Rapid Microbial Testing

Testing for the presence of microbes in an aseptic environment used to be a process that took weeks. Now, methods for rapid microbial testing can be used to detect the presence of mircoorganisms in the environment or in a finished drug product in hours or days. In this technical forum, experts describe different methods of rapid microbial testing and their applications. Contributors to the forum are: Anne Baumstummler, senior scientist; Renaud Chollet, head of biotechnologies and reagents; Hervé Meder, R&D project manager; Céline Rofel, cell culture and molecular biology technician; Adrien Venchiarutti, engineer; and Sébastien Ribault, PhD, director development & bioproduction, all at Millipore SAS; and Ruth Eden, PhD, president of BioLumix.

Microcolonies Fluorescent Staining Method for Rapid Detection of Microbial Contamination in Mammalian Cell Culture

Anne Baumstummler, Renaud Chollet, Hervé Meder, Céline Rofel, Adrien Venchiarutti, and Sébastien Ribault, Millipore SAS

Traditional testing methods for microbial contamination of mammalian cell culture require several days to detect contamination. The presence of microorganisms is typically assessed by membrane filtration and incubation on solid media or inoculation into liquid media. These culture-based methods rely on microorganisms growing and yielding visible colonies or turbidity. Time-to-result is measured in days for standard bioburden testing (1). The microcolonies fluorescent staining method (MFSM) described in this study is a robust, simple, and rapid method enabling the detection of microbial contaminants. It can be applied to cell culture samples with time-to-results that are two to five times shorter than traditional microbiology. The nondestructive feature makes it compatible with downstream identification of contaminants.

The need for speed. Many technologies and detection systems such as enzyme-linked immunosorbent assays, impedimetry, bioluminescence, flow cytometry, and polymerase chain reaction arose in the field of rapid detection during the two past decades (2–4). Fluorescent dyes have been routinely used in methods such as the direct epifluorescent filter technique (DEFT) to monitor microbial contamination (5). This technique involves the capture of microorganisms on the surface of polycarbonate membrane filters, staining, and visualization using epifluorescence microscopy. The sensitivity is directly linked to the volume filtered and the number of fields observed. A limitation of this approach is the inability to differentiate fluorescent microorganisms from autofluorescent particles. Automated and semi-automated systems have been developed to increase this discrimination and improve accuracy (6–8).

An alternative is to combine the principles of epifluorescence microscopy and flow cytometry (9). Microorganisms retained on a polycarbonate filter are fluorescently labeled and automatically counted by a laser-scanning device that makes the differentiation between fluorescent microorganisms and particles (10). Several authors have reported the use of this solid-phase cytometry technology for bioburden testing of mammalian cell culture processes (1). Results are typically obtained within five hours of sample preparation (6). One drawback of such direct detection methods is the inability to discriminate between culturable and nonculturable organisms (11). In addition, many fluorescent-labeling techniques have an impact on the viability of colony formation, thereby making it impossible to identify organisms that are detected. This is a challenge when conducting investigations or evaluating the severity of a contamination.

MFSM is a method that combines the widely used membrane filtration method with a fluorescence-based staining for the rapid detection of contaminants in filterable samples using microcolony formation. MFSM is based on a nondestructive fluorescent labeling of viable and culturable microorganisms present on cellulose membranes (MilliFlex Quantum, EMD Millipore; see Figure 1). The procedure consists of filtration of the sample, a short incubation on media to yield microcolonies, staining of microorganisms, and counting of fluorescent microcolonies with a light-emitting diode (LED) system. The microorganisms are labeled directly on the filter with a nonfluorescent substrate that is cleaved by intracellular microbial enzymes. Only metabolically active microorganisms with membrane integrity that retain the fluorescent product are stained. Because the staining solution is nondestructive, the method is compatible with standard identification methods.

Figure 1

The objective of this study was to compare MFSM with traditional epifluorescence microscopy for its ability to rapidly detect microbial contaminants in mammalian cell culture.

Results. Both methods were used to detect bacteria in Chinese hamster ovary (CHO) cell cultures. Epifluorescence microscopy was limited by filterability, media interference, and nonrobustness issues, whereas microcolonies fluorescent staining method enabled consistent detection of Bacillus cereus, Staphylococcus epidermidis, and Propionibacterium acnes after eight, nine, and 48 hours of incubation.

With epifluorescence microscopy, nonspiked treated CHO cells yielded a great deal of fluorescent debris and a very high background (see Figure 2A), which interferes with the detection of B. cereus (see Figure 2B). Because observation was difficult in the presence of cells, B. cereus was spiked in CHO medium. The fluorescent background was removed, but it was still difficult to observe bacteria on the polycarbonate filter. The detection of the microorganisms was only possible when increasing the number of contaminants retained on the filter by tenfold (see Figure 2C). Microorganisms spiked in the mammalian cell culture were easily detected with the MFSM without any background or media interference (see Figure 2, D–F). Background noise was minimized as a result of lysing the CHO cells before the filtration using the mammalian cell lysis solution; this buffer eliminates CHO cells while minimizing the impact on microorganisms. Similar results were obtained for both methods with S. epidermidis and P. acnes (data not shown).

FIgure 2

As a result of the nondestructive feature of the MFSM, stained membranes could be reincubated on culture media to yield visible colonies that can be collected for further identification using existing identification methodologies (i.e., biochemical, morphological, nucleic-acid analytics, etc.). The ability to identify the microorganism can accelerate the root cause analysis and implementation of the corrective/preventative action (CAPA) plan.

References

1. A. Onadipe and K. Ulvedal, PDAJ. Pharm. Sci. Technol. 55, 337–345 (2001).

2. R.T. Noble and S.B. Weisberg, J. Water Health 3, 381–392 (2005).

3. S. Flint et al., J. Appl. Microbiol. 102, 909–915 (2007).

4. R. Chollet et al., J. Rapid Methods Autom. Microbiol. 16, 256–272 (2008).

5. U. M. Rodrigues and R.G. Kroll, J. Appl. Bacteriol. 59, 493–499 (1985).

6. J. Moldenhauer, (2008) "Overview of rapid microbiological methods" in Principles of Bacterial Detection: Biosensors, Recognition Receptors and Microsystems M. Zourob, S. Elwary, and A. Turner, Eds. (Springer New York, 2008) pp. 49–79.

7. Y. Motoyama et al., Transfusion 48, 2364–2369 (2008).

8. Y. Motoyama et al., J. Health Sci. 55, 726–731 (2009).

9. J.T. Lisle et al., Appl. Environ. Microbiol. 70, 5343–5348 (2004).

10. K. Mignon-Godefroy, J.G. Guillet, and C. Butor, Cytometry 27, 336–344 (1997).

11. S. Asano et al., J. Biosci. Bioeng. 108, 124–129 (2009).

Growth-Based System for Rapid Microbial Testing

Ruth Eden, president, BioLumix

Currently available growth-based rapid microbiological methods provide either a quantitative cell count, estimation of viable cell concentration, information regarding the presence of a specific microorganism, or a microbial identification. However, the ability of these rapid microbiological methods technologies are limited in scope in that they cannot be used to perform all the required assays using a single technology platform. A new automated growth-based system simultaneously detects microbial growth, provides an estimation of viable cell counts, and identifies the presence of specified micro-organisms.

Technology. The system is based upon detection of optical variations due to microbial metabolism in liquid medium within a two-zone ready-to-use test vial. An optical sensor monitors optical changes within the vial's reading zone, which is physically separated from the incubation zone. This two-zone approach prevents masking of the optical pathway by product or microbial turbidity and, therefore, eliminates product interference. Separate test vials are used to automatically detect the presence of viable microorganisms and to estimate the concentration of viable counts by monitoring changes in CO₂ production during cellular growth. The CO₂ sensor is composted of a matrix of a polymeric material that is transparent to light and contains an indicator agent sensitive to CO₂ gas generated by micororganisms, changing the indicator color from blue-green to yellow. Alternatively, the hydolysis of fluorogenic synthetic substrates by bacterial enzymes can cause an increase in fluorescence. Also, pH changes can be monitored calorimetrically. Each of these applications can be simultaneously performed using the same instrumentation and at the same time.

The sensitivity of the system is a single viable cell per sample vial; when a single cell replicates to a specific detection threshold level, a positive response is recorded. The threshold level is ~100,000 cells/mL for bacteria and ~10,000 cells/mL for yeast and molds. Additionally, the system yields significantly faster results than the plate count method; one bacterial cell is usually detected within 8–18 hours, a single yeast cell is detected in 20–30 hours, and mold cells require 35–48 hours.

The system creates dynamic patterns as the microorganisms grow in the medium. As shown in Figure 1, the green curve shows a pattern where no growth had occurred. The curve is flat without any significant increase in the signal and no detection time (DT) is observed. The blue curve shows the pattern generated when the microorganism grows in the vial (a DT of 11 hours is observed).

Figure 1

In highly contaminated samples, bacteria are typically detected in 8–12 hours, yeast in 16–24 hours, and mold in 24–35 hours, providing timely warning of contamination.

The system. Each BioLumix instrument has a capacity of 32 sample locations with a single incubating temperature. Multiple instruments (up to 32 instruments) can be attached to a single computer with Windows-based program that controls the operation of the instrument(s) and is barcode capable. The software is validated to meet 21 CFR Part 11 requirements, provides an audit trail, operator identification (log in and log out), trend analysis, and provides various data reports. Detection events are automatically displayed.

A critical element of the technology is the two-zone detection vial with an upper incubation zone where the sample is added and a lower reading zone that remains optically clear and free of turbidity from microorganisms and sample components. This two-zone vial design eliminates interference of the optical pathway during color and fluorescence monitoring by the sample and microbial growth.

Comparative testing of products: The data generated for the comparison of the growth-based rapid method to USP <61> Microbial Examination of Nonsterile Products: Microbial Enumeration Tests, is summarized in Table I. The table shows the specification ranges tested, the number of samples, and the percent agreement between the new method and the USP plate count method using soybean casein digest agar.

Table I

Total aerobic count: (i) Naturally contaminated samples: Two hundred and one nonsterile over-the-counter medicine samples such as vitamins, antacids, suppositories, laxatives, ibuprofin, and aspirin were tested. All noninoculated samples were below the detection level by both methods. There was 100% agreement between the two methods in classifying samples as above or below the specification levels. This indicates the ability of the BioLumix system to yield equivalent results to the plate count method when detecting growth over the range of 10 cfu/g to 1000 cfu/g.

(ii) Inoculated products: Fifty-nine products were inoculated with all the organisms cited in USP <61>. Ten inoculated samples had counts below the specification level and all were correctly classified as the challenge organism by the BioLumix system. There was 100% agreement between the two methods in determining whether samples were above or below the specification level.

As shown in Table I, similar data were obtained for yeast and mold as well as Gram negative bile-tolerant bacteria. The data indicate the ability of the system to yield equivalent results to the plate-count method over the range of 10–1000 cfu/g for both of these assays.

The method has good specificity in detecting target organisms and excluding nontarget flora, and the detection limit for the system equals or is slightly better that the limit for the plate count method. High precision or repeatability was obtained for all three assays tested. It can be used to detect the presence or absence of organisms, total aerobic count, the presence of yeast and mold, enterobacterial count, and absence of objectionable organisms in 10 grams of product. By encompassing both USP types of testing, the system offers a complete screening solution, making the existing microbial testing simpler, faster, and automated.