The Trouble with Tribbles: Do Antibodies Against TRIB2 Cause Narcolepsy?

OVER 25 YEARS AGO, YUTAKA HONDA AND COLLEAGUES DISCOVERED THAT NARCOLEPSY IS STRONGLY ASSOCIATED WITH HLA DR2,¹ AND SINCE then, many have hypothesized that narcolepsy is caused by an autoimmune process that kills the hypocretin/orexin-producing neurons. This perspective was further reinforced last year by the observation that narcolepsy is also associated with a polymorphism in the T-cell receptor alpha gene,² but still, direct evidence for an autoimmune process has been lacking. Now, three independent studies (two of them published in this issue of SLEEP) have found that some patients with narcolepsy have elevated levels of antibodies against a protein known as Tribbles homolog 2 (TRIB2). TRIB2 is produced in hypocretin neurons, and this exciting research may be leading towards some of the firmest evidence yet for an autoimmune cause of narcolepsy.

The story began with the search by Cvetkovic-Lopes and others from the Tafti group³ for proteins that are enriched in hypocretin neurons. They identified several proteins, and among these, TRIB2 stood out as a potential autoimmune target as some patients with autoimmune uveitis have antibodies against TRIB2. They then found that among 119 European patients with narcolepsy with cataplexy, 14% had elevated levels of antibodies against TRIB2, and titers seemed higher in the first 2 years after the onset of narcolepsy.

In this issue of SLEEP, two additional groups report strikingly similar increases in anti-TRIB2 antibodies. In 90 narcoleptics with cataplexy, mainly from the US, Kawashima and colleagues⁴ found that 21% had elevated levels of antibodies against TRIB2, and most of them carried DQB1*0602. In contrast, antibody levels were increased in only 4% of narcoleptics without cataplexy and in 4% of controls. Furthermore, TRIB2 antibodies were increased in 41% of patients studied within 2.3 years of cataplexy onset, and many of these patients also had high titers of anti-streptolysin O antibodies. In the second study, Toyoda and colleagues⁵ examined 88 Japanese patients (all 0602 positive) with narcolepsy with cataplexy, and found that TRIB2 antibodies were increased in 26%, compared to only 2% of controls. Thus, three independent groups have found evidence of an immune response to TRIB2 in a sizable number of narcoleptics with cataplexy.

Though this observation is clearly a breakthrough in understanding the cause of narcolepsy, as yet, there is no direct evidence that antibodies against TRIB2 actually injure the hypocretin neurons. In fact, all three studies are consistent with three possible interpretations (Figure 1). First, TRIB2 antibodies may kill hypocretin neurons. Specifically, some inciting factor may trigger the production of anti-TRIB2 antibodies in susceptible individuals, and the antibodies directly injure the hypocretin neurons, resulting in narcolepsy. One argument against this hypothesis is that TRIB2 is seemingly widely expressed in brain and in other tissues both in mice⁶ and humans,⁷ yet it appears that only the hypocretin neurons are injured in narcolepsy.

Figure 1.

Three possible models of the role of anti-TRIB2 antibodies in narcolepsy.

Second, TRIB2 antibodies may be a consequence of hypocretin neuronal damage. In particular, after the hypocretin neurons are damaged by another mechanism, the release of normally intracellular TRIB2 protein may lead to the production of anti-TRIB2 antibodies. The actual cause of hypocretin cell death could be an unrelated autoimmune process (perhaps triggered by Streptococcal infection),⁸ a direct infection of the lateral hypothalamus by a neurotropic virus, a toxin, or a neurodegenerative process.

Third, TRIB2 antibodies may be completely incidental and unrelated to hypocretin neuronal damage. In this model, an inflammatory response may injure the hypocretin neurons and independently trigger the production of TRIB2 antibodies, but the antibodies themselves do not injure the hypocretin neurons. Any of these three scenarios could account for the association of TRIB2 antibodies with narcolepsy as well as the observation that high levels are most common soon after the onset of narcolepsy.

Whichever of these three scenarios turns out to be true, it appears that TRIB2 antibodies are neither strictly necessary nor sufficient for the development of narcolepsy. High levels of TRIB2 antibodies are elevated in a few, rare individuals without narcolepsy, suggesting that TRIB2 antibodies alone are not sufficient and that other factors must be needed. Meanwhile, the observation that most narcoleptics do not have high anti-TRIB2 levels suggests that there must be other mechanisms for developing narcolepsy. This may be analogous to the situation in myasthenia gravis, a disorder with a well-accepted, antibody-mediated mechanism. Myasthenia gravis can result from antibodies against many different proteins,⁹ as well as non-immune mechanisms.¹⁰

Archelos and Hartung have outlined five criteria to establish a causal link between an autoantibody and a neurological disease.¹¹ Applied to narcolepsy, they would be:

1. The presence of the autoantibody in the sera and ideally cerebrospinal fluid of individuals with narcolepsy but not controls.

2. The histological demonstration of immunoglobulins near the hypocretin neurons in patients with narcolepsy.

3. A beneficial response to plasma exchange.

4. The induction of narcolepsy in animals by passive transfer of the autoantibody.

5. The induction of narcolepsy by immunization of animals with purified antigen.

All five of these criteria have been satisfied for myasthenia gravis, but for narcolepsy, only the first criterion has been satisfied, and even then only in serum and not in cerebrospinal fluid. Small studies showing some benefit from intravenous immunoglobulin (IVIG) in narcolepsy¹² help satisfy criterion 3, although the mechanism of IVIG is still unclear and its efficacy has not yet been demonstrated in well-controlled trials.

The discovery of increased levels of antibodies against TRIB2 in narcolepsy with cataplexy is a significant advance, providing much needed evidence for an inflammatory process in at least some people with narcolepsy. Still, many questions must be addressed to establish whether TRIB2 antibodies are a cause or a consequence of hypocretin neuron loss. Where else in the human brain is TRIB2 expressed? Are these other neurons injured in narcolepsy? What causes narcolepsy in the majority of narcoleptics with cataplexy who seem to have normal anti-TRIB2 levels? Are antibodies against TRIB2 increased in other disorders such as Parkinson's disease and traumatic brain injury in which hypocretin neurons are lost through non-immune mechanisms?^13–15 If patients with those disorders also have high levels of TRIB2 antibodies, then they may be a result rather than a cause of hypocretin neuron loss. Much work remains to be done, but these observations provide some of the best evidence yet for an inflammatory process in narcolepsy.

DISCLOSURE STATEMENT

Dr. Scammell has consulted for Merck, GlaxoSmithKline, Novartis, Roche, Valeant, and Cephalon. Dr. Lim has indicated no financial conflicts of interest.