BACKGROUND HistoryNeuromyelitis optica (abbreviated NMO) is a rare CNS conditionoften affecting both the spinal cord and the optic nerve.
While the majority ofadvancements in this field has been made only in the past two decades,neuromyelitis optica traces its roots all the way back to the early 19thcentury. In 1804, Louis XVIII’s physician Antoine Portal (1742-1832) reportedvisual loss in a patient with spinal cord inflammation but no brain pathology;this represented the first ever account of its kind in Western literature (Jariusand Wildemann, 2012). From then on, various physicans have reported cases ofsimilar symptoms in their patients: Giovanni Battista Pescetto (1806-1884) in1844, Christopher Mercer Durrant (1814-1901) in 1850, and Jacob AugustusLockhart Clarke (1817-1880) in 1862. It wasn’t until 1894, however, when Frenchneurologist Eugène Devic (1858-1930) finally gave this syndrome (characterisedby optic neuritis and acute myelitis) a name: “neuromyelitis optica” (Jariusand Wildemann, 2013). Devic’s contribution to the discovery of this conditionis still recognised to this day; neuromyelitis optica is also known as ‘Devic’sdisease’ or ‘Devic’s syndrome’. Devic believed that neuromyelitis optica was adisease in its own right;however, over the years, people have only considered it to be a variant ofmultiple sclerosis (Papadopoulos and Verkman, 2012). It wasn’t until after theturn of the century when a major discovery definitively distinguished NMO frommultiple sclerosis. Clinical Features& DemographicsTraditionally,neuromyelitis optica could be diagnosed when the following were present: opticneuritis, acute myelitis, and at least two of three supportive criteria(contiguous spinal cord MRI lesion extending over ³3 vertebral segments; brain MRI not meetingcriteria for multiple sclerosis; NMO-IgG seropositive status) (Wingerchuk et al.
, 2006). According to Papadopoulosand Verkman’s review article (2012), neuromyelitis optica affects approximately0.3-4.4 per 100,000 individuals.
It is much more prevalent in women than in men;up to 90% of all NMO patients are female (Wingerchuk et al., 2007). Comparison toMultiple SclerosisWhile neuromyelitis optica and multiple sclerosisshare a number of clinical and radiological features (Jarius and Wildemann,2013) such as transverse myelitis and spinal cord lesions, there are manydifferences between them. A review article by Wingerchuk et al. (2007) described the median age of onset for neuromyelitisoptica to be 39 years-old, compared to 29 for multiple sclerosis; the majority(80-90%) of NMO patients have relapsing episodes of myelitis and opticneuritis, whereas MS patients usually have milder attacks with good recovery (only15% of MS patients are primary-progressive). Wingerchuk et al. also expressed that the distinction of neuromyelitis opticafrom multiple sclerosis can be made further by the use of laboratory studies,specifically the analysis of the cerebrospinal fluid: NMO-specific myelitis ischaracterised by prominent cerebrospinal fluid pleocytosis (>50´106leucocytes/L) with a predominance of neutrophils; on the other hand, attacks ofmultiple sclerosis typically involve much milder CSF pleocytosis with a highproportion of lymphocytes instead of neutrophils.
Finally, excessiveoligoclonal bands of IgG (indicating intrathecal immunoglobulin synthesis) isfound in the cerebrospinal fluid of 15-30% of NMO patients, compared to 85% ofMS patients (Wingerchuk et al.,2007). Discovery ofNMO-IgGIn the year 2004, Lennon et al. discovered acirculating autoantibody in a group of patients with neuromyelitis optica whichwas absent in MS patients; this biological marker was termed NMO-IgG (orAQP4-IgG). This breakthrough not only revolutionised our understanding of thedisease (Papadopoulos and Verkman, 2012), but it also gave conclusive evidencethat neuromyelitis optica was distinct from multiple sclerosis. In their study,Lennon et al.
tested serum samplesfrom North American and Japanese patients with suspected neuromyelitis opticaand multiple sclerosis. They found that NMO-IgG was present in 73% of thosediagnosed with NMO, 46% of those classified as high risk candidates for NMO,and 0% of those diagnosed with classic MS or miscellaneous autoimmune andparaneoplastic neurological disorders; furthermore, they concluded that NMO-IgGbinds at the blood-brain barrier (specifically microvessels in the CNS, pia,subpia, and Virchow-Robin space) (Lennon etal., 2004). It wasn’t particularly surprising to find that neuromyelitisoptica was an antibody-mediated disease, however; 78% of all patients withautoimmune diseases are women (Fairweather, Frisancho-Kiss and Rose, 2008),which is similar to the demographics of neuromyelitis optica. Discussion of AquaporinsAquaporins are integralmembrane proteins that selectively allow the passive transport of watermolecules. While there are thirteen known classes of aquaporins, the mostabundant water channel in the central nervous system is aquaporin-4 (AQP4),which is found in the perimicrovessel astrocyte foot processes, glia limitans,and ependyma (Saadoun and Papadopoulos, 2010).
The structure of aquaporin-4consists of four monomers, each with six helical transmembrane domains and twoshort helical segments surrounding an aqueous pore (Verkman et al., 2013). A key article by Saadoun et al.
(2005) discovered that aside fromfacilitating water movement into and out of the brain, aquaporin-4 plays a keyrole in enhancing astroglial cell migration in glial scar formation. The significanceand relevance of this to neuromyelitis optica will be discussed further in thearticle. Discussion of AstrocytesAstrocytes (astron = star and kytos = cell in Greek) are star-shaped glialcells found in the central nervous system. Also known as astroglia, theyconsist of a cell body (soma), a high number of branched processes, and endfeet at the end of each process.
Astrocytes are arguably one of the mostimportant cells of the central nervous system; a few of their many functionsinclude providing structural support for neurons and other glial cells, maintaininginterstitial fluid homeostasis by regulating ion concentration, clearingsynapses of used neurotransmitters, forming a glial scar as a response toinjury, and contributing to the blood-brain barrier. Discussion of the Complement SystemThe complement system is apart of the innate immune system that promotes inflammatory responses and opsonisespathogens to fight infection; its name comes from the fact that it ‘complements’and enhances the action of antibodies and phagocytic cells (Janeway et al., 2001). This system consists ofmany different plasma proteins, called “complement proteins”, which activate alarge-scale complement cascade at the onset of infection. Due to thepotentially dangerous nature of a pathway that leads to such potentinflammation and destructive effects, tight regulatory mechanisms must be putin place; for this reason, complement regulators are present at many points inthe complement cascade (Janeway et al.,2001). PATHOPHYSIOLOGY Pathogenicity of NMO-IgGA study by Hinson et al. (2007) evaluated the selectivityand consequences of immunoglobulins binding to target cells expressing aquaporin-4.
Using confocal microscopy and flow cytometry, they not only found that serumIgG (but not IgM) from patients with neuromyelitis optica binds to aquaporin-4and initiates both aquaporin-4 endocytosis/degradation and also complementactivation, but also found that aquaporin-4 is highly expressed at paranodalastrocytic endfeet; from these results, they concluded that NMO patients’ serumIgG has a selective pathologic effect on cell membranes expressing aquaporin-4(Hinson et al., 2007). Binding of NMO-IgG to Aquaporin-4In humans, aquaporin-4channels are expressed as two different isoforms formed by alternative splicing:a full-length isoform (named M1) with translation initiation at Met-1, and ashort isoform (named M23) with translation initiation at Met-23 (Jin, Rossi andVerkman, 2011). The significant difference between these two isoforms is that whileM23 forms supramolecular assemblies called orthogonal arrays of particles(OAPs), M1 does not do the same, unless it coassembles with M23 (Jin, Rossi andVerkman, 2011). Multiple studies have found that in the serum of patients withneuromyelitis optica, binding of NMO-IgG occurs more with cells expressing M23than with M1; these studies suggest that NMO-IgG preferentially binds to OAPs (Papadopoulosand Verkman, 2012).
NMO-IgG and Complement ActivationWhile the NMO-IgG antibody isthe main marker for neuromyelitis optica, there is one other key component thatmust also be present for the formation of NMO lesions: complement proteins. Astudy by Saadoun et al. (2010)involving mouse models discovered that NMO-IgG alone was unable to producelesions in mouse; however, once human complement was co-injected, they observedlesions with the following characteristic histological features of human NMOlesions: inflammatory cell infiltration, demyelination, loss of aquaporin-4 andGFAP (glial fibrillary acidic protein) expression, and perivascular depositionof activated complement components. Furthermore, since NMO-IgG with humancomplement did not produce NMO-like lesions in AQP4-null mice, it was confirmedthat autoantibodies to aquaporin-4 were indeed responsible for these lesions,instead of another autoimmune component of the IgG preparations (Saadoun et al., 2010).
Downstream Effects of Complement ActivationBoth complement-dependentand antibody-dependent cell-mediated cytotoxicities are thought to be presentin the pathogenesis of neuromyelitis optica (Nishiyama et al., 2016). Thought to be the principal mechanism ofcytotoxicity in neuromyelitis optica, complement-dependent cytotoxicity isgreatly enhanced in aquaporin-4 channels assembled in OAPs; on the other hand,antibody-dependent cell-mediated cytotoxicity involves natural killer cells andwas found not to be dependent on OAP formation (thus can be present in both M1and M23-expressing cells) (Phuan et al.,2012). This complement-dependent cytotoxicity pathway of neuromyelitis opticabegins with the multivalent interaction between complement protein C1q andarray-assembled NMO-IgG (Phuan et al.
,2012). The activation of theclassical complement pathway at the perivascular astrocytic end-feet results ina number of downstream effects. A study by Lucchinetti et al. (2002) proposed that this complement activation leads to therecruitment of activated macrophages that locally generate cytokines, proteasesand oxygen/nitrogen free radicals; this results in the non-selectivedestruction of both grey and white matter, including axons andoligodendrocytes.
This explains the demyelinating feature of neuromyelitisoptica, as oligodendrocytes are glial cells responsible for the myelination ofneuronal axons in the central nervous system. As shown by Lucchinetti et al., this axonal degeneration leadsto neuronal necrosis.
Furthermore, there is intense perivascular and meningealinfiltration by eosinophils and neutrophils in the spinal cord; these activatedeosinophils release cytotoxic granule proteins such as MBP, eosinophil-derivedneurotoxin, eosinophil cationic protein and eosinophil peroxidase (Lucchinetti et al., 2002). Animal ModelsEver since the discovery ofthe NMO-IgG auto-antibody (Lennon et al.,2004), many animal models have been developed to further investigate thepathogenesis of neuromyelitis optica. However, it is surprisingly difficult tomimic this disease in other organisms; for instance, the simple transfer ofAQP4 antibodies into naïve animals was insufficient to induce experimental NMO(similar to how human patients may possess anti-AQP4 antibodies for many yearsbefore showing any clinical signs of NMO) (Bradl and Lassmann, 2014).
Furthermore, NMO-IgG only recognises conformational epitopes of aquaporin-4that require immunogen consisting of properly folded AQP4 (which is insolubleand can only be solubilised by toxic concentrations of triton); this can’t beachieved by using recombinant proteins during immunisation (Bradl and Lassman,2014). Despite these challenges, many models have crucially improved ourunderstanding of this disease: autoimmunity to myelin oligodendrocyteglycoprotein (MOG) in rats was found to mimic the pathology in classic multiplesclerosis and neuromyelitis optica (Storch etal., 1998); in addition to antibodies to AQP4, complement proteins mustalso be present for the formation of NMO lesions (Saadoun et al., 2010); and more recently, interleukin-1 beta released inNMO lesions was found to facilitate neutrophil entry and blood-brain barrierbreakdown (Kitic et al., 2013). CLINICAL IMPLICATIONS Importance of Understanding Pathophysiology for Developing TreatmentsUnderstanding thepathophysiology of neuromyelitis optica is essential for the development ofnovel treatments, as it gives us insight into where potential therapeutictargets may be.
For example, our understanding of the presence of anti-AQP4antibodies led to the routine usage of plasmapheresis in treating this disease;one trial found that anti-AQP4 levels decreased by 85% after plasmapheresis wasused in acute attacks (Kim et al.,2013). Next, the discovery that complement was required for the formation oflesions suggested that complement inhibitors such as C1 inhibitor may be aneffective way to limit CNS injury in neuromyelitis optica (Saadoun et al.
, 2010). Finally, recent evidence havepointed to B-cell-mediated humoral immunity in the pathogenesis ofneuromyelitis optica; this led to the usage of Rituximab, an antibody againstthe CD20 antigen on B-cells, which profoundly depletes B-cells and decreasesthe frequence and severity of NMO attacks (Etemadifar et al., 2017). While the treatment methodsabove have led to marked improvement in patients with neuromyelitis optica, newtreatment options are constantly being explored. One of the most prominent approachesstill in their preclinical phase would be aquaporumab, a highly-selective,nonpathogenic human monoclonal antibody that competes against NMO-IgG to bindto aquaporin-4 channels; aquaporumab greatly reduced NMO-IgG-dependentcytotoxicity in animal and in vitromodels of neuromyelitis optica (Papadopoulos, Bennett and Verkman, 2014). Someexisting therapeutic strategies that target complement proteins, neutrophils,and eosinophils (initially developed for other indications) are currently underclinical evaluation for the repurposing for neuromyelitis optica (Papadopoulos,Bennett and Verkman, 2014).
Modified Diagnostic CriteriaRecent advancements in thefield of neuromyelitis optica research had rendered the previous diagnosticcriteria from 2006 inadequate for contemporary practice; the InternationalPanel for NMO Diagnosis (IPND) was thus assembled to develop revised diagnosticcriteria that included non-opticospinal clinical and MRI characteristics(Wingerchuk et al., 2015).Furthermore, the term ‘neuromyelitis optica’ would be merged with’neuromyelitis optica spectrum disorder’ (NMOSD), which was previously used forpatients who didn’t necessarily fit under the traditional criteria of NMO butwere vulnerable to future attacks. Table 1. (Wingerchuk et al., 2015) NMOSD diagnostic criteriafor adult patients Seronegative NMO & Anti-MOG AntibodiesPatients who don’t testpositive for the anti-AQP4 antibody may still be diagnosed with neuromyelitisoptica, but are referred to as being ‘seronegative’.
As it has been many yearssince the initial discovery of AQP4-IgG, anti-AQP4 assay sensitivity has beenimproved to near 90%; thus the concept of seronegative NMO has been challenged(Levy, 2014). Scientists and clinicians have attempted to explore thisso-called ‘seronegative NMO’ by identifying a subpopulation of patients withclinical features of neuromyelitis optica while testing positive for antibodiesagainst myelin oligodendrocyte glycoprotein (MOG); they appeared to have aslightly different disease phenotype with a demographic and clinical profilemore similar to patients with acute demyelinating encephalomyelitis: whilefemales comprised the vast majority of the anti-AQP4-seropositive group,anti-MOG seropositivity was more common in males; episodes were more severe inthe anti-MOG-seropositive group but recovery was better and more likely to bemonophasic, compared to those with anti-AQP4 seropositivity (Levy, 2014). Whilethe ambiguity of seropositive vs. seronegative NMO had posed as an issue, thenewly-revised diagnostic criteria for NMOSD in 2015 have tackled this problemby clarifying the requirements for the diagnosis of NMOSD without the presenceof AQP4-IgG. Organs Outside the Central Nervous SystemIn addition to the centralnervous system, aquaporin-4 channels are also expressed in the epithelium ofmany organs in the human body such as the kidney, intestine, salivary glands,sensory organs, and skeletal muscles (Gleiser et al., 2016).
However, these peripheral organs are usuallyunaffected in patients with NMOSD. This phenomenon can be explained by the roleof complement regulators. Astrocytic end-feet in normal brain lack complementregulators CD46, CD55 and CD59, while the kidneys, stomach and skeletal muscleexpress one or more of these complement regulators; this suggests why theseperipheral organs are generally spared from AQP4-IgG and complement-mediateddamage (Saadoun and Papadopoulos, 2015). However, the impact of NMOSD onperipheral organs warrants further investigation, as a recent study has foundthat muscle damage occurs in patients with NMOSD and is aggravated during theacute phase (Chen et al., 2017). Neuromyelitis Optica and PregnancyAs the vast majority of NMOSDpatients are female, it’s important to discuss how NMOSD affects pregnancy. Apassive mouse transfer study proposed that AQP4-IgG increases miscarriage ratesby binding to the placental syncytiotrophoblast of fetal villi and activatingthe classical complement pathway; this is followed by the deposit of C5b-9 ontothe syncytiotrophoblast plasma membrane, thus causing damage and loss of AQP4expression, followed by leukocyte infiltration into the placenta, releasingelastase and other proteases that inflame and necrotise the placenta,ultimately causing fetal death (Saadoun etal., 2013).
The link between neuromyelitis optica and pregnancy outcome wasfurther investigated in a retrospective study by Nour et al. (2016), where they found the miscarriage rate to be higherafter NMOSD onset (42.9% compared to 7.04% before NMOSD onset). They concludedthat pregnancy after NMOSD onset is a risk factor for marriage (Nour et al., 2016) and provided furtherevidence pointing to the destructive effects of AQP4-IgG on the placenta. FUTURE DIRECTIONS While the field ofneuromyelitis optica research has rapidly evolved over the past one-and-halfdecades, there is still major room for further development.
It will beinteresting to explore whether there are other antibodies that causeAQP4-IgG-like damage; the presence of anti-MOG antibodies in some patientssuggests that even more types of antibodies may be present. In addition, sinceneuromyelitis optica is currently incurable, the development of novel therapeuticdrugs (e.g. aquaporumab, as mentioned earlier) is essential.
Anotherpossibility is to modify complement regulators in order to protect the centralnervous system from complement-mediated astrocytopathy. Other questions to exploremay be how AQP4-IgG gains entry into the blood-brain barrier, why AQP4-IgGforms in the first place, and whether there are other aquaporinopathieselsewhere in the body. A link between neuromyelitis optica and myastheniagravis (a chronic, autoimmune neuromuscular disease characterised by skeletalmuscle weakness) shoud be studied. Since the thymus is often enlarged inpatients with myasthenia gravis, it will be interesting to observe how thethymus differs in patients with NMOSD from those without.
Since no treatmenthas been proven to be safe and effective for NMOSD in any clinical trials(Weinshenker et al., 2015), moreinterventional studies must be done in order for the development of novelviable therapeutic agents to be effective. Perhaps the biggest challenge of implementingplacebo-controlled trials is the severity of NMOSD attacks; however, innovativeapproaches such as shared placebo groups may overcome the hurdles associatedwith designing such trials (Weinshenker etal., 2015).Furthermore, another majordifficult aspect of designing NMOSD clinical trials is the rarity of thisdisease; since more common autoimmune disorders (such as multiple sclerosis)often overshadow neuromyelitis optica, it deserves more attention and awarenessthan it is currently getting.