Background: PgR is only expressed significantly in ER+ breast cancers and is generally considered a classical estrogen-sensitive protein, however, while ER+ cells are almost always expressed in a non-clustered pattern across the tumour bed, PgR+ cells form clusters or clumps with surrounding areas of PgR- cells when assessed by IHC in about 20-25% of PgR+ tumours. This is an ill-described and unexplained feature of intratumoural heterogeneity of a widely measured biomarker. The prognostic significance of the PgR clumpiness is also unknown.
Aims: to determine the (i) biologic and (ii) prognostic significance of PgR clumpiness in ER+ primary breast cancer.
Methods: (i) Biologic study: 40 primary ER+PgR+ tumours were identified with distinct PgR clumpiness by IHC. Areas of PgR+ and PgR- cells (PgR-rich and PgR-poor areas, respectively) were needle microdissected from unstained sections by juxtaposing them with PgR-stained sections from the same tumour. Areas were also dissected from 8 homogeneous PgR+ and 8 PgR- tumours for reference. NanoString gene expression analysis using a 50-gene code set (45 oestrogen-responsive or proliferation genes + 5 housekeeping genes) was performed on RNA extracts from the samples. (ii) Prognostic study: we developed a PgR heterogeneity score (PgR Het-score, range 0-18) as the product of (A) the number of PgR+ clumps/cm2 of tumour area (range 0 to 6) and (B) the level of overall PgR-positivity (0-4% or 96-100%, score 0; 5-19% or 80-95%, score 1; 20-39 or 60-79%, score 2; 40-59%, score 3). PgR Het-scores were derived from images (Hamamatsu NanoZoomer-XR) of PgR-stained sections from two sets of tumour biopsies taken at diagnosis after which all patients received adjuvant endocrine therapy: 322 tumours from the Royal Marsden and from Southern Sweden collected for a case:control study of risk of recurrence (recurrence:no recurrence, 1:1) and 591 patients from the TransATAC cohort study with time to recurrence as the end-point.
Results: (i) Biologic study: In unsupervised hierarchical clustering of RNA expression from all samples, 33/40 pairs of PgR-rich and PgR-poor areas from a single tumour paired together and all 8 positive and negative control samples were grouped separately. Eight genes were differentially expressed between the PgR-rich and -poor areas (FDR<0.05: PGR, SERPINA3, AURKA, MSMB, PDZK1, IGSF1, FKBP5, SLC2A3). PGR had a 35-fold difference indicating that transcriptional differences explained the PgR IHC differences. The progesterone-regulated genes FKBP3 and SERPINA3 both showed significantly higher expression in PgR-rich areas consistent with the PgR differences having functional impact. AURKA was also more highly expressed in the PgR-rich areas indicating that the PgR-rich areas may also be more proliferative. ER and most of known estrogen-regulated genes were not differentially expressed and therefore show an unexpected divergence from PgR expression and regulation downstream of ER. (ii) Prognostic study: The PgR Het-score showed no consistent association with clinical factors including tumour size, nodal status, grade, age. Each of the individual components as well as the composite PgR Het score were significantly associated with risk of recurrence in POLAR: (A) p=0.02; (B) p=0.002; (PgR Het score) p=0.008. However, there was no association of the PgR Het score with risk of recurrence in TransATAC.
Conclusion: Major, functionally significant differences in PgR expression occur within some ER+ tumours that are not explained by differences in ER expression and are not associated with differences in expression of most other estrogen-dependent genes. Further work is on-going to provide a mechanistic explanation for the PgR heterogeneity. Given the discordance in results from the two clinical studies its prognostic significance is uncertain.
This work was funded by BCRF.
Citation Format: Lila Zabaglo, Richard Buus, Eugene Schuster, Belinda Yeo, Marie Klintman, Ivana Sestak, Jack Cuzick, Ian Smith, Mitch Dowsett. Intratumoural heterogeneity in PgR expression: Molecular and prognostic significance [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS18-10.