Celsus described the four principal effects of acute inflammation nearly 2000 years ago (Table 3.1). These include redness from acute dilatation of small blood vessels within the area. Heat, or warmth, is usually seen only in the peripheral parts of the body such as the skin. It is also due to the increased blood flow or hyperemia through the region from vascular dilatation. Swelling results from edema which is the accumulation of fluid in the extravascular space and from the physical mass of the inflammatory cells migrating to the area. Pain is one of the best-known features of acute inflammation and it results partly from the stretching and distortion of tissues due to the edema in the area. Chemical mediators of acute inflammation including bradykinin, the prostaglandins, and serotonin are also known to induce pain. Loss of function is a well-known consequence of inflammation added by Virchow to the list originated by Celsus. Movement of an inflamed area is consciously and reflexively inhibited by pain, while severe swelling or local muscle spasm may limit movement of the area.

Table 3.1 Celsus’s original description of the characteristic signs of inflammation

Erythema (rubor)

Warmth (calor)

Pain (dolor)

(Loss of function was added by Virchow)

The acute inflammatory response involves three changes or processes. Changes in the vessel size and flow, increased vascular permeability and the formation of the fluid exudate, and migration or de-margination of polymorphonuclear leukocytes (PML) into the extravascular space are characteristic processes of acute inflammation. Briefly, these early stages involve small blood vessels adjacent to the area of the tissue damage, which become dilated with increased blood flow. As blood flow begins to slow, the endothelial cells swell and partially retract so that they form a leaky continuum within the blood vessel. The vessels become leaky, which permits the passage of water, salts, and small proteins into the damaged area. One of the main proteins to leak out during this period is fibrinogen. Circulating PMLs initially adhere to the swollen endothelial cells and then migrate through these channels created by the retracted endothelial cells and through the basement membrane, passing into the area of tissue damage. Later on, blood monocytes (macrophages) migrate in a similar way. The microcirculation consists of a network of small capillaries that lie between the arterioles. These microcapillaries initially experience an increased blood flow following the initial phase of arteriolar constriction, which is transient. Blood flow to the injured area may increase up to tenfold during this time, but then blood flow begins to slow down, allowing the leukocytes to de-marginate into the area. The slowing of this blood flow, which follows the phase of hyperemia, is due to increased vascular permeability and allows plasma to escape into the tissues while blood cells stay within the blood vessels. Blood viscosity is therefore relatively increased as the percentage of red cells relative to white cells and other proteins increases. The increased vascular permeability increases capillary hydrostatic pressure as well as allowing the escape of plasma proteins in the extravascular space. Instead of the usual return of fluid into the vascular space, however, proteins act to increase the colloid osmotic pressure in the extravascular space. Consequently, more fluid leaves the vessel than comes back and the net escape of protein rich fluid is called exudation.

Experimental work has demonstrated three patterns of increased vascular permeability. There is an immediate response that is transient, lasting 30–60 minutes, mediated by histamine acting on the endothelium directly. A delayed response starts 2–3 hours after injury and may last for up to 8 hours. This is mediated by factors synthesized by local cells such as bradykinin or factors from the complement cascade, or those released from dead neutrophils in the exudate. A third response that can be prolonged for more than 24 hours is seen if there is a direct necrosis of the endothelium. In the later stages of acute inflammation where movement of neutrophils becomes important, experimental evidence has shown purposeful migration of neutrophils along a concentration gradient. This movement appears to be mediated by substances known as chemotactic factors diffusing from the area of damage. The main neutrophil chemotactic factors are C5a, LTB4, and bacterial components. These factors, when bound to the receptor on the surface of a neutrophil, activate secondary messenger systems stimulating increased cytosolic calcium with the assembly of cytoskeletal specializations that are involved in their ability to move.

The spread of the inflammatory response following injury to a small area of tissue suggests that chemical substances are released from the injured tissues spreading out to uninjured areas. These chemicals are called endogenous mediators and contribute to the vasodilatation, de-margination of neutrophils, chemotaxis, and increased vascular permeability. Chemical mediators released from the cells include histamine, which is probably the best-known chemical mediator in acute inflammation. It causes vascular dilatation in the immediate transient phase of increased vascular permeability. This substance is stored in mast cells, basophils and eosinophils, as well as platelets. Histamine released from those sites is stimulated by complement components C3a and C5a, and by lysosomal proteins released from neutrophils. Lysosomal compounds are released from neutrophils and include cationic proteins that may increase vascular permeability and neutral proteases, which may activate complement. Prostaglandins are a group of long-chain fatty acids derived from arachidonic acid and synthesized by many cell types. Some prostaglandins potentiate the increase in vascular permeability caused by other compounds. Part of the antiinflammatory activity of drugs such as aspirin and nonsteroidal antiinflammatory drugs (NSAIDs) is attributable to inhibition of one of the enzymes involved in prostaglandin synthesis. Leukotrienes are also synthesized from arachidonic acid, especially in neutrophils, and appear to have vasoactive properties. SRS-A (slow-reacting substance of anaphylaxis) involved in type I hypersensitivity is a mixture of leukotrienes. Serotonin (5-hydroxytryptamine) is present in high concentration in mast cells and platelets and is a potent vasoconstrictor. Lymphokines are a family of chemical messengers released by lymphocytes. Aside from their major role in type IV sensitivity, lymphokines also have vasoactive or chemotactic properties.

Within the plasma are four enzymatic cascade systems including the complement system, the kinins, the coagulation factors, and the fibrinolytic system, which are interrelated and produce various inflammatory mediators. The complement system is a cascade system of enzymatic proteins and can be activated during the acute inflammatory reaction in various ways. In tissue necrosis, enzymes capable of activating complement are released from dying cells. During infection, the formation of antigen–antibody complexes can activate complement via the classical pathway, while endotoxins of Gram-negative bacteria activate complement via the alternative pathway. Products of the kinin, coagulation, and fibrinolytic systems can also activate complement. The products of complement activation that are most important in acute inflammation include C5a, which is chemotactic for neutrophils, increases vascular permeability, and releases histamine from mast cells. C3a has similar properties to those of C5a, but is less active. C5, 6, and 7 are all chemotactic for neutrophils and, in combination with 8 and 9, have additional cytolytic activity. Finally, C4b, C2a, and C3b are all important in the opsonization of bacteria, which facilitates phagocytosis by macrophages.

The kinin system includes the kinins, which are peptides of 9 to 11 amino acids that are important in increasing vascular permeability. The most important of these is bradykinin. The kinin system is activated by coagulation factor XII. Bradykinin is also an important chemical mediator of pain, which is a cardinal feature of acute inflammation.

Within the coagulation system, factor XII, once it has been activated by contact with extracellular materials, will activate the coagulation, kinin, and fibrinolytic systems directly. This system is responsible for the conversion of soluble fibrinogen into fibrin, which is the major component of the acute inflammatory exudate. These fibrin degradation products result from the lysis of fibrin in the presence of plasmin.

Within the fibrinolytic system, the fibrin degradation products have effects on local vascular permeability.

The PML is the characteristic cell in acute inflammation. Its ability to move in a response to a concentration gradient of chemotactic factors has been well demonstrated and is mediated by cytosolic calcium. Neutrophils are able to bind to bacterial components via their Fc receptor and are able to phagocytose various particles or organisms and partially liquefy them with toxic compounds contained within lysosomes.

Following tissue damage or loss from any cause, including damage due to the inflammatory process, there may be resolution, regeneration, or repair. All of these processes may occur in the same tissue or begin as soon as there is significant tissue damage. Healing does not wait for inflammation or other mechanisms to subside, but usually takes place concurrently. The outcome depends on which of these three processes predominate and on a number of factors. Resolution tends to occur when there is little tissue destruction as well as a limited period of inflammation and short, successful treatment. Regeneration occurs when lost tissue is replaced by a proliferation of cells of the same type reconstructing the normal architecture. Regeneration proceeds based on cell type, and cells are usually classified into three groups based on their ability to regenerate. Labile cells are those that are normally associated with high rate of loss and replacement and therefore have a high capacity for regeneration. Stable cells do not normally proliferate to a significant extent but can be stimulated to do so after they have been damaged. Permanent cells are unable to divide after their initial development and therefore cannot regenerate when lost (i.e. neurons).

Tissue architecture is also important. Simple structures are easier to reconstruct following damage than complex ones. An imperfect attempt at regeneration can have important clinical consequences such as the cirrhosis that results after damage to the liver and the resulting abnormal nodular architecture from the repair. This process is also dependent on the amount of tissue loss. There must be cells left in the area to regenerate, as well as a reasonable volume to regenerate prior to scar formation. In repair, the process results in formation of a fibrous scar from the granulation tissue. Following the acute inflammation and phagocytosis of necrotic debris and other foreign material, blood vessels proliferate and fibroblasts assemble at the edge of the damaged area. As the endothelial cells and fibroblasts grow into the damaged area, vascularization also proceeds. Fibroblasts continue to proliferate, producing collagen and giving the tissue mechanical strength; eventually a scar consisting of dense collagen results. Factors influencing healing include the rate of healing, the presence of foreign material or of continuing inflammation, inadequate blood supply, abnormal motion, or certain medications that inhibit this process. Systemically, the healing process becomes less effective and slower with increasing age. Nutritional deficiencies play an important role as well as metabolic diseases such as renal failure or diabetes mellitus. Some patients with ongoing malignancies are actually in a catabolic state and unable to heal even simple wounds. Additionally, corticosteroids are important systemic inhibitors of wound healing.

The process of inflammation

Inflammation is a complex, stereotypical reaction of the body in response to damage of cells in vascularized tissues. In avascular tissue such as the normal cornea or within the disc space, true inflammation does not occur. The cardinal signs of inflammation presented earlier, including redness, swelling, heat, pain and deranged function, have been known for thousands of years. The inflammatory response can be divided temporally into hyperacute, acute, subacute, and chronic inflammation. The response can be based on the degree of tissue damage, such as superficial or profound, or on the immunopathological mechanisms such as allergic, or inflammation mediated by cytotoxic antibodies, or inflammation mediated by immune complexes, or delayed-type hypersensitivity reactions. As presented earlier, the development of inflammatory reactions is controlled by cytokines, by products of the plasma enzyme systems (complement, the coagulation system, the kinin and fibrinolytic pathways), by lipid mediators (prostaglandins and leukotrienes) released from different cells, and by vasoactive mediators released from mast cells, basophils, and platelets. These inflammatory mediators controlling different types of reactions differ from one another. Fast-acting mediators such as the vasoactive amines and the products of the kinin system modulate the immediate response. Later, newly synthesized mediators such as leukotrienes are involved in the accumulation and activation of other cells. Once the leukocytes have arrived at the site of inflammation, they release mediators that control the later accumulation and activation of other cells. However, it is important to realize that in inflammatory reactions initiated by the immune system the ultimate control is exerted by the antigen itself, in the same way as it controls the immune response itself. For this reason, the cellular accumulation at the site of a chronic infection or in an autoimmune reaction is quite different from that at sites where the antigenic stimulus is rapidly cleared.

Inflammation can become chronic. In certain settings, the acute process, characterized by neutrophil infiltration and edema, gives way to a predominance of mononuclear phagocytes and lymphocytes. This probably occurs to some degree with the normal healing process that becomes exaggerated and chronic when there is an effective elimination of foreign material as in some infections, or introduction of foreign bodies, or deposition of crystals, or persistent inflammatory product secretions such as disc herniations.

Inflammatory cells

Mediators of inflammation

In addition to the previously mentioned cell types, there are several chemical mediators of inflammation. There is considerable redundancy of these mediators. The most important vasoactive mediators stored in mast cells and basophil granules are histamine and serotonin. These are both also present in human platelets. Histamine has diverse functions including dilation of small vessels, locally increased vascular permeability by endothelial cell contraction, chemotaxis for eosinophils, and blocking of key T-lymphocyte function. Serotonin is also capable of increasing vascular permeability, dilating capillaries, and producing contractions of nonvascular smooth muscle.

Lipid mediators

The major constituents of cell membranes are phospholipids. Cellular phospholipase, especially phospholipase A₂ and C, are activated during inflammation and degrade phospholipids to arachidonic acid. Arachidonic acid has a short half-life and can be metabolized by two major routes, the cyclooxygenase and the lipoxygenase pathways. The cyclooxygenase pathway produces prostaglandins, prostacyclins, and thromboxanes. The lipoxygenase pathway produces either leukotrienes or lipoxins.

The prostaglandins are a family of lipid-soluble hormone-like molecules produced by different cell types in the body. For example, macrophages and monocytes are large producers of both PGE₂ and PGF₂. Neutrophils produce moderate amounts of PGE₂, and mast cells produce PGD₂. PGE₂ enhances vascular permeability, is pyrogenic, and increases sensitivity to pain. Prostaglandins must be synthesized and released in response to an appropriate stimulus and do not exist free in tissues.

Thromboxin A₂ is produced by monocytes and macrophages as well as platelets. It causes platelets to aggregate and vasoconstriction. These effects are somewhat opposed by the action of prostacyclin which is a potent vasodilator. Leukotrienes LTD₄ and 5-hydroxyeicosatetranoate (5-HETE) cause the chemotaxis and chemokinesis of several cell types including neutrophils. They are spasmogenic and cause contraction of smooth muscle and have effects on mucous secretion. Lipoxins LXA₄ and LXB₄ stimulate changes in microcirculation.

Products of the complement system

Complement is a complex system containing more than 30 different glycoproteins present in the serum in the form of components, factors, or other regulators, and on the surface of different cells in the form of receptors. The components of the classical pathway are numbered 1–9 and in prefix by the letter ‘C.’ All these pathways use C5–C9 that form the membrane attack complex (MAC). Activation of each of the components results from the proteolytic cleavage event in a cascade mechanism. The complement system influences the activity of numerous cells, tissues, and physiologic mechanism of the body. The result of cytotoxic complement reaction may be beneficial or harmful to the body. The complement system is a potent mechanism for initiating and amplifying inflammation. This is mediated through fragments of complement components. Tissue injury following ischemic infarction may also cause complement activation and abundant deposition of membrane attack complex may be readily seen in tissue following ischemic injury.

Cytokines mediating inflammatory functions

Cytokines are soluble glycoproteins that act nonenzymatically through specific receptors to regulate cell functions. Cytokines make up the fourth major class of soluble intercellular signaling molecules with neurotransmitters, endocrine hormones, and autocoids. Cytokines are synthesized, stored, and transported by many different cell types. Lymphokines are cytokines that are secreted mainly by activated T lymphocytes and monokines are produced by activated macrophages and monocytes. In order to unify the terminology of these factors, the term interleukin was accepted. Besides the term expressing their origin, cytokines can also be named according to their function as are interferons and others. Cytokines are directly responsible for the temporal amplitude and duration of the immune response as well as tissue remodeling. Individual cytokines can have widely varying responses and functions depending on cell type, concentration, and the synergistic or modulating effects of other cytokines. The information that an individual cytokine conveys depends on a pattern of regulators to which a cell is exposed and not on just a single cytokine. There is no doubt that cytokines contribute to the signs, symptoms, and pathology of inflammatory, infectious, autoimmune, and malignant diseases. TNF-α is an excellent example. Locally, it has important regulatory and antitumor activities but when TNF-α circulates in higher concentrations it may be involved in the pathogenesis of endotoxic shock, cachexia, and other serious diseases. Inflammation is dependent on both pro- and antiinflammatory cytokines. Proinflammatory cytokines are produced predominantly by activated macrophages and are involved in the upregulation of inflammatory reactions. Antiinflammatory cytokines belong to the T-cell-derived cytokines and are involved in the downregulation of inflammatory reactions.

The central role in inflammatory responses involves IL-1 and TNF-α. Antagonists to IL-1 (IL-1ra) and TNF-α may become important clinically in the treatment of some rheumatologic conditions such as ankylosing spondylitis and rheumatoid arthritis. IL-1 and TNF-α with IL-6 serve as endogenous pyrogens. The upregulation of inflammatory reactions is also performed by IL-11, IFN-α, IFN-β, and especially by the members of the chemokines superfamily. On the other hand, antiinflammatory cytokines (IL-4, IL-10, and IL-13) are responsible for the downregulation of the inflammatory response. The production of most lymphokines and monokines such as IL-1, IL-6, and TNF-α is also inhibited by TGF-β. However, TGF-β has a number of proinflammatory activities including chemoattractant effects on neutrophils, T lymphocytes, and nonactivated monocytes. TGF-β has been demonstrated to have in vivo immunosuppressive and antiinflammatory effects, as well as proinflammatory and selected immunoenhancing activities. When administered systemically, TGF-β acts as an inhibitor, but if given locally it can promote inflammation. Generally, TGF-β stimulates neovascularization and the proliferation and activities of connective tissue cells, and is a pivotal factor in scar formation and wound healing. But TGF-β has antiproliferative effects on most other cells including epithelial cells, endothelial cells, smooth muscle cells, myeloid, erythroid, and lymphoid cells.

Chemokines

Chemokines represent a superfamily of chemotactic cytokines acting as initiators and potentiators of antiinflammatory reactions. They are active over a high concentration range, and are produced by a wide variety of cell types. Exogenous irritants and endogenous mediators such as IL-1, TNF-α, PDGF and IFN-γ induce the production of chemokines, and because they bind to specific cell surface receptors they are considered second-order cytokines. Additionally, most chemokine molecules share structural similarities and function, and are attractants for many different types of cells.

BIOCHEMISTRY OF DISC DEGENERATION

The biochemical events that occur with intervertebral disc degeneration and, in particular, the role of biochemical mediators of inflammation and tissue degradation, have received more attention in the literature over the last 10 years. Matrix metalloproteinases (MMPs), prostaglandin E₂ (PGE₂), and a variety of cytokines have been shown to play a role in the degradation of articular cartilage. Nitric oxide is another mediator.

The clinical presentation of acute lumbar radiculopathy is most often attributed to a compressed lumbar nerve root by a herniated intervertebral disc.

It is well-known and something of a paradox that some patients with large herniations have no radicular symptoms and, in contrast, some patients with no evidence of disc herniations have severe radiculopathy. While the mechanics of nerve root compression undoubtedly play a role in the pain, it probably only partially explains the exact pathophysiology of the radiculopathy.

MMPs, cytokines, and nitrous oxide

Matrix metalloproteinases (MMPs) are thought to be responsible for the turnover of the extracellular matrix within the nucleus pulposus and anulus fibrosus. Their activity is controlled on at least three levels. First, they are upregulated by cytokines such as interleukin-1 via gene expression, and also by TNF-αγ. Next, MMPs are latent in their proform, requiring activation prior to reaching their full degradative potential. And lastly, MMPs are inhibited in connective tissue by a number of TIMPs (tissue inhibitors of metalloproteinases).

MMPs come in several different varieties. The most commonly investigated ones in terms of intervertebral disc degeneration have been MMP₂ (gelatinase-α) and MMP₃ (stromelysin). Kang investigated stromelysin production as well as production of nitric oxide IL-6 and PGE₂, comparing 18 herniated lumbar discs with 8 control discs obtained from patients undergoing anterior surgery for scoliosis and burst fractures.² Kang examined gelatinase, stromelysin, as well as collagenase activity. His group found a nearly sixfold increase in gelatinase among the herniated disc samples compared to the controls. Collagenase production was absent in the control subjects and nonsignificantly elevated in the herniated discs. Caseinase (or stromelysin – MMP₃) showed an approximately fourfold increase in the herniated samples compared with the control discs. This early finding and the activity of MMPs in herniated disc samples was interesting, especially in the case of caseinase (stromelysin) which is known to degrade the core protein of cartilage proteoglycans. The progressive loss of these proteoglycans within the nucleus pulposus is believed to be one of the central reasons behind its desiccation and failure to retain its water content. The high levels found in the herniated discs probably repre sent the levels found in the degenerative discs compared to the lower level of MMP activity in the normal discs. It is likely that the smaller or lower activity of the MMPs in the normal discs reflects a basal amount of MMP activity responsible for ongoing remodeling of the disc architecture. The high MMP production in the herniated discs is likely a result of the increased inflammatory mediators produced within the discs or in the immediate area of the discs because of the inflammation.

IL-1 is known to have a positive modulating response on the MMPs. In the presence of a high IL-1 concentration and a low or relatively low TIMP concentration, the degradative enzymes may be expected to flourish. In a follow-up study to this article, Kang et al. reported on the effect of interleukin-1β on control and herniated discs using samples from the lumbar and cervical spine.³ They showed significantly elevated MMP production in the form of gelatinase and stromelysin by normal nondegenerated disc specimens after the addition of IL-1β. The basal levels of gelatinase and stromelysin were already increased in the lumbar and cervical degenerative disc specimens and the addition of IL-1β to these cultures did not significantly increase them. Collagenase activity was not detected.

An interesting control in this last study was the use of L-NMA (n-monomethyl-L-arginine) to block endogenously produced NO. Cells were cultured (control and diseased) in the presence of L-NMA in order to study the effects of endogenously produced nitrous oxide on the other mediators. When L-NMA was added to the nondegenerate control specimens that had been stimulated with IL-1β, the production of gelatinase was significantly decreased, but not the production of stromelysin. When this same effect was studied in herniated lumbar discs that were stimulated with IL-lβ, both gelatinase and stromelysin were significantly reduced. Interestingly, the same study done on the herniated cervical discs stimulated with IL-1β had no significant effect on gelatinase or stromelysin.³

Several other authors have studied MMPs and their association with intervertebral disc degeneration. Fujita et al. studied autopsy specimens of degenerative discs.⁴ They first discovered serine elastases with high activity in the endplate and nucleus pulposus of degenerative discs. Another group using a monoclonal antibody against MMP₃, found the MMP₃-positive cell ratio was significantly correlated with the magnetic resonance imaging grade of intervertebral disc degeneration. The MMP₃-positive cell ratio observed in prolapsed lumbar intervertebral discs was significantly higher than in nonprolapsed discs. The same study used an anti-TIMP₁ monoclonal antibody to demonstrate the normal presence of MMP₃ and TIMP₁ together in the degenerative intervertebral discs and hypothesized that an imbalance between MMP₃ and TIMP may induce degeneration.

IL-1 is a known mediator of mesenchymal cells and probably has a central role in disc degeneration. It is one of the key inflammatory mediators and it has been found in mononuclear cells responding to disc herniations. The studies on human disc tissue have had difficulty demonstrating IL-1β in the intervertebral disc tissues, but when disc cells were stimulated with lipopolysaccharide, elevated levels of IL-1β were found. Both MMP₂ (gelatinase) and MMP₃ (stromelysin) respond to IL-1. In an experiment using ovine disc cells, Shen et al. demonstrated the ability of IL-1 to enhance the in vitro production of MMP₂ and MMP₃ by cells of the nucleus pulposus.⁵ However, the active form of MMP₃ predominated over the active form of MMP₂ in this model of IL-1 activation. This suggests that, in the presence of IL-1 as an inflammatory mediator, MMP₃ may be more intimately involved with ongoing intervertebral disc degeneration than is MMP₂.

Therefore, the MMPs appear to be key factors in disc degeneration (Fig. 3.1). They are the active form of the enzymes that they produce, and are capable of degrading constituents of the extracellular matrix and basement membrane at physiologic pH values. Substrates for these MMPs are present in abundance in the disc: collagens II and III are substrates for MMP₁, MMP₈ and MMP₁₃, as well as their proteoglycans and other minor collagens which are substrates for MMP₂ and MMP₉. Compared with healthy discs, degenerative discs have been noted to have higher activities of not only MMP₃ and MMP₇, but also TIMP₁. MMP₃ activity has been correlated to the size of osteophytes present in disc degeneration. Inhibitors of MMPs have been found in low levels and are constitutively expressed. TIMP₂ appears to be released by most cell types within the discs, whereas TIMP₁ appears to be exclusively overexpressed in discs with degenerative disease. These expressions of MMPs and TIMPs have also been measured in spines with presumed abnormal biomechanical loading characteristics such as those with scoliosis. Handa et al. showed that proteoglycans and inhibitors of MMPs were produced in increased amounts under hydrostatic conditions when loads were increased to within a normal range.⁶ Taking these loads to abnormally high pressures resulted in decreased proteoglycan production and an increased production of MMP₃.

Fig. 3.1 MMPs as key factors in disc degeneration.

Much of the work involving the study of MMP activation and measurement has been done using tissue obtained from patients operated on for herniated discs. Therefore, many of these publications include the fact that the assay was done on disc tissue that had been presumably exposed to some type of a burst of inflammation after exposure to the epidural space. It has been proposed that patients with herniated, sequestered, or noncontained herniations may have a more severe inflammatory reaction and pain response. Nygaard et al. looked at 37 patients undergoing surgery for lumbar disc herniation.⁷ They divided the patients into those who had a bulging disc, a contained or incomplete herniation, or a noncontained or sequestered free disc fragment. Unfortunately, they were unable to recruit enough patients with bulging discs to investigate this phenomenon statistically. In looking at the two groups with the largest number of patients, including the contained herniation group which had 25 members and the noncontained herniation group which had 9 members, there was a significant difference in the mean concentrations of LTB₄, with the noncontained group having almost double the concentration versus the contained herniation group. As well, thromboxane B₂ was significantly higher in the noncontained versus the contained herniation group. Although the measured concentration of these two proinflammatory cytokines was lower in the bulging disc, their numbers were too small to be included in the statistical analysis. This study seems to support the theory that there are different inflammatory characteristics of different degrees of disc herniations.

One of the other paradoxes in the delineation of an inflammatory response for disc herniation has to do with the atypical cellular response when compared to inflammation occurring at other places in the body. Neutrophils are the sine qua non of acute inflammation; however, they have really only been found in noncontained or sequestered disc fragments where neovascularization may be occurring. Most of the cellular elements that have been identified and are proposed to be the source or factories for most of the inflammatory cytokines are macrophages. Gronblad,⁸ Nojara,⁹ Yasuma,¹⁰ and Haro¹¹ have identified macrophages as well as vascular proliferation in the granulation tissue of herniated discs. Haro additionally found that inflammatory cells were more abundant in the noncontained group of disc herniations than in the contained group. Inflammatory cells are known to act in an autocrine or paracrine type fashion with regard to their effect on resident cells in the inflammatory process. This must also be true for the degenerative disc. The intervertebral disc, which is normally nourished through diffusion, can become neovascularized to some extent after exposure to the epidural space. These discs display granulation tissue with macrophage and T-lymphocyte infiltration not observed in healthy discs. Haro et al. have proposed that the natural resorption of a herniated disc appears to occur by a vascularization-mediated process and is correlated with macrophage infiltration.

It is also known that chondrocytes replace proteoglycans within the nucleus pulposus and these cells have been proposed to play a very important role in the inflammatory process in regards to production of abnormal types of collagen as well as MMPs and TIMPs in response to abnormal loading characteristics. Haro et al. reported their results in a co-culture system of chondrocytes and macrophages and demonstrated a marked upregulation of MMP₃ by disc chondrocytes with the addition of macrophages to the culture system. This resulted in eventual resorption of the disc through macrophage action. They further used MMP-null mice to determine that the production of MMP₃ by the chondrocytes was required for macrophage infiltration in disc resorption. In a more recent study, this same group has shown that the production of MMP₇ by macrophages was found to be required for infiltration into disc tissue through a mechanism involving the release of soluble TNF-α.¹²

The support for macrophage-mediated cellular response in herniated disc tissue is also supported by another study by Haro. While macrophage invasion appears to accompany and participate in the inflammatory response, the likely end to this is reabsorption of the herniated disc tissue. Groups have proposed neovascularization of these disc tissues as the means by which this happens. Previous studies have shown that resorption may be mediated by neovascularization as measured through Gd-DTPA MRI. Komori showed that the tendency of these herniated disc tissues to spontaneously resorb was proportional to the degree of Gd-DTPA enhancement, which suggests that the resorption was mediated by a vascular event.¹³ Haro and his group have shown that, in an in vitro co-culture system they have used previously, an increase in macrophage VEGF protein (vascular endothelial growth factor) and mRNA expression was observed after they exposed disc tissue to the co-culture.¹⁴ They found TNF-α was required for induction of VEGF protein and conclude that this may be one mechanism for resorption for herniated disc tissue.

Further evidence for the involvement of the macrophage and its importance is shown in the paper by Burke et al.¹⁵ This group studied the production of monocyte chemoattractant protein-1 (MCP₁) and interleukin-8 (IL-8) by intervertebral discs removed after surgery. Burke found that MCP₁ and IL-8 were detected in both the control and herniated disc specimens and that the noncontained herniated samples contained higher levels of these chemokines than those with an intact anulus. They proposed that the MCP₁ production attracts the macrophages while IL-8 may influence the angiogenesis or the neovascularization that is seen in these samples. Although the stimulus for MCP₁ in this in vitro experiment was not investigated and is as yet unknown, this may represent a physiologic mechanism for initiation of macrophage infiltration after disc prolapse and the process of disc resorption. IL-8 was also strongly influenced by the noncontained morphology of these samples. In addition to the angiogenic properties of IL-8, it is also chemotactic for T cells that have been identified in the chronic inflammatory filtrate around disc herniations.

In addition to TNF activation of or paracrine/autocrine effects governing MMP production, TNF-αγ has long been regarded to be a key player in mediating the sensitization of nerve roots by material from the nucleus pulposus, and other effects such as edema, intervascular coagulation, reduction in blood flow, and the splitting of myelin. TNF-αγ is known to be released from the chondrocyte resident cells in the nucleus pulposus. In a local application of TNF-αγ, it induced a reduction in nerve conduction velocity in a porcine experiment done by Aoki et al.¹⁶ In this study, applications of interleukin-1β and interferon-γ induced a very small reduction of nerve velocity compared with epidural fat. In a follow-up study to this, Olmarker and Rydevik demonstrated that local blockers to TNF-αγ prevented the reduction of nerve conduction velocity and seemed to limit the nerve fiber injury and intercapillary thrombus formation, as well as the intraneural edema seen in the absence of the inhibitor.¹⁷ These authors have suggested that TNF-αγ inhibitors may be important therapeutically in the future. Presently, synthesis of TNF-αγ can be blocked with systemic corticosteroids, IL-10, TGF-βγ or by other drugs such as chlorpromazine, pentoxifylline, or ciclosporin. However, these drugs are non-specific inhibitors and may result in side effects that would be undesirable. Presently, there are anti-TNF agents being used in the treatment of rheumatoid arthritis. The first of these, infliximab (Remicade), was quickly followed by etanercept (Enbrel). Recently, a monoclonal antibody against TNF-αγ, adalimumab (Humira), has been released. These agents are not presently approved for treatment of sciatic pain, but have given sufferers of rheumatoid arthritis a further dimension for their treatment.

Another potent inflammatory mediator that is also induced by TNF-αγ is nitric oxide. Nitric oxide is a particularly interesting compound in that it has been shown to act in various ways depending on the tissues that in which it resides. In bone, mechanical stress affects intracellular cyclic AMP, calcium, and PGE₂ levels, as well as having effects on matrix synthesis. It has been demonstrated that nitric oxide is a key mediator of these processes. Articular chondrocytes have been shown to produce large amounts of nitric oxide. As described in the preceding sections introducing the inflammatory process, nitric oxide is produced in several forms including the inducible form that is present in chondrocytes. Kang et al. first showed the spontaneous production of nitric oxide from human lumbar discs and that this production was higher in herniated discs than normals.² In a follow-up study, Kang et al. examined the effects of IL-1β on normal and herniated disc tissue. They found that the addition of IL-1βγ caused a significant increase in the production of nitric oxide as well as IL-6 and PGE₂.³ While these inflammatory mediators were sharply increased in both normal and herniated disc tissue, the interesting point to this paper was that MMP production did not change in the herniation disc material, while the normal disc showed a sharp increase in the production of MMPs. It is also noted by this group that endogenously produced nitric oxide had a large inhibitory effect on IL-6.