Intratumoral Heterogeneity of ERBB2 Amplification and HER2 Expression in Micropapillary Urothelial Carcinoma

Abstract

Micropapillary urothelial carcinoma (MPUC) is a rare but an aggressive variant of urothelial carcinoma. MPUC has been shown to commonly exhibit ERBB2 amplification and HER2 protein overexpression, but the frequency and distribution of these findings within micropapillary (MP) and not otherwise specified (NOS) components of tumors have not been addressed. Therefore, we evaluated ERBB2 amplification and HER2 expression in 43 MPUC cases by fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC). Of the 35 tumors containing both MP and NOS components, ERBB2 amplification was present in both the MP and NOS components of 12 tumors (34.3%), in only the MP component of 11 tumors (31.4%), and exclusively in the NOS component of 4 tumors (11.4%). HER2 protein overexpression was significantly more commonly present in the MP component compared to the NOS component within the same tumor (68.6% vs. 34.3%, p=0.012). Overall, there was a moderately positive correlation between HER2 protein expression and ERBB2 amplification in both MP (rho=0.59, p<0.001) and NOS (rho=0.70, p<0.001) components. All MP/NOS areas with IHC score 3+ and none of MP/NOS areas with IHC score 0 were associated with ERBB2 amplification. We conclude that ERBB2 amplification and HER2 overexpression are preferentially but not exclusively identified in the MP component compared to the NOS component within the same tumor. Our findings identify the presence of intratumoral heterogeneity of ERBB2 amplification and HER2 expression in MPUC and provide grounds for further investigation into the mechanisms underlying the development of MPUC.

INTRODUCTION

Micropapillary urothelial carcinoma (MPUC) is an aggressive rare variant of bladder cancer and has poor long-term survival [1–5]. MPUC also has a high propensity for angiolymphatic invasion and lymph node involvement [3, 4]. The molecular pathogenesis of MPUC is poorly understood and optimal treatment strategies for MPUC patients remain controversial. Some investigators reported a poor response to intravesical BCG therapy as well as to neoadjuvant chemotherapy in MPUC patients and therefore have advocated early cystectomy, while others suggested that intravesical BCG therapy as well as neoadjuvant chemotherapy should be offered to all eligible MPUC patients [4, 6– 8]. Therefore, there is a need to identify novel therapeutic targets and treatment options for this aggressive disease.

Human epidermal growth factor receptor-2 (ERBB2, HER2) is a type-1 transmembrane growth factor receptor that functions to activate intracellular mitogenic signaling pathways in response to extracellular growth signals [9]. Treatment with monoclonal antibodies targeting the HER2 extracellular domain is currently a standard of care in HER2 positive breast, gastric and gastroesophageal junction cancers patients [10–14]. HER2 protein overexpression and ERBB2 amplification have also been reported in several other cancers though the therapeutic relevance of such alterations remains to be defined [9]. Overall, bladder cancer has low frequency of ERBB2 amplification (6.2% to 8.8%) [15–17], but higher rates of ERBB2 amplification have been reported in MPUC [18–21].

Histologically, the majority of MPUC tumors is of mixed histology with a component of classic urothelial carcinoma, “not otherwise specified” (NOS) adjacent to or intermixed with the micropapillary (MP) component. The status of ERBB2 amplification and HER2 expression in the NOS component in relation to the MP component of MPUC has not been addressed previously. To explore whether ERBB2 expression is variable in bladder tumors containing both MP and NOS components (intratumoral heterogeneity); we evaluated ERBB2 amplification as well as HER2 protein expression in both the MP and NOS components of the same tumors, in cases where the two co-exist.

MATERIALS AND METHODS

Patient samples

Following IRB approval, we identified 43 MPUC cases with tissue available at our institution. All available slides were reviewed by two urological pathologists (Huang, Al-Ahmadie) to confirm diagnosis and representative blocks were selected for further analyses. All tumors were subjected to fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC) to detect ERBB2 amplification and HER2 protein expression, respectively. Of these 43 cases, 35 cases also had an NOS component within the same tumor that was evaluable by both FISH and IHC.

Fluorescence in situ hybridization (FISH)

FISH analysis was performed using a 2-color ERBB2/Cen17 probe. The probe mix consisted of BAC clones containing the ERBB2 gene (RP11-94L15, RP11-62N23 and CTD-3211L18; labeled with Red dUTP) and a centromeric repeat plasmid for Chr17 (clone p17H8p; labeled with Green dUTP) served as the control. For each case, the corresponding H&E slide was used to identify and mark regions of MP and/or NOS. The entire section was first scanned under 63× objective magnification to assess the signal pattern and possible intratumoral heterogeneity. Non-tumor area(s) or normal tissue including stromal cells or infiltrating lymphocytes were also scanned and served as the internal control to assess the quality of hybridization. Following the initial scan, representative regions (regions that served as example(s) for the observed signal pattern(s) in the entire tissue section) within MP and/or NOS components were selected and at least five images per region captured (each image was a compressed stack of 12 z-section images taken at 0.5 µm intervals). Signal counts were performed on captured images and for each marked histologic entity or distinct region, a minimum of 50 discrete nuclei scored (range 50–200 nuclei). The ASCO/CAP 2013 guidelines for invasive breast cancer were followed to assess amplification status [22]. Amplification was defined by an Her2/Cep17 ratio of ≥2.0 or > 6 copies of the gene. Additionally, amplification was categorized as homogeneously staining region (HSR) when observed as large clustered signal. HSRs correspond to linearly integrated chromosomal/DNA segments that appear homogeneous and stain uniformly by conventional cytogenetic methods.

Immunohistochemistry (IHC)

Immunohistochemical staining for HER2 (Clone 4B5; Cat# 790-2991; Ventana, Tucson, AZ) was conducted on BenchMark XT automated stainers using NexES Version 12.10 software and iVIEW DAB Detection Kit (Ventana Medical Systems) with appropriate external and internal positive and negative controls. HER2 protein expression was scored by two urological pathologists (Huang, Al-Ahmadie) according to the ASCO/CAP 2013 guidelines for invasive breast cancer [22]. HER2 overexpression was considered with IHC score of 2+ and 3+. MP and NOS areas within the same tumor were scored independently and the results were compared.

Statistical analysis

All data was analyzed using Stata™ v12.0 statistical analyses software (Stata Corporation, College Station, TX). The McNemar's test was used to compare paired proportions between MP and NOS components. The association between ERBB2 amplification and HER2 protein expression was evaluated using Spearman’s rank correlation coefficients. Cox proportional hazards regression was used to identify significant prognostic factors for overall survival. Kaplan-Meier survival plots were created to demonstrate the ability of the pathologic variables, amount of MP component, ERBB2 amplification and HER2 expression in the MP component to predict overall survival. Statistical significance in this study was set as p ≤ 0.05.

RESULTS

Detailed demographic and pathologic information for our cohort are shown in Supplementary Table 1. Of 43 MPUC cases, 35 were of mixed histology with both MP and NOS components whereas 8 tumors contained only MP features (Table 1). The amount of MP component was variable and ranged from 5% to 100% (average 46%, median 50%, more details are available in Supplementary Table 1).

Table 1.

ERBB2 Amplification status and HER2 expression in tumors with Micropapillary (MP) and not otherwise specified (NOS) components as well as in tumors with MP features only.

Test		Tumors with MP and NOS components (N=35)		Tumors with MP component only (N=8)
		MP component	NOS component
ERBB2 Amplification	Present	23 (67.6%)	16 (45.7%)	5 (62.5%)
	HSR	18 (78.3%)	13 (81.3%)	3 (60.0%)
	High Polysomy	5 (21.7%)	3 (18.7%)	2 (40.0%)
	Absent	11 (32.4%)a	19 (54.3%)	3 (37.5%)
HER2 IHC	0	1 (2.8%)	6 (17.1%)	0 (0.0%)
	1+	10 (28.6%)	17 (48.7%)	2 (25.0%)
	2+	14 (40.0%)	6 (17.1%)	4 (50.0%)
	3+	10 (28.6%)	6 (17.1%)	2 (25.0%)

^aMP component was absent on the FISH slides for 1 case

The status of ERBB2 amplification in MP and NOS components is summarized in Table 1 (detailed description in Supplementary Table 1). Of the 35 tumors containing both MP and NOS components, ERBB2 amplification was detected in 27 tumors (77.1%). ERBB2 amplification was present in both the MP and NOS components of 12 tumors (34.3%), in only the MP component of 11 tumors (31.4%), and in only the NOS component of 4 tumors (11.4%). Eight tumors (22.9%) lacked ERBB2 amplification in either component. In one case (case no. 5), the MP component was absent on the FISH slide (Table 1). Figures 1, 2, 3 and 4 illustrate examples of different patterns of ERBB2 amplification and HER2 expression in MP and NOS components of the same tumors. Overall, ERBB2 amplification was more commonly present in the MP component compared to the NOS component within the same tumor [23 of 34 tumors (67.6%) vs. 15 of 34 tumors (44.1%), p=0.057, Figure 5].

Figure 1.

An example of HER2 expression score 3+ in the micropapillary component (MP) with ERBB2 amplification (HSR type). There is no HER2 expression in the NOS component (score 0) and no ERBB2 amplification. A and B represent panoramic view of MP and NOS components by hematoxylin and eosin (H&E) and HER2 immunohistochemistry staining, respectively. Upper panel 20× magnification and lower panel 40× magnification.

Figure 2.

An example of HER2 expression score 3+ in the NOS component with ERBB2 amplification (HSR type). There is 1+ HER2 expression in the MP component and no ERBB2 amplification. 100× magnification.

Figure 3.

An example of HER2 expression score 3+ and ERBB2 amplification (HSR type) in both the NOS and MP components. 100× magnification.

Figure 4.

An example HER2 2+ IHC expression in MP component associated with ERBB2 amplification, the high polysomy type. This tumor component has an average of 8 copies of ERBB2 and 5 copies of Chr17 (Her2/Cen17 ratio of 1.6). 100× magnification.

Figure 5.

ERBB2 Amplification in classic “not otherwise specified (NOS)” bladder cancer, NOS component of MPUC and MP component of MPUC.

In the histogram, the prevalence of ERBB2 amplification in NOS bladder cancer was derived from TCGA dataset [Reference 17 in the manuscript, 6.3% (8/127)]. ERBB2 amplification was significantly more common in MPUC, both within the MP and NOS components, as compared to pure NOS bladder tumors

ERBB2 amplification was in the form of HSR in the majority of cases. This HSR pattern was present in the MP component in 18 of 23 tumors (78.3%) and in the NOS component in 13 of 16 tumors (81.3%). Examples in Figures 1, 2 and 3 are representative of HSR type of amplification. The other type of ERBB2 amplification, higher polysomy pattern (≥6 copies of ERBB2), was present in MP components in 5 of 23 tumors (21.7%) and in the NOS components of 3 of 16 tumors (18.7%). An example of polysomy pattern ERBB2 amplification is depicted in Figure 4. In the 12 tumors with ERBB2 amplification in both MP and NOS components, concordant HSR amplification was observed in 8 tumors. Interestingly, 4 tumors displayed different patterns of ERBB2 amplification in the MP and NOS components: in 2 tumors, high polysomy type amplification was observed in the MP component and HSR in the NOS component and the reverse occurred in 2 cases. In summary, our results indicate that HSR-type amplification of ERBB2 is more prevalent in MPUC. In this cohort, the Her2/Cep17 ratio was higher in cases with the HSR compared to high polysomy type amplification (5.3 vs. 1.7). Considering the NOS component of MPUC, the rate of ERBB2 amplification was significantly higher than that of urothelial carcinoma NOS without MP component as previously reported [15–17] (p<0.001, Figure 5).

IHC results of HER2 expression in MP and NOS components are shown in Table 2. Of 35 tumors containing both MP and NOS components, HER2 overexpression (IHC score 2+ or 3+) was detected in 28 tumors (80.0%), and was identified in both the MP and NOS components of 8 tumors (22.9%), in only the MP component of 16 tumors (45.7%), and in only the NOS component of 4 tumors (11.4%). HER2 overexpression was significantly more prevalent in the MP component compared to the NOS component within the same tumor [24 of 35 tumors (68.6%) vs. 12 of 35 tumors (34.3%), p=0.012].

Table 2.

HER2 protein expression correlation with ERBB2 amplification in MP and NOS areas within the same bladder tumor (N= 35)

HER2 IHC Scores	ERBB2 Amplification
	MP	NOS
0	0% (0/1)	0% (0/6)
1+	33.3% (3/9)	29.4% (5/17)
2+	71.4% (10/14)	83.3% (5/6)
3+	100% (10/10)	100% (6/6)

For the 35 tumors containing both MP and NOS components, the correlation between ERBB2 amplification by FISH and HER2 expression by IHC is shown in Table 2. Overall, there was a moderately positive correlation between HER2 protein expression and ERBB2 amplification in both MP (rho=0.59, p<0.001) and NOS (rho=0.70, p<0.001) components. An IHC score of 3+ was assigned in 10 MP areas and in 6 NOS areas, all of which harbored ERBB2 amplification. In tumors with an IHC score of 2+, ERBB2 amplification was noted in 71.4% (10 of 14) of MP and 83.3% (5 of 6) of NOS areas respectively. In tumors with an IHC score of 1+, ERBB2 amplification was noted in 33.3% (3 of 9) of MP and 29.4% (5 of 17) of NOS components respectively. A HER2 IHC score of 0 was identified in 1 MP and 6 NOS areas, none of which harbored ERBB2 amplification.

In the 8 cases with MP histology only, 3 tumors had HSR type ERBB2 amplification; 2 of which had HER2 IHC score 3+ and 1 tumor had 2+ score. Two tumors displayed high polysomy type ERBB2 amplification, with 1 tumor characterized as HER2 IHC score 2+ and 1 tumor 1+. Of the 3 tumors without ERBB2 amplification, 2 had HER2 IHC score 2+ and 1 tumor had 1+.

Overall, HER2 expression was variable in extent and intensity in both the MP and NOS components and ranged from 0 to 100% of tumor cells, indicating the presence of heterogeneity not only between but also within the individual components. There was no correlation between amount of MP component with ERBB2 amplification (rho = 0.19, p=0.24) or HER2 expression (rho=0.04, p=0.78).

To assess the impact of the MP component on outcome, we performed cox regression analysis and found that amount of MP component, ERBB2 amplification and HER2 protein expression in the MP component were not predictive of overall survival (p = 0.92, 0.64 and 0.96 respectively) in this cohort of 43 patients. Of note, tumor stage and lymph node status were significant predictors of overall survival (p = 0.01 and <0.001 respectively). Supplementary Figure 1 shows Kaplan-Meier overall survival plots as a function of tumor stage, lymph node status, amount of MP component, ERBB2 amplification and HER2 protein expression in the MP component.

DISCUSSION

Morphologic heterogeneity is a common feature of urothelial carcinoma of the bladder. Over 10 variant histologies are now recognized and most often, these variant histologies co-exist with classic urothelial carcinoma, not otherwise specified (NOS), within the same primary tumor [23]. MPUC is a rare but highly aggressive variant histology, the molecular pathogenesis of which is poorly understood.

In this study, we examined 43 MPUC cases for evidence of ERBB2 amplification and HER2 overexpression. We identified ERBB2 amplification and HER2 overexpression in nearly two-thirds of cases, which is significantly higher than what has been reported in classic urothelial carcinoma [15– 17, 24]. The high rate of ERBB2 amplification in MPUC suggests that agents that inhibit HER2 signaling may be a potential targeted therapeutic approach for this histologic subtype [10, 11, 22]. Recently, Choudhury et al [25] reported that treatment of urothelial tumors containing ERBB2 or ERBB3 alterations with afatinib (an oral irreversible inhibitor of the ERBB family) significantly increased progression free survival duration in patients with platinum-refractory urothelial carcinoma.

Notably, this is the first study to assess ERBB2 amplification in NOS component of MPUC. It is known that in the majority of MPUC there is an identifiable component, however small, of classic urothelial carcinoma. In this study, we showed that ERBB2 amplification and HER2 overexpression are more frequently present in the MP component than in the NOS component within the same tumor (67.6% vs. 44.1%, p=0.057; 68.6% vs. 34.3%, p=0.012 respectively). Moreover, the NOS component associated with MPUC harbored ERBB2 amplification at a much higher frequency than previously reported in urothelial carcinoma without micropapillary morphology (Figure 5) [15–17]. This suggests that the NOS component of MPUC is potentially biologically different from UC, NOS or UC not associated with micropapillary variant histology. Further, this indicates that ERBB2 amplification is a common genetic event in both the MP and NOS components although it is more frequently associated with the former.

Overall, there was a moderately positive correlation between HER2 protein expression and ERBB2 amplification in both MP and NOS components.. All 18 MP/NOS areas with an IHC score 3+ were associated with gene amplification (100%). Moreover, the majority of MP/NOS regions with IHC score 2+ (17 of 24, 70.8%) and even some MP/NOS areas with IHC score 1+ harbored ERBB2 amplification (9 of 28, 32.1%). These findings differ from breast cancer where an IHC score of 1+ is considered negative for HER2 overexpression and is not investigated further for ERBB2 amplification. In our study, only MP/NOS areas with IHC scores of 0 were not associated with ERBB2 amplification. The majority of MPUC in this cohort were associated with HSR-type of ERBB2 amplification and there was no correlation between the type of amplification and clinical outcome. This is similar to breast cancer where the majority ERBB2 amplification was of the HSR type and where there is no reported correlation between the type of ERBB2 amplification and response to treatment with ERBB2 targeted therapy with trastuzumab [26, 27].

Prior publications [18, 19, 21, 28] on MPUC have reported ERBB2 amplification/HER2 expression incidence rates ranging from 15% to 74%. However, prior studies did not address the distribution or intratumoral heterogeneity of such alterations. Therefore, this study is the first to report on the presence of intratumoral heterogeneity in ERBB2 amplification and HER2 overexpression in MPUC in relation to morphologic heterogeneity. It has been recently suggested that MPUC also harbors higher rates of activating ERBB2 mutations [29], but this observation requires further independent validation. Additionally, ERBB2 was one of the top 10 upstream regulators enriched in MPUC in a recent report on the gene expression profiles of MPUC [30], further supporting a possible role for ERBB2 activation in the pathogenesis of MPUC. One of the limitations of our study is the small sample size used to assess ERBB2 status in MPUC. A large multi-institutional study incorporating multi-platform genomic analysis of MPUC is required to better delineate the prevalence, clonality, and intra- and inter-tumoral heterogeneity of ERBB2 amplification and HER2 overexpression in this rare variant histologic subtype.

In conclusion, ERBB2 amplification and HER2 overexpression are preferentially but not exclusively identified within the MP component compared to the NOS component of MPUC. Our findings identify the presence of intratumoral heterogeneity of ERBB2 amplification and HER2 expression in MPUC and provide grounds for further investigation into the mechanisms underlying the development of MPUC.

Highlights.

• ERBB2 amplification is very common in micropapillary bladder cancer.

• There is intratumoral heterogeneity of ERBB2 amplification in micropapillary bladder cancer.

• ERBB2 amplification is more common in micropapillary component of tumors with mixed histology.