Haifeng Xu

• Yin, Yanping*; Xu, Haifeng*; He, Liye; Brown, Jennifer R.; Mato, Anthony R.; Aittokallio, Tero Antero & Skånland, Sigrid S (2025) Protein Profiles Predict Treatment Responses to the PI3K Inhibitor Umbralisib in Patients with Chronic Lymphocytic Leukemia. Clinical Cancer Research. ISSN 1078-0432. 31(10), p. 1943–1955. doi: 10.1158/1078-0432.CCR-24-2911. (* co-first authors)
• Xu, Haifeng; Lillebergen, Tonje Lien; Bergholtz, Helga; Fleischer, Thomas; Djerroudi, Lounes; Vincent-Salomon, Anne; Sørlie, Therese & Aittokallio, Tero Antero (2021). Multi-omics marker analysis enables early prediction of breast tumor progression. Frontiers in Genetics. ISSN 1664-8021. 12, p. 1–14. doi: 10.3389/fgene.2021.670749.

• Xu, Haifeng; Venäläinen, Mikko S.; Berge, Tone & Elo, Laura L. (2026). A candidate CD4⁺ T-cell gene expression panel for multiple sclerosis highlights MgADP-associated genes. Multiple Sclerosis and Related Disorders. ISSN 2211-0348. p. 107944–107944. doi: 10.1016/j.msard.2026.107944. Full text in Research Archive
• Yanping, Yin; Xu, Haifeng; He, Liye; Brown, Jennifer R.; Mato, Anthony R. & Aittokallio, Tero Antero [Show all 7 contributors for this article] (2025). Protein Profiles Predict Treatment Responses to the PI3K Inhibitor Umbralisib in Patients with Chronic Lymphocytic Leukemia. Clinical Cancer Research. ISSN 1078-0432. 31(10), p. 1943–1955. doi: 10.1158/1078-0432.CCR-24-2911. Full text in Research Archive Show summary

  Purpose: The management of chronic lymphocytic leukemia (CLL) has significantly improved with targeted therapies. However, many patients experience a suboptimal response. To optimally select the best therapy, predictive biomarkers are necessary. In this study, we used the phosphoinositide 3-kinase (PI3K) inhibitor umbralisib as a model to (i) understand the impact of targeted treatment on cell signaling and immunophenotypes in responders and nonresponders, (ii) identify molecular features that predict individual treatment responses, and (iii) suggest alternative treatment options for the nonresponders. Experimental Design: We performed functional phenotyping of CLL cells from patients enrolled in two clinical trials with umbralisib, administered either as a monotherapy (NCT02742090, n = 55) or in combination with the Bruton tyrosine kinase (BTK) inhibitor acalabrutinib (NCT04624633, n = 12). Results: We found that umbralisib monotherapy led to significant changes in (phospho)protein levels, including AKT (pS473), in responders but not in nonresponders. Furthermore, the proportion of cytotoxic natural killer (NK) cells increased at the end of the study but only in responders, suggesting a role in the antitumor response. To identify molecular predictors of response, we used the baseline levels of 30 (phospho)proteins in the monotherapy cohort as input features for a machine learning model, which achieved significant prediction accuracy in cross-validation and maintained its predictive power in the combination cohort. Drug sensitivity profiling of the CLL cells at baseline suggested that PI3K + Bcl-2 inhibitors are effective in umbralisib nonresponders. Conclusions: Functional phenotyping reveals differential cellular responses to umbralisib treatment in responders and nonresponders; predicts treatment response of individual patients with CLL; and suggests alternative treatment options for the nonresponders.

• Yin, Yanping; Athanasiadis, Paschalis; Karlsen, Linda; Urban, Aleksandra; Xu, Haifeng & Murali, Ishwarya [Show all 17 contributors for this article] (2022). Functional testing to characterize and stratify PI3K inhibitor responses in chronic lymphocytic leukemia. Clinical Cancer Research. ISSN 1078-0432. 28(20), p. 4444–4455. doi: 10.1158/1078-0432.CCR-22-1221. Full text in Research Archive Show summary

  Purpose: PI3K inhibitors (PI3Ki) are approved for relapsed chronic lymphocytic leukemia (CLL). Although patients may show an initial response to these therapies, development of treatment intolerance or resistance remain clinical challenges. To overcome these, prediction of individual treatment responses based on actionable biomarkers is needed. Here, we characterized the activity and cellular effects of 10 PI3Ki and investigated whether functional analyses can identify treatment vulnerabilities in PI3Ki-refractory/intolerant CLL and stratify responders to PI3Ki. Experimental Design: Peripheral blood mononuclear cell samples (n = 51 in total) from treatment-naïve and PI3Ki-treated patients with CLL were studied. Cells were profiled against 10 PI3Ki and the Bcl-2 antagonist venetoclax. Cell signaling and immune phenotypes were analyzed by flow cytometry. Cell viability was monitored by detection of cleaved caspase-3 and the CellTiter-Glo assay. Results: pan-PI3Kis were most effective at inhibiting PI3K signaling and cell viability, and showed activity in CLL cells from both treatment-naïve and idelalisib-refractory/intolerant patients. CLL cells from idelalisib-refractory/intolerant patients showed overall reduced protein phosphorylation levels. The pan-PI3Ki copanlisib, but not the p110δ inhibitor idelalisib, inhibited PI3K signaling in CD4+ and CD8+ T cells in addition to CD19+ B cells, but did not significantly affect T-cell numbers. Combination treatment with a PI3Ki and venetoclax resulted in synergistic induction of apoptosis. Analysis of drug sensitivities to 73 drug combinations and profiling of 31 proteins stratified responders to idelalisib and umbralisib, respectively. Conclusions: Our findings suggest novel treatment vulnerabilities in idelalisib-refractory/intolerant CLL, and indicate that ex vivo functional profiling may stratify PI3Ki responders.

• Xu, Haifeng; Lillebergen, Tonje Lien; Bergholtz, Helga; Fleischer, Thomas; Djerroudi, Lounes & Vincent-Salomon, Anne [Show all 8 contributors for this article] (2021). Multi-omics marker analysis enables early prediction of breast tumor progression. Frontiers in Genetics. 12, p. 1–14. doi: 10.3389/fgene.2021.670749. Full text in Research Archive Show summary

  Ductal carcinoma in situ (DCIS) is a preinvasive form of breast cancer with a highly variable potential of becoming invasive and affecting mortality of the patients. Due to the lack of accurate markers of disease progression, many women with detected DCIS are currently overtreated. To distinguish those DCIS cases who are likely to require therapy from those who should be left untreated, there is a need for robust and predictive biomarkers extracted from molecular or genetic profiles. We developed a supervised machine learning approach that implements multi-omics feature selection and model regularization for the identification of biomarker combinations that could be used to distinguish low-risk DCIS lesions from those with a higher likelihood of progression. To investigate the genetic heterogeneity of disease progression, we applied this approach to 40 pure DCIS and 259 invasive breast cancer (IBC) samples profiled with genome-wide transcriptomics, DNA methylation, and DNA copy number variation. Feature selection using the multi-omics Lasso-regularized algorithm identified both known genes involved in breast cancer development, as well as novel markers for early detection. Even though the gene expression-based model features led to the highest classification accuracy alone, methylation data provided a complementary source of features and improved especially the sensitivity of correctly classifying DCIS cases. We also identified a number of repeatedly misclassified DCIS cases when using either the expression or methylation markers. A small panel of 10 gene markers was able to distinguish DCIS and IBC cases with high accuracy in nested cross-validation (AU-ROC = 0.99). The marker panel was not specific to any of the established breast cancer subtypes, suggesting that the 10-gene signature may provide a subtype-agnostic and cost-effective approach for breast cancer detection and patient stratification. We further confirmed high accuracy of the 10-gene signature in an external validation cohort (AU-ROC = 0.95), profiled using distinct transcriptomic assay, hence demonstrating robustness of the risk signature.

View all works in NVA