Imaging of Intratumoral Inflammation during Oncolytic Virotherapy of Tumors by 19F-Magnetic Resonance Imaging (MRI)

Background Oncolytic virotherapy of tumors is an up-coming, promising therapeutic modality of cancer therapy. Unfortunately, non-invasive techniques to evaluate the inflammatory host response to treatment are rare. Here, we evaluate 19F magnetic resonance imaging (MRI) which enables the non-invasive visualization of inflammatory processes in pathological conditions by the use of perfluorocarbon nanoemulsions (PFC) for monitoring of oncolytic virotherapy. Methodology/Principal Findings The Vaccinia virus strain GLV-1h68 was used as an oncolytic agent for the treatment of different tumor models. Systemic application of PFC emulsions followed by 1H/19F MRI of mock-infected and GLV-1h68-infected tumor-bearing mice revealed a significant accumulation of the 19F signal in the tumor rim of virus-treated mice. Histological examination of tumors confirmed a similar spatial distribution of the 19F signal hot spots and CD68+-macrophages. Thereby, the CD68+-macrophages encapsulate the GFP-positive viral infection foci. In multiple tumor models, we specifically visualized early inflammatory cell recruitment in Vaccinia virus colonized tumors. Furthermore, we documented that the 19F signal correlated with the extent of viral spreading within tumors. Conclusions/Significance These results suggest 19F MRI as a non-invasive methodology to document the tumor-associated host immune response as well as the extent of intratumoral viral replication. Thus, 19F MRI represents a new platform to non-invasively investigate the role of the host immune response for therapeutic outcome of oncolytic virotherapy and individual patient response.


Introduction
Oncolytic virotherapy of tumors is based on the lytic destruction of solid tumors mediated by infection of the malignant tissue by tumor-specific viruses [1][2][3][4]. Today, a multitude of different virus strains with oncolytic potential are described in the literature and promising pre-clinical data as well as clinical trial reports from oncolytic virotherapy are available [5][6][7]. In addition to the oncolytic tissue destruction massive inflammation within the tumor microenvironment occurs, which is primarily directed against the virus [8][9][10]. However, virus-mediated cell lysis also leads to the release of tumor-associated antigens, which may finally stimulate an anti-tumoral immune response [11,12]. So far, it is well known that oncolytic virotherapy is accompanied by a host immune response, however, the individual time course, the polarization of the immune response, and the effect on the therapeutic efficacy remain to be elucidated.
In practice, a longitudinal, non-invasive quantification of the intratumoral inflammation during oncolytic virotherapy may provide substantial benefits to therapeutic monitoring, tumor diagnostics and indirect virus imaging as well as to the optimization of new therapeutic virus strains. Recently, 19 F/ 1 H MRI was introduced as a novel and promising non-invasive imaging modality of different inflammatory conditions in rodent models of cardiac ischemia [13], allograft rejection [14], LPSinduced pulmonary inflammation [15], inflammatory bowel disease [16], neuro-inflammation [17] and bacterial abscess formation [18]. In practice a single intravenous injection of an emulsified perfluorocarbon (PFC) predominantly lead to the phagocytic uptake by the monocyte-macrophage system followed by their detection via 19 F MRI [19]. As a consequence of the progressive infiltration of PFC-labeled immune cells into inflamed tissues, the foci of inflammation can be localized as hotspots and morphologically correlated to the anatomical or patho-physiological context with the help of 1 H images [19,20].
Although recent advances in the visualization of inflammatory processes using different imaging modalities were reported [21][22][23], 19 F MRI has several advantages especially compared to contrast agent based 1 H MRI. First, no detectable fluorine signal is present in tissues enabling background-free 19 F MRI of the exogenous 19 F marker [20] and the directly acquired 19 F signal facilitates the quantification of the marker amount [24]. Second, the 1 H signal is not altered compared to other contrast agents making the assessment of other quantitative 1 H parameters feasible [18]. Therefore, 19 F MRI is a promising modality for imaging of inflammation combined with high anatomical resolution ( 1 H) in the context of oncolytic virotherapy.
The abundance of inflammatory cells is a hallmark of cancer, and their importance in cancer development and progression have been demonstrated in numerous studies [25,26]. Of particular importance are macrophages due to their high phenotypic heterogeneity reaching from cytotoxic M1-polarized macrophages, which can theoretically harm tumor tissues to M2-polarized macrophages (tumor-associated macrophages, TAMs), which promote tumorigenesis [27]. The destruction of TAMs or the reprogramming of tumor-promoting M2 to cytotoxic M1 macrophages represents a current immunotherapeutic strategy against cancer [28]. Since colonization of solid tumors with systemically injected oncolytic viruses induces macrophage-recruitment this treatment may interfere with the polarization of macrophages by confronting the host with a variety of pathogen-associated molecular patterns (PAMPs) [12].
To the best of our knowledge, PFC nanoemulsions have not yet been used to target, visualize and monitor viral tumor infection and oncolytic virotherapy of tumors. Consequently, in this proofof-principle study 19 F MRI was applied to analyze tumor-bearing mice treated with the oncolytic Vaccinia virus (VACV) GLV-1h68. This oncolytic virus strain was previously described as a powerful agent to treat various types of cancer [29][30][31]. The presented results suggest 19 F MRI as a non-invasive methodology to document the tumor-associated host immune response as well as the extent of intratumoral viral replication.

Ethics Statement
All animal experiments were carried out in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Explora Biolabs (San Diego, CA, USA, protocol number EB08-003) or the government of Unterfranken (Würzburg, Germany, protocol number AZ 55.2-2531.01-17/08). Both the Institutional Animal Care and Use Committee of Explora Biolabs and the government of Unterfranken specifically approved this study.

Virus
The construction of the attenuated Vaccinia virus strain GLV-1h68 was previously described by Zhang et al. [31]. Briefly, three expression cassettes (encoding for Renilla luciferase-GFP fusion protein, b-galactosidase and b-glucuronidase) were recombined into the F14.5L, J2R and A56R loci, respectively, of the LIVP strain virus genome. Viruses were propagated in CV-1 cells and purified through sucrose gradients.
The emulsified 19 F perfluorocarbon (PFC) solution (VS-1000H, Celsense, Inc., Pittsburgh, PA, USA) with a mean particle size of approximately 145 nm [18] was directly i.v. injected as 20% v/v emulsion (100 ml) into tumor-bearing mice. The injection time points of each individual experiment are indicated in the corresponding figure legend.

Ex vivo MRI
After completion of the in vivo experiments the 1936-MEL tumors were excised and snap-frozen in liquid N 2 and stored at 280uC. For ex vivo 19 F MRI measurements followed by histology the tumors were fixed in 4% paraformaldehyde/PBS pH 7.4 for 16 h at 4uC and rinsed in PBS followed by embedding and dehydration in 10% Sucrose/5% w/v low-melting point agarose/ PBS.
In a first set of experiments, focusing on the 1936-MEL tumor model, all infected tumors (n = 3) were placed together in a single 15 ml tube (Greiner Bio-One GmbH, Frickenhausen, Germany) embedded in 10% Sucrose/5% w/v low-melting point agarose/ PBS. The corresponding control tumors (n = 3) were placed in an additional 15 ml tube. Both tubes were placed together in the measurement coil to enable measurement of all tumor specimens in only one 1 H and one 19

Post Processing of MRI Datasets
Post processing was performed using home-written routines in MATLAB (The MathWorks Inc., Natick, MA, USA) if not mentioned otherwise.
Figure preparation. For anatomical correlation of the in vivo A549 19 F data, a summed image of all echo images was calculated for each 1 H MSE experiment. Prior to summation, the echo images were pixel-wise weighted with the calculated T 2 time assuming a mono-exponential decay (S(t) = S 0 exp(2t/T 2 ); S(t) = signal; t = time).
All 19 F datasets were zerofilled to the matrix sizes of the respective 1 H datasets. The SNR of the zerofilled 19 F data were calculated following reference [35] for low SNR data. For better visualization the maximum of the shown 2D images of the 3D 19 F datasets were scaled to a specific maximal SNR which is indicated in the respective figure legends. For 1 H/ 19 F overlay images a SNR threshold of 4.5 was chosen for the 19 F data. Additionally, remaining, isolated 19 F signal pixels were removed.
Regarding the 3D overlay reconstruction, 19 F data were zerofilled as described above. The 3D 1 H and the zerofilled 19 F data were scaled to their respective maxima. The zerofilled 3D 19 F data were set to a threshold as described above. Afterwards a 3D overlay dataset was generated. The overlay dataset was additionally masked by only selecting the animal area. The animal mask was manually generated with the help of the 1 H 3D data. After the described pre-preparation of the 3D overlay data in MATLAB, the data were transferred to MeVisLab (MeVis Medical Solutions AG, Bremen, Germany) where the 3D surface reconstruction was performed. For the visual correlation of the ex vivo MRI data and histology results only tissue slices from the tumor middle were compared.
Quantitative analysis of the 19 F MRI data. For quantitative analysis of the in vivo MRI time course in 1936-MEL tumorbearing mice, ratios of the total 19 F SNR of the tumor and the tumor volume ( 19 FTVR) were generated. Since from each group only n = 2 animals were measured at 9 dpi, this parameter was only calculated for 7 and 11 dpi (n = 4 animals per group). The tumor regions were manually selected with the help of the fully resolved 1 H images. SNR maps of the zerofilled 19 F data were generated as described above. Only 19 F signal pixels within the predefined regions-of-interest (ROI) having an SNR$4.5 were regarded and single 19 F signal pixels were removed. The total 19 F SNR was evaluated by summation of all pixels remaining after the pre-preparation/2selection procedure. The 19 F SNR data were additionally normed to the mean SNR of the corresponding reference. For each animal the corresponding 19 FTVRs of 7 and 11 dpi were individually regarded. Thus, for each pair (day 7 and 11), the 19 FTVRs were normed to the maximum value and the slope between both values was calculated assuming a linear relationship. Similar as for in vivo MRI of 1936-MEL tumor-bearing animals, ratios of the total 19 F SNR of the tumor and the tumor weight ( 19 FTWR) were generated for in vivo MRI of A549 tumors and ex vivo MRI of 1936-MEL/GI-101-A tumor-bearing animals. Thus, SNR maps of the original 19 F data were generated for quantitative analysis as described above. Overlays were generated with the help of the original 19 F data ( 19 F SNR$4.5) and downscaled 1 H data. This was done to select ROIs which excluded 19 F signal from other regions (e.g. lymph nodes).
Additionally, for VACV infected animals having 1936-MEL tumors a correlation of the ex vivo 19 F signal area with the GFP positive area was performed. Thus, for each tumor sample, the middle slice of the tumors was chosen using the original 19 F data. Furthermore, ROIs of the tumor region were manually selected with the help of downscaled 1 H data. In a next step, the number of signal pixels in the ROIs having a SNR$4.5 was calculated. Single 19 F signal pixels were removed. In a final step, the percentage of the 19 F signal containing volume in regard to the tumor volume of the selected slice was calculated.

Immunohistochemistry
Following ex vivo 19 F MRI measurements histology was performed. Briefly, the paraformaldehyde-fixed and in 10% Sucrose/5% w/v low-melting point agarose/PBS embedded tumors were further dehydrated in 30% sucrose/PBS for 12 h and finally embedded in Tissue-TekH O.C.T. (Sakura Finetek Europe B.V., Alphen aan den Rijn, Netherlands). Tumor samples were sectioned (15 mm) with the cryostat 2800 Frigocut (Leica Microsystems GmbH, Wetzlar, Germany). Antibody-labeling was performed following fixation in ice-cold acetone. The primary antibody was incubated for 1 h. After washing with PBS, sections were labeled for 30 min with the secondary antibody and finally mounted in Mowiol 4-88.

Fluorescence Microscopy
The fluorescence-labelled preparations were examined using the MZ16 FA Stereo-Fluorescence microscope (Leica) equipped with the digital DC500 CCD camera and the Leica IM1000 4.0 software (130061030 pixel RGB-color images) as well as the Leica TCS SP2 AOBS confocal laser microscope equipped with an argon, helium-neon and UV laser and the LCS 2.16 software (102461024 pixel RGB-color images). Digital images were processed with Photoshop 7.0 (Adobe Systems, Mountain View, CA, USA) and merged to yield overlay images.

Fluorescence Intensity and Immune Cell Density Measurements
The fluorescence intensity of the CD68-labelling or the percentage of the CD68-positive tumor area in 15-mm-thick cryostat sections of control and GLV-1h68-colonized tumors was measured on digital images (x10 objective or x40 objective, x1 ocular) of specimens stained for CD68 immuno-reactivity. On the fluorescence microscope, the background fluorescence was set to a barely detectable level by adjusting the gain of the CCD camera before all the images were captured with identical settings. The fluorescence intensity of the CD68-labelling was determined in RGB images and represented the mean brightness of all pixels (intensity range 20-255) and was measured using ImageJ software (http://rsbweb.nih.gov/ij). The whole area of the tumor crosssection was determined by manually drawing ROIs using ImageJ.
The immune cell density was determined either in microscopic images of whole tumor cross-sections (x10 objective, x1 ocular) by quantification of the positive labelled area using ImageJ software or by direct counting of CD68-positive immune cells in multiple areas of different tumor sections obtained with the confocal laser microscope (x20 objective, x10 ocular). On the fluorescence microscope, for each image the fluorescence signal was set to a clearly detectable level by individually adjusting the gain of the CCD camera before the images were captured. All RGB-images were equally adjusted and converted into 8-bit gray scale images using Photoshop 7.0. Positive pixels (intensity range 0-255) of the CD68-labelling were measured using ImageJ software. The immune cell density was shown as positive area/tumor section for whole tumor sections or as number of positive cells/field of view and presented as mean values with standard deviations.
The Cy3-conjugated secondary antibodies (donkey) were obtained from Jackson ImmunoResearch (West Grove, PA, USA). Hoechst 33342 (Sigma Aldrich) was used to label nuclei in tissue sections.

Statistics
A two-tailed Students t test was used for statistical analysis of different quantitative parameters described above. Pearsons rank correlation was used to determine the correlation coefficient r. p values of ,0.05 were considered statistically significant.

PFC Accumulates in 1936-MEL Melanomas during Oncolytic Virotherapy
To test the applicability of 19 F MRI to monitor virus-induced inflammation during oncolytic virotherapy, we used 1936-MEL melanoma-bearing mice and administered VACV (GLV-1h68) i.v. followed by i.v. injection of emulsified PFC at day 4 and day 6 pi and MR imaging at 8 dpi (Fig. 1A). The tumor area was localized by 1 H images followed by acquisition of anatomically matching 19 F images in mock-infected control (Fig. 1B, n = 2) as well as VACV-treated animals (Fig. 1C, n = 2). The corresponding 1 H/ 19 F overlay in vivo images revealed intense hot spots of the 19 F signal located at the tumor periphery in VACV-colonized animals compared to only few 19 F-positive hot spots in the mockinfected control animals. The 3D 1 H/ 19 F overlay reconstruction of the abdomen of a VACV-infected mouse clearly showed an accumulation of PFC at the tumor margin (partial/hollow sphere) enabling tumor localization as well as localization of the adjacent lymph nodes (Fig. 1D). Ex vivo 1 H/ 19 F MRI imaging of excised tumors followed by CD68-and Ly-6G-immunohistology revealed a similar spatial distribution pattern of the 19 F signal and CD68positive monocytes/macrophages in both groups of mice (Fig. 1E, F, n = 3). However, in VACV-colonized tumors the macrophagecontaining hot spots ( 19 F + /CD68 + ) accumulate around the viral infected GFP-positive areas in the tumor periphery, whereas in control animals the CD68-positive macrophages and the corresponding 19 F signal were generally distributed with lower signal intensity throughout the tumor tissue. Moreover, there were also Ly-6G-positive neutrophils located around the viral infected GFPpositive areas forming the inner part of the invading immune cell front. Analysis of the CD68-fluorescence signal intensity of whole tumor cross-sections by histology revealed a significantly enhanced fluorescence intensity in VACV-infected compared to mockinfected tumors (Fig. 1G; n = 3, p = 0.026).
The detailed histological analysis further showed that the 19 Fpositive/CD68-positive areas were also positive for CD11b and MHCII which confirm the affiliation to monocyte/macrophage populations ( Figure S1).
Altogether, the results demonstrate that non-invasive 19 F MRI can specifically visualize VACV-mediated spatial changes in the myeloid cell populations of tumors which may either due to recruitment or redistribution of tumoral myeloid cells.
PFC and CD68 + -macrophage Accumulation in 1936-MEL Melanomas during the Early Time Course of Infection (0-8 dpi) To identify the timing of PFC accumulation in 1936-MEL melanomas after VACV-treatment, we performed a time course study of PFC accumulation during early infection stages (0, 2, 4, 6, and 8 dpi). Each group was injected with one dose of PFC 4 days before tumors were harvested and ex vivo MRI was performed ( Fig. 2A). Immediately before (0 dpi) and 2 days pi no 19 F signal was detectable by MRI (Fig. 2B). However, 4 days pi the first 19 Fpositive patches were detectable in the VACV-treated tumors which further increased at day 6 and 8 pi (Fig. 2B). Interestingly, in contrast to the 4 dpi tumors the distribution pattern of the 19 F hot spots of the 6 and 8 dpi tumors differed and 19 F hot spots were mostly located in close proximity to the tumor rim. Therefore, for robust detection of viral-induced 19 F accumulation we suggest 2-4 dpi as the optimal injection time point for PFC in 1936-MEL tumor-bearing mice followed by 19 F MR imaging 6-8 days post VACV injection.
To confirm the MRI results, we performed immuno-histological analysis of the CD68 + -macrophage population in tumor sections at all time points. CD68-positive cells were detectable in all investigated tumors with no significant difference of the total CD68-positive tumor area during the infection time course (Fig. 2C, F). Interestingly, the distribution pattern of the CD68positive cells was notably different in 0-4 dpi and 6-8 dpi tumors changing from an intra-and peritumoral scattered pattern to a strong accumulation of macrophages at the tumor margin. Detailed analysis of the intratumoral and the peripheral tumor areas revealed a significant increase in the CD68 + -macrophage population in the tumor margin concomitant with increasing viral spreading (Fig. 2D, G) and a significant loss of the intratumoral CD68 + -macrophage population (Fig. 2E, H).

Long-term Imaging of PFC Accumulation -a Spatiotemporal Analysis
To analyze the stability and/or the spatial distribution of the 19 F signal pattern over time, we injected PFC only once at day 4 pi in mock-and VACV-treated groups of 1936-MEL tumor-bearing mice and performed 19 F MRI at day 7, 9, and 11 dpi. During these time points tumor growth arrests in VACV-treated animals in contrast to the continuously increasing tumor volume of the mock-infected group (Fig. 3A). In each group the 19 F signal distribution pattern was similar during the imaging period (Fig. 3B). However, the 19 FTWR (total 19 F SNR of the tumor/ tumor weight) was significant higher in VACV-treated at 11 dpi compared to mock-treated tumor-bearing animals as determined by ex vivo 19

Non-macrophage-related PFC Accumulation in Large, Mock-infected Control Tumors
During our study, we noticed that some large, mock-infected tumors can incorporate substantial PFC amounts after intravenous PFC application. This could interfere with the reliable detection of virus-mediated immune cell recruitment. However, this group of mainly large tumors showed a different distribution pattern of the 19 F signal compared to the previously shown scattered intratu-moral 19 F pattern of mock-infected control tumors as shown in Fig. 1E. The analysis of different 1 H/ 19 F overlay axial slices of the abdomen revealed that the 19 F signal accumulation also appeared in non-tumorous areas and was mostly located outside of the malignant tissue in abdominal cross-sections (Fig. 4A). Interestingly, the investigation of the phenotype of these large, mockinfected tumor-bearing animals revealed that these mice developed peritumoral hematoma and even bloody tumors in comparison to VACV-treated tumor-bearing mice (Fig. 4B). Direct comparison of the 1 H/ 19 F signal pattern of large, mockinfected tumors and VACV-infected tumors by ex vivo MRI showed that the 19 F signal accumulated in the outermost part of the tumor rim in the mock-infected animals, whereas 19 F hot spots accumulated deeper into the tumor tissue in VACV-treated animals (Fig. 4C). Histological analysis revealed a similar pattern of CD68 + -macrophage populations and 19 F hot spots only for the VACV-treated tumors and no matching was found for mockinfected tumors (Fig. 4D).

F-MRI-based Imaging of VACV-induced Recruitment of Macrophages in Different Tumor Models
Since different tumor types as well as individual tumors of the same origin can greatly vary in their content of phagocytic macrophages, the polarization of the macrophage status and the immune-related response to therapy [36,37], we decided to test our concept in different tumor models. Thus, we chose the A549  lung carcinoma model bearing few macrophages as well as the GI-101A breast adenocarcinoma model harboring a substantial population of tumor-associated macrophages before treatment (unpublished results). Despite these immunological differences, in both tumor types a significant increase in the 19 F signal due to VACV-treatment was detectable in comparison to the mockinfected animal group (Fig. 5, Figure S2).
In addition we wanted to emphasize, that the use of 19 F PFC as cellular contrast agent should not prevent the utilization of quantitative 1 H parameters. Exemplarily, 1 H T 2 maps were acquired for the A549 lung carcinoma model (Fig. 5B, D)  indicating the possible integration of other quantitative 1 H MRI parameters to facilitate tumor characterization [38].
In summary, the 19 F-MRI-based imaging of tumoral VACVcolonization is applicable in different tumor models and possesses high potential for monitoring oncolytic therapy. However, the great diversity of the microenvironment of different tumor models and even individual tumors implies for future work a specific and thus personalized imaging modality applying 19 F MRI before and after starting the oncolytic treatment.

Discussion
We have shown that the VACV-induced infiltration of myeloid cells into different tumor models is reliably detectable by 19 F MRI. Therefore, we suggest that 19 F MRI may be used in the future as a novel tool to quantitatively and non-invasively monitor the innate immune response in tumors following oncolytic virotherapy. Thus, this imaging technique can be readily investigated as a surrogate measure to monitor viral tumor colonization and also therapeutic response. To the best of our knowledge, this is the first report of utilizing PFC nanoparticles in combination with 19 F MRI for detection of myeloid cell infiltration into tumors undergoing oncolytic virotherapy. Other investigators have reported the use of PFC nanoparticles to monitor inflammatory conditions in different other pathological situations [13][14][15][16][17][18]. Recently, Kleijn [39] used a Gadolinium (Gd)-based agent targeting the inflammatory enzyme myeloperoxidase (MPO) in a similar approach to detect oncolytic virus-associated tumor inflammation by MRI. However, 19 F MRI has several advantages compared to contrast agent based 1 H MRI including no background in tissues [20], the possibility of direct quantification of the marker amount [24] and the unaltered 1 H signal making the assessment of other quantitative 1 H parameters more feasible [18].
In the present study, 19 F MRI as well as the histological examination of mock-infected tumors revealed a diffuse distribution of both the 19 F signal and the CD68 + -macrophages throughout the tumor. In this respect, the so called enhanced permeability and retention (EPR) effect, which is a microenvironmental characteristic of solid tumors leading to the passive and unspecific accumulation of a variety of macromolecules and nanoparticles, may enhance the accumulation of PFC in large mock-infected control tumors [40]. However, the direct comparison to VACV-treated tumors clearly showed that the viral tumor colonization significantly altered the 19 F signal distribution as well as the signal intensity. In VACV-treated tumors the 19 F-positive hot spots, which showed a similar distribution pattern as the CD68 + -macrophage population, encapsulated the viral infection focus and formed a ''partial/hollow sphere'' localized to the tumor rim. In this regard, 3D-reconstructions of the abdominal region of VACV-treated tumor-bearing mice enable a straightforward localization of the tumorous regions. Since GLV-1h68 preferentially colonizes metastases [41] this imaging modality may be useful to detect metastases. In accordance with previous findings [13,17], we further observed a strong 19 F signal in the adjacent lymph nodes of tumors in 1936-MEL-bearing animals also indicating the potential of this imaging modality for the localization of sentinel lymph nodes. In future studies, we will especially analyze 19 F-positive lymph nodes and clarify whether these lymph nodes are already metastasized or sites of an ongoing inflammatory immune cell activation.
Recent studies combining blood density gradient centrifugation and 1 H/ 19 F MRI showed that i.v. applied PFCs are predominantly taken up within the blood stream by monocytes but to a minor degree also by B-cells and neutrophils [13,15]. After uptake, especially PFC-labelled neutrophils and macrophages migrate to sites of ongoing inflammation, where monocytes/macrophages are the predominant PFC-labelled cell fraction [15,19]. In 19 F-positive hot spots of VACV-treated tumors both monocyte-derived macrophages (CD68, CD11b, MHCII) as well as neutrophils (Ly-6G) were detectable, however, clearly spatially separated in different areas of the invading immune cell front. Neutrophils were localized in the inner part of the immune cell immigrants, whereas macrophages were located in the surrounding outer area. Since neutrophils are the first actors at the infection site followed by monocytes/macrophages [42] the here observed accumulation pattern of both immune cell populations implies that both populations are recruited in a timely different fashion to the VACV-infected tumor rather than re-distributed within the tumor tissue.
The time course study from 0 to 8 dpi based on 19 F MRI and histological analysis of VACV-treated tumors revealed discrepancies between the spatio-temporal 19 F accumulation and the CD68 + -macrophage population during the course of infection. The 19 F signal was first detectable at 4 dpi further increasing in the tumor rim at day 6 and 8 pi, whereas the CD68 + -macrophage population was at all investigated time points detectable, however, either distributed throughout the whole tumor (0-4 dpi) or mostly located with a high density at the tumor rim (6-8 dpi). Further, the microscopic analysis revealed that the intratumoral macrophage population significantly decreases whereas the peritumoral population increases during the course of infection and with increasing viral spreading. These results indicate that the resident TAM population may be directly eliminated from the tumor tissue by viral infection and simultaneously a second VACV infectioninduced population of 19 F-positive macrophages was recruited from the circulation to the tumor encapsulating the infection focus. The lack of the 19 F accumulation early after infection may be responsible either to the lower phagocytic activity of the resident macrophage population [37] or to the reduced vessel permeability in these tumors avoiding significant intratumoral accumulation of PFC after injection. Since the presence of TAMs in several cancer types correlates strongly with a poor outcome [43] studies were already performed to develop a clinically applicable, non-invasive diagnostic assay for visualization of TAMs in tumors based on MRI and clinically applicable iron oxide nanoparticles [44]. In the present study, however, the results indicate that, rather than labelling the intratumoral TAMs, PFC nanoparticles may label mainly the immune cells in the circulation which immigrate into the tumor after viral colonization.
The observed decrease of the 19 FTVRs during the time course study could be mainly explained by the continuous tumor growth in the mock-infected group. However, in VACV-treated animals the average tumor volume remained stable during the imaging period. Therefore, the decrease in the 19 FTVRs in VACV-treated mice may be either due to an emigration of PFC-labelled macrophages to e.g. draining lymph nodes, which could also explain the observed 19 F signal in the sentinel lymph nodes (Fig.1  D) or to the loss of freely accumulated PFC via efferent lymphatics. For the immuno-therapeutic effect of oncolytic virotherapy, emigration of phagocytic cells from the tumor to adjacent lymph nodes may offer a way to activate an adaptive anti-tumoral immune response via tumor antigen presentation to B-and T cells and should be further investigated in future studies.
The examination of the 19 F-positive area and the extent of the viral spreading determined by the GFP-expression revealed a significant positive correlation for both parameters indicating that 19 F MRI during oncolytic virotherapy enables indirect monitoring of viral replication. This would be beneficial for non-invasive monitoring of intratumoral viral replication in pre-clinical and in future clinical studies, since 19 F MRI provides a whole organ visualization circumventing the problems associated with small sample biopsies such as false-negative or -positive results of viral replication. For future studies using 19 F MRI for monitoring of oncolytic virotherapy, we suggest a specific personalized imaging modality applying 19 F MRI before and after starting the oncolytic treatment to avoid misinterpretation of unspecific PFC accumulation as we have shown in tumor-associated hematoma.
For correlation purposes only tissue slices from the tumor middle and the same orientation were always chosen for histological sections and ex-vivo 19 F images. However, due to the different slice thickness and possible alignment errors of the histological section and the MRI data, the correlation remains limited. In the future, as has been done with 1 H MRI [45], sophisticated 19 F MRI coils should be developed to improve the correlation of 19 F MRI and histology.
Since we could show that 19 F MRI enables the detection of viral tumor colonization and immune cell recruitment to tumors in different tumor types with different immunological context, we assume that the here described imaging modality may also be useful in the future for clinical applications. Thorne previously discussed that it is also likely that systemic measurements of the level and type of immune response induced by the viral treatment may ultimately be used either as an early prognostic indicator of therapeutic response, or may help elucidate the immune properties of the tumor being treated, and so assist in the design of subsequent immunotherapeutic treatments [46]. In the same manner, we propose that 19 F MRI may have the potential to be used for therapeutic monitoring and as prognostic indicator for the therapeutic outcome. Figure S1 Co-localization of the 19 F signal with monocytes/macrophages and neutrophils. Representative ex vivo 1 H/ 19 F overlays as well as corresponding histologically prepared tumor sections demonstrating a similar distribution pattern of the CD68 + -(monocytes/macrophages), MHCII + -(antigen-presenting cells such as dendritic cells (DCs), macrophages and B cells), CD11b + -(myeloid cells), Ly-6G + -population (neutrophils) and the 19