Degradation of endothelial glycocalyx in Tanzanian children with Falciparum malaria

Abstract

A layer of glycocalyx covers the vascular endothelium serving important protective and homeostatic functions. The objective of this study was to determine if breakdown of the endothelial glycocalyx (eGC) occurs during malaria infection in children. Measures of eGC integrity, endothelial activation, and microvascular reactivity were prospectively evaluated in 146 children: 44 with moderately severe malaria (MSM), 42 with severe malaria (SM), and 60 healthy controls (HC). Biochemical measures of eGC integrity included plasma syndecan-1 and total urinary glycosaminoglycans (GAG). Side-stream dark field imaging was used to quantitatively assess integrity of eGC. Plasma angiopoietin-2 (Ang-2) was measured as a marker of endothelial activation and also as a possible mediator of eGC breakdown. Our results show that urinary GAG, syndecan-1, and Ang-2 were elevated in patients with MSM and SM compared with HC. Syndecan-1 and GAG levels correlated significantly with each other and with plasma Ang-2. The eGC breakdown products also inversely correlated significantly with hemoglobin and platelet count. In the MSM group, imaging results provided further evidence for eGC degradation. Although not correlated with markers of eGC degradation, vascular function [assessed by non-invasive near infrared spectroscopy (NIRS)] demonstrated reduced microvascular reactivity, particularly affecting the SM group. Our findings provide further evidence for breakdown of eGC in falciparum malaria that may contribute to endothelial activation and adhesion of parasitized RBC, with reduced nitric oxide formation, and vascular dysfunction.

Introduction

In 2019, malaria cases totaled approximately 229 million with 409,000 deaths.¹ Over 90% of cases and deaths occur in Africa where Plasmodium falciparum is the predominant parasite species and children are most vulnerable to death.¹ Factors including endothelial activation, binding of parasitized red blood cells (RBC) to vascular endothelium, and lysis of infected RBC lead to microvascular dysfunction with parasite sequestration, vascular obstruction, impaired perfusion, tissue dysoxia, and organ dysfunction.² Reduced nitric oxide (NO) availability, endothelial activation and microvascular dysfunction have been shown to correlate with disease severity and outcomes.^3,4,5

A glycocalyx layer is present on cells throughout the body, consisting of glycosaminoglycans (GAG) and proteoglycans.⁶ In the vasculature, the endothelial glycocalyx (eGC) serves to protect and maintain homeostasis.⁷ This gel-like protective layer is comprised of proteoglycan molecules, known as core proteins, anchored to the endothelium with various GAG molecule side chains covalently attached.^6-8 The major GAG components present in the eGC include chondroitin sulfate (CS) and heparan sulfate (HS) as well as hyaluronic acid, while the major proteoglycans include syndecans and glypicans.^6-8 The functions of the eGC include maintaining vessel integrity,⁹ and modulating signalling pathways¹⁰ and cellular adhesion (e.g., of platelets and infected RBC).^11-13 Of special interest regarding malaria, endothelial glycocalyx mediates the signal for NO production in response to shear stress of flowing blood.^14-16

Breakdown of the glycocalyx is associated with a broad range of chronic disease states including cardiovascular disease, diabetes mellitus, and chronic kidney disease.⁶ In infectious diseases, dysfunction or demonstrated breakdown of eGC has been implicated in clinical disease severity in sepsis,¹⁷ dengue,¹⁸ and malaria.^19-21 In sepsis, measures of eGC breakdown serve as a predictor of mortality. ^17,22 Breakdown of eGC is associated with increased vascular permeability and disease severity in dengue infection.¹⁸

Loss of eGC may contribute to malaria pathogenesis in multiple ways. Glycocalyx impairs cytoadhesion of P. falciparum infected RBC to the endothelial CD36 receptor in an in vitro assay.¹² In mouse experimental cerebral malaria, degradation of the glycocalyx in cerebral vasculature was reported.²³ A follow-up study using the experimental cerebral malaria model identified loss of select glycoprotein and GAG components of the eGC in the brain and reported increased vascular permeability.²⁴ In clinical studies in falciparum malaria, we reported evidence of breakdown of eGC in children with uncomplicated and severe disease and eGC breakdown associated with disease severity and mortality in adults.^19,20 Lyimo et al. imaged microvessels in the buccal mucosa of children with falciparum malaria and reported loss of eGC along with an increased prevalence of perivascular hemorrhages.²¹

Our hypothesis is that degradation of endothelial glycocalyx occurs in patients with P. falciparum malaria, leading to endothelial activation, impaired blood-flow-mediated NO synthesis and vasodilation, and increased adherence of infected RBC and other cells to vascular endothelium. These events contribute to the pathogenesis of vascular dysfunction in malaria. In this prospective study, we performed extensive evaluation of the endothelial glycocalyx in children with moderately severe and severe P. falciparum malaria and healthy control children. The evaluations include markers of disease severity; measurement of plasma and urine breakdown products of eGC; imaging of eGC with sidestream, dark field, non-invasive microscopy; endothelial activation; and microvascular function. We describe the relationship between eGC integrity and these factors.

Materials and Methods

Human Subjects:

This prospective study was conducted at Hubert Kairuki Medical University (HKMU), Dar es Salaam, Tanzania. Ethical approvals were obtained from Institutional/ethical Review Boards at HKMU, Duke University Medical Center, and the Tanzanian National Health Research Ethics Committee at the National Institute for Medical Research. Parents or guardians of all participating children provided informed consent.

Healthy control (HC) children, and children with moderately severe malaria (MSM) and severe malaria (SM) aged 1 to 11 years were prospectively enrolled from 2016 to 2019. Healthy control subjects were well and without signs or symptoms of illness including fever or presence of malaria parasitemia.

Malaria was determined by positive rapid diagnostic test (BinaxNOW Malaria, Abbott Point of Care, Princeton, NJ, USA) and confirmed with microscopy by the presence of P. falciparum >2,500 parasites/μL. We excluded children with infection with non-falciparum malaria. Those with MSM required hospitalization for intravenous fluid administration due to clinical symptoms consistent with malaria including fever, but did not exhibit any World Health Organization (WHO) criteria for severe malaria.²⁵ Severe malaria was classified by presence of 1 or more modified WHO criteria of severity.²⁵ All MSM and SM children received inpatient parenteral antimalarial drug therapy with artesunate, as well as other supportive care per standard Tanzanian national protocols and as recommended by WHO.²⁵

Clinical and laboratory assessments:

A history and clinical assessment were performed including assignment of Blantyre coma score for cerebral malaria. Clinical assessments measured at baseline and daily included vital signs, routine biochemistry (sodium, chloride, potassium, bicarbonate, urea nitrogen, glucose, calcium, total bilirubin, creatinine, and lactate) by bedside biochemical analyzer (i-STAT, Abbott Point of Care, Princeton, NJ, USA); and complete blood count measured with an automated counter (Beckman-Coulter Act 10, Brea, CA, USA). Demographic and clinical data were collected and managed during the study using the Research Electronic Data Capture (REDCap) tools ²⁶ hosted at Duke University.

Other biochemical studies:

Blood and urine were collected once for HC and once daily for up to 3 days for MSM and SM to assess changes over the period of acute illness. Urine was collected voluntarily into a sterile urine collection cup or with use of a pediatric bag (U-bag urine collector, Hollister Pediatric, Libertyville, IL, USA) affixed to the perineum with adhesive.

Blood was processed by centrifugation within 30 to 90 minutes after venipuncture. Plasma was stored frozen at −80° C and shipped frozen in a liquid nitrogen dry shipper for research analyses in Durham, NC, USA.

Markers of disease severity:

Level of parasitemia, plasma histidine rich protein-2 (HRP2), lactate, platelet count, hemoglobin (Hb), and angiopoietin2 (Ang-2) were considered markers of disease severity. Parasite counts of P. falciparum were measured using Giemsa-stained thick and thin fields, and cross-checked by experienced microscopists. To quantitate total parasite biomass, we measured plasma HRP2 by ELISA as described previously.^4,27

Glycocalyx integrity measures:

Two major constituents of eGC were measured to assess extent of eGC degradation. These included total urinary GAG and the plasma core protein syndecan-1. Total sulfated GAG level in urine was measured by the colorimetric dimethylmethylene blue (DMMB) assay, described previously.²⁸ Prior to assay, urine was thawed and centrifuged for clarification. Absorbance at 525 nm was measured, and GAG concentrations were determined based on a chondroitin sulfate calibration curve. Urinary creatinine concentration was determined by the alkaline picrate method and then used to normalize urinary GAG concentrations (g/mol Cr).²⁹ Plasma syndecan-1, a core protein of eGC, was measured by ELISA (Abcam, Cambridge, MA, USA).

Glycocalyx imaging:

To further assess integrity of the eGC, sidestream dark-field video imaging was performed with a handheld, video microscope (GlycoCheck camera and associated analysis software; Glycocheck and Microvascular Health Solutions, American Fork, UT, USA) for real-time imaging of microvasculature. Three vascular beds were identified for imaging including sublingual, axilla and pinna of the external ear. The camera allows visualization of the flowing column of RBC; lateral movement of RBCs is used to determine the perfused boundary region (PBR) which represents the part of eGC that RBCs can penetrate. The PBR is used as an indirect measure of integrity of the eGC, with increased PBR values indicating deeper penetration of RBC toward the endothelium.³⁰ Vessels assessed ranged in size from 5 to 25 microns in diameter. Duration of imaging generally took less than 5 minutes with output captured on a laptop computer. Imaging was either not done or limited to the non-sublingual vascular beds in children who could not cooperate with sublingual imaging. Parameters captured included % RBC filling, perfused boundary region (PBR) measurement for varied sizes of vessels, and density of perfused vessels. RBC filling refers to the average RBC content of the vessels.

Endothelial activation:

We used Ang-2 as a marker of endothelial activation and also disease severity. Ang-2 plasma concentration was determined by ELISA (R&D Systems, Minneapolis, MN, USA).

Microvascular function:

Near infrared spectroscopy (NIRS) was used to measure tissue hemoglobin oxygen saturation (StO2%), as described previously. ^31,32 A short period of vascular occlusion was performed using a pneumatic cuff applied to the arm followed by recovery, to evaluate the capacity for re-oxygenation of the tissue. StO2 measured at various time points and associated parameters were analyzed. These included the StO2 at baseline and during occlusion, the peak hyperemia response, as well as the slope of StO2 recovery during reperfusion that followed the ischemic period. We used the slope of recovery to assess microvascular reactivity of the tissue. We also measured the hemoglobin perfusion index (THI) and skeletal muscle tissue oxygen consumption (VO2) [defined as the difference in tissue O2 content (THI x 1.39 x StO2) before and at the end of vascular occlusion, divided by duration] as previously reported.^5,31,32

Statistical methods:

Differences between groups (HC, MSM, SM) were compared with analysis of variance (ANOVA) or Kruskal-Wallis as appropriate, depending on normality of distribution. Comparisons between means of each group (SM vs. MSM; SM vs. HC; MSM vs HC) were performed using the Sidak method adjusting for multiple comparisons. The Spearman correlation coefficient was used to explore correlations. For correlation analysis, the Bonferroni method was used to adjust for multiple comparisons, and the new alpha level was determined to be 0.01. We used linear regression analysis to determine the effects of multiple factors. Two-sided P-values of < 0.05 were considered significant. Correlation and regression analyses were performed using IBM SPSS v25; all other analyses were performed using GraphPad Prism 8.4.3.

Results

At HKMU in Dar es Salaam, Tanzania, we prospectively evaluated 146 children aged 2-11 years, including 60 HC subjects, 44 patients with MSM and 42 meeting WHO criteria for SM.²⁵ Of the 42 patients who had SM, one subject (2%) met five of the criteria for SM, 3 subjects (7%) met three criteria, 11 subjects (26%) met two criteria, and 27 subjects (64%) met one of the criteria. The breakdown of criteria met for diagnosis of SM are summarized in Supplementary Table 1.

Ages of the children ranged from 2-11 years in all groups, however the median age of children in the HC group was significantly higher than those in the malaria groups (p<0.001). The median weight showed a similar pattern being higher in the HC group compared with MSM and SM groups (p<0.001). There were no significant differences for age, weight or vital signs between the MSM and SM groups.

A blood transfusion was administered in 1/44 (2%) subjects with MSM and 9/42 (21%) with SM. Hepatosplenomegaly was present in 9/44 (20%) children with MSM and 20/42 (48%) with SM. In the SM group, prostration was present in 22/42 (52%), deep breathing in 5/42 (12%), nasal flaring in 4/42 (10%), intercostal retractions in 4/42 (10%), rigidity in 3/42 (7%), and opisthotonus in 1/42 (2%). Seizures were reported in 1/44 (2%) children with MSM and 10/42 (24%) with SM. In total, 2/42 (5%) patients with SM had coma (BCS<3), classifying as cerebral malaria. One death occurred due to malaria; this child met 5 criteria for SM on admission. Baseline demographic and clinical characteristics and hematology, biochemical and microvascular function are summarized in Tables 1 and 2, respectively.

Table 1.

Baseline demographic & clinical characteristics^a

Characteristic	Healthy Control (n=60 )	Moderately Severe Malaria (n=44)	Severe Malaria (n=42)	P valueb
Male sex, no (%)	40 (66.7)	26 (59.1)	31(73.8)	0.352
Age (years)	9 (2-11)	5 (2-11)	5 (2-11)	<0.001
Weight (kg)	20.9 (12.0-36.5)	13.2 (10.0-28.3)	14.0 (10.0-33.0)	<0.001
Temperature ( °C)	36.1 (35.1-37.5)	38.0 (34.4-40.0)	38.4 (34.0-40.5)	<0.001
Heart rate (beats per minute)	94 (28-154)	139 (99-170)	135 (78-190)	<0.001
Respiratory rate (breaths per minute)	24 (18-90)	36 (24-52)	34 (24-68)	<0.001
Systolic blood pressure (mmHg)	100 (80-120)	90 (80-120)	93 (80-120)	0.028
Diastolic blood pressure (mmHg)	60 (40-80)	50 (35-70)	55 (35-60)	<0.001

^aMedian (range), except when indicated otherwise

^bANOVA or Kruskal-Wallis test for comparison of HC, MSM and SM groups.

Table 2.

Baseline hematology, biochemical and tissue oxygen saturation^a

Parameter	Healthy Control (n=60)	Moderately Severe Malaria (n=44)	Severe Malaria (n=42)	P valueb
WBC (x109 cells/L)	7.1 (3.0-13.0)	7.4 (4.3 – 17.3)	8.0 (2.3-27.2)	0.198
Hemoglobin, g/dL	11.8 (8.7-13.8)	9.5 (6.1-15.1)	8.4 (4.5-11.9) c	<0.001
Platelets (x109 cells/L)	352.5 (98.0-542.0)	134.5 (22.0-488.0)	91.0 (13.0-329.0)	<0.001
Lactate (mM)	2.3 (0.8-4.1) (n=30)	3.1 (1.5-5.1) (n=27)	4.1 (0.02-15.0) (n=30)	<0.001
Creatinine (serum; mg/dL)	0.4 (0.2-0.6) (n=20)	0.3 (0.2-0.4) (n=18)	0.3 (0.2-0.5) (n=19)	0.056
Parasite density (# per μL; geometric mean, 95% confidence interval)	0	107,640 (86,747 – 133,565)	155,718 (112,205-216105)	0.015 d
HRP-2 (ng/mL)	0.01 (0.01-0.04)	39.1 (0.01-1317)	60.9 (3.06-1625)	<0.001
Urinary GAG (g/mol Cr)	4.24 (0.04-12.46)	10.18 (0.35-83.73.)	13.77 (4.01-25.96)	<0.001
Plasma syndecan-1 (ng/mL)	65.0 (22.80-276.3)	235.3 (20.70-899.4)	273.7 (35.0-1558.0)	<0.001
Plasma angiopoietin-2 (pg/mL)	1605 (806-6290)	3447 (1256-16021)	4612 (2186-30491)	<0.001
StO2, at baseline (%)	80 (55-93)	75 (49-93)	79 (50-94)	0.055
StO2 low, at end of occlusion (%)	39 (13-62)	37 (0-73)	50 (4-84) e	0.035
Recovery StO2 % increase/sec	1.42 (−0.02-4.17)	1.20 (0.03-5.96)	0.61 (0.01-2.55) e	0.005
Peak StO2 after release %	90 (59-98)	86 (61-96)	87 (55-97)	0.116
Difference between peak and baseline StO2%	8.2 (−4.4-26.9)	10.9 (0.8-35)	7.8 (−11.7-26.0)	0.045
O2 consumption (VO2)	170 (13-245)	132 (45-286)	122 (26-276)	0.043
Tissue Hb Index at baseline	11.4 (5-14.1)	10.8 (4.9-15.1)	10.9 (6.0-18.7)	0.425
Tissue Hb Index, end of occlusion	7.2 (3.3-11.65)	7.2 (3.5-13.6)	7.6 (3.2-13.0)	0.294

^aMedian (range), except when indicated otherwise

^bANOVA or Kruskal-Wallis test for comparison of HC, MSM and SM groups

Significant difference between SM and MSM groups:

^chemoglobin, p<0.001

^dparasitemia (by Mann-Whitney), p=0.015

^eoxygen saturation measures, p<0.05

Abbreviations: HRP-2: histidine rich protein-2; StO2 : tissue oxygen saturation

Markers of disease severity:

Parasitemia, lactate, platelet count, hemoglobin, plasma HRP2, and Ang-2 are markers of disease severity in malaria. There were significant differences between HC vs. MSM and HC vs. SM groups for each of these markers (p<0.001 for all), except for lactate (HC vs MSM; p=0.09). On pairwise comparison between the MSM and SM groups, only parasitemia (p=0.015) and hemoglobin (p<0.001) were significantly different (Table 2). Significant correlations were present between certain markers of disease severity and loss of glycocalyx, as described below.

Glycocalyx integrity measures:

Syndecan-1 and GAG are known breakdown products of eGC. We measured total GAG levels in urine at baseline for 54 HC subjects, 37 with MSM and 39 with SM (Table 2; Fig 1). Total urinary GAG levels were higher in SM and MSM compared with healthy control subjects (p<0.001). Levels of urinary GAG were not significantly higher in SM compared with MSM groups on pairwise comparison of the groups. The total urinary GAG measured by the DMMB assay represents all sulfated GAG. We previously tested the supernatant medium from cultures of P. falciparum growing in human RBC and did not find any GAG.¹⁹ This was done to confirm that the parasite or infected RBC were not serving as sources of GAG.

Fig 1.

Total urinary glycosaminoglycan levels measured by DMMB colorimetric assay in Tanzanian children. Median values indicated by horizontal bars. P<0.001 by ANOVA; multiple comparison p-values shown for MSM vs. HC and MSM vs. SM. Two GAG values were excluded from analysis, due to questionable assay results and identification as outliers by Grubb’s statistical method (1 in the Control and 1 from the MSM group).

We measured plasma syndecan-1, a predominant core protein of the eGC in 53 control subjects, 39 with MSM and 40 with SM (Table 2; Fig 2). Levels of syndecan-1 were significantly higher in SM and MSM compared to HC (p<0.001). However, there was not a significant difference between MSM and SM on pairwise comparison.

Fig 2.

Plasma syndecan-1 levels measured by ELISA. Median values indicated by horizontal bars. P<0.001 by Kruskal-Wallis test; multiple comparison p-values shown for MSM vs. HC and MSM vs. SM. Note that the y-axis is displayed on a logarithmic scale.

Correlations between markers of eGC integrity, endothelial activation, and disease severity are shown in Table 3. Total urinary GAG and plasma syndecan-1 were significantly correlated in all patients with malaria (p=0.003). There were also significant correlations between the eGC breakdown products and common markers of disease severity in malaria (e.g. blood levels of Hb, platelet count, HRP2) in patients with malaria. There were strong inverse correlations between total urinary GAG and Hb for all malaria patients (p<0.001) and SM (p=0.004), but this correlation was not significant in the MSM group (p=0.080). There were significant inverse correlations between syndecan-1 and Hb in all malaria patients (p<0.001) and SM (p=0.009). There were also significant inverse correlations between plasma syndecan-1 and platelet count in all malaria patients (p=0.001). An inverse correlation between urinary GAG and platelet count was present in all malaria patients (p=0.004) and in SM (p=0.009), but this was not significant for MSM. There were positive correlations between plasma syndecan-1 and HRP2 in all malaria (p=0.010) patients and MSM (p=0.008), but this was not significant for SM. There were no significant associations between GAG and lactate, HRP2, or quantitative parasitemia, nor between syndecan-1 and lactate or quantitative parasitemia.

Table 3:

Spearman rho correlation coefficient for markers of eGC breakdown, endothelial activation and disease severity

		Ang-2	Syndecan-1	GAG	Hemoglobin	Platelet	HRP2
Ang-2	MSM		0.379	0.192	−0.317	−0.405	0.497 **
	SM		0.653 **	0.541 **	−0.372	−0.338	−0.007
	All		0.520 **	0.410 **	−0.378 **	−0.427 **	0.254
Syn-1	MSM	0.379		0.247	−0.353	−0.365	0.424 **
	SM	0.653 **		0.395	−0.410 **	−0.242	0.167
	All	0.520 **		0.341 **	−0.403 **	−0.355 **	0.289 **
GAG	MSM	0.192	0.247		−0.300	−0.146	0.089
	SM	0.541 **	0.395		−0.452 **	−0.410**	0.116
	All	0.410 **	0.341 **		−0.441 **	−0.327 **	0.107
Hb	MSM	−0.317	−0.353	−0.300		0.486 **	−0.515 **
	SM	−0.372	−0.410 **	−0.452 **		0.239	−0.044
	All	−0.378 **	−0.403 **	−0.441 **		0.440 **	−0.330 **
Platelet	MSM	−0.405	−0.365	−0.146	0.486 **		−0.595 **
	SM	−0.338	−0.242	−0.410**	0.239		−0.399
	All	−0.427 **	−0.355 **	−0.327 **	0.440 **		−0.502 **
HRP2	MSM	0.497 **	0.424 **	0.089	−0.515 **	−0.595 **
	SM	−0.007	0.167	0.116	−0.044	−0.399
	All	0.254	0.289 **	0.107	−0.330 **	−0.502 **

Spearman rho correlation coefficients. Bold face indicates significant values (**p ≤0.01)

Bonferroni method was used to adjust for multiple comparisons and the new alpha level was determined to be 0.01.

MSM: moderately severe malaria; SM: severe malaria; “All” refers to the combined MSM and SM groups

Glycocalyx imaging:

We imaged vascular beds in the axilla of 53 HC children, 38 with MSM, and 37 with SM (Table 4), as well as in the pinna (ear) in 51 HC children, 32 with MSM, and 32 with SM. Sublingual imaging was not feasible in the majority of patients because of the inability of the younger subjects to be still long enough for the measurement. PBR is used as an indicator of eGC integrity, such that when breakdown occurs, red blood cells can perfuse closer to the endothelium resulting in increased PBR. When the 3 groups were compared, there were significant differences for several measures in axilla and pinna: increased PBR_20-25 for axilla (p=0.038); and decreased RBC filling for pinna (p=0.034). However, there were no significant differences across the 3 groups for PBR measured in pinna.

Table 4.

Glycocalyx Imaging Results, Axilla^a

Parameter	Healthy Control (n=53)	Moderately Severe Malaria (n=38)	Severe Malaria (n=37)	P valueb
PBR 5-25 (μm)	1.86 (0.63-3.09)	2.17 (1.35-3.58)	1.94 (1.39-3.54)	0.094
PBR 5-9 (μm)	1.18 (0.45-1.79)	1.20 (0.75-2.02)	1.20 (0.54-1.54)	0.460
PBR 10-19 (μm)	2.10 (0.70-3.83)	2.29 (1.46-3.78)	2.17 (1.18-4.08)	0.169
PBR 20-25 (μm)	2.17 (0.71-3.57)	2.55 (1.48-4.05)*	2.32 (1.46-4.31)	0.038
RBC filling (%)	71.7 (38.6-96.2)	64.5 (27.5-86.0)	68.9 (34.5-92.4)	0.131
Density (μm/mm2)	247 (79-816)	353 (86-869)	349 (110-739)	0.061

^aMedian (range), except when indicated otherwise

^bANOVA or Kruskal-Wallis test for comparison of HC, MSM and SM groups

^*p<0.05 between group comparison vs. HC

Due to questions about the validity of imaging in SM (see Discussion), the MSM group was explored further and compared to HC for the axilla measures. There were increases in patients with MSM compared with HC subjects for axillary PBR_5-25 (p=0.034) and PBR_20-25 (p=0.015). In addition there was an increase in microvascular density in axilla (p=0.037). There were no significant correlations between the Glycocheck imaging parameters and markers of eGC breakdown.

Endothelial activation:

We measured plasma levels of Ang-2 as a marker of endothelial activation and disease severity in malaria. Ang-2 is also a purported causative agent for eGC breakdown³³. Ang-2 levels were measured in 54 control subjects, 39 with MSM, and 40 with SM (Table 2 and Fig 3). Levels of Ang-2 were significantly increased in MSM and SM compared to HC (p<0.001). However, levels of Ang-2 were not significantly different between SM and MSM groups (p= 0.132).

Fig 3.

Plasma angiopoietin-2 levels measured by ELISA. Median values indicated by horizontal bars. P<0.001 by Kruskal-Wallis test; multiple comparison p-values shown for MSM vs. HC and MSM vs. SM. Note that the y-axis is displayed on a logarithmic scale.

In all malaria patients, increases in urinary GAG correlated significantly with increases in Ang-2 (p<0.001) (Table 3), and a comparable significant correlation was also present in SM (p<0.001). Likewise, syndecan-1 correlated with Ang-2 in all malaria patients and in the SM group (p<0.001) (Fig 4). When regression analyses were performed to control for other factors associated with malaria disease severity, significant associations remained between total GAG and Ang-2 (p=0.020) and between syndecan-1 and Ang-2 (p<0.001). In all malaria patients, there was a significant inverse correlation between Ang-2 and Hb (p=0.001). There was also a significant inverse correlation between Ang-2 and platelet count (p<0.001) in all malaria patients.

Fig 4.

Correlation (Spearman) between plasma syndecan-1, a core protein from glycocalyx, and plasma angiopoietin-2.

a. SM: Severe malaria group; b. MSM: Moderately severe malaria group; c. All malaria.

Microvascular function:

We evaluated microvascular function using NIRS in 50 healthy control children, 33 with MSM, and 33 with SM. Results are displayed in Table 2. A comparison of microvascular reactivity between healthy control and children with malaria, as measured by the slope of recovery of StO2 following occlusion, is shown in Fig 5. Median microvascular reactivity was 57% lower in SM and 15% lower in MSM compared with the HC group. The difference between SM and HC groups was significant (p=0.008) while the difference between MSM and HC groups was not. There was a significant difference between MSM and SM groups (p=0.021).

Fig 5.

Rate of increase in skeletal muscle reoxygenation (StO2) per second during recovery from ischemic stress in healthy controls and children with moderately severe malaria and severe malaria.

p=0.005 across groups by Kruskal-Wallis. Multiple comparison p-values are shown for MSM vs. HC and MSM vs. SM.

The median oxygen consumption, as measured by the difference in tissue O2 content ([THI x 1.39 x StO2] before and at the end of vascular occlusion) divided by the duration of vascular occlusion, was significantly different among groups (p=0.043). Microvascular reactivity was not significantly associated with Hb, platelet count, parasitemia, HRP2, or lactate. There were no significant associations between microvascular reactivity and syndecan-1 or GAG, the markers of eGC degradation.

Discussion

The findings of this prospective study in Tanzanian children with falciparum malaria build on our prior studies of eGC breakdown in Indonesian adults and Tanzanian children with P. falciparum malaria.^19,20,34 In this study we report results of SDF imaging to assess integrity of eGC, assessment of microvascular function, and biochemical markers of eGC breakdown, disease severity, and endothelial activation.

In summary, elevation of urinary GAG and plasma syndecan-1 (biochemical markers of eGC degradation) were present in children with malaria compared with control subjects. Elevation of plasma syndecan-1 and/or GAG have been reported previously in adults and African children with malaria.^19-21,35 Results of non-invasive imaging showing an increase in the perfused boundary region (PBR) further support the presence of eGC degradation in MSM. Decrement in microvascular reactivity was detected particularly in children with SM. The biochemical markers of eGC breakdown correlated strongly with Ang-2, Hb, and platelet count, but they were not significantly correlated with eGC imaging parameters, microvascular reactivity, or tissue oxygen consumption. The absence of correlation between eGC breakdown products with imaging or microvascular function parameters measured by NIRS may reflect broader loss of eGC across the vasculature, whereas the imaging and functional testing were limited to specific microvascular beds.

Small but significant increases in PBR were detected with imaging of axillary microvessels in children with MSM, supporting loss of eGC. Since many of the children could not comply with the several minutes of minimal movement required for sublingual imaging, we performed imaging of the axilla and pinna of the ear as alternative microvascular sites. In general, we did not find striking differences between malaria patients and controls in these parameters measured, nor were there correlations with eGC breakdown products noted. Similar absence of imaging changes in the presence of elevated biochemical markers indicative of eGC breakdown has been reported for patients with chronic kidney disease.³⁶ The lack of robust changes observed in the malaria patients may relate to the selection of the alternative vascular beds imaged (i.e., pinna and axilla). It is also possible that certain microvascular beds are preferentially impacted in malaria more so than others.³⁷ The axilla may be a more reliable measure than the pinna in malaria. Additionally, impaired blood flow as a result of RBC sequestration in malaria may reduce ability to accurately quantify changes with the Glycocheck technology, since the system relies on flowing RBC to quantify the PBR. This could be particularly problematic in SM in which more flow blockage than in HC or MSM subjects would lead to inaccurate imaging assessment of the eGC (and reduced ability to detect differences among the groups). Significant microvascular obstruction has been reported previously with orthogonal polarization spectral imaging of the rectal mucosa bed in adults with severe falciparum malaria.³⁷ More recently, using incident dark-field imaging, Lyimo and colleagues reported perivascular changes in the buccal microvascular bed in children with severe malaria, including stagnant RBC, microhemorrhages and, importantly, increased PBR.²¹ In our study, despite biochemical evidence for glycocalyx breakdown in both malaria severities, there was lack of evidence for increased PBR in severe malaria using the sidestream dark-field video imaging/analysis . Therefore, the incident dark field imaging technique used by Lyimo et al.²¹ may be an alternative glycocalyx imaging methodology for future studies in severe falciparum malaria. Incident dark field is a newer imaging technology similar to sidestream dark field, and this newer imaging technique uses a smaller device with a different method for illumination resulting in a larger field of view and higher resolution.³⁸

To assess the relationship of microvascular function and breakdown of eGC, we measured microvascular reactivity following vascular occlusion and tissue oxygen consumption. Previously, we reported decreased tissue perfusion and microvascular dysfunction in children with falciparum malaria^5,32 but with no significant difference between MSM and SM. In this study, microvascular reactivity was impaired to a greater degree in SM compared with MSM or HC, a result similar to findings in adult SM.³¹ There were no significant associations between measures of microvascular reactivity with markers of eGC integrity.

The mechanisms for eGC breakdown in malaria are not known, but multiple causal pathways have been hypothesized.^19,39 The Weibel-Palade Body product Ang-2 has been implicated in eGC breakdown in other disease settings.^22,33 Lukasz et al. reported Ang-2 induced breakdown of eGC that is heparanase-dependent in cultured endothelial cells and an in vivo mouse model³³. Ang-2 is released from endothelial cell Weibel-Palade bodies in response to inflammatory mediators, and is a ligand at the Tie2 receptor.⁴⁰ Elevation of Ang-2 is associated with endothelial activation and disease severity in malaria.^41,42 As found in adult¹⁹ and pediatric²¹ SM, we also noted a correlation of Ang-2 with the eGC breakdown products plasma syndecan-1 and urinary GAG, suggesting a possible role in degradation. In the present study, syndecan-1 and GAG had significant inverse associations with Hb and platelet count. This may relate to loss of glycocalyx from red blood cells and platelets in SM, or an increased propensity for cellular adhesion and/or decreased lifespan of erythrocytes and platelets following breakdown of eGC.

We did not find significant differences between patients with MSM and SM for most markers of disease severity, or for the markers of eGC breakdown syndecan-1 and urinary GAG. This may be related to marginal differences in clinical severity between the two groups. The MSM children were ill enough to require hospitalization. But while all of the SM children met WHO clinical criteria for SM, the majority had severity criteria with relatively low risk of mortality. There were some significant differences between the MSM and SM groups relative to parasitemia, Hb, and measures of microvascular function. However, other reliable indicators of disease severity (e.g., levels of platelets, lactate, and HRP-2) did not differ significantly between the two groups. Separate, earlier clinical studies by us and others have reported differences in degree of eGC loss related to disease severity in children²¹ and adults¹⁹ with falciparum malaria.

The DMMB assay was used to measure total sulfated GAG. While we did not analyze for specific types of GAG in the current study, our previous studies of Indonesian adults¹⁹ and Tanzanian children²⁰ with falciparum malaria, reported elevation of urinary GAG heparan sulfate (HS) and chondroitin sulfate (CS), while dermatan sulfate (DS) was not elevated. These findings provide support that the total sulfated GAG in urine of patients with malaria are from eGC since HS and CS are predominant forms of GAG found in eGC, while DS is not.^7,10 Since plasma syndecan-1, a prominent glycoprotein in eGC, was elevated and significantly correlated with GAG, this provides further evidence for an endothelial source for the elevated breakdown products.

The time course and sequence of events leading to eGC degradation in malaria are not known. When healthy adult volunteers were inoculated with Plasmodium falciparum, we noted that levels of osteoprotegerin, von Willebrand Factor, and angiopoietin-2 in plasma increased.³⁴ In that particular study, healthy volunteers were promptly treated with anti-malarial drugs upon the onset of symptoms and thus observations were limited to early events in the clinical course of infection. Despite elevation of these biomarkers of endothelial activation, there were no changes observed in markers of eGC breakdown (total urinary GAG or plasma syndecan-1), eGC sublingual videomicroscopy, or measures of microvascular dysfunction³⁴. These findings suggest that endothelial activation may occur early in infection contributing to downstream events such as eGC degradation and microvascular dysfunction. Early release of Ang-2 during malaria infection provides support for its involvement as a mediator of eGC breakdown.

While it is clear that breakdown of the eGC could adversely affect endothelial function through disrupted vascular homeostasis and cell signaling and increased cellular adhesion, it is also possible that circulating GAG could have detrimental effects of their own. Circulating fragments of GAG (e.g., heparan sulfate) have been reported to cross the blood-brain-barrier and mediate cognitive dysfunction in a mouse model of sepsis through binding to brain-derived neurotrophic factor (BDNF).⁴³ And cognitive impairment in patients who have sepsis is associated with presence of heparan sulfate with specific sulfation patterns in plasma upon admission. Heparan sulfate could have a similar contributing role to coma in pediatric CM, with plasma HS being significantly higher in children with CM having deeper coma.²¹

A limitation of this study is the age differential between the control and malaria groups, with the median age in the control group being 3 to 5 years higher than the median age in the SM and MSM groups, respectively. This factor could potentially contribute to differences between the control and malaria groups. In a study comparing different age ranges of children, urine GAG/ creatinine ratio decreased with increasing age, with children between 1-3 years having higher values.⁴⁴ The 1-3 year age range represents 22% of malaria vs. 5% of control subjects in our study. However this should not impact differences seen between the two malaria groups or the correlations reported for urinary GAG in malaria patients. Our study did not have a non-malaria fever comparison group, but it is well documented that glycocalyx degradation is not specific to malaria, with degradation also found in bacterial sepsis¹⁷, dengue¹⁸, and a variety of viral infections including COVID-19.⁴⁵

Adjunctive treatment that inhibits eGC degradation or enhances eGC synthesis may have benefit in malaria. However, identification of such agents and their role in diseases in which eGC is known to contribute is not well defined at this point, and there is no clinical evidence to date to support their clinical use. It is notable that sevuparin has recently been shown to inhibit cytoadherence and sequestration of infected RBC in falciparum malaria.^46,47 Structurally, sevuparin has close similarity to heparan sulfate, a major component of eGC, and thus could promote increased eGC.

In conclusion, our results provide further evidence for degradation of the eGC that could contribute to pathogenesis in malaria. Endothelial activation, increased adhesion of infected RBC and platelets to endothelial cells, and decreased nitric oxide formation are potential consequences of the eGC degradation and vascular dysfunction.

Nonstandard abbreviations:

Ang-2

angiopoietin 2

CS

chondroitin sulfate

DMMB

dimethylmethylene blue

DS

dermatan sulfate

eGC

endothelial glycocalyx

GAG

glycosaminoglycans

HC

healthy control

HKMU

Hubert Kairuki Medical University

HRP2

plasma histidine rich protein 2

HS

heparan sulfate

MSM

moderately severe malaria

NIRS

near infrared spectroscopy

NO

nitric oxide

PBR

perfused boundary region

REDCap

research electronic data capture

SM

severe malaria

StO2

tissue hemoglobin oxygen saturation

THI

hemoglobin perfusion index

WHO

World Health Organization