UBC Faculty Research and Publications

HOXA4 protein levels and localization in the aorta and in human abdominal aortic aneurysms Klausen, Christian; Auersperg, Nelly Dec 14, 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12899_2011_Article_186.pdf [ 258.74kB ]
JSON: 52383-1.0215900.json
JSON-LD: 52383-1.0215900-ld.json
RDF/XML (Pretty): 52383-1.0215900-rdf.xml
RDF/JSON: 52383-1.0215900-rdf.json
Turtle: 52383-1.0215900-turtle.txt
N-Triples: 52383-1.0215900-rdf-ntriples.txt
Original Record: 52383-1.0215900-source.json
Full Text

Full Text

CORRESPONDENCE Open AccessHOXA4 protein levels and localization in theaorta and in human abdominal aortic aneurysmsChristian Klausen* and Nelly AuerspergAbstractThis report presents evidence for the specificities of select commercially available HOXA4 antibodies in regards toconcerns about the specificity of the HOXA4 antibody used by Lillvis et al. (Regional expression of HOXA4 alongthe aorta and its potential role in human abdominal aortic aneurysms. BMC Physiol 2011, 11:9). Using anantibody characterized extensively by us, Lillvis et al. report detecting HOXA4 at a size of 33 kDa despite ourprevious reports that HOXA4 is detected at ~37-39 kDa and that the ~30-33 kDa band is non-specific. Using smallinterfering RNA targeting HOXA4, forced expression of full-length HOXA4 and HOXA4-positive and -negativeovarian cancer cell lines, we confirm our previous findings that the ~30-33 kDa band is non-specific and thatHOXA4 is detected at ~37-39 kDa. Moreover, we demonstrate that HOXA4 small interfering RNA reduces the ~37-39 kDa HOXA4 band, but not the ~30-33 kDa non-specific band, in a human acute monocytic leukemia cell lineused by Lillvis et al. Western blot analysis performed with two additional commercially available HOXA4 antibodiesalso detected HOXA4 at ~37-39 kDa. Lastly, immunofluorescent staining of a HOXA4-negative ovarian cancer cellline with the antibody used by Lillvis et al. yields strong perinuclear staining, similar to that observed by Lillvis etal., which cannot be attributed to HOXA4. Our results highlight and briefly discuss the importance of carefulantibody validation and selection for use in various applications.CorrespondenceWe read with interest the study of Lillvis et al. [1]regarding the expression of HOXA4 in the aorta and itspotential role in abdominal aortic aneurysms. Theauthors used microarray analysis validated by reversetranscription quantitative real-time PCR to providestrong evidence that HOXA4 mRNA levels are reducedin human abdominal aortic aneurysms relative to con-trol human abdominal aorta. However, we have signifi-cant concerns about the subsequent data regardingHOXA4 protein levels. For their studies Lillvis et al.used a commercially available rabbit polyclonal HOXA4antibody (ab26097; Abcam, Cambridge, MA) that waspreviously characterized extensively by us [2,3]. Whilethey were kind enough to reference our studies, theystate that HOXA4 was detected as a single band at ~33kDa and evidence for this is presented in their Addi-tional file three, Figure S1A. However, in both of ourprevious studies we state that the size of HOXA4 is~37-39 kDa and our first study [2] demonstrated thatthe band at ~30-33 kDa is a non-specific band.The ~30-33 kDa non-specific band is by far the mostintense band and appears as a single band at low expo-sures irrespective of blotting conditions, as was observedby Lillvis et al. in Additional file three, Figure S1A. Thiswas demonstrated previously by us [2] and we now pro-vide additional evidence of this in Figures 1A and 1B.Strong expression of the ~30-33 kDa non-specific bandis observed in the five human ovarian cancer cell linesshown in Figures 1A and 1B regardless of the fact thatHOXA4 mRNA is undetectable in both SKOV-3 andA2780 cells [2,3]. More importantly, the ~30-33 kDanon-specific band is insensitive to small interfering RNA(siRNA) targeting HOXA4 ([2] and Figure 1A) andforced expression of full-length HOXA4 (Figure 1B).Likewise, Figure 1A shows strong, siRNA-insensitiveexpression of the ~30-33 kDa non-specific band in theacute monocytic leukemia cell line used by Lillvis et al.(THP-1 cells, designated as MP1 cells by Lillvis et al. inAdditional file four, Table S3).In contrast, we have demonstrated that the ~37-39kDa HOXA4 band correlates with HOXA4 mRNA* Correspondence: klausen@mail.ubc.caDepartment of Obstetrics & Gynaecology, Child & Family Research Institute,University of British Columbia, Vancouver, British Columbia, CanadaKlausen and Auersperg BMC Physiology 2011, 11:18http://www.biomedcentral.com/1472-6793/11/18© 2011 Klausen and Auersperg; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.Figure 1 Evidence for the specificities of select commercially available HOXA4 antibodies. Immunoblot analysis was performed as in [3]with lysates from one human acute monocytic leukemia (THP-1) and five ovarian cancer (OVCAR-8, OVCAR-3, SKOV-3, A2780 and CaOV-3) celllines. Red arrowheads indicate HOXA4 (~37-39 kDa), black arrowheads indicate the ~30-33 kDa non-specific band associated with the Abcamantibody, and white arrowheads indicate select non-specific bands of equal/greater intensity. (A) SKOV-3 cells lack HOXA4 whereas OVCAR-8 andOVCAR-3 cells express high levels of HOXA4 [2,3]. Transient knockdown was performed as in [3] with 20 nM HOXA4-targeting siRNA (siHOXA4;ON-TARGETplus SMARTpool; Dharmacon), control siRNA (siControl; ON-TARGETplus Non-Targeting Pool) or transfection reagent alone (iMAX;Lipofectamine RNAiMAX). Immunoblot analysis with the Abcam HOXA4 antibody characterized by us [2,3] shows that the ~30-33 kDa non-specific band detected by Lillvis et al. [1] is insensitive to HOXA4 siRNA and is expressed by HOXA4-negative SKOV-3 cells. (B) A2780 cells lackHOXA4 whereas CaOV-3 cells express low levels of HOXA4 [2,3]. CaOV-3 cells were transfected as in [3] with control vector (CaOV-3-Vector) orvector encoding full-length HOXA4 (CaOV-3-HOXA4). Immunoblot analysis with the Abcam antibody and two other commercially availableHOXA4 antibodies shows that the ~37-39 kDa HOXA4 band detects exogenously expressed HOXA4 and is undetectable in HOXA4-negativeA2780 cells. (C) Formaldehyde-fixed SKOV-3 cells were permeabilized, blocked and incubated overnight with anti-HOXA4 (1:3200). Secondaryantibody was applied, cell nuclei were stained with Hoechst 33258, and coverslips were examined by epifluorescence microscopy. AlthoughHOXA4 mRNA [2,3] and protein (A) are undetectable in SKOV-3 cells, these cells have strong expression of the ~30-33 kDa non-specific band ((A)and [2]). Note that immunofluorescent staining of HOXA4-negative SKOV-3 cells with the Abcam antibody yields strong perinuclear stainingwhich cannot be attributed to HOXA4.Klausen and Auersperg BMC Physiology 2011, 11:18http://www.biomedcentral.com/1472-6793/11/18Page 2 of 4levels, is sensitive to HOXA4 siRNA and detects exo-genously expressed HOXA4 ([2,3] and Figures 1A and1B). Although the ~37-39 kDa HOXA4 band isexpressed at very low levels in THP-1 cells, it is theonly band that is reduced by treatment with HOXA4siRNA (Figure 1A). Furthermore, we now provide evi-dence that two additional commercially availableHOXA4 antibodies detect a ~37-39 kDa band that cor-relates with HOXA4 mRNA levels and detects exogen-ously expressed HOXA4 (Figure 1B). In this context, weare not convinced that the Western blot results pre-sented by Lillvis et al. reflect changes in HOXA4 proteinlevels. In light of their convincing mRNA data, perhapsa re-examination of the results of longer exposureswould allow Lillvis et al. to quantify the ~37-39 kDaHOXA4 band detected with this antibody. We routinelycut away the membrane just below the 37 kDa molecu-lar mass marker prior to immunoblotting in order toprevent strong signal from the ~30-33 kDa non-specificband from interfering with quantitation of the ~37-39kDa HOXA4 band.Our second, and very much related, concern has to dowith the use of the same rabbit polyclonal HOXA4 anti-body for immunohistochemistry and immunofluores-cence. In light of the intense ~30-33 kDa non-specificband and additional ~46-48 kDa non-specific bands ofequal or greater intensity than HOXA4, we have signifi-cant concerns about the use of this antibody in thesetypes of applications. Indeed, we have performed immu-nofluorescent staining of HOXA4-negative SKOV-3 cells([2,3] and Figure 1A) and found strong perinuclearstaining similar to that observed by Lillvis et al. whichcannot be attributed to HOXA4 since SKOV-3 cellsdon’t express HOXA4 (Figure 1C). While this stainingmost likely corresponds to the ~30-33 kDa non-specificband due to its intensity and the high dilution of theantibody (1:3200 vs. 1:200 used by Lillvis et al.), we can-not rule out contributions from the ~46-48 kDa non-specific bands. Nuclear localization, especially asassessed by epifluorescence rather than confocal micro-scopy, does not appear to be sufficient to identifyHOXA4 in cells with demonstrated HOXA4 expressionsince Lillvis et al. also report detecting the ~30-33 kDaband in both cytoplasmic and nuclear fractions by Wes-tern blot.In this context, we are not convinced that the immu-nohistochemistry and immunofluorescence results pre-sented by Lillvis et al. accurately depict the localizationof HOXA4 protein. Indeed, we would suggest that theuse of this rabbit polyclonal HOXA4 antibody in thesetypes of applications is unwarranted unless definitivevalidation of specificity is presented. Our preliminaryWestern blot analysis of two additional commerciallyavailable HOXA4 antibodies has similarly revealed anumber of non-specific bands of equal or greater inten-sity than HOXA4 (Figure 1B). Persuasive validation stu-dies will therefore be required to establish whether anyof the commercially available HOXA4 antibodies possessthe specificity required for use in applications such asimmunohistochemistry and immunofluorescence. Futurestudies on the biology of HOXA4 and its roles in dis-ease would benefit greatly from the development of ahighly specific antibody suitable for use in these andother applications.The example of HOXA4 highlights the need forresearchers to validate the specificity of antibodies foruse in various applications, especially in the absence ofrigorous antibody characterization. Indeed, while theAbcam HOXA4 antibody performs very well in Westernblot analysis, its use in other applications such as immu-nocytochemistry, immunohistochemistry or immunopre-cipitation is likely unwarranted unless definitivevalidation studies are performed in consideration of thesheer intensity of the signal from the ~30-33 kDa pro-tein. For many companies antibody characterizationtakes the form of ELISA or Western blot analysis withthe antigenic peptide/protein. While this does establishwhether or not the antibody can detect the particularpeptide/protein against which it was raised, it does notprove it will detect the actual protein (except whenraised and tested against full-length protein) nor does itaddress how the antibody will perform in the context ofthe entire proteome (i.e. in cells or tissues) or in differ-ent applications. Preabsorption studies with the anti-genic peptide/protein can identify true non-specificbands (i.e. cannot be eliminated by saturation), but theycannot distinguish the protein of interest from non-spe-cific proteins with similar epitopes that the antibodybinds to in a specific manner. For example, preabsorp-tion of the Abcam HOXA4 antibody with the antigenicpeptide completely eliminates the ~37-39 kDa HOXA4band, the ~30-33 kDa non-specific band and one of the~46-48 kDa non-specific bands [2]. Thus, the other~46-48 kDa non-specific band is likely a true non-speci-fic band, yet this approach fails to identify only the ~37-39 kDa band as being specific for HOXA4.The detection of a band of “appropriate” size in cell ortissue lysates can be used to confirm the identity of aband, but must involve more than one (preferably threeor more) lysates from cells or tissues known to have dif-fering basal or experimentally manipulated expressionlevels. In general, accurate up-front knowledge ofexpression levels is most easily achieved by measuringmRNA levels. An obvious, and highly effective, exten-sion of this approach is to use RNA interference-mediated knockdown or vector-based forced expressionto generate lysates with gene-specific changes in theexpression level of the protein of interest. Although useKlausen and Auersperg BMC Physiology 2011, 11:18http://www.biomedcentral.com/1472-6793/11/18Page 3 of 4of an antibody in several applications requires separatevalidation for each application, many of the approachesdiscussed above can be adapted to different applications.In addition, companies are beginning to employ theseapproaches as they strive to market gene-specific pro-duct portfolios (antibody, protein/peptide, RNA interfer-ence and forced expression) on a genome-wide scale.However, in the absence of thorough antibody charac-terization it is ultimately the responsibility of theresearcher to ensure the validity of results obtainedfrom antibody-based applications.AcknowledgementsThis work was supported by a grant from the National Cancer Institute ofCanada with funds from the Canadian Cancer Society to N.A.Authors’ contributionsCK conceived of the study, participated in its design, carried out allexperiments and drafted the manuscript. NA participated in the design ofthe study and helped to draft the manuscript. All authors read andapproved the final manuscript.Competing interestsThe authors declare that they have no competing interests.Received: 22 August 2011 Accepted: 14 December 2011Published: 14 December 2011References1. Lillvis JH, Erdman R, Schworer CM, Golden A, Derr K, Gatalica Z, Cox LA,Shen J, Vander Heide RS, Lenk GM, Hlavaty L, Li L, Elmore JR, Franklin DP,Gray JL, Garvin RP, Carey DJ, Lancaster WD, Tromp G, Kuivaniemi H:Regional expression of HOXA4 along the aorta and its potential role inhuman abdominal aortic aneurysms. BMC Physiol 2011, 11:9.2. Ota T, Klausen C, Salamanca MC, Woo HL, Leung PC, Auersperg N:Expression and function of HOXA genes in normal and neoplasticovarian epithelial cells. Differentiation 2009, 77:162-171.3. Klausen C, Leung PC, Auersperg N: Cell motility and spreading aresuppressed by HOXA4 in ovarian cancer cells: possible involvement ofbeta1 integrin. Mol Cancer Res 2009, 7:1425-1437.doi:10.1186/1472-6793-11-18Cite this article as: Klausen and Auersperg: HOXA4 protein levels andlocalization in the aorta and in human abdominal aortic aneurysms.BMC Physiology 2011 11:18.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitKlausen and Auersperg BMC Physiology 2011, 11:18http://www.biomedcentral.com/1472-6793/11/18Page 4 of 4


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items