Low Endotoxin Recovery, A Brief Overview

• Roche

Introduction

Since first reported by Chen and Vinther in 2013 (Chen and Vinther, 2013), the phenomenon known as low endotoxin recovery (LER) has been broadly observed in certain matrices commonly used for biologic formulations and certain therapeutic proteins. LER is defined as loss of detectable endotoxin activity over time using compendial LAL assays (Bacterial Endotoxin Test = BET per USP<85>/EP 2.6.14./ JP 4.01) when undiluted products are spiked with known amount of endotoxin standards.

LER is a temperature- and time-dependent process. Reduced endo- toxin recovery usually does not occur immediately; rather it is only observed upon storage (several hours to several days). BET method qualification per USP<85>/EP 2.6.14./JP 4.01does not include defined sample storage conditions, which may explain why LER has not been detected by following compendial testing guidance. Because LER poses potential risks that endotoxin contaminations in products may be underestimated or undetected by BET, the U.S. FDA’s Center for Drug Evaluation and Research (CDER) has recently started requesting that companies conduct endotoxin spike/hold recovery studies to determine whether a given biological product causes LER (Hughes, 2015; Hughes et al., 2015). These studies, however, have often produced confounding and sometimes contradictory results, likely due to variations in study designs, experimental procedures, and/or types of endotoxin used.

Recognizing the significance and complex nature of the LER issue, the PDA Biotechnology Advisory Board sanctioned the LER task force (TF) in 2015. The PDA LER TF provides a platform for subject matter experts from academia, U.S. FDA, biopharmaceutical companies, and reagent suppliers/testing contractors work together to develop a science- based and data-driven strategy in dealing with the LER phenomenon.

What Do We Know So Far?

Definitions and Terminology

LER vs. LLR

The term LER was initially used to describe the phenomenon of reduced endotoxin recovery when undiluted products containing either polysorbate/citrate or polysorbate/phosphate are spiked with a known amount of endotoxin (Chen and Vinther, 2013). Subsequently, LER has also been observed in other matrices (see Hughes et al., 2015). Several studies have reported that loss or reduced recovery linked to polysorbate and chealtor (e.g. citrate and phosphate) is only observed with purified lipopolysaccharide (LPS) such as CSE or RSE but not less purified natural occurring endotoxin (NOE). As a result, the term “Low LPS recovery” (LLR) was recently proposed in an attempt to differentiate the significantly different recovery observed in these studies, which are seemingly attributed to the use of different types of endotoxin analytes (Platco 2014). To complicate the matter, it has been recently reported that loss/reduced recovery was also observed with NOE preparations (Hughes et al., 2015, Chen et al., 2016), suggesting that the LER phenomenon is not limited to LPS. In addition, many authors are using the term “endotoxin” and “LPS” synonymously (Galanos and Lüderitz, 1984; Munford, 2016; Williams, 2007).

Endotoxin masking

The term endotoxin masking was used by Petsch et al. to describe reduced LAL reactivity of endotoxin in the presence of cationic proteins (Petsch et al., 1998). The authors argued that endotoxin masking differs from matrix interference as it can’t be mitigated by dilutions or other published procedures. LER has also been termed as endotoxin masking by some simply because it can’t be overcome by dilution (Reich et al., 2016; von Wintzingerode and Chen, 2015). Others however view the LER phenomenon as an example of matrix interference (McCullough, 2016) since it meets the defined interference criteria: (i) conditions that reduce the efficiency of the LAL reagent to detect activity, and (ii) conditions that affect the CSE/ LPS spike such that activity of the LPS itself is masked or diminished (Cooper, 1990, Dubczak, 2011). Additional studies are clearly needed to elucidate the underlying mechanism of LER and better understand the differences and similarities among these phenomenon (LER, masking, and matrix interference).

LER Root Cause

The underlying mechanism of LER remains poorly understood and more research is needed to uncover the biophysical/biochemical details and critical factors involved in LER. Studies on LER linked to the combination of polysorbate and chelating agents have shown that the process is progressive and temperature-dependent (Chen, 2013; Reich et al., 2016). It has been speculated by several groups that LER result from changes in aggregation state of endotoxin (Tsuchiya, 2014 and 2015, Zähringer, 2015, and Reich et al., 2016). Zähringer (2015) hypothesized that endotoxin is “buried” in complex aggregates consisting of surfactant, buffer, and monoclonal antibody (mAbs) products and is only detectable by LAL if released from the aggregates. A two-step model was proposed by Reich et al (2016) to explain the LER phenomenon linked to surfactants and chelating agents. In the first step, endotoxin aggregates are disrupted by chelators, resulting in an increase in the permeability of the aggregates. Subsequently, surfactants covert the supramolecular aggregates to the forms that are not reactive to LAL reagents.

It has been shown that endotoxin recovery can be significantly influenced by spiking analytes (Platco 2015, Bolden et al., 2015 and 2016). Bolden et al. (2015) reported that no LER was observed when NOE was used as the analyte instead of CSE or RSE. However, other studies showed less apparent correlation between endotoxin analytes and endotoxin recovery (Reich 2015, von Wintzingerode and Chen 2015. McCullough (2015, 2016) postulated that differences seen for CSE, RSE, and NOE are caused by the different degrees of purity (CSE and RSE are highly purified whereas NOE are crude endotoxin extracts derived from gram-negative bacteria). Different LPS structures represented by CSE, RSE, and NOEs could be an alternate explanation (CSE from vendor 1 and RSE is E. coli O113:H10:K-, CSE from vendor 2 and 3 is E. coli O55:B5 whereas NOEs are usually prepared from non- E. coli gram-negative bacteria). Gutsmann et al., (2010) have shown that different endotoxin structures exhibit different LAL reactivity. In addition, unpublished data presented during the PDA LER workshop has shown that for some products highly purified endotoxin from Salmonella does not exhibit LER whereas E. coli O55:B5 type CSE did. Finally, it has also been shown that endotoxin recovery could potentially be impacted by experimental procedures such as sampling scheme and mixing time (Bolden et al., 2016).

Approaches to Overcome LER

It has been shown that some BET based protocols are capable of overcoming LER in a product specific manner (Reich, 2015; von Wintzingerode and Chen 2015; Platco et al., 2016). These protocols modify sample preparations by: (i) addition of MgCl ₂ and Pyrosperse (Platco et al., 2016), (ii) dilution with LRW, addition of MgCl ₂ followed by dialysis (von Wintzingerode and Chen 2015), and (iii) using specific reagents provided with the Endo-RS™ kit (Hyglos, Germany) (Reich, 2015). Developing a product specific BET based protocol to overcome LER is time consuming with great uncertainty at this time due to the lack of good mechanistic understanding of the LER phenomenon. Inclusion of adequate controls is critical to ensure the validity of the outcome.

LER and Patient Safety

It was reported that LPS-spiked products while not detected by BET in vitro caused pyrogenic reaction in rabbits upon injection (J. Chen 2013). Such apparent disagreement between the in vitro (BET) and in vivo (rabbit) assays implies that a contaminated pharmaceutical product could go undetected by BET, thus representing a potential threat to patient safety as articulated by the US health authorities (Hughes et al. 2015). Monocyte activation testing (MAT) represents a useful tool in our desire to further assessing LER’s potential risk to patient safety. Individual companies should work closely with health authorities to adequately address the potential risks associated with LER. In case of confirmed LER for drug substance (DS) or drug product (DP) it is possible to mitigate the potential LER risk by implementing a risk-based in process control throughout the drug substance (DS) and drug product (DP) manufacturing processes. The risk based approach should assess endotoxin contents of the furthest downstream step that does not exhibit LER and raw materials including process water used for preparation of buffers applied downstream of the LER affected step. The risk based approach would further take into account bioburden load of final process steps and process water used for batch preparation since gram-negative bacteria might introduce endotoxins. Concerning processing materials and primary packaging components, the endotoxin contents should be specified. Further treatment, i.e. washing, sterilization, and depyrogenization should be in place to reduce potential contamination with endotoxins.

Acknowledgement

The author thanks Dayue Chen for helpful suggestions and the PDA LER Task Force for valuable discussions.

References

1. Chen J, Vinther A. Low endotoxin recovery in common biologics products. Presented at the PDA Annual Meeting, Orlando, FL. April 15-17, 2013.
2. Hughes P. “Endotoxin – A FDA Perspective.” Presented at the PDA 10th Annual Global Conference on Pharmaceutical Microbiology, Bethesda, MD. October, 19-21, 2015.
3. Hughes P, Thomas C, Suvarna K, Chi b, Candau-Chacon, Gomez-Broughton C, Narasimhan L. Low endotoxin recovery: An FDA perspective. BioPharma Asia 2015; 4 (2):14-25.
4. Bolden J, Warburton R, Phelan R, Murphy M, Smith KR, De Felippis MR, and Chen D. Endotoxin recovery using Limulus amoebocyte lysate (LAL) assay. Biologicals. 2016; in press.
5. Platco, 2014. American Pharmaceutical Review, Endotoxin Suppl. September 2014.
6. Reich et al. 2016. Biologicals. 44:417–422.
7. Reich 2015 ” Overcoming Endotoxin Masking in a Drug Product”. PharmaLab 2015, Düsseldorf, 11 November 2015.
8. Chen D., von Wintzingerode, F.; Eaton J., Levin J., Hughes P., PDA’s LER Task Force Holds its First Workshop2016 PDA Letter September
9. von Wintzingerode and Chen J. 2015 “Development of a LAL-based method to overcome LER in a Biologics product”, PharmaLab 2015, Düsseldorf, 11 November 2015
10. Gutsmann et al., 2010. Innate Immunity 16:39-47
11. McCullough K, American Pharmaceutical Review, Endotoxin Suppl. 2015
12. McCullough K, American Pharmaceutical Review, July 2016
13. Platco et al., American Pharmaceutical Review, July 2016
14. Galanos and Lüderitz. 1984. Handbook of Endotoxin, Chemistry of Endotoxin Vol. 1 Elsevier Science Publishers.
15. Williams. 2007. Endotoxins. Informa Healthcare 3rd edition.
16. Munford. J Leukoc Biol published online July 14, 2016.