Fig 2: HA-CDKL3 co-immunoprecipitates endogenous Akt.Source data are available for this figure.

Fig 3: Cyclin-dependent kinase–like 3 (CDKL3) defines poor prognosis and correlates with Akt phosphorylation in clinic.(A, B) Kaplan–Meier plots of overall survival (A) and metastasis-free survival (B) of 152 osteosarcoma (OS) patients, stratified by CDKL3 levels (− represents negative staining; +, ++, and +++ represent weak, intermediate, and strong staining, respectively). (C) Representative immunohistochemistry (IHC) images of OS biopsies with different levels of CDKL3 expression on an OS microarray containing 152 primary OS tissues samples. Scale bar = 500 μm. (D) Representative HE and IHC images of OS biopsies stained by CDKL3 and Ki67. Scale bar = 100 μm. (E) Quantitative analysis of Ki67 expression in OS patients with different levels of CDKL3 expression. (F) HE and IHC staining images of CDKL3, AKT, pAKT (p308), pAKT (ser473), and LC3 in representative CDKL3-positive and CDKL3-negative patients. Scale bar = 100 μm. (G) Western blot detection of CDKL3 and Akt phosphorylation in OS tissues and adjacent non-tumor tissues from three different patients. (H) Putative underlying mechanism that CDKL3 promotes OS progression. Overexpression of CDKL3 leads to increased phosphorylation of Akt, followed by governing mTORC1 and FoxO activities, likely independent of functions of PDK1, growth factors, or relevant receptors; this may inhibit autophagy and eventually promote OS development. *P < 0.05, **P < 0.01, ***P < 0.001, two-tailed t test.Source data are available for this figure.

Fig 4: Cyclin-dependent kinase–like 3 (CDKL3) critically regulates Akt activation in osteosarcoma (OS).(A) Akt and mTOR activation under regular or starvation conditions in parental and CDKL3-KO U2OS cells. Upon CDKL3 KO, Akt and mTORC1 activation were significantly alleviated. The numbers below each blotting strip is intensity quantification by ImageJ. (B) The expression patterns of mTORC1 and FoxO downstream target genes confirm that CDKL3 regulates both pathways at the transcription level. Cells were cultured under the regular growth condition. (C) Overexpression of CDKL3 in U2OS cells causes hyper-activation of Akt in the presence and absence of FBS. (D) General working mechanism of rapamycin, AZD5363, and MK-2206. (E) U2OS cell growth under different conditions (n = 3). (F) Co-immunoprecipitation (Co-IP) reveals the physical interaction between HA-CDKL3 and Myc-Akt1. (G) Endogenous CDKL3 coexists with endogenous Akt shown by co-IP. (H) Schematic diagrams of Akt1 constructs for mapping. (I) Co-IP of HA-CDKL3 with different Akt1 constructs shows that the Akt kinase domain is indispensable for CDKL3 interaction. (I, J) Reverse IP of CDKL3 and Akt1 confirms the findings in (I). Error bars indicate SD (n = 3). *P < 0.05 parental DMSO versus CDKL3−/− DMSO, or parental rapamycin, or parental AZD5363, or parental MK-2206; **P < 0.01 CDKL3−/− DMSO versus CDKL3−/− rapamycin, or CDKL3−/− AZD5363, or CDKL3−/− MK-2206; ##P < 0.01 parental MK-2206 versus CDKL3−/− MK-2206. Statistical significance of the differences was estimated by unpaired two-tailed t test.Source data are available for this figure.

Fig 5: Protein mapping on cyclin-dependent kinase–like 3 (CDKL3).(A) Schematic drawing for CDKL FL (full-length), K (Kinase domain), ΔK (Truncation of kinase domain) constructs. (B) Co-immunoprecipitation of Myc-Akt and HA-CDKL3 revealed that the kinase domain of CDKL3 is responsible for Akt interaction.Source data are available for this figure.