Modeling the Efficacy of HTST for Inactivating Mouse Minute Virus

Heat inactivation, and, particularly, high-temperature short-time (HTST) treatment, has been evaluated by various biologics manufacturers as a potential upstream barrier technology for mitigating the risk of introducing adventitious viral contaminants into manufacturing processes via contaminated cell-culture reagents and other process solutions (1–3). Adventitious agent testing of such reagents alone has not provided adequate risk mitigation in the past, due perhaps to the limitations of the testing methods for detection of low-level contaminants. For evaluating the efficacy of HTST, a parvovirus such as mouse minute virus (MMV) is typically used, because the parvoviruses appear to be among the most highly resistant of the virus families to heat inactivation in liquids (4, 5).

There have been varying reports of the efficacy of HTST for inactivating the parvovirus MMV. For example, Murphy, et al. reported that an HTST treatment consisting of 102 °C/10 s was able to completely inactivate MMV spiked into a proprietary LM7105 culture medium at low titer (10 tissue culture infectious dose₅₀/mL [TCID₅₀/mL]), but not the MMV spiked into the same medium at the moderately higher titer of 100 TCID₅₀/mL (3). Not surprisingly, these authors concluded that the efficacy of inactivation under these HTST conditions would not be sufficient to mitigate the risk of a bioreactor contamination if an upstream reagent was contaminated with MMV at 100 TCID₅₀/mL or higher. Schleh et al. reported that a variety of HTST treatment conditions, the most stringent being 115 °C/30 s, inactivated MMV spiked into two cell-culture medium formulations (1). At the highest temperature, the treatment resulted in ≥3.2 log₁₀ inactivation of virus in medium one and ≥4.6 log₁₀ inactivation in medium two. These authors, therefore, considered the HTST treatment to represent an efficacious and robust risk mitigation step. Which conclusion regarding the efficacy of HTST for mitigating the risk of introducing a parvovirus into a bioreactor via a contaminated upstream process reagent is correct? The current authors believe that both conclusions are correct, but would argue that relatively stringent HTST conditions will be required to routinely achieve the desired risk reduction.

To demonstrate that this is indeed the case, the authors of this exercise have employed a method of modeling the heat inactivation of viruses (6). Such modeling is necessary because most of the empirical studies of the heat inactivation of MMV have not been performed at the specific conditions of temperature and time used in the HTST studies, and both of these factors (time and temperature) are known to play important roles in determining inactivation efficacy, even in the same sample matrix. The modeling done for the purpose of predicting HTST efficacy at different temperatures and time was based on a dataset for MMV spiked into water reported by Boschetti et al. (7).

Modeling of viral inactivation by heat

In brief, the modeling approach consists of assigning a power function line fit directly to a plot of D (the time required to inactivate one log₁₀ of virus) vs. temperature. As with the more traditionally used z value approach, the power function approach requires that inactivation must have been assessed and D values obtained at three or more different temperatures. The power function line fit (Figure 1) may be obtained using Equation 1 in Microsoft Excel as:

where a and b are constants assigned during the line-fit process.

Once the power function parameters have been obtained (see Table I), the equation may be solved for D (minutes) at a given temperature (i.e., 102 °C or 115 °C in the current effort) using Equation 1. For modeling the extent of inactivation of viruses under the specific conditions of the HTST treatment (i.e., 102 °C/10 s or 115 °C/30 s), the following Equations 2 and 3 were then used:

where D_{102 °C} and  D_{115 °C} are the times (in minutes) required to inactivate one log₁₀ of virus at 102 °C  or 115 °C, and 10 s = 0.17 minutes, and 30 s = 0.5 minutes.

The use of this modeling approach is valid to the extent that the following assumptions are met: first-order kinetics apply to inactivation by heat over multiple log₁₀ reductions in titer; and the power function equations are valid for predicting inactivation at 102–115 °C even in cases where the empirical results used as the basis of the modeling were obtained only at lower temperatures. The modeling parameters derived from the MMV inactivation reference (7) utilized in the authors’ study are shown in Table I. The assumptions for first-order inactivation are generally met for heat inactivation and were met in the dataset used. As shown in Table I, the dataset used for the modeling involved temperatures as high as 90 °C. The authors have determined that the goodness of fit of the empirical data points to the power function trend line influences the accuracy of the modeling for providing results for temperatures outside of the empirical range. This goodness of fit to the power function trend line is associated as well with goodness of fit of the empirical data to the linear line fit of the log₁₀D vs. temperature plot that is used to determine the historically used z value (i.e., the value corresponding to the temperature required to cause a one-log₁₀ change in D). In either case, dramatic departures from the trend lines indicate potential inaccuracies with regard to the measurements and indicate that any modeling done using the data may also be inaccurate. In the case of the MMV dataset used for the modeling of HTST inactivation, the R² values for the power function fit of D vs. temperature and for the linear fit of the log₁₀D vs. temperature were each 0.99.

Results of Modeling of MMV inactivation  under HTST conditions

The modeling (Table II) based on the predicted D_{102 °C} value for MMV indicates that ~0.3 log₁₀ of inactivation might be expected during a 10 s treatment such as that used by Murphy et al. (3). If this prediction is reflective of the actual conditions used for the empirical HTST experiment reported by these authors (3), it is not surprising that 100 TCID₅₀/mL of MMV were not completely inactivated. These modeling data suggest that a temperature of 102 °C for 10 s would be insufficient for assuring inactivation of MMV (Table II).On the other hand, the modeling indicates that exposure to 115 °C for 30 s (Table II) should provide greater, albeit perhaps not complete, inactivation of MMV.

The modeling results shown in Table II would not be able to account for any inactivation that might occur during HTST treatment that might be associated with the time required for ramping up temperature or cooling down following HTST. The extent of such inactivation is probably minimal, however, as most HTST treatments are based on residence time in heating chambers and there is provision for rapid cooling after HTST treatment. The modeling also does not reflect the inactivation of MMV within the specific matrices used in the HTST studies. On the other hand, virus spiked into water should represent a best case for heat inactivation, as there would be few solutes present to afford protection or stabilization of the virus with respect to the effects of heat inactivation.

While Schleh et al. (1, 8) reported extensive MMV inactivation even under less stringent HTST conditions than those modeled here (i.e., shorter times at 115 °C and at temperatures as low as 102 °C), the modeling results herein and the results obtained by Murphy et al. (3) suggest that for routine HTST treatment intended to provide robust inactivation of extremely heat-resistant viruses such as MMV, the most stringent condition compatible with preservation of reagent performance should be used.

Discussion
Following the initial experience of MMV contamination of bioreactors by Garnick (9), similar contamination events have now been reported by Moody et al. (10) and Skrine (11). This indicates that MMV contamination remains a potential problem for biologics manufactured in Chinese hamster ovary cells, justifying the continued mandate (12, 13) for MMV testing of bulk harvest material derived from such manufacturing operations. Adding to the difficulty in mitigating risk of such contamination is the fact that the actual points of introduction of the MMV have not in all cases been definitively understood. Unlike many of the viral contaminants that have been encountered by biologics manufacturers(e.g., Cache valley virus, reovirus, and vesivirus), it is not believed that MMV is introduced through contaminated animal-derived raw materials. Moody et al. attributed the MMV contamination in their case to an animal-derived material-free (recombinant) medium additive (10). The lack of understanding of the root source of the MMV contamination has, in part, driven the evaluation of barrier technologies (e.g., filtration, HTST, and ultraviolet irradiation) for preventing virus introduction via upstream process reagents.

In the modeling exercise described here, the authors have attempted to contribute to the understanding of some apparently discrepant reports (1, 3) of the efficacy of HTST treatment for inactivating MMV. Extrapolation of results from a heat-inactivation study performed at lower (non-HTST) temperatures has been performed using a modeling technique developed for the purpose of affording inter-family comparisons of susceptibility to heat inactivation (4, 6). Inactivation at lower temperatures is fraught with fewer difficulties, as the duration of heating time measured in minutes is more easily controlled. The authors’ modeling results suggest that at less than 115 ˚C, reproducible and robust inactivation may not routinely be achieved for MMV.

It is important to keep in mind that inactivation efficacy is typically reported in terms of log₁₀ inactivation, implying that 100% (i.e., complete) inactivation may not be practically attainable. As with other inactivation modalities, mitigation of the risk of introducing viral contaminants through HTST will be maximized by attaining the highest possible log₁₀ inactivation result compatible with maintaining the required performance characteristics of the process material (inactivation matrix) being treated.

References
1.     M. Schleh et al., Biotechnol Prog. 25 (3), pp. 854–860 (2009).
2.     B. Weaver and S. Rosenthal, PDA J. Pharm. Sci. Technol. 64 (5), pp. 436–439 (2010).
3.     M. Murphy, G.M. Quesada, and D. Chen, Biologicals. 39 (6), pp. 438–443 (2011).
4.     R.W. Nims and M. Plavsic, J. Microb. Biochem. Technol. 5, pp. 136–141 (2013).  
5.     R.W. Nims and M. Plavsic, Bioprocess. J. 13 (2), pp. 6–13 (2014).
6.     R. Nims and M. Plavsic, Bioprocess. J. 12 (2), pp. 25–35 (2013).
7.     N. Boschetti et al., Biologicals 31 (3), pp. 181­–185 (2003).
8.     M. Schleh et al., Amer. Pharm. Rev., pp. 72–76 (2010), , accessed Feb. 5, 2015.
9.     R.L. Garnick, Dev. Biol. Stand. 88, pp. 49­–56 (1996).
10.  M. Moody et al., PDA J. Pharm. Sci. Technol. 65 (6), pp. 580–588 (2011).
11.  J. Skrine et al., PDA J. Pharm. Sci. Technol. 65 (6), pp. 599­–611 (2011).
12.  FDA, Center for Biologics Evaluation and Research, Points to Consider: Points to Consider in the Manufacture and Testing of Monoclonal Antibody Products for Human Use (Rockville, MD, Feb. 27, 1997).
13.  EMEA/CHMP/BWP, Guideline on Virus Safety Evaluation of Biotechnological Investigational Medicinal Products (London, July 24, 2008).

About the Authors
Raymond W. Nims, PhD, is a senior consultant at RMC Pharmaceutical Solutions, Inc., Longmont, CO.
Mark Plavsic, PhD, DVM, is head of Corporate Product Biosafety, Genzyme, a Sanofi Company, Framingham, MA, USA.

Article Details

BioPharm International
Vol. 29, No. 4
Pages: 42–45
Citation: When referring to this article, please cite it as R.W. Nims and M. Plavsic, "Modeling the Efficacy of HTST for Inactivating Mouse Minute Virus," BioPharm International (29) 2016.