Insulin resistance in patients with recurrent pregnancy loss is associated with lymphocyte population aberration

ABSTRACT

This study was designed to investigate the relationship of insulin resistance (IR) and cellular immune abnormalities associated with women with recurrent pregnancy loss (RPL). Women with RPL were divided into two groups according to their homeostasis model assessment for IR (HOMA-IR) scores. The IR group received metformin approximately 3 months before pregnancy. The percentage of lymphocyte subsets and other blood biochemical indices were tested. The HOMA-IR and fasting serum insulin levels were related to the percentage of lymphocyte subsets. The women with RPL had higher CD3⁺ and CD3⁺CD4⁺ cell levels while CD56⁺CD16⁺cell levels were lower. A higher likelihood of cellular immune abnormalities was observed. Women with normal lymphocyte subsets had normal pregnancy outcomes. Metformin significantly downregulated CD3⁺ and CD3⁺CD4⁺ cells and improved pregnancy outcomes. IR was associated with cellular immune abnormalities in RPL. The data suggests that metformin affected the immune/inflammatory response, which may regulate the cellular immune balance and improve pregnancy outcomes.

Abbreviations RPL: recurrent pregnancy loss; IR insulin resistance; HOMA-IR: homeostasis model assessment for IR

KEYWORDS:

• Insulin resistance
• lymphocyte population aberration
• metformin
• recurrent pregnancy loss

Introduction

Recurrent pregnancy loss (RPL) refers to two or more pregnancy losses before 20 weeks of gestation, affecting approximately 1-5% of reproductive age women [ASRM 2013]. Its pathogenesis includes parental or embryonic chromosome abnormalities, reproductive organ abnormalities, endocrine abnormalities, infection, immune abnormalities, and environmental factors [Kaur and Gupta 2016]. Many studies have shown that an aberrant lymphocyte population and inflammatory immune response may lead to complications in pregnancy, including RPL [Kim et al. 2014]. And the balance of pro- and anti-inflammatory immune responses at the maternal-fetal interface is crucial for maintaining pregnancy [Kanninen et al. 2013]. Both lymphocyte subsets and cytokine disorders are considered cellular abnormalities. Dysregulated CD3⁺ cell immunity and/or natural killer cell (CD16⁺CD56⁺cell, NK) immunity are associated with RPL. CD3⁺CD4⁺ cells are associated with immunological tolerance in pregnancy. In the classical model of differentiation of lymphocytes, various CD3⁺cell subsets, including Th17 and T regulatory cells, have been examined [Kumagai et al. 2016]. Previous studies have found that the ratio of Th1/Th2 cells and the balance between Treg cells and Th17 cells are connected with the etiology of RPL [Nakagawa et al. 2015, Sereshki et al. 2014]. Once lymphocyte subsets become abnormal, pregnancy outcome is affected.

It has been observed that PCOS may be related to autoimmunity, and insulin resistance (IR) is one of the main characteristics of PCOS [Gonzalez 2012]. IR is characterized by a decreased responsiveness of cells or tissues to normal physiological levels of insulin and has a considerable negative impact on public health. Dysregulation of glucose and fat metabolism are known to induce low-grade inflammation. Additionally, a chronic inflammatory state may contribute to the development of IR. Homeostasis model assessment for insulin resistance (HOMA-IR) is a mathematical model that takes into account the interactions between fasting plasma insulin and fasting plasma glucose concentrations, and is calculated to determine IR [Matthews et al. 1985]. Metformin is the most commonly used oral hypoglycemic agent. It functions by activating the insulin receptor and increasing glucose uptake by enhancing the translocation of glucose transporter I, consequently improving IR [Gunton et al. 2003]. Taking metformin for a period of time can substantially improve glucose tolerance and insulin sensitivity [Konopka et al. 2016]. Moreover, metformin possesses properties that are desirable for autoimmune disease therapy, including anti-inflammatory effects [Tomczynska et al. 2016]. Studies have shown that metformin interferes with the adenosine 5′-monophosphate-activated protein kinase pathway, which inhibits the stimulation of CD3⁺cells, consequently causing an anti-inflammatory effect. In addition, metformin reduces immune suppression by inhibiting the expression of pro-inflammatory mediators [Narender et al. 2009].

Under normal circumstances, the percentage of lymphocyte subset populations maintain a dynamic balance. Flow cytometry is commonly used in the clinic to test peripheral blood lymphocyte subsets. Using this data we can make a preliminary assessment of the an organism’s immunity through the lymphocyte population. Thus far, all previous studies have primarily focused on the cellular immune response alone rather than in association with the RPL and IR. In this study, we aimed to study the HOMA-IR and proportions of lymphocyte subsets of patients with RPL and investigate the effects of metformin pregnancy outcomes in women with RPL and IR.

Results

Characteristics of study participants

A total of 377 participants completed the study. The mean age, BMI, and frequency of abortion were 31.09 ± 3.92 years, 21.73 ±2.30 kg/m², and 2.86 ± 0.98 (range from 2 to 6), respectively. When we analyzed the relationship between IR and cellular immunity, HOMA-IR ≥ 1.60 was defined as IR and HOMA-IR < 1.60 was defined as non-IR. The reference ranges of CD3⁺, CD3⁺CD4⁺, CD3⁺CD8⁺, CD16⁺CD56⁺, and CD19⁺ cells were defined as 52.4-81.4%, 23.9-46.3%, 11.7-40.3%, 8.7-38.3%, and 4.7–19.3%, respectively. Characteristics of the study participants are shown in Table 1A and B. IR patients had a higher level of CD3⁺CD4⁺ cells and a lower level of CD56⁺CD16⁺ cells (p<0.05).

Table 1A. Clinical characteristics of the study population.

	Non-IR group	IR group	P
Age (years)	30.99 ± 3.87	30.75 ± 3.97	0.55
Number of miscarriages	2.90 ± 0.96	2.8 ± 1.01	0.36
BMI (kg/m^2)	21.70 ± 1.99	23.06 ± 2.42	<0.01*
HOMA-IR	1.14 ± 0.31	2.47 ± 0.87	<0.01*
F-PG(mmol/L)	4.78 ± 0.38	5.10 ± 0.39	<0.01*
F-INS (uIU/mL)	5.39 ± 1.39	10.9 ± 3.63	<0.01*
HbA1c (%)	5.06 ± 0.32	5.15 ± 0.39	0.11
CHO (mmol/L)	4.53 ± 0.80	4.42 ± 0.70	0.26
TG (mmol/L)	0.94 ± 0.49	1.11 ± 0.67	0.03*
CD3^+cells (%)	66.92 ± 6.97	69.17 ± 7.24	<0.01*
CD3^+CD4^+cells (%)	35.85 ± 5.77	40.89 ± 7.00	<0.01*
CD3^+CD8^+cells (%)	27.67 ± 7.12	26.91 ± 6.29	0.06
CD3^+CD4^+/CD3^+CD8^+ cell ratio	1.40 ± 0.47	1.72 ± 0.63	<0.01*
CD56^+CD16^+ cells (%)	16.43 ± 7.05	14.74 ± 5.95	0.03*
CD19^+cells (%)	12.15 ± 4.04	12.19 ± 4.43	0.95

BMI: body mass index; HOMA-IR: homeostasis model assessment for insulin resistance; F-PG: fasting plasma glucose; F-INS: fasting serum insulin; HbA1c: glycosylated hemoglobin; CHO: cholesterol; TG: triglyceride.

*significant difference between the two groups; t-test was used to analyze the difference between 2 groups.

Table 1B. Additional clinical characteristics of the study population.

	Lymphocyte populations abnormality group	Lymphocyte populations normality group	P
Age (years)	30.88 ± 3.89	30.89 ± 4.03	0.98
Number of miscarriages	2.87 ± 0.97	2.75 ± 1.05	0.28
BMI(kg/m^2)	22.43 ± 2.37	22.00 ± 1.97	0.38
HOMA-IR	1.85 ± 0.94	1.50 ± 0.80	<0.01*
F-PG(mmol/L)	4.94 ± 0.42	4.87 ± 0.39	0.20
F-INS (uIU/mL)	8.34 ± 3.95	6.79 ± 3.21	<0.01*
HbA1c (%)	5.12 ± 37	5.06 ± 0.33	0.39
CHO (mmol/L)	4.47 ± 0.65	4.47 ± 0.76	0.98
TG (mmol/L)	1.04 ± 0.62	0.96 ± 0.47	0.45

BMI: body mass index; HOMA-IR: homeostasis model assessment for insulin resistance; F-PG: fasting plasma glucose; F-INS: fasting serum insulin; HbA1c: glycosylated hemoglobin; CHO: cholesterol; TG: triglyceride.

*significant difference between the two groups; t-test was used to see whether there was a difference in frequency distribution.

Prevalence of aberrant lymphocyte populations in the study population

A total of 222 patients were found to have at least one or more populations of aberrant lymphocytes such as increased CD3⁺CD4⁺and CD16⁺CD56⁺cell levels. Among the patients with a HOMA-IR score of ≥1.60, 75.82% had a variety of abnormal inflammatory response. However, the percentage in the control group was 43.08%. The frequency of populations of aberrant lymphocytes was different between the two groups (p<0.05). After metformin treatment, only 58.57% of patients with IR had a high level of CD3⁺CD4⁺and/or CD16⁺CD56⁺cells. As shown in Table 2, there was a significant difference in the number of lymphocyte population aberrations between the IR and non-IR groups (p<0.01).

Table 2. Prevalence of lymphocyte population aberration in the study population.

	IR group	Non-IR group	P-value
Number of lymphocyte	138	84
Populations abnormality			P<0.05*
Number of c lymphocyte	44	111
Populations normality

*significant difference between the two groups; chi square test was used to see whether there was a difference in frequency distribution.

It may not be feasible for clinicians to calculate the HOMA-IR score in practice [Wang et al. 2011]. Saxena et al. [2011] showed that the glucose load insulin level obtained using the OGTT can be used as a reliable marker for IR. Patients with IR always had a remarkably higher level of CD3⁺ and CD3⁺CD4⁺cells than patients without IR (p<0.01; Table 3).

Table 3. Patients lymphocyte subset in sub-groups.

	Fasting insulin <10 µIU/mL	Fasting insulin >10 µIU/mL	P
CD3^+cells (%)	66.84 ± 7.33	69.40 ± 6.03	<0.01*
CD3^+CD4^+cells (%)	35.75 ± 6.02	41.19 ± 6.08	<0.01*
CD3^+CD8^+cells (%)	27.9 ± 7.34	24.92 ± 5.16	<0.01*
CD3^+CD4^+/CD3^+CD8^+ cell ratio	1.39 ± 0.50	1.74 ± 0.54	<0.01*
CD56^+CD16^+ cells (%)	16.46 ± 7.11	14.65 ± 5.71	0.02^*
CD19^+cells (%)	12.28 ± 4.23	11.81 ± 3.92	0.34

*significant difference between the two groups; T-test was used to analyze the difference between 2 groups.

Effects of metformin treatment on lymphocytes of women with RPL

The study population was divided into two groups on the basis of their HOMA-IR. Metformin treatment depends on the patients’ HOMA-IR. When the proportions of lymphocyte subsets before and after metformin treatment were compared, the results revealed that metformin treatment significantly affected the CD3⁺cell, CD3⁺CD4⁺cells, and CD19⁺ cell levels and the CD3⁺CD4⁺/CD3⁺CD8⁺ cell ratio, indicating that this drug modulates the cellular immune system. After metformin treatment, the level of CD3⁺ cells, CD3⁺CD4⁺ cells, and CD3⁺CD4⁺/CD3⁺CD8⁺ cell ratio in the IR group decreased (p<0.05). The results are shown in Figure 1. After metformin treatment, only 58.57% of IR patients had a high level of CD3⁺CD4⁺ cells and/or CD16⁺CD56⁺cells. Metformin treatment significantly improved the aberrant lymphocyte population in RPL women with IR (p<0.01).

Figure 1. Comparison and effects of metformin treatment on the proportions of lymphocyte subsets in women with RPL. The study participants were divided into IR and non-IR groups. Peripheral blood was analyzed three times, at the time of preliminary diagnosis, after three months of regular metformin treatment (in patients who had insulin resistance), and when patients became pregnant (human chorionic gonadotropin pregnancy test showed positive). We used t-test to analyze the difference between the two groups. *There was significant difference in the cells compared to that of before metformin treatment.

Relationship between cellular immunity and HOMA-IR

Based on the results of the Pearson’s correlation coefficient, we found that the level of CD3⁺CD4⁺ cells and CD3⁺CD4⁺/CD3⁺CD8⁺cell ratio before pregnancy were positively related to HOMA-IR. The correlation coefficients were 0.45 and 0.32, respectively (p<0.01). To a certain extent, an increased HOMA-IR score indicates an increased percentage of CD3⁺CD4⁺cells (Figure 2). Moreover, the results revealed that lymphocytes were correlated with triglycerides during pregnancy. The partial regression coefficient was 1.01.

Figure 2. Relationship between lymphocytes and HOMA-IR. CD3⁺CD4⁺ cells were closely correlated to the HOMA-IR score. And the CD3⁺CD4⁺ cell level and CD3⁺CD4⁺/CD3⁺CD8⁺ cell ratio increased with the HOMA-IR score; while CD3⁺CD8⁺ cells and CD16⁺CD56⁺ cells showed no significance change after metformin treatment. We used Pearson’s correlation test to analyze the correlation between HOMA-IR and lymphocytes. *There was no statistical association between lymphocyte subset and HOMA-IR.

Treatment outcomes of metformin

The pregnancy outcomes of RPL patients who received treatment for IR were found to be considerably high. Of the 377 patients with RPL, 89.66% had successful pregnancy outcomes. Successful treatment outcomes of RPL women with and without IR were compared (89.56% and 89.74%, respectively). The percentage of successful pregnancies did not significantly differ between the two groups. However, we found that the state of the progestational cellular immune reaction would make the treatment outcomes remarkably different and the pregnancy outcomes significantly differed between the subgroups. Patients with RPL who had no aberrant lymphocyte population were more likely to have successful outcomes. This percentage was even higher among the RPL patients who displayed an improved cellular immune response following metformin treatment (93.33%) than that in the control group. Our results revealed that 12.71% of the patients with IR who still had an aberrant lymphocyte population despite metformin treatment failed to give birth (Figures 3 and 4).

Figure 3. Percentage of RPL patients who delivered successfully in each subgroup. The patients were divided into four subgroups according to whether the subjects had insulin resistance or cellular immune aberration. Each bar stands for the percentage of live births in each subgroup. According to the results of follow-up, most of the patients had a normal labor, and the live birth rate could be up to more than 85% after proper treatment. The patients with PRL and IR had the highest risk of pregnancy failure due to the opposite influence.

Figure 4. Percentage of RPL patients who delivered successfully in each subgroup. The patients were divided into three subgroups according to whether the subjects had insulin resistance and whether metformin treatment improved cellular immune aberration in IR patients. Each bar stands for the percentage of live births in each subgroup. Although the live birth rate between the three groups showed no statistical significance, the patients with poor response to metformin were harder to labor successfully.

Interestingly, analysis of the RPL patients with IR who miscarried despite receiving metformin revealed an increasing trend in the percentage of CD3⁺CD4⁺ lymphocytes. The CD3⁺CD4⁺cell level among 19 of the study subjects with IR who underwent abortions was 40.89 ± 5.65%, while it was 35.09 ± 4.94% among other IR patients whose pregnancies continued for at least 20 weeks of gestation; this difference was statistically significant (p<0.01). Conversely, CD3⁺cells, CD3⁺CD8⁺cells, CD19⁺cells, and CD16⁺CD56⁺cells did not exhibit noticeable differences between the IR and control groups (p = 0.13, 0.06, 0.10, and 0.46, respectively).

Discussion

This study aimed to assess the role of the cellular immune system in RPL patients with IR although IR is not considered to be one of the etiologies of RPL [ASRM 2013]. There is currently no definite morbidity about insulin resistance in recurrent spontaneous abortion patients. According to the Shanghai First Maternity and Infant Hospital Department of Reproductive and Immunology outpatient statistics, the incidence could reach 27.74% among all RPL patients when compared to the proportion of lymphocyte subsets in RPL women with or without IR. The hyperinsuline micoroglycemic clamp technique is the gold standard in the diagnosis of IR [Aljiffry et al. 2016]. Apart from this technique, there is no consensus on the IR standard used. In reality, this technique cannot be practically implemented, and HOMA-IR is more commonly used for IR. Therefore, we analyzed the relationship between lymphocyte subsets and fasting insulin levels.

We found that a greater percentage of newly diagnosed RPL patients with IR had an aberrant lymphocyte population compared to RPL patients without IR. The population of lymphocyte subsets in peripheral blood, especially the value of CD3⁺CD4⁺ cell, are sensitive indexes which can reflect the balance of cellular immunity. Additionally, the abnormal numbers of the peripheral blood CD3⁺cell subsets and CD16⁺CD56⁺cells were related with the occurrence of RPL [Zhu et al. 2015]. Peripheral CD16⁺CD56⁺cell activity and percentage have been shown to decline during the first trimester of normal pregnancies, whereas these parameters are found to be elevated during early pregnancy in RPL women [Yamada et al. 2003]. Patients with RPL who undergo spontaneous pregnancy loss have been found to have a stronger immune reaction than patients who have a normal pregnancy (Shankarkumar et al. 2011). In the present study, we showed that RPL patients with IR had even higher CD3⁺ and CD3⁺CD4⁺cell levels, patients with aberrant lymphocyte populations were more likely to develop IR and RPL, which is consistent with Shi et al. [2008] results. T cell, NK cell, and B cell aberration together constitute the cellular immunity. As one increases, the others will be affected and then decrease. Cellular immunity is activated in PCOS patients characterized by IR. IR is often accompanied by a cellular immune response and chronic low-grade inflammation [Sainz et al. 2015]. Lymphocytes participate in the pathogenesis of IR, this has also been shown in mouse [Solano et al. 2011]. Biochemical pregnancy, recurrent implantation failure, and RPL may occur due to an alteration in the number of lymphocyte subsets, which could then influence the maternal-placental immune tolerance by triggering the immune response [Perez Leiros and Ramhorst 2013]. IR and cellular immunity interact as both cause and effect. This shows that IR in patients with RPL is closely associated with lymphocyte subset population aberration. Furthermore, abnormal cellular immunological reactions have been found to be associated with immunological failure of pregnancy. That is why women with PCOS, which is characterized by IR, have a high rate of abortion [Serdynska-Szuster et al. 2011].

It is well known that metformin improves IR and remarkably reduces the HOMA-IR score [Zawiejska et al. 2016]. Metformin can influence the expression of inflammatory factors and consequently act on the process of differentiation, migration, and infiltration of lymphocytes [Rojas and Gomes 2013]. Metformin therapy in pregnant women with PCOS has been associated with a significant reduction in the rate of early pregnancy loss [Al-Biate 2015] and improves the pregnancy rate in in vitro fertilization repeaters without PCOS [Jinno et al. 2010]. The beneficial effects of metformin therapy could increase the clinical pregnancy rates and reduce the IR in women with PCOS [Abu Hashim 2016]. Collectively, our study indicated that prescription of metformin prior to the pregnancy could not only increase insulin sensitivity and decrease the HOMA-IR score but also improve lymphocyte population aberration, which indicated that IR, cellular immunity, and RPL are interrelated. IR is the primary feature of polycystic ovary syndrome (PCOS) [Ghaffarzad et al. 2016]. Women with PCOS showed a recurrent miscarriage of 40%, while those without PCOS had a recurrent miscarriage rate of 12% [Banu et al. 2014]. The treatment outcomes of RPL patients with IR following metformin treatment was comparable to those of patients with PCOS who did not use metformin [Chakraborty et al. 2013]. Metformin treatment could potentially improve the reproductive outcomes in infertile women with RPL. A subgroup of RPL patients with IR who showed no improvement in lymphocyte population aberration despite receiving metformin treatment miscarried again and showed a trend of increase in the percentage of CD3⁺CD4⁺cells. In our study, treatment with metformin demonstrated a significant improvement in the rate of treatment outcomes.

The incidence of insulin resistance in patients with RPL is high. IR in patients with RPL is associated with lymphocyte population aberration. Metformin can modulate immune reaction imbalance to a certain extent. The results of this study showed that metformin may successfully improve the prognosis of pregnant patients with IR. Pregnancy outcomes were found to be connected with HOMA-IR and a proper cellular immune response in the parent body. In addition, further research to explore the absolute cellular immune count, CD16⁺CD56⁺cytotoxicity, NKT cell, and inflammatory cytokines during pregnancy in RPL patients with IR is warranted.

Materials and methods

Study population

The present study is a retrospective observational study. This research was approved by Ethics Committee of Shanghai First Maternity and Infant Hospital. The study subjects included patients who visited the Department of Reproductive Immunology, between January 1, 2015 and December 31, 2015. The object of the study’s exclusion criteria and excluded were: (1) abnormalities of reproductive system; (2) ureaplasma urealyticum and Chlamydia trachomatis infection; (3) patient karyotype abnormalities; and (4) thrombophilia, autoimmune disease, luteal function defect, hyperprolactinemia, hyperthyroidism, and hypothyroidism. A total of 377 RPL patients were enrolled in this study. And all the subjects signed informed consent.

Study design

There are different values to define IR. Some studies defined IR as HOMA-IR as more than 2.14, while others thought when HOMA-IR >2.69, the patients could be diagnosed as IR [Chen et al. 2006, Miehle et al. 2012]. The insulin resistance standard of HOMA-IR used in our study was established by our lab’s statistics.

Blood samples were drawn at the luteal phase for analysis of the lymphocyte populations, glucose, insulin, HbA1c, cholesterol, and triglyceride levels. The HOMA-IR was then calculated on the basis of these values:

Patients determined to have an abnormal HOMA-IR score were given 1,000 mg of metformin per d (with a meal, twice a d). After the IR patients took metformin for three months, peripheral blood was withdrawn for lymphocyte study. On the day of positive pregnancy test, blood samples were collected to measure the proportions of lymphocyte subsets. The pregnancy outcome was considered to be successful if the pregnancy could be maintained for more than 20 w.

Laboratory examination

Lymphocyte subsets

Four-color flow cytometry was used to study lymphocyte subsets. Whole blood samples were stained with monoclonal antibodies conjugated with fluorochromes (fluorescein isothiocyanate [FITC], phycoerythrin [PE], PE cyanine 5 [PE-Cy 5], and allophycocyanin [APC]) and incubated away from light for 15 min. Erythrocytes were lysed by FACS lysing solution and the blood sample was incubated again. The blood samples were analyzed with a FASCalibur flow cytometer (reagents and equipment, BD Biosciences, USA). A total of 10,000 cells were counted. Multiset software (BD Biosciences) was used for data analysis.

Blood chemistry

Venous blood samples and oral glucose tolerance test were taken on an empty stomach (overnight fasting). The fasting plasma glucose levels were determined by a glucose oxidase-based assay. Fasting insulin levels were measured by chemiluminescence. Glycosylated hemoglobin was detected using high-performance liquid chromatography. The lipid profile included cholesterol and triglyceride concentrations, and was carried out using appropriate enzymatic assays (Beckman Coulter Inc., Brea, CA, USA).

Statistical analysis

All dates are expressed as means ± SD. Lymphocytes, glucose, and insulin levels were conformed to the normal distribution. Statistical analysis was performed using SPSS Statistics (version 22.0; SPSS Inc., Chicago, USA). Independent-samples t test and chi-square test were applied to compare the difference between each group. Pearson’s correlation test and multiple-factor regression analysis were used to analyze the correlation between HOMA-IR and lymphocytes and other factors such as BMI. Statistical significance was set at p<0.05.

Declaration of interests

This work was supported by grants from the Research Grant from Shanghai Committee of Science and Technology (No. 14411966600) and the Joint Key Project of New Frontier Technology in Shanghai Municipal Hospitals (No. SHDC12014129). The authors report no conflicts of interest.

Additional information

Notes on contributors

Yan Yan

Study conception and design: YY; Acquisition of data: WT; Analysis and interpretation of data: LW; Drafting of manuscript: YY, SS; Critical revision: SB.

Shihua Bao

Study conception and design: YY; Acquisition of data: WT; Analysis and interpretation of data: LW; Drafting of manuscript: YY, SS; Critical revision: SB.

Shile Sheng

Study conception and design: YY; Acquisition of data: WT; Analysis and interpretation of data: LW; Drafting of manuscript: YY, SS; Critical revision: SB.

Liuliu Wang

Study conception and design: YY; Acquisition of data: WT; Analysis and interpretation of data: LW; Drafting of manuscript: YY, SS; Critical revision: SB.

Weiyan Tu

Study conception and design: YY; Acquisition of data: WT; Analysis and interpretation of data: LW; Drafting of manuscript: YY, SS; Critical revision: SB.