Autofluorescence of Mycobacteria as a Tool for Detection of Mycobacterium tuberculosis

Abstract

The diagnosis of tuberculosis in developing countries still relies on direct sputum examination by light microscopy, a method that is easy to perform and that is widely applied. However, because of its poor sensitivity and requirement for significant labor and training, light microscopy examination detects the bacilli in only 45 to 60% of all people whose specimens are culture positive for Mycobacterium tuberculosis. Therefore, new diagnostic methods that would enable the detection of the undiagnosed infected population and allow the early commencement of antituberculosis treatment are needed. In this work, the potential use of mycobacterial cyan autofluorescence for the detection of Mycobacterium tuberculosis was explored. The tubercle bacilli were easily visualized as brilliant fluorescent bacilli by microscopy and were easily tracked ex vivo during macrophage infection. Assays with seeded sputum and a 96-well microplate reader fluorimeter indicated that <10⁶ bacilli ml⁻¹ of sputum could be detected. Moreover, the use of microplates allowed the examination of only 200 μl of sputum per sample without a loss of sensitivity. Treatment with heat or decontaminating chemical agents did not interfere with the autofluorescence assay; on the contrary, they improved the level of bacterial detection. Autofluorescence for the detection of bacilli is rapid and easy to perform compared to other methodologies and can be performed with minimal training, making this method suitable for implementation in developing countries.

The diagnosis of tuberculosis (TB) in low- and middle-income countries, where more than 90% of TB cases occur, is mainly performed by microscopy examination of stained sputum smears for acid-fast bacilli (25). The International Standards for Tuberculosis Care considers the microscopic analysis of two or three specimens per patient to be a cornerstone for the diagnosis of TB (13). Although light microscopy is inexpensive, easy to perform, and highly specific in areas where there is a high prevalence of TB, it is relatively insensitive, requiring ≥10⁴ bacilli ml⁻¹ of sputum to achieve a positive result (15, 35). In addition, the procedure requires the observation of from 100 to 300 fields in order to obtain accurate results (31).

The sensitivity of microscopy is influenced by numerous factors, such as the prevalence and severity of disease, the quality of specimen collection, the type of specimen, the method of processing (direct or concentrated, centrifugation, liquefaction), staining, and finally, the quality of the examination (20, 23, 26, 29). In well-trained hands, the test detects tubercle bacilli in 75% of all people who have active pulmonary TB, but the sensitivity may fall to 45 to 60%, depending on the training, eye, and motivation of the laboratory technician (1). The basis of conventional light microscopy was developed more than 100 years ago, and, at present, conventional light microscopy uses carbolfuchsin Ziehl-Neelsen or Kinyoun acid-fast stains. Fluorescence microscopy with fluorochrome dyes such as auramine O or auramine-rhodamine is reputed to have higher degrees of sensitivity and specificity and is thus a more accurate test for the diagnosis of TB (16, 32). However, it is mainly performed in developed countries because it is more expensive than the conventional method, as it requires unstable and expensive fluorescent staining reagents, which increases the cost of the assay, thus hindering its widespread use in developing countries (32). Actions are required to diminish that disadvantage and to allow the wider applicability of fluorescence microscopy in developing countries.

Intrinsic fluorescence (autofluorescence) is a characteristic previously described in a few microorganisms that allows their detection without the need for fluorescence staining and has been used for the detection of bacterial cells (6). Here we report for the first time that the mycobacteria, including Mycobacterium tuberculosis, emit autofluorescence in the cyan range of the visible spectrum. Autofluorescence was first explored with cultured microorganisms; then with experimentally infected macrophages, which were used to mock a natural intracellular infection; and finally, with sputum specimens spiked with M. tuberculosis and submitted to the decontamination processes routinely used in diagnostic laboratories.

MATERIALS AND METHODS

Strains and culture conditions.

The microbial strains used in this study are listed in Table 1. Species belonging to the Mycobacterium genus, some partially acid-fast actinomycetes, as well as some pathogens frequently detected in the sputum of patients with respiratory infections were selected. The mycobacteria and the nocardia species were grown in Middlebrook 7H9 medium (Difco) supplemented with 0.5% (vol/vol) glycerol, 10% (vol/vol) albumin-dextrose-catalase (Difco), and 0.05% (vol/vol) Tween 80 (Sigma, St. Louis, MO) with gentle agitation at 30°C or 37°C, according to the optimal growth temperature for the organism. Corynebacterium glutamicum was grown in Luria broth medium (Difco) with shaking at 37°C. Haemophilus influenzae and Streptococcus pneumoniae were grown in brain heart infusion broth (Himedia, India) at 37°C without agitation. Cultures of Paracoccidioides brasiliensis (yeast phase) were kindly provided by G. Niño-Vega (Mycology Laboratory, Instituto Venezolano de Investigaciones Científicas, Venezuela). For microscopic observation, the microorganisms were harvested and washed twice with phosphate-buffered saline (PBS) solution at pH 7.4. Aliquots of 20 to 40 μl were applied to slides that had been pretreated with polylysine, as described previously (3).

TABLE 1.

Microorganisms used in this study

Microorganism	Human pathogen	Pigmentation	Strain	Classification (phylum, order)
Mycobacterium smegmatis mc2155	No	No	ATCC 700084	Actinobacteria, Actinomycetales
Mycobacterium fortuitum	Opportunistic	No	ATCC 6841	Actinobacteria, Actinomycetales
Mycobacterium avium	Opportunistic	No	ATCC 25291T	Actinobacteria, Actinomycetales
Mycobacterium gordonae	Opportunistic	Scotochromogen	Clinical isolatea	Actinobacteria, Actinomycetales
Mycobacterium kansasii	Opportunistic	Photochromogen	Clinical isolatea	Actinobacteria, Actinomycetales
Mycobacterium marinum	Yes	Photochromogen	ATCC 927	Actinobacteria, Actinomycetales
Mycobacterium tuberculosis H37Rv	Yes	No	ATCC 27294	Actinobacteria, Actinomycetales
Mycobacterium bovis BCG Pasteur	No	No	ATCC 35734	Actinobacteria, Actinomycetales
Nocardia sp.	Yes	Yes	Clinical isolatea	Actinobacteria, Actinomycetales
Corynebacterium glutamicum	No	No	ATCC 13032b	Actinobacteria, Actinomycetales
Streptococcus pneumoniae	Yes	No	Clinical isolateb	Firmicutes, “Lactobacillales”
Haemophilus influenzae	Yes	No	Clinical isolateb	Proteobacteria, Pasteurellales
Paracoccidioides brasiliensis C81	Yes	No	ATCC 32071c	Ascomycota, Onygenales

^aFrom J. García-Rodríguez, Hospital Universitario La Paz (Madrid, Spain).

^bFrom the Laboratorio Nacional Centro Venezolano de Colecciones de Microorganismos (Instituto de Biología Experimental, Venezuela).

^cFrom G. Niño-Vega, Mycology Laboratory (Instituto Venezolano de Investigaciones Científicas).

For macrophage infection, wild-type M. tuberculosis H37Rv and H37Rv transformants bearing pGFP22-4, a plasmid that expresses the green fluorescent protein (GFP) (30), were grown in a roller agitator under the conditions described above.

Pathogenic strains were grown and manipulated inside a biosafety level 3 containment facility.

Macrophage infection.

To test whether the infection of macrophages can be tracked by using cyan autofluorescence, THP-1 cells were infected with M. tuberculosis H37Rv. THP-1 cells were grown in RPMI 1640 (Gibco, BRL) supplemented with 10% fetal calf serum, 25 mM HEPES, 2 mM glutamine, penicillin (10,000 μg ml⁻¹), and streptomycin (10,000 μg ml⁻¹) at 37°C in 5% CO₂. The cells were subcultured every 3 days and passaged at least one time in fetal calf serum-supplemented antibiotic-free RPMI 1640 growth medium. Before infection, 2.5 × 10⁵ macrophages were differentiated in eight-well Nunc Lab-Tek chamber slides by the addition of 100 μM phorbol-12-myristate-13-acetate (Sigma) and 100 U ml⁻¹ of recombinant human gamma interferon (specific activity, 2 × 10⁷ U mg⁻¹; Abcam) for 24 h at 37°C in a 5% CO₂ atmosphere.

M. tuberculosis inocula were prepared by resuspending the bacteria in warm RPMI 1640 medium at a concentration of 2 × 10⁶ bacilli ml⁻¹, and the inocula were used to infect activated THP-1 cells at a ratio of 10:1 (mycobacteria to macrophages). At 3 h postinfection, the cells were washed extensively with PBS (pH 7.4) to remove the extracellular bacilli and fixed with 4% paraformaldehyde for 30 min at 37°C. Finally, infected macrophages from two wells were stained with 25% (wt/vol) fluorescein diacetate (FDA; Sigma) for microscopic observation.

Fluorescence microscopy and image analysis.

The microorganisms were photographed with a cooled Orca 12-ERG digital charge-coupled-device camera (Hamamatsu Photonics) mounted on a Nikon E600 fluorescence microscope through a ×100 (numerical aperture [NA], 1.3) oil immersion objective. Images were taken, and the images were processed with MetaMorph software (Universal Imaging Corporation). The cells were first photographed in the difference interference contrast mode and were then photographed with a cyan GFP filter (excitation, 426 to 446 nm; emission, 460 to 500 nm) to capture mycobacterial autofluorescence.

M. tuberculosis-infected macrophages grown in chamber slides were visualized with the same equipment through a ×40 (NA, 0.55) objective and photographed as described above for the detection of mycobacterial autofluorescence. A B2-A filter (excitation, 450 to 490 nm; emission, 515 nm) was used to document FDA staining and the GFP transformants.

Confocal laser scanning microscopy.

Bacteria were grown until they reached the exponential phase of growth, and aliquots were twice washed with PBS. M. tuberculosis was heat inactivated at 80°C for 20 min and washed twice in PBS. Aliquots of 100 μl were resuspended in PBS and filtered through a 0.2-μm-pore-size GTTP filter (Millipore, Bedford, MA). The filters were mounted on a slide by using mounting medium (Vectashield; Atom). Cell images were obtained with a Leica DM IRE2 inverted microscope equipped with a Leica TCS SP2 time confocal scanner. The samples were excited with an argon ion laser by using the 405-nm and 488-nm lines. A 430- to 630-nm band-pass filter was used for the scanning emissions. A HCX PL APO lambda blue ×63 (NA, 1.2) water objective was used for all scans.

Effects of heat treatment and pH on M. tuberculosis autofluorescence.

In order to investigate the effect of high temperature on autofluorescence, samples of M. tuberculosis cells from a liquid culture were harvested at the exponential phase and heat treated (20 min, 80°C). To determine the effect of pH, exponential-phase M. tuberculosis cultures were harvested by centrifugation and the cells were resuspended in the same volume of phosphate or borate solutions buffered to pHs ranging from 6 to 10. The fluorescence of aliquots of 150 μl was measured by fluorimetry.

Processing of sputum samples.

Pooled sputum samples that were smear negative for acid-fast bacilli and that were obtained from the Biomedicine Institute (Caracas, Venezuela) were irradiated with UV for 15 min to destroy microorganisms and were spiked with known concentrations of M. tuberculosis, as determined by measurement of the optical density at 640 nm (OD₆₄₀) and the numbers of CFU per ml. The number of CFU was calculated by plating serial dilutions of the M. tuberculosis suspensions on 7H10 medium (Himedia). Aliquots of 1 ml were dispensed into 2-ml screw-cap tubes. An equal volume of a liquefying reagent, such as NaOH, sodium N-acetyl cysteine (NALC-NaOH), sodium hypochlorite (NaOCl), dimethyl benzyl lauryl ammonium bromide (Gerdex), sodium dodecyl sulfate-sodium hydroxide (SDS-NaOH), or dithiothreitol (DTT), was added; and the seeded sputa were incubated at 37°C for 15 min with gentle agitation. The samples were left on the bench for 15 min at room temperature in order to allow sedimentation of the sputum (33). Duplicates were left on the bench overnight at room temperature, when indicated. The supernatant was carefully poured off, and aliquots (200 μl) of the sputum sediment were used for fluorescence measurement. The control samples for each experiment included M. tuberculosis-free sputum containing bacterial culture medium and the liquefying agents.

Fluorimetry.

The fluorescence of M. tuberculosis in culture suspensions and spiked sputa was directly measured in a 96-well microplate reader (Genius Tecan) by using a filter combination for cyan (excitation, 430 ± 20 nm; emission, 480 ± 20 nm). The emitted cyan fluorescence was expressed as the number of relative fluorescence units (RFUs), which was corrected by subtracting the fluorescence of the solution used to resuspend the bacilli in each experiment. The fluorescence emission was determined from at least three different preparations of each sample. RFUs of >15 to 20% of the background were considered to be indicative of cyan autofluorescence.

RESULTS

Autofluorescence of Mycobacterium by fluorescence microscopy.

Fluorescent Mycobacterium cells were easily visualized by simple direct observation of a suspension of bacilli under a fluorescence microscope by using the excitation and barrier filter selected. The Nocardia sp. also emitted cyan autofluorescence (Fig. 1). The remaining microorganisms, including C. glutamicum, an organism closely related to mycobacteria, did not show cyan autofluorescence.

FIG. 1.

Fluorescence microscopy. The exposure time was 200 μs. The cells were photographed by using a cyan GFP filter cube (excitation, 426 to 446 nm; emission, 460 to 500 nm). Bar, 3 μm.

M. tuberculosis cells infecting macrophages were distinguished by green fluorescence when the bacteria expressed GFP, an easily detected fluorescent molecule, or after they were stained with FDA. However, the tubercle bacilli were also easily visualized inside infected macrophages as bacilli with cyan fluorescence (Fig. 2).

FIG. 2.

Visualization of M. tuberculosis inside macrophages. THP-1 cells were infected with M. tuberculosis H37Rv (stained with FDA) or M. tuberculosis bearing plasmid pGFP22-4 and visualized by fluorescence microscopy with a B2-A filter (green fluorescence) or a cyan GFP filter (cyan fluorescence). With both filters, M. tuberculosis was observed as green as well as cyan bacilli inside the infected macrophages.

Fluorescence emission spectra.

In order to gain insight into the nature of the intrinsic mycobacterial autofluorescence, the emission spectra of four Mycobacterium species were determined by confocal laser scanner microscopy. The results are shown in Fig. 3. The mycobacteria exhibited an emission maximum at 475 nm when they were excited with a laser at a wavelength of 405 nm. No fluorescence emission was detected when the mycobacteria were excited with a laser at a wavelength of 488 nm.

FIG. 3.

Emission spectra of Mycobacterium species recorded at an excitation wavelength of 405 nm.

Influences of temperature and pH on autofluorescence of M. tuberculosis.

In order to determine the influences that standard diagnostic procedures could have on M. tuberculosis autofluorescence, the fluorescence emissions from heat-treated cultures as well as from M. tuberculosis cells resuspended in buffer solutions with different pH values were determined. The level of M. tuberculosis cyan fluorescence increased fourfold after incubation for 20 min at 80°C (Fig. 4A). The autofluorescence of M. tuberculosis did not vary over a pH range of from 6 to 9 in phosphate and borate buffers, whereas in borate buffer adjusted to pH 10, the cyan autofluorescence of the tubercle bacilli increased eightfold (Fig. 4B).

FIG. 4.

Effect of heat treatment (A) and pH (B) on the cyan autofluorescence of M. tuberculosis. The fluorescence emissions from cultures at room temperature (RT) or that had previously been heat inactivated for 20 min (80°C) were calculated once the fluorescence of the broth medium (A) or the buffer used to resuspend the bacteria (B) was subtracted. PPB, potassium phosphate buffer; SBB, sodium borate buffer; 6 to 10, pHs 6 to 10. The bars and the error bars show the means and the standard deviations, respectively, of at least three separate determinations.

The minimum amount of bacilli required for autofluorescence detection by fluorimetry was determined by using exponential-phase M. tuberculosis cultures and a 96-well microplate format. For untreated cells, 3 × 10⁶ bacilli per well (corresponding to 2 × 10⁷ CFU ml⁻¹) were required to clearly detect cyan autofluorescence. The limit of detection dropped to 4 × 10⁴ bacilli per well (2.7 × 10⁵ CFU ml⁻¹) when the same cultures were heat inactivated (at 80°C for 20 min). Therefore, the heat treatment substantially increased the emission of fluorescence, making possible the detection of approximately 75-fold fewer bacilli.

Influence of standard methods of sputum decontamination on M. tuberculosis autofluorescence.

The cyan autofluorescence of the seeded sputa was detected with all clarifying agents except NaOCl. This molecule showed a high level of intrinsic fluorescence that interfered with that of M. tuberculosis, which, as a consequence, was undetectable. When chemical reagents other than NaOCl were considered, the levels of the autofluorescence at room temperature were similar regardless of the liquefying agent used. However, drastic differences were found when seeded sputa were incubated at 80°C for 20 min (Fig. 5). The autofluorescence increased about 10 times when the seeded sputa were heat inactivated and treated with NaOH or NALC-NaOH. In the presence of DTT, the autofluorescence showed a moderate increase when the bacteria were heated. On the contrary, heat inactivation in the presence of other reagents, such as dimethyl benzyl lauryl ammonium bromide or SDS-NaOH, hindered the detection of the autofluorescence, probably due to the lysis of the bacilli under these conditions (Fig. 5).

FIG. 5.

Effect of decontamination on the detection of the M. tuberculosis cyan autofluorescence from sputum. Fluorescence emission was measured from samples without heat treatment (gray bars) and heat inactivated at 80°C for 20 min (white bars). The fluorescence emission from each seeded sputum specimen was calculated once the fluorescence of sterilized sputum in 7H9 broth without M. tuberculosis and the liquefying agents was subtracted. RT, room temperature.

Detection of M. tuberculosis in sputum by using autofluorescence.

In order to examine the potential use of the mycobacterial autofluorescence for the diagnosis of TB, we determined the detectable range of tubercle bacilli from sputum specimens by the use of autofluorescence and a 96-well microplate format. Known concentrations of M. tuberculosis H37Rv cells that had previously been heat inactivated were seeded in sterilized sputum, and an equal volume of NaOH or NALC-NaOH was added to mimic the usual manipulation of the samples that occurs during microbiological diagnostic procedures.

The autofluorescence of M. tuberculosis was detected with all the concentrations of bacilli assayed, but overnight sedimentation rendered the best results. A better correlation between the fluorescence emissions and the bacillus concentration was observed when NALC-NaOH was used. Overnight sedimentation improved the rate of detection of the autofluorescence of seeded sputum liquefied with NALC-NaOH 1.4 times (Fig. 6).

FIG. 6.

Range of detection of the M. tuberculosis cyan autofluorescence in sputum. Liquefying agents (NaOH or NALC-NaOH) were added to seeded sputa containing known concentrations of M. tuberculosis cells. The fluorescence emission was measured from sedimented samples after 15 min and overnight (o/n) sedimentation and was calculated by subtracting the fluorescence of sterilized sputum in 7H9 broth without M. tuberculosis and the liquefying agents. The bars and the error bars show the means and standard deviations, respectively, of at least three separate determinations.

DISCUSSION

The diagnosis of TB is primarily based on clinical features, histopathology, the identification of acid-fast bacilli, and the isolation of M. tuberculosis from the clinical specimens. Cultivation of the tubercle bacilli is the most sensitive method for the detection of M. tuberculosis and can detect 100 bacilli per ml of sputum (18); unfortunately, it can take several weeks to yield results. However, the routine implementation of the identification of M. tuberculosis by culture is very limited, especially in low-income countries.

WHO strongly recommends the use of the sputum smear examination as the primary tool for case detection and the diagnosis of TB as a pivotal element of the directly observed therapy strategy. Sputum smear examination is the preferred screening test for the detection of M. tuberculosis in developing countries, since this technique is simple and inexpensive and allows the detection of infectious cases. However, WHO (http://www.who.int/tb/features_archive/global_plan_to_stop_tb/) has also recognized the significant limitations of standard direct sputum smear microscopy. The requirement of time-consuming slide examination by light microscopy of two or three sputum samples and the training needed for the application of the technique imply a heavy workload on already overstretched health care systems in the developing world. The use of fluorescence microscopy decreases the run time for sample examination and shortens the time required to obtain results. Unfortunately, the cost derived from the requirement for unstable and expensive fluorescent reagents, such as auramine-rhodamine, is hindering the use of fluorescence microscopy in developing countries (32).

In this work we explored the ability of M. tuberculosis to autofluoresce in order to improve the detection of the tubercle bacilli in clinical samples. Very few studies on the use of autofluorescence for the detection of bacteria have been reported (2, 14). Some bacteria exhibit spontaneous emission spectra following excitation at specific wavelengths. The energy of the photons emitted depends on the intrinsic fluorophores retained, such as protein tryptophans, nucleic acids, and coenzymes (19), or on the production of other fluorescent substances (4, 21). This work shows for the first time that M. tuberculosis can easily be detected by its intrinsic autofluorescence in the cyan emission spectrum as a brilliant fluorescent bacillus in vitro as well ex vivo during macrophage infection (Fig. 1 and 2). In fact, the cyan fluorescence of M. tuberculosis cells was as brilliant as that of M. tuberculosis cells expressing cytoplasmic GFP or FDA-stained bacilli (Fig. 2).

Confocal laser scanning microscopy demonstrated that the mycobacteria exhibited a unique emission maximum at 475 nm when they were excited at 405 nm (Fig. 3), suggesting that the autofluorescence of mycobacteria is possibly due to the fluorescence emitted by the coenzyme F₄₂₀ or its biosynthetic intermediate, FO (5). F₄₂₀/FO shows a bright blue-green fluorescence with a maximum of about 470 nm when cells are excited with a laser at a wavelength of 405 nm (11, 12). F₄₂₀ is a two-electron transfer coenzyme which has 7,8-didemethyl-8-hydroxy 5-deazaflavin, a phospholactyl moiety, and various numbers (two to six) of glutamate residues. FO contains a deazaflavin ring and ribityl sugar but lacks the phospholactyl group and glutamate moieties (5). Coenzyme F₄₂₀ is present in some members of three distantly related groups of prokaryotes: archaea, aerobic actinomycetes, and cyanobacteria (28). The presence of F₄₂₀/FO has previously been reported in Mycobacterium avium, Mycobacterium smegmatis, Mycobacterium fortuitum, Mycobacterium bovis BCG, Mycobacterium phlei, and M. tuberculosis (5, 9, 17, 24).

The autofluorescence of the M. tuberculosis cultures was appreciably increased by high temperature and a basic pH (Fig. 4), perhaps due to spectral changes in F₄₂₀ under these conditions (27).

The role of F₄₂₀ in Mycobacterium species is poorly understood. M. tuberculosis encodes four glucose-6-phosphate dehydrogenases, two of which (fgd-1 and fgd-2) require F₄₂₀ as a cofactor (8). The loss of F₄₂₀ biosynthesis or the inability to reduce this cofactor by Fgd-1 results in resistance to nitroimidazo[2,1-b] oxazine (PA-824), a drug that shows activity against the tubercle bacilli under both aerobic and anaerobic conditions (22, 34). In addition, it has been suggested that reduced F₄₂₀ has a role in the low-redox-potential electron transfer reactions that would be associated with the survival of M. tuberculosis under anaerobiosis (7).

When M. tuberculosis cells were observed under the fluorescence microscope, the autofluorescence decayed quickly, but micrographs could be made (Fig. 1 and 2). The decay of the autofluorescence may result from the photoreduction of F₄₂₀, as was observed in methanogenic bacteria (10, 11, 12). However, fading of the autofluorescence was not detected when the emitted light was detected by spectroscopy, probably due to the lower-power lamp used in the fluorimeter.

Although the autofluorescence of mycobacteria is shared by Nocardia spp., it was not observed in other frequent respiratory pathogens, such as S. pneumoniae and H. influenzae. Bacteria in the genus Nocardia are closely related to mycobacteria; they also synthesize the coenzyme F₄₂₀ (9), and like mycobacteria, they have a cell envelope with high lipid content, which also allows the cell to be stained by carbolfuchsin Ziehl-Neelsen. Thus, the application of autofluorescence to the identification of M. tuberculosis is likely to present the same limitations in specificity as microscopy of acid-fast bacilli.

Some chemical reagents, such as NaOH, NALC-NaOH, and NaOCl, are currently used to liquefy and decontaminate the sputum, which is further concentrated by sedimentation or centrifugation to increase the sensitivity of detection of the bacilli (33). Since sputum samples are the most common clinical specimens analyzed for the diagnosis of TB and because clarification is strictly required to handle sputa, we sought to examine the effects of these and other reagents on the mycobacterial autofluorescence in seeded sputa. We found that the fluorescence emission of the seeded sputa liquefied with NaOH or NALC-NaOH and treated with heat was considerably higher that that when any other decontaminating agent was used. We speculate that the basic pH of these liquefaction agents avoids the reduction of F₄₂₀/FO and, together with heat treatment, increases the natural autofluorescence of M. tuberculosis in the seeded sputum samples.

Our preliminary results showed that under the assay conditions used, at least 7.8 × 10⁵ tubercle bacilli per ml of sputum could be detected by autofluorescence (Fig. 6). Hobby et al. (15) related the bacillus microscopy results to the mycobacterial concentrations in the sputum of patients with TB. They established that sputum samples with microscopy scores of 3+ have from 1.08 × 10⁶ to 4.02 × 10⁷ bacilli ml⁻¹, while sputum samples with scores of 2+ have from 2.4 × 10⁴ to 1.41 × 10⁶ bacilli ml⁻¹. Hence, the results of this work suggest that the autofluorescence of M. tuberculosis could have a sensitivity similar to that of light microscopy examination, but with the advantage that it avoids the staining step in the procedure.

Fluorimetry in a 96-well plate format might become a user-friendly tool for the screening of sputum samples for M. tuberculosis, thus enabling the simultaneous examination of multiple samples. It has the added advantages that it employs a lower sample volume, is easy to perform by any laboratory technician, and is fast. In addition, the results are easily interpretable with simple training. All these characteristics make the use of cyan autofluorescence a potential method for the detection of mycobacteria, including M. tuberculosis.

In order to determine the applicability of the autofluorescence method in the field, the analysis of sputum samples from TB patients from diverse regions of Venezuela is in progress. The application of autofluorescence will improve the detection of tubercle bacilli.