown and nanoparticle biodistribution using confocal microscopy

366

Visual confirmation of in vitro knockdown was performed using live cell imaging. EGFP+

367

H1299 cells were seeded at a density of 25,000 cells/well onto an 8-chamber Lab-Tek® (MatTek,

368

Ashland, MA, USA) and imaged in multitrack mode using a Zeiss LSM 510 META confocal

369

Axioplan 200 microscope (Carl Zeiss AG, Feldbach, Switzerland). For in vivo biodistribution

370

and subcellular localization of DY647 labeled siRNA, organs were cryosectioned (5 µm), actin 19 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 64

371

stained using AlexaFluor 546 phalloidin and mounted with ProLong Diamond Antifade

372

containing DAPI (Life technologies, Burlington, ON, Canada).

373

Statistical analysis

374

Data was collected and expressed as average ± standard deviation (stdev). Statistical analysis

375

was conducted using STATISTICA® 12.0 (Dell Statsoft, Tulsa, OK, USA), SigmaPlot® 13.0

376

(Systat software, San Jose, CA, USA) and GraphPad Prism® 7.0 (GraphPad Software Inc, La

377

Jolla, CA, USA) software packages. Unless otherwise stated, the General Linear Model,

378

One/Two-Factor ANOVA and multiple regression analysis were performed on collected data

379

with Dunnet and Tukey test used for multiple comparisons. The design of experiment module in

380

STATISTICA® 12.0 was used to generate full or fractional factorial designs and generate

381

multifactorial modeling. Data from the comet experiments was also subjected to non-parametric

382

analysis.

20 ACS Paragon Plus Environment

Page 21 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

383

Results

384

Chitosan DDA dictate nanoparticle uptake, target knockdown, ζ-potential and encapsulation

385

efficiency at physiological pH while increasing Mn and N:P ratios have positive effects.

386

A library of chitosans was produced, precisely characterized (Table 1), and assessed under

387

different experimental conditions (pH, ionic strength and presence/absence of serum) to

388

understand molecular parameters favoring adequate physicochemical properties (size, surface

389

charge, colloidal stability, and encapsulation efficiency) and efficient non-toxic in vitro

390

knockdown (target knockdown and metabolic toxicity). As illustrated in Figure 1, nanoparticle

391

size, measured 2.5 minutes post-incubation in buffer, increased with both polymer length (Mn)

392

and ionic strength. In both low and high ionic strength, size increased 2-3 -fold due to an

393

increase in Mn from 10 to 120 kDa. The effect of chitosan DDA and amine to phosphate ratio

394

(N:P) on size was minimal. Nanoparticle surface charge (ζ-potential) decreased with increasing

395

ionic strength, as expected due to salt-induced electrostatic screening. At low ionic strength, ζ-

396

potential increased with increased DDA, Mn and N:P ratio. As shown in Figure 1 C, and

397

confirmed using multiple regression analysis (Data not shown), DDA had the strongest effect on

398

ζ-potential followed by N:P ratio and Mn, respectively. Although ζ-potential was around 2-3 fold

399

lower at high ionic strength, the tendency of increased ζ-potential with a concomitant increase in

400

DDA, Mn and N:P ratio was conserved (Figure 1 C and D).

401

Given that DDA and the N:P ratio did not significantly alter the size, colloidal stability was

402

investigated using low and high Mn chitosan formulated at an N:P ratio of 5 (92-10-5 and 92-

403

120-5). As shown in Figure S1, nanoparticle size, and polydispersity was independent of Mn in

404

10 mM NaCl and stable up to at least 1-hour post complexation. Increasing the ionic strength to

405

150 mM had a significant impact on colloidal stability with nanoparticles aggregating rapidly to 21 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 64

406

reach the µm scale. The polydispersity index (PdI), a dimensionless measure of dispersion

407

around the mean, reached its maximum value of 1 in 150 mM NaCl around 15 min post

408

complexation indicating severe aggregation/high heterogeneity. Interestingly, colloidal

409

instability increased with the increase of molecular weight.

410 411

Figure 1 Nanoparticle size and ζ-potential as a function of DDA, Mn, and amine to phosphate

412

ratio (N:P) measured in the presence of 10 and 150 mM NaCl. A) Nanoparticle size (Z-ave

413

diameter) vs DDA, Mn and N:P ratio in the presence of low ionic strength (10 mM NaCl, pH 5.5,

22 ACS Paragon Plus Environment

Page 23 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

414

measurement at 2.5 min post incubation in medium). B) Nanoparticle surface charge (ζ-

415

potential) vs DDA, Mn and N:P ratio in the presence of low ionic strength (10 mM NaCl, pH

416

5.5). C) Nanoparticle size vs DDA, Mn and N:P ratio in the presence of high ionic strength (150

417

mM NaCl, pH 5.5). D) Nanoparticle surface charge (ζ-potential) vs DDA, Mn and N:P ratio in

418

the presence of high ionic strength (150 mM NaCl, pH 5.5). Data represent the average ±

419

standard deviation of 3 independent experiments with 2 technical replicates per experiment

420

(N=3, n=6). Measurements in 150 mM NaCl were taken immediately after adding 150 mM

421

NaCl. Measurements in 10 mM NaCl were stable over time (see Supp Info Fig 1)

422

siRNA compositions should be able to protect the nucleic acid cargo in physiological fluids

423

through a material specific mechanism of payload entrapment. As a consequence, encapsulation

424

efficiency (EE), or the percentage of siRNA incorporated into the nanoparticle, was measured as

425

a function of DDA, Mn, N:P ratio and pH using dye exclusion. In order to eliminate the effect of

426

colloidal instability on dye exclusion, the assay was performed at low ionic strength. As shown

427

in Figure 2, all formulations were able to achieve complete payload encapsulation at pH 6.5.

428

Increasing the pH from 6.5 to 8.0 resulted in dye accessing siRNA payload indicating release and

429

highlighting the importance of DDA, Mn and N:P ratio, and their interaction, on the

430

integrity/stability of the particles and payload protection. Maximization of nanoparticle integrity

431

at high pH could be achieved by increasing DDA, Mn and N:P ratio.

23 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 64

432 433

Figure 2 Effect of DDA, Mn and N:P ratio on the encapsulation efficiency at two different pH.

434

Nanoparticles were formed in water and incubated either in 25 mM MES (pH 6.5) or 1X TAE

435

(pH 8.0) for 24 hours then assayed for siRNA release using the Quant-iT™ RiboGreen® assay.

436

The percentage of siRNA release provided the percent encapsulation efficiency (% EE)

437

computed relative to naked siRNA (N:P 0). Red color corresponds to 100% encapsulation

438

efficiency (no release) while magenta corresponds to 0% encapsulation efficiency (all released). 24 ACS Paragon Plus Environment

Page 25 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

439

Average values from 2 independent experiments with 3-4 technical replicates per experiment. At

440

pH 6.5, complete encapsulation (0% release) of the payload is observed at all DDA, Mn and N:P

441

ratio used to form nanoparticles. However, at pH 8.0, chitosan glucosamine units become

442

deprotonated and their interaction with siRNA phosphate groups decreases promoting payload

443

release. At pH 8, an increase in DDA, Mn and N:P ratio is required to maintain nanoparticle

444

integrity.

445

In vitro performance of these formulations in the presence of 10% serum was assessed using +

446

the EGFP

447

illustrated in Figure 3, EGFP knockdown significantly increased with increasing charge density

448

(DDA), and to a lesser extent with increasing polymer length (Mn) and N:P ratio. The effect of

449

pH on the biological performance of nanoparticles was minimal with a slight decrease in

450

knockdown efficiency at higher pH observed for formulations with low-to-intermediate DDA

451

(i.e. 72-10, 72-120, 80-10 and 80-120). No pH dependent performance could be detected for

452

formulations with high charge density (92-10, 92-120, 98-10 and 98-120) when transfection pH

453

increased from 6.5 to 7.4 (Figure 3). In contrast, a statistically insignificant improvement in

454

knockdown at acidic pH was observed for formulations with low-to-intermediate charge density.

455

Although the effect of N:P ratio was minimal in contrast to DDA, increasing N:P ratio could

456

increase the knockdown efficiency of formulations with low-to-intermediate DDA (Figure 3).

H1299 cell line and correlated with nanoparticle physicochemical properties. As

25 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 64

457 458

Figure 3 Effect of DDA, Mn, N:P ratio and pH on the biological performance of chitosan-siRNA

459

nanoparticles in the presence of 10% FBS. Nanoparticles were formed in water following 1:1

460

(v/v) mixing of chitosan to siRNA (0.1 mg/mL). EGFP+ H1299 cells were transfected at a final

461

siRNA concentration of 100 nM. Data represent average ± standard deviation of at least 3

462

independent experiments with at least 2-3 technical replicates in each experiment (N=3, n=6-9).

463

The media pH was adjusted to pH 6.5-6.7 (left side top and bottom panels) by buffering the

464

RPMI-1640 media (pH 7.2-7.4) with 20 mM MES followed by a1N HCl titration.

465

The effect of DDA, Mn and N:P ratio on nanoparticle internalization was investigated at

466

physiological pH (7.2-7.4, 290 mOsm) and in the presence of serum. As depicted in Figure 4 A, 26 ACS Paragon Plus Environment

Page 27 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

467

nanoparticle internalization increased with increasing DDA, Mn and N:P ratio reminiscent of the

468

trend observed in Figure 3.

469

To understand the relationship between nanoparticle physicochemical characteristics and their

470

biological activity, physiochemical parameters were correlated and regressed with respect to

471

knockdown efficiency. Figure 4 shows the relationship between size, surface ζ-potential, and

472

knockdown efficiency. No correlation between size measured in 10 or 150 mM NaCl and

473

knockdown was observed (Figure 4, B). However, nanoparticle ζ-potential, a parameter found to

474

be strongly dependent on DDA, Mn and N:P ratio (Figure 1), showed strong correlation with

475

EGFP knockdown with Pearson product moment correlation coefficient (PPMCC) reaching

476

0.74-0.88.

477

In order to demonstrate that observed knockdown is independent of metabolic disturbances,

478

the viability of treated cells was determined relative to untreated controls using the alamarBlue®

479

assay. The principle of the assay is based on the mitochondrial reduction of resazurin, a blue and

480

non-fluorescent molecule, into resorufin, a red and fluorescent molecule. As such, the number of

481

cells and the incubation time were optimized before performing the assay since seeding density

482

and population doubling time are critical parameters for accurate results. Figure S2 shows

483

complete depletion of resazurin four hours post incubation at a cell density of 65,000 cells/cm2.

484

The optimal number of cells for viability testing was thereby determined to be 15,625 cell/cm2, a

485

cell density equivalent to 5,000 cells/well (96 well plate). As illustrated in Figure 4 D, a slight

486

decrease of 5-15% in metabolic activity was recorded for all formulations tested except for the

487

larger decrease observed with 98-10-5. Under identical transfection conditions where polymer

488

and lipid based nanoparticles are in contact with cells for 48 hours, all formulations

489

outperformed the lipid control. Under such conditions, DharmaFect® 2 showed strong toxicity 27 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 64

490

with around 60-80% cell death. However, it is noteworthy to mention that media replacement 5

491

hours post transfection abrogated changes in metabolic activity relative to untreated cells for

492

both chitosan and the lipid formulations and that no toxicity was observed when evaluated 4

493

hours post-transfection (data not shown).

494 495

Figure 4 Cell uptake, knockdown, correlations to size and charge, cell toxicity. A) The effects of

496

DDA, Mn and N:P ratio on uptake were measured in the EGFP+ H1299 cell line 48 hours post

497

transfection at 100 nM siRNA B) Lack of correlation between EGFP knockdown and 28 ACS Paragon Plus Environment

Page 29 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

498

nanoparticle size measured at low and high ionic strength. C) Strong correlation between EGFP

499

knockdown and nanoparticle surface charge (ζ-potential) measured at low and high ionic

500

strength. D) Effect of different formulations prepared at N:P 5 on metabolic toxicity. All

501

experiments in this figures were performed at pH 7.2-7.4 in the presence of 10% serum. All data

502

shown represent average of at least 3 independent experiments with 2-3 technical replicates per

503

experiment (N=3, n=6-9). Correlation graphs represent average values of size or ζ-potential

504

correlated with average values of EGFP knockdown.

505 506

In vitro knockdown efficiency is Mn independent above a certain threshold with chitosan found to disturb global gene expression

507

Based on the above results obtained from in vitro knockdown and the correlation between

508

chitosan DDA, Mn and N:P ratio and physicochemical properties (size, ζ-potential and EE)

509

and/or biological performance (EGFP knockdown and viability), the family of polymers with a

510

degree of deacetylation of 92% was selected for further characterization of the effect of Mn, N:P

511

ratio and serum proteins on siRNA uptake and knockdown. As depicted in Figure 5 A, in vitro

512

uptake, or internalization, of DY647 labeled siRNA requires a threshold of polymer length of 10

513

kDa, above which internalization appears to be independent of both Mn and N:P ratio. Below

514

this Mn threshold of 10 kDa, the role of N:P ratio appears critical with a two-fold increase in

515

siRNA uptake when increasing N:P ratio from 5 to 30 (Figure 5 A). As expected, EGFP

516

knockdown followed a similar pattern since uptake and knockdown are generally correlated,

517

given the ability of these nanoparticles to escape endosomal compartments

518

efficiency was independent of Mn and N:P ratio above 40 kDa (Figure 5). To demonstrate that

519

the decrease in EGFP fluorescence intensity was not related to toxic or non-specific effects of

520

chitosan itself, mock transfections with naked chitosan (M) and transfections using non-targeting

12, 41

. Knockdown

29 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

521

siRNA (siNT) were included as controls. As shown in Figure 5 C and D, the delivery of siNT

522

resulted in minimal EGFP knockdown, reaching a maximum of 10±2% with 92-40. In contrast,

523

mock transfections mediated a modest 5-10% increase in EGFP expression for some chitosans.

524

The pattern of EGFP knockdown and/or expression for both siNT and mock seem to follow a

525

trend where longer chain chitosans appears to have a slight positive effect on EGFP expression.

526 527

Figure 5 Effect of Mn at 92% DDA (92-Mn) and N:P ratio on uptake and knockdown. The

528

EGFP+ H1299 cell line was transfected in the presence of 10% serum at a final siRNA

529

concentration of 100 nM. A) Uptake of DY647 labeled siRNA expressed as median fluorescence 30 ACS Paragon Plus Environment

Page 30 of 64

Page 31 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

530

intensity (MFI). B) EGFP knockdown post transfection with anti EGFP nanoparticles. C) Lack

531

of EGFP knockdown post transfection with non-targeting siNT nanoparticles; siNT represents a

532

scrambled siRNA and is an indicator of specificity to the target siRNA sequence. D) Lack of

533

EGFP knockdown following transfection with chitosan only. Data represent the average ±

534

standard deviation of 3 independent experiments with 2 technical replicates per experiment

535

(N=3, n=6).

536

Since siRNA is a small and rigid molecule compared to other nucleic acids previously studied

537

with chitosan, we hypothesized that chitosan chain length below a threshold of ~ 60-70

538

monomers (10 kDa) have a lower affinity for siRNA and therefore release siRNA in complex

539

media (pH 7.4, high ionic strength and presence of serum). As shown in Figure S3, an increase in

540

Mn from 5 to 10 kDa or 10 to 120 kDa dramatically improves siRNA encapsulation at low N:P

541

ratios. Although the effect of physiological pH on nanoparticle stability, below the Mn 10kDa

542

threshold, is clear (Figure S3), it does not seem to totally account for the loss of internalization

543

and knockdown efficiencies observed in Figure 5. We subsequently verified the effect of serum

544

on EGFP knockdown with low versus high Mn chitosans (10 vs 120 kDa). Figure 6 shows a

545

decrease in performance in the presence of 10% serum for the 10 kDa chain which was rescued

546

by increasing the N:P ratio from 5 to 30 indicating that a threshold of at least 10 kDa is needed to

547

counter the negative effects of both pH and serum (Figure S3 and Figure 6).

31 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 64

548 549

Figure 6 Effect of FBS, chitosan Mn and N:P ratio at 92% deacetylation on EGFP knockdown in

550

H1299 cells. A) EGFP knockdown measured as the average fluorescence intensity (FI) relative

551

to untreated cells 48 h post-transfection with siEGFP. EGFP+ H1299 cells were transfected in

552

the absence or presence of 10% serum for a period of 5 hours, media aspirated and replenished

553

with complete RPMI-1640 media (pH 7.2-7.4, 290 mOsm) and incubated for 44 hours before

554

analysis. B) EGFP mRNA knockdown measured using qPCR, normalized using the geometric

555

average of EIF, PUM-1, and GAPDH and calibrated to untreated cells. EGFP+ H1299 cells were

556

treated as described in A. C) Representative confocal laser scanning microscopy (CLSM)

557

images. EGFP is indicated in green. In all experiments, siRNA was delivered at a final

558

concentration of 100 nM and data in A and B expressed as the average value of 3 independent

559

experiments with 2-3 technical replicates per experiment (N=3, n=6-9), *p-value < 0.01.

32 ACS Paragon Plus Environment

Page 33 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

560

Next, we confirmed assessed knockdown on the transcriptomic level by quantifying EGFP

561

messenger RNA (mRNA) using MIQE compliant quantitative real time PCR (qPCR). qPCR is a

562

very sensitive and powerful technique but can generate biased results in several cases where a

563

single or a combination of factors such as normalization strategy, validation of primer efficiency

564

and/or RNA integrity are not properly controlled

565

reference genes following treatments was validated in parallel with the validation of primer-

566

probe efficacy (Table S1). As seen in Figure 7, common reference genes i.e. β-actin and HPRT

567

were highly unstable with M scores above 0.5. Treatment dependent fluctuations were also

568

observed during modeling of treatment effect, with the removal of DharmaFect® 2 altering the

569

classification of reference genes (Figure 7 B). This observation confirms that treatments with

570

either lipid or chitosan based nanoparticles have, in principle, a certain impact on the

571

transcriptome or at least on these tested reference genes. The high variability of β-actin could be

572

attributed to poor assay efficiency (84%), which was confirmed to be treatment independent as

573

assays had the same amplification efficiency on cDNA amplified from total RNA extracted from

574

treated vs untreated samples (Table S1). The V-score, or the pairwise variation between the

575

normalization factors NF n and NF n+1 , showed that accurate normalization could be achieved by

576

using the geometric mean of the 2-3 most stable reference genes (Figure S4). Under optimal

577

normalization conditions and following MIQE guidelines, a similar trend for the effect of Mn

578

and N:P ratio in the presence/absence of serum was observed using quantitative PCR (Figure 6,

579

B). The knockdown of EGFP in the H1299 cells was qualitatively confirmed by confocal laser

580

scanning microscopy (CLSM), as shown in Figure 6 C.

35, 42-48

. As a consequence, the stability of

33 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 64

581 582

Figure 7 Effect of chitosan and DharmaFect® 2 treatment on reference gene stability. A panel of

583

10 genes was tested for their stability under diverse experimental conditions. EGFP+ H1299 cells

584

were transfected with the following formulations i.e. 92-10-5, 92-10-30, 92-120-5, 92-120-30,

585

98-10-5, 98-10-30 and DharmaFect® 2 at a final siRNA concentration of 100 nM untreated cells

586

were included in the analysis. A) panel shows the classification of the least to most stable (left to

587

right) reference gene based on the average expression stability values, or geNorm M-Score,

588

computed on the remaining control genes during stepwise exclusion of the least stable control 34 ACS Paragon Plus Environment

Page 35 of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

589

gene, for samples from all treatments. B) panel shows the effect of the exclusion of

590

DharmaFect® 2 from the statistical analysis. C) Panel shows the effect of the exclusion of both

591

DharmaFect® 2 and untreated cells form the analysis. The M-Scores were computed using the

592

geNorm statistical package on the average Cq of two independent experiments.

593 594

In vitro lipid nanoparticle (LNP) like potency (EC 50 ) can be achieved, in the presence of serum, using chitosans with increasing Mn and N:P ratio

595

The effect of Mn and N:P ratio on the minimum effective dose needed for EGFP knockdown

596

in the H1299 cell line was determined in the presence of serum. The half maximal effective

597

concentration (EC 50 ), a measurement of nanoparticle potency, was computed from a 4-parameter

598

sigmoidal curve (4-PL) 48 hours post-transfection. Table 2 shows that potency increased with

599

increased Mn and N:P ratio. DharmaFect® 2, a lipid control developed for siRNA delivery in

600

H1299 cells, had the lowest EC 50 (Table 2) and was able to induce meaningful knockdown at an

601

EC 50 of 23 nM. Similar potency was obtained with the 92-120-30 formulation (EC 50 of 29.3

602

nM). As shown in Figure S5, all formulations reached a plateau around 200 nM with a marginal

603

increase in EGFP knockdown observed at higher concentrations. The delivery of non-targeting

604

siRNA (siNT) at higher doses (Figure S5) compared to Figure 5, C, did not cause a meaningful

605

increase in off-target effects. Again, siNT induced a small decrease/increase in EGFP expression

606

with low and high Mn chitosan respectively indicating that off-target effects are probably due to

607

reduced metabolic activity observed with low Mn chitosan (Figure 9 A).

608

Table 2 Effect of Mn and N:P ratio on in vitro dose dependent knockdown.

35 ACS Paragon Plus Environment

Biomacromolecules 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 64

EC 50 Standard error

p-value

Curve R2

77.2

9.88

0.01

0.99

30

42.8

4.32

0.00

0.99

92-120

5

46.3

1.71

0.00

0.99

92-120

30

29.4

4.52

0.02

0.99

DharmaFect® 2

NA

23.8

3.28

0.02

0.99

Formulation

N:P ratio

EC 50 (nM)

92-10

5

92-10

2

fit

609 610 611

Reduction in knockdown efficiency due to serum can be mitigated by increasing Mn and N:P ratio.

612

Following observations regarding the effect of serum on the performance of low molecular

613

weight chitosan, coupled with the fact that one use of nanoparticle is systemic administration in

614

vivo, the effect of increasing serum concentration on biological performance was studied. In

615

particular, the effect of Mn and N:P ratio on EGFP knockdown was investigated in the presence

616

of increasing serum concentrations. As shown in Figure 8, A, nanoparticles rapidly lost their

617

performance in the presence of increasing serum concentration. This loss of performance could

618

be mitigated by higher Mn and/or N:P ratio. The latter seems to play an important role in

619

promoting transfection in the presence of high concentration of serum (i.e. 94%). N:P ratio

620

seemed to account for around 15% while Mn for around 10% as shown in Figure 8, B. The effect

621

of physiological concentrations of heparin (2.5 µg/mL)49 and albumin (25-40 mg/mL) at pH of

622

7.2-7.4 on payload release was studied by means of dye exclusion. As shown in Figure 8, C, a

623

physiological concentration of heparin, equivalent to concentrations found in 94% serum, was 2

EC50 values were derived from a 4-parameter sigmoid curve fitted to data derived from 2 independent experiments with 2 technical replicates per experiment; Figure S5 in supplemental data for more information, p-value