Abstract
Objective: Examining the unresolved relationship between the lower motor neuron disorder progressive muscular atrophy (PMA) and ALS is important in clinical practice because of emerging therapies.
Methods: Spinal and brainstem tissues donated from patients with ALS/motor neuron disorder (n = 81) were examined. Using retrospective case note review, the authors assigned patients into three categories: PMA (12), PMA progressing to ALS (6), and ALS ab initio (63). Conventional stains for long tract degeneration and immunocytochemistry for ubiquitin and the macrophage marker CD68 were examined.
Results: Rapid progression and typical ubiquitinated inclusions in lower motor neurons were present in 77 (95%) of the cases. Immunocytochemistry for CD68 was a more sensitive marker of long tract pathology in comparison with conventional stains. Half of the cases with PMA showed corticospinal tract degeneration by CD68.
Conclusion: Patients with PMA frequently have undetected long tract pathology and most have ubiquitinated inclusions typical of ALS. A patient presenting with PMA with rapid clinical evolution likely has the pathology and pathophysiology of ALS whether or not upper motor neuron signs evolve.
ALS/motor neuron disease (ALS/MND) is traditionally regarded as comprising four main subtypes: combined upper and lower motor degeneration (ALS), pure upper motor neuron (UMN) degeneration (primary lateral sclerosis [PLS]), pure bulbar lower motor neuron (LMN) degeneration (progressive bulbar palsy [PBA]), and pure spinal LMN degeneration (progressive muscular atrophy [PMA]). PBA usually evolves rapidly into classic ALS whereas PLS most commonly remains confined to the UMN pool at least for some years. In contrast, PMA shows more variable evolution, so that some cases develop into ALS but others remain a pure LMN disorder. The diagnosis of ALS/MND is defined by the El Escorial criteria, and subsequent refinements at the Airlie House meeting.1-3⇓⇓ ALS is used as the generic disease term, in contrast to the traditional UK medical terminology of MND, and is categorized into definite ALS, probable ALS, and possible ALS. In this diagnostic system, devised for research studies and clinical trials, PMA corresponds to an even less certain diagnostic category of suspected ALS.
The differential diagnosis of PMA includes other LMN disorders, including Kennedy syndrome (X-linked bulbo-spinal neuronopathy), late onset spinal muscular atrophy, and motor neuropathies including multifocal motor neuropathy with conduction block.4
Previous reports have drawn attention to cases diagnosed as PMA in life in which some corticospinal tract (CST) degeneration or microglial activation was apparent at autopsy.5-9⇓⇓⇓⇓ We hypothesize that PMA represents one end of a spectrum of motor system disorders, spanning ALS, with PLS (and ALS-dementia) at the other end. To investigate this further we have examined postmortem tissues of patients with MND to look for common molecular pathologies between PMA and ALS, along with those patients who progress from a LMN presentation to combined degeneration (PMA→ALS).
Methods.
Patients and clinical evaluations.
The participants in this study were patients under the care of the Motor Neurone Disease Clinic in Newcastle upon Tyne between 1985 and 1999 from whose relatives consent was obtained to archive the brain and spinal cord for research purposes after death. The only cases (n = 6) excluded were those in whom the clinical diagnosis was that of a pure UMN syndrome. Control spinal cord tissue from 10 tissue donors was obtained from the Newcastle Brain Tissue Bank from autopsy donations from individuals who did not have a neurologic disease. All patients under the care of the MND Clinic had progressive motor system degeneration. Participants in this study were diagnosed by a neurologist (P.J.S.) after full clinical evaluation and extensive investigation to exclude other diagnoses (hematologic and biochemical blood screen, autoimmune screen, serum immunoglobulins, CSF analysis, neurophysiology, neuroimaging, and muscle biopsy in selected cases) including other causes of motor system degeneration (genetic analysis for Kennedy disease). Five of the participants had genetic abnormalities in the SOD1 gene,10 and one had a familial LMN disease,11 described in previous clinicopathologic reports. The demographic features of these patients are summarized in table 1.
Table 1 Demographic and disease-related variables in the ALS autopsy cohort
Patients attending the MND clinic were regularly reviewed throughout the course of the disease at 2-month intervals. Patients who presented before establishment of the clinic had less systematic review and the range of intervals between last clinical review and death is given in table 1 for PMA diagnoses. For all but two cases this period was less than 3 months so that underascertainment of incident onset of new UMN signs is likely to be negligible. The patients with MND were divided into three subgroups by case note review on the basis of the presence of UMN and LMN signs. UMN involvement was defined using the El Escorial criteria2: exaggerated deep tendon jerks, brisk jaw jerk, positive Hoffmann reflex, clonus, and increased tone. LMN involvement was indicated by weakness on examination, muscle atrophy, fasciculations, and electrophysiologic evidence of chronic or active motor denervation. The resulting three diagnostic groups comprised PMA (LMN disease throughout), PMA→ALS (initial LMN presentation with evolution to ALS), and ALS from disease onset.
Neuropathology.
After death, permission to perform an autopsy and retain the brain and spinal cord for research purposes was obtained from the surviving relatives by the clinic staff. The duration between death and autopsy was less than 48 hours in all cases with a median of 14 hours. The spinal cord was removed using an anterior approach as part of a full autopsy. Among the autopsy-verified causes of death, respiratory failure and pneumonia predominated, although the latter diagnosis was less common in more recent deaths. Other infrequent causes included head injury (one case) and myocardial infarction (one case).
The brain and spinal cord were dissected while fresh for archival storage as frozen tissue and formalin-fixed tissues. The latter tissues were used for the current study, and included representative blocks of all brain and spinal cord regions. Tissues were fixed for between 3 and 12 weeks in buffered formalin before paraffin embedding. Sections for ubiquitin (polyclonal antibody, 1:100 dilution: Dako, Ely, UK) and CD68 (monoclonal antibody clone PGM-1, 1:50: Dako) immunocytochemistry were cut at 10-μm thickness and stained without antigen retrieval using a standard avidin-biotin complex method. Ubiquitin staining was performed on the following blocks as a minimum sample: upper medulla and spinal cord segments C6, C8, T6, T8, L4, and S1 or S2. For most cases additional levels were also available to include spinal and brainstem motor neuron pools. Luxol fast blue (LFB) staining was performed on sections cut at 20 μm using blocks from the cervical (C3, 4, or 5), thoracic (T2, 5, or 7), and medullary pyramid levels of the CST. A subgroup of 26 cases was sampled at high cervical (C3 segment) and high thoracic (T2 segment) levels using the osmium-based Marchi impregnation method for myelin breakdown products12 also cut as sections at 20 μm.
For ubiquitin staining of spinal motor neuron inclusions, serial ribbons of 10 sections were available at the spinal levels examined. Positive staining was judged to be present when the inclusions showed typical skein or compact or intermediate morphology as previously reported.10,13,14⇓⇓ The intensity of CST staining with the CD68, LFB, or Marchi methods was assessed semiquantitatively into four categories (corresponding to absent, mild, moderate, or severe degeneration) by a single observer. The sections were coded and assessed without knowledge of the clinical features. Figure 1 shows typical examples of these severity categories. Differences between the distributions of these categorical scores among the disease groups were examined using the Kruskal-Wallace method for nonparametric data.
Figure 1. Severity of corticospinal tract pathology: (A) macrophage permeation demonstrated by CD68 immunoreactive profiles, (B) Marchi impregnation, (C) Luxol fast blue (LFB). The columns correspond to grading scores: 0 = normal, 1 = mild, 2 = moderate, 3 = severe. Many cases with normal Marchi and LFB staining had mild or moderate CD68 scores.
The validity of the semiquantitative scoring method for CD68 immunoreactivity was assessed in a sample of 12 cases in comparison with objective measurements of the area occupied by staining in the lateral CST using digital image analysis. Five fields were captured in both left and right lateral CST regions of the cervical spinal cord using the ×40 objective (Zeiss Axioplan) and the area occupied by staining calculated (Image Pro Plus software, Media Cybernetics, MD). Figure 2 shows the mean percent area measurement for 10 fields from each of 12 cases compared to the CD68 score for each case.
Figure 2. Semiquantitative CD68 scores compared with percentage of field occupied by CD68 immunoreactivity using digital image analysis.
Results.
Ubiquitinated LMN inclusions (UBI).
None of the control cases showed UBI. Of the 81 MND cases included in this study, only 4 (∼5%) did not show any UBI, including two cases of PMA. UBI were absent from those cases despite extensive sampling, including the examination of ribbons of 10 serial sections at five limb enlargement levels of the spinal cord, together with brainstem sections including the XIIth, VIIth, and Vth motor nuclei. One patient with PMA had a protracted disease of 10 years’ duration with exclusively LMN signs. CST pathology was absent in all stains at all levels examined. UBI pathology was also absent and the possibility that this disease represented a form of adult-onset spinal muscular atrophy was considered. The second PMA case has been reported previously as familial MND associated with colon cancer.11 This patient showed CST pathology using CD68 (score 2) but no changes in myelin stains. Both these cases were excluded from further analysis. Two SOD1 familial ALS (FALS) cases showed neurofilament immunoreactive hyaline conglomerate inclusions (HCI) characteristic of some mutations in the gene encoding SOD115 rather than UBI. The clinical and pathologic features were otherwise typical of ALS and these patients were included in the analysis because of the genetic data.
These findings emphasize the near universal presence of UBI (or HCI) in ALS/MND (95%) based on this sampling protocol, and confirm the value of these pathologic findings as a diagnostic feature of the disease. All cases with UBI or HCI were therefore regarded as ALS/MND syndromes.
Clinical features of PMA cases.
Of the 12 patients with purely LMN signs two were excluded from further analysis based on the ubiquitin immunocytochemistry findings (as above). Among the remainder the clinical features are shown in table 2. Onset was usually either upper limb or bulbar region with progression to the lower limbs in almost all cases. None of these patients developed autonomic disturbance, sensory impairment, dementia, abnormal ocular movements, extrapyramidal features, or impaired coordination.
Table 2 Demographic data, anatomic site at presentation, and progression in the PMA cohort
CST pathology.
CST staining with all three methods had normal results in the 10 control cases examined. The appearance of white matter tracts showing each category of pathologic change is shown in figure 1. For the conventional stains the cases were allocated the highest score for either method (LFB or Marchi) and cases were allocated the highest score found at any anatomic level of the CST. The distributions of scores assigned to CD68 immunoreactivity and to LFB and Marchi staining are shown in figure 3. Figure 4 shows an LFB stained section of the spinal cord in a patient with a typical history of ALS (figure 4A) compared with a patient with a history of PMA (figure 4B) and illustrates the difficulty in interpreting CST changes using this method. The presence of CD68-immunoreactive profiles in the CST was unrelated to the presence of a terminal infection (i.e., pneumonia) and showed no trend related to year of death.
Figure 3. Corticospinal tract degeneration scores in each disease category using the CD68 and Luxol fast blue methods. □ = Progressive muscular atrophy (PMA); 
= PMA progressing to ALS; ▪ = ALS.
Figure 4. Luxol fast blue–stained spinal cord sections of patients with a clinical diagnosis of progressive muscular atrophy (A) and ALS (B) showing no discernible difference in the intensity of myelin staining using this method.
The CD68 method showed increased sensitivity in the detection of descending CST changes compared to LFB or Marchi methods in cases with PMA at presentation, with or without ultimate progression to ALS (see figure 1). The morphology of the CD68 immunoreactive profiles was predominantly that of circular profiles, rather than dendritic cells, and they were interpreted as phagocytic cells derived from resident microglia. Statistical analysis of the comparison between conventional stains and CD68 immunoreactivity for all disease categories shows a difference between the data sets (Kruskal-Wallis; p < 0.001). Among the PMA and PMA→ALS cases there were none showing higher scores for LFB/Marchi pathology compared with CD68. In PMA more than 80% of cases show no evidence of long tract pathology on LFB or Marchi staining whereas only 50% are unaffected in the CD68 stain. More than a third of PMA cases have moderate or severe CD68 scores despite the absence of clinically detectable UMN features. We hypothesize that this discordance between pathology and clinical observation reflects the difficulty in detecting signs of UMN involvement in the presence of severe LMN degeneration and weakness.
Among the ALS cases, 87% reached a moderate or severe CD68 score (mostly severe) compared with <75% (mostly moderate) using the conventional stains. This lesser sensitivity of the LFB and Marchi methods in detecting CST degeneration included two cases in which CST myelin changes were not detected but showed increased CD68 immunoreactivity. However, 12% of patients with ALS had marked myelin pallor, but only mild CD68 staining. These patients may reflect decreasing expression of CD68 in cases with longstanding or burnt out descending tract degeneration.
Dorsal column changes.
The presence of dorsal column (DC) changes was also assessed in these cases. In contrast to the CST changes, the presence of myelin pallor in the fasciculus gracilis in the LFB stain was the most sensitive marker of DC pathology. DC changes were present in 8% of pure PMA cases, 30% of those PMA cases that progressed to ALS, and 47% of cases presenting as ALS. The relative insensitivity of CD68 staining and the Marchi method may reflect the mildness of this pathology (few LFB scores were greater than the mild category) so that active degeneration of myelin is slight. No DC myelin changes were observed in the control cases.
Discussion.
The study highlights a subgroup of patients with ALS/MND with rapid progression of LMN degeneration (mean survival from diagnosis <2 years) who share the molecular pathology of ALS. We propose that the high relative frequency of LMN UBI in this group indicates a shared molecular pathology with typical ALS cases,13,16⇓ patients with ALS-dementia,17 and patients with frontotemporal dementia with MND type inclusions.18,19⇓ The occurrence of UBI in spinal motor neurons in PMA has previously been reported in individual case reports.20,21⇓ This shared molecular pathology is the basis of our hypothesis that these disorders form a clinicopathologic continuum in which the clinical syndrome is dependent on individual anatomic susceptibility in the CNS to degeneration involving common underlying molecular mechanisms.10 The other neuronal inclusion said to be specific to ALS is the Bunina body.22 Whereas Bunina bodies may be present in up to 85% of ALS cases, they are not considered to be as useful as UBI in defining the molecular pathology of ALS.23
The presence of clinically undetected degeneration of the CST in half of the patients with PMA further strengthens the relationship between this disorder and ALS. We postulate that the lack of clinical signs of UMN dysfunction may reflect profound terminal amyotrophy. Similar findings have been reported in patients with the A4V and D101N mutations in the gene encoding SOD1.24,25⇓ Such patients develop familial ALS characterized by rapid progression with absence or paucity of UMN signs. At autopsy the A4V cases may show only mild CST degeneration but CST changes are prominent in the D101N cases reported. A4V cases are characterized by HCI in spinal motor neurons, which are immunoreactive for neurofilaments and are found in some other SOD1 FALS mutations.10,15,26,27⇓⇓⇓ An apparently sporadic case of PMA was shown to have HCI28 without CST pathology and we suggest that this case may represent undiagnosed SOD1 FALS. Further evidence for UMN involvement in PMA includes the finding of Bunina bodies and UBI in cortical Betz cells of six patients, although CST pallor was not detected using LFB staining.29 These six patients also had spinal Bunina bodies and UBI.
Much of the previous literature on PMA predates modern molecular pathology. Physicians have used the term PMA since the early nineteenth century. Duchenne gave an account of the disorder in 1849 in the case of Prosper Lecomte, a circus owner.30 In 1850, the typical features of PMA were described under the title atrophie musculaire progressive,31 and Charcot’s account32 predates his clinicopathologic description of ALS.33 Despite advances in understanding the pathogenesis of ALS/MND there is still uncertainty regarding the classification of PMA within the spectrum of the MND, and its relationship to ALS. Cases of slowly progressive LMN illness corresponding to PMA have been regarded as a separate MND subtype34 that can be distinguished both from inherited early onset spinal muscular atrophy and from those cases of ALS presenting as PMA but showing rapid progression and evolution into typical ALS.35 Classic autopsy series of ALS include patients with a clinical diagnosis of PMA who had CST changes at autopsy.6,9⇓ However, neither of these series had access to ubiquitin immunocytochemistry.
PMA is therefore usually described as a disease characterized by slow progression and long duration with survival from onset ranging from 43 to 407 months (mean 159.2 months).35 Lifetables of PMA show a survival rate of 61.3% and 56.4% at 3 and 5 years.36 Onset and distribution of symptoms in PMA is generally described as an asymmetric distal upper limb weakness preceding proximal upper limb and lower limb involvement. Bulbar involvement is often mild and occurs late in the disease.34,36⇓ Two of our 10 patients accepted as having the PMA variant of ALS/MND presented with bulbar symptoms.
The definition of PMA and criteria for excluding UMN pathology in PMA remain major issues in clinical diagnosis. It has been suggested that the preservation of tendon reflexes in an otherwise weak, atrophic, and fasciculating limb is abnormal and represents a pathologically brisk UMN sign.4 This is a logical and reasonable view but it has not been possible to incorporate it into the diagnostic criteria for research studies and clinical trials.1,2⇓ This is because the significance of a preserved reflex must be judged against the progression of amyotrophy in the affected limb. This is a clinical judgment that is made by the attending physician, without verified published criteria on which to base the decision. Our patients who were assigned a diagnosis of PMA early in the disease did not show pathologically brisk reflexes in either their affected or unaffected limbs at that stage, in the judgment of an experienced ALS physician (P.J.S.). Six patients were reassigned to ALS as the disease progressed, taking into account their tendon reflexes together with other UMN features. We have not given detailed serial information on the state of tendon reflexes, but this study and the previously published work on CST changes in PMA and PMA/FALS justify a future prospective evaluation of UMN features in PMA. Such a study should include longitudinal evaluation of tendon reflexes compared with muscle strength, newer MRI modalities such as diffusion tensor MR imaging of the CST, and eventual postmortem neuropathology.
The current series illustrates that there is a subgroup of sporadic PMA cases with rapid progression (i.e., to death from respiratory failure in <2 years), UBI pathology, and a high prevalence of clinically unsuspected CST pathology at autopsy, in addition to cases developing into ALS from a PMA presentation. These findings are consistent with a clinicopathologic continuum of frequency and severity of CST pathology from PMA cases (mild CST pathology in ∼50%) through to ALS (severe and invariant CST pathology). Case reports in the literature support this interpretation.20,21,37⇓⇓ Pathologic evidence of Betz cell loss and CST degeneration was demonstrated in a patient diagnosed with sporadic spinal PMA who survived for 19 years,37 but UBI were absent. This case report is difficult to evaluate because the sampling strategy on which the lack of UBI was based is not defined. Cases with very profound loss of spinal motor neurons are the most difficult in which to demonstrate motor neuron inclusions and may require multiple serial sections from several blocks. The authors suggested that their case was one of ALS of long duration mimicking spinal PMA. However, they give no clinicopathologic definition of the syndrome of spinal PMA or literature references to other examples. Two cases of PMA with long duration in whom UBI were identified at autopsy have been reported.20 These cases illustrate that long duration alone in PMA is not a reliable predictor of the absence of UBI pathology.
A range of prevalence of PMA in MND populations has been reported: 1%,38 2.4%,39 9%,40 and 19%.41 This variation likely reflects different criteria and interpretations of the diagnosis. The relative frequency of cases of PMA and PMA→ALS in the current autopsy cohort (22%) does not represent prevalence in a community sample and may reflect bias in case referral for autopsy. Among these patients, 75% either developed UMN signs in life or had evidence of CST degeneration at autopsy. These patients present compelling evidence of disease that is essentially the same as ALS, including evidence for the presence of spinal motor neuron UBI.
Acknowledgments
Supported by the Medical Research Council, Wellcome Trust, Motor Neurone Disease Association, Action Research, and the ALS Association (USA).
Footnotes
-
See also page 1258
- Received September 9, 2002.
- Accepted January 17, 2003.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Hastening the Diagnosis of Amyotrophic Lateral Sclerosis
Dr. Brian Callaghan and Dr. Kellen Quigg