Evaluation of the expression of genes associated with iron metabolism in Peripheral Blood Mononuclear Cells from Type 2 Diabetes Mellitus patients

Abstract

Aims

Type 2 Diabetes (T2DM) has been linked to ferroptosis. This study aimed to assess expression levels of genes linked with iron metabolism in peripheral blood mononuclear cells (PBMCs) from T2DM patients and to investigate the association of these expression levels with anthropometric and clinical parameters.

Methods

Gene expression of iron metabolism genes (Ferritin Light Chain, FTL; Ferritin Heavy Chain, FTH1; Transferrin Receptor, TFRC; Divalent Metal Transporter 1, SLC11A2; Ferroportin, SLC40A1) in archival PBMCs was assessed using quantitative real-time PCR assays. Correlations of gene expression with anthropometric/biochemical patient data were evaluated.

Results

The study included 36 (18 male/18 female) T2DM patients and 45 (28 male/17 female) normoglycemic (NGT) subjects with a mean age of 38.1 ± 6.8 years and 47.6 ± 8.6 years respectively. Relative expression of FTL was significantly lower in T2DM females compared to that in NGT females (P=0.027). Relative expression of SLC40A1 was significantly lower in the T2DM group (P=0.043) and in the T2DM females (P=0.021). Relative expression of SLC11A2 was negatively correlated with systolic blood pressure in T2DM male patients. Relative expression of SLC40A1 was negatively associated with serum phosphorous and positively associated with serum thyroid stimulating hormone in male T2DM patients.

Conclusions

Our findings indicate a reduction in the expression of FTL in perimenopausal T2DM females. Also, in male T2DM patients and NGT subjects, biochemical markers are significantly correlated with the expression of FTL, FTH1, SLC11A2, and SLC40A1 in PBMCs.

1. Introduction

Diabetes mellitus, a chronic, metabolic disorder characterized by insulin resistance and impaired glucose homeostasis[1], has contributed tremendously to the burden of mortality and disability worldwide[2]. The epidemic of diabetes mellitus and its complications pose a major global health threat. Recently, Type 2 Diabetes (T2DM) has been linked to ferroptosis, a form of nonapoptotic regulated cell death that was first defined in 2012. Notably, ferroptosis is dependent on intracellular iron and insulin secretion dysfunction in pancreatic β-cells[3]. Iron is an indispensable element for metabolic reactions as it is involved in protein synthesis and imbalance in iron metabolism both overload or deficiency can affect vital physiological pathways[4]. Therefore, balanced iron metabolism in the body is an obligatory need of the body to perform metabolic functions properly. Taken together, the monitoring and control of iron metabolism and ferroptosis-related factors may facilitate early diagnosis and therapy of T2DM. However, to date, only a small number of studies have explored the relationship between iron metabolism, ferroptosis, and T2DM.

Developing disease-modifying therapeutics for T2DM necessitates the identification of appropriate pharmacological targets and therapeutic effectiveness measures. To further understand the function of ferroptosis in T2DM in pancreatic - cells and immune effector cells, information on dysregulation, if any, in expression of genes linked with iron metabolism in the context of T2DM, in relevant effector cells is required. Human invasive tissue samples are difficult to collect[5]. However, islet-infiltrating immune effectors are probably in balance with circulating pools and may be collected in peripheral blood mononuclear cells (PBMCs)[6,7].

In this study, we hypothesize that information on the status of iron metabolism associated metabolism might be reflected in the expression of iron metabolism genes in PBMCs. As such, the aim of this study was to assess the expression of genes linked with iron metabolism (Ferritin Light Chain, FTL; Ferritin Heavy Chain, FTH1; Transferrin Receptor, TFRC; Divalent Metal Transporter 1, SLC11A2; Ferroportin, SLC40A1) in PBMCs from T2DM patients and investigate relations between expression of iron metabolism genes in PBMCs with the anthropometric and clinical parameters of the T2DM patients.

Iron is one of the important trace mineral elements involved in several metabolic processes needed to life[8]. A recent study had indicated the role of iron in islet functions under diabetic conditions in human islet amyloid polypeptide-induced diabetes mouse model, and reports that decreased iron level was associated with the dysregulation of glucose stimulated mitochondrial respiration [9]. Also, the body's ferritin level is closely related to ferroptosis. The probable link between increased body iron accumulation and T2DM has been identified by epidemiological investigations[3]. In this view, monitoring changes in the gene expression of iron metabolism markers in PBMCs from T2DM patients may be a crucial first step in understanding how iron metabolism is affected in T2DM patients.

2. Methods

2.1. Study population

This is a retrospective study in which archival PBMC samples and patient data collected were used. Patient data and samples were selected from a larger previous study in which normoglycemic controls (NGT) and type-2 diabetics (T2DM; T2DM status according to the American Diabetes Association criteria[10]: Fasting Blood Glucose >125 mg/dL and/or HbA1c >6.5%; and/or being clinically managed for T2DM using oral hypoglycaemic agents such as metformin) adults (18-60 years old) had been recruited at the Molecular Physiology Lab, Division of Nutrition, St. John's Research Institute, Bangalore from the community in and around of St. John's Research Institute between 2016 and 2021.

T2DM patients (n=36; 18 males, 18 females) and NGT individuals (n=45; 28 males, 17 females) were selected based on complete availability of socio-demographics, anthropometry and clinical chemistry data and of sufficient PBMC samples required for extraction of RNA of appropriate quality and quantity for carrying out gene expression assays.

2.2. Ethical Approval

The samples included in this retrospective analysis were from a previous study that had been approved by the institutional ethics committee of Bangalore's St. John's Medical College and Hospital. Participant-signed, informed consent was obtained at the time of recruiting after the study's approach was explained to them in their native language. The study's procedure was followed in accordance with all applicable guidelines and regulations.

2.3. Socio-demographics, anthropometry and clinical chemistry

A systematic questionnaire was used to collect thorough socio-demographic data from each subject during subject recruitment for the earlier study, from which the samples for this retrospective analysis were taken. The respondents were given the questionnaire in their native language, and competent professionals documented their replies in English. Dual-energy X-ray absorptiometry (DXA; DPXMD 7254, Lunar Corporation, Madison, WI) was used to investigate the anthropometric body composition characteristics. Accuracy in measuring weight and height was 0.1 kg and 0.1 cm, respectively. During the recruitment process, each subject underwent thorough clinical chemistry evaluation that included standard plasma and serum clinical chemistry assays on fasting blood samples, as previously reported[11].

2.4. Collection of PBMCs

During the recruitment process of the previous study, 10 mL of blood had been collected from fasting subjects. Blood was separated into buffy-coat (containing PBMCs) and serum/plasma. The serum/plasma was used for biochemical testing after collection. Buffy-coats had been stored separately at -80°C. The gene expression experiments in this work were conducted using these frozen buffy coat samples.

2.5. RNA extraction, quality checks and cDNA preparation

RNA was extracted from frozen buffy coats using Trizol reagent (T9424; Sigma-Aldrich, St. Louis, MO, USA), followed by phase separation using chloroform. Briefly, total RNA was extracted from the aqueous phase and precipitated using an equal volume of isopropanol. The resulting RNA pellet was washed in 75% ethanol, air-dried, and solubilized in diethylpyrocarbonate treated water. The Take3 Micro-Volume Plate reader (Synergy H1 Hybrid Multi-Mode Reader, BioTek Instruments Inc, Winooski, USA) was used to determine the total RNA's purity. The extracted RNA samples with 260/280 and 260/230 ratios of 1.8 - 2.2 and ≥1.8, respectively, were chosen for further processing[12,13].

Total RNA was quantified using Ribogreen reagent (Quant-iT Ribogreen RNA Assay kit, Invitrogen, Eugene, USA) on a fluorescent microplate reader (Synergy H1 microplate reader, BioTek Instruments Inc, Winooski, USA). Following this, a DNase I kit (Sigma Aldrich, St. Louis, USA) was used to treat the obtained RNA with DNAse. The DNAse-treated samples were reverse transcribed to obtain cDNA using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Luthiana) in accordance with the kit manufacturer’s instructions. To determine the transcript abundance of the target genes in the obtained cDNA, qRT-PCR was used.

2.6. qRT-PCR Assays

Expression of the iron-metabolism-genes FTL, FTH1, TFRC, SLC11A2, SLC40A1 were assessed in PBMCs from T2DM patients and NGT subjects. Reference genes YWHAZ and ACTB were used[14], based on our earlier study on identification of stable reference genes in PBMC samples from T2DM patients and NGT subjects. Primer sequences for qRT-PCR assays were designed for the target genes (FTL, FTH1, TFRC, SLC11A2, SLC40A1) (Supplementary Table 1). Primer sequences for qRT-PCR assays of the references genes is available from Hazarika et al[14]. 10 μl qRT-PCR reactions were set up in duplicates. PCR cycling conditions were: 95°C for 10 min, then 40 cycles of 95°C for 15 s, followed by 60°C for 1 min. All qRT-PCR reactions were performed using SYBR Green master mix (Power SYBR Green PCR Master Mix, Thermo Fisher Scientific, USA) on a QuantStudio 6 Flex Real- Time PCR system (Thermo Fisher Scientific, USA).

cDNA derived from the Universal Human Reference RNA (Agilent Technologies, Santa Clara, CA, USA) was used to construct standard curves for the qRT-PCR assays. The qRT-PCR assays for each gene had PCR efficiencies and correlation coefficients for linearity that varied from 81.732 to 97.509% and 0.997 to 1.000, respectively (Supplementary Fig.S1). A modification of the ΔCt method using ACTB and YWHAZ as reference genes [ΔCt =Ct_{(target gene)} - Ct_{(Reference gene mean)}][15] was used for calculation of relative expression of the iron-metabolism-genes. The ΔCt values were converted to relative normalization units (RNU = 15 - ΔCt) according to Mukhopadhyay et. al[16].

2.7. Statistical analysis

The study subjects' baseline anthropometric and clinical parameters were represented as mean ± SD. The Shapiro-Wilk test and Q-Q plots were used to determine if distributions were normal. In order to compare the baseline parameters, and relative transcript abundances of the target genes (FTL, FTH1, TFRC, SLC11A2, SLC40A1) between the study groups; Student's t-test (for normally distributed data) or Mann Whitney U test (for non-normally distributed data) were used. The correlations between anthropometric clinical variables and the relative expression of iron-metabolism genes in PBMCs were examined using Spearman's rank correlation coefficient. Microsoft Excel 2016 (Microsoft Corporation, Redmond, WA, USA), and GraphPad Prism (Version 8, GraphPad Software, San Diego, CA, USA) were used for all statistical analyses.

3. Results

3.1. Anthropometric and clinical characteristics of the study subjects

The anthropometric and clinical characteristics of the 81 study subjects are summarized in Table 1. Weight, height, BMI and blood pressure did not have any significant difference between the NGT and T2DM groups. The T2DM group was older and had higher levels of HbA1c, fasting glucose, C-peptide and Triglyceride, amongst all 81 study subjects and separately amongst male and female subjects. The age of the T2DM females (45.9 ± 6.82 years) was significantly (P=0.025) higher as compared to the NGT females (39.7 ± 8.47 years).

Table 1.

Subject characteristics and metabolic profile of the 81 study subjects. P<0.05 are shown in bold.

	All samples (n=81)	NGT (n=45)	T2DM (n=36)	P	NGT Males (n=28)	T2DM Males (n=18)	P	NGT Females (n=17)	T2DM Females (n=18)	P
Gender (M/F)#	46/35	28/17	18/18	0.270
Age (years)	42.3 ± 8.94	38.1 ± 6.81	47.6 ± 8.57	<0.001	37.2 ± 5.53	49.3 ± 9.93	<0.001	39.7 ± 8.47	45.9 ± 6.82	0.025
Weight (kg)	65.8 ± 9.73	67.1 ± 10.2	64.1 ± 8.97	0.171	72.5 ± 8.51	67.5 ± 9	0.070	58.2 ± 5.42	60.8 ± 7.77	0.263
Height (cm)	161.6 ± 9.63	163.5 ± 10.1	159.1 ± 8.58	0.035	169.3 ± 7.56	165.8 ± 5.27	0.074	154.1 ± 5.42	152.4 ± 5.30	0.353
BMI (kg/m2)	25.1 ± 2.46	25.0 ± 2.43	25.3 ± 2.52	0.601	25.3 ± 2.16	24.5 ± 2.66	0.327	24.6 ± 2.85	26.1 ± 2.16	0.093
Waist-Hip Ratio (cm)	0.90 ± 0.07	0.90 ± 0.06	0.91 ± 0.07	0.48	0.92 ± 0.05	0.95 ± 0.05	0.029	0.86 ± 0.06	0.86 ± 0.07	0.971
Body fat (%)	35.1 ± 7.68	34.7 ± 6.91	35.6 ± 8.62	0.621	30.7 ± 4.69	28.9 ± 6.15	0.297	41.5 ± 4.15	42.4 ± 4.30	0.522
Fat mass (kg)	22.1 ± 5.31	22.2 ± 4.68	22.0 ± 6.06	0.001	21.5 ± 4.99	19.1 ± 5.82	0.154	23.4 ± 4	25.0 ± 4.83	0.282
Appendicular lean mass (kg)	41.1 ± 8.26	42.3 ± 8.71	39.7 ± 7.54	<0.001	48.1 ± 5.17	45.8 ± 4.83	0.142	32.7 ± 2.79	33.6 ± 3.87	0.432
Bone Mineral Content (kg)	2.49 ± 0.46	2.59 ± 0.49	2.37 ± 0.41	<0.001	2.87 ± 0.39	2.61 ± 0.36	0.029	2.12 ± 0.18	2.14 ± 0.3	0.854
Android Fat (%)	43.6 ± 6.62	43.5 ± 6.38	43.7 ± 7.00	0.912	42.1 ± 6.17	40.2 ± 6.74	0.322	45.8 ± 6.22	47.2 ± 5.39	0.478
Gynoid Fat (%)	39.6 ± 9.04	40.0 ± 8.18	39.0 ± 10.1	0.624	34.6 ± 4.54	30.2 ± 5.51	0.008	49.0 ± 3.41	47.8 ± 3.82	0.344
Android: Gynoid ratio	1.14 ± 0.21	1.11 ± 0.19	1.17 ± 0.23	0.286	1.22 ± 0.15	1.34 ± 0.17	0.024	0.93 ± 0.09	0.99 ± 0.11	0.105
Systolic blood pressure (mmHg)	120.1 ± 15.0	118.2 ± 13.8	122.5 ± 16.3	0.219	121.3 ± 11.7	129.8 ± 15.5	0.056	113.2 ± 15.9	115.2 ± 13.8	0.705
Diastolic blood pressure (mmHg)	80.4 ± 8.58	80.3 ± 9.52	80.6 ± 7.37	0.887	81.9 ± 8.06	82.8 ± 8.19	0.732	77.6 ± 11.3	78.3 ± 5.86	0.810
Haemoglobin, Hb% (mg/dL)	14.0 ± 1.92	14.3 ± 2.01	13.6 ± 1.75	0.123	15.4 ± 1.04	14.7 ± 0.93	0.027	12.5 ± 1.95	12.6 ± 1.74	0.901
CRP (mg/dL) α∞	0.43 ± 0.43	0.33 ± 0.38	0.54 ± 0.46	0.035	0.32 ± 0.33	0.43 ± 0.46	0.414	0.35 ± 0.44	0.65 ± 0.45	0.054
Calcium (mg/dL) α β	9.22 ± 0.58	9.13 ± 0.55	9.34 ± 0.61	0.104	9.24 ± 0.59	9.39 ± 0.65	0.435	8.94 ± 0.42	9.29 ± 0.58	0.051
Fasting Glucose (mg/dL)^	130.3 ± 74.4	85.0 ± 9.32	186.9 ± 81.3	<0.001	87.2 ± 8.61	165.8 ± 72.2	<0.001	81.3 ± 9.54	208.1 ± 86.3	<0.001
HbA1c (%)^	6.83 ± 2.13	5.26 ± 0.35	8.81 ± 1.72	<0.001	5.26 ± 0.34	8.36 ± 1.76	<0.001	5.25 ± 0.38	9.26 ± 1.61	<0.001
C-Peptide (ng/mL) α	3.00 ± 1.52	2.44 ± 1.00	3.71 ± 1.77	<0.001	2.66 ± 1.01	4.14 ± 1.81	0.004	2.08 ± 0.89	3.28 ± 1.66	0.012
Fasting Insulin (mU/L)α β	11.6 ± 7.56	10.6 ± 7.19	13.0 ± 7.91	0.161	11.0 ± 6.87	11.0 ± 4.9	0.962	9.91 ± 7.85	15.0 ± 9.93	0.105
Total Protein (g/dL) α	7.50 ± 0.63	7.43 ± 0.63	7.60 ± 0.63	0.229	7.45 ± 0.64	7.72 ± 0.54	0.143	7.39 ± 0.62	7.48 ± 0.71	0.677
Albumin (g/dL) α	4.23 ± 0.36	4.23 ± 0.35	4.24 ± 0.39	0.878	4.35 ± 0.33	4.36 ± 0.4	0.936	4.04 ± 0.29	4.13 ± 0.35	0.403
Aspartate aminotransferase, AST (IU/L) α	27.7 ± 14.1	26.4 ± 11.9	29.4 ± 16.5	0.361	28.5 ± 14.1	29.6 ± 14.1	0.789	23 ± 6.2	29.2 ± 19.1	0.203
Alanine Transaminase, ALT (IU/L) α	32.1 ± 16.5	31.4 ± 17.4	33.1 ± 15.4	0.642	35.6 ± 18.8	36.7 ± 13.1	0.820	24.4 ± 12.2	29.4 ± 17.0	0.315
Alkaline Phosphatase, ALP (U/L) α	73.6 ± 40.6	62.2 ± 20.3	87.9 ± 53.5	0.009	61.7 ± 19.3	81.9 ± 20.9	0.002	63.1 ± 22.6	93.8 ± 73.4	0.106
Gamma-glutamyl transferase, GGT (U/L) α	51.7 ± 86.5	32.2 ± 16.3	76.1 ± 125.2	0.044	37.5 ± 17.3	68.8 ± 68.5	0.073	23.4 ± 9.84	83.3 ± 165.8	0.144
Cholesterol (mg/dL) α	193.3 ± 42.2	179.2 ± 36.2	211 ± 42.9	0.001	182.5 ± 34.3	209.8 ± 46.8	0.042	173.6 ± 39.8	212.2 ± 40.0	0.007
HDL Cholesterol (mg/dL) α	43.1 ± 8.99	43.2 ± 8.25	43.1 ± 10.0	0.974	42.8 ± 7.3	40.6 ± 9.84	0.435	43.9 ± 9.82	45.6 ± 9.70	0.604
LDL Cholesterol (mg/dL) α	123.7 ± 30.2	118.4 ± 27.7	130.4 ± 32.1	0.079	120.5 ± 28	132.1 ± 38.1	0.278	114.8 ± 27.8	128.8 ± 25.8	0.134
Triglycerides (mg/dL) α	157.0 ± 88.1	117.0 ± 52.3	207.0 ± 98.4	<0.001	130.4 ± 49.7	209.1 ± 84.7	0.002	95.1 ± 50.3	204.9 ± 112.9	0.001

Blood was drawn from fasting volunteers. Data presented as mean ± SD. P values are from the independent t-test. ^#P value from Chi-square test; ^α Biochemical Analysis in serum; ^ Biochemical Analysis in plasma; ^β n=80, ^∞n=76

3.2. Relative expression of iron-metabolism-genes in NGT and T2DM PBMCs

Relative expression of FTL (using ACTB and YWHAZ as reference genes) in NGT and T2DM groups, overall, and within males was not significantly different (Fig. 1a, 1b), and that of FTH1, TFRC, and SLC11A2 in NGT and T2DM groups, either overall, or separately, within males or females, was not significantly different (Fig. 1d-1l, Supplementary table 2). Relative expression of FTL was significantly lower in T2DM females compared to that in NGT females (P=0.027) (Fig. 1c). Relative expression of SLC40A1 was significantly lower in the T2DM (All samples) group (P=0.043) (Fig.1m) and in the T2DM females (P=0.021) (Fig.1o) as compared to NGT (All samples) group and NGT females respectively.

Figure 1.

Boxplots of relative expressions of FTL (a-c), FTH1 (d-f), TFRC (g-i), SLC11A2 (j-l) and SLC40A1 (m-o) expressed as relative normalization units (RNU) with ACTB and YWHAZ as reference genes[16], in PBMC samples from: NGT (n=45) and T2DM patients (n=36) within all study subjects (n=81) (a, d, g, j, m); NGT (n=28) and T2DM (n=18) subjects within male subjects (b, e, h, k, n) and NGT (n=17) and T2DM (n=18) subjects within female subjects (c, f, i, l, o). Median values are represented as lines across the box; lower and the upper boxes represent the first and third quartile, respectively; whiskers represent the maximum and minimum value. P values are from Mann Whitney test.

3.3. Anthropometric/clinical parameters in relation to expression of iron-metabolism genes in PBMCs

As iron status and metabolism is intricately related to whole body metabolism, relations between relative expression of iron-metabolism genes in PBMCs and anthropometric and biochemical parameters of the subjects were next investigated (Fig. 2 and 3, Supplementary Table 3). Relative expression of SLC11A2 was negatively correlated with systolic blood pressure in male T2DM patients. Relative expression of SLC40A1 was negatively associated with serum phosphorous (Fig. 2f) and positively associated with serum thyroid stimulating hormone (Fig. 2i) in male T2DM patients. Though these relations were observed overall in the T2DM group, they were specific to the male T2DM patients and were absent in the female T2DM patients.

Figure 2.

Scatter plots of correlations between relative expressions (expressed as relative normalization units (RNU) with ACTB and YWHAZ as reference genes[16]) of iron-metabolism genes and anthropometric/clinical parameters. Relative expression of SLC40A1 and systolic blood pressure (a-c); Serum Phosphorous (d-f) and Serum Thyroid stimulating hormone (g-i) in PBMC samples from: NGT (n=45) (b, e, h); T2DM patients (n=36) (c, f, i) and within all study subjects (n=81) (a, d, g).

Figure 3.

Scatter plots of correlations between relative expressions (expressed as relative normalization units (RNU) with ACTB and YWHAZ as reference genes[16]) iron-metabolism genes and anthropometric/clinical parameters. Relative expression of FTL and serum albumin (a-c); of FTH1 and serum albumin (d-f); of FTL and Serum Alkaline phosphatase (g-i); of SLC11A2 and Serum Alkaline phosphatase (j-l); and of SLC40A1 (m-o) in PBMC samples from: NGT (n=45) (b, e, h, k, n) and T2DM patients (n=36) (c, f, i, l, o) within all study subjects (n=81) (a, d, g, j, m).

In the NGT group, the relative expressions of FTL (Fig. 3b) and of FTH1 (Fig.3e) were negatively correlated to serum albumin. Relative expression of FTL (Fig.3h) and SLC11A2 (Fig.3k) were negatively correlated with Serum alkaline phosphatase levels while that of SLC40A1 was negatively correlated with serum aspartate aminotransferase (Fig. 3n). These relations were observed specifically in male NGT subjects and were absent in female NGT subjects.

4. Discussion

The role of iron homeostasis and glucose metabolism is widely evidenced in recent studies[17,18]. In the current study, we have compared the relative expression levels of the iron metabolism genes FTL, FTH1, TRFC, SLC11A2 and SLC40A1 in PBMCs from T2DM and NGT subjects. We have further investigated the correlation between anthropometric and biochemical parameters with relative transcript abundance of iron-metabolism genes in PBMCs in a sex-specific manner, which to the best of our knowledge has not been reported till date.

Female T2DM patients had lower relative expression of FTL, compared to female NGT subjects. The iron storage protein ferritin's light chain component is encoded by the FTL gene. Sex-specific dysregulation of FTL in diabetic PBMCs has not been reported before, even though elevated serum ferritin levels has been underlined as one of the risk factors of T2DM in a recently published meta-analysis and systematic review[8]. Though the most prominent regulator of cellular ferritin is labile iron, it regulates cellular ferritin through a post-transcriptional mechanism[19]. In comparison, oxidative stress can increase FTL transcription[19]. The link between T2DM and oxidative stress is well known[20]. However, sex of the patient needs to be a consideration, as Grindel et al did not observe any difference in oxidative stress parameters amongst female-only T2DM patients grouped according to HbA1c cut-off />7.5%[21]. Further, lack of consideration of sex of the patients could explain equivocal results from clinical trials using antioxidant vitamins for diabetes progression or treatment[22]. Future sex-stratified longitudinal studies are needed for illuminating the metabolic implications of lower FTL expression in female T2DM PBMCs.

Relative expression of SLC40A1, that codes for Ferroportin, an iron transporter postulated to play roles in intestinal iron absorption and cellular iron release[23], was significantly lower in PBMCs from T2DM patients in our study. This could have been due to diabetic PBMCs being near constantly in a high insulin environment, as Qiu et al has recently demonstrated that in vitro insulin and/or palmitate treatment of human hepatocytes led to decrease in SLC40A1 transcription[24]. Earlier, lipopolysaccharide-treatment induced inflammation has been reported to transcriptionally repress SLC40A1 in primary mice tissue macrophages[25].Though the role of ferroportin in recycling of RBC heme iron by spleen and liver macrophages is well known[26], its role in iron homeostasis in PBMCs has not been explored in depth, especially in the context of T2DM associated dysglycaemia. Interestingly, the reduced expression of SLC40A1 was observed specifically in female T2DM patients, when compared to female NGT subjects. It remains to be experimentally assessed in future studies if female diabetic PBMCs are more vulnerable to the diabetic environment (higher blood glucose and insulin for longer periods of time) than the male ones, as relevant to dysregulation of iron-metabolism.

Relative expression of SLC40A1 was positively associated with serum thyroid stimulating hormone (TSH) in male T2DM patients. Yanshi et. al. have recently reported positive association of blood iron levels with free triiodothyronine (FT3) in a cross-sectional study on 1067 Chinese adults, that included 9.5% diabetics [27]. They also reported negative association between blood copper levels and TSH. Though T2DM and thyroid dysfunction are known to go hand-in-hand[28], their relation with PBMC iron regulation or the role of sexual dimorphism in it has not yet been explored.

Relative expression of SLC11A2 was negatively associated with systolic blood pressure in T2DM males. Tzoulaki et. al. reported an inverse association of dietary total iron intake and non-haem iron intake with blood pressure in adults, with ~1:1 male:female ratio[29]. Sexually dimorphic associations of PBMC SLC11A2 expression with blood pressure have not been reported before and the underlying mechanisms remain to be explored further.

In NGT males, we observed negative association of relative expression of FTL and FTH1 in PBMCs with serum albumin. Serum albumin levels have been reported to be negatively associated with the severity of intravenous iron-induced oxidative stress in anaemic chronic renal failure patients[30]. Whether PBMC FTL and FTH1 expression acts as a surrogate for iron status of NGT males and is similarly negatively associated with iron status-associated oxidative stress, remains to be addressed. The serum alkaline phosphatase levels were negatively associated with the expression of FTL and SLC11A2. The serum alkaline phosphatase levels are known to be associated with systemic inflammation, measured as serum CRP levels[31] but sexual dimorphism of these observations have not been reported or investigated. Wu et. al. demonstrated that decreased expression of SLC11A2 in intestinal mucosa leads to compromised absorption and transportation of iron [32]. The role of SLC11A2 is also implicated in ferroptosis [33]. In this view, the ferroptosis-associated generation of ROS could be the reason for the CRP associated inflammation in NGT males. Further research is needed to determine the underlying cause of this sex-specific observation regarding the relative expression of SLC11A2 in PBMCs.

We observed negative association between relative expression of SLC40A1 which encodes the protein ferroportin and serum AST in NGT males. Similar to our findings, Liu et. al. in their study on mice reported that administration of a drug with side effects on the liver showed disordered systemic iron homeostasis via the upregulation of the hepcidin-ferroportin axis[34]. Though ferroportin plays an important role in iron recycling from senescent red blood cells by hepatic and splenic macrophages by acting as the sole exporter of iron from these cells[35] and thereby is crucially involved in regulating hepatic iron stores[36], its role in PBMCs in regulating liver function in a sexually dimorphic manner has not been described or investigated till date.

The strengths of our study include availability of detailed anthopometric, biochemical and clinical data from the subjects that formed part of this study and in-depth analysis of gene expression status of iron metabolism genes in PBMCs from South Asian Indian NGT and T2DM subjects. In addition, subject sex-specific analyses helped us uncover sexual dimorphism in associations between regulation of iron-metabolism genes in PBMCs and glycaemic and metabolic status. Due to the limited availability of samples, one limitation of the current investigation is the number of iron metabolism genes we were able to analyse in our study samples. Also, due to logistical reasons, age matching between the NGT and T2DM groups could not be achieved. A recent study that included 1,089,270 people with T2DM registered on the Australian diabetes registry followed from July 2010 to June 2019, reported that the median age for the onset of T2DM was 58.2 years (48.7, 25th centile 67.2, 75th centile)[37]. This could be the reason for the higher age of the T2DM subjects included in this study. As such, age is a likely contributor to the differences in expression of iron metabolism genes that we observe. The associations for the gene expression of FTL and SLC40A1 in PBMCs with age of the subjects is yet to reported.

To conclude, our findings from this study show a sex-specific reduction in the expression of FTL gene in PBMCs from T2DM female patients who were of perimenopausal age (45.9 ± 6.82 years). According to reports, perimenopausal symptoms in Indian women start to appear at 44.69 ± 3.79 years[38]. Also, we report that the expression of SLC40A1, which plays a key role in ferroptosis signalling was significantly reduced in female T2DM patients. Lack of existing literature to understand the causes and consequences of female-specific reduction in FTL and SLC40A1 expression in PBMCs underlines the need for future, in-depth sex-specific mechanistic studies on T2DM in human subjects. We also conclude that biochemical parameters are closely associated with the expression of iron metabolism genes in PBMCs. We report the following correlations in male T2DM patients- negative associations of SLC11A2 expression with systolic blood pressure, of SLC40A1 expression with serum phosphorous; and a positive association of SLC40A1 expression with thyroid stimulating hormone. In normoglycemic male subjects we report the following negative correlations- FTL expression with serum albumin and serum alkaline phosphatase; FTH1 expression with serum albumin, SLC11A2 expression with serum alkaline phosphatase, SLC40A1 expression with serum aspartate aminotransferase. Further research investigating the underlying reasons for the observed sex specific correlations in male T2DM patients and male normoglycemic subjects is warranted.