BCG Infection due to MPT64-Negative Strain: A Diagnostic Challenge

INTRODUCTION

Live attenuated BCG vaccine was introduced to the world in 1921.¹ In countries like India, with a high burden of tuberculosis (TB), the WHO recommends BCG vaccination of all children at birth to prevent tuberculous meningitis and disseminated TB in young children and infants.¹ Among the various BCG strains, Glaxo 1077, Pasteur 1173, Danish 1331, and Tokyo 172 account for almost 90% of the vaccines worldwide.^1,2 Whereas Pasteur 1173 and Danish 1331 are considered as strong strains (higher vaccine efficacy), Glaxo 1077 and Tokyo 172 are weak strains (lower vaccine efficacy).^1,2 The genealogy of BCG vaccine strains based on historical data shows that BCG strains differ in the presence of mpb64 gene and the number of copies of IS6110 gene.^3,4 The ancestral BCG strain obtained before 1925 had two copies of IS6110 and mpb64 genes, but subsequently loss of one IS6110 happened in 1925–1926 followed by loss of mpb64 gene around 1927–1931.^3,4 The mpb64 gene is present in BCG-Japan, BCG-Sweden, BCG-Moreau, BCG-Birkhaug, and BCG-Russia, whereas it is absent in BCG-Danish, BCG-Pasteur, BCG-Glaxo, BCG-Tice, and BCG-Connaught. Complications due to BCG vaccination are influenced by BCG strain, dose, age, technique of immunization, and the underlying immune status of the vaccines.^2,3 The incidence of severe complications ranges from 2 to 1,000 per million vaccinations and is most often seen in patients with primary immunodeficiency disorders such as chronic granulomatous disease, severe combined immunodeficiency, type 1 cytokine axis defect, DiGeorge syndrome, and HIV infection.^1,3 Severe complications may require chemotherapy with antiTB drugs, which get complicated by the inherent resistance to pyrazinamide in all Mycobacterium bovis strains.^1,5 Because of inherent resistance to pyrazinamide, the CDC recommends a triple-drug regimen consisting of rifampicin, isoniazid, and ethambutol with treatment duration being extended to 9 months for treatment of disease due to M. bovis.⁶ Disseminated BCG infection in young children could be an indication of serious immune deficiency, which needs to be investigated and ruled out. Therefore, differentiation of M. bovis BCG from Mycobacterium tuberculosis as well as from nontuberculous mycobacteria (NTM) is crucial for appropriate treatment. Assays for MPT64 protein have been recommended by the Foundation for Innovative New Diagnostics for rapid identification of the M. tuberculosis complex using the BACTEC MGIT 960 liquid culture system (Becton Dickinson Diagnostics, Sparks, MD).⁷ The assays are routinely performed all over the world, both in developed and developing countries, as suggested by two meta-analyses on their performance in accurately identifying the M. tuberculosis complex in liquid culture.^8,9 We report two cases of BCG infection due to M. bovis BCG strains which were MPT64 negative.

SHORT REPORT

Case 2.

A 1-year-and-5-month-old male child, developmentally normal for age, presented with a limp on the left leg and soft tissue swelling of 1 month duration on the anterolateral aspect of the left knee joint. The size of the swelling progressively increased over 2 weeks from about 1 × 1 cm to 2 × 2 cm. There were no complaints of associated fever, pain over swelling, skin changes, or weight loss. The child was immunocompetent with no underlying disorders, and the BCG vaccine administered on the left deltoid region at birth healed without any complication.

Written informed consent was obtained from the legal guardians of both the patients. Pus aspirated from the left supraclavicular lymph node in the first case and tissue from the excised swelling in the second case were subjected to acid-fast stain, aerobic cultures, Mycobacterium tuberculosis complex (MTBC) PCR using Xpert^® MTB/RIF assay (Cepheid Inc., Sunnyvale, CA), and liquid mycobacterial cultures using BACTEC MGIT 960 (Becton Dickinson Diagnostics, Sparks, MD). The clinical mycobacterial isolates were compared with M. tuberculosis H37Ra, MPT64-positive M. bovis BCG-Russia (Moscow 368, Serum Institute of India), and MPT64-negative M. bovis BCG-Denmark (Danish 1331, Green Signal Bio Pharma Pvt Ltd. Chennai, India) strains because in India both these strains are used for vaccination. Genomic DNA was isolated from the cultures of clinical isolates, vaccine strains, and M. tuberculosis H37Ra. To differentiate between M. bovis and the other MTBC complex organisms, gyrB gene was amplified by PCR and the 1,020-bp PCR product was subjected to restriction fragment length polymorphism (RFLP) using TaqI and sequencing. The sequences obtained were analyzed using the BLASTN algorithm. To detect and differentiate M. bovis BCG from M. tuberculosis and M. bovis, PCR amplifications of the RD1 and RD4 regions were performed using the primers, as described earlier (Table 1).¹⁰

Table 1

List of tests performed to examine the specimens and culture isolates

Method	Test performed	Result
Microbiological identification (clinical specimen)	Acid-fast staining	Negative
	BACTEC MGIT liquid mycobacterial culturing	Positive
Rapid Methods (PCR/immunochromatographic)	Xpert MTB/RIF (clinical specimen)	Positive for MTBC/RIF sensitive
	SD Bioline MPT64 card assay (culture isolates)	Negative for MTBC
Molecular biological identification (culture isolates)	gyrB PCR-RFLP using TaqI	M. bovis
	RD1	M. bovis BCG
	RD4	M. bovis/M. bovis BCG
	DNA sequencing gyrB gene	M. bovis BCG Danish 1331

M. bovis = Mycobacterium bovis; MTBC = Mycobacterium tuberculosis complex.

Smear for acid-fast bacilli (AFB) and aerobic cultures of the pus and tissue were negative. Mycobacterial cultures grew AFB showing cord formation. PCR of pus and tissue using Xpert MTB/RIF was positive for MTBC with rifampicin resistance not detected, indicating that both the specimens belong to MTBC. But both the MGIT culture isolates were negative for the MPT64 protein immunochromatographic test (ICT) (Standard Diagnostics, Seoul,. South Korea), suggestive of NTM. To resolve these conflicting results and identify the species of these two MPT64 card–negative and Xpert MTB/RIF-positive MGIT culture isolates, PCR amplification, RFLP, and sequencing methods were carried out.

RFLP patterns of the clinical isolates were identical to those of the vaccine strains using TaqI (Figure 1A and B). The RD1 PCR was used to differentiate M. bovis BCG (196 bp) from M. tuberculosis/M. bovis (146 bp). The RD4 PCR was used to differentiate between M. tuberculosis (172 bp) and M. bovis/M. bovis BCG (268 bp) (Figure 1E and F). RD4 PCR results identified the culture isolates as M. bovis/M. bovis BCG (268 bp), and RD1 PCR identified them as M. bovis BCG. Sequencing of the gyrB and the RD4 PCR products identified the clinical isolates as M. bovis BCG Danish 1331 (Figure 1B and C).

View Full Size

(A) Schematic representation of the TaqI restriction pattern of the 1,020-bp gyrB fragment among the members of the MTBC. (B) RFLP pattern obtained after the digestion of the 1,020-bp gyrB PCR product with the TaqI restriction enzyme. Lane 1: Mycobacterium tuberculosis H37Ra, 2: Mycobacterium bovis BCG-Russia (Moscow 368) (Serum Institute of India), 3: M. bovis BCG-Denmark (Danish 1331) (Green Signal Biopharma, India), 4 and 5: MGIT culture-positive and MPT64-negative samples, Lane M: 100-bp DNA ladder. A 1020-bp PCR-amplified product was seen in both clinical isolates. (C) Multiple sequence alignment of the discriminatory regions in the gyrB gene between M. bovis and M. bovis BCG and other members of the MTBC. The black box indicates the single nucleotide variation G to A. (D) Electropherogram of the gyrB PCR product of the clinical isolate. The arrow indicates the single-nucleotide variation G to A representing M. bovis. (E) Schematic representation of RD1 and RD4 PCR amplification band sizes for differentially identifying M. bovis BCG and M. tuberculosis and M. bovis. (F) Agarose gel electrophoresis of RD1 and RD4 PCR products: Lanes 1 to 6 are RD1 PCR products; Lane M: 100-bp DNA ladder, Lane 1: no template control (NTC), Lane 2: M. tuberculosis H37Ra, Lane 3: M. bovis BCG-Russia (Moscow 368), Lane 4: M. bovis BCG-Denmark (Danish 1331), and Lanes 5 and 6: MGIT culture–positive and MPT64-negative samples; Lanes 7 to 12 are RD4 PCR products; Lane 7: NTC, Lane 8: M. tuberculosis H37Ra, Lane 9: M. bovis BCG-Russia (Moscow 368), and Lane 10: M. bovis BCG-Denmark (Danish 1331); and Lanes 11 and 12: MGIT culture–positive and MPT64-negative samples.

Citation: The American Journal of Tropical Medicine and Hygiene 103, 3; 10.4269/ajtmh.19-0853

• Download Figure
• Download figure as PowerPoint slide

Based on these observations, the first patient underwent aspiration of the supraclavicular lymph node and biopsy of the left axillary lymph node, whereas the second patient underwent excision of the soft tissue swelling and treatment with modified first-line antitubercular drugs with replacement of pyrazinamide with streptomycin. At 1-year follow-up, the first child is doing well with complete resolution of lymphadenitis without any relapse. The second child completed 6 months of antitubercular therapy with complete resolution of symptoms.

DISCUSSION

The identification of M. bovis BCG is not straightforward as it belongs to MTBC, a group of highly related organisms comprising M. tuberculosis, Mycobacterium africanum, M. bovis, Mycobacterium microti, and Mycobacterium canettii.¹¹ The viable and actively dividing cells of the members of MTBC secrete a 24-kDa secretary protein which remains stable in the culture medium for at least 1 year.^2,7 The introduction of rapid ICT for the detection of MPT64 protein has made the differentiation of MTBC from NTM very simple and economical.⁸ However, M. bovis BCG strains with deletion or mutation in the mpb64 gene cannot be detected with this assay.⁴ Single-nucleotide polymorphisms that exist in the gyrB gene of the MTBC were used to differentiate M. bovis from the other MTBC species by performing RFLP using the restriction enzyme TaqI.¹² Previous studies have shown the absence of regions of difference (RD) RD1, RD9, and RD10 in all BCG vaccine strains and the presence in TB strains, although M. bovis lacks the RD4 region. Non-BCG M. bovis is distinguished from BCG by the presence of RD1 and the absence of RD9.¹⁰ In this study, we have used the RD1 region– and RD4 region–specific amplifications to differentiate between M. tuberculosis, M. bovis, and M. bovis BCG.

The diagnostic algorithm for pediatric TB in India advocates the upfront use of cartridge-based nucleic acid testing (Xpert MTB/RIF) and cultures, if available.¹³ Both these tests will not differentiate M. tuberculosis from M. bovis and may lead to inappropriate treatment with pyrazinamide, to which M. bovis is inherently resistant. Moreover, the use of MPT64-negative BCG strain will lead to misidentification as NTM in mycobacterial cultures. Pediatric infections due to NTM are very rare and may occur in children with underlying immunodeficiency. A study from India reported that 75% of infants with BCG-related disease have regional or localized involvement, whereas only 25% had disseminated disease.¹⁴ Among patients with disseminated disease, only one patient had IFNγ-R1 gene mutation.¹⁴ The use of two different strains for vaccination in the same population, one with and the other without the mpb64 gene, further complicates the problem of identification. In India, the public sector uses MPT64-negative Danish 1331 strain which is provided free of cost, whereas the private sector uses the MPT64-positive Russian (Moscow) 368 strain and needs to be bought by the patients.¹⁵ The knowledge of various strains used for vaccination plays a crucial role in avoiding misidentification in cases of BCG-associated complications. In our case, although the patient was asplenic, she was immunocompetent, and the BCG vaccine lesion healed normally with a scar. The children were vaccinated with Danish 1331 strain; therefore, the culture isolates were negative for the MPT64 antigen. The pus and tissue samples were positive for MTBC using Xpert MTB/RIF as it targets the rpoB gene for identification. However, Xpert MTB/RIF does not differentiate between the members of MTBC, making it not possible to identify M. bovis BCG. Because M. bovis is inherently resistant to pyrazinamide, accurate identification is necessary for appropriate chemotherapy. Our case emphasizes the importance of careful use of the MPT64 kit with clinical correlation and combining morphological characteristics or nucleic acid amplification tests to identify MPT64-negative M. bovis BCG clinical isolates. Clinicians should also be aware that finding disseminated BCG infection in young children is a hallmark of serious immune deficiency which would need investigation and management.