Fluoride-related changes in the fetal cord blood proteome; a pilot study

Background: Fluoride exposure during pregnancy has been associated with various effects on offspring, including changes in behavior and IQ. To provide clues to possible mechanisms by which fluoride affects human fetal development, we completed proteomic analyses of cord blood serum collected from second-trimester pregnant women residing in Northern California with either high or low fluoride exposure, as identified by maternal serum fluoride concentrations. Objective: To identify changes in cord blood proteins associated with maternal serum fluoride concentration in pregnant women living in Northern California. Methods: The proteomes of 19 archived second-trimester cord blood samples representing highest and lowest serum fluoride concentrations from a cohort of 48 women living in Northern California, previously analyzed for serum, urine and amniotic fluoride concentrations, were characterized by mass spectrometry. Proteins highly correlated to maternal serum fluoride concentrations were identified, and further compared in a group of samples from women with the highest serum fluoride to the group with the lowest maternal serum fluoride concentrations. Results: Nine cord blood proteins were significantly correlated with maternal serum fluoride concentrations. Six of these proteins, including apolipoprotein B-100, delta homolog 1, coagulation factor X, mimecan, plasma kallikrein, and vasorin, were significantly decreased in the cord blood from women with the highest serum fluoride levels. Conclusion: Changes in the relative amounts of second trimester cord blood proteins included proteins associated with the development of the fetal hematopoetic system.


INTRODUCTION
Tooth enamel uorosis, a biomarker for uoride exposure, has steadily increased in the US [1,2] resulting in hypomineralization of tooth enamel and at higher exposures, increasing the risk for dental disease [3].
Increasing levels of uorosis are of concern in view of recent human studies showing associations between uoride exposure during either pregnancy or the rst year of life, and impaired neurodevelopment of offspring [4][5][6][7][8][9].
Rodent models show uoride-related effects on cell function and neurodevelopment [10,11], associated with increased oxidative stress and in ammation [12][13][14][15][16].However, the relevance of these studies to human health has been questioned because rodents require 5 to 10-fold higher levels of uoride in drinking water to achieve plasma uoride levels similar to those found in humans [17].Mechanisms responsible for possible uoride-related effects in humans remain poorly understood, though reports of age and sex related effect of uoride on children, identi ed by analyses of NHANES data, suggest that uoride can affect multiple cells and tissues [18][19][20][21][22].
In this study, we investigated possible effects of uoride on the developing fetus by comparing the cord blood proteome of second-trimester pregnant women in Northern California [23].Cord blood samples were obtained from a random cohort of 48 samples of healthy second trimester pregnant women that we had previously analyzed for serum, urine, and amniotic uid uoride concentrations.Urine uoride concentrations of this cohort were similar to cohorts from larger retrospective studies of Candian [23] and Mexican women [5,6] showing uoride related change in offspring neurodevelopment.

Study Samples
This study was approved by the University of California, San Francisco Committee on Human Research.
Archived cord blood serum samples, from a cohort of 48 women [23], were selected for whole proteome analysis.Ten samples were from women who had the lowest serum uoride, and 9 samples were from women with the highest serum uoride in from this cohort.All women resided in communities with water uoride levels ranging from 0.2 to 0.9 ppm.The community water uoride concentration recommended by CDC was 1.0 ppm prior to 2015, and was subsequently change to 0.7 ppm uoride in 2015, and therefore all were near or below the maxium recommended levels for arti cial water uoridation.

Proteomic analysis
The cord blood samples were doped with protease inhibitors, and after clearing and dilution, the 14 most abundant proteins, including serum albumins were removed using immobilized antibodies.Total protein concentration was determined with BCA protein assay.The samples were buffer exchanged into ammonium bicarbonate, denatured with urea, reduced and alkylated, and digested with trypsin.The resulting peptide mixtures were desalted and analyzed by liquid chromatography-tandem mass spectrometry.Protein abundance in a sample was inferred from the number of peptide identi cations.To account for the analytical variability, relative peptide abundance was calculated for each protein in each sample by dividing the number of peptides identi ed from a given protein by the total number of peptides identi ed in that sample.

Statistical analyses
Demographics were compared by Kruskal-Wallis one-way analysis of variance test.Pearson's correlation coe cients were calculated for the relative protein abundance.The samples were grouped relative to highest and lowest maternal serum uoride concentrations and and signi cance (p-value) from paired Welch's t-test were calculated.All data analyses were conducted in the R statistical computing environment (version 4.0 or newer).

RESULTS
Demographics of the study population showed community water uoride levels at the time of sample collection from the 10 women with the lowest serum uoride concentrations (0.0040 to 0.0073 ppm) ranged from 0.2 to 0.8 ppm.Community water levels at the time of sample collection, for 9 women with higher serum uoride (0.0224 to 0.0593 ppm uoride) ranged from 0.16 to 0.9 ppm.Community water uoride concentrations for all subjects were near or below 1.0 ppm, which was the optimal water uoride levels recommended by the Center for Disease Control (CDC) at the time of collection.
The group of women with higher serum uoride levels, were further divided into middle range serum uoride of 0.0224 to 0.0291 ppm uoride, and high serum uoride ranging from 0.0444 to 0.0593 ppm uoride.There were no signi cant differences in maternal age, smoking status or ethnicity between the uoride groups (Table 1).However, the higher community water uoride levels were associated with higher serum uoride concentrations, with signi cant differences between groups.Mass spectrometry analysis of the cord blood proteome of the 19 samples showed signi cant correlations (Pearson's correlation coe cient greater or less than 0.5) of 9 proteins with maternal serum uoride concentrations (see Fig. 1).
Proteins identi ed as signi cantly correlated to maternal serum uoride were grouped into low and higher serum uoride concentration (combined medium and high serum uoride) and relative protein concentrations were compared.Six of the 9 cord blood proteins were signi cantly different in the low as compared to higher uoride group (p ≤ 0.05) (see Table 2).These cord blood proteins were all signi cantly reduced in the higher as compared to lower uoride group.

DISCUSSION
Fetal circulation occurs through the umbilical vein that carries maternal blood and nutrients ltered through the placenta to the fetus, and 2 umbilical arteries that transport blood to and from the placenta.
The placenta forms a barrier to macromolecules, including most proteins, to cross from maternal to fetal circulation [30].However, uoride crosses the placenta [23,31], and is also increased in the placenta relative to cord blood uoride concentrations [32,33].Previous studies have shown that fetal cord blood uoride concentrations re ect maternal blood uoride [34], and therefore in this study we used maternal serum uoride concentrations as a biomarker for fetal uoride exposure [23].
Cord blood serum was collected between 2014 and 2016 from second trimester pregnant women residing in Northern California [23].All women were healthy without known underlying medical conditions, and their community water uoride concentrations were at or below optimal water uoride concentrations, as recommended by the US Center for Disease Control (CDC) at the time of sample collection (1.0 ppm uoride).The 9 cord proteins found to be signi cantly associated with the maternal serum uoride concentration were not correlated with risk factors for pregnancy outcomes including maternal age [35,36], smoking status [37,38], BMI [39], and race/ethnicity.These proteins included protein delta-like homolog 1 (DLK1), a transmembrane protein highly expressed by stromal cells of the placental villi [28] and by the fetal liver hepatocytes [40], which was negatively associated with maternal serum uoride.DLK1 regulates glucose metabolism [41] and placental growth hormone [42], and is reduced in small for gestational age fetuses [28].The placenta is a primary hematopoietic stem cell (HSC) niche during pregnancy, which is believed to seed the fetal liver [43,44], and depletion of DLK1 from of human fetal liver hepatoblasts in culture reduces the percentage of mature hematopoietic cells [40].
DLK1 also regulates notch signaling and affects neurodevelopment [45].In mice, DLK1 was shown to regulate hippocampal neurogenesis and cognition.Reduced DLK1 levels triggers cognitive abnormalities [46], suggesting the possibility that uoride-related effects on fetal neurodevelopment may be associated with reductions in DLK1.
The fetal liver is the primary organ for erythropoiesis during development, functioning as a vascular connection between the developing placental vessels connecting the heart [47].It does not perform the traditional digestive and ltration functions because nutrients are normally carried to the fetus from the mother via the placenta.Other proteins signi cantly downregulated relative to maternal serum uoride are also associated with hematopoiesis, and include apolipoprotein B-100 (APOB), coagulation factor X, plasma kallikrein, and vasorin.APOB, which is synthesized by the placenta, is a likely pathway for lipid transfer required for fetal growth [24].Vasorin is a transmembrane glycoprotein[48] with a possible role in the regulation of the glycogen-mediated mTOR-ULK1 signaling pathway in the liver [49], and mimecan, also known as osteoglycin which is a small leucine-rich proteoglycan (SLRPs) present in the extracellular matrix of multiple tissues, including adipose tissue.
The possibility that uoride may affect the liver in humans is supported by NHANES survey data from 2013-2016 that shows higher water uoride and plasma uoride concentrations are associated with lower blood urea nitrogen among US adolescents.However, while the fetal hematopoietic system is separate from that of the maternal circulatory system, the placenta is of dual origin, comprised of both fetally and maternally derived cells and maternal in ammation can alter maternal-fetal interactions.Maternal immune activation during pregnancy, can affect the development of the fetal hematopoietic system [50], and therefore it possible that uoride related effects to increase maternal in ammation may also have a role in fetal development.Analysis of available NHANES survey data (2013-2014 and 2015-2016) of children aged 6-19, shows plasma uoride concentrations positively correlated with total white blood cell count, segmented neutrophils and monocytes when adjusted for age, gender, and BMI as predictors.While there is no available data linking second trimester maternal white blood cell counts to uoride exposure, further studies to explore this possibility are warrented.
In summary, this is the rst study to assess possible uoride related changes associated with human fetal development at low level uoride exposure.Our nding in this pilot study, that cord blood proteins associated with the development of the fetal hematopoietic system were signi cantly downregulated relative to maternal uoride exposure, supports the possibility that uoride related mechansims affecting fetal development and may be associated with the placenta and fetal hematopoesis during second trimester fetal development.The second trimester of fetal development is also critical for development of the prefrontal cortex and integration of control of attention, thought, emotions and actions [51].Therefore, the identi cation of cord blood proteins altered by relative levels of maternal uoride exposure may direct future studies of uoride related mechanism that may affect fetal development, including neurodevelopment. Figures

Table 1
Demographics of study population (N = 19)