ANTI-INFLAMMATORY POTENCY OF NOVEL ECTO-5`-NUCLEOTIDASE/CD73 INHIBITORS IN ASTROCYTE CULTURE MODEL OF NEUROINFLAMMATION

Abstract

Three novel cytosine-derived α,β-methylene diphosphonates designated MRS4598, MRS4552, and MRS4602 were tested in the range of 1×10⁻⁹ to 1×10⁻³ M for their efficacy and potency in inhibiting membrane-bound ecto-5`-nucleotidase/CD73 activity in primary astrocytes in vitro. The compounds were also tested for their ability to attenuate the reactive astrocyte phenotype induced by proinflammatory cytokines. The main findings are as follows: A) The tested compounds induced concentration-dependent inhibition of CD73 activity, with maximal inhibition achieved at ~1×10⁻³M; B) All compounds showed high inhibitory potency, as reflected by IC₅₀ values in the submicromolar range; C) All compounds showed high binding capacity, as reflected by Ki values in the low nanomolar range; D) Among the tested compounds, MRS4598 showed the highest inhibitory efficacy and potency, as reflected by IC₅₀ and Ki values of 0.11 μM and 18.2 nM; E) Neither compound affected astrocyte proliferation and cell metabolic activity at concentrations near to IC₅₀; E) MRS4598 was able to inhibit CD73 activity in reactive astrocytes stimulated with TNF-α and to induce concentration-dependent inhibition of CD73 in reactive astrocytes stimulated with IL-1β, with an order of magnitude higher IC₅₀ value; F) MRS4598 was the only compound tested that was able to induce shedding of the CD73 from astrocyte membranes and to enhance astrocyte migration in the scratch wound migration assay, albeit at concentration well above its IC₅₀ value. Given the role of CD73 in neurodegenerative diseases, MRS4598, MRS4552, and MRS4602 are promising pharmacological tools for the treatment of neurodegeneration and neuroinflammation.

Graphical Abstract

1. INTRODUCTION

Extracellular ATP acts as a danger signal when released in large amounts from neurons and astrocytes during metabolic or traumatic brain injury. The nucleotides is degraded by stepwise hydrolysis catalyzed by membrane-bound ectonucleotidases. Ecto-nucleoside triphosphate diphosphohydrolase-1 (NTPDase1/CD39) hydrolyzes ATP/ADP to AMP, while ecto-5’-nucleotidase (CD73) catalyzes the rate-limiting step of AMP hydrolysis to adenosine (James and Richardson, 1993). CD73 is a glycoprotein attached to the plasma membrane via a glycosylphosphatidylinositol (GPI) anchor. It is constitutively expressed in astrocytes, ependymal cells, and epithelial cells of the choroid plexus (Langer et al., 2008), while neuronal and presynaptic localizations have been demonstrated by biochemical and enzyme histochemistry methods (Zimmermann, 1992). The nucleoside regulates several functions in the nervous system, including oxygen supply/demand, immune modulation and information flow through neuronal circuits (Boison et al., 2010; Cunha, 2016). Adenosine acts through four adenosine receptors subtypes (ARs), which are G-protein-coupled receptors associated with inhibition (A₁R, A₃R) or stimulation (A_2AR and A_2BR) of adenylate cyclase (Fredholm et al., 2011). ARs have distinct cellular distribution and affinity for adenosine. Accordingly, the impact on cellular processes depends on its extracellular level and the availability of ARs in a given tissue context, as expression of A₁R dominates in healthy brain, whereas the expression of other subtypes, particularly A_2AR increases under neuroinflammatory conditions (Gomes et al., 2011).

There is ample evidence that CD73/A_2AR signaling plays a critical role in neurodegeneration that precedes neuroinflammation (Cunha, 2016; Moreira-de-Sá et al., 2021). CD73 and A_2AR are spatially coupled, such that adenosine directly promotes A_2AR overactivation (Moreira-de-Sá et al., 2021). Excessive A_2AR signaling enhances glutamate excitotoxicity and contributes to synaptic and neuronal damage, demonstrated by neuroprotection through genetic or pharmacological blockade of A_2AR in animal models of neurodegeneration (Barros-Barbosa et al., 2016). Overactivation of A_2AR is sufficient to induce brain dysfunction in the absence of other disease trigger. CD73 and A_2AR establish complexes with membrane proteins and interfere with intercellular signaling in multiple neuropathologies (Borroto-Escuela et al., 2018). Therefore, targeting CD73 could be an effective potential therapeutic target in neurodegenerative diseases associated with neuroinflammation (Moreira-de-Sá et al., 2021).

The level of CD73 varies with cell proliferation and metabolic state in many cell types, including astrocytes. The Nt5e gene is developmentally regulated via TCF/LEF element, a target of Wnt signaling (Resta et al., 1993). Nt5e overexpression has been demonstrated in several tumors, due to the direct action of hypoxia-inducible factor-1 on the gene regulatory elements (Spychala et al., 1999). Moreover, adenosine production induced by ischemic conditions contributes to the immunosuppressive properties of the tumor environment (Boison and Yegutkin, 2019). Due to the above role of CD73 in cancer, a large number of novel chemical ligands have been developed in recent years (Nocentini et al., 2017), while several CD73-targeting chemicals are currently in clinical development for the treatment of cancer (www.clinicaltrials.gov).

In the present work, we investigated the inhibitory efficacy and potency of three cytosine-based nucleoside 5’-α,β-methylene diphosphonates (Scortichini et al., 2022) to inhibit CD73 activity and alter the functional properties of reactive astrocytes in vitro.

2. MATERIALS AND METHODS

2.1. Chemicals

Cytosine-based nucleoside 5’-α,β-methylene diphosphonates MRS4598, MRS4552, and MRS4602 were synthesized and purified by HPLC (≥ 95% purity) as triethylammonium salts, as previously reported (Junker et al., 2019; Scortichini et al., 2022). The chemical formulas and structural formulas are shown in Fig. 1. Inhibitors were dissolved in sterile water, in final concentrations as indicated.

Figure 1. Evaluation of the inhibition of CD73 activity by cytosine-based 5’-α,β-methylene diphosphonates MRS4598, MRS4552, and MRS4602.

Structural and chemical formulas (A, E, C) and concentration-response curve of MRS4598 (B), MRS4552 (D), and MRS4602 (F). Concentration-response curves were obtained by fitting raw data (percent enzyme inhibition vs [inhibitor concentration]) to the sigmoid curve using a four-parameter logistic regression model. Black symbols represent individual measurements, colored symbols represent mean inhibition (E) at each inhibitor concentration ± SEM from n ≥ 3 independent measurements.

2.2. Animals

One-day-old male rat pups of the Wistar strain from the local colony were used. The animal experiments were performed in compliance with European Communities Council Directive (2010/63/EU) and the Serbian Laboratory Animal Science Association for the Protection of Animals Used for Experimental and the Scientific Purposes and were approved by the Ethical Committee for the use of Laboratory Animals, Vinča Institute of Nuclear Sciences (No. 116–14/2020).

2.3. Primary astrocyte cultures

Primary astrocyte cultures were prepared as previously described (Adzic and Nedeljkovic, 2018). Briefly, cortices were isolated and mechanically dissociated under sterile conditions in Leibovitz’s L-15 isolation medium containing 2 mM L-glutamine, 100 IU/ml penicillin, 0.1 mg/ml streptomycin, and 0.1% BSA. After two centrifugation steps at 500 × g for 4 min, the cell suspension was passed through 21 G and 23 G sterile needles (to remove residual tissue aggregates). After another centrifugation step at 500 × g for 4 minutes, cells were resuspended in DMEM supplemented with 10% heat-inactivated FBS, 25 mM glucose, 2 mM L-glutamine, 1 mM sodium pyruvate, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin. Cells were then seeded in culture flasks and grown at 37°C in a humidified incubator containing 5% CO₂/95% air, with the culture medium replaced every third day. After 6 to 8 days in culture, primary microglia and oligodendrocytes were removed by vigorous shaking at 400 rpm for 16–20 hours on a plate shaker (Perkin Elmer, Turku, Finland). Adherent primary astrocytes were washed with PBS, trypsinized (0.25% trypsin and 0.02% EDTA), and propagated at a density of 1.5 × 10⁴ cells/cm² and maintained until confluence was reached. The vast majority of cells in the cultures (98.6%) were GFAP-expressing cells, whereas the remaining cells represented a microglial fraction of less than 2%, based on Iba-1 immunostaining. CNPase-expressing oligodendrocytes were not detected (data not shown).

2.4. Ecto-5`-nucleotidase/CD73 enzyme assay

Ecto-5`-nucleotidase/CD73 hydrolyzes 5`-AMP to adenosine with release of phosphate groups (Pi). Therefore, the phosphohydrolase activity of CD73 was measured by determining the amount of free phosphates (Pi) released as a product of the enzyme reaction. The 5`-AMP phosphohydrolase activity was measured in confluent astrocyte cultures seeded on 24-well plates (Adzic and Nedeljkovic, 2018). Cells were washed in phosphate-free medium (117 mM NaCl, 5.3 mM KCl, 1.8 mM MgCl₂, 26 mM NaHCO₃, 10 mM glucose, 10 mM HEPES, pH 7.4) to remove cell debris and residual free phosphates. Cells were preincubated for 15 minutes with inhibitor dissolved in sterile water at the indicated concentration. The reaction was initiated by adding 240 μl of 1.0 mM AMP to the culture medium and lasted for 30 minutes/37°C. The control culture was incubated without inhibitor. The reaction was stopped by transferring the reaction volume into tubes containing 26.5 μl of ice-cold 3 M perchloroacetic acid. The amount of Pi released as a product of the enzymatic reaction was determined by the malachite green method using KH₂PO₄ as a standard (Brisevac et al., 2012). The absorbance was measured at 620 nm. After the assay, the total protein content in each well was determined using the Micro BCA Protein Assay Kit (Thermo Fisher Scientific, Rockford, USA) in sample obtained by scraping cells in 100 μl of RIPA buffer. CD73 activity in each well was expressed as the amount of Pi (nmol) per milligram of protein per minute (± SEM), from n ≥ 3 independent determinations. The effect of inhibitor on CD73 activity was expressed as percent inhibition (%) relative to control (± SEM).

2.5. Inhibition analysys

The compounds were tested in a concentration ranging from of 1×10⁻⁹ to 1 ×10⁻³ M [I]. The enzyme activity obtained without inhibitor was defined as the maximum activity (Vmax). The difference between Vmax and the activity obtained at each compound concentration was expressed relative to Vmax as percent inhibition (E) and was plotted as a function of log [I]. The concentration-response curve was generated by fitting the raw data to the four-parameter sigmoid function with nonlinear regression analysis using the Origin 7.0 software package. The inhibitory efficacy was defined as the maximum inhibitory effect (Emax) reached at the plateau at infinite compound concentration. The inhibitory potency or functional strength of the inhibitor was evaluated by determining the concentration of compound required to produce 50% inhibition of the enzyme (IC₅₀). The IC₅₀ value was determined by Dixon linearization of concentration-response data (E vs [I]). The binding capacity of the inhibitor was evaluated by determining the inhibition constant (Ki) using the Cheng-Prusoff equation (Cheng and Prusoff, 1973) and assuming the value of Km for rat CD73 of 53.0 ± 4.1 μM (Junker et al., 2019).

2.6. MTT and proliferation assays

The effects of the compounds on the metabolic activity was assessed using the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide assay (Sigma-Aldrich, St. Louis, USA), as described (Brisevac et al., 2015). Primary astrocytes were seeded on 24-well plates (1.5×10⁴ cells/cm²) and kept under serum deprivation for 24 hours. Cultures were incubated with the drug at the indicated final concentration for 24 hours. The next day, fresh medium containing MTT (at a final concentration of 1 mg/mL) was added to the cultures. The cells were incubated at 37°C for 30 minutes, and the formazan formed was dissolved in dimethyl sulfoxide (DMSO). Optical density was measured at 570 nm using the LKB 5060 Multiplate Reader. Formazan formation as a measure of cell metabolic activity was expressed as a percentage of untreated controls ± SEM, with two separate determinations in quadruplicate.

2.7. Western blot and dot blot

For Western blot analysis, astrocyte samples were deglycosylated with PNGaseF (New England Biolabs), which is capable of removing N-linked glycan molecules from membrane glycoproteins. Aliquots of astrocyte samples containing 10 μg of proteins were resuspended with 2 μl of 10×GlycoBuffer2 and PNGaseF was added to a final concentration of 500 U/ml. The mixture was incubated at 37°C for 8 hours and then immediately used for further analysis. Equivalent amounts of native (non-deglycosylated) and deglycosylated samples were diluted with 6× Laemmli sample buffer, resolved on 7.5% SDS-PAGE gels, and transferred to a PVDF support membrane. After blocking with 5% BSA (Sigma-Aldrich, USA) in Tris-buffer saline/Tween 20 (TBST), the blot was probed overnight at 4°C with the rabbit monoclonal antibodies against CD73 (1:1500 dilution in TBST, Cell Signaling Cat. #13160). The support membrane was incubated for 2 hours at room temperature (RT) with donkey anti-rabbit IgG-horseradish peroxidase-conjugated secondary antibodies (1:10,000 dilution in TBST; sc-2305, Santa Cruz Biotechnology, Santa Cruz, CA, USA).Release of soluble CD73 in the culture medium was detected by dot blot chemiluminescence detection (Adzic et al., 2017). Aliquots of culture media (100 μl) were spotted onto a polyvinylidene fluoride (PVDF) support membrane (Immobilon-P transfer membrane, Millipore) in a vacuum-based minifold dot blot device (Schleicher & Schuell Inc., Keene, N.H.). The PVDF membrane was blocked with 5% BSA in Tris-buffered saline containing 0.1% TBST and probed with primary rabbit anti-rat CD73 (1:1500 in TBST) and donkey anti-rabbit HRP-conjugated IgG (1:10,000 in TBST).

Bands and dots were visualized on X-ray films (Kodak) using Clarity Max Western ECL Substrate (Bio Rad) and Chemi Doc-It Imaging System (UVP, Upland, CA, USA). Relative chemiluminescence intensity was determined in ImageJ.

2.8. Bright field microscopy and scratch wound migration assay

Astrocytes were seeded at a density of 1.5 ×10⁴ cells/cm² on a 24-well plate and maintained until near confluence was reached. Twenty-four hours after addition of the tested compound, images of 3 random fields per dish were acquired with Carl Zeiss AxioObserver A1 inverted microscope (A-Plan 10× objective) equipped with an EM512 digital CCD Camera System (Evolve, Photometrics).

For the scratch wound migration assay, astrocytes were seeded at a density of 2 ×10⁴ cells/cm² on 35-mm Petri dishes for adherent cells, as described previously (Adzic et al., 2017). A wound was created in a confluent astrocyte monolayer by scraping the bottom of the dish with a sterile 200-μl pipette tip. Three to four scratches were created per Petri dish in a defined geometry and only scratches with an initial width of 250–300 μm were analyzed. MRS4598 was applied to the culture subsequently and culture continued to be maintained in normal growth medium. Eight to ten random fields per dish, in n = 3 independent culture preparations, were imaged at time 0 hours time on the Carl Zeiss AxioObserver A1 inverted microscope. Consecutive images of selected microscopic fields were acquired after 8 and 24 hours and stored as digitized data. Wound area (μm²) was determined for each image and time point using the ImageJ software package. Wound closure (%) in the absence (control) or presence of MRS4598 was assessed by expressing the area uncovered at each time point as a percentage of the initial wound area (0 hours). Data are expressed as mean closed area (% ± SEM), from n > 22 random images per inhibitor concentration.

2.9. Cell proliferation

The impact of tested compounds on cell proliferation was assessed by double Ki67/DAPI fluorescence staining (Adzic et al., 2017), which determines the percentage of cells expressing the nuclear proliferation marker Ki67 within the entire cell population labeled with the nuclear dye 4’,6-diamidino-2-phenylindole (DAPI). Astrocytes were seeded at a density of 1.5×10⁴ cells/cm² on PLL-coated glass coverslips. Twenty-four hours after addition of a compound, cells were fixed and permeabilized with 0.1% Triton X-100 and then blocked in 5% BSA in 0.01 M PBS. The primary rabbit anti-Ki67 antibody (1:500, Abcam, ab15589) was applied overnight in 2% BSA at +4°C, followed by incubation with the secondary donkey anti-rabbit antibody AlexaFluor-555 (1:200, Initrogen A-21428) for 2 hours at room temperature. In the next round, DAPI (1:4,000) was applied for 10 minutes at room temperature. The coverslips were mounted on microscopic slides with Mowiol solution. Eight images of the microscopic fields were acquired with the Carl Zeiss Axio Observer A1 inverted epifluorescence microscope (A-Plan 10× objective), with the EM512 Digital CCD Camera System (Evolve, Photometrics). Numbers of Ki67+ and DAPI+ cells were counted in ImageJ (NIH, Cell counter plugin), and the proliferation rate was expressed as the mean Ki67+/DAPI+ (%) ± SEM of n = 2 separate culture preparations.

2.10. Immunofluorescence confocal microscopy

Astrocytes grown on PLL-coated glass coverslips (15 mm) were treated with inhibitors for 24 hours. Cells were prefixed in 4% PFA and blocked with 5% BSA in 0.01 M PBS for 1 hour at room temperature. Cells were incubated with rabbit anti-rat GFAP antibody (1:500 in 1% BSA in PBS; DAKO, Agilent Z0334) for 1 hour overnight and secondary donkey anti-rabbit Alexa Fluor-555 IgG antibody (1:200 in PBS; Invitrogen A21428) for 2 hours at room temperature. Nuclei were counterstained with DAPI, for 10 minutes at room temperature, and cells were mounted on slides with MOWIOL solution.

Images of the microscopic fields were acquired with a confocal laser scanning microscope (LSM 510, Carl Zeiss GmbH, Jena, Germany) using an Ar Multiline laser (457, 478, 488, and 514 nm) with a 63× DIC oil objective and an AxioCam ICm1 monochrome camera (Carl Zeiss GmbH, Germany). Cell surface area (μm²) and shape descriptors were determined by analyzing the microscopic images in ImageJ. Values are mean ± SEM, obtained from at least 22 cells in ≥ 2 separate cultures.

2.11. qPCR analysis

Gene expression was determined by qRT-PCR as previously described (Adzic and Nedeljkovic, 2018). Astrocytes were grown on 6-well plates until confluence was reached and treated with 100 ng/ml TNF-α or IL-1β for 24 hours. Samples were collected using TRIzol reagent to isolate total RNA and purity and concentration were determined using OD 260/OD280 and OD260, respectively. Complementary DNA (cDNA) was synthesized using the High-Capacity cDNA Reverse Transcription Kit (ThermoFisher Scientific, MA, USA). The qPCR reaction mix contained 5 μl QTM SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, United States), 0.5 μl primers (100 pmol/μl), 2 μL RNAse-free water, and 2 μl cDNA (20 ng in 10 μl qPCR reaction). Reactions were performed under the optimized conditions in the QuantStudio TM3 Real-Time PCR System (Applied Biosystems, Foster City, CA, United States). Expression of target genes was determined using the following primer sequences: Nt5e: F: CAAATCTGCCTCTGGAAAGC, R: ACCTTCCAGAAGGACCCTGT; Lcn2: F: GGATCAGAACATTCGTTCCA, R: GGATGGAATTGTGAGGGAGA; C3: F: GCGGTACTACCAGACCATCG, R: CTTCTGGCACGAACCTTCAGT; Gapdh (internal control): F: TGGACCTCATGGCCTACAT, R: GGATGGAATTGTGAGGGAGA). Quantification was performed using the 2^−ΔCt method, from n = 3 separate determinations.

2.12. Data analysis

All values are means ± SE of n independent determinations and separate culture preparations. Data were analyzed for normality using the Shapiro-Wilk test. Student’s t-test was used for two-sample comparisons at the 95% confidence level. For multiple comparisons, one-way parametric ANOVA was used. The Bonferroni test was used for post hoc multiple comparisons at a significance level of 0.05 and a target power of 80–95%. All statistical analyzes and graphical representations were prepared using Origin 7.0 computer software.

3. RESULTS

3.1. Inhibition analysis

Three cytosine-derived α,β-methylene diphosphonate derivatives designated MRS4598 (Fig. 1A), MRS4552 (Fig. 1C), and MRS4602 (Fig. 1E) were tested for their efficacy and potency in inhibiting membrane-bound CD73 activity in primary astrocytes. The ability of the inhibitors to affect cellular processes regulated by extracellular adenosine was also investigated. Because CD73 has 5’-AMP-phosphohydrolase activity, the enzyme activity was measured as the amount of inorganic phosphate (Pi) released by hydrolysis of AMP. The assay was performed in a phosphate-free medium, with inhibitors dissolved in sterile water and added 15 minutes before the start of the reaction. The canonical CD73 inhibitor APCP was used for comparison. The rate of CD73 activity obtained in the presence of inhibitor at each concentration [I] was expressed relative to Vmax activity obtained without inhibitor (9.1 ± 2.4 nM Pi/mg protein/min). Concentration-response curves were constructed by plotting percent enzyme inhibiton against log [I] (Fig 1B, D, F). All the inhibitors decreased CD73 activity in primary astrocytes in a concentration-dependent manner. The compounds exhibited significantly higher efficacy than APCP. MRS4598 displayed the highest inhibitory efficacy as it produced a complete inhibition of the enzyme at 1×10⁻³ M. IC₅₀ values were derived from concentration-response curves by Dixon linearization of the data (1/V versus [I]) (see inset in the graphs). Ki values were calculated using the Cheng-Prussof equation (Cheng and Prusoff, 1973). The inhibition parameters, Emax, IC₅₀ and Ki are summarized in Table 1.

Table. 1.

Inhibition parameters

Inhibitor	Emax (%)	IC50 (μM)	Ki (nM) (% inhibition at indicated concentration)
APCP	82.7 ± 10.5	2.63 ± 1.76	132 ± 89 (18.6)
MRS4598	97.3 ± 6.4	0.11 ± 0.08	18.2 ± 2.5 (53.4)
MRS4552	90.1 ± 2.8	0.68 ± 0.14	23.2 ± 0.1 (32.3)
MRS4602	95.7 ± 3.3	0.22 ± 0.07	33.5 ± 2.5 (54.8)

3.2. Impact of CD73 inhibitors on cell metabolic activity, viability and proliferation

The effect of the compounds on cell status was further tested using the MTT assay, which reflects the metabolic activity of cells (Fig. 2A). MRS4552 and MRS4602 had no effect on formazan formation even at the highest concentration tested. MRS4598 showed no effect at a concentration of 1 μM, whereas it reduced formazan formation by 40% at concentrations above 10 μM, but without apparent concentration dependence (Fig. 2A, blue line). Although the reduction in dye formation in the MTT assay primarily reflects a decrease in the metabolic activity of the cells, it may indirectly indicate cytotoxicity of the drug. Therefore, we further examined the cells by bright-field microscopy (Fig. 2B). There were no signs of cell death, such as cell shrinkage, membrane and nuclear fragmentation, or cell detachment. We next analyzed the effects on proliferation by double Ki67/DAPI fluorescence staining (Fig. 2C, D). The proliferation rate in the control culture was ~22%. MRS4598 at concentrations of 50 μM and 100 μM resulted in a modest but significant decrease in proliferation compared to the control (p < 0.05), while other compounds, including APCP, had no apparent effect on cell proliferation.

Figure 2. Impact of compounds on metabolic activity and cellular state of astrocytes.

A) Cell metabolic activity was determined by MTT assay in the range of inhibitor concentrations and concentration-response curves for MRS4598 (blue line), MRS4552 (pink line), and MRS4602 (green line) are presented. Dotted line represents basal metabolic activity of control cells, defined as 100%. Significance inside graph: *p < 0.05. B) Bright field images of astrocytes treated with the compounds at the indicated concentrations, taken 24 hours after treatment. Scale bar = 100 μm. C) The proliferation rate was determined by double Ki67/DAPI fluorescence, in the presence of inhibitors for 24 hours at the indicated concentrations. Proliferation rate (%) was expressed as the proportion of Ki67+ cells out of total number of cells (DAPI+) cells. Dotted line represents basal proliferation rate in control culture. D) Representative images of Ki67/DAPI staining in the presence of compounds at indicated concentrations. Scale bar: 100 μm.

3.3. Impact of CD73 inhibitors on astrocyte morphology

Next, we performed morphometric analysis to further evaluate the impact of tested compounds. Cells were immunostained against GFAP and counterstained with DAPI (Fig. 3). With increasing concentrations of the compounds, cells gradually changed their epithelioid morphology, which is common for cultured astrocytes, to a stellate shape. The changes in cell shape were quantitatively assessed by determining cell surface and shape descriptor indices - circularity, aspect ratio, roundness, and solidity. As summarized in Table 2, the compounds caused a significant reduction in cell surface area (p < 0.05) when present at the highest concentration. Analysis of cell shape indices point to a decrease in the circularity index, implying a more irregular morphology, and the solidity index, characterizing cells with process expansion. Thus, the morphometric analysis indicates that cells treated with the compounds at a concentration of 100 μM did not show signs of cell death, but rather assumed a shape similar to their typical in situ phenotype with small cell bodies and protruding projections.

Figure 3. The influence of compounds on astrocyte morphology.

Astrocytes treated with the indicated concentration of compound were prefixed after 24 hours and immunostained for GFAP, whereas nuclei were counterstained with DAPI. Images of the microscopic fields were acquired using confocal laser scanning microscopy. Images were used to determine cell surface area and shape descriptors (Table 1) in ImageJ. Scale bar in a): 20 μm.

Table 2.

Inhibitors affect cell size and and shape

	Area (μm2)	Circularity	Aspect ratio	Roundness	Solidity
Control	3224.7 ± 256.5	0.55 ± 0.02	2.5 ± 0.3	0.48 ± 0.03	0.85 ± 0.02
100 μM APCP	3181.9 ± 607.2	0.49 ± 0.03	2.2 ± 0.2	0.54 ± 0.07	0.75 ± 0.04*
10 μM MRS4598	3619.4 ± 552.5	0.55 ± 0.06	2.3 ± 0.2	0.47 ± 0.04	0.84 ± 0.04
100 μM MRS4598	1861.8 ± 137.5*	0.43 ± 0.03*	2.7 ± 0.2	0.42 ± 0.04	0.77 ± 0.02
10 μM MRS4552	3159.6 ± 705.7	0.40 ± 0.05*	2.9 ± 0.5	0.44 ± 0.06	0.73 ± 0.05
100 μM MRS4552	2173.9 ± 287.7*	0.37 ± 0.04*	2.7 ± 0.4	0.48 ± 0.05	0.66 ± 0.05*
10 μM MRS4602	2832.4 ± 151.3*	0.54 ± 0.04	2.4 ± 0.3	0.54 ± 0.04	0.80 ± 0.03
100 μM MRS4602	2375.4 ±352.9*	0.50 ± 0.04	2.4 ± 0.2	0.46 ± 0.05	0.82 ± 0.05

Mean area (μm²) and shape descriptor values expressed as a mean ± SE, from n ≥ 22 cells in 3 random and non-overlapping microscopic fields.

^*p < 0.05 in respect to control.

3.4. Impact of MRS4598 on CD73 activity in reactive astrocyte stimulated with inflammatory cytokines

For MRS4598, which had shown the highest inhibitory potency, the analysis was extended to evaluate the ability to inhibit CD73 variant(s) in reactive astrocytes under neuroinflammatory conditions. Astrocytes were treated with TNFα or IL −1β for 24 hours and responded with the reactive astrocyte phenotype confirmed by several-fold induction of C3 and Lcn genes, which were used as markers of neuroinflammation (Table 3). The cytokines also induced upregulation of the Nt5e gene (Table 3, p < 0.05) and the corresponding increase in CD73 at the protein (Fig. 4A) and at the functional levels (Fig. 4B, p < 0.01). To test the efficacy of MRS4598 in inhibiting CD73 in reactive astrocytes, the enzyme was assayed in the presence of 50 μM MRS4598 in cells treated with TNFα or IL-1β for 24 hours. The inhibitor decreased the catalytic activity of CD73 (Figure 4B), demonstrating the efficacy of inhibiting CD73 in both control and reactive astrocytes.

Table 3.

qRT-PCR analysis of inflammatory genes expression

Target gene-mRNA abundance (fold-increase in respect to non-stimulated control)
Gene	TNFα	IL-1β
C3	4.65 ± 2.29*	3.13 ± 0.69*
Lcn	49.04 ± 3.34*	24.22 ± 6.21*
Nt5e	2.49 ± 0.33*	1.77 0.49*

^*Statistical significance at p < 0.05

Figure 4. The inhibitory potency of MRS4598 on CD73 activity in astrocytes treated with inflammatory cytokines.

A) Representative immunoblot of astrocyte membrane samples isolated from control cultures and cultures treated with TNF-α and IL −1β. Equivalent amounts of native samples or PNGaseF-treated samples (dControl, dTNF-α, and dIL-1β) were resolved on SDS-PAGE, transffered onto PVDF, and immunoblotted with antiCD73 antibodies. B) CD73 activity in control culture and in cultures treated with TNF-α or IL-1β in the presence (blue bars) or absence (gray bars) of 50 μM MRS4598. Significance inside the graph: # p < 0.05 between cytokine-treated culture and control; * p < 0.05 between cultures exposed to MRS4598 and the corresponding control. C) Concentration-response curve to MRS4598 ranging from 1–250 μM in astrocytes treated with 0.1 μg/ml IL −1β. Data were fitted to a sigmoid curve (red lines) with four-parameter sigmoid function nonlinear regression analysis using Origin 7.0.

A more detailed analysis of the immunoblot support membrane showed that samples from cytokine-treated cells differed from that of control (Fig. 4A). Specifically, in contrast to control sample, which had one dominant protein band of ~70 kDa, two more pronaunced protein bands were seen in the sample from TNFα-treated cells and multiple protein bands were seen in the sample from IL-1β-treated cells (Fig. 4A). After subjecting samples to enzymatic deglycosylation with PNGaseF before immunoblotting, only one major ~70 kDa band was obtained, confirming that multiple bands represent CD73 glycovariants. Because pattern of glycosylation can alter kinetic properties, we next analyzed the efficacy of MRS4598 to inhibit CD73 in astrocytes stimulated with 0.1 μg/ml IL −1β (Fig. 4C). MRS4598 produced a concentration-dependent inhibition of CD73, with an IC50 value of 5.00 ± 0.26 μM.

3.5. Impact of MRS4598 on astrocytes migration in scratch wound assay

Since membrane-bound CD73 functions as a cell adhesion molecule, we further explored impact of MRS4598 on astrocyte migration in the scratch wound migration experiment. MRS4598 was applied at concentration in the a range of 1 –100 μM, and images were acquired after 0, 8 and 24 hours (Fig. 5A). Area remaining at each time point was expressed as a percentage of initial wound area (%) (Fig. 5B). MRS4598 did not affect migration velocity of astrocytes, except at the highest tested concentration of 100 μM.

Figure 5. The impact of MRS4598 on astrocyte migration in scratch wound assay.

A) Representative images of astrocytes in the presesnce of MRS4598 (1–100 μM), taken immediately after creating a wound and after 8 and 24 hours. Scale bar = 200 μm. B) The impact of MRS4598 is assessed by determining wound closure (%) as uncovered area at 8 and 24 hours, expressed as a percentage of the initial wound area at 0 hours.

3.6. Impact of MRS4598 on CD73 solubilization

Cleavage of the GPI anchor and release of the soluble enzyme molecule is an important mechanism of regulation of CD73 under physiological and pathological conditions. Therefore, we examined the ability of MRS4598 to induce the release of CD73 by detecting the presence of the CD73 molecule in astrocyte culture media collected 15 minutes after addition of the inhibitor. As shown in Fig. 6, the amount of CD73 released in the culture medium increased with MRS4598, whereas APCP was ineffective at all concentrations tested.

Figure 6. The impact of MRS4598 on CD73 release from astrocyte membranes.

A) Representative dot-blot membrane of soluble CD73 detected in culture media, 24 hours after addition of APCP or MRS4598 (both in the range of 1–100 μM). The presence of soluble CD73 was detected by anti-CD73 antibody and chemiluminescence detection. B) Integrated density values determined in ImageJ.

4. DISCUSSION

In the present study, we investigated the efficacy and inhibitory potency of three cytosine-derived α,β-methylene diphosphonates (Junker et al., 2019; Scortichini et al., 2022) in inhibiting membrane-bound CD73 and attenuating the reactive phenotype of astrocytes in vitro. CD73 is the major source of extracellular adenosine (Cunha, 2016), which plays a central role in a variety of physiological processes in the CNS (Boison, 2010), as well as in neuropathological diseases such as Alzheimer’s disease (AD), Parkinson’s disease (PD), and multiple sclerosis (MS) (Cellai et al., 2018; Meng et al., 2019). For example, in PD, increased CD73-derived adenosine and bolstered A_2AR activation are responsible for the abberant synaptic plasticity that precedes the onset of motor symptoms (Gonçalves et al., 2023). Excessive CD73-A_2AR signaling is also responsible for morphological and functional changes in the inflammatory phenotype of astrocytes and microglia (Agostinho et al., 2020). Indeed, increased astroglial CD73 and microglial A_2AR expression characterize chronic neuroinflammation in diseases such as AD and PD (Cellai et al., 2018; Meng et al., 2019). Therefore, CD73 may be an effective therapeutic target for the treatment of chronic neurodegenerative diseases, assuming that suppression of chronic glial activation can improve neurological deficits in the neurodegenerative diseases. Considering that a large number of novel CD73 ligands have been developed in the context of other medical indications, particularly cancer (Junker et al., 2019; Nocentini et al., 2021), there is immense therapeutic potential that can be repurposed for neurodegenerative diseases.

5’-(α,β-methylene)diphosphate adenosine (APCP) was the first nucleotide derivative described as a potent CD73 inhibitor (Bhattarai et al., 2012) and was therefore established as the lead molecule for the development of new adenine nucleotide-based analogs. Several anthraquinone (Baqi et al., 2010; Sharif et al., 2021) and sulfonamide derivatives (Ripphausen et al., 2012) have also been described as competitive inhibitors, while polyphenols (Braganhol et al., 2007) and polyoxometalates (Lee et al., 2015) have been described as noncompetitive CD73 inhibitors. In recent years, several monoclonal antibodies and nanobodies targeting CD73 have been developed (Bendell et al., 2023; Menzel et al., 2018; Rahimova et al., 2018), as well as dual-acting blockers containing both a thiazolo[5,4-d]pyrimidine core, which is essential for blocking A_2AR, and a benzenesulfonamide group, which is typical of CD73 inhibitors (Varano et al., 2020). Recent studies investigated the structure-activity relationships of a series of purine- and pyrimidine-based nucleoside 5’-α,β-methylene diphosphonates as inhibitors of human soluble CD73 in vitro, with Ki values in the low nanomolar range (Junker et al., 2019; Scortichini et al., 2022). These studies revealed several cytosine-derived analogs among the most potent CD73 inhibitors reported in the literature and thus good candidates for insfurther development (Junker et al., 2019; Scortichini et al., 2022). In the present study, we tested compounds MRS4598, MRS4552, and MRS4602 which showed the highest inhibitory potency against human soluble CD73 (Scortichini et al., 2022). Additional advantage of these compounds over adenine nucleotide analogs is that their parent nucleoside, cytosine, does not activate ARs, particulalry A_2AR, which may enhance the intended effect of blocking CD73-A_2AR signalization. The small molecule inhibitors may have an additional advantage in the treatment of neurological disorders because they can be engineered with significant permeability across the blood-brain barrier to be developed as anti-inflammatory agents.

All compounds produced concentration-dependent inhibition of membrane-bound CD73 activity in astrocytes with higher efficacy compared with APCP. The compounds also showed much higher inhibitory potency than APCP, with IC₅₀ values in the low micromolar range and Ki values in the low nanomolar range. MRS4598 showed both the highest efficacy and the potency, as reflected by the lowest IC₅₀ and Ki values among tested compounds. Additionally, MRS4598 showed favorable inhibitory activity compared to a number of CD73 inhibitors currently under preclinical investigation (Nocentini et al., 2019). In terms of effects on cell metabolic activity and proliferation, MRS4552 and MRS4602 had little effect even at the highest concentration tested. On the other hand, MRS4598 caused a significant decrease in the metabolic activity and cell proliferation rate at concentrations of 10 μM and higher, which may be due to its cytotoxic effect. However, morphological analysis did not confirm cytotoxicity, as the culture showed no signs of cell death, such as cell shrinkage, membrane and nuclear fragmentation, and cell detachment. Instead, all compounds, including MRS4598, induced stellation typical of astrocyte activation in culture. Considering that the IC₅₀ value for MRS4598 at which it effectively inhibits CD73 is several orders of magnitude below the concentration that could potentially reduce cell metabolism or induce cell death, it was concluded that MRS4598 has negligible cytotoxic effects on primary astrocytes at concentrations as low as 1 μM.

Adenosine is involved in a number of cellular processes under physiological and pathological conditions, including cell growth, differentiation, and immunosuppression (Schwaninger et al., 1997; Hasko and Cronstein, 2004). With respect to proliferation, adenosine can have both positive and negative effects in various in vitro systems, depending on which ARs are present. Specifically, activation of A_2ARs reduces cell viability while promoting cell proliferation, whereas activation of A₁R and A₃R reduces cell proliferation (Agostinho et al., 2020). Although astrocytes express transcripts for all ARs receptors, the expression of A₁Rs dominates under control conditions, while A_2AR and A_2BR increase upon neuroinflammation (Matos et al., 2012; Barros-Barbosa et al., 2016; Adzic and Nedeljkovic, 2018). In our study, MRS4598 decreased the basal proliferation rate in astrocyte cultures. The proliferation rate of astrocytes in vivo is limited, especially in the context of neuroinflammation. In animal models, the proliferation rate is ~ 10% in stab wound (Simon et al., 2011), ~ 3% in AD (Kamphuis et al., 2012; Sirko et al., 2013), and 7% in ALS (Lepore et al., 2008). Considerable proliferative astrocyte reactivity occurs after trauma when the reactive response forms a protective scar around the injury (Anderson et al., 2016) and in glioma and glioblastoma cells, which exhibit the most intense proliferation (Kim et al., 2013). Therefore, in the context of glioma, at higher concentrations, MRS4598 may have a useful property for selective action against the highly proliferative cancer cells.

Cells have the ability to perceive numerous cues in their physical environment and respond by dynamically changing their shape. Thus, shape reflects the physiological state of the cell, such as proliferation, differentiation, or apoptosis (Vogel and Sheetz, 2006). Cell shape is often used as a pathological indicator and prognostic marker (Yu et al., 2013). Since activation of A₁R and A_2AR modulates the activity of Rho kinase, which is involved in microfilament polimerization (Agostinho et al., 2018), the effect of the inhibitors was also investigated by morphometric analysis. No effects on cell morphology were detected at concentrations less than 10 μM, which is much above the IC₅₀ values. The compounds at 100 μM induced significant changes in cell morphology, as they decreased the average cell body area and resulted in a star-shaped morphology. The changes from polygonal to a process-bearing shape in astrocytes in vitro occur with fluctuations in intracellular cAMP and cGMP (Won and Oh, 2000), activation of AMP-dependent protein kinase (Favero and Mandell, 2007), release of IL −1β (John et al., 2004), and action of nonsteroidal anti-inflammatory drugs (Lichtenstein et al., 2010). Because all ARs are coupled to adenyl cyclase, the resulting changes in cAMP levels directly affect cell morphology via cAMP-dependent activation of Rho GTPase and Rho kinase (ROCK) and their remodeling of the actin cytoskeleton (Rosso et al., 2007). Stellation is observed during astrocyte differentiation and reactive astrogliosis. The fact that the CD73 inhibitor MRS4598 triggered stellation suggests that finely tuned levels of adenosine play a critical role in astrocyte physiology.

Because MRS4598 exhibited the highest inhibitory potency among tested compounds, its pharmacological characterization in primary astrocyte cultures was further extended. MRS4598 demonstrated the ability to inhibit CD73 in reactive astrocytes, in which overexpression of CD73 is induced by inflammatory cytokines. Reactive astrocytes have been shown to express distinct CD73 glycoforms in vivo due to altered N-glycosylation under inflammatory conditions (Lavrnja et al., 2015). The type of glycan molecules bound to the CD73 protein core primarily affects the cell adhesion properties of CD73, including cell recognition, cell-cell and/or cell-matrix interactions (Mkhikian et al., 2011), but also determines the solubility, stability and catalytic properties of the enzyme. Our data showed that MRS4598 was able to inhibit the activity of CD73 in astrocytes treated with TNF-α and IL −1β. The compound caused a concentration-dependent decrease in CD73 activity in cells treated with IL-1β. Considering that the IC₅₀ for inhibition of CD73 by MRS4598 was an order of magnitude higher in stimulated astrocytes than in control astrocytes, this result may suggest that CD73 variants expressed in reactive astrocytes as a result of altered N-glycosylation may have altered kinetic properties and reduced sensitivity to inhibiton.

Release of CD73 by cleavage of the GPI anchor is a pharmacologically and clinically relevant mechanism for the enzyme regulation (Geoghegan et al., 2016). The release may be triggered by interaction with ECM or signaling molecules that initiate enzymatic hydrolysis of the GPI anchor (Sadej et al., 2008; Terp et al., 2013). Specifically, phosphatidylinositol-specific phospholipase C (PI-PLC) is able to cleave the GPI anchor and release truncated CD73 molecule into the extracellular milieu (Lehto et al., 1993). In skeletal muscle cells, PI-PLC-mediated release of CD73 is induced by insulin (Klip et al., 1988). The physiological significance of CD73 shedding resides in the fact that the soluble enzyme exhibits catalytic activity (Lehto et al., 1998) and circulates in the bloodstream as potent auxiliary system for nucleotides degradation (Yegutkin, 2008; Laketa et al., 2015). CD73 functions as a docking molecule involved in cell adhesion and migration. In resting astrocytes, interactions with fibronectin keep astrocytes bound to the basal lamina. However, reactive astrocytes actively alter their microenvironment by releasing tenascin C, which promotes astrocyte migration (Wiese et al., 2012). Namely, tenascin C inhibits the catalytic activity of CD73 and induces its shedding ifrom membrane, which promotes astrocyte migration (Sadej et al., 2008). Previous studies have shown that ligands that reduce the exposure of CD73 on the membrane surface decrease cell proliferation and increase cell migration in vitro (Adzic and Nedeljkovic, 2018). Our study showed that MRS4598 induced the release of CD73, which contributed to increased astrocyte migration in vitro. The compound simultaneously altered cell shape and induced stellation, which may also contribute to astrocyte migration.

5. CONCLUSION

Most of the current data on CD73 inhibitors come from the field of cancer research This is the first study to demonstrate the efficacy of MRS4598, MRS4552, and MRS4602 in inhibiting CD73 in cultures of resting and reactive astrocytes in vitro, which imply the antiinflammatory potential of the cytosine-derived α,β-methylene diphosphonates in neuroinflammation. Novel CD73 inhibitors could be considered for the development of innovative therapies in neurodegenerative diseases.