Advertisement

Lost in translation: Revisiting the use of tyrosine kinase inhibitors in colorectal cancer

  • Author Footnotes
    1 Department of Medical Oncology, Radboud Institute for Health Sciences, Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Kirti K. Iyer
    Footnotes
    1 Department of Medical Oncology, Radboud Institute for Health Sciences, Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Affiliations
    Department of Medical Oncology, Radboud Institute for Health and Sciences, Radboud University Medical Centre, Nijmegen, The Netherlands

    Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Centre, Nijmegen, The Netherlands
    Search for articles by this author
  • Author Footnotes
    2 Department of Pharmacy, Radboud Institute for Health Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Nielka P. van Erp
    Footnotes
    2 Department of Pharmacy, Radboud Institute for Health Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Affiliations
    Department of Pharmacy, Radboud Institute for Health and Sciences, Radboud University Medical Centre, Nijmegen, The Netherlands
    Search for articles by this author
  • Author Footnotes
    3 Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Daniele V.F. Tauriello
    Footnotes
    3 Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Affiliations
    Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Centre, Nijmegen, The Netherlands
    Search for articles by this author
  • Author Footnotes
    4 Department of Medical Oncology, Radboud Institute for Health Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Henk M.W. Verheul
    Footnotes
    4 Department of Medical Oncology, Radboud Institute for Health Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Affiliations
    Department of Medical Oncology, Radboud Institute for Health and Sciences, Radboud University Medical Centre, Nijmegen, The Netherlands
    Search for articles by this author
  • Dennis Poel
    Correspondence
    Corresponding author at: Department of Medical Oncology, Radboud Institute for Health Sciences, Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    Affiliations
    Department of Medical Oncology, Radboud Institute for Health and Sciences, Radboud University Medical Centre, Nijmegen, The Netherlands

    Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Centre, Nijmegen, The Netherlands
    Search for articles by this author
  • Author Footnotes
    1 Department of Medical Oncology, Radboud Institute for Health Sciences, Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    2 Department of Pharmacy, Radboud Institute for Health Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    3 Department of Cell Biology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
    4 Department of Medical Oncology, Radboud Institute for Health Sciences, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands.
Open AccessPublished:September 20, 2022DOI:https://doi.org/10.1016/j.ctrv.2022.102466

      Highlights

      • Over 50 tyrosine kinase inhibitors (TKIs) have been approved by the FDA.
      • Around 42 TKIs have shown preclinical anticancer activity in colorectal cancer.
      • Only 1 TKI, regorafenib has been approved by FDA for patients with colorectal cancer.
      • There is a huge gap in translational of TKIs to the clinic for colorectal cancer.

      Abstract

      Patients with advanced or metastatic colorectal cancer ((m)CRC) have limited effective treatment options resulting in high mortality rates. A better understanding of the molecular basis of this disease has led to growing interest in small molecule tyrosine kinase inhibitors (TKIs) for its treatment. However, of around 42 TKIs demonstrating preclinical anti-tumour activity, and despite numerous clinical trials, only 1 has been approved for clinical use in mCRC. Clearly, there is a huge gap in the translation of these targeted therapies to the clinic. This underlines the limitations of preclinical models to predict clinical drug efficacy and to fully characterize the mechanism of action. Moreover, several relevant topics remain poorly resolved. Do we know the actual intracellular concentrations that are required for anticancer efficacy, and what range of intra-tumoral drug concentrations is reached in clinical setting? Are the intended targeted kinases responsible for the anti-cancer activity or are other unexpected cellular targets involved? Do we have any idea of the effect of these drugs on the tumour microenvironment and does this help explain therapy success, failure or heterogeneity? In this review, we address these questions and discuss concepts that jointly complicate the clinical translation of TKIs for CRC. Finally, we will argue that an integrated approach with more sophisticated preclinical models and techniques may improve the prediction of clinical treatment efficacy.

      Keywords

      Introduction

      Colorectal cancer (CRC) is the 4th most lethal cancer in the world approximately causing 900,000 deaths every year. Recent advances in treatment options including surgery, radiotherapy, chemotherapy, and targeted therapy—along with early diagnosis—have improved the overall survival (OS) rate of patients with CRC [
      • Dekker E.
      • et al.
      Colorectal cancer.
      ]. In the cases of metastatic disease, fluoropyrimidine with either oxaliplatin and/or irinotecan (FOLFOXIRI/FOLFOX/FOLFIRI), often combined with bevacizumab or epidermal growth receptor (EGFR; cetuximab or panitumumab) are used as first-line chemotherapy regimens [
      • Gravalos C.
      • et al.
      Role of tyrosine kinase inhibitors in the treatment of advanced colorectal cancer.
      ,
      • Rossini D.
      • et al.
      Treatments after progression to first-line FOLFOXIRI and bevacizumab in metastatic colorectal cancer: a pooled analysis of TRIBE and TRIBE2 studies by GONO.
      ]. Immunotherapy (pembrolizumab/nivolumab-ipilimumab), are used for patients with microsatellite instability-high unresectable mCRC [

      Wookey V, Grothey A. Update on the role of pembrolizumab in patients with unresectable or metastatic colorectal cancer. Therap Adv Gastroenterol 2021;14:17562848211024460.

      ,
      • Andre T.
      • et al.
      Pembrolizumab in Microsatellite-Instability-High Advanced Colorectal Cancer.
      ,
      • Casak S.J.
      • et al.
      FDA Approval Summary: Pembrolizumab for the First-line Treatment of Patients with MSI-H/dMMR Advanced Unresectable or Metastatic Colorectal Carcinoma.
      ,
      • Franke A.J.
      • et al.
      Immunotherapy for Colorectal Cancer: A Review of Current and Novel Therapeutic Approaches.
      ]. Upon relapse, available options include second, third or fourth-line chemotherapy that does not include agents that have been used in first-line, the combination of cetuximab plus encorafenib for patients with BRAFV600e mutations, TAS-102, or the multi-kinase inhibitor regorafenib [
      • Van Cutsem E.
      • et al.
      Metastatic colorectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up.
      ,
      • Bekaii-Saab T.
      • et al.
      Third- or Later-line Therapy for Metastatic Colorectal Cancer: Reviewing Best Practice.
      ]. Unfortunately, despite the initial high chance of clinical benefit from currently available therapies, the 5-year survival rate of patients with mCRC for whom no local therapy is available remains poor (<15 %) mainly due to acquired resistance to treatment [
      • Wang J.
      • et al.
      Metastatic patterns and survival outcomes in patients with stage IV colon cancer: A population-based analysis.
      ]. This points out the urgent need for novel and more effective treatment strategies.
      In parallel to targeted monoclonal antibodies, growing knowledge of the molecular basis and oncogenic signalling of CRC progression has prompted the testing of many small molecule tyrosine kinase inhibitors (TKIs). These can inhibit key enzymes that regulate signalling pathways involved in cell growth, differentiation, proliferation and survival [
      • Gravalos C.
      • et al.
      Role of tyrosine kinase inhibitors in the treatment of advanced colorectal cancer.
      ,
      • Garcia-Aranda M.
      • Redondo M.
      Targeting Receptor Kinases in Colorectal Cancer.
      ,
      • Paul M.K.
      • Mukhopadhyay A.K.
      Tyrosine kinase - Role and significance in Cancer.
      ] (Fig. 1). There are over 90 tyrosine kinases in the human genome, out of which 58 are categorized as receptor tyrosine kinases (RTKs) that are further divided into 20 subfamilies depending on their kinase domain sequence [
      • Robinson D.R.
      • Wu Y.M.
      • Lin S.F.
      The protein tyrosine kinase family of the human genome.
      ,

      Thomson RJ, Moshirfar M, Ronquillo Y. Tyrosine Kinase Inhibitors, in StatPearls. 2022: Treasure Island (FL).

      ]. Several RTKs, or the pathways these kinases operate in, are mutated or amplified in CRC, and comprise attractive therapeutic targets [
      • Huang L.
      • Jiang S.
      • Shi Y.
      Tyrosine kinase inhibitors for solid tumors in the past 20 years (2001–2020).
      ]. Indeed, many small molecule TKIs have been developed and investigated for anticancer activities in CRC.
      Figure thumbnail gr1
      Fig. 1Mechanism of action of TKIs. Transmembrane receptor tyrosine kinases (RTKs) play key roles in transmitting extracellular signals from growth factors (GF) and other molecules to important signalling pathways, including MAPK, PI3K/AKT and JAK/STAT, that can drive cellular survival, proliferation, migration, as well as other downstream effects. The prevalent contributions of these signalling cascades to colorectal cancer (CRC) has prompted extensive testing of small molecule tyrosine kinase inhibitors (TKIs), which target the cytoplasmic signalling domain of these RTKs, as potential anti-cancer therapies. Abbreviations:; MAPK – mitogen activated protein kinase; PI3K/AKT – phosphatidylinositol 3-kinase/ protein kinase B; JAK/STAT – Janus kinase/signal transducers and activators of transcription; ECM – extracellular matrix.
      Over 50 TKIs are approved by the U.S Food & Drug-Administration (FDA). Although most of these TKIs show promising effects when tested preclinically for CRC, the vast majority fail in the clinic [
      • Cohen P.
      • Cross D.
      • Janne P.A.
      Kinase drug discovery 20 years after imatinib: progress and future directions.
      ]. Multiple factors can be associated with clinical failure, including the lack of predictive complex preclinical models, inadequate knowledge of the pharmacokinetics and pharmacodynamics of the TKIs, impaired understanding of the tumour mutational background and tumour heterogeneity. Moreover, there is still a limited comprehension of the role of the tumour microenvironment (TME), which has emerged as a highly relevant factor in CRC metastasis as well as a therapeutic target and thereby widens the gap between the preclinical and clinical outcomes [
      • Tauriello D.V.F.
      • Batlle E.
      Targeting the Microenvironment in Advanced Colorectal Cancer.
      ].

      Methods and search strategy

      We prepared a list of TKIs that were tested in (m)CRC. This was done by including TKIs from the FDA approved list of small molecule protein kinase inhibitors. Serine/threonine kinases inhibitors, phosphoinositide 3-kinases and transferases were excluded from the list. Additionally, through literature search on PubMed, Web of science and clinicaltrials.gov using search terms: “colorectal cancer”, “metastatic colorectal cancer”, “TKIs”, and individual TKI names, a list of non-FDA approved TKIs was also established. Preclinical studies published in English were included up to January 2022. Registered, recruiting, recruited and finished clinical trials for m(CRC) were also included until this period.
      Further, studies for each TKI tested clinically as monotherapy or combination strategies were included. Next, we searched for the same TKIs tested as mono- and combination therapy preclinically (Supplementary Figure S1). Additionally, to determine the current state of preclinical studies, we searched for the most recently published study with each TKI.

      Lost in translation

      To gauge the magnitude of the translational failure of TKIs in CRC study design and outcomes of preclinical and clinical studies were compared.

      TKIs as monotherapy

      Fourteen TKIs that were tested as monotherapy preclinically were also tested in a clinical trial for mCRC (Table 1 & Supplementary Table S1). Of these, 13 demonstrated prominent anti-cancer effects in the preclinical setting and only 4 TKIs (apatinib, cabozantinib, cediranib, and lenvatinib) were deemed as promising in non-randomized phase I/II studies, two (fruquintinib and regorafenib) showed clinical benefit in a randomized phase III study. Only one (regorafenib) is FDA approved.
      Table 1Clinical outcomes of preclinically selected TKIs used as monotherapy. The TKIs in bold have shown encouraging results and or have been successfully translated to the clinic.
      TKITarget(s)Study ModelClinical Trial StudyClinical OutcomesReferences
      apatinibVEGFR2, c-Kit, s-SRC

      CRC cell lines

      Murine CRC cell lines

      Mouse model
      Phase II single arm, open label studyMedian PFS – 3.9 months

      Median OS – 7.9 months
      • Cai X.
      • et al.
      Therapeutic Potential of Apatinib Against Colorectal Cancer by Inhibiting VEGFR2-Mediated Angiogenesis and beta-Catenin Signaling.
      ,
      • Chen X.
      • et al.
      A Single-Arm, Phase II Study of Apatinib in Refractory Metastatic Colorectal Cancer.
      bosutinibBCR-ABL, Src, Lyn, Hck

      CRC cell lines

      Xenograft mouse models
      Phase I, prospective clinical trialORR (6 %)

      PR – 1

      CR – 0
      • Golas J.M.
      • et al.
      SKI-606, a Src/Abl inhibitor with in vivo activity in colon tumor xenograft models.
      ,
      • Daud A.I.
      • et al.
      Phase I study of bosutinib, a src/abl tyrosine kinase inhibitor, administered to patients with advanced solid tumors.
      cabozantinibMET, VEGFR1, 2 and 3, AXL, RET, ROS1, TYRO3, MER, KIT, TRKB, FLT-3, TIE-2

      CRC cell lines

      Xenograft mice models
      Phase II single-arm, two-stage studyPFS ≥ 12-week (34 % patients)PR – 1 patient

      (Best response)SD with a DCR at week 6

      (72.7 %)
      • Yang S.
      • et al.
      Cabozantinib induces PUMA-dependent apoptosis in colon cancer cells via AKT/GSK-3beta/NF-kappaB signaling pathway.
      ,
      • Scott A.J.
      • et al.
      A phase II study investigating cabozantinib in patients with refractory metastatic colorectal cancer (AGICC 17CRC01).
      cediranibVEGFR1, 2, 3, PDGFRs, FGFRsCRC cell lines

      Mouse model
      Phase I, multicentre, open-labelDCR (81 %) – 26/32 patients
      • Melsens E.
      • et al.
      The VEGFR Inhibitor Cediranib Improves the Efficacy of Fractionated Radiotherapy in a Colorectal Cancer Xenograft Model.
      ,
      • Yamamoto N.
      • et al.
      Phase I, dose escalation and pharmacokinetic study of cediranib (RECENTIN), a highly potent and selective VEGFR signaling inhibitor, in Japanese patients with advanced solid tumors.
      dasatinibBCR-ABL, SRC family (SRC, LCK, YES, FYN), c-KIT, EPHA2, PDGFRβCRC xenograft

      Mice model

      CRC cell lines
      Phase

      II multicentre

      trial
      No OR observed

      • Scott A.J.
      • et al.
      Evaluation of the efficacy of dasatinib, a Src/Abl inhibitor, in colorectal cancer cell lines and explant mouse model.
      ,
      • Sharma M.R.
      • et al.
      Dasatinib in previously treated metastatic colorectal cancer: a phase II trial of the University of Chicago Phase II Consortium.
      erlotinibEGFRPDXPhase II studyNo CR or PR
      • Townsley C.A.
      • et al.
      Phase II study of erlotinib (OSI-774) in patients with metastatic colorectal cancer.
      ,
      • Rivera M.
      • et al.
      Patient-derived xenograft (PDX) models of colorectal carcinoma (CRC) as a platform for chemosensitivity and biomarker analysis in personalized medicine.
      fruquintinibVEGFR1,2,3

      Mouse modelPhase III randomized, double-blind, placebo-controlled, multicentre, clinical trialMedian OS fruquintinib vs placebo – 9.3 vs 6.6 months

      Median PFS fruquintinib vs placebo – 3.7 vs 1.8 months
      • Wang Y.
      • et al.
      Combination of Fruquintinib and Anti-PD-1 for the Treatment of Colorectal Cancer.
      ,
      • Li J.
      • et al.
      Effect of Fruquintinib vs Placebo on Overall Survival in Patients With Previously Treated Metastatic Colorectal Cancer The FRESCO Randomized Clinical Trial.
      gefitinibEGFR exon 19 deletion or exon 21 point mutation L858R, IGF and PDGF-mediated signallingCRC cell linesPhase II Randomized Trial

      PR in 1 of 110 patients (max 2.3 months)

      Median PFS – 1.9 months
      • Rothenberg M.L.
      • et al.
      Randomized phase II trial of the clinical and biological effects of two dose levels of gefitinib in patients with recurrent colorectal adenocarcinoma.
      ,
      • Georgiou A.
      • et al.
      Inactivation of NF1 Promotes Resistance to EGFR Inhibition in KRAS/NRAS/BRAF(V600) -Wild-Type Colorectal Cancer.
      lenvatinibVEGFR1, 2, & 3, PDGFRα, FGFR, KIT, RET

      CRC cell lines

      PDX
      Phase II open-label, single centre, single-arm, studyDCR (70.0 %)

      Median PFS –3.6 months

      Median OS – 7.4 months
      • Wiegering A.
      • et al.
      E7080 (lenvatinib), a multi-targeted tyrosine kinase inhibitor, demonstrates antitumor activities against colorectal cancer xenografts.
      ,

      Lim JH et al. Patient-Derived, Drug-Resistant Colon Cancer Cells Evade Chemotherapeutic Drug Effects via the Induction of Epithelial-Mesenchymal Transition-Mediated Angiogenesis. Int J Mol Sci 2020;21(20).

      ,
      • Iwasa S.
      • et al.
      Phase II study of lenvatinib for metastatic colorectal cancer refractory to standard chemotherapy: the LEMON study (NCCH1503).
      linifanibVEGF, PDGF, FLT3

      CRC cell lines

      3D micro-Tumours

      Phase II non-randomized,

      investigator-initiated open-label study
      Zero tumour responses observed

      Primary endpoint of ORR not met
      • Sobrino A.
      • et al.
      3D microtumors in vitro supported by perfused vascular networks.
      ,
      • Chan E.
      • et al.
      Phase II study of the Multikinase inhibitor of angiogenesis, Linifanib, in patients with metastatic and refractory colorectal cancer expressing mutated KRAS.
      nintedanibPDGFR α and β, FGFR 1–3, VEGFR 1,2,3, CSF1R, FLT-3

      CRC cell linesPhase III randomized, double-blind, placebo-controlled trialThe study failed to meet both co-primary end points

      No improvement in OS

      Significant but modest improvement in PFS
      • Van Cutsem E.
      • et al.
      Nintedanib for the treatment of patients with refractory metastatic colorectal cancer (LUME-Colon 1): a phase III, international, randomized, placebo-controlled study.
      ,
      • Cheng G.
      • et al.
      The microRNA-429/DUSP4 axis regulates the sensitivity of colorectal cancer cells to nintedanib.
      regorafenibRET, VEGFR1, 2, 3, KIT, PDGFRα and β, FGFR1, FGFR2, TIE2, DDR2, Trk2A, Eph2A, RAF-1, BRAF, BRAFV600E, SAPK2, PTK5, AblCRC PDTOsPhase III randomized, placebo-controlled, studyMedian OS regorafenib vs placebo – 6.4 vs 5.0 months
      • Vlachogiannis G.
      • et al.
      Patient-derived organoids model treatment response of metastatic gastrointestinal cancers.
      ,
      • Grothey A.
      • et al.
      Regorafenib monotherapy for previously treated metastatic colorectal cancer (CORRECT): an international, multicentre, randomised, placebo-controlled, phase 3 trial.
      sunitinibPDGFRα and β,VEGFR1,2, 3,KIT, FLT3, CSF-1R, RETCRC Cell lines

      Xenograft mouse model
      Phase II open-label, multicentre, two-stage, clinical trialNo OR observed

      • Lu Z.
      • et al.
      Regulation of intercellular biomolecule transfer-driven tumor angiogenesis and responses to anticancer therapies.
      ,

      Saltz, Phase II trial of sunitinib in patients with metastatic colorectal cancer after failure of standard therapy (vol 25, pg 4793, 2007). J Clin Oncol 2008;26(3):514-514.

      vandetanibEGFR and VEGFR families, RET, BRK, TIE2, and members of the EPH receptor and Src kinase familiesCRC Cell lines

      PDCs
      Phase I open Label, randomized studyNo OR observed
      • Kim S.Y.
      • et al.
      NCOA4-RET fusion in colorectal cancer: Therapeutic challenge using patient-derived tumor cell lines.
      ,
      • Mross K.
      • et al.
      DCE-MRI assessment of the effect of vandetanib on tumor vasculature in patients with advanced colorectal cancer and liver metastases: a randomized phase I study.
      Abbreviations: CRC – colorectal cancer; mCRC - metastatic colorectal cancer; PR - partial response; SD - stable disease; CR - complete response; OR - overall response; ORR – objective response rate; DCR – disease control rate; PDTOs – patient-derived tumour organoids; PDCs – patient-derived cells; OS – overall survival; PFS – progression-free survival; PDX – patient-derived xenograft.
      Apatinib, a relatively selective inhibitor of VEGFR2, c-Kit and s-SRC, showed promising pro-apoptotic effect in vitro when tested in both human and mice CRC cell lines. It also strongly suppressed CT26 (a mouse cell line) growth in mouse xenograft models [
      • Cai X.
      • et al.
      Therapeutic Potential of Apatinib Against Colorectal Cancer by Inhibiting VEGFR2-Mediated Angiogenesis and beta-Catenin Signaling.
      ]. Similarly, in a phase II clinical study, apatinib monotherapy showed promising efficacy (3.9 months progression-free survival (PFS)) for patients with refractory CRC without liver metastases [
      • Chen X.
      • et al.
      A Single-Arm, Phase II Study of Apatinib in Refractory Metastatic Colorectal Cancer.
      ]. Further, fruquintinib, a selective VEGFR1,-2,-3 inhibitor, demonstrated slower tumour growth when tested in mouse models preclinically [
      • Wang Y.
      • et al.
      Combination of Fruquintinib and Anti-PD-1 for the Treatment of Colorectal Cancer.
      ]. Clinical efficacy was demonstrated in a randomized phase III trial where fruquintinib displayed a significant increase in OS in a Chinese population [
      • Li J.
      • et al.
      Effect of Fruquintinib vs Placebo on Overall Survival in Patients With Previously Treated Metastatic Colorectal Cancer The FRESCO Randomized Clinical Trial.
      ] (Table 1). Currently, fruquintinib has been approved in China for treating patients with mCRC who have previously failed at least two systemic anti-neoplastic therapies [
      • Shirley M.
      Fruquintinib: First Global Approval.
      ].Recently, it has also been granted FDA fast track designation for patients with mCRC and is likely to become the second TKI to be approved for mCRC.
      Multi-kinase inhibitors such as cediranib and cabozantinib have been tested in vitro on CRC cell lines and in vivo in xenograft models in immunocompromised mice [
      • Yang S.
      • et al.
      Cabozantinib induces PUMA-dependent apoptosis in colon cancer cells via AKT/GSK-3beta/NF-kappaB signaling pathway.
      ,
      • Melsens E.
      • et al.
      The VEGFR Inhibitor Cediranib Improves the Efficacy of Fractionated Radiotherapy in a Colorectal Cancer Xenograft Model.
      ]. Both these TKIs induced anti-proliferative activities and apoptosis in vitro and suppressed tumour growth in vivo. Despite encouraging activity of cediranib in a phase I clinical study, it was not further investigated for patients with CRC [
      • Yamamoto N.
      • et al.
      Phase I, dose escalation and pharmacokinetic study of cediranib (RECENTIN), a highly potent and selective VEGFR signaling inhibitor, in Japanese patients with advanced solid tumors.
      ]. Cabozantinib treatment was found to be safe and demonstrated encouraging efficacy as 34 % of the heavily pre-treated patients with CRC achieved ≥ 12 weeks PFS in an ongoing phase II clinical study [
      • Scott A.J.
      • et al.
      A phase II study investigating cabozantinib in patients with refractory metastatic colorectal cancer (AGICC 17CRC01).
      ]. Another multi-kinase inhibitor, lenvatinib did not demonstrate cytotoxic but rather cytostatic effect in vitro and gave variable results in vivo [
      • Wiegering A.
      • et al.
      E7080 (lenvatinib), a multi-targeted tyrosine kinase inhibitor, demonstrates antitumor activities against colorectal cancer xenografts.
      ,

      Lim JH et al. Patient-Derived, Drug-Resistant Colon Cancer Cells Evade Chemotherapeutic Drug Effects via the Induction of Epithelial-Mesenchymal Transition-Mediated Angiogenesis. Int J Mol Sci 2020;21(20).

      ]. In a phase II clinical trial, lenvatinib showed promising clinical activity with 70 % disease control rate (DCR) [
      • Iwasa S.
      • et al.
      Phase II study of lenvatinib for metastatic colorectal cancer refractory to standard chemotherapy: the LEMON study (NCCH1503).
      ]. Regorafenib has been tested in the above-mentioned preclinical models and also in patient-derived tumour organoids (PDTOs) [
      • Liu Y.C.
      • et al.
      Regorafenib suppresses epidermal growth factor receptor signaling-modulated progression of colorectal cancer.
      ,
      • Vlachogiannis G.
      • et al.
      Patient-derived organoids model treatment response of metastatic gastrointestinal cancers.
      ,
      • Song X.
      • et al.
      Mcl-1 inhibition overcomes intrinsic and acquired regorafenib resistance in colorectal cancer.
      ]. It induced anti-proliferative activities and apoptosis in vitro and suppressed tumour growth in vivo. Regorafenib is the first and only small molecule multi-kinase inhibitor that is FDA approved as monotherapy for treatment of mCRC. This approval is based on the results of the CORRECT trial - a randomized phase III study in patients with mCRC who progressed on all standard therapies [
      • Grothey A.
      • et al.
      Regorafenib monotherapy for previously treated metastatic colorectal cancer (CORRECT): an international, multicentre, randomised, placebo-controlled, phase 3 trial.
      ] (Table 1). Another randomized phase III (CONCUR) trial supported these encouraging results [
      • Li J.
      • et al.
      Regorafenib plus best supportive care versus placebo plus best supportive care in Asian patients with previously treated metastatic colorectal cancer (CONCUR): a randomised, double-blind, placebo-controlled, phase 3 trial.
      ].
      The other 8 TKIs, that showed promising antitumour activity preclinically, failed in the clinic. For example, the multi-kinase inhibitor sunitinib, when tested as monotherapy on CRC cell lines and in mice, resulted in strong anti-tumour activities [
      • Lu Z.
      • et al.
      Regulation of intercellular biomolecule transfer-driven tumor angiogenesis and responses to anticancer therapies.
      ]. However, in a phase II clinical trial it did not demonstrate a meaningful single-agent objective response rate (ORR) in patients with mCRC refractory to standard chemotherapy [

      Saltz, Phase II trial of sunitinib in patients with metastatic colorectal cancer after failure of standard therapy (vol 25, pg 4793, 2007). J Clin Oncol 2008;26(3):514-514.

      ]. Studies performed by Golas J.M. et al. demonstrated that bosutinib is effective both in vitro (CRC cell lines) and in vivo (mouse xenograft model) but in the subsequent phase I clinical trial, this drug had limited efficacy [
      • Golas J.M.
      • et al.
      SKI-606, a Src/Abl inhibitor with in vivo activity in colon tumor xenograft models.
      ,
      • Daud A.I.
      • et al.
      Phase I study of bosutinib, a src/abl tyrosine kinase inhibitor, administered to patients with advanced solid tumors.
      ]. Similarly, nintedanib showed efficacy in vitro (CRC cell lines) but failed to improve OS in a phase III clinical setting [
      • Van Cutsem E.
      • et al.
      Nintedanib for the treatment of patients with refractory metastatic colorectal cancer (LUME-Colon 1): a phase III, international, randomized, placebo-controlled study.
      ,
      • Cheng G.
      • et al.
      The microRNA-429/DUSP4 axis regulates the sensitivity of colorectal cancer cells to nintedanib.
      ]. Likewise, many other TKIs including linifanib (CRC cell lines, 3D microtumour), dasatinib (CRC cell lines and xenograft models), erlotinib (patient-derived xenografts), gefitinib (CRC cell lines) and vandetanib (patient-derived cells) showed promising anti-tumour activities preclinically. However, in the clinic, none of these drugs elicited a beneficial response when administered as monotherapy (Table 1).

      TKIs in combination therapies

      TKIs have been considered in therapeutic combinations for CRC. We found 17 clinical trials studying 17 different TKIs in combination strategies that were considered successful in the preclinical setting (Table 2 & Supplementary Table S1). Of these, 11 (69 %) combinations with either chemotherapy, radiotherapy, or both, did not show promising results in the clinic. As an important caveat, several of these clinical trials used different drugs with similar targets compared to the corresponding preclinical studies, i.e. different chemotherapy drugs, or antibodies with equivalent targets (Table 2). For example, a synergistic effect was observed preclinically when CRC cells were subjected to a combination of dasatinib and oxaliplatin [
      • Kopetz S.
      • et al.
      Synergistic Activity of the Src Family Kinase Inhibitor Dasatinib and Oxaliplatin in Colon Carcinoma Cells Is Mediated by Oxidative Stress.
      ]. Correspondingly, in a phase I clinical study, the combination of dasatinib with capecitabine and oxaliplatin was used. For this case, the clinical combination also showed promising efficacy [
      • Strickler J.H.
      • et al.
      Phase I study of dasatinib in combination with capecitabine, oxaliplatin and bevacizumab followed by an expanded cohort in previously untreated metastatic colorectal cancer.
      ]. Further encouraging results were reported for the preclinical testing of cediranib plus radiotherapy in CRC xenograft models and these findings are in line with the phase I clinical trial combining cediranib plus chemo-radiotherapy [
      • Melsens E.
      • et al.
      The VEGFR Inhibitor Cediranib Improves the Efficacy of Fractionated Radiotherapy in a Colorectal Cancer Xenograft Model.
      ,
      • Marti F.E.M.
      • et al.
      Novel phase I trial design to evaluate the addition of cediranib or selumetinib to preoperative chemoradiotherapy for locally advanced rectal cancer: the DREAMtherapy trial.
      ].
      Table 2Clinical outcomes of preclinically selected TKIs used as combination therapies. The TKIs highlighted in bold have shown encouraging results in the clinic.
      TKITarget(s)ModelPreclinical CombinationClinical CombinationClinical Trial DesignClinical OutcomesReferences

      apatinibVEGFR2, c-Kit, s-SRC

      Murine CRC cell line & mouse modelApatinib + anti-PD1 antibodyAnti-PD-1 antibody SHR-1210 + apatinib

      Phase II prospective, single-arm, open-labelORR (0 %)DCR

      (22.2 %)

      The median PFS – 1.83 months

      The median OS – 7.80 months
      • Cai X.
      • et al.
      Apatinib enhanced anti-PD-1 therapy for colon cancer in mice via promoting PD-L1 expression.
      ,
      • Ren C.
      • et al.
      Anti-PD-1 antibody SHR-1210 plus apatinib for metastatic colorectal cancer: a prospective, single-arm, open-label, phase II trial.
      bosutinibBCR-ABL, Src, Lyn, Hck

      CRC Cell linesbosutinib + 5-FUbosutinib + capecitabinePhase I dose-escalation, multicentre, open-labelPR or SD greater than 24 weeks (all tumour types) – (13 %)

      • Wenzel T.
      • et al.
      Restoration of MARCK enhances chemosensitivity in cancer.
      ,
      • Isakoff S.J.
      • et al.
      Bosutinib plus capecitabine for selected advanced solid tumours: results of a phase 1 dose-escalation study.
      cabozantinibMET, VEGFR1, 2 and 3, AXL, RET, ROS1, TYRO3, MER, KIT, TRKB, FLT-3, TIE-2CRC Cell lines & 3D type I collagencabozantinib + Cetuximabcabozantinib + panitumumabPhase Ib clinical trialORR (16 %)

      Median PFS – 3.7 months.

      Median OS – 12.1 months.
      • Graves-Deal R.
      • et al.
      Broad-spectrum receptor tyrosine kinase inhibitors overcome de novo and acquired modes of resistance to EGFR-targeted therapies in colorectal cancer.
      ,
      • Strickler J.H.
      • et al.
      Cabozantinib and Panitumumab for RAS Wild-Type Metastatic Colorectal Cancer.
      cediranibVEGFR1, 2, 3, PDGFRs, FGFRsXenograft mouse modelcediranib + radiotherapy

      cediranib + chemoradiotherapyPhase I,

      alternating cohort design
      CR (41 %)ECPR

      (53 %)
      • Melsens E.
      • et al.
      The VEGFR Inhibitor Cediranib Improves the Efficacy of Fractionated Radiotherapy in a Colorectal Cancer Xenograft Model.
      ,
      • Marti F.E.M.
      • et al.
      Novel phase I trial design to evaluate the addition of cediranib or selumetinib to preoperative chemoradiotherapy for locally advanced rectal cancer: the DREAMtherapy trial.
      dasatinibBCR-ABL, SRC family (SRC, LCK, YES, FYN), c-KIT, EPHA2, PDGFRβCRC Cell lines

      dasatinib + oxaliplatin

      combination of dasatinib, capecitabine, oxaliplatin, and bevacizumabPhase I dose escalation and cohort expansion studyORR (75 %) – patients with high srcact expressionORR

      (0 %) patients with low srcact expression
      • Kopetz S.
      • et al.
      Synergistic Activity of the Src Family Kinase Inhibitor Dasatinib and Oxaliplatin in Colon Carcinoma Cells Is Mediated by Oxidative Stress.
      ,
      • Strickler J.H.
      • et al.
      Phase I study of dasatinib in combination with capecitabine, oxaliplatin and bevacizumab followed by an expanded cohort in previously untreated metastatic colorectal cancer.
      erlotinibEGFRCRC cell lineserlotinib + cetuximaberlotinib + cetuximabPhase II studyRR (41 %) in KRAS wt tumours & (52 %) in KRAS/BRAF wt tumourNo responses in patients with KRAS and BRAF mutations
      • Weickhardt A.J.
      • et al.
      Dual Targeting of the Epidermal Growth Factor Receptor Using the Combination of Cetuximab and Erlotinib: Preclinical Evaluation and Results of the Phase II DUX Study in Chemotherapy-Refractory, Advanced Colorectal Cancer.
      gefitinibEGFR exon 19 deletion or exon 21 point mutation L858R, IGF and PDGF-mediated signallingCRC cell linesgefitinib + chemotherapy + radiotherapygefitinib + FOLFIRIPhase II randomized multicenter trial

      OR, PFS and OS did not improve as compared to the FOLFIRI arm
      • Palumbo I.
      • et al.
      Gefitinib enhances the effects of combined radiotherapy and 5-fluorouracil in a colorectal cancer cell line.
      ,
      • Santoro A.
      • et al.
      A phase II randomized multicenter trial of gefitinib plus FOLFIRI and FOLFIRI alone in patients with metastatic colorectal cancer.
      imatinibBcr-Abl, PDGF, CSF/c-kit

      Xenograft mouse modelimatinib + bevacizumab

      XELOX + bevacizumab and imatinibPhase I/II prospective, non-randomized, open-label6-month PFS rate (76 %)

      Median PFS – 10.6 months

      Median OS – 23.2 months
      • Schiffmann L.M.
      • et al.
      A combination of low-dose bevacizumab and imatinib enhances vascular normalisation without inducing extracellular matrix deposition.
      ,
      • Hoehler T.
      • et al.
      Phase I/II trial of capecitabine and oxaliplatin in combination with bevacizumab and imatinib in patients with metastatic colorectal cancer: AIO KRK 0205.
      lenvatinibVEGFR1, 2, 3, PDGFRα, FGFR, KIT, RETCT26 murine cell line isograftsLenvatinib + anti-PD-1 antibodyLenvatinib + pembrolizumabPhase II, multicohort studyORR (22 %)
      • Kato Y.
      • et al.
      Lenvatinib plus anti-PD-1 antibody combination treatment activates CD8(+) T cells through reduction of tumor-associated macrophage and activation of the interferon pathway.
      ,
      • Gomez-Roca C.
      • et al.
      LEAP-005: A phase II multicohort study of lenvatinib plus pembrolizumab in patients with previously treated selected solid tumors-Results from the colorectal cancer cohort.
      nintedanibPDGFR α and β, FGFR 1–3, VEGFR 1,2,3, CSF1R, FLT-3CRC Cell lines & mouse modelnintedanib + TFTD/TAS102nintedanib + mFOLFOX6Phase II randomized, placebo-controlledPrimary end point criteria were not met
      • Boland P.M.
      • et al.
      A phase I/II study of nintedanib and capecitabine in refractory metastatic colorectal cancer.
      ,
      • Suzuki N.
      • et al.
      Effect of a novel oral chemotherapeutic agent containing a combination of trifluridine, tipiracil and the novel triple angiokinase inhibitor nintedanib, on human colorectal cancer xenografts.
      pazopanibVEGFR1,2,3,FGFR 1 and 3, KIT, interleukin-2, ltk, Lck, and c-FmsPDX Mouse Modelpazopanib + TEM +



      FOLFOX
      pazopanib + FOLFOX6 or CapeOxPhase I, Open-label, 2-part, dose-findingRR (40 %) pazopanib + FOLFOX6RR

      (38 %) pazopanib + CapeOx
      • Zhu G.
      • et al.
      Temozolomide and Pazopanib Combined with FOLFOX Regressed a Primary Colorectal Cancer in a Patient-derived Orthotopic Xenograft Mouse Model.
      ,
      • Brady J.
      • et al.
      An open-label study of the safety and tolerability of pazopanib in combination with FOLFOX6 or CapeOx in patients with colorectal cancer.
      regorafenibRET, VEGFR1,2,3, KIT, PDGFR α and β, FGFR1, FGFR2, TIE2, DDR2, Trk2A, Eph2A, RAF-1, BRAF, BRAFV600E, SAPK2, PTK5, AblMurine CRC cell linesregorafenib + anti-PD-1 antibodyregorafenib + nivolumab (anti-PD-1 antibody)Phase I, dose escalationOTR (36 %) patients with CRC

      Median PFS −7.9 months

      • Doleschel D.
      • et al.
      Regorafenib enhances anti-PD1 immunotherapy efficacy in murine colorectal cancers and their combination prevents tumor regrowth.
      ,
      • Fukuoka S.
      • et al.
      Regorafenib Plus Nivolumab in Patients With Advanced Gastric or Colorectal Cancer: An Open-Label, Dose-Escalation, and Dose-Expansion Phase Ib Trial (REGONIVO, EPOC1603).
      semaxinibVEGFR2, c-kit

      CRC Cell lines & xenograft mouse modelSemaxinib + SN-38semaxinib + bolus 5-FU, leucovorin, and irinotecan (IFL)Phase I/II trialConfirmed PR (27 %)Unconfirmed PR

      (18 %)SD

      (36 %)
      • Bocci G.
      • et al.
      Antiangiogenic and anticolorectal cancer effects of metronomic irinotecan chemotherapy alone and in combination with semaxinib.
      ,
      • Lockhart A.C.
      • et al.
      Phase I/pilot study of SU5416 (semaxinib) in combination with irinotecan/bolus 5-FU/LV (IFL) in patients with metastatic colorectal cancer.
      sorafenibc-CRAF, BRAF, mBRAF, KIT, FLT- 3, RET, RET/PTC, VEGFR1,2,3,PDGFR-ßCRC cell linessorafenib + cetuximabsorafenib + cetuximabPhase II open-label, single-armMedian PFS-1.84 months

      No OR observed
      • Do K.
      • et al.
      A Phase II Study of Sorafenib Combined With Cetuximab in EGFR-Expressing, KRAS-Mutated Metastatic Colorectal Cancer.
      ,
      • Martinelli E.
      • et al.
      Synergistic antitumor activity of sorafenib in combination with epidermal growth factor receptor inhibitors in colorectal and lung cancer cells.
      sunitinibPDGFRα and β,VEGFR1,2,3,KIT, FLT3, CSF-1R, RETCRC Cell lines & mouse modelsunitinib + irinotecansunitinib + FOLFIRIPhase II Multicenter, open labelPR − 8/17 patients
      • Mross K.
      • et al.
      FOLFIRI and sunitinib as first-line treatment in metastatic colorectal cancer patients with liver metastases - a CESAR phase II study including pharmacokinetic, biomarker, and imaging data.
      ,
      • Di Desidero T.
      • et al.
      Chemotherapeutic and antiangiogenic drugs beyond tumor progression in colon cancer: Evaluation of the effects of switched schedules and related pharmacodynamics.
      vandetanibEGFR and VEGFR families, RET, BRK, TIE2, and members of the EPH receptor and Src kinase familiesXenograft Modelvandetanib + radiotherapy and irinotecanvandetanib + irinotecan + cetuximabPhase I with an expanded MTD cohortPR (7 %), SD (59 %), PD (34 %)

      Median PFS- 3.6 months

      Median OS −10.5 months
      • Wachsberger P.
      • et al.
      Combination of vandetanib, radiotherapy, and irinotecan in the LoVo human colorectal cancer xenograft model.
      ,
      • Meyerhardt J.A.
      • et al.
      Phase I Study of Cetuximab, Irinotecan, and Vandetanib (ZD6474) as Therapy for Patients with Previously Treated Metastastic Colorectal Cancer.
      vatalanibVEGFR1,2, PDGFR, c-Kit, C-Fms

      CRC cell linesVatalanib + SN-38vatalanib + FOLFOX4Phase III randomized, placebo-controlledPFS, OS, and ORR were not improved
      • To K.K.W.
      • et al.
      Vatalanib sensitizes ABCB1 and ABCG2-overexpressing multidrug resistant colon cancer cells to chemotherapy under hypoxia.
      ,
      • Hecht J.R.
      • et al.
      Randomized, Placebo-Controlled, Phase III Study of First-Line Oxaliplatin-Based Chemotherapy Plus PTK787/ZK 222584, an Oral Vascular Endothelial Growth Factor Receptor Inhibitor, in Patients With Metastatic Colorectal Adenocarcinoma.
      Abbreviations: CRC – colorectal cancer; mCRC- metastatic colorectal cancer; PDX - patient derived xenograft; MTD – maximum tolerated dose; RR - response rate; PR - partial response/partial remission; SD - stable disease; PD - progressive disease; CR - complete response; ECPR - excellent clinical or pathological response; OR - overall response; OTR - objective tumour response; ORR – objective response rate; DCR – disease control rate; OS – overall survival; PFS - progression free survival; FOLFOX4/6/mFOLFOX6 - leucovorin calcium (folinic acid), fluorouracil, and oxaliplatin; SN-38 – active metabolite of irinotecan; 5-FU − 5-Fluorouracil; CapeOx/XELOX – capecitabine, oxaliplatin; TFTD/TAS102 - trifluridine/tipiracil; FOLFIRI - leucovorin calcium (calcium folinate), 5-fluorouracil, and irinotecan.
      A second strategy can be considered by combining TKIs with antibodies targeting the epidermal growth factor receptor (EGFR)/mitogen-activated protein kinase (MAPK) pathway. Graves-Deal et al. tested the combination of cabozantinib with cetuximab, an anti-EGFR antibody, in CRC cell lines cultured in 3D in collagen and found that cabozantinib helped to overcome cetuximab resistance (Table 2) [
      • Graves-Deal R.
      • et al.
      Broad-spectrum receptor tyrosine kinase inhibitors overcome de novo and acquired modes of resistance to EGFR-targeted therapies in colorectal cancer.
      ]. Subsequently in a phase I study, the combination of cabozantinib with panitumumab, another anti-EGFR antibody, had limited efficacy with severe toxicity leading to significant dose reductions and 20 % treatment discontinuation [
      • Strickler J.H.
      • et al.
      Cabozantinib and Panitumumab for RAS Wild-Type Metastatic Colorectal Cancer.
      ]. Further, TKIs are also combined with the anti-VEGF antibody bevacizumab. A superior level of vascular normalization without the induction of extracellular matrix (ECM) deposition was observed when bevacizumab was administered in combination with imatinib in CRC xenograft tumours [
      • Schiffmann L.M.
      • et al.
      A combination of low-dose bevacizumab and imatinib enhances vascular normalisation without inducing extracellular matrix deposition.
      ]. Next, a phase I/II clinical trial showed that the combination of XELOX (capecitabine plus oxaliplatin) with bevacizumab and imatinib was tolerable and demonstrated promising efficacy (Table 2) [
      • Hoehler T.
      • et al.
      Phase I/II trial of capecitabine and oxaliplatin in combination with bevacizumab and imatinib in patients with metastatic colorectal cancer: AIO KRK 0205.
      ].
      An additional emerging approach is the combination of TKIs with immunotherapy. The past decade showed an increased interest for therapies that boost the host’s immune response against cancer, especially due to the promising but as yet limited efficacy of immune checkpoint inhibitors (ICIs) like anti-PD-1 antibody in CRC [
      • Janssen E.
      • et al.
      Combinatorial Immunotherapies for Metastatic Colorectal Cancer.
      ]. In vivo tumour growth of CT-26 isografts was significantly inhibited when treated with a combination of lenvatinib plus an anti-PD-1 antibody [
      • Kato Y.
      • et al.
      Lenvatinib plus anti-PD-1 antibody combination treatment activates CD8(+) T cells through reduction of tumor-associated macrophage and activation of the interferon pathway.
      ]. A phase II clinical trial demonstrated similar promising efficacy of the combination of lenvatinib plus pembrolizumab in patients with advanced non-MSI-H mCRC [
      • Gomez-Roca C.
      • et al.
      LEAP-005: A phase II multicohort study of lenvatinib plus pembrolizumab in patients with previously treated selected solid tumors-Results from the colorectal cancer cohort.
      ]. Based on these results, a phase III trial is currently ongoing to compare the efficacy and safety of this combination versus standard of care in patients with non-MSI-H mCRC (NCT04776148). In another case, a synergistic immunomodulatory effect was observed when regorafenib was combined with anti-PD-1 antibody in CRC cells. The combination strongly inhibited CRC regrowth preclinically (Table 2) [
      • Doleschel D.
      • et al.
      Regorafenib enhances anti-PD1 immunotherapy efficacy in murine colorectal cancers and their combination prevents tumor regrowth.
      ]. Likewise, encouraging anti-tumour activities were observed in a phase I dose escalation clinical trial (REGONIVO) that tested the combination of regorafenib with nivolumab (anti-PD-1 antibody) [
      • Fukuoka S.
      • et al.
      Regorafenib Plus Nivolumab in Patients With Advanced Gastric or Colorectal Cancer: An Open-Label, Dose-Escalation, and Dose-Expansion Phase Ib Trial (REGONIVO, EPOC1603).
      ]. However, subsequent studies investigating combinations of regorafenib with ICIs indicated limited clinical benefit [
      • Wang C.
      • et al.
      Regorafenib and Nivolumab or Pembrolizumab Combination and Circulating Tumor DNA Response Assessment in Refractory Microsatellite Stable Colorectal Cancer.
      ,
      • Cousin S.
      • et al.
      Regorafenib-Avelumab Combination in Patients with Microsatellite Stable Colorectal Cancer (REGOMUNE): A Single-arm, Open-label. Phase II Trial.
      ,
      • Kim R.D.
      • et al.
      A phase I/Ib study of regorafenib and nivolumab in mismatch repair proficient advanced refractory colorectal cancer.
      ,
      • Yang K.
      • et al.
      Real-world outcomes of regorafenib combined with immune checkpoint inhibitors in patients with advanced or metastatic microsatellite stable colorectal cancer: A multicenter study.
      ]. No randomized data have been generated so far that prove the added value of the combination regimen over standard treatment.
      Similarly, whereas pazopanib combined with FOLFOX and temozolomide appeared promising in a patient-derived orthotopic xenograft mouse model, no additive clinical benefits were observed with the combination of pazopanib and FOLFO6 or CAPOX in a phase I clinical trial (Table 2) [
      • Zhu G.
      • et al.
      Temozolomide and Pazopanib Combined with FOLFOX Regressed a Primary Colorectal Cancer in a Patient-derived Orthotopic Xenograft Mouse Model.
      ,
      • Brady J.
      • et al.
      An open-label study of the safety and tolerability of pazopanib in combination with FOLFOX6 or CapeOx in patients with colorectal cancer.
      ]. Also, the combination of vandetanib, irinotecan and radiotherapy caused strong tumour shrinkage in human colorectal xenograft models [
      • Wachsberger P.
      • et al.
      Combination of vandetanib, radiotherapy, and irinotecan in the LoVo human colorectal cancer xenograft model.
      ]; yet the combination of vandetanib with irinotecan and cetuximab did not show any increase in efficacy in comparison to historic data in previously treated mCRC patients [
      • Meyerhardt J.A.
      • et al.
      Phase I Study of Cetuximab, Irinotecan, and Vandetanib (ZD6474) as Therapy for Patients with Previously Treated Metastastic Colorectal Cancer.
      ]. Another issue arose when irinotecan or its active metabolite SN-38, combined with sunitinib, was translated to the clinic as sunitinib plus FOLFIRI. This led to severe adverse effects, while not offering an improved ORR over FOLFIRI alone [
      • Mross K.
      • et al.
      FOLFIRI and sunitinib as first-line treatment in metastatic colorectal cancer patients with liver metastases - a CESAR phase II study including pharmacokinetic, biomarker, and imaging data.
      ].
      More examples of unsuccessful translations of promising preclinical combinations arose from clinical trials—testing bosutinib, gefitinib, nintedanib, and vatalanib in combination with different forms of chemotherapies—that failed to meet their primary endpoints (Table 2). Furthermore, the combination of sorafenib with cetuximab failed to reach an objective response in the clinic [
      • Do K.
      • et al.
      A Phase II Study of Sorafenib Combined With Cetuximab in EGFR-Expressing, KRAS-Mutated Metastatic Colorectal Cancer.
      ], despite the fact that a synergistic effect was observed when tested on CRC cell lines [
      • Martinelli E.
      • et al.
      Synergistic antitumor activity of sorafenib in combination with epidermal growth factor receptor inhibitors in colorectal and lung cancer cells.
      ]. Also, when an anti-PD-1 antibody was combined with apatinib in mice, a significant inhibition in tumour growth of transplanted tumours was observed [
      • Cai X.
      • et al.
      Apatinib enhanced anti-PD-1 therapy for colon cancer in mice via promoting PD-L1 expression.
      ]. However, in the clinic, apatinib combined with anti-PD-1 antibody (SHR-1210) failed to improve the efficacy in patients with microsatellite stable (MSS) mCRC, which was probably due to intolerable toxicity [
      • Ren C.
      • et al.
      Anti-PD-1 antibody SHR-1210 plus apatinib for metastatic colorectal cancer: a prospective, single-arm, open-label, phase II trial.
      ].

      Lessons learned

      Taken together, the majority of favourable preclinical outcomes failed upon clinical translation. Out of the 42 TKIs from preclinical studies (Supplementary Table 1), 14 have been tested in the clinic as monotherapy for CRC and only a fraction showed any clinical response leading to only a single FDA approval (Table 1). Regarding TKI-combination strategies, less than one third was deemed promising in early clinical trials, although none have been approved yet. This can be attributed to the fact that first-line treatment of currently available therapies is initially very effective. Therefore, to improve efficacy, a TKI must be highly effective in combination with chemotherapy [
      • Santoro A.
      • et al.
      A phase II randomized multicenter trial of gefitinib plus FOLFIRI and FOLFIRI alone in patients with metastatic colorectal cancer.
      ]. The other reasons behind the low success rate of TKIs for the treatment of CRC are often unknown or not thoroughly investigated. Rather than simply discarding most TKIs for this cancer type—and accepting that the process of drug testing has to be this inefficient, we instead propose a renewed effort to identify and address the gaps and hurdles that currently impede successful translational research. In our view, this requires an integrated approach that adopts advanced preclinical models, pays more attention to pharmacology, and performs an in-depth assessment of mechanisms of action and resistance (Fig. 2).
      Figure thumbnail gr2
      Fig. 2Gaps in Clinical Translation. Multiple, interlinked factors contribute to poor clinical translation in CRC. Lack of better predictive preclinical models seems to be associated with all the other translational gaps. Appropriate dosing and scheduling can lead to improved efficacy of TKIs. The lack of clinical dose prediction in the preclinical models is a major barrier of translation to the clinic. Hence, when the dosing is inadequate it leads to lack of efficacy. Another challenge of treatment resistance also persists as we do not have an appropriate model in the preclinic to replicate the tumour with the TME, which could also help predict clinical responses to the TKIs.

      Complexity of preclinical models

      A wide selection of preclinical models—ranging from classical cell lines to higher-complexity 3D culture systems and animal studies—are being used to study the application of TKIs for CRC (Supplementary Fig. 2). Many studies utilize several models in parallel, starting from simple mouse cell lines and human CRC cell lines to cell lines as xenograft and mice models. Patient-derived xenografts (PDXs) and patient-derived cells have been grown and injected into mice to test the effect of different TKIs. This has been pushed further into 3D-model testing as spheroids and other 3D gel-free structures. Finally, some TKIs have also been tested preclinically in PDTO models.
      CRC cell lines have been used widely as they are cost- and time-efficient and well established. This is also evident from our search, as the most recent studies on 33 of the 42 TKIs involves this conventional modality (Supplementary Fig. 2). However, issues such as lack of heterogeneity and loss of genomic architecture following abundant passaging have eroded confidence in the predictability of CRC cell lines. Furthermore, these models are not representative of the in vivo environment due to the absence of interaction with human immune and stromal components of the TME [
      • Rizzo G.
      • et al.
      Patient-derived tumor models: a more suitable tool for pre-clinical studies in colorectal cancer.
      ,
      • Gillet J.P.
      • Varma S.
      • Gottesman M.M.
      The clinical relevance of cancer cell lines.
      ]. The use of in vivo mouse models helped to overcome some disadvantages experienced with cell lines, and facilitates the understanding of the complex molecular pathways and different cell types involved in tumour biology. However, the high costs of these models and short lifespan of mice reduce their applicability [
      • Rizzo G.
      • et al.
      Patient-derived tumor models: a more suitable tool for pre-clinical studies in colorectal cancer.
      ,
      • Burtin F.
      • Mullins C.S.
      • Linnebacher M.
      Mouse models of colorectal cancer: Past, present and future perspectives.
      ].
      The use of PDXs and PDTOs have brought us a step closer to the clinic as they both better represent individual tumours at genomic and transcriptomic levels [
      • Nunes M.
      • et al.
      Evaluating Patient-Derived Colorectal Cancer Xenografts as Preclinical Models by Comparison with Patient Clinical Data.
      ,
      • Bleijs M.
      • et al.
      Xenograft and organoid model systems in cancer research.
      ]. PDXs are attractive personalized preclinical models. However, the high cost and time required for large scale screening studies, the number of animals required, and the loss of human stromal cells are relevant considerations. PDTOs on the other hand are less time consuming and more cost effective than PDX models. To take into account interpatient heterogeneity, the use of several PDTOs is more feasible than using the same number of genetic mouse models, or in vivo implantation models. The drawback is that, currently, tumour organoids lack stromal and immune cells, as well as vasculature [
      • Bleijs M.
      • et al.
      Xenograft and organoid model systems in cancer research.
      ]. Despite the promise of PDTOs, it is important to realize that no prospective clinical proof of their relevance to predict clinical outcomes in drug development is available.
      Further analysing the applied types of preclinical models, we noted that the TKIs that were able to show any clinical response in mCRC were all tested in models of higher complexity than merely 2D cell lines. For example, apatinib, fruquintinib and imatinib were tested in xenograft mouse models; cediranib was tested in vivo in athymic mice and rats, along with cell lines; and regorafenib was tested on multiple preclinical models including PDTOs (Table 1, Table 2). In TKI combinations, cabozantinib with panitumumab was tested in a 3D model with type 1 collagen that predicted clinical response accurately [
      • Graves-Deal R.
      • et al.
      Broad-spectrum receptor tyrosine kinase inhibitors overcome de novo and acquired modes of resistance to EGFR-targeted therapies in colorectal cancer.
      ]. Interestingly, this might be attributed to the fact that an increased amount of collagen type 1 is present in the TME at increased stage of CRC [
      • Li Z.L.
      • et al.
      Changes in extracellular matrix in different stages of colorectal cancer and their effects on proliferation of cancer cells.
      ]. These results provide an indication of an added predictive value for the use of preclinical models with higher complexity. Though poor clinical translation of efficacy of TKIs in CRC is likely not only driven by the preclinical models used.

      Mechanisms of resistance

      A second challenge that limits predictability of preclinical research is a poor understanding of resistance mechanisms. Nearly 90 % of metastatic cancer treatment failure is attributed to multi-drug resistance [
      • Wu G.
      • et al.
      Overcoming treatment resistance in cancer: Current understanding and tactics.
      ]. This is also the case for mCRC, where resistance to currently available chemotherapeutic and targeted therapy is one of the major challenges in the clinic [
      • Van der Jeught K.
      • et al.
      Drug resistance and new therapies in colorectal cancer.
      ,

      Acquired and Intrinsic Resistance to Colorectal Cancer Treatment. 2018.

      ]. This problem is broadly grouped into 2 categories of intrinsic resistance and acquired resistance [
      • Holohan C.
      • et al.
      Cancer drug resistance: an evolving paradigm.
      ]. The former specifies the presence of pre-existing factors that cause the inefficacy of the treatment, whereas the latter type of resistance develops during treatment and allows an initially sensitive tumour to become recalcitrant [
      • Longley D.B.
      • Johnston P.G.
      Molecular mechanisms of drug resistance.
      ]. Since almost all of the TKIs fail in early phase clinical trials, it is especially interesting to understand the pre-existing factors that might cause intrinsic resistance.
      In addition, a better understanding of the intrinsic molecular mechanisms of resistance could improve patient selection, leading to a better translation of the TKIs to the clinic in CRC. A phase II study with the combination of erlotinib and cetuximab showed promising response rates in patients with chemotherapy-refractory advanced CRC. Yet, this was only limited to patients with KRAS/BRAF wild-type tumours. This type of limitation can be predicted when testing in CRC cell lines [
      • Weickhardt A.J.
      • et al.
      Dual Targeting of the Epidermal Growth Factor Receptor Using the Combination of Cetuximab and Erlotinib: Preclinical Evaluation and Results of the Phase II DUX Study in Chemotherapy-Refractory, Advanced Colorectal Cancer.
      ].
      Inherent tumour heterogeneity and the presence of drug resistant genetic clones within tumour cells could be a cause behind the intrinsic resistance in CRC [
      • Molinari C.
      • et al.
      Heterogeneity in Colorectal Cancer: A Challenge for Personalized Medicine?.
      ,
      • Frank M.H.
      • et al.
      Clinical Implications of Colorectal Cancer Stem Cells in the Age of Single-Cell Omics and Targeted Therapies.
      ]. Several TKIs are known to interact with targets within the cells of the TME [
      • Tan H.Y.
      • et al.
      Targeting tumour microenvironment by tyrosine kinase inhibitor.
      ]. This includes immune cells, ECM components, cancer associated fibroblasts (CAFs), and endothelial cells. Together, these TME cells have complex effects on tumour cell behaviour, progression and survival, which includes regulation of (treatment-resistant) cancer stem cells [
      • Frank M.H.
      • et al.
      Clinical Implications of Colorectal Cancer Stem Cells in the Age of Single-Cell Omics and Targeted Therapies.
      ]. Therefore, TKIs are increasingly perceived as a promising candidate for TME modulation [
      • Tan H.Y.
      • et al.
      Targeting tumour microenvironment by tyrosine kinase inhibitor.
      ]. However, due to the complex network of cellular crosstalk in the TME, it is difficult to predict the response to TKI treatment by using in vitro models that contain only cancer cells. Consequently, there is a dire need for more complex 3D models with biochemical and biomechanical properties that represent the whole tumour and its TME.
      Correspondingly, preclinical models of higher complexity that additionally recapitulate TME-specific predictive biomarkers of response would greatly help patient selection in the clinic.

      Appropriate dosing

      Administering TKIs at an effective dose without causing excess toxicity is another major hurdle to clear for smooth translation to the clinic. Currently, it is hypothesised that a consistent TKI plasma concentration enables continuous inhibition of its targeted pathway and thereby results in anti-tumour activity. Consequently, in the clinic, most of the TKIs are administered once or twice daily, continuously near their maximum tolerate dose (MTD) [
      • Gerritse S.L.
      • et al.
      High-dose administration of tyrosine kinase inhibitors to improve clinical benefit: A systematic review.
      ]. The drug concentration that is required to inhibit a biological process by 50 % is defined as the IC50 of the drug [
      • Aykul S.
      • Martinez-Hackert E.
      Determination of half-maximal inhibitory concentration using biosensor-based protein interaction analysis.
      ]. Among the preclinical studies, we found that only 14 out of 42 studies mentioned IC50 values of the TKIs tested. Moreover, none of the preclinical studies investigated intra-cellular TKI concentrations, implicitly assuming that the concentration in the medium equals the concentration in the cells. It is important to note that the actual intra-cellular concentration may differ substantially due to factors such as transporter affinity, dissociation coefficient and pKa value of the TKI [
      • Decosterd L.A.
      • et al.
      Therapeutic drug monitoring of targeted anticancer therapy.
      ]. Therefore, the IC50-values derived from the medium concentrations the model is exposed to, could be a gross under- or overestimation from required drug levels at the target site in in vivo models and patients. Indeed, the actual drug concentration that interacts with the target and causes anti-proliferative or other anti-tumour effects is mostly unexplored.
      Another factor to consider when translating preclinical studies to the clinic is that only the unbound fraction (e.g., unbound by proteins like albumin) of the TKIs can interact with the target to cause antitumour efficacy [
      • Smith D.A.
      • Di L.
      • Kerns E.H.
      The effect of plasma protein binding on in vivo efficacy: misconceptions in drug discovery.
      ]. In in vitro studies the total concentration (bound and unbound drugs) is used to determine the IC50 value, whereas the active unbound fraction might be very different between patients and ex vivo tumour models. Consequently, the TKI concentration required to induce a potent effect presents a challenge for translation into the clinic. The relevance of rational dosing can be addressed by comparing three clinical trials in which different sunitinib treatment schedules were tested in patients with mCRC. In the first, phase II clinical trial conducted by Saltz et al., the safety and efficacy of sunitinib was investigated in 84 patients with mCRC after failure of standard therapy. This study demonstrated no effect of sunitinib as a single agent, as no meaningful ORR was observed in patients with heavily pre-treated mCRC and only 16 % had a PFS of more than 22 weeks. The standard dose of 50 mg once daily for 4 consecutive weeks, followed by 2 weeks off treatment, in repeated 6 week cycles was administered in this study [

      Saltz, Phase II trial of sunitinib in patients with metastatic colorectal cancer after failure of standard therapy (vol 25, pg 4793, 2007). J Clin Oncol 2008;26(3):514-514.

      ]. A similar conclusion was drawn from a second study where Al Baghdadi et al. investigated the response to sunitinib in ten patients with mCRC with FLT-3 amplification. The dosing strategy was the same as in the study from Saltz et al. It was concluded that sunitinib as monotherapy did not show any clinical activity in mCRC patients with FLT-3 amplification [
      • Al Baghdadi T.
      • et al.
      Sunitinib in Patients with Metastatic Colorectal Cancer (mCRC) with FLT-3 Amplification: Results from the Targeted Agent and Profiling Utilization Registry (TAPUR) Study.
      ].
      However, in a third study using an alternative sunitinib dose (700 mg every 2 weeks), Rovithi et al. demonstrated that higher plasma concentrations can be reached, and that 20 % of the patients with mCRC had a PFS of more than 5 months with acceptable toxicity. In this study it was found that this regimen resulted in high intra-tumoral TKI concentrations (Gerritse et al. submitted). Consequently, the more potent anti-cancer effect could be due to the inhibition of unexpected cellular targets. This indicates the importance of not only selecting the right drug but also the right dose and schedule when translating preclinical findings in to the clinic [
      • Rovithi M.
      • et al.
      Phase I Dose-Escalation Study of Once Weekly or Once Every Two Weeks Administration of High-Dose Sunitinib in Patients With Refractory Solid Tumors.
      ]. It emphasizes that the adagio of continuous inhibition of its designated target of a TKI might be less relevant than achieving higher tumour concentrations to increase the efficacy by hitting more (off-target) kinases [
      • Rovithi M.
      • Verheul H.M.W.
      Pulsatile high-dose treatment with antiangiogenic tyrosine kinase inhibitors improves clinical antitumor activity.
      ].
      Hence, the (unbound) concentration and intracellular TKI concentrations at which antitumour activity is observed should be defined preclinically in complex in vitro models. Also, quantification of intra-tumoral TKI concentrations should be incorporated in phase I clinical trials to make translation of preclinical to clinical exposure possible and to determine the optimal dose for further research.

      PDTO-based 3D culture modalities

      The above-mentioned limitations of currently available in vitro models highlight the translational importance of reliable preclinical models to fully understand the mechanism of action or of resistance of TKIs, and to determine the appropriate dosing. Recent studies provide more evidence that the predictability of clinical success may increase when preclinical studies include 3D cultures, 3D CRC- stromal cell co-cultures, and PDTO models. Zoetemelk et al. studied the differing responses of CRC cell lines in 2D, 3D spheroids, and 3D co-cultures with fibroblasts and endothelial cells to regorafenib, erlotinib and 5-FU in varying combination strategies. Regorafenib had the same response in both 2D and 3D models, but the response to 5-FU and erlotinib was different between these models. It was reported that 3D and 3D co-culture systems were more robust, reproducible for monotherapy and combination efficacy studies [
      • Zoetemelk M.
      • et al.
      Short-term 3D culture systems of various complexity for treatment optimization of colorectal carcinoma.
      ].
      Furthermore, Ganesh and colleagues reported that rectal cancer PDTOs molecularly resemble rectal cancer and that these PDTOs are applicable for drug screening preclinically. Indeed, the PDTOs response to chemotherapy was associated with the clinical response in patient tumours. Additionally, these PDTOs were engrafted into murine rectal mucosa and displayed the heterogeneous sensitivity to clinically observed chemotherapy [
      • Ganesh K.
      • et al.
      A rectal cancer organoid platform to study individual responses to chemoradiation.
      ]. Another study investigated the predictability of PDTOs when treated with chemoradiation in advanced rectal cancer. It demonstrated a broad range of chemoradiation responses in rectal cancer organoids with some matching the patient responses [
      • Yao Y.
      • et al.
      Patient-Derived Organoids Predict Chemoradiation Responses of Locally Advanced Rectal Cancer.
      ]. However, these data do indicate a very large variation and are too inaccurate for clinical use to select therapy. It has also been reported that the treatment results from PDTOs are associated with response to irinotecan monotherapy in the clinic [
      • Ooft S.N.
      • et al.
      Patient-derived organoids can predict response to chemotherapy in metastatic colorectal cancer patients.
      ]. One caveat here is that the number of samples tested is low and therefore one should interpret these results with caution.
      There is evidence showing that PDTOs can predict the response to targeted therapies. A study performed by Young Kim et al., demonstrated the translational relevance of using PDTOs in advanced lung adenocarcinomas. The PDTOs were able to predict clinical responses of non-small cell lung cancer patients who received clinically approved targeted agents [
      • Kim S.Y.
      • et al.
      Modeling Clinical Responses to Targeted Therapies by Patient-Derived Organoids of Advanced Lung Adenocarcinoma.
      ]. A separate study exhibited the comparative responses to ruxolitinib treatment between a patient with Li Fraumeni syndrome and PDTOs generated from the patient’s tumour [
      • Reed M.R.
      • et al.
      A Functional Precision Medicine Pipeline Combines Comparative Transcriptomics and Tumor Organoid Modeling to Identify Bespoke Treatment Strategies for Glioblastoma.
      ]. Therefore, it seems that increasing the complexity of the models can bring us closer to predicting clinical responses [

      Reidy E, et al. A 3D View of Colorectal Cancer Models in Predicting Therapeutic Responses and Resistance. Cancers (Basel) 2021. 13(2).

      ].
      An interesting way to further increase the complexity of PDTO models is by including components of the TME, including fibroblasts, endothelial cells and immune cells such as macrophages. This may help understanding the role of these stromal cells on the effect of TKIs. For example, PDTOs co-cultured with stromal cells could further help to identify whether intrinsic resistance to TKIs is stimulated by the TME. Fabrizio Fontana et al. summarized various 3D co-culture studies and revealed a better prediction to drug response in 3D co-cultures as compared to 2D models [
      • Fontana F.
      • et al.
      In Vitro 3D Cultures to Model the Tumor Microenvironment.
      ]. This is also indicated by a study performed by Jin et al., in which the potency of anlotinib, a multi-kinase inhibitor, changes when gastric cancer PDTOs were co-cultured with CAFs [
      • Jin Z.J.
      • et al.
      The cross-talk between tumor cells and activated fibroblasts mediated by lactate/BDNF/TrkB signaling promotes acquired resistance to anlotinib in human gastric cancer.
      ]. The study provides insight in the interaction of the epithelium and the stroma leading to anlotinib resistance in gastric cancer. Hence, a 3D PDTO co-culture model with components of the TME, while still in early development, may potentially reproduce the cancer niche and the interactions between the tumour and its microenvironment. Such models can also facilitate the study of immunological effects of TKIs, which is critical to build a potential rationale for immunotherapeutic (combination) approaches.
      Taking above aspects of genetic heterogeneity, TME and appropriate dosing one step further towards adequately predicting clinical outcome, tissue engineering with microfluidic devices (tumour-on-a-chip models) may allow a drug delivering system in which vasculature and TME constituents are incorporated. Such a drug delivery system will add pharmacodynamics as an integrated parameter into preclinical modelling of a tumour-microenvironment-on-a-chip system by incorporating PDTOs [
      • Low L.A.
      • et al.
      Organs-on-chips: into the next decade.
      ,
      • Hofer M.
      • Lutolf M.P.
      Engineering organoids.
      ].

      Conclusion

      Dozens of TKIs have displayed significant anti-tumoral activity when preclinically tested for CRC. Unfortunately, encouraging preclinical results are rarely successfully translated into the clinic. Indeed, regorafenib is currently the only FDA-approved TKI for the treatment of mCRC. Multiple factors contribute to the early-phase clinical failure of TKIs. Among them, a lack of predictable preclinical models to stringently test efficacy, poorly defined intrinsic resistance mechanisms, the mostly unexplored role of the TME, and the hurdles in defining the actual—and optimal—exposure to the TKI all contribute to this huge gap in translating preclinical successes to the clinic. The alternative possibility, in principle, is that the vast majority of TKIs simply do not work in mCRC. We currently see this as an unlikely scenario and therefore support ongoing efforts to develop more advanced preclinical models of higher complexity that might further our understanding of mechanisms of action and of resistance. Such models should also enable us to test more efficient dosing and scheduling of TKIs—either as monotherapy or in combination with other anti-cancer agents, which we hold as essential even for already-failed drugs to determine their true potential. Preclinical prediction of early resistance mechanisms may facilitate improved patient selection for clinical trials to enhance the successful translation of TKIs for patients with mCRC. As a caveat, several challenges remain for the next generation of preclinical 3D tumour models to strike a balance in complexity, feasibility, and functionality between conventional in vitro cell culture and animal studies. Moreover, in vitro models are not expected to close the clinical translational gaps due to drug toxicity and adverse effects. Still, we envision that tumour tissue engineering from patient-specific cells (e.g., 3D assemblies of PDTOs and TME components) will provide new insight in response and resistance mechanisms to the preclinically tested drug (combinations).

      Funding

      DVFT is funded by a Hypatia Fellowship from the Radboudumc.

      Notes

      Role of the funder – None.
      Disclosure – No disclosures to report.

      Data availability statement

      Articles used in this review are from published sources and ongoing clinical trials. Tables with the most recent studies including different model types is available upon request.

      CRediT authorship contribution statement

      Kirti K. Iyer: Conceptualization, Investigation. Nielka P. van Erp: Conceptualization. Daniele V.F. Tauriello: Conceptualization. Henk M.W. Verheul: Conceptualization. Dennis Poel: Conceptualization, Investigation.

      Declaration of Competing Interest

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Appendix A. Supplementary data

      The following are the Supplementary data to this article:

      References

        • Dekker E.
        • et al.
        Colorectal cancer.
        Lancet. 2019; 394: 1467-1480
        • Gravalos C.
        • et al.
        Role of tyrosine kinase inhibitors in the treatment of advanced colorectal cancer.
        Clin Colorectal Cancer. 2007; 6: 691-699
        • Rossini D.
        • et al.
        Treatments after progression to first-line FOLFOXIRI and bevacizumab in metastatic colorectal cancer: a pooled analysis of TRIBE and TRIBE2 studies by GONO.
        Br J Cancer. 2021; 124: 183-190
      1. Wookey V, Grothey A. Update on the role of pembrolizumab in patients with unresectable or metastatic colorectal cancer. Therap Adv Gastroenterol 2021;14:17562848211024460.

        • Andre T.
        • et al.
        Pembrolizumab in Microsatellite-Instability-High Advanced Colorectal Cancer.
        N Engl J Med. 2020; 383: 2207-2218
        • Casak S.J.
        • et al.
        FDA Approval Summary: Pembrolizumab for the First-line Treatment of Patients with MSI-H/dMMR Advanced Unresectable or Metastatic Colorectal Carcinoma.
        Clin Cancer Res. 2021; 27: 4680-4684
        • Franke A.J.
        • et al.
        Immunotherapy for Colorectal Cancer: A Review of Current and Novel Therapeutic Approaches.
        J Natl Cancer Inst. 2019; 111: 1131-1141
        • Van Cutsem E.
        • et al.
        Metastatic colorectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up.
        Ann Oncol. 2014; 25 Suppl 3: p. iii1-9
        • Bekaii-Saab T.
        • et al.
        Third- or Later-line Therapy for Metastatic Colorectal Cancer: Reviewing Best Practice.
        Clin Colorectal Cancer. 2019; 18: e117-e129
        • Wang J.
        • et al.
        Metastatic patterns and survival outcomes in patients with stage IV colon cancer: A population-based analysis.
        Cancer Med. 2020; 9: 361-373
        • Garcia-Aranda M.
        • Redondo M.
        Targeting Receptor Kinases in Colorectal Cancer.
        Cancers (Basel). 2019; 11
        • Paul M.K.
        • Mukhopadhyay A.K.
        Tyrosine kinase - Role and significance in Cancer.
        Int J Med Sci. 2004; 1: 101-115
        • Robinson D.R.
        • Wu Y.M.
        • Lin S.F.
        The protein tyrosine kinase family of the human genome.
        Oncogene. 2000; 19: 5548-5557
      2. Thomson RJ, Moshirfar M, Ronquillo Y. Tyrosine Kinase Inhibitors, in StatPearls. 2022: Treasure Island (FL).

        • Huang L.
        • Jiang S.
        • Shi Y.
        Tyrosine kinase inhibitors for solid tumors in the past 20 years (2001–2020).
        J Hematol Oncol. 2020; 13: 143
        • Cohen P.
        • Cross D.
        • Janne P.A.
        Kinase drug discovery 20 years after imatinib: progress and future directions.
        Nat Rev Drug Discov. 2021;
        • Tauriello D.V.F.
        • Batlle E.
        Targeting the Microenvironment in Advanced Colorectal Cancer.
        Trends Cancer. 2016; 2: 495-504
        • Cai X.
        • et al.
        Therapeutic Potential of Apatinib Against Colorectal Cancer by Inhibiting VEGFR2-Mediated Angiogenesis and beta-Catenin Signaling.
        Onco Targets Ther. 2020; 13: 11031-11044
        • Chen X.
        • et al.
        A Single-Arm, Phase II Study of Apatinib in Refractory Metastatic Colorectal Cancer.
        Oncologist. 2019; 24: 883-e407
        • Wang Y.
        • et al.
        Combination of Fruquintinib and Anti-PD-1 for the Treatment of Colorectal Cancer.
        J Immunol. 2020; 205: 2905-2915
        • Li J.
        • et al.
        Effect of Fruquintinib vs Placebo on Overall Survival in Patients With Previously Treated Metastatic Colorectal Cancer The FRESCO Randomized Clinical Trial.
        Jama-Journal of the American Medical Association. 2018; 319: 2486-2496
        • Shirley M.
        Fruquintinib: First Global Approval.
        Drugs. 2018; 78: 1757-1761
        • Yang S.
        • et al.
        Cabozantinib induces PUMA-dependent apoptosis in colon cancer cells via AKT/GSK-3beta/NF-kappaB signaling pathway.
        Cancer Gene Ther. 2020; 27: 368-377
        • Melsens E.
        • et al.
        The VEGFR Inhibitor Cediranib Improves the Efficacy of Fractionated Radiotherapy in a Colorectal Cancer Xenograft Model.
        Eur Surg Res. 2017; 58: 95-108
        • Yamamoto N.
        • et al.
        Phase I, dose escalation and pharmacokinetic study of cediranib (RECENTIN), a highly potent and selective VEGFR signaling inhibitor, in Japanese patients with advanced solid tumors.
        Cancer Chemother Pharmacol. 2009; 64: 1165-1172
        • Scott A.J.
        • et al.
        A phase II study investigating cabozantinib in patients with refractory metastatic colorectal cancer (AGICC 17CRC01).
        J Clin Oncol. 2020; 38
        • Wiegering A.
        • et al.
        E7080 (lenvatinib), a multi-targeted tyrosine kinase inhibitor, demonstrates antitumor activities against colorectal cancer xenografts.
        Neoplasia. 2014; 16: 972-981
      3. Lim JH et al. Patient-Derived, Drug-Resistant Colon Cancer Cells Evade Chemotherapeutic Drug Effects via the Induction of Epithelial-Mesenchymal Transition-Mediated Angiogenesis. Int J Mol Sci 2020;21(20).

        • Iwasa S.
        • et al.
        Phase II study of lenvatinib for metastatic colorectal cancer refractory to standard chemotherapy: the LEMON study (NCCH1503).
        ESMO Open. 2020; 5
        • Liu Y.C.
        • et al.
        Regorafenib suppresses epidermal growth factor receptor signaling-modulated progression of colorectal cancer.
        Biomed Pharmacother. 2020; 128: 110319
        • Vlachogiannis G.
        • et al.
        Patient-derived organoids model treatment response of metastatic gastrointestinal cancers.
        Science. 2018; 359: 920-926
        • Song X.
        • et al.
        Mcl-1 inhibition overcomes intrinsic and acquired regorafenib resistance in colorectal cancer.
        Theranostics. 2020; 10: 8098-8110
        • Grothey A.
        • et al.
        Regorafenib monotherapy for previously treated metastatic colorectal cancer (CORRECT): an international, multicentre, randomised, placebo-controlled, phase 3 trial.
        Lancet. 2013; 381: 303-312
        • Li J.
        • et al.
        Regorafenib plus best supportive care versus placebo plus best supportive care in Asian patients with previously treated metastatic colorectal cancer (CONCUR): a randomised, double-blind, placebo-controlled, phase 3 trial.
        Lancet Oncol. 2015; 16: 619-629
        • Lu Z.
        • et al.
        Regulation of intercellular biomolecule transfer-driven tumor angiogenesis and responses to anticancer therapies.
        J Clin Invest. 2021; 131
      4. Saltz, Phase II trial of sunitinib in patients with metastatic colorectal cancer after failure of standard therapy (vol 25, pg 4793, 2007). J Clin Oncol 2008;26(3):514-514.

        • Golas J.M.
        • et al.
        SKI-606, a Src/Abl inhibitor with in vivo activity in colon tumor xenograft models.
        Cancer Res. 2005; 65: 5358-5364
        • Daud A.I.
        • et al.
        Phase I study of bosutinib, a src/abl tyrosine kinase inhibitor, administered to patients with advanced solid tumors.
        Clin Cancer Res. 2012; 18: 1092-1100
        • Van Cutsem E.
        • et al.
        Nintedanib for the treatment of patients with refractory metastatic colorectal cancer (LUME-Colon 1): a phase III, international, randomized, placebo-controlled study.
        Ann Oncol. 2018; 29: 1955-1963
        • Cheng G.
        • et al.
        The microRNA-429/DUSP4 axis regulates the sensitivity of colorectal cancer cells to nintedanib.
        Mol Med Rep. 2021; 23
        • Kopetz S.
        • et al.
        Synergistic Activity of the Src Family Kinase Inhibitor Dasatinib and Oxaliplatin in Colon Carcinoma Cells Is Mediated by Oxidative Stress.
        Cancer Res. 2009; 69: 3842-3849
        • Strickler J.H.
        • et al.
        Phase I study of dasatinib in combination with capecitabine, oxaliplatin and bevacizumab followed by an expanded cohort in previously untreated metastatic colorectal cancer.
        Invest New Drugs. 2014; 32: 330-339
        • Marti F.E.M.
        • et al.
        Novel phase I trial design to evaluate the addition of cediranib or selumetinib to preoperative chemoradiotherapy for locally advanced rectal cancer: the DREAMtherapy trial.
        Eur J Cancer. 2019; 117: 48-59
        • Graves-Deal R.
        • et al.
        Broad-spectrum receptor tyrosine kinase inhibitors overcome de novo and acquired modes of resistance to EGFR-targeted therapies in colorectal cancer.
        Oncotarget. 2019; 10: 1320-1333
        • Strickler J.H.
        • et al.
        Cabozantinib and Panitumumab for RAS Wild-Type Metastatic Colorectal Cancer.
        Oncologist. 2021; 26: p. 465-+
        • Schiffmann L.M.
        • et al.
        A combination of low-dose bevacizumab and imatinib enhances vascular normalisation without inducing extracellular matrix deposition.
        Br J Cancer. 2017; 116: 600-608
        • Hoehler T.
        • et al.
        Phase I/II trial of capecitabine and oxaliplatin in combination with bevacizumab and imatinib in patients with metastatic colorectal cancer: AIO KRK 0205.
        Br J Cancer. 2013; 109: 1408-1413
        • Janssen E.
        • et al.
        Combinatorial Immunotherapies for Metastatic Colorectal Cancer.
        Cancers. 2020; 12
        • Kato Y.
        • et al.
        Lenvatinib plus anti-PD-1 antibody combination treatment activates CD8(+) T cells through reduction of tumor-associated macrophage and activation of the interferon pathway.
        PLoS ONE. 2019; 14
        • Gomez-Roca C.
        • et al.
        LEAP-005: A phase II multicohort study of lenvatinib plus pembrolizumab in patients with previously treated selected solid tumors-Results from the colorectal cancer cohort.
        J Clin Oncol. 2021; 39
        • Doleschel D.
        • et al.
        Regorafenib enhances anti-PD1 immunotherapy efficacy in murine colorectal cancers and their combination prevents tumor regrowth.
        J Exp Clin Cancer Res. 2021; 40
        • Fukuoka S.
        • et al.
        Regorafenib Plus Nivolumab in Patients With Advanced Gastric or Colorectal Cancer: An Open-Label, Dose-Escalation, and Dose-Expansion Phase Ib Trial (REGONIVO, EPOC1603).
        J Clin Oncol. 2020; 38: p. 2053-+
        • Wang C.
        • et al.
        Regorafenib and Nivolumab or Pembrolizumab Combination and Circulating Tumor DNA Response Assessment in Refractory Microsatellite Stable Colorectal Cancer.
        Oncologist. 2020; 25: e1188-e1194
        • Cousin S.
        • et al.
        Regorafenib-Avelumab Combination in Patients with Microsatellite Stable Colorectal Cancer (REGOMUNE): A Single-arm, Open-label. Phase II Trial.
        Clin Cancer Res. 2021; 27: 2139-2147
        • Kim R.D.
        • et al.
        A phase I/Ib study of regorafenib and nivolumab in mismatch repair proficient advanced refractory colorectal cancer.
        Eur J Cancer. 2022; 169: 93-102
        • Yang K.
        • et al.
        Real-world outcomes of regorafenib combined with immune checkpoint inhibitors in patients with advanced or metastatic microsatellite stable colorectal cancer: A multicenter study.
        Cancer Immunol Immunother. 2022; 71: 1443-1451
        • Zhu G.
        • et al.
        Temozolomide and Pazopanib Combined with FOLFOX Regressed a Primary Colorectal Cancer in a Patient-derived Orthotopic Xenograft Mouse Model.
        Transl Oncol. 2020; 13: 100739
        • Brady J.
        • et al.
        An open-label study of the safety and tolerability of pazopanib in combination with FOLFOX6 or CapeOx in patients with colorectal cancer.
        Invest New Drugs. 2013; 31: 1228-1235
        • Wachsberger P.
        • et al.
        Combination of vandetanib, radiotherapy, and irinotecan in the LoVo human colorectal cancer xenograft model.
        Int J Radiat Oncol Biol Phys. 2009; 75: 854-861
        • Meyerhardt J.A.
        • et al.
        Phase I Study of Cetuximab, Irinotecan, and Vandetanib (ZD6474) as Therapy for Patients with Previously Treated Metastastic Colorectal Cancer.
        PLoS ONE. 2012; 7
        • Mross K.
        • et al.
        FOLFIRI and sunitinib as first-line treatment in metastatic colorectal cancer patients with liver metastases - a CESAR phase II study including pharmacokinetic, biomarker, and imaging data.
        Int J Clin Pharmacol Ther. 2014; 52: 642-652
        • Do K.
        • et al.
        A Phase II Study of Sorafenib Combined With Cetuximab in EGFR-Expressing, KRAS-Mutated Metastatic Colorectal Cancer.
        Clin Colorectal Cancer. 2015; 14: 154-161
        • Martinelli E.
        • et al.
        Synergistic antitumor activity of sorafenib in combination with epidermal growth factor receptor inhibitors in colorectal and lung cancer cells.
        Clin Cancer Res. 2010; 16: 4990-5001
        • Cai X.
        • et al.
        Apatinib enhanced anti-PD-1 therapy for colon cancer in mice via promoting PD-L1 expression.
        Int Immunopharmacol. 2020; 88: 106858
        • Ren C.
        • et al.
        Anti-PD-1 antibody SHR-1210 plus apatinib for metastatic colorectal cancer: a prospective, single-arm, open-label, phase II trial.
        Am J Cancer Res. 2020; 10: 2946-2954
        • Santoro A.
        • et al.
        A phase II randomized multicenter trial of gefitinib plus FOLFIRI and FOLFIRI alone in patients with metastatic colorectal cancer.
        Ann Oncol. 2008; 19: 1888-1893
        • Rizzo G.
        • et al.
        Patient-derived tumor models: a more suitable tool for pre-clinical studies in colorectal cancer.
        J Exp Clin Cancer Res. 2021; 40
        • Gillet J.P.
        • Varma S.
        • Gottesman M.M.
        The clinical relevance of cancer cell lines.
        J Natl Cancer Inst. 2013; 105: 452-458
        • Burtin F.
        • Mullins C.S.
        • Linnebacher M.
        Mouse models of colorectal cancer: Past, present and future perspectives.
        World J Gastroenterol. 2020; 26: 1394-1426
        • Nunes M.
        • et al.
        Evaluating Patient-Derived Colorectal Cancer Xenografts as Preclinical Models by Comparison with Patient Clinical Data.
        Cancer Res. 2015; 75: 1560-1566
        • Bleijs M.
        • et al.
        Xenograft and organoid model systems in cancer research.
        EMBO J. 2019; 38: e101654
        • Li Z.L.
        • et al.
        Changes in extracellular matrix in different stages of colorectal cancer and their effects on proliferation of cancer cells.
        World J Gastroint Oncol. 2020; 12: 267-275
        • Wu G.
        • et al.
        Overcoming treatment resistance in cancer: Current understanding and tactics.
        Cancer Lett. 2017; 387: 69-76
        • Van der Jeught K.
        • et al.
        Drug resistance and new therapies in colorectal cancer.
        World J Gastroenterol. 2018; 24: 3834-3848
      5. Acquired and Intrinsic Resistance to Colorectal Cancer Treatment. 2018.

        • Holohan C.
        • et al.
        Cancer drug resistance: an evolving paradigm.
        Nat Rev Cancer. 2013; 13: 714-726
        • Longley D.B.
        • Johnston P.G.
        Molecular mechanisms of drug resistance.
        Journal of Pathology. 2005; 205: 275-292
        • Weickhardt A.J.
        • et al.
        Dual Targeting of the Epidermal Growth Factor Receptor Using the Combination of Cetuximab and Erlotinib: Preclinical Evaluation and Results of the Phase II DUX Study in Chemotherapy-Refractory, Advanced Colorectal Cancer.
        J Clin Oncol. 2012; 30: 1505-1512
        • Molinari C.
        • et al.
        Heterogeneity in Colorectal Cancer: A Challenge for Personalized Medicine?.
        Int J Mol Sci. 2018; 19
        • Frank M.H.
        • et al.
        Clinical Implications of Colorectal Cancer Stem Cells in the Age of Single-Cell Omics and Targeted Therapies.
        Gastroenterology. 2021; 160: 1947-1960
        • Tan H.Y.
        • et al.
        Targeting tumour microenvironment by tyrosine kinase inhibitor.
        Molecular Cancer. 2018; 17
        • Gerritse S.L.
        • et al.
        High-dose administration of tyrosine kinase inhibitors to improve clinical benefit: A systematic review.
        Cancer Treat Rev. 2021; 97
        • Aykul S.
        • Martinez-Hackert E.
        Determination of half-maximal inhibitory concentration using biosensor-based protein interaction analysis.
        Anal Biochem. 2016; 508: 97-103
        • Decosterd L.A.
        • et al.
        Therapeutic drug monitoring of targeted anticancer therapy.
        Biomark Med. 2015; 9: 887-893
        • Smith D.A.
        • Di L.
        • Kerns E.H.
        The effect of plasma protein binding on in vivo efficacy: misconceptions in drug discovery.
        Nat Rev Drug Discovery. 2010; 9: 929-939
        • Al Baghdadi T.
        • et al.
        Sunitinib in Patients with Metastatic Colorectal Cancer (mCRC) with FLT-3 Amplification: Results from the Targeted Agent and Profiling Utilization Registry (TAPUR) Study.
        Target Oncol. 2020; 15: 743-750
        • Rovithi M.
        • et al.
        Phase I Dose-Escalation Study of Once Weekly or Once Every Two Weeks Administration of High-Dose Sunitinib in Patients With Refractory Solid Tumors.
        J Clin Oncol. 2019; 37: 411-418
        • Rovithi M.
        • Verheul H.M.W.
        Pulsatile high-dose treatment with antiangiogenic tyrosine kinase inhibitors improves clinical antitumor activity.
        Angiogenesis. 2017; 20: 287-289
        • Zoetemelk M.
        • et al.
        Short-term 3D culture systems of various complexity for treatment optimization of colorectal carcinoma.
        Sci Rep. 2019; 9
        • Ganesh K.
        • et al.
        A rectal cancer organoid platform to study individual responses to chemoradiation.
        Nat Med. 2019; 25: p. 1607-+
        • Yao Y.
        • et al.
        Patient-Derived Organoids Predict Chemoradiation Responses of Locally Advanced Rectal Cancer.
        Cell Stem Cell. 2020; 26: p. 17-+
        • Ooft S.N.
        • et al.
        Patient-derived organoids can predict response to chemotherapy in metastatic colorectal cancer patients.
        Sci Transl Med. 2019; 11
        • Kim S.Y.
        • et al.
        Modeling Clinical Responses to Targeted Therapies by Patient-Derived Organoids of Advanced Lung Adenocarcinoma.
        Clin Cancer Res. 2021; 27: 4397-4409
        • Reed M.R.
        • et al.
        A Functional Precision Medicine Pipeline Combines Comparative Transcriptomics and Tumor Organoid Modeling to Identify Bespoke Treatment Strategies for Glioblastoma.
        Cells. 2021; 10
      6. Reidy E, et al. A 3D View of Colorectal Cancer Models in Predicting Therapeutic Responses and Resistance. Cancers (Basel) 2021. 13(2).

        • Fontana F.
        • et al.
        In Vitro 3D Cultures to Model the Tumor Microenvironment.
        Cancers. 2021; 13
        • Jin Z.J.
        • et al.
        The cross-talk between tumor cells and activated fibroblasts mediated by lactate/BDNF/TrkB signaling promotes acquired resistance to anlotinib in human gastric cancer.
        Redox Biol. 2021; 46
        • Low L.A.
        • et al.
        Organs-on-chips: into the next decade.
        Nat Rev Drug Discov. 2021; 20: 345-361
        • Hofer M.
        • Lutolf M.P.
        Engineering organoids.
        Nat Rev Mater. 2021; 6: 402-420
        • Scott A.J.
        • et al.
        Evaluation of the efficacy of dasatinib, a Src/Abl inhibitor, in colorectal cancer cell lines and explant mouse model.
        PLoS ONE. 2017; 12e0187173
        • Sharma M.R.
        • et al.
        Dasatinib in previously treated metastatic colorectal cancer: a phase II trial of the University of Chicago Phase II Consortium.
        Invest New Drugs. 2012; 30: 1211-1215
        • Townsley C.A.
        • et al.
        Phase II study of erlotinib (OSI-774) in patients with metastatic colorectal cancer.
        Br J Cancer. 2006; 94: 1136-1143
        • Rivera M.
        • et al.
        Patient-derived xenograft (PDX) models of colorectal carcinoma (CRC) as a platform for chemosensitivity and biomarker analysis in personalized medicine.
        Neoplasia. 2021; 23: 21-35
        • Rothenberg M.L.
        • et al.
        Randomized phase II trial of the clinical and biological effects of two dose levels of gefitinib in patients with recurrent colorectal adenocarcinoma.
        J Clin Oncol. 2005; 23: 9265-9274
        • Georgiou A.
        • et al.
        Inactivation of NF1 Promotes Resistance to EGFR Inhibition in KRAS/NRAS/BRAF(V600) -Wild-Type Colorectal Cancer.
        Mol Cancer Res. 2020; 18: 835-846
        • Sobrino A.
        • et al.
        3D microtumors in vitro supported by perfused vascular networks.
        Sci Rep. 2016; 6
        • Chan E.
        • et al.
        Phase II study of the Multikinase inhibitor of angiogenesis, Linifanib, in patients with metastatic and refractory colorectal cancer expressing mutated KRAS.
        Invest New Drugs. 2017; 35: 491-498
        • Kim S.Y.
        • et al.
        NCOA4-RET fusion in colorectal cancer: Therapeutic challenge using patient-derived tumor cell lines.
        J Cancer. 2018; 9: 3032-3037
        • Mross K.
        • et al.
        DCE-MRI assessment of the effect of vandetanib on tumor vasculature in patients with advanced colorectal cancer and liver metastases: a randomized phase I study.
        J Angiogenes Res. 2009; 1: 5
        • Wenzel T.
        • et al.
        Restoration of MARCK enhances chemosensitivity in cancer.
        J Cancer Res Clin Oncol. 2020; 146: 843-858
        • Isakoff S.J.
        • et al.
        Bosutinib plus capecitabine for selected advanced solid tumours: results of a phase 1 dose-escalation study.
        Br J Cancer. 2014; 111: 2058-2066
        • Palumbo I.
        • et al.
        Gefitinib enhances the effects of combined radiotherapy and 5-fluorouracil in a colorectal cancer cell line.
        Int J Colorectal Dis. 2014; 29: 31-41
        • Boland P.M.
        • et al.
        A phase I/II study of nintedanib and capecitabine in refractory metastatic colorectal cancer.
        J Clin Oncol. 2018; 36
        • Suzuki N.
        • et al.
        Effect of a novel oral chemotherapeutic agent containing a combination of trifluridine, tipiracil and the novel triple angiokinase inhibitor nintedanib, on human colorectal cancer xenografts.
        Oncol Rep. 2016; 36: 3123-3130
        • Bocci G.
        • et al.
        Antiangiogenic and anticolorectal cancer effects of metronomic irinotecan chemotherapy alone and in combination with semaxinib.
        Br J Cancer. 2008; 98: 1619-1629
        • Lockhart A.C.
        • et al.
        Phase I/pilot study of SU5416 (semaxinib) in combination with irinotecan/bolus 5-FU/LV (IFL) in patients with metastatic colorectal cancer.
        Am J Clin Oncol-Cancer Clin Trials. 2006; 29: 109-115
        • Di Desidero T.
        • et al.
        Chemotherapeutic and antiangiogenic drugs beyond tumor progression in colon cancer: Evaluation of the effects of switched schedules and related pharmacodynamics.
        Biochem Pharmacol. 2019; 164: 94-105
        • To K.K.W.
        • et al.
        Vatalanib sensitizes ABCB1 and ABCG2-overexpressing multidrug resistant colon cancer cells to chemotherapy under hypoxia.
        Biochem Pharmacol. 2015; 97: 27-37
        • Hecht J.R.
        • et al.
        Randomized, Placebo-Controlled, Phase III Study of First-Line Oxaliplatin-Based Chemotherapy Plus PTK787/ZK 222584, an Oral Vascular Endothelial Growth Factor Receptor Inhibitor, in Patients With Metastatic Colorectal Adenocarcinoma.
        J Clin Oncol. 2011; 29: 1997-2003