Abstract P2-08-12: Development and validation of prognostic gene signatures for basal-like breast cancer and high grade serous ovarian cancer

Background: Basal-like breast cancer (BLBC) have poor prognosis. Molecular similarities have been reported between BLBC and high grade serous ovarian cancer (HGSOC). To date, there have been no prognostic biomarkers specifically developed for BLBC or HGSOC that can provide risk stratification and inform treatment selections. In this study, we utilized RNA-seq data available from The Cancer Genome Atlas (TCGA) project to develop molecular signatures for risk stratification in BLBC, and further validated these signatures in HGSOC RNA-seq data from TCGA.

Methods: Raw count of RNA-seq data were downloaded from TCGA for 190 BLBC and 374 HGSOC patients. The datasets were annotated with 56963 Ensembl gene IDs. Excluding 31375 gene IDs with no greater than 10 counts in at least 90% of the samples, totally 25228 unique Ensembl gene IDs were used. Progression-free interval (PFI) is the primary study endpoint. Analyses of differentially expressed genes were performed using 3 bioconductor packages: DESEq2, edgeR and voom/limma. Signatures based on commonly identified genes among the 3 analytic methods were established using weighted linear combination of gene expression levels. Their performance was evaluated in the BLBC and HGSOC datasets using Kaplan-Meier survival analysis with log-rank tests and Cox proportional hazard regressions.

Results: Among 190 TNBC patients, 18 had recurrences within 2 years and 40 showed no recurrences for at least 5 years. These patients were used as recurrent vs. non-recurrent cases for differential expression analysis. 307 and 343 genes were differentially expressed based on adjusted p value threshold 0.05 and 0.01 in DESeq2 and edgeR analysis, respectively. voom/limma identified no genes differentially expressed based on adjusted p values, but 228 genes had unadjusted p values < 0.01 and were used in the following analysis. Taken together, 63, 58 and 21 genes were commonly identified by DESeq2/edgeR, DESeq2/limma and edgeR/limma analysis, respectively. All 3 signatures were able to significantly stratify the TNBC full dataset (n=190) using either 20-, 50- or 80-percentile as the cut-points. When evaluated in HGSOC patients using 80-percentile cut-point, both 63- and 58-gene signatures were able to significantly stratify patients into different risk groups (HR 2.16, 95% CI: 1.4-3.34, p < 0.001; HR 2.06, 95% CI: 1.36-3.11, p < 0.001). Multivariate Cox regression adjusting for age, grade and stage showed 63- and 58-gene signatures remained to be statistically significant in stratifying HGSOC patients (p = 0.0005 and 0.001, respectively).

Conclusion: Gene signatures were specifically identified to prognosticate BLBC patients based on RNA-seq data from TCGA project. Which were able to classify HGSOC patients into differential risk groups. With further validations, these signatures may provide additional prognostic tools for clinicians to better manage triple-negative breast cancer that mostly overlap with BLBC, and HGSOC patients who are difficult-to-treat currently.

Disclaimer The views expressed in this article are those of the authors and do not reflect the official policy of the department of Army/Navy/Air Force, Department of Defense, or U.S. government.

Citation Format: Hu H, Zhang Y, Liu J, Raj-Kumar P-K, Yang H, Lee M, Kovatich AJ, Shriver CD. Development and validation of prognostic gene signatures for basal-like breast cancer and high grade serous ovarian cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-08-12.

Abstract P4-09-14: Analysis of breast cancer recurrence using gene set enrichment analysis

Background: Even after successful treatment of primary breast tumors, there is a continued risk of recurrence. The risk varies between subtypes and there are ongoing efforts that aim to improve prediction of such risks for individual patients. Detection of subclinical metastases might be achieved by biomarkers in blood. In this study, we profiled protein expression in blood plasma from patients with known clinical outcome (recurrence vs no recurrence) to identify prognostic markers of breast cancer recurrence.

Methods: The subjects and specimens were made available through the Clinical Breast Care Project using IRB-approved protocols. We analyzed blood plasma samples taken at the time of diagnosis from consented patients who subsequently relapsed (33 cases) as well as those with no disease recurrence (31 controls). Based on hormone receptor and lymph node status the samples were grouped as: ER-/HER2- (17 cases/15 controls), ER+/LN+ (10/10) and ER+/LN- (6/6). We used aptamer-based SOMAscan assay platform to study the expression of 1252 proteins. We analyzed the protein expression data by using their coding genes in order to apply the Gene Set Enrichment Analysis method (GSEA v.2, Broad Institute). Pathway databases of KEGG, REACTOME, BIOCARTA and C4 collection were used. Significant gene sets were called at 5% FDR, and overlaps and low coverage gene sets (Tags <70%) were removed. Statistical analysis and clustering were done using R.

Results: Unsupervised clustering showed some difference in signal in the ER+/LN- group. Even though there was a lack of significantly differentiated proteins between the cases and controls of this group, many significant gene sets were identified. After applying the cutoff filters and removing the overlaps, there were 5 gene sets enriched with the pathway collection, involved in B-cell receptor signaling, mRNA metabolism, tight junction and SCF-KIT signaling. Similarly, 9 gene sets from the MORF compendium were differentially expressed with the C4 collection and included neighborhood genes of NME2, ACTG1, EIF3S2, AP2M1, DAP3, UBE2I, NPM1, AATF and NPM1. In contrast, neither differentially expressed proteins nor gene sets were identified from the ER+/LN+ and ER-/HER2- groups. Since the sample size of the ER+/LN- group was small, we conducted a similar analysis by randomly choosing 6 case and control samples in the other two groups respectively. There were still no differentially expressed proteins or gene sets identified above the specified cutoff parameters.

Conclusion: Using plasma protein expression data we identified underlying gene sets differentially expressed between ER+/LN- patients who had cancer recurrence and no recurrence. Many genes in these sets were already known biomarkers (e.g. PTEN, AKT1, STAT3, SET etc.). These results can be used for understanding patterns of recurrence in different cancer subtypes. Further research is needed to estimate the clinical significance of these gene products.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Army/Navy/Air Force, the Department of Defense, or U.S. Government.

Citation Format: Praveen Kumar A, Kovatich AJ, Biancotto A, Cheung F, Davidson-Moncada JK, Kvecher L, Liu J, Ru Y, Kovatich AW, Deyarmin B, Fantacone-Campbell JL, Hooke JA, Raj Kumar PK, Rui H, Hu H, Shriver CD. Analysis of breast cancer recurrence using gene set enrichment analysis [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P4-09-14.

Abstract P2-06-04: Use of principal component analyses to select ER-balanced subset for gene centering in PAM50 subtyping

Background: PAM (Prediction Analysis of Microarray) 50 is an established gene expression-based algorithm to classify breast tumors into basal-like, HER2-enriched, luminal A (LA), and luminal B (LB) subtypes. Clinical subtyping is mainly based on immunohistochemistry (IHC) assays of estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (Her2) and Ki67 classifying tumors into triple-negative (ER-/PR-/Her2-), Her2+ (ER-/PR-/Her2+), LA (ER+/Her2-/Ki67-), LB1 (ER+/Her2-/Ki67+) and LB2 (ER+/Her2+). These two subtyping methods do not completely match even on comparable subtypes. Nevertheless, the ER-balanced subset for gene-centering in PAM50 subtyping was selected based on clinical status. Here we explored the possibility of using principal component analyses and iterative PAM50 call to refine the selection of an ER balance subset to improve consistency between these methods focusing on LB calls which is more aggressive than LA tumors.

Methods: Normalized gene expression data was obtained from TCGA research network for 712 primary tumors which had IHC status available for ER, PR and Her2. Since Ki67 status was not available LA and LB was discriminated for ER+ cases with Her2- and Her2+ respectively. In house RNA-Seq dataset had 118 primary tumors and were drawn from the Clinical Breast Care Project where breast cancer patients were consented using an IRB-approved protocol. Tumors were selected and processed by laser microdissection. RNA was extracted from tissues using the Illustra triplePrep kit (GE Healthcare). Paired-end mRNA sequencing was performed using the Illumina HiSeq platform. Sequenced reads were processed using PERL based pipeline utilizing PRINSEQ, GSNAP and HTSeq. Principal component analysis (PCA) was done using R. Wilcoxon rank sum test was used for statistical significance (p<0.05).

Results: In both datasets, the PCA map grouping of cases does not perfectly reflect the clinical subtypes. This motivated us to select ER balance subset based on the PC1 separation and IHC subtype. The resulting PAM50 subtypes on PCA map distinguished Basal and LA as two well separated components. Using all of Basal and equal number of LA cases for ER balance subset for PAM50 resulted in increased LB call and a better consistency with IHC LB calls. Among 712 cases in TCGA LB numbers increased from 142 in initial PAM50 call to 203 in ER balanced refined PAM50 call. We noticed that there was significantly higher (p-value = 4.414e-11) MKI67 expression for the 39 cases switch from LA to LB between PAM50 calls. Similar trend was observed in our in-house dataset where majority of the IHC-LB1 cases was called as LB in PAM50. The new method increased LB call from 22 to 27 which in-turn increased consistency between molecular and clinical subtypes from 73 to 79 out of the total of 118 cases.

Conclusion: We show that an iterative PAM50 call coupled with PCA for selection of ER balance set potentially enhanced the consistency of the LB calls with clinical subtyping and that the tumors switched from LA to LB have high MKI67 expression.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Army/Navy/Air Force, the Department of Defense, or U.S. Government.

Citation Format: Raj-Kumar P-K, Liu J, Kovatich AJ, Kvecher L, Shriver CD, Hu H. Use of principal component analyses to select ER-balanced subset for gene centering in PAM50 subtyping [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P2-06-04.

Abstract P3-16-01: Using the new pan-cancer clinical data resource (TCGA-CDR) to identify breast cancer genomic correlates associating with different survival outcome endpoints

Introduction The Cancer Genome Atlas (TCGA) generated abundant high quality molecular data, however its relatively short-term patient follow-up limited its immediate clinical utility. We led a PanCanAtlas effort to systematically collate, integrate, and quality check the large body of acquired clinicopathologic data, generated 4 primary clinical outcome endpoints for each case, and created a new Pan-Cancer Clinical Data Resource (TCGA-CDR) for public use. We report here on the utility and validity of this TCGA-CDR in relating breast cancer (BC) genomic information to survival endpoints.

Methods Clinicopathologic data from all data files were integrated and processed. Overall survival (OS), disease-specific survival (DSS, an approximation), progression-free interval (PFI), and disease-free interval (DFI)were derived.Tests of the adequacy of the follow-up intervals for each endpoint were performed, and quality evaluation of these endpoints was established by their comparison with different clinical features. As a case study we compared each survival endpoint for significant association (FDR <0.2) with chromosomal aneuploidy.

Results The 4 endpoints were derived for 1097 TCGA BC cases having a median follow-up time of 27.7 months. Median times to events/censorship for OS, DSS, PFI, and DFI were 41.8/25.0, 32.6/26.0, 26.0/25.0, and 25.4/25.0 months respectively. PFI and DFI passed tests for adequate follow-up times; OS and DSS partially passed the same tests signaling some caution with their use in genomic associations.

Using the endpoints, outcomes of patients with ER+ and ER- tumors were compared, along with those of patients with low (I&II) and high (III&IV) stage breast tumors. Univariate analyses suggested patients with ER+ tumors had significantly better survival than patients with ER- tumors when using PFI (p=0.005), DFI (p=0.001), and DSS (p=0.009), with OS not reaching significance (p=0.09). Patients with low stage tumors showed significantly better outcomes than patients with high stage tumors for each endpoint (p<0.001). The 4 endpoints were also evaluated for their significant associations with chromosomal arm aneuploidy. Adjusted for patient age and AJCC stage, tumors with a loss of 8q and 8p (p=0.019, FDR=0.37) had worse PFI; and those with loss of 8q, 20q, and 8p had worse DFI. Tumors with gain of 11q or loss of 14, 7q, 12q, 18q, 20q, 3p, 7p, 8p, 18p, and 20p had worse OS. In contrast, tumors with loss of 16q had better DSS, while those with loss of 3q, 12q, 17q, 18q, 19q, 20q, 3p, 8p, 12p, 18p, 19p, and 20p had worse DSS. The finding that 8p loss associated with worse survival for all 4 endpoints, while 18p loss associated with worse OS and DSS, agrees with literature reports.

Conclusion These findings confirm that PFI and DFI, as extracted from the TCGA-CDR, are valid and appropriate BC survival endpoints, while OS and DSS may be recommended with some caution when employing TCGA data to evaluate new relationships between breast cancer genomic abnormalities and clinical outcomes.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Army/Navy/Air Force, the Department of Defense, or U.S. Government.

Citation Format: Liu J, Lichtenberg T, Hoadley KA, Cherniack A, Poisson L, Kovatich AJ, Benz C, Thorsson V, TCGA PanCanAtlas Research Network, Shriver CD, Hu H. Using the new pan-cancer clinical data resource (TCGA-CDR) to identify breast cancer genomic correlates associating with different survival outcome endpoints [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P3-16-01.

Abstract P4-09-14: PhosphohistoneH3 as a prognostic marker in breast cancer: High expression is associated with younger age, triple negative subtype, and disease specific survival

BACKGROUND PhosphohistoneH3 (PPH3) is an emerging marker in breast cancer and has been linked to both patient survival and age. Phosphorylation of HistoneH3 is an important step during the cell cycle leading to proper compaction of the chromatin during late G2 and early mitosis. Here we assessed the use of PPH3 as a prognostic marker within a group of invasive breast cancers in the Clinical Breast Care Project (CBCP).

METHODS CBCP participants and their samples were collected following IRB-approved, HIPAA-compliant protocols. Samples from 157 CBCP patients were selected for tissue whole section immunohistochemistry (IHC), using antibodies to PPH3, ER, PR, Ki67, and Her2. For each sample, staining of PPH3 was assessed across 5 high powered microscope fields and was considered positive if there was on average >2 stained cells per field. ER and PR were considered positive when there was >5% nuclear staining, and Ki67 was positive when there was >15% nuclear staining. Her2 was considered positive with an IHC score of 3+ or 2+ with a FISH score above 2.2. The samples were subtyped as Luminal A (LA: ER+/HER2-/Ki67-), Luminal B1 (LB1: ER+/HER2-/Ki67+), Luminal B2 (LB2: ER+/HER2+), Her2+ (ER-/PR-/HER2+), and Triple Negative (TN: ER-/PR-/HER2-). PPH3 was tested for associations with age and subtype using a stratified univariate Wilcoxon rank-sum analysis and a multivariate analysis controlling for subtype. To test the efficacy of PPH3 as a prognostic marker, Kaplan-Meier curves for disease specific survival were analyzed and the cox proportional hazard regression model was calculated. Further analysis addressing population demographics and additional cancer characteristics is ongoing.

RESULTS Wilcoxon analysis revealed an association between higher PPH3 levels and younger age (P=.0038). Subtype was also found to be associated with PPH3, with the TN subtype 6.26 times more likely to have higher PPH3 expression than LA (P=.005). The association with age was confirmed by repeating the analysis and stratifying into non-TN subtypes (P=.05) and TN only subtype (P=.017). Non-TN subtypes positive for PPH3 expression had median age of 53.18 at diagnosis and 63.29 for negative PPH3 expression; TN subtypes that were positive for PPH3 had a median age of 50.44 and 72.9 for negative PPH3. Multivariate analysis with age and subtype as the variables also supported these results (age P=.017; TN vs LA P=.022). Disease specific survival analysis showed that a shorter survival time was associated with positive PPH3 protein levels (P=0.03; hazard ratio=6.97).

CONCLUSIONS High expression of PPH3 is associated with a younger age, poorer survival rate, and the TN subtype. These results corroborate the use of PPH3 as a prognostic marker for breast cancer patients.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Defense, or U.S. Government.

Citation Format: Craig J, Kovatich AJ, Hooke JA, Kvecher L, Liu J, Fantacone-Campbell JL, Rui H, Shriver CD, Hu H. PhosphohistoneH3 as a prognostic marker in breast cancer: High expression is associated with younger age, triple negative subtype, and disease specific survival. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P4-09-14.

Steroid induction of therapy-resistant cytokeratin-5-positive cells in estrogen receptor-positive breast cancer through a BCL6-dependent mechanism

Therapy resistance remains a major problem in estrogen receptor-α (ERα)-positive breast cancer. A subgroup of ERα-positive breast cancer is characterized by mosaic presence of a minor population of ERα-negative cancer cells expressing the basal cytokeratin-5 (CK5). These CK5-positive cells are therapy resistant and have increased tumor-initiating potential. Although a series of reports document induction of the CK5-positive cells by progestins, it is unknown if other 3-ketosteroids share this ability. We now report that glucocorticoids and mineralocorticoids effectively expand the CK5-positive cell population. CK5-positive cells induced by 3-ketosteroids lacked ERα and progesterone receptors, expressed stem cell marker, CD44, and displayed increased clonogenicity in soft agar and broad drug-resistance in vitro and in vivo. Upregulation of CK5-positive cells by 3-ketosteroids required induction of the transcriptional repressor BCL6 based on suppression of BCL6 by two independent BCL6 small hairpin RNAs or by prolactin. Prolactin also suppressed 3-ketosteroid induction of CK5+ cells in T47D xenografts in vivo. Survival analysis with recursive partitioning in node-negative ERα-positive breast cancer using quantitative CK5 and BCL6 mRNA or protein expression data identified patients at high or low risk for tumor recurrence in two independent patient cohorts. The data provide a mechanism by which common pathophysiological or pharmacologic elevations in glucocorticoids or other 3-ketosteroids may adversely affect patients with mixed ERα+/CK5+ breast cancer. The observations further suggest a cooperative diagnostic utility of CK5 and BCL6 expression levels and justify exploring efficacy of inhibitors of BCL6 and 3-ketosteroid receptors for a subset of ERα-positive breast cancers.

Abstract P4-06-03: Assays on core biopsies and surgically resected tumors may result in different subtyping of the invasive breast cancer from the same patient

Background Core biopsies (CBs) are often used for biomarker expression assays to determine the treatment regimen. However, a number of other clinically important analyses (e.g. OncoType Dx), are performed on surgically resected tumors (SRTs). A previous study has shown that biomarkers ER, PR, and Ki67 expressed higher in CBs than in SRTs. Here we analyze how this difference impacts the subtyping of ER+ breast tumors.

Methods Female patients enrolled in the Clinical Breast Care Project (CBCP) from a civilian site were selected for this study, where expression of ER, PR, HER2, and Ki67 were assayed by IHC in a reference lab on CBs; the same 4 assays were performed on SRTs by a CBCP central lab. Both labs are CLIA-certified. Patients treated with neoadjuvant chemotherapy and those with multiple tumors were excluded. 167 cases were identified for this study to compare assays performed on CBs and SRTs from the same patients. ER and PR were positive if >1% nuclear staining, HER2 was negative if IHC = 0 or 1+, positive if IHC = 3+, and for IHC = 2+ FISH was used for the final call. Ki67 was positive if > = 15% nuclear staining. LA was ER+/HER2-/Ki67-, LB1 was ER+/HER2-/Ki67+, and LB2 was ER+/HER2+. For histologic grades, only readings from the central lab on SRTs were used. Statistical analyses were performed using SAS.

Results This analysis confirmed that Ki67, ER, and PR showed higher percent nuclear staining in CBs than in SRTs from the same patients. The difference for Ki67 was more striking and unidirectional. ER and PR cases clustered at the upper percent levels. Histograms with a bin-width of 15% show a peak at 15% for Ki67 difference between CBs and SRTs, whereas the peaks for ER and PR differences were at 0%. McNemar's (or Exact McNemar’s) test showed significant differences between the binary status calls for Ki67 (p = 3.2E-15) and ER (p = 0.012), but not for PR (p = 0.65). Assays on CBs and SRTs resulted in different subtype calls for the cases (Table 1). Grade distributions were different between LA and LB (p<0.001 for both CB- and SRT-based subtypes, Chi-Square or Fisher's Exact test), but not so between LB1 and LB2 (p = 0.23 for CB, 0.31 for SRT). However, SRT-based LB1 cases concentrate more on higher grades compared to CB-based cases (p = 0.048).

Table 1. ER+ subtypes based on IHC assays (from CBs and SRTs) and corresponding grades (from SRTs) CBSRTSubtypeG1G2G3G1G2G3LA2126034518LB11435342820LB2036032

Discussion On IHC assays, Ki67 expression is strikingly higher in CBs than in SRTs, and ER expression is also higher in CBs than in SRTs. This directly resulted in more LB than LA subtypes based on CBs. SRT-based LB1 cases concentrate more on higher grades compared to CB-based cases, which is more consistent with the observation that LB subtypes have worse outcomes. A limitation of this study is that technical differences between the labs may contribute to the observed differences between CBs and SRTs. Further studies need to be performed to determine whether SRT should also be assayed in addition to CB for treatment regimen decision-making.

The views expressed in this abstract are those of the authors and do not reflect the official policy of the Department of Defense, or US Government.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-06-03.

Abstract P1-08-20: Increased risk of hormone therapy failure in breast cancers expressing low phospho-Stat5: Validation of quantitative immunofluorescence assay parameters

Previous analyses of three breast cancer cohorts revealed that loss of phospho-Stat5 in breast cancer is associated with significantly elevated risk of hormone therapy failure (1, 2). Nuclear localized tyrosine phosphorylated Stat5 (Nuc-pYStat5) may therefore have clinical value as a predictive marker. Analysis of two of the three previously reported anti-estrogen treated patient cohorts used pathologist scoring of diaminobenzidine (DAB) chromogen-stained Stat5. However the third cohort, analyzed by quantitative immunofluorescence analysis (QIF) on the Genoptix/HistoRx AQUA platform, revealed a greater hazard ratio than the cohorts analyzed by pathologist DAB-scoring. To extend and validate these observations, we applied the Nuc-pYStat5 cutpoint derived in our previous study (2) to an independent cohort of anti-estrogen-treated breast cancer patients using two distinct QIF software platforms, AQUA and Definiens Tissue Studio. Tissue Studio relies on supervised machine learning and multiparametric features of a high-resolution whole slide image to identify cancer cell regions, while AQUA software relies on costaining of a tumor marker to identify cancer cell regions. The two QIF platforms produced highly concordant Nuc-pYStat5 levels (R2 linear = 0.96, P<0.001, N = 344) and confirmed a significant elevated risk of failing antiestrogen therapy in patients whose tumors had lost Nuc-pYStat5 (Hazard ratio 3.6; 95% CI 1.8-7.4; P<0.02; N = 98). On both QIF platforms, Nuc-pYStat5 remained an independent marker after multivariate adjustment for standard pathology parameters, including ER/PR, HER2, age, node status and grade, with a hazard ratio of 5.8 (95% CI 1.3-22.2; P = 0.02; N = 52). High concordance between Nuc-pYStat5 levels produced by the two QIF platforms held up in a second independent dataset of more than 300 breast cancer specimens (R2 linear = 0.97, P<0.001, N = 382). Nuc-pYStat5 levels by the two QIF methods remained highly concordant across the entire dynamic range in both patient cohorts. Furthermore, high concordance was also observed between replicate QIF analyses of Nuc-pYStat5 on serial tumor microarray sections stained in the same run on an automated immunostainer (Concordance Correlation Coefficient (CCC) = 0.96; 95% CI 0.96-0.97). Modest inter-assay staining variation (CCC = 0.84; 95% CI 0.82-0.87) for Nuc-pYStat5 when serial tumor microarrays were stained on different runs several days apart could be corrected for by normalization procedures (CCC = 0.94; 95% CI 0.92-0.95). This progress supports the utility of QIF analysis of Nuc-pYStat5 levels in human breast cancer and further documents the potential value of Nuc-pYStat5 as a predictive marker of response to antiestrogen therapy. The study confirms that further retrospective and prospective validation studies are warranted.

References:

1) Yamashita et al. Stat5 expression predicts response to endocrine therapy and improves survival in estrogen receptor-positive breast cancer. Endocr Relat Cancer. 2006;13:885-93.

2) Peck et al. Loss of nuclear localized and tyrosine phosphorylated Stat5 in breast cancer predicts poor clinical outcome and increased risk of antiestrogen therapy failure. J Clin Oncol. 2011;29:2448-58.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-08-20.

Abstract P4-06-09: HER2+ and HER2- luminal B subtypes have similar overall survival and histologic grade distributions

Background There are multiple subtypes in invasive breast cancers (IBCs). Immunohistochemistry (IHC)-based assays using ER, PR, HER2, and Ki67 for subtyping has been developed. However, association between such subtypes and treatment outcomes and histology is not completely known, and are impacted by dataset-to-dataset and pathologist-to-pathologist variations. We report an analysis on these problems, as a pilot study of a project involving 5,000 patients and 250 protein biomarkers.

Methods Patients were enrolled for the Clinical Breast Care Project from a military site with data collected per IRB-approved protocols, from 2000 to 2010. Total of 215 female IBC cases were included in this study, with surgically resected tumors (SRT) assayed for ER, PR, HER2, and Ki67 by IHC in a central CLIA-certified lab following clinical guidelines where applicable. All slides were reviewed by a single experienced breast pathologist. ER and PR was positive if nuclear staining was >5%. HER2 was negative if IHC = 0 or 1+ and positive if IHC = 3+; For IHC = 2+, the FISH result determined the final call. Ki67 was positive if nuclear staining was > = 15%. For IBC subtypes, LA was ER+/HER2-/Ki67-; Two LB subtypes were defined, with LB1 being ER+/HER2-/Ki67+ and LB2 being ER+/HER2+; Her2+ was ER-/PR-/HER2+; TN was ER-/PR-/HER2-. Statistical analyses were performed using SAS, Kaplan-Meier estimate and log-rank test were used for survival analysis and the follow-up period was 10 years with a median of 4.6 years. Chi-Square test was used for categorical data analysis supplemented by Fisher's Exact test as appropriate.

Results 204 of the 215 cases were classified into subtypes of LA (n = 74, 7 deceased), LB1 (n = 53, 4 deceased), LB2 (n = 14, 1 deceased), Her2+ (n = 14, 1 deceased), and TN (n = 49, 16 deceased). Despite the low number of events in some subtypes, there was a significant difference in overall survival between the 5 subtypes of IBCs defined here (p = 0.0023), with TN cases showing the least favorable outcome. No difference was observed in outcome between LB1 and LB2 (p = 0.86). Overall, Ki67+ cases trended toward worse outcomes (p = 0.08), which was also observed in TN (p = 0.17) but not other subtypes. Histologic grades were significantly different among the 5 subtypes (p = 6.25E-20); 96% of LA cases were G1 or G2, over 80% of LB1 and LB2 cases were G2 or G3, and all Her2+ and 93% of TN cases were G2 or G3. Within the luminal subtypes, grade distribution for LA cases was significantly different from that for LB cases (p<0.0001) but there was no difference between LB1 and LB2 cases (p = 0.95).

Discussion In this cohort where all IHC and pathology slides were reviewed by a single pathologist, we used cell proliferation marker Ki67 to help classify luminal IBCs into LA, LB1 (HER2-), and LB2 (HER2+). Overall survival analysis result for all cases was consistent with the literature, Ki67+ cases trended toward worse outcomes, and no outcome difference was identified between LB1 and LB2. Histologic grade distributions in different subtypes were consistent with the literature; we further found no difference between LB1 and LB2 subtypes.

The views expressed in this abstract are those of the authors and do not reflect the official policy of the Department of Defense, or US Government.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-06-09.

Abstract P2-05-21: Molecular subtypes of invasive breast cancers show differential expression of the proliferation marker Aurora Kinase A (AURKA)

Background: Invasive breast cancer (IBC) has been classified into four major subtypes based on gene expression profiling. The luminal A subtype (LA) has the best prognosis, when compared to luminal B (LB), HER2+, and basal-like (Basal). Ki67 by gene expression or immunohistochemistry (IHC) is commonly used as a proliferation index. The function of Ki67 in proliferation remains unknown. AURKA (STK15) is known to play an important role in mitosis, and is a component of the 21-gene recurrence score of the Oncotype Dx. With multiple platforms of molecular data available from hundreds of IBC tissues in The Cancer Genome Atlas project (TCGA), we sought to study the association of AURKA with different IBC subtypes and explore its use as a proliferation marker in IBCs.

Methods: Gene expression (Agilent, log2 transformed), relative DNA copy number (CN, Affymetrix SNP 6.0), and exome sequence mutation (Illumina) data for 459 IBC cases were downloaded from the TCGA data portal. PAM50 classification results of all samples were obtained from the TCGA breast cancer AWG group and included 203 LA, 113 LB, 51 HER2+, 84 Basal-like, and 8 Normal-like which were not used in this study due to the low numbers. Kruskal-Wallis tests were used to evaluate the differences among four subtypes on AURKA expression and CN, followed by Wilcoxon Mann-Whitney test with Bonferroni adjustment for pairwise analyses. Pearson's Correlation Coefficient was used for correlation analyses. All statistical analyses were performed using SAS and R, and two-sided, p values <.05 were considered statistically significant.

Results: There was a significant difference among IBC subtypes, in gene expression as well as in CN (p values < 0.0001). AURKA mRNA levels were significantly lower in LA (mean±SD, −2.61±0.63) compared to LB (−1.45±0.78), HER2+ (−1.38±0.61), and Basal (−1.26±0.62) subtypes (p values all < 0.0001). No significant difference was detected between other subtype pairs. In CN analysis, Basal (0.09±0.22) was lower than HER2+ (0.32±0.308, p < 0.0002) and LB (0.33±0.41, p < 0.0001), and LA (0.14±0.28) is lower than HER2 (p < 0.0016) and LB (p < 0.0001), but no other significant CN difference between the subtypes were found. The means and SDs are provided for reference only. No correlation of p53 mutation status and AURKA expression were observed. However, AURKA gene expression level is correlated with MKI67 gene expression (R = 0.69, p < 2.2e−16), and its correlation with PAM50 proliferation score is even higher (R = 0.80, p < 2.2e−16).

Discussion: Using the TCGA data we observed that the mean gene expression level of AURKA is significantly lower in LA than the other IBC subtypes, by more than 50% (note the log2 transformation). This differential expression is not completely due to CN changes (especially for the Basal subtype). There is a strong association with other tumor cell proliferation markers such as the MKI67 gene and the PAM50 proliferation score. We are using computational and laboratorial studies to better understand the role of AURKA in the etiology of invasive breast cancers.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Defense, or U.S. Government.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P2-05-21.

Abstract P5-01-07: Fibroadenomatoid changes are more prevalent in middle-aged women and have a positive association with invasive breast cancer

Background: The role of benign breast diseases (BBDs) in the development of invasive breast cancers (IBCs) has been studied for many years. Some BBDs have been studied comprehensively (e.g., fibrocystic changes (FCC)) while less is known about other BBDs (e.g., fiboadenomatoid changes (FAC)). FAC has been considered by some researchers as a precursor of fibroadenoma (FA). Conclusions from different studies vary, partially due to different interpretation methods and diagnostic criteria when multiple hospitals and pathologists were involved. In this study, we used subjects in the Clinical Breast Care Project (CBCP) from a military medical center where pathology slides were reviewed by a single breast pathologist to study FAC, FA, and FCC in comparison to the published literature.

Methods: Subjects were enrolled in the study following IRB-approved, HIPAA-compliant protocols. All the clinicopathologic data are available from the CBCP data warehouse (DW4TR). In the CBCP, FCC is composed of 4 components: stromal fibrosis, cysts, apocrine metaplasia, and sclerosing adenosis. Two modeling studies were performed. i) For the BBDs and IBC association study, two groups of subjects were identified: 1136 subjects diagnosed with “Benign” or “Atypical” diseases, and 619 cases diagnosed with IBCs. A logistic regression model was developed for the prediction of IBCs by the 3 BBDs and 2 well-established risk factors (RF): age (younger, <=40; middle-aged, 41–60; older, >60) and race (Caucasian, African American, Asian, and other). ii) For the RF association study with the BBDs, 6 additional RFs reported to be associated with these BBDs were identified from the literature: current use of oral contraceptives, number of live births, education, body mass index, hormonal replacement therapy, and IBC family history. These 8 RFs were used to develop a logistic regression model for each of the BBDs. The analyses were performed in SAS.

Results: In the first study, age and race were confirmed as RFs for IBCs. FAC was positively associated with IBC (OR = 3.04, 95%CI=2.06 to 4.50). FA was negatively associated with IBC, and the level of the association was stronger in women without FCC (OR = 0.15, 95%CI=0.08 to 0.28), compared to women with FCC (OR = 0.40, 95%CI=0.24 to 0.65). FCC was not significantly associated with IBC. Results from the second study indicated that, age was significantly associated with FAC (p = 0.015), specifically the middle-aged women were more likely to have FAC compared to younger women (OR = 2.03, 95%CI=1.23 to 3.34), while the older women were at a non-significantly increased risk. Trends of association with FAC were also noted for the number of live birth (p = 0.095), ethnicity (p = 0.096), and current oral contraceptive pill use (p = 0.077). The FCC model results were in general consistent with the literature, and we also confirmed that age was negatively associated with the diagnosis of FA.

Discussion: Our study was consistent with FCC findings in the literature. We observed that FAC was positively associated with IBC, whereas FA was negatively associated. Also, FAC occurred more often in middle-aged women while FAs occurrence was higher in younger women. Our results suggest that FAC and FA may be two different diseases.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P5-01-07.

P1-06-24: Nuclear Localization of Stat5a Predicts Response to Antiestrogen Therapy and Prognosis of Clinical Breast Cancer Outcome

Nuclear-localized and tyrosine-phosphorylated Stat5 has been reported as a favorable prognostic marker and predictor of response to antiestrogen therapy in breast cancer. Phospho-Stat5 antibodies do not distinguish between phosphorylated Stat5a and the closely related Stat5b, but Stat5a is considered more critical for normal mammary development than Stat5b. The purpose of this study was to determine whether levels of nuclear-localized Stat5a protein (Nuc-Stat5a) were prognostic of clinical outcome or predictive of antiestrogen response. Stat5a was detected by traditional diaminobenzidine-chromogen immunohistochemistry (IHC) and pathologist scoring or by quantitative immunofluorescence in five archival cohorts of breast cancer. Levels of nuclear-localized Stat5a (Nuc-Stat5a) were evaluated by pathologist scoring of whole tissue sections detected by IHC or automated quantitative analysis (AQUA) of immunofluorescently-labeled tissue microarrays. Levels of Nuc-Stat5a were reduced in invasive breast cancer tissues and lymph node metastases compared to normal tissue and ductal carcinoma in situ when quantified by AQUA (Material I; n=180). Tissues from patients not treated with adjuvant therapy or treated with antiestrogen monotherapy were analyzed according to Nuc-Stat5a status for breast cancer-specific survival (CSS) and time to recurrence (TTR) using univariate and multivariate statistical models, adjusting for clinical features including tumor grade, size, lymph node and ER, PR and Her2 status. In two prognostic cohorts of node-negative breast cancer patients, low expression of Nuc-Stat5a, detected by standard IHC (Material II; n=223) or quantitative analysis (Material III; n=198), was prognostic of poor breast cancer outcome as measured by univariate and multivariate CSS (Material II/III) and TTR (Material II). CSS and TTR analysis of two independent materials of tumors from patients treated with antiestrogen monotherapy and analyzed by standard IHC (Material IV; n=73) or quantitative immunofluorescence (Material V; n=97) indicated that patients whose tumors expressed low levels of Nuc-Stat5a were at a greater than 4-fold risk of antiestrogen therapy failure when adjusted for hormone receptor status and clinical features (multivariate CSS: Material IV HR=4.3 (1.2,15.6), p=0.03; Material V HR=5.0 (1.87,13.06), p=0.001). In conclusion, loss of Nuc-Stat5a is a promising independent marker of poor breast cancer prognosis in node-negative, non-adjuvant treated breast cancer patients. Additionally, Nuc-Stat5a may be a useful clinical tool to predict tumor response to antiestrogen therapy.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P1-06-24.

P1-03-06: Fibroadenomatoid Changes Have a Higher Occurrence Rate in Middle-Aged Benign Breast Disease Patients with the Trend Retained in Cancer Patients

Background: Fibroadenoma (FA) is a common benign breast lesion known to have a high incidence rate in younger women. There are controversial reports whether FA elevates the risk of developing breast cancers. In clinical practice, FA may be surgically removed due to multiple reasons making it complicated to study its impact on the development of breast cancers that have a higher incidence rate in older women. Fibroadenomatoid change (FAC), also known as fibroadenomatous hyperplasia, is an uncommon lesion with histologic features similar to that of FA but lacking well-defined borders and usually discovered incidentally on breast biopsy specimens. FAC is not surgically targeted. The Walter Reed Army Medical Center, through the Clinical Breast Care Project, has enrolled over 2000 subjects undergoing a biopsy; all the pathology was reviewed by a single pathologist. These subjects provide an opportunity to study the age-dependent pattern of FAC in different patient populations.

Methods: Subjects were enrolled following IRB-approved protocols with data collected through two comprehensive questionnaires, a Core Questionnaire and a Pathology Checklist. A total of 1964 female subjects were identified for this study, including 1135 benign/atypical, 192 in situ, and 637 invasive cancer patients. Patients were divided into three age groups: <=45 years, 46–65 years, and >=66 years. Chi-Square test in the SAS was used for statistical analysis.

Results: As shown in the table, FA occurrence rate decreases significantly with increasing age in benign disease patients. FAC, on the other hand, shows a significantly higher occurrence rate in middle-aged patients with benign findings, and this trend is retained in the invasive or in situ cancer populations. FAC rate is also significantly higher in patients with cancer (invasive, or invasive and in situ combined) compared to benign patients in each age group with p-values ranging from 0.0001 to 0.019 (not shown).

P3-05-02: Subtype-Specific Co-Occurrence of Atypical Hyperplasia and In Situ Carcinoma with Invasive Breast Cancers

Background

Atypical ductal hyperplasia (ADH), lobular carcinoma in situ (LCIS), and ductal carcinoma in situ (DCIS) are considered risk factors for the development of invasive breast cancer (IBC). The co-occurrence of these lesions with IBC may provide insights into cancer initiation and development. IBC subtypes have distinct clinicopathological features. A clinically practical IHC-based subtyping classification has been developed based on the expression of ER, PR, HER2, and Ki67, defining Luminal A (LA), Luminal B (LB), HER2+, and Triple Negative (TN) subtypes. The Walter Reed Army Medical Center (WRAMC), through the Clinical Breast Care Project (CBCP), has enrolled over 500 IBC subjects with single pathologist review and central lab analysis. The co-occurrence of ADH, LCIS, and DCIS will be studied in relation to IBC subtypes.

Methods: Subjects were enrolled following IRB-approved protocols. IBC patients enrolled at WRAMC were selected and their clinical and pathology data were reviewed. ER and PR positivity is defined as > 5% nuclear staining. The HER2 result is negative if the IHC=0 or 1+ and positive if IHC=3+. For IHC=2+, the FISH result determines the final HER2 status. Ki67 is considered positive if nuclear staining is >= 15%. For IBC subtypes, LA is ER+/HER2−/Ki67-; LB is either ER+/HER2−/Ki67+, or ER+/HER2+; HER2+ is ER-/PR-/HER2+; TN is ER-/PR-/HER2−. Statistical analysis was performed using SAS, and the Chi-Square test was used for categorical data analysis supplemented by the Fisher's Exact test where appropriate. For age analysis, ANOVA was performed with Bonferroni adjustment for multi-pair t-test.

Results: A total of 459 IBC patients were identified and categorized into LA (41.6%), LB (27.7%), HER2+ (10.2%), and TN (20.5%). Many of the previously reported subtype-specific characteristics were confirmed. Age at diagnosis varied by subtype (p=0.0034) with LA being the oldest (Mean±SD=59.9+12.5 years) and TN the youngest (54±12.6 years, p=0.0048). Ethnicity distribution of African American (AA) relative to Caucasian American patients varied significantly in subtypes with AA=18% in LA, 31% in LB, 32% in Her2+, and 42% in TN (p=0.0008). The grade, the AJCC stage and its components T and N were all significantly different among the subtypes (p ranges from <0.0001 to 0.0020). The grades and stages were consistently lowest for LA, highest for HER2+ and TN. We further found that the co-occurrence of ADH, DCIS, and LCIS with IBC were subtype-specific with the following distributions: ADH—LA (25.1%), LB (18.9%), HER2+ (0%), and TN (6.4%) (p<0.0001, n=78); DCIS—LA (63.4%), LB (76.4%), HER2+ (80.9%), and TN (58.5%) (p=0.0039, n=311); LCIS—LA (36.7%), LB (19.7%), HER2+ (4.3%), and TN (6.4%) (p<0.0001, n=103).

Discussion: By including Ki67 in IHC-based IBC subtyping we confirmed many subtype-specific clinico-pathological characteristics in the CBCP WRAMC population. We further report subtype-specific co-occurrences of ADH, DCIS, and LCIS. These co-occurrence patterns may reveal distinct developmental mechanisms between the different subtypes of IBC.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P3-05-02.

Abstract P3-13-02: Ethnicity Difference of Benign Breast Diseases in Breast Cancer and Non-Cancer Patients

Background: Breast cancer (BC) is a heterogeneous disease. Ethnicity differences in BC for Caucasian (CA) and African American (AA) women have been reported, but there is no report on ethnicity differences in benign breast diseases (BBDs) though it is known that BBDs may be precursors or risk factors of BCs. In the Clinical Breast Care Project (CBCP), a comprehensive characterization of BBDs performed on breast biopsies makes it possible to conduct a study on the ethnicity difference of BBDs between CA and AA, in the Cancer and Non-Cancer groups respectively.

Method: CA and AA patients undergoing a biopsy were selected from CBCP, totaling 1,963. A Pathology Checklist is available, reporting 131 pathologic conditions including 83 BBDs. In addition, a Core Questionnaire covering information such as demographics, medical history, risk factors, life style, etc. is completed. The Cancer group is composed of 731 CAs and 170 AAs (including in situ, invasive, and malignant NOS), and the Non-Cancer group is composed of 748 CAs and 314 AAs. BBDs with a frequency of <1% in each group were removed from the study leaving a total of 26 BBDs analyzed for the Cancer and 25 BBDs for the Non-Cancer groups. Pearson's Chi-square test was used to analyze statistical significance.

Results: The Cancer group showed 6 BBDs significantly associated with the ethnicity; 2 were more frequent in CA, i.e., fat necrosis (5.5% CA vs. 0.6% AA, p =0.011), and mild intraductal hyperplasia (6.8% CA vs. 2.4% AA, p= 0.041). The other 4 BBDs were more frequently observed in AA, which were cysts (47.2% CA vs 56.5% AA, p=0.036), multiple papillomas (8.3% CA vs 15.9% AA, p=0.005), moderate intraductal hyperplasia (18.1% CA vs 28.2% AA, p=0.004), and fibroadenomatoid nodule (5.3% CA 11.8% AA, p= 0.004). These BBDs were not significantly different in the Non-Cancer group between the two ethnicities. Six BBDs were significantly associated with the ethnicity in the Non-Cancer group. Four of them were more frequent in CA, including duct ectasia (9.5% CA vs. 3.2% AA, p=0.001), microcalcifications (35.3% CA vs. 27.4% AA, p=0.015), fibocystic changes (61.0% CA vs. 44.6% AA, p= 1.30E-06), and sclerosing adenosis (21.7% CA vs. 13.4% AA, p=0.002). The other 2 BBDs were more frequent in AA, i.e., sclerosing papilloma (0.7% CA vs 2.6% AA, p=0.025), and fibroadenoma (20.9% CA vs 29.9% AA, p=0.002). These BBDs were not significantly different in the Cancer group between the two ethnicities, for example microcalcifications were 50.1% in CA and 55.9% in AA.

Discussion: Multiple papillomas and moderate intraductal hyperplasia are moderate risk factors for BC, and in the Cancer group they were more frequently detected in AA than in CA. Microcalcification as a BC risk factor did not show ethnicity difference in the Cancer group, but was detected more frequently in CA in the Non-Cancer group. It is interesting that not a single BBD was found to be significantly associated with the ethnicity (AA and CA) across the Cancer and the Non-Cancer groups. Thus, the ethnicity difference of BBDs in AA and CA reported here not only suggests possible ethnicity-specific BC risk factors but also generates new hypotheses for future studies.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P3-13-02.

Differential expression of the 150-kd oxygen-regulated protein (ORP150) in benign, pre-malignant and malignant breast tissue

10770

Background: Heat shock protein, ORP150, plays a role in hypoxia/ischemia and angiogenesis. Preliminary studies demonstrate increased ORP 150 expression in human cancer, and associate its over expression with aggressive tumor biology. This study further evaluates ORP150 expression in different stages of breast cancer, such as benign, pre-malignant and malignant breast lesions and correlates it with clinical-pathological data. Methods: Sixty-six prospectively collected paraffin-embedded breast tissue sections were reviewed for diagnostic confirmation (normal, n=25; DCIS, n=20; invasive breast cancer (Ca), n=21) and stained using ORP150 antibody immunohistochemistry (IHC). Antibody-staining levels in the benign, DCIS and invasive component of each lesion were reviewed independently by two pathologists and scored: 0 (lowest) to 3+ (highest). Clinical-pathological data was compared with ORP150 staining using ANOVA or T-Test as appropriate with significance determined by p<0.05. Results: Significant differential ORP150 staining was detected in benign-normal versus benign adjacent to invasive cancer, as well as in benign adjacent to DCIS versus benign adjacent to invasive cancer. ORP150 expression in the invasive portion of the breast cancer correlated significantly with tumor grade and absence of hormone receptor expression, presence of lymphovascular invasion and lymph node metastasis. Conclusions: ORP150 expression in breast cancer is associated with poor prognostic histological factors. As ORP150 is differentially expressed in benign tissue adjacent to invasive cancer, further study is warranted to determine its utility in detecting occult invasive cancer within benign biopsy specimens, as well as its putative role in tumor-stromal cell interactions.

[Table: see text]

No significant financial relationships to disclose.

Expression of cytokeratin 20 and CD44 protein in upper urinary tract transitional cell carcinoma: cytologic-histologic correlation

OBJECTIVE

To explore the potential utility of immunostaining for CK20 and CD44 protein isoforms in evaluating cases of upper urinary tract transitional cell carcinoma (UTTCC).

STUDY DESIGN

Of 105 consecutive patients diagnosed cytologically with UTTCC, 33 subsequently underwent open surgical procedures. Cytologic samples from these patients retrieved by aspiration and biopsy, and corresponding surgical specimens were graded and staged using World Health Organization/International Society of Urologic Pathologists criteria. Immunostaining for CK20, CD44 standard (CD44s) and CD44v6 isoform (CD44-v6) was performed on all available cytologic and surgical materials. Expression levels and distributions of these markers were correlated semiquantitatively with grade and stage.

RESULTS

Cytologically assigned grade correlated with final histologic grade in 19 of 31 cases examined (61%). However, tumor invasion was not accurately assessable in cytologic samples from the majority of these cases. Statistically significant correlations of both increasing tumor grade and stage with abnormal CK20 expression were found. In addition, a significant relationship between focal CD44 isoform expression loss and tumor grade was identified. However, CD44 isoform expression loss did not significantly correlate with increasing tumor stage.

CONCLUSION

Although cytologic tumor grading of UTTCC was accurate, invasion could not be adequately assessed. As an adjunct to morphologic analysis, immunostaining for CK20 and CD44 may aid in the clinical evaluation of UTTCC tumor stage and biologic behavior prior to definitive therapy.

Lupus profundus, indeterminate lymphocytic lobular panniculitis and subcutaneous T-cell lymphoma: a spectrum of subcuticular T-cell lymphoid dyscrasia

INTRODUCTION

The diagnosis and classification of lymphocytic lobular panniculitis (LLP) has historically proven to be a difficult challenge. We encountered 32 cases of primary LLP which could be categorized as: 1) lupus erythematosus profundus (LEP) (19 patients); 2) an indeterminate group termed indeterminate lymphocytic lobular panniculitis (ILLP) (6 patients); and 3) subcutaneous T-cell lymphoma (SCTCL) (7 patients).

OBJECTIVE

We attempted to better define the subtypes of LLP by morphologic, phenotypic and genotypic features and to correlate those features to clinical presentation and outcome.

METHOD

Skin biopsy material was studied by conventional light microscopy, through immunophenotyping performed on sections from paraffin-embedded, formalin-fixed tissue and in some cases on sections of tissue frozen after receipt in physiological (Michel's) medium, and by polymerase chain reaction single-stranded conformational polymorphism analysis to assess for clonality of T-lymphocytes. Clinical features were correlated to histologic, phenotypic, and genotypic analyses.

RESULTS

Patients with LEP had a prior diagnosis of LE or overlying skin changes which light microscopically were characteristic of LE. Patients with ILLP had no concurrent or prior history of LE, no systemic symptoms or cytopenias, and a clinical course not suggestive of lymphoma. Cases of SCTCL showed hemophagocytic syndrome and/or lesional progression with demise attributable to the disease. Lesions in all groups showed proximal extremity predilection. Females predominated in the LEP group. The average age of onset was 38, 40 and 55 years in the LEP, ILLP and SCTCL groups, respectively. Cytopenia was seen in 4 LEP patients; 1 also developed fever. In LEP and ILLP, lesions resolved with hydroxychloroquine and/or steroid therapy, with recurrences following cessation of therapy. In the SCTCL group 4 developed hemophagocytic syndrome, 4 died within 2 years of diagnosis, and 3 went into remission following chemotherapy. The LEP and SCTCL groups manifested histological similarities: dense perieccrine and lobular lymphocytic infiltration, lymphoid atypia, histiocytes with ingested debris, eosinophilic necrosis of the fat lobule and thrombosis. The atypical lymphocytes although pleomorphic did not have a cerebriform morphology. The infiltrate in ILLP had a similar cytomorphology and distribution with variable angioinvasion which in all save one case was of lesser intensity and was not associated with significant fat necrosis or vasculitis. Germinal centers, dermal/subcuticular mucin deposition and an atrophying interface dermatitis with hyperkeratosis and follicular plugging were largely confined to the LEP group. Erythrophagocytosis, characteristic of SCTCL, usually indicated a supervening subcuticular lymphoid dyscrasia when encountered in ILLP and LEP. SCTCL showed a selective loss of CD5 expression with or without diminution in CD7 and monoclonal CD3 expression. Of 4 cases studied, 3 showed a CD8 dominant infiltrate while 2 others exhibited CD56 and CD30 positivity, respectively. All cases of SCTCL with amplifiable DNA showed T-cell clonality. Similar molecular and phenotypic features indicative of subcuticular lymphoid dyscrasia were encountered in cases of LEP and ILLP including a reduction in CD5, CD7, and/or monoclonal CD3 expression, a preponderance of CD8 lymphocytes within the subcutaneous fat and T-cell clonality. These cases showed lymphoid atypia with variable erythrophagocytosis. Cases of phenotypically abnormal and/or clonal LEP showed one or more of local destruction, lesional size progression, fever, and cytopenias, but lesions responded to hydroxychloroquine and/or prednisone therapy and death attributable to panniculitis could not be documented. Cases that were phenotypically normal and without clonality had none of the aforesaid atypical clinical features.

CONCLUSION

Lymphoid atypia, erythrophagocytosis, loss of certain pan T-cell markers, a reduced CD4/8 ratio and TCR rearrangement define subcuticular T-cell lymphoid dyscrasia, including a subset of LEP and ILLP. The subcuticular lymphoid infiltrates represent a spectrum of histologic, immunophenotypic, and molecular abnormalities which range from those which are clearly benign to those which are clearly neoplastic, and also encompasses those cases which defy precise classification into the two aforesaid poles.

RB2/p130 gene-enhanced expression down-regulates vascular endothelial growth factor expression and inhibits angiogenesis in vivo

Angiogenesis is an essential step in the progression of tumor formation and development. The switch to an angiogenetic phenotype can occur as a distinct step before progression to a neoplastic phenotype and is linked to genetic changes such as mutations in key cell cycle regulatory genes. The pathogenesis of the angiogenetic phenotype may involve the inactivation of tumor suppressor genes such as the "guardian of the genome," p53, and the cyclin-dependent kinase inhibitor p16. Retinoblastoma family member RB2/p130 encodes a cell cycle regulatory protein and has been found mutated in different tumor types. Overexpression of RB2/p130 not only suppresses tumor formation in nude mice but also causes regression of established tumor grafts, suggesting that RB2/p130 may modulate the angiogenetic balance. We found that induction of RB2/p130 expression using a tetracycline-regulated gene expression system as well as retroviral and adenoviral-mediated gene delivery inhibited angiogenesis in vivo. This correlated with pRb2/p130-mediated down-regulation of vascular endothelial growth factor protein expression both in vitro and in vivo.

Colorectal carcinomas with high microsatellite instability: defining a distinct immunologic and molecular entity with respect to prognostic markers

Molecular analysis of hereditary nonpolyposis colorectal carcinomas (HNPCC) has identified DNA mismatch repair deficiencies with resulting microsatellite instability (MSI) as a pathway of carcinogenesis that appears to be relevant for prognosis, treatment, and possibly prevention. In this study, expression of cell cycle proteins and other known prognostic markers is correlated with the microsatellite status of colorectal cancers (CRC). One hundred consecutive cases from the CRC Registry at Thomas Jefferson University were analyzed for MSI. Immunohistochemistry was performed for the mismatch repair proteins hMLH1 and hMSH2, tumor suppressor p53, apoptosis inhibitor bcl-2, cell cycle proteins p21(WAF1/CIP1), and p27 and the proliferation markers Ki-67 and topoisomerase II. High MSI (MSI-H) is significantly correlated with loss of either hMLH1 or hMSH2, presence of bcl-2, and absence of p53. p21(WAF1/CIP1) is positive in all tumors with MSI-H. Previous findings of a lower proliferation rate were confirmed with a topoisomerase II stain. Microsatellite stable (MSS) tumors generally express both MSH2 and MLH1. Other highly significant differences are positive p53 in 56% of MSS cases and negative bcl-2 in 98% of MSS cases. p27 expression is found in approximately 50% of all CRCs irrespective of the microsatellite status. MSI-H tumors follow the mutator pathway, with loss of expression of one mismatch repair protein, wild-type p53, lower proliferation, and positivity for p21(WAF1/CIP1). MSS tumors follow the suppressor pathway, characterized by p53 overexpression, higher proliferation, and absence of bcl-2 expression; p21(WAF1/CIP1) expression can be variable. These data provide a molecular basis for the clinical observation that patients with HNPCC appear to have a more favorable prognosis. HUM PATHOL 31:1506-1514.

Overexpression of p21WAF1/CIP1 and MDM2 characterizes serous borderline ovarian tumors

Ovarian epithelial tumors are classically divided into benign, malignant, and borderline or of low malignant potential. It is controversial whether this last group of tumors should be considered benign or malignant. Expression of cell cycle markers has recently been linked to tumor behavior and response to treatment. It has been shown that one of the pathways through which the p53 gene controls the cell cycle is by transactivating p21WAF1/CIP1, a cyclin-dependent kinase (cdk) inhibitor. By inhibiting cdks, p21WAF1/CIP1 blocks the G-1 to S-phase transition in the cell cycle. p53 can be regulated by MDM2 (murine double minute-2) through direct inactivation or promotion of its cytoplasmic degradation. In an attempt to investigate the cell cycle checkpoint mechanisms of these tumors, we studied the expression of p53, Ki-67, MDM2, and p21WAF1/CIP1 by immunohistochemistry. We analyzed the expression of these proteins in 19 cystadenomas (8 serous and 11 mucinous), 40 borderline tumors (31 serous and 9 mucinous), and 18 serous carcinomas of the ovary. p21WAF1/CIP1 was expressed in 7 of 19 (37%) benign cystadenomas, 32 of 40 (80%) borderline tumors (93.5% of serous and 33% of mucinous), and in 9 of 18 (50%) serous carcinomas. Ki-67 was only weakly expressed in 8 of 19 (42%) benign cystadenomas, all borderline tumors showed Ki-67 staining in less than 50% of the cells, and 55% of serous carcinomas stained in more than 50% of tumor cells. p53 was absent in all but 1 of the cystadenomas, was expressed in 9 of 40 (22.5%) borderline tumors (25.8% of serous and 11% of mucinous), and in 10 of 18 (55%) carcinomas. All 11 implants of serous borderline tumors expressed p21WAF1/CIP1. Most serous borderline tumors expressed higher levels of MDM2 compared with the benign cystadenomas and carcinomas. Four of the serous borderline implants (40%) expressed MDM2. Coexpression of p21WAF1/CIP1 and MDM2 characterizes serous borderline tumors of the ovary and their implants, which suggests that these cell cycle control proteins are important in these tumors and may be related to tumor progression. Low expression of p53 protein in serous borderline tumors might be in part mediated by MDM2. This suggests that the p53 pathway is intact in most of these tumors, in contrast with carcinomas, in which high expression of p53 has been related to mutations of this gene.

Immunohistochemical localization of prostate-specific antigen in ductal epithelium of male breast. Potential diagnostic pitfall in patients with gynecomastia

Enlargement of the male breast is frequently encountered in the course of adjuvant antiandrogen therapy for advanced prostate carcinoma. The clinical differential diagnosis in this setting includes hormonal imbalance-induced gynecomastia, primary breast carcinoma, and metastasis of prostatic carcinoma. Biopsy of the lesion with the identification of prostate-specific antigen (PSA) plays an important role in establishing the correct diagnosis. Recent studies showed that female mammary epithelium may be a significant source of PSA, but its expression in male breasts has not been sufficiently studied. We found that normal and hyperplastic duct epithelium in gynecomastia exhibited focal, strong (+3) PSA immunoreactivity in 5 of 18 cases (28%). In contrast, no PSA reactivity was found in eight cases of male breast carcinoma. No reactivity was seen with antiprostatic acid phosphatase (PsAP) antibody, in either benign or malignant epithelium. Frequent expression of PSA in gynecomastia may, in an appropriate clinical setting, cause confusion with metastatic prostatic carcinoma. The lack of immunoreactivity for PsAP in male breast epithelium indicates its usefulness in the differential diagnosis.

Cyclin E expression and early cervical neoplasia in ThinPrep specimens. A feasibility study

OBJECTIVE

To investigate cyclin E expression as a possible marker for early cervical neoplasia using ThinPrep gynecologic specimens from premenopausal women.

STUDY DESIGN

Archived ThinPrep liquid-based cervical/endocervical specimens (Cytyc Corporation, Boxborough, Massachusetts, U.S.A.) diagnosed as human papillomavirus infection (HPV) (20), atypical squamous cells of undetermined significance (ASCUS) (48) and within normal limits (WNL)/benign cellular changes (BCC) (21) were resampled in duplicate, fixed in 95% ethanol, subjected to immunocytochemical staining with the cyclin E antibody (clone 13A3, Novocastra Laboratories Ltd., Newcastle upon Tyne, U.K.) and HPV antibody (clone K1H8, Dako Corporation, Carpinteria, California, U.S.A.) and the expression scored by two pathologists and correlated with the cytologic diagnosis. A case was scored as positive if it contained > 10 abnormal squamous cells with nuclear immunocytochemical staining.

RESULTS

The cylin E antibody assay was positive in 20 (100%) cases cytologically diagnosed as HPV. These cases were also anti-HPV antibody positive. Four cases (19%) cytologically diagnosed as WNL/BCC were cyclin E positive. Of these, two were anti-HPV antibody positive. Thirty-four (73%) cases cytologically diagnosed as ASCUS were positive for the cyclin E assay and for anti-HPV antibody staining.

CONCLUSION

Cyclin E expression correlates strongly with morphologic features of HPV in ThinPrep specimens and may serve as a surrogate marker for HPV infection and early cervical preneoplastic lesions.

Extramammary Paget disease is characterized by the consistent lack of estrogen and progesterone receptors but frequently expresses androgen receptor

Extramammary Paget disease (EPD) is an uncommon cutaneous malignant neoplasm that arises in areas rich in apocrine glands (perineum, vulva, and axilla). Apocrine gland origin or apocrine differentiation of cells of EPD has been suggested. Estrongen, progesterone, and androgen hormone receptors have been reported to exhibit a characteristic pattern of expression in mammary apocrine type carcinomas; however, their expression in EPD has not been elucidated fully. By using immunohistochemical methods, we studied the expression of steroid receptors in EPD on formalin-fixed paraffin-embedded tissue samples from 28 patients with EPD without associated visceral malignant neoplasms or adnexal carcinoma. Androgen receptor (AR) was identified in 15 of 28 cases. The proportion of AR-positive cells varied from 1% to more than 75%; 8 cases expressed AR in more than 10% of cells. Strong AR expression also was seen in the invasive carcinoma arising from 1 case of EPD. All cases lacked immunohistochemically detectable estrogen and progesterone receptors. The immunophenotype characteristic of apocrine carcinomas (AR-positive, estrogen receptor-negative, progesterone receptor-negative) was seen in a substantial proportion of EPD cases. Results suggest that AR expression is a factor in pathogenesis of EPD. This may be important for the therapy of recurrent or invasive disease.

Loss of FHIT expression in transitional cell carcinoma of the urinary bladder

Cytogenetic and loss of heterozygosity (LOH) studies demonstrated chromosome 3p deletions in transitional cell carcinoma (TCC). We recently cloned the tumor suppressor gene FHIT (fragile histidine triad) at 3p14.2, one of the most frequently deleted chromosomal regions in TCC of the bladder, and showed that it is the target of environmental carcinogens. Abnormalities at the FHIT locus have been found in tumors of the lung, breast, cervix, head and neck, stomach, pancreas, and clear cell carcinoma of the kidney. We examined six TCC derived cell lines (SW780, T24, Hs228T, CRL7930, CRL7833, and HTB9) and 30 primary TCC of the bladder for the integrity of the FHIT transcript, using reverse transcriptase-polymerase chain reaction (RT-PCR) to investigate a potential role of the FHIT gene in TCC of the bladder. In addition, we tested expression of the Fhit protein in the six TCC-derived cell lines by Western blot analysis and in 85 specimens of primary TCCs by immunohistochemistry. Three of the six cell lines (50%) did not show the wild-type FHIT transcript, and Fhit protein was not detected in four of the six cell lines (67%) tested. Fhit expression also was correlated with pathological and clinical status. A significant correlation was observed between reduced Fhit expression and advanced stage of the tumors. Overall, 26 of 30 (87%) primary TCCs showed abnormal transcripts. Fhit protein was absent or greatly reduced in 61% of the TCCs analyzed by immunohistochemistry. These results suggested that loss of Fhit expression may be as important in the development of bladder cancer as it is for other neoplasms caused by environmental carcinogens.

Fhit expression in gastric adenocarcinoma: correlation with disease stage and survival

BACKGROUND

The FHIT gene is inactivated by deletion in a large fraction of human tumors, including gastric carcinomas, and the Fhit protein has been proposed to act as a tumor suppressor in multiple tumor types. A large fraction of gastric adenocarcinomas have lost expression of the candidate tumor suppressor protein, Fhit, whereas normal gastric epithelial cells are strongly positive and Fhit loss has been found to correlate with alterations of the FHIT locus. Because the majority of gastric tumors in the current study were found to be entirely negative for Fhit protein, it is possible that alteration of the carcinogen-susceptible fragile region within the FHIT gene is an early event in gastric carcinoma, as it is in lung carcinoma.

METHODS

To determine whether the absence of Fhit protein correlates with expression of tumor markers or with clinical parameters, such as grade, stage, and survival time, the authors assessed Fhit expression using immunohistochemistry in a well characterized set of 55 gastric adenocarcinomas resected over several years, with longitudinal follow-up of patients for outcome.

RESULTS

In this set of 55 gastric cancers, the absence of Fhit protein correlated with higher tumor stage (P = 0.003) and higher histologic grade (P = 0.007). In addition, patients whose tumors had lost expression of Fhit died of disease significantly earlier than those with Fhit positive tumors (P = 0.017). The absence of Fhit expression did not correlate with the expression of any tumor markers.

CONCLUSIONS

Larger studies will be required to elucidate further the relation between tumor stage, grade, and Fhit loss and to determine whether inclusion of Fhit antiserum in immunophenotyping of gastric adenocarcinomas will be a useful indicator of post-diagnosis prognosis.

p53 immunohistochemistry for distinguishing reactive mesothelium from low grade ovarian carcinoma

OBJECTIVE

To determine the utility of immunohistochemical staining for p53 in cell block material for distinguishing reactive mesothelium from borderline or low grade ovarian carcinoma.

STUDY DESIGN

Paraffin-embedded cell blocks from paracentesis and pelvic wash fluid of 44 cases of ovarian carcinoma and 20 cases containing only reactive mesothelium were immunostained for p53 using monoclonal antibody DO-7. Tumor grades ranged from borderline to high grade and were serous papillary (33), clear cell (3), mucinous (2), endometrioid (2), mixed serous papillary/clear cell (3) and undifferentiated (1). The three authors independently evaluated the staining, including estimation of the percentage and intensity of positive nuclear staining.

RESULTS

A separation of positive from negative cases was seen when staining intensity was considered the critical parameter; moderate to strong staining was considered truly positive. Seventy-three percent (8/11) of borderline tumors, 80% (8/10) of low grade tumors and 65% (15/23) of intermediate to high grade tumors showed moderate to strong positivity. Percentage of staining was a less-reliable parameter as 25% of negative cases were positive by this assessment.

CONCLUSION

p53 Immunohistochemistry, using monoclonal antibody DO-7 combined with standard morphologic evaluation, may be useful in distinguishing benign reactive mesothelium from borderline or low grade ovarian carcinoma.

Loss or reduction of Fhit expression in renal neoplasias: correlation with histogenic class

Involvement of the 3p14.2 region of chromosome 3 in kidney cancers was suggested 20 years ago, when a reciprocal constitutional translocation, t(3;8)(p14.2;q24), was shown to segregate with bilateral clear cell renal carcinoma in 3 generations of 1 family. The FHITgene that is interrupted at 3p14.2 by the t(3;8) translocation has been isolated, characterized, and shown to be frequently altered, mainly by internal deletion, in carcinomas or cancer-derived cell lines of the lung, stomach, pancreas, esophagus, cervix, and colon. Although up to 90% of sporadic clear cell renal carcinomas, representing 70% of adult renal carcinomas, exhibit loss of FHIT alleles, FHIT gene alterations have been documented for only a few renal cell carcinoma-derived cell lines. Nevertheless, more than 50% of clear cell carcinomas were recently shown to express little or no Fhit protein, unlike the normal kidney tubule epithelium, which is uniformly strongly positive for Fhit expression. We have extended our immunohistochemical study of expression of Fhit protein to the spectrum of histopathologic subtypes of adult renal tumors. There is an apparent continuum of Fhit expression from the 100% strongly positive oncocytomas through mostly positive papillary and chromophobe to the mostly negative clear cell and sarcomatoid to the negative or predominantly negative collecting duct renal carcinomas. This pattern of diminishing Fhit expression correlates with reported frequency of 3p allele loss in renal carcinomas and may parallel the potential for aggressive behavior of tumors, as suggested by the abundant Fhit expression in the benign oncocytomas and the near absence of Fhit expression in sarcomatoid and collecting duct RCCs.

Gastrointestinal stromal tumors/smooth muscle tumors (GISTs) primary in the omentum and mesentery: clinicopathologic and immunohistochemical study of 26 cases

Gastrointestinal stromal tumor or smooth muscle tumor (GIST) is the designation for a major subset of gastrointestinal mesenchymal tumors that histologically, immunohistochemically, and genetically differ from typical leiomyomas, leiomyosarcomas, and schwannomas. Because GISTs, like the interstitial cells of Cajal, the gastrointestinal pacemaker cells, express CD117 (c-kit protein), the origin of GISTs from the interstitial cells of Cajal has been recently proposed. Comparison of GISTs primary in the omentum and mesentery to GISTs primary in the tubular gastrointestinal tract is of particular diagnostic and histogenetic interest in view of the possible similarity of these tumors with the GIST group. In this study, we analyzed 14 omental and 12 mesenteric primary mesenchymal tumors representing smooth muscle tumors or GISTs. These tumors were phenotypically compared with gastric and small intestinal GISTs, leiomyomas of the esophagus, and leiomyosarcomas of the retroperitoneum. Most (13 of 14) omental and mesenteric (10 of 12) tumors showed histologic features similar to GISTs with elongated spindle cells or epithelioid cells with high cellularity; most of these tumors showed low mitotic activity. Omental and mesenteric GISTs were typically positive for CD117 and less consistently for CD34. They often showed alpha-smooth muscle actin reactivity but were virtually negative for desmin and S-100 protein. One omental and two mesenteric tumors showed features of leiomyosarcoma with ovoid, less elongated nuclei, cytoplasmic eosinophilia; all these tumors had significant mitotic activity. These tumors were positive for alpha-smooth muscle actin and two of them for desmin, but all were negative for CD34 and CD117, similar to retroperitoneal leiomyosarcomas. Tumor-related mortality occurred in the group of mesenteric GISTs, but not in the group of omental GISTs. In contrast, all three patients with a true leiomyosarcoma of the omentum or mesentery had documented liver metastases or died of tumor. In summary, we show that tumors phenotypically identical with GISTs occur as primary tumors in the omentum and mesentery. The occurrence of CD117-positive tumors outside the gastrointestinal tract militates against an origin of these tumors exclusively from the interstitial cells of Cajal.

Intratumoral recombinant GM-CSF-encoding virus as gene therapy in patients with cutaneous melanoma

Seven immunocompetent, revaccinated patients with surgically incurable cutaneous melanoma underwent treatment of dermal and/or subcutaneous metastases with twice-weekly intratumoral injections of escalating doses (10(4)-2 x 10(7) plaque-forming units (PFU)/lesion; 10(4)-8 x 10(7) PFU/session) of a vaccinia/GM-CSF recombinant virus for 6 weeks. Patients with stable or responding disease were maintained on treatment until tumor resolution or progression. Systemic toxicity was infrequent, dose-related, and limited to mild flu-like symptoms that resolved within 24 hours. Local inflammation, at times with pustule formation, was consistently seen with doses of > or =10(7) PFU/lesion. Chronically treated lesions showed a dense infiltration, with CD4+ and CD8+ lymphocytes, histiocytes, and eosinophils. All seven patients developed an antivaccinia humoral immune response 14-21 days following revaccination. Despite the presence of these antivaccinia antibodies, the reporter gene was expressed, as judged by the development of anti-beta-galactosidase antibodies in all patients. Passenger cytokine gene function was evidenced by the presence of virally encoded GM-CSF mRNA at injection sites both early (weeks 1 and 5) and late (week 31) in the course of treatment. Eosinophilia at treatment sites indicated that physiologically significant levels of functional cytokine were generated. However, there were no changes in the total number of peripheral white blood cells or in the numbers or percentages of polymorphonuclear leukocytes, monocytes, or eosinophils. GM-CSF was not detected in the sera. The two patients with the largest tumor burdens failed to respond even at treatment sites. Three patients had mixed responses, with regression of treated and untreated dermal metastases and progression of disease elsewhere. One patient had a partial response, with regression of injected and uninjected regional dermal metastases. Residual melanoma was excised, rendering the patient disease free. One patient with only dermal metastases confined to the scalp achieved a complete remission. Sequential administration of escalating doses of a GM-CSF recombinant vaccinia virus is safe, effective at maintaining passenger gene function, and can induce tumor regression.

Calponin and h-caldesmon in soft tissue tumors: consistent h-caldesmon immunoreactivity in gastrointestinal stromal tumors indicates traits of smooth muscle differentiation

Currently, the immunohistochemical evaluation of smooth muscle differentiation is usually based on desmin, which also reacts with skeletal muscle and is not present in all smooth muscle tumors, and alpha-smooth muscle actin, which reacts with myoepithelial cells. Neither marker typically reacts with gastrointestinal stromal tumors (GISTs), previously classified as smooth muscle tumors or presently often classified as smooth muscle/stromal tumors. Two cytoskeleton-associated actin-binding proteins, calponin (CALP) and h-caldesmon (HCD), are putative smooth muscle markers that also react with myoepithelia. These markers are of particular interest in the immunohistochemical analysis of tumors; neither of them has been extensively documented in soft tissue tumors. In this study, we evaluated selected normal and reactive tissues and more than 250 mesenchymal tumors for CALP and HCD. Both markers were expressed in parenchymal and vascular smooth muscle cells in various organs and in myoepithelial cells. CALP also reacted with myofibroblasts of desmoplastic stroma. All of our 25 benign smooth muscle tumors from various locations were positive for CALP and HCD, as were most of the retroperitoneal and uterine leiomyosarcomas. HCD was more specific, because CALP also reacted with myofibroblastic lesions. The common reactivity of malignant fibrous histiocytomas with CALP and HCD suggests a combination of myofibroblastic and smooth muscle differentiation in these tumors. The GISTs (c-kit positive, usually actin negative) showed nearly consistent HCD reactivity, suggesting traits of smooth muscle differentiation. GISTs were usually CALP negative and showed a CALP expression pattern similar to that of alpha-smooth muscle actin. Although nonmuscle, nonmyofibroblastic tumors were negative for CALP and HCD, synovial sarcomas showed streaks of CALP-positive cells of unknown significance. CALP and HCD should be explored as markers to identify myofibroblastic and smooth muscle cell differentiation in mesenchymal tumors.

Reduced Fhit expression in sporadic and BRCA2-linked breast carcinomas

Evidence for alteration of the FHIT gene in a significant fraction of breast carcinomas has been reported, in apparent concordance with loss of heterozygosity (LOH) at chromosome region 3p14.2 in breast cancer and benign proliferative breast disease. A significantly higher frequency of LOH at the FHIT locus was reported for BRCA2-/- tumors, possibly due to misrepaired double-strand breaks at this common fragile region. To determine whether such genomic alterations lead to Fhit inactivation, we have assessed the level of Fhit expression by immunohistochemical detection in sporadic tumors and cancers occurring in BRCA2 999del5 carriers. To determine whether Fhit inactivation may have prognostic significance, we have also assessed expression of breast cancer markers and clinical features in sporadic tumors relative to Fhit expression. Of 40 consecutive sporadic breast carcinomas studied for tumor markers, 50% showed reduced Fhit expression. In these sporadic cancers, loss of Fhit expression was not correlated significantly with the presence or absence of other tumor markers. In a study of 58 sporadic and 34 BRCA2 999del5 Icelandic invasive cancers, there was a significant association of LOH at 3p14.2 with reduced expression of Fhit (P = 0.001); also the lower expression of Fhit and higher LOH at 3p14.2 in BRCA2 999del5 tumors relative to sporadic cancers was significant (P = 0.002). Thus, genetic alteration at the fragile site within the FHIT gene leads to loss of Fhit protein in a significant fraction of sporadic breast cancers and a much larger fraction of familial breast cancers with an inherited BRCA2 mutation, consistent with the idea that loss of BRCA2 function affects stability of the FHIT/FRA3B locus.

MLH1 promoter methylation and gene silencing is the primary cause of microsatellite instability in sporadic endometrial cancers

Defective DNA mismatch repair in human tumors leads to genome-wide instability of microsatellite repeats and a molecular phenotype referred to as microsatellite instability (MSI). MSI has been reported in a variety of cancers and is a consistent feature of tumors from patients with hereditary non-polyposis colorectal cancer. Approximately 20% of cancers of the uterine endometrium, the fifth most common cancer of women world-wide, exhibit MSI. Although the frequency of MSI is higher in endometrial cancers than in any other common malignancy, the genetic basis of MSI in these tumors has remained elusive. We investigated the role that methylation of the MLH1 DNA mismatch repair gene plays in the genesis of MSI in a large series of sporadic endometrial cancers. The MLH1 promoter was methylated in 41 of 53 (77%) MSI-positive cancers investigated. In MSI-negative tumors on the other hand, there was evidence for limited methylation in only one of 11 tumors studied. Immunohistochemical investigation of a subset of the tumors revealed that methylation of the MLH1 promoter in MSI-positive tumors was associated with loss of MLH1 expression. Immunohistochemistry proved that two MSI-positive tumors lacking MLH1 methylation failed to express the MSH2 mismatch repair gene. Both of these cancers came from women who had family and medical histories suggestive of inherited cancer susceptibility. These observations suggest that epigenetic changes in the MLH1 locus account for MSI in most cases of sporadic endometrial cancers and provide additional evidence that the MSH2 gene may contribute substantially to inherited forms of endometrial cancer.

The role of deletions at the FRA3B/FHIT locus in carcinogenesis

The FHIT gene, which encodes a 1-kb message and a 16.8-kDa protein that hydrolyses diadenosine triphosphate (ApppA) to ADP and AMP in vitro, covers a megabase genomic region at chromosome band 3p14.2. The gene encompasses the most active of the common human chromosomal fragile regions, FRA3B. Over the years, it has been suggested that fragile sites might be especially susceptible to carcinogen damage and that chromosomal regions of nonrandom alterations in cancer cells may coincide with defined fragile sites. Within the FRA3B region, the characteristic induced chromosome gaps can occur across the entire region, but 60% of the gaps are centered on a 300-kb region flanking FHIT exon 5, the first protein-coding exon. Numerous hemizygous and homozygous deletions, translocations and DNA insertions occur within FHIT in cancer cell lines, uncultured tumors, and even in preneoplastic lesions, especially in tissues such as lung that are targets of carcinogens. This supports the proposed cancer-fragile site connection and suggests that the FHIT gene, expression of which is frequently altered in cells showing FHIT locus damage, is a tumor suppressor gene whose inactivation may drive clonal expansion of preneoplastic and neoplastic cells. Replacement of Fhit expression in Fhit-negative cancer cells abrogates their tumorigenicity in nude mice. Analysis of the approximately 300-kb DNA sequence encompassing FHIT exon 5 in the FRA3B epicenter has provided clues to the mechanism of repair of the fragile site double strand breaks. The mechanism involves recombination between LINE 1 elements with deletion of the intervening sequence, often including FHIT exons. These studies have also shown that FHIT alterations generally entail independent deletion of both FHIT alleles. Future studies will focus on two objectives: study of (1) the in vivo function of the Fhit protein and (2) mechanisms of break and repair in the FRA3B fragile region.

hMSH5: a human MutS homologue that forms a novel heterodimer with hMSH4 and is expressed during spermatogenesis

MutS homologues have been identified in nearly all organisms examined to date. They play essential roles in maintaining mitotic genetic fidelity and meiotic segregation fidelity. MutS homologues appear to function as a molecular switch that signals genomic manipulation events. Here we describe the identification of the human homologue of the Saccharomyces cerevisiae MSH5, which is known to participate in meiotic segregation fidelity and crossing-over. The human MSH5 (hMSH5) was localized to chromosome 6p22-21 and appears to play a role in meiosis because expression is induced during spermatogenesis between the late primary spermatocytes and the elongated spermatid phase. hMSH5 interacts specifically with hMSH4, confirming the generality of functional heterodimeric interactions in the eukaryotic MutS homologue, which also includes hMSH2-hMSH3 and hMSH2-hMSH6.

Expression of p21WAF1/CIP1 in soft tissue sarcomas: a comparative immunohistochemical study with p53 and Ki-67

The p53 gene controls the cell cycle by transactivating p21WAF1/CIP1, a cyclin dependent kinase (cdk) inhibitor. By inhibiting cdks, p21WAF1/CIP1 regulates the cell cycle by blocking the G1 to S phase transition. In this study, we analyzed the immunohistochemical expression of p21WAF1/CIP1 in 66 soft tissue sarcomas and its relationship to p53 and the cell cycle proliferation antigen Ki-67. Expression of p21WAF1/CIP1, was detected in 76% of the tumors and p53 in 26%. All malignant schwannomas, synovial sarcomas, leiomyosarcomas and gastrointestinal stromal tumors expressed p21WAF1/CIP1. The majority of angiosarcomas, dermatofibrosarcomas, and fibrosarcomas showed low expression or were negative for p21WAF1/CIP1. Ewing's sarcomas, liposarcomas, and malignant fibrous histiocytomas were heterogeneous in their expression of p21WAF1/CIP1. Combining p53 and p21WAF1/CIP1 staining, the following four patterns were observed: 23% of the tumors showed the p53+/p21+ pattern; 53% showed the p53-/p21+ pattern; 3% showed the p53+/p21- pattern and 21% were negative for both p53 and p21WAF1/CIP1. There was no correlation between Ki-67 and p21WAF1/CIP1 or p53 staining. Our results show that soft tissue sarcomas, independent of their histologic subtype, frequently express p21WAF1/CIP1 which is probably important in their tumorigenesis. Additionally, p21WAF1/CIP1 may play a role in determining the efficacy of various cell cycle-directed therapies in soft tissue sarcomas.

A comparative analysis of Ki-67, p53, and p21WAF1CIP1 expression in tamoxifen associated endometrial carcinomas

OBJECTIVE

To determine clinicopathologic parameters, expression of proliferation markers, and immunohistochemical oncogene expression in endometrial cancers in patients with a history of breast cancer with and without tamoxifen use.

METHODS

Thirty endometrial carcinoma specimens were examined from patients with a previous history of breast cancer. Patients who had taken tamoxifen (15) were compared to non-users (15). Immunohistochemical staining was performed for p53, Ki-67, and p21WAF1/CIP1, overexpression was defined as greater than 10% positivity.

RESULTS

Patient populations were statistically similar. P53 was overexpressed in 73% of tamoxifen users compared to 53% of non-users. Ki-67 was overexpressed in over 90% of user and non-user specimens. p21WAF1/CIP1 was overexpressed in 33% of users and 47% of non-users. Tamoxifen users had shorter time to diagnosis of endometrial cancer than non-users.

CONCLUSIONS

In this small study, tamoxifen associated tumors expressed p53 more frequently than non-users, while the opposite was observed with p21WAF1/CIP1. This suggests that p53 mutations might play a role in development of tamoxifen associated tumors.

Molecular alterations to human chromosome 3p loci in neuroendocrine lung tumors

BACKGROUND

The origins of and interrelations between low grade and high grade neuroendocrine lung tumors, typical and atypical carcinoids, and small cell lung carcinoma (SCLC) have not been elucidated. Karyotypic and molecular genetic studies have demonstrated deletions in 3p in 100% of SCLCs and the candidate lung tumor suppressor gene, FHIT, at 3p14.2 is not expressed in the majority of SCLCs. Similar studies of typical and atypical carcinoids could clarify the interrelations among these tumors.

METHODS

For molecular genetic analyses, archival carcinoids and paired normal cells were microdissected from paraffin sections, deparaffinized, and DNA prepared. Oligonucleotide primer pairs for 12 microsatellite markers mapping between 3p14.2 and 3p21.3 were used to amplify allelic DNA fragments from 13 typical and 6 atypical carcinoids. In addition, an independent series of archival sections of carcinoids and SCLCs was tested by immunohistochemistry for expression of Fhit protein.

RESULTS

Of the six atypical carcinoids examined, three had lost an allele at all informative markers, whereas one had lost alleles in two distinct regions and two showed allele loss in a subregion of the chromosome region tested. Of the 13 typical carcinoids, 3 showed allele loss at only 1 or 2 loci each. Typical carcinoids, similar to normal lung epithelia, were strongly positive for the cytoplasmic Fhit protein, SCLCs were uniformly negative, and atypical carcinoids appeared to express an intermediate level of Fhit protein.

CONCLUSIONS

Loss of heterozygosity at 3p14.2-p21.3 is significantly more extensive in all atypical carcinoids. Atypical carcinoids, which exhibit clinicopathologic features intermediate between typical carcinoids and small cell carcinomas and have been considered well differentiated neuroendocrine carcinomas, also are intermediate between typical carcinoids and SCLC on the basis of extent of loss of 3p alleles and reduced expression of Fhit protein.

Cell-type- and tumour-type-related patterns of bcl-2 reactivity in mesenchymal cells and soft tissue tumours

Bcl-2 is one of the many proteins that regulate programmed cell death and is overexpressed in B-cell lymphomas. The expression of bcl-2 in mesenchymal cells and soft tissue tumours was the subject of this study. Normal mesenchymal tissue and representative cases of soft tissue tumours of different types (n>200) were investigated immunohistochemically for bcl-2 expression. Although bcl-2 expression was normally relatively restricted to some smooth muscle cells and neural cells, bcl-2 immunoreactivity was widespread in different types of soft tissue neoplasms, both benign and malignant. Consistently positive tumours included solitary fibrous tumour, haemangiopericytoma, schwannoma and synovial sarcoma. The few soft tissue tumours that were consistently negative for bcl-2 included nodular fasciitis and desmoid tumour. Leiomyomas and leiomyosarcomas were heterogeneous; all uterine leiomyomas were bcl-2 positive, but all oesophageal leiomyomas were negative, paralleling the reactivity observed in the smooth muscle at those sites. Gastrointestinal stromal tumours showed bcl-2 reactivity; this was less consistent in malignant tumours. Along the malignancy gradient, there was no consistent trend in the bcl-2 reactivity. Dermatofibrosarcomas showed increase of bcl-2 expression with fibrosarcomatous transformation, whereas smooth muscle sarcomas and malignant peripheral nerve sheath sarcomas were less consistently positive than the corresponding benign neoplasms. We conclude that bcl-2 expression is widespread in soft tissue tumours, but shows constitutional expression patterns that are often parallel to the normal tissue counterparts. Compared with benign soft tissue tumours, bcl-2 expression is often reduced in sarcomas, but it cannot be used as a prognostic marker without correlation of the data to its phenotypic expression patterns.

CD117: a sensitive marker for gastrointestinal stromal tumors that is more specific than CD34

Gastrointestinal stromal tumors (GISTs) represent a distinct and the most important subset of mesenchymal tumors of the GI tract. These tumors are both phenotypically and genotypically different from true leiomyomas and usually express CD34, a hematopoietic progenitor cell antigen. CD34, however, is also present in a wide variety of fibroblastic and endothelial cell tumors. In this immunohistochemical study of CD117, we evaluated 85 cases of GIST and more than 150 other mesenchymal tumors, including leiomyomas and schwannomas. CD117, the c-kit proto-oncogene product, is expressed in subsets of hematopoietic stem cells, mast cells, melanocytes, and interstitial cells of Cajal of the GI tract. CD117 was almost always (85%) expressed in both benign and malignant GISTs. CD117 was observed both in the spindle cell and epithelioid subtypes of GISTs in all locations. In addition to reacting with the CD34-positive GISTs, CD117 was positive in some CD34-negative cases. Approximately one-third of GISTs coexpressed CD117 and smooth muscle actins. In contrast, true leiomyomas (desmin and actin-positive) and schwannomas in both GI and peripheral locations were consistently negative for CD117. Solitary fibrous tumors and Kaposi's sarcomas, which are typically CD34 positive, were consistently CD117 negative. Among the CD34-positive tumors that showed occasional CD117 reactivity were dermatofibrosarcoma protuberans (1 of 7) and hemangiopericytoma (2 of 10). Other mesenchymal tumors that were variably CD 117 positive included clear cell sarcoma (7 of 15), metastatic melanoma (9 of 25), and malignant fibrous histiocytoma (1 of 20). These results indicate that CD117 is a specific marker for GIST among tumors that occur in the GI tract and adjacent regions. CD117 expression also separates GISTs from true leiomyomas and gastric schwannomas.

Immunocytochemical evaluation of HBME-1, CA 19-9, and CD-15 (Leu-M1) in fine-needle aspirates of thyroid nodules

Overlapping morphologic patterns that may be observed in goiter, follicular adenoma, and papillary carcinoma can limit the cytologic evaluation of the thyroid gland. In an attempt to develop a useful adjunctive test, the immunocytochemical reactivity of HBME-1, carcinoma antigen 19-9 (CA 19-9), and CD-15 (Leu-M1) was tested on 59 cell block preparations from fine-needle aspirations of the thyroid gland. HBME-1 monoclonal antibody was reactive in all 21 papillary carcinomas, in 4 of 18 adenomas, and in 5 of 20 goiters. CA 19-9 was identified in 13 of 21 carcinomas, 1 goiter, but none of the adenomas. CD-15 was present in 15 of 21 carcinomas, 1 goiter, and 1 adenoma. We conclude that HBME-1 is a sensitive marker of papillary thyroid carcinoma. CD-15 and CA 19-9 are less sensitive but more specific. This panel can be useful to help classify morphologically equivocal lesions. As with all immunocytochemical testing, caution must be exercised in the interpretation of results, and correlation made with morphologic and clinical data.

Keratin subsets in papillary and follicular thyroid lesions. A paraffin section analysis with diagnostic implications

Previous studies indicate that keratins 7, 8 and 18 are present in all thyroid papillary and follicular lesions, but the distribution of other keratins has been incompletely characterized. The profile of individual keratin (K) polypeptides was evaluated immunohistochemically in over 200 non-neoplastic and neoplastic thyroid papillary and follicular lesions. Monoclonal antibodies to K19, K17, K16, K5/6 and K10 were applied in paraffin sections of formaldehyde-fixed tissue. K19 was present variably, often only focally in goitres, and was present only sporadically in papillary hyperplasia. However, K19 was strongly and uniformly expressed in virtually all papillary carcinomas, indicating differential diagnostic usefulness in differentiating papillary hyperplasia and papillary carcinoma. About half of the follicular carcinomas (defined as tumours strictly excluding the follicular variant of papillary carcinoma) were also strongly K19-positive, suggesting that K19 patterns are not reliable in differentiating papillary and follicular carcinoma. K17 and K5/6 were present in cysts and squamous metaplasia of goitres, and focally in papillary but only exceptionally in follicular carcinoma in areas of squamous differentiation and tumour cells in desmoplastic stroma. K16 in turn was present only focally in well-developed squamous metaplasia in goitres but was not found in differentiated thyroid carcinomas. K10, a high-molecular-weight keratin typical of epidermal differentiation, was identified neither in non-neoplastic nor in neoplastic differentiated thyroid lesions, including squamous metaplasia. These results indicate that papillary carcinomas differ from other differentiated thyroid tumours in their varying, usually focal, expression of stratified epithelial keratins that are partly but not exclusively related to squamous differentiation in such lesions. However, papillary carcinomas do not express truly epidermally restricted keratins; their previously described reactivity with polyclonal "epidermal keratin" antibodies most probably results from the reactivity of such antibodies with K19.

Hyaline globules in renal cell carcinomas and oncocytomas

Hyaline globules (extracellular collections of amorphous material) are identified in 10 of 59 renal cell carcinomas (RCC) and in 2 of 9 oncocytomas. Immunohistochemical characterization of these PAS-positive structures revealed the presence of basement membrane material in most cases. Collagen type IV and laminin were the predominant constituents, whereas fibronectin was detected only occasionally. Electron microscopic examination of the globules showed concentric multilayered accumulations of basement membrane material. No such structures were recognized in 8 renal pelvic transitional cell carcinomas nor in 2 metanephric adenomas. RCC associated hyaline globules were always negative for alpha1-antitrypsin (AAT), alpha-fetoprotein (AFP), amyloid A, cytokeratin, vimentin, or lysozyme. These features differ from those of the hyaline globules previously described in other malignant neoplasms, notably AAT-positive hyaline globules in ovarian tumors, and AFP-positive globules in yolk sac tumors. Identification and immunohistochemical characterization of hyaline globules in metastases may be helpful in determining the origin of occult primary tumors.

Ki-67 antigen staining as an adjunct to identifying cervical intraepithelial neoplasia

The purpose of the work was to determine the feasibility and predictive value of Ki-67 immunostaining of cervical cytology and the detection of cervical dysplasia. Air-dried cervical smears were stained with MIB-1 antibody to identify the Ki-67 antigen. Nuclear decoration in abnormal squamous nuclei determined immunoreactivity. One hundred twenty-four nonpregnant patients underwent colposcopy and directed biopsies for abnormal cytology. Sensitivity (0.89), specificity (0.65), positive predictive value (0.60), and negative predictive value (0.91) were found for Ki-67 immunostaining in detection of high-grade cervical intraepithelial neoplasia (CIN) in 124 patients and positive Ki-67 staining was a significant predictor of high-grade CIN in both univariate (odds ratio 15.5 (95% CI 5.5-43.8) and multivariable (odds ratio 21.5 (95% CI 5.0-92.0) analysis. In 101 patients with ASCUS and LGSIL, Ki-67 immunostaining demonstrated the following in detection of high-grade CIN: sensitivity (0.96), specificity (0.67), positive predictive value (0.49), and negative predictive value (0.98). Ki-67 immunostaining of cervical cytology is a predictor of significant cervical pathology with high sensitivity and negative predictive value. Ki-67 immunostaining of cervical cytology may represent a new and cost-effective triage tool for patients with minor abnormalities on cytology.

Quantitative image analysis of MIB-1 reactivity in inflammatory, hyperplastic, and neoplastic endocervical lesions

The diagnosis of endocervical neoplasia can be difficult as it is sometimes mimicked by proliferative or reactive glands. MIB-1 is a proliferation marker that can aid in the diagnosis of squamous intraepithelial lesions (SIL) of the cervix and vulva, but its potential value in the diagnosis of endocervical lesions has not been fully explored. Ten formalin-fixed, paraffin-embedded cases of each of the following were obtained: morphologically normal endocervical glands from patients with cervical SIL, endocervicitis, microglandular hyperplasia (MGH), and endocervical adenocarcinomas (eight in situ, two invasive). Microwave unmasking of antigens was performed prior to immunohistochemical staining for MIB-1 using the avidin/biotin peroxidase method. Labeling indexes were calculated for 34 specimens (10 adenocarcinoma. 8 each of the other diagnoses) using image analysis (Samba 4000). There was diffuse MIB-1 reactivity in adenocarcinoma (labeling index 57-96%, mean 80%), minimal focal reactivity in normal glands underlying SIL (labeling index 0.8-4.3%, mean 2.4%), moderate spotty reactivity in MGH (labeling index 2.9-18.4%, mean 8.5%), and minimal to focally diffuse reactivity in endocervicitis (labeling index 1.0-13.3%, mean 5.7%). These data indicate that the percentage and distribution of MIB-1-reactive endocervical cells can be of diagnostic utility in distinguishing neoplastic glands from those of endocervicitis and MGH.

Immunohistochemical localization of p21(WAF1/CIP1) in normal, hyperplastic, and neoplastic uterine tissues

p21WAF1/CIP1 is a nuclear protein that binds to cyclin-dependent kinase complexes (CDKs) and inhibits the activity of multiple kinases. These CDKs are involved in the regulation of cell cycle progression at several checkpoints. In this study, the authors have analyzed by immunohistochemistry the expression of p21WAF1/CIP1 in normal uterine tissues, 12 endometrial hyperplasias, 17 endocervical adenocarcinomas, and 31 endometrial adenocarcinomas. In addition, a group of 10 leiomyomas and 10 uterine leiomyosarcomas were also stained. To evaluate cell proliferation, the monoclonal antibody Ki-67 was used in all of the available cases. Terminally differentiated epithelial endocervical and endometrial cells showed variable expression of p21WAF1/CIP1, whereas the endometrial hyperplasias, and endocervical and endometrial adenocarcinomas showed decreased expression or were negative. All of the cases of cervical squamous dysplasia were positive. Normal smooth muscle cells and 50% of leiomyomas were negative, whereas all leiomyosarcomas showed expression of p21WAF1/CIP1. These results indicate that p21WAF1/CIP1 contributes to differentiation in normal endometrial and endocervical glands. The decreased expression of p21WAF1/CIP1 in endometrial hyperplasias and carcinomas may be important in the process of neoplastic transformation. The role of certain CDK inhibitors, such as p21WAF1/CIP1, is different in epithelial and mesenchymal tumorigenesis in the uterus.

Detection of JC virus DNA sequence and expression of the viral oncoprotein, tumor antigen, in brain of immunocompetent patient with oligoastrocytoma

We describe molecular and clinical findings in an immunocompetent patient with an oligoastrocytoma and the concomitant presence of the human papovavirus, JC virus (JCV), which is the etiologic agent of the subacute, debilitating demyelinating disease, progressive multifocal leukoencephalopathy. Histologic review revealed a glial neoplasm consisting primarily of a moderately cellular oligodendroglioma with distinct areas of a fibrillary astrocytoma. Immunohistochemical analysis revealed nuclear staining of tumor cells with antibodies against the viral oncoprotein [tumor antigen (T antigen)], the proliferation marker (Ki67), and the cellular proliferation regulator (p53). Using primers specific to the JCV control region, PCR yielded amplified DNA that was identical to the control region of the Mad-4 strain of the virus. PCR analysis demonstrated the presence of the genome for the viral oncoprotein, T antigen, and results from primer extension studies revealed synthesis of the viral early RNA for T antigen in the tumor tissues. The presence of viral T antigen in the tumor tissue was further demonstrated by immunoblot assay. To our knowledge, this is the first report of the presence of JCV DNA, RNA, and T antigen in tissue in which viral T antigen is localized to tumor cell nuclei and suggests the possible association of JCV with some glial neoplasms.

Protocol for immunocytochemical detection of SIL in cervical smears using MIB-1 antibody to Ki-67 [corrected]

Ki-67 nuclear antigen is expressed in upper epithelial levels of intraepithelial neoplasia of the cervix and vulva, variably in condyloma, and in basal and parabasal cells of normal squamous mucosa in histologic preparations. The application of antibodies to Ki-67 as a marker of squamous intraepithelial lesions in cervical smears was explored using either air-dried, acetone-fixed cervical smears obtained from 106 consenting patients or a single slide from archival two-slide cases of squamous intraepithelial lesions MIB-1 monoclonal antibody to Ki-67 was tested using two immunocytochemical techniques. In one set of smears, avidin-biotin peroxidase was used for detection and diaminobenzidine with H2O2 as the chromogen. Some specimens were incubated with 0.3% H2O2 and phosphate buffered saline for blockade of endogenous peroxidase. Alternatively, other air-dried smears were stained using alkaline phosphatase antialkaline phosphatase for detection and new fuchsin as the chromogen. Nuclear staining in squamous intraepithelial lesions was identified in air-dried smears using all of the above methods. Slides stained with avidin-biotin peroxidase and blocked with 0.3% H2O2 and phosphate buffered saline showed less background staining from neutrophils and erythrocytes compared with those without blocking. Slides stained using alkaline phosphatase antialkaline phosphatase showed excessive cytoplasmic staining of endocervical cells, making intraepithelial difficult. No nuclear staining of squamous intraepithelial lesions was observed in destained archival smears. Air-dried smears blocked with 0.3% H2O2 and phosphate buffered saline, incubated with MIB-1, and stained using avidin-biotin peroxidase gave the best results for identification of Ki-67 expression in squamous intraepithelial lesions.

Metanephric adenoma: histology, immunophenotype, cytogenetics, ultrastructure

Metanephric adenoma is a rare renal neoplasm that is histologically and clinically unique. We found this neoplasm in a 54-year-old woman. This large, well-circumscribed, solid, tan tumor showed histologic similarity to developing metanephric tubular epithelium. It is composed of uniformly small epithelial cells forming tubules. Immunohistochemical evaluation showed that the tumor cells express proximal tubule marker URO-2 and wt1 gene protein product, commonly expressed in Wilms' tumors. Ultrastructurally, the epithelial cells were characterized by the presence of cilia on the luminal side and were resting on an abundant basement membrane. Cytogenetic analysis showed normal female karyotype. The unique features of metanephric adenoma should be clinically and pathologically recognized because of its invariably benign course.

Immunohistochemical staining for proliferating cell nuclear antigen, BCL2, and Ki-67 in vulvar tissues

Dysplastic cells of the uterine cervix commonly react with antibodies to Ki-67 and proliferating cell nuclear antigen (PCNA), but less is known regarding the reactivity of vulvar lesions. Paraffin-embedded slides from seven normal vulvar tissues, eight cases of lichen sclerosus, and 10 cases of vulvar intraepithelial neoplasia (VIN), many with associated condylomatous changes, were immunostained with antibodies to PCNA, Ki-67, and BCL2 protein, a protooncogene product normally expressed only in basal cells of squamous epithelium. Ki-67 and PCNA nuclear staining was largely restricted to basal and parabasal cells in normal tissues and lichen sclerosus. Focal midepithelial staining with PCNA and Ki-67 was seen in one case of lichen sclerosus; a vulvar biopsy of this patient 1 year later was negative for VIN. Both antibodies stained dysplastic cells at higher epithelial levels in VIN, but Ki-67 was more consistently reactive and showed a sharper distinction from adjacent histologically uninvolved epithelium compared to PCNA. Condylomatous changes were variably stained. The pattern of BCL2 staining was identical in normal vulva, lichen sclerosus, and VIN, but BCL2 decorated occasional mitotic figures in VIN. Overall, Ki-67 was a better marker of vulvar dysplasia than PCNA.