Abstract P042

Proteomic profiling identifies classic Hodgkin lymphoma patients at risk of bleomycin pulmonary toxicity

Background: Advances in both chemo- and radiotherapy have notably improved cure rates in classic Hodgkin lymphoma (cHL), resulting in overall survival rates surpassing 80%. Consequently, an increasing number of long-term survivors are emerging, raising concerns about the possibility of long-term complications, notably the risk of cardiac and pulmonary toxicity. Bleomycin poses a significant risk of bleomycin-induced pulmonary toxicity (BPT), with an incidence around 10%, and a mortality ranging between 10-20%. We performed proteomics as a tool for conducting a large-scale hypothesis-generating study to identify differentially expressed proteins in diagnostic cHL lymph node tumor samples from patients with and without subsequent BPT (Figure 1).

Methods: The study included patients diagnosed with cHL at Aarhus University Hospital, Denmark, during the period 2000–2018, treated with ABVD-based therapy regimens. Protein expression patterns in diagnostic lymphoma samples from patients who either developed BPT (n=23; T-cHL) or did not (n=44; nT-cHL), were analyzed by label-free quantification nano liquid chromatography-tandem mass spectrometry (LFQ nLC-MS/MS). Differential expressions of janus kinase 3 (JAK3), BH3 integrating domain death agonist (BID), matrix metallopeptidase 9 (MMP9), tumor protein D52 (TPD52), and phosphoinositide 3-kinase regulatory subunit 4 (PIK3R4) were further evaluated by immunohistochemistry (n=290).

Results: At diagnosis, lymph node samples from T-cHL patients had significantly lower expression of TPD52 (p<0.001), and PIK3R4 (p=0.006), whereas JAK3 (p=0.003), BID (p=0.003), and MMP9 (p=0.006) showed a significantly higher expression compared with samples from nT-cHL. Dividing the biomarkers into risk scores of 0 or 1, with 1 being high risk of BPT according to the individual markers, i.e. low levels of TPD52 and PIK3R4 and high levels of JAK3, BID, and MMP9, and subsequently combining the risk scores, was significantly predictive of BPT. A risk score of ≥ 4 markers predicted BPT with a sensitivity of 0.600 and specificity of 0.939 (p<0.001).

Conclusion: Upon lymphoma diagnosis, we identified differences in protein expression in pre-treatment lymph node biopsies that could identify patients at high risk of developing BPT. Although individual protein markers offer limited predictive value for BPT development, utilizing a combination of markers can improve prediction accuracy and assist in making informed treatment decisions.