Tuberculosis remains a major public health concern worldwide, although its prevalence and incidence have gradually decreased over the years. According to the Global Tuberculosis Report of the World Health Organization, over 10.3 million individuals developed tuberculosis in 2022 (1). Disease transmission can be prevented through early molecular diagnosis and treatment using a combination of anti-tuberculous agents.

Patients with tuberculosis are screened using several markers specific to Mycobacterium tuberculosis (MTB) (2). MPT64, a soluble protein encoded by a gene located in the region of difference (RD) 2, is expressed in MTB isolates but not in the live attenuated Mycobacterium bovis (Bacillus Calmette-Guérin) vaccine strain (3). Therefore, the positive MPT64 expression in laboratory cultures and clinical samples provides direct evidence of MTB infection (4, 5). In general, MPT64 is detected via an immunochromatographic test using specific antibodies. Commonly used commercial kits include Capilia TB and SD Bioline TB Ag MPT64. While MPT64 detection in tissue specimens is superior to Gene Xpert assay and mycobacterial cultures for diagnosing extrapulmonary tuberculosis (6, 7), its diagnostic performance is heterogeneous across other specimens.

Nonetheless, inconsistencies have been observed between the MPT64 and nucleic acid-based assays for MTB infections (8, 9). The negative results are largely caused by mpt64 gene mutations, as supported by a few studies (4, 9–12). For negative MPT64 results, further species identification usingmolecular techniques is necessary to confirm whether the strainbelongs to non-tuberculous mycobacteria (NTM). Certain MTB lineages appear to have a higher prevalence of mpt64 mutations; however, mutation types vary across different regions (9, 12, 13). Furthermore, whether clinical features vary across patients with infections by MTB strains with negative MPT64 detection remains unknown.

The MPT64 antigen assay is routinely used to detect MTB in mycobacterial cultures and clinical samples. However, negative results caused by mpt64 mutations can lead to misdiagnosis in the absence of other molecular assays. In practice, a cultured mycobacterial strain with a negative MPT64 assay result is considered an NTM, which requires a different drug regimen, compared with MTB (8, 22). Moreover, patients with NTM infection have not been routinely monitored in most regions because NTM rarely leads to transmission in immunocompetent individuals. Moreover, multiple antibiotic agent-containing therapy is not recommended for patients with mild symptoms (23). Thus, false-negative results in the MPT64 assay possibly resulted in MTB misdiagnosis as NTM infection, causing more susceptible individuals becoming infected by MPT64-negative patients because of a lack of therapy or resistance acquirement from an inappropriate drug regimen (24, 25). For mycobacterial strains with negative MPT64 assay results, additional time and money are warranted to conduct nucleotide acid-based assays using mycobacterial culture to identify the mycobacterial species (8). However, alternative assays (next-generation sequencing, nucleotide acid amplification, and nucleic acid probe detection) for all mycobacterial cultures are not recommended because of higher costs and the need for professional instruments and trained staff.

With the application of several molecular detection techniques in clinical laboratories for mycobacterial identification, negative MPT64 antigen assay results in MTB strains have been reported globally (8, 10–13, 26). Nevertheless, the clinical relevance of negative MPT64 antigen results has not been thoroughly investigated. In this study, a higher percentage of patients who tested negative for the MPT64 antigen received anti-tuberculosis treatment previously. This phenomenon could be partially explained by the insufficient duration of routine treatment for these patients because MPT64-negative MTB strains might have lower division activity than MPT64-positive patients (8, 27). The bacilli load influences the detection of the MPT64 antigen because the cultured MTB strains negative for the MPT64 antigen are more likely to be negative for MTB-specific nucleic acids in clinical samples, such as Gene Xpert assay (8). Moreover, the time to report tended to be longer for strains with negative MPT64 assay results, despite no significance in this study, indicating the possibility of a lower bacilli load in the specimen.

Mpt64 mutations are the major reason for a negative MPT64 result (8, 10, 12, 13). Mpt64 mutation types appear to have a specific association with certain lineages (12, 26). In this study, all strains carrying the 63 bp deletion in the mpt64 gene belonged to the non-Beijing lineage, aligning with previous data suggesting that this deletion is commonly observed in L4 and L1 lineages (10, 13). The truncated protein encoded by the mpt64 gene with a 63 bp deletion may not be recognized by the antibodies in enzyme-linked immunosorbent assay or immunochromatographic commercial kits, yielding false-negative results (10). A 63 bp deletion is predicted to drastically change the protein structure because of an alpha helix, explaining the phenomenon (13).

Single-nucleotide mutations have been reported in the mpt64 gene. Missense mutations can alter protein structure and stability; however, the corresponding strains remained positive for the MPT64 protein assay (9, 11, 13). For example, a common mutation at 477 bp that changes phenylalanine to leucine does not affect MPT64 antigen detection (9, 11). Similar results have been reported for the missense mutation at 128 T > A, which appears specific to L5 lineage strains (9). Furthermore, a longer incubation period (8, 12) or the use of L-J culture instead of liquid culture (8) may reverse the decreased sensitivity of the MPT64 assay because of a single amino acid change. For MTB strains with negative MPT64 assay results from Australia, the mpt64 mutation type is the deletion of two nucleotides, resulting in a frameshift in protein expression (26). Further analysis indicated that all the strains carrying this mutation belong to Lineage 4, which has been attributed to European migration and colonization (28). IS6110 insertion is an uncommon mpt64 mutation, and previously reported insertion sites (501 bp) (29) is different from our study (313 bp). The influence of IS6110 on protein structure or function depends on the insertion site (30). In detail, insertion at 313 bp in this study introduced a termination codon, resulting in a shorter polypeptide of 135 amino acids.

Mpt64 mutations are specific to certain MTB lineages or sublineages; nonetheless, whether these strains possess unique features concerning resistance phenotype and virulence remains elusive (28, 31). In the future, multi-omic investigations might elucidate the mechanism by which these strains interact with geographic populations (32–34). Moreover, the lower sensitivity of the MPT64 antigen-based assay could be overcome by combining molecular assays for different targets or adopting more convenient and economical methods for MTB diagnosis (35, 36). Furthermore, novel antibodies detecting both the mutated and wild MPT64 proteins would improve diagnostic efficiency in liquid cultures and minimize false negatives (37).

This study has several limitations. MTB strains were collected from a single center, which may not entirely represent broader genetic diversity. Strains carrying the wild-type mpt64 gene might have yielded positive results after longer incubation in liquid culture; however, these strains were not preserved and only the nucleic acids were available. Furthermore, clinical information for some patients was missing because of retrospective data collection, particularly for outpatients, possibly introducing bias in the comparison of clinical characteristics. Finally, detailed lineage data was inaccessible for some strains because of the lower DNA quality. In the future, whole genome sequencing of mutated MTB strains would provide a more comprehensive mutational landscape.

Patients with MTB infections who are negative for the MPT64 antigen are more likely to be misdiagnosed as having NTM because of lower positive rates in the Gene Xpert assay. Mpt64 mutations, primarily the 63 bp deletion and IS6110 insertion, were associated with negative MPT64 antigen assay results. In contrast, single nucleotide mutations in the mpt64 gene partly contributed to the lower sensitivity of antigen assays, which may be improved by prolonged incubation. Other nucleic acid-based tests for MTB should be conducted for patients who tested negative for the MPT64 antigen but positive for mycobacterial culture.