19 F-Magnetic Resonance Imaging for Tracking Bone-Marrow Macrophages in a Model of Experimental Autoimmune Myocarditis: A Pilot Study in a Model of Experimental Autoimmune A Pilot Study

In recent years, 19F-MRI has emerged as an attractive method to
detect perfluoro-carbon (PFC) compounds non-invasively in animal
experiments...


Introduction
In recent years, 19 F-MRI has emerged as an attractive method to detect perfluoro-carbon (PFC) compounds non-invasively in animal experiments [1,2]. Because 19 F background signal is effectively absent in animals and humans, 19 F-MRI can unequivocally detect an exogenous 19 F compound with high specificity of the signal. As the signal received by an MR volume coil with a homogenous B 1 field is directly proportional to the amount of 19 F nuclei present in the tissue under experimental conditions allowing uniform nuclear spin relaxation and excitation, the signal can be related to a reference of known concentration, rendering this technique quantitative [3,4]. Moreover, unlike molecular imaging methods based upon positron emission or hyperpolarized carbon-13, non-volatile PFC compounds are not limited by signal decay over time, and the time window for their detection can therefore last several days. Finally, 19 F-MRI images can be merged with conventional 1 H-MRI images to match the 19 F signal with its exact anatomic location in the body and to also correlate it with function or other tissue characteristics.
Accordingly, 19 F-MRI has been used successfully to detect and track well-defined cell populations in rodent models of inflammation, including myocardial infarction [5][6][7], cerebral ischemia [6], pneumonia [8], atherosclerosis [9], arthritis [10], and tumors infiltrated by macrophages [11]. In a recent study, we successfully visualized heart-infiltrating macrophages in-vivo by 19 F-MRI in a model of experimental autoimmune myocarditis (EAM) after IV injection of a PFC emulsion [12]. The EAM model is a CD4 + T cell-mediated disease which closely resembles the phenotype of human myocarditis, including its progression toward inflammatory dilated cardiomyopathy [13]. Inflammatory cardiomyopathy, on the other hand, is an important cause of heart failure in young patients and its prevalence is most likely largely underestimated [14]. Cardiac inflammation in EAM typically peaks approximately 21 days after immunization, as assessed by hematoxylin and eosin histological staining of inflammatory infiltrates, followed by a transition to diffuse interstitial fibrosis and end-stage heart failure [15].

Our earlier study demonstrated by immunohistochemistry
and flow cytometry the presence of PFC in macrophages within the inflamed myocardium [12]. However, the precise origin of these PFC-labeled macrophages remains unknown since the IVinjected PFC can be taken up by either circulating monocytes or by macrophages which differentiate from heart-infiltrating or resident immature precursor cells in the inflamed myocardium. Several lines of evidence suggest an immunosuppressive role of myeloid cells and macrophages in regulating disease severity and outcome in myocarditis [16,17]. If circulating cells are responsible for a major part of cardiac infiltrates detected by 19 F-MRI in the EAM model, it should then be possible in principle to follow the migration of labeled monocytes or macrophages to the inflamed myocardium.

Recent reports have demonstrated that T cells, splenocytes, stem cells, and endothelial cells can be labeled with PFC in-vitro
and then re-injected to be tracked by 19 F-MRI [4,[18][19][20][21][22]. The aim of this study was to determine the feasibility of using this method to track macrophage migration in the EAM model as a proof-ofconcept. Bone-marrow-derived macrophages (BMM) were labeled with PFC in-vitro and re-injected into EAM mice in the early phase of inflammation, before the peak, for cell tracking by 19 F-MRI. 19 F-MRI was performed on the heart as well as the lungs, spleen and liver to best characterize the fate of the injected PFC-labeled BMM, and the results were confirmed with ex-vivo 19 F-NMR spectroscopy, immunohistochemistry, and flow cytometry experiments.

Animals
Mice were maintained under specific pathogen-free conditions. EAM was induced in 52 male BALB/c (CD45.2) mice by subcutaneous injections of αMyHC peptide (alpha myosin heavy chain, Ac-RSLKLMATLFSTYASADROH; Caslo, Lyngby, Denmark) emulsified 1:1 with complete Freund's adjuvant (Difco, Franklin Lakes, NJ) as previously described [13]. Immunization with the peptide was performed at days 0 and 7, following the in-vivo protocol timeline depicted in Figure 1. In the EAM model, immunization results in acute myocarditis that peaks ~21 days after immunization.
Inflammation resolves slowly thereafter, whereas the number of cardiac fibroblasts progressively increases [15].

In-vitro 19 F-NMR Spectroscopy of Labeled Cells
In order to measure the mean 19    (A) "Long time" protocol. In-vitro: Cell preparation. Bone marrow cells were isolated from femurs of CD45.1 mice (n=2) and cultured as described in the Materials and Methods section. In-vivo: At day 0 (D0) and D7, BALB/c mice (n=6) were immunized to induce EAM. At D20 the mice received an IV injection of 19 F-labeled-BMM (from 4 to 10 × 10 6 cells /mouse). All the mice were analyzed by in vivo 1H-and 19 F-MRI at D21 and also either at D24 (n=3) or D28 (n=3). These mice were euthanized immediately after the MRI session and their hearts were analyzed by histology.

BMM Transfer into EAM Mice
(B) "Short time" protocol. In-vitro: Cell preparation. At D0 bone marrow cells were isolated cultured as before (9 mice). In-vivo: At D0 and D7, BALB/c mice (n=46) were immunized as before. At D9 all the mice received an IV injection of 19 F-labeled-BMM (from 2 to 10 × 10 6 cells /mouse). Mice were scanned and / or analyzed 36 hours after BMM injection.

Organ Collection
The mice were euthanized to collect the organs after MRI scanning. Thirteen of the hearts were immediately analyzed by 19 F-MRS prior to being embedded in optimal cutting temperature (OCT) compound for immunohistochemistry. The liver, lungs, and spleen of 8 unscanned mice were also collected and divided for FACS analysis and immunohistochemistry, the latter set embedded in OCT; their bone marrow was also collected. The hearts of 10 other mice were also embedded in OCT. The liver, lungs, and spleen of 11 unscanned mice were collected for ex-vivo 19 F NMR spectroscopic analysis.

Ex-vivo 19 F-NMR Spectroscopy of Excised Organs
The 19 F NMR measurements were performed on the homogenates prepared from liver, lungs and spleen in 1% (v/v) Triton X-100 in PBS. The cell lysates were then mixed with 250 and placed in a 5 mm borosilicate glass NMR tube. Spectra were acquired as described above in the "In-vitro 19 F-NMR spectroscopy of labeled cells" section.

Immunohistochemistry
OCT-embedded snap frozen tissues (heart, liver, lung and spleen) were sectioned in a cryostat at 6 µm thickness. Slides were air dried and fixed in acetone for 10 min at room temperature.

Statistical Analyses
Values are given as mean ± standard deviation. Analyses of differences between groups were performed using one-way analysis of variance (ANOVA) with Bonferroni correction for multiple comparisons. Normality was evaluated with the Shapiro-Wilk test.

Detection of 19 F Signal by 19 F-MRI
Protocols with early (day 9) and late (day 20) 19 F-labeled cell injection and cardiac follow-up 19 F-MRI studies were applied ( Figure 1). The late injection protocol followed a similar timeline as in our previous study where PFC was directly IV administered into EAM mice [12]. In 6 mice, PFC-labeled BMM were injected IV at day 20 after the first immunization of recipient mice, and invivo cardiac 19 F-MRI was performed at days 21, 24 and 28. Under these conditions, no 19 F-signals were detected in the heart by 19 F-MRI (data not shown). Given that heart-infiltrating T cells and monocytes in the EAM model appear as early as 7-10 days after the first immunization, we decided to change the time course of our experiments. In the early series, 19 F-labeled BMM were injected at day 9, and in-vivo cardiac 19 F-MRI was performed at day 11.
Although 19 F signal of variable intensity was detected in all animals by spectroscopy, the MRI signal detected in the heart was weak or absent (Figure 2A, arrowhead) and the majority of the 19 F-signal was identified in the lungs (solid arrow) and the liver (dashed arrow).
These in-vivo data were confirmed by post-mortem analyses of the organs. Hearts were analyzed ex-vivo by 19 F-MRS (n=13), and a faint spectral peak was detected within the combined cardiac tissue of 4 animals ( Figure 2B) 19 F recovered in the organs was significantly higher in the liver than the lungs (p = 0.0096, Figure 3), even though the liver 19 F content was below the limit of detection in two unscanned mice injected with weakly-labeled BMM (6.49 × 10 11 and 2.37 × 10 11 19 F atoms/cell). 1H-signal appears in grey, 19F-signal in orange, and the heart, liver and lungs can clearly be identified. The highest signal was detected in the liver (dashed arrows) and in the lung (solid arrows). Heart signal is weak (arrowhead) and patchy, in agreement with a patchy infiltration of the EAM heart previously reported [12].   19 F label in lungs, liver and spleen. The 19 F content of each organ is divided by the sum of the 19 F content of the three organs in a mouse. Samples with 19 F below the limit of detection by NMR were assigned the value of zero. The liver had on average the highest fraction of recovered 19 F, which was significantly higher than the lung fraction.

Immunohistochemistry
The presence of CD45.1 donor cell infiltrates was identified by immunohistochemistry in 12 out of 17 hearts, including 10 mice that underwent the same treatment but were not scanned. CD45.1 infiltrates were also found by immunohistochemistry in the liver, the lungs, and the spleen of the 3 mice analyzed (Figure 4).

Flow Cytometry
To confirm the immunohistochemistry results, flow cytometry of the CD45.1 marker was used to follow the distribution of the donor BMM in the different organs ( Figure 5). Flow cytometric analysis confirmed the presence of the injected CD45.1 cells in the organs, including bone marrow (6.2 ± 0.2 % of the gated cells), spleen (0.4 ± 0.03 %), liver (0.2 ± 0.04 %) and lungs (0.4 ± 0.2 %).

Cardiac Tissue
The presence of the PFC-labeled BMM in the liver, lungs, and spleen was readily apparent by 19 F-MRI and confirmed by ex-vivo 19 F-NMR spectroscopy. This pattern of accumulation is reminiscent of the one previously reported for labeled BMM tracked by singlephoton emission computed tomography (SPECT) and T 2 *-weighted MRI, where the cells were primarily located in the liver and spleen after initially being found mainly in the lung [25]. The similar pattern suggests that these two different labeling methods do not affect the BMM migration behavior, but a functional effect due to the labeling cannot be ruled out.

Weak Invasion of the Myocardium in the EAM Model by BBM
In earlier studies, direct IV injection of PFC nanoemulsions successfully labeled macrophages in inflamed myocardium of the EAM model [12] and in other models of myocarditis [26]. In the current study, however, PFC-labeled BMM, when injected IV, did not migrate to the site of inflammation in the heart at a high level and were rarely detectable there by in-vivo 19 F-MRI or ex-vivo 19 F-NMR spectroscopy. In-vivo 19 F-MRI and ex-vivo 19 F-NMR spectroscopy detected a 19 F signal in the heart in 2 out of 18 and 1 out of 6 experiments, respectively. However, in-vivo and ex-vivo signals were not detected in the same mice, and the apparent in-vivo 19 F cardiac signal may be due to partial-volume effects from the neighboring lung and liver. In an earlier study [18], the same in-vivo 19  in the inflamed myocardium [27][28][29] and also macrophages derived from recruited blood monocytes [30].
Reports from our laboratories clearly indicate an important role of heart-infiltrating bone-marrow-derived CD133-expressing immature monocyte-like precursor cells as a potential source of both myofibroblasts and macrophages in the inflamed heart [28,31,32]. On the one hand, the uptake of IV-infused PFC by phagocytic cells in the circulation, followed by their migration to the inflamed myocardium, provides a good rationale for the myocardial accumulation of the 19 F signal in previous myocarditis studies [12,26]. Bönner et al. have shown the ability of blood-derived leukocytes to take up PFC [33], which would support this route of myocardial PFC entry. On the other hand, it is possible for cells already present in the heart to take up infused PFC in quantities sufficient to be imaged. For example, Ding et al. demonstrated infiltration of the inflamed myocardium in an ischemia-reperfusion model not only by blood-derived monocytes but also by epicardium-derived cells (EPDC) [34].
These EPDC displayed strong endocytic activity to take up the IV-injected PFC emulsion and infiltrated the inflamed myocardium in a time-dependent manner, such that they represented the main source of 19 F-labeled cells 3 days after the ischemia-reperfusion event. If BMM migration is not affected by labeling, the present results would indicate that the injected BMM are unable to effectively infiltrate the heart in large numbers. This would be consistent with the view that accumulation of macrophages within the inflamed heart results either from differentiation of recruited or resident precursor cells or from proliferation of those already present.
Nevertheless, to better test this mechanism of PFC accumulation in the inflamed myocardium would require a separate protocol to track PFC-labeled monocytes in the murine EAM model.

Limitations
The techniques used in this study result in a number of limitations. One important limitation of the PFC-labeling technique is the considerable variability of the achieved loading, ranging from 2.37 × 10 11 to 5.34 × 10 13 19 F-atoms/cell, which translated into the variable number of injected BMM cells, ranging from 1 to 14 × 10 6 per animal. Overall, less 19 F was administered to the mice as PFC-labeled BMMs than in our previous myocarditis study with infused PFC, and this may limit sensitivity. Despite the >10fold higher average per-cell 19 F content compared to prior studies using IV-transferred PFC-labeled cells [18,20], there still was not sufficient sensitivity to detect the label in the target tissue in vivo.
Additionally, the in vivo MRI and ex vivo MRS were performed with a surface coil, which provides high sensitivity but makes absolute quantitation of the 19 F more difficult. To better quantitate the in vivo 19 F signal, a volume coil could be used with a reference standard in the field of view.
The sensitivity of the 19 F imaging sequence could also be further improved with a lower acquisition bandwidth, or by using a spoiled-gradient echo (SPGR), steady state at free precession (SSFP)-type method instead [21]. Although the procedure to differentiate bone-marrow derived macrophage precursor cells with L929-conditioned medium is a well-accepted method to generate a relatively uniform population of mature quiescent BMM in high yield [35], the phenotype of the transferred cells found in the liver, lungs, spleen, heart and bone marrow was not characterized beyond the presence of the CD45.1 marker. While PFC labeling did not affect BMM pre-infusion viability, further study is needed to understand whether it has any effect on cell migration to the inflamed myocardium in the EAM model.

Conclusion
BMM were successfully labeled in-vitro with PFC and tracked in-vivo by non-invasive 19 F-MRI. In the EAM model, non-invasive 19 F-MRI reliably demonstrates accumulation of IV-injected mature BMM primarily in the liver, spleen, and lungs, with little migration to the inflamed myocardium. Thus, this technique has the potential to non-invasively track the in-vivo migration of specific cell types in the EAM model. These findings suggest that either resident inflammatory cells, macrophages differentiating from heart-infiltrating immature precursors represent the main source of differentiated macrophages accumulating in the heart in myocarditis and point the way to further studies that directly address these questions.

Ethics Approval and Consent to Participate
All animal procedures were approved by the institutional ethics committee and performed in accordance with an authorization for