Flow cytometry has traditionally been used to identify hemato-lymphoid neoplasms. However, the flow cytometry laboratories that deal with tissues would often receive samples that have an epithelial neoplasm. In our laboratory, we use flow cytometry to identify cells with epithelial differentiation using Ber-EP4 antibody that targets CD326 (EpCAM). We have formulated an algorithm-based approach for the application of this marker. This approach has been elaborated in this article.

by Dr Pranav Dorwal and Dr Helen Moore

Introduction
The use of flow cytometry in the laboratory has traditionally been applied for diagnosing lymphomas and leukemias. The biggest advantage that flow cytometry has over histopathology is a much quicker turn-around-time, as most of the samples are fresh and can be processed right away, unlike a histopathology sample which needs to undergo fixation and processing before it is ready to be examined. Any additional testing on flow cytometry samples can be performed instantly, whereas the same usually requires another day in the histopathology lab. Many of the lymph node malignancies (primary lymphomas versus metastatic involvement) can appear undifferentiated. In these cases, the histopathologist needs the help of a plethora of immunohistochemical markers to reach a diagnosis. The ability to identify samples where non-hematological malignancies are present can be helpful for the treating physician as well as the reporting histopathologist, who can then test with a more dedicated panel. The ultimate aim of this testing is to get an early diagnosis so that patient’s treatment is not delayed.

A large number of markers have been used to identify epithelial differentiation in tumours by immunohistochemistry (IHC), including cytokeratin (CK), carcinoembryonic antigen (CEA), cancer antigen 125 (CA-125) as well as epitopes recognized by the antibodies LeuM1 (anti-CD15 antibody), and MOC-31 and Ber-EP4 antibodies, both of which recognize epitopes on EpCAM (the epithelial cell adhesion molecule). However, most of these are not available for use by flow cytometry. EpCAM (also known as CD326) was first discovered in 1979 and at that time thought to be specific for colonic carcinoma [1]. Ber-EP4 is, therefore, an anti-CD326 antibody which binds to a cell membrane glycoprotein on human epithelia. There is a comprehensive list of tumours that are Ber-EP4 positive, as described by Went et al. and Spizzo et al. [2, 3]. The traditional use of Ber-EP4 in histopathology has been limited essentially for differentiation between adenocarcinoma and malignant mesothelioma [4]. This could be due to the fact that other epithelial markers (such as CK) are expressed more often than the CD326 (EpCAM) in epithelial malignancies and thus are more helpful in lineage determination.

We use an algorithmic approach to decide the flow cytometry panel to be applied (Fig. 1). When the clinical details or radiological findings are indicative of a non-hematopoietic malignancy, we apply the CD326 panel. This panel is composed of CD326, CD56 and CD45. CD56 was included in the panel to identify myeloma cells (which may be present in the CD45-negative region) and cells with neuroendocrine differentiation. If, on analysis, there is no CD45-negative population and the sample is composed of predominantly lymphoid cells, a lymphoid screening panel is then used. Samples that are received with diagnosis of suspected lymphoma are initially processed with a routine lymphoid screening panel. In these cases, Ber-EP4 antibody is tested only if large numbers of CD45-negative events are identified.

Method for Ber-EP4 testing
The tissue and fine-needle aspirate (FNA) samples are received fresh in RPMI medium. The tissues are placed on the metal sieve and ground using a glass pestle to form a cell suspension using 2 % PBS-FCS. This suspension is subsequently filtered, which is then washed and lysed. The cell count is ascertained by the cell counter only in cases of larger tissues, where we may have to dilute the sample to adjust the cell count to approximately 10×10⁹/L. FNA and core biopsies are usually paucicellular and do not need a cell count.

The sample is stained with 5 µl of CD45-PC5 [Immunotech SAS (Beckman Coulter)], 20 µl of CD56-PE (Immunotech SAS) and 10 µl of monoclonal mouse anti-human epithelial antigen-FITC conjugated antibody (Clone: Ber-EP4) (Dako Denmark A/S). The sample is then incubated at 4 °C for 30 minutes, followed by a washing step and is ready to be run on the flow cytometer (Beckman Coulter Life Sciences). A total of 10 000 events are acquired with the time threshold set at 300 seconds for the acquisition.

Flow cytometric analysis
The flow cytometric analysis is performed using Navios and Kaluza softwares (Beckman Coulter Life Sciences). The various populations of interest are gated with the focus on identifying the expression of CD326 (with or without CD56) in the CD45-negative population.

Discussion
In our experience of testing for CD326 by flow cytometry, we have been able to comment on the presence or absence of CD326 expression in CD45-negative populations (Figs 2(a, b) and 3). The various carcinomas where we have identified CD326 positivity are: adenocarcinoma, small cell carcinoma, Merkel cell carcinoma, renal cell carcinoma, squamous cell carcinoma, prostate carcinoma, germ cell tumour of testis, and myxoma. We have observed that the expression of CD326 in melanomas can be variable, but they more frequently express CD56. The co-expression of CD326 and CD56 usually indicates a neuroendocrine tumour. Our concordance rate with histopathology using CD326 testing was found to be 97.6 %, which we have published previously [5].

CD326 expression has also been reported to be a prognostic marker with poor outcomes in epithelial ovarian and gall bladder carcinomas [6, 7]. Another important role of this testing could be application in decision making for use of monoclonal antibodies for targeted therapy. The first EpCAM targeting antibody, Catumaxomab (trade name Removab, Fresenius Biotech GmbH) received European market approval in EpCAM-positive carcinomas for the treatment of malignant ascites. Another modification that could be useful in diagnosing epithelial malignancies is to apply Ki67 testing using flow cytometry. This could be done in the same tube as CD326, and thus more information could be obtained with the same amount of sample [8].

There has been considerable data describing the use of the Ber-EP4 antibody in malignant effusions [9–11]. The literature mentions that the presence of epithelial cells in the body fluid should raise the suspicion of metastatic epithelial malignancy, as the reactive body fluids may be composed of lymphocytes and reactive mesothelial cells in varying proportions. There have been multiple studies in the past where flow cytometric CD326 testing has been applied for identifying epithelial cells in body fluid effusions. We have found that our results have a very good concordance with histopathology results. This is in keeping with the findings of Davidson et al., although their study looked at the detection of malignant cells in effusions [12].

The disadvantage of using Ber-EP4 for identifying epithelial differentiation is that there are many epithelial malignancies that do not express CD326 (EpCAM). As mentioned earlier, the use of a broader antibody like cytokeratin (pan-CK) may solve this problem. But unfortunately, such an antibody is not currently available for clinical use by flow cytometry, to the best of our knowledge. Meanwhile, Ber-EP4 should give us the answer in most of the cases. Another disadvantage is that CD326 will be negative in cases of neoplasms of mesenchymal origin, such as sarcomas.
Most flow cytometry laboratories across the world will liaise with histopathology departments for the diagnosis of non-Hodgkin lymphomas. The use of Ber-EP4-testing flow cytometry may play an important role even in epithelial malignancies. The antibody used by us is a CE-marked antibody for in vitro diagnostics and, thus, requires a limited verification process. We followed the method recommended by the manufacturer. The rapid turn-around-time of flow cytometry results makes it a useful screening tool. Our experience shows that flow cytometric testing for CD326 (EpCAM) can be a useful method for diagnosing non-lymphoid malignancies that are poorly differentiated. We suggest that this method would be more useful if the protocol for its application is set up in consultation with the histopathology department, along with setting up a channel of bilateral communication. The histopathologist, based on the flow cytometry information provided, can then set up a more directed immunohistochemical panel. We would like to emphasize at this stage that the aim of the flow cytometric CD326 testing is not to formally diagnose carcinomas, but to highlight the presence of epithelial cells which may lead to the diagnosis of carcinoma. Final classification obviously remains the role of the histopathologist.

References
1. Patriarca C, Macchi RM, Marschner AK, Mellstedt H. Epithelial cell adhesion molecule expression (CD326) in cancer: a short review. Cancer Treat Rev 2012; 38(1): 68–75.
2. Went PT, Lugli A, Meier S, Bundi M, Mirlacher M, Sauter G, Dirnhofer S. Frequent EpCam protein expression in human carcinomas. Hum Pathol 2004; 35(1): 122–128.
3. Spizzo G, Fong D, Wurm M, Ensinger C, Obrist P, Hofer C, Mazzoleni G, Gastl G, Went P. EpCAM expression in primary tumour tissues and metastases: an immunohistochemical analysis. J Clin Pathol 2011; 64(5): 415–420.
4. Sheibani K, Shin SS, Kezirian J, Weiss LM. Ber-EP4 antibody as a discriminant in the differential diagnosis of malignant mesothelioma versus adenocarcinoma. Am J Surg Pathol 1991; 15(8): 779–784.
5. Dorwal P, Moore H, Stewart P, Harrison B, Monaghan J. CD326 (EpCAM) testing by flow cytometric BerEP4 antibody is a useful and rapid adjunct to histopathology. Cytometry B Clin Cytom 2017; doi: 10.1002/cyto.b.21543.
6. Spizzo G, Went P, Dirnhofer S, Obrist P, Moch H, Baeuerle PA, Mueller-Holzner E, Marth C, Gastl G, Zeimet AG. Overexpression of epithelial cell adhesion molecule (Ep-CAM) is an independent prognostic marker for reduced survival of patients with epithelial ovarian cancer. Gynecol Oncol 2006; 103(2): 483–488.
7. Varga M, Obrist P, Schneeberger S, Mühlmann G, Felgel-Farnholz C, Fong D, Zitt M, Brunhuber T, Schäfer G, et al. Overexpression of epithelial cell adhesion molecule antigen in gallbladder carcinoma is an independent marker for poor survival. Clin Cancer Res 2004; 10(9): 3131–3136.
8. Sikora J, Dworacki G, Zeromski J. DNA ploidy, S-phase, and Ki-67 antigen expression in the evaluation of cell content of pleural effusions. Lung 1996; 174: 303-313.
9. Pillai V, Cibas ES, Dorfman DM. A simplified flow cytometric immunophenotyping procedure for the diagnosis of effusions caused by epithelial malignancies. A J Clin Pathol 2013; 139(5): 672–681.
10. Krishan A, Ganjei‐Azar P, Hamelik R, Sharma D, Reis I, Nadji M. Flow immunocytochemistry of marker expression in cells from body cavity fluids. Cytometry A 2010; 77(2): 132–143.
11. Risberg B, Davidson B, Dong HP, Nesland JM, Berner A. Flow cytometric immunophenotyping of serous effusions and peritoneal washings: comparison with immunocytochemistry and morphological findings. J Clin Pathol 2000; 53(7): 513–517.
12. Davidson B, Dong HP, Berner A, Christensen J, Nielsen S, Johansen P, Bryne M, Asschenfeldt P, Risberg B. Detection of malignant epithelial cells in effusions using flow cytometric immunophenotyping. Am J Clin Pathol 2002; 118(1): 85–92.

The authors
Pranav Dorwal* MBBS, DCP, DNB; Helen Moore MBChB, FRACP, FRCPA
Waikato Hospital, Pembroke St, Hamilton 3204, New Zealand

*Corresponding author
E-mail: Pranav.dorwal@waikatodhb.health.nz

Current measures for diagnosis and therapy of chronic kidney disease are limited. Better biomarkers are required to improve treatment by directing therapeutic intervention, tracking responses to therapy and providing greater understanding of the underlying mechanisms driving renal disease progression. We describe here the development of microRNAs as biomarkers for diabetic kidney disease, the most common etiology leading to chronic kidney disease and end-stage renal failure.

by Dr Tanya A. Smith, Dr Kate Simpson, Prof Donald J. Fraser and Dr Timothy Bowen

Diabetes, complications and biomarkers
Diabetes is a major global health challenge, with 23.1 million cases diagnosed in the US alone [1]. As described below, our laboratory is currently developing urinary microRNAs as biomarkers for diabetic kidney disease. These transcripts may also have utility as biomarkers for other complications of type 2 diabetes mellitus including diabetic retinopathy, neuropathy, cardiovascular disease, stroke, ulceration and amputation [2].

Diabetic kidney disease
Diabetic kidney disease (DKD) is the leading cause of end-stage renal disease in the United States. Clinical presentation is characterized by proteinuria, hypertension, and progressive reduction in kidney function. DKD is a progressive condition associated with around 35% of patients with type 1 and type 2 diabetes mellitus [3]. A highly significant public health concern, DKD is currently managed by targeting cardiovascular risk reduction, blood pressure management, glycemic control (hemoglobin A1c concentration), nutritional counselling, weight loss, smoking cessation, and pharmacological inhibition of the renin–angiotensin system using angiotensin-converting enzyme inhibitors or angiotensin-2 receptor blockers [4].

Despite the stabilization of the incidence of diabetes over the past 15 years, the United States Renal Data System has demonstrated increased prevalence of end-stage renal disease attributed to diabetes. However, the disease burden is such that patients often do not survive to end-stage renal disease. There is a broad spectrum of cardiovascular complications associated with DKD of which the underlying etiology remains unclear. Cardiovascular disease is the leading cause of death in this patient group, manifesting as cerebral vascular event, sudden cardiac death, myocardial infarction and diabetic cardiomyopathy. It is, therefore, essential to identify and treat patients before irreversible organ damage to reduce the medical and economic burden of disease [4].

Existing DKD biomarkers
DKD is associated with both glomerular hyperfiltration leading to progressive albuminuria, and declining glomerular filtration rate.

Albuminuria
Proteinuria is a biomarker used widely as a proxy to assess the integrity of the glomerular filtration barrier (for detailed glomerular and nephronal physiology see [5]). Quantification of urinary albuminuria excretion is a non-invasive and inexpensive method to monitor disease. Microalbuminuria is currently the primary predictive clinical DKD marker and occurs when urinary albuminuria excretion rate reaches 30–300 mg/24 h, macroalbuminuria is reached when this rate exceeds 300 mg/24 h. In the presence of diabetes mellitus, confirmation of microalbuminuria in two separate samples taken 3–6 months apart is diagnostic of DKD. Screening for albuminuria is more commonly performed using urinary albumin-to-creatinine ratio on an isolated urine sample, and is defined as >30 mg/g.

However, albuminuria is a non-specific biomarker measurable only after kidney injury has occurred and correlates poorly with clinical disease. In addition, albuminuria may be a transient DKD feature, or may occur only when widespread glomerular damage is already present [6, 7]. Recent reports have noted that up to 25% of patients with type 2 diabetes mellitus and diminished kidney function have little or no proteinuria, despite having biopsy-proven DKD [8]. There is, therefore, a need to find sensitive and specific biomarkers to predict DKD susceptibility and progression.

Estimated glomerular filtration rate
The Kidney Disease: Improving Global Outcomes (KDIGO) [4] classification is directed at adults and children over the age of 2 years old with evidence of kidney disease. Glomerular filtration rate (GFR) is considered the best measure of kidney function. Normal GFR is quantified as 100–150 ml/min and can be determined by creatinine clearance or an estimated GFR (eGFR) calculation basis on serum creatinine, age, sex and ethnicity (Table 1).

Histological features of renal biopsies, eGFR and DKD
Histological features (see [5]) correlate with functional alterations in DKD. The Renal Pathological Society system, based on glomerular changes observed in the development of DKD, groups both type 1 and type 2 diabetes mellitus patients into the four classes described below [9].

Class I: Glomerular basement membrane thickening: isolated glomerular basement membrane thickening and only mild, non-specific changes by light microscopy that do not meet the criteria of classes II–IV.
Class II: Mesangial expansion, mild (class IIa) or severe (class IIb). Glomeruli classified as mild or severe mesangial expansion but without nodular sclerosis (Kimmelstiel–Wilson lesions) or global glomerulosclerosis in >50% of glomeruli.
Class III: Nodular sclerosis (Kimmelstiel–Wilson lesions): at least one glomerulus with nodular increase in mesangial matrix (Kimmelstiel–Wilson) without changes described in class IV.
Class IV: Advanced diabetic glomerulosclerosis. Over 50% global glomerulosclerosis with other clinical or pathologic evidence showing that sclerosis is attributable to DKD.

The need for newer biomarkers
Current biomarkers do not relate well to the above pathological classification. Many potential novel biomarkers have been tested in an attempt to improve our ability to discern underlying renal pathology non-invasively, with the aim of guiding therapy. These include urinary transferrin, serum osteopontin, urinary retinol-binding protein (RBP), serum interleukin-18, serum cystatin C, serum resistin, serum TNF-α, serum interleukin-6 and urinary neutrophil gelatinase-associated lipocalin (NGAL) [reviewed in 6]. In patients with albuminuria these markers increase significantly, but their relationships with histopathological changes, eGFR, HBA1C and blood pressure is complex.

Detection and identification of microRNAs in body fluids as kidney disease biomarkers
Members of the short single-stranded endogenous RNA transcript family known as microRNAs (miRNAs) modulate the expression of most mammalian protein coding genes, thereby influencing developmental and metabolic processes, and disease phenotypes [10]. Disease-associated changes in miRNA expression profiles have been observed in cancer, cardiovascular disease, diabetes and chronic kidney disease that is treated by dialysis or transplantation [reviewed in 11–14].

To date, the majority of miRNA biomarker analyses have focused on detection of circulating transcripts [11, 13]. By contrast, the adoption into existing treatment pathways of a miRNA biomarker test on biofluid samples that can be obtained without venipuncture promises attractive reductions in time and cost [15].

We have developed RT-qPCR-based methods for precise quantification of miRNAs in urine, peritoneal dialysis effluent and renal transplantation perfusate [15–19]. The robust recovery of miRNAs from these complex analytical matrices highlights their potential utility both as non-invasive biomarkers of occurrence and/or progression of kidney disease, and as potential targets for therapeutic intervention. We have shown association of increased miR-21 with peritoneal fibrosis [17] and transplantation outcomes [18, 19]. Analysis of the renal transplantation perfusate with which the organ is supplied between donor and recipient also identified elevated miR-21 [18].

Utility of urinary miR-29b, miR-126 and miR-155 to test for DKD
Disease biomarkers are useful only when they can inform our potential to change patient treatment. The US Food and Drug Administration recommends that a reduction in eGFR of 40% over 2–3 years is a broadly acceptable effective surrogate for confirmation of CKD [20]. However, since eGFR decline is typically very gradual over the first decade or so of disease and more rapid thereafter, a biomarker that can differentiate between later stages of CKD maybe more cost-effective in detecting quantifiable responses to therapy in clinical trials [20, 21].

We have recently shown association of elevated urinary miR-29b, miR-126 and miR-155 detection predominantly in patients with type 2 diabetes mellitus and DKD [15]. We observed upregulation of these three miRNAs in two disease cohorts, obtaining an area under the curve of 0.8 in combined receiver operating characteristic curve analysis [15]. Our markers are clustered in late-stage disease (Fig. 1) and at an 80% relative quantification threshold for each miRNA, identified 48% of DKD patients with a 3.6% false positive detection rate [15]. We are currently investigating the significance of this apparent DKD patient stratification.

Utility of urinary miR-29b, miR-126 and miR-155 to investigate DKD mechanisms
We detected increased miR-29b and miR-126 in conditioned medium from cultured glomerular endothelial cells exposed to disease-related cytokines transforming growth factor-β1 and tumour necrosis factor-α, respectively [15]. It is thus conceivable that miRNAs may travel down the nephron [5] to mediate disease-related and functional effects [22]. Our data also included evidence for decreased urinary miR-192 in DKD [15], supporting our previous finding showing downregulated miR-192 expression in renal biopsies from DKD patients [23].

Conclusion
DKD is one of the most important global health challenges. Existing biomarkers provide a non-invasive approach to diagnosis and, in late-stage disease, identify the extent of kidney damage. However, there is a lack of non-invasive measures of active disease processes. New biomarkers are, therefore, required to measure risk of progressive kidney damage and to measure responses to treatment in the individual. Successful development of such biomarkers would help to individualize treatment using existing approaches, and would greatly accelerate testing of new treatments. MicroRNAs tested in urine show promise in this area.

Acknowledgments
Supported by the National Institute for Health Research Invention for Innovation (i4i) Programme grant II-LA-0712-20003 and Kidney Research UK Project grant award RP44/2014. The Wales Kidney Research Unit is funded by core support from Health and Care Research Wales.

Disclosure
TB and DF are inventors for patent WO/2017/129977 Chronic Kidney Disease Diagnostic.

References
1. National diabetes statistics report, 2017: estimates of diabetes and its burden in the United States. Centers for Disease Control and Prevention (CDC) 2017 (https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf).
2. Wang J, Chen J, Sen S. MicroRNAs as biomarkers and diagnostics. J Cell Physiol 2016; 231(1): 25–30.
3. de Boer IH, et al. Temporal trends in the prevalence of diabetic kidney disease in the United States. JAMA 2011; 305(24): 2532–2539.
4. Levin A, et al. Kidney disease: improving global outcomes (KDIGO) CKD work group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Supplements 2013; 3(1): 1–150.
5. Pollak MR, et al. The glomerulus: the sphere of influence. Clin J Am Soc Nephrol 2017; 9(8): 1461–1469.
6. Al-Rubeaan K, et al. Assessment of the diagnostic value of different biomarkers in relation to various stages of diabetic nephropathy in type 2 diabetic patients. Sci Rep 2017; 7(1): 2684.
7. Alicic RZ, et al. Diabetic kidney disease: challenges, progress, and possibilities. Clin J Am Soc Nephrol 2017; 12(12): 2032–2045.
8. Dwyer JP, Lewis JB. Nonproteinuric diabetic nephropathy: when diabetics don’t read the textbook. Med Clin North Am 2013; 97(1): 53–58.
9. Tervaert TW, et al. Pathologic classification of diabetic nephropathy. J Am Soc Nephrol 2010; 21(4): 556–563.
10. Bartel DP. Metazoan microRNAs. Cell 2018; 173(1): 20–51.
11. Simpson K, et al. MicroRNAs in diabetic nephropathy: from biomarkers to therapy. Curr Diab Rep 2016; 16(3): 35.
12. Rupaimoole R, Slack FJ. MicroRNA therapeutics: towards a new era for the management of cancer and other diseases. Nat Rev Drug Discov 2016; 16(3): 203–222.
13. Wonnacott A, et al. MicroRNAs as biomarkers in chronic kidney disease. Curr Opin in Nephrol and Hypertens 2017; 26(6): 460–466.
14. Zhao H, et al. MicroRNAs in chronic kidney disease. Clin Chim Acta 2019; 491(4): 59–65.
15. Beltrami C, et al. Association of elevated urinary miR-126, miR-155 and miR-29b with diabetic kidney disease. Am J Pathol 2018; 188(9): 1982–1992.
16. Beltrami C, et al. Stabilization of urinary microRNAs by association with exosomes and argonaute 2 protein. Noncoding RNA 2015; 1(2): 151–165.
17. Lopez Anton M, et al. MicroRNA-21 promotes fibrogenesis in peritoneal dialysis. Am J Pathol 2017; 187(7): 1537–1550.
18. Khalid U, et al. MicroRNA-21 (miR-21) expression in hypothermic machine perfusate may be predictive of early outcomes in kidney transplantation. Clinical Transplant 2016; 30(2): 99–104.
19. Khalid U, et al. A urinary microRNA panel that is an early predictive biomarker of delayed graft function following kidney transplantation. Sci Rep 2019; 9: 3584.
20. Levey AS, et al. GFR decline as an end point for clinical trials in CKD: a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration. Am J Kidney Dis 2014; 64(6): 821–835.
21. Stevens LA, et al. Surrogate end points for clinical trials of kidney disease progression. Clin J Am Soc Nephrol 2006; 1(12): 874–884.
22. Thomas MJ, et al. Biogenesis, stabilization and transport of microRNAs in kidney Health and Disease. Noncoding RNA 2018; 4(4): E30.
23. Krupa A, et al. Loss of microRNA-192 promotes fibrogenesis in diabetic nephropathy. J Am Soc Nephrol 2010; 21(3): 438–447.

The authors
Tanya A. Smith MB ChB; Kate Simpson PhD; Donald J. Fraser MB ChB, PhD; Timothy Bowen* PhD
Wales Kidney Research Unit, Cardiff University School of Medicine, Cardiff, CF14 4XN, UK
*Corresponding author
E-mail: bowent@cardiff.ac.uk