Chapter 3
For the first time STELA was optimised to measure placental telomeres. Telomere length has been previously measured in the placenta in relation to IUGR, GDM, PE, air pollution exposure and high maternal pre-pregnancy BMI using respectively Q-FISH, q-PCR and RT-PCR (Biron-Shental et al. 2010, 2010, 2015; Martens et al. 2016 and 2017). These methods provide relatively low- resolution compared to STELA and therefore, very subtle differences in telomere distributions may not be as obvious (Aubert et al. 2012). STELA has been successfully applied in several cellular types or tissues as depicted in Figure 3.2. An advantage of using this technique was that a more detailed information on telomere homeostasis in the human placenta was provided. As a result, remarkable heterogeneity not previously reported before was observed in the placenta using this technique. In addition, this study demonstrated that sampling site, mode of delivery or foetal sex did not affect the telomere distributions, indicating that this technique has a potentially broader application in pregnancy cohort regardless of study design. The placenta, an easily accessible foetal derived tissue, constitutes a proxy for the foetal exposures during pregnancy. Malfunction of this organ may contribute to adverse outcomes.
One limitation of this study was the sample size. The majority of the participants were from a Caucasian origin, which can be interpreted as a strength in terms of variable control. However, larger samples were ethnic diversity is greater may provide further information of the role that ethnicity plays in the risk of suffering certain pregnancy complications. Although in the results above described only the Xp and Yp chromosomal telomeres were examined, Roger et al. 2013 reported no differences in telomere length between sex and autosomal chromosomes, and thus the XpYp telomere is representative of the other chromosome ends. Another limitation is the sole use of placenta as foetal origin source. Although previous research has reported intra-individual synchrony in telomere length across different somatic tissues in the newborn (Youngren et al. 1998; Okuda et al. 2002), it would be of interest to reproduce the study also in CB as some studies have reported differences between these two tissues (Biron-
Shental 2015, 2016). It is also important to note that the heterogeneous telomere length profiles revealed in the placenta in some individuals appears as bimodal or trimodal distributions. It is therefore possible that differentiated distributions correspond to same cell type with different replicative history or different cell types with different replicative history. A culture of placenta may give the answer. It would be also of interest to perform a more detailed cell fractionation of the placental tissue prior to STELA. Finally, the analysis did not control for any variable and further statistical analysis controlling for other variables including gestational age, would be advantageous.
Chapter 4
This is the first time telomere length has been measured with respect to GDM using a high resolution technique. There is only one other study which observed telomere attrition in the placenta from babies exposed to GDM in utero (Biron- Shental et al. 2015). Other studies have explored the role of telomere length in GDM using q-PCR, Q-FISH and Flow-FISH (Hjort et al. 2018, Xu et al. 2014, Biron-Shental et al. 2016, Harville et al. 2010 and Cross et al. 2010). A strength of this study was that the groups made up of controls, untreated GDM participants and treated GDM participants only differ in gestational age and the metabolic characteristics linked to GDM. A rigorous statistical analysis identified a significant association between maternal ethnicity and telomere length. Further comparison of telomere length between the groups controlled for this variable. Another strength was that telomere length analysis was presented as average telomere length but also as a percentage of telomeres below 5 kb, covering the range of telomeres within the lengths that leads to senescence. In this study, a sexual dimorphism in telomere length in response to GDM was described. This highlights the importance of taking into account foetal sex when analysing the effect of environmental insults on newborn telomere length. Finally, placental telomeres have emerged as a sensitive tool for testing prenatal adversity and measuring effectiveness of intervention strategies, such as metformin/insulin use. Although this study found a correlation between intrauterine GDM exposure and placental telomere length, it failed to stablish causality. While it seems more
the offspring inherited shorter telomeres from mothers predisposed to develop GDM. It would be necessary to examine the telomere length in oocytes to address this question. Placental telomere attrition associated with GDM appeared to be prevented by maternal treatment with metformin and/or insulin in male placenta. This protective role may not be limited to boys which was not possible to assess in this study due to the smaller number of placental samples from girls. Another limitation is the fact that the placenta is a foetal derived tissue obtained at birth. These patterns of telomere length may not be present in the foetus, telomeres at this time has been reported as more vulnerable to being affected by several confounders compared to telomere length in the adolescence (Hjort et al. 2018). A further limitation was the small number of samples on treatment and it was not possible to explore how the action of insulin compared to metformin.
Chapter 5
This study is the first to report how prenatal depression and anxiety impact placental telomere features. To the best of our knowledge, only Gotlib et al. 2015 have reported an effect of maternal depression on offspring telomeres. Other studies have noted the impact of prenatal stress on new-born leukocyte telomere length (Entringer et al. 2011, 2013; Marchetto et al. 2016 and Send et al. 2017). A strength of this study was that female placenta exhibited shorter telomeres in female offspring exposed to maternal depression even after accounting for the potential effects of several variables. A hierarchical multiple regression that includes maternal age, gestational age, parity, smoking, alcohol consumption and WIMD made the analysis more robust. Another strength is the data available with respect to maternal salivary cortisol, mental health history and antidepressant treatment which could be factored into the analysis. In particular, it enables controlling for the variable of SSRI treatment when analysing the association between telomere length and depression. It is also important to note that association between placental telomere length and prenatal depression was assessed in two chromosomes: the sex chromosomes (XpYp) and an autosomal chromosome (17p).
Although this study found a correlation between intrauterine prenatal depression exposure and placental telomere length, it was not possible to establish causality. Whether the telomere impairment was inherited by offspring from women at risk of depression potentially contributing to depression or whether depression directly or indirectly caused the telomere dysfunction is unknown. Another limitation of this study is the use of self-reported questionnaires, the reliance on maternal self-report data on symptoms of depression and anxiety. The classification of women in depressed or not depressed can be subjective as EPDS cut-off can vary. In this regard, while using a cut-off point of ³ 13, a woman with EPDS 12 would be control, a cut off by range: 0-6, 7-12 and ³13 would categorised this woman as mild depressed. Another limitation is that no correlation was found between salivary cortisol and telomere length, indicating that other stress related physiological mechanism should be explored. There was no correlation between cortisol and maternally reported mood symptoms which suggest the relationship between maternal mood and cortisol is not a simple one. The data available with respect to antidepressants was not identical between the medical notes and the maternal questionnaires (Janssen et al. 2018). Some drugs may influence telomere length (Wolkowitz et al. 2011) and it would be important to increase sample size and validate compliance of taking the drugs in a future study. These factors mean that analysing the effect of cortisol and antidepressant use on telomere length has to be interpreted cautiously. Finally, the behaviour of sex chromosomes was different to the autosomal chromosome analysed. This may be due to differences in sample size (109 vs 60), which would mean the study of the autosome was underpowered. However, there may be inherent differences in telomere biology which would be interesting to explore in a larger dataset.