Background
Cerebrospinal fluid shunt placement for the treatment of hydrocephalus is one of the most common procedures performed by pediatric neurosurgeons in the USA, with tens of thousands of shunts implanted annually [
1]. Unfortunately, 30–40 % of all shunts placed in pediatric patients fail within the first year, resulting in a shunt revision to a primary placement ratio of 3:1 in many health-care centers [
1,
2]. One of the most common causes of shunt failure is infection, reported in 5–30 % of cases [
2]. This equates to approximately 2400 hospital admissions each year, with an increase in seizure frequency and intelligence quotient (IQ) loss in many children, who represent the highest risk population for these infections [
2‐
4]. The most common organism responsible for shunt infections,
Staphylococcus epidermidis (
S. epidermidis), is known to form biofilms, which are communities of bacterial cells that aggregate on the catheter surface, encased in a protective self-produced matrix [
2,
5]. The biofilm’s recalcitrance to antimicrobial agents makes it difficult to manage central nervous system (CNS) catheter infections non-surgically, such that catheter removal is currently required for effective treatment [
2,
6]. Studies designed to advance our understanding of the unique immune response to shunt infections in children are needed to develop improved diagnostic, treatment, and prevention strategies for these serious infections.
Our laboratory developed a murine model of CNS catheter infection to generate a consistent catheter-associated infection with
Staphylococcus aureus (
S. aureus), mimicking what is seen in humans with ventricular shunt infections [
7]. Using this model, we have been able to demonstrate a relative decrease in inflammation in biofilm infections in the brain, as compared to abscess or infection with biofilm-deficient strains of bacteria [
8]. This suggests that there is an alteration in the immune response to biofilm infections, possibly contributing to the persistence of these infections in patients. One potential cause of the skewed immune response to these biofilm infections is an increase in interleukin-10 (IL-10), which may play a role in regulating inflammatory responses in this setting. IL-10 is a potent anti-inflammatory cytokine that can be produced by both innate and adaptive immune cells [
9]. IL-10 polymorphisms in humans and mice have been associated with an increase in autoimmune disease, such as inflammatory bowel disease, atopic dermatitis, and wheezing, as well as an increased inflammatory response to some microbial pathogens [
9‐
12]. Interestingly, this increase in inflammatory response to microbes does not necessarily result in increased clearance of the organism, suggesting that the role of IL-10 in response to infection is likely pathogen-specific [
9]. Very little has been studied about the role of IL-10 in response to
S. epidermidis infection or biofilm infections specifically, but elevated levels of IL-10 have been associated with poor outcomes in patients with
S. aureus bacteremia, suggesting this cytokine may play a role in staphylococcal disease [
12]. In these studies, we have adapted our mouse model of
S. aureus CNS catheter infection to generate infection with
S. epidermidis and have used this model to define the contribution of IL-10 to the inflammatory response to CNS catheter infection. Our results show that alterations in IL-10 result in increases in inflammatory mediators and infiltration of peripheral immune cells but do not significantly impact bacterial burdens.
Discussion
S. epidermidis and
S. aureus are the most common biofilm-forming bacteria responsible for CSF shunt infections [
19]. These infections hinder the effectiveness of shunt therapy and inflict an unnecessary burden in the treatment of pediatric patients, who are known to be at higher risk for shunt infection, although the mechanisms responsible for this increased risk in children are not yet defined [
19]. Biofilm infections are known to be difficult to diagnose and recalcitrant to antibiotic treatment, leading to removal of the shunt [
19]. This increases the numbers of surgical procedures needed for patients, in addition to increasing future infection risks as revision itself is associated with an increased risk of shunt infection [
19,
20]. Infants are known to have the most adverse outcomes, such as increased risk of seizures, IQ loss, and up to 10 % death rate in patients with neurological handicap [
21,
22]. Shunt infections also lead to disproportionate use of hospital days and health-care dollars [
23]. Collectively, these findings highlight the significance of CNS shunt infection and the need to better understand the pathophysiology of the disease with the goal of informing better treatment strategies that would avoid the need for shunt replacement surgery.
In these studies, we adapted our established model of
S. aureus CNS catheter infection to generate infection with the more prevalent, but less virulent
S. epidermidis [
14,
24]. While
S. epidermidis is the cause of most CNS catheter infections, its primary virulence property is its ability to form biofilm, as it lacks the production of virulence factors seen in
S. aureus [
14]. Catheter-adherent biofilm loads registered peak values at day 3 post-surgery compared to immediately adjacent tissue in both models, consistent with biofilm formation and planktonic spread of the infection. As we have only evaluated the immediately adjacent tissue in these studies, we may underestimate bacterial spread throughout the brain. In the
S. aureus model of infection, this pattern of bacterial growth was visually consistent with biofilm formation on electron microscopy [
7]. Interestingly, both infection models demonstrate a shift to greater parenchymal involvement as overall bacterial burdens lessen in later days of infection [
7]. Differences between the pathogenic natures of these two closely related species were noted by the earlier decrease in bacterial burdens and reduced mortality rate in
S. epidermidis-infected mice compared to the prolonged course of infection and greater casualties in the
S. aureus model [
7]. Additionally, higher colony counts were necessary to achieve effective infecting doses with
S. epidermidis (6 × 10
7 cfu/ml) compared to
S. aureus (8.7 × 10
3 cfu/ml) [
7]. These findings are consistent with the increased morbidity attributable to
S. aureus in clinical settings and reflect the multiplicity of virulence factors known to be present in
S. aureus compared to
S. epidermidis [
25]
.
These studies also demonstrate a distinct inflammatory profile with greater evidence of attenuation in response to
S. epidermidis biofilm infection in the CNS than was seen in our prior studies with
S. aureus [
7]
. As in our prior studies with
S. aureus, we observed a brief increase in inflammatory mediators at early time points following the implantation of sterile catheters which likely results from tissue damage; importantly, this inflammation is consistently less than observed in response to infected catheters [
7,
8]. There is an increase in pro-inflammatory IL-1β, IL-6, CXCL1, and CXCL2 with
S. epidermidis, demonstrating that implant-mediated
S. epidermidis infection elicits a significant inflammatory response in the brain, similar to prior work with
S. aureus [
7]. However, we also observed a small increase in the anti-inflammatory IL-10 and a lack of pro-inflammatory IL-12p70, suggesting that there are also anti-inflammatory pathways involved in the response to
S. epidermidis CNS catheter infection. This is similar to studies of peripheral catheter-associated infections which have demonstrated anti-inflammatory immune responses to biofilm infections, suggesting a role for biofilm structure to modulate host response [
26]. Our prior studies in the brain also suggest an attenuated response to biofilm infection, with decreased levels of pro-inflammatory mediators observed when comparing biofilm infection to both parenchymal abscess infections and catheter-associated infection with a biofilm-deficient
sarA mutant [
8]. Future studies may utilize biofilm-deficient
S. epidermidis mutants, which would allow us to better dissect if there is a distinctive effect of biofilm formation in the immune response to
S. epidermidis as well. However, this approach may not be feasible as biofilm-deficient
S. epidermidis may not be able to establish device-associated infection.
Herein, we report increased IL-10 levels in infected mice following
S. epidermidis CNS catheter infection compared to the non-infected controls. This suggests a possible effect of IL-10 in curbing the CNS inflammatory environment to allow non-pathogenic
S. epidermidis to cause infection [
27,
28]. IL-10 plays a central role in chronic and acute inflammation, particularly in the CNS where prolonged contact with pro-inflammatory cytokines leads to harmful effects on neuronal function, behavior, and cognition [
27,
29‐
32]. To further study the role of IL-10 in this model, IL-10 KO mice were infected to evaluate the bacterial kinetics and immune response in the absence of this important immune regulator. Surprisingly, IL-10 deficiency had no impact in bacterial accumulation in the catheter or immediately adjacent tissue in the first 7 days post-implantation. IL-10 may play a role in bacterial clearance in later dates post-infection, which represents a limitation of the current studies. As expected IL-1β, IL-6, CXCL1, and CXCL2 levels were increased in IL-10 KO mice. This increase in pro-inflammatory mediators likely led to the increase in weight loss in the IL-10 KO mice, as IL-1β and IL-6 have been shown to play a role in sickness behaviors [
33,
34]. TNF-α also plays an important role in sickness behaviors but was not evaluated in these studies [
35,
36]. Heightened levels of inflammatory cytokines and chemokines in IL-10 KO mice demonstrate that IL-10 is important in controlling immune response in this model. However, this data also shows the limitations of enhancing pro-inflammatory responses in an effort to treat CNS catheter infections, as the increased inflammation resulted not in improved bacterial clearance but rather in increased clinical illness.
While our studies have demonstrated that IL-10 is not solely responsible for dampening the immune response to permit infection to occur, it does play a role in modulating the immune response to
S. epidermidis infection in the CNS. This may play a particularly significant role in infants, who are known to be at higher risk of CNS catheter infections [
2,
4,
19]. In humans, IL-10 appears to play an important protective role in infants, with altered IL-10 levels associated with inflammatory conditions such as bronchopulmonary dysplasia, sepsis, seizures, and colitis [
37‐
40]. Additionally, cord blood studies suggest that the neonatal immune system is primed toward increased IL-10 production, with increased IL-10 release from mononuclear cells following exposure to neonatal plasma and increased IL-10 expression from infant monocytes following TLR stimulation [
41,
42]. These in vitro studies suggest that IL-10 may play an important role in regulating the immune response to infection in young hosts, which is currently being evaluated in vivo in our laboratory. Importantly, while this disease is more common in children, it is not exclusive to children and there may be distinct differences between the adult and infant responses to CNS infection such that investigation of both adult and infant models is needed.
Defining the contribution of IL-10 to pathology in vivo is essential as this will allow us to account for both the multiple antigens and evasion capabilities of clinically relevant bacteria such as S. epidermidis, as opposed to single antigens, in addition to the impact of immune responses on multiple cell types within the brain. The unique milieu of the CNS includes multiple mechanisms of cross talk between glia and peripheral immune cells as a means of protecting the CNS from inflammatory damage as well as invading pathogens and evaluation in vivo allows for the best opportunity to model these complex interactions. It is reasonable to suspect that IL-10 will play a greater role in immune regulation in infant hosts, given its significant role in neonatal immunity, and future studies in our laboratory will expand the current studies to this population.
Acknowledgements
The authors thank Dr. Paul Fey at the University of Nebraska Medical Center for providing the S. epidermidis strain used herein. We also thank Philip Hexley and Victoria Smith in the UNMC Cell Analysis Facility for the assistance with the fluorescent-activated cell sorting (FACS) analysis. We would like to thank Tammy Kielian, PhD, for her assistance in the review and revision of this manuscript.