The GSE93984 profile (https://www.ncbi.nlm.nih.gov/pubmed/28428442) and its corresponding platform annotation files were downloaded from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) (31). This dataset consisted of 34 samples. Differential expression of genes in ibrutinib-responsive [complete response (CR; 10 cases) + partial response (PR; 14 cases)] and non-responsive [stable disease (SD; 10 cases)] of DLBCL was tested with cut-off criteria of P<0.05 and fold change (FC)|>2. Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) (https://www.genome.jp/kegg/kegg1.html) (32–34) were performed with the Database for Annotation, Visualization and Integrated Discovery (version 6.8; http://david.ncifcrf.gov/) (35,36). Through The Search Tool for Interactions of Chemicals (STITCH) database (version 4.0; http://stitch.embl.de/) (37,38), the genes interacting with ibrutinib were extracted and visualized. The Search Tool for the Retrieval of Interacting Genes (STRING) (version 11.0), which provides information for experimental and predicted interactions, is an online database (39).

Between April 2013 and March 2015, 62 patients who were histologically diagnosed with DLBCL in the Fudan University Shanghai Cancer Center were investigated (Table I). A total of 29 women (46.8%) and 33 men (53.2%) were included. The mean age was 50 years (range, 20–77 years). The inclusion criteria were as follows: Each patient was diagnosed with DLBCL by pathology, received ibrutinib therapy alone and provided informed consent. Cancer tissue samples were then collected from these patients by means of biopsy. DLBCL tissues were defined as showing a PR or CR following ibrutinib treatment, and the resistant DLBCL tissues as showing relapsed/refractory disease. This study was approved by the Research Ethics Committee of Fudan University Shanghai Cancer Center.

The TMD8 and HBL1 cell lines were obtained from the American Type Culture Collection. Cell line authentication was performed by the Cell Check service at IDEXX Laboratories, Inc. Cell lines in the logarithmic growth phase were cultured in RPMI 1640 medium containing 10% FBS (Atlanta Biologicals, Inc.; R&D Systems, Inc.), 1 mM sodium pyruvate and 1% penicillin/streptomycin (Pen/Strep); RPMI 1640 medium, sodium pyruvate and Pen/Strep were obtained from Thermo Fisher Scientific, Inc. Ibrutinib-resistant HBL1 and TMD8 cells were generated via in vitro culture of the parental cell lines for prolonged periods of time with progressively increasing concentrations of ibrutinib (0, 5, 10, 200, 500, 800 and 1,000 nM). These varying concentrations of ibrutinib were added to ibrutinib-resistant HBL1 and TMD8 cells for 24 h prior to the MTT assay. Then, 200 nM ibrutinib was added to ibrutinib-resistant HBL1 and TMD8 cells for 24 h prior to the flow cytometry assay. lentiviruses Lv-EGFR was obtained from Auragene Bioscience Corporation, Inc.

Cells were seeded in 6-well plates (1×10⁶/well). Total RNA was extracted using the TaqMan^® Fast Cells-to-CT™ kit (Thermo Fisher Scientific, Inc.) and then reverse transcribed into cDNA according to the manufacturer's instructions. quantitative PCR (qPCR) was performed on a QuantStudio 7 Flex Real-Time PCR System (Thermo Fisher Scientific, Inc.). Amplification was performed in a three-step cycle procedure, 95°C (denaturation) 10 sec, 60°C (annealing) 30 sec, and 72°C (extension) 30 sec, for 40 cycles. The primers were: PDGFD, forward 5′-GAACAGCTACCCCAGGAACC-3′, reverse 5′-CTTGTGTCCACACCATCGTC-3′; EGFR, forward 5′-CCCTCCTGAGCTCTCTGAGT-3′, reverse 5′-GTTTCCCCCTCTGGAGATGC-3′; β-actin, forward 5′-TTGTTACAGGAAGTCCCTTGCC-3′; reverse 5′-ATGCTATCACCTCCCCTGTGTG-3′. mRNA levels were quantified using the 2^−ΔΔCq method (40) and normalized to the internal reference gene β-actin. All experiments were repeated three times.

The TMD8 and HBL1 ibrutinib-resistant cells in the logarithmic growth phase were seeded in 6-well plates at a density of 3×10⁵ cells/well. The PDGF-D shRNA sequence was GCGCATCCATCAAAGCTTTGC. PDGFD short hairpin RNA (shRNA; 20 pmol) and pHBLV-U6-Puro lentiviruses (20 pmol; Auragene Bioscience Corporation, Inc.) were added to the cells for 24 h in the presence of Polybrene (Santa Cruz Biotechnology, Inc.; 5 µg/ml) after the cells had adhered to the walls of the plates. Following infection for 48 h, cells were observed under a fluorescence microscope (IX70; Olympus Corporation), and uninfected cells were killed with puromycin. The surviving cells were collected, and the PDGFD protein level was analyzed using western blotting.

Cells were seeded into 96-well flat-bottomed tissue culture plates (5–10×10³/well in 100 µl medium). The cells were cultured for 24–96 h at 37°C under 5% CO₂ before a total of 20 µl MTT (5 mg/ml) was added to cells in the logarithmic growth phase of each group for 4 h. Dimethyl sulfoxide were added to dissolve purple formazan for 10 min. The optical density (OD) value at 490 nm was measured with a microplate reader (Bio-Rad Laboratories, Inc.). The cell survival rate was calculated with a concentration-survival curve.

Tissue samples from DLBCL and ibrutinib-resistant DLBCL areas were formalin-fixed, dehydrated, cleared in xylene and embedded in paraffin. Sections (5 µm) were deparaffinized, hydrated, and 3% H₂O₂ solution was added for 15 min to remove endogenous catalase and antigen repair. Non-immune normal goat serum was incubated at room temperature for 60 min at 100 µl and then stained with anti-PDGFD in 4°C overnight (1:100, cat. no. ab181845; Abcam). Horseradish peroxidase conjugated secondary antibodies were incubated at room temperature for 30 min (1:1,000, Pv-80000, OriGene Technologies, Inc.), and 3,3-diaminobenzidine substrate was added for the development of immunostaining according to the manufacturer's instructions (Dako; Agilent Technologies, Inc.). Then the slides were counterstained with hematoxylin and eosin at room temperature for 5 min. Positive cells were counted in 10 randomly selected fields with a ×40 objective (Olympus CX23; Olympus Corporation).

Western blotting was performed as previously described (41). In brief, cells were harvested and lysed with RIPA buffer (cat. no. R0278; Sigma-Aldrich; Merck KGaA) containing 1X protease/phosphatase inhibitor. A BCA protein assay kit (Beyotime) was employed to measure the protein concentrations. Equal amounts (20 µg/well) of protein were separated by 10% SDS-PAGE and transferred to PVDF membranes. The membranes were washed by 1X TBST and blocked by Odyssey Blocking Buffer (cat. no. 927-40000; LI-COR Biosciences) for 1 h at room temperature. Then membranes were incubated with primary antibodies against PDGFD (1:1,000; cat. no. ab240960; Abcam), EGFR (1:1,000; cat. no. ab52894; Abcam) and GAPDH (1:2,000; cat. no. ab181602; Abcam) overnight at 4°C. Afterwards, membranes were washed and incubated with secondary antibodies for 1 h at room temperature. Signals were measured with a luminescent image analyzer (ImageQuant LAS4000 mini) and GAPDH served as a loading control.

The ApoDETECT Annexin V-FITC kit (Thermo Fisher Scientific, Inc.) was used to quantify the number of apoptotic cells in the indicated groups. Briefly, 1X binding buffer was used to resuspend cells (5×10⁵ cells/ml). Annexin V-FITC was added at room temperature for 10 min in the dark. Then, the cells were resuspended in 190 µl binding buffer containing 10 µl 20 µg/ml propidium iodide (PI) at 4°C for 30 min and analyzed with a flow cytometer (BD Biosciences) using ModFit LT software version 3.0 (Verity Software House, Inc.). The apoptotic rate was calculated as early apoptotic cells + late apoptotic cells.

Data are presented as the mean ± standard deviation using SPSS 17.0 software (SPSS, Inc.). The statistical significance between multiple experimental groups was analyzed using one-way ANOVA followed by Tukey's post hoc test. A Student's t-test was used for comparisons between two groups. P<0.05 was considered to indicate a statistically significant difference. Each experiment was repeated at least three times.

Through the GEO database, the GSE93984 gene expression profile was downloaded, and gene expression in patients treated with ibrutinib in the CR (10 cases), PR (14 cases) and SD (10 cases) groups were analyzed; 22,189 genes were identified. To determine the key genes affecting DLBCL, genes with increasing expression in the CR, PR and SD groups, and genes with FC>2.0 or <0.5, and P<0.05 were selected (Fig. 1A). GO classification analysis showed that DEGs were primarily involved in the negative regulation of growth and the immune response (Fig. 1B), and the gene network diagram of the ibrutinib interaction (Fig. 2A) was analyzed with the STITCH database. It was found that there was an interaction between EGFR and ibrutinib, and the known phase of DLBCL was found by searching the database and related literature. The STRING database showed that EGFR might be a downstream target gene of PDGFD (Fig. 2B).

First, to verify PDGFD expression in lymphoma drug-resistant and drug-sensitive tissues, as well as drug-resistant cell lines and parental cell lines, IHC was performed to detect the expression of PDGFD in these tissues. PDGFD was expressed in low levels in the cytoplasm of cells in the sensitive tissues. However, the expression of PDGFD in the stromal cells of lymphoma-resistant tissues was homogeneous and high (Fig. 3A). The expression of PDGFD in ibrutinib-resistant cell lines was higher than in the sensitive parental cell lines (Fig. 3B and C), as determined via qPCR and western blotting. The results indicated that PDGFD expression is elevated in both drug-resistant tissues and cell lines; thus, abnormal expression of PDGFD may be associated with ibrutinib resistance in lymphoma.

To further study the effect of PDGFD on the resistance of DLBCL cells to ibrutinib, a lentivirus carrying shPDGFD (Lv-shPDGFD) was constructed, and the TMD8-ibrutinib and HBL1-ibrutinib cell lines were infected. Western blotting was performed to verify the interference efficiency (Fig. 4A). MTT assays were then performed to calculate the IC₅₀ values of the resistant and the parental strains. The resistant strains exhibited higher IC₅₀ values than the parental strains (Fig. 4B). Conversely, after Lv-shPDGFD was used to infect the drug-resistant cells, it was found that silencing of PDGFD reduced the IC₅₀ values of ibrutinib in drug-resistant cells (Fig. 4C). Additionally, the apoptotic rate increased in ibrutinib-resistant cell lines following Lv-shPDGFD infection (Fig. 4D). The data suggested that PDGFD silencing increased the sensitivity of drug-resistant strains to PDGFD, indicating that PDGFD is related to the resistance of DLBCL cells to ibrutinib.

To compare EGFR expression between the drug-resistant cell lines and parental cell lines, qPCR was used to measure EGFR expression. Compared with the parental cell lines, the expression of EGFR mRNA in the drug-resistant cell lines was upregulated (Fig. 5A). In addition, western blotting revealed that the expression of EGFR protein in the drug-resistant cell lines was higher than that in the parental cell lines, which was consistent with the qPCR results (Fig. 5B). Compared with that in the NC group, the expression of PDGFD mRNA in the Lv-shPDGFD group was downregulated, which was accompanied by decreased EGFR expression (Fig. 5C), suggesting that EGFR was a downstream target of PDGFD. Finally, Lv-shPDGFD and an EGFR overexpression lentivirus (Lv-EGFR; Fig. 5D) were coinfected into drug-resistant cells to determine whether EGFR overexpression reversed the IC₅₀ changes induced by PDGFD silencing. Compared with that in the Lv-shPDGFD-infected cells, the IC₅₀ for ibrutinib in the Lv-shPDGFD + Lv-EGFR-infected cells was increased (Fig. 5E and F), suggesting that the overexpression of EGFR could reverse the sensitization of DLBCL to ibrutinib induced by PDGFD interference in the drug-resistant strains. Collectively, the data suggested that PDGFD could induce DLBCL ibrutinib resistance by regulating EGFR expression.