Evidence suggests it’s not the LAL test that’s broken. The real problem is more likely the insistence that the calibration standard functions like outer membrane vesicles.
As scientists, we recognize that ours is not a static discipline. We are trained to reflect on the scientific literature so that we can codify accumulated knowledge, test new hypotheses and move forward. We are innovators. In the 1970s, “But we’ve always done it this way” was a rallying call to change how we thought about pyrogen testing and work together as representatives of industry, government and academia to have the LAL test accepted as a substitute for the rabbit pyrogen test. I’m privileged, humbled and very proud to have been a part of that effort.
It was variability in LAL reagents in the 1970s, that prompted the creation of a calibration standard prepared from extracted, purified, and formulated lipopolysaccharide (LPS).1 The calibration standard (RSE) assures that all lysates work equally well in water. We make standard series for gel and quantitative assays in water. The authors of the original USP <85> and the 1987 FDA Guideline knew that the recovery of the calibration standard in undiluted product would be difficult, if not impossible for many drug products so they provided us with a work around for product specific interferences called the Maximum Valid Dilution (MVD). The MVD allows us to dilute the product (and corresponding interference) in water to the point where the recovery of the calibration standard control in diluted product (PPG) is the same as recovery in water. Yet, some expect that same calibration standard will fully recover in undiluted product. Why?
The initial FDA and European standards were not made from E. coli, and the European standard was a liquid preparation. E. coli was ultimately chosen for the calibration standard, not because it is more or less resistant to the interferences we see in LAL testing, or because it is “worst case” for pharmaceutical manufacturing, but rather because it was described as a “typical” LPS at the time. Even though the definitions of Control Standard Endotoxin (CSE) in USP XXI and in FDA’s current Q&A Guidance do not require that the CSE be extracted in hot phenol from E. coli, or that it be purified, or that it be further formulated, or that it be lyophilized, some insist that the CSE must be all of those things. Why?
Unexpected LPS results
It is interesting to note that the very first native or natural endotoxin preparation was actually made by FDA.1 Almost 40 years ago, FDA decided to “validate” their laboratories and the methodology was to inoculate an undiluted product with a known activity level of the calibration standard and send it out to the various labs, confident that all would get the same answer. Surprise! The labs could not recover LPS activity. So FDA made an extract of an E. coli culture, used the extract to inoculate undiluted product, sent it out, and lo and behold, the labs recovered the correct level of activity and were therefore validated, qualified and calibrated. Even though FDA’s experience confirms that a native source of inoculum is a very useful tool when dealing with product specific interferences, some still insist that only the calibration standard be used as the analyte for product controls and recovery. Why?
Unlike exotoxins, Lipid A and LPS are not secreted as molecular entities into the extracellular environment and therefore neither purified Lipid A nor purified LPS can be contaminants in parenteral products. Rather, as Gram negative organisms grow, they create buds called outer membrane vesicles (OMV or blebs) that pinch off from the cell membrane and are released into the external environment as free floating mini spheres comprised of outer cell membrane components, including LPS.2 It is these OMV and cell wall fragments that are the contaminants in our products. Even so, some insist that the use of a well controlled “real life” OMV preparation is an inappropriate control for recovery studies. Why?
Darwin was right. Organisms adapt to their environments. We have current confirmation that the biochemical structure of LPS purified from organisms grown in a nutrient rich medium is different than the biochemical structure of LPS embedded in the cell membranes and OMV of stressed organisms. Such a change in LPS biochemistry is essential to maintain the integrity of the organism’s outer membrane and therefore its continued survival in harsh environments such as pharmaceutical water systems and conditions found in some parenteral product formulations.2 Yet, we have a calibration standard that is made in, extracted from, and further purified from E. coli grown in a nutrient rich environment. AND, we assume and insist that this purified LPS is the biochemical, structural and functional equivalent of and therefore representative of naturally modified LPS embedded in the cell membranes of vesicles released from the adapted organisms that contaminate parenteral products. Why?
A Rallying Cry to Scientists
If the new paradigm for some products is to test them undiluted in hold/recovery studies, then we must acknowledge what FDA, industry and vendor scientists have known and published for many years. The vast majority of parenterals will fail recovery studies when we use the calibration standard as the inoculum in undiluted product. We hear that such failures foretell the possibility of false negative LAL tests resulting in pyrogenic outbreaks from product that will be released as the result of an “invalid” test. Yet, 40 years of published history and statistics regarding the LAL test as a screen for parenteral formulations do not support that fear. The published data suggest that it’s not the LAL test that’s broken. The real problem is more likely the insistence that the calibration standard functions like OMV when we know empirically that it can’t and it won’t in most environments. We knowingly and repeatedly set ourselves up for failure. Consequences of this thinking include a return to rabbit pyrogen testing and untold millions of dollars wasted trying to “fix” what many feel is the wrong problem, all with no measurable difference in patient safety. Why?
Why not use “But we’ve always done it that way”, not as a rationale to maintain the status quo, but as a rallying cry to work together again to ask the right questions, address our concerns scientifically and evolve our thinking? While the current calibration standard may be an appropriate choice for comparing LAL reagents in water, history and science tell us that we can’t expect it to calibrate product specific interferences, particularly in many undiluted products. I challenge us, as scientists, to work together and come to consensus to define and accept a properly controlled native Control Standard Endotoxin preparation that can be the appropriate choice when the calibration standard doesn’t work as the control for product recovery and depyrogenation studies.
- Stappen, Melissa and David Hussong. 2017. The LAL Assay for Pyrogen Testing of Parenteral Products: Evolution and Challenges. https://www.pharmaceuticalonline.com/doc/the-lalassay- for-pyrogen-testing-ofparenteral-products-evolution-challenges-0001
- Bonnington, Katherine E., Meta J. Keuhn. 2016. Outer Membrane Vesicle Production Facilitates LPS Remodeling and Outer Membrane Maintenance in Salmonella during Environmental Transitions. mBio. 7(5). September/October
This article first appeared March 16 in American Pharmaceutical Review. It was reprinted with permission.