Impact of RYGB surgery on plasma immunoglobulins: association between blood pressure and glucose levels six months after surgery

We aimed to study levels of natural antibodies in plasma, and their associations to clinical and fecal biomarkers, before and 6 months after Roux‐en‐Y gastric bypass (RYGB) surgery. Thirty individuals with obesity [16 type 2 diabetic, 14 non‐diabetic (ND)] had RYGB surgery. Total plasma IgA, IgG and IgM antibody levels and specific antibodies to oxidized low‐density lipoprotein (oxLDL), malondialdehyde‐acetaldehyde adducts, Porphyromonas gingivalis gingipain A hemagglutinin domain (Rgp44), and phosphocholine were measured using chemiluminescence immunoassay. Associations between plasma and fecal antibodies as well as clinical markers were analyzed. RYGB surgery reduced blood pressure, and the glycemic state was improved. A higher level of diastolic blood pressure was associated with lower plasma antibodies to oxLDL after surgery. Also, lower level of glucose markers associated with lower level of plasma antibodies to bacterial virulence factors. Antibodies to oxLDL decreased after surgery, and positive association between active serum lipopolysaccharide and specific oxLDL antibodies was detected. Total IgG levels decreased after surgery, but only in ND individuals. Reduced level of total plasma IgG, improved state of hypertension and hyperglycemia and their associations with decreased levels of specific antibodies in plasma, suggest an improved state of systemic inflammation after RYGB surgery.


INTRODUCTION
Obesity is a worldwide health problem with many comorbidities, such as type 2 diabetes (T2D), atherosclerotic cardiovascular disease, dyslipidemia, hypertension, and chronic low-grade inflammation.
All these are associated with increased oxidative stress, which along with all comorbidities, affects the immune system in a negative way [1][2][3].
Roux-en-Y gastric bypass (RYGB) surgery is a common bariatric procedure [4].It has been shown that surgery is a more effective way to lose weight, improve metabolic health and gain higher remission rate of diabetes than non-surgical treatments [3].
In recent years, the impact of RYGB surgery on hypertension and glycemic control has been in focus of the research.Studies report significant decrease in blood pressure after RYGB surgery [5][6][7] and immediate improvement of glycemic state [8].
Natural antibodies are of importance in innate immune responses [9].Most of natural antibodies are IgM isotype, and they are the first ones to be produced during immune development [9,10].IgA and IgG isotypes of natural antibodies have also been well-described [11], and recently isotype IgE of natural antibodies also has been successfully isolated [12].These natural antibodies are of importance, because they easily bind to invading pathogens, and they also recognize altered-self material and well-conserved epitopes [9].
In our earlier paper [17], we investigated the impact of RYGB surgery on natural antibodies in the gut in individuals with or without T2D.We documented that total fecal IgA levels were elevated, a possibly beneficial effect, while fecal IgM specific to oxLDL, PCho and Rgp44 decreased after surgery.We have also previously investigated the impact of RYGB gastrointestinal inflammation in individuals with and without T2D [18].That study documented altered intestinal inflammation homeostasis after RYGB surgery, with several fecal markers such as calprotectin and lipopolysaccharide (LPS) suggesting increased inflammation, possibly linked to the risk for intestinal bowel disease (IBD) and colorectal cancer [19][20][21][22][23].
The results from our earlier paper [17] prompted us to investigate how RYGB surgery affects the levels of natural antibodies in plasma.We wanted to investigate if plasma antibodies were associated with clinical markers, compare the results to fecal findings [17], and explore the effects of RYGB surgery on immunoglobulins and specific plasma antibodies during an oral liquid meal test.

Study subjects
Thirty individuals were recruited to the study, during the years 2010-2014 at Oulu University Hospital, Finland.These individuals with obesity (BMI >30 kg/m 2 ) and a medical indication for bariatric surgery, were divided into two groups: (i) T2D individuals without insulin treatment (N = 16) and (ii) non-diabetic (ND) individuals (N = 14).As comparison, six healthy individuals with BMI <30 kg/m 2 where recruited.Plasma and fecal samples were collected before and approximately 6 months after RYGB surgery.A more detailed description of the study protocol, study subjects and their criteria has been published earlier [18].
The study followed the ethical standards of Helsinki Declaration and was approved by the Ethics Committee of the Northern Ostrobothnia Hospital District.A written informed consent was obtained from each study participant.The trial was registered at ClinicalTrials.gov as NCT01330251.

Plasma sample collection and preparation for immunoassay
Plasma samples were collected before and approximately 6 months after the surgery.All study subjects participated in an oral two-hour liquid meal test at the baseline study visit, where blood samples were collected at four timepoints: 0, 30, 60, and 120 min.Healthy control individuals gave only one set of samples, no follow-up samples were taken.The meal test consisted of a 252 mL liquid mix, including 75 g carbohydrates, 17.4 g fat, 18 g protein and 528.4 kcal of energy.A more detailed description of sample collection and liquid meal test have been done earlier [18].

Chemiluminescence immunoassays and specific antigens
Chemiluminescence immunoassay was used to measure total plasma IgA, IgG and IgM immunoglobulin levels as described [24][25][26].Chemiluminescence immunoassay was also used to measure specific antibody binding to five different oxLDL and bacterial cross-reactive virulence factors [copper-oxidized low-density lipoprotein (CuOx-LDL), malondialdehyde acetaldehyde low-density lipoprotein (MAA-LDL), malondialdehyde acetaldehyde bovine serum albumin (MAA-BSA), PCho and Rgp44].The antigens and their sources are listed in detail in our previous paper [17].

Measurement of LDL aggregation susceptibility
LDL was isolated by sequential ultracentrifugation and the aggregation susceptibility of the isolated LDL particles defines as the aggregate size at 2 h after induction of LDL aggregation by sphingomyelinase as describes [27,28].

Statistics
All statistical analyses were carried out using IBM SPSS Statistics software 25.0 (IBM, Armonk, NY, USA).Wilcoxon signed rank test was used for statistical analyses.
Pearson's correlation test was used for correlations.p-Values <0.05 were considered statistically significant.

Clinical characteristics
The clinical characteristics of the study subjects discussed in this study are summarized in Table 1.Characteristics of healthy individuals are presented in Table 2.More detailed presentation of clinical characteristics has been discussed earlier [18].
The detailed clinical characteristics of study subjects and the impact of RYGB on clinical variables are discussed in our previous publications [17,18].In general, glucose and lipid metabolism improved, and diastolic blood pressure and weight were lowered after RYGB surgery.Serum LPS activity was decreased after surgery, but the surgery did not induce any significant differences in serum LDL aggregation susceptibility.Of fecal biomarkers, calprotectin increased significantly after surgery, but no significant association between fecal calprotectin and plasma immunoglobulins were detected.
Plasma total IgG levels decreased significantly after surgery, but only in ND study subjects (p = 0.004).No other significant changes in plasma total immunoglobulin levels could be detected.
Plasma antibody levels to specific epitopes decreased after surgery.Significant results were seen in antibodies to CuOx-LDL and virulence factors PCho and Rgp44.Detailed results are listed in Table 1.

Plasma immunoglobulin levels during meal test before and after surgery
Total immunoglobulin levels were measured during the meal test before and after surgery, but no statistically significant changes in their levels were detected.
Specific plasma antibody levels to CuOx-LDL and MAA-LDL at different timepoints are presented in Fig. 1.Statistically significant changes were observed in IgA, IgG, and IgM to CuOx-LDL (Fig. 1A-C) at different timepoints when analyzing all study subjects together.Significant differences in IgA to MAA-LDL or IgG to MAA-LDL (Fig. 1D,  E) were not observed.However, IgM to MAA-LDL (Fig. 1F) showed statistically significant changes at 30, 60, and 120 min after the surgery, when analyzing the whole study group together.IgA showed statistically significant change in IgA to MAA-BSA only at 30 min (p = 0.010) (data not shown).
Plasma antibody levels to Rgp44 and PCho at different timepoints are shown in Fig. 2. IgA to Rgp44 and IgM to Rgp44 decreased significantly at some timepoints after the surgery (Fig. 2A,C).IgA to Rgp44 was statistically significant at 0 and 120 min.Also, IgM to Rgp44 showed significant results at 0 and 30 min.IgG to Rgp44 showed no significant changes (Fig. 2B).
IgG to PCho decreased significantly after the surgery at 30 min only (Fig. 2E).Also, IgM to PCho was significantly lower after surgery at the 0 min timepoint (Fig. 2F).IgA to PCho showed no significant results (Fig. 2D).

Plasma immunoglobulins associated with clinical markers
Results of associations between plasma immunoglobulins to clinical markers before and after surgery are presented in Table S1.After RYGB surgery, significant associations between specific plasma immunoglobulins and clinical markers were detected.Diastolic blood pressure had a negative association with plasma IgG to CuOx-LDL and with plasma IgG to MAA-LDL (Fig. 3).Lower diastolic blood pressure associated with a higher level of specific plasma antibodies to OxLDL after surgery.No significant association was seen before surgery.Markers of glucose metabolism had positive associations with plasma IgM to PCho (Fig. 4) after surgery.Higher levels of glycosylated hemoglobin (HbA1c) and fasting plasma glucose (fPGluk) associated with higher levels of plasma IgM to PCho.No significant associations between glucose markers and specific antibodies were seen before surgery.Serum LPS activity (sLPS) had a positive association with plasma IgA to CuOx-LDL and plasma IgM to CuOx-LDL (Fig. 5).Higher levels of sLPS associated with a higher amount of plasma antibody to oxLDL after surgery, but no significant association was seen before surgery.The measurements before and after surgery of individual clinical markers can be found in detail in Table 1.Specific immunoglobulin levels presented in Figs 3-5 represents mean values of results from all four timepoints, for each specific immunoglobulin.
Before surgery, plasma IgM to Rgp44 and IgA to MAA-LDL had negative correlations with aggregated LDL in serum (sLDLaggr) (Fig. S1).Larger sLDLaggr associated with a lower level of specific plasma antibody.The association could not be detected anymore after surgery.

The association of plasma immunoglobulins with the same types of fecal immunoglobulins before and after surgery
We also investigated the association of total and specific plasma immunoglobulins (IgA, IgG, IgM) levels with similar total and specific fecal immunoglobulin levels.No significant associations were detected between similar total plasma and total fecal immunoglobulins before or after RYGB surgery (data not shown).However, one statistically significant association between specific plasma IgG to PCho and specific fecal IgG to PCho was detected after surgery.Plasma and fecal IgG to PCho had a negative correlation as higher plasma IgG to PCho associated with lower levels of fecal IgG to PCho (r = À0.376,p = 0.045).

DISCUSSION
RYGB surgery improves the state of systemic inflammation and general health.In this study we discovered that the RYGB surgery causes a decrease in specific antibodies to oxidized epitopes and microbial virulence factors in plasma samples.We also found significant associations between clinical markers, such as diastolic blood pressure and fasting plasma glucose, and specific plasma antibodies before and after surgery.Finally, we compared results from this present study with previous findings in fecal samples [17] and found that RYGB affects total IgA antibody levels in fecal samples, while IgG antibody levels are affected in plasma samples.
In our earlier study [18], we reported a significant decrease in diastolic blood pressure after RYGB surgery.In this study, we detected inverse associations between diastolic blood pressure and plasma IgG to CuOx-LDL and IgG to MAA-LDL after the surgery.Lower blood pressure associated with higher levels of specific antibodies to oxLDL.No associations were seen before surgery.Earlier studies have reported that antibodies to oxLDL  associate with hypertension [29][30][31].Wu et al. [31] reported over 20 years ago, that individuals with borderline hypertension have a decreased number of antibodies against oxLDL, but the reason for this is unknown.They suggested that the reason for this could be an increased consumption of antibodies against oxLDL, a decreased production of them, or that antibodies could be trapped in atherosclerotic lesions.This is supported by our findings.Even if lower blood pressure is associated with higher amounts of antibodies against oxLDL after RYGB surgery, the general levels of antibodies against oxLDL are decreased after surgery.It can be suggested that these antibodies to oxLDL could have protective properties, but further studies are needed.
We, and others, have reported earlier that HbA1c and fasting plasma glucose levels decrease significantly after RYGB surgery [18,32].In this study, we found that these glucose markers had a positive association with plasma IgM to PCho.Autoantibodies have previously been reported to regulate blood glucose homeostasis and insulin levels in mice [33].Also, variable associations between serum antibody levels and hyperglycemia in people with periodontal damage have been reported [34].We have previously reported positive associations between specific plasma IgA to PCho and fasting blood glucose levels [35], but no previous links between glucose metabolism and plasma IgM to PCho has been reported.In our previous study [17], we reported that specific fecal immunoglobulins (IgA, IgG and IgM) to Rgp44 were linked with fasting plasma glucose levels before surgery.RYGB surgery affects many physiological functions [36], and it could be speculated that the altered gut microbiota could explain these differing results in fecal and plasma samples before and after surgery.
We also wanted to investigate if there were any associations between plasma immunoglobulins and serum inflammatory marker, LPS.Serum LPS levels were significantly reduced by RYGB.Furthermore, after RYGB surgery, higher sLPS levels associated with higher levels of plasma IgA to CuOx-LDL and IgM to CuOx-LDL.LPS is an endotoxin located on the outer membrane of gram-negative bacteria, where it binds to lipopolysaccharide-binding protein (LBP) causing an activation of inflammatory signaling and production of proinflammatory cytokines.Our earlier study [18] reported an increased amount of fecal LPS after RYGB surgery, a finding confirmed by a more recent study [37].Increased fecal LPS suggests increased fecal bacterial activity, leading to increased inflammatory state in the gut [18].Indeed, a recent study demonstrated a shift to proinflammatory bacteria in feces 1 year after RYGB [37].However, we have previously demonstrated an increase of intestinal IgA after RYGB, possibly a beneficial compensatory antibody response, improving the state of systemic low-grade inflammation [17,37].There are reports of positive associations between oxLDL and circulating LPS [38] and findings of this study support the hypothesis of an improved state of systemic inflammation after RYGB, since antibodies to oxLDL and amount of serum active LPS decreased significantly after surgery.
In this study, we detected a negative association between plasma IgM to Rgp44 and IgA to MAA-LDL and the aggregation susceptibility of LDL particles before RYGB surgery; lower levels of specific plasma immunoglobulins were associated with higher propensity of LDL particles to aggregate.However, no association was seen after surgery.The aggregation susceptibility of LDL depends largely on the phospholipid composition of the particles [28], which can be influenced also by the diet [28,39].No previous study has shown association between antibodies and the aggregationsusceptibility of LDL, but based on our results, the specific antibodies may have protective properties against modification and aggregation of LDL particles.LDL particles can aggregate in the atherosclerotic arterial wall after modification, for example by oxidation [40,41].Aggregation susceptibility of LDL particles is associated with future cardiovascular events in patients with established atherosclerosis [28,42].It could therefore be speculated, that these specific antibodies may have protective properties against future cardiovascular events.
After RYGB surgery, total plasma IgG levels decreased significantly in individuals without diabetes, while total plasma IgA or IgM levels were not affected.When analyzing all study subjects, no Fig. 3. Diastolic blood pressure (RRDias) had negative associations with plasma IgG to CuOx-LDL and plasma IgG to MAA-LDL levels after RYGB surgery, in the whole study population (N = 30).Patients with lower blood pressure had a higher level of specific plasma antibodies to oxLDL.Specific immunoglobulin levels were calculated as mean values of the results from the samples collected at four different timepoints during the liquid oral meal test, however associations with baseline RRDias were detected at every individual timepoint.Fig. 4. Glucose metabolism markers, glycosylated hemoglobin (HbA1c) and fasting plasma glucose (fPGluk), levels had positive associations with plasma IgM to PCho levels after the RYGB surgery, in the whole study population (N = 30).Higher levels of glucose markers correlated with a higher level of plasma IgM to PCho.Plasma IgM to PCho levels were calculated as mean values of the results from the samples collected at four different timepoints during the oral liquid meal test, however associations with fasting HbA1c and glucose were detected at every individual timepoint.Fig. 5. Serum LPS activity levels had positive associations to specific antibodies IgA and IgM to CuOx levels after RYGB surgery, in the whole study population (N = 30).Higher levels of active LPS correlated with a higher level of specific antibody to CuOx.Specific plasma antibody levels present mean values of the results from samples collected at four different timepoints during the oral liquid meal test.significant changes could be detected in total plasma IgA, IgG or IgM levels.In fecal samples [17], total IgA levels increased significantly after the surgery while no changes were detected in fecal total IgG and IgM levels.There are no previous publications of comparison between plasma and fecal antibody levels after RYGB surgery.Only a few studies have reported decreased IgG levels in plasma after RYGB surgery.A study reported that total IgG antibody decreased 6 months after RYGB surgery [43], in line with our finding.Another study reported a decrease in IgG, 5 and 10 months after RYGB [44].Oxidative stress is reduced after bariatric surgery, and it is reflected in improved oxidized lipoprotein levels and reduced number of IgG autoantibodies [45].Our results show a decrease in several immunoglobulin classes to oxLDL, and they have been documented to also associate with cardiovascular diseases [46][47][48].Based on our findings, it could be suggested that a decrease in plasma IgG levels indicates an improved state of systemic low-grade inflammation after RYGB surgery.
We compared the results from this study to the results from our previous study [17] with fecal samples.In feces, total fecal IgA levels increased, while total IgG levels decreased in plasma samples after the surgery.These divergent results may reflect the different roles of immunoglobulins in feces and plasma: IgG can only be found in low concentrations at mucosal surfaces and secretions, whereas IgA is the most abundant immunoglobulin in the gut [49].A recent study suggested that RYGB surgery shifts the intestinal microbiota towards a more pro-inflammatory direction, but that the systemic inflammation is improved [37].This was explained by an improved intestinal barrier function and a compensatory IgA response against proinflammatory compounds.This suggestion is aligned also with our results.

CONCLUSION
RYGB is an effective surgical treatment for sustained weight loss and improved state of metabolic comorbidities, such as dyslipidemia, glycemic control, and hypertension.Numerous earlier reports have shown that systemic low-grade inflammation and oxidative stress are reduced after RYGB surgery [2,44,50,51].However, previous studies by us and others have suggested that fecal inflammation increase after RYGB [17,18,37].This study shows that sLPS and the antibodies to oxidized LDL decrease after surgery, supporting the view that systemic inflammation is decreased.This is further supported by positive associations between plasma glucose levels and antibodies to oxidized epitopes.
N.H. wrote the first draft of the manuscript, and performed statistical analyses and part of the experimental work.J.H., M.J.S., V.K., and S.H. contributed to the conception of the idea, design of the study, and critically evaluated the manuscript.R.A., A.E.N., M-A.H., K.A., M.R., and K. € O. were involved with the experimental work and critically evaluated the manuscript.M.L. and P.-H.G. participated in the interpretation of the results and critically evaluated the manuscript.All authors approved the final content of the manuscript.The Yrj€ o Jahnsson Foundation (20197174), Academy of Finland (275614, 316664, and 315568), Novo Nordisk Foundation (#NNF OC0013659), Signe and Ane Gyllenberg Foundation, Folkh€ alsan Research Foundation, Helsinki University Central Hospital Research Funds, the Diabetes Research Foundation, the P€ aivikki and Sakari Sohlberg Foundation, the Finnish Foundation for Cardiovascular Research, the Northern Finland Health Care Support Foundation, the Finnish Medical Foundation, Wilhelm and Else Stockmann Foundation and Nylands Nation supported this study.We thank study nurses and coordinators for their assistance in conducting the study visits.

Fig. 1 .
Fig. 1.Plasma immunoglobulins to CuOx-LDL and MAA-LDL before and after surgery.Specific plasma IgA, IgG and IgM antibody levels to oxLDL before and after Roux-en-Y (RYGB) surgery, measured at four different timepoints (0, 30, 60, 120 min) during an oral liquid meal test.Specific plasma antibodies to copper-oxidized low-density lipoprotein (CuOx-LDL) are shown in upper panels A-C.Specific plasma antibodies to malondialdehyde acetaldehyde-modified low-density lipoprotein (MAA-LDL) are shown in lower panels D-F.Plasma glucose levels at different timepoints were also measured during the meal test, and results are shown in panel G as comparison.These results represent median values at each timepoint from the whole study population (N = 30).RLU, relative light unit.** Presents p-value under 0.001 and * presents p-value under 0.05.

Fig. 2 .
Fig. 2. Specific plasma antibody binding to bacterial virulence factor Rgp44 and PCho before and after surgery.Specific plasma IgA, IgG and IgM antibody levels to bacterial virulence factors Rgp44 and PCho before and after Roux-en-Y (RYGB) surgery, measured at four different timepoints (0, 30, 60 and 120 min) during an oral liquid meal-test.Specific plasma antibodies to Porphyromonas gingivalis gingipain A hemagglutinin domain Rgp44 are shown in upper panels A-C.Specific plasma antibodies to phosphocholine (PCho) are shown in lower panels D-F.Plasma glucose levels at different timepoints were also measured during the meal test, and results are shown as comparison in panel G (same panel in Figure 1).These results represent median values at each timepoints from the whole study population (N = 30).RLU, relative light unit.**Presents p-value under 0.001 and * presents p-value under 0.05.

Table 1 .
The data are presented as mean 190 Ó 2023 The Authors.APMIS published by John Wiley & Sons Ltd on behalf of Scandinavian Societies for Pathology, Medical Microbiology and Immunology.HAPPONEN et al.

Table 2 .
Clinical characteristics from healthy individuals BP, blood pressure.The data are presented as mean AE SD.