Quantitative Analysis of S1PR1 Expression in the Postmortem Multiple Sclerosis Central Nervous System

Abstract

Multiple sclerosis (MS) is an immune-mediated disease that is characterized by demyelination and inflammation in the central nervous system (CNS). Previous studies demonstrated that sphingosine-1-phosphate receptor (S1PR) modulators effectively inhibit S1PR1 in immune cell trafficking and reduce entry of pathogenic cells into the CNS. Studies have also implicated a nonimmune, inflammatory role of S1PR1 within the CNS in MS. In this study, we explored the expression of S1PR1 in the development and progression of demyelinating pathology of MS by quantitative assessment of S1PR1 expression using our S1PR1-specific radioligand, [³H]CS1P1, in the postmortem human CNS tissues including cortex, cerebellum, and spinal cord of MS cases and age- and sex-matched healthy cases. Immunohistochemistry with whole slide scanning for S1PR1 and various myelin proteins was also performed. Autoradiographic analysis using [³H]CS1P1 showed that the expression of S1PR1 was statistically significantly elevated in lesions compared to nonlesion regions in the MS cases, as well as normal healthy controls. The uptake of [³H]CS1P1 in the gray matter and nonlesion white matter did not significantly differ between healthy and MS CNS tissues. Saturation autoradiography analysis showed an increased binding affinity (K_d) of [³H]CS1P1 to S1PR1 in both gray matter and white matter of MS brains compared to healthy brains. Our blocking study using NIBR-0213, a S1PR1 antagonist, indicated [³H]CS1P1 is highly specific to S1PR1. Our findings demonstrated the activation of S1PR1 and an increased uptake of [³H]CS1P1 in the lesions of MS CNS. In summary, our quantitative autoradiography analysis using [³H]CS1P1 on human postmortem tissues shows the feasibility of novel imaging strategies for MS by targeting S1PR1.

Graphical Abstract

INTRODUCTION

Multiple sclerosis (MS) is the most common immune-mediated disease of the central nervous system (CNS) and affects over two million people worldwide. It is characterized by inflammation, demyelination, gliosis, and neurodegeneration in the CNS, and the pathological hallmark of MS is the focal areas of demyelination in the white matter of the optic nerves, brain, and spinal cord, commonly recognized as MS lesions.^1,2 Chronic dysregulations of the peripheral immune system, blood–brain barrier permeability, and intrinsic CNS immune system contribute to the pathogenesis of MS.³ While there is no cure for MS, several disease-modifying therapies have been developed in the last few decades, including fingolimod, a nonselective sphingosine 1-phosphate (S1P) receptor modulator.⁴ Additional medications in this drug class that are more selective to specific subtypes of S1P receptors (S1PRs) such as ponesimod,⁵ siponimod,⁶ and ozanimod are in clinical use.⁷ These S1PR modulators bind to S1PR1 in lymphocytes and cause S1PR1 internalization. The internalization of S1PR1 subsequently prevents lymphocyte egress from lymphoid tissues and therefore reduces the migration of lymphocytes to the CNS.⁸

S1PR1 is the most abundant S1P receptor and is ubiquitously expressed throughout all tissues including CNS. It is a G-protein-coupled receptor primarily coupled through G_i/o and regulates a variety of physiological and cellular processes such as the egress of T cells from lymph nodes, vasculature formation, and regulation of endothelial barriers.^9,10 Thus, dysregulation of S1PR1 signaling is associated with diverse diseases such as infection,^11,12 atherosclerosis,¹³ and cancer.¹⁴ In the CNS, S1PR1 is widely expressed in different types of cells and it mediates a broad range of cellular functions, including neurogenesis, proliferation and migration of astrocytes, neurotransmission in neurons, and proinflammatory cytokine production in the microglia.^15,16 It plays important roles in many physiological and pathophysiological processes in the CNS. For instance, previous studies show activation of S1PR1 can enhance hippocampal neurogenesis;¹⁷ hypothalamic S1PR1 is important for the maintenance of energy homeostasis;¹⁸ and S1PR1 activation in astrocytes contributes to neuropathic pain.¹⁹ Additionally, the expression of S1PR1 is directly upregulated in MS²⁰ and Type 2 schizophrenia,²¹ and may play a role in other neurodegenerative and neuroinflammatory disorders. Since the introduction of fingolimod for the treatment of MS, S1PR1 has provoked broad interest as a pathological marker and therapeutic target for MS as well as many other neurological disorders. In addition to the role of immune cell trafficking, new studies have demonstrated that the clinical effectiveness of S1PR1 modulators involves neuroprotective activities in the brain^8,22 and/or direct suppression of astrocyte function and CNS innate immune responses in MS and neuropathic pain.^19,23

While numerous studies have investigated the clinical effectiveness and biological mechanisms of S1PR1 in CNS disorders, studies of S1PR1 as a pathological marker for MS and other CNS disorders, with noninvasive methodologies such as positron emission tomography (PET) have not been reported previously. To explore the potential of targeting S1PR1 for diagnosis and validating therapeutic efficacy, we reported a series of S1PR1 radioligands and performed initial preclinical in vitro and in vivo evaluations.^24–27 Our leading radioligand, [¹¹C]CS1P1 (previously named [¹¹C]TZ3321, Figure 1) has a high binding affinity toward S1PR1 with an IC₅₀ value of 2.13 nM and more than 1000-fold selectivity over other S1PRs.²⁸ We have also performed a systematic characterization of [¹¹C]CS1P1 and its tritiated version [³H]CS1P1 (Figure 1) in the human tissues and animal models including rodents and nonhuman primates. We demonstrated that [¹¹C]CS1P1 is a promising radioligand for PET imaging quantification of S1PR1 expression in the CNS.^24,25,29 Furthermore, with FDA approval for an exploratory investigational new drug (e-IND) of [¹¹C]CS1P1 for human use, we completed a whole-body PET imaging in healthy volunteers for human dosimetry estimates³⁰ and proof of concept studies in patients with MS, demonstrating that [¹¹C]CS1P1 binding was increased in MS lesions.³¹ Both our preclinical and first-in-human evaluation suggests that PET with S1PR1-specific radioligand [¹¹C]CS1P1 is a reliable methodology to assess neuroinflammation in the CNS. Our proof of concept PET study in MS patients showed an elevation of S1PR1 in MS lesions, consistent with previous studies,²⁰ but directly quantitative analysis of S1PR1 expression in MS lesions has not been explored. Quantitative autoradiography analysis enables studies of the functionality of a receptor protein by measuring the distribution of its radioligand in tissues, and it has the advantage of visualizing both spatial anatomy and functionality of targeted protein.^32–34 Furthermore, quantitative autoradiography analysis such as using a saturation autoradiography study permits the calculation of the dissociation constant (K_d) and the number of binding sites (B_max) in the tissues for a radioligand. This provides information on both the affinity of a radioligand, [³H]CS1P1 for its binding sites, and the total density (B_max) of targeted proteins in the tissues.^35,36 To further understand the expression of S1PR1 in the CNS of MS patients and the distribution of our leading radioligand CS1P1 in MS, we performed a quantitative autoradiography study in postmortem human CNS tissues using our leading S1PR1 ligand [³H]CS1P1. Our data showed a statistically significant increase in [³H]CS1P1 level in the lesion of MS CNS tissues compared to that in the healthy controls or nonlesion MS CNS tissues. We also observed a statistically higher binding affinity (K_d) of [³H]CS1P1 in the MS brain compared to that in the normal brain and a statistically lower B_max of [³H]CS1P1 in the white matter compared to that in gray matter of both the normal control and the MS brain. Our studies confirm that S1PR1 is involved in MS pathology and show that our lead S1PR1-specific ligand, CS1P1, can detect S1PR1 elevation in MS lesions.

Figure 1.

Characterization of S1PR1-specific radioligand [³H]CS1P1 and immunohistochemistry studies in the human cortical tissues. (A) Chemical structures of S1PR1 endogenous ligand S1P, our leading S1PR1 radioligand [¹¹C]CS1P1 and [³H]CS1P1, and known S1PR1 antagonist NIBR0213; (B) distribution of [³H]CS1P1 in the cortical region of the human brain was detected using autoradiography analysis and the corresponding compared to the H&E staining, distribution of neuron (NeuN), astrocyte (GFAP), and microglia (IBA1), and S1PR1. [³H]CS1P1 was mainly distributed in gray matter compared to white matter and matched well with anti-S1PR1 immunostaining and anti-GFAP staining. The specificity of [³H]CS1P1 was confirmed by blocking it with a well-validated S1PR1 antagonist, NIBR0213. Scale bar = 5000 μm.

RESULTS

Characterization and Quantitative Assessment of [³H]CS1P1 Distribution in the Human Brain.

S1PR1 has been intensively characterized in neurophysiological and different neuropathological states. Using our well-validated S1PR1-specific radioligand, [³H]CS1P1, we examined the expression of S1PR1 in fresh frozen human cortical brain tissues (Figure 1). The distribution of [³H]CS1P1 matched well with the immunohistochemistry of S1PR1 and was blocked with the S1PR1-specific antagonist, NIBR-0213, indicating that [³H]CS1P1 is highly specific to S1PR1. In general, [³H]CS1P1 was mainly localized in the gray matter of the cerebral cortex and showed diffusely fine granular staining along with S1PR1. On the other hand, although the radioactive signal of [³H]CS1P1 was much stronger in the gray matter than that in the white matter, a faint distribution of [³H]CS1P1 was also observed in the white matter regions, and the overall distribution of [³H]CS1P1 was more comparable to GFAP staining rather than NeuN and IBA1 staining. This was consistent with the fact that S1PR1 is ubiquitously expressed in various types of brain cells at relatively low-to-mild levels but highly expressed by astrocytes.^37,38

[³H]CS1P1 in Normal and MS Cortical Brain Tissues and MS Lesions.

To quantitatively compare the level of S1PR1 in the brain tissues of normal controls and MS patients, we next performed autoradiography analysis in the cortical brain tissues of the normal controls and MS subjects (Figures 2 and 3). In general, the radioactive signal of [³H]CS1P1 and expression of S1PR1 were similar in the cortical tissues of normal healthy controls and the nonlesion cortical tissues of MS patients. MS lesions were identified by hematoxylin and eosin (H&E), myelin basic protein (MBP), and myelin proteolipid protein (PLP) staining (Figure 2). Notably, a statistically significant increase in radioactivity of [³H]CS1P1 and increased expression of S1PR1 were observed in the lesion regions of MS compared to nonlesion brain regions of the MS patients and normal brain tissues of normal health controls. Interestingly, the increased radioactivity of [³H]CS1P1 and upregulation of S1PR1 were identified in the glial scar (Figure 2, MS example 1) of late stage lesions as well as in the center of active lesions (Figure 2, MS example 2). Furthermore, increased radioactivity of [³H]CS1P1 and upregulation of S1PR1 was observed in the lesions tissue from a patient with secondary progressive multiple sclerosis (SPMS) and no significant change was observed in the tested tissue from a patient with relapsing-remitting multiple sclerosis (RRMS) (Figure 2).

Figure 2.

Representative images of [³H]CS1P1 and S1PR1 in normal and MS cortical tissues, including the lesion and nonlesion regions of cortex from RRMS, and SPMS brain. H&E staining, anti-S1PR1, MBP, and PLP staining, and [³H]CS1P1 autoradiography were performed. A statistically higher level of [³H]CS1P1 and S1PR1 was identified in the lesion of white matter that was stained by MBP and PLP staining (arrows). Scale bar = 5000 μm.

Figure 3.

Quantitative assessment of S1PR1 expression in normal and MS cortical tissues using S1PR1-specific ligand [³H]CS1P1 showed a statistically significant increase of [³H]CS1P1 uptake in the MS lesions. (A,B) Identical levels of [³H]CS1P1 were identical in the gray (A) and white (B) matter of cortical tissues between normal and MS patients with p values of 0.239 and 0.890, respectively; (C) one-way ANOVA showed a statistical difference of [³H]CS1P1 uptake among groups in cortex (F(2,20) = 34.58, P < 0.0001), Tukey’s multiple comparison test following ANOVA showed no significant change in the white matter between normal and MS brain, whereas a statistically higher level of [³H]CS1P1 was observed in MS lesion compared to normal and MS nonlesion white matter with a p values <0.0001 for both.

For normal healthy controls, quantitative autoradiography showed the uptake of [³H]CS1P1 was 120.0 ± 9.20 fmol/mg in gray matter (Figure 3A), 1.8-fold higher than that in the white matter (67.87 ± 14.97 fmol/mg) with a p < 0.0001 using unpaired t-test (Figure 3B). Similarly, in the brain of MS, the uptake of [³H]CS1P1 was 1.8-fold higher in gray matter at 124.8 ± 5.52 fmol/mg than that in the nonlesion white matter regions at 68.83 ± 11.39 fmol/mg with a p < 0.0001 using unpaired t-test. The uptake of [³H]CS1P1 in the gray matter of normal healthy control and MS brain tissue was nearly identical (Figure 3A). Similarly, the uptake of [³H]CS1P1 in white mater of healthy control brain tissue was almost the same as nonlesion white matter of MS brain tissue (Figure 3B). As stated above, a statistically significant increase of [³H]CS1P1 uptake and S1PR1 expression was observed in the MS lesions in cortical tissues compared to nonlesion of MS brain and normal healthy control brain (Figures 2 and 4). The uptake of [³H]CS1P1 in MS lesions was 115.9 ± 10.21 fmol/ mg, ~1.7-fold higher than the uptake in the white matter of the normal healthy control brain (67.87 ± 14.97 fmol/mg) and the uptake in the white matter of nonlesion MS brain tissue (68.83 ± 11.39 fmol/mg). One-way ANOVA showed that the uptake was statistically different among the lesion and nonlesion white matter regions with a P value <0.0001 (F(2,20) = 34.58). Tukey’s multiple comparison test indicated [³H]CS1P1 was statistically higher in MS lesions compared to normal white matter (p < 0.0001) and nonlesional MS white matter (p < 0.0001), whereas no statistical difference was observed between normal white matter and nonlesional white matter in MS brain (p = 0.9876) (Figure 3C).

Figure 4.

Representative images of [³H]CS1P1 and S1PR1 in normal and MS cerebellum tissues, including the lesion and nonlesion regions of cerebellum from RRMS and SPMS brain. H&E staining, anti-S1PR1, MBP, and PLP staining, and [³H]CS1P1 autoradiography were performed. A statistically higher level of [³H]CS1P1 and S1PR1 was identified in the lesion of white matter that was identified by MBP and PLP staining (arrows). Scale bar = 5000 μm.

[³H]CS1P1 in Normal and MS Cerebellum Tissues and MS Lesions.

In the cerebellum, similar to the cortical regions, high uptake of [³H]CS1P1 and high expression of S1PR1 were observed in the gray matter compared to that in the white matter, and an increase of S1PR1 expression and [³H]CS1P1 uptake was also observed in MS lesions (Figure 4). While no significant change was observed in the tested tissue from a subject with RRMS, increased uptake of [³H]CS1P1 and upregulation of S1PR1 were observed in the lesions from SPMS tissue (Figure 4).

Quantitative autoradiography showed in cerebellar brain tissues of normal healthy controls, [³H]CS1P1 had an uptake of 133.9 ± 7.655 fmol/mg in the molecular layer, 120.7 ± 9.521 fmol/mg in the granule layer, 115.5 ± 7.969 fmol/mg nucleus dentate, and 82.71 ± 7.217 fmol/mg in the white matter (Figure 5). In contrast, in the MS brain cerebellum, [³H]CS1P1 had an uptake of 123.9 ± 10.87 fmol/mg in the molecular layer, 116.7 ± 12.22 fmol/mg in the granule layer, 93.11 ± 16.76 fmol/mg in the nucleus dentate, and 77.63 ± 5.938 fmol/mg in the white matter (Figure 5). Interestingly, compared to the normal healthy control brain, the molecular layer in the MS brain cerebellum had a statistically lower uptake of [³H]CS1P1 than in the normal healthy control brain (p = 0.0347) (Figure 5D), although the difference was small. In contrast, no statistical difference was identified between the normal healthy control and MS brain in the granule layer (p = 0.450), nucleus dentate (p = 0.069), and white matter (p = 0.129) of the cerebellum (Figure 5A–C).

Figure 5.

Quantitative assessment of S1PR1 expression in normal and MS cerebellum using the S1PR1 specific ligand [³H]CS1P1. (A–D) [³H]CS1P1 uptake in different regions of the cerebellum: [³H]CS1P1 was statistically lower in the molecular layer of the cerebellum from MS compared to the normal brain p values of 0.035 (A), whereas no significant difference was observed in the granule layer (B), nucleus dentate (C), and white matter (D) of the cerebellum from MS compared to the normal brain p values of 0.450, 0.069, and 0.129, respectively; (E) one-way ANOVA showed a statistical difference of [³H]CS1P1 uptake among groups in cerebellum (F(2,23) = 33.91, P < 0.0001); Tukey’s multiple comparison test following ANOVA showed no significant in the white matter between normal and MS brain, whereas a statistically higher level of [³H]CS1P1 was observed in MS lesion compared to normal and MS nonlesion white matter with a p values <0.0001 for both.

A clear increase in the [³H]CS1P1 uptake and the S1PR1 expression was detected in MS lesions within cortical tissues. Among the normal brain white matter, MS brain nonlesion white matter, and MS lesion white matter, the highest uptake of [³H]CS1P1 was observed in the white matter lesion of the MS brain at 105.8 ± 8.700 fmol/mg, followed by 82.71 ± 7.217 fmol/mg in the white matter of normal healthy brain, and 77.63 ± 5.938 fmol/mg in the nonlesion white matter of MS brain. One-way ANOVA showed the uptake of [³H]CS1P1 was statistically different among three groups with a P value <0.0001 (F(2,23) = 33.91). Tukey’s multiple comparison tests showed a statistical higher of [³H]CS1P1 in MS lesion compared to normal white matter (~1.3-fold, p < 0.0001) and nonlesion white matter regions of MS brain (~1.4-fold, p < 0.0001) and no difference between normal white matter and nonlesion white matter in MS brain (p = 0.3299) (Figure 5D).

[³H]cS1P1 in the Spinal Cord of Normal and MS Tissues.

In addition to cortical and cerebellar brain tissues, we also tested the expression of S1PR1 and uptake of [³H]CS1P1 in the cervical region of the spinal cord between normal healthy controls and MS cases from a separate study cohort. The distribution of S1PR1 and the uptake of [³H]CS1P1 in the normal healthy control spinal cord and MS spinal cord were similar to the observations in the brain but with slightly higher uptake in the white matter (Figure 6A). The uptake of [³H]CS1P1 was 119.0 ± 3.362 fmol/mg in the gray matter, ~17% higher than 102.3 ± 1.027 fmol/mg in the white matter of the normal healthy spinal cord with a p < 0.0001 using unpaired t-test. Similarly, in the MS spinal cord, the uptake of [³H]CS1P1 was 119.2 ± 2.786 fmol/mg in the gray matter, ~21% higher than 98.32 ± 13.85 fmol/mg in the white matter with a p = 0.0254 using the unpaired t-test. No statistical difference was observed in the gray matter of spinal cord between normal healthy control and MS with a p value of 0.9407 (Figure 6B), and in the white matter with a p value of 0.5377 (Figure 6C). Furthermore, the uptake of [³H]CS1P1 was higher in the MS lesion region of the spinal cord at 116.6 ± 2.919 fmol/mg, similar to that in the brain as expected. Oneway ANOVA showed a statistical difference in the white matter of normal healthy control, nonlesion MS, and MS lesions with a P value of 0.0415 (F(2,9) = 4.624) and Tukey’s multiple comparison tests showed the uptake of [³H]CS1P1 was statistically higher in the MS lesion compared to that in MS nonlesion white matter with a p value of 0.0402 (Figure 6D).

Figure 6.

Comparison of [³H]CS1P1 in the spinal cord of normal, MS nonlesion region and MS lesion region. (A) Representative images of distribution of S1PR1 and [³H]CS1P1 in the normal and MS spinal cord; (B,C) the uptake of [³H]CS1P1 was identical between the gray matter (B) and the white matter (C) of normal and MS spinal cord with p values of 0.941 and 0.538, respectively; (C) one-way ANOVA showed a statistical difference of [³H]CS1P1 uptake among groups in spinal cord F(2,9) = 4.624, P = 0.0415, Tukey’s multiple comparison test following ANOVA showed no significant in the white matter between normal and MS spinal cord (p = 0.756) and normal and MS lesions (p = 0.092), whereas a statistically higher level of [³H]CS1P1 was observed in MS lesion compared to MS nonlesion white matter with a p value of 0.040. Scale bar: 1000 μm.

Characterization of K_d and B_max in the Normal and MS Brain.

Compared to the immunohistochemistry analysis, ligand–receptor autoradiography analysis permits the calculation of the affinity (dissociation constant, K_d) and maximal number of binding sites (B_max) of specific binding sites. We next performed a saturation autoradiography analysis and measured the K_d (nM) and B_max (fmol/mg of wet tissues) of [³H]CS1P1 to S1PR1 in cortical brain tissues and compared the difference between the normal healthy control and MS brain (Figure 7). Saturation autoradiography analysis showed a K_d of 2.7 ± 0.2 nM in the gray matter of normal control brain indicating [³H]CS1P1 is highly potent to S1PR1. A B_max value of 133.4 ± 5.0 fmol/mg was detected in the gray matter of the normal control brain (Figure 7A). On the other hand, a K_d of 2.6 ± 0.3 nM and a B_max of 110.3 ± 12.1 fmol/mg were observed in the white matter of the normal healthy control brain (Figure 7B). For the MS patient brain, a K_d of 3.8 ± 0.4 nM and a B_max of 127.1 ± 4.1 fmol/mg were detected in the gray matter (Figure 7C) and a K_d of 4.1 ± 0.4 nM and a B_max of 110.8 ± 2.4 fmol/mg were detected in the nonlesion white matter (Figure 7D).

Figure 7.

In vivo characterization of K_d and B_max of [³H]CS1P1 in normal and MS brains. Saturation autoradiography analysis of [³H]CS1P1 was performed in the normal and MS cortical tissues. (A,B) In the normal brain, [³H]CS1P1 has a K_b of 2.7 ± 0.2 nM and a B_max of 133.4 ± 5.0 fmol/ mg in the gray matter, whereas a K_b of 2.6 ± 0.3 nM and a B_max of 110.3 ± 12.1 fmol/mg in the white matter; (C,D) in the MS brain, [³H]CS1P1 has a K_d of 3.8 ± 0.4 nM and a B_max of 127.1 ± 4.1 nM and a B_max of 110.8 ± 2.4 fmol/mg in the white matter; (E) representative image of saturation autoradiography analysis in the normal and MS cortical brain.

Interestingly, the K_d values were statistically higher in the gray matter (Figure 8A) and white matter regions of the MS brain (Figure 8B) than the K_d values in both the gray matter and white matter regions of the normal healthy control brain (p = 0.0057 and 0.0016, respectively). The higher binding affinity in the white matter and the gray matter of the MS brain versus the normal healthy control brain indicates a physical, such as S1PR1 structure and activity, or biochemical differences in the MS brain versus in the normal healthy control brain that may interfere with the binding of [³H]CS1P1 to S1PR1. In contrast, K_d was nearly identical between gray and white matter in normal healthy control and MS brain with p values of 0.5384 and 0.5419, respectively, indicating no physical or biochemical differences between gray and white matter of normal healthy control brain as well as MS brain. Together, our data suggest that the K_d values in both gray matter and white matter of the MS brain are higher than the K_d value in the normal healthy control brain, but the K_d values between the gray matter and white matter are similar in the normal healthy control brain or in the MS brain.

Figure 8.

Comparison of K_d and B_max of [³H]CS1P1 in normal and MS brain. (A,B) K_d was statistically higher in the MS brain compared to normal brain in both gray (A) and white matter (B) regions with a p values of 0.0057 and 0.0016 respectively; (C,D) similar level of B_max was identified between normal and MS brain in both gray (C) and white matter (D) regions.

In contrast, the B_max values in the gray matter of the normal healthy control brain and MS brain were almost identical (Figure 8C) with a p value of 0.1434. Similarly, identical B_max values were observed in white matter of the normal healthy control brain and MS brain (Figure 8D) with a p value of 0.9486. Nevertheless, a statistically higher B_max value was observed gray matter versus the white matter in both normal healthy control brain and MS brain brains as expected with p values of 0.0222 and 0.0010 respectively.

DISCUSSION

MS is the most common immune-mediated disease of the CNS and is characterized by demyelination and inflammation.³⁹ While several effective disease-modifying S1PR modulators approved in the last two decades for the treatment of relapsing-remitting MS, the fundamental pathological process that drives the progression of MS remains poorly understood.^3,40 In addition, the early detection of MS remains challenging due to its clinical heterogeneity and lengthy differential diagnosis.¹ Lesion regions determined by MRI and clinical characteristics are key tools and widely used for the diagnosis and prognosis in larger patient groups of MS with difficulties in characterization and prediction in individual patients. A few molecular biomarkers showed initial successes in prediction of diagnosis, prognosis, and the monitoring of treatment response, and longterm and large-scale evaluations are necessary for clinical practice. The 18-kDa translocator protein (TSPO)-binding radiotracers are the most widely used radiopharmaceuticals for imaging neuroinflammation; [¹¹C]PK11195 PET has been used to evaluate microglial response in fingolimod-treated MS patients.⁴¹ Its poor signal-to-noise ratio, high lipophilicity, and low blood–brain barrier penetration prevent the implementation of [¹¹C]PK11195 on clinical practice.^42,43 Consequently, a few second-generation PET radiotracers, such as [¹¹C]PBR28, [¹⁸F]GE180, [¹⁸F]DPA-714, and [¹⁸F]F-DPA for imaging TSPO, have been reported.^44–49 Nevertheless, the Ala147Thr genotype polymorphism of TSPO influences the brain uptake of most of the TSPO radiotracers, resulting in the need for genetic prescreening to prevent high within-group variation,⁵⁰ which creates substantial complexity for clinical studies. S1PR1 has gained significant interest since the approval of the S1PR modulators for the treatment of MS. Recent studies demonstrated S1PR1 ubiquitously expresses in all types of brain cells including astrocytes, oligodendrocytes, neurons, and microglia, and is involved in neuronal plasticity, including myelination, neurogenesis, and neuroprotection in the CNS.¹⁶ PET with a specific S1PR1 radioligand could offer a noninvasive methodology to quantitatively measure S1PR1 expression in vivo and to evaluate the target engagement and therapeutic response of S1PR1-targeting drugs.

We previously reported CS1P1 is a potent and selective S1PR1 ligand, radiolabeled [¹¹C]CS1P1, and tritiated [³H]CS1P1 are suitable quantitative investigations of S1PR1 expression in animal model and human tissues.^24,25,28,29 Once the FDA approved [¹¹C]CS1P1 for human studies, we completed dosimetry estimates in normal healthy volunteers and preliminary data in people with MS revealed its potential for mechanistic investigations of MS.³⁰ To continue our efforts to characterize CS1P1 in the CNS of MS, we focused on the investigation of S1PR1 expression in the postmortem tissues from MS cases and normal healthy control cases using autoradiography analysis with [³H]CS1P1, and immunohistochemistry analysis using corresponding antibodies and cell biomarkers in the present study. Our data show the distribution of [³H]CS1P1 and the expression of S1PR1 in normal healthy control cortical tissues (Figure 1) is consistent with our previous studies in rodent, nonhuman primate, and human CNS tissues.²⁴ [³H]CS1P1 was mainly distributed in the gray matter of the brain, and the distribution of [³H]CS1P1 matched well with the immunohistochemistry analysis of S1PR1, and the uptake of [³H]CS1P1 was able to be blocked with well-validated S1PR1 antagonist NIBR-0213, further demonstrating CS1P1 is a S1PR1-specific radioligand. In addition, the distribution of [³H]CS1P1 matched well with that of astrocyte staining. In fact, in CNS, S1PR1 is primarily expressed in peripheral processes of astrocytes, even greater than in brain endothelial cells.³⁷ Immunohistochemical investigation showed S1PR1 is mainly distributed in the gray matter rather than white matter of CNS with a strong expression in astrocytes rather than neurons.⁵¹ Furthermore, recent studies suggest S1PR1 activation in astrocytes may contribute to neuropathic pain,¹⁹ MS and EAE animal models,^20,22 and neuroinflammation.^23,52 More importantly, recent studies demonstrated that the clinical effectiveness of S1PR1 modulators is not limited to immune cell trafficking, but also suggest a direct neuroprotection effect in the CNS.^8,9,40

We used quantitative autoradiography to quantify the level of [³H]CS1P1 in postmortem CNS tissues from normal healthy control cases and MS cases (Table 1). Compared with the immunohistochemistry method that measures certain epitopes in fixed tissues, autoradiography analysis can evaluate the expression of activated S1PR1 in fresh frozen tissue sections and can thus provide the functionality of a receptor protein on top of information about their anatomical distribution.^32–34 In our studies, we found a higher uptake of [³H]CS1P1 in gray matter compared to white matter in normal healthy control and MS cortical, cerebellum, and spinal cord tissues, as expected. Intriguingly, a statistically higher level of [³H]CS1P1 was observed in the MS lesions in all the tested tissues and regions, whereas an identical level of [³H]CS1P1 in the gray and nonlesion white matter regions was observed between normal and MS CNS tissues (Figures 2–6). Tukey’s multiple comparison tests showed [³H]CS1P1 in the lesions was higher than white matter in both normal healthy and nonlesion MS patients’ cortex and cerebellum. Although in the spinal cord, a statistically higher level of [³H]CS1P1 was observed in the MS lesions compared to MS nonlesion white matter regions and a less but not significant level of [³H]CS1P1 was found in comparison to white matter from the normal healthy control spinal cord, this might result from the fact that the expression of S1PR1 in the white matter of spinal cord is considered higher than the expression of S1PR1 in other CNS brain tissue.⁵¹ Our study is consistent with the previous immunohistochemistry studies²⁰ and demonstrates a direct involvement of S1PR1 in the MS lesion pathology. As mentioned above, S1PR1 is mainly expressed in the astrocytes and is involved in various astrocytic physiological and pathophysiological processes in the CNS. Recent studies also show multiple roles of astrocytes in the evolution of MS lesions.^37,53–56 In our case, we observed an increased uptake of [³H]CS1P1 and expression of S1PR1 in both regions that surround the lesion (Figure 2, MS example 1) and the center of lesion (Figure 2, MS example 2), indicating that S1PR1 may be involved in both late and early stages of lesions. In fact, recent studies demonstrated they are early and active participants in demyelination.^57–60 While the precise mechanism of astrocytic S1PR1 in MS lesion remains not fully understood, our result provides further evidence that activated S1PR1 is distributed and elevated in the early stage and then retained even in the late-stage MS lesions. In addition, this study demonstrates our leading S1PR1-specific radioligand [³H]CS1P1 is promising to quantify S1PR1 expression in the CNS, and it also provides a proof of concept for PET imaging measurement of MS lesions using an S1PR1-specific radioligand. Our data show that the uptake of [³H]CS1P1 in MS lesions in both cortical and cerebellum brain tissues is similar to that in the gray matter. Overall, the MS brain only showed moderate increase on S1PR1 expression compared to normal control brain. This may attribute to the widely expressed S1PR1 in CNS cells, especially highly expressed in gray matters; the radiotracer uptake in the whole cortical gray matter and white matter had no significant changes in the MS versus normal controls. Importantly, the radiotracer uptake in MS lesions is significantly higher than that in the surrounding white matter regions, suggesting that PET with [¹¹C]CS1P1 has high sensitivity in detecting the lesion regions, the inflammatory sites in the MS brain. PET with [¹¹C]CS1P1 may provide a new tool to detect neuroinflammation in MS and other neurological diseases. Together, our current findings are consistent with our findings in animal EAE model of MS and our ongoing PET with [¹¹C]CS1P1 measure of the MS brain.

Table 1.

Characteristics of the Human Brain and Spinal Cord Samples

cohort 1 (cortex and cerebellum samples)
sample ID	sex	age, yrs	diagnosis
1	M	63	control
2	M	65	control
3	F	57	control
4	M	74	control
5	M	74	control
6	F	56	control
7	F	46	control
8	F	73	control
9	M	49	control
10	F	55	control
11	M	70	MS
12	F	46	MS
13	F	58	MS
14	M	49	MS
15	M	59	MS
16	F	74	MS
17	M	61	MS
18	M	73	MS
19	F	56	MS
20	F	59	MS
cohort 2 (spinal cord samples)
sample ID	sex	age, yrs	diagnosis
21	F	56	control
22	M	51	control
23	F	84	control
24	F	62	control
25	M	81	control
26	F	39	control
27	F	39	MS
28	F	54	MS
29	F	69	MS
30	F	79	MS
31	M	54	MS

Quantitative radioligand–receptor autoradiography analysis can provide binding parameters such as the affinity (dissociation constant, K_d) and maximal number of binding sites (B_max) of receptors. To further characterize our leading ligand CS1P1 for potential diagnostic usage in MS and monitor the therapeutic efficacy of treating MS by targeting S1PR1, we also determined and compared the K_d and B_max of [³H]CS1P1 between normal healthy and MS brains (Figures 7 and 8). Similar to our autoradiographic analysis, an identical level of B_max was observed in both gray and nonlesion white matter regions in normal health control and MS brains. In contrast, B_max was statistically lower in the nonlesion white matter than that in gray matter in both normal and MS brains as S1PR1 is highly expressed in gray matter. On the other hand, K_d did not significantly differ between gray and white matter for either normal or MS brains, indicating no physical or biochemical difference exists between gray and white matter that affects [³H]CS1P1 binding in healthy and disease conditions. However, gray and white matter K_d was statistically higher in the MS brain than in gray and white matter in normal brains. This interesting finding suggests that potential physical or biochemical changes in the MS brain can interfere with the binding of [³H]CS1P1 to S1PR1. This change of binding affinity in MS brain could be due to the pathophysiological changes in MS condition that result in structure and/or activity changes in S1PR1, and thus affect the interaction between CS1P1 and S1PR1 molecules. Further characterization of its mechanisms underlines the MS condition and its relevance to MS pathology is necessary.

In summary, we conducted a quantitative assessment of S1PR1 in human post-mortem CNS tissues of MS cases using autoradiography analysis with our leading S1PR1-specific ligand [³H]CS1P1. The level of [³H]CS1P1 was statistically elevated in the lesion of MS compared to that in white matter of healthy normal control brain and the white matter in nonlesion regions of the MS. Although no significant difference was observed in both gray and white matter of cortex, cerebellum, and spinal cord between normal healthy control and nonlesion MS samples, a high binding affinity (K_d) was identified in both gray and white matter of MS brains compared to healthy control brains. A lower B_max was also observed in the white matter of both healthy normal and MS brains. Further investigation of molecular mechanisms of shift K_d in MS and B_max in white matter is of interest to further understand the role of S1PR1MS pathology. Our study further suggests that PET with our S1PR1-specific radiolabeled CS1P1 is a promising clinical methodology for imaging MS lesions and S1PR1-associated neuroinflammation in the CNS.

METHODS

Human Brain Tissue.

Human brain tissues were obtained from the Rocky Mountain Multiple Sclerosis Center Tissue Bank at the University of Colorado, and human spinal cord tissues were obtained from the John L. Trotter MS Center at Washington University in Saint Louis School of Medicine. Fresh frozen brain and spinal cord tissues were used in this study. For brain tissues, whole cerebellum slices and cortical tissue slices that include superior, middle, and inferior regions were obtained at autopsy from 10 MS patients and 10 age-matched healthy normal controls (Table 1). For normal and MS groups, an equal number of male and female samples (n = 5) were included. Normal samples had an average age of 61.20 ± 3.25 (mean ± SEM) years and MS samples had an average age of 60.50 ± 2.99 years. For spinal tissues, cervical spine slides were obtained from five MS patients and six age-matched normal controls (Table 1) with average ages of 59.00 ± 7.83 and 60.50 ± 6.89 years, respectively. All tissues were used in accordance with Washington University guidelines for using postmortem tissues. All samples were snap-frozen and stored at −80 °C until used.

Radioligand Preparation.

[³H]CS1P1 (Figure 1) was custom synthesized by NOVANDI Chemistry AB (Sweden) using our homemade precursors from our laboratory. The final tritiated compounds have a specific radioactivity of 3.0 TBq/mmol, a radiochemical purity >99%, and a radioactive concentration at 37 MBq/mL in ethanol. The radioligand was stored at −80 °C until use.

Immunohistochemistry.

Immunohistochemistry staining was carried out in frozen sections from various regions of the postmortem human CNS tissues.²⁴ In brief, 16-μm sections were prepared and stored at −80 °C until used. For S1PR1, neuronal nuclear protein (NeuN), glial fibrillary acidic protein (GFAP), and ionized calcium binding adaptor molecule 1 (IBA1) staining, sections were air-dried and fixed in 4% paraformaldehyde for 10 min at room temperature (RT), and then washed for 5 min in PBS three times. Sections were then blocked with BLOXALL blocking solution (Vector Laboratories, Burlingame, CA) for 20 min, rinsed in PBS, and then blocked with 10% horse serum for 1 h at RT. All sections were then stained using anti-S1PR1 antibody (Thermo Fisher, Waltham, MA) for S1PR1 staining, anti-NeuN antibody (Novus Biologicals, Centennial, CO) for neuron staining, anti-GFAP antibody (Novus Biologicals, Centennial, CO) for astrocyte staining, or anti-IBA1 antibody (Novus Biologicals, Centennial, CO) for microglia staining at 4 °C overnight. All sections were then washed for 10 min for three times at RT and incubated with ImmPRESS HRP horse antimouse or antirabbit polymer or HRP conjugated horse antigoat IgG antibody (Vector Laboratories, Burlingame, CA) for 1 h at RT, washed for 10 min for three times at RT, and then developed using ImmPACT DAB Substrate Kit (Vector Laboratories, Burlingame, CA). Sections were then rinsed in tap water, dehydrated and cleaned in ethanol and xylene, and mounted with CYTOSEAL XYL (Thermo Fisher, Waltham, MA). For MBP and PLP staining, sections were air-dried and fixed in ice-cold acetone for 10 min at −20 °C, and then air-dried for 20 min at RT. All sections were then stained with anti-MBP (Santa Cruz Biotechnology, Dallas, TX) for MBP or anti-PLP (Bio-Rad Laboratories, Hercules, CA) as described above. For H&E staining, sections were air-dried and fixed in 4% paraformaldehyde for 10 min at RT, and then washed for 10 min in PBS for three times. All sections were then stained with H&E staining kit (Newcomersupply, Middleton, WI) following the manufacturer’s instructions. For all stains, whole slide scan was performed on tissue sections using Zeiss Axio Scan.Z1 (Carl Zeiss, Jena, Germany) and analyzed using ZEN 2.3 (Carl Zeiss, Jena, Germany). All representative images showed were processed using “RGB-high-contrast-Auto-Best-Fit” method in ZEN.

Autoradiography Study.

In vitro autoradiography study was carried out using [³H]CS1P1 in frozen sections of various regions of postmortem human CNS tissues.²⁴ In brief, 14 μm sections were preincubated with HBSS buffer containing 10 mM HEPES, 5 mM MgCl₂, 1 mM CaCl₂, 0.5% BSA, and 0.1 mM EDTA at pH 7.4 for 5 min at RT. All sections were then incubated with 5 nM of [³H]CS1P1 in buffer for 1 h at RT with gentle shaking. After that, all sections were washed with buffer for 5 min at RT three times and then dipped into ice-cold H₂O for 1 min and air-dried overnight. Dried slides were incubated with the Carestream BioMax autoradiography film (Carestream, Rochester, NY) in a Hypercassette autoradiography cassette (Cytiva, Amersham, UK) for 30 days along with an ART-123 Tritium Standards (American Radiolabeled Chemicals, St Louis, MO). The film was then processed using a Kodak film developer (Kodak, Rochester, NY). To determine the nonspecific binding, 10 μM of S1PR1-specific antagonist, NIBR-0213 (Cayman, Ann Arbor, MI), was used. All images were processed and analyzed using Fiji ImageJ, and all representative images showed were processed using a continuous l ut “physics.” CNS regions were identified from the autoradiography images according to the H&E staining in the adjacent slide. Regions of interest (ROIs) were randomly selected in the target regions, and the intensity was measured.

Saturation Autoradiography Analysis.

Saturation autoradiography analysis was carried out using [³H]CS1P1 in cortical brain samples.²⁴ The sections were incubated, washed, and developed as described in the autoradiography study with a serial dilution of [³H]CS1P1 ranging from 30.86 to 0.48 nM. In brief, 14 μm sections were preconditioned by preincubating with buffer followed by incubation with dilutions of [³H]CS1P1 including blank, 0.48, 0.96, 1.93, 3.86, 7.72, 15.43, and 30.86 nM of [³H]CS1P1 in buffer, and then washed and dried overnight. Nonspecific binding was determined by the addition of 10 μM of S1PR1-specific antagonist, NIBR-0213 (Cayman, Ann Arbor, MI). Slides were then developed and processed as described above. ROIs were randomly selected in the target regions, and the intensity was measured.

Statistical Analysis.

All data were analyzed with Prism 9.1 (GraphPad Software, San Diego, CA). For quantitative autoradiography assessment of [³H]CS1P1 in tissues, >8 ROIs (region of interest such as gray or white matter) were randomly selected and measured, and the average intensity of ROIs was used for statistical analysis. Unpaired parametric t-test was used to compare the intensity of [³H]CS1P1 between normal and MS or gray and white matter. One-way ANOVA was used to compare the intensity of [³H]CS1P1 in the white matter among normal, MS nonlesion, and MS lesion; Tukey’s multiple comparison test following one-way ANOVA was used to compare the difference between each pair. For saturation autoradiography analysis, the equilibrium dissociation constant (K_d in nM) and the maximum number of binding sites (B_max in femtomoles per milligram of wet tissues) were determined by nonlinear regression analysis of one site saturation binding model in Prism 9.1. Unpaired parametric t-test was used to compare the K_d and B_max between normal and MS or gray and white matter.