Gene markers of normal villous maturation and their expression in placentas with maturational pathology

Construction of a pathway-level model for preeclampsia based on gene expression data

Preeclampsia (PE) is a heterogeneous disease that seriously affects the health of mothers and fetuses. Lack of detection assays, its diagnosis and intervention are often delayed when the clinical symptoms are atypical. Using personalized pathway-based analysis and machine learning algorithms, we built a PE diagnosis model consisting of nine core pathways using multiple cohorts from the Gene Expression Omnibus database. The model showed an area under the receiver operating characteristic (AUROC) curve of 0.959 with the data from the placental tissue samples in the development cohort. In the two validation cohorts, the AUROCs were 0.898 and 0.876, respectively. The model also performed well with the maternal plasma data in another validation cohort (AUROC: 0.815). Moreover, we identified tyrosine-protein kinase Lck (LCK) as the hub gene in this model and found that LCK and pLCK proteins were downregulated in placentas from PE patients. The pathway-level model for PE can provide a novel direction to develop molecular diagnostic assay and investigate potential mechanisms of PE in future studies.

High-throughput mRNA-seq atlas of human placenta shows vast transcriptome remodeling from first to third trimester†

The placenta, composed of chorionic villi, changes dramatically across gestation. Understanding differences in ongoing pregnancies are essential to identify the role of chorionic villi at specific times in gestation and develop biomarkers and prognostic indicators of maternal-fetal health. The normative mRNA profile is established using next-generation sequencing of 124 first trimester and 43 third trimester human placentas from ongoing healthy pregnancies. Stably expressed genes (SEGs) not different between trimesters and with low variability are identified. Differential expression analysis of first versus third trimester adjusted for fetal sex is performed, followed by a subanalysis with 23 matched pregnancies to control for subject variability using the same genetic and environmental background. Placenta expresses 14,979 polyadenylated genes above sequencing noise (transcripts per million > 0.66), with 10.7% SEGs across gestation. Differentially expressed genes (DEGs) account for 86.7% of genes in the full cohort [false discovery rate (FDR) < 0.05]. Fold changes highly correlate between the full cohort and subanalysis (Pearson = 0.98). At stricter thresholds (FDR < 0.001, fold change > 1.5), there remains 50.1% DEGs (3353 upregulated in first and 4155 upregulated in third trimester). This is the largest mRNA atlas of healthy human placenta across gestation, controlling for genetic and environmental factors, demonstrating substantial changes from first to third trimester in chorionic villi. Specific differences and SEGs may be used to understand the specific role of the chorionic villi throughout gestation and develop first trimester biomarkers of placental health that transpire across gestation, which can be used for future development of biomarkers for maternal-fetal health.

Tissue Contamination Challenges the Credibility of Machine Learning Models in Real World Digital Pathology

Machine learning (ML) models are poised to transform surgical pathology practice. The most successful use attention mechanisms to examine whole slides, identify which areas of tissue are diagnostic, and use them to guide diagnosis. Tissue contaminants, such as floaters, represent unexpected tissue. Although human pathologists are extensively trained to consider and detect tissue contaminants, we examined their impact on ML models. We trained 4 whole-slide models. Three operate in placenta for the following functions: (1) detection of decidual arteriopathy, (2) estimation of gestational age, and (3) classification of macroscopic placental lesions. We also developed a model to detect prostate cancer in needle biopsies. We designed experiments wherein patches of contaminant tissue are randomly sampled from known slides and digitally added to patient slides and measured model performance. We measured the proportion of attention given to contaminants and examined the impact of contaminants in the t-distributed stochastic neighbor embedding feature space. Every model showed performance degradation in response to one or more tissue contaminants. Decidual arteriopathy detection--balanced accuracy decreased from 0.74 to 0.69 ± 0.01 with addition of 1 patch of prostate tissue for every 100 patches of placenta (1% contaminant). Bladder, added at 10% contaminant, raised the mean absolute error in estimating gestational age from 1.626 weeks to 2.371 ± 0.003 weeks. Blood, incorporated into placental sections, induced false-negative diagnoses of intervillous thrombi. Addition of bladder to prostate cancer needle biopsies induced false positives, a selection of high-attention patches, representing 0.033 mm2, and resulted in a 97% false-positive rate when added to needle biopsies. Contaminant patches received attention at or above the rate of the average patch of patient tissue. Tissue contaminants induce errors in modern ML models. The high level of attention given to contaminants indicates a failure to encode biological phenomena. Practitioners should move to quantify and ameliorate this problem.

High-throughput mRNA-seq atlas of human placenta shows vast transcriptome remodeling from first to third trimester

Background

The placenta, composed of chorionic villi, changes dramatically across gestation. Understanding differences in ongoing pregnancies are essential to identify the role of chorionic villi at specific times in gestation and develop biomarkers and prognostic indicators of maternal- fetal health.

Methods

The normative mRNA profile is established using next-generation sequencing of 124 first trimester and 43 third trimester human placentas from ongoing healthy pregnancies. Stably expressed genes not different between trimesters and with low variability are identified. Differential expression analysis of first versus third trimester adjusted for fetal sex is performed, followed by a subanalysis with 23 matched pregnancies to control for subject variability using the same genetic and environmental background.

Results

Placenta expresses 14,979 mRNAs above sequencing noise (TPM>0.66), with 1,545 stably expressed genes across gestation. Differentially expressed genes account for 86.7% of genes in the full cohort (FDR<0.05). Fold changes highly correlate between the full cohort and subanalysis (Pearson = 0.98). At stricter thresholds (FDR<0.001, fold change>1.5), there are 6,941 differentially expressed protein coding genes (3,206 upregulated in first and 3,735 upregulated in third trimester).

Conclusion

This is the largest mRNA atlas of healthy human placenta across gestation, controlling for genetic and environmental factors, demonstrating substantial changes from first to third trimester in chorionic villi. Specific differences and stably expressed genes may be used to understand the specific role of the chorionic villi throughout gestation and develop first trimester biomarkers of placental health that transpire across gestation, which can be used for future development of biomarkers in maternal-fetal disease.

Tissue contamination challenges the credibility of machine learning models in real world digital pathology

Machine learning (ML) models are poised to transform surgical pathology practice. The most successful use attention mechanisms to examine whole slides, identify which areas of tissue are diagnostic, and use them to guide diagnosis. Tissue contaminants, such as floaters, represent unexpected tissue. While human pathologists are extensively trained to consider and detect tissue contaminants, we examined their impact on ML models. We trained 4 whole slide models. Three operate in placenta for 1) detection of decidual arteriopathy (DA), 2) estimation of gestational age (GA), and 3) classification of macroscopic placental lesions. We also developed a model to detect prostate cancer in needle biopsies. We designed experiments wherein patches of contaminant tissue are randomly sampled from known slides and digitally added to patient slides and measured model performance. We measured the proportion of attention given to contaminants and examined the impact of contaminants in T-distributed Stochastic Neighbor Embedding (tSNE) feature space. Every model showed performance degradation in response to one or more tissue contaminants. DA detection balanced accuracy decreased from 0.74 to 0.69 +/- 0.01 with addition of 1 patch of prostate tissue for every 100 patches of placenta (1% contaminant). Bladder, added at 10% contaminant raised the mean absolute error in estimating gestation age from 1.626 weeks to 2.371 +/ 0.003 weeks. Blood, incorporated into placental sections, induced false negative diagnoses of intervillous thrombi. Addition of bladder to prostate cancer needle biopsies induced false positives, a selection of high-attention patches, representing 0.033mm2, resulted in a 97% false positive rate when added to needle biopsies. Contaminant patches received attention at or above the rate of the average patch of patient tissue. Tissue contaminants induce errors in modern ML models. The high level of attention given to contaminants indicates a failure to encode biological phenomena. Practitioners should move to quantify and ameliorate this problem.

Histopathologic and Transcriptomic Profiling Identifies Novel Trophoblast Defects in Patients With Preeclampsia and Maternal Vascular Malperfusion

Preeclampsia (PE) is a heterogeneous disease for which the current clinical classification system is based on the presence or absence of specific clinical features. PE-associated placentas also show heterogeneous findings on pathologic examination, suggesting that further subclassification is possible. We combined clinical, pathologic, immunohistochemical, and transcriptomic profiling of placentas to develop integrated signatures for multiple subclasses of PE. In total, 303 PE and 1388 nonhypertensive control placentas were included. We found that maternal vascular malperfusion (MVM) in the placenta was associated with preterm PE with severe features and with small-for-gestational-age neonates. Interestingly, PE placentas with either MVM or no histologic pattern of injury showed a linear decrease in proliferative (p63+) cytotrophoblast per villous area with increasing gestational age, similar to placentas obtained from the nonhypertensive patient cohort; however, PE placentas with fetal vascular malperfusion or villitis of unknown etiology lost this phenotype. This is mainly because of cases of fetal vascular malperfusion in placentas of patients with preterm PE and villitis of unknown etiology in placentas of patients with term PE, which are associated with a decrease or increase, respectively, in the cytotrophoblast per villous area. Finally, a transcriptomic analysis identified pathways associated with hypoxia, inflammation, and reduced cell proliferation in PE-MVM placentas and further subclassified this group into extravillous trophoblast-high and extravillous trophoblast-low PE, confirmed using an immunohistochemical analysis of trophoblast lineage-specific markers. Our findings suggest that within specific histopathologic patterns of placental injury, PE can be subclassified based on specific cellular and molecular defects, allowing the identification of pathways that may be targeted for diagnostic and therapeutic purposes.

Epigenetic clocks and female fertility timeline: A new approach to an old issue?

Worldwide increase in life expectancy has boosted research on aging. Overcoming the concept of chronological age, higher attention has been addressed to biological age, which reflects a person's real health state, and which may be the resulting combination of both intrinsic and environmental factors. As epigenetics may exert a pivotal role in the biological aging, epigenetic clocks were developed. They are based on mathematical models aimed at identifying DNA methylation patterns that can define the biological age and that can be adopted for different clinical scopes (i.e., estimation of the risks of developing age-related disorders or predicting lifespan). Recently, epigenetic clocks have gained a peculiar attention in the fertility research field, in particular in the female counterpart. The insight into the possible relations between epigenetic aging and women's infertility might glean additional information about certain conditions that are still not completely understood. Moreover, they could disclose significant implications for health promotion programs in infertile women. Of relevance here is that the impact of biological age and epigenetics may not be limited to fertility status but could translate into pregnancy issues. Indeed, epigenetic alterations of the mother may transfer into the offspring, and pregnancy itself as well as related complications could contribute to epigenetic modifications in both the mother and newborn. However, even if the growing interest has culminated in the conspicuous production of studies on these topics, a global overview and the availability of validated instruments for diagnosis is still missing. The present narrative review aims to explore the possible bonds between epigenetic aging and fertility timeline. In the "infertility" section, we will discuss the advances on epigenetic clocks focusing on the different tissues examined (endometrium, peripheral blood, ovaries). In the "pregnancy" section, we will discuss the results obtained from placenta, umbilical cord and peripheral blood. The possible role of epigenetic aging on infertility mechanisms and pregnancy outcomes represents a question that may configure epigenetic clock as a bond between two apparently opposite worlds: infertility and pregnancy.

GestAltNet: aggregation and attention to improve deep learning of gestational age from placental whole-slide images

The placenta is the first organ to form and performs the functions of the lung, gut, kidney, and endocrine systems. Abnormalities in the placenta cause or reflect most abnormalities in gestation and can have life-long consequences for the mother and infant. Placental villi undergo a complex but reproducible sequence of maturation across the third-trimester. Abnormalities of villous maturation are a feature of gestational diabetes and preeclampsia, among others, but there is significant interobserver variability in their diagnosis. Machine learning has emerged as a powerful tool for research in pathology. To capture the volume of data and manage heterogeneity within the placenta, we developed GestaltNet, which emulates human attention to high-yield areas and aggregation across regions. We used this network to estimate the gestational age (GA) of scanned placental slides and compared it to a baseline model lacking the attention and aggregation functions. In the test set, GestaltNet showed a higher r2 (0.9444 vs. 0.9220) than the baseline model. The mean absolute error (MAE) between the estimated and actual GA was also better in the GestaltNet (1.0847 weeks vs. 1.4505 weeks). On whole-slide images, we found the attention sub-network discriminates areas of terminal villi from other placental structures. Using this behavior, we estimated GA for 36 whole slides not previously seen by the model. In this task, similar to that faced by human pathologists, the model showed an r2 of 0.8859 with an MAE of 1.3671 weeks. We show that villous maturation is machine-recognizable. Machine-estimated GA could be useful when GA is unknown or to study abnormalities of villous maturation, including those in gestational diabetes or preeclampsia. GestaltNet points toward a future of genuinely whole-slide digital pathology by incorporating human-like behaviors of attention and aggregation.

Current approaches and developments in transcript profiling of the human placenta

BACKGROUND

The placenta is the active interface between mother and foetus, bearing the molecular marks of rapid development and exposures in utero. The placenta is routinely discarded at delivery, providing a valuable resource to explore maternal-offspring health and disease in pregnancy. Genome-wide profiling of the human placental transcriptome provides an unbiased approach to study normal maternal–placental–foetal physiology and pathologies.

OBJECTIVE AND RATIONALE

To date, many studies have examined the human placental transcriptome, but often within a narrow focus. This review aims to provide a comprehensive overview of human placental transcriptome studies, encompassing those from the cellular to tissue levels and contextualize current findings from a broader perspective. We have consolidated studies into overarching themes, summarized key research findings and addressed important considerations in study design, as a means to promote wider data sharing and support larger meta-analysis of already available data and greater collaboration between researchers in order to fully capitalize on the potential of transcript profiling in future studies.

SEARCH METHODS

The PubMed database, National Center for Biotechnology Information and European Bioinformatics Institute dataset repositories were searched, to identify all relevant human studies using ‘placenta’, ‘decidua’, ‘trophoblast’, ‘transcriptome’, ‘microarray’ and ‘RNA sequencing’ as search terms until May 2019. Additional studies were found from bibliographies of identified studies.

OUTCOMES

The 179 identified studies were classifiable into four broad themes: healthy placental development, pregnancy complications, exposures during pregnancy and in vitro placental cultures. The median sample size was 13 (interquartile range 8–29). Transcriptome studies prior to 2015 were predominantly performed using microarrays, while RNA sequencing became the preferred choice in more recent studies. Development of fluidics technology, combined with RNA sequencing, has enabled transcript profiles to be generated of single cells throughout pregnancy, in contrast to previous studies relying on isolated cells. There are several key study aspects, such as sample selection criteria, sample processing and data analysis methods that may represent pitfalls and limitations, which need to be carefully considered as they influence interpretation of findings and conclusions. Furthermore, several areas of growing importance, such as maternal mental health and maternal obesity are understudied and the profiling of placentas from these conditions should be prioritized.

WIDER IMPLICATIONS

Integrative analysis of placental transcriptomics with other ‘omics’ (methylome, proteome and metabolome) and linkage with future outcomes from longitudinal studies is crucial in enhancing knowledge of healthy placental development and function, and in enabling the underlying causal mechanisms of pregnancy complications to be identified. Such understanding could help in predicting risk of future adversity and in designing interventions that can improve the health outcomes of both mothers and their offspring. Wider collaboration and sharing of placental transcriptome data, overcoming the challenges in obtaining sufficient numbers of quality samples with well-defined clinical characteristics, and dedication of resources to understudied areas of pregnancy will undoubtedly help drive the field forward.

Integrated Learning: Screening Optimal Biomarkers for Identifying Preeclampsia in Placental mRNA Samples

Preeclampsia (PE) is a maternal disease that causes maternal and child death. Treatment and preventive measures are not sound enough. The problem of PE screening has attracted much attention. The purpose of this study is to screen placental mRNA to obtain the best PE biomarkers for identifying patients with PE. We use Limma in the R language to screen out the 48 differentially expressed genes with the largest differences and used correlation-based feature selection algorithms to reduce the dimensionality and avoid attribute redundancy arising from too many mRNA samples participating in the classification. After reducing the mRNA attributes, the mRNA samples are sorted from large to small according to information gain. In this study, a classifier model is designed to identify whether samples had PE through mRNA in the placenta. To improve the accuracy of classification and avoid overfitting, three classifiers, including C4.5, AdaBoost, and multilayer perceptron, are used. We use the majority voting strategy integrated with the differentially expressed genes and the genes filtered by the best subset method as comparison methods to train the classifier. The results show that the classification accuracy rate has increased from 79% to 82.2%, and the number of mRNA features has decreased from 48 to 13. This study provides clues for the main PE biomarkers of mRNA in the placenta and provides ideas for the treatment and screening of PE.

Disorders of placental villous maturation in fetal death

Objective The aims of this study were to ascertain the frequency of disorders of villous maturation in fetal death and to also delineate other placental histopathologic lesions in fetal death. Methods This was a retrospective observational cohort study of fetal deaths occurring among women between January 2004 and January 2016 at Hutzel Women's Hospital, Detroit, MI, USA. Cases comprised fetuses with death beyond 20 weeks' gestation. Fetal deaths with congenital anomalies and multiple gestations were excluded. Controls included pregnant women without medical/obstetrical complications and delivered singleton, term (37-42 weeks) neonate with 5-min Apgar score ≥7 and birthweight between the 10th and 90th percentiles. Results Ninety-two percent (132/143) of placentas with fetal death showed placental histologic lesions. Fetal deaths were associated with (1) higher frequency of disorders of villous maturation [44.0% (64/143) vs. 1.0% (4/405), P < 0.0001, prevalence ratio, 44.6; delayed villous maturation, 22% (31/143); accelerated villous maturation, 20% (28/143); and maturation arrest, 4% (5/143)]; (2) higher frequency of maternal vascular malperfusion lesions [75.5% (108/143) vs. 35.7% (337/944), P < 0.0001, prevalence ratio, 2.1] and fetal vascular malperfusion lesions [88.1% (126/143) vs. 19.7% (186/944), P < 0.0001, prevalence ratio, 4.5]; (3) higher frequency of placental histologic patterns suggestive of hypoxia [59.0% (85/143) vs. 9.3% (82/942), P < 0.0001, prevalence ratio, 6.8]; and (4) higher frequency of chronic inflammatory lesions [53.1% (76/143) vs. 29.9% (282/944), P < 0.001, prevalence ratio 1.8]. Conclusion This study demonstrates that placentas of women with fetal death were 44 times more likely to present disorders of villous maturation compared to placentas of those with normal pregnancy. This suggests that the burden of placental disorders of villous maturation lesions is substantial.

Long noncoding RNA LRRC75A-AS1 inhibits cell proliferation and migration in colorectal carcinoma

With the continuous improvement of technology in the molecular field, more and more evidence indicated that long noncoding RNAs (lncRNAs) are widely expressed in a broad spectrum of human tumors, playing an important role in the development and progression of tumors. Most studies reported that lncRNAs might serve as reliable biomarkers and effective clinical therapeutic target. Leucine-rich repeat containing 75 A-antisense RNA1 (LRRC75A-AS1) was reported to be relevant to many types of cancers and indicated to do influence on colorectal carcinoma (CRC). This research firstly examined the role of LRRC75A-AS1 in CRC and analyzed its association with the biological behaviors of CRC cells. Quantitative reverse transcription polymerase chain reaction (qRT-PCR) brought to light that LRRC75A-AS1 was remarkably expressed at low levels in CRC tissues. We also found that LRRC75A-AS1 was localized in the cytoplasm. In addition, LRRC75A-AS1 knockdown also notably promoted CRC cell proliferation, metastasis, invasion, and colony formation. To summarize, these experimental results showed that LRRC75A-AS1 might serve as an anti-oncogene for CRC tumorigenesis and advancement, and it may become a novel molecular marker for clinical diagnosis.

Impact statement

It is reported that colorectal cancer has seriously threatened human health. The incidence of colorectal cancer in China is increasing year by year. At present, the treatment of cancer is gradually developing towards individualized treatment whose core is targeted therapy, and molecular pathology is the basis of targeted therapy. Previous studies have shown that in addition to protein-coding genes that regulate tumor invasion and metastasis, there are also some non-coding genes involved in tumor encroachment and spread. Our study found that long noncoding RNA LRRC75A-AS1 is closely related to CRC and is related to its proliferation and migration. And it may become a novel molecular marker for clinical diagnosis and treatment.

Placental epigenetic clocks: estimating gestational age using placental DNA methylation levels

The human pan-tissue epigenetic clock is widely used for estimating age across the entire lifespan, but it does not lend itself well to estimating gestational age (GA) based on placental DNAm methylation (DNAm) data. We replicate previous findings demonstrating a strong correlation between GA and genome-wide DNAm changes. Using substantially more DNAm arrays (n=1,102 in the training set) than a previous study, we present three new placental epigenetic clocks: 1) a robust placental clock (RPC) which is unaffected by common pregnancy complications (e.g., gestational diabetes, preeclampsia), and 2) a control placental clock (CPC) constructed using placental samples from pregnancies without known placental pathology, and 3) a refined RPC for uncomplicated term pregnancies. These placental clocks are highly accurate estimators of GA based on placental tissue; e.g., predicted GA based on RPC is highly correlated with actual GA (r>0.95 in test data, median error less than one week). We show that epigenetic clocks derived from cord blood or other tissues do not accurately estimate GA in placental samples. While fundamentally different from Horvath's pan-tissue epigenetic clock, placental clocks closely track fetal age during development and may have interesting applications.

Long noncoding RNAs in vascular smooth muscle cells regulate vascular calcification

Vascular calcification is characterized by the accumulation of hydroxyapatite crystals, which is a result of aberrant mineral metabolism. Although many clinical studies have reported its adverse effects on cardiovascular morbidity, the molecular mechanism of vascular calcification, especially the involvement of long noncoding RNAs (lncRNAs), is not yet reported. From the transcriptomic analysis, we discovered hundreds of lncRNAs differentially expressed in rat vascular smooth muscle cells (VSMCs) treated with inorganic phosphate, which mimics vascular calcification. We focused on Lrrc75a-as1 and elucidated its transcript structure and confirmed its cytoplasmic localization. Our results showed that calcium deposition was elevated after knockdown of Lrrc75a-as1, while its overexpression inhibited calcium accumulation in A10 cells. In addition, Lrrc75a-as1 attenuated VSMCs calcification by decreasing the expression of osteoblast-related factors. These findings suggest that Lrrc75a-as1 acts as a negative regulator of vascular calcification, and may serve as a possible therapeutic target in vascular calcification.

Gestational diabetes mellitus is associated with increased CD163 expression and iron storage in the placenta

PROBLEM

GDM has been associated with disturbances in iron homeostasis and exaggerated immune activation. We sought to investigate the extent to which placental iron storage and macrophage accumulations were altered in GDM.

METHOD OF STUDY

We conducted a retrospective, case-control study of archived placental tissues obtained from 22 pregnancies complicated by GDM and 22 unaffected controls. Controls were matched to cases based on maternal age, gestational age at birth, and method of delivery. Placental tissues were assessed for altered histology and CD68 and CD163 staining. Tissue iron was assessed using Prussian blue staining.

RESULTS

Maternal hematocrit levels were higher in GDM participants compared to controls (P = 0.02). The presence of meconium-laden macrophages was significantly greater within the amnion of GDM cases (adjusted odds ratio (OR) 12.51). Although the total abundance of CD68-expressing macrophages was not significantly different between groups, we detected a significantly greater abundance of CD163 expression within the chorion and decidua of cases. The total area staining positive for iron was 24% (95% confidence intervals of 2%-46%) greater in GDM placentae versus controls.

CONCLUSION

GDM is associated with altered placental histology and increases in meconium-laden macrophages. Greater iron stores within the placentae of women with GDM is consistent with reports that iron excess is associated with an increased risk for GDM. The higher level of expression of CD163 on macrophage-like cells of the chorion and decidua in GDM suggests an increase in M2-like macrophages. Overall, our results add to growing evidence that GDM has direct effects on placental structure.

Re-view and view on maturation disorders in the placenta

Until delivery, the placenta plays an important mediator role between mother and fetus. This unit is affected by peristatic conditions, such as acute or chronic maternal diseases, malnutrition, drugs, and others. But also genetic factors and fetal malformations due to embryonic developmental disorders may contribute to macroscopically visible changes and functional disorders of the placenta. In a constantly ongoing progress of maturation, the placenta records and saves changes due to fetal distress partly as maturation disorders. Understanding of maturation disorders might, therefore, be an important contribution to a better understanding of influences on villous differentiation and might improve follow up and fetal outcome to reduce recurrence risk. However, an internationally unified classification system of maturation disorders does not exist. In this review, terminology, trials, and classifications of villous maturation disorders are summed up and compared, to pinpoint the need of agreement on an international unified and reproducible classification of maturation disorders.