ABSTRACT
Objective
To elucidate the features of plexin D1‐immunoglobulin (Ig)G‐associated neuropathic pain and its relationship to atopic myelitis (AM) in a nationwide Japanese survey.
Methods
A preliminary survey questionnaire was sent to 1574 selected departments (neurology and pediatrics/pediatric neurology) to explore the numbers of AM and plexin D1‐IgG‐positive patients between 2018 and 2022. A secondary survey collected detailed patient data via a questionnaire.
Results
In the preliminary survey, 987 (62.7%) institutions responded, reporting 87 AM patients (49 women) and 11 plexin D1‐IgG‐positive non‐AM patients (8 women). The secondary survey collected 71 AM (plexin D1‐IgG‐positive: 6/31) and 11 plexin D1‐IgG‐positive non‐AM patients (83.7% recovery rate). In AM, paresthesia/dysesthesia was most frequently experienced (> 90%), followed by pain (> 70%). The underlying diseases in 17 plexin D1‐IgG‐positive patients, all of whom had neuropathic pain, were AM and small fiber neuropathy in 6 each, neuromyelitis optica spectrum disorder with aquaporin‐4‐IgG in 2, and painful trigeminal neuropathy, erythromelalgia, and multiple sclerosis in 1 each. When 14 plexin D1‐IgG‐positive patients (excluding 3 patients with established demyelinating diseases) were compared with 25 plexin D1‐IgG‐negative AM patients, onset ≥ 50 years old, pain at onset, and allodynia/erythromelalgia/facial pain during the entire disease course were significantly more common in the plexin D1‐IgG‐positive group. Conversely, atopic disorders and hyperIgEemia were associated with plexin D1‐IgG‐negative AM but not plexin D1‐IgG‐positive patients.
Interpretation
Both AM and plexin D1‐IgG‐positive patients present long‐standing neuropathic pain, whereas plexin D1‐IgG is particularly associated with aged‐onset neuropathic pain, allodynia, erythromelalgia, and facial pain, but not atopy.
Keywords: atopic myelitis, autoantibody, nationwide survey, neuropathic pain, plexin D1atopic myelitis
1. Introduction
Atopic myelitis (AM) is a form of myelitis with unique characteristics that occurs in patients with long‐standing atopic disorders, such as atopic dermatitis, bronchial asthma, allergic rhinitis, allergic conjunctivitis, and food allergy [1, 2, 3]; the risk of AM is particularly elevated in those with atopic dermatitis [4]. The cardinal features of this condition encompass long‐lasting paresthesia and dysesthesia in addition to motor weakness in the extremities [1, 2, 3, 4, 5]. Although patients with AM most commonly present with persistent focal cervical spinal cord lesions that mimic spinal cord tumors [1, 2, 3, 4, 5], subclinical peripheral nerve involvement is also reported in approximately one‐quarter of cases [5, 6]. Biopsy and autopsy studies of cervical spinal cord lesions have revealed that both myelin and axons are lost; this is accompanied by the perivascular infiltration of inflammatory cells, including various degrees of eosinophils in acute lesions and the parenchymatous infiltration of activated microglia and astroglia in chronic lesions [7, 8, 9]. AM is rare; however, two Japanese nationwide surveys performed in 2001 [4] and 2006 [5] revealed that AM occurs throughout Japan. On the basis of these observations, diagnostic criteria for AM were proposed [10]. More recently, similar cases have been reported in not only Asian but also Western countries [11, 12, 13, 14].
In 2018, a population of AM patients was discovered to have immunoglobulin (Ig)G autoantibodies that recognize plexin D1 expressed on the membranes of small dorsal root ganglion (DRG) neurons conducting pain and on their nerve terminals in the dorsal horns [15]. Subsequently, plexin D1‐IgG was also identified in patients with small fiber neuropathy (SFN) and painful trigeminal neuropathy (PTN) presenting neuropathic pain in the limbs/body and face, respectively [16, 17, 18]. Intrathecally administered plexin D1‐IgG selectively activates pain‐conducting small DRG neurons and induces allodynia in mice [17]. Together, these findings have contributed to conceptualizing autoantibody‐mediated neuropathic pain [19]. However, neuropathic pain and allodynia were not studied in the previous nationwide AM [4, 5] surveys because they were conducted before the discovery of plexin D1‐IgG [15]. Furthermore, because the clinical manifestations of only a few plexin D1‐IgG‐positive patients have been reported to date [15, 16, 17, 18], the full spectrum and risk factors of plexin D1‐IgG‐associated disease remain ill‐defined. Although plexin D1‐IgG is detected in a fraction of AM patients [15], plexin D1‐IgG‐positive patients do not necessarily fulfill the AM diagnostic criteria [16, 17, 18]. Thus, the relationship between AM and plexin D1‐IgG‐associated diseases remains unknown. In the current study, we therefore aimed to characterize the clinical and laboratory features of AM and plexin D1‐IgG‐associated neurological disease, with special reference to pain and risk factors, using the first nationwide survey in Japan since the discovery of plexin D1‐IgG.
2. Methods
2.1. Survey Procedures
The survey was a cross‐sectional study that was undertaken in two steps (similar to the two previous surveys) [4, 5]. A preliminary survey was performed in June 2023 to ascertain the approximate number of patients with AM or plexin D1‐IgG. A secondary survey was then conducted between October 2023 and January 2024 using a questionnaire for each patient. All responses were provided by neurologists/pediatric neurologists at each facility and were based on the hospital discharge records of patients and the medical records of outpatients. The preliminary survey form was mailed to 892 facilities comprising educational facilities accredited by the Japanese Society of Neurology, neurology departments with two or more board‐certified neurologists, and neurology departments in hospitals with more than 500 beds. Additionally, 682 pediatric or pediatric neurology departments with at least one board‐certified pediatric neurologist were sent the preliminary survey form. AM was defined as myelitis of unknown cause with either (1) hyperIgEemia (> 240 U/mL) and antigen‐specific IgE positivity or (2) coexistent or past atopic diseases following the diagnostic criteria, excluding other diseases (Supporting Information) [4, 5, 10]. The existence of myelitis was confirmed by spinal cord MRI, motor‐evoked potentials (MEPs), somatosensory‐evoked potentials (SSEPs), or neurological findings of either pyramidal tract signs or sensory levels. Additionally, plexin D1‐IgG‐positive cases were also collated, irrespective of the fulfillment of the AM diagnostic criteria [4, 5, 10].
The questionnaire used in the preliminary survey requested the number and sex of patients with AM and those with plexin D1‐IgG between January 1, 2018, and December 31, 2022. The secondary questionnaire was forwarded to the facilities that reported patients in the preliminary survey. It requested detailed clinical information regarding individual patients, including their clinical history, symptomatology, neurological examination findings, retrospectively assessed disability score using the Kurtzke Expanded Disability Status Scale (EDSS) [20], and allergologic, electrophysiologic, and neuroimaging data (Table [Link], [Link]). Probable or definite neuropathic pain [21] as determined based on these data was statistically analyzed. Treatment efficacy was also evaluated by the neurologists/pediatric neurologists in attendance on the basis of subjective and objective clinical symptomatology (Table S1). During the study period, plexin D1‐IgG was measured using a tissue‐based assay (TBA) of mouse DRGs [15] in only one laboratory (Department of Neurology, Kyushu University) in Japan. The TBA (detailed methodology described in the Supporting Information) is regarded as the gold standard for the plexin D1‐IgG assay [15, 16, 17]. The research protocol for this retrospective study and the data privacy procedures for human samples were reviewed and approved by the Ethics Committees of the International University of Health and Welfare (23‐Ifh‐006).
2.2. Statistical Analysis
Differences in ratios among groups were tested for significance using Fisher's exact probability test or Pearson's chi‐squared test (two sided), whereas differences in numerical variables between two groups were examined using Wilcoxon's rank sum test. Wilcoxon's signed‐rank sum test was used to compare EDSS scores at the peak and the last examination. Spearman's rank correlation test was used to analyze the correlation between EDSS scores and disease duration. All analyses were performed using JMP 6.0.3 (SAS Institute, Cary, NC, USA). The significance level was set at p < 0.05.
3. Results
3.1. Response Rates and Numbers of Collated Patients
In the preliminary survey, we received responses from 62.7% of the facilities (987/1571; 66.7% [595/892] from neurology departments and 57.5% [392/682] from pediatric/pediatric neurology departments); 47 facilities had one or more patients with AM and/or plexin D1‐IgG‐positive cases (Table S2). There were 87 cases of AM (40 men and 47 women), which is a similar result to the 2006 survey, accounting for the difference in the study periods [5]. There were also 11 plexin D1‐IgG‐positive patients who did not fulfill the AM criteria (3 men and 8 women). In the secondary survey, detailed information was collated from 81.6% (71/87) of the AM patients and 100% (11/11) of the additional plexin D1‐IgG‐positive patients; the overall recovery rate was 83.7% (82/98). There were no significant differences in response rates according to the size and location of the hospitals (p > 0.05).
3.2. Clinical and Laboratory Features of All AM Patients
The demographic and clinical features of all AM patients collated in the secondary survey (Tables 1, S3 and S4) were similar to those of the previous studies, except for a longer disease duration in the present study [4, 5]. AM preferentially affected young adults of both sexes (Table 1). Approximately 70% of AM patients showed acute onset, including sudden and subacute onset, and nearly 30% exhibited chronic onset. The frequent initial symptoms were (in descending order) paresthesia/dysesthesia, motor weakness, neuropathic pain, and hypesthesia (Table 1). During illness, paresthesia/dysesthesia was most frequently experienced (> 90%) followed by hypesthesia, motor weakness, and neuropathic pain (> 70%). Pyramidal tract signs were also frequently observed (approximately 70%). Allodynia, sphincter disturbance, chronic fatigue, and hyporeflexia were occasionally observed, whereas muscle atrophy, erythromelalgia, visual impairment, and Lhermitte's sign were rare (Table 1). A relapsing remitting/fluctuating course was the most common (> 60%), followed by chronic progressive and monophasic courses. AM patients had moderate disability at the peak of illness (median EDSS = 3.75; nearly half of AM patients ≥ 4.5), and EDSS scores at the last examination were significantly improved from peak EDSS scores (median EDSS = 3.0, p < 0.0001; one‐third of AM patients ≥ 4.5) (Table 1). EDSS scores were significantly positively correlated with disease duration (ρ = 0.3965, p = 0.0009) (Figure 1A).
TABLE 1.
Clinical features of patients with AM.
| Survey items | AM (n = 71) |
|---|---|
| Age at onset, y, median (IQR) | 31 (23, 45) |
| Mean ± SD | 33.4 ± 13.7 |
| Onset age ≥ 50 y (%) | 12/71 (16.9%) |
| Disease duration, y, median (IQR) | 8 (3, 15) |
| Mean ± SD | 11.0 ± 11.2 |
| Men: women (% women) | 28: 43 (60.6) |
| Mode of onset (%) | |
| Sudden | 5/71 (7.0) |
| Acute/subacute | 45/71 (63.3) |
| Chronic | 21/71 (29.6) |
| Onset symptom (%) | |
| Paresthesia/dysesthesia | 36/70 (51.4) |
| Neuropathic pain | 21/70 (30.0) |
| Hypesthesia | 13/70 (18.6) |
| Motor weakness | 25/70 (35.7) |
| Gait disturbance | 9/70 (12.9) |
| Others | 5/70 (7.1) a |
| Symptoms and signs during illness (%) | |
| Paresthesia/dysesthesia | 65/71 (91.5) |
| Neuropathic pain | 49/70 (70.0) |
| Allodynia | 12/69 (17.4) |
| Erythromelalgia | 3/71 (4.2) |
| Hypesthesia | 58/70 (82.9) |
| Motor weakness | 53/71 (74.6) |
| Muscle atrophy | 7/71 (9.9) |
| Pyramidal tract signs | 49/71 (69.0) |
| Hyporeflexia | 11/70 (15.7) |
| Sphincter disturbance | 21/71 (29.6) |
| Chronic fatigue | 17/69 (24.6) |
| Lhermitte sign | 4/69 (5.8) |
| Visual impairment | 5/71 (7.0) |
| Clinical course | |
| Monophasic | 12/70 (17.1) |
| Relapsing remitting/fluctuating | 44/70 (62.9) |
| Chronic progressive | 14/70 (20.0) |
|
Peak EDSS score, median (IQR) b Mean ± SD |
3.75 (3.5, 6.0) 4.5 ± 2.1 |
| Peak EDSS ≥ 4.5 b | 32/68 (47.1) |
| Final EDSS score, median (IQR) b | 3.0 (2.0, 5.5) |
| Mean ± SD | 3.6 ± 2.2 |
| Final EDSS ≥ 4.5 | 25/67 (37.3) |
Abbreviations: AM, atopic myelitis; EDSS, Kurtzke Expanded Disability Status Scale; Ig, immunoglobulin; IQR, interquartile range; y, years.
Others included two cases of micturition disturbance and one case each of dystonia, visual impairment, and respiratory disturbance.
Wilcoxon signed‐rank sum test was used for the comparison between EDSS scores at the peak and at the last examination. EDSS scores were significantly higher at the peak than at the last examination in the present survey (p < 0.0001).
FIGURE 1.
Correlations between EDSS scores and disease duration. (A) In all AM patients, EDSS scores exhibited a significant positive correlation with disease duration (ρ = 0.3965, p = 0.0009). (B) In PANPS patients, EDSS scores showed a strong positive correlation with disease duration (ρ = 0.7584, p = 0.0042). (C) In plexin D1‐IgG‐negative AM patients, no significant correlation between EDSS scores and disease duration was identified. p values were obtained using Spearman's rank correlation test. AM, atopic myelitis; EDSS, Kurtzke Expanded Disability Status Scale; PANPS, plexin D1‐associated neuropathic pain syndrome.
Nearly half of AM patients had spinal cord MRI lesions; these preferentially affected the cervical spinal cord with occasional contrast enhancement. By contrast, brain MRI lesions were rare (Table S3). MEPs frequently showed abnormalities (> 60%), particularly central conduction abnormalities, followed by abnormalities in SSEPs, mostly presenting as central conduction abnormalities (Table S3). Peripheral conduction abnormalities in MEPs and SSEPs were rare. Occasionally, AM patients showed visual evoked potential (VEP) abnormalities (approximately 30%), although apparent visual impairment was rare (7.0%) (Table S3). Abnormalities in nerve conduction studies and needle EMG were infrequent. The current perception threshold test (CPT) frequently exhibited abnormalities affecting all fiber sizes (total > 70%; Aβ, Aδ, and C fiber abnormalities, 55%–60% each). CPTs showed increased thresholds more often than decreased thresholds in all fibers. CSF pleocytosis and protein increases were infrequent, and oligoclonal IgG bands and increased IgG indexes were rarely seen (Table S3).
Regarding atopic diathesis, most AM patients had preceding and/or concomitant atopic disorders (> 90%); these were most commonly atopic dermatitis, followed by allergic rhinitis, bronchial asthma, food allergy, and allergic conjunctivitis (Table S3). Most patients had specific IgE to mite antigens (> 80%) such as Dermatophagoides pteronyssinus and D. farinae, followed by IgE to cedar pollen, dog and/or cat dander, and food allergens. The median IgE level (683 U/mL) was higher than the normal range (normal < 240 U/mL), and nearly two‐thirds of patients had hyperIgEemia (Table S3). By contrast, the median eosinophil count was within the normal range (241/μL), and eosinophilia was infrequent.
Concerning treatment, intravenous methylprednisolone (IVMP) pulse therapy was most frequently used (87.1%) and was generally effective (86.4%), although it was less effective for pain (71.1%) (Table S4). Plasmapheresis was applied in approximately half of AM cases and exhibited the best general and pain effectiveness (≥ 90%). Oral corticosteroids, immunosuppressants, and intravenous (IV)Ig were less frequently used and showed similar effects to oral corticosteroids (70%–90%). Half of AM patients received anti‐allergic drugs, which were generally effective (80%) but were not effective for pain (35.7%) (Table S4). Pregabalin/mirogabalin, other anti‐epileptic drugs, and anti‐depressants were occasionally used and showed some beneficial effects on pain (Table S4).
3.3. Plexin D1‐IgG Assay Results
In addition to the 11 plexin D1‐IgG patients who did not fulfill the AM criteria, plexin D1‐IgG was examined in 31 of 71 (43.7%) AM patients collated in the secondary survey; 6/31 (19.4%) AM patients were positive for plexin D1‐IgG. Representative results are shown (Figure S1). The underlying diseases of the 17 plexin D1‐IgG‐positive patients were AM and SFN in 6 patients each (35.3%), neuromyelitis optica spectrum disorder (NMOSD) with aquaporin 4‐IgG in 2 patients (11.8%), and PTN, erythromelalgia, and multiple sclerosis (MS) in 1 patient each (5.9%) (Table S5). To characterize the clinical features and risk factors of plexin D1‐IgG‐positive patients, 14 plexin D1‐IgG‐positive patients (excluding 3 patients with established demyelinating diseases; i.e., NMOSD or MS) were tentatively combined into a plexin D1‐associated neuropathic pain syndrome (PANPS) group and were compared with the 25 plexin D1‐IgG‐negative AM patients or the 65 plexin D1‐IgG‐negative/unknown AM patients (Figure S2).
3.4. Comparison of Demographic and Clinical Features Between Patients With PANPS and Those With Plexin D1‐IgG‐Negative AM
Patients with PANPS had an older median age at onset than those with plexin D1‐IgG‐negative AM (nearly 20 years older) and a higher frequency of onset age ≥ 50 years old (57.1% vs. 12.0%, p = 0.0027) (Table 2). The disease duration, proportion of women, and mode of onset did not significantly differ between the two conditions. Of the onset symptoms, neuropathic pain was more common in PANPS than in plexin D1‐IgG‐negative AM patients (p = 0.0005), whereas motor weakness was more frequent in plexin D1‐IgG‐negative AM than in PANPS patients (p = 0.0050) (Table 2). Over the entire clinical course, paresthesia/dysesthesia and neuropathic pain were most commonly observed in both conditions (both > 90%) (Table 2). However, the distribution of pain sites differed significantly between the two conditions (p = 0.0168); overall, facial pain was much more frequently encountered for PANPS than for plexin D1‐IgG‐negative AM (50.0% vs. 4.6%, p = 0.0026) (Table 3). The distribution of paresthesia/dysesthesia sites showed a similar trend but did not reach significance. Allodynia and erythromelalgia were more frequently encountered in PANPS than in plexin D1‐IgG‐negative AM patients (p = 0.0207 and p = 0.0469, respectively), whereas hypesthesia was frequently observed in both conditions (Table 2). By contrast, motor weakness and pyramidal tract signs were more commonly noted for plexin D1‐IgG‐negative AM than for PANPS (p = 0.0004 and p = 0.0024, respectively). The frequencies of other symptoms/signs, such as muscle atrophy, hyporeflexia, sphincter disturbance, chronic fatigue, Lhermitte's sign, and visual impairment, did not differ significantly between the two conditions. Regarding the clinical course, PANPS patients showed a chronic progressive course more frequently and a relapsing remitting/fluctuating course less frequently than did plexin D1‐IgG‐negative AM patients (p = 0.0090). The peak and final EDSS scores were significantly lower in PANPS than in plexin D1‐IgG‐negative AM patients (p = 0.0191 and p = 0.0378, respectively) (Table 2). Although EDSS scores were higher at the peak than at the last examination for plexin D1‐IgG‐negative AM (p = 0.0006), no significant difference in EDSS scores was identified between the peak and last examination in PANPS patients. In PANPS, EDSS scores exhibited a strong positive correlation with disease duration (ρ = 0.7584, p = 0.0042), whereas no significant correlation was identified between EDSS scores and disease duration for plexin D1‐IgG‐negative AM (Figure 1B,C).
TABLE 2.
Comparison of clinical features between patients with PANPS and plexin D1‐IgG‐negative AM.
| PANPS (n = 14) | Plexin D1‐IgG‐negative AM (n = 25) | p | |
|---|---|---|---|
|
Age at onset, y, median (IQR) Mean ± SD |
51.5 (14, 65) 41.7 ± 23.1 |
31 (25.5, 40.5) 32.7 ± 11.0 |
NS a |
| Onset age ≥ 50 y (%) | 8/14 (57.1) | 3/25 (12.0) | 0.0027 b |
| Disease duration, y, median (IQR) | 16.4 (4.4, 26.0) | 13.4 (5.5, 22.2) | NS a |
| Mean ± SD | 18.6 ± 16.7 | 14.1 ± 9.3 | |
| Men: women (% women) | 4:10 (71.4) | 3: 22 (88.0) | NS b |
| Mode of onset (%) | NS c | ||
| Sudden | 0/14 (0) | 0/25 (0) | |
| Acute/subacute | 8/14 (57.1) | 15/25 (60.0) | |
| Chronic | 6/14 (42.9) | 10/25 (40.0) | |
| Onset symptom (%) | |||
| Paresthesia/dysesthesia | 6/14 (42.9) | 15/24 (62.5) | NS b |
| Neuropathic pain | 12/14 (85.7) | 6/24 (25.0) | 0.0005 b |
| Hypesthesia | 1/14 (7.1) | 3/24 (12.5) | NS b |
| Motor weakness | 1/14 (7.1) | 13/24 (54.2) | 0.0050 b |
| Gait disturbance | 0/14 (0) | 1/24 (4.2) | NS b |
| Others | 0/14 (0) | 0/25 (0) | NS b |
| Symptoms and signs during illness (%) | |||
| Paresthesia/dysesthesia | 13/14 (92.9) | 24/25 (96.0) | NS b |
| Neuropathic pain | 14/14 (100) | 23/25 (92.0) | NS b |
| Allodynia | 7/14 (50.0) | 3/24 (12.5) | 0.0207 b |
| Erythromelalgia | 4/14 (28.6) | 1/25 (4.0) | 0.0469 b |
| Hypesthesia | 11/14 (78.6) | 23/25 (92.0) | NS b |
| Motor weakness | 6/14 (42.9) | 24/25 (96.0) | 0.0004 b |
| Muscle atrophy | 0/14 (0) | 2/25 (8.0) | NS b |
| Pyramidal tract signs | 4/14 (28.6) | 20/25 (80.0) | 0.0024 b |
| Hyporeflexia | 1/14 (7.1) | 1/25 (4.0) | NS b |
| Sphincter disturbance | 3/14 (21.4) | 10/25 (40.0) | NS b |
| Chronic fatigue | 3/14 (21.4) | 6/23 (26.1) | NS b |
| Lhermitte sign | 0/14 (0) | 1/25 (4.0) | NS b |
| Visual impairment | 1/14 (7.1) | 4/25 (16.0) | NS b |
| Clinical course (%) | 0.0090 c | ||
| Monophasic | 0/14 (0) | 2/25 (8.0) | |
| Relapsing remitting/fluctuating | 7/14 (50) | 21/25 (84.0) | |
| Chronic progressive | 7/14 (50) | 2/25 (8.0) | |
| Peak EDSS score, median (IQR) d | 3.25 (2.0, 4.5) d | 4.0 (3.5, 6.0) | 0.0191 a |
| Mean ± SD | 3.2 ± 1.5 | 4.7 ± 1.6 | |
| Peak EDSS ≥ 4.5 d | 4/12 (33.3) | 12/25 (48.0) | NS b |
| Final EDSS score, median (IQR) d | 2.75 (2.0, 4.25) | 3.5 (3.0, 6.0) | 0.0378 a |
| Mean ± SD | 3.0 ± 1.5 | 4.3 ± 1.8 | |
| Final EDSS ≥ 4.5 | 3/12 (25.0) | 12/25 (48.0) | NS b |
Abbreviations: AM, atopic myelitis; EDSS, Kurtzke Expanded Disability Status Scale; IQR, interquartile range; NS, not significant; PANPS, plexin D1‐IgG‐associated neuropathic pain syndrome; y, years.
Wilcoxon test (two‐sided test).
Fisher's exact test (two‐sided test).
Pearson's chi‐squared test (two‐sided test).
Wilcoxon signed‐rank sum test was used for the comparison between EDSS scores at the peak and last examination. EDSS scores were higher at the peak than at the last examination in plexin D1‐IgG‐negative AM (p = 0.0006) but not in PANPS (p > 0.1).
TABLE 3.
Comparison of the distribution of neuropathic pain and paresthesia/dysesthesia between patients with PANPS and plexin D1‐IgG‐negative AM.
| Body part (%) | Neuropathic pain | Paresthesia/dysesthesia | ||||
|---|---|---|---|---|---|---|
| PANPS | Plexin D1‐IgG‐negative AM | p a | PANPS | Plexin D1‐IgG‐negative AM | p a | |
| Limb alone |
6 (42.9) |
17 (77.3) |
0.0168 b |
7 (53.9) |
16 (69.6) |
0.0738 b |
| Limb & trunk |
1 (7.1) |
3 (13.6) |
1 (7.7) |
2 (8.7) |
||
| Trunk alone |
0 (0) |
1 (4.6) |
0 (0) |
1 (4.4) |
||
| Limb & trunk & face |
1 (7.1) |
1 (4.6) |
1 (7.7) |
4 (17.4) |
||
| Limb & face |
6 (42.9) |
0 (0) |
4 (30.8) |
0 (0) |
||
| Face alone |
0 (0) |
0 (0) |
0 (0) |
0 (0) |
||
| Total | 14 | 22 | 13 | 23 | ||
| All facial involvement (%) |
7/14 (50.0) |
1/22 (4.6) | 0.0026 c |
5/13 (38.5) |
4/23 (17.4) | NS c |
Abbreviations: AM, atopic myelitis; NS, not significant; PANPS, plexin D1‐IgG‐associated neuropathic pain syndrome.
Comparison between PANPS and plexin D1‐AM.
Pearson's chi‐squared test (two‐sided test).
Fisher's exact test (two‐sided test).
When PANPS was compared with plexin D1‐IgG‐negative/unknown AM, the results were essentially the same except that the longer disease duration and higher frequencies of neuropathic pain and facial paresthesia/dysesthesia in PANPS than in plexin D1‐IgG‐negative/unknown AM became significant (p = 0.0270, p = 0.0161, and p = 0.0081, respectively) (Tables S6 and S7).
3.5. Comparison of Laboratory Findings Between Patients With PANPS and Plexin D1‐IgG‐Negative AM
There were no significant differences in the frequencies of spinal cord and brain MRI lesions, nerve conduction study abnormalities, needle EMG abnormalities, and evoked potential abnormalities, including MEPs, SSEPs, and VEPs (Table 4). In the CPT, although both PANPS and plexin D1‐IgG‐negative AM patients frequently showed abnormalities in all fibers, decreased thresholds were not identified in any plexin D1‐IgG‐negative AM patients in any fibers (including Aβ, Aδ, and C fibers). By contrast, PANPS patients frequently showed decreased thresholds in all of these fibers (25%–50%). As a result, C fiber thresholds tended to decrease more often in PANPS patients than in plexin D1‐IgG‐negative AM patients (37.5% vs. 0%, p = 0.0686) (Table 4). CSF tests were mostly normal in both conditions, including no oligoclonal IgG band positivity, except for some protein increases in a minority of patients (Table 4). Comparisons between PANPS and plexin D1‐IgG‐negative/unknown AM revealed an increased frequency of decreased C fiber thresholds (37.5% vs. 0%, p = 0.0421) and increased peripheral abnormalities on SSEPs in PANPS compared with those in plexin D1‐IgG‐negative/unknown AM (p = 0.0860) (Table S8). No other significant differences between the two conditions were noted in the examined laboratory tests.
TABLE 4.
Comparison of laboratory findings between patients with PANPS and plexin D1‐IgG‐negative AM.
| PANPS (n = 14) | Plexin D1‐IgG‐negative AM (n = 25) | p value a | |
|---|---|---|---|
| Spinal cord MRI lesion (%) | 4/13 (30.8) | 8/25 (32.0) | NS |
| Cervical spinal cord lesion | 1/2 (50.0) | 3/8 (37.5) | NS |
| Gadolinium‐enhancement | 1/4 (25.0) | 0/12 (0) | NS |
| Brain MRI lesion (%) | 2/12 (16.7) | 3/25 (12.0) | NS |
| Brainstem lesion | 0/12 (0) | 1/25 (4.0) | NS |
| Gadolinium enhancement | 0/9 (0) | 0/19 (0) | NS |
| NCS abnormalities (%) | 1/10 (10.0) | 1/20 (5.0) | NS |
| Needle EMG abnormalities (%) | 0/3 (0) | 0/3 (0) | NS |
| Evoked potentials (%) | |||
| MEP, total abnormality | 4/8 (50.0) | 15/22 (68.2) | NS |
| MEP, central abnormalities | 4/8 (50.0) | 14/22 (63.6) | NS |
| MEP, peripheral abnormalities | 0/8 (0) | 1/22 (4.6) | NS |
| SSEP, total abnormality | 3/9 (33.3) | 6/19 (31.6) | NS |
| SSEP, central abnormality | 1/9 (11.1) | 6/19 (31.6) | NS |
| SSEP, peripheral abnormality | 2/9 (22.2) | 1/19 (5.3) | NS |
| VEP abnormality | 1/2 (50.0) | 5/12 (41.7) | NS |
| CPT (%) | |||
| All fibers, total abnormality | 6/8 (75.0) | 6/10 (60.0) | NS |
| All fibers, increased threshold | 5/8 (62.5) | 6/10 (60.0) | NS |
| All fibers, decreased threshold | 3/8 (37.5) | 0/10 (0) | 0.0686 |
| Aβ fibers, total abnormality | 4/8 (50.0) | 4/10 (40.0) | NS |
| Aβ fibers, increased threshold | 4/8 (50.0) | 4/10 (40.0) | NS |
| Aβ fibers, decreased threshold | 1/8 (12.5) | 0/10 (0) | NS |
| Aδ fibers, total abnormality | 3/8 (37.5) | 5/10 (50.0) | NS |
| Aδ fibers, increased threshold | 3/8 (37.5) | 5/10 (50.0) | NS |
| Aδ fibers, decreased threshold | 2/8 (25.0) | 0/10 (0) | NS |
| C fibers, total abnormality | 6/8 (75.0) | 5/10 (50.0) | NS |
| C fibers, increased threshold | 4/8 (50.0) | 5/10 (50.0) | NS |
| C fibers, decreased threshold | 3/8 (37.5) | 0/10 (0) | 0.0686 |
| CSF (%) | |||
| Pleocytosis | 0/7 (0) | 0/23 (0) | NS |
| Protein increase | 1/7 (14.3) | 4/23 (17.4) | NS |
| OCB | 0/5 (0) | 0/21 (0) | NS |
| Increased IgG index | 0/5 (0) | 0/21 (0) | NS |
Abbreviations: AM, atopic myelitis; CPT, current perception threshold test; Ig, immunoglobulin; MEP, motor evoked potential; NCS, nerve conduction study; NS, not significant; OCB, oligoclonal IgG bands; PANPS, plexin D1‐IgG‐associated neuropathic pain syndrome; SSEP, somatosensory evoked potential; VEP, visual evoked potential.
All p values were obtained using Fisher's exact test (two‐sided test).
3.6. Comparison of Atopic Diathesis Between Patients With PANPS and Plexin D1‐IgG‐Negative AM
High frequencies of atopic/allergic diseases, as observed for plexin D1‐IgG‐negative AM, were not noted for PANPS. The frequencies of any preceding/concomitant atopic/allergic diseases were much lower for PANPS than for plexin D1‐IgG‐negative AM (p < 0.0001), particularly for atopic dermatitis and allergic rhinitis (p = 0.0007 and p = 0.0031, respectively) (Table 5). Total serum IgE levels and hyperIgEemia frequency were lower in PANPS than in plexin D1‐IgG‐negative AM patients (p = 0.0315 and p = 0.0590, respectively). However, blood eosinophil counts, the frequency of eosinophilia, and the frequencies of specific IgE to D. pteronyssinus and D. farinae, cedar pollen, food allergens, and dog and/or cat dander did not significantly differ between the two conditions (Table 5). Comparisons between PANPS and plexin D1‐IgG‐negative/unknown AM showed similar trends (Table S9).
TABLE 5.
Comparison of atopic diathesis between patients with PANPS and plexin D1‐IgG‐negative AM.
| PANPS (n = 14) | Plexin D1‐IgG‐negative AM (n = 25) | p | |
|---|---|---|---|
| Any preceding/concomitant atopic/allergic diseases (%) | 5/13 (38.5) | 24/24 (100) | < 0.0001 a |
| Atopic dermatitis | 1/13 (7.7) | 13/18 (72.2) | 0.0007 a |
| Allergic rhinitis | 3/13 (23.1) | 16/20 (80.0) | 0.0031 a |
| Bronchial asthma | 4/12 (33.3) | 10/20 (50.0) | NS a |
| Food allergy | 1/12 (8.3) | 3/15 (20.0) | NS a |
| Allergic conjunctivitis | 1/13 (7.7) | 6/17 (35.3) | NS a |
|
Total serum IgE, U/mL, median (IQR) Mean ± SD |
90.2 (42.3, 307.5) 180.2 ± 234.5 |
523.0 (168.5, 2103.5) 2132.2 ± 3649.0 |
0.0315 b |
| HyperIgEemia (%) | 1/6 (16.7) | 17/25 (68.0) | 0.0590 a |
| Specific IgE to (%) | |||
| D. pteronyssinus and/or D. farinae | 5/6 (83.3) | 20/24 (83.3) | NS a |
| Cedar pollen | 5/5 (100) | 18/24 (75.0) | NS a |
| Any food allergen | 0/4 (0) | 3/17 (17.7) | NS a |
| Dog and/or cat dander | 2/5 (40.0) | 5/13 (38.5) | NS a |
| Blood eosinophils, /μL, median (IQR) |
146.7 (131.9, 223.3) |
190.6 (79.0, 378.6) |
NS b |
| Mean ± SD | 175.9 ± 70.6 | 274.5 ± 272.3 | |
| Eosinophilia (≥ 500/μL) | 0/12 (0) | 4/24 (16.7) | NS a |
| Other autoantibodies | 2/12 (16.7) c | 1/22 (4.6) c | NS a |
Abbreviations: AM, atopic myelitis; D., Dermatophagoides; Ig, immunoglobulin; IQR, interquartile range; NS, not significant; PANPS, plexin D1‐IgG‐associated neuropathic pain syndrome.
Fisher's exact test (two‐sided test).
Wilcoxon signed‐rank sum test.
Other autoantibodies detected included anti‐nuclear antibodies in two patients with PANPS and in one patient with plexin D1‐IgG‐negative AM.
3.7. Comparison of Treatment Modalities and Responses Between Patients With PANPS and Plexin D1‐IgG‐Negative AM
Immunotherapies, including oral corticosteroids (p = 0.0244), IVMP pulse therapy (p = 0.0086), immunosuppressants (p = 0.0089), plasmapheresis (p = 0.0604), and IVIg (p = 0.0201), were more frequently administered for plexin D1‐IgG‐negative AM than for PANPS (Table 6). However, when they were administered, oral corticosteroids, IVMP pulse therapy, and plasmapheresis were similarly effective for both conditions (Table 6). Anti‐allergic drugs were more frequently used for plexin D1‐IgG‐negative AM than for PANPS (p = 0.0002) and were less effective for PANPS than for plexin D1‐IgG‐negative AM (p = 0.0071) (Table 6). Pregabalin and mirogabalin were frequently used and were partly effective for pain in both conditions. Comparisons between PANPS and plexin D1‐IgG‐negative/unknown AM showed similar trends (Table S10).
TABLE 6.
Comparison of treatment modalities and responses between patients with PANPS and plexin‐IgG‐negative AM.
| PANPS (n = 14) | Plexin D1‐IgG‐negative AM (n = 25) | p a | |
|---|---|---|---|
| Oral corticosteroids (%) | 4/12 (33.3) | 18/23 (78.3) | 0.0244 |
| Generally effective | 3/4 (75.0) | 15/18 (83.3) | NS |
| Effective for pain | 3/4 (75.0) | 13/15 (86.7) | NS |
| IVMP pulse therapy (%) | 5/13 (38.5) | 21/25 (84.0) | 0.0086 |
| Generally effective | 4/5 (80.0) | 18/21 (85.7) | NS |
| Effective for pain | 4/5 (80.0) | 15/18 (83.3) | NS |
| Immunosuppressants b (%) | 1/11 (9.1) | 7/16 (43.8) | 0.0089 |
| Generally effective | 0/1 (0) | 6/6 (100) | NA |
| Effective for pain | 0/1 (0) | 6/6 (100) | NA |
| Plasmapheresis (%) | 3/12 (25.0) | 11/17 (64.7) | 0.0604 |
| Generally effective | 3/3 (100) | 11/11 (100) | NS |
| Effective for pain | 2/3 (66.7) | 11/11 (100) | NS |
| IVIg (%) | 1/12 (8.3) | 10/19 (52.6) | 0.0201 |
| Generally effective | 0/1 (0) | 8/10 (80.0) | NA |
| Effective for pain | 0/1 (0) | 7/8 (87.5) | NA |
| Anti‐allergic drug (%) | 4/12 (33.3) | 15/15 (100) | 0.0002 |
| Generally effective | 0/3 (0) | 12/13 (92.3) | 0.0071 |
| Effective for pain | 0/3 (0) | 4/10 (40.0) | NS |
| Pregabalin/mirogabalin (%) | 10/13 (76.9) | 13/19 (68.4) | NS |
| Effective for pain | 9/10 (90.0) | 12/13 (92.3) | NS |
| Other anti‐epileptic drugs c (%) | 2/12 (16.7) | 7/16 (43.8) | NS |
| Effective for pain | 0/2 (0) | 7/7 (100) | NA |
| Anti‐depressants d (%) | 2/13 (15.4) | 5/18 (27.8) | NA |
| Effective for pain | 1/2 (50.0) | 5/5 (100) | NA |
Abbreviations: AM, atopic myelitis; IVIg, intravenous immunoglobulin; IVMP, intravenous methylprednisolone; NA, not applicable; NS, not significant; PANPS, plexin D1‐IgG‐associated neuropathic pain syndrome.
Fisher's exact test (two‐sided test).
In plexin D1‐IgG‐negative AM, cyclosporin was used in seven cases (100%), whereas in PANPS, azathioprine was used in one case.
In plexin D1‐IgG‐negative AM, clonazepam was used in five cases (71.4%), and gabapentin and carbamazepine were used in one case each (14.3%), whereas in PANPS, clonazepam and gabapentin were used in one case each.
In plexin D1‐IgG‐negative AM, amitriptyline was used in four cases and nortriptyline was used in one case, whereas in PANPS, nortriptyline and duloxetine/amitriptyline were used in one case each.
Even when the three plexin D1‐IgG‐positive patients with established demyelinating diseases were included in the PANPS group, similar results were obtained in comparisons between PANPS and plexin D1‐IgG‐negative AM concerning the key clinical and laboratory findings (Supporting Information, Table S11).
3.8. Comparison of Clinical and Laboratory Findings Between AM Patients With and Without Plexin D1‐IgG
Finally, clinical and laboratory findings were compared between AM patients with and without plexin D1‐IgG, although the number of AM patients with plexin D1‐IgG was small in the present study. Practically, similar results were obtained to those of comparisons between PANPS and plexin D1‐IgG‐negative AM patients (Tables S12–16). Importantly, the frequencies of neuropathic pain at disease onset, allodynia, and facial pain were significantly higher in AM patients with plexin D1‐IgG than in those without plexin D1‐IgG (p = 0.0156, p = 0.0157, and p = 0.0222, respectively). Compared with AM patients without plexin D1‐IgG, total IgE levels and the use of anti‐allergic drugs were significantly lower (p = 0.0422 and p = 0.0088, respectively), and the frequency of atopic dermatitis was nearly significantly lower (p = 0.0501) in AM patients with plexin D1‐IgG. Additionally, the frequencies of aged onset (≥ 50 years old), erythromelalgia, facial paresthesia/dysesthesia, progressive course, and decreased C fiber threshold on the CPT were 2‐ to 4‐fold higher in AM patients with plexin D1‐IgG than in those without, while the frequencies of motor weakness at onset and hyperIgEemia were 2‐ to 3‐fold lower in AM patients with plexin D1‐IgG than in those without, although these differences did not reach statistical significance.
4. Discussion
The present nationwide survey characterized the clinical and laboratory features of AM and PANPS (summary in Supporting Information). This study had a higher response rate in the preliminary survey (62.7%) than in previous surveys (52.3% and 45.9% in the 2001 and 2006 surveys, respectively) [4, 5] and a similar recovery rate in the secondary survey (100%, 73.7%, and 83.7% in the 2001, 2006, and present surveys, respectively) [4, 5], suggesting a similar capture rate between this and the previous surveys [4, 5]. The number of AM patients collated in the present survey (2023) is similar to that of the 2006 survey [5], considering the difference in survey periods (5 vs. 10 years). This finding suggests that a relatively stable number of AM patients exists in Japan, although it remains a rare condition. AM exhibited similar cardinal features to those identified in the previous surveys: [4, 5] predominant paresthesia/dysesthesia and hypesthesia in addition to pyramidal motor weakness in the limbs. These symptoms/signs are consistent with the relatively high frequencies of central MEP and SSEP abnormalities together with frequent cervical cord lesions on MRI, suggesting that the cervical spinal cord involving the posterior column and pyramidal tracts is the site responsible for lesions in this condition. For the first time, the present study revealed high frequencies of neuropathic pain during illness and CPT abnormalities, with particularly increased thresholds in all fibers. These findings suggest that the dorsal horns (or DRG/dorsal roots) are also frequently damaged in AM.
More than 80% of AM patients had a long‐standing illness with a relapsing remitting/fluctuating or chronic progressive course. The disease duration in the present survey was much longer than those in the previous two surveys (mean: 11.0 years in the present survey vs. 4.5 in the 2001 survey [4] and 6.6 years in the 2006 survey [5]). Moreover, the present study revealed that EDSS scores were positively correlated with disease duration, in accord with a previous small case study, in which disability progression was reported to be mainly caused by progressive sensory functional score worsening [22]. This may also explain the observation that the final EDSS scores were greater in the present survey than in the 2006 survey (mean: 3.0 in the present survey vs. 2.3 in the 2006 survey [5]; EDSS scores were not available in the 2001 survey [4]), reflecting a longer disease duration in the present survey. Compared with the 2006 survey [6], various immunotherapy modalities were introduced in AM patients in the current survey; most AM patients were treated with only IVMP and/or oral corticosteroids in the 2006 survey, whereas plasmapheresis, IVIg, and immunosuppressants other than corticosteroids were frequently introduced in the present survey (approximately 2%–10% in the 2006 survey [6] vs. 30%–50% in the present survey). Overall, immunotherapies generally had beneficial effects on AM symptoms/signs, although their effectiveness on pain was more favorable with plasmapheresis than with other treatments in the current survey. In the 2006 survey, few (three) AM patients received plasmapheresis, whereas in the present survey, 25 patients reported the most favorable effects of plasmapheresis among the immunotherapies, a finding that is also consistent with a previous small case series [23]. These results suggest that some humoral factors may be involved, particularly in generating pain in AM.
The main underlying plexin D1‐IgG diseases were AM and neuropathic pain‐related conditions, such as SFN, PTN, and erythromelalgia, in the present study, which is consistent with previous reports [15, 16, 17]. The comparison between PANPS and plexin D1‐IgG‐negative AM patients revealed that plexin D1‐IgG was associated with neuropathic pain at onset and with allodynia and erythromelalgia during the entire course, irrespective of the inclusion or exclusion of three patients with established demyelinating diseases. These findings suggest that plexin D1 expressed in small pain‐conducting neurons as well as in sympathetic nerve terminals is targeted by plexin D1‐IgG (Figure 2) [15, 17]. Of note, facial pain was much more common in plexin D1‐IgG‐positive patients than in plexin D1‐IgG‐negative AM patients, which suggests that the involvement of small pain‐conducting neurons in the trigeminal ganglia is characteristic of PANPS (Figure 2). By contrast, pyramidal motor weakness, which was experienced by most plexin D1‐IgG‐negative AM patients, was less frequent in PANPS. This may explain the significantly lower EDSS scores in PANPS than in plexin D1‐IgG‐negative AM patients, because the EDSS predominantly reflects motor impairment [20].
FIGURE 2.
Schema of the targets and manifestations of plexin D1‐IgG‐associated neuropathic pain syndrome (PANPS). Plexin D1‐IgG targets small pain‐conducting neurons in the DRG, causing SFN that presents as limb pain, and in the trigeminal ganglia, causing PTN that presents as facial pain. In a fraction of AM patients, plexin D1‐IgG targets small pain‐conducting neuron terminals in spinal dorsal horn layers I and II, presenting as limb and trunk pain. A small percentage of MS and NMOSD patients also have plexin D1‐IgG, which may also target pain‐conducting neuron terminals in the spinal dorsal horns. Plexin D1‐IgG might target sympathetic nerve terminals that express plexin D1 and may cause erythromelalgia. AM, atopic myelitis; DRG, dorsal root ganglion; Ig, immunoglobulin; MS, multiple sclerosis; NMOSD, neuromyelitis optica spectrum disorder; PTN, painful trigeminal neuropathy; SFN, small fiber neuropathy.
Interestingly, plexin D1‐IgG‐negative AM patients preferentially showed increased thresholds in all fibers including Aβ, Aδ, and C fibers, whereas PANPS patients demonstrated not only increased but also decreased thresholds in all fibers, particularly in C fibers. Decreased thresholds in C and Aδ fibers may contribute to developing neuropathic pain in PANPS. These findings are consistent with the previous observation that intrathecal injection of patients' plexin D1‐IgG in mice induced mechanical and thermal hypersensitivity and activation of small DRG neurons that express plexin D1 [17]. Application of plexin D1‐IgG to cultured DRG neurons increased membrane permeability, leading to cellular swelling [15]. Semaphorin 3E‐plexin D1 signaling mediates repulsive axon guidance [24, 25], altering actin cytoskeleton dynamics [25], which are involved in receptor distribution and synapse plasticity [26]. Plexin D1 interaction inhibits inappropriate synapse formation [27], which might be influenced by anti‐plexin D1 antibodies. We are currently investigating the mechanisms of neuronal hyper‐excitability induced by plexin D1‐IgG‐triggered actin reorganization in cultured DRG neurons.
Neuropathic pain is a constellation of heterogeneous conditions. Among them, autoantibody‐mediated neuropathic pain comprises a unique subgroup involving immune mechanisms. Plexin D1‐IgG was particularly associated with pain onset, allodynia, facial pain, erythromelalgia, and a chronic progressive course in this survey. These characteristics were also associated with plexin D1‐IgG positivity even in comparisons between AM patients with and without plexin D1‐IgG. The female preponderance in PANPS reported in our single institutional study [15] was also confirmed by this nationwide survey. Importantly, the presence of anti‐plexin D1 antibodies was also reported in a fraction of SFN patients among not only non‐Japanese Asians, such as Koreans [17] and Singaporeans [28], but also Westerners such as North Americans (9% of SFN) [18, 29, 30], who also showed a female preponderance and chronic persistent pain [18, 29]. We therefore propose that PANPS is a new clinical entity of autoantibody‐mediated neuropathic pain presenting unique features, although further clinical and laboratory studies together with mechanistic studies are required.
The strong positive correlation between EDSS scores and disease duration in PANPS may be partly related to the chronic progressive nature of this condition and the general absence of therapeutic interventions. Plasmapheresis and corticosteroids had some beneficial effects in the limited number of PANPS patients in the present survey. IVIg and rituximab depletion of B cells were reportedly effective for alleviating neuropathic pain in patients with plexin D1‐IgG‐positive SFN [28, 30]. Given that an anti‐plexin D1 antibody assay is available in Japan (Department of Neurology, Kyushu University) [15, 16, 17] and outside Japan (Neuromuscular Laboratory, Washington University) [28, 29, 30], we consider anti‐plexin D1 antibody measurement a promising potential biomarker for implementing immunotherapy, at least in the context of neuropathic pain and SFN, worldwide.
The present study revealed that atopic diathesis was not associated with PANPS. Atopic dermatitis and allergic rhinitis—two major atopic/allergic conditions associated with AM—were significantly less common in PANPS than in AM patients, which was consistent with the normal range of serum IgE levels and the less frequent occurrence of hyperIgEemia in PANPS. Anti‐allergic drugs that were generally effective in most plexin D1‐IgG‐negative AM patients were infrequently used in PANPS patients and had no beneficial effects. These findings suggest that atopy, which is defined as enhanced IgE responsiveness to common environmental antigens [31], is not a risk or contributing factor for PANPS. The risk factors for plexin D1‐IgG emergence should be elucidated in the future.
The current study has several limitations. First, the number of collated cases was relatively small because of the rarity of the diseases, which resulted in low statistical power. Second, the plexin D1‐IgG assay was only available in one academic laboratory in Japan during the study period. Thus, more than half of the AM patients collected did not undergo plexin D1‐IgG measurement in the present survey, which further limited the number of plexin D1‐IgG‐negative AM and plexin D1‐IgG‐positive cases. Third, we did not standardize the treatment efficacy measures, which were subjectively evaluated by physicians, because this was a retrospective large‐scale nationwide survey involving more than 1500 neurology/pediatric neurology departments. However, the significant decrease in EDSS scores from the peak of illness to the last examination in this survey supports the beneficial effects of immunotherapies. Fourth, as the timing of symptomatic treatments in relation to immunotherapy was not specified in this survey, the effects of symptomatic therapy should be cautiously interpreted. Our study results should therefore be regarded as exploratory, although we have presented their significance in the comparative analyses of the collated cases.
In conclusion, both AM and plexin D1‐IgG‐positive patients presented long‐lasting paresthesia/dysesthesia and neuropathic pain. Nearly half of all AM patients had moderate disability (EDSS scores ≥ 4.5) after an average of 11 years of illness, and one‐quarter of plexin D1‐IgG‐positive patients had moderate disability after an average of 18.6 years of illness. The early introduction of immunotherapies should be recommended in these conditions.
Author Contributions
All authors contributed to the study design and reviewed the manuscript. J.‐i.K., T.F., M.M., M.U., S.Y., A.S., T.I., S.K., N.I., and Y.N. collected the clinical data. J.‐i.K., X.Z., T.F., M.M., M.U., S.Y., A.S., T.I., N.I., and Y.N. analyzed the results. J.‐i.K., X.Z., T.F., M.M., M.U., S.K., N.I., and Y.N. were involved in the interpretation of the results and drafted the manuscript.
Conflicts of Interest
J.‐i.K. reports grants from the Japan Agency for Medical Research and Development (AMED), Japan (Grants 21ek0109547h0001, 22ek0109547h0002, 23ek0109547h0003, 23ek0109626h0001, 24ek0109626h0002, 25ek0109626h0003, and 25ek0109817h0001); a Grant‐in‐Aid for Scientific Research (B) (JSPS KAKENHI Grant 22H02985); a Grant‐in‐Aid for Challenging Research (Exploratory) (JSPS KAKENHI Grant 23K18266); research funds from Sumitomo Pharma; and consultancy fees, speaking fees, and/or honoraria from Novartis Pharma, Argenx, Mitsubishi Tanabe Pharma, Biogen Japan, the Takeda Pharmaceutical Co. Ltd., Chugai Pharmaceutical Co. Ltd., Daiichi Sankyo Co. Ltd., and Alexion Pharma. X.Z. reports grants from JSPS KAKENHI (Grants JP21K15703 and JP23K14783). T.F. reports grants from JSPS KAKENHI (Grants JP23K06945 and JP24H00067), AMED Japan (Grant 23ek0109626h0001); research funds from Yamasa Corporation; and speaking fees from Daiichi Sankyo. A.S. reports a grant from JSPS KAKENHI (Grant JP23K14761). T.I. reports a grant from JSPS KAKENHI (Grant JP22K07499) and speaking fees from Tsumura & Co. and Kowa Co. Ltd. N.I. reports grants from JSPS KAKENHI (Grant JP21K07464), AMED Japan (Grant 23zf0127004h0003), the Health and Labour Sciences Research Grant on Intractable Diseases (Neuroimmunological Diseases) from the Ministry of Health, Labour and Welfare of Japan (23FC1009), Sumitomo Pharma, Daiichi Sankyo Co. Ltd., Mitsubishi Tanabe Pharma, Novartis Pharma, Biogen Japan, Yamasa Corporation, Kyowa Kirin Co. Ltd., and Nippon Boehringer Ingelheim Co. Ltd.; and honoraria from Alexion Pharma, Novartis Pharma, Argenx, Mitsubishi Tanabe Pharma, Biogen Japan, the Takeda Pharmaceutical Co. Ltd., UCB Japan, Ono Pharmaceutical Co. Ltd., Chugai Pharmaceutical Co. Ltd., Daiichi Sankyo Co. Ltd., Teijin Healthcare Co. Ltd., Amgen Inc., and Eisai Co. Ltd. Y.N. reports grants from JSPS KAKENHI (Grants JP21K07467 and JP24K10666) and speaking fees and/or honoraria from Novartis Pharma, Biogen Japan, the Takeda Pharmaceutical Co. Ltd., Chugai Pharmaceutical Co. Ltd., and Chugai‐Igakusha. M.M., M.U., S.Y., and S.K. have nothing to declare.
Supporting information
Figure S1: Representative TBA results for plexin D1‐IgG. Positivity for plexin D1‐IgG was detected by a TBA using mouse lumbar DRGs, as previously described [15, 16, 17]. (A) Sera from a representative AM patient without plexin D1‐IgG, (B) sera from a representative AM patient with plexin D1‐IgG, and (C) sera from a representative AM patient with plexin D1‐IgG after incubation with recombinant plexin D1 are shown. Note the membranous staining patterns of the plexin D1‐IgG‐positive patient in (B), which disappeared after incubation with recombinant plexin D1 in (C). AM, atopic myelitis; DRG, dorsal root ganglion; Ig, immunoglobulin; TBA, tissue‐based assay.
Figure S2: Secondary survey flow chart. In the secondary survey, detailed information was collated from 71 AM patients, as well as 11 plexin D1‐IgG‐positive patients who did not meet the AM criteria [10]. Among the 71 AM patients, plexin D1‐IgG was examined in 31 patients, of whom 6 (19.4%) were positive. Of the 17 plexin D1‐IgG‐positive patients, 3 patients with established demyelinating diseases were excluded, and 14 patients were tentatively combined in the PANPS group. These PANPS patients were compared with 25 plexin D1‐IgG‐negative AM patients or 65 plexin D1‐IgG‐negative/unknown AM patients. AM, atopic myelitis; Ig, immunoglobulin; PANPS, plexin D1‐associated neuropathic pain syndrome.
Table S1: 1 Survey items for AM in the secondary survey questionnaire.
Table S2: Comparison of the numbers of collated cases between the 2006 and present nationwide surveys.
Table S3: Laboratory features of patients with AM.
Table S4: Treatment modalities and responses in patients with AM.
Table S5: Underlying diseases in plexin D1‐IgG‐positive cases.
Table S6: Comparison of clinical features between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S7: Comparison of the distribution of neuropathic pain and paresthesia/dysesthesia between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S8: Comparison of laboratory findings between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S9: Comparison of atopic diathesis between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S10: Comparison of treatment modalities and responses between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S11: Comparison of key clinical and laboratory features between patients with PANPS including three with demyelinating disease and plexin D1‐IgG‐negative AM.
Table S12: Comparison of clinical features between AM patients with and without plexin D1‐IgG.
Table S13: Comparison of the distribution of neuropathic pain and paresthesia/dysesthesia between AM patients with and without plexin D1‐IgG.
Table S14: Comparison of laboratory findings between AM patients with and without plexin D1‐IgG.
Table S15: Comparison of atopic diathesis between AM patients with and without plexin D1‐IgG.
Table S16: Comparison of treatment modalities and responses between AM patients with and without plexin D1‐IgG.
Data S1: acn370232‐sup‐0004‐Supinfo.docx.
Acknowledgments
This study was funded in part by grants from the Japan Society for the Promotion of Science (JSPS) KAKENHI (Grants JP22H02985 and JP23K14783) and from the Japan Agency for Medical Research and Development (AMED) (Grants 23ek0109547h0003, 23ek0109626h0001, 24ek0109626h0002, 25ek0109626h0003, and 25ek0109817h0001). We thank all of the neurologists, pediatricians, and pediatric neurologists who responded to our surveys. We also thank Bronwen Gardner, PhD, from Edanz (https://jp.edanz.com/ac) for editing the English text of a draft of this manuscript.
Funding: This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI (Grants 22H02985, 23K18266, JP21K15703, JP23K06945, JP24H00067, JP22K07499, JP21K07467, JP24K10666, JP21K07464, JP23K14761, JP22H02985 and JP23K14783) and from the Japan Agency for Medical Research and Development (AMED) (Grants 21ek0109547h0001, 22ek0109547h0002, 23ek0109547h0003, 23ek0109626h0001, 24ek0109626h0002, 25ek0109626h0003, and 25ek0109817h0001, 23zf0127004h0003), Ministry of Health, Labour and Welfare of Japan (Grant 23FC1009).
Funding Statement
This work was funded by Japan Agency for Medical Research and Development grants 23ek0109547h0003, 23ek0109626h0001, 24ek0109626h0002, 25ek0109626h0003, 25ek0109817h0001, 21ek0109547h0001, 22ek0109547h0002, and 23zf0127004h0003; Japan Society for the Promotion of Science grants JP22H02985, JP23K14783, 22H02985, 23K18266, JP21K15703, JP23K06945, JP24H00067, JP22K07499, JP21K07467, JP24K10666, JP21K07464, and JP23K14761; Ministry of Health, Labour and Welfare of Japan grant 23FC1009.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Kira J., Yamasaki K., Kawano Y., and Kobayashi T., “Acute Myelitis Associated With hyperIgEemia and Atopic Dermatitis,” Journal of the Neurological Sciences 148 (1997): 199–203. [DOI] [PubMed] [Google Scholar]
- 2. Kira J., Kawano Y., Yamasaki K., and Tobimatsu S., “Acute Myelitis With hyperIgEaemia and Mite Antigen Specific IgE: Atopic Myelitis,” Journal of Neurology, Neurosurgery, and Psychiatry 64 (1998): 676–679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Kira J., Kawano Y., Horiuchi I., et al., “Clinical, Immunological and MRI Features of Myelitis With Atopic Dermatitis (Atopic Myelitis),” Journal of the Neurological Sciences 162 (1999): 56–61. [DOI] [PubMed] [Google Scholar]
- 4. Osoegawa M., Ochi H., Minohara M., et al., “Myelitis With Atopic Diathesis: A Nationwide Survey of 79 Cases in Japan,” Journal of the Neurological Sciences 209 (2003): 5–11. [DOI] [PubMed] [Google Scholar]
- 5. Isobe N., Kira J., Kawamura N., Ishizu T., Arimura K., and Kawano Y., “Neural Damage Associated With Atopic Diathesis: A Nationwide Survey in Japan,” Neurology 73 (2009): 790–797. [DOI] [PubMed] [Google Scholar]
- 6. Osoegawa M., Ochi H., Yamada T., et al., “High Incidence of Subclinical Peripheral Neuropathy in Myelitis With hyperIgEaemia and Mite Antigen‐Specific IgE (Atopic Myelitis): An Electrophysiological Study,” Internal Medicine 41 (2002): 684–691. [DOI] [PubMed] [Google Scholar]
- 7. Kikuchi H., Osoegawa M., Ochi H., et al., “Spinal Cord Lesions of Myelitis With hyperIgEemia and Mite Antigen Specific IgE (Atopic Myelitis) Manifest Eosinophilic Inflammation,” Journal of the Neurological Sciences 183 (2001): 73–78. [DOI] [PubMed] [Google Scholar]
- 8. Osoegawa M., Ochi H., Kikuchi H., et al., “Eosinophilic Myelitis Associated With Atopic Diathesis: A Combined Neuroimaging and Histologic Study,” Acta Neuropathologica 105 (2003): 289–295. [DOI] [PubMed] [Google Scholar]
- 9. Yamasaki R., Fujii T., Wang B., et al., “Allergic Inflammation Leads to Neuropathic Pain via Glial Cell Activation,” Journal of Neuroscience 36, no. 47 (2016): 11929–11945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Isobe N., Kanamori Y., Yonekawa T., et al., “First Diagnostic Criteria for Atopic Myelitis With Special Reference to Discrimination From Myelitis‐Onset Multiple Sclerosis,” Journal of the Neurological Sciences 316 (2012): 30–35. [DOI] [PubMed] [Google Scholar]
- 11. Zoli A., Mariano M., Fusari A., et al., “Atopic Myelitis: First Case Report Outside Japan?,” Allergy 60 (2005): 410–411. [DOI] [PubMed] [Google Scholar]
- 12. Gregoire S. M., Mormont E., Laloux P., Godfraind C., and Gilliard C., “Atopic Myelitis: A Clinical, Biological, Radiological and Histopathological Diagnosis,” Journal of the Neurological Sciences 247 (2006): 231–235. [DOI] [PubMed] [Google Scholar]
- 13. Yoon J. H., Joo I. S., Li W. Y., and Sohn S. Y., “Clinical and Laboratory Characteristics of Atopic Myelitis: Korean Experience,” Journal of the Neurological Sciences 285 (2009): 154–158. [DOI] [PubMed] [Google Scholar]
- 14. Isaacs J. D., Bodini B., Ciccarelli O., Scadding G. K., and Thompson A. J., “Atopic Myelitis in a European Woman Residing in Japan,” Journal of Neurology, Neurosurgery, and Psychiatry 82, no. 9 (2010): 1022–1024, 10.1136/jnnp.2009.186775. [DOI] [PubMed] [Google Scholar]
- 15. Fujii T., Yamasaki R., Iinuma K., et al., “A Novel Autoantibody Against Plexin D1 in Patients With Neuropathic Pain,” Annals of Neurology 84, no. 2 (2018): 208–224. [DOI] [PubMed] [Google Scholar]
- 16. Fujii T., Yamasaki R., Miyachi Y., et al., “Painful Trigeminal Neuropathy Associated With Anti‐Plexin D1 Antibody,” Neurology: Neuroimmunology & Neuroinflammation 7 (2020): e819, 10.1212/NXI.0000000000000819. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Fujii T., Lee E.‐J., Miyachi Y., et al., “Anti‐Plexin D1 Antibodies Relate to Small Fiber Neuropathy and Induce Neuropathic Pain in Animals,” Neurology: Neuroimmunology & Neuroinflammation 8, no. 5 (2021): e1028, 10.1212/NXI.0000000000001028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Zeidman L. A., Saini P., and Mai P., “Immune‐Mediated Small Fiber Neuropathy With Trisulfated Heparin Disaccharide, Fibroblast Growth Factor Receptor 3, or Plexin D1 Antibodies: Presentation and Treatment With Intravenous Immunoglobulin,” Journal of Clinical Neuromuscular Disease 24, no. 1 (2022): 26–37, 10.1097/CND.0000000000000423. [DOI] [PubMed] [Google Scholar]
- 19. Daifallah O., Adham Farah A., and Dawes J. M., “A Role for Pathogenic Autoantibodies in Small Fiber Neuropathy?,” Frontiers in Molecular Neuroscience 16 (2023): 1254854, 10.3389/fnmol.2023.1254854. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Kurtzke J. F., “Rating Neurologic Impairment in Multiple Sclerosis: An Expanded Disability Status Scale (EDSS),” Neurology 33 (1983): 1444–1452. [DOI] [PubMed] [Google Scholar]
- 21. Finnerup N. B., Haroutounian S., Kamerman P., et al., “Neuropathic Pain: An Updated Grading System for Research and Clinical Practice,” Pain 157 (2016): 1599–1606, 10.1097/j.pain.0000000000000492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Ainiding G., Yamashita K., Torii T., et al., “Clinical Disability Progression and Platelet GP IIb/IIIa Values in Patients With Atopic Myelitis,” Journal of Neuroimmunology 246 (2012): 108–112. [DOI] [PubMed] [Google Scholar]
- 23. Murai H., Arahata H., Osoegawa M., et al., “Effect of Immunotherapy in Myelitis With Atopic Diathesis,” Journal of the Neurological Sciences 227 (2004): 39–47. [DOI] [PubMed] [Google Scholar]
- 24. Pecho‐Vrieseling E., Sigrist M., Yoshida Y., Jessell T. M., and Arber S., “Specificity of Sensory‐Motor Connections Encoded by Sema3e‐Plxnd1 Recognition,” Nature 459, no. 7248 (2009): 842–846, 10.1038/nature08000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Kim N., Li Y., Yu R., et al., “Repulsive Sema3E‐Plexin‐D1 Signaling Coordinates Both Axonal Extension and Steering via Activating an Autoregulatory Factor, Mtss1,” eLife 13 (2024): e96891, 10.7554/eLife.96891. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Okabe S., “Regulation of Actin Dynamics in Dendritic Spines: Nanostructure, Molecular Mobility, and Signaling Mechanisms,” Molecular and Cellular Neurosciences 109 (2020): 103564, 10.1016/j.mcn.2020.103564. [DOI] [PubMed] [Google Scholar]
- 27. Wang F., Eagleson K. L., and Levitt P., “Positive Regulation of Neocortical Synapse Formation by the Plexin‐D1 Receptor,” Brain Research 1616 (2015): 157–165, 10.1016/j.brainres.2015.05.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Chan A. C. Y., Shahana R., Siah K. T. H., et al., “Efficacy of Rituximab on Antibody‐Positive Small Fiber Neuropathy: A Series of 5 Cases,” Journal of Neuroimmunology 401 (2025): 578559, 10.1016/j.jneuroim.2025.578559. [DOI] [PubMed] [Google Scholar]
- 29. Murin P. J., Massabki I., and Kafaie J., “Anti–Plexin‐D1 Seropositive Small Fiber Neuropathy: Clinical Phenotype, Demographics, and Literature Review,” Journal of Clinical Neuromuscular Disease 25, no. 3 (2024): 132–140. [DOI] [PubMed] [Google Scholar]
- 30. Zeidman L. A., “Effectiveness of IVIG on Non–Length‐Dependent Skin Biopsies in Small Fiber Neuropathy With Plexin D1, Trisulfated Heparin Disaccharide, and Fibroblast Growth Factor Receptor 3 Autoantibodies,” Journal of Clinical Neuromuscular Disease 4 (2024): 184–196. [DOI] [PubMed] [Google Scholar]
- 31. Stenius B., Wide L., Seymour W. M., et al., “Clinical Significance of Specific IgE to Common Allergens. I. Relationship of Specific IgE Against Dermatophagoides spp and Grass Pollen to Skin and Nasal Tests and History. Clin,” Allergy 1 (1971): 37–55. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Representative TBA results for plexin D1‐IgG. Positivity for plexin D1‐IgG was detected by a TBA using mouse lumbar DRGs, as previously described [15, 16, 17]. (A) Sera from a representative AM patient without plexin D1‐IgG, (B) sera from a representative AM patient with plexin D1‐IgG, and (C) sera from a representative AM patient with plexin D1‐IgG after incubation with recombinant plexin D1 are shown. Note the membranous staining patterns of the plexin D1‐IgG‐positive patient in (B), which disappeared after incubation with recombinant plexin D1 in (C). AM, atopic myelitis; DRG, dorsal root ganglion; Ig, immunoglobulin; TBA, tissue‐based assay.
Figure S2: Secondary survey flow chart. In the secondary survey, detailed information was collated from 71 AM patients, as well as 11 plexin D1‐IgG‐positive patients who did not meet the AM criteria [10]. Among the 71 AM patients, plexin D1‐IgG was examined in 31 patients, of whom 6 (19.4%) were positive. Of the 17 plexin D1‐IgG‐positive patients, 3 patients with established demyelinating diseases were excluded, and 14 patients were tentatively combined in the PANPS group. These PANPS patients were compared with 25 plexin D1‐IgG‐negative AM patients or 65 plexin D1‐IgG‐negative/unknown AM patients. AM, atopic myelitis; Ig, immunoglobulin; PANPS, plexin D1‐associated neuropathic pain syndrome.
Table S1: 1 Survey items for AM in the secondary survey questionnaire.
Table S2: Comparison of the numbers of collated cases between the 2006 and present nationwide surveys.
Table S3: Laboratory features of patients with AM.
Table S4: Treatment modalities and responses in patients with AM.
Table S5: Underlying diseases in plexin D1‐IgG‐positive cases.
Table S6: Comparison of clinical features between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S7: Comparison of the distribution of neuropathic pain and paresthesia/dysesthesia between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S8: Comparison of laboratory findings between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S9: Comparison of atopic diathesis between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S10: Comparison of treatment modalities and responses between patients with PANPS and plexin D1‐IgG‐negative/unknown AM.
Table S11: Comparison of key clinical and laboratory features between patients with PANPS including three with demyelinating disease and plexin D1‐IgG‐negative AM.
Table S12: Comparison of clinical features between AM patients with and without plexin D1‐IgG.
Table S13: Comparison of the distribution of neuropathic pain and paresthesia/dysesthesia between AM patients with and without plexin D1‐IgG.
Table S14: Comparison of laboratory findings between AM patients with and without plexin D1‐IgG.
Table S15: Comparison of atopic diathesis between AM patients with and without plexin D1‐IgG.
Table S16: Comparison of treatment modalities and responses between AM patients with and without plexin D1‐IgG.
Data S1: acn370232‐sup‐0004‐Supinfo.docx.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.