Depleting the 19S proteasome regulatory PSMD1 subunit as a cancer therapy strategy

Abstract

Background: Proteasome inhibitors are in use in treating certain types of cancers. These drugs inhibit the catalytic activity of the 20S proteasome, shared by all the different proteasome complexes. Inhibitors of the 26S-associated deubiquitinating activity explicitly inhibit the 26S proteasomal degradation of ubiquitinylated substrates. We have previously reported an alternative strategy that is based on reducing the 26S/20S ratio by depleting PSMD1, 6, and 11, the subunits of the 19S proteasome regulatory complex. Given the addiction of the many cancer types to a high 26S/20S ratio, the depletion strategy is highly effective in killing many aggressive cancer cell lines but not mouse and human immortalized and normal cells.

Methods: We used two aggressive cell lines, MDA-MB-231, a triple-negative breast tumor cell line, and OVCAR8, a high-grade ovary adenocarcinoma. Cell culture, mouse MDA-MB-231, OVCAR8 xenografts, and patient-derived ovarian cancer xenograft (PDX) models were transduced with lentivectors expressing PSMD1 shRNA. Tumor size was measured to follow treatment efficacy.

Results: Using different experimental strategies of expressing shRNA, we found that PSMD1 depletion, either by expressing PSMD1 shRNA in an inducible manner or in a constitutive manner, robustly inhibited MDA-MB-231, and OVCAR8 xenograft tumor growth. Furthermore, the PSMD1 depletion strategy compromised the growth of the PDX of primary ovarian cancer.

Conclusion: Our results suggest that reducing the 26S/20S ratio might be a valuable strategy for treating drug-resistant aggressive types of cancers.

Keywords: 19S regulatory particle; 26S proteasome; breast cancer; cancer therapy; mouse xenografts; ovarian cancer PDX.

FIGURE 1

Triple‐negative breast cancer MDA‐MB‐231 and the high‐grade ovary adenocarcinoma OVCAR8 cell lines are addicted to high 26S proteasome levels. (A) The levels of PSMD1 and polyubiquitinated proteins in MDA‐MB‐231 and OVCAR8 cells harboring a doxycycline‐inducible PSMD1 shRNA before and after doxycycline (dox) treatment. The extracts were immunoblotted with PSMD1, ubiquitin, and GAPDH antibodies. The quantifications of three replicates are shown. (B) Microscopic images of control and dox‐treated microscopic images of cells described under panel A, shown by bright field (BF) and RFP expression. RFP is expressed only when doxycycline induction is effective and thus monitors shRNA expression. (C) Growth of MDA‐MB‐231 and OVCAR8 cells expressing dox‐inducible PSMD1 shRNA was analyzed using the XTT assay (N = 3). (D) The results obtained by the XTT, cell viability assay, were confirmed by in situ cell monitoring using the Incucyte® SX1 live‐cell analysis system (N = 3). In (C) and (D) standard error means are lower than 5%, and p < 0.001 from Day 4 and on. (E) Cell death was quantified by measuring the subG1 fraction (marked by an arrow) by FACS analysis.

FIGURE 2

shPSMD1 depletion effectively reduced the tumor size in a xenograft model. (A) Schematic presentation of the experimental design of the three treated mice groups labeled 1–3 in all the panels. Ten female nude mice (HsdHli:CD1‐Foxn1nu; 6–7 weeks old) were obtained from Envigo per group. The red line indicates the dox‐dependent induction of expression, as visualized by live imaging (cf. the mice of Group 2 to Group 3). Either control or inducible shPSMD1 cassette harboring cells (4 × 10⁶) were subcutaneously injected into the right back of each mouse on Day 1. To induce shPSMD1 expression, mice were treated with doxycycline (1 mg/mL) in drinking water starting from Day 5 after injection. (B) Caliper measurements examined tumor growth rate. Tumor volume was calculated as X2Y/2 (X is the smallest tumor dimension). (C) Statistical calculation of the average tumor volume, on Day 29, by boxplot demonstration. *p < 0.05, **p < 0.01. (D) Luciferase live imaging of the three groups of mice was conducted at the end of the experiment (Day 26 after injection of tumor cells) using IVIS SPECTRUM. *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 3

Intratumor LV‐shPSMD1 injection reduced tumor size. (A) GFP expression (the red spot) in the tumors after intratumor injection of the lentivector expressing GFP. (B) Schematic presentation of the design of the experiments. MDA‐MB‐231 cells (4 × 10⁶) were injected subcutaneously into the right back of each mouse on Day 1. Group 1 was intratumorally injected with the control lentivector, and Group 2 with LV‐shPSMD1 and GFP (2–10 × 10⁶TU). The second control group was injected with HBSS physiological solution only. The injection was repeated on the days labeled by an arrow. The expression of GFP, as detected by live imaging, validated the presence of the LV‐shPSMD1 in the tumor site. (C) Caliper measurements examined tumor volume. Tumor volume was calculated as X2Y/2 (X is the smallest tumor dimension). (D) The statistically calculated average tumor size by boxplot on Day 36. *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 4

LV‐shPSMD1 expression effectively reduced the OVCAR8 tumor size in the xenograft mice model. (A) Schematic presentation of the experimental design of the three different treated mice groups, numbered 1–3 in all the panels. The red line indicates the dox‐dependent induction. (B) Either control or inducible LV‐shPSMD1 cassette cells (8 × 10⁶) were subcutaneously injected into the right back of each mouse on Day 1. To activate LV‐shPSMD1, we used doxycycline (dox) supplementation (1 mg/mL) in drinking water starting from Day 5 after injection. Caliper measurements examined tumor growth. (C) The statistically calculated average tumor size by boxplot on Day 70. *p < 0.05, **p < 0.01.

FIGURE 5

Treating the mice bearing PDX xenograft tumors with LV‐shPSMD1 injections. (A) NSG NSG mice were implanted with metastatic high‐grade ovarian carcinoma PDX model. Treatments started when the tumors reached ~20 mm³ sizes. Either constitutively active LV‐shPSMD1 or irrelevant sequence‐based virions (~5 × 10⁶TU) were injected each time (the blue arrows) intratumorally. The volume of the tumors was measured as described above. For each group of treatment, 9–11 mice were used. (B) The boxplot's statistically calculated average tumor size is indicated in the right panel. The growth rate with the time of the tumor is lower in the treatment compared to the control (p = 0.0001). The p‐value of the growth rate is of the interaction term in a mixed effect linear model with the time and the interaction between the time and the treatment as the fixed effects, and the mouse as the random effect.