The lung microvasculature promotes alveolar type 2 cell differentiation via secreted SPARCL1

Abstract

Lung endothelial cells (ECs) and pericytes are closely juxtaposed with the respiratory epithelium before birth and thus may have instructive roles during development. To test this hypothesis, we screened EC-secreted proteins for their ability to alter cell differentiation in alveolar organoids. We identified secreted protein acidic and rich in cysteine-like protein 1 (SPARCL1) as an extracellular matrix molecule that can promote alveolar type 2 (AT2) cell differentiation in vitro. SPARCL1-treated organoids display lysozyme upregulation and a doubling in the number of AT2 cells at the expense of intermediate progenitors. SPARCL1 also induces the upregulation of nuclear factor κB (NF-κB) target genes, and suppression of NF-κB activation in lung organoids blocked SPARCL1 effects. NF-κB activation by lipopolysaccharide (LPS) was sufficient to induce AT2 cell differentiation; however, pharmacological inhibition of the pathway alone did not prevent it. These data support a role for SPARCL1 and NF-κB in alveolar cell differentiation and suggest a potential value in targeting this signaling axis to promote alveolar maturation and regeneration.

Keywords: SPARCL1; alveolar type 2 cells; cell differentiation; endothelial cells; extracellular matrix; lung alveologenesis; lung development; lung organoids; pericytes; vascular-epithelial crosstalk.

Figure 1

SPARCL1 promotes AT2 cell differentiation in fetal lung organoids (A) Endothelial cells (ECs) are closely juxtaposed with the alveolar epithelium in saccular lungs. Immunostaining of 150 μm precision-cut lung slices from E16.5 and E18.5 mouse embryos. CDH1 (magenta, epithelium), RAGE (magenta, basolateral membranes of AT1 cells), EMCN (green, ECs). Insets: high magnification of endothelial-epithelial contacts (yellow arrowheads). Scale bars: 100 μm, 50 μm (inset). (B) Shortlist of candidate genes encoding secreted proteins. Candidates’ expression levels at E18.5 compared with E15.5 in sorted (KDR⁺) lung ECs (Daniel et al., 2018). (C) Recombinant protein screen (n ≥ 3). SPARCL1 treatments lead to an upregulation of Lyz2. FGF7 control treatments lead to an upregulation of Sftpc (adj. p = 0.0371) and a downregulation of Ager and Lyz2 (adj. p = 0.0002). Relative mRNA levels for markers of alveolar epithelial cell identity: Ager (AT1 and intermediate progenitor [IP] cells), Sftpc (AT2 and IP), Lyz2 (AT2), and Lamp3 (AT2). Data are presented as mean ± SD. p values are from one-way ANOVA, Tukey’s multiple comparison testing. (D) Lyz2 mRNA levels increase dose dependently in organoids treated with SPARCL1, but not SPARC (n = 5 dams, at least 12 organoids per condition). Data are presented as mean ± SD. (E) Quantification of alveolar cell proportions in recombinant protein-treated organoids (n = 3 dams, at least 6 organoids per condition). SPARCL1 induces an increase in AT2 cell counts. FGF7 control treatments lead to an increase in IP cell counts (adj. p < 0.0001). Data are presented as mean ± SD. p values are from one-way ANOVA, Tukey’s multiple comparison testing. (F) Increased AT2 cell counts in SPARCL1-treated organoids. Maximum intensity projections of representative control and SPARCL1-treated organoids (4 μg/mL) immunostained for RAGE (red, AT1 cells) and SFTPC (white, AT2 cells). Increased numbers of SFTPC⁺/RAGE⁻ cells (yellow arrowheads) in SPARCL1-treated organoids. Insets: details of organoid branches. Scale bars: 50 μm, 10 μm (insets).

Figure 2

SPARCL1 is a marker of the E18 microvasculature (A) Sparcl1 expression peaks in lung capillary ECs at E18 and is maintained in pericytes postnatally. gCap, general capillary cell; aCap, alveolar capillary cell/aerocyte; data from Negretti et al., 2021. (B) Leiden clustering of single lung EC transcriptomes from 6 developmental stages (E12-P3). Left: pre-saccular EC transcriptomes (blue, yellow, green) separate from postnatal ones (purple, brown) in the uniform manifold approximation and projection (UMAP) space. Right: cluster 3 (green) is distinct from cluster 1 (blue, pre-saccular ECs) and clusters 2 and 4 (yellow and red, postnatal ECs); data from Negretti et al., 2021. (C) Cluster 3 ECs derive from the lungs at E18. Cluster identity of ECs quantified as proportion of all ECs profiled per developmental stage. Each scRNA-seq replicate is quantified independently. A majority of ECs profiled at E12, E15, and E16 are represented in cluster 1. Clusters 2 and 4 are populated by postnatal ECs. The majority of E18 ECs are in cluster 3, and a minority in clusters 1 and 2. p values are from empirical Bayes moderated ANOVA test. (D) High levels of Ly6e and Sparcl1 expression identify cluster 3 ECs. Top 15 differentially expressed genes between EC clusters 1 and 3. (E) SPARCL1 (green) co-localizes with membranes of microvascular ECs (EMCN, magenta) and pericytes (CSPG4, magenta) in distal lungs. Immunostaining of E18 (top row) and P0 (bottom row) lung cryosections. Left: overview of ECs in distal lung regions. Scale bars: 50 μm (left), 10 μm (middle and right).

Figure 3

SPARCL1 triggers AT2 cell differentiation via TLR4 and NF-κB activation in lung organoids (A) Workflow schematic for RNA-seq of organoids. Control and SPARCL1-treated (2 μg/mL) organoids (n = 3 dams, at least 16 organoids per condition) were collected after 24 h (RNA-seq) and 48 h (RT-qPCR to validate recombinant protein activity). (B) Upregulated genes in SPARCL1-treated organoids compared with control (24 h). Heatmap showing SPARCL1-regulated genes (adj. p < 0.2). (C) The SPARCL1-induced gene signature is enriched in NF-κB pathway and inflammatory signaling genes. GSEA hallmark collections by high normalized enrichment score (NES): TNFA signaling via NF-κB (NES: 1.94), inflammatory response (NES: 1.94), epithelial mesenchymal transition (NES: 1.92). GSEA, gene set enrichment analysis; FWER, family-wise error rate. (D) mRNA levels for Nfkbia and Nfkbiz, but not for Nfkbib, are increased in SPARCL1-treated organoids (48 h, n = 4 dams, at least 16 organoids per condition). Data are presented as mean ± SD. (E) Bacterial LPS mimics SPARCL1 transcriptional effects in lung organoids. mRNA levels for the mature AT2 marker genes Lyz1, Lyz2, and Lamp3, as well as for Nfkbia, are increased by LPS treatment (n = 4 dams, at least 12 organoids per condition). Sftpc and Ager mRNA levels remained unchanged. Data are presented as mean ± SD. (F) TLR4⁺ alveolar epithelial cells in lung organoids on culture days 6 and 8, in areas of active cell differentiation. Scale bars: 50 μm. (G) Pharmacological inhibition of TLR4 blocks SPARCL1 transcriptional effects in lung organoids. 1 μM TAK-242 reduced and 10 μM TAK-242 blunted the SPARCL1-induced upregulation of Lyz1, Lyz2, and Lamp3, as well as of Nfkbia (n = 3 dams, at least 12 organoids per condition). Sftpc and Ager mRNA levels remained unchanged. Data are presented as mean ± SD. (H) Pharmacological inhibition of RELA nuclear translocation blocks SPARCL1 transcriptional effects in lung organoids. 10 μM JSH-23 profoundly reduced Lyz1, Lyz2, and Lamp3 mRNA levels (n = 3 dams, at least 11 organoids per condition). Sftpc and Ager mRNA levels remained unchanged. Nfkbia expression did not change significantly upon JSH-23 treatment. Data are presented as mean ± SD. (I) Pharmacological inhibition of IKKβ blocks SPARCL1 transcriptional effects in lung organoids. 10 μM BI-605906 blunted the SPARCL1-induced upregulation of Lyz1, Lyz2, and Lamp3, as well as of Nfkbia (n = 2 dams, at least 12 organoids per condition). Controls were treated with the inactive and structurally similar compound BI-5026. Data are presented as mean ± SD.

Figure 4

NF-κB target gene transcription correlates with AT2 cell maturity, and NF-κB activation is sufficient for AT2 cell differentiation in organoids (A) Transcriptional diversity of alveolar epithelial cells in developing mouse lungs (E12-P14). scRNA-seq data and cell type annotation from Negretti et al., 2021. Red: AT2 cells; green: AT1 cells. (B) SPARCL1/NF-κB target gene expression maps to AT2 cells beginning at E18. Mean expression levels for the SPARCL1/NF-κB target genes in alveolar epithelial cells profiled at E16-P7. (C) Stage-resolved comparison of the mean expression level for SPARCL1/NF-κB target genes between AT1 and AT2 cells. (D) TLR4 (green) is localized on the plasma membrane of distal airway epithelial cells at early saccular stages and marks the basolateral membrane of a subset of AT2 cells (asterisks; LAMP3, yellow). TLR4⁺/LAMP3⁺ cells are in contact with SPARCL1⁺ membranes (magenta). Immunostaining of E17 and E18 lung cryosections. Left: low-magnification overviews. Right: single channel and merged views. Scale bars: 100 μm (left), 10 μm (right). (E) NF-κB signaling activation in organoids is sufficient for AT2 cell differentiation. LPS stimulation of organoids increased the proportion of AT2 cells (blue bars) and decreased the number of IP cells (yellow bars). JSH-23 treatment alone did not prevent AT2 cell differentiation (n = 2 dams, at least 6 organoids per condition). Data are presented as mean ± SD. p values are from one-way ANOVA, Tukey’s multiple comparison testing. (F) Proposed signaling model. SPARCL1 is secreted by lung endothelial cells (ECs, purple) and pericytes (PCs, purple) and promotes AT2 cell differentiation (blue) via TLR4 and NF-κB. Cells marked by dashed outlines represent FGF-expressing fibroblasts.