A comprehensive and extensive IHC analysis with mAbs of a large number of FFPE normal tissues has shown that, contrary to what is generally assumed, HLA class I subunits are not expressed by nucleated cells in all the normal tissues [23,24,25]. Many of them are barely or not stained by mAbs reacting with HLA-A heavy chain, HLA-B heavy chain, HLA-C heavy chain, and β2-M.

Thanks to the availability of innovative and specific antibodies, we propose that healthy human tissues be classified into four different groups, based on the mAb HC-A2, mAb HC-10, and β2-M mAb staining intensity and cellular localization: (i) tissues with moderate/strong staining patterns by HLA class I subunit-specific mAbs, (ii) tissues showing a barely detectable staining pattern by HLA class I subunit-specific mAbs, (iii) tissues showing a differential staining pattern by HLA class I subunit-specific mAbs, (iv) and tissues with no detectable staining by HLA class I subunit-specific mAbs.

Our results parallel those obtained by testing normal frozen tissues [14,15,16] with mAb PA 2.6 and mAb 06/64 for HLA ABC highlighting some differences. Differences include the use of frozen vs. FFPE tissues [14,15,16], antibodies with diverse sensitivity, and methodological variability in the employed assays. In particular, previous studies used a manual peroxidase-antiperoxidase (PAP) method with the monoclonal antibody PA2.6, which reacts similarly to w6/32 that recognizes an epitope on the HLA-A,B,C heavy chain/β2 micro-globulin complex in the membrane of cells [14, 15]. While in the present study, we employed a more accurate and sensitive automated IHC procedure, and we used monoclonal antibodies with higher specificity that although they are unable to detect the HLA-A,B,C heavy chain/β2 micro-globulin complex, they allowed us to discriminate the subcellular localization of the single sub-units in FFPE tissues [26, 27].

Another point of strength of our study was the use of TMA, which is a rapid and high-throughput technique to assay numerous tissues arrayed on a single slide [28]. TMA allows a reliable semiquantitative scoring of the intensity of the staining because all tissue samples on a TMA slide are exposed to the same amount of primary and secondary antibody and chromogen. Simultaneous analysis of a large number of specimens decreases time and cost since only a small amount of each reagent is needed to assay all cores at the same time. A potential caveat is a more comprehensive analysis of tissues in the presence of tissue heterogeneity, particularly for small cores. Nevertheless, by using normal and not tumoral tissues, this could represent a partial boundary; however, we re-tested on the whole section of every sample that was negative for the three staining or not evaluable on TMAs.

The reported data clearly showed that the tissue distribution of HLA A, B, C heavy chains and β2-M is not uniform all over the tested tissues. In this context, our study is relevant and timely by providing the most extensive and comprehensive evaluation of HLA class I subunit expression (A, B, C heavy chains and β2-M) in different human healthy tissues. This represents a backbone analysis for future functional studies in cancer tissues with the aim to elucidate the baseline staining expression of HLA A, B, C heavy chains and β2-M in healthy tissues. The expression of the gene products of HLA class I loci is controlled by different regulatory mechanisms [17, 18]. One of the most important pathways is interferon-γ (IFN-γ)/Janus kinase (JAK)/signal transducer and activator of transcription (STAT1). The IFN-γ/JAK/STAT1 pathway plays a crucial role in the antigen processing pathway and the subsequent dynamic change of downstream signals, including the HLA class I subunit [17,18,19,20,21,22,23,24,25,26,27,28,29]. The differential basal expression of HLA A, B, C heavy chain and β2-microglobulin in normal tissues could be justified by the parallel differential basal expression of the transcription factor STAT1. In this paper, we show that there is an interesting association between the expression of HLA A, B, C heavy chains and β2-M and the expression of STAT1.

Although these results strongly support our thesis that HLA class I subunits are not ubiquitously expressed, on the other hand, we expected to find differences in the levels of phosphorylated STAT1 as well. Instead, all healthy tissues tested were negative. This unexpected result may be due to the basal activity of STAT1. Activated in response to many different cytokines and growth factors by phosphorylation of specific tyrosine residues, STAT1 enhances its transcription factor effect. This strong activation has been shown to be present in many pathological conditions such as injury, ischemia, and tumors [30,31,32]. We could assume that in healthy tissues, this activation does not occur and that the non-phosphorylated form STAT1 handles the basal expression of proteins such as HLA class I components [33].

Expression of HLA class I subunits in healthy human tissues should be considered when evaluating the increase, reduction, or loss of HLA I expression on malignant cells reported in many types of cancer [32,33,34,35]. The amount of these molecules expressed at the cell surface varies significantly depending on the level of gene transcription, transduction, and epigenetic regulation (8). In immunohistochemical studies, however, the boundary between complete, irreversible loss, and down-regulation (or low expression) of HLA in tumor tissues may be unclear if the baseline level of expression in the respective normal tissue is not taken into consideration.

In conclusion, our data elucidated the selective distribution and patterns of expression of HLA A, B, C heavy chains and β2-microglobulin in healthy human tissues. We showed that the expression of the three main subunits of the HLA class I is tissue-specific and can vary within the same districts, emphasizing again the caution that is needed to explain the changing levels of expression of these antigens in respective tumor tissues. Thus, our data could add new insight for the interpretation of immunohistochemical studies in different cancer types, and it may improve the understanding of the mechanisms of escape adopted by malignant cells during tumor development and progression.