The adhesion molecule ICAM-1 in diffuse large B-cell lymphoma post-rituximab era: relationship with prognostic importance and rituximab resistance

Demographic and baseline patient characteristics

Among 102 patients who were all hospitalized, median age was 52 years (range, 19 to 72 years) at the time of diagnosis and 52 (51%) were male. According to Ann Arbor staging criteria, 67 patients (65.7%) had stage I/II diseases, and 35 patients (34.3%) had stage III/IV diseases. 80 patients (78.4%) were IPI 0 or 1 and 22 patients (21.6%) were IPI 2 or more. Eastern Cooperative Oncology Group (ECOG) scores of patients were 0-1. Concerning of cell of origin, 41 patients were germinal center B-cell lymphoma (GCB) and 56 patients were non-germinal center B-cell lymphoma (non-GCB) according to Han’s algorithm. 59 patients (57.8%) received R-CHOP and 43 patients (42.2%) received CHOP regimen as the first-line treatment. Basic features of the patients are recorded concerning the clinical and pathological parameters (Table 1).

ICAM-1 expression in lymphoma tissue samples

ICAM-1 expression was observed on lymphoma cell membrane. The protein highly expressed in 28 (27.5%) cases. Image Figure 1A, 1B represented negative and positive expression of ICAM-1. Overexpression of ICAM-1 was observed more frequently in patients with GCB subtype than non-GCB subtype (41.5% vs. 19.6%, P=0.019) (Figure 1C).

Figure 1. ICAM-1 expression in DLBCL patients. Representative immunohistochemical staining of ICAM-1 with negative (A) and positive (B) expression. (C) Overexpression of ICAM-1 in GCB and non-GCB subtypes.

The association of ICAM-1 expression with ORR, PFS, and OS in CHOP and R-CHOP treated groups

Previous studies reported that lower ICAM-1 expression correlated with inferior prognosis in non-Hodgkin's lymphoma before rituximab era [19]. Our study showed results consistent with previous studies, that CHOP patients with lower expression of ICAM-1 had a worse PFS (61% vs. 100%, P = 0.03), and a worse OS (64% vs. 100%, P=0.05) compared to those with higher expression of ICAM-1 by Kaplan-Meier analysis (Figure 2A, 2B).

Figure 2. Progression-free survival (PFS) and overall survival (OS) of DLBCL patients in our study. Kaplan-Meier curves showing the association between ICAM-1 expression and PFS (A), OS (B) of DLBCL patients treated with CHOP regimen. R-CHOP could significantly improve PFS (C) and OS (D) in ICAM-1 negative expression patient compared to CHOP patients. In ICAM-1 positive patient, no statistical difference of PFS (E) and OS (F) between R-CHOP and CHOP group. All the P values are shown in the graph, by log-rank test.

To our surprise, in post-rituximab era, R-CHOP significantly improved PFS and OS in ICAM-1 negative expression patients compared to CHOP patients (P=0.01, P=0.02; respectively) (Figure 2C, 2D). However, in the ICAM-1 positive patient, there was no statistical difference of PFS and OS between the two groups (Figure 2E, 2F).

As for overall response rate, in ICAM-1 positive group, there was no statistical difference between patients treated with CHOP or R-CHOP regimen (77.7% vs. 89.4%, P=0.409); however, in ICAM-1 negative group, an improved ORR was found in patients treated with R-CHOP compared to those with CHOP regimen (92.5% vs. 73.5%, P=0.019) (Table 2).

Downregulation of ICAM-1 and its correlation to aggregation and anti-CD20 antibody activity in rituximab sensitive and resistant cell lines (RSCL and RRCL)

To investigate the mechanism underlying that low ICAM-1 expression group had better ORR, PFS and OS treated by R-CHOP compared to CHOP regimen, we used RSCL and RRCL generated in our lab as previous described.

We found the level of surface ICAM-1 protein decreased in RRCL by flow cytometry staining. However, ICAM-1 binding ligand CD11a did not have significant decrease level (Figure 3A). Western blot also validated the lower expression levels of ICAM-1 in RRCL than in RSCL (Figure 3B). Since ICAM-1 is the major adhesion molecule, further investigation revealed that RRCL did not aggregate as a cluster together as RSCL did (Figure 3C). We also used shRNA to knockdown ICAM-1 in RSCL and RRCL, and found that cell clustering is dependent on the ICAM-1 expression levels (Figure 3D). Anti-CD20 antibodies (rituximab, ofatumumab, TG20, R603 and GA101) dependent cellular toxicity (ADCC) and complement dependent cytotoxicity (CMC) were all decreased in rituximab resistant cells compared to its parental sensitive cell lines (Table 3). All these suggested decreased ICAM-1 expression correlated with loss of cellular aggregation and decreased CD20 antibody activity in RRCL.

Figure 3. Level of ICAM-1 and its correlation with aggregation in rituximab sensitive and resistant cell lines. (A) Levels of ICAM-1 and its binding ligand CD11a in rituximab sensitive and resistant cell lines by flow cytometry staining. (B) Western blot showed lower expression levels of ICAM-1 in rituximab resistant cell lines (RRCL) than in rituximab sensitive cell lines (RSCL). (C) Ability of aggregate as cluster in RRCL and RSCL at different time points. (D) Knockdown of ICAM-1 by shRNA and the ability of aggregation in RSCL and RRCL.

Rituximab enhances cellular aggregation in both ICAM-1 high expressed RSCL and ICAM-1 low expressed RRCL

The major cellular killing activity of rituximab is by CMC and ADCC, both of which depend on cellular interaction with complement and/or other cells. We observed that both RSCL and RRCL aggregated after rituximab treatment, regardless of ICAM-1 expression levels (Figure 4). Furthermore, We performed western blot to examine expression levels of other adhesion molecules (ICAM-3, VCAM-1 and E-cad) in RSCL, and found all of them showed low or no expression (data not shown).

Figure 4. Cellular aggregation in RSCL and RRCL treated with rituximab. Rituximab enhances cellular aggregation in both ICAM-1 high expressed RSCL and ICAM-1 low expressed RRCL.

Knockdown of ICAM-1 causes rituximab resistance

We neutralized ICAM-1 in Raji cell line, which had high expression of ICAM-1, and found there were no statistical differences of rituximab mediated CMC and ADCC between the control group and ICAM-1 neutralization group in vitro (P> 0.05, Figure 5A, 5B). However, MTT assay showed that downregulation of ICAM-1 could cause resistance of rituximab (P values are shown in Figure 5C). In vivo, rituximab combined with ICAM-1 neutralization did not affect rituximab killing activity (P> 0.05) (Figure 5D).

Figure 5. Impacts of neutralization of ICAM-1 in vitro and in vivo. (A, B) Neutralization of ICAM-1 did not affect rituximab and other anti-CD20 antibodies mediated CMC and ADCC in rituximab sensitive cell lines. (C) MTT assay showed that downregulation of ICAM-1 could cause resistance of rituximab. (D) Kaplan-Meier analyses showing the survival curves of neutralization of ICAM-1 alone and in combination with rituximab in vivo. (*P<0.05, **P <0.01, ***P <0.001,****P <0.0001).

Patients and samples

A total of 102 patients were included in this retrospective study that were newly diagnosed of DLBCL and treated in Fudan University Shanghai Cancer Center, Shanghai, China from March 2009 to December 2012. All patients were pathologically confirmed as DLBCL through surgical or biopsy samples, and all pathological results were reviewed by experienced pathologists in pathology department of Fudan University Shanghai Cancer Center. The demographic details, clinical and laboratory features of the patients are summarized in Table 1. This study was approved by the Institutional Review Board of the Fudan University Shanghai Cancer Center. Survival data were available with a median follow-up of 1,545 days (range 45~2,598 days).

Immunohistochemistry (IHC)

Immunohistochemical staining was carried out with the Dako Envision System (Dako, Glostrup, Denmark). In brief, formalin-fixed, paraffin-embedded tissue sections were incubated overnight at 4° C with mouse monoclonal antibody against ICAM-1 (Santa Cruz, CA, USA) at 1:100 dilutions. Slides with no primary antibodies added served as negative controls.

Immunohistochemical assessment

The expression level of ICAM-1 was assessed by percentage of positive cells. Tumors were considered positive when at least 75% of tumor cells expressed ICAM-1. Assessment was done by three pathologists without prior knowledge of the clinical features or follow-up data of the patients.

Cell lines

Human lymphoma cell line Raji and RL were purchased from American Type Culture Collection (ATCC, Manassas, VA). Rituximab resistant cell line Raji 4RH, RL 4RH were generated by exposure sensitive cells to rituximab at escalating dosages over times in the presence of human complement as previously described [32].

All the cell lines were maintained in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin and 100 g/ml streptomycin, maintained in 5% CO2 at 37° C.

Disruption of cell cluster aggregation for spontaneous re-aggregation study

Disruption of rituximab sensitive cell line (RSCL) clusters at the concentration of 1×10⁶ cells/ml were achieved by gentle pipetting before plating on a 6-well plate. Re-aggregation of cell cluster was then observed, and pictures were taken under microscope at different time intervals.

Effects of rituximab on spontaneous cell cluster aggregation

1×10⁶ cells were centrifuged and RPMI 1640-10%FBS media was removed, leaving 500ul in the tube with cells. Calculated volume of rituximab, or isotype controls was added to each tube of cells. Cells were resuspended briefly so that they mixed well with antibody, and then incubated for 1hr in 37° C incubator for binding of antibody to cell surface to occur. After 1hr incubation, 1ml of RPMI1640-10% FBS media was added back to each tube. Cells were again resuspended well by gentle pipetting before being transferred to a 6-well plate.

Flow cytometry for ICAM-1 detection

RSCL and rituximab resistant cell lines (RRCL) were treated with rituximab or Herceptin isotype control for 48h. Then the cells were stained with either FITC-conjugated mouse anti-human ICAM-1 (BD Invitrogen) or Isotype control (mouse IgG2α-FITC) (BD Invitrogen) for 30 minutes. Cells were then washed with PBS, fixed with 2% paraformaldehyde and analyzed by flow cytometry using FACSCCalibur (BD Biosciences). Cells were analyzed using the FCS express software (De Novo Software, Los Angeles, CA, USA).

Western blot

Cells were lysed in 200 μl lysis buffer (12.5mM Tris-HCL PH 6.8, 4% SDS, 20% glycerol, 0.004% bromophenol blue). The protein extracts (20 μg) were separated by 10% SDS–PAGE and transferred to polyvinylidene fluoride membranes (Millipore, Temecula, CA, USA). The membrane was blocked for 1 hour by 5% non-fat milk and then incubated with primary antibodies overnight at 4 ° C. After washed 3 times with Tris-buffered saline/0.1% Tween 20, the membrane was incubated with horseradish peroxidase-linked secondary antibodies for 1 hour. Proteins were visualized using Tanon full-automatic light detecting system with the BeyoECL Star (Ultra hypersensitive ECL chemiluminescence kit). All the data were confirmed by three individual experiments.

Plasmid, lentivirus production and infection

The forward sequences of shRNAs targeting ICAM1 were as follows: sh-1: CCGGCCGGTATGAGATTGTCATCATCTCGAGATGATGACAATCTCATACCGGTTTTTG;sh-2:CCGGGCCAACCAATGTGCTATTCAACTCGAGTTGAATAGCACATTGGTTGGCTTTTTG. The shRNA oligo sequences were cloned into a lentiviral shRNA expressing plasmid GV248 (GeneChem, Shanghai). A scramble sequence TTCTCCGAACGTGTCACGT constructed as stem-loop-stem structure were also cloned into the same plasmid and were used as control (SC). For the production of lentivirus, 293T cells were transfected with shRNA-expressing plasmids (SC, shICAM1-1, or shICAM1-2) and packaging plasmids (PSPAX2 and pMD2G) by Lipofectamine® 3000 (Invitrogen). Virus supernatants were harvested at 48h, 72h and 96h after transfection and were used to infect cultured cell lines. The infected cells were then isolated by puromycin selection.

Antibody-dependent cellular cytotoxicity (ADCC) and complement-mediated cytotoxicity (CMC)

The ADCC and CDC were accessed by ⁵¹Cr release assay. In brief, cells were labeled with ⁵¹Cr at 37° C, 5% CO₂ for 2 hours and then were plated at a cell concentration of 1x10⁵ cells/well (CMC assay) or 1x10⁴ cells/well (ADCC assay). Cells were then exposed to rituximab (10μg/ml), ofatumumab, TG20, R603 and GA101 or isotype (10μg/ml) and human serum (CMC, 1:4 dilution) or PBMCs (ADCC, 40:1 effector: target ratio) for six hours at 37° C and 5% CO₂. ⁵¹Cr release was measured and percentage of cell-lysis was calculated as previously described.

Cell viability assay (MTT assay)

RSCL was plated at a cell density of 1 × 10⁵ cells/ml in 96 well plates and was exposed to 10ug/ml of rituximab for different time intervals. At the indicated time points, 20 μl of 0.5 mg/ml MTT (Thiazolyl Blue Tetrazolium Bromide, Sigma) was added to each well and incubated at 37° C for 4 h. In the end, the culture was replaced with 150 μL DMSO (dimethyl sulfoxide, D8418, Sigma) to stop the reaction. The absorbance values (OD 570 nm) were measured using a spectrophotometer.

Analyses of the effect of ICAM-1 neutralization in RSCL on tumor growth in SCID mouse model

Following 1hr incubation with the indicated antibodies (10μg/ml rituximab, 0.25 μg/ml ICAM-1 neutralizing antibody (Calbiochem), or isotype controls), Raji cells were inoculated into SCID mice through intraperitoneally. The growth of tumor was measured twice per week and mice were sacrificed when one dimension of the tumor reached 2cm, or when necrotic/ulcerated tumor developed. The average cumulative days of survival were calculated.

Statistical analysis

The analyses were performed using SPSS 23.0 (SPSS Inc., Chicago) or GraphPad Prism 8 software (GraphPad Software, Inc., La Jolla, CA). Associations between protein expression and clinical/laboratory parameters or treatment response were assessed by χ² test. Differences of the cell viability between control and sh-ICAM-1 cell lines treated with rituximab were analyzed using students t-test. Kaplan-Meier survival curves were constructed for survival analyses, and differences were tested by the log-rank test. OS was defined as the time between the date of surgery or biopsy and the date of death or the date of last contact. PFS refers to the period from the beginning of treatment to the observed progression of the disease or the occurrence of death for any reason. The data of patients alive at the end of the study were censored. All P values were two-sided, and the results were considered significant if P < 0.05.