Endotoxin and other pyrogenic contaminants

Pyrogens are substances that, when they enter the blood stream, influence hypothalamic regulation of body temperature, usually resulting in fever. Medical control of pyrogen-induced fever proves very difficult, and in severe cases results in patient death.

Pyrogens represent a diverse group of substances, including various chemicals, particulate matter and endotoxin (LPS), a molecule derived from the outer membrane of Gram-negative bacteria. Such Gram-negative organisms harbour 3-4 million LPS molecules on their surface, representing in the region of 75 per cent of their outer membrane surface area. Gram-negative bacteria




Figure 7.6 Photograph of a modern protein sequencing system. Photograph courtesy of Perkin Elmer Applied Biosystems Ltd, UK

clinically significant in human medicine include E. coli, Haemophilus influenzae, Salmonella enterica, Klebsiella pneumoniae, Bordetella pertussis, Pseudomonas aeruginosa, Chylamydia psittaci and Legionella pneumophila.

In many instances the influence of pyrogens on body temperature is indirect. For example, entry of endotoxin into the bloodstream stimulates the production of IL-1 (Chapter 9) by macrophages. It is the IL-1 that directly initiates the fever response (hence its alternative name, 'endogenous pyrogen').

Although entry of any pyrogenic substance into the bloodstream can have serious medical consequences, endotoxin receives most attention because of its ubiquitous nature. Therefore, it is the pyrogen most likely to contaminate parenteral (bio)pharmaceutical products. Effective implementation of GMP minimizes the likelihood of product contamination by pyrogens. For example, GMP dictates that chemical reagents used in the manufacture of process buffers be extremely pure. Such raw materials, therefore, are unlikely to contain chemical contaminants displaying pyrogenic activity. Furthermore, GMP encourages filtration of virtually all parenteral products through a 0.45 or 0.22 ^m filter at points during processing and prior to filling in final product containers (even if the product can subsequently be sterilized by autoclaving). Filtration ensures removal of all particulate matter from the product. In addition, most final product containers are rendered particle free immediately prior to filling by an automatic pre-rinse using WFI. As an additional safeguard, the final product will usually be subject to a particulate matter test by QC before final product release. The simplest format for such a test could involve visual inspection of vial contents, although specific particle detecting and counting equipment is more routinely used.

Contamination of the final product with endotoxin is more difficult to control because:

• Many recombinant biopharmaceuticals are produced in Gram-negative bacterial systems; thus, the product source is also a source of endotoxin.

• Despite rigorous implementation of GMP, most biopharmaceutical preparations will be contaminated with low levels of Gram-negative bacteria at some stage of manufacture. These bacteria shed endotoxin into the product stream, which is not removed during subsequent bacterial filtration steps. This is one of many reasons why GMP dictates that the level of bioburden in the product stream should be minimized at all stages of manufacture.

• The heat stability exhibited by endotoxin (see Section 7.6.1) means that autoclaving of process equipment will not destroy endotoxin present on such equipment.

• Adverse medical reactions caused by endotoxin are witnessed in humans at dosage rates as low as 0.5 ng per kilogram body weight.

7.6.1 Endotoxin, the molecule

The structural detail of a generalized endotoxin (LPS) molecule is presented in Figure 7.7. As its name suggests, LPS consists of a complex polysaccharide component linked to a lipid (lipid A) moiety. The polysaccharide moiety is generally composed of 50 or more monosaccharide units linked by glycosidic bonds. Sugar moieties often found in LPS include glucose, glucosamine, mannose and galactose, as well as more extensive structures such as L-glycero-mannoheptose. The polysaccharide component of LPS may be divided into several structural domains. The inner (core) domains vary relatively little between LPS molecules isolated from different Gram-negative bacteria. The outer (O-specific) domain is usually bacterial-strain specific.

Most of the LPS biological activity (pyrogenicity) is associated with its lipid A moiety. This usually consists of six or more fatty acids attached directly to sugars such as glucosamine. Again, as is the case in relation to the carbohydrate component, lipid A moieties of LPS isolated from different bacteria can vary somewhat. The structure of E coli's lipid A has been studied in the greatest detail; its exact structure has been elucidated and it can be chemically synthesized.

7.6.2 Pyrogen detection

Pyrogens may be detected in parenteral preparations (or other substances) by a number of methods. Two such methods are widely employed in the pharmaceutical industry.

Historically, the rabbit pyrogen test constituted the most widely used method. This entails parenteral administration of the product to a group of healthy rabbits, with subsequent monitoring of rabbit temperature using rectal probes. Increased rabbit temperature above a certain point suggests the presence of pyrogenic substances. The basic rabbit method, as outlined in the European Pharmacopoeia, entails initial administration of the product to three rabbits. The product is considered to have passed the test if the total (summed) increase of the temperature of all three animals is less than 1.15 °C. If the total increase recorded is greater than 2.65 °C then the product has failed. However, if the response observed falls between these two limits

Figure 7.7 Structure of a generalized LPS molecule. LPS constitutes the major structural component of the outer membrane of Gram-negative bacteria. Although LPSs of different Gram-negative organisms differ in their chemical structure, each consists of a complex polysaccharide component, linked to a lipid component. Refer to text for specific details y Polysaccharide

' component

Lipid A component

Figure 7.7 Structure of a generalized LPS molecule. LPS constitutes the major structural component of the outer membrane of Gram-negative bacteria. Although LPSs of different Gram-negative organisms differ in their chemical structure, each consists of a complex polysaccharide component, linked to a lipid component. Refer to text for specific details the result is considered inconclusive, and the test must be repeated using a further batch of animals.

This test is popular because it detects a wide spectrum of pyrogenic substances. However, it is also subject to a number of disadvantages, including:

• it is expensive (there is a requirement for animals, animal facilities and animal technicians);

excitation/poor handling of the rabbits can affect the results obtained, usually prompting a false positive result;

• subclinical infection/poor overall animal health can also lead to false positive results;

• use of different rabbit colonies/breeds can yield variable results.

Another issue of relevance is that certain biopharmaceuticals (e.g. cytokines such as 1L-1 and TNF; Chapter 9) themselves induce a natural pyrogenic response. This rules out use of the rabbit-based assay for detection of exogenous pyrogens in such products. Such difficulties have led to the increased use of an in vitro assay; the Limulus ameobocyte lysate (LAL) test. This is based upon endotoxin-stimulated coagulation of amoebocyte lysate obtained from horseshoe crabs. This test is now the most widely used assay for the detection of endotoxins in biopharmaceutical and other pharmaceutical preparations.

Development of the LAL assay was based upon the observation that the presence of Gramnegative bacteria in the vascular system of the American horseshoe crab, Limulus polyphemus, resulted in the clotting of its blood. Tests on fractionated blood showed that the factor responsible for coagulation resided within the crab's circulating blood cells, i.e. the amoebocytes. Further research revealed that the bacterial agent responsible of initiation of clot formation was endotoxin.

The endotoxin molecule activates a coagulation cascade quite similar in design to the mammalian blood coagulation cascade (Figure 7.8). Activation of the cascade also requires the presence of divalent cations such as calcium or magnesium. The final steps of this pathway entail the proteolytic cleavage of the polypeptide coagulogen, forming coagulin, and a smaller peptide fragment. Coagulin molecules then interact non-covalently, forming a 'clot' or 'gel'.

The LAL-based assay for endotoxin became commercially available in the 1970s. The LAL reagent is prepared by extraction of blood from the horseshoe crab, followed by isolation of its amoebocytes by centrifugation. After a washing step, the amoebocytes are lysed and the lysate dispensed into pyrogen-free vials. The assay is normally performed by making a series of 1:2 dilutions of the test sample using (pyrogen-free) WFI (and pyrogen-free test tubes; see later). A reference standard endotoxin preparation is treated similarly. LAL reagent is added to all tubes, incubated for 1 h, and these tubes are then inverted to test for gel (i.e. clot) formation, which would indicate presence of endotoxin.

More recently, a colorimetric-based LAL procedure has been devised. This entails addition to the LAL reagent of a short peptide, susceptible to hydrolysis by the LAL clotting enzyme. This synthetic peptide contains a chromogenic tag (usually para-nitroaniline, pNA) which is released free into solution by the clotting enzyme. This allows spectrophotometric analysis of the test sample, facilitating more accurate end-point determination.

The LAL system displays several advantages when compared with the rabbit test, most notably:

• sensitivity - endotoxin levels as low as a few picograms per millilitre of sample assayed will be detected;

• cost - the assay is far less expensive than the rabbit assay;

• speed - depending upon the format used, the LAL assay may be conducted within 15-60 min.


Factor C Active Factor C

Factor B Active Factor B

Factor A

Active Factor A


Coagulin & peptide C



Figure 7.8 Activation of clot formation by endotoxin. The presence of endotoxin causes stepwise, sequential activation of various clotting factors present naturally within the amoebocytes of the American horseshoe crab. The net result is the generation of the polypeptide fragment coagulin, which polymerizes, thus forming a gel or clot

Its major disadvantage is its selectivity: it only detects endotoxin-based pyrogens. In practice, however, endotoxin represents the pyrogen that is by far the most likely to be present in pharmaceutical products. The LAL method is used extensively within the industry. It is used not only to detect endotoxin in finished parenteral preparations, but also in WFI and in biological fluids, such as serum or cerebrospinal fluid.

Before the LAL assay is routinely used to detect/quantify endotoxin in any product, its effective functioning in the presence of that product must be demonstrated by validation studies. Such studies are required to prove that the product (or, more likely, excipients present in the product) do not interfere with the rate/extent of clot formation (i.e. are neither inhibitors nor activators of the LAL-based enzymes). LAL enzyme inhibition could facilitate false-negative results upon sample assay. Validation studies entail, for example, observing the effect of spiking endotoxin-negative product with know quantities of endotoxin, or spiking endotoxin with varying quantities of product, before assay with the LAL reagents.

All ancillary reagents used in the LAL assay system (e.g. WFI, test tubes, pipette tips for liquid transfer, etc.) must obviously be endotoxin free. Such items can be rendered endotoxin free by heat. Its heat-stable nature, however, renders very vigorous heating necessary in order to destroy contaminant endotoxin. A single autoclave cycle is insufficient, with total destruction requiring three consecutive autoclave cycles. Dry heat may also be used (180 °C for 3 h or 240 °C for 1 h).

GMP requires that, where practicable, process equipment coming into direct contact with the biopharmaceutical product stream should be rendered endotoxin free (depyrogenated) before use. Autoclaving, steam or dry heat can effectively be used on many process vessels, pipework, etc., which are usually manufactured from stainless steel or other heat-resistant material. Such an approach is not routinely practicable in the case of some items of process equipment, such as chromatographic systems. Fortunately, endotoxin is sensitive to strongly alkaline conditions; thus, routine cleaning in place of chromatographic systems using 1 mol l_1 NaOH represents an effective depyrogenation step. Gentler approaches, such as exhaustive rinsing with WFI (until an LAL test shows the eluate to be endotoxin free), can also be surprisingly effective.

It is generally unnecessary to introduce specific measures aimed at endotoxin removal from the product during downstream processing. Endotoxin present in the earlier stages of production is often effectively removed from the product during chromatographic fractionation. The endotoxin molecule's highly negative charge often facilitates its effective removal from the product stream by ion-exchange chromatography. Gel-filtration chromatography also serves to remove endotoxin from the product. Although individual LPS molecules exhibit an average molecular mass of less than 20 kDa, these molecules aggregate in aqueous environments and generate supramolecular structures of molecular mass 100-1000 kDa.

The molecular mass of most biopharmaceuticals is considerably less than 100 kDa (Table 7.4). The proteins would thus elute from gel-filtration columns much later than contaminating endo-toxin aggregates. Should the biopharmaceutical exhibit a molecular mass approaching or exceeding 100 kDa, then effective separation can still be achieved by inclusion of a chelating agent such as EDTA in the running buffer. This promotes depolymerization of the endotoxin aggregates into monomeric (20 kDa) form.

Additional techniques capable of separating biomolecules on the basis of molecular mass (e.g. ultrafiltration) may also be used to remove endotoxin from the product stream.

The clinical significance of DNA-based contaminants in biopharmaceutical products remains unclear. The concerns relating to the presence of DNA in modern biopharmaceuticals focus

Table 7.4 The molecular mass of some polypeptide biopharmaceuticals. Many are glycosylated, thereby exhibiting a range of molecular masses due to differential glycosylation


Molecular mass (kDa)


Molecular mass (kDa)


Molecular mass (kDa)





































"Biologically active, trimeric form.

TPO: thrombopoietin; EGF: epidermal growth factor; NGF: nerve growth factor; LH: luteinizing hormone.

"Biologically active, trimeric form.

TPO: thrombopoietin; EGF: epidermal growth factor; NGF: nerve growth factor; LH: luteinizing hormone.

primarily upon the presence of active oncogenes in the genome of several producer cell types (e.g. monoclonal antibody production in hybridoma cell lines). Parenteral administration of DNA contaminants containing active oncogenes to patients is considered undesirable. The concern is that uptake and expression of such DNA in human cells could occur. There is some evidence to suggest that naked DNA can be assimilated by some cells at least, under certain conditions (Chapter 14). Guidelines to date state that an acceptable level of residual DNA in recombinant products is of the order of 10 pg per therapeutic dose.

DNA hybridization studies (e.g. the 'dot blot' assay) utilizing radiolabelled DNA probes allows detection of DNA contaminants in the product, to levels in the nanogram range. The process begins with isolation of the contaminating DNA from the product. This can be achieved, for example, by phenol and chloroform extraction and ethanol precipitation. The isolated DNA is then applied as a spot (i.e. a 'dot') onto nitrocellulose filter paper, with subsequent baking of the filter at 80 °C under vacuum. This promotes (a) DNA denaturation, yielding single strands, and (b) binding of the DNA to the filter.

A sample of total DNA derived from the cells in which the product is produced is then radiolabelled with 32P using the process of nick translation. It is heated to 90 °C (promotes denaturation, forming single strands) and incubated with the baked filter for several hours at 40 °C. Lowering the temperature allows reannealing of single strands via complementary base-pairing to occur. Labelled DNA will reanneal with any complementary DNA strands immobilized on the filter. After the filter is washed (to remove non-specifically bound radiolabelled probe) it is subjected to autoradiography, which allows detection of any bound probe.

Quantification of the DNA isolated from the product involves concurrent inclusion in the dot blot assay of a set of spots, containing known quantities of DNA, and being derived from the producer cell. After autoradiography, the intensity of the test spot is compared with the standards.

In many instances there is little need to incorporate specific DNA removal steps during downstream processing. Endogenous nucleases liberated upon cellular homogenization come into direct contact with cellular DNA, resulting in its degradation. Commercial DNase's are sometimes added to crude homogenate to reduce DNA-associated product viscosity (Chapter 6). Most chro-matographic steps are also effective in separating DNA from the product stream. Ion-exchange chromatography is particularly effective, as DNA exhibits a large overall negative charge (due to the phosphate constituent of its nucleotide backbone; Chapter 3).

7.6.4 Microbial and viral contaminants

Finished-product biopharmaceuticals, along with other pharmaceuticals intended for parenteral administration, must be sterile (the one exception being live bacterial vaccines). The presence of microorganisms in the final product is unacceptable for a number of reasons:

• Parenteral administration of contaminated product would likely lead to the establishment of a severe infection in the recipient patient.

• Microorganisms may be capable of metabolizing the product itself, thus reducing its potency. This is particularly true of protein-based biopharmaceuticals, as most microbes produce an array of extracellular proteases.

• Microbial-derived substances secreted into the product could adversely affect the recipient's health. Examples include endotoxin secreted from Gram-negative bacteria, or microbial proteins that would stimulate an immune response.

Terminal sterilization by autoclaving guarantees product sterility. Heat sterilization, however, is not a viable option in the case of biopharmaceuticals. Sterilization of biopharmaceuticals by filtration, followed by aseptic filling into a sterile final-product container, inherently carries a greater risk of product contamination. Finished-product sterility testing of such preparations thus represents one of the most critical product tests undertaken by QC. Specific guidelines relating to sterility testing of finished products are given in international pharmacopoeias.

Biopharmaceutical products are also subjected to screening for the presence of viral particles prior to final product release. Although viruses could be introduced, for example, via infected personnel during downstream processing, proper implementation of GMP minimizes such risk. Any viral particles found in the finished product are most likely derived from raw material sources. Examples could include HIV or hepatitis viruses present in blood used in the manufacture of blood products. Such raw materials must be screened before processing for the presence of likely viral contaminants.

A variety of murine (mouse) and other mammalian cell lines have become popular host systems for the production of recombinant human biopharmaceuticals. Moreover, most monoclonal antibodies used for therapeutic purposes are produced by murine-derived hybridoma cells. These cell lines are sensitive to infection by various viral particles. Producer cell lines are screened during product development studies to ensure freedom from a variety of pathogenic advantageous agents, including various species of bacteria, fungi, yeast, mycoplasma, protozoa, parasites, viruses and prions. Suitable microbiological precautions must subsequently be undertaken to prevent producer cell banks from becoming contaminated with such pathogens.

Removal of viruses from the product stream can be achieved in a number of ways. The physi-cochemical properties of viral particles differ greatly from most proteins, ensuring that effective fractionation is automatically achieved by most chromatographic techniques. Gel-filtration chro-matography, for example, effectively separates viral particles from most proteins on the basis of differences in size.

In addition to chromatographic separation, downstream processing steps may be undertaken that are specifically aimed at removal or inactivation of viral particles potentially present in the product stream. Significantly, many are 'blanket' procedures, equally capable of removing known or potentially likely viral contaminants and any uncharacterized/undetected viruses. Filtration through a 0.22 ^m filter effectively removes microbial agents from the product stream, but fails to remove most viral types. Repeat filtration through a 0.1 ^m filter is more effective in this regard. Alternatively, incorporation of an ultrafiltration step (preferably at the terminal stages of downstream processing) also proves effective.

Incorporation of downstream processing steps known to inactivate a wide variety of viral types provides further assurance that the final product is unlikely to harbour active virus. Heating and irradiation are amongst the two most popular such approaches. Heating the product to between 40 and 60°C for several hours inactivates a broad range of viruses. Many biopharmaceuticals can be heated to such temperatures without being denatured themselves. Such an approach has been used extensively to inactivate blood-borne viruses in blood products. Exposure of product to controlled levels of UV radiation can also be quite effective, while having no adverse effect on the product itself.

7.6.5 Viral assays

A range of assay techniques may be used to detect and quantify viral contaminants in both raw materials and finished-biopharmaceutical products. No generic assay exists that is capable of detecting all viral types potentially present in a given sample. Viral assays currently available will detect only a specific virus, or at best a family of closely related viruses. The strategy adopted, therefore, usually entails screening product for viral particles known to be capable of infecting the biopharmaceutical source material. Such assays will not normally detect newly evolved viral strains, or uncharacterized/unknown viral contaminants. This fact underlines the importance of including at least one step in downstream processing that is likely to inactivate or remove viruses indiscriminately from the product. This acts as a safety net.

Current viral assays fall into one of three categories:

• immunoassays;

• assays based on viral DNA probes;

Generation of antibodies that can recognize and bind to specific viruses is straightforward. A sample of live or attenuated virus, or a purified component of the viral caspid, can be injected into animals to stimulate polyclonal antibody production (or to facilitate monoclonal antibody production by hybridoma technology). Harvested antibodies are then employed to develop specific immunoassays that can be used to screen test samples routinely for the presence of that specific virus. Immunoassays capable of detecting a wide range of viruses are available commercially. The sensitivity, ease, speed and relative inexpensiveness of these assays render them particularly attractive.

An alternative assay format entails the use of virus-specific DNA probes. These can be used to screen the biopharmaceutical product for the presence of viral DNA. The assay strategy is similar to the dot blot assays used to detect host-cell-derived DNA contaminants, as discussed earlier.

Viral bioassays of various different formats have also been developed. One format entails incubation of the final product with cell lines sensitive to a range of viruses. The cells are subsequently monitored for cytopathic effects or other obvious signs of viral infection.

A range of mouse-, rabbit- or hamster-antibody production tests may also be undertaken. These bioassays entail administration of the product to a test animal. Any viral agents present will elicit production of antiviral antibodies in that animal. Serum samples (withdrawn from the animal approximately 4 weeks after product administration) are screened for the presence of antibodies recognizing a range of viral antigens. This can be achieved by enzyme immu-noassay, in which immobilized antigen is used to screen for the virus-specific antibodies. These assay systems are extremely sensitive, as minute quantities of viral antigen will elicit strong antibody production. A single serum sample can also be screened for antibodies specific to a wide range of viral particles. Time and expense factors, however, militate against this particular assay format.

7.6.6 Miscellaneous contaminants

In addition to those already discussed, biopharmaceutical products may harbour other contaminants, some of which may be intentionally added to the product stream during the initial stages of downstream processing. Examples could include buffer components, precipitants (ethanol or other solvents, salts, etc.), proteolytic inhibitors, glycerol, anti-foam agents, etc. In addition to these, other contaminants may enter the product during downstream processing in a less controlled way. Examples could include metal ions leached from product-holding tanks/pipework, or breakdown products leaking from chromatographic media. The final product containers must also be chosen carefully. They must be chemically inert and be of suitable quality to eliminate the possibility of leaching of any substance from the container during product storage. For this reason, high-quality glass vials are often used.

In some instances it may be necessary to demonstrate that all traces of specific contaminants have been removed prior to final product filling. This would be true, for example, of many proteo-lytic inhibitors added during the initial stages of downstream processing to prevent proteolysis by endogenous proteases. Some such inhibitors may be inherently toxic, and many could (inappropriately) inhibit endogenous proteases of the recipient patient.

Demonstration of absence from the product of breakdown products from chromatographic columns may be necessary in certain instances. This is particularly true with regard to some affinity chromatography columns. Various chemical-coupling methods may be used to attach affinity ligands to the chromatographic support material. Some such procedures entail the use of toxic reagents, which, if not entirely removed after coupling, could leach into the product. In some cases ligands can also subsequently leach from the columns, particularly after sustained usage or overvigorous sanitation procedures. Improvements in the chemical stability of modern chromatographic media, however, have reduced such difficulties, and most manufacturers have carried out extensive validation studies regarding the stability of their product.

Sophisticated analytical methodologies facilitate detection of vanishingly low levels of many contaminants in biopharmaceutical preparations. The possibility exists, however, that uncharac-terized contaminants may persist, remaining undetected in the final product. As an additional safety measure, finished products are often subjected to 'abnormal toxicity' or 'general safety' tests. Standardized protocols for such tests are outlined in various international pharmacopoeias. These normally entail parenteral administration of the product to at least five healthy mice. The animals are placed under observation for 48 h and should exhibit no ill effects (other than expected symptoms). The death or illness of one or more animals signals a requirement for further investigation, usually using a larger number of animals. Such toxicity testing represents a safety net, designed to expose any unexpected activities in the product that could compromise the health of the recipient.

7.6.7 Validation studies

Validation can be defined as 'the act of proving that any procedure, process, equipment, material, activity or system leads to the expected results'. Routine and adequate validation studies form a core principle of GMP as applied to (bio)pharmaceutical manufacture, as such studies help assure the overall safety of the finished product (Box 7.2).

All validation procedures must be carefully designed and fully documented in written format (Box 7.2). The results of all validation studies undertaken must also be documented, and retained in the plant files. As part of their routine inspection of manufacturing facilities, regulatory personnel will usually inspect a sample of these records, to ensure conformance to GMP.

Validation studies encompass all aspects of (bio)pharmaceutical manufacture. All new items of equipment must be validated before being routinely used. Initial validation studies should be

Diabetes 2

Diabetes 2

Diabetes is a disease that affects the way your body uses food. Normally, your body converts sugars, starches and other foods into a form of sugar called glucose. Your body uses glucose for fuel. The cells receive the glucose through the bloodstream. They then use insulin a hormone made by the pancreas to absorb the glucose, convert it into energy, and either use it or store it for later use. Learn more...

Get My Free Ebook

Post a comment