Backtracking cryptic recurrence of esophageal cancer from membranous nephropathy: the detection of glomerular NELL-1 and IgG4

ABSTRACT

We reported the detection of neural epidermal growth factor-like 1 (NELL-1) and immunoglobulin G4 (IgG4) on glomerular capillary wall in membrane nephropathy (MN), which led to the discovery of early post-operative recurrence of esophageal squamous cell cancer (ESCC) in a 68-year-old man. Further, NELL-1 was also identified in the cancerous tissue sampled by esophagoscope. Moreover, serum IgG4 percentage seemed to be higher when comparing with both previous reports and another age-matched male with NELL-1-negative MN upon fully recovered ESCC. Therefore, the finding of NELL-1 in a renal biopsy should trigger a detailed workup in search of malignancy, especially with concomitant IgG4 dominance.

BACKGROUND

Neural epidermal growth factor-like 1 (NELL-1) is a recently described antigen in membranous nephropathy (MN) [1]. Reportedly, NELL-1-positive MN accounts for an approximate 5%–10% and 2.5% prevalence in primary cases and across the entire spectrum of MN, respectively [2, 3]. Further, this specific entity had unique histopathologic features and was more often associated with malignancy [4]. In fact, Caza et al. confirmed that NELL-1 could be the first antigen in malignancy-associated MN and their cohort of 91 NELL-1-positive MN manifested 30 distinct types of primary malignancy, but short of esophageal cancer [5]. Intriguingly, both NELL-1 and immunoglobulin G4 (IgG4) were precisely implicated in the tumorigenesis of esophageal cancer [6, 7]. As such, we hereby described the detection of glomerular NELL-1 and IgG4 in MN, which subsequently led to the uncovering of asymptomatic thus occult recurrence of esophageal squamous cell carcinoma (ESCC). Of importance, NELL-1 was simultaneously identified in the cancerous tissue sampled by esophagoscope. These findings offered a unique chance to explore the role of NELL-1 and IgG subclass in malignancy and the associated MN, especially when cross-validated in patients with and without recurrent ESCC.

THE CASES

A 68-year-old male was admitted due to edema of both lower limbs for more than 3 months. He had diabetes mellitus and underwent radical esophagectomy at a local hospital 5 years and 1 year ago, respectively. During the hospital stay, his pre-operational urinalysis and renal function were normal and post-operational pathology found poorly differentiated ESCC, and he subsequently received four cycles of capecitabine and oxaliplatin regimen (Fig. 1). Thereafter, the patient claimed to be in “good shape” until the occurrence of edema. On arrival, he was normotensive but felt general asthenia and the following workup showed (Supplementary data, Table) slight anemia, nephrotic range proteinuria, hypoalbuminemia, normal serum creatinine and no evidence of secondary MN including paraproteinemia. Of note, his serum IgG1 and IgG4 took up 63.4% and 16.0% of total IgG subclasses. Chest computed tomography (CT) scan showed an ostensible lesion in the anterior segment of the right upper lobe which was grossly the same on the pre-operational CT image. However, we have noticed a less discernible and unreported anomaly around the esophageal anastomotic site (Fig. 2), and proceeded with renal biopsy. Upon the detection of MN with positive NELL-1 on the glomerular capillary wall (Fig. 1), positron emission tomography–CT was promptly ordered. Consistently, irregular thickening of the above site with hypermetabolism was found, which strongly indicated recurrence of the ESCC. Eventually, ESCC and NELL-1 were respectively identified by pathological examination and immunohistochemical staining in tissue samples taken from close proximity of the anastomotic site by esophagoscope. Fully aware of the medical situation, the patient elected to leave with further therapeutic options at his discretion. With angiotensin receptor blockade but the malignancy left unattended, his urinary protein excretion and serum albumin 2 and 6 weeks after discharge were 7.41 g/24 h/23.6 g/L and 10.5 g/24 h/21.4 g/L, respectively. The patient accepted radiotherapy shortly thereafter and the prescription dose was: pGTV and pGTVnd five times per week with 200 cGy each for 6 weeks that totaled 6000 cGy/30 times. One month after the completion of radiotherapy, his urinary protein excretion was 2.38 g/24 h and serum albumin 27.9 g/L.

Figure 1:

Top panel: a brief overview of the patient's disease progression and major clinical examinations. (A1–A4) Light microscope showing glomerular features consistent with membranous nephropathy. (B1–B4) Immunofluorescence microscopy demonstrating segmental IgG/IgG4, C3 and NELL-1 staining along the glomerular capillary wall. (C1) Electron microscopy showing electron-dense deposits along the glomerular capillary wall and diffuse effacement of the podocyte foot process. (C2) Esophageal tissue under light microscope showing chronic inflammation of mucosa, infiltrative growth of atypical squamous epithelial nests under the squamous and glandular epithelium. These findings are consistent with ESCC. (C3 and C4) Immunohistochemistry and immunofluorescence showing positive staining for NELL-1, respectively, in esophageal tissue samples. For the latter, the control had zero fluorescent signal. Upro: 24-h urinary excretion of protein; E NELL-1: testing NELL-1 in esophageal tissue sample by methods in C3 and C4.

Figure 2:

(A1 and A2) Lung and mediastinal window, with the latter showing an aberrant thickening of esophageal wall around the anastomotic site (arrow). (A3) View under the esophagoscope showing local protrusion with mucosal hyperemia. (Column B, C and D) Triaxial view depicting increased soft tissue shadow at the anastomotic site and abnormally augmented 18-fluoro fluorodeoxyglucose (¹⁸F-FDG) uptake (cross-hair).

In an age-matched male who also had a history of ESCC and biopsy-confirmed NELL-1-negative MN, Rituximab was used as there was no sign of recurrence (Supplementary data, Figs S1 and S2). He had acquired complete remission eventually: urinary protein excretion 0.276 g/24 h and serum albumin 40.7 g/L. Of note, his serum IgG1 and IgG4 on admission took up 69.2% and 2.5% of total IgG subclasses.

DISCUSSION

NELL-1 is a secreted, non-membrane bound 90-kDa glycoprotein expressed in osteoblasts and promotes bone regeneration [8]. By nature, it contains six epidermal growth factor-like repeats, in addition to a thrombospondin N-terminal domain, histidine- and cysteine-rich domains, and several von Willebrand factors [4]. By function, NELL-1 was involved in a wide range of human diseases including tumorigenesis [9]. It is now clear that it also had an active role in the pathogenesis of MN in general and a malignancy-associated one in particular, thanks to the latest work of Sethi et al. and Caza et al. [1, 5]. By confocal microscopy, both IgG and NELL-1 co-localized to the glomerular basement membrane [1]. By histopathology, moreover, this glomerular distribution had a unique appearance characterized by segmental to incomplete IgG capillary loop staining (93.4%) and IgG1 predominance (95.5%), and one-third of patients with NELL-1-positive MN had a history of malignancy [4, 5]. Unlike the PLA2R, THSD7A and SEMA3B, however, NELL-1 was not normally expressed on the podocyte surface [10]. This fact indicated that NELL-1 is either “planted” as a neo-epitope or “entrapped” as part of an immune complex in the subepithelial space. In this regard, we and others have previously reported glomerular THSD7A deposit in the malignancy-associated MN [10, 11]. Arguably, our current findings were in general agreement with the above perspectives.

Nonetheless, some obvious inconsistence must be addressed. As a known tumor suppressor gene (TSG), silencing of NELL-1 gene transcription by promoter hypermethylation is a common event in early-stage esophageal cancer [6]. However, tumor-based studies have shown that tumor suppressors may paradoxically divert into oncogenic factors, which is an epigenetic-regulatory, type-specific and stage-dependent phenomenon [12]. To be exact, Nell-1 was observed to have pro-angiogenic effects, both in vitro and in vivo [13]. This notion may help explain the normal urinalysis (as surrogate of absent NELL-1) in the first episode of symptomatic ESCC and detectable tissue NELL-1 during the early recurrence. Likewise, it may be readily extended to our discovery of the glomerular IgG4 deposit.

IgG1 and IgG4 are the most and least IgG subclass in term of abundance in serum (61.0% vs 3.0%) [14] and predominance in NELL-1-positive malignancy-associated MN (73.1% vs 4.5%) [5], respectively. By function, IgG4 has the lowest capacity to engage C1q and Fc receptors, and to activate immune effector cell [14]. Furthermore, IgG4-containing B lymphocytes and IgG4 concentration were significantly increased in the neoplastic tissue in patients with esophageal cancer [7]. Namely the IgG4 blockade in cancer, locally increased IgG4 in the cancer microenvironment could inhibit the IgG1-mediated anticancer responses, help cancer to evade local immune attack and indirectly promote cancer growth. Indeed, our patient with recurrent ESCC had lower serum IgG1 concentration and higher IgG4 concentration/percentage comparing with corresponding value in the patient without recurrent ESCC. These observations, particularly the IgG4 variations, corresponded well with the glomerular NELL-1 and IgG4 deposit and were the rationale for our successful backtracking of the recurrence of ESCC. Conversely, IgG4 was most frequent in malignancy-associated MN with either anti-PLA2R or anti-THSD7A antibodies [7]. This discrepancy indicated that subtle features of the glomerular lesion depended on the biological behavior of malignancy and the patient's immune reaction. Nevertheless, mechanisms underlying anti-podocyte autoimmunity induced by cancer warrants further research.

We detected NELL-1 in glomerulus but not in the serum. According to the kidney as a sink theory, a small amount of circulating antigen may be accumulatively enriched in the glomerulus and render detection possible [3]. In parallel, immunohistochemistry showed definite positive NELL-1 staining in the endoscope-harvested tissue and, to a lesser extent, immunofluorescence displayed positive spots as well. Presumably, this deviation in signal strength was most likely due to the use of paraffin sections in both procedures because of limited amount of the sampled tissue. Taken together, all the above lines of evidence pointed to the recurrence of ESCC as the source of NELL-1. Consistently, effective treatment of the underlying malignancy usually elicited good response of the NELL-1-positive MN per se [5]. From the professional perspective, possible nephrotoxicity of the chemotherapy was also carefully contemplated and we found no previous report of MN elicited by the capecitabine/oxaliplatin regimen [15].

CONCLUSION

In summary, our work is not nephrology in the simple sense, but rather a blend of immunology, oncology and the unique experience of renal pathology. We hence tried to use it to implicate the interplay across the functional transformation of NELL-1 as a TSG, involvement of IgG4 and tumorigenesis, and accurately discover the underlying malignancy behind the finding of NELL-1-positive MN in a biopsy.

CONFLICT OF INTEREST STATEMENT

No conflict of interest for any of the authors.

ETHICAL APPROVAL

The study was approved by ethical committee of the First Hospital of Hebei Medical University (20 220 744) in accordance with the Helsinki Declaration.

PATIENT CONSENT

The authors would like to thank the patients who had given their informed consent to publish these cases.