Gene Therapy (2009) 16, 78–92 & 2009 Macmillan Publishers Limited All rights reserved 0969-7128/09 $32.00 www.nature.com/gt
ORIGINAL ARTICLE
Microvascular free tissue transfer for gene delivery: in vivo evaluation of different routes of plasmid and adenoviral delivery VK Agrawal1,2, KM Copeland1, Y Barbachano3, A Rahim1, R Seth1, CL White1, M Hingorani1, CM Nutting2, M Kelly2, P Harris2, H Pandha4, AA Melcher5, RG Vile6, C Porter1 and KJ Harrington1,2 1
Chester Beatty Laboratories, The Institute of Cancer Research, London, UK; 2Head and Neck Unit, Royal Marsden Hospital, London, UK; 3Department of Statistics, Royal Marsden Hospital, Sutton, UK; 4University of Surrey, Guildford, UK; 5St James’s University Hospital, Leeds, UK and 6St George’s Hospital, Cranmer Terrace, London, UK
Transfer of healthy autologous tissue as a microvascular free flap facilitates reconstruction during ablative cancer surgery. In addition to filling surgical defects, free flaps might concentrate viral vectors at the tumour bed and mediate local therapeutic effects. We evaluated the magnitude, topography and duration of luciferase gene expression after plasmid and adenoviral delivery in rat superficial inferior epigastric (SIE) flaps. For plasmid delivery, luciferase expression was significantly increased by all transduction routes (topical, intraflap injection, intravascular) ( Po0.01) at day 1, but not at day 7. The spread of luciferase expression was significantly different between the 4 groups at 1 day ( P ¼ 0.026) and was greatest for flaps
transduced by intravascular injection. For adenoviral transduction, total radiance was significantly different between the transduced groups at 1, 14 and 28 days (Po0.05 for all comparisons). The highest levels of radiance were seen in the intravascular group. There was a statistically significant difference in the spread of light emission between the 3 groups at 1 ( P ¼ 0.009) and 14 (P ¼ 0.013) days, but this was no longer evident at 28 days. Intravascular adenoviral delivery yields high-level, diffuse and durable gene expression in rat SIE flaps and is suitable for examination in therapeutic models. Gene Therapy (2009) 16, 78–92; doi:10.1038/gt.2008.140; published online 11 September 2008
Keywords: adenovirus; free flap; gene expression; luciferase; plasmid
Introduction For many tumours, such as head and neck cancer, treatment of the primary (and locoregional lymph nodes) by surgery is potentially curative.1–3 However, with locally invasive disease, there is a significant chance of leaving microscopic residual disease (MRD) in the surgical bed.4 The risk of MRD contaminating the surgical field is greatest with large, locally advanced tumours and necessitates resection of large volumes of tissue en bloc to achieve clear margins of excision.5 Such ablative procedures leave large defects that cannot simply be closed by suturing the wound edges together. Instead, fresh autologous tissue must be mobilized (either from adjacent or distant sites) and transferred to the surgical site during reconstructive surgery.6 Recent advances in microsurgical techniques have ushered in a new era for reconstruction following head and neck cancer ablative surgery.7 Microvascular free tissue Correspondence: Dr KJ Harrington, Cancer Research UK Centre for Cell and Molecular Biology, The Institute of Cancer Research, 237 Fulham Road, London SW3 6JB, UK. E-mail:
[email protected] The authors wish to recognise the key contribution made by the late Mr Martin Kelly to these studies. Martin was instrumental in developing our concepts of using flaps for therapeutic purposes in patients with head and neck cancer and we will deeply miss his insight in further developments in this area. Received 28 November 2007; revised 27 June 2008; accepted 30 June 2008; published online 11 September 2008
transfer (of so-called flaps) enables reconstruction of major defects that were irreparable earlier, as well as markedly improving function and cosmesis.8 Thus, surgeons can undertake aggressive surgical resections of large tumours, secure in the knowledge that the resulting tissue defect can be filled with healthy autologous flap tissue harvested from any appropriate part of the body.9 Flap surgery involves transferring tissue from the immediate vicinity of the primary defect (local flaps) or from distant sites (distant flaps).10 Distant tissue transfer involves transplanting donor tissue from one part of the body to another.11 The tissue can either be transferred as a pedicled flap, without cutting and rejoining (anastomosing) the blood vessels, or as a free flap, whereby a block of tissue based on an arterial and venous vascular territory is removed from a donor site and reattached to the distant recipient site, where its circulation is restored by microvascular anastomosis.12 The transferred tissue may contain skin, subcutaneous connective tissue and fascia, muscle, bone as well as the vasculature that supplies it. Factors to consider when choosing a flap include: size and location of the defect; underlying or exposed structures; size of available donor tissue and its pedicle length; resulting donor site defect or disability; viability of surrounding tissue and the shape and contour of the potential reconstruction.13 Tissues selected for transfer are chosen on the basis of type and volume of tissue required to fill the surgical defect suitably, the ease of
Gene therapy by microvascular free tissue transfer VK Agrawal et al
harvest and associated morbidity14,15 and the length of pedicle available13,16 As such, free tissue transfer simply fills a surgical defect and does not contribute directly to therapy. However, transferring fresh, tumour-free tissue from a distant site to a surgical bed that may be contaminated by MRD is a therapeutic opportunity in itself. Once detached from the blood vessels at the donor site, the free tissue flap will remain viable ex vivo for a period of time (up to many hours) that depends on its tissue composition.17 The period between detachment from the donor site to re-anastomosis at the surgical bed represents a window of opportunity for transducing the flap with plasmid or viral vectors that may exert a therapeutic effect on MRD in the surgical defect. For flaps that are commonly used in the clinic, the window of opportunity in which transduction can take place is at least 45–60 min (and may be longer for some flaps). As a first step towards investigating therapeutic free flap transfer in our laboratory, we have studied the relative transduction efficiencies of plasmid and adenoviral vectors in a rat model. Preliminary qualitative data were derived from studies using lacZ-expressing vectors. Thereafter, detailed analysis of the magnitude, topography and duration of gene expression was carried out using a non-invasive technique on the basis of imaging luciferase expression. These data provide a strong rationale for further preclinical development of this approach.
Results Establishment of a rat superficial inferior epigastric artery free flap model A detailed account of the surgical technique is provided in Materials and methods and Figures 1a–f. In brief, under surgical anaesthesia, the flap was demarcated (Figure 1a), raised (Figure 1b) and the vascular pedicle was isolated (Figure 1c) before the flap was removed and transduced ex vivo (Figure 1d) with the appropriate plasmid or viral vector. Thereafter, the flap was reanastomosed to the vasculature (Figure 1e) and the flap was sutured back in place (Figure 1f). Intravascular transduction of SIE flaps with lacZ-expressing plasmid and adenoviral vectors LacZ expression was defined qualitatively at a macroscopic and microscopic level. Representative images for plasmid and adenoviral transductions (with appropriate phosphate buffered saline (PBS) controls) are presented in Figures 2a–h. The macroscopic images show diffuse transduction after intravascular administration of both pIRES-Neo.lacZ (Figure 2b) and Ad.lacZ (Figure 2d) vectors. X-gal staining showed that, if anything, lacZ expression was greatest after plasmid transduction, although this method did not lend itself readily to quantitative assessment. In contrast, control flaps showed little or no evidence of endogenous lacZ expression (Figures 2a and c). Attempts to quantitate the level of transduction and to define its distribution were complicated by heterogeneous staining within individual flaps. However, the representative microscopic images (Figures 2f and h) show that lacZ staining was seen in a predominantly perivascular distribution.
Therefore, to define further the extent of transduction throughout the superficial inferior epigastric (SIE) flap and its duration, we used non-invasive imaging of luciferase.
79
Analysis of total radiance after transduction with plasmid expressing luciferase Superficial inferior epigastric flaps were raised and transduced with 50 mg of the luciferase-expressing plasmid (p.Luc). administered by one of four routes: topical bathing, direct intraflap injection, intravascular injection and intravascular injection plus ultrasound and microbubbles. Animals were imaged at 1 and 7 days post-transduction. At day 1, luciferase expression in all the transduced groups was significantly greater than in untransduced controls (Mack-Skilling’s test, Po0.01).18 Within the transduced groups, variability was seen in the magnitude and pattern of light emission. Luciferase expression was lowest in the bathed group, with low-level light emission limited to the periphery of the flap and notably in the area of the sutures (Figure 3b). Direct intraflap injection was associated with high-level luciferase expression at the site of the injection, but with little spread of signal elsewhere in the flap (Figure 3c). Both groups that received intravascular injections (both with and without microbubble injection and ultrasound) had diffuse luciferase expression throughout the flap (Figures 3d and e). Comparing the groups using the Kruskal– Wallis test a significant difference was seen between the groups (P ¼ 0.032). Using the Mann–Whitney test, both the intravascular and the intravascular plus ultrasound and microbubbles groups demonstrated significantly higher transduction than the bathed group (P ¼ 0.029 in both cases). No other differences were significant. At day 7, luciferase expression was reduced in all experimental groups (Figures 3g–j), suggesting that plasmid transduction was not associated with longevity of gene expression in this model. Nonetheless, there were statistically significant differences in transduction between the various groups. The total radiance in the group that received the direct intraflap injection (Figure 3h) was significantly higher than that in the groups that were transduced by topical bathing (Figure 3g) or intravascular injection with ultrasound and microbubbles (Figure 3j) (Mann–Whitney P-value ¼ 0.029 in both cases). The difference between the group that received the direct intraflap injection and the group that received intravascular injection (Figure 3i) was not significant. Analysis of the distribution of luciferase expression after transduction with p.Luc The distribution of luciferase expression was quantified by measuring the light emission from each voxel on a grid of 96 equal-sized units (Figure 4). To compare the distribution of radiance between groups, the number of voxels above a pre-determined threshold was recorded for each flap and groups were compared using the Kruskal–Wallis test. There was a significant difference (P ¼ 0.026) in the spread of the radiance between the four groups at the 24 h time point (Figures 5a–d). At 24 h, the largest numbers of voxels with radiance above the threshold were recorded for the flaps transduced by intravascular injection with or without ultrasound Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
80
d
a
MVC T Caudal
b
AC
Cranial
Cranial
e
VA
Caudal
c
f
Cranial
AC
MVC
Caudal Figure 1 Photographic images of the surgical technique for generation of superficial inferior epigastric flaps. (a) The skin over the ventral abdominal surface of anaesthetized rats is epilated and the territory of the flap is marked. The additional marks at the corners of the flap and the triangular mark (T) assist with re-orientation when the flap is inset and sutured back into place. The orientation of the cranial (head) and caudal (tail) ends of the animal are denoted. (b) The flap is mobilized—note the marked skin contraction. (c) The vascular pedicle is isolated and the periadventitial fat is removed. After clamping the proximal femoral artery, an arteriotomy is made and an arterial cannula (AC) is introduced. After flushing the vasculature of the flap of residual blood, a microvascular clamp (MVC) is placed on the femoral vein. (d) The flap is placed in a Petri dish and transduced with plasmid or adenoviral vectors (or appropriate controls) administered by a number of different routes. The positions of the arterial cannula (AC) and microvascular clamp (MVC) are indicated. (e) After 45 min, the artery and vein are re-anastomosed. The vascular anastomosis (VA) is indicated. (f) The flap is inset and sutured back into position.
and microbubbles (Figure 6a), although this effect had disappeared by 7 days (P ¼ 0.388) (Figure 6b). Intravascularly transduced flaps demonstrated the most even distribution of transduction, although the site of maximal transduction varied according to the entry point of the pedicle into the flap. The addition of microbubbles and focused ultrasound did not deliver any significant additional transduction, improving neither the quantity nor the distribution of transduction. Direct intraflap injection caused transduction that was centralized around the point of the needle on the lateral aspect of the flap (Figures 3c and 5b). In keeping with the data on the lowest levels of total radiance after topical bathing with p.Luc, this group demonstrated very little evidence of dissemination of transduction within the flap (Figures 5a, 6a and b). Gene Therapy
Analysis of total radiance after transduction with adenovirus expressing luciferase Superficial inferior epigastric flaps were raised and transduced with Ad.Luc. administered by one of three routes: topical bathing; direct intraflap injection or intravascular injection. After reviewing the results with p.Luc, it was decided not to test intravascular injection plus ultrasound and microbubbles with Ad.Luc. Animals were imaged at 1, 2, 3, 7, 14 and 28 days post-transduction. Data were initially analysed using repeated measures analysis of variance. The experimental group to which the animal was assigned (topical bathing, intraflap injection or intravascular injection), whether it had been transduced or not, and the time at which the radiance
Gene therapy by microvascular free tissue transfer VK Agrawal et al
81
a
e
BV
AT
b
f AT
BV
c
g
BV
d
h
Figure 2 Macroscopic and microscopic images of flaps transduced by intravascular injection of lacZ-expressing vectors (or appropriate controls). (a) Macroscopic view of the deep surface of a control flap injected with phosphate-buffered saline (PBS) during the series of experiments with pCMV.lacZ. There is no evidence of positive staining with X-gal. (b) Macroscopic view of the deep surface of a flap injected with p.lacZ demonstrating widespread positive X-gal staining. (c) Macroscopic view of the deep surface of a control flap injected with PBS during the series of experiments with Ad.lacZ. There is no evidence of positive staining with X-gal. (d) Macroscopic view of a flap injected with Ad.lacZ demonstrating diffuse positive X-gal staining. (e) Photomicrograph of flap from (a) that was injected with PBS. There is no positive staining for lacZ. (f) Photomicrograph of flap from (b) that was injected with p.lacZ. Areas of perivascular lacZ staining are indicated (red arrows). (g) Photomicrograph of flap from (c) that was injected with PBS. There is no evidence of positive staining with X-gal. (h) Photomicrograph of flap from (d) that was injected with Ad.lacZ. Areas of perivascular lacZ staining are indicated (red arrows). Adipose tissue (AT) and blood vessels (BV) are indicated. Flaps were harvested 3 days after transduction. Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
82
was measured were the factors affecting the total radiance (interaction P ¼ 0.023). Time had a much greater influence in the intravascular group than in the others (Figures 7c, f, i and 8).
There was a significant difference in total radiance between the three experimental groups at 24 h (Kruskal– Wallis test, P ¼ 0.018) (Figures 7a–c). Total radiance for the intravascular group was significantly higher than
Day 1
Day 7
a
f
b
g
c
h
d
i
e
j
Control
Topical Bathing
Injection Intraflap
Intravascular
Intravascular + US and Microbubbles
Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
that seen in the other two groups (Mann–Whitney, P ¼ 0.029 for each comparison). Total radiance after topical bathing was indistinguishable from untransduced controls (Figure 7a, d and g). At 2 weeks, there was still a significant difference in total radiance between the three transduced groups (Kruskal–Wallis test, P ¼ 0.037) (Figure 9d, e and f) and this difference was maintained at 4 weeks (Kruskal–Wallis test, P ¼ 0.039) (Figure 9g, h and i), with the highest levels of radiance seen in the intravascular group.
distribution of luciferase expression was maintained at 2 weeks (Kruskal–Wallis, P ¼ 0.013) (Figures 9d–f and 10b) but was absent at 4 weeks (Kruskal–Wallis, P ¼ 0.086) (Figures 9g–i and 10c). When the individual groups at 2 weeks were compared, a significant difference was seen between the intravascular and topical bathing groups (Mann–Whitney, P ¼ 0.029), but not between the intraflap and bathed groups. At 4 weeks, there was no significant difference between any of the groups.
Analysis of the distribution of luciferase expression after transduction with Ad.Luc There was a significant difference in the spread of the radiance between the three groups at the 24 h time point (Kruskal–Wallis test, P ¼ 0.009) (Figures 7a–c and 9a–c). Intravascular transduction yielded the largest numbers of voxels with radiance above the threshold (Figures 9c and 10a) and the most even distribution of transduction (Figures 7c, f and i) although, once again, the location of the point of maximal transduction varied according to the entry point of the pedicle. Direct intraflap injection yielded transduction centralized around the point of the needle on the lateral aspect of the flap (Figures 7b, e and h). At 24 h, the largest numbers of voxels with radiance above threshold were recorded for the flaps transduced by intravascular injection (P ¼ 0.009) (Figures 9c and 10a). The significant difference in
Analysis of systemic distribution of adenoviral vector Images from the IVIS camera failed to demonstrate quantifiable luciferase expression beyond the boundaries of the transduced flap tissue (data not shown). However, the sensitivity of such an assessment may have been impaired by tissue attenuation of light emission from more deep-seated organs. Therefore, in line with the methodology used by Michaels et al.,19 we used reverse transcription PCR to look for lacZ expression in normal tissues. Representative data are shown for flap tissue and normal organs (heart, lung, liver, spleen, kidney, testis) harvested 5 days after Ad.lacZ injection (Figure 11). A strong PCR signal was seen from the flap. A much weaker signal was seen from lung tissue in one rat (Figure 11) and from kidney and testis in one rat (data not shown).
Cranial
83
Photons per second
C Lateral
D Medial
B A
D
Caudal B A
Figure 4 Methodology for generation of topographic intensity maps of gene expression across free tissue flaps. Captured images were divided into 96 voxels of equal size and orientated as follows for all analyses: A ¼ caudal/medial; B ¼ caudal/lateral; C ¼ cranial/lateral and D ¼ cranial/medial (left hand image). Individual measures from each of the voxels from a group of animals were then used to construct topographic intensity maps in the form of so-called ‘Manhattan plots’ (the right hand image). Note the manhattan plot indicates average data from four animals.
Figure 3 The IVIS camera images taken at days 1 and 7 of representative rats with flaps that were transduced with p.Luc (or phosphatebuffered saline (PBS) control) administered by topical bathing, direct intraflap injection, intravascular injection or intravascular injection plus microbubbles and ultrasound. (a) PBS control (through intravascular route) at day 1. PBS controls for each of the other routes gave similar negative images (data not shown). (b) p.Luc by topical bathing at day 1 demonstrating luciferase expression limited to the sites of re-suturing of the flap (red arrow). (c) p.Luc by intraflap injection at day 1 showing high-level luciferase expression at the injection site (white arrow) and the lower level expression at the suture sites (red arrow). (d) p.Luc by intravascular injection at day 1 demonstrating high-level, diffuse luciferase expression (white arrow). (e) p.Luc by intravascular injection plus microbubbles and ultrasound at day 1 showing high-level, diffuse luciferase expression (white arrow). Note that the apparent extension of luciferase expression beyond the margins of the flap is probably because of anatomical displacement of the subcutaneous portion of the flap. (f) PBS control (through intravascular route) at day 7. PBS controls for each of the other routes gave similar negative images (data not shown). (g) p.Luc by topical bathing at day 7. (h) p.Luc by intraflap injection at day 7. (i) p.Luc by intravascular injection at day 7. (j) p.Luc by intravascular injection plus microbubbles and ultrasound at day 7. Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
84
a
b
25 20 15 10
Cranial
5
35 30 25 20 15 10
Cranial
5
0 Lateral
0 Lateral Caudal
Caudal
Medial
c
Photons per second (x103)
30
Photons per second (x103)
35
Medial
d 35
20 15 10
Cranial
5
35 30 25 20 15 10
Cranial
5
0 Lateral Caudal
Medial
0
Photons per second (x103)
25
Photons per second (x103)
30
Lateral Caudal
Medial
Figure 5 Manhattan plots of luciferase expression at 24 h from free tissue flaps transduced with p.Luc. (a) Topical bathing. (b) Direct intraflap injection. (c) Intravascular injection. (d) Intravascular injection plus ultrasound and microbubbles. A significant difference was seen between the groups (Kruskal–Wallis’ test, P ¼ 0.032). Both the intravascular and the intravascular plus ultrasound and microbubbles groups demonstrated significantly higher transduction than the bathed group (Mann–Whitney test, P ¼ 0.029 in both cases).
Discussion Free tissue transfer is a routine part of surgical cancer management, especially in head and neck, breast and gynaecological tumours. In a recent clinical study in 72 patients, we demonstrated that free tissue transfer opens up new therapeutic options for patients with relapsed head and neck cancer.20 During wide-field ablative surgery for recurrent cervical lymphadenopathy in a previously irradiated subcutaneous tissue and skin area, transferring fresh autologous muscle, subcutaneous and skin allowed us to re-irradiate the tumour bed to 60 Gy with 192Ir-brachytherapy. In-field control rates of 40–58% at 5 years were achieved without significant normal tissue damage. In this approach, the free tissue flap was a facilitator of further treatment, rather than the direct mediator of the therapeutic effect. In this paper, we describe a means of refining the use of flaps to endow transferred tissue with the potential to contribute directly to an anticancer therapy. As such, free tissue flaps that express therapeutic genes potentially represent a powerful tool. Gene therapy approaches that may act against MRD in a post-surgical tumour bed Gene Therapy
include virally directed enzyme prodrug therapy, immunomodulatory therapy and radionuclide uptake therapy.21 Virally directed enzyme prodrug therapy is an attractive option mediated by generation of activated prodrug metabolites in normal cells at the flap/tumour bed interface. These agents would have the potential of diffusing to areas containing MRD. The use of S-phase specific systems, such as HSV thymidine kinase/ganciclovir or cytosine deaminase/5-fluorocytosine, would result in antitumour specificity because flap tissues would not be actively dividing. Similarly, expression of immunomodulatory cytokines capable of concentrating effector T cells and/or antigen-presenting cells at the tumour bed would be an interesting therapeutic strategy. The use of NIS-mediated radioisotopic irradiation (131I, 186 Re, 188Re or 211At)22,23 of the tumour bed has direct parallels with our recent experience with 192Ir-brachytherapy in patients, but with the potential advantages of more homogeneous dose delivery, the use of fractionated doses and the ability to avoid implanting radioactive wires into the tumour bed. Before undertaking therapeutic studies, it is necessary to evaluate the effect of different vectors and routes
Gene therapy by microvascular free tissue transfer VK Agrawal et al
85 100
No of Voxels above Threshold
80
60
40
20
0
80
60
40
20
Group
No. of voxels above threshold
Group
Intravascular, US and microbubbles
Intravascular
Intraflap injection
Intravascular, US and microbubbles
Intravascular
Intraflap injection
Bathed
0 Bathed
No of Voxels above Threshold
100
No. of voxels above threshold
Bathed
1
0
4
6
Bathed
Injected
2
15
44
11
Injected
2
16
0
17
Intravascular
55
35
18
23
Intravascular
8
0
1
0
Intravascular, US, microbubbles
32
27
23
53
Intravascular, US, microbubbles
1
0
0
0
p = 0.026 (Kruskal-Wallis)
0
0
20
0
p = 0.388 (Kruskal-Wallis)
Figure 6 Distribution of luciferase expression after transduction with p.Luc as a function of the number of voxels with light emission above a threshold of 141706.2 p s1 cm2. (a) The spread of radiance was significantly different between the four groups at 24 h, with the largest numbers of voxels exceeding threshold in the flaps transduced by intravascular injection with or without ultrasound and microbubbles (Kruskal–Wallis test, P ¼ 0.026). (b) At 7 days, there was no significant difference between the flaps (Kruskal–Wallis test, P ¼ 0.388).
of administration on the magnitude, topography and duration of gene expression. In this paper, we demonstrate that non-invasive bioluminescent imaging can be used to evaluate these parameters in a rat SIE artery free flap model. These studies confirm data generated in previous work in quadratus femoris and SIE flap models, which demonstrated efficient gene delivery by the intravascular route.19,24 That work examined the effect of a number of variables, including intravascular and intraparenchymal routes of injection, the effect of supraphysiological perfusion pressure, dwell time, viral concentrations and temperature on transduction and formed the initial starting point for our examination of the system. The SIE flap has a similar composition to flaps used in the clinic and also possesses a number of additional advantages. As the flap is located superficially, surgical access is relatively easy allowing the flap to be raised comparatively quickly. Additionally, following vascular re-anastomosis, the skin paddle facilitates reliable post-operative monitoring of the vascular sufficiency of the flap. The flap is based on the femoral vessels whose length and calibre both confer advantages. The calibre of the pedicle, with a diameter of approximately 1 mm, allows reliable re-anastomosis, and the long pedicle permits division of the vessels, segmental excision and re-anastomosis without flap compromise.
As the muscular component is very thin, it has a higher ischaemic tolerance than a pure muscle flap, such as the quadratus femoris flap.25 This carries significance as our experimental design required an ex vivo period of 45 min (a time that is achievable in the clinic). Initial studies demonstrated that intravascular delivery of lacZ-expressing plasmid and adenoviral vectors yielded transduction of a rat SIE flap. These data are in line with those of a previous study, which demonstrated that intravascular delivery of Ad.lacZ in SIE flaps transduced all cell types within the flap and was strictly limited to the flap.19 Our analyses also showed lacZ expression, but highlighted the fact that this methodology was not well-suited to detailed quantitative analysis because of the patchy nature of the lacZ staining and the difficulties in obtaining representative three-dimensional reconstructions of levels of gene expression throughout the flap. Another shortcoming of lacZ staining was the fact that each rat could only be studied once after necropsy at a particular time point. Similar limitations exist for the use of reverse transcription-PCR or enzyme-linked immunosorbent assay-based analyses. Therefore, we chose to assess gene expression non-invasively by serial bioluminescent measurements, using the widely studied methodology of an IVIS camera. Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
86
Topical Bathing
Intraflap Injection
Intravascular
b
c
d
e
f
g
h
i
Day 28
Day 14
Day 1
a
Figure 7 The IVIS camera images taken at days 1, 14 and 28 of representative rats with flaps that were transduced with Ad.Luc administered by topical bathing, direct intraflap injection or intravascular injection. Note healing of the flaps and regrowth of hair over the 4-week period of study. Images from rats in each group treated with phos[phate-buffered saline (PBS) showed no evidence of luciferase expression (data not shown). (a, d and g) Topical bathing group imaged at days 1, 14 and 28, respectively. (b, e and h) Direct intraflap injection group imaged at days 1, 14 and 28, respectively. High-level luciferase expression is seen focally at the site of injection (white arrows) and persists over 28 days. (c, f and i) Intravascular injection group imaged at days 1, 14 and 28, respectively. Diffuse high-level luciferase expression is seen throughout the flap at days 1 and 14 (red arrows) and adjacent to the site of the vascular pedicle at day 28 (white arrow).
10 Topical Bathing
400
Intraflap Injection Intravascular
300 200 100 0
Estimated Marginal Means (x10 3)
Estimated Marginal Means (x10 3)
500
Topical Bathing Intraflap Injection Intravascular
8
6
4
2 Day 1
Day Day Week Week Week 2 3 1 2 4 Time after transduction
Day 1
Day Day Week Week Week 2 3 1 2 4 Time after transduction
Figure 8 Time course for estimated marginal means of luciferase expression after transduction with Ad.Luc (a) or phosphate-buffered saline (PBS) controls (b). Note 50-fold difference between the scale of the Y axis in (a) and (b) and the time-dependency of luciferase expression after Ad.Luc transduction.
In the first instance, gene expression was measured after administration of vectors expressing firefly luciferase by a number of different routes. This initial evaluation aimed to assess the value of intravascular administration compared to more direct applications of the vector (intraflap injection or topical bathing). These studies represent the first attempt to compare intravascular with other routes of vector delivery. For both Gene Therapy
plasmid and adenoviral vectors, intravascular administration yielded greater total levels of luciferase expression. This observation is important because the more direct approaches to vector administration would certainly be easier to use in a clinical scenario and would not require specialist microsurgical techniques. Nonetheless, it appears that intra-arterial infusion of vectors is superior to other routes and provides a strong rationale
Gene therapy by microvascular free tissue transfer VK Agrawal et al
87 Topical Bathing
Intraflap Injection
Day 1
Day 28
Photons pe r second (x 3) 10
Photons per seco nd (x103)
35 30 25 20 15 10 5 0
35 30 25 20 15 10 5 0
Photons per second (x1 3 0 )
Photons per second (x103 )
Photons per seco nd (x103)
35 30 25 20 15 10 5 0
35 30 25 20 15 10 5 0
35 30 25 20 15 10 5 0
35 30 25 20 15 10 5 0
35 30 25 20 15 10 5 0
Photons per seco nd (x103)
Day 14
15 10 5 0
35 30 25 20 15 10 5 0
Photons per second (x104 )
25 20
Photons per second (x103)
Photons per seco nd (x103)
35 30
Intravascular
Figure 9 Manhattan plots of luciferase expression at 24 h, 14 and 28 days from free tissue flaps transduced with Ad.Luc.). (a, d and g) Topical bathing group imaged at days 1, 14 and 28, respectively. (b, e and h) Direct intraflap injection group imaged at days 1, 14 and 28, respectively. High-level luciferase expression is seen focally at the site of injection and persists over 28 days. (c, f and i) Intravascular injection group imaged at days 1, 14 and 28, respectively. Note x axis in 9C is photons per second ( 104). A significant difference in total radiance was seen between the groups (Kruskal–Wallis test, P ¼ 0.037). The greatest level of total radiance and the most diffuse pattern of gene expression were seen in the intravascular group.
for developing this approach in the setting of free flap transfer (where surgical access to the arterial supply is required). These findings mirror those of Michaels et al.19 In terms of therapeutic importance, it is arguable that the topography of gene transduction, rather than its absolute level, will be a key factor. This issue can be considered in terms of the anatomical proximity between a transduced area in the flap and the underlying tumour bed harbouring MRD. For cytotoxic gene therapy approaches, such as virally directed enzyme prodrug therapy or radionuclide uptake therapy, a therapeutic effect mediated by transduction of normal tissues within a flap will be entirely reliant on the bystander effect. Therefore, the activated prodrugs generated by the flap will need to diffuse to foci of MRD and radioisotopes taken up in the flap will need to decay and release b-particles with sufficient path length to deliver radiation to the tumour bed. Success will be more likely if expression of the therapeutic gene occurs adjacent to
MRD. Therefore, a diffuse pattern of moderate gene expression is probably more desirable than a very high level of expression in a small part of the flap. In this regard, intravascular administration of plasmid and viral vectors yielded the greatest spread of gene expression. Topical bathing of the deep surface of the flap initially seemed an attractive option given the fact that, after transfer to the surgical bed, this surface lies directly over the area at greatest risk of harbouring MRD. Unfortunately, the poor levels of transduction achieved with this method do not support its further use. Similarly, the lack of diffusion of vector from the site of direct intraflap injections argues strongly against the further development of this approach. The duration of gene expression is another key variable in this system. In clinical settings, it will be necessary for patients to recover from the surgical procedure before any attempt is made to exploit the therapeutic potential of the transduced flap tissue. Our Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
88 100
40 20
60 40 20
40 20
Bathed
Intravascular
Intraflap injection
Bathed
Intravascular
Intraflap injection
Bathed
60
0
0
0
80
Intravascular
60
80
Intraflap injection
80
100 No of Voxels above Threshold
No of Voxels above Threshold
No of Voxels above Threshold
100
Group
No. of voxels above threshold
Group
No. of voxels above threshold
Group
Bathed
0
0
0
0
Bathed
0
0
0
0
Bathed
0
0
0
0
Injected
31
25
16
0
Injected
0
32
13
6
Injected
0
15
0
0
Intravascular
96
91
96
87
Intravascular
60
24
57
58
Intravascular
17
0
37
6
p = 0.009 (Kruskal-Wallis)
p = 0.013 (Kruskal-Wallis)
No. of voxels above threshold
p = 0.086 (Kruskal-Wallis)
Testes
Spleen Kidney
Liver
Lung
Heart
Flap
H2O control
Plasmid DNA
RNA +ve control
DNA ladder
Figure 10 Distribution of luciferase expression after transduction with Ad.Luc as a function of the number of voxels with light emission above a threshold of 141706.2 p s1 cm2. (a) The spread of radiance was significantly different between the three groups at 24 h, with the largest numbers of voxels exceeding threshold in the flaps transduced by intravascular injection (Kruskal–Wallis test, P ¼ 0.009). (b) At 14 days, there was still a significant difference between the groups (Kruskal–Wallis test, P ¼ 0.013). (c) At 28 days, there was no significant difference in the spread of luciferase expression between the groups (Kruskal–Wallis test, P ¼ 0.086).
Figure 11 Reverse transcriptase PCR from flap and normal organs in a representative rat with an superficial inferior epigastric flap transduced with Ad.lacZ. Positive signal was seen from flap and to a lesser extent from lung tissue. RNA positive control represents RNA harvested from 293A cells infected with Ad.lacZ. Plasmid DNA represents 20 ng pIRES-Neo.lacZ.
previous experience with the use of free tissue flaps in the head and neck region showed that a post-surgical recovery time of 5–7 days was required.20 Therefore, maintenance of gene expression at and beyond this time point will be necessary if therapeutic benefit is to be derived. The data for plasmid administration suggest that gene expression is short-lived with almost complete silencing after 7 days, irrespective of the route of administration. For intravascular adenoviral delivery, gene expression also decreased over time but was still maintained at levels significantly above controls at 2 weeks. By using IVIS imaging, we were able to track gene expression in flaps in individual animals sequentially over time, without subjecting them to necropsy. These data provide clear evidence that this approach has potential clinical utility, and ongoing studies in our Gene Therapy
laboratory are assessing the efficacy of intravascular adenovirus-mediated therapeutic gene delivery in a model of MRD. An important component of the development of the strategy of using free tissue flaps will be evaluation of specific interventions to improve the efficiency of the transduction process. Our studies involving ultrasound aimed to improve flap transduction without jeopardising its viability. Unfortunately, although ultrasound was well tolerated, it did not improve the magnitude, topography or duration of gene expression in the plasmid experiments, and its use was not pursued with adenovirus. Further experiments will now focus on such variables as the effect of concomitant recombinant VEGF26 and vascular permeabilizing cytokines and physical factors such as variation in the temperature, dose and pressure of the perfusate. We will also consider the effect of dwelltime with these different scenarios and compare the data with previous reports.19,24
Materials and methods Plasmid vectors A lacZ expression plasmid was constructed as follows. LacZ cDNA was amplified by PCR with primers GCTCAGGATCCTTATTTTTGACACCAGACCAACTGG and ATAAGATTGCGGCCGCATGGTCGTTTTACAACG TCGTGACTGG that contained unique BamH1 and Not1 (both from New England Biolabs (UK) Ltd.) restriction enzyme sites, respectively (underlined). The PCR product was digested with BamHI and NotI and gel purified using the QIAquick gel purification kit (Qiagen, Crawley, UK). Plasmid pIRES-Neo (Clontech-Takara Bio Europe, France) was also digested with NotI and BamHI and the linearized plasmid DNA gel was purified. The
Gene therapy by microvascular free tissue transfer VK Agrawal et al
purified PCR and backbone plasmid products were ligated (T4 DNA ligase, Invitrogen Ltd, Paisley, UK) and transformed into E. coli TOP10 (Invitrogen). pIRESNeo.lacZ plasmid DNA was extracted and assessed for the presence of the lacZ gene by PCR and by transfection into HCT116 tumour cells followed by staining for b-galactosidase expression. Positive clones were grown up and DNA extracted (Maxi Prep, Qiagen) for subsequent in vivo analyses. Plasmid p.Luc was provided as a gift by Dr M Smalley (Breakthrough for Breast Cancer Centre, The Institute of Cancer Research). In this plasmid, firefly luciferase (Luc) is expressed under the control of the cytomegalovirus promoter. It was constructed by insertion of the Luc open reading frame from pGL3basic (Promega UK, Southampton, UK) between the NheI and XbaI restriction sites of vector pcDNA3.1+ (Invitrogen). Chemically competent DH5a´ E. coli (Invitrogen) were transformed and resultant cultures were grown in the presence of 100 mg ml1 ampicillin. pCMV-Luc was extracted and purified using a Midiprep Kit (Qiagen). Plasmid concentrations (typically 0.5 mg ml1) and purity (ratio of optical densities at 260 and 280 nm of 1.7–1.9) were evaluated using spectrophotometry.
Adenoviral vectors Ad.lacZ (Adeno-X-LacZ) in which b-galactosidase was driven by a human cytomegalovirus immediate/early promoter was purchased from Clontech (Mountain View, CA, USA) at a titre of 1010 plaque-forming units (PFU) per ml. Ad.Luc contained a luciferase expression cassette driven by the Rous sarcoma virus promoter that was produced in 293 cells. Titration of the virus stock to determine the number of infectious units (IU) per ml was carried out by plaque assay on 293 cells. All viral stocks were stored in single use aliquots at 80 1C until needed. b-galactosidase staining Cell monolayers were washed with PBS and then fixed with formaldehyde (BDH Laboratory Supplies, Poole, UK), 0.2%. glutaraldehyde (Sigma-Aldrich, Poole, UK) for 5 min at room temperature. The fixed cell monolayers were then washed twice with PBS and stained with 1 mg ml1 X-gal (5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside) (Merck Biosciences, Darmstadt, Germany), 5 mM ferricyanide (Sigma-Aldrich Ltd, Poole, UK), 5 mM ferrocyanide (Sigma), 2 mM MgCl2 (BDH Laboratory Supplies, Poole, UK), 0.02% (v/v) NP40 (Igepal CA-630, Sigma), 0.01% (v/v) SDS (BDH Laboratory Supplies) in PBS for 8–4 h in a 37 1C humidified incubator. Animals Adult male Fisher 344 rats (Harlan, Bicester, UK) weighing between 150 and 250 g were used. Rats were maintained under specific pathogen-free conditions in sterile filter-top cages on sterile bedding, and fed an irradiated diet of standard rodent chow and autoclaved, acidified water (pH 2.8) ad libitum. All animal procedures were performed under licence from the Home Office and were approved by the institutional ethics review board.
89
Surgical technique to generate SIE flaps The adipo-fascio-myocutaneous SIE flap as first described by Strauch and Murray27 and modified by Nishikawa et al.28 was selected for these studies. Surgical anaesthesia was induced and maintained with isoflurane (Abbott Laboratories, Queenborough, UK). The ventral inferior skin was depilated with Veet cream (Veet, Clevedon, Eire), cleaned and the flap margins were marked on the skin (Figure 1a). The flap was raised with a scalpel and surgical blade (Swann-Morton, Sheffield, UK) (Figure 1b). Using an operating microscope (Zeiss, Welwyn Garden City, UK) and surgical instruments (Vessel Dilators/Forceps D-5a.3, Microscissors SAS-15, Needle Holder: B-15-8.3, Microclamps: B-3 00400 V, HD-S (S&T, Neuhausen, Switzerland)), the pedicle was isolated and all periadventitial fat removed. Particular care was taken to avoid handling the origin of the SIE artery. The surgical site and vessels were bathed in 2% lidocaine (Hameln Pharmaceuticals, Gloucester, UK) to provide local anaesthesia and to relieve arterial spasm. Following clamping of the proximal femoral artery with a microvascular clamp size B-3 (S&T), an arteriotomy was made to permit insertion of a 30 gauge ophthalmic cannula (Altomed, Boldon, UK). The cannula was secured with a double-tied 10/0 W2850 Ethilon suture (Ethicon, Edinburgh, UK) and the flap flushed with 1 ml of Dulbecco’s PBS to ensure that the vasculature was cleared of residual blood. The femoral vein was then clamped and both the artery and vein transected midway along their length (Figure 1c). The flap was removed from the body with the cannula in situ and placed in a Petri dish (Nunc, Roskilde, Denmark). A vascular clamp was applied to the vein (Figure 1d). Flaps were transduced as described below and then the artery and vein were re-anastomosed using 10/0 nylon (Figure 1e) and the flap was inset and sutured into position with interrupted 3/0 W2511T silk sutures (Ethicon) (Figure 1f). The rats were then placed in a protective soft collar made from cage liner and taped together by autoclave tape (Westfield Medical, Radstock, UK) and allowed to recover. Pre-, per- and post-operative analgesic drugs were administered to minimise the morbidity of the surgical procedure. Animals were given buprenorphine (Schering-Plough, Harefield, UK) 0.1 ml per 100 g weight preoperatively, in the immediate post-operative period and 12 hourly for 5 days after surgery. In addition, they were given carpofen (Pfizer, Tadworth, UK) 0.1 ml per 100 g of a dilution of 0.5 ml in 9 ml sterile water immediately post-operatively. Transduction of tissue flaps Flaps were transduced ex vivo in sterile Petri dishes (Nunc) (Figure 1d). A 1 ml syringe (Terumo, Leuven, Belgium) fitted with a 12G needle (Terumo) was primed with PBS and used to draw up the appropriate experimental preparation (plasmid DNA (pIRES-Neo. LacZ or p.Luc)—50 mg in 500 ml PBS; adenovirus (Ad.lacZ or Ad.Luc)—1 108 PFU in 500 ml PBS)). Following mixing by gentle inversion, the solution was delivered to the flap by a number of different routes, as follows: (i) bathed flaps were transduced by adding the solution containing the plasmid or adenovirus drop-wise onto the deep surface of the flap in the Petri dish. The Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
90
flap was coated with the solution and then placed with the deep surface face down in the covered dish to minimise evaporative losses; (ii) injected flaps were transduced by direct introduction of a 12G needle in to the centre of the flap using a lateral approach to minimise the risk of pedicle damage. The solution was deposited under the skin surface and the flap placed in a Petri dish and covered; (iii) intravascular gene delivery was performed as an intra-arterial injection with the vein cross-clamped at the pedicle using a B-3 microvascular clamp (S&T). To maximise efficiency of intravascular transfer, a further 500 ml of PBS was delivered using the same needle and syringe. This had the effect of ensuring that the entire vascular tree of the flap was perfused; (iv) intravascular transduction with plasmid DNA in association with microbubbles and ultrasound delivery was performed, as described in the section on Ultrasound Treatment (see below). All four routes of administration were tested with the plasmid DNA, but for adenoviral gene delivery the last experimental group (intravascular plus microbubbles and ultrasound) was omitted. In all experiments, control flaps were treated in an identical manner (surgical technique, vascular cannulation, ex vivo time separated from vasculature) to the experimental flaps, with the exception that the transduction solution contained only PBS. For the pIRES-Neo.LacZ and Ad.lacZ experiments, groups of three rats were transduced for each condition. For the p.Luc and Ad.Luc experiments, groups of four rats were transduced for each condition. Following an incubation period of 45 min at room temperature, the cannula was used to flush through 1 ml of PBS and the surface of the flap was gently irrigated. Thereafter, the flap was reanastomosed to the vasculature, inset and sutured as described above.
Ultrasonic exposure of the SIE flap Assembly of the apparatus used to generate the ultrasound field has been described previously.29–31 Briefly, signal from a waveform generator (model 33220A, Agilent Technologies Inc., Loveland, CO, USA) was fed into an RF power amplifier (58 dB gain, model BT00500 BETA, Tomco Technologies, Norwood, Australia) and thence to a custom-made 1 MHz ultrasound transducer (Imasonics, Besancon, France). The transducer was spherically focused, of 20 mm diameter and 68% bandwidth, with the focal point of the ultrasound beam 67 mm from the transducer face. Dosimetry of the 1 MHz transducer in its focal plane has been previously described.29,30 To expose the tissue flap at the focal distance, a cylindrical perspex ‘stand-off’ was made to fit around the transducer: a layer of cling-film was used to retain a column of water in this tube through which the ultrasound was transmitted. The internal diameter of 38 mm was such that there was negligible disturbance of the acoustic field at the focal distance (that is, at the end of the tube). Acoustic contact gel was used to provide good transmission into the tissue flap, which was placed (again with contact gel) on a 14 mm-thick polyurethanebased anechoic absorber (National Physical Laboratory, Teddington, UK). The latter was used to prevent reflections and standing wave generation leading to unquantifiable exposure conditions; the effectiveness of Gene Therapy
this material in eradicating reflections has been demonstrated.29,30 Before exposure of the tissue flap to ultrasound, a total volume of 0.4 ml of 0.9% (w/v) sodium chloride containing 50 mg p.Luc and 4% (v/v) SonoVue microbubbles (Bracco, High Wycombe, UK) was injected into the tissue flap through the cannulated artery. A further 0.6 ml 0.9% (w/v) sodium chloride was injected and the artery was clamped. SonoVue is an ultrasound contrast agent consisting of phospholipid-encapsulated microbubbles (1–12 mm) filled with SF6 gas.30,31 SonoVue was prepared according to the manufacturer’s protocol to yield a solution containing 2 108–5 108 microbubbles per ml. The tissue flap was then exposed to pulsed 1 MHz ultrasound at the transducer focal distance for 10 s, followed by an interval of 10 s and a second 10 s exposure period. The pulse repetition frequency was 1 kHz, each pulse consisting of 40 cycles of pressure amplitude 1.4 MPa (peak-to-peak) at the transducer focus. Either side of the focus the amplitude falls, with a peak width at half-height of 8 mm, over which region transfection efficiency in vitro is maximal. As the tissue flap was slightly larger than this region, the transducer was ‘scanned’ during the exposure periods for uniform exposure.
Immunohistochemistry for Lac Z Flaps were harvested, rinsed with PBS and immersed immediately in fixing solution (1% formaldehyde, 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM EGTA, 0.02% NP-40 made up to volume of 200 ml with PBS) overnight at room temperature. They were then washed thrice in a solution of PBS/0.02% NP-40 at room temperature for 5 min at a time on a rocking platform. The flaps were then transferred to X-Gal staining solution (5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 2 mM MgCl2, 0.01% sodium deoxycholate, 0.02% NP-40, 1 mg ml1 X-Gal) made up to a volume of 200 ml with PBS and incubated at 37 1C overnight. Following staining, the flaps were washed thrice for 5 min with PBS at room temperature on a rocking platform. The flaps were then post-fixed in 4% paraformaldehyde solution (200 ml of deionized water, 2.68 g sodium cacodylate (SigmaAldrich), 1.9 g sodium chloride, 11.8 g paraformaldehyde (Fisher, Loughborough, UK)) at 4 1C overnight. Tissue was then washed with PBS thrice for 10 min each and serially dehydrated in 10, 15 and 20% sucrose in PBS. The flaps were then incubated overnight at 4 1C in a 1:1 mix of 20% sucrose:OCT (Raymond A Lamb, Eastbourne, UK). They were then embedded in OCT embedding medium and stored at 80 1C. The frozen section blocks were then sectioned on a cryostat (Leica CM 3050, Wetzlar, Germany) for microscopic examination. In vivo luciferase imaging protocol To quantitate Luc gene expression, an IVIS system (Xenogen, Cranbury, New Jersey) was used to measure radiance and data were analysed using Living Image 2.60.1 software (Xenogen). Rats were induced and maintained with isoflurane anaesthesia and an intraperitoneal injection of 15 mg kg1 body weight D-luciferin firefly potassium salt (Xenogen) was administered. After 5 min, the animal was imaged twice at exposures lasting 5 min each with the following settings: sensitivity— large/high sensitivity; f/stop—1; emission filter—open;
Gene therapy by microvascular free tissue transfer VK Agrawal et al
photo—0.2, auto; median—8; overlay; field of view d— 12 cm and subject height 1.5 cm. Rats treated with p.Luc were imaged on days 1 and 7 and rats treated with Ad.Luc were imaged on days 1, 2, 3, 7, 14 and 28.
Analysis of intensity and distribution of luciferase expression To assess the intensity and distribution of luciferase expression within the flap, captured images were divided into 96 voxels of equal size. Total radiance (as a measure of the total transduction of the flap) was calculated by the Living Image software as an integrated function of all 96 voxels. Individual measures from each of the voxels were used to construct topographic intensity maps for the flaps transduced by the different vectors and the different routes of administration. Analysis of systemic distribution of adenoviral vector To assess systemic leakage of adenoviral vectors, flaps were transduced with Ad.lacZ. in three rats according to the standard protocol and the animals were killed at 5 days. The flaps and normal organs (heart, lung, liver, spleen, kidney, testis) were harvested, snap frozen and stored at 80 1C. RNA was extracted using Qiagen RNeasy (Qiagen), and 60 ng from each sample was assayed directly by reverse transcriptase PCR) using the OneStep RT-PCR Enzyme Mix Kit (Qiagen). Primers for lacZ were used: forward, 50 GATCAAATCTGTCGATCCT TCC and reverse, CAAAGACCAGACCGTTCATACA. PCR reaction conditions were: 50 1C for 30 min (for reverse transcription); 95 1C for 15 min; and then 40 cycles of 94 1C for 30 s, 55 1C for 30 s, 72 1C for 45 s; followed by 72 1C for 10 min. Following the PCR, 5 ml of the reaction product was run on a 2% agarose gel. Statistical analysis Data were analysed using Microsoft Excel and the Statistical Program for Social Sciences (SPSS 2006 SPSS Inc. Chicago, IL, USA). The total radiance produced by the different transduced groups was compared by using Kruskal–Wallis’ test, a non-parametric version of the one-way analysis of variance. When a significant difference was found between the groups, separate pairs were compared with Mann–Whitney’s test and a P-value less than 5% was considered significant. With only four rats per group and using Mann–Whitney’s test, the lowest possible P-value we can obtain is 0.029, which corresponds to all the rats in one group having lower radiance to the rats in the other group. Any other outcome would result in a P-value greater than 5%. The total radiance of the transduced groups was compared against the controls by using Mack–Skilling’s test,18 a non-parametric alternative to the two-way analysis of variance. For the Ad.Luc data, repeated measures analysis of variance was used to take into account the effect of time. We quantified the spread of the transduction within a flap by determining the number of voxels that were above a threshold value. This threshold (141706.2 p s1 cm2) was derived by Living Image using an automatic function on the first rat that was imaged and remained the same for each rat. The experimental groups were then compared, as above, using Kruskal–Wallis’ test.
91
References 1 Al-Sarraf M. Treatment of locally advanced head and neck cancer: historical and critical review. Cancer Control 2002; 9: 387–399. 2 Capote A, Escorial V, Munoz Guerra MF, Rodriguez-Campo FJ, Gamallo C, Naval L. Elective neck dissection in early-stage oral squamous cell carcinoma—does it influence recurrence and survival? Head and Neck 2006; 29: 3–11. 3 Ambrosch P, Freudenberg L, Kron M, Steiner W. Selective neck dissection in the management of squamous cell carcinoma of the upper digestive tract. Eur Arch Otorhinolaryngol 1996; 253: 329–335. 4 Zbaren P, Nuyens M, Caversaccio M, Stauffer E. Elective neck dissection for carcinomas of the oral cavity: occult metastases, neck recurrences and adjuvant treatment of pathologically positive necks. Am J Surg 2006; 191: 756–760. 5 El-Marakby HH. The reliability of pectoralis major myocutaneous flap in head and neck reconstruction. J Egypt Natl Canc Inst 2006; 18: 41–50. 6 Issing PR, Kempf HG, Heppt W, Schonermark M, Lenarz T. Reconstructive surgery in the head—neck area with regional and free tissue transfer. Laryngorhinootolgie 1996; 75: 183–192. 7 Rinaldo A, Shaha AR, Wei WI, Silver CE, Ferlito A. Microvascular free flaps: a major advance in head and neck reconstruction. Acta Otolaryngol 2002; 122: 779–784. 8 Blackwell KE. Unsurpassed reliability of free flaps for head and neck reconstruction. Arch Otolaryngol Head Neck Surg 1999; 125: 295–299. 9 Jones NF, Hardesty RA, Swartz WM, Ramasastry SS, Heckler FR, Newton Ed. Extensive and complex defects of the scalp, middle third of the face and palate: the role of microvascular reconstruction. Plast Reconstr Surg 1988; 82: 937–952. 10 McGregor AD, McGreogor IA. Fundamental Techniques of Plastic Surgery and their Surgical Applications, 10th edn. Churchill Livingstone, 2000, pp 61–75. 11 Disa JJ, Santamaria E, Cordeiro PG. General principles of reconstructive surgery for head and neck cancer. In: Shah JP (ed). Cancer of the Head and Neck, Chapter 18. BC Decker Inc.: Lewiston, NY, USA, 2001, pp 330–357. 12 Jacobson JM, Suarez El. Microsurgery in anastomosis of small vessels. Surg Forum 1960; 11: 243. 13 Mathes SJ, Nahai F. Reconstructive Surgery, Principles, Anatomy and Technique. Vol. 1 Quality Medical Publishing, 1995, pp 30–55. 14 Schusterman MA, Horndeski G. Analysis of the morbidity associated with immediate microvascular reconstruction in head and neck cancer patients. Head Neck 1991; 13: 51–55. 15 Schusterman MA, Miller MJ, Rece GP, Kroll S, Marchi M, Goepfert H. A single centre’s experience with 308 free flaps for repair of head and neck cancer defects. Plast Reconstr Surg 1994; 93: 472–478. 16 Eckardt A, Fokas K. Microsurgical reconstruction in the head and neck region: an 18 year experience with 500 consecutive cases. J Craniomaxillofac Surg 2003; 31: 197–201. 17 Sloan GM, Reinisch JF. Flap physiology and the prediction of flap viability. Hand Clin 1985; 1: 609–619. 18 Mack GA, Skillings JH. A Friedman-type rank test for main effects in a two factor ANOVA. J Amer Stat Assoc 1980; 75: 947–951. 19 Michaels J, Levine JP, Hazen A, Ceradini DJ, Galiano RD, Soltanian H et al. Biologic brachytherapy: Ex vivo transduction of microvascular beds for efficient, targeted gene therapy. Plast Reconstr Surg 2006; 118: 54–65. 20 Nutting CM, Horlock N, A’Hern R, Searle A, Henk JM, Rhys Evans P et al. Manually afterloaded 192Ir low-dose rate brachytherapy after subtotal excision and flap reconstruction of recurrent cervical lymphadenopathy from head and neck cancer. Radiother Oncol 2006; 80: 39–42. 21 Harrington KJ, Melcher AA, Bateman AR, Ahmed A, Vile RG. Cancer gene therapy: Part 2. Candidate transgenes Gene Therapy
Gene therapy by microvascular free tissue transfer VK Agrawal et al
92 22
23
24
25
Gene Therapy
and their clinical development. Clin Oncol (R Coll Radiol) 2002; 14: 148–169. Spitzweg C, Harrington KJ, Pinke LA, Vile RG, Morris JC. Clinical review 132: The sodium iodide symporter and its potential role in cancer therapy. J Clin Endocrinol Metab 2001; 86: 3327–3335. Willhauck MJ, Sharif Samani BR, Gildehaus FJ, Wolf I, Senekowitsch-Schmidtke R, Stark HJ et al. Application of 188Rhenium as an alternative radionuclide for treatment of prostate cancer after tumor-specific sodium iodide symporter gene expression. J Clin Endocrinol Metab 2007; 92: 4451–4458. Michaels J, Dobryansky M, Galiano RD, Ceradini DJ, Bonillas R, Jones D et al. Ex vivo transduction of microvascular free flaps for localised peptide delivery. Ann Plast Surg 2004; 52: 581–584. Dogan T, Kryger Z, Zhang F, Shi DY, Komorowska-Timek E, Lineweaver WC et al. Quadriceps femoris muscle flap: largest muscle flap model in the rat. J Reconstr Microsurg 1999; 15: 433–437.
26 Gregorevic P, Blankinship MJ, Allen JM, Crawford RW, Meuse L, Miller DG et al. Systemic delivery of genes to striated muscles using adeno-associated viral vectors. Nat Med 2004; 10: 828–834. 27 Strauch B, Murray DE. Transfer of composite graft with immediate suture of its vascular pedicle measuring less than 1 mm in external diameter using microsurgical techniques. Plast Reconstr Surg 1967; 40: 325–329. 28 Nishikawa H, Manek S, Green CJ. The oblique rat groin flap. Br J Plast Surg 1991; 44: 295–298. 29 Rahim A, Taylor SL, Bush NL, ter Haar GR, Bamber JC, Porter CD. Physical parameters affecting ultrasound/microbubblemediated gene delivery efficiency in vitro. Ultrasound Med Biol 2006; 32: 1269–1279. 30 Rahim AA, Taylor SL, Bush NL, ter Haar GR, Bamber JC, Porter CD. Spatial and acoustic pressure dependence of microbubblemediated gene delivery targeted using focused ultrasound. J Gene Med 2006; 8: 1347–1357. 31 Taylor SL, Rahim AA, Bush NL, Bamber JC, Porter CD. Targeted retroviral gene delivery using ultrasound. J Gene Med 2007; 9: 77–87.
