Smoking during pregnancy increases chemerin expression in neonatal tissue

Abstract

It has been shown in utero tobacco exposure increases offspring risk for obesity, but the mechanisms responsible for this increased risk are not well understood. Chemerin is an adipokine that regulates adipocyte differentiation. This chemokine is elevated in obese individuals and with smoke exposure, but its levels have not been measured in neonates exposed to cigarette smoke in utero. We examined chemerin gene expression [n = 31 non-smoker (NS) and 15 smoker (S)] and DNA methylation (n = 28 NS and n = 11 S) in skin collected from babies born to mothers who smoked during pregnancy as compared to non–smoking controls. Quality RNA and DNA were isolated from foreskin tissue following circumcision, and chemerin gene expression and DNA methylation were assessed. Further, in a second cohort, we utilized primary dermal foreskin fibroblasts as a functional measure of adipogenesis in living cells (n = 11 NS and n = 8 S). Cells were stimulated with an adipogenic cocktail, mRNA was isolated from cells after 14 days, and chemerin gene expression assessed via real-time PCR. Chemerin mRNA was elevated in both whole tissue (NS: 2409.20±555.28 counts and S: 2966.72±636.84 counts) and primary fibroblasts (NS: 1.12±0.55 2^ΔΔct and S: 2.13±1.34 2^ΔΔct) collected from infants born to smoking mothers. Chemerin DNA methylation was reduced in whole tissue of offspring born to smokers (NS: 4.18±1.28 and S: 3.07±1.31), which may contribute to the increased gene expression. Neonates born to mothers who smoke during pregnancy exhibit distinct changes in chemerin gene expression in response to in utero tobacco smoke exposure which are regulated in part by epigenetic alterations.

Introduction:

Nearly 35% of all American adults and roughly 20% of children between the ages of 6–19 years old are obese (Ogden et al. 2014). It is estimated that $200 billion is spent annually on obesity-related healthcare costs (Cawley & Meyerhoefer 2012). Thus, understanding how to prevent and treat obesity and obesity-related diseases is crucial in stalling or reversing this epidemic. Though multiple factors play a part in the development of obesity and metabolic disorders, one potential contribution is the in utero environment during pregnancy. The developmental origins of health and disease hypothesis posits that the perinatal environment can potentially lead to long term disease risk in offspring (Wadhwa et al. 2009). Specific environmental exposures during fetal development, such as tobacco smoke, negatively impact offspring health.

It has been consistently shown that tobacco smoke exposure during fetal development increases the risk of pediatric and adult offspring obesity and adult offspring type 2 diabetes (T2D)(Cupul-Uicab et al. 2012; Ino 2010; Power & Jefferis 2002; Thomas et al. 2007). Despite this knowledge, 15–18% of pregnant women in the United States continue to smoke throughout pregnancy and nursing (Results from the 2013 National Survey on Drug Use and Health: Summary of National Findings 2013). However, many unanswered questions remain in humans about the mechanisms that are driving obesity and diabetes risk in offspring exposed to tobacco smoke in utero.

Chemerin is an inflammatory adipokine that has been implicated in adipocyte differentiation (Goralski et al. 2007) and its protein levels are found to be elevated in serum of obese individuals (Shin et al. 2012). However, Shin et al. (Shin et al. 2012) and Hatziagelaki et al. (Hatziagelaki et al. 2015) demonstrate a weak association between serum chemerin concentrations and body mass index. Thus, the exact role of chemerin in the pathophysiology of obesity remains unclear. Chemerin is a ligand for the G-protein coupled receptor, chemokine like receptor 1 (CMKLR1), which is highly expressed in adipocytes (Goralski et al. 2007). Chemerin protein expression has been shown to be elevated in the broncho alveolar lavage fluid of mice exposed to cigarette smoke (Demoor et al. 2011), but whether this translates into elevated chemerin expression in human smokers or in offspring born to smoking mothers is not yet known. These attributes make chemerin a reasonable biomarker that could be associated with in utero smoke exposure and may explain why babies born to mothers who smoke during pregnancy are at an increased risk of developing obesity later in life. Understanding potential mechanisms that are altered in offspring exposed to tobacco smoke during development may provide potential pharmacological targets for therapy.

Foreskin tissue has been used previously as a surrogate tissue to study a number of cellular adaptations including wound healing, developmental abnormalities, inflammation, insulin signaling, and oxidative stress (De Corte et al. 2012; Mendez et al. 1999; Qiao et al. 2012; Reynolds, Dickens, et al. 2017; Reynolds, Pollack, et al. 2017; Vottero et al. 2011). Thus, this tissue provides an opportunity to study developmental programming utilizing a neonatal tissue that comes directly from the infant (Reynolds, Dickens, et al. 2017). The purpose of this study was to examine if chemerin gene expression was altered in neonatal tissue of babies born to smokers compared to non-smoking mothers, and to examine any epigenetic modifications that may contribute to differential chemerin expression.

Methods:

Ethical Approval:

Experiments were approved by the University of Kentucky Institutional Review Board (approval reference numbers 15-0197-F3R and 12-0111-P2H). Written informed consent was obtained on all subjects prior to data and sample collection. The experiments conform to the standards set by the Declaration of Helsinki, except for registration in a database.

Subjects:

Women at the University of Kentucky (UK) Chandler Hospital, Labor and Delivery Unit were recruited into the study post-partum. A random population of women from 2 different cohorts was consented from 2012 to 2013 (Cohort 1) and from 2015–2016 (Cohort 2). Subjects were then identified as smokers or non-smokers by self-report. All women in the study reported smoking <1 pack of cigarettes/day, with the exception of 1 woman who smoked 2 packs of cigarettes/day. Inclusion criteria for this study included: full term gestation (≥ 37 weeks) at delivery (vaginal or cesarean), a non-anomalous newborn, singleton male infants with the circumcision performed < 72 hours after birth. Only English speaking mothers were consented for the study.

Experimental Design:

Whole Tissue Experiments:

For the whole tissue experiments (cohort 1), foreskins (n=49) from babies were collected following circumcision and grossly dissected to separate the dartos (hypodermis) from the epidermal/dermal layers. Samples were snap frozen and stored at −80°C until analysis. We previously published on the effects of birthweight on gene expression in the dartos layer of the foreskin (Reynolds, Pollack, et al. 2017). As part of the present study, we utilized the dermal/epidermal layers for analyses. Three out of 49 samples collected were not utilized from cohort 1 due to degraded RNA. Fifteen of the remaining 46 samples were from newborns whose mothers smoked during pregnancy. Chemerin mRNA expression was analyzed in all 46 samples (31 non-smoker and 15 smoker) while DNA methylation analyses of chemerin (28 non-smoker and 11 smoker) and LINE1 (12 non-smoker and 7 smoker), a global measure of DNA methylation, were performed in a subset of the 46 samples analyzed for mRNA expression due to limited tissue availability.

NanoString:

RNA levels were quantified in the dermal/epidermal layer by NanoString as previously described by our lab where we analyzed the dartos (hypodermal) layer of the foreskin (Reynolds, Pollack, et al. 2017). The dermis and epidermis cannot be mechanically separated at the time of collection so were analyzed together. Briefly, 500 ng RNA was loaded per sample for each NanoString run. Results were normalized for positive controls and pre-selected housekeeping genes (the geometric mean was calculated for 13 housekeeping genes for each sample). The NanoString nCounter system is highly reproducible and provides similar expression patterns to real-time qPCR (Geiss et al. 2008).

DNA Methylation: Nucleic Acid Extraction and Bisulfite Modification:

DNA was extracted from whole tissues using the Qiagen DNeasy Blood and Tissue Kit (Qiagen, Inc., Valencia, CA). Purified DNA was quantified using a Qubit 3.0 Fluorometer (Thermo Fisher Scientific Inc., Waltham, MA), and DNA samples (500 ng) were bisulfite-modified using the EZ DNA Methylation Kit (Zymo Research, Irvine, CA, USA.) and stored at −20°C. Samples were randomized for analysis.

Bisulfite pyrosequencing DNA methylation analysis:

Chemerin pyrosequencing assay was conducted with Qiagen Pyromark CpG assay Hs_RARRES2_02_PM (PM00029904). Additionally, we utilized a LINE-1 pyrosequencing assay as a surrogate marker for global genomic DNA methylation. LINE-1 sequences constitute >15% of the human genome and are highly repeated human retrotransposon sequences. Repetitive elements make up roughly one-third of genomic methylation. Pyrosequencing primers for LINE-1 were as follows: Forward primer 5’-TTTTGAGTTAGGTGTGGGATATA-3’; Reverse primer 5’-AAAATCAAAAAATTCCCTTTC-3’; Sequencing primer 5’-AGTTAGGTGTGGGATATAGT-3’. Samples were run in triplicate and data points were rejected if they were greater or less than 2 times the standard deviation of the mean of the triplicate group. DNA methylation across all CpG sites within the LINE-1 assay was averaged to obtain average percent methylation for analysis.

Cell Culture Experiments:

Foreskins were collected from infants from an additional population (cohort 2) of smoking and non-smoking mothers as described above. From the 24 samples collected for cohort 2, 13 were from newborns to smoking mothers. However, only 8 of these samples were utilized in the data analysis due to cell contamination or lack of cell growth in cohort 2.

Cells were plated as previously described by our lab (Reynolds, Dickens, et al. 2017). Briefly, the dartos layer was grossly dissected away from the dermal/epidermal layer. The dermal/epidermal layer was then immediately put in phosphate buffered saline (1x) and transported to the laboratory. Under sterile conditions, the epidermal/dermal layer was incubated in 5 mL dispase II solution (25 U/mL) (Life Technologies, Carlsbad, CA) for 16–20 hours at 4 degrees Celsius. The epidermis was then separated from the dermal layer and discarded. The dermal layer was minced and homogenized in 10 mL collagenase type IV solution (1500 U/mL) (Life Technologies, Carlsbad, CA) for 1 hour. The cells were then pelleted by centrifugation, the collagenase aspirated, and cells re-suspended in media (Gibco (Thermo Fisher Scientific) Gaithersburg, MD, catalogue #10569–010) supplemented with 10% fetal bovine serum and gentamicin/amphotericin (Life Technologies, Carlsbad, CA). The cells were filtered onto a plate and additional media was added if necessary. Media was changed 24 hours after plating and every 48 hours following. Once the cells reached 80% confluency they were passaged onto a 12-well plate for adipogenesis experiments.

Adipogenesis:

Primary dermal fibroblasts from newborns of smoking and non-smoking mothers were plated onto 12-well plates. The following adipogenesis protocol was implemented to induce adipocyte differentiation as previously described by our lab (Reynolds, Dickens, et al. 2017). Forty-eight hours post-confluency, the cells were induced in a cocktail of media (Gibco (Thermo Fisher Scientific), Gaithersburg, MD, catalogue #10569–010), 10% fetal bovine serum, gentamicin/amphotericin, 1 µM dexamethasone, 0.5 mM 3-isobutyl-1-methylxanthine, 10 µg/mL insulin and 1.0 µM rosiglitazone for 3 days. Insulin (10 µg/mL), rosiglitazone (1.0 µM), and cell media were refreshed every other day for an additional 11 days. RNA was collected and isolated using standard procedures from the Qiagen RNeasy kit (“RNeasy Mini Handbook” 2016). Chemerin gene expression was assessed via qPCR using the Step One Plus Real-Time PCR System (Applied Biosystems, Life Technologies, Carlsbad, CA). 20 ng cDNA per reaction was used with chemerin TaqMan Probes (Applied Biosystems, Life Technologies, Carlsbad, CA). Tubulin, beta class I (TUBB) was selected as the housekeeping gene. Data are reported as 2^ΔΔCt.

Statistics:

Unpaired t-tests were performed on maternal and infant characteristics listed in Table 1 and 2 and chemerin mRNA (Figures 1A and 2), chemerin DNA methylation (Figure 1B) and LINE1 DNA methylation (Figure 1D). The pre-pregnancy BMI data in Cohort 1 (Table 1) were not normally distributed. Thus, a Mann-Whitney Rank Sum Test was performed. Pearson’s correlation was performed on the chemerin DNA methylation and chemerin mRNA in Figure 1C. Data are presented as mean ± S.D.

Table 1.

Maternal and Infant Characteristics for Study 1

Characteristics	No Smoke	Smoke	P value
Maternal
N	31	15
Age (years)	28±4.47	28±4.68	0.61
Pre-Pregnancy BMI (kg/m2)
Overall Average	25±5.84	27±7.05	0.15
<18 kg/m2 (N)	1	1
18.1–24.9 kg/m2 (N)	19	5
25–29.9 kg/m2 (N)	4	3
>30 kg/m2 (N)	7	6
Infant
Birth Weight (g)	3509±383.92	3014±393.71	<0.01
Birth Length (cm)	52±2.03	49±2.34	<0.01

BMI:Body Mass Index

N=Sample Size

Table 2.

Maternal and Infant Characteristics for Study 2

Characteristics	No Smoke	Smoke	P value
Maternal
N	11	8
Age (years)	28±7.45	25±3.62	0.27
Pre-Pregnancy BMI (kg/m2)
Overall Average	26±5.86	24±6.10	0.62
<18 kg/m2 (N)	0	1
18.1–24.9 kg/m2 (N)	6	3
25–29.9 kg/m2 (N)	3	3
>30 kg/m2 (N)	2	1
Infant
Birth Weight (g)	3459±526.93	2909±228.04	0.01
Birth Length (cm)	51±2.27	48±2.33	0.01

BMI:Body Mass Index

N=Sample Size

Figure 1.

Chemerin expression in the epidermal/dermal layer of the foreskin of babies born to non-smokers and smokers. A. Chemerin mRNA expression in babies not exposed or exposed to in utero cigarette smoke (non-smoking and smoking, respectively) assessed via NanoString, B. Chemerin DNA methylation of CpG3 site in babies not exposed or exposed to in utero cigarette smoke, C. Correlation between chemerin DNA methylation expressed in percentage and mRNA expression expressed in counts in all samples from panels 1A and 1B, and D. LINE1 DNA methylation in babies not exposed or exposed to in utero cigarette smoke.

Figure 2.

Chemerin mRNA expression in primary foreskin fibroblasts isolated from babies born to non-smokers or smokers.

Results:

Maternal Characteristics:

Maternal characteristics of mother/infant pairs utilized in the study are listed in Table 1 and 2. Maternal age and pre-pregnancy BMI were not different between the smoker and non-smoker groups in cohort 1 or 2 (p>0.05); however, in both cohorts infant birth weight and length were significantly reduced in the infants exposed in utero to cigarette smoke (p<0.05).

Whole Tissue Experiments:

Whole tissue from babies exposed in utero to cigarette smoke demonstrated increased chemerin gene expression (Figure 1A). The geometric mean of the 13 housekeeping genes used was not significantly different (NS: 14880.90±1148.46 counts and S: 14464.48±931.65 counts, p>0.05). Chemerin CpG methylation averaged across all sites examined appeared reduced amongst in utero smoke exposed infants (p=0.073, data not shown), with CpG site 3 (chr7:150038291 (in Ensembl Release 75 GRCh37)) specifically demonstrating a significant reduction of methylation (Figure 1B) (p<0.05). CpG site 1 (Non-Smoking: 7.57±1.30, Smoking: 7.22±1.04) and site 2 (Non-Smoking: 10.67±2.42, Smoking: 10.22±1.33) did not show statistical significance (p>0.05). Chemerin DNA methylation at site 3 was significantly and negatively correlated to chemerin mRNA expression (Figure 1C) (p<0.05), which is not due to gross alterations of DNA methylation as LINE1 DNA methylation, a marker of global genomic methylation, was not significantly different between the two groups (Figure 1D) (p>0.05).

Cell Culture Experiments:

Primary dermal fibroblasts were grown in culture and stimulated with an adipogenic cocktail. Cells that were collected from babies born to smokers demonstrated elevated chemerin mRNA expression compared to those cells isolated from babies born to non-smokers (Figure 2) (p<0.05). Of note, cycle counts of the housekeeping gene, TUBB, were not significantly different between the Non-Smoking (26.83±0.80) and Smoking (26.92±0.46) groups (p>0.05).

Discussion:

Our results suggest that in utero cigarette smoke exposure may contribute to increased chemerin gene expression in whole tissue and primary cells collected from neonates. These data also suggest its increased expression, may be, in part, epigenetically regulated as we saw a decrease in chemerin DNA methylation at the CpG3 site in whole tissues of newborns born to mothers who smoked during pregnancy. A previous experiment by Zhang et al. revealed that chemerin DNA methylation was negatively correlated with chemerin mRNA concentration in multiple tissues (Zhang et al. 2016), supporting the role of DNA methylation in regulating chemerin gene expression. Zhang et al. demonstrated in adipose tissue of CD1 mice a correlation of −0.893 between chemerin methylation and chemerin expression, which is a stronger correlation than the results from our study. However, given that humans are a much more heterogeneous population than laboratory mice, this is not surprising. In the present study, the changes in DNA methylation of chemerin do not appear to be due to global alterations in DNA methylation, as LINE1 DNA methylation was unchanged between the smoking and non-smoking groups. As expected, our cohort of exposed newborns had reduced birth weight and length compared to newborns not exposed in utero to cigarette smoke.

While individuals who smoke generally weigh less than their non-smoking counterparts, individuals who smoke tend to have greater central adiposity (Barrett-Connor & Khaw 1989; Canoy et al. 2005; Shimokata et al. 1989). Other factors such as age, sedentary lifestyle, gender, and lack of education, to name a few, are also associated with increased central adiposity (Ortega et al. 2007; Wang & Beydoun 2007). Previous studies have demonstrated that adipogenesis is increased following cigarette smoke extract exposure in primary cultured orbital fibroblasts (Cawood et al. 2007; Yoon et al. 2013) suggesting a potential mechanism by which smoking may lead to individuals with greater adiposity in distinct locations. Whether this increased adipogenesis occurs in multiple tissue types in vivo following smoke exposure has not been elucidated. The present data support a potential mechanism whereby children or adults exposed in utero to cigarette smoke could demonstrate greater rates of obesity later in life. Others have shown that although newborns exposed in utero to cigarette smoke tend to be smaller, they do have greater rates of obesity later in life (Power & Jefferis 2002) suggesting altered developmental programming, as extensively reviewed by Janesick et al. (Janesick & Blumberg 2011). Indeed, in the present study, we demonstrate that the infants born to mothers who smoked during pregnancy had reduced birth weight and length compared to the babies born to mothers who did not smoke during pregnancy. These data are in consonance with what others have shown in that smoke exposure during pregnancy alters infant size (Varvarigou et al. 2010).

We demonstrated that chemerin mRNA expression was increased in dermal foreskin tissue and primary dermal fibroblasts, and that DNA methylation of chemerin was decreased in dermal foreskin tissue of human neonates exposed in utero to cigarette smoke. Cigarette smoke has been shown to induce greater chemerin and chemerin receptor protein expression in the broncho alveolar lavage fluid of mice (Demoor et al. 2011). Our study demonstrates that these findings translate into human neonates of smoking mothers and provide evidence of epigenetic regulation as DNA methylation of chemerin was decreased in babies exposed to cigarette smoke. While we did not measure chemerin receptor directly, we would anticipate that chemerin receptor would also be increased. Numerous studies demonstrate chemerin is associated with increased adipogenesis (Goralski et al. 2007; Muruganandan et al. 2011; Roh et al. 2007) which may provide a mechanism behind why infants exposed in utero to cigarette smoke have greater rates of obesity later in life. However, defining causation for this mechanism would require additional studies of dose response and of knockout or transgenic mouse models of chemerin expression. Despite the present limitations, these results provide important new evidence for a link between maternal smoking during pregnancy and increased chemerin mRNA expression.

Low birth weight babies often demonstrate greater catch up growth than normal weight babies which places them at a greater risk of developing diseases, such as cardiovascular disease, type 2 diabetes, and obesity later in life (Ekelund et al. 2007; Eriksson et al. 1999). Indeed, the full term babies born to mothers who smoked in the present study had lower birth weights than the babies born to mothers who did not smoke during pregnancy. Vik et al. (Vik et al. 1996) found that babies born to smoking mothers displayed lower birth weight and length. However, by 5 years of age, these children had caught up to their non-smoke exposed counterparts and even displayed increased skinfold thickness suggestive of greater amounts of adiposity. Although not measured, we propose that increased chemerin expression may play a role in catch up growth. To our knowledge this has never been examined, and thus is purely speculation.

The study does have limitations. Given that newborns in the smoking group were smaller than control group, it is possible that other factors controlling gestational weight gain, instead of maternal smoking, are impacting the increased chemerin expression. However, others have demonstrated that chemerin in cord blood is positively correlated with birth weight which would not support such a concern (Mazaki-Tovi et al. 2012). Secondly, all samples were derived from term deliveries to minimize possible variation related to gestational age. The gene expression we observed may vary at different gestational ages or vary based upon the extent of nicotine exposure (we did not measure nicotine or cotinine directly in the mom or neonate). Thirdly, it is possible that some of the women who reported not smoking were exposed to second hand smoke during their pregnancy as this was not examined in this study. However, the inclusion of second hand smoke exposed samples in our non-smoking group would likely result in smaller differences between the groups. Thus it is possible that there would be a greater difference in chemerin expression between smokers and non-smokers if no second hand smoke samples were included. Lastly, we are making these measures in epidermal/dermal samples and predict that the adipose tissue would respond in a similar fashion which may not be the case. Due to limited tissue availability, we were only able to assess DNA methylation and mRNA expression of chemerin rather than protein expression in our samples so this should be investigated in the future. Finally, in the present study, we only collected samples from male newborns; thus, we cannot comment on the potential for sex bias in this study and whether female neonates would demonstrate similar chemerin expression patterns related to tobacco exposure.

In summary, we demonstrate that chemerin expression is altered in neonatal tissue and primary fibroblasts from newborns of mothers who smoked during pregnancy. Neonates born to mothers who smoke during pregnancy exhibit distinct changes as a response to in utero smoke exposure which are likely regulated in part by epigenetics. Our data provide a potentially novel mechanism behind increased later-life obesity risk in babies born to mothers who smoke during pregnancy.

New Findings.

• What is the central question of this study?

  • ∘Babies born to mothers who smoke are at an increased risk of obesity later in life. This study determined if chemerin, an adipokine implicated in obesity, is increased in neonates following in utero cigarette smoke exposure.

• What is the main finding and its importance?

  • ∘Chemerin mRNA expression was increased and chemerin DNA methylation was decreased in babies born to mothers who smoked during pregnancy. These data provide a potential mechanism which may be mediating the increased obesity risk in individuals who are born to mothers who smoked during pregnancy.