monitoring angiogenesis in soft-tissue engineered constructs ......monitoring angiogenesis in...

TSpace Research Repository tspace.library.utoronto.ca

Monitoring angiogenesis in soft-tissue engineered constructs forcalvarium bone

regeneration: an in-vivolongitudinal DCE-MRI study

Marine Beaumont, Marc G. DuVal, Yasir Loai, Walid A. Farhat, George K. Sándor, Hai-Ling Margaret Cheng

Version Post-print/accepted manuscript

Citation (published version)

Beaumont, Marine, Marc G. DuVal, Yasir Loai, Walid A. Farhat,

George K. Sandor, and Hai‐Ling Margaret Cheng. "Monitoring

angiogenesis in soft‐tissue engineered constructs for calvarium bone

regeneration: an in vivo longitudinal DCE‐MRI study." NMR in

Biomedicine: An International Journal Devoted to the Development and Application of Magnetic Resonance In vivo 23, no. 1 (2010): 48-55.

Publisher’s Statement This is the peer reviewed version of the following article:

Beaumont, Marine, Marc G. DuVal, Yasir Loai, Walid A. Farhat,

George K. Sandor, and Hai‐Ling Margaret Cheng. "Monitoring

angiogenesis in soft‐tissue engineered constructs for calvarium bone

regeneration: an in vivo longitudinal DCE‐MRI study." NMR in

Biomedicine: An International Journal Devoted to the Development and Application of Magnetic Resonance In vivo 23, no. 1 (2010): 48-55. which has been published in final form at 10.1002/nbm.1425. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions.

How to cite TSpace items

Always cite the published version, so the author(s) will receive recognition through services that track

citation counts, e.g. Scopus. If you need to cite the page number of the author manuscript from TSpace because you cannot access the published version, then cite the TSpace version in addition to the published

version using the permanent URI (handle) found on the record page.

This article was made openly accessible by U of T Faculty.

Please tell us how this access benefits you. Your story matters.

https://doi.org/10.1002/nbm.1425

https://tspace.library.utoronto.ca/feedback

Monitoring angiogenesis in soft-tissue engineered constructs for

calvarium bone regeneration: an in-vivo longitudinal DCE-MRI study

Marine Beaumont, PhD1,2

, Marc G. DuVal, MD3, Yasir Loai

4, Walid A. Farhat, MD

1,4, George

K. Sándor, MD3,5,6

, Hai-Ling Margaret Cheng, PhD1,2,7

1The Research Institute, The Hospital for Sick Children, Toronto, Canada

2Diagnostic Imaging, The Hospital for Sick Children, Toronto, Canada

3Department of Oral and Maxillofacial Surgery, University of Toronto, Toronto, Canada

4Division of Urology, The Hospital for Sick Children, Toronto, Canada

5Regea Institute for Regenerative Medicine, University of Tampere, Tampere, Finland

6Department of Oral and Maxillofacial Surgery, University of Oulu, Oulu, Finland

7Department of Medical Biophysics, University of Toronto, Toronto, Canada

Corresponding author:

Hai-Ling Margaret Cheng, Ph.D.

The Hospital for Sick Children

555 University Avenue, Toronto, Ontario M5G 1X8 Canada

Phone: 416-813-5415

Fax: 416-813-7362

[email protected]

Grants: Canadian Institutes of Health Research, Pediatric Regenerative Medicine Fund

Running title: DCE-MRI in tissue-engineered calvarium bone.

DCE-MRI in tissue-engineered calvarium bone 2

ABSTRACT

Tissue engineering is a promising technique for bone repair and can overcome the major

drawbacks of conventional autogenous bone grafting. In this in-vivo longitudinal study, we

proposed a new tissue-engineering paradigm: inserting a biological soft-tissue construct within

the bone defect to enhance angiogenesis for improved bone regeneration. The construct acts as a

resorbable scaffold to support desired angiogenesis and cellular activity and as a vector of

vascular endothelial growth factor, known to promote both vessel and bone growth. Dynamic

contrast-enhanced magnetic resonance imaging was performed to investigate and characterize

angiogenesis necessary for bone formation following the proposed paradigm of inserting a

VEGF-impregnated tissue-engineered construct within the critical-sized calvarial defect in the

membranous parietal bone of the rabbit. Results show that a model-free quantitative approach,

the normalized initial area under the curve metric, provides sensitive and reproducible measures

of vascularity that is consistent with known temporal evolution of angiogenesis during bone

regeneration.

Keywords: Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), tissue

engineering, angiogenesis, bone regeneration.

Abbreviations used: DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging, AIF,

arterial input function, IAUC, initial area under the curve, VEGF, vascular endothelial growth

factor, ACM, acellular matrix, HA, hyaluronic acid, IV, intravenous, FA, flip angle, BW,

bandwidth, NEX, number of experiments, 3D, three dimensional, SPGR, spoiled gradient

recalled echo, ROI, region of interest, micro-CT, micro-computed tomography.


GRAPHICAL ABSTRACT

Monitoring angiogenesis in soft-tissue engineered constructs for calvarium bone

regeneration: an in-vivo longitudinal DCE-MRI study

Marine Beaumont, Marc G. DuVal, Yasir Loai, Walid A. Farhat, George K. Sándor, Hai-Ling

Margaret Cheng.

DCE-MRI was performed to

investigate and characterize

angiogenesis necessary for bone

formation following the proposed

paradigm of inserting a VEGF-

impregnated tissue-engineered

construct within the critical-sized

calvarial defect in the rabbit.

Results show that a model-free

quantitative approach, the

normalized initial area under the

curve metric, provides sensitive and

reproducible measures of

vascularity that is consistent with

known temporal evolution of

angiogenesis during bone

regeneration.


INTRODUCTION

Autogenous bone grafting is considered the gold standard for repair and regeneration of

extensive bony defects following trauma, cancer resection, or non-united fractures (1). However,

grafting bone is non-ideal: a second surgical site, usually involving the iliac crest, tibia,

calvarium, or scapula, is required; the bony skeleton does not offer a limitless supply of bone to

harvest; the procedure is associated with increased operating room time usage and increased

patient morbidity (2). Tissue engineering is one of the most promising techniques to overcome

these major drawbacks of autogenous bone grafting (3). Successful tissue regeneration mainly

depends on the establishment of a vascular bed, which connects the implanted construct to the

host tissue and supplies necessary nutrients for bone regeneration (4). The different mechanisms

underlying bone repair and regeneration have been extensively studied and identified (5,6), but

very few studies have focused on monitoring the angiogenesis process during bone wound

healing.

As a noninvasive imaging technique, MRI is well suited to in-vivo longitudinal

evaluation of tissue repair following grafting procedures. In tissue-engineered constructs, MRI

has been used to determine cellularization and maturation of the construct in vivo in cartilage

(7,8), but bony constructs have been studied only in vitro (9,10). In either case, the vascular

status and evolution of angiogenesis with tissue growth were not investigated. Angiogenesis in

tissue-engineered constructs was first evaluated in bladder constructs in vivo (11-13) using

dynamic contrast-enhanced MRI (DCE-MRI). The foray of DCE-MRI into tissue-engineering

applications remains limited despite its well-established role in assessing the microvasculature in

clinical and research studies (14). A few exceptions are found in in-vivo longitudinal studies

conducted using DCE-MRI to evaluate the wound-healing process following subcutaneous


insertion of wound chambers (15) and massive bone allograft (16). In this latter study,

revascularization was assessed at the osteotomy site but not at the graft site. To our knowledge,

longitudinal in-vivo characterization of angiogenesis in bony constructs has never been reported.

In order to quantify angiogenesis using DCE-MRI derived data, one can use numerous

analysis techniques. The quantitative approach using pharmacokinetic modeling of tracer

concentration (17) provides parameters related to physiology, such as blood volume and vascular

permeability. This approach depends on the knowledge of the tracer concentration evolution in

the blood, or arterial input function (AIF). However, determination of the AIF may be

compromised by tracer concentration-MR signal non-linearity, low temporal resolution, and

partial volume effects, which are especially problematic in small animal brain imaging.

Moreover, it has been shown that errors in AIF measurement can strongly influence

pharmacokinetic parameter accuracy (18,19). To overcome this limitation, a model-free

quantitative approach has been proposed (20), which uses the initial area under the tracer

concentration-time curve (IAUC) as an empirical quantitative parameter. Change in the IAUC

has been shown to correlate with vessel density (13) and tumor response to anti-vascular

treatment (21,22). Also, this non-model-based analysis is very robust, and reproducibility is often

comparable to or even better than that obtained with traditional model-based parameters (23-25).

In this in-vivo longitudinal study, we propose a new tissue-engineering paradigm:

inserting a biological soft-tissue construct within the bone defect to enhance angiogenesis for

improved bone regeneration. The construct acts as a resorbable scaffold to support desired

angiogenesis and cellular activity and as a vector of vascular growth factor (VEGF), known to

promote both vessel and bone growth (26). DCE-MRI was performed to investigate and

characterize angiogenesis necessary for bone formation following the proposed paradigm of


inserting a VEGF-impregnated tissue-engineered construct within the critical-sized calvarial

defect in the membranous parietal bone of the rabbit (27). Results show that an IAUC approach

provides sensitive and reproducible measures of vascularity that is consistent with known

temporal evolution of angiogenesis during bone regeneration.

EXPERIMENTAL

This study was approved by the institutional Animal Care Committee (protocol #7578).

Construct Preparation and Implantation

Tissue-engineered construct preparation began with the acellularization of urinary

bladder harvested from 20-50 kg porcines. Bladders were washed with sterile phosphate buffer

saline, longitudinally sectioned and stirred in hypotonic solution for 48 hours at 4°C to

effectively break down cellular structures and inhibit proteases. Bladders were placed in

hypertonic solution for 48 hours at 4°C to denature protein residues and degrade all DNA and

RNA components. Bladders were washed with Hank’s Balanced Salt Solution (Invitrogen,

#14175, USA) containing 2 U/mL Benzonase (Novagen, #: 70746, Germany) overnight at 37°C

and transferred to 0.25% CHAPS-containing detergent based solution. Acellular matrices

(ACMs) were repeatedly washed with sterile dH2O and stored in 70% ethanol prior to use. H&E

staining was used to confirm acellularity. ACMs were cut into 1.5 cm diameter discs, weighed,

and immersed in gradually increasing concentrations of ethanol for full dehydration. ACMs were

lyophilized (ViTis-temp, 120 millitorr and vacuum) for 24 hours, then rehydrated with increasing

concentrations of HA (0.05, 0.1, 0.2 and 0.5 mg/100mL) (Sigma, product #H5388, USA). Alcian

blue staining was used to confirm HA incorporation. In addition, HA-ACMs were dehydrated in

alcohol, lyophilized, and rehydrated with VEGF121 (Sigma, #V3388, USA) 10 ng/g of ACM.


Finally, a critical-sized (non-spontaneous healing) calvarial defect (15 mm in diameter) was

created surgically in the parietal bones of the rabbits, the centre of the defect was 9.5 mm from

the midline to allow 2 mm for the sagittal sinus. The tissue-engineered HA-VEGF containing

construct was grafted to the calvarium.

Experimental Protocol

Surgery and MR imaging were performed on twelve New Zealand white rabbits (3.5-

4 kg). Imaging sessions were held 1, 2, 3, 6 and 12 weeks after the surgery. All MR experiments

were performed under anesthesia using two different protocols. Five rabbits were anesthetized

with an intravenous (IV) injection of pentobarbital (25 mg/kg) through the ear vein. During the

MRI (about 1 hour), anesthesia was maintained with an IV dose of 0.25 mg/kg/min; and the

animals were kept under 100% oxygen with a face mask. For the other seven rabbits, anesthesia

was performed with 5% of isoflurane for induction, and 2±0.2% for maintenance, in 100%

oxygen using a face mask. All the animals were equipped with a catheter in the ear vein for the

contrast agent injection (Gd-DTPA – Magnevist, Berlex Canada, Lachine, Canada) and

continuous hydration (4 mL of saline/kg/h). Heart rate and pO2 were monitored during the scan

time. Five animals were euthanized after the fourth imaging session (6 weeks after surgery)

while the remaining animals were euthanized after the last imaging session (12 weeks after

surgery). In each animal, the calvarium was excised for subsequent histological procedures.

MRI

MRI was performed on a 1.5 T GE scanner (Signa EXCITE TwinSpeed; GE Healthcare,

Milwaukee, WI, USA) using a three-inch diameter receive surface coil and a body transmit coil.

Animals were placed in prone position and an axial orientation was chosen for all the slices.

Construct localization was achieved using a gradient echo sequence: TR/TE = 6.5/2.3 ms,


FA = 30°, bandwidth (BW) = 23.4 kHz, 16 x 12 x 8 slices (thickness = 2 mm, space between

slices = 1 mm), matrix = 192 x 192, FOV = 12 x 12 cm2, number of experiments (NEX) = 2. A

rapid 3D T1-mapping method, based on variable flip angles and integrating B1 correction (28),

was employed prior to Gd-DTPA injection to acquire the baseline T1 map: 3D fast spoiled

gradient recalled echo (SPGR), TR/TE = 8.8/3.4 ms, FA = 2°, 10° and 20°, bandwidth

(BW) = 31.2 kHz, matrix = 256 x 160 x 10, FOV = 10 x 10 x 3 cm3, number of experiments

(NEX) = 4. Dynamic T1-weighted images were acquired using a 3D fast SPGR sequence:

TR/TE = 10.2/3.4 ms, FA = 60°, BW = 31.2 kHz, matrix = 256 x 128 x 10,

FOV = 10 x 10 x 3 cm3, NEX = 0.75. After the acquisition of 5 baseline images, Gd-DTPA was

administered as a rapid bolus (0.1 mmol of Gd/kg) and imaging continued for 6 min post-

injection with a time resolution of 8.2 s (300 images in total).

MRI data analysis

All analyses were performed on a pixel-wise basis using in-house programs developed

with Matlab (v.7.0, Mathworks Inc., Natick, MA, USA). Analysis of the 3D T1 maps followed

the method described previously (28). DCE-MRI data were converted into Gd concentration

maps using the pre-injection T1 map, the SPGR signal equation (29) and the Gd-DTPA relaxivity

r1 (4.1 s-1

.mM-1

, in plasma at 1.5 T and 37°C (30)). Finally, the initial area under the Gd

concentration time curve for 60s after contrast agent injection (IAUC60) was computed for each

pixel. Only physiologically meaningful pixels were retained. Pixels were excluded if they met

any of the following conditions: T1 values outside of the range 0-4000 ms, Gd concentration

mean value over time outside of the range 0-2 mM and for which the standard deviation of Gd

concentration along the baseline was greater than 0.1 mM. Three regions of interest (ROI) were

manually drawn. On the IAUC60 map, two ROIs were defined on the construct: the periphery,


which corresponds to the early enhancing region, and the centre (Figure 1c); and one ROI

surrounding the whole brain. For each ROI and for each MR acquisition time point, the mean

IAUC60 value was computed as well as the percentage of physiologically meaningful pixels in

the ROI involved in this calculation. To minimize effects from inter-individual physiological

variation and contrast agent dose variability, each IAUC60 mean value was normalized to a

reference, the IAUC60 mean value measured in the brain.

Statistical analysis was based on the permutation test (also called randomization test,

(31)), which, unlike the widely used t-test, is suited to data originating from small-sized groups

and with unknown distributions. This test determines the significance of the difference of means

d between two populations A and B with nA and nB samples, respectively, as follows. A

difference of means d’ is calculated between all possible groupings A’ and B’, where A’ and B’

represent a different partition of values from A+B into two groups of size nA and nB. The P-value

P is equal to the ratio of the number of events where d’>d to the total number of events. Paired

permutation tests were applied to compare IAUC60 at different time points, and statistical

significance was set at P < 0.05. Evolution of the IAUC60 mean value with time was also

determined for each individual animal and in each ROI by comparing the mean value of the

estimate at week (i) with the mean±SEM value at week (i-1).

Micro-Computed Tomography

Calvarial specimens were scanned by an Explore Locus SP micro-CT scanner (GE

Medical Systems, London, Ontario, Canada) using the fast mode with 0.05 mm sections.

Reconstruction of scanned images was done using Microview software (GE Medical Systems,

London, Ontario, Canada) after calibrating using the bone water and air standard values. The

reconstructed 3D image was then traced in 3 dimensions to the circumference of the original


defect margins. This allowed the creation of a 3D reconstruction of the defect, referred to as the

ROI. A threshold level was selected manually based on 25% of the bone standard provided by

the manufacturer.

RESULTS

MRI

Figure 1 illustrates that the soft-tissue construct was clearly identifiable on MR images

even prior to contrast agent injection (Fig.1a). The T1-weighted MR signal increased in the

whole construct after contrast agent injection with a more pronounced enhancement in the

periphery than in the centre of the graft (Fig.1b). Note that contrast uptake was observed from

the first imaging time point (1 week after surgery) up to the last (12 weeks after surgery) in all

rabbits. The IAUC60 map provided the best support to draw the different ROIs with very clean

contours of the construct (Fig.1c).

Table 1 shows the evolution of the percentage of physiologically meaningful pixels in

these ROIs that are retained for DCE-MRI analysis. The percentage of pixels retained is

generally highest ( 91%) in the construct periphery compared to the centre or normal brain

tissue. An exception occurs at 12 weeks, where fewer useful pixels are detected in the entire

construct. In the centre, the percentage of retained pixels increases with angiogenic growth. In

the normal brain where contrast agent uptake is lowest, at least 75% of pixels were retained

consistently throughout the 12 weeks, which supports the robustness of the proposed method for

pixel selection. Table 2 reports the evolution of the size of the ROIs normalized to week 1. For

both periphery and centre of the construct, the size of the ROI 6 weeks after the surgery is almost

half of the initial size. The size of the ROI defined on the brain remains constant with time.


Figure 2 illustrates in one rabbit typical uptake curves in the periphery and the centre of

the construct at four time points (1, 3, 6 and 12 weeks after surgery). The uptake curves in the

periphery were characteristic of highly permeable vessels usually found in presence of active

angiogenesis. At the early time point, the periphery enhancement curve (Fig.2, Week 1) exhibits

the most rapid initial uptake followed by a distinct washout. Later time points may also show

significant enhancement, but the initial uptake is more gradual and washout is not always evident

(e.g. Fig.2, Week 6). In the centre of the construct, the uptake dynamics were distinct from those

in the periphery, characterized by a much lower and gradual uptake. Both ROIs exhibited less Gd

uptake at the last time point of 12 weeks after graft implantation.

Figure 3 shows the evolution of the mean IAUC60 values normalized to the reference

(IAUC60 in the brain) in both the periphery and centre of the construct throughout the 12 weeks.

The periphery exhibited the highest value at the early time point, followed by a plateau and a

two-fold decrease at 12 weeks post-surgery. The centre of the construct exhibited different

characteristics. Normalized IAUC60 values were significantly lower at all time points, and a

progressive increase was observed at early times up to 3 weeks post-surgery. The one common

trait observed in both the centre and periphery was a decrease in uptake at 12 weeks. Values for

all animals are shown in Table 3.

Figure 4 shows how angiogenesis evolves in individual animals at time points where

significant changes were observed, namely, between weeks 1 and 3 and weeks 6 and 12. It also

distinguishes groups administered different anesthetics (solid versus dashed lines). Several key

results are noted in this figure. First, the proposed DCE-MRI approach yields reproducible trends

in angiogenic development in most animals: a significant early decrease in the periphery versus

an increase in the centre (in each case, the trend was observed on 8 animals out of 12), followed


by a decrease in the entire construct at 12 weeks. Second, this approach is insensitive to

differential vascular response to anesthesia that is independent of angiogenesis. An exception to

this is seen at 12 weeks (Table 3), where normalized IAUC60 values are significantly higher in

the isoflurane group. Although the sample size was small (n=2), the possibility of anesthesia

influence on vascular development at 12 weeks could not be disregarded, and these two animals

were excluded from the twelve weeks mean values in Fig. 3 and Table 3. It is important to note

that aside from the 12-week discrepancy, these two animals followed the general trend at earlier

time points and were part of the 8 out of 12 animals showing the same behavior between week 1

and week 3. Further investigations are required to confirm this hypothesis regarding anesthesia-

induced differences.

Micro-Computed Tomography

Figure 5 shows representative micro-CT images of excised calvarium. Micro-CT

evaluation supported the MRI findings. There was new bone noted at 6 weeks, mostly at the

periphery of the bony defects containing the HA-VEGF constructs, while at 12 weeks islands of

newly formed bone had appeared in the center of the defects.

DISCUSSION

Tissue engineering is a promising technique for bone repair and can overcome the major

drawbacks of conventional autogenous bone grafting. Instead of using a surgically harvested

bone graft, a biological soft-tissue construct is placed within the bony defect. This construct acts

as a resorbable scaffold to support two intimately linked processes during tissue regeneration:

cellular colonization and angiogenesis. Monitoring angiogenesis longitudinally, for instance,

using DCE-MRI, provides a means to assess bone development non-invasively. In this study, a


model-free analysis of DCE-MRI data was used to investigate and characterize angiogenesis in a

VEGF-impregnated tissue-engineered construct within the rabbit calvarial defect. Vessel

network initiation and establishment within the construct, leading to successful bone

regeneration, were demonstrated in vivo using normalized IAUC60 measurements. The proposed

approach provides reproducible and sensitive results and is robust to challenging experimental

conditions.

Natural wound healing and bone regeneration after autogeneous bone grafting have been

extensively described in the literature. Bone regeneration results from a combination of different

processes, or phases, which occur usually in a dynamic equilibrium (6). Although our proposed

paradigm uses a soft-tissue construct instead of a bone graft, the same phases are required to

ensure successful bone regeneration. In fact, the collagen-based scaffold may be more easily

resorbed than traditional bone grafts to allow new bone formation. Also, the soft-tissue construct

provides the ideal three-dimensional framework to guide desired angiogenesis (4), which is a

major determinant to bone formation. The different processes involved in bone formation are as

follows. The early inflammatory phase is characterized by the different components of wound

healing: inflammation, hemorrhage, hematoma, and blood clot formation. These combined

phenomena stimulate macrophages to release growth factors (such as VEGF), which induces

angiogenesis (32). Vascularization is the second stage in graft healing and is strongly associated

with the previous phase. It starts shortly after graft implantation, and the mature vessel network

is established within 2 to 3 weeks following surgery. The third phase significantly overlaps with

the vascularization phase and may last up to 6 weeks. During this time span, the graft is

colonized by cells, nutrients, and other growth factors necessary for new bone formation via the

newly established vasculature. The final phase, the remodeling phase, is initiated between the


fourth and the sixth week post-surgery and continues for several months. It involves remodeling

and mineralization of the immature randomly-oriented bone. The two latter phases are also

accompanied by the resorption of the graft material (6).

In our study, periphery and centre regions of the soft-tissue construct exhibited different

behaviors with regards to angiogenic activity, but in each case, the different phases observed

were consistent with expected angiogenic progression, as described in the literature. In the

periphery, the high IAUC60 value observed one week after the surgery is consistent with the first

two phases of inflammation and angiogenesis, since both lead to high contrast agent uptake due

to high vessel leakiness. Then, between the second and the sixth week after insertion, when both

angiogenesis and the third phase are expected to occur, the IAUC60 is lower and remains

relatively constant. The lower contrast agent uptake is consistent with the completion of the

inflammatory phase. The plateau in IAUC60 needs to be interpreted carefully, since different

uptake curve dynamics (at weeks 3 and 6) suggest distinct vascular behavior that possibly

indicates a transition from phase 2 to 3. The IAUC60 contribution comes mainly from

inflammatory/angiogenic activity initially and later from a mature vasculature that is functioning

fully to supply the graft with elements to the new bony tissue. Finally, the drop at twelve weeks

is expected since the remodeling phase no longer requires a large vascular supply, perhaps

resulting in subsequent vessel pruning (33). In the centre, the inflammation phase is not

distinguishable because of the distance from the actual surgery site. As for the three other phases,

they are clearly separated: the slow IAUC60 increase reflects the growing vessels during

angiogenesis, the plateau corresponds to the new bone formation phase, and the decrease is the

same as in the periphery. Nevertheless, it should be pointed out that the uptake curve pattern

observed in the centre may reflect primarily diffusion of the contrast agent from the construct


periphery to the core and is not a purely vascular-driven response. In fact, diffusion is a

fundamental means of transport in tissue regeneration since some elements, like cells, are not

vehicle from the host tissue onto the graft via the blood, but using diffusion along the scaffold.

Finally, the size of ROIs defined on the construct decreased significantly from week 1, indicating

resorption of the graft while new bony tissue is created as observed on the micro-CT results.

In this work, the IAUC approach was preferred to more common model-based

approaches to analyze DCE-MRI data. Indeed, our experimental conditions (small animal brain

imaging with clinical magnetic field, low temporal resolution) rendered the determination of the

AIF very challenging and would have compromised a model-based analysis (19). Although

empirical approaches do not have clear physiological associations, our results show that our

normalized IAUC60 metric is very sensitive to the different stages that occur during graft healing

and bone regeneration, and can monitor the regeneration process and its success. Moreover,

when quantitative model-based approaches are used, numerous factors (AIF errors, limited

temporal resolution, choice of analysis model) may impair robustness and accuracy (18). When

using an IAUC metric, several approaches have been taken. Walker-Samuel et al. (34)

recommended the use of IAUC90 for in-vivo studies. But they also mentioned that the integration

period does not matter if the interval is long enough to ensure good signal-to-noise ratio. In

another study, Morgan et al. (24) showed there were no major changes by using IAUC60 or

IAUC180. Furthermore, they demonstrated that reproducibility was improved when IAUC was

computed from the T1 values rather than from the MR signal intensities. Finally, Evelhoch et al.

(21) proposed normalizing the data with an AIF for longitudinal studies. Considering all these

elements, we used normalized IAUC60 values derived from the Gd concentration time curve to

achieve accuracy. Although our normalization was performed using the whole brain IAUC60


value rather than the AIF one, the automatic selection of enhancing pixels implemented provided

results sensitive enough to minimize effects from inter-individual physiological variation and

contrast agent dose variability. Muscle tissue was not chosen as a reference in this study because

of low signal-to-noise ratio due to greater distance from the surface coil. The good sensitivity of

the IAUC approach demonstrated in this study warrants further investigation to achieve

improved distinction of uptake curve patterns.

The presented normalized IAUC60 method also demonstrated robustness and insensitivity

to the anesthetic procedure (one exception occured at the last time point), whereas the non-

normalized IAUC60 values presented significant differences (results not shown). Indeed, the two

anesthetics used in our protocol are well known to induce antagonistic physiological effects,

mainly on the vasculature (35). Those effects have been investigated using MRI and they have

been shown to influence the MR signal (36). As demonstrated by Hendrich et al. (37), using

arterial spin labeling on rat, cerebral blood flow, as well as pO2 and mean arterial blood pressure,

measured on animals undergoing isoflurane anesthesia are higher than the ones of animals

anesthetized with pentobarbital. If the proposed approach successfully homogenized the two

groups without hiding the relevant changes in angiogenesis activity, a discrepancy appeared at 12

weeks after surgery. Several hypotheses may be raised to explain this result. First, one can

suggest that the sample size of the isoflurane group at week 12 (n = 2), is not large enough to

reflect an actual difference, and this result may be biased by inter-animal variability. Another

hypothesis, as mentioned above, is long-term remodeling of the newly established vasculature,

which normally leads to partial vessel pruning. The repeated isoflurane anesthesia could have an

anti-pruning effect on the vasculature, resulting in a higher blood volume and, thus, a higher

normalized IAUC60 value. This explanation is likely, since isoflurane is known for having


preconditioning properties, as shown in stroke (38). Finally, as isoflurane and pentobarbital

generate vasodilatation and vasoconstriction, respectively, the trend observed may be simply an

emphasis of the natural trend, because vessels are more mature and more reactive to those

chemical stimulations. A combination of these three hypotheses may also explain the reported

difference. Understanding this phenomenon requires further investigation.

In conclusion, this study provides the first report of an in vivo longitudinal DCE-MRI

study in tissue-engineered calvarium bone. The soft-tissue construct acts as a resorbable support

for promoting angiogenesis and achieving complete bone regeneration. The normalized IAUC60

approach used for the DCE-MRI analysis was capable of characterizing angiogenesis and

discriminating between the different phases of bone regeneration, and was robust to challenging

experimental conditions.

ACKNOWLEDGMENTS

We thank Marvin Estrada for animal care support and Tammy Rayner, Ruth Weiss, and

Garry Detzler for technical assistance during MR scanning.


REFERENCES

1. Giannoudis PV, Dinopoulos H, Tsiridis E. Bone substitutes: an update. Injury 2005;36

Suppl 3:S20-27.

2. Ebraheim NA, Elgafy H, Xu R. Bone-graft harvesting from iliac and fibular donor sites:

techniques and complications. J Am Acad Orthop Surg 2001;9(3):210-218.

3. Kneser U, Schaefer DJ, Polykandriotis E, Horch RE. Tissue engineering of bone: the

reconstructive surgeon's point of view. J Cell Mol Med 2006;10(1):7-19.

4. Kanczler JM, Oreffo RO. Osteogenesis and angiogenesis: the potential for engineering

bone. Eur Cell Mater 2008;15:100-114.

5. Axhausen W. The osteogenetic phases of regeneration of bone; a historial and

experimental study. J Bone Joint Surg Am 1956;38-A(3):593-600.

6. Goldberg VM, Stevenson S. Natural history of autografts and allografts. Clin Orthop

Relat Res 1987(225):7-16.

7. Watrin-Pinzano A, Ruaud JP, Cheli Y, Gonord P, Grossin L, Bettembourg-Brault I, Gillet

P, Payan E, Guillot G, Netter P, Loeuille D. Evaluation of cartilage repair tissue after

biomaterial implantation in rat patella by using T2 mapping. MAGMA 2004;17(3-6):219-

228.

8. Trattnig S, Mamisch TC, Welsch GH, Glaser C, Szomolanyi P, Gebetsroither S, Stastny

O, Horger W, Millington S, Marlovits S. Quantitative T2 mapping of matrix-associated

autologous chondrocyte transplantation at 3 Tesla: an in vivo cross-sectional study. Invest

Radiol 2007;42(6):442-448.

9. Moinnes JJ, Vidula N, Halim N, Othman SF. Ultrasound accelerated bone tissue

engineering monitored with magnetic resonance microscopy. Conf Proc IEEE Eng Med

Biol Soc 2006;1:484-488.

10. Xu H, Othman SF, Hong L, Peptan IA, Magin RL. Magnetic resonance microscopy for

monitoring osteogenesis in tissue-engineered construct in vitro. Phys Med Biol

2006;51(3):719-732.


11. Cartwright L, Farhat WA, Sherman C, Chen J, Babyn P, Yeger H, Cheng HL. Dynamic

contrast-enhanced MRI to quantify VEGF-enhanced tissue-engineered bladder graft

neovascularization: pilot study. J Biomed Mater Res A 2006;77(2):390-395.

12. Cheng HL, Chen J, Babyn PS, Farhat WA. Dynamic Gd-DTPA enhanced MRI as a

surrogate marker of angiogenesis in tissue-engineered bladder constructs: a feasibility

study in rabbits. J Magn Reson Imaging 2005;21(4):415-423.

13. Cheng HL, Wallis C, Shou Z, Farhat WA. Quantifying angiogenesis in VEGF-enhanced

tissue-engineered bladder constructs by dynamic contrast-enhanced MRI using contrast

agents of different molecular weights. J Magn Reson Imaging 2007;25(1):137-145.

14. Cheng HL. Dynamic contrast-enhanced MRI in oncology drug development. Curr Clin

Pharmacol 2007;2(2):111-122.

15. Helbich TH, Roberts TP, Rollins MD, Shames DM, Turetschek K, Hopf HW, Muhler M,

Hunt TK, Brasch RC. Noninvasive assessment of wound-healing angiogenesis with

contrast-enhanced MRI. Acad Radiol 2002;9 Suppl 1:S145-147.

16. Ehrhart N, Kraft S, Conover D, Rosier RN, Schwarz EM. Quantification of massive

allograft healing with dynamic contrast enhanced-MRI and cone beam-CT: a pilot study.

Clin Orthop Relat Res 2008;466(8):1897-1904.

17. Tofts PS. Modeling tracer kinetics in dynamic Gd-DTPA MR imaging. J Magn Reson

Imaging 1997;7(1):91-101.

18. Cheng HL. Investigation and optimization of parameter accuracy in dynamic contrast-

enhanced MRI. J Magn Reson Imaging 2008;28(3):736-743.

19. Port RE, Knopp MV, Brix G. Dynamic contrast-enhanced MRI using Gd-DTPA:

interindividual variability of the arterial input function and consequences for the

assessment of kinetics in tumors. Magn Reson Med 2001;45(6):1030-1038.

20. Evelhoch JL. Key factors in the acquisition of contrast kinetic data for oncology. J Magn

Reson Imaging 1999;10(3):254-259.

21. Evelhoch JL, LoRusso PM, He Z, DelProposto Z, Polin L, Corbett TH, Langmuir P,

Wheeler C, Stone A, Leadbetter J, Ryan AJ, Blakey DC, Waterton JC. Magnetic

resonance imaging measurements of the response of murine and human tumors to the

vascular-targeting agent ZD6126. Clin Cancer Res 2004;10(11):3650-3657.


22. Galbraith SM, Maxwell RJ, Lodge MA, Tozer GM, Wilson J, Taylor NJ, Stirling JJ, Sena

L, Padhani AR, Rustin GJ. Combretastatin A4 phosphate has tumor antivascular activity

in rat and man as demonstrated by dynamic magnetic resonance imaging. J Clin Oncol

2003;21(15):2831-2842.

23. Galbraith SM, Lodge MA, Taylor NJ, Rustin GJ, Bentzen S, Stirling JJ, Padhani AR.

Reproducibility of dynamic contrast-enhanced MRI in human muscle and tumours:

comparison of quantitative and semi-quantitative analysis. NMR Biomed

2002;15(2):132-142.

24. Morgan B, Utting JF, Higginson A, Thomas AL, Steward WP, Horsfield MA. A simple,

reproducible method for monitoring the treatment of tumours using dynamic contrast-

enhanced MR imaging. Br J Cancer 2006;94(10):1420-1427.

25. Roberts C, Issa B, Stone A, Jackson A, Waterton JC, Parker GJ. Comparative study into

the robustness of compartmental modeling and model-free analysis in DCE-MRI studies.

J Magn Reson Imaging 2006;23(4):554-563.

26. Fok TC, Jan A, Peel SA, Evans AW, Clokie CM, Sandor GK. Hyperbaric oxygen results

in increased vascular endothelial growth factor (VEGF) protein expression in rabbit

calvarial critical-sized defects. Oral Surg Oral Med Oral Pathol Oral Radiol Endod

2008;105(4):417-422.

27. Jan AM, Sandor GK, Iera D, Mhawi A, Peel S, Evans AW, Clokie CM. Hyperbaric

oxygen results in an increase in rabbit calvarial critical sized defects. Oral Surg Oral Med

Oral Pathol Oral Radiol Endod 2006;101(2):144-149.

28. Cheng HL, Wright GA. Rapid high-resolution T(1) mapping by variable flip angles:

accurate and precise measurements in the presence of radiofrequency field

inhomogeneity. Magn Reson Med 2006;55(3):566-574.

29. Haase A. Snapshot FLASH MRI. Applications to T1, T2, and chemical-shift imaging.

Magn Reson Med 1990;13(1):77-89.

30. Rohrer M, Bauer H, Mintorovitch J, Requardt M, Weinmann HJ. Comparison of

magnetic properties of MRI contrast media solutions at different magnetic field strengths.

Invest Radiol 2005;40(11):715-724.

31. Pitman EJG. Significance Tests Which May be Applied to Samples From any

Populations. Suppl J R Statist Soc 1937;4(1):119-130.


32. Marx RE, Ehler WJ, Peleg M. "Mandibular and facial reconstruction" rehabilitation of

the head and neck cancer patient. Bone 1996;19(1 Suppl):59S-82S.

33. Ghajar CM, George SC, Putnam AJ. Matrix metalloproteinase control of capillary

morphogenesis. Crit Rev Eukaryot Gene Expr 2008;18(3):251-278.

34. Walker-Samuel S, Leach MO, Collins DJ. Evaluation of response to treatment using

DCE-MRI: the relationship between initial area under the gadolinium curve (IAUGC)

and quantitative pharmacokinetic analysis. Phys Med Biol 2006;51(14):3593-3602.

35. Todd MM, Wu B, Warner DS, Maktabi M. The dose-related effects of nitric oxide

synthase inhibition on cerebral blood flow during isoflurane and pentobarbital anesthesia.

Anesthesiology 1994;80(5):1128-1136.

36. Baudelet C, Gallez B. Effect of anesthesia on the signal intensity in tumors using BOLD-

MRI: comparison with flow measurements by Laser Doppler flowmetry and oxygen

measurements by luminescence-based probes. Magn Reson Imaging 2004;22(7):905-912.

37. Hendrich KS, Kochanek PM, Melick JA, Schiding JK, Statler KD, Williams DS, Marion

DW, Ho C. Cerebral perfusion during anesthesia with fentanyl, isoflurane, or

pentobarbital in normal rats studied by arterial spin-labeled MRI. Magn Reson Med

2001;46(1):202-206.

38. Kitano H, Kirsch JR, Hurn PD, Murphy SJ. Inhalational anesthetics as neuroprotectants

or chemical preconditioning agents in ischemic brain. J Cereb Blood Flow Metab

2007;27(6):1108-1128.


Table 1: Evolution of the percentage of physiologically meaningful pixels (within the ranges of

validity defined in MRI data analysis).

Physiologically meaningful pixels (% - mean ± SEM)

Time after surgery (weeks)

ROI n 1 2 3 6 12 n

Periphery 12 94 ± 2 94 ± 3 92 ± 1 91 ± 2 75 ± 6 5

Centre 12 77 ± 4 80 ± 5 93 ± 3 91 ± 4 69 ± 6 5

Brain 12 75 ± 2 75 ± 2 71 ± 3 78 ± 2 77 ± 3 5


Table 2: Evolution of the size of the ROI normalized to week 1.

Size of the ROI normalized to week 1 (% - mean ± SEM)


ROI n 1 2 3 6 12 n

Periphery 12 100 88 ± 6 * 88 ± 6 57 ± 6 * 70 ± 12 5

Centre 12 100 73 ± 9 * 72 ± 10 56 ± 7 * 34 ± 6 5

Brain 12 100 101 ± 1 103 ± 2 100 ± 2 102 ± 2 5

* P < 0.05 week (i) versus week (i-1)


Table 3: Evolution of normalized IAUC60 values presented in all animals and in each anesthetic

group.

IAUC60ROI

/ IAUC60brain

(mean ± SEM)


ROI n 1 2 3 6 12 n

Periphery

All 12 8.42 ± 0.48 6.44 ± 0.47 6.56 ± 0.39 6.40 ± 0.57 3.44 ± 0.45 5

Pentobarbital 5 7.93 ± 0.83 5.14 ± 0.41 * 6.73 ± 0.70 5.41 ± 0.44 3.44 ± 0.45 * 5

Isoflurane 7 8.78 ± 0.60 7.37 ± 0.53 * 6.44 ± 0.50 7.10 ± 0.86 7.46 ± 0.69 * 2

Centre

All 12 1.92 ± 0.29 2.15 ± 0.26 2.44 ± 0.22 2.41 ± 0.34 1.35 ± 0.22 5

Pentobarbital 5 1.49 ± 0.13 1.64 ± 0.13 2.13 ± 0.37 2.19 ± 0.73 1.35 ± 0.22 * 5

Isoflurane 7 2.23 ± 0.47 2.51 ± 0.39 2.65 ± 0.26 2.57 ± 0.32 2.95 ± 0.61 * 2

* P < 0.05 Pentobarbital versus Isoflurane


FIGURES LEGENDS

Figure 1: T1-weighted images (a) before, (b) 60s after Gd-DTPA injection and (c) the

corresponding IAUC60 map. Imaging time is 3 weeks post-surgery. A zoomed-in version of the

implanted VEGF-impregnated soft-tissue construct is shown in insets. Note the two distinct

regions of the construct (periphery and centre) on images (b) and (c). The periphery ROI is

delineated on the inset of (c).

Figure 2: Representative Gd-DTPA concentration-time curves in one animal between 1 and 12

weeks post-surgery. Mean values in the periphery and the centre of the implanted construct are

shown.

Figure 3: Evolution of angiogenesis in the periphery and centre of the implanted construct in all

animals (n=12, n=5 for 12 weeks) using the normalized IAUC60 metric. Results are presented as

mean±SEM. P<0.05: different relative to week 1, # different relative to week 6.

Figure 4: Time points showing significant changes in angiogenic development in all animals.

Evolution of the vascular response does not differ between the two anesthetic procedures

(isoflurane (dashed lines) and pentobarbital (solid lines)). On each graph, the number of rabbits

showing an increased (↑), decreased (↓), or constant (=) change is indicated.

Figure 5: Evaluation of bone growth with micro-CT images. (a) At 6 weeks after the surgery,

new bone is present mostly at the periphery of the bony defect containing the HA-VEGF

construct. (b) At 12 weeks, there were islands of newly formed bone in the center of the defect.

The dotted line indicates the midline and the circle is an estimate of the original defect.

Figure 1

Figure 2

b a c

Week 1

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 1 2 3 4 5 6

Time (min)

[Gd

] (m

M)

Periphery

CentreWeek 3

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 1 2 3 4 5 6

Time (min)

[Gd] (m

M)

Periphery

Centre

Week 6

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 1 2 3 4 5 6

Time (min)

[Gd] (m

M)

Periphery

CentreWeek 12

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 1 2 3 4 5 6

Time (min)

[Gd] (m

M)

Periphery

Centre

1

2

3

4

5

6

7

8

9

0 2 4 6 8 10 12


IAU

C60R

OI / IA

UC

60b

rain

Periphery

Centre

Figure 3

#

#

Figure 4

0

1

2

3

4

5

6 12


IAU

C6

0R

OI /

IA

UC

60

bra

in

0

1

2

3

4

5

1 3


IAU

C6

0R

OI /

IA

UC

60

bra

in

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

6 12

Time after surgery (week)

IAU

C60R

OI /

IAU

C60b

rain

(a.u

.) Pentobarbital

Isoflurane

2

4

6

8

10

12

1 3


IAU

C6

0R

OI /

IA

UC

60

bra

in

0

2

4

6

8

10

6 12


IAU

C6

0R

OI /

IA

UC

60

bra

in

Centre Periphery a

c

b

d

↑ 2 = 2 ↓ 8

↑ 8 = 1 ↓ 3

↑ 1 = 2 ↓ 4

↑ 0 = 1 ↓ 6

Figure 5

a b

monitoring angiogenesis in soft-tissue engineered constructs ......monitoring angiogenesis in...

Documents