KIF26A is an atypical member of the kinesin superfamily proteins (KIFs). It lacks the motor ATPase activity and functions as a signaling regulator rather than a molecular motor. Here I show that mice with a homozygous deletion of Kif26a were suffered from hyperalgesia, showing increased sensitivity to noxious heat and mechanical stimuli. Morphologically, peripheral innervation and branching of DRG neuronal axons were enhanced. In the cultured dorsal root ganglion (DRG) neurons of Kif26a(-/-) mice, capsaicin-triggered [Ca(2+)]i elevation was increased and abnormally persisted after capsaicin was washed out. Consistently, expression level and phosphorylation level of TRPV1 channel that sensitizes this capsaicin receptor, and tyrosine phosphorylation level of PMCA calcium pump that suppresses the calcium efflux, were both significantly elevated. To the upstream to these phosphorylations and abnormal development, MAPK and PI3K responses to capsaicin or NGF stimulation were significantly exaggerated in Kif26a(-/-) DRG neurons. KIF26A was associated with Grb2 and inhibited its interaction with SH3 in DRG cells to negatively regulate the MAPK and PI3K signal transductions. Therefore, deficiency of KIF26A resulted in the hypersensitivity of nociceptor neurons. These results propose a new role of KIF26A in the function and development of sensory neurons.

In this study I have reexamined the behavior of Kif26a(-/-) mice and first describe the hyperalgesia of these animals. I show that deletion of the Kif26a gene led to hyperalgesia, which is abnormal pain activity to noxious heat and mechanical stimuli. In situ hybridization revealed high-level expression of KIF26A in DRG neurons. Through physiological analyses of primary cultured DRG neurons, I propose a new role of KIF26A in eliminating the capsaicin-induced intracellular calcium elevation, possibly due to elevated PI3K and MAPK signaling. Furthermore, KIF26A deficiency led to peripheral innervation and branching of sensory nerves. These results suggested a new role of KIF26A in modulating the intracellular signals essential for development and function of the PANs.

According to behavioral analyses of these mice, sensory responses were significantly exaggerated in Kif26a(-/-) mice (KO). These results suggested that Kif26a(-/-)mice are hypersensitive to the noxious stimuli. In order to elucidate the responsible tissue of this hyperalgesia, I examined the expression of Kif26a in DRG. In situ hybridization using a specific probe revealed significant expression of Kif26a mRNA in DRG of WT mouse pups of postnatal 1 day old, compared with that of KO, suggesting that dysfunction of DRG neurons could cause the hyperalgesia of KIF26A-deficient mice.

Because the previous report described abnormalities in the enteric nervous system development, I investigated possible changes in development of DRG neurons. First, cryosections of Kif26a(-/-) and Kif26a(+/+) ganglia were compared using a pan-neuronal marker Islet-1. As a result, no apparent morphological changes were found at the age of postnatal 1 day. Furthermore, the total number of Islet-1+ neurons in T10 DRG was not significantly changed. These data suggested that KIF26A is dispensable for the regulation of the number of mouse DRG neurons at this stage.

Previous studies have shown that DRG neuron size was increased after exposure to exogenously applied NGF. Transgenic mice overexpressing NGF in skin also showed significant hypertrophy of TrkA positive neurons in DRG. To quantify the changes in neuronal size, the soma area of TrkA positive as well as the CGRP positive neurons were measured in DRG using sections from both KIF26A mutant and WT mice. The results of size-frequency analysis indicated that the size of the TrkA+ and CGRP+ neurons was increased in Kif26a(-/-) DRG. This result also suggested that KIF26A suppress the NGF-TrkA signaling in the DRG sensory neurons. To further study the morphological changes of Kif26a(-/-) DRG, I observed the innervation of free nerve ending in the footpad skin of Kif26a(-/-) and WT mice. The number of fibers innervating to the epidermis of Kif26a(-/-) footpad skin at postnatal 7 days was significantly increased compare to that of Kif26a(+/+)skin. I next investigated whether KIF26A deficiency could affect axonal extension or branching in vivo. A whole-mount immunofluorescent staining assay using antibody against neurofilament-M was performed. I found the brancing of the NF-M positive fibers was increased in Kif26a(-/-) paws relative to controls at embryonic-14.5-day. The number of NF-M positive branches per nerve trunk and the number of orders of branching were both increased in KIF26A mutants compared to WT. However, the central projections of different nociceptive neurons innervating the dorsal horn of the spinal cord were not affected in Kif26a(-/-) and control mouse lines at postnatal 1day. These findings suggest that KIF26A negatively regulates the innervation and branching of peripheral but not central projection of DRG sensory neurons.

Next, I investigated the physiological performance of primary cultured DRG neurons on postnatal 15 days by measuring the [Ca(2+)]i response challenged by capsaicin, because the ion-channel TRPV1 is believed to be a major sensor of noxious heat. Rapid activation of TRPV1 receptor was studied by monitoring the increase in intracellular calcium concentration [Ca(2+)]i following to exposure to capsaicin. I stimulated the neurons with Capsaicin using a perfusion system on the confocal microscopy and measured the time course of [Ca(2+)]i elevation using the fluorescence intensity of calcium indicator, Fluo-4, AM. Perfusion on the neurons with MEM containing 1 μM capsaicin resulted in an abrupt elevation of [Ca(2+)]i levels similarly in both Kif26a(-/-) and Kif26a(+/+) neurons. After short stimulation with capsaicin, the elevated [Ca(2+)]i level rapidly decreased to the basal level in the wild-type neurons after intensive wash with the complete medium. Surprisingly, in the Kif26a(-/-) neurons, the [Ca(2+)]i elevation was persisted significantly. This cellular abnormality of Kif26a(-/-) neurons can explain the hyperalgesia at the individual level.

As a molecular basis of these elevated and persistent calcium responses in Kif26a(-/-) DRG neurons, I respectively conducted phosphorylation assays of TRPV1 and PMCA proteins. First, it has been reported that expression and phosphorylation levels of TRPV1 regulate the sensitivity of TRPV1 responsible for the sensation of noxious heat stimuli. The protein amount and phosphorylation level of TRPV1 were compared in DRG neurons of Kif26a(-/-) and Kif26a(+/+) 15-day-old mice. The expression levels of TRPV1 and pTRPV1 were compared after the amount of samples was normalized by the amount of KIF5A by immunoblotting. As a result, I found that the amount of TRPV1 was slightly increased in Kif26a(-/-) DRGs while phosphorylation levels of TRPV1 was significantly increased in Kif26a(-/-) DRGs. This could explain the molecular mechanism for enhanced capsaicin response of DRG neurons. Second, the plasma membrane Ca(2+) ATPase (PMCA) provides the predominant mechanism for extrusion of Ca(2+) from cytoplasm of DRG neurons. Because previous studies demonstrated that tyrosine phosphorylation of PMCA inhibits pump activity, I sought to biochemically investigate the tyrosine phosphorylation level of PMCA in KIF26A-knockdown neurons. A Kif26a miRNA vector could successfully knockdown KIF26A in F11 cells, compared with a scrambled miRNA vector as a negative control. Then, PMCA was immonoprecipitated from KIF26A-knockdown F11 cells and labeled with anti-phosphotyrosine antibody. As a result, the knockdown cells significantly elevated tyrosine phosphorylated levels of PMCA compared with the negative control suggesting that PMCA was inactivated by KIF26A deficiency. These data will explain the persistent [Ca(2+)]i elevation after withdrawal of capsaicin stimulation in Kif26a(-/-) DRG neurons.

Previous studies reported that PI3K and MAPK pathways could be activated by capsaicin in various types of cells including DRG neurons. Therefore, responsiveness of DRG neurons to the capsaicin stimulation could be tested by observing the phosphorylation level of Akt and Erk1/2. Accordingly, the activation of PI3K and MAPK signaling pathways were tested using an anti-phospho-Akt (pAkt) and an anti-phospho-Erk1/2 (pErk) antibodies respectively. The cultured DRG neurons were stimulated with capsaicin, with or without extensive washout of capsaicin, fixed and subjected to immunofluorescence staining against pAkt and pErk1/2. As a result, the fluorescence intensities of pAkt and pErk1/2 in Substance-P+ Kif26a(+/+) neurons significantly decreased after capsaicin washout. pAkt and pErk1/2 levels of Kif26a(-/-) neurons were already higher than that of Kif26a(+/+) neurons at the capsaicin activation phase. Furthermore, they were significantly failed in inactivation of these signaling pathways after capsaicin washout. These results demonstrated that PI3K and MAPK signaling pathways of Kif26a(-/-) DRG neurons were hypersensitive and persistently activated by capsaicin stimulation.

I have further investigated whether Kif26a(-/-) neurons were hypersensive in NGF/TrkA-mediated MAPK and PI3K stimulation, because NGF has been demonstrated as an important extracellular signaling molecule in enhancing the sensation of pain. Injection of NGF could induce thermal and mechanical hyperalgesia in adult rats and result in increased sensitivity to noxious heat and mechanical stimuli in humans. Meanwhile, NGF-TrkA signaling has been well characterized to be essential for the sensitization of TRPV1. The cultured DRG neurons were treated with NGF for and stained using antibodies that recognize pAkt and pErk. As a result, the fluorescent intensities of pAkt and pErk were found significantly stronger in Kif26a(-/-) neurons than those of Kif26a(+/+) neurons. These results suggested that a hyperactivated NGF signaling could also enhance the hyperalgesia of Kif26a(-/-) PANs.

In a previous study from our lab, KIF26A negatively regulated the GDNF-Ret-mediated Grb2-SHC interaction and suppressed the PI3K and MAPK signal transductions in enteric neurons. Because NGF-TrkA and GDNF-Ret signaling share the similar downstream signaling pathways mediated by the SHC-Grb2 complex, I sought to investigate whether KIF26A negatively regulates the MAPK and PI3K signaling via affecting the Grb2-SHC interaction also in the DRG neurons. Immunoprecipitation of a DRG cell line F11 showed that the level of SHC-Grb2 complex formation upon NGF stimulation was significantly increased by KIF26A deficiency. Furthermore, co-immunoprecipitation of endogenous KIF26A and endogenous Grb2 was verified using F11 cell lysates. These results suggested that negative regulation of Grb2 by KIF26A plays an important role in suppressing the MAPK and PI3K signaling also in DRG neurons, thereby properly controlling the noxious sensitivity of the PANs.

In conclusion, my study resolved a new function of kinesin superfamily protein 26A as a regulator of pain sensation and also established a relationship among KIF26A, NGF-TrkA signaling, and calcium regulation in the sensory nervous system through its general role in modulating Grb2 signaling. It provided evidence that the signaling modulatory role of this atypical kinesin can work in general and be responsible for wider range of disease and development than previously expected, which will be available for use in basic research of cell signaling and in development of therapeutic applications.

本研究は細胞内情報伝達において重要な機能をもつと考えられるKIF26Aタンパク質の末梢神経における役割を明らかにするため、KIF26A欠失マウスの行動解析、病理組織学的解析、ならびに一次培養した脊髄神経節ニューロンの生理学的及び細胞生物学的解析を一連のものとして行ったものであり、下記の結果を得ている。

1.Kif26a(+/+)マウスとKif26a(-/-)マウスのHot Plate TestならびにRandall Selitto Testによる行動解析を行なったところ、Kif26a(-/-)マウスでは侵害刺激に対する反応潜時が有意に低下し、通常より過敏な反応が認められた。このことから、Kif26a(-/-)マウスは痛覚過敏症の新たなモデルマウスとして利用できることが示された。

2.切片に対するin situ hybridization法によって、マウス脊髄神経節ニューロンに有意にKIF26Aが発現されていることが示された。また、この脊髄神経節全体の細胞数には大きな異常はなかった。

3.脊髄神経節ニューロンの切片において、NGFレセプターTrkAを発現している痛覚神経細胞を免疫組織化学法によって染色してみると、KIF26A欠失マウスではTrkAニューロンの細胞体が、有意に肥大化していることが認められた。

4.同様に免疫組織化学によって皮膚の抗PGP9.5染色、手掌のNF160染色を施行してみると、感覚神経の末梢部が増殖し、また異常な分枝が現われていることが明らかとなった。しかし、脊髄のCGRP, Ret, TrkA染色では、感覚神経の中枢側の形態には特に異常はみられなかった。

5.脊髄神経節ニューロンの刺激に対する反応性を調べるため、脊髄神経節ニューロンをカバースリップ上に一次培養し、これを新たに組み上げたコンフォーカル顕微鏡のステージ上の潅流システムにマウントし、痛覚刺激物質カプサイシンに対する細胞内カルシウムの反応を記録した。その結果、(1) KIF26A欠失ニューロンは刺激に対する反応性が有意に上昇しているとともに、(2)カプサイシン洗浄後も細胞内の高カルシウムレベルが持続していることが明らかとなった。このことは、KIF26Aの欠失によって脊髄神経節ニューロンの刺激反応性が昂進および異常持続することを示しておりcell autonomousな生理学的異常によって上記個体レベルの痛覚過敏を十分に説明できるものである。

6.脊髄神経節の感覚神経細胞のカプサイシン受容体TRPV1の刺激反応性昂進の分子機序として、過剰なリン酸化の存在が考えられる。そこで、脊髄神経節におけるTRPV1の発現量およびTRPV1のリン酸化の程度を生化学的に解析したところ、これらはいずれも有意に上昇していることが示された。

7.脊髄神経節の感覚神経細胞の細胞内カルシウム上昇の異常持続性の分子機序として、細胞膜上のカルシウムポンプPMCAの過剰なチロシンリン酸化の存在が考えられる。そこで、脊髄神経節細胞株F11細胞においてKIF26Aのノックダウン系を立ち上げ、その細胞上清からPMCAを免疫沈降し、抗リン酸化チロシン抗体でイムノブロッティングを行なってKIF26A欠失細胞におけるPMCAのチロシンリン酸化レベルを測定した。その結果、コントロール細胞に比してKIF26A欠失細胞では有意にPMCAのチロシンリン酸化レベルが上昇していた。すなわち、KIF26Aの欠失によりPMCAポンプが不活性化することによって細胞内カルシウムの除去機構が阻害されることが示唆された。

8.以上のようなタンパク質リン酸化の上流分子機構として、プライマリー培養脊髄神経節細胞のPI3K及びMAPKの活性を、抗pAkt抗体ならびに抗pErk抗体の免疫染色によって定量した。まずカプサイシンで野生型細胞を刺激すると、PI3K及びMAPK活性は刺激中一過性に上昇し、カプサイシン洗浄によって速やかにバックグラウンドレベルに復帰した。ノックアウト細胞のカプサイシン刺激中のPI3K/MAPK活性は、有意に野生型細胞のものより上昇していた。さらにカプサイシン洗浄後も、この高レベルのPI3K/MAPK活性の異常継続が示された。次にNGFでTrkA受容体を刺激した場合においても、ノックアウト細胞は野生型に比べて有意に高レベルのPI3K/MAPK反応を生じることが示された。このことは、受容体結合型ならびに受容体非結合型のチロシンキナーゼ下流のシグナル伝達がKIF26Aの欠失によって異常に昂進し、細胞内カルシウムの異常上昇によって痛覚過敏が生じていることを示唆するものである。

9.このPI3K/MAPK活性異常昂進の分子メカニズムとして、KIF26Aは腸管神経系の神経細胞のGDNF/Retシグナリングにおいて、Grb2とSHCの結合を競合阻害することでチロシンキナーゼ下流のシグナル伝達を負に制御していることが知られている。この機構の有効性を脊髄神経節細胞で調べるため、NGF刺激下におけるSHC-Grb2の免疫沈降を行ったところ、ノックダウン細胞では有意にこれらの結合能が上昇していることが示された。また、この細胞においてもKIF26AとGrb2は共沈することが示され、この負の制御機構は生体各所においてユニバーサルに働いていることが示唆された。

以上、本論文はKIF26Aノックアウトマウスにおいて、痛覚過敏のモデル系としてのその新しい役割を見出した。本研究は持続性疼痛に関与する脊髄神経節のチロシンキナーゼシグナル伝達における新しい制御因子を、キネシンモーター蛋白の細胞生物学という観点から同定したもので、細胞のシグナル伝達機構ならびに疼痛の病態生理の解明と治療に重要な貢献をなすと考えられ、学位の授与に値するものと考えられる。