Contrast-enhanced T₁-weighted MRI, ¹¹C-DPA-713 PET and ¹¹C-CPPC PET as predictive imaging biomarkers of neuroinflammation in radiotherapy-induced brain injury

Abstract

Radiotherapy-induced brain injury (RIBI) is a chronic side effect that affects brain tumor survivors treated with radiotherapy. Neuroinflammation is a key contributor to RIBI. Thus, imaging methods capable of noninvasively monitoring neuroinflammation are needed. Although positron emission tomography (PET)-based radiotracers exist for imaging neuroinflammation, PET involves ionizing radiation which could be detrimental to pediatric patients already facing the risk of RIBI. Here, we evaluated the feasibility of developing contrast-enhanced T₁W MRI as a predictive biomarker of neuroinflammation in RIBI. Four groups of eight-week-old female BALB/c mice were stereotactically irradiated at 80 Gy and monitored longitudinally for neuroinflammation using ¹¹C-DPA-713 PET; ¹¹C-CPPC PET; gadoteridol-based contrast-enhanced T₁-weighted MRI; and TSPO, CD68, IBA1 immunohistochemistry. Our results showed that contrast-enhanced T₁W MRI was as effective as ¹¹C-DPA-713 PET; ¹¹C-CPPC PET and immunohistochemistry (P < 0.05, n = 3) in predicting neuroinflammation, by detecting subtle changes in the blood-brain barrier permeability that affected neuroinflammation changes.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-37264-w.

Introduction

Radiotherapy-induced brain injury (RIBI) is a chronic and debilitating side effect that causes cognitive impairment, and affects up to 90% of brain tumor survivors treated with radiotherapy^1–3. Given the greater vulnerability of developing brains to ionizing radiation, survivors of childhood brain tumors who were treated with radiotherapy at a young age are particularly prone to developing RIBI^4–7. Recent diagnostic and therapeutic advances in the management of pediatric brain tumors have greatly improved the overall survival rates of pediatric brain tumor patients⁵. Since more pediatric brain tumor patients currently survive childhood cancers, and live into adulthood, there has been an increase in the incidence of RIBI^4–8.

RIBI detection usually relies on neuropsychological evaluations^9–12. However, neuropsychological evaluations typically detect RIBI at the later stages of injury, when it is already symptomatic^9–12. Consequently, there is a need for noninvasive diagnostic methods capable of detecting RIBI early on, for its effective management^9,10. There is also a need for diagnostic methods capable of monitoring patient responses to potential RIBI therapies^13–15.

Neuroinflammation has been shown to play a critical role in the progression of RIBI and RIBI-associated cognitive impairment^1–3. Briefly, irradiation generates reactive oxygen species (ROS) which cause oxidative stress and destroy vulnerable endothelial cells^16,17. The damaged endothelial cells then release cytokines, which activate the neuroimmune system into a pro-inflammatory state via the activation of resident microglial cells and astrocytes^16,17. The damaged endothelial cells also cause the disruption of the blood-brain barrier (BBB), which further enhances neuroinflammation by facilitating the infiltration of peripheral macrophages^16,17. Additionally, radiation-induce oxidative stress directly activates resident microglial cells and astrocytes into a pro-inflammatory state^1,3. In this activated pro-inflammatory state, microglial cells and astrocytes play an important role in the phagocytosis of damaged cells, and also secrete cytokines which further exacerbate neuroinflammation by recruiting infiltrative peripheral macrophages^1,3. Although activated microglial cells and astrocytes also help with tissue repair, prolonged activation of these microglial cells and astrocytes causes further cerebrovascular damage, which leads to ischemia, glial and neuronal cell death, neurodegeneration, and cognitive impairment^1,3.

Thus, there is a great need for the development of clinically translatable imaging strategies capable of noninvasively and longitudinally monitoring the status of neuroinflammation during RIBI progression, for the effective management of RIBI^1,17,18. There is also a need for the development of clinically translatable neuroinflammation imaging strategies capable of enabling the noninvasive evaluation of potential RIBI prophylactic and therapeutic agents that are currently being developed^19–22. Preclinical rodent models of RIBI could play a critical role in enabling the development and evaluation of clinically translatable neuroinflammation imaging strategies for RIBI^1,13.

Several positron emission tomography (PET)-based radioligands have been developed and used to noninvasively study the role of neuroinflammation in a number of neurodegenerative disorders^23–26. These PET radioligands generally target endogenous receptors that are overexpressed during the inflammatory cascade such as the 18-kDa translocator protein (TSPO)^23–26. The 18-KDa translocator protein (TSPO), formerly known as the peripheral benzodiazepine receptor (PBR), is a mitochondrial transmembrane protein that is expressed on the outer mitochondrial membrane and has been postulated to have various functional roles including immune modulation^23–26. Although TSPO is generally expressed at low levels on neuroimmune cells such as microglia cells and astrocytes, the TSPO expression levels on these neuroimmune cells is upregulated when the neuroimmune cells are activated to a pro-inflammatory state in response to brain injury and in a number of CNS disorders^23–26. Furthermore peripheral macrophages also express TSPO and infiltrate the brain when the BBB is permeable, and this further exacerbates neuroinflammation^24,26. Thus, high TSPO levels are considered a viable early biomarker of neuroinflammation in CNS disorders including brain injuries^23–26.

Several PET radioligands that targets TSPO have been developed and used to study the role of neuroinflammation in a number of neurodegenerative and psychiatric disorders^23–28. ¹¹C-N, N-diethyl-2-[2-(4-methoxyphenyl)-5,7dimethyl-pyrazolo[1,5-a]pyrimidin-3-yl]-acetamide (¹¹C-DPA-713) is a second generation TSPO PET radioligand that has been extensively used to study the role of neuroinflammation in a number of CNS disorders^29–34. However, because most TSPO radioligands are not specific to activated microglial cells but also target upregulated TSPO expression on activated astrocytes and infiltrative macrophages, other PET radioligands that target receptors specific to activated microglial cells have been developed to evaluate the specific contributions of activated microglia cells in neuroinflammation^24,26.

The macrophage colony-stimulating factor 1 receptor (CSF1R) is a cell surface receptor that is specifically expressed on microglial cells, and not on other brain cells^35,36. It controls the survival, development, and maintenance of microglia, and plays a pivotal role in neuroinflammation. 5-cyano-N-(4-(4-[11 C]methylpiperazin-1-yl)-2-(piperidin1-yl)phenyl)furan-2-carboxamide (¹¹C-CPPC) is a PET radiotracer for CSF1R that has been used to specifically image upregulated CSF1R expression in activated microglia during disease progression³⁵.

However, despite the availability of these PET-based radioligands for imaging neuroinflammation, PET involves the use of ionizing radiation, which could be particularly detrimental to pediatric patients, especially pediatric brain tumor survivors who were treated at a young age with radiotherapy and already face the risk of RIBI.

Magnetic resonance imaging (MRI) is the premier imaging modality used in the diagnosis and management of brain tumors. MRI is particularly suitable for monitoring pediatric patients, especially pediatric brain tumor survivors who were treated at a young age with radiotherapy and already face the risk of RIBI, since MRI doesn’t involve the use of ionizing radiation. Thus, MRI offers the possibility of noninvasively and longitudinally monitoring pediatric brain tumor survivors in the absence of ionizing radiation, and at high spatial resolution^37–43. Given the significant role of neuroinflammation in the progression of RIBI, developing clinically translatable MRI strategies capable of noninvasively and longitudinally monitoring the status of neuroinflammation during RIBI progression could significantly improve the management of RIBI in pediatric brain tumor survivors^17,18. These MRI strategies could also enable the evaluation of potential RIBI prophylactic and therapeutic agents, that are currently being developed^19–22.

Previously, we showed that contrast-enhanced T₁-weighted (T₁W) MRI using gadoteridol (ProHance^®) as a contrast agent was the most sensitive of the conventional MRI pulse sequences for detecting the onset of RIBI, via the detection of the blood-brain barrier (BBB) permeability⁴⁴. We also showed that contrast-enhanced T₁W MRI could be used to longitudinally detect subtle changes in the BBB permeability over a 12-month post-irradiation period. Furthermore, we showed that the maximum BBB permeability detected using contrast-enhanced T₁W MRI coincided with the maximum neuroinflammation detected using cluster of differentiation 68 (CD68) and ionized calcium binding adaptor molecule 1 (IBA1) immunohistochemistry. Finally, we showed that the detection of this simultaneous maximum BBB permeability and maximum neuroinflammation detection also coincided with the detection of the onset of cognitive impairment in our preclinical model of RIBI⁴⁴. These findings suggested the possibility of using contrast-enhanced T₁W MRI as a predictive biomarker of neuroinflammation in RIBI.

In this study, we hypothesized that the indirect detection of neuroinflammation via the detection of BBB permeability, using contrast-enhanced T₁W MRI, will be as effective as the direct detection of neuroinflammation using ¹¹C-DPA-713 PET and ¹¹C-CPPC PET, respectively, in our previously developed preclinical model of RIBI. Thus, we expanded on our previous findings and carried out a comparative study to see if the indirect detection of neuroinflammation via the detection of BBB permeability, using contrast-enhanced T₁W MRI was as effective as the detection of neuroinflammation using ¹¹C-DPA-713 PET and ¹¹C-CPPC PET, respectively, in our previously developed preclinical model of RIBI (Fig. 1).

Fig. 1.

Schematic of the experimental design.

Ultimately, this method could provide a simple yet effective means of noninvasively probing the status of neuroinflammation in RIBI progression using conventional MRI, which is already used routinely in the clinical setting to monitor and manage brain tumor patients and survivors. Additionally, this method could provide a simple preclinical and clinically translatable means of longitudinally monitoring the therapeutic efficacy of potential RIBI prophylactic and therapeutic agents that are currently in development^1,13,45.

Materials and methods

Ethics Declarations for the Use of Live Vertebrates in Research. All live animal procedures and methods were approved prior to the experiments by the Johns Hopkins University Animal Care and Use Committee (JHU ACUC). All live animal procedures and methods were carried out in accordance with all relevant guidelines and regulations. Live animal procedures and methods are reported in accordance with ARRIVE 2.0 guidelines⁴⁶.

Animals. Female three-to-four-week-old immune competent BALB/cAnNTac (BALB/c) mice were obtained from Taconic Biosciences. The mice were imaged and then irradiated at eight weeks old. The mice irradiated at eight weeks old are roughly the equivalent of humans irradiated as middle to late adolescents (15-21-years-old)^47–49. Female mice were used in our current study to expand on our previous finding with this group of mice⁴⁴. All animal procedures were approved by the Johns Hopkins University Animal Care and Use Committee (JHU ACUC).

Study Design. Four groups of eight-week-old female BALB/c mice were used in this study (Fig. 1). The mice were imaged and then irradiated at eight weeks old, as previously described^50–53. The mice were then monitored longitudinally for neuroinflammation over a period of nine months post-irradiation (IR) at the following time points: (1) One-month post-IR; (2) two months post-IR; (3) six months post-IR; and (4) nine months post-IR. Whereas Group 1 mice were monitored using ¹¹C-DPA-713 PET, Group 2 mice were monitored using ¹¹C-CPPC PET, and Group 3 mice were monitored using gadoteridol (ProHance^®)-based contrast-enhanced T₁-weighted (T₁W) MRI. Finally, Group 4 mice were monitored with immunohistochemistry (IHC), using 18-KDa translocator protein (TSPO) and ionized calcium binding adaptor molecule 1 (IBA1) IHC respectively for the detection of activated microglial cells, and cluster of differentiation 68 (CD68) for the detection of infiltrative peripheral macrophages. All the mice were euthanized at the end of the study period (nine months after irradiation).

Irradiation. All mice were anesthetized with a 2% isoflurane and oxygen mixture, and then irradiated using a small animal radiation research platform (SARRP, Xstrahl Inc., Suwanee, GA) as previously described^50–53. Computed tomography (CT) images were acquired and a target in the right brain hemisphere (with iso-center at the hippocampus), was chosen on the CT images using the coordinates below: Anterior posterior (AP) = + 3 mm relative to the lambda; medial lateral (ML) = + 0.5 mm relative to the midline; and dorsal ventral (DV) = − 2 mm relative to the skull base. The mice were then irradiated at a dose of 80 Gy and a dose rate of 1.7 Gy per minute, using a single image-guided focal X-ray beam with a 3 mm x 3 mm collimator.

Radiopharmaceutical Preparation of ¹¹C-DPA-713 and ¹¹C-CPPC. ¹¹C-DPA-713 and ¹¹C-CPPC were synthesized at the Johns Hopkins Positron Emission Tomography Radiotracer Center under Investigational New Drug number 78,283, and were obtained with radioactive purity greater than 95%, as previously reported^33,35,54. Carbon-11 labeled tracers (¹¹C-DPA-713 and ¹¹C-CPPC) were chosen over of fluorine-18 labeled tracers, despite their shorter half-lives, because these carbon-11-labeled tracers are readily available at our institution and are routinely used in clinical trials at our institution^{31–33,35,54}.

¹¹C-DPA-713 and ¹¹C-CPPC Positron Emission Tomography and Magnetic Resonance Imaging (PET/MRI). All ¹¹C-DPA-713 and ¹¹C-CPPC PET/MRIs were acquired using a preclinical 7T Bruker Biospec 70/30 horizontal bore simultaneous PET/MRI scanner, equipped with a three-ring Bruker Si198 PET insert and an ultra-shield refrigerated magnet (Ettlingen, Germany). A 72 mm, PET-optimized proton transmit-receive Bruker volume radiofrequency coil centered in the PET detector was used in this study. The Paravision 360 software was used for the acquisition of all images.

All mice were anesthetized prior to each scan and maintained under anesthesia throughout the scans using a 2% isoflurane and air mixture. A catheter was placed in the tail veins of the anesthetized mice and then the mice placed in the scanner. Dynamic PET scans were used in this study, and dynamic frames were acquired as follows (frames × time(s)): 3 × 10 s, 2 × 15 s, 2 × 30 s, 3 × 60 s, and 10 × 150 s. 30-minute dynamic PET scans were initiated 30 s prior to the intravenous administration of the PET radiotracers. Each mouse was intravenously administered 150 µCi (5.55 MBq) of a radiotracer 30 s after the start of the dynamic PET scans.

Simultaneous anatomical brain T₂-weighted (T₂W) MRI scans were also acquired during the 30-minute dynamic PET scans as described below. Anatomical in vivo axial T₂W MRIs were acquired using a rapid acquisition with refocused echoes (RARE) pulse sequence with the following acquisitioning parameters: echo time = 25 milliseconds (ms); RARE factor = 8; repetition time = 2500 ms; number of averages = 10; field of view = 25 × 30 mm; matrix size = 100 × 120 mm; and slice thickness = 0.5 mm. Anatomical in vivo coronal T₂W brain MRIs were acquired using a RARE pulse sequence with the following acquisitioning parameters: echo time = 30 ms; RARE factor = 8; repetition time = 3000 ms; number of averages = 10; field of view = 25 × 25 mm; matrix size = 133 × 133 mm; and slice thickness = 0.5 mm. The acquired PET images were reconstructed using a three-dimensional maximum-likelihood expectation-maximization (MLEM) iterative image reconstruction algorithm with a pixel size of 0.5 mm and 18 iterations. The PET images were spatially co-registered using a geometry-based approach with anatomical T₂W brain MRI images as a reference in the Bruker Paravision 360 software and converted to the Digital Imaging and Communications in Medicine format (DICOM).

Final image analyses were performed using the PMOD software version 4.004 (PMOD Technologies, Zurich, Switzerland) as described below. Briefly, for each mouse, two regions of interest (ROIs) were drawn on each axial MRI image to cover the whole brain for the analysis (Fig. S1). The first ROI was drawn around the whole right-brain (irradiated) hemisphere, while the second ROI was drawn around the whole left-brain (non-irradiated) hemisphere. We opted to use whole brain hemisphere ROIs in this comparative study, so we could keep the ROI sizes constant throughout the nine-month study period. This was important since the PET signal changed over the nine-month study period. Time-activity curves were then generated for the respective ROIs of each mouse (Fig. S2 and S3). The data was then presented as a ratio of the radioactivity in the irradiated brain hemisphere (ROI 1) divided by the radioactivity in the non-irradiated hemisphere (ROI 2) at each time point after the administration of the respective ¹¹C-DPA-713 and ¹¹C-CPPC PET radiotracers (Fig. S2 and S3). A sample size of three was used in the analysis of each time point of the PET/MRI study.

In this study, volumes of interest (VOI) analyses were not performed on the respective ¹¹C-DPA-713 and ¹¹C-CPPC PET/MRI data. Since the contrast-enhanced T₁-weighted MRI data analysis and the immunohistochemistry data analysis were all done using the ROI analysis method, the same ROI analysis method was used for the respective ¹¹C-DPA-713 PET and ¹¹C-CPPC PET data, in order to standardize the comparison.

Contrast-enhanced T₁-weighted Magnetic Resonance Imaging (MRI). All mice were anesthetized with a 2% isoflurane and air mixture. The contrast-enhanced T₁-weighted (T₁W) MRIs were then acquired using an 11.7T Bruker Biospec horizontal bore scanner (Billerica, MA), equipped with a 23 mm Bruker volume radiofrequency coil and the Paravision 6.1.0 software, as previously reported⁴⁴. Briefly, in vivo contrast-enhanced T₁W MRIs were acquired after the bolus intravenous administration of 200 µL of a 0.25 M (2 mmol/Kg) gadoteridol (ProHance^®) solution (Bracco Diagnostics Inc., Singen, Germany), prepared in 0.01 M phosphate buffered saline (PBS). The images were then acquired using a rapid acquisition with refocused echoes (RARE) pulse sequence, with the acquisition parameters below: Echo time = 5.5 milliseconds (ms); effective echo time = 5.5 ms; RARE factor = 4; repetition time = 557 ms; number of averages = 3; field of view = 18 × 18 mm; matrix size = 256 × 256 pixels; and slice thickness = 0.5 mm. The contrast-enhanced T₁-weighted MRI analyses were next performed using the NIH ImageJ software, as previously reported⁴⁴. In brief, two regions of interest (ROIs) were drawn on each axial image to cover the whole brain for the analyses (Fig. S1). The first ROI was drawn around the whole right brain (irradiated) hemisphere, while the second ROI was drawn around the whole left brain (contralateral non-irradiated) hemisphere. We opted to use whole brain hemisphere ROIs to keep the sizes of the ROIs constant throughout the nine-month study period. This was necessary, since the size of the regions of contrast enhancement changed over the nine-month study period. Bright pixel analyses of the defined ROIs were then performed, as previously reported^44,55–58. A sample size of three was used in the analysis of each time point of the contrast-enhanced T₁W MRI study.

Immunohistochemistry. All Group 4 mice were transcardially perfused with 30 mL of heparinized phosphate buffered saline (1% heparin in 0.01 M PBS), followed by 30 mL of a 4% paraformaldehyde (PFA) solution (4% PFA in 0.01 M PBS), at a flow rate of 10 mL/min. The mouse skulls were next stored overnight in 4% PFA at 4 °C, then in PBS at 4 °C for 48 h. The brains were next extracted from the skulls, paraffin-embedded, sectioned into 4 μm slices, and stained for immunohistochemistry. For neuroinflammation detection, a rabbit anti-TSPO (1:100, ab109497, Abcam) and a rat anti-IBA1 (1:100, ab283346, Abcam) primary antibody were used, followed by an Alexa Fluor^® 488 goat anti-rabbit (1:200, ab150077, Abcam) and an Alexa Fluor^® 647 goat anti-rat (1:1000, ab150167, Abcam) secondary antibody, respectively. All slides were then counter stained with DAPI in fluoroshield mounting media (ab104139, Abcam). Fluorescence image acquisitioning was performed using a Zeiss Axio Scan.Z.1 Slide Scanner, and all image analyses were performed using the NIH ImageJ software. Two regions of interest (ROIs) were drawn on each fluorescence image to cover the whole brain for the analysis, (Fig. S1). Pixel analysis of the defined ROIs was then performed, as previously reported^44,56. A sample size of three was used in the analysis of each time point of the immunohistochemistry study.

Euthanasia Method. All mice used for in vivo imaging were euthanized at nine months post-irradiation using carbon dioxide, in accordance with the American Veterinary Medical Association (AVMA). The euthanasia protocol was approved by the Johns Hopkins University Animal Care and Use Committee (JHU ACUC). Briefly, the mouse cage lid was removed, and the cage was placed in a gas chamber without prefilling the chamber. Carbon dioxide (100%) was then introduced into the gas chamber at a displacement rate of 50% of the gas chamber volume per minute. The carbon dioxide was allowed to flow for four minutes, and then the mice were maintained in the carbon dioxide filled gas chamber for another 10 min, for a total of 14 min of carbon dioxide exposure. We then verified that the mice were dead by the absence of response to stimuli. The mice were then disposed of appropriately.

Statistical Analyses. All ¹¹C-DPA-713 and ¹¹C-CPPC PET/MRI, contrast-enhanced T₁W MRI, and immunohistochemistry data points were presented as the mean ± standard deviation of three independent experiments. All statistical analyses were done using the GraphPad Prism 10.4.1 software. Two dataset comparisons were made using either unpaired or paired two-tailed student t-tests (except when otherwise stated), while multiple comparisons were made using the ordinary two-way ANOVA test. All results were considered statistically significant at P < 0.05.

Results

Detection of neuroinflammation in RIBI using ¹¹C-DPA-713 PET/MRI. In non-irradiated mice, PET/MRIs of the spatial biodistribution of the ¹¹C-DPA-713 radiotracer in the mouse brains showed no differences between the brain hemispheres at any time point after the bolus intravenous administration of ¹¹C-DPA-713 (Figs. 2 and 3 and Fig. S2). This indicated the absence of neuroinflammation in the non-irradiated mice, as was expected. Quantification of the time-radioactivity kinetic curve ratios of the right brain hemispheres normalized to the corresponding contralateral left-brain hemispheres of the non-irradiated mice, after the bolus intravenous administration of ¹¹C-DPA-713, showed an average radioactivity uptake ratio of 1.21 ± 0.06 between 10 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 (Figs. 2 and 3, and Fig. S2).

Fig. 2.

Detection of neuroinflammation in RIBI using ¹¹C-DPA-713 PET/MRI. PET/MRIs of the radioactivity biodistribution of the ¹¹C-DPA-713 radiotracer in the brain before, 2, 5, 10, 15 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713: (a) Before irradiation (non-irradiated control mice); (b) one month post-irradiation; (c) two months post-irradiation; (d) six months post- irradiation; and (e) nine months post-irradiation. A high accumulation and retention of ¹¹C-DPA-713 was detected in the irradiated brain hemispheres of the mice at one month and two months post-irradiation (red arrows). This ¹¹C-DPA-713 accumulation in the irradiated brain hemispheres of the mice at these early time points post-irradiation was indicative of the presence of neuroinflammation. Quantification of the time-radioactivity kinetic curve uptake ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the mice after the bolus intravenous administration of ¹¹C-DPA-713: (f) Before irradiation (non-irradiated control mice); (g) one month post-irradiation versus before irradiation (P = 0.0001, n = 3); (h) two months post-irradiation versus before irradiation (P = 0.0260, n = 3); (i) six months post-irradiation versus before irradiation; and (j) nine months post-irradiation versus before irradiation. Statistically significant accumulation and retention of the ¹¹C-DPA-713 radiotracer was detected in the irradiated brain hemispheres compared to the corresponding contralateral non-irradiated brain hemispheres of the mice at one month (P = 0.0001, n = 3) and two months (P = 0.026, n = 3) post-irradiation. This high ¹¹C-DPA-713 accumulation and retention at these early time points post-irradiation were indicative of the presence of neuroinflammation in the irradiated brain hemispheres.

Fig. 3.

Summary of the detection of neuroinflammation in RIBI using ¹¹C-DPA-713 PET/MRI over a nine-month post-irradiation period. Representative PET/MRIs of the radioactivity biodistribution of the ¹¹C-DPA-713 radiotracer in mouse brains over the nine-month post-irradiation study period. The PET/MRIs were acquired before, 10, 15 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713: (a) Coronal view; and (b) Axial view. (c) Quantification of the time-radioactivity kinetic curve uptake ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the respective mice, over 30 min after the bolus intravenous administration of ¹¹C-DPA-713, at different time points before and after irradiation (P < 0.05). (d) Radioactivity uptake ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the respective mice, 0, 10, 13, 15, 18, and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 at different time points before and after irradiation (P < 0.05).

In irradiated mice, the highest accumulation and retention of the ¹¹C-DPA-713 radiotracer in the irradiated brain hemispheres of the mice throughout the nine-month study period was detected at one month post-irradiation. At one month post-irradiation, the greatest differences between the irradiated brain hemispheres (right) and the corresponding contralateral non-irradiated (left) brain hemispheres were detected between 10 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 (Figs. 2 and 3, Fig. S2, Table S1, S2, and S3). Quantification of the time-radioactivity kinetic curve ratios of the irradiated brain hemispheres normalized to the corresponding contralateral non-irradiated brain hemispheres of the respective mice showed a statistically significant increase in the radioactivity uptake ratios at one month post-irradiation, compared to the ratios detected in the non-irradiated control mice (P = 0.0001, n = 3). An average radioactivity uptake ratio of 1.67 ± 0.11 was detected between 10 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713, at one-month post irradiation (Figs. 2 and 3, Fig. S2, Table S1, S2, and S3). This ratio detected at one month post-irradiation was the highest radioactivity uptake ratio detected throughout the nine-month study period. This indicated the presence of maximum neuroinflammation in the irradiated brain hemispheres of the mice at one month post-irradiation in the preclinical mouse model. This finding is in accordance with our previous report⁴⁴.

At two months post-irradiation, the ¹¹C-DPA-713 radiotracer was still detected in the irradiated brain hemispheres compared to the corresponding contralateral non-irradiated brain hemispheres of the mice between 10 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 (Figs. 2 and 3, Fig. S2, Table S1, S2, and S3). Quantification of the time-radioactivity kinetic curve ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral non-irradiated (left) brain hemispheres of the mice after the bolus intravenous administration of ¹¹C-DPA-713 showed a statistically significant increase in the radioactivity uptake ratios at two months post-irradiation, compared to the ratios detected in the non-irradiated control mice (P = 0.026, n = 3). An average radioactivity uptake ratio of 1.35 ± 0.19 was detected between 10 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 at two months post irradiation (Figs. 2 and 3, Fig. S2, Table S1, S2, and S3). Although this radioactivity uptake ratio was higher than that detected in the non-irradiated mice, it was lower than the ratio detected at one month post-irradiation (Figs. 2 and 3, Fig. S2, Table S1, S2 and S3). This indicated a reduction in the neuroinflammation signal detected in the irradiated brain hemispheres at two months post-irradiation compared to that detected at one month post-irradiation in the preclinical mouse model. This finding is also in accordance with our previous report⁴⁴.

At six months and nine months post-irradiation, no statistically significant differences were detected in the spatial biodistribution of the ¹¹C-DPA-713 radiotracer in the irradiated brain hemispheres compared to the corresponding contralateral non-irradiated brain hemispheres of the mice, at any time point after the bolus intravenous administration of ¹¹C-DPA-713 (Figs. 2 and 3, Fig. S2, Table S1, S2 and S3). Average radioactivity uptake ratios of 1.18 ± 0.07 and 1.17 ± 0.04 were detected between 10 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 at six months and nine months post-irradiation, respectively (Figs. 2 and 3, Fig. S2, Table S1, S2, and S3). These respective average radioactivity uptake ratios were similar to the average radioactivity uptake ratio detected in the non-irradiated control mice. Thus, this indicated a further reduction in the neuroinflammation signal detected in the irradiated brain hemispheres of the mice at six months and nine months post-irradiation compared to that detected at two months post-irradiation. This finding is also in accordance with our previous report⁴⁴.

Next, to evaluate the specificity of the ¹¹C-DPA-713 radiotracer to detect neuroinflammation in RIBI, a blocking study was carried out in the irradiated mice at two months post-irradiation (Figs. 4 and 5, and Table S4). Quantification of the time-radioactivity kinetic curve ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral non-irradiated (left) brain hemispheres of the mice that received the blocking agent (unlabeled DPA-713) before the bolus intravenous administration of ¹¹C-DPA-713 at two months post-irradiation showed a statistically significant decrease in the radioactivity uptake ratios, compared to the radioactivity uptake ratios in the mice that did not receive the blocking agent before the bolus intravenous administration of ¹¹C-DPA-713 (P < 0.0001, n = 3). An average radioactivity uptake ratio of 1.19 ± 0.09 was detected between 10 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 in the mice that received the blocking agent compared to 1.35 ± 0.09 in the mice that did not receive the blocking agent at two month post irradiation. Additionally, the average radioactivity uptake ratio in the blocking study was similar to that detected in the non-irradiated control mice. This confirmed the specificity of the ¹¹C-DPA-713 radiotracer to detect TSPO in the preclinical mouse model of RIBI.

Fig. 4.

¹¹C-DPA-713 PET/MRI blocking study. PET/MRIs of the radioactivity biodistribution of the ¹¹C-DPA-713 radiotracer in the brain before, 2, 5, 10, 15 and 20 min after the bolus intravenous administration of ¹¹C-DPA-713: (a) Before irradiation (non-irradiated control mice); (b) two months post-irradiation; (c) two months post-irradiation, 15 min after the bolus intravenous administration of 0.025 mg (1 mg/Kg) of the blocking agent (unlabeled DPA-713). Quantification of the time-radioactivity kinetic curve ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the mice after the bolus intravenous administration of ¹¹C-DPA-713: (d) Before irradiation (non-irradiated control mice); (e) two months post-irradiation versus before irradiation (P = 0.0001, n = 3); (f) two months post-irradiation with the blocker versus two months post-irradiation without the blocker versus before irradiation (P = 0.0260, n = 3). A decrease was detected in the ¹¹C-DPA-713 uptake ratio of the mice that received the blocking agent compared to the mice that did not receive the blocking agent at two months post-irradiation. This ¹¹C-DPA-713 uptake ratio in the mice that received the blocking agent at two months post-irradiation was similar to that detected in the non-irradiated control mice. This confirmed the specificity of the ¹¹C-DPA-713 to detect TSPO in the preclinical mouse model of RIBI.

Fig. 5.

Summary of the ¹¹C-DPA-713 PET/MRI blocking study. PET/MRIs of the radioactivity biodistribution of the ¹¹C-DPA-713 radiotracer in the brain before, 10, and 20 min after the bolus intravenous administration of ¹¹C-DPA-713 in non-irradiated control mice; in irradiated mice at two months post-irradiation; and in irradiated mice at two months post-irradiation, 15 min after the bolus intravenous administration of 0.025 mg (1 mg/Kg) of the blocking agent (unlabeled DPA-713), respectively: (a) Coronal view; and (b) Axial view. (c) Quantification of the time-radioactivity kinetic curve uptake ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the mice, after the bolus intravenous administration of ¹¹C-DPA-713 in non-irradiated control mice; in irradiated mice at two months post-irradiation (P = 0.0001, n = 3); and in irradiated mice at two months post-irradiation, 15 min after the bolus intravenous administration of 0.025 mg (1 mg/Kg) of the blocking agent (unlabeled DPA-713). A decrease was detected in the ¹¹C-DPA-713 uptake ratio of the mice that received the blocking agent compared to the mice that did not receive the blocking agent at two months post-irradiation. This confirmed the specificity of the ¹¹C-DPA-713 to detect TSPO in the preclinical mouse model of RIBI.

Detection of neuroinflammation in RIBI using C-CPPC PET/MRI. Overall, a similar neuroinflammation trend to that detected using ¹¹C-DPA-713 was observed using the ¹¹C-CPPC radiotracer. However, the radioactivity uptake ratios quantified for the time-radioactivity kinetic curves were lower for the ¹¹C-CPPC radiotracer compared to those quantified for the ¹¹C-DPA-713 radiotracer. Furthermore, the greatest differences between the irradiated brain hemispheres (right) and the corresponding contralateral non-irradiated (left) brain hemispheres were detected between 20 and 30 min after the bolus intravenous administration of the ¹¹C-CPPC radiotracer, instead of between 10 and 20 min of radiotracer administration as observed with the ¹¹C-DPA-713 (Figs. 6 and 7 and Fig. S3).

Fig. 6.

Detection of neuroinflammation in RIBI using ¹¹C-CPPC PET-MRI. PET-MRIs of the radioactivity biodistribution of the ¹¹C-CPPC radiotracer in the brain before, 2, 17.5, 20, 22.5, 25, 27.5, and 30 min after the bolus intravenous administration of ¹¹C-CPPC: (a) Before irradiation (non-irradiated control mice); (b) one month post-irradiation; (c) two months post-irradiation; (d) six months post-irradiation; and (e) nine months post-irradiation. A high accumulation and retention of ¹¹C-CPPC was detected in the irradiated brain hemispheres of the mice, at one month and two months post-irradiation (red arrows). This ¹¹C-CPPC accumulation in the irradiated brain hemispheres of the mice at these early time points post-irradiation was indicative of the presence of neuroinflammation. Quantification of the time-radioactivity kinetic curve ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the mice after the bolus intravenous administration of ¹¹C-CPPC: (f) Before irradiation (non-irradiated control mice); (g) one month post-irradiation versus before irradiation (P = 0.0002, n = 3, between 20 to 30 min post-tracer administration); (h) two months post-irradiation versus before irradiation (P < 0.0001, n = 3, between 20 to 30 min post-tracer administration); (i) six months post-irradiation versus before irradiation; and (j) nine months post-irradiation versus before irradiation. Statistically significant accumulation and retention of the ¹¹C-CPPC radiotracer was detected in the irradiated brain hemispheres compared to the corresponding contralateral non-irradiated brain hemispheres of the mice at one month and two months post-irradiation between 20- to 30- minutes post-tracer administration. This high ¹¹C-CPPC accumulation and retention at these early time points post-irradiation were indicative of the presence of neuroinflammation in the irradiated brain hemispheres.

Fig. 7.

Summary of the detection of neuroinflammation in RIBI using ¹¹C-CPPC PET/MRI over a nine-month post-irradiation period. Representative PET/MRIs of the radioactivity biodistribution of the ¹¹C-CPPC radiotracer in mouse brains over the nine-month post-irradiation study period. The PET/MRIs were acquired 30 min after the bolus intravenous administration of ¹¹C-CPPC: (a) Coronal view; and (b) Axial view. (c) Quantification of the time-radioactivity kinetic curve uptake ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the respective mice, over 30 min after the bolus intravenous administration of ¹¹C-CPPC at different time points before and after irradiation (P < 0.05, n = 3). (d) Radioactivity ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the respective mice, 10, 15, 20, 25, 27.5 and 30 min after the bolus intravenous administration of ¹¹C-CPPC at different time points before and after irradiation (P < 0.05, n = 3, unpaired, one-tailed).

In the non-irradiated mice, PET/MRIs of the spatial biodistribution of the ¹¹C-CPPC radiotracer in the mouse brains showed no differences between the brain hemispheres at any time point after the bolus intravenous administration of ¹¹C-CPPC (Figs. 6 and 7 and Fig. S3). This indicated the absence of neuroinflammation in the non-irradiated mice, as was expected, and similar to what was observed with the ¹¹C-DPA-713 radiotracer. Quantification of the time-radioactivity kinetic curve ratios of the right-brain hemispheres normalized to the corresponding contralateral left-brain hemispheres of the non-irradiated mice, after the bolus intravenous administration of ¹¹C-CPPC showed an average radioactivity uptake ratio of 1.02 ± 0.07 between 20 and 30 min after the bolus intravenous administration of ¹¹C-CPPC (Figs. 6 and 7, and Fig. S3).

In irradiated mice, the greatest accumulation and retention of the ¹¹C-CPPC radiotracer in the irradiated brain hemispheres of the mice throughout the nine-month study period was detected at one month post-irradiation. This finding is similar to what was observed with the ¹¹C-DPA-713 radiotracer. At one month post-irradiation, the greatest differences between the irradiated brain hemispheres (right) and the corresponding contralateral non-irradiated (left) brain hemispheres were detected between 20 and 30 min after the bolus intravenous administration of the ¹¹C-CPPC radiotracer (Figs. 6 and 7, Fig. S3, Table S5, S6, and S7). Quantification of the time-radioactivity kinetic curve ratios of the irradiated brain hemispheres normalized to the corresponding contralateral non-irradiated brain hemispheres of the mice showed a statistically significant increase in the radioactivity uptake ratios at one month post-irradiation, compared to the radioactivity uptake ratios detected in the non-irradiated control mice (P = 0.0002, n = 3). An average radioactivity uptake ratio of 1.30 ± 0.15 was detected between 20 and 30 min after the bolus intravenous administration of ¹¹C-CPPC at one month post-irradiation (Figs. 6 and 7, Fig. S3, Table S5, S6 and S7). This average radioactivity uptake ratio detected at one month post-irradiation was the highest radioactivity uptake ratio detected throughout the nine-month study period. This indicated the presence of maximum neuroinflammation in the irradiated brain hemispheres of the mice at one month post-irradiation in the preclinical mouse model. This finding is in accordance with the ¹¹C-DPA-713 findings and our previous report⁴⁴.

At two months post-irradiation, the ¹¹C-CPPC radiotracer was still detected in the irradiated brain hemispheres compared to the corresponding contralateral non-irradiated brain hemispheres of the mice between 20 and 30 min after the bolus intravenous administration of the radiotracer (Figs. 6 and 7, Fig. S3, Table S5, S6, and S7). Quantification of the time-radioactivity kinetic curve ratios of the irradiated brain hemispheres normalized to the corresponding non-irradiated contralateral brain hemispheres of the mice after the bolus intravenous administration of ¹¹C-CPPC showed a statistically significant increase in the radioactivity uptake ratios between 20 and 30 min after the bolus intravenous administration of the radiotracer at two months post-irradiation, compared to the ratios detected in the non-irradiated control mice (P < 0.0001, n = 3) between 20 and 30 min after the bolus intravenous administration of the radiotracer. An average radioactivity uptake ratio of 1.14 ± 0.11 was detected between 20 and 30 min after the bolus intravenous administration of ¹¹C-CPPC at two months post-irradiation (Figs. 6 and 7, Fig. S3, Table S5, S6 and S7). Although this radioactivity uptake ratio was higher than that detected in the non-irradiated mice, it was lower than the radioactivity uptake ratio detected at one month post-irradiation (Figs. 6 and 7, Fig. S3, Table S4 and S5). This indicated a decrease in the neuroinflammation signal detected in the irradiated brain hemispheres of the mice at two months post-irradiation compared to that detected at one month post-irradiation in the preclinical mouse model. This finding is also in accordance with the ¹¹C-DPA-713 findings and our previous report⁴⁴.

At six months and nine months post-irradiation, no statistically significant differences were detected in the spatial biodistribution of the ¹¹C-CPPC radiotracer in the irradiated brain hemispheres compared to the corresponding contralateral non-irradiated brain hemispheres of the mice between 20 and 30 min after the bolus intravenous administration of ¹¹C-CPPC (Figs. 6 and 7, Fig. S6, Table S5, S6, and S7). Average radioactivity uptake ratios of 1.06 ± 0.06 and 1.03 ± 0.04 were detected between 20 and 30 min after the bolus intravenous administration of ¹¹C-CPPC at six months and nine months post-irradiation, respectively (Figs. 6 and 7, Fig. S2, Table S5, S6, and S6). These respective radioactivity uptake ratios were similar to the radioactivity uptake ratio detected in the non-irradiated control mice. This indicated a further reduction in the neuroinflammation signal detected in the irradiated brain hemispheres of the mice at six months and at nine months post-irradiation, compared to that detected at two months post-irradiation. This finding is also in accordance with the ¹¹C-DPA-713 findings and our previous report⁴⁴.

Next, to evaluate the specificity of the ¹¹C-CPPC radiotracer to detect neuroinflammation in RIBI, a blocking study was carried out in the irradiated mice at two months post-irradiation (Figs. 8 and 9, and Table S8). Quantification of the time-radioactivity kinetic curve ratios of the irradiated brain hemispheres normalized to the corresponding contralateral non-irradiated brain hemispheres of the mice that received the blocking agent (unlabeled CPPC) before the bolus intravenous administration of ¹¹C-CPPC at two months post-irradiation, showed a statistically significant decrease in the radioactivity uptake ratios between 20 and 30 min after the bolus intravenous administration of ¹¹C-CPPC, compared to the radioactivity uptake ratios of the mice that did not receive the blocking agent before the bolus intravenous administration of ¹¹C-CPPC (P = 0.0001, n = 3). An average radioactivity uptake ratio of 1.02 ± 0.03 was detected between 20 and 30 min after the bolus intravenous administration of ¹¹C-CPPC in the mice that received the blocking agent compared to 1.14 ± 0.11 in the mice that did not receive the blocking agent at two months post-irradiation. The radioactivity uptake ratio from the blocking study was similar to that detected in the non-irradiated control mice. This confirmed the specificity of the ¹¹C-CPPC radiotracer to detect neuroinflammation via CSF1R detection in the preclinical mouse model of RIBI.

Fig. 8.

¹¹C-CPPC PET/MRI blocking study. PET/MRIs of the radioactivity biodistribution of the ¹¹C-CPPC radiotracer in the brain before, 2, 17.5, 20, 22.5, 25, 27.5 and 30 min after the bolus intravenous administration of ¹¹C-CPPC: (a) Before irradiation (non-irradiated control mice); (b) two months post-irradiation; and (c) two months post-irradiation, 15 min after the bolus intravenous administration of 0.025 mg (1 mg/Kg) of the blocking agent (unlabeled CPPC). Quantification of the time-radioactivity kinetic curve uptake ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the mice, after the bolus intravenous administration of ¹¹C-CPPC: (d) Before irradiation (non-irradiated control mice); (e) two months post-irradiation versus before irradiation (P = 0.0001, n = 3); (f) two months post-irradiation (with the blocker) versus two months post-irradiation (without the blocker) versus before irradiation (P = 0.0001, n = 3). A decrease was detected in the ¹¹C-CPPC uptake ratios of the irradiated mice that received the blocking agent compared to the irradiated mice that did not receive the blocking agent at two months post-irradiation. Furthermore, the ¹¹C-CPPC uptake ratios in the irradiated mice that received the blocking agent at two months post-irradiation were similar to those detected in the non-irradiated control mice. This confirmed the specificity of the ¹¹C-CPPC radiotracer to detect CSF1R in the preclinical mouse model of RIBI.

Fig. 9.

Summary of the ¹¹C-CPPC PET/MRI blocking study. PET/MRIs of the radioactivity biodistribution of the ¹¹C-CPPC radiotracer in the brain before, 25, and 27.5 min after the bolus intravenous administration of ¹¹C-CPPC in irradiated mice at two month post-irradiation; versus in irradiated mice at two months post-irradiation, 15 min after the bolus intravenous administration of 0.025 mg (1 mg/Kg) of the blocking agent (unlabeled CPPC); versus in non-irradiated control mice, respectively: (a) Coronal view; and (b) Axial view. (c) Quantification of the time-radioactivity kinetic curve uptake ratios of the irradiated (right) brain hemispheres normalized to the corresponding contralateral (left) brain hemispheres of the mice, after the bolus intravenous administration of ¹¹C-CPPC in irradiated mice at two months post-irradiation (P = 0.0001, n = 3) versus in irradiated mice at two months post-irradiation, 15 min after the bolus intravenous administration of 0.025 mg (1 mg/Kg) of the blocking agent (unlabeled CPPC), versus in non-irradiated control mice. A decrease was detected in the ¹¹C-CPPC uptake ratios of the irradiated mice that received the blocking agent compared to the irradiated mice that did not receive the blocking agent at two months post-irradiation. Additionally, the ¹¹C-CPPC uptake ratios in the irradiated mice that received the blocking agent at two months post-irradiation were similar to those detected in the non-irradiated control mice. This confirmed the specificity of the ¹¹C-CPPC radiotracer to detect CSF1R in the preclinical mouse model of RIBI.

Detection of longitudinal changes in the blood-brain barrier permeability after irradiation using contrast-enhanced T₁-weighted MRI. To evaluate how the permeability of the blood-brain barrier (BBB) changed after irradiation, the kinetics of T₁W MRI contrast enhancement of the irradiated mouse brain was determined over a 60-minute post-contrast administration period at two different time points: one month and two months post-irradiation. A similar contrast enhancement kinetic trend was detected at both time points (Fig. 10 and Fig. S4). Quantification of the contrast-enhanced T₁W MRI signal intensity ratio of the irradiated brain hemispheres compared to the contralateral non-irradiated brain hemispheres of the respective mice showed higher contrast enhancement at one month post-irradiation, compared to two months post-irradiation (P = 0.005, n = 3). More importantly, a plateaued signal was detected between 15- and 60-minutes post-contrast administration (Fig. 10 and Fig. S4). This indicated that contrast-enhanced T₁W MRIs acquired between 15 and 30 min post-contrast administration could be used to reliably to evaluate changes in the permeability of the BBB at different time points after irradiation.

Fig. 10.

The kinetics of contrast-enhanced T₁W MRI. Color-coded contrast-enhanced T₁W MRIs showing the kinetics of contrast enhancement at one month (a) and two months (b) post-irradiation in mice irradiated with 80 Gy of an image-guided focal X-ray beam at age eight-weeks-old. Color-coded z-projection images of six 0.5 mm coronal slices and six 0.5 mm axial slices were used to generate each respective image. A contrast enhancement equilibrium was detected between 15 to 60 min after the administration of the contrast agent at both time points. Quantification of the kinetics of T₁W MRI contrast enhancement one month (c) and two months (d) post-irradiation in mice irradiated with 80 Gy of an image-guided focal X-ray beam at age eight-weeks-old. A contrast enhancement equilibrium was detected between 15 to 60 min after the administration of the contrast agent at both time points. Overall, a similar kinetic trend was observed at both time points, even though the signal at one month post-irradiation was statistically higher than that at two months post-irradiation (P < 0.005, n = 3).

Next, using contrast-enhanced T₁W MRI, longitudinal changes in the BBB permeability were evaluated over a nine-month post-irradiation period in mice irradiated at eight-weeks of age compared to non-irradiated control mice. As expected, no contrast-enhanced T₁W MRI signal was detected in the control non-irradiated mice (Fig. 11, Fig. S5, and Tables S9 - S12). A maximum contrast-enhanced T₁W MRI signal was detected at one month post-irradiation (Fig. 11, Fig. S5,Tables S9 - S12) in the irradiated mice. Quantification of the contrast-enhanced T₁W MRI signal intensity ratio of the irradiated brain hemispheres compared to the contralateral non-irradiated brain hemispheres of the respective mice revealed an 18.95 ± 1.75-fold increase in the irradiated mice at one month post-irradiation compared to the control non-irradiated mice (P < 0.0001, n = 3). This contrast-enhanced T₁W MRI signal intensity ratio then gradually decreased to a 13.50 ± 1.35-fold increase in the irradiated mice at two months post-irradiation compared to the control non-irradiated mice (P < 0.0001, n = 3). This decrease in the BBB permeability continued to a 9.53 ± 0.95-fold increase in the irradiated mice at six months post-irradiation compared to the control non-irradiated mice (P = 0.0001, n = 3). A further decrease in the BBB permeability to an 8.42 ± 2.54-fold increase in the irradiated mice at nine months post-irradiation compared to the control non-irradiated mice (P = 0.0073, n = 3) was detected in the study.

Fig. 11.

Detection of longitudinal changes in blood-brain barrier (BBB) permeability after stereotactic irradiation using contrast-enhanced T₁W MRI. In vivo contrast-enhanced T₁W MRIs: (a) Coronal view of Z-projections of six slices (color-coded); (b) Axial view of Z-projections of six slices (color-coded). The images show that the degree of BBB permeability changed over the nine-month post-irradiation study period. A maximum BBB permeability was detected at one month post-irradiation during the study. (c) Quantification of the contrast-enhanced T₁W MRI signal in the irradiated brain hemisphere normalized to the non-irradiated brain hemisphere at 0, 15 and 30 min post-contrast administration in non-irradiated control mice and in mice irradiated at eight-weeks-old and monitored over the nine-month post-irradiation study period (P < 0.05, n = 3). (d) Quantification of the changes in the contrast-enhanced T₁W MRI signal in the irradiated brain hemisphere normalized to the non-irradiated brain hemisphere over a period of nine-months post-irradiation (P < 0.05, n = 3). This confirmed that the degree of BBB permeability changed statistically significantly (P < 0.05, n = 3) over the nine-month post-irradiation period.

Detection of neuroinflammation in RIBI using TSPO, IBA1, and CD68 immunohistochemistry. Neuroinflammation was detected in the Group 4 irradiated mouse brain tissue sections using cluster of differentiation 68 (CD68), ionized calcium binding adaptor molecule 1 (IBA1), and 18-kDa translocator protein (TSPO) immunohistochemistry. Whereas CD68 was used to detect infiltrating macrophages, IBA1 was used to detect microglial activation. Neuroinflammation was further validated with TSPO immunohistochemistry.

Fluorescence microscopy images of brain tissue sections showed maximum CD68 and IBA1 fluorescence signals at one month post-irradiation (Fig. 12 and Tables S13 - S16). This was validated with TSPO immunohistochemistry, which also showed a similar maximum fluorescence signals at one month post-irradiation (Fig. 12, Table S17 and S18).

Fig. 12.

Detection of longitudinal changes in neuroinflammation after stereotactic irradiation using CD68, IBA1, and TSPO immunohistochemistry. Detection of longitudinal changes in neuroinflammation with CD68 (green) and IBA1 (red) immunohistochemistry in mice irradiated at eight-weeks-old and monitored for nine months post-irradiation (a) Fluorescence images of whole brain tissue sections showing the irradiation path (2 mm scale bar). (b) Fluorescence images of representative brain regions along the irradiation path (100 μm scale bar). Maximum neuroinflammation was detected at one month post-irradiation. Detection of longitudinal changes in neuroinflammation using TSPO (green) and IBA1 (red) immunohistochemistry in mice irradiated at eight-weeks-old and monitored for nine months post-irradiation (c) Fluorescence images of whole brain tissue sections showing the irradiation path (2 mm scale bar). (d) Fluorescence images of representative brain regions along the irradiation path (100 μm scale bar). Maximum neuroinflammation was detected at one month post-irradiation. Quantification of the respective fluorescence signals in the irradiated brain hemispheres normalized to the non-irradiated brain hemispheres at different time points post-irradiation. e) Quantification of the respective fluorescence signals in tissue sections stained using CD68 (green) and IBA1 (red) immunohistochemistry (P < 0.05, n = 3). f) Quantification of the respective fluorescence signals in tissue sections stained using TSPO (green) and IBA1 (red) immunohistochemistry (P < 0.05, n = 3). g) Comparison of the respective CD68, IBA1, and TSPO fluorescence signals in the tissue sections (P < 0.05, n = 3). Maximum neuroinflammation was detected at one month post-irradiation and then decreased over the following months of the study period.

Quantification of the respective fluorescence signal intensity ratios of the irradiated brain hemispheres normalized to the contralateral non-irradiated brain hemispheres of the respective mice revealed the following respective ratio increases in the irradiated mice at one month post-irradiation compared to the control non-irradiated mice: a 24.48 ± 4.40-fold increase for CD68 (P = 0.0007, n = 3); a 24.14 ± 6.72-fold increase for IBA1 (P = 0.0039, n = 3); and a 96.71 ± 1.76-fold increase for TSPO (P < 0.0001, n = 3). These fluorescence signal intensity ratios then gradually decreased to the following respective ratio increases in the irradiated mice at six months post-irradiation compared to the control non-irradiated mice: a 6.12 ± 2.12-fold increase for CD68 (P = 0.0128, n = 3); a 5.11 ± 3.84-fold increase for IBA1; and a 16.41 ± 5.78-fold increase for TSPO (P < 0.0010, n = 3). These ratios further decreased to a 2.08 ± 1.40-fold increase for CD68; and a 2.97 ± 0.81-fold increase for IBA1 (P = 0.0125, n = 3) in the irradiated mice at nine months post-irradiation compared to the control non-irradiated mice.

Correlation of the PET and immunohistochemical biomarkers of neuroinflammation to the contrast-enhanced T₁W MRI biomarker of BBB permeability in RIBI. Overall, the neuroinflammation trend detected using PET (¹¹C-DPA-713 and ¹¹C-CPPC radiotracers) and immunohistochemistry (CD68, IBA1 and TSPO) coincided with the BBB permeability trend detected using contrast enhanced T₁W MRI (Fig. 13 and Table S19). Correlating the respective ¹¹C-DPA-713 PET signals, ¹¹C-CPPC PET signals, and contrast-enhanced T₁W MRI signals to the TSPO fluorescence signals detected in the irradiated mice gave correlation constants close to one (Fig. 13). The following correlation constants (R) were obtained: R = 0.9998 for ¹¹C-CPPC and TSPO; R = 0.9958 for ¹¹C-DPA-713 and TSPO; and R = 0.9932 for contrast-enhanced T₁W MRI and TSPO. These correlation constants (R) close to one confirmed a strong correlation between the respective ¹¹C-DPA-713 PET signal; ¹¹C-CPPC PET signal, and contrast-enhanced T₁W MRI signal and the TSPO fluorescence signal. However, the contrast-enhanced T₁W MRI / TSPO correlation constant was slightly lower than that of ¹¹C-CPPC PET / TSPO and ¹¹C-DPA-713 PET / TSPO, respectively. This suggested that contrast-enhanced T₁W MRI is slightly less specific to TSPO detection compared to ¹¹C-CPPC PET and¹¹C-DPA-713 PET, as was expected. Nonetheless, this suggested that contrast-enhanced T₁W MRI has a comparable neuroinflammation detection capability compared to ¹¹C-CPPC PET and ¹¹C-DPA-713 PET.

Fig. 13.

Correlation of the PET, immunohistochemical and contrast-enhanced T₁W MRI biomarkers of neuroinflammation in RIBI. (a) Immunohistochemistry (CD68 and IBA1), ¹¹C-DPA-713 PET, ¹¹C-CPPC PET and contrast-enhanced T₁W MRI images over the nine-month post-irradiation study period. Quantification of the respective signals over the nine-month post-irradiation study period. (b) Quantification of the CD68 and IBA1 immunohistochemistry, ¹¹C-DPA-713 PET, and ¹¹C-CPPC PET signals (P < 0.05, n = 3). (c) Quantification of the CD68 and IBA1 immunohistochemistry and contrast-enhanced T₁W MRI signals (P < 0.05, n = 3). (d) Quantification of the TSPO immunohistochemistry, ¹¹C-DPA-713 PET, and ¹¹C-CPPC PET signals (P < 0.05, n = 3). (e) Quantification of the TSPO immunohistochemistry and contrast-enhanced T₁W MRI signals (P < 0.05, n = 3). (f) Quantification of the respective ¹¹C-DPA-713 PET, ¹¹C-CPPC PET and contrast-enhanced T₁W MRI signals (P < 0.05, n = 3). (g) Quantification of the respective immunohistochemistry (CD68, IBA1 and TSPO) signals (P < 0.05, n = 3). (h) Correlation of the respective ¹¹C-DPA-713 PET signals, ¹¹C-CPPC PET signals, and contrast-enhanced T₁W MRI signals to the TSPO fluorescence signals detected in the irradiated mice over the nine-month post-irradiation study period. (i) Correlation of the respective ¹¹C-DPA-713 PET signals, ¹¹C-CPPC PET signals, and contrast-enhanced T₁W MRI signals to the IBA1 fluorescence signals detected in the irradiated mice over the nine-month post-irradiation study period. (j) Correlation of the respective ¹¹C-DPA-713 PET signals, ¹¹C-CPPC PET signals, and contrast-enhanced T₁W MRI signals to the CD68 fluorescence signals detected in the irradiated mice over the nine-month post-irradiation study period.

Furthermore, correlating the respective ¹¹C-DPA-713 PET; ¹¹C-CPPC PET, and contrast-enhanced T₁W MRI signals to the IBA1 fluorescence signals detected in the irradiated mice also gave correlation constants close to one. The following correlation constants (R) were obtained: R = 0.9988 for ¹¹C-CPPC and IBA1; R = 0.9931 for ¹¹C-DPA-713 and IBA1; and R = 0.9958 for contrast-enhanced T₁W MRI and IBA1. Although the contrast-enhanced T₁W MRI / IBA1 correlation constant was slightly lower than that of ¹¹C-CPPC PET / IBA1, it was unexpectedly slightly higher than that of ¹¹C-DPA-713 PET / IBA1. This suggested that contrast-enhanced T₁W MRI has a comparable neuroinflammation detection capability compared to ¹¹C-CPPC PET and ¹¹C-DPA-713 PET.

Finally, correlating the respective ¹¹C-DPA-713 PET; ¹¹C-CPPC PET, and contrast-enhanced T₁W MRI signals to the CD68 fluorescence signals detected in the irradiated mice also gave correlation constants close to one. However, the correlation constants obtained with CD68 were lower than the correlation constants calculated for TSPO and IBA1, respectively. The following correlation constants (R) were obtained: R = 0.9672 for ¹¹C-CPPC and CD68; R = 0.9608 for ¹¹C-DPA-713 and CD68; and R = 0.9891 for contrast-enhanced T₁W MRI and CD68.

Interestingly, the contrast-enhanced T₁W MRI / CD68 correlation constant was higher than that of ¹¹C-CPPC PET / CD68 and ¹¹C-DPA-713 PET / CD68. This is understandable especially since CD68 could be more highly expressed on peripheral macrophages that infiltrate the brain following BBB disruption than on resident activated microglia cells (especially depending on their activation state). This further suggested that contrast-enhanced T₁W MRI can be used to detect neuroinflammation arising from a variety of neuroimmune cells. Collectively, these findings suggest that contrast-enhanced T₁W MRI could be used as a predictive biomarker of neuroinflammation in RIBI.

Discussion

In this comparative study, we investigated if the indirect detection of neuroinflammation via the detection of BBB permeability using contrast-enhanced T₁-weighted magnetic resonance imaging (T₁W MRI), was as effective as the direct detection of neuroinflammation using ¹¹C-DPA-713 PET and ¹¹C-CPPC PET, respectively, in a preclinical mouse model of RIBI.

Overall, our results showed that both ¹¹C-DPA-713 PET and ¹¹C-CPPC PET detected changes in neuroinflammation in the preclinical model of RIBI over the nine-month post-irradiation study period. In summary, maximum ¹¹C-DPA-713 PET and ¹¹C-CPPC PET signals were detected in the irradiated brain hemispheres of the mice at one month post-irradiation (Table S19). These maximum ¹¹C-DPA-713 PET and ¹¹C-CPPC PET signals at one month post-irradiation were indicative of maximum neuroinflammation in the irradiated brain hemispheres of the mice. These ¹¹C-DPA-713 PET and ¹¹C-CPPC PET signals then decreased at two months post-irradiation and continued to decrease through six and nine months post-irradiation in the irradiated mice (Table S19). Although both the ¹¹C-DPA-713 PET and ¹¹C-CPPC PET radiotracers showed similar neuroinflammation detection trends, the time-radioactivity kinetic curve uptake ratios of the ¹¹C-CPPC radiotracer were lower than those of the ¹¹C-DPA-713 radiotracer. This difference could be because while the ¹¹C-CPPC radiotracer detects the CSFR1 receptor, which is expressed only on activated microglial cells, the ¹¹C-DPA-713 radiotracer detects TSPO, which is expressed on a variety of cells involved in the neuroinflammatory cascade including activated microglia cells, astrocytes, and infiltrative peripheral macrophages^24,26.

The ¹¹C-DPA-713 PET and ¹¹C-CPPC PET direct detection of neuroinflammation changes over the nine-month post-irradiation study period in the preclinical RIBI mouse model was validated with TSPO, IBA1, and CD68 immunohistochemistry. TSPO, IBA1 and CD68 immunohistochemistry detected similar changes in neuroinflammation to those detected with ¹¹C-DPA-713 PET and ¹¹C-CPPC PET, respectively (Table S19). Interestingly, the TSPO fluorescence signal detected at each point of the study was much higher than the IBA1 and the CD68 fluorescence signals detected. This difference in signal intensity could be because TSPO is expressed on a variety of cells involved in the neuroinflammatory cascade including activated microglia cells, activated astrocytes, and infiltrative peripheral macrophages; while IBA1 and CD68 are only expressed on activated microglia cells and infiltrative peripheral macrophages but not on activated astrocytes^24,26,59,60.

Overall, the neuroinflammation trend detected using PET (¹¹C-DPA-713 and ¹¹C-CPPC tracers) and immunohistochemistry (CD68, IBA1, and TSPO staining) coincided with the BBB permeability trend detected using contrast enhanced T₁W MRI (Table S19). Correlating the contrast-enhanced T₁W MRI signal to the respective TSPO, IBA1, and CD68 fluorescence signals detected in the irradiated mice gave correlation constants close to one. This strong correlation detected between BBB permeability and neuroinflammation suggests that disruption of the BBB after irradiation leads to neuroinflammation by the activation of resident microglia cells and astrocytes and enables the recruitment of infiltrative peripheral macrophages. Collectively, these findings suggest the possibility of using contrast-enhanced T₁W MRI to detect BBB permeability as a predictive biomarker of neuroinflammation in RIBI.

However, there are a few limitations associated with using contrast-enhanced T₁W MRI to detect BBB permeability as a predictive biomarker of neuroinflammation in RIBI. These include the following: (1) Since this method is not specific to neuroinflammation, but rather to BBB permeability changes, its use could be challenging especially in tumor-bearing models where the blood-tumor barrier is sometimes permeable. (2) It might also be difficult to distinguish neuroinflammation detected using this method from tumor recurrence. Thus, caution should be exercised when using this method in tumor-bearing models of RIBI. We envision that this method would best serve as an excellent noninvasive and simple screening tool to evaluate the efficacy of potential RIBI prophylactic and therapeutic agents in preclinical non-tumor bearing models of RIBI⁶¹.

Our future studies include evaluating the feasibility of using this method to distinguish RIBI-associated neuroinflammation from tumor recurrence in a tumor-bearing mouse model of RIBI.

Conclusion

In this study, we showed that the indirect detection of neuroinflammation via the detection of blood-brain barrier permeability, using gadoteridol (ProHance^®)-based contrast-enhanced T₁W MRI, was as effective as the direct detection of neuroinflammation using ¹¹C-DPA PET and ¹¹C-CPPC PET, respectively. Ultimately, this contrast-enhanced T₁W MRI method could serve as a predictive biomarker of neuroinflammation and provide a simple yet effective means of noninvasively and indirectly probing the status of neuroinflammation during RIBI progression in pediatric brain tumor patients and survivors. Additionally, this method could provide a simple indirect preclinical and clinically translatable means of longitudinally monitoring the therapeutic efficacy of RIBI prophylactic and therapeutic agents that are currently in development.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

SM coordinated the study and performed all the PET dosing experiments; SKY acquired and processed the PET/MRI data and images; MT acquired and processed the contrast-enhance T1W MRI and immunohistochemistry data and images; EV irradiated the mice; SL prepared the tissue sections for immunohistochemistry; KL acquired the immunohistochemistry fluorescence images; AB established the MRI pulse sequences; EJN conceptualized and designed the study; acquired funding for the study; analyzed the data; prepared the figures and wrote the manuscript. All authors reviewed and edited the manuscript.

Funding

Research reported in this publication was supported by the National Cancer Institute (NCI) and the Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD) of the National Institutes of Health under Awards Number R01CA262887 and R21HD097357, respectively. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Data availability

All the datasets and data processing methods used in this study are available for sharing upon the receipt of reasonable requests made to the corresponding author.

Declarations

Competing Interests

The authors declare no competing interests.