Inhibition of asparagine synthetase effectively retards polycystic kidney disease progression

Abstract

Polycystic kidney disease (PKD) is a genetic disorder characterized by bilateral cyst formation. We showed that PKD cells and kidneys display metabolic alterations, including the Warburg effect and glutaminolysis, sustained in vitro by the enzyme asparagine synthetase (ASNS). Here, we used antisense oligonucleotides (ASO) against Asns in orthologous and slowly progressive PKD murine models and show that treatment leads to a drastic reduction of total kidney volume (measured by MRI) and a prominent rescue of renal function in the mouse. Mechanistically, the upregulation of an ATF4-ASNS axis in PKD is driven by the amino acid response (AAR) branch of the integrated stress response (ISR). Metabolic profiling of PKD or control kidneys treated with Asns-ASO or Scr-ASO revealed major changes in the mutants, several of which are rescued by Asns silencing in vivo. Indeed, ASNS drives glutamine-dependent de novo pyrimidine synthesis and proliferation in cystic epithelia. Notably, while several metabolic pathways were completely corrected by Asns-ASO, glycolysis was only partially restored. Accordingly, combining the glycolytic inhibitor 2DG with Asns-ASO further improved efficacy. Our studies identify a new therapeutic target and novel metabolic vulnerabilities in PKD.

Keywords: ADPKD; Antisense Oligonucleotides; Glutamine Metabolism; Glycolysis; Metabolic Reprogramming.

Figure 1. ASNS is upregulated in PKD models and in human ADPKD.

(A) Asns mRNA expression in Pkd1^−/− and control MEF, cultured in full medium or under glucose deprivation. Representative of n = 3 independent experiments. (B) ASNS expression in Pkd1^-/- and control MEF, cultured in full medium or under glucose deprivation. Representative of n = 3 independent experiments. (C) ASNS expression in Pkd1^−/− and control mCCD, cultured in 0% FBS medium under glucose deprivation for 24 h (low and high confluency). Representative of n = 3 independent experiments. (D) ASNS expression in KspCre;Pkd1^ΔC/flox (cystic) and relative control kidneys at P4. c cytoplasmatic fraction, n nuclear fraction. (E) Asns expression in P10 Pkd1^v/v mice microarray. (F) Asns expression in P12 Pkd1^RC/RC mice RNA-seq. (G) ASNS expression in ADPKD human samples microarray. (H) Dot plot of snRNA-seq dataset showing ASNS expression in clusters identified in human ADPKD cystic and normal kidney tissues. PT proximal tubule, FR-PTC failed-repair proximal tubular cells, PEC parietal epithelial cells, TAL thick ascending limb of Henle’s loop, DCT distal convoluted tubule, CNT_PC connecting tubule and principal cells, ICA Type A intercalated cells, ICB Type B intercalated cells, PODO podocytes, ENDO endothelial cells, FIB fibroblasts, LEUK leukocytes. Data information: in (A) data are shown as mean ± SD. One-way ANOVA. *P < 0.05; **P < 0.01; ***P < 0.001. Source data are available online for this figure.

Figure 2. Targeting ASNS ameliorates PKD phenotype.

(A) Experimental design of Asns-ASO-treatment study in Tam-Cre;Pkd1^ΔC/flox model (n = 8 (4M-4F) ctrls Scr-ASO; n = 4 (2M-2F) ctrls Asns-ASO; n = 12 (3M-9F) cystic Scr-ASO; n = 11 (5M-6F) cystic Asns-ASO). (B) Asns mRNA expression in cystic kidneys compared to controls at P160, treated with either Scr-ASO or Asns-ASO (n = 3 ctrls Scr-ASO; n = 3 ctrls Asns-ASO; n = 9 cystic Scr-ASO; n = 9 cystic Asns-ASO). (C) ASNS protein expression in cystic kidneys, treated with either PBS, Scr-ASO or Asns-ASO, compared to controls. (D) Representative images of IHC ASNS staining in cystic renal epithelium of KspCre;Pkd1^ΔC/flox (P4), and Tam-Cre;Pkd1^ΔC/flox (P160) mice treated with Scr-ASO or Asns-ASO. Scale bar (50 μm). (E) MRI representative images of cystic kidneys treated with Scr-ASO, Asns-ASO. Images were acquired at P40 (before tamoxifen induction) and at P130. Scale bar (1 cm). (F) Total kidneys volume of cystic and control mice treated with Scr-ASO or Asns-ASO, calculated at P40, P100 and P130 (n = 8 ctrls Scr-ASO; n = 4 ctrls Asns-ASO; n = 12 cystic Scr-ASO; n = 11 cystic Asns-ASO). (G) BUN concentration in cystic and control mice treated with Scr-ASO or Asns-ASO, measured at P160 (n = 7 ctrls Scr-ASO; n = 4 ctrls Asns-ASO; n = 12 cystic Scr-ASO; n = 11 cystic Asns-ASO). (H) Creatinine concentration in cystic and control mice treated with Scr-ASO or Asns-ASO, measured at P160 (n = 6 ctrls Scr-ASO; n = 4 ctrls Asns-ASO; n = 12 cystic Scr-ASO; n = 11 cystic Asns-ASO). (I) Representative images of cystic kidneys harvested at P160 from mice treated with Scr-ASO, Asns-ASO or PBS. Scale bar (1 cm). (J) Percentage of kidney weight normalized on total body weight at P160 of mice treated with Scr-ASO or Asns-ASO (n = 7 ctrls Scr-ASO; n = 4 ctrls Asns-ASO; n = 12 cystic Scr-ASO; n = 11 cystic Asns-ASO). (K) Quantification of the cystic area percentage of the total kidney area measured in sections of cystic Scr-ASO or Asns-ASO groups (n = 9 cystic Scr-ASO; n = 8 cystic Asns-ASO). (L) Representative images of H&E stained cross sections of P160 cystic kidneys, treated with Scr-ASO or Asns-ASO. Scale bar (2 mm). Data information: in (B, F–H, J), data are shown as mean ± SD. One-way ANOVA. ns not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. In (K) data are shown as mean ± SD. Student’s unpaired two-tailed t test. **P < 0.01. Source data are available online for this figure.

Figure 3. ASNS is upregulated via GCN2-dependent activation of ATF4.

(A) Schematic representation of the integrated stress response-dependent transcription of Asns. (B) ATF4 protein expression in KspCre;Pkd1^ΔC/flox cystic kidneys and relative controls at P4. (C) ATF4 protein expression in Scr-ASO-treated Tam-Cre;Pkd1^ΔC/flox mice and relative controls at P160. (D) Eif2ak4 expression in Pkd1^−/− and control MEF cells upon silencing (representative out of three independent experiments; n = 3 biological replicates). (E) Asns expression in Pkd1^−/− and control MEF cells upon Eif2ak4 silencing (representative of three independent experiments; n = 3 biological replicates). (F) P-GCN2, ATF4 and ASNS protein expression in Pkd1^−/− and control MEF ± silencing of Eif2ak4 (n = 2). Data information: in (D, E) data are shown as mean ± SD. One-way ANOVA. ns not significant; *P < 0.05; **P < 0.01. Source data are available online for this figure.

Figure 4. Asns-ASO treatment rescues the metabolic reprogramming occurring in PKD.

(A) MRI scan of P130 cystic Tam-Cre;Pkd1^ΔC/flox kidneys, treated with ASOs, included in the metabolomic analysis. Scale bar (1 cm). (B) Principal Component Analysis (PCA) of targeted metabolomics of Tam-Cre;Pkd1^ΔC/flox cystic kidneys and relative controls, treated with Scr- or Asns-ASO (n = 5 ctrl Scr-ASO; n = 3 ctrl Asns-ASO; n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (C) Schematic representation and analysis of metabolites involved in ASNS-dependent fueling of TCA cycle (reductive carboxylation) and glycolysis (n = 5 ctrls Scr-ASO; n = 3 ctrls Asns-ASO; n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (D) Pathway Enrichment Analysis of up- and downregulated pathways (Cystic Scr-ASO respect to the other groups), and rescue of Asns-ASO treatment compared to Scr-ASO group. Pathways non-statistically (ns) different from Ctrl group upon Asns-ASO treatment are indicated as completely rescued. Data information: In (C) data are shown as mean ± SD. One-way ANOVA, corrected with Tukey’s multiple comparisons. ns not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. In (D) statistical significance of Pathway Enrichment and Impact was evaluated with Hypergeometric Test with Relative-betweenness Centrality, based on KEGG Database with P < 0.05, FDR < 0.1. Source data are available online for this figure.

Figure 5. Targeting ASNS hampers CAD-dependent pyrimidine biosynthesis in the PKD model.

(A) Schematic representation CAD-dependent de novo pyrimidine biosynthesis pathway and targeted metabolomic profile of relative intermediates normalized on protein content, of P160 Tam-Cre;Pkd1^ΔC/flox cystic and control kidneys treated with ASOs (n = 5 ctrls Scr-ASO; n = 3 ctrls Asns-ASO; n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (B) Quantification of Ki-67-positive nuclei in the epithelium lining the cysts in the kidney cortex of Tam-Cre;Pkd1^ΔC/flox mice at P94, treated with Scr- or Asns-ASO (n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (C) Targeted metabolomic of main products of purine de novo biosynthesis pathway normalized on protein content, of P160 Tam-Cre;Pkd1^ΔC/flox cystic and control kidneys treated with ASOs (n = 5 ctrls Scr-ASO; n = 3 ctrls Asns-ASO; n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). Data information: In (A, C) data are shown as mean ± SD. One-way ANOVA, corrected with Tukey’s multiple comparisons. In (B) data are shown as mean ± SD. Student’s unpaired two-tailed t test. ns not significant; *P < 0.05; **P < 0.01; ***P < 0.001. Source data are available online for this figure.

Figure 6. CAD-dependent pathway is upregulated in different models of PKD and rescued by Asns silencing in vitro.

(A) Western blot analysis of Tam-Cre;Pkd1^ΔC/flox and relative controls at P160 treated with Asns-ASO or Scr-ASO. (B) Cad expression in microarray from cystic kidneys of P10 Pkd1^v/v mice, compared to relative controls. (C) Cad expression in RNA-seq of Pkd1^RC/RC cystic mice at P12, compared to controls. (D) CAD expression in microarray of ADPKD patients samples. (E) Dot plot of snRNA-seq dataset showing CAD expression in clusters identified in human ADPKD cystic and normal kidney tissues. PT proximal tubule, FR-PTC failed-repair proximal tubular cells, PEC parietal epithelial cells, TAL thick ascending limb of Henle’s loop, DCT distal convoluted tubule, CNT_PC connecting tubule and principal cells, ICA Type A intercalated cells, ICB Type B intercalated cells, PODO podocytes, ENDO endothelial cells, FIB fibroblasts, LEUK leukocytes. (F) Tracing metabolomics analysis (¹³C₅-glutamine) evaluating labeled Carbamoyl Aspartate (M + 4) in Pkd1^−/− and control MEF cells ± silencing of Asns (n = 6 biological replicates in one experiment). (G) Tracing metabolomics analysis (¹⁵N₂-glutamine) evaluating labeled metabolites of the de novo pyrimidine biosynthesis pathway in Pkd1^-/- and control MEF cells ± silencing of Asns (n = 6 biological replicates in one experiment). Data information: in (F, G) data are shown as mean ± SEM. One-way ANOVA. ns not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are available online for this figure.

Figure 7. Combined targeting of glutamine usage via Asns and glycolysis delays the PKD progression in Pkd1^ΔC/f;Tam-Cre mice and ameliorates the endpoint phenotype and function.

(A) Up- and downregulated Pathway Analysis Enrichment (cystic Asns-ASO not completely rescued compared to cystic Scr-ASO). (B) Experimental design of Tam-Cre;Pkd1^ΔC/flox and relative controls treated with Asns-ASO or Scr-ASO. (C) MRI representative images of cystic kidneys treated with Scr-ASO + PBS or Asns-ASO + 2DG. Images were acquired at P40 (before tamoxifen induction), P100 and P130. Scale bar (1 cm). (D) Total kidneys volume of cystic and control mice treated with Scr-ASO + PBS (n = 7, 5M-2F) or Asns-ASO + 2DG (n = 12, 5M-7F), calculated at P40, P100 and P130. (E) Percentage of kidneys weight normalized to total body weight of P160 mice treated with Scr-ASO + PBS (n = 7, 5M-2F) or Asns-ASO + 2DG (n = 12, 5M-2F). (F) BUN of cystic and control mice treated with Scr-ASO + PBS (n = 7, 5M-2F) or Asns-ASO + 2DG (n = 12, 5M-7F), measured P160. (G) Comparison of total kidneys volume measurement at P100 and P130 between cystic animals of the two presented studies: the single treatment with Asns-ASO (ASO study) and the combinatory treatment Asns-ASO + 2DG (COMBO study) (n = 12 cystic Scr-ASO; n = 11 cystic Asns-ASO; n = 7 cystic Scr-ASO + PBS; n = 12 cystic Asns-ASO + 2DG). ASO study bars correspond to data shown in Fig. 2E, while COMBO study bars correspond to data shown in (D). Data information: in A statistical significance of Pathway Enrichment and Impact was evaluated with Hypergeometric Test with relative-betweenness Centrality, based on KEGG Database with P < 0.05, FDR < 0.1. In (D–G) data are shown as mean ± SD. Unpaired two-tailed Student’s t test. ns not significant. *P < 0.05, **P < 0.01. Source data are available online for this figure.

Figure 8. Schematic representation of metabolic rewiring supported by ASNS in PKD and proposed strategy for therapy.

The scheme summarizes how glucose and glutamine usage supports the progression of PKD. Increased glycolysis and lactate production (Rowe et al, 2013) reduce the fueling of glucose-derived carbons to the TCA cycle (Podrini et al, 2018) (yellow boxes). Glutamine utilization partially compensates for this deficit. Glutamine utilization is ASNS-dependent and fuels TCA cycle both through reductive carboxylation and oxidative phosphorylation (Podrini et al, 2018) (yellow boxes). Here we show that ASNS upregulation is downstream of the amino acid unbalance and GCN2-dependent AAR (blue box). In line with this, targeting Asns in vivo in PKD models retards disease progression. Metabolomic profiling in these tissues reveals a prominent upregulation of de novo pyrimidine biosynthesis as a consequence of ASNS enzyme upregulation and glutamine utilization (blue box). Furthermore, combining targeting of ASNS and glycolysis via treatment with ASO and 2DG, respectively, breaks the energetic balance acquired in PKD resulting in the amelioration of cystic phenotype. Bold arrows and text highlight processes found upregulated in PKD.

Figure EV1. Validation of cell lines KO for Pkd1 and of antibodies directed against ASNS.

(A) Immunoblotting on control and Pkd1 KO mCCD clones for the detection of PC1 protein, using PC1 7E12 and E8 antibodies. FL, PC1 full-length protein; NTF, PC1 N-terminal fragment; CTF, PC1 C-terminal fragment. PC1 bands were detected in control cells and not in Pkd1 KO clones. Vinculin was used as loading control to show equal loading of protein samples. (B) Representative images of immunofluorescence staining of ASNS in Pkd1^−/− and Pkd1^+/+ MEF cells, untreated (NT), scrambled (Scr) or silenced for Asns. Scale bar (10 μm). (C) Asns expression in Pkd1^−/− and control MEF cells upon silencing (n = 3 biological replicates). (D) Representative images of IHC ASNS staining in cystic renal epithelium of Tam-Cre;Pkd1^ΔC/flox (P160) mice treated with Scr-ASO or Asns-ASO. Scale bar (50 μm). Data information: In (C) data are shown as mean ± SD. One-way ANOVA, corrected with Tukey’s multiple comparisons. ****P < 0.0001.

Figure EV2. Pilot test of Asns-ASO on a medium-term PKD model.

(A) Experimental design of pilot study on Tam-Cre;Pkd1^ΔC/flox and relative controls treated with Asns-ASO or Scr-ASO (n = 4 ctrl Scr-ASO; n = 2 ctrl Asns-ASO; n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (B) Asns mRNA expression in Tam-Cre;Pkd1^ΔC/flox and relative controls treated with Asns-ASO or Scr-ASO (n = 4 ctrl Scr-ASO; n = 2 ctrl Asns-ASO; n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (C) Representative images of cystic kidneys and relative controls at P94 treated with Asns-ASO or Scr-ASO. (D) Percentage of kidneys weight normalized to body weight of cystic and relative controls treated with Asns-ASO or Scr-ASO (n = 4 ctrl Scr-ASO; n = 2 ctrl Asns-ASO; n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (E) BUN of cystic and relative control kidneys treated with Asns-ASO or Scr-ASO (n = 4 ctrl Scr-ASO; n = 2 ctrl Asns-ASO; n = 4 cystic Scr-ASO; n = 4 cystic Asns-ASO). (F) Quantification of the cystic area percentage of the total kidney area measured in transversial sections of cystic Scr-ASO or Asns-ASO groups (n = 5 cystic Scr-ASO; n = 5 cystic Asns-ASO). (G) ASO distribution in cystic kidneys harvested from Asns-ASO-treated mice at P95 from three different litters. Representative images of ASO distribution in total kidneys (upper panel, scale bar (1 mm)), renal cortex middle panel, scale bar (50 μm)), and renal medulla (lower panel, scale bar (50 μm)). Arrows indicate the ASO-positive cystic epithelium. Data information: in (B, D, E) data are shown as mean ± SD. One-way ANOVA. ns non-significant; *P < 0.05; **P < 0.01. In (F) data are shown as mean ± SD. Student’s unpaired two-tailed t test. **P < 0.01.

Figure EV3. Hierarchical clustering of metabolomics data.

(A) Dedrogram diagram showing the hierachical clustering of 4 ASO-treated groups of samples analyzed through LC–MS. (B) Heatmap based on the HCA of the metabolom (265 metabolites) comparing Scr- and Asns-ASO-treated cystic and control kidneys. (C) Heatmap based on the HCA of amino acids detected in PKD cystic and control kidneys treated with Scr-ASO or Asns-ASO. Data information: in (A–C) clustering based on t test/ANOVA result performed with Metaboanalyst 5.0.

Figure EV4. Validation of the CAD-pyrimidine biosynthesis pathway in different models.

(A) Intermediate metabolites of de novo pyrimidine biosynthesis pathway analyzed through untargeted metabolomics of KspCre;Pkd1^ΔC/flox cystic kidneys and relative controls at P4 (n = 8 ctrls; n = 8 cystic). (B) P-CAD, ASNS and P-S6RP protein expression in Pkd1^−/− and control MEF cells treated for 4 h, 8 h, or 24 h with rapamycin (50 nM), after overnight serum starvation (n = 3). Data information: in (A) data are shown as mean ± SEM. Student’s t test. **P < 0.01; ****P < 0.0001.