Endotoxin testing for peptides: LAL, recombinant factor C, limits and interference

A peptide can pass HPLC purity at 99%, pass a sterility test, and still be unsuitable for a cell assay because it carries bacterial endotoxin. Endotoxin is not a living contaminant and not a peptide impurity: it is a lipid-sugar molecule shed by Gram-negative bacteria, and it is measured with its own test. This guide explains what that test is, how the LAL, recombinant factor C and monocyte activation methods differ, how endotoxin limits and dilutions are calculated, why peptide samples interfere, and where sterility and bioburden testing fit.

What are endotoxins, and why do they matter for peptides?

Bacterial endotoxins are lipopolysaccharides (LPS). A typical LPS has a hydrophobic domain called lipid A (the endotoxic part), a non-repeating core oligosaccharide and a distal O-antigen polysaccharide [1]. In animal cells, the plasma membrane protein TLR4 is the lipid A signalling receptor [1], which is why tiny amounts of LPS switch on innate immune cells.

That matters for research peptides in two ways:

• Safety of parenteral products. Endotoxin is a pyrogen, and testing for it is a standard release test for injectable medicines. EMA lists “bacterial endotoxins” among the possible specification tests for a synthetic peptide active substance [20].
• Reliability of in vitro data. Schwarz and colleagues showed that primary human CD1c+ dendritic cells can be activated by “minimal amounts of LPS”, at levels they detected in some commercial recombinant proteins, and warned that “even low endotoxin contamination may generate erroneous data” in work with LPS-sensitive cells [15]. That was an in vitro study on recombinant proteins, but the principle applies to any material added to immune-cell cultures, including synthetic peptides.

No bacteria are used to make a synthetic peptide, but endotoxin can still arrive with materials and process steps: FDA asks manufacturers to consider contamination in raw materials, in-process materials and the finished product, and notes that many firms monitor incoming ingredients, including processing water [4]. In a 2024 test-purchase study of lyophilized semaglutide vials sold by illegal online pharmacies, all three vials were free of viable micro-organisms, but the authors reported endotoxin in all of them, at levels they gave as 2.1645 to 8.9511 EU/mg; two of the three results were upper bounds (below 2.8658 and below 2.1645 EU/mg) [21].

Endotoxin, pyrogen, sterility and bioburden: which test answers which question?

These four tests are often confused on certificates. They measure different things and none substitutes for another.

Test	Question	Detects	Typical method	Ref
Bacterial endotoxins test (BET)	How much Gram-negative endotoxin is present?	LPS (lipid A activity)	LAL gel-clot, turbidimetric or chromogenic; rFC	[2], [5]
Pyrogen test	Does the sample cause a fever response?	Endotoxin and non-endotoxin pyrogens	Monocyte activation test; historically the rabbit test	[4], [13]
Sterility test	Are viable micro-organisms present?	Living bacteria and fungi that grow in media	Membrane filtration or direct inoculation, 14 days	[18]
Bioburden (microbial enumeration)	How many viable micro-organisms are present in a non-sterile material?	Countable viable micro-organisms	Ph. Eur. 2.6.12 / USP <61> / JP 4.05	[19]

The key distinction is between living organisms and what they leave behind. Endotoxin “is found in the outer membrane of Gram-negative bacteria” [7] and remains after the bacteria die or are filtered out. A sterile-filtered solution can therefore fail the endotoxin test, and an endotoxin-free solution can fail sterility.

How does the LAL test work?

The Limulus amebocyte lysate (LAL) is an extract of horseshoe crab blood cells. It contains a cascade of serine protease zymogens: factor C, factor B and the proclotting enzyme, plus coagulogen, a clottable protein [8]. Endotoxin activates factor C, factor C activates factor B, and factor B converts the proclotting enzyme to the clotting enzyme, which gels coagulogen [8]. Factor C, “the primer of this coagulation cascade”, is the endotoxin sensor [7].

The harmonised pharmacopoeial text describes three techniques [2]:

Technique	Read-out	Result type	Notes from the harmonised text	Ref
Gel-clot limit test	Firm gel that stays in place when the tube is inverted through about 180°	Pass/fail at the limit	Incubation usually 37 ± 1 °C for 60 ± 2 minutes; the referee method in disputes	[2], [3]
Gel-clot quantitative (semi-quantitative)	Endpoint of a dilution series	Concentration from endpoint dilution × λ	Same incubation; replicate dilution series	[2]
Turbidimetric (end-point or kinetic)	Turbidity as the gel forms	Quantitative against a standard curve	Usually 37 ± 1 °C	[2]
Chromogenic (end-point or kinetic)	Colour released from a synthetic peptide-chromogen substrate	Quantitative against a standard curve	Usually 37 ± 1 °C	[2]

ICH Q4B Annex 14 confirms that Ph. Eur. 2.6.14, JP 4.01 and USP <85> “can be used as interchangeable in the ICH regions”, that “any of the three techniques can be used”, and that “in the event of doubt or dispute, the gel-clot limit test should be used to make the final decision” [3]. The three pharmacopoeias’ reference standards are interchangeable because each is calibrated against the WHO International Standard for Endotoxin [3], and one EU equals one International Unit of endotoxin [2].

For photometric methods, the standard curve must have an absolute correlation coefficient |r| of at least 0.980, and spiked endotoxin must be recovered at 50% to 200% for the sample to be considered free of interfering factors [2]. FDA notes that quantitative testing “may be preferable to limit testing” because it allows trending and earlier correction of excursions [4].

LAL vs rFC vs MAT: how do the methods compare?

	LAL (USP <85>, Ph. Eur. 2.6.14)	Recombinant factor C (Ph. Eur. 2.6.32, USP <86>)	Monocyte activation test (MAT)
What it uses	Horseshoe crab lysate (whole cascade)	Recombinant factor C based on the horseshoe crab gene sequence, fluorimetric detection	Human monocytic cells
What it detects	Endotoxin; also activated by beta-glucans via factor G	Endotoxin; no factor G, so no glucan pathway	Endotoxin and non-endotoxin pyrogens, including Gram-positive pyrogens
Animal source	Yes	No	No
Replaces	Rabbit pyrogen test, for endotoxin	LAL (as an alternative method)	Rabbit pyrogen test
Status	Harmonised compendial method (ICH Q4B Annex 14)	Ph. Eur. general chapter effective 1 January 2021; USP <86> scheduled official May 2025, an alternative to <85>	Named by EDQM as an example in vitro test once the rabbit test was removed
Refs	[2], [3]	[5], [6], [8]	[11], [13]

Why choose rFC?

Recombinant factor C was produced as a 132 kDa proenzyme that is activated by trace endotoxin [7]. Because it “does not contain glucan-sensitive factor G”, the rFC assay is not subject to the glucan false positives that affect LAL [8]. Maloney and colleagues also estimated that using rFC for water and other common manufacturing materials could cut the use of horseshoe-crab-derived reagents by 90% [8]. In an industry validation across several pharmaceutical products, Bolden and Smith found the end-point fluorescence rFC method “equivalent or superior to the compendia bacterial endotoxins test method” [9].

The regulatory position differs slightly by region. EDQM states that an rFC test “can be used in the same way as LAL-based methods, after demonstration of its fitness for use for the specific substance or product”, and does not itself need validation, although replacing an LAL method prescribed in a monograph counts as using an alternative method [5]. In the US, USP describes the methods in <86> as “alternatives to Chapter <85>” [6], and FDA says alternative methods should be validated under USP <1225> and shown to give equivalent or better results, with the compendial gel-clot method deciding disputes unless the monograph says otherwise [4].

Is the rabbit pyrogen test still used?

Less and less. FDA has accepted LAL “in lieu of the rabbit pyrogens test” for more than 30 years [4], but it still advises using the rabbit test where non-endotoxin pyrogens may be present or where interference cannot be overcome by dilution up to the MVD [4]. Europe has gone further. In June 2024 the Ph. Eur. Commission adopted 57 revised texts with the rabbit pyrogen test deleted, plus a new general chapter on Pyrogenicity (5.1.13), with implementation on 1 July 2025 [11]. It then decided to suppress the general chapter Pyrogens (2.6.8) itself, a decision in force from 1 January 2026 [12]. Users now choose “a suitable in vitro test (e.g. the monocyte-activation test)” based on a risk assessment [11]. An international validation study in 2005 found that monocyte-based pyrogen tests had a lower limit of detection than the rabbit test and, unlike the bacterial endotoxins test, could detect Gram-positive pyrogens [13].

How are endotoxin limits calculated? EU/mL, EU/mg and the MVD

Endotoxin limits are not part of the general test chapter. ICH Q4B notes that the harmonised texts “did not contain acceptance criteria”; limits are set in the application dossier or an individual monograph [3].

For injectable medicines, the harmonised text defines the endotoxin limit as K/M, where K is “a minimum pyrogenic dose of endotoxin per kg body mass (EU/kg)” and M is the maximum bolus dose of product per kilogram, or the maximum total dose in one hour for frequent or continuous administration [2]. FDA advises that limits be calculated with the USP or AAMI methods and that monograph limits be rechecked against current product strengths [4].

The limit can be expressed in different units, and the sample concentration in the dilution formula follows the same basis [2]:

Limit expressed as	Sample concentration expressed as	Typical use
EU/mg	mg/mL	Peptide powders, APIs
EU/mL	mL/mL	Solutions, water
EU/Unit	Units/mL	Products dosed in biological units
EU/mEq	mEq/mL	Electrolytes

The maximum valid dilution (MVD) is “the maximum allowable dilution of a sample solution at which the endotoxin limit can be determined” [2]:

MVD = (endotoxin limit × concentration of sample solution) / λ

where λ is the labelled lysate sensitivity in the gel-clot method, or the lowest point of the standard curve in photometric methods [2]. Dilution is the main tool for overcoming interference, so the MVD sets how far you can go. A more sensitive lysate (smaller λ) allows a greater dilution [2].

What endotoxin limit should a research peptide meet?

For research-use material there is no compendial limit, so the right question is what the experiment can tolerate. Schwarz and colleagues reported that most suppliers of recombinant proteins guarantee less than 1 EU, “approximately equivalent to 100 pg of E. coli LPS per microgram” of protein, and found that CD1c+ dendritic cells could be activated by LPS amounts equivalent to the contamination they detected in some commercial proteins [15]. For immune-cell, macrophage or in vivo work, ask for a quantitative endotoxin result in EU/mg on the batch certificate, not only “passes”, and check that the reported detection limit is below the level your assay can tolerate.

Sample preparation and interference: why peptide samples inhibit or enhance the test

The BET is an enzyme assay, and anything that changes enzyme activity, pH or the physical state of LPS can distort it. The harmonised text sets several preparation rules [2]:

• Depyrogenate glassware. “Commonly used minimum time and temperature settings are 30 minutes at 250 °C.” Plastics must be shown free of detectable endotoxin and non-interfering.
• Use water for BET for dissolving and diluting samples.
• Control pH. Adjust so that the lysate-sample mixture falls within the lysate’s specified range, “usually” for a sample solution with pH 6.0 to 8.0.
• Run the test for interfering factors. Spike the sample with standard endotoxin and confirm recovery: for gel-clot, the endpoint must fall between 0.5λ and 2λ; for photometric methods, recovery must be 50% to 200%.
• Dilute or treat. If the sample interferes, repeat at a greater dilution not exceeding the MVD, or remove the interference “by suitable treatment, such as filtration, neutralization, dialysis or heat treatment”, proving the treatment does not lose endotoxin.

FDA adds a practical rule for routine screening: once you know the lowest dilution that neutralises interference, screen just above it. Its example is a product with an MVD of 1:100 that inhibits at 1:10 but not 1:20, which “may be best” screened at 1:30 [4]. If a test fails below the MVD, it should be repeated at a greater dilution not exceeding the MVD; an out-of-specification result at the MVD that is not due to testing error means the lot should be rejected [4].

Beta-glucans and false positives

LAL can respond to substances other than endotoxin. Roslansky and Novitsky found all tested LAL reagents were “at least 1,000-fold more sensitive to endotoxin than to LRGs” (LAL-reactive glucans), but glucans such as curdlan also amplified the LAL response to endotoxin, depending on concentration [10]. Maloney and colleagues list peptidoglycan, some exotoxins, yeast mannans, bacterial dextrans and dithiols among LAL activators, and note that common glucan sources include fungi and yeast hydrolysate [8]. In the 1991 study, one LAL-reactive material was cellulosic, extracted from a hollow-fibre haemodialyser [10]. rFC avoids the glucan pathway entirely [8].

Low endotoxin recovery (LER)

The opposite problem is masking. Low endotoxin recovery describes “the inability of LAL assays, in some cases, to detect LPS due to a masking effect caused by interaction with formulation excipients”, thought to involve disruption of the supramolecular aggregates LPS forms in solution [14]. The mechanism “has not been fully determined”, and the review discusses the possibility that masked LPS remains biologically active [14]. FDA’s guidance asks firms to support sample storage and handling with data showing assayable endotoxin is stable, noting that purified endotoxin “might react differently from native sources” [4]. For a peptide formulated with excipients, a spiked hold-time study is one way to check that endotoxin stays detectable.

How do you remove endotoxins from peptides?

The honest answer is that prevention is much easier than removal. Petsch and Anspach concluded that “a general method for the removal of endotoxins from protein solutions is not available” [16]. Methods used for water, such as ultrafiltration, “have little effect on endotoxin levels in protein solutions” [16].

Approach	What the sources say	Limitation	Ref
Prevention	Water for BET, glassware depyrogenated (commonly 30 minutes at 250 °C), monitoring of incoming ingredients and processing water, in-process testing at points of risk	Needs discipline at every step	[2], [4]
Ion exchangers	Among the main techniques applied to protein solutions	Success varies by product	[16]
Affinity and endotoxin-selective adsorbers	Applied “with varying success”; tailor-made adsorber matrices discussed	Product-specific	[16]
Two-phase extraction, including aqueous micellar systems	Micellar systems give a mild environment and suit small-scale work	Scale and clean-up	[16], [17]
Adsorption, ultrafiltration and chromatography	Reviewed as economical routes with relatively high recoveries in some cases	No method works for every product	[17]
Ultrafiltration (as used for water)	“Little effect” on endotoxin in protein solutions	Not a general fix	[16]

Magalhães and colleagues reach the same conclusion in their review: efficient, cost-effective endotoxin removal “is challenging”, and more research is needed [17]. For a synthetic peptide, this is why endotoxin control belongs in the process, with in-process testing at points of risk, rather than in a clean-up step at the end. FDA describes exactly this approach: in-process testing at production areas at risk, monitoring of incoming ingredients and processing water, and in-process limits derived from the release specification [4].

Sterility testing for peptides

The sterility test is harmonised as Ph. Eur. 2.6.1, JP 4.06 and USP <71>, which ICH Q4B Annex 8 treats as interchangeable [18]. Its key features [18]:

• Two media. Fluid thioglycollate medium, primarily for anaerobic bacteria but also detecting aerobes, incubated at 30–35 °C; and soya-bean casein digest medium, for fungi and aerobic bacteria, incubated at 20–25 °C.
• Two techniques. Membrane filtration, using filters with a nominal pore size not greater than 0.45 µm, or direct inoculation of the media.
• Incubation for not less than 14 days, with growth promotion and method suitability tests to show the product does not inhibit growth.
• A statistical limit. The text itself warns that “a satisfactory result only indicates that no contaminating micro-organism has been found in the sample examined in the conditions of the test.”

Sterility testing is relevant to peptides supplied as sterile products, such as filled vials. It is not a test of a bulk powder’s general cleanliness; that is what bioburden testing is for.

Bioburden (microbial enumeration) testing

For non-sterile materials, including bulk peptide APIs, the relevant test is microbial enumeration, harmonised as Ph. Eur. 2.6.12, JP 4.05 and USP <61> [19]. It counts viable micro-organisms in the material. Like the endotoxin chapter, the harmonised text contains no acceptance criteria [19]; limits come from the specification. EMA lists “microbiological purity” alongside bacterial endotoxins as possible specification tests for a synthetic peptide [20].

How to read endotoxin results on a peptide COA

When you check the endotoxin line on a certificate of analysis, look for:

1. The method. LAL gel-clot, kinetic turbidimetric, kinetic chromogenic or rFC, and the chapter cited (USP <85>, Ph. Eur. 2.6.14, Ph. Eur. 2.6.32 or USP <86>).
2. A number with units. EU/mg for a powder. “Complies” without a value hides the detection limit.
3. A “less than” result that makes sense. ”< x EU/mg” means nothing was detected at the dilution used; x depends on λ and the dilution.
4. A limit that suits your use. A limit for an API destined for injectables is calculated from K/M; a research limit should reflect your assay’s sensitivity.
5. Evidence of a valid test. A positive product control (spike recovery within 50–200% for photometric methods) shows the sample did not mask endotoxin [2].
6. Sterility and bioburden reported separately, if they matter for your application.

For the analytical side of the same certificate (HPLC purity, identity and peptide content), see our guide to peptide purity testing. For the API case, see the semaglutide API COA guide.

How Winhycare Pharma supplies research peptides

Winhycare Pharma supplies research peptides from Hyderabad, India, as research-use-only material at purity of ≥99%, as stated in our product brochure, and offers custom peptide synthesis. Semaglutide, tirzepatide, retatrutide, cagrilintide and BPC-157 are also supplied as API to licensed manufacturers; see peptide APIs. If your work is endotoxin-sensitive, tell us the limit you need in EU/mg when you ask for a quote and request the batch Certificate of Analysis: request research peptides.

For laboratory research use only. Not for human or veterinary use.