Impaired slow axonal transport in diabetic peripheralnerve is independent of RAGE
Judyta K. Juranek,1,2 Matthew S. Geddis,1,3 Rosa Rosario1,2 and Ann Marie Schmidt1,21Department of Surgery, Columbia University Medical Center, New York, NY, USA2Diabetes Research Center, Department of Medicine, NYU Langone Medical Center, Smilow Building 906, 550 First Avenue,New York, NY 10016, USA3Department of Science, BMCC-City University of New York, New York, NY, USA
Diabetic peripheral nerve dysfunction is a common complication occurring in 30–50% of long-term diabetic patients. The patho-genesis of this dysfunction remains unclear but growing evidence suggests that it might be attributed, in part, to alteration in axo-nal transport. Our previous studies demonstrated that RAGE (Receptor for Advanced Glycation Endproducts) contributes to thepathogenesis of diabetic peripheral neuropathy and impairs nerve regeneration consequent to sciatic nerve crush, particularly indiabetes. We hypothesize that RAGE plays a role in axonal transport impairment via the interaction of its cytoplasmic domain withmammalian Diaphanous 1 (mDia1) – actin interacting molecule. Studies showed that mDia1–RAGE interaction is necessary forRAGE-ligand-dependent cellular migration, AKT phosphorylation, macrophage inflammatory response and smooth muscle migra-tion. Here, we studied RAGE, mDia1 and markers of axonal transport rates in the peripheral nerves of wild-type C57BL/6 andRAGE null control and streptozotocin-injected diabetic mice at 1, 3 and 6 h after sciatic nerve crush. The results show that in bothcontrol and diabetic nerves, the amount of RAGE accumulated at the proximal and distal side of the crush area is similar, indicat-ing that the recycling rate for RAGE is very high and that it is evenly transported from and towards the neuronal cell body. Fur-thermore, we show that slow axonal transport of proteins such as Neurofilament is affected by diabetes in a RAGE-independentmanner. Finally, our study demonstrates that mDia1 axonal transport is impaired in diabetes, suggesting that diabetes-relatedchanges affecting actin binding proteins occur early in the course of the disease.
Introduction
Impairment of peripheral nerve function is a common complicationof diabetes, occurring in 30–50% of all long-term diabetic patients(Deshpande et al., 2008; Smith & Singleton, 2012). Histopathologi-cal observations reveal that the most prevalent structural changesunderlying nerve dysfunction in diabetic peripheral nerve includeaxonal atrophy and degeneration (Medori et al., 1985). Althoughthe pathogenesis of these changes remains unclear, there is growingevidence that such impairment might be attributed to alterations inaxonal transport (Vitadello et al., 1985; Medori et al., 1988a;Larsen & Sidenius, 1989; McLean et al., 1992; McLean, 1997).Studies in streptozotocin-injected type 1 diabetic rodents show thatimpaired Neurofilament and actin transport often correlateswith observed histopathological and electrophysiological (nerveconduction velocity) changes in the diabetic nerve (Medori et al.,1985, 1988a; McLean et al., 1992), probably contributing tosymptomatic nerve dysfunction and diabetic neuropathy.Our study primarily focuses on axonal transport of RAGE
(Receptor for Advanced Glycation Endproducts) and its actin-bind-
ing cytoplasmic domain interacting partner (mDia1, mammaliandiaphanous 1) in diabetic peripheral nerves of streptozotocin-injectedwild-type C57BL/6 and RAGE null mice. Our previous studies dem-onstrated that RAGE contributes to the pathogenesis of diabeticperipheral neuropathy (Rong et al., 2004) and impairs nerve regen-eration consequent to sciatic nerve crush, particularly in diabetes(Juranek et al., 2013), by interacting with AGEs (Advanced Glyca-tion Endproducts) or other RAGE ligands and modulating adaptiveinflammatory responses at the nerve site. Furthermore, studies showthat glycation of cytoskeletal proteins such as Neurofilament or actinmight result in changes in axonal structure and function, leading tolong-term neurodegeneration (Sugimoto et al., 2008).Here, we hypothesize that RAGE plays a role in axonal transport
impairment via interaction with its cytoplasmic domain binding part-ner (Hudson et al., 2008; Rai et al., 2012), mDia1. mDia1 belongsto the family of Rho-GTPase formins (Mao, 2011) involved in actinstructure modification and modulation of microtubule dynamics.Studies showed that mDia1–RAGE interaction is necessary forRAGE-ligand-dependent cellular migration (Hudson et al., 2008),macrophage inflammatory responses (Xu et al., 2010), AKT phos-phorylation (Rai et al., 2012) and migration of smooth muscle cells(Toure et al., 2012). In the nervous system, mDia1 is thought toplay a role in neuronal migration and morphogenesis during devel-
Correspondence: Dr J. K. Juranek, 2Diabetes Research Center, as above.E-mail: [email protected]
Received 25 March 2013, revised 29 May 2013, accepted 11 July 2013
European Journal of Neuroscience, pp. 1–10, 2013 doi:10.1111/ejn.12333
European Journal of Neuroscience
opment (Arakawa et al., 2003; Shinohara et al., 2012). A recentreport on mDia1 and mDia3 double knockout revealed that mDia1deficiency in the brain leads to severe morphological and physiolog-ical changes in the neuroepithelial cell layer, resulting in productionof periventricular dysplastic masses and development of hydrocepha-lus (Thumkeo et al., 2011). Moreover, mDia1 is known to affectactin stress fiber formation (Watanabe & Higashida, 2004) and assuch it is crucial for proper functioning of the cellular cytoskeleton.The interaction of mDia1 with both actin and RAGE cytoplasmicdomain makes this protein of particular interest in studies of diabe-tes-related neuronal changes in the context of axonal transport.
Materials and methods
Animals
Eight-week-old male wild-type C57BL/6 (Jackson Laboratories, BarHarbor, ME, USA) and homozygous RAGE null mice (RKO), in theC57BL/6 background, generated as previously described (Sakaguchiet al., 2003; Origlia et al., 2008), were used in the study. BothC57BL/6 and RKO were weight-matched and divided into two groups:control and diabetic. All animals in the diabetic groups were given i.p.injections of streptozotocin (STZ) (5 days, 50 mg/kg in citrate buffer,100 mmol/L, pH 4.5; Sigma, St. Louis, MO, USA) and all controlmice were injected with an equal volume of vehicle (citrate buffer) asdescribed previously (Juranek et al., 2013). To confirm the diabeticstatus of the animals, blood glucose levels were measured 7 days afterthe first STZ injection and then every third week throughout theduration of the experiment and on the final day of the study. Animalswith a blood glucose level > 13 mmol/L were considered diabetic andused for further experiments. Animals were killed 2 months after thediabetic status was confirmed. All animal procedures were approvedby the Columbia University and New York University InstitutionalAnimal Care and Use Committees and were performed in accordancewith the National Institutes of Health Animal Care Guidelines.
Surgical procedures
Two months after induction of diabetes, mice were anesthetized viaan i.p. injection of ketamine/xylazine and the sciatic nerve wasexposed unilaterally and crushed as previously described (Wanget al., 2002; Juranek et al., 2013).
Immunofluorescence
To investigate the rate of anterograde (proximal site to crush) andretrograde axonal transport (distal site to crush) of studied proteins,sciatic nerves were collected at 1, 3 and 6 h after the crush (Li et al.,1996a; Wang et al., 2002), transferred immediately to 4% parafor-maldehyde in 0.1 M phosphate-buffered saline (PBS) and post-fixedfor 12 h at 4 °C. The following day, sections were rinsed in PBS,transferred to 20% sucrose and stored at �20 °C for further process-ing. Nerve samples were cut longitudinally at 10 lm thickness on acryostat (Microm HM 550; Thermo Scientific, Waltham, MA, USA)and collected on polylysine-coated slides (SuperFrost Plus; FisherScientific, Pittsburgh, PA, USA). Immunofluorescence staining wasperformed according to a standard laboratory protocol; sections wereincubated with the following antibodies – primary: goat anti-RAGE(1 : 100; Genetex, Irvine, CA, USA), rabbit anti-DIAPH1 (mDia1,1 : 100; Abcam, Cambridge, MA, USA), rabbit anti-Synaptophysin(SYP, 1 : 200; Synaptic System, Goettingen, Germany; fast axonaltransport), chicken anti-Neurofilament (NF, 1 : 100; Abcam; slow
axonal transport), rabbit anti-CML (carboxymethyllysine, representa-tive and prototypic AGE, 1 : 100; Abcam) and mouse anti-actin(1 : 100; Sigma); secondary: chicken anti-goat Alexa 568, chickenanti-rabbit and rabbit anti-chicken Alexa 488 and/or Alexa 564(Alexa 633 for triple staining) (1 : 300 and 1 : 200, respectively;Invitrogen, Grand Island, NY, USA). To control for the specificityof antibodies and to minimize the risk of false positive results, stan-dard immunostaining procedures with omission or replacement ofprimary antibodies on sections from each tissue sample set was car-ried out parallel to the experimental staining. All sections from eachtime-point series were immunostained in parallel as described(Li et al., 1996b) and the number of immunopositive axons at alength of 0.5–3 mm from the crush at both proximal and distal sideswas examined with a confocal microscope (Leica SP5, G€ottingen,Germany) with 20 9 and 63 9 objectives (Leica 20 9 Plan-Apo-chromat air and 63 9 Plan Apochromat oil objective) at one focalplane. Quantification of immunopositive fibers was performed usingImage J open source software (http://rsbweb.nih.gov/ij/); for quanti-tative study, proximal side values were used as a reference.
Statistical analysis
All values are presented as mean � standard error of the mean(SEM). The statistical significance of differences (P < 0.05) betweenall four studied groups (WT and RKO control and diabetic) was
A
B
Fig. 1. RAGE immunostaining in wild-type control (A) and diabetic (B) sci-atic nerve. A progressive increase of RAGE-immunopositive fibers, reflectingrates of anterograde (proximal side) and retrograde (distal side) axonal trans-port, is observed. A slight increase in RAGE immunostaining is observedin the diabetic nerve at all times points, 1, 3 and 6 h post-crush. Scalebar = 100 lm.
Fig. 2. Statistical analysis of RAGE immunostaining at the crush sides of control (A) and diabetic (B) sciatic nerve. As depicted on the diagrams, the percent-age of RAGE-positive fibers is similar on both sides of the crush (proximal side used as a reference, 100%), indicating that there are no noticeable losses ofRAGE on its way back from the axonal terminal. (C) RAGE accumulation on both sides of the crush increases steadily over time (all normalized to the 1 hWT control time point). Significant differences are observed between 1 and 6 h and slightly higher values of accumulation are noted for diabetic nerves. n = 5mice per group.
RAGE does not affect slow axonal transport in diabetes 3
evaluated by non-parametric ANOVA (Kruskal–Wallis test correctedfor ties) with Dunn multiple comparison post-test (GraphPad Instat,La Jolla, CA, USA).
Results
RAGE axonal transport in control and diabetic peripheralnerves
Sciatic nerve RAGE immunostaining was similar between control,non-diabetic and diabetic nerves, revealing similar patterns of pro-gressive accumulation at each site relative to crush for both nerveconditions over the studied time points (1–6 h; Fig. 1A and B).Detailed analysis of RAGE accumulation at the proximal and distalsides of the crush showed that the level of accumulation measuredas a number of RAGE-positive fibers was similar at both sides ofthe crush at each given time point in control and diabetic nerve(Fig. 2A and B, Table 1). Analysis of the immunostaining ratiobetween control and diabetic nerves over time [values for wild-type(WT) control at 1 h were used as a reference] revealed that therewas no significant difference between control and diabetic nerves atany studied time point. However, significant differences wereobserved between time points. The number of RAGE-positive fibersfor 1 h control nerves was significantly lower compared with 3 hdiabetic and 6 h control and diabetic nerves; values for 1 h diabeticnerve were significantly lower compared with 6 h control and dia-betic nerves.
Similarly, at the distal side there was no significant differencebetween control and diabetic nerves at any studied time point, butsignificant differences were observed between time points. The num-ber of RAGE-positive fibers at 1 h for control nerves was signifi-cantly lower compared with 6 h control and diabetic nerves.Similarly, 1 h diabetic nerve values were significantly lower com-pared with the 6 h diabetic condition (Fig. 2C).
Differences between fast (SYP) and slow (NF) axonaltransport markers in WT and RKO control and diabetic sciaticnerves
Detailed analysis of immunofluorescent accumulation of SYP, a fastaxonal transport marker, measured as the number of SYP-immuno-positive fibers at the proximal and distal side, respectively, revealedno significant difference between WT and RKO control and diabeticnerves at any time point (Fig. 3A). Similar analysis conducted forNF, a slow axonal transport marker, revealed statistically significantdifferences in accumulation within and between conditions andgenotypes (Fig. 3B, Table 1). At the proximal side, significant dif-ferences in the number of NF-positive fibers were observed 6 h afterthe crush between WT control and WT diabetic nerves. Statisticallysignificant differences were also observed between genotypes, i.e.between WT control and RKO diabetic nerves and WT diabetic andRKO control nerves. On the distal side, significant differences in thenumber of NF-positive fibers were first observed 3 h after the crushbetween WT control and WT and RKO diabetic nerves and were
Table 1. Summarized results of immunostaining quantification for all studied proteins
Time point
1 h 3 h 6 h
Groups WT control WT diabetic WT control WT diabetic WT control WT diabeticRAGEProximal No differences among groups No differences among groups No differences among groupsDistal No differences among groups No differences among groups No differences among groups
Groups WT control RKO control WT control RKO control WT control RKO controlWT diabetic RKO diabetic WT diabetic RKO diabetic WT diabetic RKO diabetic
SYPProximal No differences among groups No differences among groups No differences among groupsDistal No differences among groups No differences among groups No differences among groups
NFProximal No differences among groups No differences among groups Difference between:
WT control and WT/RKO diabeticRKO control and WT diabetic
Distal No differences among groups Difference between:WT control and WT/RKO diabetic
Difference between:WT control and WT/RKO diabeticRKO control and WT/RKO diabetic
mDia1Proximal No differences among groups No differences among groups Difference between:
WT control and WT/RKO diabeticRKO control and WT diabetic
Distal No differences among groups Difference between:WT/RKO control and RKO diabetic
Difference between:WT control and RKO diabeticRKO control and WT/RKO diabetic
CMLProximal Difference between:
RKO control and WT diabeticDifference between:RKO control and WT diabetic
Difference between:RKO control and WT diabeticWT diabetic and RKO diabetic
Distal Difference between:RKO control and WT diabetic
Difference between:RKO control and WT diabetic
Difference between:RKO control and WT diabeticWT diabetic and RKO diabetic
‘No differences’ signifies that no statistically significant differences were observed between and within groups. ‘Difference’ means that there was a statisticallysignificant difference (P < 0.05, non-parametric ANOVA with Dunn post hoc test) between given groups.
Fig. 3. Quantitative analysis of fast and slow axonal transport rates in the control and diabetic groups as measured by Synaptophysin (A) and Neurofilament(B) staining, respectively. No statistically significant changes were observed for Synaptophysin at any time point on either side of the nerve for all groups stud-ied (A). In contrast, Neurofilament staining revealed statistical differences in the accumulation rates on the proximal side between control and diabetic groups6 h after the crush and at the distal side between control and diabetic groups 3 and 6 h after the crush (B). The differences were dependent on diabetes but notdependent on RAGE genotype. n = 5 mice per group.
RAGE does not affect slow axonal transport in diabetes 5
more pronounced 6 h after the crush between WT control and dia-betic nerves and RKO control and diabetic nerves. Statistical differ-ences in the number of NF-positive fibers were also noted acrossgenotypes, between WT control and RKO diabetic and RKO controland WT diabetic nerves. However, no RAGE-dependent differencesin control or diabetic state were noted (Fig. 3B, Table 1).
mDia1 axonal transport impairment in WT and RKO diabeticnerves
At 1 h after the crush, no noticeable differences in immunostainingpatterns between conditions and genotypes were observed. At 3 h,
mDia1 immunostaining patterns for both WT and RKO remainedsimilar; clear staining along the fiber length was noticeable at bothsides of the crush. At the distal side, visible differences in the num-ber of stained fibers between control and diabetic groups wereobserved. Six hours after the crush, staining differences betweencontrol and diabetic groups were more pronounced; control samplesshowed higher staining intensity than their diabetic counterparts(Fig. 4A–C). Quantitative analysis of the mDia1-positive fiber num-ber at each studied time point revealed statistically significant differ-ences between control and diabetic groups 3 and 6 h after the crush.No statistically significant RAGE-dependent effects were observedat any time point (Fig. 5A and B, Table 1).
A
B
C
Fig. 4. mDia1 immunostaining in WT and RKO control and diabetic groups at the 1 h (A), 3 h (B) and 6 h (C) time points. (A) At 1 h, no noticeable differ-ences in immunostaining between groups are visible. (B) At 3 h, noticeable changes in mDia1 accumulation are visible at the distal side of both WT and RKOdiabetic nerves. (C) At 6 h, immunostaining differences in mDia1 accumulation are more pronounced, as compared with the 3 h time point, between controland diabetic groups. Scale bar = 100 lm.
CML axonal transport is lower in RKO mice and increases in adiabetic state
Immunostaining for CML, a prototypical AGE that is a specific ligandof RAGE, revealed noticeable differences in the number of immuno-positive axons between WT and RKO tissue samples (Fig. 6A).Quantification analysis revealed that on average, the number of
CML-positive neurons was significantly lower in RKO sciatic nervesat all studied time points and a trend towards a higher number ofaxons in both WT and RKO diabetic nerves compared with controlsamples (Fig. 6B, Table 1).
Actin-mDia1–RAGE colocalization – immunostaining
Double immunostaining of actin and mDia1 in WT control sciaticnerves, 6 h after the crush, revealed that both proteins colocalize in anumber of sciatic nerve fibers (Fig. 7A). Triple immunostaining foractin, mDia1 and RAGE revealed that a high number of sciatic nervefibers are positive both for RAGE and mDia1 and that all three pro-teins are present together in a smaller subset of nerve fibers (Fig. 7B).
Discussion
Growing evidence indicates that axonal transport is impaired in dia-betes, probably contributing to the development of neurologicalcomplications and subsequently diabetic neuropathy (McLean,
1997). However, many of the available reports are not precise indefining the axonal transport per se and do not provide conclusiveresults as to the specific type of transport, fast or slow, antero- orretrograde, and if they are affected in diabetic nerves.To gain insight into diabetes-related axonal transport impairment as
a likely contributor to the pathogenesis of diabetes-related neurologi-cal complications, here we studied expression of RAGE, a majorplayer in the pathogenesis of diabetic complications, and its actin-binding intracellular ligand mDia1, along with axonal transport mark-ers, in WT and RKO control and diabetic nerves. Our results revealedthat RAGE accumulation at the proximal and distal side of the crush issimilar in both control and diabetic nerves, indicating that the amountof RAGE delivered from perykaryon to the axonal ending is to a greatextent recycled and in similar quantity transported back from theaxonal ending to the neuronal cell body via retrograde transport, inde-pendently of the condition. Furthermore, we observed a progressiveincrease in the number of RAGE-positive fibers over time in bothcontrol and diabetic nerves at proximal and distal side of the crush,trending towards noticeably higher RAGE expression at the distal sidein the diabetic nerve. This observation is consistent with our previousstudies showing increased RAGE expression in a porcine intact andmouse long-term crushed diabetic nerve (Juranek et al., 2010). How-ever, due to the different time points investigated in the present study,changes in RAGE expression observed here were less pronounced.Studies of the fast (SYP) axonal transport markers revealed, as
reported previously (Dahlin et al., 1986; Abbate et al., 1991), that
A
B
Fig. 5. Quantitative analysis of mDia1 axonal transport in WT and RKO control and diabetic nerves. (A) On the proximal side, differences were noted 6 hafter the crush between control and diabetic groups and on the distal side (B) differences were observed 3 and 6 h after the crush. A trend toward lower accu-mulation numbers was observed for RKO diabetic nerve at all studied time points. n = 5 mice per group.
RAGE does not affect slow axonal transport in diabetes 7
rates of fast antero- and retrograde axonal transport were notaffected by diabetes. It might be speculated that the observed lackof effect on fast axonal transport in a diabetic nerve is due to mech-anistic differences underlying both types of transport as well as adifferent molecular structure of transported proteins. Fast axonaltransport might be simply too fast for hyperglycemia to mediate aneffect, and diabetes-related oxidative stress and protein glycation areboth lengthy processes to which rapidly transported proteins mightnot be susceptible. By contrast, slow axonal transport, studied hereby analysing NF expression at short intervals, reveals significant dif-ferences at the proximal side, 6 h after the crush and at the distalside, and 3 and 6 h after the crush between control and diabetes,indicating that hyperglycemia-related systemic changes affect slowaxonal transport even at the early stage of diabetes, by altering ratesof protein recycling (distal side accumulation) and accompanyingstructural changes to NF proteins (Pekiner & McLean, 1991; Miller
et al., 2002). No genotype-related differences were observed withinthe WT and RKO control or diabetic groups, indicating that RAGEdeficiency does not affect NF transport per se; rather it is secondaryto the diabetes phenomenon. Alterations of NF axonal transportwere previously reported in diabetic rats at the early stage of thedisease (Larsen & Sidenius, 1989; Macioce et al., 1989; Tomlinsonet al., 1990). Reports suggest that the observed retardation of NFtransport might contribute to the axonal degeneration and conse-quently to the development of neurological complications of diabe-tes (Medori et al., 1985, 1988b; Bomers et al., 1996; Fernyhough &Schmidt, 2002). Furthermore, studies in diabetic human and animalperipheral nerves showed a reduction in neurofilament mRNA levelthat, in addition, correlates with increased glycation and post-transla-tional modification of neurofilament protein, thus probably impairingaxonal transport and nerve regeneration in diabetic peripheral nerves(McLean, 1997).
A
B
Fig. 6. Qualitative and quantitative analysis of CML immunostaining in WT and RKO control and diabetic nerves. (A) Differences in immunostaining betweenWT and RKO tissue samples were clearly noticeable 6 h after the crush on both the proximal and the distal side. (B) Statistical differences were noted betweenRKO and WT nerves at all time points and a trend toward relatively higher CML immunostaining was noted for diabetic samples compared with their controlcounterparts. n = 5 mice per group.
Our study indicated the axonal transport of mDia1, RAGE-interacting and actin binding protein, is impaired in the diabeticnerve at 3 and 6 h after crush. The impairment is slightly more pro-nounced in the RKO nerve, indicating that hyperglycemia-relatedstructural changes in actin and actin binding proteins occur early inthe course of the disease and to some extent might be affected byRAGE deficiency. However, at this point in our study, it is difficultto discern whether these observed changes are related to or simplycoincide with RAGE deficiency. The observed trend towards a lowerpercentage of mDia1-positive fibers in the diabetic RKO nerve mightsuggest RAGE null genotype-related changes, although more detailedinvestigations are required to further support this notion. It is likelythat the observed trend reflects general, systemic and genotype-related neurological changes in diabetes, leading to alteration ofmDia1 axonal transport. It is plausible to hypothesize that theobserved impairment in mDia1 axonal transport is tightly correlatedwith the level of diabetes-evoked actin glycation (Pekiner et al.,1993) and, as such, reflects a general rather than individual, protein-specific trend in diabetic nerve axonal transport patterns. It has beenshown that at a very early stage of diabetes in rodents, 2 weeks afterSTZ injection (Pekiner et al., 1993), glycated actin was detected inbrain homogenates of diabetic animals and in blood platelets ofearly-stage diabetic patients. It is possible that changes in actin struc-ture affect the mDia1–actin interaction and thus lead to impairedtransport of both. Furthermore, studies on extracellular matrix pro-teins present in diabetic nerve endoneurium showed that early glyca-tion of these proteins affects neurite outgrowth and as such mightcontribute to the axonal regeneration impairment observed in diabeticneuropathy (Duran-Jimenez et al., 2009).Finally, we found that the level of CML, an AGE representative
protein, was on average lower in both control and diabetic RKOanimals compared with their WT counterparts. However, a trendtowards higher relative expression of CML was observed in bothWT and RKO diabetic nerves irrespective of their geneticbackground. These observations support our finding that the deletionof RAGE positively correlates with decreased AGE level and thus
might be beneficial in inflammatory states by reducing proteinglycation (Ramasamy et al., 2005). One potential explanation is ourobservation that in diabetic kidney of mice devoid of RAGE, levelsof the enzyme glyoxalase 1, which detoxifies AGE precursor meth-ylglyoxal, were higher than diabetic animals expressing RAGE, evenin the face of equivalent degrees of hyperglycemia (Reiniger et al.,2010). Hence, lower levels of CML in RAGE-null nerve mightreflect a greater degree of glyoxalase 1 activity. Furthermore, givenour own and others’ studies on cytoskeleton proteins and glycationin diabetic peripheral neuropathy, it might be speculated that thereis a certain threshold level of unbound AGE proteins that contrib-utes to increased cytoskeleton protein glycation irrespective of ani-mal genetic background and thus affecting, at least partially, axonaltransport of cytoskeleton proteins. These observations are in agree-ment with the observed lack of RAGE effect on NF and mDia1 axo-nal transport, indicating that the mechanisms governing axonaltransport alteration in diabetes are complex and not limited to onecause. It has been shown that the glycation of tubulin, one part ofthe microtubule-governed transport machinery, was highly increasedin the peripheral nerve of STZ-injected diabetic rats probably affect-ing the transport, although measurement of brain tubulin glycationlevel in the same animal did not show any increase indicating themicrotubule assembly inhibition in the brain was unrelated to theglycation level (Cullum et al., 1991). It is possible to hypothesizethat different pathomechanisms underlie the observed impairment ofaxonal transport in different parts of the nervous system in a dia-betic state, and that in the diabetic peripheral nerve, the observedimpairment reflects a combination of increased glycation and other,as yet undetermined factors.
Acknowledgements
We acknowledge funding for this research from the US Public HealthService (17P01AG017490-11). We thank Mrs Yan Lu for technical assis-tance with sample preparation and Ms Latoya Woods for excellent technicalassistance with formatting and figure preparation. The authors declare noconflict of interest.
A B
Fig. 7. Immunofluorescent analysis of actin-mDia1–RAGE interaction in WT control sciatic nerve, 6 h after the crush, on the proximal side. High-magnifica-tion images (one focal plane) of (A) actin-mDia1 and (B) actin-mDia1–RAGE staining. (A) Both actin and mDia1 are colocalized in a number of nerve fibers(arrows). Scale bar = 50 lm. (B) Triple immunostaining shows colocalization between actin, mDia1 and RAGE (arrows). Noticeable similarities in staining pat-terns between mDia1 and RAGE are observed. Scale bar = 20 lm.
Abbate, S.L., Atkinson, M.B. & Breuer, A.C. (1991) Amount and speed offast axonal transport in diabetes. Diabetes, 40, 111–117.
Arakawa, Y., Bito, H., Furuyashiki, T., Tsuji, T., Takemoto-Kimura, S.,Kimura, K., Nozaki, K., Hashimoto, N. & Narumiya, S. (2003) Control ofaxon elongation via an SDF-1a/Rho/mDia pathway in cultured cerebellargranule neurons. J. Cell Biol., 161, 381–391.
Bomers, K., Braendgaard, H., Flyvbjerg, A. & Sidenius, P. (1996) Redistri-bution of axoplasm in the motor root in experimental diabetes. Acta Neu-ropathol., 92, 98–101.
Dahlin, L.B., Meiri, K.F., McLean, W.G., Rydevik, B. & Sjostrand, J.(1986) Effects of nerve compression on fast axonal transport in streptozo-tocin-induced diabetes mellitus. An experimental study in the sciatic nerveof rats. Diabetologia, 29, 181–185.
Deshpande, A.D., Harris-Hayes, M. & Schootman, M. (2008) Epidemiologyof diabetes and diabetes-related complications. Phys. Ther., 88, 1254–1264.
Duran-Jimenez, B., Dobler, D., Moffatt, S., Rabbani, N., Streuli, C.H.,Thornalley, P.J., Tomlinson, D.R. & Gardiner, N.J. (2009) Advanced gly-cation end products in extracellular matrix proteins contribute to the failureof sensory nerve regeneration in diabetes. Diabetes, 58, 2893–2903.
Fernyhough, P. & Schmidt, R.E. (2002) Neurofilaments in diabetic neuropa-thy. Int. Rev. Neurobiol., 50, 115–144.
Hudson, B.I., Kalea, A.Z., Del Mar Arriero, M., Harja, E., Boulanger, E.,D’Agati, V. & Schmidt, A.M. (2008) Interaction of the RAGE cytoplas-mic domain with diaphanous-1 is required for ligand-stimulated cellularmigration through activation of Rac1 and Cdc42. J. Biol. Chem., 283,34457–34468.
Juranek, J.K., Aleshin, A., Rattigan, E.M., Johnson, L., Qu, W., Song, F.,Ananthakrishnan, R., Quadri, N., Yan, S.D., Ramasamy, R., Schmidt,A.M. & Geddis, M.S. (2010) Morphological changes and immunohisto-chemical expression of RAGE and its ligands in the sciatic nerve of hyper-glycemic pig (Sus scrofa). Biochem. Insights, 2010, 47–59.
Larsen, J.R. & Sidenius, P. (1989) Slow axonal transport of structural poly-peptides in rat, early changes in streptozocin diabetes, and effect of insulintreatment. J. Neurochem., 52, 390–401.
Li, J.Y., Jahn, R., Hou, X.E., Kling-Petersen, A. & Dahlstr€om, A. (1996a)Distribution of Rab3a in rat nervous system: comparison with other synap-tic vesicle proteins and neuropeptides. Brain Res., 706, 103–112.
Li, J.Y., Edelmann, L., Jahn, R. & Dahlstrom, A. (1996b) Axonal transportand distribution of synaptobrevin I and II in the rat peripheral nervoussystem. J. Neurosci., 16, 137–147.
Macioce, P., Filliatreau, G., Figliomeni, B., Hassig, R., Thiery, J. & DiGiamberardino, L. (1989) Slow axonal transport impairment ofcytoskeletal proteins in streptozocin-induced diabetic neuropathy. J. Neuro-chem., 53, 1261–1267.
Mao, Y. (2011) FORMIN a link between kinetochores and microtubule ends.Trends Cell Biol., 21, 625–629.
McLean, W.G. (1997) The role of axonal cytoskeleton in diabetic neuropa-thy. Neurochem. Res., 22, 951–956.
Medori, R., Autilio-Gambetti, L., Monaco, S. & Gambetti, P. (1985)Experimental diabetic neuropathy: impairment of slow transport withchanges in axon cross-sectional area. Proc. Natl. Acad. Sci. USA, 82,7716–7720.
Medori, R., Autilio-Gambetti, L., Jenich, H. & Gambetti, P. (1988a) Changesin axon size and slow axonal transport are related in experimental diabeticneuropathy. Neurology, 38, 597–601.
Medori, R., Jenich, H., Autilio-Gambetti, L. & Gambetti, P. (1988b) Experi-mental diabetic neuropathy: similar changes of slow axonal transport andaxonal size in different animal models. J. Neurosci., 8, 1814–1821.
Miller, C.C., Ackerley, S., Brownlees, J., Grierson, A.J., Jacobsen, N.J. &Thornhill, P. (2002) Axonal transport of neurofilaments in normal and dis-ease states. Cell. Mol. Life Sci., 59, 323–330.
Origlia, N., Righi, M., Capsoni, S., Cattaneo, A., Fang, F., Stern, D.M.,Chen, J.X., Schmidt, A.M., Arancio, O., Yan, S.D. & Domenici, L. (2008)Receptor for advanced glycation end product-dependent activation of p38mitogen-activated protein kinase contributes to amyloid-beta-mediated cor-tical synaptic dysfunction. J. Neurosci., 28, 3521–3530.
Pekiner, C. & McLean, W.G. (1991) Neurofilament protein phosphorylation inspinal cord of experimentally diabetic rats. J. Neurochem., 56, 1362–1367.
Pekiner, C., Cullum, N.A., Hughes, J.N., Hargreaves, A.J., Mahon, J., Cas-son, I.F. & McLean, W.G. (1993) Glycation of brain actin in experimentaldiabetes. J. Neurochem., 61, 436–442.
Rai, V., Maldonado, A.Y., Burz, D.S., Reverdatto, S., Yan, S.F., Schmidt,A.M. & Shekhtman, A. (2012) Signal transduction in receptor foradvanced glycation end products (RAGE): solution structure of C-terminalrage (ctRAGE) and its binding to mDia1. J. Biol. Chem., 287, 5133–5144.
Ramasamy, R., Vannucci, S.J., Yan, S.S., Herold, K., Yan, S.F. & Schmidt,A.M. (2005) Advanced glycation end products and RAGE: a commonthread in aging, diabetes, neurodegeneration, and inflammation. Glycobiol-ogy, 15, 16R–28R.
Reiniger, N., Lau, K., McCalla, D., Eby, B., Cheng, B., Lu, Y., Qu, W.,Quadri, N., Ananthakrishnan, R., Furmansky, M., Rosario, R., Song, F.,Rai, V., Weinberg, A., Friedman, R., Ramasamy, R., D’Agati, V. &Schmidt, A.M. (2010) Deletion of the receptor for advanced glycation endproducts reduces glomerulosclerosis and preserves renal function in thediabetic OVE26 mouse. Diabetes, 59, 2043–2054.
Rong, L.L., Yan, S.F., Wendt, T., Hans, D., Pachydaki, S., Bucciarelli,L.G., Adebayo, A., Qu, W., Lu, Y., Kostov, K., Lalla, E., Yan, S.D.,Gooch, C., Szabolcs, M., Trojaborg, W., Hays, A.P. & Schmidt, A.M.(2004) RAGE modulates peripheral nerve regeneration via recruitmentof both inflammatory and axonal outgrowth pathways. Faseb J., 18,1818–1825.
Sakaguchi, T., Yan, S.F., Yan, S.D., Belov, D., Rong, L.L., Sousa, M., And-rassy, M., Marso, S.P., Duda, S., Arnold, B., Liliensiek, B., Nawroth,P.P., Stern, D.M., Schmidt, A.M. & Naka, Y. (2003) Central role ofRAGE-dependent neointimal expansion in arterial restenosis. J. Clin.Invest., 111, 959–972.
Shinohara, R., Thumkeo, D., Kamijo, H., Kaneko, N., Sawamoto, K., Watan-abe, K., Takebayashi, H., Kiyonari, H., Ishizaki, T., Furuyashiki, T. &Narumiya, S. (2012) A role for mDia, a Rho-regulated actin nucleator,in tangential migration of interneuron precursors. Nat. Neurosci., 15,373–380, S1–S2.
Sugimoto, K., Yasujima, M. & Yagihashi, S. (2008) Role of advancedglycation end products in diabetic neuropathy. Curr. Pharm. Design, 14,953–961.
Thumkeo, D., Shinohara, R., Watanabe, K., Takebayashi, H., Toyoda, Y.,Tohyama, K., Ishizaki, T., Furuyashiki, T. & Narumiya, S. (2011) Defi-ciency of mDia, an actin nucleator, disrupts integrity of neuroepitheliumand causes periventricular dysplasia. PLoS ONE, 6, e25465.
Tomlinson, D.R., Filliatreau, G., Figliomeni, B., Hassig, R., Di Giamberardi-no, L. & Willars, G.B. (1990) Proteins of slow axonal transport in sciaticmotoneurones of rats with streptozotocin-induced diabetes or galactosa-emia. Diabetes Res. Clin. Pr., 9, 15–21.
Vitadello, M., Filliatreau, G., Dupont, J.L., Hassig, R., Gorio, A. & DiGiamberardino, L. (1985) Altered axonal transport of cytoskeletal proteinsin the mutant diabetic mouse. J. Neurochem., 45, 860–868.
Wang, Z.Y., Li, J.Y., Danscher, G. & Dahlstrom, A. (2002) Localization ofzinc-enriched neurons in the mouse peripheral sympathetic system. BrainRes., 928, 165–174.
Watanabe, N. & Higashida, C. (2004) Formins: processive cappers ofgrowing actin filaments. Exp. Cell Res., 301, 16–22.
Xu, Y., Toure, F., Qu, W., Lin, L., Song, F., Shen, X., Rosario, R., Garcia,J., Schmidt, A.M. & Yan, S.F. (2010) Advanced glycation end product(AGE)-receptor for AGE (RAGE) signaling and up-regulation of Egr-1 inhypoxic macrophages. J. Biol. Chem., 285, 23233–23240.
