Differences in in vitro microglial accumulation of the energy metabolism tracers [18F]FDG and [18F]BCPP-EF during LPS- and IL4 stimulation

The positron emission tomography probes 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) and 2-tert-butyl-4-chloro-5-{6-[2-(2-[18F]fluoroethoxy)-ethoxy]-pyridin-3-ylmethoxy}-2H-pyridazin-3-one ([18F]BCPP-EF) are designed to evaluate glycolysis and oxidative phosphorylation, respectively, and are both used to estimate neuronal activity. However, previous studies have shown a discrepancy in these probes’ accumulation in the compromised region, possibly due to the presence of activated microglia acting like deleterious or neuroprotective phenotypes. Hence, we evaluated lipopolysaccharide (LPS)- and interleukin 4 (IL4)-stimulated microglial uptake of [14C]2DG and [18F]BCPP-EF to give a new insight into the hypothesis that different uptake of [18F]FDG and [18F]BCPP-EF can be ascribed to the different metabolic pathways activated during microglial activation. LPS or IL4 stimulation increased the proinflammatory or anti-inflammatory marker gene expression in microglial cells. In LPS-stimulated cells, [14C]2DG uptake and glycolysis related gene expression were elevated, and [18F]BCPP-EF uptake was reduced. In IL4-stimulated cells, [18F]BCPP-EF uptake was increased, and [14C]2DG uptake was decreased. The expression of genes involved in glycolysis and mitochondrial complex I subunits was not changed by IL4 stimulation. The uptake of [14C]2DG and [18F]BCPP-EF differs in LPS- and IL4-stimulated polarized microglial cells. The present results suggest that the in vivo accumulation of metabolic tracers [18F]FDG and [18F]BCPP-EF can be influenced by the different aspects of neuroinflammation.


Scientific Reports
| (2021) 11:13200 | https://doi.org/10.1038/s41598-021-92436-0 www.nature.com/scientificreports/ diseases due to the overwhelming neuronal loss. The degree of [ 18 F]FDG uptake in the tissue depends on the balance of activities of neurons and glial cells, and also on the conditions of acute or chronic insults. Thus, [ 18 F] FDG should be re-evaluated to determine how it specifically accumulates in the brain region affected by acute injury in the presence of different types of neuroinflammation. In vivo energy metabolism can also be evaluated with our newly developed PET tracer, 2-tert-butyl-4-chloro-5-{6- [2-(2-[ 18 F]fluoroethoxy)-ethoxy]-pyridin-3-ylmethoxy}-2H-pyridazin-3-one ([ 18 F]BCPP-EF) 14,15 , which binds to the rotenone binding site on mitochondrial complex I (MC-I) 11,14,16 , the rate-limiting enzyme of mitochondrial oxidative phosphorylation (OXPHOS) 17 . In contrast to [ 18 F]FDG, which reflects glycolysis (especially anaerobic), [ 18 F]BCPP-EF is considered an indicator of mitochondrial activity, reflecting oxidative metabolism. In our previous study, [ 18 F]BCPP-EF accumulation in the compromised brain region was reduced, whereas the uptake of [ 18 F]FDG was increased, suggesting that [ 18 F]BCPP-EF might be a good indicator of neuronal damage or insensitivity to inflammatory glial activity 11,12 . However, the time course of proinflammatory and anti-inflammatory effects of microglia has been reported to be different following acute brain injury 18 , and the energy metabolism pathway activity is shown to change according to microglial activation states as follows: the glucose consumption rate is increased in LPS and interferon γ (INFγ)-stimulated microglia and decreased in IL4 stimulated microglia 19 , while the oxygen consumption rate (OCR), an OXPHOS index, is decreased in IFNγ-and LPS + IFNγ-stimulated microglia and increased in IL4-stimulated microglia 20,21 . Thus, microglial [ 18 F]FDG and [ 18 F]BCPP-EF accumulation could be determined by the microglial activation states according to the changes in energy metabolism activity. However, the effects of neuroinflammation on the brain accumulation of [ 18 F] FDG and [ 18 F]BCPP-EF was not been able to be identified, because neurons were mixed with glial cells in in vivo experiments such as our previous study.
Here, the aim of this study was to investigate whether the in vivo metabolic tracers [ 18 F]FDG and [ 18 F]BCPP-EF differ in their accumulations in differentially activated microglia. To specify the cellular accumulation of radiotracers in such microglia independently of neuron and the other glial cells, we used chemically-induced activated microglia in an in vitro experimental setting instead of in vivo settings where neuron and glial cells coexist.
Glycolysis-related gene expression in LPS-or IL4-stimulated microglial cells. The expression levels of glucose transporter (GLUT) 1 and 3 in LPS-or IL4-stimulated BV-2 and MG5 cells are shown in Fig. 2. GLUT1 mRNA levels increased with time in LPS-stimulated cells and did not change in IL4-stimulated cells ( Fig. 2A,B). On the other hand, GLUT3 mRNA levels were not changed in either LPS-or IL4-stimulated BV-2 and MG5 cells compared with control cells (p > 0.05; Fig. 2C,D).

Discussion
The present in vitro study of cultured microglial cells showed that the accumulation of the in vivo metabolic tracers [ 18 F]FDG and [ 18 F]BCPP-EF may depend on differences in the neuroinflammatory state (proinflammatory or anti-inflammatory) in accordance with differences in the activation of energy metabolism.
While these PET tracers used in brain research are designed to evaluate energy metabolism in the living brain, a variety of intrinsic cells in the brain, such as neurons and glia, share common energy metabolism pathways,  In the present study, we used LPS to induce proinflammatory activated states and IL4 to induce antiinflammatory activated states. We used two types of microglial cells, i.e., BV-2 cells, which were immortalized using v-raf/v-myc retrovirus 22 , and MG5 cells, which were derived from a p53-deficient mouse 23 , to exclude the effects of the immortalization procedure on energy metabolism profiles. To confirm the characteristics of the www.nature.com/scientificreports/ activated microglial cells, we measured changes in the expression of widely available pro-and anti-inflammatory markers through biochemical methods 24 . As shown in Fig. 1, LPS stimulation increased the expression levels of the proinflammatory markers TNF-α and iNOS from early time points after stimulation without increasing the expression of their anti-inflammatory counterparts, CD206 and Arg1. In contrast, compared with control cells, IL4-stimulated microglial cells showed increased expression levels of anti-inflammatory markers and unaltered levels of pro-inflammatory markers. These results convinced us to proceed with further experiments to evaluate radiotracer uptake in activated microglia. Prior to the radiotracer experiments, we confirmed that the uptake of two radiolabeled glucose analogs, i.e., [ (Fig. 5A,B). To investigate whether the change in [ 14 C]2DG uptake was in parallel with the concomitant change in the expression of glycolysisrelated genes in polarized microglial cells, we focused on changes in the gene expression of GLUTs and HKs during LPS-and IL4-stimulation (Figs. 2 and 3). We found that [ 14 C]2DG uptake in LPS-stimulated microglial cells increased as the gene expression of GLUT1 and HKs was elevated. Microglial GLUT3 expression levels www.nature.com/scientificreports/ were not affected by LPS or IL4 stimulation, in agreement with a previous report 21 , whereas GLUT3 expression in peripheral inflammatory cells has been reported to be increased via their activation 29 . In addition, several indicators of glycolytic activity, such as glucose consumption, extracellular lactate release, and the extracellular acidification rate (ECAR), were elevated in LPS-stimulated microglial cells, as reported in previous studies   19 . Glycolysis suppression in IL4-stimulated microglial cells might be caused by factors other than gene expression regulation, such as negative feedback allosteric regulation 31 and/or intracellular localization of HKs 32,33 . Although the basic mechanisms of glycolytic recruitment in LPS-stimulated microglia and reduced recruitment in IL4-stimulated microglia remain to be elucidated, it is likely that proinflammatory microglia can proliferate when the glycolytic metabolic pathway is enhanced possible under anaerobic conditions, whereas metabolic activities other than glycolysis are needed for the growth of anti-inflammatory microglia. The [ 18 F]BCPP-EF uptake study showed reduced uptake in LPS-stimulated microglial cells and increased uptake in IL4-stimulated microglial cells with time (Fig. 5C,D). To investigate whether [ 18 F]BCPP-EF uptake was associated with the gene expression levels of MC-I subunits in polarized microglia, we analyzed the expression levels of two MC-I subunit genes, NDUFS2 coding rotenone binding site 34 and NDUFA2 coding essential subunit for MC-I assembly 35 . Our assumption was that the expression levels of these genes could affect [ 18 F]BCPP-EF binding because [ 18 F]BCPP-EF is thought to bind to the rotenone binding site 11,14,16 that is encoded in NDUFA2 and because knockdown of NDUFS2 or NDUFA2 has been reported to reduce MC-I activity 35,36 . Contrary to our conjecture, it was shown that NDUFS2 and NDUFA2 gene expression levels were not markedly changed during LPS-and IL4-stimulation (Fig. 4) and that they were not correlated with [ 18 F]BCPP-EF uptake levels (Figs. 4,  5C,D). Despite this, it has been reported that OXPHOS activities vary according to the activated states. Indeed, both the basal OCR and the OXPHOS reserve are reportedly decreased in LPS-stimulated BV-2 cells 21,30 , while the basal OCR is increased in IL4-stimulated cells 20 . [ 18 F]BCPP-EF uptake in LPS-or IL4-stimulated microglia was in accordance with these OXPHOS activity changes. Since we evaluated gene expression levels but not MC-I activity in polarized microglia in the present study, it is too early to conclude that the different in [ 18 F]BCPP-EF uptake between polarized microglia reflects only a difference in MC-I activity. However, a previous study showing that LPS impairs MC-I activity without changing MC-I protein levels in mouse liver tissue 37 suggests that MC-I activity may be decreased due to the presence of inactive MC-I 38 without accompanying changes in MC-I gene or protein expression levels. Taken together, these findings suggest that the mechanism of low [ 18 F] BCPP-EF uptake in LPS-stimulated microglia and its high uptake in IL4-stimulated microglia might reside in the mitochondrial respiratory capacity irrespective of MC-I related gene expression or the amount of the protein, rather depending on the degree of presence of active or inactive form of MC-I (Fig. 6).
Different uptake patterns of [ 14 C]2DG and [ 18 F]BCPP-EF can be thought to result from the different degrees of dependance on anaerobic or aerobic metabolism in differentially polarized microglia. We estimated the degree of ATP production that shifted to anaerobic glycolysis or mitochondrial OXPHOS in LPS-or IL4-stimulated microglia by calculating the ratio of [ 18 F]BCPP-EF uptake to [ 14 C]2DG uptake in the same cells. It was found that the [ 18 F]BCPP-EF/[ 14 C]2DG uptake ratio decreased in LPS-stimulated microglia and increased in IL4-stimulated microglia (Fig. 5E,F). The decrease in [ 18 F]BCPP-EF/[ 14 C]2DG ratio in LPS-stimulated microglia was in line with a previous study showing that the OCR/ECAR ratio is reduced in rat primary cultured microglial cells after stimulation with LPS and IFN-γ 39 . In contrast, the [ 18 F]BCPP-EF/[ 14 C]2DG ratio was found to be increased in IL4-stimulated microglia in the present study, while the OCR/ECAR ratio has been reported to remain stable in rat primary cultured microglial cells after stimulation with IL4 + IL13 39 . The inconsistency between the present increase in the [ 18 F]BCPP-EF/[ 14 C]2DG ratio and the lack of change in the OCR/ECAR ratio in previously reported IL4 + IL13 stimulated microglia might be attributed to two things: misidentification of glycolysis activity by ECAR because of extracellular acidification irrelevant to glycolysis 40 and the different polarization states of IL4 alone-and IL4/IL13-stimulated cells. Despite this inconsistency, it is very likely that a change in the [ 18 F] BCPP-EF/[ 14 C]2DG uptake ratio reflects an alteration in metabolic dependence on glycolysis or mitochondrial OXPHOS in differentially activated microglia states.
In conclusion, the present in vitro uptake experiments demonstrated that [ 18 F]FDG uptake was increased in LPS-stimulated microglia and decreased in IL4-stimulated microglia and that [ 18 F]BCPP-EF uptake was decreased in LPS-stimulated microglia and increased in IL4-stimulated microglia, which is in agreement with previously reported metabolic changes. The current findings indicates that in vivo microglial accumulation of [ 18 F]FDG and [ 18 F]BCPP-EF could be changed according to the microglial activation states, which might cause the discrepancy between neural activity and brain [ 18 23 , were obtained from the Japanese Collection of Research Bioresources Cell Bank (Osaka, Japan). A1 cells were cultured in A1 growth medium (low-glucose DMEM (Nacalai Tesque) containing 10% heat inactivated FBS, 100 units/mL penicillin, and 100 µg/mL streptomycin) to prepare A1-conditioned medium 23 . MG5 cells were maintained in a 1:1 mixture of A1-conditioned medium and A1 growth medium and were seeded in 12-well plates 28 h before the experiments at a density of 2 × 10 5 cells/well in FBS and antibiotic-free low-glucose DMEM. BV-2 cells were seeded in 12-well plates 27 h before the experiments at the same density as MG5 cells in FBS and antibiotic-free high-glucose DMEM. The cells were then stimulated for 1, 6, 12, or 24 h before the experiments with 1000 ng/ mL LPS and 60 ng/mL IL4 to induce proinflammatory and anti-inflammatory activated microglia, respectively. Unstimulated cells were prepared under the same conditions and used as controls. All cells were incubated at 37 °C in humidified 95% air and 5% CO 2 .
Reverse transcription polymerase chain reaction (RT-PCR) analysis. Total RNA was extracted and purified from microglial cells using RNAiso Plus (Takara Bio, Shiga, Japan) according to the manufacturer's protocol. cDNA was obtained from 500 ng of total RNA using a ReverTra Ace RT-PCR RT Master Mix kit (TOYOBO, Osaka, Japan) following the manufacturer's instructions. cDNA (10 ng/well) was then subjected to PCR on a Thermal Cycler Dice Real Time System III instrument (Takara Bio) in a 20 µL reaction volume containing TB Green Premix Ex Taq II (Takara Bio) and 800 pmol each of the respective forward and reverse primers. The sequences of primers used are listed in Table 1. Samples were subjected to 40 cycles of amplification at 95 °C for 5 s and 60 °C for 30 s. All reactions were performed in duplicate, relative expression was quantified by the 2 − ΔΔCT method, and gene expression was normalized to the expression of β-actin as a housekeeping gene (n = 3). . For the blocking studies, 10 µmol/L rotenone or 10 mmol/L unlabeled 2-deoxy-D-glucose was coincubated with the radiotracers to assess the nonspecific uptake of the radiotracers. After three rapid washes with ice-cold PBS (+), the cells were lysed with 400 µL of 0.1 M sodium hydroxide aqueous solution. A portion of each lysate (300 µL) was transferred to a scintillation vial, and 18 F radioactivity was measured with an auto-well γ-counter (1480 Wizard 3, PerkinElmer). The protein concentration of another portion of each lysate was determined by the DC protein assay (Bio-Rad Laboratories, CA, USA). After the radioactivity of 18 F had decreased to the background level, the samples were diluted with 3 mL scintillation cocktail (Aquasol-2, PerkinElmer), and 14 C radioactivity was measured with a scintillation counter (Aloka LSC-6100, Hitachi, Tokyo, Japan). The net uptake of the radiotracers was normalized to the protein content. Specific uptake was calculated by subtracting nonspecific uptake, which was determined by the blocking studies, from total uptake. Statistical analysis. All data are expressed as the mean ± standard deviation (SD) of the fold change of the control group. Statistical analyses were performed with GraphPad Prism version 5 (GraphPad Software, CA, USA). The significance of differences between the control and drug-stimulated groups was determined by oneway analysis of variance (ANOVA) followed by Dunnett's multiple comparison test. The correlation between [ 18 F]FDG and [ 14 C]2DG uptake was analyzed by Pearson's correlation test. A p value of < 0.05 was regarded as statistically significant.

Data availability
All data supporting this study are available within the article and its supplementary material.

Gene
Forward primer (5′-3′) Reverse primer (5′-3′) CAT CCG TAA AGA CCT CTA TGC CAA C ATG GAG CCA CCG ATC CAC A   TNF-α  GTT CTA TGG CCC AGA CCC TCAC  GGC ACC ACT AGT TGG TTG TCT TTG   iNOS  GAC AGC ACA GAA TGT TCC AG  GAC TTG GAC TTG CAA GTG   CD206  CTG CAA GCA GCA GAA TGC TG  CAA CCA CTG CGT ACA CTC AG   Arg1  CAG AGT ATG ACG TGA GAG AC  GAG ATG CTT CCA ACT GCC AG   GLUT1  CAG GTG TTT GGC TTA GAC TC  CTC GAA GCT TCT TCA GCA CAC   GLUT3  CAG ACG CAA CTC TAT GCT TC  CTC TCC AAT GTA CAT AGG CAC   HK1  CTG GAA GAC ATT CGA ACA GAG  GAG CAG CTC TGG AGT AAT G   HK2  TCA TTG TTG GCA CTG GAA GC  TTG CCA GGG TTG AGA GAG AG   HK3  CTG GAA TGC TTG CAG CAA TTC  GTG TGG ACC TCA CGT ATG TG   NDUFS2  CAC GGA GAA GCT CAT TGA G  CTA TCG AAT AGG CCT GTT C   NDUFA2 CTG GGG TTG CGT GAG ATT C GAT CAG AAT GGG CAG GTT G