Warning! You won't be able to use the quotation basket until you enable cookies in your Web browser.
Warning! Your Web browser is no longer supported. Please upgrade to a modern browser.

DOG-1 – A Breed Showing K9 Excellence

Gastrointestinal stromal tumour (GIST) is a common mesenchymal tumour found in the gastrointestinal tract.  Morphologically it may resemble a wide variety of other tumours.

The incidence of GIST is estimated at 10- 20 cases per million per year.  Since the detection of CD117 / c-kit gene proto-oncogene expression in these tumours, the classification and diagnosis of GISTs has changed drastically.

While surgery is the main course of treatment for resectable GISTs, other modalities such as receptor tyrosine kinase (RTK) inhibitors show promising results.

DOG-1 (Discovered on GIST) is a protein of unknown function which is highly expressed in GISTs.

While the immunohistochemical expression of CD117 plays a vital role in the diagnosis and treatment of GISTs, DOG-1 has shown to stain 13% more GISTs than CD117 and thus renders it a more sensitive and specific marker than CD117.

Introduction

Gastrointestinal stromal tumours (GISTs) are a subset of gastrointestinal mesenchymal tumours of varying differentiation 1.  Historically, these lesions were classified as leiomyomas or leiomyosarcomas as they possessed smooth muscle features when examined by light microscopy 2.

Since the term GIST was introduced by Mazur and Clark in 1983, laboratory investigations which were aimed at the subcellular and molecular levels, have demonstrated that GISTs do not possess the ultra-structural and immunocytochemical features characteristic of smooth muscle differentiation as are seen in leiomyomas and leiomyosarcomas 2.

In all, there are approximately 3500-5000 cases of GIST per year in the United States. This makes GIST the most common form of stromal tumour, which constitutes more than 70 types of cancer, but in all forms constitutes less than 1% of all cancer 1.  Review of the literature indicates that 70% of these tumours occur in the stomach, 20% in the small intestine and less than 10% in the esophagus 2.

A small percentage of GISTs occur outside the  gastrointestinal tract (GIT), in the abdominal cavity, mesentery, omentum, retroperitoneum and pelvic cavity; and are known as extra-gastrointestinal stromal tumours (EGIST) 3.

Kindblom et al, 1998 and others believe that the interstitial cell of Cajal (ICC), the proposed pacemaker cells function in controlling motility, represent the cell of origin of GIST 4.

Perhaps the most critical development that distinguished GIST as a unique clinical entity was the discovery of c-kit proto-oncogene mutations in these tumours by Hirota and colleagues in 1998 3.

Currently, accepted terminology for mesenchymal tumours of the GIT assigns true smooth muscle and neural tumours into separate categories from spindle or epithelioid GISTs which demonstrate CD117 and / or CD34 immunochemical staining 3, 4, 5.

Therefore, the determination was made that GISTs do not arise from smooth muscle cells, but from another mesenchymal derivative such as the progenitors of spindle and epithelioid cells 3.

GISTs have either a spindle, epithelioid or mixed cell appearance.  The spindle cell pattern consists of uniform spindle cells arranged in whorls or short intersecting fascicles.  Nuclear palisading and extensive stromal hyalinisation are common 3, 4.

The epithelioid pattern is composed mainly of epithelioid cells with a morphology of scattered multinucleate cells with well-defined cell membranes and round nuclei with small nucleoli.  The epithelioid cells have a perinuclear, condensed rim of eosinophillic cytoplasm, often with a peripheral zone of cytoplasmic clearing. The mixed pattern is a combination of the spindle and epithelioid cells.

CD117/ c-kit; a trans-membrane tyrosine kinase growth factor receptor involved in  cellular differentiation; is expressed by the ICC, melanocytes, mast cells and in GISTs 4.

Mutations in the receptor tyrosine kinase (RTK) genes c-kit or Platelet derived growth factor receptor alpha (PDGFRA), cause activation of tyrosine kinase, which promote proliferation and decrease apoptosis.  Therefore, the CD117 expression by GIST plays a key role in the diagnosis and treatment of these patients 2, 5, 6.

About 80% of GISTs have c-kit mutations and approximately 5-8% PDGFRA mutations.  Another 12-15%, known as wild-type GIST, lacks the mutation of both these genes 2, 5, 6.  Until recently, the choice of treatment for GIST was limited to radiotherapy and surgery 2, 6.

Imatinib mesylate (Gleevec) a RTK inhibitor, which was originally developed for the treatment of chronic myeloid leukaemia, has proved to be an effective treatment of GISTs 2, 7, 8.  Its action directly suppresses c-kit expression and PDGFRA activity.  A CD117 immunochemical positive tumour therefore indicates Imatinib sensitivity 7, 8, 9.

Wild type GISTs are resistant to Imatinib, but may respond to other RTK inhibitors such as Sunitinib (Sutent).  Sunitinib is an alternative option for patients who either fail Imatinib therapy or develop resistance (secondary mutations) to Imatinib 5, 6.

While the diagnosis of GIST relies heavily on the anatomical location of the tumour, histological appearance and immunohistochemical expression of CD117 and CD34, the 4 - 15 % of GISTs that are weakly positive or negative for CD117, remain a significant diagnostic challenge 1, 5, 6.

While CD117 remains a relatively specific marker for GIST within the stromal tumours, it can stain positive for a significant number of seminomas and melanomas 1, 3.

DOG-1 is a protein found to be selectively expressed in GISTs.  While its function remains unclear, it cited as a possible ion channel, due to the presence of 8 transmembrane regions.  The DOG-1 gene is localised on chromosome 11 (11q13) 7, 8.

In a study undertaken by Espinosa et al., 2008, the staining of DOG-1 when compared with other non-sarcomatous lesions and a variety of carcinomas from 25 primary sites that included lymphomas, brain tumours, and other types of tumours, showed that DOG-1 was expressed weakly with variable intensity in 10% breast tumours and 6% colon tumours 8.

Most of these cases were negative for DOG-1, while CD117 on the other hand, stained 85% seminomas, 66% melanomas as well as occasional cases of liver, pancreas, kidney, bladder and endometrial carcinoma 8.

Espinosa et al., 2008 report that only 0.1% desmoplastic melanoma were positive for DOG-1. No seminomas stained positively.  As seminomas and melanomas are part of the differential diagnosis of GIST, this data has clinical significance.  DOG -1 antibody stained 87% of GISTs with the c-kit mutation, whereas CD117 antibody stained 74% and CD34 only 59%.

This group further report that of the GISTs harbouring PDGFRA mutations, 79% stained with DOG-1, 9% with CD117 and 27% with CD34 8 (Table 1, below).

Review of the literature reports that about a third of patients who harbour PDGFRA mutations fail to stain with CD117 5.  It was predicted that these cases could be sensitive to Imatinib.  The other two thirds of cases may respond to other RTK inhibitor therapy 8, 9, 10.

In the wild-type GISTs, DOG-1 staining is reported in 89%, CD117 in 83% and CD34 in 51% cases respectively 8, 10.

Table 1. The expression of DOG-1, CD117 and CD34 in GISTs with c-kit, PDGFRA and wild-type mutations (Espinosa et al., 2008).

Mutations

DOG-1

CD117

CD34

Total GISTs

87%

74%

59%

C-KIT

92%

81%

64%

PDGFRA

79%

9%

27%

Wild-type

89%

83%

51%

Materials and Methods

Tissues were fixed in 10% neutral-buffered formalin and processed to paraffin wax by conventional laboratory methods.  Sections were cut a 3μm, mounted on commercially available positively charged glass slides and incubated at 60ºC for 30 minutes.

Twelve cases of GISTs which had been diagnosed on H+E; were stained (n=6 stomach, n=3 small bowel, n=1 colon, n=1 anal rectal area and n=1 retroperitoneal tumour).

Of these GISTs, n=8 spindle cell tumours, n=2 epithelioid and n=2 a mixture of spindle and epithelioid cells (Table 2).

At initial optimisation, the appropriate positive and negative controls were included to exclude false positive and non-specific staining.

The Novocastra DOG-1 monoclonal antibody, (NCL-L-DOG1; clone K9) was optimised using sections of normal gastric mucosa, normal breast and spindle cell GIST.  Staining of these controls produce the expected pattern of staining which was localised to the membranes and cytoplasm of gastric mucosa, as well as glands of the breast.  DOG-1 antibody staining is not expressed in ICC.

Antigen retrieval and staining was performed on the Leica BOND-MAX™ instrument using the BOND Polymer Refine Detection System™ and a DAB Enhancer™.

Antigen retrieval with either BOND Epitope Retrieval 1 (ER1™) or BOND Epitope Retrieval 2 (ER2™) solutions at 96°C for 20 minutes gave optimal results on the FFPE material from this site.

A primary antibody dilution of 1:100 produced optimal staining when applied for 15 minutes at ambient temperature.

Hydrogen peroxide treatment was carried out after the polymer step, minimising any non-specific endogenous peroxidase reactivity.

Based on the clinical information provided, DOG-1 was included for staining as part of a panel that included, but was not restricted to the following markers: CD117, CD34, SMA and Desmin.  A pilot study is currently investigating the mutation status of these cases.

Results

Table 2.  Immunohistochemical results of DOG-1, CD117, CD34 and others in various types and locations of GISTs.

Case nr.

Type of GIST

Site

Risk

DOG-1

CD117

OTHERS

1

Spindle

Stomach

LR

+++

+++

CD34 +++; SMA +/w; S100 -; Caldesmon -

2

Spindle

Stomach

LR

+++

+++

S100 -; Desmin -; SMA -

3

Spindle

Stomach

LR

+++

+++

 

4

Spindle

Stomach subserosa

LR

++

++

 

5

Spindle

Small bowel

LR

+++

+/w

 

6

Spindle

Small bowel

HR

+++

+++

SMA -; Desmin -; S100 -

7

Spindle

Small bowel

LR

+++

+++

CD34 -; SMA -; S100 -; Desmin -

8

Spindle

Retroperitoneal

HR

+++

+++

CD34 +++; VIM +++; MYF4 -; SMA -; S100 -; Desmin -

9

Epithelioid

Stomach

LR

++

+/w

 

10

Epithelioid

Anal/rectal

 

-

+++

CD34 -; MNF -; CD45 -; CD20 -; CD3 -; CD30 -; MPO -; S100 -; Melan A -; Desmin -; SMA -; MYF4 -; PLAP -; Synaptophysin -; Chromogranin -

11

Mixed

Colon

HR

+++

+/w

Vim +/w; Desmin -; S100 -;   AE1/3 -; CAM 5.2 +/w

12

Mixed

Stomach

LR

+/w

-

CD34 +/w; MNF -; CD45 -

 

Key:  Intensely positive: +++; moderately positive: ++; weakly positive with focal positivity: +/w; Negative: -; Low risk: LR; High risk: HR

Discussion

The staining pattern of DOG-1 was predominantly membranous in the epithelioid variant of GIST.  It was noted that membranous accentuation of the spindle cell GISTs was often accompanied by cytoplasmic staining.

The overall staining of DOG-1 demonstrated superior, crisp localisation when compared with the diffuse staining pattern of the CD117.

A stronger staining intensity with DOG-1 was observed in all of the GISTs with the exception of one case that was negative for DOG-1, when compared with the staining of the CD117.  The staining in this case was repeated to exclude any technical issues.  It is thought that this case could be negative due to genetic mutation.  No data on patient therapeutic intervention such as radiotherapy was available.

In three of the cases, which were from no specific morphologic subtype, CD117 showed partial weak staining of the tumour and some areas were completely negative (Figures 1 and 3).  This can potentially lead to false negative results if biopsy specimens are taken from these areas.  C-kit mutations appear to occur preferentially in the spindle rather than the epithelioid cell variant of GIST.  Consistent and even staining of the entire tumour was noted with DOG-1 antibody (Figures 2 and 4).

Figure 1: CD117 showed partial weak staining
Figure 1: Image illustrating that CD117 showed partial weak staining
Figure 2: DOG-1 showed consistent and even staining
Figure 2: Image illustrating that DOG-1 showed consistent and even staining
Figure 31: CD117 showed partial weak staining
Figure 31: Image illustrating that CD117 showed partial weak staining
Figure 4: DOG-1 showed consistent and even staining
Figure 4: Image illustrating that DOG-1 showed consistent and even staining

Conclusion

The staining of DOG-1 antibody has produced an intense and well-localised immuno-reaction that was easy to interpret.

DOG-1 antibody has also consistently stained the entire tumour, whereas with CD117 there are some areas that are weak or completely negative.  It is unclear if this is due to particular variants of GIST.  Despite current targeted RTK therapies, the accurate diagnosis of GIST plays a vital role in the treatment of patients.

Unfortunately, routine screening for c-kit and PDGFRA mutations can be helpful, but it is expensive, time consuming and skill dependent.

While the numbers of cases compared are limited, we have concluded that the staining of DOG-1 antibody can potentially diagnose approximately 13% more GIST’s on IHC than CD117 staining.  Additionally, due to its failure to react with most other tumours, this makes it a sensitive and specific marker for GIST, warranting further investigation for the inclusion of this marker as an adjuvant to the panel of antibodies for the accurate identification and diagnosis of GIST.

References

  1. Miettinen M, Lasota J. Gastrointestinal stromal tumors: review on morphology, molecular pathology, prognosis, and differential diagnosis. Arch. Pathol. Lab. Med 2006; 130 (10): 1466–78
  2. News: Genetic Variations in GI Tumors Determine Which Medications Are Efficacious. Genetic Engineering & Biotechnology News. 13 Nov 2008.
  3. Mazur MT, Clark HB. Gastric stromal tumors. Reappraisal of histogenesis. Am J Surg Pathol. Sep 1983; 7(6):507-19
  4. Kindblom LG, Remotti HE, Aldenborg F, Meis-Kindblom JM. Gastrointestinal pacemaker cell tumor (GIPACT): gastrointestinal stromal tumors show phenotypic characteristics of the interstitial cells of Cajal. Am J Pathol. May 1998; 152(5):1259-69
  5. Hirota S, Isozaki K, Moriyama Y, Hashimoto K, Nishida T, Ishiguro S, et al. Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science. Jan 23 1998; 279(5350):577-80.
  6. Stomatakos M, Douzinas E, et al. Gastrointestinal Stromal Tumour. World Journal of Surgical Oncology 2009; 10.1186/1477-7819-7-61
  7. Berman, J, O'Leary, T. J. Gastrointestinal stromal tumour workshop. Human Pathology 2001; 32:578-582
  8. Espinosa I, Lee CH, Kim MK, et al. A novel monoclonal antibody against DOG1 is a sensitive and specific marker for gastrointestinal stromal tumours. Am J Surg Pathol 2008; 32:210-8
  9. Heinrich MC, Coreless CL, et al. PDGFRA activating mutations in gastrointestinal stromal tumours. Science 2003; 299:708-710
  10. Rubin B. Gastrointestinal stromal tumours: an update. Histopathology 2006; 48(1):83-96