The Internally Truncated LRP5 Receptor Presents a Therapeutic Target in Breast Cancer

Background Breast cancer is a common malignant disease, which may be caused by a number of genes deregulated by genomic or epigenomic events. Deregulated WNT/β-catenin signaling with accumulation of β-catenin is common in breast tumors, but mutations in WNT signaling pathway components have been rare. An aberrantly spliced internally truncated LRP5 receptor (LRP5Δ666–809, LRP5Δ) was shown recently to be resistant to DKK1 inhibition, and was required for β-catenin accumulation in hyperparathyroid tumors and parathyroid tumor growth. Methodology/Principal Findings Here we show, by reverse transcription PCR and Western blot analysis, that LRP5Δ is frequently expressed in breast tumors of different cancer stage (58–100%), including carcinoma in situ and metastatic carcinoma. LRP5Δ was required in MCF7 breast cancer cells for the non-phosphorylated active β-catenin level, transcription activity of β-catenin, cell growth in vitro, and breast tumor growth in a xenograft SCID mouse model. WNT3 ligand, but not WNT1 and WNT3A augmented the endogenous β-catenin activity of MCF7 cells in a DKK1-insensitive manner. Furthermore, an anti-LRP5 antibody attenuated β-catenin activity, inhibited cell growth, and induced apoptosis in LRP5Δ-positive MCF7 and T-47D breast cancer cells, but not in control cells. Conclusions/Significance Our results suggest that the LRP5Δ receptor is strongly implicated in mammary gland tumorigenesis and that its aberrant expression present an early event during disease progression. LRP5 antibody therapy may have a significant role in the treatment of breast cancer.


Introduction
Deregulated Wnt signaling with accumulation of b-catenin in the cytoplasm/nucleus plays an important role in a variety of human cancers. The binding of WNT ligand to frizzled and LRP5/6 cell surface receptors normally leads to inhibition of a ''destruction complex'' consisting of APC/Axin/GSK-3b/Ck1/Dvl and other factors, with subsequent accumulation of dephosphorylated stabilized b-catenin, and regulation of its target genes [1][2][3][4][5][6][7]. Wnt signaling is involved in mammary gland development and mouse WNT1, WNT3, and WNT10b are clearly implicated in MMTVinduced breast tumorigenesis [8,9]. The LRP5 receptor was recently shown to be required for mammary ductal stem cell activity and WNT1-induced tumorigenesis in the mouse [10]. This may be of particular importance since subtypes of breast tumors have been suggested to originate from stem cell populations [11][12][13].
A novel mechanism for aberrant activation of Wnt signaling was recently disclosed by us in parathyroid tumors [26]. An aberrantly spliced internally truncated LRP5 receptor (LRP5D666-809, LRP5D) was found to be expressed in the majority of analyzed parathyroid tumors of both primary and secondary origin, which all displayed cytoplasmic/nuclear b-catenin. The LRP5D receptor responded strongly to WNT3 ligand and was shown to be required for accumulation of nonphosphorylated transcriptionally active bcatenin, MYC expression, parathyroid cell growth in vitro, and parathyroid tumor growth in vivo in SCID mice. Furthermore, LRP5D and stabilizing mutation of CTNNB1 was found to be mutually exclusive in those tumors [26,27]. The 142 amino acid (666-809) extracellular truncation of LRP5D overlaps a binding domain for the LRP5 antagonist DKK1 [28][29][30][31], and consequently LRP5D was found to be insensitive to inhibition by the DKK1 ligand [26].
We have now extended our analysis of aberrant Wnt signaling to breast carcinoma, and demonstrate a fundamental role for the internally truncated LRP5 receptor in deregulated b-catenin signaling and tumor growth.

Results and Discussion
The LRP5D receptor is expressed in breast carcinoma To investigate whether the internally truncated LRP5D666-809 receptor (LRP5D) was expressed in breast carcinoma we initially analyzed nineteen breast cancer specimens (T1-T19) that had been randomly selected in a previous study regarding 25hydroxyvitamin D 3 1a-hydroxylase [32]. RT-PCR analysis, using primers located in exons 9 and 13 of LRP5 [26], revealed expression of LRP5D in 15 out of 17 ductal breast carcinoma and 1 out of 2 lobular carcinoma ( Figure 1A). Thus, 84% of the analyzed tumors expressed LRP5D. Normal LRP5 (LRP5wt) transcripts were seen in all tumors, and in normal breast tissue specimens (N1-N4) as anticipated ( Figure 1A) [26]. Immunoprecipitation and Western blot analysis confirmed expression of LRP5wt and LRP5D in the tumors ( Figure 1B).
In order to relate accumulation of b-catenin to expression of LRP5D, Western blotting analysis was done on cryosections of the Figure 1. The LRP5D receptor is expressed in breast tumors. Accumulation of active b-catenin. (A) PCR analysis of cDNA from normal breast tissue (N1-N4) and primary breast cancer (T1-T19). Primers were located in exons 9 and 13 of LRP5 as described [26]. The PCR reaction is not quantitative as the primers compete for the two fragments. A total of sixty-two LRP5 truncated fragments (see also tumor material, including two normal breast tissue specimens. The sixteen tumors with LRP5D showed increased accumulation of non-phosphorylated active b-catenin when compared to the normal breast tissues, and the tumors expressing LRP5wt only (T3, T9, T11) showed similar relative amounts ( Figure 1C). Thus, correlation of LRP5D expression and aberrant accumulation of non-phosphorylated active b-catenin was observed in these breast tumors. Stabilizing mutation in b-catenin exon 3 was not found in the nineteen breast carcinomas (not shown).
Encouraged by the above results we then screened a commercially available cDNA panel of 96 breast tissue specimens that covered eight cancer stages. LRP5wt and LRP5D was observed in 79 out of 95 (83%) samples ( Table 1). The remaining samples expressed only LRP5wt and one sample was negative. LRP5D was detected in all disease stages including carcinoma in situ and metastases. Thus, expression of LRP5D was very common in carcinoma of the breast and may be an early event during cancer progression.

LRP5D causes active b-catenin signaling in MCF7 cells
In order to study the consequences of LRP5D receptor expression we employed the widely used MCF7 mammary adenocarcinoma cell line which expressed the LRP5D receptor but not LRP5wt (Figure 2A), and has been shown to accumulate b-catenin in the nucleus [25,33]. We first chose to test the TOPFLASH/FOPFLASH and the pTOPGlow/pFOPGlow TCF luciferase reporters, although conflicting results regarding detection of active b-catenin signaling using these reporters in breast cancer cell lines have been published [14,34,35].
Control siRNAs and three highly specific siRNAs directed against LRP5 mRNA were transfected to MCF7 cells. The specificity and silencing potential of the LRP5 siRNAs were ascertained at the mRNA and protein level in MCF7 cells ( Figure 2B), as we have shown previously in sHPT-1 parathyroid tumor cells [26]. Transfection of siLRP5D, specific for the internally truncated LRP5 receptor, as well as of siLRP5tot directed against exon 13 present in both LRP5 wt and LRP5D transcripts, resulted in markedly reduced non-phosphorylated active b-catenin level, compared to control siRNAs and siLRP5wt ( Figure 2C). siLRP5wt was directed to exon 10, not included in the LRP5D transcript. sib-catenin was included as positive control. Similarily, the endogenous b-catenin activity in MCF7 ( Figure 2D, left panel), as measured by using the TOPFLASH/FOPFLASH or the pTOP/Glow/pFOPGlow TCF/b-catenin luciferase reporters (6-8 fold), was dependent on maintained expression of LRP5D and b-catenin ( Figure 2D, right panels). The various siRNAs displayed no effect on FOPFLASH [26] or pFOPGlow reporter activities (data not shown), which contain mutated TCF binding elements in their promoters [34]. The experiments were done using our in house strain of MCF7 or fresh MCF7 cells from ATCC, which similarily expressed LRP5D and showed b-catenin activity by the TOPFLASH and pTOPGlow assays.
Next we employed the natural b-catenin responsive DKK1 promoter [36][37][38] instead of the synthetic minimal promoters of TOPFLASH and pTOPGlow. Clearly, the DKK1 promoter activity in transfected MCF7 cells was dependent on the TCF binding elements (TBEs), and also on maintained expression of LRP5D and b-catenin ( Figure 3A). Thus, this confirmed the results obtained with the TOPFLASH and pTOPGlow reporters ( Figure 2D). Furthermore, we transfected the various siRNAs to MCF7 cells and determined the endogenous DKK1 mRNA expression level. In accordance with the above results, transfection of siLRP5D and siLRP5tot, but not of siLRP5wt or Control siRNA resulted in significantly reduced endogenous DKK1 expression ( Figure 3B).
In an attempt to address possible explanations for conflicting results regarding detection of b-catenin activity by the TOP-FLASH and pTOPGlow assays [14,34,35], we performed transfections in the presence of varying cell densities. We routinely plate cells rather sparsely (2610 5 cells/35 mm dish, 10 4 cells/96well microplates) as compared to for example 80% confluency [34]. As expected, the transfection efficiency (b-galactosidase activity) was reduced with increasing cell density (not shown), and clearly the TOPFLASH/FOPFLASH as well as the pTOPGlow/ pFOPGlow ratios decreased with increasing cell density ( Figure 3C). Thus, cell density seemed to be an important determinant of WNT/b-catenin signaling constituting one possible explanation for published inconsistencies.
In summary, the results showed that maintained expression of the internally truncated LRP5 receptor in MCF7 cells appeared necessary for accumulation of transcriptionally active b-catenin.

WNT3 ligand and LRP5D activate transcription synergistically in a DKK1-insensitive manner
We reported previously that WNT3 ligand conditioned medium (CM), but not WNT1 and WNT3A, further activated endogenous b-catenin driven TOPFLASH reporter transcription in sHPT-1 parathyroid tumor cells. WNT1, WNT3 and WNT3A are expressed in MCF7 cells [25], and their effects were determined by transient cotransfections of the TOPFLASH reporter, expression plasmids for LRP5wt or LRP5D, and incubation with WNT CMs ( Figure 4A). Only WNT3 CM activated the endogenous bcatenin activity in MCF7 (10-fold). Transfection with LRP5D increased the endogenous b-catenin activity by 8-fold, and this was further strongly enhanced by WNT3 (12-fold), to a total of 96-fold activation compared to control transfected and unstimulated cells (1.0). Transfection of LRP5wt and stimulation with WNT3 CM resulted in 8-fold activation. WNT1 and WNT3A CM activated transcription in the presence of cotransfected LRP5wt (2.5 and 9fold) and LRP5D (3.5 and 2.5-fold). Thus, WNT1 activated transcription in the presence of cotransfected LRP5D in these breast tumor cells, while only WNT3 activity was observed in parathyroid tumors cells [26]. This may reflect the collection of expressed frizzled receptors or other cofactors in the two cell lines.
The LRP5 antagonist DKK1 requires several amino acid residues included in the LRP5D truncation, and indeed the DKK1 ligand could not inhibit endogenous LRP5D-induced transcriptional activity in parathyroid tumor cells or in LRP5D-transfected  PLoS ONE | www.plosone.org MCF7 cells. siLRP5wt is directed to wild type transcripts, siLRP5D to the truncated transcript, and siLRP5tot to both transcripts [26]. Quantitative realtime PCR of both LRP5 transcripts (LRP5tot, left panel) and immunoprecipitation and Western blot analysis of LRP5 (right panel). (C) Western blot analysis of non-phosphorylated active b-catenin after siRNA transfection. (D) Transient cotransfections of TOPFLASH/FOPFLASH or pTOPGlow/ pFOPGlow TCF/b-catenin reporter, the CMV-LacZ reference plasmid (left panel), and the indicated siRNAs (right panels) to MCF7 cells. FOPFLASH and pFOPGlow contain mutated binding sites for TCFs, while TOPFLASH and pTOPGlow do not. Luciferase activities were normalized to b-galactosidase activities. The siRNAs displayed no effect on pFOPGlow (not shown) and FOPFLASH reporter activity [26]. doi:10.1371/journal.pone.0004243.g002 HEK 293T cells [26]. Similarily, the presence of DKK1 CM did not inhibit WNT3-induced transcriptional activation in MCF7 cells, while inhibition of WNT3-induced endogenous b-catenin activity was observed as expected in HEK 293T control cells ( Figure 4B). Thus, the DKK1 ligand failed to inhibit b-catenin signaling in vitro, and this may contribute to the total aberrant signaling level in LRP5D-positive breast tumors. DKK1 was found to be expressed in all (n = 16) analyzed breast tumors, as determined by immunohistochemistry (not shown). The data presented so far provide compelling evidence for a major role of the internally truncated LRP5 receptor in sustained aberrant b-catenin signaling in breast cancer.

LRP5D is required for breast tumor cell growth in vitro and in vivo in SCID mice
Control siRNA, siLRP5wt, siLRP5D, and siLRP5tot were transfected to MCF7 cells and the cell viability was determined 24 hrs and 48 hrs later. Growth inhibition and induction of cell death were seen only with siLRP5D, and siLRP5tot at both time points ( Figure 5A). Reduced expression of LRP5D also lead to growth inhibition and cell death in sHPT-1 parathyroid tumor cells, but not in HeLa cells that do express LRP5wt and not LRP5D [26].
Tumor growth was then evaluated in a xenograft SCID mouse model. Tumor growth was significantly reduced in transplants of MCF7 cells pretransfected with siLRP5D and siLRP5tot, but not with control siRNA when compared to non-transfected cells ( Figure 5B). Thus, LRP5D appeared to be necessary for breast tumor cell growth both in cell culture and in SCID mice.
If breast tumor cell growth is dependent on LRP5D, as strongly suggested by the above experiments, an appropriate anti-LRP5 antibody may reduce cell viability as well as b-catenin activity. The anti-LRP5 goat polyclonal antibody, that immunoprecipitated both LRP5wt and LRP5D ( Figure 1B, Figure 2B, and Figure 5C), significantly attenuated the non-phosphorylated active b-catenin level and the b-catenin activity in MCF7 cells, and caused reduced cell viability ( Figure 5D). This was not observed in HeLa cells that only express LRP5wt ( Figure 5D). We also determined effects of the LRP5 antibody on T-47D breast cancer cells, since these cells expressed both LRP5wt and LRP5D, showed nuclear accumulation of b-catenin [25], and displayed endogenous b-catenin activity as determined by the TOPFLASH assay ( Figure 5C) [14]. The LRP5 antibody significantly inhibited cell growth and attenuated b-catenin activity also in these cells ( Figure 5D). Treatment with the anti-LRP5 antibody induced significant apoptosis in both breast cancer cell lines, and not in HeLa cells ( Figure 5E).
Thus, breast tumor cell growth was dependent on maintained expression of LRP5D and continued b-catenin signaling by the receptor. The latter result is in line with the observation that WNT antagonist SFRPs were shown to suppress MCF7 and T-47D cell colony formation [25]. Compared to LRP5wt, LRP5D activated b-catenin driven transcription more strongly in the presence of WNT3 ligand and in a DKK1-insensitive way, both likely contributing to the oncogenic potential of LRP5D. In addition to dephosphorylation/phosphorylation, specific ubiquitination by Rad6B has been suggested to control b-catenin stabilization in MDA-MB-231 breast cancer cells [35]. LRP5D was found not to be expressed in this cell line (not shown).
Although the mechanism by which the LRP5D receptor is made and how b-catenin signaling is activated remain to be understood, the results presented here strongly support an important role of the truncated LRP5 receptor in mammary gland tumorigenesis, and its expression may present an early event during disease progression. Our results furthermore suggest that antibody therapy directed against LRP5D, possibly in combination with chemotherapy [39][40][41], present future treatment options of breast cancer.

Tissue specimens
Cryosections of nineteen breast cancer specimens, including seventeen invasive ductal carcinoma and two invasive lobular carcinoma [32] were analyzed. Four apparently normal breast tissue specimens from patients with breast cancer were also included in the analyses. Written informed consent was obtained from the patients and approval was obtained from the local ethical committee, Uppsala. cDNA panels containing 96 breast samples covering eight (0, I, IIA, IIB, IIIA, IIIB Detection of LRP5D and DNA sequencing DNA-free total RNA was prepared from breast tissue cryosections and from MCF7 or T-47D cells, and analyzed by RT-PCR using primary and nested primers spanning positions 1992-2932 of LRP5 mRNA as described [26]. RT-PCR analysis of the two breast tissue cDNA panels was performed using the same primers as above, and in addition the nested PCR was also done using a novel forward primer specific for the D666-809 truncation: forward primer (position 2012), 59-TCCACAGGATCTCCCTCGAGACCAA-TAACAACGACC-39; and reverse (position 2543), 59-TTGGAC-GACTCGATCATGTTGGTGTCCAGGTCGGTC-39 (Gen-Bank accession number AF064548; http://www.ncbi.nlm.nih. gov/Genbank). The underlined C residue in the forward primer is unique to the D666-809 truncation point [26]. The PCR resulted in a 105 base pair fragment. The following conditions were used for the LRP5D-specific PCR: The nested PCR amplification was performed with 5 ul primary PCR product, 10 pmol of each primer, 0.2 mM dNTPs, 16 PCR buffer, 1,5 mM MgCl 2 and 0.25 U Platinum Taq DNA polymerase (Invitrogen Corporation, Carlsbad, California, USA). Denaturation at 95uC for 60 s, followed by 40 cycles of denaturation for 10 s, annealing at 60uC for 20 s and extension at 72uC for 10 s and a final extension at 72uC for 7 min. The LRP5D-specific PCR is more sensitive since no competition between LRP5wt and LRP5D fragments occurs during the PCR. Additional positive tumors of the cDNA panel was detected by this protocol. A total of sixty-two LRP5 truncated fragments were directly sequenced on ABI 373A using the ABI Prism Dye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems, Foster City, California, USA), and all contained the same in-frame deletion (D666-809).
Quantitative PCR analysis of LRP5tot and DKK1 mRNA cDNA from siRNA transfected MCF7 cells was prepared as described above. The following mRNA-specific PCR primers and labeled probe (59FAM-sequence-39TAMRA) were used for quantitative real-time RT-PCR analysis. For LRP5tot: forward, 59-ATCGACTGTATCCCC GGGGC-39; reverse, 59-CACCACG-CGCTGGCACACAA-39; and probe, 59-CGGACTGTG ACGC-CATCTGCC TGC-39. For 28S rRNA, the Ribosomal RNA Control Reagents (VIC probe) was used (Applied Biosystems, Foster City, California, USA). The following PCR primers were used for DKK1: forward, 59-TTCTCCCTCTTGAGTCCTTCTG-39; and reverse, 59-AGGAGTTCACTGCATTTGGAT-39. PCR reactions were performed on MyiQ Single-Color Real-Time PCR Detection System (Bio-Rad Laboratories, Inc., Hercules, California, USA) using the TaqMan PCR core Reagent Kit (Applied Biosystems) or the iQ SYBR Green Supermix (Bio-Rad Laboratories, Inc.). Each cDNA sample was analyzed in triplicate. Standard curves for the expressed genes were established by amplifying a purified PCR fragment covering the sites for probes and primers.

Mouse xenograft model
Two to three week old female Fox Chase severe combined immunodeficient mice (SCID) were used (Taconic, Denmark). The mice were anesthetized with isoflurane (Forene; Abbott, Abbott Park, IL, USA) during the manipulations. Both flanks of each animal were injected subcutaneously (total 200 ml) with MCF7 cells together (1:1) with BD Matrigel Matrix (BD Biosciences Clontech, Palo Alto, California, USA), after transfection of 10 6 cells for 24 hours. The animals were monitored every day and sacrificed after 5 weeks. The animal experiments was approved by the Uppsala University board of animal experimentation and was performed according to the United Kingdom Coordinating Committee on Cancer Research guidelines for the welfare of animals in experimental neoplasia [43].

Statistical analysis
Unpaired t test was used for all statistical analyses. Values are presented as arithmetrical mean6SEM. A p value of ,0.05 was considered significant.