**1. Introduction**

Cholangiocellular carcinomas (CCCs) are rare but aggressive tumors that display features of biliary di fferentiation. CCCs comprise approximately 3% of gastrointestinal tumors and have an overall incidence of less than two per 100,000 [1]. According to their anatomical location, CCCs are commonly classified as intrahepatic and extrahepatic tumors, the latter entity being further subdivided into perihilar and distal tumors. In Germany, mortality from intrahepatic CCCs more than tripled between 1998 and 2008 [2]. In line with the increased mortality data, the number of reported cases of intrahepatic CCC also increased between 1970 and 2006 [3]. CCC is one of the most fatal cancers: the median overall survival is 20–28 months and five-year survival rates are about 25% [4,5]. In the United States, the number of CCC-related deaths per annum increased dramatically in the past two

decades, surpassing 7000 by 2013 [6]. CCC mortality for those aged 25+ increased 36% between 1999 and 2014 [6]. The only curative option for CCC is surgical resection, but most patients develop a recurrence after resection.

Increased glucose metabolism and uptake is a hallmark of aggressive cancer. Accordingly, elevated 18Fluorodesoxyglucose (FDG) uptake was present in 92% of CCCs [7]. This phenomenon, also called the Warburg effect, indicates that cancer cells generate energy predominantly via glycolysis even if sufficient oxygen is present. At first, it seems paradoxical that tumors use 'inefficient' glycolysis instead of OXPHOS for energy production. However, there are several explanations for why ATP generation is reprogrammed in cancer cells. First, aerobic glycolysis is not as inefficient as is often reported. Although it is correct that the amount of ATP generated per molecule of glucose is low, the rate of glucose metabolism is high in cancer cells. The production of lactate from glucose occurs 10–100 times faster than the complete oxidation in mitochondria, and the level of ATP production is similar [8]. Thus, rewiring energy metabolism toward glycolysis causes increased generation of lactate. Several studies suggested that lactate facilitates metastasis via production of a microenvironment toxic to normal cells and stimulation of tissue lysis [9,10]. Lactate dehydrogenase A (LDHA) catalyzes the inter-conversion of pyruvate and L-lactate with concomitant inter-conversion of nicotinamide adenine dinucleotide (NADH) and NAD<sup>+</sup> [11]. Secondly, the Warburg effect has been proposed to be an adaptive mechanism to support the biosynthetic requirements of uncontrolled proliferation. Glucose metabolites serve as carbon sources for anabolic processes. The excess carbon is diverted into branching pathways emanating from glycolysis and is used for the generation of cellular building blocks such as nucleotides, lipids, and proteins [12–15]. Another theory proposes that tumors shut down OXPHOS to reduce the reactive oxygen species (ROS) burden on their own biomolecules, while maintaining a level necessary for cell signaling [14]. The respiratory chain complexes are the main production sites of ROS, although their individual contributions to this process likely differ, with complex I being the most important in this vicious circle [16].

Immunohistochemical (IHC) staining of OXPHOS complexes of homogeneous tissue samples correlates well with enzymatic analysis, as the OXPHOS system is mainly regulated via the protein amount [17–19]. In the present study, we used IHC for technical reasons, foremost because it is nearly impossible to obtain sufficient amounts of frozen CCC tissue for functional evaluation of OXPHOS enzymes. In addition, substantial cellular heterogeneity is present within single tumors. Furthermore, a tumor cell content of over 80% is needed to generate reliable functional data on OXPHOS enzyme activity. IHC staining of heterogeneous samples is the method of choice because it reliably reflects the in vivo situation at the cellular level [18,19].

Therefore, the aim of the present study was to characterize the OXPHOS phenotype of CCC by IHC using tissue microarrays.

#### **2. Materials and Methods**
