LRP5 promotes cancer stem cell traits and chemoresistance in colorectal cancer

Abstract The overactivation of canonical Wnt/β‐catenin pathway and the maintenance of cancer stem cells (CSCs) are essential for the onset and malignant progression of most human cancers. However, their regulatory mechanism in colorectal cancer (CRC) has not yet been well demonstrated. Low‐density lipoprotein receptor‐related protein 5 (LRP5) has been identified as an indispensable co‐receptor with frizzled family members for the canonical Wnt/β‐catenin signal transduction. Herein, we show that activation of LRP5 gene promotes CSCs‐like phenotypes, including tumorigenicity and drug resistance in CRC cells, through activating the canonical Wnt/β‐catenin and IL‐6/STAT3 signalling pathways. Clinically, the expression of LRP5 is upregulated in human CRC tissues and closely associated with clinical stages of patients with CRC. Further analysis showed silencing of endogenous LRP5 gene is sufficient to suppress the CSCs‐like phenotypes of CRC through inhibiting these two pathways. In conclusion, our findings not only reveal a regulatory cross‐talk between canonical Wnt/β‐catenin signalling pathway, IL‐6/STAT3 signalling pathway and CD133‐related stemness that promote the malignant behaviour of CRC, but also provide a valuable target for the diagnosis and treatment of CRC.


| INTRODUC TI ON
Colorectal cancer (CRC) is the third most common cancer and the fourth leading cause of cancer-related mortality worldwide, with almost 1.93 million newly diagnosed cases and 935,000 deaths occurred worldwide in 2020. 1 The routine use of colonoscopy and systematic therapies, such as chemo-and radiation therapy as well as immunotherapy have effectively improved the diagnosis rate, controlled many localized CRC and evidently improved clinical prognosis. 2 However, when it progresses to advanced stages, because of the metastasis, acquisition of chemoresistance and recurrence, the 5-year survival rate sharply declines to be less than 30%. Therefore, it is imperative to explore accurate biomarkers and elucidate mechanisms underlying tumorigenesis of CRC for developing novel and more effective therapeutic strategies.
As a small subpopulation of cancer-initiating cells, cancer stem cells (CSCs) will undergo asymmetric division to maintain their own pool and differentiate into ordinary common tumour cells and are considered being responsible for the initiation, maintenance, progression, metastasis, relapse, chemo-/radioresistance and other phenotypes of many types of human malignancies, including CRC. [3][4][5] CSCs in CRC were first reported in 2007 that CRC was created and propagated by a rare population of undifferentiated tumorigenic cells expressing CD133. 6,7 Subsequently, a series of studies showed that the maintenance of colorectal CSCs-like phenotypes was also closely related to the specifical expression of other protein markers such as aldehyde dehydrogenase (ALDH1A1), Lgr5, Nanog and Bmi1, and the overactivation of several important signalling pathways including Notch, OCT3/4, Hedgehog and Wnt etc. [8][9][10][11] Among them, the hyperactivation of Wnt pathways is considered being one of the most important events for the occurrence and progression of a variety of cancers. Specifically, activation of canonical Wnt/βcatenin via upregulation of its key components β-catenin and c-Myc could augment the formation of CRC cellular colonospheres, which are particularly concentrated in CSCs. 12 Consistent with these observations, hypoxia-induced canonical Wnt/β-catenin/Id2 cascade increased the occurrence of CSCs-like phenotypes of CRC cells. 13 Conversely, blocking the canonical Wnt pathway by phenethyl isothiocyanate, histone demethylase inhibitor JIB-04 or its inhibitor IC-2 could effectively inhibit colorectal CSCs and sensitize CRC cells to chemotherapeutic agents. [14][15][16] However, the underlying mechanisms of canonical Wnt/β-catenin pathway dysregulation in colorectal CSCs have not been clearly elucidated.
Low-density lipoprotein receptor-related protein 5 (LRP5) serves as a key co-receptor with Frizzled (FZD) proteins for transmitting signals by Wnt ligands. 17 The canonical Wnt/β-catenin pathway is activated upon binding of secreted Wnt ligands to the FZD and LRP5/6; then, Dishevelled (DVL) proteins are activated and recruited to form DVL polymers that can deactivate the 'destruction complex' composed of Axin, adenomatous polyposis coli (APC) and glycogen synthase kinase 3β (GSK3β), thereby β-catenin is stabilized and translocated into nucleus to form β-catenin-LEF/TCF transcriptional complex complexes, thus to activate downstream oncogenes such as CCND1 and c-Myc, ultimately brings about the development of human malignancies. 18,19 Inversely, in the absence of Wnt ligands, β-catenin is sequestered by the 'destruction complex' and prone to be phosphorylate followed by the degradation via ubiquitination, which ultimately inactivates the β-catenin mediated Wnt signalling.
LRP5 has been proved to be less effective than LRP6 in transducing Wnt signal, perhaps owing to the fact that the PPPSPxS motifs, which are the cytoplasmic domains of LRP5/6 and provide inducible docking sites for Axin, are easier to be phosphorylated in LRP6 than in LRP5. 20,21 However, mutation or polymorphism of LRP5 gene is still closely associated with the defect of tissue homeostasis and pathogenesis of multiple diseases. [22][23][24][25] The aberrant expression of LRP5 has been reported to participate in the development of several human malignancies. [26][27][28][29] However, its role in the progression of different cancers has shown some discrepancies and needs to be determined in specific type of cancers. At present, the expression pattern and role of LRP5 in CRC have not been documented and urgently need investigation.
Here, we report LRP5 plays a vital oncogenic role in CRC development and contributes to the chemoresistance of CRC cells, partially through activating the canonical Wnt/β-catenin and IL-6/ STAT3 signalling pathways and consequently promoting CSCs phenotypes. Clinically, LRP5 is overexpressed in CRC tissues and some corresponding cell lines, the expression of endogenous LRP5 gene is closely associated with that of CSCs-related gene PROM1 (encoding CD133) and represents a valuable diagnostic biomarker for CRC.
Moreover, silencing the endogenous LRP5 gene significantly inhibits the tumorigenicity of CRC, increases the chemosensitivity and apoptosis of CRC cells. Therefore, our study proposed a novel insight into the mechanism of CRC and provides a promising biomarker and target for the diagnosis and treatment of CRC.

| Immunohistochemistry (IHC)
The CRC tissue microarray embedded in paraffin was purchased from Shanghai Outdo biotech (Shanghai, China), and IHC was carried out as described. In brief, the section was deparaffinized by incubation at 65°C for 1 h and further rehydrated through graded ethanols, and the antigen retrieval was performed by boiling the section for 30 min in 1-mM EDTA buffer (pH 7.5), followed by treated with 3% H 2 O 2 to quench the endogenous peroxidase.
After incubation with 5% sheep serum albumin for 30 min to block the nonspecific binding, the section was incubated overnight with a primary antibody against LRP5 protein (Abcam, UK) (1:500) at 4°C; after rinsed with PBS for three times, the section was then incubated with the secondary antibody for 30 min at 37°C.
Finally, the section was incubated with 3′3-diaminobenzidine tetra-hydrochloride (DAB) to develop colour and then counterstained with haematoxylin. Meanwhile, IHC images of LRP5 protein expression in normal colorectal tissues and CRC ones were downloaded and examined from the Human Protein Atlas (HPA) (http://www.prote inatl as.org/).

| Selecting stable CRC cells with the activation or knockdown of LRP5
The CRISPR/Cas9 KO plasmid or CRISPR activation plasmid that disrupts or activates the endogenous LRP5 gene was purchased (Santa Cruz, USA). HCT-116 cells were plated at a density of 3 × 10 5 cells per well of a 6-well plate 24 h before transfection. LRP5-CRISPR/ Cas9 KO plasmid (LRP5-KO), LRP5-CRISPR activation plasmid (LRP5-ACT) or corresponding controls were transfected into cells separately by using lipofectamine 3000 (Invitrogen, USA) in serumfree medium; then, cells were cultured in medium containing 5 μg/ ml puromycin for two weeks, and single cell was selectively by using flow cytometer followed by transferred to 96-well plate for expansion. Cell clones exhibiting the most efficient downregulation or upregulation of LRP5 compared with corresponding controls were verified by qRT-PCR and immunoblot analysis and selected for subsequent analysis.

| RNA isolation and qRT-PCR analysis
RNA extraction and qRT-PCR were performed according to our previous work. 30 Sequences of primers used for real-time PCR are listed in Table S1.

| Protein extraction and immunoblot
Protein extraction from cells or tumour tissues and immunoblot were carried out according to our previous work. 31

| Cell proliferation assays
To evaluate cell proliferation ability, stable cell clones were plated at a density of 4 × 10 3 cells per well (ACT groups) or 9 ×  To analyse the influence on apoptosis-related genes, stable LRP5-KO or LRP5-ACT HCT-116 cells and their controls were seeding into 6-well plate and incubated for 24 h, followed by treated with 8μg/ml or 12μg/ml cisplatin for another 36 h. Cells were harvested for analysing the mRNA levels of pro-apoptotic genes, such as BAX,

| Wound healing migration assay
Cells were seeded into 6-well plates (2 × 10 5 cells/well) and cultured till reaching 80% confluence; then, cells were scraped by a 100μl pipette tip. The plate was washed with PBS to remove the floating cells completely and replaced with fresh medium contains 2% FBS. Cells were cultured and photographed every 24 h to monitor the migrated distance. The relative migration rate was calculated by deducting w 48h/96 h (the final width) from w 0 (the initial width) and normalizing to the initial width.

| Colony formation assay
A total of 500 cells (ACT groups) or 800 cells (KO groups) per well were seeded into a 6-well plate and cultured at 37°C for 14 days (ACT groups) or 10 days (KO groups). After washing with PBS for three times and fixed with 75% ethanol, cells were stained with 0.1% crystal violet. The number of colonies was counted automatically by using Image Pro Plus (IPP) software (Media Cybernetics).

| Xenograft tumour assay in nude mice
Female BALB/c nude mice at 4 weeks old were purchased from Beijing Vital River Laboratories (Beijing, China) and fed with standard rodent chow under specific pathogen-free (SPF) conditions. Mice were subcutaneously injected with 2 × 10 6 (ACT groups) or 5 × 10 6 (KO groups) HCT-116 cells resuspended in 150 μl of serumfree PBS. Tumour size was measured every three days to estimate tumour volume (length × width × width) using a Vernier calliper.
Mice were anaesthetized 24 days later and sacrificed to harvest all tumours. To evaluate the effect of LRP5 activation on the drug resistance of CRC cells to chemotherapeutic agents in vivo, the nude mice were implanted with 4 × 10 6 LRP5-ACT HCT-116 cells or control cells resuspended in 150 μl of serum-free PBS, followed by received i.p. treatment of 1 mg/kg cisplatin every three days from the third day of transplantation. The body weight of the nude mice was measured every three days, and all mice were humanely sacrificed on day 24 and the tumour weight was measured. All animal studies were conducted with approval from the Biomedical Research Ethics Committee of Henan University. Kaplan-Meier curves of RFS and OS were drawn to evaluate the difference in survival outcomes between patients with low or high LRP5 expression. The odds ratio (OR) with 95% confidence interval (CI) and p-value for the LogRank test were calculated to evaluate the association between the clinicopathological parameters and RFS/OS in CRC patients. The correlation among the expression of LRP5, PROM1 (CD133) and STAT3 was evaluated by Pearson correlation analysis, and error bar indicates the standard deviation of mean value. A p-value less than 0.05 was considered statistically significant. One, two and three asterisks represent p-values of < 0.05, 0.01 and 0.001 respectively.

| LRP5 is overexpressed in CRC cell lines and tissues
Several studies have reported that LRP5 is crucial for the tumori- Importantly, we reconfirmed the results through analysing data from TCGA ( Figure 1C). Other oncomine analysis of CRC tissues versus normal ones also proved LRP5 was overexpressed in CRC (Table 1).
However, there was no obvious difference in LRP5 mRNA levels in patients with different survival outcomes ( Figure 1D). In addition, we performed IHC in tissue microarray containing paired CRC specimens and normal ones. As shown in Figure 1E, a similar upregulation of LRP5 protein was detected in CRC tissues with different pathological grades compared with that in normal colorectal tissues ( Figure 1E). Similarly, after analysing the images for IHC staining from the online HPA database, we found CRC tissues had strong LRP5 staining but normal colorectal tissues had low LRP5 staining ( Figure 1F).

| Correlation between the LRP5 mRNA levels and clinicopathological parameters in CRC tissues
The mRNA levels of LRP5 in different pathological stages were further assessed by analysing the data for CRC from TCGA database.
The results showed that the LRP5 mRNA level was obviously elevated in advanced stages (stage Ⅲ and stage Ⅳ) compared with that in early stages (stage Ⅰ and stage Ⅱ) (Figure 2A). However, as shown in Figure 2B, no difference in LRP5 mRNA level was found between tumour grade in small (T1 and T2) and tumour grade in large (T3 and T4). CRC tissues with involvement of local lymph node (N1/N2/N3) or metastasis displayed a higher level of LRP5 mRNA level compared with that without involvement of local lymph node or metastasis ( Figure 2C, D). To validate the association between LRP5 mRNA level and OS or RFS, we mapped the Kaplan-Meier curves of OS and RFS and no obvious association was found between them, even though it seemed CRC patients with low LRP5 mRNA level had superior OS and RFS compared with those with high LRP5 mRNA level ( Figure 2E, F).
We subsequently assessed the independent prognostic value of LRP5 mRNA level in CRC. The association between the LRP5 expression levels and various clinicopathological characteristics of CRC patients was summarized in Table 2

| Activation of LRP5 promotes the tumorigenicity of CRC cells
To investigate whether activation of LRP5 has a potential carcinogenic effect on CRC cells, we used LRP5-CRISPR activation plasmid to construct cell lines with stable activation of endogenous transcription of LRP5 gene in HCT-116 cells with puromycin selection. As shown in Figure 3A, B, both the transcriptional and translational levels of LRP5 were elevated in LRP5-ACT group. To determine the effect of LRP5 activation on proliferative ability of CRC cells, the cellular proliferative indices were real-time monitored by using the RTCA system. Activation of LRP5 obviously accelerated the proliferation of HCT-116 cells ( Figure 3C). We further performed the wound-healing assay to test the effect of LRP5 activation on cellular migration. As shown in Figure 3D, E, activation of LRP5 led to a significant increase in migratory capacity.
To explore the effect of LRP5 activation on self-renewal capacity of CRC cells, a key feature of CSCs, we cultured the cells in serum-free medium to generate typical tumourspheres. As shown in Figure 3F Figure 4B). This might be the main reasons for the changes in cellular activities, including the enhanced tumoursphere-formation and with those in CD133 − ones ( Figure 4C). By analysing the GSE34053 data set, 33 we further confirmed that LRP5 mRNA level was upregulated in CD133 + CRC cells, compared with that in CD133 − ones (p = 0.071) and carcinoma-associated fibroblasts (p = 0.017) isolated from the same stage II colon cancer patient specimen ( Figure 4D).
Interestingly, we found an obviously positive correlation between the mRNA levels of LRP5 and PROM1 (CD133) (p = 0.015, R = 0.117) through analysing the data on CRC and normal tissues from the same TCGA database ( Figure 4E), further confirmed the positive regulation of LRP5 activation on stemness of CRC.
The IL-6/STAT3 inflammatory signalling axis was proved to promote stem-like properties of CRC, and we further questioned whether the upregulation of stemness genes induced by the activation of LRP5 could be partially attributed to the activation of IL-6/ STAT3 pathway. As shown in Figure 5A, B, we were surprised to find that the mRNA levels of IL-6 and STAT3 were both upregulated in

| Activation of LRP5 induces the drug resistance of CRC cells to chemotherapeutic agents
The chemotherapeutic resistance is the key characteristics of

| Silencing of LRP5 suppresses the tumorigenicity of CRC cells
Since LRP5 has a carcinogenic role in CRC development, we ex-

| Silencing of LRP5 suppresses the stemness of CRC by inhibiting the Canonical Wnt/β-catenin pathway and IL-6/STA3 pathway
We further explored the effect of LRP5 knockdown on the canonical Wnt/β-catenin pathway. As shown in Figure 8A, knockdown of LRP5 inhibited the mRNA levels of some crucial genes in this cascade, including CTNNB1, c-Myc, CCND1 and COX2. In contrast, the expression of its upstream molecule, such as Wnt3, was not affected. We   Figure 9C). Taken together, these results illustrate that silencing of LRP5 promotes the sensitivity of CRC cells to chemotherapeutic agent and its induced cell apoptosis, and targeting LRP5 could be a valuable strategy for CRC treatment.

| DISCUSS ION
At present, the role of canonical Wnt/β-catenin signalling pathway in the initiation and progression of malignancies has been determined and attracted much attention, whereas the components of  and provide a tumour-promoting microenvironment for tumour progression and metastasis. 54 An obvious positive correlation was also found between the mRNA levels of LRP5 and PROM1 (CD133) in CRC tissues. These data suggested that the overactivation of canonical Wnt/β-catenin signalling in colorectal CSCs could be attributed to the upregulation of LRP5. Activation of LRP5 also promoted the self-renewal capacity of CSCs by upregulating the expression of stemness-related genes, including PROM1 (CD133). We speculate that the activation of canonical Wnt/β-catenin signalling induced by LRP5 overexpression expands the CSCs population and strengthens their stemness-related functions, which further upregulate LRP5 expression and eventually forms a vicious circle. Therefore, LRP5 is the intermediary factor that mediates the mutual activation between the canonical Wnt signalling and the stemness of CSCs in CRC.
Angiogenesis also plays an important role in the progression of CRC, and CSCs derived from CRC can generate functional cancer blood vessels and enhance tumour neovascularization by expressing some angiogenic factors. 55 The expression level of certain well-known human CSC marker, such as CD133, is also positively associated with the angiogenesis of CRC. 56 Therefore, the activation of LRP5 might induce the angiogenesis of CRC by facilitating the self-renewal of CSCs and needs further investigation. Our data also suggested that properties. 61 In lung cancer, activation of Akt1/IL-6/STAT3 pathway also contributes to maintaining the stemness of tumour initiating cells. 62 Therefore, we speculated that the enhancement of stemness phenotype induced by LRP5 activation might be partly ascribed to the overactivation of IL-6/STAT3 signalling. In order to explore this hypothesis, we analysed the relationship among the expression of LRP5, STAT3 and PROM1 (CD133) genes in TCGA database. As expected, we found the mRNA levels of LRP5 and PROM1 (CD133) were both positively associated with the mRNA levels of STAT3. We subsequently treated the LRP5-KO CRC cells with exogenous IL-6 protein. However, the downregulation of the expression of stemness-related genes, induced by silencing of LRP5, was not reversed after the IL-6 treatment (data not shown). Taken together, these findings suggest that silencing of LRP5 could inhibit CSC-like phenotype by blocking the canonical Wnt/β-catenin and IL-6/STAT3 pathways, and targeting the overexpressed LRP5 might be a promising therapeutic approach for CRC.

| CON CLUS IONS
In summary, we demonstrated that the expression of LRP5 was upregulated in CRC tissues and cells, and activation of LRP5 exerted a carcinogenic effect on the progression of CRC. On the contrary, silencing of LRP5 suppressed the tumorigenicity and sensitized CRC to cisplatin treatment by targeting Wnt/β-catenin mediated CSC properties and IL-6/STAT3 pathway. Therefore, these findings broaden our knowledge of CRC development and provide a valuable therapeutic target for CRC research.