The GDF5 mutant BB-1 enhances the bone formation induced by an injectable, poly(l-lactide-co-glycolide) acid (PLGA) fiber-reinforced, brushite-forming cement in a sheep defect model of lumbar osteopenia

Accepted Manuscript Title: The GDF5 mutant BB-1 enhances the bone formation induced by an injectable, PLGA-fiber reinforced, brushite-forming cement in a sheep defect model of lumbar osteopenia Author: Francesca Gunnella, Elke Kunisch, Stefan Maenz, Victoria Horbert, Long Xin, Joerg Mika, Juliane Borowski, Sabine Bischoff, Harald Schubert, Andre Sachse, Bernhard Illerhaus, Jens Günster, Jörg Bossert, Klaus D. Jandt, Frank Plöger, Raimund W. Kinne, Olaf Brinkmann, Matthias Bungartz PII: DOI: Reference:

S1529-9430(17)31051-3 https://doi.org/doi:10.1016/j.spinee.2017.10.002 SPINEE 57510

To appear in:

The Spine Journal

Received date: Revised date: Accepted date:

18-7-2017 15-9-2017 2-10-2017

Please cite this article as: Francesca Gunnella, Elke Kunisch, Stefan Maenz, Victoria Horbert, Long Xin, Joerg Mika, Juliane Borowski, Sabine Bischoff, Harald Schubert, Andre Sachse, Bernhard Illerhaus, Jens Günster, Jörg Bossert, Klaus D. Jandt, Frank Plöger, Raimund W. Kinne, Olaf Brinkmann, Matthias Bungartz, The GDF5 mutant BB-1 enhances the bone formation induced by an injectable, PLGA-fiber reinforced, brushite-forming cement in a sheep defect model of lumbar osteopenia, The Spine Journal (2017), https://doi.org/doi:10.1016/j.spinee.2017.10.002. 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.

1

The GDF5 mutant BB-1 enhances the bone formation induced by an injectable, PLGA-fiber

2

reinforced, brushite-forming cement in a sheep defect model of lumbar osteopenia

3 4

Francesca Gunnellaa, Elke Kunischa, Stefan Maenzc,f, Victoria Horberta, Long Xina, Joerg Mikaa,1,2,

5

Juliane Borowskia, Sabine Bischoffd, Harald Schubertd, Andre Sachseb, Bernhard Illerhause, Jens

6

Günstere, Jörg Bossertc, Klaus D. Jandtc,f,g, Frank Plögerh, Raimund W. Kinnea*, Olaf Brinkmanna,b,

7

Matthias Bungartza,b

8

a

9

Waldkrankenhaus “Rudolf Elle,” Eisenberg, Germany

Experimental Rheumatology Unit, Department of Orthopedics, Jena University Hospital,

10

b

11

“Rudolf Elle”, Eisenberg, Germany

12

c

13

Jena, Germany

14

d

Institute of Laboratory Animal Sciences and Welfare, Jena University Hospital, Jena, Germany

15

e

BAM Bundesanstalt für Materialforschung und –prüfung (BAM), Berlin, Germany

16

f

Jena Center for Soft Matter (JCSM), Friedrich Schiller University Jena, Germany

17

g

Jena School for Microbial Communication (JSMC), Friedrich Schiller University Jena, Germany

18

h

BIOPHARM GmbH, Heidelberg, Germany

Chair of Orthopedics, Department of Orthopedics, Jena University Hospital, Waldkrankenhaus

Chair of Materials Science, Otto Schott Institute of Materials Research, Friedrich Schiller University

19 20

Current address: 1Laboratory of Experimental Trauma Surgery, Justus-Liebig-University Giessen,

21

35394 Giessen, Germany; 2Department of Trauma, Hand and Reconstructive Surgery Giessen,

22

University Hospital Giessen-Marburg, Campus Giessen, Rudolf-Buchheim-Str. 7, 35385 Giessen,

23

Germany

24

*Corresponding author: Raimund W. Kinne; Experimental Rheumatology Unit, Department of

25

Orthopedics, Jena University Hospital, Waldkrankenhaus “Rudolf Elle”, Klosterlausnitzer Str. 81,

26

07607 Eisenberg; Germany. E-Mail: [email protected], phone: 0049 (0) 36691

27

81228, fax: 0049 (0) 36691 81226.

28 1 Page 1 of 25

1

Abstract

2

Background context: Targeted delivery of osteoinductive bone morphogenetic proteins (BMPs; e.g.,

3

GDF5)

4

vertebroplasty/kyphoplasty of osteoporotic vertebral fractures, may be required to counteract

5

augmented local bone catabolism and support complete bone regeneration. The biologically optimized

6

GDF5 mutant BB-1 may represent an attractive drug candidate for this purpose.

7

Purpose: The aim of the current study was to test an injectable, poly (l-lactide-co-glycolide) acid

8

(PLGA)-fiber reinforced, brushite-forming cement (CPC) containing low-dose BB-1 in a sheep lumbar

9

osteopenia model.

in

bioresorbable

calcium

phosphate

cement

(CPC),

potentially

suitable

for

10

Study Design/ Setting: This is a prospective experimental animal study.

11

Methods: Bone defects (diameter 5 mm) were generated in aged, osteopenic female sheep and filled

12

with fiber-reinforced CPC alone (L4; CPC+fibers) or with CPC containing different dosages of BB-1

13

(L5; CPC+fibers+BB-1; 5, 100, and 500 µg BB-1; n = 6 each). The results were compared to those of

14

untouched controls (L1). Three and 9 months after the operation, structural and functional effects of

15

the CPC (± BB-1) were analyzed ex vivo by measuring: i) bone mineral density; ii) bone structure,

16

i.e., bone volume/total volume (assessed by micro-CT and histomorphometry), trabecular thickness,

17

and trabecular number; iii) bone formation, i.e., osteoid volume/bone volume, osteoid surface/bone

18

surface, osteoid thickness, mineralizing surface/bone surface, mineral apposition rate, and bone

19

formation rate/bone surface; iv) bone resorption, i.e., eroded surface/bone surface; and v)

20

compressive strength.

21

Results: Compared to untouched controls (L1), CPC+fibers (L4) and/or CPC+fibers+BB-1 (L5)

22

significantly improved all parameters of bone formation, bone resorption, and bone structure. These

23

effects were observed at 3 and 9 months, but were less pronounced for some parameters at 9 months.

24

Compared to CPC without BB-1, additional significant effects of BB-1 were demonstrated for bone

25

mineral density; bone structure (bone volume/total volume, trabecular thickness, trabecular number);

26

and bone formation (osteoid surface/bone surface and mineralizing surface/bone surface). The BB-1

27

effects on bone formation at 3 and/or 9 months were dose-dependent, with 100 µg as the potentially

28

optimal dosage.

29

Conclusions: BB-1 significantly enhanced the bone formation induced by a PLGA-fiber reinforced

30

CPC in sheep lumbar osteopenia. A single local dose as low as 100 µg BB-1 was sufficient to augment

31

middle to long-term bone formation. CPC containing the novel GDF5 mutant BB-1 may thus represent

32

an alternative to the bioinert, supraphysiologically stiff polymethylmethacrylate cement presently used

33

to treat osteoporotic vertebral fractures by vertebroplasty/kyphoplasty. Keywords: bone regeneration; GDF-5 mutant; calcium phosphate cement; bone morphogenetic protein; large animal model sheep; osteoporotic vertebral fracture 2 Page 2 of 25

1

1.

Introduction

2

Vertebral compression fractures are a common condition, especially among the elderly population, and

3

account for approx. 45% of all fractures in osteoporosis [1]. Such fractures often cause severe pain and

4

difficulties in daily activities and substantially reduce the quality of life.

5

The available treatment options include minimally-invasive procedures for the improvement of lumbar

6

biomechanics and stability, such as vertebroplasty and balloon kyphoplasty [2, 3], in which a

7

biomaterial is injected into the damaged vertebral body. Recently biodegradable, osteoconductive,

8

calcium phosphate cements (CPC) have been proposed as an alternative to the bioinert

9

polymethylmethacrylate (PMMA) cement, which is currently the most widely used bone void filling

10

material in vertebroplasty and kyphoplasty. In addition, the inclusion of fibers in the CPC can

11

considerably improve their mechanical strength and fracture toughness [4, 5]. For this reason, a poly

12

(l-lactide-co-glycolide) acid (PLGA)-fiber reinforced CPC was used in the present study.

13

CPC can be used as a carrier system to apply osteoinductive molecules such as bone morphogenetic

14

proteins [6]. Until now, only few growth factors including BMP-2 and BMP-7 have been clinically

15

approved for the stimulation of bone regeneration, however with growing safety concerns [7, 8].

16

Recently, GDF5 (also called BMP-14) has shown promising effects in an in vivo sheep model of

17

lumbar osteopenia by enhancing middle to long-term bone formation via changes in bone structure,

18

formation, resorption, and compressive strength [9]. GDF5 clearly and strongly stimulates

19

angiogenesis [10], while its osteoconductive potential may be less pronounced compared to that of

20

BMP-2 [10, 11]. For that reason, a mutant GDF5 protein called BB-1 has been recently produced,

21

which shows augmented bone formation capacity in comparison to the wild-type GDF5 [11]. The

22

mutant was obtained by introducing two point mutations (methionine to valine at positions 453 and

23

456) in the GDF5 alpha-helix structure, the portion of the molecule involved in binding to the BMP-

24

type1 receptor (BMPR-I). This mutation led to an increased affinity of BB-1 to the BMPR-IA subunit

25

and provided the mutant with an osteogenic capacity similar or even superior to BMP-2, while

26

maintaining its angiogenic properties [12, 13].

27

The present study was focused on comparing bone defects in lumbar vertebrae injected with fiber-

28

reinforced CPC containing different dosages of BB-1 (5, 100, and 500 µg) to bone defects injected

29

with fiber-reinforced CPC alone and to lumbar vertebrae without any defect, in order to assess

30

differences among these groups concerning bone mineral density, compressive strength, as well as

31

parameters of bone structure, bone formation, and bone resorption at 3 and 9 months post treatment.

32 33

2.

Materials and Methods

34

Bone defects placed in lumbar vertebrae of aged, osteopenic female sheep and treated with fiber-

35

reinforced CPC alone (sacrifice: 3, 9 months) or with CPC containing 5, 100, and 500 µg BB-1 were

36

compared to controls without a defect. Bone mineral density, bone structure, bone formation, bone

37

resorption, and compressive strength were measured to assess structural and functional effects. In 3 Page 3 of 25

1

accordance with clear recommendations of the United States Food and Drug Administration [14] and

2

the European Medicines Agency [15] to use appropriate animal models for the development of new

3

drugs in the context of osteoporosis, the large animal sheep was used, regarded as highly suitable due

4

to its availability, low cost, high homogeneity, large size, and good comparability of its bone

5

structures with those of humans ([15] and references therein).

6

2.1.

7

growth factor-loaded CPC

8

PLGA fibers (diameter of 25 µm) were prepared and characterized as described previously [5]. In

9

brief, PLGA fibers were extruded from the granulate material PURASORB PLG 1017 (Purac,

10

Gorinchem, Netherlands) using a mini extrusion system (RANDCASTLE EXTRUSION SYSTEMS

11

INC, Cedar Grove, USA). Fibers were then cut to a length of 597 ± 362 µm using a cutting mill

12

PULVERISETTE 19 (FRITSCH GmbH, Idar-Oberstein, Germany) with a 1 mm sieve insert. CPC

13

powder of a commercially available bone cement (Conformité Européenne (CE)-certified, brushite-

14

forming CPC JectOS+; Kasios, L’Union, France) with 10% fiber content (w/w) and a fiber length of 1

15

mm was obtained by mixing defined amounts of fibers and CPC powder.

16

The different CPC samples were prepared during surgery, shortly before injection into the defect. For

17

PLGA fiber-reinforced CPC, the cement powder was thoroughly mixed with the liquid compound of

18

JectOS+ in a powder-to-liquid ratio of 2.2. To obtain the BB-1-loaded, PLGA fiber-reinforced cement,

19

the lyophilized growth factor (BB1: BioPharm GmbH, Heidelberg, Germany) was first dissolved in

20

the liquid and then added to the cement powder.

21

2.2.

22

Aged, osteopenic, female sheep were used [30 Merino animals; 6 – 10 years old (7.40 ± 0.20 years,

23

mean ± standard deviation); 63 – 110 kg body weight (79.80 ± 3.70 kg)], with a significantly lower

24

bone volume/total volume than young sheep aged 2-4 years (see Figure S1 in [16], as well as [17, 18]).

25

Animals were allocated to the different experimental groups in order to achieve an equal mean age.

26

The groups were treated with 5 µg (3 months only; total n = 6), or 100 and 500 µg BB-1 (3 and 9

27

months; total n = 24). Power analyses (G*power; [19]) for dose-dependent differences among the

28

various doses of BB-1 for the structural parameter bone volume/total volume (BV/TV), the bone

29

formation parameters osteoid volume/total volume (OV/BV) and mineralizing surface/bone surface

30

(MS/BS) confirmed that group sizes between 3 and 5 animals were sufficient to detect differences

31

with an alpha error probability of 0.05, a power (1- error probability) of 0.80, and an effect size

32

between 1.75 and 2.97. Permission was obtained from the governmental commission for animal

33

protection, Free State of Thuringia, Germany (registration number 02-036/11).

34

2.3.

35

Anesthesia and surgery were performed as previously published [15]. The surgical technique used in

36

this study has been already described as a minimally-invasive ventrolateral vertebroplasty [15], in

Fabrication of PLGA fibers and preparation of the injectable, PLGA fiber-reinforced,

Animals

Anesthesia and surgical technique

4 Page 4 of 25

1

which lumbar defects (diameter 5 mm; depth approximately 14 mm) were created by a ventrolateral

2

percutaneous approach and subsequently filled following a standardized protocol with a bone

3

filler/pestle system (Medtronic, Minneapolis, USA) under radiolucent control. L1 remained without a

4

defect (untouched) and served as normal control, the defects created in L4 were injected with

5

CPC+fibers, those in L5 with CPC+fibers+BB-1. For selected parameters, the results were also

6

compared to those in L3, which was injected with the validated, CE-certified JectOS+ CPC without

7

fiber reinforcement (pure CPC; compare with Supplementary Fig. 1).

8

Since the effects of pure CPC and CPC+fibers have been partially reported in previous publications

9

[15, 16], however, the current manuscript is focused on the comparison of L5 and L4 with the

10

untouched control L1.

11

After surgery, animals were kept in separate pens for 1-2 weeks and then returned to long-care

12

paddocks for 3 (short-term) or 9 months (long-term). Postoperative medication included antibiotic

13

protection (ampicillin sodium, twice daily 10 mg/kg body weight for 4 days, Ratiopharm GmbH, Ulm,

14

Germany; Enrofloxacin, once daily 2.5 mg/kg for 4 days, Bayer, Leverkusen, Germany), as well as

15

antiphlogistic therapy (metamizole sodium, twice daily 2 mg/kg for 4 days, Wirtschaftsgenossenschaft

16

Deutscher Tierärzte - WDT -, Garbsen, Germany; carprofen, twice daily 2 mg/kg for 2.5 days, Pfizer

17

Animal Health, Berlin, Germany).

18

To determine dynamic histomorphometric indices, sheep were given alternate intramuscular injections

19

of oxytretracycline (20 mg/kg body weight and 10 mg lidocaine) and alizarin red (12 mg/kg body

20

weight; both Sigma, Deisenhofen, Germany) at equal 3 months distance. Short-term animals thus

21

received a total of 2 injections (oxytretracycline 7 days post-operatively and alizarin red 10 days prior

22

to sacrifice at 3 months), long-term animals instead received a total of 4 injections (oxytretracycline 7

23

days and 6 months post-operatively; alizarin red 3 months post-operatively and 10 days before

24

sacrifice at 9 months).

25

2.4.

26

The animals were sacrificed by intravenous injection of overdosed barbiturate (Pentobarbital™, Essex

27

Pharma GmbH, Munich, Germany) and subsequent application of magnesium sulfate (MgSO4). The

28

lumbar spine was then removed using an oscillating bone saw and kept frozen until further use.

29

2.5.

30

Osteodensitometry was conducted using a software-guided digital bone density measuring instrument

31

(DEXA QDR 4500 Elite™; Hologic, Waltham, MA, USA; [20, 21]). In order to avoid artifacts, the

32

lumbar spine was sawn into individual vertebrae L1, L4, and L5. The spinous and transverse

33

processes, as well as the covering and base plates, were removed (final height in each case 15 mm)

34

and the central, spongious area of the vertebral bodies was imaged longitudinally. The hyperdensity

35

region of the injected cement was excluded from quantification after identifying the respective pixel

36

regions by hand, and a rectangular region of interest in a transversal orientation was chosen as a

Sample excision

Measurement of bone mineral density (BMD) via digital osteodensitometry

5 Page 5 of 25

1

measuring field (size 9 × 11 mm; Fig. 1A). Osteodensitometric measurements were therefore limited

2

to the central, spongious area of the vertebral bodies.

3

2.6.

4

The acquisition of 3-D images was achieved using an X-RAY WorX 225 kV tube micro-CT system

5

with a minimal spot size below 5 µm and a flat panel detector (PerkinElmer 1621; CsJ as scintillator;

6

2048 × 2048 pixel). A total of 8 frozen samples were analyzed under dry ice within 6 h. After

7

reconstruction with the original Feldkamp code, the resulting voxel size was 66.6 µm (Fig. 1A).

8

Quantitative analysis of the micro-CT data was carried out using the 3-D software VGSTUDIO MAX

9

2.2 (Volume Graphics GmbH, Heidelberg, Germany) and applying half-cylinders with a radius of 3.0,

10

4.0, and 5.0 mm. The longitudinal axis and length of the half cylinders for each individual vertebral

11

body were defined on the basis of the respective parameters of the drill channel. The exact 3-D

12

position of the half cylinders (e.g., rotation and depth) was chosen to exclude cortical bone. Bone

13

volume (BV) and total volume (TV) were separately determined in the drill channel (maximal radius

14

of 2.5 mm) and the adjacent half cylinder segments by global threshold determination (principle of

15

onion shell). For this purpose, the respective maxima of the grey values for soft tissue and bone tissue

16

were determined and the mean values between the maxima were used as thresholds for the volume

17

determination of the individual components. This allowed the high density injected cement to be

18

excluded from the analysis.

19

2.7.

20

The lumbar vertebral bodies were cut in 2 parts directly along the axis of the cement injection channel

21

and the analyses were carried out using 2 different types of histological sections: i) decalcified paraffin

22

sections stained by hematoxylin-eosin [22, 23]; or ii) plastic-embedded sections obtained by fixation

23

in acetone and dehydration in ascending alcohol series without demineralization. For this purpose, the

24

samples were embedded in Technovit 9100 according to the instructions of the supplier (Heraeus

25

Kulzer, Wehrheim, Germany; [24]). Sections were then cut to a thickness of approx. 7 μm. For static

26

histomorphometry, the sections were stained with trichrome stain according to Masson-Goldner [22],

27

for dynamic histomorphometry the sections were left unstained (Fig. 2A).

28

Static histomorphometry of each individual vertebral body was performed in the immediate vicinity of

29

the injection channel using a standard microscope (Axiovert 200 M, Carl Zeiss Microimaging GmbH,

30

Oberkochen, Germany) with a 200-fold magnification. The respective parameters were determined

31

according to published procedures [25, 26], excluding the injected CPC from the analysis. For

32

dynamic histomorphometry, bone induction was assessed in 10 fields of each individual vertebral

33

body by determining the histomorphometric parameters using a standard microscope (Axiovert 200

34

M™, Carl Zeiss Microimaging GmbH), 100-fold magnification, and the corresponding software

35

(Axioversion 3.1, Carl Zeiss Microimaging GmbH). Total bone surface, labeled surface (single and

36

double fluorescence lines), and inter-label distance were quantified using the ImageJ program

Analysis of bone structure via high-resolution microcomputed tomography (micro-CT)

Histological and static/dynamic histomorphometrical measurements

6 Page 6 of 25

1

(imagej.nih.gov). These parameters were then used to calculate the mineralizing surface/bone surface

2

(MS/BS), mineral apposition rate (MAR), and bone formation rate surface-based (BFR/BS; [26]).

3

2.8.

4

Biomechanical compressive strength measurements were conducted using a universal material testing

5

machine (FPG 7/20-010™, Kögel, Leipzig, Germany) and the corresponding software (FRK Quicktest

6

2004.01™). For biomechanical testing, frozen cancellous bone cylinders (10 mm diameter x 15 mm

7

height) were obtained from the central part of the vertebral bodies (L1, L4, L5), using a surgical

8

diamond hollow milling cutter (diameter 10 mm). After a defined thawing time of exactly 30 minutes,

9

samples were semi-confined in 2 semilunar clamps (minimal inner diameter of 10.1 mm; length 9.8

10

mm) and then compressed along their longitudinal axis until fracturing. This axis was chosen because

11

it represents the main loading axis in humans and is therefore of major interest for future clinical

12

application.

13

2.9.

14

All results were expressed as means ± SEM. Since the populations did not consistently show a normal

15

distribution (as determined using the Shapiro Wilk test), the non-parametric Wilcoxon signed rank test

16

was used for the comparison of dependent groups, the Kruskall-Wallis multi-group and Mann Whitney

17

U tests for the comparison of independent groups. In order to address the problem of multiple testing

18

and to reduce the number of individual statistical comparisons, only the parameters showing

19

statistically significant differences among different doses of BB-1 in the Kruskall-Wallis test were

20

further analyzed for significant differences between individual groups. Significance was accepted at p

21

< 0.05.

22

In addition, as all outcome parameters were expected to undergo the same direction of changes, the

23

possibility of a type 1 error was expected to decrease.

Biomechanical testing (compressive strength)

Statistical methods

24 25

3.

Results

26

3.1.

Bone densitometry

27

Numerical or significant increases of the bone mineral density (BMD) in L4 (CPC+fibers) and L5

28

(CPC+fibers+BB-1) compared to L1 (untouched control) were observed in all groups. The increase

29

was significant between L4 (- BB-1) and L1 (untouched control) in the 500 µg group at 3 months, and

30

between L5 (+ BB-1) and L1 (untouched control) in the 500 µg groups at 3 and 9 months. In the

31

former group, the increase was also significant between L5 and L4 (Fig. 1B).

32

In general, the BMD values for both L4 and L5 decreased from 3 months to 9 months, with a

33

significant difference for L5 at 500 µg BB-1.

34

There were no dose-dependent BB-1 effects for this parameter at 3 or 9 months (Fig. 1B).

35

3.2.

Micro computed tomography (bone volume/total volume; BV/TV)

7 Page 7 of 25

1

The BV/TV in L4 (CPC+fibers) and L5 (CPC+fibers+BB-1) compared to L1 (untouched control) was

2

numerically or significantly increased in almost all groups and at all distances from the edge of the

3

drill channel (0.5, 1.5, or 2.5 mm; with decreased effects at larger distances). However, a significant

4

difference between L5 (+ BB-1) and L4 (– BB-1) was only observed at 9 months for 500 µg BB-1 at

5

the distance of 1.5 mm (Fig. 1C).

6

The values for both L4 and L5 decreased from 3 months to 9 months, but no significant differences

7

were observed between the 2 time points (Fig. 1C).

8

As for the BMD, there were no dose-dependent BB-1 effects at 3 or 9 months.

9

3.3.

Histological analysis

10

3.3.1. Bone volume/total volume (BV/TV)

11

In representative 9 month histology images of untouched controls (L1) and treated vertebral bodies

12

(L4: CPC+fibers; L5: CPC+fibers+BB-1), L4 and L5 showed a dense bony layer around the injection

13

channel (Fig. 2A). Also, there was bone formation inside the cement region of the vertebral bodies. No

14

signs of inflammatory infiltration and osteolysis close to or distant from the injection channel were

15

observed in either L4 or L5 at any time point.

16

There was a significant increase of the BV/TV in L4 (CPC+fibers) and L5 (CPC+fibers+BB-1)

17

compared to L1 (untouched control) in all groups (Fig. 2B). This significant increase was also

18

observed when comparing L3 (pure CPC) to L1 (Supplementary Fig. 1A).

19

In addition, all BV/TV values for L5 (+ BB-1) were numerically higher than those for L4 (– BB-1)

20

and for L3 (pure CPC) for all doses of BB-1 (Fig. 2B; Supplementary Fig. 1A). Significant differences

21

between L5 and L4 were observed for 5 µg and 100 µg BB-1 at 3 months, as well as for 500 µg BB-1

22

at 9 months (p < 0.05; Fig. 2B). Significant differences between L5 and L3 were observed for 500 µg

23

BB-1 at 3 months, as well as for 100 and 500 µg BB-1 at 9 months (Supplementary Fig. 1A).

24

The BV/TV values for L3, L4, and L5 did not significantly change from 3 months to 9 months (Fig.

25

2B; Supplementary Fig. 1A).

26

At 3 months, maximal BB-1 effects were observed at 5 µg and 100 µg, while at 9 months the highest

27

dose (500 µg) resulted in maximal effects (Fig. 2B).

28

3.3.2. Trabecular thickness (Tb.Th)

29

Regarding the comparison between L3, L4, and L5 versus L1, the results were identical to those for

30

the BV/TV (compare Fig. 2C with Fig. 2B and Supplementary Fig. 1A with Supplementary Fig. 1B).

31

As for the BV/TV, all 3 month and 9 month Tb.Th values for L5 (+ BB-1) were higher than those for

32

L4 (– BB-1) and L3 (pure CPC) for all doses of BB-1 (p < 0.05 versus L4 for 5 µg and 100 µg BB-1 at

33

3 months, as well as for 500 µg at 9 months; p < 0.05 versus L3 for all groups; Fig. 2C;

34

Supplementary Fig. 1B). The highest fold-increases in the comparison between L5 and L4 were

35

observed for 5 µg BB-1 at 3 months (1.21-fold) and for 100 µg at 3 months (1.23-fold).

8 Page 8 of 25

1

There was no significant decrease of the Tb.Th values from 3 months to 9 months in any group (Fig.

2

2C; Supplementary Fig. 1B).

3

Similar to the BV/TV, maximal BB-1 effects were observed at 3 months for 5 µg and 100 µg.

4

3.3.3. Trabecular number (Tb.N)

5

In contrast to the results for BV/TV and Tb.Th, the Tb.N significantly decreased in L3 (pure CPC), L4

6

(CPC+fibers), and L5 (CPC+fibers+BB-1) compared to L1 (untouched control) in all groups (Fig. 2D;

7

Supplementary Fig. 1C).

8

In addition, L5 (+ BB-1) showed lower values in comparison to L4 (– BB-1) and L3 (pure CPC) in all

9

groups (p < 0.05 versus L4 for 500 µg BB-1 at 9 months; p < 0.05 versus L3 for all groups except for

10

the 100 µg group at 9 months; Fig. 2D; Supplementary Fig. 1C).

11

There were no significant changes of the Tb.N from 3 months to 9 months in any group (Fig. 2D;

12

Supplementary Fig. 1C).

13

As already noted for BV/TV and Tb.Th, maximal BB-1 effects for the Tb.N were observed at 3

14

months for 5 µg and 100 µg (Fig. 2D).

15

3.3.4. Osteoid volume/total volume (OV/BV)

16

Significant increases of the OV/BV in L4 (CPC+fibers) and L5 (CPC+fibers+BB-1) compared to L1

17

(untouched control) were observed in the 100 µg and 500 µg BB-1 groups at 3 months and the 100 µg

18

BB-1 group at 9 months. For this parameter, there were no significant differences between L5 (+ BB-

19

1) and L4 (– BB-1).

20

A significant decrease from 3 to 9 months was observed in the 500 µg BB-1 group (Fig. 3A).

21

Maximal BB-1 effects at 3 and 9 months were found for 100 µg [with a significant increase at 3

22

months from 5 µg to 100 µg BB-1 (Fig. 3A)].

23

3.3.5. Osteoid surface/bone surface (OS/BS)

24

Significant increases of the OS/BS in comparison to L1 (untouched control) were observed for L5

25

(CPC+fibers+BB-1) at 3 months in the 100 and 500 µg BB-1 group (Fig. 3B) and for L3 (pure CPC)

26

at 9 months in the 500 µg BB-1 group (Supplementary Fig. 1D).

27

At 5 and 500 µg BB-1 after 3 months, the OS/BS in L5 (+ BB-1) was also significantly higher than in

28

L4 (– BB-1; Fig. 3B); in addition, the OS/BS in L5 (+ BB-1) was significantly higher than in L3 (pure

29

CPC) at 100 µg and 500 µg BB-1 after 3 months and at 100 µg BB-1 after 9 months (Supplementary

30

Fig. 1D).

31

For the 500 µg BB-1 group, there was a significant decrease of the OS/BS values from 3 months to 9

32

months (Fig. 3B).

33

As for the OV/BV, maximal BB-1 effects were observed for 100 µg at 3 and 9 months [with a

34

significant increase from 5 µg to 100 µg and 500 µg BB-1 at 3 months (Fig. 3B)].

35

3.3.6. Osteoid thickness (O.Th)

9 Page 9 of 25

1

Similar to the OS/BS, the O.Th displayed significant increases in comparison to L1 (untouched

2

control) for L4 (CPC+fibers) in the 100 and 500 µg BB-1 group at 3 months and the 100 µg BB-1

3

group at 9 months, as well as the for L5 (CPC+fibers+BB-1) in the 500 µg BB-1 group at 3 months

4

and the 100 µg group at 9 months (Fig. 3C). For the O.Th, there were no significant differences

5

between L5 (+ BB-1) and L4 (– BB-1).

6

Significant differences between the 3 and 9 months values were only observed for L5

7

(CPC+fibers+BB-1) in the 500 µg BB-1 group (Fig. 3C).

8

Maximal BB-1 effects for this parameter were detected for 5 µg at 3 months (Fig. 3C)].

9

3.3.7. Eroded surface/bone surface (ES/BS)

10

Interestingly, the ES/BS was significantly decreased compared to L1 (untouched control) in both L4

11

(CPC+fibers) and L5 (CPC+fibers+BB-1) in the 5 µg BB-1 group at 3 months and the 100 µg BB-1

12

group at 9 months (Fig. 3D).

13

For this parameter, there were no significant differences between L5 (+ BB-1) and L4 (– BB-1),

14

between the 3 and 9 months values, or among the different BB-1 doses (Fig. 3D).

15

3.3.8. Mineralizing surface/bone surface (MS/BS)

16

In parallel to a significant increase of the MS/BS in L4 (CPC+fibers) and L5 (CPC+fibers+BB-1)

17

compared to L1 (untouched control) in all groups, the MS/BS in L5 (+ BB-1) for 500 µg BB-1 at 3

18

months was also significantly higher than in L4 (– BB-1; Fig. 4A).

19

The respective MS/BS values were stable over time or even increased from 3 to 9 months (Fig. 4A).

20

Maximal BB-1 effects were observed at 3 months for 5 µg and 500 µg (Fig. 3A).

21

3.3.9. Mineral apposition rate (MAR) and bone formation rate surface based (BFR/BS)

22

In the 3 month groups, no double lines were observed and the MAR and BFR/BS could therefore not

23

be calculated. At 9 months, there was a significant increase of the MAR and BFR/BS compared to L1

24

(untouched control) for L4 (CPC+fibers) in the 100 and 500 µg BB-1 groups (Figs. 4B and 4C).

25

For these parameters, there were no significant differences between L5 (+ BB-1) and L4 (– BB-1),

26

between the 3 and 9 months values, or among the different doses of BB-1 (maximum effects for both

27

parameters at 100 µg; Figs. 4B and 4C).

28

3.4.

29

Interestingly, for the 100 µg BB-1 group at 9 months both L4 (– BB-1) and L5 (+ BB-1) showed a

30

numerically higher compressive strength than L1 (untouched control; 1.37 and 1.38-fold increase,

31

respectively, compared to L1), however without significant differences between L5 (+ BB-1) and L4

32

(– BB-1). There was a significant increase in the 100 µg BB-1 group from 3 to 9 months, and this

33

group at 9 months also showed a significantly higher compressive strength than the 500 µg BB-1

34

group (Fig. 5).

Compressive strength

35 36

4.

Discussion 10 Page 10 of 25

1

On the basis of the extended clinical use of BMPs for the promotion of bone repair [27], the present

2

study aimed at investigating the influence of loading BB-1, a mutant GDF5 protein with augmented

3

bone formation capacity in comparison to wild-type GDF5 [11], into a PLGA fiber-reinforced CPC on

4

bone regeneration in a sheep model of lumbar osteopenia.

5

The present results indicated the following aspects: i) both BB-1-free and BB-1-loaded, PLGA fiber-

6

reinforced CPC significantly enhanced bone formation via changes in bone structure, formation, and

7

resorption (e.g., BV/TV and Tb.Th; OS/BS and MS/BS; as well as ES/BS); ii) in comparison to BB-1-

8

free CPC, BB-1-loaded CPC even significantly augmented BMD, bone structure (BV/TV, Tb.Th, and

9

Tb.N) and bone formation (OS/BS and MS/BS); iii) BB-1 effects showed a clear dose dependence,

10

with maximal effects between 5 µg and 100 µg; iv) all the BB-1-free or BB-1-loaded CPC

11

components are highly biocompatible since there were no signs of inflammatory infiltration after the

12

injection [27]. The present study thus shows that the application of BB-1 for the treatment of lumbar

13

vertebral defects enhances bone formation and bone structure. With the limitation of different load

14

sharing forces in sheep and human [28], it also confirms the suitability of the present physiological

15

and convenient osteopenia model for the analysis of vertebral augmentation with resorbable and

16

osteoconductive CPC [15], as well as for dose-response studies with different BMPs ([9, 29]; present

17

study). Despite the considerably more time- and resource-consuming induction procedure, further

18

support for the efficacy of doses as low as 100 µg BB-1 to augment middle to long-term bone

19

formation would be derived from the confirmation of the effects in models with induced osteoporosis

20

and/or osteoporotic compression fractures [30, 31].

21

4.1.

22

The enhancement of bone formation by pure CPC or PLGA fiber-reinforced CPC is in line with the

23

known osteoconductive effects of CPC with different compositions [32, 33], without any suppressive

24

effect on bone formation [9, 16, 29, 33].

25

4.2.

26

Some of the parameters measured in the present study indicated exclusively early significant effects of

27

BB-1 (OS/BS, MS/BS) or exclusively late effects of BB-1 [BV/TV (micro-CT), Tb.N], whereas

28

several parameters supported an inductive effect of BB-1 at both the early and the late time point

29

[BV/TV (histology), Tb.Th]. It is interesting to note that bone formation parameters such as OS/BS

30

and MS/BS showed a reaction already at the early time point of 3 months. This finding is in agreement

31

with the knowledge that bone healing initiates very rapidly after injury and leads to load-bearing

32

fracture healing already after approx. 6 weeks ([34, 35]; own unpublished work). On the other hand,

33

several bone structure parameters still showed significant effects of BB-1 at 9 months, suggesting a

34

middle to long-term effect of BB-1 after a single therapeutic application (in analogy to GDF5 and

35

BMP-2 [9, 29, 36].

Effects of the PLGA fiber-reinforced CPC

Early versus late effects of BB-1

11 Page 11 of 25

1

4.3.

Dose-dependence of BB-1 effects

2

Some parameters indicated a dose-dependence of the BB-1 effects, with maximal effects of 5 µg for

3

the bone structure parameters BV/TV (histology) and Tb.Th at 3 months, but maximal effects of 100

4

µg for some bone formation parameters, e.g., OV/BV, OS/BS (3 and 9 months), and O.Th (9 months).

5

Since the application of 500 µg did not further enhance the bone regeneration in a significant fashion,

6

100 µg BB-1 (i.e., 20 to 400-fold less than the dosage of BMP-2 previously used in clinical studies;

7

[7]) may be the maximal local dose needed for the augmentation of bone formation by BB-1 in the

8

present system.

9

The results obtained with this particular BB-1 dose may also be clinically relevant. In fact due to the

10

highly catabolic situation in the spongiosa and corticalis of the osteoporotic bone, it is crucial to

11

improve the mechanical properties of the bone and its stability to avoid refracturing of the stabilized

12

fracture and to revert or at least diminish the local catabolic process. To achieve this purpose, even

13

small improvements may be critical for the long-term improvement of the quality of life of the

14

individual and may thus justify the additional cost [8].

15

This view is supported by the significant changes (remarkable increases of up to 1.63-fold) especially

16

in the bone formation parameters, e.g. OV/BV, OS/BS, and MS/BS, as well as the high effect sizes for

17

dose-dependent differences among the various doses of BB-1 (between 1.75 and 2.97 despite the

18

limited group size of n = 6; i.e., well above the effect sizes between 0.5 and 1.0 usually assumed in

19

clinical studies). So far only two in vivo studies have been performed in small animals using the newly

20

created GDF5 mutant BB-1, identifying a concentration comparable to the one in the present study,

21

e.g. 10 µg for the implantation of a β-tricalcium phosphate scaffolds loaded with BB-1 in SCID mice

22

[11] or 20 µg for the treatment of rabbit radius defects with a BB-1 coated collagen carrier [13].

23

Interestingly, in our previous study on the same sheep model of osteopenia, a similar dose range (≤

24

100 µg) was sufficient to augment bone formation upon treatment with CPC+fibers loaded with wild-

25

type GDF5, indicating that in the present model wild-type and mutant GDF5 may be equally potent

26

[9].

27

It remains to be determined whether the present formulation of the BB-1-loaded CPC, in which a

28

homogeneous distribution of the BB-1 with sustained release can be assumed, is suitable for therapy or

29

whether modifications of the delivery system are required to optimize the release kinetics of the

30

biologically active molecule [8, 37, 38].

31

4.4.

32

The present results also showed that BB-1 influenced both bone formation (OV/BV, OS/BS, O.Th)

33

and bone resorption parameters such as ES/BS. This is in agreement with the accepted role of BMPs in

34

inducing the differentiation and/or activation of both osteoblasts and osteoclasts [39] and suggests a

35

combined direct or indirect influence of BB-1 on anabolic and catabolic aspects of bone regeneration,

36

as previously shown for other growth factors of the same family such as BMP-7 [40], GDF-5 [9], and

37

BMP-2 [29]. This combined effect may further enhance the positive action of BB-1 on bone

Cellular effects of BB-1

12 Page 12 of 25

1

regeneration. On the other hand, the fact that the structural parameters BV/TV and Tb.Th are still

2

strongly elevated above the level of the untouched bone at 9 months indicates that bone remodeling

3

and the return to physiological levels is not yet completed.

4

For some parameters, doses, and time points (e.g., BV/TV assessed by micro-CT, OV/BV, OS/BS,

5

O.Th, MS/BS, MAR, BFR/BS, and compressive strength), the addition of BB-1 to the CPC led to

6

numerically lower values, suggesting the possibility of negative effects of BB-1 on bone formation

7

and/or bone structure. This indicates that the local dose of BB-1 will have to be carefully established

8

in order to exclude inhibiting effects of this BMP on the structure or quality of bone [9, 29].

9

4.5.

Adverse effects of BB-1

10

Since this was not designed as a safety study, a systematic assessment of the safety profile of the

11

PLGA fiber-reinforced, BB-1-loaded CPC was not possible. Up to now no clinical data on BB-1 are

12

available; however, there were no signs of the main adverse effects (e.g., ectopic bone formation or

13

osteolysis, disability or poor outcome, leg or back pain, swelling or wound complications, subsidence

14

or displacement of the implant, neurologic events, urogenital events, and radiculitis) previously

15

reported after the spinal application of a high-dose of the clinically used BMP-2 (as high as 1.95 to 40

16

mg; [7, 8]). In addition, there were no signs of local inflammatory infiltration after the use of the

17

PLGA fiber-reinforced, BB-1-loaded CPC at any time point, indicating that the cement components

18

have no or negligible pro-inflammatory effects. Furthermore, as previously published, there were no

19

signs of systemic intravascular leakage after CPC injection and both pure CPC and PLGA fiber-

20

reinforced CPC showed a significantly lower cement extrusion from the initial bony void than the

21

commonly applied PMMA cement ex vivo and in vivo [41]. Finally, the suggested local dose of 100

22

µg BB-1 is between 20 and 400 times lower than the previously applied clinical doses of BMP-2.

23

Although systematic safety studies are clearly needed, the safety profile of low-dose BB-1 in the

24

present sheep model appears to be favorable.

25 26

5.

Conclusion

27

Our study demonstrated that the new mutant GDF-5V453/V456 protein (BB-1) significantly enhanced the

28

bone formation induced by a PLGA-fiber reinforced CPC in sheep lumbar osteopenia for at least 3

29

months, showing long-term beneficial effects on the bone structure after a single therapeutic

30

application. The results indicated that local doses of ≤ 100 µg BB-1 may be sufficient for the induction

31

of bone tissue formation. Since there were no signs of local or systemic inflammation, all components

32

of BB-1-free and BB-1-loaded CPC appeared to be highly biocompatible.

33

The novel BB-1-loaded CPC may thus represents an opportunity to replace the bioinert, non-

34

resorbable,

35

vertebroplasty/kyphoplasty of osteoporotic vertebral fractures, and its use may also be extended to

36

broader clinical applications.

supraphysiologically

stiff

PMMA

cements

currently

applied

in

the

13 Page 13 of 25

1

1.

2

We gratefully acknowledge the financial support by the Carl Zeiss Foundation (doctoral candidate

3

scholarship to S.M.) and the German Federal Ministry of Education and Research (BMBF FKZ

4

0316205C to J.B and K.D.J.; BMBF FKZ 035577D, 0316205B, and 13N12601 to R.W.K). We

5

gratefully acknowledge the partial financial support of the Deutsche Forschungsgemeinschaft (DFG),

6

grant reference INST 275/241-1 FUGG to K.D.J., and the Thüringer Ministerium für Bildung,

7

Wissenschaft und Kultur (TMBWK), grant reference 62-4264 925/1/10/1/01 to K.D.J.

8

Nicolas Guena and Alain Lerch, Kasios, are gratefully acknowledged for providing Jectos cement. We

9

gratefully acknowledge E. Mark for performing the osteodensitometry analyses and S. Födisch, G.

10

Acknowledgement

Grunert, U. Körner, C. Müller, and B. Ukena for their excellent technical assistance.

11 12

References

13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

[1] Hernlund E, Svedbom A, Ivergard M, Compston J, Cooper C, Stenmark J, et al. Osteoporosis in the European Union: medical management, epidemiology and economic burden. A report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Federation of Pharmaceutical Industry Associations (EFPIA). Arch Osteoporos. 2013;8:136. [2] Teyssedou S, Saget M, Pries P. Kyphopasty and vertebroplasty. Orthopaedics & traumatology, surgery & research : OTSR. 2014;100:S169-79. [3] Ates A, Gemalmaz HC, Deveci MA, Simsek SA, Cetin E, Senkoylu A. Comparison of effectiveness of kyphoplasty and vertebroplasty in patients with osteoporotic vertebra fractures. Acta orthopaedica et traumatologica turcica. 2016;50:619-22. [4] Gorst NJ, Perrie Y, Gbureck U, Hutton AL, Hofmann MP, Grover LM, et al. Effects of fibre reinforcement on the mechanical properties of brushite cement. Acta Biomater. 2006;2:95-102. [5] Maenz S, Kunisch E, Muhlstadt M, Bohm A, Kopsch V, Bossert J, et al. Enhanced mechanical properties of a novel, injectable, fiber-reinforced brushite cement. J Mech Behav Biomed Mater. 2014;39:328-38. [6] Verron E, Khairoun I, Guicheux J, Bouler JM. Calcium phosphate biomaterials as bone drug delivery systems: a review. Drug Discov Today. 2010;15:547-52. [7] Carragee EJ, Hurwitz EL, Weiner BK. A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned. The spine journal : official journal of the North American Spine Society. 2011;11:471-91. [8] James AW, LaChaud G, Shen J, Asatrian G, Nguyen V, Zhang X, et al. A Review of the Clinical Side Effects of Bone Morphogenetic Protein-2. Tissue engineering Part B, Reviews. 2016. [9] Bungartz M, Kunisch E, Maenz S, Horbert V, Xin L, Gunnella F, et al. GDF5 significantly augments the bone formation induced by an injectable, PLGA-fiber reinforced, brushite-forming cement in a sheep defect model of lumbar osteopenia. The spine journal : official journal of the North American Spine Society. 2017. [10] Kadomatsu H, Matsuyama T, Yoshimoto T, Negishi Y, Sekiya H, Yamamoto M, et al. Injectable growth/differentiation factor-5-recombinant human collagen composite induces endochondral ossification via Sry-related HMG box 9 (Sox9)expression and angiogenesis in murine calvariae. Journal of periodontal research. 2008;43:483-9. [11] Kasten P, Beyen I, Bormann D, Luginbuhl R, Ploger F, Richter W. The effect of two point mutations in GDF-5 on ectopic bone formation in a beta-tricalciumphosphate scaffold. Biomaterials. 2010;31:3878-84.

14 Page 14 of 25

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

[12] Kleinschmidt K, Wagner-Ecker M, Bartek B, Holschbach J, Richter W. Superior angiogenic potential of GDF-5 and GDF-5(V453/V456) compared with BMP-2 in a rabbit long-bone defect model. The Journal of bone and joint surgery American volume. 2014;96:1699-707. [13] Kleinschmidt K, Ploeger F, Nickel J, Glockenmeier J, Kunz P, Richter W. Enhanced reconstruction of long bone architecture by a growth factor mutant combining positive features of GDF-5 and BMP-2. Biomaterials. 2013;34:5926-36. [14] Thompson DD, Simmons HA, Pirie CM, Ke HZ. FDA Guidelines and animal models for osteoporosis. Bone. 1995;17:125S-33S. [15] Bungartz M, Maenz S, Kunisch E, Horbert V, Xin L, Gunnella F, et al. First-time systematic postoperative clinical assessment of a minimally invasive approach for lumbar ventrolateral vertebroplasty in the large animal model sheep. The spine journal : official journal of the North American Spine Society. 2016;16:1263-75. [16] Maenz S, Brinkmann O, Kunisch E, Horbert V, Gunnella F, Bischoff S, et al. Enhanced bone formation in sheep vertebral bodies after minimally invasive treatment with a novel, PLGA fiberreinforced brushite cement. The spine journal : official journal of the North American Spine Society. 2016. [17] Borsari V, Fini M, Giavaresi G, Rimondini L, Consolo U, Chiusoli L, et al. Osteointegration of titanium and hydroxyapatite rough surfaces in healthy and compromised cortical and trabecular bone: in vivo comparative study on young, aged, and estrogen-deficient sheep. J Orthop Res. 2007;25:125060. [18] Sachse A, Wagner A, Keller M, Wagner O, Wetzel WD, Layher F, et al. Osteointegration of hydroxyapatite-titanium implants coated with nonglycosylated recombinant human bone morphogenetic protein-2 (BMP-2) in aged sheep. Bone. 2005;37:699-710. [19] Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behavior research methods. 2009;41:1149-60. [20] Turner AS. The sheep as a model for osteoporosis in humans. Vet J. 2002;163:232-9. [21] Syed Z, Khan A. Bone densitometry: applications and limitations. J Obstet Gynaecol Can. 2002;24:476-84. [22] Plenck Hj. Untersuchung des Binde- und Stützgewebes (Trichrom-Färbung nach Goldner). In: Böck P, editor. Romeis – Mikroskopische Technik. Munich, Germany: Urban und Schwarzenberg Verlag; 1989. p. 499. [23] Knutsen G, Engebretsen L, Ludvigsen TC, Drogset JO, Grontvedt T, Solheim E, et al. Autologous chondrocyte implantation compared with microfracture in the knee. A randomized trial. The Journal of bone and joint surgery American volume. 2004;86-A:455-64. [24] Donath K, Breuner G. A method for the study of undecalcified bones and teeth with attached soft tissues. The Sage-Schliff (sawing and grinding) technique. J Oral Pathol. 1982;11:318-26. [25] Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res. 1987;2:595-610. [26] Dempster DW, Compston JE, Drezner MK, Glorieux FH, Kanis JA, Malluche H, et al. Standardized nomenclature, symbols, and units for bone histomorphometry: a 2012 update of the report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res. 2013;28:2-17. [27] Scarfi S. Use of bone morphogenetic proteins in mesenchymal stem cell stimulation of cartilage and bone repair. World journal of stem cells. 2016;8:1-12. [28] Mageed M, Berner D, Julke H, Hohaus C, Brehm W, Gerlach K. Is sheep lumbar spine a suitable alternative model for human spinal researches? Morphometrical comparison study. Laboratory animal research. 2013;29:183-9. [29] Gunnella F, Kunisch E, Bungartz M, Maenz S, Horbert V, Xin L, et al. Low-dose BMP-2 is sufficient to enhance the bone formation induced by an injectable, PLGA-fiber reinforced, brushiteforming cement in a sheep defect model of lumbar osteopenia. The spine journal : official journal of the North American Spine Society. 2017. [30] Eschler A, Roepenack P, Roesner J, Herlyn PK, Martin H, Reichel M, et al. Cementless Titanium Mesh Fixation of Osteoporotic Burst Fractures of the Lumbar Spine Leads to Bony Healing: Results of an Experimental Sheep Model. BioMed research international. 2016;2016:4094161. [31] Eschler A, Ropenack P, Herlyn PK, Roesner J, Pille K, Busing K, et al. The standardized creation of a lumbar spine vertebral compression fracture in a sheep osteoporosis model induced by 15 Page 15 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

ovariectomy, corticosteroid therapy and calcium/phosphorus/vitamin D-deficient diet. Injury. 2015;46 Suppl 4:S17-23. [32] Theiss F, Apelt D, Brand B, Kutter A, Zlinszky K, Bohner M, et al. Biocompatibility and resorption of a brushite calcium phosphate cement. Biomaterials. 2005;26:4383-94. [33] Dagang G, Haoliang S, Kewei X, Yong H. Long-term variations in mechanical properties and in vivo degradability of CPC/PLGA composite. J Biomed Mater Res B Appl Biomater. 2007;82:533-44. [34] Einhorn TA, Gerstenfeld LC. Fracture healing: mechanisms and interventions. Nature reviews Rheumatology. 2015;11:45-54. [35] Rüedi TB BR, Moran CG. Principles of fracture management. 2nd ed Stuttgart. 2007. [36] Gruber RM, Ludwig A, Merten HA, Pippig S, Kramer FJ, Schliephake H. Sinus floor augmentation with recombinant human growth and differentiation factor-5 (rhGDF-5): a pilot study in the Goettingen miniature pig comparing autogenous bone and rhGDF-5. Clin Oral Implants Res. 2009;20:175-82. [37] King WJ, Krebsbach PH. Growth factor delivery: how surface interactions modulate release in vitro and in vivo. Advanced drug delivery reviews. 2012;64:1239-56. [38] Luginbuehl V, Meinel L, Merkle HP, Gander B. Localized delivery of growth factors for bone repair. European journal of pharmaceutics and biopharmaceutics : official journal of Arbeitsgemeinschaft fur Pharmazeutische Verfahrenstechnik eV. 2004;58:197-208. [39] Giannoudis PV, Kanakaris NK, Einhorn TA. Interaction of bone morphogenetic proteins with cells of the osteoclast lineage: review of the existing evidence. Osteoporosis international : a journal established as result of cooperation between the European Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA. 2007;18:1565-81. [40] Maurer T, Zimmermann G, Maurer S, Stegmaier S, Wagner C, Hansch GM. Inhibition of osteoclast generation: a novel function of the bone morphogenetic protein 7/osteogenic protein 1. Mediators of inflammation. 2012;2012:171209. [41] Xin L, Bungartz M, Maenz S, Horbert V, Hennig M, Illerhaus B, et al. Decreased extrusion of calcium phosphate cement versus high viscosity PMMA cement into spongious bone marrow-an ex vivo and in vivo study in sheep vertebrae. The spine journal : official journal of the North American Spine Society. 2016;16:1468-77.

30

16 Page 16 of 25

1

Figure 1: (A) Definition of region of interests (ROI) for osteodensitometry (BMD); (B) bone mineral

2

density (BMD) 3 months and 9 months after surgery (n = 6 for each group), as determined by

3

osteodensitometry in differently treated lumbar vertebral bodies; * p ≤ 0.05 Wilcoxon test versus L1; +

4

p ≤ 0.05 Wilcoxon test versus L4; ° p ≤ 0.05 Mann Whitney U test versus 3 months; data are

5

expressed as means ± SEM; the fold-change in L5 (+ BB-1) versus L4 (– BB-1) is indicated; n.d.: not

6

determined.

7

Figure 2: (A) Control of CPC placement by micro-CT (determination of BV/TV using onion shell

8

regions as depicted in the four lower panels); (B) bone volume/total volume (BV/TV) after 3 months

9

and 9 months, as determined by micro-CT in differently treated lumbar vertebral bodies; * p ≤ 0.05

10

Wilcoxon test versus L1; + p ≤ 0.05 Wilcoxon test versus L4; data are expressed as means ± SEM; the

11

fold-change in L5 (+ BB-1) versus L4 (– BB-1) is indicated; n.d.: not determined.

12

Figure 3: (A) Representative histology images of differently treated lumbar vertebral bodies 9 months

13

after surgery stained by hematoxylin-eosin (paraffin sections, second panel: 12.5x magnification, with

14

the scale bar indicating the depth of cement penetration into the bone marrow; third panel 100x

15

magnification) and Masson Goldner staining (plastic-embedded tissue sections, right panel: 200x

16

magnification; white arrow: osteoid; mb: mineralized bone); untreated control (L1), CPC+fibers (L4),

17

and CPC+fibers+BB-1 (L5). (B – D) Structural bone parameters 3 months and 9 months after surgery

18

(n = 6 for each group), as determined by static histomorphometry in differently treated lumbar

19

vertebral bodies: (B) bone volume/total volume (BV/TV); (C) trabecular thickness (Tb.Th); (D)

20

trabecular number (Tb.N). * p ≤ 0.05 Wilcoxon test versus L1; + p ≤ 0.05 Wilcoxon test versus L4;

21

data are expressed as means ± SEM; the fold-change in L5 (+ BB-1) versus L4 (– BB-1) is indicated;

22

n.d.: not determined.

23

Figure 4: (A - D) Bone formation and erosion parameters 3 months and 9 months after surgery (n = 6

24

for each group), as determined by static histomorphometry in differently treated lumbar vertebral

25

bodies: (A) osteoid volume (OV/BV); (B) osteoid surface (OS/BS); (C) osteoid thickness (O.Th); (D)

26

eroded surface (ES/BS). * p ≤ 0.05 Wilcoxon test versus L1; + p ≤ 0.05 Wilcoxon test versus L4; ° p ≤

27

0.05 Mann Whitney U test versus 3 months; # p ≤ 0.05 Mann Whitney U test versus indicated BB-1

28

concentration; data are expressed as means ± SEM; the fold-change in L5 (+ BB-1) versus L4 (– BB-

29

1) is indicated. n.d.: not determined.

30

Figure 5: (A - C) Bone formation parameters 3 months and 9 months after surgery (n = 6 for each

31

group), as determined by dynamic histomorphometry in differently treated lumbar vertebral bodies;

32

(A) mineralizing surface/bone surface (MS/BS); (B) mineral apposition rate (MAR); (C) bone

33

formation rate (BFR/BS). * p ≤ 0.05 Wilcoxon test versus L1; + p ≤ 0.05 Wilcoxon test versus L4;

34

data are expressed as means ± SEM; the fold-change in L5 (+ BB-1) versus L4 (– BB-1) is indicated;

35

n.d.: not determined.

17 Page 17 of 25

1

Figure 6: Compressive strength of spongious bone cylinders 3 months and 9 months after surgery (n =

2

6 for each group) in differently treated lumbar vertebral bodies. * p ≤ 0.05 Wilcoxon test versus L1; °

3

p ≤ 0.05 Mann Whitney U test versus 3 months; # p ≤ 0.05 Mann Whitney U test versus indicated BB-

4

1 concentration; data are expressed as means ± SEM; NB: in this case, the fold-change versus L1

5

(untouched control) in L5 (+ BB-1) and L4 (– BB-1) is indicated; n.d.: not determined.

6 7

Supplementary Figure 1: (A – D) Structural bone parameters and the bone formation parameter

8

osteoid surface 3 months and 9 months after surgery (n = 6 for each group), as determined by static

9

histomorphometry in differently treated lumbar vertebral bodies: (A) bone volume/total volume

10

(BV/TV); (B) trabecular thickness (Tb.Th); (C) trabecular number (Tb.N); (D) osteoid surface

11

(OS/BS); * p ≤ 0.05 Wilcoxon test versus L1; § p ≤ 0.05 Wilcoxon test versus L3; + p ≤ 0.05

12

Wilcoxon test versus L4; ° p ≤ 0.05 Mann Whitney U test versus 3 months; # p ≤ 0.05 Mann Whitney

13

U test versus indicated BB-1 concentration; data are expressed as means ± SEM; n.d.: not determined.

14 15 16

18 Page 18 of 25

1 2

Fig. 1_bestsetConverted.png

3

19 Page 19 of 25

1 2

Fig. 2_bestsetConverted.png

3

20 Page 20 of 25

1 2

Fig. 3_bestsetConverted.png

3

21 Page 21 of 25

1 2

Fig. 4_bestsetConverted.png

3

22 Page 22 of 25

1 2

Fig. 5_bestsetConverted.png

3

23 Page 23 of 25

1 2

Fig. 6_bestsetConverted.png

3

24 Page 24 of 25

1 2

SuppData Fig. 1_bestsetConverted.png

25 Page 25 of 25