Gemcitabine Confers Comparable FOLFIRINOX Pancreatic Cancer Survival in Those with High hENT1 Levels
The drug gemcitabine is generally well tolerated, but its efficacy for pancreatic cancer when used as a solitary chemotherapy is now understood to not be as great as that of use in combination with Abraxane, or as that of the four-drug chemotherapy combination regimen – FOLFIRINOX. However these combination chemotherapy regimens for pancreatic cancer tend to carry more frequent and severe adverse effects.
Gemcitabine is a hydrophilic nucleoside analog, and consequently must be carried by cell membrane transporters to penetrate the hydrophobic cell wall in order to initiate its cytotoxic anti-pancreatic-tumor activity. The main membrane carrier for gemcitabine is considered to be hENT1 (human equilibrative nucleoside transporter 1). Past retrospective studies with gemcitabine used as adjuvant therapy for pancreatic cancer, though somewhat conflicting, generally appear to suggest that increased levels of hENT1 may be associated with a greater than expected survival advantage.
The authors, from the Università Cattolica del Sacro Cuore in Rome, Italy, conducted a retrospective analysis to better understand the relationship between hENT1 expression as an indicator of gemcitabine chemotherapy outcomes in pancreatic cancer. And more specifically, to determine the survival characteristics in pancreatic cancer of the less toxic monotherapy as compared to the generally more efficacious but more toxic combination regimen of FOLFIRINOX, especially in terms of those pancreatic cancer patients demonstrating hyperexpression of hENT1. The results of this study were published in the October 2016 issue of Clinical & Translational Oncology (the official journal of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico).
The researchers reviewed the records of 149 patients with locally advanced and metastatic pancreatic cancer who had been treated in their institution from 2009 until 2013. They found that 70 of patients’ records were evaluable (36 FOLFIRINOX and 34 gemcitabine monotherapy). The authors also found that 31 patients had stored pancreatic cancer tumor tissue with enough RNA to evaluate hENT1 levels. The median overall survival was significantly longer in duration in the FOLFIRINOX as compared to the gemcitabine group (11 months vs. 8 months). Also, the FOLFIRINOX arm showed a statistically greater incidence of Grade 3 and 4 adverse effects over gemcitabine, especially neutropenia, diarrhea, and sensory neuropathy. These outcomes were seen as consistent with what has come to be expected in the treatment of pancreatic cancer.
The authors now did something very clever. They assayed hENT1 levels from pancreatic tissue in three healthy controls. In this manner, they established a median hENT1 level which served as a border for the purposes of the remainder of the study. Results found to lie below this median level were assumed to represent low hENT1 expression, and those above this cut-off were assumed to be hyperexpression of hENT1.
Of the 31 patients who had pancreatic tissue for analysis, 12 were classified as having low hENT1 levels, and 19 patients were found to carry hENT1 hyperexpression. Seven of the hENT1 hyperexpressors had received FOLFIRINOX; twelve received gemcitabine. Seven of the hENT1 hypoexpressors had received FOLFIRINOX, five received gemcitabine.
The key finding of this study is that within the hENT1 hyperexpressors in pancreatic cancer there was no statistical difference in overall survival and in progression free survival between the FOLFIRINOX and the gemcitabine monotherapy groups. (OS FOLFIRINOX 8.5 months; gemcitabine 7 months). Further, the overall survival of the 12 hENT1 hyperexpressors treated with gemcitabine was significantly greater than the 5 hypoexpressors treated with gemcitabine (8 months vs. 2 months). The FOLFIRINOX expression subgroups with pancreatic cancer did not show any statistical difference in survival.
In the discussion section of the paper, the authors note the small sample size as a possible limitation of the study. The implications of these results certainly appear to be worth future replication attempts with a much larger study population. The identification of those pancreatic cancer patients who might do as well with less toxic gemcitabine monotherapy over combination chemotherapy could confer a considerable quality of life benefit.
Dale O’Brien, MD
In two separate studies, Boston researchers (primarily) have discovered and begun to characterize the intriguing and likely important finding that protein metabolites found in human serum are elevated typically years before the diagnosis of pancreatic cancer (adenocarcinoma) is established. The implications of this discovery have cast a new light on the early development of pancreatic cancer.
The central finding is that amino acids in the proteinogenic class known as branched-chain amino acids appear to be elevated in serum most often years prior to the diagnosis of pancreatic cancer. Specifically, these metabolic factors are the three amino acids isoleucine, leucine, and valine.
The research team for the initial study was led by scientists from the Dana-Farber Cancer Institute (and other Harvard institutions) and MIT. Also included among these scientists were those from other American universities and Canada. The results from this primary paper were published in the October 2014 issue of NATURE Medicine journal. The researchers gained access to blood samples that had been earlier collected (at least two years) from about 1,500 people in other medical tracking studies. They analyzed these samples for more than 100 metabolites and discovered 15 that were present in those people who eventually were diagnosed with pancreatic cancer as compared with matched controls (p ≤ 0.05). On further inspection, the remarkable probabilities of the noted three branched-chain amino acid elevations were determined as significant at p ≤ 0.0006 in those later diagnosed with pancreatic cancer.
The branched-chain amino acid elevations were most strongly noted during the period between two to five years before the diagnosis of pancreatic cancer, and the range was 2 to 25 years prior. The median duration between the branched-chain amino acid elevations and the diagnosis of pancreatic cancer was 8.7 years.
The causation of amino acid elevations is not fully clear. Circulation serum branched-chain amino acid elevations are seen in obese individuals and in those with insulin resistance. Subsequent tests by the researchers using mice did not appear to find a relationship between the elevated branched-chain amino acid levels and diabetes mellitus or with pancreatitis. There did again appear to be a relationship between the elevated amino acid levels and early pancreatic cancer, though this was essentially only seen in mice with Kras associated pancreatic cancer tumors. The authors note that unlike other amino acids, branched-chain amino acids are not regulated by the liver. That the levels are typically determined by dietary intake, tissue metabolism, and the breakdown of muscle and other bodily proteins. Further mice studies appeared to indicate that the early elevated branched-chain amino acid load associated with later pancreatic cancer may be related to muscle catabolism.
The second related study included some of the same aforementioned researchers, comprised primarily of Boston located scientists mainly from Harvard and MIT institutions. This work, published in the September 2016 issue of the journal Science (American Association for the Advancement of Science), built on an aspect of the earlier study to further explicate the tissue context related to tumor development by determining and comparing metabolic use patterns of branched-chain amino acids in pancreatic cancer as compared to non-small lung cancer. The authors note that both pancreatic cancer and non-small cell lung cancer often include mutations of Kras and p53, but that each of these tumors use branched-chain amino acids differently. NSCLC utilizes these amino acids as a nitrogen source (thus, increased uptake), whereas pancreatic cancer demonstrates diminished uptake of branched-chain amino acids. The differences in the levels of branched-chain amino acids between these tumors were shown to be at least in part metabolically based though enzyme variation. The different amino acid levels in these two tumor types appear to hold in mice as well as humans. The authors conclude that tissue origin appears to determine how tumors fulfil their metabolic needs.
The finding and explication related to elevations of branched-chain amino acids in serum years prior to the diagnosis of pancreatic cancer has to be judged as highly significant. It is not clear at this point the direction that this line of inquiry will assume, but one hopes that it continues and that the final destination will lead to the earlier diagnosis of pancreatic cancer.
Dale O’Brien, MD
Two interesting studies with different methods and results have been published this year by international teams as led by Australian researchers with the aim of trying to establish key sub-types of pancreatic cancer to aid in improved understanding, management and treatment of patients with adenocarcinoma of the pancreas.
The more extensive study included 149 collaborators from Australia, Britain, Europe and the U.S. under the aegis of the Australian Pancreatic Cancer Genome Initiative, who analyzed the genome of 456 separate tumors diagnosed as adenocarcinoma of the pancreas, finding common genetic mutations in 32 genes that were eventually devolved, per the authors’ analyses, into FOUR pancreatic cancer subtypes. This study was published in the March 3rd 2016 issue of the journal Nature. The four subtypes identified include 1). Squamous, related to mutations in TP53 and KDM6A which appeared to carry a poor pancreatic cancer prognosis, 2). Pancreatic Progenitor Tumors, related to genes found in the early development of pancreatic cancer (FOXA2/3, PDX1, and MNX1), Immunogenic Tumors, related to pancreatic cancer immune suppression, and ADEX Tumors, related to KRAS gene mutation activity in pancreatic cancer. The authors suggest that these sub-types of pancreatic cancer carry different natural histories and vulnerabilities that may allow for more precise and efficacious treatment congruence, as more is learned.
Thirteen international authors (including five from the above discussed study) encompassing Scottish and German researchers, again as led by Australians published an intriguing article in the August 2016 issue of Molecular & Cellular Proteomics (The American Society for Biochemistry and Molecular Biology) that postulates THREE subtypes of pancreatic cancer as identified by using a different method. The authors used mass spectrometry to explicate tyrosine phosphorylation patterns in pancreatic cancer cell lines (ductal adenocarcinoma). They found three distinct patterns that appeared related to characteristics of signaling networks. The authors acknowledge this as a novel taxonomy for pancreatic cancer, but indicate that these results provide insight about how increased understanding of the biology of pancreatic ductal adenocarcinoma may offer distinctions for more tailored treatment and care of pancreatic cancer.
These are two highly interesting research studies that do credit to our colleagues from down under. Further practical extension of this work would seem in order. The age of precision medicine for pancreatic cancer is dawning.
Dale O’Brien, MD