Available via license: CC BY 4.0
Content may be subject to copyright.
Diagnostics 2023, 13, 187. https://doi.org/10.3390/diagnostics13020187 www.mdpi.com/journal/diagnostics
Article
Clinical Value of Glycan Changes in Cerebrospinal Fluid for
Evaluation of Post-Neurosurgical Bacterial Meningitis with
Hemorrhagic Stroke Patients
Lei Ye
1,
*
,†
, Xuefei Ji
1,†
, Zijian Song
2
, Liao Guan
1
, Liang Zhao
1
, Wenwen Wang
3,4
and Weidong Du
3,
*
1
Department of Neurosurgery, The First Affiliated Hospital of Anhui Medical University, Jixi Road 218,
Hefei 230022, China
2
Department of Orthopaedics, Xuzhou Municipal First People’s Hospital, Daxue Road 269,
Xuzhou 221116, China
3
Department of Pathology, Anhui Medical University, Meishan Road 81, Hefei 230032, China
4
School of Clinical Medicine, Anhui Medical University, Meishan Road 81, Hefei 230032, China
* Correspondence: yelei@ahmu.edu.cn (L.Y.); weidong.du@ahmu.edu.cn (W.D.);
Tel.: +86-551-6292-2114 (L.Y.); +86-551-6516-1011 (W.D.);
Fax: +86-551-6363-3742 (L.Y.); +86-551-6516-5628 (W.D.)
† These authors contributed equally to this work.
Abstract: Post-neurosurgical bacterial meningitis (PNBM) is one of the severe complications in pa-
tients receiving neurosurgical procedures. Recent studies have found microbe-related glycans play
important roles in adhesion, invasion, and toxicity toward innate immunological reactions. In this
study, we aimed to investigate the glycomic profile and its potential diagnostic efficacy in post-
neurosurgical bacterial meningitis (PNBM) patients with hemorrhagic stroke. A total of 136 cere-
brospinal fluid (CSF) samples were recruited and divided into a PNBM group and a non-PNBM
group based on the clinical diagnostic criteria. A lectin biochip-based method was established for
the detection of glycans in CSF. The clinicopathological data and biochemical parameters in CSF
from all patients were analyzed. Two models for multivariate analysis investigating glycan changes
in the CSF were conducted, aiming at determining the specific expression and diagnostic efficacy
of lectin-probing glycans (LPGs) for PNBM. In univariate analysis, we found that 8 out of 11 LPGs
were significantly correlated with PNBM. Model 1 multivariate analysis revealed that PNA (p =
0.034), Jacalin (p = 0.034) and LTL (p = 0.001) were differentially expressed in the CSF of PNBM
patients compared with those of non-PNBM patients. Model 2 multivariate analysis further dis-
closed that LTL (p = 0.021) and CSF glucose (p < 0.001) had independent diagnostic efficacies in
PNBM, with areas under the curve (AUC) of 0.703 and 0.922, respectively. In summary, this study
provided a new insight into the subject of CSF glycomics concerning bacterial infection in patients
with hemorrhagic stroke.
Keywords: bacterial meningitis; neurosurgery; hemorrhagic stroke; glycans; biochip
1. Introduction
Post-neurosurgical bacterial meningitis (PNBM) is one of the severe complications
among patients receiving neurosurgical treatment for the diseases of the central nervous
system (CNS) [1]. Regardless of applications of advanced aseptic technology worldwide,
PNBM seems inevitable, with the incidence ranging from 0.3–10% in different neurosur-
gical diseases [2,3]. Due to the low positive rate in bacterial culture, the etiological diag-
nosis for PNBM has been a great challenge. Clinical decision-making for therapeutic ap-
plication of antibiotics in the disease is difficult. On the other hand, excessive neuroin-
flammatory reactions are commonly observed in neurological diseases, especially for
Citation: Ye, L.; Ji, X.; Song, Z.;
Guan, L.; Zhao, L.; Wang, W.; Du, W.
Clinical Value of Glycan Changes in
Cerebrospinal Fluid for
Evaluation of Post-Neurosurgical
Bacterial Meningitis with
Hemorrhagic Stroke Patients.
Diagnostics 2023, 13, 187.
https://doi.org/10.3390/
diagnostics13020187
Academic Editor: Ludmilla
Morozova-Roche
Received: 26 October 2022
Revised: 26 December 2022
Accepted: 29 December 2022
Published: 4 January 2023
Copyright: © 2023 by the authors. Li-
censee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and con-
ditions of the Creative Commons At-
tribution (CC BY) license (http://crea-
tivecommons.org/licenses/by/4.0/).
Diagnostics 2023, 13, 187 2 of 11
those receiving invasive therapies [4]. Therefore, biochemical, and immunological com-
ponents in cerebrospinal fluid (CSF) are often influenced by various confounding factors
that may be resulted from both primary neurological disease and surgical procedure. Re-
cent studies have proposed that the procalcitonin (PCT), lactate and neuron-specific eno-
lase in CSF were promising biomarkers for diagnosis of PNBM [5,6]. However, molecular
levels can be also influenced by physical and chemical changes of CSF, such as hemolysis,
which is a common phenomenon during the procedure of lumbar puncture [7]. Therefore,
novel biomarkers that can objectively and effectively reflect the infectious status in CSF
are urgently demanded in clinical practice.
Glycosylation is a common procedure that is post-translational modification (PTM)
in nature and is widely observed among mammals and microbes [8,9]. Glycan has differ-
ent structures and permutations conjugating proteins, lipids or other biological compo-
nents via numerous glycosyltransferases [10]. It has been reported that glycan contributes
important biological functions to cell adhesion, molecular trafficking, clearance, signaling
transduction and endocytosis [11]. In recent decades, increasing studies have focused on
both clinical diagnostic values and molecular mechanisms of glycan in different diseases,
especially in the infectious disease [12,13].
The bacterial surface is abundantly exposing glycans and glycoconjugates, which
provide unique antigens against innate immunological reaction for recognition and clear-
ance [14]. Importantly, the glycans or glycocojugates on the bacterial surface have been
demonstrated to have functions in bacterial motility and adhesive ability, consequently
influencing the virulence [15]. In addition, these glycans mediate the interactions of bac-
teria with the environment, towards both host and other bacteria, facilitating the coloni-
zation and survival [16]. In another aspect, bacteria also produces numerous glycosylated
structures and polysaccharides, including capsule polysaccharide (CPS), lipopolysaccha-
ride (LPS) and teichoic acid, which are involved in the immunological reaction and eva-
sion [17]. However, the profiling studies of bacterial glycan were only illustrated in some
pathological bacteria, such as Neisseria gonorrhoeae [18], Neisseria meningitides [16], Group
B Streptococcus [19], Campylobacter jejuni [20], and Burkholderia cepacia [15], as well as the
commensal bacteria Lactobacillus plantarum [21]. An evaluation of the association between
glycan level and bacterial infection in CSF not only helps to understand the invasion and
virulence properties but also provides a novel insight into diagnosis of infectious types
and even vaccine development.
In this study, we detected different glycan levels in CSF among PNBM and non-
PNBM patients receiving neurosurgery owing to hemorrhagic stroke by a well-estab-
lished lectin biochip. We aimed at discovering potential biomarkers for the diagnostic val-
ues in PNBM. The results may also provide a novel concept about the profile of glycan in
infectious CSF into development of glycoconjugate vaccines.
2. Materials and Methods
2.1. Patients and Sample Collection
A total of 136 patients were randomly recruited in this study. All patients were diag-
nosed with hemorrhagic stroke by two senior doctors and received neurosurgical treat-
ments. CSF samples were collected via lumbar cistern drainage or lumbar puncture with
aseptic technique and were stored at −80 °C. The study was conducted according to the
Declaration of Helsinki and was approved by the Institutional Ethics Board of the First
Affiliated Hospital of Anhui Medical University. Informed consent was obtained from all
participants or the appropriate relatives.
2.2. Sample Treatment
Isolation of high-abundant proteins in CSF was introduced in our previous study
[22]. According to the manufacturer’s protocol, high-abundant proteins, including albu-
min and IgG proteins, were filtrated with a Proteo Prep Immunoaffinity Albumin and IgG
Diagnostics 2023, 13, 187 3 of 11
Depletion Kit. After that, the CSF samples were labeled with Cy3 using the protocol that
was described elsewhere [22]. Redundant Cy3 was removed with a PD MiniTrap G-25
column (GE, Massachusetts, USA).
2.3. Profiling Glycosylation of Sera in ICH
The protocol for chemical modification on biochip was described according to our
previous study [22]. Lectins, including Wheat Germ Agglutinin (WGA, L-1020), Lens
Culinaris Agglutinin (LCA, L-1040), Peanut Agglutinin (PNA, L-1070), Ricinus Com-
munis Agglutinin I (RCA-I, L-1080), Jacalin (L-1150), Vicia Villosa Lectin (VVL, L-1230),
Sambucus Nigra Lectin (SNA, L-1300), Maackia Amurensis Lectin I (MAL-I, L-1310), Lo-
tus Tetragonolobus Lectin (LTL, L-1320), Narcissus Pseudonarcissus Lectin (NPL, L-1370),
and Phaseolus Vulgaris Leucoagglutinin (PVL, L-1110), were commercially obtained from
Vector Laboratories Inc. (Newark, California, USA).
Each lectin was resolved in 10 mM HEPES buffer (pH 8.5) containing 0.001% BSA to
the concentration of 1 mg/mL and then was immobilized on individual spot on the biochip
surface, incubating at room temperature (RT) for 2 h to have robust conjugations with the
surface. The biochips were then washed with a 0.01 M PBST buffer (pH 7.4) and dried
with a nitrogen flow. The Cy3-labelled sera from 53 PNBM patients and 83 non-PNBM
controls were individually supplied on the lectin-probed biochips and incubated in a hu-
midity chamber and dark environment at RT for 0.5 h. After rinsed in PBST buffer twice
at RT for 3 min and dried with a flow of nitrogen, the biochips were scanned with a mi-
croarray scanner (LuxscanTM 10K-A microscanner, Capitalbio Co., Ltd., Beijing, China).
The fluorescence intensities on the biochips were recorded.
2.4. Statistical Analysis
Statistical analyses were performed using SPSS 19.0 software. All continuous data
were summarized as the mean ± standard deviation (Mean ± SD) or the median with in-
terquartile range (IQR). Univariate analysis was performed using Student’s t-test or the
Mann–Whitney U test. The dependent variables that were statistically significant between
PNBM and non-PNBM cohort in the univariate analysis were further analyzed with mul-
tivariate analysis. Two models of multivariate analyses were conducted using lineage re-
gression analysis. Model 1 investigated the glycan changes that were differentially ex-
pressed in the CSF of PNBM patients. Model 2 analyzed all the lectin-probing glycans
(LPGs) and biochemical parameters in the CSF to discover independent risks for PNBM.
B value in multivariate analysis represents the partial regression coefficient of arguments
in the equation of regression. The negative value of B represents that the argument has a
negative effect on the dependent variable. Binary data were analyzed using the chi-
squared test. The p values reported in the study were two-sided and p < 0.05 was consid-
ered significant.
3. Results
3.1. Characteristics of PNBM Patients
This study recruited 136 patients with hemorrhagic stroke of whom 62 were diag-
nosed with intracerebral hemorrhage (ICH) and 74 were diagnosed with aneurysmal sub-
arachnoid hemorrhage (aSAH). PNBM was diagnosed in 53 out of the 136 patients based
on the diagnostic criteria issued by IDSA’s Clinical Practice Guidelines for Healthcare-
Associated Ventriculitis and Meningitis 2017 [23] and a Chinese Expert Consensus of Di-
agnostic and Therapy for the Neurosurgical Central Nervous System Infections in 2021.
Demographic and clinicopathological features are summarized in Table 1. Six patients had
positive results for bacterial cultures, accounting for 11.3% (6/53 cases), three of whom
were infected with Stenotrophomonas maltophilia, Moderate thermophiles, and Streptococcus
agalactiae, respectively. Two patients were infected with Acinetobacter baumannii, and a pa-
tient was jointly infected with the bacteria of Pseudomonas aeruginosa and Aeromonas caviae.
Diagnostics 2023, 13, 187 4 of 11
The remaining 47 patients were diagnosed with PNBM according to the biochemical char-
acters of CSF.
Table 1. Clinicopathological characteristics of patients recruited in this study. (PNBM: post-neuro-
surgical bacterial meningitis; CSF: cerebrospinal fluid; cGlu: CSF glucose; bGlu: blood glucose; ICP:
intracranial pressure; aSAH: aneurysmal subarachnoid hemorrhage; ICH: intracerebral hemor-
rhage; IQR: interquartile range).
PNBM
(n = 53)
Non-PNBM
(n = 83) p Value
Age (years, mean ± SD) 55.70 ± 17.09 57.65 ± 15.40 0.491
Gender 0.568
Male 28 45
Female 25 38
CSF
Glucose (mmol/L, IQR) 1.59 (1.26, 1.95) 3.47 (2.90, 4.33) <0.001
Protein (g/L, IQR) 3.00 (2.00, 4.35) 1.00 (0.63, 1.90) <0.001
White blood cells (×106/L, IQR) 972 (345, 4768) 64 (18, 291) <0.001
Red blood cells (×106/L, IQR) 14,000 (550, 94500) 5000 (500, 16,000) 0.038
Proportion of multinuclear cell
(IQR) 78.2 (58.6, 88.9) 46.2 (20.0, 79.1) <0.001
Chlorine (mmol/L, mean ± SD) 120.30 ± 10.63 124.13 ± 8.94 0.025
Pandy tests (negative/positive) 12/41 22/61 0.612
Blood glucose(mmol/L,
IQR) 6.43 (5.43, 7.55) 7.08 (5.72, 9.22) 0.296
cGlu/bGlu ratio (IQR) 0.23 (0.16, 0.33) 0.51 (0.39, 0.64) <0.001
ICP (mmH2O, IQR) 180 (115, 260) 153 (110, 210) 0.595
Primary disease 0.264
aSAH 32 42
ICH 21 41
3.2. Glycosylation Profile of CSF
We probed 11 lectins on the biochip surface to detect the differential expressions of
glycans in the CSF. We found 8 out of 11 LPG levels were significantly higher in PNBM
group than in non-PNBM group by the univariate analysis. They were WGA (p = 0.002),
LCA (p = 0.002), PNA (p = 0.002), RCA-I (p = 0.027), Jacalin (p < 0.001), SNA (p = 0.009),
MAL-I (p = 0.048) and LTL (p < 0.001). The expressions of the remaining LPGs did not
reveal statistical significances between the two groups, but two revealed marginal p val-
ues, NPL (p = 0.073) and PVL (p = 0.086) (Figure 1). We conducted multivariate analyses
of two models. In Model 1, we analyzed only the LPGs that were significantly different
between PNBM and non-PNBM patients. In this model, we found that PNA (p = 0.034),
Jacalin (p = 0.034) and LTL (p = 0.001) in CSF revealed significantly differences between
the two groups. In Model 2, biochemical parameters and LPGs in the CSF that were dif-
ferentially expressed between PNBM and non-PNBM groups in univariate analysis were
further analyzed. The results showed that LTL (p = 0.021) and CSF glucose (p < 0.001) had
statistical significances for independently distinguishing PNBM from non-PNBM (Table
2).
Diagnostics 2023, 13, 187 5 of 11
Figure 1. A representative fluorescence detection (A) and quantification comparison (B) of 11 kinds
of lectin-probing glycans between 53 PNBM patients and 83 non-PNBM controls. Samples 1–8 be-
long to PNBM patients. PBS was incubated on the lectin-probed biochip instead of serum samples
as negative control. (WGA: Wheat Germ Agglutinin; LCA: Lens Culinaris Agglutinin; PNA: Peanut
Agglutinin; RCAI: Ricinus Communis Agglutinin I; VVL: Vicia Villosa Lectin; SNA: Sambucus
Nigra Lectin; MAL-I: Maackia Amurensis Lectin I; LTL: Lotus Tetragonolobus Lectin; NPL: Narcis-
sus Pseudonarcissus Lectin; PVL: Phaseolus Vulgaris Leucoagglutinin; PNBM: post-neurosurgical
bacterial meningitis. * p < 0.05; ** p < 0.01).
Table 2. Multivariate analysis of lectin probing glycans and biochemical characteristics in PNBM.
Model 1 of multivariate analysis included all lectin probing glycans; Model 2 of multivariate analy-
sis was involved in combination of all the lectin-probing glycans and diagnostic characteristics for
PNBM. The B value represents the partial regression coefficient of arguments in the equation of
regression. (WGA: Wheat Germ Agglutinin; LCA: Lens Culinaris Agglutinin; PNA: Peanut Agglu-
tinin; RCA-I: Ricinus Communis Agglutinin I; SNA: Sambucus Nigra Lectin; MAL-I: Maackia
Amurensis Lectin I; LTL: Lotus Tetragonolobus Lectin; CSF: cerebrospinal fluid; cGlu: CSF glucose;
bGlu: blood glucose).
Model 1 Model 2
B p Value B p Value
WGA −0.034 0.830 −0.107 0.409
LCA 0.132 0.423 0.130 0.333
Diagnostics 2023, 13, 187 6 of 11
PNA −0.356 0.034 −0.074 0.606
RCA-I 0.078 0.676 0.053 0.729
Jacalin −0.200 0.034 −0.128 0.096
SNA −0.129 0.431 −0.960 0.468
MAL-I 0.189 0.220 0.032 0.795
LTL −0.324 0.001 −0.200 0.021
CSF glucose - - 0.394 <0.001
CSF protein - - −0.108 0.198
CSF white blood cells - - −0.106 0.207
CSF red blood cells - - −0.036 0.661
CSF proportion of multinu-
clear cell - - −0.097 0.199
CSF chlorine - - 0.100 0.146
cGlu/bGlu ratio - - 0.107 0.306
3.3. Diagnostic Values for PNBM
Based on the data from the Model 2 multivariate analysis, we found LTL and glucose
in CSF were the independently hazard risks for PNBM. Therefore, we performed a re-
ceiver operator characteristic (ROC) analysis for further evaluating the diagnostic values
of LTL and CSF glucose between the PNBM group and non-PNBM group. The area under
the curve (AUC) was 0.703 for LTL, with sensitivity of 54.2% and specificity of 84.9% (Fig-
ure 2A), and for CSF glucose was 0.922, with sensitivity of 98.8% and specificity of 88.7%
(Figure 2B).
Figure 2. Receiver operating curve analysis for evaluating clinical values of Lotus Tetragonolobus
Lectin (LTL) (A) and glucose (B) in cerebrospinal fluid in patients with post-neurosurgical bacterial
meningitis.
3.4. Correlation Analyses for Biochemical Parameters of CSF
We performed correlation analyses between the biochemical parameters of CSF and
LPGs in patients with PNBM (Figure 3A) and without PNBM (Figure 3B), respectively.
We found some moderate-to-strong correlations (R > 0.7) within LPGs in both groups,
indicating that there were interactions among the glycans regardless of infection status.
However, there were few correlations between the CSF glucose and LPGs in both the
groups, indicating that glucose might have had minor effects on glycans. Interestingly, we
found some moderate correlations of LPGs with CSF protein in patients with PNBM, ra-
ther than those without PNBM (for example R = 0.511 vs. 0.163 for PNA; R = 0.435 vs. 0.184
Diagnostics 2023, 13, 187 7 of 11
for RCA-I; R = 0.322 vs. 0.083 for VVL R = 0.295 vs. 0.105 for SNA; R = 0.298 vs. 0.100 for
MAL; R = 0.461 vs. 0.043 for NPL; R = 0.324 vs. 0.077 for PVL). This result provided evi-
dence that the CSF glycans in the PNBM patients might be derived from the glycoproteins.
Figure 3. Correlation analyses between biochemical parameters and lectin-probing glycans of cere-
brospinal fluid (CSF) in 53 patients with post-neurosurgical bacterial meningitis (PNBM) (A) and 83
patients without PNBM (B). (WGA: Wheat Germ Agglutinin; LCA: Lens Culinaris Agglutinin; PNA:
Peanut Agglutinin; RCA-I: Ricinus Communis Agglutinin I; VVL: Vicia Villosa Lectin; SNA: Sam-
bucus Nigra Lectin; MAL-I: Maackia Amurensis Lectin I; LTL: Lotus Tetragonolobus Lectin; NPL:
Narcissus Pseudonarcissus Lectin; PVL: Phaseolus Vulgaris Leucoagglutinin; PNBM: post-neuro-
surgical bacterial meningitis; c/b Glu: the ratio of CSF and blood glucose; cRBC: CSF red blood cell;
bGlu: blood glucose; cCl: CSF chlorine; cMNC: CSF proportion of multinuclear cell; cWBC: CSF
white blood cell; cPro: CSF total protein; cGlu: CSF glucose).
4. Discussion
In this study, we profiled glycan levels in CSF in patients with hemorrhagic stroke
via a well-established lectin biochip. Our results indicated that PNA, Jacalin and LTL-
probing glycans were differentially expressed in the CSF of PNBM patients, rather than
that of non-PNBM patients. In addition, we found that LTL-probing glycans and CSF glu-
cose might have independently diagnostic efficacies in PNBM. This study also provided
a new concept about the glycomic profiles of nosocomial bacterial infection in CSF regard-
less of bacterial types.
Glycosylation is an important post-transcriptional modification, functioning in cell–
cell adhesion, protein folding and protein trafficking. Glycan covers almost all surfaces of
cells and microbes [24]. Recent studies have demonstrated that glycans or glycoconjugates
might serve as biomarkers in diagnosis of multiple diseases, especially in infectious dis-
eases. Tang et al. [25] found that N-acetyl-d-lactosamine was mainly increased on macro-
phages during virulent Mycobacterium tuberculosis infection and could serve as a potential
novel diagnostic and therapeutic biomarker. Moreover, Wang et al. [26] demonstrated
that O-glycosylation of CtxB-BCAL2737a could be used to detect anti-O-glycan antibodies
in human serum samples from patients with Burkholderia-associated infections.
The composition of bacterial cell wall varies between taxonomic groups and species,
depending on the cell type and the developmental stage. Bacterial surface is enriched in
glycocojugates. Molecular structures of glycan on some bacterial cell walls vary from
those of glycan produced by humans [24]. Previous studies demonstrated that these gly-
cans or glycoconjugates had significant interactions with their environments, including
Diagnostics 2023, 13, 187 8 of 11
the host and the commensal bacteria. Lectin, which could preferentially combine gly-
coconjugates, such as glycolipids or glycoproteins with a specific pattern as antigen-anti-
body interaction, has been reported to be expressed at innate immunocytes. These glycans
represent an important class of pattern recognition receptors (PRRs) and have the ability
of carbohydrate recognition [27]. The initial recognition and combination between lectin
of innate immunocytes and glycans on bacterial surface provide early host defense. Fur-
thermore, bacterial glycans or glycoconjugates significantly correlate with virulence. Mu-
baiwa et al. reported that in the processes of Neisseria spp. Infection, such as N. meningitis
and N. gonorrhoeae, glycans involved in all stages of colonization and progression [16]. On
the other hand, the bacterial glycoconjugates have also been demonstrated to be involved
in immune evasion. Some Klebsiella serotypes averted the recognition by host lectin via
altering sugars, such as mannobiose and rhamnobiose in their capsule, leading to a de-
creasing response of phagocytes in the respiratory tract [28].
In our study, we used two models of multiple regression analysis to investigate the
differentially expressed LPGs in CSF. In Model 1, we found 3 LPGs, PNA-probing, Jacalin-
probing and LTL-probing glycans, Galβ3GalNAc, Galβ3GalNAc-Ser/Thr and α-Fucose,
were differentially expressed in the CSF between PNBM and non-PNBM cohorts. These
results reflected the differentially expressed molecules concerning the infection, regard-
less of diagnostic efficacy. In previous studies, fucosylation of bacterial related proteins or
lipids, which LTL was conjugated, has been widely investigated in the bacterial infections.
Vimonish et al. [29] found that Anaplasma marginale regulated the differential expression
of tick α-(1,3)-fucosyltransferases in midgut cells, indicating that the pathogen utilized
core α-(1,3)-fucose of N-glycan to infect tick midgut cells. Sun et al. [30] observed a Salmo-
nella neuraminidase-associated compositional shift of macrophage glycocalyx using nano-
LC-chip QTOF mass cytometry technology. The shift led to an increase in fucosylation,
and a subsequently impairment for both macrophage phagocytosis and galvanotaxis.
However, we did find any direct evidence for the association of α-fucose or fucosylation
with bacterial infections in CNS. In addition, a previous report indicated that Galβ3Gal-
NAc, which PNA conjugated to, had an intensive interaction with cell wall-glyceralde-
hyde-3-phosphate dehydrogenases in Lactobacillus reuteri [31]. Although we failed to find
the involvement of Jacalin-probing Galβ3GalNAc-Ser/Thr in bacterial infections in
PNBM, our data supported the hypothesis that the glycan profiling would potentially re-
flect PNBM status.
The information about the infection-related glycomics in CSF is vital for development
of glycoconjugate vaccines and antibody-based drugs. As early as 1923, capsular polysac-
charide (CPS) of Streptococcus pneumoniae was found to be reactive with anti-pneumococ-
cus sera and then it was developed as a hexavalent CPS vaccine, which received a brief
license by US Food and Drug Administration [32]. So far, approximately 10 glycan-related
vaccines have been licensed for use in US, offering protection strategies against numerous
pathogenic strains of Gram negative bacterium [33]. Furthermore, in recent years, mono-
clonal antibody (mAb) has been developed as a biological tool for clinical diagnosis and
treatment of infectious diseases, especially when anti-microbial resistant occurs. Several
mAb drugs have been in phase I or II clinical trials, such as F598 for N. gonorrhoeae target-
ing poly-N-acetylglucosamine [34], AR-105/Aerucin for P. aeruginosa targeting alginate
[35], and DSTA4637S for S. aureus targeting β-WTA of S. aureus [36]. Even though we are
not clear about the specific infectious bacterial types in this study, our results provide a
new insight into concept for PNBM. Some glycans could be applied in the development
of universal vaccines toward the patients with neurosurgical treatments for prevention of
bacterial infection.
In Model 2, the biochemical characters of CSF, which were related to the diagnostic
criteria of meningitis, were included. The results indicated that LTL-probing glycans and
glucose in CSF might be served as independently diagnostic biomarkers for PNBM. Clin-
ical decision-making for PNBM is mainly based on the bacterial culture and biochemical
Diagnostics 2023, 13, 187 9 of 11
factors in CSF. However, low positive rate in conventional bacterial culture and the com-
plicated CSF environment in neurosurgical patients make the diagnosis of PNBM perplex-
ing. We noticed that polymerase chain reaction (PCR) and next generation sequencing
(NGS) can also detect the microbial species with a higher sensitivity. However, we did not
apply these techniques for detection of the infectious bacteria in CSF. Because of trace
abundance of bacterial load, the host DNA may disturb the PCR amplification of bacterial
DNA, thus leading to a complex background. On the other hand, due to the high-sensi-
tivity of NGS, multiple microbes can be detected, including commensal microbes and
pathogenic microbes. Moreover, because of the microbial interactions, we could not dis-
tinguish which microbe contributes to the infection. CSF glucose, a monosaccharide, is
routinely tested as one of the well-established biomarkers estimating CNS infections but
is less served as an independent application for the diagnosis of PNBM. In this study, we
found that both glucose and LTL-probing glycan in CSF would significantly be the inde-
pendent biomarkers for PNBM, indicating that the glycometabolism might play a key role
in the pathogenesis of PNBM. However, whether a combinational test of LTL-probing
glycan and glucose in CSF could be served as a novel criterion for the independent diag-
nosis of PNBM should be further validated in an augmented population.
Some limitations in the study should be carefully considered. First, although 6 out of
53 patients showed positive bacterial culture, it was too few to do statistical analysis for
individual microbial stains. Moreover, it was not clear what was the detailed pathogenic
microbes in the remaining 47 patients negative with PNBM according to the biochemical
characters of CSF, thus it is improper to make the comparison of fluorescent intensity be-
tween the culture-positive and culture-negative samples. Therefore, the results of the gly-
can profile in this study did not reflect relationships to a certain bacterial type but in gen-
eral were the universal biomarkers for bacterial infection. Bacteria-specific glycan profil-
ing analysis will be expected when enough CSF samples with definite information about
bacterial types are recruited, and second, the LPGs in our study referred to the molecules
that contained structures of specific glycans. However, we did not know about detailed
information about the molecules. As indicated from the result of correlation analysis of
the LPGs with biochemical characters of CSF, a combination study of proteomics and gly-
comics for CSF should be conducted to investigate the functional molecules in the patho-
genesis of PNBM.
In summary, this study provided new information in CSF glycomics concerning bac-
terial infection in patients with PNBM, which would be of potential value in development
of glyco-conjugated vaccines in prevention of meningitis. Additionally, we observed that
the LTL-probing glycan and glucose in CSF were two independent risk factors for clinical
diagnosis of PNBM.
Author Contributions: Conceptualization, L.Y. and W.D.; methodology, X.J. and L.Z.; resources
and visualization, X.J., L.G. and W.W.; software and formal analysis, Z.S. and L.G.; supervision, L.Y.
and W.D.; funding acquisition, L.Y. and W.D.; administration, W.D.; writing—review and editing,
L.Y. and W.D. All authors approved the critical revision of the manuscript for important intellectual
content. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the National Natural Science Foundation of China
(81901238), Project of Institute of Translational Medicine of Anhui Province (2017zhyx37), Scientific
Foundation of Anhui Medical University (2022xjk207) and National Natural Science Foundation of
Anhui Province (2208085MH224).
Institutional Review Board Statement: The study was conducted according to the Declaration of
Helsinki and was approved by the Institutional Ethics Board of the First Affiliated Hospital of Anhui
Medical University.
Informed Consent Statement: Informed consent was obtained from all individual participants or
the appropriate relatives.
Data Availability Statement: The data that support the findings of this study are available from the
corresponding author upon reasonable request.
Diagnostics 2023, 13, 187 10 of 11
Conflicts of Interest: The authors declared no conflict of interest.
References
1. Hussein, K.; Bitterman, R.; Shofty, B.; Paul, M.; Neuberger, A. Management of post-neurosurgical meningitis: Narrative review.
Clin. Microbiol. Infect. 2017, 23, 621–628. https://doi.org/10.1016/j.cmi.2017.05.013.
2. Blomstedt, G.C. Infections in neurosurgery: A retrospective study of 1143 patients and 1517 operations. Acta Neurochir. 1985,
78, 81–90.
3. Zhang, Y.; Xiao, X.; Zhang, J.; Gao, Z.; Ji, N.; Zhang, L. Diagnostic accuracy of routine blood examinations and CSF lactate level
for post-neurosurgical bacterial meningitis. Int. J. Infect. Dis. 2017, 59, 50–54.
4. Yin, L.; Han, Y.; Miao, G.; Jiang, L.; Xie, S.; Liu, B. CSF leukocyte, polykaryocyte, protein and glucose: Their cut-offs of judging
whether post-neurosurgical bacterial meningitis has been cured. Clin. Neurol. Neurosurg. 2018, 174, 198–202.
5. Alons, I.M.; Verheul, R.J.; Kuipers, I.; Jellema, K.; Wermer, M.J.; Algra, A.; Ponjee, G. Procalcitonin in cerebrospinal fluid in
meningitis: A prospective diagnostic study. Brain. Behav. 2016, 6, e00545.
6. Lotfi, R.; Ines, B.; Aziz, D.M.; Mohamed, B. Cerebrospinal Fluid Lactate as an Indicator for Post-neurosurgical Bacterial Menin-
gitis. Indian J. Crit. Care Med. 2019, 23, 127–130.
7. Lippi, G.; Salvagno, G.L.; Montagnana, M.; Brocco, G.; Guidi, G.C. Influence of hemolysis on routine clinical chemistry testing.
Clin. Chem. Lab. Med. 2006, 44, 311–316.
8. Schjoldager, K.T.; Narimatsu, Y.; Joshi, H.J.; Clausen, H. Global view of human protein glycosylation pathways and functions.
Nat. Rev. Mol. Cell. Biol. 2020, 21, 729–749.
9. Trempel, F.; Kajiura, H.; Ranf, S.; Grimmer, J.; Westphal, L.; Zipfel, C.; Scheel, D.; Fujiyama, K.; Lee, J. Altered glycosylation of
exported proteins, including surface immune receptors, compromises calcium and downstream signaling responses to microbe-
associated molecular patterns in Arabidopsis thaliana. BMC Plant Biol. 2016, 16, 31.
10. Borud, B.; Barnes, G.K.; Brynildsrud, O.B.; Fritzsonn, E.; Caugant, D.A. Genotypic and Phenotypic Characterization of the O-
Linked Protein Glycosylation System Reveals High Glycan Diversity in Paired Meningococcal Carriage Isolates. J. Bacteriol.
2018, 200, e00794-17.
11. Deng, W.; Cao, J.; Chen, L.; McMullin, D.; Januzzi, J.L.; Jr., Buonanno, F.S.; Lo, E.H.; Ning, M.M. Plasma Glycoproteomic Study
of Therapeutic Hypothermia Reveals Novel Markers Predicting Neurologic Outcome Post-cardiac Arrest. Transl. Stroke Res.
2018, 9, 64–73.
12. De, P.; Amin, A.G.; Flores, D.; Simpson, A.; Dobos, K.; Chatterjee, D. Structural implications of lipoarabinomannan glycans
from global clinical isolates in diagnosis of Mycobacterium tuberculosis infection. J. Biol. Chem. 2021, 297, 101265.
13. Wu, Z.L.; Ertelt, J.M. Fluorescent glycan fingerprinting of SARS2 spike proteins. Sci. Rep. 2021, 11, 20428.
14. Swamy, M.; Pathak, S.; Grzes, K.M.; Damerow, S.; Sinclair, L.V.; van Aalten, D.M.; Cantrell, D.A. Glucose and glutamine fuel
protein O-GlcNAcylation to control T cell self-renewal and malignancy. Nat. Immunol. 2016, 17, 712–720.
15. Lithgow, K.V.; Scott, N.E.; Iwashkiw, J.A.; Thomson, E.L.; Foster, L.J.; Feldman, M.F.; Dennis, J.J. A general protein O-glycosyl-
ation system within the Burkholderia cepacia complex is involved in motility and virulence. Mol. Microbiol. 2014, 92, 116–137.
16. Mubaiwa, T.D.; Semchenko, E.A.; Hartley-Tassell, L.E.; Day, C.J.; Jennings, M.P.; Seib, K.L. The sweet side of the pathogenic
Neisseria: The role of glycan interactions in colonisation and disease. Pathog. Dis. 2017, 75, ftx063.
17. Tan, F.Y.; Tang, C.M.; Exley, R.M. Sugar coating: Bacterial protein glycosylation and host-microbe interactions. Trends Biochem.
Sci. 2015, 40, 342–350.
18. Vik, A.; Aas, F.E.; Anonsen, J.H.; Bilsborough, S.; Schneider, A.; Egge-Jacobsen, W.; Koomey, M. Broad spectrum O-linked pro-
tein glycosylation in the human pathogen Neisseria gonorrhoeae. Proc. Natl. Acad. Sci. USA 2009, 106, 4447–4452.
19. van Sorge, N.M.; Quach, D.; Gurney, M.A.; Sullam, P.M.; Nizet, V.; Doran, K.S. The group B streptococcal serine-rich repeat 1
glycoprotein mediates penetration of the blood-brain barrier. J. Infect. Dis. 2009, 199, 1479–1487.
20. Nothaft, H.; Szymanski, C.M. Bacterial protein N-glycosylation: New perspectives and applications. J. Biol. Chem. 2013, 288,
6912–6920.
21. Fredriksen, L.; Moen, A.; Adzhubei, A.A.; Mathiesen, G.; Eijsink, V.G.; Egge-Jacobsen, W. Lactobacillus plantarum WCFS1 O-
linked protein glycosylation: An extended spectrum of target proteins and modification sites detected by mass spectrometry.
Glycobiology 2013, 23, 1439–1451.
22. Tunkel, A.R.; Hasbun, R.; Bhimraj, A.; Byers, K.; Kaplan, S.L.; Scheld, W.M.; van de Beek, D.; Bleck, T.P.; Garton, H.J.L.; Zunt,
J.R. 2017 Infectious Diseases Society of America's Clinical Practice Guidelines for Healthcare-Associated Ventriculitis and Men-
ingitis. Clin. Infect. Dis. 2017, 64, e34–e65.
23. Ye, L.; Fang, Y.S.; Li, X.X.; Gao, Y.; Liu, S.S.; Chen, Q.; Wu, Q.; Cheng, H.W.; Du, W.D. A simple lectin-based biochip might
display the potential clinical value of glycomics in patients with spontaneous intracerebral hemorrhage. Ann. Transl. Med. 2021,
9, 544.
24. Imperiali, B. Bacterial carbohydrate diversity—A Brave New World. Curr. Opin. Chem. Biol. 2019, 53, 1–8.
25. Tang, X.L.; Yuan, C.H.; Ding, Q.; Zhou, Y.; Pan, Q.; Zhang, X.L. Selection and identification of specific glycoproteins and glycan
biomarkers of macrophages involved in Mycobacterium tuberculosis infection. Tuberculosis 2017, 104, 95–106.
Diagnostics 2023, 13, 187 11 of 11
26. Wang, G.; Glaser, L.; Scott, N.E.; Fathy Mohamed, Y.; Ingram, R.; Laroucau, K.; Valvano, M.A. A glycoengineered antigen ex-
ploiting a conserved protein O-glycosylation pathway in the Burkholderia genus for detection of glanders infections. Virulence
2021, 12, 493–506.
27. Cao, X. Self-regulation and cross-regulation of pattern-recognition receptor signalling in health and disease. Nat. Rev. Immunol.
2016, 16, 35–50.
28. Sahly, H.; Keisari, Y.; Ofek, I. Manno(rhamno)biose-containing capsular polysaccharides of Klebsiella pneumoniae enhance
opsono-stimulation of human polymorphonuclear leukocytes. J. Innate Immun. 2009, 1, 136–144.
29. Vimonish, R.; Capelli-Peixoto, J.; Johnson, W.C.; Hussein, H.E.; Taus, N.S.; Brayton, K.A.; Munderloh, U.G.; Noh, S.M.; Ueti,
M.W. Anaplasma marginale Infection of Dermacentor andersoni Primary Midgut Cell Culture Is Dependent on Fucosylated
Glycans. Front. Cell. Infect. Microbiol. 2022, 12, 877525.
30. Sun, Y.H.; Luxardi, G.; Xu, G.; Zhu, K.; Reid, B.; Guo, B.P.; Lebrilla, C.B.; Maverakis, E.; Zhao, M. Surface Glycans Regulate
Salmonella Infection-Dependent Directional Switch in Macrophage Galvanotaxis Independent of NanH. Infect. Immun. 2022, 90,
e0051621.
31. Deng, Z.; Dai, T.; Zhang, W.; Zhu, J.; Luo, X.M.; Fu, D.; Liu, J.; Wang, H. Glyceraldehyde-3-Phosphate Dehydrogenase Increases
the Adhesion of Lactobacillus reuteri to Host Mucin to Enhance Probiotic Effects. Int. J. Mol. Sci. 2020, 21, 9756.
32. Avery, O.T.; Heidelberger, M. Immunological Relationships of Cell Constituents of Pneumococcus. J. Exp. Med. 1923, 38, 81–85.
33. Temme, J.S.; Butler, D.L.; Gildersleeve, J.C. Anti-glycan antibodies: Roles in human disease. Biochem. J. 2021, 478, 1485–1509.
34. Soliman, C.; Walduck, A.K.; Yuriev, E.; Richards, J.S.; Cywes-Bentley, C.; Pier, G.B.; Ramsland, P.A. Structural basis for antibody
targeting of the broadly expressed microbial polysaccharide poly-N-acetylglucosamine. J. Biol. Chem. 2018, 293, 5079–5089.
35. Pier, G.B.; Boyer, D.; Preston, M.; Coleman, F.T.; Llosa, N.; Mueschenborn-Koglin, S.; Theilacker, C.; Goldenberg, H.; Uchin, J.;
Priebe, G.P.; et al. Human monoclonal antibodies to Pseudomonas aeruginosa alginate that protect against infection by both
mucoid and nonmucoid strains. J. Immunol. 2004, 173, 5671–5678.
36. Peck, M.; Rothenberg, M.E.; Deng, R.; Lewin-Koh, N.; She, G.; Kamath, A.V.; Carrasco-Triguero, M.; Saad, O.; Castro, A.; Teufel,
L.; et al. A Phase 1, Randomized, Single-Ascending-Dose Study to Investigate the Safety, Tolerability, and Pharmacokinetics of
DSTA4637S, an Anti-Staphylococcus aureus Thiomab Antibody-Antibiotic Conjugate, in Healthy Volunteers. Antimicrob.
Agents Chemother. 2019, 63, e02588-18.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual au-
thor(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.
