Detection of different tumor growth kinetics in single transgenic mice with oncogene-induced mammary carcinomas by flat-panel volume computed tomography

Authors

  • Katharina Jannasch,

    Corresponding author
    1. Department of Hematology and Oncology, University Medicine Göttingen, 37075 Göttingen, Germany
    • Department of Hematology and Oncology, University Medicine, Robert-Koch-Str. 40, 37075 Göttingen, Germany
    Search for more papers by this author
    • Fax: ++49 (0)551 25123.

  • Christian Dullin,

    1. Department of Diagnostic Radiology, University Medicine, 37075 Göttingen, Germany
    Search for more papers by this author
  • Christina Heinlein,

    1. Department of Tumor Virology, Heinrich-Pette-Institute for Experimental Virology and Immunology, 20251 Hamburg, Germany
    Current affiliation:
    1. Children's Medical Research Institute, Westmead, New South Wales 2145, Australia
    Search for more papers by this author
  • Frauke Krepulat,

    1. Department of Tumor Virology, Heinrich-Pette-Institute for Experimental Virology and Immunology, 20251 Hamburg, Germany
    Search for more papers by this author
  • Florian Wegwitz,

    1. Department of Tumor Virology, Heinrich-Pette-Institute for Experimental Virology and Immunology, 20251 Hamburg, Germany
    Search for more papers by this author
  • Wolfgang Deppert,

    1. Department of Tumor Virology, Heinrich-Pette-Institute for Experimental Virology and Immunology, 20251 Hamburg, Germany
    Search for more papers by this author
  • Frauke Alves

    1. Department of Hematology and Oncology, University Medicine Göttingen, 37075 Göttingen, Germany
    2. Department of Molecular Biology of Neuronal Signals, Max-Planck-Institute of Experimental Medicine, 37075 Göttingen, Germany
    Search for more papers by this author

Abstract

Transgenic mouse models offer an excellent opportunity for studying the molecular basis of cancer development and progression. Here we applied flat-panel volume computed tomography (fpVCT) to monitor tumor progression as well as the development of tumor vasculature in vivo in a transgenic mouse model for oncogene-induced mammary carcinogenesis (WAP-T mice). WAP-T mice develop multiple mammary carcinomas on oncogene induction within 3 to 5 months. Following induction, 3-dimensional fpVCT data sets were obtained by serial single scans of entire mice in combination with iodine containing contrast agents and served as basis for precise measurements of tumor volumes. Thereby, we were able to depict tumors within the mammary glands at a very early stage of the development. Tumors of small sizes (0.001 cm3) were detected by fpVCT before being palpable or visible by inspection. The capability to determine early tumor onset combined with longitudinal noninvasive imaging identified diverse time points of tumor onset for each mammary carcinoma and different tumor growth kinetics for multiple breast carcinomas that developed in single mice. Furthermore, blood supply to the breast tumors, as well as blood vessels around and within the tumors, were clearly visible over time by fpVCT. Three-dimensional visualization of tumor vessels in high resolution was enhanced by the use of a novel blood pool contrast agent. Here, we demonstrate by longitudinal fpVCT imaging that mammary carcinomas develop at different time points in each WAP-T mouse, and thereafter show divergent growth rates and distinct vascularization patterns. © 2009 UICC

Accompanying the rapid progress in deciphering the human and mouse genome, many transgenic mouse models have been generated. Weissleder and Mahmood already assumed in 2001 that more than 25 million transgenic and knockout mice would be raised that year for experimental studies, accounting for over 90% of all mammals in research.1 These mouse models offer the opportunity to study gene functions in their biological context in a much more physiological way than any other approach. Especially cancer research benefits from the advantages of these transgenic mice in elucidating the role of a specific gene in tumor growth and spread as well as angiogenesis. One promising model in addressing such questions is the WAP-T transgenic mouse.2–4 In these mice, the SV40 early gene region is induced specifically in differentiating mammary epithelial cells upon activation of the WAP-promoter during late pregnancy and lactation, and drives mammary carcinogenesis in mammary epithelial cells re-expressing the transgene after involution of the mammary glands. The SV40 early region mainly serves to functionally eliminate the p53 and pRb tumor suppressors by binding to the SV40 large T-antigen5 and to change the specificity of the phosphatase PP2A6 by its interaction with SV40 small T-antigen, thereby initiating tumor development. Contrary to implantation tumor models, WAP-T mice develop multiple endogenous mammary carcinomas at various time points after involution. To date, these tumors have only been analyzed at experimental end points at dissection. There is a high demand for noninvasive imaging techniques to determine precisely the onset of tumor growth and to characterize growth over time, especially in spontaneous or induced murine tumor models such as WAP-T mice. Therefore, techniques have been applied that enable imaging and exact monitoring of tumor growth and progression in small animals, such as magnetic resonance imaging (MRI7), ultrasound,8–10 computed tomography (CT11) and nuclear imaging.12, 13 Each modality possesses a unique combination of advantages and disadvantages that affect their selection for use in a particular study.

Here we used flat-panel volume computed tomography (fpVCT) in longitudinal studies. Contrary to μCT that has long acquisition times and high x-ray exposure but works with extremely high resolution of 5–10 μm,14, 15 fpVCT combines the advantages of 3-dimensional (3D) imaging with a resolution of about 150 μm to visualize anatomical structures of small animals in vivo with a short acquisition time, thereby allowing the use of contrast agents which would otherwise be excreted during the scan. Furthermore, with a rather low radiation dose of 40 ± 5 mGy a fpVCT scan is far less harmful for the animal than a μCT scan that delivers a radiation dose of about 100 mGy.16 FpVCT has already been described as a system for noninvasive imaging of different organ systems in preclinical research.17 In combination with the administration of contrast agents, fpVCT allows accurate measurements of tumor volumes, vessel formation18 and morphological changes within the tumor over time in subcutaneous and orthotopic tumor models in longitudinal studies.19 FpVCT has also been used in monitoring skeletal structures,20, 21 including progression of osteolytic lesions,22 as well as for the evaluation of therapy responses in vivo over time.23

Vascular imaging with x-rays requires the use of special contrast enhancement agents. Scans using conventional small-molecular contrast agents such as Isovist 300 need a complex time management because of the rapid distribution of the contrast agent between the intravascular and extracellular fluid compartments of the body. As a result of such distribution, concentration of small-molecular contrast agents in plasma decreases rapidly, leading to virtually equal vascular and interstitial concentrations, an effect that compromises the duration and quality of vascular contrast enhancement. On the contrary, the newly available eXia 160 is an aqueous colloidal poly-dispersed contrast medium, which avoids renal filtration and diffusion from the intravascular into the interstitial space, thereby providing the opportunity of using it as a blood pool agent. It stays in the blood circulation for a prolonged period of time with a half life of about 1 hr and is suited for characterization of blood vessels and certain types of pathologic abnormalities found in cancer (communication from Binitio Biomedical Inc.). Both approaches were used in this study.

In this study, we applied fpVCT in an oncogene-based inducible tumor mouse model for assessing tumor onset, growth rates and vessel development as well as morphological characteristics such as formation of cysts in vivo. We demonstrate that fpVCT exhibits excellent sensitivity in identifying diverse time points of tumor onset as well as monitoring distinct growth characteristics and vessel formation of multiple mammary carcinomas that developed within single transgenic mice.

Abbreviations:

fpVCT, flat-panel volume computed tomography; H&E, haematoxylin and eosin; MIN, multifocal intraepithelial neoplasia; MRI, magnetic resonance imaging; 3D, three-dimensional; WAP, whey acidic protein.

Material and methods

Animals

All animals were handled according to German regulations for animal experimentations and all animal protocols were approved by the administrations of Hamburg and Lower Saxony, Germany. Ten induced female WAP-T-NP8 mice2 at the age of 3 months were used in this study. In these mice the SV40 early gene region, driven by the whey acidic protein (WAP)–promoter, was introduced into the germline of CB6F2 mice, which then were back-crossed into BALB/c mice (for details see Ref. 2). Expression of the transgene was induced by lactotrophic hormones during late pregnancy and lactation and is directed to differentiating mammary epithelial cells of all mammary glands. By the time mice exhibited any adverse effects because of tumor growth that led to suffering, such as weight loss, the experiment was terminated. Subsequently, WAP-T mice were sacrificed, tumors excised, weighed and the abdomen and thoracic cavity were examined systematically for the presence of metastases. Visible tumors were externally measured with a caliper after excision. Tumors were collected and placed in phosphate—buffered 4% formalin for 16 hr at room temperature and embedded in paraffin. Tissue sections (2.5 μm) were obtained, stained with haematoxylin and eosin (H&E), and were inspected by routine microscopic examinations.

fpVCT imaging

Mice were imaged with a nonclinical flat panel-based volume computed tomograph prototype, the fpVCT (GE Global Research, Niskayuna, NY). Scanning started between 60 and 70 days after induction and was repeated regularly. Mice were anesthetized with vaporized isoflurane at 0.8–1% concentration throughout the imaging session and placed belly-down perpendicular to the fpVCT gantry axis of rotation. Thereby, it was possible to scan the whole mouse with 1 rotation. The nonionic isoosmolar small-molecular iodinated contrast agent, Isovist 300 that contains 300 mg of iodine per milliliter (Bayer-Schering, Berlin, Germany) was routinely used for scanning. It was administered intravenously via the tail vein using an insuline syringe at 30 g × [1/2]” (Braun, Melsungen, Germany) at 150 μl 30 sec before each scan. For better demonstration of small blood vessels, Isovist 300 was replaced by the blood pool agent eXia 160 that contains 160 mg of iodine per milliliter in a polymolecular colloidal complex (Binitio Biomedical Inc., Ottawa, Canada) at selected time points. It was administered intravenously at 200 μl 90 sec before each scan according to the manufacturer's instructions. For comparison of these 2 contrast agents some mice were scanned using both contrast agents consecutively during the last days of the experiment. EXia 160 was administered 2 hr after the scan using Isovist 300 in order to allow the clearing of the first contrast agent. All fpVCT data sets were acquired with the following protocol: 500 views per rotation, 4 sec rotation time, 360 used detector rows, tube voltage of 80 kVp and a current of 100 mA. A modified Feldkamp algorithm was used for image reconstruction resulting in isotropic high-resolution volume data sets (512 × 512 matrix with a voxel size of about 100 μm to prevent the creation of resampling artifacts). Tumors were first detectable from a size of 0.001 cm3 which corresponds to 1000 voxels. For tumor segmentation and volume estimation data sets were analyzed with voxtools 3.0.64 Advantage Workstation 4.2 (GE Healthcare, UK), which uses an algorithm similar to region-growing. Density histograms were performed according to Missbach-Guntner et al.19

Results

Determination of growth rates of multiple mammary carcinomas within single mice by fpVCT

To monitor onset and progression of mammary carcinomas in the inducible transgenic tumor model, whole bodies of 10 living WAP-T transgenic mice were scanned at the time points described in material and methods over a period of several months after induction of transgene expression. Multiple carcinomas developed within the mammary glands (for nomenclature see Fig. 1) of induced WAP-T transgenic mice within 91–170 days. Table I shows that the mice investigated developed between 2 and 6 tumors. Correlation between the number of tumors and the mammary gland in which they developed, showed, that mammary glands 3 and 7 hosted most tumors, whereas mammary glands 4, 5 and 6 showed only few tumors. Interestingly, when the number of tumors was correlated to the region of the mouse in which they developed, most tumors developed within the thoracal mammary glands and only few were found in inguinal mammary glands. Comparison of size, localization and onset of the tumors did not show any correlation. Tumors with volumes greater than 1 cm3 were found in cervical, thoracal, abdominal as well as in inguinal mammary glands and all tumors became increasingly irregular in shape during development. Figure 1 focuses on a representative WAP-T transgenic mouse and demonstrates the use of contrast-enhanced fpVCT to visualize all mammary carcinomas that developed in this mouse over time. Tumor sites were numbered referring to the sites of mammary glands (Fig. 1a). Two tumors were clearly visible by inspection at the day of dissection. Histological examination revealed that small areas of necrosis were dispersed within most WAP-T mammary carcinomas (Fig. 1a). As shown in Figure 1b, fpVCT enabled the visualization of the mammary carcinomas in 3D. After precise delineation of the tumors in vivo, they were segmented in fpVCT data sets for the determination of tumor volumes (Fig. 1b). For each time point, tumor volumes were calculated (Fig. 1c, upper panel) generating precise growth curves for each tumor within each mouse over time (Fig. 1c, lower panel). In this mouse, 6 mammary carcinomas developed at distinct times with differences up to 40 days in divergent mammary glands, reaching tumor volumes between 0.004 and 1.227 cm3 at the end of the experiment. Their growth kinetics were comparable up to a tumor volume of 0.2 cm3. Because of early detection, tumor progression could be monitored by fpVCT over a long period of time, even up to 90 days. Since mammary carcinomas in these mice developed with increasingly irregular shapes or lobulated, calculation of tumor volumes by fpVCT was much more accurate than measurements with a caliper, which were far from representative throughout the whole experiment (data not shown).

Figure 1.

Growth rate of multiple mammary tumors within one mouse depicted in vivo by 3D reconstructions of fpVCT data sets. (a) Representation of a tumor-bearing WAP-T transgenic mouse. Visible mammary carcinomas that developed at different sites are indicated by circles. The nomenclature of the tumors depending on their site of origin is displayed on the right side. Section of a representative mammary tumor stained with H&E is depicted (middle). Note, that the tumor comprises several small areas of necrosis within intact tumor tissue (arrows). (b) Shows a 3D fpVCT image of a mammary tumor 182 days after induction in situ (top), followed by another visualization protocol (middle), displaying particularly soft tissue and vessels. After segmentation only the tumor is depicted (bottom). (c) Top panel: Visualization of the enlargement of the virtually isolated mammary carcinomas by repeated fpVCT scans in combination with contrast agent performed between day 143 and 182 after induction. Bottom panel: Resulting tumor growth curves are displayed. Note, that each mammary carcinoma is depicted at first by fpVCT at different time points but all tumors display similar growth kinetics up to a volume of 0.2 cm3. Scale bars: (a) 2 mm, (b) 10 mm. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Table 1. Development of Mammary Carcinomas in WAP-T Transgenic Mice
Mouse no.1ntumorFirst onset (days after induction)Vlargest tumorMammary gland no.largest tumor
  • 1

    Listing of number (n) and first detection of mammary carcinomas that developed in 10 induced WAP-T transgenic mice. Each mouse developed more than one mammary carcinoma between 89 and 188 days after induction. Volume and tumor site of the largest mammary carcinoma are also shown for each mouse. Mice were sacrificed when exhibiting any adverse effects which lead to major differences in the volumes of the largest tumors.

  • 2

    Representation of tumors that developed in different mammary glands. Upper panel: Number of tumors developing in single glands, lower panel: number of tumors that developed in different regions of the mouse.

12910.1503
261431.2272
32921.2443
431220.2432
54891.5464
62971.7062
731881.2596
851141.6244
931120.5138
1031321.0687
Mammary gland no.2  ntumor
1  4
2  5
3  8
4  2
5  2
6  1
7  7
8  5
Cervical  9
Thoracal  12
Abdominal  9
Inguinal  4

Assessment of tumor growth characteristics

To assess tumor progression in WAP-T mice, animals were scanned regularly by fpVCT over up to 5 months. Interestingly, mammary carcinomas developing in these transgenic mice exhibited different growth characteristics. This is illustrated by volume representations of fpVCT data sets depicting segmented mammary carcinomas at distinct time points from 2 representative mice (Fig. 2). In one induced WAP-T mouse, shown in Figure 2a, 2 large mammary carcinomas that developed within the left cervical and abdominal glands displayed comparable tumor sizes of 1.050 cm3 and 1.244 cm3 respectively at the end of the experiment. However, measurements of tumor volumes during disease progression demonstrated diverse tumor growth rates for these tumors. While the cervical tumor was already detectable 92 days after induction and developed steadily, the abdominal tumor appeared only 140 days after induction and showed a highly exponential growth, reaching an even slightly larger volume of 1.244 cm3 at the end of the experiment. In contrast to the growth behavior of these mammary carcinomas (Fig. 2a), a thoracal and an abdominal tumor of a different mouse were first detected exactly at the same time point, 132 days after induction with a volume of 0.001 cm3 as shown in Figure 2b. The tumor in the thoracal mammary complex increased its volume to 1.068 cm3 within 48 days while the tumor at the abdominal site grew much slower displaying a volume of 0.18 cm3 at the same time point. These distinct tumor growth characteristics within one mouse could easily be missed using standard methods such as measurements with a caliper.

Figure 2.

Tumor growth characteristics over time visualized in vivo using 3D fpVCT data sets. Illustration of segmented mammary carcinomas over time and graphic representations of their different growth rates monitored with repeated scans of contrast-agent based fpVCT. (a) Comparison of the development of 2 mammary carcinomas within one mouse at 2 different sites. Note that comparable tumor volumes are depicted at dissection, although these tumors had different time points of tumor onset and thereby divergent growth rates. Top, isolated tumors from 3D fpVCT data sets; bottom, tumor growth curves using automatically determined tumor volumes from fpVCT data sets. (b) Visualization of 2 tumors over time within one mouse. Note, that initially tumors at the same day show comparable small volumes, but display distinct growth kinetics with divergent tumor volumes at dissection.

Visualization of a cyst that developed within a mammary carcinoma

In one of the WAP-T transgenic mice a cyst was macroscopically visible within a group of 3 tumor nodules at the thoracal side of the mouse on the day of dissection (Fig. 3a). FpVCT data sets from the same day show the group of these 3 tumor nodules (Fig. 3b). A coronal section at the end of the experiment (Fig. 3c) depicts the cyst visible as a dark cavity (arrow) within these tumor nodules. The development of the cyst can be monitored by virtual transverse sections at distinct time points (Fig. 3d). Tumor nodules with a solid appearance can already be seen (circle) 167 days after induction. The corresponding transverse section on day 174 demonstrates the enlargement of the tumor and a region within 1 tumor nodule that appeared darker than the surrounding mammary carcinoma, and is therefore indicative for the development of a cavity (arrow). Eight days later, at dissection, the cyst is clearly delineated against the surrounding tumor tissue (arrow). These results underscore the potential of fpVCT to discriminate between tumor tissue and a cyst and thus to detect and monitor the development of cysts within tumor tissue in vivo over time.

Figure 3.

Development of a cyst monitored in vivo by fpVCT. (a) Representation of a cyst within a thoracal tumor (circle) marked by arrows during dissection and (b–d) during development using fpVCT data sets with different visualization protocols. (b) Visualization of the mammary carcinoma at the end of the experiment by fpVCT. Note, that in this protocols only different tumor nodules can be distinguished. (c) In contrast, a virtual coronal section clearly visualizes the cyst at this time point, characterized through reduced density. (d) In a coronal section of day 167 a nodule is visible which does not differ from the surrounding tumor tissue, on day 174 the density has decreased, identifying the nodule as a cyst. By day 182 the cyst is clearly delineated against the surrounding tumor tissue (arrow). Scale bars: 10 mm.

Visualization of blood vessel development by fpVCT

FpVCT analysis of WAP-T transgenic mice was also used to monitor the recruitment and formation of blood vessels accompanying tumor development. Visualization of blood vessels can be achieved through the use of contrast agents, but is limited due to the resolution of the system. Blood vessels were depicted over time in arbitrary planes of the tumor periphery from fpVCT data sets. As shown in the representative example presented in Figure 4, fpVCT images already visualized a distinct blood supply to a thoracal mammary carcinoma with a rather small volume of 0.073 cm3 150 days after induction. Only 1 tumor vessel can be seen that discharges into the right brachiocephalic vessel. Compared with vessels seen on the left—nontumor—side of the body, the diameter of the vessel supplying the tumor was clearly enlarged. The tumor developed rapidly to a size of 0.459 cm3 within only 16 days (Fig. 4, middle; compare inserts) accompanied by further enlargement of the diameter of the corresponding tumor vessel. The same vessel that appeared straight at day 150, exhibited a rather corkscrew-like shape 16 days later. At this time point tumor-surrounding and tumor-invading small vessels could be clearly depicted by fpVCT. The macroscopic appearance of the tumor vessels during dissection 14 days later confirmed the findings visualized by fpVCT. The corkscrew-like vessel on the apical side of the tumor (Fig. 4, right, arrow) is located next to an enlarged vessel discharging into the brachiocephalic vessel. From day 150 until the day of dissection (day 180), the tumor volume had increased by 130% to 1.068 cm3. All thoracal tumors that were investigated showed this or a similar blood vessel development. Thus, the dynamic process of blood vessel formation during tumor growth could be qualitatively monitored over time using fpVCT.

Figure 4.

Tumor vessel development over time visualized in vivo using 3D fpVCT data sets. Depiction of contrast agent-containing tumor vessels within a developing mammary carcinoma by repeated fpVCT scans of a representative mouse. Bottom panel: Tumor volumes (V) are calculated and locations of the tumors are illustrated using a different visualization protocol. Top panel: At day 150 a large efferent vessel discharges in the brachiocephalic vessel, no other vessels are depicted. Sixteen days later this vessel appears to be enlarged with a corkscrew-like structure. At this time point small tumor-surrounding and invading vessels became visible. The corkscrew-like vessel can be related to a vessel seen macroscopically at dissection at day 180 (arrow). Scale bars: 10 mm.

Depiction of tumor vessels using a blood pool contrast agent

Although conventional small-molecular iodinated contrast agents are widely used for tumor and vascular imaging in both angiography and computed tomography, they diffuse massively from the intravascular space within seconds after administration, which rapidly reduces the differential contrast of vessels to be imaged. To further improve the visualization of tumor vessel formation we used the blood pool contrast agent eXia 160, which has the characteristic feature to extravasate slowly out of the vessel, and directly compared data sets acquired with the blood pool agent eXia 160 to images obtained after administering the conventional small-molecular contrast agent Isovist 300. Tumor-bearing mice were scanned with both contrast agents within 2 hrs during the last days of the experiment. Figure 5 shows different visualization protocols for fpVCT scan data of a representative mouse that developed a tumor in the left thoracal mammary complex with a volume of 1.350 cm3. Although the dose of Isovist 300 was administered at an iodine excess compared with eXia 160, only few and mainly rudimentary vessels were depicted, located lateral of the mouse and at the apical side of the tumor (Fig. 5a). In comparison, fpVCT scans obtained with eXia 160 resulted in more detailed views of tumor vessels (Fig. 5b,c) visualizing even a diffuse network of small vessels inside the tumor. Views from different angles demonstrated that 1 major vessel supplies the tumor and discharges into the brachiocephalic vessel. Furthermore, a corkscrew-like vessel is clearly seen in the basal part of the tumor (Fig. 5c, arrow).

Figure 5.

Comparison of vessel images generated by fpVCT using 2 different iodine-based contrast agents. Representative fpVCT images of a mouse scanned with Isovist 300 as well as eXia 160 within 2 hr 138 days after induction. (a) Different visualization protocols of the anterior chest of a mouse scanned with Isovist 300 and in (b, c) scanned with the blood pool agent eXia 160 are displayed. (c) Shows a view from a different angle as well as a view from the posterior side of the mouse from the 3D representation of the data sets acquired with eXia 160. Note, that in comparison to the picture received using Isovist 300, a much higher visibility of small blood vessels is obtained by fpVCT. Thereby tumor-surrounding and invading small vessels can be depicted. The main tumor supplying vessel discharges into the brachiocephalic vein (red arrow) and a corkscrew-like vessel can be depicted in the basal part of the tumor (white arrow). Scale bars: 10 mm.

Discussion

In this study, analysis of tumor growth using noninvasive VCT imaging with flat-panel detectors allowed us to monitor the different growth kinetics of each mammary carcinoma that developed at different time points over up to 5 months within individual WAP-T transgenic mice. By analyzing isotropic high-resolution 3D-volume data sets we simultaneously detected tumors with comparable growth kinetics but different tumor onsets, as well as tumors with resembling start or end points, but deviating growth kinetics. All WAP-T mice examined in this study developed more than 1 tumor (Table I). Tumors were located in all mammary glands, but never in all mammary glands at the same time. Interestingly, more tumors developed in the thoracal mammary glands while fewer tumors could be found in the inguinal mammary glands. Further studies have to clarify whether this might relate to differences in tumor vascularization. No correlation between size and localization of tumors could be observed. WAP-T transgenic mice have been established by Schulze-Garg et al.2 and have been proposed as a model to study the molecular events leading to ductal carcinoma in situ (DCIS) and its progression to invasive disease. So far, tumors of these mice had only been investigated at the end point of the experiment at dissection. Here, we report new insights into the growth kinetics and vessel formation of these mammary carcinomas during tumor development and progression. Although WAP-T transgenic mice are genetically identical, the tumor growth patterns were unexpectedly diverse. This fact would not have been depicted without fpVCT in vivo monitoring. The unexpected differences in the tumor growth pattern of individual tumors in a single WAP-T mouse is at first glance surprising. However, whole mount analysis of mammary glands of induced mice showed that about 90 days after involution virtually all terminal end buds within each mammary complex had been converted to intraepithelial neoplasias (MIN, corresponding to human DCIS; unpublished data). Thus progression from MIN is the decisive event in the development of invasive mammary carcinomas in these mice. This progression normally is an extremely rare event, probably requiring as yet unknown additional events, including genetic and epigenetic changes as well as altered microenvironmental, metabolic and immune system driven influences.

Mammary carcinomas from WAP-T mice are mostly irregularly shaped and often consist of several nodules. Therefore, it can be expected that volumes obtained from fpVCT data sets are more representative compared with diameter-based volumetric measurements using a caliper which assumes an ellipsoid configuration of the tumor. The feasibility of fpVCT imaging for an accurate real-time assessment for growing carcinoma was already shown and validated in different tumor models as well as through the use of phantoms where the relative measurement errors were between 0.99 and 1.01%.19 Although a precise indication of tumor volumes to the third decimal does not give major information on the growing tumor, it allows the early detection and quantification of very small tumors. Furthermore, in the same publication, these authors identified and quantified central necrosis that developed within mammary carcinomas in this implanted tumor mouse model by density histograms.19 In our study, density histograms calculated from fpVCT data sets of WAP-T mice showed uniformly distributed densities not enabling the delineation of necrosis within tumors during tumor development (data not shown). This confirms histological analyses indicating that endogenous WAP-T tumors contain multiple but small areas of necrosis.

Other tumor constituents that do not reflect viable tumor tissue are cysts, which can often be observed within breast tissue.24 For the evaluation of the efficacy of anticancer treatment it is necessary to obtain information of the tumor tissue composition and their changes during tumor development and in response to therapy. Since cysts contain air, fluids or semisolid material and thereby display reduced density compared to the surrounding tissue, fpVCT is very suitable to monitor the development of cysts in vivo and provides valuable information about the tumor composition. Thereby scanning via fpVCT offers excellent prospects for the preclinical evaluation of the therapeutic success, e.g. effects on volumes of tumors as well as of cysts.

Since fpVCT enables the determination of the onset of tumor growth even before the tumor is palpable, this technique will be of enormous impact when arranging therapeutic studies. Preclinical trials using transgenic mice can be more easily managed, since the start of therapy can be defined in each mouse at a certain tumor volume and growth rates can be depicted in response to therapy. At present, longitudinal studies of mouse cancer models require large cohorts, since autopsy at different tumor stages has been the only reliable method to evaluate tumor progression as well as anticancer treatment efficacy. With the ability of fpVCT to monitor tumor growth and progression over time in living mice, significant data sets are obtained with a markedly reduced number of animals. An influence of radiation on tumor growth cannot be observed. Moreover, in therapeutic studies the observed biological variability in these mice would probably not have been recognized using traditional time-point sacrifice methods. This problem can therefore be minimized due to the analysis of one and the same animal for several measuring points. Up to date, fpVCT has already successfully been applied in heterozygous Ptchneo67/+ transgenic mice to evaluate therapy responses in vivo over time.23

Angiogenesis is critical for the growth of malignant tumors. Different attempts have been made to visualize tumor angiogenesis. MRI, CT, positron emission tomography, ultrasound and optical imaging provide noninvasive images of angiogenesis in animals and humans.25, 26 CT and MRI are used frequently to evaluate therapeutic efficacy in patients,27, 28 but microvessels inside tumors cannot be assessed.

Excellent contrast in visualizing microvessels has been achieved using barium sulfate29 as well as the new lipophilic contrast agent Angiofil30 in μCT. However, as injection of both substances is lethal, they are not suited for long-term studies. Here we report the successful use of fpVCT to illustrate and monitor the 3D formation of tumor blood vessels over time. The fpVCT data sets clearly define origin and distribution of blood vessels supplying a mammary carcinoma. One of the most challenging problems using fpVCT is the spatial resolution. At 10% MTF (modulation transfer function) the resolution constitutes 200 μm according to the manufacturer. The perceptibility of details in areas of high contrast could be determined at 150 μm using line grading. However, this is not high enough for the visualization of small vessels. Enhanced visibility could be achieved by using the newly available blood pool agent eXia 160. In comparison to other imaging techniques such as μCT or MRI, fpVCT has the advantage of providing high-resolution images of small vessels within a very short acquisition time of 4 sec,31 thereby avoiding the need to maintain body temperature and to administer long-lasting anesthesia, and thus enabling serial scans. However, scans using conventional small-molecular contrast agents like Isovist 300 need high amounts of iodine to provide adequate contrast enhancement and a complex time management because of the rapid distribution of the contrast agent between the intravascular and extracellular fluid compartments of the body. As a result of such distribution, concentration of small-molecular contrast agents in plasma decreases rapidly, leading to virtually equal vascular and interstitial concentrations, an effect that compromises the duration and quality of vascular contrast enhancement. Moreover, enlargement effects from leaking contrast agents are common. To optimize the visualization of blood vessels, there is a strong need for blood-pool contrast agents that remain intravascular.32, 33 Comparison of the blood pool agent eXia 160 with the conventional small-molecular contrast agent Isovist 300 resulted in a significantly higher visibility of smaller vessels in images obtained with eXia 160. Isovist 300 is a nonspecific water-soluble contrast medium that is commonly used in clinical settings to increase the contrast on CT images. Following intravascular injection, Isovist 300 is immediately diluted in the circulating plasma and then distributed rapidly between circulating blood and other extracellular fluids. Equilibration with the extracellular compartments is reached within a short time. Isovist 300 is excreted mainly through the kidneys and can be visualized in the renal parenchyma already within seconds following intravenous administration. Rapid renal clearance and rapid dilution of Isovist 300 in the extravascular fluid compartments causes a sharp decline in Isovist 300 plasma concentration, which severely impairs the degree of enhancement inside the blood vessels and reduces the differential contrast of structures to be imaged. The ability of small water-soluble contrast agents, like Isovist 300, to enhance vascular compartments of small laboratory animals, is additionally impaired by a higher heart rate of mice which is at least 10 times faster than in humans. Contrary to Isovist 300, eXia 160 is a colloidal polydisperse iodinated contrast agent. After intravascular administration eXia 160 shows a pharmacokinetic profile consistent with that of a slow-clearance blood-pool agent. Because of its large molecular weight, eXia 160 does not diffuse through the endothelium of the capillary wall and is not cleared through the kidneys by glomerular filtration. Hepatobiliary excretion is the major pathway of eXia 160 elimination from the intravascular compartment. Most of eXia 160 is cleared by the liver within the first 6 hr and only insignificant increase in attenuation remains at 24 hr. EXia 160 shows a dose-dependent enhancement in normal liver parenchyma, but it does not enhance primary or metastatic tumorous liver tissue. In some cases of hepatic insufficiency or fatty liver infiltration liver uptake of eXia 160 can be decreased (communication from Binitio Biomedical Inc.). Till date there are only few iodine-containing blood pool agents available. EXia 160 contains 160 mg iodine/ml which is a higher amount than hitherto used blood pool agents, promising an enhanced visibility. Here we report for the first time the use of an iodine-containing blood pool agents for the visualization of mammary carcinoma vascularization. Monitoring alterations in tumor vessel development in genetically engineered mice in comparison to wild type mice over time might offer insights in the role of different genes in vascularization34 and will help to assess the efficacy of antiangiogenic treatments.

In summary, our study demonstrates the use of serial fpVCT scans to depict early and noninvasively tumor onset, to identify distinct growth kinetics as well as to monitor dynamic processes of blood vessel formation over the course of each mammary carcinoma that developed in WAP-T transgenic mice. Comparative gene analysis of tumors with different growth kinetics depicted by fpVCT should enable researchers to correlate gene expression and function. FpVCT is thus an important new tool for the identification of novel genes influencing tumor cell proliferation in future studies. Furthermore, by defining distinct time points of tumor onsets and calculating precise tumor volumes it is ideally suited for follow-up studies on tumor progression in a physiological environment and for preclinical studies of tumor treatment.

Acknowledgements

The authors acknowledge the excellent technical assistance of Ms. Sarah Greco and Ms. Roswitha Streich as well as Ms. Birgit Grübbeling and Ms. Karin Stapp-Kurz for technical support running the fpVCT. Furthermore, the authors thank Dr. Ekatarina Rizhevskaya for providing information about eXia 160. We state that the authors do not have any duality of interest (both financial and personal).

Ancillary