Commercially available antibodies against human and murine histamine H4-receptor lack specificity
- First Online:
- Cite this article as:
- Beermann, S., Seifert, R. & Neumann, D. Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385: 125. doi:10.1007/s00210-011-0700-4
- 246 Views
Antibodies are important tools to detect expression and localization of proteins within the living cell. However, for a series of commercially available antibodies which are supposed to recognize G-protein-coupled receptors (GPCR), lack of specificity has been described. In recent publications, antisera against the histamine H4-receptor (H4R), which is a member of the GPCR family, have been used to demonstrate receptor expression. However, a comprehensive characterization of these antisera has not been performed yet. Therefore, the purpose of our study was to evaluate the specificity of three commercially available H4R antibodies. Sf9 insect cells and HEK293 cells expressing recombinant murine and human H4R, spleen cells obtained from H4R−/− and from wild-type mice, and human CD20+ and CD20− peripheral blood cells were analyzed by flow cytometry and Western blot using three commercially available H4R antibodies. Our results show that all tested H4R antibodies bind to virtually all cells, independently of the expression of H4R, thus in an unspecific fashion. Also in Western blot, the H4R antibodies do not bind to the specified protein. Our data underscore the importance of stringent evaluation of antibodies using valid controls, such as cells of H4R−/− mice, to show true receptor expression and antigen specificity. Improved validation of commercially available antibodies prior to release to the market would avoid time-consuming and expensive validation assays by the user.
KeywordsAntibody specificityHistamine receptorFlow cytometryWestern blotting
The biogenic amine histamine is an important local mediator and neurotransmitter. Histamine is generated in vivo by decarboxylation of the amino acid l-histidine. It is produced, e.g., by mast cells, where it is stored in an inactive form in cellular granules bound to heparin, but in low amounts also in basophils, leukocytes, and thrombocytes. In an acute allergic reaction, histamine is released from intracellular stores of mast cells, leading to vasodilatation and increased vascular permeability (Hill et al. 1997; Thurmond et al. 2008; Leurs et al. 2011).
On the target cells, histamine acts via histamine receptors, of which four different subtypes (H1R–H4R) have been identified so far and which all belong to the G-protein-coupled receptors (GPCR) family (Hill et al. 1997; Hough 2001; Liu et al. 2001b; Leurs et al. 2011; Lieberman 2011). The H4R is the most recently discovered histamine receptor, which was identified independently by several research groups (Nakamura et al. 2000; Oda et al. 2000; Liu et al. 2001b; Morse et al. 2001; Zhu et al. 2001). Compared to the other histamine receptors, the H4R shows a homology of 37% to the H3R and only a ∼19% homology to the H1R and H2R. The sequence homology of the human H4R compared to other species ranges between 67% and 72% (Liu et al. 2001b; Leurs et al. 2009). The sequence of the H4R encodes a 390 amino acid protein which forms seven transmembrane helices. By coupling to pertussis toxin-sensitive Gαi/0 proteins, activation of the H4R leads to inhibition of forskolin-induced cAMP production (Oda et al. 2000), increase in extracellular signal-related kinase phosphorylation (Morse et al. 2001), and calcium elevation via activation of phospholipase C (Hofstra et al. 2003).
The H4R is expressed at low levels in most tissues. By reverse transcriptase-PCR analysis, the highest levels of the receptor were detected in spleen, lung, liver, bone marrow, thymus, small intestine, colon, heart, lymph node, kidney, and peripheral blood (Nakamura et al. 2000; Cogé et al. 2001; Liu et al. 2001a; Zhu et al. 2001; de Esch et al. 2005). At the cellular level, H4R mRNA was found in several immune cells such as mast cells, T cells, dendritic cells, monocytes, eosinophils, neutrophils, and basophils (Liu et al. 2001a; Morse et al. 2001; Zhu et al. 2001; Hofstra et al. 2003).
Since mRNA occurrence and protein expression at the cell surface do not necessarily correlate (de Sousa Abreu et al. 2009), receptor detection at the protein level is essential for functional expression studies. Expression of recombinant receptor proteins, which are genetically modified with an epitope-tag, is easily detectable using antibodies against the corresponding eptitope-tag (Klein et al. 1997; Jongsma et al. 2007). However, for the detection of endogenous receptor expression, ligands labeled by, e.g., isotopes or fluorescent groups (Myslivecek et al. 2008; Bylund and Toews 2011; Keller et al. 2011), or functional assays are commonly used. Unfortunately, not for all receptors, selective labeled ligands are available (Vrydag and Michel 2007). Another possibility to detect the receptor protein itself, without any ligand, is the use of specific antibodies. These antibodies would also be helpful to investigate the expression density of the receptor, the subcellular localization, and receptor internalization (Michel et al. 2009).
For detection of the H4R, several commercial antisera are available. Although these antibodies are not well characterized, there are several published studies using them (van Rijn et al. 2006; Bäumer et al. 2008; Dijkstra et al. 2008; Gschwandtner et al. 2010; Lethbridge and Chazot 2010). In most cases, no intrinsic controls, such as cells from receptor knockout animals, receptor-selective siRNA-treated cells, cells transfected with the targeted and related receptors, or use of various antibodies against different epitopes of the receptor or an added epitope-tag (Miyauchi et al. 2009), were performed to examine their specificity. Therefore, in this study, we investigated the quality of three commercially available H4R antibodies using different methods and systems, including H4R−/− mice, expression of tagged receptors, and different receptor subtypes in HEK293 and Sf9 cells.
H4R-deficient (H4R−/−) mice were kindly provided by Dr. Robin Thurmond (Johnson & Johnson Research and Development, La Jolla, CA, USA). BALB/c wild-type mice were obtained from Janvier (Le Genest, France). Baculoviruses encoding mH4R, hH4R, and hH3R were prepared as described (Schneider et al. 2009; Schnell et al. 2010; Schnell et al. 2011). The cDNAs of mH4R and hH4R were labeled N-terminally by a Flag-epitope and C-terminally by a hexahistidine (His6)-tag and cloned into pcDNA3.1. If not stated otherwise, all chemicals were obtained from Sigma-Aldrich (Munich, Germany).
The following antibodies were used: anti-H4R sc M120 (Santa Cruz Biotechnology, Santa Cruz, CA, USA): rabbit polyclonal antibody raised against amino acids 194–313, mapping within a region in the third intracellular loop of the human H4R, tested to recognize the human, murine, and rat isoforms, concentration: 200 μg/ml (http://www.scbt.com/datasheet-50313-histamine-h4-receptor-m-120--antibody.html); anti-H4R sc H110 (Santa Cruz Biotechnology): rabbit polyclonal antibody raised against amino acids 194–303, mapping within a region in the third intracellular loop of the human H4R, tested for the human protein only, concentration: 200 μg/ml (http://www.scbt.com/datasheet-50313-histamine-h4-receptor-h-110--antibody.html; both Santa Cruz antibodies are intended for the use in Western blot, immunoprecipitation, isoelectric focusing, and solid-phase ELISA); anti-H4R ab13183 (Abcam, Cambridge, UK): rabbit polyclonal antibody raised against a peptide from the first cytoplasmic domain of the human H4R, tested for the human protein only, concentration: 1 mg/ml (tested for the application in immunohistochemistry of formalin/paraformalin-fixed and paraffin-embedded tissues only; meanwhile withdrawn from the market); anti-Flag M2, biotinylated (Sigma-Aldrich, Munich, Germany): mouse monoclonal, concentration: 3-mg/ml anti-His6 (Dianova, Hamburg, Germany): mouse monoclonal, concentration: 200 μg/ml; anti-CD20 PE (Miltenyi Biotech, Bergisch Gladbach, Germany): concentration not reported.
Streptavidin PE (Jackson ImmunoResearch, Suffolk, UK); goat anti-mouse fluorescein isothiocyanate (FITC) (Jackson ImmunoReaserch); goat anti-rabbit FITC (Sigma-Aldrich); streptavidin horseradish peroxidase (HRP) (Jackson ImmunoReaserch); and goat anti-rabbit HRP (Dako, Hamburg, Germany) were used as secondary antibodies.
RNA extraction and RT-PCR
Total RNA was extracted from 5 × 107 cells using the NucleoSpin RNA II kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s protocol. Two-microgram RNA was reverse-transcribed into cDNA using oligo dT primers (Fermentas, Meridian, Rockford, USA) and RevertAid Reverse Transcriptase (Fermentas).
Specific sequences of the mH4R were amplified by 35 cycles of PCR. Primer sequences (MWG Biotech, Ebersberg, Germany) were as follows: mH4R forward 5′-GGCTATTTCTGACTTCCTCG-3′ and mH4R reverse 5′-ACTGACTGGTATCGATCGTA-3′. The resulting products were separated in 2% (w/v) agarose gels and visualized by ethidium bromide staining.
Infection of Sf9 cells
Sf9 cells were grown in suspension culture in single-use polycarbonate Erlenmeyer flasks at 28°C under rotation at 150 rpm. The cells were maintained in Insect Xpress Medium with 5% (v/v) fetal calf serum (FCS) and 0.1 mg/ml gentamycin (all from Lonza Walkersville, MD, USA) at a density of 0.5–3 × 106 cells/ml. High-titer baculovirus stocks encoding recombinant proteins were generated in Sf9 cells using the BaculoGOLD transfection kit (BD PharMingen, San Diego, CA, USA) according to the manufacturer’s instruction. For infection, Sf9 cells were sedimented by centrifugation, suspended in fresh medium, and seeded at 3 × 106 cells/ml (Schneider and Seifert 2010a; Schneider and Seifert 2010b; Schneider and Seifert 2010c). Baculovirus stocks encoding mH4R, hH4R, hH3R, and empty bacolovirus (mock) were used at a dilution of 1:100 to infect the cells. The cells were cultured 48 h before they were used for flow cytometric analysis or membrane preparation. At this time, the infection was checked by light microscopy. Most of the cells showed the typical signs of infection (e.g., altered morphology and size) and were still intact (Schneider and Seifert 2010c).
Transient transfection of HEK29 cells
HEK293 cells were maintained in DMEM medium (PAA Laboratories, Pasching, Austria) supplemented with 10% (v/v) FCS (Lonza), penicillin/streptomycin, and l-glutamine at 37°C in a 7% (v/v) CO2 environment. For transfection, the cells were seeded with 3 × 105 cells/ml and 2 ml/well in 6-well plates 24 h before transfection. Cells were grown to 60–80% confluency, and transfection was performed with Fugene HD (Roche Diagnostics, Manheim, Germany) according to the manufacturer’s protocol. Following transfection, the cells were incubated for another 48 h before being used for FACS analysis or lysed for Western blotting.
Preparation of spleen cell suspension
Mice were killed according to §4 of the German Animal Welfare Protection Act after prior report to the State Government of Lower Saxony. Spleens were excised and processed to single cell suspension in RPMI containing 5% (v/v) FCS (Lonza), penicillin/streptomycin, l-glutamine, nonessential amino acids, pyruvate, and β-mercaptoethanol. Cell suspensions were transferred into tubes and kept on ice for 10 min for sedimentation of crude pieces of the tissue. The supernatant containing single cells was removed and centrifuged. Lysis of erythrocytes was performed by suspending the cell pellet in ACK solution (155 mMNH4Cl, 10 mM KHCO3) and incubation for 10 min at room temperature. Cells were washed twice in RPMI and afterwards used for FACS analysis.
Isolation of PBMCs from human blood
Fresh human blood from healthy volunteers was diluted 1:2 in phosphate-buffered saline (PBS). Blood donation was approved by the ethics committee of the Hannover Medical School. The mixture was carefully placed on a Ficoll layer (Biochrom, Berlin, Germany), followed by centrifugation at 400×g for 20 min. The resulting leukocyte layer was transferred into another tube and washed twice with RPMI containing 5% FCS (Lonza), penicillin/streptomycin, l-glutamine, nonessential amino acids, pyruvate, and β-mercaptoethanol by centrifugation at 400×g for 10 min.
Extracellular flow cytometry staining
For each sample, 1 × 106 cells were washed twice (400 g, 10 min) in AutoMacs Running buffer (Miltenyi Biotech, Bergisch Gladbach, Germany) and then incubated for 15 min in AutoMacs Running buffer containing a previously determined optimal amount [20% (v/v)] of supernatant fluid of 2.4G2 hybridoma cells in order to block Fc receptors. Expression of recombinant receptors in infected Sf9 and transfected HEK293 cells was verified by staining with anti-Flag (1:100) followed by streptavidin PE (1:100), both for 30 min at 4°C. To distinguish B cells from other human leukocytes within the peripheral blood mononuclear cell (PBMC) preparation, cells were stained with anti-CD20 PE (1:10) for 30 min at 4°C.
Intracellular flow cytometry staining
For each sample, 1 × 106 cells were washed and blocked as described above, followed by fixation in 4% (w/v) paraformaldehyde in PBS for 10 min at 4°C. After permeabilization with PBS containing 0.5% (w/v) BSA, 0.01% NaN3 (w/v), and 0.5% (w/v) saponin, recombinant receptors in Sf9 and HEK293 cells were stained with mouse anti-His6-tag (1:50) followed by goat anti-mouse FITC (1:500) for 1 h at RT to verify receptor expression. Intracellular staining with the H4R antibodies sc M120, sc H110, and ab 13283 at multiple concentrations or rabbit IgG isotype control at corresponding concentrations was performed for 1 h at RT followed by detection with goat anti-rabbit FITC (1:500). Stained cells were analyzed by flow cytometry (FACScan, Becton Dickinson, Heidelberg, Germany or MacsQuant, Mitenyi Biotech). Reported are data obtained with antibodies applied at concentrations of 4 μg/ml (sc M120 and sc H110) and 20 μg/ml (ab 13283), which did not differ from data obtained with other concentrations.
Membrane preparation of Sf9 cells
Sf9 membranes were prepared as described (Schneider et al. 2009), using 1 mM ethylenediaminetetraacetic acid (EDTA), 0.2 mM phenylmethylsulfonyl fluoride, 10 μg/ml benzamidine, and 10 μg/ml leupeptin as protease inhibitors. The membranes were suspended in binding buffer (12.5 mM MgCl2, 1 mM EDTA, and 75 mM Tris/HCl, pH 7.4) and stored at −80°C until use.
Generation of HEK293 cell lysates
Transfected HEK293 cells were washed once with PBS and lysed directly in the P6 plate with lysis buffer, containing 0.5% (v/v) NP-40, 50 mM Hepes, 250 mM NaCl, 1 mM EDTA, 10% (v/v) glycerol, 20 mM β-glycerophosphate, 5 mM p-nitrophenylphophate, 100 mM Na-orthovanadate, 500 mM dithiothreitol, and complete protease-inhibitor cocktail (Roche, Mannheim), pH 7.6. Lysates were transferred into tubes and incubated for 1 h under rotation at 4°C to allow complete cell lysis. Afterwards, the cells were treated with ultrasound for 15 s to disrupt the membranes. Total protein concentration was determined by Bradford protein assay (BioRad Laboratories, Munich, Germany) in a dilution of 1:200 in PBS.
SDS-PAGE and immunoblot analysis
Proteins were diluted in Laemmli buffer without boiling. One hundred micrograms of HEK293 cell lysate and 40 μg of membrane proteins of Sf9 cells per lane were separated on a 10% (w/v) Sodium dodecyl sulfate (SDS) polyacrylamide gel. Separated proteins were transferred onto a PVDF membrane (Pierce, Rockford, IL, USA) and the membrane was blocked with 5% (w/v) milk powder in TBS/T (Tris, borate, SDS, Triton X-100). Recombinantly expressed H4R was identified by staining with anti-Flag (1:500) overnight followed by streptavidin HRP (1:10,000) for 2 h at 4°C. Identical blots were stained either with H4R antibodies sc M120, sc H110, or with rabbit IgG isotype control. Bound primary antibodies were detected with goat anti rabbit HRP (1:2,000) and electrochemoluminescence, followed by exposition to radiography films (Kodak, Stuttgart, Germany). Differing concentrations of primary antibodies were examined, however, with identical results. Reported are results obtained by applying the primary antibodies or rabbit IgG iosotype control at a concentration of 2 μg/ml.
FACS data were analyzed with CellQuest Pro Software (Becton Dickinson) or MacsQuantify Version 2.2 (Miltenyi Biotech).
Evaluation of the specificity of H4R antibodies by flow cytometric analysis of infected Sf9 cells
Evaluation of the specificity of H4R antibodies by Western blot analysis of infected Sf9 cells
Evaluation of the specificity of H4R antibodies by flow cytometric analysis of transfected HEK293 cells
Evaluation of the specificity of H4R antibodies by Western blot analysis of transfected HEK293 cells
By comparison of the anti-H4R Western blot results obtained with Sf9 insect cells and with HEK293 cells, marked differences appeared. In these analyses, the Sf9 cells (Fig. 3) lead to only very faint bands while the Western blots of HEK293 cells resulted in multiple intense bands (Fig. 5), although comparable exposition times were used. The reduction of such bands in Fig. 3 may be due to the cellular expression system used, Sf9 insect cells, probably displaying membrane structures the antibodies do not cross-react with.
Evaluation of the specificity of H4R antibodies by flow cytometric analysis of splenocytes from wild-type and H4R−/− mice
Evaluation of the specificity of H4R antibodies by flow cytometric analysis of human PBMCs
We investigated the specificity of three commercially available H4R polyclonal antibodies in various cellular systems using different parameters. The H4R antibodies were raised against peptides with amino acid sequences obtained from intracellular loops of the human H4R. The Santa Cruz antibodies sc M120 and sc H110 have been raised against two very similar peptides, thus they are expected to behave very similarly. In addition, the immunizing peptides were rather long, theoretically enhancing the probability to obtain antisera recognizing the respective full-length proteins. However, in the case of membrane-integrated proteins such as receptors, the secondary and tertiary structures are influenced by their interactions with membrane components, probably resulting in epitopes different to those displayed in the corresponding soluble peptides. For all these antibodies, their reactivity against the human protein was stated in the data sheets, and only one, sc M120, was indicated to recognize the mouse and the rat proteins, too (http://datasheets.scbt.com/sc-50313.pdf; http://datasheets.scbt.com/sc-50312.pdf). The intended applications of the Santa Cruz antibodies sc M120 and sc H110 were Western blot, immunoprecipitation, isoelectric focusing, and solid-phase ELISA. Their use in other applications was not analyzed. The abcam antibody ab 13183 was tested in immunohistochemistry of formalin/paraformalin-fixed and paraffin-embedded tissues only.
In flow cytometric analysis of baculovirus-infected Sf9 cells as well as of transfected HEK293 cells, the complete H4R expression from N-terminus to C-terminus could be proved by use of Flag and His6 antibodies. Staining with anti-Flag and anti-His6 was also used to verify efficient extra- and intracellular detection, respectively. In comparison to the respective controls, anti-Flag- and anti-His6-stained infected/transfected cells showed a distinct increase in fluorescence intensity, resulting in two populations, one with fluorescence intensity similar to untransfected cells and a second with a shifted fluorescence peak. This pattern indicates that after infection/transfection the resulting population is a mixed one, consisting of cells which do not and cells which do express recombinant receptors, respectively (Figs. 2a–h and and 4a–f).
The samples stained with anti H4R-antibodies also showed a signal with higher fluorescence intensity as compared to isotype or unstained controls. However, this increase was independent of infected/transfected receptors, since there was almost an identical signal in mH4R-, hH4R-, or hH3R-expressing cells as well as in control cells. Moreover, after anti-H4R staining, an increase in fluorescence intensity appeared in the whole population, in contrast to data generated with the Flag and His6 antibodies which demonstrate two separate populations of infected and uninfected cells (Figs. 2i–t and 4g–l). Similar results, cells stained with the H4R antibodies regardless of the presence or absence of the receptor protein, were obtained by analyzing splenocytes from wild-type and H4R−/− mice (Fig. 6) and human non-B and B leukocytes (Fig. 7), of which the latter do not express H4R while it can be detected in non B-cell leukocytes (Hofstra et al. 2003).
In Western blot analysis, none of the bands detected by the H4R antibodies showed the expected size of the H4R, which could be detected using the M2 antibody (Figs. 3 and 5). Collectively, all these findings indicate that the signals obtained by the use of the H4R antibodies do not result from specific binding to the receptor, but rather from unspecific binding to cells. Hence, the three antibodies evaluated in this study are not useful for a specific detection of the H4R, neither in native form (flow cytometric) nor in denatured form (SDS PAGE).
In published data, staining with isotype-matched Ig and preabsorbtion with the cognate peptide is used frequently as the only control (Bäumer et al. 2008; Dijkstra et al. 2008). This procedure led to the conclusion that the antibody specifically detected the receptor, since the fluorescence intensity of the controls was lower than that of the anti-H4R-stained cells. However, as discussed above, such an increase in fluorescence intensity appears also in samples of uninfected/-transfected cells or cells from H4R−/− mice. These results indicate that cells which do not express the respective receptor are a crucial control and that many previously published data need reinterpretation.
The criteria most often used to proof specificity of antisera are the absence of staining in the presence of the peptide against which the antiserum was raised and a band on Western blot of expected size, independent of other unspecific signals (Holmseth et al. 2006; Pradidarcheep et al. 2008). Several groups reported the lack of specificity of different anti-GPCR antibodies, e.g., for muscarinic receptors and β-adrenoreceptors (Pradidarcheep et al. 2009), the α1-adrenergic receptor (Jensen et al. 2009), or dopamine receptors (Bodei et al. 2009). In order to provide evidence for the specificity and selectivity of these antibodies against GPCRs, a correct validation including intrinsic controls is necessary and is explicitly requested by many reviewers of scientific journals.
Pradidarcheep et al. (2008) and Michel et al. (2009) defined four criteria to characterize antibodies of which at least one has to be fulfilled for the antibody being useful. First, specificity can be assumed if staining by a given antibody is absent in cells or tissue from knockout animals for the respective receptor. We tested this criterion using cells from H4R−/− mice and showed lack of specificity of the H4R antibodies (Fig. 6). Second, another indication for receptor specificity is a distinct decrease of staining in cell lines were receptor expression is genetically knocked down, for example by siRNA. We did not analyze this criterion in our study. The third criterion to provide evidence for selectivity is the transfection of several related GPCR subtypes of a given family in the same host cell line. A selective antibody only recognizes the target receptor but not the related subtypes. We tested this criterion by analyzing the H3R which is closely related to the H4R (Fig. 2). Fourth, evidence for receptor specificity of an antibody can be checked by using various antibodies raised against different epitopes of the receptor or an additional antibody against a specific tag after transfection with a tagged receptor (Miyauchi et al. 2009). All antibodies should show an almost similar staining pattern in immunhistochemistry or Western blotting. We tested this criterion by tagging the recombinant receptors with Flag- and His6-epitope tags and using tag-specific antibodies (Figs. 2, 3, 4, and 5). In summary, we tested three of the four criteria proposed (Michel et al. 2009; Pradidarcheep et al. 2009), however, the antibodies fulfilled none of them. Thus, we did not obtain evidence for specificity of the analyzed H4R antibodies. To our best knowledge, a comparable complex analysis of the specificity of GPCR antibodies has not been reported before.
We fully agree with Pradidarcheep et al. (2008) and Michel et al. (2009) who stated that it would be helpful to have “certified” commercially available antibodies or antisera which fulfill at least one of the above-listed criteria to demonstrate sufficient specificity. Such validation should become a gold standard for antibody suppliers and in datasheets specificity should be shown by appropriate methods including the corresponding controls. If such antibodies were available, several time-consuming and expensive validation assays by the user would be unnecessary. In conclusion, for the time being, caution must be exerted when interpreting studies using commercially available H4R as well as other GPCR antibodies.
We thank Dr. Christina Hartwig for her valuable help and stimulating discussions. The excellent technical assistance of Mrs. Renate Schottmann is appreciated. H4R cDNAs and the H4R-deficient mouse were kind gifts from Dr. Rob Thurmond (Johnson and Johnson Research & Development, La Jolla, CA, USA). This work was supported by a grant of the Deutsche Forschungsgemeinschaft (DFG, SFB 587) and by the EU COST action BMBS 0806.