Herceptin is a humanized monoclonal antibody targeted against the extracellular domain of the HER2 oncogene, which is amplified and overexpressed in 10–34% of breast cancers. Herceptin therapy provides effective treatment in HER2-positive metastatic breast cancer, although a favorable treatment response is not achieved in all cases. Here, we show that Herceptin treatment induces a dose-dependent growth reduction in breast cancer cell lines with HER2 amplification, whereas nonamplified cell lines are practically resistant. Time-course analysis of global gene expression patterns in amplified and nonamplified cell lines indicated a major change in transcript levels between 24 and 48 h of Herceptin treatment. A step-wise gene selection algorithm revealed a set of 439 genes whose temporal expression profiles differed most between the amplified and nonamplified cell lines. The discriminatory power of these genes was confirmed by both hierarchical clustering and self-organizing map analyses. In the amplified cell lines, the Herceptin treatment induced the expression of several genes involved in RNA processing and DNA repair, while cell adhesion mediators and known oncogenes, such as c-FOS and c-KIT, were downregulated. These results provide additional clues to the downstream effects of blocking the HER2 pathway in breast cancer and may provide new targets for more effective treatment.
The HER2 proto-oncogene encodes a 185 kDa transmembrane glycoprotein, which is one component of the four-member growth factor receptor family, including closely related EGFR, HER3, and HER4 proteins. The amplification and overexpression of HER2 is found in 10–34% of human breast cancers, and has been associated with poor clinical outcome in several studies (reviewed in Ross and Fletcher, 1999). Recently, targeted antibody-based therapy against HER2 protein was developed (Carter et al., 1992). The recombinant humanized monoclonal antibody Herceptin binds to the extracellular domain of the HER2 receptor with high affinity and disrupts various downstream signaling events leading ultimately to reduced cell growth (Molina et al., 2001; Yakes et al., 2002). Several studies have demonstrated the effectiveness of this new anticancer therapy against a subset of advanced breast tumors with HER2 overexpression, although for unknown reasons all tumors with HER2 activation do not respond favorably (Baselga et al., 1996; Cobleigh et al., 1999; Slamon et al., 2001; Vogel et al., 2002). In this study, we investigated the effects of Herceptin treatment on gene expression patterns in HER2-amplified and nonamplified breast cancer cell lines to elucidate the molecular consequences of blocking the HER2 pathway.
Breast cancer cell lines BT-474, SK-BR-3, and ZR-75-30 previously shown to display high-level amplification and overexpression of HER2, MDA-453 harboring low-level amplification (2.8-fold relative to the chromosome 17 centromere) accompanied by only a modest increase in HER2 expression, as well as MCF7 and MDA-436 with neither amplification nor increased expression, were included in this study (Kauraniemi et al., 2001; Hyman et al., 2002). Cells were harvested at the exponential growth phase and plated on 96-well plates. After an initial growth period of 48 h (to achieve an at least 70% confluency) in recommended medium and culture conditions, Herceptin was added to the test cells at various concentrations (1, 10, 100, and 500 μg/ml). Antibody treatment was carried out for 72 h, followed by cell growth analysis. As expected, the cell lines with HER2 amplification showed a clear reduction in growth after Herceptin treatment (Figure 1). This growth reduction was dose-dependent and already detectable with 1 μg/ml Herceptin. The most prominent effect was seen in SK-BR-3 cells that illustrated a 20% growth suppression with 1 μg/ml Herceptin and a 70% growth reduction with 500 μg/ml Herceptin (Figure 1). The HER2 nonamplified cell lines showed negligible growth suppression with low doses of Herceptin, but MCF7 cells did illustrate a 20% inhibition with the highest dose used. MDA-436 cells were most resistant and showed a noticeable growth suppression only with 500 μg/ml Herceptin (Figure 1). Similar to the nonamplified cell lines, the MDA-453 cell line with low-level HER2 amplification showed virtually no response with 1 μg/ml Herceptin. However, a clear growth reduction was observed in MDA-453 cells with higher doses of Herceptin, although with only a small difference between doses of 100 and 500 μg/ml (Figure 1). Overall, the difference in mean growth suppression between the HER2 amplified (BT-474, SK-BR-3, ZR-75-30, MDA-453) and nonamplified (MCF7, MDA-436) cell line groups was statistically significant (P<0.05, two-tailed t-test) with Herceptin doses of 10, 100, and 500 μg/ml.
Next, we explored the specific effects of Herceptin treatment in the HER2-amplified and nonamplified breast cancer cell lines using a cDNA microarray containing 14 380 clones. The cells were treated for 24, 48, and 72 h with 10 μg/ml Herceptin, a dose showing a statistically significant difference in growth suppression between the amplified and nonamplified cell lines with a maximum number of viable cells. The expression levels in the treated cells at specific time points were compared to those of untreated cells of the same cell line. Hierarchical clustering was performed to evaluate the overall expression changes caused by Herceptin treatment. For each cell line, the three different time points clustered together indicating that the global expression patterns within a cell line were more similar than those between different samples (Figure 2). Moreover, the hierarchical clustering dendrogram also revealed that for all cell lines the 48 and 72 h time points were more similar to one another than to the 24 h time point (Figure 2). This result suggests that the main effect of Herceptin on transcript levels occurs between 24 and 48 h of drug treatment.
A combination of statistical methods (Hautaniemi et al., 2003a) was then used to identify a specific set of genes whose expression levels either increased or decreased as a function of time during Herceptin treatment. First, equalization transformation (Bolstad et al., 2003) was performed to obtain a similar range for ratios across the time-series experiments. Linear regression analysis was then applied to convert the three time-series observations to a single number T, where the sign of T reflects the direction of change in the expression level (either decrease or increase) and the value of T corresponds to the magnitude of expression change. After calculation of the T-values, a stepwise gene selection algorithm was utilized to identify genes that would best separate the HER2 amplified cell lines from those with no amplification. For this purpose, we assigned the BT-474, SK-BR-3, and ZR-75-30 cell lines as amplified and MDA-436 and MCF7 as nonamplified. MDA-453 was excluded because it did not fit into either category. Genes where the difference in the median expression level between the amplified and nonamplified groups was less than one were excluded, leaving 2990 genes for the analysis. In the stepwise gene selection, the weight (W) for each gene was computed using Fisher's linear discriminant and new genes were added to the set until W reduced to three. As a result, a set of 439 discriminatory genes were identified representing 15% of the genes included in the stepwise gene selection procedure (Supplemental Table 1). Of these, 212 and 227 genes showed decreasing and increasing expression levels, respectively, after the Herceptin treatment in the amplified vs nonamplified cell lines, demonstrating an almost equal partitioning of up- and downregulated discriminatory genes. A randomization test (Hautaniemi et al., 2003a) was applied for the 2990 genes to verify that the result was statistically significant (P<10−9). The expression level of the HER2 gene varied less than twofold during the antibody treatment in all cell lines, indicating that the treatment did not have a significant effect on the expression of the HER2 receptor itself.
Two different clustering approaches were used to illustrate the temporal gene expression changes among the 439 discriminatory genes (Figure 3). In hierarchical clustering, the amplified and nonamplified cell lines were clustered at the opposite ends of the dendrogram. The self-organizing map revealed two major clusters of genes, whose expression profiles were almost opposite in the amplified and nonamplified cell lines. These results demonstrate that the expression profiles of the discriminatory genes were clearly distinct between the HER2-amplified and nonamplified cell lines, thereby confirming the validity of the set of discriminatory genes. Interestingly, both clustering methods indicated that the expression changes in the MDA-453 cell line were different from those in the other cell lines with HER2 amplification, with a large fraction of genes showing temporal gene expression patterns similar to those seen in the nonamplified cell lines. This transcriptional profile might be explained by the fact that the low-level HER2 amplification in MDA-453 leads only to a slight increase in HER2 expression (Kauraniemi et al., 2001; Hyman et al., 2002), and therefore the effects of Herceptin treatment in MDA-453 mimic those seen in cells with no HER2 amplification.
The publicly available gene ontology information (http://source.stanford.edu/cgi-bin/sourceSearch) provided information on the functional roles of 59% of the discriminatory genes. The remaining 179 genes represented hypothetical proteins or known genes with no associated functional information. Closer examination of the gene ontology data revealed that a large number of the discriminatory genes encode for proteins involved in key cellular processes like transcription, signal transduction, protein processing, cell metabolism, and transport. Genes encoding proteins involved in various RNA processing reactions, for example, biogenesis of spliceosomal snRNAs (SIP1), RNA 3′-end processing (CPSF4), and nuclear pre-mRNA splicing (PRPF31) were upregulated by Herceptin treatment in the amplified cell lines. Herceptin treatment also induced the expression of genes involved in various DNA repair pathways, for example, nucleotide-excision (ERCC2), double-strand break (MRE11), and DNA mismatch repair (MSH5). On the other hand, genes encoding cell adhesion proteins (e.g. JUP, CTNND2, and CNTN1) or well-characterized oncoproteins (e.g. FOS and KIT) showed decreasing expression levels in the amplified cell lines during the drug treatment. Several studies have shown that estrogen-induced mitogenesis of breast cancer cells is partly mediated by the increased expression of c-FOS, indicating an important role for c-FOS in breast cancer pathogenesis (Wilding et al., 1988; Van der Burg et al., 1989; Duan et al., 2002). Although c-KIT oncogene, encoding a transmembrane receptor tyrosine kinase, has not been previously implicated in breast cancer, it is often activated in a number of human malignancies, including mast cell leukemia (Furitsu et al., 1993), acute myeloid leukemia (Longley et al., 2001), and gastrointestinal stromal tumors (Hirota et al., 1998).
In conclusion, the Herceptin treatment induced a dose-dependent growth reduction in HER2 amplified breast cancer cell lines. The expression profiling identified a set of 439 genes whose temporal expression patterns discriminated between the amplified and nonamplified cell lines. These included several genes involved in key cellular processes, such as RNA processing, DNA repair, cell adhesion, and oncogenesis. Detailed analysis of these genes is likely to provide additional information on the downstream effects of blocking the HER2 pathway in breast cancer.
Baselga J, Tripathy D, Mendelsohn J, Baughman S, Benz CC, Dantis L, Sklarin NT, Seidman AD, Hudis CA, Moore J, Rosen PP, Twaddell T, Henderson IC and Norton L . (1996). J. Clin. Oncol., 14, 737–744.
Bolstad B, Irizarry R, Astrand M and Speed TA . (2003). Bioinformatics, 19, 185–193.
Carter P, Presta L, Gorman CM, Ridgway JB, Henner D, Wong WL, Rowland AM, Kotts C, Carver ME and Shepard HM . (1992). Proc. Natl. Acad. Sci. USA, 89, 4285–4289.
Chen Y, Dougherty ER and Bittner ML . (1997). J. Biomed. Optics, 2, 364–374.
Cleveland W . (1979). J. Am. Stat. Assoc., 74, 829–836.
Cobleigh MA, Vogel CL, Tripathy D, Robert NJ, Scholl S, Fehrenbacher L, Wolter JM, Paton V, Shak S, Lieberman G and Slamon DJ . (1999). J. Clin. Oncol., 17, 2639–2648.
Duan R, Xie W, Li X, McDougal A and Safe S . (2002). Biochem. Biophys. Res. Commun., 294, 384–394.
Eisen M, Spellman P, Brown P and Botstein D . (1998). Proc. Natl. Acad. Sci. USA, 95, 14863–14868.
Furitsu T, Tsujimura T, Tono T, Ikeda H, Kitayama H, Koshimizu U, Sugahara H, Butterfield JH, Ashman LK, Kanayama Y, Matsuzawa Y, Kitamura Y and Kanakura Y . (1993). J. Clin. Invest., 92, 1736–1744.
Hautaniemi S, Kauraniemi P, Rämö P, Yli-Harja O, Astola J and Kallioniemi A . (2003a). A Strategy for Identifying Class-Separating Genes in Drug-Treatment Microarray Data. Report 1. Institute of Signal Processing, Tampere University of Technology: Finland, ISBN 952-15-1067-6.
Hautaniemi S, Yli-Harja O, Astola J, Kauraniemi P, Kallioniemi A, Wolf M, Ruiz J, Mousses S and Kallioniemi O-P . (2003b). Mach. Learn., 52, 45–66.
Hirota S, Isozaki K, Moriyama Y, Hashimoto K, Nishida T, Ishiguro S, Kawano K, Hanada M, Kurata A, Takeda M, Muhammad Tunio G, Matsuzawa Y, Kanakura Y, Shinomura Y and Kitamura Y . (1998). Science, 279, 577–580.
Hyman E, Kauraniemi P, Hautaniemi S, Wolf M, Mousses S, Rozenblum E, Ringner M, Sauter G, Monni O, Elkahloun A, Kallioniemi OP and Kallioniemi A . (2002). Cancer Res., 62, 6240–6245.
Kauraniemi P, Bärlund M, Monni O and Kallioniemi A . (2001). Cancer Res., 61, 8235–8240.
Kohonen T (ed). (2001). Self-Organizing Maps, 3rd edn Springer: Heidelberg, Germany.
Longley BJ, Reguera MJ and Ma Y . (2001). Leuk. Res., 25, 571–576.
Molina MA, Codony-Servat J, Albanell J, Rojo F, Arribas J and Baselga J . (2001). Cancer Res., 61, 4744–4749.
Monni O, Bärlund M, Mousses S, Kononen J, Sauter G, Heiskanen M, Paavola P, Avela K, Chen Y, Bittner ML and Kallioniemi A . (2001). Proc. Natl. Acad. Sci. USA, 98, 5711–5716.
Mousses S, Bittner ML, Chen Y, Dougherty ER, Baxevanis A, Meltzer PS and Trent JM . (2000). Functional Genomics: Gene Expression Analysis by cDNA Microarrays, Livesey FJ and Hunt SP (eds). Oxford University Press: Oxford, pp. 113–137.
Ross JS and Fletcher JA . (1999). Semin. Cancer Biol., 9, 125–138.
Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, Fleming T, Eiermann W, Wolter J, Pegram M, Baselga J and Norton L . (2001). N. Engl. J. Med., 344, 783–792.
van der Burg B, van Selm-Miltenburg AJ, de Laat SW and van Zoelen EJ . (1989). Mol. Cell. Endocrinol., 64, 223–228.
Vesanto J, Himberg J, Alhoniemi E and Parhankangas J . (2000). SOM Toolbox for Matlab 5. Technical Report A57. Helsinki University of Technology: Finland.
Vogel CL, Cobleigh MA, Tripathy D, Gutheil JC, Harris LN, Fehrenbacher L, Slamon DJ, Murphy M, Novotny WF, Burchmore M, Shak S, Stewart SJ and Press M . (2002). J. Clin. Oncol., 20, 719–726.
Wilding G, Lippman ME and Gelmann EP . (1988). Cancer Res., 48, 802–805.
Yakes FM, Chinratanalab W, Ritter CA, King W, Seelig S and Arteaga CL . (2002). Cancer Res., 62, 4132–4141.
Yang Y, Dudoit S, Luu P and Speed T . (2001). Microarrays: Optical Technologies and Informatics, Bittner M, Chen Y, Dorsel A and Dougherty E (eds). SPIE, Society for Optical Engineering: San Jose, CA, pp. 141–152.
We thank Ms Kati Rouhento for excellent technical assistance. This work was supported by the Academy of Finland, Foundation for Finnish Cancer Institute, the Medical Research Fund of the Tampere University Hospital, the Science Fund of Tampere, as well as Finnish and Pirkanmaa Cultural Foundations.
About this article
Single-cell RNA sequencing reveals novel gene expression signatures of trastuzumab treatment in HER2+ breast cancer
Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization (2015)
Scientific Reports (2015)
Poly (ADP-ribose) polymerase inhibition enhances trastuzumab antitumour activity in HER2 overexpressing breast cancer
European Journal of Cancer (2014)
Journal of Proteome Research (2013)