Cherie Schneider is associate research scientist-microbiologist at Pfizer CentreOne.

Cherie Schneider

Recent industry data indicating that biological formulations containing polysorbate and a divalent-cation chelating buffer may exhibit low endotoxin recovery (LER) (1) have resulted from a 2012 FDA guidance document that requires the development of a method “for storing and handling (including product mixing) samples for bacterial endotoxins analysis using laboratory data that demonstrate the stability of assayable endotoxins content” (2).

	Representatives of the biopharma industry--biopharmaceutical companies and suppliers-have argued that LER is more accurately described as low lipopolysaccharide recovery (LLR) and may be overcome by using naturally occurring endotoxin (NOE) (3). FDA, however, is reluctant to accept NOE studies due to lack of NOE standardization, leaving the pharma industry in a state of limbo.

	Until it is determined whether a standardized NOE will become available and if FDA will accept this approach, the fastest way to market for a new biologic drug is to use traditional approaches to overcome endotoxin masking using control standard endotoxin (CSE)/reference standard endotoxin (RSE) methods before resorting to more drastic and regulatory-risky measures involving NOE. This strategy is consistent with the industry’s “worst-case” approach for validations; a method resulting in the more difficult recovery of CSE/RSE would likely result in recovery of NOE.

	LER studies conducted by Pfizer CentreOne, Pfizer’s contract manufacturing organization, were used to develop an approach for overcoming LER using CSE/RSE. The studies revealed that there is no one-size-fits-all solution and resulted in multiple, molecule-specific solutions to overcome LER. Consequently, a deep understanding of the endotoxin assay and a variety of endotoxin detection technologies are both essential for successful completion of LER studies as required in question 3 of the FDA endotoxin guideline (2).

	Endotoxin impurities in biologic drugs can induce dose-dependent responses with the potential to be life threatening. Endotoxin is present in the outer layer of gram-negative bacteria and consists of surface proteins, lipoproteins, and phospholipids surrounding lipopolysaccharide (LPS) molecules. LPS in pure form does not exist naturally but is the component of endotoxin that exhibits biologic activity.

	LER studies are a recent industry requirement since FDA issued 

Guidance for Industry, Pyrogen and Endotoxins Testing: Question and Answer

s, question 3 (2). Genentech researcher Joseph Chen gave a presentation at the 2013 PDA Annual Meeting in which he revealed that spiked amounts of LPS were not recovered from some undiluted formulated biologics (1).

	According to the 2012 guidance document, biopharmaceutical manufacturers are required to demonstrate that there is no loss of measurable limulus amebocyte lysate (LAL) activity (i.e., masking) over time when endotoxin is added to drugs formulated with chelating buffers. The loss of biological activity has been demonstrated to occur when using a chemically purified LPS (CSE and RSE), but is less likely to occur when using NOE (4) due to structural/chemical differences of the NOE molecule. Unlike the chemically purified LPS, which are unprotected and smaller in size, endotoxin is present in cell-wall vesicles and thus experiences limited exposure to chelating agents and buffer solutions. More specifically, it is the Lipid A portion of LPS which is responsible for the biological activity and is what is masked in the presence of chelating buffers.

	Benefits of using a traditional approach

	The US Pharmacopeial Convention (USP) Microbiology Expert Committee has proposed the use of a well-characterized NOE, which could potentially lead to a solution and save costly sample pretreatment studies (5). Questions remain from FDA about using NOE testing, however, such as what material to use as a standard for hold-time studies.

	Therefore, until such time as FDA accepts the use of a NOE standard for hold studies, the use of traditional CSE/RSE-based methods is expected to provide the fastest pathway to commercialization of drug products in the United States. Following protocols that meet the requirements established in 

) Chapter <85> “Bacterial Endotoxin Test” (6) keeps validation work to a minimum compared to approaches based on NOE methods.

	Pfizer CentreOne’s approach involves the use of treatment studies and diluents as outlined in  

 Chapter <85> that aim to restore Lipid A activity. It is important to note, however, that the mechanism for LLR varies from one drug to another. Consequently, molecule-specific sample treatments must be developed and tests performed to demonstrate that the altered sample treatments used in hold studies do not interfere with endotoxin recovery.

	Needed: Information, reagents, and methods

	An understanding of possible reasons for LLR for a given molecule/formulated product is crucial to shortening testing timelines and getting drugs to market more quickly. Because mechanisms can vary significantly from molecule to molecule, experience with different types of biologic drug substances provides a significant advantage.

	Multiple studies across a wide range of compounds and formulations revealed that the characteristics of biologic molecules have a direct impact on their interaction with endotoxin. It is therefore necessary to have as much information as possible on the structure and physicochemical properties of each drug to be evaluated in order to be able to more accurately predict how they will behave and to select sample pretreatment conditions that will have the greatest likelihood of being successful.

					Table I. Potential methods and pretreatment options.

					Dilution in pH buffer/cation/dispersant buffer

					Dilution in cation/dispersant buffer

					Dilution in buffer with intermittent vortexing over specified amount of time*

					Dilution in buffer and heat treatment plus vortexing*

					*Vortexing and heat can help disrupt binding between endotoxin and proteins.

	In addition, because in most cases the pretreatment conditions--including the detection method--are molecule-specific, having access to a variety of reagents and detection methods is essential for identifying the optimum test protocol for a given product. 

 lists potential methods and pretreatment conditions.

	Access to reagents from multiple vendors is as important as access to different reagent types. Even slight variations in buffering capacities, for instance, can have a significant impact on study results. The author has observed different recoveries using the same method and pretreatment conditions but with materials from different vendors (

 lists detection methods and sample treatments used for eight different products evaluated in the studies as part of biologic license application submissions. Six were monoclonal antibodies and the other two were proteins. All samples had some component of buffers known to cause LER (citrate/phosphate buffer with polysorbate) and showed some form of masking that was overcome by sample pre-treatment.

					Table II. Examples of products for which overcoming low endotoxin recovery was achieved using sample pretreatment. LAL is limulus amebocyte lysate. LRW is LAL reagent water.

					Dilution in cation buffer for specified amount of time/temperature

					Dilution in cation buffer with dispersant/specified vortex time

					Dilution in cation buffer with dispersant, vendor specific

					*For these studies successful recovery was achieved at three days.

					**LAL reagents from different vendors should be evaluated for the detection method chosen as they each have slightly different buffering capacities.

	In most cases, a combination of several treatments (e.g., dilution in cation/dispersant buffer over time with heat and vortexing) was required to achieve a successful result. Passing results (>50% recovery of endotoxin) were targeted for seven days after CSE was spiked into each sample. A comparison of CSE recovery for different detection methods is shown in 

. The final product vial was used in each study and samples were stored at 2–8 °C after spiking.

	The general protocol begins with a method screen to determine which approved endotoxin detection method and LAL buffer is most compatible with the product matrix. The most compatible endotoxin detection method is then used to screen for a non-interfering test dilution by evaluating positive product control recovery via traditional inhibition/enhancement methods. Finally, the non-interfering dilution is used in an LER/LLR screen by evaluating CSE/RSE recovery via different sample pretreatments and conditions. Results for four drugs are shown in 

	The spike of CSE/RSE is made directly into the finished product vial without diluting out the drug. The endotoxin spike should be at a concentration such that dilution to the non-interfering dilution will result in an endotoxin concentration at the middle of the standard curve. The choice of conditions is based on any available information about the drug. Once the optimum sample pretreatment and conditions are identified, the actual study is performed.

	Development of optimized sample pretreatments and conditions that are matched to the specific characteristics of each product often allows for the use of traditional endotoxin technology, resulting in a less risky regulatory pathway and a higher probability of receiving approval. Such an approach, however, requires experience with many different molecules and endotoxin test methods, but ultimately results in a quicker path to market.

	1. J. Chen and A. Vinther, “Low Endotoxin Recovery (LER) in Common Biologics Products.” Presented at the 2013 PDA Annual Meeting, Orlando, FL, April 2013.
	2. FDA, 

Guidance for Industry: Pyrogen and Endotoxins Testing: Questions and Answers

, No 3. (Rockville, MD, June 2012).
	3. J. Bolden, et al., 

PDA Journal of Pharmaceutical Science and Technology

 68 (2014) 472-477.
	4. J. Bolden, et al., “The use of endotoxin as an analyte in biopharmaceutical product hold time studies.” 

, 41 (2015).
	5. USP General Announcement, General Chapters-Microbiology Expert Committee, “

Early Input Sought on Proposed Naturally Occurring Endoxtoxin (NOE) Reference Standard

,” May 27, 2016, accessed Nov. 11, 2016.
	6. 

 36NF 34 General Chapter <85>, “Bacterial Endotoxins Test”.

	All Figures Are Courtesy of the Author. 

	When referring to this article, please cite as C. Schneider, "Overcoming Low Endotoxin Recovery," 

	This article was revised on June 12, 2017, after its original publication in the print edition of 

Articles

Understanding of endotoxin assays and a range of detection technologies are essential for effective testing.