The Autocrine Secreted PDZD2 Functions As a Novel Tumor Suppressor in AML, Inducing Growth Inhibition and Cell Cycle Arrest

Epigenetic deregulation is a hallmark of acute myeloid leukemia (AML) and we have previously reported that recurring epigenetic abnormalities can be found across all AML subtypes, irrespective of cytogenetic or molecular drivers. Among the genes recurrently hypermethylated in AML we identified PDZD2 (PDZ domain-containing protein 2), which was silenced through DNA hypermethylation in over 90% of AML patients. PDZD2 processing generates two functional proteins: full-length PDZD2 and secreted PDZD2 (sPDZD2), a product of PDZD2 cleavage by caspase 3 at its C-terminal. While little is known about this protein's role in AML, work in solid tumors has shown that sPDZD2 can act as an autocrine tumor suppressor leading to senescence or quiescence of malignant cells. Given the role of PDZD2 in cancer and its almost universal loss in AML, we hypothesize that PDZD2 is a novel tumor suppressor in AML and that treatment with recombinant sPDZD2 represents a novel avenue for therapeutic targeting of AMLs.

To test this hypothesis, we first characterized PDZD2 expression pattern in a panel of 13 AML cell lines and normal hematopoietic CD34 ⁺ cells. Quantification of PDZD2 mRNA levels in AML cell lines confirmed its almost universal downregulation when compared to normal CD34 ⁺ cells. Subcellular localization analysis in normal hematopoietic stem and progenitor cells demonstrated the presence of full-length PDZD2 in the cytoplasmic and nuclear fractions. In addition, treatment of a subset of the cell lines with the hypomethylating agent Decitabine at 500 nM for 72 h resulted in upregulation of PDZD2 by a range of 2.5 to 100-fold change, validating the observation that PDZD2 gets silenced through DNA hypermethylation. At the protein level, while full-length PDZD2 was detectable in approximately half the AML cell lines tested, detailed analysis revealed that in these cell lines PDZD2 was misprocessed, failing to get cleaved, and resulting in accumulation of the full-length version of the protein in the membranous (ER) fraction of the cell, while this is not observed in normal CD34 ⁺ cells.

We next tested the effect of re-exposing AML cell lines to sPDZD2 by treating a panel of 10 AML cell lines with increasing doses of recombinant sPDZD2. sPDZD2 was added to the culture media every 12 h and cell growth was evaluated over a period of 8 days. We analyzed cell proliferation and viability every 48 h as well as cell cycle and expression of differentiation markers (CD11b and CD15) at the end of the treatment period. Even at relatively low doses (10-100nM), we observed variable degrees of dose-dependent growth inhibition in 8 out of 10 AML cell lines tested, with some cell lines showing greater and/or faster sensitivity than others. Cell cycle analysis showed that after 8 days of treatment, 5/8 sensitive cell lines (Kasumi1, KG1, SKNO1, MV4-11, HL60, K562) displayed G ₀/G ₁ arrest compared to vehicle control. Notably, 7/8 sensitive cell lines showed signs of a dose-dependent increase in the expression of either CD11b or CD15 differentiation markers, while the resistant cell line MOLM13 did not increase expression of either of these markers. In addition, patient-derived xenograft lines corresponding to AMLs carrying MLL-AF9 and 5 (out of 6 tested) primary human AML specimens cultured in vitro showed clear sensitivity to treatment with sPDZD2 compared to vehicle control. In contrast, normal CD34 ⁺ cells isolated from healthy donors and treated with sPDZD2 at the same doses showed no signs of toxicity, confirming sPDZD2's selectivity against AML cells. Finally, we sought to determine whether endogenous re-expression of full-length PDZD2 would likewise result in cell growth inhibition, indicating intact function of the cleavage mechanism. Transient activation of PDZD2 using a dCas9-VP64 fusion system resulted in a block in cell proliferation and G ₀/G ₁ cell cycle arrest in K562 cells, indicating that re-expression of PDZD2 phenocopies the results seen with sPDZD2 treatment.

In summary, our findings are the first to demonstrate that sPDZD2 functions as a tumor suppressor in AML, inhibiting cell cycle and inducing differentiation, and that treatment with recombinant sPDZD2 can induce growth inhibition in a wide variety of AML cell lines and primary specimens, irrespective of their molecular and cytogenetic subtypes.