Familial hemiplegic migraine (FHM) is an autosomal dominant type of hemiplegic migraine that typically includes weakness of half the body which can last for hours, days, or weeks. It can be accompanied by other symptoms, such as ataxia, coma, and paralysis. Migraine attacks may be provoked by minor head trauma. Some cases of minor head trauma in patients with hemiplegic migraine can develop into delayed cerebral edema, a life-threatening medical emergency. Clinical overlap occurs in some FHM patients with episodic ataxia type 2 and spinocerebellar ataxia type 6, benign familial infantile epilepsy, and alternating hemiplegia of childhood.

Three genetic loci for FHM are known. FHM1, which accounts for about 50% of FHM patients, is caused by mutations in a gene coding for the P/Q-type calcium channel α subunit, CACNA1A. FHM1 is also associated with cerebellar degeneration. FHM2, which accounts for less than 25% of cases, is caused by mutations in the /-ATPase gene ATP1A2. FHM3 is a rare subtype of FHM and is caused by mutations in a sodium channel α-subunit coding gene, SCN1A. These three subtypes do not account for all cases of FHM, suggesting the existence of at least one other locus (FHM4).

Also, nonfamilial cases of hemiplegic migraine are seen, termed sporadic hemiplegic migraine. These cases seem to have the same causes as the familial cases and represent de novo mutations. Sporadic cases are also clinically identical to familial cases with the exception of a lack of known family history of attacks.

Signs and symptoms FHM signs overlap significantly with those of migraine with aura. In short, FHM is typified by migraine with aura associated with hemiparesis, and in FHM1, cerebellar degeneration, which can result in episodic or progressive ataxia. FHM can also present with the same signs as benign familial infantile convulsions and alternating hemiplegia of childhood. Other symptoms are altered consciousness (in fact, some cases seem related to head trauma), gaze-evoked nystagmus, and coma. Aura symptoms, such as numbness and blurring of vision, typically persist for 30–60 minutes, but can last for weeks to months. An attack resembles a stroke, but unlike a stroke, it resolves in time. These signs typically first manifest themselves in the first or second decade of life.

Causes See the equivalent section in the main migraine article.

FHM mutations are believed to lead to migraine susceptibility by lowering the threshold for cortical-spreading-depression generation. The FHM1 and FHM3 mutations occur in ion channels expressed in neurons. These mutations may lead to both the hyper- and hypoexcitable neurons that might underlie cortical-spreading-depression. How the mutations seen in FHM2 patients might lead to FHM symptoms is even less clear, as the gene mutated in FHM2 is expressed primarily in astrocytes. One proposal states that the depolarization of astrocytes caused by haploinsufficiency of the ATP1A2 /-ATPase causes increased release of compounds such as adenosine from astrocytes. These compounds then interact with neighboring neurons, altering their excitability and leading to cortical-spreading-depression and migraine.

Pathophysiology

FHM1 (CACNA1A) The first discovered FHM locus was the CACNA1A gene (originally named CACNL1A4), which encodes the P/Q-type calcium channel CaV2.1. Currently, 17 mutations in this channel are known (table 1), and these mutations are distributed throughout the channel. Some of these mutations result in patients with notable cerebellar degeneration or other dysfunction, including one mutation (S218L), which may be related to severe responses to mild concussion, up to and including delayed cerebral edema, coma, and death. Fifteen of these mutants have received at least some further analysis at the electrophysiological level to attempt to determine how they might lead to the FHM1 phenotype. Contradiction in the literature is increasing as to the end result of these mutations on channel kinetics and neuronal excitability.

A good example of this contradiction can be seen in the literature regarding the R192Q mutation. The first investigation of this mutation, using the rabbit isoform of the channel expressed in oocytes, found that it did not alter any measured channel properties. A subsequent report, using human channels expressed in HEK293 cells, found a small, hyperpolarizing shift in the midpoint for activation, a result common among FHM1 mutants. This shift results in channels that open at more negative potentials, thus have a higher open probability than wild-type channels at most potentials. This report also found that the R192Q mutant produced almost twice as much whole-cell current compared to wild-type channels. This is not due to a change in single channel conductance, but to an equivalent increase in channel density. A subsequent group noticed that this mutation is in a region important for modulation by G protein-coupled receptors (GPCRs). GPCR activation leads to inhibition of wild-type CaV2.1 currents. R192Q mutant channel currents are also decreased by GPCR activation, but by a smaller amount. A more recent group has confirmed some of these results by creating a R192Q knock-in mouse. They confirmed that the R192Q mutant activates at more negative potentials and that neurons producing these channels have much larger whole-cell current. This resulted in a much larger quantal content (the number of neurotransmitter packets released per action potential) and generally enhanced neurotransmitter release in R192Q-expressing neurons versus wild-type. Consequently, these mutant mice were more susceptible to cortical-spreading-depression than their wild-type counterparts. The most recent experiments on this mutant, however, have contradicted some of these results. In CaV2.1 knockout neurons transfected with human channels, P/Q-type currents from mutant channels are actually smaller than their wild-type counterpart. They also found a significant decrease in calcium influx during depolarization, leading to decreased quantal content, in mutant versus wild-type expressing neurons. Neurons expressing mutant channels were also less able to mediate inhibitory input and have smaller inhibitory postsynaptic currents through P/Q-type channels. Further testing with this and other mutants is required to determine their end effect on human physiology.

+Table 1. Summary of mutations in CACNA1A found in patients diagnosed with FHM type 1 Mutation Position Effect Cerebellar signs Reference Nucleotide Amino acid c.G575A R192Q D1S4 Decreases G-protein mediated inhibition, activates at more negative potentials, increased expression, faster recovery from inactivation. In mice: greater current, activates at more negative potentials, enhances transmitter release ? c.G584A R195K D1S4 No c.C653T S218L D1S4-5 Increases sojourns to subconductances, activates at more negative potentials, decreased slow inactivation, increased fast inactivation Yes c.G1748A R583Q* D2S4 Activates at more negative potentials, faster current decay, faster inactivation, slower recovery from inactivation Yes c.C1997T T666M D2-pore Activates at more negative potentials, faster current decay, slowed recovery from inactivation, smaller single channel conductance, higher i*Po, slower recovery from inactivation, Increased G-protein mediated inhibition, decreased gating charge (fewer channels available to open) Yes c.T2141C V714A D2S6 Activates at more negative potentials, faster current decay, faster recovery from inactivation, decreases expression, faster recovery from inactivation, increases G-protein-mediated inhibition No c.C2145G D715E D2S6 Activates at more negative potentials, faster current decay, faster inactivation Yes c.A4003G K1335E D3S3-4 Activates at more negative potentials, inactivates at more negative potentials, slowed recovery from inactivation, increased frequency dependent rundown No c.G4037A R1346Q D3S4 Yes c.A4151G Y1384C D3S5 Yes c.G4366T V1456L D3-pore Activates at more negative potentials, slower current decay, slower recovery from inactivation No c.C4636T R1546X** D4S1 Decreased current Yes c.C4999T R1667W D4S4 Yes c.T5047C W1683R D4S4-5 Activates at more negative potentials, inactivates at more negative potentials, slowed recovery from inactivation, increased frequency dependent rundown Yes c.G5083A V1695I D4S5 Slowed recovery from inactivation, increased frequency dependent rundown No c.T5126C I1709T D4S5 Yes c.A5428C I1810L D4S6 Activates at more negative potentials, faster recovery from inactivation, decreased expression, faster recovery from inactivation, Increased G-protein mediated inhibition Yes *** Also diagnosed as spinocerebellar ataxia type-6Also diagnosed as episodic ataxia type-2 Sequence numbering according to NCBI reference sequence NM_000068.2. Cerebellar signs refers to findings of cerebellar degeneration or ataxia upon clinical examination.

FHM2 (ATP1A2)

The second subtype of familial hemiplegic migraine, FHM2, is caused by mutations in the gene ATP1A2 that encodes a /-ATPase. This /-ATPase is heavily expressed in astrocytes and helps to set and maintain their reversal potential. Twenty-nine known mutations in this gene are associated with FHM2 (table 2), many clustering in the large intracellular loop between membrane-spanning segments 4 and 5 (figure 1). Twelve of these mutations have been studied by expression in model cells. All but one have shown either complete loss of function or more complex decreases in ATPase activity or potassium sensitivity. Astrocytes expressing these mutant ion pumps will have much higher resting potentials and are believed to lead to disease through a poorly understood mechanism.

+Table 2. Summary of mutations in ATP1A2 found in patients diagnosed with FHM type 2MutationLocationPhysiological resultReference(s) E174K M2-3 No change T263M M2-3 G301R M3 T345A M4-5 Decreased K influx T376M M4-5 R383H M4-5 T378N M4-5 C515Y M4-5 Loss of function (haploinsufficiency) R548H M4-5 R593W M4-5 Loss of function (haploinsufficiency) A606T M4-5 G615R M4-5 Loss of function (haploinsufficiency) V628M M4-5 Loss of function (haploinsufficiency) R689Q M4-5 Decreased catalytic turnover E700K M4-5 D718N M4-5 Loss of function (haploinsufficiency) M731T M4-5 Decreased catalytic turnover R763H M4-5 Loss of function (haploinsufficiency) L764P M4-5 Loss of function (haploinsufficiency) P796R M5-6 M829R M6-7 R834Q M6-7 W887R M7-8 Loss of function (haploinsufficiency) E902K M7-8 935K_940SdelinsI M8-9 R937P M8-9 S966LfsX998 M9 P979L M9-10 X1021RextX28 C-Terminus Numbering according to the NCBI reference sequence NM_000702.2.

FHM3 (SCN1A) The final known locus FHM3 is the SCN1A gene, which encodes a sodium channel α subunit. The only study so far that has found mutations in this gene discovered the same Q1489K mutation in three of 20 families (15%) with 11 other kindreds (55%) already having mutations in CACNA1A or ATP1A2. This mutation is located in a highly conserved region of an intracellular loop connecting domains three and four. This mutation results in a greatly hastened (two- to four-fold) recovery from inactivation compared to wild-type. As this channel is important for action potential generation in neurons, the Q1489K mutant is expected to result in hyperexcitable neurons.

FHM4 candidate The axonal protein PRRT2, associated with the exocytosis complex, may be a fourth causal gene for FHM, but this is still under debate.

Other genetic associations Other genes associated with this condition are proline-rich transmembrane protein 2 (PRRT2) and SLC4A4, which encodes the electrogenic NaHCO3 cotransporter NBCe1.

Diagnosis Diagnosis of FHM is made according to these criteria: Two attacks of each of: