Literature DB >> 29035371

lncRNA MIR100HG-derived miR-100 and miR-125b mediate cetuximab resistance via Wnt/β-catenin signaling.

Yuanyuan Lu1,2, Xiaodi Zhao1,2, Qi Liu3, Cunxi Li4, Ramona Graves-Deal1, Zheng Cao1, Bhuminder Singh1, Jeffrey L Franklin1, Jing Wang3, Huaying Hu4, Tianying Wei4, Mingli Yang5, Timothy J Yeatman5, Ethan Lee6, Kenyi Saito-Diaz6, Scott Hinger7, James G Patton7, Christine H Chung8, Stephan Emmrich9, Jan-Henning Klusmann9, Daiming Fan2, Robert J Coffey1,10.   

Abstract

De novo and acquired resistance, which are largely attributed to genetic alterations, are barriers to effective anti-epidermal-growth-factor-receptor (EGFR) therapy. To generate cetuximab-resistant cells, we exposed cetuximab-sensitive colorectal cancer cells to cetuximab in three-dimensional culture. Using whole-exome sequencing and transcriptional profiling, we found that the long non-coding RNA MIR100HG and two embedded microRNAs, miR-100 and miR-125b, were overexpressed in the absence of known genetic events linked to cetuximab resistance. MIR100HG, miR-100 and miR-125b overexpression was also observed in cetuximab-resistant colorectal cancer and head and neck squamous cell cancer cell lines and in tumors from colorectal cancer patients that progressed on cetuximab. miR-100 and miR-125b coordinately repressed five Wnt/β-catenin negative regulators, resulting in increased Wnt signaling, and Wnt inhibition in cetuximab-resistant cells restored cetuximab responsiveness. Our results describe a double-negative feedback loop between MIR100HG and the transcription factor GATA6, whereby GATA6 represses MIR100HG, but this repression is relieved by miR-125b targeting of GATA6. These findings identify a clinically actionable, epigenetic cause of cetuximab resistance.

Entities:  

Mesh：

Substances：

Year:  2017        PMID： 29035371      PMCID： PMC5961502          DOI： 10.1038/nm.4424

Source DB:  PubMed          Journal:  Nat Med        ISSN： 1078-8956            Impact factor:   53.440


Colorectal cancer (CRC) remains a leading cause of cancer-related death worldwide[1]. Cetuximab and panitumumab are EGF receptor (EGFR) monoclonal antibodies (mAbs) that bind the extracellular domain of EGFR and enhance receptor internalization and degradation. These EGFR mAbs are common targeted agents for patients with wild-type KRAS metastatic CRC. As monotherapy, 12–17% patients have durable responses[2] and up to 72% response rates are reported when combined with chemotherapy[3]. However, drug resistance frequently arises. Intense efforts have led to identification of many de novo and acquired genetic mechanisms of resistance to EGFR mAb therapy, including KRAS, NRAS, BRAF, PIK3CA, and EGFR mutations[2,4,5]. However, little is known about non-genetic resistance mechanisms. Non-coding RNAs (ncRNAs), in particular long non-coding RNAs (lncRNAs) and microRNAs (miRNAs), play crucial roles in epigenetic regulation[6,7]. Recently, a complex interplay between these two classes of regulatory ncRNAs has been discovered in which some lncRNAs are processed to produce miRNAs that repress target mRNAs[8,9]. For example, lncRNA H19-derived miR-675 suppresses translation of insulin growth factor receptor (Igf1r), inhibiting cell proliferation in response to cellular stress or oncogenic signals[10]. miR-17∼92, generated from the lncRNA MIR17HG locus, attenuates TGF-β signaling to stimulate angiogenesis and tumor growth[11]. The lncRNA MIR100HG-derived miR-100/let-7a-2/miR-125b-1 and MIR99AHG-derived miR-99a/let-7c/miR-125b-2 clusters participate in the pathogenesis of acute megakaryoblastic leukemia[12,13]. However, whether these lncRNAs or derived miRNAs contribute to drug resistance is largely unknown. Herein, we identify a role for lncRNA MIR100HG and two embedded miRNAs, miR-100 and miR-125b, in conferring cetuximab resistance. We show that MIR100HG and miR-100/125b are overexpressed in the setting of de novo and acquired cetuximab resistance in CRC and head and neck squamous cell cancer (HNSCC) cell lines. miR-100 and miR-125b coordinately downregulate five negative regulators (DKK1, DKK3, ZNRF3, RNF43, APC2) of canonical Wnt/β-catenin signaling (hereafter Wnt signaling), leading to increased Wnt signaling. Wnt inhibition restores responsiveness to cetuximab in vitro and in vivo. We show these events occur in CRC patients whose tumors progressed on cetuximab. We also discovered that MIR100HG overexpression is reinforced by miR-125b suppression of GATA6, which in turn represses MIR100HG. These studies identify an epigenetic cause of cetuximab resistance with diagnostic and therapeutic implications.

Results

Establishment of cetuximab-resistant cells in three-dimensional (3D) culture

By placing single cells from a human KRAS/NRAS/BRAF wild-type, microsatellite unstable CRC cell line, HCA-7, into 3D culture in type-1 collagen, a line was derived from colonies with cystic morphology and designated cystic colonies (CC) [14,15]. Proliferation of CC was inhibited by cetuximab in 3D culture but not in 2D plastic culture[14]. Upon continuous exposure to cetuximab in 3D culture for approximately 4 months, a line was generated and designated CC-cetuximab resistant (CC-CR) (Fig. 1a). In 2D culture, CC and CC-CR were morphologically indistinguishable. In 3D, however, CC formed hollow cysts with a central lumen lined by a monolayer of polarized cells, whereas CC-CR formed solid disorganized colonies (Fig. 1b). As expected, cetuximab inhibited CC growth in 3D, while CC-CR remained refractory to cetuximab up to 200 μg/ml (Fig. 1c; Extended Data Fig. 1a and b). Decreased expression of the proliferative marker, Ki-67, and increased expression of the apoptotic marker, cleaved Caspase-3, were observed in CC 24 h after cetuximab treatment, but these indices were unaffected in CC-CR (Fig. 1d). In cetuximab-treated CC, we observed reduced levels of p-EGFR, p-ERK1/2, p-AKT and Cyclin D1, as well as increased cleaved Caspase-3 and the pro-apoptotic marker, BIM; these markers were largely unaffected in cetuximab-treated CC-CR (Fig. 1e). Next, CC and CC-CR were stably transduced with a green fluorescent protein (GFP)-expressing lentiviral vector and injected subcutaneously into athymic nude mice. CC tumors were well differentiated and regressed upon administration of cetuximab. In contrast, CC-CR tumors were poorly differentiated and continued to grow in the presence of cetuximab, although not to the extent of untreated tumors (Fig. 1f and g; Extended Data Fig. 1c-e).
Figure 1

Characterization of cetuximab-resistant CC (CC-CR) in 3D

(a) Schematic of experimental approach to establish cetuximab (CTX)-resistant cells in 3D. In the presence of CTX (3 μg/ml) in 3D type-1 collagen culture, greater than 95% of CC colonies die. Residual colonies were isolated and iteratively passaged in 2D and 3D in the continued presence of CTX over approximately 4 months. These colonies were designated CC-CR. (b) Top: differential interference contrast (DIC) and confocal images of representative CC and CC-CR in 2D and 3D. F-actin was stained with phalloidin (red). Scale bars: 400, 1000, 200, 50 μm, respectively (from left to right). Bottom: left, number of nuclei in the midplane of each colony; right, the morphology of colonies was divided into those with luminal, multi-layered, or solid morphology. n=4 independent experiments, *P<0.05 by Student's t test. (c) CC and CC-CR were cultured in 3D in the presence or absence of CTX (3 μg/ml) and colonies were counted after 18 days. n=3 independent experiments performed in triplicate. **P<0.01 by Student's t test. (d) CC and CC-CR cells were cultured in 3D for 12 days and treated with CTX (10 μg/ml) for 24 h. Ki-67 (red) and Cleaved Caspase-3 (Cleaved Casp-3, green) staining were imaged by confocal microscopy. Representative of 4 independent experiments. Scale bar, 50 μm. Quantification is shown on the right (n=4). **P<0.01 by Student's t test. (e) Immunoblots of 3D cell lysates from CC and CC-CR treated with CTX (10 μg/ml) for indicated time. β-actin served as the loading control. A representative blot from 3 independent experiments is shown. (f) Nude mice (n=8) bearing subcutaneous tumors were treated with control saline or CTX at a dose of 1 mg/mouse, intraperitoneal (i.p.) injection, every 3 days. Tumor volumes were measured every 3 days using calipers. **P<0.01 by repeated-measures ANOVA test followed by LSD post-hoc test. (g) Representative immunohistochemical images of Ki-67 and Cleaved Casp-3 from CC and CC-CR xenografts before and after CTX treatment. Scale bar: 50 μm. Data represent mean ± s.d. in b-d and f. n.s., not significant.

Extended Data Fig. 1

Establishment of cetuximab-resistant cells in 3D culture.

(a) Five thousand cells/ml were cultured in type-1 collagen for 17 days. Fresh medium was added with different concentrations of CTX every 2 days, and colony number was determined using a GelCount plate reader. n=2 independent experiments performed in triplicate. (b) Twelve-day old CC and CR were treated with CTX (10 μg/ml) for 3 days. Representative images from 3 independent experiments are shown. Scale bars, 1000 μm. (c) Left: Representative fluorescence images of GFP signals captured from subcutaneous tumors, generated by injection of CC and CC-CR stably transduced with GFP-expressing lentivirus. Right: Quantification of radiant efficiency from tumors. n=8. **P<0.01 by paired Student's t test. (d) Representative H&E staining of the tumor xenografts from the indicated groups. Scale bar, 100 μm. (e) Quantification of IHC staining in Fig. 1g. n=8 mice. **P<0.01 by Student's t test. (f) CC and CC-CR cells grown on Transwell filters were incubated with Alexa Fluor 488-labeled C225 mAb directed against the extracellular domain of EGFR and then stained for F-actin (Phalloidin) and nuclei (DAPI). Scale bars, 20 μm. Data represent mean ± s.d. in a, c, and e.

Upregulation of lncRNA MIR100HG and embedded miR-100/125b in cetuximab-resistant cells

We first considered known mechanisms of cetuximab resistance in this 3D model. By whole exome sequencing and RNA Sequencing (RNA-Seq), no known genetic events linked to cetuximab resistance were found, including all reported gene mutations, copy number changes and gene fusion events (Extended Data Table 1). By RNA-Seq, we found 141 transcripts upregulated and 220 transcripts downregulated in CC-CR compared to CC (fold change>2 and false-discovery rate, FDR<0.01). Expression levels of ERBB1-4, the 7 EGFR ligands, and MET were comparable between CC and CC-CR (Extended Data Table 2). Immunofluorescence also showed equivalent cell-surface EGFR staining in CC and CC-CR (Extended Data Fig. 1f). Small RNA-Seq detected 7 miRNAs upregulated and 24 miRNAs downregulated in CC-CR compared to CC (fold change>2 and FDR<0.01). Of note, the most upregulated transcript in CC-CR was lncRNA MIR100HG, and the two most upregulated miRNAs were miR-125b and miR-100 (Fig. 2a).
Extended Data Table 1
ChrPosRefAltTypeAAchangeDNAchangeRefseqIDGeneCC_GTCC_CR_GT
chr11.17E+08GANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)T875I;T895IaCc/aTc;aCc/aTcNM_001007237.1;NM_001542.2IGSF3;IGSF3Homo_refHomo_ALT
chr1674537530CANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)A214S;A225S;A225SGcc/Tcc;Gcc/Tcc;Gcc/TccNM_001145666.1;NM_001145667.1;NM_012201.5GLG1;GLG1;GLG1Homo_refHomo_ALT
chrX48762192TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)I332V;I332VAtt/Gtt;Att/GttNM_001042498.2;NM_005660.1SLC35A2;SLC35A2Homo_refHomo_ALT
chrX1.34E+08CANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)R22I;R22I;R22I;R22IaGa/aTa;aGa/aTa;aGa/aTa;aGa/aTaNM_001166600.2;NM_001166599.2;NM_145284.5;NM_001170756.1FAM122B;FAM122B;FAM122B;FAM122BHomo_refHomo_ALT
chrX1.35E+08AGNON_SYNONYMOUS_CODING(MODERATE)I1667MatA/atGNM_153834.3GPR112Homo_refHomo_ALT
chr11254773GTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)A111D;A117D;A82DgCc/gAc;gCc/gAc;gCc/gAcNM_017871.5;NM_001256456.1;NM_001256460.1CPSF3L;CPSF3L;CPSF3LHomo_refHeter
chr117740089CTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)R384H;R384HcGt/cAt;cGt/cAtNM_001136204.2;NM_018715.3RCC2;RCC2Homo_refHeter
chr132653705CTNON_SYNONYMOUS_CODING(MODERATE)R250WCgg/TggNM_175852.3TXLNAHomo_refHeter
chr147024318GANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)R402C;R443CCgc/Tgc;Cgc/TgcNM_001135553.2;NM_003684.5MKNK1;MKNK1Homo_refHeter
chr151210391ACNON_SYNONYMOUS_CODING(MODERATE)L142VTta/GtaNM_007051.2FAF1Homo_refHeter
chr193303020CTSTOP_GAINED(HIGH)R179*Cga/TgaNM_000969.3RPL5Homo_refHeter
chr11.46E+08GANON_SYNONYMOUS_CODING(MODERATE)G410EgGa/gAaNM_144698.3ANKRD35Homo_refHeter
chr11.55E+08TANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)M411K;M427K;M427K;M427K;M427K;M427K;M427K;M427K;M437KaTg/aAg;aTg/aAg;aTg/aAg;aTg/aAg;aTg/aAg;aTg/aAg;aTg/aAg;aTg/aAg;aTg/aAgNM_001261466.1;NM_207191.2;NM_001261465.1;NM_003815.4;NM_207195.2;NM_207194.2;NM_207196.2;NM_207197.2;NM_001261464.1ADAM15;ADAM15;ADAM15;ADAM15;ADAM15;ADAM15;ADAM15;ADAM15;ADAM15Homo_refHeter
chr11.86E+08AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)K263E;K272E;K295EAaa/Gaa;Aaa/Gaa;Aaa/GaaNM_001164246.1;NM_001164245.1;NM_017847.5C1orf27;C1orf27;C1orf27Homo_refHeter
chr12.27E+08CANON_SYNONYMOUS_CODING(MODERATE)E690DgaG/gaTNM_001618.3PARP1Homo_refHeter
chr12.49E+08GANON_SYNONYMOUS_CODING(MODERATE)G293RGgg/AggNM_024836.1ZNF672Homo_refHeter
chr256420411CTNON_SYNONYMOUS_CODING(MODERATE)P359LcCg/cTgNM_001080433.1CCDC85AHomo_refHeter
chr265541153CTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)A244T;A247TGcg/Acg;Gcg/AcgNM_001128210.1;NM_181784.2SPRED2;SPRED2Homo_refHeter
chr284668385GTNON_SYNONYMOUS_CODING(MODERATE)P173TCct/ActNM_003849.3SUCLG1Homo_refHeter
chr298999879AGNON_SYNONYMOUS_CODING(MODERATE)N142DAac/GacNM_001298.2CNGA3Homo_refHeter
chr299797317TANON_SYNONYMOUS_CODING(MODERATE)D43VgAt/gTtNM_138798.1MITD1Homo_refHeter
chr21.44E+08GTNON_SYNONYMOUS_CODING(MODERATE)S217IaGt/aTtNM_018460.3ARHGAP15Homo_refHeter
chr21.6E+08GTNON_SYNONYMOUS_CODING(MODERATE)S2108YtCc/tAcNM_013450.2BAZ2BHomo_refHeter
chr21.98E+08ATNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)F317I;F317ITtc/Atc;Ttc/AtcNM_001206774.1;NM_012086.4GTF3C3;GTF3C3Homo_refHeter
chr22.36E+08ATNON_SYNONYMOUS_CODING(MODERATE)Y479FtAc/tTcNM_014521.2SH3BP4Homo_refHeter
chr330713286GANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)G204D;G229DgGc/gAc;gGc/gAcNM_003242.5;NM_001024847.2TGFBR2;TGFBR2Homo_refHeter
chr348604415ACNON_SYNONYMOUS_CODING(MODERATE)N2717KaaT/aaGNM_000094.3COL7A1Homo_refHeter
chr348604422CANON_SYNONYMOUS_CODING(MODERATE)S2715IaGt/aTtNM_000094.3COL7A1Homo_refHeter
chr349689732GANON_SYNONYMOUS_CODING(MODERATE)A915TGca/AcaNM_003458.3BSNHomo_refHeter
chr349765037GTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)R117S;R282S;R282SCgt/Agt;Cgt/Agt;Cgt/AgtNM_001006115.2;NM_001242829.1;NM_153273.3IP6K1;IP6K1;IP6K1Homo_refHeter
chr375788432TANON_SYNONYMOUS_CODING(MODERATE)Q114HcaA/caTNM_001128223.1ZNF717Homo_refHeter
chr31.01E+08AGNON_SYNONYMOUS_CODING(MODERATE)I247TaTc/aCcNM_016247.3IMPG2Homo_refHeter
chr31.88E+08GANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)A303T;A303TGca/Aca;Gca/AcaNM_001167671.1;NM_005578.3LPP;LPPHomo_refHeter
chr31.9E+08TCNON_SYNONYMOUS_CODING(MODERATE)W653RTgg/CggNM_001167931.1IL1RAPHomo_refHeter
chr45577982TANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)E1006V;E1086VgAg/gTg;gAg/gTgNM_001166136.1;NM_147127.4EVC2;EVC2Homo_refHeter
chr474021398AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)V317A;V317AgTg/gCg;gTg/gCgNM_198889.1;NM_032217.3ANKRD17;ANKRD17Homo_refHeter
chr489709004CTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)D37N;D391N;D51N;D65N;D65NGac/Aac;Gac/Aac;Gac/Aac;Gac/Aac;Gac/AacNM_001265580.1;NM_014883.3;NM_001265578.1;NM_001265579.1;NM_001015045.2FAM13A;FAM13A;FAM13A;FAM13A;FAM13AHomo_refHeter
chr41.26E+08TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)Q1212R;Q1391RcAg/cGg;cAg/cGgNM_001167882.1;NM_020337.2ANKRD50;ANKRD50Homo_refHeter
chr41.55E+08AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)F177L;F771LTtt/Ctt;Ttt/CttNM_017639.3;NM_001142552.1DCHS2;DCHS2Homo_refHeter
chr41.56E+08TANON_SYNONYMOUS_CODING(MODERATE)I187KaTa/aAaNM_004744.3LRATHomo_refHeter
chr41.56E+08CTSPLICE_SITE_DONOR(HIGH)NM_001039580.1MAP9Homo_refHeter
chr5434878GANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)G679R;G697RGgg/Agg;Ggg/AggNM_001242412.1;NM_020731.4AHRR;AHRRHomo_refHeter
chr523527585GACACACACAGACACACACACAFRAME_SHIFT(HIGH)-800?-/CANM_020227.2PRDM9Homo_refHeter
chr531799521ATNON_SYNONYMOUS_CODING(MODERATE)T56SAcg/TcgNM_178140.2PDZD2Homo_refHeter
chr554624530TCNON_SYNONYMOUS_CODING(MODERATE)Y136HTac/CacNM_015360.4SKIV2L2Homo_refHeter
chr51.32E+08AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)I124T;I124T;I124T;I64TaTc/aCc;aTc/aCc;aTc/aCc;aTc/aCcNM_015146.1;NM_001098812.1;NM_001098811.1;NM_001098813.1SEPT8;SEPT8;SEPT8;SEPT8Homo_refHeter
chr51.4E+08TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)V60A;V60AgTg/gCg;gTg/gCgNM_031500.1;NM_018907.2PCDHA4;PCDHA4Homo_refHeter
chr51.4E+08TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)V552A;V552AgTg/gCg;gTg/gCgNM_031501.1;NM_018908.2PCDHA5;PCDHA5Homo_refHeter
chr51.41E+08ATNON_SYNONYMOUS_CODING(MODERATE)E38VgAa/gTaNM_020957.1PCDHB16Homo_refHeter
chr51.79E+08CGNON_SYNONYMOUS_CODING(MODERATE)D22HGac/CacNM_001142306.1C5orf60Homo_refHeter
chr6553900GANON_SYNONYMOUS_CODING(MODERATE)S692FtCc/tTcNM_018303.5EXOC2Homo_refHeter
chr625972305GANON_SYNONYMOUS_CODING(MODERATE)G239DgGc/gAcNM_006355.3TRIM38Homo_refHeter
chr639073484GTSTOP_GAINED(HIGH)C92*tgC/tgANM_018322.1SAYSD1Homo_refHeter
chr643267448TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)V194A;V196AgTa/gCa;gTa/gCaNM_006672.3;NM_153320.2SLC22A7;SLC22A7Homo_refHeter
chr61.61E+08ATNON_SYNONYMOUS_CODING(MODERATE)K2078IaAa/aTaNM_000876.2IGF2RHomo_refHeter
chr61.71E+08AGNON_SYNONYMOUS_CODING(MODERATE)W159RTgg/CggNM_005618.3DLL1Homo_refHeter
chr71538645AGNON_SYNONYMOUS_CODING(MODERATE)V368AgTg/gCgNM_001080453.2INTS1Homo_refHeter
chr748349667CTNON_SYNONYMOUS_CODING(MODERATE)P3149SCca/TcaNM_152701.3ABCA13Homo_refHeter
chr791603099ACNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)K41N;K41NaaA/aaC;aaA/aaCNM_147185.2;NM_005751.4AKAP9;AKAP9Homo_refHeter
chr791756945CANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)V157F;V52FGtt/Ttt;Gtt/TttNM_000786.3;NM_001146152.1CYP51A1;CYP51A1Homo_refHeter
chr797863036TCNON_SYNONYMOUS_CODING(MODERATE)T457AAcc/GccNM_015395.2TECPR1Homo_refHeter
chr71E+08TTGFRAME_SHIFT(HIGH);FRAME_SHIFT(HIGH)-1923?;-1923?-/G;-/GNM_173059.1;NM_003386.1ZAN;ZANHomo_refHeter
chr71.05E+08ACNON_SYNONYMOUS_CODING(MODERATE)V613GgTt/gGtNM_019042.3PUS7Homo_refHeter
chr71.27E+08CASTOP_GAINED(HIGH)E48*Gag/TagNM_176814.3ZNF800Homo_refHeter
chr71.42E+08GANON_SYNONYMOUS_CODING(MODERATE)R1437KaGg/aAgNM_004668.2MGAMHomo_refHeter
chr71.49E+08GANON_SYNONYMOUS_CODING(MODERATE)A66VgCc/gTcNM_207336.1ZNF467Homo_refHeter
chr8623772CTNON_SYNONYMOUS_CODING(MODERATE)D194NGac/AacNM_207332.1ERICH1Homo_refHeter
chr835453086TGNON_SYNONYMOUS_CODING(MODERATE)F161VTtt/GttNM_080872.2UNC5DHomo_refHeter
chr81.42E+08GANON_SYNONYMOUS_CODING(MODERATE)R388CCgc/TgcNM_001080431.1SLC45A4Homo_refHeter
chr81.45E+08GTNON_SYNONYMOUS_CODING(MODERATE)M60IatG/atTNM_001916.3CYC1Homo_refHeter
chr934255869AGNON_SYNONYMOUS_CODING(MODERATE)C1246RTgc/CgcNM_194313.2KIF24Homo_refHeter
chr986452279ATNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)S1184R;S1215R;S1281RagT/agA;agT/agA;agT/agANM_001271928.1;NM_001271927.1;NM_017576.2KIF27;KIF27;KIF27Homo_refHeter
chr996052320TCNON_SYNONYMOUS_CODING(MODERATE)S1643PTca/CcaNM_006648.3WNK2Homo_refHeter
chr91.16E+08TGNON_SYNONYMOUS_CODING(MODERATE)Q119PcAg/cCgNM_015258.1FKBP15Homo_refHeter
chr91.25E+08AGNON_SYNONYMOUS_CODING(MODERATE)K81RaAg/aGgNM_001005235.1OR1L4Homo_refHeter
chr91.28E+08CGNON_SYNONYMOUS_CODING(MODERATE)R438TaGa/aCaNM_002077.3GOLGA1Homo_refHeter
chr91.34E+08AGNON_SYNONYMOUS_CODING(MODERATE)M114VAtg/GtgNM_021619.2PRDM12Homo_refHeter
chr91.34E+08TANON_SYNONYMOUS_CODING(MODERATE)S1441TTcc/AccNM_006059.3LAMC3Homo_refHeter
chr91.37E+08TCNON_SYNONYMOUS_CODING(MODERATE)Y189HTac/CacNM_002957.4RXRAHomo_refHeter
chr91.39E+08CTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)C58Y;C63YtGc/tAc;tGc/tAcNM_178138.4;NM_014564.3LHX3;LHX3Homo_refHeter
chr1017199504GASTOP_GAINED(HIGH)R275*Cga/TgaNM_004412.5TRDMT1Homo_refHeter
chr1026455049CTNON_SYNONYMOUS_CODING(MODERATE)T1018IaCc/aTcNM_017433.4MYO3AHomo_refHeter
chr1032304529ACNON_SYNONYMOUS_CODING(MODERATE)I940MatT/atGNM_004521.2KIF5BHomo_refHeter
chr1050819232TCNON_SYNONYMOUS_CODING(MODERATE)M149TaTg/aCgNM_003055.2SLC18A3Homo_refHeter
chr1098129916ATNON_SYNONYMOUS_CODING(MODERATE)V940DgTt/gAtNM_012465.3TLL2Homo_refHeter
chr101.23E+08TCNON_SYNONYMOUS_CODING(MODERATE)V597AgTt/gCtNM_018117.11WDR11Homo_refHeter
chr101.35E+08TCNON_SYNONYMOUS_CODING(MODERATE)K508RaAg/aGgNM_001200049.2TTC40Homo_refHeter
chr11612778AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)V431A;V460A;V473AgTg/gCg;gTg/gCg;gTg/gCgNM_004029.2;NM_001572.3;NM_004031.2IRF7;IRF7;IRF7Homo_refHeter
chr11773623CTNON_SYNONYMOUS_CODING(MODERATE)R85QcGg/cAgNM_182612.2PDDC1Homo_refHeter
chr111078307GTNON_SYNONYMOUS_CODING(MODERATE)K198NaaG/aaTNM_002457.2MUC2Homo_refHeter
chr112427314AGNON_SYNONYMOUS_CODING(MODERATE)F1087LTtc/CtcNM_014555.3TRPM5Homo_refHeter
chr1166619985TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)T584A;T584A;T584AAcc/Gcc;Acc/Gcc;Acc/GccNM_000920.3;NM_001040716.1;NM_022172.2PC;PC;PCHomo_refHeter
chr1177378107GCNON_SYNONYMOUS_CODING(MODERATE)P1394RcCc/cGcNM_016578.3RSF1Homo_refHeter
chr111.19E+08CANON_SYNONYMOUS_CODING(MODERATE)S662YtCt/tAtNM_021729.4VPS11Homo_refHeter
chr111.19E+08ATSTOP_GAINED(HIGH);STOP_GAINED(HIGH)K86*;K86*Aag/Tag;Aag/TagNM_001142505.1;NM_022169.4ABCG4;ABCG4Homo_refHeter
chr126933276CANON_SYNONYMOUS_CODING(MODERATE)A71DgCc/gAcNM_019858.1GPR162Homo_refHeter
chr129260161CTNON_SYNONYMOUS_CODING(MODERATE)G280SGgt/AgtNM_000014.4A2MHomo_refHeter
chr1253453359GTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)R521L;R645L;R655LcGa/cTa;cGa/cTa;cGa/cTaNM_198316.1;NM_170754.2;NM_015319.2TENC1;TENC1;TENC1Homo_refHeter
chr1258197052ATNON_SYNONYMOUS_CODING(MODERATE)M647KaTg/aAgNM_006576.3AVILHomo_refHeter
chr1269126435GTNON_SYNONYMOUS_CODING(MODERATE)D673YGat/TatNM_020401.2NUP107Homo_refHeter
chr1277235849CTNON_SYNONYMOUS_CODING(MODERATE)P410LcCa/cTaNM_015336.2ZDHHC17Homo_refHeter
chr1339425170CTSTOP_GAINED(HIGH)Q2223*Caa/TaaNM_207361.4FREM2Homo_refHeter
chr1473957946TGNON_SYNONYMOUS_CODING(MODERATE)L75RcTg/cGgNM_024644.3C14orf169Homo_refHeter
chr141.04E+08TCNON_SYNONYMOUS_CODING(MODERATE)S218GAgt/GgtNM_015316.2PPP1R13BHomo_refHeter
chr1566601046GTSPLICE_SITE_ACCEPTOR(HIGH);SPLICE_SITE_ACCEPTOR(HIGH);;NM_001143688.1;NM_133375.3DIS3L;DIS3LHomo_refHeter
chr1590801386GANON_SYNONYMOUS_CODING(MODERATE)R351HcGc/cAcNM_001029964.2TTLL13Homo_refHeter
chr16613451TCNON_SYNONYMOUS_CODING(MODERATE)Y53HTac/CacNM_145270.2C16orf11Homo_refHeter
chr1631230697CTNON_SYNONYMOUS_CODING(MODERATE)R192CCgc/TgcNM_001008274.3TRIM72Homo_refHeter
chr174167225AGSPLICE_SITE_DONOR(HIGH);SPLICE_SITE_DONOR(HIGH);;NM_016376.3;NR_047571.1ANKFY1;ANKFY1Homo_refHeter
chr178045692GANON_SYNONYMOUS_CODING(MODERATE)A1115VgCt/gTtNM_002616.2PER1Homo_refHeter
chr1737565110AGNON_SYNONYMOUS_CODING(MODERATE)S1122PTca/CcaNM_004774.3MED1Homo_refHeter
chr1762856060AGNON_SYNONYMOUS_CODING(MODERATE)F1402LTtt/CttNM_199340.2LRRC37A3Homo_refHeter
chr1773499968AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)Y433H;Y515HTat/Cat;Tat/CatNM_001142643.1;NM_020753.3CASKIN2;CASKIN2Homo_refHeter
chr1776449531CTNON_SYNONYMOUS_CODING(MODERATE)E3475KGag/AagNM_173628.3DNAH17Homo_refHeter
chr1779555984CANON_SYNONYMOUS_CODING(MODERATE)D423YGat/TatNM_017921.2NPLOC4Homo_refHeter
chr183272991AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)Q32R;Q32R;Q32RcAg/cGg;cAg/cGg;cAg/cGgNM_001144944.1;NM_001144945.1;NM_033546.3MYL12B;MYL12B;MYL12BHomo_refHeter
chr1829867458CTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)V368M;V368MGtg/Atg;Gtg/AtgNM_022751.2;NM_001242409.1GAREM;GAREMHomo_refHeter
chr191295566GANON_SYNONYMOUS_CODING(MODERATE)D55NGac/AacNM_001405.3EFNA2Homo_refHeter
chr199059365CASTOP_GAINED(HIGH)E9361*Gaa/TaaNM_024690.2MUC16Homo_refHeter
chr1916982170CTTCTTTFRAME_SHIFT(HIGH)-884?-/TNM_015260.2SIN3BHomo_refHeter
chr1919414563AGNON_SYNONYMOUS_CODING(MODERATE)M211TaTg/aCgNM_172231.3SUGP1Homo_refHeter
chr1948305970GANON_SYNONYMOUS_CODING(MODERATE)P100SCct/TctNM_198479.2TPRX1Homo_refHeter
chr1948922916CTNON_SYNONYMOUS_CODING(MODERATE)P646SCcc/TccNM_000836.2GRIN2DHomo_refHeter
chr1949377522TGSTOP_GAINED(HIGH)Y344*taT/taGNM_014330.3PPP1R15AHomo_refHeter
chr1954759960TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)K201E;K201EAaa/Gaa;Aaa/GaaNM_006840.3;NM_001081442.1LILRB5;LILRB5Homo_refHeter
chr1956185411TANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)Y465N;Y469NTat/Aat;Tat/AatNM_001012478.1;NM_007279.2U2AF2;U2AF2Homo_refHeter
chr1957325702CTNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)A1244T;A1246T;A1370T;A1370T;A1370TGca/Aca;Gca/Aca;Gca/Aca;Gca/Aca;Gca/AcaNM_001146185.1;NM_001146187.1;NM_001146184.1;NM_001146186.1;NM_006210.2PEG3;PEG3;PEG3;PEG3;PEG3Homo_refHeter
chr1957640078TANON_SYNONYMOUS_CODING(MODERATE)I12NaTt/aAtNM_020903.2USP29Homo_refHeter
chr2017928162TANON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)K349I;K349IaAa/aTa;aAa/aTaNM_014426.2;NM_152227.1SNX5;SNX5Homo_refHeter
chr2031656753CTNON_SYNONYMOUS_CODING(MODERATE)P375SCct/TctNM_182658.1BPIFB3Homo_refHeter
chr2043052973TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)L381P;L403PcTc/cCc;cTc/cCcNM_001030004.2;NM_178850.2HNF4A;HNF4AHomo_refHeter
chr2048808388ACNON_SYNONYMOUS_CODING(MODERATE)E273AgAg/gCgNM_005194.3CEBPBHomo_refHeter
chr2050342443ACNON_SYNONYMOUS_CODING(MODERATE)F81CtTc/tGcNM_006045.1ATP9AHomo_refHeter
chr2054970702AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)N32D;N32D;N32DAat/Gat;Aat/Gat;Aat/GatNM_001033521.1;NM_001033522.1;NM_001324.2CSTF1;CSTF1;CSTF1Homo_refHeter
chr2141559073AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)C922R;C922RTgc/Cgc;Tgc/CgcNM_001271534.1;NM_001389.3DSCAM;DSCAMHomo_refHeter
chr2143867183TCNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)C512R;M584T;M622TTgt/Cgt;aTg/aCg;aTg/aCgNM_001243467.1;NM_001001895.2;NM_018961.3UBASH3A;UBASH3A;UBASH3AHomo_refHeter
chr2145994548CTNON_SYNONYMOUS_CODING(MODERATE)P305SCca/TcaNM_198687.1KRTAP10-4Homo_refHeter
chr2146086757GANON_SYNONYMOUS_CODING(MODERATE)A16VgCg/gTgNM_181684.2KRTAP12-2Homo_refHeter
chr2146086758CTNON_SYNONYMOUS_CODING(MODERATE)A16TGcg/AcgNM_181684.2KRTAP12-2Homo_refHeter
chr2223596000AGNON_SYNONYMOUS_CODING(MODERATE);NON_SYNONYMOUS_CODING(MODERATE)T432A;T432AAca/Gca;Aca/GcaNM_021574.2;NM_004327.3BCR;BCRHomo_refHeter
chr315613277CTTAACSPLICE_SITE_ACCEPTOR(HIGH)NM_012260.2HACL1HeterHomo_ALT
Extended Data Table 2

Relative expression of ERBB1-4, seven EGFR ligands, and MET in CC and CC-CR

Gene NameFPKM1 CC 1FPKM CC 2FPKM CC 3FPKM CC-CR 1FPKM CC-CR 2FPKM CC-CR 3Log2FC2 (CC-CR vs CC)FDR
EGFR7.1137927.2163156.909610.535678.8389119.7404460.4574686.32E-06
ERBB224.2338324.5219223.9776223.0725225.5493821.22859-0.056560.730888
ERBB318.7960718.9125217.9086919.0704120.0056814.89755-0.041980.859801
ERBB400.00510700.01606900.004937NANA
AREG21.571621.7664620.7611613.4858513.7375417.10571-0.52760.000618
BTC6.4134085.3076376.7237835.59145.3758648.2320570.0625810.881371
EGF0.5688420.7697710.5746850.4889510.3227590.366821-0.696270.012365
EPGN0.1397330.1098590.18311900.2631410.176997NANA
EREG31.8230431.8603537.4071221.3140723.2711133.68543-0.365290.117939
HBEGF3.1705482.7073052.5755393.4165463.660363.2979670.2956920.175018
TGFA13.6748312.7957311.8066319.2037716.2955315.215870.4073070.000454
MET135.2686157.1884149.2257213.2824228.8418221.76580.5903343.12E-09

FPKM stands for Fragments Per Kilobase of transcript per Million mapped reads.

Differential expression analysis was performed on RNA-Seq counts using edgeR.

Figure 2

Transcriptome profiling of CC and CC-CR in 3D

(a) Left, heatmap of top 50 differentially expressed transcripts in CC-CR versus CC from 3 independent 3D culture experiments. Gene expression values are gene-wise z-transformed and are colored red for high abundance and blue for low abundance, as indicated in the scale bar. Right, miRNA heatmap showing miRNAs altered (>2-fold and FDR<0.01) in CC-CR versus CC. (b) Genomic organization of lncRNA MIR100HG, host gene of miR-100/let-7a-2/miR-125b-1 cluster, on human chromosome 11 (hsa chr11). (c) qRT-PCR showing upregulation of lncRNA MIR100HG, miR-100 and miR-125b in CC-CR compared to CC in 3D. In CC-CR, cells were treated with CTX (CTX+, 3 μg/ml) or normal culture medium (CTX-) for 14 consecutive days in 3D. ACTB or U6 snRNA served as the internal control, respectively. n=3 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test. (d) Scatter plots of MIR100HG versus miR-100 or miR-125b expression in TCGA CRC data repository. Pearson correlation coefficients (r) and P values are shown. (e) RNA FISH showing high MIR100HG (red) expression in CC-CR mouse tumor xenografts compared to CC xenografts. Concomitantly, high miR-100 (green) and miR-125b (red) signals were observed in CC-CR tumors; the yellow fluorescent signal indicates co-expression of miR-100 and miR-125b. Scale bars, 50 μm. (f) qRT-PCR analysis of MIR100HG, miR-100 and miR-125b expression levels among a panel of 30 CRC cell lines ranked by their responsiveness to cetuximab (see Extended Data Table 3). ACTB or U6 snRNA served as the internal control. Fold changes were normalized to CC. n.d., not detected. n=3 independent experiments performed in triplicate. Data represent mean ± s.d. in c and f.

MIR100HG is the host gene of the miR-100/let-7a-2/miR-125b-1 cluster on chromosome 11 (Fig. 2b). qRT-PCR analysis confirmed upregulation of endogenous MIR100HG expression in CC-CR in the presence or absence of cetuximab (Fig. 2c). pri-miR-100, pri-miR-125b-1, and their corresponding mature miRNA, miR-100 and miR-125b, were also enriched in CC-CR (Fig. 2c and Extended Data Fig. 2a). Although pri-let-7a-2 was upregulated in CC-CR, mature let-7a was unchanged compared to CC (Extended Data Fig. 2b). The transcriptional start site (TSS) of MIR100HG was confirmed by 5′ RACE-PCR (Extended Data Fig. 2c). Analysis of The Cancer Genome Atlas (TCGA) CRC data repository revealed that miR-100 and miR-125b expression is tightly correlated with MIR100HG expression (Fig. 2d). RNA fluorescence in situ hybridization (FISH) showed highly enriched MIR100HG and miR-100/125b expression in CC-CR tumor xenografts (Fig. 2e). In contrast, let-7a expression did not correlate with that of MIR100HG (Extended Data Fig. 2d).
Extended Data Fig. 2

MIR100HG and miR-100/125b overexpression in cetuximab-resistant colorectal cancer cell lines.

(a) qRT-PCR showing upregulation of pri-miR-100 and pri-miR-125b-1 in CC-CR compared to CC grown in 3D. In CC-CR, cells were treated with CTX (CTX+, 3 μg/ml) or normal culture medium (CTX-) for consecutive 14 days in 3D. (b) qRT-PCR showing upregulation of pri-let-7a-2 expression in CC-CR but unchanged expression of mature let-7a between the 2 cell lines. n=3 independent experiments performed in triplicate in a and b. Data represent mean ± s.d. **P<0.01 by one-way ANOVA followed by Dunnett's test compared with CC. (c) Left: a schematic diagram showing the PCR primers used in the 5′ RACE. Right: MIR100HG TSS was validated by 5′ RACE nested PCR in CC-CR with subsequent sequencing of the cloned fragments. Arrow indicates band of expected size. M, DNA marker. (d) Scatter plots of MIR100HG versus let-7a expression in TCGA CRC data repository. No correlation was found between those 2 molecules. (e, f) Expression of MIR100HG and miR-100/125b negatively correlates with cetuximab growth inhibition regardless of KRAS/BRAF mutational status. (e) Scatter plot of MIR100HG and miR-100/125b expression versus cetuximab inhibition rate in a panel of 30 CRC cell lines. (f) Twenty-one cell lines harbor KRAS or BRAF mutation, and 9 cell lines are KRAS/BRAF wild-type (WT). Pearson correlation coefficients (r) and P values are shown.

To assess whether MIR100HG and miR-100/125b overexpression extended beyond this one cell line, we examined their expression in a panel of 30 CRC cell lines placed upon a continuum of cetuximab sensitivity and resistance based upon published reports[16,17] (Extended Data Table 3). Expression of MIR100HG and miR-100/125b were enriched in more cetuximab-resistant lines compared to the more sensitive lines (Fig. 2f). Their expression inversely correlated with cetuximab resistance, regardless of KRAS/BRAF mutational status (Extended Data Fig. 2e and f). For example, two of the cetuximab-sensitive lines (GEO and SW403) expressed low levels of MIR100HG and miR-100/125b despite harboring mutant KRAS. In addition, we also observed upregulation of MIR100HG and miR-100/125b in the setting of cetuximab resistance in HNSCC cell lines (Extended Data Fig. 3a). Thus, MIR100HG and miR-100/125b are upregulated in the setting of cetuximab resistance in CRC and HNSCC cell lines and this phenomenon occurs in both acquired and de novo resistance. These findings led us to further explore the function of MIR100HG and miR-100/125b in cetuximab resistance.
Extended Data Table 3

Mutational status of 30 CRC cell lines used in Fig. 2f and their response to cetuximab

Cell lineMutational statusMean CTX inhibition rate (%)CTX response category
NCI-H508BRAF83.41sensitive
V9PWT82.03sensitive
DiFiWT80.81sensitive
LIM1215WT79.62sensitive
GEOKRAS68.72sensitive
SW403KRAS66.02sensitive
SNUC4WT48.31partially responsive
Caco-2WT47.72partially responsive
SW948KRAS42.72partially responsive
HT29BRAF36.52partially responsive
SK-CO-1KRAS33.92partially responsive
DLD-1KRAS24.92resistant
SW480KRAS23.72resistant
SW837KRAS21.82resistant
SW48WT21.82resistant
SW620KRAS14.52resistant
LoVoKRAS14.31resistant
COLO205BRAF12.51resistant
T84KRAS11.22resistant
LS174TKRAS9.71resistant
NCI-H716WT9.71resistant
HCT8KRAS8.41resistant
HCT15KRAS4.52resistant
SW1116KRAS2.01resistant
LIM2405KRAS2.02resistant
RKOBRAF0.41resistant
COLO320DMWT-3.21resistant
HuTu80WT-5.41resistant
LS123KRAS-4.81resistant
HCT116KRAS-14.12resistant

Data came from Medico, E. et al, Nat Commun, 2015[17].

Data came from Jhawer, M. et al, Cancer Res, 2008[16].

Experimental data from the present study.

Extended Data Fig. 3

MIR100HG and miR-100/125b expression in head and neck squamous cell cancer cell lines and modulation of miR-100 and/or miR-125b in CC and CC-CR cells.

(a) qRT-PCR analysis of MIR100HG, miR-100, and miR-125b expression among the CTX-sensitive head and neck squamous cell carcinoma (HNSCC) cell line SCC25 and its derived CTX-resistant sublines (CTX-R1, R3, R4, R5, R7, and R8) upon continuous exposure to cetuximab, as well as UNC10, a de novo CTX-resistant cell line. n=3 independent experiments performed in triplicate. *P<0.05, **P<0.01 by one-way ANOVA followed by Dunnett's test compared with SCC25. (b) qRT-PCR of indicated miRNA expression in CC stably overexpressing miR-100, miR-125b, or Bicistron. (c) qRT-PCR of indicated miRNA expression in CC-CR stably expressing miR-100 sponge (100-Sp), miR-125b sponge (125b-Sp), or bicistron sponge (Bicistron-Sp). Values were normalized to U6 snRNA. n=3 experiments performed in triplicate. **P<0.01 by Student's t test. (d, e) Quantification of Ki-67 and Cleaved Casp-3 in Fig. 3c and d. n=4 independent experiments. *P<0.05, **P<0.01 by Student's t test. Data represent mean ± s.d. n.s., not significant.

miR-100 and miR-125b cooperativity drives cetuximab resistance

Since a major role of certain lncRNAs is production of embedded miRNAs[10,18], we asked whether cetuximab resistance is mediated by miR-100 and miR-125b overexpression. To this end, we delivered lentiviral-based overexpression or sponge constructs into CC and CC-CR, respectively, to generate stable cell lines expressing each miRNA, the miR-100/125b bicistron, or their corresponding sponges (Extended Data Fig. 3b and c). Although the miR-100 sponge had no significant effect on colony number in CC-CR in 3D culture, both the miR-125b and bicistron sponges significantly reduced colony number (Fig. 3a). In the presence of cetuximab, the miR-100 sponge modestly reduced colony number, whereas the reduction in colony number was more pronounced with the miR-125b sponge and the bicistron sponge (Fig. 3a). In contrast, opposite effects were observed in CC upon overexpressing miR-100 and miR-125b individually and together. The miR-100/125b bicistron, but not individual miRNAs, increased colony number in CC (Fig. 3b). Upon cetuximab treatment, the miR-100/125b bicistron conferred the strongest pro-survival effect; when introduced individually, miR-125b had a greater effect than miR-100 (Fig. 3b). Similar opposing effects were observed in morphological changes, as well as Ki-67 and cleaved Caspase-3 staining upon expressing the different sponges in CC-CR and the different miRNAs in CC (Fig. 3c and d; Extended Data Fig. 3d and e). Additionally, overexpression of the miR-100/125b bicistron in Caco-2 cells (low endogenous miR-100/125b expression) rendered cells less responsive to cetuximab, whereas inhibition of the miR-100/125b bicistron in DLD-1 cells and HNSCC SCC25-derived CTX-R7 cells (both with high endogenous miR-100/125b expression) restored cetuximab responsiveness (Extended Data Fig. 4). Similar results were observed when CC-CR and CC cells with the differing manipulations were established as subcutaneous xenografts in nude mice and treated with cetuximab (Fig. 3e and f; Extended Data Fig. 5). Together, these results are consistent with a model in which miR-100 and miR-125b cooperate to confer cetuximab resistance.
Figure 3

Cooperativity of miR-100 and miR-125b in CTX resistance

(a, b) Indicated cells were grown in 3D in normal medium (CTL) or treated with CTX (3 μg/ml) in 3D. The resultant colonies were counted after 18 days. Sponge (Sp). n=3 independent experiments performed in triplicate. *P<0.05, **P<0.01 by one-way ANOVA followed by Dunnett's test compared with CTL-Sp or miR-CTL. (c, d) Left: indicated cells were cultured in 3D for 12 days and CTX (10 μg/ml) was added for 24 h before cells were fixed, stained for Cleaved Casp-3 (cyan) and Ki-67 (magenta). Scale bars, 50 μm. Right: quantification of the morphological changes among indicated cell lines. n=4 independent experiments. (e, f) Left: indicated cells were injected subcutaneously into nude mice (n=8). After tumor size reached approximately 100 mm3, the mice received CTX treatment (1 mg/mouse, i.p. injection every 3 days). Representative fluorescent images of GFP signals captured from subcutaneous tumors are shown. Middle: growth curve of tumors in nude mice (n=8) injected with cells as indicated. **P<0.01 by repeated-measures ANOVA test followed by Dunnett's test. Right: tumors (n=8) were isolated on day 28 after treatment and tumor weight was calculated. **P<0.01 by one-way ANOVA followed by Dunnett's test compared with CTL-Sp or miR-CTL. Data represent mean ± s.d. n.s., not significant.

Extended Data Fig. 4

miR-100 and miR-125b cooperativity drives cetuximab resistance in colorectal cancer and head and neck squamous cell cancer cell lines.

(a) Caco-2 cells stably overexpressing Bicistron or control (miR-CTL) were cultured in 3D for 5 days and treated with CTX (50 μg/ml) for 24 h. Immunofluorescence was performed for Cleaved Casp-3 (cyan) and Ki-67 (magenta) with quantification shown on the right. Scale bar, 50 μm. n=3 independent experiments. *P<0.05, **P<0.01 by Student's t test. (b) DLD-1 cells stably expressing Bicistron-Sp or control (CTL-Sp) were cultured in 3D for 10 days and treated with CTX (200 μg/ml) for 24 h. Staining of Cleaved Casp-3 (cyan) and Ki-67 (magenta) were shown. Scale bars, 50 μm. Quantification is shown on the right. n=3 independent experiments. *P<0.05, **P<0.01 by Student's t test. (c, d) Indicated cells were grown in 3D in normal medium (CTL) or treated with CTX (50 μg/ml for Caco-2, and 200 μg/ml for DLD-1) in 3D. The resultant colonies were counted. n=2 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test compared with miR-CTL or CTL-Sp. (e) Left: CTX-R7 cells stably expressing miR-100 and/or miR-125b sponges were grown in normal medium (CTL) or treated with CTX (30 μg/ml). Cell viability was measured by cell counting kit-8 (CCK-8) assays after 72 h. n=3 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test compared with CTL-Sp. Middle: qRT-PCR analysis of Wnt target genes in the stable bicistron sponge-transduced CTX-R7 cells. n=2 independent experiments performed in triplicate. *P<0.05, **P<0.01 by Student's t test. Right: CTX-R7 cells were treated with CTX (30 μg/ml) and/or ICG-001 (2 μM) for 72 h, and cell viability was measured by CCK-8 assays. n=2 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by by LSD post-hoc test. Data represent mean ± s.d. n.s., not significant.

Extended Data Fig. 5

Effects of differential modulation of miR-100 and/or miR-125b on cetuximab responsiveness in CC and CC-CR in vivo.

(a, b) Quantification of radiant efficiency from tumors (n=8) represented on Fig. 3e and f. **P<0.01 by paired Student's t test. (c-f) Representative IHC images and quantification of Ki-67 and Cleaved Casp-3 from indicated xenografts (n=8) treated with CTX. Scale bars, 50 μm. **P<0.01 by one-way ANOVA followed by Dunnett's test in e and f. Data represent mean ± s.d. in a, b, e, and f. n.s., not significant.

miR-100 and miR-125b repress multiple Wnt negative regulators and increase Wnt signaling

To understand how miR-100 and miR-125b influence cetuximab responsiveness, we considered the most downregulated genes in CC-CR to be potential targets of these miRNAs. Two negative regulators of Wnt signaling, DKK1 and DKK3, were decreased over 30-fold in CC-CR compared to CC (Fig. 2a). Meanwhile, Wnt activity was enhanced in CC-CR by analysis of 64 consensus β-catenin target genes[19] (Extended Data Fig. 6a). Analysis of a large human CRC dataset (n=458) also revealed that MIR100HG expression levels positively correlated with the Wnt score[19], whereas no correlation was observed between MIR100HG and the Ras-Az score[20], which measures MEK activation as a downstream index of RAS signaling (Extended Data Fig. 6a and data not shown). Functional enrichment analysis further identified Wnt pathway enrichment in miR-100 and miR-125b putative targets (Extended Data Table 4). We thus considered whether miR-100 and miR-125b might target components of Wnt signaling. Through computational target prediction, we found that 3′ UTRs of DKK1 and DKK3 contain binding sites for miR-100 and miR-125b, respectively (Fig. 4a). Since clustered miRNAs are co-expressed and often coordinately regulate molecular pathways by targeting different components of the same pathway[21], we searched for other negative regulators of Wnt signaling that contain putative binding sites for miR-100 or miR-125b, and identified zinc and ring finger 3 (ZNRF3), ring finger protein 43 (RNF43), and APC2 as potential targets of miR-100 or miR-125b alone or in combination (Fig. 4a and Extended Data Table 5). Decreased protein levels of these five Wnt negative regulators in CC-CR compared to CC were confirmed by both immunoblots in cell lines and immunostaining in xenografts (Extended Data Fig. 6b and c). Using 3′ UTR luciferase reporter assays, we confirmed these five candidates are direct targets of miR-100 and/or miR-125b in both CC and Caco-2 cells; repression of these genes was rescued by mutations in the corresponding binding sites (Fig. 4b and c; Extended Data Fig. 6d). Immunoblots confirmed the regulation of the predicted targets by miR-100 and miR-125b alone or in combination in CC and CC-CR (Fig. 4d). Consistently, this regulation was also observed in Caco-2 cells (low endogenous miR-100/125b expression) and HuTu80 cells (high endogenous miR-100/125b expression) (Extended Data Fig. 6e).
Extended Data Fig. 6

miR-100/125b coordinately represses five Wnt/β-catenin negative regulators, resulting in increased Wnt signaling.

(a) Left: Wnt activation in CC and CC-CR cells was measured by the 64 Wnt/β -catenin target genes (Wnt signature score). **P<0.01 by Student's t test. Right: Scatter plots of MIR100HG expression versus 64-gene Wnt signature score on 458 CRC. Pearson correlation coefficients (r) and P values are shown. (b) Immunoblots of DKK1, DKK3, ZNRF3, RNF43, and APC2 levels from 3D cell lysates of CC and CC-CR. In CC-CR, cells were treated with CTX (CTX+, 3 μg/ml) or normal culture medium (CTX-) for 14 days in 3D before protein extraction. Representative of 3 independent experiments. (c) Top : representative IHC images of DKK1, DKK3, ZNRF3, RNF43, and APC2 in CC and CC-CR xenografts (n=8). Bottom: measurement of protein expression, **P<0.01 by Mann-Whitney U test. (d) Dual luciferase assays of genes predicted to be regulated by miR-100 or miR-125b in Caco-2. Renilla luciferase activity was normalized to firefly activity. n=2 independent experiments. **P<0.01 by Student's t test. (e) Immunoblots of indicated proteins in stable miRNA-transduced Caco-2 and sponge (Sp)-transduced HuTu80. Representative of 2 independent experiments. (f) Immunoblot of nuclear and cytoplasmic extracts for β-catenin and p-β-catenin (S552). Loading controls were GAPDH for cytoplasmic fractions and Lamin A/C for nuclear fractions. (g) CC and CC-CR in 3D were treated with CTX (10 μg/ml) and/or Wnt3a (100 ng/ml). Immunoblots of indicated proteins after 48 h of treatment are shown. Representative of 3 independent experiments. (h) qRT-PCR analysis of Wnt targets CCND1, CD44, FOSL1, and NKD1 mRNAs at indicated time points following CTX (10 μg/ml) treatment in 3D. n=2 independent experiments performed in triplicate. **P<0.01 by two-way ANOVA test. Data represent mean ± s.d. in a, c, d, and h.

Extended Data Table 4
miRNAPathNamePathFgPathBgGenomeFGGenomeBGpvalBH
hsa-miR-100-5pPathways in cancer1973308828197472.49E-084.84E-06
hsa-miR-100-5pPancreatic cancer52758828197471.39E-050.00264553
hsa-miR-100-5pWnt signaling pathway941528828197471.49E-050.00285334
hsa-miR-100-5pmTOR signaling pathway39538828197471.87E-050.00354634
hsa-miR-100-5pUbiquitin mediated proteolysis831348828197474.24E-050.00797254
hsa-miR-100-5pGlioma45658828197475.51E-050.01025001
hsa-miR-100-5pEndocytosis1101878828197476.79E-050.0124941
hsa-miR-100-5pChronic myeloid leukemia50758828197470.000100820.01834838
hsa-miR-100-5pCalcium signaling pathway1041788828197470.000152220.02739988
hsa-miR-100-5pFocal adhesion1162038828197470.000232680.04141736
hsa-miR-100-5pLeukocyte transendothelial migration711168828197470.000248120.04416472
hsa-miR-100-5pInsulin signaling pathway831398828197470.000255070.04514807
hsa-miR-100-5pNon small cell lung cancer37548828197470.000346890.06035915
hsa-miR-100-5pApoptosis55878828197470.000380210.06577642
hsa-miR-100-5pMAPK signaling pathway1492728828197470.00049660.0849182
hsa-miR-100-5pRenal cell carcinoma46718828197470.000506510.08661247
hsa-miR-100-5pLong term potentiation46718828197470.000506510.08661247
hsa-miR-100-5pAxon guidance761298828197470.000793890.13257995
hsa-miR-100-5pSmall cell lung cancer52848828197470.001105230.17904734
hsa-miR-100-5pColorectal cancer53868828197470.001154010.18694947
hsa-miR-100-5pAldosterone regulated sodium reabsorption29428828197470.001240640.19974264
hsa-miR-100-5pPhosphatidylinositol signaling system47768828197470.001935950.30200877
hsa-miR-100-5pChondroitin sulfate biosynthesis17228828197470.001954060.30483395
hsa-miR-100-5pVEGF signaling pathway48788828197470.002019370.31502171
hsa-miR-100-5pHedgehog signaling pathway36568828197470.00245470.37556932
hsa-miR-100-5pCell adhesion molecules CAMs761338828197470.00256930.39310273
hsa-miR-100-5pVascular smooth muscle contraction671168828197470.00313070.47273626
hsa-miR-100-5pGlycosphingolipid biosynthesis lacto and neolacto series19268828197470.003203560.48373791
hsa-miR-100-5pAmyotrophic lateral sclerosis ALS35558828197470.003598940.53624172
hsa-miR-100-5pFc gamma R mediated phagocytosis57978828197470.003668450.54659853
hsa-miR-100-5pLysosome691218828197470.00422010.61613498
hsa-miR-100-5pDilated cardiomyopathy55948828197470.004843920.6975252
hsa-miR-100-5pMelanoma43718828197470.005141810.73527865
hsa-miR-100-5pHeparan sulfate biosynthesis18268828197470.010125351
hsa-miR-100-5pLysine degradation28458828197470.013505591
hsa-miR-100-5pMelanogenesis571028828197470.015002761
hsa-miR-100-5pGnRH signaling pathway561058828197470.046413941
hsa-miR-100-5pArrhythmogenic right ventricular cardiomyopathy ARVC41748828197470.041520651
hsa-miR-100-5pKeratan sulfate biosynthesis11158828197470.024075781
hsa-miR-100-5pRegulation of actin cytoskeleton1102128828197470.020680521
hsa-miR-100-5pType II diabetes mellitus29498828197470.029225261
hsa-miR-100-5pNeurotrophin signaling pathway701298828197470.01804861
hsa-miR-100-5pO Glycan biosynthesis20308828197470.012665371
hsa-miR-100-5pValine leucine and isoleucine degradation28458828197470.013505591
hsa-miR-100-5pAdherens junction45768828197470.00763721
hsa-miR-100-5pABC transporters26448828197470.038754651
hsa-miR-100-5pBasal cell carcinoma34558828197470.007869331
hsa-miR-100-5pProstate cancer48898828197470.050071061
hsa-miR-100-5pTGF beta signaling pathway47868828197470.040396751
hsa-miR-100-5pDorso ventral axis formation16248828197470.025090591
hsa-miR-100-5pNicotinate and nicotinamide metabolism17248828197470.00876571
hsa-miR-100-5pAdipocytokine signaling pathway41708828197470.013523831
hsa-miR-100-5pFc epsilon RI signaling pathway46828828197470.02485631
hsa-miR-100-5pButanoate metabolism21358828197470.049700631
hsa-miR-100-5pNotch signaling pathway28478828197470.028644521
hsa-miR-100-5pT cell receptor signaling pathway581108828197470.055086981
hsa-miR-100-5pp53 signaling pathway40688828197470.013302841
hsa-miR-100-5pErbB signaling pathway51898828197470.01123981
hsa-miR-100-5pHypertrophic cardiomyopathy HCM49868828197470.014674351
miRNAPathNamePathFgPathBgGenomeFGGenomeBGpvalBH
hsa-miR-125b-5pMAPK signaling pathway21727212145197475.52E-111.07E-08
hsa-miR-125b-5pAxon guidance10812912145197473.01E-085.77E-06
hsa-miR-125b-5pPathways in cancer24933012145197474.35E-088.35E-06
hsa-miR-125b-5pRegulation of actin cytoskeleton16521212145197472.56E-074.84E-05
hsa-miR-125b-5pInsulin signaling pathway11313912145197473.55E-076.71E-05
hsa-miR-125b-5pGlioma586512145197475.88E-070.0001106
hsa-miR-125b-5pCell adhesion molecules CAMs10813312145197477.17E-070.00013411
hsa-miR-125b-5pWnt signaling pathway12115212145194741.1E-060.00022601
hsa-miR-125b-5pErbB signaling pathway748912145197477.66E-060.0014179
hsa-miR-125b-5pHedgehog signaling pathway495612145197471.64E-050.00298105
hsa-miR-125b-5pEndocytosis14218712145197471.87E-050.00339254
hsa-miR-125b-5pNeurotrophin signaling pathway10112912145197473.36E-050.00602008
hsa-miR-125b-5pChronic myeloid leukemia627512145197476.00E-050.01061902
hsa-miR-125b-5pPancreatic cancer627512145197476.00E-050.01061902
hsa-miR-125b-5pMelanoma597112145197476.69E-050.01177116
hsa-miR-125b-5pFocal adhesion15120312145197476.80E-050.01196783
hsa-miR-125b-5pNotch signaling pathway414712145197479.72E-050.0170169
hsa-miR-125b-5pNon small cell lung cancer465412145197470.00012960.02255083
hsa-miR-125b-5pLeukocyte transendothelial migration9011612145197470.000162250.02774454
hsa-miR-125b-5pLong term potentiation587112145197470.000199260.03367434
hsa-miR-125b-5pColorectal cancer688612145197470.000377180.06185754
hsa-miR-125b-5pCalcium signaling pathway13117812145197470.000435970.07106345
hsa-miR-125b-5pGlycosphingolipid biosynthesis lacto and neolacto series242612145197470.000466250.07596159
hsa-miR-125b-5pmTOR signaling pathway445312145197470.000599050.09644727
hsa-miR-125b-5pEndometrial cancer435212145197470.000816810.12905666
hsa-miR-125b-5pPhosphatidylinositol signaling system607612145197470.000887610.14024252
hsa-miR-125b-5pProstate cancer698912145197470.000954270.14982021
hsa-miR-125b-5pVEGF signaling pathway617812145197470.001254920.19576823
hsa-miR-125b-5pT cell receptor signaling pathway8311012145197470.00136530.21298668
hsa-miR-125b-5pVascular smooth muscle contraction8711612145197470.001460850.22643156
hsa-miR-125b-5pLysosome9012112145197470.001897720.28891677
hsa-miR-125b-5pLong term depression577312145197470.001923660.29239561
hsa-miR-125b-5pMelanogenesis7710212145197470.001962110.29824123
hsa-miR-125b-5pGnRH signaling pathway7910512145197470.002032790.30695151
hsa-miR-125b-5pAdherens junction597612145197470.002071140.31274149
hsa-miR-125b-5pBasal cell carcinoma445512145197470.002650510.38962493
hsa-miR-125b-5pRenal cell carcinoma557112145197470.003168480.45635002
hsa-miR-125b-5pApoptosis668712145197470.003262150.46975023
hsa-miR-125b-5pChemokine signaling pathway13418912145197470.004226690.59596343
hsa-miR-125b-5pFc epsilon RI signaling pathway628212145197470.004905390.68675477
hsa-miR-125b-5pType II diabetes mellitus394912145197470.005436610.75568815
hsa-miR-125b-5pFc gamma R mediated phagocytosis729712145197470.005653610.78585178
hsa-miR-125b-5pArrhythmogenic right ventricular cardiomyopathy ARVC567412145197470.007112680.98155051
hsa-miR-125b-5pEpithelial cell signaling in Helicobacter pylori infection527112145197470.02577271
hsa-miR-125b-5pChondroitin sulfate biosynthesis182212145197470.036349671
hsa-miR-125b-5pSNARE interactions in vesicular transport303912145197470.031571981
hsa-miR-125b-5pGlycerophospholipid metabolism517012145197470.031307651
hsa-miR-125b-5pO Glycan biosynthesis253012145197470.008822871
hsa-miR-125b-5pTGF beta signaling pathway628612145197470.026143411
hsa-miR-125b-5pp53 signaling pathway506812145197470.025531461
hsa-miR-125b-5pHypertrophic cardiomyopathy HCM628612145197470.026143411
hsa-miR-125b-5pGap junction649012145197470.036625251
hsa-miR-125b-5pUbiquitin mediated proteolysis9513412145197470.014482071
hsa-miR-125b-5pInositol phosphate metabolism415412145197470.018346021
hsa-miR-125b-5pKeratan sulfate biosynthesis131512145197470.035063661
hsa-miR-125b-5pAcute myeloid leukemia435812145197470.030119361
hsa-miR-125b-5pB cell receptor signaling pathway547512145197470.03776691
hsa-miR-125b-5pTight junction9413212145197470.012414541
hsa-miR-125b-5pBladder cancer334312145197470.025921051
hsa-miR-125b-5pVibrio cholerae infection405512145197470.055365721
hsa-miR-125b-5pSmall cell lung cancer618412145197470.021745431
hsa-miR-125b-5pAdipocytokine signaling pathway527012145197470.01694511
hsa-miR-125b-5pDilated cardiomyopathy669412145197470.049465031
Figure 4

miR-100 and miR-125b augment Wnt signaling by repressing multiple Wnt negative regulators

(a) Predicted miR-100 (red) and miR-125b (blue) binding sites in 3′ untranslated regions (3′ UTRs) of human DKK1, DKK3, ZNRF3, RNF43, and APC2. CDS, coding sequence. (b, c) Dual luciferase assays of candidates predicted to be regulated by miR-100 or miR-125b. Renilla luciferase activity was normalized to firefly activity and presented as relative luciferase activity. n=2 independent experiments. **P<0.01 by Student's t test. (d) Immunoblots of indicated proteins in stable miRNA-transduced CC and sponge (Sp)-transduced CC-CR. Representative of 3 independent experiments. (e) Immunofluorescence of p-β-catenin (Y489). Scale bars, 50 μm. Right, quantification of 4 independent experiments. **P<0.01 by Student's t test. (f) Representative IHC of β-catenin in CC and CC-CR xenografts (n=8). Scale bars: 50 μm (main); 20 μm (inset). Quantification of nuclear β-catenin-positive cells is shown. **P<0.01 by Student's t test. (g) qRT-PCR analysis of Wnt target genes in CC and CC-CR cells. n=3 independent experiments performed in triplicate. *P<0.05, **P<0.01 by Student's t test. (h) qRT-PCR analysis of Wnt target genes in the indicated stable miRNA-transduced CC cells. n=3 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test. (i) CC-CR were injected subcutaneously into nude mice (n=6). When tumor size reached around 100 mm3, mice were treated with control saline, or CTX (1 mg/mouse, i.p. injection every 3 days) and/or ICG-001 (150 mg/kg i.p. injection daily). Representative in vivo fluorescent images are shown. (j) Growth curve of tumors in nude mice (n=6) treated with different compounds. **P<0.01 by repeated-measures ANOVA test followed by LSD post-hoc test. (k) Tumors (n=6) were isolated on day 28 after treatment and tumor weight was measured. **P<0.01 by one-way ANOVA followed by LSD post-hoc test. Data represent mean ± s.d.

Extended Data Table 5

Schema of miR-100 and/or miR-125b binding sites in predicted target 3′ UTR sequences of human genes

GeneTarget Site13′ UTR Position
DKK1 88-98
DKK1 422-421
DKK3 641-647
ZNRF3 1164-1172
ZNRF3 475-481
ZNRF3 1529-1535
RNF43 1308-1314
RNF43 1462-1468
APC2 625-631
APC2 1612-1618
GATA6 851-857

Putative binding sites of miR-100 or miR-125b were mutated and highlighted in red.

We next examined whether miR-100/125b-induced downregulation of these Wnt negative regulators resulted in increased Wnt signaling. Although total β-catenin levels were not significantly altered, CC-CR exhibited increased active tyrosine phosphorylated p-Y489 β-catenin and increased nuclear β-catenin compared to CC in 3D (Fig. 4e and Extended Data Fig. 6f). Consistently, β-catenin was largely confined to the plasma membrane in CC xenografts, whereas it was largely nuclear in CC-CR xenografts (Fig. 4f). Moreover, mRNA expression of a panel of Wnt target genes was significantly enriched in CC-CR versus CC (Fig. 4g). Cetuximab blocked Wnt3a-induced Wnt activation in CC, but had no obvious effect on Wnt3a-induced Wnt signaling in CC-CR (Extended Data Fig. 6g). Cetuximab also led to a marked and persistent decrease in Wnt target genes in CC over 48 h, whereas expression of those genes in CC-CR was only modestly decreased at early time points after treatment before rebounding at later time points (Extended Data Fig. 6h). Furthermore, nuclear β-catenin levels increased in CC and Caco-2 cells stably overexpressing either miR-100 or miR-125b, and the increase was greater in cells expressing the miR-100/125b bicistron (Extended Data Fig. 7a). In contrast, nuclear β-catenin levels were reduced upon overexpressing the miR-100/125b bicistron sponge in CC-CR, DLD-1, and CTX-R7 cells (Extended Data Fig. 7a). Corresponding changes of Wnt target genes were also observed (Fig. 4h and Extended Data Fig. 4e). Consistent with these findings, nuclear β-catenin immunoreactivity increased in CC nude mouse xenografts expressing the miR-100/125b bicistron and decreased in their CC-CR counterparts expressing the bicistronic sponge (Extended Data Fig. 7b).
Extended Data Fig. 7

Effects of differential modulation of miR-100 and/or miR-125b on nuclear β-catenin expression levels.

(a) Immunoblots for β-catenin from nuclear fractions in the CC and Caco-2 cells overexpressing miR-100 and/or miR-125b, or CC-CR, DLD-1 and CTX-R7 cells expressing miR-100 and/or miR-125b sponges. Lamin A/C served as the control for nuclear fractions. Representative of 2 independent experiments. (b) Representative IHC of β-catenin in the indicated xenografts (n=8). Scale bars: 50 μm. Quantification of nuclear β-catenin positive cells is shown. Data represent mean ± s.d. **P<0.01 by Student's t test.

Based on our findings that Wnt signaling is increased in CC-CR, we hypothesized that cetuximab responsiveness may be restored by suppression of Wnt signaling. Since DKK1 and DKK3 are secreted Wnt antagonists and among the most downregulated genes in CC-CR, we tested whether their overexpression could overcome cetuximab resistance using a doxycycline-inducible lentiviral system[22]. Although induction of DKK1 or DKK3 resulted in a slight reduction in colony number, this effect was augmented with addition of cetuximab (Extended Data Fig. 8a and b). Moreover, administration of recombinant DKK1 and DKK3 enhanced the ability of cetuximab to decrease proliferation and increase apoptosis (Extended Data Fig. 8c). Furthermore, nuclear β-catenin expression decreased when DKK1 or DKK3 was inducibly expressed in CC-CR in the presence of cetuximab (Extended Data Fig. 8d). We next tested whether pharmacological inhibition of Wnt activity sensitized CC-CR to cetuximab using a tankyrase inhibitor, XAV-939[23], and a β-catenin/CBP inhibitor, ICG-001[24]. Both compounds caused a concentration-dependent reduction in colony number, and cetuximab growth inhibition was enhanced by their addition (Extended Data Fig. 8e). ICG-001 also enhanced the growth inhibitory effects of cetuximab in other CRC and HNSCC cell lines with high expression of MIR100HG (Extended Data Fig. 8f and 4e). In CC-CR nude mouse xenografts, administration of cetuximab and ICG-001 individually only slowed tumor growth; however, combined treatment resulted in tumor regression (Fig. 4i-k; Extended Data Fig. 8g and h). Thus, blockade of Wnt signaling, either upstream or downstream of the APC/β-catenin degradation complex, restores cetuximab responsiveness to cetuximab-resistant cells.
Extended Data Fig. 8

Blockade of Wnt signaling restores cetuximab responsiveness to cetuximab-resistant cells.

(a, b) Left: CC-CR doxycycline (Dox)-on DKK1 or DKK3 cells were cultured in the presence or absence of Dox (1 μg/ml) and harvested at 48 h. Total cell lysates and conditioned media were harvested and subjected to immunoblot analysis. Right: indicated cells were grown in 3D in normal medium or treated with CTX (3 μg/ml). The resultant colonies were counted after 18 days. n=3 experiments performed in triplicate. **P<0.01 by Student's t test. (c) CC-CR cells were grown in 3D in normal medium (CTL), treated with CTX (3 μg/ml) or in combination with recombinant DKK1 (rDKK1) and DKK3 (rDKK3) in 3D every 2 days. The resultant colonies were stained after 18 days for Cleaved Casp-3 (green) and Ki-67 (red). Scale bar, 50 μm. Quantification was shown. n=3 independent experiments. (d) Immunoblots for β-catenin from nuclear and cytoplasmic fractions of indicated cells upon CTX (10 μg/ml) treatment. Loading controls were GAPDH for cytoplasmic fractions and Lamin A/C for nuclear fractions. (e) CC-CR were treated with CTX (3 μg/ml), and/or XAV-939 (1, 5, 10 μM), and/or ICG-001 (1, 2.5, 5 μM) in 3D for 18 days, and colony number was determined. n=3 experiments performed in triplicate. (f) DLD-1 and HCT8 cells were treated with CTX (200 μg/ml) and/or ICG-001 (4 μM) for 14 days in 3D, and colony number was determined. n=2 independent experiments performed in triplicate. (g) Quantification of radiant efficiency from tumors (n=6) represented on Fig. 4i. **P<0.01 by paired Student's t test. (h) Representative IHC images and quantification of Ki-67 and Cleaved Casp-3 from CC-CR xenografts (n=6) treated with control saline (CTL), or CTX (1 mg/mouse, i.p. injection, every 3 days), and/or ICG-001 (150 mg/kg, i.p. injection, daily). Scale bar, 50 μm. *P<0.05, **P<0.01 by one-way ANOVA followed by Dunnett's test in c, e, and h, and one-way ANOVA followed by LSD post-hoc test in f. Data represent mean ± s.d. in a-c and e-h. n.s., not significant.

Reciprocal negative regulation between GATA6 and MIR100HG/miR-125b

To explore mechanism(s) by which miR-100 and miR-125b are upregulated in CC-CR, we investigated transcriptional regulation of the host gene, MIR100HG. Possible transcription factors containing binding sites within the 2.5 kb promoter of MIR100HG were mapped in silico using the Match program (version 1.0)[25] and cross-referenced with the RNA-Seq dataset (Extended Data Table 6). Among these transcription factors, we focused on the zinc-finger transcription factor GATA6, which was downregulated at both the mRNA and protein level in CC-CR in 3D culture and in nude mouse xenografts (Fig. 2a and Fig. 5a-c).
Extended Data Table 6

Predicted transcription factors for MIR100HG

FeatureName1logFC2logCPMLRp-ValueFDR
ENSG00000141448GATA6-5.43013.5893538.96450.00000.0000
ENSG00000156127BATF-4.9970-0.574332.75720.00000.0000
ENSG00000089225TBX5-4.7068-0.659331.49590.00000.0000
ENSG00000054598FOXC1-1.20823.042742.67540.00000.0000
ENSG00000007372PAX6-0.84702.084321.59470.00000.0001
ENSG00000160973FOXH1-0.79940.82746.30420.01200.0741
ENSG00000114315HES1-0.79921.436210.57550.00110.0124
ENSG00000150907FOXO10.61644.837832.52110.00000.0000
ENSG00000185630PBX10.62175.352827.55870.00000.0000
ENSG00000132170PPARG0.64025.537823.05510.00000.0000
ENSG00000139515PDX10.67140.18862.60650.10640.3237
ENSG00000134954ETS10.70104.857820.74110.00000.0001
ENSG00000182759MAFA0.72971.18769.74870.00180.0177
ENSG00000138378STAT40.75301.04135.53930.01860.1010
ENSG00000113916BCL60.75505.501325.03000.00000.0000
ENSG00000115415STAT10.77567.678445.76580.00000.0000
ENSG00000179388EGR31.16681.809329.37490.00000.0000
ENSG00000179348GATA21.16815.763724.04120.00000.0000
ENSG00000165556CDX23.12885.099271.73250.00000.0000

Possible transcription factor binding sites within the 2.5 kb promoter region of MIR100HG were predicted by the Match program (version 1.0).

Candidate transcription factors differentially expressed between CC and CC-CR (fold change >1.5) were listed.

Figure 5

GATA6 transcriptionally represses MIR100HG and is targeted by miR-125b in a double-negative feedback loop

(a) Immunoblot of GATA6 in CC and CC-CR cells cultured in 3D. In CC-CR, cells were treated with CTX (CTX+, 3 μg/ml) or normal culture medium (CTX-) for consecutive 14 days in 3D before protein extraction. Representative of 3 independent experiments. (b) Immunofluorescence of GATA6 (green) and nuclei (blue). Scale bar, 50 μm. (c) Representative IHC of GATA6 in CC and CC-CR xenografts (n=8). (d) qRT-PCR analysis of MIR100HG and GATA6 expression at indicated time points following CTX treatment (10 μg/ml) in CC cultured in 3D. n=3 independent experiments. (e) CC cells were transfected with two independent siRNAs against GATA6 or control (siCTL), treated with CTX (10 μg/ml) and subjected to qRT-PCR analysis. n=2 independent experiments performed in triplicate. **P<0.01 by Student's t test. (f) Luciferase reporter assays were performed by co-transfection of pGL3-MIR100HG promoter luciferase reporter with increasing concentrations of pcDNA3.1-GATA6 plasmid or empty vector control (CTL), along with a Renilla luciferase reporter. Luciferase activity was measured 36 h post-transfection and normalized to Renilla values. n=3 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test. (g) A schematic representation of consecutive deletion and mutation constructs spanning the -2000∼+500 region of MIR100HG promoter. The putative GATA6-binding sites within MIR100HG promoter are shown in black boxes. (h) The luciferase vector pGL3 driven by either wild-type, deletion or mutant (MUT) promoter was transfected in CC-CR, and luciferase activity was measured. n=3 independent experiments. *P<0.05, **P<0.01 by Student's t test. (i) Luciferase reporter analysis of a wild-type (WT) or mutant (MUT) GATA6 3′ UTR activity upon addition of either synthetic miR-125b or a negative control miR-CTL. **P<0.01 by Student's t test. (j) Immunoblots of GATA6 in stable miR-125b-transduced CC and 125b-Sp-transduced CC-CR. Representative of 3 independent experiments. (k) Box plots showing expression of GATA6 (left) and MIR100HG (middle) by stage from the TCGA CRC data repository. Right panel depicts MIR100HG expression in the lower (<25%) and the higher (>75%) quartile of GATA6 expression. *P<0.05, **P<0.01 by Mann-Whitney U test. n.s., not significant. Data represent mean ± s.d. in d-f, h, and i.

GATA6 is critical for gut endoderm development, and it both promotes and suppresses gastrointestinal and pancreatic neoplasia[26-29]. We found that MIR100HG expression decreased in cetuximab-treated CC, while GATA6 mRNA progressively increased over 48 h (Fig. 5d); however, this phenomenon did not occur in CC-CR (data not shown). GATA6 knockdown in CC (Extended Data Fig. 9a and Fig. 5e, top) caused MIR100HG upregulation and its expression no longer decreased upon cetuximab treatment (Fig. 5e, bottom), suggesting a repressive effect of GATA6 on MIR100HG. Luciferase reporter assays showed overexpression of GATA6 (Extended Data Fig. 9b) resulted in a concentration-dependent inhibition of MIR100HG promoter activity (Fig. 5f). Four putative GATA binding sites (G/A)GATA(A/T) were identified in the MIR100HG promoter region (Fig. 5g). Sequential deletions and mutations of these binding sites revealed that GATA binding site 2 (-1198 upstream of the TSS) is the major site for GATA6 repression of MIR100HG transcriptional activity (Fig. 5h). GATA6 repression of MIR100HG was also validated in HuTu80 cells with low expression of GATA6 and high expression of MIR100HG (Extended Data Fig. 9c and d). Chromatin occupancy of GATA6 at GATA-binding site 2 was confirmed by chromatin immunoprecipitation (ChIP) and electromobility shift assay (EMSA) using nuclear extracts from CC cells (Extended Data Fig. 9e and f).
Extended Data Fig. 9

A double-negative feedback loop between MIR100HG/miR-125b and GATA6. (a) Immunoblots of GATA6 expression in CC transfected with 2 independent siRNAs against GATA6 or control siRNA (siCTL). (b) Immunoblots of GATA6 expression in CC-CR transfected with either pcDNA3.1-GATA6 (WT GATA6), or pcDNA3.1-mutant GATA6 (MUT GATA6), or empty vector (CTL). (c) Luciferase reporter assays were performed in HuTu80 by co-transfection of pGL3-MIR100HG promoter luciferase reporter with increasing concentrations of pcDNA3.1-GATA6, and a Renilla control. Luciferase activity was normalized to Renilla values. n=3 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test. (d) The luciferase vector pGL3 driven by either wild-type, deletion, or mutant (MUT) promoter was transfected in HuTu80, and luciferase activity was measured. n=3 independent experiments. *P<0.05, **P<0.01 by Student's t test. (e) ChIP assays were performed with anti-GATA6 antibody or control IgG in CC-CR overexpressing either WT GATA6, MUT GATA6, or CTL. The abundance of DNA within the MIR100HG promoter region was assessed by qRT-PCR with a primer pair spanning the GATA-binding site 2. A primer pair 6.4 kb distal to the MIR100HG promoter (Distal) was used as control. Data are presented as relative enrichment normalized to control IgG. **P<0.01 by one-way ANOVA followed by LSD post-hoc test. (f) EMSA using nuclear extracts from CC and the indicated probes. Ab, antibody. Representative of 3 independent experiments. (g) Luciferase reporter analysis of a wild-type (WT) or MUT GATA6 3′ UTR activity upon addition of synthetic miR-125b in Caco-2. n=2 independent experiments. **P<0.01 by Student's t test. (h) Immunoblots of GATA6 in stable miR-125b-transduced Caco-2 and 125b-Sp-transduced HuTu80. (i) Box plots showing MIR100HG expression in the lower (<25%) and the higher (>75%) quartiles of GATA6 expression from GEO CRC datasets GSE14333 and GSE39582. **P<0.01 by Mann–Whitney U test. (j) MET genomic status detected by FISH assay. There was no obvious change in MET copy number in 10 paired tumor specimens pre- and post-cetuximab treatment. Representative images are shown. Red, MET locus; green, chromosome 7 centromere (CEP7). Scale bar, 20 μm. Data represent mean ± s.d. in c-e and g.

Of interest, we found that the 3′ UTR of GATA6 harbors a putative binding site for miR-125b (Extended Data Table 5). In both CC and Caco-2 cells, introduction of miR-125b reduced luciferase activity of the wild-type 3′ UTR reporter construct, but not when the miR-125b site was mutated (Fig. 5i and Extended Data Fig. 9g). As predicted, GATA6 levels were reduced in CC and Caco-2 cells stably expressing miR-125b, and conversely increased in CC-CR and HuTu80 cells expressing the miR-125b sponge (Fig. 5j and Extended Data Fig. 9h). Further analysis of the TCGA data repository indicated that GATA6 is significantly downregulated, whereas MIR100HG is significantly upregulated in stage IV CRC patients (Fig. 5k). Also, CRCs with lower quartile expression of GATA6 tend to have higher expression of MIR100HG in the TCGA data repository (Fig. 5k), as well as in two additional CRC datasets (Extended Data Fig. 9i). Together, these findings suggest a double-negative regulatory circuit between GATA6 and MIR100HG/miR-125b underlies cetuximab resistance.

Increased MIR100HG, miR-100, and miR-125b are found in CRC specimens at time of progression on cetuximab

To examine whether this mode of cetuximab resistance occurs in human CRC, we obtained paired tumor specimens from ten individuals prior to the start of cetuximab treatment and at the time of tumor progression (Extended Data Table 7). KRAS/NRAS/BRAF mutations had been excluded in tumor specimens obtained prior to treatment with cetuximab. qRT-PCR showed that miR-100 and miR-125b were coordinately overexpressed (rs=0.842, P<0.01) in tumors that progressed on treatment compared to pre-treatment levels (P<0.05, Fig. 6a). In addition, nuclear β-catenin immunoreactivity was significantly higher in tumors that progressed on cetuximab (Fig. 6b). miR-125b expression directly correlated with nuclear β-catenin staining (r=0.636, P<0.05); the correlation between miR-100 expression and nuclear β-catenin staining did not reach statistical significance (r=0.612, P=0.06). Conversely and consistent with our pre-clinical findings, there was reduced nuclear GATA6 expression in tumors that advanced on cetuximab (Fig. 6c). However, we did not find a significant inverse correlation between miR-100 and GATA6 (r=-0.455, P=0.187), or miR-125b and GATA6 (r=-0.515, P=0.128). By FISH analysis, the MIR100HG, miR-100, and miR-125b signals increased in tumors that progressed on treatment. In these same samples, there was increased β-catenin staining and reduced GATA6 staining (Fig. 6d). We excluded MET amplification by FISH in all ten paired specimens and sequenced the post-treatment tumors for mutations in KRAS/NRAS/BRAF (Extended Data Fig. 9j and Extended Data Table 8). NRAS and KRAS mutations were detected in 2 cases, respectively; we confirmed that these were likely acquired events by re-sequencing the pre-treatment DNA. In both cases, MIR100HG and miR-100/125b were increased. In the remaining 8 cases that lacked genetic resistance events, 5 cases exhibited upregulated MIR100HG and miR-100/125b in the tumors post-treatment. These clinical data support our pre-clinical findings and demonstrate that upregulation of MIR100HG and miR-100/125b occur in the setting of acquired cetuximab resistance in CRC patients, and this upregulation may both coincide with and be independent of genetic mutations associated with cetuximab resistance.
Extended Data Table 7

Clinic-pathological characteristics of metastatic CRC patients with paired specimen pre- and post-cetuximab treatment

No.1GenderAgePrimary siteMetastatic sitesDifferentiation2Cetuximab regimen3Best response to cetuximab4Site of specimen (pre)Site of specimen (post)
1M56left colonlung, liver, peritoneal cavityG2FOLFOX4+cetuximabPRsigmoidliver lesion
2F39left colonleft ovary and adnexa, lungG1FOLFIRI+cetuximabSDsigmoidleft adnexal mass
3M52right colonlung, liver, peritoneal cavityG1-G2FOLFIRI+cetuximabSDHepatic flexure of colonabdominal wall mass
4M52left colonliverG2-G3FOLFIRI+cetuximabSDleft colonliver lesion
5F55left colonliverG1FOLFOX4+cetuximabPRsigmoidliver lesion
6M55rectosigmoidliverG2FOLFOX4+cetuximabPRrectumrectum
7F48left colonlung, liver, peritoneal cavityG2mFOLFOX6+cetuximabSDsigmoidomental mass
8F81transverse colonliver, boneG3Cetuximab aloneSDtransverse colontransverse colon
9M57left colonliver and lungG2FOLFOX4+cetuximabPRsigmoidsigmoid
10M72rectumliverG2FOLFOX4+cetuximabPRrectumliver lesion

Cases 1, 2, 3 in Fig. 6d denote subject No. 2, 3, 5 in this table.

G1, well-differentiated, G2, moderately differentiated, G3, poorly differentiated.

Cetuximab 400 mg/m2 initial dose followed by 250 mg/m2 weekly thereafter with cetuximab dose intensity>90% were given to all subjects. Chemotherapy regimens: FOLFOX4 or mFOLFOX6, modified FOLFOX6 (5-fluorouracil, leucovorin, and oxaliplatin).

PR, Partial Response; SD, Stable Disease.

Figure 6

Increased MIR100HG and miR-100/125b are found in CRC patient specimens at time of progression on cetuximab

(a) qRT-PCR of miR-100 and miR-125b levels in 10 pairs of matched human CRC specimens pre- and post-cetuximab resistance. Each symbol represents mean value of an individual patient. *P<0.05 by Wilcoxon matched-pairs signed rank test. (b, c) Frequency of nuclear β-catenin-positive cells (b) and GATA6-positive cells (c) in 10 pairs of matched human CRC specimens pre- and post-cetuximab resistance. *P<0.05 by Wilcoxon matched-pairs signed rank test. (d) Representative FISH images of MIR100HG, miR-100/125b and corresponding IHC images of β-catenin and GATA6 staining in representative three paired human CRC specimens obtained pre- and post-cetuximab resistance. Scale bars, 50 μm (main); 500 μm (inset).

Extended Data Table 8

KRAS, NRAS, and BRAF mutational status and MET amplification status in CRC patients with paired specimens obtained prior to cetuximab (Pre) and at time of tumor progression (Post)

No.1Combined analysis of KRAS, NRAS, and BRAF2MET/CEP7 ratio3miR-100/125b expression change (Post vs Pre)
PrePostPrePost
1WTWT<2<2Up
2WTNRAS c.182A>T p.Q61L<2<2Up
3WTWTn.d.<2Up
4WTKRAS c.34G>A p.G12S<2<2Up
5WTWT<2<2Up
6WTWT<2<2Down
7WTWT<2<2Down
8WTWTn.d.<2Up
9WTWT<2<2Down
10WTWTn.d.<2Up

Cases 1, 2, 3 in Fig. 6d denote Subjects No. 2, 3, 5 in this table.

NRAS Q61L and KRAS G12S were identified in post-treatment specimens of Subject No. 2 and 4, respectively. In these two cases, DNA isolated from both the pre-and post-treatment samples was sequenced in parallel for KRAS, NRAS, and BRAF.

WT, wild-type; n.d., not detected.

Discussion

MIR100HG is a polycistronic miRNA host gene, which encodes miR-100, let-7a-2, and miR-125b-1 within its third intron. MIR100HG was first reported to participate in fate determination of human mesenchymal stem cells[30], and later found to be highly expressed in acute megakaryoblastic leukemia[12,13]. Increased MIR100HG expression is associated with a poor prognosis in cervical cancer[31], whereas its expression is reduced in breast cancer due to hypermethylation[32]. Increased expression of miR-100 and miR-125b are also correlated with gastric cancer progression in clinical samples[33]. In our study, concomitant upregulation of MIR100HG and miR-100/125b occurs in the setting of acquired and de novo cetuximab resistance in CRC and HNSCC cell lines. Moreover, we show these events can co-occur with KRAS/NRAS/BRAF mutation and in tumors of CRC patients that progressed on cetuximab. Analysis of TCGA CRC data repository revealed a stage-dependent increase of MIR100HG expression. These data support the hypothesis that MIR100HG and miR-100/125b are potential predictive biomarkers for cetuximab resistance. We identified that miR-100 and miR-125b coordinately contribute to cetuximab resistance by targeting five negative regulators of Wnt signaling. miR-100 targets DKK1 and ZNRF3; miR-125b also targets ZNRF3, as well as RNF43, DKK3, and APC2. Wnt signaling is tightly regulated and negative regulators act at many different levels[34,35]. DKK1 and DKK3 are secreted Wnt signaling antagonists of the Dickkopf family. DKK1 acts by binding and internalizing the Wnt co-receptor LRP5/6[36], while it is unclear how DKK3 attenuates Wnt signaling[37]. ZNRF3 and RNF43 are two closely related transmembrane E3 ubiquitin ligases that antagonize Wnt signaling through ubiquitylation and degradation of the Wnt receptor Frizzled and its co-receptor LRP5/6[38,39]. Although inactivating mutations have been reported for ZNRF3/RNF43[40], our data suggest that downregulation by miR-100/125b may represent an alternative mechanism of attenuating ZNRF3/RNF43 function. APC2 targets β-catenin for destruction and is functionally complementary to APC[41]; it was recently reported that APC2 recruits TNKS into the β-catenin destruction complex to regulate β-catenin proteolysis[42]. We previously identified APC as a target of miR-125b in leukemia cells[12], but did not examine APC in CC or CC-CR since it is mutated in these cells. However, it was recently reported that miR-125b targets APC in mutant β-catenin HCT116 cells that have wild-type APC[43]. The present study shows that miR-100 and miR-125b work together to target these five Wnt negative regulators, providing a novel regulatory mode for clustered miRNAs to cooperatively regulate this pathway. We cannot exclude that miR-100/125b contributes to cetuximab resistance through means other than Wnt signaling. For example, miR-125b can enhance tumor formation in the skin by targeting vacuolar protein-sorting 4 homolog B (Vps4b) and indirectly prolonging EGFR activity[44]. However, we observed no differences in VPS4b expression between CC and CC-CR. We also have not excluded an effect of the full-length 3 kb MIR100HG transcript on Wnt signaling. The present study adds to the literature describing crosstalk between EGFR and Wnt signaling[45,46]. For example, in APC-mutant CRC, increased EGFR signaling enhances Wnt activity, supporting the notion that Wnt signaling is further modulated in the setting of an impaired β-catenin degradation complex[47]. In a reciprocal manner, binding of Wnt ligands to their GPCR Frizzled receptors results in EGFR transactivation via metalloprotease-dependent, cell-surface ectodomain cleavage of EGFR ligands[48]. Moreover, increased Wnt signaling confers resistance to EGFR tyrosine kinase inhibitors in lung cancer[49,50]. The precise mechanism by which increased Wnt signaling confers cetuximab resistance is uncertain. It has been reported that Wnt signaling increases EGFR expression in liver[51]. We observed that cetuximab did not reduce p-EGFR, pERK1/2, or p-AKT in CC-CR as it does in CC. Although there is equivalent cell-surface EGFR staining in CC and CC-CR, this does not exclude differences in rates of EGFR internalization, recycling and degradation. Going forward, a system-wide approach should prove useful to help unravel mechanisms underlying the EGFR/Wnt crosstalk in this system. The role of GATA6 in cancer is complex and context-dependent; even in the same tumor type, conflicting evidence exists. For example, GATA6 promotes pancreatic carcinogenesis by activating Wnt signaling[52]. In separate studies, it serves a tumor suppressive role by maintaining a pancreatic differentiation program[53,54]. Likewise, in colonic neoplasia, opposing actions are reported. In colonic adenomas, GATA6 represses BMP expression, thereby enabling stem cell self-renewal[27], and in CRC cell lines, it enhances expression of Lgr5 and REG4 to promote clonogenicity and growth, respectively[55,56]. In contrast, our study supports a tumor-suppressive role for GATA6 in CRC. Analysis of the TCGA CRC repository reveals reduction in GATA6 expression in stage IV CRC along with increased MIR100HG expression. Reduced expression of GATA6 would permit increased expression of MIR100HG, and the corresponding increased expression of miR-125b would reinforce repression of GATA6. In this context, GATA6 serves a permissive tumor suppressive role by preventing Wnt signaling-enhanced cetuximab resistance. The present study has important therapeutic implications for CRC and HNSCC. It is increasingly appreciated that there are gradients of Wnt signaling in CRC[57] and that Wnt signaling can be modulated in the setting of APC loss-of-function[58,59]. In our model, MIR100HG- and miR-100/125b-mediated Wnt activation represents an adaption of activating compensatory pathways for cells to survive under EGFR inhibition. We show induction of DKK1 or DKK3 individually, or the combined addition of recombinant DKK1 and DKK3, overcomes cetuximab resistance in CC-CR. Both XAV-939, a tankyrase inhibitor, and ICG-001, a β-catenin-CBP inhibitor, augment the growth inhibitory effects of cetuximab. We propose that future trials in individuals with wild-type KRAS/NRAS/BRAF CRC should consider the levels of MIR100HG expression. In summary, we have identified a complex circuitry underlying cetuximab resistance by upregulation of MIR100HG and its embedded miRNAs (see Extended Data Fig. 10). miR-100 and miR-125b coordinately activate Wnt signaling by reducing expression of five negative regulators of Wnt signaling. miR-125b reinforces upregulation of MIR100HG by inhibiting GATA6 expression, which normally suppresses MIR100HG. We show that inhibition of Wnt signaling can overcome this mode of cetuximab resistance, underscoring the potential clinical relevance of the interactions between EGFR and Wnt signaling.
Extended Data Fig. 10

Model of a new mode of acquired and de novo cetuximab resistance. We propose a complex circuitry in which the lncRNA MIR100HG through embedded miR-100 and miR-125b confers cetuximab resistance by targeting and decreasing expression of five negative regulators of Wnt signaling, DKK1, DKK3, ZNRF3, RNF43, and APC2. This results in increased Wnt signaling and cetuximab resistance; this resistance can be overcome by blockade of Wnt signaling. We present evidence that GATA6 represses MIR100HG expression, but that miR-125b targets GATA6 to relieve this repression.

Data availability

RNA-Seq and small RNA-Seq data are available at the NCBI Gene Expression Omnibus (GEO) repository with accession code GSE82236. Whole-exome sequencing data are available at the GEO repository with accession code GSE76352. A Life Sciences Reporting Summary is available online.

Online Methods

2D and 3D cell culture

NCI-H508, Caco-2, SW403, SW948, HT29, SK-CO-1, DLD-1, SW480, SW837, SW48, SW620, LoVo, COLO205, T84, LS174T, NCI-H716, HCT8, HCT15, SW1116, RKO, COLO320DM, HuTu80, LS123, and HCT116 cell lines were from the American Type Culture Collection (ATCC). HCA-7, its derivatives CC and CC-CR, DiFi, GEO, LIM1215, and LIM2405 were maintained in the Coffey lab. The SNUC4 cell line was from the Korean Cell Line Bank and the V9P cell line was provided by John Mariadason (Olivia Newton-John Cancer Research Institute, Melbourne, Australia). HNSCC cell lines SCC25, its derived cetuximab-resistant sublines (CTX-R1, R3, R4, R5, R7, and R8), and UNC10 were maintained in Christine Chung's laboratory. All cell lines were confirmed to be free of mycoplasma contamination. Cells were grown in Dulbecco's Modified Eagle's Medium (DMEM, Corning) supplemented with 10% bovine growth serum, glutamine, nonessential amino acids, 100 U/ml penicillin and 100 μg/ml streptomycin (HyClone) in 5% CO2 at 37°C. 3D collagen cultures were set up using 3 layers of type-I collagen PureCol (Advanced BioMatrix) in triplicate as previously described[14,15]. Human recombinant DKK1 (rDKK1) and DKK3 (rDKK3) are from R&D Systems. Drugs are used as follows: cetuximab (Merck KGaA), Wnt pathway inhibitor ICG-001 and XAV-939 (Selleck Chemicals). Colonies were counted by GelCount colony counter (Oxford Optronix).

RNA-Seq analysis

Total RNA from cells embedded in collagen was isolated by TRIzol Reagent (Invitrogen) and then purified using RNeasy Mini Kit (Qiagen). The concentration and integrity of total RNA were estimated using the Qubit 2.0 Fluorometer (Invitrogen) and Agilent 2100 Bioanalyzer (Agilent Technologies), respectively. Polyadenylated RNAs were isolated using NEBNext Magnetic Oligo d(T)25 Beads. First strand synthesis was performed using NEBNext RNA First Strand Synthesis Module (New England BioLabs). Directional second strand synthesis was performed using NEBNext Ultra Directional Second Strand Synthesis Module. The NEBNext DNA Library Prep Master Mix Set for Illumina was used to prepare next-generation sequencing expression libraries per manufacturer's protocol. Accurate quantification for sequencing applications was determined using the qPCR-based KAPA Biosystems Library Quantification Kit (Kapa Biosystems). Paired-end (PE) sequencing (75bp) was performed on the NextSeq 500 sequencer (Illumina). RNA-Seq reads were aligned to the human genome hg19 using TopHat2[60], and the number of reads mapped to each gene was calculated by HTseq (http://www-huber.embl.de/users/anders/HTSeq/). Differentially expressed genes between CC and CC-CR were detected by edgeR based on negative binomial distribution[61]. The p-values were adjusted by Benjamini and Hochberg's multiple test correction procedures. Differential expression was determined based on fold-change (FC) and false discovery rate (FDR) with |log2(FC)| >1 and FDR<0.01.

Small RNA-Seq analysis

Approximately 1 μg of total RNA from each sample was utilized for small RNA library preparation using NEBNext Small RNA Library Prep Set for Illumina (New England BioLabs) following the manufacturer's protocol. Post PCR material was purified using QIAquick PCR Purification Kit (Qiagen). Post PCR yield and concentration of the prepared libraries were assessed using Qubit 2.0 Fluorometer and DNA 1000 chip on Agilent 2100 Bioanalyzer. Size selection of small RNA was done on the Pippin Prep instrument (Sage Science). Accurate quantification for sequencing applications was performed using the qPCR-based KAPA Biosystems Library Quantification Kit. Single end sequencing (50BP) was performed on a NextSeq 500 Sequencer (Illumina). Adapters from 3′ end of small RNA-Seq reads were trimmed by Cutadpt (http://code.google.com/p/cutadapt/). Sequences shorter than 15 bp were excluded from the downstream analysis. Reads were aligned to the human genome hg19 using Bowtie. Mapped reads were annotated, and miRNA expression was quantified using ncPRO-seq (version v1.5.1)[62] based on miRbase v19. Differentially expressed miRNAs between CC-CR and CC cells were detected by edgeR[61]. The p-values were adjusted by Benjamini and Hochberg's multiple test correction procedures. Differential expression was determined based on FC and FDR with |log2(FC)| >1 and FDR <0.01.

Whole-exome sequencing

DNA extraction was performed by the QIAamp DNA Mini Kit (Qiagen) following the manufacturer's instructions. The DNA was quantified by Nanodrop spectrophotometer (Thermo Fisher Scientific). Genomic DNA was sequenced using Illumina HiSeq 2500. Reads were aligned to the human genome hg19 with BWA, sorted, and indexed with SAMtools. Duplicated reads were marked by Picard (http://picard.sourceforge.net/). SNPs and Indels were called simultaneously on CC and CC-CR samples by SAMtools with base quality ≥30, reads with mapping quality ≥30, and mapping quality downgrading coefficient of 50. SNPs and Indels with strand bias P<0.01, base quality bias P<0.01, mapping quality bias P<0.01, or end distance bias P<0.01 were filtered out. Furthermore, SNPs within 3 bp around a gap were removed. SNPs and Indels were annotated, and their effects were predicted by snpEff and snpSift[63]. Strekla[64] was used to detect SNVs and indels that were present at a significantly different frequency between CC and CC-CR samples with default parameters except turning off the depth filter for exome sequencing data.

Constructs, oligonucleotides, infection and transfection of human cell lines

miRNA expression lentiviral vectors LeGO-cO:miR-100, LeGO-cO:miR-125b, LeGO-cO:miR-100/125b bicistron, and control empty vector were used as described previously[12]. Lentivirus produced in HEK293 cells was generated and collected using standard protocols[12]. GFP-positive infected cells were selected in Blasticidin S (10 μg/ml) followed by flow sorting. Stable miRNA knockdown was achieved by introducing the lentiviral miRNA sponge constructs that target either miR-100, miR-125b alone or both. Briefly, eight repeats of anti-sense miR-100 (5′-CACAAGTTCGGATCTACGGGTT-3′) or/and anti-sense miR-125b (5′-TCACAAGTTAGGGTCTCAGGGA-3′) were designed and synthesized following standard protocols[65], and then cloned into the pGLV3/H1/GFP vector (GenePharma). A control sponge was used, which includes eight repeats of an artificial miRNA (5′-AAGTTTTCAGAAAGCTAACA-3′)[66] that is not complementary to any known miRNA. GFP-positive infected cells were selected in puromycin (1 μg/ml) followed by flow sorting. Human DKK1 expression vector pcDNA3-DKK1-FLAG was kindly provided by Dr. Stuart Aaronson (Mount Sinai School of Medicine). Human DKK3 expression vector pCS2-DKK3-flag was from Addgene (plasmid #15496). Lentiviral-inducible expression constructs containing DKK1 or DKK3 under the control of a doxycycline-inducible promoter were constructed by transferring each ORF into the pInducer-20 lentiviral vector[67]. Infected cells were selected in G418 (200 μg/ml) to generate stable cell lines. Tetracycline-reduced FBS (Clontech) was substituted for all media for cells transduced with the pInducer-20 vectors. To induce expression of DKK1 or DKK3, 1 μg/ml doxycycline (Sigma-Aldrich) was added to the culture medium. pcDNA3.1-GATA6 and pcDNA3.1-mutant (mut) GATA6 by site-directed mutagenesis were kindly provided by Dr. Christine A. Iacobuzio-Donahue (Memorial Sloan Kettering Cancer Center). GATA6 Silencer Select siRNAs (ID s5605, s5606) and the Silencer Select negative control siRNA were used for transient transfection (Life Technologies). GATA6 expression plasmids or siRNAs were transfected into indicated cells using Lipofectamine 2000 or RNAiMAX Reagent (Thermo Fisher Scientific), respectively. Experiments were performed 48 h after transfection. For 3′ UTR luciferase reporter assay, the 3′ UTR fragments of DKK1, DKK3, ZNRF3, RNF43, and APC2 containing miR-100 or miR-125b putative target sites were amplified and cloned downstream of the SV40 promoter-driven Renilla luciferase cassette in psiCHECK-2 (Promega). For luciferase reporter assays to measure promoter activities, PCR products of sequential deletion fragments of human MIR100HG promoter were cloned into pGL3-Basic vector (Promega). A site-directed mutagenesis kit (Agilent Technologies) was used to mutate the miR-100, miR-125b, or GATA6 binding sites of these vectors. All sequences were confirmed by sequencing.

Quantitative RT-PCR

Analysis of mRNA and miRNA levels was performed on the StepOnePlus Real-Time PCR System (Applied Biosystems). For mRNA detection, cDNA was generated with the QuantiTect Reverse Transcription Kit (Qiagen). Diluted cDNA samples were amplified to establish a standard curve for calculation of relative target concentrations using Express SYBR GreenER qPCR SuperMix with Premixed ROX (Life Technologies). The housekeeping gene ACTB was used as an internal control. The primers for the genes of interest were synthesized by RealTimePrimers.com or Sigma-Aldrich (Extended Data Table 9). Analysis of lncRNA and miRNA levels was performed with the use of the TaqMan fast advanced master mix (Applied Biosystems). TaqMan lncRNA, miRNA, and Pri-miRNA expression assays (Life technologies) were used according to the manufacturer's instructions, with ACTB or U6 small nuclear RNA (U6 snRNA) as the internal control (Extended Data Table 10). The relative expression of RNAs was calculated using the comparative Ct method.
Extended Data Table 9

Primers used in the qRT-PCR assays for indicated genes

Gene NameForward Primer Sequence (5′ to 3′)Reverse Primer Sequence (5′ to 3′)
ACTBGGACTTCGAGCAAGAGATGGAGCACTGTGTTGGCGTACAG
DKK1AACAGCTATCCAAATGCAGTCACAGGGGAGTTCCATAAA
DKK3CTGGGAGCTAGAGCCTGATGTCATACTCATCGGGGACCTC
CCND1TTCAAATGTGTGCAGAAGGAGGGATGGTCTCCTTCATCTT
KLF4CGAACCCACACAGGTGAGAATACGGTAGTGCCTGGTCAGTTC
MYCACCAGAGAAACCTAACAGTGCCTCTTTCATTTCGGCCAGTTC
NKD1TGCCTCCTGAGAAGACTGACCATAGATGGTGTGCAGCAAG
PROX1TCACCTTATTCGGGAAGTGCGTACTGGTGACCCCATCGTT
S100A4AACTAAAGGAGCTGCTGACCCTGTTGCTGTCCAAGTTGCTC
CD44TAGGAGAAGGTGTGGGCAGAAGAGCTCACTGGGTTTCCTGTCTT
FOSL1AGTCAGGAGCTGCAGTGGATGGTTCAGTTCCTTCCTCCGGTTCCTGC
GATA6TGCAATGCTTGTGGACTCTAGTGGGGGAAGTATTTTTGCT
AXIN2TACCGGAGGATGCTGAAGGCCCACTGGCCGATTCTTCCTT
Extended Data Table 10

Primers used in the qRT-PCR assays for indicated miRNA or lncRNA

NameLife Technologies IDCategorySpecies
MIR100HGHs03680804_m1TaqMan ® LncRNA Assayhomo sapiens
pri-mir-100Hs03302731_priTaqMan ® Pri-miRNA Assayhomo sapiens
pri-mir-125b-1Hs03303095_priTaqMan ® Pri-miRNA Assayhomo sapiens
pri-let-7a-2Hs03302539_priTaqMan ® Pri-miRNA Assayhomo sapiens
miR-100# 4427975 000437TaqMan® MicroRNA Assayshomo sapiens
miR-125b# 4427975 000449TaqMan® MicroRNA Assayshomo sapiens
let-7a# 4427975 000377TaqMan® MicroRNA Assayshomo sapiens
ACTBHs01060665_g1TaqMan ® gene expression Assayhomo sapiens
U6 snRNA# 4427975 001973TaqMan® microRNA Control Assayshomo sapiens

5′ Rapid amplification of cDNA ends (RACE)

5′ RACE was used to determine transcriptional initiation sites of lncRNA MIR100HG using FirstChoice RLM-RACE Kit (Thermo Fisher Scientific), and Zero Blunt TOPO PCR Cloning Kit (Life Technologies) was used for sequencing according to the manufacturer's instructions. Two reverse primers for the TSS of MIR100HG were used in a nested PCR with the two 5′ primers from the kit. Outer primer: 5′-AAACCGGGCCCTCCAGTTCACTAT C-3′; Inner primer: 5′-TCTTTTCCATCCCCTTTGCATGTGG-3.

Western blot analysis

Whole cell lysates were prepared using RIPA buffer supplemented with protease inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitor (Roche). The nuclear extract was isolated using NE-PER nuclear and cytoplasmic extraction reagents (Thermo Fisher Scientific). Protein concentration was determined using the BCA Protein Assay Kit (Thermo Fisher Scientific). Primary antibodies were against DKK1 (Santa Cruz sc-25516), ZNRF3 (sc-86958), RNF43 (sc-165398), GATA6 (sc-7244), Lamin A/C (sc-7292), and GAPDH (sc-20357); DKK3 (Abcam #2459), APC2 (Abcam #80018), ZNRF3 (Abcam #122353), RNF43 (Abcam #84125), and p-EGFR Y1068 (Abcam #5644); β-Catenin (BD Biosciences #610154); EGFR (Millipore #06-847) and p-LRP6 (Millipore #07-2187); ERK1/2 (Cell Signaling Technology, CST, #4695), p-ERK1/2 (CST #4370), AKT (CST #9272), p-AKT (CST #4060), cleaved Caspase-3 (CST #9664), Cyclin-D1 (CST #2978), BIM (CST #2933), p-β-catenin (S552) (CST #9566), LRP6 (CST #2560), and GATA6 (CST #5851); β-actin (Sigma-Aldrich A1978). Immunoreactivity was detected and the signals were analyzed under nonsaturating conditions with an image densitometer (ChemiDoc MP Imager, Bio-Rad) and Image Lab software (Bio-Rad). All immunoblot analyses were performed at least three times.

Immunofluorescent staining

Immunofluorescent staining and confocal analysis were performed as described[14]. Primary antibodies were against cleaved Caspase-3 antibody (Abcam #2302), Ki-67 (Dako M7240), p-β-catenin (Y489) (Developmental Studies Hybridoma Bank, IA), and GATA6 (CST #5851). Alexa Fluor 568 Phalloidin was from Life Technologies. Secondary antibodies included Alexa Fluor 488, 568, or 647-conjugated goat anti-mouse or anti-rabbit IgGs (Life Technologies). Nuclei were stained with Hoechst 33342. Slides were mounted with Prolong Gold Antifade Reagent (Life Technologies) prior to imaging on a Zeiss LSM 710 confocal microscope. Quantification was done by manually counting at least 5 randomly chosen high power fields (HPFs) per sample.

Luciferase reporter assays

For 3′ UTR luciferase reporter assays, indicated cells cultured in 24-well plates were co-transfected with miR-100 or miR-125b mimic or negative control (Ambion) and indicated psiCHECK-2-3′ UTR wild-type or mutant plasmids using Lipofectamine 2000 (Thermo Fisher Scientific). Renilla and firefly luciferase activities were measured after 48 h with the dual-luciferase reporter assay system (Promega). Renilla luciferase activity was normalized to firefly activity and presented as relative luciferase activity. For luciferase reporter assay to measure promoter activities, indicated cells were co-transfected with pGL3-MIR100HG promoter fragment, pRL-SV40 Renilla luciferase reporter, and pcDNA3.1-GATA6 or mut GATA6 expression plasmid or empty vector control. The firefly and Renilla luciferase activity was measured with the dual-luciferase reporter assay system (Promega). Firefly luciferase activity was normalized to Renilla activity and presented as relative luciferase activity. All assays were performed in triplicate three times.

Chromatin immunoprecipitation (ChIP)

ChIP assays were performed using a Pierce Agarose ChIP Kit (Thermo Fisher Scientific) following the manufacturer's instructions, and ChIP-enriched DNA samples were analyzed by qPCR. Cells were cross-linked with 1% formaldehyde for 10 min at room temperature and quenched in glycine. Rabbit anti-GATA6 antibody (CST #5851) or normal rabbit IgG (BD Biosciences) were used for immunoprecipitation. The DNA was recovered and subjected to qPCR to amplify the binding sites of the MIR100HG promoter region. Data are presented as relative enrichment normalized to control IgG. The qPCR primer sets are: for GATA-binding site-2 (-1198 upstream of the TSS): forward, 5′-ACCTATCTCTGCTACTTATTTTATG-3′, reverse, 5′-CTATTTATCAGCACAGTTACTGG -3′; distal region primer sets: forward, 5′-GAATGCAGTAGTGGCTAG GAATG-3′, reverse, 5′-CTAACTCTCTAGGCTGTTATCTG-3′.

Electrophoretic mobility shift assay (EMSA)

Nuclear and cytoplasmic proteins were extracted using the NE-PER Nuclear and Cytoplasmic Extraction Reagents Kit (Thermo Fisher Scientific). LightShift Chemiluminescent EMSA Kit (Thermo Fisher Scientific) was used according to the manufacturer's protocol. Double-stranded biotin-labeled probe (5′-GCTCTCTATTTATCA GCACAGTTA-biotin-3′) was used, which corresponds to GATA-binding site-2 of the human MIR100HG gene promoter (-1198 upstream of the TSS). Nuclear extracts (8 μg) were incubated with labeled probe, poly (dI-dC), and the binding buffer for 30 min at room temperature. For supershift experiments, 1 μg of monoclonal rabbit anti-GATA6 (CST #5851) was added to the reaction mixture before the addition of labeled probe. For the binding competition experiment, an excess (200-fold) of unlabeled cold competitor probe or mutant probe (5′-GCTCTCTGCCCGCTGATACAGTTA-3′) was added into the reaction mixture. Bound DNA complexes were resolved on polyacrylamide gels and transferred to a nylon membrane (Thermo Fisher Scientific). Nylon membranes were cross-linked and chemiluminescent detection was performed.

Immunohistochemistry (IHC)

IHC for target molecules was performed on serial sections from tumor tissues of nude mice xenografts and CRC patients. Tissue sections were deparaffinized, subjected to antigen retrieval using target antigen retrieval solution (Dako), and incubated with primary antibodies against Ki67 (Dako M7240), cleaved Caspase-3 (CST #9661), β-catenin (BD Biosciences #610154), GATA6 (Abcam #22600), DKK1 (Abcam #61034), DKK3 (Abcam #115869), ZNRF3 (Abcam #122353), RNF43 (Abcam #129401), and APC2 (Abcam #113370). Then sections were incubated with Envision System HRP-labeled polymer anti-rabbit or anti-mouse secondary antibodies (Dako). The results of IHC were scored by two independent observers. Ki-67 and cleaved Caspase-3 staining was quantified by calculating positively stained cells in at least 5 randomly chosen HPFs of each sample. Quantification of other molecules was based on intensity and extent of staining according to the histological scoring method as previously described[68].

In vivo tumor growth in xenograft model

In vivo cetuximab treatment was performed using 6∼8-week-old female athymic BALB/c nude mice. All experiments were conducted under protocols approved by the Fourth Military Medical University Institutional Animal Care and Use Committee. Suspensions of the corresponding cells were subcutaneously injected into the flanks (6×106 tumor cells/150 μl PBS per spot; 6-8 mice in each group). Animals were weighed, and the tumor size was measured using bilateral caliper measurements. Tumor volume was calculated using the formula: Tumor maximum diameter (L) × the right angle diameter to that axis (W)2/2. When the tumors reached the determined size (around 100 mm3), mice were randomized into control and treatment groups. Cetuximab treatment was given at a dose of 1 mg/mouse, intraperitoneal (i.p.) injection, every 3 days. For Wnt pathway inhibitor ICG-001 in vivo treatment, the sodium phosphate form of ICG-001, synthesized by Vanderbilt Institute of Chemical Biology (VICB) Synthesis Core, was administered at a dose of 150 mg/kg body weight, i.p. injection every day. In vivo imaging system (IVIS, PerkinElmer) was used to detect GFP fluorescence in tumor-bearing mice. After 4 weeks of treatment, mice were sacrificed according to institutional ethical guidelines. Postmortem examination included tumor size and weight measurements, and then tumors were paraffin-embedded to perform hematoxylin and eosin (H&E) staining. The sample size for the experiments was based on the pilot studies and determined to ensure a power at 0.8 with type 1 error (α) at 0.05 of expected difference. Postmortem examination and data analysis were done by two investigators blinded to the group allocations.

Human CRC samples and subjects

All human CRC samples were obtained from the Xijing Hospital of Digestive Diseases. The study was approved by the Ethics Committee of Xijing Hospital with written informed consent obtained from all subjects. The pathological status of the specimens was provided by the Department of Pathology. In total, we analyzed 10 pairs of tumor specimen pre- and post-cetuximab treatment. Pre-cetuximab treated specimens were retrospectively obtained during surgical or biopsy under colonoscopy on subjects with CRC. After computed tomography (CT) of tumor lesions demonstrated tumor re-growth (disease progression) following initial response to cetuximab-based therapy, post-cetuximab treated specimens were collected whenever possible at the time of progression. RNA was extracted with RecoverAll Total Nucleic Acid Isolation Kit for FFPE (Ambion). IHC and FISH analyses were performed on formalin-fixed, paraffin-embedded (FFPE) tumor tissue sections. Blind evaluation was done by two pathologists.

Fluorescence in situ hybridization (FISH) assays

Locked nucleic acid-in situ hybridization (LNA-ISH) with tyramide signal amplification (TSA) was performed to detect lncRNA and miRNA as previously described[13,69]. All LNA probes were synthesized (Exiqon) including double biotin-labeled probe against MIR100HG, double digoxigenin (DIG)-labeled probe against miR-125b, double fluorescein-labeled probe against miR-100, DIG-labeled probe against U6 snRNA, and DIG-labeled scramble probe. Anti-Digoxigenin HRP Conjugate, anti-Fluorescein HRP Conjugate, Streptavidin-HRP Conjugate, and TSA Cy3 and Fluorescein Kit (all from PerkinElmer) were used for TSA methods. Confocal fluorescence microscopy was performed using a Zeiss LSM 710 confocal microscope. For detection of MET amplification, MET/CEP7 dual-color probes (Cytotest) were used for recognizing the MET gene status following the manufacturer's protocol. Analysis was according to the University of Colorado Cancer Center (UCCC) criteria. A MET/CEP7 ratio was established based on counting at least 200 cells.

Targeted Sanger sequencing of KRAS, NRAS and BRAF

KRAS/NRAS/BRAF mutations had been excluded in tumor specimens obtained prior to treatment with cetuximab. This study was carried out on 12 FFPE blocks of colorectal carcinomas (10 blocks obtained after disease progression upon cetuximab treatment and case 2 and 4 blocks before treatment). Genomic DNA was isolated using QIAamp DNA FFPE Tissue Kit (QIAGEN). Oncogenic alleles of KRAS (codon G12, G13, Q61, K117, and A146), NRAS (codon G12, G13, and Q61), and BRAF (codon G465, G468, Y472, D593, F594, L596, L597, T598, V600, and K601) were sequenced by targeted Sanger sequencing with PCR primers listed (Extended Data Table 11). The PCR products were then sequenced using BigDye Terminator 3.1 Cycle Sequencing Kit on a 96-capillary 3730XL DNA Sequencer (Applied Biosystems).
Extended Data Table 11

Primers used for Sanger sequencing of KRAS, NRAS, and BRAF

LocusPrimer CodePrime Sequence (5′ to 3′)
KRAS Codon G12, G13KRAS-exon2-FGTTCTAATATAGTCACATTTTCA
KRAS-exon2-RTCTATTGTTGGATCATATTCG
KRAS Codon Q61KRAS-exon3-FTCTCCCTTCTCAGGATTC
KRAS-exon3-RATTATTTATGGCAAATACACAAAG
KRAS Codon A146KRAS-exon4-FTTCTAGAACAGTAGACACAAAAC
KRAS-exon4-RGAGAGAAAAACTGATATATTAAATGAC
KRAS Codon K117KRAS-exon4-2FCTTTCCCAGAGAACAAATTAAAAG
KRAS-exon4-2RTCAATAAAAGGAATTCCATAACTTCT
NRAS Codon G12, G13NRAS-exon2-FCTGATTACTGGTTTCCAACAG
NRAS-exon2-RCCTCTATGGTGGGATCATATTC
NRAS Codon Q61NRAS-exon3-FCCCCAGGATTCTTACAGAAAA
NRAS-exon3-RTTGATGGCAAATACACAGAG
BRAF Codon G465, G468, Y472BRAF-exon11-2FGGGACTCGAGTGATGA
BRAF-exon11-2RAAAAGTTGTTAAACATATCCTATT
BRAF Codon D593, F594, L596, L597, T598, V600, K601BRAF-exon15-2FATGAGATCTACTGTTTTCCTTTACT
BRAF-exon15-2RCCTCAATTCTTACCATCCACA

Statistical analysis

Statistical analysis was performed by the SPSS 18.0 (SPSS Inc.) and R (version 3.3.1). The statistical significance between data sets was expressed as P-values, and P<0.05 was considered statistically significant. Two-tailed unpaired or paired Student's t test, ANOVA (Dunnett's or LSD post-hoc test), non-parametric signed rank test, Mann– Whitney U test, and Pearson correlation coefficients were used according to the type of experiment. For Pearson correlation of MIR100HG expression with 64-gene Wnt scores and RAS_AZ scores on 458 CRCs, the Log2 expression values of MIR100HG were obtained on 458 CRC tumors (10 samples without suitable microarray data were excluded from 468 CRCs previously reported[19,70]). A set of 64 “consensus” β-catenin (upregulated) targeted genes were adopted from a recent study[71], and a mean Log2 expression of the 64 genes was calculated as the Wnt/β-catenin pathway score on 458 CRC samples as previously described[19]. The RAS_AZ signature score, which measures MEK activation as a downstream index of RAS activity, was previously developed[20] and was pre-calculated on 458 CRC samples[19,70]. For comparison of the Wnt scores between CC and CC-CR, the 64-gene Wnt scores (Log2 expression) were calculated for CC and CC-CR (each with 3 replicates) similarly as described above and then subjected to the two-tailed unpaired Student's t test. RNA-Seq and small RNA-Seq data of CRC was obtained from TCGA Firehose developed by the Broad GDAC (https://confluence.broadinstitute.org/display/GDAC/ Dashboard-Stddata). mRNA expression array data of CRC/colon cancer was acquired from published studies (GEO accession: GSE14333, GSE39582)[72,73]. The gene expression abundances were Log2 transformed, and Pearson correlation coefficients were used to measure the correlation between MIR100HG and miR-100, miR-125b, and let-7a. Mann-Whitney U test was used to determine the expression difference of MIR100HG in the lower (0-25%) and the higher (>75%) quartiles of GATA6 expression. Establishment of cetuximab-resistant cells in 3D culture. (a) Five thousand cells/ml were cultured in type-1 collagen for 17 days. Fresh medium was added with different concentrations of CTX every 2 days, and colony number was determined using a GelCount plate reader. n=2 independent experiments performed in triplicate. (b) Twelve-day old CC and CR were treated with CTX (10 μg/ml) for 3 days. Representative images from 3 independent experiments are shown. Scale bars, 1000 μm. (c) Left: Representative fluorescence images of GFP signals captured from subcutaneous tumors, generated by injection of CC and CC-CR stably transduced with GFP-expressing lentivirus. Right: Quantification of radiant efficiency from tumors. n=8. **P<0.01 by paired Student's t test. (d) Representative H&E staining of the tumor xenografts from the indicated groups. Scale bar, 100 μm. (e) Quantification of IHC staining in Fig. 1g. n=8 mice. **P<0.01 by Student's t test. (f) CC and CC-CR cells grown on Transwell filters were incubated with Alexa Fluor 488-labeled C225 mAb directed against the extracellular domain of EGFR and then stained for F-actin (Phalloidin) and nuclei (DAPI). Scale bars, 20 μm. Data represent mean ± s.d. in a, c, and e. MIR100HG and miR-100/125b overexpression in cetuximab-resistant colorectal cancer cell lines. (a) qRT-PCR showing upregulation of pri-miR-100 and pri-miR-125b-1 in CC-CR compared to CC grown in 3D. In CC-CR, cells were treated with CTX (CTX+, 3 μg/ml) or normal culture medium (CTX-) for consecutive 14 days in 3D. (b) qRT-PCR showing upregulation of pri-let-7a-2 expression in CC-CR but unchanged expression of mature let-7a between the 2 cell lines. n=3 independent experiments performed in triplicate in a and b. Data represent mean ± s.d. **P<0.01 by one-way ANOVA followed by Dunnett's test compared with CC. (c) Left: a schematic diagram showing the PCR primers used in the 5′ RACE. Right: MIR100HG TSS was validated by 5′ RACE nested PCR in CC-CR with subsequent sequencing of the cloned fragments. Arrow indicates band of expected size. M, DNA marker. (d) Scatter plots of MIR100HG versus let-7a expression in TCGA CRC data repository. No correlation was found between those 2 molecules. (e, f) Expression of MIR100HG and miR-100/125b negatively correlates with cetuximab growth inhibition regardless of KRAS/BRAF mutational status. (e) Scatter plot of MIR100HG and miR-100/125b expression versus cetuximab inhibition rate in a panel of 30 CRC cell lines. (f) Twenty-one cell lines harbor KRAS or BRAF mutation, and 9 cell lines are KRAS/BRAF wild-type (WT). Pearson correlation coefficients (r) and P values are shown. MIR100HG and miR-100/125b expression in head and neck squamous cell cancer cell lines and modulation of miR-100 and/or miR-125b in CC and CC-CR cells. (a) qRT-PCR analysis of MIR100HG, miR-100, and miR-125b expression among the CTX-sensitive head and neck squamous cell carcinoma (HNSCC) cell line SCC25 and its derived CTX-resistant sublines (CTX-R1, R3, R4, R5, R7, and R8) upon continuous exposure to cetuximab, as well as UNC10, a de novo CTX-resistant cell line. n=3 independent experiments performed in triplicate. *P<0.05, **P<0.01 by one-way ANOVA followed by Dunnett's test compared with SCC25. (b) qRT-PCR of indicated miRNA expression in CC stably overexpressing miR-100, miR-125b, or Bicistron. (c) qRT-PCR of indicated miRNA expression in CC-CR stably expressing miR-100 sponge (100-Sp), miR-125b sponge (125b-Sp), or bicistron sponge (Bicistron-Sp). Values were normalized to U6 snRNA. n=3 experiments performed in triplicate. **P<0.01 by Student's t test. (d, e) Quantification of Ki-67 and Cleaved Casp-3 in Fig. 3c and d. n=4 independent experiments. *P<0.05, **P<0.01 by Student's t test. Data represent mean ± s.d. n.s., not significant. miR-100 and miR-125b cooperativity drives cetuximab resistance in colorectal cancer and head and neck squamous cell cancer cell lines. (a) Caco-2 cells stably overexpressing Bicistron or control (miR-CTL) were cultured in 3D for 5 days and treated with CTX (50 μg/ml) for 24 h. Immunofluorescence was performed for Cleaved Casp-3 (cyan) and Ki-67 (magenta) with quantification shown on the right. Scale bar, 50 μm. n=3 independent experiments. *P<0.05, **P<0.01 by Student's t test. (b) DLD-1 cells stably expressing Bicistron-Sp or control (CTL-Sp) were cultured in 3D for 10 days and treated with CTX (200 μg/ml) for 24 h. Staining of Cleaved Casp-3 (cyan) and Ki-67 (magenta) were shown. Scale bars, 50 μm. Quantification is shown on the right. n=3 independent experiments. *P<0.05, **P<0.01 by Student's t test. (c, d) Indicated cells were grown in 3D in normal medium (CTL) or treated with CTX (50 μg/ml for Caco-2, and 200 μg/ml for DLD-1) in 3D. The resultant colonies were counted. n=2 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test compared with miR-CTL or CTL-Sp. (e) Left: CTX-R7 cells stably expressing miR-100 and/or miR-125b sponges were grown in normal medium (CTL) or treated with CTX (30 μg/ml). Cell viability was measured by cell counting kit-8 (CCK-8) assays after 72 h. n=3 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test compared with CTL-Sp. Middle: qRT-PCR analysis of Wnt target genes in the stable bicistron sponge-transduced CTX-R7 cells. n=2 independent experiments performed in triplicate. *P<0.05, **P<0.01 by Student's t test. Right: CTX-R7 cells were treated with CTX (30 μg/ml) and/or ICG-001 (2 μM) for 72 h, and cell viability was measured by CCK-8 assays. n=2 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by by LSD post-hoc test. Data represent mean ± s.d. n.s., not significant. Effects of differential modulation of miR-100 and/or miR-125b on cetuximab responsiveness in CC and CC-CR in vivo. (a, b) Quantification of radiant efficiency from tumors (n=8) represented on Fig. 3e and f. **P<0.01 by paired Student's t test. (c-f) Representative IHC images and quantification of Ki-67 and Cleaved Casp-3 from indicated xenografts (n=8) treated with CTX. Scale bars, 50 μm. **P<0.01 by one-way ANOVA followed by Dunnett's test in e and f. Data represent mean ± s.d. in a, b, e, and f. n.s., not significant. miR-100/125b coordinately represses five Wnt/β-catenin negative regulators, resulting in increased Wnt signaling. (a) Left: Wnt activation in CC and CC-CR cells was measured by the 64 Wnt/β -catenin target genes (Wnt signature score). **P<0.01 by Student's t test. Right: Scatter plots of MIR100HG expression versus 64-gene Wnt signature score on 458 CRC. Pearson correlation coefficients (r) and P values are shown. (b) Immunoblots of DKK1, DKK3, ZNRF3, RNF43, and APC2 levels from 3D cell lysates of CC and CC-CR. In CC-CR, cells were treated with CTX (CTX+, 3 μg/ml) or normal culture medium (CTX-) for 14 days in 3D before protein extraction. Representative of 3 independent experiments. (c) Top : representative IHC images of DKK1, DKK3, ZNRF3, RNF43, and APC2 in CC and CC-CR xenografts (n=8). Bottom: measurement of protein expression, **P<0.01 by Mann-Whitney U test. (d) Dual luciferase assays of genes predicted to be regulated by miR-100 or miR-125b in Caco-2. Renilla luciferase activity was normalized to firefly activity. n=2 independent experiments. **P<0.01 by Student's t test. (e) Immunoblots of indicated proteins in stable miRNA-transduced Caco-2 and sponge (Sp)-transduced HuTu80. Representative of 2 independent experiments. (f) Immunoblot of nuclear and cytoplasmic extracts for β-catenin and p-β-catenin (S552). Loading controls were GAPDH for cytoplasmic fractions and Lamin A/C for nuclear fractions. (g) CC and CC-CR in 3D were treated with CTX (10 μg/ml) and/or Wnt3a (100 ng/ml). Immunoblots of indicated proteins after 48 h of treatment are shown. Representative of 3 independent experiments. (h) qRT-PCR analysis of Wnt targets CCND1, CD44, FOSL1, and NKD1 mRNAs at indicated time points following CTX (10 μg/ml) treatment in 3D. n=2 independent experiments performed in triplicate. **P<0.01 by two-way ANOVA test. Data represent mean ± s.d. in a, c, d, and h. Effects of differential modulation of miR-100 and/or miR-125b on nuclear β-catenin expression levels. (a) Immunoblots for β-catenin from nuclear fractions in the CC and Caco-2 cells overexpressing miR-100 and/or miR-125b, or CC-CR, DLD-1 and CTX-R7 cells expressing miR-100 and/or miR-125b sponges. Lamin A/C served as the control for nuclear fractions. Representative of 2 independent experiments. (b) Representative IHC of β-catenin in the indicated xenografts (n=8). Scale bars: 50 μm. Quantification of nuclear β-catenin positive cells is shown. Data represent mean ± s.d. **P<0.01 by Student's t test. Blockade of Wnt signaling restores cetuximab responsiveness to cetuximab-resistant cells. (a, b) Left: CC-CR doxycycline (Dox)-on DKK1 or DKK3 cells were cultured in the presence or absence of Dox (1 μg/ml) and harvested at 48 h. Total cell lysates and conditioned media were harvested and subjected to immunoblot analysis. Right: indicated cells were grown in 3D in normal medium or treated with CTX (3 μg/ml). The resultant colonies were counted after 18 days. n=3 experiments performed in triplicate. **P<0.01 by Student's t test. (c) CC-CR cells were grown in 3D in normal medium (CTL), treated with CTX (3 μg/ml) or in combination with recombinant DKK1 (rDKK1) and DKK3 (rDKK3) in 3D every 2 days. The resultant colonies were stained after 18 days for Cleaved Casp-3 (green) and Ki-67 (red). Scale bar, 50 μm. Quantification was shown. n=3 independent experiments. (d) Immunoblots for β-catenin from nuclear and cytoplasmic fractions of indicated cells upon CTX (10 μg/ml) treatment. Loading controls were GAPDH for cytoplasmic fractions and Lamin A/C for nuclear fractions. (e) CC-CR were treated with CTX (3 μg/ml), and/or XAV-939 (1, 5, 10 μM), and/or ICG-001 (1, 2.5, 5 μM) in 3D for 18 days, and colony number was determined. n=3 experiments performed in triplicate. (f) DLD-1 and HCT8 cells were treated with CTX (200 μg/ml) and/or ICG-001 (4 μM) for 14 days in 3D, and colony number was determined. n=2 independent experiments performed in triplicate. (g) Quantification of radiant efficiency from tumors (n=6) represented on Fig. 4i. **P<0.01 by paired Student's t test. (h) Representative IHC images and quantification of Ki-67 and Cleaved Casp-3 from CC-CR xenografts (n=6) treated with control saline (CTL), or CTX (1 mg/mouse, i.p. injection, every 3 days), and/or ICG-001 (150 mg/kg, i.p. injection, daily). Scale bar, 50 μm. *P<0.05, **P<0.01 by one-way ANOVA followed by Dunnett's test in c, e, and h, and one-way ANOVA followed by LSD post-hoc test in f. Data represent mean ± s.d. in a-c and e-h. n.s., not significant. A double-negative feedback loop between MIR100HG/miR-125b and GATA6. (a) Immunoblots of GATA6 expression in CC transfected with 2 independent siRNAs against GATA6 or control siRNA (siCTL). (b) Immunoblots of GATA6 expression in CC-CR transfected with either pcDNA3.1-GATA6 (WT GATA6), or pcDNA3.1-mutant GATA6 (MUT GATA6), or empty vector (CTL). (c) Luciferase reporter assays were performed in HuTu80 by co-transfection of pGL3-MIR100HG promoter luciferase reporter with increasing concentrations of pcDNA3.1-GATA6, and a Renilla control. Luciferase activity was normalized to Renilla values. n=3 independent experiments performed in triplicate. **P<0.01 by one-way ANOVA followed by Dunnett's test. (d) The luciferase vector pGL3 driven by either wild-type, deletion, or mutant (MUT) promoter was transfected in HuTu80, and luciferase activity was measured. n=3 independent experiments. *P<0.05, **P<0.01 by Student's t test. (e) ChIP assays were performed with anti-GATA6 antibody or control IgG in CC-CR overexpressing either WT GATA6, MUT GATA6, or CTL. The abundance of DNA within the MIR100HG promoter region was assessed by qRT-PCR with a primer pair spanning the GATA-binding site 2. A primer pair 6.4 kb distal to the MIR100HG promoter (Distal) was used as control. Data are presented as relative enrichment normalized to control IgG. **P<0.01 by one-way ANOVA followed by LSD post-hoc test. (f) EMSA using nuclear extracts from CC and the indicated probes. Ab, antibody. Representative of 3 independent experiments. (g) Luciferase reporter analysis of a wild-type (WT) or MUT GATA6 3′ UTR activity upon addition of synthetic miR-125b in Caco-2. n=2 independent experiments. **P<0.01 by Student's t test. (h) Immunoblots of GATA6 in stable miR-125b-transduced Caco-2 and 125b-Sp-transduced HuTu80. (i) Box plots showing MIR100HG expression in the lower (<25%) and the higher (>75%) quartiles of GATA6 expression from GEO CRC datasets GSE14333 and GSE39582. **P<0.01 by Mann–Whitney U test. (j) MET genomic status detected by FISH assay. There was no obvious change in MET copy number in 10 paired tumor specimens pre- and post-cetuximab treatment. Representative images are shown. Red, MET locus; green, chromosome 7 centromere (CEP7). Scale bar, 20 μm. Data represent mean ± s.d. in c-e and g. Model of a new mode of acquired and de novo cetuximab resistance. We propose a complex circuitry in which the lncRNA MIR100HG through embedded miR-100 and miR-125b confers cetuximab resistance by targeting and decreasing expression of five negative regulators of Wnt signaling, DKK1, DKK3, ZNRF3, RNF43, and APC2. This results in increased Wnt signaling and cetuximab resistance; this resistance can be overcome by blockade of Wnt signaling. We present evidence that GATA6 represses MIR100HG expression, but that miR-125b targets GATA6 to relieve this repression. Original images of immunoblots with molecular weight standards. FPKM stands for Fragments Per Kilobase of transcript per Million mapped reads. Differential expression analysis was performed on RNA-Seq counts using edgeR. Mutational status of 30 CRC cell lines used in Fig. 2f and their response to cetuximab Data came from Medico, E. et al, Nat Commun, 2015[17]. Data came from Jhawer, M. et al, Cancer Res, 2008[16]. Experimental data from the present study. Putative binding sites of miR-100 or miR-125b were mutated and highlighted in red. Possible transcription factor binding sites within the 2.5 kb promoter region of MIR100HG were predicted by the Match program (version 1.0). Candidate transcription factors differentially expressed between CC and CC-CR (fold change >1.5) were listed. Cases 1, 2, 3 in Fig. 6d denote subject No. 2, 3, 5 in this table. G1, well-differentiated, G2, moderately differentiated, G3, poorly differentiated. Cetuximab 400 mg/m2 initial dose followed by 250 mg/m2 weekly thereafter with cetuximab dose intensity>90% were given to all subjects. Chemotherapy regimens: FOLFOX4 or mFOLFOX6, modified FOLFOX6 (5-fluorouracil, leucovorin, and oxaliplatin). PR, Partial Response; SD, Stable Disease. KRAS, NRAS, and BRAF mutational status and MET amplification status in CRC patients with paired specimens obtained prior to cetuximab (Pre) and at time of tumor progression (Post) Cases 1, 2, 3 in Fig. 6d denote Subjects No. 2, 3, 5 in this table. NRAS Q61L and KRAS G12S were identified in post-treatment specimens of Subject No. 2 and 4, respectively. In these two cases, DNA isolated from both the pre-and post-treatment samples was sequenced in parallel for KRAS, NRAS, and BRAF. WT, wild-type; n.d., not detected.
  71 in total

1.  The miR-363-GATA6-Lgr5 pathway is critical for colorectal tumourigenesis.

Authors:  Shinnosuke Tsuji; Yoshihiro Kawasaki; Shiori Furukawa; Kenzui Taniue; Tomoatsu Hayashi; Masumi Okuno; Masaya Hiyoshi; Joji Kitayama; Tetsu Akiyama
Journal:  Nat Commun       Date:  2014       Impact factor: 14.919

2.  ncPRO-seq: a tool for annotation and profiling of ncRNAs in sRNA-seq data.

Authors:  Chong-Jian Chen; Nicolas Servant; Joern Toedling; Alexis Sarazin; Antonin Marchais; Evelyne Duvernois-Berthet; Valérie Cognat; Vincent Colot; Olivier Voinnet; Edith Heard; Constance Ciaudo; Emmanuel Barillot
Journal:  Bioinformatics       Date:  2012-10-07       Impact factor: 6.937

Review 3.  Secreted and transmembrane wnt inhibitors and activators.

Authors:  Cristina-Maria Cruciat; Christof Niehrs
Journal:  Cold Spring Harb Perspect Biol       Date:  2013-03-01       Impact factor: 10.005

4.  Emergence of Multiple EGFR Extracellular Mutations during Cetuximab Treatment in Colorectal Cancer.

Authors:  Sabrina Arena; Beatriz Bellosillo; Giulia Siravegna; Alejandro Martínez; Israel Cañadas; Luca Lazzari; Noelia Ferruz; Mariangela Russo; Sandra Misale; Iria González; Mar Iglesias; Elena Gavilan; Giorgio Corti; Sebastijan Hobor; Giovanni Crisafulli; Marta Salido; Juan Sánchez; Alba Dalmases; Joaquim Bellmunt; Gianni De Fabritiis; Ana Rovira; Federica Di Nicolantonio; Joan Albanell; Alberto Bardelli; Clara Montagut
Journal:  Clin Cancer Res       Date:  2015-01-26       Impact factor: 12.531

5.  Dishevelled promotes Wnt receptor degradation through recruitment of ZNRF3/RNF43 E3 ubiquitin ligases.

Authors:  Xiaomo Jiang; Olga Charlat; Raffaella Zamponi; Yi Yang; Feng Cong
Journal:  Mol Cell       Date:  2015-04-16       Impact factor: 17.970

Review 6.  Wnt/beta-catenin signaling in cancer stemness and malignant behavior.

Authors:  Riccardo Fodde; Thomas Brabletz
Journal:  Curr Opin Cell Biol       Date:  2007-02-16       Impact factor: 8.382

7.  The pINDUCER lentiviral toolkit for inducible RNA interference in vitro and in vivo.

Authors:  Kristen L Meerbrey; Guang Hu; Jessica D Kessler; Kevin Roarty; Mamie Z Li; Justin E Fang; Jason I Herschkowitz; Anna E Burrows; Alberto Ciccia; Tingting Sun; Earlene M Schmitt; Ronald J Bernardi; Xiaoyong Fu; Christopher S Bland; Thomas A Cooper; Rachel Schiff; Jeffrey M Rosen; Thomas F Westbrook; Stephen J Elledge
Journal:  Proc Natl Acad Sci U S A       Date:  2011-02-09       Impact factor: 11.205

8.  A small molecule inhibitor of beta-catenin/CREB-binding protein transcription [corrected].

Authors:  Katayoon H Emami; Cu Nguyen; Hong Ma; Dae Hoon Kim; Kwang Won Jeong; Masakatsu Eguchi; Randall T Moon; Jia-Ling Teo; Se Woong Oh; Hak Yeop Kim; Sung Hwan Moon; Jong Ryul Ha; Michael Kahn
Journal:  Proc Natl Acad Sci U S A       Date:  2004-08-16       Impact factor: 11.205

9.  Transcription factors GATA-4 and GATA-6 in normal and neoplastic human gastrointestinal mucosa.

Authors:  Hanna Haveri; Mia Westerholm-Ormio; Katri Lindfors; Markku Mäki; Erkki Savilahti; Leif C Andersson; Markku Heikinheimo
Journal:  BMC Gastroenterol       Date:  2008-04-11       Impact factor: 3.067

10.  GATA6 regulates EMT and tumour dissemination, and is a marker of response to adjuvant chemotherapy in pancreatic cancer.

Authors:  Paola Martinelli; Enrique Carrillo-de Santa Pau; Trevor Cox; Bruno Sainz; Nelson Dusetti; William Greenhalf; Lorenzo Rinaldi; Eithne Costello; Paula Ghaneh; Núria Malats; Markus Büchler; Marina Pajic; Andrew V Biankin; Juan Iovanna; John Neoptolemos; Francisco X Real
Journal:  Gut       Date:  2016-06-20       Impact factor: 23.059
View more
  125 in total

1.  Identification of functional long non-coding RNAs in gastric cancer by bioinformatics analysis.

Authors:  Yang Li; Dongyang Ma; Tang Li; Yuan Yin
Journal:  Int J Exp Pathol       Date:  2020-07-01       Impact factor: 1.925

2.  MicroRNA-Related Genetic Variants Associated with Survival of Head and Neck Squamous Cell Carcinoma.

Authors:  Owen M Wilkins; Alexander J Titus; Lucas A Salas; Jiang Gui; Melissa Eliot; Rondi A Butler; Erich M Sturgis; Guojun Li; Karl T Kelsey; Brock C Christensen
Journal:  Cancer Epidemiol Biomarkers Prev       Date:  2018-06-07       Impact factor: 4.254

3.  Colorectal cancer: miR-100 and miR-125b induce cetuximab resistance in CRC.

Authors:  Hugh Thomas
Journal:  Nat Rev Gastroenterol Hepatol       Date:  2017-11-02       Impact factor: 46.802

4.  LncRNA-NBAT-1 modulates esophageal cancer proliferation via PKM2.

Authors:  Bo Zhao; Peng Cao; Shan Hu; Fan Li; Kangle Kong; Yukun Zu
Journal:  Am J Transl Res       Date:  2019-09-15       Impact factor: 4.060

5.  MiR-218 and miR-100 polymorphisms as markers of irinotecan-based chemotherapy response in metastatic colorectal cancer.

Authors:  Dimitra-Ioanna Lampropoulou; Gerasimos Aravantinos; Konstantinos Laschos; Theodosis Theodosopoulos; Christos Papadimitriou; Maria Gazouli
Journal:  Int J Colorectal Dis       Date:  2019-10-09       Impact factor: 2.571

6.  miRNA-independent function of long noncoding pri-miRNA loci.

Authors:  Daniel He; David Wu; Soren Muller; Lin Wang; Parna Saha; Sajad Hamid Ahanger; Siyuan John Liu; Miao Cui; Sung Jun Hong; Miten Jain; Hugh E Olson; Mark Akeson; Joseph F Costello; Aaron Diaz; Daniel A Lim
Journal:  Proc Natl Acad Sci U S A       Date:  2021-03-30       Impact factor: 11.205

7.  Plasma exosomal microRNA-125b as a monitoring biomarker of resistance to mFOLFOX6-based chemotherapy in advanced and recurrent colorectal cancer patients.

Authors:  Takahiro Yagi; Hisae Iinuma; Tamuro Hayama; Keiji Matsuda; Keijirou Nozawa; Mitsuo Tsukamoto; Ryu Shimada; Takuya Akahane; Takeshi Tsuchiya; Tsuyoshi Ozawa; Yojiro Hashiguchi
Journal:  Mol Clin Oncol       Date:  2019-08-14

Review 8.  One locus with two roles: microRNA-independent functions of microRNA-host-gene locus-encoded long noncoding RNAs.

Authors:  Qinyu Sun; You Jin Song; Kannanganattu V Prasanth
Journal:  Wiley Interdiscip Rev RNA       Date:  2020-09-17       Impact factor: 9.957

9.  LncRNA MIR155HG Promotes Temozolomide Resistance by Activating the Wnt/β-Catenin Pathway Via Binding to PTBP1 in Glioma.

Authors:  Xin He; Jie Sheng; Wei Yu; Kejian Wang; Shujuan Zhu; Qian Liu
Journal:  Cell Mol Neurobiol       Date:  2020-06-11       Impact factor: 5.046

10.  Exosomal transfer of miR‑25‑3p promotes the proliferation and temozolomide resistance of glioblastoma cells by targeting FBXW7.

Authors:  Jianxin Wang; Tianxiao Li; Bin Wang
Journal:  Int J Oncol       Date:  2021-07-19       Impact factor: 5.650
View more

北京卡尤迪生物科技股份有限公司 © 2022-2023.