Background
The blood-brain barrier (BBB) is a highly specialized structure crucial for the maintenance of central nervous system (CNS) homeostasis [
1,
2]. The basis of the barrier in the brain, and the corresponding barrier in the spinal cord - the blood-spinal cord barrier (BSCB), is a network of endothelial cells joined by tight junctions that line the blood vessels within the CNS [
3,
4]. The core 'neurovascular unit' comprises endothelial cells, pericytes and astrocytic endfeet embedded within their basal laminae. The space between the astrocytic endfeet, which make up the abluminal surface of CNS capillaries, and the endothelial cells/pericytes represents the interface between the blood and CNS. The BBB is highly restrictive with only a subset of small molecular weight, diffusible molecules readily crossing from the blood into the CNS parenchyma. Thus, most substances are normally precluded from entering the CNS by the BBB. However, in many CNS pathological states the BBB becomes disrupted, allowing entry of substances from blood into the CNS, and this disruption is considered a key step for disorders such as traumatic injury, stroke and neurodegeneration.
In peripheral tissues, vascular permeability is normally greater than in the CNS, although there is a vascular-tissue barrier that excludes, for example, large proteins from entering the tissue. It has long been known that peripheral vascular permeability in skin and other tissues can be markedly increased by antidromic discharges in primary sensory neurons, allowing large proteins to leak through capillaries thereby causing plasma extravasation [
5‐
7]. This plasma extravasation, together with the vasodilation that is produced by sensory nerve stimulation, comprise neurogenic inflammation which is mediated by the release of the peptides substance P and calcitonin gene-related peptide (CGRP) from peripheral terminals of peptidergic C-fibers [
6,
8,
9]. In accordance with Dale's Principle [
10], discharge activity in peptidergic C-fibers also releases substance P and CGRP from the central terminals of primary afferents in the spinal cord dorsal horn [
11,
12]. However, it has been found that vascular permeability in the dorsal horn is not increased by activating C-fibers, at least over the time course of peripheral neurogenic inflammation [
13]. Therefore, it has been assumed that although activity in sensory nerves causes rapid increases in vascular permeability in peripheral tissues, this activity is not capable of causing vascular permeability to increase in the CNS.
Discharge activity in primary afferents, particularly that initiated by injury to sensory nerves, may have slowly developing and persistent consequences in the nervous system that can lead to chronic pain states [
14]. Emerging evidence indicates that one such consequence of peripheral nerve injury (PNI) is the entry of monocytes [
15] and T cells [
16,
17] from the circulation into the spinal dorsal horn. Because the accumulation of these normally circulating cells is apparent many hours or days after the nerve injury, we wondered whether the BSCB may be disrupted at these times, as has been suggested at longer times after spinal nerve transection [
18]. Therefore, here we investigated whether PNI or activation of primary afferent C-fibers may cause increased vascular permeability in the CNS but over a time course beyond that of neurogenic inflammation. Preliminary results of portions of this work have been reported [
19].
Discussion
Here we have discovered that peripheral nerve injury or electrical stimulation of C-fibers in the sciatic nerve produce opening of the blood-spinal cord and blood-brain barriers. The increase in BSCB permeability is prevented by applying lidocaine to the nerve prior to the nerve injury or electrical stimulation, indicating that action potential discharge is required. By contrast, applying lidocaine directly after the electrical stimulation had no effect on the subsequent increase in BSCB permeability. Increased BSCB permeability was also produced by direct application of capsaicin to the sciatic nerve. Taking our findings together, the most parsimonious explanation is that discharge activity in TRPV1-expressing C-fibers triggers a cascade of events which, after many hours, leads to an increase in BSCB and BBB permeability. Our findings thus reveal a previously unknown function of TRPV1-expressing C-fibers.
This function of these afferents to cause opening of the BSCB was unanticipated as peripheral neurogenic plasma extravasation, which is also initiated by TRPV1-expressing C-fibers, develops rapidly upon electrical activation. Plasma extravasation has been found to begin within tens of seconds of the start of C fiber stimulation [
20,
23]. By contrast, BSCB permeability has been found to not increase during or within minutes after C fiber stimulation [
13] or with intravenous administration of capsaicin [
24]. Thus, the onset of the effect of TRPV1-expressing afferents on vascular permeability in peripheral tissues is dramatically different than that of the effect of these fibers on vascular permeability in the CNS. A second major difference between the effects of TRPV1-afferents on peripheral versus CNS vascular permeability is the duration of the increase: peripheral plasma extravasation ends within minutes of terminating electrical C-fiber stimulation [
23], whereas the increase in BSCB permeability persists for days after the stimulation. A third difference is the localization of the increase in vascular permeability: plasma extravasation in the periphery is highly localized to the territory innervated by the nerve that is stimulated [
25] whereas the increase in permeability in the CNS is evoked far beyond the anatomical distribution of the central terminals, occurring throughout the spinal cord. These major differences imply that, beyond being initiated by TRPV1-expressing C-fibers, the mechanisms for the sensory neuron-evoked increase in vascular permeability in the CNS are highly divergent from those producing plasma extravasation in the periphery.
Because the increase in BSCB permeability is delayed many hours after electrical C-fiber stimulation and the increase is not prevented by applying lidocaine to the nerve immediately after the stimulation, we infer that the activity evoked during the stimulation acts to trigger a cascade of events that culminates in the opening of the barrier. It is conceivable that the cascade may include transcription of critical genes and translation of the relevant gene products that act as mediators. In addition, or alternatively, to increased production of certain gene products, key elements of the BBB might become reduced, and in preliminary experiments we have found a decrease in the level of aquaporin-4, a component of the BBB, preceding the increase in permeability [
19]. Given the widespread increase in vascular permeability in the CNS, a possible scenario may be that the cascade of events involves the production and release of a humoral mediator, or mediators, that act on the cellular and/or intercellular elements that maintain the intact barrier. Alternatively, there could be rostral spread of BSCB permeability from an initiation site in the lumbar dorsal horn or central release of diffusible mediator(s) that circulate within the cerebrospinal fluid. Or, the increased permeability might be caused by synaptically released mediators from neuronal pathways having widespread projections, such as those from brainstem or hypothalamic regions.
The increase in BSCB permeability might contribute to the development of pain hypersensitivity after PNI. Increased permeability may facilitate the entry of circulating cells such as monocytes [
15] and T cells [
17] that home to the dorsal horn near the site of termination of the central endings of injured primary afferents. Increased BSCB permeability could also allow entry of soluble factors that are normally excluded but that could contribute to pain hypersensitivity in the dorsal horn. For example, fibronectin is known to cross the opened BBB [
26] and this is known to stimulate P2X4R expression in microglia [
27], and this increase in P2X4 is critical for mechanical hypersensitivity after PNI [
28,
29]. Also, matrix metalloproteinases (MMPs) have been implicated in opening the BBB [
30], and MMPs in the dorsal horn are critical for pain hypersensitivity after PNI [
31,
32]. However, the PNI-induced increase in BBB permeability is not sufficient on its own to cause pain hypersensitivity because permeability is increased throughout the spinal cord whereas pain hypersensitivity is typically restricted to the region near the nerve injury.
Our finding that the PNI-induced increase in BSCB permeability began within 24 hours and the permeability returning to the basal level by 7 days after PNI, is at earlier time points than a previous report using transection of the L4 spinal nerve [
18]. It was found that L4 nerve injury caused an endogenous albumin accumulation at time points of one to ten weeks after the transection. The accumulation observed at those time points may reflect albumin that entered during the period of opening that we have discovered here. The recovery of BSCB integrity appears to be due to a gradual loss of the signals that are driving the opening, rather than to resistance of the BSCB to sustained signals to open, because we found here that the permeability increased upon a second nerve injury 7 days after the first. Thus, while the BSCB is only open for several days after nerve injury, repeated nerve injury may re-open the BSCB after it has recovered.
Traumatic and ischemic injuries to the CNS are well-known to cause localized disruption of the BBB which is considered critical to the pathophysiology [
1,
33,
34]. With the stimuli used presently -- PNI, electrical C-fiber stimulation or capsaicin applied to the nerve -- there is no direct injury to areas where the BSCB and BBB permeability is increased. Thus, CNS injury cannot account for the increases in BSCB and BBB permeability we have found presently.
Our findings open up the possibility that PNI and activating TRPV1-expressing primary afferents may cause increased BSCB and BBB permeability in humans. If this is found to be the case, then there are several potential clinical implications of our findings. First, the BSCB and BBB would be much more permeable in situations of peripheral trauma, accidental or surgical, particularly when there is nerve damage. From this, one may expect that the pharmacokinetics of systemically administered drugs may be altered and the penetration into CNS might be enhanced. This enhanced drug penetration might increase the potency of agents where BBB permeability is limiting, but might also increase centrally-mediated side effects. A second potential clinical implication is that our findings may point to development of approaches to purposefully open the BBB in order to facilitate entry of drugs that normally have limited access to central targets. While it is unlikely that direct nerve injury would be used therapeutically, it is conceivable that short-duration C-fiber stimulation, done under general anaesthesia as in the studies here in rats, might be tolerated. However, by far the best approach would be through ascertaining the mediator or mediators involved in the afferent-induced opening of the BBB and to then develop treatments based on mimicking, in a safe way, the mechanism. A third possibility is that widespread opening of the BBB caused by PNI or activation of TRPV1-expressing C-fibers might contribute to the so-called 'sickness syndrome' [
35] that follows injury. A final, and much more speculative, possibility to consider pertains to non-traumatic, non-ischemic disorders where disruption of the BBB appears critical for the pathophysiology, such as demyelinating, neuroimmune or neurodegenerative disorders. Our findings raise the possibility that in some such diseases activity of TRPV1-expressing C-fibers may contribute to the BBB disruption.
Methods
Animals
All animals were used in accordance with the guidelines of the Canadian Council on Animal Care. All protocols were approved by the Animal Care Committee of the Hospital for Sick Children. For all experiments male 250-300 g Sprague-Dawley (Charles River) were used.
Peripheral Nerve injury models
Spared nerve injury (SNI) was performed as described previously [
36]. Briefly, rats were anaesthetized by isofluorane inhalation and the left sciatic nerve exposed under aseptic conditions. The distal trifurcation of the sciatic nerve was identified and the tibial and common peroneal branches ligated and cut, leaving the sural branch intact. The wound was sutured closed and the animals allowed to recover and returned to their housing. Chronic constriction injury (CCI) was performed as described previously [
37]. A polyethylene cuff (PE-60, 2 mm in length) was placed around the exposed left sciatic nerve and the wound closed as above. Sham surgeries, exposure of the sciatic nerve only, were also performed. To assess the effect of sequential nerve injury on the permeability of the BSCB, CCI was performed on the left sciatic nerve followed by a further CCI on the right nerve 7 days later.
Electrical stimulation of sciatic nerve
The sciatic nerve was isolated as described above and a bipolar stimulating hook electrode was used for electrical stimulation. The stimulation parameters for C-fiber strength stimulation were: 5 min train duration, 500 μsec stimulus pulse duration, 10 Hz, 10 mA. The stimulation parameters for A-fiber strength stimulation were: 5 min train duration, 150 μsec stimulus pulse duration, 10 Hz, 1 mA.
Lidocaine block
0.1 ml of 2% lidocaine (Novocol Pharmaceutical, Canada) was applied topically to the sciatic nerve for 30 minutes before nerve injury or C-fiber stimulation. The efficacy of local lidocaine block was confirmed by lack of peripheral plasma extravasations 45 minutes after C-fiber stimulation.
Capsaicin stimulation
Capsaicin (1% in Ethanol: Tween 80: saline 1:1:8, Sigma-Aldrich) or vehicle was applied to the sciatic nerve.
Evans Blue assay
Blood spinal cord permeability was determined by Evans Blue extravasation into the spinal cord. Evan's blue dye (2%, 4 ml/kg) was infused through the jugular vein of anaesthetized rats. After 45 minutes animals were perfused with PBS. The spinal cord and supraspinal tissues were immediately dissected and dura mater removed. The lumbar spinal cord (L4-6), and middle thoracic cord were further dissected. Tissue was incubated in 600 μl of formamide (Sigma-Aldrich) at 60C for 72 h. Evan's Blue concentration was then determind by spectrophotometry at 620 nm.
Horseradish peroxidise assay
BSCB permeability was further assayed using a horseradish peroxidise (HRP) assay. HRP (3 mg/kg) was infused through the jugular vein of anaesthetized animals. 45 minutes later animals were perfused transcardially with 10% formalin solution. Following post-fixation and sucrose cryoprotection, spinal cord sections (50 μm) were cut on a microtome. Sections were incubated in cy3-labelled tyramide solution (Perkin Elmer) for 7 minutes to measure HRP activity. Resulting fluorescence images were captured and integrated pixel density measured using Volocity software.
Statistics
Evan's Blue concentration was assessed across all groups using one way ANOVA and Newman-Keuls post-hoc multiple comparisons. Where appropriate, naïve and experimental groups were compared using t-tests.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SB participated in the design of the study, carried out experiments, analyzed data and wrote the paper. XJL carried out experiments, analyzed data and participated in writing of the paper. CK designed and carried out experiments, and analyzed data. MWS conceived of the study, participated in its design and coordination, and wrote the paper. All authors read and approved the text.