Sir,

We have read with particular attention the recent article published in the BJC by Alvarez et al (2013).

They, through endoscopic ultrasound elastography system and conventional immuno-histochemical assays, perform an intrinsic characterisation of tumour stroma’s stiffness, fibroblast density and architecture focusing in type I collagen fibre arrangement, cancer-associated fibroblasts (CAFs) and activated CAF quantification. The analysis of post-treatment tissues from patients treated with nab-paclitaxel and gemcitabine combination showed a decrease in CAF number and changes in collagen architecture.

These results reaffirm the potential ability of nab-paclitaxel to target the stroma and to change its phenotype, altering the ‘hard of the rock’, as described by Garber (2010), intensifying its permeability to deliver cytotoxic agents such as gemcitabine by increasing tumour vascularisation or enzymatic inhibition (Frese et al, 2012) that would suppose a better pharmacokinetic profile (Von Hoff et al, 2011). Even though it is known the clinical benefit of nab-paclitaxel plus gemcitabine in metastatic pancreatic cancer, the cellular or molecular mechanisms that are expressly addressing this stromal involution have not been appointed yet (Alvarez et al, 2013). In this sense, authors analysed the role of secreted protein, acidic and rich in cysteine (SPARC) although they could not find the correlation between the degree of SPARC expression and clinical or pathological responses due to its high affinity to albumin protein of nab-paclitaxel. SPARC strengthens the accumulation of nab-paclitaxel mediated by albumin in the ‘tumour’s Achilles’ heel’ (Von Hoff, Annual meeting of the American Society of Clinical Oncology, 2009); however, the mere use of this endogenous transport system would not justify the clinical impact of this combined treatment. From our point of view the key may be in what Omary et al (2007) described as ‘a star on the rise’ in pancreatic disease: the pancreatic stellate cells (PaSCs; α-SMA+ Nestin++ Vimentin++) because of their role as the main manufacturers of profibrotic extracellular matrix (ECM) components of the pancreatic tumour stroma, and which should match with the activated CAFs (SMA+ Vimentin+ fibroblasts) in Alvareźs paper.

In pancreatic cancer, PaSCs show increased proliferation and migration properties, and so they could be a suitable target for nab-paclitaxel because of their ability to interfere with the mitotic activity (Gradishar, 2006). Alvarez et al (2013) demonstrate that although CAFs number decreases in patients treated with nab-paclitaxel plus gemcitabine in neoadjuvant setting, the average of activated CAFs remains unchanged.

We would like to explain our hypothesis related to what is happening in the stroma: it could be a transient blockade of activated CAFs metabolism, a quiescent status forced for a pharmacologically active substance. Bachem et al (2005) show that cancer cells (CCs) induce a desmoplastic reaction in pancreatic adenocarcinoma by stimulating PaSCs in a paracrine way. So it could be an indirect elimination of principal fibrogenic mediators that stimulate proliferation (platelet-derived growth factor) and ECM synthesis (fibroblast growth factor -2 and transforming growth factor -ß1) of activated PaSCs through the abrogation of CCs. In this sense, the hypothetical presence of surface cellular receptors for nab-paclitaxel in CCs could be an interesting pathway biomarker for the effectiveness of the drug as authors related. On the other hand, the ablation of physiological PaSCs functioning after nab-paclitaxel inclusion, in turn, would break the two-way communication between PaSCs and CCs (Apte et al, 2013). In this context, under normal conditions CCs recruit new PaSCs to their vicinity (Vonlaufen et al, 2008), whereas nab-paclitaxel would temporarily inhibit the main mechanism that rules the desmoplastic reaction.

All these molecular mechanisms would allow to maintain and stabilise the activated CAF's number despite them being dysfunctional, and also would decline its CAF progeny. In this sense, Apte et al (2013) and Bachem et al (2005) propose that activated PaSCs can transform into a myofibroblast-like phenotype sub-population with the ability of secreting excess amounts of ECM.

In conclusion, due to the dynamic nature of the stromal compartment, it is critically involved in the development and progresion of pancreatic tumours (Heinemann et al, 2013). Before using neadjuvant treatment it may be important to know the stroma’s cellular activation grade with regard to PaSCs plus their number. In this sense, PaSCs or activated CAFs could give us ample information about tumoral potential of the stroma itself and so could have an important contribution for the patient's prognosis. Also, the fact that activated CAFs do not decrease after nab-paclitaxel treatment could not mean that these cells are not one of the main actors of stromal disruption but the primary target of nab-paclitaxel.