Pasteurization of Mycobacterium paratuberculosis in whole milk

AF Hope, PA Tulk, and RJ Condron

Victorian Institute of Animal Science,
475 Mickleham Road,
Attwood 3049,
Australia

Table of contents.

• Abstract
• Introduction
• Materials and Methods
• Results
• Discussion
• Acknowledgements
• References
• Table One: Pasteurisation of raw whole milk artificially seeded with Mycobacterium paratuberculosis using a small-scale commercial unit
• Table Two: Level of detection of culture for Mycobacterium paratuberculosis in milk

Abstract

It is preferable to use commercial pasteurizing units to assess bacterial thermosusceptibility and establish the efficacy of pasteurization because the heating and cooling differentials generated in laboratory heat treatment may not simulate commercial conditions. Seventeen batches of raw milk were loaded with 10²-10⁵ CFU/mL of Mycobacterium paratuberculosis and pasteurized in a small scale commercial pasteurizing unit at temperatures ranging from 72-90^oC for 15-35 seconds. Up to 20 samples were tested from each batch and M. paratuberculosis was not isolated from 96% (275/286) of pasteurized milk samples, representing 10⁴ reductions in mycobacterial concentrations as radiometric culture could detect about one colony forming unit per milliliter. Viable mycobacteria were not recovered when raw whole milk was loaded with less than 10⁴ mycobacteria per milliliter. They were not detected in any of five batches of milk pasteurized at 72-73^oC for 25-35 seconds, which are the minimum conditions applied when this machine is used commercially to correct for laminar flow in the holding tube. At shorter times than recommended no viable bacteria were isolated from seven batches, however small numbers of viable bacteria were cultured from four of eight batches heat treated at 72-73^oC for 15 seconds and one of four batches treated at 82-92^oC for 15 seconds. In the five batches where M. paratuberculosis was recovered the raw whole milk was loaded with more than 10⁴ mycobacteria per milliliter. Survival of M. paratuberculosis in experimentally inoculated batches of milk in the small-scale commercial unit cannot be directly extrapolated to commercial pasteurization of naturally infected milk in dairy factories because of artificially high mycobacterial loads used in these experiments, possible differences in the thermosusceptibility of laboratory cultured mycobacteria, and features of the small-scale unit. Pasteurization in a small-scale commercial unit used in these experiments appears to be more efficient at killing mycobacteria than laboratory heat treatment systems.

Introduction

Mycobacterium paratuberculosis is not a recognized agent of food-borne disease, however there have been recent investigations of its thermosusceptiblity following its detection in patients with Crohn's disease³.

M. paratuberculosis has been isolated from raw milk. Organisms excreted in feces are the most likely source of milk contamination, although endogenous infection has also been demonstrated by culture of the mycobacteria from asceptically collected milk^17,20,23, supramammary lymph nodes¹⁷, and deep udder tissue⁶. M. paratuberculosis has been isolated from the milk of up to 12% of subclinically infected cows^16,17,19, from the milk of between 5% and 82% of clinically infected cattle^4,17,19,23, and from colostrum¹⁶. Using a highly sensitive method of detection, DNA segments specific to M. paratuberculosis were found in 6.25% of 336 retail milk samples in the United Kingdom¹².

Pasteurization, heating to 72.4^oC for at least 15 seconds and immediately cooling to 3.5^oC, destroys milk-borne pathogens and delays development of spoilage bacteria. Viable M. paratuberculosis were recovered after heating at 72^oC for 15 seconds in two recent laboratory experiments^3,8. Estimates of bacterial thermal death time are influenced by the heating and cooling lags generated within the heat treatment systems^13,18. It is preferable to establish bacterial thermosusceptibility using commercial pasteurization units, although in practice their use is limited by their availability and the large volumes of milk required for each experiment. The aims of this study were to establish the effect of pasteurization under commercial conditions on the viability of M. paratuberculosis, using a small-scale continuous flow pasteurizing machine.

Materials and Methods

Seventeen batches of 5 L to 15 L of raw milk, artificially seeded with M. paratuberculosis, were pasteurized by the high-temperature short-time (HTST) method in a small-scale commercial pasteurizing unit

The pasteurizing unit featured a positive displacement pump^a, Alfa-Laval P20-HB heat exchange unit with a bank of 30 plates, an Alfa-Laval GM2 hot water pump with thermostat, a 117 cm stainless steel holding line with thermometer at the terminal end, and glycol chilling. The small-scale pasteurizing unit used in this study differed from commercial units in the following features; it had a linear holding tube, it did not have filters to remove large particulate debris, an homogenizer, or flow diversion valve. Holding times were altered by changing the rate of flow of milk through the pasteurizing unit. Milk particles flowing through straight sections of the holding tube were subject to laminar flow rather than turbulent flow generated in the holding tubes of larger units.

M. paratuberculosis was obtained from field cases of bovine and caprine Johne's disease in Australia, an isolate from an Australian Crohn's patient, and a reference culture ATCC 19698. Large quantities of mycobacteria used to inoculate batches of milk were cultured in Watson Reid synthetic medium, adjusted to pH 5.8, and supplemented with 2mg/L mycobactin J^b. The bovine field isolates were passed no more than five times. M. paratuberculosis was recovered from the grossly thickened and corrugated ileum of two dairy cattle in advanced stages of Johne's disease (in vivo inoculum). A total of 1,200 g of macerated mucosa was digested in 0.5% trypsin at pH 7.8, centrifuged, and the pellet decontaminated in 0.75% hexadecylpridiniumchloride (HPC).

Initial mycobacterial concentrations in raw milk were estimated from five 50 mL samples. Following centrifugation the pellets were decontaminated in 0.75% HPC overnight and an additional incubation at 37^oC for four hours with 0.2 mL of a commercial antibiotic mix (PANTA Plus^c)in 5 mL water. Five 50 mL samples from ten-fold dilutions were cultured onto Herrold Egg Yolk slopes containing mycobactin (HEYJ).

Pasteurized milk samples (50 mL) were collected and immediately processed for radiometric culture. 0.1 mL of suspension was injected into modified 12B Bactec^d bottles supplemented with 1.0 mL of egg yolk/PBS suspension, 4 mg of mycobactin J, and 0.2 mL of PANTA Plus. The bottles were incubated at 37^oC and monitored weekly for evidence of bacterial growth. Results from samples containing viable M. paratuberculosis were considered to be binomially distributed in each batch and their 95% confidence intervals were estimated using two-tailed tests in batches yielding M. paratuberculosis and one-tailed tests in batches that did not yield growth¹⁵.

The sensitivity of the radiometric culture system was determined for a bovine field strain, in vivo mycobacteria pooled from two clinical cases, an isolate from a Crohn's patient, and Strain 316V. These mycobacteria were placed in 10 mL of pasteurized whole milk, and ten-fold dilutions were cultured by radiometric and conventional methods. Mean mycobacterial concentrations (CFU/mL) in raw milk were estimated using arithmetic means from ten-fold sample dilutions which yielded 10-80 colonies per slope. A small random selection of pasteurized milk samples (11) which were negative on radiometric culture were examined for the presence of the DNA insertion sequence (IS900) by polymerase chain reaction (PCR)²² and all gave positive reactions.

Results

M. paratuberculosis was added to batches of raw milk resulting in concentrations of 10² to 10⁵ CFU/mL (Table 1). M. paratuberculosis was not recovered by radiometric culture in 96% (275 of 286) of pasteurized milk samples. Twelve of 17 batches of pasteurized milk tested negative, including ten batches loaded with 10²-10⁵ CFU/mL bovine field strains, and single batches containing 10² CFU/mL ATCC 19698, and an isolate from a Crohn's patient 10⁴ CFU/mL. Viable mycobacteria were not recovered from any batch containing initial loads of less than 10⁴ CFU/mL. M. paratuberculosis DNA was detected by PCR in seven pasteurized milk samples which were negative by radiometric culture.

Eleven samples yielding viable M. paratuberculosis were from Batches A, G, K, M, and O (Table 1) comprising 4 batches pasteurized at 72-73^oC for 15 seconds, and one batch pasteurized at 82-90^oC for 15 seconds. The samples were collected from the unit's outlet during early (4), mid (3) and late (4) processing of the batches. The highest rate of recovery was from a batch of skim milk.

Radiometric culture was more sensitive than conventional culture and was able to detect M. paratuberculosis at concentrations of about 1 CFU/mL (Table 2). Conventional culture was not suitable for enumeration of organisms in pasteurized samples.

Discussion

The commercial pasteurizing unit efficiently killed mycobacteria, with no M. paratuberculosis recovered by radiometric culture from 96% of pasteurized milk samples, representing in the order of 10⁴ reductions in mycobacterial concentrations. Commercial pasteurization appeared to be more effective than heat treatment in laboratory systems, as mycobacteria were recovered from a higher percentage of milk samples treated in laboratory models of pasteurization^3,8.

Small numbers of M. paratuberculosis survived pasteurization in half of the batches of milk at 72-73^oC for 15 seconds in the small-scale continuous flow pasteurizing machine used in these experiments. Important functional differences between the small-scale commercial pasteurizing unit and larger commercial units were the lack of a homogenizer and a straight holding tube which results in laminar flow of milk particles rather than turbulent flow. To correct for laminar flow effects, the minimum holding time of 25 seconds is applied when this machine is used commercially. M. paratuberculosis was not recovered from any of five batches of milk held at 72-73^oC for 25 seconds. In the small number of batches treated adequate holding time appeared to be more effective in killing M. paratuberculosis than higher temperatures.

Survival of M. paratuberculosis in some batches of milk cannot be directly extrapolated to commercial pasteurization of naturally infected milk in dairy factories because of the artificially high mycobacterial loads used in the experiments, possible differences in the thermosusceptibility of laboratory cultured mycobacteria, and physical limitations of the small-scale unit. The thermosusceptibility of laboratory cultured mycobacteria varies with their culture media and age and growth phase of cultures through mechanisms such as changes in cell membrane viscosity and the production of heat shock proteins^2,6,9,24. Laboratory cultured mycobacteria may have greater thermotolerance compared with in vivo mycobacteria¹¹. Mycobacterial concentrations used in this experiment were one million times higher than reported concentrations of M. paratuberculosis in naturally infected milk¹⁹.

The proportion of pasteurized milk samples containing viable mycobacteria was greater in a skim milk compared to whole milk batches. This may reflect a larger inoculum of mycobacteria in the skim milk batch, improved repair of sublethally affected mycobacteria in skim milk, or increased efficiency of recovery from skim milk. Pasteurization efficacy in whole milk may differ from skim milk or homogenized milk if killing of lipophilic mycobacteria is affected by the distribution of milk fat globules in these liquids.

Pasteurization was effective in reducing concentrations of M. paratuberculosis to below the detectable limit (about 1 CFU/mL) in twelve batches of raw whole milk. The inability to recover bacteria post-pasteurization does not ensure sterility. Criticisms of pasteurization trials yielding negative results are; suboptimal cultural conditions for bacterial recovery, employment of sampling methods unlikely to detect viable bacteria, and extrapolation of heat treatment of bacterial suspensions to bacteria that are usually found within cells. Although it has been postulated that bacteria located within bovine phagocytes will have a higher heat tolerance⁵, many studies have found pasteurization is still effective against intracellular pathogens^1,7,10. There was no evidence from these experiments that M. paratuberculosis in tissues (in vivo inoculum) was resistant to heat treatment. It has not been established whether radiometric culture is optimum for the recovery of heat damaged mycobacteria. Injured organisms are sensitive to their immediate chemical environment, and it is possible the radiometric liquid media was more conducive to recovery of mycobacteria than conventional culture media despite its antibiotic supplementation. Sample volumes of 50 mL were selected to increase the likelihood of detecting bacteria¹⁴.

Pasteurization studies investigating the control of potentially hazardous milk-borne bacteria often give conflicting results. Kill rates in pasteurization studies depend on heat differentials generated in the heat treatment systems, and varying come-up times, and cooling lags explain contrasting results between laboratory and commercial systems¹⁸. Further studies using commercial pasteurization of milk from naturally infected cows and herds will provide confidence in the assessment of pasteurization efficacy against M. paratuberculosis.

Acknowledgments

This work was financially supported by the Dairy Research and Development Corporation. The authors would like to thank the Victorian College of Agriculture and Horticulture for access to their pasteurizing unit, with special thanks to Selwyn Stokes for his assistance and guidance.

References

Table One: Pasteurisation of raw whole milk artificially seeded with Mycobacterium paratuberculosis using a small-scale commercial unit

^a Estimate based on counts <10
^c Contaminated samples excluded from report
^r Raw skim used in this batch

Table Two: Level of detection of culture for Mycobacterium paratuberculosis in milk

^a On any of five HEYJ slopes
^b In a single modified Bactec 12B bottle