Antibody Specificity?

“No, there is no such thing as a monoclonal antibody that, because it is monoclonal, recognizes only one protein or only one virus. It will bind to any protein having the same (or a very similar) sequence.”

-Clifford Saper, one of the world’s leading authorities on monoclonal antibodies, Harvard Medical School professor

https://off-guardian.org/2021/03/06/the-antibody-deception/

One of the primary selling points for the use of antibodies is their supposed (presumed) specificity. This is the ability of the antibody to discriminate between similar or even dissimilar antigens. This is how researchers try to claim that the measure of a “specific” antibody means one has been exposed to a certain “virus” in the past or that they are currently “immune” after vaccination. These measurements are used in order to state that vaccination is successful if they see a rise in a “specific” antibody over a set (and seemingly arbitrary) amount of titers.

However, it has been known for decades that antibodies are in fact not as specific as they are claimed to be. Numerous scientific papers have been withdrawn due to unreproducible and irreplicable results stemming from promised antibody specificity that was not to be. Instead of binding to the desired target, these antibodies are said to regularly attach to similar or even completely unrelated targets. They often cross-react to other “viruses” or fail to show up in appreciable amounts at all in individuals apparently exposed. This lack of specificity is one of many reasons antibody research is facing a disastrous reproducibility crisis that is only getting worse as time goes on.

One of the people trying to clean up this insurmountable mess is antibody expert Clifford Saper. He was editor-in-chief of the Journal of Comparative Neurology between 1994 and 2011. In that time, Saper came across numerous research papers which had to be withdrawn due to erroneous results from faulty antibodies. Excerpts from the below Nature article from 2015 provides further details about Saper’s involvement in an attempt to rectify this mess and the continued problems facing antibody research:

Antibody anarchy: A call to order

Antibodies used in research often give murky results. Broader awareness and advanced technologies promise clarity.

“A mouse first alerted Clifford Saper to the fact that antibodies were misleading the scientific community. As editor-in-chief of the Journal of Comparative Neurology between 1994 and 2011, he handled scores of papers in which scientists relied on antibodies to flag the locations of neurotransmitters and their receptors. Around the turn of the century, related investigations began to roll in from researchers using knockout mice, animals genetically engineered to not express a target gene. The results were unsettling. Antibody staining in knockout animals should have shown radically different patterns from those in unmodified animals. But all too often the images were identical. “As we saw more and more retractions due to this, I began to realize that we had no systematic way to evaluate papers that used antibodies,” recalls Saper, now chair of neurology at Beth Israel Deaconess Medical Center in Boston, Massachusetts.

Thus began a one-journal revolution. Saper and his editorial colleagues set up a policy of requiring extensive validation data on each antibody¹. The policy was good for rigour, but not submissions, he recalls. “Many authors were caught in the middle, and found it easier to publish their papers elsewhere.” But Saper persisted. His efforts eventually culminated in the JCN Antibody Database, an inventory of a few thousand antibodies that can be trusted for neuroanatomy.

Today, biomedical researchers still collect tales of antibody woe faster than country-music labels spin out sad songs. The most common grumble is the cheating reagent: the antibody purchased to detect protein X surreptitiously binds protein Y (and perhaps ignores X altogether). Another complaint is ‘lost treasure’: a run of promising experiments that stalls when a new batch of antibodies fails to reproduce previous findings (see ‘A market in a bind’).”

“It is alarming, then, to discover that antibodies can be unreliable reagents. Insufficient specificity, sensitivity and lot-to-lot consistency have resulted in false findings and wasted efforts. Antibody unreliability has taken its toll across studies in cancer, metabolism, ageing, immunology and cell signalling, and in any field concerned with researching complex biomolecules. The waste, in terms of time and resources, is colossal. Losses from purchasing poorly characterized antibodies have been estimated at $800 million per year, not counting the impact of false conclusions, uninterpretable (or misinterpreted) experiments, wasted patient samples and fruitless research time.²“

https://www.nature.com/articles/527545a?proof=t

According to the Nature article, antibodies are “unreliable reagents.” They lack specificity and sensitivity and there is wide lot-to-lot variability hindering reproducibility. Clifford Saper noticed this trend and in 2005, he wrote an open letter to the readers of the Journal of Comparative Neurology to address this growing threat to scientific integrity. In it, he described the problems facing antibody research: mainly the lack of specificity and the inability to reproduce results. He laid forth some criteria that needed to be met. His full letter is presented below:

An Open Letter to Our Readers on the Use of Antibodies

“The use of immunohistochemistry has become ubiquitous in neuroscience. A large majority of papers now published in The Journal of Comparative Neurology use immunohistochemistry, and some papers may employ a battery of ten or more antibodies to examine issues of colocalization or cell typing. This pattern has resulted in a flood of new information. . . .but also a flood of misinformation.

The Journal has repeatedly, over the last few years, received distressed communications from authors, who have had to withdraw papers because an antibody against a novel marker was found to stain tissue in knockout animals, who lack that target protein. In many cases these papers contained careful characterization of the antibodies and immunocytochemical controls. This issue has sensitized the Editors to the problem of antibody specificity, and we soon realized that many of the papers we were publishing had very limited characterization or controls for antibodies that were used. Subsequently the Editors noticed that a number of commercially available antisera, particularly against G-protein coupled receptors, gave staining patterns that did not match mRNA distributions, and that these antibodies still stained tissue from animals in which the receptor had been knocked out. We felt that the integrity of scientific communication was being threatened by the proliferation of poorly characterized antibodies that produce artifactual staining patterns, and therefore came up with a minimal set of rules for identification, characterization, and controls for immunohistochemistry that we are attempting to apply to all manuscripts that we publish (Saper and Sawchenko, 2003).

The result has been a substantial degree of confusion about what information is necessary for a paper in JCN (or any other journal, for that matter), to provide a reasonable level of assurance that an antibody is actually recognizing what it is supposed to be staining. In the interest of demystifying this procedure further, we present the following brief description of the three basic elements that are necessary in describing an antibody for use in neuroscience:

1.) Complete information on the antibody. It is important to recognize that antibodies are not simple reagents that always identify the same thing. Thus, to describe an antibody as completely as possible, so that other scientists can replicate your work, you need to provide information of two kinds:

—Identification: What was the source of the antibody? If another lab has donated the antibody, give the antiserum code number, and if possible, the bleed. If it was obtained from commercial sources, give the catalog, and if possible the lot number.

—Preparation of the antibody: What was the antibody actually raised against? Give the precise structure of the immunizing antigen, not just vague information about the part of the molecule that was used. What species was the antiserum raised in? Was it a polyclonal or monoclonal preparation? Please note that some antibody manufacturers deliberately try to obscure the information about the structure of the antigen. We have received complaints from authors that several manufacturers have claimed that sequence information on the antigen was “proprietary,” and they would not provide it. Because work using these antibodies is inherently not repeatable, such papers are not acceptable for publication. Our advice is:

NEVER BUY ANTIBODIES FOR WHICH THE MANUFACTURER WILL NOT DISCLOSE THE STRUCTURE OF THE IMMUNIZING ANTIGEN. THESE REAGENTS ARE NOT FIT FOR SCIENTIFIC WORK, AND THE WORK YOU DO WITH THEM WILL NOT BE PUBLISHABLE.

Examples of antibody descriptions that would be acceptable:
“This rabbit antiserum (Company XYZ #30248) was prepared against a synthetic peptide representing amino acids 121-142 from tyrosine hydroxylase.”
“This mouse monoclonal antibody, kindly donated by Dr. John Smith, University of Alabama, was raised against human placental choline acetyltransferase.”

2.) How has the specificity of the antibody been characterized? If the antiserum is against a large protein, it is important to know what it stains on a gel from the tissue and species you are using. Ideally, an antibody should stain a single band (or several bands if the antigen has several known molecular configurations) of appropriate molecular weight. This information is often included by the manufacturer in the technical information, and can be cited, as can previous studies that provided this information. However, it is not sufficient just to state “this antiserum was previously characterized (Jones et al., 2002).” You need to provide a reasonably critical reviewer with information on what characterization was done, and what it showed.

Examples of antibody characterization that would be acceptable:
“The antiserum stains a single band of 55 kD molecular weight on Western blot (manufacturer’s technical information).”
“This antiserum stains the 150kD but not the 130kD or 110kD forms of the molecule on Western blot (Fig. 1).

3.) What controls are necessary for immunostaining? We are constantly surprised by the oddly trusting nature of many of our colleagues, who seem to believe (or want to believe) that an antibody will stain what the manufacturer claims. In fact, in many cases nothing could be further from the truth. It is always important to provide the highest level of control you can to assure that the antibody is staining what you want it to stain.

The gold standard in this regard is: does the antibody stain the tissue of interest from which the molecule of interest has been removed? This standard is simple to achieve for antibodies against tracers, or bromodeoxyuridine, or green fluorescent protein, which are not normally present in tissue. A similar level of assurance can be obtained when a knockout mouse (or other animal) is available (i.e., if the antibody stains tissue in a wild-type animal and not the animals from which the gene has been deleted). But this is only applicable if you are looking at tissue from an animal for which a knockout exists.

Most of us, most of the time, must make due with lesser degrees of certainty. A useful control, which is especially important for antibodies raised against synthetic peptides, is to preadsorb the antiserum against the peptide. If the staining disappears, it is at least likely that the staining component of the antiserum is indeed raised against that antigen. This does not protect you from other tissue proteins that may cross-react, however, and we have had a number of papers eventually retracted when an anti-serum that passed the preadsorption test also were shown to stain tissue from knockout animals. This strategy cannot be used for monoclonal antibodies, which always will be adsorbed out by their antigen, even if they are staining something entirely different in the tissue.

Lacking the antigen (e.g., if it is a large protein), there are several alternative strategies for establishing staining specificity. If the pattern of staining of the antigen is well known (e.g., antibodies against glial fibrillary acidic protein or tyrosine hydroxylase), it is reasonable to show that tissue of the type and species you are studying, when stained with your antibody, produces a pattern that is identical to that previously reported. Alternatively, it can sometimes be demonstrated that one antibody (e.g., a monoclonal antiserum against a large protein) produces a pattern of staining that is identical to another antiserum (e..g., a polyclonal serum against a synthetic peptide) that is better characterized (and does pass the adsorption test). Antisera against different parts of the same molecule which produce the same staining pattern, or can be shown to be colocalized in double label studies, provide an important strategy for establishing specificity. Similarly, comparison with the pattern of mRNA expression by using in situ hybridization histochemistry can demonstrate that the immunostaining for the protein is genuine.

Note that omission controls (staining without the primary antibody) are NOT controls for specificity of the primary antibody at all. They control only for the specificity of the secondary antiserum.

Examples of acceptable controls for immunostaining (in roughly descending order of rigor):
“No staining was seen when the antibody was used to stain tissue from an orexin knockout mouse.”
“All staining was abolished when 1 ml of the diluted primary antibody was preincubated with 50 g of the immunizing peptide.”
“Staining with this antiserum was colocalized with in situ hybridization for the mRNA for the same protein (Fig. 3).”
“This antiserum against the N-terminal 13 amino acids of the protein gave the same staining pattern as another antiserum against the C-terminal cyanogen bromide fragment of the protein (Smith et al., 2001).”
“Staining of sections through the pons produced a pattern of tyrosine hydroxylase immunoreactivity that was identical with previous descriptions (Smith et al., 1992).”
“The GFAP antiserum only stained cells with the classic morphology and distribution of fibrillary astrocytes (Fig. 2; see Smith and Jones, 1998).”

The full information for any antibody usually requires only two or three sentences. Often this information can be contained within a table. However, it is critical that for each antibody used, the authors must supply sufficient information to assure that the result is replicable and likely to be correct. Blind faith that the antibody will stain whatever the manufacturer claims is not consistent with good science.

doi: 10.1002/cne.20839.

It is evident that the unreliability of the research papers being submitted was a problem Saper felt had huge implications for scientific integrity. False results were leading to a wave of misinformation flooding the market. Obviously, this is rather distressing as this creates a pattern of faulty research being built upon faulty research. Flawed results were and still are being presented as fact. Saper tried to implement some semblance of quality guidelines for researchers to follow in order to be accepted in his journal. Sadly, many ignored his pleas and looked elsewhere to publish their research.

In January 2009, Saper wrote a more in-depth article breaking down the issues regarding antibody specificity leading to faulty research. In it, he reiterated the same three criteria he advised others to follow in 2005 in order to produce “accurate” research. He broke down the differences between monoclonal and polyclonal antibodies and their respective issues relating to specificity. He explained that an over reliance on immunohistochemical staining has blinded researchers to the fact that this technique can not identify molecular targets accurately. While Saper offered many methods to attempt to validate antibodies, each of them are problematic in their own ways. He ultimately concluded that researchers are always uncovering new ways that nature can fool and because of this, no antibody localization is perfect:

A Guide to the Perplexed on the Specificity of Antibodies

Abstract

“Many investigators are unaware of the potential problems with specificity of antibodies and the need to document antibody characterization meticulously for each antibody that is used. In this review, I consider the principles of antibody action and how they define a set of rules for what information should be obtained by the investigator before using an antibody in a serious scientific investigation. (J Histochem Cytochem 57:1–5, 2009)

Since the description of indirect immunohistochemical (IHC) staining by Coons (1958), IHC staining has become a standard method used in most laboratories doing cellular or systems level localization of proteins and other cellular constituents. In fact, the methods have become so mundane that many current practitioners take for granted that an antibody that is sold to localize a particular molecular target will be both sensitive and specific. In the current era of very accurate DNA analyses by in situ hybridization, DNA chip analyses, and deep sequencing, it is often assumed that IHC has an analogous ability to identify molecular targets accurately.

Nothing could be further from the truth.

In fact, IHC methods remain as primitive, in terms of both sensitivity and specificity, as they were in the days when DNA sequencing was done by hand using sequencing gels. The fundamental principles on which antibody localization is based have not improved at all in the last two decades, and if anything, the slope occupied by IHC has become more slippery than ever.”

By fusing individual antibody-producing cells with antibody-producing myeloma cells, individual cells can be immortalized, so that they divide into colonies of “hybridoma” cells, all of which produce the same, identical immune globulin, with the same variable region. These monoclonal antibodies have the property that they will only bind to molecules that bind that single variable site. Although this relationship imparts specificity to the interaction, it is possible that the variable site may bind to a variety of different targets, particularly when tested in different tissues, and that these may be quite different from the molecule against which the antibody was raised.”

Monoclonal Antibodies vs Polyclonal Antisera

“As indicated above, access to monoclonal antibodies has provided us with antibodies that are pure reagents. The monoclonal antibodies are derived from hybridoma cells, which are grown either in culture or by injecting them intraperitoneally in a host animal. When the hybridoma cells are grown intraperitoneally, the host animals build up fluid, which is called ascites and which can be drawn off from the abdomen and contains high concentrations of the monoclonal antibodies. Either the culture fluid or ascites fluid containing the antibodies can be subjected to purification by precipitating the antibodies with protein A. The resulting relatively pure antibody preparations are quantified based on the micrograms of protein.

Polyclonal antisera, in contrast, are derived by bleeding animals a few weeks after they have been immunized. Usually several “booster immunizations” are given, and several bleeds are taken. Blood volume in a mammal is usually ∼7% of body weight, and typically ∼10–15% of total blood volume may be exsanguinated at any one time without injury to the animal. Hence, a single bleed from a 3-kg rabbit may be 25 ml, whereas a bleed from a 30-kg goat can be 250 ml. When the red blood cells are spun down from the clotted blood, the remaining serum is usually about one half this volume. As a result, a single bleed from a larger animal can be used for a much larger number of IHC reactions than a bleed from a smaller animal. The advantage of having the larger amount of serum per bleed is that each bleed is essentially a unique combination of antibody clones. Even when boosting the same animal with repeated immunizations with the same antigen, the antibody content in sequential bleeds may differ markedly. Hence, the lot for a polyclonal antiserum is critical, and even another batch from the same animal may have entirely different staining properties. For this reason, experienced immunohistochemists write down the lot numbers for each vial of antiserum and, when they have a good lot, buy up as much of that lot as they are likely to need in the foreseeable future to avoid inability to finish a project.

Synthetic Peptide Antigens and Antigen Mapping

The ability to create synthetic proteins and peptides has revolutionized the way in which antibodies may be made and how they can be characterized. Synthetic peptides are usually from a few amino acids up to ∼25 or 30 in length. The current peptide synthesis technology results in decreasing yields as the peptide lengthens, so that synthetic peptides much longer than this, while possible, are not practical. On the other hand, much longer amino acid sequences can be prepared by recombinant technology, in which a corresponding nucleic acid sequence is expressed either in a cellular or cell-free protein expression system. It seems obvious that the exact sequence used to create the antibody is critical to its properties, and hence, we will return to this issue in the criteria for antibody suitability.

At the same time, the availability of amino acid sequences from different parts of the parent target molecule has allowed us to identify the target sites in the native molecule to which the antibody binds. When the antibody binds to a partial sequence or a partial sequence competes against binding to the native molecule, the epitope, or structural features that the antibody recognizes, is presumed to be located in that sequence. This method is used to map the epitope that the antibody binds. However, this does not indicate what the sequence was of the original immunogen, because the antibody may have been made against an overlapping sequence.”

Antibodies Against Different Portions of the Same Molecule

A related topic is the ability to generate antibodies against synthetic peptides that are derived from different components of the same molecule. Thus, it is possible, for example, to have antibodies against a large protein target that specifically bind to the N- or the C-terminal portions of the protein. This possibility gives us a powerful potential tool to use in determining antibody specificity. When the two different antibodies stain exactly the same pattern, it is highly likely that they are staining the correct target.

Antibodies Against Phosphorylated or Glycosylated Epitopes

Another possibility provided by the use of synthetic antigens is to prepare immunogens that are specifically altered, for example, with phosphorylation, glycosylation, or some other post-translational modification. Antibodies prepared in this way may be able to distinguish between different modified forms of the same molecule with great accuracy. However, showing this specificity requires appropriate controls (such as staining after dephosphorylation).”

Rules for Judging Whether an Antibody Is Showing What Is Expected in Tissue

“Most investigators want to use antibodies to localize cellular components and do not want to have to become experts in immunology or IHC to do so. Hence, it is useful to have a set of criteria for what constitutes a reasonable degree of assurance that the antibody being used is actually targeting its correct antigen. The answers to the questions that follow are ones that investigators should ask for each antibody they are acquiring, before they ever use it in an experiment (why waste time on an invalid antibody?). If all investigators followed these rules, the literature would be much more accurate, and investigators would avoid wasting a lot of time on invalid antibodies.

What Immunogen Is Used to Raise the Antibody?

The first critical criterion in locating a valid antibody is that the immunogen against which the antibody was raised must be known. A key principle of science is that the work must be repeatable. Hence, if the antibody is raised against a “proprietary” antigen (usually a secret amino acid sequence, to avoid competitors from copying the product), it simply is not valid for serious scientific work. Some manufacturers have claimed that their “intellectual property” must be protected if they are to provide antibodies in the future, but in fact, this has become a routine process, and for most antibodies there are multiple manufacturers who do provide the sequence for their antigens. More importantly, if protecting their profits interferes with science, it is the use of their product that must be eliminated. Other manufacturers have claimed that they will provide their proprietary product to other laboratories in the future, so that the result of the experiment is repeatable. However, there is tremendous turnover in this field, and companies frankly are in business to make profits and not to protect scientific integrity. If they find tomorrow morning that they can make more profit selling shoes than antibodies, that is exactly what they will do, and no one will be able to repeat the work. Hence, a key issue in buying any new antibody is to avoid products for which the identity of the immunogen is not provided at the time it is purchased.

What is the Evidence That the Antibody Binds Specifically to the Expected Target Molecule in the Tissue of Interest?

The second key criterion for using an antibody in a scientific project should be to obtain at least reasonable evidence that the antibody does bind to its expected target in the tissue in which it will be studied and not to something else. This is often provided by a Western blot, which should show that the antibody stains a single band (or a set of bands) of appropriate molecular mass for that target. Note that if extraneous bands are stained, this indicates that the antibody has other additional targets in the tissue and should raise red flags against using that antibody for IHC, unless you have taken additional precautions. For example, we have seen authors take tissue from mice in which the target protein was deleted (as shown by Western blot) and preadsorb the antiserum against tissue from the knockout mouse before using it to stain the brain. This is a lovely control that removes the extraneous staining and provides strong confidence that what is stained is the target molecule.

Note the importance of doing the Western blot in the same tissue and species as the antibody will be applied for IHC. It is quite possible for the antibody to see only one band in some tissues but to see multiple extraneous bands in other tissues from the same animal. Similarly, manufacturers often try to “prove” specificity by running the antibody against a gel preparation of purified or recombinant protein. This may show that the antibody can bind to its target but does not tell anything about what else it may bind to in tissue.

What Controls Can Be Done to Insure That the Antibody Binds in Fixed Tissue Only to Its Target Molecule?

Despite our best attempts to insure specificity of the antibody against native proteins in the aqueous phase, ultimately we have to apply it to fixed tissue. In the fixed state, it is possible that the antibody that works well in a Western blot will find that its target antigen is distorted by the fixation process and no longer recognizable. In fact, this occurs so often that most manufacturers mark antisera as usable for Western blotting or IHC, and the latter are by far the rarer.

When polyclonal antisera are raised against a peptide antigen, it is common that most of the antisera that are produced will stain fixed tissue poorly or not at all. In one case in which the author screened antisera, we found only 2 of 31 against a common peptide hormone that could be used to stain brain tissue. If one applies the mathematics of a Poisson distribution to this problem (i.e., assume that the probability of stimulating a single antibody clone that recognizes the fixed molecule is an independent event), it is likely that, in most polyclonal sera, the antiserum is staining the tissue with only one or at most a small number of antibody clones (i.e., that the polyclonal, which may contain thousands of clones against other antigens the host animal encountered in its lifetime, is functionally a monoclonal or oligoclonal for this purpose).

One of the best tests to show that the antibody can identify its target in fixed tissue is to transfect the DNA for the target protein into cells that normally do not make it in tissue culture. The transfected and untransfected controls can then both be fixed and stained, and the presence of staining in the transfected cells shows that the antibody really does stain its target. However, this control does not prove that the antibody will only stain its target in the tissue of interest.

Another control for specific staining in tissue is the preadsorption test. Mixing the diluted antibody with an excess of the immunogen should completely block staining. This shows that the staining in the tissue is against something that is at least cross-reactive with the original protein (although it does not prove that this is what the target in the tissue actually is). In general, when the original immunogen is readily available, such as for a synthetic peptide, the preadsorption test should be run as a matter of course. This is less practical for large protein molecules and antibodies against partially purified tissue components. Note that the preadsorption control is meaningless for a monoclonal antibody (which is produced by screening for its binding to the target, and therefore will always bind it and always pass a preadsorption test, by definition) and for antibodies that have already been affinity purified (for the same reason).

As a practical matter, the best controls for assuring that the staining in the tissue is the target molecule involve one of two approaches (Lorincz and Nusser 2008). First, if the staining is being evaluated in mice or a closely related rodent species, and there is a strain in which the target molecule is deleted, the absence of staining is a strong confirmation of specificity. Unfortunately, this is not a perfect test, because the target that is stained in the tissue may be a related molecule that is downregulated in the knockout animal. In addition, this approach only applies to situations where there is a knockout strategy available, which limits it to a few model species. Finally, in many so-called knockout mice, the original protein is not entirely eliminated. If only a portion of that protein is still expressed, it may have no functional presence but still stain with your antibody. Hence, it is critical in a knockout control to make sure what the actual gene construct is and what is actually expressed.

The second molecular approach to confirming identity of the staining was alluded to above in the section on making antibodies against different components of the same target molecule. When the two antibodies are made in the same species, showing that the staining patterns are very similar is a strong control. When the two antibodies are made in different species, simultaneous staining and showing colocalization is an even more satisfying and persuasive control.

The methods described above are not by any means exhaustively detailed. There are many clever and innovative ways that are identified by investigators to test their antibodies each year. Science is endlessly creative, and we are always finding new methods and ways of improving older methods. At the same time, we are always uncovering new ways that nature can fool us. Thus, no antibody localization is really perfect, although following the practical guide provided here should help investigators, especially those who are new to the mysteries of IHC, to insure the scientific integrity of their work.”

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2605712/

Highlights from the next two studies do not involve Clifford Saper but they do help to show that this lack of specificity is a problem that not only he recognized. In 2014, an article examined the critically overlooked and understudied problem of cross-reactivity in antibodies. Cross reactivity refers to the reactivity of an observed agent which initiates reactions outside the main reaction expected. In other words, the antibody is targeting another antigen with a similar structure rather than the one the researchers are targeting. Obviously, this would be a problem when trying to determine specificity which deals with an antibody binding to the correct target only. This leads to false results and wrong conclusions:

Cross-reactivity in antibody microarrays and multiplexed sandwich assays: shedding light on the dark side of multiplexing

“Immunoassays are indispensable for research and clinical analysis, and following the emergence of the omics paradigm, multiplexing of immunoassays is more needed than ever. Cross-reactivity (CR) in multiplexed immunoassays has been unexpectedly difficult to mitigate, preventing scaling up of multiplexing, limiting assay performance, and resulting in inaccurate and even false results, and wrong conclusions.”

Introduction

“Cross-reactivity (CR) to non-target proteins is ubiquitous and widespread for antibodies (Abs) [1,2-], and together with a lack of Abs against many targets, arguably the biggest obstacle in establishing high performance and large scale multiplexed immunoassays. Unless CR is adequately addressed and suppressed, or at least mitigated, it can be devastating to the performance and reliability of immunoassays. CR is not a mainstream scientific area of research, and rather seen as an impediment, and hence only receives little attention compared to that devoted to the developments of new assay technologies and methods, such as antibody microarrays and high sensitivity assays. Ironically, progress in the development and application of novel assay technologies is often stumped by CR.

Immunoassays depend on an affinity binder — a polyclonal or a monoclonal Ab, a recombinant binder, an aptamer, or a receptor — that binds a target protein with high specificity and affinity. The binding is transduced and amplified into a detectable signal, and in the ideal scenario, the intensity of the signal isratiometric with the concentration of target analyte. Improving immunoassays is predicated on the availability and quality of the affinity binders, and the need for more, better, and cheaper binders is widely recognized [3,4]. CR of affinity binders can be tested using random peptide arrays for example, but CR to non-homologous amino-acid sequences was found to be widespread [5]. The suppression of CR is further complicated by the large parameter space of possible three dimensional conformations adopted by proteins [6]. The Human Protein Atlas has established a polyclonal Ab production pipeline with rigorous quality control standards, and in a Herculean effort, produced Abs against 15 000 of the 20 000 human proteins (Human Protein Atlas; URL: http://www.proteinatlas.org/). Yet, when Schwenk et al. evaluated a preselection of 11,000 affinity-purified, monospecific Abs, only 531 Abs produced a single band on a Western blot, indicating that 95% bound to proteins outside of the expected band [7]. Whereas some of the binding might be ascribed to protein isoforms, cleaved proteins, or post-translational modifications, much is likely caused by CR. Collectively, these studies underline that affinity binders often cross-react.”

“CR and non-specific binding has been studied extensively for single-plex assays, and whereas it may not be possible to eliminate it completely, it is fairly well understood and managed [10]. Dual Ab assays, also called sandwich assays, and often simply referred to as enzyme linked immunosorbent assays (ELISAs) embody an effective strategy to mitigate CR by binding two distinct epitopes on the same protein: a capture Ab (cAb) immobilizes and concentrates the analyte, while the simultaneous binding of a labeled detection Ab (dAb) transduces the binding into a detectable signal (Figure 1c). The strength of the sandwich assay stems from its tolerance to CR because a single CR (or non-specific binding) does not result in a detectable (false positive) signal (Figure 1d).

Indeed, two simultaneous spurious binding events are required to lead to detectable CR, but the odds for it to occur are very low. This point highlights the importance to distinguish between CR that leads to false positive signals and CR that does not lead to a signal, and which can be tolerated, but should not be ignored. In all cases, CR can be further minimized by seeking affinity binders with high specificity and affinity, and by developing assays protocols that minimize CR. For example, binders with low dissociation constants (and low off-binding rates) have long been used, because they can withstand harsh wash steps in ELISA and other assays, while weakly bound and cross-reacting species are washed off [4].”

“In the most favorable case, CR contributes to increased background noise, compromising the LOD of the assay, but in the worst case it generates a false positive signal. The vulnerability is also expressed by the fact that a single contaminated dAb, or an additive in the mixture, can compromise all assays as it interacts with the entire array. We define CR arising because of reagent mixing as reagent-driven CR.”

“End-users of commercial MSAs conducted numerous studies evaluating the performance of various kits, and initial studies with limited multiplexing found good correlations and concluded MSAs to be reproducible [22]. More recently, studies with higher number of targets and a more critical analysis found a lack of reproducibility and correlation between kits from the same [20] and from different vendors [23–26]. In addition, one study found significant differences depending on whether an assay was run in single-plex or in a multiplex format, indicating that MSAs may not be accurate [27]. Because of these findings and additional issues [28-], the use of MSAs with reagent mixing is often not recommended for quantitative analysis and clinical studies [20,23–27,28-]. The authors of these studies are seemingly unaware of the vulnerability to reagent-driven CR, which could explain the observed variability and contradictory conclusions.”

Discussion

“CR is hard to eliminate from immunoassays as Abs are imperfect and often a ‘black box’, yet assays with ever more multiplexing and higher sensitivities are sought after. Much of the discussion in this opinion is based on incidental observations of CR and reasoning. Indeed, systematic studies of CR are rare [5,18-] and the source of CR or assay interference has been difficult to identify.”

doi: 10.1016/j.cbpa.2013.11.012

This final study is from August 2020. I included it to show that 15 years after Saper’s plea to the scientific community, his guidelines were still not being followed as the accuracy and validation of antibodies is still of a questionable nature with no standardized guidelines to validate them. It is even estimated that only 52% of labs adopted some or all of the laboratory guidelines that were laid out by another researcher in 2014. Without a central body certifying the results, it falls upon the faulty peer review system to keep the scientific literature “accurate.” The lack of verifying whether the antibodies used actually target the desired agent is the driving factor behind the current reproducibility crisis in research relying on these theoretical particles:

Antibody validation for protein expression on tissue slides: a protocol for immunohistochemistry

“Antibodies play a crucial role in basic research and clinical decision-making. However, there are no standardized algorithms or guidelines to ensure their accuracy and validity. There have been efforts to generate consensus, but, with the exception of clinical labs, antibody validation remains variable in the literature and sometimes in clinical practice. Here we focus on immunohistochemistry, an example of a scientific and clinical tool where validation of antibodies is critical. We describe a protocol that we use to validate antibodies specifically for immunohistochemistry, including some of the pillars of antibody validation from Uhlen et al. 2016, as an example of a rigorous approach to build antibody-based tests for both basic and translational science labs and for the clinic.”

“Data from the Human Protein Atlas indicate that at least 50% of over 2500 commercially available antibodies did not perform as expected in their intended assay [2]. Improperly validated antibodies, such as several which target estrogen receptor β, have led to nonrigorous research in promising fields.”

“The FDA defines validation as ‘the collection and evaluation of data, from the process design stage through commercial production, which establishes scientific evidence that a process is capable of consistently delivering quality product’. In 2014, Fitzgibbons et al. developed laboratory practice guidelines for analytical validation and revalidation of IHC assays used in anatomic pathology clinical services [6]. Still, it is estimated that only 52% of laboratories have adopted some or all of the recommendations [7]. Sfanos et al. explained that much of the ‘reproducibility crisis’ involved with antibody use in IHC stems from end users being unaware of the need to properly validate an antibody [8].”

“For use in an IHC assay, an antibody must be highly sensitive and also highly specific to the target antigen. The best antibodies have a very high affinity and very low cross-reactivity. It is also beneficial for them to have a fast on-rate and slow off-rate. However, even with recombinant methods, it is hard and expensive to make a perfect antibody, especially for use in IHC; these assays in particular involve unique conditions for antigens, as tissue fixation can hide epitopes exposed in native or denatured forms and expose epitopes that are not exposed when the protein is in its native form in vivo [10,11].

For long-term use in a diagnostic test, and even for reproducible scientific studies, monoclonal antibodies are the best choice. Polyclonal antibodies are produced by several B-cell clones and essentially represent a pool of different antibody clones, each binding to a distinct epitope on the target antigen. While they may be good for rapid proof-of-concept studies, they represent a batch-specific mix that will ultimately be exhausted and is then impossible to reproduce exactly. Monoclonal antibodies are produced by identical B cell clones from a single parent cell and bind to a single, definable epitope on the target antigen; thus they are highly specific and consistent between experiments. Furthermore, with recombinant technologies, they can be optimized and exactly reproduced. Hence, for scientifically rigorous work, or for clinically useful diagnostic assays, a monoclonal antibody is required. As has been shown, however, monoclonal antibodies directed at a target are not all devoid of their own specificity and reproducibility issues, and care must be taken in selecting the best monoclonal antibody, which should always be validated before use [12,13].”

“Going even further, it has been proposed that because the specific sequence for each antibody can be defined and reported, disclosure of the sequence of the epitope would allow for unambiguous knowledge of the antibody being used [42]. Some authors note that a strength of recombinant antibody approaches is that sequences would be more easily reportable and would streamline validation efforts [36,43]. While reporting sequences seems like a good idea from a scientific perspective, many commercial vendors point out that the epitope is the ‘secret sauce’ that makes their antibody better than other competitors. Rather than try to patent each epitope (which is not economically feasible), they maintain confidentiality about their epitope and its sequence [44].”

“Because research labs in the academic sector are not subject to any certification and peer review is the only ‘regulatory body’ in that sphere, it is perhaps unrealistic to expect strict adherence to any standardization requirements.”

“Another key concept likely to be important in IHC in the future is antibody biochemistry. In general, the affinity and dissociation constants for most monoclonal antibodies are not known or not easily available. Optimal antibodies for IHC have a fast on-rate and slow off-rate, but the kinetics of antibody binding for IHC are essentially never published.”

https://www.future-science.com/doi/10.2144/btn-2020-0095

In Summary:

• According to antibody expert Clifford Saper: “No, there is no such thing as a monoclonal antibody that, because it is monoclonal, recognizes only one protein or only one virus. It will bind to any protein having the same (or a very similar) sequence.”
• A 2015 article highlighting Saper stated antibodies used in research often give murky results
• A mouse first alerted Clifford Saper to the fact that antibodies were misleading the scientific community
• Results using knockout mice were unsettling as antibody staining in knockout animals should have shown radically different patterns from those in unmodified animals, but all too often the images were identical
• Due to numerous retractions, Saper and his editorial colleagues set up a policy of requiring extensive validation data on each antibody
• The policy was good for rigour, but not submissions as many researchers looked elsewhere to publish results
• Biomedical researchers still collect tales of antibody woe 
• The most common grumble is the cheating reagent: the antibody purchased to detect protein X surreptitiously binds protein Y (and perhaps ignores X altogether)
• Another complaint is ‘lost treasure’: a run of promising experiments that stalls when a new batch of antibodies fails to reproduce previous findings
• It is alarming to discover that antibodies can be unreliable reagents
• Insufficient specificity, sensitivity and lot-to-lot consistency have resulted in false findings and wasted efforts
• Antibody unreliability has taken its toll across studies in:
  1. Cancer
  2. Metabolism
  3. Ageing
  4. Immunology and cell signalling
  5. Any field concerned with researching complex biomolecules
• Losses from purchasing poorly characterized antibodies have been estimated at $800 million per year, not counting the impact of false conclusions, uninterpretable (or misinterpreted) experiments, wasted patient samples and fruitless research time

• In an open letter Saper wrote in The Journal of Comparative Neurology in 2005, he stated that the use of immunohistochemistry and the employment of a battery of ten or more antibodies to examine issues of colocalization or cell typing has resulted in a flood of misinformation
• The Journal received numerous distressed communications from authors who had to withdraw papers because an antibody against a novel marker was found to stain tissue in knockout animals, who lack that target protein
• In many cases, these papers contained careful characterization of the antibodies and immunocytochemical controls
• The editors felt that the integrity of scientific communication was being threatened by the proliferation of poorly characterized antibodies that produce artifactual staining patterns
• There was confusion about how to provide a reasonable level of assurance that an antibody is actually recognizing what it is supposed to be staining
• They decided three basic criteria needed to describe antibodies in papers
  1. Complete information on the antibody.
     • It is important to recognize that antibodies are not simple reagents that always identify the same thing
     • Identification: What was the source of the antibody?
     • Preparation of the antibody: What was the antibody actually raised against?
     • Some antibody manufacturers deliberately try to obscure the information about the structure of the antigen
     • They received complaints from authors that several manufacturers have claimed that sequence information on the antigen was “proprietary,” and they would not provide it
     • Because work using these antibodies is inherently not repeatable, such papers are not acceptable for publication
  2. How has the specificity of the antibody been characterized?
     • Ideally, an antibody should stain a single band (or several bands if the antigen has several known molecular configurations) of appropriate molecular weight
     • It is not sufficient just to state “this antiserum was previously characterized (Jones et al., 2002)”
  3. What controls are necessary for immunostaining?
     • They are constantly surprised by the oddly trusting nature of many of their colleagues, who seem to believe (or want to believe) that an antibody will stain what the manufacturer claims yet, in many cases, nothing could be further from the truth
     • The gold standard in this regard is: does the antibody stain the tissue of interest from which the molecule of interest has been removed?
     • This is only applicable if you are looking at tissue from an animal for which a knockout exists
     • Saper admits most of them, most of the time, must make due with lesser degrees of certainty
     • This does not protect from other tissue proteins that may cross-react and they have had a number of papers eventually retracted when an anti-serum that passed the preadsorption test also were shown to stain tissue from knockout animals
     • This strategy cannot be used for monoclonal antibodies, which always will be adsorbed out by their antigen, even if they are staining something entirely different in the tissue
     • He notes that omission controls (staining without the primary antibody) are NOT controls for specificity of the primary antibody at all
• Saper concludes that it is critical that for each antibody used, the authors must supply sufficient information to assure that the result is replicable and likely to be correct
• Blind faith that the antibody will stain whatever the manufacturer claims is not consistent with good science

• Many investigators are unaware of the potential problems with specificity of antibodies and the need to document antibody characterization meticulously for each antibody that is used
• Immunohistochemical staining has become so mundane it is often assumed that IHC has an analogous ability to identify molecular targets accurately but nothing could be further from the truth
• IHC methods remain primitive in terms of both sensitivity and specificity
• The fundamental principles on which antibody localization is based have not improved at all in the last two decades, and if anything, the slope occupied by IHC has become more slippery than ever
• In monoclonal antibodies (laboratory-made proteins made from cancer cells that mimic the immune system’s ability to fight off harmful pathogens such as “viruses“), it is possible that the variable site may bind to a variety of different targets, and that these may be quite different from the molecule against which the antibody was raised
• Monoclonal antibodies are derived from hybridoma cells (hybrid cells produced by the fusion of an antibody-producing lymphocyte with a tumor cell) which are grown either in culture or by injecting them intraperitoneally in a host animal
• Either the culture fluid or ascites fluid from animals containing the antibodies can be subjected to purification by precipitating the antibodies with protein A and the resulting relatively pure antibody preparations (i.e. not pure) are quantified based on the micrograms of protein
• Polyclonal antisera, in contrast, are derived by bleeding animals a few weeks after they have been immunized and usually several “booster immunizations” are given with several bleeds taken
• When boosting the same animal with repeated immunizations with the same antigen, the antibody content in sequential bleeds may differ markedly
• The lot for a polyclonal antiserum is critical, and even another batch from the same animal may have entirely different staining properties
• The ability to create synthetic proteins and peptides has “revolutionized” the way in which antibodies may be made and how they can be characterized
• When the antibody binds to a partial sequence or a partial sequence competes against binding to the native molecule, the epitope, or structural features that the antibody recognizes, is presumed (i.e. suppose that something is the case on the basis of probability) to be located in that sequence
• This method is used to map the epitope that the antibody binds, however, this does not indicate what the sequence was of the original immunogen, because the antibody may have been made against an overlapping sequence
• When the two different antibodies stain exactly the same pattern, it is highly likely (i.e not 100% certain) that they are staining the correct target
• Specifically altered synthetic antibodies prepared in this way may be able to distinguish between different modified forms of the same molecule with great accuracy
• Most investigators want to use antibodies to localize cellular components and do not want to have to become experts in immunology or IHC to do so
• Saper states it is useful to have a set of criteria for what constitutes a reasonable degree of assurance that the antibody being used is actually targeting its correct antigen
• If all investigators followed the rules, the literature would be much more accurate
• The questions in need of asking are:
  1. What Immunogen Is Used to Raise the Antibody?
     • A key principle of science is that the work must be repeatable
     • If the antibody is raised against a “proprietary” antigen (usually a secret amino acid sequence, to avoid competitors from copying the product), it simply is not valid for serious scientific work
  2. What is the Evidence That the Antibody Binds Specifically to the Expected Target Molecule in the Tissue of Interest?
     • Researchers should obtain at least reasonable evidence that the antibody does bind to its expected target in the tissue in which it will be studied and not to something else
     • If extraneous bands are stained, this indicates that the antibody has other additional targets in the tissue and should raise red flags against using that antibody for IHC
     • It is quite possible for the antibody to see only one band in some tissues but to see multiple extraneous bands in other tissues from the same animal
     • Similarly, manufacturers often try to “prove” specificity by running the antibody against a gel preparation of purified or recombinant protein which may show that the antibody can bind to its target but does not tell anything about what else it may bind to in tissue
  3. What Controls Can Be Done to Insure That the Antibody Binds in Fixed Tissue Only to Its Target Molecule?
     • In the fixed state, it is possible that the antibody that works well in a Western blot will find that its target antigen is distorted by the fixation process and no longer recognizable
     • When polyclonal antisera are raised against a peptide antigen, it is common that most of the antisera that are produced will stain fixed tissue poorly or not at all
     • One of the best controls is to transfect the DNA for the target protein into cells that normally do not make it in tissue culture, however, this control does not prove that the antibody will only stain its target in the tissue of interest
     • Another control for specific staining in tissue is the preadsorption test, which involves mixing the diluted antibody with an excess of the immunogen which should completely block staining
     • This shows that the staining in the tissue is against something that is at least cross-reactive with the original protein (although it does not prove that this is what the target in the tissue actually is)
     • The preadsorption control is meaningless for a monoclonal antibody (which is produced by screening for its binding to the target, and therefore will always bind it and always pass a preadsorption test, by definition) and for antibodies that have already been affinity purified (for the same reason)
     • The absence of staining is a strong confirmation of specificity but unfortunately this is not a perfect test because the target that is stained in the tissue may be a related molecule that is downregulated in the knockout animal
     • In addition, this approach only applies to situations where there is a knockout strategy available, which limits it to a few model species
     • Finally, in many so-called knockout mice, the original protein is not entirely eliminated and if only a portion of that protein is still expressed, it may have no functional presence but still stain with the antibody
     • When two antibodies are made in the same species, showing that the staining patterns are very similar (i.e. not identical) is a strong control
• Saper concludes that we are always uncovering new ways that nature can fool us and thus, no antibody localization is really perfect

• Cross-reactivity (CR) in multiplexed immunoassays has been unexpectedly difficult to mitigate, preventing scaling up of multiplexing, limiting assay performance, and resulting in inaccurate and even false results, and wrong conclusions
• Cross-reactivity (CR) to non-target proteins is ubiquitous and widespread for antibodies and together with a lack of Abs against many targets, arguably the biggest obstacle in establishing high performance and large scale multiplexed immunoassays
• Unless CR is adequately addressed and suppressed, or at least mitigated, it can be devastating to the performance and reliability of immunoassays
• CR to non-homologous amino-acid sequences was found to be widespread
• The Human Protein Atlas has established a polyclonal Ab production pipeline with rigorous quality control standards, and in a Herculean effort, produced Abs against 15,000 of the 20,000 human proteins (Human Protein Atlas; URL: http://www.proteinatlas.org/)
• Yet, when Schwenk et al. evaluated a preselection of 11,000 affinity-purified, monospecific Abs, only 531 Abs produced a single band on a Western blot, indicating that 95% bound to proteins outside of the expected band
• Collectively, these studies underline that affinity binders often cross-react
• CR and non-specific binding has been studied extensively for single-plex assays and it may not be possible to eliminate it completely
• Two simultaneous spurious binding events are required to lead to detectable CR, but the odds for it to occur are very low
• This point highlights the importance to distinguish between CR that leads to false positive signals and CR that does not lead to a signal, and which can be tolerated, (why should CR be tolerated?) but should not be ignored
• In the most favorable case, CR contributes to increased background noise, compromising the LOD of the assay, but in the worst case it generates a false positive signal
• The vulnerability is also expressed by the fact that a single contaminated dAb, or an additive in the mixture, can compromise all assays as it interacts with the entire array
• Studies with higher number of targets and a more critical analysis found a lack of reproducibility and correlation between kits from the same and from different vendors
• In addition, one study found significant differences depending on whether an assay was run in single-plex or in a multiplex format, indicating that MSAs may not be accurate
• Because of these findings and additional issues, the use of MSAs with reagent mixing is often not recommended for quantitative analysis and clinical studies
• The authors of these studies are seemingly unaware of the vulnerability to reagent-driven CR, which could explain the observed variability and contradictory conclusions
• CR is hard to eliminate from immunoassays as antibodies are imperfect and often a ‘black box’
• Systematic studies of CR are rare and the source of CR or assay interference has been difficult to identify

• There are no standardized algorithms or guidelines to ensure antibody accuracy and validity
• There have been efforts to generate consensus, but, with the exception of clinical labs, antibody validation remains variable in the literature and sometimes in clinical practice
• Data from the Human Protein Atlas indicate that at least 50% of over 2500 commercially available antibodies did not perform as expected in their intended assay
• The FDA defines validation as ‘the collection and evaluation of data, from the process design stage through commercial production, which establishes scientific evidence that a process is capable of consistently delivering quality product‘
• Quality guidelines were established in 2014 yet it is estimated that only 52% of laboratories have adopted some or all of the recommendations
• Sfanos et al. explained that much of the ‘reproducibility crisis’ involved with antibody use in IHC stems from end users being unaware of the need to properly validate an antibody
• The best antibodies have a very high affinity and very low cross-reactivity
• Even with recombinant methods, it is hard and expensive to make a perfect antibody, especially for use in IHC
• Polyclonal antibodies may be good for rapid proof-of-concept studies, but they represent a batch-specific mix that will ultimately be exhausted and is then impossible to reproduce exactly
• Monoclonal antibodies are said to be best for long-term diagnostic use and reproducible studies, however, monoclonal antibodies directed at a target are not all devoid of their own specificity and reproducibility issues
• In other words, the synthetically-created from cancer cells monoclonal antibodies are “better” than the cultured lab-created from animal cells polyclonal antibodies yet both have specificity/reproducibility problems
• Reporting sequences has been proposed as validation yet many commercial vendors point out that the epitope is the ‘secret sauce’ that makes their antibody better than other competitors and keep the sequences confidential
• Because research labs in the academic sector are not subject to any certification and peer review is the only ‘regulatory body’ in that sphere, it is unrealistic to expect strict adherence to any standardization requirements
• The affinity and dissociation constants for most monoclonal antibodies are not known or not easily available
• Optimal antibodies for IHC have a fast on-rate and slow off-rate, but the kinetics of antibody binding for IHC are essentially never published

If the evidence continually shows that antibodies are not specific as is commonly claimed by the scientific community and that they cross-react with other targets, how can any measurements gained from their use be considered valid? If polyclonal antibodies are only as good as the current batch and the results can not be reproduced once it is used up, what value do they have as a research tool? If monoclonal antibodies are said to be more reliable yet they can not bind to just one “virus” and regularly target other proteins of a similar structure, how can the results be trusted?

These questions are easy to answer once it is understood that antibodies still remain unseen theoretical particles that are assumed to be present in the body and have been assigned unproven form and functions to fit the mold of the prevalent paradigm. There are no ways in which the researchers can directly see these tiny proteins so they rely on using indirect chemical reactions such as immunohistochemical staining, haemagglutination inhibition tests, neutralization assays, complement fixation tests, ELISA tests, etc. in order to claim that the resulting reactions are proof antibodies are present. They claim these measurements have meaning yet all of these measures are utterly useless and have consistently been found to be unreliable.

Contrary to the popular claims of the scientific community, antibodies are not specific. This is why an antibody test for HIV can return a positive result for over 70 other conditions (such as tuberculosis, pregnancy, or even the flu vaccine) that are said to trigger the same antibody response. The results are inaccurate. Next time they try to state that “specific” antibodies are targeting certain “viruses” or their “spike proteins,” hopefully you will now know better and realize that this is never the case.