In multiple sclerosis, the body's own immune cells attack the brain and spinal cord. But how they get there from peripheral tissues has been a mystery. Surprisingly, the lungs might be a key transit point. See Letter p.675
The immune system has one cardinal function: to protect the body against pathogens. To accomplish this task, immune cells develop in such a way that they are activated only by foreign structures. This discrimination is imperfect, however, and when it goes wrong, our bodies can suffer autoimmune inflammation and disease. In many cases, this inflammation is largely or entirely limited to a single organ — multiple sclerosis, for example, affects the brain, optic nerves and spinal cord of the central nervous system (CNS). Understanding how autoimmune cells migrate to and accumulate within an affected organ is vital to developing effective treatment strategies. On page 675 of this issue, Odoardi and colleagues1 use a rat model of multiple sclerosis to show that, before their entry into the CNS, inflammatory autoimmune cells transiently settle in the lungs — an organ not previously associated with immune-cell trafficking to the CNS.
Multiple sclerosis affects around 2.5 million people worldwide. The disease targets myelin, a membranous sheath that wraps around the axon fibres of nerve cells. Damage to myelin leads to myriad symptoms, including sensory disturbance, impaired balance and difficulty thinking. To study multiple sclerosis, researchers have developed several animal models, one of which is called adoptive-transfer experimental autoimmune encephalomyelitis (EAE). In this model, rats are first immunized with myelin basic protein (MBP), a major component of myelin. The immunization activates certain immune cells (T cells) that are specific for MBP, and induces them to proliferate. These T cells can then be collected from the immunized rats and injected into other rats, where they act as autoimmune cells and cause inflammation that centres on the spinal cord. Because these 'encephalitogenic' T cells specifically attack myelin, and not other CNS structures2, the rat disease closely mimics certain aspects of the human condition.
However, researchers have been puzzled by the fact that it takes four to five days following transfer of the encephalitogenic T cells for rats to develop disease, despite the cells being activated in vitro before transfer. The research group presenting the current paper also previously reported another strange phenomenon of the model: the T cells must be infused into the recipient rat's circulation to induce EAE; direct administration into the cerebrospinal fluid, which bathes the CNS, not only fails to hasten disease onset but is in fact ineffective in inducing the disease3.
In this earlier study, the authors showed that the injected T cells migrate from the bloodstream to immune-system organs, including lymph nodes and the spleen, and that the gene-expression pattern of the cells changes during this time: a sharp reduction in the expression of activation- and proliferation-related genes is accompanied by a striking increase in the expression of genes involved in cell migration3 (Fig. 1). Then, at the onset of disease, millions of T cells accumulate abruptly and simultaneously in the CNS3; the same research group has also characterized the events taking place at this time of CNS entry4. So the first steps following T-cell infusion, and the events immediately preceding disease onset, had both been established — but what happens to the T cells between those time points remained unclear.
Odoardi and colleagues set out to clarify this 'black box'. They used encephalitogenic T cells that express green fluorescent protein, which allows the cells to be tracked, and they developed the ability to image these fluorescent cells in the spinal cord of living animals using a technique called two-photon microscopy5, an advance that has also been made by other research groups. Imaging of the intact brain is likewise possible6,7, and together these methods have revolutionized our understanding of the processes by which immune cells gain entry to the CNS during disease and immune responses8.
The authors first asked whether the injection of activated encephalitogenic T cells could itself result in systemic inflammation that might affect the CNS or its associated vasculature, so as to induce chemical signals that attract the T cells to that region. To assess this possibility, the researchers conducted an experiment in which they joined the circulatory systems of two rats, one of which had received encephalitogenic T cells 48 hours earlier, and then monitored the arrival of the cells in the CNS of each animal. They found this to occur at approximately the same time in both rats, suggesting that the CNS of the rat that had received the cells earlier was not preconditioned to harbour an inflammatory reaction.
The researchers next sought to identify where the transferred cells reside before arriving in the CNS. Surprisingly, the two-photon microscopy experiments revealed the vast majority of the T cells to be in the rats' lungs. The cells were initially located in the lung bronchi (air passages) and alveoli (terminal air sacs), before accumulating in dense immune-cell clusters known as bronchus-associated lymphoid tissue (BALT; Fig. 1). The authors also show that, unlike encephalitogenic T cells injected into the bloodstream, T cells injected into the lungs' bronchi by means of the trachea can rapidly leave the lungs and cause disease. This finding suggests that the time the T cells spend in the lung environment 'licenses' them in a way that allows them to move to the CNS. They found, for example, that the changes in gene-expression profile previously observed in these cells were accentuated during the time spent in the lungs.
Importantly, the authors also found encephalitogenic T cells in BALT of rats that had been injected with the cells 2–3 months earlier, as newborns, and that had not developed signs of disease. When the researchers stimulated these dormant T cells by introducing MBP as an aerosol into the rats' tracheas, the cells became activated and accumulated in the CNS, causing disease.
Interpretation of this latter finding is assisted by considering the concept of tissue-resident immune memory cells, an idea stemming from the immune system's remarkable ability to mount an accelerated response to a pathogen to which it was previously exposed even decades before. Although it has long been known that memory T cells circulate through blood and lymphoid tissues, it is now clear that vast numbers of memory T cells also persist in other tissues, particularly those exposed to the outside world, such as the skin, gut and lung9. Not surprisingly, the distribution of tissue-resident memory cells mirrors sites of pathogen predilection; for example, influenza-specific memory cells are found mainly in the lungs.
Odoardi and colleagues have now shown that tissue-resident immune memory cells are not limited to pathogen-responsive populations but also include autoimmune cells, which can be activated to emigrate and mediate disease in a distant organ. Although the relevance of these findings to human disease remains speculative, distinct possibilities are evident. For example, healthy humans lack substantial BALT, but cigarette smoking, which is a potent risk factor for developing multiple sclerosis, is also known to induce BALT formation10,11,12. Furthermore, disease activity in patients with multiple sclerosis can be triggered by respiratory infections13. Both of these observations might be explained by a process in which myelin-specific autoimmune cells transit the lungs before establishing disease in the CNS, with some of these cells forming a resident lung population that can be subsequently reactivated.
Odoardi, F. et al. Nature 488, 675–679 (2012).
Kojima, K. et al. J. Exp. Med. 180, 817–829 (1994).
Flügel, A. et al. Immunity 14, 547–560 (2001).
Bartholomäus, I. et al. Nature 462, 94–98 (2009).
Denk, W., Strickler, J. H. & Webb, W. W. Science 248, 73–76 (1990).
Laudanna, C. & Constantin, G. J. Immunol. Methods 273, 115–123 (2003).
Johnston, B. et al. J. Immunol. 164, 3337–3344 (2000).
Ransohoff, R. M. Nature 462, 41–42 (2009).
Purwar, R. et al. PLoS ONE 6, e16245 (2011).
Moyron-Quiroz, J. E. et al. Nature Med. 10, 927–934 (2004).
Richmond, I. et al. Thorax 48, 1130–1134 (1993).
van der Mei, I. A. et al. Neurol. Clin. 29, 233–255 (2011).
Sibley, W. A., Bamford, C. R. & Clark, K. Lancet 1, 1313–1315 (1985).
About this article
Pulmonary Administration of Soluble Antigen Arrays Is Superior to Antigen in Treatment of Experimental Autoimmune Encephalomyelitis
Journal of Pharmaceutical Sciences (2017)
Journal of Autoimmunity (2016)
Inflammatory Bowel Diseases (2013)