An immune mechanism that makes babies more likely than adults to die from sepsis has been identified by scientists affiliated with the Center for Research on Inflammatory Diseases (CRID in Ribeirão Preto, São Paulo State (Brazil).
The study is published) in Critical Care.
The scientists are planning to test new therapeutic approaches based on the discovery.
“We’re designing a clinical trial with drugs that have been approved for human use and are known to induce this immune mechanism.
The goal is to improve the survival rate for infants with sepsis,” said Fernando de Queiroz Cunha, CRID’s principal investigator. CRID is one of the Research, Innovation and Dissemination Centers (RIDCs funded by São Paulo Research Foundation – FAPESP.
Sepsis (sometimes referred to as blood poisoning) is systemic inflammation usually triggered by a localized bacterial infection that spins out of control.
The body’s immune response to combat the pathogen ends up damaging multiple organs and tissues.
Symptoms include fever or low temperature, difficulty breathing, low blood pressure, a fast heart rate, and an abnormally high or low white blood cell count.
The condition may remain active even after the initial threat has been eliminated.
Its most severe form can lead to lesions that impair the function of vital organs, septic shock and death.
“In any experimental animal model of sepsis, all the parameters used to measure the severity of the condition are higher in infants.
There’s more systemic inflammatory response, more organ impairment, and higher mortality,” said Cunha, who is a Full Professor in the Department of Pharmacology at the University of São Paulo’s Ribeirão Preto Medical School (FMRP-USP).
In humans, it is more difficult to compare infant and adult mortality rates, he explained, because, before contracting sepsis, the adult patient may have been weakened by diseases such as diabetes, cancer, heart failure or hypertension (high blood pressure).
“Most adults who die as a result of septic shock already had serious health problems,” Cunha told.
Given their knowledge that organ injury is more severe in young individuals, both human and murine, the group decided to determine exactly what substances are produced by the immune system during sepsis.
Their hypothesis was that defense cells in infants must produce more oxidizing substances, such as oxygen and nitrogen free radicals.
What they found, however, was the opposite.
“We took a long time to understand why infants have more tissue injury if they produce smaller amounts of free radicals.
Finally, we decided to investigate NETs [neutrophil extracellular traps],” Cunha said.
Neutrophils are white blood cells that form the front line of the immune system, phagocytosing (killing) bacteria, fungi and viruses.
NETs are structures composed of DNA and granular proteins that rapidly trap and kill pathogens.
“This immune mechanism was first described about ten years ago.
In some situations, for poorly understood reasons, the immune system activates an enzyme called PAD-4, which increases the permeability of the neutrophil nucleus.
When this happens, the genetic material in the nucleus decondenses and forms networks, which are released by the cell into the extracellular medium to trap and kill bacteria,” Cunha said.
NETs are typically activated by bacterial infections, he added, as well as some viruses, including chikungunya, the arbovirus that causes the most tissue injury.
The mechanism also occurs in some autoimmune disorders.
“The main problem is that NETs aren’t just toxic for pathogens: they also damage human cells.
In fact, they do more damage than oxygen and nitrogen free radicals.”
Tests involving pediatric patients were conducted in collaboration with a research group led by Professor Ana Paula Carlotti, attached to the ICU at FMRP-USP’s teaching and general hospital (Hospital das Clínicas).
Laboratory analysis showed that neutrophils from infants produced 40 percent more NETs than those taken from adults, in the case of humans.
The difference was 60 percent in mice.
The group then set out to use experimental models to understand how this immune mechanism works in sepsis.
Traps deactivated
The experiments with mice involved a group of two-week-old infants and a group of healthy young adults.
Both received an intraperitoneal injection of intestinal bacteria and developed sepsis.
“A dose of bacteria sufficient to kill 100 percent of infants killed only 50 percent of the adults.
That’s a significant difference.
Moreover, in the days following the injection, the infant mice displayed higher levels of bacteremia [bacteria in the bloodstream] and of biochemical markers indicating organ injury,” Cunha said.
When NETs were broken down with recombinant human DNase (a drug used to treat cystic fibrosis), the survival rate jumped from 0 to 50 percent in the infant group.
In the adult group, the proportion of mice that survived sepsis rose from 50 percent to 60 percent.
“The difference between the groups when treated with DNase was small, clearly showing that greater infant susceptibility is associated with higher levels of NETs,” Cunha said.
In another experiment, the group replaced DNase with a compound designed to inhibit PAD-4, the enzyme that triggers the activation of NETs.
In this case, the survival rate for the infant group was 40 percent.
“It was somewhat less effective than DNase because it’s not actually a specific PAD-4 inhibitor.
One of our goals for future research is the development of a specific drug to inhibit PAD-4,” Cunha said.
The group analyzed the expression of the PAD-4 gene, which encodes the PAD-4 enzyme, in neutrophils from patients and from mice.
In both cases, PAD-4 expression was higher in infants with sepsis than in adults with the same condition.
The reasons are unknown and are currently being sought by David Fernando Colón Morelo, the first author of the article. Cunha is Morelo’s Ph.D. supervisor.
Morelo has a doctoral scholarship from FAPESP and is now doing a research internship at Bonn University in Germany.
“We’re also studying the role of NETs in other diseases involving organ injury, such as rheumatoid arthritis and lupus,” Cunha said.
Neutrophils are the most abundant white blood cells in the circulation and serve antimicrobial functions.
One of their antimicrobial mechanisms involves the release of neutrophil extracellular traps (NETs), long chromatin fibers decorated with antimicrobial granular proteins that contribute to the elimination of pathogens.
However, the release of NETs has also been associated with disease processes.
While recent research has focused on biochemical reactions catalyzed by NETs, significantly less is known about the mechanical effect of NETs in circulation.
Here, microfluidic devices and biophysical models are employed to study the consequences of the interactions between NETs trapped in channels and red blood cells (RBCs) flowing in blood over the NETs.
It has been found that the RBCs can be deformed and ruptured after interactions with NETs, generating RBC fragments.
Significant increases in the number of RBC fragments have also been found in the circulation of patients with conditions in which NETs have been demonstrated to be present in circulation, including sepsis and kidney transplant.
Further studies will probe the potential utility of RBC fragments in the diagnostic, monitoring, and treatment of diseases associated with the presence of NETs in circulation.
Neutrophils are the most abundant immune cells in the circulation and are the first line of defense during infections. Neutrophils employ several strategies to eliminate invading microbes including the engulfing of pathogens, production of antimicrobial compounds, and formation of neutrophil extracellular traps (NETs).[1]
The NETs consist primarily of DNA fibers, histones, and granule proteins.
They could immobilize pathogens in an environment rich in antimicrobial compounds.[2–4]
However, in addition to the antimicrobial role, it is also known that NETs contribute to the pathology of several diseases, such as sepsis,[5,6] asthma,[7] thrombosis,[8–10]lupus erythematosus,[11] type 2 diabetes,[12] cystic fibrosis.[13–15] Although NETs provide an important mechanism for clearing pathogens, excessive NETs release can lead to organ failure and even death.[16]
Moreover, NETs release and accumulation within blood vessels have been described recently[17] as well as the ability of NETs in circulation to block capillaries, impair the microcirculation, damage tissues, and promote inflammation.[6,18]
A microfluidic device designed to mimic a capillary plexus has shown that NETs can block capillary networks in vitro and divert RBCs flow, resulting in large areas void of RBCs.[19] Interestingly, almost half a century ago, the mechanical interaction between RBCs and fibrin fibers was reported to damage the RBCs, resulting in schistocyte formation.[20]
It was proposed that schistocytes are generated when there is damage to the blood vessel and a clot begins to form. The formation of the fibrin strands in the vessels, as part of the clot formation process, can trap red blood cells and the blood flow can shear the trapped RBCs.[20]
Schistocytes are often found in the blood of patients with hemolytic anemia.
It was also reported that the schistocyte count of >1% is suggestive of disseminated intravascular coagulation, which is caused by a systemic response to a specific condition including sepsis. It is likely that the RBC fragments are cleared out by the spleen. Recent experiments in animals have shown that schistocytes are present in the blood in the absence of DNase to degrade NETs.[21]
Together, these observations led us to hypothesize that RBCs may be damaged by NETs trapped in arteriole and capillary networks.
If this is true, a correlation may exist between RBC damage and diseases associated with the presence of increased amounts of NETs in the circulation. Such large amounts of NETs have been measured in the blood from various patients[6,14,16] and it is possible that they contribute to the larger frequency of schistocytes in these patients.
Here, we designed a microfluidic device to capture chromatin fibers and to study the mechanical interactions between RBCs and chromatin fibers under flow.
We employed chromatin fibers both in the form of NETs released from human neutrophils as well as ligated lambda DNA (λ DNA), a simplified biomimetic model that shares mechanical features with NETs but without the associated enzymes and histones.
We found that λ DNA and NETs when trapped inside channels can damage RBCs, producing RBC fragments. We also used a finite element method to study the effect of NETs on trapped RBCs and calculated shear stresses compatible with the RBC damage. Moreover, we measured the presence of larger numbers of RBC fragments in blood samples from ICU patients with and without sepsis as well as patients receiving immunosuppressive therapy posttransplant compared to healthy controls.
A better understanding of the interaction between excessive NETs release and circulating RBCs will enhance our ability to diagnose, monitor, and treat chronic inflammatory processes associated with NETs release in the circulation such as sepsis.
We designed devices with arrays of 50 μm diameter posts spaced at various distances inside microscale channels (10, 20, 40, and 80 μm). We employed these devices to mechanically trap chromatin fibers, either NETs or λ DNA, across the post arrays (Figure 1a,b). We imaged the trapped chromatin fibers, using a DNA intercalating dye as well as scanning electron microscopy (SEM) (Figure 1c-e).
The amount of chromatin trapped by the posts is significantly larger than the amount of chromatin that may stick to the glass.
Mechanical entanglement of chromatin fibers on post-arrays in microfluidic chips. a,b) Schematics of the microfluidic device for in vitro studies of RBCs interaction with NETs. The cross-section of channels is 1066 × 35 μm. The diameter of the posts is 50 μm. The spacing between posts in different designs ranged from 10 to 80 μm. c,d) Fluorescence images show λ DNA (c), and NETs (d) trapped by posts inside the device and stained with Sytox orange dye. Both λ DNA and NETs wrap around the posts in the microfluidic device. Images were recorded under flow conditions, at 40 μL min−1 flow rate. Scale bars are 100 μm. e) SEM representative image of NETs inside a microfluidic device. Scale bar 5 μm. f) Bright field images extracted from a high-speed video (captured at 1000 frames per second) and show the breaking of a red blood cell trapped between λDNA under flow conditions (40 μL min−1) in the microfluidic device over 50 ms. The λ DNA fibers and trapped RBC fragments are visible in Movie S1 in the Supporting Information.
Following entrapment of chromatin fibers, we injected human blood through the microfluidic device, at 40 μL mi−1 flow rate. We monitored the integrity of RBCs and found that RBCs are mechanically damaged after the interactions with the chromatin fibers. High-speed microscopy documented directly the trapping of RBCs inside the chromatin fibers network, followed by retention, deformation, and fragmentation (Figure 1f and Movie S1, Supporting Information).
Using finite element method with Comsol Multiphysics, we studied the effect of NETs on the shear stress around trapped cells. Without NETs, RBCs follow the flow lines between the posts. They experience minimal shear stress due to negligible difference between their velocity and the velocity of the fluid.
In the presence of NETs, two processes take place at the same time. Firstly, RBCs are trapped in the NETs. The trapped RBCs experience very high shear stress due to the flow of blood around them. Secondly, NETs and RBCs can block the spaces between posts diverting the flow through the unblocked gaps between posts.
This situation leads to a higher shear stress on the RBCs trapped in those regions.
To calculate shear stress around trapped RBCs, we modeled RBCs as biconcave discoid and simulated different blocking configurations, as shown in Figure 2f.
Readers are referred to Experimental Section for more discussion on the computational model and its validation. Figure 2-e shows contour plots of shear stresses around a trapped RBC as the percentage of the blocked interpillar channels increases.
In these simulations, interpillar spacing was 20 μm, and the flow rate was 40 μL min−1.
In the absence of NETs, the maximum wall shear stress was ≈60 Pa which is well below the threshold shear stress of 150 Pa for hemolysis.[22–24]
Additionally, cells moving with the flow will experience even, lower shear stress. With 50% of the interpillar channels blocked, maximum shear stress rises to 100 Pa around the mid-plane of a trapped cell. When 75% of the interpillar channels are blocked, maximum shear stress goes up to ≈200 Pa. At this shear stress, 10 min of exposure will ensue hemolysis.[23,25] Our results are consistent with previous studies showing RBCs damage after interactions with fine fibrin strands.[20]
The shear stress around RBCs increases with the reduction of interpillar channel size (Figure 2g) and can reach as high as ≈1268 Pa (in devices with 10 μm interpillar gap when 90% interpillar channels are blocked).
Computational simulation of velocity and shear stress around trapped RBCs with NETs. a) Contour plot of the velocity field between pillars from the computational simulations. b) Corresponding experimental plot of the velocity field measured using μ-PIV. c) Contour plot of shear stress when NETs are not present in the device. Maximum shear stress is ≈60 Pa. d) The effect of NETs on shear stress around trapped RBCs. NETs are modeled by blocking interpillar channels. Contour plot of shear stress around the mid-plane of an RBC when 50% of the interpillar channels are blocked. Maximum shear stress is ≈100 Pa. The interpillar spacing is 20 μm and the flow rate is 40 μL min−1. e) Contour plot of shear stress around the mid-plane of an RBC when 75% of the interpillar channels are blocked. Maximum shear stress is ≈200 Pa. f) Schematics of the 3D model of the device used in finite element simulations when 50% of the interpillar channels are blocked. Inset Figure shows one RBC modeled as a biconcave discoid. Schematic of the devices used in the finite element simulations corresponding to 0, 50, and 75% blocking. g) Increase in shear stress around RBCs for different interpillar gaps as the percentage of the blocked channels increases. Shear stress is ≈1268 Pa in a 10 μm interpillar gap device.
We also employed imaging flow cytometry and examined the presence of RBC fragments in the effluent from the microfluidic devices.
We distinguished the platelets from RBC fragments and intact RBCs as well as other small vesicles by size and the binding of fluorescent antibodies against specific cell surface markers. We found a significant increase in the number of RBC fragments in the blood after passage through chromatin fibers mechanically trapped in the microfluidic channels (Figure 3a,b).
We observed increases in the number of RBC fragments both in the presence of NETs and λ DNA. While the amount of DNA introduced in the devices as NETs and λ DNA were similar, we noted differences in the number of RBC fragments generated. One possible explanation for these differences is the organization of the fibers over the posts, where the λ DNA appears to form smaller number of thicker bundles (Figure 1c,d).
Additionally, we calculated that the decrease in the fraction of intact RBCs matches the increase in RBC fragments in the presence of chromatin fibers. This result suggests that RBC fragments were generated from intact RBCs, after interacting with chromatin fibers. We also found that a decrease in post distance from 80 to 10 μm increases the number of RBC fragments by 9.5% (Figure 3c).
Quantification of RBC fragments in blood. a) Imaging flow cytometry representative scatter plots showing the populations of RBCs and RBC fragments after passing blood through the microfluidic device with 10 μm—spaced posts without chromatin fibers (control), with λ DNA, and with NETs. b) Representative images of one RBC and two RBC fragments using fluorescence (CD235a+) and bright field. Scale bar 10 μm. c) The percentage of RBC fragments relative to the total number of CD235a+ events decreases with the increased spacing between posts. Three conditions are compared: control (blue), λ DNA (red), and NETs (green), (N = 4 experimental repeats). Two-way ANOVA followed by Dunnett’s multiple comparisons test. *P < 0.05; ****P < 0.0001. d) The concentration of free hemoglobin released in plasma decreases with the increasing spacing between posts. Empty controls (blue), λ DNA (red), and NETs (green) conditions are compared. (N = 4 experimental repeats). Two-way ANOVA followed by Dunnett’s multiple comparisons test. **P < 0.01; ***P < 0.001; ****P < 0.0001. e) The number of RBC fragments (blue square) remains constant for up to 48 h (left axis) while the free hemoglobin in plasma (red circle) becomes undetectable within hours (right axis). The interpillar spacing is 20 μm, the flow rate is 40 μL min−1, and blood samples were kept at 4 °C over 48 h.
We measured an increase in the free hemoglobin concentration in the blood after passage through microfluidic devices when chromatin fibers are entrapped in the post arrays (Figure 3d). The source of this hemoglobin is likely the RBC fragmentation.
When the integrity of RBC membrane is transiently disrupted, some hemoglobin can leak out from RBCs. In the absence of chromatin, the free hemoglobin concentration remained between 4 and 5 mg dL−1 for all devices with different post distances. However, in the presence of NETs, the free hemoglobin concentration increases with decreased post distance (Figure 3d).
This increase in free hemoglobin correlates with the increase in RBC fragments. We also found that the concentration of free hemoglobin of the blood passed through the microfluidic devices decreases quickly over time. The effect is likely due to the neutralization of free hemoglobin by proteins in the whole blood, as measured over 48 h (Figure 3e).
While the number of RBC fragments does not change over the same time, the RBC fragments may represent a better indicator of RBC damage compared to the free hemoglobin.
In addition to RBCs, other blood cells have a high chance to get damaged when passing through a device after NETs trapping
Although neutrophils are known to be able to squeeze in minutes through narrow spaces during chemotaxis, the deformation inside the devices under flow is significantly faster, in the order of milliseconds.
The shorter time may not allow neutrophils to deform sufficiently to pass through the NETs mesh. Moreover, this deformation is under the effect of external forces, rather than due to active remodeling of the actin cytoskeleton.
Thus, similar to RBCs, neutrophils can be damaged by high shear forces, releasing nuclear chromatin and small particles.
The release of chromatin from damaged neutrophils may further contribute to the increasing amount of NETs trapped in the microfluidic device. We tested this possibility by passing neutrophils through the microfluidic device after trapping NETs across the posts. We found that the area of trapped NETs increased over time, indicating neutrophils were sheared, releasing chromatin that was trapped by the posts.
We utilized imaging flow cytometry to compare the geometry of neutrophils before and after injection through the microfluidic device in the presence of trapped NETs.
The neutrophils were stained with anti-human CD66b+ antibody. We found that a significant number of neutrophils was damaged after passing through the trapped NETs, resulting in deformed neutrophils and a large number of particles carrying the CD66b neutrophil markers (Figure S2e-f, Supporting Information). Other white blood cells may undergo similar damages when flowing into trapped NETs.
However, besides the qualitative similarities, one would have to consider that the white blood cells are present in the blood at 1000-fold smaller concentration compared to RBCs. Thus, at a similar rate of damage, the number of white blood cells derived particles will always be in the orders of magnitude lower than the number of RBC fragments.
We explored the presence of RBC fragments in blood derived from healthy individuals (Table S1, Supporting Information) and patients with different medical conditions, including ICU patients without sepsis or with sepsis (Table S2, Supporting Information), and posttransplant patients (Table S3, Supporting Information) using imaging flow cytometry. We found that the number of RBCs fragments from ICU patients with sepsis was higher compared to ICU patients without sepsis and to healthy controls (Figure 4).
The number of RBC fragments was also higher in patients receiving immunosuppressive therapy after an organ transplant, compared to healthy individuals (Figure 4).
The increase in RBC fragments in these patients is intriguing considering that the immunosuppressive treatment can lower the leukocyte counts,[26] reducing the presence of NETs in circulation. One potential explanation takes into account the role of chronic inflammation in activating the neutrophils and the release of NETs.[16]
The number of patient samples analyzed in this study is small and therefore more extensive studies are necessary to clarify if the differences between RBC fragments in different patient groups correlate with clinically-relevant endpoints. While the patients and healthy individuals in our pilot measurements were not matched for age and sex, further studies will also have to take these variables into account.
Quantification of RBC fragments in patient samples, using flow cytometry. The fraction of RBC fragments increases in ICU patients with and without sepsis and kidney transplant patients compared to healthy individuals. The vertical axis represents the percentages of RBC fragments relative to the total number of CD235a positive events in blood. N = 7 samples from N = 7 healthy donors, N = 6 samples from N = 5 ICU patients without sepsis, N = 13 samples from N = 9 ICU patients with sepsis, and N = 4 samples from N = 3 kidney transplant patients. Samples from the same patient were collected on different days and assumed to be independent. One way ANOVA followed by Dunnett’s multiple comparisons test. *P < 0.05.
n summary, we employed devices with microfluidic channels that mimic mechanical features of capillary networks and studied the interaction between mechanically trapped chromatin fibers and circulating RBCs.
We found that the flow of blood over trapped NETs could damage the RBCs and generate significant numbers of RBCs fragments. Our in vitro experiments and finite element simulations validate the role of higher shear stresses (>150 Pa) in the RBC damage by trapped chromatin fibers.
The flow cytometry studies of patient blood samples indicate that the presence of RBC fragments could be more ubiquitous in patients than previously considered.
Further studies will clarify if any correlations exist between the presence of RBC fragments in the circulation, release of NETs in the circulation and the pathological processes associated with the NETs trapping in the vasculature.
More information: David F. Colón et al, Neutrophil extracellular traps (NETs) exacerbate severity of infant sepsis, Critical Care (2019). DOI: 10.1186/s13054-019-2407-8
Journal information: Critical Care
Provided by FAPESP
