The use of high-frequency ultrasound imaging and biofluorescence for in vivoevaluation of gene therapy vectors
Non-invasive imaging of the biodistribution of novel therapeutics including gene therapy vectors in animal models is essential.
This study assessed the utility of high-frequency ultrasound (HF-US) combined with biofluoresence imaging (BFI) to determine the longitudinal impact of a Herpesvirus saimiri amplicon on human colorectal cancer xenograft growth.
HF-US imaging of xenografts resulted in an accurate and informative xenograft volume in a longitudinal study. The volumes correlated better with final ex vivo volume than mechanical callipers (R2 = 0.7993, p = 0.0002 vs. R2 = 0.7867, p = 0.0014). HF-US showed that the amplicon caused lobe formation. BFI demonstrated retention and expression of the amplicon in the xenografts and quantitation of the fluorescence levels also correlated with tumour volumes.
The use of multi-modal imaging provided useful and enhanced insights into the behaviour of gene therapy vectors in vivo in real-time. These relatively inexpensive technologies are easy to incorporate into pre-clinical studies.
KeywordsBiofluorescence Ultrasound Gene therapy Imaging Multi-modal Colorectal cancer
Green fluorescent protein
The use of non-invasive and accurate methods to determine tumour volume, as well as biodistribution and transduction imaging of novel therapeutics, is essential in experimental models in vivo. In particular, for gene therapy studies, knowledge of maintenance, expression and efficacy of the vector is a fundamental part of the testing process . However, this is rarely achieved during the in vivo study of a novel gene therapy strategy, as often only longitudinal calliper measurements of xenograft growth or final histology after treatment are carried out. The spread or loss of a vector is rarely detected during the course of the experiment and for cancer treatment, not all therapies will result in a reduction in tumour volume. Therefore it is important to be able to examine the impact of a gene therapy vector during the in vivo testing phase using different assessment criteria, whilst being mindful of adhering to the principles of reduction, refinement and replacement in animal experiments.
Ultrasound is a non-invasive method that has been utilised recently for tumour growth studies in vivo and is used in the clinic for staging colorectal cancer among others [2, 3]. High-frequency ultrasound (HF-US) machines are available for small animal imaging. They are relatively easy to use and give high resolution greyscale images of mouse anatomy . They also give functional information on the vascular structure of xenografts through the use of contrast agents and are relatively inexpensive and portable compared to MRI machines . Mechanical callipers, however, are still utilised extensively for therapeutic agent testing, especially in gene therapy applications on xenografts . These are very cheap, non-invasive and allow multiple repeated measurements with no anaesthetic required. However, mechanical callipers assume that the growth of xenografts is always ellipsoid and can only measure growth above the skin surface of the animal. In addition, calliper measurements are also affected by skin thickness, subcutaneous fat layer thickness and compressibility of the tumour . From our experience of xenograft growth in gene therapy and other therapeutic studies, we know that this ellipsoid growth pattern is rarely observed, especially as the tumour volume becomes large (above approximately 300mm3).
A gene encoding a fluorescent or luminescent protein is often incorporated into gene therapy vectors in order to enumerate transduction efficiencies in vitro[8, 9]. Moreover, these markers are also very useful for in vivo studies. Optical imaging chambers can be used to image the biodistribution of a vector when administered and can give an indication of the transduction efficiency in the target cells . Optical imaging systems also allow the maintenance of a vector to be determined throughout the course of treatment, as well as examining the genetic stability of the vector over time. The first paper to prove that optical imaging could be used to measure tumour growth used bioluminescence of tumour cells in rat brain and was compared to MRI scans for tumour volume . Imaging of stably-transfected cell lines containing red or green fluorescent protein (RFP or GFP) has been used to measure tumour and metastatic growth [12, 13]. Recent work has also shown that fluorescent intensity correlates better with tumour volume than fluorescent area .
In the study described herein, we aimed to determine whether the use of HF-US measurements were more accurate than mechanical callipers in assessing xenograft volumes of tumour cells which were infected before injection with an experimental gene therapy vector. The use of HF-US to provide anatomical information on tumour growth and BFI to monitor expression of a gene therapy vector in longitudinal studies, were also analysed. The vector we used was a Herpesvirus saimiri (HVS) amplicon which contains the minimal elements for episomal maintenance without infectious capabilities [9, 15]. This gamma-2 Herpesvirus amplicon can incorporate large amounts of heterologous DNA using a HVS-BAC (bacterial artificial chromosome) system and infects a broad range of human cells. The amplicon was previously stably transfected into the SW480 colorectal cancer cell line and contains a constitutively active GFP gene . The presence of the GFP gene enabled monitoring of its persistence during xenograft growth in this study.
The colorectal cancer cell line, HCT116 was stably-transfected with an episomally-maintained Herpesvirus saimiri amplicon incorporating the GFP gene under the control of the Cytomegalovirus (CMV) promoter. These cells were grown in Dulbecco’s Modified Eagle Medium (DMEM, Invitrogen) supplemented with 10% (v/v) foetal calf serum, and 4ul/ml Hygromycin B (Sigma, Poole U.K.) in 5% CO2 at 37oC until there were enough cells for xenograft set up (approximately 3-4 weeks from infection). Parental cell lines were grown in DMEM and serum but no Hygromycin B. Two days before injection the amplicon-transfected cells were transferred to medium without any Hygromycin B.
1 × 106 each of the parental and amplicon-containing cells were collected in 100ul of serum-free DMEM and injected subcutaneously into the right flank of 8-10 week old female CD1 nude mice to form xenografts. 6 mice per group were used. All experiments were performed following local ethical approval and in accordance with the Home Office Animal Scientific Procedures Act 1986.
Tumour volume measurement with mechanical callipers
Anatomical imaging and tumour volume measurement using HF-US
Once per week, mice were anaesthetised using 3% (v/v) isofluorane and xenografts were imaged using a Vevo 770 high-frequency ultrasound machine (FUJIFILM VisualSonics, Inc, Toronto, Canada) equipped with a 40 MHz transducer. The focal depth of the transducer was placed at the mid-point of the centre of the tumour whilst scanning. A 3D scan of the tumour was then performed using the minimum step size possible for the length of tumour and regions of interest were drawn around the xenograft at approximately every 5 frames by an operator with extensive experience of HF-US and analysis . A tumour volume was then calculated using the Vevo 770 version 3 software by creating a 3D reconstruction of these xenografts.
Measurement of biofluorescence
Before sacrifice at day 40, xenografts were imaged in an IVIS Spectrum (PerkinElmer, Inc, Massachusetts, USA). Standard settings for GFP were used (excitation 500nm and emission detected at 540nm) in epi-illumination at high intensity. Binning was set at 8, field of view was 13.1cm and f stop was 2. Regions of interest of the same size were drawn around each xenograft and the total radiant efficiency ([photons/s]/[μW/cm2]) was calculated within this using Living Image version 4.2 software(PerkinElmer, Inc, Massachusetts, USA).
Histology and morphology of xenografts
Once the xenographs were excised, photographs were taken of the intact tumours. The tumours were then cut in half and fixed in 4% (w/v) paraformaldehyde in PBS overnight. After processing and embedding in wax, sections were dewaxed, rehydrated and stained with haematoxylin and eosin. Sections were assessed by an experienced histopathologist.
Analysis of the tumour volumes and vector expression obtained by these methods used Pearson correlations. Positive correlations produced a positive R2 value and were considered significant if p < 0.05. Agreement between the methods was then further analysed by Bland-Altman plots where the central line (mean of differences or bias) and 2 standard deviation (SD) limits of agreement were generated. The bias was considered significant if 0 was not included within these standard deviation lines. These calculations were carried out using GraphPad Prism version 5 (GraphPad Software, Inc, La Jolla, California, USA).
Comparison of tumour growth curves generated using mechanical callipers or HF-US
Comparison of tumour volume measurement methods to the volume calculated using ex vivocalliper measurements
Pearson correlation coefficients of xenograft tumour volumes using different ellipsoid formulae and measured using mechanical callipers, HF-US or mass
π/6 × L × W × H
0.5 × L × W × H
0.5 × L × W2
π/6 × L × W2
π/6 × (L × W)3/2
Comparison of tumour volume measurement methods to final tumour mass
HF-US imaging and BFI of tumour anatomy and gene therapy vector expression
Correlation of total radiant efficiency (fluorescence) and tumour volume measurements
Multimodal imaging in gene therapy applications is a useful tool to shed light on the behaviour of vectors during in vivo testing. In this study, the use of HF-US imaging identified anatomical differences during growth between the parental cell line and the vector-transfected cell line in a xenograft model, even from day 8 after implantation. It has been shown that HF-US can more accurately measure tumour volume compared to the traditional mechanical callipers, as demonstrated in this paper and by others [2, 18]. The use of different ellipsoid volume formulae to generate the tumour volumes from calliper measurements made small differences in accuracy where the highest correlation to mass was found using π/6 × (L × W)3/2 rather than the more commonly used 0.5 × L × W2 as described previously (although based on only one paper ). Correlation to determining volume by water displacement would be the gold standard and would be a useful addition to this study. HF-US volume generation and mechanical calliper measurements by multiple operators would also be valuable for determining variability as these measurements are subject to bias from operators. Jensen and colleagues compared volumes determined by microCT, 18F-FDG-microPET and external callipers, to an ex vivo reference volume calculated by weight and density . They demonstrated that micro-CT was more accurate and reproducible between observers than either external callipers or 18F-FDG-microPET. They also showed that 18F-FDG-microPET was not so useful for determining tumour size, although there was some correlation (R2 = 0.75). This was similar to our findings with biofluorescence imaging. As with our study, this functional tumour imaging modality is useful for metabolic imaging and should give an indication of the effect of a gene therapy vector on tumour viability. In the current study, HF-US accurately showed the slower tumour growth of the vector-transfected cell line compared to the parental cell line, as predicted from in vitro cell growth curves . However, lobe formation was unexpected. We are currently investigating whether this is due to the GFP gene or other components of the vector backbone. We also demonstrated the utility of the different greyscale textures in monitoring different patterns of growth. The discrimination of areas of necrosis and high vascularity (using contrast agents) was also possible. This should allow real-time monitoring of agents that currently have little apparent effect on tumour volume but may have useful effects of anti-angiogenesis or inducing cell senescence. HF-US would be of particular use for very small xenografts, orthotopic models to in transgenic mice such as the Apc Min/+ mouse, where callipers cannot access the tumour. Indeed, gene therapy vectors are also used in non-cancer applications such as diabetes or organ regeneration, where callipers may not be used to measure disease progress or regression. In these cases, HF-US would be invaluable in monitoring progress longitudinally without sacrifice of mice.
In addition to HF-US images, the use of biofluorescence allowed monitoring of tumour growth patterns and correlated well with final tumour volumes (although it must be noted this was based on small numbers with a wide variation). This technique is a simple and very quick method of visualising the tumour and much less expensive than 18F-FDG-microPET, for example. Bio-fluorescence is also applicable to patients. It is currently being trialled in surgery on human tumours to define tumour margins for resection . The monitoring of these two cell lines grown as xenografts showed that the presence and expression of the vector was maintained within the tumour over the duration of the experiment. This information is of great value for gene therapy applications as silencing of the vector can occur, which may not be evident from growth curves or even from immunohistochemistry on ex vivo tumour sections for vector proteins. Linkage of the therapeutic gene of interest to a fluorescent marker gene via an IRES (internal ribosomal entry site) sequence or as a fusion protein would yield valuable information on the efficacy of expression during the time course of an in vivo experiment. It may also be used to reduce costs by eliminating animals in which the introduction of a vector by injection has not been successful.
In conclusion we believe that multi-modal imaging provides useful and enhanced insights into the behaviour of gene therapy vectors in vivo. Addition of imaging to gene therapy protocols would be straightforward especially in the case of relatively inexpensive ultrasound and biofluorescence imaging. The use of multi-modal imaging can give important information on the behaviour of gene therapy vectors in real-time, rather than traditional calliper measurements and final histological examination.
This study was funded by EPSRC grant number EP/I000623/1 and YCR grant number L332.
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2342/13/35/prepub
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.