The determination of colony-forming units (CFU) is the gold-standard technique for quantifying bacteria, including Mycobacterium tuberculosis which can take weeks to form visible colonies. Here we describe a micro-CFU for CFU determination with increased time efficiency, reduced lab space and reagent cost, and scalability to medium and high throughput experiments.
Tuberculosis (TB), the leading cause of death worldwide by an infectious agent, killed 1.6 million people in 2022, only being surpassed by COVID-19 during the 2019-2021 pandemic. The disease is caused by the bacterium Mycobacterium tuberculosis (M.tb). The Mycobacterium bovis strain Bacillus Calmette-Guérin (BCG), the only TB vaccine, is the oldest licensed vaccine in the world, still in use. Currently, there are 12 vaccines in clinical trials and dozens of vaccines under pre-clinical development. The method of choice used to assess the efficacy of TB vaccines in pre-clinical studies is the enumeration of bacterial colonies by the colony-forming units (CFU) assay. This time-consuming assay takes 4 to 6 weeks to conclude, requires substantial laboratory and incubator space, has high reagent costs, and is prone to contamination. Here we describe an optimized method for colony enumeration, the micro-CFU (mCFU), that offers a simple and rapid solution to analyze M.tb vaccine efficacy results. The mCFU assay requires tenfold fewer reagents, reduces the incubation period threefold, taking 1 to 2 weeks to conclude, reduces lab space and reagent cost, and minimizes the health and safety risks associated with working with large numbers of M.tb. Moreover, to evaluate the efficacy of a TB vaccine, samples may be obtained from a variety of sources, including tissues from vaccinated animals infected with Mycobacteria. We also describe an optimized method to produce a unicellular, uniform, and high-quality mycobacterial culture for infection studies. Finally, we propose that these methods should be universally adopted for pre-clinical studies of vaccine efficacy determination, ultimately leading to time reduction in the development of vaccines against TB.
Tuberculosis (TB) is the leading cause of death worldwide by a single infectious agent, bacterium Mycobacterium tuberculosis (M.tb), killing more people than any other pathogen. In 2021, TB was responsible for 1.6 million deaths and was surpassed by COVID-19 during the 2019-2021 pandemic1. Moreover, according to the World Health Organization´s global TB report of 2022, the COVID-19 pandemic was responsible for an increase in new TB cases. The WHO also reports large drops in the number of people diagnosed with TB during this period, which could increase further the number of TB cases1.
The Bacillus Calmette-Guérin (BCG) is a live-attenuated strain of the pathogenic Mycobacterium bovis, used for the first time as a vaccine more than 100 years ago. This is the only vaccine against TB and is the oldest licensed vaccine in the world still in use2,3. Currently, there are 12 vaccines in different phases of clinical trials4, and dozens of vaccines are under pre-clinical development5,6. Pre-clinical assessment of vaccines against TB includes the evaluation of the safety and immunogenicity7, which can be obtained in diverse animal models such as zebrafish, mice, guinea pigs, rabbits, cattle, and non-human primates8,9,10. Additionally, assessing the capacity of a vaccine to induce protection against M.tb infection and/or transmission, i.e., the vaccine efficacy, requires an M.tb challenge in vivo5,11. Interestingly, BCG vaccination induces non-specific effects that affect the survival of other bacterial and viral pathogens12,13 through the mechanism of trained immunity14. To quantify the viable bacterial burden in an infected animal, the method of choice is the enumeration of bacterial colonies through the colony-forming units (CFU) assay5,15. CFU is a unit that estimates the number of microorganisms (bacteria or fungi) that form colonies under specific growth conditions. CFUs originate from viable and replicative microorganisms, and the absolute number of living microorganisms within each colony is difficult to estimate. It is uncertain whether a colony has originated from one or more microorganisms. The CFU unit reflects this uncertainty, hence a great variability can be observed in replicates of the same sample. This time-consuming assay requires specialized technicians trained to work in a biosafety level 3 (BSL3) facility, substantial laboratory and incubator space, takes from 4 to 6 weeks to conclude, and is prone to contamination.
In this study, we describe an optimized method for colony enumeration, the micro-CFU (mCFU), and offer a simple and rapid solution to analyze the results15,16,17,18,19,20. The mCFU assay requires tenfold fewer reagents, reduces the incubation period threefold, taking 1 to 2 weeks to conclude, reduces lab space and reagent cost, and minimizes the health and safety risks associated with working with large numbers of M.tb. We propose that this method should be universally adopted for pre-clinical studies of vaccine efficacy determination, ultimately leading to time reduction in the development of vaccines against TB. Finally, this optimized method of CFU enumeration has been used to quantify not only Mycobacteria but also other bacteria, such as Escherichia coli and Ralstonia solanacearum21.
NOTE: The protocol described here is for BCG but can be applied to any Mycobacteria. BCG can be used as a surrogate bacterium for TB experiments when BSL3 facilities are not available22. The following procedures using BCG should be performed under a biosafety level 2 (BSL2) laboratory and follow the appropriate biosafety guidelines and good laboratory practices for the manipulation of hazard group 2 microorganisms.
1. Culture media preparation
2. Sample preparation
3. Production of BCG culture
NOTE: For in vivo studies of TB vaccines, the aim is to improve the efficacy of BCG. Therefore, BCG-vaccinated groups are usually used as control. BCG strains used for human vaccination are ideal for testing in animal models. In this case, a culture of BCG must be reconstituted according to the supplier's instructions27. However, a BCG culture for in vivo studies can also be produced in-house11. The production of unicellular, uniform, and high-quality BCG culture for in vitro infection protocols has been produced very successfully in several studies11, 16, 18,19,20, 26, 28, 29, using the following protocol, which can also be used for animal challenge studies.
4. Micro-colony forming unit assay
NOTE: After an in vivo or in vitro infection experiment is completed, the enumeration of bacteria can be performed by mCFU. For in vivo studies, samples must be first homogenized in a bead beater or another tissue homogenizer. For in vitro cultures of macrophages/dendritic cells/neutrophils infected with BCG, samples must be lysed using a non-ionic detergent (e.g., 0.05% solution of non-ionic, non-denaturing detergent).
Figure 1. Schematic representation of the mCFU protocol. (A) Serial 10-fold dilutions of the BCG-containing lysates in a 96-well plate. (B) Square Petri dish containing solid culture medium and overlayed by 96 droplets of 5 µL each. Droplets are pipetted directly from the 96-well plate using a multichannel pipette. Created with BioRender.com. Please click here to view a larger version of this figure.
Figure 2. Micro-colony forming units of BCG following 10 days of incubation. On the left, a photo of a square Petri dish overlayed by 96 droplets of 5 µL each, as previously represented in Figure 1B. On the right, individual photos of 3 droplets correspond to an original lysate (100) and two dilutions (101, 102). Photos were taken using a DSRL camera equipped with an 18-55 mm zoom lens (plate) or a 105 mm macro lens (droplets). Please click here to view a larger version of this figure.
5. Micro-colony forming unit counting in Fiji (ImageJ)
NOTE: The mCFU method allows CFU quantification of large sets of samples. Pictures of the droplets may be recorded for posterior analysis to facilitate colony counting. Several photographic devices can produce images with sufficient quality for this purpose. These include digital cameras, webcams, camera-attached microscopes and magnifying glasses, and cell phones. Free image analysis software such as ImageJ offers the possibility of manual or automated colony counting in those images. To demonstrate both methods, Fiji will be used, which is a distribution of ImageJ that packages several tools for scientific image analysis30. Fiji can be downloaded from https://fiji.sc/.
Figure 3. A manual method for counting mCFU using the cell counter plugin on Fiji software. The blue dots indicate colonies already clicked on by the user. The menu on the right displays the number of colonies counted so far (the count is 41). Please click here to view a larger version of this figure.
Figure 4. An automated method for counting mCFU using Fiji software. (A, B) The region of interest with the colonies is selected using the oval selection tool, and the outside area is removed using the clear outside command. (C, D) A black-and-white image of the colonies is generated using the threshold tool. (E, F) The number of colonies is quantified using the analyze particles tool. Please click here to view a larger version of this figure.
The mCFU assay described here increases the amount of information that can be retrieved from a single Petri dish to at least 96-fold. Figure 5 depicts a comparison of two drug-delivery methods for the repurposed use of saquinavir (SQV)31,32 as a host-directed drug to treat tuberculosis. In this assay, four different strains of Mycobacterium tuberculosis were used to infect primary human macrophages. M. tuberculosis H37Rv laboratory strain and three clinical strains isolated from patients with active TB by the Portuguese National Institute of Health's Dr. Ricardo Jorge (INSA): a drug-susceptible strain (INSA code 33427), a multiple drug-resistant (MDR) strain (INSA code 34192), and an extensively drug-resistant (XDR) strain (INSA code 163761). Each infection by each strain was further multiplied into eight different treatment conditions comparing three different concentrations of free-drug or liposome-loaded drug to the respective non-treated controls. Finally, each condition was analyzed at four different time points post-infection. To quantify intracellular bacterial growth, the total amount of lysates extracted from each time point was 32. This was followed by three serial dilutions to each lysate increasing the number to 128 samples. Multiplied by the four-time points analyzed results in 512 samples. Since the experiment was repeated using macrophages from at least three different blood donors, the total number of analyzed samples increased to 1536. Using the mCFU method described here, this experiment accounted for only 16 square Petri dishes versus 1536 that would be necessary using the standard CFU protocol. As shown in Figure 5, the results obtained with this method can demonstrate statistically significant differences between treatments.
Figure 5. The mCFU method can produce high amounts of data from a small number of agar plates. This set of experiments tested the efficacy of the drug saquinavir (SQV) to induce intracellular killing of M.tb laboratory and clinical strains in human macrophages. SQV was administered to the macrophage cultures in its free form or loaded in liposomes (LipSQV). The treatments were performed in three different concentrations: 50, 20, and 10 µg/mL. Macrophages were infected with different M.tb strains for 3 h and then treated with selected concentrations of LipSQV and free SQV. Liposomes without the drug (LipUnloaded) and DMSO were used as controls. To evaluate bacterial intracellular survival, at discrete time points, macrophages were lysed, and serial dilutions of the bacterial suspension were plated on 7H10 agar plates. mCFU units were counted following 2-3 weeks. Lines depict the average mCFU per sample from at least 3 independent experiments. Bars represent the average mCFU percentage calculated relative to the respective controls at day 7 post-infection. Symbols represent each experiment with macrophages from a different donor. Error bars represent the standard error of the mean. Multiple group comparisons were performed using one-way ANOVA followed by a Holm-Sidak post-hoc test. * p ≤0.05, ** p ≤0.01, *** p ≤0.001. This figure has been modified from33. Please click here to view a larger version of this figure.
TB is an important public health problem with increasing importance, particularly in low and middle-income countries. The disruption of healthcare settings to diagnose and treat TB during the COVID-19 pandemic caused a negative impact on the incidence of new cases1. In addition, the multi-drug and extensively-drug resistant M.tb strains, and the co-infection of M.tb and HIV must be urgently addressed to control this epidemic1,34. New vaccines to replace or improve BCG are currently in clinical trials, and new candidates could hit the market in the coming years35. The development of new vaccine candidates against TB relies on pre-clinical studies using diverse animal models. The vaccine efficacy determination in animal models requires bacterial burden quantification in the infected organs to demonstrate the protective effect of the vaccine5,36. Currently, there are direct and indirect methods to quantify the bacterial burden in infected mice and non-human primates. The direct methods include histological preparations and bacterial staining using fluorescent auramine-o and Ziehl-Neelsen37, the quantification of mycobacterial load in samples based on RT-qPCR38, and others.
The mycobacterial growth inhibition assay (MGIA) is an in vitro assay based on the BACTEC MGIT system and is a sensitive method used to quantify live mycobacterial burden in vaccinated and unvaccinated animals. The system is a fully automated instrument that can test for the presence of Mycobacteria in specifically designed test tubes. The vial contains a fluorescent reporter that reacts to the concentration of oxygen in the sample. The fluorescence is proportional to the amount of oxygen consumed, giving an indication of the number of live bacteria present in the sample, every hour39. In the MGIA method, PBMCs and autologous serum samples are collected from animals and co-cultured with mycobacteria for 96 h. Then, adherent monocytes are lysed to release intracellular mycobacteria. The system will quantify the bacterial burden indirectly by measuring oxygen consumption40. This method is typically used for drug resistance studies but has been applied recently to the TB vaccine efficacy evaluation40.
These methods can be complemented with the standard CFU to determine the total number of viable mycobacteria. For example, the evaluation of vaccine candidates in a murine aerosol M.tb challenge experiment is achieved by enumerating colonies obtained from tissues by CFU11.
The mCFU method described here can be a substantial improvement over the classic CFU method, as it is more rapid to perform and inexpensive in terms of reagents used and in terms of laboratory space required. This method allows the simultaneous comparison of a large set of samples in a single experiment, thus decreasing the effects of assay-to-assay variations. For example, as shown in the experiment described in Figure 5, the CFU was determined for 1536 samples using the mCFU method in 16 square Petri dishes, which represents a 96-fold reduction in the number of plates used. This is more relevant for experiments using primary biological and clinical samples, where availability may be an issue. The compact form in which the colonies are displayed on an agar plate conveniently allows the use of specialized or common image-recording equipment. This, in combination with automated colony counting methods such as the one described here using the Fiji software, helps reduce the reliance on the user's ability to identify and quantify the CFU. Even so, like the standard CFU technique, this method suffers from the inability to distinguish merged colonies leading to a potential under-representation of the CFU. However, this can be minimized by selecting a higher dilution to perform the quantification or by counting the colonies under the microscope before their overgrowth results in colony merger. This is more relevant if the sample produces colonies with heterogeneous growth rates. Researchers should validate mCFU implementation for their specific experimental conditions in those cases.
The critical steps within the protocol are the following points: the serial dilutions made should be performed using a 96-well plate, and the dilutions must be thoroughly mixed; the microdroplet plating should be performed with a 0.5-10 µL multichannel pipette to transfer 5 µL from each row of the 96-well plate to the solid medium square plate, without touching and damaging the medium; the micro-colony counting must be performed between 6-10 days of incubation. However, at the time of counting, care must be taken to ensure the colonies are not excessively large or too small.
The modifications and troubleshooting of the technique include the utilization of different bacterial strains. Therefore, for each strain, the optimal growth medium should be used. The limitations of the technique are identical to the limitations of the conventional CFU method, which include the difficulty of enumerating bacteria in low dilutions. Additionally, specific limitations of this method include the necessity of using a microscope to enumerate the colonies.
Finally, this method can be used to quantify the bacterial burden in TB vaccination studies and for test drug resistance/susceptibility studies and host-pathogen interaction studies. For example, Figure 5 shows the application of mCFU for the determination of intracellular killing of M.tb strains, including clinical strains, multi-drug resistant and extensively-drug resistant strains treated with different drug formulations, in human macrophages. Importantly, this method can also be applied to any microorganism, provided that the culturing conditions for each organism are met.
The authors have nothing to disclose.
This work was supported by internal funding from the Faculty of Medicine, Universidade Católica Portuguesa, and external funding from Fundação para a Ciência e a Tecnologia (FCT), under the grants UIDP/04279/2020, UIDB/04279/2020, and EXPL/SAU-INF/0742/2021.
96-well plates | VWR | 734-2781 | |
DSLR 15-55 mm lens | Nikon | AF-P DX NIKKOR 18-55mm f/3.5-5.6G VR | |
DSLR camera | Nikon | D3400 | |
DSLR macro lens | Sigma | MACRO 105mm F2.8 EX DG OS HSM | |
Fetal calf serum | Gibco | 10270106 | |
Fiji Software | https://fiji.sc/ | Fiji is an open-source software supported by several laboratories, institutions, and individuals. All the required plugins are included. | |
Igepal CA-630 | Sigma-Aldrich | 18896 | |
L-glutamine | Gibco | 25030-081 | |
Middlebrook 7H10 | BD | 262710 | |
Middlebrook 7H9 | BD | 271310 | |
Multichannel pipette (0.5 – 10 µl) | Gilson | FA10013 | |
Multichannel pipette (20 – 200 µl) | Gilson | FA10011 | |
Mycobacterium bovis BCG | American Type Culture Collection | ATCC35734 | strain TMC 1011 [BCG Pasteur] |
OADC enrichment | BD | 211886 | |
Phosphate buffered saline (PBS) | NZYTech | MB25201 | |
RPMI 1640 medium | Gibco | 21875091 | |
Sodium pyruvate | Gibco | 11360-070 | |
Spectrophotometer UV-6300PC | VWR | 634-6041 | |
Square Petri dish 120 x 120 mm | Corning | BP124-05 | |
Tyloxapol | Sigma-Aldrich | T8761 | |
Ultrasound bath Elma P 30 H | VWR | 142-0051 |