[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to main content
Download PDF

Whole-exome profiles of inflammatory breast cancer and pathological response to neoadjuvant chemotherapy

Journal of Translational Medicine volume 22, Article number: 969 (2024) Cite this article

Abstract

Background

Neoadjuvant chemotherapy (NACT) became a standard treatment strategy for patients with inflammatory breast cancer (IBC) because of high disease aggressiveness. However, given the heterogeneity of IBC, no molecular feature reliably predicts the response to chemotherapy. Whole-exome sequencing (WES) of clinical tumor samples provides an opportunity to identify genomic alterations associated with chemosensitivity.

Methods

We retrospectively applied WES to 44 untreated IBC primary tumor samples and matched normal DNA. The pathological response to NACT, assessed on operative specimen, distinguished the patients with versus without pathological complete response (pCR versus no-pCR respectively). We compared the mutational profiles, spectra and signatures, pathway mutations, copy number alterations (CNAs), HRD, and heterogeneity scores between pCR versus no-pCR patients.

Results

The TMB, HRD, and mutational spectra were not different between the complete (N = 13) versus non-complete (N = 31) responders. The two most frequently mutated genes were TP53 and PIK3CA. They were more frequently mutated in the complete responders, but the difference was not significant. Only two genes, NLRP3 and SLC9B1, were significantly more frequently mutated in the complete responders (23% vs. 0%). By contrast, several biological pathways involved in protein translation, PI3K pathway, and signal transduction showed significantly higher mutation frequency in the patients with pCR. We observed a higher abundance of COSMIC signature 7 (due to ultraviolet light exposure) in tumors from complete responders. The comparison of CNAs of the 3808 genes included in the GISTIC regions between both patients’ groups identified 234 genes as differentially altered. The CIN signatures were not differentially represented between the complete versus non-complete responders. Based on the H-index, the patients with heterogeneous tumors displayed a lower pCR rate (11%) than those with less heterogeneous tumors (35%).

Conclusions

This is the first study aiming at identifying correlations between the WES data of IBC samples and the achievement of pCR to NACT. Our results, obtained in this 44-sample series, suggest a few subtle genomic alterations associated with pathological response. Additional investigations are required in larger series.

Background

Inflammatory breast cancer (IBC) represents the most aggressive form of breast cancer (BC) due to high metastatic potential [1]. If the disease is relatively rare, compared to non-IBCs (~ 3% of all BCs), it accounts for a higher proportion of BC-related deaths as compared to non-IBCs (8–10%). Despite progresses in the multidisciplinary treatment based on neoadjuvant chemotherapy (NACT) followed by surgery, radiation therapy, then systemic therapy [2], the long-term survival remains poor, around 50%. The rate of pathological complete response (pCR) after NACT ranges from 15 to 40%, and is associated with survival. However, the response and its benefit are heterogeneous between patients, and to date, no validated predictive factor is used in clinical routine.

Identification of biomarkers predictive for pathological response to NACT is a major research focus. But because IBC is rare and the pre-treatment biopsy samples are small, past molecular studies have not been conclusive [3]. During the last 20 years, “omics” studies were applied to clinical samples [4], mainly based on gene expression profiling [5]. In this context, the largest series reported to date remains the one we reported within the World IBC Consortium [6, 7], in which the high number of IBC (n = 137) allowed us to identify a robust gene expression signature associated with pCR [6]. In parallel, studies on the association of whole-exome sequencing (WES) profiles of BCs with response to NACT are much more limited, with to our knowledge, 10 studies reported in the literature in non-IBC and including 21 to 405 (median 134) patients [8,9,10,11,12,13,14,15,16,17]. None of these studies was dedicated to IBC. Among the three studies that applied WES to clinical IBC samples [18,19,20], (respectively 22 patients with advanced HER2+ IBC and 6 patients with hormone receptor-positive (HR+) IBC), none addressed this issue. The objective of our present study is to search for somatic or germline genomic alterations associated with pCR to NACT in the 44 informative samples included in our recent retrospective study of 54 IBC tumor samples [20].

Methods

Patients and samples selection

We retrospectively collected pre-treatment tumor samples and paired peripheral blood samples from 52 patients with IBC treated in our institutions: 28 patients at Institut Paoli-Calmettes (Marseille, France) and 24 at Institut Curie (Paris and Saint-Cloud, France). The tumor samples corresponded to diagnostic biopsy taken before any systemic treatment: The study was approved by our institutional review boards. IBC was clinically defined as T4d according to the international consensus criteria [1]. Extraction and quality control of tumor DNA were done as described [21]. Other selection criteria included female patients, available frozen sample, tumor cellularity (> 70%), good quality of extracted tumor DNA and germline DNA, available clinicopathological data, and available patients’ written informed consent for somatic and constitutional genomic analysis. The final series included 54 IBCs. The tumor molecular status was based upon immunohistochemistry (IHC) and sometimes in situ hybridization (ISH) or HER2: the hormone receptor (HR) status was positive when estrogen receptor (ER) and/or progesterone receptor (PR) expression were scored ≥ 10% by IHC, and the HER2 status was positive when IHC score was 3 + (more than 10% tumor cells with intense, complete and homogeneous membrane staining of HER2) or 2 + with a positive ISH result. The tumors were then classified as HR+/HER2− when HR was positive and HER2 negative, HER2+ when HER2 was positive, and TN when the three receptors were negative. All patients were treated with anthracycline-based NACT, often including taxane, and coupled with trastuzumab in one third of cases with HER2 amplification. NACT was followed by mastectomy and axillary lymph node dissection for clinically non-progressive and consenting patients. The pathological response to NACT was defined on the surgical specimens of both primary tumor and lymph nodes. Among the 52 patients, 44 underwent surgery after NACT and were assessed in term of pathological response to NACT: they represent the analysis population of the present study.

Whole-exome sequencing

Methods were previously described in [20]. Briefly, after extraction of tumor and germline DNA [21], WES was performed using Illumina HiSeq 2500 sequencing systems. The 150 bp paired-end libraries were prepared using the Sureselect Human All Exome capturing kit (Agilent, Santa Clara, CA, USA). We then aligned the sequence to the human reference genome (UCSC hg19) using the Burrows-Wheeler Aligner [22]. Tumor and normal samples were sequenced at a median depth of 188x (range, 49 to 462) and 63x (range, 29 to 1087, respectively. Bam files were deduplicated, realigned and base recalibration was applied with GATK version 3.7 [23]. Somatic single nucleotide variants (SNVs) calling were done with Mutect [24]. Somatic insertions/deletions (indels) calling were done with Strelka2 [25]. The variants, i.e., SNVs and indels, were annotated with the Annotate Variation Software (ANNOVAR, version 2013-11-12) [26]. The TMB was defined as the number of somatic coding mutations including missense, nonsense, silent, and indel divided by the panel size. COSMIC mutational signatures were computed with the Python program mutation-signatures (https://github.com/mskcc/mutation-signatures). We collected 1576 biological pathways from the HumanCyc, KEGG, PANTHER, pid, and Reactome pathway databases and considered a pathway as mutated if it had one or more member genes with somatic mutation. To assess significance, we first calculated the odds ratio (OR) of the response group (pCR or no-pCR) versus the pathway status (mutated versus wild-type) using logistic regression; next we randomly permuted the pCR or no-pCR labels for 1000 iterations and recalculated the OR for each pathway. The proportion of random permutations showing an OR greater than the OR of the actual data defined the p-value. The regions significantly gained/amplified or lost/deleted across samples (q < 0.25) were identified using the GISTIC2 software [27] with the alteration threshold set at 0.2. Allelic copy numbers were estimated with FACETS [28]. We also assessed the distribution of the 17 CIN signatures [29] in all tumor samples. The Homologous Recombination Deficiency (HRD) score was determined [30] by considering three independent measures of genomic instability: the number of loss of heterozygosity (LOH), the number of telomeric-allelic imbalances (TAI), and the number of large-scale state transitions (LST), scored from FACETS results. The score was the sum of the TAI, LST, and LOH scores, and the profile was defined as HRD-high when the HRD score was ≥ 42 [31]. The Shannon's Index (H), used to estimate intra-tumor heterogeneity, was assessed using the SciClone package [32]. The calculation of the Cancer Cell Fraction (CCF) for each mutation was based on the purity, allele frequency, and total copy number. Subsequently, credible intervals were computed using the bayestestR package and the Highest Density Interval (HDI) method. If the value 1 (i.e., CCF = 1, clonal) fell within the credible interval, the mutation was classified as "Clonal"; otherwise, it was categorized as "Subclonal”.

Statistical analysis

The primary endpoint, pCR to NACT, was defined as absence of invasive cancer in both breast and axillary lymph nodes (ypT0/Tis ypN0) in the surgical specimen. The continuous variables were described by median and range, and the binary variables by numbers and percentages. Correlations between the clinicopathological (pCR/no-pCR) variables and molecular WES variables were analyzed using Wilcoxon test or Fisher's exact test when appropriate. Correlations between the pCR/no-pCR status and gene expression levels were calculated using a generalized linear model (GLM) with a binomial family and logit link function; we used the test of homogeneity in a fixed-effects model to assess whether the effect sizes measured a common effect size where heterogeneity was assessed using the I2 statistic [33] using the meta package in R Cran. Variables with a p-value inferior to 0.05 were considered as significant. Analyses were done using the R-software (version 4.2.1: http://www.cran.r-project.org/).

Results

Patients and pathological response to NACT

We analyzed 44 IBC samples (Table 1) from patients treated with NACT period over a study period extended from May 1996 to May 2012. All cases were ductal pathological type and from patients with AJCC stage III. The median patients’ age at diagnosis was 48 years (range, 31 to 78). The pathological grade was 3 in 71% of cases. The molecular subtypes were HER2+ in 41% of cases, HR+/HER2− in 39%, and TN in 20%. All patients had been treated with anthracycline-based NACT, followed by mastectomy and axillary lymph node dissection. Assessment of operative specimen identified 13 patients (30%) who achieved pCR after NACT (pCR group) and 31 (70%) who did not (no-pCR group). In univariate analysis for pCR prediction (Supplementary Table 1), and as expected, the pathological grade, and HR and HER2 statutes tended to be associated with pCR (p < 0.10, Fisher's exact test), with higher pCR rate in grade 3 vs. 2 tumors, in HR-negative vs. HR-positive tumors, and in HER2-positive vs. HER2-negative tumors.

Table 1 Clinicopathological characteristics of patients and samples
Full size table

Somatic and germline mutations and response to NACT

WES analysis identified 3,632 somatic mutations in 2937 genes in the 44 IBC samples (Supplementary Table 2). Among these mutations, 95% were single nucleotide variants (SNVs) and 5% were insertions/deletions (indels). In the whole population, the median number of somatic mutations per sample and the median TMB were 62,5 (range, 3 to 942) and 1,24 (range, 0,06 to 21) respectively. The median number of somatic mutations per sample was not significantly different between patients with pCR (62 range, 27 to 188) and those without pCR (63, range, 3 to 942), p = 0.662, Wilcoxon test). Similarly, the median TMB was not different between both patients’ groups: 1.15 mutations/MB (range, 0,60 to 3.48) in complete responders vs. 1.26 (range, 0.06 to 21) in non-complete responders (p = 0.653, Wilcoxon test; Fig. 1A).

Fig. 1
figure 1

TMB and distribution of mutations in the 44 IBC samples. A Violin plots showing the distribution of the TMB in all informative patients (N = 44) and according to the response to NACT (with versus without pCR). The p-value is for the Wilcoxon test. B Oncoprint of the 66 genes mutated in at least 3 IBC samples. Top: Number of somatic mutations in each sample. IHC-based molecular subtypes and pCR/no-pCR groups are color-coded as indicated in the legend. Middle: somatic gene mutations color-coded according to the legend. The genes are ordered from top to bottom by decreasing percentage of altered IBCs right panel. *: genes differentially mutated in complete (pCR) versus non-complete responders (no-pCR) to NACT. Bottom: germline gene mutations. The percentages of mutation in tumor samples are shown to the right of the Oncoprint

Full size image

Regarding somatic mutations, our comparative analysis concerned the 66 genes with at least three samples mutated out of 44. The three most frequently mutated genes were TP53 (55%), PIK3CA (25%), and TTN (25%). PIK3CA was more frequently mutated in the patients with pCR (38%) than in the patients without pCR (19%), but the difference was not significant (Odds Ratio OR = 2.54; 95% CI [0.48–13.44]; p = 0.255, Fisher’s exact test). The same was true for TP53 mutations: 69% of patients with pCR displayed mutations versus 48% of other patients (OR = 2.35; 95% CI 0.52–12.78]; p = 0.321, Fisher’s exact test). Among the 66 genes, only two genes were significantly more frequently mutated in complete responders than in non-complete responders: NLRP3 and SLC9B1 (23% vs. 0% respectively; p = 0.021, Fisher’s exact test; Fig. 1B).

As germline mutations may affect the sensitivity to chemotherapy, we then assessed the presence of pathogenic or likely pathogenic germline mutations of the 67 cancer predisposition genes included in the BROCA list (https://testguide.labmed.uw.edu/public/view/BROCA). Five patients out of 44 displayed such mutation, involving four different genes: ATM, CHEK2, and HOXB13 (one patient each), and MITF (two patients) (Fig. 1B). These mutations were observed in five patients without pCR (p = 0.305).

Mutational spectra and processes of somatic SNVs and response to NACT

The proportions of six possible base substitutions across SNVs are shown in Fig. 2A, B. In the whole series, the most frequent base change was C > T (median of 50% of substitutions, range 18% to 68%) with respect to single-nucleotide-mutation contexts (Fig. 2A). For each base substitution, the proportions were similar between the patients with pCR and those without pCR (p > 0.05). When considering the tri-nucleotide mutation contexts (Fig. 2B), the most frequent base substitution was G[C > T]G in the whole series, and no difference was observed according to the degree of response to NACT.

Fig. 2
figure 2

Mutational processes of somatic SNVs and correlations with pCR. A Proportions of base substitutions with respect to single-nucleotide-mutation contexts in patients with versus without pCR (N = 44). B Similar to A but with respect to tri-nucleotide mutation contexts. C Proportions of the most represented COSMIC mutational signatures in the whole population: age-related (signature S1); APOBEC activation (signatures S2, S13); HRD (signature S3); mismatch repair (signatures S6, S20, S26); POLE (signature S10); ultra-violet UV-related (signature S7); unknown (signatures S5, S8, S17, S18, S30); others (other signatures). The signatures, IHC-based molecular subtypes and pCR/no-pCR groups are color-coded according to the legend

Full size image

We also assessed the distribution of the 30 COSMIC mutational signatures. In the whole series, the most represented signatures were, as expected, the signatures 1 (age-related), 3 (homologous recombination deficiency), and 2 and 13 (APOBEC activation) (Fig. 2C). The comparison between complete vs. non-complete responders identified only one signature differentially represented between both groups: the signature 7 that was more abundant in the complete responders (8% vs. 3%; p = 0,015, Wilcoxon test).

Pathway-level somatic mutations and response to NACT

We assessed associations between pCR and pathway-level mutations. Among the 1576 pathways analyzed, 1322 were mutated in at least three out of 44 samples (Supplementary Table S3) and were submitted to comparative analysis. Twenty-five pathways were significantly more frequently mutated in patients with pCR than in patients without pCR (Table 2, Fig. 3), with mutation rates OR (pCR/no-pCR) ranging from ~ 13 to 4. The “reactome_Formation of a pool of free 40S subunits” showed the strongest correlation with pCR (OR =  ~ 13). Interestingly, 5 of the 94 genes included in this pathway had at least one mutation in the 5 out of 44 samples but different genes were affected in different samples. Similarly, 12 of the 27 genes included in the “HumanCyc_3-phosphoinositide biosynthesis” pathway had at least one mutation in 15 out of 44 samples but different genes were affected in different samples. Consequently, none of the genes alone showed significant association with pCR, conversely to the corresponding whole pathways (Supplementary Figure S1).

Table 2 List of somatic mutation pathways associated with pCR to NACT
Full size table
Fig. 3
figure 3

Pathways more frequently altered in patients with pCR. Dot-plots showing the 25 pathways significantly more frequently mutated in patients with pCR than in patients without pCR (p < 0.05). The pathways are plot tfrom top to bottom by increasing Odds Ratio. The size of the dots represents the number of patients with pCR and pathway alteration and the color represents the p-value as indicated by the color scale

Full size image

Copy number alterations and response to NACT

In our 44 samples, the most frequently gained regions were on 1q, 8p, 8q, 11q, and 17q chromosomal arms, whereas the regions frequently lost were on 1p, 6q, 8p, and 16q. GISTIC analysis identified 30 chromosomal cytobands significantly (q < 0.25) gained/amplified and 14 chromosomal cytobands significantly (q < 0.25) lost/deleted (Fig. 4A). The gained/amplified cytobands comprised 1,014 genes including driver alterations: AKT3, CCND1, EGFR, FGFR1, HER2, MYC, and ZNF703. The two most significant gained/amplified cytobands (17q12 and 11q13.3) were regions classically amplified in breast cancer (q < 1E−05). The lost/deleted cytobands comprised 2,794 genes, including driver genes such as CDKN1A, STK11, BAP1, and ARID1A (Supplementary Table 4). The same analysis was done separately in patients with pCR and in patients without pCR (Fig. 4B). We compared the alteration frequencies of these 3808 genes between complete responders to NACT and non-complete responders (Supplementary Table 5). A total of 234 genes were differentially altered (p < 0.05, Fisher's exact test), including 10 more frequently gained/amplified in patients with pCR, 212 more frequently lost/deleted in patients without pCR, and 12 that were both more frequently gained/amplified in patients with pCR and more frequently lost/deleted in patients without pCR. In order to validate association of these latter 12 gene with the response to NACT, we analyzed the correlation between their mRNA expression level and the pCR/no-pCR status in an independent series of 87 informative IBCs (28 displaying pCR and 59 no-pCR) from the World IBC Consortium [6]. As shown in Figure S2, 8 out of 11 informative genes showed a pCR/no-pCR OR superior to 1, and the fixed-effect model showed a combined OR of 1.38 (95% CI 1.11–1.72), indicating a statistically significant positive effect on pCR (p = 4.4E−03) with no statistically observed heterogeneity among the ORs (I2 = 0.0%, p = 0.646). Three genes, CDC40, FIG4, and AMD1 tended towards significance (p < 0.1).

Fig. 4
figure 4

Profiles of CNAs and correlation with pCR. A GISTIC profiles of CNAs in the whole population (N = 44) and according to pCR. The profiles are plotted as a function of chromosome location. The CNAs are color-coded as follows: gains/amplifications (red), losses/deletions (green). Two significant regions are annotated. B CIN signatures based on CNAs in the whole series (N = 29 informative samples) and comparison between patient with versus without pCR. Only the five signatures with median proportion ≥ 2% in at least one patients’ group are shown

Full size image

We also assessed the distribution of the 17 CIN signatures [29] in the 29 informative samples. As expected, the most represented signatures were the signatures CX1 (median of 54%: associated with chromosome segregation via defective mitosis and/or telomere dysfunction), CX3 (median of 20%: associated with impaired homologous recombination with replication stress and impaired damage sensing) and CX5 (median of 7%: associated with impaired homologous recombination with replication stress) (Fig. 4B). The comparison between complete vs. non-complete responders did not identify any signature differentially represented between both groups (p > 0.05, Wilcoxon test).

Homologous recombination deficiency and response to NACT

The median HRD score was equal to 28 in the 40 informative samples (range 1 to 99) (Fig. 5A). As expected, the score was higher in the TN subtype than in the HR+/HER2− subtype (p = 0.007, Wilcoxon test) and the HER2+ subtype (p = 0.005, Wilcoxon test), and was higher in grade 3 vs grade 2 tumors (p = 0.09, Wilcoxon test; Fig. 5A). The HRD score was higher in the tumors displaying strong COSMIC signature 3 (p = 0.005, Wilcoxon test; Fig. 5B). The median HRD score was slightly higher in patients with pCR (median 30, 1 to 99) than in patients without pCR (median 28, 1 to 78), but the difference was not significant (p = 0.444, Wilcoxon test; Fig. 5C).

Fig. 5
figure 5

HRD score and correlations with pCR. A Box-plot of HRD score in the whole series (N = 40 informative samples), and according to the molecular subtypes and the tumor grade. B Similar to A, but according to the abundance of COSMIC signature S3 (cut-off 10%). C Similar to A, but according to the response to NACT. The p-values are for the Wilcoxon test

Full size image

By using the classical positivity cut-off (score = 42), 28% of samples (11/40) were defined as HRD-high, including 6/8 TN (75%), 2/15 HR+/HER2 (13%), and 3/17 HER2+ (18%; p = 0.005). Patients with HRD-high tumors displayed a 31% pCR rate, similar to those with HRD-low tumors (26%; p = 1, Fisher’s exact test).

Intratumor heterogeneity and clonality and response to NACT

The median heterogeneity index (H-index) was 0.65 in the 43 informative IBC samples (range 0 to 1.93), and 0.66 (range 0 to 1.5) in the patients with pCR versus 0.61 (range 0 to 1.92) in those without pCR (p = 0.989, Wilcoxon test; Fig. 6A). Twenty-three percent of them (10/43) displayed an H-index superior or equal to 1, corresponding to more heterogeneous tumors than the ones with an H-index inferior to 1. The pCR rate was lower in the patients with more heterogeneous tumors than in those with less heterogeneous tumors (10% versus 36%), but the difference was not significant (p = 0.237, Fisher’s exact test; Fig. 6B).

Fig. 6
figure 6

Heterogeneity index and mutational clonality and correlations with pCR. A Box-plot of Heterogeneity H-index in the whole series (N = 43 informative samples), and according to the response to NACT. B Contingency table between the tumor heterogeneity status and the complete vs non-complete responder status. C Box-plot of the percentages of clonal and subclonal mutations in the whole series (N = 42 informative samples) and according to the response to NACT. The p-values are for the Wilcoxon test

Full size image

We then assessed the median percentage of clonal or subclonal non-synonymous mutations in the 42 informative samples (Fig. 6C): the rate was 54% for clonal mutations and 46% for subclonal mutations. The percentage of clonal mutations (reported to the total number of non-synonymous mutations) was not different between the complete responders and the non-complete responders (59% in pCR vs 54% in no-pCR; p < 0.05, Wilcoxon exact), nor the percentage of subclonal mutations (41% in pCR vs 46% in no-pCR; p < 0.05, Wilcoxon exact; Fig. 6C).

Discussion

We defined the WES profiles of untreated primary tumors of 44 IBCs and compared the profiles of patients displaying pCR after NACT versus those of patients not displaying pCR. To our knowledge, this is the first series of this type published in the literature. To avoid DNA biases induced by previous treatments or resistance mechanisms [34, 35], we analyzed untreated primary tumors only. For each patient, normal DNA was sequenced. The pCR rate observed in our IBC series (30%) was consistent with the literature data in IBC, and the classical predictive clinicopathological tumor features (grade, HR, and HER2 statutes) confirmed their predictive value, even if non-significantly probably because of the sample size.

The TMB was similar to the one reported in our whole series of 54 IBC [20] and to the one reported in the literature in non-IBC [36]. It was not differential between the complete responders and the non-complete responders to chemotherapy, whereas it has been associated with response to immune therapy [37]. In our series, no patient had received neoadjuvant immune checkpoint inhibitor. Such absence of correlation between TMB and pathological response to NACT was already reported in series including all molecular subtypes like here [8, 11, 16], whereas a correlation was reported in series dedicated to TN subtype only [14, 15, 17]. In our series, the number of samples was too small to search for correlation per molecular subtype.

As expected in BC, the most frequently mutated genes were TP53 as tumor suppressor gene and PIK3CA as oncogene. TP53 was more frequently mutated in the complete responders than in non-complete responders (69% versus 48% respectively) as previously reported [38, 39], but the difference was not significant likely because of the small number of patients. A total of 220 patients would be needed to found this difference as significant with a statistical power of 80%. PIK3CA was more frequently mutated in the patients with pCR (38%) than in those without pCR (19%), and the difference was not significant (p = 0.255, Fisher’s exact test). In the literature, the unfavorable predictive role of PIK3CA mutation on pCR is clear in HER2+ BCs after anti-HER2-based NACT [12, 16, 40, 41]. For example, in a meta-analysis of 22 studies covering 3,361 HER2+ patients treated with NACT and anti-HER2 therapy in the neoadjuvant setting [41], PIK3CA mutation was significantly associated with a lower pCR rate (OR = 0.23, 95% CI 0.19–0.27, p < 0.001). This association remained significant irrespective of the type of antiHER2 therapy (single-agent or dual-agent) and hormone receptor status, and there was no evidence of inter-study heterogeneity. Our observation is not in agreement with these results, but the number of samples is too low in our series before concluding that it is related to IBC phenotype, and only one third of HER2+ patients had received anti-HER2 drug in the neoadjuvant setting. Only two genes, NLRP3 [42, 43] and SLC9B1, were significantly more frequently mutated in the complete responders than in non-complete responders (23% vs. 0%). The NLRP3 inflammasome is involved in antitumor immunity. Chemotherapy treatment of cancer cells (e.g., with anthracyclines) causes cells to release ATP, thus activating the NLRP3 inflammasome and the IL-1β/IL-1 receptor (IL1R) signaling axis in dendritic cells (DCs), then activating CD8+ T cells [44]. The germline mutations were too rare and our series was too small to find any significant association with pCR to NACT.

Regarding the proportions of possible base substitutions in SNVs, no difference was observed according to the degree of response to NACT. By contrast, the proportion of C > T substitution was lower in the responsive group (~ 39%) compared with the nonresponsive group (49%), especially when the substitution site was flanked by C and G [11]. The pattern of COSMIC signatures has been associated to response to chemotherapy in different cancers such as BC [8, 11] and bladder carcinoma [45]. We observed a higher abundance of signature 7 in tumors from complete responders. This signature, likely due to ultraviolet light exposure, has been found predominantly in skin cancers. The link with response to chemotherapy is not clear. Because our series is small, further research into these mutational spectra and signatures is needed before any conclusion.

Because the scarcity of most gene mutations limits the statistical power to detect significant clinical association and because pathway-level alterations may be more important than recurrent single gene mutations, we assessed mutations at the pathway level. Significant correlations with 25 pathways were found, whereas none of the included genes alone showed significant correlation. These pathways corresponded to major biological processes and were more frequently mutated in patients with pCR than in patients without pCR. Several of them were related to protein translation (“reactome_Formation of a pool of free 40S subunits”, “reactome_3'-UTR-mediated translational regulation”, “reactome_Cap-dependent Translation Initiation”, “reactome_GTP hydrolysis and joining of the 60S ribosomal subunit”), to PI3K (“HumanCyc_superpathway of inositol phosphate compounds”, “HumanCyc_3-phosphoinositide biosynthesis”, “Synthesis of PIPs at the plasma membrane”, “KEGG_inositol phosphate metabolism”), signal transduction (“PANTHER_EGF receptor signaling pathway”, “PANTHER_FGF signaling pathway”, “reactome_Signaling by ERBB2″,” reactome_Signaling by FGFR”), to angiogenesis (“PANTHER_VEGF signaling pathway”, “pid_Signaling events mediated by VEGFR1 and VEGFR2″), and to immunity (“pid_IL2-mediated signaling events”). These different processes might lead or testify to higher proliferation of cancer cells, and higher immune response against tumor cells, all features known to be associated with better chemosensitivity. For example, differences for ribosomal genes between patients with and without pCR were recently reported by some of us in an independent clinical 85-sample series of IBC [46]. The differential expression of ribosomal proteins and translation factors involved in protein synthesis in response to stress can impair cell survival, cell proliferation, DNA repair during cellular stress, accumulation of reactive oxygen species, and apoptotic sensitivity to therapeutic agents. In the literature, we identified four WES-based studies of non-IBC samples [9, 13, 14, 16] that searched for correlation between pCR and alteration of biological pathways. None of them identified correlation with ribosomal or protein translation pathways. By contrast, we recently observed in IBC that genes associated with ribosomal processes were significantly enriched amongst the group of genes overexpressed in IBC patients with pCR [46]. We thus completed this observation in a large series of 1203 patients with non-IBCs and treated with NACT: in multivariate analysis including the molecular subtypes, a ribosomal activation score based on the “leading-edge genes” was positively associated with achievement of pCR, as observed in IBC (p = 2.56E−03, (glm function, logit link; personal data). Higher alteration of the PIK3 pathway in patients with pCR was already reported by others in non-IBC. Basho et al. [13] analysed a series of 177 TNBC patients treated with an anthracycline-taxane-based NACT using WES; they reported pCR rates of 46% in patients with PIK3 pathway alteration versus 36% in those without PIK3 pathway alteration (p = 0.2). Similarly, Blenman et al. [14] analysed a series of 57 TNBC patients treated with anthracycline-taxane-based NACT coupled with durvalumab using WES; they found higher mutation rates in the PIK3 pathway in patients with pCR (96%) than in patients without (77%; p = 0.03). By contrast, by analysing WES from 67 HER2+ patients treated with trastuzumab in the neo-ALTTO trial [16], Shi et al. showed that mutations in the PIK3 pathway were associated with lower pCR rate as compared to no mutation (4% versus 56%, p < 0.001). Immune pathways were also enriched in patients with pCR in the Blenman et al.’s study [14], as we observed in our series with the “pid_IL2-mediated signalling events” pathway.

GISTIC analysis identified 30 recurrently gained/amplified regions (1014 genes) and 14 recurrently lost/deleted regions (2794 genes), including regions and genes classically altered in BC. A total of 234 of these 3808 genes were differentially altered between the complete responders to NACT and the non-complete responders. HER2 was not included but tended to be more frequently gained/amplified in the complete versus non-complete responders (p = 0.099) as reported [47]. Among the 234 significant genes, 10 were more frequently gained/amplified in patients with pCR, and 212 were more frequently lost/deleted in patients without pCR, notably CCNC, PRDM1, and TNFAIP3. A recent pre-clinical study showed that the depletion of CCNC (cyclin C) decreased mitochondrial ROS production and reduced cell apoptosis under cisplatin treatment [48]. In a study of TNBC treated with NACT, amplification of PRDM1, a regulator of both B-cell and T-cell differentiation, in pre-treatment samples significantly characterized the tumors of patients with pCR [49]. A20/TNFAIP3 is involved in the regulation of DNA damage response and the cancer cell resistance to DNA-damaging therapy [50]. Interestingly, 12 out of 234 significant genes were more frequently gained/amplified in patients with pCR and/or more frequently lost/deleted in patients without pCR. This correlation was validated for most of them at the gene expression level in an independent series of 87 IBC samples treated with NACT. The positive odds ratios were found as homogeneous between genes. These 12 genes included CDC40, FIG7, and AMD1. CDC40 protein has been suggested to control cell cycle through splicing of intron-containing pre-mRNAs that encode proteins important for cell cycle progression [51]. FIG4 encodes for SAC3 phosphatase, a protein involved in TNBC cell proliferation [52]. AMD1 or S-adenosylmethionine (AdoMet) decarboxylase is a key enzyme in polyamine biosynthesis, which promotes breast cancer cell proliferation [53, 54]. Their positive correlation with chemosensitivity might be related to their role in cell proliferation. As expected for BC samples, the most represented CIN signatures [29] were the signatures CX1, CX3, and CX5, but no signature was differentially represented between the tumor samples from complete versus non-complete responders. More samples are needed to address this issue that has never been tested in BC.

The median HRD score was based on the three classical components provided by WES (LOH, TAI, LST). As expected, it was associated with aggressive tumor (TN molecular subtype and grade 3), and tumors enriched in COSMIC signature 3. However, even if the HRD score was slightly higher in patients with pCR than in patients without pCR, the difference was not significant in our series. We identified two studies in the literature that searched for correlations between the HRD score and the pathological response to NACT in non-IBC. The same result as ours was reported in a small series of 21 TNBCs, but the difference was not significant (p = 0.27) likely because of the small series size [17]. In a lager series of 94 non-IBCs including all molecular subtypes and treated with NACT (paclitaxel followed by FEC), Kim et al. reported higher pCR rate (39.1% versus 9.9%; p = 0.003) to NACT in patients with HRD-high BCs than in patients with HRD-low BCs [10]. Clearly, because of the small size of our series, analysis of larger IBC sample cohort is required before drawing any conclusion.

Intratumor heterogeneity is a prevalent feature in many cancer types and represents a considerable challenge to optimize prognosis and treatment. In our series, the heterogeneity H-index was not associated with achievement of pCR. However, even if not significant, the patients with heterogeneous tumors (H-index ≥ 1) displayed a lower pCR rate (11%) than those with less heterogeneous tumors (35%). In their small series of 21 TNBCs, Zhu et al. reported higher pCR rate in the patients with high intratumor heterogeneity as compared to those with low intratumor heterogeneity (40% versus 17%; p = 0.36). Here too, the difference was not significant likely because of the small series size [17]. In an ancillary study of neo-ALTTO trial including 203 HER2+ non-IBCs [16], the mean clonal heterogeneity score was significantly higher in cases without pCR compared with pCR, suggesting that greater clonal heterogeneity is associated with greater resistance to therapy. Our result in IBCs is in agreement with those reported in non-IBCs, even if non-significant very likely because of the small number of patients. Whether such observation is related to the higher number of potentially resistant subclones within heterogeneous tumors deserves investigation. The percentages of clonal mutations and subclonal mutations were not associated with the pCR in our series.

Our study suffers from a few limitations. First: its retrospective nature induces biases such as the selection bias due to our inclusion criteria. Because of the disease scarcity and the difficulty to obtain and to profile small pre-treatment diagnostic biopsy samples, we had to select patients over a long period study (16 years) during which the NACT regimens changed. For example, only one third of HER2+ patients received neoadjuvant trastuzumab that became standard practice from 2006. Second, the small number of samples precluded statistical correction for multiple tests in all comparative analyses. In the pathway analysis concerning 1,576 biological pathways, we applied random permutation of the pCR/no-pCR label across 1000 iterations and recalculated the OR for each pathway. The proportion of random permutations showing an OR greater than the OR of the actual data defined the p-value. But we acknowledge that our approach may have led to false positives in the absence of actual multiple testing correction. Furthermore, the small sample size decreases the statistical power of our findings likely leading to false negatives by impeding the significant detection of actual differences, as shown for TP53 and PIK3CA mutations. The Supplementary Table 6 shows the genome alterations that we found as significant with the p-value and the statistical power of our analysis using the 44-sample set, as well as the theoretical number of samples needed to demonstrate the same differences with a power of 80%. Third, no profiling of the residual tumor after NACT in the patients without pCR was done. That might have been interested to track the cancer clones resistant to NACT.

Conclusions

We report here the first study aiming at identifying correlations between the WES data of untreated IBC tumor samples and the achievement of pCR to NACT. Global genomic scores such as TMB, HRD, and H-index were not significantly different between the patients with pCR and those without pCR. Analysis of somatic mutations identified only two genes more frequently mutated in the complete responder patients, but several significant biological pathways involved in protein translation, PI3K pathway, or signal transduction. Several CNAs were also found as differentially observed between both patients’ groups. Besides its originality, our study displays several strengths: largest WES study of IBCs, consensual definition of IBCs, previously untreated primary tumors, profiling of matched normal samples, and homogeneous definition of pCR. But of course, and because of the limitations above-quoted, notably the low number of patients, these results must be regarded as descriptive and hypothesis generating in terms of clinical and biological importance, and thus require validation in larger and independent series in future studies. Additional analyses, both functional and clinical, in larger series are warranted. The results could help better understanding the mechanisms of response/resistance to NACT and better personalizing the neoadjuvant systemic treatment in IBC. First, the patients of the predicted no-pCR group might be enrolled in clinical trials testing non-anthracycline/taxane-based regimens. Second, the functional validation of predictive gene alterations could lead to the development of new therapies combined with chemotherapy and able to avoid the resistance.

Availability of data and materials

All clinicopathological data and genomic data analyzed in the present study will be available after paper acceptance upon reasonable request.

Abbreviations

AJCC:

American Joint Committee on Cancer

CNAs:

Copy number alteration

HR+:

Hormone receptor-positive

HRD:

Homologous recombination deficiency

IBC:

Inflammatory breast cancer

IHC:

Immunohistochemistry

Mb:

Megabase

non-IBC:

Non-inflammatory breast cancer

NACT:

Neoadjuvant chemotherapy

pCR:

Pathological complete response

SNV:

Single nucleotide variant

TMB:

Tumor mutational burden

TN:

Triple-negative

WES:

Whole-exome sequencing

References

  1. Dawood S, Merajver SD, Viens P, Vermeulen PB, Swain SM, Buchholz TA, et al. International expert panel on inflammatory breast cancer: consensus statement for standardized diagnosis and treatment. Ann Oncol. 2011;22:515–23.

    Article  CAS  PubMed  Google Scholar 

  2. Viens P, Tarpin C, Roche H, Bertucci F. Systemic therapy of inflammatory breast cancer from high-dose chemotherapy to targeted therapies: the French experience. Cancer. 2010;116:2829–36.

    Article  CAS  PubMed  Google Scholar 

  3. Charafe-Jauffret E, Tarpin C, Viens P, Bertucci F. Defining the molecular biology of inflammatory breast cancer. Semin Oncol. 2008;35:41–50.

    Article  PubMed  Google Scholar 

  4. Bertucci F, Finetti P, Vermeulen P, Van Dam P, Dirix L, Birnbaum D, et al. Genomic profiling of inflammatory breast cancer: a review. Breast. 2014;23:538–45.

    Article  PubMed  Google Scholar 

  5. Bertucci F, Finetti P, Birnbaum D, Viens P. Gene expression profiling of inflammatory breast cancer. Cancer. 2010;116:2783–93.

    Article  CAS  PubMed  Google Scholar 

  6. Bertucci F, Ueno NT, Finetti P, Vermeulen P, Lucci A, Robertson FM, et al. Gene expression profiles of inflammatory breast cancer: correlation with response to neoadjuvant chemotherapy and metastasis-free survival. Ann Oncol. 2014;25:358–65.

    Article  CAS  PubMed  Google Scholar 

  7. Van Laere S, Ueno NT, Finetti P, Vermeulen PB, Lucci A, Robertson F, et al. Uncovering the molecular secrets of inflammatory breast cancer biology: an integrated analysis of three distinct affymetrix gene expression data sets. Clin Cancer Res. 2013;19:4685–96.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Denkert C, Untch M, Benz S, Schneeweiss A, Weber KE, Schmatloch S, et al. Reconstructing tumor history in breast cancer: signatures of mutational processes and response to neoadjuvant chemotherapy (small star, filled). Ann Oncol. 2021;32:500–11.

    Article  CAS  PubMed  Google Scholar 

  9. Dong M, Shan B, Han X, Zhao X, Wang F, Zhu L, et al. Baseline mutations and up-regulation of PI3K-AKT pathway serve as potential indicators of lack of response to neoadjuvant chemotherapy in stage II/III breast cancer. Front Oncol. 2021;11: 784985.

    Article  CAS  PubMed  Google Scholar 

  10. Kim SJ, Sota Y, Naoi Y, Honma K, Kagara N, Miyake T, et al. Determining homologous recombination deficiency scores with whole exome sequencing and their association with responses to neoadjuvant chemotherapy in breast cancer. Transl Oncol. 2021;14: 100986.

    Article  CAS  PubMed  Google Scholar 

  11. Yin G, Liu L, Yu T, Yu L, Feng M, Zhou C, et al. Genomic and transcriptomic analysis of breast cancer identifies novel signatures associated with response to neoadjuvant chemotherapy. Genome Med. 2024;16:11.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Zagami P, Fernandez-Martinez A, Rashid NU, Hoadley KA, Spears PA, Curigliano G, et al. Association of PIK3CA mutation with pathologic complete response and outcome by hormone receptor status and intrinsic subtype in early-stage ERBB2/HER2-positive breast cancer. JAMA Netw Open. 2023;6: e2348814.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Basho RK, Zhao L, White JB, Huo L, Bassett RL, Mittendorf EA, et al. Comprehensive analysis identifies variability in PI3K pathway alterations in triple-negative breast cancer subtypes. JCO Precis Oncol. 2024;8: e2300124.

    Article  PubMed  Google Scholar 

  14. Blenman KRM, Marczyk M, Karn T, Qing T, Li X, Gunasekharan V, et al. Predictive markers of response to neoadjuvant durvalumab with nab-paclitaxel and dose-dense doxorubicin/cyclophosphamide in basal-like triple-negative breast cancer. Clin Cancer Res. 2022;28:2587–97.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Karn T, Denkert C, Weber KE, Holtrich U, Hanusch C, Sinn BV, et al. Tumor mutational burden and immune infiltration as independent predictors of response to neoadjuvant immune checkpoint inhibition in early TNBC in GeparNuevo. Ann Oncol. 2020;31:1216–22.

    Article  CAS  PubMed  Google Scholar 

  16. Shi W, Jiang T, Nuciforo P, Hatzis C, Holmes E, Harbeck N, et al. Pathway level alterations rather than mutations in single genes predict response to HER2-targeted therapies in the neo-ALTTO trial. Ann Oncol. 2017;28:128–35.

    Article  CAS  PubMed  Google Scholar 

  17. Zhu Y, Zhang H, Pan C, He G, Cui X, Yu X, et al. Integrated tumor genomic and immune microenvironment analysis identifies predictive biomarkers associated with the efficacy of neoadjuvant therapy for triple-negative breast cancer. Cancer Med. 2023;12:5846–58.

    Article  CAS  PubMed  Google Scholar 

  18. Goh G, Schmid R, Guiver K, Arpornwirat W, Chitapanarux I, Ganju V, et al. Clonal evolutionary analysis during HER2 blockade in HER2-positive inflammatory breast cancer: a phase II open-label clinical trial of afatinib +/- vinorelbine. PLoS Med. 2016;13: e1002136.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Luo R, Chong W, Wei Q, Zhang Z, Wang C, Ye Z, et al. Whole-exome sequencing identifies somatic mutations and intratumor heterogeneity in inflammatory breast cancer. NPJ Breast Cancer. 2021;7:72.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Bertucci F, Lerebours F, Ceccarelli M, Guille A, Syed N, Finetti P, et al. Mutational landscape of inflammatory breast cancer. J Transl Med. 2024;22:374.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Bertucci F, Finetti P, Guille A, Adelaide J, Garnier S, Carbuccia N, et al. Comparative genomic analysis of primary tumors and metastases in breast cancer. Oncotarget. 2016;7:27208–19.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–60.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011;43:491–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Cibulskis K, Lawrence MS, Carter SL, Sivachenko A, Jaffe D, Sougnez C, et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat Biotechnol. 2013;31:213–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Kim S, Scheffler K, Halpern AL, Bekritsky MA, Noh E, Kallberg M, et al. Strelka2: fast and accurate calling of germline and somatic variants. Nat Methods. 2018;15:591–4.

    Article  CAS  PubMed  Google Scholar 

  26. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38: e164.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, Getz G. GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biol. 2011;12:R41.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Shen R, Seshan VE. FACETS: allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing. Nucleic Acids Res. 2016;44: e131.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Drews RM, Hernando B, Tarabichi M, Haase K, Lesluyes T, Smith PS, et al. A pan-cancer compendium of chromosomal instability. Nature. 2022;606:976–83.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Marquard AM, Eklund AC, Joshi T, Krzystanek M, Favero F, Wang ZC, et al. Pan-cancer analysis of genomic scar signatures associated with homologous recombination deficiency suggests novel indications for existing cancer drugs. Biomark Res. 2015;3:9.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Telli ML, Timms KM, Reid J, Hennessy B, Mills GB, Jensen KC, et al. Homologous Recombination Deficiency (HRD) score predicts response to platinum-containing neoadjuvant chemotherapy in patients with triple-negative breast cancer. Clin Cancer Res. 2016;22:3764–73.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Miller CA, White BS, Dees ND, Griffith M, Welch JS, Griffith OL, et al. SciClone: inferring clonal architecture and tracking the spatial and temporal patterns of tumor evolution. PLoS Comput Biol. 2014;10: e1003665.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539–58.

    Article  PubMed  Google Scholar 

  34. McGranahan N, Favero F, de Bruin EC, Birkbak NJ, Szallasi Z, Swanton C. Clonal status of actionable driver events and the timing of mutational processes in cancer evolution. Sci Transl Med. 2015;7: 283ra54.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Bertucci F, Ng CKY, Patsouris A, Droin N, Piscuoglio S, Carbuccia N, et al. Genomic characterization of metastatic breast cancers. Nature. 2019;569:560–4.

    Article  CAS  PubMed  Google Scholar 

  36. Kandoth C, McLellan MD, Vandin F, Ye K, Niu B, Lu C, et al. Mutational landscape and significance across 12 major cancer types. Nature. 2013;502:333–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Valero C, Lee M, Hoen D, Wang J, Nadeem Z, Patel N, et al. The association between tumor mutational burden and prognosis is dependent on treatment context. Nat Genet. 2021;53:11–5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Chen MB, Zhu YQ, Xu JY, Wang LQ, Liu CY, Ji ZY, et al. Value of TP53 status for predicting response to neoadjuvant chemotherapy in breast cancer: a meta-analysis. PLoS ONE. 2012;7: e39655.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Yang L, Ye F, Bao L, Zhou X, Wang Z, Hu P, et al. Somatic alterations of TP53, ERBB2, PIK3CA and CCND1 are associated with chemosensitivity for breast cancers. Cancer Sci. 2019;110:1389–400.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Jank P, Karn T, van Mackelenbergh M, Lindner J, Treue D, Huober J, et al. An analysis of PIK3CA hotspot mutations and response to neoadjuvant therapy in breast cancer patients from four prospective clinical trials. Clin Cancer Res. 2024;30:3868–80.

    Article  PubMed  Google Scholar 

  41. Chen H, Hu X, Wang D, Wang Y, Yu Y, Tao H. Association of PIK3CA mutation with outcomes in HER2-positive breast cancer treated with anti-HER2 therapy: a meta-analysis and bioinformatics analysis of TCGA-BRCA data. Transl Oncol. 2023;37: 101738.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Karki R, Man SM, Kanneganti TD. Inflammasomes and cancer. Cancer Immunol Res. 2017;5:94–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Huang Y, Wang H, Hao Y, Lin H, Dong M, Ye J, et al. Myeloid PTEN promotes chemotherapy-induced NLRP3-inflammasome activation and antitumour immunity. Nat Cell Biol. 2020;22:716–27.

    Article  CAS  PubMed  Google Scholar 

  44. Lee PH, Yamamoto TN, Gurusamy D, Sukumar M, Yu Z, Hu-Li J, et al. Host conditioning with IL-1beta improves the antitumor function of adoptively transferred T cells. J Exp Med. 2019;216:2619–34.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Liu D, Abbosh P, Keliher D, Reardon B, Miao D, Mouw K, et al. Mutational patterns in chemotherapy resistant muscle-invasive bladder cancer. Nat Commun. 2017;8:2193.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Devi GR, Pai P, Lee S, Foster MW, Sannareddy DS, Bertucci F, et al. Altered ribosomal profile in acquired resistance and reversal associates with pathological response to chemotherapy in inflammatory breast cancer. NPJ Breast Cancer. 2024;10:65.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Greenwell K, Hussain L, Lee D, Bramlage M, Bills G, Mehta A, et al. Complete pathologic response rate to neoadjuvant chemotherapy increases with increasing HER2/CEP17 ratio in HER2 overexpressing breast cancer: analysis of the National Cancer Database (NCDB). Breast Cancer Res Treat. 2020;181:249–54.

    Article  CAS  PubMed  Google Scholar 

  48. Jiang HY, Chen YL, Xu XX, Li CY, Chen Y, Li DP, et al. Ubiquitylation of cyclin C by HACE1 regulates cisplatin-associated sensitivity in gastric cancer. Clin Transl Med. 2022;12: e770.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. Silvestri M, Dugo M, Vismara M, De Cecco L, Lanzoni D, Vingiani A, et al. Copy number alterations analysis of primary tumor tissue and circulating tumor cells from patients with early-stage triple negative breast cancer. Sci Rep. 2022;12:1470.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  50. Yang C, Zang W, Tang Z, Ji Y, Xu R, Yang Y, et al. A20/TNFAIP3 regulates the DNA damage response and mediates tumor cell resistance to DNA-damaging therapy. Cancer Res. 2018;78:1069–82.

    Article  CAS  PubMed  Google Scholar 

  51. Ben-Yehuda S, Russell CS, Dix I, Beggs JD, Kupiec M. Extensive genetic interactions between PRP8 and PRP17/CDC40, two yeast genes involved in pre-mRNA splicing and cell cycle progression. Genetics. 2000;154:61–71.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Ikonomov OC, Filios C, Sbrissa D, Chen X, Shisheva A. The PIKfyve-ArPIKfyve-Sac3 triad in human breast cancer: functional link between elevated Sac3 phosphatase and enhanced proliferation of triple negative cell lines. Biochem Biophys Res Commun. 2013;440:342–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Liao R, Chen X, Cao Q, Bai L, Ma C, Dai Z, et al. AMD1 promotes breast cancer aggressiveness via a spermidine-eIF5A hypusination-TCF4 axis. Breast Cancer Res. 2024;26:70.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. Gao H, Li H, Wang J, Xu C, Zhu Y, Tuluhong D, et al. Polyamine synthesis enzyme AMD1 is closely related to the tumorigenesis and prognosis of human breast cancer. Exp Cell Res. 2022;417: 113235.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We thank the computing facilities DISC (Datacenter IT and Scientific Computing, CRCM) and the Institut Paoli-Calmettes biobank (authorization number AC-2018-1905) for their support.

Funding

This work was supported by Ligue Nationale Contre Le Cancer (EL2022/FB), and Association Ruban Rose (Prix Ruban Rose 2020, FB).

Author information

Author notes

  1. Arnaud Guille, Florence Lerebours, Michele Ceccarelli, Najeeb Syed are equal second authors.

  2. Céline Callens, Davide Bedognetti, Emilie Mamessier are equal last authors.

Authors and Affiliations

  1. Predictive Oncology Laboratory, Centre de Recherche en Cancérologie de Marseille (CRCM), Inserm, U1068, CNRS UMR7258, Institut Paoli-Calmettes, Aix-Marseille Université, 232, Boulevard de Sainte-Marguerite, 13009, Marseille, France

    François Bertucci, Arnaud Guille, José Adélaïde, Pascal Finetti, Daniel Birnbaum & Emilie Mamessier

  2. Department of Medical Oncology, Institut Paoli-Calmettes, Aix-Marseille Université, Marseille, France

    François Bertucci, Patrice Viens, Alexandre De Nonneville & Anthony Goncalves

  3. Department of Medical Oncology, Institut Curie Saint-Cloud, Paris, France

    Florence Lerebours & Céline Callens

  4. Sylvester Comprehensive Cancer Center, University of Miami, Miami, USA

    Michele Ceccarelli

  5. Department of Public Health Sciences, University of Miami, Miami, USA

    Michele Ceccarelli

  6. University of Hawai’i Cancer Center, Honolulu, HI, USA

    Najeeb Syed & Naoto T. Ueno

  7. Center for Oncological Research (CORE), Integrated Personalized and Precision Oncology Network (IPPON), University of Antwerp, Universiteitsplein 1, Wilrijk, Belgium

    Steven Van Laere

  8. Tumor Biology and Immunology Laboratory, Research Branch, Sidra Medicine, Doha, Qatar

    Davide Bedognetti

Authors
  1. François Bertucci
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Arnaud Guille
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Florence Lerebours
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Michele Ceccarelli
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Najeeb Syed
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. José Adélaïde
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Pascal Finetti
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Naoto T. Ueno
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Steven Van Laere
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Patrice Viens
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Alexandre De Nonneville
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Anthony Goncalves
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Daniel Birnbaum
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Céline Callens
    View author publications

    You can also search for this author in PubMed Google Scholar

  15. Davide Bedognetti
    View author publications

    You can also search for this author in PubMed Google Scholar

  16. Emilie Mamessier
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Conceptualization, design, and supervision: FB. Data acquisition, analysis and interpretation: all authors. Writing of manuscript: FB. Review and editing of manuscript: FB, FL, MC, AdN, AG, NU, SVL, DBi, CC, DBe, EM. Reading and approval of the final manuscript: all authors. Funding acquisition: FB, FL, MC, DBi, CC, DB, EM.

Corresponding author

Correspondence to François Bertucci.

Ethics declarations

Ethics approval and consent to participate

The study was performed in accordance with the Declaration of Helsinki. Each patient had given written informed consent for somatic and constitutional genomic analysis. The study was approved by the Institut Paoli-Calmettes review board and the Institut Curie ethics committee.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no conflict of interest related to this paper.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary Material 1: Table S1. Univariate analysis for pCR in the 44 IBC samples

Supplementary Material 2: Table S2. List of somatic mutations in the 44 IBC samples

12967_2024_5790_MOESM3_ESM.xlsx

Supplementary Material 3: Table S3. List of 1576 biological pathways collected from the HumanCyc, KEGG, PANTHER, pid, and Reactome databases

12967_2024_5790_MOESM4_ESM.pdf

Supplementary Material 4: Figure S1. Mutations in genes included in the three pathways most significantly mutated in patients with pCR to NACT. Top: somatic gene mutations. The genes are ordered from top to bottom by decreasing percentage of mutated IBC samples. Bottom: pCR/no-pCR groups, mutated/no-mutated groups, and IHC-based molecular subtypes are color-coded as indicated in the legend. The percentages of gene mutation in tumor samples are shown to the right of the Oncoprint

12967_2024_5790_MOESM5_ESM.xlsx

Supplementary Material 5: Table S4. List of gained/amplified and lost/deleted regions according to the GISTIC analysis of CNAs

12967_2024_5790_MOESM6_ESM.xlsx

Supplementary Material 6: Table S5. Comparison between patients with pCR and patients without pCR of alteration frequencies of genes included in the GISTIC regions

12967_2024_5790_MOESM7_ESM.pdf

Supplementary Material 7: Figure S2. Correlation of gene expression level and the pathological response to NACT. Forest plot of ORfor pCR/no-pCR in the World IBC Consortium datasetfor mRNA expression level of 11 genes that, in our present series, were both more frequently gained/amplified in patients with pCR and more frequently lost/deleted in patients without pCR IBCs. Analysis of homogeneity of OR between genes in a fixed-effects model revealed homogeneity in term of correlation with pCR/no-pCR status

12967_2024_5790_MOESM8_ESM.xlsx

Supplementary Material 8: Table S6. List of genome alterations found as significant. For each alteration, the p-value and the statistical power of our analysis using the 44-sample set are shown, as well as the theoretical number of samples needed to demonstrate the same differences with a power of 80%

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Bertucci, F., Guille, A., Lerebours, F. et al. Whole-exome profiles of inflammatory breast cancer and pathological response to neoadjuvant chemotherapy. J Transl Med 22, 969 (2024). https://doi.org/10.1186/s12967-024-05790-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12967-024-05790-8

Keywords

Journal of Translational Medicine

ISSN: 1479-5876

Contact us