NRAS Mutations May Be Involved in the Pathogenesis of Cutaneous Rosai Dorfman Disease: A Pilot Study

Background: Purely cutaneous Rosai-Dorfman disease (RDD) is a rare histiocytic proliferative disorder limited to the skin. To date, its pathogenesis remains unclear. Owing to recent findings of specific mutations in the mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) pathway in histiocytic proliferative disorders, it provides a novel perspective on the pathomechanism of cutaneous RDD. We aim to investigate the genomic mutations in MAPK/ERK pathway in cutaneous RDD. Methods: We retrospectively recruited all cases of cutaneous RDD from two hospitals in Taiwan from January 2010 to March 2020 with the clinicopathologic features, immunohistochemistry, and treatment. Mutations of neuroblastoma RAS viral oncogene homolog (NRAS), Kirsten rat sarcoma 2 viral oncogene homolog (KRAS), and v-raf murine sarcoma viral oncogene homolog B1 (BRAF) in MAPK/ERK pathway were investigated by the highly sensitive polymerase chain reaction with Sanger sequencing. Results: Seven patients with cutaneous RDD were recruited with nine biopsy specimens. The median age was 46 years (range: 17–62 years). Four of seven patients (57.1%) received tumor excision, while the other three chose oral and/or topical or intralesional steroids. NRAS mutation was detected in 4 of 7 cases (4/7; 51.7%), and NRAS A146T was the most common mutant point (n = 4/7), followed by NRAS G13S (n = 2/7). There is no KRAS or BRAF mutation detected. Conclusions: We report the NRAS mutation is common in cutaneous RDD, and NRAS A146T was the most frequent mutation in this cohort. Mutations in the NRAS gene can activate the RAS/MAPK signaling and have been reported to be associated with various cancers. It indicates that NRAS mutation in MAPK/ERK pathway may involve the pathogenesis of cutaneous RDD.


Introduction
Rosai-Dorfman disease (RDD), also known as "sinus histiocytosis with massive lymphadenopathy", is a rare histiocytic proliferative disorder. It was first described in 1965 by Destombes, and later recognized as a distinct clinicopathologic entity of pale-staining histiocytes in 1969 by Rosai and Dorfman [1,2]. RDD was traditionally classified as non-Langerhans cell histiocytosis (non-LCH) [3], and now belongs to the "R" group of the

Patient Enrollment
This case study had retrospectively reviewed all medical records with a pathological diagnosis of cutaneous RDD in the Kaohsiung Veterans General Hospital and the Ditmanson Medical Foundation Chia-Yi Christian Hospital in Taiwan during the period from January 2010 to March 2020. All cases were histopathologically confirmed, and would be included if the following criteria were all met: (1) a pure cutaneous lesion; (2) proliferation of RD cells; (3) the RD cells were immunohistochemically positive for S100 and CD68 but negative for CD1a. Data were retrospectively collected and included age at presentation, sex, clinical symptoms, histopathological findings, immunohistochemical profiles, clinical management, and outcome. Patients with systemic manifestations or lymphovascular involvements were excluded. The study was approved by the Ethics Committees of Ditmanson Medical Foundation Chia-Yi Christian Hospital (CYCH-IRB-2021008, on 20 January 2021) and was conducted under the ethical guidelines of the Declaration of Helsinki.

Histopathology and Immunohistochemistry
The specimens of resected primary tumors were retrieved and examined with hematoxylin and eosin staining and immunohistochemistry including at least S-100, CD68, CD1a, and LCA to confirm the histopathologic diagnosis. With paraffin-embedded blocks of all RDD patients, formalin-fixed, paraffin-embedded sections were prepared with 4 µm thick and stained with hematoxylin and eosin for histological evaluation. Immunohistochemistry was performed on deparaffinized sections using an autostainer (Bond-Max autostainer; Leica Biosystems Newcastle Ltd., Melbourne, Australia) with hematoxylin as counterstain. The primary antibodies applied for confirming diagnosis were as follows: LCA (M0701, 1:200), CD1a (NCL-L-CD1a-235, 1:30), S100 (Z0311, 1:3000), and CD68 (M0814, 1:100). Appropriate positive and negative controls were utilized. The diagnoses of all patients had been confirmed by two experienced pathologists (J.-B. Liao and C.-C. Chen).

Mutation Analysis for KRAS, NRAS, and BRAF
Genomic DNA was extracted from formalin-fixed, paraffin-embedded blocks of all 7 cases using the QIAamp ® DNA FFPE Kit (Qiagen, Vialencia, CA, USA) according to manufacturer instructions. MAPK/ERK pathway mutations in either the KRAS, NRAS or BRAF gene were further screened by highly sensitive polymerase chain reaction (PCR) amplification followed by Sanger sequencing, using the KRAS mutant-enriched PCR Kits (FemtoPath ® , HongJing Biotech Inc, New Taipei City, Taiwan), the NRAS mutation screen PCR Kits (FemtoPath ® ), and the BRAF mutation screen PCR Kits (FemtoPath ® ) according to manufacturer recommendations. Since sample nucleic acids with the mutation sequence were preferentially amplified over the wild type sequence, we subsequently sequenced the regions of known mutations in KRAS, NRAS, and BRAF using Sanger sequencing. PCR products were sequenced in the reverse direction with primers of KRAS, NRAS, and BRAF mentioned above and followed by Sanger sequencing (ABI3730XL, Genomics ® ,Genomics, New Taipei City, Taiwan). The sequence was interpreted by visual inspection of the DNA sequence generated by Chromas software (Technelysium Pty Ltd., Queensland, Australia).

Results
A total of 7 patients with cutaneous RDD were retrospectively recruited and analyzed with histological features, clinical characteristics, and follow-up data. The patients were four men and three women with a median age of 46 years (range: 17-62 years). All the cases suffered from persistent cutaneous lesions for at least one month, and three of them for more than one year. The cutaneous RDD was solitary in 6 patients and multifocal in 1 patient. Two patients had anemia, but one of them had a history of thalassemia. None of the patients had a fever or severe hematopoietic abnormalities. Four of seven patients (57.1%) received tumor excision without evidence of recurrence, while the other three chose oral and/or topical or intralesional corticosteroids. Case 7 received oral corticosteroids with significant clinical improvement, but case 6 received oral and topical corticosteroids with refractory treatment response. In our experience, surgery was the first-line treatment for resectable skin lesions, while topical and systemic corticosteroid therapy should be considered alternatively for unresectable cases.
A total of 9 biopsy specimens were obtained from 8 cutaneous tumors of these 7 patients. The eight tumors located on variable regions from head to lower extremities (head and neck: n = 3, 37.5%; lower extremities: n = 3, 37.5%; trunk: n = 2, 25%). In the skin biopsy specimens, aggregates of medium to large histiocytic cells with pale staining cytoplasm, large nuclei, and prominent nucleoli were seen in a lymphoid-rich background with occasional emperipolesis (Figure 1a,b). These histiocytic cells, the so-called RD cells, were immunohistochemically positive for S100 and negative for CD1a (Figure 1c,d). A total of 9 specimens from 7 patients were analyzed with BRAF, KRAS, and NRAS gene mutation by highly sensitive PCR with Sanger sequencing. Mutations were detected in 5 specimens from 4 patients (4/7; 51.7%), and the genetic alterations were all detected in the NRAS gene. Among the 4 patients, NRAS A146T was the most common mutant point detected in 4 patients, followed by NRAS G13S detected in 2 patients. Three patients had more than one focus of genetic mutation. Table 1 summarized patients' clinical characteristics and molecular results. Interestingly, the two biopsy specimens A and B from the same cutaneous tumor of case 6 revealed different genetic alterations, probably caused by tumor heterogeneity (Figure 2a-d).  A total of 9 specimens from 7 patients were analyzed with BRAF, KRAS, and NRAS gene mutation by highly sensitive PCR with Sanger sequencing. Mutations were detected in 5 specimens from 4 patients (4/7; 51.7%), and the genetic alterations were all detected in the NRAS gene. Among the 4 patients, NRAS A146T was the most common mutant point detected in 4 patients, followed by NRAS G13S detected in 2 patients. Three patients had more than one focus of genetic mutation. Table 1 summarized patients' clinical characteristics and molecular results. Interestingly, the two biopsy specimens A and B from the same cutaneous tumor of case 6 revealed different genetic alterations, probably caused by tumor heterogeneity (Figure 2a-d).

Discussion
RDD is a rare, indolent, and self-limiting histiocytic proliferative disorder that commonly arises in young adults in the first two decades of life with a slight male predominance (M/F = 1.4:1) [5,6]. Extranodal involvement had been reported in approximately 40% of cases [7,8]. Of the extranodal involvements, the skin has been reported to be the most common site. Though extranodal involvement is not uncommon, purely extranodal RDD without lymph node involvement is rare. In 1978, Thawerani et al. firstly reported ten RDD patients with cutaneous manifestation and one of them had purely cutaneous involvement. In the ten cases, a wide range of clinical presentations were noted, including xanthomatous or erythematous rash/papules, exanthema with brownish pigmentation, pink-bluish infiltrative plaque, and nodules, etc., that can be solitary, clustered, or dispersed [10]. Later, some studied were done in purely cutaneous RDD, which were found to be distinct from typical RDD [9,13,36]. Nodules, papules, and plaques seem to be the main clinical patterns of cutaneous RDD, which may associate with purple/brown discoloration, erythema, or hyperpigmentation. Besides, it tends to affect older people (mean age: in the fourth decades) with a slight female preponderance (1.3 to 2:1). Moreover, the occurrence of associated systemic symptoms, cervical lymphadenopathy, and hematological and immunological disorders was lower [9][10][11][12][13].
Histopathological features of typical RDD are characterized by proliferation of RD cells, accompanying background infiltration of plasma cells, lymphocytes, eosinophils, or neutrophils. Emperipolesis is another classic finding in RDD, meaning the presence of inflammatory cells, such as lymphocytes or neutrophils, in the cytoplasm of RD cells [5,9,13,37]. The immunophenotype of the RD cells uniquely expresses S100, and CD68, but not CD1a in contrast to LCH. Other positive markers, including CD 163, CD 14, CD163, α1-antichymotrypsin, α1-antitrypsin, fascin, and HAM-56, had also been reported [37][38][39][40]. Compared with typical RDD, purely cutaneous RDD has similar histopathological characteris-tics [41]. In detail, it seems to have less emperipolesis, more frequent plasma cell infiltrates, and a greater degree of stromal fibrosis, which may result in a strikingly storiform growth pattern or a lobulated pattern [9][10][11]13,41]. Histological diagnostic criteria for cutaneous RDD were reported by Chu et al. (Table 2) [11]. A previous study published by Wang et al. demonstrated the different histological stages of purely cutaneous RDD, showing that the early or growing skin lesions presented as hypervascular architectures with mild or free of fibrosis, and relatively more plasma cell infiltrates when compared with old lesions [9]. 46

Discussion
RDD is a rare, indolent, and self-limiting histiocytic proliferative disorder t monly arises in young adults in the first two decades of life with a slight male p nance (M/F = 1.4:1) [5,6]. Extranodal involvement had been reported in approxima of cases [7,8]. Of the extranodal involvements, the skin has been reported to be common site. Though extranodal involvement is not uncommon, purely extrano To date, there is no standard treatment developed for RDD. This disease is thought to be self-limiting, and aggressive intervention is not required. However, some patients may present with refractory or persistent lesions. Surgery, systemic corticosteroids, cryotherapy, radiotherapy, intralesional or topical corticosteroid, interferon, dapsone, acitretin, or methotrexate had been reported for cutaneous RDD [9,[42][43][44][45][46][47][48][49][50]. Surgery seems to be the exclusively effective treatment [43], though recurrence after excision had been reported [9]. Besides, surgical intervention is more suitable for those with a single or smaller lesion, and systemic treatments should be considered in cases with a larger lesion.
The pathophysiology of RDD remains unclear. Viral infection such as herpesvirus-6 and Epstein-Barr virus is supposed to be one of the etiologies resulting in host immune dysregulation [14][15][16]. Germline mutation in SLC29A3 was also reported in some familial cases with controversy [17,18]. In the current literature, investigations for gene mutations in RDD are limited and the results are still controversial. One previous study reported a Biology 2021, 10, 396 7 of 15 patient with autoimmune lymphoproliferative syndrome developing histiocytic sarcoma in a background of RDD with a germline missense mutation in exon 9 of the TNFRSF6 gene encoding Fas, which elucidated a probable Fas/Fas ligand pathway-mediated pathogenesis [51]. Table 2. A histologic summary of cutaneous Rosai-Dorfman disease (modified from reference [11]).

1.
Infiltrates or aggregates of atypical histiocytic cells with abundant cytoplasm with irregular borders, round vesicular nuclei, and small nucleoli. Emperipolesis (intact cells within the cytoplasm and surrounded by small halos) of atypical histiocytic cells may be seen.

2.
Dense infiltrates of lymphocytes and plasma cells, admixed with neutrophils and eosinophils, and stromal fibrosis.
3. Prominent venules infiltrated by plasma cells, more prominent at peripheral areas.

4.
Lymphoid aggregates and germinal center formation, more prominent at peripheral areas.

5.
Lymphatic involvement of atypical histiocytic cells, more frequent at central areas.
The MAPK/ERK pathway, also known as the RAS-RAF-MEK-ERK pathway, regulates many fundamental cellular processes, including cell proliferation, survival, differentiation, apoptosis, motility, and metabolism. A small G protein (RAS) and three protein kinases (RAF, MEK, ERK) are the general structures involved in this pathway [52,53]. Aberrant activation of the pathway is frequently seen in human cancers, including colon cancer, pancreatic cancer, melanoma, papillary thyroid cancer, etc. With the application of recent breakthroughs in molecular pathology, genomic mutations were identified in more and more diseases. Recently, some specific mutations in the MAPK/ERK pathway have been identified in LCH, ECD, and other non-LCH diseases [20][21][22][23][24][25][26][27][28]. These results indicated that the genomic mutation in the MAPK/ERK pathway may play a role in the pathogenesis of these histiocytic disorders. Because of the high association between the MAPK/ERK pathway and histiocytosis, RDD, an entity of histiocytic proliferative disorders, was investigated. Table 3 summarizes the literature review of gene mutations described in RDD [21,[28][29][30][31][32][33][34][35]51,[54][55][56]. Approximately, 33-50% of RDD cases were detected with the genomic mutation involving KRAS, NRAS, ARAF, and MAP2K1 [21, [28][29][30][31][32][33], and only two cases with BRAF mutation were reported [34,35]. With the possible pathomechanism of MAPK/ERK pathway in RDD, the potential risks of malignant associations were concerned as in LCH [57] and non-LCH, such as ECD [58]. In RDD, malignant transformation to histiocytic sarcoma was reported in two cases [51,59], and RDD with concomitant malignant lymphoma was identified in 27 cases [60,61]. However, as mentioned earlier, one of the cases transforming to histiocytic sarcoma had a history of autoimmune lymphoproliferative syndrome with a germline missense mutation in the TNFRSF6 gene, making the pathogenesis of malignant transformation more complicated [51]. The mechanism of RDD coexisting with lymphoma is still unclear, which could be an aberrant histiocytic response to EBV infection. But, the most common type of lymphoma concurrently involving the same lymph node in RDD was nodular lymphocyte-predominant Hodgkin lymphoma, a histological subtype of Hodgkin lymphoma rarely associated with EBV [61]. Another conceivable mechanism is the development of RDD-like changes resulting from the secretion of certain cytokines (such as interleukin) and growth factors by lymphoma cells. However, there is some doubt as to whether it is a morphologic mimic of RDD or truly RDD [61][62][63]. Furthermore, environmental exposures may also contribute to tumorigenesis or act as the second hit to the susceptible individuals. In LCH, for example, more than 50% of cases are identified with BRAF mutation [20], and approximately 30% of BRAF wild-type LCH with MAP2K1 mutation [64], but the prevalence of additional malignancies (32%) is lower than that of genome mutations. This result implied that genome mutation alone may not be enough to cause additional malignancies. Ma J et al. demonstrated that adult LCH patients with additional malignancy of lung cancer had a high percentage of smokers (5/6, 83%) [57], indicating the additional role of environmental exposures in tumorigenesis. Although the relationship between mutations in the MAPK/ERK pathway and potential malignant change or concomitant malignancy of RDD remains unclear, it provides a clue of the possible instability of the genome. Combining the shreds of evidence discussed above, the MAPK/ERK pathway in RDD may direct a novel and promising aspect of treatments in the future. Jacobsen et al. reported a case of RDD with KRAS mutation successfully treated with cobimetinib, a MEK inhibitor, which provided further support of the importance of the MAPK/ERK pathway in RDD [33].
In this study, the median age of our cases was 46 years which is compatible with the previous studies of purely cutaneous RDD. Our case study had a slight male predominance (four men and three women), and this opposite result may be related to the small case number. The clinical lesions were variable but nodules or plaques were the main presentations. A palpable lymph node was not identified in our cases, although small reactive lymphadenopathy over bilateral inguinal areas was seen in case 6. Anemia was noted in two patients, one of whom had a past history of thalassemia. Excluding the underlying disease, 4 out of 7 patients had no accompanying systemic symptoms, lymphadenopathy, and hematological and immunological abnormalities. After follow-up, none of our cases has evidence of malignant transformation.
In our cases, 4 out of 7 patients (57.1%) were detected with genetic mutations and all the mutant regions occurred in the NRAS gene. No mutation was identified in KRAS and BRAF. All 4 cases with NRAS mutations had point mutation at NRAS A146T. NRAS G13S mutation was found in 2 patients, while NRAS G12D, NRAS G12N, and NRAS A59V were observed once in different specimens. Interestingly, different genetic mutations were found in case 6, who had two specimens (specimen A with NRAS G12D, and specimen B with NRAS G12N and NRAS A146T). Referring to her clinical presentation, the two specimens were taken from different components of the cutaneous tumor. The different gene mutations in the same patient may be related to the heterogeneity of tumor cells. There was no significant difference in clinical presentation in cases with or without NRAS mutation. Comparing with previous studies, we could find that NRAS mutation was not frequently identified in patients with RDD. Only Diamond et al. [28] reported one patient with NRAS mutation, but there was no further information on the involved area, genomic coordinates, and variants. Furthermore, most of the mutant reported cases were not presented with cutaneous lesions. Therefore, the role of NRAS mutation in RDD, or specifically cutaneous RDD remains unknown. The NRAS protein encoded by the NRAS gene is a GTPase, acting as a switch for pathway activation and playing a crucial role in cell proliferation and survival. When it gets mutated, it can act as an oncogene, which can be seen in some malignancies, such as melanomas and some hematopoietic malignancies [65]. Congenital melanocytic nevi were also found to frequently harbor NRAS mutations [66]. Some studies mentioned that NRAS mutation seemed to be a poor prognosis in leptomeningeal melanosis and melanomas because of the activation of both ERK and the phosphatidylinositol 3-kinase (PI3K)/AKT pathway, in contrast to ERK signaling alone in BRAF mutation cases [67,68]. Whereas, in metastatic colorectal cancer, BRAF mutation seems to have the poorest prognosis compared with KRAS and NRAS [69]. Of note, most of the reported NRAS mutations are located at codons 12, 13, and/or 61, not NRAS A146T [70]. Besides, Garces et al. found that in RDD, immunostain for phospho-ERK overexpression was seen in all MAP2K1-mutated cases of RDD, supporting the activation of MAPK/ERK pathway, but not in KRAS mutated cases. These results may indicate that there may be some differences in the downstream cascade activation or even an alternative pathway in KRAS-mutated RDD [30]. Is there also an alternative pathway in NRAS-mutated RDD? Does this alternative mechanism make any difference in the clinical presentation and prognosis or let tumor cells be more confined to the skin? These questions remain unclear.
Compared with other mutations in MAPK/ERK pathway, NRAS mutation is less frequent [69,71], either in histiocytic disorders or other malignancies. However, the percentage of NRAS mutation is considerably high (57.1%) in our study. All the cases in our study were Taiwanese people. Can NRAS mutation be an ethnic mutation or a common single nucleotide polymorphism in Asians? Unfortunately, there is no population-based genomic study of NRAS mutation in the general Asian population or other ethnicities so far. However, in malignancies with frequent mutations in the MAPK/ERK pathway, such as melanomas and colon cancers, the percentage of NRAS and KRAS mutation in Asians is not higher than those in Caucasians. In fact, NRAS mutation is even higher in Caucasian patients with melanomas. In addition, our case number is too small to demonstrate the possible ethnic effect. Besides, the highly sensitive PCR were all performed by a Taiwanese company, so machine bias cannot be totally excluded, but which is less likely because there is no NRAS mutation detected in 3 cases [72][73][74][75]. Owing to the limited data, further studies and more cases are needed for the impact of NRAS mutation in tumorigenesis and prognosis.
To our knowledge, this is the first study to demonstrate gene mutations in cutaneous RDD. NRAS mutation was frequent in cutaneous RDD, especially at NRAS A146T. Our results indicate that NRAS mutation in MAPK/ERK pathway may be involved in the pathogenesis of RDD, especially in cutaneous RDD. To sum up our findings, MAPK/ERK pathway seems to be an important pathogenic pathway in a part of the patients with RDD. Our hypothetical thesis of molecular pathogenesis for cutaneous RDD is shown in Figure 3.
There are several limitations in the present study. First, this study is performed retrospectively and the DNA quality of paraffin blocks might not be as good as fresh tissues. Second, the case number is small, and the findings needs to be confirmed in large case series. Third, all the cases in this study are Asians. Whether these results can be applied in other populations such as Caucasian remains uncertain. Future studies are needed. Fourth, the follow-up period is short that most of the patients lost follow-up several months after the diagnosis. Our hypothetical thesis of molecular pathogenesis of cutaneous RDD. In MAPK/ERK pathway, a signaling cascade is initiated when ligands bind to two receptor tyrosine kinases (RTK) subunits and induce receptor dimerization. Adaptor proteins such as growth factor receptor binding protein 2 (GRB2) result in the recruitment of guanine nucleotide exchange factors (GEFs) such as Son of Sevenless (SOS) protein to the cell membrane. SOS then activates RAS by facilitating the small G proteins, including RAS subfamily GTPases such as HRAS, KRAS, and NRAS, to dissociate from GDP and bind to GTP. GPT-bound RAS then activates RAF, a subfamily of MAPKKK, which phosphorylates MEK (MAPKK). Phosphorylated MEK in turn activates ERK (MAPK). Activated ERK enters the nucleus, where it phosphorylates numerous nuclear transcription factors and modulates gene expression, which regulates many cellular processes, including cell proliferation, survival, differentiation, apoptosis, motility, and metabolism. When NRAS gene gets mutated, the signal pathway will lose its normal regulation, which activates the downstream cascade resulting in the proliferation of distinctive histiocytic cells in cutaneous RDD.

Conclusions
In conclusion, our study supports that cutaneous RDD is a distinct group apart from typical RDD. Cutaneous RDD tends to affect older people, and the histopathological features showed less emperipolesis, more dominant plasma cell infiltrate, and relatively more stromal fibrosis. The etiology of RDD remains unclear, but our molecular results provide an implication that MAPK/ERK pathway may play an important role in the pathogenesis of cutaneous RDD, especially NRAS mutation. NRAS A146T was the most frequent genomic mutation in our cases. Patients with refractory cutaneous RDD having mutation of MAPK/ERK pathway may benefit from targeted treatment. Further large studies with a greater sample size of patients are required to ascertain the etiology and pathogenesis of RDD or cutaneous RDD.  Informed Consent Statement: Informed consent was waived because of the retrospective nature of the study and the analysis used anonymous data.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy.

Conflicts of Interest:
The authors declare no conflict of interest.