TRAF1 S146 is constitutively phosphorylated in OP9-cultured primary chronic lymphocytic leukemia cells by PKN1/2

Abstract

TNF receptor–associated factor 1 (TRAF1) is a prosurvival signaling adaptor that contributes to NF-κB activation downstream of a subset of TNF receptor superfamily members. TRAF1 is overexpressed in many cancers of mature B cells, including chronic lymphocytic leukemia (CLL). Previous studies have established that TRAF1 S146 is a target of phosphorylation by the kinase PKN1 and that PKN1 is required to prevent cellular inhibitor of apoptosis protein (cIAP)–dependent degradation of TRAF1 in the CD40 signaling complex. The kinase inhibitor OTSSP167 inhibits PKN1 in the nanomolar range and its addition to primary CLL cells was previously shown to induce dose-dependent loss of TRAF1 and concomitant increases in activated caspase 3 and cell death. These studies identified PKN1 as a target for therapy of CLL. To identify more potent and specific PKN1 inhibitors for therapy of B-cell cancers, it is important to measure a direct target of PKN1, such as phospho-TRAF1. To this end, here we use overexpression of an S146A mutant of human TRAF1 in HEK293 cells to validate a recently generated TRAF1 phospho (p)S146-specific antibody and to confirm that this phosphorylation is lost upon treatment with OTSSP167. Using CRISPR/Cas9 knockout in RAJI cells, we also show that both PKN1 and the closely related family member PKN2 can phosphorylate TRAF1 S146. We further show that TRAF1 S146 is constitutively phosphorylated in OP9 cultured primary human CLL cells, including those with p53 mutations, and that this phosphorylation is sensitive to inhibition with OTSSP167. These findings provide support the development of more potent PKN1/2 inhibitors for CLL.

Introduction

Chronic lymphocytic leukemia (CLL) is the most prevalent adult leukemia. In recent years, first-line CLL therapy has dramatically shifted toward targeted therapies including drugs such as ibrutinib to inhibit B-cell receptor (BCR) signaling or venetoclax to target cellular survival through antagonizing Bcl-2 (for review see¹). Despite the promise of these new therapies, resistance still occurs and there is a need for additional approaches to target venetoclax-resistant CLL survival signaling.²

TNF receptor–associated factor 1 (TRAF1) is an NF-κB inducible signaling adaptor that contributes to feedback enhancement of NF-κB signaling downstream of several TNF receptor (TNFR) superfamily members.³ In addition to NF-κB, TRAF1 has been implicated in the regulation of JNK and MAPK signaling pathways, although its precise role appears to be context-dependent.⁴ TRAF1 is normally undetectable in resting normal immune cells but overexpressed in cancers of mature B-cell origin, including CLL.⁵ TRAF1 has been directly implicated in B lymphomagenesis using mice expressing constitutively active NF-κB2, where deletion of TRAF1 restored normal B-cell homeostasis.⁶ Mechanistically, TRAF1 forms a complex with TRAF2 and the cellular inhibitors of apoptosis proteins (cIAPs).⁷ TRAF2, but not TRAF1, possesses E3 ubiquitin ligase activity within this complex. However, structural analysis has shown that a (TRAF2)₂ TRAF1 heterotrimer more efficiently recruits the cIAPs than TRAF2 alone.⁷ TRAF1 also stabilizes TRAF2 against degradation.⁸ The TRAF1/2/cIAP complex functions to increase NF-κB activation downstream of TNFRs such as TNFR2, CD40, CD30, and 4-1BB.^9–12 This in turn leads to induction of Bcl-2 family members such as Mcl-1 that can contribute to venetoclax resistance (Fig. 1).^2,3,13

Figure 1.

Schematic showing the PKN1/2–TRAF1 axis in TNFR signaling pathways in CLL. CLL cells exhibit constitutive signaling through the BCR, leading to NF-κB signaling and induction of prosurvival Bcl-2 family members. CD40 and related TNFR superfamily members demonstrate autocrine signaling in CLL cells leading to NF-κB and MAPK activation, providing another source of survival signals. Venetoclax antagonizes the effects of Bcl-2. TRAF1 is constitutively phosphorylated by PKN1 in lymphoma and CLL cells. A PKN1 inhibitor can induce TRAF1 degradation and loss of TRAF1-dependent signals in B-cell cancers.

TRAF1 serine 146 (S146) is a target of phosphorylation by the protein kinase C–related kinase PKN1.¹⁴ We previously showed that in the absence of PKN1, TRAF1 is targeted for degradation by cIAPs during CD40 signaling.¹³ Conversely, phosphorylation by PKN1 allows TRAF1 protein levels to be maintained during TNFR signaling, resulting in increased NF-κB signaling and induction of prosurvival Bcl-2 family members.¹³ Thus, PKN1-mediated phosphorylation of TRAF1 S146 is critical for maintaining TRAF1 protein levels and the resulting prosurvival signaling in B cells. Previous work identified the kinase inhibitor OTSSP167 as inhibiting PKN1 with a half-maximal inhibitory concentration (IC₅₀) of 18 nM.¹³ Consistently, when added to primary CLL cells, OTSSP167 led to a dose-dependent loss of TRAF1 protein, with concomitant loss of Bcl-2 and Mcl-1, and increased levels of activated caspase 3, leading to cell death.¹³ Moreover, OTSSP167 synergized with venetoclax in inducing CLL cell death.¹³ These findings identified PKN1 as a potential target for inhibition for CLL. However, OTSSP167 also inhibits other kinases, most notably maternal embryonic leucine zipper kinase (MELK)¹⁵ with an IC₅₀ in the picomolar range. Thus, it is of interest to find more selective PKN1 inhibitors. TRAF1 is regulated both transcriptionally through NF-κB¹⁶ and at the level of protein stability through PKN1.¹³ Therefore to identify more potent and specific inhibitors of PKN1, it would be important to measure a direct target of PKN1, namely TRAF1 phospho-S146 (pS146). To this end, in this study, we used human embryonic kidney 293 (HEK293) cells overexpressing wild-type (WT) or S146A TRAF1 to validate a recently generated antibody specific for the pS146 form of TRAF1.

PKN1 is part of a family that includes 3 related kinases (PKN1, 2, and 3), with PKN1 and PKN2 reported to be ubiquitously expressed and 50% identical at the amino acid level.¹⁷ Here we show that either PKN1 or PKN2 can phosphorylate TRAF1 on S146. We also show that TRAF1 S146 is constitutively phosphorylated in primary patient CLL cells after overnight culture on OP9 stromal cells and that this phosphorylation is inhibited by OTSSP167, previously identified as inhibiting PKN1.¹³ Together our studies suggest constitutive PKN1/2 activity in CLL cells and provide a means of monitoring this activity through a TRAF1 pS146–specific antibody.

Materials and methods

Human subjects

Peripheral blood mononuclear cells (PBMCs) from patients with a diagnosis of CLL were obtained from the Leukemia Tissue Bank at Princess Margaret Cancer Centre/University Health Network (UHN), Toronto, Ontario. Informed consent for tissue bank donation was in compliance with the Declaration of Helsinki and in agreement with the UHN Research Ethics Review Board (protocol #01-0573). Research with these samples was conducted at the University of Toronto, with approval of the University of Toronto ethics board (protocol #60791). The clinical and cytogenetic characteristics (where available) of the CLL patient samples used in this study are summarized in Table 1.

Table 1.

Patient characteristics.

ID	Exp #	Cytogenetics	Sex	Vital status	Date Dx	DOS	WBC	Lymph	Neut	PLT	Figure
794	CLL-6	17p del-p53 mut	F	dec	2-2007	1-2008	172	166	4.8	116	Fig. 3C
80049	CLL-3	13q del	M	alive	6-1998	4-2008	105	99.9	3.6	155	Fig. 3C
140984	CLL-4	ND	F	dec	06-1990	10-2014	202	191	2.5	61	Fig. 3C
967557	CLL-15	ND	M	alive	6-2022	6-2022	11.3	6.9	2.5	122	Fig. 3A, lane 2
977098	CLL-16	13q del	M	alive	9-2019	8-2022	45	40.8	2.6	99	Fig. 3A, lane 4
981116	CLL-18	TP53	M	alive	8-2022	9-2022	63.2	56.5	5.1	259	Fig. 3A, lane 6
982520	CLL-17	ND	M	alive	2018	9-2022	24.5	21.5	2.5	115	Fig. 3A, lane 3
991996	CLL-14	ND	F	alive	2-2019	12-2022	77.9	65.2	5.7	186	Fig. 3A, lane 1 Fig. 3B, lanes 1, 2, 3
992061	CLL-20	TP53	M	alive	12-2022	12-2022	91.2	78.3	3.5	20	Fig. 3A, lane 7 Fig. 3B, lanes 4, 5, 6
1003519	CLL-19	13q del	F	alive	9-2021	3-2023	68.4	63.6	2.4	151	Fig. 3A, lane 5
1016318	CLL-21	TP53	F	dec	2019	7-2023	70.5	64.3	3.1	93	Fig. 3A, lane 8 Fig. 3B, lanes 7, 8, 9

Abbreviations: CLL, chronic lymphocytic leukemia; Date Dx, date of diagnosis (month-year); dec, deceased; DOS, date of sample (month-year); Exp, Experiment; F, female; Lymph, lymphocyte count (10⁹/L); M, male; ND, Not Determined; Neut, neutrophil count (10⁹/L); PLT, platelet count (10⁹/L); WBC, white blood cell count (10⁹/L).

Cell lines

HEK293 cells (ATCC) were cultured in Eagle’s minimum essential medium (MEM) with Earle’s salts (purchased from Wisent Inc.) supplemented with 10% fetal calf serum (FCS, Wisent Inc.) and 1% of 100× glutamine-penicillin-streptomycin (Sigma-Aldrich, Oakville, Canada). RAJI B cells (ATCC) were cultured in RPMI with 25 mM HEPES (purchased from Wisent Inc.) supplemented with 10% FCS from Wisent and 1% of 100× glutamine-pyruvate-penicillin-streptomycin (Sigma-Aldrich). OP9 stromal cells, a mouse bone marrow–derived stromal cell line used to support the survival of primary CLL cells ex vivo, were kindly provided by J.C. Zuniga-Pflücker (Sunnybrook Research Institute, Toronto, Canada). We previously found that these cell lines supported CLL primary culture for a least 1 week.¹³ All of the cell lines tested negative for mycoplasma (Mycoplasma detection kit, Millipore-Sigma, Oakville, Canada). TRAF1-overexpressing cells were generated as described previously.¹³ In brief, HEK293 cells were transiently transfected with pcDNA3 expression constructs encoding WT human TRAF1 or the S146A mutant of TRAF1 using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. HEK293 cells do not express endogenous TRAF1, allowing unambiguous assessment of WT versus S146A TRAF1 phosphorylation. RAJI cells, a Burkitt lymphoma cell line, were used for CRISPR/Cas9 knockout studies because they endogenously express both TRAF1 and PKN1 and PKN2.

Antibodies

Anti-TRAF1 pS146 (EPR25987-11), BSA and azide free, was kindly provided by Abcam (Cambridge, UK). Anti-TRAF1 (Cell Signaling Technology [CST] clone 45D3, catalog [cat] #4715), anti-PKN2 (CST, cat #2612), and anti-GAPDH (CST, cat #2118) were purchased from New England Biolabs (Whitby, Canada). Anti-PKN1 (cat #610687) was purchased from BD Biosciences (Canada) until it was discontinued, after which we used anti-PKN1 from Thermo Fisher Scientific (cat #PA587454) purchased from Life Technologies (Burlington, Canada). Goat anti-rabbit-HRP and goat anti-mouse-HRP from Jackson Immunoresearch were purchased from Cedarlane. Goat anti-actin-HRP (A3854) was purchased from Millipore-Sigma Canada.

Knockout of PKN1 and PKN2

Cell lines with gene disruptions in PKN1 and/or PKN2 were generated at the Genomic Engineering and Molecular Biology core facility in the Faculty of Medicine at the University of Ottawa (RRID: SCR_022954). In brief, guide sequences targeting gene ID 5585 (PKN1) or 5586 (PKN2) were cloned into pLentiCRISPRv2 (puromycin resistant) and pLentiCRISPRv2BLAST (blasticidin resistant), respectively. A nonmammalian targeting guide sequence targeting Renilla luciferase was cloned into both these vectors as a control. Guide sequences were chosen to minimize the potential of off-target cutting¹⁸ and cloned as previously described.¹⁹ In brief, complementary oligonucleotides encoding each single guide RNA (sgRNA) sequence were phosphorylated using T4 polynucleotide kinase, annealed by heating to 95 °C and slowly cooling to room temperature, and then ligated into Bacillus stearothermophilus BBI (BsmBI)-digested pLentiCRISPRv2 (puromycin resistant) or pLentiCRISPRv2-Blast (blasticidin resistant) backbone vectors using T4 DNA ligase. Guide sequences are listed in Table 2. Lentiviral particles were generated as previously described¹⁹ and were used to transduce RAJI cells, followed by selection with puromycin (1 µg/mL) or blasticidin (5 µg/mL) as appropriate. In brief, sgRNA oligonucleotides were annealed, phosphorylated, and ligated into the BsmBI-digested lentiCRISPRv2 or lentiCRISPRv2-Blast backbone. Lentiviral particles were produced by co-transfection of HEK293 cells with the lentiCRISPR construct and packaging plasmids psPAX2 and pVSVg, and viral supernatants were collected 48 to 72 hours posttransfection. Editing efficiencies for guides targeting gene ID 5585 or 5586 were measured by generating PCR products flanking the edit sites using primers 5′-CTTAATGTGGGGGACGCTGT and 5′- TCTTGAGTGTTACCGGGTGC or 5′-AACTGCTGATCCAGACGTGTT and 5′-AATCGCCTTAGCTGTTAGCG, respectively, and sequencing these primers using Oxford Nanopore technology (sequencing performed by Plasmidsaurus). Subsequently, editing efficiencies were measured using the CRISPRESSO2 software suite,²⁰ subtracting sequencing noise measured in the control population. Independent pairs of guides were then used to generate lines in which both genes were disrupted (5585 g2 + 5586 g1 and 5585 g3 + 5586 g3). Editing rate represents a snapshot of the polyclonal population at one point in time and likely increased over passaging from this initial measured value as the lines stably express the CRISPR/Cas9 system.

Table 2.

Guide sequences and Nanopore-measured editing rates for PKN1 and PKN2.

Target gene	Guide	Sequence 5′-3′	Editing rate
5585 (human PKN1)	g1	GCTGCGGCGGGAAATCCGCA	31%
	g2	GGTCAGTGGTGGCCCGCCGC	45%
	g3	CGTGGTGCTTCCCGACCCGG	72%
5586 (human PKN2)	g1	AAGCTCGATAATACTGTGGT	87% (89% in double)
	g2	CTGTGGTTGGCCAAACTAGC	24%
	g3	AGTTTACACTGGAACTGGAC	67% (74% in double)
Renilla (nontargeting control)	gREN	GGTATAATACACCGCGCTAC	Not applicable

Western blot

Media was removed and cells were washed once with cold PBS (Wisent Inc), then lysed for protein extraction. Cells were lysed in RIPA lysis buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl 1.0% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS [Millipore-Sigma]) with phosphatase and protease inhibitor mix (Roche, Basel, Switzerland). Total protein concentration was quantified by a colorimetric assay (Bio-Rad, Berkeley, CA, USA), then subjected to SDS-PAGE (4% to 20% gradient gel) and transferred to polyvinylidene difluoride membranes (wet transfer; transfer with 10% methanol, Bio-Rad). After blocking with 5% non-fat dry milk solution in TBST (TBS with 0.1% Tween), membranes were probed with antibodies specific for pS146 of TRAF1, TRAF1 for 1 hour at room temperature, PKN1 overnight at 4 °C, and GAPDH for 30 minutes at room temperature, followed by HRP-conjugated anti-rabbit or anti-mouse antibody, and signals were detected with a chemiluminescence substrate (Clarity Western ECL substrate from Bio-Rad Canada and Immobilon Forte Western HRP substrate from Millipore-Sigma Canada).

Primary CLL cell culture and treatment with inhibitors

OP9 cells were resuspended at 10⁵ cells/mL in OP9 media (α-MEM from Thermo Fisher Scientific with added 20% FCS, glutamine, penicillin, and streptomycin) and seeded at 5 × 10⁴ cells per well into a 24-well plate. When 80% to 90% confluent, CLL patient samples (PMBCs) were thawed and resuspended at 8 × 10⁶ cells/mL in high-glucose complete media (RPMI 1640 supplemented with an additional 2.5 g/L total glucose [total 4.5 g/L final] and other additions as indicated for RAJI cells above). The OP9 media was removed and CLL cells were plated at 2 × 10⁶ cells per well on a 24-well plate containing confluent OP9 and rested overnight. Samples were treated with OTSSP167 (dose indicated in the figures) or 0.1% DMSO media control for 24 hours, then samples were gently removed from the well and prepared for western blot as described above.

Results

Validation of TRAF1 pS146–specific antibody

Anti-TRAF1 pS146 (clone EPR25987-11 Abcam) is reported to bind pTRAF1 containing peptide and detects a protein the size of TRAF1 in RAJI cells. To validate that this antibody is specific for the pS146 form of TRAF1, we used parental HEK293 cells that do not express TRAF1, as well as HEK293 cells transfected with WT or a mutant form of human TRAF1 in which serine 146 is replaced with alanine (S146A TRAF1) (Fig. 2A). The results show that anti-TRAF1 pS146 antibody detects WT but not S146A TRAF1. We next used OTSSP167, a kinase inhibitor previously shown to inhibit PKN1,¹³ to determine if total TRAF1 as well as TRAF1 pS146 was sensitive to this inhibitor (Fig. 2B–D). We found that administration of OTSSP167 for 24 hours resulted in loss of TRAF1 pS146 with an IC₅₀ of 29.4 nM (Fig. 2B, D, E).

Figure 2.

OTSSP167 inhibits TRAF1 phosphorylation in a dose-dependent manner. (A) HEK293 cells were either nontransfected or transfected with wild-type TRAF1 (TRAF1 WT) or S146A mutant TRAF1 (TRAF1 S146A). Whole cell lysates were then immunoblotted for pS146-TRAF1 (p-TRAF1), total TRAF1 (TRAF1), and β-actin (data are representative of at least 3 independent experiments). (B) HEK293 cells overexpressing WT TRAF1 were treated with 10 to 100 nM OTSSP167 for 24 hours. Whole cell lysates were then immunoblotted for pS146-TRAF1 (p-TRAF1), total TRAF1 (TRAF1), PKN1, and β-actin. Lysates from TRAF1 WT and S146A-overexpressing HEK293 cells were also included on the blots as controls for the p-TRAF1 assay. (C, D) Densitometry of TRAF1 and p-TRAF1 band intensity from 3 independent experiments, as in (B), were quantified by ImageLab software and (E) used to calculate the IC₅₀ for PKN1 based on reduction of expression of TRAF1 pS146. *P < 0.05, **P < 0.01, One-Way ANOVA with Dunnett's multiple comparisons test.

PKN1 and PKN2 both contribute to TRAF1 phosphorylation at S146

PKN1 and PKN2 are highly similar, with both being expressed by RAJI lymphoma cells (Fig. 3). Therefore, to determine whether PKN2 can also contribute to phosphorylation of TRAF1, we generated mutant versions of RAJI in which PKN1 or PKN2 or both were knocked out by CRISPR/Cas9 technology. Knockout of either PKN1 or PKN2 each partially reduced the level of TRAF1 pS146, with additive effects of knockout of both kinases (Fig. 3). These data suggest that the highly similar PKN1 and PKN2 can each contribute to TRAF1 S146 phosphorylation.

Figure 3.

PKN1 and PKN2 are required for TRAF1 phosphorylation. Five or 10 µg of whole cell lysates from WT, bulk PKN1 knockout (PKN1 KO), bulk PKN2 knockout (PKN2 KO), or bulk PKN1 and PKN2 double knockout (PKN1/2 KO) RAJI cells were immunoblotted for TRAF1 pS146 (p-TRAF1), total TRAF1 (TRAF1), PKN1, PKN2 and GAPDH. Lysates from TRAF1 WT and S146A-overexpressing HEK293 cells were loaded separately as controls for TRAF1 pS146 specificity. These data are representative of 3 similar experiments, in which 2 of 3 experiments showed a similar role for PKN1 and PKN2 and one experiment showed a greater role for PKN1 than PKN2 in TRAF1 S146 phosphorylation.

TRAF1 S146 is constitutively phosphorylated in primary CLL cell lines cocultured with OP9 stromal cells

CLL cells have been reported to have constitutive signaling through TNFRs such as CD40 and CD30, resulting in constitutive NF-κB activation,^21,22 which induces TRAF1¹⁰ as well as Bcl-2 and Mcl-1.¹³ To determine whether TRAF1 is phosphorylated in CLL cells, we conducted western blot analysis of PBMC from CLL patients, after overnight culture of the cells with OP9 stromal cells, to maintain CLL viability during the recovering time after thawing cells. The majority of cells in the PBMC are CLL cells and our previous results have shown that neither OP9 stromal cells or resting normal lymphocytes express TRAF1.¹³ We analyzed CLL cells harboring TP53 mutations, which are associated with a poor prognosis; CLL cells with the 13q translocation, which has a more favorable prognosis²³; and CLL samples for which cytogenetic information was not available (Fig. 4A). The results show that TRAF1 pS146 is constitutively present in all OP9-cultured CLL patient samples analyzed, albeit with some variability between samples in the ratio of phospho-TRAF1 to total TRAF1, likely reflecting differences in the relative contributions of transcriptional (NF-κB dependent) versus posttranslational (PKN1/2 dependent) regulation across individual patients. Moreover, this TRAF1 pS146 signal was sensitive to treatment with OTSSP167, with an IC₅₀ of 39 nM (Fig. 4B–F) in CLL cells with 13q or 17p translocations, as well as in CLL cells with unknown cytogenetics.

Figure 4.

OTSSP167 decreases total and phospho-specific TRAF1 protein levels in primary CLL cells. (A) Lysates from CLL cells from patients with cytogenetics not available (unknown; n = 3 patients), 13q (n = 2 patients), or p53 mutations (TP53) (n = 3 patients) were subjected to western blot analysis for pS146 specific TRAF1 (p-TRAF1), total TRAF1 (TRAF1), PKN1, or GAPDH. (B) CLL cells from representative donors with unknown orTP53 mutations (n = 2 different donors) from (A) were treated with DMSO control or OTSSP167 for 24 hours at the indicated concentrations. Lysates were then immunoblotted for phospho-TRAF1, total TRAF1 (TRAF1), PKN1, or GAPDH. (C) CLL patient cells with cytogenetics not available, 13q deletion, or 17p deletion (del[17p]; results in loss of WT p53) were treated with OTSSP167 at doses ranging from 10 to 50 nM for 24 hours. Lysates were then immunoblotted for phospho-TRAF1, total TRAF1 (TRAF1), PKN1, or β-actin. (D, E) Densitometry of TRAF1 (D) and p-TRAF1 (E) band intensity from 3 CLL patient cells as in (C) were quantified by ImageLab software. (F) IC₅₀ of total TRAF1 (TRAF1) and phospho-S146 TRAF1 (p-TRAF1) were calculated from densitometry analysis shown in (D) and (E). **P < 0.01; ***P < 0.001, ****P < 0.0001, One-Way ANOVA with Dunnett's multiple comparisons test.

Discussion

The results presented in this study validate that a TRAF1 pS146–specific antibody raised against a phospho-peptide (Abcam clone EPR25987-11) is completely dependent on the phosphate modification of S146 on the intact TRAF1 protein in cells. Using gene knockout in the RAJI lymphoma cell line, we show that both PKN1 and PKN2 can contribute to this phosphorylation. The kinase inhibitor OTSSP167 has a picomolar affinity for MELK,¹⁵ but an IC₅₀ for PKN1 in the nanomolar range.¹³ We previously showed that the IC₅₀ for TRAF1 loss upon inhibition with OTSSP167 was around 30 nM in primary CLL cells, which is similar to the IC₅₀ for PKN1.¹³ Here we show that treatment of HEK293 cells or primary CLL with OTSSP167 results in loss of the S146 phosphate group on TRAF1, with a similar IC₅₀. These data are consistent with PKN1 phosphorylating TRAF1 at serine 146 in primary CLL to promote TRAF1 stability.

Our previous study showed that loss of TRAF1 in response to OTSSP167 resulted in loss of Bcl2, Mcl-1, increased activated caspase 3, and cell death, with a similar IC₅₀ for each readout.¹³ As PKN1 and PKN2 are 50% identical in their nucelotide binding domain, the location of OTSSP167 binding, it seems likely that OTSSP167 is targeting both kinases.¹⁷ PKN2 is reported to be ubiquitously expressed in human tissues¹⁷ and was present in the RAJI cell line used here, although the relative contributions of PKN1 and PKN2 to constitutive TRAF1 phosphorylation in primary CLL remain to be determined. Taken together, these studies validate TRAF1 pS146 staining as a readout for inhibition of PKN1/2 in cell lines as well as primary CLL cells. This antibody should therefore be useful for studies of TRAF1 function in CLL and for screening new PKN1/2 inhibitors for on target effects in CLL.

Our previous work used loss of total TRAF1 protein in response to OTSSP167 to show that PKN1 is important for TRAF1 stability during constitutive signaling in primary CLL cells. Here we validate this through direct measurement of TRAF1 pS146. We also provide evidence here that TRAF1 is constitutively phosphorylated in primary CLL cultured on OP9 stromal cells, including those with 13q translocations that have a more favorable prognosis, and those with del(17p) or p53 mutations that are associated with poor response to conventional therapies.

TRAF1 is known to form heterotrimers with TRAF2 that bind activated TNFRs, with TRAF2 providing E3 ubiquitin ligase activity that modulates downstream signaling.^7,12 Our previous study showed that addition of the OTSSP167 PKN/MELK inhibitor reduces the level of pS6, pErk, pNF-κB p65, Mcl-1, and Bcl-2 in RAJI and primary CLL.¹³ Although not formally demonstrated here, our previous work showed that PKN1 knockdown leads to cIAP-dependent degradation of TRAF1 during CD40 signaling,¹³ suggesting a model in which S146 phosphorylation protects TRAF1 from cIAP-mediated ubiquitination and degradation, thereby potentiating CLL survival.

We note that OP9 co-culture was used to maintain CLL cell viability ex vivo, as primary CLL cells undergo rapid spontaneous apoptosis without stromal support. However, we cannot completely exclude the possibility that OP9 co-culture modulates the level of S146 phosphorylation in the CLL cells.¹³

We observed some variability in the ratio of TRAF1 pS146 to total TRAF1 across CLL patient samples, which may reflect the combination of regulation at both the transcriptional and protein levels. Studies with a larger number of samples are required to determine if TRAF1 levels differ among different subgroups of CLL.

Author contributions

T.H.W. and A.A.A.-S. conceptualized and obtained funding for the study and wrote and edited the manuscript. M.D.M. provided access to CLL samples, advised on interpretation, and edited the manuscript. B.G. conducted the investigations and edited the manuscript. L.N. assisted with investigation and manuscript preparation.

Birinder Ghumman (Data curation [Lead], Formal analysis [Equal], Writing—review & editing [Supporting]), Laura Nicolucci (Investigation [Supporting], Writing—review & editing [Supporting]), Mark D. Minden (Resources [Supporting]), Tania H. Watts (Conceptualization [Equal], Formal analysis [Equal], Funding acquisition [Equal], Project administration [Equal], Supervision [Equal], Writing—original draft [Equal], Writing—review & editing [Equal]), and Ali A. Abdul-Sater (Conceptualization [Equal], Formal analysis [Equal], Funding acquisition [Equal], Investigation [Equal], Project administration [Equal], Supervision [Equal], Writing—original draft [Equal], Writing—review & editing [Equal])

Funding

This research was supported by Canadian Institutes of Health Research Grant FDN-143250 to T.H.W. and York Research Chair to A.A.A.-S. In addition, T.H.W. holds a Tier I Canada Research Chair in antiviral immunity at the University of Toronto. The GEMb facility is supported by Canada Foundation for Innovation CFI-36490, CFI-37607, and CFI-36940.

Conflicts of interest

The authors have no conflicts of interest to declare.

Data availability

Raw data will be made available upon request.