Serial Kinetics of the Antibody Response against the Complete ...

Serial Kinetics of the Antibody Response against the Complete Brucella melitensis ORFeome in Focal Vertebral Brucellosis Anthony P. Cannella,a Jennifer C. Lin,a Li Liang,b Vidya Atluri,c Eduardo Gotuzzo,d Philip L. Felgner,b Renee M. Tsolis,c and Joseph M. Vinetza,d,e Division of Infectious Diseases, Department of Medicine, University of California—San Diego School of Medicine, La Jolla, California, USAa; Division of Infectious Diseases, Department of Medicine, University of California—Irvine School of Medicine, Irvine, California, USAb; Department of Medical Microbiology and Immunology, University of California—Davis, Davis, California, USAc; Alexander von Humboldt Institute of Tropical Medicine, Universidad Peruana Cayetano Heredia, Lima, Perud; and Department of Microbiology, Faculty of Sciences, Laboratory of Research and Development, Universidad Peruana Cayetano Heredia, Lima, Perue

Human brucellosis is a common zoonosis worldwide. Here we present a case of focal vertebral brucellosis in a 71-year-old Mexican-American woman who contracted infection from unpasteurized goat milk. Standard agglutination serology was negative; the diagnosis was established by the isolation of Brucella melitensis from abscess fluid. A B. melitensis protein microarray comprised of nearly all proteins encoded by the bacterial genome was used to determine the kinetics of this patient’s antibody responses to the complete collection of open reading frames existing in the genome (ORFeome). Three patterns of antibody responses against B. melitensis antigens were seen for serum samples obtained on days 0 (pretreatment), 14, 49, 100, and 180: (i) stable titers over time, (ii) a steady fall in titers, and (iii) an initial rise in titers followed by declining titers. Sera from this patient with chronic brucellosis recognized some of the same B. melitensis proteins as those recognized by sera from acute/subacute, blood culture-positive brucellosis patients but also recognized a distinct set of proteins. This study is the first to determine the kinetics of the human antibody responses to the complete repertoire of proteins encoded by a bacterial genome and demonstrates fundamentally different immunopathogenetic mechanisms between acute human brucellosis and chronic human brucellosis. While an extension of these findings to a larger patient population is necessary, these findings have important clinical and diagnostic implications and lead toward new insights into the fundamental immunopathogenesis of brucellosis.

H

uman brucellosis is a common bacterial zoonosis worldwide. Every year, more than 500,000 new cases are reported globally, with risk factors associated primarily with the consumption of unpasteurized dairy products and occupationally involving direct contact with infected animals (slaughterhouse workers and veterinarians) and among microbiologists handling cultures in the laboratory (12, 13). Brucella melitensis (associated primarily with goats and sheep) and Brucella abortus (associated primarily with cattle) are the two species of Brucella found most frequently in human cases (11). While brucellosis is classically associated with diverse clinical manifestations, fever is invariably present to some degree. Clinical manifestations are classically divided into acute/subacute and chronic presentations, yet some authors have called such a distinction of clinically limited utility (12). Acute/subacute febrile illness is characterized by generalized symptoms (fever, sweats, and weight loss) over weeks to months. Brucellosis may also manifest as focal disease localized to a number of organ systems, most commonly involving the axial and large joints (e.g., spondylitis, sacroiliitis, and arthritis) but also liver, spleen, brain, and cardiac valves. The initial diagnosis of brucellosis is often made by various agglutination tests, including the Rose Bengal screening test (4, 13) and the quantitative tube agglutination test; in cases of high suspicion where standard agglutination tests are negative, an indirect Coombs test or the more recent Brucella enzyme-linked immunosorbent assay (ELISA) may be used (5, 13). The gold standard of diagnosis is the isolation of the organism from a culture, either from blood (usually in acute/subacute disease) or from tissue/body fluids in cases of focal disease. Blood cultures are typically negative in musculoskeletal brucellosis (13). While serum agglutination tests are useful for the diagnosis of brucellosis, they are imperfect and cannot necessarily diagnose

922

jcm.asm.org

chronic infection, which may require cultures of affected sites or bone marrow to prove the diagnosis (6). False-negative serologies may occur with serum agglutination tests due to the “prozone phenomenon,” where an excess of antibodies results in the inhibition of agglutination at low serum dilutions (12). However, dilution of serum to rule out the prozone phenomenon is warranted. The Coombs test, although not routine diagnostic practice, and the recently developed BrucellaCapt assay can be useful alternatives for the detection of anti-Brucella antibodies even when agglutination tests are negative (1, 2). Here we report the case of a patient with Brucella melitensis vertebral osteomyelitis and psoas abscess complicated by urinary retention due to spinal cord compression. The serial kinetics of this patient’s antibody responses against the complete set of proteins predicted to be encoded by the B. melitensis genome (the so-called ORFeome) were determined by use of a protein microarray (9). While we (9) and other groups (17) have examined the Brucella ORFeome previously in relation to humoral immunity and molecular pathogenesis, a comprehensive analysis of the kinetics of host antibody responses to the B. melitensis ORFeome has not been done. In the presently reported case of chronic, focal

Received 1 August 2011 Returned for modification 9 September 2011 Accepted 21 December 2011 Published ahead of print 4 January 2012 Address correspondence to Joseph M. Vinetz, [email protected]. Supplemental material for this article may be found at http://jcm.asm.org/. Copyright © 2012, American Society for Microbiology. All Rights Reserved. doi:10.1128/JCM.05298-11

0095-1137/12/$12.00

Journal of Clinical Microbiology

p. 922–926

Protein Microarray Analysis in Vertebral Brucellosis

FIG 1 Radiographic imaging of a case of brucellar spondylodiscitis. (A) Plain lumbosacral radiograph at initial presentation, demonstrating vertebral collapse at L1-L2. (B) T1-weighted sagittal magnetic resonance image (MRI) demonstrating edematous changes centered at the L1-L2 disk with surrounding destructive changes involving the inferior L1 vertebral body and superior L2 vertebral body. Not shown is a ⬃4-cm left-sided psoas abscess. (C) T2-weighted MRI demonstrating spinal cord compression. (D) Plain lumbosacral radiograph 6 months after initial presentation, showing posterolateral fusion with pedicle screw placement at T11, T12, L3, and L4 with bilateral fusion rods and anterior corpectomy and cage placement.

musculoskeletal brucellosis, the patient’s humoral immune responses differed fundamentally from those of patients with blood culture-positive, acute/subacute brucellosis. A comprehensive examination of the universe of antibody responses against the com-

plete B. melitensis ORFeome reveals novel insights into human immune responses to this intracellular bacterium. These findings have potential relevance to the immunopathogeneses of other intracellular bacterial pathogens.

FIG 2 Comparison of ORFeome-level antibody responses of a patient with chronic, vertebral brucellosis with those of Peruvian acute/subacute brucellosis patients by protein microarray analysis. Responses are classified into three groups: group I, 23 antigens recognized both by sera from our patient and by sera from confirmed Peruvian acute/subacute brucellosis patients (9); group II, 50 antigens recognized by sera from our patient but not by sera from Peruvian acute/ subacute brucellosis patients (BMEI1079 was cross-reactive); and group III, 11 antigens recognized by sera from Peruvian acute/subacute brucellosis patients but not by sera from our patient (BMEI0228 was cross-reactive). The signal intensity (reported in arbitrary units) is indicated by the color bar (where red is the highest signal intensity and green is the lowest), which corresponds to relative antibody titers. C-RB⫹, culture negative, Rose Bengal positive.

March 2012 Volume 50 Number 3

jcm.asm.org 923

Cannella et al.

924

jcm.asm.org


Protein Microarray Analysis in Vertebral Brucellosis

CASE REPORT

A 71-year-old Mexican-American woman was admitted to the University of California—San Diego (UCSD) Medical Center with progressive lower back pain, fever, night sweats, anorexia, and a 10-kg weight loss over a 2- to 3-month period. The patient denied weakness in the lower extremities, numbness/tingling, and urinary/bowel incontinence but had difficulty voiding urine. The patient’s past medical history was significant for a remote history of left ovary tuberculosis resulting in oophorectomy at the age of 19 years. She was born in central Mexico and moved to the United States in the 1960s. She had frequently visited Mexico, staying with relatives for weeks to months, and while there, she regularly ingested unpasteurized goat milk and cheese from a local farmer. Upon admission, she had fever, did not have malodorous sweating, and was in moderate distress from pain. Her musculoskeletal examination was significant for severe kyphosis at L1 and tenderness over the lower spine. Straight-leg raise of the right leg was painful; the right hip had a limited range of motion. The physical examination was not otherwise remarkable. Laboratory testing showed a leukocyte count of 7.2 ⫻ 103 leukocytes/mm3, with a predominance of neutrophils, and a hemoglobin level of 10.4 g/dl. The C-reactive protein level was 9.7 mg/dl; the erythrocyte sedimentation rate was 59 mm/h. Radiography showed significant inflammatory and destructive findings for the lumbar spine (Fig. 1). The patient underwent emergent laminectomy with a bone biopsy and culture that yielded a pure culture of B. melitensis at 5 days. Blood cultures were negative. Other testing showed Brucella serum agglutination negative at a dilution of ⬍1:20; a reevaluation of the patient’s serum agglutination test at a dilution of 1:100 yielded a positive test, indicating a prior false-negative result, possibly related to interfering antibodies, as has been seen previously for chronic forms of brucellosis. IgM and IgG ELISAs (GenWay Biotech, San Diego, CA) were both positive (IgM, 1.39 units [positive, ⬎1.11 units]; IgG, 2.88 units [positive, ⬎1.11 units]). Indirect Coombs and BrucellaCapt assays were not performed, given the positive culture result that confirmed infection. The patient was given 2 weeks of intravenous gentamicin at 5 mg/kg of body weight/day for 14 days and a total of 6 months of oral doxycycline (100 mg/day) and rifampin (600 mg/day). Debilitating back and radiating radicular pain in both legs resolved after surgical intervention 3 and 4 months after the initial presentation, and at 9 months, she had complete clinical recovery. MATERIALS AND METHODS Serum samples were collected on days 0, 14, 49, 100, and 180 (day 0 was the initial presentation prior to surgery and initiation of antimicrobials). The patient’s sera were simultaneously used to probe a newly fabricated B. melitensis microarray comprised of ⬃3,200 proteins, and the results were compared with those of a recently reported analysis of acute and subacute blood culture-positive brucellosis patients in Peru (9, 10). The patient’s responses were compared to results from blood culture-positive Peruvian

B. melitensis patients and from asymptomatic subjects from a region of Peru where brucellosis is endemic (9). The study of human subjects was approved by the Humans Research Protections Program of the University of California—San Diego and the Comité de Ética of the Universidad Peruana Cayetano Heredia, Lima, Perú (9).

RESULTS AND DISCUSSION

A protein microarray analysis carried out on serum samples from days 0, 14, 49, 100, and 180 showed three distinct sets of antigens recognized by sera from our patient with focal vertebral brucellosis (Fig. 2 and see Table S1 in the supplemental material) compared to those of previously reported Peruvian patients and controls (9): 23 antigens were recognized in common with Peruvian patients with acute/subacute brucellosis, 50 antigens were recognized by sera from our patient but not sera from the Peruvian patients, and 11 antigens were recognized by sera from the Peruvian patients but not our patient. The kinetics of the antibody responses to the complete ORFeome of B. melitensis antigens (Fig. 2) were further divided into three groups of antigens characterized based on kinetic changes (Fig. 3). The group III antigens shown in Fig. 2 were not differentially recognized by the patient’s sera from this case study. These three different categories of antigens recognized by the patient’s antibody response were as follows (Fig. 3A and see Table S1 in the supplemental material): group I, stable, high-level antibody responses throughout the 180 days of treatment; group II, antibody responses that decreased after treatment; and group III, antibody responses that initially rose at least 2- to 5-fold after treatment was initiated and then had variable kinetics thereafter. One protein of particular interest in group III (Fig. 3A), BMEI1938, is an oligopeptide transport ATP-binding protein (OppD) from the B. melitensis BMEI1938 gene. The top hit for this protein from a BLAST search identified a Mycobacterium tuberculosis oligopeptide transport ATPbinding protein (42% homology and an E value of 10⫺103) (GenBank accession number NP_218180.1) shown previously to alter cytokine release by macrophages and to prevent apoptosis (3). Compared to the other antigen-specific antibody responses that either remained stable at high titers or declined over time (groups I and II, respectively), the kinetics of antibody responses to OppD and other group III proteins are notable and require explanation. Low-level antibody responses to the remaining 3,000 or more proteins in the B. melitensis ORFeome were seen not only among the patients with acute brucellosis but also for our patient described here and among negativecontrol patients from a country where this disease is endemic. We postulate two possible but contrasting mechanisms that could explain this pattern. On the one hand, it is possible that once antimicrobial therapy has killed sufficient numbers of Brucella organisms, their proteins and other components are released, which induces antibody responses in this delayed pattern. We believe that this is unlikely, because delayed rises in levels of antibody responses were observed for only a minority of proteins encoded by the B. melitensis

FIG 3 Analysis of the serial kinetics of ORFeome-level antibody responses. The group I and II antigens shown in Fig. 2 are further divided into 3 groups of antigens characterized based on kinetic changes. The antibody responses are classified into three groups: group I, stable, high-level antibody responses throughout the 180 days of treatment; group II, antibody responses that decreased after treatment; and group III, antibody responses that initially rose at least 2- to 5-fold after treatment was initiated then had variable kinetics thereafter. (A) Heat map as described in the legend of Fig. 2. (B) Graphical representation of examples of antigens from each group. The signal intensity (reported in arbitrary units) is indicated by the color bar (where red is the highest signal intensity and green is the lowest), which corresponds to relative antibody titers. Panels I to III are representative of groups I to III; panel IV is a graphical representation of the averages for all three groups throughout the 180-day timeline. LPS, lipopolysaccharide.

March 2012 Volume 50 Number 3

jcm.asm.org 925

Cannella et al.

genome. Alternatively, and perhaps more convincingly, we hypothesize that key macrophage functions (phagolysosomal fusion to kill pathogens and antigen presentation) are actively suppressed by live Brucella organisms but that once the bacteria within macrophages are no longer viable, macrophages are released from immune stasis. Such molecules might be involved in the long-hypothesized but poorly understood mechanisms of T cell anergy in chronic brucellosis (15, 16). We hypothesize that B. melitensis OppD may have a function similar to that found for the M. tuberculosis protein, since both are intracellular pathogens of macrophages occupying phagolysosomes (3); further work will be required to test this hypothesis. Therefore, the latter explanation would lead to the hypothesis that group III proteins may be involved in the ability of B. melitensis to actively downmodulate macrophages, contributing to intracellular persistence using mechanisms reminiscent of those of pathogens such as M. tuberculosis, Salmonella enterica serovar Typhi, and others. The 26-kDa immunogenic protein is known to be strongly recognized as a serological diagnostic antigen in cases of human brucellosis (18) and, notably, has been associated with a protective T cell response in murine models of brucellosis (8). The BMEII0032 (VirB8) protein has been noted to be an important virulence factor in Brucella spp. (14), most recently demonstrated with a lethal murine model of Brucella microti infection as well as for other newly identified Brucella species (7). There are important limitations of this study that require further investigation. First, the data presented here are from one patient with chronic vertebral brucellosis; a larger number of well-characterized patients with other focal and chronic brucellosis syndromes will need be to be compared with acute/subacute brucellosis patients to delineate differences in antibody responses between the various forms of brucellosis. Nonetheless, while it is premature to draw precise conclusions about the mechanisms of immunity and pathogenesis in relation to the proteins described in the present study, genomic-level investigations of both antibody- and cell-mediated immune responses in human brucellosis are feasible and promise to yield new insights into different clinical forms of this disease. This study is the first to analyze the kinetics of the humoral immune response to the complete ORFeome of a bacterial pathogen. We report the case of a patient with chronic vertebral brucellosis whose antibody responses to the complete protein antigen repertoire of B. melitensis had important similarities with and differences compared to blood culture-positive B. melitensis-infected patients with acute/subacute disease. Furthermore, in an analysis of the kinetics of this patient’s antibody responses over 6 months of antimicrobial treatment, we found that an important and novel pattern of antibody reactivity included a substantial rise in titers against a subset of B. melitensis antigens after antibiotic treatment was initiated. This genomic-level analysis provides new leads for diagnostic development applicable to different forms of brucellosis. These data set the stage for identifying potential roles of hitherto-unknown Brucella proteins in mediating host-pathogen interactions and mechanisms of immunopathogenesis. Future work will be needed to define the diagnostic utility and performance of the set of antigens described in this report for difficult-

926

jcm.asm.org

to-diagnose syndromes of brucellosis, including focal and chronic cases. ACKNOWLEDGMENTS We thank Randy Taplitz of the University of California, San Diego, Division of Infectious Diseases, Weijen Chang of the Division of Hospital Medicine, and the medical and surgical teams for their excellent care of the patient reported here. We also thank Kalina Campos for performing serologies, Mahesh Atluri for technical assistance, and Paula Maguina of the University of California, San Diego, who made essential research contributions in terms of ethics management and international coordination as well as logistical support for this project. We declare that we have no conflicts of interest. This work was supported by U.S. National Institutes of Health grants K24AI06890 (J.M.V.), U01AI075420 (J.M.V.), T32AI007036, U01AI078213 (P.L.F.), and U54AI065359 (P.L.F.) and by SBIR grant R43AI06816601 (P.L.F.).

REFERENCES 1. Araj GF. 2010. Update on laboratory diagnosis of human brucellosis. Int. J. Antimicrob. Agents 36(Suppl 1):S12–S17. 2. Casanova A, Ariza J, Rubio M, Masuet C, Diaz R. 2009. BrucellaCapt versus classical tests in the serological diagnosis and management of human brucellosis. Clin. Vaccine Immunol. 16:844 – 851. 3. Dasgupta A, et al. 2010. An oligopeptide transporter of Mycobacterium tuberculosis regulates cytokine release and apoptosis of infected macrophages. PLoS One 5:e12225. 4. Diaz R, Casanova A, Ariza J, Moriyo I. 2011. The Rose Bengal test in human brucellosis: a neglected test for the diagnosis of a neglected disease. PLoS Negl. Trop. Dis. 5:e950. 5. Franco MP, Mulder M, Gilman RH, Smits HL. 2007. Human brucellosis. Lancet Infect. Dis. 7:775–786. 6. Gotuzzo E, Carrillo C, Guerra J, Llosa L. 1986. An evaluation of diagnostic methods for brucellosis—the value of bone marrow culture. J. Infect. Dis. 153:122–125. 7. Hanna N, et al. 2011. The virB operon is essential for lethality of Brucella microti in the Balb/c murine model of infection. J. Infect. Dis. 203:1129 – 1135. 8. Kumar S, Tuteja U, Batra HV. 2007. Generation and characterization of murine monoclonal antibodies to recombinant 26-kDa periplasmic protein of Brucella abortus. Hybridoma (Larchmt.) 26:322–327. 9. Liang L, et al. 2010. Large scale immune profiling of infected humans and goats reveals differential recognition of Brucella melitensis antigens. PLoS Negl. Trop. Dis. 4:e673. 10. Liang L, et al. 2011. Systems biology approach predicts antibody signature associated with Brucella melitensis infection in humans. J. Proteome Res. 10:4813– 4824. 11. Luna-Martinez JE, Mejia-Teran C. 2002. Brucellosis in Mexico: current status and trends. Vet. Microbiol. 90:19 –30. 12. Pappas G, Akritidis N, Bosilkovski M, Tsianos E. 2005. Brucellosis. N. Engl. J. Med. 352:2325–2336. 13. Pappas G, Papadimitriou P, Akritidis N, Christou L, Tsianos EV. 2006. The new global map of human brucellosis. Lancet Infect. Dis. 6:91–99. 14. Paschos A, et al. 2011. An in vivo high-throughput screening approach targeting the type IV secretion system component VirB8 identified inhibitors of Brucella abortus 2308 proliferation. Infect. Immun. 79:1033–1043. 15. Skendros P, Pappas G, Boura P. 2011. Cell-mediated immunity in human brucellosis. Microbes Infect. 13:134 –142. 16. Thornes RD. 1977. The anergy of chronic human brucellosis. Ir. Med. J. 70:480 – 483. 17. Viadas C, Rodriguez MC, Garcia-Lobo JM, Sangari FJ, Lopez-Goni I. 2009. Construction and evaluation of an ORFeome-based Brucella wholegenome DNA microarray. Microb. Pathog. 47:189 –195. 18. Yang X, Hudson M, Walters N, Bargatze RF, Pascual DW. 2005. Selection of protective epitopes for Brucella melitensis by DNA vaccination. Infect. Immun. 73:7297–7303.
