An audit of electron microscopy in the diagnosis of focal segmental glomerulosclerosis – are current pathological techniques missing important abnormalities in the glomerular basement membrane?

Focal segmental glomerulosclerosis (FSGS) is a pathological diagnosis which underpins a variety of progressive renal diseases that commonly result in proteinuria, chronic kidney disease (CKD) and potential end stage kidney disease (ESKD) requiring renal replacement therapy (1). FSGS occurs in either a primary. secondary, or genetic form. Primary FSGS is thought to be largely immunological in nature perhaps driven by an elusive permeability factor and secondary FSGS caused by compensatory hyperfilitration due to a previous glomerular injury (2). In recent years there has been a greater appreciation of the contribution of underlying genetic causes of this histological pattern. Many of these genetic aetiologies for FSGS have potential clinical significance for at-risk family members, subsequent genetic counselling, future living related kidney transplantation and therapeutics including potential avoidance of immunosuppressive therapies otherwise used for primary immunologically-mediated FSGS. Although initial focus has been on genes that are involved in the maintenance of podocyte structure and function it has become apparent that abnormalities in genes responsible for the structural integrity of the glomerular basement membrane (GBM) may underpin a significant minority of cases of adult-onset FSGS (3). Variants in COL4A3, COL4A4 and COL4A5 which encode the α3, α4 and α5 chains of collagen IV respectively, the major constituent of the GBM, previously linked to both Alport syndrome and thin basement membrane nephropathy (TBMN) may also underlie cases of FSGS (4). This suggestion has biological plausibility. Podocytes, the glomerular epithelial cells responsible for maintenance of the filtration barrier of the glomerulus through which plasma is ultrafiltrated, are not only a source for the α-chains


Abstract
Background Background Focal segmental glomerulosclerosis (FSGS) is a pathological diagnosis which underpins a variety of progressive renal diseases that commonly result in proteinuria, chronic kidney disease (CKD) and potential end stage kidney disease (ESKD) requiring renal replacement therapy (1). FSGS occurs in either a primary. secondary, or genetic form. Primary FSGS is thought to be largely immunological in nature perhaps driven by an elusive permeability factor and secondary FSGS caused by compensatory hyperfilitration due to a previous glomerular injury (2). In recent years there has been a greater appreciation of the contribution of underlying genetic causes of this histological pattern. Many of these genetic aetiologies for FSGS have potential clinical significance for at-risk family members, subsequent genetic counselling, future living related kidney transplantation and therapeutics including potential avoidance of immunosuppressive therapies otherwise used for primary immunologically-mediated FSGS. Although initial focus has been on genes that are involved in the maintenance of podocyte structure and function it has become apparent that abnormalities in genes responsible for the structural integrity of the glomerular basement membrane (GBM) may underpin a significant minority of cases of adult-onset FSGS (3). Variants in COL4A3, COL4A4 and COL4A5 which encode the α3, α4 and α5 chains of collagen IV respectively, the major constituent of the GBM, previously linked to both Alport syndrome and thin basement membrane nephropathy (TBMN) may also underlie cases of FSGS (4). This suggestion has biological plausibility. Podocytes, the glomerular epithelial cells responsible for maintenance of the filtration barrier of the glomerulus through which plasma is ultrafiltrated, are not only a source for the α-chains secreted to form the GBM but also are required to be anchored to this structure in order to maintain the aforementioned filtration barrier (2,5). It is not inconceivable then that inherited structural abnormalities of the GBM may lead to subsequent podocyte dysfunction and ultimately the lesion histologically characterised as FSGS.
The relationship between the three renal conditions intertwined around variants in the collagen IV genes; Alport syndrome, TBMN, and FSGS is complex and incompletely understood. Whilst previously it was believed that patients heterozygous for variants in COL4A3 and COL4A4 would develop TBMN with persistent microscopic haematuria and an otherwise benign prognosis, this traditional thinking has been overturned by the discovery that some pedigrees with these variants will go on to develop significant proteinuria, FSGS lesions and the potential for progressive CKD or ESRD (6,7). More recently, targeted gene sequencing of adults thought to have primary FSGS or steroid resistant nephrotic syndrome (the paediatric equivalent) found that pathogenic or likely pathogenic variants in collagen IV genes were present in up to 38% of families with familial FSGS, and 3% of those with sporadic FSGS (3,8). These findings highlight the importance of considering variants in these genes in the diagnosis and subsequent management of patients found to have FSGS histologically.
FSGS is a clinicopathological diagnosis that is made on the review of renal tissue obtained by percutaneous biopsy. This tissue is processed and subjected to several different staining techniques and methods of microscopy, including light, immunofluorescence and electron microscopy (9). FSGS, suggested on light microscopy by the finding of focal, segmental lesions with sclerosis, and an increase in the glomerular matrix with obliteration of the capillary lumens only requires this finding in one glomerulus and light microscopy to diagnose (2). As such, some specimens may previously have not been sent for further processing and review with other techniques such as electron microscopy even though this histopathology has heterogenous causative aetiologies. The other potential renal lesions that may be caused by collagen IV gene variants, Alport syndrome and TBMN, cause thinning, lamellation and fraying of the GBM and may also be associated with podocyte foot process effacement, all of which require electron microscopy to visualise (4). It should be noted that for Alport syndrome there are other techniques aside from electron microscopy available for identifying the diagnosis. Immunostaining of a renal biopsy specimen for type IV collagen may demonstrate the absence of alpha 3,4 or 5 chains in up to two thirds of patients, or a skin biopsy may show an absence of an alpha 5 chain (10,11), in addition to the aforementioned genetic testing. Such absence by immunostaining does not appear be present in FSGS (12). Given that the pathological diagnosis of FSGS does not routinely require electron microscopy it is therefore conceivable that GBM lesions potentially associated with an underlying collagen 4 variant may have been missed. This represents an opportunity to audit our prior clinical behaviour to see if our multidisciplinary diagnostic practice may require improvement. We conducted a retrospective audit of prior renal biopsy results in two tertiary centres to see how many of those that had been given a histological diagnosis of FSGS were sent for subsequent electron microscopy.

Methods
Our study was a retrospective cohort analysis across two tertiary hospital sites involving the review of prior renal biopsy results that had been given the histological diagnosis of FSGS on light microscopy and how many of these samples subsequently went on to be processed for electron microscopy. In addition, of those samples that were subjected to electron microscopy we reviewed how many displayed evidence of a potential collagen IV disorder. Our audit had human ethics approval at the participating hospitals (Barwon HREC 18/131; HREC/18/QRBW/258). Samples for biopsy were placed in the following fixatives for processing; formalin for light microscopy, saline soaked gauze with freezing for immunofluorescence, and glutaraldehyde for electron microscopy. Samples for light microscopy are embedded and sectioned at 2 to 3µm thickness with hematoxylin-eosin, Jones silver, periodic acid-schiff and trichrome staining performed. Immunofluorescence samples are sectioned within a cryostat and placed on prelabelled air-dried slides of the antigen in question. Electron microscopy tissue is processed into plastic, trimmed, cut into a 1µm section and stained with toludine blue. The images are reviewed in a digital medium (13).
A search strategy was developed for this retrospective cohort study to identify appropriate patients through local electronic medical record (EMR) systems using the search term 'focal segmental glomerulosclerosis' with the AND operator to combine with 'glomerulonephritis' or 'hereditary nephritis'. These patient results were subsequently reviewed for evidence of a previous percutaneous renal biopsy and report. The data was independently extracted from the included patients by the primary reviewer and collated in a Microsoft excel document that included information on age, primary diagnosis, included use of electron microscopy and any changes that may be consistent with an underlying collagen disorder including thinning, lamellation and fraying of the GBM. The Of the 43 patients identified 30 underwent electron microscopy after initial light microscopy and immunofluorescence. Two of these samples showed signs on electron microscopy that might be consistent with an underlying collagen IV glomerular basement membrane disorder. Microscopic haematuria was not noted in either of these two patients.
Of the 13 samples which did not undergo electron microscopy, four had no glomeruli present within the processed core. The remaining 9 had no documented reason.
The overall percentage of biopsy samples for which data was available that were not subjected to electron microscopy was 30%, of which close to 21% had no documented reason for not undergoing electron microscopy. The annualised rate of biopsy samples that were not subjected to electron microscopy varied with time ( Figure 1). The number of biopsy reports available for analysis prior to 2013 was only one per year at most, however from 2013 onwards the rate progressively increased to between 3 to 11 cases of FSGS diagnosed at histopathology between the two institutions per year. The annualised rate of electron microscopy from 2013 onwards varied each year, ranging from 50 to 87%. Overall 30% of biopsy samples did not receive electron microscopy. The overall rate of electron microscopy did not significantly change across the study period between the two centres.

Discussion
This retrospective cohort analysis demonstrated that about two thirds of native kidney biopsy samples across two institutions that were deemed to have primary FSGS underwent subsequent electron microscopy. Of those that did, two were reported to have characteristics that might be consistent with an underlying collagen IV disorder. In both samples electron microscopy revealed a diffusely thin GBM, with the second additionally identifying early focal splitting of the GBM. The first sample was suggested to be consistent with TBMN whereas there was no pathological comment made about the second. FSGS (NOS) was the most common lesion described in this study, which is consistent with prior reports (2). Importantly close to one in three cases of primary FSGS were not proceeding to electron microscopy despite an indication to do so and 1 in 20 cases within our cohort had structural changes that were consistent with an underlying collagen IV variant. Whilst some samples were unable to undergo electron microscopy due to a lack of glomeruli in the biopsy core, in the majority of the others it is unclear why subsequent electron microscopy did not occur. The annualised rate of biopsy samples not subjected to electron microscopy varied by year, but on average around one in three samples were not subjected to electron microscopy despite receiving a histological diagnosis of FSGS.
Curiously, neither of the two patients in whom electron microscopy was suggestive of an underlying collagen IV disorder was noted to have haematuria on their urinalysis at the time of presentation. Both of these samples were noted to have a thin basement membrane on electron microscopy, with focal splitting noted in one with an average thickness of 230.72nm given. Unfortunately, due to the retrospective nature of this study we were unable to send any samples for immunostaining of collagen IV.
There is an increasing body of evidence indicating that inheritable variants in collagen IV genes may underlie a proportion of cases of FSGS, with up to 12.5% cases of autosomal dominant FSGS attributable to COL4A3 in some cohorts (14). Not subjecting these renal biopsy samples to electron microscopy represents a potential gap in the investigation and subsequent management of such patients given they are much less likely to respond to immunosuppressive therapy (15) which has otherwise been classically indicated.
This study was designed as a retrospective cohort study looking at the number of samples sent for electron microscopy as well as any potential changes which might be consistent with a collagen IV glomerular basement membrane disorder. It is important to recognise that not all groups have found the characteristic changes associated with the collagen IV disorders such as Alport's Syndrome or TBMN on electron microscopy. One study described the typical pathological changes of FSGS but not the glomerular basement membrane abnormalities characterising Alport syndrome or TBMN in patients known to have variants in either COL4A3 or COL4A4 (7). It is thus possible that a lack of classical findings for a collagen IV glomerular basement membrane disorder may have accounted for the low number of those with GBM features on electron microscopy consistent with a collagen 4 disorder noted within our study. Suggesting against this however another study suggested that biopsy samples from patients with the classical features of Alport Syndrome or TBMN showed podocyte detachment which might be expected and subsequently cause FSGS type lesions (4). Other studies which have looked at electron microscopy in FSGS cases have similarly found low numbers of abnormalities that may be consistent with an underlying collagen 4 disorder (3,8) which suggests the overall number of abnormalities to be found via electron microscopy may be low.
The process by which variants within the collagen IV genes might cause FSGS remains unclear, particularly given their clear association with Alport Syndrome and TBMN. One proposal is that the ultrastructural changes induced by the collagen IV variants, perhaps under the influence of modifier genes such as laminin, result in impaired podocyte attachment to the glomerular basement membrane which leads to accelerated podocyte detachment, subsequent foot process effacement as a response to the increased shear stress induced by the denuded basement membrane and at a critical level of podocyte loss collapse of the capillary network with the appearance of the classical segmental sclerotic lesion (2,4,16). It also remains unclear as to whether the changes of FSGS are a secondary process occurring in those with TBMN or whether the collagen 4 variants are capable of causing primary FSGS (7,17). FSGS occurring as a secondary process to other basement membrane abnormalities may explain why immunosuppressive therapy has traditionally been less effective in inherited forms of FSGS, although there are case reports of the successful use of the calcineurin inhibitor cyclosporine for some patients harbouring inheritable collagen 4 disorders (6).
In summary, this study has found that not all biopsy samples that had primary FSGS as a histological diagnosis were subjected to subsequent electron microscopy. This may have potentially led to inadvertently overlooking characteristic basement membrane abnormalities which may suggest an underlying and heritable collagen IV disorder. These findings reflect an opportunity to change practice in order to better investigate, counsel and provide clinical management to these and future patients.

Consent for publication
Not applicable.

Availability of data and materials
All data generated and/or analysed during the current study are contained within an endnote user file and Microsoft excel documents. The former are available from the corresponding author upon reasonable request, the latter is supplied as a separate appendix file.

Competing interests
Justin Davis, Alwie Tjipto, Katherine Hegarty and Andrew Mallett declare no competing interests.

Funding
No funding was obtained for this study. There was no internal or external funding provided for the production of this manuscript.

Authors contributions
All authors made substantial intellectual contributions to the manuscript. The audit was overseen by the lead authors. All authors participated in the interpretation of data and vouch for the completeness and accuracy of the data. The first author (JD) wrote the first draft with all authors involved in the conception and design of the manuscript. All authors participated in the development of this manuscript and made the decision to publish the results.
The acquisition, analysis and interpretation of data was led by the corresponding author JD with equally important contributions and oversight including the resourcing of key references from AT, KH and AM. All authors were both involved in the drafting of the manuscript and critical revision for important intellectual content. All authors have given final approval for the version to be published. Each author has participated sufficiently in the work to take public responsibility for appropriate portions of the content and both authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of the work are appropriately investigated and resolved.

Conflict of interest statement
None declared.  Figure 1 Annualised rate of biopsy samples subjected to electron microscopy.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.