Metastasis Prediction: Mechanobiology-Based Early Determination of Metastatic Risk in Pancreatic Tumors

Introduction

The main cause of cancer-related deaths is metastasis. A critical step in metastases formation is invasion of cancer cells through tissue. Invading metastatic cells are dynamic, rapidly changing morphology and applying forces to their surroundings. Evaluating forceful interactions of cells on an impenetrable, synthetic gel, we have previously shown that subsets of invasive breast-cancer cells indent custom gels while benign/non-invasive cells do not. Specifically, we rapidly and quantitatively evaluate the cells’ invasive capacity by the amounts of indenting cells and their attained depths; the mechanical invasiveness reliably separates cells with high and low metastatic potential (MP). Here, we show that the mechanical invasiveness of fresh, human, primary-site pancreatic tumors agrees with the clinical histopathology and matches results in established cell lines.

Materials and methods

We have seeded cells from 6 pancreatic cell lines (ATCC, Manassas, VA) or from fresh, human pancreatic tissue samples (enzymatically degraded) on physiological-stiffness (2.4 kPa) polyacrylamide gels. Within 1-hr of seeding, the fraction of invasive cells indent the gels, their amounts and depths are determined by fluorescence microscopy and provide the mechanical invasiveness. In cell lines, we compare to amount of cells trespassing 8µm pores of trans-well migration assay within 72 hrs. For tumor samples, clinical histopathology and patient follow-up are used as gold-standard.

Results and discussion

The mechanical invasiveness of pancreatic cell-lines with high- and low-MP differed in a statistically significant manner, respectively, with 83-85% and 46-53% percent indenting cells and attained depths of 4.7-5 µm and of 2.9-4.1 µm; results agree with the trans-well migration assay. We show agreement with fresh, human tumor samples, where invasive samples (e.g. metastatic adenocarcinoma) exhibit >40% indenting cells and benign/non-invasive samples (e.g. non-invasive fibrotic tumors or pre-cancerous lesions) are at <20%.

Conclusions

Our unique mechanobiology-based approach provides rapid (2hr) cancer identification and prediction of metastatic likelihood of a tumor, already during the first diagnosis, which agrees with the clinical histopathology and outcome in patients. Early prognosis can critically affect choice of patient-specific treatment protocols and increase life expectancy.