DNA AND CELL BIOLOGY Volume 11, Number 3, 1992 Mary Ann Liebert, Inc., Publishers
Pp. 207-213
Molecular
Diagnosis of Borrelia burgdorferi (Lyme Disease)
Infection
WILLIAM V. WILLIAMS,*'! PETER CALLEGARI,* BRUCE FREUNDLICH,* GREG KEENAN,*>t DANIEL ELDRIDGE,*4 HYONAH SHIN,* MARTIN KREITMAN,§ DANIEL McCALLUS,*4 and DAVID B. WEINERÎ
ABSTRACT
spite of significant advances in immunologically based testing, accurate diagnosis of Lyme borreliosis remains problematic. To address this issue, a DNA amplification-based diagnostic test was developed utilizing the polymerase chain reaction (PCR) and oligonucleotide primers specific for the OspA and OspB genes of Borrelia burgdorferi. In this approach, a relatively large DNA fragment is amplified with an outer set of primers, and a "nested" internal sequence of the PCR product subsequently rea m pi Hied with an inner set of primers. This nexted approach coupled with simple differential centrifugation allowed specific detection of as few as four B. burgdorferi organisms mixed in 2 ml of blood. This methodology was utilized on patients' samples, and it allowed detection of B. burgdorferi in the peripheral blood and urine of several individuals with clinical evidence of Lyme borreliosis. PCR became negative and symptoms improved following antibiotic therapy of treated individuals. These studies suggest that direct detection of Borrelia in infected individuals can aid in diagnosis and evaluation of therapy for Lyme borreliosis. In
INTRODUCTION Lyme disease is a spirochete, first isolated by Barbour and burgdorferi, Burgdorfer (Barbour, 1984). The spirochetes contain at least 30 different proteins, but the function of only a few of these is known. Important bacterial proteins include: outer surface protein (OSP) A and B (Coleman and Benach, 1987), 66-kD outer membrane protein, 41-kD flagellum antigen, and a 58- or 60-kD heat shock protein (Luft et al, 1989). Two other proteins (19 kD and 26 kD) appear to elicit immune responses in patients with Lyme borreliosis, although their role is not known (Luft et al., 1989). There are differences in American and European isolates, with more diversity in proteins noted in the European isolates (Barbour et al, 1985). These isolate-specific differences may account for different clinical presentations seen in Lyme disease. Borrelia burgdorferi remains extremely difficult to isolate and culture from blood, cerebrospinal fluid, or synovial fluid (Barbour, 1984), and diagnosis of Lyme bor-
The
agent that causes
Borrelia
*The
Rheumatology Section, Department
of
reliosis remains problematic. Detection of an antibody response has been the basis of testing for Lyme borreliosis. Indirect immunofluorescence and enzyme-linked immunoabsorbance assays (ELISA) are used to quantify the immune response. IgM titers after infection reach a peak between 3 and 6 weeks, occasionally having a second peak later in the illness. IgG titers rise more slowly and are generally higher months to years later, especially in persistent Lyme. Detection of patient antibody directed at one specific, unique surface antigen of the organism using the various "enriched" ELISA assays and the use of capture IgM ELISA have also been reported (Duffy et al, 1988). Immunoblotting (Western blot analyses) of blood or cerebrospinal fluid (CSF), primarily to address false positive assays, has also been used (Grodzicki and Steere, 1988). Each of these assays is limited by the host's ability to mount an antibody response within the range of detection of these assays. This has led to particular difficulty in diagnosis early in the disease and after antibiotic therapy (Datt-
wyler et al, 1988). The polymerase chain reaction (PCR) has
been used to
Medicine, University of Pennsylvania School of Medicine, Philadelphila,
tThe Pédiatrie Rheumatology Section, The Childrens' Hospital of Philadelphia, Philadelphia, íThe Biotechnology Center, The Wistar Institute of Anatomy and Biology, Philadelphia, PA §The Department of Molecular Biology, Princeton University, Princeton, NJ 08540. 207
PA 19104. 19104.
PA 19104.
208
WILLIAMS ET AL.
amplify and detect
burgdorferi DNA sequences (Rosa Malloy et al, 1990; Neilsen et al, 1990; Persing et al, 1990a,b; Wallich et al, 1990). PCR is able directly to detect portions of the spirochete genome in body fluids and pathologic specimens. Using Thermus aquaticus DNA polymerase and oligonucleotide primers constructed to hybridize with a target gene, it is possible to multiply the number of copies of the target gene over a million-fold, allowing easier detection. PCR has been used B.
and Schwan, 1989;
aid in the detection of the infectious agents in various in different tissues. The purpose of the studies described herein is to determine (i) the feasibility of applying PCR to the specific detection of DNA derived from B. burgdorferi spirochetes and (ii) the ability of this technique to detect B. burgdorferi DNA in infected individuals. Detection of B. burgdorferi DNA in clinical specimens may provide initial information regarding the potential utility of this methodology in clinical diagnosis. to
body fluids and
MATERIALS AND METHODS
Organisms Borrelia
burgdorferi spirochetes strains 35210 and
35211, and B. hermsii strain 35209 were obtained from the American Type Culture Collection (Rockville, MD). Spirochetes were grown in BSK media (Barbour, 1984) at 37°C in a 5% C02 atmosphere. Typical motile spirochetes were observed and enumerated on phase-contrast microscopy. Spirochetes were pelleted by centrifugation at 10,000 rpm for 30 min. Pellets or suspensions of spirochetes in BSK media were stored at -70°C prior to use. DNA extraction All samples were handled under sterile conditions in a laminar flow hood in closed containers, with only one specimen opened at a time. DNA was extracted by a modification of the protocol described for isolation of genomic DNA from mammalian tissue (Ausubel et al, 1989). Briefly, bacterial pellets or cell pellets (see below) were resuspended in 0.3 ml of digestion buffer (100 mMNaCl, 10 mM Tris-HCl pH 8, 25 rnM EDTA pH 8, 0.5% NaDodS04, 0.1 mg/ml proteinase K), and incubated at 65 CC for 1-2 hr. The samples were extracted with phenol/ chloroform/isoamyl alcohol 3-4 x, and microfuged. The aqueous layer was removed and DNA was precipitated in 1/2 volume 7.5 M ammonium acetate with 2 volumes 100% ethanol, microfuged, washed in 70% ethanol, and rotary evaporated. The DNA was resuspended in 50 /¿l of distilled water and 5-10 ¡A was utilized for DNA amplification.
DNA
amplification and detection
DNA was amplified utilizing Thermus aquaticus DNA polymerase ( Taq polymerase) and standard reaction conditions suggested by the manufacturer (Perkin-Elmer Cetus Corp., Norwalk, CT). The reaction mixture contained 10
lA of 10 x reaction buffer, 16 ¡à 1.25 mM dNTPs (final concentration 200 fiMin each dNTP), 5 fà of each oligonucleotide primer at 20 fiM (final 1 ¡tM in each primer), 5 fi\ of DNA, 0.5 ¡à of Taq polymerase, and 58.5 Ml distilled/
deionized water. The samples were covered with a drop of mineral oil, and amplified for 30-40 cycles (melting at 95°C for 1 min, annealing at 52°C for 2 min, elongation at 72° C for 2 min) on a Programmable Thermal Cycler (MJ Research, Watertown, MA). The reaction products were run on a 3% agarose gel, stained with ethidium bromide, and photographed under UV light.
Blotting
and
oligonucleotide probing
Agarose gels were transferred to Gene Screen Plus filters (Whatman Limited, England) by standard capillary transfer procedures as suggested by the manufacturer. For dot or slot blots, DNA amplification products were applied to the nylon filters utilizing an ABN Vacudot-VS dot blot apparatus or a Vacuslot-VS slot blot apparatus (American Bionetics, Hayward, CA), according to the manufacturer's instructions. Hybridization was with primer B noted in Table 1 and Fig. 1. Oligonucleotide labeling employed 100 ng of DNA, 75 pCi [32P]ATP, 2.5 fi\ of 10 x kinase buffer (500 mM Tris HC1 pH 7.6, 100 mM MgCl2, 50 mM dithiothreitol, 1 mM spermidine, 1 mM EDTA), 10 units of T4 DNA kinase adjusted to a final volume of 25 ¡à with distilled water, labeling was carried out by incubation at 37°C for 30 min prior to use. Blots were prehybridized in 5 x SSC, 5x Denhardt's solution, 0.1% NaDodS04 for 11.5 hr at 55°C in Seal-a-Meal bags, the solution was poured off, 32P-labeled oligonucleotide was added (75 /¿Ci) in prehybridization solution and hybridized for 2-3 hr at 42°C or overnight at 4°C, the blot washed 1 x in 2 x SSC, 0.1% NaDodS04 for 20 min at 45 °C, then 3 x in 5 x SSC, 0.1% NaDodS04 for 20 min at 45°C, and exposed to Kodak XRP film at -70°C for 2-72 hr.
RESULTS
Primer design The primers utilized were designed based on sequence homology between the outer surface proteins (OspA and OspB) of B. burgdorferi (Bergstrom et al, 1989) as shown in Fig. 1. Where there were discrepancies in the sequences, degeneracy was included to allow annealing to both gene products. The oligonucleotides were synthesized by the Wistar Institute oligonucleotide synthesis facility, and their sequences
are
shown in Table 1.
Amplification of B. burgdorfei
DNA
These primers are able to amplify B. burgdorferi DNA shown in Fig. 2. DNA was extracted from 1 ml of culture media containing B. burgdorferi or B. hermsii organisms and subjected to amplification by PCR with the various primers. All of the available primers were tested against both B. burgdorferi and B. hermsii in 40 cycles of amplification. All of the primers gave good results on B. as
MOLECULAR DIAGNOSIS OF LYME DISEASE
209
Table 1. Oligonucleotide Primers
Sequence (5'—3')
Primer A »Primer B Primer B' »Primer C »Primer D Primer E
TTGTAAGCAAAGAAAAAAA TTTAAAGAAGGAACTGTTACT AGTAACAGTTCCTTCTTTAA TTAAAAACGCTTTAAAATAA
(GT)(AC)CGGCAA(GA)TA(CT)GAT(CT)TAG(CT)(TC)G ATTTCA(AC)(CT)TGCTGA(CT)CC(CT)TC
Primers
size
Primers
A-C D-C A-E
701 bp 679 bp 664 bp
D-E A-B D-B
»Denotes anti-sense
OSPA
OSPA
bp bp bp
size
B'-C B'-E
285 247
bp bp
Oligonucleotide Primers
550 520 530 560 570 540 GCTGAAAAAACAACATTGGTGGTTAAAGAAGGAACTGTTACTTTAAGCAAAAATATTTCA
I
I I I I I I I I I I
I I I
I I I I I I I I I I
I I I I I I I I I
I I I
I
I
I
I I I
I
GTCGGAAAAACAACAGTGGAAATTAAAGAAGGTACTGTTACTCTAAAAAGAGAAATTGAA 650 620 630 640 600 610
160 140 150 110 120 130 CCTGGTGAAATGAAAGTTCTTGTAAGCAAAGAAAAAAACAAAGACGGCAAGTACGATCTAGCTG I I I I I I I I I I I I I I I I II I I I I I I I I I I M I I I I I I I I I I I I I I TTTAATGGTAATAAAATTTTTGTAAGCAAAGAAAAAAATAGCTCCGGCAAATATGATTTAGTCG 210 180 190 220 230 200
I OSPB
Product size
primer.
Position of
OSPB
Expected Product Sizes
Designation
Product
OSPA
and
I I I I
II
I I I I I I
I I I I I I M
III
III
II
I I I I I I I I I I I I I I I I I I I I
ACCAGCCTAGAAGGATCAGCAAGTGAAATTAAAAATCTTTCAGAGCTTAAAAACGCTTTAAAATAA -" 880 830 840 850 860 870
Nucleotide sequences of OspA and OspB aligned. The nucleotide sequences of OspA and OspB are shown aligned for maximal homology, with the base numbering immediately above and below the sequences. The regions corresponding to the primers utilized in these studies are underlined, and their designation is indicated above the primers. (—) Sense (coding sequence) primer; (—) anti-sense (noncoding strand) primer.
FIG. 1.
burgdorferi DNA, with a few exhibiting "cross-reactivity" on B. hermsii DNA, specifically A&C, A&B, and D&B. Interestingly, when these reaction products were transferred to nitrocellulose and probed with "P-labeled primer B, hybridization was seen only to those B. hermsii amplified bands that contained primer B in the amplification process (data not shown). Results were also compared for DNA extracted from E. coli human peripheral blood mononuclear cells, and salmon sperm DNA (data not shown). These templates did not give rise to PCR products, further indicating the specificity of the primers used.
Amplification
in the presence
of extraneous DNA
For this technique to be useful for clinical specimens, DNA must also be detectable in the presence of extraneous DNA. Therefore, B. burgdorferi organisms were mixed in varying dilutions in 1 ml of whole heparinized blood and DNA extracted from the mixture. Following ethanol precipitation, the DNA was resuspended in 50 fi\ of water, and 5 fil utilized for PCR analysis as noted above. Using this protocol, only the highest concentration of the highest dilution of DNA amplified with the A—C primers gave a specific product. This was likely due to the interfering ef-
210
12
WILLIAMS ET AL.
3 4
5 6
7 8
9 10 11 12 1314 1516
1718
Utility of oligonucleotide primers. DNA from B. burgdorferi strain 38210 (odd-numbered lanes) or B. hermsii strain 38209 (even-numbered lanes) was amplified with the PCR primer pairs noted in Table 1. The primer pairs were: lanes 1 and 2, A—C; lanes 3 and 4, D—C; FIG. 2.
lanes 5 and 6, A—E; lanes 7 and 8, D—E; lanes 9 and 10, A-B; lanes 11 and 12, D-B; lanes 13 and 14, B'-C; lanes 15 and 16, B'-E; lanes 17 and 18, ¿X174 Hae III and \Hind III molecular weight markers. The additional bands seen in some lanes may be due to multimerization of PCR products or mispriming of nontarget sequences.
feet of extraneous DNA in the sample. A nested set of priwas next utilized to allow detection of smaller amounts of DNA. Similarly prepared samples were subjected to nested PCR utilizing first A—C followed by D—E primers. This strategy allowed the B primer to be utilized as a probe in hybridization studies. The results are presented in Fig. 3. DNA extracted from heparinized blood (1 ml) spiked with decreasing numbers of B. burgdorferi organisms was amplified at 1:1 and 1:10 dilutions. In the concentration range used, only the 1:10 dilution of DNA elicited an appropriate sized band (Fig. 3A). The presence of high concentrations of cellular DNA may interfere with specific priming at other dilutions. As few as 400 organisms were detected in this procedure. When the gel was transferred to nylon filters and probed with "P-labeled primer B, only the appropriate lanes could be visualized by hybridization (Fig. 3B). When the remaining reaction products were directly dotted onto nylon filters and probed, again only the appropriate samples showed hybridization (Fig. 3C). The hybridization procedure was approximately fivefold more sensitive than electrophoresis followed by ethidium bromide staining, but both techmers
niques appeared equally specific.
effort to reduce the amount of contaminating protein/DNA, an additional method to increase the specific signal was tested. In these experiments, blood was drawn into tubes containing EDTA instead of heparin, organisms were added at varying concentrations, and the red blood cells were lysed with a 5- to 10-fold excess of Geys' solution. Differential centrifugation was used to separate partially the remaining white blood cells and spirochetes. The white blood cells were pelleted at 1,000 rpm for 10 min (first pellet), and the supernatant then spun at 10,000 rpm for 30 min (second pellet). DNA was extracted from each pellet and the DNA subjected to amplification by the A—C followed by the D—E primers. As controls, highspeed pellets of organisms diluted in phosphate-buffered saline (PBS) or urine were also tested. The results are shown in Fig. 4. Both the first and second pellets of B. In
an
c
I
#
Ä
«¡I
FIG. 3. Amplification of B. burgdorferi DNA sequences from organisms diluted in heparinized blood. Borrelia burgdorferi was grown as described in Materials and Methods. One milliliter containing 4 x 10* spirochetes, was serially diluted in 1-ml aliquots of heparinized human blood. The cells were lysed in DNA digestion buffer, extracted with phenol/chloroform/isoamyl alcohol, and the DNA precipitated and resuspended in 50 fil of water. Five microliters of DNA (even lanes) or 0.5 ¡a of DNA (odd lanes) was amplified initially with primers A and C. Five microliters of the reaction products were reamplified with primers D and E. Ethidium bromide-stained reaction products are shown in Fig. 3A. The gel was transferred to nylon filters and probed with 32P-labeled primer B, with the results shown in Fig. 3B. In addition, 50 pi of the reaction products were dotted onto nylon filters and similarly probed, with the results shown in Fig. 3C. The lanes correspond to the following number of spirochetes in the original reaction mixture: 1, 4 X 10*; 2, 4 x 10*; 3, 4 x 103; 4, 4 x 104; 5, 4 x 102; 6, 4 x 103; 7, 4 x 10; 8, 4 x 102. The higher-molecular-weight bands present in B5 and B7 are probably multimers of the PCR products inapparent on ethidium bromide staining. There is also a smudge artifact overlying B3.
burgdorferi produced clear, distinct bands when amplified, although the intensity of the bands for the second pellet was of greater intensity at higher dilutions of organisms. The organisms diluted in PBS or urine gave similar results. The strategy of lysing the red blood cells, employing differential centrifugation and avoiding heparin, added to the sensitivity of detection. In any case, this approach allowed detection of as few as 100 organisms in mock clinical samples from whole blood.
MOLECULAR DIAGNOSIS OF LYME DISEASE
12 3 4 5
211
6 7 8 9 10
1112131415
16 17 18 19 20
FIG. 4.
Differential centrifugation to enhance PCR detection of B. burgdorferi DNA. Aliquots of 106 B. burgdorferi organisms were serially 10-fold-diluted in 2 ml of EDTA-anticoagulated blood (lanes 1-10), urine (lanes 11-15), or phosphate-buffered saline (PBS) (lanes 16-20). The urine and PBS were centrifuged at 10,000 rpm for 30 min, and pellets utilized. The blood was combined with 2 ml of Gey's solution on ice for 5 min to lyse most of the red blood cells, and the cells were centrifuged at 1,000 rpm for 5 min (cell pellet 1-5). The supernatant was further centrifuged at 10,000 rpm for 30 min (bacterial pellet 6-10). DNA was extracted from the pellets as noted above, and following precipitation, resuspended in 50 ¡A of water. Five microliters of the reatcion product was amplified with primers A and C, and 10 fi\ of the reaction product reamplified with primers D and E. The reaction products were analyzed on a 3% agarose gel with ethidium bromide staining. The following numbers of spirochetes in the original reaction mixture correspond to the lanes: 105, ~
=
=
lanes 1, 6, 11, and 16; 10", lanes 2, 7, and 12; 103, lanes 3, 8, 13, and 17; 102, lanes 4, 9, 14, and 18; 10, lanes 5, 10, 15, and 19; 1, lane 20.
of B. burgdorferi from patients Detection
in blood and urine
5
6 7
8 9 10
We next examined blood and urine specimens from 2 individuals. In these cases, both blood and urine specimens gave positive results, whereas negative controls remained Detection of Borrelia DNA sequences in the penegative. The clinical characteristics of two of these cases FIG. 5. blood of an infected individual. EDTA anti-coagripheral are summarized here. ulated blood was obtained from patient J.D., and proJ.D. is a 57-year-old white male machinist with a 12-year cessed to the scheme described in the legend to history of rheumatoid factor-positive palindromic arthri- Fig. 4. according DNA was amplified with primers A and C followed tis. The patient was first seen by us in December, 1990, by D and E, as described in Fig. 4. Reaction products were with a change in the pattern from his previous course of run on a 3% agarose gel and stained with ethidium broarthritis. This consisted of 9 months of constant severe mide (A), followed by capillary transfer to nylon filters pain and swelling in the small joints of the hands and both and probing with 32P-labeled primer B (B). The DNA amankles that prevented him from working. Physical exami- plified shown for each lane was: lane 1, negative control nation confirmed a symmetrical swelling, erythema, and human DNA; lanes 2-4, positive control B. burgdorferi tenderness in multiple small joints of the hand along with DNA at three 10 X dilutions; lanes 5-7, three 10 x dilulimited range of shoulder motion bilaterally due to pain, tions of cell pellet DNA; 8-10, three 10 x dilutions of bacterial pellet DNA. Positive controls were obtained from but was otherwise unremarkable. DNA extracted from 100 pi of B. burgdorferi organisms Initial evaluation revealed a rheumatoid factor of grown in culture. 1:1,280 (normal < 1:80) by latex fixation, antinuclear antibody 1:640, speckled pattern, and normal CH50. A chest radiograph demonstrated increased interstitial markings. The patient initially received 20 mg of Prednisone daily CG. is an 11-year-old white male who noted swelling of and 1 gram of sulfasalazine daily with no improvement his left knee 5 months prior to our evaluation. There was over 1 month. Further history revealed that the patient no history of trauma, fever, rash, recent immunization, lived in a wooded area and had had a circular red rash on tick bite, any other joint involvement, back pain, or an ankle several months before his arthritis had intensidysuria. Past medical history and family history were unrefied. An ELISA for Lyme, which had been negative in markable. The pediatrician noted a mild effusion of the June, 1990, was now positive. PCR results from blood are left knee. Radiograph was negative as was ANA. The Hgb shown in Fig. 5. Note that both the cellular and bacterial was 12.6 g, WBC 7,700/cu mm, platelet count of 474,000/ blood samples were positive. Faint positive results in the cu. mm, and the ESR was 89 mm/hr. A Lyme titer oburine were most easily detectable following Southern trans- tained 1 month after presentation was reportedly positive fer of the blot and subsequent probing and hybridization by ELISA. The patient was diagnosed as having Lyme (data not shown). Within 1 week of receiving the Ceftria- arthritis and was treated for 6 weeks with amoxicillin. xone, the patient had a marked improvement of joint pain Swelling in the left knee improved, occurring only with and swelling. Repeat PCR analysis following antibiotic physical activity. During this therapy, the right knee subsetherapy was negative. quently developed an effusion. The patient was treated
212
WILLIAMS ET AL.
with on ASA three times daily for 1 month without resolution and was referred to the Children's Hospital of Pennsylvania for evaluation. At our initial evaluation, his exam was remarkable for a moderated effusion, warmth, and a reduction in range of motion of the right knee. Laboratories revealed a Hb of 12.6 g, a WBC of 9,900/cu mm, platelet count of 356,000/ cu mm, and an ESR of 55 mm/hr. Gram stain was negative for white cells and bacteria and bacterial culture of the synovial fluid was negative. Lyme titer by ELISA was positive at 43.2 SD above the mean (normal is 2 SD). Western blot was also positive. Blood, urine, and synovial fluid were positive by PCR (data not shown). A 2-week course of daily intravenous ceftriaxone was initiated. Follow-up 1 month later the patient was entirely asymptomatic. Repeat PCR following therapy was negative. Together, these results indicate the feasibility of detecting B. burgdorferi sequences in clinical specimens using a nested approach to PCR.
DISCUSSION Several groups have suggested PCR as a potential diagnostic tool for Lyme borreliosis (Rosa and Schwan, 1989; Neilsen et al, 1990; Wallich et al, 1990). Borrelia burgdorferi DNA has recently been detected in clinical specimens by PCR. Persing et al used primers derived from the OspA sequence, and were able to detect B. burgdorferi sequences in Ixodces dammini ticks collected from the field, as well as museum specimens (Persing et al, 1990a,b). Malloy et al utilized primers derived from the OspA sequence to amplify a 309-bp fragment followed by oligonucleotide probing (Malloy et al, 1990). They utilized this technique to detect B. burgdorferi DNA in the blood and urine of an infected canine. Other groups have utilized PCR for detection of B. burgdorferi DNA from patient samples (Nishio et al, 1990; Goodman et al, 1991). Their studies included detection of B. burgdorferi in the urine of infected individuals. This approach eliminated the potential of contaminating cellular DNA and gave excellent results, as 4 patients with suspected Lyme disease who gave positive results on PCR showed an excellent clinical response to antibiotic therapy. The utility of urine testing is quite apparent for Lyme disease, as this is the basis for immunodetection assays for spirochetal antigens (Hyde et al, 1989). While this group did not concentrate their samples by centrifugation, simple extraction and amplification was sufficient for detection. We utilized three complementary strategies to address the problem of contaminating DNA. The first is the use of primers derived from sequences present in > 1 copy. Our primers were designed based on sequence similarity between the OspA and OspB genes. This gives the primers a twofold advantage as functionally two copies of the gene product would serve as templates for amplification. Our approach also targets highly conserved sequences that might be less susceptible to genetic drift. The second is the use of nested PCR. By preamplifying DNA extracted from clinical samples, the percentage of B. burgdorferi DNA
template is significantly enriched prior to the second amplification. This allows detection of relatively fewer organisms in the presence of large amounts of extraneous DNA. We also utilized differential centrifugation to eliminate as much extraneous cellular DNA as possible. We sought to take advantage of the substantial size difference between the spirochetes and peripheral blood cells. Red blood cells are first dispensed with by lysis in hypotonie
ammonium chloride solution. White blood cells were then pelleted at low speed, followed by a high-speed spin to pellet potential smaller pathogens. Under these conditions, DNA was extracted from as much as 10-20 ml of blood without the concomitant large quantities of contaminating cellular DNA present. When clinical samples are processed by the methods described here, we observed patients with positive results for both blood and urine (Fig. 5). These were the results observed in patients with clinical histories suggestive of Lyme disease. In one case, the clinical scenario was quite difficult to sort out, as the patient had preexisting palindromic rheumatism, and this was markedly exacerbated by tick bite exposure. In such an individual, with polyclonal hypergammaglobulinemia, positive antibody tests were difficult to interpret. This problem was even more apparent in the second individual, who had previous clinical and sérologie evidence of Lyme disease, was treated, and was thought to be reinfected. In such cases, sérologie testing and tests for other immune responses are difficult to interpret. Thus, direct detection of B. burgdorferi spirochetes can be of clinical utility in clinical scenarios where other tests are difficult to interpret. The data presented here demonstrates that, with respect to B. burgdorferi, PCR diagnosis is the method of choice for clarification of difficult clinical reexposure cases and, similarly, may be of great utility in early disease. The protocols described herein may also be of general utility in detection of other infectious agents.
ACKNOWLEDGMENTS W.V.W. is supported in part by grants from the NIH. P.C. is supported in part by the NIH, and D.B.W. is supported in part by grants from AmFAR and the CTR.
REFERENCES AUSUBEL, F., BRENT, R., KINGSTON, R., MOORE, D., SEIDMAN, J., SMITH, J., and STRUHL, K. (1989). Current Protocols in Molecular Biology. (Greene Publishing Associates and Wiley-Interscience, John Wiley & Sons, New York). BARBOUR, A.G. (1984). Isolation and cultivation of Lyme disease spirochetes. Yale J. Biol. Med. 57, 521-525. BARBOUR, A., HEILAND, R., and HOWE, T. (1985). Heterogeneity of major proteins in Lyme disease borreliae: A molecular analysis of North American and European isolates. J. Infect. Dis. 152, 478-484. BERGSTROM, S., BUNDOC, V., and BARBOUR, A. (1989). Molecular analysis of linear plasmid-encoded major surface proteins, OspA and OspB, of the Lyme disease spirocheaete Borrelia burgdorferi. Mol. Microbiol. 3, 479-486.
213
MOLECULAR DIAGNOSIS OF LYME DISEASE COLEMAN, J., and BENACH, J. (1987). Isolation of antigenic components from the lyme disease spirochete: Their role in early diagnosis. J. Infect. Dis. 155, 756-765.
DATTWYLER, R.J., VOLKMAN, D.J., LUFT, B.J., HALPERIN, J.J., THOMAS, J., and GOLIGHTLY, M.G. (1988).
Seronegative lyme disease: Dissociation of specific T- and B-lymphocyte responses to Borrelia burgdorferi. N. Engl. J.
Med. 319, 1441. DUFFY, J., MERTZ, L., WOBIG, G., and KATZMANN, J.
(1988). Diagnosing lyme disease: The contribution of sérologie testing. Mayo Clin. Proc. 63, 1116-1121. GOODMAN, J., JURKOVICH, P., KRAMBER, J., and JOHNSON, R. (1991). Molecular detection of persistent Borrelia burgdorferi in the urine of patients with active Lyme disease. Infect. Immun. 59, 269-278. GRODZICKI, R., and STEERE, A. (1988). Comparison of immunoblotting and indirect enzyme-linked immunosorbent assay using different antigen preparations for diagnosing early Lyme disease. J. Infect. Dis. 157, 790-797. HYDE, F., JOHNSON, R., WHITE, T., and SHELBURNE, C. (1989). Detection of antigens in the urine of mice and humans infected with Borrelia burgdorferi, etiologic agent for lyme disease. J. Clin. Microbiol. 27, 58-61. LUFT, B., JIANG, W., MUNOZ, P., DATWYLER, R., and GOREVIC, P. (1989). Biochemical and immunological characterization of the surface proteins of Borrelia burgdorferi. Infect. Immun. 57, 3637-3645. MALLOY, D., NAUMAN, R., and PAXTON, H. (1990). Detection of Borrelia burgdorferi using the polymerase chain reaction. J. Clin. Microbiol. 28, 1089-1093. NEILSEN, S., YOUNG, K., and BARBOUR, A. (1990). Detection of Borrelia burgdorferi DNA by the polymerase chain re-
action. Mol. Cell. Probes 4, 71-79. NISHIO, M.J., LIEBLING, M.R., RODRIGUEZ, A., SIGAL, L.H., and LOUIE, J.S. (1990). Identification of Borrelia burgdorferi using interrupted polymerase chain reaction. Arthritis Rheum. 33, S84. PERSING, D., TELFORD, S., RHYS, P., DODGE, D., WHITE, T., MALAWISTA, S., and SPEILMAN, A. (1990a).
Detection of the Borrelia burgdorferi DNA in museum mens of Ixodes dammini ticks. Science 249, 1420-1423.
speci-
PERSING, D., TELFORD, S., SPIELMAN, A., and BARTHOLD, S. (1990b). Detection of Borrelia burgdorferi infection in Ixodes dammini ticks with the polymerase chain reaction. J. Clin. Microbiol. 28, 566-572. ROSA, P., and SCHWAN, T. (1989). A specific and sensitive assay for the Lyme disease spirochete Borrelia burgdorferi using the polymerase chain reaction. J. Infect. Dis. 160, 1018-1029. WALLICH, R., MOTER, S., SIMON, M., EBNET, K., HEIBERGER, A., and KRAMER, M. (1990). The Borrelia burgdorferi flagellum-associated 41-kilodalton antigen (flagellin): Molecular cloning, expression and amplification of the gene. Infect. Immun. 58, 1711-1719. Address
reprint requests
to:
Dr. David B. Weiner
The
Biotechnology Center
The Wistar Institute 3600 Spruce Street Philadelphia, PA 19104-4268 Received for publication November 29, 1991; accepted January 15, 1992.
Pp. 207-213
Molecular
Diagnosis of Borrelia burgdorferi (Lyme Disease)
Infection
WILLIAM V. WILLIAMS,*'! PETER CALLEGARI,* BRUCE FREUNDLICH,* GREG KEENAN,*>t DANIEL ELDRIDGE,*4 HYONAH SHIN,* MARTIN KREITMAN,§ DANIEL McCALLUS,*4 and DAVID B. WEINERÎ
ABSTRACT
spite of significant advances in immunologically based testing, accurate diagnosis of Lyme borreliosis remains problematic. To address this issue, a DNA amplification-based diagnostic test was developed utilizing the polymerase chain reaction (PCR) and oligonucleotide primers specific for the OspA and OspB genes of Borrelia burgdorferi. In this approach, a relatively large DNA fragment is amplified with an outer set of primers, and a "nested" internal sequence of the PCR product subsequently rea m pi Hied with an inner set of primers. This nexted approach coupled with simple differential centrifugation allowed specific detection of as few as four B. burgdorferi organisms mixed in 2 ml of blood. This methodology was utilized on patients' samples, and it allowed detection of B. burgdorferi in the peripheral blood and urine of several individuals with clinical evidence of Lyme borreliosis. PCR became negative and symptoms improved following antibiotic therapy of treated individuals. These studies suggest that direct detection of Borrelia in infected individuals can aid in diagnosis and evaluation of therapy for Lyme borreliosis. In
INTRODUCTION Lyme disease is a spirochete, first isolated by Barbour and burgdorferi, Burgdorfer (Barbour, 1984). The spirochetes contain at least 30 different proteins, but the function of only a few of these is known. Important bacterial proteins include: outer surface protein (OSP) A and B (Coleman and Benach, 1987), 66-kD outer membrane protein, 41-kD flagellum antigen, and a 58- or 60-kD heat shock protein (Luft et al, 1989). Two other proteins (19 kD and 26 kD) appear to elicit immune responses in patients with Lyme borreliosis, although their role is not known (Luft et al., 1989). There are differences in American and European isolates, with more diversity in proteins noted in the European isolates (Barbour et al, 1985). These isolate-specific differences may account for different clinical presentations seen in Lyme disease. Borrelia burgdorferi remains extremely difficult to isolate and culture from blood, cerebrospinal fluid, or synovial fluid (Barbour, 1984), and diagnosis of Lyme bor-
The
agent that causes
Borrelia
*The
Rheumatology Section, Department
of
reliosis remains problematic. Detection of an antibody response has been the basis of testing for Lyme borreliosis. Indirect immunofluorescence and enzyme-linked immunoabsorbance assays (ELISA) are used to quantify the immune response. IgM titers after infection reach a peak between 3 and 6 weeks, occasionally having a second peak later in the illness. IgG titers rise more slowly and are generally higher months to years later, especially in persistent Lyme. Detection of patient antibody directed at one specific, unique surface antigen of the organism using the various "enriched" ELISA assays and the use of capture IgM ELISA have also been reported (Duffy et al, 1988). Immunoblotting (Western blot analyses) of blood or cerebrospinal fluid (CSF), primarily to address false positive assays, has also been used (Grodzicki and Steere, 1988). Each of these assays is limited by the host's ability to mount an antibody response within the range of detection of these assays. This has led to particular difficulty in diagnosis early in the disease and after antibiotic therapy (Datt-
wyler et al, 1988). The polymerase chain reaction (PCR) has
been used to
Medicine, University of Pennsylvania School of Medicine, Philadelphila,
tThe Pédiatrie Rheumatology Section, The Childrens' Hospital of Philadelphia, Philadelphia, íThe Biotechnology Center, The Wistar Institute of Anatomy and Biology, Philadelphia, PA §The Department of Molecular Biology, Princeton University, Princeton, NJ 08540. 207
PA 19104. 19104.
PA 19104.
208
WILLIAMS ET AL.
amplify and detect
burgdorferi DNA sequences (Rosa Malloy et al, 1990; Neilsen et al, 1990; Persing et al, 1990a,b; Wallich et al, 1990). PCR is able directly to detect portions of the spirochete genome in body fluids and pathologic specimens. Using Thermus aquaticus DNA polymerase and oligonucleotide primers constructed to hybridize with a target gene, it is possible to multiply the number of copies of the target gene over a million-fold, allowing easier detection. PCR has been used B.
and Schwan, 1989;
aid in the detection of the infectious agents in various in different tissues. The purpose of the studies described herein is to determine (i) the feasibility of applying PCR to the specific detection of DNA derived from B. burgdorferi spirochetes and (ii) the ability of this technique to detect B. burgdorferi DNA in infected individuals. Detection of B. burgdorferi DNA in clinical specimens may provide initial information regarding the potential utility of this methodology in clinical diagnosis. to
body fluids and
MATERIALS AND METHODS
Organisms Borrelia
burgdorferi spirochetes strains 35210 and
35211, and B. hermsii strain 35209 were obtained from the American Type Culture Collection (Rockville, MD). Spirochetes were grown in BSK media (Barbour, 1984) at 37°C in a 5% C02 atmosphere. Typical motile spirochetes were observed and enumerated on phase-contrast microscopy. Spirochetes were pelleted by centrifugation at 10,000 rpm for 30 min. Pellets or suspensions of spirochetes in BSK media were stored at -70°C prior to use. DNA extraction All samples were handled under sterile conditions in a laminar flow hood in closed containers, with only one specimen opened at a time. DNA was extracted by a modification of the protocol described for isolation of genomic DNA from mammalian tissue (Ausubel et al, 1989). Briefly, bacterial pellets or cell pellets (see below) were resuspended in 0.3 ml of digestion buffer (100 mMNaCl, 10 mM Tris-HCl pH 8, 25 rnM EDTA pH 8, 0.5% NaDodS04, 0.1 mg/ml proteinase K), and incubated at 65 CC for 1-2 hr. The samples were extracted with phenol/ chloroform/isoamyl alcohol 3-4 x, and microfuged. The aqueous layer was removed and DNA was precipitated in 1/2 volume 7.5 M ammonium acetate with 2 volumes 100% ethanol, microfuged, washed in 70% ethanol, and rotary evaporated. The DNA was resuspended in 50 /¿l of distilled water and 5-10 ¡A was utilized for DNA amplification.
DNA
amplification and detection
DNA was amplified utilizing Thermus aquaticus DNA polymerase ( Taq polymerase) and standard reaction conditions suggested by the manufacturer (Perkin-Elmer Cetus Corp., Norwalk, CT). The reaction mixture contained 10
lA of 10 x reaction buffer, 16 ¡à 1.25 mM dNTPs (final concentration 200 fiMin each dNTP), 5 fà of each oligonucleotide primer at 20 fiM (final 1 ¡tM in each primer), 5 fi\ of DNA, 0.5 ¡à of Taq polymerase, and 58.5 Ml distilled/
deionized water. The samples were covered with a drop of mineral oil, and amplified for 30-40 cycles (melting at 95°C for 1 min, annealing at 52°C for 2 min, elongation at 72° C for 2 min) on a Programmable Thermal Cycler (MJ Research, Watertown, MA). The reaction products were run on a 3% agarose gel, stained with ethidium bromide, and photographed under UV light.
Blotting
and
oligonucleotide probing
Agarose gels were transferred to Gene Screen Plus filters (Whatman Limited, England) by standard capillary transfer procedures as suggested by the manufacturer. For dot or slot blots, DNA amplification products were applied to the nylon filters utilizing an ABN Vacudot-VS dot blot apparatus or a Vacuslot-VS slot blot apparatus (American Bionetics, Hayward, CA), according to the manufacturer's instructions. Hybridization was with primer B noted in Table 1 and Fig. 1. Oligonucleotide labeling employed 100 ng of DNA, 75 pCi [32P]ATP, 2.5 fi\ of 10 x kinase buffer (500 mM Tris HC1 pH 7.6, 100 mM MgCl2, 50 mM dithiothreitol, 1 mM spermidine, 1 mM EDTA), 10 units of T4 DNA kinase adjusted to a final volume of 25 ¡à with distilled water, labeling was carried out by incubation at 37°C for 30 min prior to use. Blots were prehybridized in 5 x SSC, 5x Denhardt's solution, 0.1% NaDodS04 for 11.5 hr at 55°C in Seal-a-Meal bags, the solution was poured off, 32P-labeled oligonucleotide was added (75 /¿Ci) in prehybridization solution and hybridized for 2-3 hr at 42°C or overnight at 4°C, the blot washed 1 x in 2 x SSC, 0.1% NaDodS04 for 20 min at 45 °C, then 3 x in 5 x SSC, 0.1% NaDodS04 for 20 min at 45°C, and exposed to Kodak XRP film at -70°C for 2-72 hr.
RESULTS
Primer design The primers utilized were designed based on sequence homology between the outer surface proteins (OspA and OspB) of B. burgdorferi (Bergstrom et al, 1989) as shown in Fig. 1. Where there were discrepancies in the sequences, degeneracy was included to allow annealing to both gene products. The oligonucleotides were synthesized by the Wistar Institute oligonucleotide synthesis facility, and their sequences
are
shown in Table 1.
Amplification of B. burgdorfei
DNA
These primers are able to amplify B. burgdorferi DNA shown in Fig. 2. DNA was extracted from 1 ml of culture media containing B. burgdorferi or B. hermsii organisms and subjected to amplification by PCR with the various primers. All of the available primers were tested against both B. burgdorferi and B. hermsii in 40 cycles of amplification. All of the primers gave good results on B. as
MOLECULAR DIAGNOSIS OF LYME DISEASE
209
Table 1. Oligonucleotide Primers
Sequence (5'—3')
Primer A »Primer B Primer B' »Primer C »Primer D Primer E
TTGTAAGCAAAGAAAAAAA TTTAAAGAAGGAACTGTTACT AGTAACAGTTCCTTCTTTAA TTAAAAACGCTTTAAAATAA
(GT)(AC)CGGCAA(GA)TA(CT)GAT(CT)TAG(CT)(TC)G ATTTCA(AC)(CT)TGCTGA(CT)CC(CT)TC
Primers
size
Primers
A-C D-C A-E
701 bp 679 bp 664 bp
D-E A-B D-B
»Denotes anti-sense
OSPA
OSPA
bp bp bp
size
B'-C B'-E
285 247
bp bp
Oligonucleotide Primers
550 520 530 560 570 540 GCTGAAAAAACAACATTGGTGGTTAAAGAAGGAACTGTTACTTTAAGCAAAAATATTTCA
I
I I I I I I I I I I
I I I
I I I I I I I I I I
I I I I I I I I I
I I I
I
I
I
I I I
I
GTCGGAAAAACAACAGTGGAAATTAAAGAAGGTACTGTTACTCTAAAAAGAGAAATTGAA 650 620 630 640 600 610
160 140 150 110 120 130 CCTGGTGAAATGAAAGTTCTTGTAAGCAAAGAAAAAAACAAAGACGGCAAGTACGATCTAGCTG I I I I I I I I I I I I I I I I II I I I I I I I I I I M I I I I I I I I I I I I I I TTTAATGGTAATAAAATTTTTGTAAGCAAAGAAAAAAATAGCTCCGGCAAATATGATTTAGTCG 210 180 190 220 230 200
I OSPB
Product size
primer.
Position of
OSPB
Expected Product Sizes
Designation
Product
OSPA
and
I I I I
II
I I I I I I
I I I I I I M
III
III
II
I I I I I I I I I I I I I I I I I I I I
ACCAGCCTAGAAGGATCAGCAAGTGAAATTAAAAATCTTTCAGAGCTTAAAAACGCTTTAAAATAA -" 880 830 840 850 860 870
Nucleotide sequences of OspA and OspB aligned. The nucleotide sequences of OspA and OspB are shown aligned for maximal homology, with the base numbering immediately above and below the sequences. The regions corresponding to the primers utilized in these studies are underlined, and their designation is indicated above the primers. (—) Sense (coding sequence) primer; (—) anti-sense (noncoding strand) primer.
FIG. 1.
burgdorferi DNA, with a few exhibiting "cross-reactivity" on B. hermsii DNA, specifically A&C, A&B, and D&B. Interestingly, when these reaction products were transferred to nitrocellulose and probed with "P-labeled primer B, hybridization was seen only to those B. hermsii amplified bands that contained primer B in the amplification process (data not shown). Results were also compared for DNA extracted from E. coli human peripheral blood mononuclear cells, and salmon sperm DNA (data not shown). These templates did not give rise to PCR products, further indicating the specificity of the primers used.
Amplification
in the presence
of extraneous DNA
For this technique to be useful for clinical specimens, DNA must also be detectable in the presence of extraneous DNA. Therefore, B. burgdorferi organisms were mixed in varying dilutions in 1 ml of whole heparinized blood and DNA extracted from the mixture. Following ethanol precipitation, the DNA was resuspended in 50 fi\ of water, and 5 fil utilized for PCR analysis as noted above. Using this protocol, only the highest concentration of the highest dilution of DNA amplified with the A—C primers gave a specific product. This was likely due to the interfering ef-
210
12
WILLIAMS ET AL.
3 4
5 6
7 8
9 10 11 12 1314 1516
1718
Utility of oligonucleotide primers. DNA from B. burgdorferi strain 38210 (odd-numbered lanes) or B. hermsii strain 38209 (even-numbered lanes) was amplified with the PCR primer pairs noted in Table 1. The primer pairs were: lanes 1 and 2, A—C; lanes 3 and 4, D—C; FIG. 2.
lanes 5 and 6, A—E; lanes 7 and 8, D—E; lanes 9 and 10, A-B; lanes 11 and 12, D-B; lanes 13 and 14, B'-C; lanes 15 and 16, B'-E; lanes 17 and 18, ¿X174 Hae III and \Hind III molecular weight markers. The additional bands seen in some lanes may be due to multimerization of PCR products or mispriming of nontarget sequences.
feet of extraneous DNA in the sample. A nested set of priwas next utilized to allow detection of smaller amounts of DNA. Similarly prepared samples were subjected to nested PCR utilizing first A—C followed by D—E primers. This strategy allowed the B primer to be utilized as a probe in hybridization studies. The results are presented in Fig. 3. DNA extracted from heparinized blood (1 ml) spiked with decreasing numbers of B. burgdorferi organisms was amplified at 1:1 and 1:10 dilutions. In the concentration range used, only the 1:10 dilution of DNA elicited an appropriate sized band (Fig. 3A). The presence of high concentrations of cellular DNA may interfere with specific priming at other dilutions. As few as 400 organisms were detected in this procedure. When the gel was transferred to nylon filters and probed with "P-labeled primer B, only the appropriate lanes could be visualized by hybridization (Fig. 3B). When the remaining reaction products were directly dotted onto nylon filters and probed, again only the appropriate samples showed hybridization (Fig. 3C). The hybridization procedure was approximately fivefold more sensitive than electrophoresis followed by ethidium bromide staining, but both techmers
niques appeared equally specific.
effort to reduce the amount of contaminating protein/DNA, an additional method to increase the specific signal was tested. In these experiments, blood was drawn into tubes containing EDTA instead of heparin, organisms were added at varying concentrations, and the red blood cells were lysed with a 5- to 10-fold excess of Geys' solution. Differential centrifugation was used to separate partially the remaining white blood cells and spirochetes. The white blood cells were pelleted at 1,000 rpm for 10 min (first pellet), and the supernatant then spun at 10,000 rpm for 30 min (second pellet). DNA was extracted from each pellet and the DNA subjected to amplification by the A—C followed by the D—E primers. As controls, highspeed pellets of organisms diluted in phosphate-buffered saline (PBS) or urine were also tested. The results are shown in Fig. 4. Both the first and second pellets of B. In
an
c
I
#
Ä
«¡I
FIG. 3. Amplification of B. burgdorferi DNA sequences from organisms diluted in heparinized blood. Borrelia burgdorferi was grown as described in Materials and Methods. One milliliter containing 4 x 10* spirochetes, was serially diluted in 1-ml aliquots of heparinized human blood. The cells were lysed in DNA digestion buffer, extracted with phenol/chloroform/isoamyl alcohol, and the DNA precipitated and resuspended in 50 fil of water. Five microliters of DNA (even lanes) or 0.5 ¡a of DNA (odd lanes) was amplified initially with primers A and C. Five microliters of the reaction products were reamplified with primers D and E. Ethidium bromide-stained reaction products are shown in Fig. 3A. The gel was transferred to nylon filters and probed with 32P-labeled primer B, with the results shown in Fig. 3B. In addition, 50 pi of the reaction products were dotted onto nylon filters and similarly probed, with the results shown in Fig. 3C. The lanes correspond to the following number of spirochetes in the original reaction mixture: 1, 4 X 10*; 2, 4 x 10*; 3, 4 x 103; 4, 4 x 104; 5, 4 x 102; 6, 4 x 103; 7, 4 x 10; 8, 4 x 102. The higher-molecular-weight bands present in B5 and B7 are probably multimers of the PCR products inapparent on ethidium bromide staining. There is also a smudge artifact overlying B3.
burgdorferi produced clear, distinct bands when amplified, although the intensity of the bands for the second pellet was of greater intensity at higher dilutions of organisms. The organisms diluted in PBS or urine gave similar results. The strategy of lysing the red blood cells, employing differential centrifugation and avoiding heparin, added to the sensitivity of detection. In any case, this approach allowed detection of as few as 100 organisms in mock clinical samples from whole blood.
MOLECULAR DIAGNOSIS OF LYME DISEASE
12 3 4 5
211
6 7 8 9 10
1112131415
16 17 18 19 20
FIG. 4.
Differential centrifugation to enhance PCR detection of B. burgdorferi DNA. Aliquots of 106 B. burgdorferi organisms were serially 10-fold-diluted in 2 ml of EDTA-anticoagulated blood (lanes 1-10), urine (lanes 11-15), or phosphate-buffered saline (PBS) (lanes 16-20). The urine and PBS were centrifuged at 10,000 rpm for 30 min, and pellets utilized. The blood was combined with 2 ml of Gey's solution on ice for 5 min to lyse most of the red blood cells, and the cells were centrifuged at 1,000 rpm for 5 min (cell pellet 1-5). The supernatant was further centrifuged at 10,000 rpm for 30 min (bacterial pellet 6-10). DNA was extracted from the pellets as noted above, and following precipitation, resuspended in 50 ¡A of water. Five microliters of the reatcion product was amplified with primers A and C, and 10 fi\ of the reaction product reamplified with primers D and E. The reaction products were analyzed on a 3% agarose gel with ethidium bromide staining. The following numbers of spirochetes in the original reaction mixture correspond to the lanes: 105, ~
=
=
lanes 1, 6, 11, and 16; 10", lanes 2, 7, and 12; 103, lanes 3, 8, 13, and 17; 102, lanes 4, 9, 14, and 18; 10, lanes 5, 10, 15, and 19; 1, lane 20.
of B. burgdorferi from patients Detection
in blood and urine
5
6 7
8 9 10
We next examined blood and urine specimens from 2 individuals. In these cases, both blood and urine specimens gave positive results, whereas negative controls remained Detection of Borrelia DNA sequences in the penegative. The clinical characteristics of two of these cases FIG. 5. blood of an infected individual. EDTA anti-coagripheral are summarized here. ulated blood was obtained from patient J.D., and proJ.D. is a 57-year-old white male machinist with a 12-year cessed to the scheme described in the legend to history of rheumatoid factor-positive palindromic arthri- Fig. 4. according DNA was amplified with primers A and C followed tis. The patient was first seen by us in December, 1990, by D and E, as described in Fig. 4. Reaction products were with a change in the pattern from his previous course of run on a 3% agarose gel and stained with ethidium broarthritis. This consisted of 9 months of constant severe mide (A), followed by capillary transfer to nylon filters pain and swelling in the small joints of the hands and both and probing with 32P-labeled primer B (B). The DNA amankles that prevented him from working. Physical exami- plified shown for each lane was: lane 1, negative control nation confirmed a symmetrical swelling, erythema, and human DNA; lanes 2-4, positive control B. burgdorferi tenderness in multiple small joints of the hand along with DNA at three 10 X dilutions; lanes 5-7, three 10 x dilulimited range of shoulder motion bilaterally due to pain, tions of cell pellet DNA; 8-10, three 10 x dilutions of bacterial pellet DNA. Positive controls were obtained from but was otherwise unremarkable. DNA extracted from 100 pi of B. burgdorferi organisms Initial evaluation revealed a rheumatoid factor of grown in culture. 1:1,280 (normal < 1:80) by latex fixation, antinuclear antibody 1:640, speckled pattern, and normal CH50. A chest radiograph demonstrated increased interstitial markings. The patient initially received 20 mg of Prednisone daily CG. is an 11-year-old white male who noted swelling of and 1 gram of sulfasalazine daily with no improvement his left knee 5 months prior to our evaluation. There was over 1 month. Further history revealed that the patient no history of trauma, fever, rash, recent immunization, lived in a wooded area and had had a circular red rash on tick bite, any other joint involvement, back pain, or an ankle several months before his arthritis had intensidysuria. Past medical history and family history were unrefied. An ELISA for Lyme, which had been negative in markable. The pediatrician noted a mild effusion of the June, 1990, was now positive. PCR results from blood are left knee. Radiograph was negative as was ANA. The Hgb shown in Fig. 5. Note that both the cellular and bacterial was 12.6 g, WBC 7,700/cu mm, platelet count of 474,000/ blood samples were positive. Faint positive results in the cu. mm, and the ESR was 89 mm/hr. A Lyme titer oburine were most easily detectable following Southern trans- tained 1 month after presentation was reportedly positive fer of the blot and subsequent probing and hybridization by ELISA. The patient was diagnosed as having Lyme (data not shown). Within 1 week of receiving the Ceftria- arthritis and was treated for 6 weeks with amoxicillin. xone, the patient had a marked improvement of joint pain Swelling in the left knee improved, occurring only with and swelling. Repeat PCR analysis following antibiotic physical activity. During this therapy, the right knee subsetherapy was negative. quently developed an effusion. The patient was treated
212
WILLIAMS ET AL.
with on ASA three times daily for 1 month without resolution and was referred to the Children's Hospital of Pennsylvania for evaluation. At our initial evaluation, his exam was remarkable for a moderated effusion, warmth, and a reduction in range of motion of the right knee. Laboratories revealed a Hb of 12.6 g, a WBC of 9,900/cu mm, platelet count of 356,000/ cu mm, and an ESR of 55 mm/hr. Gram stain was negative for white cells and bacteria and bacterial culture of the synovial fluid was negative. Lyme titer by ELISA was positive at 43.2 SD above the mean (normal is 2 SD). Western blot was also positive. Blood, urine, and synovial fluid were positive by PCR (data not shown). A 2-week course of daily intravenous ceftriaxone was initiated. Follow-up 1 month later the patient was entirely asymptomatic. Repeat PCR following therapy was negative. Together, these results indicate the feasibility of detecting B. burgdorferi sequences in clinical specimens using a nested approach to PCR.
DISCUSSION Several groups have suggested PCR as a potential diagnostic tool for Lyme borreliosis (Rosa and Schwan, 1989; Neilsen et al, 1990; Wallich et al, 1990). Borrelia burgdorferi DNA has recently been detected in clinical specimens by PCR. Persing et al used primers derived from the OspA sequence, and were able to detect B. burgdorferi sequences in Ixodces dammini ticks collected from the field, as well as museum specimens (Persing et al, 1990a,b). Malloy et al utilized primers derived from the OspA sequence to amplify a 309-bp fragment followed by oligonucleotide probing (Malloy et al, 1990). They utilized this technique to detect B. burgdorferi DNA in the blood and urine of an infected canine. Other groups have utilized PCR for detection of B. burgdorferi DNA from patient samples (Nishio et al, 1990; Goodman et al, 1991). Their studies included detection of B. burgdorferi in the urine of infected individuals. This approach eliminated the potential of contaminating cellular DNA and gave excellent results, as 4 patients with suspected Lyme disease who gave positive results on PCR showed an excellent clinical response to antibiotic therapy. The utility of urine testing is quite apparent for Lyme disease, as this is the basis for immunodetection assays for spirochetal antigens (Hyde et al, 1989). While this group did not concentrate their samples by centrifugation, simple extraction and amplification was sufficient for detection. We utilized three complementary strategies to address the problem of contaminating DNA. The first is the use of primers derived from sequences present in > 1 copy. Our primers were designed based on sequence similarity between the OspA and OspB genes. This gives the primers a twofold advantage as functionally two copies of the gene product would serve as templates for amplification. Our approach also targets highly conserved sequences that might be less susceptible to genetic drift. The second is the use of nested PCR. By preamplifying DNA extracted from clinical samples, the percentage of B. burgdorferi DNA
template is significantly enriched prior to the second amplification. This allows detection of relatively fewer organisms in the presence of large amounts of extraneous DNA. We also utilized differential centrifugation to eliminate as much extraneous cellular DNA as possible. We sought to take advantage of the substantial size difference between the spirochetes and peripheral blood cells. Red blood cells are first dispensed with by lysis in hypotonie
ammonium chloride solution. White blood cells were then pelleted at low speed, followed by a high-speed spin to pellet potential smaller pathogens. Under these conditions, DNA was extracted from as much as 10-20 ml of blood without the concomitant large quantities of contaminating cellular DNA present. When clinical samples are processed by the methods described here, we observed patients with positive results for both blood and urine (Fig. 5). These were the results observed in patients with clinical histories suggestive of Lyme disease. In one case, the clinical scenario was quite difficult to sort out, as the patient had preexisting palindromic rheumatism, and this was markedly exacerbated by tick bite exposure. In such an individual, with polyclonal hypergammaglobulinemia, positive antibody tests were difficult to interpret. This problem was even more apparent in the second individual, who had previous clinical and sérologie evidence of Lyme disease, was treated, and was thought to be reinfected. In such cases, sérologie testing and tests for other immune responses are difficult to interpret. Thus, direct detection of B. burgdorferi spirochetes can be of clinical utility in clinical scenarios where other tests are difficult to interpret. The data presented here demonstrates that, with respect to B. burgdorferi, PCR diagnosis is the method of choice for clarification of difficult clinical reexposure cases and, similarly, may be of great utility in early disease. The protocols described herein may also be of general utility in detection of other infectious agents.
ACKNOWLEDGMENTS W.V.W. is supported in part by grants from the NIH. P.C. is supported in part by the NIH, and D.B.W. is supported in part by grants from AmFAR and the CTR.
REFERENCES AUSUBEL, F., BRENT, R., KINGSTON, R., MOORE, D., SEIDMAN, J., SMITH, J., and STRUHL, K. (1989). Current Protocols in Molecular Biology. (Greene Publishing Associates and Wiley-Interscience, John Wiley & Sons, New York). BARBOUR, A.G. (1984). Isolation and cultivation of Lyme disease spirochetes. Yale J. Biol. Med. 57, 521-525. BARBOUR, A., HEILAND, R., and HOWE, T. (1985). Heterogeneity of major proteins in Lyme disease borreliae: A molecular analysis of North American and European isolates. J. Infect. Dis. 152, 478-484. BERGSTROM, S., BUNDOC, V., and BARBOUR, A. (1989). Molecular analysis of linear plasmid-encoded major surface proteins, OspA and OspB, of the Lyme disease spirocheaete Borrelia burgdorferi. Mol. Microbiol. 3, 479-486.
213
MOLECULAR DIAGNOSIS OF LYME DISEASE COLEMAN, J., and BENACH, J. (1987). Isolation of antigenic components from the lyme disease spirochete: Their role in early diagnosis. J. Infect. Dis. 155, 756-765.
DATTWYLER, R.J., VOLKMAN, D.J., LUFT, B.J., HALPERIN, J.J., THOMAS, J., and GOLIGHTLY, M.G. (1988).
Seronegative lyme disease: Dissociation of specific T- and B-lymphocyte responses to Borrelia burgdorferi. N. Engl. J.
Med. 319, 1441. DUFFY, J., MERTZ, L., WOBIG, G., and KATZMANN, J.
(1988). Diagnosing lyme disease: The contribution of sérologie testing. Mayo Clin. Proc. 63, 1116-1121. GOODMAN, J., JURKOVICH, P., KRAMBER, J., and JOHNSON, R. (1991). Molecular detection of persistent Borrelia burgdorferi in the urine of patients with active Lyme disease. Infect. Immun. 59, 269-278. GRODZICKI, R., and STEERE, A. (1988). Comparison of immunoblotting and indirect enzyme-linked immunosorbent assay using different antigen preparations for diagnosing early Lyme disease. J. Infect. Dis. 157, 790-797. HYDE, F., JOHNSON, R., WHITE, T., and SHELBURNE, C. (1989). Detection of antigens in the urine of mice and humans infected with Borrelia burgdorferi, etiologic agent for lyme disease. J. Clin. Microbiol. 27, 58-61. LUFT, B., JIANG, W., MUNOZ, P., DATWYLER, R., and GOREVIC, P. (1989). Biochemical and immunological characterization of the surface proteins of Borrelia burgdorferi. Infect. Immun. 57, 3637-3645. MALLOY, D., NAUMAN, R., and PAXTON, H. (1990). Detection of Borrelia burgdorferi using the polymerase chain reaction. J. Clin. Microbiol. 28, 1089-1093. NEILSEN, S., YOUNG, K., and BARBOUR, A. (1990). Detection of Borrelia burgdorferi DNA by the polymerase chain re-
action. Mol. Cell. Probes 4, 71-79. NISHIO, M.J., LIEBLING, M.R., RODRIGUEZ, A., SIGAL, L.H., and LOUIE, J.S. (1990). Identification of Borrelia burgdorferi using interrupted polymerase chain reaction. Arthritis Rheum. 33, S84. PERSING, D., TELFORD, S., RHYS, P., DODGE, D., WHITE, T., MALAWISTA, S., and SPEILMAN, A. (1990a).
Detection of the Borrelia burgdorferi DNA in museum mens of Ixodes dammini ticks. Science 249, 1420-1423.
speci-
PERSING, D., TELFORD, S., SPIELMAN, A., and BARTHOLD, S. (1990b). Detection of Borrelia burgdorferi infection in Ixodes dammini ticks with the polymerase chain reaction. J. Clin. Microbiol. 28, 566-572. ROSA, P., and SCHWAN, T. (1989). A specific and sensitive assay for the Lyme disease spirochete Borrelia burgdorferi using the polymerase chain reaction. J. Infect. Dis. 160, 1018-1029. WALLICH, R., MOTER, S., SIMON, M., EBNET, K., HEIBERGER, A., and KRAMER, M. (1990). The Borrelia burgdorferi flagellum-associated 41-kilodalton antigen (flagellin): Molecular cloning, expression and amplification of the gene. Infect. Immun. 58, 1711-1719. Address
reprint requests
to:
Dr. David B. Weiner
The
Biotechnology Center
The Wistar Institute 3600 Spruce Street Philadelphia, PA 19104-4268 Received for publication November 29, 1991; accepted January 15, 1992.
