Targeting of the E3 ubiquitin-protein ligase HUWE1 impairs DNA repair capacity and tumor growth in preclinical multiple myeloma models

Experimental evidence suggests that ubiquitin-protein ligases regulate a number of cellular processes involved in tumorigenesis. We analysed the role of the E3 ubiquitin-protein ligase HUWE1 for pathobiology of multiple myeloma (MM), a still incurable blood cancer. mRNA expression analysis indicates an increase in HUWE1 expression levels correlated with advanced stages of myeloma. Pharmacologic as well as RNAi-mediated HUWE1 inhibition caused anti-proliferative effects in MM cell lines in vitro and in an MM1.S xenotransplantation mouse model. Cell cycle analysis upon HUWE1 inhibition revealed decreased S phase cell fractions. Analyses of potential HUWE1-dependent molecular functions did not show involvement in MYC-dependent gene regulation. However, HUWE1 depleted MM cells displayed increased DNA tail length by comet assay, as well as changes in the levels of DNA damage response mediators such as pBRCA1, DNA-polymerase β, γH2AX and Mcl-1. Our finding that HUWE1 might thus be involved in endogenous DNA repair is further supported by strongly enhanced apoptotic effects of the DNA-damaging agent melphalan in HUWE1 depleted cells in vitro and in vivo. These data suggest that HUWE1 might contribute to tumour growth by endogenous repair of DNA, and could therefore potentially be exploitable in future treatment developments.

Sustained inducible knockdown of HUWE1 strongly reduces proliferation of HCMLs in longer-term culture. We therefore decided to also analyse the role of HUWE1 in MM by molecular genetic approaches and generated sublines of JJN3, MM1.S, U266 and U266-MYC stably transfected with a doxycyclineinducible shRNA expression cassette targeting HUWE1. Western blot analysis revealed effective depletion of HUWE1 protein in all HMCL sublines 5 days after addition of doxycycline to the cell culture (Fig. 2c). Accordingly, real time qPCR analysis demonstrated about 80-90% decreases of HUWE1-mRNA levels (Fig. 2d). Over the 18-day study period HUWE1 knockdown strongly affected the growth curves of all MM cell lines tested, although to different degrees (Fig. 2e). Whereas HUWE1 knockdown in JJN3 cells resulted in a slight increase of their doubling time from 1.51/1.82 days (95% CI) to 1.80/2.13 days, the effect in MM1.S cells was a shift from 1.41/2.23 days to 13.06 days/infinity. Analysis of the live cell fraction in the treated cultures showed only in MM1.S a decrease of viable cells, whereas in JJN3, U266 and U266-MYC no decrease of viability could be detected ( Fig. 2f and supplementary Fig. 4a). Doxycycline treatment of parental cells was without discernible www.nature.com/scientificreports/ effect over the same time frame (shown for MM1.S and U266 cells in Supplementary Fig. 1c). U266 cells, which lack endogenous MYC expression, were more sensitive to HUWE1 knockdown (  . 3a). We therefore conclude that the effects of HUWE1 depletion on MM cell growth are not primarily mediated via MYC and MIZ1. To further characterize the observed anti-proliferative effect upon HUWE1 knockdown, we used bromodeoxyuridine (BrdU) and 4′,6-diamidino-2-phenylindole (DAPI) staining for FACS-based cell cycle analysis. HUWE1 knockdown entailed a significant reduction of the S phase in all HMCLs tested (Fig. 2f,g; supplementary Fig. 4a). In MM1.S we detected a decrease of viable cells over the analysed period of 18 days and an increase of the sub-G1 fraction in the cell cycle analysis.
HUWE1 knockdown inhibits malignant MM growth in the MM1.S/NOD scid mouse model in vivo. Next, we asked whether HUWE1 might also support malignant growth of MM cells in vivo. For these experiments, MM1.S cells, stably transfected to express luciferase for bioimaging, and harbouring a doxycyclineinducible shRNA expression cassette targeting HUWE1, were injected into the tail veins of NOD scid gamma (NSG) mice (Fig. 3a). Prior to MM cell inoculation, the mice were divided into a control group which was fed a normal grain-based diet, and a test group which received a diet containing doxycycline (200 mg/kg). Notably, in the immediate time period following injection, when MM1.S cells migrate into the bone marrow, they still express HUWE1 because it takes about 3 to 4 days of exposure to doxycycline to implement full-scale HUWE1 knockdown ( Fig. 2 and supplementary Fig. 2c). At day 7 post-inoculation the control group displayed significantly larger tumour mass than the HUWE1 knockdown group (22.14 × 10 9 ± 4.70 × 10 9 vs. 13.98 × 10 9 ± 5.61 × 10 9 photons/sec). This difference was further magnified at 14 days post injection, when the control group was at 119.38 × 10 9 ± 37.69 × 10 9 photons/sec and HUWE1 knockdown group at 70.92 × 10 9 ± 20.35 × 10 9 photons/sec (Fig. 3b). These data demonstrate that HUWE1 loss reduces tumour growth in vivo.

Loss of HUWE1 expression impairs DNA repair capacity. Based on its reported involvement in DDR
regulation, we hypothesized that the observed effects of HUWE1 knockdown on cell cycle regulation and proliferation are functionally linked to its role in DDR 2,17 . First, we assessed the effects of HUWE1 depletion on ubiquitination of γH2AX which has been shown to be important for both DSB signalling and restarting of stalled replication forks 10,17 . After irradiation of MM1.S cells (10 Gy) both γH2AX and ubiquitinated γH2AX increased to slightly lesser extents in HUWE1 knockdown cells (Fig. 4a), but this still constituted functional DSB repair signalling. Similarly, HUWE1 knockdown did not affect the capacity for DNA repair after acute irradiation damage, as measured by the % DNA tail-length in a Comet assay over a 45-min time-window, in any HMCL tested (exemplarily shown for MM1.S cells; Fig. 4b). However, we also observed that the steady state level of ubiquitinated γH2AX was decreased in MM1.S cells in the absence of HUWE1 expression, whereas γH2AX was increased (Fig. 4a, lanes 1 and 3). Next, we therefore analysed the steady state levels of DNA tail-length in different HMCLs with or without prolonged HUWE1 knockdown and found significant increases in all HUWE1depleted conditions, indicative for increased endogenous DNA damage (Fig. 4c). Finally, we analysed a number of regulatory factors known to interact with HUWE1 and to be associated with DDR (e.g. γH2AX, BRCA1, Mcl1, p53, DNA-Polymerase β [8][9][10]17 ) by Western blotting of MM cells with and without HUWE1 depletion (Fig. 4d). Whereas a strong increase of γH2AX was found in MM1.S cells (the HMCL most sensitive for growth inhibition by HUWE1 knockdown), this effect was less pronounced in U266 and absent in JJN3 or U266-MYC cells. In contrast, phospho-BRCA1 levels decreased and Mcl-1 expression increased upon HUWE1 knockdown in all HMCLs analysed (Fig. 4d). In contrast to other reports, we found no consistent evidence that HUWE1 is involved in the degradation of H2AX, HDM2, p53 or p21 (supplementary Fig. 2c), or of PCNA, BRCA1 or DNA-Pol β (Fig. 4d). Indeed, in the HMCLs tested the levels of DNA-Pol β were slightly increased upon HUWE1 knockdown (Fig. 4d). Taken together, these data indicate that the endogenous DNA damage repair is-at least in functionally distinguishable parts-negatively affected by the absence of HUWE1.
HUWE1 knockdown strongly enhances growth-inhibitory effects of the DNA-damaging anti-MM agent melphalan in the MM1.S/NSG mouse model. Melphalan, an established chemotherapeutic agent in the treatment of MM patients, damages DNA which, if not repaired, leads to induction of apoptosis in HMCLs (for review 18 ). Therefore, we investigated if HUWE1 depletion may potentiate the antimyeloma activity of melphalan. In the presence of 2.5 µm melphalan, knockdown of HUWE1 decreased proliferation of MM1.S cells in vitro from 69.41% ± 5.67 to 35.04% ± 1.24 compared to control as measured by MTT ( Supplementary Fig. 4b). We then asked, if the increased melphalan sensitivity in vitro would also translate into an enhanced anti-tumour effect in vivo. Two groups of NSG mice (either with or without doxycycline-containing diet) were injected i.v. with luciferase expressing MM1.S cells endowed with an inducible expression cassette for shRNA targeting HUWE1 (Day 0), and the cells were allowed to migrate to the bone. Thereafter, starting at day 3 all mice were injected i.p. with 5 mg/kg melphalan (at days 3, 6 and 9), and bioluminescencent imaging was performed at days 3, 7 and 10 ( Fig. 5a). Concomitant melphalan application and HUWE1 knockdown resulted in robust suppression of tumour growth (at day 7 post knockdown induction: 10.00 × 10 8 ± 1.12 × 10 8 photons/ sec vs 38.09 × 10 8 ± 12.07 × 10 8 photons/sec in the melphalan-only control group; Fig. 5a,b). Moreover, this inhibition continued over 10 days (6.57 × 10 8 ± 2.02 × 10 8 photons/sec without HUWE1 and 20.81 × 10 8 ± 6.30 × 10 8 photons/sec with HUWE1) until the termination of the experiment due to animal health. These data support our www.nature.com/scientificreports/ finding that HUWE1 is involved in endogenous DNA damage repair which represents a salvage mechanism for DNA-damaged cells after treatment with alkylating agents like melphalan.

Discussion
Therapeutic intervention targeting molecular machineries that involve the ubiquitin proteasome system has been central to recent clinical advances in MM treatment and now forms the bedrock of current MM therapies 15 . However, many functional aspects of ubiquitination, in particular its role in tumorigenesis and as a potential therapeutic target remain to be better explored. Here we have investigated the HECT-E3 ligase HUWE1, which has been implicated in a number of oncogenic processes such as MYC regulation 6,7 and DNA damage response (DDR) 6,12,17 , for its role in multiple myeloma. HUWE1 expression at either RNA-or protein level was detected in every MM cell line tested, which is in accordance with its pattern of ubiquitous expression in either normal or cancer tissue (www.prote in-atlas .org/ENSG0 00000 86758 -HUWE1 /patho logy). Analysis of HUWE1 mRNA expression data from plasma cells also showed a moderate but still marked tendency for increase from the pre-malignant stages (normal PCs, MGUS) through intramedullary MM to plasma cell leukaemia suggesting a pathogenetic involvement of HUWE1 during MM progression to more aggressive forms. Inhibition of HUWE1 activity either by the pharmacologic compound BI8622 or by shRNA-mediated knockdown led to decreased growth and/or viability in MM cell lines as well as in a substantial fraction of primary MM samples, supporting a potential role of HUWE1 for malignant expansion. Both approaches varied somewhat in the extent of their effects and the pronounced decrease in viability across the board in MM cell lines after incubation with 10 µM BI8622 may in parts be due to different on-target effects to paralogs of HUWE1, e. g. HECTD4, or to off-target effects. However, the knockdown approach also demonstrated that HUWE1 depletion entails reduced proliferation rates, as shown by reduced S-phase components in cell cycle analyses. Accordingly, application of inducible HUWE1 knockdown in an MM1.S xenotransplantation model showed for the first time that blockade of HUWE1 expression is also effective at slowing tumour growth in vivo. Our data thus support findings in other cancer entities showing a role for HUWE1 for malignant growth [5][6][7] . It was therefore surprising that neither in MM cell lines engineered for doxycycline-inducible HUWE1 knockdown nor in the MYC-overexpressing subline U266-MYC, moderate to strong HUWE1 depletion had any discernible effects on the expression levels of MYC target genes, nor on MIZ protein, an inhibitor of MYC activity. Unlike its role in colon cancer, our data thus do not imply HUWE1 as important regulator of MYC activity in MM. However, our results are compatible with recent in-depth analyses of the genetic mechanisms of MYC dysregulation in MM, which suggest that-unlike in other cancer entities-MYC expression is not correlated with proliferation but may rather support other physiological needs of plasma cell tumors 19 .
Different authors have implicated HUWE1 function in the DNA damage response. Interestingly, although we found no impairment of the immediate response to irradiation-induced acute DNA damage in HUWE1 depleted MM cells, we show for the first time that prolonged HUWE1 knockdown always led to significant accumulation of damaged DNA as witnessed by Comet-assay, clearly indicating a role for HUWE1 in the response to endogenous DNA damage. Yet, only in MM1.S cells, this effect was translated into a notable increase in γH2AX levels. At this point, it remains unclear if absence of increased levels of γH2AX in HUWE1-depleted cells might be a result of general impairment of the detection of damaged DNA, or if only in MM1.S cells single strand breaks are converted to the double strand breaks that induce formation of γH2AX.
Amongst a host of potential DDR mediators analysed, increased levels of Mcl1 and DNA polymerase β and decreased levels of phospho-BRCA1 were the only changes consistently observed in HUWE1-depleted MM cells. Mcl1 has indeed recently been shown to be targeted for degradation by HUWE1 20 , but a connection to impaired DDR has so far only been shown in the context of Mcl1 depletion 21 . In addition it has been reported that Mcl1 acts as a functional switch between NHEJ and HR DNA repair pathways 22 . Likewise, the functional relationship between HUWE1 and BRCA1 remains puzzling, because in contrast to breast cancer cells 23 , HUWE1 depletion did not affect levels of BRCA1 itself. Similar to the absence of increased γH2AX, though, the strong decrease of  www.nature.com/scientificreports/ phosphorylated BRCA1 again implies defective detection of DNA damage or impaired initiation of DNA damage responses in HUWE1-depleted cells. Our suggestion is that the BRCA1 phosphorylation may be facilitated by a ubiquitin tag from HUWE1, but that this modification is not essential for DDR after IR exposure.
Recently, it was demonstrated that inhibition of nucleotide excision repair increases sensitivity of HMCLs to the DNA-alkylating anti-MM agent melphalan 24 . Because the observed antiproliferative effects of HUWE1 depletion in MM cells also appear to be mediated by impaired repair of endogenous DNA damage, it was rational to test melphalan effectivity within the context of HUWE1 knockdown. While this combination only produced a modest effect on viability in vitro, MM tumour growth in vivo continually decreased over the length of the study indicating that knockdown of HUWE1 sensitized the cells to melphalan-induced cell death. Genome-wide mutation analysis in newly diagnosed MM patients identified driver mutations in HUWE1. The majority of the reported mutations affect splice sites in the HUWE1 mRNA with the consequence of HUWE1 inactivation 25 . In light of our data, these would suggest, that MM patients with HUWE1 inactivation display a reduced DNA repair capacity. Therefore HUWE1 would act as a tumor suppressor assisting in DNA repair in MM cells which due to the high replicative stress display increased DNA damage 26,27 .
In summary, our data support a role for HUWE1 in the regulation or execution of endogenous and melphalan-induced DDR, and in this capacity may contribute to the malignant phenotype of MM. Pharmacological targeting of HUWE1 could be an attractive option to increase effectivity of MM therapies mechanistically related to induction of DNA damage. www.nature.com/scientificreports/ 1 mM sodium pyruvate, 2 mM L-glutamine, 2 mg/l glucose, 100 µg/ml gentamicin and with 2 ng/ml interleukin 6 30 .

Lentivirus production and HMCL transduction. Lentivirus production and transduction of MM cell
lines was performed as previously described 28 .
Western blot. Protein expression analysis was performed as previously described 32   MTT assay. Cell sensitivity to melphalan with and without HUWE1 depletion was measured using an MTTbased assay. Metabolic activity was quantified by adding 10 μl Thiazolyl Blue Tetrazolium Bromide (final concentration 0.5 mg/ml) (Sigma-Aldrich). After 4 h 100 μl solubilization solution (10% SDS in 0.01 M HCl) was added and after overnight incubation at 37 °C absorbance of the solubilized reduced formazan was measured at 570 nm with a microtiter plate reader (Sunrise, Tecan).
Alkaline single-cell electrophoresis "Comet Assay". Detection and quantification of DNA damage was assayed using the Comet Assay essentially as described 33 . qPCR, real time PCR. RNA extraction was performed using the NucleoSpin RNA isolation kit (Macherey-Nagel). cDNA was synthesized from 1 μg total RNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). Each cDNA sample was pre-diluted 1:20 and 4 μl were used with 4 × Luminaris Color HiGreen qPCR Master Mix (Thermo Fisher Scientific) and 250 nM primer for the qPCR reaction. A two-step cycling protocol was performed as follows: UDG pre-treatment for 2 min (50 °C), initial denaturation for 10 min (95 °C), 40 cycles of 15 s at 95 °C, 40 s at 60 °C (CFX Connect Thermal Cycler, Bio-Rad). The relative quantity of the target mRNA was normalized to beta-2-microglobulin). The fold changes in RNA expression were calculated using the 2 −ΔΔCt method 28 . Primer sequences are listed in supplementary information.

Statistical analysis.
A two-tailed Student's t-test was applied to perform statistical analysis. Results were considered significant at p < 0.05. All experiments were performed as independent biological replicates at least three times. Calculations were performed with Prism GraphPad 8.0. All data are presented as mean ± SD.

Ethics declaration.
All animal experiments were carried out in accordance with relevant guidelines and regulations and approved by the Regierung von Unterfranken AZ 55.2 2532-2-335. Primary human MM cells were obtained from routine diagnostic bone marrow aspirates of patients after informed consent. All experimental protocols were approved by the Ethics Committee of the University of Würzburg (AZ 76/13) and carried out in accordance with relevant guidelines and regulations. www.nature.com/scientificreports/