Article Text

Download PDFPDF

Cytokeratin expression by CD34 positive blasts in a case of refractory anaemia with excess of blasts in transformation (RAEB-t)
  1. J-L Dargent,
  2. K Jochmans,
  3. M De Waele,
  4. R Schots,
  5. C Bourgain
  1. Department of Pathology, CHU Saint-Pierre/ULB Institut Jules Bordet, 1 rue Héger-Bordet, B-1000 Brussels, Belgium jldargent{at}
  2. Laboratory of Haematology, Academisch Ziekenhuis-VUB, 101 Laarbeeklaan, B-1090 Brussels
  3. Department of Haemato-Oncology, Academisch Ziekenhuis-VUB
  4. Department of Pathology, Academisch Ziekenhuis-VUB

    Statistics from

    Immunohistochemistry has become a very important, and in some cases indispensable, tool in diagnostic pathology, enabling the precise identification of tumours, the detection of micrometastases in a given sample, and the evaluation of various prognosis factors. However, in some cases, the use of multiple but distinct immunostains can lead to some unforeseen results—for example, the expression of an apparently aberrant marker by a neoplasm can sometimes be seen. In this context, we report our experience with a case of refractory anaemia with excess of blasts in transformation (RAEB-t) in which the blasts were unexpectedly found to express cytokeratin (CK).

    An 86 year old woman with a past medical history of breast carcinoma treated by mastectomy and adjuvant radiotherapy was admitted to our institution because of worsening anaemia. The following haematological indices were noticed: haemoglobin, 8.6 g/litre; erythrocytes, 2.5 × 1012/litre; white blood cells, 3 × 109/litre; and platelets, 465 × 109/litre. Blasts were also found in peripheral blood (11%). Both bone marrow aspirate and biopsy demonstrated features suggestive of a myelodysplastic syndrome (MDS) (fig 1), identified as an RAEB-t according to the criteria of the French–American–British cooperative group (FAB).1 Cytochemical study of the blasts revealed the presence of α-napthyl acetate esterase but not of myeloperoxidase (MPO) or napthol ASD chloroacetate esterase. On flow cytometry, the blasts were found to be CD34 positive with expression to some extent of CD13, CD33, CD45, CD45RA, CD117 (c-kit gene product), and CD56. No expression of CD10, CD19, CD45RO or CD90 could be demonstrated. Immunohistochemical studies performed on paraffin wax embedded sections demonstrated CD34 and CD45 positivity in the blasts (fig 2A). However, these cells were MPO, CD3, and CD20 negative. CD68 expression was variable. To exclude with certainty the possibility of an unnoticeable bone marrow infiltration by the underlying breast carcinoma, complementary anti-CK stains using KL-1 and CAM 5.2 antibodies were performed. Surprisingly, the blasts showed a strong perinuclear or punctuate (dot-like) staining pattern (fig 2B). However, these cells did not react with the anti-CK19 antibody, further demonstrating the absence of bone marrow infiltration by the breast carcinoma.

    CK expression by myeloid blasts is a very uncommon finding—only three publications (two case reports and one in vitro study) dealing with this matter have been published so far.2–4 In this setting, the comparison between our findings and those described in these reports allows us to make some interesting comments. First, similar to CK detection reported in various lymphomatous or plasma cell disorders,4 the immunohistochemical pattern of CK expression in myeloid blasts is also dot-like or perinuclear. Another similarity between these studies and our case is the use of antibodies that recognise a wide spectrum of CKs (prekeratin, KL-1, AE1/AE3 cocktail, or CAM 5.2).2–4 Indeed, the KL-1 antibody reacts with the following CK polypeptides: CK1, CK2, CK5, CK6, CK7, CK8, CK11, CK14, CK16, CK17, and CK18. The antibody cocktail AE1/AE3 recognises numerous acidic and basic keratins, namely: CK10, CK14, CK15, CK16, CK19 and CK1, CK2, CK3, CK4, CK5, CK6, CK7, CK8, respectively. Finally, CK8 and CK18 are labelled by the CAM 5.2 antibody. In this regard, it is worth noting that these blasts are stained by AE1 but not by AE3 when these antibodies are used separately.2 Another point of interest is the similar differentiation that characterises CK expressing blasts. Indeed, the two reported cases of CK positive acute myeloid leukaemia (AML) reported belonged to the category of AML FAB M4.3,4 Although precise subtyping of AML arising from MDS may be difficult, the morphological, cytochemical, and phenotypic features seen in our patient are also consistent with a myelomonocytic differentiation. Like Turner and Milliken,4 we found no CK19 expression by the blasts. Therefore, this observation shows that CK19 is a more specific marker of carcinomas showing glandular differentiation than are various pan-CK immunostains.5 Whatever the precise explanation for this unusual phenomenon may be, in addition to the two cases published previously, our observation illustrates that CK expression can be seen, albeit rarely, in AML.

    Figure 1

    Bone marrow trephine biopsy. The picture shows a hypercellular bone marrow featuring dysplastic megakaryocytes, abnormal erythropoiesis, and increased numbers of blasts. These are often clustered.

    Figure 2

    (A) Immunostaining with an antibody directed against CD34 and (B) against CAM5.2. The stained cells look very similar. Note the dot-like staining pattern of CAM 5.2.


    The Vesalius Foundation supported this study (thanks to a grant from “La Loterie Nationale”).


    View Abstract

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.