All animal experiments were approved by the National Animal Experiment Board in Finland and the Regional State Administrative Agency for Southern Finland. Autoimmune myocarditis was induced in rats as previously described (4). Briefly, eight male Lewis rats aged 6 to 8 weeks (weight in g, 287 ± 17) were immunized with subcutaneous injections of 5 mg/ml pig cardiac myosin (M0531; Sigma Aldrich) in an equal volume of complete Freund's adjuvant supplemented with 1 mg/ml heat-killed and dried mycobacterium tuberculosis (F5881; Sigma Aldrich) into the hock of the left foot twice 7 days apart (days 0 and 7). To enhance the immunization effect, intraperitoneal injection of 250 ng/ml pertussis toxin (P2980; Sigma Aldrich) was also made on day 0. Eight male control Lewis rats (weight in g, 293.2 ± 23) received the adjuvant alone into the hock of the left foot. All procedures were carried out under isoflurane anesthesia (1.5–2.5%) and buprenorphine (0.03 mg/kg) was being administered two times a day for 2 days after immunization for analgesia.

Positron emission tomography (PET) imaging with [⁶⁸Ga]Ga-NODAGA-RGD was performed 3 weeks after the first immunization (day 21). To visualize the myocardium, contrast-enhanced CT or [¹⁸F]FDG PET imaging was carried out. At 100 min after imaging with [⁶⁸Ga]Ga-NODAGA-RGD, rats were euthanized and the heart and other organs were excised, weighed, and analyzed for the biodistribution of [⁶⁸Ga]Ga-NODAGA-RGD utilizing a gamma counter (Triathler 3″; Hidex, Turku, Finland) (4). The heart was frozen and sliced into serial transverse 20 and 8 μm cryosections at 1 mm intervals from base to apex for autoradiography, histology, and immunostainings.

⁶⁸Gallium-labeled 1,4,7-triazacyclononane-1-glutaric acid-4,7-diacetic acid conjugated RGD peptide ([⁶⁸Ga]Ga-NODAGA-RGD) was synthesized according to previously described procedures (27). The radiochemical purity was ≥ 99% and the molar activity was 18 ± 7.3 GBq/μmol (n = 9). A non-targeted peptide [⁶⁸Ga]Ga-DOTA-E[c(RGEfK)]₂, where E[c(RGEfk)]₂ = glutamic acid-[cyclo(arginyl-glycyl-glutamic acid-D-phenylalanine-lysine)], was used as a negative control as previously described (28).

The rats were imaged using a small animal PET/CT device (Inveon Multimodality; Siemens Medical Solutions) after [⁶⁸Ga]Ga-NODAGA-RGD injection under isoflurane anesthesia as previously described (4). The rats were injected with 50.8 ± 2.8 MBq [⁶⁸Ga]Ga-NODAGA-RGD intravenously and either a 10 min static PET acquisition starting at 60 min post-injection (n = 4 immunized rats and n = 8 controls) or a 90 min dynamic PET acquisition starting at the time of injection (6 × 10 s, 4 × 60 s, 5 × 300 s, and 6 × 600 s frames, n = 4 immunized rats) was performed to study the kinetics of the tracer.

In order to localize the myocardium, a high-resolution CT was acquired in all animals immediately after the [⁶⁸Ga]Ga-NODAGA-RGD PET injection 300 μL intravascular iodinated contrast agent (eXia™ 160XL; Binitio Biomedical Inc.) as described earlier (4). In addition to CT, a 40 min static [¹⁸F]FDG (38.5 ± 2.9 MBq) PET acquisition starting at 20 min post-injection was carried out in 11 rats (n = 6 immunized rats and n = 5 controls) on day 20 to facilitate localization of the myocardium.

[⁶⁸Ga]Ga-NODAGA-RGD PET images were co-registered with CT or [¹⁸F]FDG images using Carimas 2.9 software (Turku PET Center, Turku, Finland), and regions of interest (ROI) with uniform size were defined in the left ventricle (LV) myocardium in [⁶⁸Ga]Ga-NODAGA-RGD images. In immunized rats, ROIs were drawn in myocardial areas roughly corresponding to inflammatory lesions in histology. In controls rats, ROIs were drawn in similar areas as in immunized rats. The uptake is expressed as mean standardized uptake value (SUV_mean) and was compared between inflammatory lesions of immunized rats and the non-inflamed myocardium of control rats. Blood pool activity was measured as the average of ROIs drawn in the LV cavity and inferior vena cava. Additional ROIs were drawn in the lung, liver, kidney, and foreleg muscle. Decay-corrected time-activity curves (TACs) were extracted from dynamic image data.

For histology, 20 μm cryosections were stained with H&E. For immunohistochemistry, adjacent serial 8 μm cryosections were stained with a monoclonal mouse anti-rat CD68 (1:1000, MCA341GA, Bio-Rad, Hercules, CA, USA) to detect macrophages, a monoclonal mouse anti-rat CD61 antibody (dilution 1:500, MCA1773, Bio-Rad, Hercules, CA, USA) to detect integrin β3 chain, a polyclonal rabbit anti-rat CD31 antibody (1:200, NB100-2284, Novus Biologicals, Centennial, CO, USA) to detect endothelial cells, and monoclonal mouse anti- α-smooth muscle actin (α-SMA) antibody (1:12,000, A5228-200, Sigma-Aldrich, St. Louis, MO, USA) to detect myofibroblasts.

Uptake of [⁶⁸Ga]Ga-NODAGA-RGD in the inflamed and non-inflamed myocardium was analyzed by ex vivo digital autoradiography of 20 μm tissue sections as previously described (4). Multiple ROIs were drawn in the inflamed and non-inflamed myocardium in co-registered autoradiographs and images of H&E stainings of the same sections using TINA™ 2.10f software (Raytest Isotopenmessgeräte GmbH., Straubenhardt, Germany). Inflamed myocardium was defined as the presence of inflammatory cell infiltrate and signs of myocyte necrosis in the H&E-stained section. Non-inflamed myocardium was defined as the absence of both inflammations in the H&E-stained section and CD68 positive macrophages in a serial tissue section. The whole area of histologically inflamed myocardium was sampled, whereas the non-inflamed myocardium was sampled using multiple, uniformly sized ROIs. The results are as average photo-stimulated luminescence per square millimeter (PSL/mm²).

To study the specificity of [⁶⁸Ga]Ga-NODAGA-RGD uptake in cardiac inflammatory lesions, uptake of a non-specific control peptide [⁶⁸Ga]Ga-DOTA-E[c(RGEfK)]₂ was analyzed in three immunized rats (weight in g, 310.3 ± 6.6). The rats were injected with 52.4 ± 0.8 MBq [⁶⁸Ga]Ga-DOTA-E[c(RGEfK)]₂ via the tail vein and euthanized at 90 min post-injection. Then, tracer biodistribution was analyzed and ex vivo autoradiography of 20 μm cryosections of the myocardium was performed as described above.

Formal power calculation was not performed for this exploratory study. However, uptake of [⁶⁸Ga]DOTA-RGD in healthy myocardium (SUV 0.19 ± 0.02) in a previous study (26) indicated that it was possible to detect at least 20% difference in tracer uptake with a sample size of 6 rats at 90% probability and the alpha value of 0.05. All data are expressed as mean ± SD. Statistical analysis was performed with GraphPad Prism 6 Software. Unpaired or paired Student's t-test was used for single comparisons between two groups with normally distributed data. The statistical significance threshold was P < 0.05.

Histological analysis showed that of the eight immunized rats, seven (78%) developed wide-spread cardiac inflammatory lesions, whereas none of the eight control rats developed any inflammatory lesions (Figure 1).

The inflammatory lesions were characterized by intense inflammatory cell infiltration and myocyte necrosis. Consistent with our previous findings in this model (4), the immunohistochemical staining revealed that CD68-positive macrophages are the most abundant inflammatory cells in the lesions. Positive α-SMA staining indicates the presence of myofibroblasts and CD31 positive staining shows the presence of capillary network within the inflammatory lesions (Figure 1). CD61 staining co-localized in the same areas with CD31-positive staining and there were capillary-like structures showing positive staining (Figure 1).

In vivo [⁶⁸Ga]Ga-NODAGA-RGD PET/CT images showed visually increased tracer uptake in the myocardium of all immunized rats with histological inflammation (n = 7), while there was no visible tracer uptake in the myocardium of the immunized rat without inflammation (n = 1) or control rats (n = 8) (Figure 2).

Analysis of the kinetics of [⁶⁸Ga]Ga-NODAGA-RGD uptake based on TACs demonstrated that tracer uptake in the inflamed myocardium was most stable after 30 min post-injection (Figure 3A).

The uptake of the [⁶⁸Ga]Ga-NODAGA-RGD at 60–80 min post-injection was significantly higher in the inflamed myocardium compared to the non-inflamed myocardium of control rats (SUV_mean, 0.4 ± 0.1 vs. 0.1 ± 0.02; P = 0.00006, Figure 3B). The blood SUV_mean, was 0.3 ± 0.1 (p = 0.09 vs. inflamed myocardium) and the uptake ratio between inflamed myocardium and blood was 1.4 ± 0.5.

The results of [⁶⁸Ga]Ga-NODAGA-RGD biodistribution at 100 min are shown in Table 1. Tracer uptake was higher in the myocardium than in blood in both immunized and control rats (P < 0.01). Compared with control rats, tracer accumulation was higher in the hearts and lymph nodes (mediastinal and axillary) in the immunized rats. Notably, in addition to the heart, inflammation was evident in the lymph nodes of immunized rats by histology (Supplementary Figure 1).

Autoradiography showed focally increased [⁶⁸Ga]Ga-NODAGA-RGD uptake in the LV myocardium of immunized rats colocalizing with histological inflammatory lesions. Tracer uptake was low in the non-inflamed myocardium of immunized rats or myocardium of controls. Quantitatively, the average [⁶⁸Ga]Ga-NODAGA-RGD uptake in the inflamed area was 6.7 ± 3.0-fold higher than the uptake in the non-inflamed area of immunized rats (19 ± 9.3 vs. 2.9 ± 1.1 PSL/mm²; P = 0.004) and 7.1 ± 3.5-fold higher than controls (2.7 ± 0.4 PSL/mm² n = 8; P = 0.0005; Figure 3C).

Ex vivo autoradiography showed that uptake of non-specific [⁶⁸Ga]Ga-DOTA-E[c(RGEfK)]₂ in histologically confirmed inflammatory lesions of three immunized rats was 76% lower as compared to [⁶⁸Ga]Ga-NODAGA-RGD in inflammatory lesions of seven immunized rats (4.5 ± 0.6 vs. 19 ± 9.3 PSL/mm²; P = 0.04). The average [⁶⁸Ga]Ga-DOTA-E[c(RGEfK)]₂ uptakes in the inflamed area was higher than the uptake in the non-inflamed area of those three immunized rats (4.5 ± 0.6 vs. 0.9 ± 0.3 PSL/mm²; P = 0.004) but lower than the uptake of [⁶⁸Ga]Ga-NODAGA-RGD in the inflamed area.