вход по аккаунту


Coculture with autologous myotubes of cytotoxic T cells isolated from muscle in inflammatory myopathies.

код для вставкиСкачать
Coculture with Autologous Myotubes
of Cytotoxic T Cells Isolated from Muscle
in Inflammatory Myopathm
Reinhard Hohlfeld, MD, and Andrew G. Engel, M D
T-cell lines were expanded from muscle of 10 patients with polymyositis, 5 with inclusion body myositis, 5 with
dermatomyositis, and 5 with other muscle diseases. All cell lines uniformly expressed T-cell antigens, but not natural
killer cell or B-cell antigens. The proportion of helper (CD4+)and cytotoxic (CD8+)T cells in the expanded lines was
variable and showed no correlation with the diagnosis. Sixteen cell lines (6 polymyositis, 4 inclusion body myositis, 5
dermatomyositis, 1 other muscle disease) consisted predominantly of CD8+ T cells. None of these lines displayed
natural killer-like cytotoxicity but all were capable of lectin-dependent cytotoxicity. Three of 6 polymyositis, 1 of 4
inclusion body myositis, and 1 of 5 dermatomyositis lines showed low but statistically significant cytotoxiciry against
autologous myotubes (6 to 27% specific 51Crrelease; effector-target ratiu, 20: 1). The results demonstrate that functionally competent cytotoxic T cells can be expanded from muscle affected by inflammatory myopathies and are consistent
with the hypothesis that some cytotoxic T cells recognize an autoantigen on myotubes. Further studies of this experimental system may define the molecular mechanism of T cell-mediated muscle fiber injury and may help to identify
the relevant antigens.
Hohlfeld R, Engel AG. Coculture with autologous myotubes of cytotoxic T cells isolated from
muscle in inflammatory myopathies. Ann Neurol 1991;29:498-507
Immunohistochemical studies have provided evidence
for a T cell-mediated cytotoxic response against muscle fiber-associated antigens in polymyositis (PM) and
inclusion body myositis (IBM), and an antibody- or
immune complex-mediated response against a vascular component in dermatomyositis (DM) [l-8). The
antigens involved in the pathogenesis of PM, DM, and
IBM are not known. A number of investigators postulated that self-antigens on muscle are recognized by
cytotoxic cells in DM and PM, and supported this norion by reporting that peripheral blood cells of patients
with PM or DM had strong cytotoxic effects on c d tured xenogeneic C9-121 or embryonic human [9, 131
myotubes. However, some of these observations could
not be reproduced [14}, and the experiments were
conducted when very little was known about the mechanisms of antigen recognition by T cells.
T cells recognize an antigen associated with a polymorphic determinant of the self-major histocompatibility complex (MHC) [151. Therefore, the antigen
specificity of T cells can be assessed only in the presence of MHC-compatible (preferably autologous)
antigen-expressing cells. Further, the early studies used
poorly defined, blood-derived effector cells. In con-
trast, it is now possible to expand T cells from target
tissues [ l b , 17) and to ctiaracterize T-cell subpopulations precisely [18). Finally, the peripheral blood contains natural killer (NK) cells, that is, lymphocytes that
can kill certain target cells without prior sensitization
and without MHC restriction 1191. We recently
showed that human myotubes and myoblasts are highly
susceptible to killing by NK cells 1201. It follows that
evaluation of anrigen-specific, T cell-mediated killing
of myotubes must be under conditions that exclude
antigen-nonspecific NK-cell cytotoxicity.
The present study had two major objectives: (1) to
design an experimental system to characterize the interactions that occur in cocultures of human cytotoxic
T lymphocytes (CTLs) and myotubes, and (2) to critically reevaluate the question of whether or not T cells
from muscles of patients with different myopathies can
kill autofogow cultured myotubes.
From the Department of Neurology and Neuromuscular Research
Laboratory, hfayo Clinic and Mayo Foundation, Rochester, MN.
Address correspondence to Dr Engel, Neuromuscular Research Laboratory, Guggenheim Building G 801, Mayo Clinic, Rochester, MN
Received Aug 7, 1990, and in revised form Nov 13. Accepted for
publication Nov 15, 1990.
Materials and Methods
Reagents and Sappfies
Chick embryo extract, Dulbecco’s modified Eagle medium
(D-MEM), Dulbecco’s phosphate-buffered saline solution
(D-PBS), calcium and magnesium-free D-PBS, certified fetal
498 Copyright 0 1991 by the American Neurological Association
Table I . List of Monoclonal Antibodies (mAbs)
Ig Subclass
Antigen Distribution
WT3 1
Ti A
Anti-Leu-1l a
Anti-Leu-I l b
Becton Dickinson
Becton Dickinson
Becton Dickinson
Ortho Pharmaceutical
Becton Dickinson
Becton Dickinson
Dr Th. Hercend I291
T Cell Sciences
Becton Dickinson
Becton Dickinson
Becton Dickinson
Becton Dickinson
Becton Dickinson
IgG 1
IgG 1
IgG 1
TCR, chain
TCR, chain
HNK- 1
NKH- 1
CD1 l b
Becton Dickinson
All T cells and subset of N K cells
All T cells
Helperiinducer T cells
Helperiinducer T cells
Cytotoxic/suppressor T cells
Cytotoxic/suppressor T cells
Majority of T cells
Minority of T cells
Minority of T cells
Subset of N K cells and CTL
NK cells and neutrophils (Fc receptor for IgG)
NK cells and neutrophils (Fc receptor for 1gG)
N K cells and subset of T cells
NK cells, granulocytes, macrophages, suppressor T-cell subset
B cells
Wide tissue distribution
natural killer; CTLs = cytotoxic T lymphocytes.
calf serum (FCS), Ham’s F-10 medium, horse serum, ~-glutamine, and RPMI-1640 medium were obtained from Gibco
(Grand Island, NY). Bovine serum albumin (BSA), dimethylsulfoxide (DMSO), 0.02% ethylenediaminetetraacetic acid
(EDTAI-PBS, Histopaque-1077, 2% gelatin, penicillinstreptomycin, phytohemagglutinin (PHA-P), poly-r-lysine,
p-phenylenediamine, trypan blue, and 0.05 % trypsin-O.O2%
EDTA-PBS were from Sigma (St Louis, MO). Lectin-free
interleukin-2 (IL2; Lymphocult T-LF) was from Biotest
(Fairfield,NJ). Na,51Cr04was obtained from Amersham (Arlington Heights, IL). Rhodamine-labeled avidin-D, biotinylated horse anti-mouse IgG, biotinylated nonimmune mouse
IgG, and the avidin-biotin complex (ABC) immunoperoxidase kit was from Vector (Burlingame, CA). Rhodaminelabeled goat anti-mouse IgM was from Kirkegaard and Perry
(Gaithersburg, MD). Fluorescein isothiocyanate (F1TC)labeled nonimmune mouse IgG was from Coulter (Hialeah,
FL). The control antibodies for flow cytometry were from
Becton Dickinson (Mountain View, CA). Human AB + senun, prescreened baby rabbit complement, and human buffy
coat preparations were obtained from the Mayo Clinic Blood
Bank. K-562 cells were purchased from the American Type
Culture Collection (Rockville, MD). The plastic culture supplies were from Costar (Cambridge, MA). The sources of the
monoclonal antibodies (mAbs) are given in Table 1.
Muscle Specimens
Twenty-five muscle specimens were studied: 10 from patients with PM, 5 from patients with IBM, 5 from patients
with DM, and 5 from patients with other muscle diseases
(OM). The clinical and pathological criteria for the diagnosis
of PM, IBM, and DM have been described previously [2}.
The OM diseases included sarcoid myopathy (Patient O M l ) ,
denervation atrophy (OM2 and 4), necrotizing myopathy
(OM3), and a slight myopathy of undetermined cause
(OM5). Patients (and cell lines) in each group were coded
with a consecutive number (e.g., PM1 or DM5). The suffix
“-L”for each cell line code stands for lymphocytes, and the
suffix “ - M for myoblastsimyotubes.
Dissociation of Muscle
Muscle specimens, weighing 200 to 500 mg, were homogenized mechanically and trypsinited as described by Blau and
Webster 121). The resulting cell suspensions typically contained between lo4 and lo5 mononuclear cells, and small
muscle fragments sometimes associated with adhering or invading mononuclear cells. The cell suspensions were cultured
separately for expansion of myoblasts and of lymphoblasts,
as will be described. The experimental design is summarized
in Figure 1.
Myoblmt and Myotube Cultures
Myoblasts and myotubes were cultured as described previously [20-22).
T-cell Cultures
were designed to find the conditions required for the expansion of T-cell lines from muscle. The cultures were stimulated (A) with lectin-free IL-2 (10% Lymphocult T-LF) plus
autologous irradiated feeder cells, or (B) with 10 pg/ml of
PHA-P plus autologous or allogeneic irradiated feeder cells.
In most cases, procedure A was unsuccessful and therefore
procedure B was adopted. Peripheral blood mononuclear
cells (PBMCs), prepared by standard gradient centrifugation
1231 from freshly drawn autologous blood or from buffy coat
preparations, served as feeder cells. Immediately before use
in culture, feeder cells were irradiated with 6,000 rad at 4 19
radirnin from a 1 3 T Csource.
For initial stimulation (day 0), the cells were cultured in
round-bottom, 6.4-mm microwells in lymphocyte growth
medium (LGM; RPMI-1640 with 10% heat-inactivated hu-
Hohlfeld and Engel: Coculture of Cytotoxic T Cells and Myotubes
Myoblast culture
Stimulate with PHA and
inadiated feeder cells
Growth medium
(3-4 wk)
Fusion medium
(-7 days)
Coculture In autologous and
allogenelc comblnatlon
man A B serum, 2 mM L-glutamine, 100 unitsiml of penicillin, 100 pgiml of streptomycin), containing 4 x lo5irradiated feeder cells and 10% Lymphocult T-LF, or 10 pgiml of
PHA-P. On days 6 to 8, the microwell cultures were pooled,
resuspended in 1 ml of fresh LGM, and placed into a 16-mm
culture well. By days 10 to 14, the cultures contained vigorously proliferating lymphoblasts and muscle fiber debris. The
cultures were fed every 3 days with fresh LGM plus IL-2 for
about 20 days, when the lymphoblasts began to revert to
small lymphocytes. At this time, the cultures were replated
at 0.5 to 1.0 x lo5 lymphoblasts plus 2.5 x lo6 irradiated
autologous or allogeneic feeder cells per milliliter of LGM
containing 10 pgiml of PHA-P. The cells were expanded in
LGM plus IL-2 for 4 to 6 weeks so that approximately 10’
lymphoblasts were available for phenotypic and functional
analysis (cultures containing only irradiated feeder cells never
gave rise to lymphocyte lines).
(4-6 Wk)
----- 3
Clone by limiting dilution
T cells
Fig 1 , Experzmentuldeszgnof the study PHA = phytohemagglutznrn: NK = natural kzller cell; LDCC = lecttn-dependent cyto-
T cell cutture
NK activity
LDCC activity
Positive wells were identified after 10 to 14 days and expanded as described for polyclonal T-cell lines. The cloning
efficiency was approximately 10%.
of Leukocyte Surface Antigens
The mAbs used are listed in Table 1. Immunoreagents were
dissolved in solutions A, B, or C. Solution A consisted of
PBS, p H 7.2, and 2% BSA. Solution B consisted of PBS,
2% BSA, 10% heat-inactivated serum from A B + donors,
and 5% heat-inactivated horse serum. Solution C consisted
of PBS, 2% BSA, and 5% heat-inactivated goat serum.
Cells were suspended in solution A and
divided into aliquots of 0.5 to 1.0 x lo6 cells. Each aliquot
was treated with a different pair of phycoerythrin (PE)- and
FITC-labeled d b s according to a standard protocol provided by Becton Dickinson. The following antibody combinations were used: PE-Leu-2a
+ FITC-Leu-4, PE-Leu-15 FITC-Leu-Za, PE-Leu-19
FITC-Leu- 11, and PE-Leu- 15 FITC-Leu- 11. Control samples were treated with nonimmune, fluorochrome-conjugated, isotype-matched antibodies.
CLONING BY LIMITING DILUTION. T cells were placed into
6.4-mm round-bottom microwells at 1 and 3 cells per well,
and stimulated with 10 pg/ml of PHA-P plus 4 X 10’ irradiated feeder cells in LGM [24]. After 48 hours, IL-2 was
added to a hnal concentration of 10% Lymphocult T-LF [24].
FLUORESCENCE MICROSCOPY.Fresh cells in suspension
were immunostained with mAbs according to standard procedures using antibody concentrations suggested by the suppliers. After the last rinse, cells were suspended in 10 pl of
solution A, placed on a coverslip pretreated with 50 pgiml
of poly-~-lysine,and dried. The coverslips were mounted in
a glycerol-based mounting medium containing 1 mgiml of
For staining cells on coverslips, 1 to 2 x lo5 cells were
smeared on poly-L-lysine-coated coverslips, air-dried, and
fixed in acetone for 10 minutes at 4°C. The cell smears were
immunostained like tissue sections as described [2,31. After
the last rinse, the coverslips were mounted in the same medium as the cell suspensions.
CD4’ CELLS. Five predominantly CD4 *
polyclonal T-cell lines were converted into predominantly
CD8+ T-cell lines by treatment with OKT4 and complement
as suggested by the supplier (Ortho Pharmaceutical, Raritan,
NJ). The procedure killed at least 75% of the CD4’ cells
as assessed by trypan blue staining and by a fluorescenceactivated cell sorter (FACS). The remaining cells were expanded by stimulation with PHA-P and irradiated feeder
cells, as described already. After restimulation, these T-cell
lines contained more than 80% CD8+ T cells.
500 Annals of Neurology
Vol 29 No 5 May 1791
Unlabeled antibodies, biotinylated Leu-4, and biotinylated
WT31 were diluted in solution B and applied in an indirect
procedure with biotinylated horse anti-mouse IgG ( 5 p.g/ml
in solution B) and rhodamine-labeled avidin-D (1.7 pg/ml in
solution A) as the secondary reagents. For Leu-7 and Leu-1 1b
(IgM isotype), rhodamine-labeled goat anti-mouse IgM (1.25
pg/ml in solution C) was used as the secondary antibody.
WT31 could not be used on fixed cells.
The preparations were examined in a Zeiss photomicroscope equipped with epifluorescence, phase, and bright field
optics. At least 250 cells were counted in each preparation.
In control experiments, isotype-matched nonimmune Ig was
used at the same concentration as the primary antibody. Biotinylated nonimmune mouse IgG was substituted for biotinylated antibodies, and FITC-labeled nonimmune mouse IgG
for FlTC-labeled antibodies. In most instances, PBMCs were
treated in parallel and used as an additional control.
cJJtO&?XiCig hSdJ’.f
All cytotoxicity assays were performed with effector cells that
had been cultured in lectin-free medium for at least 7 days.
N K ACTIVITY. NK-sensitive K-562 cells were used as targets in a 4-hour 5’Cr-release assay. K-562 cells (2 x lo6)
were labeled by incubation for 1 hour at 37°C in 1.0 ml of
RPMI-1640 containing 10% FCS (FCS-RPMI) and 200 CCi
of Na251CrOl(specific activity, 250 to 500 mCi/mg of Cr).
After labeling, target cells were washed 2 times in D-PBS,
resuspended in 10 ml of FCS-RPMI, incubated for 30
minutes at 37“C, and washed once more with D-PBS. 51Crlabeled target cells (5 x lo3) were cocultured with musclederived lymphocytes at the effector-target (E/T) ratios indicated in the Results section. The cytotoxicity assays were
performed in triplicates in round-bottom, 6.4-mm microwells
in a total volume of 200 p.1 of FCS-RPMI per well. Spontaneous release was measured in six wells that contained only
effector cells and medium, and maximal release was measured in six wells that contained target cells and 5% Triton
X-100. The plates were centrifuged at 300 g for 5 minutes
and were incubated at 37°C in 5% carbon dioxide, 95% air.
After 4 hours, the plates were centrifuged at 600 g for 10
minutes and 100 p1 of supernatant from each well was
counted in a Beckman-Gamma-4000 counter.
assay was identical to the N K assay but activity against K-562
target cells (or against myotube target cells) was measured in
the presence of 10 p.g/ml of PHA-P. In the N K and lectindependent cell-mediated cytotoxicity (LDCC) assays, background release ranged from 10 to 28% of the maximal release.
MYOCYTOTOXICITY.Cytotoxic activity against autologous
and allogeneic myotubes was measured in an 8-hour 5’Crrelease assay. Target myotubes were prepared as follows. Myoblasts were plated at 5 x l o 3 cellsiwell in muscle growth
medium into gelatin-coated, flat-bottom, 6.4-mm microwells.
After 24 hours, muscle fusion medium was substituted for
muscle growth medium. On days 6 to 8, muscle fusion medium was removed, and 30 pl of FCS-RPMI containing 6
p.Ci of Na,51Cr04 was added to each well. After incubation
for 1 hour at 3 7 T , the wells were washed twice with D-PBS,
and 200 p,l of FCS-RPMI (assay medium) was added to each
well. The plates were again incubated for 30 minutes, washed
with D-PBS, and the 51Cr-labeledmyotubes were then used
as targets in the cytotoxicity cxperiments.
The E/T ratios were based on the numbers of nuclei. If
sufficient numbers of responder and target cells were a v d able, cytotoxicity was measured at the E/T ratios of 5 , 10,
and 20. If the number of cells available for analysis was small,
cytotoxicity was meawred only at the E/T ratio of 20. For
each E/T ratio, the assay was done in triplicate in a total
volume of 200 IJ.~of FCS-RPMI per well. Spontaneous release and maximal release were determined as in the other
cytotoxicity assays. Background release ranged from 10 to
24% of the maximal release.
In all cytotoxicity assays, the mean percent
specific cytotoxicity, c, was calculated for each set of triplicates according to
with x being mean experimental release; b, mean background
release; and m, mean maximal release. The composite standard deviation SD, was calculated according to Bevington
1251 (SD, incorporates the standard deviations of the three
measured variables [SD,, SD,, SD,]). Cytotoxicity was considered significant if c was greater than twice the value of
SD, [25].
HLA Typing
HLA-ABC typing was carried out on the T-cell lines or on
PBMCs using a standard National Institutes of Health cytotoxicity assay (Mayo Clinic Blood Bank). In 1 patient, both
freshly obtained PBMCs and muscle-derived T line cells
were typed, leading to identical results.
A/lorphology of CTL-hfyotube Interaction
For observation under phase optics, myotubes and CTLs
were cocultured in 16-mm wells. A grid of perpendicular
lines was drawn on the outer surface of the bottom of the
well to help locate individual myotubes. Cultures were
viewed and photographed with a Nikon MS inverted microscope.
IL-2 Responsiveness of Muscle-Derived T Cells
in Primay Cultwe
Using protocol A (see Materials and Methods), of 10
specimens (PM, 6 ; IBM, 2; OM, 2), T cells could be
expanded from muscle only in 1. These cells were from
a patient with PM with an unusually heavy endomysial
inflammatory exudate (PM9 in Table 2). By contrast,
using protocol B, T cells were expanded from 9 of 10
cultures. The exception was a specimen of IBM with
a mild endomysial inflammatory exudare. Therefore,
protocol B was used for all subsequent specimens.
Hohlfeld and Engel: Coculture of Cytotoxic T Cells and Myotubcs
Table 2.CD4 and CD8 Expression in 25 Muscle-Derived T-cell Lines'
Phenotype (%)
Cell Line (Code)
Polymyositis (n = 10)
PM 1-L
PM 5 -L
PM 7-L
PM 10-L
Inclusion body myositis (n = 5)
Dermdtomyositis (n = 5 )
DM 5-L
Other muscle diseases (n = 5)
OM 1-L
OM 5-L
< I
< I
< 1
< 1
< I
< I
< 1
aAnalyzed by flow cytometry after 4-6 weeks in culture (IBM2-L was analyzed by fluorescence microscopy).
bAfter 4 weeks in culture (compare with Table 3).
not done.
Surface Antigens of Muscle-Derived T-cell Lines
Polyclonal T-cell lines were expanded from 25 muscle
specimens (see Table 2). Unless stated otherwise, all
phenotypic and functional analyses of the T-cell lines
were done after 4 to 6 weeks in culture. All cell lines
were CD2 'CD3+, and were consistently negative for
antigens associated with N K cells (HNKlICD57,
CD16, NKHl/CD56, CDll b), B cells (CD22), granulocytes (CD1 Ib), monocytes (CD1Ib), and a subset
of suppressor cells (CDl lb) (see Table 1 for a complete list of the mAbs used).
CD4 and CD8 expression was determined by flow
cytometry (Fig 2, see Table 2). One cell line (PM9-L)
contained 18% of cells with the uncommon phenotype
C D 4 - C D K . The CD4-CD8- cells expressed the
gamma-delta T-cell receptor (TCR) [26, 27). Because
the gamma-delta TCR could be expressed not only by
CD4-CD8- T cells, but also by CD8+ cells [28), we
also examined the predominantly CD8+ T-cell lines
for alpha-beta TCR and gamma-delta TCR expression,
using WT31 and T,gamma-A mAbs [29}. A11 CD8'
502 Annals of Neurology
Vol 29 No 5 May 1991
cells were positive for WT31, but did not react with
T,gamma-A. The results show that in all specimens
except PM9 the muscle-derived CD8+ cells had the
phenotype of classic cytotoxic T cells.
NK and LDCC Activity of Muscle-Derived T Cells
Sixteen predominantly CD8+ lines (PM, 6; IBM, 4;
DM, 5 ; OM, 1) and one predominantly CD4+ line
(OM2) were tested for N K activity and LDCC against
the NK-susceptible K-562 cell line (Table 3). Five of
the 16 predominantly CD8' lines (PM1-L, PM3-L,
DM2-L, DM4-L, and DM5-L) were derived from their
predominantly CD4+ parent lines (see Table 2) by
treatment with OKT4 and complement.
None of the 17 T-cell lines had appreciable N K
activity against K-562 target cells at an E/T ratio of
20:l (Fig 3, see Table 3). By contrast, all T-cell lines
had significant LDCC activity against K-562 cells at an
E/T ratio of 20:l (see Fig 3 and Tabie 3). LDCC activity was also demonstrated with myotube targets if
PHA-P was added to the cocultures (not shown).
Fig 2. Two-colorjou' cytometric analysir of CD4 and CD8 expression in 3 representative muscle-derived T-cell lines. Red juorescence iphycoerythrin [PE]) corresponds to CD8 expression (ordinate) and green Jluorescence Huorescein isothiocyanate [FITC)) to
CD4 expession (abscissa). Contour maps are divided into quadrants to represent iells stained only with PE (upper left), with
FITC and PE (upper right), and only with FITC (lower right),
and unstained cells (lower left). (A)Line IBM4-L (9794
CD8'CD4-). (BI Line PM6-L (99% CD4'CD8-). (Ci Line
PM8-L (81% CD8'CD4-, 12% CD4+CD8-). Compare
with Table 2.
These results established that the muscle-derived Tcell lines had functional cytotoxic capability, but were
distinct from NK cells by phenotypic and functional
Adhesion of T Cells to Cultured Myotubes
In either autologous or allogeneic cocultures, T cells
of both major phenotypes (CD4+,CDS') adhered to
myotubes (Fig 4). Adherence was independent of the
antigen-specific receptor of the T cells because it occurred across histocompatibility barriers (see Fig 4D),
Table 3. Natural Killer Cell Activity, Lectin-Dependent Cytotoxicity,and Autoreactive Myocytotoxicitji of Muscle-Derived T-cell Lines
Cell Line
Polymyositis (n = 6)
PM 1-Lc
PM 10-L
Inclusion body rnyositis (n = 4 )
Derrnatomyositis (n = 5)
Other muscle diseases (n = 2)
Percentage of
CD8+CD4- Cells
Natural Killer Cell
LDCC Activity'
0.8 ? 0.4
33.3 2 2.6
36.1 2 2.7
55.4 5.4
18.6 +- 1.1
44.1 +- 2.4
11.6 f 2.9"
5.8 f 2.1d
1.4 t 0.8
f 0.5
0.8 0.7
37.3 +- 2.0
39.2 +- 3.4
61.4 2.9
61.0 3.4
7.9 t 3.4d
3.5 f 1.9
4.1 +- 3.4
2.2 t 0.6
0.7 f 0.6
5.9 t 2.ld
0.5 i- 0.2
3.4 i 1.7
60.5 +- 3.6
82.0 3.9
62.2 f 4.7
4.4 2.2
27.2 2 5.Sd
< 1
1.1 i 0.8
1.7 f 0.1
13.0 f 1.6
62.1 t 2.4
3.3 t 3.1
"Analyzed after 4-6 weeks in culture. Cytotoxicity values represent mean
SD of specific "Cr release (effector-target ratios
bK-S62 target cells.
'Derived from predominantly CD4 - lines by treatment with OKT4 and complement.
'Statistically significant cytotoxicity against autologous myotubes.
'After 6 weeks in culture (compare with Table 2).
LDCC = lectin-dcpendcnt cytotoxicity.
Hohlfeld and Engel: Coculture of Cytotoxic T Cells and Myotubes
CD4+ effector cells
CD8+ effector cells
Effector:target ratio
Fig 3. Lectin-dependent cytotoxzcity (LDCC) and natural killer
cell (NK) activity of a representative CD4+ (OM2-L, left) and
CD8' (IBM4-L, right) muscle-derived T-cell line icompare with
Table 31. Cytotoxicity values represent mean f SD of speczfic "Cr
release from K-562 target cells.
Fig 4. Adhejion of T celh t o cultured myotubes. T celh and myotubes were coculturedfor at leaJt 6 hours. I n A, B, and D , the
nonadherent cells were rinsed o f f before the photograph was taken.
(A)Autologow coculture, predominantly CD4' T cells (OM2).
(B) Autologous coculture, mixed CD4' and CD8' T cells
(PMS). (C) Autologous coculture, predominantly CD8' cells
504 Annals of Neurology Vol 29 No 5
May 1991
(IBM3). ID) Allogeneic coculture. mixed CD4' and CD8' T
cells (mitogen-stimulatedperipheral blood T cells from one of the
authors {R. H.)). The variation of celldensities between the different cultures rejects chance variations of cell density in indikidual
cultures. T h m was no evidence of myotube lysis during the experimental period (phase microscopy, x 230).
and with peripheral blood T cells obtained from a normal donor (see Fig 4D). Further, the adherence of T
cells to myotubes was completely independent of their
cytotoxicity (see Figs 4A through 4D). These findings
indicate that T cells adhere to myotubes independent
of their target specificity or cytotoxicity.
Autoreactive Myocytotoxicity
One of the main questions asked in the present study
was whether muscle-derived T cells can kill autologous
myotubes. T-cell lines were analyzed for autoreactive
cytotoxicity after 4 to 6 weeks in culture (see Table 3).
Twelve of the 17 T-cell lines, including 1 predominantly CD4' control line (OM2-L), showed no or no
significant cytotoxicity for autologous myotubes.
Five T-cell lines (3 of 6 PM, 1 of 4 IBM, and 1 of
5 DM) had low to moderate but statistically significant
cytotoxicity against autologous myotubes. The cytotoxicity values ranged from 5.8 to 27.2% lysis at an EIT
ratio of 20 :1.
The autoreactive potential of the T cells declined
when the cells were retested after prolonged culture.
For example, the specific lysis effected by T-cell line
DM5-L declined from 27.2 k 5.8% at 6 weeks in
culture to 9.4
3.1% at 10 weeks in culture. During
this time, the phenotype of the line remained stable.
After 10 weeks of continuous culture, however, this
cell line began to coexpress the CD8 and CD56iLeu19 antigens and acquired NK-like cytotoxicity against
K-562 cells. The observed instability in autoreactive
potential could be due to clonal shifts within the polyclonal cell lines or to change in the properties of a
given clone. To distinguish between these possibilities,
we isolated C D 8 + N K H l sublines from DM5-L by
limiting dilution cloning, and followed the properties
of the clones for 3 months. Like the polyclonal lines,
the isolated clones acquired the NKHl+/CD56' phenotype, indicating de novo expression of the Leu-19
antigen on individual cells [30}.
In brief, the results show that in each of the inflammatory myopathies, some muscle-derived CTLs had a
statistically significant cytotoxic effect against autologous myotubes in the complete absence of N K activity.
Methodological Considerations
The two main aims of the present study were (1) to
develop and characterize an in vitro system for the
study of CTL-myotube interactions, and (2) to use this
system to investigate whether muscle-derived CTLs
can kill autologous myotubes.
Because the T cells in muscle lacked responsiveness
to IL-2 and because the antigens against which the T
cells are sensitized are unknown, we used PHA-P, a
mitogenic lectin, for the polyclonal activation of T cells.
An advantage of this protocol {l6f is that it allows the
establishment and screening of a large number of T-cell
lines from many patients. The disadvantage is that the
polyclonal T-cell lines do not remain stable over an
extended period of time, and relevant CD4+ or CD8+
T-cell clones can be lost during culture.
Properties of the T-cell Lines
There was no correlation between the proportion of
CD4' and CD8 ' cells in muscle and in the T-cell
lines. This is not surprising. Even at the time of initial
testing (i.e., after 4 to 6 weeks in culture), the cell
lines must have become enriched in T cells that grew
optimally under the culture conditions, and these were
not necessarily the cells that were pathogenic in vivo.
The T-cell lines studied here were predominantly
WT31+ T,gamma-A-. This implies that most cells expressed the common TCR isotype, the alpha-beta
TCR. When these cells are cytotoxic, they operate in
an MHC-restricted, antigen-specific manner. N o evidence was found for MHC and antigen-unrestricted
NK-like activity against K-562 target cells. This was
not due to lack of cytotoxic potential, since all T-cell
lines displayed LDCC. Thus, the CD8' musclederived T cells were classic CTLs by phenotypic and
functional criteria.
Three of 6 PM, 1 of 4 IBM, and 1 of 5 DM lines
showed statistically significant low to moderate cytotoxicity against autologous myotubes. The lack of autoreactive cytotoxicity by some PM and IBM cell lines
in vitro, in the face of myocytotoxicity in these cases
in vivo, could be related to incomplete expansion or
nonexpansion of those CD8' T cells that were pathogenic in vivo. Incomplete expansion of relevant CD8+
cells may also account for the fact that those T-cell
lines that were cytotoxic to autologous myotubes
showed only low to moderate cytotoxicity. Another
reason for lack of autoreactivity by PM and IBM T-cell
lines enriched in CD8' cells could be insufficient expression of the appropriate antigen on the autologous
myotubes, as could occur if antigen expression was developmentally regulated, or was constitutively low on
immature fibers and cultured myotubes.
The autoreactive myocytotoxicity in vitro, when
present, may indicate the recognition of an antigen on
the myotubes, or a form of nonspecific killing. The
antigen-specific mechanism is more likely because 3 of
the 5 autoreactive CTL lines (PM8-L, IBM1-L, DM5L) were cytotoxic for autologous myotubes but completely spared several allogeneic myotube targets (not
shown). Also, at the time of testing, the autoreactive
CTL lines lacked the surface markers of N K cells and
had no significant N K activity against K-562 target cells
(see Table 3).
The antigen recognized by the autoreactive CTL
Hohlfeld and Engel: Coculture of Cytotoxic T Cells and Myotuhes
lines could be a genuine autoantigen, or a foreign antigen. CD8+ T cells usually react to peptides derived
from endogenously synthesized antigen (e.g., viral proteins) that are associated with MHC class I molecules
(reviewed by Germain 13 11). There is evidence indicating that the antigen-binding site of most MHC
class 1 and 11 molecules is constitutively occupied by
a processed self-peptide r32-34 1. The MHC-bound
self-peptides may vary between different tissues [ 3 5 J
Theoretically, a tissue-specific self-peptide could be
recognized by autoaggressive CTLs, leading to autoimmune lysis of the target cells. At present, we cannot
distinguish between autoreactive myocytotoxic T cells
recognizing isolated MHC molecules, a tissue-specific
antigen, a more widely distributed self-antigen, or a
foreign (perhaps latent viral) antigen. One way to further define the antigen(s) is to use other autologous
cells as targets for the CTLs.
The T-cell line with the highest autoreactive cytotoxicity (DMS-L) was derived from a patient with early
DM. This was unexpected because CTLmediated
muscle fiber injury is not a feature of DM [2}, and
because frozen sections of muscle from which the Tcell line was derived showed no invasion of muscle
fibers by T cells. Further studies are required to determine whether CD8 T cells showing myocytotoxicity
can be isolated from muscles of some patients with
One morphological aspect of T cell-mediated muscle fiber injury is the extensive replacement of nonnecrotic muscle fibers by CD8+ CT’Ls 12-41. This was
not seen in vitro where lysis of myotubes occurred
rapidly. Possible reasons for this discrepancy include
the following: (1) Mature muscle fibers may be capable
of repair mechanisms lacking in myotubes. (2) The cultured CTLs may have changed their properties in vitro.
(3) The CTLs in vivo act in concert with macrophages
and other mononuclear cells, including regulatory T
cells r2-41, that are absent in the in vitro system. (4)
Other factors in the local microenvironment in vivo
may not be reproduced in vitro.
Other Observations
A nonspecific aspect of our cocultures was tight adhesion of T cells to myotubes. These cell to cell contacts
were independent of the phenotype, antigen specificity, and lytic activity of the T cells. This observation is
in agreement with the idea that adhesion of effector
cells to target cells is a necessary but not a sufficient
requirement for lysis [36, 371. The adhesion is likely
to be mediated by “adhesion molecules” expressed on
both T-cell and myotube surfaces. The experimental
system used would be suitable for manipulating T
cell-myotube adhesion in vitro with mAbs against candidate adhesion molecules, such as ICAM-1, LFA-1,
CD2, and LFA-3 {38).
Annals of Neurology
Vol 29 No 5
May 1991
T cells with the phenotypic and functional properties
of classic CTLs can be recovered from muscle in inflammatory myopathies. We established and characterized an in vitro system in which T cell-myotube interactions can be investigated. Using this system we
systematically investigated the effects of CT’Ls on autologozls myotubes. A low but significant autoreactive cytotoxic effect was observed in several cases. The cyrotoxic effects in this defined system were much smaller
than those reported in earlier studies that had overlooked the role of MHC restriction and the mechanism
of nonspecific NK-like killing. We are now designing
experiments to prove or disprove the hypothesis that
the CTLs recognize an autoantigen on the myotubes.
This work was supported by National Institutes of Health grant
NS-6277 and a research center grant from the Muscular Dystrophy
Association. Dr Hohlfeld was a recipient of a Heisenberg grant from
the Deutsche Forschungsgemeinscha.
The authors thank Dr J. A. Katzmann and Ms T. Kimlinger ior help
with the flow cytometry, and Drs Jelko Bajzer and Bernhard
Fleischer for valuable suggestions.
1. Whitaker JN. Inflammatory myopathy: a review of etiologic and
pathogenetic factors. Muscle Nerve 1982;5:573-592
2. Arahdta K, Engel AG. Monoclonal antibody analysis of mononuclear cells in myopathies. I: quantitation of subsets according
to diagnosis and sites of accumulation and demonstration and
counts of muscle fibers invaded by T cells. Ann Neurol
3. Engel AG, Arahata K. Monoclonal antibody analysis of mononuclear cells in myopathies. 11: phenotype of autoinvasive cells
in polymyositis and inclusion body myositis. Ann Neurol
1984;16:209-2 15
4. Arahata K, Engel AG. Mononuclear anribody analysis of mononuclear cells in myopathies. 111: immunoelectron microscopy
aspects of cell-mediated muscle fiber injury. Ann Neurol
1986;lO:112-1 25
5. Emslie-Smith AM, Arahata K, Engel AG. Major histocompatibility complex class I antigen expression, immunolocalization of
interferon subtypes, and T cell-mediated cytotoxicity in myopathies. Hum Pathol 1989;20:224-23 1
6. Whitaker JN, Engel WK. Vascular deposits of immunoglobulin
and complement in idiopathic inflammarory myopathy. M Engl
J Med 1972;286:333-338
7. Kissel JT, Mendell JR, Rammohan KW. Microvascular deposition of complement membrane attack complex in dermatomyositis. N Engl J Med 1986;314:329-334
8. Emslie-Smith A, Engel AG. Microvascular changes in early and
advanced adult dermatomyositis. Ann Neurol 1990;27:343356
9. Currie S, Saunders M, Knowles M, Brown AE. Immunological
aspects of polymyositis: the in vitro activity of lymphocytes on
incubation with muscle antigen and with muscle cultures. Q J
Med 1971;157:63-84
10. Dawkins RL,Mastaglia FL. Cell-mediated cytotoxicity to muscle
in polymyositis: effect of immunosuppression. N End J Med
11. Haas DC, Arnason BGW. Cell-mediated immuniry in polymyo-
sitis: creatinine phosphohnase release from muscle cultures.
Arch Neurol 1974;31:192-196
Rosenschein U, Radnay J, Shoham D, et al. Human musclederived, tissue-specific, myocytotoxic T cell lines in dermatomyositis. Clin Exp Immunol 1987;67:309-318
Kakulas BA, Shute GH, Leclere ALF. In vitro desrruction of
human foetal muscle cultures by peripheral blood lymphocytes
from patients with polymyositis and lupus erythematosus. Proc
Ausc Assoc Neurol 1971;8:85-92
Haas DC. Absence of cell-mediated cytotoxicity to muscle cultures in polymyositis. J Rheumatol 1980;7:671-676
Hohlfeld R. Neurological autoimmune disease and the rrimolecular complex of T-lymphocytes. Ann Neurol 1989;25:
Mayer TG, Fuller AA, Fuller TC, et al. Characterization of in
vivo-activated allospecific T lymphocytes propagated from human renal allograft biopsies undergoing rejection. J Immunol
Bonazzo GF, Todd I, Mirakian R, et al. Organ-specific autoimmunity: a 1986 overview. Immunol Rev 1986;94:137169
Knapp W, Rieber P, Dorken B, et al. Towards a better definition of human leukocyte surface molecules. Immunol Today
Trinchieri G. Biology of natural killer cells. Adv Immunol
Hohlfeld R, Engel AG. Lysis of myotubes by alloreactive cytoroxic T cells and natural killer cells: relevance to myoblasr transplantation. J Clin Invest 1990;86:370-374
Blau HA{, Webster C. Isolation and characterization of human
muscle cells. Proc Natl Acad Sci USA 1981;78:5623-5627
Hohlfeld R, Engel AG. Induction of HLA-DR expression on
human myoblasts with interferon-gamma. Am J Pathol 1990;
Boyum A. Isolation of mononuclear cells and granulocytes from
human blood. Scand J Lab Invest 1968;21:77-89
Morena A, Pantaleo G, Moretta L, et al. Direct demonstration
of the clonogenic potential of every peripheral blood T cell:
clonal analysis of HLA-DR expression and cytolytic activity. J
Exp Med 1983;157:743-754
25. Bevington PR. Data reduction and error analysis for the physical
sciences. New York: McGraw-Hill, 1969
26. Brenner MB, Strominger JL, Krangel MS. The gammddelta T
cell receptor. Adv lmmunol 1988;34:133-192
27. Hohlfeld R, Ii K, Engel AG. T cells expressing the gamma-delta
receptor can mediate polymyositis. Neurology 1990;4O(suppl
1):203 (Abstract)
28. Groh V, Porcelli S, Fabbi M, et al. Human lymphocytes bearing
T cell receptor gamma/delta are phenotypically diverse and
evenly distributed throughout the lymphoid system. J Exp Med
20. Jitsukawa S, Faure F, Lipinski M, et al. A novel subset of human
lymphocytes with a T cell receptor-gamma complex. J Exp Med
30. Lanier LL, Le AM, Ding EL, et al. Expression of Leu19 (NKH-1)
antigen on IL-2 dependent cytotoxic and non-cytotoxic T cell
lines. J Immunol 1987;138:2019-2033
31. Germain RN. The ins and outs of antigen processing and presentation. Nature 1986;322:687-688
32. Babbitt BP, Matsueda G, Haber E, et al. Antigenic competition
at the level of peptide-Ia binding. Proc Natl Acad Sci USA
1986;83:45O9-45 13
33. Bjorkman PJ, Saper MA, Samraoui B, et al. The foreign antigen
binding site and T cell recognition regions of class 1histocompatibility antigens. Nature 1987;329:5 12-5 18
34. Buus S, Sette A, Colon SM. Grey HM. Autologous peptides
constitutively occupy the antigen binding sire on Ia. Science
35. Marrack P, Kappler J. T cells can distinguish between dogeneic
major histocompatibility complex producrs on different cell
types. Nature 1988;332:840-843
36. Berke G. The cytolytic T lymphocyte and its mode of action.
Immunol Lett 1989;20:169-178
37. Young JDE, Liu CC, Persechini PM, Cohn ZA. Perforindependent and -independent pathways of cytoroxicity mediated
by lymphocytes. lmmunol Rev 1988;103:161-202
38. Springer TA, Dustin ML, Kishimoto TK, Marlin SD. The lymphocyte function-associated LFA-1, CD2, and MA-3 molecules:
cell adhesion receptors of the immune system. Annu Rev Immuno1 1987;5:223-252
Hohlfeld and Engel: Coculture of Cytotoxic T Cells and Myotubes
Без категории
Размер файла
940 Кб
myopathic, muscle, isolated, inflammatory, cytotoxic, myotubes, coculture, autologous, cells
Пожаловаться на содержимое документа