d":"\ KURZMITTEILUNG ---~------

Endotoxin Detection in Pharmaceuticals and Medical Devices with Kinetic-QCL, a Kinetic- Quantitative Chromogenic Limulus Amebocyte Lysate Assay

Ronald N. Berzofsky BioWhittaker, Inc., USA-Walkersville

ZusammenJassung Summary

Die Beobachtung, daj3 Endotoxin in Extrakten von The observation that endotoxin caused gelation in Limulus Amobozyten ein Gel erzeugt, wurde zu einem in extracts of Limulus amebocytes has been expanded to the vitro Test ausgebaut, einem kinetischen, quantitativen development of an in vitro kinetic, quantitative chromo- Chromogen-LAL-Assay (Kinetic-QCL), mit dem in genic LAL assay (Kinetic-QCL) for the detection of wdssrigen Losungen Endotoxine bestimmt werden endotoxin in aqueous fluids. Within the last 15 years, the konnen. In den letzten 15 lahren hat der Limulus Amo- use of Limulus amebocyte lysate to detect and control bozyten Lysat Test (LAL) zur Bestimmung von Endotoxi- the presence of pyrogenic substances in pharmaceuticals nen in Pharmaka und medizinischen Hilfsmitteln breite and medical devices has gained wide international internationale Anerkennung gefunden. Sowohl die acceptance. Both the United States and European Pharmakopoen der USA als auch Europas enthalten Pharmacopoeias contain descriptions of and require- Beschreibungen und Vorschriften rum Einsat; des LAL- ments for the LAL Bacterial Endotoxin Test. Both Tests fur die Bestimmung bakterieller Endotoxine. In pharmacopoeias have begun to remove the rabbit beiden Pharmakopoen wurde damit begonnen, die pyrogen test requirement in a majority of drug mono- erforderlichen Kaninchen-Pyrogen- Tests in Arzneimit- graphs and have substituted endotoxin limits to be tel-Monographien durch Endotoxin-Grenzwerte zu determined by LAL. The use of LAL has proved invalua- ersetzen, die mit dem LAL-Test bestimmt werden konnen. ble in controlling the level of endotoxin in finished Der Einsat: des LAL-Tests zur Endotoxin-Kontrolle von product. The endotoxin contribution of raw materials Endprodukten hat sich als unschdtzbar erwiesen. Eben- and packaging material can be monitored as well. In- sogut kann auch der Endotoxin-Gehalt in Ausgangsstof- process testing at critical production steps can identify fen und Verpackungsmaterialien bestimmt werden. Bei additional sources of endotoxin contamination, and der Kontrolle wdhrend kritischer Produktionsabldufe depyrogenation processes can be validated by quantita- kiinnen die Ursachen von Endotoxin-Kontaminationen ting the degradation of endotoxin challenges. The speed, gefunden und Entpyrogenisierungsprozesse durch die reproducibility, sensitivity, and economics of the Kinetic- Quantifizierung des Endotoxin-Abbaus validiert werden. QCL assay, in conjunction with the appropriate equip- Schnelligkeit, Reproduzierbarkeit, Empfindlichkeit und ment and software, over both the in vivo rabbit pyrogen Wirtschaftlichkeit, verbunden mit der entsprechenden test and the more traditional LAL gel-clot assay allow a Ausriistung und Software, machen aus dem kinetischen more in-depth approach to the control of endotoxin in QCL-Assay eine sowohl dem in vivo Kaninchen-Pyro- pharmaceuticals and medical devices. gentest als auch dem traditionellen Limulus-Geliertest uberlegene Methode, wenn es um die Kontrolle von Keywords: Limulus amebocyte lysat test, kinetic quanti- Endotoxinen in Pharmazeutika und medizinischen tative chromogenic LAL assay, endotoxins, pharma- Hilfsmitteln geht. copeias, pharmaceuticals, medical devices

1 Introduction ence of a heat-stable pyrogenic mate- the nature of this fever-inducing rial present in the injected fluid as substance suggested that it was of The association of pyrogenic reac- the causative agent. Subsequently, bacterial origin and exclusively asso- tions with the intravenous injection several investigators, Billroth ciated with Gram-negative bacteria. of fluids is over 200 years old. In (1862), Burdon-Sanderson (1876), Pfeiffer is reported to have proposed 1874, Panum postulated the exist- Jona and Centanni (1916), studying the term "endotoxin" to describe the

ALTEX 12,2/95 93 BERZOFSKY wr.:'\ ------h~·-- ~v~~.. membrane-associated toxin of Vibrio endotoxin. Endotoxin catalyzes the toxin activated clotting enzyme in (cited in Pearson, 1985). activation of a proenzyme contained place of the natural clotting protein. The concern over the presence of within the LAL (Young et aI., 1972). As with the gel-clot assay, the endotoxin in pharmaceuticals was The initial rate of activation is deter- amount of active clotting enzyme raised by the studies of Wechsel- mined by the concentration of endo- produced is proportional to the mann (1911), Muller (1911), Samel- toxin present. The activated enzyme amount of endotoxin in the sample. son (1913), and Bendix and Berg- cleaves the clotting protein, also However, instead of resulting in a man (1913). Hort and Penfold present in the LAL, resulting in the simple yes/no, qualitative, assay; the (1912a, 1912b, 1912c) expanded the formation of an insoluble clot. use of the chromogenic substrate findings that a Gram-negative bacte- results in an assay which allows for rial substance was responsible for the endotoxin concentration In pyrogenicity. These workers are 2.1 Gel-Clot Assay unknowns to be quantitated. credited also with designing the first The combination of LAL and the rabbit pyrogen assay. The most commonly employed LAL chromogenic substrate (Kinetic-QCL Subsequently, confirmatory obser- method is the gel-clot assay. This Reagent) provides a better under- vations by Seibert (1923, 1925), assay utilizes the entire endotoxin- standing of the endotoxin mediated Seibert and Mendel (1923), and Ra- mediated cascade in addition to the reaction (Figure 1). Solutions contai- demaker (1930, 1932) reinforced the clotting protein to product a gelati- ning known amounts of endotoxin concept that bacterial contamination nous clot after incubation with endo- ranging in concentration from 50 to was present in all pyrogenic injecta- toxin. Basically equal volumes of 0.005 EU/ml were combined with an bles, and that through careful tech- sample in LAL (typically 0.1 ml equal volume of the Kinetic-QCL nique to avoid bacterial contaminati- each) are combined in a 10 x 75 mm Reagent, and incubated at 37°C whi- on, non-pyrogenic injectables could glass tube. After an incubation peri- le the absorbance of the reaction at be produced. Their work lead to a od of 60 minutes at 37°C, the tubes 405 nm monitored. The graph illu- collaborative study, which standardi- are inverted 180°. A positive result is strates the time, on the X-axis, before zed the rabbit pyrogen test, and indicated by a clot, which withstands the creation of the active enzyme as allowed pyrogen testing to obtain the inversion. By titrating the lysate seen by the increase in absorbance, pharmacopeial recognition in 1942. with an endotoxin of known potency, on the Y-axis, at different endotoxin The rabbit pyrogen test remained the minimum concentration of endo- concentrations. The concentration of the exclusive official pyrogen assay, toxin required to yield a positive clot endotoxin is expressed in Endotoxin referenced in the U.S. Pharamcopeia can be determined. This minimum Units (EU) based on the current (USP) for over 25 years, and is still endotoxin concentration, or end- FDA/USP Reference Endotoxin referenced exclusively in many of point, is referred to as the lysate Standard, EC-S. With high concen- the USP drug monographs. Howe- sensitivity. trations of endotoxin, i.e. 50 EU/ml, ver, events occurring in the late The gel-clot assay can be used as a there is a short lag time followed by 1950s and early 1960s were to lead purely qualitative limits test to rank a rapid increase in absorbance and an to a significant in vitro alternative to samples as either positive or negati- eventual plateau when all the sub- the rabbit pyrogen test for determi- ve, i.e. greater than or less than the strate has been consumed. As the ning pyrogenicity in pharmaceuticals lysate sensitivity. However, by titra- endotoxin concentrations decrease, and medical devices. ting positive samples one can obtain the corresponding lag time signifi- a semi-quantitative measure of the cantly increase and the rates of color endotoxin concentration in un- generation become slower. 2 Limulus Amebocyte Lysate knowns by multiplying the last posi- This family of curves is different Assay tive sample dilution by the lysate from what one normally sees when sensitivity. The gel-clot assay suffers performing a kinetic enzyme analy- The basis for using Limulus amebo- from the disadvantage that they are sis, where the reaction begins imme- cyte lysate in the detection of endo- best semi-quantitative and require diately and different concentrations toxin lies entirely in the coagulation preparing multiple dilutions of sam- of analyte only cause different reac- reaction inherent in Limulus blood. ples to determine a positive/negative tion rates. It is important to keep in In 1956, Frederic Bang, reported that end-point. mind, that unlike "classical" kinetics, infecting the horseshoe crab, Limu- in the kinetic-quantitative chromoge- Ius polyphemus with Gram-negative nic LAL (Kinetic-QCL) assay the bacteria resulted in fatal intravascu- 2.2 Kinetic-QCL Assay enzyme does not exist but is created lar coagulation. By 1964, Levin and during the reaction by the activation Bang had demonstrated that extract A truly quantitative assay has been of endotoxin. of the circulating amebocytes would developed by substituting a synthetic In order to estimate the endotoxin gel in the presence of Gram-negative chromogenic substrate for the endo- concentration in unknown samples,

94 ALTEX 12,2/95 d'~"\ BERZOFSKY --~~~------~v~

it is necessary to develop a standard 2000 curve which correlates a unique reac- • • ••• • 1 a • A A A *• *A *• t tion parameter with a corresponding • • * * endotoxin concentration. It was deci- 6' 1500 • * 0 • * ded to focus on the time required for .s • * • • * the creation of the active enzyme. Q) o * This time, designated the Reaction c 1000 • • 0 ItS * 0 .0 0 Time, was arbitrarily defined as the •.. • * 0 0 • .01. 0 time when the absorbance of the (/) * .0 * 0 reaction mixture increased by 200

ALTEX 12.2/95 95 BERZOFSKY ~r.:"\ ------~~~-- ~~ "'~. 10000 successfully validated using the LAL assay. Slope = -0.204 The specific validation require- (jJ Y int. = 3.008 ments for analyst qualification using D c Corr. = -0.996 0 the Kinetic-QCL assay involve con- o Q) firming the linearity of a standard 2- curve generated using an endotoxin ill E 1000 preparation of known potency. The F c coefficient of correlation, r, determi- 0 f) • ned by linear regression must fall cu ill between -0.980 and -1.000. These IT: values are negative because of the inverse relationship between Reac- tion Time and endotoxin concentrati- 100 on. The product validation require- 0,5 50 0,005 0,05 5 ments using the Kinetic-QCL assay Endotoxin Concentration involve spiking a sample of the Figure 2: Kinetic-QCL assay, standard curve product with a known amount of Standard curve for Kinetic-QCL assay. The log endotoxin concentration is plotted endotoxin, typically 0.5 EU/ml, and, against the log reaction time. The curve can quantitate endotoxin in unknowns over a after accounting for any background concentration range of 50 to 0.005 EU/ml endotoxin contained in the unspiked sample, recovering +/- 50% of the spike pharmaceuticals and medical de- 3.2 Validation vices. Collectively, these documents outline the acceptable procedures for Validation of the LAL assay is com- References 1) determining endotoxin limits, posed of two parts, Bang, F. (1956). A bacterial disease of 2) validating the LAL tests and 1) qualifying the laboratory analyst Limulus polyphemus. Bull. Johns 3) developing a routine testing pro- and Hopkins Hosp. 98, 325-35l. tocol. 2) qualifying the specific product. Bendix, B. und Bergmann, 1. (1913). Analyst qualification involves con- Uber das sogenannte Kochsalzfie- firming the endotoxin sensitivity of ber. Monatsschr. Kinderheilkd. 11, 3.1 Endotoxin Limits the assay method using an endotoxin 387-404. preparation of known potency. Pro- Billroth, T. (1862). Beobachtungstudi- In the United States, endotoxin pot- duct qualification involves spiking en uber das Wundfieber und acci- encies are expressed in Endotoxin the product with known amounts of dentelle Wundkrankheiten. Arch. Units (EU) based on the current endotoxin and subsequently detec- Klin. Chir. 2, 578 (cited in Pearson, FDAfUSP Reference Endotoxin ting an acceptable percentage of the 1985). Standard, EC-5. In Europe, endoto- spike. This recovery experiment in- Burdon-Sanderson, J. (1876). On the xin potencies are expressed in Inter- dicates that the product does not process of fever. Part III. Pyrexia. national Units (IU) based on the adversely interfere with the LAL Practitioner, 417 (cited in Pearson, WHO Endotoxin Standard, 84-650. assay. Many pharmaceutical pro- 1985). Centanni, E. (1921). La immunita sto- For pharmaceuticals, endotoxin li- ducts interfere with the performance mits are based on the product's of LAL test. Guilfoyle and Munson mogene (terza immunita). In Tratta- to di Immunologia, (149) Rome: human dose. A typical parenteral (1982) reported that of 587 pharma- Societa Endrice Libraria (cited in solution would have an endotoxin ceuticals tested at use concentration, Pearson, 1985). limit of 5 units of endotoxin per 78% interfered with the LAL assay Friberger, P., Knos, M. and Mellstam, kilogram body dose of the product. in some manner. Interference may be L. (1982). A quantitative endotoxin Alternative procedures exist for in- defined as the inability to recover an assay utilizing LAL and a chromo- trathecals, radiopharmaceuticals and endotoxin spiked within acceptable genic substrate. In S. W. Watson, J. anti-neoplastics. Medical devices are limits. Enhancement, on the other Levin and T. 1. Novitsky (eds.), typically tested by extracting the hand, is characterized by the appa- Endotoxins and their Detection with device with 40 mls of an appropriate rent recovery of more than the the Limulus Amebocyte Lysate Test extraction solution. Each ml of the known endotoxin spike. However, (195-206). New York: Alan R. Liss. extraction fluid must contain less the authors demonstrated that all of Guilfoyle, D. E. and Munson, T. E. than 0.5 units of endotoxin. these products could be pre-treated, (1982). Procedures for improving most notably by dilution alone, and detection of endotoxin in products

96 ALTEX 12,2/95 wl-'\ BERZOFSKY --~~~------~~

found incompatible for direct analy- Seibert, F. B. and Mendel, L. B. Address sis with Limulus amebocyte lysate. (1923). Protein fevers, with special Ronald N. Berzofsky, Ph. D. In S. W. Watson, J. Levin and T. J. reference to casein. Am. 1. Physiol. BioWhittaker, Inc. Novitsky (eds.), Endotoxins and 67, 105-123. 8830 Biggs Ford Road their Detection with the Limulus Wechselmann, W. (1911). Neuere Er- Walkersville, MD 21793-0127, USA Lysate Test (79-90). New York: fahrungen uber intravenose Salvar- Alan R. Liss. san-Injektionen ohne Reaktionser- Hort, E. C. and Penfold, W. J. (l912a). scheinungen. Munch. Med. Wochen- The relation of salvarsan fever to schr. 58, 1510-1511. other forms of infection fever. Proc. R. Soc. Med. (Part III) 5 (Path. Sect.),131-139. Hort, E. C. and Penfold, W. J. (1912b). Microorganisms and their relation to fever. 1. Hyg. 12,361-390. Hort, E. C. and Penfold, W. J. (1912c). A critical study of experimental fe- ver. Proc R. Soc. London, Ser. B 85, 174-186. Jona, 1. L. (1916). A contribution to the experimental study of fever. 1. Hyg. 15, 169 (cited in Pearson, 1985). Levin, J. and Bang, F. B. (1964). The role of endotoxin in the extracellular coagulation of Limulus blood. Bull. Johns Hopkins Hosp. 115,265-274. MUller, P. T. (1911). Uber den Bakte- riengehalt des in Apotheken erhaltli- chen destillierten Wassers. Munch. Richtigstellung: Med. Wochenschr. 58,2739-2740. Panum, P. L. (1874). Das putride Gift, Die Prasidenten der Schweizerischen Akademie der Medizinischen die Bakterien, die putride Infektion Wissenschaften (SAMW; Prof. A. F. Muller) und der Schweizerischen oder Intoxikation und die Septik- Akademie der Naturwissenschaften (SANW, Prof. B. Hauck), weisen amie. Arch. Pathol. Anat. Pysiol. Klin. Med. 60, 301 (cited in Pearson, uns auf einen Irrtum bei der Besprechung der Ethischen Grundsatze 1985). und Richtlinien (EGR) in ALTEX 1114, 1994, S. 220ff hin. Pearson, F. C. III (1985). Pyrogens, Endotoxins, LAL Testing, and De- Da war zu lesen: pyrogenation. In J. R. Robinson (Hrsg.), Advances in Parenteral Sci- "Sie (die EGR) sind fi.ir Mitglieder der beiden Akademien freiwillig ences 2. New York: Marcel Dekker. und unverbindlich. Die Industrie muf diese Empfehlungen nicht Rademaker, L. (1930). The cause and berucksichtigen. " elimination of reactions after intra- venous infusions. Ann. Surg. 92, Richtig miisse es heiBen: 195-201. Rademaker, L. (1933). Reactions after Die SAMW und die SANW haben die EGR bereits 1983 als Kodex fur intravenous infusion. A further re- aIle in der Schweiz tatigen Wissenschaftler und Wissenschaftlerinnen port on their elimination. Surg. und deren Mitarbeiter und Mitarbeiterinnen verbindlich erklart (1). Gynecol. Obstet. 56, 956-958. Auch die chemische Industrie der Schweiz steht uneingeschrankt Samelson, S. (1913). Uber das soge- hinter diesen Richtlinien und erachtet sie als Basis fur ihre Tatigkeit nannte Kochsalzfieber. Monatsschr. (2). Kinderheilkd. 11, 125-133. Seibert, F. B. (1923). Fever-producing (1) Ethische Grundsatze und Richtlinien fur wissenschaftliche Tier- substance found in some distilld versuche (1994). Schweizerische Arstezeitung 75, Heft 33, 1255- waters, Am. 1. Physiol. 67, 90-104. 1263. Seibert, F. B. (1925). The cause of (2) Tierversuche sind notwendig (1992). Eine Dokumentation des many febrile reactions following in- Arbeitskreises Gesundheit und Forschung, Zi.irich, Seite 21. travenous injections. Am. 1. Physiol. 71,621-652.

ALTEX 12,2/95 97