Techniques for Quantitative LC–MS/MS Analysis of Protein Therapeutics: Advances in Enzyme Digestion and Immunocapture

847 Bioanalysis (2016) 8(8), 847–856 ISSN 1757-6180

Review part of 10.4155/bio.16.24 © 2016 Future Science Ltd

LC–MS/MS has been investigated to quantify protein therapeutics in biological matrices. The protein therapeutics is digested by an enzyme to generate surrogate peptide(s) before LC–MS/MS analysis. One challenge is isolating protein therapeutics in the presence of large number of endogenous proteins in biological matrices.

Immunocapture, in which a capture agent is used to preferentially bind the protein therapeutics over other proteins, is gaining traction. The protein therapeutics is eluted for digestion and LC–MS/MS analysis. One area of tremendous potential for immunocapture-LC–MS/MS is to obtain quantitative data where ligand-binding assay alone is not sufficient, for example, quantitation of antidrug antibody complexes. Herein, we present an overview of recent advance in enzyme digestion and immunocapture applicable to protein quantitation.

First draft submitted: 25 September 2015; Accepted for publication: 19 February 2016;

Published online: 23 March 2016 Keywords: acid hydrolysis • enzyme digestion • immunocapture • LC–MS/MS • protein quantitation • regulated bioanalysis

Ligand-binding assays (LBA) have tradition- ally been used to quantify protein therapeu- tics in support of drug discovery and devel- opment  [1,2]. The selection/detection of the protein therapeutic (or protein analyte) in a complex matrix (e.g., serum) is accomplished by the specific binding of the protein analyte to the capture antibody/detection antibody.

Recently, LC–MS/MS has been investigated as a complementary technique to quantify protein therapeutics in biological matrices because of its unique mass selectivity, as selec- tion/detection of protein analyte is accom- plished by its unique m/z ratio. Due to the limited sensitivity of analyzing intact protein by MS, unlike in LBA, the protein analyte is usually digested by an endoprotease such as trypsin to generate one or more surrogate peptides. These surrogate peptides are then analyzed by LC–MS/MS  [3–10]. For quanti- tative purpose, one surrogate peptide is used as the ‘surrogate’ of the protein. Therefore, it is important to generate a surrogate peptide that is unique to the protein. In the case of antibodies, it is preferable that the surrogate peptide is located in the complementarity determining region (CDR) of the protein.

Another challenge for analyzing protein therapeutics in biological matrices is separat- ing the protein analyte from a large number of proteins with similar physical–chemical properties. For LBA, this is accomplished by utilizing a highly selective capture anti- body. For LC–MS/MS, the traditional sample preparation methods such as SPE or liquid–liquid extraction usually are not suffi- ciently selective and likely result in significant loss of analyte. It is highly desirable to mini- mize the number of proteins going into the enzyme digestion mixture in order to reduce interference. One methodology reported is the differential precipitation by organic solvent in which the different solubility of pegylated proteins and nonpegylated pro- Techniques for quantitative LC–MS/MS analysis of protein therapeutics: advances in enzyme digestion and immunocapture

Eliza N Fung*,1, Peter Bryan2 & Alexander Kozhich3

1Research and Development, Bristol- Myers Squibb Company, 1 Squibb Drive,

New Brunswick, NJ 08943, USA 2B2S Consulting, Mendham, NJ 07945,

USA 3Research and Development, Bristol- Myers Squibb Company, Route

206/Province Line Road, Princeton, NJ 08543, USA *Author for correspondence:

Tel.: +1 732 227 7725 ngakiteliza.fung@bms.com For reprint orders, please contact reprints@future-science.com

848 Bioanalysis (2016) 8(8) Figure 1. A flowchart depicting the workflow of immunocapture. future science group

Review  Fung, Bryan & Kozhich teins in an organic solvent are exploited. For example,

Wu et al. [11]. reported that pegylated proteins were sol- ubilized in 0.1% formic acid in 2-propanol while other endogenous proteins were not. The serum samples were then treated with 0.1% formic acid in 2-propanol to precipitate the endogenous proteins. Another meth- odology for isolating the protein analyte is to precipi- tate the protein analyte and all other proteins with an organic solvent, for example, methanol, while leaving other endogenous components (e.g.,  small-molecule entities that are soluble in the organic solvent) in solu- tion. The precipitated proteins are then resuspended in digestion buffer for enzyme digestion [12,13]. It is noted that this method does not result in clean extracted samples for subsequent enzyme digestion.

Another methodology that is gaining traction is immunocapture (or immunoaffinity capture), in which a capture agent, usually an antibody, is used to capture the protein analyte, or the surrogate peptide after enzyme digestion, essentially augmenting the selectivity of LC–MS/MS with the orthogonal selec- tivity of ligand-binding assay. The analyte (protein or surrogate peptide) is then eluted for LC–MS/MS analysis (Figure 1).

Herein, we discuss recent advance in the tech- niques used for quantitative LC–MS/MS analysis of protein therapeutics, especially in the areas of enzyme digestion and immunocapture.

Advances in enzyme digestion Trypsin has been the endoprotease of choice for quan- titative work, while other endoproteases such as chy- motrypsin, Asp-N, Glu-C, Lys-C, protease K and pep- sin have also been used in proteomics and quantitative work  [14–20]. Trypsin specifically hydrolyzes peptide bonds at the carboxyl side (or so called C-terminal) of lysine and arginine residues and tends to yield surro- gate peptides typically in the 5–40 amino acid range.

Other endoproteases hydrolyze peptide bonds at other specific amino acids. A list of endoproteases and their specific cleavage sites is presented in Table 1. It is, there- fore, possible to target a specific region of the protein, for example, the CDR in an antibody, by choosing the appropriate endoprotease (or a mixture of endoprote- ases) to generate a surrogate peptide that encompasses the region of interest. For example, Hager et al.  [6] used Asp-N to generate peptides at the C-terminus of various FGF21 modalities to study FGF21 C-termi- nus clipping in vivo, while trypsin failed to generate a suitable surrogate peptide. By making the appropri- ate choice of surrogate peptide and endoproteases, it is feasible to obtain high selectivity against a complex mixture of proteins in typical biological matrices. It is noted that the choice of enzyme(s) can be guided by using online tools such as ExPASy [21], and should be confirmed experimentally for reasons detailed below. It is the authors’ experience that the best surrogate pep- tides for good retention on reversed-phase LC columns and reasonable MS sensitivity are between 10 and 30 amino acids in length and basic in nature for optimal

ESI.

Recently, cysteine proteinase from Streptococcus pyo- genes (IdeS) has been adopted to cleave IgG at a single site below the hinge region, yielding F(ab’)2 and Fc frag- ments for protein characterization, making it attractive in the area of antibody–drug conjugates (ADCs) devel- opment  [22,23]. It could potentially be combined with other endoproteases in quantitative work. Aside from these enzymes, other new enzymes are being utilized for proteomic work, which can be readily adopted for quantitative protein analysis. For example, Kadek et al. reported the production of aspartic protease Nepenthe- sin-1 using recombinant technology as an alternative to the endoprotease pepsin [24].

For protein quantitation by LC–MS/MS, the enzyme digestion step is crucial to the reproducibility and sensitivity of the analytical assay, especially when a good internal standard that can compensate for vari- ability of digestion between samples, in the form of stable-isotope labeled protein is not always readily available. Enzymatic digestion is a series of complex chemical reactions with the enzyme serving as cata- lyst, and there are different parameters that can affect the reproducibility of enzyme cleavage. Trypsin, being the mostly commonly used endoprotease in quantita- tive work and proteomics, has been extensively investi- gated. While these investigations have been conducted

Protein (peptide) of interest in biological matrix, e.g., serum, plasma

Immunocapture with capture agent on solid support Washing with buffer

Elution with addition of acid www.future-science.com

849 future science group Techniques for quantitative LC–MS/MS analysis of protein therapeutics Review with the goal of applying to proteomics, the same principles are applicable to quantitative analysis of proteins. Trypsin is a relatively well-behaved endopro- tease, yet, there is published report on ‘missed’ or non- specific cleavage with the use of trypsin [25]. In other words, it cleaves peptide bonds at residues other than the C-terminal of lysine and arginine residues. Since most of the peptides are generated by at least two or more (the exceptions are the ones at the C-terminus and N-terminus) cleavages, any ‘missed’ or ‘nonspe- cific’ cleavage can affect the generation of the desired surrogate peptide (and hence quantitation of the pro- tein analyte) especially if the ‘other’ cleavage is located within the surrogate peptide. Although online in silico tools such as ExPASy are invaluable to identify the potential cleavage positions, bioanalysts are encour- aged to confirm the cleavage experimentally by iden- tifying the formation of the desired surrogate peptides with tools such as high-resolution mass spectrometer.

For some proteins, it was reported that the use of surfactants, reduction of the disulfide bonds on pro- teins followed by alkylation of free thiol groups are needed to achieve efficient trypsin digestion by expos- ing the desired cleavage site to trypsin. Obviously, these additional chemicals, especially surfactants may introduce undesired matrix effects to the mass spec- trometric analysis and the additional sample cleanup would likely result in sample loss and lower sensitiv- ity. A number of recent studies have been focused on identifying surfactants that are compatible with mass spectrometric analysis [26–28].

Another important, albeit not intuitive, parameter is the quality of trypsin, given that the native trypsin is susceptible to autolysis in which the trypsin cleaves itself generating pseudotrypsin, which exhibits a broad- ened specificity including a chymotrypsin-like activity.

Such autolysis products, together with contaminants [29] present in a trypsin preparation, would result in addi- tional peptide fragments that could interfere with the detection of the target surrogate peptide. In addition, the autolysis of trypsin could result in lowering of tryp- sin concentration over time. This can potentially affect the digestion efficiency and specificity and hence, the reproducibility of the quantitation work. Some com- mercial trypsin suppliers modify the lysine residues in the porcine trypsin by reductive methylation, yielding a highly active and stable molecule that is not susceptible to autolysis. The specificity of purified trypsin can also be further improved by tosyl phenylalanyl chloromethyl ketone (TPCK), a protease inhibitor treatment, which inactivates chymotrypsin. Multiple groups  [29–32] have evaluated the quality of the commercially available tryp- sin and it was generally agreed that it had significant impact on enzymatic digestion. Burkhart et al. [32] pro- posed a procedure to evaluate the digestion efficiency and specificity of the trypsin. The primary drawback of

Table 1. A list of commonly used endoproteases.

Enzyme Biological source Cleavage site Comments Arg-C

Clostridium histolyticum Cleaves peptide bonds at the C-terminal of arginine, including sites next to proline.

Cleaves also at lysine residue Requires dithiothreitol, cysteine or another reducing agent, and

CaCl2 to activate Asp-N Pseudomonas fragi Cleaves peptide bonds at the N-terminal side of aspartic acid and cysteic acid residues

Not applicable Chymotrypsin Bovine pancreas Cleaves peptide bonds at the C-terminal of tyrosine, phenylalanine, tryptophan and leucine. Methionine, alanine, aspartic acid and glutamine may be cleaved at a lower rate

Not applicable Glu-C Staphylococcus aureus Cleaves peptide bonds at the C-terminal of glutamine and aspartic acid

Not applicable Lys-C Lysobacter enzymogenes Cleaves peptide bonds at the C-terminal of lysine

Not applicable Pepsin Porcine stomach Cleaves peptide bonds at the C-terminal of phenylalanine, leucine, tyrosine and tryptophan.

Not applicable Proteinase K Tritirachium album Limber

Cleaves peptide bonds adjacent to the carboxylic group of aliphatic and aromatic amino acid

Useful in general digestion of proteins Trypsin Bovine or porcine pancreas

Cleaves peptide bonds at the C-terminal of lysine and arginine residues

Not applicable 850 Bioanalysis (2016) 8(8) future science group

Review  Fung, Bryan & Kozhich using modified trypsin is the higher cost compared with native trypsin, especially for quantitation of protein therapeutics in pharmacokinetic/toxicokinetic samples in which a large number of samples are processed.

Additional parameters such as digestion buffer com- position/pH, ratio of protein to enzyme and combina- tion with other endoproteases such as Lys-C have been evaluated [26–33]. While it is possible that some of these parameters are protein-specific (or peptide-specific), it is certainly worthwhile to investigate the effect of these parameters on the tryptic digestion of the pro- tein analyte and carefully optimize as needed during method development. In addition, though most of the investigative work has been performed extensively with trypsin because it is the most widely used endoprote- ases in proteomics and protein quantitation, the same parameters should be carefully considered when other endoproteases are used.

In an effort to improve reproducibility and efficiency, research has been undertaken to immobilize endopro- teases such as trypsin, pepsin and protease K on solid support such as magnetic beads. The immobilized enzymes have been reported to have improved reproduc- ibility and efficiency by reducing nonspecific cleavage, and making online digestion feasible and amenable to automation [14–15,34].

Besides enzyme digestion, it is possible to generate surrogate peptides with chemical means such as cyano- gen bromide and dilute formic acid. Fung  et  al. and

Wang et al. [35,36] reported protein quantitation with the use of dilute formic acid at elevated temperatures. The advantages of using chemical means include the relative ease of use and low cost. The major drawback is lack of specificity compared with endoproteases. Nonetheless, it is a valuable tool for protein quantitative work. Just like endoproteases, the potential cleavage positions can be identified by using online in silico tools such as ExPASy.

As in the case of enzymatic digestion, bioanalysts are encouraged to confirm the cleavage experimentally, and evaluated the optimal parameters such as concentrations of cyanogen bromide, formic acid and temperature for the protein analyte.

With the advance in protein engineering and purifi- cation, and the wider use of LC–MS in proteomics, pro- tein characterization and protein quantitation, there are opportunities for manufacturers to further improve the quality and specificity of the endoproteases, and identify new endoproteases with specific cleavages.

Advances in immunocapture Protein therapeutics in general have similar physio- chemical properties as other endogenous proteins and very different physiochemical properties from small molecules, therefore, traditional sample cleanup tech- niques for small molecules such as liquid–liquid extrac- tion with water-immiscible solvents such as ethyl acetate and protein precipitation may not be suitable.

Recently, immunocapture has been more widely employed as a highly selective sample cleanup method by taking advantage of the unique immunoaffinity of the target analyte (either protein therapeutic or its sur- rogate peptide) and the capture agent, and thus provides unique selectivity. It is similar to the capture step used in

LBA. The capture agent is usually an antibody specific for the target analyte and binds to solid support. In this procedure, the target analyte binds specifically to the capture agent, which is immobilized on a solid support (e.g., magnetic beads, agarose beads or column packing material), and thus is separated from other endogenous proteins and peptides, which do not bind very tightly to the capture agent. The mixture is then washed with a buffer to remove unbound proteins and other endog- enous components. The analyte is then eluted from the capture agent by the addition of acid, and followed by digestion with an endoprotease, or hydrolyzed by dilute acid at elevated temperature. The resulting surrogate peptide is then analyzed by LC–MS/MS. As expected, this sample cleanup produces a very clean extract and greatly reduces the matrix effect to the LC–MS/MS analysis. Besides removing other endogenous proteins and components, the immunocapture step can also serve as an enrichment step, and hence, improve the sensitivity of the assay as detailed by Wang et al. [10].

As the name implies, the crucial component to suc- cessful immunocapture is the capture agent, be it an antibody, a protein or a fragment of a protein. The ideal capture agent, as in LBA, binds the protein analyte alone with high affinity (but not irreversible binding to allow dissociation of protein from the capture agent) and with minimal affinities to other potential interfer- ing components at much higher concentrations than the protein analyte such as peptides, endogenous proteins and co-administered protein therapeutics. Commercial availability and low cost are additional attributes to an ideal capture agent. In practice, a less than ideal capture agent can be used successfully with LC–MS/MS detec- tion because unlike LBA with nonspecific detection antibody and detection techniques such as fluorescence,

LC–MS/MS provides a high-level of specificity as the detection of surrogate peptide is based on its intrinsic m/z ratio and in theory, as little as 1 amu difference can be detected. This allows the use of less specific capture agent in which small amount of other proteins with various affinities to the capture agent are copurified and digested by endoproteases, and the unique surro- gate peptide from the protein analyte is then analyzed by LC–MS/MS. Table 2 summarizes different capture agents used for immunoaffinity enrichment. www.future-science.com

851 future science group Techniques for quantitative LC–MS/MS analysis of protein therapeutics Review

Commercially available Protein A, G, A/G and L, immobilized on agarose beads or magnetic beads, have been used as the capture agents. These proteins binds to different areas (Fc, Fab, or κ light chain in the case of protein L) of the immunoglobulins, especially IgG of many species with different affinities. They are very useful for the protein analytes containing appropriate fragments (Fc, Fab, κ light chain etc.) of IgG as dem- onstrated by Chenau el al. [17]. and Bronesma et al. [37]

Another commercially available class of capture agents, anti-human IgG (Fc-specific) antibodies from goat or mouse have also been successfully deployed [7]. These antibodies bind proteins analytes containing human

IgG, and specifically the Fc domain of IgG. They offer the advantages of being commercially available in immobilized, high-throughput format, with estab- lished protocol, ease of use and amenable to automa- tion. These antibodies are especially useful during the discovery phase of drug development when multiple protein therapeutic candidates are evaluated and lim- ited resources are available to generate the antibodies specific for the protein analytes in a timely manner.

The major drawback is that they also bind to other endogenous proteins containing IgG. Cross-reactivity with other endogenous proteins should, therefore, be carefully evaluated when these capture agents are used.

The best capture agents are the ones that specifi- cally bind the protein analytes. They have been suc- Table 2. A list of commonly used capture agents for immunocapture.

Capture agent Target Pros Cons Protein A

IgG of many mammalian species, specifically the heavy chain within the Fc region of most immunoglobulins and also within the Fab region of the human VH3 family

– Commercially available – Binds to many proteins that contain

IgG – Good capture agent after immobilized on agarose beads or magnetic beads

– Not highly selective due to cross-reactivity with other proteins that contain IgG

Protein G

Binds to the Fc and Fab region of immunoglobulins

– Commercially available – Binds to many proteins that contain

IgG, with different affinity than Protein A – Good capture agent after immobilized on agarose beads or magnetic beads

– Not highly selective due to cross-reactivity with other proteins that contain IgG

Protein A/G

Protein A/G is a recombinant fusion protein that combines IgG- binding domains of both Protein

A and G. It combines the binding affinity of Protein A and G, and is lesser pH-dependent than Protein

A – Commercially available – Combines the affinity of Protein A and

G – Good capture agent after immobilized on agarose beads or magnetic beads

– Not highly selective due to cross-reactivity with other proteins that contain IgG

Protein L

Protein L binds antibodies through light chain interactions, specifically those with κ light chain. Protein

L binds to representatives of all antibody classes, including IgG,

IgM, IgA, IgE and IgD. Single chain variable fragments (scFv) and Fab fragments also bind to Protein L

– Commercially available – Binds to antibodies with κ light chain, thus, offer alternative to Protein A and

G – Good capture agent after immobilized on agarose beads or magnetic beads

– Not highly selective due to cross-reactivity with other proteins that contain κ light chain

– Comparatively higher cost

Anti-human IgG Fc specific

Targets human IgG and does not bind other human immunoglobulins

– Commercially available – Binds specifically with proteins containing human IgG (Fc) with no significant reactivity with human IgG (Fab2), IgM or other serum proteins. It is therefore, potentially useful to fusion proteins containing human IgG

– May need to test cross- reactivity with other species

IgG

Target capture

Antibodies targets uniquely the analyte of interest

– Highly specific with desired binding affinity – Minimal crossreactivity with other proteins in the biological matrices

– Time consuming and costly to produce, especially for analytes at discovery stage

852 Bioanalysis (2016) 8(8) future science group Review  Fung, Bryan & Kozhich cessfully used in many cases, for example, the use of anti-neuron-specific enolase to capture neuron-specific enolase  [3,5,6,8,9]. These capture agents in general are custom-made, and can be either polyclonal or mono- clonal. They produce the cleanest extract. They are usually used in later stage of drug development because it is in general time-consuming and costly to generate them. The capture agents can bind to the CDR of the therapeutic antibody or a unique region of the protein therapeutic. The key to the selection of a suitable capture agent is a thorough understanding of difference between the protein therapeutic and other proteins (or interfer- ing components) in the biological matrix. For example,

Fung et al. [35] used an antibody that targeted the adnec- tin region of an FGF-21-adnectin fusion protein even though the surrogate peptide was in the FGF-21 portion of the protein. Xu et al. [3] used anti-PEG antibody to capture a pegylated protein therapeutic by binding to the

PEG component of the protein therapeutic. It is recom- mended to screen for a number of different antibodies to select the one with highest recovery of the protein therapeutic.

In recent years, the specificity of immunocapture has been further explored to detect/quantify analytes of biological interest that could not have been accom- plished otherwise by LBA alone, for example, quantita- tion of antidrug antibody (ADA) complexes. Bronsema et al. [37] reported the use of immunocapture with Pro- tein G as the capture agent to quantify ADA-human α-glucosidase complex in human plasma. In this work, the ADA-human α-glucosidase complex is captured by Protein G because of its unique binding affinity to the constant region of the immunoglobulin of the

ADA, and thus separating from the unbound human α-glucosidase (with no bound ADA).

Another area for applying immunocapture is in the pharmacokinetic assays of  ADCs. Dere  et  al.,

Kaur et al. and Myler et al. [38–40] reported the use of immunocapture-LC–MS/MS to quantify the active payload (the active drug in the modality) and its metabolite conjugated to the antibody in an ADC.

The ADC (with the active payload and its metabolite) was bound to the capture agent (an anti-ID antibody) and separated from the unconjugated payload. The active payload and its metabolite were then analyzed by

LC–MS/MS after cleavage from the ADC by enzymes such as cathepsin B. Besides proteins, immunocapture can be adopted to capture other types of analytes. For example, Chenau et al. [17] reported the use of immu- nocapture to detect Bacillus anthracis spores, in which an antibody that was specific to B. anthracis spores was used to capture the spores, which was then analyzed by

LC–MS/MS, following trypsin and Glu-C digestion.

Besides taking advantage of the unique immuno- affinity between the capture agent and the protein therapeutic to isolate the protein therapeutic before enzyme digestion, it is also possible to explore the unique immunoaffinity between the capture agent and the surrogate peptide(s) to perform immunocapture after enzyme digestion, and achieve additional sample cleanup and improved sensitivity [10]. Neubert et al. [41] reported performing two immunocaptures to quantify total human β-nerve growth factor, the first immuno- capture step was to capture the human β-nerve growth factor before trypsin digestion, followed by the second immunocapture step to capture the surrogate pep- tide produced by the trypsin digestion with a differ- ent capture agent. Palandra et al. [42] also successfully employed the same strategy to quantify human and monkey IL-21.

Despite the unique selectivity of immunocapture, there are occasions where a more universal approach is desired, especially during the discovery phase in which a single drug candidate has not been finalized and there is interest in quantifying a class of proteins instead of a single one. Li et al. and Zhang et al. [43,44] reported the use of anti-human fragment (anti-Fc) antibody that recognized human monoclonal antibody protein thera- peutics but not the endogenous immunoglobulins in the preclinical samples (e.g., monkey serum). Another possibility is to utilize more than one capture agent to capture different multiple analytes [45].

As expected, one of the major drawbacks (or bottle- neck) of employing target-specific capture agent is the availability of the appropriate capture agent. In order to overcome the long lead time in generating the cap- ture agent, research work has been pursued to expedite the generation of capture agents. Säll et al. [46] reported the use of AFFIRM – a multiplexed immunoaffinity platform that utilized recombinant antibody fragments (in this case, scFv), generated by phage display technol- ogy to produce capture agents against different target proteins, while Whiteaker et al. demonstrated that Fab alone can be used as the capture agents instead of mono- clonal antibodies  [47]. In addition, Bostrőm et al.  [48] investigated the applicability of antibodies generated with Human Protein Atlas as the capture agents.

On another front, agarose beads and magnetic beads have been widely used to solid support for the immobi- lization of capture agents [3,5–9,37,43–44]. Agarose beads require the use of a centrifugation step or chromato- graphic setup for isolating the protein therapeutic and less amenable for high-throughput sample processing.

In recent years, magnetic beads are gaining popular- ity due to its ease of isolating/removing the magnetic beads, comparability of high-throughput sample pro- cessing and shorter processing time. Their major dis- advantages are time-consuming and labor-intensive www.future-science.com

853 future science group Techniques for quantitative LC–MS/MS analysis of protein therapeutics Review washing step and the associated cost. Yang et al.  [49] investigated the use of ELISA microplate as a cost- effective alternative to magnetic beads. Another possi- bility is to reuse the antibody and additional work will need to be done on this end.

Another area that can impact the quantitation of protein therapeutic is the choice of internal standard to correct for the variability of the immunocapture between different samples. The internal standard would need to be a protein of very similar proper- ties to the analyte so it binds to the capture agent in the same manner as the analyte. It is not likely that an analog surrogate peptide (stable-isotope labeled or otherwise) would have similar binding affinity to the capture agent as the protein analyte and thus compen- sate properly for the variability. Nonetheless, based on the published results [3,5–9,37,41,43–44], acceptable accu- racy (within ±20%) and precision (≤20%), good lin- earity could still be achieved even in the absence of a stable-isotope labeled protein internal standard [50].

Conclusion Nowadays, protein therapeutics make up for a significant portion of the portfolio of many pharmaceutical/biotechnology companies.

With these new modalities, new analytical technologies are needed to properly characterize and quantify them.

LC–MS/MS, especially in combination with immu- nocapture, has emerged as a viable technique to quan- tify protein therapeutics in biological matrices. With

LBA being the cost-effective gold standard of quanti- fying proteins, one area of great potential is to apply immunocapture-LC–MS/MS to answer biological questions where LBA data alone are not sufficient, for example, quantitation of ADA–protein complexes, conjugated payload in ADC. As summarized in this work, tremendous advancement and understanding of the important parameters that can significantly impact the enzymatic digestion have been made in recent years, be it the quality of the trypsin or identification of new endoproteases. As for immunocapture, select- ing a suitable capture agent requires a thorough under- standing of difference between the protein therapeutic and other, often interfering proteins in the biological matrix, stage of drug development and appreciation of the biological questions that need to be answered. The commercially available capture agents such as Protein

A, anti-human Fc IgG and custom-made target cap- ture agents, with their pros and cons are all important tools in the endeavor to answer important biological questions about the absorption, distribution, metabo- lism and excretion of the protein therapeutics. These quickly become valuable tool kits in the toolbox of bioanalysts.

At the time of publication, most of the work pub- lished (with few exceptions) has not been conducted in regulated environment, or used in Biologics License

Applications filings yet, since the field is still at a rela- tively early stage. There is need for continuous dialogues with the regulators before submitting pharmacokinetic data generated by immunocapture-LC–MS/MS and how they correlate with the data generated by LBA, if it is deemed possible. Toward that end, one potential area of future development is the generation of internal standards that can compensate for the variability of immunocapture and enzyme digestion, and therefore improve the reproducibility and ruggedness of the bio- analytical methods, which is important for filing pur- pose. Stable-isotope labeled proteins are the ideal inter- nal standards because they have same immunoaffinity to the capture agents, enzyme digestion efficiency and mass spectrometric properties as the protein analytes, but it is time-consuming and costly to generate them.

Advancement in protein engineering can result in cost reduction and decrease in production time. Another possibility is to carefully control the immunocap- ture procedure to minimize variability, for example, improvement in instrumentation, and it is certainly another area of focus for bioanalysts.

Future perspective In the next 5–10 years, immunocapture-LC–MS/MS will continue to mature and will likely be more widely adopted and routinely used for protein bioanalysis. We expect further advancement in the biology of endopro- teases and improvement in immunocapture technol- ogy, both in the instrumentation and the generation of capture agents, thus resulting in reduction of the cost of sample analysis and lead time for method devel- opment. With the advancement of instrumentation used for automation, the immunocapture and enzyme digestion steps can be further improved to reduce the analysis time and increase throughput. Addition dia- logues with regulators will facilitate the inclusion of data generated by this technology platform in filing applications.

Financial & competing interests disclosure The authors (EN Fung and A Kozhich) of this article are cur- rent employees of Bristol-Myers Squibb Company (BMS). All financial support for the studies reported herein was provided by BMS. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject mat- ter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

854 Bioanalysis (2016) 8(8) future science group Review  Fung, Bryan & Kozhich

Executive summary Background • LC–MS, especially in combination with immunocapture has emerged as a viable technique to measure protein therapeutics in biological matrices. One area of great potential is to apply immunocapture-LC–MS/MS to answer biological questions where ligand-binding assays data alone are not sufficient, for example, quantitation of antidrug antibody–protein complex, conjugated payload in antibody drug conjugate.

Advance in enzyme digestion & immunocapture • Tremendous advancement has been made in recent years to both the enzyme digestion and immunocapture.

Conclusion • At the time of publication, most of the work published (with few exceptions) has not been conducted in regulated environment, or used in Biologics License Applications filings yet since the field is still at a relatively early stage. There is need for continuous dialogues with the regulators before submitting pharmacokinetic data generated by immunocapture-LC–MS/MS and how they correlate with the data generated by ligand-binding assays, if it is deemed possible.

References Papers of special note have been highlighted as:

• of interest 1 DeSilva B, Smith W, Weiner R et al. Recommendations for the bioanalytical method validation of ligand-binding assays to support pharmacokinetic assessments of macromolecules.

Pharm. Res. 20(11), 1885–1900 (2003).

2 Myler HA, Given A, Kolz H, Mora JR, Hristopoulos G.

Biotherapeutic bioanalysis: a multi-indication case study review. Bioanalysis 3, 623–643 (2011).

3 Xu Y, Mehl JT, Bakhtiar R, Woolf EJ. Immunoaffinity purification using anti-PEG antibody followed by two- dimensional liquid chromatography/tandem mass spectrometry for the quantification of a PEGylated therapeutic peptide in human plasma. Anal. Chem. 82,

6877–6886 (2010).

• Summarizes the comprehensive method development of quantifying a PEGylated peptide by LC–MS/MS.

4 Beasley-Green A, Burris NM, Bunk DM, Phinney KW. Multiplexed LC–MS/MS assay for urine albumin.

J. Proteome Res. 13, 3930–3939 (2014).

5 Cingöz A, Hugon-Chapuis F, Pichon V. Total on- line analysis of a target protein from plasma by immunoextraction, digestion and liquid chromatography- mass spectrometry. J. Chromatogr. B Analyt. Technol. Biomed.

Life Sci. 878, 213–221 (2010).

6 Hager T, Spahr C, Xu J, Salimi-Moosavi H, Hall M.

Differential enzyme-linked immunosorbent assay and ligand- binding mass spectrometry for analysis of biotransformation of protein therapeutics: application to various FGF21 modalities. Anal. Chem. 85, 2731–2738 (2013).

7 Onami I, Ayabe M, Murao N, Ishigai M. A versatile method for protein-based antigen bioanalysis in non-clinical pharmacokinetics studies of a human monoclonal antibody drug by an immunoaffinity liquid chromatography-tandem mass spectrometry. J. Chromatogr. A. 1334, 64–71 (2014).

8 Torsetnes SB, LØvbak SG, Claus C et al. Immunocapture and

LC–MS/MS for selective quantification and differentiation of the isozymes of the biomarker neuron-specific enolase in serum. J. Chromatogr. B Analyt. Technol. Biomed. Life

Sci. 929, 125–132 (2013).

9 Wang Y, Heilig JS. Differentiation and quantification of endogenous and recombinant-methionyl human leptin in clinical plasma samples by immunocapture/mass spectrometry.

J. Pharm. Biomed. Anal. 70, 440–446 (2012).

10 Wang W, Walker ND, Zhu LJ et al. Quantification of circulating D-dimer by peptide immunoaffinity enrichment and tandem mass spectrometry. Anal. Chem. 84, 6891–6898 (2012).

11 Wu ST, Ouyang Z, Olah TV, Jemal M. A strategy for liquid chromatography/tandem mass spectrometry based quantitation of pegylated protein drugs in plasma using plasma protein precipitation with water-miscible organic solvents and subsequent trypsin digestion to generate surrogate peptides for detection. Rapid Commun. Mass Spectrom. 25, 281–290 (2011).

12 Ouyang Z, Furlong MT, Wu S et al. Pellet digestion: a simple and efficient sample preparation technique for LC–MS/

MS quantification of large therapeutic proteins in plasma.

Bioanalysis 4, 17–28 (2012).

13 Liu G, Zhao Y, Angeles A, Hamuro LL, Arnold ME, Shen JX.

A novel and cost effective method of removing excess albumin from plasma/serum samples and its impacts on LC–MS/MS bioanalysis of therapeutic proteins. Anal. Chem. 86, 8336–

8343 (2014).

14 López-Ferrer D, Petritis K, Robinson EW et al. Pressurized pepsin digestion in proteomics: an automatable alternative to trypsin for integrated top-down bottom-up proteomics. Mol.

Cell. Proteomics 10(2), M110.001479 (2011).

15 Slováková M, Peyrin JM, Bílková Z, Juklícková M,

Hernychová L, Viovy JL. Magnetic proteinase K reactor as a new tool for reproducible limited protein digestion. Bioconjug.

Chem.19, 966–972 (2008).

16 van den Broek I, Sparidans RW, Schellens JH, Beijnen

JH. Enzymatic digestion as a tool for the LC–MS/MS quantification of large peptides in biological matrices: measurement of chymotryptic fragments from the HIV-1 fusion inhibitor enfuvirtide and its metabolite M-20 in human plasma. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 854,

245–259 (2007).

17 Chenau J, Fenaille F, Ezan E et al. Sensitive detection of

Bacillus anthracis spores by immunocapture and liquid chromatography-tandem mass spectrometry. Anal. Chem. 83,

8675–8686 (2011). www.future-science.com 855 future science group

Techniques for quantitative LC–MS/MS analysis of protein therapeutics Review

18 Doucet J, Avrameas A. A novel method for quantitative measurement of a therapeutic monoclonal antibody in the presence of its target protein using enzymatic digestion.

J. Pharm. Biomed. Anal. 52, 565–570 (2010).

19 Gilmore JM, Kettenbach AN, Gerber SA. Increasing phosphoproteomic coverage through sequential digestion by complementary proteases. Anal. BioAnal. Chem. 402,

711–720 (2012).

20 Neubert H, Gale J, Muirhead D. Online high-flow peptide immunoaffinity enrichment and nanoflow LC–MS/MS: assay development for total salivary pepsin/pepsinogen. Clin.

Chem. 56, 1413–1423 (2010).

21 Peptide Cutter on ExPASy. http://web.expasy.org/peptide_cutter

22 Leblanc Y, Romanin M, Bihoreau N, Chevreux G. LC-MS analysis of polyclonal IgGs using IdeS enzymatic proteolysis for oxidation monitoring. J. Chromatogr. B Analyt. Technol.

Biomed. Life Sci. 961, 1–4 (2014).

23 Firth D, Bell L, Squires M, Estdale S, McKee C. A rapid approach for characterization of thiol-conjugated antibody–drug conjugates and calculation of drug-antibody ratio by liquid chromatography mass spectrometry. Anal.

Biochem. 485, 34–42 (2015).

24 Kadek A, Mrazek H, Halada P, Rey M, Schriemer DC, Man

P. Aspartic protease nepenthesin-1 as a tool for digestion in hydrogen/deuterium exchange mass spectrometry. Anal.

Chem. 86, 4287–4294 (2014).

25 Chiva C, Ortega M, Sabidó E. Influence of the digestion technique, protease, and missed cleavage peptides in protein quantitation. J. Proteome Res. 13, 3979–3986 (2014).

26 An B, Zhang M, Johnson RW, Qu J. Surfactant-aided precipitation/on-pellet-digestion (SOD) procedure provides robust and rapid sample preparation for reproducible, accurate and sensitive LC/MS quantification of therapeutic protein in plasma and tissues. Anal. Chem. 87, 4023–4029 (2015).

27 Yu YQ, Gilar M, Lee PJ, Bouvier ES, Gebler JC. Enzyme- friendly, mass spectrometry-compatible surfactant for in-solution enzymatic digestion of proteins. Anal. Chem. 75,

6023–6028 (2003).

28 Chang YH, Gregorich ZR, Chen AJ et al. New mass- spectrometry-compatible degradable surfactant for tissue proteomics. J. Proteome Res. 14, 1587–1599 (2015).

29 Bunkenborg J, Espadas G, Molina H. Cutting edge proteomics: benchmarking of six commercial trypsins.

J. Proteome Res. 12, 3631–3641 (2013).

• Describes how the quality of trypsin affects the digestion efficiency and reproducibility.

30 Brownridge P, Beynon RJ. The importance of the digest: proteolysis and absolute quantification in proteomics.

Methods 54, 351–360 (2011).

31 Hao P, Ren Y, Datta A, Tam JP, Sze SK. Evaluation of the effect of trypsin digestion buffers on artificial deamidation.

J. Proteome Res. 14, 1308–1314 (2015).

32 Burkhart JM, Schumbrutzki C, Wortelkamp S, Sickmann A,

Zahedi RP. Systematic and quantitative comparison of digest efficiency and specificity reveals the impact of trypsin quality on MS-based proteomics. J. Proteomics 75, 1454–1462 (2012).

33 Bronsema KJ, Bischoff R, van de Merbel NC. High-sensitivity

LC–MS/MS quantification of peptides and proteins in complex biological samples: the impact of enzymatic digestion and internal standard selection on method performance. Anal.

Chem. 85, 9528–9535 (2013).

34 Sun L, Li Y, Yang P, Zhu G, Dovichi NJ. High efficiency and quantitatively reproducible protein digestion by trypsin- immobilized magnetic microspheres. J. Chromatogr. A. 1220,

68–74 (2012).

35 Fung EN, Zambito F, Haulenbeek J et al. Targeting an acid labile aspartyl-prolyl amide bond as a viable alternative to trypsin digestion to generate a surrogate peptide for

LC–MS/MS analysis. Bioanalysis 6, 2985–2998 (2014).

36 Wang N, Li L. Reproducible microwave-assisted acid hydrolysis of proteins using a household microwave oven and its combination with LC-ESI MS/MS for mapping protein sequences and modifications. J. Am. Soc. Mass Spectrom. 21,

1573–1587 (2010).

37 Bronsema KJ, Bischoff R, Pijnappel WW, van der Ploeg AT, van de Merbel NC. Absolute quantification of the total and antidrug antibody-bound concentrations of recombinant human α-glucosidase in human plasma using protein G extraction and LC–MS/MS. Anal. Chem. 87, 4394–4401 (2015).

38 Dere R, Yi JH, Lei C et al. PK assays for antibody–drug conjugates: case study with ado-trastuzumab emtansine.

Bioanalysis 5, 1025–1040 (2013).

39 Kaur S, Xu K, Saad OM, Dere RC, Carrasco-Triguero

M. Bioanalytical assay strategies for the development of antibody–drug conjugate biotherapeutics. Bioanalysis 5,

201–226 (2013).

40 Myler H, Rangan VS, Wang J et al. An integrated multiplatform bioanalytical strategy for antibody–drug conjugates: a novel case study. Bioanalysis 7, 1569–1582 (2015).

41 Neubert H, Muirhead D, Kabir M, Grace C, Cleton A,

Arends R. Sequential protein and peptide immunoaffinity capture for mass spectrometry-based quantification of total human β-nerve growth factor. Anal. Chem. 85, 1719–1726 (2012).

42 Palandra J, Finelli A, Zhu M, Masferrer J, Neubert H. Highly specific and sensitive measurements of human and monkey interleukin 21 using sequential protein and tryptic peptide immunoaffinity LC–MS/MS. Anal. Chem. 85, 5522–5529 (2013).

43 Li H, Ortiz R, Tran L et al. General LC–MS/MS method approach to quantify therapeutic monoclonal antibodies using a common whole antibody internal standard with application to preclinical studies. Anal. Chem. 84, 1267–1273 (2012).

44 Zhang Q1, Spellman DS, Song Y et al. Generic automated method for liquid chromatography-multiple reaction monitoring mass spectrometry based monoclonal antibody quantitation for preclinical pharmacokinetic studies. Anal.

Chem. 86, 8776–8784 (2014).

45 Torsetnes SB, Levernæs MS, Broughton MN, Paus E,

Halvorsen TG, Reubsaet L. Multiplexing determination of

856 Bioanalysis (2016) 8(8) future science group Review  Fung, Bryan & Kozhich small cell lung cancer biomarkers and their isovariants in serum by immunocapture LC–MS/MS. Anal. Chem. 86,

6983–6992 (2014).

46 Säll A, Carlsson F, Olsson N et al. AFFIRM – a multiplexed immunoaffinity platform that combines recombinant antibody fragments and LC-SRM analysis. J. Proteome

Res. 13, 5837–5847 (2014).

47 Whiteaker JR, Zhao L, Frisch C et al. High-affinity recombinant antibody fragments (Fabs) can be applied in peptide enrichment immuno-MRM assays. J. Proteome

Res. 13, 2187–2196 (2014).

48 Boström T, Johansson HJ, Lehtiö J, Uhlén M, Hober S.

Investigating the applicability of antibodies generated within the human protein atlas as capture agents in immunoenrichment coupled to mass spectrometry.

J. Proteome Res. 13, 4424–4435 (2014).

49 Yang W, Kernstock R, Simmons N, Alak A. ELISA microplate: a viable immunocapture platform over magnetic beads for immunoaffinity-LC–MS/MS quantitation of protein therapeutics? Bioanalysis 7, 307–318 (2015).

50 Zhao L, Whiteaker JR, Voytovich UJ, Ivey RG, Paulovich

AG. Antibody-coupled magnetic beads can be reused in immuno-mrm assays to reduce cost and extend antibody supply. J. Proteome Res. 14, 4425–4431 (2015).