Biphasic actions of the adenosine A2a receptor antagonist istradefylline in mice with obstructive uropathy

Abstract

Regardless of its underlying causes, progressive chronic kidney disease is characterized by progressive kidney fibrosis. The adenosine A2a receptor is expressed by activated kidney fibroblasts. Istradefylline is an adenosine A2a receptor antagonist that has received regulatory authority approval as add-on treatment of Parkinson’s disease. Here, we examined the potential of repurposing istradefylline for kidney disease by administering istradefylline to mice with kidney fibrosis caused by unilateral ureteral obstruction (UUO). C57BL/6N mice underwent sham or UUO surgery and were followed for 14 days. Mice were treated with vehicle or istradefylline, with istradefylline initiated either the day after surgery (early intervention) or seven days after surgery (late intervention). Early intervention with istradefylline increased mortality in UUO mice and upregulated the expression of some inflammatory genes. In contrast, late intervention with istradefylline attenuated upregulation of the profibrotic cytokine Il11 and of extracellular matrix proteins in UUO kidneys. Istradefylline similarly attenuated transforming growth factor-β1 induced upregulation of Il11, Col1a1 and Fn1 in NRK-49F fibroblasts. The adenosine A2a receptor has both anti-inflammatory and profibrotic effects. The therapeutic window by which adenosine A2a receptor antagonism may slow CKD progression is thus relatively narrow, even when employing an agent with immediate clinical applicability.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-24912-w.

Introduction

Drug repurposing saves time, cuts costs, and eases regulatory approval processes, offering an efficient process through which new treatments can be made available to patients. The history of biomedicine is littered with examples of successful drug repurposing that has saved millions of lives. Aspirin, for instance, was originally developed as an antipyretic and anti-inflammatory agent, but it is now routinely used to prevent heart attacks and strokes¹. Sodium-glucose cotransporter-2 (SGLT2) inhibitors are glucose-lowering therapies used in the treatment of Type 2 diabetes. However, owing to their glucose-independent effects, SGLT2 inhibitors now also have indications for the treatment of chronic kidney disease (CKD) and heart failure even in those without diabetes^2,3. Other drug repurposing opportunities are being keenly investigated by researchers from both the academic and industrial communities. Almost regardless of its underlying etiology, CKD is characterized by progressive kidney fibrosis⁴. In a search for new repurposing opportunities for kidney fibrosis, we recently performed a screen of G-protein coupled receptors (GPCRs) that are expressed by activated kidney fibroblasts⁵. Among the 56 GPCRs that we identified, the adenosine A2a receptor was enriched ~ 20-fold in single activated kidney fibroblasts isolated from mice with kidney fibrosis caused by unilateral ureteral obstruction (UUO)⁵. This was interesting to us because an orally active specific adenosine A2a receptor antagonist, istradefylline, has received regulatory authority approval as adjunctive therapy for the treatment of Parkinson’s disease⁶. Istradefylline is hepatically metabolized, and no adjustment of its dose is needed down to a GFR of 15 ml/min⁶. It has also recently been shown to protect against cisplatin-induced nephrotoxicity and peripheral neuropathy⁷. This led us to speculate that istradefylline could be repurposed for the treatment of kidney fibrosis.

The adenosine A2a receptor is one of four GPCRs (A1, A2a, A2b and A3) for the purine nucleoside adenosine^8–10. Adenosine itself is a ubiquitous extracellular signaling molecule with a short half-life that is produced by ectonucleotidase-mediated hydrolysis of ATP¹¹. Adenosine is normally present at low concentration in the pericellular space¹². However, local adenosine concentrations increase when tissue is damaged, and ATP is released into the extracellular space by injured or dying cells or by inflammatory cells¹². Prior research has linked the adenosine A2a receptor to several different kidney pathologies including ischemia reperfusion injury, diabetic kidney disease, glomerulonephritis and fibrosis^13–17. However, whereas adenosine A2a receptor antagonism has been reported to have antifibrotic actions in the liver¹⁸, dermis^19–22, and peritoneum²³, most (albeit not all⁷) studies have described a protective role for adenosine A2a receptors in kidney disease^13,15,16. This likely reflects the established, predominantly anti-inflammatory actions of the adenosine A2a receptor²⁴, which may be distinct from any role it may have in regulating kidney fibrosis.

Given the reported antifibrotic actions of adenosine A2a receptor antagonism in other organs, the enrichment of adenosine A2a receptors in activated kidney fibroblasts, and the existence of an orally active antagonist amenable to repurposing, in the present study we set out to test the effects of istradefylline in mice with kidney inflammation and fibrosis caused by UUO. In light of the reported multifaceted actions of the adenosine receptors, we compared the effects of administering istradefylline early after UUO, before fibrosis has occurred, and later after UUO, when kidney disease is established.

Results

Kidney Adora2a mRNA levels are increased in male and female mice 14 days after unilateral ureteral obstruction (UUO)

In our first experiments, we measured mRNA levels of Adora2a (the gene that encodes the adenosine A2a receptor) in the kidneys of male and female mice 14 days after sham or UUO surgery²⁵. As expected⁵, Adora2a levels were increased in the kidneys of mice with UUO (Fig. 1A). In the kidneys of sham-operated mice, Adora2a transcripts were evident in tubule cells, interstitial cells, glomerular cells, and vascular smooth muscle cells. Adora2a transcripts were more abundant in UUO kidneys, especially within accumulating interstitial cells (Fig. 1B), likely fibroblasts, immune cells (including macrophages, dendritic cells and T cells) and peritubular endothelial cells, each of these cell-types being known to express Adora2a^5,26,27.

Fig. 1.

Kidney Adora2a mRNA levels are increased in mice with unilateral ureteral obstruction (UUO). (A) qRT-PCR for Adora2a in male and female C57BL6N mice two weeks after sham or UUO surgery. n = 6/group. (B) RNAscope in situ hybridization for Adora2a in the kidneys of mice two weeks after sham or UUO surgery. Original magnification × 400. Scale bar = 50 µm. n = 4/group. In situ hybridization for the bacterial gene dapB in UUO kidneys is the negative control. n = 3. Values are mean ± S.D.. **P < 0.01, ***P < 0.001, ****P < 0.0001 by one-way ANOVA with Tukey’s post-hoc test. Values were log-transformed before statistical comparison.

Early intervention, but not late intervention, with istradefylline increases mortality in mice after unilateral ureteral obstruction (UUO)

C57BL/6N mice were subsequently randomized to undergo sham or UUO surgery and were then randomized to receive treatment with vehicle or istradefylline by daily gavage. The mice were followed for 14 days from surgery prior to kidney phenotyping. Mice randomized to receive treatment with istradefylline were subdivided into early intervention and late intervention arms. Those in the early intervention arm received treatment with istradefylline beginning the day after surgery. Mice in the late intervention arm received treatment with istradefylline beginning seven days after surgery. The study design is shown in Fig. 2A. Vehicle-treated mice were gavaged from the day after surgery, along with those in the early intervention with istradefylline arm. The animal numbers, body weights and weights of the obstructed (left) kidneys are shown in Table 1. Six of 15 mice in the UUO early intervention arm died before study end, whereas mortality was lower in the other treatment groups (Table 1). Survival curves are shown in Fig. 2B. Mortality was significantly increased in UUO mice treated with istradefylline as an early intervention in comparison to those receiving vehicle (P = 0.0007), whereas there were no mortalities in UUO mice when istradefylline was started seven days after surgery [late intervention (Table 1 and Fig. 2B)]. At the end of the study period, weights of the obstructed left kidneys were lower in the late intervention group than in UUO mice treated with vehicle or with early intervention istradefylline (Table 1). Representative Masson’s trichrome staining from each group is shown in Fig. 2C. Kidney tubule injury, determined semi-quantitively on H&E-stained kidney sections was expectedly elevated in UUO kidneys and was numerically (albeit non-significantly) lower with late (but not early) intervention with istradefylline (S.I. Fig. 1).

Fig. 2.

Early intervention with istradefylline increases mortality in mice after unilateral ureteral obstruction (UUO). (A) Study design. Mice were randomized to undergo sham or UUO surgery. Vehicle-treated mice received vehicle (40% DMSO, 30% cremaphor, and 30% mineral oil, diluted 1:6 in 2% sucrose water) by daily gavage beginning the day after surgery and were followed for 14 days from the day of surgery. Mice in the early intervention arm received istradefylline (3 mg/kg) in vehicle by daily gavage beginning the day after surgery and were followed for 14 days from the day of surgery. Mice in the late intervention arm received istradefylline (3 mg/kg) in vehicle by daily gavage beginning 7 days after surgery and were followed for 14 days from the day of surgery. (B) Survival curves. Note: no mortalities were observed in the sham + vehicle, UUO + vehicle, and UUO + late intervention istradefylline groups. Statistical comparison between groups determined by log-rank (Mantel-Cox) test. (C) Masson’s trichrome staining of kidney sections. n = 4 per group. Original magnification × 100. Scale bar = 100 µm.

Table 1.

Functional characteristics of C57BL/6N mice undergoing sham or unilateral ureteral obstruction (UUO) surgery followed for 14 days and treated with istradefylline (3 mg/kg) starting the day after surgery (early intervention) or 7 days after surgery (late intervention).

	Alive/dead	Body weight (g)	Left kidney weight (g)	Left kidney weight:body weight (%)
Sham + vehicle	15/0	23 ± 2	0.14 ± 0.03	0.60 ± 0.13
Sham + Istradefylline (early intervention)	12/2	24 ± 3	0.13 ± 0.02	0.55 ± 0.05
Sham + Istradefylline (late intervention)	12/1	25 ± 2	0.13 ± 0.02	0.54 ± 0.04
UUO + vehicle	13/0	24 ± 2	0.22 ± 0.02abc	0.94 ± 0.19abc
UUO + Istradefylline (early intervention)	9/6	24 ± 2	0.21 ± 0.02bcd	0.86 ± 0.11bcd
UUO + Istradefylline (late intervention)	15/0	23 ± 2	0.18 ± 0.04efgh	0.80 ± 0.18cij

Values are mean ± S.D.. ^aP < 0.0001 versus sham + vehicle, ^bP < 0.0001 versus sham + early intervention, ^cP < 0.0001 versus sham + late intervention, ^dP < 0.001 versus sham + vehicle, ^eP < 0.05 versus sham + vehicle, ^fP < 0.01 versus sham + early intervention, ^gP < 0.01 versus sham + late intervention, ^hP < 0.05 versus UUO + vehicle, ⁱP < 0.01 versus sham + vehicle, ^jP < 0.001 versus sham + early intervention.

RNA sequencing of the kidneys of UUO mice treated with istradefylline as either an early intervention or late intervention

Next, we performed unbiased RNA sequencing to determine what kidney gene expression changes occur with istradefylline treatment in UUO mice. Volcano plots are shown in Fig. 3A–C. Using a fold change cut off of 1.5 and an unadjusted p_value ≤ 0.05, 4400 genes were upregulated, and 1372 genes were downregulated in a comparison between the kidneys of vehicle-treated UUO and sham-operated mice (Fig. 3A). Table 2 shows the top 10 upregulated and downregulated genes between vehicle-treated UUO mice and sham-operated mice. The top upregulated gene was Havcr1, which encodes kidney injury molecule-1 (KIM-1) and which was increased 176-fold in UUO kidneys (q_value 6.76964E-05). The top downregulated gene in UUO kidneys was Pvalb (fold change 0.03, q_value 2.73241E-05) which is a marker of distal convoluted tubules and connecting tubules. In comparison, few genes were differentially expressed with either early or late intervention with istradefylline in UUO kidneys (Fig. 3B,C). Table 3 shows the differentially expressed genes in the early and late intervention arms of the study in comparison to vehicle-treated mice for both sham-operated and UUO kidneys, after controlling for multiple groups comparisons. Thirty seven genes were differentially downregulated in the comparison of late intervention istradefylline and vehicle in sham-operated mice. In contrast, q_values achieved < 0.05 for no more than 3 genes in the other comparisons. Interestingly, the two downregulated genes in the comparison of early intervention istradefylline with vehicle in UUO mice have both been linked to antioxidant defenses. The most downregulated gene in any of the comparisons was Gpx1 (fold change 0.19, q_value 0.034193684) which encodes the antioxidant protein glutathione peroxidase 1. The other downregulated gene was Fbxl17 (fold change 0.48, q_value 0.034193684) which encodes F-box and leucine rich repeat protein 17 (FBXL17) which is a regulator of the nuclear factor (erythroid-derived 2)-like 2 (NRF2) antioxidant pathway²⁸. Accordingly, given the increase in mortality and downregulation of antioxidant genes in UUO mice with early intervention with istradefylline, we speculated that istradefylline could, under some circumstances, augment inflammation. To determine whether this is the case, we isolated and differentiated bone marrow-derived macrophages (BMDMs) from mice and preincubated them with istradefylline prior to stimulation with lipopolysaccharide (LPS). Istradefylline pre-treatment caused a marked upregulation in induction of the inflammatory genes Tnfa and C3 induced by LPS in BMDMs (Fig. 3D,E).

Fig. 3.

Volcano plots of differentially expressed genes determined by RNA sequencing of the kidneys of sham-operated mice and mice with unilateral ureteral obstruction (UUO) and treated with istradefylline (3 mg/kg) beginning on day 1 (early intervention) or day 7 (late intervention) after surgery and followed for 14 days (A–C); and istradefylline augments inflammatory gene upregulation in bone marrow-derived macrophages (BMDMs) (D,E). (A) Comparison of differentially expression genes in UUO + vehicle versus sham + vehicle. (B) Comparison of differentially expression genes in UUO + early intervention versus UUO + vehicle. (C) Comparison of differentially expression genes in UUO + late intervention versus UUO + vehicle. Fold change cut off 1.5, p_value ≤ 0.05. (D,E) qRT-PCR for Tnfa (D) and C3 (E) in BMDMs pretreated in the presence or absence of 10 µmol/L istradefylline for 2 h before the addition of 1 ng/mL lipopolysaccharide (LPS) or vehicle for a further 16 h. n = 6/condition. Values are mean ± S.D.. *P < 0.05, **P < 0.01, ****P < 0.0001 by one-way ANOVA followed by Tukey’s post-test.

Table 2.

Top 10 differentially expressed genes determined by RNA sequencing in the kidneys of vehicle-treated sham-operated mice and mice with unilateral ureteral obstruction (UUO).

Gene	Fold Change	p_value	q_value
UUO vehicle versus sham vehicle up
Havcr1	175.9858618	2.03017E-07	6.76964E-05
Timp1	77.68942052	1.266E-08	2.73241E-05
Sprr2f	66.97956639	1.09677E-07	5.2527E-05
Ccl8	49.17743175	7.17229E-06	0.000217807
Lyz2	44.26348946	1.97266E-06	0.000139594
Lcn2	43.93366224	3.55716E-05	0.000490628
Mmp7	40.56220919	7.04784E-06	0.000216737
Ubd	37.74631103	2.07873E-06	0.000139594
C3	36.4791501	0.000200585	0.001333205
Col1a1	34.11794127	1.70988E-06	0.000135248
UUO vehicle versus sham vehicle down
Pvalb	0.029063938	7.48045E-09	2.73241E-05
Nccrp1	0.049465858	1.28639E-05	0.000297233
Egf	0.051092462	6.9134E-06	0.000214542
Cyp2d9	0.072064201	0.000244166	0.001503907
Gpx6	0.076771801	5.2947E-05	0.000605526
Aqp2	0.085032585	0.000156238	0.001136862
Cyp2d12	0.087668792	0.003267891	0.009124647
Calb1	0.095470392	0.001020748	0.003893223
Dnase1	0.096217864	0.000176714	0.00123063
Mep1b	0.097054865	6.50764E-05	0.000676875

q_value = false discovery rate adjusted p_value.

Table 3.

Differentially expressed genes in the kidneys of sham-operated mice and mice with unilateral ureteral obstruction (UUO) treated with vehicle or istradefylline (3 mg/kg) starting the day after surgery (early intervention) or 7 days after surgery (late intervention).

Gene	Fold Change	p_value	q_value
Sham early intervention versus sham vehicle up
Number of differentially expressed genes (q_value < 0.05) 3
Golm1	2.626301319	4.46448E-06	0.02840038
Micu3	2.134663295	1.56608E-05	0.038870074
Atp4a	1.807190743	9.53545E-06	0.02840038
Sham early intervention versus sham vehicle down
Number of differentially expressed genes (q_value < 0.05) 2
Zfp422	0.227680852	4.12124E-06	0.02840038
Dctn5	0.403431848	8.06796E-06	0.02840038
Sham late intervention versus sham vehicle up
Number of differentially expressed genes (q_value < 0.05) 1
Golm1	2.329469711	4.70251E-06	0.029123823
Sham late intervention versus sham vehicle down
Number of differentially expressed genes (q_value < 0.05) 37
Zfp708	0.388987191	2.04917E-05	0.029123823
Ccl21a	0.514703545	1.19469E-05	0.029123823
Ankrd23	0.554907903	2.88833E-05	0.029123823
Fnbp4	0.587466448	1.77802E-05	0.029123823
Appl2	0.622215377	2.02664E-05	0.029123823
UUO early intervention versus UUO vehicle up
Number of differentially expressed genes (q_value) < 0.05) 1
Dsel	4.308487395	1.27384E-06	0.010811739
UUO early intervention versus UUO vehicle down
Number of differentially expressed genes (q_value < 0.05) 2
Gpx1	0.192826966	8.05742E-06	0.034193684
Fbxl17	0.4878973	7.71483E-06	0.034193684
UUO late intervention versus UUO vehicle up
Number of differentially expressed genes (q_value < 0.05) 1
Micu3	2.350138078	3.42013E-06	0.018225115
UUO late intervention versus UUO vehicle down
Number of differentially expressed genes (q_value < 0.05) 3
Zfp317	0.345443993	4.29939E-06	0.018225115
Syvn1	0.493209925	3.12117E-06	0.018225115
Swi5	0.627195133	1.25826E-05	0.04266996

NB. For sham late intervention versus sham vehicle down, the top 5 of 37 differentially expressed genes are shown. q_value = false discovery rate adjusted p_value.

Late intervention with istradefylline attenuates kidney fibrotic gene upregulation in UUO mice

We next examined the expression of injury markers and fibroinflammatory genes in the kidneys of the six groups of mice by quantitative PCR. Istradefylline had no effect on the upregulation of Havcr1, Lcn2, Il6 or Ccl2 in UUO kidneys (Fig. 4A–D). mRNA levels of the inflammatory genes C3, Il1b and Ccl8 were higher in UUO kidneys of mice treated with early intervention with istradefylline than those receiving istadefylline as a late intervention (Fig. 4E–G). Similarly, mRNA levels of Adora2a itself were increased in UUO kidneys with early (but not late) intervention with istradefylline (S.I. Fig. 2). In contrast, the fibrotic genes Timp1, Tgfb1, Col1a1, Col3a1 and Fn1 were all reduced in UUO mice when istradefylline was administered as a late intervention (Fig. 4H–L).

Fig. 4.

Late intervention with istradefylline attenuates fibroinflammatory gene upregulation in mice with unilateral ureteral obstruction (UUO). Sham-operated and UUO mice were treated with vehicle or istradefylline (3 mg/kg) beginning on day 1 (early intervention (early int.)) or day 7 (late intervention (late int.)) after surgery and were followed for 14 days. (A–L) qRT-PCR for (A) Havcr1, (B) Lcn2, (C) Il6, (D) Ccl2, (E) C3, (F) Il1b, (G) Ccl8, (H) Timp1, (I) Tgbf1, (J) Col1a1, (K) Col3a1 and (L) Fn1. Sham + vehicle n = 12; sham + early intervention istradefylline n = 10; sham + late intervention istradefylline n = 12; UUO + vehicle n = 12; UUO + early intervention istradefylline n = 9; UUO + late intervention istradefylline n = 12; except Havcr1 and Ccl2 sham + vehicle, UUO + vehicle and UUO + late intervention istadefylline n = 13, sham + early intervention istradefylline n = 12; and Tgfb1 UUO + vehicle n = 11. Values are mean ± S.D.. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by one-way ANOVA followed by Tukey’s post-test. Data were log-transformed before statistical comparison.

Late intervention with istradefylline attenuates Il11 upregulation in UUO mice

Having observed an attenuation in fibrotic gene mRNA levels in the kidneys of UUO mice with late (but not early) intervention, we set out to determine whether late intervention with istradefylline does reduce kidney fibrosis. Type 1 collagen protein levels were numerically reduced (albeit non-significantly) and fibronectin levels were significantly reduced in UUO kidneys with late intervention with istradefylline than in vehicle-treated UUO mice (Fig. 5A,B). In considering mechanisms by which late intervention with istradefylline may affect fibrotic gene upregulation in UUO mice, we went back to our RNA sequencing dataset and we observed that mRNA levels of the profibrotic cytokine, Il11 tended to be lower in UUO kidneys following late intervention with istradefylline in comparison to the kidneys from vehicle-treated UUO mice (fold change 0.66, unadjusted p_value 0.014073781), which was confirmed by qRT-PCR (Fig. 5C). Cognizant that IL-11 mediates matrix protein production through STAT3 signaling²⁹, we immunoblotted mouse kidneys observing an increase in levels of phosphorylated STAT3 in UUO kidneys that was attenuated by late intervention with istradefylline (Fig. 5D). To determine whether adenosine A2a receptor signaling does mediate fibroblast IL-11 and fibrotic gene expression, we stimulated NRK-49F cells with transforming growth factor-β1 (TGF-β1) in the presence or absence of istradefylline, observing that the increases in Il11, Col1a1 and Fn1 mRNA levels induced by TGF-β1 were each attenuated with istradefylline (Fig. 5E–G).

Fig. 5.

Late intervention with istradefylline attenuates Il11 upregulation and kidney fibrosis in mice with unilateral ureteral obstruction (UUO). Sham-operated mice and mice with UUO were treated with vehicle beginning the day after surgery or istradefylline (3 mg/kg) beginning on day 7 (late intervention (late int.)) after surgery and were followed for 14 days from the time of surgery. (A) Immunostaining for type 1 collagen and quantitation of proportional staining in sham + vehicle n = 12, sham + late intervention istradefylline n = 12, UUO + vehicle n = 13 and UUO + late intervention istradefylline n = 15. Original magnification × 100. Scale bar = 100 µm. (B) Immunostaining for fibronectin and quantitation of fibronectin immunopositivity in sham + vehicle n = 13, sham + late intervention istradefylline n = 12, UUO + vehicle n = 13 and UUO + late intervention istradefylline n = 15. Original magnification × 100. Scale bar = 100 µm. In A and B the IgG controls are from the UUO + vehicle group n = 4. (C) qRT-PCR for Il11 in kidneys of sham-operated mice and mice with UUO followed for 14 days and treated with vehicle or istradefylline (3 mg/kg) as a late intervention (late int.) beginning seven days after surgery and followed for 14 days from the time of surgery. n = 12/group. (D) Immunoblotting for phosphorylated STAT3 and total STAT3. n = 6/group. (E–G) qRT-PCR for Il11 (E), Col1a1 (F) and Fn1 (G) in NRK-49F cells treated with 10 ng/mL TGF-β1 (or vehicle) for 2 h, before 10 ng/mL TGF-β1 (or vehicle) +/− 10 µmol/L istradefylline for a further 24 h. n = 6/condition except (E) istradefylline n = 5 and (G) TGF-β1 n = 5, Outliers ROUT Q = 1. Values are mean ± S.D.. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by one-way ANOVA followed by Tukey’s post-test test (A, C, D, E, F and G) or Fisher’s least significant difference test (B). Data in (A and B) were log-transformed before statistical comparison.

Discussion

Here we report the effects of treating mice with obstructive uropathy with the specific adenosine A2a receptor antagonist, istradefylline. We observed that initiation of istradefylline the day after UUO surgery decreased mouse survival, whereas delaying administration of istradefylline to seven days after surgery resulted in improved tolerability and had a marginal, albeit non-negligible, antifibrotic effect. This biphasic response highlights the intricate biological actions of the adenosine A2a receptor and the potentially narrow therapeutic window in which the adenosine A2a receptor may be targeted for the treatment of chronic diseases characterized by inflammation or fibrosis.

The design of our study was informed by our observations of Adora2a enrichment in activated kidney fibroblasts, of a small but statistically significant upregulation in mRNA levels of the profibrotic growth factor Ccn2 with adenosine A2a receptor agonism, and of an antifibrotic effect of treatment of UUO mice with caffeine, which is a non-specific adenosine receptor antagonist⁵. All four adenosine receptor subtypes are expressed in the kidney³⁰, and several studies have explored the effects of adenosine receptor agonism, antagonism or knockout in different models of kidney disease (reviewed in³¹). Our value-add is the testing of a compound that has already received regulatory authority approval for another indication, representing a legitimate repurposing opportunity. Our findings are generally aligned with the prior literature which points to a predominantly protective effect of adenosine signaling through the adenosine A2a receptor (especially in the acute setting), with some lesser profibrotic effects of persistent adenosine A2a receptor signaling.

Although it was approved by the U.S. Food and Drug Administration (FDA) in 2019 for use as adjunctive therapy to levodopa/carbidopa in adult patients with Parkinson’s disease experiencing “OFF” episodes⁶, istradefylline (also known as KW-6002) is not a new compound. Its clinical program was initiated around 1996³². This long interval from discovery to FDA approval illustrates the reality and complexity of drug development, trial design and end-point determination^33,34. Istradefylline is highly selective for the adenosine A2a receptor (K_i 2.2 nM, 70-fold greater than for the adenosine A₁ receptor³⁵). Although adenosine A2a receptors are expressed by many cells and organs, their expression is highest in immune cells and in the striatopallidal system in the brain²⁴. That istradefylline has found a clinical niche for a disease that affects the striatopallidal system is illustrative that adenosine A2a receptor antagonism is most likely to induce biological effects caused by altering the actions of cells most enriched for the adenosine A2a receptor. These would thus be diseases of the striatopallidal system, and diseases characterized by immune dysregulation.

Adenosine A2a receptors are densely expressed by almost all immune cells, and the adenosine A2a receptor is the major immunoregulator of the adenosine-adenosine receptor system, mediating generally anti-inflammatory effects²⁴. In obstructive uropathy, following surgical ligation of the kidneys, there is a transient increase in kidney blood flow that is followed by vasoconstriction and an increase in vascular resistance within hours³⁶. Immune cell infiltration peaks within 2–3 days, and myofibroblasts accumulate within the interstitium inducing tissue fibrosis from around seven days³⁷. In the kidneys, adenosine A2a receptors mediate vasodilatation³⁸, and adenosine A1 receptors mediate afferent arteriolar vasoconstriction³⁹. We started early intervention istradefylline treatment one day after sham or UUO surgery and observed mortalities in UUO mice between days four and 11. RNA sequencing of kidneys of early intervention UUO mice that survived to the end of the study identified Gpx1 and Fbxl17 as being differentially downregulated with istradefylline. Both of these genes encode proteins linked to antioxidant pathways^28,40. Furthermore, we observed that istradefylline pretreatment augmented LPS-induced upregulation of the mediators of inflammation Tnfa and C3 in BMDMs. Accordingly, we speculate that the excess mortalities that we observed when adenosine A2a receptor antagonism was initiated early after UUO surgery was because of a potentiation of a heightened pro-oxidant, inflammatory response to UUO. This would be aligned with the anti-inflammatory effects of adenosine A2a receptor agonism or adenosine kinase inhibition reported in other models of kidney disease^{15,16,41–44}.

Whereas we observed diminished mouse survival when adenosine A2a receptor antagonism was initiated one day after surgery, we did not observe any mortalities when istradefylline treatment was delayed to seven days after UUO surgery. Furthermore, we observed a significant reduction in kidney fibronectin upregulation and of Il11 mRNA levels in UUO mice treated with istradefylline. We similarly observed that istradefylline decreased Il11 induction by TGF-β1 in cultured NRK-49F cells. This is interesting because IL-11 has recently emerged as a pivotal player in organ fibrosis⁴⁵, and IL-11 antagonists are being developed for the treatment of kidney fibrosis⁴⁶. Adenosine A2a receptor intracellular signaling is well-characterized. Upon activation, the adenosine A2a receptor couples to the Gs protein and signals through adenylyl cyclase, protein kinase A (PKA) and CREB. PKA can augment TGF-β signaling⁴⁷, and correspondingly PKA inhibition has been reported to inhibit fibronectin induction by TGF-β⁴⁸. TGF-β1 is also a potent inducer of IL-11⁴⁵. IL-11, in turn, mediates its profibrotic effects by signaling through STAT3²⁹. In our studies, we observed a diminution in Il11, STAT3 phosphorylation and fibronectin upregulation in the kidneys of UUO mice treated with istradefylline as a late intervention. Thus, signaling through the adenosine A2a receptor may potentiate kidney fibrosis through the induction of IL-11 by kidney fibroblasts. This effect would need to be balanced against the established anti-inflammatory effects of adenosine-adenosine A2a receptor signaling, especially in the acute setting.

This study has several limitations that warrant particular emphasis. Firstly, we chose to administer istradefylline orally because we wished to model the scenario in which the drug may be repurposed for kidney disease in patients. We selected a dose of istradefylline of 3 mg/kg/day because this oral dosing has been used previously to attenuate choroidal neovascularization⁴⁹, and it also prevented long-term episodic memory impairment⁵⁰ in mice. However, in toxicology studies, mice have been dosed with istradefylline up to 250 mg/kg/day, which results in plasma exposures ~ 20 fold higher than the maximum dose of istradefylline recommended in humans (40 mg/day)⁶. The comparatively low dosing regimen we selected may be responsible for the relatively small changes we observed in differentially expressed genes in our RNA sequencing. However, the increase in mouse mortality in the early intervention arm of the study, and the consistent changes we observed in fibrotic gene expression and Il11 in the late intervention arm, illustrate that the drug did have biological effects in mice at this dosage. It is possible that had we used a higher dose of istradefylline, or studied the mice at a different timepoint, we would have unmasked a larger effect on fibrosis. However, this could have similarly resulted in a larger effect on mouse survival when administered early after UUO. This illustrates the narrow therapeutic window available for adenosine A2a receptor antagonism in the treatment of kidney disease, that is likely to be both time- and dose-dependent. Secondly, whereas we have inferred that the increase in mouse mortality with early intervention with istradefylline may be due to an augmentation in the early inflammatory response to UUO, this has not been proven. Furthermore, there were three mortalities among sham-operated mice treated with istradefylline either as an early- or late- intervention. The causes of these mortalities are unclear and it is possible that the mortalities we observed in sham-operated mice may be a delayed response to gavaging accident. Istradefylline has been shown to have a wide safety window in preclinical toxicology studies with no evidence of carcinogenesis, mutagenesis or brain mineralization⁶. The most common side effects of istradefylline in humans are dyskinesia, dizziness, constipation, nausea, hallucinations and insomnia⁶. Lastly, it has previously been proposed that of the two Gs coupled adenosine receptors (A2a and A2b), adenosine A2a receptor primarily exerts anti-inflammatory effects in the kidney whereas the adenosine A2b receptor is the principal mediator of kidney fibrosis. The adenosine A2b receptor has lower affinity for adenosine (EC₅₀ 24 µM vs. 0.7 µM for A2a) with levels of adenosine > 1 µM only likely achieved under pathophysiological conditions²⁴. Indeed, global knockout of adenosine A2b receptor has been reported to attenuate kidney fibrosis in UUO mice⁵¹. Despite its limitations though, this study has merits, most notably the testing of a compound with a proven safety and tolerability profile that was recently proposed as a repurposing opportunity for cisplatin-nephrotoxicity⁷.

In summary, although adenosine A2a receptor antagonism has been reported to have antifibrotic effects in other disease settings^18,23,52, and it is expressed by kidney fibroblasts and regulates profibrotic pathways, its therapeutic targeting in vivo is challenging. Inflammatory and fibrotic processes are intimately intertwined in chronic kidney diseases; the former begetting the latter. The pathway to regulatory approval for istradefylline as adjunctive treatment for Parkinson’s disease was long and complicated. Our study in a mouse model of kidney disease similarly illustrates the intricacies involved in identifying repurposing opportunities that involve adenosine receptor signaling.

Methods

Animal studies

Kidney Adora2a mRNA levels were measured by qRT-PCR in male and female C57BL/6N mice 14 days after sham or UUO surgery. The details of these mice have been reported before²⁵. Localization of Adora2a mRNA transcripts was performed in kidney sections from C57BL/6N mice that had undergone sham or UUO surgery and had been followed for 14 days, For studies of istradefylline, sham and UUO surgeries were performed in male C57BL/6N mice (C57BL/6N/Crl; Charles River Laboratories, Senneville, Quebec, Canada) aged ~ 7–8 weeks as we have previously reported⁵³. Mice were anesthetized with 2% isoflurane and an incision was made in the left flank before occlusion of the left ureter distal to its origin using two 5–0 silk sutures. Sham-operated mice underwent the same procedure without ligation of the left ureter. Analgesia was achieved with slow-release buprenorphine (0.5 mg/kg subcutaneously) administered pre-operatively. Sham and UUO mice were randomized to daily gavage with istradefylline (3 mg/kg; catalog no. HY-10888, lot no. 58314, MedChemExpress, Monmouth Junction, NJ) or vehicle made up of in 40% DMSO, 30% cremaphor, and 30% mineral oil, diluted 1:6 in 2% sucrose water⁵⁴. In the early intervention arm of the study, sham and UUO mice began treatment one day after surgery, and in the late intervention arm of the study, treatment was started seven days after surgery. The mice were followed for 14 days. The study was conducted in accordance with ARRIVE guidelines. All experimental procedures adhered to the guidelines of the Canadian Council of Animal Care and were approved by the St. Michael’s Hospital Animal Care Committee (ACC257).

RNA sequencing

RNA was isolated from kidney homogenates (n = 4/group) using TRIzol Reagent (Life Technologies, Thermo Fisher Scientific, Waltham, MA). RNA sequencing was performed using the 6G RNA Sequencing Service (150 bp paired-end, 40 million reads) from ArrayStar (Rockville, MD), as previously described⁵⁵. Briefly, after quantitation of RNA using a Nanodrop ND-1000, RNA was enriched using oligo (dT) magnetic beads and sequencing libraries were prepared using a KAPA Stranded RNA-Seq Library prep Kit (Illumina, San Diego, CA). Sequencing was performed on an Illumina Novaseq 6000 (150 cycles for both ends). Solexa pipeline v1.8 was used for image analysis. Sequence quality was assessed using FastQC. Hisat2 software was used to align trimmed reads (trimmed 5′, 3′-adaptor bases using cutadapt) to the GRCm38 reference genome⁵⁶. Transcript abundances were estimated using StringTie⁵⁷, and fragments per kilobase of exon per million mapped fragments (FPKM) and differential gene expression were determined with Ballgown⁵⁸. Volcano plots were generated with the differentially expressed genes in R, Python or shell environment. Data are deposited to Gene Expression Omnibus (accession number GSE299417).

Bone marrow-derived macrophages (BMDMs)

BMDMs were isolated and differentiated as we have previously reported⁵⁹ and based on the protocol previously described by Toda et al.⁶⁰. In brief, bone marrow cells collected from the femurs and tibiae of C57BL/6 J mice were allowed to differentiate in phenol red–free (high-glucose) DMEM containing 10% FBS, 10 ng/mL M-CSF (M9170, MilliporeSigma, Oakville, Ontario, Canada), and 1% penicillin–streptomycin for seven days. BMDMs were treated with 10 µmol/L istradefylline for 2 h before the addition of 1 ng/mL LPS (MilliporeSigma) for a further 16 h prior to qRT-PCR.

qRT-PCR

RNA was isolated from mouse whole kidney tissue or NRK-49F cells using TRIzol reagent (Thermo Fisher Scientific), and cDNA was reverse transcribed using a High-Capacity cDNA Reverse Transcription Kit (catalog no. 4368814, lot no. 2959983, Thermo Fisher Scientific). Custom-designed primers were from Integrated DNA technologies (Coralville, IA) and had the following sequences: mouse Adora2a forward 5′-TACATCGCCATCCGAATTCCA-3′, reverse 5′-GAATGACAGCACCCAGCAAA-3′; mouse Tnfa forward 5′-TGATCGGTCCCCAAAGGGAT-3′, reverse 5′-TGTCTTTGAGATCCATGCCGT-3′; mouse C3 forward 5′-CATATGCTCCAGCACTGAGAAC-3′, reverse 5′-TGCCTCTTTAGGAAGTCTTG-3′; mouse Havcr1 forward 5′-GGAAGGCAACCACGCTTAGAGA-3′, reverse 5′-AGGGAAGCCGCAGAAAAACCCTA-3; mouse Lcn2 forward 5′-AAGTCACCTCCATCCTGGTCA-3′, reverse 5′-GCGAACTGGTTGTAGTCCGTGGT-3′; mouse Ccl2 forward 5′-GACCCGTAAATCTGAAGCTAA-3′, reverse, 5′-CACACTGGTCACTCCTACAGAA-3′; mouse Il1b forward 5′-CAGGATGAGGACATGAGCACC-3′, reverse 5′_CTCTGCAGACTCAAACTCCAC-3′; mouse Ccl8 forward 5′-CCAGATAAGGCTCCAGTCAC, reverse, 5′-AGAGAGACATACCCTGCTTG-3′; mouse Timp1 forward 5′-CGAGACCACCTTATACCAGCG-3′, reverse 5′-GGCGTACCGGATATCTGCG-3′; mouse Tgfb1 forward 5′-GCTGCGCTTGCAGAGATTAA-3′, reverse 5′-GTAACGCCAGGAATTGTTGCTA-3′; mouse Col1a1 forward 5′-TTCAGGGAATGCCTGGTGAA-3′, reverse 5′-ACCTTTGGGACCAGCATCA-3′; mouse Col3a1 forward 5′-TGATGTCAAGTCTGGAGTGG-3′, reverse 5′-TCCTGACTCTCCATCCTTTC-3′; mouse Fn1 forward 5′-ATGTGGACCCCTCCTGATAGT-3′, reverse 5′-GCCCAGTGATTTCAGCAAAGG-3′; mouse Il11 forward 5′-CTGCACAGATGAGAGACAAATTCC-3′, reverse 5′-GAAGCTGCAAAGATCCCAATG-3′; mouse Rpl13a forward 5′-GCTCTCAAGGTTGTTCGGCTGA-3′, reverse 5′-AGATCTGCTTCTTCTTCCGATA-3′; rat Il11 forward 5′-GGGACACTGGGATCTTTGCA-3′, reverse 5′-GGCGAGACATCAAGAGCTGT-3′; rat Col1a1 forward 5′-AAGTCTCAAGATGGTGGCCG-3′, reverse 5′-TCTCCGCTCTTCCAGTCAGA-3′; rat Fn1 forward 5 ‘-GGCCACTTCCGAATCTGTCA-3′, reverse 5′-GCTCATCTCCTTCCTCGCTC-3′; rat Gapdh forward 5′- GAACGGGAAGCTCACTGG-3′, reverse, 5′-GCCTGCTTCACCACCTTCT-3′. Mouse Il6 primers were from OriGene (catalog no. MP206798, lot no. PM18AP2502, OriGene, Technologies Inc., Rockville, MD). SYBR green-based quantitative RT-PCR (qRT-PCR) was performed on a QuantStudio 7 Flex Real-Time PCR System (Thermo Fisher Scientific). Data analysis was performed using the comparative ΔΔC_T method.

RNAscope in situ hybridization

RNAscope in situ hybridization (Advanced Cell Diagnostics, Hayward, CA) was performed according to the manufacturer’s instructions as previously described^61,62 with probesets for Adora2a (catalog no. 409431, lot no. 21315A) and the bacterial gene dapB as a negative control (catalog no. 310043, lot no. 2017575). Hybridization signals were detected using Fast Red and RNA staining was identified as red puncta on light microscopy.

Histology

Masson’s trichrome staining was performed by members of the Pathology Research Program at Toronto General Hospital (Toronto, Ontario, Canada). Kidney tubule injury was assessed by calculating a semi-quantitative tubule injury score on H&E-stained kidney sections in 10 nonoverlapping fields (magnification × 400) as previously described⁶³.

Immunohistochemistry

Immunohistochemistry was performed with an anti-collagen 1 antibody at 1:100 dilution (catalog no. 1310-01, lot no. B2918-TD09, SouthernBiotech, Birmingham, AL) and an anti-fibronectin antibody used at 1:800 dilution (NB110-56990, lot no. C112307, Novus Biologicals, Centennial, CO). Quantitation of immunostaining was performed on digitized images (Axio Scan.z1) using HALO® (Indica Labs, Albuquerque, MA). Equal concentrations of goat IgG (catalog no. ab37373, lot no. GR296103-1) and rabbit IgG (catalog no. ab172730, lot no. GR3235749-21) served as negative controls.

Immunoblotting

Immunoblotting was performed using the following antibodies: anti-phospho-STAT3 (Tyr705) 1:1000 dilution (#44-380G, Thermo Fisher Scientific), anti-STAT3 1:1000 dilution (#9139 (124H6), Cell Signaling Technology], and anti-GAPDH 1:1000 dilution (#2188, clone 14C10, lot no. 14, Cell Signaling Technology). Densitometry was performed using ImageJ version 1.53.

NRK-49F cells

NRK-49F renal fibroblasts (CRL-1570, ATCC, Manassas, VA) were treated with 10 ng/mL TGF-β1 (catalog no. CA59, lot no. 0332259MC16, Novoprotein Scientific Inc., Summit, NJ)⁶⁴ for 2 h, before 10 ng/mL TGF-β1 with or without 10 µmol/L istradefylline for a further 24 h⁶⁵.

Statistics

Data are expressed as mean ± S.D.. Animals were randomly allocated to the study groups. Analyses of data were performed in a masked manner where feasible. Statistical analyses were performed using GraphPad Prism 10 for macOS (GraphPad Software Inc., San Diego, CA). Prior to statistical comparison, data were tested for normality by Shapiro–Wilk test. Normally distributed data were compared by Student t test (two groups) or one-way ANOVA with Tukey’s post-test (for multiple groups comparisons) unless otherwise stated. Data that were not normally distributed were compared by Mann–Whitney test (two groups) or were log-transformed prior to statistical comparison by one-way ANOVA, or compared by Kruskal–Wallis test followed by Dunn’s post-test as stated. For multiple comparisons, post hoc testing was only conducted if F in ANOVA achieved P < 0.05. All analyses were two-tailed. P < 0.05 was considered statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceived and designed the experiments: S.N.B., D.A,Y., A.A.. Performed experiments: E.S.H.Y., D.T.T., S.L.A., Y.L., H.K., R.A.. Analyzed the data: E.S.H.Y., S.N.B., H.K., A.A.. Wrote/edited the paper: E.S.H.Y., A.A.. Contributed materials, reagents and instruments: D.A.Y., C.T.L., A.A.. All authors reviewed the manuscript.

Data availability

RNA sequencing data have been deposited with Gene Expression Omnibus (accession number GSE299417). Other data are available from the corresponding author upon reasonable request.

Declarations

Competing interests

A.A. has received research support through his institution from Boehringer Ingelheim. D.A.Y. is a scientific cofounder and consultant for Fibrocor Therapeutics. Other authors declare no competing interests.