Towards targeted combinatorial therapy design for the treatment of castration-resistant prostate cancer

Abstract

Background: Prostate cancer is one of the most prevalent cancers in males in the United States and amongst the leading causes of cancer related deaths. A particularly virulent form of this disease is castration-resistant prostate cancer (CRPC), where patients no longer respond to medical or surgical castration. CRPC is a complex, multifaceted and heterogeneous malady with limited standard treatment options.

Results: The growth and progression of prostate cancer is a complicated process that involves multiple pathways. The signaling network comprising the integral constituents of the signature pathways involved in the development and progression of prostate cancer is modeled as a combinatorial circuit. The failures in the gene regulatory network that lead to cancer are abstracted as faults in the equivalent circuit and the Boolean circuit model is then used to design therapies tailored to counteract the effect of each molecular abnormality and to propose potentially efficacious combinatorial therapy regimens. Furthermore, stochastic computational modeling is utilized to identify potentially vulnerable components in the network that may serve as viable candidates for drug development.

Conclusion: The results presented herein can aid in the design of scientifically well-grounded targeted therapies that can be employed for the treatment of prostate cancer patients.

Keywords: Boolean modeling; Combination therapy; Gene regulatory networks; Prostate cancer; Stochastic logic; Vulnerability assessment.

Fig. 1

Prostate cancer signal transduction network. A schematic diagram of key signaling pathways deregulated in prostate cancer. Black and red lines represent activating and inhibiting interactions respectively whereas the red boxes depict prostate cancer drugs at their corresponding points of intervention in the network

Fig. 2

Boolean model. Combinational circuit model of prostate cancer signaling pathways. Each node is assigned a numeric label in parentheses. These labels also serve to enumerate the fault locations with stuck-at-one and stuck-at-zero faults in black and red numerals respectively. The dotted arrows indicate the intervention points for the respective drugs

Fig. 3

Circuit with stuck-at fault. An example of a stuck-at fault. In the absence of the stuck-at fault, the output is zero. If there is a stuck-at-one fault at the location marked with a cross, the output of the faulty circuit becomes one

Fig. 4

A stochastic logic circuit. An example of a stochastic logic circuit

Fig. 5

Computation of node vulnerability. Depicts the architecture used to compute the vulnerability of a node. x ₁ to x ₇ are the input stochastic bit streams for each of the seven primary inputs in the Boolean network model. The output bit streams for each of the six output components when these input sequences are propagated through the circuit with a dysfunctional node (whose vulnerability we want to compute) are denoted by $y_1^{\ast }$ to $y_6^{\ast }$ whereas those for the fault-free circuit are labeled as y ₁ to y ₆