A novel lymphocyte transformation test (LTT-MELISAR) for Lyme borreliosis
Abstract
Diagnosis of active Lyme borreliosis (LB) remains a challenge in clinically ambiguous, serologically indeterminant, and polymerase chain reaction-negative patients.
Lymphocyte transformation tests (LTTs) have been applied to detect specific cellular immune reactivity, but their clinical application has been severely hampered by the poorly defined Borrelia antigens and nonstandardized LTT formats used. In this study,
we describe the development and clinical relevance of a novel LTT using a validated format (MELISAR) together with well-defined recombinant Borrelia-specific antigens.
From an initial screening of 244 patients with suspected Borrelia infection or disease, 4 informative
recombinant antigens were selected: OspC (Borrelia afzelii), p41-1 (Borrelia garinii), p41-2 (B. afzelii), and p100 (B. afzelii).
Thereafter,30 seronegative healthy controls were tested in LTT-MELISAR to determine specificity, 68 patients were tested in parallel to determine reproducibility, and 54 lymphocyte-reactive symptomatic patients were tested before and after antibiotic therapy to assess clinical relevance.
Most (86.2%) of the 36.9% (90/244) LTT-MELISAR-positive patients were seropositive and showed symptoms of active LB. Specificity was 96.7% and reproducibility 92.6%. After therapy, most patients (90.7%) showed negative or markedly reduced lymphocyte reactivity correlating with clinical improvement.
This novel LTT-MELISAR assay appears to correlate with active LB and may have diagnostic
relevance in confirming LB in clinically and serologically ambiguous cases.
D 2007 Elsevier Inc. All rights reserved
Discussion
The present study describes the development, specificity,
reproducibility, and clinical relevance for LB of a novel LTT
based on 2 improvements. First, a standardized validated
LTT was applied. Developed in the 1980s, the MELISAR
LTT modification has since been shown to improve the
sensitivity and specificity of conventional LTTs for detecting
metal sensitivity by utilizing a higher number of
lymphocytes per well (1 106), preselected nonmitogenic
pooled human serum, nonmitogenic and noncytotoxic
concentrations of antigens, and reduced numbers of monocytic
cells (Stejskal et al., 1994, 1999; Valentine-Thon and
Schiwara, 2003). Furthermore, quality control is enhanced
by supplementing the radiologic with a morphologic
analysis of proliferating cells to reduce possible falsepositive
or false-negative results (Stejskal et al., 1999). Its
technical validity and clinical relevance for detecting and
monitoring immunologic sensibilization to drugs, medications,
and heavy metals have been well documented
(Stejskal et al., 1986, 1990, 1994, 1996, 1999; Tibbling
et al., 1995; Stejskal, 1997; Sterzl et al., 1999; Regland et al.,
2001; Valentine-Thon and Schiwara, 2003; Prochazkova
et al., 2004; Kakuschke et al., 2005). The present study
describes the 1st application of LTT-MELISAR to detect
cellular immune reactivity to infectious organisms.
In the 2nd improvement, only well-characterized recombinant
Borrelia antigens used in commercial serologic tests
were used. In most previous studies, the antigenic source
was bacterial lysates or culture supernatants (Table 1).
Besides containing known mitogenic substances (e.g.,
bacterial lipopolysaccharides) and multiple antigenic epitopes,
such sources exclude the in vivo expressed antigens
recently demonstrated to improve the serologic diagnosis of
Borrelia infections (Wilske et al., 1993a; Schulte-Spechtel
et al., 2003; Wilske, 2003; Aguero-Rosenfeld et al., 2005).
The use of recombinant antigens in LTTs, therefore, can
reduce nonspecific background proliferation and at the same
time improve the specificity and clinical relevance of the
results. On the other hand, in 2 more recent studies using
exclusively recombinant Borrelia antigens, specificity was
improved but was still problematic (Bauer et al., 2001;
Kalish et al., 2003). Apparently, the use of recombinant
antigens together with an optimized LTT format is critical.
Because the 4 antigens selected in this study are derived
from B. afzelii and B. garinii but do not include
B. burgdorferi sensu stricto, their relevance for detecting LB
in countries with low or nonexistent prevalences of B. afzelii
and/or B. garinii (i.e., United States) is unclear. In 124 tests
of patients outside Germany, 40 were reactive and included
patients from England (n = 28), Luxembourg (n = 4),
Switzerland (n = 2), France (n = 2), Austria (n = 1),
Holland (n = 1), as well as the United States (n = 2) (data
not shown). Results from the latter 2 patients support a
potential diagnostic relevance even in the United States.
Alternatively, these 2 seropositive patients (frequent travelers)
may have become infected outside the United States.
A study with LB patients from within the United States
should clarify this important question.
Although most of the LTT-MELISAR–reactive patients
among the 244 initially tested were seropositive and
symptomatic, supporting a good correlation between cellular
reactivity and both Borrelia infection and, apparently,
active LB, 9 lymphocyte reactive patients were seronegative.
Seven of these showed clinical symptoms characteristic
of LB and may exemplify the controversial bseronegative
LB patientsQ reported in previous LTT studies, in particular,
in Dattwyler et al. (1988). Seronegativity may result from
immunosuppression, delayed humoral response, sequestration
of antibodies, or inadequate assay systems (Stanek
and Strle, 2003; Wilske, 2003; Aguero-Rosenfeld et al.,
2005). The latter is supported by our own observations that
many ELISA-negative patients are, in fact, positive in
Western blot (unpublished data, see also patients 1 and 5).
The current 2-tiered assay system will fail to identify a
Borrelia infection in such patients, because Western blot is
generally performed (i.e., reimbursed!) only when ELISA is
positive or equivocal. ELISA-negative persons with clinical
suspicion of LB should, therefore, except in unequivocal
EM, be tested in Western blot or, alternatively, in an
optimized LTT.
Based on the results obtained with control persons, a
new working cutoff was defined (i.e., patient positive only
with multiple reactions of SI z 3) in analogy to that
applied for identifying beryllium sensitization with the
beryllium lymphocyte proliferation test (Be-LPT) (United
States Department of Energy, 2001; Stange et al., 2004).
The high specificity obtained (96.7%) is in contrast to the
poor specificity of most previous studies and underlines
the relevance of a standardized LTT format with an
appropriate cutoff in addition to recombinant antigens.
bEquivocal Q patients (only 1 reaction of SI z3) may be
showing bfalse positiveQ or true very low-level specific
cellular reactivity, and should be repeat tested. Repetitive
bequivocal Q results appear to support active LB. Patient
6 in this study, for example, had classic early stage LB
(EM) with seroconversion, tested only bequivocal Q (not
positive) in 3 successive blood samples, and became LTTMELISAR
negative, seronegative, and asymptomatic after antibiotic therapy
The good reproducibility rate of 92.6% based on
replicate plate testing of 68 patients supports the reliability
of the assay system. This rigorous approach (as opposed to
testing replicate wells of the same plate) was deliberately
designed to account for the many variations arising from
the multiple manual steps in LTT-MELISAR. The
5 bdiscordan tQ patients showed low-level responses characterizing
the bgray zoneQ typical of most laboratory tests. A
similarly high reproducibility rate of 94% was reported for
the LTT-MELISAR as applied to detecting metal sensitivity
(Valentine-Thon and Schiwara, 2003), whereas comparable
reproducibility data for other LTTs have been generally
lacking. Recently, in an extensive analysis of Be-LPT results
obtained from split samples, an interlaboratory agreement of
up to 64.7% and intralaboratory agreement of up to 91.9%
was reported (Stange et al., 2004).
Although the sensitivity of LTT-MELISAR could not be
statistically evaluated in this study (appropriate samples for
PCR or culture authentication were not available from most
patients), the clinical relevance of this assay system for
identifying patients who can benefit from antibiotic therapy
was clearly demonstrated. The large majority of lymphocyte-
reactive symptomatic patients showed a significant
reduction in lymphocyte reactivity correlating with clinical
improvement after 1 or 2 courses of treatment. The 7 patients
still (albeit less) reactive (e.g., patients 4 and 5) may require
more time for depletion of specific memory cells. Similarly,
the 5 bequivocalQ patients who remained unchanged may
still respond with time, be therapy resistant, have reinfections,
or have an alternative etiology for their clinical state.
Although the kinetics of both humoral and cellular
immunologic responses is clearly variable, T-cell reactivity
appears to diminish more rapidly than antibody titers in
most patients. Consequently, the assay system described
here may provide an early marker of therapeutic success. At
the same time, persistent lymphocyte reactivity after a 1st
course of therapy (as seen in patients 2, 3, 4, and 5) may
indicate the need for more intensive and/or alternative
antimicrobial treatment. Reduced (but not negative) lymphocyte
reactivity after therapy was reported in 3 cases by
Breier et al. (1995) and in 6 cases by Krause et al. (1991)
but not in most of the 39 cases described more recently by
Vaz et al. (2001). Major methodological, protocol, and
cutoff differences in the latter study probably account for
this discrepancy
Because of the complexity of lymphocyte proliferation
assays and the controversy surrounding their use for
diagnosing LB, we strongly recommend that the LTTMELISAR
described here, or comparable tests, be applied
only in accredited laboratories with proven cell culture
expertise. Such laboratories should establish their own
working cutoff for defining a positive patient, perform
rigorous reproducibility testing at regular intervals for
quality control, and interpret the results in conjunction with
available anamnestic, clinical, and serologic data. Under
these conditions, LTT-MELISAR or equivalent LTTs should
have diagnostic relevance for clarifying serologically and
clinically ambiguous cases of LB and, where appropriate, in
confirming therapeutic success
