Quantification of regional murine ozone-induced lung inflammation using [18F]F-FDG microPET/CT imaging

Ozone (O3) is a highly potent and reactive air pollutant. It has been linked to acute and chronic respiratory diseases in humans by inducing inflammation. Our studies have found evidence that 0.05 ppm of O3, within the threshold of air quality standards, is capable of inducing acute lung injury. This study was undertaken to examine O3-induced lung damage using [18F]F-FDG (2-deoxy-2-[18F]fluoro-D-glucose) microPET/CT in wild-type mice. [18F]F-FDG is a known PET tracer for inflammation. Sequential [18F]F-FDG microPET/CT was performed at baseline (i.e. before O3 exposure), immediately (0 h), at 24 h and at 28 h following 2 h of 0.05 ppm O3 exposure. The images were quantified to determine O3 induced spatial standard uptake ratio of [18F]F-FDG in relation to lung tissue density and compared with baseline values. Immediately after O3 exposure, we detected a 72.21 ± 0.79% increase in lung [18F]F-FDG uptake ratio when compared to baseline measures. At 24 h post-O3 exposure, the [18F]F-FDG uptake becomes highly variable (S.D. in [18F]F-FDG = 5.174 × 10–4 units) with a 42.54 ± 0.33% increase in lung [18F]F-FDG compared to baseline. At 28 h time-point, [18F]F-FDG uptake ratio was similar to baseline values. However, the pattern of [18F]F-FDG distribution varied and was interspersed with zones of minimal uptake. Our microPET/CT imaging protocol can quantify and identify atypical regional lung uptake of [18F]F-FDG to understand the lung response to O3 exposure.

Ozone (O 3 ) is a toxic and highly reactive gaseous oxidizing chemical with well-documented adverse health effects in humans. On the basis of animal and human data, environmental guidelines and air quality standards recommend a threshold for exposure of no more than 0.063 ppm of O 3 (average daily concentrations). Experiments done in animal models have shown that O 3 induces acute lung injury, albeit at much higher and for longer O 3 exposures (near 2 ppm for 3-6 h). Our research has standardized a sensitive model of sterile murine lung inflammation by exposing mice for two hours at 0.05 ppm O 3 , a level below the current recommendations for what is considered a safe or "ambient" O 3 concentration for humans [1][2][3] . 0.05 ppm O 3 exposure causes immediate lung neutrophil recruitment, release of IL-1β dependent cytokines in broncho-alveolar lavage and bone marrow mobilization of pan-leukocyte chemokine, SDF1α 2,3 . Thus, it is imperative to understand the progression of O 3 induced early lung metabolic changes, at concentrations feasible in the environment.
[ 18 F]F-FDG (2-deoxy-2-[ 18 F]fluoro-D-glucose) microPET/CT imaging is a sensitive method to detect lung cancer and study various inflammatory diseases. The current quantification methods focus on tracer kinetics and compartment modelling such as the Patlak and Sokoloff methods 4,5 . However, the compartmental modelling does not convey the regional [ 18 F]F-FDG activity pattern unless images are acquired through dynamic gating protocols and invasive blood sampling. Many pre-clinical microPET scanners resolve only up to 80 µm structures, which cannot resolve functional activity in the murine alveolar septa. Thus, we sought to fast-track the imaging protocol without invasive blood sampling. For understanding disease progression and regional tracer localization in relation to specific organs such as lungs, we developed an imaging protocol cum image analysis strategy to quantify the sequential uptake and distribution of [ 18 F]F-FDG in murine lungs. We imaged each animal at four separate image time-points that spanned 2 days. Ozone exposures. For O 3 exposures, mice were continuously exposed in an induction box for 2 h 1,2 . These mice were housed in custom induction box and had free access to food and water. O 3 (0.05 ± 0.02 ppm) was generated, at 3 L/min, from ultra-high-purity air using a silent-arc discharge O 3 calibrator cum generator (2B Technologies, CO, USA). Constant chamber air temperature (72 ± 3°F) and relative humidity (50 ± 15%) were maintained. O 3 concentrations were calibrated in small box using a real-time O 3 monitor (2B Technologies, CO, USA). Experiment design. Three pre-weighed mice were acclimatized to the cyclotron facility, overnight, and fasted for at least 4 h before imaging. The experiment design is shown in the schematic (Fig. 1).
Mice were imaged individually. At the beginning of the experiment, mice were prepared for baseline (i.e. before O 3 exposure) imaging. Immediately after inducing isoflurane (5%) anesthesia, mice were maintained under 1.5-2.5% isoflurane on a warm water circulated heating pad and monitored for vital signs (BioVet, breathing rate 40-100 breaths per min, body temperature 36.5-38 °C and blood oxygen saturation by pulse oximetry i.e. SpO 2 98-100%). Mice were injected with 2-5 MBq of [ 18 F]F-FDG via a tail vein. Within 15-30 min post injection, mice were imaged for 3-5 full-body time-frames (800 µm voxel size), where each frame spanned 5 min (VECTor 4 CT, MI Labs). Thus, for every time-point, mice were imaged for over 25 min. [ 18 F]F-FDG microPET imaging was followed by a 2 min X-ray CT (2 bed positions). After baseline imaging, the mice were continuously exposed for 2 h at 0.05 ppm O 3 as explained in Image processing and analysis. The acquired X-ray and PET-CT image data sets were processed for flat and dark current normalization, reconstruction, co-registration (MI Labs software) and quantification by Pmod (pmod.com). As all the time-points, before and after O 3 exposure, were acquired through the same imaging protocol, the [ 18 F]F-FDG counts were decay corrected in order to plot tissue [ 18 F]F-FDG uptake or elimination, and not decay, for the 25-30 min imaging time. Thereafter, the images were quantified and analyzed on Image J (https ://fiji.sc/#). The image stacks from X-ray CT were threshold-selected to segment out the lungs and confirmed with the PET-CT stacks. Depending upon the data-set, anywhere from 130 to 160 ortho slices spanned the entire lung region. The selected regions of interest (ROIs) were then copied on to the corresponding [ 18 F] F-FDG PET image slices across multiple frames (F0-F4). The ROIs, from X-ray CT as well as the PET images, were simultaneously quantified for the various image parameters such as the area, perimeter, mean, median, mode, standard deviation (SD), range and integrated counts. After exporting the data to excel file, data was sorted, filtered and analyzed for corresponding imaging time-points, frames and/or CT parameters (https ://doi. org/10.6084/m9.figsh are.12233 576). Finally, a SUM of the full-body parameters of Z-stacks were analyzed for every frame in order to calculate the full body [ 18 F]F-FDG SUV (Standard Uptake Value). After [ 18 F]F-FDG SUV is quantified for lung slice(s) as well as the full-body, the ratio of these values, termed as the Standard Uptake   (Fig. 3, Movie 1) compared to the full-body uptake (Fig. 4a), with majority excreted through the kidneys (Suppl. Fig. 1). Notice the steady uptake in bladder (up to 37 min of imaging) versus already peaked concentrations in lungs (by 17 min) and heart (by 24 min) (Supp. Figs. 1a, a1). It must be noted that O 3 induces instant lung tissue damage, as indicated by attenuated lung CT grey values, at 0 and 24 h post-exposure, compared to baseline values (Fig. 4j). Despite the loss of lung tissue, O 3 induced a 72.21 ± 0.79% increase (p < 0.05) in lung [ 18 F]F-FDG at 0 h compared to baseline (Fig. 4a, b, e, f, i). This is not very obvious in the single ortho-slice shown but evident in the volume rendered [ 18 F]F-FDG uptake in lungs shown in Fig. 3 Fig. 5a).

Discussion
Positron emission tomographic (PET) imaging with [ 18 F]F-FDG is a promising technique that may serve as a more sensitive outcome measure for pulmonary inflammation. The advantages of PET imaging include its noninvasiveness, ease of quantification, and ability to assess the entire lung. [ 18 F]F-FDG-PET exploits the "Warburg effect", the observation that many cancers use aerobic cytoplasmic glycolysis as opposed to mitochondrial glucose oxidation as a major energy source, a process that requires increased cellular glucose uptake. Aerobic glycolysis is also a characteristic of nonmalignant proliferating cells and is observed in acute lung inflammation 6,7 . Evidence to date suggests that neutrophils contribute primarily to the increased uptake of [ 18 F]F-FDG in lung inflammation and that the [ 18 F]F-FDG-PET signal correlates with the presence of activated neutrophils 8 . Clinical studies have also demonstrated that [ 18 F]F-FDG-PET imaging can be used to assess the neutrophilic inflammatory burden in the lungs in cystic fibrosis, pneumonia, and experimentally induced lung inflammation [8][9][10][11] . These results together indicate that [ 18 F]F-FDG-PET imaging can potentially be used to measure changes in pulmonary inflammation in response to anti-inflammatory treatments. We set out to explore 2 questions: (1)    We did not report the uptake or elimination [ 18 F]F-FDG rate constants in our current study, owing to fewer data points, spanning a total of 30-35 min imaging in 5 min increments, which adds to the lack of time resolution for effective uptake and elimination rate constant calculations. Future studies will aim at modifying the acquisition parameters for plotting [ 18 F]F-FDG time activity curves and repeated blood sampling. Our sequential [ 18 F] F-FDG imaging strategy is capable of objective analysis of the lung [ 18 F]F-FDG retention before as well as up to 28 h after O 3 exposure.

Conclusion
Thus, our study reveals that longitudinal [ 18 F]F-FDG-PET imaging may offer a tool for deep phenotyping of lung inflammation to understand the response to new targeted treatments in animal models and later in clinical trials.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.