This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lindsay, R. S. Articles by Walker, J. D. Search for Related Content PubMed PubMed Citation Articles by Lindsay, R. S. Articles by Walker, J. D.

Type 1 Diabetes-Related Antibodies in the Fetal Circulation: Prevalence and Influence on Cord Insulin and Birth Weight in Offspring of Mothers with Type 1 Diabetes

British Heart Foundation Cardiovascular Research Centre (R.S.L.), University of Glasgow, Western Infirmary, Glasgow G11 6NT, United Kingdom; Diabetes Research Institute and 3rd Medical Department (A.-G.Z.), Krankenhaus Munchen-Schwabing, Munich, Germany; University Department of Obstetrics and Gynaecology (B.A.H., A.A.C., F.D.J.), Centre for Reproductive Biology, University of Edinburgh, Edinburgh, United Kingdom; and Diabetic Department (J.D.W.), St John’s Hospital at Howden, West Lothian National Health Service Trust Livingston, West Lothian EH54 6PP, United Kingdom

Address all correspondence and requests for reprints to: Robert Lindsay, British Heart Foundation Cardiovascular Research Centre, University of Glasgow, Western Infirmary, 44 Church Street, Glasgow G11 6NT, United Kingdom. E-mail: rsl3c{at}clinmed.gla.ac.uk.

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
During pregnancy, maternal type 1 diabetes-associated autoantibodies may cross the placenta. It is proposed that insulin antibodies (IA) allow transfer of insulin across the placenta, contributing to fetal hyperinsulinemia and macrosomia.

We assessed the prevalence of IA, the tyrosine phosphatase IA-2, and glutamic acid decarboxylase (GADA) in cord blood from offspring of mothers with type 1 diabetes (ODM, n = 138) and control mothers (control, n = 47) and further assessed cross-sectional relationships of antibody titers to birth weight and fetal insulin.

In ODM, antibodies were frequently present in cord blood; 124 ODM (95%) were positive for IA, 82 (59%) were positive for GADA antibodies, and 61 (44%) were positive for IA-2 antibodies. In controls, GADA and IA-2 antibodies were absent, whereas seven controls (15%) were positive for IA at low titers (P < 0.0001 ODM vs. controls for all).

ODM with IA (IA positive) or without IA (IA negative) had similar birth weights (mean ± SD: IA positive, 3.8 ± 0.7 kg; IA negative, 4.0 ± 0.6 kg; P = 0.31) and cord insulin concentrations (IA positive: median, 112 pmol/liter; interquartile range, 62–219 pmol/liter; IA negative: median, 114 pmol/liter; interquartile range, 59–194 pmol/liter; P = 0.96). Similarly, the presence of GADA and/or IA-2 autoantibodies (n = 103) was not associated with differences in birth weight or insulin concentrations. Antibody titers were not associated with birth weight or insulin as continuous variables in either controls or ODM.

Islet autoantibodies and IA are a common finding in cord blood of ODM, but we found no evidence that they influence offspring insulin concentrations or weight at birth.

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
IN TYPE 1 DIABETES, maternal islet autoantibodies and antibodies to exogenous insulin are known to cross the placenta and are detectable in cord blood at birth. Little is known about the functional effects of these antibodies on the developing fetus. Studies in mothers with type 1 diabetes treated with nonhuman insulin have shown that anti-insulin antibodies (IA) allow passage of maternal insulin (which normally does not cross the placenta) into the fetal circulation (1). Surprisingly, in this study, an average of 27% of insulin in umbilical vein plasma was found to be of nonhuman origin, representing transfer of exogenous insulin from the mother. In the light of this, it was proposed that transfer of insulin by anti-IA might contribute to the development of fetal hyperinsulinemia and macrosomia in pregnancy complicated by type 1 diabetes (1). This notion was not supported in a subsequent study of IA measured in offspring of mothers with either type 1 diabetes or gestational diabetes (2); however, in this study, an older (ELISA) technique was used to measure IA, the prevalence of positive IA was relatively low (23%), and antibodies to IA-2 and glutamic acid decarboxylase (GADA) were not measured (2).

We have assessed the prevalence of islet autoantibodies, antibodies to the tyrosine phosphatase IA-2 (IA-2A) and to GADA, as well as IA in cord blood samples from offspring of mothers with type 1 diabetes (ODM) using stored samples from the Fetal Insulin and Glycemia Study (3). It should be noted that current assays do not distinguish between insulin autoantibodies and the antibodies to exogenous insulin that are frequently present after insulin administration. For clarity, we refer to antibodies to insulin (IA) throughout. We have examined the hypothesis that the presence of antibodies would have functional consequences on growth in utero or fetal insulinemia as assessed by weight and cord insulin at birth. More specifically, we hypothesized that the presence of anti-IA might result in increased fetal insulin and birth weight, whereas autoantibodies to islet cell antigens, IA-2A and GADA, would be associated with reduced insulin secretion and thus lower birth weight and cord insulin concentrations.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Recruitment and clinical protocol

The cohort has been previously described in detail (3). A total of 250 women with type 1 diabetes at eight hospital-based antenatal centers consented to take part in the study (94% of those enrolling with and planning to deliver in the centers); cord blood samples were obtained in 200 (80%) of these women. The 200 samples were further restricted to women in whom there was no evidence of hemolysis of cord blood (by visual inspection), in whom cord blood had been collected within 20 min [median collection time for remaining samples, 2 min; interquartile range (IQR), 1–7 min], and whose cord blood had been centrifuged and plasma frozen within 60 min (median time from collection to freezing for remaining samples, 14 min; IQR, 6–23 min) (3). Of 151 ODM potentially available after these restrictions, stored samples were available for assay of autoantibodies in 140. For the main analysis, 138 participants are included in whom all three autoantibodies were successfully measured. Review of charts revealed that 27 mothers (20%) had been hospitalized during pregnancy due to hypertensive problems (nine women with pregnancy-induced hypertension and 18 women with preeclampsia). Twenty mothers (15%) had previously diagnosed thyroid disease, seven (5%) had asthma, and one had previously diagnosed epilepsy.

For comparison, autoantibodies were also assessed in a convenience sample of 47 control mothers who had no history of obstetric or metabolic disease and in whom routine screening for gestational diabetes (using national guidelines: http://www.sign.ac.uk/guidelines/fulltext/55/section8.html) was negative. Mothers were recruited from routine obstetric follow-up clinics after the 34th week of pregnancy in the same centers (3).

All clinics offer antenatal care delivered by a multidisciplinary team comprising obstetricians, diabetologists, specialist midwives, diabetes specialist nurses, and dietetic support. Local management protocols for treatment of type 1 diabetes in pregnancy were followed. Data on clinical outcome, including caesarian section, intercurrent medical conditions, and hypertensive conditions of pregnancy, were obtained by chart review. Gestational ages were calculated from estimated dates of delivery from chart review. Estimated dates of delivery were derived from dates of last menstrual period (LMP), when available, or by ultrasound when there was either conflict with dates as assessed by LMP (>6 d) or LMP was unavailable.

All mothers gave informed consent. Protocols were approved by local ethical committees.

Collections of cord bloods and assays

After delivery, 20 ml of cord blood was collected from the umbilical vein, after cord clamping, into lithium heparin at ambient temperature. Depending on local circumstances, samples were then either transferred to local laboratories for centrifugation or centrifuged in situ, before initial storage of plasma at –20 C and eventual central storage at –70 C [after a median of 11 d (IQR, 5–21 d)]. Previous analyses in this group have established that insulin is unstable when there are delays in collection of blood from umbilical vein of more than 20 min or delays in freezing of more than 60 min, whereas by contrast, insulin propeptides are stable for at least 30 min of delay after cord sampling and 24 h before freezing (3). For the purposes of this investigation, hormonal measures were only included if collected from cord within 20 min and frozen within 60 min.

Plasma insulin, 32–33 split proinsulin, and HbA1c (at a central reference laboratory: Bio-Rad Variant, nondiabetic reference range 4.4–5.7%; Bio-Rad, Hercules, CA) were assayed as previously described (3). In two cases (both controls), insulin concentrations were below the limit of detection of the assay (2 pmol/liter). These data have been included with an assigned value of 1.9 pmol/liter. These assays display low cross-reactivity to other insulin species. The insulin assay shows cross-reactivity less than 0.2% with intact proinsulin and less than 0.5% with 32–33 split proinsulin at concentrations of 2736 pmol/liter and 2800 pmol/liter, respectively. The intact proinsulin assay shows less than 0.05% cross-reaction with insulin and less than 1% with 32–33 split proinsulin at concentrations of 6000 pmol/liter and 400 pmol/liter, respectively. The 32–33 split proinsulin assay shows 100% cross-reaction with intact proinsulin. To obtain a specific measure of 32–33 split proinsulin, it is necessary to take account of the intact proinsulin concentration of the specimen. Cross-reaction with insulin is less than 0.05% at 6000 pmol/liter.

Antibodies (IA, GADA, and IA-2A) were determined by radiobinding assays (4, 5). For the purposes of this analysis, autoantibodies were considered positive at titers greater than the 99th percentile of control children, corresponding to titers more than 8.5 local U/ml (GADA), more than 2.5 local U/ml (IA-2A) and more than 1.5 local U/ml (IA) as previously published (4, 5). Assays had sensitivities and specificities of 84 and 96% (GADA), 66 and 100% (IA-2A), and 64 and 99% (IA), respectively, in the Third Diabetes Autoantibodies Standardization Program Proficiency Workshop.

Statistical analysis

Data were analyzed using standard software (SAS Institute Inc., Cary, NC). In several cases (insulin and 32–33 split proinsulin), measures were not normally distributed, and therefore, log-transformed values were used to approximate normal distributions. Differences between control women and women with type 1 diabetes were assessed by unpaired t test or, when further predictor variables were included, by general linear models. Because we have previously demonstrated that 32–33 split proinsulin concentrations are more stable than insulin to sampling conditions at birth (3), analysis of the relationship of antibodies to both 32–33 split proinsulin and insulin are included. With the exception of IA in ODM, antibody titers were frequently below the limit of detection of the assay, resulting in a truncated distribution (antibody titers below limit of detection in ODM: IA, one of 138; GADA, 13 of 138; IA-2A, 19 of 138; antibody titers below limit of detection in controls: IA, 10 of 47; GADA, 20 of 47; IA-2A, nine of 47). Thus, for models using IA, GADA, or IA-2A as the dependent variable, a Tobit model was used. In models where IA, GADA, or IA-2A were entered as predictors, log-transformed values were used after assignment of samples below the level of detection (0.1 local U/ml) with values of 0.05 local U/ml.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
In ODM, islet autoantibodies and anti-IA were frequently present in cord blood; 124 ODM (95%) were positive for IA, 82 (59%) were positive for GADA, and 61 (44%) were positive for IA-2A (P < 0.0001 ODM vs. controls for each). GADA and IA-2A antibodies were absent from all control cord bloods. Nominally positive IA were found in seven controls, although all were at low titers (six controls between 1.5 and 2.8 local U/ml and one control with 5.6 local U/ml). By comparison, median IA in ODM was 30 local U/ml, and 76% had values greater than 5.6 local U/ml.

Cord GADA titers were inversely related to duration of maternal diabetes (P = 0.003) but were unrelated to maternal age at delivery, gestational age at delivery, and maternal glycosylated hemoglobin (taken at wk 30 of gestation ± 4 wk). Cord IA-2A and IA were not significantly related to any of the above variables.

ODM with IA (IA positive) or without IA (IA negative) had similar birth weights (mean ± SD: IA positive, 3.8 ± 0.7 kg; IA negative, 4.0 ± 0.6 kg; P = 0.31 and P = 0.15 after adjustment for sex and gestational age at delivery) and cord insulin concentrations (IA positive: median, 112 pmol/liter; IQR, 62–219 pmol/liter; IA negative: median, 114 pmol/liter; IQR, 59–194 pmol/liter; P = 0.96). When analyzed as continuous variables, IA showed a marginal negative relationship with birth weight in ODM (P = 0.07 with adjustment for gestational age at delivery and sex; Fig. 1) but no significant relationship with cord insulin (P = 0.55) or 32–33 split proinsulin (P = 0.15).

View larger version (18K): [in this window] [in a new window]   FIG. 1. Birth weight (adjusted for sex and gestational age) by quintile of IA in ODM. Quintiles (n = 27 or 28 in each group) of IA are as follows (lowest to highest): 0–4.6 local U/ml, 4.8–20.6 local U/ml, 22.7–40.2 local U/ml, 40.7–98.2 local U/ml, and 99.6–1404 local U/ml. There is a marginal negative trend of IA titer as a continuous variable on birth weight (P = 0.07) after adjustment for sex and gestational age at delivery, but there is no relationship with cord insulin (P = 0.55).

Autoantibodies were not significantly related to birth weight, insulin, or 32–33 split proinsulin in controls or in controls and ODM combined (data not shown) with the possible exception of birth weight, which again showed a negative relationship with cord IA as observed in ODM alone (P = 0.07, with adjustment for presence of diabetes, gestational age at delivery, and sex).

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Our study confirms that islet autoantibodies and IA are a frequent finding in cord blood at birth; however, we have not found any support for functional consequences of this with regard to either birth weight or insulin at birth.

Several previous studies have documented a high prevalence of positive titers of IA, GADA, and IA-2A in cord blood specimens taken from ODM (6, 7, 8). In the most recent analysis of the BABY-DIAB study (using the same assays as this study), antibodies were present in the majority of cord samples from ODM at similar titers to those described here (IA, 86%; GADA, 56%; and IA-2A, 37%) (8). Antibodies are subsequently eliminated from the fetal circulation by 6–12 months of age (5, 6, 7). GADA and IA-2A antibodies are infrequently found in offspring of diabetic fathers at birth (6), and titers in maternal and cord blood are highly correlated (6), supporting the concept that antibodies are passed from mother to child. Similarly, when specific anti-IA assays (4) are used (as in this series), IA are present in cord blood taken from ODM but almost never in cord blood from controls or offspring of fathers with type 1 diabetes and, thus, are likely to represent IA transferred from mother (4, 8).

Evidence for the functional effects of autoantibodies found in cord blood is more limited. In the nonobese diabetic mouse, transplacental passage of maternal autoantibodies clearly increases the risk of autoimmune diabetes in offspring (9). By contrast, recent follow up of the BABY-DIAB study showed that the presence of GADA and IA-2A antibodies in cord blood samples from ODM was associated with a reduction in later incidence of type 1 diabetes (8). Presence of IA did not influence later autoimmune diabetes in the child (8). Others have postulated an effect of anti-IA on fetal insulin and macrosomia. Menon et al. (1) reported that anti-insulin autoantibodies acted to pass maternal insulin (which does not usually cross the placenta) from maternal to fetal circulation. In support of this, they noted that, in the presence of anti-IA, bovine and porcine insulin could be detected in the fetal circulation when mothers were treated with nonhuman insulins (1). They suggested that passage of maternal insulin into the fetal circulation might be an important mechanism in production of fetal hyperinsulinemia and subsequent stimulation of fetal growth (1). They were not able to rigorously test this hypothesis. Although they noted that ODM who had macrosomia were more often hyperinsulinemic, they did not detect an excess proportion of offspring who were autoantibody positive in the presence of macrosomia, but the numbers were too small (n = 28) for detailed analysis (1). By contrast, Weiss et al. (2) reported no relationship of cord blood IA measured by ELISA to birth weight in offspring of 59 mothers with diabetes (either type 1 or insulin-treated gestational diabetes). ELISA is generally less sensitive for measurement of IA (10); in keeping with this, the prevalence of positive IA was relatively low in the series of Weiss et al. (2) (23% of offspring of mothers with type 1 diabetes), and notably, a similar percentage (20%) of offspring of insulin-treated gestational diabetes was IA positive, although the numbers were small. Despite these caveats, the results of our larger study, representing only ODM, are in agreement and would suggest that IA do not influence weight or insulin levels at birth. Indeed, if anything, there was a trend to a negative relationship between the presence of IA and birth weight, opposite to our original hypothesis.

Theoretically, the presence of IA could interfere with assay of insulin and thus our examination of whether IA resulted in higher cord insulin. We do not believe that this is a major problem. The insulin assay used in this study has not been found to be influenced by IA (Halsall, I., personal communication) and detects total insulin (i.e. free and IA-bound insulin). Of course, it remains possible that, in isolated cases, antibodies will have functional effects in utero, perhaps reflecting differences in insulin binding on an individual basis and perhaps, given our data, acting to reduce birth weight. Nevertheless, our data would not suggest that this is a common phenomenon. Our study is also limited to examination of weight and insulin at birth and not earlier in pregnancy. Maternal type 1 diabetes may be associated with complex effects on growth, notably a reduction in growth in early pregnancy as well as the well-documented later increases in birth weight (11). It remains possible then that islet autoantibodies or IA might exert influences on growth and function earlier in pregnancy but have little influence in later pregnancy, when effects of fetal production of insulin will dominate.

	Acknowledgments

 
This study would not have been possible without the contributions of many people to data collection at the different centers as previously noted (3 ). We acknowledge the expert technical help of Annette Knopff and Karolina von Dalwigk in measurement of antibodies for this study.

	Footnotes

 
This work was supported by Project Grant K/MRS/50/C2726 from the Chief Scientist Office of the Scottish Office.

Abbreviations: GADA, Glutamic acid decarboxylase; IA, insulin antibodies; IA-2A, antibodies to the tyrosine phosphatase IA-2; IQR, interquartile range; LMP, last menstrual period; ODM, offspring of mothers with type 1 diabetes.

Received February 4, 2004.

Accepted April 1, 2004.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. Menon RK, Cohen RM, Sperling MA, Cutfield WS, Mimouni F, Khoury JC 1990 Transplacental passage of insulin in pregnant women with insulin-dependent diabetes mellitus. Its role in fetal macrosomia. N Engl J Med 323:309–315[Abstract]
2. Weiss PA, Kainer F, Purstner P, Zehetleitner G, Huttner U, Haas J 1998 Anti-insulin antibodies and birth weight in pregnancies complicated by diabetes. Early Hum Dev 53:145–154[Medline]
3. Lindsay RS, Walker JD, Halsall I, Hales CN, Calder AA, Hamilton BA, Johnstone FD, on behalf of The Scottish Multicentre Study of Diabetes in Pregnancy 2003 Insulin and insulin propeptides at birth in offspring of diabetic mothers. J Clin Endocrinol Metab 88:1664–1671[Abstract/Free Full Text]
4. Naserke HE, Bonifacio E, Ziegler AG 1999 Immunoglobulin G insulin autoantibodies in BABYDIAB offspring appear postnatally: sensitive early detection using a protein A/G-based radiobinding assay. J Clin Endocrinol Metab 84:1239–1243[Abstract/Free Full Text]
5. Ziegler AG, Hummel M, Schenker M, Bonifacio E 1999 Autoantibody appearance and risk for development of childhood diabetes in offspring of parents with type 1 diabetes: the 2-year analysis of the German BABYDIAB Study. Diabetes 48:460–468[Abstract]
6. Roll U, Christie MR, Fuchtenbusch M, Payton MA, Hawkes CJ, Ziegler AG 1996 Perinatal autoimmunity in offspring of diabetic parents. The German Multicenter BABY-DIAB Study: detection of humoral immune responses to islet antigens in early childhood. Diabetes 45:967–973[Abstract]
7. Hamalainen AM, Ronkainen MS, Akerblom HK, Knip M 2000 Postnatal elimination of transplacentally acquired disease-associated antibodies in infants born to families with type 1 diabetes. The Finnish TRIGR Study Group. Trial to Reduce IDDM in the Genetically at Risk. J Clin Endocrinol Metab 85:4249–4253[Abstract/Free Full Text]
8. Koczwara K, Bonifacio E, Ziegler AG 2004 Transmission of maternal islet antibodies and risk of autoimmune diabetes in offspring of mothers with type 1 diabetes. Diabetes 53:1–4[Abstract/Free Full Text]
9. Greeley SA, Katsumata M, Yu L, Eisenbarth GS, Moore DJ, Goodarzi H, Barker CF, Naji A, Noorchashm H 2002 Elimination of maternally transmitted autoantibodies prevents diabetes in nonobese diabetic mice. Nat Med 8:399–402[CrossRef][Medline]
10. Levy-Marchal C, Bridel MP, Sodoyez-Goffaux F, Koch M, Tichet J, Czernichow P, Sodoyez JC 1991 Superiority of radiobinding assay over ELISA for detection of IAAs in newly diagnosed type I diabetic children. Diabetes Care 14:61–63[Abstract]
11. Steel JM, Wu PS, Johnstone FD, Muir BB, Sweeting VM, Hillier SG 1995 Does early growth delay occur in diabetic pregnancy? Br J Obstet Gynaecol 102:224–227[Medline]

This article has been cited by other articles:

				S. E. Fineberg, T. T. Kawabata, D. Finco-Kent, R. J. Fountaine, G. L. Finch, and A. S. Krasner Immunological Responses to Exogenous Insulin Endocr. Rev., October 1, 2007; 28(6): 625 - 652. [Abstract] [Full Text] [PDF]

				A. de Leiva, D. Mauricio, and R. Corcoy Diabetes-Related Autoantibodies and Gestational Diabetes Diabetes Care, July 1, 2007; 30(Supplement_2): S127 - S133. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lindsay, R. S. Articles by Walker, J. D. Search for Related Content PubMed PubMed Citation Articles by Lindsay, R. S. Articles by Walker, J. D.