Introduction:
Growth factors are regulatory proteins and peptides involved in paracrine communication. In contrast to hormones, growth factors exert their actions in local rather than distal manner.(1) Growth factors are among the most notable mitogenic signaling polypeptides and thus contribute significantly to embryogenesis by promoting both proliferation and differentiation.(2)
Growth/differentiation factor 15 (GDF15) is one of the most well-studied TGF-beta superfamily members.(3–8) Similarly to many other TGF-beta proteins, GDF15 modulates cellular migration and adhesion, and tissue remodeling. All these processes are critical for pregnancy, and these two facts suggest that GDF15 may control implantation and trophoblast invasion in early pregnancy. As GDF15 is a regulatory immunosuppressive protein, it may also participate in the further pregnancy maintenance by blocking the maternal immunologic reactions towards the semi-allogenic fetus. These considerations are supported by several reports proving GDF15 as a marker for miscarriage.(9–11)
Case-control approach is a common and reliable method for studying physiological roles of intrinsic biological factors. However, assets of this approach may be abated in cases of rare diseases, syndrome heterogeneity and due to the problem of biological significance. This is particularly fair when studying the highly dynamic organismal state - pregnancy - having numerous and versatile disturbances. In this case, analysis of pregnancy-associated dynamics of the factor may be a convenient alternative or even required step advancing our understanding of the fundamental involvement of a factor in pregnancy. Thus, we aimed at assessing the dynamics of the serum GDF15 protein in the course of the physiological human pregnancy. Taking into consideration data from our study and those collected in other labs, we planned to put some light on the problem of cause-effect relationship of GDF15 and human gestation.
Materials and Methods
Systematic literature review: GDF15 has several aliases: GDF-15, MIC1, PDF, PLAB, PTGFB, NAG-1, NRG-1, MIC-1.
To find papers concerning GDF15 levels in human pregnancy (of whatever variant), we initially used NCBI PubMed search with the “GDF15 AND pregnancy” request. As the GDF15 term is present in MeSH database, the request yielded the following search engine translation (Search Details section):
("growth differentiation factor 15"[MeSH Terms] OR "growth differentiation factor 15"[All Fields] OR "gdf15"[All Fields]) AND ("pregnancy"[MeSH Terms] OR "pregnancy"[All Fields]).
The search returned 21 results supplying the most of information available. Additional searches were used to account the GDF15 aliases (Table 1). Please note that duplicate search hits are provided only once - at first appearance.
| Table 1: Search results and inclusion status of the papers found for the systematic review on the GDF15 dynamics in normal human pregnancy |
Request |
No. of results |
Inclusion status and non-redundant PMIDs |
Included |
Excluded but related |
Excluded |
GDF15 AND pregnancy |
21 |
22525911;
22365732;
16484622;
14726168;
12900512;
12495665;
11134143 |
22064648, GDF15 in tissues: decidual stromal cells and trophoblast;
22005158, GDF15 in placentas in labor;
19470878, GDF15 in preeclampsia and diabetes mellitus;
15057958, GDF15 in Down syndrome |
23507375, GDF15 after physical activity;
22688311, GDF15 in CVD;
21193529, Gdf15 in rats;
20578239, GDF15 in cancer;
20534737, GDF15 in cancer;
20529285, GDF15 in CVD;
15550334, review only;
15081668, author reply;
10660478, cell-specific GDF15 expression only;
9375789, cell-specific GDF15 expression only |
"GDF-15" AND pregnancy |
3 |
None |
None |
21815188, GDF15 in neonates |
"MIC1" AND pregnancy |
4 |
None |
None |
20385082, T. gondii mic1 gene is analyzed;
19447110, N. caninum NcMIC1 gene is mentioned;
17575983, N. caninum MIC1 gene is mentioned |
(((pdf) OR Prostate differentiation factor) AND pregnancy) NOT ".pdf" |
32 |
None |
None |
19095262, GDF15 in prostate;
18536641, hit by ‘prostate’ and ‘differentiation’;
17413841, GDF15 in prostate;
16724947, hit by ‘prostate’ and ‘differentiation’;
12039075, hit by ‘prostate’ and ‘differentiation’;
10583383, hit by ‘prostate’ and ‘differentiation’;
10188193, hit by ‘prostate’ and ‘differentiation’;
9842220, hit by ‘prostate’ and ‘differentiation’;
9697307, hit by ‘prostate’ and ‘differentiation’;
8981132, hit by ‘prostate’ and ‘differentiation’;
7974495, hit by ‘prostate’ and ‘differentiation’; |
"PLAB" AND pregnancy |
6 |
11440548 |
15297175, GDF15 in choriocarcinoma |
12485621, hit by e-mail,
9678666, hit by e-mail |
"NAG-1" AND pregnancy |
0 |
None |
None |
None |
"NRG-1" AND pregnancy |
5 |
None |
None |
23380917, neuregulin 1 meant;
22412988, neuregulin 1 meant;
18076984, neuregulin 1 meant;
15524405, neuregulin 1 meant;
15306553, neuregulin 1 meant |
"MIC-1" AND pregnancy |
14 |
24069146;
23824055 |
24699265, MIC-1 is in the introduction and not tested; the same research group as in (Hromas et al., 1997) |
16885531, review only;
2145816, N. gonorrhoeae study;
2660572, hit by “MIC” |
Participants: The present experimental study was a part of a screening study aimed at revealing markers of complications of the human pregnancy over its whole course (11). In that study, four trimester-time-points were covered: 5-7 weeks, 11-13 weeks, 15-18 weeks and 38-41 weeks of gestation. These time-points were analyzed without the follow-up of the same subjects, as the primary aim was to study several gestational age-specific pregnancy complications, i.e. physiological gestation groups were formed so as to match gestational ages of the complicated pregnancies groups and to serve as controls. For the purposes of the current study, we analyzed only the control groups - as those generally representing the dynamics of the normal human pregnancy.
The research project plan and study protocols were approved by Rostov state medical university bioethical committee, and the study was undertaken in accordance with the Helsinki Declaration of 1975, as revised in 2000. All participants signed the informed consent. A total of 124 healthy gravid women aged 19-39 years having similar reproductive and general anamneses were enrolled in the investigation. Characteristics of the studied groups are presented in Table 2.
Table 2: Primary characteristics of the groups |
Pregnancy status |
Group size |
Gestational age range (weeks) |
Age (years)a |
Median; 5-95% |
Mean±SD |
I trimester (mid) |
25 |
5-6 |
27; 21.8-40 |
28.5±5.27 |
I trimester (late) |
30 |
11-13 |
29; 24-35 |
29.3±4.02 |
II trimester (early) |
32 |
15-18 |
28; 23-33.7 |
27.7±2.7 |
III trimester (term) |
37 |
38-41 |
28; 21.3-38.6 |
28.9±5.14 |
| a - no assumption was made initially on the age data distribution, so the data were analyzed using Kruskal-Wallis test; no significant differences between all the groups were found. Note that the age data were normally distributed according to the D'Agostino-Pearson test; thus reverse statistics is applicable, and we therefore provide parametric data for the readers’ convenience. |
The women were under care of medical staff of Rostov state medical university. Delivery mode in the group of the women at term was cesarean section due to medical reasons not classified as pregnancy complications. All participants were residents of Rostov-on-Don. All participants signed informed consents.
Blood collection:Fasting blood samples were collected only once from each patient by the venipuncture procedure into the vacuum collection tubes without anti-coagulant in the morning of a doctor’s appointment or in the morning of the day of the operation. Blood samples were centrifuged to obtain serum. The samples were aliquoted and deep-frozen until analysis (-80 0C).
GDF15 level measurement:Serum concentrations of GDF15 were determined using the “GDF-15/MIC-1 Human ELISA” kit (BioVendor R&D, Czech Republic) and the automatic ELISA-analyzer “Alisei” (Radim, Italy), following the manufacturer’s protocols.
Statistical analysis and data representation:As this was not a follow-up study, we used only the independent samples statistics still allowing characterizing the population values variations. All results, including significant anamnesis and surveying data were statistically analyzed using the appropriate criteria in MedCalc 11.4. Sequential GDF15 level comparisons were performed using the Mann-Whitney criterion after Kruskal-Wallis test for the multiple comparisons eligibility. Acceptable maximum false-positive error probability (p-level) used was 0.05.
Results
GDF15 levels at the 4 time-points of the normal human pregnancy: Considering the involvement of GDF 15 in cellular and tissue processes significantly resembling those taking place during the pregnancy, we addressed whether GDF15 has a pregnancy advancing-associated dynamics. The results of the GDF15 level measurements at the mid- and late-first, early second and late third (at term) trimesters are presented in Figure 1.
 |
Figure 1: The dynamics of GDF15 in the course of the physiological human pregnancy. GW - weeks of gestation. GDF15 changes are represented with the black line connecting the gestational age-specific medians (black squares), while the gray lines are the 25th-75th percentile boundaries for the medians. The Kruskal-Wallis test revealed intra-group differences with p<0.0001. Further testing was performed using the Mann-Whitney post-hoc analysis. * - p<0.0001 - GDF15 levels significantly differ at the time-points of the mid- and late-first trimester. † - p<0.0001 - GDF15 levels significantly differ between the early second trimester and at term |
As seen from Figure 1, the concentration of the GDF15 protein in the present study increased by almost two orders of magnitude with the progression of the normal human pregnancy from the mid-first trimester toward the term. We observed the incline by one order of magnitude over the period from 5/6 to 11/13 weeks but did not observe any significant differences between gestational weeks 11/13 and 15/18. GDF15 concentration further increased drastically toward the term with an almost five-fold increment. To account for the observed results, we further performed a systematic analysis of the related data from the literature.
Systematic review: GDF15 in normal human pregnancy:Due to significant intersection of aliases of various factors both within and between species, we had to filter and select only the papers that were concerned with the subject. However, for the purpose of the systematic review, we provide below all available to us (to date of the manuscript preparation) information on the GDF15 levels in non-pregnant females and in healthy pregnant women. Table 3 summarizes information that was found using the PubMed index, with addition of the data from the present study.
As seen from Table 3, the estimates of serum/plasma GDF15 content varied by multiple orders of magnitude from study to study, and characteristics of the dynamics also significantly differed. In total, only 7 studies (including ours) presented data on the GDF15 dynamics in the normal human pregnancy, while the gestational age ranges covered in the studies are mostly non-intersecting.
Table 3: The state of knowledge in GDF15 levels in the normal human pregnancy. Please note that some values were acquired from charts, so the values are approximate |
Pregnancy status |
Samples |
Method |
Data |
GDF15 level (ng/ml) |
Ref. |
Non-pregnant |
N=6; pooled serum |
In-house ELISA kit |
Mean; SD |
0.36; 0.04 |
(Moore, et al., 2000) |
5 GW |
N= 19; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
2.5; 0.5 |
(Kaitu’u-Lino, et al., 2013) |
5-6 GW |
N=26; maternal serum |
ELISA kit from BioVendor R&D |
Median; 25-75% |
4.4; 2.8-10.7 |
This study |
6 GW |
N= 132; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
6.5; 0.1 |
(Kaitu’u-Lino, et al., 2013) |
6 GW |
N=36; maternal serum |
ELISA kit from R&D Systems |
Mean; SEM |
9500; 1000 |
(Tong, et al., 2012) |
7 GW |
N= 95; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
10; 0.3 |
(Kaitu’u-Lino, et al., 2013) |
7 GW |
N=17; maternal serum |
In-house ELISA kit |
Median; 95% CI |
4.0*109; 3.0*109-5.0*109 (sic) |
(Tong, et al., 2004) |
7 GW |
N=92; maternal serum |
ELISA kit from R&D Systems |
Mean; SEM |
11000; 500 |
(Tong, et al., 2012) |
8 GW |
N= 65; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
14; 0.6 |
(Kaitu’u-Lino, et al., 2013) |
8 GW |
N=25; maternal serum |
In-house ELISA kit |
Median; 95% CI |
5.0*109; 4.2*109-5.8*109 (sic) |
(Tong, et al., 2004) |
8 GW |
N=209; maternal serum |
ELISA kit from R&D Systems |
Mean; SEM |
13500; 350 |
(Tong, et al., 2012) |
9 GW |
N=54; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
12.2; 0.6 |
(Kaitu’u-Lino, et al., 2013) |
9 GW |
N=26; maternal serum |
In-house ELISA kit |
Median; 95% CI |
5.0*109; 4.1*109-5.9*109 (sic) |
(Tong, et al., 2004) |
9 GW |
N=250; maternal serum |
ELISA kit from R&D Systems |
Mean; SEM |
13500; 230 |
(Tong, et al., 2012) |
10 GW |
N=35; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
14; 1.0 |
(Kaitu’u-Lino, et al., 2013) |
10 GW |
N=174; maternal serum |
ELISA kit from R&D Systems |
Mean; SEM |
14500; 500 |
(Tong, et al., 2012) |
10 GW |
N=32; maternal serum |
In-house ELISA kit |
Median; 95% CI |
5.0*109; 4.2*109-5.8*109 (sic) |
(Tong, et al., 2004) |
11 GW |
N=21; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
14.5; 1.0 |
(Kaitu’u-Lino, et al., 2013) |
11 GW |
N=28; maternal serum |
In-house ELISA kit |
Median; 95% CI |
5.0*109; 4.1*109-5.9*109 (sic) |
(Tong, et al., 2004) |
12 GW |
N=6; maternal EDTA plasma |
ELISA kit from R&D Systems |
Mean; SEM |
13.5; 2.8 |
(Kaitu’u-Lino, et al., 2013) |
12 GW |
N=33; maternal
serum |
In-house ELISA kit |
Median; 95% CI |
6.0*109; 5.8*109-6.2*109 (sic) |
(Tong, et al., 2004) |
13 GW |
N=29; maternal
serum |
In-house ELISA kit |
Median; 95% CI |
5.9*109; 4.9*109-6.9*109 (sic) |
(Tong, et al., 2004) |
10-14 GW |
N=6; maternal serum; pooled |
In-house ELISA kit |
Mean; SD |
6.3; 0.02 |
(Moore, et al., 2000) |
10-13 GW |
N=28..33; maternal serum |
In-house ELISA kit |
Median ranged |
5.0*109..6.0*109 (sic) |
(Tong, et al., 2004) |
11-13 GW |
N=26; maternal serum |
ELISA kit from BioVendor R&D |
Median; 25-75% |
33.3; 23.8-45.9 |
This study |
15-18 GW |
N=32; maternal serum |
ELISA kit from BioVendor R&D |
Median; 25-75% |
38.6; 30.5-40.9 |
This study |
6-24 GW |
N=116; maternal serum |
In-house ELISA kit |
Median; SD |
11.4;4.2 |
(Marjono, et al., 2003) |
26-30 GW |
N=8; maternal serum; pooled |
In-house ELISA kit |
Mean; SD |
12.2; 0.5 |
(Moore, et al., 2000) |
33-35 GW |
N=12; maternal serum |
In-house ELISA kit |
Median; SD |
38.0; 9.4 |
(Marjono, et al., 2003) |
37-40 GW |
N=8; maternal serum; pooled |
In-house ELISA kit |
Mean; SD |
15.3; 1.3 |
(Moore, et al., 2000) |
38-40 GW |
N=39; maternal serum |
ELISA kit from BioVendor R&D |
Median; 25-75% |
181.2; 137.8-217.6 |
This study |
I trimester |
N=42; maternal serum |
ELISA kit from R&D Systems |
Median; range |
10.9; 2.69-22.6 |
(Finkenstedt, et al., 2012) |
II trimester |
N=12; maternal serum |
ELISA kit from R&D Systems |
Median; range |
14.3; 12.0-56.8 |
(Finkenstedt, et al., 2012) |
III trimester |
N=42; maternal serum |
ELISA kit from R&D Systems |
Median; range |
22.6; 8.3-118.9 |
(Finkenstedt, et al., 2012) |
Discussion
We observed a significant increase in the GDF15 levels over the physiological pregnancy course. However, there was no significant change of the levels between the gestational weeks 11/13 and 15/18. This fact may be accounted for by the oxidative burst which takes place upon spiral artery unplugging at 10–12 weeks of gestation (12, 13) and affects the inflammatory state of placenta, thus causing the anti-inflammatory response in macrophages with GDF15 increase being a part of it. The revealed dynamics of GDF15 is in a good agreement with the hypotheses of involvement of the protein in the pregnancy maintenance or its dependence on placenta and membranes growth. Additional experimental data would be highly helpful in choosing the most likely cause-effect relationship. However and surprisingly, data present in the literature appeared to be insufficient to answer the question, and somewhat contradictory.
S. Tong and co-authors found neglectful oscillations of the maternal serum GDF15 concentrations between 7 and 13 weeks of gestation (9), while Kaitu’u-Lino and colleagues (10) revealed somewhat pronounced GDF15 dynamics, especially in earlier stages of pregnancy. We observed an incline by one order of magnitude over the period between 5/6 and 11/13 weeks. S. Tong and colleagues, later - in 2012, (14) generally found slight changes in GDF15 at weeks 6-10.
Data on the GDF15 dynamics between the first and second trimesters were scarce. A.G. Moore and co-authors (15) found a duplication in GDF15 concentrations over the period between the ends of the first and second trimesters. We did not observe any significant increase in the GDF15 levels between the late-first trimester and early-second trimesters.
Further challenges were revealed during the analysis of the second-to-third trimester transition GDF15 levels. A.G. Moore and co-authors (15) did not observe any increase in GDF15 concentration, while Marjono and colleagues (6) and we found nearly a quadruplication.
Finkenstedt and co-authors (16) found a 1.5-fold increase in the GDF15 levels between the second and third trimesters.
A.G. Moore and colleagues (15) found a 2.5-fold increase of the GDF15 concentrations over the gestation period ranging between 10 and 40 weeks. Finkenstedt and co-authors (16) found a duplication of GDF15 levels between the first and third trimesters. We observed the striking ˜40-fold GDF15 increase between 5 and 40 weeks of gestation. Interestingly, Poggi and colleagues (17) showed that placental GDF15 mRNA expression is stable at gestation weeks 7 through 42.
In our opinion, the lack of consistency in the estimates of GDF15 dynamics is dramatic for the fundamental understanding of the roles of this factor in the normal human pregnancy and for further developments in studying pregnancy complications of the second and third trimesters. Although the absolute values of GDF15 in the maternal serum are highly questionable to date due to inter-laboratory data variation (a well-known problem discussed elsewhere (18)), the major challenge seems to be in the character of dynamics, i.e. the shape and the slope of the GDF15 incline during the physiological human pregnancy, as, though surprisingly, data from existing studies appears to be contradictory.
In conclusion, we found that the maternal serum GDF15 levels increase in the course of the human physiological pregnancy, rising by almost two orders of magnitude with the advance from the mid-first trimester toward the early second trimester and then toward the term. This character of the dynamics implies GDF15 as a factor participating in pregnancy maintenance and (or) dependent on the fetal growth, in any case suggesting significant practical value of GDF15 as a gestational age-specific pregnancy state marker. However, these practical considerations are limited by the significant controversy present in the literature on the circulating maternal levels of GDF15 in normal human pregnancies.
Acknowledgements
This study was supported by the federal assignment # 6.703.2014/K “New targets for the predictive diagnostics of reproductive disorders” from Russian Ministry of Science and Education. Funding source had no influence on study design and results. Analytical work was carried out on the equipment of Centers for collective use of Southern Federal University "High Technology", grant 14.594.21.0002.
Funding Information: This article was funded by the federal assignment from Russian Ministry of Science and Education. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Conflict of interest: None.
Ethics: No ethical issues relate to the present study.
References
- MeSH database. www.ncbi.nlm.nih.gov/mesh (Accessed on October 17, 2014).
- Perona R. Cell signalling: growth factors and tyrosine kinase receptors. Clin Transl Oncol. 2006;8(2):77-82. URL: http://www.ncbi.nlm.nih.gov/pubmed/16632420.
- Bootcov MR, Bauskin AR, Valenzuela SM, et al. MIC-1, a novel macrophage inhibitory cytokine, is a divergent member of the TGF-beta superfamily. Proc Natl Acad Sci U S A. 1997;94(21):11514-11519. URL: http://www.pnas.org/content/94/21/11514.long.
- Hromas R, Hufford M, Sutton J, et al. PLAB, a novel placental bone morphogenetic protein. Biochim Biophys Acta. 1997;1354(1):40-44. URL: http://www.sciencedirect.com/science/article/pii/S016747819700122X.
- Keelan, JA, Wang K, Chaiworapongsa T, et al. inhibitory cytokine 1 in fetal membranes and amniotic fluid from pregnancies with and without preterm labour and premature rupture of membranes. Mol Hum Reprod. 2003;9(9):535-540. URL: http://molehr.oxfordjournals.org/content/9/9/535.long.
- Marjono AB, Brown DA, Horton KE, et al. Macrophage inhibitory cytokine-1 in gestational tissues and maternal serum in normal and pre-eclamptic pregnancy. Placenta 2003;24(1):100-106. URL: http://www.sciencedirect.com/science/article/pii/S0143400402908817.
- Ago T, Sadoshima J. GDF15, a cardioprotective TGF-beta superfamily protein. Circ Res. 2006;98(3):294-297. URL: http://circres.ahajournals.org/content/98/3/294.long.
- Roth P, Junker M, Tritschler I, et al. GDF-15 contributes to proliferation and immune escape of malignant gliomas. Clin Cancer Res. 2010;16(15):3851-3859. DOI: 10.1158/1078-0432.CCR-10-0705.
- Tong S, Marjono B, Brown DA, et al. Serum concentrations of macrophage inhibitory cytokine 1 (MIC 1) as a predictor of miscarriage. Lancet 2004;363(9403):129-130. DOI: 10.1016/S0140-6736(03)15265-8.
- Kaitu'u-Lino TJ, Bambang K, Onwude J, et al. Plasma MIC-1 and PAPP-a levels are decreased among women presenting to an early pregnancy assessment unit, have fetal viability confirmed but later miscarry. PLoS One 2013;8(9):e72437. DOI: 10.1371/journal.pone.0072437.
- Skubisz MM, Brown JK, Tong S, et al. Maternal serum macrophage inhibitory cytokine-1 as a biomarker for ectopic pregnancy in women with a pregnancy of unknown location. PLoS One 2013;8(6):e66339. DOI: 10.1371/journal.pone.0066339.
- Zolotukhin P, Aleksandrova A, Goncharova A, et al. Oxidative status shifts in uterine cervical incompetence patients. Syst Biol Reprod Med. 2014;60(2):98-104. DOI: 10.3109/19396368.2013.864343.
- Hustin J, Schaaps JP. Echographic and anatomic studies of the maternotrophoblastic border during the first trimester of pregnancy. Am J Obstet Gynecol. 1987;157(1):162-168. URL: http://www.ncbi.nlm.nih.gov/pubmed/3300349.
- Tong S, Ngian GL, Onwude JL, et al. Diagnostic accuracy of maternal serum macrophage inhibitory cytokine-1 and pregnancy-associated plasma protein-A at 6-10 weeks of gestation to predict miscarriage. Obstet Gynecol. 2012;119(5):1000-1008. DOI: 10.1097/AOG.0b013e3182518fd3.
- Moore AG, Brown DA, Fairlie WD, et al. The transforming growth factor-[beta] superfamily cytokine macrophage inhibitory cytokine-1 is present in high concentrations in the serum of pregnant women. J Clin Endocrinol Metab. 2000;85(12):4781-4788. DOI: 10.1210/jcem.85.12.7007.
- Finkenstedt A, Widschwendter A, Brasse-Lagnel CG, et al. Hepcidin is correlated to soluble hemojuvelin but not to increased GDF15 during pregnancy. Blood Cells Mol Dis. 2012;48(4):233-237. DOI: 10.1016/j.bcmd.2012.02.001.
- Poggi SH, Bostrom KI, Demer LL, et al. Placental calcification: a metastatic process? Placenta 2001;22(6):591-596. DOI: 10.1053/plac.2001.0688.
- Zuvela E, Walls M, Matson P. Within-laboratory and between-laboratory variability in the measurement of anti-müllerian hormone determined within an external quality assurance scheme. Reprod Biol. 2013;13(3):255-257. DOI: 10.1016/j.repbio.2013.04.005.
|
