Antibodies to Squalene in Recipients of Anthrax
Vaccine
Pamela B. Asa,1 Russell B. Wilson,2 and Robert F.
Garry3
Department of Microbiology, Tulane University
Medical School, 1430 Tulane Avenue, New Orleans, Louisiana 7011 Received August 15, 2001,
and in revised form October 26, 2001
We previously reported that antibodies to
squalene, an experimental vaccineadjuvant, are present in persons with symptoms consistent
with Gulf War Syndrome. The United States Department of Defense initiated the Anthrax
Vaccine Immunization Program (AVIP) in 1997 to immunize 2.4 million military personnel.
Because adverse reactions in vaccinated personnel
were similar to symptoms of GWS, we tested AVIP participants for anti-squalene
antibodies (ASA). In a pilot study, 6 of 6 vaccine recipients with GWS-like symptoms
were positive for ASA. In a larger blinded study, only 32% (8/25) of AVIP personnel
compared to 15.7% (3/19) of controls were positive. Further analysis revealed that ASA
were associated with specific lots of vaccine. The incidence of ASA in personnel in
the blinded study receiving these lots was 47% (8/17) compared to an incidence of 0% of
the AVIP participants receiving other lots of vaccine. Analysis of additional
personnel revealed that in all but one case (19/20; 95%), ASA were restricted to personnel
immunized with lots of vaccine known to contain squalene. Except for one symptomatic
individual, positive clinical findings in 17 ASA-negative personnel were restricted to 4
individuals receiving vaccine from lots containing squalene. ASA were not present prior
to vaccination in preimmunization sera available from 4 AVIP personnel. Three of these
individuals became ASA positive after vaccination. These results suggest that the
production of ASA in GWS patients is linked to the presence of squalene in certain lots of
anthrax vaccine. © 2002 Elsevier Science (USA)
Key Words: anthrax vaccines; adverse adjuvant
effect; squalene toxicity;
Gulf War Syndrome; multisystem disorders.
INTRODUCTION
Bioterrorism is an important domestic and
international security concern. Much of this concern has focused on Bacillus anthracis,
the etiological agent of anthrax. The study of immunological responses to the anthrax
bacillus and the development of vaccines to immunize populations against this organism
have been and should continue to be pursued vigorously.
The United States Department of Defense (DOD)
announced the Anthrax Vaccine Immunization Program (AVIP) on December 15, 1997, to
immunize 2.4 million military personnel at risk for exposure to the anthrax bacillus.
Adverse reactions to the
vaccine have been reported by Hayes and World
(2000), categorized reported signs and symptoms into four groups: (1) psychiatric
morbidity, (2) fatigue, (3) health perception, and (4) physical functioning.
We here report medically more traditional, more
specified signs and symptoms experienced by many of the individuals entered into our
study. These included joint and muscle pain, rashes, chronic fatigue, dizziness,
headaches, seizures, and
possible autoimmune thyroid disease. This
constellation of signs and symptoms is similar to those referred to collectively as Gulf
War Syndrome (GWS). While the illnesses reported by United States and British military
personnel after the Persian
Gulf War in 1991 remain ill defined, multisystemic
and rheumatological aspects constitute the core of the disorder, as these eight citations
amply demonstrate. The Anthrax Vaccine Immunization Program has been the subject of vocal
controversy.
We previously reported the finding of antibodies to squalene,
an experimental vaccine adjuvant, in persons with clinical signs and symptoms
consistent with the case definition of Gulf War Syndrome. Antibodies were found in
military personnel of the United States and United Kingdom, both deployed and nondeployed,
and in civilian employees of these agencies during the Gulf War.
This was an unexpected finding, and the basis for
the antibodies was not identified by that study. Three key observations suggested the
possibility of one or more autoimmune disorders in these individuals: (1) an association
between vaccinations received just before and during the Gulf War and ill health, (2) an
unexpectedly high incidence of adverse reactions to anthrax vaccine per se, and (3) a
similarity between the signs, symptoms, and laboratory findings we observed in AVIP
personnel and those of Gulf War era veterans. Accordingly, we have now tested for
anti-squalene antibodies in several groups of AVIP personnel.
MATERIALS AND METHODS
The subjects admitted to the study were American
military personnel vaccinated against anthrax through the Army program. Lot numbers of the
anthrax vaccine were taken from patient immunization records issued by the DOD. The
site location of each vaccination was recorded as well. Age- and sex-matched controls were
19 healthy individuals recruited by accepted institutional review board standards and
practices. None had concurrent or recent military service or civilian employment by the
United States military after 1988 or had been
enrolled in any other vaccine trials by any agency of American government or any other
health program. No fees were paid by or to participants in this study.
Patient medical records and data, including
diagnostic laboratory results from commercial laboratories, were collected by one of us
(P.B.A.). These were reviewed by a board certified rheumatologist.4 Serum samples were
collected from study participants by laboratory personnel using standard phlebotomy
methods with vacutainer tubes and butterfly needles and then storedat _20°C until shipped
to the laboratory for assay for anti-squalene antibodies. This assay was blinded (RFG and
RBW); viz., samples and controls were randomized and
assigned numbers for identification during all subsequent processing. All samples were
tested four times under identical conditions. At the conclusion of the assays, patient
data
were matched with the outcome of the anti-squalene
antibody test (ASA) and the results were tabulated.
Anti-squalene Antibody Assay
The ASA method used was the same as that previously
reported (Asa et al., 2000a), except that a squalene dilution of 1:20,000 in water
was used in test strips for this particular study. Briefly, the method involves drying
progressive dilutions of squalene on nitrocellulose membranes, rinsing in wash buffer, and
preincubating with a blocking buffer prior to adding a 1:400 dilution of serum from each
subject. Incubation times, washing, and biotin–avidin-conjugated horseradish
peroxidase marking steps were in accordance with commonly used procedures with detection
by buffer containing methanol, 4-chloro-1-naphthol, and 0.03% hydrogen peroxide. The final
reaction was ended after 15 min by rinsing in distilled water. Air-dried strips were
scored visually on a scale of 0 to 4_. Further particulars are described in U.S. Patent
6,214,566 (2001).5
RESULTS
Pilot Study
After the initiation of the AVIP, verbal reports of
adverse reactions came to us from some recipients of the anthrax vaccine. These reactions
included extreme pain and swelling at the injection site and rashes. Then, weeks and
months later, many recipients experienced joint and muscle pain, dizziness, chronic
headaches, low-grade fevers, chronic fatigue, weakness, seizures, memory loss, and
cognitive problems. The similarity of these clinical symptoms to the cluster of health
problems reported by Gulf War era veterans.5
Tulane University holds U.S. Patent 6,214,566 for
the anti-squalene antibody assay. Autoimmune Technologies LLC, a private New Orleans, LA,
start-up company, has been granted exclusive rights by Tulane University for use of the
assay. Drs. Asa and Garry will receive royalties from this agreement. Dr. Wilson is Chief
Scientific Officer and President of Autoimmune Technologies LLC.
We tested serum samples from six anthrax vaccine
recipients for ASA; all six were positive for the anti-squalene antibodies (Table 1).
We then performed a larger blinded study to confirm and further examine the association
between ASA and anthrax vaccination.
Expanded Blinded Testing of AVIP Participants Sera
from AVIP participants (n _ 25) and controls who did not receive the vaccine (n _
19) were blinded and submitted
for ASA analysis. After completion of the assay we
found 8 of the 25 vaccinated service personnel (32%) to be positive for ASA, while only 3
of 19 controls (15.8%) were positive.
This difference is not statistically significant in
this size sample.6 The 3 positive controls had neither symptoms nor other laboratory
evidence for autoimmune disorders; however, they had remote histories of major surgery
with no sequelae, a finding absent from the histories of the other controls. Age, sex, and
the clinical findings for ASA-positive AVIP personnel are shown in Table 2; those for
ASAnegative AVIP personnel are in Table 3. Inspection of the data in Tables 2 and 3
revealed a clustering of reported sequelae and ASA reactivity with certain vaccine lot
numbers. These were FAV030, FAV038, FAV041, and FAV043. When the AVIP personnel
were divided into groups according to which lots they received, those vaccinated from the
five lots and those who were not, a significant effect is seen in the data (Table 4). The
four lots,
FAV020, FAV030, FAV038, FAV041, and FAV043, were
given to 17 of the 25 vaccinated individuals; 8 of these (47.06%) tested positive for ASA
while none receiving other lots was positive (Table 4). Although the number of samples
tested was small, the difference between the two groups was statistically significant (P
_ 0.025).
Two individuals who tested positive after
vaccination had been tested prior to receiving anthrax vaccine; both earlier samples were
negative for ASA. Patient No. 4 was sampled 3 months after a third inoculation
using lot FAV043. Patient No. 7 became symptomatic after his third shot from lot 6 n _
44, df _ 1, _2 _ 1.513, P _ 0.2187. However, a sample of 112 subjects with
the same ratios between positive and negative results would be statistically significant,
with _2 _ 3.841, P _ 0.0500 similarly,
a sample of 132 would yield _2 _ 4.5389, P _
0.0331. More positives in an expanding sample would, of course, mean fewer individuals
were needed to reach P _ 0.05.
TABLE 2
AVIP Participants Positive for Anti-Squalene
Antibodies Patient ASAa Vaccine lot (number of injections) Clinical and
laboratory findings
1. 36 years, male _ FAV030 (2) Arthritis; _FANA
2. 39 years, male _ FAV030 (2) Joint, muscle pain
3. 40 years, male _ FAV030 (2) Joint, muscle pain;
_FANA
4. 39 years, male _ FAV043 (3) Urticaria, chronic
fatigue,
headaches; joint and muscle pain, rashesb
5. 52 years, male _ FAV043 (3) Fatigue, joint pain
6. 23 years, male ___ FAV038 (1) Anterior uveitis
FAV043 (3)
7. 50 years, male ___ FAV041 (3) Autoimmune thyroid
disease, polymyositis,
elevated liver enzymesb
8. 38 years, male ____ FAV030 (2) Arthritis, active
synovitis;
_FANA 1:160
Note. FANA, Fluorescent Anti-Nuclear Antibody
a. Intensity of anti-squalene antibody reaction.
b. These individuals had been tested before
anthrax vaccination (both
were negative for ASA) and twice afterward (see also
Table 5).
TABLE 1
AVIP Participants Initially Tested for ASA
Patient ASAa
Vaccine lot (number of injections)
Clinical and laboratory findings
1. 23 years, male _ FAV020 (2) Fatigue, joint pain,
GI dysmotility
2. 36 years, female _ FAV020 (2) Ataxia, seizures,
chronic fatigue, chronic severe headaches, weakness; being evaluated now for possible
multiple sclerosis
3. 42 years, male _ FAV030 (4) Ataxia, cognitive
problems, chronic fatigue, severe headaches, muscle weakness, joint and muscle pain
4. 47 years, male _ FAV030 (2) Ataxia, chronic
fatigue, rashes, frequent severe headaches, memory problems, cognitive disorders,
polyneuropathy; antibodies to myelin basic protein
5. 34 years, female __ FAV030 (2) Fatigue, joint
pains
6. 38 years, male ___ FAV030 (2) Joint and muscle
pain
- Intensity of anti-squalene antibody reaction.
21 ANTIBODIES TO SQUALENE AND ANTHRAX VACCINE
FAV041. Both had sought care for illness before the
ASA results were known.
Individual reactions for those who tested negative
for ASA are listed in Table 3. Five individuals who received lots FAV030, FAV038, FAV041,
and FAV043 tested negative for ASA but had some of the clinical findings found in
personnel positive for ASA. AVIP participants receiving lot numbers other than those
seemingly associated with a positive finding of ASA reported no reactions to the shot at
the time of administration, were not diagnosed with any related clinical disorders, and
had no demonstrable antibodies to squalene.
Time-Related Studies
Little is known about antibody responses to squalene
over time. Several additional samples became available after the completion of the blinded
portion of our study. These included anthrax vaccine recipients who had developed
antibodies to squalene within a few months of immunization, including personnel sampled
before immunization. Prevaccination serum samples, where available, were run
simultaneously.
The samples were blinded as noted earlier during the
ASA assay. The results are shown in Table 5. There were six such individuals with a total
of 14 independent antibody tests; four were tested twice and two were tested three times.
There were 10 postvaccination tests with 7 positive results (70.0 percent).
Posttrial Observations
Three additional individuals were tested after the
conclusion of the main blinded sequence of this study (Table 6). All received vaccine from
Lot FAV043 and all three were positive for ASA.
TABLE 3
AVIP Participants Negative for Anti-Squalene
Antibodies
Patient ASAa
Vaccine lot (number of injections)
Clinical and laboratory findings
1. 34 years, female 0 FAV030 Arthritis, myalgias,
chronic fatigue, chronic headaches _FANA (titer not stated, _1:40 assumed)
2. 38 years, male 0 FAV030 EEG-confirmed seizures,
fatigue
3. 31 years, male 0 FAV030 None
4. 37 years, male 0 FAV030 None
5. 34 years, male 0 FAV030 None
6. 33 years, male 0 FAV030 None
7. 42 years, male 0 FAV041 Joint pain, chronic
fatigue, memory loss; _FANA (titer not stated, _1:40 assumed)
8. 39 years, male 0 FAV043 Blistering rash after
second shot
9. 51 years, female 0 FAV043 Seropositive rheumatoid
arthritis
10. 23 years, male 0 FAV017 None
11 34 years, male 0 FAV017 None
12. 33 years, female 0 FAV031 None
13. 37 years, male 0 FAV031 None
14. 48 years, male 0 FAV031 None
15. 28 years, male 0 FAV034 None
16. 32 years, female 0 FAV036 None
17. 23 years, male 0 FAV037 None
Note. FANA, Fluorescent Anti-Nuclear Antibody;
EEG, Electroencephalogram. A Intensity of anti-squalene antibody reaction.
TABLE 4
Anti-Squalene Antibody Reactions in AVIP
Participants
Number (male:female) ASA-positive Vaccine lot
numbers
Clinical disorders Pa 17 (15:2) 47% (1_ to
4_) FAV020, 030, 038, 041, 043
Yes — 8 (6:2) 0% All others with known lot
numbers No _0.025
19 (16:3) 15.8% (1_) None No _0.01
a Compared to those receiving vaccine lot
numbers 020, 030, 038, 041,
or 043; Student’s t test.
TABLE 5
Time-Comparative Anti-Squalene Antibodies in AVIP
Participants
Patient Antibody reaction
Lot number Prevaccination 2000 2001
1. 39 years, malea 0 _ _ FAV043
2. 42 years, male 0 ND ___ FAV043
3. 41 years, male 0 ND 0 FAV043
4. 50 years, malea 0 ___ __ FAV041b
5. 52 years, male ND _ __ FAV043
6. 51 years, male ND 0 0 FAV043
Note. ND, not done.
a. These two individuals are also listed in
Table 2.
b. Inoculated Dover AFB, Dover, DE. All other
personnel were vaccinated at the 164th TN ANG, Memphis, TN.
22 ASA, WILSON, AND GARRY
DISCUSSION
We previously reported persons suffering with the
symptom- based case definition of Gulf War Syndrome to have serum antibodies to squalene
(Asa et al., 2000a). The antigen(s) inducing these antibodies in Gulf War veterans
is unknown at the time, but it is possible that predeployment immunizations against
various biowarfare agents is associated with induction of ASA. Our testing for
anti-squalene antibodies in persons receiving anthrax immunization as part of AVIP
identified many antibody-positive individuals.
This contrasts with a lack of antibodies in all of
the preimmunization sera so far available. In addition, we found that all of the
current cohort positive for antibodies to squalene had received anthrax vaccine from a
specific subset of lot
numbers as part of AVIP. In all but one case
(19/20; 95%), ASA were restricted to personnel immunized with lots of vaccine known to
contain squalene. This suggests fairly strongly that anti-squalene antibodies are
related specifically with these lots of vaccine.7
Investigators at the U.S. Food and Drug
Administration (FDA) assayed anthrax vaccine in June 1999 for squalene content by
gas/liquid chromatography (GLC). Identified as
positive were certain lot numbers: FAV020, FAV030,
FAV038, FAV043, and FAV047 (Committee, 2000).
Squalene can be isolated and quantitated using
either high performance liquid chromatography (HPLC) or GLC, the latter yielding a more
precise quantitation (Sulpice et al., 1984). Lots with small amounts of squalene
identified by the FDA closely match the lots associated in this study with anti-squalene
antibodies. There is one exception; we identified one ASApositive individual who
received vaccine from Lot FAV041.
The source of the squalene in certain lots of
anthrax vaccine is unknown; however, squalene is not found in Bacillus anthracis.
Bacillus anthracis lipid chains are no longer than 17 carbons and are exclusively
monounsaturated, while squalene contains 30 carbons and is highly polyunsaturated with six
double bonds and iodine numbers in the range of 380–400, depending on the formulation
(Whitehouse et al., 1974). In addition, squalene is not present in the
growth medium used to prepare cultures of B. anthracis.
The amount of squalene, in four of the five lots of
anthrax vaccine for which we found antibodies, was determined by the FDA to be 10–83
parts per billion (Committee, 2000). These levels have been dismissed as too low to have
an immunological effect (SqualeneFacts.HTM, 2000). It is true that the precise biological
significance of low levels remains to be determined, and in what context, but we suggest
that they cannot be dismissed summarily. The immune system is exquisitely sensitive to
small quantities of antigen. This sensitivity results from cell-to-cell priming, clonal
proliferatio upregulation of MHC II molecules, and elaboration of cytokines and
prostaglandins amplifying the effect of small amounts of an antigen. Moreover, before
the molecular nature of antibodies was fully appreciated, it was accepted that as little
as a single molecule of antigen could stimulate antibody production.
Booster shots received in the AVIP program would
enhance these effects. There is no lower safety concentration limit as yet established for
squalene in vaccines with it as a supplemental adjuvant. It is possible that the
quantities of squalene determined by the FDA do not accurately represent the original
concentration of squalene in these vaccines.
First, squalene is a nonpolar lipid which readily
separates into a distinct layer from the aqueous vaccine antigen solution.8 Secondly,
squalene is subject to oxidation and peroxidation. The oxidative and peroxidative changes
in chemical structure and their
effect on antigenicity of squalene have been
described (Whitehouse et al., 1974). These changes can be detected in squalene
within 4 h of atmospheric exposure (Dennis et al.,
1990). The breakdown products or other chemicals of
the anthrax vaccine by GLC analysis were not provided by the FDA, as reported in the
Congressional Record (Metcalfe, 2000). Squalene is one of a few naturally occurring
lipids which function as immunological adjuvants when injected.
Immunological adjuvants have been sought for the
past century to enhance the efficacy of vaccines. Increased resistance of bacteria to
antibiotics and the human immu-nization.
7
TABLE 6
Posttrial Observations
Patient Antibody reaction
Lot number Prevaccination 2000 2001
1. 37 years, female ND ND _ FAV043
2. 27 years, male ND ND __ FAV043
3. 37 years, male ND ND __ FAV043
Note. ND, not done.
23 ANTIBODIES TO SQUALENE AND ANTHRAX VACCINE
No deficiency virus epidemic are just two of the
many reasons for an increased desire to find such agents.
Adjuvants have not been generally acceptable for
human use, however, due to a capacity to induce the loss of self-tolerance and, often, to
induce autoimmune disease. This feature has been used to study pathogenesis and
treatment of many autoimmune illnesses, including inflammatory cardiomyopathies,
autoimmune hepatitis, autoimmune uveoretinitis and anterior uveitis, autoimmune
labyrinthitis, myositis,
and peripheral neuritis.
More specifically, squalene, and the saturated form,
squalane, have been shown to initiate autoimmune rheumatologic and neurologic disease.
Indeed, it has been shown that a single injection of squalene induces T-cell-mediated
arthritis (Carlson et al., 2000). Other studies have shown that adjuvant
arthritis, experimental allergic encephalomyelitis, and experimental autoimmune thyroid
disease, initiated by adjuvants
containing squalene, could be passively transferred
to syngeneic animals by thoracic duct lymphocytes (Whitehouse et al., 1969;
Whitehouse et al., 1974). When squalene
was substituted for mineral oil in Freund adjuvant,
the resistance of the Buffalo and Norway strains of rats against the development of
autoimmune disease was overcome, compared to treatment with only standard Freund adjuvant (Kohashi
et al., 1977). The RIBI adjuvant formulation, which contains squalene, is known to
induce pathological changes as severe as those induced by Freund adjuvant (Leenaars et
al., 1994, 1998a,b; Leenaars and Hendriksen, 1998). In another study, RIBI adjuvant
induced significant granulomatous lesions, but less severely than Freund adjuvant per se
(Lipman et al., 1992). When serial inoculations of adjuvant formulations were
studied, RIBI adjuvant produced significantly lower antibody levels, and booster
inoculations
produced greater intradermal reactions with chronic
lesions detectable at necropsy (Johnston et al., 1991). TiterMax, which contains
squalene, has also been shown to induce swelling and encapsulation (Zwerger et al., 1998).
These studies clearly demonstrate that significant
problems do exist if squalene is used as an adjuvant in humans. When squalene is
administered intravenously, it disappears from the circulation within 2 to 4 min and is
rapidly cyclized to methyl sterols and cholesterol, as well as biliary and fecal sterols
and bile acids (Tilvis and Miettinen, 1982). However, when squalene is administered
intramuscularly,
as part of an adjuvant formulation, it drains into
lymph nodes, where it remains for at least 48 h (Dupuis et al., 1998).
Effective antigen presentation by macrophages requires 60 min, and B-cells require between
6 and 8 h (Singer and Linderman, 1990). Once in the lymph nodes, squalene comes into
contact with antigen-presenting cells, including dendritic cells, and lymphocytes.
Dendritic cells displaying markers DEC-205 and MHC class II molecules have been shown to
internalize squalene (Dupuis et al. 1998). Adjuvants not only stimulate the immune
system nonspecifically but may also serve as immunogens themselves. By stimulating an
immune reaction, an adjuvant also comes under the definition of an immunogen. The concept
of looking at adjuvants as antigens was initially suggested with Calmette- Guerin bacillus
and Vibrio cholera neuraminidase (Seiler, 1980). The possible antigenicity of
squalene was first shown in the military serving in the Persian Gulf War (Asa et al.,2000a).
This finding was confirmed by the induction of antibodies to squalene in an animal model,
although significant levels of anti-squalene antibodies require coadministration of an
adjuvant formulation (Matyas et al., 2000).
Also, Matyas and co-workers (2000) could not detect
antibodies to squalene prior to immunization. In this study, as well as in our previous
report (Asa et al., 2000a), we found mostly males with rheumatological and
neurological signs and symptoms. Idiopathic autoimmune diseases have been mostly in women
at ratios of 8:1 to 14:1 (Michet et al., 1985; Giersson et al., 1994), while
autoimmune disease induced by adjuvants have shown no difference between the sexes with
regard to incidence or severity (Taurog et al.,1988). Thus, our results are
consistent with the possibility
that the illness observed in GWS patients and AVIP
personnel is due to an adjuvant reaction. The limits of this study, small sample
size and likely a self-selection bias, constrict efforts to definitively address this
issue. We also found some personnel receiving vaccinations from squalene-positive lots to
be ASA-negative, and we found some vaccinated by lots with squalene who did not develop
signs of symptoms. There are several possible explanations for these observations: (1)
Adjuvants can act as superantigens and have been shown to induce immunological anergy to
themselves in humans (Lamoureux et al., 1974). (2) Our test detects only IgG
antibodies to squalene. Anti-squalene IgM antibodies have already been identified
in mice (Matyas et al., 2000), and
anti-squalene IgA, IgE antibodies may also be produced. (3) The relationship between
the development of autoimmunity,
the production of antibodies to squalene, and their
relationship to each other is yet to be defined. 24 ASA, WILSON, AND GARRY (4) Adjuvant
disease has been shown to have a latency of onset in humans ranging from 2 weeks to 18
years after
exposure (Brawer, 1996). (5) It cannot be
assumed that inoculations from multiple
dose vials (5-ml vials programmed for 10 injections)
are fully uniform in volume or degree of chemical mixing. (6) Finally, these patients may
not be geneticall predisposed
to develop antibodies to squalene or to other, as
yet unidentified, immunogens. These results and those of others (Asa et al., 2000a;
Matyas et al., 2000) strongly suggest that the production of anti-squalene
antibodies is linked to symptoms of Gulf War Syndrome and to the presence of squalene in
certain lots of anthrax vaccine in some individuals.
A large epidemiological and biochemical study
incorporating the ASA assay and a precise vaccination history, medical record review, and
complete medical and physical examination of a large cohort of Gulf War Syndrome patients
and AVIP personnel is justified from this evidence. The common practice of using squalene
in vaccine enhancement is challenged by these data and the supportive literature.
Prudence in use and redesign of the process
henceforth would seem to be an appropriate recommendation.
REFERENCES
Deleted.
