Abstract
Assessing the phenotypic diversity underlying tumour progression requires the identification of variations in the respective molecular interaction networks. Here we report proof-of-concept for a platform called poly-ligand profiling (PLP) that surveys these system states and distinguishes breast cancer patients who did or did not derive benefit from trastuzumab. We perform tissue-SELEX on breast cancer specimens to enrich single-stranded DNA (ssDNA) libraries that preferentially interact with molecular components associated with the two clinical phenotypes. Testing of independent sample sets verifies the ability of PLP to classify trastuzumab-treated patients according to their clinical outcomes with ROC-AUC of 0.78. Standard HER2 testing of the same patients gives a ROC-AUC of 0.47. Kaplan–Meier analysis reveals a median increase in benefit from trastuzumab-containing treatments of 300 days for PLP-positive compared to PLP-negative patients. If prospectively validated, PLP may increase success rates in precision oncology and clinical trials, thus improving both patient care and drug development.
Introduction
Precision oncology currently operates under a "one target, one drug" paradigm. A primary example is the use of trastuzumab (Herceptin) for breast cancer patients who overexpress the HER2 protein. However, this reductionist model is imperfect: approximately 38–50% of HER2-positive patients fail to derive clinical benefit, while some HER2-negative patients unexpectedly respond.
The authors propose that a tumor’s drug response is governed by its "interactome"—the complex network of molecular interactions and system states—rather than a single protein. Poly-ligand profiling (PLP) was developed to measure these global states without requiring prior knowledge of specific target proteins.
Methods: Tissue-SELEX and Library Enrichment
The core of the PLP platform is tissue-SELEX (Systematic Evolution of Ligands by Exponential Enrichment). This process generates "aptamers"—short, single-stranded DNA (ssDNA) sequences that fold into unique 3D shapes to bind molecular targets within tissue samples.
Researchers began with a massive random library of $10^{13}$ unique ssDNA sequences. Through iterative rounds of selection using formalin-fixed paraffin-embedded (FFPE) tumor samples, they refined the library into two specialized pools: EL-B (enriched for binding to "Benefit" tissue) and EL-NB (enriched for "No Benefit" tissue). These libraries were then used as histological stains. A pathologist, blinded to patient outcomes, assigned "H-scores" based on binding intensity, which were then processed via a logistic regression model to classify new patients.
Study Design and Datasets
The study utilized three distinct cohorts to train and validate the PLP platform:
- Discovery/Enrichment Set: 33 breast cancer cases with known Time to Next Therapy (TTNT) used to train the libraries.
- Independent Test Set: 61 trastuzumab-treated patients used to validate predictive accuracy.
- Control Set: 63 patients treated with chemotherapy without trastuzumab to determine if the test was specifically predictive of drug response or merely prognostic of general survival.
Key Results
PLP demonstrated a significant improvement over the current clinical standard. In the independent test set, the platform achieved an ROC-AUC of 0.78, whereas standard HER2 immunohistochemistry (IHC) testing on the same samples resulted in an AUC of only 0.47—essentially no better than a coin flip in this metastatic cohort.
Clinical impact was further highlighted by Kaplan–Meier analysis. Patients identified as "PLP-positive" experienced a median benefit of 429 days on trastuzumab-containing treatments, compared to 129 days for "PLP-negative" patients. This 300-day difference represents a substantial clinical improvement in progression-free survival.
Crucially, the PLP test failed to differentiate outcomes in the control group (patients not receiving trastuzumab). This confirms that the platform is measuring a specific molecular state related to the drug's mechanism of action, rather than general patient health.
Discussion and Limitations
PLP offers a "black box" advantage: it identifies relevant molecular differences between responders and non-responders without needing to pre-identify the target proteins. This system-wide view captures the complexity of resistance mechanisms that single-protein biomarkers miss.
However, several hurdles remain before clinical adoption. This was a retrospective study with a relatively small test set (n=61). Prospective clinical trials and larger cohorts are necessary to ensure the model generalizes across diverse breast cancer subtypes. Additionally, identifying the specific molecular targets bound by the EL-B and EL-NB libraries could provide new biological insights into trastuzumab resistance.
Summary updated