Dear Editor,

We would like to thank Gutzmer et al. (2012) for their comments on our paper “Commercially available antibodies against the human and murine histamine H4-receptor lack specificity” (Beermann et al. 2011), discussing the validation of commercially available antibodies directed against the histamine H4 receptor.

In their communication (Gutzmer et al. 2012), the authors correctly point out that in the publications by van Rijn et al. (2006), Dijkstra et al. (2008), and Bäumer et al. (2008), a non-commercially available antibody, generated by van Rijn et al. (2006), was used. Thus, we honestly apologize that in our paper we conveyed the wrong impression that this non-commercial antibody has been analyzed too. We only analyzed commercially available antibodies in our study (Beermann et al. 2011).

Gutzmer et al. (2012) also state that this non-commercial antibody as well as the sc H110 antibody from Santa Cruz used by Gschwandtner et al. (2010) have been characterized carefully. Indeed, the authors performed experiments to validate the antibodies using cells and tissue of H4R-deficient mice. These experiments show reduced immunofluorescence staining with the sc H110 antibody in cells from H4R-deficient Balb/c animals compared to wild-type controls.

Nevertheless, we would like to discuss several additional aspects concerning the use of commercial and non-commercial antibodies:

  1. 1.

    Preabsorption with the peptide used for immunization is not a reliable method to demonstrate an antibody’s specificity. Preabsorption with the cognate peptide inhibits nonspecific cross-binding of the antibody to cellular structures as well as specific binding (Grube 1980). Thus, these experiments just indicate the specificity of the antibody against the peptide, not necessarily the specificity against the holo-protein in its native form. In line with this, we also tried to generate H4R antibodies by peptide immunization using a peptide comprising a sequence from the second extracellular loop (peptide #4948) of the murine H4R. We obtained several antisera and also monoclonal antibodies, which perfectly recognize the peptide used for immunization. However, none of the reagents specifically reacted with H4R-expressing cells (Fig. 1). This indicates that peptide immunization is not a generally reliable method for the generation of antibodies recognizing native, membrane-integrated G-protein coupled receptors (GPCRs) such as the H4R, what may be due to their highly complex secondary structure (Kobilka and Deupi 2007; Haga et al. 2012).

    Fig. 1
    figure 1

    Generation and evaluation of monoclonal mH4R antibodies. H4R-deficient C57Bl/6 mice were immunized with a peptide (#4948) encompassing a sequence from the second extracellular loop (a) of the mouse H4R, and monoclonal antibodies (mAbs) were generated using standard techniques. Four of the mAbs demonstrate reactivity against the immunizing peptide as analyzed by ELISA (b). However, in flow cytometry using HEK 239 cells expressing mH4R (black histograms) and untransfected HEK cells (gray histograms), none of the generated mAbs was able to specifically stain the mH4R-expressing cells (c). As positive control, cells were stained with the M2 mAb (αFlag, Sigma Aldrich, St. Louis, MO, USA) or with the secondary reagent only, goat αmIgG-FITC (Jackson ImmunoResearch, Suffolk, UK)

    Full size image

    In order to overcome these methodological problems, we are currently planning to generate H4R antibodies by genetic immunization (Allard et al. 2011). This method has the advantage that the membrane-integrated receptor in its native structure is used for immunization, instead of a short peptide deduced from the receptors amino acid sequence in a probably mis- or unfolded structure.

  2. 2.

    Isotype-matched antibodies are commonly used to demonstrate specificity of an antibody, also by us (Beermann et al. 2011). However, as we have shown (Beermann et al. 2011, Figs. 2, 4, and 6), the H4R antibodies tested bind very well also to cells deficient in H4R expression, while the isotype-matched control antibodies do not bind. This indicates that the nonspecific binding of the H4R antibodies is cross-binding, which is not mediated via the Fc portion. Thus, the use of isotype-matched control antibodies only provides a hint for specificity (as far as unspecific binding is due to the presence of Fc-receptors on the target cell), but does not conclusively prove it. Such a proof is achieved only by the use of cells definitively deficient in the expression of the respective target molecule to be recognized by the antibody.

  3. 3.

    The antibody generated against a peptide comprising a sequence out of the C-terminal tail of the human (h)H4R by van Rijn et al. (2006) was analyzed for specificity by western blot of recombinantly expressed hH4R in HEK 293 cells. Several species with apparent masses of 34–36, 65–72, and 45 kDa, the latter one occurring only occasionally, were detected by the H4R antibody as well as by epitope-tag antibodies upon expression of tagged receptor constructs. The calculated molecular mass of the human H4R is 44 kDa (Nakamura et al. 2000). In our hands, the Flag-tagged recombinant human, murine, and canine H4R apparently migrate at the expected molecular mass (Schneider et al. 2009; Beermann et al. 2011; Schnell et al. 2011). Differences in the methodologies may be responsible for the different results. While van Rijn et al. (2006) performed western blots with chloroform/methanol-precipitated protein samples, we used non-precipitated whole cell lysates solubilized with NP-40.

    Protein precipitation may lead to aggregation of receptor molecules. In our hands, heat denaturation of cellular lysates containing recombinantly expressed Flag-tagged H4R, as determined by flow cytometry of the H4R-expressing cells, prevented detection of the H4R in western blots (data not shown). Hence, protein precipitation likely is a much more critical factor in immunological GPCR analysis than is generally appreciated. Unfortunately, we did not find studies by third party authors analyzing the apparent migration of the H4R in western blot experiments to reconcile these apparent discrepancies.

  4. 4.

    Gutzmer et al. (2012) state that the Santa Cruz sc H110 antibody (raised against a peptide composed of a sequence out of the third intracellular loop of the hH4R) was “characterized for specificity in flow cytometry, immunofluorescence and Western blot experiments.” The western blot experiments show immunoreactivities at 72 kDa in lysates from murine spleen cells and at 36 and 72 kDa in lysates from canine mastocytoma cells, but not at 45 kDa (R. Gutzmer and W. Bäumer, personal communication). These results, which are not shown in the paper by Gschwandtner et al. (2010), resemble those found by van Rijn et al. (2006) which are discussed above (section 3). Nevertheless, according to these data by Gutzmer and Bäumer, a cross-reactivity of the hH4R antibody sc H110 with H4R of other species could be proposed. However, in our study (Beermann et al. 2011), we used both the human and the murine recombinantly expressed H4R and the native human and murine H4R in primary cells, but found no specific binding of the sc H110 antibody, neither in western blot nor in flow cytometry experiments.

  5. 5.

    With regard to Lethbridge and Chazot (2010) and Connelly et al. (2009), no data using tissue from H4R-deficient mice are shown or are reported as “data not shown” in order to characterize the murine (m)H4R antibody’s specificity, but have been presented on several symposia. These data, however, are problematic since the antibody’s global staining pattern differs between different mouse strains, i.e., C3H and C57Bl/6, which have been used for functional analysis (Connelly et al. 2009) and for antibody validation, respectively (P. Chazot, personal communication).

  6. 6.

    For characterization of the mH4R antibody’s specificity, Lethbridge and Chazot (2010) also provide a western blot experiment comparing mock- and mH4R-transfected HEK293 cells. In both lanes comprising the lysates of mock- and H4R-transfected HEK293 cells, an identical pattern of reactive bands is visible, only being different in their intensities. This, however, is indicative for differences in the loading of the gel, but not for specificity of the antibody used. In the paper by Connelly et al. (2009), who used the antibody generated by van Rijn et al. (2006), specificity of the antibody was assessed by peptide preabsorption (see above, section 1) and by immunohistochemical control stains omitting the H4R antibody. This latter experiment, however, just indicates that the secondary reagent does not bind nonspecifically; it does not determine specificity of the H4R antibody.

  7. 7.

    Gutzmer et al. (2012) suggest that discrepancies in the results may be due to technical differences. Therefore, we have conducted additional flow cytometry experiments (Fig. 2a, b) under conditions matching as closely as possible the ones from Gschwandtner et al. (2010). Unfortunately, we found no specific staining using the sc H110 antibody (Fig. 2a, b). In addition, we have performed immunohistochemical studies of tissues obtained from wild-type and H4R-deficient Balb/c mice (Fig. 2c). The H4R-deficient genotype, obtained by genetic deletion of the major portion of exon 1 of the H4R gene (Hofstra et al. 2003), thereby ablating the essential N-terminal part of the H4R protein, was extensively backcrossed to the wild-type Balb/c genetic background before using the mice in our analyses. Moreover, the H4R-deficient progeny is continuously controlled for their H4R-deficient genotype, resulting in a clearly controlled H4R-deficient mouse strain and a genetically matching control wild-type strain. Both antibodies sc H110 and sc M120, stained sections of murine tissues (the spleen (Fig. 2c), the ear, and the lung (data not shown)), irrespective of whether they were obtained from either wild-type or from H4R-deficient mice to the same extent.

    Fig. 2
    figure 2

    Analysis of commercially available H4R antibodies. HEK 293 cells (a) expressing either the mH1R or the mH4R as indicated above the panel or spleen cells (b) obtained from either wild-type (Balb/c) or H4R-deficient (H4R−/−) Balb/c mice as indicated above the panel were stained with the indicated antibodies (sc M120, sc H110; Santa Cruz, Biotechnology, Santa Cruz, CA, USA) or an isotype-matched control antibody (Jackson ImmunoResearch) and analyzed by flow cytometry. The staining procedure was carried as described by Gschwandtner et al. (2010), using the Cell Fixation and Permeabilisation Kit (eBioscience, San Diego, CA, USA) and 3 μg/ml (w/v) of the primary antibodies. Black histograms (a) or black lines (b) represent staining with the sc M120 or the sc H110 antibodies, as indicated in the histograms. Light gray and dark gray histograms (a) or lines (b) represent unstained cells or cells stained with the isotype-matched control antibody, respectively. In (c), immunohistochemical staining of formalin-fixed and paraffin-embedded spleen sections obtained from either wild-type (Balb/c) or H4R-deficient (H4R−/−) Balb/c mice as indicated above the panel are shown. Sections were dried, deparaffinized, incubated in hydrogen peroxide, and rehydrated. Heat-induced epitope retrieval was performed using Tris–EDTA buffer at 95 °C. Samples were further processed with an automated horizontal slide-processing system (Dako Autostainer Plus, Dako Colorade, Ft. Collins, CO, USA). Slides were blocked [reagent 1, Zytochem Plus (HRP) Kit, Zytomed Systems, Berlin, Germany] and thereafter incubated with primary antibodies as indicated on the left (ctr without primary antibody). Bound primary antibodies were visualized with the HRP detection system [Zytochem Plus (HRP) Kit, Zytomed Systems] and DAB substrate (DAB Substrate Kit High Contrast, Zytomed Systems; brown color) according to the manufacturers’ instructions. Slides were counterstained for 2 min with hematoxylin (blue color). In both analyses, flow cytometry and histochemistry, the antibodies sc M120 and sc H110 stain cells which express the H4R as well as cells which are deficient in H4R expression, thus in a unspecific manner

    Full size image

    How can these discrepancies be explained? Besides the discussion about the species specificity of the H4R antibody sc 110 (section 4), it should be noted that the experiments by Gschwandtner et al. (2010) were performed earlier than our experiments. Hence, it is possible that the quality of the commercial antibody preparation declined over time so that initial positive results could not be reproduced. Accordingly, stringent quality control of GPCR antibodies supplied by commercial manufactures is essential, regardless of whether previous experiments with identical antibodies were successfully performed. As the discussion thread in the papers by Beermann et al. (2012), Gutzmer et al. (2012), and in this letter shows, continuous and rigorous antibody validation is absolutely essential to avoid confusion.

  8. 8.

    We also kindly asked for provision of the antibodies generated by van Rijn et al. (2006) and by Lethbridge and Chazot (2010), in order to characterize them using our test systems and to reconcile the discrepancies. The authors would have provided the antibodies to us, but unfortunately, the antibodies were already exhausted and thus were any more available to us. Hence, we cannot finally answer the question whether or not these antibodies are specific also in our test systems.

In summary, we appreciate the agreement of Gutzmer et al. (2012) that each single preparation of polyclonal H4R antibody has to be rigorously validated for each application and that there is an urgent scientific need for new H4R detection tools. Unfortunately, the antibodies analyzed in Beermann et al. (2011) and in this communication are not specific from our point of view.

Comparable nonspecific reactivities of a series of antibodies supposedly recognizing other biogenic amine receptors have already been described in a special cluster of publications edited by Michel et al. (2009). These structurally related receptors include the β1- and β2-adrenergic receptor, α-adrenergic receptors, muscarinic receptors, and dopamine receptors (Hamdani and van der Velden 2009; Pradidarcheep et al. 2009; Jensen et al. 2009; Bodei et al. 2009; Jositsch et al. 2009). A recent study published in this issue of Naunyn-Schmiedeberg’s Archives of Pharmacology extends these findings to antibodies directed against the β3-adrenoceptor (Cernecka et al. 2012). This publication, however, also indicates that the specificity of an antibody strictly depends on the application it is used for, e.g., while a given antibody may be suitable for immunocytochemistry, it may completely lack specificity in western blot. Thus, it is conceivable that the conformation of the epitopes recognized by the antibodies under various experimental conditions is different, resulting in different apparent antibody specificities.

The problems regarding the specificity of GPCR antibodies are certainly associated also with the fact that, in general, GPCRs are expressed at very low levels in native cells, e.g., as compared to most G-proteins (Alousi et al. 1991; Defer et al. 2000; Seifert et al. 1998; Wenzel-Seifert et al. 1999). As a consequence, the signal-to-noise ratio of GPCR antibodies is decreased profoundly.

At the time being, we recommend validated selective H4R agonists and antagonists as tools to identify the H4R functionally. In a recent study, we have used a combination of selective H1-, H2-, and H4R agonists as well as H1-, H2-, and H4R antagonists to comprehensively characterize the H4R in highly purified human eosinophils. In fact, use of a broad spectrum of validated pharmacological tools readily compensated for the deficiency of suitable H4R antibodies (Reher et al. 2012).

Collectively, the critical use of selective H X R agonists and antagonists and the H4R-deficient mouse on a controlled genetic background are currently the best experimental tools to study the involvement of the H4R in biological processes in native cells (Hofstra et al. 2003; Dunford et al. 2006; Beermann et al. 2012; Reher et al. 2012). In a broader perspective, publications using antibodies directed against biogenic amine receptors should be read critically with regard to the specificity of the used reagents.