Keywords
platelets - complement activation - lectin - systemic lupus erythematosus - antiphospholipid
syndrome
Introduction
Systemic lupus erythematosus (SLE) and antiphospholipid syndrome (APS) are chronic
autoimmune diseases. APS is characterized by venous and/or arterial thromboses, or
placenta-mediated pregnancy complications, combined with the persistent presence of
antiphospholipid antibodies (aPLs).[1] APS can be primary and secondary. Secondary APS is related to autoimmune diseases
like SLE.[2] Among patients with SLE, 30 to 40% have aPLs, and 10 to 15% have clinical APS.[3] SLE primarily affects young women. It is a heterogeneous disease with manifestations
in multiple organ systems, including the kidneys, skin, central nervous system, lungs,
joints, and cardiovascular system.[4]
Thromboembolic disease is one of the leading causes of death among patients suffering
from SLE or APS.[5]
[6] Patients suffering from SLE develop thromboembolic disease earlier and more frequently
than the general population. They have a three to four times higher risk of developing
venous thromboembolism than age- and gender-matched healthy individuals.[7] In a cohort of 1,000 APS patients from 13 different countries, 31.7% experienced
deep vein thrombosis, 9.1% pulmonary emboli, and 13.1% experienced stroke.[3]
The increased thromboembolic risk in SLE and APS is strongly associated with the presence
of aPL.[8] However, despite the strong association, the mechanism of aPL-mediated thrombosis
has not been fully elucidated.[9] New research suggests a multifactorial mechanism of thrombosis, and experimental
animal models and in vitro studies indicate the involvement of multiple cellular and
inflammatory systems, including platelets and the complement system.[10]
[11]
Complement activation occurs through three distinct pathways, the classical, the lectin,
and the alternative pathway, ultimately leading to the cleavage of C3 to bioactive
C3a and C3b. This initiates the common terminal pathway, leading to the membrane attack
complex.[12] Activation of the complement system leads to inflammation, opsonization, phagocytosis,
and adaptive immune system activation.[13] Although normally well-regulated, the system can be both inefficient, contributing
to susceptibility to infections, and overstimulated, furthering autoimmunity, inflammation,
and possibly thrombosis.[13]
The lectin pathway of the complement system includes five pattern-recognition molecules
(PRMs), including mannan-binding lectin (MBL; mannose-binding lectin), M-ficolin (ficolin-1),
L-ficolin (ficolin-2), H-ficolin (ficolin-3), and CL-LK. CL-LK is found in plasma
as a heteromer of the two CL-L1 and CL-K1 polypeptide chains. MBL-associated serine
proteases 1, 2, and 3 (MASP-1, MASP-2, and MASP-3) are in complex with the PRMs and
are activated upon binding of the PRMs to a fitting target, e.g., microorganisms or
altered self-structures. Additionally, MBL-associated proteins 44 (MAp44) and 19 (MAp19)
are associated with the PRMs.[14]
The role of the complement system in SLE is well-known, and its importance is underlined
by its inclusion in the SLE classification criteria.[15] In APS, recent studies reported hypocomplementemia and increased levels of complement
activation products,[16] and animal models of aPL-induced thrombosis showed diminished thrombus formation
in complement-deficient rats.[16]
[17]
Increased activation of platelets has been found in SLE and APS. Thrombocytopenia
correlates with disease activity in SLE, and platelet immune-modulating responses
have been suggested to contribute to SLE pathogenesis.[18] In APS, aPLs bind to and activate platelets and potentiate agonist-stimulated aggregation.[19]
Several studies have demonstrated complement–platelet interactions.[20]
[21] In SLE, complement deposition on platelets has been associated with vascular events,
both venous and arterial, and the association was strengthened by the presence of
aPLs.[22] Moreover, Peerschke et al found complement fixation on platelets to associate with
the presence of aPLs and increased platelet activation, suggesting an aPL-mediated
link between platelet–complement interactions.[23]
This leads to the hypothesis that crosstalk between complement and platelet plays
a role in the thrombotic pathophysiology in SLE and APS. However, studies focusing
on the lectin pathway of complement and its interactions with platelets are sparse.
To further explore the prothrombotic pathophysiology in SLE and APS, the present study
investigated the correlations between lectin pathway protein (LPP) concentrations,
complement activation, platelet aggregation, and platelet activation in SLE, APS,
and healthy controls.
Methods
Study Population and Design
This cross-sectional cohort study was conducted at Aarhus University Hospital, Denmark,
from October 2019 to August 2020. Patients with SLE were included at the Rheumatology
outpatient clinic, and patients with APS were included at the outpatient clinic at
the Thrombosis and Haemostasis Clinic. Healthy controls were recruited at the Blood
Bank, Aarhus University Hospital.
Inclusion criteria for the patients were age ≥18 years with either a diagnosis of
SLE or APS. SLE diagnosis required at least 4 out of 11 criteria from the American
College of Rheumatology's revised classification criteria of 1997 to be fulfilled.[24] APS diagnosis required the Sydney Classification Criteria for Definite Antiphospholipid
Syndrome to be fulfilled.[25] Patients were excluded if they met one or more of the following criteria: active
cancer, liver failure (> Child–Pugh class A), pregnancy, use of anticoagulants or
antiplatelet drugs, except for low-molecular-weight heparin (LMWH). LMWH was never
administered on the day of blood sampling. SLE patients were excluded if they met
both clinical and biochemical criteria for secondary APS, as they all received regular
anticoagulant treatment.
Clinical information regarding diagnosis, disease activity, medical history, and clinical
manifestations, including previous thromboembolic disease, was systematically collected
through interviews supplemented by electronic medical and pharmacy records.
The inclusion criteria for healthy controls were age between 17 and 75 years. Exclusion
criteria were autoimmune disease of the muscle, joints, skin, or connective tissue,
hemophilia or coagulation defects, active cancer, and use of anticoagulants or antiplatelet
drugs.
The study was approved by the Central Denmark Region Committees on Health Research
Ethics (case no. 1-10-72-91-19) and the Danish Data Protection Agency (case no. 1-16-02-346-19).
Written informed consent was obtained from all patients before inclusion.
Blood Sampling and Laboratory Analyses
Blood was drawn from the antecubital vein through a 21-gauge needle. The first tube
was discarded.
For the APS cohort, aPL antibodies were detected at two times with at least 12 weeks
apart, before project blood samples were acquired, and blood samples were acquired
at least 3 months after any clinical event.
Samples for hemoglobin, platelet count, immature platelet fraction, immature platelet
count, and leucocytes were collected in 2 mL ethylenediaminetetraacetic acid (EDTA)
tubes and measured by Sysmex XN9000 (Sysmex, Kobe, Japan). Likewise, C3 and C4 concentration
samples were collected in EDTA tubes and determined by turbidimetry using ADVIA Chemistry
XPT Systems (Siemens, Chicago, Illinois, United States). C-reactive protein (CRP)
and creatinine were collected in 4 mL lithium-heparin tubes and analyzed employing
a Cobas 6000 (Roche Diagnostics, Indianapolis, Indiana, United States). Samples for
tests of fibrinogen (functional, Clauss method), fibrin D-dimer, activated partial
thromboplastin time (aPTT), and international normalized ratio (INR) were collected
in 1.8 mL citrate 3.2% tubes and analyzed using a Sysmex CS 5100i (Sysmex, Kobe, Japan).
Antinuclear antibodies (ANAs) were determined in serum by fluoroenzyme-immunoassay
using a Phadia 250 (Thermo Scientific, Uppsala, Sweden) or indirect immunofluorescence
with human epithelial cells (Hep-2 cells) as substrate. A positive ANA was defined
as a fluoroenzyme-immunoassay ratio above 1 or a Hep-2 assay with a dilution ratio
above or equal to 1:160.
Lupus anticoagulant (LA) was analyzed by the dilute Russell Viper Venom test and a
lupus-sensitive aPTT using a CS 2100i Sysmex (Siemens/Dade Behring, Marburg, Germany).
The concentration of immunoglobulin G (IgG) and IgM anticardiolipin antibodies (aCLs)
was determined using a Bio-Flash (Inova Diagnostics, San Diego, United States), and
IgG and IgM anti-β2GPI antibodies were determined using a Phadia 250 (Thermo Scientific, Uppsala, Sweden).
aPL positivity was defined as either an LA1/LA2 ratio of or above 1.4, a measurement
of anti-β2GPI antibodies above 10*103 int. units/L, and measurement of aCL antibodies above 40*103 int. units/L, as defined by the APS criteria.[25]
Platelet Aggregation
Blood was collected in sodium citrate 3.2% tubes and analyzed within 1 hour of sampling
by light transmission aggregometry (Bio/Data Corporation, Pennsylvania, United States).
Calf Skin Collagen agonist (type 1) lyophilized collagen (0.19 mg/mL, Bio/Data Corporation,
Pennsylvania, United States) and adenosine diphosphate (ADP) (2 µM, Bio/Data Corporation,
Pennsylvania, United States) were used as agonists. The final concentrations were
0.19 mg/mL (collagen) and 200 µM (ADP). Platelet-rich and platelet-poor plasma were
prepared using PDQ Platelet Function Centrifuge (Bio/data Corporation, Pennsylvania,
United States). Platelet aggregation was monitored for 6 minutes. Results were expressed
as maximum aggregation (%).
Platelet Activation
Blood was collected in sodium citrate 3.2% tubes and rested for 1 hour. Flow cytometry
measurement of platelet activation was performed on whole blood employing Navios EX
Flow Cytometer (Beckman Coulter, Miami, United States). Fluorescent-labelled antibodies
specific for P-selectin (eBioscience, San Diego, California, United States), CD63
(PE Cy7, BD Bioscience, San Jose, California, United States), and bound fibrinogen
(FITC, Polyclonal chicken, Diapensia HB, Linköping, Sweden) were used.
P-selectin is an indicator of α-granule secretion, whereas CD63 is an indicator of
δ-granule secretion, both being activation-dependent markers. The glycoprotein IIb–IIIa
complex is only capable of binding to fibrinogen following a conformational change
after activation. Consequently, the use of specific antibodies against fibrinogen
allows us to explore platelet activation-dependent surface changes.
The four agonists were ADP (140 μM, Sigma-Aldrich, St. Louis, United States), arachidonic
acid (AA; 7.5 mM, Sigma-Aldrich, St. Louis, United States), collagen-related peptide,
platelet GpVI ligand (collagen) (1.5 μg/mL, University of Cambridge, United Kingdom),
and thrombin receptor activating peptide (TRAP; 371 μM, JPT, Berlin, Germany).[26] The final concentrations of the four agonists were 10.8 µM (ADP), 0.58 mM (AA),
0.12 µg/mL (collagen), and 28.6 µM (TRAP). A pool of 35 µL HEPES buffer, 5µL agonist,
and 5 µL off each antibody was made. 5 µL blood was added, and samples were incubated
for 10 minutes at room temperature in darkness, giving a final dilution of 1:13. Fixation
was performed using 2 mL 0.2% PFA-PBS. Preparation and fixation were completed within
2 hours of sampling. Platelet activation was expressed as the percentage of positive
platelets within a gate (%-positive platelets) and median fluorescence intensity (MFI)
of all platelets within the gate. For each sample, gates were set to include 1 to
2% positive events for CD63 and bound fibrinogen and 0.1 to 0.2% for P-selectin on
the negative control using single-stained platelets and matching isotype controls,
as previously optimized in our lab.[27] The gating strategy is shown in [Supplementary Fig. S1]. Daily quality control of particle size and fluorescence was performed according
to the manufacturer's instructions using Flow-Check Pro and Flow-set Pro (Beckman
Coulter, Florida, United States). In performance of the flow cytometric analysis,
the MIFlowCyt guideline, as outlined by Lee et al,[28] was followed including compensation, type of flow cell, and signal characteristics,
with the information also being available in our protocol publication.[27]
Lectin Pathway Protein Concentrations
Plasma was frozen after collection and stored at −80°C until analysis. The concentrations
of LPPs (MBL, CL-L1, CL-K1, M-ficolin, H-ficolin, MASP-1, -2 and -3, MAp19 and MAp44)
and the complement activation fragment C3dg were determined in EDTA plasma using time-resolved
immunofluorometric assays at the Department of Biomedicine, Aarhus University, employing
protocols previously described.[29]
[30] Briefly, microtiter wells were coated with mannan (for the MBL assay), acetylated
bovine serum albumin (for the H-ficolin assay), or relevant capture antibodies for
the remaining assays. Plasma was thawed and diluted in assay-specific buffers and
added to the coated microtiter wells. Biotinylated assay-specific antibodies followed
by europium-labelled streptavidin were used to detect proteins bound in the wells.
Signals were compared to a standard curve with known protein concentration, and three
quality controls were used for each microtiter plate. Intra- and interassay coefficients
of variation were below 15% for all assays. The concentration of L-ficolin was measured
by a commercial immunoassay, following the instructions of the manufacturer (HK336-02,
HycultBiotech, Uden, The Netherlands). As mentioned in the introduction, CL-LK is
a heteromer of the two CL-L1 and CL-K1 polypeptide chains. In the present report we
measure the concentration of both of the CL-L1 and CL-K1 polypeptide chains.
Statistical Analyses
Data were visually assessed for Gaussian distribution by QQ-plots and histograms.
Data were log-transformed if Gaussian distribution could not be assumed on the normal
scale. For comparison across all three groups, one-way ANOVA (analysis of variance)
or the Kruskal–Wallis test was used for data following and not following Gaussian
distribution. When a significant difference, defined as a p-value <0.05, was found across groups, a post-hoc unpaired t-test was performed between groups for data following Gaussian distribution and Mann–Whitney
U-test for data not following Gaussian distribution. Spearman's rank correlation coefficient
was calculated for correlation analyses.
As this was considered an exploratory study, no corrections for multiple testing were
made.
CRP values below 4 mg/L were noted as 4 mg/L, INR values below 1 were noted as 1,
and fibrin D-dimer values below 0.25 mg/L were noted as 0.25 mg/L.
Data were described with median and interquartile range (IQR) and presented graphically
as dot plots with bars marking median and IQR unless otherwise stated.
Statistical analyses were performed using Stata Statistical Software version 16.1,
and graphical illustrations were created using GraphPad Prism version 8.4.3.
Results
Study Population
In total, 20 SLE patients, 17 patients with APS, and 39 age-matched healthy controls
were included. Age was similar across the three groups ([Table 1]).
Table 1
Baseline characteristics for patients with SLE or APS and healthy controls
|
Demographics
|
SLE, n = 20
|
APS, n = 17
|
Healthy controls, n = 39
|
|
Age at inclusion, y
|
44 [35; 55]
|
43 [32; 51]
|
44 [33; 55]
|
|
Age at diagnosis, y
|
40 [28; 49]
|
43 [31; 51]
|
|
|
Gender, female
|
20 (100)
|
8 (47)
|
28 (72)
|
|
Race, white
|
18 (90)
|
17 (100)
|
|
|
Body mass index, kg/m2
|
23 [21; 24]
|
27 [26; 31]
|
|
|
Smoker status
|
|
Current smoker
|
1 (5)
|
3 (18)
|
|
|
Previous smoker
|
10 (50)
|
8 (47)
|
|
|
Biochemical data at day of inclusion
|
|
|
Reference interval
|
|
Hemoglobin, mM
|
8.0 [7.6; 8.5]
|
8.8 [8.1; 9.3]
|
7.3–10.5[a]
|
|
Platelet count, 109/L
|
244 [202; 293]
|
238 [199; 285]
|
145–400[a]
|
|
Immature platelet fraction
|
0.032 [0.023; 0.048]
|
0.024 [0.015; 0.043]
|
0.016–0.126
|
|
Immature platelet count, 109/L
|
9.4 [6.8; 12.5]
|
5.7 [4.6; 8.8]
|
4.4–26.7
|
|
Fibrinogen, µmol/L
|
8.1 [7.2; 9.5]
|
9.9 [8.2; 12.0]
|
5.5–12.0
|
|
Fibrin D-dimer, mg/L
|
0.3 [0.3; 0.6]
|
0.3 [0.3; 0.4]
|
< 0.5[b]
|
|
aPTT, s
|
25 [22; 27]
|
24 [22; 25]
|
20–29
|
|
INR
|
1.1 [1.0; 1,2]
|
1.0 [1.0; 1.0]
|
< 1.2
|
|
White blood cell count, 109/L
|
4.4 [3.7; 5.6]
|
6.6 [5.0; 7.9]
|
3.3–10.0
|
|
Creatinine, µmol/L
|
57 [49; 64]
|
68 [62; 73]
|
45–105[a]
|
|
CRP, mg/L
|
4 [4; 5]
|
4 [4; 6]
|
< 8
|
|
Positive ANA
|
20 (100)
|
2 (12)
|
|
|
aPL status
|
|
aPL present at any time
|
4 (20)
|
17 (100)
|
|
|
aPL present 2 times 12 weeks apart
|
2 (10)
|
17 (100)
|
|
|
aPL subtype (present 2 times 12 weeks apart)
|
|
|
|
|
Lupus anticoagulant
|
2 (10)
|
15 (88)
|
|
|
Anticardiolipin IgM or IgG
|
1 (5)
|
3 (18)
|
|
|
Anti-beta2-glycoprotein-I IgM or IgG
|
1 (5)
|
1 (6)
|
|
|
aPL-related complications
|
|
Arterial thrombosis
|
0 (0)
|
3 (18)
|
|
|
Venous thrombosis
|
1 (5)
|
7 (41)
|
|
|
Obstetric complications
|
4 (20)
|
7 (88)
|
|
|
Concomitant disease
|
|
Concomitant chronic disease[c]
|
4 (20)
|
3 (18)
|
|
|
Hereditary thrombophilia[d]
|
1 (5)
|
2 (12)
|
|
Abbreviations: ANA, antinuclear antibodies; APS, antiphospholipid syndrome; aPTT,
activated partial thromboplastin time; CRP, C-reactive protein; INR, international
normalized ratio; SLE, systemic lupus erythematosus.
Note: Results are provided as median [IQR] or n (%).
a Combined RI for females and males.
b Reference interval increases with 0.1 mg/L per 10 years of age, from age 55 and above.
c Including the autoimmune or inflammatory diseases: Sjögren's syndrome, rheumatoid
or juvenile chronic arthritis, autoimmune thrombocytopenia, autoimmune hypo- or hyperthyroidism,
poly- or dermatomyositis, myasthenia gravis, coeliac disease, fibrosing alveolitis,
chronic active hepatitis, chronic urticarial, psoriasis, colitis ulcerosa, asthma
requiring systemic treatment.
d Hereditary thrombophilia includes factor V Leiden, prothrombin G20210A mutation,
protein S, C, or antithrombin deficiency.
Regarding hereditary thrombophilia, only three SLE patients had been systematically
examined for hereditary thrombophilia, and one of these was heterozygous for factor
V Leiden. All APS patients were investigated at the time of APS diagnosis, and three
patients were heterozygous for Factor V Leiden.
One SLE and one APS patient in the cohort were triple aPL positive.
Besides the eight SLE patients receiving prednisolone treatment on the day of inclusion,
one APS patient also received prednisolone treatment due to rheumatic polymyalgia.
Seven patients with APS received LMWH outside the day of blood sampling. APS patients
did not take any other medication.
The SLE patients had low disease activity, indicated by a median SLEDAI score of 4,
and were managed on hydroxychloroquine (HCQ; 85%) and low-dose prednisolone (40%)
([Supplementary Table S1]).
Three arterial and seven venous thromboembolic events had previously occurred in the
APS cohort ([Table 1]). All arterial cases were ischemic strokes. Of the seven patients with APS-related
venous thromboembolic events, two were in vena porta and one in vena jugularis. The
remaining four patients had deep venous thrombosis, and three had a simultaneous pulmonary
embolism.
In the SLE cohort, four patients had experienced unexplained fetal death after the
10th week of gestation, and one experienced three or more consecutive spontaneous
abortions before the 10th week of gestation. For the APS cohort, two patients had
experienced unexplained fetal death after the 10th week of gestation, two patients
experienced three or more spontaneous abortions before the 10th week of gestation,
and three patients experienced premature birth before the 34th week of gestation due
to severe pre-eclampsia.
Lectin Pathway Protein Levels in Patients with SLE or APS and Healthy Controls
Levels of LPPs are illustrated in [Fig. 1], and all values are presented in [Supplementary Table S2]. Four LPPs (L-, M-, H-ficolin, and MAp19) differed in concentrations between patients
with SLE, APS, and healthy controls. Higher concentrations of L-ficolin were found
in patients with APS than in patients with SLE (p = 0.004), and higher H-ficolin concentrations were found in both SLE (p = 0.025) and APS (p = 0.004) patients compared to healthy controls. The concentration of M-ficolin was
lower in the SLE cohort compared to the APS cohort (p = 0.004) and healthy controls (p = 0.032), and the MAp19 concentration was higher in APS patients than in SLE patients
(p = 0.007) and healthy controls (p < 0.001).
Fig. 1 Concentrations of lectin pathway proteins and C3dg in patients with SLE (n = 20), patients with APS (n = 17), and healthy controls (n = 39). Bars indicate median and interquartile range. Capped lines indicate p-values across all three groups. Tick-down lines indicate p-values between two groups when significant. APS, antiphospholipid syndrome; CL-K1;
collectin kidney 1; CL-L1, collectin liver 1; MAp19/44, MBL-associated protein of
19/44 kDa; MASP, MBL-associated protease; MBL, mannose-binding lectin; SLE, systemic
lupus erythematosus.
A positive correlation between C3dg and M-ficolin was found in the APS population
(r = 0.50, p = 0.043). No other significant correlations were observed between C3dg and the proteins
with a concentration difference.
Platelet Aggregation and Activation in SLE and APS Patients and Healthy Controls
We found no differences in platelet aggregation between the three cohorts, except
for a significantly lower ADP-induced platelet aggregation in the SLE cohort than
in healthy controls (p = 0.009) ([Supplementary Fig. S2] and [Supplementary Table S2]).
Platelet activation is also illustrated in [Fig. 2] and presented in [Supplementary Table S2]. The %-positive platelets were significantly higher in the SLE cohort than in healthy
controls when using AA as an agonist for bound fibrinogen (p = 0.001), CD63 (p = 0.007), and P-selectin (p < 0.001). A similar increase in AA-induced %-positive platelets was found in the
APS cohort; however, only significant for the surface marker-bound fibrinogen (p = 0.014). The %-positive platelets for bound fibrinogen, CD63, and P-selectin did
not differ significantly (p-values > 0.05). The median (IQR) preactivation, as assessed by %-positive platelets
for P-selectin after the addition of HEPES buffer, was 6% (5–11%), 7% (5–9%), and
10% (9–14%) for the SLE, APS, and healthy controls, respectively.
Fig. 2 Flow cytometry analyses of platelet activation in patients with SLE (n = 20), patients with APS (n = 17), and healthy controls (n = 39). Expressions of activation-dependent platelet surface markers bound fibrinogen,
CD63, and P-selectin are shown. Graphs to the left illustrate the %-positive platelets.
Graphs to the right illustrate the MFI of the platelet surface markers. Bars indicate
median with IQR. The tick-down lines indicate p-values between two groups when significant. AA, arachidonic acid; ADP, adenosine
diphosphate; APS, antiphospholipid syndrome; Collagen, collagen-related peptide; MFI,
median fluorescence intensity; SLE, systemic lupus erythematosus; TRAP, thrombin receptor
activating peptide-6.
Lower MFIs were found in SLE patients compared to healthy controls. The MFI of bound
fibrinogen was significantly lower using ADP (p =0.003) and TRAP (p < 0.001) as agonists. For CD63, the MFI was significantly lower using ADP (p = 0.014), TRAP (p = 0.009), and AA (p = 0.002) as agonists.
Similarly, lower MFIs of bound fibrinogen were found in APS patients than in healthy
controls when using ADP (p = 0.008) and TRAP (p = 0.011) as agonists, but this was not the case for MFIs of CD63. When comparing
APS and SLE patients, the MFIs of bound fibrinogen, CD63, and P-selectin did not differ
significantly, except for CD63 with ADP as an agonist, where SLE patients showed lower
MFI than patients with APS (p = 0.049).
SLE patients not treated with prednisolone (n = 12) had a higher percentage of platelets positive for bound fibrinogen (p = 0.002), CD63 (p < 0.001), and P-selectin (p < 0.001) when using AA as an agonist compared to healthy controls. Similar results
were found when comparing SLE patients with and without prednisolone treatment. However,
this was only significant for the platelet marker CD63 (p = 0.01) ([Fig. 3]). No significant differences were found for MFI.
Fig. 3 Flow cytometry analyses of platelet activation, measured as %-positive platelets
for bound fibrinogen, CD63, and P-selectin, using arachidonic acid as an agonist,
in patients suffering from SLE not receiving prednisolone treatment (n = 12), patients suffering from SLE receiving prednisolone treatment (n = 8), and healthy controls (n = 39). Bars indicate median with IQR. Capped lines indicate p-values across all three groups. Tick-down lines indicate p-values between two groups when significant. APS, antiphospholipid syndrome; SLE,
systemic lupus erythematosus.
In the platelet aggregation and activation analyses, a few low outliers can be seen
in [Fig. 2] and [Supplementary Fig. S2]. A systematic review of outlying data points in the analyses was performed to clarify
whether these outliers could be explained by the use of drugs inhibiting platelet
function (suppression of platelet function under AA stimulation could indicate the
use of aspirin or nonsteroidal anti-inflammatory drug (NSAID), and suppression of
platelet function during ADP stimulation could indicate the use of P2Y12s inhibitors).
One SLE patient showed low platelet activation following stimulation but not in platelet
aggregation. The patient did not report any use of aspirin or NSAID, and no drugs
were administered, according to the medical database.
One APS patient showed low platelet activation and aggregation following stimulation
by ADP and TRAP. According to medical records, the patient did not use any ADP inhibitors
or ASA-containing drugs at the time of inclusion.
To inspect the possible effect of LMWH, all patients' aPTT was reviewed. One of the
seven patients receiving LMWH had a prolonged aPTT of 38 as well as prolonged thrombin
time. As so, we cannot rule out an effect of LMWH on the remaining analyses in this
individual.
Correlation between LPPs and Platelet Activation in APS Patients
In the APS cohort, MASP-2 concentrations correlated negatively with both %-positive
platelets and MFI for bound fibrinogen with all agonists (all r < − 0.3), except for %-positive platelets when using collagen as an agonist. No significant
correlation between MASP-2 and plasma fibrinogen was found in the APS cohort (r = 0.34, p = 0.18). No correlation was found in the SLE or control cohorts ([Fig. 4] and [Supplementary Table S3]).
Fig. 4 Correlations between MASP-2 concentrations and MFI of platelet-bound fibrinogen after
agonist stimulation in patients with APS (n = 17). The agonist is defined in the titles. AA, arachidonic acid; ADP, adenosine
diphosphate; APS, antiphospholipid syndrome; Collagen, collagen-related peptide; MASP,
MBL-associated protease; MFI, median fluorescence intensity; TRAP, thrombin receptor
activating peptide-6.
Correlation between Complement Activation and Platelet Activation in APS Patients
C3dg concentrations correlated negatively with both %-positive platelets and MFI across
all surface markers and agonists in the APS cohort, except for MFI of CD63 after collagen
stimulation ([Fig. 5]). The strongest correlations were found between C3dg and P-selectin expression,
measured as both MFI and %-positive platelets (r between −0.18 and −0.82). However, only three out of eight correlation coefficients
were significant (p < 0.05).
Fig.
5 Correlations between C3dg concentrations and MFI of platelet surface markers after
agonist stimulation in patients with APS (n = 17). One outlier with C3dg concentrations at 194 mUnits/mL was excluded in the
graphical illustrations, as it did not affect the overall conclusions. See [Supplementary Table S4] for further correlation analyses. AA, arachidonic acid; ADP, adenosine diphosphate;
APS, antiphospholipid syndrome; Collagen, collagen-related peptide; MFI, median fluorescence
intensity; TRAP, thrombin receptor activating peptide-6.
No strong correlations between platelet activity and C3dg were found in healthy controls
or the SLE cohort ([Supplementary Table S4]).
Discussion and Conclusion
Discussion and Conclusion
This study is the first to investigate and compare LPP levels, complement activation,
and platelet function in three distinct cohorts of patients with SLE, APS, and healthy
controls.
The main finding was that LPP concentrations and platelet activation differed between
patients with SLE or APS. We demonstrated that increased platelet activation in the
SLE cohort was only observed in patients without prednisolone treatment. We showed
that the concentration of MASP-2 and C3dg correlated negatively with platelet activation
in the APS cohort. These findings suggest that complement and platelet activation
interactions occur in both SLE and APS, but the interactions are different.
The two patient cohorts in the present study varied significantly in gender distribution,
with the SLE cohort consisting solely of women. Previous studies from our lab demonstrated
no significant differences in platelet function across gender.[31] However, several LPPs vary between gender, with higher L- and H-ficolin, MASP-2
and -3 concentrations, and lower M- ficolin and MASP-1 concentrations in women.[32] Regarding other demographic parameters, the two cohorts were comparable.
It was intended to include a third patient cohort consisting of SLE patients with
secondary APS. However, these patients received lifelong anticoagulant therapy and
were excluded according to the project exclusion criteria.
Our group has previously quantified the concentrations of LPPs in a large cohort of
SLE patients. Overall, our findings align with those previously reported.[29] However, our previous studies showed increased levels of C3dg in patients with SLE,
which was not found in the current study, suggesting that the SLE patients had low
disease activity.
Interestingly, comparisons of LPP concentrations between the SLE and APS cohorts showed
significant differences in M- and L-ficolin as well as MAp19, whereas both cohorts
had increased levels of H-ficolin compared with healthy controls. This could indicate
genetic differences in LP regulating genes in the two cohorts or that LP proteins
serve other immune functions in patients unrelated to complement activation.
Only a few previous studies have quantified the concentrations of LPPs in a cohort
with APS or persistent presence of aPLs. In 2014, Breen et al investigated the concentration
of MBL, L- and H-ficolin in a cohort of 100 patients, of which 69 patients had primary
APS and 31 had isolated aPL antibodies present but no clinical APS outcomes. They
found no difference in LPP concentrations between the patient cohort APS and healthy
controls.[33] An important difference in Breen et al's approach is their use of serum for sampling
collection. In contrast, we used EDTA plasma to avoid further complement activation
by Ca2+ and Mg2+ chelation. Further, complement enzymes are less prone to auto-activation in EDTA
compared to serum.[34] A previous study from our laboratory determined LPP concentrations in serum and
plasma and showed serum concentrations of H-ficolin significantly higher than plasma
concentrations.[32] Likewise, previous studies have described L-ficolins' ability to bind to the silica
particles in serum tubes, leading to lower concentrations in serum than plasma.[35] Thus, the comparison of L- and H-ficolin concentrations between our study and the
study by Breen et al is hampered by sampling differences.
APS patients had significantly higher levels of MAp19 than healthy controls and SLE
patients. MAp19 is suggested to be a modulator of the lectin pathway; however, the
biological functions of MAp19 are yet to be elucidated.[14] Increased levels of MAp19 in plasma could be due to increased production, decreased
usage, or reduced ability to bind to cell surfaces. MAp19 and MASP-2 are splice variants
from the same gene.[36] As MASP-2 was not found to be increased in APS patients, high levels of MAp19 are
presumably not due to increased transcription but could be due to increased alternative
splicing of the gene. Concentrations of MAp19 have been measured in a few other cohorts.
In a cohort of septic shock patients, no significant differences in MAp19 concentrations
were found between septic shock patients and healthy controls.[37] In patients suffering out-of-hospital cardiac arrest, Bro-Jeppesen et al demonstrated
a significantly lower concentration of MAp19 compared with healthy controls, and low
MAp19 levels were associated with increased mortality.[38]
Previous studies have shown increased platelet activation in patients with SLE.[39] Contrarily, our study shows an equal or decreased platelet activation and aggregation
in SLE compared to healthy controls on almost all parameters, except %-positive platelets
when stimulated with AA. A recent study by Cornwell et al found a significant reduction
in platelet aggregation and decreased P-selectin expression in SLE patients receiving
HCQ compared to SLE patients not receiving HCQ.[40] As 17 out of 20 patients in our cohort were treated with HCQ, this could be a factor
worth considering when interpreting our results.
SLE patients not receiving prednisolone had increased platelet activation to agonist
stimulation compared to patients treated with prednisolone and healthy controls, indicating
platelet activation is inhibited by prednisolone treatment. Two previous studies on
the in vitro effect of prednisolone on platelets found that prednisolone interacts
with glucocorticoid platelet receptors leading to reduced platelet adhesion and aggregation.[41]
Both cohorts showed decreased platelet activation measured as MFI. This could be due
to reduced platelet size, as previously found in an SLE cohort,[42] leading to fewer receptors per platelet. It has been theorized that constant in
vivo platelet activation can lead to partial degranulation and exhausted platelets
with decreased responsiveness to agonist stimulation ex vivo. Previous studies have
found platelets with a high degree of in vivo activation to have a reduced in vitro
function, supporting the theory of exhausted platelets.[43] Such constant activation in the circulation has been suggested in both SLE and APS,[44]
[45] and several reasons, including the presence of platelet-activating autoantibodies
like double-stranded DNA antibodies and aPLs, have been proprosed.[11]
[46]
In our APS cohort, we found a reverse correlation between platelet activation and
complement activation. This does not comply with previous studies. Peerschke et al
proposed that circulating aPL antibodies form immune complexes on the surface of platelets,
leading to activation of the classical pathway and deposition of complement proteins
on the surface of platelets, followed by complement-mediated platelet activation.[47] Likewise, complement fragments generated during activation of the complement system
can bind directly to platelet receptors. C3dg, measured in the present report, binds
to complement receptor 2 on platelets, leading to platelet activation.[48] Recently, Lonati et al measured increased platelet-bound C4d, a measure of complement
activation, in APS patients and found positive correlations between platelet-bound
C4d and aPL concentrations. Additionally, they showed in vitro binding of recombinant
anti-β2GPI antibodies to activated platelets, inducing C4d binding to the surface.[49] Svenungsson et al showed C4d deposition on platelets in SLE and an association with
vascular thrombotic events.[22] A plausible explanation for our results could be that constant aPL-mediated activation
of platelets and platelet-related complement activation is followed by C3dg generation
and exhaustion, leading to a decreased in vitro platelet activation potential.
Our study found a negative correlation between platelet activation, measured as bound
fibrinogen, and concentrations of MASP-2 in the APS cohort. Increased deposition of
MASP-2 has previously been demonstrated on activated platelets.[50] The lower concentration of MASP-2 with high platelet activation could represent
deposition on platelets and activation of the lectin pathway in APS. In our study,
this observation was APS-specific and could not be replicated in the SLE or control
cohort. A study of MASP-2s proteolytic effect on the coagulation proteins demonstrated
MASP-2's ability to mediate fibrinogen turnover through activation of prothrombin.[51] Kozarcanin et al showed that activated platelets bind ficolins, which is associated
with MASP-1 and -2 activation. Further, the cleavage product from fibrinogen and fibrin
could bind and activate both MASPs.[50] Our results support the theory of a link between the LPPs, coagulation proteins
such as fibrinogen, and platelet activation in APS.
Some advantages to our study are the well-characterized patient groups and the exclusion
of patients in antithrombotic treatment, making it possible to perform detailed investigations
of platelet aggregation and stimulus activation.
We recognize that our study has limitations. Our control cohort is age-matched to
both patient cohorts. Still, due to significant gender variations between the two
cohorts, the control cohort is possibly not an optimal match for either. It cannot
be excluded that gender is a potential confounding factor for our complement and platelet
results. As it is an exploratory study, multiple parameters were tested without correcting
for multiple testing. Also, our correlations are primarily hypothesis-generating and
require more studies to elucidate the interactions between complement and platelets
and the pathological role in thrombosis. Measurement of platelet activation during
nonstimulated conditions would have been relevant to elucidate platelet activation
in circulation further. We lack a measure of mean corpuscular volume of the platelets,
as this measurement has been associated with disease activity and platelet activation
in SLE cohorts.[52] There is a possible selection bias in the SLE cohort, as SLE patients with a history
of thromboembolic disease would be on antithrombotic treatment and excluded from our
study. Likewise, our SLE cohort had low disease activity, measured as SLEDAI, limiting
our results to SLE patients with low disease activity and excluding potential correlations
and associations with high disease activity. Although information regarding medical
use was collected through medical records and interviews, the use of over-the-counter
drugs cannot be ruled out, which could affect the platelet function analyses.
In summary, the LP represents a potential link between complement activation and platelet
activation.[16]
[50] However, complement and platelet interactions are different in the two diseases
we studied. We found reduced platelet aggregation and platelet activation in SLE patients
compared to healthy controls, which could partly explain the use of prednisolone.
In the APS cohort, we found negative correlations between MASP-2 and C3dg concentrations
and platelet activation. The findings suggest that complement and platelet interactions
occur in both SLE and APS although the mechanisms behind the interactions differ.
What Is Known about This Topic?
What Does This Paper Add?
-
Differences between APS and SLE patients, observed in both platelet activation and
complement activation, indicate that the prothrombotic state in these conditions has
different driving mechanisms.
-
In patients with APS, platelet activation showed a negative correlation with both
C3dg and MASP-2, implicating a role of the lectin pathway in platelet activation in
APS.
