| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Center for Human Genetics,* the Laboratory for Morphology and Molecular Pathology, the Department of Pathology, and the Department of Pediatrics, University of Leuven, Leuven; and the Center for Human Genetics, Flanders Interuniversity Institute for Biotechnology (VIB), University of Leuven, Leuven, Belgium
 |
Abstract |
|---|
 Top Abstract Case ReportMaterials and MethodsResultsDiscussionReferences |
|---|
Anaplastic lymphoma kinase (ALK) is a member of the insulin receptor family of receptor tyrosine kinases and its expression is normally restricted to the nervous system.12 ALK gene fusion was first identified in anaplastic large-cell lymphoma (ALCL), in which most commonly a t(2;5) chromosomal translocation generates an NPM-ALK gene fusion and constitutively activated chimeric protein.13 Activation of ALK by variant ALK fusions have also been described in a subset of inflammatory myofibroblastic tumors (IMTs) occurring in children and young adults.14-17 These fusions are associated with striking ALK expression in the myofibroblastic spindle cells in a significant (up to 40%) subset of tumors.18 Genetic changes that underlie ALK-negative IPTs remain unclear. It is feasible that ALK-negative IPTs might represent a spectrum of entities with overlapping histological features but different underlying transforming mechanisms.15,18 Some investigators have proposed a role for viruses in the pathogenesis of IPTs, including EBV and human herpesvirus-8.6,19-21
This report describes a 10-year-old patient who developed a post-transplant EBV-related liver tumor with histologically combined smooth-muscle and IPT characteristics. Tumor cells revealed a spectrum of serial genomic rearrangements, involving the ALK gene and integrated EBV sequences, with the subsequent expansion of morphologically and molecularly distinct but cytogenetically related clones. Our findings indicate that the ALK loci may be a site of EBV virus integration. Virus integration could facilitate secondary genome rearrangements, leading consequently to the development of tumors with morphological features resembling IPTs and to related smooth-muscle tumors.
|
Case Report |
|---|
|
TopAbstractCase ReportMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractCase ReportMaterials and MethodsResultsDiscussionReferences |
|---|
After sampling for frozen material, tumor samples were fixed in 6% buffered formalin, embedded in paraffin, and stained with hematoxylin and eosin (H&E) for routine histological examination. Indirect immunohistochemistry was performed, either on paraffin or frozen sections, using monoclonal desmin (dilution 1:20; ICN-Cappel, Gent, Belgium), 
-smooth-muscle actin (-SMA, dilution 1:40; DAKO, Glostrup, Denmark), cytokeratin (dilution 1:50; Biodesign, Saco, ME), CD21 (dilution 1:10; DAKO, Merelbek, Belgium), CD35 (dilution 1:10; DAKO), myogenin (dilution 1:30, Novocastra, Gent, Belgium) and ALK-1 (dilution 1:10, DAKO) antibodies.
Electron Microscopy
Small fragments of the biopsy were immediately fixed in 2.5% glutaraldehyde, 0.1 mol/L phosphate buffer. After postfixation in 1% osmiumtetroxide 0.1 mol/L phosphate buffer, the samples were dehydrated in graded series of alcohol and embedded in epoxy resin. Ultrathin sections were cut, stained with uranyl acetate and lead citrate, and examined using a Zeiss EM10 electron microscope.
Cytogenetic and FISH Analysis
Chromosome metaphases were obtained after short-term culture of a primary tumor sample, using standard procedures. G-banded chromosomes were evaluated and classified according to the International System of Human Cytogenetic Nomenclature.22
FISH was performed on either metaphase chromosomes or interphase nuclei using probes homologous to 3' or 5' regions of ALK gene, as shown in Figure 1
. The ALK locus was evaluated with the LSI ALK Dual-Color Rearrangement Probe (Vysis, Inc., Downer’s Grove, IL), composed of a mixture of two differentially labeled DNAs that map either 3' telomeric (SpectrumOrange) or 5'centromeric (SpectrumGreen) to ALK. For subsequent study, the digoxigenin-labeled ALK P1 (PAC RP11–1111H1) (specific for 3'end of ALK catalytic domain) or PAC RP11–223F18 (specific for 5' end of ALK) clones were used.23
Additional FISH experiments used the EcoRI-B fragment of EBV DNA probe (kindly provided by Dr. John Arrand from the Paterson Institute for Cancer Research, Manchester, UK) and a probe specific for the subtelomeric sequences of chromosome 3q. FISH was performed according to the protocol previously described.24
The FISH data were obtained on a Leica DMRB (Leica, Wetzlar, Germany) fluorescence microscope equipped with a cooled black and white charged couple device camera (Photometrics, Tuscon, AZ), run by Quips SmartCapture FISH Imaging Software (Vysis, Bergisch-Gladbach, Germany).
|
ALK RT-PCR Analysis
Total RNA was extracted from the frozen tissue of the present case and from an IMT with ALK-ATIC fusion (positive control) using Trizol (Invitrogen, Carlsbad, CA), according to manufacturer’s protocol. The RNA was reverse-transcribed using MMLV reverse transcriptase. A positive control reverse transcription-polymerase chain reaction (RT-PCR) for the quality of the cDNA synthesis, using the primers 5'-CCCATTTCGGCTCCCATGATT and 5'-GGTCTCGGTGTGTGACTTGGA, was performed to amplify the CIZ gene. To analyze ALK transcripts, two primer sets were developed. The first, ALK3f (GCAACATCAGCCTGAAGACA) and ALK3r (GCCTGTTGAGAGACCAGGAG), amplifies a 192-bp fragment derived from the 3' end of the ALK transcript. The second set, ALK5f (CTCAGCGAGCTGTTCAGTTG) and ALK5r (GGAGAAGGCATGTTTGTTGG), amplifies a 200-bp fragment derived from the 5' end of the ALK transcript.
|
Results |
|---|
|
TopAbstractCase ReportMaterials and MethodsResultsDiscussionReferences |
|---|
Generally, the tumor was relatively well delineated from the surrounding parenchyma, although nodular extensions were occasionally seen (Figure 2b)
. A striking phenotypic variability was present in the lesion (Figure 2a)
. Part of the tumor consisted of a typical smooth-muscle tumor, showing fascicles of eosinophilic cells with cigar-like shaped nuclei and strong desmin and -SMA expression (Figure 2, c and e)
. Some mononuclear inflammatory cells were present between the smooth-muscle cells. Other parts of the tumor revealed an inflammatory, loosely textured background, in which -SMA-positive and desmin-negative myofibroblasts were present (Figure 2, d and f)
. The inflammatory component consisted mainly of lymphocytes and plasma cells (Figure 2d)
. This part of the tumor showed characteristics of an IPT. Both phenotypes were easily discriminated in most parts of the tumor, albeit that both components were intermingled in some areas (Figure 2)
. Quantitatively, the IPT component outnumbered the smooth-muscle component and the latter was mainly present at the edge of the lesion. Mitotic figures were occasionally present, mainly in the smooth-muscle component. Necrosis was focally seen. Immunostaining on frozen tumor sections using the ALK-1antibody showed no reactivity in the IPT component or in the smooth-muscle compartment. Frozen sections as well as paraffin sections were repeatedly negative on CD21 and CD35 staining. No expression of cytokeratin ormyogenin was found.
|
The tumor consisted of scattered spindle-shaped cells lying in an inconspicuous stroma with some collagen fibers and amorphous matrix components. Two types of spindle-shaped tumor cells could be distinguished. The most frequent type in our sections presented characteristics of myofibroblastic cells. The cells had a large nucleocytoplasmic ratio with an oval nucleus and prominent nucleolus.
The cytoplasm contained some rough endoplasmatic reticulum (RER) cisternae, numerous free ribosomes and polysomes, a small fat droplet and sometimes bundles of peripherally located actin microfilaments (Figure 3)
. The other cell type was more oblong and presented with a more voluminous cytoplasm almost completely filled with bundles of actin microfilaments and densifications. Plasmalemmal attachment plaques became obvious on the peripheral cell membrane as well as caveolae (Figure 3)
. These cells presented characteristics of smooth-muscle cells.
|
Cytogenetic Analysis
A total of 13 metaphase spreads were analyzed. Six metaphases revealed an abnormal karyotype: 46,XY,del(2)(p2;3),der(3)t(2,3)(p23;q29),der(21)t(Y;21)(q12;p13) (Figure 4)
. Seven cells confirmed a constitutionally normal karyotype.
|
Dual-color FISH analysis using LSI ALK assay showed two different patterns of hybridization in cells with abnormal karyotype. Sixteen out of 25 analyzed metaphases revealed only one fused (orange/green) signal on normal chromosome 2p23 (Figure 5A)
, while the remaining nine spreads, besides the normal (fused) signal on chromosome 2, disclosed the presence of the 3'end ALK signal (orange) on derivative chromosome 3 (Figure 5B)
. This result indicated the presence of two major clones, both showing a common loss of sequences 5'end to ALK (green). However, in the first clone the region 3'end to ALK (orange) was lost, leading to the loss of the entire ALK allele, while in the second clone it was translocated on 3q29. The existence of two clones was supported also by the results of interphase FISH, using the same dual-color LSI ALK probe. The touch preparation of the tumor specimen showed normal, adjacent (orange/green) signals (two fusions) in 15% of cells. The remaining 85% were abnormal, with 30% of nuclei showing only one fused (orange/green) signal, 39% of nuclei having one fused signal and one orange signal, and 16% of cells with variable number of fused and orange signals (Figure 5C)
.
|
and Table 2
.
|
|
Co-hybridization of digoxigenin-labeled ALKP1 together with biotin-labeled EBV DNA probes with the interphase nuclei of tumor cells revealed one red and one fused green/red (yellow) signal in 43% (Figure 5D)
, one red and one green signal in 35% and two red signals in 17% of cells. While the abnormal chromosome spreads were evaluated, a single red signal at normal chromosome 2p23 and overlapping red/green hybridization signal at the breakpoint of a der(3) was present (Figure 5E)
. This result indicated the integration of EBV DNA in the translocated 3'end ALK region in the molecularly distinctive clone. In contrast, the metaphases that failed to show translocated 3' end ALK did not show chromosomal integration of the virus DNA. In five of 14 of these metaphases, however, the overlapping green/red signal was found outside the counterstained chromosomal chromatin borders, pointing the presence of co-localized EBV and ALK sequences in an autonomous (non-integrated) episomal form (Figure 5F)
. Normal metaphases showed only red signals from ALK P1 probe on two normal homologous chromosomes 2p23 and no green signals from EBV probe.
Combined Immunofluorescence and FISH Study
The immunofluorescence and subsequent FISH study was performed to find out if the molecular heterogeneity was in any way associated with the phenotypic diversity of tumor cells. Using anti-desmin antibodies, positive cytoplasmatic immunoreactivity was found in 27% of cells. Immunopositive cells were randomly dispersed among negative ones. Twelve out of 14 desmin immunopositive cells that were evaluated demonstrated loss of one ALK allele using LSI ALK probe (Figure 5, G and H)
; two cells showed two fused, red and green, signals. Eleven out of 16 desmin immunonegative cells revealed the presence of one fused plus one single red signal. The remaining five cells showed the varying number of fused and/or single red signals.
ALK RT-PCR Analysis
To confirm ALK expression status, we performed RT-PCR analysis of total RNA prepared from the tumor specimen and from control IMT with ATIC-ALK chimeric fusion. As shown in Figure 6
, the tumor cells of the present case did not show ALK specific transcripts using primers for either the 5' or the 3'end of the ALK transcript. A control experiment demonstrated RT-PCR detection of cDNAs for an unrelated gene, CIZ, in the same sample. In contrast, when cDNA derived was used from an ATIC-ALK positive control case, a PCR product was obtained with the primers amplifying 3' ALK but not with 5' ALK primers. These findings confirm the lack of detectable ALK expression in the tumor under the study.
|
|
Discussion |
|---|
|
TopAbstractCase ReportMaterials and MethodsResultsDiscussionReferences |
|---|
-SMA-positive spindle-cell population, admixed with mononuclear inflammatory cells, including plasma cells and vessels, in a myxoid background. IPTs have been described in immunosuppressed patients and IPTs from the liver, spleen, and epidydimis have been associated with EBV.27-30
A major proportion of EBV-related liver and spleen IPTs correspond to EBV-associated follicular dendritic reticulum cell proliferations, but this is not the case in the tumor we describe, since specific follicular dendritic reticulum cell markers (CD21 and CD35) were negative. In addition, no ultrastructural features of follicular dendritic reticulum cells could be found. In the differential diagnosis of the IPT component, an inflammatory leiomyosarcoma could be considered. However, the latter entity generally does not show the loosely textured stroma as seen in our case and tend to be more cellular. In addition, one would expect desmin to be expressed in an inflammatory leiomyosarcoma. Moreover, these tumors often show foam cells, or even psammoma bodies, which are completely lacking. Finally, inflammatory leiomyosarcoma shows a near haploid karyotype, which is not the case in the tumor we present.31 Cytogenetic analysis demonstrated an unbalanced translocation involving bands 2p23 (ALK locus region) and 3q29. FISH analysis confirmed that the 2p23 rearrangement was within the ALK gene in a significant proportion of cells. Furthermore, among visually the same abnormal metaphase cells two molecularly distinctive clones were recognized: one with loss and the second with retained 3' ALK catalytic domain on derivative chromosome 3. The former resulted probably from the latter by a secondary rearrangement during aberrant cell expansion and tumor formation. Using an EBV-specific probe, FISH revealed integration of viral EBV DNA sequences at the 3q29 translocation breakpoint site and EBV co-localization with the translocated 3' end ALK kinase domain in the latter. The clone that revealed loss of translocated ALK contained viral sequences together with 3' end ALK in the nuclear episomal form. It is unknown whether the chromosomal integration of EBV sequences occurred before or after the translocation. However, the EBV integration site itself was shown to be a chromosomal region prone to enhanced fragility due to an altered chromatin structure.32 The loss of the integrated EBV viral genome due to chromosomal breaks proximal to the integration site was documented in BL cells and NPC cells in culture.33-34 This observation led to the hypothesis that under certain conditions EBV might act by a so-called "hit and run" mechanism during oncogenic transformation. Accordingly, due to increased chromosomal instability and a strong positive selection pressure toward cells retaining only episomal EBV, transient EBV integration into the host genome might be followed by loss of the virus together with the adjacent host cell chromosome fragments. Similarly, the loss of integrated EBV with the adjacent translocated ALK sequences in a subset of cells was observed in our tumor, suggesting the hit-and-run mechanisms of EBV-driven transformation that occurred in vivo.
In this context, the biological implications of EBV infection associated with tumor formation are of special interest. An integration of the viral particles into the human genome in EBV-associated disease has been well documented.34 The integration of EBV within the vicinity of the ALK receptor tyrosine kinase gene was never reported. As is known from other RNA and DNA tumor viruses, integration of viral DNA might modify the expression of viral as well as of cellular genes and might also cause chromosomal instability. The known mechanism of ALK activation in ALCLs and IMTs is formation of chimeric fusions that result in constitutive, ligand-independent receptor oligomerization. In the present case ALK kinase domain was also involved in the translocation. However, in contrast to the subset of IMTs with known variant chromosomal translocations involving the ALK gene, activation of ALK could not be demonstrated in myofibroblastic cells of the tumor, as shown by immunostaining and as confirmed by RT-PCR analysis. Thus, the IPT component of the present case belongs to the category of ALK-negative IPTs, although the possibility of transient ALK expression during the EBV integration near or within ALK gene or during the subsequent ALK secondary rearrangements cannot be excluded. Alternatively, the clonal expansion of cells harboring the ALK translocation could gain the required proliferative advantage through EBV association, and not by direct ALK activation. This hypothesis is supported by the recent finding that a small proportion of normal lymphoid cells in non-neoplastic lymph nodes are positive for chimeric ALK transcripts, suggesting that ALK gene rearrangement by itself might be insufficient to induce tumor formation.35 Furthermore, it is known that EBV-associated smooth-muscle tumors contain oncogenic latent membrane proteins that alone may implicate neoplastic cellular transformation.36 The transforming ability of LMP-1 is well recognized in lymphoid malignancies and LMP-1 might play a role in the pathogenesis of ALK-negative ALCLs.11,38
The very intriguing relation between distinct cellular genotypes, with respect to ALK changes and desmin-positive/-negative phenotypes that was observed within tumoral cells of the present case, requires further consideration. Both metaphase and interphase FISH confirmed the spectrum of serial genomic changes involving the ALK gene and integrated EBV sequences. Immunofluorescence anti-desmin staining combined with subsequent FISH analysis demonstrated the presence and the loss of translocated 3' end of ALK tyrosine kinase domain in desmin-negative and -positive cells, respectively. The desmin-negative cells most likely correspond to IPT-related myofibroblasts while the desmin-positive cells represent the smooth-muscle tumor component. Although myofibroblasts can also express desmin, on the histological sections the desmin-positive cells were only seen in the histologically typical smooth-muscle component of the tumor. The combined cytogenetic, immunofluorescence and morphological findings indicate that, although phenotypically different, both components of the tumor shared a common single-cell origin. We believe that this case represents a single immunosuppression-related EBV-driven lesion, which shows a phenotypic spectrum, ranging from a smooth-muscle tumor to an IPT. Thus, the observed phenotypic heterogeneity most likely reflects divergence during neoplastic progression, with the subsequent expansion of morphologically and molecularly distinct but cytogenetically related clones. The reason for this diversity remains obscure. One may hypothesize, however, that secondary genomic rearrangements induced by virus integration and subsequent relocation could impose somatic changes of cellular genes encoding functions responsible for tumor’s gene expression profile.
Notably, the EBV-ALK interaction in EBV-associated smooth-muscle tumors or IPTs has never been investigated. Whether these tumors might be induced by direct ALK-EBV fusion warrants further studies.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
This text presents research results of the Belgian program on Interuniversity Poles of Attraction initiated by the Belgian Sate, Prime Minister’s Office, Science Policy Programming. Its authors assume the scientific responsibility. The authors would like to acknowledge the COST support through the COST ACTION B19 "Molecular cytogenetics of solid tumors" in carrying out this work.
Accepted for publication May 14, 2003.
|
References |
|---|
|
TopAbstractCase ReportMaterials and MethodsResultsDiscussionReferences |
|---|
ß. Oncogene 1999, 18:6959-6964[Medline]
This article has been cited by other articles:
 |
|
 |
|
  W. Y. Chay, A. Penafiel, J. Raghuram, K. L. Chuah, C. M. Loo, and P. Lee Dyspnea in a Transplant Recipient With Pulmonary Nodules Chest, March 1, 2009; 135(3): 860 - 865. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
