close

Вход

Забыли?

вход по аккаунту

?

Demyelination induced by serum form patients with Guillain-Barr syndrome.

код для вставкиСкачать
Demyelination Induced by Serum from
Patients with Gdain-Barr6 Syndrome
B. M. Harrison, PhD, L. A. Hansen, BA, J. D. Pollard, FRACP, and J. G. McLeod, FRACP
Sera from 16 patients with acute Guillain-Barre syndrome (GBS)and 14 healthy control subjects were injected into rat
sciatic nerve and assessed for demyelinating activity by electrophysiological and histological techniques. Only fresh
GBS serum, and not GBS serum stored at - 20°C or - 70°C, blocked conduction to a significantly greater extent than
did fresh control serum. Conduction block developed gradually, starting within 24 hours of injection and reaching a
maximum between days 3 and 6. Recovery of conduction commenced thereafter, and conduction returned to normal by
day 33. Quantitative histological studies on day 6 showed that fresh GBS serum produced significantly more widespread demyelination t%andid stored GBS serum (p < 0.01). Stored GBS serum showed residual demyelinating activity
when compared with fresh control serum (p < 0.01). Fresh serum obtained from 4 patients after recovery from GBS did
not produce conduction block, despite it having done so during the acute phase of the disease.
Harrison BM, Hansen LA, Pollard JD, McLeod JG: Demyelination induced by serum from patients with
Guillain-Barre syndrome. Ann Neurol 15: 163-170, 1984
The pathological hallmark of Guillain-BarrC syndrome
(GBS) is perivenous demyelination of peripheral nerve
in association with infiltration of lymphocytes and macrophages [I). Most pathological studies have ascribed
myelin damage to the activity of the inflammatory
infiltrate 11, 12-14}. A possible role for humoral demyelinating factors has been suggested by several lines
of evidence reviewed recently by Cook and Dowling
C3); increased titers of antineural antibodies have been
reported in GBS, and some studies, utilizing myelinated cultures of peripheral nerve, have demonstrated in
vitro demyelination with GBS sera. More recently,
serum from animals with experimental allergic neuritis
has been shown to produce demyelination when injected directly into normal peripheral nerve ClS).
Pathological studies by Feasby and colleagues { S , 61
and Saida and associates { 161 have demonstrated that
the intraneural injection of GBS serum into rat sciatic
nerve produces demyelination greater than that produced by normal human serum. The results of Tandon
and co-workers 121) and Pollard and associates Ell),
however, do not support this view. In an electrophysiological study by Sumner and colleagues 1191, the injection of GBS serum into rat sciatic nerve resulted in a
significant degree of conduction block when compared
with control serum. Low and colleagues 19, 10) were
unable to confirm this finding. The present study was
undertaken in an attempt to resolve the conflicting evidence concerning the demyelinating effect of GBS
serum in this in vivo experimental system.
From the Department of Medicine, University of Sydney, Sydney,
New South Wales 2006, Australia.
Materials and Methods
Male Wistar rats weighing 300 to 400 gm were anesthetized
with intraperitoneal pentobarbitone sodium, 35 mg per kilogram of body weight. The sciatic nerve was exposed through
a skin incision placed laterally in the thigh from the sciatic
notch to the popliteal fossa. A micrometer syringe fitted with
a 30-gauge needle was used to inject 50 ~1 of serum subperineurially into the sciatic nerve at a point 1 cm proximal to
its division into peroneal and tibia1 branches. All injections
were performed with the aid of an operating microscope.
Serum was obtained from 16 patients with acute GBS during the first month of illness, and from 4 of these patients 6 to
15 months later. All patients satisfied the criteria for GBS
established by an ad hoc committee of the National Institute
of Neurological and Communicative Disorders and Stroke
12). GBS serum was injected either fresh, within 4 hours of
collection (13 sera), or after storage at - 70°C (11 sera) or
-20°C (6 sera). Storage time ranged from 3 to 14 months,
with a mean of 8 months. Serum was also obtained from 14
healthy control subjects. Serum was drawn from 4 of these
normal donors on two or more occasions. All control sera
were injected fresh, within 4 hours of collection.Each rat was
injected with GBS serum in one sciatic nerve and with normal human serum in the other. At least two and frequently
three rats were injected with each patient's serum.
Electrapbysiological Studies
The compound muscle action potential was recorded from
the small muscles on the plantar surface of the rat foot by
means of a pair of steel needle recording electrodes as described by Sumner and colleagues {20). The active electrode
was placed just beneath the skin on the dorsum of the foot,
Received May 9, 1983, and in revised form June 21. Accepted for
publication June 21, 1983.
Address reprint requests to Dr Harrison.
163
and the remote electrode inserted into the most lateral toe.
The sciatic nerve was stimulated through two pairs of 30gauge steel needle electrodes inserted alongside the nerve
at the sciatic notch and ankle. Square-wave pulses 0.05 ms
in duration were delivered through an isolated voltage
stimulator (Devices type MKlV) triggered by a Digitimer
(Devices Sales Ltd) at l-second intervals. Care was taken to
deliver supramaximal stimuli. The muscle action potential
was amplified and displayed on one beam of a Tektronics
storage cathode ray oscilloscope, and a timing trace generated
by the Digitimer was displayed on a second beam. Permanent
photographic recordings were made of superimposed images
from the oscilloscope screen. The latency of the muscle action potential was measured from stimulus artifact to the first
negative deflection, and the amplitude was measured from
baseline to peak. Measurements were made on all animals
prior to operation, at 2-day intervals from postoperative days
1 or 2 during the first week, and then once weekly for up to 5
weeks. During all experiments the temperature of the animal's hind limb was maintained at 37°C +- 0.2"C by means of
a heat lamp and a thermistor probe inserted into leg muscles.
For each study, the latency and amplitude of the evoked
muscle action potential elicited by stimulation at both proximal and distal sires were recorded. The distance between the
two sites of stimulation was measured along the skin surface
with the leg fully extended, and the conduction velocity calculated. Conduction block was manifested by the occurrence
of a reduction in the amplitude of the evoked muscle potential on stimulation proximal to the injection site while the
amplitude of the evoked response on distal stimulation at the
ankle remained normal. The ratio of the amplitude of the response evoked by sciatic notch stimulation divided by the
amplitude of the response evoked by ankle stimulation was
used to compare the extent of conduction block from animal
to animal 119, 20). This amplitude ratio was calculated for
each nerve studied.
Histological Studies
Serum taken from 10 patients with GBS during the acute
phase of the disease and within 4 weeks of onset of clinical
symptoms was injected fresh or after storage into rat sciatic
nerves by the technique outlined. Forty microliters of test
serum was mixed with 10 p1 of fresh guinea pig serum as a
source of complement. Both sciatic nerves of each rat were
injected, one with GBS serum and the other with fresh
serum from 1 healthy individual or with fresh serum pooled
from 4 to 6 healthy subjects. Six days after intraneural injection the rats were anesthetized with an overdose of sodium
pentobarbitone and perfused through the abdominal aorta
with 2.5% paraformaldehyde and 4% glutaraldehyde in
phosphate buffer. The injected sciatic nerves were removed,
and each nerve was cut carefully into four segments each
approximately 5 mm long: two proximal to, one at, and one
distal to the site of injection. Nerve segments were postfixed
in Dalton's chrome-osmium solution and then dehydrated in
graded concentrations of ethyl alcohol and embedded in
Spurr's resin. Cross sections I p m thick were stained with
toluidine blue and examined by light microscopy. Ultrathin
sections were mounted on copper mesh grids, double stained
with lead and umnyl acetate, and examined with a Philips 400
electron microscope.
164 Annals of Neurology
Vol 15 N o 2
February 1984
Toluidine blue-stained 1 (*.m sections from the proximal
segments of each sciatic nerve were examined by light microscopy. The numbers of demyelinated fibers were counted
in the segment of nerve with the most pronounced abnormalities. The entire cross section of each nerve, which usually
consisted of a single fascicle, was examined. All tissue blocks
were coded, and the code was not broken until after microscopic examination of sections. Counts obtained from nerves
injected with the same serum were averaged, and the results
from nerves injected with fresh GBS serum, stored GBS
serum, and control serum were compared statistically using
the Mann-Whirney test. Results are expressed as means
standard deviations.
*
Results
Electmphysiological Findings
Nerves injected with fresh normal human serum had
an initial mean amplitude ratio of 0.80 ? 0.07 and
showed little change by day 1. The mean amplitude
ratio dropped to 0.62 k 0.01 on day 4 and returned to
0.78 t 0.08 by day 22 (Fig 1).
Fresh serum from 9 patients with acute GBS consistently produced conduction block when injected intraneurally (Fig 2; see Fig 1). The mean amplitude ratio
before injection was 0.82
0.06. Conduction block
developed gradually from day 1 and was most severe
on day 5, when the mean amplitude ratio fell to 0.36 -+
0.07. The results differed significantly from those in
controls on days 1, 3, 5, and 6 Cp 5 0.01). Improvement commenced between days 7 and 10. There
was no significant difference in mean amplitude ratio
between control and GBS serum-injected nerves by
week 2 (days 14 to 20 postinjection). The mean amplitude ratio in GBS serum-injected nerves returned
to 0.85 +- 0.01 by day 33.
The ability of acute GBS serum to cause conduction
block was affected by storage. Whereas fresh GBS
serum produced severe conduction block, changes
caused by acute GBS serum stored at -70°C were
much smaller. At no time was there a significant difference in mean amplitude ratio between nerves injected
with GBS serum stored at - 70°C and nerves injected
with normal human serum (Fig 3). Four of the six GBS
sera stored at -7O"C, when injected fresh, had produced conduction block. The results from injections of
acute GBS serum stored at - 20°C were indistinguishable from results obtained from normal human serum
injections. One of these serum samples, when injected
fresh, had produced conduction block.
The demyelinating effect of GBS serum was specific
to the acute phase of the disease. Serum taken from 4
patients 6 to 15 months after the onset of GBS produced no conduction block when the serum was injected fresh. There was no significant difference in results with serum from patients after recovery from
GBS and serum from controls at any time (Fig 4).
Serum taken from 2 of these patients 4 to 6 weeks after
*
I
I
23
I
I
I
I
-P
I
N
ro
5
J
E
T
I
0
N
Fresh GBS Serum 9 Patients
o
I
'
initial
study
,';
1
0
I
5
I
,
10
Fresh Normal Serum 14 Donors
I
15
20
25
30
35
40
Days Post - Injection
Fig I . Mean amplitude ratios from 9 fresh acute GBS Jera and
14 fresh normalsera. Numbers indicate the total number of
nerves studied on each day for GBS serum and for normalserum.
(Vertical bars = 1 standard deviation.)
L
onset had produced mild conduction block when compared with acute phase serum.
INITIAL
Histological Findings
Fresh acute GBS serum consistently produced more
extensive demyelination after intraneural injection into
rat sciatic nerve than did fresh serum from normal,
healthy subjects. The mean number of fibers demyelinated by fresh acute GBS serum was 110 ? 32,
compared with 29 ? 16 by fresh control serum. GBS
serum stored at - 70°C produced a mean of 5 5
20
demyelinated fibers, significantly less demyelination
than produced by fresh GBS serum (p < 0.01) but
more than produced by fresh control serum (p < 0.01)
(Fig 5 ) .
Similar quantitative morphological changes were
seen in all serum-injected nerves examined on day 6.
The segment of nerve proximal to the injection zone
was characterized by edema, causing increased spacing
between nerve fibers. Cellular infiltration occurred
around small blood vessels and beneath the perineurium but was never extensive. In nerves injected
with GBS serum, it was particularly noticeable that
demyelinated axons were distributed in groups either
along channels within the nerve fascicles or subperineurially or, occasionally, around small intrafascicular capillary blood vessels (Fig 6). Around some axons
myelin changes, such as the vesicular disruption of
*
DAY
A
N
STUDY
3
I
/I
r,
DAY
S
28
Fig 2. Serial electrical recordingsfrom rat sciatic nerve injected
with fresh acute GBS serum. Evoked muscle potentiaisfrom
stimulation at the sciatic notch (SN) and at the ankle (A) are
shown for the initialstudy, day 3, and duy 28. On the third day
after injection, the ankle response remained normal, whereas sciatic notch stimulation proximal t o the injection site demonstrated
partial conduction block. By day 28 the sciatic evoked response returned to normal. (Vertical bar = 5 m v ; time scale = 0.1 ms.)
Harrison et al: Demyelination with GBS Serum
165
0 Fresh Normal Serum
Fresh GBS Serum
-70' C stored GBS Serum
-20
'C
stored GBS Serum
initial
study
Fig 3. Mean amplitude ratios for fresh nornlalserum (14 donors),fresh acute GBS serum 19 patients), GBS serum .stored at
- 70°C (6 patientsi, and GBS serum stored at -20°C (6 patients). Results from fresh acute GBS serum are sign$cantly d;ffirent ip 5 0.01, Mann-Whitney)from controlfindings on &ys
day 1
day 6
day 3
week 2
Days Post - Injection
1 . 3, and 6 (indicated b.y asterisk). No other resultj. shou~riare
sig nificanti) different from con troI 1;ndings. Week 2 include^.
data from day 14 through day 20. (Shaded bars = meun amplitude ratio for each group; vertical lines = 1 standard deviation.)
0 Fresh Normal Serum
Fresh Acute GBS Serum
Fresh Recovery GBS Serum
1
1
initial
study
Fig 4.Mean amplitude ratios for fresh normalserum 114 donorsi, fresh acute GBS serum 19 patientJ), andfresh serum from
patients after recovery from GBS (4patients). Serum from rerovered patients ma taken 6 t o 15 months afrer disease onset.
Only fresh acute GBS serum results were significantly dif/erent
166 Annals of Neurology
Vol 15 No 2
February 1084
day 1
day 3
day 6-7
week 2
Days Post - Injection
ip S 0.01, Mann-Whitneji test)from contm(findings on days 1 ,
3 , and 6 (asterisks). Week 2 includes data from day 14 through
day 20. (Shaded bars = mean amplitude ratio for each group;
vertical lines = 1 standard daiution.)
0
**
a
*
*
.
*m
* *
I
0
.
c
Fresh
Stored
GBS
GBS
(-70'Cl
Fresh
NHS
Fig 5 . The number of demyelinated axons counted in nerves injected with fresh GBS serum. GBS serum stored at - 70"C,
and fresh normal human serum. Each plot represents the average
count of demyelinated axons in nerves injected with the same
serum. The horizontaI burs represent the mean number of demyelinated axons in each experimental group. (Shaded regions =
1 standard deuiati0n.i
F t g 6. Rat sciatic newe. Fresh acute GBS serum, 6 days postinjection. Demyelinated axons (arrows) are scattered in small
groups thronghout this region ofthe nerve fascicle. (Toluidine
bluestain; bar = 20 Fm.)
167
lamellae, were seen in the absence of any infiltrating
cells in the region. However, most demyelinated fibers
were surrounded by phagocytic cell processes containing myelin debris (Fig 7). Schwann cells generally appeared normal, but occasionally changes associated
with cellular degeneration, such as edematous and
vesiculated cytoplasm and dilated rough endoplasmic
reticulum, were observed within these cells. These
Schwann cell changes were no more prominent in
nerves injected with GBS serum than in those injected
with control serum. Schwann cell processes of unmyelinated fibers were only rarely involved, although in
some regions retraction of Schwann cell processes surrounding unmyelinated fibers resulted in the axolemma of adjacent fibers coming into contact. This
change was seen in all serum-injected nerves. Small
numbers of fibers undergoing axonal degeneration
were found, mainly in groups at the site of injection
along and close to the region of the needle track.
Discussion
When GBS serum that has been collected during the
active stage of the disease is injected directly into rat
sciatic nerve, it produces conduction block and demyelination. However, serum taken during recovery from
the disease does not block conduction to a significantly
168 Annals of Neurology
Vol 15 No 2
February 1984
Fig 7. Rat sciatic nerve. Fresh acute GBS serum, 6 days gostinjection. A phagorytic cell containing myelin debris lies within a
Schwann cell basal lamina (arrows), adjacent to a demyelinated
axon (Ax). l B m = 2 pm.)
greater extent than does serum from healthy subjects.
These results indicate that demyelinating factors are
present in the serum of patients with GBS, specifically
during the acute phase of the disease. The activity of
these factors is markedly reduced after the serum is
stored for comparatively short periods of time, even at
- 70°C.
In this study and those of Low and colleagues IS.
lo], there was both electrophysiological and histologi-.
cal evidence of demyelination in nerves injected with
serum from healthy subjects. We have noted (unpublished observations, 1982), as have Dyck and coworkers [4], that this effect is greater than that in
nerves injected with physiological saline. The demyelination produced by normal serum, therefore, cannot
result solely from the trauma of intramural injection.
Low and colleagues [9, 101 discussed possible mechanisms by which normal serum could result in demyelination, and raised the possibility that myelinotoxic substances, such as lysolecithin, are present in serum.
Lysolecithin, which has no effect when given systemi-
cally, produces severe demyelination when applied in
minute quantities intraneurally [7), presumably because it then circumvents the blood-nerve barrier, If, as
seems likely, a primary event in GBS is a defect in the
blood-nerve barrier, then some of the pathological
changes seen in GBS lesions may be accounted for by
the leakage into nerve of nomal serum components,
which are usually denied access to nerve by an intact
blood-nerve barrier. The results of this study, however,
confirm the findings of others [ S , 6, 19) that fresh GBS
serum is more myelinotoxic than is normal serum.
The present work provides an explanation for the
contradictory results of Low and colleagues 19, 10) and
Sumner and associates {19}. The results in our electrophysiological study of injection of stored serum are in
agreement with the findings of Low and colleagues (9,
lo), who also used stored serum, and of a preliminary
study from this laboratory in which stored serum was
used (11). In none of these studies did stored GBS
serum cause significant conduction block. We were
able to demonstrate this effect only with fresh acute
GBS serum. Sumner and colleagues reported conduction block [19) and used both fresh and stored serum
(A. Sumner, personal communication, 1982).
Saida and collaborators [ l G ) mainly used stored
serum (the sera of only 2 of 27 patients were injected
fresh) to assess histologically the in vivo demyelinating
activity of acute GBS serum. In both their study and
our own, stored serum was found to produce more
extensive demyelination after intraneural injection
than was control serum. The number of demyelinated
fibers produced by both normal human serum and
stored GBS serum, however, was considerably greater
in our study. This may be the result of a combination of
factors, such as slight variations in the intraneural injection technique practiced in the two laboratories and the
discrepancy in the time of assessment of demyelination
(2 days after intraneural injection in the study of Saida
and colleagues and 6 days after injection in our study,
when conduction block using fresh GBS serum was
found to be maximal).
There is good internal consistency in our study between the results obtained electrophysiologically and
those obtained histologically; both techniques demonstrate the reduced demyelinating activity of GBS
serum after storage. For the electrophysiological studies stored serum was injected intraneurally without the
addition of extra complement, whereas for the histological study, fresh guinea pig serum was added to
stored GBS serum (in the ratio 1 : 4 ) for that purpose.
It therefore appears most unlikely that the failure of
stored GBS serum to produce significant conduction
block resulted from a lack of complement necessary to
mediate the activity of serum demyelinating factors. A
much more reasonable explanation is that there is degradation of the myelinotoxic factors themselves on
storage and that the activity of these factors is not restored by the addition of complement. Feasby and coworkers [b} showed that the demyelinating activity of
GBS serum is not restored by the addition of complement after heat inactivation of serum.
Antigalactocerebroside antibodies are strongly suspected to be prominent agents in the development of
demyelination resulting from the intraneural injection
of serum from animals with experimental allergic
neuritis, the experimental analog of GBS [ l 5 , 17, 20).
It remains to be determined whether antigalactocerebroside antibody titers are elevated in GBS. Recent
evidence indicates that anti-P2 antimyelin antibody titers are not elevated in GBS {8, 22). Nevertheless, the
possibility remains that the demyelinating factors in
GBS serum are antibodies with specificity against other
myelin components or against other constituents of peripheral nerve, such as Schwann cells. Molecules secreted by inflammatory infiltrates, such as proteinases
and lymphokines, must also be considered as possible
demyelinating agents.
Presented at the Third Annual Australian Neuroscience Meeting,
Melbourne, Feb 2-4, 1983, and printed in abstract form in Neuroscience Letters (suppl) 11:S46, 1983.
Supported by grants from the National Health and Medical Research
Council and the National Multiple Sclerosis Society of Australia.
The authors are grateful to Mrs V. Antidormi and Mrs M. Verheijden for their skilled technical assistance, and to the Electron Microscope Unit, University of Sydney, for its services.
References
1. Asbury AK, Arnason BG, Adams RD: The inflammatory lesion
in idiopathic polyneuritis: its role in pathogenesis. Medicine
(Baltimore) 48:173-215, 1969
2. Asbury AK, Arnason BGW, Karp HR, McFarlin DE: Criteria
for diagnosis of Guillain-Barre syndrome. Ann Neurol 3:565566, 1978
3. Cook SD, Dowling PC: The role of autoantibody and immune
complexes in the pathogenesis of Guillain-Barre syndrome. Ann
Neurol 9 (suppl):70-79, 1981
4. Dyck PJ, Lais AC, Sparks MF, Hansen SM: Technique assessment of demyelinating activity from endoneurially injected test
solutions and sera: demyelinating activity of HMSN 111 sera.
Neurology (NY) 31:154, 1981
5. Feasby TE, Hahn AF, Gilbert JJ: Passive transfer of demyelinating activity in Guillain-Barre polyneuropathy. Neurology (NY)
30:363, 1980
6. Feasby TE, Hahn AF, Gilbert JJ: Passive transfer studies in
Guillain-Barre polyneuropathy. Neurology (NY) 32: 11591167, 1982
7. Hall SM, Gregson NA: The in vivo and ultrastructural effects of
injection of lysophosphatidyl choline into myelinated peripheral
nerve fibres of the adult mouse. J Cell Sci 9769-789, 1971
8. Iqbal A, Oger JJ-F, Arnason BGW: Cell-mediated immunity in
idiopathic polyneuritis. Ann Neurol 9 (suppl):65-69, 1981
9. Low PA, Schmelzer JD, Dyck PJ: Results of endoneural injection of Guillain-Barre serum in Lewis rats. Mayo Clin Proc
57:360-364, 1982
10. Low PA, Schmelzer J, Dyck PJ, Kelly JJ: Endoneurial effects of
Harrison et al: Demyelination with GBS Serum
169
11.
12.
13.
14.
15.
16.
17.
sera from patients with acute inflammatory polyradiculoneuropathy: electrophysiologic studies on normal and demyelinated rat nerves. Neurology (NY) 32:720-724, 1982
Pollard JD, Harrison BM, Gatenby P: Serum-induced demyelination: an electrophysiological and histological study. Clin Exp
Neurol 18:70-80, i98l
Prineas JW: Acute idiopathic polyneuritis: an electron microscope study. Lab Invest 26:133-147, 1972
Prineas JW: Pathology of the Guillain-Barre syndrome. Ann
Neurol 9 (suppl):6-19, 1981
Prineas JW, McLeod JG: Chronic relapsing polyneuriris. J
Neurol Sci 27:427-458, 1976
Saida K, Saida T, Brown MJ, Silberberg DH: In vivo demyelination induced by intraneural injection of anti-galactocerebroside
serum. Am J Pathol 95:99-116, 1979
Saida T, Saida K, Lisak RP, et al: In vivo demyelinating activity
of sera from patients with Guillain-Barre syndrome. Ann
Neurol 11:69-75, 1982
Saida T, Saida K, Silberberg DH: Demyelinarion produced by
170 Annals of Neurology
Vol 15 No 2 February 1984
experimental allergic neuritis serum and anti-galactocerebroside
antiserum in CNS cultures. Acta Neuropathol (Bed) 48:19-25,
1979
18. SaidaT, Saida K, Silberberg DH, Brown MJ: Transfer of demyelinarion by incraneural injection of experimental allergic
neuritis serum. Nature 272:639-641, 1978
19. Sumner A, Said G, Idy I, Metral S: Syndrome de Guillain-Barre:
effets Clectrophysiologiques et morphologiques du serum humain introduit dans I'espace endoneural du nerf sciatique du rat:
r6sultacs prkliminaires. Rev Neurol (Paris) 138:17-24, 1982
20. Sumner A, Saida K, Saida T, et al: Acute conduction block
associated with experimental antiserum-mediated demyelination
of peripheral nerve. Ann Neurol 11:469-477, 1982
21. Tandon DS, Griffin JW, Drachman DB, et al: Studies on
the humoral mechanisms of inflammatory demyelinating
neuropathies. Neurology (NY) 30:362, 1980
22. Zweiman B, Rostami A, Lsak RP, et al: Immune reactions ro P2
protein in the human inflammatory demyelinative neuropathies.
Neurology (NY) 33:234-236, 1983
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 892 Кб
Теги
barry, forma, patients, induced, syndrome, serum, guillain, demyelination
1/--страниц
Пожаловаться на содержимое документа