N‐acetylaspartate release by glutaminolytic ovarian cancer cells sustains protumoral macrophages

Abstract Glutaminolysis is known to correlate with ovarian cancer aggressiveness and invasion. However, how this affects the tumor microenvironment is elusive. Here, we show that ovarian cancer cells become addicted to extracellular glutamine when silenced for glutamine synthetase (GS), similar to naturally occurring GS‐low, glutaminolysis‐high ovarian cancer cells. Glutamine addiction elicits a crosstalk mechanism whereby cancer cells release N‐acetylaspartate (NAA) which, through the inhibition of the NMDA receptor, and synergistically with IL‐10, enforces GS expression in macrophages. In turn, GS‐high macrophages acquire M2‐like, tumorigenic features. Supporting this in␣vitro model, in silico data and the analysis of ascitic fluid isolated from ovarian cancer patients prove that an M2‐like macrophage phenotype, IL‐10 release, and NAA levels positively correlate with disease stage. Our study uncovers the unprecedented role of glutamine metabolism in modulating macrophage polarization in highly invasive ovarian cancer and highlights the anti‐inflammatory, protumoral function of NAA.

Data information: Data are displayed as mean AE SEM. Statistical significance was calculated by one-way ANOVA analyses with Tukey correction (A, B), unpaired t-test (C) and defined as *P <0.05. Figure EV2. Effect of NAA on macrophage polarization by FACS.

A B
A, B Flow cytometric quantification of the percentage of CD80 + (A) and MHCII + (B) cells after specific treatments. Mφ macrophages were used as a control. Representative overlaid flow cytometry histograms normalized to cell count (n = 3 biological replicates) are shown.
Data information: Data are displayed as mean AE SEM. Statistical significance was calculated by one-way ANOVA analyses with Tukey correction and defined as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Figure EV3. Effect of NAA on macrophages in different conditions.

▸
A qRT-PCR quantification of ASPA mRNA levels in LPS/IFNc macrophages treated with NAA (10 lM) and/or IL-10 and/or NMDA and/or with siASPA (n = 3 biological replicates). B-E qRT-PCR quantification of (B) CD206, (C) CD163, (D) CD80, and (E) GLUL mRNA levels in resting (Mφ) macrophages and LPS/IFNc macrophages treated with NAA (10 lM) and/or IL-10 and/or NMDA (n = 3 biological replicates) in the presence of Gln. F-I Comparison between qRT-PCR quantifications of (F) CD206, (G) CD163, (H) CD80, and (I) GLUL mRNA levels in LPS/IFNc macrophages treated with NAA (10 lM) and/or IL-10 and/or NMDA (n = 3 biological replicates) in the absence of Gln, with qRT-PCR quantifications of the same markers in LPS/IFNc macrophages treated with NAA (10 lM) and/or IL-10 and/or NMDA (n = 3 biological replicates) in the presence of Gln (n = 3 biological replicates). J-L qRT-PCR quantification of (J) CD206, (K) CD163, (L) GLUL mRNA levels in IL-10 macrophages treated with NAA (10 lM) and/or NMDA (n = 3 biological replicates). A Lateral view of the human cryo-EM solved NMDAR structure (6irg.pdb), consisting of 2 GluN1 subunits (white cartoon) and 2 GluN2A (orange cartoon) in complex with Glu, Gly, Ro25-6981, and MK-801 ligands, reported in white, magenta, and cyan sphere representation, respectively. B Scheme representation of the human NMDAR interacting with agonists and/or inhibitors. C, D Modeling and zoomed views of the GluN2D ligand-binding core showing residues involved in direct interactions with Glu (white sticks), NMDA (green sticks), and NAA (yellow sticks).
▸ Figure EV5. Ascitic and cancer tissue arrays ASPA, GS, GLS1, Ki67, and CD163 evaluations. stained with anti-GS, GLS1, Ki67, and CD163. The representative merged tissue images ("merge" panels, also shown in Fig 5G) are shown here together with those in splitted channels. The images were collected at 20× magnification using Nikon Eclipse 80i-ViCO. The scale bars indicate 100 lm.
Data information: Data are displayed as mean AE SEM. For all panels, statistical significance was calculated by one-way ANOVA analyses with Dunn's correction. Data information: Transcripts per million (TPM) values were used as measure of mRNA expression. Data are displayed as mean AE SEM. Statistical significance was calculated by paired t-test (A, B, D, E), Pearson correlation coefficients (C) and defined as *P < 0.05, **P < 0.01, ****P < 0.0001, ns, not significant.