Distinct cytokine patterns may regulate the severity of neonatal asphyxia—an observational study

Neonatal asphyxia evokes the injury of the central nervous system (CNS) due to the severe lack of oxygen and perfusion during labor and delivery, resulting in moderate to severe neurological dysfunction. Asphyxia primarily affects term and post-term neonates. It occurs in 2–4 of every 1000 live-born term neonates and is responsible for approximately 23% of neonatal deaths worldwide [1]. While some surviving children show favorable neurological outcome, others sustain severe neurodevelopmental disabilities such as mental retardation, sensory impairment, cerebral palsy, and seizures [2]. Identifying the factors responsible for such extent of individual variability regarding outcome would be critical; however, to date, no definitive predictive factors have been validated.

Inflammation of the CNS, or neuroinflammation, is now recognized to be a feature of all neurological disorders, including that related to neonatal asphyxia. Microglia and astrocytes become activated and release pro-inflammatory cytokines and chemokines. Disruption of the blood-brain barrier allows infiltration of peripheral monocytes into the brain that further enhances the inflammatory response, leading to neuronal injury and apoptosis. However, the inflammatory reaction following asphyxia is not limited to the CNS, but can also be detected in the periphery. Systemic immune activation is characterized by increased synthesis of pro-inflammatory cytokines [3]. A key player in the mediation of the inflammatory response both in the brain and peripheral blood during asphyxia is the subset of T lymphocytes. T lymphocytes have a pivotal role in the evolution of hypoxic injury. The mechanisms by which T cells are neurotoxic include the production of perforin and granzyme B, the release of free radicals, the triggering of apoptotic pathways within neurons, and most importantly, the production of pro- and anti-inflammatory cytokines [4, 5].

An extensive dataset describes neuroinflammation to have detrimental consequences, but results have indicated over the past decade that some aspects of the inflammatory response are beneficial for the CNS outcomes [6, 7]. Benefits of neuroinflammation include neuroprotection, the mobilization of neural precursors for repair, remyelination, and axonal regeneration. In vitro studies demonstrated that pro-inflammatory cytokines, such as TNF-α and IFN-γ, are toxic for oligodendrocytes [8‐10]. Although inflammatory cytokines contribute to injury progression, they also play a vital role in the fast elimination of cellular debris, and in the processes of growth and repair, contributing to functional recovery [11, 12]. The results of Saliba et al. support the positive role of certain cytokines in neuronal regeneration [13]. In addition to its toxic effect, TNF-α also plays a role in neuronal progenitor cell proliferation, lineage commitment, and cellular differentiation. IL-1 also has neurotrophic properties which might be mediated by the stimulation of nerve growth factor production. Direct intracerebral injection of IL-1 or TNF-α has been shown to stimulate astrogliosis and angiogenesis in the developing rodent brain [14]. The TGF-β family consists of pleiotropic proteins with potent immunoregulatory properties, which might also play key roles in the development, repair, and survival of neurons [13].

Previous investigations in asphyxia demonstrated that pro-inflammatory IL-1β, TNF-α, and IFN-γ play an outstanding role in the pathophysiology. IL-6, IL-8, and IL-17 (Th17 cells) also have an important contribution [13, 15‐17]. On the other hand, anti-inflammatory TGF-β and IL-10 have a protective role and are important for regenerative processes [18]. Prolonged moderate hypothermia improves neurological outcome and has become standard care for term infants with hypoxic-ischemic encephalopathy over the recent years [19]. One mechanism by which hypothermia exerts a neuroprotective effect may be by reducing systemic inflammation [20]. In an earlier study, we measured cytokine levels at the 6th, 12th, and 24th postnatal hours in neonates with asphyxia treated with hypothermia or standard intensive care on normothermia [21]. Our results indicated that IL-6 levels (at 6 h of age) and IL-4 levels (at all time points) were significantly lower in asphyxiated neonates treated with hypothermia compared to normothermic neonates. The duration of hypothermia initiated before 6 h of age correlated with lower levels of IL-6, TNF-α, and IFN-γ measured at 6 h of age. These data suggest that therapeutic hypothermia may rapidly suppress and modify the immediate cytokine response in asphyxia.

The permeability of the BBB is higher in neonates compared to adults and is further disrupted by the hypoxic injury itself. The release of IL-1β, TNF-α, and IFN-γ also increase the permeability of the BBB [22, 23]. CD49d is part of the VLA-4 antigen which mediates the migration of activated leukocytes to the site of tissue inflammation via binding to VCAM-1, expressed by endothelial cells [24]. VLA-4 is thus crucial for the migration of activated T lymphocytes through the BBB to the site of inflammation in the brain [25, 26], making it a primary therapeutic target in multiple sclerosis [27] and in primary neuroinflammatory brain disease in murine ischemic stroke models [28]. Although VCAM-1 is not exclusively expressed in the CNS, the level of CD49d expression can be correlated with the capacity of T lymphocytes to enter the site of inflammation, more specifically the brain tissue in the case of neuroinflammation [29].

The interplay between the kynurenine system and cytokines is a regulator of both innate and adaptive immune responses, and it plays an important role in the interactions between the central nervous and the immune systems [30, 31]. Indoleamine 2,3-dioxygenase (IDO), the rate-limiting enzyme in the degradation of tryptophan, plays a central role in regulating these interactions. IDO degrades TRP to kynurenine (KYN), which is then metabolized by the enzymes within the kynurenine pathway into other catabolites, such as Kynurenic acid (KYNA). While some TRP metabolites may have neurotoxic potential, KYNA appears to be a potent neuroprotective agent as it ameliorates NMDA receptor-mediated excitotoxicity [32] and acts as a potent free radical scavenger and endogenous antioxidant [33]. Induced by pro-inflammatory stimuli (such as IFN-γ), IDO is primarily produced by antigen presenting cells (APCs) and has several immunosuppressive effects, thus maintaining the balance between pro- and anti-inflammatory impulses. The rate of TRP degradation, expressed by the ratio of KYN to TRP (K/T), allows a good estimate of its enzymatic activity [34]. The induction of IDO and the kynurenine system results in the inhibition of T cell functions, the activation of regulatory T cells, and the inhibition of natural killer cells. The alterations of the kynurenine system appear to play a role in the pathophysiology of a broad spectrum of neurological disorders [35], including ischemic brain injury; however, its role has previously not been investigated in perinatal asphyxia.

The challenge in neonatal asphyxia is to harness the beneficial aspects of neuroinflammation following the insult to allow neuroprotection and regeneration within the CNS while at the same time minimizing its harmful effects. Significant barriers remain in understanding the benefits of inflammation in contrast to its detriments following neonatal asphyxia. Identification of factors that differentiate between infants with an extensive and potentially damaging neuroinflammatory response and infants with moderate inflammation would present new options for a more individualized therapeutic approach in neonatal asphyxia. In this study, we aimed to assess the prevalence and cytokine production of T lymphocyte subsets in moderate and severe perinatal asphyxia in order to identify players of the inflammatory response that may influence patient outcome. In contrast to previous studies, we aimed to determine the intracellular cytokine levels of T lymphocytes besides plasma cytokine levels. We also aimed to describe plasma TRP, KYN, and KYNA levels. We also expanded our investigation to 1 month of age following the CNS insult to understand longer-term consequences of the hypoxic event.

Several studies have described the importance of cytokines in normal neuronal differentiation and survival [41‐43]. The perinatal brain might be particularly susceptible to alterations in cytokine concentrations, and experimental data suggest that cytokines play a pivotal role in the regulatory network orchestrating neuroinflammation [44]. Cytokines have been in the limelight of research focusing on asphyxia. The inflammatory response following hypoxic brain injury has been shown to have dual effects. A certain level of inflammation appears to be necessary for the adequate regeneration of the brain tissue [6], while extensive neuroinflammation might contribute to further CNS injury and be an important factor in worse functional outcome. We aimed therefore to determine factors that might differentiate between infants who have an adequate level of inflammatory response which is necessary for neuroregeneration from infants in whom the uncompensated inflammatory response contributes to brain injury. Previous studies have focused on determining the level of cytokines from plasma. However, plasma cytokine levels have been shown to have a great variability compared with intracellular cytokine levels, which closely reflect cytokine production at a cellular level and show more stable kinetics in time, and thus open new possibilities for more precise characterization of the cytokine network in immune disorders. The relationship between cellular cytokine production and serum cytokine levels is undefined; cytokines in the serum originate from various different sources and show less stable kinetics. The advantage of intracellular cytokine analysis by flow cytometry is that with this method, the cytokine production of each cell type can be accurately described. This method opens up the opportunity for precise characterization of the function of each cell type in a physiological setting (i.e., maintaining autocrine and cell-cell interactions), which could be of great value in identifying key cellular players of various inflammatory conditions [45, 46]. We therefore primarily aimed to describe intracellular cytokine values; however, we also measured cytokine levels from plasma to gain a more comprehensive picture of the immunologic alterations following the hypoxic insult. Furthermore, contrary to previous studies, we expanded sampling of infants to 1 month of age to obtain data on longer-term changes in inflammatory parameters evoked by asphyxia.

In conclusion, the need for more specific prognostic markers other than clinical assessment in neonatal asphyxia is clear, since clinical signs often do not correlate with neurological outcome and do not enable differentiation between moderate and severe hypoxic-ischemic encephalopathy. The role of various cytokines in neuroinflammation following hypoxic-ischemic injury is supported by a rapidly expanding body of evidence. IL-1β and IL-6 appear to play a key role in the early events of the inflammatory response, while TNF-α seems to be responsible for triggering a prolonged inflammation, potentially contributing to a worse outcome. On the other hand, TGF-β has a compensatory role in decreasing the level of inflammation from an early stage following the insult (Table 3). Based on ROC analysis, the assessment of the prevalence of CD4+ IL-1β+ and CD4+ IL-1β+ CD49d+ cells at 6 h appears to be able to predict severity at an early stage in asphyxia (Fig. 5). Our current results open a potentially fruitful area of research as well as diagnostic and therapeutic development in neonatal asphyxia.