In the mid-1980s, Brenner and coworkers [1,2] hypothesized that a glomerular deficit, either inherited or acquired, may increase the subsequent risk of developing hypertension. At about the same time, Barker and coworkers proposed the fetal origins hypothesis of adult disease, now known as the developmental origins of health and disease hypothesis, which originally postulated that suboptimal events during prenatal development increase the risk of cardiovascular disease in adulthood. The range of adult chronic diseases linked to fetal and early postnatal origins has now expanded to include cardiovascular disease [3,4], renal disease [5,6], obesity [7,8], and diabetes [9,10].
Similarly, the concept of podocyte depletion was conceived as a result of seminal contributions from Kriz [11–13], Wiggins [14–18], and others [19–24]. In short, the podocyte depletion hypothesis proposes that ‘absolute’ podocyte depletion (loss of podocytes resulting in a decrease in the total number of podocytes in a glomerulus) or ‘relative’ podocyte depletion (a decrease in the number of podocytes per unit volume of glomerulus) are both direct causes of focal and segmental glomerulosclerosis [25▪]
To identify the genetic, developmental, and environmental factors that lead to low nephron endowment, nephron loss, and podocyte depletion, it is of fundamental importance to be able to count glomeruli and their podocytes in an accurate (no bias) and precise (low variance) manner. Surprisingly, this has proven to be difficult and controversial. The purpose of this review is to consider current valid methods for counting glomeruli and podocytes, the pros and cons of these methods, and potential new approaches.
Brenner et al.[26,27] were the first to report a link between a glomerular deficit and hypertension in adulthood, and subsequently identified associations between low glomerular number and the development of renal disease. At roughly the same time, Barker and Osmond  identified links between low birth weight and adult disease, including cardiovascular disease. Given that human birth weight and glomerular number are directly correlated , it seems likely that low birth weight results in low glomerular number, which may increase the susceptibility to cardiovascular and renal disease in adulthood . In this context, glomerular number serves as a surrogate marker of the fetomaternal environment using glomerular endowment at the completion of nephrogenesis, which is around birth or 36 weeks of gestation in humans, and glomerular loss during childhood/adulthood using total number of glomeruli, which represents the number of glomeruli at a specific timepoint (nephron endowment minus the number of glomeruli subsequently lost during postnatal life). Thus, the study of glomerular number has the potential to provide important insights into kidney health both before and after birth. Although many methods for estimating glomerular number have been published, we briefly review the recently described approaches for estimating this key parameter in further sections.
Most researchers, and even renal pathologists, use the terms glomerular cross-sections and glomeruli interchangeably. However, there is a big difference between glomerular cross-sections (two-dimensional samples of glomeruli – essentially glomerular bits and pieces as seen on histological sections) and whole glomeruli.
The study of glomerular density would appear to be the most pragmatic approach for counting glomeruli. In short, glomerular cross-sections observed in histological sections are counted and then expressed as glomerular number per unit area of section. Tsuboi et al.[30–32] showed that low glomerular density was associated with increases in glomerular volume among patients with IgA nephropathy and obesity-related glomerulopathy. This group has also proposed that low glomerular density is directly associated with progression of IgA nephropathy [30,31], idiopathic membranous nephropathy , and response to corticosteroid therapy in adult patients with minimal change nephrotic syndrome .
Despite the fact that glomerular density is frequently reported, and may be informative in particular scenarios, it introduces a series of potential confounders. For example, it is very clear that the number of glomerular cross-sections in a histological section depends not only on the number of glomeruli present but also on glomerular size and shape [35,36]. Indeed, an estimate of glomerular density from a histological section may have significant bias, because large glomeruli have a greater chance than small glomeruli of being sampled. Thus, in a setting of glomerular hypertrophy, the glomerular density will be seen to have increased, although in fact glomerular number has not increased at all. Furthermore, the number of glomeruli seen in a histological section is influenced by section thickness. Hence, the relationship between glomerular density in a section and the number of glomeruli in a kidney is complex and difficult to predict. Importantly, knowing the number of glomerular profiles per unit area of section tells us nothing about the total number of glomeruli in the kidney. This is therefore a far from ideal method for estimating glomerular number when sufficient tissue is available for adequate analysis.
The current gold-standard method for estimating glomerular number is based upon the disector principle described by Sterio . This approach requires no knowledge or assumptions of glomerular geometry (size, shape), and when used correctly, provides unbiased estimates. Two general disector-based methods are available for counting glomeruli: the disector/Cavalieri principle  and the disector/fractionator principle . These methods have been used to estimate glomerular number in a range of species, including humans and rodents [40,41]. Although these methods are considered the current gold-standard techniques, they have a number of limitations, as follows: both require the analysis of a representative portion of the kidney, which means they are destructive and applicable only to terminal experiments – they can only provide cross-sectional data; most studies have used a plastic embedding medium (such as glycolmethacrylate) for dimensional stability, which requires expertise for tissue processing that may not be regularly available in laboratories; both require systematic slicing and exhaustive sectioning, which are not only expensive but also require significant skill; and even if all of these requirements are met, the hands-on counting time ranges from approximately 6 h for a rat kidney to 8 h for a human kidney. For all of these reasons, very few laboratories have adopted these approaches. Therefore, it is clear that a new method is urgently required.
A new approach
A new glomerular counting method utilizing MRI has recently been described [42,43,44▪▪]. Briefly, cationic ferritin perfused intravenously (rodents) or into excised kidneys (human) binds to anionic charges of the glomerular basement membrane, rendering the glomeruli visible with MRI. MRI of ex-vivo cationic ferritin-labeled kidneys provides estimates of glomerular number that are in excellent agreement with estimates obtained using the disector/fractionator approach .
To date, MRI has been used to quantify the total number of glomeruli in rat and human kidneys [42,43,44▪▪]. Magnetic resonance images of a human kidney are shown in Fig. 1a (cationic ferritin labeled) and 1b (negative control), respectively. Advantages of this new MRI approach include the following: the kidney is imaged whole and therefore the need for embedding, slicing, and sectioning is avoided; the estimates can be obtained in approximately one-sixth of the hands-on time of stereology; and as every labeled glomerulus is recorded, data on the glomerular size distribution are available, providing a potentially new and powerful technique for assessing glomerular growth, hypertrophy, shrinkage, and size variability within individuals. However, a major disadvantage is that one needs access to a high-field-strength MRI scanner.
Whether MRI can be reliably used to count and measure the size of glomeruli in vivo remains to be determined. A recent study by Qian et al.[46▪] reported in-vivo observation of rat glomeruli through signal amplification by a wireless amplified nuclear magnetic resonance (NMR) detector. Individual glomeruli were positively visualized by blood flow and negatively visualized by ferritin and Mn2+. This is an exciting advance in the field, but further work is required to identify better/alternative contrast agents and to provide sufficient evidence of efficacy and safety. Bennett et al. have discussed in greater detail the technical and regulatory challenges involved in developing MRI-based techniques for glomerular imaging.
The podocyte depletion hypothesis has gained considerable attention in the last decade, mainly because it represents a unifying concept of renal pathology . Podocyte depletion can be defined as ‘absolute’ when the total number of healthy podocytes per glomerulus has decreased (Fig. 2a), and as ‘relative’ when there is no reduction in podocyte number but there is an increase in the glomerular filtration surface area (or glomerular volume) that leads to an effective reduction in podocyte density (Fig. 2b).
In 2005, Wharram et al. described a transgenic rat model in which the human diphtheria toxin receptor was specifically expressed in podocytes in order to achieve dose-dependent podocyte depletion. This landmark study showed that while death of less than 40% of podocytes lead to reversible reductions in renal function and transient proteinuria, loss of more than 40% of podocytes resulted in segmental sclerosis with sustained proteinuria and reduced renal function. More importantly, this study provided evidence that direct podocyte injury was sufficient for the development of progressive glomerular disease, and defined a critical threshold of podocyte depletion in rodents. In 2007, Wiggins  proposed a podocyte depletion theory, placing the podocyte as the key cell in the development of glomerular diseases.
In the early 1990s, Nagata et al. showed that in a setting of glomerular hypertrophy, podocytes are able to undergo hypertrophy in order to cover an enlarged capillary surface area, which leads to podocyte stress, podocyte loss, areas of denuded glomerular basement membrane, and development of adhesions. Recently, Fukuda et al. confirmed these findings in a landmark study that highlighted how a ‘mismatch’ between podocyte volume and glomerular tuft growth can cause proteinuria, glomerulosclerosis, and progression to renal failure.
The key parameter when studying podocyte depletion is podocyte number. A variety of methods have been used over the past 20–30 years to estimate podocyte number. These are briefly considered in further sections, together with several new approaches described in the past 12 months.
Podocyte number per glomerular cross-section
The number of podocytes per glomerular cross-section is the most commonly used parameter for reporting podocyte number. However, although commonly reported, this parameter has serious flaws and the data are easily and frequently misinterpreted. The main problem is that podocyte nuclear profiles are counted. The number of these profiles is related not only to the number of podocytes present but also to the nuclear shape and size of the podocyte size, as well as the section thickness. Moreover, this method does not provide an estimate of the total number of podocytes in a glomerulus. Therefore, this parameter is of limited value when assessing podocyte depletion and should be avoided .
The stereological method of Weibel and Gomez  is also commonly used to estimate podocyte number. Although this method provides estimates of total podocyte number per glomerulus, it is designated as ‘model based’ because it requires knowledge of the geometry (size, size distribution, shape) of podocyte nuclei. Generally, values for these geometric features are assumed rather than measured, and to the extent that these assumptions are inaccurate, there is the potential for the introduction of systematic bias. This can be acceptable in certain circumstances wherein insufficient tissue is available for a more comprehensive approach. These difficulties in counting podocytes were considered recently by Lemley et al., who concluded that the disector/fractionator method was the preferred method when sufficient tissue was available, as, for example, in the case of whole autopsy kidneys.
A disector/fractionator method for estimating podocyte number was recently described by Puelles et al.[25▪]. This method combines serial sectioning of paraffin-embedded tissue, immunohistochemistry, confocal microscopy, the disector/fractionator principle (to count podocytes), and the disector/Cavalieri principle (to sample glomeruli and estimate their volume). This method has numerous advantages over previous methods that include the following: estimates of total podocyte number in glomeruli of known volume (and thereby podocyte density) are obtained; data describing the heterogeneity in total podocyte number and podocyte density between glomeruli from the same kidney are obtained; total numbers of other cell types (such as endothelial and parietal epithelial cells), and thereby cell number ratios (i.e., parietal epithelial cell/podocyte ratio) are obtained; and additional information such as the cortical location of glomeruli is also available. Using this approach, Puelles et al.[50▪▪] recently reported that large adult glomeruli contained more podocytes than glomeruli from young children, raising questions about the postnatal origin of these additional podocytes (Fig. 3a). Puelles et al.[50▪▪] also showed that despite an increased number of podocytes, large adult glomeruli had lower podocyte densities than smaller glomeruli, indicating that these large glomeruli had relative podocyte depletion, possibly placing them at greater risk of subsequent pathological change (Fig. 3b). These new insights into podocyte number and density were only obtained through the application of this new design-based method for podocyte counting.
However, despite the powerful insights provided by this new approach, it has a number of limitations that include the following: a considerable amount of time is required to obtain the confocal images (approximately 1.5 h per glomerulus) and count the podocytes (approximately 6 h per glomerulus); access to a laser confocal microscope is required; significant technical expertise is required for the serial sectioning and podocyte counting; and the method relies on specific immunostaining for podocyte identification, which may not always be reliable in pathological settings. Given the considerable amount of time to obtain estimates of podocyte number with this method, more cost-efficient methods are required. The two recently described methods detailed in the next sections are therefore of interest.
Counting podocytes with flow cytometry
Wanner et al.[51▪] recently described a new method for counting podocytes in which glomeruli are isolated using magnetic beads, a single cell suspension is then obtained, and podocytes are directly labeled with an antipodocin antibody. Podocyte number is then determined using flow cytometry. The authors report a 99% overlap between genetically labeled podocytes and podocytes detected with the antibody, suggesting that this is a highly specific technique. Although this is an excellent alternative that provides high-throughput data, the final output is podocyte number per kidney rather than per glomerulus. Even if glomerular number could be calculated first, the final result would be an average representation of the whole kidney, without any insights on the variations between glomeruli, cortical zones, or even the relationships between podocyte numbers and glomerular volume or numbers of other glomerular cell types. Nevertheless, this approach represents a potentially useful and high-throughput technique for assessing global renal podocyte number.
Podocyte number in clinical biopsies
Estimating podocyte numbers in renal biopsies has also proved challenging, primarily because of the limited amount of tissue available. Venkatareddy et al.[52▪▪] recently revived the stereological concept originally proposed by Abercrombie . A key requirement of this method is estimation of the mean caliper diameter of podocyte nuclei. Venkatareddy et al.[52▪▪] directly measured this caliper diameter and then calculated a correction factor to correct podocyte counts for section thickness. Podocyte density was then estimated, and assuming glomeruli were spherical, average podocyte number per glomerulus was obtained. Although design-based methods for estimating podocyte density may be theoretically preferable, they are also time consuming and often not suitable for estimating podocyte number or density in renal biopsies. The method of Venkatareddy et al.[52▪▪] is robust, simple to use, utilizes commonly available technologies, can be applied to large numbers of glomeruli in a biopsy or kidney section, and can be adapted for automated biopsy analysis. As long as the limitations and potential sources of bias are acknowledged and understood, this method may be a useful tool for estimating podocyte number in clinical samples.
There is currently much interest in understanding the importance of glomerular and podocyte number to adult health and disease. Obtaining accurate and precise estimates of glomerular and podocyte number in a timely fashion has proven difficult, although the past 2–3 years has witnessed the development of several new methods. More methodological advancements can be expected in the near future. These developments should improve our understanding of glomerular and podocyte numbers in adult health and disease, and ultimately contribute to the development of improved diagnostic and therapeutic options.
Financial support and sponsorship
Our glomerular studies were funded by grants from the National Institutes of Health (NIH 1 R01 DK065970-01), the NIH Center of Excellence in Minority Health (5P20M000534-02), the National Health and Medical Research Council of Australia (NHMRC), and the American Heart Association (Southeastern Affiliate). Our podocyte research has been funded by grants from the NHMRC (grant numbers 606619 and 1065902). V.G.P. received a Monash Research Graduate School Scholarship to support his PhD candidature.
Conflicts of interest
There are no conflicts of interest.
REFERENCES AND RECOMMENDED READING
Papers of particular interest, published within the annual period of review, have been highlighted as:
- ▪ of special interest
- ▪▪ of outstanding interest
1. Brenner BM. Nephron adaptation to renal injury or ablation. Am J Physiol 1985; 249 (3 Pt 2):F324–F337.
2. Hostetter TH, Olson JL, Rennke HG, et al. Hyperfiltration in remnant nephrons: a potentially adverse response to renal ablation. Am J Physiol 1981; 241:F85–F93.
3. Yeung EH, Robledo C, Boghossian N, et al. Developmental origins of cardiovascular disease. Curr Epidemiol Rep 2014; 1:9–16.
4. Ingelfinger JR, Nuyt AM. Impact of fetal programming, birth weight, and infant feeding on later hypertension. J Clin Hypertens (Greenwich) 2012; 14:365–371.
5. Dorey ES, Pantaleon M, Weir KA, Moritz KM. Adverse prenatal environment and kidney development: implications for programing of adult disease. Reproduction 2014; 147:R189–R198.
6. Luyckx VA, Bertram JF, Brenner BM, et al. Effect of fetal and child health on kidney development and long-term risk of hypertension and kidney disease. Lancet 2013; 382:273–283.
7. Desai M, Beall M, Ross MG. Developmental origins of obesity: programmed adipogenesis. Curr Diab Rep 2013; 13:27–33.
8. Yajnik CS. Transmission of obesity-adiposity and related disorders from the mother to the baby. Ann Nutr Metab 2014; 64 (Suppl 1):8–17.
9. Simmons RA. Developmental origins of diabetes: the role of oxidative stress. Best Pract Res Clin Endocrinol Metab 2012; 26:701–708.
10. Thompson JA, Regnault TR. In utero
origins of adult insulin resistance and vascular dysfunction. Semin Reprod Med 2011; 29:211–224.
11. Kretzler M, Koeppen-Hagemann I, Kriz W. Podocyte damage is a critical step in the development of glomerulosclerosis in the uninephrectomised-desoxycorticosterone hypertensive rat. Virchows Arch 1994; 425:181–193.
12. Kriz W, Gretz N, Lemley KV. Progression of glomerular diseases: is the podocyte the culprit? Kidney Int 1998; 54:687–697.
13. Kriz W, Endlich K. Hypertrophy of podocytes: a mechanism to cope with increased glomerular capillary pressures? Kidney Int 2005; 67:373–374.
14. Kim YH, Goyal M, Kurnit D, et al. Podocyte depletion
and glomerulosclerosis have a direct relationship in the PAN-treated rat. Kidney Int 2001; 60:957–968.
15. Wharram BL, Goyal M, Wiggins JE, et al. Podocyte depletion
causes glomerulosclerosis: diphtheria toxin-induced podocyte depletion
in rats expressing human diphtheria toxin receptor transgene. J Am Soc Nephrol 2005; 16:2941–2952.
16. Wiggins JE, Goyal M, Sanden SK, et al. Podocyte hypertrophy, “adaptation,” and “decompensation” associated with glomerular enlargement and glomerulosclerosis in the aging rat: prevention by calorie restriction. J Am Soc Nephrol 2005; 16:2953–2966.
17. Wiggins RC. The spectrum of podocytopathies: a unifying view of glomerular diseases. Kidney Int 2007; 71:1205–1214.
18. Fukuda A, Chowdhury MA, Venkatareddy MP, et al. Growth-dependent podocyte failure causes glomerulosclerosis. J Am Soc Nephrol 2012; 23:1351–1363.
19. Sison K, Eremina V, Baelde H, et al. Glomerular structure and function require paracrine, not autocrine, VEGF-VEGFR-2 signaling. J Am Soc Nephrol 2010; 21:1691–1701.
20. Eremina V, Cui S, Gerber H, et al. Vascular endothelial growth factor a signaling in the podocyte-endothelial compartment is required for mesangial cell migration and survival. J Am Soc Nephrol 2006; 17:724–735.
21. Eremina V, Sood M, Haigh J, et al. Glomerular-specific alterations of VEGF-A expression lead to distinct congenital and acquired renal diseases. J Clin Invest 2003; 111:707–716.
22. Shankland SJ. The podocyte's response to injury: role in proteinuria and glomerulosclerosis. Kidney Int 2006; 69:2131–2147.
23. Fiorina P, Vergani A, Bassi R, et al. Role of podocyte B7-1 in diabetic nephropathy. J Am Soc Nephrol 2014; 25:1415–1429.
24. Tharaux PL, Huber TB. How many ways can a podocyte die? Semin Nephrol 2012; 32:394–404.
25▪. Puelles VG, Douglas-Denton RN, Cullen-McEwen L, et al. Design-based stereological methods for estimating numbers of glomerular podocytes. Ann Anat 2014; 196:48–56.
This method provides unbiased estimates of total podocyte number in glomeruli of known volume, as well as estimates of podocyte density and ratios of podocyte numbers to other glomerular cell populations.
26. Brenner BM. The etiology of adult hypertension and progressive renal injury: an hypothesis. Bull Mem Acad R Med Belg 1994; 149:121–125.
27. Brenner BM, Mackenzie HS. Nephron mass as a risk factor for progression of renal disease. Kidney Int Suppl 1997; 63:S124–S127.
28. Barker DJ, Osmond C. Low birth weight and hypertension. BMJ 1988; 297:134–135.
29. Hughson M, Farris AB 3rd, Douglas-Denton R, et al. Glomerular number and size in autopsy kidneys: the relationship to birth weight. Kidney Int 2003; 63:2113–2122.
30. Tsuboi N, Kawamura T, Ishii T, et al. Changes in the glomerular density and size in serial renal biopsies during the progression of IgA nephropathy. Nephrol Dial Transplant 2009; 24:892–899.
31. Tsuboi N, Kawamura T, Koike K, et al. Glomerular density in renal biopsy specimens predicts the long-term prognosis of IgA nephropathy. Clin J Am Soc Nephrol 2010; 5:39–44.
32. Tsuboi N, Utsunomiya Y, Kanzaki G, et al. Low glomerular density with glomerulomegaly in obesity-related glomerulopathy. Clin J Am Soc Nephrol 2012; 7:735–741.
33. Tsuboi N, Kawamura T, Miyazaki Y, et al. Low glomerular density is a risk factor for progression in idiopathic membranous nephropathy. Nephrol Dial Transplant 2011; 26:3555–3560.
34. Koike K, Tsuboi N, Utsunomiya Y, et al. Glomerular density-associated changes in clinicopathological features of minimal change nephrotic syndrome in adults. Am J Nephrol 2011; 34:542–548.
35. Bertram JF. Analyzing renal glomeruli with the new stereology
. Int Rev Cytol 1995; 161:111–172.
36. Nyengaard JR. Stereologic methods and their application in kidney research. J Am Soc Nephrol 1999; 10:1100–1123.
37. Sterio DC. The unbiased estimation of number and sizes of arbitrary particles using the disector. J Microsc 1984; 134 (Pt 2):127–136.
38. Hinchliffe SA, Sargent PH, Howard CV, et al. Human intrauterine renal growth expressed in absolute number of glomeruli assessed by the disector method and Cavalieri principle. Lab Invest 1991; 64:777–784.
39. Bertram JF, Cullen-McEwen LA, Egan GF, et al. Why and how we determine nephron number
. Pediatr Nephrol 2014; 29:575–580.
40. Cullen-McEwen LA, Douglas-Denton RN, Bertram JF. Estimating total nephron number
in the adult kidney using the physical disector/fractionator combination. Methods Mol Biol 2012; 886:333–350.
41. Cullen-McEwen LA, Armitage JA, Nyengaard JR, Bertram JF. Estimating nephron number
in the developing kidney using the physical disector/fractionator combination. Methods Mol Biol 2012; 886:109–119.
42. Beeman SC, Zhang M, Gubhaju L, et al. Measuring glomerular number and size in perfused kidneys using MRI. Am J Physiol Renal Physiol 2011; 300:F1454–F1457.
43. Heilmann M, Neudecker S, Wolf I, et al. Quantification of glomerular number and size distribution in normal rat kidneys using magnetic resonance imaging. Nephrol Dial Transplant 2012; 27:100–107.
44▪▪. Beeman SC, Cullen-McEwen LA, Puelles VG, et al. MRI-based glomerular morphology and pathology in whole human kidneys. Am J Physiol Renal Physiol 2014; 306:F1381–F1390.
The first report to use MRI to count and measure the size of human glomeruli. Represents an important step toward ultimately estimating glomerular number in patients with a suspected nephron deficit.
45. Bennett KM, Bertram JF, Beeman SC, Gretz N. The emerging role of MRI in quantitative renal glomerular morphology. Am J Physiol: Renal Physiol 2013; 304:F1252–F1257.
46▪. Qian C, Yu X, Pothayee N, et al. Live nephron imaging by MRI. Am J Physiol: Renal Physiol 2014; 307:F1162–F1168.
A landmark report that for the first time used MRI to image glomeruli in vivo, both with and without contrast agent.
47. Nagata M, Scharer K, Kriz W. Glomerular damage after uninephrectomy in young rats. I. Hypertrophy and distortion of capillary architecture. Kidney Int 1992; 42:136–147.
48. Lemley KV, Bertram JF, Nicholas SB, White K. Estimation of glomerular podocyte number
: a selection of valid methods. J Am Soc Nephrol 2013; 24:1193–1202.
49. Weibel ER, Gomez DM. A principle for counting tissue structures on random sections. J Appl Physiol 1962; 17:343–348.
50▪▪. Puelles VG, Douglas-Denton RN, Cullen-McEwen LA, et al. Podocyte number
in children and adults: associations with glomerular size and numbers of other glomerular resident cells. J Am Soc Nephrol 2015; doi: 10.1681/ASN.2014070641. [Epub ahead of print].
This article provides new insights into podocyte numbers in children and adults. Interestingly, although large adult glomeruli contain more podocytes than small glomeruli, their podocyte density is lower, suggesting that these large glomeruli may be at a heightened risk of pathological change.
51▪. Wanner N, Hartleben B, Herbach N, et al. Unraveling the role of podocyte turnover in glomerular aging and injury. J Am Soc Nephrol 2014; 25:707–716.
This article describes a new technique utilizing magnetic beads and flow cytometry to estimate the total number of podocytes in whole mouse kidneys.
52▪▪. Venkatareddy M, Wang S, Yang Y, et al. Estimating podocyte number
and density using a single histologic section. J Am Soc Nephrol 2014; 25:1118–1129.
This article extends the method of Abercrombie  to count podocytes using single paraffin sections. The method is ideally suited for the assessment of podocyte density in renal biopsies.
53. Abercrombie M. Estimation of nuclear population from microtome sections. Anat Rec 1946; 94:239–247.