KIR and HLA-C : Immunogenetic regulation of human birth weight

Pregnancies resulting in very small or very large babies are at higher risk of obstetric complications with increased morbidity for both mother and baby. Using data from the Medical Birth Registry of Norway we have shown how human birth weight is still subject to stabilizing selection. Particular combinations of maternal/fetal immune genes have been implicated in pregnancies resulting in a low birth weight baby (<5 birth weight centile). More specifically, an inhibitory maternal KIRAA genotype with a paternally derived fetal HLA-C2 ligand. At the other end of the birth weight spectrum the presence of an activating maternal KIR2DS1 gene is associated with increased birth weight in linear or logistic regression analyses of all pregnancies >5 centile (p=0.005, OR=2.65). Thus, inhibitory maternal KIR combined with fetal HLAC2 is more frequently associated with low birth weight, whereas activating maternal KIR with fetal HLAC2 ligand is associated with increasing birth weight. Our findings using the MoBa cohort have replicated the association of KIR and HLA-C seen in poor placentation, and confirm the importance of maternal/fetal immune gene interactions in determining the outcome of pregnancy. This is an open access article distributed under the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Our work aims to understand how the immune system contributes to successful pregnancy. The placenta forms the interface between the mother and baby in the uterus. Fetal extravillous trophoblast cells (EVT) invade into the maternal decidua where they remodel uterine spiral artery walls converting them into high conductance vessels, thus securing a good blood supply to the feto-placental unit [1]. Insufficient transformation of the maternal spiral arteries results in poor placental development which consequently gives rise to disorders of pregnancy such as pre-eclampsia, fetal growth restriction (FGR), still birth and recurrent miscarriage [2-4]. Excessive invasion by EVT can also be detrimental in cases such as placenta accreta. High birth weight is potentially dangerous to both mother and baby through prolonged obstructed labour [5,6] with risk of asphyxia and injury in the baby [7-9], and lacerations and post-partum haemorrhage in the mother [5,9]. It is therefore crucially important that this process is carefully regulated maintaining a balance between over and under invasion. Using data from 795,068 first pregnancies resulting in a live birth from the Medical Birth Registry of Norway, we plotted the distribution of birth weight against frequency of transfer to special care baby unit [10]. This illustrated clearly the risk of morbidity in neonates at either end of the birth weight spectrum, even in a modern medicalised setting (Fig. 1). Large numbers of a specialised type of lymphocyte known as uterine NK cells are found in the decidua during placentation [11]. Uterine NK cells express Killer cell Ig-like Receptors (KIR) and fetal EVT express their cognate ligand HLA-C, the only classical HLA class I molecule found on trophoblast [12]. We have proposed that this maternal KIR/fetal HLA-C interaction functions to mediate uterine NK cell control of trophoblast invasion [11,13-15]. This receptor ligand interaction is unusual in that both KIR and HLA are highly polymorphic gene systems. They also segregate independently and are encoded on separate chromosomes. KIR genotypes vary with content, copy number and at allelic variation at individual KIR loci. Around 500 different genotypes have already been described to date [16]. To simplify this complexity, KIR haplotypes are classified as either A or B based on gene content. The KIR A haplotype is highly stable, varying little at the gene content level and carrying fewer genes. Notably it encodes KIR2DL1 and KIR2DL3, both inhibitory receptors for HLA-C. The KIR B haplotype is much more variable with the potential to encode inhibitory KIR2DL1 and KIR2DL2, and also activating KIR2DS1, all of which bind HLA-C [11]. HLA-C alleles can be subdivided into two groups C1 and C2 based on a dimorphism at position 80 of the 108 L.E. FARRELL ET AL. Figure 1. Distribution of birth weights in the Norwegian population with percentage of babies transferred to the special care baby unit for the years 1967-2010

Our work aims to understand how the immune system contributes to successful pregnancy.The placenta forms the interface between the mother and baby in the uterus.Fetal extravillous trophoblast cells (EVT) invade into the maternal decidua where they remodel uterine spiral artery walls converting them into high conductance vessels, thus securing a good blood supply to the feto-placental unit [1].Insufficient transformation of the maternal spiral arteries results in poor placental development which consequently gives rise to disorders of pregnancy such as pre-eclampsia, fetal growth restriction (FGR), still birth and recurrent miscarriage [2][3][4].Excessive invasion by EVT can also be detrimental in cases such as placenta accreta.High birth weight is potentially dangerous to both mother and baby through prolonged obstructed labour [5,6] with risk of asphyxia and injury in the baby [7][8][9], and lacerations and post-partum haemorrhage in the mother [5,9].It is therefore crucially important that this process is carefully regulated maintaining a balance between over and under invasion.Using data from 795,068 first pregnancies resulting in a live birth from the Medical Birth Registry of Norway, we plotted the distribution of birth weight against frequency of transfer to special care baby unit [10].This illustrated clearly the risk of morbidity in neonates at either end of the birth weight spectrum, even in a modern medicalised setting (Fig. 1).
Large numbers of a specialised type of lymphocyte known as uterine NK cells are found in the decidua during placentation [11].Uterine NK cells express Killer cell Ig-like Receptors (KIR) and fetal EVT express their cognate ligand HLA-C, the only classical HLA class I molecule found on trophoblast [12].We have proposed that this maternal KIR/fetal HLA-C interaction functions to mediate uterine NK cell control of trophoblast invasion [11,[13][14][15].
This receptor ligand interaction is unusual in that both KIR and HLA are highly polymorphic gene systems.They also segregate independently and are encoded on separate chromosomes.KIR genotypes vary with content, copy number and at allelic variation at individual KIR loci.Around 500 different genotypes have already been described to date [16].To simplify this complexity, KIR haplotypes are classified as either A or B based on gene content.The KIR A haplotype is highly stable, varying little at the gene content level and carrying fewer genes.Notably it encodes KIR2DL1 and KIR2DL3, both inhibitory receptors for HLA-C.The KIR B haplotype is much more variable with the potential to encode inhibitory KIR2DL1 and KIR2DL2, and also activating KIR2DS1, all of which bind HLA-C [11].
HLA-C alleles can be subdivided into two groups C1 and C2 based on a dimorphism at position 80 of the α1 domain [17].In this way KIR distinguish between C1 and C2 as mutually exclusive epitopes (Table 1).
To determine if maternal KIR and fetal HLA-C variation is important in pregnancy we have looked for consistent genetic associations of particular maternal KIR and fetal HLA-C combinations in a range of obstetric disorders (pre-eclampsia, FGR, and recurrent miscarriage) all of which result from defective placentation.All these conditions share a genetic association characterised by mothers with two KIR A haplotypes (AA genotype), combined with an HLA-C2 group allele in the fetus, particularly when the fetal C2 is paternally derived.Conversely, women who carry a KIR B haplotype encoding KIR2DS1, the activating receptor for C2, are significantly protected from these disorders when a fetal C2 is present [11,18,19].
Because we were intrigued by the high fetal and maternal mortality and morbidity seen in pregnancies with high birth weights [5][6][7][8][9] we went on to question what maternal KIR and fetal HLA-C combinations were associated with these pregnancies.
Mother and baby DNA pairs were genotyped for KIR and HLA-C from two separate cohorts.From our United Kingdom cohort we analysed 747 preeclamptic pregnancies, 118 pregnancies with FGR (≤5 th birth weight centile) and 404 normal pregnancies [10].Selected from the MoBa cohort [20] were 995 normal (including 66 >90 th birth weight centile), and 141 pre-eclamptic first pregnancies.
Replicating our previous findings from the UK, increased maternal KIR AA and low KIR2DS1 frequencies associated with pregnancies with low birth weight and/or pre-eclampsia in the MoBa cohort.Conversely, high birth weight pregnancies (≥90 th centile) had low KIR AA and high KIR2DS1 frequencies [10].
The effect of KIR2DS1 on birth weight was tested in both categorical and continuous analysis.Using birth weight in grams as a continuous variable, the presence of a maternal KIR2DS1 conferred an average birth weight increase of 78g (p=0.005) in a linear regression model.The frequency of maternal KIR2DS1 was significantly higher in pregnancies with high compared with median birth weight in categorical analysis (Table 2, [10]).
The effect of KIR AA genotypes on pre-eclampsia and FGR was previously observed particularly in pregnancies with a fetus carrying a paternally derived C2.This was shown by both categorical and continuous analysis across the birth weight spectrum.The effect of maternal KIR2DS1 on birth weight is thus dependent on the presence of fetal C2, particularly paternally derived fetal C2.
In other words the presence of fetal C2 amplifies the effect of maternal KIR2DS1.When the fetus has more C2 groups than the mother (C1/C2 fetus with C1/C1 mother, and C2/C2 fetus with C1/C2 mother), an average increase of 245g (p=0.002) was seen.Furthermore, presence or absence of KIR2DS1 as a categorical variable when combined with more C2 in the fetus showed that maternal KIR2DS1 has a significant effect when the fetus is carrying more C2 than the mother, (OR 2.93, 95% CI 1.66-5.18,p=0.0002).In pregnancies where the fetus was C1/C2 we could determine parent of origin of the C2 group.In both categorical and continuous analyses, the presence of maternal KIR2DS1 only has an effect on birth weight  2).Our findings using MoBa subjects have replicated the association of KIR and HLA-C with poor placentation (pre-eclampsia, and low birth weight) and confirm the importance of maternal KIR/fetal HLA-C interactions in determining the outcome of pregnancy.Of importance is that we now also show an effect in high birth weight pregnancies, implicating a role in the regulation of placentation in normal or excessive invasion.This effect has a clear direct impact on birth weight as a continuous variable, with an effect comparable or even greater than smoking during pregnancy, high altitude and sex of the baby [21][22][23][24].
KIR2DS1 is expressed by uterine NK cells and is functional although ascertaining exactly how NK cells operate to subtly define the extent of arterial transformation by trophoblast is an exciting challenge for the future.Nonetheless our findings indicate that a balance of KIR inhibitory and activating stimuli is necessary for optimal trophoblast invasion.NK derived cytokines such as GM-CSF, released in response to activation of KIR2DS1 by binding C2, are one possible mechanism [13].
Not all women with KIR AA genotypes and fetal C2 have a pregnancy disorder.We are now selecting patients from the MoBa cohort to study women who have recurrent pre-eclampsia as well as those who have a normal pregnancy followed by a pre-eclamptic pregnancy and vice versa.We aim to focus on KIR2DL1, the inhibitory KIR for HLA-C2 because there are 4 different KIR2DL1 alleles in the Norwegian population.Our prediction is that particular KIR2DL1 alleles will confer most risk.
Perhaps in the long term KIR/HLA-C genotyping might be a genetic predictor of birth weight to identify those at risk of FGR, pre-eclampsia or macrosomia.

Figure 1 .
Figure 1.Distribution of birth weights in the Norwegian population with percentage of babies transferred to the special care baby unit for the years 1967-2010 (n=795,068).(Originally published in Journal of Immunology: Hiby S, Apps R, Chazara O, et al.Maternal KIR in combination with paternal HLA-C2 regulate human birth weight.J. Immunol 2014; 192: 5069-5073.Copyright © [2014] The American Association of Immunologists, Inc.)

Table 1 .
KIR known to bind HLA-C and their HLA-C ligands.

Table 2 .
Presence of maternal KIR2DS1 associates with increased birth weight in both categorical and continuous analysis n=1316.This is enhanced when the fetus has more HLA-C2 epitopes than the mother n=304 and specifically paternally derived C2 n=204.Data summarised from tables I, II, and IV-VII in Hiby S, Apps R, Chazara O, et al.Maternal KIR in combination with paternal HLA-C2 regulate human birth weight.J. Immunol 2014; 192: 5069-5073.