Abstract
Aims
To evaluate the risk of pleural disorders (PD) associated with 33 protein kinase (PK) inhibitors (PKIs) through a disproportionality analysis and to identify which PKs and pathways are involved in PKI‐induced PD.
Methods
To evaluate the risk of PD, reporting odds ratios (RORs) were calculated for 33 PKIs through data registered in the World Health Organization safety report database (VigiBase). We undertook a literature review to identify PKs that were possibly involved in PD caused by PKIs. Pearson correlation coefficients (r) between RORs and affinity data of 19 PKIs were calculated to identify the cellular target most likely to be involved in PKI‐induced PD.
Results
A total of 235 110 individual case safety reports were extracted from the database for 33 available PKIs. Among these reports, 5001 concerned PD (2.1%). Significant and positive disproportionality for PD was found for 29 of 33 PKI included in our study with top values for dasatinib [ROR = 115.3; 95% confidence interval (CI): 110.1–120.8], bosutinib (ROR = 20.4; 95% CI: 15.8–26.4) and ponatinib (ROR = 12; 95% CI: 9.2–15.6). Correlation analyses between the product of dissociation constant and ROR highlighted possibly Lyn involvement in PD with PKI (r = 0.73, P = 0.0004).
Conclusions
Our study showed that 28 of the 33 tested PKIs were associated with PD. Besides, the study highlighted the role of Lyn in PD caused by PKIs through an immune‐mediated process.
Keywords: drug safety, pharmacodynamics, pharmacovigilance
What is Already Known about this Subject
Pleural disorders (PD) are an adverse drug reaction (ADR) shared by several protein kinase inhibitors (PKIs).
PKI‐induced PD in real life conditions of use has never been assessed.
Cellular target (s) and pathway (s) leading to clinical PD with PKIs are not known.
What this Study Adds
We assessed the pharmacovigilance signal of PD associated with 33 PKIs from VigiBase, the largest pharmacovigilance database in the world.
We used an approach, based on pharmacovigilance and pharmacodynamics data, to identify cellular pathways involved in PD.
We identified the role of protein kinase Lyn in PD caused by PKIs through a possible immune mediated mechanism.
Introduction
Kinases have emerged as one of the most intensely pursued targets in current pharmacological research (especially for cancer) due to their critical role in both cell signalling and death. The human protein kinase (PK) gene family consists of 518 members, and the strategy behind kinase inhibition was initiated during late the 1980s when inhibitors against http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1797 were described 1, 2. Since then, many kinase inhibitors with diverse molecular and pharmacological profiles have been approved for several indications. Improperly named targeted therapies, PK inhibitors (PKIs) often have a complex affinity profile while interacting with multiple targets because of critical conserved sequences between PKs genes. Biological consequences of multikinase activities are poorly understood, and investigation of interactions between PKIs and the human kinome could help to elucidate the relationships among selectivity, efficacy and safety of these drugs 3. Imatinib is one excellent example of this point. Initially, imatinib was first approved in 2001 for chronic myeloid leukaemia as a highly selective inhibitor of Bcr‐abl. In the second stage, imatinib was discovered to have significant activity against several other clinically relevant kinases such as http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1805, which led to the approval for the second indication, gastrointestinal stromal tumours, in 2012 4, 5. However, if multikinase activity is unusual for efficacy, it may contribute to adverse drug reaction (ADR) occurrence. Indeed, ADRs can be caused by two separate mechanisms: (i) on‐target toxicity, which leads to an ADR being induced by inhibition of the PKI's target of interest; and (ii) off‐target toxicity, leading to an ADR caused by inhibition of a secondary or unexpected target (s) of the drug 6. In this context, pharmacovigilance (PV) data can help to characterize the ADR‐associated PKI profiles in clinically relevant situations but also provide elements to allow the identification of pathophysiological pathways that may be involved in these cases. Recently, the study of Patras‐de‐Campaigno et al. highlighted the role of http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1923 tyrosine kinases in cardiac failure related to anticancer PKI exposition 7. Apart from cardiac toxicity, PV PKI reports revealed that pulmonary ADRs such as pleural disorders (PD) could lead to serious effects and/or premature treatment discontinuation with consequent loss of chance for the patient's recovery. Identifying the most at‐risk PKIs and understanding the pharmacodynamic mechanisms causing PD would allow ADR developmental mechanisms to be used in order to consider management strategies but would need to anticipate ADR drug candidate profiles. In the present study, we investigated which PKs and pathways are involved in PKI‐induced PD. For this purpose we evaluated several approaches: (i) PV data in order to perform a descriptive and disproportionality analysis to identify the PKIs associated with a pharmacovigilance signal; (ii) from a literature review, we determined the PKs for which there is a pharmacological hypothesis linking their inhibition to the occurrence of a PD; and (iii) using PV disproportionality analysis and affinity PK identified profiles, we identified the cellular targets/pathways most likely to be involved in PKI‐induced PD.
Methods
Data sources
Pharmacovigilance database
The Uppsala Monitoring Centre receives individual case safety reports (ICSRs) of suspected ADRs sent by national pharmacovigilance centres, which are stored in the World Health Organization's (WHO's) global safety database (VigiBase) 8. In January 2018, VigiBase contained more than 16 million ICSRs from 127 countries. Each ICSR consists of description of drugs that are suspected to cause ADRs and contains information on patient age, gender, medical history, country, drugs taken, and drug initiation and stop dates. Drugs are coded using the WHO drug dictionary, covering over 150 000 medicines and vaccines.
Data extraction and selection
PKI exposure
PKIs were selected according to their ability to inhibit multiple molecular targets. Also, to be included in our analysis, a PKI had to have at least 100 ICSRs reported in VigiBase and at least one PD‐related ICSR. Table 1 presents the 33 PKIs meeting the chosen criteria and selected for descriptive and disproportionality analysis. For each PKI, the description of PD was searched in the adverse reaction section of the Summary of Product Characteristics (European Medicines Agency) or the labelling (US Food and Drug Administration). We carried out linear correlation analyses for those with affinity map screened by Davis et al. 9.
Table 1.
Pharmacodynamics properties of 33 protein kinase inhibitors and adverse drug reactions contained in VigiBase
| Drug name (year of approval) | Indication | Main targets | ICSR all effect n (%) N = 235 110 | ICRS of PD n (%) N = 5001 | PD described in SMP (incidence) |
|---|---|---|---|---|---|
| Afatinib * (2013) | NSCLC | EGFR, HER2 | 3040 (1.29) | 57 (1.14) | No |
| Alectinib (2015) | NSCLC | http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1839 | 269 (0.11) | 5 (0.10) | No |
| Axitinib * (2012) | Renal carcinoma | Kit, PDGFR‐β. http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=324 | 4198 (1.79) | 21 (0.42) | No |
| Bosutinib * (2012) | CML | Bcr‐abl | 1565 (0.67) | 61 (1.22) | Yes (common) |
| Cabozantinib (2012) | Thyroid carcinoma. Renal carcinoma | http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1807, Kit, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1815, Rek, Tek, VEGFR2 | 1084 (0.46) | 11 (0.22) | No |
| Ceritinib (2014) | NSCLC | ALK | 758 (0.32) | 10 (0.20) | No |
| Cobimetinib (2015) | Melanoma | MEK | 309 (0.13) | 1 (0.02) | No |
| Crizotinib * (2011) | NSCLC | ALK, HGFR, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1840 | 4619 (1.96) | 75 (1.50) | No |
| Dabrafenib (2013) | Melanoma | http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1943 | 2217 (0.94) | 5 (0.10) | No |
| Dasatinib * (2006) | CML | ABL, Kit, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2206 | 13 566 (5.77) | 2354 (47.07) | Yes (very common) |
| Erlotinib * (2004) | NSCLC, pancreatic cancer | EGFR | 28 340 (12.05) | 372 (7.44) | No |
| Gefitinib * (2003) | NSCLC | EGFR | 5450 (2.32) | 68 (1.36) | No |
| Ibrutinib (2013) | CLL, mantle cell lymphoma | BTK | 8105 (3.45) | 181 (3.62) | Non |
| Idelalisib (2014) | CLL, follicular lymphoma | http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2155 | 2350 (1.00) | 21 (0.42) | No |
| Imatinib * (2001) | CML, ALL, MD/MPD, HES/CEL, GIST, DFSP | ABL, PDGFR. Kit | 33 043 (14.05) | 596 (11.92) | Yes (uncommon) |
| Lapatinib * (2007) | Breast cancer | HER2 | 13 176 (5.60) | 97 (1.94) | No |
| Lenvatinib (2015) | Thyroid carcinoma | VEGFR2 | 512 (0.22) | 6 (0.12) | No |
| Lestaurtinib * (2006) | AML | Flt3, http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=581. http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1817 | 171 (0.07) | 3 (0.06) | NA |
| Midostaurin * (2004) | AML, aggressive/systemic mastocytosis, mast cell leukaemia | VEGFR2, Flt3. Kit, PDGFR‐α. http://www.guidetopharmacology.org/GRAC/ReceptorFamiliesForward?type=ENZYME&familyId=286 | 341 (0.15) | 4 (0.08) | Yes (common) |
| Nilotinib * (2007) | CML | ABL | 11 373 (4.84) | 222 (4.44) | Yes (uncommon) |
| Nintedanib * (2014) | Idiopathic pulmonary fibrosis, NSCLC | http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=323, PDGFR, VEGFR1/2/3 | 2138 (0.91) | 13 (0.26) | No |
| Osimertinib (2015) | NSCLC | EGFR | 431 (0.18) | 9 (0.18) | No |
| Palbociclib (2015) | Breast cancer | http://www.guidetopharmacology.org/GRAC/ReceptorFamiliesForward?type=ENZYME&familyId=453 | 5333 (2.27) | 31 (0.62) | No |
| Pazopanib * (2009) | Renal carcinoma, soft‐tissue sarcoma | Kit, PDGFR. VEGFR1/2/3 | 12 837 (5.46) | 98 (1.96) | No |
| Ponatinib (2012) | CML, ALL | VEGFR2, ABL. FGFR1/2/3, Flt3 | 2451 (1.04) | 57 (1.14) | Yes (common) |
| Regorafenib (2012) | Colorectal cancer, GIST I | Kit, PDGFR‐β. http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=610, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2185. VEGFR1/2/3 | 4798 (2.04) | 24 (0.48) | No |
| Ruxolitinib * (2011) | Myelofibrosis, | http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=581/2 | 9781 (4.16) | 51 (1.02) | No |
| Sorafenib * (2005) | Thyroid, renal and hepatocellular carcinoma | Flt3, PDGFR‐β. RAF, RET. VEGFR2 | 19 857 (8.45) | 163 (3.26) | No |
| Sunitinib * (2006) | GIST, pancreatic cancer | Flt3, Kit. PDGFR‐β, RET. VEGFR2 | 22 766 (9.68) | 320 (6.40) | Yes (common) |
| Tofacitinib * (2012) | Rheumatoid arthritis | http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=581 | 12 711 (5.40) | 21 (0.42) | No |
| Trametinib (2013) | Melanoma | MEK | 1869 (0.79) | 8 (0.16) | No |
| Vandetanib * (2011) | Thyroid carcinoma | EGFR, RET. VEGFR2 | 802 (0.34) | 4 (0.08) | No |
| Vemurafenib (2011) | Melanoma | BRAF | 4850 (2.06) | 32 (0.62) | No |
Protein kinase inhibitors selected for linear correlation analysis
ALK, anaplastic lymphoma receptor tyrosine kinase; ALL, acute lymphoblastic leukaemia; B‐Raf, B‐ Raf proto‐oncogene, serine/threonine kinase; BTK, Bruton agammaglobulinemia tyrosine kinase; Bcr‐abl, Bcr‐abl fusion protein; CDK, cyclin‐dependent kinase inhibitor protein; CEL, chronic eosinophilic leukaemia; CML, chronic myeloid leukaemia; CLL, chronic lymphocytic leukaemia; DFSP, dermofibro sarcoma protuberans; EGFR, epidermal growth factor receptor; FGFR1/2/3, fibroblast growth factor receptors; FLT3, receptor‐type tyrosine protein kinase FLT3; GIST, gastrointestinal stromal tumour; HES, hypereosinophilic syndrome; HER: human epidermal growth factor receptor; HGFR, hepatocyte growth factor receptor; ICRS, individual case safety report; JAK1/2, tyrosine protein ki‐ nase JAK1/2; KIT, KIT proto‐oncogene receptor tyrosine kinase; Lyn: LYN proto‐oncogene, Src family tyrosine kinase; MD/MPD, myelodysplastic/myeloproliferative syndrome; NSCLC, non‐small cell lung cancer; NTRK1, neurotrophic receptor tyrosine kinase 1; PD, pleural disorder; PDGFRα/β, platelet‐derived growth factor receptor alpha/beta; PKC‐α, protein kinase C; PI3, phosphatidylinositol‐4,5‐bisphosphate 3‐kinase; Raf, Raf‐1 proto‐oncogene, serine/threonine kinase; RET, RET receptor tyrosine kinase; ROS, reactive oxygen species; SMP, summary of product characteristics; c‐SRC, v‐src sarcoma viral oncogene homologue; TEK, TEK receptor tyrosine kinase; yes: Yamaguchi sarcoma viral related oncogene homologue; VEGFR‐1/2/3, vascular endothelial growth factor 1/2/3
PDs
ADRs are coded according to the Medical Dictionary for Regulatory Activities (MedDRA). There are five levels in the MedDRA hierarchy, arranged from system organ classes with terms very general, to preferred term with specific terms. We identified PD in VigiBase using the following MedDRA preferred term as validated by a pulmonologist: (i) “pleurisy”; (ii) “pleural disorder” (iii) “pleuropericarditis”; and (iv) “pleural effusion».
Affinity data
Data on PKI binding affinities for PKs involved in PD were extracted from the study by Davis et al. 9. This in vitro study tested the interactions of 72 kinase inhibitors with 442 kinases covering more than 80% of the human catalytic protein kinome. Dissociation constants [expressed as Kd (nmol l–1)] were converted into products of dissociation constants (pKd = – log10 (Kd × 10–9; Supporting Information Appendix S1). Although other affinity studies between PKI and PK exist, the study of Davis et al. 9 was selected for the number of tested protein kinases but also to overcome measurement bias related to the use of other methods for the measurement of affinity (such as conversion of half maximal inhibitory concentration to Kd).
Identification of PKs involved in PDs
A literature search (MEDLINE database and WEB OF SCIENCE) was performed with the algorithm presented in the Supporting Information Appendix S2. This search was completed by including references cited in the selected articles. In first step, we selected studies according to the presence of the identification of a PK as a potential mediator for a PD. In the second step, only cellular targets for which a pharmacological mechanism for PD occurrence was discussed or evaluated were selected such as the presence of a link between target inhibition and pulmonary involvement at the cellular level.
Analyses
Descriptive analysis
A descriptive analysis was performed on available data through ICSRs: (i) patient data (sex, age); (ii) characteristics of PD, including seriousness in which seriousness was defined as any untoward medical occurrence that results in congenital anomalies or birth defects; death; life‐threatening event; hospital admission and/or prolongation of existing hospital stay; persistent and/or significant disability/incapacity; a case judged as clinically relevant by the physician who reports the case 10; (iii) time of onset; (iv) outcome; and (v) PKI characteristics, including posology and indication. The number and proportion of patients were presented for qualitative variables in addition to median and interquartile range or mean and standard deviation for quantitative variables.
Disproportionality analysis
Disproportionality analysis compares the proportion of specific ADRs reported for a single drug with the proportion of the same ADR for all other drugs or for a selected panel of control drugs. Briefly, if the proportion of Y ADRs in patients exposed to drug X is greater than the proportion of Y ADRs in patients not exposed to drug X, this suggests an association between the specific drug and reaction and is a potential signal for safety 11, 12. In the present study, disproportionality was calculated using the reporting odds ratio (ROR) and its 95% confidence interval (CIs): for each PKI, the ROR is the odds of exposure to the PKI among PD reports divided by the odds of exposure to all the other drugs recorded in the database among all others ADRs reports during the same period. Woolf's method was used to calculate the 95% CI. To detect and quantify a possible masking effect in our study, we undertook sensitivity analysis excluding a subset of reports suspected to induce such an effect. The masking effect is the effect by which a signal of disproportionate reporting for a given drug‐event pair might be suppressed by the presence of another product in the same database 13.
Linear correlation analysis
To determine which suspected cellular targets were involved in PKI‐induced PD, we calculated the Pearson correlation coefficients (r) between the affinities of the PKIs (expressed as pKd) for each selected target and their disproportionality (shown as the ROR) for PD. To handle multiple comparisons, the threshold P‐value of the test was adapted using a Bonferroni correction. This correction reduced the risk of finding a spurious association with an increased type I error. Scatter plots with trend line, estimate by regression line, for each tested PK were constructed to analyse the characteristics of the linear relationship. Sensitivity analyse, with the exclusion of dasatinib, was carried out to evaluate the weighting of the outlier data points in the linear relationship. All analyses were performed using the statistical software package SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA).
Regulatory aspects
This study was registered with the European Network of Pharmacoepidemiology and Pharmacovigilance Centers (ENCePP) coordinated by EMA (EPAS18319). The use of anonymised data from VigiBase abides by the rules for the protection of personal data.
Nomenclature of targets and ligands
Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 14, and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 15, 16.
Results
Pharmacovigilance data
We extracted 235 110 ICSRs pertaining to the 33 selected PKIs (Table 1). Imatinib, erlotinib and sunitinib were the three drugs most frequently reported in all ICSRs (14.05%, 12.05% and 9.68%, respectively). A total of 5001 ICSRs contained at least one MedDRA preferred term for PD. Dasatinib accounted for almost half of PD ICSRs (47.07%) followed by imatinib (11.92%), erlotinib (7.44%), sunitinib (6.40%) and nilotinib (4.44%). Baseline information contained in the ICSRs is presented in Table 2. For PD ICSRs, the mean age was 63.9 years (standard deviation 15.4; missing in 32.1% of ICSRs) with a predominance of male patients (50.3%) vs. 2145 females (42.9%). The reaction was serious for 80.8%, including 617 deaths (15.26%). Among the 230 109 other ICSRs, ADRs equally involved patients of both sexes with 107 247 males (46.6%) and 107, 233 females (46.6%). They had a reported mean age of 61.3 years (standard deviation 14.7; missing in 36.8% of ICSRs). The reactions were serious for 61.2% of ICSRs including 40 451 deaths (28.7%). A more detailed description of the information contained in the PD ICSRs has been completed and the results are presented in Supporting Information Appendix S3.
Table 2.
Baseline information reported in the 235 110 individual case safety reports (ICSRs) concerning the 33 PKIs in VigiBase
| ICSRs with PD (n = 5001) | ICSRs with other ADR (n = 230 109) | P value | |
|---|---|---|---|
| Sex, n (%) | <0.0001 | ||
| male | 2516 (50.31) | 107 247 (46.61) | |
| female | 2145 (42.89) | 107 233 (46.60) | |
| unknown | 338 (6.76) | 15 537 (6.75) | |
| Age (years) | |||
| mean (SD) | 63.89 ( ± 15.38) | 61.33 ( ± 14.75) | <0.0001 |
| n (%) 0–53 | 628 (12.56) | 34 806 (15.13) | |
| 53–63 | 754 (15.08) | 36 164 (15.72) | |
| 63–72 | 965 (19.30) | 37 954 (16.49) | |
| > 72 | 1047 (20.94) | 36 534 (15.88) | |
| unknown | 1607 (32.13) | 84 651 (36.79) | |
| Seriousness, n (%) | <0.0001 | ||
| unserious | 741 (14.82) | 79 011 (34.34) | |
| serious | 4043 (80.84) | 140 904 (61.23) | |
| unknown | 217 (4.34) | 10 194 (4.43) | |
| Seriousness pattern (among serious), n (%) death | 617 (15.26) | 40 451 (28.71) | |
| life threatening | 193 (4.77) | 4165 (2.96) | |
| other | 3619 (89.51) | 96 288 (68.33) | |
ADR, adverse drug reaction; ICSR, individual case safety report; PD, pleural disorder; PKI, protein kinase inhibitor; SD, standard deviation
Disproportionality analysis
Table 3 presents the RORs for PD with their 95% CIs for each PKI compared with all other drugs registered in VigiLyze. A significant positive disproportionality was found for 29 of 33 PKI included in our study with top values for dasatinib, bosutinib, ponatinib and ibrutinib. Dasatinib disproportionality for PD was mainly higher than for other PKIs with ROR of 115.34 (95% CI: 110.14–120.78). Given these results and to avoid a possible masking effect due to dasatinib, a sensitivity analysis excluding dasatinib data was performed to calculate the ROR. This report confirmed significant disproportionality for 30 of 33 PKI with similar ranking [the ROR associated to vandetanib became significant (ROR = 2.75, 95% CI: 1.03–7.35); Supporting Information Appendix S4].
Table 3.
Disproportionality analysis for pleural disorder with 33 protein kinase inhibitors, showing the reporting odds ratios (RORs) with confidence intervals (CIs), by descending order
| Drug name | ROR (95% CI) | |
|---|---|---|
| 1 | Dasatinib | 115.34 (110.14–120.78) |
| 2 | Bosutinib | 20.45 (15.83–26.43) |
| 3 | Ponatinib | 12.00 (9.23–15.61) |
| 4 | Ibrutinib | 11.56 (9.97–13.40) |
| 5 | Osimertinib | 10.73 (5.55–20.78) |
| 6 | Nilotinib | 10.09 (8.83–11.53) |
| 7 | Afatinib | 9.63 (7.41–12.52) |
| 8 | Alectinib | 9.53 (3.93–23.09) |
| 9 | Imatinib | 9.42 (8.68–10.23) |
| 10 | Lestaurtinib | 8.98 (2.87–28.15) |
| 11 | Crizotinib | 8.32 (6.62–10.46) |
| 12 | Sunitinib | 7.25 (6.48–8.10) |
| 13 | Erlotinib | 6.77 (6.11–7.50) |
| 14 | Ceritinib | 6.73 (3.60–12.56) |
| 15 | Gefitinib | 6.37 (5.01–8.09) |
| 16 | Lenvatinib | 5.97 (2.67–13.35) |
| 17 | Midostaurin | 5.97 (2.23–16.01) |
| 18 | Cabozantinib | 5.16 (2.85–9.35) |
| 19 | Idélalisib | 4.54 (2.95–6.98) |
| 20 | Sorafenib | 4.18 (3.58–4.88) |
| 21 | Pazopanib | 3.88 (3.18–4.74) |
| 22 | Lapatinib | 3.74 (3.06–4.57) |
| 23 | Vemurafenib | 3.34 (2.36–4.74) |
| 24 | Nintedanib | 3.08 (1.78–5.31) |
| 25 | Palbociclib | 2.94 (2.07–4.19) |
| 26 | Ruxolitinib | 2.64 (2.00–3.47) |
| 27 | Axitinib | 2.53 (1.65–3.89) |
| 28 | Regorafenib | 2.53 (1.69–3.78) |
| 29 | Vandetanib | 2.52 (0.94–6.74) |
| 30 | Trametinib | 2.16 (1.08–4.33) |
| 31 | Cobimetinib | 1.63 (0.23–11.64) |
| 32 | Dabrafenib | 1.14 (0.47–2.73) |
| 33 | Tofacitinib | 0.83 (0.54–1.28) |
Selection of PKs involved in PD
We identified 11 PKs possibly involved in PDs in the literature review based on the presence of a pharmacological hypothesis. For each PK, the pharmacological hypothesis is assumed in Table 4.
Table 4.
List of selected molecular targets and associated pharmacological hypothesis
| Protein kinase | Physiological process involved | Pharmacological hypothesis | Reference |
|---|---|---|---|
| PDGFR‐β | Interstitial tissue pressure 46 | PDGFR‐β is expressed in the pericytes and its inhibition could cause a decrease in the interstitial pressure of the pleural space leading to an increase of the transfer of liquid, thus favouring the appearance of pleural disorder. | 18, 25, 35, 36, 37 |
| Lyn | Permeability and endothelial stability 47 | Inhibition of Lyn could destabilize endothelial barrier leading to an increase endothelial permeability in pulmonary tissue. | 35, 36, 37 |
| TEC, BTK, Lck, Lyn, kit | Immune response 48, 49 | Cell‐mediated immune‐mechanism. | 23, 24, 35, 42, 50, 51, 52 |
| DDR1 | adhesion of epithelial cells to the extracellular matrix 53 | DDR1 is expressed on bronchial epithelial cells, transducing signals from the extracellular matrix and plays a role in cell adhesion suggesting it may be a mediator of the pulmonary toxicities. | 35 |
| HER1–4 | Angiogenesis 54 | Inhibition of these protein kinases (EGF receptors) could affect the vascular pleural endothelium. | 55 |
BTK, Bruton agammaglobulinemia tyrosine kinase; DDR1, discoidin domain receptor family, member 1; EGF, epidermal growth factor; HER, human epidermal growth factor receptor; c‐kit, proto‐oncogene tyrosine kinase; Lck lymphocyte‐specific protein tyrosine kinase; Lyn, v‐yes‐1 Yamaguchi sarcoma viral related oncogene homologue; PDGFR β, platelet‐derived growth factor receptor β; SRC, v‐src sarcoma viral oncogene homologue; VEGF, vascular endothelial growth factor; yes, Yamaguchi sarcoma viral related oncogene homologue
Relationship between disproportionality for PD and PKI affinity
Nineteen of the 33 PKIs available in VigiBase were selected according to their pharmacodynamic screening in the dataset of Davis et al. (Table 1) 9. The Pearson correlation coefficients (r) between RORs and pKd values are presented in Table 5. A significant correlation coefficient between pKd and ROR was observed for four proteins kinases: (i) http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2238 (r = 0.95, P < 0.0001); (ii) http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2060 (r = 0.73, P = 0.0004); (iii) http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1948 (r = 0.72, P = 0.0005); and (iv) http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2053 (r = 0.62, P = 0.0042). However, after a sensitivity analysis with exclusion of dasatinib, the results showed a significant correlation coefficient only for Lyn (r’ = 0.71, P = 0.0008; Table 5). The scatter plots with trend lines for Lyn (with or without dasatinib) are presented in Figure 1 and for the other PKs in Supporting Information Appendix S5.
Table 5.
Results of Pearson correlation coefficients with (r) or without dasatinib (r’; with P‐values) between reporting odds ratios and product of dissociation constants (pKd; threshold P‐value adjusted with Bonferroni's correction = 0.0045)
| PK | r | P‐value | r’ | P‐value | |
|---|---|---|---|---|---|
| 1 | TEC | 0.95 | <0.0001 | 0.50 | 0.0347 |
| 2 | Lyn | 0.73 | 0.0004 | 0.71 | 0.0008 |
| 3 | BTK | 0,72 | 0.0005 | 0.61 | 0.0067 |
| 4 | Lck | 0.62 | 0.0042 | 0.61 | 0.0073 |
| 5 | HER3 | 0.53 | 0.0193 | 0.58 | 0.0114 |
| 6 | DDR1 | 0.37 | 0.1131 | 0.12 | 0.6354 |
| 7 | kit | 0.37 | 0.1206 | –0.018 | 0.9432 |
| 8 | HER4 | 0.37 | 0.1181 | 0.39 | 0.1050 |
| 9 | PDGFR‐Β | 0.35 | 0.1397 | 0.03 | 0.9036 |
| 10 | HER1 | 0.11 | 0.6358 | 0.19 | 0.4513 |
| 11 | HER2 | 0.10 | 0.6762 | 0.06 | 0.8135 |
BTK, Bruton agammaglobulinemia tyrosine kinase; DDR1, discoidin domain receptor family, member 1; EGF, epidermal growth factor; HER, human epidermal growth factor receptor; c‐kit, proto‐oncogene tyrosine kinase; Lck lymphocyte‐specific protein tyrosine kinase; Lyn, v‐yes‐1 Yamaguchi sarcoma viral related oncogene homologue; PDGFR‐β, platelet‐derived growth factor receptor β; PK, protein kinase; c‐SRC, v‐src sarcoma viral oncogene homologue; yes, Yamaguchi sarcoma viral related oncogene homologue
Figure 1.
(A) Scatter plots with trend line between RORs of pleural disorder and PKI affinity (pKd) for protein kinase Lyn. (B) Scatter plots with trend line between RORs of pleural disorder and PKI's affinity (pKd) for protein kinase Lyn (sensitivity analysis without dasatinib) Lyn, v‐yes‐1 Yamaguchi sarcoma viral related oncogene homolog; pKd, product of the dissociation constant; PKI, protein kinase inhibitor; ROR, reporting odds ratio
Discussion
Main results
Disproportionality analysis showed that 29 of 33 PKIs have significant RORs for PD. Dasatinib (ROR = 114.27, 95% CI: 109.11–119.66) is by far the highest value followed by bosutinib (ROR = 20.45, 95% CI: 15.83–26.43) and ponatinib (ROR = 12.00, 95% CI: 9.23–15.61). Based on the pharmacological hypothesis, we identified 11 PKs possibly involved in PDs in the literature review. In the analysis of the relationship between pKd and ROR, we observed four suspected proteins kinases: (i) TEC (r = 0.95, P < 0.0001); (ii) Lyn (r = 0.73, P = 0.0004); (iii) Btk (r = 0.72, P = 0.0005); and (iv) Lck (r = 0.62, P = 0.0042). Finally, after sensitivity analysis performed [with (r) and without (r’) dasatinib] a significant correlation was maintained for the proteins kinase Lyn (r = 0.73, P = 0.0004 and r’ = 0.71, P = 0.0008).
These results demonstrate some discordance with data in the literature regarding the potential for the PD induction by PKIs. Indeed, data have been found for dasatinib 17, 18, 19, 20, 21, 22, 23, 24, imatinib 4, 25, 26, 27, 28, erlotinib 29, nilotinib 30, 31, sunitinib 32 and bosutinib 33, occurrence of PD is mentioned in only seven summary product characteristics among the 33 PKIs of our study that included bosutinib, dasatinib, imatinib, midostaurin, nilotinib, ponatinib and sunitinib. However, PKIs ranking with the highest RORs were those with data listed in literature, expect for ibrutinib, afatinib and alectinib, which were ranked fourth, seventh and eighth, respectively, for which no data were found.
Description of PDs
Dasatinib dose schedule is risk factor for PD 18. Indeed, posology of 100 mg once daily had a better safety profile with the same efficacy of higher doses; the incidence of pleural effusion was also reduced and, therefore, 100 mg once daily became the recommended initial dosage. Our results showed that for PD reports, the median dose of dasatinib was 100 mg once daily, which is consistent with current recommendations in CML (a majority indication found in our study) but suggesting that the dosage regimen is not yet optimized to reduce the occurrence of PD with dasatinib. This is supported by Rousselot et al. who conducted a study with a therapeutic drug monitoring strategy for dasatinib dose in CML patients during first‐line therapy 34. The incidence of PD decreased from 45% to 11% at five years, and the mean dose of dasatinib after adaptation was 50 mg once daily without efficacy loss.
Numerous missing data were associated with the time of onset and outcome of PD (>60%). On average, PD appears 142 days after the initiation of the PKI, but our results were heterogeneous and depended on the PKI. Comparisons with data from the literature can be made only with dasatinib. In our study, a delayed PD onset with dasatinib was found to be an average of 449 days ( ± 568), which was higher than that found in the literature [median of 5 weeks (1–107); median 179 days (20–756)] 24, 34. Only the study of Porkka et al. found a median time to PD onset of 315 days (136–289) with an increased PD incidence between the 24th and 36th months of treatment 18.
ICSRs with PD were more serious than other ADRs (80.84% vs. 61.23%; P < 0.001). The PD‐related deaths (15.26%) occurring with PKIs were indicated in lung cancer (afatinib (7.01%), crizotinib (5.33%), erlotinib (4.84%) and gefitinib (4.41%), except for imatinib (5.20%), which was indicated in CML and sunitinib (5.94%), which was indicated in kidney cancer. This may reflect greater symptom severity with these molecules but also the presence of another aetiology such as the progression of the underlying pathology. Other risk factors for PD such hypertension, history of cardiac diseases or autoimmune diseases, hypercholesterolaemia and/or a history of rash with dasatinib have been identified 19, 25, 34, 35, but the information gathered in VigiBase did not allow us to analyse these data.
Involvement of protein kinase Lyn in PDs
Our study failed to demonstrate a potential effect from http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1843, epidermal growth factor receptors http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=320 and protein kinase http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1804, for which no significant Pearson correlation coefficients were found. Our results refute one of the most widely held assumptions in the literature, which is the involvement of PDGFR‐β in PD pathogenesis of PD 18, 25, 35, 36, 37. Our results show a strong linear correlation for protein kinases of the TEC family tyrosine kinases, including TEC (r = 0.95, P < 0.0001) and BTK (r = 0.72, P = 0.0005). These proteins are recognized as essential antigen receptor signalling mediators in lymphocytes 38 and may be involved in PD pathogenesis via an immune‐mediated mechanism 35. However, the relationship between RORs and pKd is not conclusive for these targets because the slope of the trend lines may have influenced by the outlier dasatinib point (Supporting Information Appendix S5). Sensitivity analysis, with the exclusion of dasatinib, showed a decrease in the correlation coefficients for TEC and BTK (rTEC = 0.50 and rBTK = 0.61, respectively), which have become non‐significant (Table 5). These results suggest that BTK and TEC do not seem to be involved in a common pathogenesis PD development with PKIs.
Finally, after the sensitivity analysis was performed (with and without dasatinib), a significant correlation was maintained for protein kinase Lyn (r = 0.73, P = 0.0004). The slopes of the trend lines were positive, indicating a positive association between RORs and pKd values (Figure 1). The scatters had a relatively balanced spread. The relationship between RORs and pKd value seemed to be linear, even though some of the outlier points may have influenced the shape of the line as dasatinib.
The PK Lyn belongs to the Src‐family of tyrosine kinases, whose members play critical roles in a variety of cellular signal transduction pathways, regulating such diverse processes as cell division, motility, adhesion, angiogenesis and survival 39. Specifically, Lyn tyrosine kinase is expressed in B and T lymphocytes and may be involved in the pathogenesis of auto‐immune diseases such systemic lupus erythematosus: in these patients, reduced activity of Lyn protein was found 40. Interestingly, some patients with CML treated with PKIs (imatinib, dasatinib or nilotinib) can develop T cell expansion as an oligoclonal expansion of large granular lymphocytes 41. These clonal expansions concern CD56+ lymphocytes (natural killer) or CD3+/CD8+ lymphocytes (cytotoxic T lymphocytes), and Lyn inhibition may modulate these immune cells activation 42, 43.
This clonal large granular lymphocyte expansion have a cytotoxic character on endothelial cells and may be responsible for some ADRs such as fever, colitis, cytomegalovirus reactivation, pulmonary arterial hypertension or PD 43, 44. Analysis of lymphocytes from pleural fluid showed a similar phenotype (CD3+, CD8+, CD57+) and genotype (identical T‐cell clone) as concurrently observed in peripheral blood. Thus, PD induced by PKIs may be part of an antihost response driven by expansion of cytotoxic T and natural killer cells caused by inhibition of Lyn 35, 43.
Study strengths
RORs were calculated using VigiBase data, which is the most important PV database worldwide. The ROR is a reproducible tool for evaluating PV disproportionality. We performed a thorough literature review to select the most probable PKs involved in drug‐induced PD and we assumed the existence of a pharmacological hypothesis concerning the mechanism of PD occurrence. Indeed, a clinical interpretation is always necessary to exclude random or biased signals. The binding properties of a PKI have been screened in in vitro competitive binding assays 9. For each PKI, Kd values were determined using 11 serial three‐fold dilutions and a dimethyl sulfoxide control. This method, used to measure affinity, is accurate and reproducible. We chose to focus our analysis on the tested PKIs using this specific method to avoid bias and errors caused by a change in measurement methods. We gave priority to the most significantly published interaction map for PKs.
Study limitations
PV programs are mostly based on concurrent reporting systems. Consequently, accuracy and the amount of information reported in cases may not be optimal. This could have led to misclassification of PD cases. PV data also suffer from other biases such as underreporting, halo bias, and a lack of information on the exposed population and sales data for the drugs. Indeed, PD may have a tumour aetiology with an over‐risk in haematological malignancies and lung and breast cancers. These pathologies are the indications most represented by PKIs in our study and we could not ensure exclusion of another non‐drug aetiology. Not being able to return to each notification to ensure that an exhaustive search for aetiologies has been carried out leads to an information bias. Also, in the present study, disproportionality was used as a very early proxy of the relative risk although this relationship is not clear 45. The screening study of Davis et al. 9 was published in 2011, and only 19 out of the 33 PKIs registered in VigiBase included data to perform the correlation analysis. A more substantial number of points would have made it possible to increase the correlation model robustness. The model used in the analyses to link PV and pharmacodynamics data was based on three hypotheses: (i) PD of a PKI was caused by a single PK. The model was not able to detect coinhibition of multiple PKs or inhibition/activation of non‐PK cellular targets (such as proteasomes, G protein‐coupled receptors, voltage‐gated ion channels or ligand‐gated ion channels). Some authors suggest that PKI could cause PD by their ability to inhibit multiple targets as PDGFR‐β and SRC family tyrosine kinases 36. Considering the interactions between PK via an adapted statistical model would have modelled the physiological reality; (ii) we assumed that this mechanism was similar for all the selected PKIs even though there is heterogeneity in this family; and (iii) the relationship between ROR and pKd is simple and straightforward. The relationship between these two parameters may not be linear and, thus, the model would not be able to highlight the molecular targets of interest under these conditions. Accordingly, we observed some heterogeneity within the scatter plots between disproportionality and the affinity of PKIs for Lyn. This difference might have been caused by the uncertainty of measuring two measured variables but also the inadequacy of the model for these drugs. Other cellular mechanisms for PD associated with some PKIs cannot be excluded.
Conclusion
The present study was the first to investigate the disproportionality of PD induced by PKIs in the WHO global safety database. We observed a disproportionality for 29 PKI out of the 33 included in our study. This PV signal remains to be confirmed by appropriate pharmaco‐epidemiological studies to inform health professionals and patients about the real PKI risks. Also, we used an approach to combine our PV and pharmacodynamics knowledge in order to identify ADR‐related cellular targets. This study highlights the role of Lyn in PD caused by PKIs through an immune‐mediated process.
Identification of cellular targets with PV and pharmacodynamics knowledge of approved drugs could allow very early prediction of new drug toxicity, according to the affinity data measured during the preclinical developmental stage.
Competing Interests
There are no competing interests to declare.
This work received support from the National Research Agency [Agence Nationale de la Recherche (ANR)] for the investissement d'avenir (Investment in the Future; ANR‐11‐PHUC‐001).
Supporting information
Appendix S1 Dissociation constants (Kd in nM) and product of dissociation constants (pKd) of the 19 PKI for the 13 protein kinases involved in pleural disorder according to study of Davis et al. [9]
Appendix S2 Algorithm used for the literature search in Medline database (Mesh terms with the PubMed tool) and Web of science
Appendix S3 Description of Individual Safety Reports (ICSRs) from VigiBase® of Pleural Disorders (PD) with 33 protein kinase inhibitors
Appendix S4 Disproportionality analysis for pleural disorder with 32 protein kinase inhibitors (out dasatinib), showing the reporting odds ratios (RORs) with confidence intervals, by descending order
Appendix S5 Scatter plots with trend lines between ROR of pleural disorder and PKI's affinity (pKd) for the 11 proteins kinase of our study. pKd, product of the dissociation constant; PKI, protein kinase inhibitor; ROR, reporting odds ratio
Mahé, J. , de Campaigno, E. P. , Chené, A.‐L. , Montastruc, J.‐L. , Despas, F. , and Jolliet, P. (2018) Pleural adverse drugs reactions and protein kinase inhibitors: Identification of suspicious targets by disproportionality analysis from VigiBase. Br J Clin Pharmacol, 84: 2373–2383. 10.1111/bcp.13693.
References
- 1. Manning G, Whyte DB, Martinez R, Hunter T, Sudarsanam S. The protein kinase complement of the human genome. Science 2002; 298: 1912–1934. [DOI] [PubMed] [Google Scholar]
- 2. Yaish P, Gazit A, Gilon C, Levitzki A. Blocking of EGF‐dependent cell proliferation by EGF receptor kinase inhibitors. Science 1988; 242: 933–935. [DOI] [PubMed] [Google Scholar]
- 3. Karaman MW, Herrgard S, Treiber DK, Gallant P, Atteridge CE, Campbell BT, et al A quantitative analysis of kinase inhibitor selectivity. Nat Biotechnol 2008; 26: 127–132. [DOI] [PubMed] [Google Scholar]
- 4. Druker BJ, Talpaz M, Resta DJ, Peng B, Buchdunger E, Ford JM, et al Efficacy and safety of a specific inhibitor of the BCR‐ABL tyrosine kinase in chronic myeloid leukemia. N Engl J Med 2001; 344: 1031–1037. [DOI] [PubMed] [Google Scholar]
- 5. Siehl J, Thiel E. C‐kit, GIST, and imatinib. Recent Results Cancer Res Fortschritte Krebsforsch Progres Dans Rech Sur Cancer 2007; 176: 145–151. [DOI] [PubMed] [Google Scholar]
- 6. Cheng H, Force T. Why do kinase inhibitors cause cardiotoxicity and what can be done about it? Prog Cardiovasc Dis 2010; 53: 114–120. [DOI] [PubMed] [Google Scholar]
- 7. Patras de Campaigno E, Bondon‐Guitton E, Laurent G, Montastruc F, Montastruc JL, Lapeyre‐Mestre M, et al Identification of cellular targets involved in cardiac failure caused by PKI in oncology: an approach combining pharmacovigilance and pharmacodynamics. Br J Clin Pharmacol 2017; 83: 1544–1555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Lindquist M. VigiBase, the WHO global ICSR database system: basic facts. Drug Inf J 2008; 42: 409–419. [Google Scholar]
- 9. Davis MI, Hunt JP, Herrgard S, Ciceri P, Wodicka LM, Pallares G, et al Comprehensive analysis of kinase inhibitor selectivity. Nat Biotechnol 2011; 29: 1046–1051. [DOI] [PubMed] [Google Scholar]
- 10. Edwards IR, Aronson JK. Adverse drug reactions: definitions, diagnosis, and management. Lancet Lond Engl 2000; 356: 1255–1259. [DOI] [PubMed] [Google Scholar]
- 11. Montastruc J‐L, Sommet A, Bagheri H, Lapeyre‐Mestre M. Benefits and strengths of the disproportionality analysis for identification of adverse drug reactions in a pharmacovigilance database. Br J Clin Pharmacol 2011; 72: 905–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Chavant F, Favrelière S, Lafay‐Chebassier C, Plazanet C, Pérault‐Pochat MC. Memory disorders associated with consumption of drugs: updating through a case/noncase study in the French PharmacoVigilance database. Br J Clin Pharmacol 2011; 72: 898–904. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Maignen F, Hauben M, Hung E, Van Holle L, Dogne JM. Assessing the extent and impact of the masking effect of disproportionality analyses on two spontaneous reporting systems databases. Pharmacoepidemiol Drug Saf 2014; 23: 195–207. [DOI] [PubMed] [Google Scholar]
- 14. Harding SD, Sharman JL, Faccenda E, Southan C, Pawson AJ, Ireland S, et al The IUPHAR/BPS guide to PHARMACOLOGY in 2018: updates and expansion to encompass the new guide to IMMUNOPHARMACOLOGY. Nucl Acids Res 2018; 46: D1091–D1106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Alexander SPH, Fabbro D, Kelly E, Marrion NV, Peters JA, Faccenda E, et al The Concise Guide to PHARMACOLOGY 2017/18: Enzymes. Br J Pharmacol 2017; 174: S272–S359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Alexander SPH, Fabbro D, Kelly E, Marrion NV, Peters JA, Faccenda E, et al The Concise Guide to PHARMACOLOGY 2017/18: Catalytic receptors. Br J Pharmacol 2017; 174: S225–S271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Apperley JF, Cortes JE, Kim D‐W, Roy L, Roboz GJ, Rosti G, et al Dasatinib in the treatment of chronic myeloid leukemia in accelerated phase after imatinib failure: the START a trial. J Clin Oncol 2009; 27: 3472–3479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Porkka K, Khoury HJ, Paquette RL, Matloub Y, Sinha R, Cortes JE. Dasatinib 100 mg once daily minimizes the occurrence of pleural effusion in patients with chronic myeloid leukemia in chronic phase and efficacy is unaffected in patients who develop pleural effusion. Cancer 2010; 116: 377–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Kantarjian H, Pasquini R, Lévy V, Jootar S, Holowiecki J, Hamerschlak N, et al Dasatinib or high‐dose imatinib for chronic‐phase chronic myeloid leukemia resistant to imatinib at a dose of 400 to 600 milligrams daily: two‐year follow‐up of a randomized phase 2 study (START‐R). Cancer 2009; 115: 4136–4147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Kantarjian H, Shah NP, Hochhaus A, Cortes J, Shah S, Ayala M, et al Dasatinib versus imatinib in newly diagnosed chronic‐phase chronic myeloid leukemia. N Engl J Med 2010; 362: 2260–2270. [DOI] [PubMed] [Google Scholar]
- 21. Hochhaus A, Kantarjian HM, Baccarani M, Lipton JH, Apperley JF, Druker BJ, et al Dasatinib induces notable hematologic and cytogenetic responses in chronic‐phase chronic myeloid leukemia after failure of imatinib therapy. Blood 2007; 109: 2303–2309. [DOI] [PubMed] [Google Scholar]
- 22. Cortes JE, Saglio G, Kantarjian HM, Baccarani M, Mayer J, Boqué C, et al Final 5‐year study results of DASISION: the Dasatinib versus Imatinib study in treatment‐Naïve chronic myeloid leukemia patients trial. J Clin Oncol 2016; 34: 2333–2340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Bergeron A, Réa D, Levy V, Picard C, Meignin V, Tamburini J, et al Lung abnormalities after Dasatinib treatment for chronic myeloid leukemia. Am J Respir Crit Care Med 2007; 176: 814–818. [DOI] [PubMed] [Google Scholar]
- 24. de Lavallade H, Punnialingam S, Milojkovic D, Bua M, Khorashad JS, Gabriel IH, et al Pleural effusions in patients with chronic myeloid leukaemia treated with dasatinib may have an immune‐mediated pathogenesis. Br J Haematol 2008; 141: 745–747. [DOI] [PubMed] [Google Scholar]
- 25. Breccia M, D'Elia GM, D'Andrea M, Latagliata R, Alimena G. Pleural‐pericardic effusion as uncommon complication in CML patients treated with Imatinib. Eur J Haematol 2005; 74: 89–90. [DOI] [PubMed] [Google Scholar]
- 26. Banka R, Udwadia Z. Imatinib‐induced pleural effusion: a case report. J Postgrad Med 2017; 63: 55–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Bongiovanni A, Ricci M, Riva N, Calpona S, Oboldi D, Pieri F, et al Pleural effusion in a patient with metastatic gastrointestinal stromal tumor treated with imatinib: case report. Future Oncol Lond Engl 2014; 10: 2423–2427. [DOI] [PubMed] [Google Scholar]
- 28. Posadas EM, Kwitkowski V, Kotz HL, Espina V, Minasian L, Tchabo N, et al A prospective analysis of imatinib‐induced c‐KIT modulation in ovarian cancer: a phase II clinical study with proteomic profiling. Cancer 2007; 110: 309–317. [DOI] [PubMed] [Google Scholar]
- 29. Tachihara M, Misa K, Uematsu M, Minemura H, Katsuura Y, Ishida T, et al Increase of ascites and pleural effusion misleading assessment of antitumor response to erlotinib in adenocarcinoma of the lung. J Clin Oncol Off J Am Soc Clin Oncol 2011; 29: e675–e677. [DOI] [PubMed] [Google Scholar]
- 30. Chakraborty K, Bossaer JB, Patel R, Krishnan K. Successful treatment of nilotinib‐induced pleural effusion with prednisone. J Oncol Pharm Pract 2013; 19: 175–177. [DOI] [PubMed] [Google Scholar]
- 31. Ustun C, Randall N, Podgaetz E, Amin K, Dincer HE. Severe dyspnoea in a patient with chronic myelogenous leukaemia on a tyrosine kinase inhibitor. Thorax 2015; 70: 701–704. [DOI] [PubMed] [Google Scholar]
- 32. Kust D, Kruljac I, Peternac AŠ, Ostojić J, Prpić M, Čaržavec D, et al Pleural and pericardial effusions combined with ascites in a patient with severe sunitinib‐induced hypothyroidism. Acta Clin Belg 2016; 71: 175–177. [DOI] [PubMed] [Google Scholar]
- 33. Riou M, Seferian A, Savale L, Chaumais MC, Guignabert C, Canuet M, et al Deterioration of pulmonary hypertension and pleural effusion with bosutinib following dasatinib lung toxicity. Eur Respir J 2016; 48: 1517–1519. [DOI] [PubMed] [Google Scholar]
- 34. Rousselot P, Mollica L, Etienne G, Bouchet S, Guerci A, Nicolini FE, et al Pharmacologic monitoring of Dasatinib as first line therapy in newly diagnosed chronic phase chronic Myelogenous leukemia (CP‐CML) identifies patients at higher risk of pleural effusion: a sub‐analysis of the OPTIM‐Dasatinib trial. Blood 2012; 120: 3770. [Google Scholar]
- 35. Kelly K, Swords R, Mahalingam D, Padmanabhan S, Giles FJ. Serosal inflammation (pleural and pericardial effusions) related to tyrosine kinase inhibitors. Target Oncol 2009; 4: 99–105. [DOI] [PubMed] [Google Scholar]
- 36. Quintás‐Cardama A, Kantarjian H, O'brien S, Borthakur G, Bruzzi J, Munden R, et al Pleural effusion in patients with chronic myelogenous leukemia treated with dasatinib after imatinib failure. J Clin Oncol Off J Am Soc Clin Oncol 2007; 25: 3908–3914. [DOI] [PubMed] [Google Scholar]
- 37. Giles FJ, O'Dwyer M, Swords R. Class effects of tyrosine kinase inhibitors in the treatment of chronic myeloid leukemia. Leukemia 2009; 23: 1698–1707. [DOI] [PubMed] [Google Scholar]
- 38. Berg LJ, Finkelstein LD, Lucas JA, Schwartzberg PL. Tec family kinases in T lymphocyte development and function. Annu Rev Immunol 2005; 23: 549–600. [DOI] [PubMed] [Google Scholar]
- 39. Summy JM, Gallick GE. Src family kinases in tumor progression and metastasis. Cancer Metastasis Rev 2003; 22: 337–358. [DOI] [PubMed] [Google Scholar]
- 40. Jenks SA, Sanz I. Altered B cell receptor signaling in human systemic lupus erythematosus. Autoimmun Rev 2009; 8: 209–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Powers JJ, Dubovsky JA, Epling‐Burnette PK, Moscinski L, Zhang L, Mustjoki S, et al A molecular and functional analysis of large granular lymphocyte expansions in patients with chronic myelogenous leukemia treated with tyrosine kinase inhibitors. Leuk Lymphoma 2011; 52: 668–679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Paydas S. Dasatinib, large granular lymphocytosis, and pleural effusion: useful or adverse effect? Crit Rev Oncol Hematol 2014; 89: 242–247. [DOI] [PubMed] [Google Scholar]
- 43. Mustjoki S, Ekblom M, Arstila TP, Dybedal I, Epling‐Burnette PK, Guilhot F, et al Clonal expansion of T/NK‐cells during tyrosine kinase inhibitor dasatinib therapy. Leukemia 2009; 23: 1398–1405. [DOI] [PubMed] [Google Scholar]
- 44. Minami M, Arita T, Iwasaki H, Muta T, Aoki T, Aoki K, et al Comparative analysis of pulmonary hypertension in patients treated with imatinib, nilotinib and dasatinib. Br J Haematol 2017; 177: 578–587. [DOI] [PubMed] [Google Scholar]
- 45. Maciá‐Martínez MA, de Abajo FJ, Roberts G, Slattery J, Thakrar B, Wisniewski AFZ. An empirical approach to explore the relationship between measures of disproportionate reporting and relative risks from analytical studies. Drug Saf 2016; 39: 29–43. [DOI] [PubMed] [Google Scholar]
- 46. Rodt SA, Ahlén K, Berg A, Rubin K. Reed RK. A novel physiological function for platelet‐derived growth factor‐BB in rat dermis. J Physiol 1996; 495 (Pt 1): 193–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Han J, Zhang G, Welch EJ, Liang Y, Fu J, Vogel SM, et al A critical role for Lyn kinase in strengthening endothelial integrity and barrier function. Blood 2013; 122: 4140–4149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Mano H. Tec family of protein‐tyrosine kinases: an overview of their structure and function. Cytokine Growth Factor Rev 1999; 10: 267–280. [DOI] [PubMed] [Google Scholar]
- 49. Lowell CA. Src‐family kinases: rheostats of immune cell signaling. Mol Immunol 2004; 41: 631–643. [DOI] [PubMed] [Google Scholar]
- 50. Eskazan AE, Eyice D, Kurt EA, Elverdi T, Yalniz FF, Salihoglu A, et al Chronic myeloid leukemia patients who develop grade I/II pleural effusion under second‐line dasatinib have better responses and outcomes than patients without pleural effusion. Leuk Res 2014; 38: 781–787. [DOI] [PubMed] [Google Scholar]
- 51. Krauth M‐T, Herndlhofer S, Schmook M‐T, Mitterbauer‐Hohendanner G, Schlögl E, Valent P. Extensive pleural and pericardial effusion in chronic myeloid leukemia during treatment with dasatinib at 100 mg or 50 mg daily. Haematologica 2011; 96: 163–166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Thachil J. Adverse effects of imatinib – dissecting heart from the rest. Cancer 2011; 117: 228 author reply 229. [DOI] [PubMed] [Google Scholar]
- 53. Sakamoto O, Suga M, Suda T, Ando M. Expression of discoidin domain receptor 1 tyrosine kinase on the human bronchial epithelium. Eur Respir J 2001; 17: 969–974. [DOI] [PubMed] [Google Scholar]
- 54. van Cruijsen H, Giaccone G, Hoekman K. Epidermal growth factor receptor and angiogenesis: opportunities for combined anticancer strategies. Int J Cancer 2005; 117: 883–888. [DOI] [PubMed] [Google Scholar]
- 55. Pneumotox [Internet] . The drug‐induced respiratory disease website. Available at http://www.pneumotox.com (last accessed 19 May 2017).
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1 Dissociation constants (Kd in nM) and product of dissociation constants (pKd) of the 19 PKI for the 13 protein kinases involved in pleural disorder according to study of Davis et al. [9]
Appendix S2 Algorithm used for the literature search in Medline database (Mesh terms with the PubMed tool) and Web of science
Appendix S3 Description of Individual Safety Reports (ICSRs) from VigiBase® of Pleural Disorders (PD) with 33 protein kinase inhibitors
Appendix S4 Disproportionality analysis for pleural disorder with 32 protein kinase inhibitors (out dasatinib), showing the reporting odds ratios (RORs) with confidence intervals, by descending order
Appendix S5 Scatter plots with trend lines between ROR of pleural disorder and PKI's affinity (pKd) for the 11 proteins kinase of our study. pKd, product of the dissociation constant; PKI, protein kinase inhibitor; ROR, reporting odds ratio