Improving Understanding and Outcomes of Traumatic Brain Injury Using Bidirectional Translational Research

Abstract

Recent clinical trials in traumatic brain injury (TBI) have failed to demonstrate therapeutic effects even when there appears to be good evidence for efficacy in one or more appropriate pre-clinical models. While existing animal models mimic the injury, difficulties in translating promising therapeutics are exacerbated by the lack of alignment of discrete measures of the underlying injury pathology between the animal models and human subjects. To address this mismatch, we have incorporated reverse translation of bedside experience to inform pre-clinical studies in a large animal (pig) model of TBI that mirror practical clinical assessments. Cerebral autoregulation is impaired after TBI, contributing to poor outcome. Cerebral perfusion pressure (CPP) is often normalized by use of vasoactive agents to increase mean arterial pressure (MAP) and thereby limit impairment of cerebral autoregulation and neurological deficits. Vasoactive agents clinically used to elevate MAP to increase CPP after TBI, such as phenylephrine (Phe), dopamine (DA), norepinephrine (NE), and epinephrine (EPI), however, have not been compared sufficiently regarding effect on CPP, autoregulation, and survival after TBI, and clinically, current vasoactive agent use is variable. The cerebral effects of these clinically commonly used vasoactive agents are not known. This review will emphasize pediatric work and will describe bidirectional translational studies using a more human-like animal model of TBI to identify better therapeutic strategies to improve outcome post-injury. These studies in addition investigated the mechanism(s) involved in improvement of outcome in the setting of TBI.

Introduction

Recent clinical trials in traumatic brain injury (TBI) have failed to demonstrate therapeutic effects even when there appears to be good evidence for efficacy in one or more appropriate pre-clinical models. Translational activities have been dependent largely on outcomes of efficacy that correlate with injury severity in the animal models but do not mimic practical clinical measures of the underlying pathology that could facilitate stratification for clinical trial design.

To address this mismatch, we have assembled a multi-disciplinary team of pre-clinical and clinical researchers to identify and develop new measures of discrete neurological function in a pre-clinical pig model that are reliable and have relevant and practical correlates for clinical assessment in TBI. The opportunity to monitor closely and use background intensive care unit therapies in pig TBI models may be helpful in identifying novel therapeutic strategies to improve outcome.

We have incorporated reverse translation of bedside experience to inform pre-clinical studies that mirror practical clinical assessments. Using this bidirectional translational model, we have studied vascular function, cerebral autoregulation, brain histopathology, and neurological outcome. Basic science insights into mechanism have been turned back to inform design of newer therapeutic strategies to improve outcome after TBI. We have been interested particularly in the role of age and sex in determining outcome after TBI and have observed parallelism between our basic science and clinical studies.

In humans, TBI is characterized by a vast spectrum of evolving neuropathologies that span varying degrees of brain swelling, hemorrhage, contusions, diffuse axonal injury, and ischemia. This heterogeneity is too complex to allow for investigation and development of effective treatment strategies in the clinical scenario. In addition, the primary mechanical injury initiates a secondary injury cascade that is often dependent on the type of initial insult.¹ Accordingly, experimental models of TBI have been developed to reproduce particular specific aspects of the clinical entity to support development of therapeutic strategies that target secondary injury cascades that have been upregulated.²

Historically, each of the models available for use in pre-clinical work have been categorized in terms of the model having characteristics of one of the two types of brain injury: focal and diffuse.³ Focal models of TBI produce either a contusion or a localized regional area of injury. Clinically relevant focal brain injuries include cerebral contusion and laceration, as well as hemorrhages and hematomas. Diffuse brain injuries include cerebral concussion and prolonged traumatic coma caused by diffuse axonal injury (DAI).

Some of the focal models of TBI include weight drop, controlled cortical impact (CCI), and subdural hematoma.⁴ An example of a diffuse model of TBI is impact acceleration, while fluid percussion brain injury (FPI) has elements of both diffuse and focal injuries.⁴ The FPI, in fact, can be of two types: lateral and medial. A lateral injury includes primarily cerebral cortical damage whereas a medial injury involves the brainstem in injury.^5,6

Use of a large animal model more similar to humans in pre-clinical TBI studies

Pre-clinical studies often have used rodents to model human TBI.^2,7 They are less expensive than larger animals and do offer some possibilities for physiological monitoring such as intracranial pressure (ICP) and cerebral blood flow (CBF) via arterial spin label magnetic resonance imaging (MRI). In addition, there is a large fund of literature on normative data for rodents that has been collated over many years.²

There have been a number of excellent reviews published recently that summarize the modeling of specific human TBI outcomes.^8–10 These studies have focused largely on molecular pathology, with some positive therapeutic results that unfortunately have not translated well into being successful in the human condition. Many have thought that this translational failure resulted from the predominant use of rodents in TBI compared with larger, more human-like species.^11,12

A key limitation toward translation of rodent observations to the human relates to the fact that rodents have lissencephalic brains, whereas humans have gyrencephalic brains. This difference will impact the degree of brain deformation experienced with an insult to the brain. For example, the presence of gyri influences the movement of the brain within the skull, with much more brain deformation occurring in gyrencephalic brains compared with those not containing gyri.¹³ In addition, gyri and sulci influence the points of maximal mechanical stress in tissue during angular acceleration.^14,15

In a lissencephalic brain, therefore, maximal stress would be uniformly experienced in tissue nearer the surface of the brain. In a gyrencephalic brain, however, the presence of sulci will force the maximal stress toward the base of the sulci, the result being that the maximal stress would be experienced farther away from the surface of the brain.

Brain size also plays an important role in that the smaller lissencephalic brain of the rat can withstand greater angular forces than animals with bigger gyrencephalic brains because shearing force is related to brain mass.¹³ This may have unintended consequences when, for example, TBI research is conducted in a higher species such as the marmoset monkey that also has a relatively small lissencephalic brain, confounding the attempt to investigate TBI in a more human-like species. Caution, therefore, should be exercised when interpreting primate TBI work.

Other brain characteristics vary among species. For example, rodents have relatively little white matter compared with larger animals such as the pig and sheep, which have large white matter domains and white/gray ratio much more similar to the human. This has implications when investigating diffuse axonal injury or edema. The bony structures of the skull vary by species and also influence the nature of the primary insult.

For example, ICP and, hence, edema can vary depending on the presence/absence of a rigid tentorium cerebelli, which separates the cerebrum from the cerebellum. This is quite rigid in the human. Elevations in edema after TBI generally occur in the cerebrum, and the presence of a rigid tentorium plays an important role in restricting such increases in ICP to the cerebrum.¹⁶ While larger species such as the pig and the sheep equally have a rigid tentorium, rodents do not, predisposing to a different ICP/edema pattern (spread to other compartments) in rodents compared with pigs/sheep.¹⁶

Another bony structure that varies by species is the sphenoid ridge, present in humans, but absent in rodents. This structure is an important source of brain trauma in acceleration/deceleration injury, leading to divergent injury patterns in humans and rodents.¹⁶ A final element is that physical size in large animal models more readily facilitates clinically relevant monitoring. For example, clinically relevant monitoring of variables such as ICP, brain tissue oxygen, and CBF can be performed in both large animals and humans. Importantly, such variables are exceedingly similar in large animals to age-matched humans in contrast to that obtained in rodents, thus reinforcing their clinical relevance.¹⁶

The overall shape, gyral pattern, and distribution of gray and white matter are similar in pigs/sheep and humans.^17,18 The response of the pig to hypoxia and ischemia appears parallel to that observed in humans.¹⁹ In addition, CBF and metabolism are similar, and the pig matures in a manner similar to the human with respect to myelination and brain electrical activity.²⁰ Selective vulnerability in the white matter similar to that observed in the human has been demonstrated in pigs with a model of acute subdural hematoma.²¹

Therefore, species such as the pig, sheep, and non-human primate all appear to be acceptable choices to model human TBI. For these and other reasons, investigators have developed large animal TBI models. To date, however, no new therapy for TBI has been translated successfully in a gyrencephalic animal.

Early studies focused on use of non-human primates. One of the earliest investigations involved study of CCI in the non-human primate by Ommaya and colleagues²² in 1966. The CCI technique requires a craniotomy to gain access to the brain and uses a pneumatic or electromechanically driven metallic piston, although more recently, there has been somewhat greater use of a closed head CCI approach.^23,24 It typically produces focal injury with vascular injury, edema, and neuronal cell loss, with more hemorrhage than that observed in FPI.²⁴ The severity of the injury can be varied by changing the velocity and/or the depth of the impactor penetration, along with the time of tissue contact.²⁴

Several decades later, seminal work by Gennarelli and associates²⁵ characterized DAI after TBI in the primate using a unique device, the Penn II device, which produces non-impact head rotational acceleration that results in coma. Angular acceleration was produced by moving the head rapidly (11–22 msec) through a 60-degree arc. The shape of the resultant wave consisted of a long acceleration phase followed by a short deceleration component, thought to mimic a fall or a motor vehicle accident.

The direction of head movement could be made sagittal (posterior to anterior), oblique (head turned 30 degrees to the left and moving posterior to anterior), or lateral (head turned 90 degrees to the left and moving left to right).²⁵ In each case, the amount of head movement was the same (60 degrees) and the acceleration magnitude similar (1 to 2 X 10⁵ radians/sec²).²⁵

In more recent years, emphasis in TBI research using large animals largely has shifted away from use of non-human primates in favor of use of swine and sheep models of TBI. These studies have used CCI, Penn II device, FPI, manual shaking, stunning, penetrating missile strategies, and acceleration methods for production of injury. There are numerous representative CCI studies in swine,^26–28 but an early study was performed in the ferret that initially supported its consideration in modeling human TBI.²⁹

A more contemporaneous study in the ferret,³⁰ however, provided little more support for claims of the value of ferrets in TBI research. In particular, in that study, a total of only 15 ferrets were used, a single naïve control animal, nine of the 15 to optimize the surgery, and five to study varying injury severity. Behavioral outcome graphs (Fig. 7 and 8) were illustrated with results from a single ferret comparing a single injured ferret with a single naïve control. There has been one subsequent ferret MRI article (in mild TBI)³¹ and no therapy testing. For these and other reasons, rationale for use of the ferret should be tempered.

The FPI consists of a device containing a fluid-filled column, which is fluid coupled to the brain. A pendulum is allowed to swing and strike a piston on the opposite end of the column, causing a pressure wave to strike the brain parenchyma. As arc length on the pendulum is increased, the degree of the insult similarly is increased. The FPI has been used in a number of larger animal species including the cat³² and the pig.^33–35

While there are a number of variants in the blast injury technique, many investigators use some form of a shock tube that is used to generate an overpressure shockwave driven by compressed gas (e.g., air, nitrogen) that is thought to simulate free-field blasts.^36,37 Other studies have employed explosive charge-driven shock tubes. To date, researchers have not standardized shock tube parameters, use of body shielding, head mobility, and location of the animal, all of which presumably may alter greatly the nature of the injury.

In the sheep, TBI modeling has been performed using a variety of generally less widely used techniques. Some, for example, have investigated manually shaking the head of sheep in a manner believed to occur with most human abused infants, but there was no head impact.³⁸ Others have used a non-penetrating stunner that delivers an increase in atmospheric pressure to the brain of sheep.³⁹ A final technique for sheep has been to cause TBI via a penetrating missile.⁴⁰

Head acceleration models (impact and non-impact) were developed in response to the mechanical limitations involved in injury techniques that require a craniotomy—e.g., CCI and FPI—as well as the lack of a rational component. In TBI resulting from falls and/or collisions, the head often experiences a rapid acceleration and/or deceleration. By introducing an angular component, the injury will reproduce the biomechanical forces that result in DAI.

The impact acceleration model has been used in sheep where it has characterized graded insults successfully.⁴¹ The latter model in the sheep used a humane stunner that was targeted at the temporal region of the ovine head. The severity of the impact was varied via altering the device explosive charge, which determines the velocity of the impactor. The head is unrestrained, allowing free rotation with impact. The DAI was observed in this model, which correlated with the severity of cerebral and systemic responses.⁴²

Because impact models are by nature prone to skull bone fracture, a non-impact method was developed (an example being the Penn device) and later adapted for use in the pig. These adaptations included scaling to accommodate for the difference in brain mass between swine and humans and use of bite plates so that the controlled angular head rotation in the pig could be achieved.⁴³ In contrast to the coronal rotation used in primates, in the pigs it was observed that the axial rotation produced widespread DAI.⁴⁴

Given the advantages associated with the pig in pediatric studies (size of animal that allows for determination of cerebrovascular indices of outcome with ease), adaptation of the non-impact acceleration model to the piglet soon followed. Significant decreases in cerebral perfusion pressure and brain oxygenation along with sustained increase in ICP were observed after axial plane rotation injury,⁴⁵ consistent with that observed clinically. Others have used FPI to study age-dependent effects of TBI in the pig.⁴⁶ On the basis of growth curves comparing maturation of humans with other species, a pig that is 1–5 days old roughly approximates the age of the human neonate while a four week old pig approximates that of the young child-adolescent.⁴⁷

There are, in addition, a few other models that have been used in large animals to study TBI. One is the freeze lesion model, which has been used in cats, dogs, and macaque monkys.^48–50 In this model, a liquid nitrogen cooled probe is applied to the brain via a craniotomy, which causes a vasogenic edema and altered blood flow and metabolism. Because it did not mimic other aspects of TBI seen clinically, it has had limited translational value. Penetrating injury models designed to mimic a gunshot wound have been developed and used in cats and sheep.^51,52 The latter method produced hemorrhage, vasogenic edema, elevated ICP, and reduced CPP.

Functional outcome measures

In addition to the morphological changes induced by TBI such as lesion volume, hippocampal neuronal cell loss, and others described above, a number of functional outcome measures have been developed and are used in rodent models of injury. Neurobehavioral and cognitive assessment tools utilized in rodents include neuroreflexes, beam balance, beam walk, rotorod, Barnes circular maze, and Morris water maze working memory, acquisition, and retention.

An advantage to the use of a large animal model of TBI is that the greater brain mass allows for a much better fidelity in calculating the mechanical stresses induced by the insult. Also, the larger brain size is easier to image with MRI and positron emission tomography methodologies. Disadvantages to use of large animals also exist. The ability to recapitulate the intense interest in characterization of molecular/cellular changes associated with TBI that has occurred in the rodent is limited by lack of appropriate molecular probes in large animals such as the pig, along with the associated greater cost of doing such studies in these species.

In general, behavioral and cognitive studies are lacking because of lack of development of such indices of outcome in large animal models of TBI. Some progress has been made with porcine models of TBI, however. For example, a neurological injury severity score using a 32-point scale that assesses the level of consciousness, behavior, appetite, head position, standing position, utterance, gait, and motor function has been developed for a porcine model of CCI.⁵³ Cognitive function in the latter porcine study was assessed via a learned task that is remembered on the basis of operant conditioning, generating information on memory retention, recall, color recognition, spatial memory, and prioritization.⁵³

Another research group, using CCI and acceleration models of porcine TBI, has established a clinically relevant large animal model platform for TBI therapy development.⁵⁴ These researchers have also developed porcine behavioral indices of outcome in the setting of CCI and non-impact sagittal rotation models of TBI.^55–57 The battery of measures developed included open field behaviors of sniffing and moving a toy, locomotion measures of Lempel-Ziv complexity, and the probability of remaining in the current location, and a metric for evaluating motor performance.

A composite porcine disability score was able to detect brain injury that correlated with the percent axonal injury in these animals. Interestingly, it was observed that the persistence of axonal injury and behavioral deficits is dependent on the rotational plane, with sagittal rotations causing more prolonged injury than axial rotations.⁵⁸

Clinical observations inform basic science modeling of TBI

Cerebral autoregulation is a homeostatic mechanism that regulates CBF across a range of blood pressures. Cerebral arterioles dilate and constrict to maintain CBF nearly constant over a range of blood pressures. In healthy adults, changes in mean arterial pressure (MAP) between 60–160 mm Hg or cerebral perfusion pressure (CPP) between 50–150 mm Hg produce little or no change in CBF.⁵⁹ Outside of these lower and upper limits of autoregulation, CBF will depend on CPP.

Hypotension may result in cerebral ischemia, and hypertension may cause cerebral hyperemia. Hypotension in the setting of pediatric TBI is associated with poor outcome.^60–62 Young children with TBI and younger than 4 years are more at risk of impaired cerebral autoregulation than older children, and after moderate-severe TBI, boys also have more impairment of autoregulation compared with girls after adjusting for injury severity.⁶³

After TBI, ICP is increased, which further decreases CPP, leading to more cerebral vasodilation, causing a vicious cycle. There is uncertainty as to whether it would be useful to empirically increase MAP to prevent cerebral ischemia as a means to protect cerebral autoregulation, because theoretically, augmenting MAP in the hyperemic brain could result in cerebral hemorrhage.^64–66 Impaired cerebral autoregulation has been associated with poor six-month outcomes after pediatric TBI,⁶³ but there is controversy as to whether impairment is a marker of injury or whether it contributes to poor outcome.

Assessment of cerebral autoregulation can be accomplished via either static or dynamic means. In the static method, only steady state relationships between CBF and MAP are considered without taking into account the time course of changes in these two parameters. Determination of a steady state relationship can be accomplished through administration of drugs that change MAP, yielding two values of CBF and their difference in relation to the MAP change being indicative of autoregulation.

In the dynamic method, assessment is based on determination of dynamic changes of CBF in response to dynamic changes in MAP. One approach considers study of CBF response to slow oscillations in MAP induced by head up tilting, paced breathing, or thigh-cuff inflation.^67–69 Transfer function analysis is then performed using beat-to-beat MAP measurements as input and CBF as output.⁷⁰

The time delay of phase difference between MAP and CBF as a function of frequency can be used to determine the degree of intactness of autoregulation. The pressure reactivity index (PRx) is one such measure of dynamic auotoregulation that uses this approach. PRx is defined as the moving Pearson correlation coefficient between MAP and ICP,⁷¹ and derives from the notion that in the normal brain, in low compliance states such as are frequently seen after TBI, increases in MAP will lead to constriction of the cerebral blood vessels to maintain constant flow. This results in a decrease in ICP because of a decrease in intracranial blood volume. The PRx values range from −1 (perfectly intact autoregulation) to +1 (complete absence of autoregulation).

The optimal MAP (MAP_opt) and CPP (CPP_opt) are thought to be in the center of the autoregulatory plateau and represent the pressures where vascular responsiveness is most robust. Optimization of CPP and MAP may limit secondary brain injury.

Studies in both adults and children have demonstrated a strong relationship between PRx and outcome, with values greater than 0.2 associated with high morbidity and death.^70–75 The PRx has been shown to vary with CPP, making it is a promising target for clinical therapeutics aimed at improving autoregulation and outcome.

The typical critical care pathway used to normalize CPP and limit impairment of cerebral autoregulation after TBI is to administer a vasoactive agent that increases MAP above the ICP that has been elevated by the injury. Vasoactive agents clinically used to elevate MAP to increase CPP after TBI, however, such as phenylephrine (Phe), dopamine (DA), norepinephrine (NE) and epinephrine (EPI),^76–78 have not been compared sufficiently regarding effect on CPP, CBF, autoregulation, and survival after TBI, and clinically, current vasoactive agent use is variable. In addition, the cerebral effects of these clinically commonly used vasoactive agents are not well characterized. Consequently, there are no recommendations regarding preferred choice of vasoactive agents in management of TBI.

Cerebral hemodynamic effects of TBI in large animal models of injury

Using the non-impact rotational model of closed head injury, it was observed that CBF was reduced by approximately 50% in the newborn pig, consistent with other laboratory findings in rats as well as human TBI patients.^79–81 Interestingly, the direction of the rotation yields influences the magnitude of the CBF reduction, with significant reductions observed for the sagittal but not coronal or horizontal injury.⁸² Similarly, sagittal rotation produced the longest duration of unconsciousness, highest incidence of apnea, and largest ICP increase while coronal rotation produced little change and horizontal rotation produced variable and intermediate change.⁸²

Substantial axonal injury occurred with both sagittal and horizontal rotations. It was speculated by the authors of this article that the direction dependent differences in injury outcomes related to differences in tissue mechanical loading produced during head rotation.⁸²

Clinically, current 2012 Pediatric Guidelines recommend maintaining CPP above 40 mm Hg, noting that an age-related continuum for the optimal CPP is between 40 and 65 mm Hg.⁸³ Clinically, however, not only is there no common agreement as to what vasoactive agent should be used to reach this CPP after TBI, in addition there is also no commonly accepted MAP value. One early study designed to inform this area noted that targeting a slightly higher MAP (e.g., 70 mm Hg) through use of Phe achieved a better outcome (cell injury volume and lactate/pyruvate ratio as an index of metabolic crisis) after rotational non-impact injury in the pig.⁸⁴

Using a fluid percussion brain injury model, other investigators performed a systematic comparison of the effects of several vasoactive agents on outcome in newborn and juvenile pigs.⁸⁵ Male and female newborn (1–5 days old) and juvenile (3–4 weeks old) pigs correspond to the human ages of younger than four and greater than four years of age.⁴⁷ These studies were designed to correspond with human studies demonstrating age and sex dependency in TBI effects.⁸⁶

In newborn males, Phe potentiated impairment of cerebral autoregulation but prevented impairment of cerebral autoregulation and histopathology in newborn females and older males and females after TBI.^86,87 Norepinephrine yielded a similar outcome to Phe, but it augmented impairment while impairment was the same as vehicle after TBI with NE.^88,89 Outcome was protected in young males and females and older females, but not older males after EPI.⁹⁰ In contrast, DA prevented impairment of autoregulation and histopathology after TBI in both ages and sexes.^91,92

These data indicate that the clinical impact is significant in that autoregulation therapy and particularly CPP optimal therapy is affected robustly by the choice of vasoactive drug. As described below, these data support a precision medicine approach toward management of TBI.

These studies advance the field in several ways. First, these studies indicate that there are age and sex dependent differences in ability to improve outcome when using drugs to achieve an equivalent CPP post-injury. To mimic the clinical situation, CPP was targeted (55–60 and 65–70 mm Hg for young and older pigs per 2012 Pediatric Guidelines) to determine the dose of the intravenous infusion of the particular vasoactive agent.

A second advance involves the use of cerebral autoregulation to individualize blood pressure management after TBI.⁹² Indeed, we have observed that the degree of intactness of autoregulation is associated with the Glasgow Coma Scale (GCS) score in pediatric patients with TBI.⁶² A correlate to GCS is the degree of neuronal cell necrosis in the CA1 and CA3 hippocampus after TBI. Protection of autoregulation was associated with prevention of hippocampal neuronal cell necrosis after TBI as a function of vasoactive agent choice. These data advance the field in that they support manipulation of cerebral autoregulation via targeted use of vasoactive agents in the setting of TBI as a means to effect improved cognitive outcome.

Other studies have been focused on characterizing the mechanism whereby some vasoactive agents improve outcome and others do not as a function of age and sex in the pig. Mitogen activated protein kinase (MAPK) is an important signaling pathway and exists in at least three isoforms: extracellular signal-related kinase (ERK), p38, and c-Jun N-terminal kinase (JNK). The phosphorylated (activated) ERK and JNK isoforms are increased in concentration after TBI and contribute to reduced CBF, impaired cerebral autoregulation, and histopathology after TBI.^85,87,89

Cerebral autoregulation is impaired more in males compared with females and in the younger compared with the older pig after TBI, because of greater upregulation of phosphorylated ERK and JNK.^85,87,88 Phe blocked phosphorylation of ERK in young female and older male and female pigs after TBI, but actually potentiated this upregulation in young males, resulting in potentiated impairment of cerebral autoregulation in young males and protection of cerebral autoregulation in all other groups after TBI.^85,87 NE blocked phosphorylation of ERK in young female and older male and female pigs after TBI, but not in young males, leading to impairment of cerebral autoregulation in young males and protection in all other groups.^88,89

Interleukin-6 (IL-6) concentration was elevated after TBI and contributed to increased phosphorylation of ERK.^88,89 On the other hand, EPI blocked phosphorylation of JNK in young male, young female, and older female but not older male pigs after TBI, resulting in impairment of autoregulation in older males but protection of autoegulation in all other groups.⁹⁰ In both sexes and ages, DA blocked phosphorylation of ERK after TBI, yielding protection of cerebral autoregulation after TBI in both sexes and age groups.^91,92

Additional studies were designed to mechanistically link modulation of MAPK isoforms with cerebral hemodynamic outcome after TBI in the pig. Relaxation of blood vessels can be mediated by several mechanisms, including cyclic adenosine monophosphate (cAMP), cyclic guanosine monophosphate (cGMP), and K⁺ channels.⁹³ Activation of K channels increases K⁺ efflux, producing hyperpolarization of vascular smooth muscle and vasodilation. Pharmacological studies using selective activators and inhibitors have provided evidence that K⁺ channels, especially adenosine triphosphate (ATP) sensitive (Katp) and calcium sensitive (Kca) channels, regulate cerebrovascular tone.⁹³

Vasodilation can be used as an index of the intactness of K channel function after TBI.⁹⁴ Pial artery dilation in response to hypotension is because of activation of Katp and Kca channels⁹⁵ giving functional significance to intactness of K channel function. Because pial artery dilation in response to Katp and Kca channel agonists is blunted more in the male than the female after FPI,⁹⁶ sex dependent greater reductions in CBF and impairment of cerebral autoregulation in the male compared with the female may relate to more aggravated impairment of K channel function in that sex.

We observed that the K channel agonist, adrenomedullin (ADM), is upregulated in the female but not the male after FPI, while exogenous ADM administration prevents impairment of autoregulation in both sexes.⁹⁷ These data indicate that the presence of an endogenous neuroprotectant in the female contributes to differential sex dependent outcome post-insult.^97–99

In an unrelated study, it was observed that endothelin-1 (ET-1) contributes to blunted K channel agonist mediated dilation after FPI via release of activated oxygen (O₂^-), which can then activate ERK.⁹⁹ Therefore, Phe, for example, impairs cerebral autoregulation in the newborn pig because of sequential upregulation of ET-1, superoxide, ERK MAPK, and impairment of K channel mediated cerebrovasodilation.⁹⁹

Our data support a precision medicine approach toward treatment of TBI. Of the vasoactive agents investigated, however, including NE, EPI, and Phe, DA is the only one demonstrated to improve outcome after TBI in both sexes and ages. These data suggest that DA should be considered as a first line treatment to protect cerebral autoregulation and promote cerebral outcomes in pediatric TBI irrespective of age and sex.

Emerging clinical perspectives regarding vasoactive agent choice in management of TBI

The idea of an invisible link between the heart and the brain has been recognized for centuries. In addition, a growing body of medical evidence has implicated neurological injuries, particularly subarachnoid hemorrhage, as a cause of cardiac dysfunction.^100,101 Despite the frequency with which TBI affects the general population, its potential impact on cardiac function has received little attention.¹⁰² Dr. Vavilala's group was the first to describe cardiac dysfunction after TBI.^103–106 Preliminary data show systolic dysfunction early after TBI, and related to TBI severity. Such cardiac dysfunction was observed for up to two weeks after TBI.¹⁰³

Cardiac enzymes (troponin, creatine kinase) were also observed to be elevated and associated with abnormal echocardiography (Fig. 1).¹⁰³ Cardiac dysfunction was defined as left ventricular ejection fraction as less than 50%. Should cardiac output be optimized to provide adequate CPP, cerebral ischemia may be prevented.

FIG. 1.

Mechanisms and clinical implications of cardiac dysfunction after neurologic illness. Adapted from Chest.¹⁰¹ ECG, electrocardiography. Color image is available online.

Additional to the above brain-heart interactions, heart-brain interactions can also occur wherein the primary injury to the brain is driven by the heart (Fig. 1). Low cardiac output and systemic hypotension, for example, may decrease CBF. Because TBI, an entity in which hypotension has been linked exquisitely to poor outcomes,⁶³ is associated with cardiac dysfunction,¹⁰⁷ there is the possibility for generation of a vicious cycle in which TBI affects heart function, feeding back to affect neurologic outcome (Fig. 1).

Although not shown in this figure, there has been an historic idea that there are intracranial mechanisms for preserving blood flow in the setting of CNS disease.¹⁰⁸ A substantial body of work supports the role of astrocytes as “neurovascular communicators,” sensing and responding to changes in neuronal activity and environment by regulating CBF.¹⁰⁹

In particular, the idea of a “brain baroreceptor” located in the brainstem has been suggested to involve astrocytes that signal to directly affect sympathetic nerve activity and blood pressure.^107,109 The potential existence of brain baroreceptors therefore offers an additional way for the brain and heart to communicate with one another.

Conclusions and Future Directions

Increasingly, health-related research involves complex teams composed of members from diverse disciplines with complementary expertise, and often in different locations, resulting in diversity of thought and a more robust end product. A bedside to bench and back again approach that we have taken toward study of the injured pediatric brain has sped transfer of knowledge between the clinical and basic science areas, allowing for breakdown of traditional silos of expertise. As a result, we have revealed an underappreciated concept that use of vasoactive agents to normalize CPP is age and sex dependent; a precision-based medicine approach should be used in management of pediatric TBI.

The concept that cross talk between brain and heart in TBI patients influences outcome has been appreciated recently. In future work to be conducted in the pig, we will first focus on brain-heart interactions after TBI in the absence of therapeutic management of brain injury to better understand the physiologic and pathologic interactions between these two organs. Next, we will superimpose the infusion of vasoactive agents (Phe, NE, DA, and EPI) to normalize CPP and determine how brain-heart interactions are modified. We speculate that therapeutic approaches that both limit cardiac dysfunction as well as improve cerebral hemodynamics will yield an improved outcome after TBI.

Author Disclosure Statement

No competing financial interests exist.