- Email: [email protected]
Enhanced Achilles tendon healing by fibromodulin gene transfer
Enhanced achilles tendon healing by fibromodulin gene transfer Anthony Delalande Ph.D, Marie-Pierre Gosselin MSc, Arnaud Suwalski MSc, William Guilmain Ph.D, Chlo´e Leduc MSc, Mathieu Berchel Ph.D, Paul-Alain Jaffr`es, Patrick Baril Ph.D, Patrick Midoux Ph.D, Chantal Pichon PII: DOI: Reference:
S1549-9634(15)00119-7 doi: 10.1016/j.nano.2015.05.004 NANO 1131
To appear in:
Nanomedicine: Nanotechnology, Biology, and Medicine
Received date: Revised date: Accepted date:
27 January 2015 30 April 2015 18 May 2015
Please cite this article as: Delalande Anthony, Gosselin Marie-Pierre, Suwalski Arnaud, Guilmain William, Leduc Chlo´e, Berchel Mathieu, Jaffr`es Paul-Alain, Baril Patrick, Midoux Patrick, Pichon Chantal, Enhanced achilles tendon healing by fibromodulin gene transfer, Nanomedicine: Nanotechnology, Biology, and Medicine (2015), doi: 10.1016/j.nano.2015.05.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Enhanced Achilles tendon healing by fibromodulin gene transfer. Anthony Delalande1*, Marie-Pierre Gosselin1*, Arnaud Suwalski1, William Guilmain1,
T
Chloé Leduc1, Mathieu Berchel2, Paul-Alain Jaffrès2, Patrick Baril1, Patrick Midoux1
RI P
and Chantal Pichon1. 1
Centre de Biophysique Moléculaire, rue Charles Sadron, Orléans 45071 CEDEX 2,
SC
France. 2
Université de Brest, Brest, France
MA NU
CEMCA, CNRS UMR 6521, IFR148 ScInBioS, Université Européenne de Bretagne,
*: These authors contributed equally to this work.
Abstract: 147 words
PT
Body text: 4962 words
AC
3 tables
CE
37 references 6 figures
ED
Corresponding author: [email protected]
1 supplementary material The authors disclose no conflict of interest. This work was supported by the Association Française contre les Myopathies (A.F.M).
1
ACCEPTED MANUSCRIPT Abstract Tendon injury is a major musculoskeletal disorder with a high public health impact.
T
We propose a non-viral based strategy of gene therapy for the treatment of tendon
RI P
injuries using histidylated vectors. Gene delivery of fibromodulin, a proteoglycan involved in collagen assembly was found to promote rat Achilles tendon repair in vivo
SC
and in vitro. In vivo liposome-based transfection of fibromodulin led to a better healing after surgical injury, biomechanical properties were better restored compared
MA NU
to untransfected control. These measures were confirmed by histological observations and scoring. To get better understandings of the mechanisms underlying fibromodulin transfection, an in vitro tendon healing model was developed.
ED
In vitro, polymer-based transfection of fibromodulin led to the best wound enclosure speed and a pronounced migration of tenocytes primary cultures was observed.
PT
These results suggest that fibromodulin non-viral gene therapy could be proposed as
Key words
CE
a new therapeutic strategy to accelerate tendon healing.
AC
Tendon healing ; Gene therapy ; Lipoplex ; Polyplex
2
ACCEPTED MANUSCRIPT 1. Background The incidence of work-related musculoskeletal disorders (WMSD) is increasing every
T
year due to modern life. Tendon injuries represent the main pathology of WMSD and
RI P
healing requires a long rest period 1. It has been proven that growth factors like PDGF, TGF-β or VEGF can considerably accelerate this wound-healing process by
proliferation and migration
SC
regulating the inflammation phase, the production of extracellular matrix and the cell 2-5
. The delivery of these growth factors has some
MA NU
limitations due to their short half-life time requiring repetitive injections and the extensive cost of recombinant protein production and purification. Gene therapy is the best approach to produce growth factors in situ and can be exploited in tendon disorders 6.
ED
Transient gene expression after gene transfer by non-viral vectors would be useful
PT
for wound-healing because molecules involved in tissue repair have to be expressed only in a short-time period 7. Synthetic vectors are chemically controlled compounds
CE
easy to handle, and show a low immunogenicity with a weak risk of transgene
AC
integration.
These last years, few reports have shown the potentiality of in vivo gene delivery for tendon regeneration. They mainly concern the delivery of genes encoding growth factors
by
using
electroporation
liposomes,
adenoviral
vectors,
silica
nanoparticles
and
8-11
. The main effect of the expression of growth factors is the
induction of tenocytes proliferation and collagen I production. But, the tendon strength depends also on a good matrix assembly which relies on the collagen fibrillogenesis that depends on the activity of proteoglycans like fibromodulin (Fmod) or lumican (Lum)
12, 13
. Moreover, proteoglycans have been shown to enhance the
3
ACCEPTED MANUSCRIPT repair and remodelling of injured cornea
14
, skin15, and ligament
16
. They have a
specific expression profile in injured tendons compared to normal tissue
17
.
T
In this study, we investigated as nanomedicine approach whether Fmod or Lum gene
RI P
transfer could improve rat Achilles tendon healing. Gene transfer was performed either with imidazole cationic lipids (Lip 100 liposomes) or histidylated linear
SC
polyethyleneimine (PTG1) in Achilles tendons in vivo and in primary cultures of tenocytes. Wound-healing effect was assessed by histological analyses and stiffness
MA NU
measurements after transfection of fibromodulin. Finally, we performed in vitro experiments including wound-healing, cell proliferation and cell migration assays to get a better understanding of the benefic effect brought by fibromodulin gene
AC
CE
PT
ED
transfer.
4
ACCEPTED MANUSCRIPT 2. Methods 2.1. Plasmids
T
pNFCMV-Luc (pLuc), a 7.5 kb homemade plasmid DNA (pDNA) encoding the
RI P
firefly luciferase under control of the strong cytomegalovirus promoter was used as a reporter gene. This plasmid has five consecutive NFκB motifs (termed NF that
SC
recognize the NFB transcription factor) inserted upstream of the promoter. The second reporter pDNA was pMaxGFPTM, a 3.5 kb plasmid encoding the eGFP gene
MA NU
(Lonza, Basel, Switzerland). pPDGF plasmid (Invitrogen, Cergy-Pontoise, France) was encoding the human PDGFB gene. Plasmids pCEP4-FBM (pFmod) and pCEP4LUM (pLum) plasmids (generous gift from Dr. Peter Roughley) were encoding human 18
. pQE30, a mock plasmid (pMock)
ED
fibromodulin and lumican genes, respectively
that did not encode any gene was used as control in wound-healing experiments.
PT
2.2. Liposomes and polymers.
Histidylated liposomes (Lip100) were prepared by mixing O,O-dioleyl-N-[3Niodide)propylene]
CE
(N-methylimidazolium
phosphoramidate
and
O,O-dioleyl-N-
19, 20
AC
histamine phosphoramidate. These lipids were synthetized as previously described . Histidylated linear polyethylenimine (PTG1) was produced by Polytheragene
(Evry, France)
21
. Lipoplexes and polyplexes were formed at vector/DNA weight ratio
of 3:1 and 6:1, respectively. The detailed procedure for DNA complexes preparation is presented in supplementary methods. 2.3. Animal studies 2.3.1. In vivo transfection Adult Wistar rats were bred in CBM animal facility at 22°C for at least one week before experiments and groups were randomly made. Experiences were conducted according to the guidelines of the French Ministry of Agriculture for 5
ACCEPTED MANUSCRIPT experiments with laboratory (law 87848, C. Pichon accreditation). Rats were anesthetized by intraperitoneal injection of ketamine (50 mg/kg) and xylazine (10
T
mg/kg) in 0.9% NaCl. Skin was incised on few millimeters in Achilles tendon region.
RI P
For tendon healing study, Achilles tendon was incised through the tendon sheath along its total length three times using a carbon steel surgical blade n°12 (Swann-
SC
Morton, Sheffield, UK) (figure 1A). Fifty microliters of lipoplexes or polyplexes containing 20 µg of plasmid DNA were slowly injected in the middle section of
MA NU
Achilles tendon using a 31G Hamilton PBS600 (Hamilton, Bonaduz, Switzerland) repeater delivery system. Skin was sutured and rats were closely observed until their wake-up. Painkiller treatments were managed by subcutaneous injection of Finadyne at 250 µl/kg (Schering-Plough, Kenilworth, NJ, USA).
ED
2.3.2. Ex vivo luciferase assay
PT
The transfection efficiency obtained in tendons of pLuc-injected rats was determined as follows. Rats were euthanized by lethal CO 2 inhalation at indicated
CE
day post-surgery. At day 1, 3, 6 and 9 post-treatment, tendons were harvested and
AC
crushed in liquid nitrogen with a mortar and a pestle. The powder was added in an ice-cold lysis buffer (Luciferase Cell Culture Lysis, Promega, Madison, USA) and the mixture was kept for 3 hours on ice before luminescence measurement. The level of luciferase activity was measured with the Lumat LB9507 luminometer (Berthold, Wildbarch, Germany) after adding 100 µl of luciferin (Luciferase Assay System, Promega) to 20 µl of tendon lysates and expressed as relative luminescence unit (RLU) per mg of proteins. Proteins content of tendons were quantified by bicinchoninic acid assay. 2.3.3. In vivo bioluminescence imaging and quantification
6
ACCEPTED MANUSCRIPT Bioluminescence imaging was conducted in vivo using a Hamamatsu Photonics cooled CCD camera mounted on a dark box chamber (Hamamatsu,
T
Japon). Signal intensity was recorded 2 minutes after luciferin injection (100 µg,
RI P
Promega) in the tendon area and quantified as the mean of photons per second (p/s) of time exposure within a region of interest prescribed over the tendon.
SC
2.3.4. Biomechanical tests
Rats were euthanized by lethal CO2 inhalation at indicated times. Achilles
MA NU
tendons were harvested and stored in PBS at -20°C before use. They were dissected and cleaned to remove muscular and adherent surrounding tissues. Then the tendon was mounted vertically on a mechanical test machine (MTS synergie 400, Eden Prairie, MN, USA) and preloaded with a velocity of 0.6 mm/min by a triple cyclic
ED
loading, stretching up to 110% of its initial length (L0). After pre-conditioning,
PT
specimens were stretched to failure at the same stretching speed and the force was recorded. The stiffness of the tendons was calculated by the linear regression of the
CE
curve when plotting the displacement (mm) versus the load (N). During all the
AC
experiments the tendons were kept wet by spraying distilled water. Each group was composed of 5 tendons (healthy group was composed of 20 tendons). Values are expressed as means ± SEM. 2.3.5. Histological analysis Tendons were harvested 14 days post-injury and were fixed in 10% pformaldehyde solution, paraffin embedded, sliced in serial sections (thickness: 4 μm), mounted on glass slides and counterstained with HES (5 tendons per group were used for histological analyses). Histological process and blind analyses were carried out by In-Cell-Art (Nantes, France). The wound healing status of the tendons was evaluated by 3 criteria: the number of nuclei (hypercellularity), the matrix continuity
7
ACCEPTED MANUSCRIPT and the presence of inflammation. For each criteria a score from 1 to 3 was given, 1: poor, 2: average, 3: good 22.
T
2.4. In vitro studies
RI P
2.4.1. Primary culture of tenocytes
Tenocytes were obtained from adult Wistar rat Achilles tendon explants.
SC
Tendons were harvested, cut in small pieces in PBS and placed in tenocyte culture medium composed of MEM alpha, 10% Fetal Calf Serum, ascorbic acid (50 µg/ml),
MA NU
primocin (100 µg/ml, InvivoGen, Toulouse, France). Tenocytes were cultured at 37°C under humidified atmosphere containing 5% CO2. Half of the medium volume was replaced every two to three days. 2.4.2. In vitro transfection
ED
Two days before experiment, 105 tenocytes were seeded on 24-well plates.
PT
Then, tenocytes were transfected with indicated DNA formulations as described in supplementary material. After 4 hours, transfection medium was removed and cells
CE
were cultured for two days in tenocytes medium. The transfection efficiency was
23
.
AC
determined by measuring the luciferase activity in cell lysates as described previously
2.4.3. Monolayer wound-healing assay The tenocyte wound-healing model was created by plating 9×104 tenocytes two days before experiments on a CytoSelect™ 24-Well Wound-Healing Assay (Cell Biolabs Inc., San Diego, CA, USA) allowing creation of a 0.9 mm-wide standardized gap between confluent tenocytes monolayers. Inserts were removed and cells were transfected with indicated polyplexes formulations for 4 hours at 37°C. Transfection medium was replaced by fresh medium and the wound closure was monitored by videomicroscopy using a Zeiss Axiovert 200M (Carl Zeiss Inc., Thornwood, NY, USA)
8
ACCEPTED MANUSCRIPT fully motorized microscope during 72 hours (scan speed: 2 images/hour). Cells were incubated in an atmosphere and temperature controlled chamber at 37°C and 5%
T
CO2. Quantitative analysis of the cell free area was performed using the Axiovision
RI P
Rel. 4.7 (Carl Zeiss Inc.). The level of wound-healing for a time x was evaluated by calculating the percentage of the cell free area at tx divided by the cell free area at
SC
the initial state. Experiments were at least repeated 6 times. In a set of experiments wound closure was followed when the tenocyte wound-healing model was incubated
2.4.4. Quantitative real-time PCR
MA NU
with medium from transfected tenocytes, collected 24 hours after transfection.
Tenocytes were harvested 18 and 43 hours post-transfection and total RNA
ED
was extracted using an RNA II NucleoSpin® mini kit (Macherey-Nagel EURL, Hoerdt, France). For all experiments, 50 ng of RNA was used for reverse transcription using
PT
200 units of M-MLV Reverse Transcriptase (New England Biolabs, MA, USA). The gene expression was evaluated from 100 ng of cDNA by qPCR using Quantifast®
CE
PCR Master Mix (Qiagen, Venlo, Netherlands) with 0.1 µM of Fmod, Lum, PDGF or
AC
RNA6S specific primers (Table 1). The qPCR was performed on a Light Cycler II 480 (Roche Applied Science, Meylan, France), the cycles consisted of 10 seconds at 95°C for DNA denaturation followed by 30 seconds at 60°C for primer annealing and polymerization repeated 40 times. Fmod, Lum and PDGF primers were specific to the human transcripts of cDNA inserted into the pFmod, pLum and pPDGF plasmids. The relative quantification was made using the ΔΔCt method using the RNA6S housekeeping gene expression.
2.4.5. Tenocytes proliferation
9
ACCEPTED MANUSCRIPT The evaluation of tenocytes proliferation was monitored by MTT colorimetric assay 18, 43 and 72 hours after transfection. Cells were incubated during 4 hours at
T
37°C, 5% CO2 in presence of 500 µg/ml of 3-(4,5-dimethylthiazol-2-yl)-2,5-
RI P
diphenyltetrazolium bromide (Sigma Aldrich, Saint Louis, MO, USA). Cells were washed and lysed with acidified isopropanol containing 3% of SDS (Sigma Aldrich)
SC
allowing the solubilization of tetrazolium salts. Quantification was made measuring optical density at 560 nm; the control was made using cells having received the same
MA NU
manipulations without transfection. Experiments were done in triplicate and repeated two times.
2.4.6. Migration analysis of tenocytes
Migration analysis was defined as the capacity of tenocytes to migrate through
ED
8 µm cell culture insert membranes (BD Falcon, Franklin Lakes, NJ, USA). Bottom
PT
wells were filled with complete medium and inserts seeded with 10 5 transfected tenocytes. Cells were incubated at 37°C in a 5% CO2 atmosphere during 18 hours.
CE
Inserts were cut and the number of cells present on the other side of the insert was
AC
evaluated by microscopy. The migration factor was calculated as the number of cells present at the bottom side of the insert after transfection divided by the number of cells present in non-transfected condition. Each condition was repeated two times in triplicate. 2.5. Statistical analysis Data were expressed by means ± SD (in vitro experiments) or SEM (in vivo experiments). All statistical differences were analyzed by a bilateral Mann–Whitney U-test when statistical difference was found by Kruskal-Wallis statistical test. XLStat 2014 software was used (Addinsoft, Paris, France), significance was defined as pvalue <0.05 (asymptotic p-value was calculated).
10
ACCEPTED MANUSCRIPT 3. Results 3.1. Achilles tendon healing model
T
We have established a tendon injury model by making a triple longitudinal
RI P
laceration wound (Figure 1A). This wound modality was chosen to keep tendon connections between muscle and bone; moreover this lesion was a good model to 24
. The timeline of the study is presented in figure 1A.
SC
reproduce tendon injury
Histological analysis and stiffness measurement of injured tendons were performed
MA NU
to validate our model. Healthy tendons are mainly composed of curly formed and parallel collagen fibers (Figure 1C). They contain very few tenocytes that are positioned along the collagen fibers. Fourteen days after tendon injury, histological
ED
analysis showed hypercellularity areas in the axis of the lesion and a disorganized collagen fibers network. Figure 1B shows biomechanical analyses of the tendon
PT
stiffness. The stiffness value reflects the state of healthiness of tendons. This value increased slowly from 3 to 14 days post-injury; tendons had a stiffness of 17.76 ±
CE
3.38 N/mm after 3 days, 20.78 ± 2.81 N/mm after 7 days and 25.75 ± 3.80 N/mm at
AC
14 days post-lesion. As comparison, healthy tendons exhibited a much greater stiffness of 33.17 ± 3.77 N/mm. The abnormal architecture of injured tendons was correlated with their reduced stiffness. Based on these data, we decided to transfer genes after lesion and to assess the effects 14 days post-injury. This time point was chosen to leave enough time for transfected genes to act on the healing process. 3.2. Gene transfer in injured Achilles tendon First, Lip100 and PTG1 vectors were used to transfer the luciferase reporter gene in injured Achilles tendons and bioluminescence was used as a read out. These two transfection reagents have never been used for in vivo transfection of Achilles
11
ACCEPTED MANUSCRIPT tendon and in vitro transfection of primary cultures of tenocytes. A summary of the physico-chemical properties of pDNA complexed with Lip100 (lipoplexes) or PTG1
23
. The chosen vector/pDNA weight ratios were those that
RI P
been previously described
T
(polyplexes) is presented in Table 2. The detailed structure of those carriers has
permitted to get a good balance between efficacy and toxicity in previous studies 24
20,
SC
. PTG1/pDNA polyplexes at ratio of 6:1 (µg:µg) were slightly positive with a (ζ
potential of +11±1.6 mV. Lip100/pDNA lipoplexes at a ratio of 3:1 exhibited a
with PTG1 than with Lip100
MA NU
negative ζ potential of -31±0.2 mV. As previously shown, pDNA was more compact 21, 25
. Rats Achilles tendons were injected with pLuc
plasmid complexed either with Lip100 or PTG1. The transfection efficiency was assessed 24 hours later (Figure 2A). The results indicate that the luciferase activity
ED
was 85 times higher after Lip100 lipoplexes than PTG1 polyplexes injection (1.7 ×
PT
106 RLU/mg of proteins vs. 2 × 104 RLU/mg of proteins). Compared to lipofection, the injection of naked pLuc gave a higher level of luciferase expression one day post-
CE
injection (Figure 2B). However, the luciferase activity dropped dramatically after 3
AC
days and was almost undetectable at day 6. By contrast, the luciferase activity that also decreased at day 3 post-lipofection was maintained at day 6 with a similar value (9.7 × 105 ± 4.95 × 104 RLU/mg of proteins vs. 1.33 × 105 ± 1.56 × 104 RLU/mg of proteins). Therefore, the luciferase expression was more stable when injection was performed with Lip100 lipoplexes compared to naked pLuc. Figure 2C shows bioluminescence representative pictures of the luciferase expression following injection of lipoplexes in the right Achilles tendon. The bioluminescence was located in the Achilles tendon area at days 1, 3 and 6 post-injection. The decrease of the luminescence intensity followed the same trend as the luciferase activity assessed ex
12
ACCEPTED MANUSCRIPT vivo (Figure 2B). The signal decreased from day 1 to day 3 but was maintained at day 6.
T
3.3. In vivo transfection of injured Achilles tendons with therapeutic genes.
RI P
Based on these results, we decided to transfer therapeutic genes by using Lip100 liposomes. Plasmids DNA encoding PDGF gene (pPDGF), Lum gene (pLum)
SC
or Fmod gene (pFmod) were used to treat Achilles tendon injured by surgical lesions. The effects of these treatments were evaluated after 14 days both by biomechanical
MA NU
and histological analyses. In figure 3A, shown are representative micrographs of sliced tendons sections treated with indicated therapeutic gene. Tendons treated either with pLum or pPDGF exhibited a higher hypercellularity area (as shown by the high number of nuclei in the healing region) compared to those treated with pFmod.
ED
However, the collagen matrix appeared much better structurally organized after the
PT
latter treatment. The tendon appearance was closer to that observed in control tendons with the presence of more aligned fibers. The histological analyses are
CE
summarized in the Table 3.
AC
Next, we measured the stiffness parameter of injured tendons 14 days posttreatment either with pFmod, pLum or pPDGF to confirm histological observations (Figure 3B). The stiffness of injured tendons was 25.75 ± 3.80 N/mm which was significantly lower than that of healthy tendons (33.17 ± 3.77 N/mm, p<0.05). Tendons treated with pFmod were stiffer than those treated either with pLum or pPDGF. Note that their stiffness value was also significantly higher than those of healthy group (44.62 ± 3.64 N/mm vs. 33.17 ± 3.77 N/mm, p<0.05). In contrast, overexpression of lumican did not restore tendon stiffness which was similar to that of untreated injured tendons (25.71 ± 3.87 N/mm vs. 25.75 ± 3.80 N/mm). Interestingly, PDGF gene transfer restored the tendon stiffness to a value close to that of healthy
13
ACCEPTED MANUSCRIPT tendon (30.00 ± 2.32 N/mm). This was in line with our previous observations after PDGF gene transfer using mesoporous silica nanoparticles25.
T
Altogether, the measurement of the stiffness parameter and the histology of treated
RI P
tendons confirm that the transfer of the fibromodulin gene improves tendon healing. 3.4. Primary tenocytes transfection with PTG1 and Lip100
SC
The transfection efficiency of primary cultures of tenocytes was tested with Lip100 and PTG1. As shown in Figure 4A, the polyplexes efficiency was 241-fold
MA NU
higher than that of lipoplexes (8.78 × 107 ± 6.70 × 106 and 3.64 × 105 ± 7.07 × 104 RLU/mg of proteins). This result confirms that most of the time in vivo efficiency of one type of vector does not predict about its in vitro efficiency and reciprocally. By
ED
using pDNA encoding eGFP reporter gene, the number of transfected tenocytes determined both by fluorescence imaging and flow cytometry was 31% after 24 hours
PT
(Figure 4C). Only 13% of cells were stained by propidium iodide indicating that PTG1 polyplexes transfection induced low cytotoxicity. Tenocytes were then transfected
CE
with PTG1 complexed either with pPDGF, pFmod or pLum. The transgene
AC
expression was assessed after 18 hours by quantitative real-time PCR (qPCR). Results indicate that all transgenes were highly expressed with a similar level (Figure 4B).
3.5. In vitro wound-healing upon Fibromodulin, Lumican and PDGF transfection The closure of the wound model in response to Fmod, Lum and PDGF gene expression after polyplexes transfection was monitored for up to 72 hours by videomicroscopy and the gap closure percentage was calculated at 18, 43 and 72 hours (Figure 5A). An accelerated gap filling was observed at early time points whatever the transfected gene. For the sake of clarity, data obtained at 18 hours were displayed on figure 5B. After pFmod transfection, 43 ± 10% of the gap area was
14
ACCEPTED MANUSCRIPT filled compared to 32 ± 9% after mock transfection (p<0.05). When tenocytes were transfected with pLum and pPDGF, the gap was filled at 44 ± 10% and 40 ± 5%,
T
respectively. These values were significantly higher compared to mock-transfected
RI P
cells (p<0.05 for both). 43 hours post-transfection, no significant difference was observed between cells transfected either with pFmod, pLum or pMock.
SC
3.6. Assessment of proliferation and migration involvement during wound-healing process of transfected tenocytes.
MA NU
The in vitro wound-healing assays suggested that the transfection of genes encoding proteoglycans improved the gap closure kinetic at early stages. This effect could occur either via an enhancement of the cell proliferation or the cell migration or both. Figure 6A shows the tenocytes relative proliferation analysis after 18, 43 and 72
ED
hours transfection either with pFmod or pLum. At 18 hours, an increased cell
PT
proliferation was seen for each transfection condition compared to the untreated control. No significant differences were observed between mock-transfected cells
CE
(1.7 ± 0.0) and those transfected with therapeutic genes, pFmod (1.8 ± 0.3) and
AC
pLum (1.9 ± 0.1) however a slight proliferation increase was obtained. Migration assays were carried out to know whether differences could occur after transfection of genes encoding proteoglycans (Figure 6B). pFmod-transfected cells exhibited a significant higher number of migrating cells through the insert compared to mock-transfected cells (227 ± 36% vs 108 ± 42%, p-value <0.05); the number of migrating non-transfected cells being taken as 100%. To go further, the effect of conditioned medium recovered from tenocytes overexpressing Fmod, Lum or PDGF was assessed. Indeed, those proteins are meant to be secreted in the cell medium allowing to get rid of possible transfection side effects. The wound closure was measured 18 hours upon incubation with
15
ACCEPTED MANUSCRIPT conditioned medium (Figure 5D). The Fmod-conditioned medium was more efficient to enhance the wound closure compared to non-treated cells (53 ± 7% vs. 35 ± 4%,
T
p<0.05). Incubation with Lum or PDGF conditioned media was not efficient as the
RI P
latter on wound-healing. The wound enclosure (38 ± 7% and 39 ± 3%, respectively) was slightly higher than that obtained with medium from mock-transfected cells.
SC
4. Discussion
MA NU
The growing number of publications related to nanomedicine applied to tendon pathologies shows and consolidates the potentiality of this application. Most of them report engraftment of genetically modified mesenchymal stem cells. Very few studies concern gene delivery approach by using non-viral systems. Recently, we reported
ED
that PDGF gene transfer by using silica mesoporous MCM-41 nanoparticles enhanced the healing of injured rat Achilles tendons25. Two injections of 10 µg of
PT
pPDGF spaced by one week were required to improve the histological architecture of
CE
injured tendons and most importantly their biomechanical properties. Comparatively, injection of pPDGF did not provide significant benefit effect in this study. By contrast,
AC
we found that one injection of 20 µg pFmod complexed with Lip100 liposomes was sufficient to induce improved structural matrix organization and stiffness of injured tendons in only 14 days. We assessed in vivo tendon and in vitro tenocytes gene delivery by using two kinds of pDNA complexes. Accordingly with the results, two remarks can be point out. First, lipoplexes made with Lip100 liposomes that provided efficient gene transfer in Achilles tendons, were weakly efficient to transfect primary cultures of tenocytes. Note that they were able to give a great transfection level of C2C12 murine myoblasts26. Thus the effectiveness of a gene delivery system is highly celldependent. Second, PTG1 polyplexes which were efficient in vitro (30% of 16
ACCEPTED MANUSCRIPT transfected tenocytes) did not transfect tendons in vivo (almost 100-times less potent than Lip100 lipoplexes). This highlights that in vitro gene transfer efficiency cannot
T
predict in vivo effectiveness. Compared to lipoplexes, the absence of efficacy of
RI P
PTG1 polyplexes could be ascribed to their positive charges which could favor interaction with negatively charged proteoglycans and thereby hamper their diffusion
SC
inside the tendon matrix.
The majority of studies related to the treatment of tendon injuries concerns the
MA NU
use of growth factors known to induce tenocytes proliferation and collagen I production. Overexpression of growth factors as IGF-I may indeed increase the risk factor of several diseases like cancer due to their mitogenic properties their expression has to be finely regulated
27, 28
, therefore
29
.
ED
Levels of gene expression obtained following in vivo Lip100 lipofection are not
PT
as high as viral vectors transduction but remains fairly stable from 3 to 6 days which is suitable for tendon healing. It is worth to note that stable and high levels of
CE
proteoglycans or growth factor expression are not always necessary since they could 30
. Fibromodulin expression and collagen fiber stiffness is closely
AC
provoke fibrosis
connected. The stiffness of Fmod-treated tendon is significantly increased compared to the healthy tendon possibly due to an overexpression of this proteoglycan. The half-life time of proteoglycans proteins as fibromodulin is known to be long (several days) as well as the half-life time of collagen proteins (more than 70 years)31. An accumulation of fibromodulin at higher levels to normal could explain the tendon stiffness increase. A well-organized matrix assembly is also needed and it can be driven by acting on fibrillogenesis. This process is regulated by proteoglycans which play a major role in maintenance of tendon structure. These molecules are also involved in
17
ACCEPTED MANUSCRIPT tenocytes proliferation by retaining/protecting growth factors as TGF-β via specific interactions32. Amongst proteoglycans, Fmod and Lum are of interest because they
14
, ligament
16
and skin
15, 33
but their action
RI P
repair and remodeling of injured cornea
T
are involved in collagen fibrillogenesis. They have been reported to play a role in the
in tendons has never been reported. When Fmod gene is knocked out, the collagen
SC
fibril architecture of connective tissues such as cornea, skin and tendon is altered leading to weak biomechanical parameters 34, 35.
MA NU
Our in vivo data indicate that the effect on wound-healing after transfection of Achilles tendons with Fmod gene was superior compared to that of Lum and PDGF gene. Injection of pLum induced high hypercellularity areas compared to
ED
untransfected tendons. This observation is often correlated with a good prognosis in the healing process as these cells will produce the new matrix. However, lumican
PT
overexpression did not improve tendon stiffness after injury. Data from in vitro wound-healing and migration assays indicate that Fmod is
CE
more potent to promote cell migration compared to Lum and PDGF. The involvement 15, 32, 33
. When Fmod is absent,
AC
of Fmod on healing has been reported in skin wound
a delayed healing was observed. The delivery of Fmod gene by adenoviral vectors has been found to improve the healing of rabbit skin wound
36
. In this study, two
injections of adenoviral vectors expressing Fmod were required to obtain an improvement of skin wound-healing. In addition to cell migration and cell proliferation, angiogenesis and inflammation processes are also crucial for wound-healing. Fibromodulin has been reported to improve angiogenesis during wound-healing
37
. Similar effect of Fmod
could likely be involved in our study. Fmod expression has also been found to be inversely correlated with a scar formation during both fetal skin development and
18
ACCEPTED MANUSCRIPT adult wound repair
38
. In a scarless fetal wound, Fmod is 3 to 5 fold upregulated
whilst there is only a small increase (1.3 fold) of its overexpression in adult wound
T
between 3 and 5 days post-injury. These results combined with ours suggest that
RI P
overexpressing Fmod could be a promising strategy to manipulate adult wound to a scarless repair.
SC
5. Conclusion
MA NU
Improved repair of injured tendon was obtained upon only one injection of 20 µg of pDNA encoding fibromodulin gene complexed with histidylated liposomes. Based on the stiffness parameter and the histological analysis of fibromodulin-treated tendons, the repair is better and closer to that of healthy tendons than after lumican
ED
or PDGF gene transfer. These results could be explained by the induction of cell migration and with a lesser extent of cell proliferation observed in vitro after
PT
fibromodulin overexpression. This repairing activity could be directly ascribed to
CE
fibromodulin expression. Therefore, fibromodulin could be a good alternative to
AC
growth factors for tendon healing.
19
ACCEPTED MANUSCRIPT Figure captions Figure 1 – In vivo Achilles tendon healing model. (A) The lesion was induced at day
T
0 by a triple longitudinal laceration of the Achilles tendon. Then, lipoplexes or
RI P
polyplexes were injected before suturing the tendon. When luciferase reporter gene was used, luciferase activity assay was performed at day 1, 3 and 6 to evaluate the
SC
gene transfer efficiency. (B) Stiffness of healthy and injured tendons was measured to assess the stiffness recovery kinetics after injury. At day 14 tendon stiffness was
MA NU
still significantly lower than healthy tendon. Data represent means ± SEM; *: p-value < 0.05. (C) Histological analyses of Achilles tendons performed 14 days after injection showed hypercellularity in the healing region and disorganized collagen
ED
fibers induced by the lesion.
Figure 2 – In vivo gene transfer efficiency in injured Achilles tendons. (A) Luciferase
PT
activity in injured Achilles tendons 24h transfection of luciferase gene complexed with PTG1 or Lip100. (B) Kinetic of luciferase expression in injured Achilles tendons upon
CE
transfection with 20 µg of pLuc naked or complexed with Lip100 liposomes(C).
AC
Representative images of bioluminescence after injection of Lip100 lipoplexes on the right hand side of the injured Achilles tendon.
Figure 3 – Wound-healing analyses 14 days after transfection of injured Achilles tendons with therapeutic genes. (A) Representative micrographs of sliced tendons from each treated and controls groups. (B) Biomechanical analyses are presented; stiffness was calculated for each group. HES coloration, observation at 4× objective. Data represent means ± SEM; *: p-value < 0.05; **: p-value < 0.01.
20
ACCEPTED MANUSCRIPT Figure 4 – Transfection efficiency of primary tenocytes cell cultures. (A) Cells were transfected with Lip100 lipoplexes and PTG1 polyplexes containing pLuc and the
T
luciferase activity was measured 24 hours after transfection. (B) Cells were
RI P
transfected with PTG1 polyplexes containing plasmids encoding either PDGF (pPDGF), fibromodulin (pFmod) or lumican (pLum). Detection of the PDGF,
SC
Fibromodulin and Lumican human transcripts 18 hours after transfection was performed by qPCR. Data represent means of 3 experiments ± SD. (C)
MA NU
Representative fluorescence images and flow cytometry analysis of tenocytes 24 hours post-transfection with PTG1 polyplexes containing pEGFP. FL1 channel was used for GFP detection and FL3 channel for detection of the cell death marker
ED
propidium iodide.
PT
Figure 5 – Wound-healing effect on PDGF, Fmod and Lum transfected primary tenocytes cell cultures. (A) Representative images of the wound-healing model for
CE
transfections at 18-hour time point. (B) The corresponding curves of the wound
AC
closure percentage measured at 18, 43 and 72 h after transfection. (C) The main differences of wound closure were observed at 18 hours. (D) Effect of conditioned medium recovered from transfected tenocytes overexpressing the indicated genes on wound-healing. Shown are the means of percentages of wound closure obtained at 18 hours after treatment ± SD. *: p-value < 0.05%.
Figure 6 – Proliferation and migration analyses after transfection of primary tenocytes cell cultures. (A) Proliferation analysis of tenocytes after PTG1 polyplexes transfection of pFmod or pLum performed by MTT assay. (B) Tenocytes transwell
21
ACCEPTED MANUSCRIPT migration analysis 18 hours after PTG1 polyplexes transfection of pFmod or pLum.
AC
CE
PT
ED
MA NU
SC
RI P
T
Data represent means of two independent experiments ± SD.
22
ACCEPTED MANUSCRIPT References
1.
M. Spallek, W. Kuhn, S. Uibel, A. van Mark and D. Quarcoo. Work-related
T
musculoskeletal disorders in the automotive industry due to repetitive work -
2.
RI P
implications for rehabilitation. J Occup Med Toxicol 2010;5:6-12.
D. G. Greenhalgh, K. H. Sprugel, M. J. Murray and R. Ross. PDGF and FGF
SC
stimulate wound healing in the genetically diabetic mouse. Am J Pathol
3.
MA NU
1990;136:1235-46.
Y. Hou, Z. Mao, X. Wei, L. Lin, L. Chen, H. Wang, et al. Effects of transforming
growth factor-beta1 and vascular endothelial growth factor 165 gene transfer on Achilles tendon healing. Matrix Biol 2009;28:324-35. S. Thomopoulos, R. Das, M. J. Silva, S. Sakiyama-Elbert, F. L. Harwood, E.
ED
4.
Zampiakis, et al. Enhanced flexor tendon healing through controlled delivery of
S. Thomopoulos, M. Zaegel, R. Das, F. L. Harwood, M. J. Silva, D. Amiel, et
CE
5.
PT
PDGF-BB. J Orthop Res 2009;27:1209-15.
al. PDGF-BB released in tendon repair using a novel delivery system promotes cell
6.
AC
proliferation and collagen remodeling. J Orthop Res 2007;25:1358-68. K. A. Hildebrand, C. B. Frank and D. A. Hart. Gene intervention in ligament
and tendon: current status, challenges, future directions. Gene Ther 2004;11:368-78. 7.
C. H. Chen, Y. Cao, Y. F. Wu, A. J. Bais, J. S. Gao and J. B. Tang. Tendon
healing in vivo: gene expression and production of multiple growth factors in early tendon healing period. J Hand Surg Am 2008;33:1834-42. 8.
I. Ozkan, K. Shino, N. Nakamura, T. Natsuume, N. Matsumoto, S. Horibe, et
al. Direct in vivo gene transfer to healing rat patellar ligament by intra-arterial delivery of haemagglutinating virus of Japan liposomes. Eur J Clin Invest 1999;29:63-7.
23
ACCEPTED MANUSCRIPT 9.
N. Nakamura, K. Shino, T. Natsuume, S. Horibe, N. Matsumoto, Y. Kaneda, et
al. Early biological effect of in vivo gene transfer of platelet-derived growth factor
P. Bolt, A. N. Clerk, H. H. Luu, Q. Kang, J. L. Kummer, Z. L. Deng, et al. BMP-
RI P
10.
T
(PDGF)-B into healing patellar ligament. Gene Ther 1998;5:1165-70.
14 gene therapy increases tendon tensile strength in a rat model of Achilles tendon
11.
SC
injury. J Bone Joint Surg Am 2007;89:1315-20.
X. T. Wang, P. Y. Liu and J. B. Tang. Tendon healing in vitro: genetic
MA NU
modification of tenocytes with exogenous PDGF gene and promotion of collagen gene expression. J Hand Surg [Am] 2004;29:884-90. 12.
J. H. Yoon and J. Halper. Tendon proteoglycans: biochemistry and function. J
Musculoskelet Neuronal Interact 2005;5:22-34.
Y. Ezura, S. Chakravarti, A. Oldberg, I. Chervoneva and D. E. Birk. Differential
ED
13.
PT
expression of lumican and fibromodulin regulate collagen fibrillogenesis in developing mouse tendons. J Cell Biol 2000;151:779-88. S. Saika, Y. Ohnishi, A. Ooshima, C. Y. Liu and W. W. Kao. Epithelial repair:
CE
14.
15.
AC
roles of extracellular matrix. Cornea 2002;21:S23-9. Z. Zheng, C. Nguyen, X. Zhang, H. Khorasani, J. Z. Wang, J. N. Zara, et al.
Delayed wound closure in fibromodulin-deficient mice is associated with increased TGF-beta3 signaling. J Invest Dermatol 2011;131:769-78. 16.
R. Boykiw, P. Sciore, C. Reno, L. Marchuk, C. B. Frank and D. A. Hart. Altered
levels of extracellular matrix molecule mRNA in healing rabbit ligaments. Matrix Biol 1998;17:371-8. 17.
J. Pingel, U. Fredberg, K. Qvortrup, J. O. Larsen, P. Schjerling, K. Heinemeier,
et al. Local biochemical and morphological differences in human Achilles tendinopathy: a case control study. BMC Musculoskelet Disord 2012;13:53-66.
24
ACCEPTED MANUSCRIPT 18.
Y. Geng, D. McQuillan and P. J. Roughley. SLRP interaction can protect
collagen fibrils from cleavage by collagenases. Matrix Biol 2006;25:484-91. M. Mevel, C. Neveu, C. Goncalves, J. J. Yaouanc, C. Pichon, P. A. Jaffres, et
T
19.
RI P
al. Novel neutral imidazole-lipophosphoramides for transfection assays. Chem Commun (Camb) 2008:3124-6.
M. Mevel, G. Breuzard, J. J. Yaouanc, J. C. Clement, P. Lehn, C. Pichon, et al.
SC
20.
Synthesis and transfection activity of new cationic phosphoramidate lipids: high
21.
MA NU
efficiency of an imidazolium derivative. Chembiochem 2008;9:1462-71. E. Bertrand, C. Goncalves, L. Billiet, J. P. Gomez, C. Pichon, H. Cheradame,
et al. Histidinylated linear PEI: a new efficient non-toxic polymer for gene transfer. Chem Commun (Camb) 2011;47:12547-9.
J. Sultana, M. R. Molla, M. Kamal, M. Shahidullah, F. Begum and M. A.
ED
22.
PT
Bashar. Histological differences in wound healing in Maxillofacial region in patients with or without risk factors. Bangladesh J Pathol 2009; 24:3-8. P. Midoux and M. Monsigny. Efficient gene transfer by histidylated
CE
23.
24.
AC
polylysine/pDNA complexes. Bioconjug Chem 1999;10:406-11. S. J. Warden. Animal models for the study of tendinopathy. Br J Sports Med
2007;41:232-40. 25.
A. Suwalski, H. Dabboue, A. Delalande, S. F. Bensamoun, F. Canon, P.
Midoux, et al. Accelerated Achilles tendon healing by PDGF gene delivery with mesoporous silica nanoparticles. Biomaterials 2010;31:5237-45. 26.
L. Billiet, J. P. Gomez, M. Berchel, P. A. Jaffres, T. Le Gall, T. Montier, et al.
Gene transfer by chemical vectors, and endocytosis routes of polyplexes, lipoplexes and lipopolyplexes in a myoblast cell line. Biomaterials 2012;33:2980-90.
25
ACCEPTED MANUSCRIPT 27.
K. D. Burroughs, S. E. Dunn, J. C. Barrett and J. A. Taylor. Insulin-like growth
factor-I: a key regulator of human cancer risk? J Natl Cancer Inst 1999;91:579-81. A. G. Renehan, M. Zwahlen, C. Minder, S. T. O'Dwyer, S. M. Shalet and M.
T
28.
RI P
Egger. Insulin-like growth factor (IGF)-I, IGF binding protein-3, and cancer risk: systematic review and meta-regression analysis. Lancet 2004;363:1346-53. S. Werner and R. Grose. Regulation of wound healing by growth factors and
cytokines. Physiol Rev 2003;83:835-70.
E. Mormone, Y. Lu, X. Ge, M. I. Fiel and N. Nieto. Fibromodulin, an oxidative
MA NU
30.
SC
29.
stress-sensitive proteoglycan, regulates the fibrogenic response to liver injury in mice. Gastroenterology 2012;142:612-621 e5. 31.
W. F. Libby, R. Berger, J. F. Mead, G. V. Alexander and J. F. Ross.
ED
Replacement Rates for Human Tissue from Atmospheric Radiocarbon. Science
32.
PT
1964;146:1170-2.
A. Hildebrand, M. Romaris, L. M. Rasmussen, D. Heinegard, D. R. Twardzik,
CE
W. A. Border, et al. Interaction of the small interstitial proteoglycans biglycan, decorin
AC
and fibromodulin with transforming growth factor beta. Biochem J 1994;302 ( Pt 2):527-34. 33.
Z. Zheng, K. S. Lee, X. Zhang, C. Nguyen, C. Hsu, J. Z. Wang, et al.
Fibromodulin-deficiency alters temporospatial expression patterns of transforming growth factor-beta ligands and receptors during adult mouse skin wound healing. PloS one 2014;9:e90817. 34.
S. Chakravarti. Functions of lumican and fibromodulin: lessons from knockout
mice. Glycoconj J 2002;19:287-93.
26
ACCEPTED MANUSCRIPT 35.
K. J. Jepsen, F. Wu, J. H. Peragallo, J. Paul, L. Roberts, Y. Ezura, et al. A
syndrome of joint laxity and impaired tendon integrity in lumican- and fibromodulin-
A. Stoff, A. A. Rivera, J. M. Mathis, S. T. Moore, N. S. Banerjee, M. Everts, et
RI P
36.
T
deficient mice. J Biol Chem 2002;277:35532-40.
al. Effect of adenoviral mediated overexpression of fibromodulin on human dermal
SC
fibroblasts and scar formation in full-thickness incisional wounds. J Mol Med 2007;85:481-96.
J. Jian, Z. Zheng, K. Zhang, T. M. Rackohn, C. Hsu, A. Levin, et al.
MA NU
37.
Fibromodulin promoted in vitro and in vivo angiogenesis. Biochem Biophys Res Commun 2013;436:530-5. 38.
C. Soo, F. Y. Hu, X. Zhang, Y. Wang, S. R. Beanes, H. P. Lorenz, et al.
ED
Differential expression of fibromodulin, a transforming growth factor-beta modulator,
AC
CE
PT
in fetal skin development and scarless repair. Am J Pathol 2000;157:423-33.
27
CE AC
Fig. 1
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
28
AC
Fig. 2
CE
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
29
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Fig. 3
30
CE AC
Fig. 4
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
31
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Fig. 5
32
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Fig. 6
33
ACCEPTED MANUSCRIPT Table 1. Primers used for qPCR. Primers
Detected transcript
Forward : 5’-CCCTCCAGATCCCCGCGACT-3’
Human Fmod
T
Reverse : 5’-CGGGAGGGAACGAAGGGCAG-3’ Forward : 5’- CCCTGGTTGAGCTGGATCTGTCCT-3’ Reverse : 5’- GGCCCCAGGATCTTGCAGAAGC-3’ Forward : 5’-AGACAGTGGCAGCTGCACGG-3’
Human PDGF
Reverse : 5’-TTCCGGTGCTTGCCCTTGGG-3’
Rat RNA6S
2331
4060
5155
Tm (°C) 68.6 68.9 67.6 69.0 65.8 70.2 65.7 68.0
AC
CE
PT
ED
MA NU
SC
Forward : 5’- CCAAGCTTATTCAGCGTCTTGTTACTCC-3’ Reverse : 5’- CCCTCGAGTCCTTCATTCTCTTGGC-3’
RI P
Human Lumican
Gene ID
34
ACCEPTED MANUSCRIPT Table 2. Physico-chemical features of Lip100 lipoplexes and PTG1 polyplexes. Lip100 liposome
PTG1 polymer
Chemical structure
O,O-dioleyl-N-[3N-(Nmethylimidazolium iodide)propylene] phosphoramidate (KLN25) and O,Odioleyl-N-histamine phosphoramidate (MM27)
linear polyethylenimine grafted with 16% histidine residues
3:1
6:1
-31 ± 0.2
RI P
ζ potential (mV)
SC
253 ± 27
134 ± 5 11 ± 1.6
AC
CE
PT
ED
MA NU
Vector/pDNA weight ratio Size (nm)
T
Name Type
35
ACCEPTED MANUSCRIPT Table 3. Histological tendon wound-healing score evaluation. Scores are ranging from 1 to 3 (1: poor, 3: good). Healthy
Untreated control
Fmod
Hypercellularity
1
3
1
Matrix continuity
3
1
Inflammation
1
3
Lum
T
Criteria evaluated
PDGF 2
2
1
1
1
2
2
AC
CE
PT
ED
MA NU
SC
RI P
2
36
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
ED
Graphical abstract
Gene transfer by histidylated vectors of fibromodulin, a proteoglycan involved
PT
in collagen fibrillogenesis, was found to enhance wound healing on a rat Achilles
CE
tendon injury model. Fourteen days after the treatment, biomechanical and histological parameters showed a healing improvement compared to the non-treated
AC
tendons. In vitro wound healing and chemotaxis assays revealed that this effect is ascribed to an induction of cell migration. In summary, fibromodulin non-viral gene therapy could be a promising new therapeutic strategy for tendon healing.
37
S1549-9634(15)00119-7 doi: 10.1016/j.nano.2015.05.004 NANO 1131
To appear in:
Nanomedicine: Nanotechnology, Biology, and Medicine
Received date: Revised date: Accepted date:
27 January 2015 30 April 2015 18 May 2015
Please cite this article as: Delalande Anthony, Gosselin Marie-Pierre, Suwalski Arnaud, Guilmain William, Leduc Chlo´e, Berchel Mathieu, Jaffr`es Paul-Alain, Baril Patrick, Midoux Patrick, Pichon Chantal, Enhanced achilles tendon healing by fibromodulin gene transfer, Nanomedicine: Nanotechnology, Biology, and Medicine (2015), doi: 10.1016/j.nano.2015.05.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Enhanced Achilles tendon healing by fibromodulin gene transfer. Anthony Delalande1*, Marie-Pierre Gosselin1*, Arnaud Suwalski1, William Guilmain1,
T
Chloé Leduc1, Mathieu Berchel2, Paul-Alain Jaffrès2, Patrick Baril1, Patrick Midoux1
RI P
and Chantal Pichon1. 1
Centre de Biophysique Moléculaire, rue Charles Sadron, Orléans 45071 CEDEX 2,
SC
France. 2
Université de Brest, Brest, France
MA NU
CEMCA, CNRS UMR 6521, IFR148 ScInBioS, Université Européenne de Bretagne,
*: These authors contributed equally to this work.
Abstract: 147 words
PT
Body text: 4962 words
AC
3 tables
CE
37 references 6 figures
ED
Corresponding author: [email protected]
1 supplementary material The authors disclose no conflict of interest. This work was supported by the Association Française contre les Myopathies (A.F.M).
1
ACCEPTED MANUSCRIPT Abstract Tendon injury is a major musculoskeletal disorder with a high public health impact.
T
We propose a non-viral based strategy of gene therapy for the treatment of tendon
RI P
injuries using histidylated vectors. Gene delivery of fibromodulin, a proteoglycan involved in collagen assembly was found to promote rat Achilles tendon repair in vivo
SC
and in vitro. In vivo liposome-based transfection of fibromodulin led to a better healing after surgical injury, biomechanical properties were better restored compared
MA NU
to untransfected control. These measures were confirmed by histological observations and scoring. To get better understandings of the mechanisms underlying fibromodulin transfection, an in vitro tendon healing model was developed.
ED
In vitro, polymer-based transfection of fibromodulin led to the best wound enclosure speed and a pronounced migration of tenocytes primary cultures was observed.
PT
These results suggest that fibromodulin non-viral gene therapy could be proposed as
Key words
CE
a new therapeutic strategy to accelerate tendon healing.
AC
Tendon healing ; Gene therapy ; Lipoplex ; Polyplex
2
ACCEPTED MANUSCRIPT 1. Background The incidence of work-related musculoskeletal disorders (WMSD) is increasing every
T
year due to modern life. Tendon injuries represent the main pathology of WMSD and
RI P
healing requires a long rest period 1. It has been proven that growth factors like PDGF, TGF-β or VEGF can considerably accelerate this wound-healing process by
proliferation and migration
SC
regulating the inflammation phase, the production of extracellular matrix and the cell 2-5
. The delivery of these growth factors has some
MA NU
limitations due to their short half-life time requiring repetitive injections and the extensive cost of recombinant protein production and purification. Gene therapy is the best approach to produce growth factors in situ and can be exploited in tendon disorders 6.
ED
Transient gene expression after gene transfer by non-viral vectors would be useful
PT
for wound-healing because molecules involved in tissue repair have to be expressed only in a short-time period 7. Synthetic vectors are chemically controlled compounds
CE
easy to handle, and show a low immunogenicity with a weak risk of transgene
AC
integration.
These last years, few reports have shown the potentiality of in vivo gene delivery for tendon regeneration. They mainly concern the delivery of genes encoding growth factors
by
using
electroporation
liposomes,
adenoviral
vectors,
silica
nanoparticles
and
8-11
. The main effect of the expression of growth factors is the
induction of tenocytes proliferation and collagen I production. But, the tendon strength depends also on a good matrix assembly which relies on the collagen fibrillogenesis that depends on the activity of proteoglycans like fibromodulin (Fmod) or lumican (Lum)
12, 13
. Moreover, proteoglycans have been shown to enhance the
3
ACCEPTED MANUSCRIPT repair and remodelling of injured cornea
14
, skin15, and ligament
16
. They have a
specific expression profile in injured tendons compared to normal tissue
17
.
T
In this study, we investigated as nanomedicine approach whether Fmod or Lum gene
RI P
transfer could improve rat Achilles tendon healing. Gene transfer was performed either with imidazole cationic lipids (Lip 100 liposomes) or histidylated linear
SC
polyethyleneimine (PTG1) in Achilles tendons in vivo and in primary cultures of tenocytes. Wound-healing effect was assessed by histological analyses and stiffness
MA NU
measurements after transfection of fibromodulin. Finally, we performed in vitro experiments including wound-healing, cell proliferation and cell migration assays to get a better understanding of the benefic effect brought by fibromodulin gene
AC
CE
PT
ED
transfer.
4
ACCEPTED MANUSCRIPT 2. Methods 2.1. Plasmids
T
pNFCMV-Luc (pLuc), a 7.5 kb homemade plasmid DNA (pDNA) encoding the
RI P
firefly luciferase under control of the strong cytomegalovirus promoter was used as a reporter gene. This plasmid has five consecutive NFκB motifs (termed NF that
SC
recognize the NFB transcription factor) inserted upstream of the promoter. The second reporter pDNA was pMaxGFPTM, a 3.5 kb plasmid encoding the eGFP gene
MA NU
(Lonza, Basel, Switzerland). pPDGF plasmid (Invitrogen, Cergy-Pontoise, France) was encoding the human PDGFB gene. Plasmids pCEP4-FBM (pFmod) and pCEP4LUM (pLum) plasmids (generous gift from Dr. Peter Roughley) were encoding human 18
. pQE30, a mock plasmid (pMock)
ED
fibromodulin and lumican genes, respectively
that did not encode any gene was used as control in wound-healing experiments.
PT
2.2. Liposomes and polymers.
Histidylated liposomes (Lip100) were prepared by mixing O,O-dioleyl-N-[3Niodide)propylene]
CE
(N-methylimidazolium
phosphoramidate
and
O,O-dioleyl-N-
19, 20
AC
histamine phosphoramidate. These lipids were synthetized as previously described . Histidylated linear polyethylenimine (PTG1) was produced by Polytheragene
(Evry, France)
21
. Lipoplexes and polyplexes were formed at vector/DNA weight ratio
of 3:1 and 6:1, respectively. The detailed procedure for DNA complexes preparation is presented in supplementary methods. 2.3. Animal studies 2.3.1. In vivo transfection Adult Wistar rats were bred in CBM animal facility at 22°C for at least one week before experiments and groups were randomly made. Experiences were conducted according to the guidelines of the French Ministry of Agriculture for 5
ACCEPTED MANUSCRIPT experiments with laboratory (law 87848, C. Pichon accreditation). Rats were anesthetized by intraperitoneal injection of ketamine (50 mg/kg) and xylazine (10
T
mg/kg) in 0.9% NaCl. Skin was incised on few millimeters in Achilles tendon region.
RI P
For tendon healing study, Achilles tendon was incised through the tendon sheath along its total length three times using a carbon steel surgical blade n°12 (Swann-
SC
Morton, Sheffield, UK) (figure 1A). Fifty microliters of lipoplexes or polyplexes containing 20 µg of plasmid DNA were slowly injected in the middle section of
MA NU
Achilles tendon using a 31G Hamilton PBS600 (Hamilton, Bonaduz, Switzerland) repeater delivery system. Skin was sutured and rats were closely observed until their wake-up. Painkiller treatments were managed by subcutaneous injection of Finadyne at 250 µl/kg (Schering-Plough, Kenilworth, NJ, USA).
ED
2.3.2. Ex vivo luciferase assay
PT
The transfection efficiency obtained in tendons of pLuc-injected rats was determined as follows. Rats were euthanized by lethal CO 2 inhalation at indicated
CE
day post-surgery. At day 1, 3, 6 and 9 post-treatment, tendons were harvested and
AC
crushed in liquid nitrogen with a mortar and a pestle. The powder was added in an ice-cold lysis buffer (Luciferase Cell Culture Lysis, Promega, Madison, USA) and the mixture was kept for 3 hours on ice before luminescence measurement. The level of luciferase activity was measured with the Lumat LB9507 luminometer (Berthold, Wildbarch, Germany) after adding 100 µl of luciferin (Luciferase Assay System, Promega) to 20 µl of tendon lysates and expressed as relative luminescence unit (RLU) per mg of proteins. Proteins content of tendons were quantified by bicinchoninic acid assay. 2.3.3. In vivo bioluminescence imaging and quantification
6
ACCEPTED MANUSCRIPT Bioluminescence imaging was conducted in vivo using a Hamamatsu Photonics cooled CCD camera mounted on a dark box chamber (Hamamatsu,
T
Japon). Signal intensity was recorded 2 minutes after luciferin injection (100 µg,
RI P
Promega) in the tendon area and quantified as the mean of photons per second (p/s) of time exposure within a region of interest prescribed over the tendon.
SC
2.3.4. Biomechanical tests
Rats were euthanized by lethal CO2 inhalation at indicated times. Achilles
MA NU
tendons were harvested and stored in PBS at -20°C before use. They were dissected and cleaned to remove muscular and adherent surrounding tissues. Then the tendon was mounted vertically on a mechanical test machine (MTS synergie 400, Eden Prairie, MN, USA) and preloaded with a velocity of 0.6 mm/min by a triple cyclic
ED
loading, stretching up to 110% of its initial length (L0). After pre-conditioning,
PT
specimens were stretched to failure at the same stretching speed and the force was recorded. The stiffness of the tendons was calculated by the linear regression of the
CE
curve when plotting the displacement (mm) versus the load (N). During all the
AC
experiments the tendons were kept wet by spraying distilled water. Each group was composed of 5 tendons (healthy group was composed of 20 tendons). Values are expressed as means ± SEM. 2.3.5. Histological analysis Tendons were harvested 14 days post-injury and were fixed in 10% pformaldehyde solution, paraffin embedded, sliced in serial sections (thickness: 4 μm), mounted on glass slides and counterstained with HES (5 tendons per group were used for histological analyses). Histological process and blind analyses were carried out by In-Cell-Art (Nantes, France). The wound healing status of the tendons was evaluated by 3 criteria: the number of nuclei (hypercellularity), the matrix continuity
7
ACCEPTED MANUSCRIPT and the presence of inflammation. For each criteria a score from 1 to 3 was given, 1: poor, 2: average, 3: good 22.
T
2.4. In vitro studies
RI P
2.4.1. Primary culture of tenocytes
Tenocytes were obtained from adult Wistar rat Achilles tendon explants.
SC
Tendons were harvested, cut in small pieces in PBS and placed in tenocyte culture medium composed of MEM alpha, 10% Fetal Calf Serum, ascorbic acid (50 µg/ml),
MA NU
primocin (100 µg/ml, InvivoGen, Toulouse, France). Tenocytes were cultured at 37°C under humidified atmosphere containing 5% CO2. Half of the medium volume was replaced every two to three days. 2.4.2. In vitro transfection
ED
Two days before experiment, 105 tenocytes were seeded on 24-well plates.
PT
Then, tenocytes were transfected with indicated DNA formulations as described in supplementary material. After 4 hours, transfection medium was removed and cells
CE
were cultured for two days in tenocytes medium. The transfection efficiency was
23
.
AC
determined by measuring the luciferase activity in cell lysates as described previously
2.4.3. Monolayer wound-healing assay The tenocyte wound-healing model was created by plating 9×104 tenocytes two days before experiments on a CytoSelect™ 24-Well Wound-Healing Assay (Cell Biolabs Inc., San Diego, CA, USA) allowing creation of a 0.9 mm-wide standardized gap between confluent tenocytes monolayers. Inserts were removed and cells were transfected with indicated polyplexes formulations for 4 hours at 37°C. Transfection medium was replaced by fresh medium and the wound closure was monitored by videomicroscopy using a Zeiss Axiovert 200M (Carl Zeiss Inc., Thornwood, NY, USA)
8
ACCEPTED MANUSCRIPT fully motorized microscope during 72 hours (scan speed: 2 images/hour). Cells were incubated in an atmosphere and temperature controlled chamber at 37°C and 5%
T
CO2. Quantitative analysis of the cell free area was performed using the Axiovision
RI P
Rel. 4.7 (Carl Zeiss Inc.). The level of wound-healing for a time x was evaluated by calculating the percentage of the cell free area at tx divided by the cell free area at
SC
the initial state. Experiments were at least repeated 6 times. In a set of experiments wound closure was followed when the tenocyte wound-healing model was incubated
2.4.4. Quantitative real-time PCR
MA NU
with medium from transfected tenocytes, collected 24 hours after transfection.
Tenocytes were harvested 18 and 43 hours post-transfection and total RNA
ED
was extracted using an RNA II NucleoSpin® mini kit (Macherey-Nagel EURL, Hoerdt, France). For all experiments, 50 ng of RNA was used for reverse transcription using
PT
200 units of M-MLV Reverse Transcriptase (New England Biolabs, MA, USA). The gene expression was evaluated from 100 ng of cDNA by qPCR using Quantifast®
CE
PCR Master Mix (Qiagen, Venlo, Netherlands) with 0.1 µM of Fmod, Lum, PDGF or
AC
RNA6S specific primers (Table 1). The qPCR was performed on a Light Cycler II 480 (Roche Applied Science, Meylan, France), the cycles consisted of 10 seconds at 95°C for DNA denaturation followed by 30 seconds at 60°C for primer annealing and polymerization repeated 40 times. Fmod, Lum and PDGF primers were specific to the human transcripts of cDNA inserted into the pFmod, pLum and pPDGF plasmids. The relative quantification was made using the ΔΔCt method using the RNA6S housekeeping gene expression.
2.4.5. Tenocytes proliferation
9
ACCEPTED MANUSCRIPT The evaluation of tenocytes proliferation was monitored by MTT colorimetric assay 18, 43 and 72 hours after transfection. Cells were incubated during 4 hours at
T
37°C, 5% CO2 in presence of 500 µg/ml of 3-(4,5-dimethylthiazol-2-yl)-2,5-
RI P
diphenyltetrazolium bromide (Sigma Aldrich, Saint Louis, MO, USA). Cells were washed and lysed with acidified isopropanol containing 3% of SDS (Sigma Aldrich)
SC
allowing the solubilization of tetrazolium salts. Quantification was made measuring optical density at 560 nm; the control was made using cells having received the same
MA NU
manipulations without transfection. Experiments were done in triplicate and repeated two times.
2.4.6. Migration analysis of tenocytes
Migration analysis was defined as the capacity of tenocytes to migrate through
ED
8 µm cell culture insert membranes (BD Falcon, Franklin Lakes, NJ, USA). Bottom
PT
wells were filled with complete medium and inserts seeded with 10 5 transfected tenocytes. Cells were incubated at 37°C in a 5% CO2 atmosphere during 18 hours.
CE
Inserts were cut and the number of cells present on the other side of the insert was
AC
evaluated by microscopy. The migration factor was calculated as the number of cells present at the bottom side of the insert after transfection divided by the number of cells present in non-transfected condition. Each condition was repeated two times in triplicate. 2.5. Statistical analysis Data were expressed by means ± SD (in vitro experiments) or SEM (in vivo experiments). All statistical differences were analyzed by a bilateral Mann–Whitney U-test when statistical difference was found by Kruskal-Wallis statistical test. XLStat 2014 software was used (Addinsoft, Paris, France), significance was defined as pvalue <0.05 (asymptotic p-value was calculated).
10
ACCEPTED MANUSCRIPT 3. Results 3.1. Achilles tendon healing model
T
We have established a tendon injury model by making a triple longitudinal
RI P
laceration wound (Figure 1A). This wound modality was chosen to keep tendon connections between muscle and bone; moreover this lesion was a good model to 24
. The timeline of the study is presented in figure 1A.
SC
reproduce tendon injury
Histological analysis and stiffness measurement of injured tendons were performed
MA NU
to validate our model. Healthy tendons are mainly composed of curly formed and parallel collagen fibers (Figure 1C). They contain very few tenocytes that are positioned along the collagen fibers. Fourteen days after tendon injury, histological
ED
analysis showed hypercellularity areas in the axis of the lesion and a disorganized collagen fibers network. Figure 1B shows biomechanical analyses of the tendon
PT
stiffness. The stiffness value reflects the state of healthiness of tendons. This value increased slowly from 3 to 14 days post-injury; tendons had a stiffness of 17.76 ±
CE
3.38 N/mm after 3 days, 20.78 ± 2.81 N/mm after 7 days and 25.75 ± 3.80 N/mm at
AC
14 days post-lesion. As comparison, healthy tendons exhibited a much greater stiffness of 33.17 ± 3.77 N/mm. The abnormal architecture of injured tendons was correlated with their reduced stiffness. Based on these data, we decided to transfer genes after lesion and to assess the effects 14 days post-injury. This time point was chosen to leave enough time for transfected genes to act on the healing process. 3.2. Gene transfer in injured Achilles tendon First, Lip100 and PTG1 vectors were used to transfer the luciferase reporter gene in injured Achilles tendons and bioluminescence was used as a read out. These two transfection reagents have never been used for in vivo transfection of Achilles
11
ACCEPTED MANUSCRIPT tendon and in vitro transfection of primary cultures of tenocytes. A summary of the physico-chemical properties of pDNA complexed with Lip100 (lipoplexes) or PTG1
23
. The chosen vector/pDNA weight ratios were those that
RI P
been previously described
T
(polyplexes) is presented in Table 2. The detailed structure of those carriers has
permitted to get a good balance between efficacy and toxicity in previous studies 24
20,
SC
. PTG1/pDNA polyplexes at ratio of 6:1 (µg:µg) were slightly positive with a (ζ
potential of +11±1.6 mV. Lip100/pDNA lipoplexes at a ratio of 3:1 exhibited a
with PTG1 than with Lip100
MA NU
negative ζ potential of -31±0.2 mV. As previously shown, pDNA was more compact 21, 25
. Rats Achilles tendons were injected with pLuc
plasmid complexed either with Lip100 or PTG1. The transfection efficiency was assessed 24 hours later (Figure 2A). The results indicate that the luciferase activity
ED
was 85 times higher after Lip100 lipoplexes than PTG1 polyplexes injection (1.7 ×
PT
106 RLU/mg of proteins vs. 2 × 104 RLU/mg of proteins). Compared to lipofection, the injection of naked pLuc gave a higher level of luciferase expression one day post-
CE
injection (Figure 2B). However, the luciferase activity dropped dramatically after 3
AC
days and was almost undetectable at day 6. By contrast, the luciferase activity that also decreased at day 3 post-lipofection was maintained at day 6 with a similar value (9.7 × 105 ± 4.95 × 104 RLU/mg of proteins vs. 1.33 × 105 ± 1.56 × 104 RLU/mg of proteins). Therefore, the luciferase expression was more stable when injection was performed with Lip100 lipoplexes compared to naked pLuc. Figure 2C shows bioluminescence representative pictures of the luciferase expression following injection of lipoplexes in the right Achilles tendon. The bioluminescence was located in the Achilles tendon area at days 1, 3 and 6 post-injection. The decrease of the luminescence intensity followed the same trend as the luciferase activity assessed ex
12
ACCEPTED MANUSCRIPT vivo (Figure 2B). The signal decreased from day 1 to day 3 but was maintained at day 6.
T
3.3. In vivo transfection of injured Achilles tendons with therapeutic genes.
RI P
Based on these results, we decided to transfer therapeutic genes by using Lip100 liposomes. Plasmids DNA encoding PDGF gene (pPDGF), Lum gene (pLum)
SC
or Fmod gene (pFmod) were used to treat Achilles tendon injured by surgical lesions. The effects of these treatments were evaluated after 14 days both by biomechanical
MA NU
and histological analyses. In figure 3A, shown are representative micrographs of sliced tendons sections treated with indicated therapeutic gene. Tendons treated either with pLum or pPDGF exhibited a higher hypercellularity area (as shown by the high number of nuclei in the healing region) compared to those treated with pFmod.
ED
However, the collagen matrix appeared much better structurally organized after the
PT
latter treatment. The tendon appearance was closer to that observed in control tendons with the presence of more aligned fibers. The histological analyses are
CE
summarized in the Table 3.
AC
Next, we measured the stiffness parameter of injured tendons 14 days posttreatment either with pFmod, pLum or pPDGF to confirm histological observations (Figure 3B). The stiffness of injured tendons was 25.75 ± 3.80 N/mm which was significantly lower than that of healthy tendons (33.17 ± 3.77 N/mm, p<0.05). Tendons treated with pFmod were stiffer than those treated either with pLum or pPDGF. Note that their stiffness value was also significantly higher than those of healthy group (44.62 ± 3.64 N/mm vs. 33.17 ± 3.77 N/mm, p<0.05). In contrast, overexpression of lumican did not restore tendon stiffness which was similar to that of untreated injured tendons (25.71 ± 3.87 N/mm vs. 25.75 ± 3.80 N/mm). Interestingly, PDGF gene transfer restored the tendon stiffness to a value close to that of healthy
13
ACCEPTED MANUSCRIPT tendon (30.00 ± 2.32 N/mm). This was in line with our previous observations after PDGF gene transfer using mesoporous silica nanoparticles25.
T
Altogether, the measurement of the stiffness parameter and the histology of treated
RI P
tendons confirm that the transfer of the fibromodulin gene improves tendon healing. 3.4. Primary tenocytes transfection with PTG1 and Lip100
SC
The transfection efficiency of primary cultures of tenocytes was tested with Lip100 and PTG1. As shown in Figure 4A, the polyplexes efficiency was 241-fold
MA NU
higher than that of lipoplexes (8.78 × 107 ± 6.70 × 106 and 3.64 × 105 ± 7.07 × 104 RLU/mg of proteins). This result confirms that most of the time in vivo efficiency of one type of vector does not predict about its in vitro efficiency and reciprocally. By
ED
using pDNA encoding eGFP reporter gene, the number of transfected tenocytes determined both by fluorescence imaging and flow cytometry was 31% after 24 hours
PT
(Figure 4C). Only 13% of cells were stained by propidium iodide indicating that PTG1 polyplexes transfection induced low cytotoxicity. Tenocytes were then transfected
CE
with PTG1 complexed either with pPDGF, pFmod or pLum. The transgene
AC
expression was assessed after 18 hours by quantitative real-time PCR (qPCR). Results indicate that all transgenes were highly expressed with a similar level (Figure 4B).
3.5. In vitro wound-healing upon Fibromodulin, Lumican and PDGF transfection The closure of the wound model in response to Fmod, Lum and PDGF gene expression after polyplexes transfection was monitored for up to 72 hours by videomicroscopy and the gap closure percentage was calculated at 18, 43 and 72 hours (Figure 5A). An accelerated gap filling was observed at early time points whatever the transfected gene. For the sake of clarity, data obtained at 18 hours were displayed on figure 5B. After pFmod transfection, 43 ± 10% of the gap area was
14
ACCEPTED MANUSCRIPT filled compared to 32 ± 9% after mock transfection (p<0.05). When tenocytes were transfected with pLum and pPDGF, the gap was filled at 44 ± 10% and 40 ± 5%,
T
respectively. These values were significantly higher compared to mock-transfected
RI P
cells (p<0.05 for both). 43 hours post-transfection, no significant difference was observed between cells transfected either with pFmod, pLum or pMock.
SC
3.6. Assessment of proliferation and migration involvement during wound-healing process of transfected tenocytes.
MA NU
The in vitro wound-healing assays suggested that the transfection of genes encoding proteoglycans improved the gap closure kinetic at early stages. This effect could occur either via an enhancement of the cell proliferation or the cell migration or both. Figure 6A shows the tenocytes relative proliferation analysis after 18, 43 and 72
ED
hours transfection either with pFmod or pLum. At 18 hours, an increased cell
PT
proliferation was seen for each transfection condition compared to the untreated control. No significant differences were observed between mock-transfected cells
CE
(1.7 ± 0.0) and those transfected with therapeutic genes, pFmod (1.8 ± 0.3) and
AC
pLum (1.9 ± 0.1) however a slight proliferation increase was obtained. Migration assays were carried out to know whether differences could occur after transfection of genes encoding proteoglycans (Figure 6B). pFmod-transfected cells exhibited a significant higher number of migrating cells through the insert compared to mock-transfected cells (227 ± 36% vs 108 ± 42%, p-value <0.05); the number of migrating non-transfected cells being taken as 100%. To go further, the effect of conditioned medium recovered from tenocytes overexpressing Fmod, Lum or PDGF was assessed. Indeed, those proteins are meant to be secreted in the cell medium allowing to get rid of possible transfection side effects. The wound closure was measured 18 hours upon incubation with
15
ACCEPTED MANUSCRIPT conditioned medium (Figure 5D). The Fmod-conditioned medium was more efficient to enhance the wound closure compared to non-treated cells (53 ± 7% vs. 35 ± 4%,
T
p<0.05). Incubation with Lum or PDGF conditioned media was not efficient as the
RI P
latter on wound-healing. The wound enclosure (38 ± 7% and 39 ± 3%, respectively) was slightly higher than that obtained with medium from mock-transfected cells.
SC
4. Discussion
MA NU
The growing number of publications related to nanomedicine applied to tendon pathologies shows and consolidates the potentiality of this application. Most of them report engraftment of genetically modified mesenchymal stem cells. Very few studies concern gene delivery approach by using non-viral systems. Recently, we reported
ED
that PDGF gene transfer by using silica mesoporous MCM-41 nanoparticles enhanced the healing of injured rat Achilles tendons25. Two injections of 10 µg of
PT
pPDGF spaced by one week were required to improve the histological architecture of
CE
injured tendons and most importantly their biomechanical properties. Comparatively, injection of pPDGF did not provide significant benefit effect in this study. By contrast,
AC
we found that one injection of 20 µg pFmod complexed with Lip100 liposomes was sufficient to induce improved structural matrix organization and stiffness of injured tendons in only 14 days. We assessed in vivo tendon and in vitro tenocytes gene delivery by using two kinds of pDNA complexes. Accordingly with the results, two remarks can be point out. First, lipoplexes made with Lip100 liposomes that provided efficient gene transfer in Achilles tendons, were weakly efficient to transfect primary cultures of tenocytes. Note that they were able to give a great transfection level of C2C12 murine myoblasts26. Thus the effectiveness of a gene delivery system is highly celldependent. Second, PTG1 polyplexes which were efficient in vitro (30% of 16
ACCEPTED MANUSCRIPT transfected tenocytes) did not transfect tendons in vivo (almost 100-times less potent than Lip100 lipoplexes). This highlights that in vitro gene transfer efficiency cannot
T
predict in vivo effectiveness. Compared to lipoplexes, the absence of efficacy of
RI P
PTG1 polyplexes could be ascribed to their positive charges which could favor interaction with negatively charged proteoglycans and thereby hamper their diffusion
SC
inside the tendon matrix.
The majority of studies related to the treatment of tendon injuries concerns the
MA NU
use of growth factors known to induce tenocytes proliferation and collagen I production. Overexpression of growth factors as IGF-I may indeed increase the risk factor of several diseases like cancer due to their mitogenic properties their expression has to be finely regulated
27, 28
, therefore
29
.
ED
Levels of gene expression obtained following in vivo Lip100 lipofection are not
PT
as high as viral vectors transduction but remains fairly stable from 3 to 6 days which is suitable for tendon healing. It is worth to note that stable and high levels of
CE
proteoglycans or growth factor expression are not always necessary since they could 30
. Fibromodulin expression and collagen fiber stiffness is closely
AC
provoke fibrosis
connected. The stiffness of Fmod-treated tendon is significantly increased compared to the healthy tendon possibly due to an overexpression of this proteoglycan. The half-life time of proteoglycans proteins as fibromodulin is known to be long (several days) as well as the half-life time of collagen proteins (more than 70 years)31. An accumulation of fibromodulin at higher levels to normal could explain the tendon stiffness increase. A well-organized matrix assembly is also needed and it can be driven by acting on fibrillogenesis. This process is regulated by proteoglycans which play a major role in maintenance of tendon structure. These molecules are also involved in
17
ACCEPTED MANUSCRIPT tenocytes proliferation by retaining/protecting growth factors as TGF-β via specific interactions32. Amongst proteoglycans, Fmod and Lum are of interest because they
14
, ligament
16
and skin
15, 33
but their action
RI P
repair and remodeling of injured cornea
T
are involved in collagen fibrillogenesis. They have been reported to play a role in the
in tendons has never been reported. When Fmod gene is knocked out, the collagen
SC
fibril architecture of connective tissues such as cornea, skin and tendon is altered leading to weak biomechanical parameters 34, 35.
MA NU
Our in vivo data indicate that the effect on wound-healing after transfection of Achilles tendons with Fmod gene was superior compared to that of Lum and PDGF gene. Injection of pLum induced high hypercellularity areas compared to
ED
untransfected tendons. This observation is often correlated with a good prognosis in the healing process as these cells will produce the new matrix. However, lumican
PT
overexpression did not improve tendon stiffness after injury. Data from in vitro wound-healing and migration assays indicate that Fmod is
CE
more potent to promote cell migration compared to Lum and PDGF. The involvement 15, 32, 33
. When Fmod is absent,
AC
of Fmod on healing has been reported in skin wound
a delayed healing was observed. The delivery of Fmod gene by adenoviral vectors has been found to improve the healing of rabbit skin wound
36
. In this study, two
injections of adenoviral vectors expressing Fmod were required to obtain an improvement of skin wound-healing. In addition to cell migration and cell proliferation, angiogenesis and inflammation processes are also crucial for wound-healing. Fibromodulin has been reported to improve angiogenesis during wound-healing
37
. Similar effect of Fmod
could likely be involved in our study. Fmod expression has also been found to be inversely correlated with a scar formation during both fetal skin development and
18
ACCEPTED MANUSCRIPT adult wound repair
38
. In a scarless fetal wound, Fmod is 3 to 5 fold upregulated
whilst there is only a small increase (1.3 fold) of its overexpression in adult wound
T
between 3 and 5 days post-injury. These results combined with ours suggest that
RI P
overexpressing Fmod could be a promising strategy to manipulate adult wound to a scarless repair.
SC
5. Conclusion
MA NU
Improved repair of injured tendon was obtained upon only one injection of 20 µg of pDNA encoding fibromodulin gene complexed with histidylated liposomes. Based on the stiffness parameter and the histological analysis of fibromodulin-treated tendons, the repair is better and closer to that of healthy tendons than after lumican
ED
or PDGF gene transfer. These results could be explained by the induction of cell migration and with a lesser extent of cell proliferation observed in vitro after
PT
fibromodulin overexpression. This repairing activity could be directly ascribed to
CE
fibromodulin expression. Therefore, fibromodulin could be a good alternative to
AC
growth factors for tendon healing.
19
ACCEPTED MANUSCRIPT Figure captions Figure 1 – In vivo Achilles tendon healing model. (A) The lesion was induced at day
T
0 by a triple longitudinal laceration of the Achilles tendon. Then, lipoplexes or
RI P
polyplexes were injected before suturing the tendon. When luciferase reporter gene was used, luciferase activity assay was performed at day 1, 3 and 6 to evaluate the
SC
gene transfer efficiency. (B) Stiffness of healthy and injured tendons was measured to assess the stiffness recovery kinetics after injury. At day 14 tendon stiffness was
MA NU
still significantly lower than healthy tendon. Data represent means ± SEM; *: p-value < 0.05. (C) Histological analyses of Achilles tendons performed 14 days after injection showed hypercellularity in the healing region and disorganized collagen
ED
fibers induced by the lesion.
Figure 2 – In vivo gene transfer efficiency in injured Achilles tendons. (A) Luciferase
PT
activity in injured Achilles tendons 24h transfection of luciferase gene complexed with PTG1 or Lip100. (B) Kinetic of luciferase expression in injured Achilles tendons upon
CE
transfection with 20 µg of pLuc naked or complexed with Lip100 liposomes(C).
AC
Representative images of bioluminescence after injection of Lip100 lipoplexes on the right hand side of the injured Achilles tendon.
Figure 3 – Wound-healing analyses 14 days after transfection of injured Achilles tendons with therapeutic genes. (A) Representative micrographs of sliced tendons from each treated and controls groups. (B) Biomechanical analyses are presented; stiffness was calculated for each group. HES coloration, observation at 4× objective. Data represent means ± SEM; *: p-value < 0.05; **: p-value < 0.01.
20
ACCEPTED MANUSCRIPT Figure 4 – Transfection efficiency of primary tenocytes cell cultures. (A) Cells were transfected with Lip100 lipoplexes and PTG1 polyplexes containing pLuc and the
T
luciferase activity was measured 24 hours after transfection. (B) Cells were
RI P
transfected with PTG1 polyplexes containing plasmids encoding either PDGF (pPDGF), fibromodulin (pFmod) or lumican (pLum). Detection of the PDGF,
SC
Fibromodulin and Lumican human transcripts 18 hours after transfection was performed by qPCR. Data represent means of 3 experiments ± SD. (C)
MA NU
Representative fluorescence images and flow cytometry analysis of tenocytes 24 hours post-transfection with PTG1 polyplexes containing pEGFP. FL1 channel was used for GFP detection and FL3 channel for detection of the cell death marker
ED
propidium iodide.
PT
Figure 5 – Wound-healing effect on PDGF, Fmod and Lum transfected primary tenocytes cell cultures. (A) Representative images of the wound-healing model for
CE
transfections at 18-hour time point. (B) The corresponding curves of the wound
AC
closure percentage measured at 18, 43 and 72 h after transfection. (C) The main differences of wound closure were observed at 18 hours. (D) Effect of conditioned medium recovered from transfected tenocytes overexpressing the indicated genes on wound-healing. Shown are the means of percentages of wound closure obtained at 18 hours after treatment ± SD. *: p-value < 0.05%.
Figure 6 – Proliferation and migration analyses after transfection of primary tenocytes cell cultures. (A) Proliferation analysis of tenocytes after PTG1 polyplexes transfection of pFmod or pLum performed by MTT assay. (B) Tenocytes transwell
21
ACCEPTED MANUSCRIPT migration analysis 18 hours after PTG1 polyplexes transfection of pFmod or pLum.
AC
CE
PT
ED
MA NU
SC
RI P
T
Data represent means of two independent experiments ± SD.
22
ACCEPTED MANUSCRIPT References
1.
M. Spallek, W. Kuhn, S. Uibel, A. van Mark and D. Quarcoo. Work-related
T
musculoskeletal disorders in the automotive industry due to repetitive work -
2.
RI P
implications for rehabilitation. J Occup Med Toxicol 2010;5:6-12.
D. G. Greenhalgh, K. H. Sprugel, M. J. Murray and R. Ross. PDGF and FGF
SC
stimulate wound healing in the genetically diabetic mouse. Am J Pathol
3.
MA NU
1990;136:1235-46.
Y. Hou, Z. Mao, X. Wei, L. Lin, L. Chen, H. Wang, et al. Effects of transforming
growth factor-beta1 and vascular endothelial growth factor 165 gene transfer on Achilles tendon healing. Matrix Biol 2009;28:324-35. S. Thomopoulos, R. Das, M. J. Silva, S. Sakiyama-Elbert, F. L. Harwood, E.
ED
4.
Zampiakis, et al. Enhanced flexor tendon healing through controlled delivery of
S. Thomopoulos, M. Zaegel, R. Das, F. L. Harwood, M. J. Silva, D. Amiel, et
CE
5.
PT
PDGF-BB. J Orthop Res 2009;27:1209-15.
al. PDGF-BB released in tendon repair using a novel delivery system promotes cell
6.
AC
proliferation and collagen remodeling. J Orthop Res 2007;25:1358-68. K. A. Hildebrand, C. B. Frank and D. A. Hart. Gene intervention in ligament
and tendon: current status, challenges, future directions. Gene Ther 2004;11:368-78. 7.
C. H. Chen, Y. Cao, Y. F. Wu, A. J. Bais, J. S. Gao and J. B. Tang. Tendon
healing in vivo: gene expression and production of multiple growth factors in early tendon healing period. J Hand Surg Am 2008;33:1834-42. 8.
I. Ozkan, K. Shino, N. Nakamura, T. Natsuume, N. Matsumoto, S. Horibe, et
al. Direct in vivo gene transfer to healing rat patellar ligament by intra-arterial delivery of haemagglutinating virus of Japan liposomes. Eur J Clin Invest 1999;29:63-7.
23
ACCEPTED MANUSCRIPT 9.
N. Nakamura, K. Shino, T. Natsuume, S. Horibe, N. Matsumoto, Y. Kaneda, et
al. Early biological effect of in vivo gene transfer of platelet-derived growth factor
P. Bolt, A. N. Clerk, H. H. Luu, Q. Kang, J. L. Kummer, Z. L. Deng, et al. BMP-
RI P
10.
T
(PDGF)-B into healing patellar ligament. Gene Ther 1998;5:1165-70.
14 gene therapy increases tendon tensile strength in a rat model of Achilles tendon
11.
SC
injury. J Bone Joint Surg Am 2007;89:1315-20.
X. T. Wang, P. Y. Liu and J. B. Tang. Tendon healing in vitro: genetic
MA NU
modification of tenocytes with exogenous PDGF gene and promotion of collagen gene expression. J Hand Surg [Am] 2004;29:884-90. 12.
J. H. Yoon and J. Halper. Tendon proteoglycans: biochemistry and function. J
Musculoskelet Neuronal Interact 2005;5:22-34.
Y. Ezura, S. Chakravarti, A. Oldberg, I. Chervoneva and D. E. Birk. Differential
ED
13.
PT
expression of lumican and fibromodulin regulate collagen fibrillogenesis in developing mouse tendons. J Cell Biol 2000;151:779-88. S. Saika, Y. Ohnishi, A. Ooshima, C. Y. Liu and W. W. Kao. Epithelial repair:
CE
14.
15.
AC
roles of extracellular matrix. Cornea 2002;21:S23-9. Z. Zheng, C. Nguyen, X. Zhang, H. Khorasani, J. Z. Wang, J. N. Zara, et al.
Delayed wound closure in fibromodulin-deficient mice is associated with increased TGF-beta3 signaling. J Invest Dermatol 2011;131:769-78. 16.
R. Boykiw, P. Sciore, C. Reno, L. Marchuk, C. B. Frank and D. A. Hart. Altered
levels of extracellular matrix molecule mRNA in healing rabbit ligaments. Matrix Biol 1998;17:371-8. 17.
J. Pingel, U. Fredberg, K. Qvortrup, J. O. Larsen, P. Schjerling, K. Heinemeier,
et al. Local biochemical and morphological differences in human Achilles tendinopathy: a case control study. BMC Musculoskelet Disord 2012;13:53-66.
24
ACCEPTED MANUSCRIPT 18.
Y. Geng, D. McQuillan and P. J. Roughley. SLRP interaction can protect
collagen fibrils from cleavage by collagenases. Matrix Biol 2006;25:484-91. M. Mevel, C. Neveu, C. Goncalves, J. J. Yaouanc, C. Pichon, P. A. Jaffres, et
T
19.
RI P
al. Novel neutral imidazole-lipophosphoramides for transfection assays. Chem Commun (Camb) 2008:3124-6.
M. Mevel, G. Breuzard, J. J. Yaouanc, J. C. Clement, P. Lehn, C. Pichon, et al.
SC
20.
Synthesis and transfection activity of new cationic phosphoramidate lipids: high
21.
MA NU
efficiency of an imidazolium derivative. Chembiochem 2008;9:1462-71. E. Bertrand, C. Goncalves, L. Billiet, J. P. Gomez, C. Pichon, H. Cheradame,
et al. Histidinylated linear PEI: a new efficient non-toxic polymer for gene transfer. Chem Commun (Camb) 2011;47:12547-9.
J. Sultana, M. R. Molla, M. Kamal, M. Shahidullah, F. Begum and M. A.
ED
22.
PT
Bashar. Histological differences in wound healing in Maxillofacial region in patients with or without risk factors. Bangladesh J Pathol 2009; 24:3-8. P. Midoux and M. Monsigny. Efficient gene transfer by histidylated
CE
23.
24.
AC
polylysine/pDNA complexes. Bioconjug Chem 1999;10:406-11. S. J. Warden. Animal models for the study of tendinopathy. Br J Sports Med
2007;41:232-40. 25.
A. Suwalski, H. Dabboue, A. Delalande, S. F. Bensamoun, F. Canon, P.
Midoux, et al. Accelerated Achilles tendon healing by PDGF gene delivery with mesoporous silica nanoparticles. Biomaterials 2010;31:5237-45. 26.
L. Billiet, J. P. Gomez, M. Berchel, P. A. Jaffres, T. Le Gall, T. Montier, et al.
Gene transfer by chemical vectors, and endocytosis routes of polyplexes, lipoplexes and lipopolyplexes in a myoblast cell line. Biomaterials 2012;33:2980-90.
25
ACCEPTED MANUSCRIPT 27.
K. D. Burroughs, S. E. Dunn, J. C. Barrett and J. A. Taylor. Insulin-like growth
factor-I: a key regulator of human cancer risk? J Natl Cancer Inst 1999;91:579-81. A. G. Renehan, M. Zwahlen, C. Minder, S. T. O'Dwyer, S. M. Shalet and M.
T
28.
RI P
Egger. Insulin-like growth factor (IGF)-I, IGF binding protein-3, and cancer risk: systematic review and meta-regression analysis. Lancet 2004;363:1346-53. S. Werner and R. Grose. Regulation of wound healing by growth factors and
cytokines. Physiol Rev 2003;83:835-70.
E. Mormone, Y. Lu, X. Ge, M. I. Fiel and N. Nieto. Fibromodulin, an oxidative
MA NU
30.
SC
29.
stress-sensitive proteoglycan, regulates the fibrogenic response to liver injury in mice. Gastroenterology 2012;142:612-621 e5. 31.
W. F. Libby, R. Berger, J. F. Mead, G. V. Alexander and J. F. Ross.
ED
Replacement Rates for Human Tissue from Atmospheric Radiocarbon. Science
32.
PT
1964;146:1170-2.
A. Hildebrand, M. Romaris, L. M. Rasmussen, D. Heinegard, D. R. Twardzik,
CE
W. A. Border, et al. Interaction of the small interstitial proteoglycans biglycan, decorin
AC
and fibromodulin with transforming growth factor beta. Biochem J 1994;302 ( Pt 2):527-34. 33.
Z. Zheng, K. S. Lee, X. Zhang, C. Nguyen, C. Hsu, J. Z. Wang, et al.
Fibromodulin-deficiency alters temporospatial expression patterns of transforming growth factor-beta ligands and receptors during adult mouse skin wound healing. PloS one 2014;9:e90817. 34.
S. Chakravarti. Functions of lumican and fibromodulin: lessons from knockout
mice. Glycoconj J 2002;19:287-93.
26
ACCEPTED MANUSCRIPT 35.
K. J. Jepsen, F. Wu, J. H. Peragallo, J. Paul, L. Roberts, Y. Ezura, et al. A
syndrome of joint laxity and impaired tendon integrity in lumican- and fibromodulin-
A. Stoff, A. A. Rivera, J. M. Mathis, S. T. Moore, N. S. Banerjee, M. Everts, et
RI P
36.
T
deficient mice. J Biol Chem 2002;277:35532-40.
al. Effect of adenoviral mediated overexpression of fibromodulin on human dermal
SC
fibroblasts and scar formation in full-thickness incisional wounds. J Mol Med 2007;85:481-96.
J. Jian, Z. Zheng, K. Zhang, T. M. Rackohn, C. Hsu, A. Levin, et al.
MA NU
37.
Fibromodulin promoted in vitro and in vivo angiogenesis. Biochem Biophys Res Commun 2013;436:530-5. 38.
C. Soo, F. Y. Hu, X. Zhang, Y. Wang, S. R. Beanes, H. P. Lorenz, et al.
ED
Differential expression of fibromodulin, a transforming growth factor-beta modulator,
AC
CE
PT
in fetal skin development and scarless repair. Am J Pathol 2000;157:423-33.
27
CE AC
Fig. 1
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
28
AC
Fig. 2
CE
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
29
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Fig. 3
30
CE AC
Fig. 4
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
31
PT
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Fig. 5
32
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Fig. 6
33
ACCEPTED MANUSCRIPT Table 1. Primers used for qPCR. Primers
Detected transcript
Forward : 5’-CCCTCCAGATCCCCGCGACT-3’
Human Fmod
T
Reverse : 5’-CGGGAGGGAACGAAGGGCAG-3’ Forward : 5’- CCCTGGTTGAGCTGGATCTGTCCT-3’ Reverse : 5’- GGCCCCAGGATCTTGCAGAAGC-3’ Forward : 5’-AGACAGTGGCAGCTGCACGG-3’
Human PDGF
Reverse : 5’-TTCCGGTGCTTGCCCTTGGG-3’
Rat RNA6S
2331
4060
5155
Tm (°C) 68.6 68.9 67.6 69.0 65.8 70.2 65.7 68.0
AC
CE
PT
ED
MA NU
SC
Forward : 5’- CCAAGCTTATTCAGCGTCTTGTTACTCC-3’ Reverse : 5’- CCCTCGAGTCCTTCATTCTCTTGGC-3’
RI P
Human Lumican
Gene ID
34
ACCEPTED MANUSCRIPT Table 2. Physico-chemical features of Lip100 lipoplexes and PTG1 polyplexes. Lip100 liposome
PTG1 polymer
Chemical structure
O,O-dioleyl-N-[3N-(Nmethylimidazolium iodide)propylene] phosphoramidate (KLN25) and O,Odioleyl-N-histamine phosphoramidate (MM27)
linear polyethylenimine grafted with 16% histidine residues
3:1
6:1
-31 ± 0.2
RI P
ζ potential (mV)
SC
253 ± 27
134 ± 5 11 ± 1.6
AC
CE
PT
ED
MA NU
Vector/pDNA weight ratio Size (nm)
T
Name Type
35
ACCEPTED MANUSCRIPT Table 3. Histological tendon wound-healing score evaluation. Scores are ranging from 1 to 3 (1: poor, 3: good). Healthy
Untreated control
Fmod
Hypercellularity
1
3
1
Matrix continuity
3
1
Inflammation
1
3
Lum
T
Criteria evaluated
PDGF 2
2
1
1
1
2
2
AC
CE
PT
ED
MA NU
SC
RI P
2
36
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
ED
Graphical abstract
Gene transfer by histidylated vectors of fibromodulin, a proteoglycan involved
PT
in collagen fibrillogenesis, was found to enhance wound healing on a rat Achilles
CE
tendon injury model. Fourteen days after the treatment, biomechanical and histological parameters showed a healing improvement compared to the non-treated
AC
tendons. In vitro wound healing and chemotaxis assays revealed that this effect is ascribed to an induction of cell migration. In summary, fibromodulin non-viral gene therapy could be a promising new therapeutic strategy for tendon healing.
37