Bench to bedside advances in the 21st century for primary headache disorders: migraine treatments for migraine patients: Table 1

Primary headache disorders represent a remarkable opportunity for clinical neuroscience to improve patients’ lives in substantial numbers and reduce societal economic burden ( Olesen et al. , 2012 ). Of the neurological maladies ( Goadsby et al. , 2002 ; Akerman et al. , 2011 ) one could never be accused of bloviating when presenting the impact or opining on the need for investment in research in headache disorders, particularly migraine ( Shapiro and Goadsby, 2007 ). Migraine alone is the sixth leading cause of disability of any condition, and was the leading cause of neurological disability, globally in 1990; it is exactly sixth again in 2013 ( Global Burden of Disease Study, 2015 ). It seems a scandal that major funders have not addressed the issue. As an example, the US National Institutes of Health – NIH-Wide Strategic Plan ( https://www.nih.gov/about-nih/nih-wide-strategic-plan ) identifies headache disorders and migraine as the most underfunded of all conditions when the burden of the disorder is considered; yet we await any serious attempt to resolve this conundrum. However, as we move into the 21st century there is much to be excited about as we strive to treat migraine patients with migraine medicines, as opposed, particularly in preventive management, to the rag-tag collection of hand-me-downs currently available ( Goadsby and Sprenger, 2010 ). Of the strategies outlined in Table 1 , those targeted at calcitonin gene-related peptide (CGRP)-based mechanisms are clearly front-running in terms of both excitement and likely impact. The CGRP story is set out elsewhere in detail ( Ho et al. , 2010 ). In brief, studies in the 1980s ( Goadsby et al. , 1983 ; McCulloch et al. , 1986 ) led the authors to be interested in the question of whether neuropeptides, such as CGRP or substance P, could be used to track trigeminovascular activation in experimental animals in a manner that could be translated to humans. It was shown that trigeminal ganglion stimulation, which was gaining interest in a parallel fashion to explore migraine as an inflammatory disorder ( Markowitz et al. , 1987 ), could increase CGRP and substance P in the cranial circulation of cat and human ( Goadsby et al. , 1988 ). Given that stimulation of the dura mater evokes pain in humans ( Feindel et al. , 1960 ), its stimulation in experimental animals predicted elevation of CGRP and not substance P during migraine ( Zagami et al. , 1990 ). Indeed during acute migraine, CGRP was observed to be elevated in the cranial circulation, and not substance P ( Goadsby et al. , 1990 ), a finding shown later to be reversed by sumatriptan ( Goadsby and Edvinsson, 1993 ). These data predicted both the subsequent failure of substance P/neurokinin-1 receptor antagonists ( Table 1 , May and Goadsby, 2001 ), and the success of blocking CGRP mechanisms ( Table 1 ). Migraine therapies (after Goadsby, 2015 a ) F = study failed. a Liver toxicity led to termination of development. Migraine therapies (after Goadsby, 2015 a ) F = study failed. a Liver toxicity led to termination of development. For acute migraine six CGRP receptor antagonists, also called gepants, have been developed ( Table 1 ), with two being stopped for liver enzyme elevations, most notably telcagepant in a preventive study ( Ho et al. , 2014 ); two being stopped for lack of development interest; and one, ubrogepant, now in phase III ( Table 1 ). Remarkably each reported study has been positive at its primary endpoint of pain freedom compared to placebo. These medicines offer a broadly safer profile compared to triptans, serotonin 5-HT 1B/1D receptor agonists ( Goadsby, 2000 ), whose cardiovascular and cerebrovascular liability remains small but present ( Dodick et al. , 2004 ); in addition they have better tolerability. Interestingly, their population effect is, in aggregate, between 3 and 5% less than triptans, although the individual patient response is equal: responders are pain-free. These molecules offer remarkably interesting questions. Will triptan non-responders respond to gepants? Would there be an additive effect of triptans and gepants given together? Will the new molecules avoid the earlier liver toxicity issues? Will gepants be associated with medication overuse problems? Given CGRP antagonism as a principle seems anti-migraine, it would be truly a leap forward to have an acute medicine in migraine with no propensity to produce medication overuse issues. Probably more exciting is the prospect of CGRP mechanism-based preventive strategies. Remarkably, monoclonal antibodies to either CGRP peptide or its canonical receptor have proven effective in episodic migraine in five placebo-controlled, randomized studies and in chronic migraine in two studies ( Table 1 ). Even before considering the implications for migraine, it is appropriate to pause and note the monoclonal antibodies here are not being used for any form of immunomodulation, rather as pure immunopharmacological approaches to the difficult chemistry issue or peptides and their receptors. This approach has broader implications in migraine, where other chemically difficult yet implicated targets, such as pituitary adenylate cyclase-activating peptide (PACAP) ( Table 1 , Miyata et al. , 1989 ), could be targeted in the future. For migraine therapy the CGRP mechanism challenges the concept of Acute Attack and Preventive Treatment class distinction. Here one mechanism has proven capable to have both roles; a cherished notion immediately needs revision. Moreover, as with all such clinical developments the underlying science evolves. Just as Humphrey et al. (1991) had started on a search for a cranial vascular serotonin receptor when only perhaps three receptors were known; the development of sumatriptan spearheaded the identification of five times as many serotonin receptors than had been considered. Similarly, the comfortable notion of a single CGRP receptor has recently been challenged with data to suggest CGRP may act at what has been called an amylin receptor ( Walker et al. , 2015 ). Such a development offering a first glimpse at possible mechanistic, indeed therapeutic differences, between the CGRP monoclonals. Lastly, there is a very reasonable rationale to test these medicines in cluster headache ( Goadsby and Edvinsson, 1994 ). Galcanezumab ( Table 1 ) is currently being studied for the preventive management of both episodic (NCT02397473) and chronic cluster headache (NCT02438826). This would be a significant boon for patients. Lastly, and most important, will patients benefit? If only to show that mechanism-based therapies derived from bench research can ultimately work for some patients: the CGRP development is already a great success. Looking at the data, what can we expect? In episodic migraine the range of therapeutic gains (active minus placebo response rate) using the 50% responder rates for migraine/probable migraine days ranges from 17 to 31 for the CGRP monoclonal antibodies (MABS, Table 1 ). In the topiramate comparator trial propranolol’s therapeutic gain was 21 and topiramate’s 15 ( Diener et al. , 2004 ), and in the two large US studies, topiramate’s was 26 and 31, respectively ( Brandes et al. , 2004 ; Silberstein et al. , 2004 ). Considering chronic migraine, the therapeutic gains for MABs in reported studies are 16 and 24 ( Table 1 ). This compares to 10 for topiramate ( Silberstein et al. , 2006 ) and 12 in the botulinum toxin type A meta-analysis ( Dodick et al. , 2010 ). Clearly one could cite differences between these various studies, but at the end of phase II, the MABs are, as a class, comparably effective to anything we currently use. Will patients not responding to current therapies have good effects with MABs? Will there be long-term attenuation of effect? Will there be long-term side effects? We have many questions at least driven by the unmet need for treatment and the positive results of each study reported. As mentioned above, a lingering issue from the triptan era, and clouding the use of these medicines today, is cardiovascular and cerebrovascular safety. While certainly safe, in aggregate ( Dodick et al. , 2004 ), this issue remains troublesome in clinical practice. First identified in the early 1990s ( Adham et al. , 1993 ); the 5-HT 1F receptor was clearly known to be distributed in the brain ( Bruinvels et al. , 1994 ). Studies with 3 H-sumatriptan and suitable masks confirmed its presence in the brain ( Castro et al. , 1997 ). Subsequent identification that more than one triptan could interact at the 5-HT 1F receptor at concentrations that might be expected in clinical use ( Goadsby, 2000 ), set the stage for development of specific agonists. One such was identified ( Phebus et al. , 1997 ) and proved effected in a placebo-controlled trial ( Goldstein et al. , 2001 b ); unfortunately it was not progressed because of toxicology issues. It is clear from laboratory work that the 5-HT 1F receptor does not have either coronary ( Bouchelet et al. , 2000 ) or carotid constrictor effects ( Goadsby and Classey, 2003 ). Moreover, it can be shown that naratriptan, a 5-HT 1B/1D/1F agonist, has residual trigeminocervical antinociceptive effects when the 5-HT 1B/1D receptors are blocked, although the exclusive 5-HT 1B/1D receptor agonist alniditan ( Leysen et al. , 1996 ) does not ( Goadsby and Classey, 2003 ). It can be shown that a 5-HT 1F agonist has inhibitory effects on trigeminovascular nociceptive neurons in experimental animals ( Storer and Goadsby, 2011 ). Lasmiditan is a next generation 5-HT1F receptor agonist that has no vascular effects in vitro ( Nelson et al. , 2010 ). Lasmiditan has been shown by intravenous and oral administration in two placebo controlled trials ( Table 1 ) to be effective for the acute treatment of migraine. Phase III studies are now underway ( Table 1 ) with what would prove a useful addition to therapy. There is no reason to expect cardiovascular or cerebrovascular issues with this mechanism, and indeed none were seen with intravenous lasmiditan ( Ferrari et al. , 2010 ) in contrast to intravenous sumatriptan ( Doenicke et al. , 1988 ). For patients and physicians alike, a new acute migraine treatment without vascular contraindications would be a boon. Lasmiditan has reported side effects of dizziness, fatigue, vertigo and somnolence, some no doubt medicine-related, and some linked to resolution of attack, as demonstrated in analysis of the placebo data from triptan studies ( Goadsby et al. , 2007 ). Will these medicines work in patients not responding to triptans? Will they be synergistic with CGRP receptor antagonists acutely or monoclonal antibodies? Will they have a propensity for medication overuse issues, as they are receptor agonists ( Goadsby, 2006 )? Here again we can be sure that advances will bring much to patients and to the science of headache. It seems unfair to some extent to lump the plethora of other opportunities into one section, although it does reflect the development status. Some broad class comments seem reasonable. Glutamatergic targets have been considered for some time across both the ionotropic and metabotropic glutamate receptors. It is likely that the NMDA receptor has some role in migraine aura ( Afridi et al. , 2013 ) but it is too troublesome to block in migraine itself. The mGluR5 receptor is perhaps the most promising of these ( Table 1 , Waung et al. , 2016 ) and its development towards a further option for prevention would be welcome. Nitric oxide synthase (NOS) remains a tantalizing target, despite negative studies with a combination 5HT 1B/1D receptor agonist and neuronal NOS inhibitor NXN-188 ( Table 1 ). The negative study with filorexant was disappointing given the positive preclinical studies ( Hoffmann et al. , 2015 ), although the medicine was dosed once a day as it produces somnolence and preclinical data suggest a differential effect of the two orexin receptors ( Goadsby, 2015 b ). Lastly, it seems reasonably clear from at least 10 different studies ( Table 1 ) that neuroinflammatory approaches to migraine treatment, either acute or preventive, fail consistently. Perhaps this is not a great way forward. This brief walk across the landscape of migraine therapeutics is very encouraging. For the most common cause of neurological disability and certainly the most common cause for referral to neurologists in most parts of the world, new therapies are always welcome. Patients have spent countless hours of frustration wondering about the utility of medicines clearly not initially meant for them and laced with myriad caveats on side effects. Clinicians have spent much of their practising lives explaining to migraine patients variously that they do not have high blood pressure, epilepsy or depression but we are just using drugs from those classes … While the new medicines will certainly not be useful to everybody, they will improve the lives of many, and demonstrate to patients, and others so far too deaf to hear, that migraine science is on the move and great treatments are around the corner so we can treat migraine patients with migraine medicines.