Measurement of hepatic glucose (18F-fluorodeoxyglucose) uptake with positron emission tomography-magnetic resonance imaging in fumonisin B intoxicated rabbit bucks

Rabbit bucks (bodyweight 5 kg) underwent dietary intoxication with fumonisin B series mycotoxins (FB1 + FB2 + FB3, 15 mg/kg diet) for 14 days to test the applicability of positron emission tomography-magnetic resonance (PET MR) hybrid imaging in characterizing experimentally induced mild hepatotoxicosis. 18F-fluorodeoxyglucose (18F-FDG) radiotracer-aided imaging was performed before and after FBs administration on identical animals, and at both time points, blood was sampled for haematology and clinical chemistry. Kinetic PET image analysis revealed time-activity curves with uptake maxima below 1 min in the liver, renal cortex, portal vein, lung and coarctatio aortae. In the frame of static PET image analysis, based on the standardized uptake value (SUV), the so-called metabolic liver volume (MLV, liver volume defined by over 0.9 × average liver SUV) and the total liver glycolysis (TLG, MLV multiplied by the SUVmean) were calculated. Mycotoxicosis increased total liver glycolysis (p < 0.04) after 14 days and liver tissue TLG inhomogeneity was minimal. Pearson correlation between TLG and alkaline phosphatase (ALP) was positive (0.515), while negative with LDH and AST (− 0.721 and − 0.491, respectively). Results indicate a slight hepatic mycotoxin effect and significantly increased glucose uptake intensity, which has been sensitively detected with molecular imaging (18F-FDG PET MRI) in the rabbit model.

characterized 2 .Fumonisin occurrence is primarily frequent in maize commodities, but may also occur on other crops and food items 3,4 ; the prevalence was 68% in 2022 in the tested corn samples 5 .The EU Commission Regulation 6 on the maximum allowed levels for FB 1+2 in unprocessed maize kernels is 4 mg/kg (food and feed also), while for rabbit feed this is 5 mg/kg 7 .
The harmful effects of FB 1 are species-specific in vertebrates 8,9 .FB 1 -induced hepatotoxic effects in rabbits have been reported at 7.5-10 mg/kg feed doses for 196 days (Ewuloa, 2009) and also at 10 mg/kg for 4 weeks 10 .The mode of action is mostly well understood 9 , providing organ specificity even within rodents 9 , with the liver and/or kidney being the target organs.FB 1 is a ceramide conformational analogue and a competitive inhibitor of CoA-dependent ceramide synthase 11 .Harmful effects of FB 1 in the liver and kidney of most domestic animal species are exerted via alterations of the sphingolipid metabolism, ultimately causing apoptosis and oncotic necrosis, and as well carcinogenesis in rodents 9 .
Fumonisins' effects on the liver are quite widespread 9 .The in vivo hepatic effects are based on two toxic actions: the drastic disruption of lipid metabolism and the induction of further molecular changes leading to diseases/ malfunctions.Lipids undergoing quantitative or qualitative changes are per se primarily the sphingolipids 12 , as FB 1 not only blocks the biosynthesis of complex sphingolipids, but induces sphinganine accumulation.According to Riley et al. 9 , sphinganine is partly metabolized to sphinganine-1-phosphate and degraded to hexadecanal and ethanolamine phosphate; the latter is incorporated into phosphatidylethanolamine (PE).Quantitative alterations of the hepatic PE (and P-choline) lipid fraction were shown in vitro in primary rat hepatocytes 13 , while in vivo, our team provided evidence for the perturbation of the fatty acid metabolism of PE and PC fractions in rats 14 .The facts above support that FBs are inducers of a complex change in multiple lipid fractions' composition 14 , some of them being involved in the regulation of energy metabolism 15 .
With regards to glucose metabolism, in the development of insulin resistance (either in the liver or muscle), bioactive lipids play determinant roles, with ceramides being one the most potent inhibitors of insulin signal transduction 16 .Meanwhile, hepatic glucose intolerance is primarily associated with increased diacyl-glycerol levels.The inter-relationship between hepatic tissue ceramide level and glucose intolerance has also been highlighted in the skeletal muscle 17 .According to Babenko and Kharchenko 18 , stress (and aging) associated modification (decrease) of hepatic ceramide levels/sphingolipid metabolism partly improved insulin regulation of glucose metabolism in vitro in rat hepatocytes.Ceramide synthesis inhibitors (e.g.fumonisin B 1 ) improved insulinmediated glucose metabolism 18 , thus enhancing cellular insulin-induced glucose uptake.As a minor addition, it shall be noted that in vivo fumonisin feeding studies often report the alterations in plasma glucose levels, more specifically a decrease (supporting insulin-mediated glucose uptake of the liver and muscles), as observed in chicks 19 and heifers 20 , co-exposure with aflatoxin).While in vitro, in porcine intestinal cells (IPEC-J2), FB 1 reduced the expression of glucose transporters, Na + /glucose co-transporter 1 (SGLT1) and glucose transporter 2 (GLUT2) 21 , in vivo Na-dependent glucose absorption is up-regulated in pigs after acute or longterm exposure to FB 1 22 .Our hypothesis was set up on the above basis, namely fumonisin B 1 is a characteristic ceramide synthesis inhibitor 23 , and ceramide efficiently lowers insulin sensitivity.Thus, the reduction of hepatic ceramide level shall-theoretically-up-regulate hepatic glucose uptake.

18
F-fluorodeoxyglucose positron emission tomography ( 18 F-FDG PET) is a reliable, functional imaging tool that provides valuable information for detecting, staging, and evaluating the therapeutic responses 24 , applicable also on small pet animals such as rabbits 25 .As a glucose analogue, the uptake of 18 F-FDG is a biological process of glucose consumption at cellular levels through intra-cellular transportation and phosphorylation.Since the liver is one of the organs in the body with relatively intensive glucose metabolism, 18 F-FDG is taken up by liver cells.Various factors, including toxicity, can influence this uptake, such as certain medications or diseases (hepatitis, fatty liver disease, cirrhosis, infections, and tumours).Our study focused exclusively on toxicity in a controlled cohort of rabbits, utilizing PET MRI imaging with 18 F-FDG.We gathered accurate anatomical information and tracked the uptake stage and changes of 18 F-FDG by using a bimodal PET MRI scanner, which provided simultaneously acquired MRI and kinetic PET images 26 .
To test and possibly quantify the fumonisin B series' effect on the hepatic glucose uptake (and its alterations), we used a rodent model of known hepatic reaction for this intoxication 23 .We invented a coupled/hybrid imaging technique (PET MR imaging) for analyzing hepatic glucose uptake in not severely fumonisin-intoxicated adult rabbit bucks.

Animals
Rabbit bucks of ca. 5 kg bodyweight (n = 4) of the Pannon White breed were enrolled in the study.Animals were kept individually in cages, and were fasted for exactly 6 h before the scanning protocol initiation to reach a nearly euglycaemic status (mean ± SD of the 1st and 2nd timepoints: 7.48 ± 0.88 and 7.21 ± 0.66 mmol/L, p = 0.57) 27 , leading to possibly uniform glucose baseline values on which tracer glucose was superposed.All animals were individually marked and scanned twice, before and after mycotoxin exposure, i.e. serving as their own individual controls.The bodyweight of animals was measured weekly, individually during the study (Fig. 2).

Feeding and intoxication
A total exposure period of 14 days was performed.A commercial rabbit feed (composition: Table 1) was mixed with a fungal culture containing fumonisin B series.The FBs (FB 1 + FB 2 + FB 3 ) feed concentration was 15 mg/kg.The mycotoxin concentration of the control and the experimental feed was determined with LC-MS 28 .The limit of detection (LOD) for FB 1 was 3 µg/kg.The diet fed to the control group did not contain detectable amounts of FBs (the full absence of deoxinivalenol, zearalenone and T-2 toxin was as well controlled and confirmed).Feed consumption was measured individually (Fig. 2).
The radiopharmaceutical tracer (allowance number: https:// ogyei.gov.hu/ gyogy szera datba zis& action= show_ detai ls& item= 133670) was produced using a TRASIS AllInOne universal radiochemistry synthesizer with a TRASIS S4000-7558 FDG (Trasis, Liége, Belgium) cassette in the early morning hours of the day of the measurements.The planned dose amount was stored in a sterile, rubber-sealed boro-silicate ampule isolated inside a shielded lead canister.
Just before the injection (into the anaesthetised animal), the radiopharmaceutical tracer was drawn into a B Braun Omnifix 3 ml Luer Lock syringe by hand.The dose amount (ca. 100 MBq / 2 ml) was measured with a daily-calibrated CURIEMENTOR 4 isotope calibrator (PTW, Freiburg, Germany).The syringe was inserted inside MR-compatible, tungsten syringe shield (Biodex W/Glass 3 ml, Loc ID: MA139, 007-961, NY, USA).The shielded syringe was carried in-house in an additional lead canister for additional radioactive protection.

Anaesthesia, radiotracer administration and PET MR hybrid image acquisition
The PET MR imaging was performed before and after the 14-day mycotoxin exposure, on the same rabbit experimental cohort/identical individuals.Animals underwent PET MR imaging under anaesthesia induced with 3% Isofluran (Abbott Lab., IL, USA) via a small animal narcotic mask.To avoid urine (and 18 F-FDG) contamination of the scanners, animals wore commercial human baby diapers during the scanning protocol.
The right hind leg was cannulated via the external iliac vein with an intravenous catheter of a Vasofix Braun-üle® 20Gx1 ¼" (B.Braun Melsungen AG, Melsungen Germany).
Examinations were performed using a hybrid PET MR scanner (Biograph mMR, Siemens Healthcare GmbH, Erlangen, Germany).Blood glucose levels were checked before tracer injection to ensure that rabbits were ca.euglycemic.The rabbits received an administration of 100 MBq activity/individual of 18 F-FDG (single tracer injection) intravenously, using an MR compatible, tungsten syringe shield (Biodex W/Glass 3 ml, Loc ID: MA139, 007-961), with 20 ml saline flush.The radiotracer was administered in a 2 ml bolus, containing the ca. 100 MBq dose.
18 F-FDG PET MR image acquisition and reconstruction 18 F-FDG PET MR image acquisition was performed on a 3T hybrid PET MR scanner (Biograph mMR, Siemens Healthcare GmbH, Erlangen, Germany) with a PET MR compatible Siemens mMR (24 channel) body coil and a Siemens mMR Spine (6 channel) spine coil.The rabbits were positioned in a head-first-prone position.After the T2 localization images, a GRAPPA MRAC image (TE = 3.56 ms, TR = 2.46 ms, flip angle = 10°, voxel size = 2.6 × 2.6 × 3.1 mm) was acquired for the MR-based PET attenuation correction.The μMAP used for attenuation correction of the PET images was calculated on-site by the Syngo MR E11 software (Siemens Healthcare GmbH, Erlangen, Germany), using a four compartment (air, water, fat, lung adaptive) model.After this, a 60 min list mode PET acquisition was started.The injection of 100 MBq 18 F-FDG was administered as a bolus injection after 5 s from the start of the PET-measurement into the external iliac vein.After the PET scan, a bolus injection of 0.4-0.5 ml Gadovist solution diluted with 1 ml of water was administered, followed by a 20 ml saline flush.For anatomical purposes, a T1-weighted starvibe fs transversal post-contrast 3D sequence was acquired (TE = 1.97 ms, TR = 3.79 ms, flip angle = 9°, voxel size = 0.8 × 0.8 × 0.8 mm).
PET images (voxel size = 1.669 × 1.669 × 2.03 mm) were reconstructed retrospectively using the Ordinary Poisson OSEM algorithm with general purpose settings (iteration = 3, subset = 21) with model-based scatter correction and delayed event subtraction for random correction.A Gaussian kernel (FWHM = 4.0 mm) was used for spatial filtering during the reconstruction.
We used static and dynamic 18 F-FDG images scaled to Standard Uptake Value (SUV) for PET image analysis, where the SUV was calculated as SUV = tissue activity (kBq/ml)/injected dose (kBq)/body weight (kg).Based on the injection time, a 40-60 min time window was used to reconstruct static images from the list-mode PET data.In contrast, the following time resolutions were used for evaluating the dynamic images: 10 × 3, 18 × 15, 10 × 60 and 9 × 300 s.

Kinetic PET image analysis
We conducted a kinetic analysis that served to test the applicability of the reversible two-tissue compartment model, investigated using the blood supply from both the hepatic artery and portal vein for accurate kinetic modelling of 18 F-FDG uptake of the liver, as proposed by Wang et al. 29 .To complete this task, we established eight spherical Regions of Interest (ROIs) with a 3-8 mm radius for the coarctatio aortae, lung, liver aorta, portal vein, left liver lobe, right liver lobe, renal cortex, and gut.During the image analysis stage, we utilized the T1-weighted MR images from the PET MRI system to identify the correct anatomical regions in the combined PET and MRI images.Using the delineated ROIs, we generated the appropriate Time Activity Curves (TACs) from the dynamic 18 F-FDG PET image series.

Static PET image analysis
In tumour diagnosis, the standardized uptake value (SUV) is the most widely used measure for quantifying metabolic activity.It is calculated as the ratio of tissue radioactivity concentration to the injected dose, normalised by BW.The maximum SUV (SUVmax) is the highest voxel value of the SUV found in the tumour.The SUVmax measurement of a tumour's metabolic burden is limited because it is based on only one voxel.Our study aimed to analyse the whole liver's entire glucose metabolism.To achieve this, we used parameters that combine glucose uptake volume and uptake intensity instead of SUVmax.
In 18 F-FDG PET image analysis, two other parameters are commonly used: metabolic tumour volume (MTV) and total lesion glycolysis (TLG, but in further cases, this abbreviation is used to designate total liver glycolysis).MTV measures the size of the tumour with high metabolism using a specific SUV threshold, while TLG is calculated by multiplying the SUVmean and MTV.Using this concept, we have developed two new parameters based on the liver's metabolic activity: the metabolic liver volume ROI (MLV ROI) and the total liver glycolysis (TLG).We used a region-growing technique with a pre-defined threshold to generate the MLV ROI, a specific region in the liver where the standardized uptake value (SUV) is higher than a predetermined level.We were unable to apply any suggested/literature-based SUVmean or SUVpeak-based threshold successfully 30 .Nonetheless, the limited number of images allowed for a visual examination of the mask-fitting to the liver, which was fused on the PET and MRI images of the PET MRI scan.By utilizing this image-fusion technique, we established the threshold at 0.9 of the mean SUV value of a manually set spherical ROI.To calculate the TLG, we multiplied the volume of the MLV ROI (measured in cubic cm) by the SUVmean value of the MLV ROI 31 .Thus TLG considers not only the FDG uptake but also the liver volume.
Figure 1 shows the fusions of 18 F-FDG PET, T1-weighted MRI and the MLV ROI.

Ethical allowance statement
The experiment was carried out according to the Hungarian Animal Protection Act regulations.The allowance number for the studies was SOI/

Statistical evaluation
Bodyweight (BW), feed intake, clinical chemical and hematological data were evaluated with paired samples t-tests and Wilcoxon tests.For calculated indices and biochemical parameters relationship testing, Pearson correlation was calculated, using IBM SPSS Statistics for Windows (version 27.0) 32 .

Growth and feed intake
After the first day (scanning), the BW decreased temporarily, being suggestive of a temporary growth retardation (Fig. 2).Feed consumption increased slightly, and after the 4th measurement (study day 10) it dropped to the baseline level.

Blood plasma clinical chemistry
From the plasma nitrogenous compounds, total protein and within this, albumin concentration decreased significantly during the study period, while plasma creatinine concentration increased.Urea and uric acid concentrations were unchanged.Within the lipids, triglyceride (TG) and total cholesterol decreased significantly, meanwhile HDL cholesterol was unchanged.In ions, potassium (K) and total calcium (Ca) decreased during the study, while none of the other ions (Na, P, Cl, Mg) provided any significant change.Plasma enzyme activities showed a decrease in AST activity and an increase in gamma-GT and alkaline phosphatase (ALP).ALT, cholinesterase and creatine kinase (CK) levels remained unaltered (Table 2).

Haemogram
Mean data on the cell distribution in the haemograms are given in Table 3. Irrespective of treatment (sampling), erythrocytes exhibited a mild anisocytosis, but no marked shape difference was identified.The leukocyte distribution, specifically that of heterophilic granulocytes and lymphocytes, varied between individuals, but fluctuations were slight.After FBs treatment, high and low proportional tendencies of heterophil granulocytes and lymphocytes were observed without statistical significance.The distributions of other leukocytes in the blood (eosinophil and basophil granulocytes, monocytes) were not significantly different before and after treatment.

PET MR hybrid imaging
Figure 3 shows the typical anatomical locations where 18 F-FDG accumulation has been profound, i.e., in the digestive system, the aortic arch, and the kidney.
For the same animal, the hepatic ROI system is given in Fig. 4.

Time activity curves
Using the delineated spherical ROIs (see Fig. 4.), we generated Time Activity Curves (TACs) from all dynamic PET scans.Figure 5 shows one representative TAC system in the range of 0-60 min (Fig. 5).This figure of TACs shows that 18 F-FDG uptake in all areas of interest was rapid, reaching its maximum within 1 min, significantly faster than in humans (30-40 min), but after the maximum started a slight decrease of SUV  values and the relative changes of the different ROIs' TAC minimized after 40 min.According to this observation, we applied the 40-60-min time window for generating static 18 F-FDG images.During the ROI analysis, we found delineating artery and vein ROIs challenging; in some cases, we could not create proper carotid or liver arterial ROIs even using good anatomical information from MR images.This practice may be explained by the PET imaging technique's partial volume effect (PVE), i.e., the artery and vein size are lower below the 3.5 mm spatial resolution of the PET camera system.

Metabolic liver volume and total liver glycolysis
Based on MLV and TLG calculations, the total hepatic 18 F-FDG uptake and the difference (before and after intoxication) were calculated.Before and after intoxication group median values are shown in Fig. 6, providing a significant (p = 0.031) difference.The median MLV data were 100.4 cm 3 and 116.2 cm 3 , respectively, and the median of SUVmean values were 2.03, both in the control and FBs cases.

Inter-relationship between hepatic glucose uptake and clinical chemistry
As a final step, the possible inter-relationship between blood clinical chemical (direct biochemical traits) data and TLG was analyzed with Pearson correlation analysis (Table 4).
The closest positive correlation of TLG (referring to the cases showing as well inter-group difference) was found with ALP, while the closest negative correlation was with total cholesterol, LDH and AST.

Growth and feed intake
In a recent study, we fed a rabbit cohort (of the same breed, Pannon White) a feed contaminated with 10 and 20 mg/kg FBs, respectively 33 .The rabbits in the cited study were enrolled into a reproduction study (for 65 days), with minimal handling; these animals showed slow but balanced growth without fallback.In another FB 1 intoxication study, 10 mg/kg FB 1 for 4 weeks did not compromise rabbit buck growth 34 .Ewuola 35 fed rabbit bucks FB 1 (0.13, 5, 7.5, and 10 mg/kg diet) for 196 days and found no fallback in BW.At the same FB 1 concentration range for 175 days, the onset of puberty was delayed by the two higher doses (i.e., 7.5 and 10 mg/kg), but the mycotoxin did not significantly influence bodyweight at puberty.Most probably, the BW loss recorded in the first 3 days shall be a treatment/handling (anaesthesia and subsequent PET MR imaging protocol) associated effect, and it was clearly visible that the growth curve continued after this initial fallback.
From the literature, the effects of FBs on FI seem controversial.Ewuola et al. 36 (12.3 and 24.5 mg FB/kg diet for 5 weeks) reported a contradicting (decrease in FI) finding, a similar result has been reported by Szabó et al. 33 .Since our recorded high feed intake was temporary, it fell back to the baseline at day 14.This may be supported  by the finding of Hafner et al. 10 , where 10 mg/kg dietary FB 1 increased feed intake of growing rabbits temporarily (during the first week) compared to non-intoxicated controls.

Nitrogenous compounds
After 14 days of FBs exposure, total protein and albumin concentrations fell considerably in the plasma, whereas the plasma creatinine concentration increased, latter is a widely accepted indicator of fumonisin-induced nephropathy (in rabbits: 10 , in farm animals: 37 ).Ewuola and Egbunike 36 reported that in rabbits, serum total protein showed an FB 1 dose-dependent concentration drop.The authors explained this with possibly altered dietary protein metabolism, including compromised digestibility and absorption of this nutrient.This has been more recently supported by Zeebone et al. 38 , where the apparent ileal digestibility of crude protein and more particularly that of arginine, histidine, and tyrosine was altered (exposure time × dose interaction) by 40 mg/ kg dietary FBs for 7 and 21 days in piglets.In the present study, plasma total protein concentration values of both groups (control and FB 1 ) fell into the physiological reference range (5.5-8.0 g/L), according to Mitruka and Rawnsley 39 .Indeed, at the dose applied, we did not prove either hypoproteinemia or hypoalbuminemia, and the unchanged levels of urea and uric acid in both groups refer to a balanced protein metabolism (supported by data of 36 ), while creatinine referred to mycotoxin associated compromised renal function.As the ultimate reason of the findings on TP and albumin, we rather suppose altered/minimally compromised hepatic function (as indicated by AST and gamma-GT).

Plasma lipids
Within the lipid metabolites, TG and total cholesterol concentrations decreased significantly, as well as LDL cholesterol.These alterations refer to a very mild extent of FB intoxication.Since serum TG concentration is direct correlate of dietary fat intake, its discussion in relation to FBs is void.In contrast, relevant total cholesterol results of Ewuola 40 provided a dose-dependent concentration increase in rabbits (0.1, 5, 7.5, and 10 mg/kg doses), and this hypercholesterolemic effect of FB 1 is widely published in mammals 41 .
An earlier in vitro study by Gelderblom et al. 11 referred to the strong compositional modification of hepatocellular membrane phospholipids, modifying the membrane structure and altering membrane cholesterol concentration.In multiple studies (e.g.Bondy et al. and Szabó et al. 42,43 ) on FB 1 effects on rats, blood plasma total cholesterol (of which free cholesterol is a major component) was reported to increase.However, fumonisinassociated hypercholesterolemia aetiology is not fully elucidated.According to Ridgway 44 , this effect is not a result of FB 1 -induced sphingosine-or sphinganine mediated inhibition of cholesterol esterification.Our results (total cholesterol decrease) hypothesize only a mild FB 1 effect on the hepatic lipid transport mechanism (towards extrahepatic sites).Furthermore, our data do not support the presence of hypercholesterolemia since recorded levels remain within the physiological range (0.3-3 mmol/l), according to Harcourt-Brown 45 .

Enyzmes
From the plasma enzymes, AST activity decreased during the study, while gamma-GT and alkaline phosphatase (ALP) increased.Relevant literature refers to a profound hepatic malfunction in rabbits as an effect of FB 1 .According to Orsi et al. 46,47 , AST, ALT, gamma-GT and ALP are responsive.A possible reason for the lack of FB 1 -response of AST and ALT may be that those are not strictly liver-specific.In contrast, gamma-GT is a sensitive indicator of liver cell injury and already a single FB 1 gavage dose (at 630 mg/kg feed dose) increased its activity in rabbits 46 .Hepatic injury was also demonstrated by a significant loss of body and liver weight, and authors interpreted ALP and gamma-GT levels as indicators suggestive of hepatic and/or biliary injury, sharing a common point with our results.Orsi et al. 46 reported that a single FB 1 dose vs. prolonged exposure (1.5 mg/ www.nature.com/scientificreports/kg for 21 days) induced a shift in organ-level intoxication, namely an early phase that is typically nephrotoxic, and this coupled with hepatotoxicity later, as corroborated by the findings of Gumprecht et al. 48and supporting the creatinine results.
Analysing the metabolic and enzymatic data comprehensively, we suppose mild hepatotoxic effect of dietary FBs (hepatocellular damage and consequent intracellular enzyme leakage: ALP and gamma-GT) with probably compromised hepatic albumin synthesis, and as well nephropathy development (increased creatinine level and ALP activity) with probable urinary protein loss (total protein and albumin level decrease in the serum).

Haemogram
The effects of fumonisins on rabbit erythrocytes are not yet fully elucidated; the results of earlier studies are controversial.In some cases, FB 1 induced 36,49,50 and could not induce 36,46 proportional haemogram shifts.Anisocytosis is a condition that frequently occurs in rabbits, resulting in a cell diameter of up to one-quarter of the normal erythrocyte.Szabó et al. 51 demonstrated that a 10 mg FB 1 /kg diet for 28 days markedly elevated rabbit erythrocytes' total Na + /K + ATPase activity.Szabó et al. 52 confirmed similar findings in piglets fed a 30 mg FB 1 /kg diet for 21 days.The authors proposed that alterations in sodium pump activity and membrane lipids occurred as a consequence of the inhibition of ceramide synthase activity, a characteristic FB mode of action.The findings of Szabó et al. 51,52 postulate that FB1 may alter erythrocyte surface/volume ratio and osmotic properties.In this study, mild anisocytosis was observed, suggestive of the negligible effect of FBs on erythrocyte morphology.Our findings corroborate those of Ewuola et al. 36 , who found that a 24.56 mg FB/kg diet for 5 weeks did not alter the haematological parameters of rabbits 36 .
To our knowledge, the in vivo effects of FBs on blood leukocyte counts are rarely discussed, especially in rabbits.From the available literature on mice 53 , swine 54 , and humans 55,56 , the effects of FB on the immune system vary according to species, sex, mode of toxin application, exposure period and mycotoxin dose.Although our findings showed only tendencies in leukocyte distributions, no statistically significant difference was proven.This finding implies that FBs have no or only a negligible effect on leukocytes.However, further investigations are needed to determine the implications of FBs on leukocytes following a prolonged exposure approach.

PET MR hybrid imaging
In the 1 st step of image analysis, we primarily focused on seeking the anatomical locations where radiotracer uptake is the most intense in the rabbits, aided by the PET MR hybrid imaging technique.
Relevant rabbit-based literature is scarce, but Probst et al. 57 reported basically similar 18 F-FDG biodistribution data in a rabbit model developing papillomavirus-induced tumors.More specifically, healthy skin, blood, liver, spleen, kidneys, heart, lung and fat were noted before infection and tumor development.In contrast, during tumor development, the tumor itself and the liver provided increased relative uptake of 18 F-FDG.Similar to our results, skeletal muscle was found to provide low 18 F-FDG uptake, and it was not changed during the study interval.Basically, similar results were described by Li et al. 27 , in that kidneys, spleen, liver and genitals were primary glucose uptake sites.Based on the portal vein and liver ROI data, we found that liver 18 F-FDG is a determinant proportion of the total uptake and might sensitively reflect the mycotoxin-induced metabolic disturbance.

Time activity curves
The primary goal of kinetic analysis was to check the 18 F-FDG uptake of different tissue types and to determine the time needed to reach the uptake maximum.Indeed, this has been found to be very rapid (below 2 min).Our results are consonant with those of Probst et al. 57 in rabbits, reporting peaking 18 F-FDG uptake immediately after administration (below ~ 5 min), and explaining this result with the high blood-pool concentration of FDG.Probst et al. 57 used 55-min acquisition time and described an early peak and a slight but stable decay during this interval for hepatic glucose uptake.We used 5 min only, but reported identical findings.Checking the data in Fig. 5, it is clearly visible that blood, arteries and veins showed the quickest and highest uptake, and from the organs, the liver was the primary uptake site (followed by the lung and intestines), being consonant with the findings of Probst et al. 57 , where liver uptake rate exceeded that of the tumour and skeletal muscle as well.
The second aim of the kinetic analysis was to check the usability of the two-compartment tissue model of the liver 29 .To use this model, two blood TACs (one from the liver artery and one from the portal vein) and two liver TACs (one from the lesion and one from the reference ROIs) were required.In all cases, we successfully delineated the two tissue ROIs and the portal vein.However, we encountered difficulty locating the liver artery TAC in almost all cases.This is due not only to the partial volume effect but also to the anatomical fact that the artery and the aorta abdominalis are located too close.As a result, we could not separate the kinetics of the liver artery from the a. abdominalis kinetics, as the latter transported the bolus-injected 18 F-FDG from the rabbit's leg to the heart.
After analyzing the kinetic data in detail, we concluded that the liver FDG uptake cannot be accurately quantified using any tracer kinetic model.To successfully complete similar experiments it is crucial to have more standardized procedures for administering 18 F-FDG injections and selecting the input location.

Metabolic liver volume and total liver glycolysis
Standardized uptake value, metabolic tumor (liver) volume, and total lesion (here: liver) glycolysis are considered as efficient prognostic factors in many cancerous cases.Basically, we intended to seek possible correlations between a well-predictable hepatic malfunction and the volumetric, functional metabolic parameters.Since the hepatotoxic effect of FBs in rodents is already well predictable, but hepatic glucose uptake has not yet been addressed as a function of FB dosage, we tested our hypothesis that FBs may increase glucose uptake rate.A specific point of our study was that MLV and TLG are used in tumor staging, while we used not only intra-individual https://doi.org/10.1038/s41598-024-68210-3

Figure 1 .Figure 2 .
Figure 1.(A) Image fusion of 18 F-FDG PET and T1 weighted MR images in three orthogonal sections.(B) The mean liver volume (MLV) generated with the region growing technique.

Figure 3 .
Figure 3. Fused 18 F-FDG PET and T1-weighted MR images of a control-phase (before intoxication) rabbit that received 100 MBq 18 F-FDG.The white circle contours show the section of 3D spherical ROIs.

Figure 4 .
Figure 4. PET MR fused images (maximum intensity projection) of a control rabbit that received 100 MBq 18 F-FDG.White circles are the hand-positioned ROIs.

Table 1 .
Chemical composition of the diet.

Table 4 .
The Pearson correlation coefficients between TLG and the analysed clinical chemical endpoints.ALP alkaline phosphatase, ALT alanine transaminase, AST aspartate aminotransferase, CK creatine kinase, gamma-GT glutamyl transferase, HDL high density lipoprotien, LDH lactate dehydrogenase.