Thursday, February 25, 2010

Hot Tubs and Bioreactors

By Dr. Ray Nims

Contaminating organisms which most commonly are under the radar for biopharmaceutical manufacturing operations include bacteria, mollicutes (mycoplasmas and acholeplasmas), and viruses. Various in-process and lot-release detection assays are mandated by the FDA and the International Conference on Harmonisation to ensure that such contaminants are detected in bulk harvests and/or final products as part of assuring patient safety (specified in ICH Q5A R1 and the 1993 Points to Consider in the Characterization of Cell Lines used to Produce Biologics). There is, however, an additional group of organisms which may threaten biologics production (one which is not normally associated with such manufacturing activities) namely, the Mycobacterium fortuitum complex.

The what?? The fortuitum complex is a group of relatively rapid-growing (non-tuberculosis) mycobacteria which is more typically associated with hot tub disease, and the contamination of industrial cutting fluids and foot baths used for pedicures. The group includes M. fortuitum, M. chelonae, M. abscessus, M. immunogenum, M. mucogenicum, M. peregrinum, and a few others. These mycobacteria, as well as other groups of non-tuberculosis mycobacteria, can be pathogenic in humans, even those who are immuno-competent. The organisms of the fortuitum complex represent a potential risk to the biopharma industry due to their propensity for forming biofilms, their ability to proliferate in water under relatively low nutrient conditions, their resistance to typical water disinfection methods, and their relatively slow growth in nutrient media.

                    Mycobacteria growing at the liquid/air
                       interface of a growth medium.

These characteristics render the organisms capable of existing in water piping and other surfaces in contact with water or aqueous media. Their slow growth in nutrient media may result in these agents being overlooked in biopharmaceutical manufacturing operations, especially when surveillance methods such as short-term bioburden assays are employed. A few cases of contaminated vaccines and tissue extracts have been reported in the literature (Mycobacterium chelonei in abscesses after injection of diphtheria-pertussis-tetanus-polio vaccine. Am. Rev. Respir. Dis. 1973 Jan; 107:1-8; Abscesses due to Mycobacterium abscessus linked to injection of unapproved alternative medication. Emerg. Inf. Dis. 1999; 5: 681-687).

Are there other examples? It is, unfortunately, difficult to estimate the frequency of occurrence of mycobacterial contamination in biologics manufacturing, since many episodes may lead to premature bioreactor termination, with little evidence to implicate a mycobacterium. It is also likely that episodes may have occurred without being reported in the literature.

Thursday, February 18, 2010

Can't See Your Team?

By Dr. Scott Rudge

The world is shrinking and collaborations growing. If you’re like me, you’re getting less sleep as collaborations go global. It’s all made possible by the internet, but is it good or bad?

I don’t think that it is either good or bad, it’s a new and evolving reality, and figuring out how to collaborate effectively using the internet is a job skill that everyone needs these days. In this blog, there are three tips for collaborating on the internet with your colleagues, local and distant.

Set rules, enforce them, obey them. Every good collaboration needs a good set of rules for how collaborations are to take place. Very few teams take time to communicate the rules, which fails as a strategy for getting people to obey the rules. An example is the use of a shared document repository. Folder names and structure should be thoughtfully laid out. The status of a particular piece of work should be clear from its file name. If the repository features check-in / check-out capability, it should be used religiously. Once set, collaborators should refrain from making additional folders, unless the workspace has been created for brainstorming, rather than, say, assembling a regulatory filing. Once the rules have been set, enforce them. If a team member doesn’t know how to use the technology, teach them; resist the temptation to do file management for them, for example. If you are not leading the collaboration, but contributing, learn the rules and follow them. If you’re confused or uncertain, ask for help. Be a good citizen in your collaboration community.

Don’t use too much email. Email is great for quick communication, horrible for collaborating on technology development. The fastest way to lose critical comments and revisions to your work is to use email to distribute it. You will get as many versions back as you’ve sent out, and collecting all the information back in to one piece of work is laborious and prone to error. Furthermore, emails cross in the ether, and are not always copied to everyone (and when they are copied to everyone, they become even more painful). It is much better practice to post your work in a repository and email a link to it, to let everyone know that your part is done, or in progress, and that they are free to look at it and contribute. Free up your inbox, start posting!

Be transparent. The old days when you can walk down the hallway and get a quick status update from everyone are over. And status updates are needed by management with very little notice. You want your update to be accurate and reflect the true progress of your collaboration, but you don’t have time to call everyone, and quite often it’s inappropriate to do so, because of time zone differences. To make updates easy and accurate, it is again critical to make use of the internet repository for all your “updateable” work. In addition, creating a tracking sheet that is also available in the internet repository is a good idea. Furthermore, it should be the responsibility of every member of the collaboration team to update the tracking sheet, at a frequency mandated by the collaboration leader (see “rules”, above). If you keep your work posted, and your status updated on the tracking sheet, you will be valued as a collaborator, and your work will be appropriately communicated outside the team.

There will be future tips, but these three should get you started towards being an excellent collaborator, much valued by your team. If you are a collaboration leader, you can get better results with less heartache. Good luck!

Wednesday, February 10, 2010

USP 63 Mycoplasma Update

By Dr. Ray Nims

The United States Pharmacopeia’s (USP) new chapter <63> Mycoplasma Tests was planned to become effective on May 1, 2010 as part of USP 33. This new chapter was intended to fill a void in the USP for mycoplasma testing, which had been addressed previously within the FDA’s 1993 Points to Consider in the Characterization of Cell Lines Used to Produce Biologicals and the European Pharmacopoeia (EP) chapter 2.6.7 Mycoplasmas (a separate document applying only to mycoplasma testing of live and inactivated viral vaccines, 21 CFR 610.30 will not be considered here). Since there were some methodological differences between the FDA guidance and the EP chapter, it was hoped by the industry that the USP guidance would serve to harmonize mycoplasma testing as much as practically possible.

Indeed, a quick look at the new USP chapter <63> indicates that the chapter was based in large part on EP chapter 2.6.7. A comparison of the three documents (USP, FDA, and EP) reveals methodological differences in only a few areas. These include the assessment of nutritive properties of the solid growth media (agar) used for mycoplasma testing, the assessment of inhibitory substances in the test material, the incubation temperature ranges to be used, and the number of positive controls to be used.

The EP chapter 2.6.7 states that “The solid medium complies with the test if adequate growth is found for each test micro-organism (growth obtained does not differ by a factor greater than 5 from the value calculated with respect to the inoculum)”. There is a different requirement within USP chapter <63>: “The solid medium complies with the test if a count within a 0.5-log unit range of the inoculate amount is found for each test microorganism”. Assuming an inoculate of 100 colony forming units (CFU), the acceptable ranges for the recovered organisms would be 32-316 CFU for the USP version vs 20-500 CFU for the EP version. The USP version is therefore more stringent in this respect.

Similarly, for the assessment of inhibitory substances, EP chapter 2.6.7 states that “…if plates directly inoculated with the product to be examined have fewer than 1/5 of the number of colonies of those inoculated without the product to be examined” there are inhibitory substances in the test material. The USP version indicates that there are inhibitory substances “…if plates directly inoculated with the test article/material are not within a 0.5-log unit range of the number of colonies of those inoculated without the test article/material.” So the USP version is again more stringent in this respect.

Minor differences in incubation temperature for test cultures exist between the documents (36 ± 1°C for the USP and PTC documents vs 35-38°C for the EP chapter). The USP and PTC documents specify the number and types of positive controls to be used in the assays: at least two known Mycoplasma species or strains should be included as positive controls (one a dextrose fermenter and one an arginine hydrolyzer). The EP chapter specifies that at least one of the six Mycoplasma species listed in the chapter be used as a positive control.

The USP chapter <63> differs from EP chapter 2.6.7 also in that the former does not provide requirements for validation of a nucleic acid-based test for mycoplasma. The USP chapter mentions the possibility of replacing the culture method with an alternative (nucleic-acid or enzymatic) method, stating that the alternative method must be validated and shown to be comparable to the agar/broth and cell culture methods. The EP chapter laid the foundation for validation of a nucleic acid-based mycoplasma detection test for the first time in version 5.8 (effective July 2007). This provided the industry with expectations for implementation of a rapid alternative test to the approved culture test for mycoplasma, which is 28 days in duration. Similar guidance is not yet forthcoming from the FDA or USP.

The issues of nutritive properties and inhibitory substances are not addressed within the FDA’s 1993 Points to Consider guidance. In order to now be compliant with the FDA and EP requirements as well as the new USP chapter, testing labs will have to make adjustment within their protocols to account for the stricter USP criteria for assessing nutritive properties and inhibitory substances. Due to errors within some of the monographs to appear in the issuance of the USP to become effective May 1, 2010, this issuance of USP 33, including chapter <63>, was retracted in January of 2010. However, it will be re-issued in March 2010 with an official date six months after reissue, and the methodological differences may need to be accounted for in testing protocols used by quality control laboratories for which USP compliance is applicable.

Once this chapter becomes effective, the mycoplasma test methods described will be considered compendial, meaning that labs following the methods outlined in the chapter will not be required to perform method validation. The labs will only be required to perform method verification for each test sample type (matrix qualification) per USP <1226>.

Wednesday, February 3, 2010

Mixed Up?

By Dr. Scott Rudge

Determining and defending mixing time is a common nuisance in process validation. There are rarely data existing from process development, and there is rarely time or enthusiasm for actually studying the tank dynamics to set mixing times appropriately. Although there typically are design criteria for tank and impeller dimensions, motor size and power input into the tank, these design criteria are rarely translated to process development and process validation functionaries.

There are resources for mixing time determination if the basic initial work has been done. A very elegant study is included in the recent PQLI A-Mab Case Study produced by the ISPE Biotech Working Group. This study shows how to scale mixing from a lab scale 50 L vessel where a correlation between power and Reynolds number has been developed, to mixing vessels of 500 and 1500 L scales. The study requires that two critical dimensional ratios remain nearly constant on scale up, the diameter of the impeller divided by the diameter of the tank, and the height of the fluid level to the diameter of the tank. The study shows very close agreement between predicted mixing time and actual mixing time. The basis for the scale up is that the power input per unit mixing volume should be constant from scale to scale.

When the dimensional ratios cannot be kept constant, there are still rules for scale up. For example, as shown in the chart below (from Perry and Chilton’s Chemical Engineers’ Handbook, 5th edition, 1973), as the impeller diameter increases relative to the tank diameter, the relative power requirement declines, but the torque required to turn the impeller increases. This correlation can be used to adjust the power requirement to the scale up condition.

Additionally, there is “general agreement that the effect of mixer power level on mass-transfer coefficient is greater before than after off-bottom motion of all particles in a solute-solvent suspension is achieved (op.cit.)”.

In other words, once particles have been fluidized off the bottom of the vessel, whether they are carried all the way to the top of the vessel or not is not so important when it comes to predicting complete dissolution of the solids. At that point, the mass transfer coefficient is related only weakly to the power input, as shown below. Mass transfer coefficients for the dissolution of solids can be easily determined in the lab, and do not have to be determined again and again for new processes.
Knowing the minimum power requirements for particle suspension and the mass transfer coefficients for the solids being dissolved allows estimation of mixing times required for preparing a buffer. Knowing the mixing time allows the manufacturer to schedule buffer or medium preparation more precisely, eliminating over-processing or incorrect processing (a principle of lean manufacturing) and helps to guarantee a quality reagent/intermediate is produced each time, on time, and ready to implement in the next manufacturing step.