Fig 2: Summary of primary findings and predicted effects in the central nervous system. ACE1 is a membrane-bound dipeptidyl-peptidase but can be cleaved primarily by ADAM10 to release the two catalytic domains into the extracellular space (far left panel). Under normal conditions, membrane-bound ACE1 cleaves Ang I to form Ang II (yellow peptide icon) which can bind to receptors such as AT1R and activate intracellular signaling cascades. Ang II can also be further degraded into other angiotensin peptide products by proteases like ACE2 and aminopeptidase A. sACE1 can also cleave Ang I to Ang II, albeit at a lower rate than membrane-bound ACE1 depending on the microenvironment. Previous work has shown the AD-associated R1279Q mutation in knock-in mice increases ACE1 presence at the neuronal membrane which caused increased Ang II production and AT1R activation, leading to neuronal death (far right panel). We have demonstrated in this work that R1279Q stable cell lines have increased membrane-bound ACE1. Two additional AD-associated ACE1 mutations, T916M and N1036K, increase Ang II production through a different mechanism than R1279Q. These mutations increased the catalytic activity of membrane-bound ACE1 which led to an increase in Ang II (middle left and middle right panels). In the central nervous system, this increase in Ang II can over-activate AT1R and lead to neuron death, as was shown in the R1279Q ACE1KI/KI mice. Notably, T916M had decreased ACE1 at the plasma membrane which was complimented by increased ACE1 shedding. This may contribute to our proposed mechanism of neuron death by increasing the amount of sACE1 available for Ang II production in the extracellular space, or soluble mutant ACE1 may play an additional role as a signaling molecule. Further work in vivo is required to understand the way in which these two mutations affect ACE1, RAS, and AD pathogenesis in the central nervous system. The schematic was created with BioRender.

Fig 3: T916M and N1036K mutants have increased membrane-bound ACE1 catalytic activity.A, figure chematic of membrane-bound ACE1. Created with BioRender. B, production of fluorescent o-aminobenzoyl peptide over 3 hours, three differentiations (2 differentiations for R1279Q) (two-way ANOVA, Dunnett’s multiple comparisons test, see Data Set 1 for details). C, ACE1 activity in WCE measured by area under the curve (AUC) of o-aminobenzoyl peptide over 3 hours. Plotted as fold of Wild-type (WT). Brown-Forsythe ANOVA test (p < 0.0001), Dunnett’s T3 multiple comparisons test Wild-type vs each cell line (vs Vector p < 0.0001, vs T916M p < 0.0001, vs N1036K p = 0.0006, vs R1279Q p = 0.0004). Captopril-treated sample average: 0.0465 AUC Fold WT. D, AUC normalized to WCE ACE1 protein per reaction (2 differentiations for R1279Q). Plotted as Fold WT. One-way ANOVA (p < 0.0001), Dunnett’s multiple comparisons test Wild-type vs each mutant (vs T916M p = 0.0024, vs N1036K p < 0.0001, vs R1279Q p = 0.0693). E, pg/ml of Ang II in stable cell line conditioned media after cells were treated with 500 nM Ang I for 1 h, plotted as fold of WT. Brown-Forsythe ANOVA test (p < 0.0001), Dunnett’s T3 multiple comparisons test all cell lines vs each other (Vector vs Wild-type p < 0.0001, Vector vs T916M p = 0.0005, Vector vs N1036K p < 0.0001, Vector vs R1279Q p = 0.0005, all other comparisons p > 0.05). Captopril-treated sample average: 0.0059 pg/ml Ang II Fold WT. F, pg/ml Ang II measured in conditioned media after 1 h treatment with 500 nM Ang I, normalized to WCE ACE1 protein per reaction, two differentiations. Plotted as Fold WT. Brown-Forsythe ANOVA test (p < 0.0001), Dunnett’s T3 multiple comparisons test Wild-type vs each mutant (vs T916M p = 0.0031, vs N1036K p < 0.0001, vs R1279Q p = 0.8669). G, pg/ml Ang II produced from incubating 3 ng sACE1 and 100 nM Ang I for 1 h in vitro, two differentiations. Plotted as Fold WT. Brown-Forsythe ANOVA test (p = 0.0242), Dunnett’s T3 multiple comparisons test Wild-type vs each mutant (vs T916M p = 0.4315, vs N1036K p = 0.0254, vs R1279Q p = 0.9927). Captopril-treated sample average: 0.0115 pg/ml Ang II Fold WT. H) pg/ml of degraded Aβ42 measured by ELISA in conditioned media after 24-h treatment with 1 nM Aβ42, two differentiations. Brown-Forsythe ANOVA test (p = 0.072), Dunnett’s multiple comparisons test Wild-type vs each mutant (vs Vector p = 0.8787, vs T916M p = 0.9936, vs N1036K p = 0.223, vs R1279Q p = 0.2669). Captopril-treated sample average: 178.76 pg/ml Aβ42. I) pg/ml of degraded Aβ42 measured by ELISA after incubating 3 ng sACE1 and 1 nM Aβ42 for 24 h in vitro, two differentiations. Plotted as Fold WT. Brown-Forsythe ANOVA test (p = 0.3583), Dunnett’s T3 multiple comparisons test Wild-type vs each mutant (vs Vector p = 0.6515, vs T916M p = 0.9068, vs N1036K p = 0.9067, vs R1279Q p = 0.6765). Captopril-treated sample average: 92.8 pg/ml Aβ42.

Fig 4: The R1279Q ACE1 mutant causes increased abundance of ACE1 protein on the cell-surface.A, ACE1 protein measured in WCE via ELISA, three differentiations (two differentiations for R1279Q). Brown-Forsythe ANOVA test (p < 0.0001), Dunnett’s T3 multiple comparisons test Wild-type vs each mutant (vs Vector p < 0.0001, vs T916M p < 0.0001, vs N1036K p = 0.0001, vs R1279Q p = 0.0096). B, surface ACE1 protein measured by flow cytometry. Each data point represents the mean fluorescence intensity (MFI) of single cells from 20,000 cells counted per well, n = 3 wells per cell line. One representative experiment showed two separate differentiations. One-way ANOVA (p < 0.0001), Dunnett’s multiple comparisons test Wild-type vs each mutant (vs Vector p < 0.0001, vs T916M p < 0.0001, vs N1036K p = 0.5619, vs R1279Q p = 0.0013). C, ACE1 shedding assessed via ELISA measurement of ACE1 in conditioned media and normalized to ACE1 in WCE, three differentiations (two differentiations for R1279Q). Brown-Forsythe ANOVA test p < 0.0001), Dunnett’s T3 multiple comparisons test Wild-type versus each mutant (vs T916M p < 0.0001, vs N1036K p = 0.1427, vs R1279Q p = 0.0548).