Immune cell infiltration correlates with intestinal permeability, inflammation, and gastrointestinal symptoms in type 1 diabetes

ABSTRACT

Aims/Introduction

Our study explores interconnections between the occurrence of gastrointestinal (GI) symptoms with immunohistochemical analysis of colon biopsies, markers of intestinal permeability, and inflammation in individuals with type 1 diabetes.

Material and Methods

Twenty subjects with type 1 diabetes and seven healthy controls underwent colonoscopy. Colon biopsy materials were analyzed immunohistochemically for CD3+, CD20+, CD4+, CD8+, CD138+, and CD68+ cell counts. The levels of lipopolysaccharide (LPS), LPS‐binding protein (LBP), and endogenous anti‐endotoxin core antibodies (EndoCAb IgG and IgM) were measured in serum. Fecal calprotectin, immunoglobulin A (IgA), albumin, protein, and intestinal alkaline phosphatase (IAP) activity were analyzed in patient subgroups.

Results

Immune cell infiltration in lamina propria did not differ between type 1 diabetes and control subjects. In type 1 diabetes, the number of CD20+, CD8+, CD138+, and CD68+ cells correlated with several GI symptoms and usage of medications for diarrhea. Fecal calprotectin correlated positively with the number of CD20+ B cells, CD3+ T cells, and CD138+ plasma cells. A negative correlation was found between CD20+ B‐cell number and fecal IgA, while CD68+ macrophage number correlated positively with fecal albumin. Serum EndoCAb IgM correlated negatively with CD138+ plasma cells and CD4+ T cells, and positively with CD68+ macrophages. Serum LBP correlated negatively with CD4+ T cells, demonstrating links between gut mucosal inflammation and the systemic response to endotoxin.

Conclusions

In type 1 diabetes, CD immune cell infiltration in the colon mucosa tended to correlate with fecal and systemic markers of inflammation and gastrointestinal symptoms. A key direction for future studies will be to elucidate the underlying pathogenic mechanisms.

In individuals with type 1 diabetes, immune cell infiltration in the colon mucosa tended to correlate with fecal and systemic markers of intestinal permeability and inflammation, as well as gastrointestinal symptoms. These findings suggest that low‐grade gut immune activation might link the intestinal barrier dysfunction with systemic endotoxin responses.

INTRODUCTION

Type 1 diabetes is an autoimmune disease initiated by T‐cell‐mediated destruction of the insulin‐producing β‐cells in the pancreas ¹ . In 2025, there are about 9.5 million people living with type 1 diabetes globally, and the incidence is increasing. Type 1 diabetes leads to chronic hyperglycemia and potential organ damage due to vascular complications (diabetic kidney disease, diabetic retinopathy, peripheral neuropathy, and cardiovascular disease) ² . The onset of type 1 diabetes has been linked to a combination of genetic, epigenetic, environmental, and lifestyle factors, including gut microbiota dysbiosis ³ , ⁴ , ⁵ , which contributes to immune system dysregulation and increased intestinal permeability ⁶ , ⁷ .

The occurrence of gut permeability in type 1 diabetes was found in functional and histological research conducted in animal models and patients ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² . Gastrointestinal (GI) symptoms are twice as prevalent in individuals with type 1 diabetes as compared with the general population ¹³ and may be linked to intestinal permeability. Intestinal permeability ¹⁴ potentially elevates the risk of diabetes complications via promotion of low‐grade inflammation and poor glycemic control ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ .

The pathogenesis of GI disorders in type 1 diabetes and their interrelations with systemic inflammation are little studied. Existing studies show that the gut mucosa in type 1 diabetes displays disease‐specific abnormalities in the inflammatory profile and microbiota ²⁰ , ultrastructural alterations ¹⁰ , and signs of chronic inflammation ²¹ , ²² . However, the correlation between infiltrating immune cells, for example, CD20+, CD3+, CD4+, CD8+ lymphocytes, and CD68+ macrophages involved in gut mucosal immunity and gut inflammatory conditions ¹⁰ , ²³ , ²⁴ , ²⁵ has not previously been studied in individuals with type 1 diabetes.

Compelling evidence suggests that the concentrations of fecal biomarkers, shown to be associated with low‐grade gastrointestinal inflammation and gut permeability (fecal propionate, butyrate, calprotectin, intestinal alkaline phosphatase (IAP), fecal immunoglobulin (IgA)), are associated with type 1 diabetes itself or its complications ¹⁷ , ²⁶ .

A disrupted intestinal barrier can lead to leakage of gram‐negative bacteria‐derived lipopolysaccharides (LPS), also frequently called endotoxins, into the circulation and thereby promoting chronic inflammation and oxidative stress ²⁷ . The resulting low‐grade inflammation is characterized by increased circulating LPS and its binding proteins (endogenous anti‐endotoxin core antibodies (EndoCAb IgG and IgM), and LPS‐binding protein (LBP)) ²⁸ . To the best of our knowledge, no study has so far investigated the immune cell subsets of the colon mucosal low‐grade inflammation in correlation with functional markers of intestinal permeability (LPS and its binding proteins) and fecal markers associated with gut permeability in individuals with type 1 diabetes. To fill this gap, our study explores the interconnections between the occurrence of gastrointestinal symptoms with histopathological and immunohistochemical analysis of colon biopsies, fecal markers of inflammation, and variables associated with systemic endotoxemia in subjects with type 1 diabetes.

MATERIALS AND METHODS

Participants and study design

This study is part of the longitudinal LatDiane study, initiated in 2013 (and participating in the international InterDiane consortium). LatDiane recruits adult individuals with type 1 diabetes diagnosed before the age of 40 with insulin treatment initiated within 1 year of diagnosis and C‐peptide levels below 0.3 nmol/L ²⁹ , ³⁰ . The protocol of the general LatDiane and the sub‐study devoted to gut health were approved by the Latvian Central Ethics Committee No. 01‐29.1/3 (dated July 10, 2013), and its further amendment (Nr. 01‐29.1.2/3274 issued on 23.04.2021). Recruitment of the study participants, biobanking, and sample storage were performed in accordance with the procedures of the Genome Database of the Latvian population ³¹ . This study is in line with the 1964 Declaration of Helsinki and its later amendments. Written informed consent was obtained from all study participants prior to inclusion in this study.

Recruitment for this sub‐study took place between 4th of November 2021 and 20th of February 2023 in Latvia, Riga. The inclusion criteria were: for the type 1 diabetes group – diabetes duration of at least 8 years; for the control group—normal glucose metabolism, defined as fasting glucose ≤5.6 mmol and glycated hemoglobin (HbA1c) ≤ 5.7% ³² . The exclusion criteria for both groups were pregnancy, a history of inflammatory bowel disease, coeliac disease, acute intestinal infection within 2 months of the planned fecal sample collection, clinical signs of acute inflammation, and fever. The study design is summarized in Figure 1. During the prescreening phase, exclusion criteria were evaluated, and eligible subjects were invited to a screening visit.

Figure 1.

Scheme of study design. T1D, type 1 diabetes.

On the screening visit day, the anthropometric measures were collected, a previously validated questionnaire ²² consisting of questions about previous gastrointestinal conditions, the usage of medications for gastrointestinal diseases and symptoms, and a scale including 17 questions (Appendix S1) was filled out with the assistance of the study personnel, blood samples were collected, and the participants received instructions and tubes for the collection of the fecal samples.

Blood, urine and fecal samples, and estimated glucose disposal rate

Blood samples were collected by standard methods from an antecubital vein for serum and ethylenediaminetetraacetic acid plasma. Urine was collected for the determination of albumin/creatinine ratio. Blood and morning spot urine samples were sent for assessment to a certified clinical laboratory. Stool samples were collected within 2 weeks after blood collection and assessment of gastrointestinal symptoms. Participants collected their fecal samples at home using sterile collection tubes without buffer (the collection date and time were recorded). Upon arrival to the laboratory, fecal samples were frozen as soon as possible at −80° until further analysis.

Endoscopic investigation

Participants with increased calprotectin levels (>50 μg/g) and/or gastrointestinal tract symptoms had a consultation by a gastroenterologist for further recommendations and planning of endoscopic investigations. Clinically relevant criteria, such as weight loss, abdominal pain, anemia, diarrhea, constipation, or any changes in bowel movements, bloating, family history of polyposis, inflammatory bowel disease, or colorectal cancer, previous diagnosis of gastrointestinal polyps, or a screening colonoscopy were performed for participants over the age of 50 years. Of the 47 participants with type 1 diabetes who were offered a colonoscopy, only 21 accepted the invitation. One participant underwent endoscopy outside the study setting and was therefore excluded. The main reasons for nonacceptance were complicated preparation procedure for the colonoscopy with the necessity of fasting and intensive glucose control, fear of complications during the preparations for the endoscopy and the procedure itself, or poor health status. As a result, 20 participants with type 1 diabetes underwent colonoscopy. Among the control subjects, 11 were offered/referred to colonoscopy, and 7 accepted the invitation. Median time between screening visit and the colonoscopy was 216 (167–473) days, average time 296 (SD = 189) days, minimal time 3 days, and maximal time 621 days.

Participants were prescribed a polyethylene glycol (PEG)‐based osmotic laxative (PEG‐3350, sodium sulfate, sodium chloride, potassium chloride, sodium ascorbate, and ascorbic acid for oral solution) for bowel cleansing. Two intestinal mucosa biopsies were taken from each investigated gut region, according to accepted protocols.

Histological examination

The morphological analysis of the biopsies was done by a certificated pathologist according to international guidelines to exclude microscopic colitis (collagenous or lymphocytic colitis), eosinophilic colitis, and other potential causes of inflammation, as well as to characterize the mucosa and to perform immunohistochemical analyses. Colon biopsy analyses were performed with the hematoxylin and eosin (HE) and the immunochemical staining methods. HE staining was performed with an automatic Daco CoverStainer CS 100 staining machine (Agilent, Santa Clara, USA) and immunohistochemical staining with Autostainer Link 48 (Agilent, Santa Clara, USA), using standard protocols. Both internal and external controls were used. Internal controls included lymphoid aggregates within the biopsy sections. As an external positive control, appendix tissue was used for all markers. Cells were considered positive when a clear chromogenic (brown) signal was observed in the expected cellular compartment. The characteristics of the primary antibodies used in the study and the approach to interpret the positive immunohistochemical reactions are summarized in Appendices S2 and S3.

Markers of intestinal permeability and inflammation

For the measurement of serum intestinal permeability markers, blood samples were incubated undisturbed for 30 min at room temperature and then centrifuged. Serum was removed from the pellet and transferred into fresh 2 mL tubes, frozen and stored at −80°C until analysis. Serum LPS activity was measured using a Hycult LAL chromogenic endpoint assay (HIT302, Hycult Biotech, Uden, The Netherlands). EndoCAb IgG and IgM were measured by Hycult EndoCAb IgG and IgM Elisa kits (HK504‐IGG; HK504‐IGM; Hycult Biotech, Uden, The Netherlands). LBP was measured using a Hycult LBP Human Elisa kit (HK315‐02, Hycult Biotech, Uden, The Netherlands). All measurements were performed according to the manufacturer's instructions. Fecal calprotectin was measured using an Alegria® Calprotectin Elisa kit (REF ORG280, Organotech Diagnostika GmbH, Budapest, Hungary) in a certified clinical laboratory. IAP activities in fecal samples were measured by an in‐house colorimetric method ³³ .

Fecal IgA levels were measured by a Secretory IgA ELISA (IC6100, Lot: IC6100‐221220, Immuchrom GmbH, Hessen, Germany), while fecal albumin was measured using Human Albumin ELISA Kit (E88‐129, Lot: E88‐129‐230,406, Bethyl Laboratories, Waltham, MA, USA). IAP activities, IgA, and albumin levels were normalized to the protein concentrations in fecal extracts. Protein levels were determined using DC Protein Assay (BioRad, Hercules CA, USA).

Statistical analyses

Statistical analyses were performed using R statistical software (version 4.5.2; last accessed on December 2, 2025) ³⁴ . Continuous data were tested for normality using the Shapiro–Wilk test. Most of the analyzed variables violated the normality assumption; therefore, the results are reported as medians and interquartile ranges (Q1–Q3).

Comparisons between two independent samples were performed using the Mann–Whitney U test. Equality of proportions in categorical variables was tested using Fisher's exact test for count data.

Pairwise correlations were evaluated between six immune cell markers and six intestinal biomarkers, as well as among the immune cell markers themselves, resulting in two predefined families of correlation analyses comprising 36 and 15 correlation pairs, respectively. Spearman's rank correlation coefficients were calculated separately for the entire cohort and for the subgroup of participants with type 1 diabetes. To account for multiple testing, the Benjamini–Hochberg false discovery rate (FDR) correction was applied within each family of tests for both the whole cohort and the type 1 diabetes subgroup. For each correlation, 95% percentile bootstrap confidence intervals were computed using 1,000. Both unadjusted P‐values and FDR‐adjusted q‐values are reported, and correlations with FDR < 0.10 were considered statistically significant for exploratory interpretation. Additional correlations outside these predefined test families are reported for exploratory purposes only.

The GI symptom score was calculated as a mean score of the reported symptoms, including the severity and frequency of abdominal pain and/or discomfort, appetite loss, heartburn, bitter or sour taste, abdominal fullness or bloating, difficulty swallowing, nausea, vomiting, and diarrhea.

P‐values below the 5% significance level (P < 0.05) were considered statistically significant, while P‐values between 5% and 10% (0.05 < P < 0.1) were considered marginally significant, considering the small sample size of the study groups.

RESULTS

Description of study participants

The study sample consists of 20 individuals with type 1 diabetes and seven controls. Characteristics of the study groups are summarized in Table 1. The majority of the type 1 diabetes participants (n = 12; 60%) had arterial hypertension as well as other complications of diabetes (diabetic retinopathy and sensorimotor peripheral neuropathy n = 17 (70%), dialysis n = 3 (15%), kidney transplantation n = 2 (10%)). Notably, eight participants in the type 1 diabetes group had calprotectin levels above 50 μg/g, while none of the subjects in the control group had elevated calprotectin levels. Assessment of GI symptoms showed no differences between study groups in symptom scores, symptom prevalence, or the use of medications for GI disorders (Appendices S4 and S5).

Table 1.

Study group description

	Control (N = 7)	Type 1 diabetes (N = 20)	P‐value
Male sex (n, %)	3 (42.9%)	6 (30%)	>0.99
Age, years	39 (36–46)	51 (37–56)	0.24
Diabetes duration, years	–	29 (23–39)	–
BMI, kg/m2	22.8 (21.3–28.6)	25.9 (23.9–30.4)	0.38
Hypertension (n, %)	1 (14.3%)	12 (60%)	>0.99
Smoking (n, %)	2 (28.6%)	5 (25%)	>0.99
Diabetic retinopathy (n, %)	–	14 (70%)	–
Diabetic sensorimotor peripheral neuropathy (n, %)	–	14 (70%)	–
Dialysis (n, %)	–	3 (15%)	–
Kidney transplantation (n, %)	–	2 (10%)	–
Autoimmune thyroiditis (n, %)	1 (14.3%)	7 (35%)	>0.99
Psoriasis (n, %)	–	2 (10%)	–
History of gastritis/duodenitis (n, %)	3 (42.9%)	9 (45%)	>0.99
History of gastrointestinal surgery (n, %)	3 (42.9%)	10 (50%)	>0.99
Plasma glucose, mmol/L	4.8 (4.7–5.1)	8.5 (6.3–12.2)	<0.001
eGFR, mL/min/1.73 m2	92 (91–100)	98 (80–106)	0.81
HbA1c, %	5.4 (5.4–5.5)	8.4 (7.4–10.4)	<0.001
Total bilirubin, mmol/L	10.0 (6.5–15.7)	8.3 (6.0–10.1)	0.33
AST, U/L	25 (21–27)	25 (21–36)	0.66
ALT, U/L	20 (18–22)	23 (18–35)	0.41
GGT, U/L	20 (15–26)	20 (15–30)	0.91
Total cholesterol, mmol/L	5.30 (3.93–5.43)	4.52 (4.11–5.03)	0.61
HDL cholesterol, mmol/L	1.30 (1.21–1.60)	1.79 (1.40–2.07)	0.13
LDL cholesterol, mmol/L	2.81 (1.99–3.82)	2.28 (1.86–2.73)	0.18
Triglycerides, mmol/L	0.89 (0.70–1.18)	1.17 (0.74–1.46)	0.39
CRP, mg/L	0.50 (0.50–0.74)	0.85 (0.59–3.32)	0.11
ESR, mm/h	8.0 (5.0–10.5)	9.50 (6.25–19.25)	0.39
Hemoglobin, g/L	148 (129–157)	135 (123–145)	0.32
Calprotectin, ug/g	19.7 (16.6–25.0)	24.7 (4.0–101.7)	0.69
Calprotectin >50 (n, %)	0 (0%)	8 (40%)	>0.99
Albumine/creatinine ratio in urine, mg/mmol	0.33 (0.21–0.94)	1.16 (0.67–2.30)	0.053
Gastrointestinal symptome score	1.12 (1.07–1.26)	1.21 (1.06–1.54)	0.74

Continuous data are presented as medians with interquartile ranges (Q1–Q3), and categorical data are presented as counts and percentages. P‐values are reported from the Mann–Whitney U test for continuous variables and Fisher's exact test for categorical variables. ALT, alanine transaminase; AST, aspartate transferase; BMI, body mass index; CRP, C‐reactive protein; ESR, erythrocyte sedimentation rate; GGT, gamma‐glutamyl transferase; GFR, glomerular filtration rate; HbA1c, glycated hemoglobin; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein.

Colonoscopy biopsy results and immunohistochemical interpretation

Median biopsy area was 8.8 (5.9–11.6) mm²; in the type 1 diabetes group 8.5 (6.0–12.6) mm² and 9.8 (7.0–10.2) mm² in control group (P = 0.91). For each biopsy, the number of crypts was counted. The median number of crypts was smaller in the type 1 diabetes group compared with the control group, but the difference was not statistically significant (type 1 diabetes: n = 448 (308–601); control: n = 516 (315–750); P = 0.89). The histopathological findings in the HE staining are given in Table 2.

Table 2.

Histopathological findings in hematoxylin and eosin staining

Histopathological findings	Control group (N = 7)	Type 1 diabetes group (N = 20)
Infiltration with eosinophils	2 (29%)	7 (35%)
Lymphoid follicles/lymphoid aggregates	4 (57%)	6 (30%)
Lymphoplasmacytic infiltration	2 (29%)	4 (20%)
Lymphocyte and macrophage infiltration	1 (14%)	3 (15%)
Mononuclear cell infiltration	1 (14%)	2 (10%)
Stromal fibrosis	2 (29%)	3 (15%)

Data are presented as counts (%). Fisher's exact test revealed no differences in any histopathological findings between study groups (all P > 0.99).

The number of CD3+, CD4+, CD8+, CD20+, CD68+, CD138+ cells in lamina propria did not differ significantly between controls and individuals with type 1 diabetes (Table 3; Figure 2). Several examples of histopathological findings and immune cell infiltration in intestinal mucosa are depicted in Appendix S6.

Table 3.

Immunophenotype of the infiltrated cells in intestinal mucosa

CD antigen positive cells per mm2	Whole cohort (N = 27)	Control (N = 7)	Type 1 diabetes (N = 20)	r rb	P‐value
CD3/mm2	912 (816–1,048)	896 (808–1,000)	928 (816–1,116)	−0.15 (small)	0.58
CD4/mm2	608 (456–704)	624 (472–680)	592 (460–840)	−0.03 (negligible)	0.93
CD8/mm2	352 (304–464)	320 (312–368)	416 (300–480)	−0.19 (small)	0.49
CD20/mm2	10 (6–16)	4 (2–12)	11 (8–17)	−0.42 (moderate)	0.108
CD68/mm2	2 (1–2)	2 (1.5–2)	2 (1–2)	0.09 (negligible)	0.71
CD138/mm2	1,552 (1168–1920)	1,184 (1024–1,568)	1,680 (1296–2008)	−0.34 (moderate)	0.20

Data are presented as median (Q1–Q3). Between‐group differences were assessed using the Mann–Whitney U test. Effect sizes are expressed as the rank‐biserial correlation ($r_{\mathrm{rb}}$).

Figure 2.

Immune cells infiltration in the colon mucosa in individuals with type 1 diabetes and controls.

Markers of intestinal permeability and inflammation in serum and fecal samples

Neither did the serum concentrations of LPS, LBP, EndoCAb IgM and IgG, nor the calprotectin concentrations differ between the individuals with type 1 diabetes and the controls. It is however noted that fecal IAP, IgA, and albumin could be performed in only 16 participants (type 1 diabetes: n = 13; control: n = 3). Therefore, these results should be interpreted with caution although the results demonstrated that the levels of fecal IAP and fecal IgA were lower in the individuals with type 1 diabetes as compared with controls (Table 4).

Table 4.

Markers of intestinal permeability and inflammation in serum and feces

	Intestinal permeability and inflammation markers	Whole cohort (N = 26)	Control (N = 6)	Type 1 diabetes (N = 20)	r rb	P‐value
Serum markers	LPS, EU/mL	0.54 (0.47–0.67)	0.44 (0.40–0.53)	0.58 (0.49–0.69)	−0.43 (moderate)	0.24
	LBP, μg/mL	9.2 (6.9–13.5)	9.2 (6.3–13.7)	9.4 (7.0–12.7)	−0.21 (small)	0.79
	EndoCAb IgM, MMU/mL	41 (25–58)	39 (30–54)	41 (21–59)	−0.01 (negligible)	0.61
	EndoCAb IgG, GMU/mL	98 (55–169)	77 (58–101)	112 (53–174)	−0.34 (moderate)	0.42
Fecal markers	Calprotectin, μg/g	20.0 (7.4–84.8)	19.7 (16.6–25.0)	24.7 (4.0–101.7)	−0.24 (small)	0.69
	IAP, U/mg*	79 (40–275)	2,103 (1,137–2,710)	59 (24–117)	−0.49 (moderate)	0.025
	IgA, μg/g*	159 (63–408)	1,080 (739–1,786)	88 (58–257)	−0.50 (large)	0.043
	Albumin, ng/mg*	17.6 (6.6–89.5)	23.3 (12.6–29.0)	12.1 (7.6–170.5)	−0.77 (large)	0.69
	Protein, mg/mL*	2.36 (1.90–2.44)	2.25 (2.19–2.29)	2.40 (1.91–2.45)	−0.81 (large)	0.44

Data are presented as median (Q1–Q3). Between‐group differences were assessed using the Mann–Whitney U test. Effect sizes are expressed as the rank‐biserial correlation ($r_{\mathrm{rb}}$).

^*For biomarkers fecal IAP, fecal IgA, fecal IgA/protein, fecal albumin, fecal total protein N = 16 (T1D = 13, control = 3).

EU/mL, endotoxin units per milliliter; IgA, immunoglobulin A; IgG, immunoglobulin G; IgM, immunoglobulin M; IAP, intestinal alkaline phosphatase; LPS, lipopolysaccharides; LBP, LPS‐binding protein; μg/mL, micrograms per milliliter; μg/g, micrograms to grams; mg/mL, milligramms to milliliter; MMU/mL, milli‐Merck units per milliliter; ng/mg, nanograms to milligrams; U/mg, units per milligram of protein. Bold values indicate statistically significant P‐value results.

Correlations between immune cell infiltration and markers of intestinal inflammation and permeability

In patients with type 1 diabetes, we observed a positive correlation between CD20+ for B‐lymphocytes and CD3+ for T cells in lamina propria (r = 0.47; 95% CI: −0.08, 0.81), as well as between CD3+ and CD4+ for T cells (r = 0.58; 95% CI: 0.06, 0.89), CD3+ and CD8+ for T cells (r = 0.55; 95% CI: 0.09, 0.84). CD138+ for T cells correlated positively with CD4+ (r = 0.45; 95% CI: 0.03, 0.76).

Moreover, in the individuals with type 1 diabetes, fecal calprotectin correlated positively with the infiltration of CD20+ cells (r = 0.42; 95% CI: −0.07, 0.81), CD3+ cells (r = 0.42; 95% CI: −0.02, 0.74), and with CD138+ cells (r = 0.42; 95% CI: 0.01, 0.79).

In individuals with type 1 diabetes, we found a negative correlation between CD4+ and EndoCAb IgM (r = −0.53; 95% CI: −0.86, −0.02) and LBP (r = −0.51, 95% CI: −0.81, −0.12). Additionally, EndoCAb IgM correlated negatively with CD138+ (r = −0.50; 95% CI: −0.81, −0.04) and positively with macrophages/histiocytes (r = 0.51; 95% CI: 0.13, 0.77). CD8+ correlated positively with EndoCAb IgG (r = 0.47; 95% CI: −0.01, 0.81). Correlations between immune cell infiltration and markers of intestinal inflammation and permeability are presented in Figure 3, together with unadjusted P‐values and Benjamini–Hochberg FDR‐adjusted q‐values.

Figure 3.

Pairwise Spearman correlations between mucosal immune cell infiltration and biomarkers of intestinal inflammation and permeability (a), as well as among the mucosal immune cell infiltration markers themselves (b) in patients with type 1 diabetes. Data are presented as the Spearman correlation coefficient (r) with its 95% bootstrap confidence interval, together with the corresponding unadjusted P‐value and FDR‐adjusted q‐value (Benjamini–Hochberg procedure). Cal, calprotectin (μg/g); LBP, lipopolysaccharide‐binding protein (μg/mL); EndoCAb IgM, immunoglobulin M (MMU/mL); EndoCAb IgG, immunoglobulin G (GMU/mL); IgA, immunoglobulin A (μg/g); Albumin: (ng/mg).

Correlations between immune cell infiltration, clinical factors, gastrointestinal symptoms, and medication usage

In individuals with type 1 diabetes, a negative correlation was found between the frequency of use of medication for diarrhea and CD20+ B‐lymphocytes (r = −0.49; 95% CI: −0.77, −0.06) as well as CD138+ plasma cells (r = −0.47; 95% CI: −0.75, −0.03) in the lamina propria. The frequency of usage of probiotic capsules correlated positively with the CD68+ macrophage/histiocyte infiltration (r = 0.52; 95% CI: 0.11, 0.78). The frequency of difficulties to swallow correlated positively with CD8+ cells in lamina propria (r = 0.45; 95% CI: 0.01, 0.74), and macrophages/histiocytes CD68+ (r = 0.45; 95% CI: 0.01, 0.75) as well as with CD3+ lymphocyte expression (r = 0.39; 95% CI: −0.06, 0.71).

Gastrointestinal symptom scores showed a positive correlation with CD138+ cell infiltration in the lamina propria (r = 0.45; 95% CI: 0.08, 0.71) and with CD8+ cells (r = 0.34; 95% CI: −0.05, 0.64) in the whole cohort. A similar association between symptom scores and CD8+ infiltration was observed in the type 1 diabetes subgroup (r = 0.41; 95% CI: −0.05, 0.72). No clear associations were observed between gastrointestinal symptom scores and infiltration by other CD+ immune cell populations.

DISCUSSION

Our study evaluated the correlations between infiltration of T‐ and B‐lymphocyte subtypes, as well as CD68+ macrophages/histiocytes into the colon mucosa and markers of intestinal permeability and inflammation, as well as gastrointestinal symptoms in individuals with type 1 diabetes.

To the best of our knowledge, this is the first study to analyze CD3+, CD4+, CD8+ T lymphocytes, CD20+ B lymphocytes, CD138+ plasma cells and CD68+ macrophages/histiocytes in the colon mucosa along with the serum concentrations of endotoxin (LPS) and its binding proteins (LBP, EndoCAb IgG and IgM) as well as fecal markers calprotectin, IAP, IgA, and albumin in individuals with type 1 diabetes.

The main finding was that the representation of lymphocyte subtypes in the colon mucosa correlated with serum and fecal intestinal permeability and inflammation markers in type 1 diabetes. For example, a negative correlation between CD4+ and CD138+ with EndoCAb IgM, as well as CD4+ and LBP might indicate a link between colonic expansion of pro‐inflammatory cells and endotoxemia.

Previously, it was shown that endotoxemia can modulate CD4+ T‐cell differentiation in the gut mucosa in animal models ³⁵ . On the other hand, CD4+ T‐cell expansion can promote colonic inflammation, potentially promoting endotoxemia ³⁶ . Increased mucosal presence of CD138+ plasma cells was previously observed in subjects with gut inflammatory conditions and was associated with nonresponsiveness to anti‐inflammatory treatment ³⁷ . Serum EndoCAb IgM levels exhibit dynamic changes in response to endotoxemia. Initially, endotoxemia may lead to a depletion of LPS‐binding antibodies, followed by a transient increase in both EndoCAb IgM and LBP. As the condition resolves, EndoCAb IgM levels tend to decrease further ³⁸ . The observed negative correlations between EndoCAb IgM or LBP and colonic CD4⁺ T cells and CD138⁺ plasma cells, along with the positive correlations between EndoCAb IgM and CD68⁺ macrophages/histiocytes, and between EndoCAb IgG and CD8⁺ T cells, may suggest a relationship between gut inflammation, endotoxin translocation into the bloodstream, and the subsequent immune response ³⁹ , ⁴⁰ .

Our study also demonstrated positive correlations between fecal calprotectin, and the number of CD20+ and CD3+ lymphocytes as well as CD138+ plasma cell expression in colon mucosa in individuals with type 1 diabetes. Calprotectin, sensitive as an inflammatory marker but with low specificity, has been found to be elevated in stool samples in a variety of digestive tract disorders ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ and even in nondigestive diseases like septicemia, meningitis, pneumonia, juvenile idiopathic arthritis, and glomerulonephritis ⁴⁶ , ⁴⁷ . In fact, studies have suggested that calprotectin can help to distinguish between functional and organic gastrointestinal disorders ⁴⁴ . In our study, inflammatory bowel diseases and coeliac disease, as well as acute inflammatory disease were exclusion criteria, therefore the correlations between calprotectin and the presence of lymphocyte subpopulations in the gut mucosa identified by us could be explained by chronic low‐grade mucosal inflammation driven by diabetic gastroenteropathy. To support the link between diabetic gastroenteropathy manifestations and inflammatory infiltration in bowel mucosa, we observed positive correlations between difficulties to swallow, and CD8+, CD68+, and CD3+ cell infiltration, as well as between the gastrointestinal symptom scores and CD138+ plasma cell number. Others have observed that lymphoplasmacytic infiltration of the gut mucosa can be a predictor of more severe colitis ⁴⁸ . This fact, together with our data discussed above, indicates that even normal‐range variability in the calprotectin levels might be associated with activation of the mucosal immune system and gastrointestinal symptoms. Although endothelial injury and dysfunctional endothelial–leukocyte interactions promote neutrophil recruitment and activation within the mucosa and microvessels, that can raise calprotectin ⁴⁹ , ⁵⁰ .

We observed a negative correlation between CD20+ cells and fecal IgA, suggesting that a more pronounced mucosal inflammation is linked to a lower level of protective IgA antibodies, secreted by CD138 plasma cells ⁵¹ . According to published evidence, an increased number of CD20+ lymphocytes correlates with histologic disease activity in ulcerative colitis ⁵² , ⁵³ . In addition, by demonstrating the positive correlation between CD68+ macrophages/histiocyte numbers in the gut mucosa and fecal albumin, a potential marker of intestinal permeability ⁵⁴ , ⁵⁵ , our study supports previous data on associations between mucosal permeability and inflammation ⁵⁶ . Previously, an increased number of CD68+ macrophages was reported in the colon of patients with Crohn's disease ⁵⁷ .

We did not observe any statistically significant differences in the counts of the investigated inflammatory cells, serum intestinal permeability markers, and fecal calprotectin between those with TID and controls. We think that the lack of difference in the studied markers might be explained by the low number of control subjects (n = 7). In addition, although subjects in the control group had normal glucose metabolism, they were referred to colonoscopy based on the same indications as for those with type 1 diabetes (mainly gastrointestinal symptoms), which means that their gastrointestinal health was not optimal.

We also examined the levels of fecal IAP, IgA, and albumin in a subset of our cohort (13 type 1 diabetes and 3 controls). We found lower levels of the fecal protective factors IAP and IgA in type 1 diabetes as compared with controls, which is in agreement with previous studies ¹⁷ , ³⁴ , ⁵³ . However, these data should be interpreted with caution due to the small number of participants in this sub‐analysis.

The limitations of our study include the cross‐sectional nature of this study, which did not allow us to evaluate any causal relationships between immune cell infiltration in the gut mucosa and markers of intestinal permeability and inflammation. The relatively low number of subjects studied (especially in the control group) also warrants cautious interpretation of the results.

Another limitation is that the serum and fecal sample collections for the determination of intestinal permeability and inflammation markers took place on average 296 (SD = 189) days before the colonoscopy, due to the logistics of the study. However, the participants received no treatment for their gastrointestinal complaints (such as anti‐inflammatory medications), which might have impacted the mucosal condition during the waiting time for colonoscopy. In addition, previous research shows that subclinical histopathological changes in the gut mucosa (e.g., following infection) can be stable over months in patients with irritable bowel syndrome and sustained gastrointestinal symptoms ⁵⁸ , ⁵⁹ . The histological findings of the colon mucosa were also stable in the majority of individuals undergoing colon interposition for a mean of 57 months after the procedure ⁶⁰ . Finally, we are aware that the bowel preparation procedure before the colonoscopy using osmotic laxatives may cause a variety of histological alterations itself ⁶¹ . Potentially, this limitation an be mitigated by substituting the colonoscopy with sigmoidoscopy in the future studies, which does not require bowel preparation with laxatives.

Intestinal dysfunction in type 1 diabetes likely arises from the combined effects of autonomic (and enteric) neuropathy and microvascular alterations, although current studies do not provide a direct quantitative partitioning of their relative contributions ⁶² , ⁶³ , ⁶⁴ .

The major strength of the study is the immunohistochemical analysis of colon biopsies along with the assessment of serum LPS, LBP, EndoCAb IgM, EndoCAb IgG, fecal calprotectin, fecal IAP, and fecal IgA, as well as evaluation of gastrointestinal symptoms in individuals with type 1 diabetes and a control group with normal glucose metabolism.

In type 1 diabetes, CD immune cell infiltration in the colon mucosa tended to correlate with fecal and systemic markers of inflammation and gastrointestinal symptoms. This suggests a connection between intestinal immune system activation and systemic inflammatory response. A key direction for future studies will be to elucidate the underlying pathogenic mechanisms driving these immune–epithelial interactions, which may open new avenues for targeted intervention.

FUNDING

Project lzp‐2020/1–0138 ‘Dissecting the associations between glucose variability, intestinal derangements and progression of diabetic nephropathy in type 1 diabetes’ funded by the Latvian Council of Science. Research was supported by Folkhälsan Research Foundation, Finnish Diabetes Research Foundation, Wilhelm and Else Stockmann Foundation, and Liv och Hälsa Society.

AUTHOR CONTRIBUTIONS

PZ: analysis, endoscopic procedures, data curation, writing – original draft, editing, methodology. RI, LP: analysis, data curation, methodology, writing – review and editing. A.F.: study coordination, recruitment, analysis, data curation, methodology, writing – review and editing. E.K.: endoscopic procedures, writing – review and editing. JN, SD: histological analysis, methodology, writing – review and editing. HS, ML: analysis of fecal markers, writing – review and editing. K.J.: analysis of serum markers, writing – review and editing. P‐H.G., UR: writing – review and editing. NS: review and editing. JS: conceptualization, data curation, formal analysis, methodology, writing – original draft, review and editing. All authors have read and approved the final version of the manuscript for submission.

DISCLOSURE

PHG has received lecture honoraria from Astellas, Astra Zeneca, Bayer, Berlin Chemie, Boehringer Ingelheim, Eli Lilly, EloWater, Genzyme, Medscape, MSD, Mundipharma, Novartis, Novo Nordisk, Peer Voice, Sanofi, and Sciarc. PHG is an advisory board member for AbbVie, Astellas, Astra Zeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Medscape, MSD, Mundipharma, Nestlé, Novartis, Novo Nordisk, and Sanofi. PHG has received investigator‐initiated grants from Eli Lilly and Roche. The other authors declare no conflict of interest.

Approval of the research protocol: The protocol for this research project has been approved by a suitably constituted Ethics Committee of the institution and it conforms to the provisions of the Declaration of Helsinki.

Informed consent: All informed consent was obtained from the subjects.

Registry and the registration no. of the study/trial: Committee of the Latvian Central Ethics No 01–29.1/3 (dated July 10, 2013), and its further amendment (Nr. 01–29.1.2/3274 issued on 23.04.2021).

Animal studies: N/A.

Supporting information

Appendix S1. Gastrointestinal symptoms questionnaire.

Appendix S2. Characteristics of the primary antibodies used in the study.

Appendix S3. Characteristics of the interpretation of a positive immunohistochemical reaction.

Appendix S4. Comparison of gastrointestinal symptoms in type 1 diabetes and control groups.

Appendix S5. Usage of medications related to gastrointestinal disorders.

Appendix S6. Histopathological findings and immune cell infiltration in intestinal mucosa.

DATA AVAILABILITY STATEMENT

The data that supports the findings of this study are available in the Appendices S1– S6 of this article.