The research for markers of therapeutic response could help us to individualize therapies in bladder cancer. In this work the authors studied a population of 71 patients submitted to neodjuvant chemotherapy and posterior radical cystectomy, and divided the on responders and non-responders based on final pathological report. By DNA-based technology, the authors concluded that ERBB2 mutations are a prognostic factor of good response to cisplatin base chemotherapy. Although this constitutes a very promising marker for selecting patients for NAC, further validation of their results is warranted in order to establish the reliability of their results.
A pathologic complete response to neoadjuvant chemotherapy (NAC) containing platinum is associated with superior clinical outcome in patients with muscle-invasive bladder cancer (MIBC)[1], [2], and [3]. No molecular markers or baseline clinical characteristics that can predict the response to NAC are clinically validated. In the present study, we found an unexpected association between mutations in the erb-b2 receptor tyrosine kinase 2 (ERBB2) gene, also known asHER2, and a complete response to chemotherapy.
We collected pre- and postchemotherapy specimens from 110 prospectively registered MIBC patients treated with NAC. Good-quality DNA from the pretreatment transurethral resection specimens was available for 94 patients. In this cohort, we identified 38 pathologic complete responders (ypT0N0), 23 partial responders (ranging from a minor response to a near-complete response), and 33 nonresponders (higher than ypT2) to NAC. Full study methods are described in the supplement. No significant differences in baseline clinical characteristics were identified between the three groups (Supplementary Table 1). Complete responders had a superior recurrence-free and cancer-specific survival compared with nonresponders, whereas the partial responders had an in-between survival (p < 0.001) (Supplementary Fig. 1).
We sequenced 178 cancer-associated genes (Supplementary Table 2) on pretreatment tumor DNA from 16 complete responders and 16 nonresponders (discovery cohort). Genes with a differential mutation frequency in complete responders compared with nonresponders were identified by contrasting analysis ( Fig. 1 A).ERBB2had the highest enrichment for mutations in complete responders ( Fig. 1 A and 1B). We therefore tested the association betweenERBB2mutations and chemotherapy sensitivity in a validation cohort consisting of the remaining 22 complete responders and 17 nonresponders to NAC in our patient series. We identified another fiveERBB2missense mutations in four complete responders and none in the nonresponders. Taken together, we identifiedERBB2missense mutations in 9 of the 38 complete responders (24%) and in none of the 33 nonresponders to NAC (p = 0.003) ( Fig. 2 ).
Fig. 1 Gene enrichment analysis of mutated genes in complete responders and nonresponders. (A) Volcano plot of effect size (log-odds ratio) and significance (−log10pvalue) of the 25 genes mutated in more than two samples. Mutated genes enriched in complete responders are labeled green; mutated genes enriched in nonresponders are labeled red. (B) Pyramid plot showing the number of mutated samples in 16 complete responders in green and the number of mutated samples in 16 nonresponders in red. (C) Plot showing the distribution ofERBB2missense mutations identified in this study by target enrichment next-generation sequencing (dark blue circles) or Sanger sequencing (orange circles).ERBB2missense mutations cluster at the S310 position in the extracellular domain and in the tyrosine kinase domain. (D) Graph showing thatERBB2missense mutations are significantly enriched in responders and significantly depleted in nonresponders to neoadjuvant chemotherapy compared with the unselected Cancer Genome Atlas urothelial bladder cancer cohort (*p < 0.05). (E, F) Sequencing results from twoERCC2-mutant nonresponders showing theERCC2missense mutation (E86Q and S44L) in DNA isolated from the pretreatment transurethral resection and post-treatment tissue but not in the germline DNA. This demonstrates that these somaticERCC2mutations were not counterselected during chemotherapy. NGS = next-generation sequencing; TCGA = The Cancer Genome Atlas; TKD = tyrosine kinase domain; TM = transmembrane domain; TUR = transurethral resection.
Fig. 2 OncoPrint showingERBB2missense mutations,ERBB2amplifications, andERCC2missense mutations in the 38 complete responders and 33 nonresponders to neoadjuvant chemotherapy in this study. Individual patients are represented as columns. N/A = not available; pCR = pathologic complete response.
Five of the 10 identifiedERBB2missense mutations cluster at amino acid 310 in the extracellular domain ( Fig. 1 C; Supplementary Table 3). The S310 position is also a mutational hotspot in The Cancer Genome Atlas (TCGA) urothelial bladder cancer cohort because approximately 40% of allERBB2missense mutations cluster at this position (TCGA Data Portal; http://cancergenome.nih.gov ). Previous functional studies onERBB2mutations have shown that the S310F, D769H, and V842I variants identified here are activating mutations that support cellular transformation[4] and [5]. To our knowledge, the R678L and V777M mutations have not been functionally characterized. However, a different amino acid substitution at the same V777 position was found to be activating [5] . It was recently reported that micropapillary urothelial carcinomas (UCs) carry a high frequency (40%) of activating extracellular domainERBB2mutations [6] . We therefore reviewed all ourERBB2-mutant UCs. None of these had micropapillary variant histology.
We confirmed the association betweenERBB2missense mutations and responsiveness to platinum-containing chemotherapy in the recently published MIBC data set of Van Allen et al [7] . In this external validation cohort, all three patients with anERBB2missense mutation responded to NAC [7] . In addition, they reported a fourth complete responder who had anERBB2mutation with an allelic fraction of 0.04.
ERBB2missense mutations are significantly enriched in the chemotherapy responders from our cohort and the Van Allen et al cohort [7] compared with the unselected TCGA cohort (8% in TCGA;p = 0.02) ( Fig. 1 D). Conversely, we found thatERBB2missense mutations were significantly depleted in the nonresponder cohorts compared with the unselected TCGA cohort (p = 0.02) ( Fig. 1 D).
Having found an association betweenERBB2missense mutations and platinum response, we next tested the association ofERBB2amplification with platinum response.ERBB2amplifications were identified in complete responders as well as in nonresponders and were not associated with response to NAC (p = 0.52) ( Fig. 2 ).ERBB2amplification was always accompanied by protein overexpression. In four patients, amplification ofERBB2was found in combination with a missense mutation inERBB2. Strikingly, in all of these cases, theERBB2-mutant allele was found amplified, stressing once more the relevance ofERBB2mutations for MIBC oncogenesis.
Van Allen et al recently reported that missense mutations inERCC2, a nucleotide excision repair gene, were selectively present in 9 of 25 MIBC patients with complete response to cisplatin-containing NAC, whereasERCC2missense mutations were absent in 25 nonresponders [7] . In our discovery cohort, we found sixERCC2missense mutations present in four complete responders and in two nonresponders ( Fig. 1 B). Sanger sequencing of the postchemotherapy-resistant tumors of the twoERCC2-mutant nonresponders demonstrated in both cases that theERCC2missense mutation was still present in the postchemotherapy-resistant tumor ( Fig. 1 E and 1F), indicating that the mutation was not counterselected during chemotherapy. In our validation cohort, we identifiedERCC2missense mutations in two tumor samples from 22 complete responders and in none of the 17 nonresponders. In total, 6 somaticERCC2missense mutations were identified in 38 complete responders (16%) and 2 in 33 nonresponders (6%;p = 0.27) ( Fig. 2 ; Supplementary Table 4). Five of the sixERCC2mutations in complete responders were present in patients with wild-type nonamplifiedERBB2( Fig. 2 ).
Finally, we also sequencedERBB2andERCC2in the pretreatment transurethral resection specimens from the remaining group of 23 patients who had responses to NAC ranging from a minor response to a near-complete response. We identified twoERBB2missense mutations in tumors from patients with a partial response. Both patients are still alive >10 yr after NAC without any signs of disease recurrence (Supplementary Table 5). This supports our finding thatERBB2missense mutations are associated with a favorable response to NAC. We identified four tumors (17%) with anERCC2missense mutation in the remaining group of patients. Two of these patients are alive without signs of disease recurrence, and two patients died due to distant recurrences (Supplementary Table 5).
Responses to NAC can also be plotted as complete response versus noncomplete response, as shown in Supplementary Figure 2. In this comparison,ERBB2mutations are strongly associated with response (p = 0.006), whereasERCC2mutations are not.
Our findings indicate thatERBB2missense mutations could assist in selecting patients responding to NAC. Furthermore, these results suggest thatHER2-directed therapies forERBB2mutant bladder cancers are unlikely to replace chemotherapy in the neoadjuvant setting because these tumors have highly favorable responses to NAC containing platinum. However, despite a pathologic complete response to NAC, 3 of the 11 patients withERBB2-mutant MIBC developed a distant recurrence. These patients may benefit from ERBB2 tyrosine kinase inhibitors, alone or in combination with chemotherapy. In contrast to the findings of Van Allen et al, we showed that the presence of anERCC2mutation does not always confer sensitivity to platinum-based therapy.
A possible limitation of this study is the heterogeneity of our cohort. Patients were treated with different platinum-containing chemotherapy regimens: MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin), gemcitabine and cisplatin, or gemcitabine and carboplatin. Although evidence for the benefit of gemcitabine and carboplatin in the neoadjuvant setting in terms of cancer-specific or overall survival is lacking, pathologic complete response rates appear to be similar [8] . Our cohort also contained more advanced cases than most neoadjuvant studies in bladder cancer. However, these patients reflect common clinical practice because many clinics would specifically treat this high-risk patient group with chemotherapy, followed by resection if possible.
In conclusion, we found thatERBB2missense mutations in MIBC are associated with an excellent response to NAC.
Author contributions:Michiel S. van der Heijden had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design:Groenendijk, de Jong, van Rhijn, Bernards, van der Heijden.
Acquisition of data:Groenendijk, van de Putte, Michaut, Schlicker.
Analysis and interpretation of data:Groenendijk, de Jong, van de Putte, Michaut, Velds, Bernards, van der Heijden.
Drafting of the manuscript:Groenendijk, Bernards, van der Heijden.
Critical revision of the manuscript for important intellectual content:de Jong, Michaut, Wessels, Broeks, van Rhijn.
Statistical analysis:Groenendijk, van de Putte, Michaut.
Obtaining funding:van der Heijden, Bernards.
Administrative, technical, or material support:Peters, Nieuwland, Kerkhoven, Broeks.
Supervision:Wessels, van Rhijn, Bernards, van der Heijden.
Other(specify): van de Heuvel (development of methodology).
Financial disclosures:Michiel S. van der Heijden certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:This work was supported by The Netherlands Organization for Scientific Research (to Michiel S. van der Heijden), the European Research Council (grant number 250043 to René Bernards), and the Cancer Genomics Netherlands consortium (to René Bernards).
Acknowledgment statement:The authors wish to acknowledge all patients who contributed tissue for research. We thank the Core Facility for Molecular Pathology & Biobanking of our institute for their assistance. We acknowledge Laura S. Mertens (Department of Urology) and Joyce Sanders (Department of Pathology) for the selection of patients and tissues. We have used data generated by the TCGA Research Network ( http://cancergenome.nih.gov/ ).
A pathologic complete response to neoadjuvant chemotherapy (NAC) containing platinum is associated with superior clinical outcome in patients with muscle-invasive bladder cancer (MIBC)[1], [2], and [3]. No molecular markers or baseline clinical characteristics that can predict the response to NAC are clinically validated. In the present study, we found an unexpected association between mutations in the erb-b2 receptor tyrosine kinase 2 (ERBB2) gene, also known asHER2, and a complete response to chemotherapy.
We collected pre- and postchemotherapy specimens from 110 prospectively registered MIBC patients treated with NAC. Good-quality DNA from the pretreatment transurethral resection specimens was available for 94 patients. In this cohort, we identified 38 pathologic complete responders (ypT0N0), 23 partial responders (ranging from a minor response to a near-complete response), and 33 nonresponders (higher than ypT2) to NAC. Full study methods are described in the supplement. No significant differences in baseline clinical characteristics were identified between the three groups (Supplementary Table 1). Complete responders had a superior recurrence-free and cancer-specific survival compared with nonresponders, whereas the partial responders had an in-between survival (p < 0.001) (Supplementary Fig. 1).
We sequenced 178 cancer-associated genes (Supplementary Table 2) on pretreatment tumor DNA from 16 complete responders and 16 nonresponders (discovery cohort). Genes with a differential mutation frequency in complete responders compared with nonresponders were identified by contrasting analysis ( Fig. 1 A).ERBB2had the highest enrichment for mutations in complete responders ( Fig. 1 A and 1B). We therefore tested the association betweenERBB2mutations and chemotherapy sensitivity in a validation cohort consisting of the remaining 22 complete responders and 17 nonresponders to NAC in our patient series. We identified another fiveERBB2missense mutations in four complete responders and none in the nonresponders. Taken together, we identifiedERBB2missense mutations in 9 of the 38 complete responders (24%) and in none of the 33 nonresponders to NAC (p = 0.003) ( Fig. 2 ).
Fig. 1 Gene enrichment analysis of mutated genes in complete responders and nonresponders. (A) Volcano plot of effect size (log-odds ratio) and significance (−log10pvalue) of the 25 genes mutated in more than two samples. Mutated genes enriched in complete responders are labeled green; mutated genes enriched in nonresponders are labeled red. (B) Pyramid plot showing the number of mutated samples in 16 complete responders in green and the number of mutated samples in 16 nonresponders in red. (C) Plot showing the distribution ofERBB2missense mutations identified in this study by target enrichment next-generation sequencing (dark blue circles) or Sanger sequencing (orange circles).ERBB2missense mutations cluster at the S310 position in the extracellular domain and in the tyrosine kinase domain. (D) Graph showing thatERBB2missense mutations are significantly enriched in responders and significantly depleted in nonresponders to neoadjuvant chemotherapy compared with the unselected Cancer Genome Atlas urothelial bladder cancer cohort (*p < 0.05). (E, F) Sequencing results from twoERCC2-mutant nonresponders showing theERCC2missense mutation (E86Q and S44L) in DNA isolated from the pretreatment transurethral resection and post-treatment tissue but not in the germline DNA. This demonstrates that these somaticERCC2mutations were not counterselected during chemotherapy. NGS = next-generation sequencing; TCGA = The Cancer Genome Atlas; TKD = tyrosine kinase domain; TM = transmembrane domain; TUR = transurethral resection.
Fig. 2 OncoPrint showingERBB2missense mutations,ERBB2amplifications, andERCC2missense mutations in the 38 complete responders and 33 nonresponders to neoadjuvant chemotherapy in this study. Individual patients are represented as columns. N/A = not available; pCR = pathologic complete response.
Five of the 10 identifiedERBB2missense mutations cluster at amino acid 310 in the extracellular domain ( Fig. 1 C; Supplementary Table 3). The S310 position is also a mutational hotspot in The Cancer Genome Atlas (TCGA) urothelial bladder cancer cohort because approximately 40% of allERBB2missense mutations cluster at this position (TCGA Data Portal; http://cancergenome.nih.gov ). Previous functional studies onERBB2mutations have shown that the S310F, D769H, and V842I variants identified here are activating mutations that support cellular transformation[4] and [5]. To our knowledge, the R678L and V777M mutations have not been functionally characterized. However, a different amino acid substitution at the same V777 position was found to be activating [5] . It was recently reported that micropapillary urothelial carcinomas (UCs) carry a high frequency (40%) of activating extracellular domainERBB2mutations [6] . We therefore reviewed all ourERBB2-mutant UCs. None of these had micropapillary variant histology.
We confirmed the association betweenERBB2missense mutations and responsiveness to platinum-containing chemotherapy in the recently published MIBC data set of Van Allen et al [7] . In this external validation cohort, all three patients with anERBB2missense mutation responded to NAC [7] . In addition, they reported a fourth complete responder who had anERBB2mutation with an allelic fraction of 0.04.
ERBB2missense mutations are significantly enriched in the chemotherapy responders from our cohort and the Van Allen et al cohort [7] compared with the unselected TCGA cohort (8% in TCGA;p = 0.02) ( Fig. 1 D). Conversely, we found thatERBB2missense mutations were significantly depleted in the nonresponder cohorts compared with the unselected TCGA cohort (p = 0.02) ( Fig. 1 D).
Having found an association betweenERBB2missense mutations and platinum response, we next tested the association ofERBB2amplification with platinum response.ERBB2amplifications were identified in complete responders as well as in nonresponders and were not associated with response to NAC (p = 0.52) ( Fig. 2 ).ERBB2amplification was always accompanied by protein overexpression. In four patients, amplification ofERBB2was found in combination with a missense mutation inERBB2. Strikingly, in all of these cases, theERBB2-mutant allele was found amplified, stressing once more the relevance ofERBB2mutations for MIBC oncogenesis.
Van Allen et al recently reported that missense mutations inERCC2, a nucleotide excision repair gene, were selectively present in 9 of 25 MIBC patients with complete response to cisplatin-containing NAC, whereasERCC2missense mutations were absent in 25 nonresponders [7] . In our discovery cohort, we found sixERCC2missense mutations present in four complete responders and in two nonresponders ( Fig. 1 B). Sanger sequencing of the postchemotherapy-resistant tumors of the twoERCC2-mutant nonresponders demonstrated in both cases that theERCC2missense mutation was still present in the postchemotherapy-resistant tumor ( Fig. 1 E and 1F), indicating that the mutation was not counterselected during chemotherapy. In our validation cohort, we identifiedERCC2missense mutations in two tumor samples from 22 complete responders and in none of the 17 nonresponders. In total, 6 somaticERCC2missense mutations were identified in 38 complete responders (16%) and 2 in 33 nonresponders (6%;p = 0.27) ( Fig. 2 ; Supplementary Table 4). Five of the sixERCC2mutations in complete responders were present in patients with wild-type nonamplifiedERBB2( Fig. 2 ).
Finally, we also sequencedERBB2andERCC2in the pretreatment transurethral resection specimens from the remaining group of 23 patients who had responses to NAC ranging from a minor response to a near-complete response. We identified twoERBB2missense mutations in tumors from patients with a partial response. Both patients are still alive >10 yr after NAC without any signs of disease recurrence (Supplementary Table 5). This supports our finding thatERBB2missense mutations are associated with a favorable response to NAC. We identified four tumors (17%) with anERCC2missense mutation in the remaining group of patients. Two of these patients are alive without signs of disease recurrence, and two patients died due to distant recurrences (Supplementary Table 5).
Responses to NAC can also be plotted as complete response versus noncomplete response, as shown in Supplementary Figure 2. In this comparison,ERBB2mutations are strongly associated with response (p = 0.006), whereasERCC2mutations are not.
Our findings indicate thatERBB2missense mutations could assist in selecting patients responding to NAC. Furthermore, these results suggest thatHER2-directed therapies forERBB2mutant bladder cancers are unlikely to replace chemotherapy in the neoadjuvant setting because these tumors have highly favorable responses to NAC containing platinum. However, despite a pathologic complete response to NAC, 3 of the 11 patients withERBB2-mutant MIBC developed a distant recurrence. These patients may benefit from ERBB2 tyrosine kinase inhibitors, alone or in combination with chemotherapy. In contrast to the findings of Van Allen et al, we showed that the presence of anERCC2mutation does not always confer sensitivity to platinum-based therapy.
A possible limitation of this study is the heterogeneity of our cohort. Patients were treated with different platinum-containing chemotherapy regimens: MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin), gemcitabine and cisplatin, or gemcitabine and carboplatin. Although evidence for the benefit of gemcitabine and carboplatin in the neoadjuvant setting in terms of cancer-specific or overall survival is lacking, pathologic complete response rates appear to be similar [8] . Our cohort also contained more advanced cases than most neoadjuvant studies in bladder cancer. However, these patients reflect common clinical practice because many clinics would specifically treat this high-risk patient group with chemotherapy, followed by resection if possible.
In conclusion, we found thatERBB2missense mutations in MIBC are associated with an excellent response to NAC.
Author contributions:Michiel S. van der Heijden had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design:Groenendijk, de Jong, van Rhijn, Bernards, van der Heijden.
Acquisition of data:Groenendijk, van de Putte, Michaut, Schlicker.
Analysis and interpretation of data:Groenendijk, de Jong, van de Putte, Michaut, Velds, Bernards, van der Heijden.
Drafting of the manuscript:Groenendijk, Bernards, van der Heijden.
Critical revision of the manuscript for important intellectual content:de Jong, Michaut, Wessels, Broeks, van Rhijn.
Statistical analysis:Groenendijk, van de Putte, Michaut.
Obtaining funding:van der Heijden, Bernards.
Administrative, technical, or material support:Peters, Nieuwland, Kerkhoven, Broeks.
Supervision:Wessels, van Rhijn, Bernards, van der Heijden.
Other(specify): van de Heuvel (development of methodology).
Financial disclosures:Michiel S. van der Heijden certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:This work was supported by The Netherlands Organization for Scientific Research (to Michiel S. van der Heijden), the European Research Council (grant number 250043 to René Bernards), and the Cancer Genomics Netherlands consortium (to René Bernards).
Acknowledgment statement:The authors wish to acknowledge all patients who contributed tissue for research. We thank the Core Facility for Molecular Pathology & Biobanking of our institute for their assistance. We acknowledge Laura S. Mertens (Department of Urology) and Joyce Sanders (Department of Pathology) for the selection of patients and tissues. We have used data generated by the TCGA Research Network ( http://cancergenome.nih.gov/ ).
A pathologic complete response to neoadjuvant chemotherapy (NAC) containing platinum is associated with superior clinical outcome in patients with muscle-invasive bladder cancer (MIBC)[1], [2], and [3]. No molecular markers or baseline clinical characteristics that can predict the response to NAC are clinically validated. In the present study, we found an unexpected association between mutations in the erb-b2 receptor tyrosine kinase 2 (ERBB2) gene, also known asHER2, and a complete response to chemotherapy.
We collected pre- and postchemotherapy specimens from 110 prospectively registered MIBC patients treated with NAC. Good-quality DNA from the pretreatment transurethral resection specimens was available for 94 patients. In this cohort, we identified 38 pathologic complete responders (ypT0N0), 23 partial responders (ranging from a minor response to a near-complete response), and 33 nonresponders (higher than ypT2) to NAC. Full study methods are described in the supplement. No significant differences in baseline clinical characteristics were identified between the three groups (Supplementary Table 1). Complete responders had a superior recurrence-free and cancer-specific survival compared with nonresponders, whereas the partial responders had an in-between survival (p < 0.001) (Supplementary Fig. 1).
We sequenced 178 cancer-associated genes (Supplementary Table 2) on pretreatment tumor DNA from 16 complete responders and 16 nonresponders (discovery cohort). Genes with a differential mutation frequency in complete responders compared with nonresponders were identified by contrasting analysis ( Fig. 1 A).ERBB2had the highest enrichment for mutations in complete responders ( Fig. 1 A and 1B). We therefore tested the association betweenERBB2mutations and chemotherapy sensitivity in a validation cohort consisting of the remaining 22 complete responders and 17 nonresponders to NAC in our patient series. We identified another fiveERBB2missense mutations in four complete responders and none in the nonresponders. Taken together, we identifiedERBB2missense mutations in 9 of the 38 complete responders (24%) and in none of the 33 nonresponders to NAC (p = 0.003) ( Fig. 2 ).
Fig. 1 Gene enrichment analysis of mutated genes in complete responders and nonresponders. (A) Volcano plot of effect size (log-odds ratio) and significance (−log10pvalue) of the 25 genes mutated in more than two samples. Mutated genes enriched in complete responders are labeled green; mutated genes enriched in nonresponders are labeled red. (B) Pyramid plot showing the number of mutated samples in 16 complete responders in green and the number of mutated samples in 16 nonresponders in red. (C) Plot showing the distribution ofERBB2missense mutations identified in this study by target enrichment next-generation sequencing (dark blue circles) or Sanger sequencing (orange circles).ERBB2missense mutations cluster at the S310 position in the extracellular domain and in the tyrosine kinase domain. (D) Graph showing thatERBB2missense mutations are significantly enriched in responders and significantly depleted in nonresponders to neoadjuvant chemotherapy compared with the unselected Cancer Genome Atlas urothelial bladder cancer cohort (*p < 0.05). (E, F) Sequencing results from twoERCC2-mutant nonresponders showing theERCC2missense mutation (E86Q and S44L) in DNA isolated from the pretreatment transurethral resection and post-treatment tissue but not in the germline DNA. This demonstrates that these somaticERCC2mutations were not counterselected during chemotherapy. NGS = next-generation sequencing; TCGA = The Cancer Genome Atlas; TKD = tyrosine kinase domain; TM = transmembrane domain; TUR = transurethral resection.
Fig. 2 OncoPrint showingERBB2missense mutations,ERBB2amplifications, andERCC2missense mutations in the 38 complete responders and 33 nonresponders to neoadjuvant chemotherapy in this study. Individual patients are represented as columns. N/A = not available; pCR = pathologic complete response.
Five of the 10 identifiedERBB2missense mutations cluster at amino acid 310 in the extracellular domain ( Fig. 1 C; Supplementary Table 3). The S310 position is also a mutational hotspot in The Cancer Genome Atlas (TCGA) urothelial bladder cancer cohort because approximately 40% of allERBB2missense mutations cluster at this position (TCGA Data Portal; http://cancergenome.nih.gov ). Previous functional studies onERBB2mutations have shown that the S310F, D769H, and V842I variants identified here are activating mutations that support cellular transformation[4] and [5]. To our knowledge, the R678L and V777M mutations have not been functionally characterized. However, a different amino acid substitution at the same V777 position was found to be activating [5] . It was recently reported that micropapillary urothelial carcinomas (UCs) carry a high frequency (40%) of activating extracellular domainERBB2mutations [6] . We therefore reviewed all ourERBB2-mutant UCs. None of these had micropapillary variant histology.
We confirmed the association betweenERBB2missense mutations and responsiveness to platinum-containing chemotherapy in the recently published MIBC data set of Van Allen et al [7] . In this external validation cohort, all three patients with anERBB2missense mutation responded to NAC [7] . In addition, they reported a fourth complete responder who had anERBB2mutation with an allelic fraction of 0.04.
ERBB2missense mutations are significantly enriched in the chemotherapy responders from our cohort and the Van Allen et al cohort [7] compared with the unselected TCGA cohort (8% in TCGA;p = 0.02) ( Fig. 1 D). Conversely, we found thatERBB2missense mutations were significantly depleted in the nonresponder cohorts compared with the unselected TCGA cohort (p = 0.02) ( Fig. 1 D).
Having found an association betweenERBB2missense mutations and platinum response, we next tested the association ofERBB2amplification with platinum response.ERBB2amplifications were identified in complete responders as well as in nonresponders and were not associated with response to NAC (p = 0.52) ( Fig. 2 ).ERBB2amplification was always accompanied by protein overexpression. In four patients, amplification ofERBB2was found in combination with a missense mutation inERBB2. Strikingly, in all of these cases, theERBB2-mutant allele was found amplified, stressing once more the relevance ofERBB2mutations for MIBC oncogenesis.
Van Allen et al recently reported that missense mutations inERCC2, a nucleotide excision repair gene, were selectively present in 9 of 25 MIBC patients with complete response to cisplatin-containing NAC, whereasERCC2missense mutations were absent in 25 nonresponders [7] . In our discovery cohort, we found sixERCC2missense mutations present in four complete responders and in two nonresponders ( Fig. 1 B). Sanger sequencing of the postchemotherapy-resistant tumors of the twoERCC2-mutant nonresponders demonstrated in both cases that theERCC2missense mutation was still present in the postchemotherapy-resistant tumor ( Fig. 1 E and 1F), indicating that the mutation was not counterselected during chemotherapy. In our validation cohort, we identifiedERCC2missense mutations in two tumor samples from 22 complete responders and in none of the 17 nonresponders. In total, 6 somaticERCC2missense mutations were identified in 38 complete responders (16%) and 2 in 33 nonresponders (6%;p = 0.27) ( Fig. 2 ; Supplementary Table 4). Five of the sixERCC2mutations in complete responders were present in patients with wild-type nonamplifiedERBB2( Fig. 2 ).
Finally, we also sequencedERBB2andERCC2in the pretreatment transurethral resection specimens from the remaining group of 23 patients who had responses to NAC ranging from a minor response to a near-complete response. We identified twoERBB2missense mutations in tumors from patients with a partial response. Both patients are still alive >10 yr after NAC without any signs of disease recurrence (Supplementary Table 5). This supports our finding thatERBB2missense mutations are associated with a favorable response to NAC. We identified four tumors (17%) with anERCC2missense mutation in the remaining group of patients. Two of these patients are alive without signs of disease recurrence, and two patients died due to distant recurrences (Supplementary Table 5).
Responses to NAC can also be plotted as complete response versus noncomplete response, as shown in Supplementary Figure 2. In this comparison,ERBB2mutations are strongly associated with response (p = 0.006), whereasERCC2mutations are not.
Our findings indicate thatERBB2missense mutations could assist in selecting patients responding to NAC. Furthermore, these results suggest thatHER2-directed therapies forERBB2mutant bladder cancers are unlikely to replace chemotherapy in the neoadjuvant setting because these tumors have highly favorable responses to NAC containing platinum. However, despite a pathologic complete response to NAC, 3 of the 11 patients withERBB2-mutant MIBC developed a distant recurrence. These patients may benefit from ERBB2 tyrosine kinase inhibitors, alone or in combination with chemotherapy. In contrast to the findings of Van Allen et al, we showed that the presence of anERCC2mutation does not always confer sensitivity to platinum-based therapy.
A possible limitation of this study is the heterogeneity of our cohort. Patients were treated with different platinum-containing chemotherapy regimens: MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin), gemcitabine and cisplatin, or gemcitabine and carboplatin. Although evidence for the benefit of gemcitabine and carboplatin in the neoadjuvant setting in terms of cancer-specific or overall survival is lacking, pathologic complete response rates appear to be similar [8] . Our cohort also contained more advanced cases than most neoadjuvant studies in bladder cancer. However, these patients reflect common clinical practice because many clinics would specifically treat this high-risk patient group with chemotherapy, followed by resection if possible.
In conclusion, we found thatERBB2missense mutations in MIBC are associated with an excellent response to NAC.
Author contributions:Michiel S. van der Heijden had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design:Groenendijk, de Jong, van Rhijn, Bernards, van der Heijden.
Acquisition of data:Groenendijk, van de Putte, Michaut, Schlicker.
Analysis and interpretation of data:Groenendijk, de Jong, van de Putte, Michaut, Velds, Bernards, van der Heijden.
Drafting of the manuscript:Groenendijk, Bernards, van der Heijden.
Critical revision of the manuscript for important intellectual content:de Jong, Michaut, Wessels, Broeks, van Rhijn.
Statistical analysis:Groenendijk, van de Putte, Michaut.
Obtaining funding:van der Heijden, Bernards.
Administrative, technical, or material support:Peters, Nieuwland, Kerkhoven, Broeks.
Supervision:Wessels, van Rhijn, Bernards, van der Heijden.
Other(specify): van de Heuvel (development of methodology).
Financial disclosures:Michiel S. van der Heijden certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:This work was supported by The Netherlands Organization for Scientific Research (to Michiel S. van der Heijden), the European Research Council (grant number 250043 to René Bernards), and the Cancer Genomics Netherlands consortium (to René Bernards).
Acknowledgment statement:The authors wish to acknowledge all patients who contributed tissue for research. We thank the Core Facility for Molecular Pathology & Biobanking of our institute for their assistance. We acknowledge Laura S. Mertens (Department of Urology) and Joyce Sanders (Department of Pathology) for the selection of patients and tissues. We have used data generated by the TCGA Research Network ( http://cancergenome.nih.gov/ ).
A pathologic complete response to neoadjuvant chemotherapy (NAC) containing platinum is associated with superior clinical outcome in patients with muscle-invasive bladder cancer (MIBC)[1], [2], and [3]. No molecular markers or baseline clinical characteristics that can predict the response to NAC are clinically validated. In the present study, we found an unexpected association between mutations in the erb-b2 receptor tyrosine kinase 2 (ERBB2) gene, also known asHER2, and a complete response to chemotherapy.
We collected pre- and postchemotherapy specimens from 110 prospectively registered MIBC patients treated with NAC. Good-quality DNA from the pretreatment transurethral resection specimens was available for 94 patients. In this cohort, we identified 38 pathologic complete responders (ypT0N0), 23 partial responders (ranging from a minor response to a near-complete response), and 33 nonresponders (higher than ypT2) to NAC. Full study methods are described in the supplement. No significant differences in baseline clinical characteristics were identified between the three groups (Supplementary Table 1). Complete responders had a superior recurrence-free and cancer-specific survival compared with nonresponders, whereas the partial responders had an in-between survival (p < 0.001) (Supplementary Fig. 1).
We sequenced 178 cancer-associated genes (Supplementary Table 2) on pretreatment tumor DNA from 16 complete responders and 16 nonresponders (discovery cohort). Genes with a differential mutation frequency in complete responders compared with nonresponders were identified by contrasting analysis ( Fig. 1 A).ERBB2had the highest enrichment for mutations in complete responders ( Fig. 1 A and 1B). We therefore tested the association betweenERBB2mutations and chemotherapy sensitivity in a validation cohort consisting of the remaining 22 complete responders and 17 nonresponders to NAC in our patient series. We identified another fiveERBB2missense mutations in four complete responders and none in the nonresponders. Taken together, we identifiedERBB2missense mutations in 9 of the 38 complete responders (24%) and in none of the 33 nonresponders to NAC (p = 0.003) ( Fig. 2 ).
Fig. 1 Gene enrichment analysis of mutated genes in complete responders and nonresponders. (A) Volcano plot of effect size (log-odds ratio) and significance (−log10pvalue) of the 25 genes mutated in more than two samples. Mutated genes enriched in complete responders are labeled green; mutated genes enriched in nonresponders are labeled red. (B) Pyramid plot showing the number of mutated samples in 16 complete responders in green and the number of mutated samples in 16 nonresponders in red. (C) Plot showing the distribution ofERBB2missense mutations identified in this study by target enrichment next-generation sequencing (dark blue circles) or Sanger sequencing (orange circles).ERBB2missense mutations cluster at the S310 position in the extracellular domain and in the tyrosine kinase domain. (D) Graph showing thatERBB2missense mutations are significantly enriched in responders and significantly depleted in nonresponders to neoadjuvant chemotherapy compared with the unselected Cancer Genome Atlas urothelial bladder cancer cohort (*p < 0.05). (E, F) Sequencing results from twoERCC2-mutant nonresponders showing theERCC2missense mutation (E86Q and S44L) in DNA isolated from the pretreatment transurethral resection and post-treatment tissue but not in the germline DNA. This demonstrates that these somaticERCC2mutations were not counterselected during chemotherapy. NGS = next-generation sequencing; TCGA = The Cancer Genome Atlas; TKD = tyrosine kinase domain; TM = transmembrane domain; TUR = transurethral resection.
Fig. 2 OncoPrint showingERBB2missense mutations,ERBB2amplifications, andERCC2missense mutations in the 38 complete responders and 33 nonresponders to neoadjuvant chemotherapy in this study. Individual patients are represented as columns. N/A = not available; pCR = pathologic complete response.
Five of the 10 identifiedERBB2missense mutations cluster at amino acid 310 in the extracellular domain ( Fig. 1 C; Supplementary Table 3). The S310 position is also a mutational hotspot in The Cancer Genome Atlas (TCGA) urothelial bladder cancer cohort because approximately 40% of allERBB2missense mutations cluster at this position (TCGA Data Portal; http://cancergenome.nih.gov ). Previous functional studies onERBB2mutations have shown that the S310F, D769H, and V842I variants identified here are activating mutations that support cellular transformation[4] and [5]. To our knowledge, the R678L and V777M mutations have not been functionally characterized. However, a different amino acid substitution at the same V777 position was found to be activating [5] . It was recently reported that micropapillary urothelial carcinomas (UCs) carry a high frequency (40%) of activating extracellular domainERBB2mutations [6] . We therefore reviewed all ourERBB2-mutant UCs. None of these had micropapillary variant histology.
We confirmed the association betweenERBB2missense mutations and responsiveness to platinum-containing chemotherapy in the recently published MIBC data set of Van Allen et al [7] . In this external validation cohort, all three patients with anERBB2missense mutation responded to NAC [7] . In addition, they reported a fourth complete responder who had anERBB2mutation with an allelic fraction of 0.04.
ERBB2missense mutations are significantly enriched in the chemotherapy responders from our cohort and the Van Allen et al cohort [7] compared with the unselected TCGA cohort (8% in TCGA;p = 0.02) ( Fig. 1 D). Conversely, we found thatERBB2missense mutations were significantly depleted in the nonresponder cohorts compared with the unselected TCGA cohort (p = 0.02) ( Fig. 1 D).
Having found an association betweenERBB2missense mutations and platinum response, we next tested the association ofERBB2amplification with platinum response.ERBB2amplifications were identified in complete responders as well as in nonresponders and were not associated with response to NAC (p = 0.52) ( Fig. 2 ).ERBB2amplification was always accompanied by protein overexpression. In four patients, amplification ofERBB2was found in combination with a missense mutation inERBB2. Strikingly, in all of these cases, theERBB2-mutant allele was found amplified, stressing once more the relevance ofERBB2mutations for MIBC oncogenesis.
Van Allen et al recently reported that missense mutations inERCC2, a nucleotide excision repair gene, were selectively present in 9 of 25 MIBC patients with complete response to cisplatin-containing NAC, whereasERCC2missense mutations were absent in 25 nonresponders [7] . In our discovery cohort, we found sixERCC2missense mutations present in four complete responders and in two nonresponders ( Fig. 1 B). Sanger sequencing of the postchemotherapy-resistant tumors of the twoERCC2-mutant nonresponders demonstrated in both cases that theERCC2missense mutation was still present in the postchemotherapy-resistant tumor ( Fig. 1 E and 1F), indicating that the mutation was not counterselected during chemotherapy. In our validation cohort, we identifiedERCC2missense mutations in two tumor samples from 22 complete responders and in none of the 17 nonresponders. In total, 6 somaticERCC2missense mutations were identified in 38 complete responders (16%) and 2 in 33 nonresponders (6%;p = 0.27) ( Fig. 2 ; Supplementary Table 4). Five of the sixERCC2mutations in complete responders were present in patients with wild-type nonamplifiedERBB2( Fig. 2 ).
Finally, we also sequencedERBB2andERCC2in the pretreatment transurethral resection specimens from the remaining group of 23 patients who had responses to NAC ranging from a minor response to a near-complete response. We identified twoERBB2missense mutations in tumors from patients with a partial response. Both patients are still alive >10 yr after NAC without any signs of disease recurrence (Supplementary Table 5). This supports our finding thatERBB2missense mutations are associated with a favorable response to NAC. We identified four tumors (17%) with anERCC2missense mutation in the remaining group of patients. Two of these patients are alive without signs of disease recurrence, and two patients died due to distant recurrences (Supplementary Table 5).
Responses to NAC can also be plotted as complete response versus noncomplete response, as shown in Supplementary Figure 2. In this comparison,ERBB2mutations are strongly associated with response (p = 0.006), whereasERCC2mutations are not.
Our findings indicate thatERBB2missense mutations could assist in selecting patients responding to NAC. Furthermore, these results suggest thatHER2-directed therapies forERBB2mutant bladder cancers are unlikely to replace chemotherapy in the neoadjuvant setting because these tumors have highly favorable responses to NAC containing platinum. However, despite a pathologic complete response to NAC, 3 of the 11 patients withERBB2-mutant MIBC developed a distant recurrence. These patients may benefit from ERBB2 tyrosine kinase inhibitors, alone or in combination with chemotherapy. In contrast to the findings of Van Allen et al, we showed that the presence of anERCC2mutation does not always confer sensitivity to platinum-based therapy.
A possible limitation of this study is the heterogeneity of our cohort. Patients were treated with different platinum-containing chemotherapy regimens: MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin), gemcitabine and cisplatin, or gemcitabine and carboplatin. Although evidence for the benefit of gemcitabine and carboplatin in the neoadjuvant setting in terms of cancer-specific or overall survival is lacking, pathologic complete response rates appear to be similar [8] . Our cohort also contained more advanced cases than most neoadjuvant studies in bladder cancer. However, these patients reflect common clinical practice because many clinics would specifically treat this high-risk patient group with chemotherapy, followed by resection if possible.
In conclusion, we found thatERBB2missense mutations in MIBC are associated with an excellent response to NAC.
Author contributions:Michiel S. van der Heijden had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design:Groenendijk, de Jong, van Rhijn, Bernards, van der Heijden.
Acquisition of data:Groenendijk, van de Putte, Michaut, Schlicker.
Analysis and interpretation of data:Groenendijk, de Jong, van de Putte, Michaut, Velds, Bernards, van der Heijden.
Drafting of the manuscript:Groenendijk, Bernards, van der Heijden.
Critical revision of the manuscript for important intellectual content:de Jong, Michaut, Wessels, Broeks, van Rhijn.
Statistical analysis:Groenendijk, van de Putte, Michaut.
Obtaining funding:van der Heijden, Bernards.
Administrative, technical, or material support:Peters, Nieuwland, Kerkhoven, Broeks.
Supervision:Wessels, van Rhijn, Bernards, van der Heijden.
Other(specify): van de Heuvel (development of methodology).
Financial disclosures:Michiel S. van der Heijden certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:This work was supported by The Netherlands Organization for Scientific Research (to Michiel S. van der Heijden), the European Research Council (grant number 250043 to René Bernards), and the Cancer Genomics Netherlands consortium (to René Bernards).
Acknowledgment statement:The authors wish to acknowledge all patients who contributed tissue for research. We thank the Core Facility for Molecular Pathology & Biobanking of our institute for their assistance. We acknowledge Laura S. Mertens (Department of Urology) and Joyce Sanders (Department of Pathology) for the selection of patients and tissues. We have used data generated by the TCGA Research Network ( http://cancergenome.nih.gov/ ).
A pathologic complete response to neoadjuvant chemotherapy (NAC) containing platinum is associated with superior clinical outcome in patients with muscle-invasive bladder cancer (MIBC)[1], [2], and [3]. No molecular markers or baseline clinical characteristics that can predict the response to NAC are clinically validated. In the present study, we found an unexpected association between mutations in the erb-b2 receptor tyrosine kinase 2 (ERBB2) gene, also known asHER2, and a complete response to chemotherapy.
We collected pre- and postchemotherapy specimens from 110 prospectively registered MIBC patients treated with NAC. Good-quality DNA from the pretreatment transurethral resection specimens was available for 94 patients. In this cohort, we identified 38 pathologic complete responders (ypT0N0), 23 partial responders (ranging from a minor response to a near-complete response), and 33 nonresponders (higher than ypT2) to NAC. Full study methods are described in the supplement. No significant differences in baseline clinical characteristics were identified between the three groups (Supplementary Table 1). Complete responders had a superior recurrence-free and cancer-specific survival compared with nonresponders, whereas the partial responders had an in-between survival (p < 0.001) (Supplementary Fig. 1).
We sequenced 178 cancer-associated genes (Supplementary Table 2) on pretreatment tumor DNA from 16 complete responders and 16 nonresponders (discovery cohort). Genes with a differential mutation frequency in complete responders compared with nonresponders were identified by contrasting analysis ( Fig. 1 A).ERBB2had the highest enrichment for mutations in complete responders ( Fig. 1 A and 1B). We therefore tested the association betweenERBB2mutations and chemotherapy sensitivity in a validation cohort consisting of the remaining 22 complete responders and 17 nonresponders to NAC in our patient series. We identified another fiveERBB2missense mutations in four complete responders and none in the nonresponders. Taken together, we identifiedERBB2missense mutations in 9 of the 38 complete responders (24%) and in none of the 33 nonresponders to NAC (p = 0.003) ( Fig. 2 ).
Fig. 1 Gene enrichment analysis of mutated genes in complete responders and nonresponders. (A) Volcano plot of effect size (log-odds ratio) and significance (−log10pvalue) of the 25 genes mutated in more than two samples. Mutated genes enriched in complete responders are labeled green; mutated genes enriched in nonresponders are labeled red. (B) Pyramid plot showing the number of mutated samples in 16 complete responders in green and the number of mutated samples in 16 nonresponders in red. (C) Plot showing the distribution ofERBB2missense mutations identified in this study by target enrichment next-generation sequencing (dark blue circles) or Sanger sequencing (orange circles).ERBB2missense mutations cluster at the S310 position in the extracellular domain and in the tyrosine kinase domain. (D) Graph showing thatERBB2missense mutations are significantly enriched in responders and significantly depleted in nonresponders to neoadjuvant chemotherapy compared with the unselected Cancer Genome Atlas urothelial bladder cancer cohort (*p < 0.05). (E, F) Sequencing results from twoERCC2-mutant nonresponders showing theERCC2missense mutation (E86Q and S44L) in DNA isolated from the pretreatment transurethral resection and post-treatment tissue but not in the germline DNA. This demonstrates that these somaticERCC2mutations were not counterselected during chemotherapy. NGS = next-generation sequencing; TCGA = The Cancer Genome Atlas; TKD = tyrosine kinase domain; TM = transmembrane domain; TUR = transurethral resection.
Fig. 2 OncoPrint showingERBB2missense mutations,ERBB2amplifications, andERCC2missense mutations in the 38 complete responders and 33 nonresponders to neoadjuvant chemotherapy in this study. Individual patients are represented as columns. N/A = not available; pCR = pathologic complete response.
Five of the 10 identifiedERBB2missense mutations cluster at amino acid 310 in the extracellular domain ( Fig. 1 C; Supplementary Table 3). The S310 position is also a mutational hotspot in The Cancer Genome Atlas (TCGA) urothelial bladder cancer cohort because approximately 40% of allERBB2missense mutations cluster at this position (TCGA Data Portal; http://cancergenome.nih.gov ). Previous functional studies onERBB2mutations have shown that the S310F, D769H, and V842I variants identified here are activating mutations that support cellular transformation[4] and [5]. To our knowledge, the R678L and V777M mutations have not been functionally characterized. However, a different amino acid substitution at the same V777 position was found to be activating [5] . It was recently reported that micropapillary urothelial carcinomas (UCs) carry a high frequency (40%) of activating extracellular domainERBB2mutations [6] . We therefore reviewed all ourERBB2-mutant UCs. None of these had micropapillary variant histology.
We confirmed the association betweenERBB2missense mutations and responsiveness to platinum-containing chemotherapy in the recently published MIBC data set of Van Allen et al [7] . In this external validation cohort, all three patients with anERBB2missense mutation responded to NAC [7] . In addition, they reported a fourth complete responder who had anERBB2mutation with an allelic fraction of 0.04.
ERBB2missense mutations are significantly enriched in the chemotherapy responders from our cohort and the Van Allen et al cohort [7] compared with the unselected TCGA cohort (8% in TCGA;p = 0.02) ( Fig. 1 D). Conversely, we found thatERBB2missense mutations were significantly depleted in the nonresponder cohorts compared with the unselected TCGA cohort (p = 0.02) ( Fig. 1 D).
Having found an association betweenERBB2missense mutations and platinum response, we next tested the association ofERBB2amplification with platinum response.ERBB2amplifications were identified in complete responders as well as in nonresponders and were not associated with response to NAC (p = 0.52) ( Fig. 2 ).ERBB2amplification was always accompanied by protein overexpression. In four patients, amplification ofERBB2was found in combination with a missense mutation inERBB2. Strikingly, in all of these cases, theERBB2-mutant allele was found amplified, stressing once more the relevance ofERBB2mutations for MIBC oncogenesis.
Van Allen et al recently reported that missense mutations inERCC2, a nucleotide excision repair gene, were selectively present in 9 of 25 MIBC patients with complete response to cisplatin-containing NAC, whereasERCC2missense mutations were absent in 25 nonresponders [7] . In our discovery cohort, we found sixERCC2missense mutations present in four complete responders and in two nonresponders ( Fig. 1 B). Sanger sequencing of the postchemotherapy-resistant tumors of the twoERCC2-mutant nonresponders demonstrated in both cases that theERCC2missense mutation was still present in the postchemotherapy-resistant tumor ( Fig. 1 E and 1F), indicating that the mutation was not counterselected during chemotherapy. In our validation cohort, we identifiedERCC2missense mutations in two tumor samples from 22 complete responders and in none of the 17 nonresponders. In total, 6 somaticERCC2missense mutations were identified in 38 complete responders (16%) and 2 in 33 nonresponders (6%;p = 0.27) ( Fig. 2 ; Supplementary Table 4). Five of the sixERCC2mutations in complete responders were present in patients with wild-type nonamplifiedERBB2( Fig. 2 ).
Finally, we also sequencedERBB2andERCC2in the pretreatment transurethral resection specimens from the remaining group of 23 patients who had responses to NAC ranging from a minor response to a near-complete response. We identified twoERBB2missense mutations in tumors from patients with a partial response. Both patients are still alive >10 yr after NAC without any signs of disease recurrence (Supplementary Table 5). This supports our finding thatERBB2missense mutations are associated with a favorable response to NAC. We identified four tumors (17%) with anERCC2missense mutation in the remaining group of patients. Two of these patients are alive without signs of disease recurrence, and two patients died due to distant recurrences (Supplementary Table 5).
Responses to NAC can also be plotted as complete response versus noncomplete response, as shown in Supplementary Figure 2. In this comparison,ERBB2mutations are strongly associated with response (p = 0.006), whereasERCC2mutations are not.
Our findings indicate thatERBB2missense mutations could assist in selecting patients responding to NAC. Furthermore, these results suggest thatHER2-directed therapies forERBB2mutant bladder cancers are unlikely to replace chemotherapy in the neoadjuvant setting because these tumors have highly favorable responses to NAC containing platinum. However, despite a pathologic complete response to NAC, 3 of the 11 patients withERBB2-mutant MIBC developed a distant recurrence. These patients may benefit from ERBB2 tyrosine kinase inhibitors, alone or in combination with chemotherapy. In contrast to the findings of Van Allen et al, we showed that the presence of anERCC2mutation does not always confer sensitivity to platinum-based therapy.
A possible limitation of this study is the heterogeneity of our cohort. Patients were treated with different platinum-containing chemotherapy regimens: MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin), gemcitabine and cisplatin, or gemcitabine and carboplatin. Although evidence for the benefit of gemcitabine and carboplatin in the neoadjuvant setting in terms of cancer-specific or overall survival is lacking, pathologic complete response rates appear to be similar [8] . Our cohort also contained more advanced cases than most neoadjuvant studies in bladder cancer. However, these patients reflect common clinical practice because many clinics would specifically treat this high-risk patient group with chemotherapy, followed by resection if possible.
In conclusion, we found thatERBB2missense mutations in MIBC are associated with an excellent response to NAC.
Author contributions:Michiel S. van der Heijden had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design:Groenendijk, de Jong, van Rhijn, Bernards, van der Heijden.
Acquisition of data:Groenendijk, van de Putte, Michaut, Schlicker.
Analysis and interpretation of data:Groenendijk, de Jong, van de Putte, Michaut, Velds, Bernards, van der Heijden.
Drafting of the manuscript:Groenendijk, Bernards, van der Heijden.
Critical revision of the manuscript for important intellectual content:de Jong, Michaut, Wessels, Broeks, van Rhijn.
Statistical analysis:Groenendijk, van de Putte, Michaut.
Obtaining funding:van der Heijden, Bernards.
Administrative, technical, or material support:Peters, Nieuwland, Kerkhoven, Broeks.
Supervision:Wessels, van Rhijn, Bernards, van der Heijden.
Other(specify): van de Heuvel (development of methodology).
Financial disclosures:Michiel S. van der Heijden certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:This work was supported by The Netherlands Organization for Scientific Research (to Michiel S. van der Heijden), the European Research Council (grant number 250043 to René Bernards), and the Cancer Genomics Netherlands consortium (to René Bernards).
Acknowledgment statement:The authors wish to acknowledge all patients who contributed tissue for research. We thank the Core Facility for Molecular Pathology & Biobanking of our institute for their assistance. We acknowledge Laura S. Mertens (Department of Urology) and Joyce Sanders (Department of Pathology) for the selection of patients and tissues. We have used data generated by the TCGA Research Network ( http://cancergenome.nih.gov/ ).
A pathologic complete response to neoadjuvant chemotherapy (NAC) containing platinum is associated with superior clinical outcome in patients with muscle-invasive bladder cancer (MIBC)[1], [2], and [3]. No molecular markers or baseline clinical characteristics that can predict the response to NAC are clinically validated. In the present study, we found an unexpected association between mutations in the erb-b2 receptor tyrosine kinase 2 (ERBB2) gene, also known asHER2, and a complete response to chemotherapy.
We collected pre- and postchemotherapy specimens from 110 prospectively registered MIBC patients treated with NAC. Good-quality DNA from the pretreatment transurethral resection specimens was available for 94 patients. In this cohort, we identified 38 pathologic complete responders (ypT0N0), 23 partial responders (ranging from a minor response to a near-complete response), and 33 nonresponders (higher than ypT2) to NAC. Full study methods are described in the supplement. No significant differences in baseline clinical characteristics were identified between the three groups (Supplementary Table 1). Complete responders had a superior recurrence-free and cancer-specific survival compared with nonresponders, whereas the partial responders had an in-between survival (p < 0.001) (Supplementary Fig. 1).
We sequenced 178 cancer-associated genes (Supplementary Table 2) on pretreatment tumor DNA from 16 complete responders and 16 nonresponders (discovery cohort). Genes with a differential mutation frequency in complete responders compared with nonresponders were identified by contrasting analysis ( Fig. 1 A).ERBB2had the highest enrichment for mutations in complete responders ( Fig. 1 A and 1B). We therefore tested the association betweenERBB2mutations and chemotherapy sensitivity in a validation cohort consisting of the remaining 22 complete responders and 17 nonresponders to NAC in our patient series. We identified another fiveERBB2missense mutations in four complete responders and none in the nonresponders. Taken together, we identifiedERBB2missense mutations in 9 of the 38 complete responders (24%) and in none of the 33 nonresponders to NAC (p = 0.003) ( Fig. 2 ).
Fig. 1 Gene enrichment analysis of mutated genes in complete responders and nonresponders. (A) Volcano plot of effect size (log-odds ratio) and significance (−log10pvalue) of the 25 genes mutated in more than two samples. Mutated genes enriched in complete responders are labeled green; mutated genes enriched in nonresponders are labeled red. (B) Pyramid plot showing the number of mutated samples in 16 complete responders in green and the number of mutated samples in 16 nonresponders in red. (C) Plot showing the distribution ofERBB2missense mutations identified in this study by target enrichment next-generation sequencing (dark blue circles) or Sanger sequencing (orange circles).ERBB2missense mutations cluster at the S310 position in the extracellular domain and in the tyrosine kinase domain. (D) Graph showing thatERBB2missense mutations are significantly enriched in responders and significantly depleted in nonresponders to neoadjuvant chemotherapy compared with the unselected Cancer Genome Atlas urothelial bladder cancer cohort (*p < 0.05). (E, F) Sequencing results from twoERCC2-mutant nonresponders showing theERCC2missense mutation (E86Q and S44L) in DNA isolated from the pretreatment transurethral resection and post-treatment tissue but not in the germline DNA. This demonstrates that these somaticERCC2mutations were not counterselected during chemotherapy. NGS = next-generation sequencing; TCGA = The Cancer Genome Atlas; TKD = tyrosine kinase domain; TM = transmembrane domain; TUR = transurethral resection.
Fig. 2 OncoPrint showingERBB2missense mutations,ERBB2amplifications, andERCC2missense mutations in the 38 complete responders and 33 nonresponders to neoadjuvant chemotherapy in this study. Individual patients are represented as columns. N/A = not available; pCR = pathologic complete response.
Five of the 10 identifiedERBB2missense mutations cluster at amino acid 310 in the extracellular domain ( Fig. 1 C; Supplementary Table 3). The S310 position is also a mutational hotspot in The Cancer Genome Atlas (TCGA) urothelial bladder cancer cohort because approximately 40% of allERBB2missense mutations cluster at this position (TCGA Data Portal; http://cancergenome.nih.gov ). Previous functional studies onERBB2mutations have shown that the S310F, D769H, and V842I variants identified here are activating mutations that support cellular transformation[4] and [5]. To our knowledge, the R678L and V777M mutations have not been functionally characterized. However, a different amino acid substitution at the same V777 position was found to be activating [5] . It was recently reported that micropapillary urothelial carcinomas (UCs) carry a high frequency (40%) of activating extracellular domainERBB2mutations [6] . We therefore reviewed all ourERBB2-mutant UCs. None of these had micropapillary variant histology.
We confirmed the association betweenERBB2missense mutations and responsiveness to platinum-containing chemotherapy in the recently published MIBC data set of Van Allen et al [7] . In this external validation cohort, all three patients with anERBB2missense mutation responded to NAC [7] . In addition, they reported a fourth complete responder who had anERBB2mutation with an allelic fraction of 0.04.
ERBB2missense mutations are significantly enriched in the chemotherapy responders from our cohort and the Van Allen et al cohort [7] compared with the unselected TCGA cohort (8% in TCGA;p = 0.02) ( Fig. 1 D). Conversely, we found thatERBB2missense mutations were significantly depleted in the nonresponder cohorts compared with the unselected TCGA cohort (p = 0.02) ( Fig. 1 D).
Having found an association betweenERBB2missense mutations and platinum response, we next tested the association ofERBB2amplification with platinum response.ERBB2amplifications were identified in complete responders as well as in nonresponders and were not associated with response to NAC (p = 0.52) ( Fig. 2 ).ERBB2amplification was always accompanied by protein overexpression. In four patients, amplification ofERBB2was found in combination with a missense mutation inERBB2. Strikingly, in all of these cases, theERBB2-mutant allele was found amplified, stressing once more the relevance ofERBB2mutations for MIBC oncogenesis.
Van Allen et al recently reported that missense mutations inERCC2, a nucleotide excision repair gene, were selectively present in 9 of 25 MIBC patients with complete response to cisplatin-containing NAC, whereasERCC2missense mutations were absent in 25 nonresponders [7] . In our discovery cohort, we found sixERCC2missense mutations present in four complete responders and in two nonresponders ( Fig. 1 B). Sanger sequencing of the postchemotherapy-resistant tumors of the twoERCC2-mutant nonresponders demonstrated in both cases that theERCC2missense mutation was still present in the postchemotherapy-resistant tumor ( Fig. 1 E and 1F), indicating that the mutation was not counterselected during chemotherapy. In our validation cohort, we identifiedERCC2missense mutations in two tumor samples from 22 complete responders and in none of the 17 nonresponders. In total, 6 somaticERCC2missense mutations were identified in 38 complete responders (16%) and 2 in 33 nonresponders (6%;p = 0.27) ( Fig. 2 ; Supplementary Table 4). Five of the sixERCC2mutations in complete responders were present in patients with wild-type nonamplifiedERBB2( Fig. 2 ).
Finally, we also sequencedERBB2andERCC2in the pretreatment transurethral resection specimens from the remaining group of 23 patients who had responses to NAC ranging from a minor response to a near-complete response. We identified twoERBB2missense mutations in tumors from patients with a partial response. Both patients are still alive >10 yr after NAC without any signs of disease recurrence (Supplementary Table 5). This supports our finding thatERBB2missense mutations are associated with a favorable response to NAC. We identified four tumors (17%) with anERCC2missense mutation in the remaining group of patients. Two of these patients are alive without signs of disease recurrence, and two patients died due to distant recurrences (Supplementary Table 5).
Responses to NAC can also be plotted as complete response versus noncomplete response, as shown in Supplementary Figure 2. In this comparison,ERBB2mutations are strongly associated with response (p = 0.006), whereasERCC2mutations are not.
Our findings indicate thatERBB2missense mutations could assist in selecting patients responding to NAC. Furthermore, these results suggest thatHER2-directed therapies forERBB2mutant bladder cancers are unlikely to replace chemotherapy in the neoadjuvant setting because these tumors have highly favorable responses to NAC containing platinum. However, despite a pathologic complete response to NAC, 3 of the 11 patients withERBB2-mutant MIBC developed a distant recurrence. These patients may benefit from ERBB2 tyrosine kinase inhibitors, alone or in combination with chemotherapy. In contrast to the findings of Van Allen et al, we showed that the presence of anERCC2mutation does not always confer sensitivity to platinum-based therapy.
A possible limitation of this study is the heterogeneity of our cohort. Patients were treated with different platinum-containing chemotherapy regimens: MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin), gemcitabine and cisplatin, or gemcitabine and carboplatin. Although evidence for the benefit of gemcitabine and carboplatin in the neoadjuvant setting in terms of cancer-specific or overall survival is lacking, pathologic complete response rates appear to be similar [8] . Our cohort also contained more advanced cases than most neoadjuvant studies in bladder cancer. However, these patients reflect common clinical practice because many clinics would specifically treat this high-risk patient group with chemotherapy, followed by resection if possible.
In conclusion, we found thatERBB2missense mutations in MIBC are associated with an excellent response to NAC.
Author contributions:Michiel S. van der Heijden had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design:Groenendijk, de Jong, van Rhijn, Bernards, van der Heijden.
Acquisition of data:Groenendijk, van de Putte, Michaut, Schlicker.
Analysis and interpretation of data:Groenendijk, de Jong, van de Putte, Michaut, Velds, Bernards, van der Heijden.
Drafting of the manuscript:Groenendijk, Bernards, van der Heijden.
Critical revision of the manuscript for important intellectual content:de Jong, Michaut, Wessels, Broeks, van Rhijn.
Statistical analysis:Groenendijk, van de Putte, Michaut.
Obtaining funding:van der Heijden, Bernards.
Administrative, technical, or material support:Peters, Nieuwland, Kerkhoven, Broeks.
Supervision:Wessels, van Rhijn, Bernards, van der Heijden.
Other(specify): van de Heuvel (development of methodology).
Financial disclosures:Michiel S. van der Heijden certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:This work was supported by The Netherlands Organization for Scientific Research (to Michiel S. van der Heijden), the European Research Council (grant number 250043 to René Bernards), and the Cancer Genomics Netherlands consortium (to René Bernards).
Acknowledgment statement:The authors wish to acknowledge all patients who contributed tissue for research. We thank the Core Facility for Molecular Pathology & Biobanking of our institute for their assistance. We acknowledge Laura S. Mertens (Department of Urology) and Joyce Sanders (Department of Pathology) for the selection of patients and tissues. We have used data generated by the TCGA Research Network ( http://cancergenome.nih.gov/ ).
The research for markers of therapeutic response could help us to individualize therapies in bladder cancer. In this work the authors studied a population of 71 patients submitted to neodjuvant chemotherapy and posterior radical cystectomy, and divided the on responders and non-responders based on final pathological report. By DNA-based technology, the authors concluded that ERBB2 mutations are a prognostic factor of good response to cisplatin base chemotherapy. Although this constitutes a very promising marker for selecting patients for NAC, further validation of their results is warranted in order to establish the reliability of their results.