Bariatric Surgery Induces Alterations in the Immune Profile of Peripheral Blood T Cells

Low-grade inflammation is closely linked to obesity and obesity-related comorbidities; therefore, immune cells have become an important topic in obesity research. Here, we performed a deep phenotypic characterization of circulating T cells in people with obesity, using flow cytometry. Forty-one individuals with obesity (OB) and clinical criteria for bariatric surgery were enrolled in this study. We identified and quantified 44 different circulating T cell subsets and assessed their activation status and the expression of immune-checkpoint molecules, immediately before (T1) and 7–18 months after (T2) the bariatric surgery. Twelve age- and sex-matched healthy individuals (nOB) were also recruited. The OB participants showed higher leukocyte counts and a higher percentage of neutrophils. The percentage of circulating Th1 cells were negatively correlated to HbA1c and insulin levels. OB Th1 cells displayed a higher activation status and lower PD-1 expression. The percentage of Th17 and Th1/17 cells were increased in OB, whereas the CD4+ Tregs’ percentage was decreased. Interestingly, a higher proportion of OB CD4+ Tregs were polarized toward Th1- and Th1/17-like cells and expressed higher levels of CCR5. Bariatric surgery induced the recovery of CD4+ Treg cell levels and the expansion and activation of Tfh and B cells. Our results show alterations in the distribution and phenotype of circulating T cells from OB people, including activation markers and immune-checkpoint proteins, demonstrating that different metabolic profiles are associated to distinct immune profiles, and both are modulated by bariatric surgery.


Introduction
Obesity is a major health concern.It can trigger a number of diseases and conditions, including insulin resistance (IR), type 2 diabetes mellitus (T2D), and cardiovascular disease (CVD) [1,2].Obesity-associated chronic low-grade inflammation is a common feature of these comorbidities [1]; therefore, the immune system has become a major focus of investigation of obesity [2,3].
Innate and adaptive immune cells play an important role in maintaining homeostasis and a perfect balance between anti-inflammatory and pro-inflammatory cell populations, including those infiltrated in the adipose tissue (AT) microenvironment.However, the adipokine-releasing patterns associated with dysfunctional adipocytes, present in obesity, prompts an imbalance within the AT-infiltrated immune cell populations.Increased production of interleukin (IL)-6, tumor necrosis factor (TNF)-α, and leptin by dysfunctional adipocytes not only modulates the local inflammation by induction of pro-inflammatory polarization of immune cells but also plays an important role in systemic inflammation [4][5][6].Furthermore, a chronic pro-inflammatory profile of immune cells may have a deleterious impact on metabolism, contributing to insulin resistance and T2D development.Neutrophils and, to a higher extent, monocytes/macrophages have been described as major contributors to obesity-related low-grade inflammation and insulin resistance development [7][8][9].
Adaptive immune cells also play an important role in obesity-related inflammation.CD4 + (Th) and CD8 + (Tc) T lymphocytes have been described as displaying an altered phenotype and function in obesity and metabolic dysfunction [10,11].In fact, studies suggest that T cells display a metabolic switch toward pro-inflammatory polarization, such as Th1 and Th17 [8,12,13].Moreover, both the number and the percentage of circulating CD4 + Treg cells within CD4 + T cells are reduced in adults with morbid obesity [14].In turn, important phenotypic alterations are present in CD8 + Treg cells from individuals with obesity (our unpublished data).And changes in the functional compartments of CD8 + T cells have also been described in people with morbid obesity and the metabolic syndrome [6,15].Despite these findings, little is known about the CD8 + T cell phenotype, either in obesity or T2D.Moreover, the information available about less-represented T cell subsets is even scarcer.
In fact, most of the available literature focuses on a restricted type and number of T cell subsets, such as Th1, Th17, or Treg [8,14].Thus, there is lack of information on less-represented T cell subsets, which may play an important role in obesity-associated low-grade inflammation.In addition, a deep evaluation of the circulating immune cells in obesity, covering most T cell subsets and providing the big picture of T cell changes in obesity, is missing.Therefore, the present study describes to a large extent the circulating T cells in people with obesity at different metabolic states.The absolute counts and percentages of 44 different T cell subsets, which are further characterized in terms of activation status and expression of immunosuppressive molecules, are assessed in this study, in participants with obesity and in a non-obese control group.Additionally, the impact of bariatric surgery on each of the circulating T cell subsets analyzed has also been studied.

Participants
hypertension, clinical diagnosis of T2D, or sleep apnea.Acute inflammatory conditions, cancer, and neurodegenerative and autoimmune diseases, as well as the use of immunosuppressive and anti-inflammatory drugs on a daily basis, were exclusion criteria.From the recruited participants, 35 were studied before bariatric surgery and 14 after bariatric surgery, as indicated in Figure 1 and Table 1.Eight of the participants were studied both before and after bariatric surgery, while six were studied exclusively after the bariatric surgery (Figure 1 and Supplementary Table S1).Only individuals with Class IV obesity undergo a second bariatric surgery.An age-and sex-matched group of healthy individuals (n = 12; 8 women and 4 men; mean age: 43 ± 11.9 years old) were also recruited to participate as a control group (non-obese group, nOB).A volume of 12 mL of EDTA-anticoagulated fasting peripheral blood (PB) was collected from all participants.

Anthropometric and Biochemical Characterization of the Participants
The anthropometric and biochemical/metabolic characteristics of the participants are detailed in Table 1.A scale (seca 515 mBCA, seca, Hamburg, Germany) was used to measure the body weight (kg).The body mass index (BMI, kg/m 2 ) was calculated as previously described [16]: BMI = BW (kg)/height 2 (m 2 ).The waist (WC), hips (HC), and neck (NC) circumferences were measured using a flexible measuring tape (cm), while a digital sphygmomanometer (Philips IntelliVue MP20, Philips, Boebligen, Germany) was used to measure the systolic and diastolic blood pressure (mmHg) after 5 min of rest.
Leptin and adiponectin levels were measured in plasma, by immunoassay, following the manufacturer's instructions (RayBio Tech Life Inc., Peachtree Corners, GA, USA) (Table 1).

Identification and Characterization of the Major Immune Cell Populations and T Cell Subsets
Flow cytometry data were analyzed with the Infinicyt software (version 2.0.5;Cytognos SL, Salamanca, Spain).Supplementary Figure S1 describes the gate strategy used to identify the major immune cell populations (lymphocytes, monocytes, neutrophils, and eosinophils) and the distinct T cell subsets.Immune cell analysis was performed after excluding debris and cell doublets, considering their forward scatter (FSC)-A and FSC-H properties.Neutrophils were identified based on their FSC and side scatter (SSC) light dispersion characteristics, while monocytes were also identified based on CD4 expression (Figure 2A).T and B cells were identified by their positivity for CD3 (Supplementary Figure S1B) and CD20 (Supplementary Figure S1C), respectively.Our mAb combination (Table 4) was designed to perform a deep characterization of T lymphocytes, allowing us to identify 44 different subpopulations.Thus, T lymphocytes were primarily divided into CD4 + (Th), CD8 + (Tc), and CD4 + CD8 + T cells (Supplementary Figure S1D).As CD4 − CD8 − T cells comprised both TCRαβ CD4 − CD8 − T cells and γδ T cells, our mAb combination did not allow for the distinction of these cell groups; therefore, they were not further analyzed.As demonstrated in Supplementary Figure S1, within each one of the three major T cell subpopulations (Th, Tc, and CD4 + CD8 + T cells), we identified the following T cell subsets: CD25 high CD127 low/− regulatory T (Treg) cells (Supplementary Figure S1E); CXCR5 + follicular T (Tf) cells (Supplementary Figure S1G); and, based on CCR5 (CD195) and CCR6 (CD196) expression (Supplementary Figure S1H), CCR5 + CCR6 − T cells (T1), CCR5 − CCR6 + T cells (T17), CCR5 + CCR6 + T cells (T1/17), and CCR5 − CCR6 − T cells.Then, Treg cells were subdivided into CXCR5 + regulatory T (Tfr) cells (Supplementary Figure S1F).According to the expression of CCR5 and CCR6, T1-like, T17-like, T1/17-like, and CCR5 − CCR6 − T cells were further identified within Treg, Tf, and Tfr cells.Finally, the percentage of early activated T cells (CD25 + , Supplementary Figure S1I) and the expression of the immune-checkpoint proteins PD-1 (CD279, Supplementary Figure S1J) and TIM-3 (CD366, Supplementary Figure S1K) were further identified within each one of the 44 T cell subsets identified in this study, whenever the number of events acquired was enough to enable this analysis.
Figure 2. White blood cells count (WBC) in peripheral blood (A) and percentage of neutrophils (B) and monocytes within peripheral blood (PB) leukocytes (C) in nOB and OB, as well as among OB participants stratified by obesity class and metabolic profile.Percentage of monocytes expressing CCR5 + within all the studied groups (D).nOB: healthy participants (without obesity); OB: participants with obesity; IS: insulin sensitive; IRn: insulin resistant and normoglycemic; Pre-T2D: pre-diabetes; T2D: type 2 diabetes.Statistical differences were considered when p < 0.05.A p < 0.05 vs. nOB; B p < 0.05 vs. Class II; C p < 0.05 vs. Class III; D p < 0.05 vs. IS; E p < 0.05 vs. IRn; F p < 0.05 vs. Pre-T2D.

Statistical Analysis
Data were presented using mean ± standard deviation or median and inter-quartile range (iqr) for all studied variables.The comparison between groups was performed using the Wilcoxon rank sum and Kruskal-Wallis test from package coin version 1.4.2[23,24], followed by Dunn's test from package rstatix version 0.6.0[25], as appropriate.Furthermore, for the participants studied both before and after bariatric surgery, paired-sample comparisons were performed using the Wilcoxon signed-rank test [26].Spearman's correlations were performed between two continuous variables, using the rcorr() function Figure 2. White blood cells count (WBC) in peripheral blood (A) and percentage of neutrophils (B) and monocytes within peripheral blood (PB) leukocytes (C) in nOB and OB, as well as among OB participants stratified by obesity class and metabolic profile.Percentage of monocytes expressing CCR5 + within all the studied groups (D).nOB: healthy participants (without obesity); OB: participants with obesity; IS: insulin sensitive; IRn: insulin resistant and normoglycemic; Pre-T2D: pre-diabetes; T2D: type 2 diabetes.Statistical differences were considered when p < 0.05.A p < 0.05 vs. nOB; B p < 0.05 vs. Class II; C p < 0.05 vs. Class III; D p < 0.05 vs. IS; E p < 0.05 vs. IRn; F p < 0.05 vs. Pre-T2D.

Statistical Analysis
Data were presented using mean ± standard deviation or median and inter-quartile range (iqr) for all studied variables.The comparison between groups was performed using the Wilcoxon rank sum and Kruskal-Wallis test from package coin version 1.4.2[23,24], followed by Dunn's test from package rstatix version 0.6.0[25], as appropriate.Furthermore, for the participants studied both before and after bariatric surgery, paired-sample comparisons were performed using the Wilcoxon signed-rank test [26].Spearman's correlations were performed between two continuous variables, using the rcorr() function from package Hmisc version 4.4-2 [27].Statistically significant differences were considered when p < 0.05.The plots were performed using the package ggplot2 [28].All statistical analysis was performed using R (version 4.0.2;R Foundation for Statistical Computing, Vienna, Austria) [26,29].

Obesity Induces Alterations in Peripheral Blood Neutrophils and Monocytes
People with obesity (OB) presented higher white blood cell (WBC) counts (7.6 cells/µL ± 2.2) in comparison to the nOB group (6.2 × 10 3 cells/µL ± 1.5, p < 0.05; Figure 2A and Supplementary Table S2).Additionally, the stratification of BMI into obesity classes indicated that Class II obesity (6.0 × 10 3 cells/µL ± 2.1) displayed a similar WBC count to the nOB group.On the other hand, Classes III (7.9 x10 3 cells/µL ± 1.7, p < 0.05 vs. nOB) and IV (8.2 × 10 3 cells/µL ± 2.4, IV vs. II and IV vs. nOB, p < 0.05) were significantly increased (Figure 2A and Supplementary Table S2).Furthermore, the BMI was positively correlated to WBC (rho = 0.39, p < 0.05), as shown in Figure 3.A 1.4-fold increase in the percentage of neutrophils in the OB group compared to nOB (p < 0.05, Figure 2B and Supplementary Table S2) was also detected, which translated into an increase of 1.7× the absolute number of neutrophils (Supplementary Table S3).Additionally, the BMI and neck circumference were highly correlated to the percentage of neutrophils (rho = 0.62 and rho = 0.58, respectively, p < 0.05, Figure 3).Monocytes from the OB group showed a higher percentage of CCR5 + cells (Figure 2D and Supplementary Table S2) and a positive correlation between the percentage of monocytes and the HbA1c % for the OB group (rho = 0.36, p < 0.05; Figure 3).Considering the metabolic profile, the landscape was quite different.The IS group displayed a similar WBC count as the nOB group (nOB: 6.2 cells/µL ± 1.5; IS: 6.1 cells/µL ± 1.8; Figure 2A and Supplementary Table S2), while the non-IS OB groups showed higher WBC counts, with Pre-T2D (Pre: 8.8 cells/µL ± 3.0) displaying a significant difference vs. the nOB group (Figure 2A and Supplementary Table S2).

Obesity Induces Alterations in Peripheral Blood Neutrophils and Monocytes
People with obesity (OB) presented higher white blood cell (WBC) counts (7.6 cells/µL ± 2.2) in comparison to the nOB group (6.2 × 10 3 cells/µL ± 1.5, p < 0.05; Figure 2A and Supplementary Table S2).Additionally, the stratification of BMI into obesity classes indicated that Class II obesity (6.0 × 10 3 cells/µL ± 2.1) displayed a similar WBC count to the nOB group.On the other hand, Classes III (7.9 x10 3 cells/µL ± 1.7, p < 0.05 vs. nOB) and IV (8.2 × 10 3 cells/µL ± 2.4, IV vs. II and IV vs. nOB, p < 0.05) were significantly increased (Figure 2A and Supplementary Table S2).Furthermore, the BMI was positively correlated to WBC (rho = 0.39, p < 0.05), as shown in Figure 3.A 1.4-fold increase in the percentage of neutrophils in the OB group compared to nOB (p < 0.05, Figure 2B and Supplementary Table S2) was also detected, which translated into an increase of 1.7× the absolute number of neutrophils (Supplementary Table S3).Additionally, the BMI and neck circumference were highly correlated to the percentage of neutrophils (rho = 0.62 and rho = 0.58, respectively, p < 0.05, Figure 3).Monocytes from the OB group showed a higher percentage of CCR5 + cells (Figure 2D and Supplementary Table S2) and a positive correlation between the percentage of monocytes and the HbA1c % for the OB group (rho = 0.36, p < 0.05; Figure 3).Considering the metabolic profile, the landscape was quite different.The IS group displayed a similar WBC count as the nOB group (nOB: 6.2 cells/µL ± 1.5; IS: 6.1 cells/µL ± 1.8; Figure 2A and Supplementary Table S2), while the non-IS OB groups showed higher WBC counts, with Pre-T2D (Pre: 8.8 cells/µL ± 3.0) displaying a significant difference vs. the nOB group (Figure 2A and Supplementary Table S2).

Obesity Alters the T Cell Percentage and Their Subsets, but Not That of B Cells, in Peripheral Blood
The percentage of lymphocytes in whole blood was decreased by 1.5-fold in the OB group when compared to nOB (p < 0.05; Figure 4A and Supplementary Table S2) and was negatively correlated to BMI and neck circumference (rho = −0.53 and rho = −0.44,respectively, p < 0.05), as displayed in Figure 5. Similarly, the percentage of T lymphocytes in whole blood was also decreased in the OB group (19% ± 6.8 vs. nOB: 26% ± 7.8, p < 0.05, Figure 4B and Supplementary Table S2) and negatively correlated to BMI and neck circumference (rho = −0.34 and rho = −0.46,respectively, p < 0.05; Figure 5).Importantly, the levels of plasma glucose showed a weak negative correlation to the percentage of lymphocytes and T lymphocytes (rho = −0.35 and rho = −0.31,respectively, p < 0.05; Figure 5).B lymphocytes did not show any differences among the studied groups (Figure 4C and Supplementary Table S2).

CD4 + T Lymphocytes
The phenotype of the CD4 + T cells, as well as the percentage of activated T cells (identified as CD25 + ), and the expression of immune regulatory molecules were summarized in Figures 6-9, and more details are provided in Supplementary Tables S4, S5, and S7.A significant increase in the percentage of CD4 + T cells in the OB group (65% ± 12.0) vs. nOB (54% ± 7.1, p < 0.05; Figure 4D and Supplementary Table S2) was observed, while the percentage of CD8 + T cells was significantly reduced (nOB: 38% ± 8.6 vs. OB: 30% ± 11.1, p < 0.05; Figure 4E and Supplementary Table S2).No differences were detected on CD4 + CD8 + T cells (Figure 4F and Supplementary Table S2).BMI showed a moderate positive correlation with the percentage of CD4 + T cells (rho = 0.30, p < 0.05) (Figure 5).
No differences were found in the percentage of Th17-like Treg between OB and nOB groups.However, the IRn presented a reduction in the percentage of Th17-like Tregs (16% ± 8.2) when compared to the nOB group (25% ± 10.1, p < 0.05), as shown in Figure 8D and Supplementary Table S5.
Interestingly, when looking at the percentages of PD-1 + Th17-like Tregs, these were reduced in Pre-T2D in comparison to the nOB and the remaining OB groups (Figure 8J and Supplementary Table S5).On the other hand, the IRn showed a higher expression of PD-1 by Th17-like Treg as compared to nOB, Pre-T2D, and T2D (p < 0.05).The percentage of TIM-3 + Th17-like Treg was increased in the Class II and Pre-T2D groups, when compared to nOB (p > 0.05), and it was decreased in the IS group (p > 0.05), as seen in Figure 8L and Supplementary Table S5.Of note, the IS group showed a similar distribution in the Treg phenotypes as the nOB group, except in the percentage of CCR5 − CCR6 − Th-like Treg, which was significantly reduced (nOB: 18% ± 5.6 vs. IS: 10% ± 4.5, p < 0.05; Supplementary Table S5).
Biomolecules 2024, 14, x FOR PEER REVIEW 16 of 31 Interestingly, when looking at the percentages of PD-1 + Th17-like Tregs, these were reduced in Pre-T2D in comparison to the nOB and the remaining OB groups (Figure 8J and Supplementary Table S5).On the other hand, the IRn showed a higher expression of PD-1 by Th17-like Treg as compared to nOB, Pre-T2D, and T2D (p < 0.05).The percentage of TIM-3 + Th17-like Treg was increased in the Class II and Pre-T2D groups, when compared to nOB (p > 0.05), and it was decreased in the IS group (p > 0.05), as seen in Figure 8L and Supplementary Table S5.Of note, the IS group showed a similar distribution in the Treg phenotypes as the nOB group, except in the percentage of CCR5 − CCR6 − Th-like Treg, which was significantly reduced (nOB: 18% ± 5.6 vs. IS: 10% ± 4.5, p < 0.05; Supplementary Table S5).

CD8 + T Lymphocytes
The phenotypic features of CD8 + T cells, the percentage of activated cells (CD25 + ), and the expression of immune regulatory proteins were summarized in Figures 11-13, and more details are provided in Supplementary Tables S7 and S8.

Tc1 Cells Display Higher Expression of CCR5 in the OB Group and It Was Correlated to Insulin Levels
In similarity to the Th1 cells, there were no differences in the percentages of Tc1 among groups (Figure 11A and Supplementary Table S7).However, OB participants presented higher expression of CCR5 (measured as MFI) in Tc1 cells (3703 ± 1853), compared to nOB (2646 ± 1091, p < 0.05), particularly in the IS group (4164 ± 1671, p < 0.05) and T2D (3884 ± 2039, p < 0.05), as seen in Supplementary Table S7.Additionally, the percentage of PD-1 + Tc1 cells was decreased in the IS (27% ± 11.4), while the remaining groups showed similar percentages to the nOB (41% ± 17.3, p > 0.05; Figure 11E and Supplementary Table S7).Interestingly, the percentage of PD-1 + Tc1 cells was highly correlated to insulin levels and HOMA-IR (rho= 0.50 and rho = 0.51, p < 0.05, respectively; Figure 12) for the nOB and nT2D groups.

CD8 + T Lymphocytes
The phenotypic features of CD8 + T cells, the percentage of activated cells (CD25 + ), and the expression of immune regulatory proteins were summarized in Figures 11-13, and more details are provided in Supplementary Tables S7 and S8.

Tc1 Cells Display Higher Expression of CCR5 in the OB Group and It Was Correlated to Insulin Levels
In similarity to the Th1 cells, there were no differences in the percentages of Tc1 among groups (Figure 11A and Supplementary Table S7).However, OB participants presented higher expression of CCR5 (measured as MFI) in Tc1 cells (3703 ± 1853), compared to nOB (2646 ± 1091, p < 0.05), particularly in the IS group (4164 ± 1671, p < 0.05) and T2D (3884 ± 2039, p < 0.05), as seen in Supplementary Table S7.Additionally, the percentage of PD-1 + Tc1 cells was decreased in the IS (27% ± 11.4), while the remaining groups showed similar percentages to the nOB (41% ± 17.3, p > 0.05; Figure 11E and Supplementary Table S7).Interestingly, the percentage of PD-1 + Tc1 cells was highly correlated to insulin levels and HOMA-IR (rho= 0.50 and rho = 0.51, p < 0.05, respectively; Figure 12) for the nOB and nT2D groups.

CD8 + Regulatory T Cells Are Expanded in Participants with Obesity
A deep characterization of CD8 + Treg cells has been described in detail in another study [30].Briefly, an increase in the percentage of CD8 + Treg cells in the OB group (nOB: 0.14% ± 0.18 vs. OB: 0.27% ± 0.24, p < 0.05) was observed.This was more pronounced in pre-diabetes (Pre-T2D: 0.42% ± 0.31, p < 0.05 vs. nOB).Additionally, positive correlations between the percentage of circulating CD8 + Treg cells and fasting insulin (p < 0.05) or CRP (p < 0.05) levels [30] were found.Moreover, the CD8 + Treg phenotype from OB individuals were more prone to differentiate into Tc1-and Tc1/17-like cells and displayed increased expression of CCR5 [30].

Follicular Cytotoxic T Cells Present an Increased Activation Status in the OB Group
A slight increase in the percentage of Tfc was detected in the OB (2.6% ± 1.8 vs. nOB: 1.9% ± 0.88, p > 0.05), particularly in the T2D (3.1% ± 2.0, p = 0.054 vs. nOB) (Figure 13A and Supplementary Table S8).In similarity to the Tfh, the Tfc also display higher expression of CXCR5 in OB (p < 0.05; Supplementary Table S8), particularly in Class II obesity, as well as a higher percentage of activated Tfc cells (OB: 10% ± 5.3 vs. nOB: 6.6% ± 4.0, p < 0.05; Figure 13B and Supplementary Table S8).
Supplementary Table S9).Interestingly, the IRn and Pre-T2D also displayed a slight reduction in the percentage of the PD-1 + CD4 + CD8 + Tf cells, compared to the nOB group (p > 0.05; Supplementary Table S9).

Differences within Metabolic Groups
Participants with obesity and insulin resistance, not taking medication for T2D, showed a higher percentage of CD4 + CD8 + T cells within T cells (IRn: 2.8% ± 1.1; Pre-T2D: 2.2% ± 2.0) compared to T2D patients under treatment (1.6% ± 1.9, p < 0.05 for IRn vs. T2D; Figure 5 and Supplementary Table S2).Furthermore, insulin-resistant groups (including T2D) displayed a higher polarization pattern toward the Th1-like, namely in CD4 + Treg cells (Figure 8 and Supplementary Table S5) and CD4 + Tfr cells (Figure 9 and Supplementary Table S5), when compared to the IS group.Importantly, groups with higher percentages of HbA1c displayed higher percentages of CD4 + CD8 + Tf cells, while the polarization of these cells presented different patterns among the groups, wherein Pre-T2D displayed the highest percentage of CD4 + CD8 + Tf toward T17-like (Supplementary Table S9).Additionally, the Pre-T2D also displayed the highest percentage of CD8 + Treg cells among groups [30].

Differences within Metabolic Groups
Participants with obesity and insulin resistance, not taking medication for T2D, showed a higher percentage of CD4 + CD8 + T cells within T cells (IRn: 2.8% ± 1.1; Pre-T2D: 2.2% ± 2.0) compared to T2D patients under treatment (1.6% ± 1.9, p < 0.05 for IRn vs. T2D; Figure 5 and Supplementary Table S2).Furthermore, insulin-resistant groups (including T2D) displayed a higher polarization pattern toward the Th1-like, namely in CD4 + Treg cells (Figure 8 and Supplementary Table S5) and CD4 + Tfr cells (Figure 9 and Supplementary Table S5), when compared to the IS group.Importantly, groups with higher percentages of HbA1c displayed higher percentages of CD4 + CD8 + Tf cells, while the polarization of these cells presented different patterns among the groups, wherein Pre-T2D displayed the highest percentage of CD4 + CD8 + Tf toward T17-like (Supplementary Table S9).Additionally, the Pre-T2D also displayed the highest percentage of CD8 + Treg cells among groups [30].
Interestingly, the IS group showed an overall reduction in the percentage of cells expressing PD-1, as well as in the amount of PD-1 protein expressed per cell (measured as MFI), while a sequential increase was observed with increasing IR (Supplementary Tables S4-S9).A similar pattern was observed regarding the percentage of cells expressing TIM-3.These variations were especially observed in Th1 (Figure 6H and Supplementary Table S4) and Tc1 (Supplementary Table S7), as well as CD4 + Treg cells regarding TIM-3 (Figure 9G and Supplementary Table S5).On the other hand, the IS group had a higher percentage of PD-1 + Th1-like Treg cells (60% ± 8.7), while the Pre-T2D had the lowest percentage (41% ± 8.1, p < 0.05; Figure 8H and Supplementary Table S5) of these cells.As previously observed, the Pre-T2D group had important alterations in the CD8 + Treg cells, regarding their phenotype and the expression of immune checkpoint molecules [30].

Effect of Bariatric Surgery on Metabolic and Immune Profiles
To understand the effect of bariatric surgery on the obesity-associated immune profile, an unpaired analysis was performed, comparing 15 individuals with Class IV obesity before surgery (at baseline or T1) and a set of 14 participants, evaluated 7 to 18 months post-surgery (T2).Additionally, a follow-up study with paired comparisons including eight individuals with Class IV obesity, analyzed both at T1 and T2, was also performed (Figure 1).In this follow-up study, the time period between T1 and T2 ranged from 9 to 18 months (Figure 16 and Supplementary Table S10).Interestingly, the IS group showed an overall reduction in the percentage of cells expressing PD-1, as well as in the amount of PD-1 protein expressed per cell (measured as MFI), while a sequential increase was observed with increasing IR (Supplementary Tables S4-S9).A similar pattern was observed regarding the percentage of cells expressing TIM-3.These variations were especially observed in Th1 (Figure 6H and Supplementary Table S4) and Tc1 (Supplementary Table S7), as well as CD4 + Treg cells regarding TIM-3 (Figure 9G and Supplementary Table S5).On the other hand, the IS group had a higher percentage of PD-1 + Th1-like Treg cells (60% ± 8.7), while the Pre-T2D had the lowest percentage (41% ± 8.1, p < 0.05; Figure 8H and Supplementary Table S5) of these cells.As previously observed, the Pre-T2D group had important alterations in the CD8 + Treg cells, regarding their phenotype and the expression of immune checkpoint molecules [30].

Effect of Bariatric Surgery on Metabolic and Immune Profiles
To understand the effect of bariatric surgery on the obesity-associated immune profile, an unpaired analysis was performed, comparing 15 individuals with Class IV obesity before surgery (at baseline or T1) and a set of 14 participants, evaluated 7 to 18 months post-surgery (T2).Additionally, a follow-up study with paired comparisons including eight individuals with Class IV obesity, analyzed both at T1 and T2, was also performed (Figure 1).In this follow-up study, the time period between T1 and T2 ranged from 9 to 18 months (Figure 16 and Supplementary Table S10).

Discussion
Obesity is a potential trigger for different comorbidities, with chronic low-grade inflammation perpetuating and underlying these conditions [1,2].The immune system is paramount in stimulating and maintaining inflammatory processes.It plays an important role in the onset of obesity-related comorbidities [2].In addition, the interaction of immune cells and the extracellular matrix (ECM) also plays an important role in the onset of lowgrade inflammation in obesity.This is driven by the release of different molecules, such as small fragments of hyaluronan, which induce the activation of innate immune cells via toll-like receptors.Moreover, the involvement of these molecules in the development of insulin resistance has previously been suggested [31].
The present study brings new light in regard to the phenotype and function of circulating immune cells in people with obesity, with obesity-associated insulin resistance and T2D.The study aimed to obtain a deep characterization of circulating T cells.The impact of obesity on the imbalance of immune homeostasis is evident in both innate and adaptive immune systems.Higher white blood cell counts were observed in people with obesity, except for those in the insulin-sensitive group who displayed similar WBC counts as the nOB.Increased percentages of neutrophils and a reduction in T lymphocytes, with consequent reduction in total lymphocytes, in people with obesity were also observed.In previous studies, the important role of neutrophils was reflected in the onset of the inflammatory cascade in obesity [32] and in the onset of insulin resistance by the release of elastase, promoting the degradation of insulin receptors [7,33].Moreover, different authors indicated the impact of obesity on the adaptive immune cells, with emphasis on T cells, either in circulation or infiltrated in adipose tissue [11].In our cohort, increases in CD4 + T cells, accompanied by reductions in CD8 + T cells, were observed within circulating immune cells.
In 2015, Łuczy ński and collaborators found an increased count of circulating CD4 + T cells with the Th17 phenotype in children with obesity (10-18 years old) and suggested these cells played a key role in low-grade inflammation and established the link between obesity and diabetes [34].In addition, they also observed a higher percentage of Th17 cells expressing IFN-γ [34] that have been further identified as Th1/17 cells [35].Importantly, our data showed that both populations were increased in the peripheral blood of people with obesity, accompanied by a reduction in CCR5 − CCR6 − Th cells.This profile was observed independently of their metabolic status.However, despite no difference being observed between the Th1 cell percentages, the IS and IRn participants presented lower expression of PD-1 by Th1 cells, as compared to T2D participants.The reduction in the expression of PD-1 could be indicative of an increased function of these cells in those participants.According to previous studies [reviewed [36]], the expression of PD-1 by Th cells was related to loss of function [36] and exhaustion [37].In fact, studies indicated that the blockade of PD-1, as well as the blockade of TIM-3, induced an increase in IFN-γ production by PD-1 + CD4 + T cells [36].On the other hand, CD8 + T and CD4 + CD8 + T cells from the OB group showed a different polarization tendency, with preferential Tc1-and T1-like polarization, respectively.Noteworthy, the Tc1 cells from individuals with obesity showed higher expression of CCR5, allowing these cells to migrate into inflammatory sites [38].
Some controversy remains regarding the circulating CD4 + Treg cells in people with obesity.Some studies indicate that CD4 + Treg cells are reduced in circulation in people with obesity [14], while others do not [39].In the present study, no differences were identified in the percentage of CD4 + Treg cells within CD4 + T cells, although a significant reduction was observed in the percentage of CD4 + Treg cells measured within whole blood.It is important to mention that this reduction can be attributed to the increased white blood cell count in people with obesity, rather than to an actual reduction in the number of CD4 + Tregs in the PB.Therefore, we evaluated the absolute number of CD4 + Tregs in whole blood and found that the absolute count was not significantly different between the nOB and the OB groups, which displayed similar counts (Supplementary Table S3).Interestingly, CD8 + Treg cells follow the opposite direction, displaying an increase in their percentage in people with obesity-more marked in people with prediabetes [30].It is noteworthy that this cell population was previously described as being positively correlated to BMI [40].The expression of chemokine receptors by Treg cells allows them to migrate into inflammatory locations [38].As previously reported, CD4 + T cells infiltrated into adipose tissue display a higher tendency to polarize toward Th1 cells [13] (and our unpublished data).It is important to acknowledge that the increase in the percentage of circulating Th1-and Th1/17-like Treg cells in people with obesity described in our study may indicate active migration of these Treg cells into adipose tissue as an attempt to mitigate the inflammation in the tissue microenvironment.Additionally, this increase was accompanied by a worsened metabolic condition.Furthermore, regarding the expression of the immune checkpoint proteins, PD-1 has commonly been associated with Treg homeostasis and immunosuppressive capability [reviewed [41]], as well as TIM-3 [42].In fact, people with obesity and pre-diabetes showed a reduced percentage of PD-1 + Th1like Treg in circulation but also a reduction in PD-1 + CD8 + Treg cells and TIM-3 + CD8 + Treg cells [30].This fact supports a previous study reporting the impaired immune suppressive function of Tregs during insulin resistance and T2D [43].
Follicular T cells play an important role in the maintenance and differentiation of B cells [44].Guo and collaborators previously reported that people with T2D display higher percentages of Tfh cells in circulation, in agreement with our study.It is important to mention that the research conducted by Guo and collaborators (2020) was performed in people without obesity (BMI of 25.90 ±.99) [44].Regardless of obesity, T2D can induce alterations in the percentage of Tfh cells.Interestingly, Tfc displayed a similar tendency, despite the expression of CXCR5 being reduced in T2D compared to the IS group.
People with Pre-T2D and T2D displayed a higher expression of TIM-3 in some CD4 + T and CD8 + T cell subsets, compared to the remaining OB groups.However, Sun and collaborators (2020) reported a reduction in the expression of TIM-3 within CD4 + and CD8 + T cells, when comparing people with obesity and people with obesity and T2D [45].In fact, the authors attributed the difference in the expression of TIM-3 to a transient upregulation of plasma glucose levels during the early stages of T2D, with a restoration of the T cells function in long-term T2D.Moreover, an increase in TIM-3 has also been previously described in Tfh cells [44].
Importantly, bariatric surgery has been used as a gold standard in the treatment of obesity and, consequently, obesity-related comorbidities [15].The impact of this procedure has been well described regarding its positive effects on metabolic profiles.However, the effects of bariatric surgery have not been well studied regarding immunometabolism and the modulation of the immune system.Here, we present a deep characterization of T cells before and post-surgery.Our results show that bariatric surgery has important modulatory effects on the circulatory immune profile of people with obesity.Wijngaarden et al. (2022) postulated a recovery of the CD4 + T cell compartment in patients three months after bariatric surgery toward the lean control phenotype and found no changes in CD8 + T cells [15].Our study analyzes the circulating immune profile of people with obesity, 7 to 18 months post-bariatric surgery.A higher percentage of CD4 + Tregs and B cells within whole blood was observed, with an increased activation profile, also accompanied by an expansion of Tfh.
A large heterogeneity found among participants and the medication used to treat/ mitigate the obesity comorbidities may interfere directly or indirectly with the immune system.The most prescribed anti-diabetic, metformin, greatly impacts immune modulation, as reviewed in [46].Furthermore, statins also play an important role in immune regulation by inhibiting T cell proliferation [47].
Despite these confounding factors, we demonstrated important new alterations on several T cell subsets associated with obesity.How distinct immune profiles impact the development of comorbidities and the obesity itself emerges as an interesting research area.Similarly, it is important to understand whether the outcome of bariatric surgery differs for patients with different immune profiles.Finally, our study demonstrated that obesity modulates differently distinct T cell subsets; therefore, it is mandatory to accurately identify these T cell subsets to understand which ones are altered.Otherwise, some differences observed in obesity would go unnoticed.

Conclusions
We describe significant and novel changes in circulating T cells associated with obesity.Remarkably, CD4 + and CD8 + T cells exhibited distinct alterations, even though both subsets exhibited increased CCR5 expression, potentially serving as a common link between the two populations.Notably, T cells from individuals with obesity displayed a pro-inflammatory profile and presented changes in the expression of immune regulatory molecules, which may serve as a counterbalance to mitigate the inflammation exacerbation.The precise impact of the immune alterations reported in this study on the development of insulin resistance and type 2 diabetes remains uncertain.The effects of bariatric surgery on the immune system were observed around 12 months after surgery, but it is mandatory to

Figure 1 .
Figure 1.Distribution of the participants according to the time-point of sample collection.Thirtyfive individuals were studied before bariatric surgery (T1).Eight of them, with BMI ≥50 kg/m 2 at T1, were also studied 9 to 18 months after the surgery (T2).Six individuals were only studied at T2.N/A: not analyzed.

Figure 1 .
Figure 1.Distribution of the participants according to the time-point of sample collection.Thirty-five individuals were studied before bariatric surgery (T1).Eight of them, with BMI ≥50 kg/m 2 at T1, were also studied 9 to 18 months after the surgery (T2).Six individuals were only studied at T2.N/A: not analyzed.

Figure 3 .
Figure 3. Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index; WBC: white blood cells.Statistical differences were considered when p < 0.05.

Figure 3 .
Figure 3. Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index; WBC: white blood cells.Statistical differences were considered when p < 0.05.

Figure 5 .
Figure 5. Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index; PB: peripheral blood.Statistical differences were considered when p < 0.05.

Figure 7 .
Figure 7. Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index; MFI: mean fluorescence intensity.Statistical differences were considered when p < 0.05.

3. 3 . 2 .
The OB Group Displays an Increased Percentage of Th17 and Th1/17 Cells That Is Associated with a Reduced Expression of TIM-3

3. 3 . 2 .
The OB Group Displays an Increased Percentage of Th17 and Th1/17 Cells That Is Associated with a Reduced Expression of TIM-3

Figure 12 .
Figure 12.Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index; MFI: mean fluorescence intensity.Statistical differences were considered when p < 0.05.

Figure 12 .
Figure 12.Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index; MFI: mean fluorescence intensity.Statistical differences were considered when p < 0.05.

Figure 15 .
Figure 15.Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index.Statistical differences were considered when p < 0.05.

Figure 15 .
Figure 15.Spearman's correlation analysis between immune and anthropometric or metabolic parameters.BMI: body mass index.Statistical differences were considered when p < 0.05.

Figure 16 .Figure 16 .
Figure 16.Phenotypical changes in T cells from individuals with obesity, before (T1) and after (T2) bariatric surgery.The analysis of unpaired samples comprised all individuals with Class IV obesity in T1 (n = 15) and 14 individuals after bariatric surgery (T2).The analysis of paired samples consisted in a follow-up of 8 Figure 16.Phenotypical changes in T cells from individuals with obesity, before (T1) and after (T2) bariatric surgery.The analysis of unpaired samples comprised all individuals with Class IV obesity in T1 (n = 15) and 14 individuals after bariatric surgery (T2).The analysis of paired samples consisted in a follow-up of 8 individuals before (T1) and after (T2) bariatric surgery.The nOB group was used as reference.Distribution of white blood cells count (WBC) (cells/µL) before (T1)

Table 2 .
Criteria used to classify people with obesity according to their metabolic profiles.

Table 3
shows how participants from the different metabolic groups are distributed into obesity classes.

Table 3 .
Distribution of participants with obesity according to their metabolic group and obesity class.

Table 4 .
Monoclonal antibody combination used to identify the different T cell subsets, presenting the information on each respective fluorochrome, clone, and commercial source.