The potential utility of (2S,4R)-4-[18F] fluoroglutamine as a novel metabolic imaging marker for inflammation explored by rat models of arthritis and paw edema

Purpose (2S,4R)-4-[18F]fluoroglutamine ([18F]FGln) is a promising metabolic imaging marker in cancer. Based on the fact that major inflammatory cells are heavily dependent on glutamine metabolism like cancer cells, we explored the potential utility of [18F]FGln as a metabolic imaging marker for inflammation in two rat models: carrageenan-induced paw edema (CIPE) and collagen-induced arthritis (CIA). Procedures: The CIPE model (n = 4) was generated by injecting 200 μL of 3% carrageenan solution into the left hind paw three hours before the PET. The CIA model (n = 4) was generated by injecting 200 μg of collagen emulsion subcutaneously at the tail base 3–4 weeks before the PET. A qualitative scoring system was used to assess the severity of paw inflammation. After a CT scan, 15.7 ± 4.9 MBq of [18F]FGln was injected via the tail vein, followed by a dynamic micro-PET scan for 90 minutes under anesthesia with isoflurane. The standard uptake value of [18F]FGln was measured by placing a volume of interest in each paw. The non-injected right hind paws of the CIPE model rats served as controls for both models. The paws with CIA were pathologically examined after PET. Results In CIPE models, uptake in the injected paw was higher compared to the non-injected paw by 52–83%. In CIA models, uptake in the paws with severe inflammation was higher than the averaged controls by 54–173%, while that with mild and no inflammation was slightly higher (33%) and lower (−7%), respectively. Combined overall, the [18F]FGln uptake in CIA showed a significant positive correlation with inflammation severity (r = 0.88, P = 0.009). The pathological findings confirmed profound inflammation in CIA. Conclusions [18F]FGln uptake was increased in both acute and chronic inflammation, and the uptake level was significantly correlated with the severity, suggesting its potential utility as a novel metabolic imaging marker for inflammation.


Introduction
(2S,4R)-4-[ 18 F] uoroglutamine ([ 18 F]FGln) has shown promise as a metabolic imaging marker using positron emission tomography (PET) technique in various types of cancer [1][2][3][4].For instance, a recent study in human subjects suggested that [ 18 F]FGln is more speci c to the metabolic activities of cancer cells avid to glutamine compared to [ 18 F] uoro-D-glucose ([ 18 F]FDG), which is currently the standard PET imaging marker to monitor cancer [5].However, [ 18 F]FGln has been rarely evaluated in non-cancerous disease conditions such as in ammation, despite the ample evidence of the heavy metabolic dependence of major in ammatory cells on glutamine as well.Speci cally, glutaminolysis, a conversion process from glutamine into glutamate, aspartate, and alanine, is an essential metabolic step to provide energy sources for the proliferation of in ammatory cells and their reactions against external or internal pathogens [6].Thus, [ 18 F]FGln has great potential to be used to monitor in ammation in addition to cancer, and we aimed to explore the utility of the [ 18 F]FGln as a novel metabolic imaging marker for in ammation in well-established rat models of acute and chronic in ammation.
Two distinct rat models were used in this study: the carrageenan-induced paw edema (CIPE) model for local and acute in ammation, and the collagen-induced arthritis (CIA) model for systemic and chronic in ammation.Carrageenan is a seaweed-derived sulfated polysaccharide that has been extensively used to induce in ammation in experimental animal models to study novel anti-in ammatory and analgesic drugs [7].In the CIPE model, carrageenan injected into the rodents' hind paws causes a classical innate immune response characterized by paw edema, neutrophil migration, and pain [8][9].CIA is an experimental model of autoimmune arthritis and has been shown to exhibit similar histological, immunological, and clinical characteristics and genetic linkage to human rheumatoid arthritis (RA) [10][11].This model has been widely used in studies on the pathogenesis of RA and the preclinical evaluation of novel therapeutics for RA [12].The protocol for the generation of the CIA model is well-established with high reproducibility [13].

Radiosynthesis of [ 18 F]FGln
[ 18 F]FGln was prepared through a semi-automated synthesis process (Fig. 1) [14], [15]: 1) The 18 F-labeled intermediate was prepared following an automated production protocol reported by Zhang et al. using GE™ TRACERlab FXNPro module; 2) The acidic deprotection and puri cation were conducted in a manual operation manner following the initial [ 18 F]FGln synthesis report with minor modi cation.After the [ 18 F]FGln intermediate was collected from semi-prep HPLC, it was rst concentrated using the solid phase extraction (SPE) method.The trapped radioactivity was next eluted with EtOH (1 mL), and the solvent was removed by azeotropic drying of the intermediate solution under a gentle nitrogen stream.Once all solvent was removed, the acidic deprotection (tri uoroacetic acid, TFA) and the following ionretardation resin removal of TFA provided the desired product [ 18 F]FGln.The total production time ranged from 86 to 110 min.Starting with 9.1-15.2GBq uorine-18 activity, 166-411 mBq amount of [ 18 F]FGln was produced at the end of synthesis with the radiochemical purities of the nal product range from 89-94%.

Generation of animal models
The experimental protocol for the generation of the CIPE and CIA models and micro-PET scans was approved by the Institutional Animal Care and Use Committee at Stony Brook University (IACUC2022-00034).
The CIPE model (n = 4, Lewis female rats, Charles River Laboratories) was generated by injecting 200 µL of 3% carrageenan solution into the left hind paw subplantarly as previously described [16] under general anesthesia with iso urane three hours before the PET scan.Carrageenan solution was made from Lambda Carrageenan (Sigma-Aldrich) mixed with saline shortly before injection.
The CIA model (n = 4, Lewis female rats, Charles River Laboratories) was generated by injecting collagen emulsion subcutaneously at the tail base 3-4 weeks before the PET scan [13].In brief, bovine type II collagen 2 mg/ml solution (Chondrex) was mixed with Incomplete Freund's Adjuvant (MP Biomedicals) using a homogenizer in an ice water bath to make an emulsion shortly before injection.In each animal, 200 µg of collagen emulsion was injected subcutaneously at the tail base under general anesthesia with iso urane.The injected animal was monitored daily for the development of arthritis.

Arthritis/paw edema rating
The severity of arthritis or paw edema was assessed using a qualitative scoring system (Table 1) [13].It is based on the edema and erythema in the hind paws of the rats.
Table 1 A qualitative scoring system used to assess the severity of paw in ammation [13] Score Condition 0 No evidence of erythema and swelling 1 Erythema and mild swelling con ned to the tarsals or ankle joint 2 Erythema and mild swelling extending from the ankle to the tarsals 3 Erythema and moderate swelling extending from the ankle to metatarsal joints 4 Erythema and severe swelling encompass the ankle, foot, and digits, or ankylosis of the limb

Micro-PET scans
After a transmission CT scan, 15.7 ± 4.9 MBq of [ 18 F]FGln was intravenously injected via the tail vein, followed by a dynamic micro-PET scan for 90 minutes with the Siemens Inveon integrated trimodal imaging system under general anesthesia with iso urane.In each [ 18 F]FGln micro-PET image, a 3dimensional spherical volume of interest (VOI) was placed in the ankle joint of each paw, based on the anatomical CT image using PMOD software v 4.2 (PMOD Technologies Ltd., Zurich, Switzerland).The VOIs had equal volumes between the right and left sides and did not include any part of the bone, which tends to be biased with radiometabolite accumulation.The standard uptake value in a pixel was calculated as tissue radioactivity concentration from the calibrated PET image divided by injected radioactivity/body weight.The maximum intensity projection images averaged from 10 to 40 minutes were used for visual comparison between conditions (presented in Fig. 2c and d).For quantitative comparisons, the Area Under the Curve (AUC) from 0 to 90 minutes was calculated in the [ 18 F]FGln timeactivity curve from each VOI.The injected left hind paw of each animal was compared to the noninjected right hind paw of the same animal in the CIPE models.Four non-injected right hind paws of the CIPE models served as controls of the CIA models for comparison of the [ 18 F]FGln uptake.

Statistical analysis
The correlation between the %increase of AUC in the [ 18 F]FGln time-activity curve and in ammation severity score in a total of eight hind paws of four CIA rats was analyzed in IBM SPSS v.29 with a nonparametric partial correlation method controlling for subjects (the degree of freedom = 5).

Postmortem exam
After obtaining micro-PET scans, the animals with CIA were sacri ced, and the paws were amputated and sent for tissue preparation.After decalci cation, xation, sectioning, and Hematoxylin & Eosin (H & E) staining following the published guideline [17], the section slides were digitalized for the visual assessment of in ammation.

Results
All four rats with the CIPE model developed severe in ammation with erythema and swelling that encompassed the ankle, foot, and digits (score 4 in Table 1) in the carrageenan-injected left hind paws in three hours from the injection (Table 2 and Fig. 2a).Among four rats with the CIA model, two developed severe in ammation (score 4) in both hind paws, one developed asymmetric in ammation with severe (score 4) and mild (score 2) in ammation in each hind paw, and the other one developed severe in ammation in only one hind paw (score 4 and score 0, presented in Fig. 2b) in 3-4 weeks from the injection of the type II collagen emulsion (Table 2).a For the CIPE model, the injected left side was compared to the non-injected right side in each animal.For the CIA model, each side was compared to the average value of the non-injected right side of the CIPE model rats (n = 4).
*%increase of AUC in the [ 18 F]FGln time-activity curve showed a signi cant positive correlation with the in ammation severity score (r = 0.88, P = 0.009).CIPE = Carrageenan-Induced Paw Edema; CIA = Collagen-Induced Arthritis; AUC = Area Under the Curve.
On the micro-PET scans of the CIPE models, [ 18 F]FGln uptake was higher in the injected hind paw with severe in ammation compared to the non-injected hind paw of the same animal by 52-83% (Table 2 and Fig. 2c).In the CIA models, [ 18 F]FGln uptake was higher in the hind paw with severe in ammation (score 4) compared to the averaged controls by 54-173% (Table 2 and Fig. 2d).In contrast, [ 18 F]FGln uptake in the paw with mild in ammation (score 2) and no in ammation (score 0) of the CIA models was slightly higher (33%) and lower (-7%), respectively, compared to the averaged controls (Table 2).Combined overall, the [ 18 F]FGln uptake and in ammation severity score in a total of eight hind paws of four CIA rats showed a signi cant positive correlation (r = 0.88, P = 0.009).
The post-mortem pathological exams of the rats with the CIA model con rmed profound in ammation in the paws with severe arthritis (Fig. 2e).Speci cally, H & E staining of the section showed typical synovial joint in ammation characterized by in ammatory cell in ltrates, increase in synoviocytes, thickening of synovial lining and sub-lining, and invasion of pannus tissue, as previously described in the CIA model [17].

Discussion
In our study, [ 18 F]FGln uptake was increased in both acute and chronic in ammatory conditions in the rat arthritis and paw edema models.Particularly, in rats with the CIA models, the [ 18 F]FGln uptake level showed a signi cant positive correlation with the severity of induced in ammation despite the small sample size.Our ndings suggest the utility of [ 18 F]FGln as a potential metabolic imaging marker for in ammatory conditions.
The involvement of glutamine metabolism in the development of joint in ammation has been consistently reported in previous studies with rodent CIA models or human RA cases.For instance, inhibition of the system A family of amino acids transporters, which has glutamine as one of the main substrates, was shown to reduce the development of CIA in rats [18].The results were mainly attributed to glutamine deprivation with subsequent decrease of circulating monocytes and neutrophils.In addition, glutaminase 1, the metabolizing enzyme from glutamine to glutamate, was shown to be increased in cultured synoviocytes harvested from synovial tissue of RA patients, and deprivation of glutamine led to suppression of the cell proliferation [19].Another study showed that glutamate is increased in the synovial uid of RA patients, and the level was positively correlated to the proin ammatory cytokine levels [20].Thus, the metabolic status of synoviocytes and other in ammatory cells in RA shows signi cant shifting from normal pathways of aerobic glycolysis to up-regulation of glutaminolysis, which is quite similar to that of the Warburg effect in cancer cells [6].Indeed, substantial evidence indicates that the metabolic characteristics of synoviocytes in RA have similarities to those of cancer cells [21].
In addition to the synoviocytes in RA, extensive study ndings have indicated that major in ammatory cells use glutamine as a dominant energy source in active in ammatory conditions.In previous studies, T-lymphocytes were shown to be dependent on glutamine metabolism in mice models with genetic arthritis or experimental autoimmune encephalomyelitis [22][23].Glutamine metabolism is also critical in macrophages, particularly for the modulation of the polarization between pro-in ammatory (M1) and anti-in ammatory (M2) phases [24][25].A recent study with [ 18 F]FGln in mice also presented the co-localization of [ 18 F]FGln uptake with glutamine transporter-positive macrophages in in amed atherosclerotic lesions [26].Neutrophils are also heavily dependent on glutamine to meet their energy demands at in amed sites where nutrients may be limited [27].Of note, neutrophils are the most abundant cells present in synovial uid in patients with RA and have pivotal roles in synovial tissue destruction [28].Both neutrophils and macrophages also contribute to the hyperacute local in ammation induced by carrageenan in the CIPE model [7][8][9].Based on this extensive evidence on the essential roles of glutamine in the metabolic status of in ammatory cells, immunosuppressant drugs, such as methotrexate, le unomide, and tofacitinib, have been developed and are clinically used for the treatment of RA or in ammatory myositis, with mechanisms to reduce the pro-in ammatory activity of immune cells via modi cation of metabolic pathways [21].

Conclusion
Glutamine metabolism is not only essential in cancer cells but also in major in ammatory cells with correlations to activity levels.Our preclinical ndings in rat models with arthritis and paw edema suggest the potential utility of [ 18 F]FGln PET as a novel metabolic imaging marker in various in ammatory conditions such as in ammatory arthritis and myositis.

Declarations
Figures

Figure 2 The
Figure 2

Table 2
Summary of [ 18 F]FGln micro-PET ndings and the severity of in ammation