三阴性乳腺癌的雌激素受体、孕激素受体、HER2均为阴性,对传统的内分泌治疗和HER2靶向治疗无效,过去主要依靠化疗,而且疗效不持久。近年来,化疗联合免疫治疗已被证实对三阴性乳腺癌疗效持久,但是对部分患者仍然无效或疗效不持久。其中,化疗通常被作为主角,免疫治疗通常被作为配角,联合治疗效果不佳通常由免疫治疗背锅下岗。
2026年1月10日,施普林格自然旗下生物医学中心出版社《分子癌症》在线发表法国马赛大学的研究报告,发现三阴性乳腺癌存在四种免疫分子分型,而且不同分型对化疗联合免疫治疗的效果截然相反,某些分型不用化疗单用免疫治疗效果反而更好,这直接挑战了当前一刀切的治疗模式。
该研究首先对免疫功能正常的三阴性乳腺癌转基因小鼠进行多参数分析,结合组织学、基因组学、转录组学、蛋白质组学和免疫组化,并将结果与人类三阴性乳腺癌队列患者数据和三阴性乳腺癌肿瘤基因大数据进行比较。随后,为了进行治疗测试和功能验证,建立来自原发肿瘤的同源移植模型,并用其评价化疗药物表柔比星联合免疫检查点抑制剂帕博利珠单抗的效果。
结果发现,根据不同的内在(基因)和外在(免疫)特征,三阴性乳腺癌可被分为四个不同亚型,与患者亚型(包括罕见的化生型)以及临床结局密切相关。
其中一类如同中性粒细胞军团:肿瘤里塞满中性粒细胞,就像横冲直撞的杂牌军,单用免疫治疗根本无效,必须先用化疗炸弹炸开防线。另二类如同巨噬细胞城堡:巨噬细胞像城墙一般包裹肿瘤,反而对单用免疫治疗敏感,就像城堡里有内应,化疗强攻反而坏事。
更令人震撼的是,把这些相同基因的肿瘤移植到其他小鼠身上,连续传了7代,结果每次长出来的肿瘤,免疫特征居然像指纹一样稳定,这说明这些肿瘤的性格是先天的,不是后天养成的。原发肿瘤的基因和免疫特征在连续同基因移植都得以保留,证实该临床前平台的转化价值。
此外,该研究小鼠模型其中40%是罕见的肉瘤样癌,这种人类预后极差的亚型,以往由于样本太少根本无法研究。而现在,科学家居然在实验室里定制出这种癌症,这意味着我们正在从追着癌症跑变成等着癌症来。
因此,该研究结果表明,三阴性乳腺癌的治疗效果,受到肿瘤基因特征和免疫特征相互作用的影响,为精准免疫肿瘤学提供了强大的临床前平台,能够对三阴性乳腺癌患者进行分层,从而制定个性化的肿瘤免疫治疗。同样的治疗方案,在不同免疫环境下效果截然相反,就像给沙漠和雨林浇同样的水,结果完全不一样。以前认为免疫细胞多就是好事,现在发现中性粒细胞反而是伪积极分子,故有必要进一步将小鼠实验室研究结果转化到人类临床研究。当前90%的癌症药物临床研究失败,因为我们用平均主义对付极端个性化的敌人,该研究如同给三阴性乳腺癌治疗安装导航系统:先通过活检确定肿瘤属于哪类免疫性格,再决定派坦克还是伞兵。对产业界的启示更直接:未来药企的核心竞争力,可能不再是单一药物,而是诊断+治疗的综合解决方案,就像卖咖啡机不能只提供机器,还要提供个性化的咖啡。将来癌症治疗的关键词,不再是特效药,而是匹配度,就像找对象,没有最好的人,只有最对的相处模式。
Mol Cancer. 2026 Jan 10. IF: 33.9
Multi-omics profiling uncovers immune-molecular clusters with distinct chemo-immunotherapeutic vulnerabilities in a mouse model of triple-negative breast cancer.
Castellanet O, Monatte J, Corvaisier N, Kaoutari AE, Kaya M, Ferreira L, Audebert S, Camoin L, de Nonneville A, Goncalves A, Borg JP, Michea P, Maina F, Lamballe F.
Aix Marseille Univ, Institut Paoli-Calmettes, Centre de Recherche en Cancérologie de Marseille, Marseille, France; Institut Universitaire de France, Paris, France; Aix Marseille Univ, Turing Center for Living Systems, Marseille, France.
BACKGROUND: Triple-negative breast cancer (TNBC) is a highly aggressive and heterogeneous breast cancer subtype with limited treatment options. Predicting patient response to chemo-immunotherapy remains challenging, highlighting the need for robust stratification strategies.
METHODS: We performed a multi-parametric analysis combining histological, genomic, transcriptomic, proteomic, and immune profiling in the immunocompetent MMTV-R26Met TNBC mouse model and compared outcomes with patient data from human TNBC cohorts and TNBC tumor microarray. To enable therapeutic testing and functional validation, we established syngeneic grafts from primary tumors and used them to evaluate combined chemotherapy (epirubicin) and anti-PD-1 immunotherapy.
RESULTS: Multi-parametric analysis of TNBC heterogeneity modeled by the MMTV-R26Met mice identified four distinct TNBC clusters, defined by unique intrinsic (molecular/genomic) and extrinsic (immune) features, which closely parallel patient subtypes, including rare metaplastic forms, and correlate with clinical outcomes. Both intrinsic and immune hallmarks of primary tumors were conserved across serial syngeneic transplantations, confirming the translational value of this preclinical platform. Treatment assessments indicated cluster-specific therapeutic vulnerabilities associated with molecular and immune traits. Specifically, whereas chemo-immunotherapy is beneficial to neutrophil-enriched tumors, immunotherapy alone appears to be more effective in macrophage-enriched tumors. Our findings indicate that TNBC treatment response is shaped by the interplay between tumor-intrinsic and immune features.
CONCLUSION: Our study provides a robust preclinical platform for precision immuno-oncology, enabling stratification of TNBC patients for tailored onco-immunotherapies.
KEYWORDS: Breast cancer models; Chemo-immunotherapy response; Immunocompetent mouse model; Molecular and immune stratification; Multi-parametric analysis; PD-1 blockade; Precision immuno-oncology; Syngeneic graft models; Triple-Negative Breast Cancer; Tumor heterogeneity; Tumor immune microenvironment
PMID: 41520125
DOI: 10.1186/s12943-025-02547-9