Adrenomedullin restores the human cortical interneurons migration defects induced by hypoxia

The hFAs recapitulate key developmental aspects of cortical interneurons migration from the medial ganglionic eminences (MGE) into the dorsal forebrain. To study the effects of hypoxia on the migration of human cortical interneurons, we established a model for hypoxic injury using hFAs. Using standardized published protocols, we previously contributed to developing (Sloan et al., 2018; Birey et al., 2017), we generated hFAs through the differentiation of hiPSCs into human cortical organoids (hCO) containing dorsal forebrain excitatory neurons (Paşca et al., 2015), and human subpallial organoids (hSO) containing cortical interneurons (Birey et al., 2017). To visualize migrating cortical interneurons, we first fluorescently tagged cells within hSOs with a forebrain interneurons cell-type-specific lentiviral reporter (Dlxi1/2b::eGFP; Potter et al., 2009). Subsequently, we fused the labeled hSOs with hCOs to generate hFAs. At approximately 10–14 days after fusion into hFAs, we observed substantial and active migration of Dlxi1/2b::eGFP⁺ cortical interneurons on the hCO side of hFAs (Figure 1A, Figure 1—figure supplement 1A).

Figure 1 with 1 supplement see all

Download asset Open asset

To monitor the migration patterns of cortical interneurons under control and hypoxic stress, we established a long-term live imaging microscopy setup in intact hFAs using a confocal microscope with a motorized stage (Zeiss LSM 980 with Airyscan 2) equipped with an environmentally controlled chamber (37 °C, 5% CO₂) and an oxygen level controller (Okolab Bold Line). For this, the hFAs were transferred to the confocal microscope in a glass-bottom 96-well plate with fresh cell culture media. We imaged individual Dlxi1/2b::eGFP-tagged interneurons within the hCOs for a total of 48 hr: 24 hr (0–24 hr) under control conditions (37 °C, 5% CO₂, 21% O₂), followed by 24 hr (24–48 hr) in pre-equilibrated hypoxic media (37 °C, 5% CO₂, 94.5% N₂,<1% O₂; Figure 1A).

To validate the decrease in partial pressure of O₂ (PO₂) in the culture media in hypoxic conditions, we used an oxygen optical microsensor OXB50 (50  μm, PyroScience) attached to a fiberoptic multi-analyte meter (FireStingO₂, PyroSciences). In line with our previous reports (Pașca et al., 2019), we observed a decrease in the media PO₂ level from ~150 mmHg in control conditions to ~25–30 mmHg in the confocal microscope environmental chamber in hypoxic conditions (Figure 1—figure supplement 1B).

Next, we confirmed that these conditions are sufficient to activate the hypoxia response pathway in hSOs within 24 hr of exposure, without inducing extensive death of cortical interneurons. Specifically, to validate the induction of hypoxia, we demonstrated the stabilization of HIF1α hypoxia-marker protein using western blot analyses in hSOs derived from four individual cell lines (two-tailed paired t-test, p=0.01; Figure 1—figure supplement 1C, D). Separately, we validated the activation of the hypoxia-response pathway using targeted transcriptional analyses for hypoxia-responsive genes, including PFKP (two-tailed paired t-test, p=0.004), PDK1 (two-tailed paired t-test, p=0.001), VEGFA (two-tailed paired t-test, p=0.003; Figure 1—figure supplement 1E, Supplementary file 2). To check whether Dlxi1/2b::eGFP-tagged interneurons start expressing cell markers for cell death, we used two different assays: we performed FACS-based Annexin V analyses in fluorescently tagged interneurons, as well as immunostainings for cleaved caspase 3 (c-CAS3). We found no significant difference between control and hypoxic conditions using Annexin V (unpaired t-test, p=0.362; Figure 1—figure supplement 1F) nor c-CAS3 (unpaired t-test, p=0.125) (Figure 1—figure supplement 1G, H).

These results demonstrate that our target hypoxic conditions (PO₂ ~25–30 mmHg) are sufficient to induce and activate the hypoxia response pathway, but do not induce cell death. This experimental design aligns with our goal of targeting mild hypoxic stress, partly because mild hypoxic episodes are much more common than the severe ones in preterm infants, and partly because we wanted to define, for the first time, the altered molecular pathways in hypoxic (but live) cortical interneurons.

The migration of cortical interneurons from the ganglionic eminences into the dorsal forebrain is at its peak during the second half of pregnancy in humans (Anderson et al., 1997). It is characterized by repetitive cycles of nuclear saltations (nucleokinesis events) in the direction of the leading process (Mayer et al., 2018; Lim et al., 2018). To study migration patterns of hypoxic cortical interneurons, we followed the movement of the same individual Dlxi1/2b::eGFP⁺ cortical interneurons for a total of 48 hr (control conditions: hr 0–24; hypoxia: hr 24–48). We analyzed the number of nuclear saltations, saltation length, and the directionality of the migration in interneurons that had already migrated into the hCO at the time of the hypoxia exposure (Figure 1B and C).

To quantify the number of saltations, we focused on interneurons that had at least one saltation within the first 24 hr under control conditions and remained visible throughout the entire 48 hr of imaging. This experimental approach was specifically chosen to focus the analyses on actively migrating cells, while excluding those that may have ceased migration and begun integrating into synaptic networks at the time of imaging. To quantify saltation length and directionality, we narrowed the cell population to the subset of cells that had at least one saltation in both control and hypoxic conditions, as a minimum of 1 saltation/condition is a prerequisite for direct comparisons for these parameters.

To identify defects in the migration patterns of interneurons in hypoxia, we analyzed migrating interneurons for 24 hr in control conditions (hr 0–24), followed by another 24 hr in hypoxic conditions (hr 24–48). We observed a ~58% decrease in the number of saltations of individual interneurons (paired Wilcoxon test, p<0.0001), and demonstrated that the phenotype is present in hFAs derived using hiPSCs from four separate individuals (two-tailed paired t-test, p=0.003; Figure 1D). Importantly, 70 out of 129 actively migrating cortical interneurons not only decreased the number of saltations but became completely stationary upon exposure to hypoxia for 24 hr. However, we found no significant differences in the average length of saltation of cells when analyzed by individual cells (paired Wilcoxon test, p=0.29) and by hiPSC line (two-tailed paired t-test, p=0.15; Figure 1E). In addition, we did not observe any differences in the directionality of migration measured by the ratio of Euclidean/accumulated distance when analyzed by individual cells (paired Wilcoxon test, p=0.36) and by hiPSC line (two-tailed paired t-test, p=0.47; Figure 1F).

To ensure that the confocal chamber environment during the prolonged live imaging does not result by itself in defects in migration of the cortical interneurons, or that a possible hypoxia phenotype could be related to a physiologic saltatory pause in the second half of imaging, we performed two separate types of experiments. First, we compared the migration patterns in control conditions (37 °C, 5% CO_2, 21% O₂) in the first 24 hr (hr 0–24) versus the second 24 hr (hr 24–48). We identified no differences in the total number of saltations/24 hr when analyzed by individual cells (paired Wilcoxon test, p=0.09) and hiPSC line (two-tailed paired t-test, p=0.2; Figure 1G); no difference in average saltation length when analyzed by individual cells (two-tailed paired t-test, p=0.79) and by hiPSC line (two-tailed paired t-test, p=0.47; Figure 1H); no difference in directionality of migration when analyzed by individual cells (paired Wilcoxon test, p=0.38) and by hiPSC line (two-tailed paired t-test, p=0.62; Figure 1I). Second, to ensure that the deliberate exclusion from analyses of cortical interneurons that display no saltations in control conditions does not bias the migration phenotype by excluding cells that might otherwise migrate more in the 24–48 hr of imaging, we identified all cortical interneurons that displayed no migration in control conditions (0–24 hr) and quantified their migration in hypoxia. Out of 113 total cells, 108 (95.6%) remained stationary (Figure 1—figure supplement 1I), suggesting these cells are not in a highly migratory phase at the time of imaging.

Overall, the experiments in Figure 1, Figure 1—figure supplement 1 suggest mild hypoxia preferentially affects the total migration length of cortical interneurons by decreasing the number of saltations; however, it does not interfere with the length of individual saltations nor the directionality of the migration. The additional validation experiments demonstrate the robustness of the experimental design and strengthen the validity of saltation deficit phenotype in hypoxia. Details on the number of cells, experiments, and number of hiPSC lines used for each experiment are presented in Supplementary file 1.

To investigate the transcriptional changes induced by mild hypoxia in hSOs, we performed bulk transcriptome-wide RNA-Sequencing (RNA-Seq) at 46 days in culture (before fusion with hCOs), following exposure to hypoxia for 12 hr or 24 hr, as well as at 72 hr after reoxygenation (Figure 2A). For these experiments, we induced hypoxia using the C-chamber hypoxia sub-chamber (Biospherix). We confirmed a significant decrease in PO₂ between control and hypoxia conditions (two-tailed paired t-test, p<0.0001; Figure 2—figure supplement 1A), and showed that PO₂ levels were similar to the ones measured in the confocal environmental chamber (Figure 1B).

Figure 2 with 1 supplement see all

Download asset Open asset

Hierarchical clustering of the gene expression profiles showed clear separation between hSOs from the control (control 12 hr, control 24 hr, control 72 hr), hypoxia-exposed (hypoxia 12 hr, hypoxia 24 hr), and hypoxia followed by reoxygenation samples (hypoxia 72 hr after <1%; Figure 2—figure supplement 1B), suggesting that hypoxia exposure induces strong transcriptional profile changes in hSOs, and these largely recover following 72 hr of reoxygenation. For this experiment, we used hSOs derived from 4 individual hiPSC lines, two XX and two XY genotypes (details by hiPSC line are presented in Supplementary file 1).

Next, we identified differentially expressed genes (DEGs) between control, hypoxia-exposed, and reoxygenated samples (padj <0.05, fold change >1.5), controlling for potential confounding variables as described in detail in the methods section. We identified 734 differentially expressed genes at 12 hr, 985 genes at 24 hr, and 21 genes at 72 hr after reoxygenation (Figure 2B and C, Supplementary file 3).

Gene set enrichment analyses (GSEA) demonstrated changes in molecular pathways associated with exposure to hypoxia (including glycolysis, mTORC1, KRAS), inflammatory response pathways (represented by nonspecific immune-related cell surface receptors), estrogen response pathways, and unfolded protein response. In addition, we identified changes in genes associated with general cellular migration (e.g. AXL, GLIPR1, etc; Figure 2D; Supplementary file 2).

Interestingly, we found that the ADM gene had the highest fold change among DEGs at 24 hr of hypoxia exposure (log₂FC = 4.06, padj = 6.08E-19; Figure 2E). To validate the increase in gene expression level of ADM under hypoxic conditions observed in bulk RNA-sequencing, we performed qPCR analyses using hypoxia-exposed intact hSOs. We confirmed the significant transcriptional upregulation of ADM gene expression (two-tailed paired t-test, p=0.02; Figure 2F).

To check whether the transcriptional upregulation of ADM gene expression translates into changes at a protein level, we quantified ADM peptide levels using quantitative enzyme immunoassay (EIA; Phoenix Pharmaceuticals, EK-010–01) in cell culture media from intact hSOs exposed and non-exposed to hypoxia for 24 hr. We observed a significant increase in the protein concentration in the hypoxia-exposed hSOs compared to control conditions (unpaired Mann-Whitney test, p<0.0001; Figure 2G). This experiment was performed in three separate biological replicates of hSOs, each derived from four hiPSC lines in one differentiation experiment (details by hiPSC line are presented in Supplementary file 1).

To bring novel insight into the cell-type-specific expression of ADM under control and hypoxic conditions in human brain cells, we performed single-cell RNA-sequencing in hSOs and hCOs, separated from hFAs that were previously exposed or non-exposed to hypoxia. To do this, we cut the hFAs under microscopic guidance (Figure 3A). We used hFAs derived from two different hiPSC lines.

Figure 3 with 1 supplement see all

Download asset Open asset

Figure 3—source data 1

Original files for western blot analysis are shown in Figure 3F and H.

https://cdn.elifesciences.org/articles/108134/elife-108134-fig3-data1-v1.zip

Download elife-108134-fig3-data1-v1.zip

Figure 3—source data 2

PDF file containing original western blots Figure 3F and H, including the relevant bands, iPSC lines, and experimental conditions.

https://cdn.elifesciences.org/articles/108134/elife-108134-fig3-data2-v1.zip

Download elife-108134-fig3-data2-v1.zip

First, we performed Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction for hSOs non-exposed (n=8691 cells) and exposed to hypoxia (n=7331 cells). We identified five distinct cell clusters including ventral forebrain progenitors and early cortical interneurons (NKX2.1⁺, DLX2⁺, GAD1⁺), cycling progenitors (TOP2A⁺), astrocytes (S100A10⁺), and a small cluster of radial glia (SOX9⁺); as expected due to the previous fusion into hFA, we identified a cluster of cortical glutamatergic neurons (STMN2⁺, NEUROD2⁺; Figure 3B, Figure 3—figure supplement 1A). For hCOs non-exposed (n=6786 cells) and exposed to hypoxia (n=5781 cells), we also identified five cellular clusters consisting of dorsal forebrain progenitors (SOX2⁺), cycling progenitors (TOP2A⁺), cortical glutamatergic neurons (STMN2⁺, NEUROD2⁺), and a small cluster of choroid plexus cells (TTR⁺); as expected due to the migration of interneurons from the hSO side, we identified a cluster of cortical interneurons (NKX2.1⁺, GAD1⁺; Figure 3C, Figure 3—figure supplement 1B). All these cell clusters from hSOs and hCOs demonstrated increased expression of hypoxia-related genes (PDK1, PFKP; Figure 3—figure supplement 1C, D). Moreover, we performed subcluster analyses of GABAergic interneurons on hSO and identified that they primarily express markers for MGE and CGE including NKX2.1, LHX6, somatostatin (SST) or calretinin (CALB2), less calbindin (CALB1), and minimal parvalbumin (PVALB), and we validated these findings and performed immunocytochemistry for the following interneuron subtypes: somatostatin (SST), calbindin (CALB), and calretinin (CALB2; Figure 3—figure supplement 1E, F).

Next, we assessed the expression of ADM in individual cell types. We found that under hypoxic conditions, the expression level of ADM increases in all cell types, but more so in ventral neural progenitors and astrocytes within hSOs, and in dorsal neural progenitors and choroid plexus cells in hCOs (Figure 3D, Figure 3—figure supplement 1G).

Lastly, we evaluated the cell-type-specific expression of RAMP1, RAMP2, and RAMP3, the known receptors for ADM peptide (Kuwasako et al., 2011; Kuwasako et al., 2004). Since the previous literature suggests RAMP2 is the main receptor for ADM, we checked the expression of the RAMP1-3 receptors in control conditions and found that RAMP2 is indeed the most expressed receptor. Interestingly, we found this receptor is preferentially expressed by interneurons (hSOs) and glutamatergic neurons (hCOs) when compared to other cell types and its transcription is increased in hypoxia (Figure 3E, Figure 3—figure supplement 1H).

To validate the findings about the higher expression of RAMP2 in control conditions at a protein level, we performed western blot analyses for RAMP1 and RAMP2 in hSOs in control conditions and found that, indeed, RAMP2 was significantly more expressed at baseline than RAMP1 by individual cell (two-tailed paired t-test, p=0.0004) and by hiPSC line (two-tailed paired t-test, p=0.0036; Figure 3F and G). To validate the findings about the increase in RAMP2 expression in hypoxia in hSOs, we checked for changes in the expression of RAMP2 at a protein level and identified a significant increase when analyzed by individual cell (two-tailed paired t-test, p=0.0027) and by hiPSC line (two-tailed paired t-test, p=0.0235; Figure 3H and I). These experiments were performed in three separate biological replicates of hSOs, each derived from four hiPSC lines in one differentiation experiment. Details by hiPSC line are presented in Supplementary file 1.

Together, these results provide the first cell type-specific characterization of the expression of ADM and its receptors in the brain in control and hypoxic conditions. Importantly, we identify complementary responses of different cell types, with neurons increasing the expression of the receptors for ADM, while astrocytes, choroid plexus, and progenitors increasing the expression of ADM, suggesting a possible evolutionary adaptation mechanism of cells to protect neurons during hypoxia as an environmental stressor.

Based on recent reports about the protective role of exogenous ADM administration in various disease processes (Zhang et al., 2009; Pedreño et al., 2014; Gao et al., 2018; Ashizuka et al., 2016; Clementi et al., 1998), and in the light of our findings described above, we tested the rescue potential of exogenous ADM administration under hypoxic conditions.

We exposed hFAs to hypoxia in the presence of 0.5 μM human ADM (Anaspec, AS-60447; Figure 4A) and imaged individual migrating cortical interneurons (Figure 4B). This dose was chosen based on previous reports from in vitro cultures (Temmesfeld-Wollbrück et al., 2009). We again observed a significant decrease in the number of saltations under hypoxic conditions when analyzed by individual cells (paired Wilcoxon test, p<0.0001) and by hiPSC line (two-tailed paired t-test, p=0.02), and identified rescue of the number of saltations in hypoxia +ADM conditions when analyzed by individual cells (paired Wilcoxon test, p=0.3) and by hiPSC line (two-tailed paired t-test, p=0.9; Figure 4C).

Figure 4 with 1 supplement see all

Download asset Open asset

To check whether exogenous ADM peptide changes the migration of cortical interneurons under control conditions, or whether this is hypoxia-specific, we analyzed the number of saltations of cortical interneurons under control conditions for a total of 48 hr: 24 hr under control conditions (0–24 hr), followed by 24 hr in control +ADM (24–48 hr). We identified no difference in the total number of saltations/24 hr when analyzed by individual cells (paired Wilcoxon test, p=0.23), and hiPSC line (two-tailed paired t-test, p=0.77; Figure 4D), suggesting ADM is not a main mechanism for migration under baseline conditions. These results align with the data about the low baseline expression of ADM (Figure 2G) in control conditions, and its known increase as an acute phase reactant in hypoxia and inflammation (Ishiyama et al., 2023; Valenzuela-Sánchez et al., 2016; Solé-Ribalta et al., 2022).

To verify that the saltation rescue is directly linked to the administration of exogenous ADM, we next performed two types of experiments.

First, we disrupted the disulfide bond between Cys16 and Cys21, which has been shown to be important for ADMs’ biological activity (Fischer et al., 2019), by denaturing and alkylating it using iodoacetamide (Figure 4E, Figure 4—figure supplement 1A). Next, we repeated the hypoxia experiments in the presence of 0.5 μM human denatured ADM (dADM) and demonstrated that it does not rescue the saltation deficit under hypoxic conditions, when analyzed by individual cells (paired Wilcoxon test, p<0.0001) and by hiPSC line (paired t-test, p=0.04; Figure 4F).

Second, we performed pharmacological blocking of RAMP2, the primary recognized receptor for ADM in the literature and in our data. For this, we added 10 μM ADM fragment 22–52 (ADM_22-52; Cayman Chemical Company 24892) to the hypoxic media already supplemented with exogenous ADM (Figure 4G). We validated the saltation deficit in hypoxia and the rescue by exogenous ADM and found that the addition of ADM_22-52 to the hypoxia +ADM condition was sufficient to prevent the saltation rescue when analyzed in individual cells (paired Wilcoxon test, p<0.0001) and by hiPSC line (paired t-test, p=0.003; Figure 4H).

To investigate whether migration of interneurons is restored upon reoxygenation, we quantified the saltations of individual migrating cortical interneurons for a total of 72 hr: 24 hr in control (0–24 hr), 24 hr in hypoxia (24–48 hr), followed by 24 hr in reoxygenation (48–72 hr). Similar to our experiments from Figures 1 and 4, we observed a ~53% decrease in saltations in hypoxia compared to control (Friedman test, p<0.0001), and a 65% decrease in the first 24 hr of reoxygenation compared to control (Friedman test, p<0.0001) when analyzed by individual cells (Figure 4—figure supplement 1B). In contrast, the presence of exogenous ADM during hypoxia exposure (not during reoxygenation) fully rescued the migration in hypoxia as previously shown in Figure 3 and provided partial protection in the first 24 hr after reoxygenation by decreasing the migration of interneurons by 35% compared to control when analyzed by individual cells (Friedman test, p=0.05; Figure 4—figure supplement 1C) and by hiPSC line (one-way ANOVA, p=0.84; Figure 4—figure supplement 1C); again, this is in contrast to the hypoxia conditions without ADM that see a 65% decrease in saltations during reoxygenation.

While not the primary focus of the current study because reoxygenation insult has a different biological mechanism of injury (Datta et al., 2020; Kalogeris et al., 2012) and in-depth analyses would require a different experimental setup, these results reinforce clinical and preclinical data from animal models that reoxygenation injury is important to understand for therapeutic intervention, and that organoid models are able to mimic complex in vivo pathophysiology. Details on the number of cells, experiments, and number of hiPSC lines used for each experiment are presented in Supplementary file 1.

To verify our findings in an independent human model, we used ex vivo developing human brain tissue at ~20 PCW and followed a similar experimental design used for hFA (Figure 5A). Using previously published scRNA-sequencing data from mid-gestation human fetal tissue, we performed subcluster analyses for GABAergic interneurons (Valenzuela-Sánchez et al., 2016; Ashizuka et al., 2016). We identified three subclusters. One small subcluster expressed doublets and was excluded from subsequent analyses (Figure 5—figure supplement 1A; Clementi et al., 1998). The remaining two sub-clusters were found to express MGE- and CGE-derived markers consisting of NKX2.1, LHX6, CALB2, SST, but minimal CALB1 and PVALB similar to the ones in hFAs (Figure 5—figure supplement 1B).

Figure 5 with 1 supplement see all

Download asset Open asset

First, we sectioned human cerebral cortex sections into 400 μm slices (Figure 5B) and cultured them on live imaging-compatible cell culture inserts (Robinson et al., 2006) (diameter, 23.1 mm; pore size, 0.4 μm; Falcon, 353090). To visualize migrating cortical interneurons, we fluorescently tagged cells using the same forebrain interneuron cell-type specific lentiviral reporter Dlxi1/2b::eGFP (Figure 5C; Potter et al., 2009).

To assess whether exposure to hypoxia is sufficient to induce a migration deficit in ex vivo human cortical interneurons as observed in hFA, we employed the live-imaging setup established for hFA and monitored the migration patterns of the Dlxi1/2b::eGFP⁺-tagged interneurons for 24 hr under control conditions (37 °C, 5% CO_2, 21% O₂). We then transitioned the sections into pre-equilibrated hypoxic media (37 °C, 5% CO₂, 94.5% N₂,<1% O₂) and imaged the same cells for another 24 hr. Similar to hypoxic hSOs, we observed a significant transcriptional increase in ADM gene by qPCR (two-tailed unpaired t-test, p=0.0005; Figure 5D). Following exposure to hypoxia for 24 hr, we identified a ~55% decrease in the number of saltations (paired Wilcoxon test, p<0.0001; Figure 5E), but no changes in the average saltation length (paired Wilcoxon test, p=0.88; Figure 5F) nor in directionality (Wilcoxon test, p=0.3; Figure 5G).

Lastly, we found the addition of 0.5 μM ADM peptide during the hypoxia exposure was sufficient to rescue the saltation numbers (paired Wilcoxon test, p=0.09; Figure 5H).

First, we performed EIA analyses focused on cAMP/PKA, AKT/pAKT, and ERK/pERK pathways, which have been previously shown to be modulated by ADM (Ozcelik et al., 2019; Figure 6A). We found that the concentration of cAMP was not affected by hypoxia but was significantly increased in the hypoxia +ADM conditions, both when analyzed by individual samples (one-way ANOVA test, p=0.0007), and by hiPSC line (one-way ANOVA test, p=0.002; Figure 6B). These results were complemented by increased PKA activity in hypoxia +ADM conditions when analyzed by individual samples (one-way ANOVA test, p<0.0001), and by hiPSC line (one-way ANOVA, p<0.0001; Figure 6C). In contrast, the ratio of pAKT/AKT was significantly decreased in hypoxia samples when analyzed by individual samples (one-way ANOVA, p<0.0001) and by hiPSC line (Kruskal-Wallis test, p=0.01), but was not rescued in hypoxia +ADM (Figure 6D). These findings were similar for the ratio of pERK/ERK in individual samples (one-way ANOVA test, p=0.001) and by hiPSC line (one-way ANOVA, p=0.02; Figure 6E). Lastly, we quantified the ratio of pCREB/CREB, which is downstream of all these pathways. We found that hypoxia induced a significant decrease in the ratio of pCREB/CREB when compared to control when analyzed by individual samples (one-way ANOVA test, p=0.003) and by hiPSC line (p=0.048). However, this ratio was restored in hypoxia +ADM conditions when analyzed by individual samples (one-way ANOVA test, p=0.27) and by hiPSC lines (one-way ANOVA, p=0.6; Figure 6F). These results suggest the cAMP/PKA/pCREB pathway is modulated by exogenous ADM administration in hypoxia in hSOs.

Figure 6 with 1 supplement see all

Download asset Open asset

Previous reports demonstrated that increased ratio of pCREB/CREB is directly involved in increasing the expression of GABA_A receptors (Figure 6G; Fukuchi et al., 2014; Birey et al., 2022), which in turn, are well-documented to facilitate interneuron migration (Luhmann et al., 2015). To check whether the expression of GABA_A receptors is affected under hypoxic conditions and rescued by ADM, we quantified the expression of several GABA_A receptor subunits in the presence or absence of ADM during hypoxia exposure. We specifically focused on the most common pentameric combination, α1β2γ2, encoded by GABRA1, GABRB2, and GABRG2 (Connolly et al., 1996; Figure 6G). We found a significant decrease in expression in hypoxia (one-way ANOVA test; control vs hypoxia: p=0.002) and rescue by ADM for GABRA1 (one-way ANOVA test; p=0.17); no significant change in expression in GABRB2 (Kruskal-Wallis test; control versus hypoxia: p=0.08; control versus hypoxia +ADM: p=0.25); a significant decrease in expression in hypoxia (one-way ANOVA test; control versus hypoxia: p=0.007) and rescue by ADM for GABRG2 (one-way ANOVA test; p=0.23; Figure 6H). Moreover, we analyzed other receptor subunits and found mild but still significant decrease in hypoxia and rescue by ADM for receptors GABRB3 (Kruskal-Wallis test; control versus hypoxia: p=0.02; control versus hypoxia +ADM: p=0.05) and GABRG3 (one-way ANOVA test; control versus hypoxia: p=0.01; control versus hypoxia +ADM: p=0.1); no change in expression under hypoxic conditions for receptors GABRA2 (one-way ANOVA; control versus hypoxia: p=0.8; control versus hypoxia +ADM: p=0.75; Figure 6—figure supplement 1A).

Separately, we investigated the expression of CXCR4, CXCR7, and CXCL12, as some of the best-known chemokines/chemokine receptors involved in interneuron migration (Wang et al., 2011; Toudji et al., 2023). Interestingly, we found a significant increase of CXCR4 in hypoxia but no rescue by ADM (one-way ANOVA test; control versus hypoxia: p=0.0005; control versus hypoxia +ADM: p=0.0031), a significant decrease of CXCR7 in hypoxia but no rescue by ADM (one-way ANOVA test; control versus hypoxia: p=0.026; control versus hypoxia +ADM: p=0.043), and a significant decrease of CXCL12 in hypoxia but no rescue by ADM (one-way ANOVA test; control versus hypoxia: p=0.024; control versus hypoxia +ADM: p=0.0039; Figure 6I).

Overall, these results suggest that rescue of GABA_A receptor expression is sufficient to restore migration patterns in hypoxic cortical interneurons, even in the absence of rescue of cytokine receptors.

A graphical summary of the proposed mechanims of protection by ADM in presented in Figure 7A.

Figure 7

Download asset Open asset

The hiPSC lines used in this study were validated using standardized methods we previously described (Robinson et al., 2006; Takano, 2015; Pedreño et al., 2014; Connolly et al., 1996). Cultures were regularly tested for, and maintained Mycoplasma free. SNP-array experiments were performed for genomic integrity. The differentiation experiments were performed using four control hiPSC lines derived from fibroblasts harvested from four healthy individuals (two XX and two XY). Informed consent for fibroblast collection was obtained from all individuals, and this process was approved by the Stanford IRB Panel.

Brain region-specific human cortical organoids (hCO) and human subpallium organoids (hSO) were differentiated using feeder-free grown hiPSCs using validated protocols we previously reported and contributed to developing (Robinson et al., 2006; Takano, 2015; Zhang et al., 2009; Pedreño et al., 2014; Connolly et al., 1996; Wang et al., 2011; Toudji et al., 2023; Kepecs and Fishell, 2014; Marín, 2012). Briefly, hiPSCs were maintained in six-well plates coated with recombinant human vitronectin (VTN-N, Thermo Fisher Scientific, A14700) in Essential 8 medium (Thermo Fisher Scientific, A1517001). For differentiation, hiPSC colonies were lifted from the plates using Accutase (Innovate Cell Technologies, AT-104). Approximately 3 million cells were transferred to each well in AggreWell 800 (STEMCELL Technologies, 34815) and centrifuged at 100 × g for 3 min in hiPSC medium supplemented with ROCK inhibitor Y-27632 (Selleck Chemicals, S1049). After 24 hr, the newly formed organoids were transferred into ultra-low-attachment plastic dishes (Corning, 430293) in Essential 6 medium (Thermo Fisher Scientific, A1516401) supplemented with dorsomorphin (2.5  μM, SigmaAldrich, P5499) and SB-431542 (10  μM, Tocris, 1614) and additional XAV-939 (hCO 0.5  μM, hSO: 2.5 μM Tocris, 3748). This medium was replaced daily for the first five days. On the sixth day in suspension, organoids were transferred to neural medium containing Neurobasal A (Thermo Fisher Scientific, 10888022), B-27 supplement without vitamin A (Thermo Fisher Scientific, 12587010), GlutaMax (2 mM, Thermo Fisher Scientific, 35050061) and penicillin and streptomycin (100 U mL^–1, Thermo Fisher Scientific, 15140122). For hCO, the neural medium was supplemented with the growth factors EGF (20 ng mL^–1, R&D Systems, 236-EG) and FGF2 (20 ng mL^–1, R&D Systems, 233-FB) every day until day 15, then every-other day until day 23. For hSO, we included continued supplementation with XAV-939 (2.5 μM) on days 6–23, and SHH pathway agonist SAG (smoothened agonist; 100 nM; Thermo Fisher Scientific, 566660) on days 12–23. Between days 25 and 43, the neural medium was supplemented with neurotrophic factors BDNF (20 ng mL^–1, Peprotech, 450–02) and NT3 (20 ng mL^–1, Peprotech, 450–03), to promote differentiation of the neural progenitors into neurons in both hCO and hSO; media change was performed every other day. After day 43, hCO and hSO were maintained in neural media.

The viral infection of the hSO for visualization of migrating cortical interneurons was performed as previously described (Robinson et al., 2006; Ozcelik et al., 2019). Briefly, at approximately 45–55 days in culture, hSO were transferred to a 24-well plate (Corning, 3474) containing 500 μL of neural medium with 10 μL of virus (Lenti-Dlxi1/2b::eGFP; construct reported and applied in reference; Gao et al., 2018). After 24 hr, 500 μL of neural medium was added. On day 3, all media was removed and replaced with 1 mL of neurobasal medium. The next day, hSO were transferred into fresh neural media in ultra-low attachment plates. After 5–7 days, hCO and virally infected hSO were used to generate hFA. To do this, one hCO and one hSO were transferred together into a 1.5 mL microcentrifuge Eppendorf tube and maintained in direct contact for 3 days. More than 95% of hCO and hSO fused. Subsequently, hFA were carefully transferred into 24-well ultra-low attachment plates (Corning, 3474), and media changes were performed very gently every two to three days. Lentivirus construct (lenti-Dlxi1/2b::eGFP) was received as a gift from J. L. Rubenstein, and virus was generated by transfecting HEK293T cells with PEI Max (Polysciences, 24765); after 48 hr the media was collected and ultracentrifuged at 17,000 rpm at 6 °C for 1 hr.

Ex vivo developing human cortical tissue at 20 PCW was obtained under a protocol approved by the Research Compliance Office at Stanford University. The tissue was processed within 4 hr after collection using a previously described protocol (Robinson et al., 2006; Ozcelik et al., 2019; Connolly et al., 1996). In brief, cortical tissue was embedded in 4% low melting-point agarose in PBS and cut using a Leica VT1200 Vibratome at 400 μm. The sections were transferred onto cell culture membrane inserts (diameter, 23.1 mm; pore size, 0.4 μm; Falcon, 353090) and incubated in culture media (66% BME, 25% Hanks, 5% FBS, 1% N-2, 1% penicillin, streptomycin, and glutamine; all from Invitrogen) and 0.66% D-(+)-Glucose (Sigma) at 5% CO₂, 37  °C with the Dlxi1/2b::eGFP lentivirus for 24  hr. Sections were then transferred to fresh cell culture media, and half media changes were performed every other day. After ~5 days in culture, Dlxi1/2b::eGFP + cells could be detected and imaging was performed within 7–10 days in culture.

For imaging, hFA were transferred intact into a 96–well plate (glass-bottom, Cellvis, Catalog #: P96-0N) in 300 μL of neural media. Imaging was performed under environmentally controlled conditions (37 °C, 5% CO_2, 21% O₂) using a confocal microscope with a motorized stage (Zeiss LSM 980) and equipped with a cage incubator for temperature, gas (including O₂ levels), and humidity control (Okolab Bold Line). During each recording session, we imaged up to 12 hFA at ×10 magnification, at a depth of 350–400 μm, and at a rate of 20 min/frame. The imaging field was focused on the hCO side of the fusion.

Samples were imaged for 24 hr under control conditions (37 °C, 5% CO₂, 21% O₂). To induce acute hypoxia, media was carefully removed and replaced with pre-equilibrated hypoxic media. All procedures were performed in the controlled hypoxic environment of the HypoxyLab (Oxford Optronix). The field of view was readjusted to capture the previous region of interest, and the same cells were imaged for an additional 24 hr while maintaining a hypoxic environment (37 °C, 5% CO₂,<1% O₂). For pharmacological rescue experiments, the pre-equilibrated hypoxic media containing the peptide ADM (0.5 μM, Anaspec, AS-60447)+/-10 uM ADM_22-52 (the RAMP2 receptor blocker [Cayman Chemical Company, cat. no. 24892]), were added to the hFA cell culture media during hypoxia exposure.

The imaging of migration of ex vivo human developing cortical interneurons was done with the same settings as described above, with the exception of image depth being 150–200 μm. Slices were kept on the cell culture inserts during imaging.

Post-acquisition analyses of cell mobility were performed using ImageJ (v:2.00-rc-69/1.52 n). Individual cells were identified and followed both in control and hypoxic conditions. Analyses were performed on all cells that had at least one saltation within the first 24 hrs of imaging (control conditions) and remained visible within the field of view until the end of the imaging period (up to 72 hrs). Paired analyses of number of saltations and saltation length were performed as described in the main text. To estimate the length of individual saltations, we manually tracked and identified the swelling of the soma of Dlxi1/2b::eGFP tagged cortical interneurons before and after saltation and measured the distance (in μm) from the initial position to the new position. The Z-stack alignment present in Zen 3.1 Blue Edition (Zeiss) was used to correct for minor drifts during imaging. To assess changes in directionality of the movement, we extracted the x and y coordinates of each cell per frame and time using the Manual Tracking plugin (ImageJ). The Chemotaxis & Migration Tool (Ibidi) was then used to calculate the Accumulated (A) and Euclidean (E) distances traveled per cell over time. Data is presented by calculating the E/A ratio. Only cells which had at least one saltation within each of the conditions were included in this analysis.

Organoids or developing human cortical tissue used for in vitro hypoxia experiments were transferred to culture medium pre-equilibrated in a hypoxic C-chamber (Biospherix) or HypoxyLab (Oxford). Hypoxic C-chamber was connected to a premixed gas tank containing 5% CO₂ and balancing N₂. Oxygen level in the C-chamber was controlled and monitored using a Proox 110 Compact Oxygen Controller to reach <1% O₂. HypoxyLab was connected to O₂, N₂, and CO₂ gas tanks and built-in gas sensors control each gas level in real-time to achieve 30 mmHg with 5% CO₂. After 24 hrs treatment by hypoxia, organoids or primary tissues were collected immediately for assays.

The O₂ tension of the media was measured using the oxygen optical microsensor OXB50 (50  μm, PyroScience) attached to a fiber-optic multi-analyte meter (FireStingO₂, PyroScience), as previously described (Fukuchi et al., 2014; Marín, 2012).

At 46 days in culture, three hSO from four individual hiPSC lines were exposed to hypoxic conditions for 12 and 24 hr using a hypoxia C-chamber as described above. For acute induction of hypoxia, the media was previously equilibrated overnight at <1% O₂, 5% CO₂, 37 °C with a resulting PO₂ of 25–30 mmHg. After 12 hr and 24 hr of hypoxic exposure, hSO were immediately collected and snap-frozen in dry ice for analyses. Control hSO maintained in regular incubator conditions was collected from each hiPSC line at each experimental timepoint. Additionally, extra hSO exposed to 24 hr of hypoxia was transferred back to baseline conditions and reoxygenated for 72 hr. At this time point, samples were collected, matched with control hSO, and snap-frozen on dry ice. All samples were kept at –80 °C until they were processed for RNA-sequencing.

mRNA was isolated using the RNeasy Mini kit (Qiagen) and sequence libraries were prepared by Admera Health (https://www.admerahealth.com) using the KAPA Hyper Prep Stranded RNA-Seq with RiboErase. Sequencing was performed on the Illumina HiSeq 400 system with paired-end 150 base pair reads. Samples averaged 40 million reads, each representing 20 x genome coverage. Trimmed RNA-Seq reads were aligned to the Ensembl GRCh38 human genome reference using STAR v2.7.3 in gene annotation mode (Lim et al., 2018). Alignment and RNA-Seq quality control metrics were generated with Picard (v 2.21.1). Principal component analysis (PCA) of the quality control matrices did not indicate any sample outliers.

Gene-level counts were filtered to remove lowly expressed genes which did not have at least 1 count per million in at least 4 out of the total 24 samples; this resulted in 17,777 remaining genes. We noticed a slight GC content bias across samples, so gene counts were further normalized by GC content, gene length, and sequencing depth using the CQN R package v1.36.0 (Luhmann et al., 2015). The resulting estimation parameters were provided to the negative binomial generalized linear model while performing the differential expression (DEG) analysis. This was carried out using DESeq2 v1.30.1 to identify significant genes between hypoxic conditions at each time point while also accounting for unwanted variation due to cell line (Wang et al., 2011). The resulting P values were corrected for multiple comparisons using the Benjamini-Hochberg method. Genes were considered significant at FDR  ≤ 0.05 and absolute fold change >1.5 (log2FC of 0.6). In total, we identified 734 DEX genes as a consequence of the hypoxic conditions at 12 hr, 985 genes at 24 hr, and only 21 genes at 72 hr after reoxygenation (Figure 2). We generated the dendrogram in Figure 2B using expression data from all 24 samples, narrowed to the union of the genes differentially expressed between hypoxic and control conditions at 12 and 24 hr or 72 hr after reoxygenation (n=1473). Hierarchical clustering, using complete linkage of Euclidean distance, determined the order of the dendrogram. Z-score normalized expression values are depicted on a continuous scale from lower values (purple) to higher values (orange). The notable reduction of DEG at 72 hr after reoxygenation reflects what we visually see in the PCA of the normalized gene expression, in that the control and hypoxia conditions cluster together at 72 hr post reoxygenation (not shown). Indeed, hierarchical clustering using the Ward’s criterion on the first three principal components separates hypoxic samples at 12 and 24 hr from the control 12 and 24 hr samples as well as the reoxygenated samples (Figure 2—figure supplement 1B).

To gain insight about putatively dysregulated biological pathways, we conducted gene set enrichment analysis that was computed using the fgsea R package v1.18.0 where all genes were ranked by their absolute log₂ fold change (Toudji et al., 2023). We evaluated the Molecular Signatures Database “’allmark’ gene sets and supplemented the WU_cell migration set based on its presence in a preliminary gene ontology enrichment. The top 10 significantly enriched gene sets (padj <0.05) that occur across all timepoints are shown in Figure 2D. Interested in discovering which genes had the largest change in response between 12 hr and 24 hr, we identified genes significant at both timepoints and ranked them by absolute difference in fold change. The top few genes increasing and decreasing in expression are shown in Figure 2E.

For single-cell gene expression analyses, human forebrain assembloids (hFA) were infected with Lenti Dlx1/2b::eGFP as described above. Samples were incubated either under a normoxic environment (37 °C, 5% CO₂, 21% O₂), or under the controlled hypoxic environment of the HypoxyLab (Oxford Optronix). Following the environmental exposure, hFA were dissected under a microscope to separate the hFA into hSO and hCO. To analyze only interneurons migrated within the hCO and not the ones still at the edges of the fusion, we intentionally performed the separation of hFA more toward the hCO side. For single-cell gene expression analysis, organoids were dissociated as previously described (Valenzuela-Sánchez et al., 2016; Marín, 2016). Dissociated cells were resuspended in ice-cold PBS containing 0.02% BSA and loaded onto a Chromium Single cell 3′ chip (with an estimated recovery of 10,000 cells per channel) to generate gel beads in emulsion (GEMs). scRNA-seq libraries were prepared with the Chromium Single cell 3′ GEM, Library & Gel Bead Kit v2 (10 x Genomics, PN: 120237). Libraries from different samples were pooled and sequenced on a NovaSeq S4 (Illumina) using 150 × 2 chemistry. Sample demultiplexing, barcode processing, and unique molecular identifiers counting were performed using the Cell Ranger software suite (v6.1.2 with default settings excluding introns). Reads were then aligned to the human reference genome (GRCh38), filtered, and counted. Expression matrices for the samples were processed using the R (v4.1.2) package Seurat (v4.2). We excluded cells that express less than 2500 genes as well as those with a proportion of mitochondrial reads higher than 15%. Gene expression was then normalized using a global-scaling normalization method (normalization.method=‘LogNormalize’, scale.factor = 10,000), and the 2000 most variable genes were then selected (selection.method=‘versust’). The Harmony package (Marín, 2012) was used to integrate control and hypoxia samples for downstream analysis. The top 7 and 6 principal components for hCO and hSO samples, respectively, were utilized to perform a UMAP projection, implemented with the RunUMAP function with default settings. The clusters were generated with resolutions of 0.1 and 0.3, respectively, for hCO and hSO samples using the FindNeighbors and FindClusters functions. We identified clusters based on differential expression of known markers (Kepecs and Fishell, 2014) identified using the FindAllMarkers function with default settings. In some cases, using expressions of known markers, we grouped together clusters that were originally separate based on the Seurat clustering. For visualization purposes, the color palette of UMAP plots for hSO and hCO was adjusted using ‘magic wand’ in non-contiguous mode for selection of all cells from respective clusters, for subsequent synchronization of the color for the given cell cluster between brain regions (Adobe Photoshop CS 2018). Original plots are available upon request. Markers used to verify the hypoxic response were based on PDK1 and PFKP. Proportions of cells expressing ADM (Expression Level >0) out of all cells per cluster were calculated using (Fukuchi et al., 2014).

Whole hSOs that were infected with Dlxi1/2b::eGFP⁺ were fixed with 4% paraformaldehyde overnight. They were then washed with PBS and transferred to 30% sucrose for up to 72 hr. The samples were transferred into an embedding medium block (Tissue-Tek OCT Compound 4583, Sakura Finetek), snap frozen and sectioned at 10-μm-thick microns using a cryostat (Leica).

The cryosections of hSO were washed with PBS three times for 5 min each to remove excess OCT and then blocked for 1 hr in PBS +0.5% triton X-100+5% BSA+0.5% DMSO at room temperature. The slides were then incubated with cleaved-caspase 3 (anti-rabbit, 1:100, Cell Signaling Technology, 9661T) overnight at 4 °C. After incubation, the slides were washed three times with PBS +0.5% triton X-100 for 5 min each. Secondary antibody incubation was then performed with Alexa Fluor 594 (anti-rabbit, 1:1000, Invitrogen, Cat # A-21207), anti-GFP (1:1000, GeneTex, GTX13970) for 1 hr at room temperature. After, slides were washed two times with PBS for 5 min each and then stained with Hoescht 33258 (1:10,000, Life Technologies) for 5 min at room temperature. Cryosections were mounted on glass coverslips using aquamount (Thermo Fisher Scientific) and were imaged using (Zeiss Axio Observer). Images were visualized and the number of cleaved-caspase 3 (c-CAS3) and Dlxi1/2b::eGFP⁺ positive cells were quantified using Fiji ImageJ Version 1.54 p.

To assess the levels of apoptotic Dlxi1/2b::mScarlet cells, FACS was conducted on NovoCyte Penteon instrument using NovoExpress 1.5.6 software (Stanford FACS Facility). To label inhibitory interneurons in hSOs at later stages of development, hiPSC lines were engineered to express mScarlet under the dlx1/2 promoter. Briefly, LV-Dlxi1/2b-mScarlet lentiviral particles were produced according to the protocol described in this manuscript. Control iPSC lines were plated sparsely and infected with the lentivirus at 1:300 dilution in E8 medium for 24 hr, following which the media was changed to E8 medium without lentivirus. The cells were dissociated using Accutase (Innovate Cell Technologies, AT-104) and plated sparsely in six-well plates for Puromycin (Sigma-Aldrich, P8833) selection of single cell clones with successful genomic integration of the LV-Dlxi1/2b-mScarlet construct. Selected clones were used to make hSO organoids according to the previously described protocol and used for further experiments. On the day of analysis, 3 spheres per line were placed in an Eppendorf tube and dissociated in 300 μL of Accutase at 37 °C for 30 min. After incubation, 1200 μL of neurobasal medium was added to each tube to inactivate the Accutase. The cells were then centrifuged at 200 × g for 4 min and resuspended in DPBS (Thermo Fisher Scientific, 14190144) and strained using a 40 μM cell strainer (Biologix Research Company, 15–1040). Using the Annexin V kit (Invitrogen, V13246), the cells were stained with Annexin V, Alexa Fluor-647 (Invitrogen, A23204) in RT for 15 min. The analysis of Annexin V signals in Dlxi1/2b::mScarlet cells was done using FlowJo (v.10.8.1).

mRNA was isolated using the RNeasy Mini Kit and RNase-Free DNase set (Qiagen, 74136), and template cDNA was prepared by reverse transcription using the PowerUp SYBR Green master mix for qRT-PCR (Life Technologies, A25742). Real-time qPCR was performed on a CFX384 Real-time system (Bio-Rad). Data was processed using the Bio-Rad CFX Maestro (Bio-Rad). Primers used are listed in Supplementary file 2.

Whole hSOs were rapidly lysed by gentle agitation on ice using the RIPA buffer system with protease and a phosphatase inhibitor cocktail (Santacruz, sc-24948A) for HIF1α, RAMP1, and RAMP2 analyses. Whole cell lysates were incubated at 4 °C for 1 hr, centrifuged at 14,000 ⅹ g for 15 min. After the supernatant was collected, and the protein concentration of the supernatant was measured using the Pierce BCA Protein Assay Kits (Thermo Fisher Scientific, 23225). The lysate was denatured in Bolt LDS Sample Buffer (Invitrogen, B0007) at 95 °C for 5 min. The samples (at 7 μg) were loaded on a Bolt 4–12% Bis-Tris Protein Gels (Invitrogen, NW04120BOX) and then transferred to a PVDF membrane using iBlot 2 Transfer Stacks (Invitrogen, IB24002) and an iBlot2 (Method: P3; Thermo Fisher Scientific, IB24002). After the samples were blocked with 5% skim milk (BD, 232100) in TBST (Tris-buffered saline [Boston BioProducts, BM-301X] containing 0.1% Tween 20 [Sigma-Aldrich, P1379 TBST]) for 1 hr, they were incubated in the blocking solution (TBST containing 5% BSA [Gendepot, A0100-005]) with the following primary antibodies: β-actin (anti-mouse, Clone 13E5, 4970 S), HIF1α (anti-rabbit, Clone D2U3T, 14179 S), RAMP1 (rabbit, 1:500, Abcam, ab156575), RAMP2 (mouse, 1:500, Santacruz, sc-365240), α-tubulin (anti-rat, 1:1000; Abcam, ab6160) at 4 °C overnight. After the blot was washed with 5% TBST, the membranes were incubated in the blocking solution with the following horseradish peroxidase-conjugated secondary antibodies: Anti-rat IgG (1:2000; Cell Signaling, #7077), Anti-mouse IgG (1:2000; Cell Signaling, #7076), Anti-rabbit IgG (1:2000; Cell Signaling, #7074) at RT for 1 hr. The SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific, 34095) was used to develop the membranes. The iBright 1500 (Thermo Fisher Scientific, A44114) was used to detect and analyze protein bands. The intensity of each band was quantified with ImageJ (1.53t) software (NIMH, Bethesda, MD) with normalization to background β-actin or α-tubulin.

ADM levels were quantified using the competitive enzyme immunoassay for human ADM (Phoenix Pharmaceuticals, EK-010–01) and assays were performed according to the manufacturer’s recommendation. Calculations of sample concentration were performed based on peptide standard solution in the range of 0–100 ng/mL. The concentrations of the samples were log₁₀ transformed, and concentrations of the samples were interpolated from the sigmoidal curve. The range of reliable detection was between 0.01100 ng/mL. Samples with concentration <0.01 ng/mL were quantified as 0.

The quantification of the pAKT:AKT(pS473) ratio and pERK:ERK (pT202/pY204, pT185/pY187) ratio was done using the AKT (Total/Phospho) InstantOne ELISA Kit (Thermo Fisher Scientific, 85-86046-11) for pAKT:AKT, and the ERK1/2 (Total/Phospho) InstantOne ELISA Kit (Thermo Fisher Scientific, 8586013–11) for pERK:ERK. Both assays were performed according to manufacturers’ protocol. For both assays, samples were lysed in a 200 μL lysis buffer provided by the manufacturer. Samples were incubated for 1 hr, and 50 μL of lysate per well was used each of the assays. Quantification of cAMP levels was performed using the cAMP Assay Kit (Competitive ELISA) (Abcam, ab65355) according to manufacturers’ protocol. For the analysis, samples were lysed in 250 μL lysis buffer, and 100 μL was used for the assay. The concentration of cAMP was calculated based on the standard curve obtained from kit standards. The quantification of PKA was done using the PKA (Protein Kinase A) Colorimetric Activity Kit (Thermo Fisher Scientific, EIA PKA) according to manufacturers’ protocol. For the assay, samples were lysed in 200 μL lysis buffer containing protease inhibitor cocktail (1 μL/mL, Sigma-Aldrich, P1860), phenylmethylsulfonyl fluoride (1 mM, Sigma-Aldrich, 10837091001), and Activated Sodium Orthovanadate (10 mM, Calibiochem, 567540). Samples were incubated for 1 hr under shaking (250 rpm) and centrifuged at 10,000 rpm for 10 min at 4 °C. Samples were then diluted 1 x with kit dilution buffer, and 40 μL of diluted sample was used for analysis. The quantification of the pCREB:CREB ratio was done using the phosphoELISA CREB (pS133) (Thermo Scientific, KHO0241) and phosphoELISA CREB (Total; Thermo Scientific, KHO0231) according to manufacturer’s protocol. For both assays, samples were lysed in 100 μl RIPA lysis buffer containing phosphatase inhibitor cocktail (Santa Cruz Biotechnology, sc-45065) and a 1:50 dilution of the lysates in standard diluent buffer was used for the assay. The concentration of pCREB and CREB (total) was calculated based on the standard curve obtained from the kit standards using the Tecan Infinite M1000 Pro and i-control 1.10 software.

The synthetic peptide ADM (2 nmol, Anaspec, AS-60447) was incubated with either ddH₂0 (for control condition) or Dithiothreitol (DTT; 5 mM, Sigma-Aldrich, D5545), at 56 °C for 30 min. Following reduction of disulfide bridges, alkylation of the free SH-groups was performed with iodoacetamide (IAM; 20 mM, Sigma-Aldrich, I1149) at 23 °C for 45 min in the dark. The unreacted IAM was quenched with 5 mM final DTT. The solution was desalted and concentrated using Amicon Ultra (3kD, EMD Millipore, UFC500324), and the volume was adjusted to the starting volume with UltraPure DNase/RNase Free Distilled Water (Thermo Fisher Scientific, 10-977-015). LC-MS analysis was performed using Vanquish LC system coupled with an ID-X Tribrid mass spectrometry equipped with a heated electrospray ionization source (HESI-II, Thermo Fisher Scientific). For the LC system, a ZORBAX RRHD diphenyl 100x2.1 mm column (Agilent, 858750–944) was used. The mobile phases were 0.1% formic acid (Fisher Scientific, A117-50) in ddH₂O for A and 0.1% formic acid in acetonitrile (Sigma-Aldrich, 34998) for B. The chromatographic gradient was as follows: 0–2 min 10% B, linear increase 2–10 min up to 60% B, linear increase 10–15 min up to 98% B, hold 15–17 min at 98% B, linear decrease 17–18 min to 10% B, equilibrate 18–21 min at 10% B. The mobile phase flow rate was set to 250 μL/min with an injection volume of 3 μL. The autosampler temperature was set to 4 °C, and the column chamber temperature to ambient. Data acquisition on the IDX MS was acquired at 120,000 resolution setting with positive ion voltage of 3.5 kV, vaporizer and transfer tube temperature of 325 °C; AGC target set to standard, RF lens set to 45%, sheath gas set to 40 AU, aux gas set to 12 AU, maximum injection time of 100ms, with 5 microscans/acquisition, and scan range of 500–2000 m/z. Examination of the LC-MS data was performed using Xcalibur FreeStyle 1.6 (Thermo Fisher Scientific). Here, we monitored the formation of carbamidomethyl-cysteines at Cys16-Cys21 (predicted M+1: 6142.9973), or the reduced form of the peptide (predicted M+1: 6028.9544).

Data are presented as mean  ± s.e.m. unless otherwise indicated. Distribution of the raw data was tested for normality; statistical analyses were performed as indicated in figure legends. Sample sizes were estimated empirically or based on power calculations. Blinding was used for analyses comparing control and treatment samples.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Stanford Maternal & Child Health Research Institute (MCHRI) Postdoctoral Fellowship (to LL), Knut and Alice Wallenberg Foundation Postdoctoral Fellowship (to WPM), Swedish Research Council IPD (to WPM), Bill and Melinda Gates Foundation (to AMP).

1. Sent for peer review: July 11, 2025
2. Preprint posted: September 11, 2025
3. Reviewed Preprint version 1: September 24, 2025
4. Reviewed Preprint version 2: April 2, 2026
5. Version of Record published: May 15, 2026

You can cite all versions using the DOI https://doi.org/10.7554/eLife.108134. This DOI represents all versions, and will always resolve to the latest one.

© 2025, Puno, Michno et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.