Plasma concentrations of syndecan- 1 are dependent on kidney function

Background: Elevated plasma concentrations of syndecan- 1 and heparan sulfate in studies of trauma, sepsis, and major surgery are commonly assumed to indicate acute glycocalyx degradation. We explored a possible role of the kidneys for these elevations. Methods: Plasma and urine concentrations of syndecan- 1, heparan sulfate, and biomarkers of inflammation were measured over 5 hours in 15 hospital patients treated for post- burn injury. The renal clearances of syndecan- 1 and heparan sulfate ( CL R ) were calculated and their influence on the plasma concentration predicted by simulation. Results: The urine/plasma CL R varied Multiple linear regression analysis


| INTRODUC TI ON
Measuring syndecan-1 and heparan sulfate concentrations in plasma is a common way of assessing the integrity of the endothelial glycocalyx layer. More than 200 studies have been published that show the frequent occurrence of threefold to fourfold elevations of the plasma concentrations of endothelial surface proteins in acute and chronic inflammation, trauma, sepsis, and major surgery. 1-3 A major challenge is to find different ways to avoid these elevations, as they are interpreted to imply acute injury to the glycocalyx layer. 4 However, the metabolism of these glycocalyx components is poorly understood, and events other than glycocalyx degradation might possibly explain a rise in their plasma concentrations. For example, their production could be upregulated or their elimination could become less efficient. 4,5 In this study, we assessed the renal clearance (CL R ) as a determinant of the plasma concentration of glycocalyx degradation products. The data were derived from post-burn patients with a marked inflammatory response and elevated levels of syndecan-1 and heparan sulfate.
The hypothesis was that CL R is a major determinant of the plasma concentration. For this purpose, the CL R of syndecan-1 and heparan sulfate were calculated (primary outcome), their variability was examined, and simulations were performed to predict how the plasma concentrations are likely to change when CL R is varied (secondary outcomes).

| ME THODS
This present study is a secondary report of a prospective singlecenter open-label trial of the volume effects and capillary leakage of 20% albumin in 15 patients recruited from the Burns Unit of Linköping University Hospital, Sweden, between October 2016 and January 2019. 6,7 The study was approved by the Regional Ethics Committee of Linköping (Dnr 2016/333-32) and the Swedish Medical Products Agency (Eu-nr 2016-000996-26) and was registered at clinicaltrials.gov (NCT02952378).
Inclusion criteria were a burned Total Body Surface Area (TBSA) of > 6% and age 18-80 years. Exclusion criteria were unconscious patients or those with severe allergies, kidney failure, or heart failure. The patients were recruited during a visit 1-2 days before the trial, which took place between 4 and 14 (mean 7) days after the burn injury. They gave their informed consent both orally and in writing.

| Sampling and analysis
The study was performed on 14 patients in the morning and on 1 patient in the evening. All patients were hemodynamically stable.
They had fasted overnight but were allowed to ingest 1 sandwich and drink 1 glass (2 dL) of liquid 90 minutes prior to the experiment.
The patients were placed in the supine position for at least 30 minutes before baseline measurements were taken.
Each patient received an intravenous infusion of 3 mL/kg 20% albumin over 30 minutes. Blood was collected in lithium-heparin plasma gel tubes and used for the measurement of the plasma albumin, creatinine, interleukin-6 (IL-6), and C-reactive protein (CRP) concentrations at 0, 60, and 300 minutes. The analyses were performed on a Cobas ® 8000 system (Roche Diagnostics, Basel, Schweiz) at the certified clinical chemistry laboratory at Linköping University Hospital.
The plasma concentrations of 2 endothelial shedding products, syndecan-1 and heparan sulfate, were analyzed at the research laboratory of Södertälje Hospital at Biovation Park, Södertälje, using commercially available ELISA kits (from Diaclone, France, and Amsbio, Abingdon, UK) with coefficients of variation (CV) of 6.2% and < 10%, respectively.
The syndecan-1 assay uses an antibody that is highly specific for the syndecan-1 (CD-138) molecule. The heparan sulfate kit is based on an antibody that reacts with the 10E4 epitope, which is present in many types of heparan sulfates. Samples with high concentrations of syndecan-1 were re-analyzed after dilution to 1:3 and to 1:10, and heparan sulfate was re-analyzed with a dilution of 1:3.
Urine samples were collected from a catheter bag or through voluntarily voiding at 0 and 300 minutes. The first urine collection was made just before the infusion of 20% albumin was initiated. Urine was analyzed for creatinine, syndecan-1, heparan sulfate, albumin, and α-microglobulin by the same technical methods used for the plasma samples.
Urine was also collected to assess the risk of acute kidney injury by Nephrocheck ® (Astute Medical, San Diego, CA), which uses the product of 2 cell-cycle arrest biomarkers, insulin-like growth factor binding protein 7 (IGFBP7) and tissue inhibitor of metalloproteinase 2 (TIMP-2), to calculate an index called AKIRisk ™ . 8,9 Risk values > 0.3 increase and > 2.0 greatly increase the risk of developing acute kidney injury within the subsequent 24 hours.

| Calculations
The renal clearance (CL R ) of syndecan-1, heparan sulfate, and creatinine during the 5 hours experiment was calculated as the product of their urinary concentration and the excreted urine volume divided by the average plasma concentration at 0 and 300 minutes. The CL R is the body volume that is completely cleared from the measured molecule by renal excretion per minute. The CL R for syndecan-1 between 0 and 5 hours was given by: The fractional excretion (FE) of syndecan-1 was calculated as follows at 0 and 5 hours: The FE of heparan sulfate was calculated in the same way. The FE is the renal clearance of the studied molecule relative to the creatinine clearance.

Editorial Comment
Measurements of syndecan-1 and heparan sulfate are commonly used to assess degree of glycocalyx degradation, a marker for the endothelial response to inflammation. In this study, concentrations of both syndecan-1 and heparan sulfate in both plasma and urine were serially measured in a cohort of burn patients, to assess the renal clearance of both proteins. Renal clearance was highly variable among patients, and this variability can potentially explain multifold differences in measurements not accounting for differential clearance. The results demonstrate an important limitation in using syndecan-1 and heparan sulfate measurements to assess glycocalyx degradation.
Worked-through examples of the influence of CL R changes on the plasma concentration were constructed based on a onecompartment kinetic model using 3 L as an arbitrary volume of dis- The plasma concentration at steady state (C ss ) was given by the The time required to reach C ss was set to equal four half-times (T 1/2 ), each of which is given by ln 2 (=0.693) V d /CL R . The plasma concentration C after any time t was 10 :

| Statistics
Data were presented as the median and interquartile range.
Relationships between variables were evaluated by simple and multiple linear regression analysis, where r = correlation coefficient.
Stepwise linear regression was used to identify predictors for the multiple regression analyses. Univariate analysis was used except where noted. P <.05 was considered statistically significant.

| Plasma and urine concentrations
Demographic and basic biochemical data are shown in Table 1.
There was a reciprocal correlation between the plasma concentration and the urinary excretion of syndecan-1, both at baseline and during the experiment ( Figure 1A). By contrast, the excretion of heparan sulfate increased directly with its plasma concentration ( Figure 1B).
The inflammatory markers IL-6 and C-reactive protein correlated only vaguely with the plasma concentrations of syndecan-1 and heparan sulfate. However, the urinary excretion of heparan sulfate increased with the mean IL-6 concentration (r = 0.66, P <.01).

| Urine/plasma ratio
When the study started, the urine/plasma concentration ratio was

| Renal clearance
The CL R of syndecan-1 during the 5 hours experiment was 2.26 (0.66-4.00) mL/min, and it varied 250-fold. The CL R of heparan TA B L E 1 Basic data for the 15 patients. Data are the median and 25th-75th quartile range for the value obtained just before initiation of the infusion of 20% albumin except where noted  Using multiple linear regression, the CL R of syndecan-1 was positively associated with creatinine clearance (P <.0032; Figure 1D) and with the urine flow (P <.015). The latter amounted to 1.4 (1.3-3.3) mL/min.
Similarly, the CL R of heparan sulfate was independently and positively associated with the plasma concentration of IL-6 (P <.003; Figure 1E) and the urine flow (P <.01).
By contrast, the creatinine clearance was on the high side, 198 (158-239) mL/min.

| Fractional excretion
The fractional excretion (FE) is the CL R divided by the creatinine clearance.

| Cell-cycle arrest biomarkers
Urinary concentrations of IGFBP7 and TIMP2 correlated strongly with each other (r = 0.86, P <.0001). Their relative changes during the experiments agreed well with the dilution of the urine, but not with IL-6 levels or with the creatinine clearance.
The product of these biomarkers, the AKIRisk ™ score, decreased from 0.23 (0.15-0.82) to 0.15 (0.14-0.25) during the study. Six patients before infusion and 3 patients at the end of the study scored > 0.3, which is the cut-off indicating a risk for the development of acute kidney injury.
No statistically significant correlation with other biomarkers was found.
The change in AKIRisk during the study correlated closely with the dilution of the urine (r = 0.83; P <.001), and the numerical ratios agreed even better when AKIRisk was compared with the squared dilution (r = 0.86; P <.001, Figure 1F). We tested the latter because the 2 components of the AKIRisk (IGFBP7 and TIMP2) should both be corrected for urine dilution.

| Complications
One patient suffered complications during the subsequent hospital care. This was a male patient with the largest burn wound (48%) and highest CRP (300 µg/L) in the series. He developed chills and fever

| Worked-through examples
Mathematical examples were used to illustrate the consequence of the variability in CL R on the plasma concentration C. Our study also included measurements of the urinary concentrations of 2 renal cell-cycle arrest biomarkers, IGFBP7 and TIMP2.
Their product yields an index, AKIRisk, for which high values predict an increased risk for the development of acute kidney injury.
The only finding was that the AKIRisk value correlated closely with the urine dilution ( Figure 1F). We and others have suggested that AKIRisk should routinely be corrected for dilution, as is the common practice for most other biochemical measurements performed on sampled urine. 13,14 The manufacturer and experts on kidney injury claim that the results stand out even when corrected for dilution.
An insight relevant to this debate might be that the AKIRisk value shown in Figure 1F should be corrected by the square of the urine dilution to yield similar proportions because both biomarkers that make up the AKIRisk should separately be corrected for dilution.
Urine creatinine can be used as marker of dilution in patients with a reasonably normal creatinine clearance, while other biomarkers of dilution, such as urine specific weight or urine color, might be more suitable in other settings. 15,16 Urine creatinine was used here because the glomerular filtration is known to be elevated during the 2nd week after burn injury, 17 which places the creatinine clearance in a high range.
The limitations of this study include that the excreted syndecan-1 and heparan sulfate is assumed to stem from the bloodstream, although syndecan-1 is expressed in the renal tubules as well. The predictions of changes in plasma concentrations resulting from a reduction of CL R assume that the studied glycoproteins follow one-compartment kinetics, that their release to the circulation is unchanged, and that the plasma volume is 3 L. Moreover the data represent a secondary publication to a study of the clinical efficacy of 20% albumin in burn injury. 6,7 The patients had overcome the acute stage of burn injury and were studied approximately 1 week after the burn incident. 1 patient had suffered from inhalation injury along with skin burns. At the time of the study, their clinical challenges were inflammation and wound infection. The small size of the study group should be noted, and the relatively high BMI is probably a remnant of the early volume loading performed during the acute phase of the burn injury.
Our testing of how glycocalyx shedding products survive storage of plasma samples until analysis also shows that syndecan-1 is quite stable, whereas heparan sulfate levels decrease over time.
This fact might explain why the heparan sulfate concentrations reported here are lower than those reported in earlier work by our group. 18

| CON CLUS IONS
Urinary excretion is an important route for elimination of syndecan-1 and heparan sulfate. Their CL R values vary greatly and are associated with kidney function variables, such as creatinine clearance and urine flow. Excretion of heparan sulfate seems to be facilitated by inflammation. The simulations illustrate that sudden reductions in the CL R values for syndecan-1 and heparan sulfate are likely to cause several-fold elevations in the plasma concentrations of these substances even without an assumption of increased glycocalyx degradation.

Associate professor Camilla Krizhanovskii and PhD student Stelia
Ntika performed the analyses of sydecan-1 and heparan sulfate.
The authors thank the staff of the Burns Unit at the Linköping University Hospital in Linköping, Sweden, for assistance during the data collection.

CO N FLI C T O F I NTE R E S T
RGH holds a grant from Grifols for the study of 20% albumin as infusion fluid. MZ and JZ declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data are available as a File S1.xls and the calibration curves for the ELISA kits as File S2.docx.