Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2025 May 15.
Published in final edited form as: Platelets. 2024 Dec 16;35(1):2437241. doi: 10.1080/09537104.2024.2437241

Platelet aggregation responses to Salmonella Typhimurium are determined by host anti-Salmonella antibody levels

Rachel E Lamerton 1,2, Samantha J Montague 1, Marisol Perez-Toledo 2, Steve P Watson 1,*, Adam F Cunningham 2,*,✉
PMCID: PMC7617673  EMSID: EMS205385  PMID: 39681834

Abstract

Invasive non-typhoidal Salmonella infections are responsible for >75 000 deaths/year and >500 000 cases/year globally. Seventy-five percent of these cases occur in Sub-Saharan Africa, an increasing number of which are from multi-drug resistant strains. Interactions between bacteria and platelets can lead to thrombus formation, which can be beneficial for control of infection (immunothrombosis), or harmful through uncontrolled inflammation and organ damage (thromboinflammation). It is unknown whether Salmonella Typhimurium can activate human platelets. To assess this, light transmission aggregometry was used to measure platelet activation by two different Salmonella Typhimurium strains in 26 healthy donors in platelet-rich plasma and washed platelets. In platelet-rich plasma, but not in washed platelets, Salmonella Typhimurium activated platelets in a donor- and strain-dependent manner mediated through the low affinity immune receptor FcγRIIA and the feedback agonists, ADP and thromboxane A2. Plasma swap studies between strong and weak responders demonstrated a plasma component was responsible for the variation between donors. Depletion of anti-Salmonella antibodies from plasma abolished Salmonella-induced platelet aggregation responses, and addition of polyclonal anti-Salmonella antibody allowed aggregation in washed platelets. Correlating levels of anti-Salmonella total IgG or the IgG1, IgG2, IgG3 and IgG4 subclasses to platelet responses revealed total IgG levels, rather than levels of individual subclasses, positively correlated with maximum platelet aggregation results, and negatively with lag times. Overall, we show that anti-Salmonella IgG antibodies are responsible for donor variation in platelet aggregation responses to Salmonella and mediate this activity through FcγRIIA.

Keywords: Antibodies, host–pathogen interactions, platelets, Salmonella, thromboinflammation

Introduction

As well as their role in hemostasis and thrombosis, platelets play an important role in immunity, with the ability to influence the immune response in multiple ways, from secretion of immunomodulatory molecules to interactions with leukocytes.1 Furthermore, the formation of platelet-containing clots can help trap bacteria, inhibiting their dissemination, in a process termed immunothrombosis.2 However, these clots can also influence the local inflammatory environment, and dysregulation of this can lead to a thromboinflammatory response in which thrombus formation becomes pathological.3 Therefore, platelet activation can result in beneficial or harmful outcomes for the host.

Invasive non-typhoidal Salmonella (iNTS) infections are a major cause of bacteremia in sub-Saharan Africa.4,5 Patients typically present with fever, with a range of other diverse symptoms possible, from splenomegaly to severe anemia.6 The highest incidence of infection occurs in the under 5s, with the immunocompromised and the elderly also disproportionately affected.6,7 In contrast to the non-typhoidal Salmonella serovars present across the world that mainly cause self-limiting gastroenteritis in immunocompetent individuals, iNTS bacteria are able to cross the gut barrier and enter the bloodstream. These strains are now the dominant isolates in many sub-Saharan regions and are becoming increasingly multi-drug resistant.8 Salmonella enteritidis serovar Typhimurium (hereon written as Salmonella Typhimurium, STm) accounts for over 77% of iNTS isolates serotyped in Africa7 and the representative epidemic STm strain D23580 is resistant to ampicillin, chloramphenicol, streptomycin, trimethoprim and sulfonamide, making treatment of this life-threatening infection problematic. The Global Burden of Disease Study estimated there were 535 000 cases of iNTS worldwide in 2017, leading to over 77 500 deaths,9 with many case reports describing thrombotic events concurrent with Salmonella infections of various serovars.10–15

Previous studies assessing the molecular interactions between bacteria and platelets have identified platelet receptors GPIbα, αIIbβ3, TLR2 and FcγRIIA as being involved in binding to certain bacterial strains.16–19 However, whilst there have been many studies looking at human platelet interactions with Gram-positive bacterial strains, there are far fewer studies investigating Gram-negative bacterial strains, with little research into non-typhoidal Salmonella strains. The only studies have looked at the effect of Salmonella Typhimurium on ADP-induced platelet aggregation, with one finding that purified filtered STm could inhibit ADP-induced aggregation (strain dependent),20 and the other finding that purified STm porins could enhance both ADP and thrombin-induced platelet aggregation.21

The aim of this study therefore was to assess whether two different strains of STm could activate human platelets.

Methods

Preparation of bacteria

One colony of Salmonella Typhimurium SL1344 (a standard laboratory strain22) or D23580 (an invasive African isolate23,24) was grown up overnight in 10 ml LB broth at 37°C. Cultures were then diluted 1:5 in LB broth and incubated for 2 h in an orbital shaking incubator at 37°C, 200 rpm, to obtain log-phase bacteria. Bacteria were spun at 10 000 g for 5 min, pellets resuspended in PBS, then centrifuged and resuspended twice more to wash bacteria. At the final resuspension, PBS was added to give a final OD600 of 1, relating to approximately 1 × 109 CFU/ml bacteria.

Blood collection and ethics

Blood was taken from healthy consenting volunteers under ethical approval ERN_11–0175 granted by the University of Birmingham Internal Ethical Review Panel and in accordance with the Declaration of Helsinki.

Venous blood (from donors who had not taken aspirin or non-steroidal anti-inflammatory drugs in the previous 10 days) was collected into 3.2% or 3.8% sodium citrate vacuette containers using a 21 gauge needle.

Preparation of platelets

Platelet rich and platelet poor plasma

Blood taken into 3.2% sodium citrate was centrifuged at room temperature for 20 min at 200 g. The platelet-rich plasma (PRP) supernatant was removed before centrifuging the remaining buffy coat layer and red cells for 10 min at 1000 g to collect platelet poor plasma (PPP).

Washed platelets

Blood taken into 3.8% sodium citrate was further anticoagulated with 10% acid citrate dextrose (ACD, 120 mM sodium citrate, 110 mM glucose, 80 mM citrate dextrose, pH 4.4) before centrifugation for 20 min at 200 g. The PRP supernatant was collected and 0.2 μg/ml prostacyclin added before further centrifugation for 10 min at 1000 g. After discarding the supernatant, the pellet was resuspended in modified Tyrode’s (134 mM NaCl, 0.34 mM Na2HPO4.12 h2O, 2.9 mM KCl, 12 mM NaHCO3, 20 mM HEPES, 5 mM glucose, 1 mM MgCl2; pH 7.4) plus 3 ml ACD and 0.2 μg/ml prostacyclin and centrifuged for 10 min at 1000 g. The supernatant was discarded and platelet pellet resuspended in modified Tyrode’s to a concentration of 2 × 108 platelets/ml. Platelets were rested for 30 min prior to experiments.

ADP sensitive washed platelets

Blood taken into 3.2% sodium citrate was further anticoagulated with 22% ACD and centrifuged for 20 min at 200 g. Supernatant (PRP) was collected and a further 10% ACD was added before centrifugation for 15 min at 500 g. The supernatant was removed, and the surface of the platelet pellet washed gently 3 times with Krebs-RingercGlucose buffer (120 mM NaCl, 4.9 mM KCl, 1.2 mM MgSO4, 1.7 mM KH2PO4, 8.3 mM Na2HPO4, 10 mM glucose, pH 7.4) supplemented with 0.01 U/ml grade 7 apyrase. For washed platelet assays, the pellet was resuspended in the buffer used to wash to a concentration of 2 × 108 platelets/ml and 1 mM calcium chloride was added. For donor swap and antibody depletion assays, pellets were resuspended in ACD-free PPP supplemented withc 0.01 U/ml grade 7 apyrase (antibody depleted where stated), and adjusted to 2–3 × 108 platelets/ml.

Platelet counts were measured using either a Coulter Counter or XP-300 (Sysmex).

Antibody depletion of PPP

PPP was obtained as above, aliquoted and frozen at −20°C for up to 2 weeks. STm SL1344 was grown as above and normalized to OD600 = 1. Bacteria and thawed PPP were combined at a 1:1 volume ratio and rotated gently for 2 h at 4°C. Centrifugation for 5 min at 10 000 g was carried out to remove bacteria (and associated antibodies), with the resulting SL1344-binding anti-body depleted (dPPP) supernatant stored overnight at 4°C before use the following day. The ADP sensitive platelet preparation above was used to wash platelets, before resuspending pellet in either thawed PPP or dPPP.

Platelet aggregometry

A PAP-8 aggregometer (Bio-data Corp) was used for all experiments, with 1200 rpm stirring and 37°C maintained throughout. Traces were recorded for at least 30 min. Donor-specific PPP was used as a blank for PRP experiments, or ADP sensitive platelet assays resuspended in PPP. Tyrode’s buffer was used as a blank for washed platelet assays. Where indicated, platelets were incubated with inhibitor for at least 3 min prior to start of recording.

Anti-Salmonella antibody ELISA

To detect antibodies to Salmonella Typhimurium, bacteria were cultured as above and the pellets resuspended in PBS with 0.01% azide. The protein concentrations of washed bacteria were quantified using a Pierce™ BCA protein assay kit. 1 μg Salmonella in PBS was added per well to Nunc MaxiSorp 96 well plates and incubated at 4°C overnight. All subsequent steps were carried out at room temperature. Plates were washed 3 times with PBS before blocking with 2% (w/v) BSA in PBS-0.1% Tween 20 for 1 h. Plates were washed with PBS-0.1% Tween 20, three times, with this wash being carried out between all subsequent steps. Serum was diluted 1 in 20 in 2% (w/v) BSA in PBS-0.1% Tween 20, and incubated on plate for 1 h. 100 μl of the appropriate mouse anti-human HRP-conjugated secondary antibody diluted 1:4000 (unless stated) was added and plates incubated for 1 h: Total IgG (clone 2040–05; 1:8000), IgG1-Fc (clone HP6001), IgG2-Fc (clone 31-7-4), IgG3-hinge (clone HP6050) and IgG4-Fc (clone HP6025), all from Southern Biotech. Plates were developed with 100 μl TMB-Core (Bio-Rad) for up to 10 min, and the reaction stopped with 50 μl 0.2 M H2SO4. OD450 was read using a SpectraMax ABS Plus plate reader.

Statistics

Statistical analysis was performed using Graph Pad Prism 9.4.1. Results are presented as stated in figure legends, either as mean ± SD or median ± interquartile range (IQR). Significance assumed at p < .05. Shapiro-Wilk was used to assess normality of data. To test for differences between two groups, paired t-tests and Wilcoxon tests were carried out for parametric and non-parametric data respectively. To test for correlations, Pearsons correlation coefficient or Spearman rank correlation were calculated for parametric and non-parametric data respectively.

Results

Salmonella can activate platelets in platelet rich plasma in a donor- and strain-dependent manner

In this study, we have used two well-characterized strains of STm. SL1344 is a standard laboratory, virulent, noninvasive strain isolated from cattle,22 and D23580 is an invasive epidemic strain originally isolated from a child in Malawi.24

We first showed that neither of the strains caused an increase in light transmission in washed platelets demonstrating that they are unable to directly induce platelet aggregation (Supplementary Figure S1). In line with results in Arman et al,18 the addition of fibrinogen did not lead to aggregation (data not shown). Addition of pooled human IgG with or without fibrinogen also did not lead to aggregation (data not shown), consistent with the results obtained by Arman et al. for S. pneumoniae.18

Many bacterial species can indirectly activate platelets through bridging proteins such as IgG, fibrinogen or complement.25 To establish whether this was the case for STm, we monitored activation in platelet-rich plasma (PRP) using light transmission aggregometry. Both strains were able to cause platelet aggregation, but with marked differences between individual donors, as well as between the STm strains themselves (Figure 1). In all cases, the increase in light transmission was blocked in the presence of the αIIbβ3 blocker eptifibatide, demonstrating that responses were true aggregation and not agglutination (Supplementary Figure S2). To ensure we were using the optimal concentration of bacteria, we initially tested final concentrations from 1 × 107 CFU/ml to 3.3 × 108 CFU/ml. No responses were seen at 1 × 107 CFU/ml, but similar aggregation responses were observed at 1 × 108 CFU/ml and 3.3 × 108 CFU/ml for nearly all donors. Therefore, the bacterial concentration of 3.3 × 108 CFU/ml was chosen so we could rule out bacterial concentration as a factor in the variation in responses

Figure 1.

Figure 1

Open in a new tab

Salmonella activates platelets in a donor and strain dependent manner in platelet rich plasma (PRP). PRP was stimulated with a final concentration of ~3.3 x 108 CFU/ml STm SL1344 or STm D23580. Colour coding: strong responders (green; >60% maximum aggregation), mid responders (orange; 20-60% maximum aggregation), weak responders (cream, <20% maximum aggregation, but still with a curve), and nonresponders (red, no aggregation). (A) Representative traces from the different colourcoded categories to STm SL1344. The arrow shows the point at which the secondary wave of aggregation begins, demonstrating the biphasic response in this donor. (B) Table summarising responses to SL1344 and D23580 from n = 26 healthy donors. (C) Numerical maximum light transmission results for responses to SL1344 and D23580. n =26. Bars are median and inter-quartile range. Wilcoxon test, *** p ≤0.001. (D) Lag times taken for aggregation to occur to (n = 25) and D23580 (n = 22). Bars are median and inter-quartile range. Unpaired t-test ** p ≤ 0.01. (E) Maximum light transmission percentages of donors over time n = 5. ~6 month time point runs from 182-196 days after initial measurement, >15 months covers tests >450 days after initial measurement (F) Lag times of donors over time. n = 4.

To characterize the variation in further detail, we categorized the donors into different response groups (Figure 1A,B), with example aggregation curves for SL1344 shown in Figure 1A. Strong responders, shown in green, had a maximum platelet aggregation that was similar in magnitude to that of classical platelet agonists such as ADP. The two examples of strong responders shown in Figure 1A differed in lag time and pattern of response, with a biphasic curve observed in one donor. Mid responders are shown in orange and had maximum aggregation values between 20% and 60%. Weak responders had a maximal aggregation value of <20%, with two examples shown in cream (one with a more obvious curve than the other). Non-responders showed no response to the bacteria at all (despite normal responses to the platelet agonist, TRAP-6, data not shown), with an example trace shown in red. Figure 1B shows that while some donor responses are in the same category for both strains (e.g. Donors A and B), this was not the case for all donors (e.g. Donors D, G and H). In a total of 26 donors (age range 20–64 years; 11 female, 15 male), the median maximum aggregation response to SL1344 was 42% with an inter-quartile range (IQR) of 17.5–84.9% (Figure 1C). Bacterial-induced platelet aggregations have a characteristic lag time before platelet aggregation occurs, unlike traditional agonists,26 and this was observed in aggregations to STm. The lag times for onset of aggregation ranged from 2.5 to 12.5 min (median 8 min, IQR 6–9.5 min; Figure 1D). Responses tended to be weaker to D23580 than to SL1344, with a median aggregation response of 22% (IQR 8.6–57.6%) (Figure 1C) and a longer lag time median of 10 min (IQR 9–12 min) (Figure 1D). There was no clear link between sex or age and whether a donor was a responder or non-responder.

With such variation between donors, it was important to assess whether donor responses remained stable over time or not. The responses of donors to the two STm stimuli were tested at 6 and/or 15 months after the first test. As shown in Figure 1E, platelet aggregation responses remained in the same response category across both strains. Lag times also remained broadly similar (Figure 1F).

Salmonella-induced platelet aggregation is mediated by FcγRIIA and secondary mediators

We next set out to characterize the STm-platelet interactions. The receptor FcγRIIA and its ligand IgG have been implicated in many bacterial-platelet interactions.18,27,28 We therefore tested whether FcγRIIA was required for STm-induced platelet aggregation. The use of the F(ab) fragment of FcγRIIA blocking monoclonal antibody IV.3 reduced almost all aggregation responses to basal levels (Figure 2A). Where this was not achieved, use of the higher affinity F(ab)2 fragment was used, which caused full inhibition.

Figure 2. Salmonella–induced platelet aggregation is mediated by FCγRIIA and secondary mediators.

Figure 2

Open in a new tab

(A) PRP was pre-incubated with 20 μM F(ab) IV.3 or 10 μM F(ab)2 IV.3 or vehicle for 3 min before stimulation with a final concentration of ~3.3x108 CFU/ml of STm SL1344 or D23580. n ≥3 (B) PRP was pre-incubated with either 20 μMMRS2179, 1 μM cangrelor, 10 μM indomethacin, vehicle control or both indomethacin and cangrelor for 3 min. STm SL1344 was then added at a final concentration of ~3.3x108 CFU/ml. n = 3.

As shown in Figure 1A, in some donors the aggregation response is biphasic, suggesting the need for feedback agonists. To investigate this, we assessed the effect of inhibitors of the two major feedback agonists, ADP and thromboxane A2, in strongly responding individuals (Figure 2B). Inhibiting signaling through the ADP receptor P2Y1 with MRS2179 had no significant effect on the time course of the response or the level of aggregation. In contrast, addition of the P2Y12 inhibitor cangrelor blocked aggregation in two out of three donors and reduced the response in the remaining donor. The cyclooxygenase inhibitor indomethacin had a similar effect to that of cangrelor, with the magnitude of response unchanged in one donor but with a slight increase in lag time. When both cangrelor and indomethacin were used in combination, platelet aggregation was abolished in all three donors, confirming the requirement for secondary mediators (Figure 2B).

A plasma component is responsible for donor variation in platelet aggregation responses to Salmonella

Our next goal was to identify what was behind the variation in donor platelet responses to STm. As the PRP used was not normalized to a standard platelet count (as recommended in ISTH guidelines29,30), we first checked that differing platelet counts in PRP were not accounting for the inter-donor variability. Platelet counts obtained at the time of testing were plotted against maximum light transmission levels and this showed there was no correlation between platelet count and levels of aggregation (Suplementary Figure S3).

To establish whether the donor variation in platelet aggregation to STm was due to donor differences in platelets, plasma or both, “donor swap” assays were carried out. ADP-sensitive washed platelets and platelet poor plasma (PPP) were obtained from strong and weak responders. As controls, the strong responders’ platelets were resuspended in the strong responders’ plasma, and the same was carried out with the weak responder. To test whether the platelet or plasma component had a stronger influence over the donor variability, the strong responders’ platelets were resuspended in the weak responders’ plasma, and vice versa.

As shown in Figures 3A,B, the combination of the strong responders’ platelets with the weak responders’ plasma caused aggregation levels equivalent to that of the weak responder control. In contrast, the weak responders’ platelets resuspended in the strong responders’ plasma gave equivalent aggregation responses to the strong responder control. Together, these results suggest that a plasma component is responsible for the inter-donor variation in response to STm.

Figure 3.

Figure 3

Open in a new tab

A plasma component is responsible for donor variation in platelet aggregation responses to Salmonella. Two washed platelet pellets from a strong and weak responder were obtained using an ADP-sensitive preparation and resuspended in ACD-free PPP from both donors. Platelet concentrations were 2.3-2.5 x 108/ml depending on PPP amounts obtained for dilution. The final concentration of bacteria added was ~3.3x108 CFU/ml. (A) Representative traces of a donor swap experiment using SL1344. (B) Labels show platelet donor first, followed by plasma donor for SL1344 and D23580, bars represent mean. n =3.

Anti-Salmonella antibodies are required for aggregation, and their levels correlate with the strength of the platelet response

With the focus on identifying a plasma component responsible for donor variation in responses, and the knowledge that inhibiting FcγRIIA abolished aggregation responses, we examined the requirement of anti-STm IgG antibodies for platelet aggregation. Using a similar method to the donor swap assay above, we took PPP from donors and depleted it of anti-STm antibodies. We then resuspended the ADP sensitive washed platelet pellets in either their normal PPP (WP/PPP), or the anti-STm antibody depleted plasma (WP/dPPP). As shown in Figure 4A, the washed platelets resuspended in normal plasma aggregated as expected, but depleting the anti-STm antibodies from plasma abolished aggregation. Consistent with this, we show that the addition of a rabbit anti-Salmonella polyclonal antibody (STmIgG) to washed platelets before stimulating with STm-induced aggregation in a dose-dependent manner (Figure 4B).

Figure 4. Salmonella -binding antibodies are required for aggregation and their levels correlate with the strength of platelet response.

Figure 4

Open in a new tab

(A) ADP sensitive washed platelets were resuspended in either normal plasma (WP/PPP) or Salmonella-binding antibody depleted plasma (WP/dPPP). Bacteria was added to a final concentration of ~3.3x108 CFU/ml. Ai) Representative traces from a strong responder. Aii) Pooled data from n = 3. Colours represent separate donors – blue and purple were strong responders, teal a weak responder. Data are presented as WP/dPPP responses to the three bacterial relative to the control WP/PPP level taken as 100% maximum aggregation. (B) Dose response curves to the Salmonella strains using an ADP sensitive washed platelet preparation with the addition of 100 μg/ml rabbit anti-Salmonella IgG (STmIgG). n = 3. Bars are mean ± SD. (C) Total IgG levels to both strains were measured using an in-house ELISA, and results correlated against maximum light transmission levels (top 2 graphs, n = 12) and lag time (bottom 2 graphs, n ≥ 10) from Fig 1. Spearman rank correlation coefficient was calculated for the non-parametric maximum light transmission results. Pearsons correlation coefficient was calculated for the parametric lag times.

The levels of total IgG binding to whole STm bacteria by our donor serum were then tested by ELISA and levels correlated against the maximum platelet aggregation level and lag time. This revealed higher anti-STm IgG levels correlated with higher platelet aggregation levels, and faster lag times (Figure 4C).

To further investigate whether a specific IgG subclass was linked to platelet aggregation, whole bacteria ELISAs were carried out to assess levels of anti-STm IgG1, 2, 3 and 4. As seen in Figure 5A and Supplementary Figure S4A, none of the individual subclasses show clear correlation patterns. However, when combining subclasses, correlations were found, suggesting that the presence and total amounts of antibody are more important for platelet aggregation than the presence of a specific IgG subclass (Figure 5B and Supplementary Figure S4B).

Figure 5.

Figure 5

Open in a new tab

Individual Salmonella-binding IgG subclass levels do not correlate with platelet aggregation responses to Salmonella, but combining subclasses reveals a correlation. In house ELISAs were carried out to measure levels of anti-Salmonella IgG subclass antibodies to each of the three strains. Results were correlated against maximum aggregation. ODs were combined from multiple subclasses, and their totals correlated against maximum aggregation in the bottom half of the figure. n = 16. Spearman correlations for these graphs are in Supp Fig 4. Lines are simple linear regression.

Discussion

We show that STm can activate platelets from healthy donors in a donor- and strain-dependent manner. Binding of STm by antibody is needed to activate platelets via FcγRIIA, with higher levels of antibody leading to stronger platelet aggregation and shorter lag times.

Traditionally, bacterial-induced platelet aggregations have been viewed as all-or-nothing responses, with either full aggregation or no aggregation observed.25 The STm-induced platelet aggregation responses observed here challenge this concept, with differing levels of activation being dependent on the donor, and the aggregation levels varying in washed platelets according to the concentration of anti-Salmonella antibody added. While studies on various Gram-positive strains have revealed donor variations in lag time,31,32 the only study finding variation in maximum aggregation responses looked at platelet responses to heat aggregated human IgG.33 Variation between strains however is a more common finding, being shown in bacteria such as Cutibacterium acnes, Streptococcus oralis and Helicobacter pylori.28,34,35

Physiologically, the average blood culture result of iNTS patients was shown in one study to be just 1 CFU/ml.4 This is several magnitudes lower than the amount used in this study. However, it is not uncommon for bacterial-platelet interaction studies to use high concentrations of bacteria. The concentration of Salmonella used in this study is in line with other bacterial-platelet interaction studies, including Arman et al. (3 × 108 CFU/ml, S gordonii), Corcoran et al. (4 × 108 CFU/ml H pylori) and Kerrigan et al. (7 × 109 CFU/ml, S sanguinis). The differences in the levels of bacteria required likely reflect the timescales involved in the aggregation events. In our assays, aggregation takes place within 30 min. While the duration of illness in humans with iNTS has limited study data,36 in pre-clinical models it can be in the order of weeks. In a mouse model of Salmonella infection, thrombi occur in the spleen and liver after 1 and 7 days respectively.37,38 The high concentrations of bacteria used in this assay allow a fast response, providing a model in which to elucidate the potential mechanisms involved at the molecular level.

Although we showed correlations between total anti-STm IgG levels, no role for individual subclasses was revealed. Instead, the overall combination of IgG antibody subclasses binding to STm allows aggregation proportional to the levels present. Nevertheless, FcγRIIA has different affinities for the different subclasses (IgG3>IgG1>IgG4>IgG239), and these affinities are further influenced by the single nucleotide polymorphism (SNP) at position 131. For example, if an individual is homozygous for histidine then FcγRIIA has an increased affinity to IgG1 and IgG2, compared to when they are homozygous for arginine at this position.40,41 These affinity differences could potentially be enough to alter the amount of antibody bound by a platelet at any one time, influencing whether a strong or weak response is provoked. Despite the donor swap assays ruling out the platelets as a cause of donor variation, the donors used in the assay were the stronger and weaker responders, so the high levels of antibody present in the stronger responders could have compensated for any differences in platelet responses. We were unable to test mid-responders in this assay due to a reduction in platelet aggregation observed upon resuspending platelet pellets in PPP. This is similar to the phenomenon reported by Cattaneo et al.,30 where dilution of PRP with PPP to normalize platelet count inhibited the platelet aggregation responses. To circumvent this, we optimized the platelet preparation by adding apyrase, which helped limit the reduction in platelet sensitivity upon dilution in PPP, but still not to the extent whereby mid-responders were able to respond in their own PPP to a level close to their PRP response, limiting our experiments to strong/weak responders only. Therefore, further work would be needed to see whether FcγRIIA levels on the platelet surface and their polymorphisms could also be playing a role in donor variation.

As well as FcγRIIA, other platelet receptors could be playing a role in Salmonella-induced platelet aggregations. In the mouse model of Salmonella infection, platelet CLEC-2 was identified as being vital to the thrombotic response in the liver.38 However, mouse platelets have over 10× the amount of CLEC-2 on their platelets as humans,42,43 and do not express FcγRIIA. This highlights the importance of human studies alongside mouse studies, and further work would be required to assess whether CLEC-2 is involved in the mechanism in human platelets. Furthermore, human platelets also express complement receptors, such as gC1q/p33 and cC1qR. Complement protein C1q can bind antibody-bacteria complexes, which could provide a further mechanism of bacterial-recognition and activation of the platelet. Complement has been implicated in platelet aggregation responses to S sanguinis, with removal of complement abolishing the aggregation responses seen.32 Again, work beyond the scope of this study would need to be carried out to assess this.

Another area that requires further exploration is the source and antigen-specificity of the anti-STm antibodies involved in this process. Nearly all adults have anti-STm IgG antibodies.44,45 The O-antigen component of LPS is an immunodominant antigen and major target of antibodies against whole STm bacteria, reflecting the sheer prevalence of this antigen on the bacterial surface.46 Since depletion of anti-STm whole bacteria antibodies, which efficiently removes anti-LPS O-antigen antibodies,47 results in loss of platelet aggregation it suggests that anti-LPS antibodies induced to STm are important for FcγRIIA activation. However, antibodies induced to the O-antigen in other bacterial pathogens can potentially cross-react with STm44,48,49 and these may also contribute. As aggregation responses to STm remained constant over the sampling period, implying antibody levels are remaining constant, it may be that a combination of STm-specific and cross-reactive antibodies are present and can contribute to platelet activation.

Differences in platelet aggregation occurred between the two STm strains used here, with lower average maximum aggregation and longer average lag times for D23580 than SL1344. Although in many ways these strains are highly similar, hence both being STm, there are differences between the two and they are from different sequence type groups. The invasive disease-associated D23580 strain was first isolated from the bloodstream of a Malawian child, whereas SL1344 is of bovine origin and is widely used in laboratories around the world. D23580 shows some evidence of host adaptation to humans with some genome degradation relative to SL1344. Two of the genes degraded in comparison to SL1344 are putative outer or integral membrane proteins, and one deletion of a secreted protein has also been noted.24 Whether or how these could influence platelet aggregation responses directly is unclear.

The observation that higher amounts of anti-STm antibody cause higher levels of platelet aggregation has implications for the immunothrombotic or thromboinflammatory responses to STm infection. In someone who already has high amounts of anti-STm antibody present at the time of pathogen encounter then antibodies can bind the bacteria leading to platelet activation and thrombus formation around the bacteria. This immunothrombotic response to a minimal number of bacteria could help stop further replication and dissemination of the bacteria throughout the bloodstream. This entrapping of Salmonella in human platelet aggregates in vitro has previously been shown by Beristain-Covarrubias et al.37 In this instance, having a high level of anti-STm antibodies is likely to be helpful. On the other hand, in those who have low or no anti-STm antibodies at the time of infection, higher levels of antibody would not be present until later, when infection is established. By this point, there is the potential for bacterial numbers to be at high levels. Here, high levels of antibody plus high levels of bacteria could be providing more opportunities for platelet activation and thrombus formation, with the potential of leading to a dysregulated thromboinflammatory response, worsening disease.

To conclude, we show that STm can activate human platelets in the presence of anti-STm antibodies. Variation in donor responses to STm is linked to the levels of anti-STm antibody present. These findings begin to reveal the factors behind differences in individual thromboinflammatory responses to infectious agents, unveiling potential avenues for therapeutic intervention.

Supplementary Material

Supplementary Figures 1-2
Supplementary Figures 3-4

Plain Language Summary.

  • Salmonella is widely known as a bacterial pathogen that causes diarrhea, with recovery occurring after about a week. However, there exists another, less well known and more severe form of Salmonella, called invasive non-typhoidal Salmonella. This type of Salmonella infection gets across the gut barrier, enters the bloodstream and spreads through the body. There are over half a million of these kinds of invasive Salmonella infections globally every year, leading to over 75 000 deaths. The majority of these are in sub-Saharan Africa but also are significant causes of harm in the elderly or those with compromised immune systems. When the bacteria are in the bloodstream, they can come into contact with various different types of cells, including platelets, the cells that cause blood to clot. Sometimes the presence of bacteria in the blood can cause platelets to clump together, leading to blood clots, which can block the blood vessels and cause organ damage, and death. The ability of Salmonella bacteria to cause clots has not been investigated in depth.

What is new?

  • We show that the bacteria that cause invasive Salmonella infections are able to trigger platelets to clump together and form a blood clot. However, the response of platelets in blood plasma from donors varies widely - some people’s platelets do not clump together much, whilst others form large clots. We show that the amount people’s platelets clot is linked to the level of antibodies that bind to Salmonella.

What is the impact?

  • There are limited treatment options for invasive non-typhoidal Salmonella infections due to increasing antimicrobial resistance. Therefore, understanding the mechanisms behind how Salmonella interacts with blood cells such as platelets could help direct further research to identify new treatment options for these infections.

Funding

This work was supported by the Wellcome Trust Mechanisms of Inflammatory Disease (MIDAS) PhD Program [grant number 222389/Z/21/Z, part of 108871/B/15/Z] to R.E.L. S.P.W is a BHF Chair [Ch03/003]. S.J.M. is supported by a BHF project grant [PG/23/11230]. The National Institute of Health and Care Research (NIHR) Birmingham Biomedical Research Centre [NIHR203326] and the British Heart Foundation Accelerator (BHF) [AA/18/2/34218] have supported the University of Birmingham Institute of Cardiovascular Sciences where this research is based. We would like to thank Dr Caroline Kardeby for useful discussion.

Footnotes

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

REL designed the research, performed experiments, analyzed data, and wrote the manuscript. SJM provided supervision, designed the research and reviewed data. MPT performed preliminary experiments and designed the research. SPW and AFC provided supervision, obtained funding, designed the research, reviewed data and edited the manuscript

References

  • 1.Ali RA, Wuescher LM, Worth RG. Platelets: essential components of the immune system. Curr Trends Immunol. 2015;16:65–78. [PMC free article] [PubMed] [Google Scholar]
  • 2.Engelmann B, Massberg S. Thrombosis as an intravascular effector of innate immunity. Nat Rev Immunol. 2013;13(1):34–45. doi: 10.1038/nri3345. [DOI] [PubMed] [Google Scholar]
  • 3.Schrottmaier WC, Assinger A. The concept of thromboinflammation. Hamostaseologie. 2024;44(1):21–30. doi: 10.1055/a-2178-6491. [DOI] [PubMed] [Google Scholar]
  • 4.Gordon MA, Kankwatira AMK, Mwafulirwa G, Walsh AL, Hopkins MJ, Parry CM, Faragher EB, Zijlstra EE, Heyderman RS, Molyneux ME. Invasive non-typhoid salmonella establish systemic intracellular infection in HIV-Infected adults: an emerging disease pathogenesis. Clin Infect Dis. 2010;50(7):953–62. doi: 10.1086/651080. [DOI] [PubMed] [Google Scholar]
  • 5.Gilchrist JJ, MacLennan CA, Donnenberg MS, Bäumler AJ. Invasive nontyphoidal salmonella disease in Africa. EcoSal Plus. 2019;8(2) doi: 10.1128/ecosalplus.esp-0007-2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Feasey NA, Dougan G, Kingsley RA, Heyderman RS, Gordon MA. Invasive non-typhoidal salmonella disease: an emerging and neglected tropical disease in Africa. Lancet. 2012;379(9835):2489–99. doi: 10.1016/S0140-6736(11)61752-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Marchello CS, Fiorino F, Pettini E, Crump JA, Vacc-i NTSCC. Incidence of non-typhoidal Salmonella invasive disease: a systematic review and meta-analysis. J Infect. 2021;83(5):523–32. doi: 10.1016/j.jinf.2021.06.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Van Puyvelde S, Pickard D, Vandelannoote K, Heinz E, Barbe B, de Block T, Clare S, Coomber EL, Harcourt K, Sridhar S, et al. An African Salmonella Typhimurium ST313 sublineage with extensive drug-resistance and signatures of host adaptation. Nat Commun. 2019;10(1):4280. doi: 10.1038/s41467-019-11844-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Stanaway JD, Parisi A, Sarkar K, Blacker BF, Reiner RC, Hay SI, Nixon MR, Dolecek C, James SL, Mokdad AH, et al. The global burden of non-typhoidal salmonella invasive disease: a systematic analysis for the global burden of disease study 2017. Lancet Infect Dis. 2019;19(12):1312–24. doi: 10.1016/S1473-3099(19)30418-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Taneda K, Adachi T, Watanabe Y, Hanajima R. Cerebral venous thrombosis due to nontyphoidal Salmonella bacteremia. Intern Med. 2019;58(13):1943–6. doi: 10.2169/internalmedicine.2266-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.García-Fernández-Bravo I, González-Munera A, Ordieres-Ortega L, Ruiz Chiriboga D, González-Leyte M. Trombosis de aneurisma aórtico infrarrenal secundario a infección por Salmonella enteritidis. Revista de Gastroenterología de México. 2019;84(4):522–4. doi: 10.1016/j.rgmx.2018.12.002. [DOI] [PubMed] [Google Scholar]
  • 12.Prakash PS, Nomellini V, Kaplan LJ. Lower extremity Salmonella abscess with vascular thrombosis and wall destruction. Surg Infections Case Rep. 2016;1(1):79–81. doi: 10.1089/crsi.2016.0017. [DOI] [Google Scholar]
  • 13.Suppamutharwyam M, Radhakrishnan AP. An unusual case of deep vein thrombosis and mycotic aneurysms secondary to Salmonella bacteraemia. Case Rep Clin Med. 2022;11(3):37–47. doi: 10.4236/crcm.2022.113007. [DOI] [Google Scholar]
  • 14.Strutynskaya AD, Karnaushkina MA, Tyurin IE, Komarova IS. Salmonella sepsis in a patient with TLR4 gene polymorphisms. Radiol Case Rep. 2020;15(9):1545–51. doi: 10.1016/j.radcr.2020.05.076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Ghosh JB, Samanta S. Venous thrombosis in enteric fever. Indian Pediatrics. 1994;31(2):230–1. [PubMed] [Google Scholar]
  • 16.Takamatsu D, Bensing BA, Cheng H, Jarvis GA, Siboo IR, López JA, Griffiss JM, Sullam PM. Binding of the Streptococcus gordonii surface glycoproteins GspB and Hsa to specific carbohydrate structures on platelet membrane glycoprotein Ibα. Mol Microbiol. 2005;58(2):380–92. doi: 10.1111/j.1365-2958.2005.04830.x. [DOI] [PubMed] [Google Scholar]
  • 17.Tilley DO, Arman M, Smolenski A, Cox D, O’Donnell JS, Douglas CWI, Watson SP, Kerrigan SW. Glycoprotein Ibα and FcγRIIa play key roles in platelet activation by the colonizing bacterium, Streptococcus oralis. J Thromb Haemost. 2013;11(5):941–50. doi: 10.1111/jth.12175. [DOI] [PubMed] [Google Scholar]
  • 18.Arman M, Krauel K, Tilley DO, Weber C, Cox D, Greinacher A, Kerrigan SW, Watson SP. Amplification of bacteria-induced platelet activation is triggered by FcγRIIA, integrin αIIbβ3, and platelet factor 4. Blood. 2014;123(20):3166–74. doi: 10.1182/blood-2013-11-540526. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Keane C, Tilley D, Cunningham A, Smolenski A, Kadioglu A, Cox D, Jenkinson HF, Kerrigan SW. Invasive Streptococcus pneumoniae trigger platelet activation via Toll-like receptor 2. J Thromb Haemost. 2010;8(12):2757–65. doi: 10.1111/j.1538-7836.2010.04093.x. [DOI] [PubMed] [Google Scholar]
  • 20.Fumarola D, Miragliota G, Palma R, Panaro A. Experimental contribution to study of enterotoxigenic activity of some Salmonella Typhimurium strains - investigations with model of platelet aggregation by ADP. Annali Sclavo. 1977;19:1242–7. [PubMed] [Google Scholar]
  • 21.Matera C, Falzarano C, Berrino L, Rossi F. Effects of tetanus toxin, Salmonella Typhimurium porin, and bacterial lipopolysaccharide on platelet aggregation. J Med. 1992;23(5):327–38. [PubMed] [Google Scholar]
  • 22.Hoiseth SK, Stocker BAD. Aromatic-dependent Salmonella Typhimurium are non-virulent and effective as live vaccines. Nature. 1981;291(5812):238–9. doi: 10.1038/291238a0. [DOI] [PubMed] [Google Scholar]
  • 23.MacLennan CA, Gondwe EN, Msefula CL, Kingsley RA, Thomson NR, White SA, Goodall M, Pickard DJ, Graham SM, Dougan G, et al. The neglected role of antibody in protection against bacteremia caused by nontyphoidal strains of Salmonella in African children. J Clin Invest. 2008;118(4):1553–62. doi: 10.1172/JCI33998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kingsley RA, Msefula CL, Thomson NR, Kariuki S, Holt KE, Gordon MA, Harris D, Clarke L, Whitehead S, Sangal V, et al. Epidemic multiple drug resistant Salmonella Typhimurium causing invasive disease in sub-Saharan Africa have a distinct genotype. Genome Res. 2009;19(12):2279–87. doi: 10.1101/gr.091017.109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Kerrigan SW. In: The non-thrombotic role of platelets in health and disease. Kerrigan S, Moran N, editors. Vol. 4. IntechOpen; Rijeka: 2015. Platelet interactions with bacteria. [Google Scholar]
  • 26.Kerrigan SW, Cox D. Platelet–bacterial interactions. Cell Mol Life Sci. 2010;67(4):513–23. doi: 10.1007/s00018-009-0207-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Shannon O, Mörgelin M, Rasmussen M. Platelet activation and biofilm formation by Aerococcus urinae, an endocarditis-causing pathogen. Infect Immun. 2010;78(10):4268–75. doi: 10.1128/IAI.00469-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Petersson F, Kilsgard O, Shannon O, Lood R. Platelet activation and aggregation by the opportunistic pathogen Cutibacterium (Propionibacterium) acnes. PLOS ONE. 2018;13(1):e0192051. doi: 10.1371/journal.pone.0192051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Cattaneo M, Cerletti C, Harrison P, Hayward CP, Kenny D, Nugent D, Nurden P, Rao AK, Schmaier AH, Watson SP, et al. Recommendations for the standardization of light transmission aggregometry: a consensus of the working party from the platelet physiology subcommittee of SSC/ISTH. J Thromb Haemost. 2013;11(6):1183–9. doi: 10.1111/jth.12231. [DOI] [PubMed] [Google Scholar]
  • 30.Cattaneo M, Lecchi A, Zighetti ML, Lussana F. Platelet aggregation studies: autologous platelet-poor plasma inhibits platelet aggregation when added to platelet-rich plasma to normalize platelet count. Haematologica. 2007;92(5):694–7. doi: 10.3324/haematol.10999. [DOI] [PubMed] [Google Scholar]
  • 31.Ford I, Douglas CWI, Cox D, Rees DGC, Heath J, Preston FE. The role of immunoglobulin G and fibrinogen in platelet aggregation by Streptococcus sanguis. Br J Haematol. 1997;97(4):737–46. doi: 10.1046/j.1365-2141.1997.1342950.x. [DOI] [PubMed] [Google Scholar]
  • 32.Ford I, Douglas CWI, Heath J, Rees C, Preston FE. Evidence for the involvement of complement proteins in platelet aggregation by Streptococcus sanguis NCTC 7863. Br J Haematol. 1996;94(4):729–39. doi: 10.1046/j.1365-2141.1996.d01-1857.x. [DOI] [PubMed] [Google Scholar]
  • 33.Rosenfeld SI, Ryan DH, Looney RJ, Anderson CL, Abraham GN, Leddy JP. Human Fc gamma receptors: stable inter-donor variation in quantitative expression on platelets correlates with functional responses. J Immunol. 1987;138(9):2869–73. [PubMed] [Google Scholar]
  • 34.Tilley DO. Pathways to human platelet activation : convergence of streptococcus pneumoniae and streptococcus oralis at a universal actor, Fc gamma RIIA. School of pharmacy, Royal College of Surgeons in Ireland; Dublin: 2013. https://repository.rcsi.com/articles/thesis/Pathways_to_human_platelet_activation_convergence_of_streptococcus_pneumoniae_and_streptococcus_oralis_at_a_universal_actor_Fc_gamma_RIIA/10805768?file=19316612 . [Google Scholar]
  • 35.Corcoran PA, Atherton JC, Kerrigan SW, Wadstrom T, Murray FE, Peek RM, Fitzgerald DJ, Cox DM, Byrne MF. The effect of different strains of Helicobacter pylori on platelet aggregation. Can J Gastroenterol Hepatol. 2007;21(6):367–70. doi: 10.1155/2007/490852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Marchello CS, Birkhold M, Crump JA NTScc V-I. Complications and mortality of non-typhoidal salmonella invasive disease: a global systematic review and meta-analysis. Lancet Infect Dis. 2022;22(5):692–705. doi: 10.1016/S1473-3099(21)00615-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Beristain-Covarrubias N, Perez-Toledo M, Flores-Langarica A, Zuidscherwoude M, Hitchcock JR, Channell WM, King LDW, Thomas MR, Henderson IR, Rayes J, et al. Salmonella-induced thrombi in mice develop asynchronously in the spleen and liver and are not effective bacterial traps. Blood. 2019;133(6):600–4. doi: 10.1182/blood-2018-08-867267. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Hitchcock JR, Cook CN, Bobat S, Ross EA, Flores-Langarica A, Lowe KL, Khan M, Dominguez-Medina CC, Lax S, Carvalho-Gaspar M, et al. Inflammation drives thrombosis after Salmonella infection via CLEC-2 on platelets. J Clin Invest. 2015;125(12):4429–46. doi: 10.1172/JCI79070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Vidarsson G, Dekkers G, Rispens T. IgG subclasses and allotypes: from structure to effector functions. Front Immunol. 2014;5:520. doi: 10.3389/fimmu.2014.00520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Anania JC, Chenoweth AM, Wines BD, Hogarth PM. The human FcγRII (CD32) family of leukocyte FcR in health and disease. Front Immunol. 2019;10:464. doi: 10.3389/fimmu.2019.00464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Bruhns P, Iannascoli B, England P, Mancardi DA, Fernandez N, Jorieux S, Daeron M. Specificity and affinity of human Fcγ receptors and their polymorphic variants for human IgG sub-classes. Blood. 2009;113(16):3716–25. doi: 10.1182/blood-2008-09-179754. [DOI] [PubMed] [Google Scholar]
  • 42.Zeiler M, Moser M, Mann M. Copy number analysis of the murine platelet proteome spanning the complete abundance range. Mol Cell Proteomics. 2014;13(12):3435–45. doi: 10.1074/mcp.M114.038513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Gitz E, Pollitt AY, Gitz-Francois JJ, Alshehri O, Mori J, Montague S, Nash GB, Douglas MR, Gardiner EE, Andrews RK, et al. CLEC-2 expression is maintained on activated platelets and on platelet microparticles. Blood. 2014;124(14):2262–70. doi: 10.1182/blood-2014-05-572818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Trebicka E, Jacob S, Pirzai W, Hurley BP, Cherayil BJ. Role of antilipopolysaccharide antibodies in serum bactericidal activity against Salmonella enterica Serovar Typhimurium in healthy adults and children in the United States. Clin Vaccine Immunol. 2013;20(10):1491–8. doi: 10.1128/CVI.00289-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Wall NA, Dominguez-Medina CC, Faustini SE, Cook CN, McClean A, Jesky MD, Perez-Toledo M, Morgan MD, Richter AG, Ferro CJ, et al. Humoral immunity to memory antigens and pathogens is maintained in patients with chronic kidney disease. PLoS One. 2018;13(4):e0195730. doi: 10.1371/journal.pone.0195730. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Mastroeni P, Rossi O. Antibodies and protection in systemic salmonella infections: do we still have more questions than answers? Infect Immun. 2020;88(10) doi: 10.1128/IAI.00219-20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.MacLennan CA, Gilchrist JJ, Gordon MA, Cunningham AF, Cobbold M, Goodall M, Kingsley RA, van Oosterhout JJ, Msefula CL, Mandala WL, et al. Dysregulated humoral immunity to nontyphoidal Salmonella in hiv-infected African adults. Science. 2010;328(5977):508–12. doi: 10.1126/science.1180346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Dalby T, Strid MA, Beyer NH, Blom J, Molbak K, Krogfelt KA. Rapid decay of Salmonella flagella antibodies during human gastro-enteritis: a follow up study. J Microbiological Methods. 2005;62(2):233–43. doi: 10.1016/j.mimet.2005.02.006. [DOI] [PubMed] [Google Scholar]
  • 49.Strid MA, Dalby T, Molbak K, Krogfelt KA. Kinetics of the human antibody response against Salmonella enterica Serovars Enteritidis and Typhimurium determined by lipopolysaccharide enzyme-linked immunosorbent assay. Clin Vaccine Immunol. 2007;14(6):741–7. doi: 10.1128/CVI.00192-06. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Figures 1-2
EMS205385-supplement-Supplementary_Figures_1_2.TIF (896.5KB, TIF)
Supplementary Figures 3-4
EMS205385-supplement-Supplementary_Figures_3_4_.TIF (1.2MB, TIF)

RESOURCES