Almost 25 years ago we first employed antisense oligonucleotides (oligos) against human derived prostatic LNCaP cells using both in vitro  and in vivo  models; initially targeting the epidermal growth factor receptor, and its ligand, transforming growth factor-alpha. Similar studies and results were found when we treated in vitro and in vivo breast cancer models . Clinically, oligos have been employed to treat cancer patients (including those with prostate tumors), targeting apoptosis inhibitors (bcl-2, clusterin) in attempts to restore tumor chemo-  or radio-sensitivity .
Yet, in spite of these patient trials and additional advances in early detection of prostate cancer, its treatment has not greatly improved in recent years. In the last few years gene therapy and newly discovered immune checkpoint blockade has given some indication that they could provide some additional improvement, particularly when administered following surgery, chemotherapy or irradiation.
As the role of the immune system in cancer treatment becomes better understood (and augmented), some oligos have been found to alter the immune response towards prostate tumors, in previously un-thought of ways. Using oligos (either mono- or bispecific) targeting bcl-2 we found that certain antisense oligos have secondary conformations based on intra strand sequences which permit complementary base pair binding within the oligo, that can induce interferon , enhance cell surface antigen expression [prostate specific membrane antigen (PSMA)] , and potentially increase tumor recognition and targeting by the immune system.
However, our studies also indicate that gene therapy employing oligos directed towards bcl-2 (in LNCaP cells) frequently are compensated for by altered regulation of apoptosis, increased androgen sensitivity, enhanced oncogene activity [7-9] and chromosomal instability as indicated by enhanced expression of fusion protein TMPRS22  and fusion transcription protein FLI-1 . All compensation mechanisms noted here suggest increased proliferation, chromosomal instability and greater mitotic activity, producing a more aggressive tumor, or selection of a subset of such cells. This is confirmed by finding greatly increased expression of proliferation factor KI-67  and mitosis related cyclin D1  after treatment directed at bcl-2.
We now believe that immunologic recognition can be an additional pathway for compensation following suppressive bcl-2 treatment and suggest that this type of gene therapy could also influence proteins associated with immune checkpoint blockade, altering its efficacy. Immune checkpoint blockade therapy has become the “standard of care” treatment for melanoma and is now being evaluated (and sometimes recommended for first line treatment) against kidney, lung, hematologic cancers as well as solid tumors (including those of the prostate). Melanoma has a long history of response to immunotherapy, first using in vitro cultivated lymphokine activated killer (LAK) T cells and then employing similarly cultivated and expanded tumor infiltrating lymphocytes (TIL). The addition of additional cytokines (or their cellular production) enhanced the technique. More recently melanoma became the first tumor to be extensively treated by monoclonal agents in checkpoint blockade, where its prolonged survival and even produced some cures in a small percentage of patients. Studies are now in progress to identify those patients most susceptible for this treatment (those having greater tumor surface expression of PD-1) and even ways to alter susceptibility through evaluation of the patient’s colon microbiota (with possible alteration using fecal transplantation).
In a continuation of these studies, and also to see whether compensation has additional effects on the immune response, we found that PD-1 and its ligand PD-Ll were significantly enhanced in their expression following treatment with both mono- and bispecific oligos directed towards bcl-2. Similarly, the prognostic indicator cyclin derived kinase-12 (CDK-12), which identifies a subpopulation of prostate tumors susceptible to immunotherapy was similarly enhanced. The lack of CDK-12 expression in untreated LNCaP cells suggests that this human derived tumor would be susceptible to immunotherapy for in vivo studies. The fact that treatment enhanced both the expression of PD-1 and PD-L1 targets, in addition to CDK-12 shows the complexity of immune therapy using checkpoint inhibitors, which is far from understood. The greatly enhanced expression of CDK-12 would suggest that the LNCaP cells would no longer be susceptible to checkpoint inhibition at the PD-1/PD-L1 level even though these proteins (and particularly PD-1) present a better target. The huge increase in CDK-12 found in the treatment’s groups would probably overwhelm such recognition.
To identify a compensatory response to evade apoptosis in the presence of bcl-2 suppression, levels of mRNA encoding non-targeted bax, caspase-3, clusterin and VDAC1were next evaluated. We found that specific suppression of the apoptosis inhibitor bcl-2 in LNCaP cells does not affect (non-targeted) bax expression nor (non-targeted) clusterin (non-targeted). Non-targeted caspase-3 expression was suppressed while non-targeted VDAC1 appeared to increase. This suggested that tumor cell variants develop which resist apoptosis through diminished expression of promoters, while other mechanisms which facilitate apoptosis may be enhanced. This study suggests that compensatory changes in the regulation of apoptosis can vary or be limited to apoptosis promoters (caspase-3), since the expression of the non-targeted apoptosis inhibitor clusterin is not affected and there is suggestion that the activator of apoptosis VDCA1 could be enhanced. Should bcl-2 suppression be clinically employed with antisense oligos it may require maintenance (or replacement) of caspase-3 activity.View pdf