Fig 1.
RNA-seq analysis of FOLR1 expression across clinically relevant breast cancer subtypes.
(A). Log2 transformed data of FOLR1 abundance from a TCGA dataset of 691 breast cancers classified as basal, Luminal A, Luminal B, and HER2+. Bars represent 95% confidence levels for difference between the means. *Statistical significance calculated by two sided unpaired t-test, assuming unequal variances of basal versus Luminal A, Luminal B, and HER2+. (B). Distribution of FOLR1 expression across tumors of each subtype. Binned data shows frequency of tumors relative to log2 FOLR1 gene counts. Right shift in the trendline of the basal subtype illustrates a higher proportion of tumors have elevated FOLR1 levels.
Fig 2.
RNA-seq analysis of independent TNBC tumor datasets.
Distribution of FOLR1 gene counts (log2) from RNA-seq analysis of 31 TNBC tumors (Mayo Clinic) and 80 triple negative tumors from RNA-seq dataset of Shah et al. [31]. Of note, gene counts cannot be compared across the independent tumor datasets as data were generated differently (library protocols, sequencing depth, normalization etc.).
Fig 3.
Representative FOLR1 staining from breast cancer subtypes.
(A) benign breast tissue (B9) with negative staining (IHC score 0), (B) estrogen receptor positive invasive ductal carcinoma (ER+IDC) with negative staining (IHC score 0), (C) estrogen receptor positive invasive lobular carcinoma (ER+ILC) with negative staining (IHC score 0), (D) human epidermal growth factor receptor positive (HER2+) with strong 3+ staining (left core) and weak 1+ staining (middle/right cores), and (E) triple negative (TNBC) tumors with strong 3+ staining (left core) and moderate 2+ staining (middle/right cores).
Fig 4.
Analysis of FOLR1 immunohistochemical staining of breast cancer subtype tissue microarrays.
TMAs comprised of 131 breast cancer tumors of different subtypes from patients with classified breast cancers subtypes that include DCIS (n = 4), ER+ (n = 33), HER2+ (n = 26) and TNBC (n = 68). TMAs were evaluated using high affinity FOLR1 antibody Mab 26B3.F2 and IHC staining intensities calculated as described in the “Materials and Methods”. (A). The percentage of breast tumors with > 30% 3+ staining by subtype. P-values calculated Mann-Whitney test of TNBC versus benign, ER+ and HER2+. (B). H-score distribution of FOLR1 expression. H-scores were calculated with the formula: Hscore = 3 * (% at 3 +) + 2 * (% at 2 +) + 1 * (% at 1 +) + 0 * (% at 0). *Statistical significance determined by Mann-Whitney test of TNBC versus benign, ER+ and HER2+ for both (A) and (B).
Fig 5.
Distribution of FOLR1 tumor expression across TNBC tumor subtypes.
FOLR1 mRNA expression from 131 TN tumors from TCGA RNA-seq dataset were plotted versus the six TNBC subtypes that include basal-like (BL1 and BL2), immunomodulatory (IM), mesenchymal (M), mesenchymal stem-like (MSL), and luminal androgen receptor (LAR) reported by Lehman et al. [9]. Tumors were assigned to an unclassified group (UNC) if they had low correlations (<0.1) for any subtype or were similar between multiple subtypes (P<.05) [9]. Statistical significance across groups was determined by ANOVA Kruskal-Wallis test followed by Dunn’s Multiple Comparison Post test. *Significance at the P values shown was confirmed by Mann Whitney test of IM versus BL1, BL2, M, MSL, and LAR subtypes.
Fig 6.
Expression of FOLR1 influences growth of TNBC breast cancer cell lines.
(A). FOLR1 mRNA abundance in 15 well-established TNBC cell lines as determined by qPCR. (B). FOLR1 expression was depleted by stable shRNA knockdown (KD) in a subset of established TNBC cell lines expressing high levels of FOLR1 mRNA. The relative level of FOLR1 KD was determined by quantitative PCR (left panel). The growth rates of FOLR1-depleted cells vs. control non-target cells were determined over 5 days using an MTT assay (right panel). Error bars represent ± SD. (C). Rescue of the growth phenotype in HCC1806 FOLR1 KD cells by forced overexpression of FOLR1. HCC1806 FOLR1 KD cells were infected with FOLR1 or NT retrovirus and selected for 7 days with 500ug/ml G418. Growth was measured by the MTT assay. Error bars represent ± SD. (D). BrdU incorporation in HCC1806 FOLR1 KD and NT cells after a 24 hr pulse. Data plotted as mean ± SD fluorescence represents relative BrdU incorporation. (E). Measurement of apoptosis in HCC1806 FOLR1 KD and NT cells using the luminescent Caspase-Glo 3/7 assay. Error bars represent ± SD.
Fig 7.
Folate receptor overexpression increases cell growth and folate uptake.
(A). Growth of HS578T breast cancer cells engineered to stably overexpress FOLR1 or empty vector (NT). Cells were cultured in low (0–40 nM) and super-physiological (160 nM) concentrations of folic acid and growth determined after 120 hr using the MTT assay. Error bars represent ± SD. **Statistical significance calculated by two sided unpaired t-test, assuming unequal variances. (B). BrdU incorporation in HS578T/NT and HS578T/FOLR1 cells grown in 10 nM folic acid. **Statistical significance calculated by two sided unpaired t-test, assuming unequal variances. (C). Folate uptake in HS578T/NT versus HS578T/FOLR1 cells. Briefly, flow cytometry was used to measure internalized fluorescent folate (fluorescent tagged folate agent, FolateRSense 690) after 1 hr of exposure. Dot plots show forward scatter area (FSC-A) vs. FolateRSense fluorescence (APC). The percent of cells with internalized fluorescent folate and the mean fluorescence per cell type are shown.