Abstract
Background
Amniotic epithelial cells are fetal-derived stem cells, capable of differentiating into all three germ layers, including mature epithelial cell populations. Here, we hypothesised that the amniotic epithelium might serve as a surrogate tissue source for investigating transcriptional profiles in the respiratory epithelium of newborns.
Methods
In this study, we compared gene expression profiles and weighted gene coexpression network structure in paired amniotic and newborn nasal epithelial samples from 85 participants in the Airway Epithelium Respiratory Illnesses and Allergy (AERIAL) cohort.
Results
In total, 11 867 genes (79.7%) were commonly expressed in both amniotic and nasal epithelium, with uniquely expressed genes (2563 and 458, respectively) enriched for biological functions related to the specialist function of each tissue (e.g. developmental programs and ciliation, respectively). We observed a strong overlap in weighted gene coexpression network structure between both tissues, with 10 coexpression modules identified by consensus network analysis. Genes commonly expressed in both tissues and/or found in the consensus network modules were enriched for biologically relevant gene signatures and pathways related to airway function. We observed significant overlap in gene expression and network structure between the amniotic epithelium and published datasets of epithelial samples from lower airways and other epithelial tissues, including skin and oesophagus, suggesting a global epithelial signature.
Conclusion
We observed significant overlap in gene expression and network structure between paired amniotic and nasal epithelial samples, supporting the potential of amnion as a noninvasive and abundant tissue surrogate. Future studies investigating amnion-based biomarkers for respiratory exposures in utero and childhood disease outcomes are needed to extend these results towards clinical translation.
Shareable abstract
We observed significant overlap in gene expression and network structure between paired amniotic and nasal epithelial samples, supporting the amnion as a useful surrogate tissue for the developing respiratory epithelium https://bit.ly/4niuHPY
Introduction
The amniotic epithelium offers a promising complementary tissue source for examining epithelial biology in the newborn. The amnion is a fetal-derived tissue, consisting of a single layer of pluripotent human amniotic epithelial stem cells, which can be differentiated into all three germ layers [1–3]. This includes the ability to differentiate into mature epithelial cells in culture, including respiratory-like epithelial cells capable of seeding lung tissue in animal models [1–3]. In addition, both the amniotic epithelium and the developing epithelium of the fetus and are exposed to the same in utero environment including the immune/inflammatory mediators present in amniotic fluid [4–6].
While mature epithelial tissues have been increasingly recognised for their integral function in health and disease, the amniotic epithelium has not been sufficiently studied in this context. In particular, the respiratory epithelium has been consistently identified as a key player in human respiratory disease pathogenesis, including in conditions such as allergic rhinitis [7], preterm lung disease [8], asthma [9, 10] and COPD [11]. However, sampling both the upper and lower airway can be technically challenging, particularly in young children [12–16].
Therefore, we hypothesised that the amnion, when sampled from the placenta after delivery, could serve as a noninvasive and abundant source of epithelial cells and offer a viable alternative for studying respiratory epithelial biology, while overcoming the limitations of airway sampling. We aimed to assess similarities and differences in gene expression and coexpression network structure between amniotic epithelium isolated from the placenta at birth and paired newborn nasal epithelial samples.
Materials and methods
Study setting and ethics
The Airway Epithelium Respiratory Illnesses and Allergy (AERIAL) birth cohort study [17] is nested within ORIGINS [18, 19] and collects longitudinal samples and data from recruited participants from birth to 5 years. This human study was approved by Ramsay Health Care WA-SA HREC (approval no. 1908) and was performed in accordance with the Declaration of Helsinki. All parents, guardians or next of kin provided written informed consent for the newborn participants to participate in this study.
Sample collection and processing
This pilot study utilised 85 matched amniotic membrane biopsies and newborn nasal brushings from the AERIAL cohort. Matched placentae were processed within 48 h post-birth with the amniotic membrane manually separated from the chorion and sectioned into strips [17, 20]. Nasal epithelial cells from newborns were collected within 6 weeks post-birth, as previously described [17]. All samples were cryopreserved in DNA/RNA Shield™ (Zymo Research) at −80°C until batch extraction.
Amniotic samples were thawed, transferred into a Precellys® CK14 soft tissue ceramic bead homogenising tube (Bertin Instruments) and homogenised at 12 833×g for 30 s using the Precellys® 24 Tissue Homogeniser (Bertin Instruments). The homogenate was mixed with QIAzol® (1:4 v/v) and 20% (v/v) chloroform, centrifuged, the aqueous phase collected and an equal volume of 70% (v/v) ethanol was added for total RNA extraction using the PureLink™ mini RNA extraction kit (Thermo Fisher Scientific) as per the manufacturer's instructions.
Total RNA from nasal brushings was extracted using Zymo Quick-RNA™ Microprep Kit (Zymo Research). Samples were thawed, treated with proteinase K for 1 h at room temperature, and centrifuged. Supernatants were mixed with RNA Lysis buffer (1:3 v/v) and 100% ethanol (1:1 v/v). Samples were treated as per manufacturer's instructions, except for an extra 400 µL Wash Buffer spin before DNase I. Kit DNase I was replaced with PureLink™ DNase I Reaction Mix (80 µL reaction volume; 30-min incubation). Elution was in 15 µL of DNase/RNase-free water.
For all samples, total RNA purity was determined by NanoDrop™ 1000 spectrophotometer (A260/280>2.0), yield by Qubit fluorometer and quality by Agilent High Sensitivity RNA ScreenTape Analysis kit and 4200 TapeStation, according to the manufacturer's instructions.
Extracted RNA was sent to the Australian Genome Research Facility for whole transcriptome sequencing. Briefly, whole transcriptome library preparation was performed using the Illumina Stranded total RNA with Ribo-Zero Plus kit. The libraries were sequenced using an Illumina NovaSeq 6000 (Illumina) platform generating at least 50 million paired-end reads (150 bp).
Data analysis
The raw FASTQ files for gene expression were processed using the nf-core/rnaseq v.3.12.0 pipeline on Setonix (Pawsey Supercomputer), supported by The Australian BioCommons Leadership Share (ABLeS) [21]. The pipeline was executed using Nextflow v.23.04.2 [22], using the human Genome Reference Consortium Human Build 38 (GRCh38), alignment with STAR [23], and quantification with RSEM [24].
Outliers were detected independently for each tissue using the standardised connectivity approach from the weighted gene coexpression network analysis (WGCNA) authors with an outlier threshold of −2 [25]. In total, five nasal and two amnion samples were identified as outliers and were removed leaving 163 samples (156 paired) for downstream analysis.
“Expressed” genes in each tissue were defined using a count per million threshold corresponding to an absolute count of at least 10 in more than 90% of samples, as previously published [14]. Asthma [26] and aberrant wound repair signatures [9] were extracted from the literature. Functional annotation of gene lists was performed using gprofiler2, filtering out terms with more than 2000 genes [27].
Raw counts, expressed in both tissues, were transformed using the DESeq2 variance stabilised transformation, which adjusts for library size and transforms into homoscedastic data [28]. Transformed data were used for WGCNA [29, 30], on each tissue independently and then in combination for consensus network analysis. For all networks, a “signed” network with settings deepSplit 2, mergeCutHeight 0.4, and minModuleSize 50 was used [29, 30]. Soft threshold powers of 22 and 16 were chosen for amnion and nasal networks, respectively, to achieve scale free topology. Module preservation was calculated using the modulePreservation function in WGCNA [31]. Modules with a summary z-score statistic >10 are considered strongly preserved, 2–10 as weakly–moderately preserved, and <2 as not preserved. Amnion modules were labelled A1–A9, nasal modules as N1–N10 and consensus modules as C1–C10. Module assignment was performed independently in each analysis so module N1 does not necessarily correlate to modules A1 or C1. N0, A0 and C0 represent the set of genes not allocated to a module by the WGCNA algorithm.
For comparison with other epithelial tissues, we extracted transcriptomic data from published datasets [14, 32]. Nasal and tracheal epithelial samples from older children were extracted from GSE118761 [14]. The Genotype-Tissue Expression (GTEx) Project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by the National Cancer Institute (NCI), the National Human Genome Research Institute (NHGRI), the National Heart, Lung, and Blood Institute (NHLBI), the National Institute on Drug Abuse (NIDA), the National institute of Mental Health (NIMH) and the National Institute of Neurological Disorders and Stroke (NINDS). The data used for the analyses described in this manuscript were obtained from GTEx Analysis v.8 release from the GTEx Portal on 20 May 2024 [32]. Both published datasets were processed and analysed using the same methods as for the AERIAL dataset.
Availability of data and materials
The datasets generated and analysed during this study are available in the Mendeley Data repository [33]. Due to ethical considerations regarding participant confidentiality, raw sequencing files cannot be made publicly available, however the deposited de-identified datasets are sufficient for reproducing all reported findings.
Results
Comparison of gene expression profiles between amniotic and nasal epithelium
Demographic data for the included AERIAL participants are presented in table 1. In total, 163 samples remained in the analysis after removal of outliers, with variability observed in RNA quantity, quality, and mapped counts per sample (table 1).
TABLE 1.
Demographic characteristics of the Airway Epithelium Respiratory Illnesses and Allergy (AERIAL) study
| Number of participants | 85 |
| Sex | Female, 41; male, 44 |
| Maternal ethnicity | White, 70; non-White, 12; not disclosed, 3 |
| Method of delivery | Vaginal, 41; Caesarean section, 41; unknown, 3 |
| Age at newborn nasal brushing, days | 14 (0–42) |
| RNA quantity, ng | Amnion, 3938 (119–18 480); nasal, 480 (24–2148) |
| RNA integrity number | Amnion, 4.7 (1.6–8.8); nasal, 6.8 (4.8–8.9) |
| RNA sequencing counts per sample, million | Amnion, 23.7 (10.7–44.4); nasal, 28.2 (12.7–51.6) |
Data are presented as mean (range) or n.
To characterise the relationship between amniotic and nasal epithelial samples, we first performed principal-component analysis, which, as expected, revealed tissue as the primary source of variation in the data, with the first principal component accounting for 80% of the variation in the dataset (figure 1a). We next compared the overlap of expressed genes between the two tissues (Methods). We identified 12 325 and 14 430 genes expressed in amniotic and nasal epithelium, respectively, of which 11 867 (79.7%) genes were expressed in both tissues (figure 1b). Pathway overrepresentation analysis on the 2563 genes uniquely expressed in nasal epithelium identified terms related to ciliated cells (e.g. “axoneme assembly”), while the 458 genes uniquely expressed in amniotic epithelium were enriched for terms related to other developmental processes (e.g. “circulatory system development” and “embryonic morphogenesis”), consistent with the known specialist roles of these tissues. Overall, gene expression was modestly correlated between the amniotic and nasal epithelium (Kendall's tau=0.54; p-value <2.2×10−16) (figure 1c).
FIGURE 1.
Conservation of global gene expression between amniotic (red) and nasal (blue) epithelium. a) Principal-component analysis of DESeq2 variance stabilised transformation count data. b) Venn Diagram showing overlap in expressed genes in each tissue. Genes were defined as “expressed” in each tissue separately using a count per million threshold corresponding to an absolute count of at least 10 in more than 90% of each tissue sample. c) Correlation plot of the mean variance stabilised transformation counts in each tissue.
We then assessed the relative expression patterns of previously published epithelial markers of asthma [26] and abnormal epithelial wound repair [9] between the two tissues. Among the 91 gene asthma signature, 65 genes (71.4%; hypergeometric p-value=0.022) were expressed in both amnion and nasal epithelium, whereas 16 genes were expressed only in the nasal epithelium, three in the amniotic epithelium and seven in neither tissue. Among the 1060 abnormal wound repair genes, 771 (72.7%; hypergeometric p-value=5.4×10−8) were expressed in both the amnion and nasal epithelium, while 167 genes were expressed only in the nasal epithelium, 38 only in the amniotic epithelium and 84 in neither tissue.
Conservation of gene coexpression networks between amniotic and nasal epithelium
To identify preserved patterns of gene expression between amniotic and nasal epithelial samples, we performed WGCNA on each tissue in parallel using the 11 867 genes expressed in both tissues as input. We identified 10 network modules within the amniotic epithelium samples (table 2) and 11 modules within the nasal epithelium samples (table 3). Overall, we observed strong overlap in network structure with high summary preservation z-scores, with most amnion networks corresponding to one or more nasal networks and vice versa (figure 2, tables 2 and 3). This included two network modules related to the immune response in both tissues (tables 2 and 3). The exceptions were the “A8” network module from the amniotic epithelium (summary preservation z-scores 5.6) and the “N1” network modules from the nasal epithelium (z-score 4.2), which showed only moderate preservation. The “A8” amnion network was enriched with genes involved in the hypoxia response (e.g. “cellular response to hypoxia”; table 2), while the “N1” nasal network was enriched for genes related to cilia (e.g. “cilium organisation”; table 3).
TABLE 2.
Weighted gene coexpression network modules identified in amnion samples
| Amnion network module | Generalised biological function | Module size | Preservation z-score summary in nasal |
|---|---|---|---|
| A1 | Cytoskeleton organisation, regulation of cellular component biogenesis, actin filament-based process, epidermis development, actin cytoskeleton organisation | 1425 | 15.6 |
| A2 | Organelle assembly, cilium organisation, cilium assembly, protein–DNA complex organisation, chromatin organisation | 1393 | 47.1 |
| A3 | Cell migration, cell motility, locomotion, positive regulation of multicellular organismal process, intracellular signalling cassette | 1268 | 17.8 |
| A4 | Aerobic respiration, oxidative phosphorylation, cellular respiration, proton motive force-driven ATP synthesis, generation of precursor metabolites and energy | 1218 | 16.9 |
| A5 | Cytoplasmic translation, translation, peptide biosynthetic process, peptide metabolic process, amide biosynthetic process | 592 | 21.9 |
| A6 | Defence response, inflammatory response, response to cytokine, cellular response to cytokine stimulus, response to biotic stimulus | 365 | 38.9 |
| A7 | Positive regulation of RNA metabolic process, protein–DNA complex organisation, positive regulation of DNA-templated transcription, positive regulation of RNA biosynthetic process, chromatin organisation | 175 | 12.2 |
| A8 | Cellular response to hypoxia, response to hypoxia, cellular response to decreased oxygen levels, response to decreased oxygen levels, cellular response to oxygen levels | 173 | 5.6 |
| A9 | Defence response to other organism, defence response to symbiont, response to other organism, response to external biotic stimulus, innate immune response | 111 | 40.6 |
| A0 | Contains genes not belonging to a specific module | 5147 | NA |
The top five generalised biological functions for genes in each network module were identified using gprofiler2. The preservation summary z-score was calculated for the identified amnion network modules in the paired nasal samples. NA: not available.
TABLE 3.
Weighted gene coexpression network modules identified in nasal samples
| Nasal network module | Generalised biological function | Module size | Summary preservation z-score in amnion |
|---|---|---|---|
| N1 | Cilium organisation, cilium assembly, plasma membrane bounded cell projection assembly, microtubule-based process, cell projection assembly | 1942 | 4.2 |
| N2 | Intracellular transport, establishment of localisation in cell, vesicle-mediated transport, protein transport, establishment of protein localisation | 1845 | 21.7 |
| N3 | Regulation of immune system process, intracellular signalling cassette, regulation of intracellular signal transduction, cell activation, positive regulation of immune system process | 1679 | 27.7 |
| N4 | DNA damage response, DNA metabolic process, DNA repair, cellular catabolic process, interstrand cross-link repair | 1540 | 34.2 |
| N5 | Translation, peptide biosynthetic process, peptide metabolic process, amide biosynthetic process, ribosome biogenesis | 1425 | 31.3 |
| N6 | Organonitrogen compound biosynthetic process, carbohydrate derivative biosynthetic process, mitochondrial transport, cellular lipid metabolic process, lipid metabolic process | 626 | 20.9 |
| N7 | Cell adhesion, enzyme-linked receptor protein signalling pathway, neurogenesis, generation of neurons, cell migration | 577 | 15.6 |
| N8 | Cytoplasmic translation, translation, peptide biosynthetic process, peptide metabolic process, amide biosynthetic process | 289 | 25.9 |
| N9 | Regulation of mRNA metabolic process, regulation of RNA splicing, regulation of mRNA processing | 282 | 40.3 |
| N10 | Defence response to other organism, defence response to symbiont, innate immune response, response to other organism, response to external biotic stimulus | 128 | 36.5 |
| N0 | Contains genes not belonging to a specific module | 1534 | NA |
The top five generalised biological functions for genes in each network module were identified using gprofiler2. The preservation summary z-score was calculated for the identified nasal network modules in the paired amnion samples. NA: not available.
FIGURE 2.
Overlap in network gene membership between amnion and nasal weighted coexpression network modules. The y-axis contains amnion modules A1 to A0; the x-axis contains nasal modules N1 to N0. Numeric values presented are the number of genes in common between each pair of amnion and nasal modules, while the red colour denotes the −log(p-value) from a Fisher's exact test.
Given the similarities between the amniotic and nasal network modules, we next built a consensus network to identify robust gene coexpression patterns common to both epithelial tissues. The consensus network contained 10 modules of coexpressed genes (figure 3a). Of these, six were associated with generalised cellular processes (“C2”, “C3”, “C5”, “C6”, “C7” and “C8”), two were associated with immune responses (“C1” and “C9”) and one with epithelial development (“C4”) (table 4). To assess the relative expression patterns of module genes between tissues, the median log2 fold change difference between the amnion and nasal epithelium was calculated across all genes in each module (figure 3b). Only the “C7” consensus network module (“smoothened signalling pathway”) had a log2 fold change either greater than or less than 1 (fold change of −1.07), supporting an overall similar pattern of gene expression between tissues within the conserved modules.
FIGURE 3.
Consensus coexpression network. a) Consensus weighted gene coexpression network dendrogram from amnion and nasal samples. b) Median log2 fold change in network module gene expression between the nasal and amnion samples. Each conserved module (C1–C0) has been assigned a unique colour for visualisation, with grey (corresponding to C0) referencing genes not assigned to a specific module by the weighted gene coexpression network analysis (WGCNA) algorithm.
TABLE 4.
Consensus weighted gene coexpression network modules identified in paired amnion and nasal samples
| Consensus network modules | Generalised biological function | Module size | Median log2 fold change |
|---|---|---|---|
| C1 | Immune response, regulation of immune system process, defence response, cell activation, leukocyte activation | 596 | −0.393 |
| C2 | Organelle assembly, cilium organisation, cilium assembly, microtubule-based process, cell cycle process | 473 | 0.132 |
| C3 | Cytoplasmic translation, translation, peptide biosynthetic process, organonitrogen compound biosynthetic process, peptide metabolic process | 427 | 0.272 |
| C4 | Epidermis development, skin development, epidermal cell differentiation, keratinocyte differentiation, epithelial cell differentiation | 241 | 0.773 |
| C5 | Cilium organisation, cilium assembly, plasma membrane bounded cell projection assembly, cell projection assembly, organelle assembly | 182 | −0.120 |
| C6 | Microtubule-based movement, microtubule-based process, microtubule-based transport, cilium organisation, cilium assembly | 120 | −0.573 |
| C7 | Smoothened signalling pathway | 118 | −1.07 |
| C8 | Extracellular matrix organisation, extracellular structure organisation, external encapsulating structure organisation, cell migration, tube development | 112 | 0.349 |
| C9 | Defence response to virus, response to virus, response to other organism, response to external biotic stimulus, defence response to symbiont | 76 | −0.813 |
| C0 | Contains genes not belonging to a specific module | 9522 | 0.0102 |
The top five generalised biological functions for genes in each network module were identified using gprofiler2. Differential gene expression analysis was performed between amnion and nasal samples, with the median log2 fold change presented across the genes in each network module.
Global epithelial gene signatures in the amnion
Given that the amniotic epithelium consists of stem cells capable of differentiating into all mature epithelial tissues in the body, we next queried the expression of global epithelial markers using cell-type enrichment analysis with annotations from the Human Protein Atlas. While many of the consensus coexpression network modules were enriched for genes annotated to the nasopharynx (i.e. “C2”, “C3”, “C4” and “C6” consensus modules) as expected, we also observed significant enrichment in consensus module genes that were annotated to the lower airways (e.g. lung/bronchus) as well as other epithelial tissues including skin, gastrointestinal tract (e.g. oesophagus/intestine) and genitourinary tract (e.g. kidney/vagina).
To explore consistency with the unified airway hypothesis (i.e. upper and lower airways as a single connected system) [14], we next assessed for conservation in gene expression and coexpression network structure with our previously published epithelial dataset of matched nasal and tracheal epithelium in older children [14]. Consistent with this hypothesis [14], 11 392 of the 11 867 genes (96.0%) expressed in both tissues in the current study were also expressed in both nasal and tracheal epithelium in an older children dataset. Overall, the consensus network structure was also strongly preserved within nasal and tracheal samples (table 5). We also performed the same analyses in the epithelial-dominant samples from the adult GTEx dataset [32], and identified a high degree of overlap in gene expression between our study and the published lung (11 124 of 11 867; 93.7%), skin (11 124 of 11 867; 93.7%) and oesophagus (11 107 of 11 867; 93.6%) samples, with overall strong preservation of network structure (table 5).
TABLE 5.
Preservation of consensus weighted gene coexpression network modules from our study within other epithelial datasets from the published literature
| Summary preservation z-score | |||||
|---|---|---|---|---|---|
| Consensus network modules |
GSE118761 nasal |
GSE118761 tracheal |
GTEx lung |
GTEx skin |
GTEx oesophagus |
| C1 | 19.3 | 11.9 | 4.6# | 9.3# | 7.1 |
| C2 | 22.4 | 18.7 | 12.5 | 27.3 | 14.3 |
| C3 | 25.0 | 32.5 | 21.0 | 51.3 | 23.7 |
| C4 | 39.6 | 33.6 | 13.8 | 10.4 | 6.0# |
| C5 | 42.7 | 40.1 | 10.3 | 14.9 | 15.7 |
| C6 | 5.5# | 5.8# | 8.1# | 18.1 | 13 |
| C7 | 38.2 | 32.9 | 10.5 | 1.4# | 1.7# |
| C8 | 39.1 | 54.3 | 12.7 | 19.0 | 18.6 |
| C9 | 18.2 | 8.6# | 6.8# | 18.6 | 29.7 |
GTEx: Genotype-Tissue Expression. #: Modules with no (z-score <2) or moderate preservation (z-score 2–10).
Discussion
The airway epithelium plays a central role in the pathogenesis of respiratory disease, but investigations into epithelial physiology are limited by the challenges of sampling airway tissues. We hypothesised that the amniotic epithelium may serve as a useful tissue surrogate for characterising aspects of the respiratory epithelium in the newborn. In support of this hypothesis, we observed significant overlap in global gene expression, expression of epithelial biomarkers of asthma/wound repair and gene coexpression network structure between paired amniotic and nasal epithelial samples.
To our knowledge, this is the first study to compare gene expression patterns in the amniotic and respiratory epithelium. However, previous studies have highlighted the relevance of the amnion to understand the effect of pregnancy conditions, such as pre-eclampsia, on the fetus [5, 34, 35]. In general, the prenatal environment has been increasingly recognised as a critical factor in the origins of chronic respiratory disease in children and adults such as asthma and COPD [36, 37]. For example, prenatal exposures (e.g. cigarette smoke, pollution and diet) and poorly controlled maternal asthma during pregnancy significantly increased risk of asthma development and respiratory morbidity throughout life [4, 36, 38, 39]. Given that the amniotic and respiratory epithelium of the fetus are both in direct contact with amniotic fluid, which contains immune mediators such as cytokines and chemokines [6], it is likely the epithelial cells in both tissues would undergo similar transcriptional changes to these factors. However, exposures such as cigarette smoking during pregnancy were rare in this study (n=4), and so we did not have the power to quantify their shared influence. While our study supports the global conservation of many core biological processes between tissues, additional work is required to test whether biomarkers of in utero exposures are similarly conserved between tissues.
One of the key roles of the respiratory epithelium that mechanistically links it to respiratory disease pathogenesis is its role in directing airway immune responses to inhaled pathogens [40]. In keeping with this function, we observed multiple coexpression network modules in the nasal samples that were enriched for genes related to innate immunity and antimicrobial defence responses, as expected. Reassuringly, these network modules exhibited strong preservation statistics in the amnion samples and these terms were also enriched in the “C1” and “C9” network modules from the consensus coexpression network analysis. The relative expression of genes in these modules across nasal and amnion samples was also similar (i.e. log2 fold change <1; figure 3b), further supporting surrogacy. Another central epithelial process that has been linked to respiratory disease development by our group and others is defective wound repair [9, 41–48]. This epithelial endotype is characterised by aberrant cell migration and wound repair due to extracellular matrix dysregulation [44, 48], abnormal barrier function [47] and impaired innate immunity to viral infection [43, 45], and associated with higher respiratory morbidity [9, 41, 42, 46]. Importantly, 72.7% of genes significantly differentially expressed in basal epithelial cells from children with and without abnormal wound repair [9], were also expressed in both nasal and amniotic epithelial cells in this study. Similarly, genes associated with asthma in a nasal epithelial transcriptomic meta-analysis [26] were also largely expressed in both nose and amnion. Collectively, these data suggest that the amnion may serve as a useful surrogate for assessing core airway functions of the respiratory epithelium.
While a large degree of overlap in gene expression and network structure was observed between tissues, significant differences were still observed as expected given the different functions of amniotic and nasal epithelium. In keeping with the stem cell characteristics of the amnion [2, 3], genes uniquely expressed in the amniotic epithelium were enriched for terms related to morphogenesis and development. Similarly, genes and network modules uniquely identified in the nose were enriched for genes and processes related to ciliated cells, reflecting the terminal differentiation of the nasal epithelium into specialised airway cells (e.g. multiciliated epithelial cells) [13]. While the overlap in biologically relevant processes related to immunity, wound repair and asthma support the potential use of the amnion as a surrogate, these differences between tissues in terms of cellular composition, function, and gene expression will remain an important consideration when studying epithelial biology or developing predictive biomarkers in each tissue.
While this study was focused on assessing the ability of the amniotic epithelium to serve as a surrogate for the nasal epithelium, the ability of the amniotic epithelium stem-like cells to differentiate to all mature epithelium in the body, would support broader surrogacy of the amnion for other epithelial tissues [2, 3]. In keeping with this, we observed preservation of network structure and gene expression between nose/amnion in this study with other previously published epithelial tissue datasets including lung, skin and oesophagus [14, 32], consistent with the significant conservation observed between nasal and tracheal gene expression in older children [14]. Similarly, many of the conserved network modules identified in our dataset were enriched for Human Protein Atlas terms related to other epithelial tissues, suggesting a strong global epithelial signature present in the amnion. Given that many epithelial tissues including those in the genitourinary and gastrointestinal tract have similar challenges to the respiratory epithelium in terms of sampling difficulty, the amnion may additionally serve as a noninvasive tissue surrogate for investigating global epithelial biology at birth.
Our study has some limitations to consider. In this study we performed bulk tissue RNA sequencing to investigate the overlap in global gene expression in each tissue. However, both tissues have complex cellular compositions with the amnion consisting of epithelial stem-like cells (predominant cell type), fibroblasts and immunocytes [49] and the nasal epithelium containing multiple specialised epithelial subsets (e.g. basal, hillock-like, multiciliated, club and goblet cells) [50]. In addition, there is also the likelihood of a degree of cell contamination in amnion and nasal samples with either maternal-derived chorion or circulating immune cells, respectively. Future studies performing single-cell sequencing would provide additional information about the complexity of each tissue and direct overlap in cell-to-cell gene expression patterns [32, 49]. In addition, due to the introduction of COVID-19 pandemic restrictions that limited hospital access for research [17], the AERIAL study was changed so that newborn nasal sampling was performed at home and up to 6 weeks of age. Given that gene expression in the nasal epithelium will be modulated by environmental exposures postnatally, this enforced change likely introduced a source of variability into the dataset that needs to be considered. Finally, given the inherent challenges of sampling the newborn nasal epithelium, we observed a wide range of RNA quality and quantity that may have partially influenced the results of this study, including relatively lower RNA integrity number values in amnion versus nasal samples. However, even with these additional sources of variation, the observation of a relevant overlap between tissues independent of cell-type composition, sampling time and RNA quality/quantity supports the potential for surrogacy.
In summary, we have demonstrated significant overlap in gene expression and coexpression network structure to support the amniotic epithelium as a meaningful surrogate for a newborn's respiratory epithelium. Future studies aimed at identifying biomarkers of in utero exposures and/or future respiratory disease risk are needed to extend these observations through to clinical translation.
Acknowledgements
We would like to thank the AERIAL families, together with the cohort study investigators not listed as authors: José A Caparros-Martin, Peter Le Souëf, Anthony Bosco, Susan Prescott and Desiree Silva. We also extend our gratitude to the hardworking and dedicated research team, for the recruitment, liaising and sample collection over the duration of the cohort study, including the AERIAL Laboratory team: Minda Amin, Bailee Renouf, Courtney Kidd, and Ashleigh Heng-Chin. AERIAL is a sub-project of the ORIGINS birth cohort. This unique long-term study, a collaboration between The Kids Research Institute Australia and Joondalup Health Campus, is one of the most comprehensive studies of pregnant women and their families in Australia to date, recruiting 10 000 families over a decade from the Joondalup and Wanneroo communities of Western Australia. We are grateful to all the ORIGINS families who support the project. We would also like to acknowledge and thank the following teams and individuals who have made The ORIGINS possible: The ORIGINS project team; Joondalup Health Campus (JHC); members of ORIGINS Community Reference and Participant Reference Groups; Research Interest Groups and the ORIGINS Scientific Committee; The Kids Research Institute Australia; City of Wanneroo; City of Joondalup; and Professor Fiona Stanley. The ORIGINS has received core funding support from the Telethon Perth Children's Hospital Research Fund, JHC, the Paul Ramsay Foundation, and the Commonwealth Government of Australia through the Channel 7 Telethon Trust. Substantial in-kind support has been provided by The Kids and JHC.
Footnotes
Provenance: Submitted article, peer reviewed.
Ethics statement: This human study was approved by Ramsay Health Care WA-SA HREC (approval no. 1908) and was performed in accordance with the Declaration of Helsinki. All parents, guardians or next of kin provided written informed consent for the newborn participants to participate in this study.
Author contributions: D.G. Hancock, S.P. Agudelo-Romero, T. Iosifidis, Y.V. Karpievitch, G. Zhang, A. Kicic and S.M. Stick developed the methods and analysis plan. E. Kicic-Starcevich, T. Iosifidis, D.J. Martino, A. Kicic, and S.M. Stick designed and funded the original AERIAL cohort. T. Iosifidis, E. Kicic-Starcevich, A. Kicic, and S.M. Stick designed and funded the addition of amnion sampling. E. Kicic-Starcevich managed the AERIAL cohort. E. Kicic-Starcevich, J. Lim and T. Iosifidis, collected the data. D.G. Hancock and S.P. Agudelo-Romero performed the data analysis. D.G. Hancock drafted the manuscript. S.P. Agudelo-Romero, E. Kicic-Starcevich, J. Lim, T. Iosifidis, Y.V. Karpievitch, G. Zhang, D.J. Martino, A. Kicic and S.M. Stick reviewed and revised the manuscript. All authors read and approved the final manuscript.
Conflict of interest: All authors have confirmed that they have no conflicts of interest to declare.
Support statement: This work was supported by grants from the National Health and Medical Research Council of Australia (NHMRC 1157548), Future Health, Research and Innovation Fund WA Near-Miss Awards, and Wal-yan Respiratory Research Centre Inspiration Award (2023). S.M. Stick holds an NHMRC Investigator Grant, T. Iosifidis is supported by the Stan Perron Charitable Foundation, A. Kicic is a Rothwell Family Fellow and D.G. Hancock is a Stan Perron/Perth Children's Hospital Foundation Early Career Research Fellow. Funding information for this article has been deposited with the Open Funder Registry.
References
- 1.Moodley Y, Ilancheran S, Samuel C, et al. Human amnion epithelial cell transplantation abrogates lung fibrosis and augments repair. Am J Respir Crit Care Med 2010; 182: 643–651. doi: 10.1164/rccm.201001-0014OC [DOI] [PubMed] [Google Scholar]
- 2.Pratama G, Vaghjiani V, Tee JY, et al. Changes in culture expanded human amniotic epithelial cells: implications for potential therapeutic applications. PLoS ONE 2011; 6: e26136. doi: 10.1371/journal.pone.0026136 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Qiu C, Ge Z, Cui W, et al. Human amniotic epithelial stem cells: a promising seed cell for clinical applications. Int J Mol Sci 2020; 21: 7730. doi: 10.3390/ijms21207730 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Liu X, Agerbo E, Schlunssen V, et al. Maternal asthma severity and control during pregnancy and risk of offspring asthma. J Allergy Clin Immunol 2018; 141: 886–892.e3. doi: 10.1016/j.jaci.2017.05.016 [DOI] [PubMed] [Google Scholar]
- 5.Suzuki M, Maekawa R, Patterson NE, et al. Amnion as a surrogate tissue reporter of the effects of maternal preeclampsia on the fetus. Clin Epigenetics 2016; 8: 67. doi: 10.1186/s13148-016-0234-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Weissenbacher T, Laubender RP, Witkin SS, et al. Influence of maternal age, gestational age and fetal gender on expression of immune mediators in amniotic fluid. BMC Res Notes 2012; 5: 375. doi: 10.1186/1756-0500-5-375 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Berni Canani R, Caminati M, Carucci L, et al. Skin, gut, and lung barrier: physiological interface and target of intervention for preventing and treating allergic diseases. Allergy 2024; 79: 1485–1500. doi: 10.1111/all.16092 [DOI] [PubMed] [Google Scholar]
- 8.Looi K, Evans DJ, Garratt LW, et al. Preterm birth: born too soon for the developing airway epithelium? Paediatr Respir Rev 2019; 31: 82–88. doi: 10.1016/j.prrv.2018.11.003 [DOI] [PubMed] [Google Scholar]
- 9.Iosifidis T, Sutanto EN, Buckley AG, et al. Aberrant cell migration contributes to defective airway epithelial repair in childhood wheeze. JCI Insight 2020; 5: e133125. doi: 10.1172/jci.insight.133125 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Iosifidis T, Sutanto EN, Montgomery ST, et al. Dysregulated notch signaling in the airway epithelium of children with wheeze. J Pers Med 2021; 11: 1323. doi: 10.3390/jpm11121323 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Love ME, Proud D. Respiratory viral and bacterial exacerbations of COPD-the role of the airway epithelium. Cells 2022; 11: 1416. doi: 10.3390/cells11091416 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Baschal EE, Larson ED, Bootpetch Roberts TC, et al. Identification of novel genes and biological pathways that overlap in infectious and nonallergic diseases of the upper and lower airways using network analyses. Front Genet 2019; 10: 1352. doi: 10.3389/fgene.2019.01352 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Deprez M, Zaragosi LE, Truchi M, et al. A single-cell atlas of the human healthy airways. Am J Respir Crit Care Med 2020; 202: 1636–1645. doi: 10.1164/rccm.201911-2199OC [DOI] [PubMed] [Google Scholar]
- 14.Kicic A, de Jong E, Ling KM, WAERP, et al. Assessing the unified airway hypothesis in children via transcriptional profiling of the airway epithelium. J Allergy Clin Immunol 2020; 145: 1562–1573. doi: 10.1016/j.jaci.2020.02.018 [DOI] [PubMed] [Google Scholar]
- 15.Miller D, Turner SW, Spiteri-Cornish D, et al. Culture of airway epithelial cells from neonates sampled within 48-hours of birth. PLoS ONE 2013; 8: e78321. doi: 10.1371/journal.pone.0078321 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Stokes AB, Kieninger E, Schogler A, et al. Comparison of three different brushing techniques to isolate and culture primary nasal epithelial cells from human subjects. Exp Lung Res 2014; 40: 327–332. doi: 10.3109/01902148.2014.925987 [DOI] [PubMed] [Google Scholar]
- 17.Kicic-Starcevich E, Hancock DG, Iosifidis T, et al. Airway epithelium respiratory illnesses and allergy (AERIAL) birth cohort: study protocol. Front Allergy 2024; 5: 1349741. doi: 10.3389/falgy.2024.1349741 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.D'Vaz N, Kidd C, Miller S, et al. The ORIGINS Project Biobank: a collaborative bio resource for investigating the developmental origins of health and disease. Int J Environ Res Public Health 2023; 20: 6297. doi: 10.3390/ijerph20136297 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Silva DT, Hagemann E, Davis JA, et al. Introducing the ORIGINS project: a community-based interventional birth cohort. Rev Environ Health 2020; 35: 281–293. doi: 10.1515/reveh-2020-0057 [DOI] [PubMed] [Google Scholar]
- 20.Magatti M, Caruso M, De Munari S, et al. Human amniotic membrane-derived mesenchymal and epithelial cells exert different effects on monocyte-derived dendritic cell differentiation and function. Cell Transplant 2015; 24: 1733–1752. doi: 10.3727/096368914X684033 [DOI] [PubMed] [Google Scholar]
- 21.Gustafsson OJR., Al Bkhetan Z., Francis R, Manos S.. Enabling national step changes in bioinformatics through ABLeS, the Australian BioCommons Leadership Share (3.0). Zenodo 2023. 10.5281/zenodo.10139651 [DOI]
- 22.Di Tommaso P, Chatzou M, Floden EW, et al. Nextflow enables reproducible computational workflows. Nat Biotechnol 2017; 35: 316–319. doi: 10.1038/nbt.3820 [DOI] [PubMed] [Google Scholar]
- 23.Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 2013; 29: 15–21. doi: 10.1093/bioinformatics/bts635 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Li B, Dewey CN. RSEM: accurate transcript quantification from RNA-seq data with or without a reference genome. BMC Bioinform 2011; 12: 323. doi: 10.1186/1471-2105-12-323 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Oldham MC, Langfelder P, Horvath S. Network methods for describing sample relationships in genomic datasets: application to Huntington's disease. BMC Syst Biol 2012; 6: 63. doi: 10.1186/1752-0509-6-63 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sajuthi SP, Everman JL, Jackson ND, et al. Nasal airway transcriptome-wide association study of asthma reveals genetically driven mucus pathobiology. Nat Commun 2022; 13: 1632. doi: 10.1038/s41467-022-28973-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kolberg L, Raudvere U, Kuzmin I, et al. g:Profiler-interoperable web service for functional enrichment analysis and gene identifier mapping (2023 update). Nucleic Acids Res 2023; 51: W207–W212. doi: 10.1093/nar/gkad347 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014; 15: 550. doi: 10.1186/s13059-014-0550-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinform 2008; 9: 559. doi: 10.1186/1471-2105-9-559 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Langfelder P, Horvath S. Fast R functions for robust correlations and hierarchical clustering. J Stat Softw 2012; 46: i11. doi: 10.18637/jss.v046.i11 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Langfelder P, Luo R, Oldham MC, et al. Is my network module preserved and reproducible? PLoS Comput Biol 2011; 7: e1001057. doi: 10.1371/journal.pcbi.1001057 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lonsdale J, Thomas J, Salvatore M, et al. The Genotype-Tissue Expression (GTEx) project. Nat Genet 2013; 45: 580–585. doi: 10.1038/ng.2653 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Delahoz-Dominguez E, Zuluaga R, Fontalvo-Herrera T. Dataset of academic performance evolution for engineering students. Data Brief 2020; 30: 105537–585. doi: 10.17632/4v6yv7bsdw.1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ching T, Song MA, Tiirikainen M, et al. Genome-wide hypermethylation coupled with promoter hypomethylation in the chorioamniotic membranes of early onset pre-eclampsia. Mol Hum Reprod 2014; 20: 885–904. doi: 10.1093/molehr/gau046 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kim J, Pitlick MM, Christine PJ, et al. Genome-wide analysis of DNA methylation in human amnion. Sci World J 2013; 2013: 678156. doi: 10.1155/2013/678156 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Altman MC, Calatroni A, Ramratnam S, et al. Endotype of allergic asthma with airway obstruction in urban children. J Allergy Clin Immunol 2021; 148: 1198–1209. doi: 10.1016/j.jaci.2021.02.040 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Krauss-Etschmann S, Bush A, Bellusci S, et al. Of flies, mice and men: a systematic approach to understanding the early life origins of chronic lung disease. Thorax 2013; 68: 380–384. doi: 10.1136/thoraxjnl-2012-201902 [DOI] [PubMed] [Google Scholar]
- 38.Wada T, Adachi Y, Murakami S, et al. Maternal exposure to smoking and infant's wheeze and asthma: Japan Environment and Children's Study. Allergol Int 2021; 70: 445–451. doi: 10.1016/j.alit.2021.04.008 [DOI] [PubMed] [Google Scholar]
- 39.Zazara DE, Wegmann M, Giannou AD, et al. A prenatally disrupted airway epithelium orchestrates the fetal origin of asthma in mice. J Allergy Clin Immunol 2020; 145: 1641–1654. doi: 10.1016/j.jaci.2020.01.050 [DOI] [PubMed] [Google Scholar]
- 40.Knight DA, Holgate ST. The airway epithelium: structural and functional properties in health and disease. Respirology 2003; 8: 432–446. doi: 10.1046/j.1440-1843.2003.00493.x [DOI] [PubMed] [Google Scholar]
- 41.Bizzintino J, Lee WM, Laing IA, et al. Association between human rhinovirus C and severity of acute asthma in children. Eur Respir J 2011; 37: 1037–1042. doi: 10.1183/09031936.00092410 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Hillas J, Evens D, Hemy N, et al. Neonatal predictors of aberrant wound repair in very preterm infants. Respirology 2019; 24: 103–104. doi: 10.1111/resp.13700_1830467927 [DOI] [Google Scholar]
- 43.Kicic A, Sutanto EN, Stevens PT, et al. Intrinsic biochemical and functional differences in bronchial epithelial cells of children with asthma. Am J Respir Crit Care Med 2006; 174: 1110–1118. doi: 10.1164/rccm.200603-392OC [DOI] [PubMed] [Google Scholar]
- 44.Kicic A, Hallstrand TS, Sutanto EN, et al. Decreased fibronectin production significantly contributes to dysregulated repair of asthmatic epithelium. Am J Respir Crit Care Med 2010; 181: 889–898. doi: 10.1164/rccm.200907-1071OC [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Kicic A, Stevens PT, Sutanto EN, et al. Impaired airway epithelial cell responses from children with asthma to rhinoviral infection. Clin Exp Allergy 2016; 46: 1441–1455. doi: 10.1111/cea.12767 [DOI] [PubMed] [Google Scholar]
- 46.Khoo SK, Read J, Franks K, et al. Upper airway cell transcriptomics identify a major new immunological phenotype with strong clinical correlates in young children with acute wheezing. J Immunol 2019; 202: 1845–1858. doi: 10.4049/jimmunol.1800178 [DOI] [PubMed] [Google Scholar]
- 47.Looi K, Buckley AG, Rigby PJ, et al. Effects of human rhinovirus on epithelial barrier integrity and function in children with asthma. Clin Exp Allergy 2018; 48: 513–524. doi: 10.1111/cea.13097 [DOI] [PubMed] [Google Scholar]
- 48.Stevens PT, Kicic A, Sutanto EN, et al. Dysregulated repair in asthmatic paediatric airway epithelial cells: the role of plasminogen activator inhibitor-1. Clin Exp Allergy 2008; 38: 1901–1910. doi: 10.1111/j.1365-2222.2008.03093.x [DOI] [PubMed] [Google Scholar]
- 49.Wang WS, Lin YK, Zhang F, et al. Single cell transcriptomic analysis of human amnion identifies cell-specific signatures associated with membrane rupture and parturition. Cell Biosci 2022; 12: 64. doi: 10.1186/s13578-022-00797-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Sikkema L, Ramirez-Suastegui C, Strobl DC, et al. An integrated cell atlas of the lung in health and disease. Nat Med 2023; 29: 1563–1577. doi: 10.1038/s41591-023-02327-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated and analysed during this study are available in the Mendeley Data repository [33]. Due to ethical considerations regarding participant confidentiality, raw sequencing files cannot be made publicly available, however the deposited de-identified datasets are sufficient for reproducing all reported findings.