Procainamide-lymphocyte reactions. a possible explanation for drug-induced autoimmunity

1019
PROCAINAMIDE-LYMPHOCYTE REACTIONS
A Possible Explanation for Drug-induced Autoimmunity
HARRY G. BLUESTEIN, DOUG REDELMAN, and NATHAN J. ZVAIFLER
Procainamide therapy has been associated with a
lupus erythematosus (LE) like illness. At least 50% of
patients who take the drug for a long time develop antinuclear antibodies and approximately 10%of these develop the complete clinical syndrome (1-3). There has
been considerable speculation about the pathogenesis of
the drug-induced disease. Most theories implicate drugnucleoprotein interactions, which result in the induction
of antinuclear antibodies, but these explanations do not
account for other autoantibodies often present in a patient with drug-induced lupus.
The autoimmune phenomena of idiopathic lupus
have been attributed to impaired immune regulation
(4). To determine if similar mechanisms operate in the
drug-induced syndrome, we looked for evidence of procainamide-induced lymphocyte abnormalities and
found that many individuals taking the drug develop
antibodies to cell surface antigens on normal lymphocytes. The frequency with which these antilymphocyte
(ALA) antibodies develop and their concentration in
the circulation are significantly increased in those patients with the procainamide-induced lupus syndrome.
When the drug is discontinued, the antilymphocyte activity diminishes rapidly, in parallel with clinical improvement. We also found that procainamide alters the
in vitro proliferative response of normal human lym-
From the Rheumatology Division, Department of Medicine,
University of California, San Diego.
Supported in part by grants from the PHS (AM-07062, AM14916) and an Arthritis Foundation Center Grant.
Address reprints to Nathan J. Zvaifler, Rheumatology Division, Department of Medicine, 225 Dickinson Street, San Diego CA
92103
Arthritis and Rheumatism, Vol. 24,No. 8 (August 1981)
phocytes to mitogenic and antigenic stimulation. Taken
together, these findings raise the possibility that procainamide induces the autoimmune lupus-like syndrome by its direct interaction with lymphocyte membranes.
Clinical observations
Seventeen patients were identified as having procainamide-induced lupus as evidenced by a history of
continued procainamide therapy, the development of
new symptoms compatible with lupus, a serum antinuclear antibody titer of 1 :10 or greater, and resolution
of symptoms after discontinuation of the drug. The patients ranged in age from 49 to 83 years and 15 were
male. The average total procainamide dose per patient
in those with the drug-induced lupus syndrome was 695
gm, ranging from 60 to 1320.
The clinical features of the procainamide-induced syndrome in our 17 patients included rheumatic
complaints (arthralgia, diffuse myalgia, and arthritis) in
16; pleurisy or chest pain in 12; fever and weight loss in
7; and skin rash, mental confusion or headache, and
Raynaud's phenomenon in 3 patients each. All subjects
had a positive antinuclear antibody test, most with titers
of 1:320 or greater. Serum antibodies to native doublestranded DNA (Farr technique) were not detected in
any samples. Other immunologic abnormalities included positive direct and indirect Coombs tests, total
leukocyte counts of less than 3000/mm3, and mild depressions in C3 or C4. A surprising finding was a circulating anticoagulant in 3 patients and an unexplained
prolonged partial thromboplastin time (PTT) in another.
BLUESTEIN ET AL
1020
Lymphocytotoxic antibodies
Cold reactive (15OC) lymphocytotoxic antibodies
were measured by a modification of the Terasaki microdroplet dye exclusion assay described in detail elsewhere (5). Peripheral blood lymphocytes (PBL) obtained from venous blood of normal volunteers by
Ficoll-Hypaque density centrifugation were used as target cells. Each serum was tested against lymphocytes
from at least 3 donors. The mean percentage of dead
cells in triplicate assays was expressed as the percent
cytotoxicity. To compare the relative cytotoxic capacity
of each serum, a 50% cytotoxicity titer (CC,,) was determined. The CC,, is defined as the log, of the reciprocal
of the greatest serum dilution producing 50% cytotoxicity. All sera were tested at an initial 1:2 dilution.
Lymphocytotoxic antibodies (CC,, L 1) were detected in the serum of l l of the 17 subjects (65%) with
procainamide-induced lupus, and one-half of the active
sera had 5090 cytotoxicity titers of 1 :8 or greater. In
contrast, antilymphocyte activity was found in sera
from only 20% of 15 control subjects on long-term procainamide therapy who did not have the lupus syndrome, and none was positive at serum dilutions greater
than 1 :2. Another control group consisted of 63 men
between the ages of 45 and 65 who had not received
procainamide. Only 3 of their sera produced 50% lymphocytotoxicity, none at greater than a 1 :2 serum dilution. There was no correlation between the titer of lymphocytotoxic antibody and the total amount of
procainamide taken by the patients with drug-induced
disease. Neither was there an association of the antibody titer and the number or severity of symptoms or
other laboratory abnormalities.
Sera obtained at varying intervals after cessation
of drug therapy were available from 6 of the procainamide-induced lupus patients whose serum was
Table 1. Sequential serum lymphocytotoxic antibody titers in
patients with procainamide-induced lupus
_____
~
Lymphocytotoxicity(CC,,)
Patient
Sample 1*
Sample 2*
2
3
5
0
0
1
5
5
2
3
6
3
Interval
(weeks)
~~
RF
WK
LR
GK
GE
PR
16
7
6
8
3
4
* Sample 1 was obtained at time of diagnosis before procainamide
was discontinued. Sample 2 was obtained at the interval indicated after discontinuing procainamide therapy.
lymphocytotoxic at the time of diagnosis. The titer of
lymphocytotoxic antibody fell abruptly after discontinuation of procainamide (Table 1). Cytotoxic activity in
all cases decreased 75% or more on retest after 3 to 16
weeks. The reduction was concomitant with the clearing
of symptoms, but was independent of the antinuclear
antibody titers which remained elevated and unchanged
for many months.
Effect of procainamide on mitogen and antigen
responses of normal human lymphocytes
Peripheral blood lymphocytes from healthy donors were cultured at a density of lo6cells/ml in RPMI1640 medium supplemented with 1oo/o pooled normal
human plasma. Purified PHA was added to the cultures
to a final concentration of 1, 0.1, or 0.05 pg/ml. Procainamide-HC1 obtained as a sterile aqueous solution
was added to the cultures in the concentrations indicated. Cultures were incubated at 37°C in a humidified 5% C0,-in-air atmosphere for 72 hours and 18 to 24
hours before harvesting were pulsed with 1 pCi of 3Hthymidine. The cultured cells were collected onto glass
fiber filter papers using an automated multiple sample
harvester and the 3H-thymidine incorporated into acid
insoluble material enumerated by liquid scintillation
spectrometry. The effect of procainamide on PHA-stimulated 3H-thymidineincorporation is expressed as a percentage of the response obtained in the absence of the
drug according to the formula:
% response =
PHA induced ’H CPM with drug
PHA induced 3HCPM without drug
loo
The response of primed lymphocytes to antigenic
stimulation in the presence of procainamide was also
studied. Peripheral blood lymphocytes from individuals
with positive tuberculin skin tests were cultured as described above with purified protein derivative (PPD,
Connaught Laboratories) at 2 @culture and the percent response calculated the same as for PHA.
Procainamide altered PHA-induced activation of
normal human PBL. A biphasic dose response was observed with marked suppression at higher concentrations but enhancement at lower doses (Figure 1). Concentrations of procainamide of 3.75 x lO-’M or greater
completely inhibited PHA-induced 3H-thymidineincorporation. However, at a tenfold lower drug concentration the response to PHA was enhanced 60% on the average beyond that obtained with PHA alone (P < 0.01
by Wilcoxon test; 0.02 > P > 0.01 by r-test). Higher
PROCAINAMIDE-LYMPHOCYTE REACTIONS
mean responses continued as the concentration of procainamide decreased to 10-6M but the differences were
not as striking (P< 0.05 by Wilcoxon test and 0.1 > P >
0.05 by paired t-test at 3.75 X lO-$M). Enhancement of
the response to PHA was observed best at suboptimal
stimulating concentrations of the mitogen (0.05-0.0 1
pg/ml). A similar pattern of complete suppression at
higher doses (3.75 X lO-’M and greater) and enhancement at lower concentrations was observed in studies of
antigen stimulated lymphocyte activation. The in vitro
response of PBL from tuberculin skin test positive individuals to PPD was 180% of normal at 3.75 x 10-4M
procainamide (P < 0.01 by Wilcoxon and paired ttests).
Effect of procainamide on in vitro antigenspecific antibody producing cell induction in
rabbit spleen cells
Procainamide has two effects on in vitro plaqueforming cell (PFC) responses of sheep erythrocyte
(SRBC) primed rabbit lymphocytes. The normal kinetics of antiSRBC responses are a sharp peak on day 4,
with a rapid decline by day 5 to less than 15% of the
peak (Table 2). At t 2 mM procainamide, no effect on
the peak response was observed at day 4 but there was a
significant enhancement of the number of PFC on days
5 and 6. As the procainamide concentration increased,
the total PFC response dropped dramatically. At 2.7
mM the day 4 response fell 8896, while the day 5 and
day 6 responses rose 2- to 3-fold higher than the cultures that did not receive the drug.
Discussion
Procainamide administration to humans induces
antinuclear antibodies and a disease resembling systemic lupus erythematosus. Antibodies to histones and
denatured, but not native, DNA are a regular finding
(6-8). These autoantibodies appear to result from
shared drug-nucleoprotein antigenicity or drug-nucleoprotein interactions which enhance the immunogenicity of nucleic acid antigens (9-13). The relationship between the antinuclear antibodies and the
drug-induced disease is ill defined. Those antibodies appear before clinical findings and persist in high titers for
months to years after the drug is withdrawn and all
symptoms have gone. Moreover, patients with procainamide-lupus show few stigmata of immune complex
disease (3,14,15).
Attributing the pathogenesis of procainamidelupus to antinuclear antibody-mediated immune com-
1021
w
m
z
0
n
m
w
CT
I-
z
W
80
0
a
w
n
1 I
10-6
10-5
10-4
10-3
PROCAINAMIDE CONC. (3.75 X M)
Figure 1. The effect of procainamide on the in vitro response of normal human peripheral blood lymphocytes to phytohemagglutinin
(PHA). For each procainamide concentration, the results are expressed as a percentage of the response to PHA (0.5 pg/ml) in the absence of added procainamide. Each point represents the mean of experiments, and the vertical bars represent the standard error (*).
plexes disregards the findings of other autoantibodies in
the drug-induced syndrome. As Blomgren and Vaughan
noted, some patients have antibodies to IgG (rheumatoid factor) and red blood cells (13). Others have reported false positive serologic tests for syphilis, leukopenia, and thrombocytopenia (15). Three of our 17
Table 2. Effect of procainamide on sheep erythrocyte plaqueforming cell responses of primed splenic lymphocytes from the rabbit
Procainamide
concentration
(M)
0
0
1.2 x
1.7 x
2.2 x
2.7 x
3.2 x
PFC/106 cultured cells
SRBC
0
10-3
10-3
10-3
10-3
10-3
+
+
+
+
+
+
Day3
Day4
Day5
Day6
158
226
3309
3800
2744
1629
713
99
20
523
1110
1346
985
1054
332
194
430
579
743
716
310
867
440
321
202
83
24
11
1022
patients had a circulating lupus anticoagulant. These
observations suggest impaired immune regulation in patients with procainamide lupus analogous to that seen in
the idiopathic form of the disease.
Antilymphocyte antibodies are autoantibodies
that can alter lymphocyte function (16). Individuals
treated with procainamide have a 5 fold increased frequency of lymphocytotoxicantibodies compared to ageand sex-matched control subjects. Most patients with
the lupus-like syndrome have circulating antilymphocyte activity at titers considerably higher than
the clinically unaffected individuals. Furthermore, procainamide withdrawal is followed by a rapid disappearance of the lymphocytotoxic antibody concomitant with clinical improvement. Thus the presence,
amount, and persistence of lymphocytotoxic antibodies
parallel the exposure to procainamide and the development and resolution of disease.
Procainamide-lymphocyte interactions have not
been documented by others, but chlorpromazine and
lidocaine were shown to inhibit cytotoxic murine T cells
and the activation of murine spleen cells by various
mitogens and bacterial lipopolysaccharides (17). The effects are dose-related, reversible, and not due to interference with mitogen binding to the lymphocyte membrane or impaired cell viability. Suppression of
mitogen-induced activation of human peripheral blood
lymphocytes and antigen-specific antibody responses of
rabbit splenocytes by procainamide occurs at similar
molar concentrations. Hyperresponsiveness of human
lymphocytes to PHA was seen at procainamide concentrations that approximate therapeutic blood levels. A
similar exaggerated effect on the activation of human
lymphocytes has been observed with chlorpromazine
(18)The biphasic effects of procainamide raise the
possibility that at lower concentrations the drug interacts preferentially with a subpopulation of immunoregulatory cells. The effects of procainamide on the in
vitro antibody-producing cell assay are under investigation to determine if selective alteration of helper or suppressor cell function can be documented. The preliminary results using sheep erythrocyte primed rabbit
splenic lymphocytes document a profound suppression
at drug concentrations above 3mM. Under 2mM, however, helper function is preserved. The increase in the
PFC response later in the time course raises the possibility of selective loss of suppressor cell activity, since
the normally rapid fall-off of the PFC response from its
peak has been attributed to T cell-mediated suppression. However, a more detailed analysis of the effects of
BLUESTEIN ET AL
procainamide on isolated lymphocyte subpopulations
will be needed to document selective immunoregulatory
activity.
The procainamide effects on lymphocyte activation are predicted by its “anesthetic” characteristics.
Anesthetics as a class are lipid soluble drugs that are incorporated into, and exert their effect upon, cell membranes. Many tranquilizers, anticonvulsants, and antiarrhythmics are included in this class of compounds
(19,20). Presumably procainamide intercalates into the
lipid membrane bilayer, thereby altering its architecture. As a result, the immune system may be continuously exposed to new lymphocyte membrane determinants that stimulate the production of autoreactive
lymphocyte antibodies. The antigenic stimulus should
cease when the drug is removed.
Based on this model, we would propose the following sequence of events in procainamide-induced
lupus. Antinucleic acid antibodies appear either because of drug-nucleoprotein interactions, as previously
suggested (6), or as an early indication of drug-induced
alteration of immune regulation. The antinuclear antibodies serve as a marker of the illness but are not sufficient to induce the clinical symptoms. Further deterioration in immune regulation, caused by the direct effect
of procainamide on lymphocyte membranes leads to development of disease. Antilymphocyte antibodies are a
consequence of the drug-induced membrane alteration.
While in the circulation they further impair normal
lymphocyte function, as has been shown for the antilymphocyte antibodies in the idiopathic form of the disease (21,22). Thus a positive feedback loop, amplifying
abnormalities of immune function, is established and
persists only so long as the drug is present.
The actual immunologic events responsible for
tissue injury still require explanation, but we suggest
looking beyond antinucleoprotein antibodies for the answer.
REFERENCES
1. Blomgren SE, Condemi JJ, Bignall MC, Vaughan JH:
Antinuclear antibody induced by procainamide: a prospective study. N Engl J Med 281:64-66, 1969
2. Molina J, Dubois EL, Bilitch M: Procainamide-induced
serologic changes in asymptomatic patients. Arthritis
Rheum 12:608-614, 1969
3. Lee SL, Chase PH: Drug-induced systemic lupus erythematosus: a critical review. Semin Arthritis Rheum
5:83-103, 1975
4. Tala1 N: Disordered immunologic regulation and autoimmunity. Transplant Rev 3 1:240-263, 1976
PROCAINAMIDE-LYMPHOCYTE REACTIONS
5 . Zvaifler NJ, Bluestein HG: Lymphocytotoxic antibody
activity in cryoprecipitates in serum of patients with SLE.
Arthritis Rheum 19:844-850, 1976
6. Tan EM: Drug-induced autoimmune disease. Fed Proc
33:1894-1897, 1974
7. Winfield JB, Davis JS: Anti-DNA antibody in procainamide-induced lupus erythematosus. Arthritis Rheum
17:97-110, 1974
8. Fitzler MJ, Tan EM: Antibodies to histones in drug-induced and idiopathic lupus erythematosus. J Clin Invest
i2:560-567, 1978
9. Yamauchi Y, Litwin A, Adams K, Zimmer H, Hess EV:
Induction of antibodies to nuclear antigen in rabbits by
immunization with hydralazine-human serum albumin
conjugates. J Clin Invest 56:958-969, 1975
10. Gold EF, Ben-Efraim S, Faivisewitz A, Steiner Z, Klajman A: Experimental studies on the mechanism of induction of antinuclear antibodies by procainamide. Clin Immunol Immunopathol7: 176-186, 1977
11. Waring MJ: Drugs which affect the structure and function
of DNA. Nature 219:132&1325, 1968
12. Eldredge NT, Robertson WB, Miller JJ 111: The interaction of lupus inducing drugs with deoxyribonucleic
acid. Clin Immunol Immunopathol 3:263-271, 1964
13. Blomgren SE, Condemi JJ, Vaughan JH: Procainamideinduced lupus erythematosus. Am J Med 52:338-348,
1972
14. Utsinger PD, Zvaifler NJ, Bluestein HG: Hypocomple-
DISCUSSION
Dr. Weigle: The sequence system of cellular events with
pokeweed mitogen is complex. An accessory T cell requirement involves interactions among T cells, macrophages, and other accessory cells.
Dr.ZvaMer: Our studies so far suggest that each of the
factors is affected to some degree. Unanue has examined the effect of both lidocaine and chlorpromazine
on macrophage function and on capping and patching in lymphocytes (Nature 25056, 1976). A group in
Minneapolis (Ferguson RM,Schmidtke JR, Simmons
RL: J Immunol 116:627-634, 1976) has looked extensively at pokeweed mitogen, phytohemagglutinin,
and other responses. Lidocaine particularly and
chlorpromazine are also good drugs for these studies.
Dr. Reidenberg: Is the effect of NAPA identical to that
of procainamide?
Dr. Zvaifler: Yes. We originally thought it would be interesting to dissect the two, but the effect was essentially the same with each.
Dr. Stollar: If there is an effect on lymphocytes, would
one expect a wide spectrum of increased antibody or
1023
mentemia in procainamide-associated SLE. Ann Intern
Med 84293-294, 1976
IS. Alarcon-Segovia D: Drug-induced lupus syndromes.
Mayo Clin Proc 44:664-681, 1969
16. Bluestein HG: Autoantibodies to lymphocyte membrane
antigens: pathogenetic implications. Clin Rheum Dis
4643459, 1978
17. Ferguson RM, Schmidtke JR, Simmons RL: Inhibition of
mitogen produced lymphocyte transformation by local
anesthetics. J Immunol 116:627-634, 1976
18. Ferguson RM, Schmidtke JR, Simmons RL: Effects of
psychoactive drugs on in vitro lymphocyte activation.
Original Article Series, XIV,pp 379406, 1978
19. Seeman P: The membrane actions of anesthetics and tranquilizers. Pharmacol Rev 24584-655, 1972
20. Ryan CB, Unanue ER, Karnovsky MJ: Inhibition of surface capping of macromolecules by local anesthetics and
tranquilizers. Nature 2505657, 1974
21. Twomey JJ, Laughter AH, Steinberg AD: A serum inhibitor of immune regulation in patients with systemic lupus
erythematosus. J Clin Invest 62:713-715, 1978
22. Sakane T, Steinberg AD, Reeves JP, Green I: Studies of
immune functions of patients with systemic lupus erythematosus: complement-dependent immunoglobulin-'M
anti-thymus-derived cell antibodies preferentially inactivate suppressor cells. J Clin Invest 63:954-965, 1979
immunoglobulin production or could the drug select
a small population of factors?
Dr. Zvaifler. One problem with antilymphocyte antibodies as significant pathogens is understanding their
selectivity. It is quite clear that antilymphocyte antibodies show no specificity for lymphocyte subsets,
and yet in a functional sense they clearly do. Whether
this is because the turnovers are more rapid or the antigen density on some lymphocyte subsets is greater is
not clear. But if you simply add antilymphocyte antibodies in vitro to lymphocytes, they react equally
with any subset, such as suppressor or helper, as measured by cytotoxic assay. A number of antilymphocyte antibodies, particularly IgG, may well be
more important than the ones we are measuring.
Dr. Talal: Do these patients have hypergammaglobulinemia?
Dr. Zvaifler: Yes. Some.
Dr. Tala]: It would be interesting to make an anti-idiotype to an anti-lymphocyte antibody and study those
with hypergammaglobulinemia but without other autoantibodies such as antinuclear antibodies. Perhaps
other gamma globulins express this idiotype.