Downregulation of PIK3CA via antibody-esiRNA-complexes suppresses human xenograft tumor growth

Abstract

Precision cancer therapy requires on the one hand detailed knowledge about a tumor's driver oncogenes and on the other hand an effective targeted therapy that specifically inhibits these oncogenes. While the determination of genomic landscape of a tumor has reached a very precise level, the respective therapy options are scarce. The application of small inhibitory (si) RNAs is a promising field of investigation. Here, we present the effective in vivo-treatment of colorectal cancer (CRC) xenograft tumors with antibody-complexed, endoribonuclease-prepared small inhibitory (esi)RNAs. We chose heterogeneous endoribonuclease-prepared siRNA pools (esiRNAs) against the frequently mutated genes PIK3CA and KRAS and coupled them to the anti-EGFR antibody cetuximab, which was internalized specifically into the tumor cells. esiRNA pools have been shown to exhibit superior specificity in target gene knockdown compared to classic siRNAs. We identified a significant decrease in tumor growth upon this treatment due to decreased tumor cell proliferation. The ex vivo-analysis of the xenograft CRC tumors revealed the expected downregulation of the intended direct targets PIK3CA and KRAS on protein level. Moreover, known downstream targets for EGFR signaling such as p-ERK, p-AKT, and c-MYC were decreased as well. We therefore propose the use of antibody-esiRNA complexes as a novel experimental treatment option against key components of the EGFR signaling cascade.

Fig 1. Cetuximab-sulfo-SMCC-protamine ability to transport two independent siRNAs.

A-I. Alexa 488 and Alexa 555 control-siRNA were mixed, coupled to cetuximab-sulfo-SMCC-protamine and incubated with the indicated cell lines at 60 nM end concentration overnight (Fig 1 A to I) or 48 h (J to R). Cells were washed, fixed, stained with Hoechst and mounted. Fluorescent micrographs showed vesicular enrichment of both Alexa488 and -555 signals within the same vesicles. S. SW480 cells were treated with cetuximab (“C”)-sulfo-SMCC-protamine coupled to control- (“C-cntr-esiRNA”), KRAS-specific or PIK3CA-specific esiRNAs at 60 nM end concentration for 3 consecutive days. Cells were washed, harvested and lysates were subjected to Western blot against KRAS, PIK3CA and actin as loading control. C-KRAS-esiRNA treatment reduced KRAS protein expression, whereas C-PIK3CA-esiRNA treatment reduced PIK3CA protein expression.T—V. Soft agar colony forming assays of CRC cell lines DLD1 (T) and SW480 (U) treated with cetuximab-coupled control esiRNA ("C-scrm" = C-cntr-esiRNA), anti-KRAS-, anti-PIK3CA- or both esiRNAs. Results showed significant reduction of colony growth with C-KRAS-esiRNA and C-PIK3CA-esiRNA, but only SW480 showed additive effect with both esiRNAs. HT29 (BRAF-mutated, PIK3CA-mutated) cells (V) were affected by C-PIK3CA-esiRNA, but not by C-KRAS-esiRNA treatment. Shown are number of colonies normalized to PBS-treated cells of three independent experiments, mean +/- SD. * p<0.05.

Fig 2. CRC xenograft growth control by cetuximab-sulfo-SMCC-protamine coupled to control-, KRAS- and/or PIK3CA-esiRNA treatment.

A. Time schedule and treatment regimen. CRC cell lines were transplanted subcutaneously to CD1 nude mice and tumor growth observed to a size as indicated. Treatment started day 0 after randomization of the animals to the different groups and was continued twice a week by i.p. applications of PBS, cetuximab-sulfo-SMCC-protamine coupled to control esiRNA (C-cntr-esiRNA) or the indicated esiRNAs directed against KRAS (C-KRAS-esiRNA) and PIK3CA (C-PIK3CA-esiRNA) or both (C-KRAS/PIK3CA-esiRNA). C-KRAS-esiRNA treatment was effective in DLD1 (B) and SW480 tumor (C) size control, but not in HT29 (BRAF-mutated, PIK3CA-mutated; D) as compared to controls. C-PIK3CA-esiRNA treatment was effective in SW480 and HT29, but less effective in DLD1. C-KRAS/PIK3CA-esiRNA combination treatment reduced tumor growth in the xenografts of all cell lines. Graph shows median tumor volumes normalized to day 0 +/- SEM. Detailed statistical analysis is presented in Table 1. Growth curves and tumor weight for C, C-cntr-esiRNA (C-scr in E-G) and C-KRAS-esiRNA treatment in B-G have been published before [35] and are reproduced here as controls with permission of the AACR to reduce the number of control animals according to the 3R rules [37]. E-G. Ex vivo photographs of dissected tumors and excised tumor weight statistics reveal significant reduction of tumor size and weight compared to controls in C-KRAS-esiRNA treated DLD1 and SW480 tumors, but not in HT29. In turn, C-PIK3CA-esiRNA reduced tumor weight significantly in all tumors along with additive effects of C-KRAS/PIK3CA-esiRNA combination treatment. Graph shows absolute tumor weight +/- SD. Significance * = p < 0.05.

Fig 3. Treatment of CRC xenografts with cetuximab-protamine siRNA controls tumor cell proliferation.

Immunohistochemical (IHC) analysis of Ki67 proliferation marker in treated tumor sections. Tumor samples from the experiments described in Fig 2 were embedded in paraffin, sectioned by microtome, rehydrated and subjected to IHC for Ki67 marker. Sections are presented at 20x magnification. All tumors treated with control esiRNA coupled to cetuximab (A, E, I: C-cntr-esiRNA) showed Ki67 nuclear signals, whereas C-KRAS-esiRNA treatment reduced Ki67 staining markedly in DLD1 (B) and SW480 (F), but not in HT29 (BRAF-mutated, PIK3CA-mutated; J). C-PIK3CA-esiRNA treatment reduced Ki67 staining in all tumor groups (C, G, K) with some further reduction by C-KRAS/PIK3CA-esiRNA combination treatment (D, H). In HT29 samples KI67 staining was further reduced by the combination (L) as compared to C-PIK3CA-esiRNA (K).

Fig 4. Treatment of CRC xenografts with cetuximab-protamine-siRNA markedly reduces siRNA target gene expression.

A-P. Paraffin sections from DLD1 tumors (A-H) and HT29 tumors (I-P) were prepared for immunohistochemical (IHC) analysis for siRNA targets KRAS and PIK3CA with antibody detecting KRAS (A, C, E, G and I, K, M, O) and PIK3CA (B, D, F, H and J, L, N, P) combined with suitable secondary antibodies and stained with diaminobenzidine (DAB) and hematoxylin and pictures taken at 20x magnification from regions without signs of necrosis. The application of KRAS siRNA coupled to cetuximab-sulfo-SMCC-protamine (C-KRAS-esiRNA) markedly reduced KRAS immunostaining (C) in DLD1 tumors, but not in HT29 (BRAF-mutated, PIK3CA-mutated; K) compared to control C-cntr-esiRNA (A, I). The treatment with C-PIK3CA-esiRNA reduced PIK3CA immunostaining in DLD1 (F) as well as HT29 tumors (N), as compared to control C-cntr-esiRNA in both cell lines (B, J). Interestingly, C-KRAS-esiRNA treatment also reduced PIK3CA staining in DLD1 tumors (D), but not in HT29 tumors (L). The combination treatment of tumors with KRAS- and PIK3CA-esiRNAs (C-KRAS/PIK3CA-esiRNA) resulted in even less KRAS and PIK3CA staining than the C-KRAS-esiRNA treatment in DLD1 tumors (G-H) and similar stainings of both in HT29 cells(O-P) as the C-PIK3CA-esiRNA monotherapy (M-N). Q-S. Western blots indicating siRNA target gene induced protein synthesis control in xenograft tumor tissue of cetuximab-protamine-esiRNA treated mice. Tumor tissue was processed for western blot as described, applied to SDS-PAGE, blotted and exposed for immunodetection by antibodies raised against KRAS, PIK3CA and actin as loading control. Application of cetuximab-protamine coupled to KRAS-esiRNA (C-KRAS-esiRNA) reduced KRAS protein levels in DLD1 (Q, top row), HT29 (R, top row) and SW480 (S, top row) tumor xenografts as compared to controls (actin row). In addition, C-KRAS-esiRNA treatment showed certain crosstalk to PI3K pathway signaling (third row from above) as indicated by reduced PIK3CA expression in DLD1 (Q), enhanced PIK3CA expression in HT29 (R) tumors and indifferent PIK3CA expression effect in SW480 (S) as compared to actin loading controls. C-PIK3CA-esiRNA treatment lead to reduced PIK3CA detection levels in all three xenograft tumor types (Q-S, third row from above) with even more pronounced suppression of PIK3CA protein by C-KRAS/PIK3CA-esiRNA combination in DLD1 and HT29 (Q-R, third row from above).

Fig 5. MAPK and AKT pathway inactivation by cetuximab-KRAS- and cetuximab-PIK3CA-esiRNA treatment.

Tumor lysate samples from xenograft experiments were processed for SDS PAGE and probed in Western blots for expression of KRAS downstream MAPK pathway effector phospho-ERK and PIK3CA downstream phospho-AKT along with their unphosphorylated (total) counterparts in DLD1 (A) and HT29 (B) tumors. In DLD1 tumors (A) treated with C-KRAS-, C-PIK3CA- and C-KRAS/PIK3CA-esiRNA, there was a clear reduction of ERK phosphorylation visible in contrary to a control siRNA treatment (top row) and compared to total ERK levels, whereas BRAF-mutated HT29 tumors (B) treated with C-KRAS-esiRNA did not response in terms of ERK phosphorylation. Here, C-PIK3CA- and C-KRAS/PIK3CA-esiRNA treatment elicited reduced phosphorylation. The same western blot membranes were stripped and probed for phosphorylated AKT (p-AKT, third row from above) and total AKT (fourth row from above). Here, C-KRAS-esiRNA treatment did not change AKT phosphorylation in DLD1 (A) as well as HT29 (B) tumors as compared to control siRNA treatment, but C-PIK3CA-esiRNA treatment and C-KRAS/PIK3CA-esiRNA markedly reduced phosphorylation of AKT indicating deactivation of PI3K signaling pathway. Actin was detected as loading control for total protein.

Fig 6. MAPK pathway suppression by cetuximab-KRAS siRNA treatment regulates c-MYC transcription factor levels in treated tumor samples.

(A-B) Cetuximab-siRNA treated tumor lysate samples from xenograft experiments were processed for Western blot and probed for c-MYC protein expression controlled by actin loading control. In DLD1 tumor samples (A) treated with C-KRAS-esiRNA and C-KRAS/PIK3CA-esiRNA, c-MYC protein level almost completely disappeared (top row), whereas the c-MYC protein expression in HT29 tumors was lower and was not consistently reduced by C-KRAS-, but by C-PIK3CA-esiRNA and C-KRAS/PIK3CA-esiRNA treatment. (C-Q) DLD1 tumor samples showing a clear c-MYC dependency on active MAPK-signaling as shown above in Western blot were paraffin embedded, sectioned and processed for double-immunofluorescence imaging. Simultaneous detection of c-MYC (red) and KRAS (green) protein on the same sections and frames with respective primary antibodies showed a correlation between markedly reduced KRAS cytoplasmic stain by C-KRAS-esiRNA treatment (G) and diminished c-MYC nuclear stain (F). C-PIK3CA-esiRNA treatment did not reduce KRAS (J) or c-MYC (I) levels. The combination treatment C-KRAS/PIK3CA-esiRNA showed less pronounced effects as compared to C-KRAS-esiRNA on both KRAS (M) and c-MYC expression (L). The right column shows Hoechst nuclear stain of the left and middle frames.O-Q show secondary antibody controls.