Striatal efferent involvement and its correlation to levodopa efficacy in patients with multiple system atrophy

The basal ganglia and related structures play an important role in motor operations via structurally and functionally organized neural interconnections. ^[1-5] These connections form several closed circuits such as the nigrostriatal loop, which consists of nigrostriate fibers and striopallidal and strionigral projections. We previously applied immunohistochemical techniques to study the striatal efferent pathway in patients with parkinsonism, including idiopathic Parkinson's disease, parkinsonism-dementia complex on Guam, progressive supranuclear palsy, and striatonigral degeneration (SND) and found that the pattern of striatal efferent involvement is distinct in each disorder, even though the degeneration of nigrostriate projections is common to the four diseases. ^[6-11]

The distribution pattern of striatal efferent degeneration in SND is characteristic; the expression of markers such as met-enkephalin (MEnk), ^[6,7] substance P (SP), ^[6] calbindin-D28k (CaBP), ^[10] and calcineurin ^[6,12] being reduced in the dorsolateral striatum and the ventrolateral portions of both segments of the globus pallidus and the substantia nigra. Based on anatomic data, ^[1,2,13] these topographic alterations seem to represent changes of the projection fibers and their terminals originating from the putamen. Consequently, these findings indicate that in SND, the nigrostriatal loop could be disrupted not only at the nigrostriate connection, but also at the putaminal efferent pathway. As the putaminal projection neurons are the main effective sites of dopamine, we presented the idea that the poor response to levodopa by SND patients may be due to striatal efferent alterations. ^[7,14]

SND often shares clinicopathologic features with olivopontocerebellar atrophy (OPCA) and Shy-Drager syndrome (SDS); Graham and Oppenheimer ^[15] have proposed the comprehensive term "multiple system atrophy" (MSA) for these clinical entities. We have now extended our immunohistochemical studies by investigating the striatal efferent system in patients with histologically proven OPCA and SDS, and addressed the possible correlation between the clinical effect of levodopa and the striatal efferent alterations.

Materials and methods.

This study was carried out on brain tissue obtained at autopsy from five patients with clinicopathologically confirmed MSA and from five neurologically normal control subjects. Patients and control individuals were matched with respect to age. The clinical diagnoses of the MSA patients were OPCA in two, SND in two, and SDS in one. Their clinical features are summarized in Table 1. There was no family history of neurologic disease in any of the patients. Their clinical symptoms started in the fifth or sixth decade of life, and the cause of death was either pneumonia or sepsis after several years of medical management. Although the degree of severity varied from one patient to another, their three major symptoms were cerebellar, extrapyramidal, and autonomic dysfunctions. Without exception, cranial CT and/or MRI disclosed pontocerebellar atrophy.

Both OPCA patients gradually developed wide-based gait, followed by dysuria and impotence within a year. The neurologic examination revealed ataxic dysarthria, limb and truncal ataxia, brisk deep tendon reflexes, and hypotonic bladder. Thorough follow-up examination disclosed mild rigidity of the four limbs and resting tremor 8 years after onset of symptoms in OPCA case 1, and mild rigidity and propulsion with loss of postural reflexes 5 years after onset in OPCA case 2. No beneficial effect on rigidity and tremor was observed following administration of a sufficient dose of levodopa in case 1; his muscular tone increased progressively, and the patient was confined to bed within the following year. By comparison, the extrapyramidal signs of OPCA case 2 were distinctly ameliorated by levodopa; there was no demonstrable rigidity or pulsions, and the effect persisted until several months before death.

The initial symptom of both SND patients was bradykinesia. SND case 1 was diagnosed at another institution 5 months later as having Parkinson's disease, and treated with levodopa, but without any improvement. Within the following years she developed finger tremor in action, monotonous speech, and shuffling gait. Neurologically, moderate bradykinesia and mild rigidity were evident. There was no resting tremor. She was barely able to walk due to marked postural instability. Besides these extrapyramidal signs, there was evidence of generalized hyperreflexia with extensor plantar responses and slight limb ataxia. Combined treatment with appropriate doses of levodopa and dopamine agonist was found to be ineffective. The patient was confined to bed 5 years after onset, when dysuria and postprandial hypotension were noted. By contrast, SND case 2 gradually manifested finger tremor while writing and shuffling gait with marked retropulsion. The neurologic examination disclosed definite parkinsonian features (moderate degree of bradykinesia, generalized rigidity and gait disturbance, no resting tremor) accompanied by mild limb ataxia, orthostatic hypotension, urinary incontinence, and exaggerated tendon reflexes. His bradykinesia and gait disturbance were moderately ameliorated by levodopa, and the patient was ambulatory until 1 month before death.

The initial symptom of the SDS patient was urinary incontinence. A neurologic assessment carried out 3 years later revealed hypotonic bladder and orthostatic hypotension only. During the following 2 years she developed gait unsteadiness and difficulty of fine finger movements. Neurologically, moderate bradykinesia and mild rigidity, limb ataxia, and wide-based gait were demonstrated, and levodopa treatment was instituted. However, the bradykinesia and rigidity worsened, the deep tendon reflexes gradually became hyperactive, and extensor plantar responses developed. She was bedridden for 3 years until she died of sepsis.

Brain tissue was fixed for several weeks in 10% neutral formalin, sliced coronally, and embedded in paraffin. Conventional neuropathologic examination of all patients revealed pronounced neuronal loss and gliosis in the putamen (the dorsolateral portion was more severely affected), substantia nigra, inferior olives, pontine nuclei, cerebellar Purkinje cells, and intermediolateral cells of the spinal cord. However, neuronal loss in the putamen of OPCA case 2, in the Purkinje cells of SND case 2, and in the inferior olives of the two SND patients was mild. The caudate nucleus was less affected than the putamen in all five MSA patients.

The immunohistochemical studies were performed as described before ^[10,11] on 6-micro meter thick sections from the basal ganglia at the levels of nucleus accumbens and of mammillary body and from the midbrain through the level of the superior colliculus and caudal red nucleus. The primary antibodies used were rabbit antisera raised against MEnk (Chemicon International Inc., Temecula, CA, USA, diluted 1:2,000 in phosphate-buffered saline [pH 7.2] containing 3% bovine serum albumin [PBS-BSA]), against SP (Cambridge Research Biochemicals Ltd., Cambridge, UK, diluted 1:1,500 in PBS-BSA), against CaBP (kind gift from Dr. M.R. Celio, ^[16] diluted 1:1,500 in PBS-BSA) and against tyrosine hydroxylase (Chemicon International Inc., diluted 1:300 in PBS-BSA). Incubation was carried out overnight at 4 degrees C. Visualization of bound antibodies was done with the appropriate Vectastain Elite ABC Kit for rabbit IgG (Vector Laboratories, Burlingame, CA, USA) following the manufacturer's protocol; 3,3 prime-diaminobenzidine tetrahydrochloride was the final chromogen. Staining specificity was assessed by replacing the primary antisera with appropriate amounts of non-immune rabbit serum. No product deposits were seen in sections thus treated.

The total number of dopaminergic neurons of the substantia nigra represented the mean number of neurons containing a nucleus and tyrosine hydroxylase immunore-activity in their perikarya in two sections. The substantia nigra was equally divided into medial and lateral portions.

Results.

The immunohistochemical staining patterns of the brain specimens of normal controls were qualitatively similar. In the striatum, an immunohistochemically identifiable striosomal organization was evident throughout the caudate nucleus and the putamen of the controls with the antibodies to MEnk Figure 1a and CaBP. By contrast, topographically depleted immunoreactivity with these antibodies and disorganized striosomal arrangements were seen in the striatum of all MSA patients; the alterations were most pronounced in the dorsolateral portions Figure 1b. The extent of the affected area differed somewhat from patient to patient. The reaction product deposits appeared to be decreased in almost the entire putamen of OPCA case 1 and of both SND patients. A reduction was also evident in the dorsal half of the putamen of the SDS patient, and in the most dorsal portion of that of OPCA case 2. As noted previously, ^[11,17,18] it was difficult to detect SP-positive strio-somes in the 6-micro meter sections of brain specimens of control subjects and MSA patients.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Striatum of a normal individual (a) and of the patient with Shy-Drager syndrome (SDS) (b) cut at the nucleus accumbens level. Immunostaining with anti-met-enkephalin (MEnk) antibody. (a) The striosomal organization is identifiable (arrows) throughout the caudate and the putamen of the control subject (times 3). (b) There is significant reduction in MEnk immunoreactivity at the dorsal portion of the putamen (arrowhead) with the striosomes being preserved in the caudate (arrows) of the SDS patient (times 3).

Strong SP-immunoreactivity was observed throughout the internal segment of the globus pallidus (GPi) of the control individual Figure 2a. By comparison, topographic depletion of SP-immunoreactivity was invariably evident in the ventrolateral GPi portion of all five MSA patients Figure 2b, c, d. Similarly, the entire external segment of the globus pallidus (GPe) from the controls was intensely immunolabeled with the anti-MEnk antibody Figure 3a. By contrast, decreased MEnk expression was evident in the lateral GPe of the MSA patients, with the topographic distribution patterns of the affected regions appearing essentially similar in OPCA Figure 3b, SND Figure 3c, and SDS Figure 3d. A relative involvement between SP- and MEnk-positive striatal efferents was not evident in any of the cases examined. Significant immunostaining with the anti-CaBP antibody was seen in both segments of the globus pallidus of all cases examined, but as documented before, ^[10] evaluation of pathologic changes was difficult with this reagent.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. The internal segment of the globus pallidus (GPi) of a normal subject (a) and of patients with multiple system atrophy (MSA) (b-d). Immunostaining for substance P (SP). (a) Strong immunoreactivity is seen throughout the GPi of the control subject (times 7). (b) Deposits of immunoreaction products are markedly decreased, except at the most dorsomedial GPi portion (arrowheads) of a patient with OPCA (times 10). (c) Topographic depletion of immunoreactivity in the ventrolateral GPi is evident in an SND patient (times 8). (d) A significant reduction of immunoreactivity is seen in the ventrolateral GP (arrows) of the SDS patient (times 8).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. External segment of the globus pallidus (GPe) of a normal subject (a) and of MSA patients (b-d). Immunostaining with anti-MEnk antibody. (a) Immunoreaction product deposits are distributed throughout the GPe of the control individual (times 7). (b-d) Similar immunoreactivity distribution patterns with significant depletion at the lateral GPe portion are seen in patients with OPCA (b), SND (c), and SDS (d). (b, times 9; c, times 8; d, times 7.5).

SP-reaction product deposits were observed throughout the substantia nigra, ^[8,19] and CaBP immunoreactivity was evident in the pars reticulata, ^[10] of the control subjects. By comparison, the staining patterns observed in the substantia nigra of the MSA patients, except OPCA case 2, were similar. There was marked reduction in immunoreactivity with the anti-SP and the anti-CaBP antibodies, with the reduction being most pronounced in the ventrolateral portion Figure 4a, b. In OPCA case 2, the SP immunoreaction products appeared to be slightly decreased in the ventrolateral substantia nigra Figure 4c. However, the involvement of the region was more obvious when anti-CaBP antibody was used Figure 4d. Since CaBP is a specific marker of the striatal matrix compartment and its efferents, ^[10,20] these findings on OPCA case 2, who had the least nigrostriatal pathology of the patients studied, may indicate that the projection neurons in the striatal matrix are indeed more susceptible than those in the striosomes. ^[21]

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Substantia nigra (SN) of MSA patients. (a) SP-immunoreactivity is markedly diminished in the SN of OPCA case 1, being most pronounced at the ventrolateral portion (arrows) (times 9). (b) The anti-SP antibody immunolabels only the medial SN (arrow) and a portion of the dorsal SN (arrowhead) of the SDS patient (times 8). (c) A slight reduction of immunoreactivity with the antibody to SP is seen at the ventrolateral portion of the SN (arrow) of OPCA case 2 (times 6.5). (d) An adjacent section to (c) stained with the antibody to CaBP. A topographic depletion of reaction products is readily identified in the ventrolateral SN (arrow) (times 6.5).

Under higher magnification, the reaction product deposits in the globus pallidus of normal controls were arranged in numerous "woolly fiber" patterns. These patterns were seen throughout the GPi with the anti-SP antibody, the GPe with the antibody to MEnk, and both segments with the anti-CaBP antiserum (data not shown). The immunopositive portions of each segment of the globus pallidus appeared similar in all MSA patients; abundant woolly fiber arrangements were visualized by the anti-SP Figure 5a and the anti-MEnk Figure 5b antibodies. The appearance and number of woolly fibers resembled those seen in the control subjects. On the other hand, in the areas where the SP- and MEnk-immunoreaction product deposits appeared to be reduced at the gross level, there was a significant difference between the MSA patients who had a response to levodopa and those who did not. Immunostaining of the specimen of the responsive OPCA case 2 with anti-SP and anti-MEnk antibodies revealed the presence of woolly fibers in the ventrolateral portions of the GPi and GPe, respectively Figure 5c, d, but their numbers and size were decreased. Similarly, albeit poorly organized, woolly fiber arrangements were also identified in SND case 2 Figure 5e, f. By contrast, in the remaining three patients whose parkinsonian symptoms had failed to respond to levodopa, there were almost no immunoreaction product deposits in the ventrolateral portions of the GPi and the lateral GPe upon incubation with the antibodies to SP and MEnk, respectively. These findings on the nonresponsive cases would suggest that these patients have severe topographic degeneration of the putaminal efferent terminals in the globus pallidus.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Globus pallidus of MSA patients immunostained for SP and for MEnk. (a, b) The dorsomedial portions of the GPi of OPCA case 2 stained for SP (a) and of the GPe stained for MEnk (b). Numerous woolly fiber arrangements can be seen. Their appearance is similar to those seen in the pallidum of control subjects. (c, d) The ventrolateral portions of SP- stained GPi (c) and MEnk-stained GPe (d) of OPCA case 2 have residual woolly fibers, but they are shorter and less numerous than in (a) and (b). (e, f) The ventrolateral portions of GPi (e) and GPe (f) of SND case 2, immunostained for SP and MEnk, respectively. As in OPCA case 2 (c, d), remaining woolly fiber arrangements can be seen in these areas (a-f, times 170).

The total number of dopaminergic neurons in the medial and lateral portions of the substantia nigra of the normal subjects was 206 +/- 16 and 211 +/- 25 (mean +/- SD), respectively. By comparison, a marked loss of these neurons was observed in all five MSA patients. The mean numbers of remaining dopaminergic neurons of the medial/lateral substantia nigra of each patient were quite similar. The respective numbers were: OPCA case 1: 72/32; OPCA case 2: 77/30; SND case 1: 67/34; SND case 2: 63/24; and SDS: 67/36. These results indicate that the extent of the surviving endogenous dopaminergic inputs would not be a determining factor for the efficacy of levodopa in the patients studied.

Discussion.

We demonstrated that topographic striatal efferent degeneration in the brains of OPCA and SDS patients is similar to that in SND. The three disorders present with any combination of cerebellar ataxia, parkinsonism, and autonomic failure of diverse severity during the clinical course. Irrespective of clinical form, neuropathologic investigations of the patients' brains usually disclose involvement of structures of the olivopontocerebellar, nigrostriatal, and autonomic systems. ^[22] Based on the overlapping clinical and histologic features, the term MSA ^[15] encompasses the three disorders. The glial cytoplasmic inclusions in MSA brains ^[23] strongly suggest that MSA may represent a distinct neurodegenerative entity. Our demonstration that characteristic striatal efferent degeneration patterns commonly underlie all clinical forms of MSA provides further evidence that the three clinical conditions could represent variable expressions of the same disorder.

Other than in SND, the typical topographic alterations only occur in adult-onset motor neuron disease with basophilic inclusions (MND/BIs), ^[19] but this disorder manifests itself clinically by upper and lower motor neuron signs and not by extrapyramidal or cerebellar dysfunction. Moreover, neuropathologic investigations revealed the widespread occurrence of basophilic intracytoplasmic neuronal inclusions instead of glial cytoplasmic inclusions. ^[24,25] Therefore, MND/BIs appears to be distinct from MSA.

Efficacy of levodopa for the parkinsonism in MSA patients has been inconsistent, with earlier observations being generally disappointing. ^[26-28] Lees ^[29] found poor, or only a transient, benefit of the extrapyramidal features in histologically proven MSA. Quinn ^[30] included non/poorly levodopa-responsive parkinsonism in his proposed diagnostic criteria for MSA. Although the majority of MSA patients do not have response to levodopa, several reports describe subsets of patients whose disease is improved by the drug, at least during the initial period of treatment. ^[29,31-37] Even though the pathologic basis for the heterogeneous response to levodopa remains speculative, some have hypothesized that the variability may be due to the relative degree of nigral versus striatal involvement. ^{[31,35,38,39]} Thus, a poor response may be attributable to the relative severity of striatal degeneration and the consequent loss of postsynaptic dopamine receptors, with a possible contribution from pallidal lesions. The underlying basis of this hypothesis was addressed by Fearnley and Lees, ^[33] who found a clear correlation between putaminal cell count and clinical efficacy of levodopa in MSA patients. However, in exceptional cases, a patient will have a response to levodopa despite severe putaminal degeneration. ^[22] Investigations on postsynaptic striatal dopamine receptors could be another approach for obtaining additional data on the inconsistent levodopa responsiveness in MSA patients. Unfortunately, there is only limited information on receptors in MSA patients, ^[34,40-44] with Brooks et al. ^[42] reporting results that provide no convincing relationship between clinical levodopa efficacy and the density of striatal D₂ receptors, and suggesting a contribution of the additional involvement of other basal ganglia connections.

Among the several functional circuits in the basal ganglia, ^[1-5] those essential for motor control are likely to be the "motor loop" of the corticosubcortical circuit, ^[3] and lateral nigrostriatal loop. ^[8,14] Loop disruption at any portion possibly results in clinically evident movement disorders. ^[4,5] Idiopathic Parkinson's disease would be a simple example where the clinical symptoms are mainly attributable to the disconnection of nigrostriatal fibers. On the other hand, as shown in this immunohistochemical study, the putaminal efferent system, along with the nigrostriatal projections, appears to be involved in MSA. The putaminal output pathway to the globus pallidus and substantia nigra is a common component of both the motor loop of the corticosubcortical circuit and the lateral nigrostriatal loop. Therefore, when the striatal efferents are totally damaged, the functional reconnection of nigrostriatal fibers with levodopa supplementation might not be sufficient to ameliorate the clinical symptoms.

Our results are in accord with the observations by Fearnley and Lees ^[33] in that the degree of alterations of the striatal projection system would correlate with levodopa efficacy in MSA patients. Moreover, since the immunohistochemical assays used in our study allow the visualization of striatal efferent fibers and their terminals, they provide direct evidence on the pathologic alterations of fiber connections between striatum and its target structures. These techniques are sensitive enough to detect surviving striatal efferents, even in patients whose loss of striatal cells is pronounced, as was the case of our SND patient 2. Our data do not contradict those of Brooks et al., ^[42] since the present investigation focused on the striatal projection terminals in the globus pallidus, which could degenerate whether the overall D₂ receptor density of the striatum is preserved or not. Furthermore, our results may also imply that postsynaptic alterations of striatal projections do occur. As the principal neurotransmitter of the striatal efferents is gamma-aminobutyric acid (GABA), which colocalizes with MEnk or SP, ^[45] it would be of interest to examine the globus pallidus and substantia nigra of MSA patients for the presence of receptors for GABA, ^[46] MEnk, and SP, particularly in view of levodopa responsiveness, and the potential pharmacologic reconnections of striatopallidal and striatonigral fibers to complete the motor-related basal ganglia circuits.

The present study provides evidence that topographic degeneration of striatal efferents underlies neuropathologic features that are common for OPCA, SND, and SDS, and that the degree of their alterations appears to correlate with clinical efficacy of levodopa for the extrapyramidal symptoms in certain MSA patients.

Acknowledgments

The authors thank Dr. M.R. Celio for generously providing the anti-CaBP antibody, Prof. A. Hirano of Montefiore Medical Center and Dr. Y. Makita of Kitano Hospital for their helpful comments and encouragement, and Dr. F. Herz for reviewing the manuscript. They are also grateful to Mr. S. Fukui, Mr. T. Hirai, and Ms. A. Asada for technical assistance.