Download High-Resolution Array-Based Comparative Genomic Hybridization

ORIGINAL ARTICLE
High-Resolution Array-Based Comparative Genomic
Hybridization for Distinguishing Paraffin-Embedded Spitz Nevi
and Melanomas
Jeff D. Harvell, MD,* Sabine Kohler MD,* Shirley Zhu, MD,* Tina Hernandez-Boussard, PhD,†
Jonathan R. Pollack, MD, PhD,* and Matt van de Rijn, MD, PhD*
Abstract: Distinguishing between Spitz nevus and melanoma
presents a challenging task for clinicians and pathologists. Most of
these lesions are submitted entirely in formalin for histologic analysis
by conventional hematoxylin and eosin–stained sections, and freshfrozen material for ancillary studies is rarely collected. Molecular
techniques, such as comparative genomic hybridization (CGH), can
detect chromosomal alterations in tumor DNA that differ between
these 2 lesions. This study investigated the ability of high-resolution
array-based CGH to serve as a diagnostic test in distinguishing Spitz
nevus and melanoma using DNA isolated from formalin-fixed and
paraffin-embedded samples. Two of 3 Spitz nevi exhibited no significant chromosomal alterations, while the third showed gain of the short
arm of chromosome 11p. The latter finding has previously been described as characteristic of a subset of Spitz nevi. The 2 melanomas
showed multiple copy number alterations characteristic of melanoma
such as 1q amplification and chromosome 9 deletion. This study has
shown the utility of array-based CGH as a potential molecular test in
distinguishing Spitz nevus from melanoma. The assay is capable of
using archival paraffin-embedded, formalin-fixed material; is technically easier to perform as compared with conventional CGH; is more
sensitive than conventional CGH in being able to detect focal alterations; and can detect copy number alterations even with relatively
small amounts of lesional tissue as is typical of many skin tumors.
Key Words: array-based comparative genomic hybridization, Spitz
nevus, melanoma, paraffin-embedded samples, archival samples
(Diagn Mol Pathol 2004;13:22–25)
T
he distinction between Spitz nevus and melanoma can be
difficult even for expert dermatopathologists.1 Because
these lesions are relatively small as compared with neoplasms
of other organ systems, the entire lesion is usually submitted
for formalin fixation and paraffin embedding, with no frozen
tissue remaining for molecular studies. A molecular technique
that could use formalin-fixed, paraffin-embedded tissue to
help identify melanoma would be a welcome addition to the
diagnostic armamentarium of pathologists. Comparative genomic hybridization (CGH) is a technique that can detect copy
number changes in tumor DNA and can identify differences
between Spitz nevi and melanomas.2 Conventional CGH entails the comparative hybridization of tumor and normal genomic DNA onto normal metaphase chromosomes and as such
has a limited cytogenetic-based mapping resolution of approximately 10-20 Mb.
Array-based CGH (aCGH) is a new method that weds
traditional CGH and microarray technology. In aCGH, tumor
and normal genomic DNAs are differentially fluorescently labeled and co-hybridized onto an array containing mapped
DNA sequences, providing measurements of tumor copy number changes at high resolution across the genome. Arrays containing up to 2,500 BAC clones have been used for aCGH.3
Gene arrays, which represent an alternative to individual BAC
clones, contain up to 30,000 cDNAs that span the human genome and correspond to individual genes or expressed sequence tags (ESTs). The resolution of gene microarrays thus is
significantly higher compared with conventional CGH and
even better than that seen with current BAC clone arrays. In
this study, we have investigated the utility of high-resolution
aCGH with gene microarrays as a potential diagnostic tool in
distinguishing Spitz nevus from melanoma, using paraffinembedded material.
MATERIALS AND METHODS
Case Selection
From the *Departments of Pathology and †Genetics, Stanford University
School of Medicine, Stanford, CA.
Reprints: Matt van de Rijn, MD, PhD, Department of Pathology Stanford University Medical Center 300 Pasteur Drive Stanford, CA 94305 (e-mail:
[email protected]).
Copyright © 2004 by Lippincott Williams & Wilkins
22
Permission to use patient material for these studies was
granted according to guidelines set and reviewed by the
Stanford University institutional review board for the study
of human subjects. Three cases of Spitz nevus and 2 cases
of invasive melanoma were culled from the Stanford University pathology files. Case 1 was originally seen at UCSF
Diagn Mol Pathol • Volume 13, Number 1, March 2004
Diagn Mol Pathol • Volume 13, Number 1, March 2004
Array-Based CGH for Melanocytic Neoplasms
Dermatopathology and previously reported (case 11).4 In each
case, the original diagnosis was based on conventional light
microscopy. The Spitz nevi and invasive melanomas were
classic examples, readily agreed on by 2 dermatopathologists
(JDH and SK) (Fig. 1). The clinical characteristics of the individual patients are summarized in Table 1.
Array Comparative Genomic Hybridization
Gene arrays were obtained from the Stanford Core Facility (http://www.microarray.org/sfgf/jsp/home.jsp) and contained 41,000 cDNA sequences spotted on Corning GAPS II
coated microarray slides (Corning Life Sciences, Acton, MA).
The chromosomal localization of these genes was uniquely assignable for 35,151 distinct cDNAs, which represented 24,540
different UniGene clusters and 3,225 cDNAs not yet represented in UniGene clusters. Tumor DNA from formalin-fixed,
paraffin-embedded tissue and reference DNA (normal gendermatched human leukocytes) were extracted using Qiagen genomic DNA columns. Gender-matched normal reference
DNA was isolated from peripheral blood leukocytes and digested with DpnII before further processing. Gel electrophoresis of digested and nondigested DNA isolated from formalinfixed, paraffin-embedded tissue was performed to confirm an
appropriate DNA fragment size distribution of approximately
0.5 to 2.0 Kb. Random primer labeling of DNA isolated from
tumor samples was performed as described previously 5
(http://www.microarrays.org/protocols.html). Briefly, 2 to 4
micrograms of tumor DNA was fluorescently labeled with
Cy5, mixed with reference DNA labeled with Cy3 and an excess of human cot-1, and hybridized overnight to the array.
After washing, the array slides were scanned on a GenePix
Scanner (Axon Instruments, Foster City, CA) and fluorescence
ratios (test/control) calculated using GenePix software. In our
experience with this study and other paraffin samples,6 we
found that very small samples, less than 5 mm in diameter,
could be used if several 10- to 15-µm sections were used to
isolate DNA. In our experience, amplification of the genomic
DNA was not necessary. For certain types of lesions where the
lesional cells are rare compared with surrounding stromal cells
laser, capture microdissection may be considered, but for our
FIGURE 1. Histopathology of (A) Spitz nevus with 11p amplification and (B) melanoma. A. Nested spindle shaped nevus
cells at the dermo-epidermal junction and epithelioid nevus
cells within the dermis characterize this Spitz nevus (case 3).
(B) Dermal nests of cytologically atypical melanocytes with
cytoplasmic melanin and obvious mitotic activity characterize
this invasive melanoma (case 4).
specimens this did not appear to be necessary and this probably
holds true for most melanocytic lesions.
Data were uploaded in the Stanford Microarray Database (http://genome-www5.stanford.edu/MicroArray/SMD/)7
for subsequent analysis. Only cDNA spots with a ratio of
signal over background of at least 1.5 in the Cy3 channel
TABLE 1. Array-Based CGH Changes of Spitz Nevi and Melanoma
Case
1*
2
3
4
5
Age/Sex
Histologic Diagnosis
9/F
30/M
33/F
77/M
74/M
Compound Spitz nevus
Intradermal Spitz nevus
Compound Spitz nevus
Invasive melanoma, superficial spreading type, Breslow 7.0 mm
Invasive melanoma, superficial spreading type, Breslow 4.2 mm
aCGH
Amplifications
None
None
11p
1q, 15q
6p, 17q
aCGH
Deletions
None
None
None
9, 10, 11, 15p
1p, 11, 15p
*Previously reported.4
© 2004 Lippincott Williams & Wilkins
23
Diagn Mol Pathol • Volume 13, Number 1, March 2004
Harvell et al
were included for analysis. Chromosomal localization of
the mapped genes was assigned by alignment against the
“Goldenpath” genome assembly data. Copy number changes
across chromosomes were represented using a smoothing
function based on a moving average of 5 adjacent genes8 and
were visualized using “CaryoScope” (http://genome-www5.
stanford.edu/cgi-bin/dev/rees/nph-aCGH-highres.pl), a software program written by Christian Rees.
RESULTS AND DISCUSSION
Despite nearly 60 years of study, the accurate diagnosis
of Spitz nevus remains a challenge and its exact biologic potential, a controversy. Nonetheless, recent studies comparing
both gene microarray expression data from freshly procured
mRNA and a CGH data using archival formalin-fixed breast
specimens have shown significant correlation between actual
gene expression and highly amplified DNA segments.5,6 Many
of these difficult lesions find their way to a specialist consultant. While there have been advances in immunohistochemistry and refined histologic criteria for distinguishing Spitz nevus from melanoma, there remains a subset of Spitz nevi that
cannot be reliably distinguished from melanoma even by expert dermatopathologists.1 Recently, studies of Spitz nevus
and melanoma by conventional CGH have uncovered molecular differences between the 2.2,9 Spitz nevi generally show no
genetic changes by conventional CGH, but a subset can show
an 11p gain.2,4 In contrast, melanomas show a number of cytogenetic abnormalities, including deletions of chromosomes
9, 10, 6q and 8p, and amplifications of chromosomes 7, 8q, 6p,
1q, 20, 17, and 2.9 Moreover, melanomas do not exhibit the
11p amplification seen in some Spitz nevi.2 Prior studies have
already shown the ability of aCGH to accurately distinguish
between benign and malignant renal and prostate tumors.3,10
These studies used up to 2460 cDNA BAC clones that spanned
the human genome, and in the second study, DNA isolated
from formalin-fixed, paraffin-embedded material was used.3
In the present study we used gene microarrays that contained
41,000 individual cDNAs spotted on a single microarray slide,
representing 24,540 different mapped UniGene clusters and
3,225 mapped cDNAs not yet represented in UniGene clusters.
We analyzed 3 Spitz nevi and 2 melanomas by aCGH.
The histologic features of 2 representative cases are shown in
Figure 1. Two of 3 Spitz nevi showed no significant chromosomal amplifications or deletions, whereas the third showed
a distinct gain of chromosome 11p (Table 1, Fig. 2). Case 1
had been previously studied by conventional CGH4 (case 11)
and likewise showed no detectable chromosomal alterations. Both melanomas displayed several of the chromosomal
amplifications/deletions characteristic of melanoma9 (Table 1, Fig. 2). The full data set of these experiments is
available from SMD (http://genome-www5.stanford.
edu/MicroArray/SMD/).7 In addition, histologic images from
all cases studied and PDF files from the aCGH experiments
24
FIGURE 2. Array-based DNA copy number profiles of (A) Spitz
nevus and (B) melanoma. Individual cDNAs are arranged according to their known location along human chromosomes
to produce a representation of tumor DNA copy number
changes mapped onto the normal human genome. Each vertical line (red or green) indicates the position of a single cDNA
clone, and the length of the line corresponds to the 1og10
fluorescence ratio of tumor versus reference DNA for that
locus, reported as a moving average of 5 adjacent clones. Red
lines indicate areas of amplification and green lines indicate
areas of deletion. (A) Case 3 shows a distinct gain of 11p, as
is characteristic of a subset of Spitz nevi. The low level red
and green signals on chromosomes 1, 9, and 10 are within
the variation seen in normal versus normal hybridizations
and therefore are not interpreted as evidence of copy number changes. (B) Case 4 shows alterations characteristic of
melanoma, including deletions of chromosome 9, 10, and
11 and amplification of 1q. Raw data from all experiments
can be downloaded from SMD, and figures of all aCGH experiments and histology in this paper can be seen at the accompanying web site (http://microarray-pubs.stanford.
edu/Spitz).
performed can be found at the accompanying web site
(http://microarray-pubs.stanford.edu/Spitz).
Until very recently, gene array studies of human cancers
have been partially limited in their scope due to dependence
upon mRNA transcripts, which require the acquisition of fresh
tissue. The need for a sizable piece of fresh tumor tissue precluded from study many cutaneous tumors, which are often
small, and as a result, cannot be apportioned for research. This
is especially the case with melanocytic lesions where the entire
specimen is submitted for histologic analysis to determine important prognostic features such as depth of invasion, vascular
invasion, regression, etc. The present study used punch biopsies as small as 5 mm in diameter yielding DNA amounts
greater than 4.0 µg, thus illustrating that meaningful results can
be obtained even with tiny amounts of lesional tissue.
Since molecular-based gene expression studies are still
considered new technology and since many of the published
© 2004 Lippincott Williams & Wilkins
Diagn Mol Pathol • Volume 13, Number 1, March 2004
genomic studies have relied on fresh tissue, these typically
have only 1 or 2 years clinical follow-up information, which is
inadequate to allow for meaningful correlation of gene expression changes with outcome. Case 1 was originally diagnosed
11 years ago illustrating that relatively antiquated paraffin
blocks can provide meaningful results that can be correlated
with long-term follow-up data. This is especially important for
borderline melanocytic lesions where prolonged follow-up
over many years may be required to ensure benign biologic
behavior (ie, absence of metastasis).
Gene expression profiling experiments use mRNA harvested from fresh tumor tissue, from which cDNAs are produced and used for hybridization. They have the benefit of detecting which genes are being actively expressed (ie, transcribed), but have the disadvantage of requiring fresh tumor
tissue. Genomic DNA is used in aCGH, but since the procured
DNA comprises both transcribed and nontranscribed genes,
the ability to determine which genes are actually being expressed within a tumor is less apparent. Nonetheless, recent
studies comparing both gene microarray expression data from
freshly procured mRNA and aCGH data using archival formalin-fixed breast specimens have shown significant correlation
between actual gene expression and highly amplified DNA
segments.5 As reflected in the present study, aCGH has the
further advantage of using DNA from formalin-fixed and paraffin-embedded biopsies. Furthermore, aCGH is technically
easier to perform than conventional CGH since the investigator does not have to be expert in cytogenetics to interpret results. Finally, the sensitivity of aCGH for the detection of localized alterations is superior to traditional CGH because of
the increased physical mapping resolution afforded by measuring changes across 41,000 individual genes or short genomic
segments. This increased sensitivity may uncover other genetic changes characteristic of melanoma, which are not detectable by conventional CGH or aCGH with fewer DNA targets. For instance, in this study, both melanomas displayed deletions of chromosome 11, which to our knowledge, has not
been previously emphasized.
In summary, we have demonstrated the possibility of using genome wide array-based CGH as a diagnostic aid in dif-
© 2004 Lippincott Williams & Wilkins
Array-Based CGH for Melanocytic Neoplasms
ferentiating Spitz nevus from melanoma. The procedure can be
applied to archival formalin-fixed and paraffin-embedded biopsies and can provide meaningful results even from small
punch biopsies. The present study is small and used classic
examples of Spitz nevus and melanoma and needs to be confirmed by a larger study that includes histologically questionable cases with long-term follow-up. We believe that aCGH
will prove to be a valuable adjunct in correctly classifying melanocytic tumors, in much the same way that PCR-based gene
rearrangement studies have aided the classification of cutaneous lymphoproliferative disorders.11
REFERENCES
1. Barnhill RL, Argenyi ZB, From L, et al. Atypical Spitz nevi/tumors: lack
of consensus for diagnosis, discrimination from melanoma, and prediction of outcome. Hum Pathol. 1999;30:513–520.
2. Bastian BC, Wesselmann U, Pinkel D, et al. Molecular cytogenetic analysis of Spitz nevi shows clear differences to melanoma. J Invest Dermatol.
1999;113:1065–1069.
3. Paris PL, Albertson DG, Alers JC, et al. High resolution analysis of paraffin-embedded and formalin-fixed prostate tumors using comparative
genomic hybridization to genomic microarrays. Am J Pathol. 2003;162:
763–770.
4. Harvell JD, Bastian BC, LeBoit PE. Persistent (recurrent) Spitz nevi: a
histopathologic, immunohistochemical, and molecular pathologic study
of 22 cases. Am J Surg Pathol. 2002;26:654–661.
5. Pollack JR, Sørlie T, Perou CM, et al. Microarray analysis reveals a major
direct role of DNA copy number alteration in the transcriptional program
of human breast tumors. Proc Natl Acad Sci U S A. 2003;99:12963–
12968.
6. Linn SC, West RB, Pollack JR et al. Gene expression patterns and gene
copy number charges in dermatofibrosarcoma protuberans. Am J Path.
2003;163:2383–2395.
7. Gollub J, Ball CA, Binkley G, et al. The Stanford Microarray Database:
data access and quality assessment tools. Nucleic Acids Res. 2003;31:
94–96.
8. Pollack JR, Perou CM, Alizadeh AA, et al. Genome-wide analysis of
DNA copy-number changes using cDNA microarrays. Nat Genet. 1999;
23:41–46.
9. Bastian BC, LeBoit PE, Hamm H, et al. Chromosomal gains and losses in
primary cutaneous melanomas detected by comparative genomic hybridization. Cancer Res. 1998;58:2170–2175.
10. Wilhelm M, Veltman JA, Olshen AB, et al. Array-based comparative genomic hybridization for the differential diagnosis of renal cell cancer.
Cancer Res. 2002;62:957–960.
11. Kohler S, Jones CD, Warnke RA, et al. PCR-heteroduplex analysis of
T-cell receptor gamma gene rearrangements in paraffin-embedded skin
biopsies. Am J Dermatopathol. 2000;22:321–327.
25