Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Gene Expression Profiling of Histiocytic Sarcomas in a Canine Model: The Predisposed Flatcoated Retriever Dog

  • Kim M. Boerkamp ,

    Affiliation Faculty of Veterinary Medicine, Department of Clinical Sciences Companion Animals, Utrecht University, Utrecht, The Netherlands

  • Marieke van der Kooij ,

    Contributed equally to this work with: Marieke van der Kooij, Frank G. van Steenbeek

    Affiliation Faculty of Veterinary Medicine, Department of Clinical Sciences Companion Animals, Utrecht University, Utrecht, The Netherlands

  • Frank G. van Steenbeek ,

    Contributed equally to this work with: Marieke van der Kooij, Frank G. van Steenbeek

    Affiliation Faculty of Veterinary Medicine, Department of Clinical Sciences Companion Animals, Utrecht University, Utrecht, The Netherlands

  • Monique E. van Wolferen,

    Affiliation Faculty of Veterinary Medicine, Department of Clinical Sciences Companion Animals, Utrecht University, Utrecht, The Netherlands

  • Marian J. A. Groot Koerkamp,

    Affiliation Molecular Cancer Research, University Medical Centre Utrecht, Utrecht, The Netherlands

  • Dik van Leenen,

    Affiliation Molecular Cancer Research, University Medical Centre Utrecht, Utrecht, The Netherlands

  • Guy C. M. Grinwis,

    Affiliation Faculty of Veterinary Medicine, Department of Pathobiology, Utrecht University, Utrecht, The Netherlands

  • Louis C. Penning,

    Affiliation Faculty of Veterinary Medicine, Department of Clinical Sciences Companion Animals, Utrecht University, Utrecht, The Netherlands

  • Erik A. C. Wiemer,

    Affiliation Dept. of Medical Oncology, Daniel den Hoed Cancer Center, Erasmus University Medical Center, Rotterdam, The Netherlands

  • Gerard R. Rutteman

    Affiliation Faculty of Veterinary Medicine, Department of Clinical Sciences Companion Animals, Utrecht University, Utrecht, The Netherlands

Gene Expression Profiling of Histiocytic Sarcomas in a Canine Model: The Predisposed Flatcoated Retriever Dog

  • Kim M. Boerkamp, 
  • Marieke van der Kooij, 
  • Frank G. van Steenbeek, 
  • Monique E. van Wolferen, 
  • Marian J. A. Groot Koerkamp, 
  • Dik van Leenen, 
  • Guy C. M. Grinwis, 
  • Louis C. Penning, 
  • Erik A. C. Wiemer, 
  • Gerard R. Rutteman



The determination of altered expression of genes in specific tumor types and their effect upon cellular processes may create insight in tumorigenesis and help to design better treatments. The Flatcoated retriever is a dog breed with an exceptionally high incidence of histiocytic sarcomas. The breed develops two distinct entities of histiocytic neoplasia, a soft tissue form and a visceral form. Gene expression studies of these tumors have value for comparable human diseases such as histiocytic/dendritic cell sarcoma for which knowledge is difficult to accrue due to their rare occurrence. In addition, such studies may help in the search for genetic aberrations underlying the genetic predisposition in this dog breed.


Microarray analysis and pathway analyses were performed on fresh-frozen tissues obtained from Flatcoated retrievers with localized, soft tissue histiocytic sarcomas (STHS) and disseminated, visceral histiocytic sarcomas (VHS) and on normal canine spleens from various breeds. Expression differences of nine genes were validated with quantitative real-time PCR (qPCR) analyses.


QPCR analyses identified the significantly altered expression of nine genes; PPBP, SpiC, VCAM1, ENPEP, ITGAD (down-regulated), and GTSF1, Col3a1, CD90 and LUM (up-regulated) in the comparison of both the soft tissue and the visceral form with healthy spleen. DAVID pathway analyses revealed 24 pathways that were significantly involved in the development of HS in general, most of which were involved in the DNA repair and replication process.


This study identified altered expression of nine genes not yet implicated in histiocytic sarcoma manifestations in the dog nor in comparable human histiocytic/dendritic sarcomas. Exploration of the downside effect of canine inbreeding strategies for the study of similar sarcomas in humans might also lead to the identification of genes related to these rare malignancies in the human.


Fundamental research on rare human diseases is not only hampered by minimal grant supply; lack of sufficient sample numbers is of equal disadvantage. One way to overcome this last catch, is to investigate other species in which a similar disease occurs at a much higher frequency, like in specific dog breeds. In dogs, a downside of selection for breeding purposes is the occurrence of a very large number of breed-specific hereditary diseases ( Rare human diseases, such as histiocytic malignancies, might therefore be common in specific dog breeds [1].

The Flatcoated retriever (FCR) breed has a strongly increased risk for histiocytic sarcoma (HS) development. In the UK, it is likely to account for about 36% of all malignant neoplasms diagnosed in this breed [2], [3].

Canine histiocytic malignant disorders were as such first described in the late 1970s [4]. Included in the name ‘histiocytic sarcoma’, which was given to the complex of malignant histiocytic disorders [5], there is a range of malignant tumors derived from CD34-committed stem cell precursors that may develop into dendritic cells (DC) such as Langerhans cells, interstitial DC and macrophages [6], [7]. In addition to morphological features, most histiocytic sarcomas can be recognized by positive immunostaining for the cell surface marker CD18 [6][8]. Canine HS has resemblance to the rare and often lethal human histiocytic malignancies, including dendritic cell and histiocytic sarcomas and disseminated Langerhans cell histiocytosis (LCH) [9][12]. Histiocytic sarcomas in dogs almost inevitably metastasize to various organs [10] and have a very poor prognosis [2], [10], [11]. A median survival of four months has been reported in the FCR [3].

In dogs, there are several clinical manifestations of HS. One common form is the localized, soft tissue histiocytic sarcoma (STHS), which manifests itself as a tumor arising in the deeply seated soft tissues of limbs often in association with joints [2], . A second common form is manifested in internal organs and often multifocal and named disseminated, visceral histiocytic sarcoma (VHS), with neoplastic changes that can be found in either spleen, liver, lung and/or bone marrow [6].

Past research using Comparative Genomic Hybridization has already shown various aberrations in HS in the Bernese Mountain Dog, another breed predisposed to histiocytic malignancies, with cases showing numerous shared Copy Number Alterations (CNAs) both gains and losses, throughout the genome. These included deletions of the tumor suppressor genes CDKN2A/B, RB1 and PTEN [10]. Furthermore, an associated constitutional haplotype in a locus near to the highly cited tumor suppressor locus MTAP-CDKN2A has recently been identified in this breed [1]. Another study concluded that deregulation of the expression of the glycation end products (Receptor for Advanced Glycation Endproducts; RAGE) and the high mobility group box1 protein (HMGB1) potentially have a major effect on the progression of malignant histiocytic disorders [14].

cDNA microarrays have become powerful tools in the study of gene expression which has enabled improved classification of various naturally occurring cancers [15], [16] and have, once the canine genome sequence became available [17] already proven their value in the research of various canine sarcomas [18][20] but not yet in HS. Thus, we examined shared genetic functional aberrations of HS by comparing both forms of HS with normal tissue, for which spleen was chosen.

The study of spontaneously occurring tumors in the dog, a species which has a genetically stronger relationship to the human than mice [21], [22] can enrich the knowledge of rare human cancers, and lead to more insight in the pathogenesis of the disease and facilitate the identification of therapeutic targets valuable for dog and human [1], [11], [18], [20], [21]. The outcome of this study provides evidence of the existence of common differences in gene activity between HS and normal spleen.

Materials and Methods

The experimental protocol (ID 2007.III.08.110) was peer-reviewed by the scientific committee of the Department of Animals in Science & Society, Utrecht University, The Netherlands, and approved by the Animal Experiments Committee of the Academic Biomedical Centre, Utrecht, The Netherlands. The Animal Experiments Committee based its decision on 'De Wet op de Dierproeven' (The Dutch 'Experiments on Animals Act', 1996) and on the 'Dierproevenbesluit' (the Dutch 'animal experiments decree', 1996). Both documents are available online at

Case Recruitment and Histopathological Evaluation

All tumor samples were confirmed as being spontaneously occurring histiocytic malignancies and were obtained from family-owned FCR with informed owner consent. All tumor material used originated from the Dutch FCR, that had not received radiotherapeutical or cytostatic treatment. Tumors were obtained under sterile conditions, either as part of a routine diagnostic or therapeutic surgical procedure, or immediately following euthanasia. Directly after excision, samples were snap frozen in liquid nitrogen, or alternatively by primary preservation in RNA-later, in both instances followed by storage at minus 70°C. Tumor samples collected adjacent to the site of the frozen or RNA-later preserved samples were fixed in 10% neutral buffered formalin and routinely processed for histological examination.

At the time of surgery or necropsy, the evident anatomical location of all tumors was recorded for each individual and categorized as either VHS, if a tumor was present in internal organs (n = 7) or STHS, if the tumor was localized in a limb only without identifiable metastases (n = 6) [8].

Histological specimens were classified by a board-certified veterinary pathologist (GCMG) according to the recommendations and classification scheme defined by Affolter and Moore [6]. In all cases immunohistochemical staining with antibodies against CD18 protein (the common subset of β2 adhesion integrins, expressed in histiocytes, dendritic cells (DC), lymphocytes, and polymorphonuclear leukocytes [23]) were used to confirm the suspected histiocytic origin [8]. Results of this staining were divided in two categories: negative or positive. If the differential diagnosis based upon morphology included the potential origin of other malignant round cell tumors (malignant lymphoma, mastocytoma, melanoma, myeloma) appropriate immunostaining to examine such potential histogentic origin was performed and had to be negative. All tumor samples selected for the genetic study contained over 50% tumor cells as assessed in histological sections of biopsies of adjacent tissue. Patient details are listed in Table 1.

As control tissue, normal spleen from (healthy) crossbreed dogs (n  = 6) was used as obtained at postmortem immediately following euthanasia that was not related to neoplastic, endocrine or metabolic diseases.

As a common reference pool a multitude of canine organs (testis, liver, spleen, prostate, duodenum, lung, kidney and brain) were used that had been obtained from healthy crossbreeds (n = 8) euthanized for non-metabolic, non-tumorous lethal conditions.

The procedures were approved by the local ethics committee, as required under Dutch legislation (ID 2007.III.08.110).

RNA Isolation

Approximately 30 mg of frozen tumor was transferred to a container with 600 µl of Buffer RLT and was disrupted/lysed and homogenized using a dismembrator (Braun Biotech Int., Melsungen, Germany) for 45 s at 2200 rpm. Total RNA was isolated and treated with DNase using the RNeasy mini kit (Qiagen, The Netherlands) according to the manufacturer's protocol. Quantity and integrity were assessed with the Bioanalyzer Agilent BioAnalyzer-2100 (Bioanalyzer, Agilent Technologies, Santa Clara, CA) in combination with an RNA 6000 Pico-LabChip. The average RNA integrity number 8.5 (range: 7.2–9.8) was found to be appropriate [24]. RNA concentration was quantified using a NanoDrop ND-1000 (Isogen Life Science) spectrophotometer.

Expression Profiling

RNA was labeled twice and hybridized against the common reference RNA on dual channel arrays. RNA concentrations were 0.6 µg/µl at a minimum amount of 3 µg per sample.

RNA amplifications and labeling were performed on an automated system (Caliper Life Sciences NV/SA, Belgium) as described [25]. Dye swap of Cy3 and Cy5 was performed to reduce dye bias. Hybridizations were done on a HS4800PRO system supplemented with QuadChambers (Tecan Benelux B.V.B.A.) using 500–1000 ng labeled cRNA per channel as described [26]. Microarrays used were Agilent Canine Gene Expression Microarrays V1 (Agilent Technologies, Belgium) representing 42,035 canine 60-mer probes in a 4×44 K layout.

Hybridized slides were scanned on an Agilent scanner (G2565BA) at 100% laser power, 30% photomultiplier tube voltage. After automated data extraction using Imagene 8.0 (BioDiscovery), printtip Loess normalization was performed [27] on mean spot-intensities. Dye-bias was corrected based on a within-set estimate as described [28].

Analyses were performed to detect common differences in gene expression between the two groups of HS and healthy spleen tissue. Data were analyzed using ANOVA [29] (R version 2.2.1/MAANOVA version 0.98–7) ( In a fixed effect analysis, sample, array and dye effects were modeled. P-values were determined by a permutation F2-test, in which residuals were permutated 5,000 times globally. Genes with P<0.0002 after either family wise error correction (FWER) or determination of false discovery rate (FDR) were considered significantly changed. Genes with log2-fold changes of more than 0.5 or less than −0.5 were then selected to ensure that only robust changes were considered.

The Gene Ontology (GO) database ( was used to check gene molecular and biological functions of the remaining genes.

Functional Annotation

In general, pathway analysis in dogs has its restraints because pathway identification relies heavily on existing functional annotation, which is still limited for this species. Still, pathway analysis provides an additional way to analyze expression data across species. This may shed light on common pathways important for tumor behavior and on finding new therapeutic targets. To examine whether certain pathways are over- or under-represented in the gene list, all genes significantly differentially expressed between either STHS, VHS and normal spleen, were introduced into DAVID (

Quantitative Real Time PCR

Gene selection.

Following the outcome of the microarray expression profiling, ten genes were selected. Selection of ten genes was based on significantly differently expression, M-fold changes and potential biological function in relation with tumor development.

These genes including their optimum temperature are listed in Table 2.

Table 2. QPCR primers for genes of interest based on Microarray and pathway analyses.

RNA Isolation and cDNA Synthesis

Tissues from all but one patient (of which the insufficient tissue remained for the qPCR experiment) were used in the microarray experiment (six spleens, six STHS and seven VHS), furthermore three additional samples (one normal spleen, one STHS and one VHS that met the inclusion criteria) were added thus creating 3 groups of seven samples for the qPCR experiment. Total RNA from these samples was isolated. After isolation, total RNA was treated with DNase using the RNeasy mini kit (Qiagen, The Netherlands) according to the manufacturer’s protocol.

Reverse transcription (RT) was performed of all 20 samples in a 80 µl reaction using 2000 ng total RNA, 16 µl iScript Reaction mix and 4 µl iScript Reverse Transcriptase (iScript cDNA Synthesis kit, Bio Rad, Veenendaal, The Netherlands). This includes a mixture of oligo-dT random hexamer primers. The mixture was incubated 5 min. at 25°C, 30 min. at 42°C and followed by 5 min. at 85°C. Minus RT controls were prepared from 500 ng of the same RNA under the same conditions, but without addition of reverse transcriptase.

Reference Genes and Primer Development

Reference genes were selected as non-regulated reference genes for normalization based on their stable expression in canine tissue [30], .

Using Ensembl, through annotated transcripts, PCR primers were designed using the Perl Primer software (version and primer3 software (version 0.4.0) according to the parameters outlined in the Bio-Rad i-cycler manual. The specificity of each primer pair was confirmed by sequencing its product and also in qPCR by checking the meltcurve and reaction efficiency. GeneNorm was used to establish expression stability [32]. Amplicon sequence-reactions were performed using BigDye v3.1 according to the manufacturer’s (Life Technologies, Bleiswijk, The Netherlands) instructions on an ABI3130×L and analyzed in Lasergene (version 9.1 DNASTAR) and confirmed the specificity of each amplicon. Using RefFinder, the stability of nine reference genes were checked. In Figure 1 they are listed according to their stability. Four of the more stable primers were chosen for further data analysis, namely ribosomal protein S5 (RPS5), signal recognition particle receptor (SRPR), ribosomal protein L13(RPL13), hypoxanthine phosphoribosyltransferase (HPRT).

Figure 1. Gene stability by Genorm for all nine reference genes.

Horizontal axis: Least stable genes (left) and most stable genes (right) Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), β-2-Microglobulin (B2MG), Ribosomal protein S5 (RPS5), Ribosomal protein S19 (RPS19), Hypoxanthine phosphoribosyltransferase (HPRT), Ribosomalprotein L8 (RPL8), B-Glucuronidase (GUSB), Signal recognition particle receptor (SRPR), and Ribosomal protein L13 (RPL13).

Primers for all nine reference genes, including their optimum temperature are listed in Table 3.

Quantative PCR

For qPCR, the Bio-Rad detection system (Bio-Rad.) with SYBR green fluorophore was used. Reactions were performed in a total volume of 25 µl containing 12.5 µl of 2×SYBR green super mixes (Bio-Rad Laboratories Ltd.), 1 µl of each primer at 400 nM concentration, 0.8 µl of cDNA and 9.7 µl Rnase and Dnase free water as previously described [33], [34]. Q-PCR reactions for each primer set were optimized by performing reactions under a gradient of annealing temperature using three serial 16×dilutions of pooled cDNA from all tissue samples. Cycling conditions were as follows: Denaturation (95°C for 5 min.), amplification cycle repeated 45 times (95°C for 10 sec, 30 sec at the primer specific annealing temperature (Table 2) and 30 sec at 72°C. The last step, 30 sec at 72°C was omitted when the annealing temperature was higher than 58°C. A melting curve analysis was performed following every run to ensure a single amplified product for every reaction. All reactions were performed in duplicate. The reference standard dilution series was repeated on every plate. Duplicate negative controls, both a minus RT and a water control, were run with every experimental plate to assess the specificity and to identify any potential contamination.

Data analysis was performed with IQ5 Real-Time PCR detection system software (BioRad). Expression levels were normalized using the average relative amount of the reference genes. Log-values of normalized relative expression were used to obtain normal distribution.

A Wilcoxon rank sum test was performed to determine the significance of differential gene expression. All results were Bonferroni corrected.

Results and Discussion

Differentiation between STHS and VHS was based on physical examination and radiographic (thorax)/ultrasound examination (thorax/abdomen). All dogs with visceral organ involvement were euthanized followed by immediate autopsy. Histomorphology and immunohistochemical staining for CD18 confirmed the suspected histiocytic origin of tumors studied.

The Microarray enabled analysis of the expression of 42,034 features. Since only 21,682 (51% ) were annotated (CanFam 2.0), it is possible that important genes are missed. When comparing VHS and spleen, 4,235 features were significantly differentially expressed. When only looking at 4-foldchanges or larger, 352 features remained. When comparing STHS with spleen, 5,779 features were significantly differentially expressed. In this comparison, when only looking at 4-foldchanges or larger, 437 features remain.

Of the total of altered genes, 3,394 features were significantly differentially expressed in both forms of HS versus normal spleen, and 319 features remained when only 4-foldchanges or larger are taking into account.

Figure 2 visualizes the heatmap of the ten genes that were chosen for qPCR confirmation.

Figure 2. Microarray-based heatmap of the ten genes chosen for qPCR.

PPBP: Pro-platelet basic protein (chemokine (C-X-C motif) ligand 7), GTSF1: Gametocyte specific factor 1, SPIC: Spi-C transcription factor, VCAM1: Vascular cell adhesion molecule 1, ENPEP: Glutamyl aminopeptidase (aminopeptidase A), LUM: Lumican, ITGAD: Integrin, alpha D, Col3A1: Collagen, type III, alpha 1, MYH11: Myosin, heavy chain 11, smooth muscle, Thy-1 (CD90): Thy-1 cell surface antigen.

In order to improve knowledge on the genetic basis of HS and to best exploit logistic and financial resources, we chose to examine a selection of the significantly altered genes found in the microarray experiment for confirmation by qPCR. Selection of genes was based on the statistical significance of their differential expression and their potential involvement in tumor development.

Since it is not possible to obtain pure samples of sedentary, non-tumorous histiocytes for expression profiling, spleen tissue was chosen as the normal equivalent of HS. HS arises from interstitial DC, and hence emanate from lymphoid domains when these arise from lymphoid organs such as the spleen [6], choosing spleen as a control tissue is a logical choice.

As a result from the qPCR experiment, significantly altered expression was confirmed for nine of the ten genes analyzed. PPBP, SpiC, VCAM1, ENPEP and ITGAD were downregulated and GTSF1, LUM, Thy1 and Col3a1 were upregulated in both STHS and VHS compared to normal spleen. Table 4 shows the adjusted p-values and Fold Change.

Table 4. Genes identified as potential interesting using gene profiling in both VHS and STHS compared to spleen.

The use of spleen as the healthy equivalent of HS does raise some concern as to how observed differences in Spi-C and VCAM1 gene expression between tumors and healthy spleen must be explained. It is possible that some of these expression differences are based on differences in tissue-origin rather than on actual tumor development. Spi-C plays a critical role in the development of splenic iron homeostasis. It is highly expressed in red pulp macrophages, but not monocytes, dendritic cells or other tissue macrophages. Spi-C is therefore highly expressed in spleen [35], and could thus lead to cause a seemingly relative down-regulation in HS tissues. Spi-C is also known to regulate VCAM1 expression [35]. This gene is thought to be involved in angiogenesis and is induced by cytokines on endothelial cells [36]. Since spleen tissue contains abundant endothelium, this could cause the relative high expression of VCAM1 in the spleen. ITGAD (also known as CD11d) is a receptor for VCAM1. In our study, we detected a lowered expression of CD11d in HS as compared to normal spleen.

Staining by immunohistochemistry for the presence of CD11d in both STHS and VHS was found negative in a first study by Moore et al, including16 splenic HS [6] and this absence of the CD11d protein in HS was seen as one of the phenotypically characteristics of a myeloid dendritic antigen-presenting cell lineage, making many HS to be likely myeloid dendritic cell sarcomas [6]. In contrast, a more recent study in Flatcoated Retrievers found the majority (12/20) of splenic HS positive for CD11d [8] and was interpreted by the investigators as marker of a likely red-pulp macrophagocytic origin of these splenic HS. A similar rate of CD11d positivity was noticed by another study by Moore et al examining hemophagocytic HS in spleen in a series of dogs from 6 breeds [7]. CD11d proteins appear to be strongly expressed only on mature granulocytes, monocytes, and certain lymphocytes, but not significantly on myeloid committed precursor cells [37] and hence, the low expression in many HS is no surprise as is its positivity in hemophagocytic HS. Still, the finding of positive expression in splenic HS by immunohistochemistry [8] is at variance with earlier studies [6] and was interpreted as marker of a likely red-pulp macrophagocytic origin of these splenic HS. In our study, we detected a lowered expression of CD11d in HS as compared to normal spleen.

The down-regulation of PPBP as was found in this study, has also been reported to play a role in the development of myelomas [38] as well as pancreatic cancers [39], in the latter report the PPBP (CXCL7) plasma-level was even postulated to be an interesting biomarker for early detection [39]. The close relation in the origin of myeloma and histiocytic sarcoma, both stemming from lineages of hematoproliferative compartments makes this downregulation of PPBP an intriguing observation worthy of further pursuit.

Expression and overexpression of LUM has been observed in various types of cancer cells (colorectal, pancreatic, and breast cancer, melanoma, neuroendocrine cell tumors) - with contrasting findings on the relation with type of growth and/or tumor progression or metastasis [40][46] - and in activated synoviocytes from rheumatoid arthritis [47]. The function of lumican - a member of the family of small leucine rich proteoglycans - in the organization of the extracellular matrix composition as well as in migration and proliferation- in relation to the observed overexpression in HS warrants further investigations, including the site of overexpression (tumor cell or tumor-associated fibroblasts) [48].

In humans, an up regulation of GTSF1 is already known in the occurrence of mycosis fungoides, which is the most common type of primary cutaneous T-cell lymphoma, in which GTSF1 was proposed as a gene for which expression appears to be restricted to mycosis fungoides tumor stage and that might even serve as diagnostic (bio)marker [49].

ENPEP probably plays a role in regulating growth and differentiation of early B-lineage cells ( and down regulation may thus also be involved in HS development.

Finally, Col3a1 also has been shown to be overexpressed in other types of tumors, such as malignant mesothelioma [50], as well as in human sarcoma xenotransplants [51]. Its overexpression could have a significant influence upon extracellular matrix composition [52].

Thy1 (CD90) is an important marker of many types of stem cells [53], including mesenchymal stem cells [54]. CD90 has already been identified as a candidate marker for cancer stem cells in primary high-grade gliomas using tissue microarrays [55], [56]. For human hepatocellular carcinoma, CD90 has even been shown to provide a clinical prognostic marker [53]. Our observation of overexpression of CD90 in HS might herald stem cell characteristics of the type of cancer.

For technical reasons, no qPCR data could be obtained for Myh11.

As a result of the pathway analyses in DAVID, 24 pathways were significantly involved in the development of HS (P<0.05). Most were involved in the DNA repair and replication process. The biological functions of ten of these pathways, amongst which the P53 signaling pathway was one of the most relevant, are listed in Table 5.

Table 5. Ten of the most significant pathways that resulted from the pathway analyses in DAVID.

Our observations provide evidence of an association between altered expression of nine genes (PPBP, SpiC, VCAM1, ENPEP, ITGAD, GTSF1, COL3A1, CD90 and LUM) and development of HS in the Flatcoated retriever dog, irrespective of the disseminated or localized form.

Based upon a fundamental and evolutionarily conserved association between cytogenetic abnormalities and tumor phenotype in different species [10], [57] these genes may be of major interest in the study of histiocytic malignancies in the human as well. There exists a great difference in incidence of histiocytic malignancies between this specific dog breed as compared to the human. This implies a genetic make-up predisposing the Flatcoated retriever (and the Bernese Mountain Dog); which is uncommon in many other dog breeds and humans alike. For the Bernese Mountain Dog, a first genetic locus has been identified. Our group takes part in a study aiming to identify predisposing genes in the Flat Coated Retriever in the hope, that these findings may provide clues for the related cancer in the human.

The current study provides the most comprehensive database of genome alterations in histiocytic malignancies to date, revealing genes and signaling pathways not previously associated with this disease. Although mRNA levels do not necessary reflect differences in protein levels, it is very well conceivable that the large difference in mRNA levels of specific genes will result in quantitative differences in protein expression. Lack of verified and specific antibodies for all nine gene products of interest let us to restrain this expression profiles to mRNA levels only. The study of Hedan [10] was able to locate recurrent genomic imbalances using CGH. As indicated in Table 3 of their publication [10], 808 genes found to be located in their regions of interest. This covers about 4% of the total number of genes. In the Agilent Canine Gene Expression Microarrays V1 that were used in our study, 430 of these 808 genes were annotated. Three of these annotated genes; ENSCAFG00000007012 (SPIC), ENSCAFG00000001046 (DESI1) and ENSCAFG00000006138 (LUM) were found to be commonly involved in our study as well as in Hedans study. In our study, eventually two of these genes, SPIC and LUM, were chosen for qPCR confirmation. We were indeed able to identify the significantly altered expression of these two genes.


This is the first study to compare gene expression in HS (both STHS and VHS) and normal spleen using both traditional fold change analysis as well as disease-based pathway analyses using DAVID.

This study provides evidence for involvement of several genes in HS, irrespective of the form of manifestation, some of which are also related with to other cancers. On the basis of quantitative differences in expression, we consider PPBP, SpiC, VCAM1, ENPEP, ITGAD (down-regulated), and GTSF1, Col3a1, CD90 and LUM (up-regulated) to be associated with the HS genotype.

Extrapolation of this data to human samples may help to further our understanding of the propagation and oncogenesis of histiocytic cells. Eventually, this will contribute to the development of effective therapeutic modalities for both species.

Author Contributions

Conceived and designed the experiments: KB FvS MGK DvL LP GR. Performed the experiments: KB MvdK MvW MGK DvL EW GG. Analyzed the data: KB MvW FvS MGK DvL EW GR. Contributed reagents/materials/analysis tools: KB MGK DvL GR GG FvS. Wrote the paper: KB MvW FvS LP GG GR MGK DvL MvdK EW.


  1. 1. Shearin AL, Hedan B, Cadieu E, Erich SA, Schmidt EV, et al.. (2012) The MTAP-CDKN2A locus confers susceptibility to a naturally occurring canine cancer. Cancer Epidemiol Biomarkers Prev. 10.1158/1055–9965.EPI-12–0190-T.
  2. 2. Dobson J, Hoather T, McKinley TJ, Wood JL (2009) Mortality in a cohort of flat-coated retrievers in the UK. Vet Comp Oncol 7: 115–121 10.1111/j.1476–5829.2009.00181.x.
  3. 3. Fidel J, Schiller I, Hauser B, Jausi Y, Rohrer-Bley C, et al. (2006) Histiocytic sarcomas in flat-coated retrievers: A summary of 37 cases (november 1998-march 2005). Vet Comp Oncol 4: 63–74 10.1111/j.1476–5810.2006.00090.x.
  4. 4. Fulmer AK, Mauldin GE (2007) Canine histiocytic neoplasia: An overview. Can Vet J 48: 1041–3, 1046–50.
  5. 5. Scott DW, Miller WH Jr, Tasker JB, Schultz RD, Meuten DJ (1979) Lymphoreticular neoplasia in a dog resembling malignant histiocytosis (histiocytic medullary reticulosis) in man. Cornell Vet 69: 176–197.
  6. 6. Affolter VK, Moore PF (2002) Localized and disseminated histiocytic sarcoma of dendritic cell origin in dogs. Vet Pathol 39: 74–83.
  7. 7. Moore PF, Affolter VK, Vernau W (2006) Canine hemophagocytic histiocytic sarcoma: A proliferative disorder of CD11d+ macrophages. Vet Pathol 43: 632–645 10.1354/vp.43-5-632.
  8. 8. Constantino-Casas F, Mayhew D, Hoather TM, Dobson JM (2011) The clinical presentation and histopathologic-immunohistochemical classification of histiocytic sarcomas in the flat coated retriever. Vet Pathol 48: 764–771 10.1177/0300985810385153.
  9. 9. Beverley PC, Egeler RM, Arceci RJ, Pritchard J (2005) The nikolas symposia and histiocytosis. Nat Rev Cancer 5: 488–494 10.1038/nrc1632.
  10. 10. Hedan B, Thomas R, Motsinger-Reif A, Abadie J, Andre C, et al. (2011) Molecular cytogenetic characterization of canine histiocytic sarcoma: A spontaneous model for human histiocytic cancer identifies deletion of tumor suppressor genes and highlights influence of genetic background on tumor behavior. BMC Cancer 11: 201 10.1186/1471–2407–11–201.
  11. 11. Abadie J, Hedan B, Cadieu E, De Brito C, Devauchelle P, et al. (2009) Epidemiology, pathology, and genetics of histiocytic sarcoma in the bernese mountain dog breed. J Hered 100 Suppl 1S19–27 10.1093/jhered/esp039.
  12. 12. Pileri SA, Grogan TM, Harris NL, Banks P, Campo E, et al. (2002) Tumours of histiocytes and accessory dendritic cells: An immunohistochemical approach to classification from the international lymphoma study group based on 61 cases. Histopathology 41: 1–29.
  13. 13. Meuten DJ (2002) Tumors in domestic animals. : Iowa State Press; A Blackwell Publishing Company. 788 p.
  14. 14. Sterenczak KA, Kleinschmidt S, Wefstaedt P, Eberle N, Hewicker-Trautwein M, et al. (2011) Quantitative PCR and immunohistochemical analyses of HMGB1 and RAGE expression in canine disseminated histiocytic sarcoma (malignant histiocytosis). Anticancer Res 31: 1541–1548.
  15. 15. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, et al. (2001) Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 98: 10869–10874 10.1073/pnas.191367098.
  16. 16. Baird K, Davis S, Antonescu CR, Harper UL, Walker RL, et al. (2005) Gene expression profiling of human sarcomas: Insights into sarcoma biology. Cancer Res 65: 9226–9235 10.1158/0008–5472.CAN-05–1699.
  17. 17. Lindblad-Toh K, Wade CM, Mikkelsen TS, Karlsson EK, Jaffe DB, et al. (2005) Genome sequence, comparative analysis and haplotype structure of the domestic dog. Nature 438: 803–819 10.1038/nature04338.
  18. 18. Tamburini BA, Phang TL, Fosmire SP, Scott MC, Trapp SC, et al. (2010) Gene expression profiling identifies inflammation and angiogenesis as distinguishing features of canine hemangiosarcoma. BMC Cancer 10: 619 10.1186/1471–2407–10–619.
  19. 19. Mahoney JA, Fisher JC, Snyder SA, Hauck ML (2010) Feasibility of using gene expression analysis to study canine soft tissue sarcomas. Mamm Genome 21: 577–582 10.1007/s00335-010-9298-y.
  20. 20. Tamburini BA, Trapp S, Phang TL, Schappa JT, Hunter LE, et al. (2009) Gene expression profiles of sporadic canine hemangiosarcoma are uniquely associated with breed. PLoS One 4: e5549 10.1371/journal.pone.0005549.
  21. 21. Paoloni M, Khanna C (2008) Translation of new cancer treatments from pet dogs to humans. Nat Rev Cancer 8: 147–156 10.1038/nrc2273.
  22. 22. Hoffman MM, Birney E (2007) Estimating the neutral rate of nucleotide substitution using introns. Mol Biol Evol 24: 522–531 10.1093/molbev/msl179.
  23. 23. Danilenko DM, Rossitto PV, Van der Vieren M, Le Trong H, McDonough SP, et al. (1995) A novel canine leukointegrin, alpha d beta 2, is expressed by specific macrophage subpopulations in tissue and a minor CD8+ lymphocyte subpopulation in peripheral blood. J Immunol 155: 35–44.
  24. 24. Opitz L, Salinas-Riester G, Grade M, Jung K, Jo P, et al. (2010) Impact of RNA degradation on gene expression profiling. BMC Med Genomics 3: 36 10.1186/1755-8794-3-36.
  25. 25. Roepman P, Wessels LF, Kettelarij N, Kemmeren P, Miles AJ, et al. (2005) An expression profile for diagnosis of lymph node metastases from primary head and neck squamous cell carcinomas. Nat Genet 37: 182–186 10.1038/ng1502.
  26. 26. van de Peppel J, Kemmeren P, van Bakel H, Radonjic M, van Leenen D, et al. (2003) Monitoring global messenger RNA changes in externally controlled microarray experiments. EMBO Rep 4: 387–393 10.1038/sj.embor.embor798.
  27. 27. Yang YH, Dudoit S, Luu P, Lin DM, Peng V, et al. (2002) Normalization for cDNA microarray data: A robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res 30: e15.
  28. 28. Margaritis T, Lijnzaad P, van Leenen D, Bouwmeester D, Kemmeren P, et al. (2009) Adaptable gene-specific dye bias correction for two-channel DNA microarrays. Mol Syst Biol 5: 266 10.1038/msb.2009.21.
  29. 29. Wu H, Kerr M, Cui X, Churchill G (2003) MAANOVA: A software package for the analysis of spotted cDNA microarray experiments. In: Anonymous The Analysis of Gene Expression Data.. 313–313–341.
  30. 30. Schlotter YM, Veenhof EZ, Brinkhof B, Rutten VP, Spee B, et al. (2009) A GeNorm algorithm-based selection of reference genes for quantitative real-time PCR in skin biopsies of healthy dogs and dogs with atopic dermatitis. Vet Immunol Immunopathol 129: 115–118 10.1016/j.vetimm.2008.12.004.
  31. 31. Brinkhof B, Spee B, Rothuizen J, Penning LC (2006) Development and evaluation of canine reference genes for accurate quantification of gene expression. Anal Biochem 356: 36–43 10.1016/j.ab.2006.06.001.
  32. 32. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, et al. (2002) Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol 3: RESEARCH0034.
  33. 33. van Steenbeek FG, Van den Bossche L, Grinwis GC, Kummeling A, van Gils IH, et al. (2013) Aberrant gene expression in dogs with portosystemic shunts. PLoS One 8: e57662 10.1371/journal.pone.0057662; 10.1371/journal.pone.0057662.
  34. 34. van Steenbeek FG, Spee B, Penning LC, Kummeling A, van Gils IH, et al. (2013) Altered subcellular localization of heat shock protein 90 is associated with impaired expression of the aryl hydrocarbon receptor pathway in dogs. PLoS One 8: e57973 10.1371/journal.pone.0057973; 10.1371/journal.pone.0057973.
  35. 35. Kohyama M, Ise W, Edelson BT, Wilker PR, Hildner K, et al. (2009) Role for spi-C in the development of red pulp macrophages and splenic iron homeostasis. Nature 457: 318–321 10.1038/nature07472.
  36. 36. Osborn L, Hession C, Tizard R, Vassallo C, Luhowskyj S, et al. (1989) Direct expression cloning of vascular cell adhesion molecule 1, a cytokine-induced endothelial protein that binds to lymphocytes. Cell 59: 1203–1211.
  37. 37. Falkenburg JH, Koning F, Duinkerken N, Fibbe WE, Voogt PJ, et al. (1986) Expression of CD11, CDw15, and transferrin receptor antigens on human hematopoietic progenitor cells. Exp Hematol 14: 90–96.
  38. 38. Davies FE, Dring AM, Li C, Rawstron AC, Shammas MA, et al. (2003) Insights into the multistep transformation of MGUS to myeloma using microarray expression analysis. Blood 102: 4504–4511 10.1182/blood-2003-01-0016.
  39. 39. Matsubara J, Honda K, Ono M, Tanaka Y, Kobayashi M, et al. (2011) Reduced plasma level of CXC chemokine ligand 7 in patients with pancreatic cancer. Cancer Epidemiol Biomarkers Prev 20: 160–171 10.1158/1055-9965.EPI-10-0397.
  40. 40. Seya T, Tanaka N, Shinji S, Yokoi K, Koizumi M, et al. (2006) Lumican expression in advanced colorectal cancer with nodal metastasis correlates with poor prognosis. Oncol Rep 16: 1225–1230.
  41. 41. Sifaki M, Assouti M, Nikitovic D, Krasagakis K, Karamanos NK, et al. (2006) Lumican, a small leucine-rich proteoglycan substituted with keratan sulfate chains is expressed and secreted by human melanoma cells and not normal melanocytes. IUBMB Life 58: 606–610 10.1080/15216540600951605.
  42. 42. Troup S, Njue C, Kliewer EV, Parisien M, Roskelley C, et al. (2003) Reduced expression of the small leucine-rich proteoglycans, lumican, and decorin is associated with poor outcome in node-negative invasive breast cancer. Clin Cancer Res 9: 207–214.
  43. 43. Vuillermoz B, Khoruzhenko A, D'Onofrio MF, Ramont L, Venteo L, et al. (2004) The small leucine-rich proteoglycan lumican inhibits melanoma progression. Exp Cell Res 296: 294–306 10.1016/j.yexcr.2004.02.005.
  44. 44. Shinji S, Tajiri T, Ishiwata T, Seya T, Tanaka N, et al. (2005) Different expression levels of lumican in human carcinoid tumor and neuroendocrine cell carcinoma. Int J Oncol 26: 873–880.
  45. 45. D'Onofrio MF, Brezillon S, Baranek T, Perreau C, Roughley PJ, et al. (2008) Identification of beta1 integrin as mediator of melanoma cell adhesion to lumican. Biochem Biophys Res Commun 365: 266–272 10.1016/j.bbrc.2007.10.155.
  46. 46. Ishiwata T, Cho K, Kawahara K, Yamamoto T, Fujiwara Y, et al. (2007) Role of lumican in cancer cells and adjacent stromal tissues in human pancreatic cancer. Oncol Rep 18: 537–543.
  47. 47. Mori H, Nishida K, Ozaki T, Inoue H, Nakanishi T (2008) Isolation of a mRNA preferentially expressed in synoviocytes from rheumatoid arthritis that is identical with lumican, which encodes a collagen binding, extracellular matrix protein. 17: 125–126–130.
  48. 48. Nikitovic D, Katonis P, Tsatsakis A, Karamanos NK, Tzanakakis GN (2008) Lumican, a small leucine-rich proteoglycan. IUBMB Life 60: 818–823 10.1002/iub.131.
  49. 49. van Kester MS, Borg MK, Zoutman WH, Out-Luiting JJ, Jansen PM, et al. (2012) A meta-analysis of gene expression data identifies a molecular signature characteristic for tumor-stage mycosis fungoides. J Invest Dermatol 132: 2050–2059 10.1038/jid.2012.117.
  50. 50. Gueugnon F, Leclercq S, Blanquart C, Sagan C, Cellerin L, et al. (2011) Identification of novel markers for the diagnosis of malignant pleural mesothelioma. Am J Pathol 178: 1033–1042 10.1016/j.ajpath.2010.12.014.
  51. 51. Berner JM, Müller CR, Holden M, Wang J, Hovig E, et al. (2006) Sampling effects on gene expression data from a human tumour xenograft. 33: 17–18–30.
  52. 52. Ross DT, Scherf U, Eisen MB, Perou CM, Rees C, et al. (2000) Systematic variation in gene expression patterns in human cancer cell lines. Nat Genet 24: 227–235 10.1038/73432.
  53. 53. Lu JW, Chang JG, Yeh KT, Chen RM, Tsai JJ, et al. (2011) Overexpression of Thy1/CD90 in human hepatocellular carcinoma is associated with HBV infection and poor prognosis. Acta Histochem 113: 833–838 10.1016/j.acthis.2011.01.001.
  54. 54. Mercati F, Pascucci L, Ceccarelli P, Dall'Aglio C, Pedini V, et al. (2009) Expression of mesenchymal stem cell marker CD90 on dermal sheath cells of the anagen hair follicle in canine species. Eur J Histochem 53: 159–166.
  55. 55. He J, Liu Y, Zhu T, Zhu J, Dimeco F, et al. (2012) CD90 is identified as a candidate marker for cancer stem cells in primary high-grade gliomas using tissue microarrays. Mol Cell Proteomics 11: M111.010744 10.1074/mcp.M111.010744.
  56. 56. Parry PV, Engh JA (2012) CD90 is identified as a marker for cancer stem cells in high-grade gliomas using tissue microarrays. Neurosurgery 70: N23–4 10.1227/01.neu.0000413227.80467.92; 10.1227/01.neu.0000413227.80467.92.
  57. 57. Breen M, Modiano JF (2008) Evolutionarily conserved cytogenetic changes in hematological malignancies of dogs and humans–man and his best friend share more than companionship. Chromosome Res 16: 145–154 10.1007/s10577-007-1212-4.