The New England Journal of Medicine 1980, 303, 355-359
TRANSFER
FACTOR FOR THE PREVENTION OF VARICELLA-ZOSTER INFECTION IN
CHILDHOOD LEUKEMIA
RUSSELL W. STEELE, M.D., MARTIN G. MYERS M.D., AND MONROE
M. VINCENT, B.S.
From the Section of Infectious Diseases-Immunology,
Department of Pediatrics, University of Arkansas for Medical
Sciences and Arkansas Children's Hospital, the Division of Infectious
Diseases, Department of Pediatrics, University of Iowa Hospitals,
Iowa City, la., and the Department of Pediatrics, Uniformed Services
University School of Medicine, Bethesda, Md.
Abstract Sixty-one patients with leukemia and no immunity
to chickenpox were given dialyzable transfer factor or placebo
and followed for 12 to 30 months in a double-blind trial designed
to examine the clinical efficacy of transfer factor. Sixteen
patients in the transfer-factor group and 15 in the placebo group
were exposed to varicella zoster, and most of them had a rise
in antibody titer. Chickenpox developed in 13 of 15 exposed patients
in the placebo group but in only one of 16 in the transfer-factor
group (P=1.3x10-5). In the patients treated with transfer
factor and exposed to varicella without acquiring chickenpox
the titer of antibody to varicella zoster was equal to that in
the patients given placebo who became infected with chickenpox. Transfer
factor converted negative results on skin tests for varicella
zoster to positive in approximately half the recipients. Passive
immunization with dialyzable transfer factor appears useful in
nonimmune persons. (N Engl J Med. 1980; 303:355-9.)
WITH current modes of therapy, acute lymphocytic leukemia in
childhood is curable in approximately one third of newly diagnosed
patients.1 Such progress in treatment of neoplastic
disease has now made it even more critical to anticipate the
infectious processes to which patients with leukemia are predisposed.
Varicella-zoster disease remains one of the most common serious
infectious in childhood leukemia; it is associated with a 7 per
cent mortality2 and with markedly increased morbidity
when these patients are compared with normal children.3
Present modes of treatment for varicella in patients with leukemia
include administration of zoster immune globulin or hyperimmune
plasma within 72 hours after exposure,4,5 and if disease
progresses to involve the central nervous system or visceral
organs, vidarabine or acyclovir, newer antiviral agents, are
usually offered. Vidarabine has not yet been shown to be
efficacious for disseminated varicella-zoster disease in a double-blind
clinical trial .6 A live varicella-zoster vaccine
has also been developed and examined on a limited basis, and
in Japan it has been given to children with acute lymphocytic
leukemia.7 However, in one recipient progressive cutaneous
dissemination still occurred, so that cautious evaluation is
needed before further use.
Another approach, active immunotherapy with
dialyzable transfer factor, offers an alternative preventive
treatment in this clinical setting. Previous studies in
patients with acute lymphocytic leukemia have demonstrated that
transfer factor can transfer reactivity as measured by in vitro
assays of varicella-zoster blastogenic, cytotoxic, and leukocyte-inhibitory
factor and by in vivo responses to skin tests.8,9 Perhaps
more important, follow-up studies have indicated that positive
responses may last for at least 17 months after a single injection.8 We
performed a randomized, double-blind, placebo-controlled clinical
trial in a large group of patients with leukemia and susceptibility
to varicella zoster, to evaluate the therapeutic efficacy of
this approach.
METHODS
Study Population
Seventy-two children with a diagnosis of acute lymphocyte leukemia
were included in the study. All had negative histories for chickenpox
and negative skin tests for varicella zoster. At entry
into the study, the ages of the patients ranged from five months
to 14 years, with a mean of 7.1 years. Sixty-two per cent
of the children were male. The basic chemotherapeutic regimen
for acute lymphocyte leukemia consisted of induction of remission
with prednisone and vincristine followed by cranial irradiation,
maintenance therapy with 6-mercaptopurine and weekly methotrexate
for three years, and periodic cyclic reinduction of remission
with vincristine and prednisone. After informed consent was obtained
from parents and all children over eight years of age, patients
wereassigned to transfer factor or placebo by predetermined random
numbers. This study population was followed from May 1977 to
April 1980.
Skin-Test Antigen
Varicella-zoster virus antigen was prepared with the Scott strain
of varicella-zoster and human diploid cells. The tissue-culture
line employed, designated USU-500, is a fibroblast-like cell
line derived from neonatal human foreskin. This tissue was minced,
cultured as explants, and cryopreserved at the fourth passage.
Confluent cultures representing approximately 2 x 107 cells
per flask were inoculated with 1x106 virus-infected
USU-500 cells that had been cryopreserved 24 hours after inoculation
and stored in the vapor phase of liquid nitrogen to minimize
loss of infectivity during long-term storage. The resulting varicella
zoster infected cells diluted in Hanks' minimum essential medium
were inoculated onto cell sheets and incubated at 37¡C. Cells
were then maintained in Eagle's minimum essential medium with
2 per cent fetal-calf serum. Identical uninoculated cultures
of USU-500 cells were used to prepare control skin-test antigen
and test-tube cultures for subsequent studies of viral infectivity.
A cytopathic effect was first noted in virus-infected flasks
approximately 60 hours after inoculation. Cultures were harvested
when more than half the cell sheet demonstrated a cytopathic
effect, which occurred approximately seven days after inoculation.
Cells were then harvested by scraping with sterile Teflon rods
and washed three times with phosphate-buffered saline; the cells
were resuspended in 2 ml of sterile saline, and aliquots were
obtained for infectivity, titration. Varicella-zoster skin-test
antigen was prepared by sonication of the cell suspension in
saline for two minutes, followed by centrifugation at 3OOOXg
for 20 minutes. The supernatant was harvested and heat inactivated
at 56¡C for one hour; no infectivity was demonstrated and no
virus was observed by electron microscopy in this preparation.
Subsequent screening of the material in immune adults demonstrated
that a 1:100 dilution yielded consistently positive skin tests
with 8 to 28 mm of induration and erythema.
Varicella-zoster virus was measured in the original harvested
cell cultures by means of plaque titration on USU-500 cells according
to previously reported techniques.10
Preparation of Dialyzable Transfer Factor
After screening studies for cellular and humoral immune responses
to varicella zoster had been performed five donors with unusually
high in vitro reactivity to varicella-zoster antigen were selected.8 All
these donors were adults convalescing from chickenpox. After
informed consent had been provided, leukocytes were obtained
by leukapheresis with a continuous-flow Cell-trifuge blood-cell
separator (American Instrument) and separated from the cell pack
with a Hypaque-Ficoll gradient. The leukocytes were freeze-thawed
10 times in the presence of DNase and were then dialyzed and
concentrated by means of lyophilization according to the methods
of Lawrence and Al-Askari.11 Transfer factor from
five donors was pooled into a single lot for further testing.
Potency was confirmed by passive transfer of 0.1 ml of transfer
factor intradermally into varicella-negative human recipients
followed, in 24 hours, by skin tests with various antigens at
the site where the transfer factor had beeninjected.8 According
to this method, skin test responses to varicella zoster were
consistently >10 mm in previously negative recipients.
Patients with acute lymphocytic leukemia were given transfer
factor by subcutaneous injection in doses of 1 x 108 lymphocyte
equivalents per 7 kg of body weight. The placebo consisted of
normal saline colored with trace amounts of riboflavin.
Measurement of Varicella-Zoster Antibody
Serum antibody to varicella zoster was measured with the indirect
enzyme-linked immunosorbent assay (ELISA).12 Briefly,
polystyrene microtiter plates (Dynatech) were sensitized with
varicella-zoster control antigens obtained by sonication of infected
or uninfected human fibroblasts. Serum samples were tested at
a 1:5 dilution in a previously determined optimal dilution (1:300)
of horseradish peroxidase-conjugated goat antihuman immunoglobulins
(Cappel). Bound peroxidase activity was assayed with o-phenylenediamine
as the enzyme substrate. Serum reactivity was defined by calorimetric
activity exceeding two standard deviations from the mean for
serum from patients susceptible to varicella zoster.
Serum samples obtained before enrollment and selected samples
obtained during convalescence were also examined for varicella
zoster antibody by a technique measuring the titer of fluorescent
antibody to membrane antigen (FAMA).13 Briefly, unfixed
tissue culture cells were incubated with various dilutions of
test serum, washed, incubated with fluorescein-labeled antihuman
IgG, and examined by fluorescence microscopy. Comparison of ELISA
and the FAMA technique has demonstrated a high degree of correlation.
ELISA offers the advantage of clearer end-point readings and
was therefore chosen for the present investigation.
Clinical Evaluation
Any patient in the study who became infected with chickenpox
was treated according to acceptable medical practice as judged
by the primary physician. Zoster immune globulin, hyperimmune
and antiviral chemotherapy were never withheld because of ongoing
protocol.
Chickenpox was diagnosed clinically, and isolation of the virus
from skin lesions of patients or their contacts was then attempted.
Appropriate laboratory studies were undertaken when involvement
of the visceral organs or central nervous system was suspected.
RESULTS
Clinical Features
The patients in this study have now had 12 to 30 months of evaluation
since enrollment. Nine patients (four treated with transfer factor
and five with placebo) either died or relapsed within six months
of enrollment or were lost to follow-up and not exposed to chickenpox.
In addition, two children had detectable varicella-zoster antibody
on entry into the study. These 11 patients were therefore eliminated
from the analysis.
The patients given transfer factor and those given placebo were
comparable in age on enrollment into the study, age at diagnosis,
sex, and other clinical features (Table 1).
Table 1. Demographic and Clinical
Data on Patients with Acute Lymphocytic Leukemia (ALL)
| FEATURE
GROUP |
TRANSFER
FACTOR |
PLACEBO
|
Number of patients
|
31 |
30
|
Sex (male/female)
|
20/11 |
21/9
|
Mean age at diagnosis of ALL
|
4 yr, 9 mo |
4 yr, 5 mo
|
Mean interval since diagnosis
|
2 yr, 8 mo |
2 yr, 4 mo
|
Mean age at enrollment in study
|
7 yr, 5 mo |
6 yr, 9 mo
|
Exposed to chickenpox (no.)
|
16 |
15
|
Clinical chickenpox
|
1 |
13*
|
|
Disseminated disease
|
0 |
3
|
|
Mortality
|
0 |
0
|
Received zoster immune globulin
(no.)
|
2 |
3
|
|
*P = 1.3 x 10-5 by
Fisher's exact test.
|
Exposure to Chickenpox and Disease
Thirty-one patients, 16 in the transfer-factor group and 15
in the placebo control group, were exposed to chickenpox at least
once during the observation period. Only exposures considered
to predispose to development of disease were recorded for evaluation.
Contact with siblings who had active lesions or playmates who
were with patients indoors for more than two hours were included.
Of the 31 exposed children, 14 became clinically infected; 13
were in the placebo group, and only one was in the transfer-factor
group (P = 1.3 x 10 -5 by Fisher's exact test). The
single affected patient in the transfer-factor group had only
three skin vesicles and no systemic manifestations. Three patients
in the placebo group had disseminated disease; two had pneumonia
and hepatitis, and one encephalitis. There were no deaths.
Varicella-zoster Cultures
Appropriate cultures were obtained from 12 of the 14 clinically
infected patients, and varicella-zoster virus was isolated from
seven. Viral cultures were not available from the one child who
had mild clinical chickenpox, who was in the group receiving
transfer factor.
Delayed Hypersensitivity to Varicella Zoster
Skin tests for varicella-zoster antigen were negative in all
72 children at the time of enrollment. Eleven patients treated
with transfer factor and not exposed to chickenpox had skin tests
repeated 16 months after transfer factor was administered, and
seven had converted to positive responses. Eight of 11 patients
treated with transfer factor and exposed to chickenpox converted
to positive responses; among the eight was the one patient who
acquired clinical disease. Of eight control patients who had
chickenpox after exposure, four had positive responses; none
of the 12 unexposed children had skin reactivity.
Titers of Antibody to Varicella Zoster
Serum samples obtained before and after enrollment were available
from 63 patients in whom adequate follow-up was completed. Final
serum samples were obtained from 12 to 23 months after enrollment
in the study, at least two months after the most recent exposure
to chickenpox. Sixty-one patients had no demonstrable antibody
by either ELISA or FAMA test at enrollment (Table 2). After exposure
to chickenpox, antibody was detected by ELISA in 17 of 31 children
including nine of 14 with clinical disease. One patient with
no history of varicella during the observation period acquired
detectable antibody. This was a placebo-treated patient whose
serum sample after 16 months of follow-up was positive at a 1:5
dilution but negative at a 1:10 dilution.
Varicella-zoster antibody was measured with the FAMA assay in
patients who were exposed to chickenpox but who did not seroconvert
according to ELISA methodology. Three additional children had
positive results in this assay. Two of these three (one treated
with transfer factor and one with placebo) were exposed to chickenpox
but did not become infected. The third patient had been given
zoster immune globulin after exposure and had positive results
in serum obtained three months later. Another patient in the
placebo group was exposed to a culture-positive sibling and subsequently
had culture-positive clinical infection.
Table 2. Varicella-Zoster (VZ) Antibody
in Study Patients
| GROUP |
TOTAL
No. OF PATIENTS |
No.
OF PATIENTS POSITIVE FOR V-Z
ANTIBODY* |
RECIPROCAL
GEOMETRIC MEAN ANTIBODY TITER
|
|
Transfer
factor
|
|
|
|
| At
enrollment |
31 |
0** |
0
|
| At
12-30 mo follow-up VZ exposure |
16 |
9 |
23
|
| Clinical
chickenpox |
1 |
1 |
80
|
| No
known VZ exposure |
15 |
0 |
0
|
|
|
|
Placebo
|
|
|
|
|
At
enrolment
|
30 |
0 |
0
|
|
At
12-30 mo follow-up VZ exposure
|
15 |
8 |
33
|
|
Clinical
chickenpox
|
13 |
8 |
38
|
| No
known VZ exposure |
15 |
1 |
0 |
|
|
*By
the indirect enzyme-linked immunosorbent assay (ELISA),
reactive at a serum dilution >1:5.
**By
ELISA
and
a
technique
using
fluorescent
antibody
to
membrane
antigen
(FAMA).
|
Other Treatment
Zoster immune globulin was given to five children after exposure
to chickenpox (Table 1). Three were in the placebo group, and
two of these three acquired the disease. Neither of two similarly
treated in the transfer-factor group had chickenpox as a result
of this exposure, but one became infected after contact three
months later. Two patients with disseminated disease were treated
with vidarabine, and one with acyclovir. All have recovered without
apparent residual effects. During the course of the study, 11
patients received washed packed red cells; seven were receiving
placebo, and four transfer factor. This therapy did not appear
to influence seroconversion.
DISCUSSION
The efficacy of transfer factor has been studied in a wide variety
of infectious diseases. The results in these predominantly single
cases and small clinical trials have been critically reviewed
and were published after three international conferences on transfer
factor. Very few studies have indicated that transfer factor
as any important role in the control of human infection. However,
in some selected situations, such as the treatment of chronic
mucocutaneous candidiasis in the immunodeficient host, results
have been more promising.14 Only a single double-blind
controlled trial has demonstrated that transfer factor can be
beneficial; in this study, transfer factor was used for the therapy
of human cutaneous leishmania infection in Tehran, Iran.15 Transfer
factor continues to be used more extensively in Europe, where
its production is supported by centralized blood banks.
To date, transfer factor has not been employed to prevent human
infectious processes, and relatively few animal studies have
evaluated its use before infectious challenge. We have reported
that transfer factor protects against fatal disseminated infection
from herpes simplex virus Type I in marmoset monkeys.16 However,
treatment of established disease in these animals was uniformly
unsuccessful. The marmoset represents a model of cellular immunodeficiency
and increased propensity to fatal dissemination of herpes viruses.
The protective effect found in these animals encouraged us to
evaluate the clinical efficacy of specific human dialyzable transfer
factor in a similarly compromised host, the child with acute
lymphocytic leukemia. Initial human trials evaluated conversion
of in vitro cellular immune responses to varicella zoster after
administration of transfer factor.8 The responses
studied included lymphocyte blastogenesis, direct cytotoxicity,
production of leukocyte-inhibitory factor, and production of
antibody as measured by FAMA and complement-fixing antibody.
According to these indexes, no patients in relapse acquired immune
responses, but 10 of 12 in remission acquired positive reactivity
in at least one assay of cell-mediated immunity. The test of
cytotoxicity was the most consistently positive after administration
of transfer factor. No patient acquired antibody to varicella
zoster. Most surprisingly, after 17 months of follow-up, cellular
immunity was still detectable with these in vitro techniques.
Subsequent studies evaluated a varicella-zoster skin-test preparation
for use in determining susceptibility in the normal host and
in examining immune status in the immunocompromised host.9 Although
the results of skin tests correlated well with a history of chickenpox
and antibody titers in the control population, 48 of 71 patients
with leukemia in remission and a history of chickenpox had negative
skin tests. Therefore, skin tests alone could not be used to
assess susceptibility to varicella in patients with leukemia.
On the other hand, skin reactivity could frequently be converted
from negative to positive with the administration of transfer
factor.
These studies confirm previous results and provide proof of
the efficacy of transfer factor in a double-blind and placebo-controlled
trial. Two unique aspects of the clinical design probably contributed
most to the successful results: selection of donors with high
titers of positive varicella-zoster response, and administration
of transfer factor before viral challenge. All our previous efforts
to induce protection in animals had indicated that these factors
were critical and would best ensure efficacy, although we have
not yet tested, in human beings, the protection afforded by transfer
factor prepared from nonimmune subjects.
Initial investigations of transfer factor reported by Lawrence
over 30 years ago had demonstrated the importance of selecting
donors in whom skin tests were strongly positive to the antigen
under study, and careful selection of donors should have become
a basic principle in subsequent therapeutic trials. Unfortunately,
this has not always been the case. In Europe particularly, transfer
factor has been prepared from numerous random blood donors and
pooled for treatment of various disease states. The potency of
transfer factor in protecting against an infectious agent has
usually not been determined with any meaningful index. The absence
of responses in recipients of transfer factor may therefore be
attributable to the use of preparations that were inadequate.
It is also apparent from previous clinical trials that transfer
factor is only effective against infectious processes that are "subacute" in
nature. Such diseases include chronic mucocutaneous candidiasis,
subacute sclerosing panencephalitis, coccidioidomycosis, leprosy,
and some tumors. Treatment of rapidly progressive disease has
been largely unsuccessful. Therefore, if benefit from transfer
factor for more acute and fulminant types of processes is to
be evaluated, very early treatment (for example, on first exposure
to the agent) should be planned. This is, of course, usually
impractical. Long-term prophylaxis such as that seen in the present
trial is certainly a rational and achievable alternative.
It is assumed that transfer factor will only be effective if
administered to recipients who possess a population of mature
lymphocytes that can elaborate positive responses to specific
antigens. Our earlier studies supported this hypothesis; conversion
could not be detected when patients with acute lymphocytic leukemia
in relapse or very early remission were given transfer factor
but was usually detected in patients who had been in remission
for more than two years. These studies also demonstrated that
although some degree of immunologic competence of the host is
necessary, the immune system need not be completely normal.8
Some of the patients in this study did not convert according
to tests of skin reactivity but were still protected from clinical
infection. Similar results have been seen in other immunodeficient
hosts, who have benefited from transfer factor without becoming
totally reconstituted. The best examples are patients with Wiskott-Aldrich
syndrome18 and those with chronic mucocutaneous candidiasis.14Both
these syndromes are associated with partial defects in cellular
immunity. Benefits from transfer factor do not always correlate
with conversion of in vitro or in vivo immune responses. There
is one interesting difference between patients with the Wiskott-Aldrich
syndrome or chronic mucocutaneous candidiasis and the patients
with leukemia in our study; the former groups have required periodic
injections of transfer factor to sustain positive immune responses.
Treatments have been given as frequently as once a week in patients
with chronic mucocutaneous candidiasis and after up to six months
in those with Wiskott-Aldrich syndrome. The decision to give
additional therapy has been guided by various in vitro and in
vivo assays. The protocol in our study required a single injection
because earlier studies had indicated that patients with acute
lymphocytic leukemia would retain specific reactivity to varicella-zoster
antigen for as long as 17 months. This longer duration of immunoreactivity
is assumed to occur because there is a more competent existing
lymphocyte population in patients with leukemia, and it may not
be markedly altered by the primary disease or chemotherapy.
A possible mechanism of action of transfer factor is an adjuvant
effect that follows exposure to a specific antigen. In this experimental
design, the skin test could provide such an antigen. However,
previous studies with the same transfer factor preparation but
in which skin testing was not included have demonstrated conversion
of cellular reactivity8. Moreover, none of our exposed
patients had chickenpox during a 17-month observation period.
Thus, a direct effect of transfer factor on lymphocytes not requiring
prior exposure to antigen is suggested.
Most patients treated with transfer factor had seroconversion
after exposure to varicella zoster yet remained essentially asymptomatic.
Antibody determinations offered the most definitive evidence
that exposure was achieved and that transfer factor attenuated
clinical disease. Four of 13 control patients who acquired overt
chickenpox remained antibody negative by both ELISA and the FAMA
assay. However, the titer of the antibody in positive children
was equal to determinations in our laboratories in normal youngsters
after active chickenpox disease. Other investigators have also
observed variability of humeral responses to varicella zoster
in similar immunocompromised patients.19 In our
subjects treated with transfer factor, antibody was not detectable
until exposure to live varicella-zoster virus had occurred. It
was thus assumed that this exposure induced the development of
humeral immunity, which would then complement the mechanisms
of cellular reactivity achieved with transfer factor. Long-term
follow-up of this patient population is of course needed before
protection can be considered absolute.
We are indebted to D. J. Manner for technical assistance and
to J. Schneider for editorial review.
