Determining the immune environment of cutaneous T-cell lymphoma lesions through the assessment of lesional blood drops

Detailed analysis of the cells that infiltrate lesional skin cannot be performed in skin biopsy specimens using immunohistochemistry or cell separation techniques because enzyme treatments applied during the isolation step can destroy small amounts of protein and minor cell populations in the biopsy specimen. Here, we describe a method for isolating T cells from drops of whole blood obtained from lesions during skin biopsy in patients with cutaneous T-cell lymphoma. Lesional blood is assumed to contain lesional resident cells, cells from capillary vessels, and blood overflowing from capillary vessels into the lesion area. The lesional blood showed substantial increases in distinct cell populations, chemokines, and the expression of various genes. The proportion of CD8+CD45RO+ T cells in the lesional blood negatively correlated with the modified severity-weighted assessment tool scores. CD4+CD45RO+ T cells in the lesional blood expressed genes associated with the development of cancer and progression of cutaneous T-cell lymphoma. In addition, CD8+CD45RO+ T cells in lesional blood had unique T-cell receptor repertoires in lesions of each stage. Assessment of lesional blood drops might provide new insight into the pathogenesis of mycosis fungoides and facilitate evaluation of the treatment efficacy for mycosis fungoides as well as other skin inflammatory diseases.

The tissue environment surrounding skin lesions has an important role in skin disease. Harvesting cells from skin lesions can be time-consuming and challenging, however, due to the considerable cell and protein loss caused by tissue degradation. Various techniques are thus used to analyze skin lesion cells and the surrounding environment, including multiphoton excitation microscopy 1 , dermal open-flow microperfusion 2 , and immersion of skin samples in medium to extract cells 3 . The cell isolation processes required for these techniques, however, may lead to the loss of critical information. For example, while lymphocytes can be isolated from skin tissue obtained by punch biopsy, the small amount of tissue contains too few cells and thus provides limited information.
Alternatively, lesional blood samples could provide valuable information about the surrounding environment, including the levels and types of cytokines and inflammatory cells, without the need for enzyme treatment. In fact, a previous study reported the successful use of sera from peripheral blood and blood obtained from psoriasis lesions to assess the skin lesion environment 4 . Lesional blood samples might therefore be useful for isolating and analyzing lesional cellular components and serum.
Skin biopsies are regularly obtained to diagnose and assess treatment efficacy in patients with cutaneous T-cell lymphoma (CTCL). Diagnosis of CTCL is relatively difficult 5 , however, and effective treatments are not yet clearly established. Better methods that allow for rapid isolation and analysis of resident and systemic pathogenic T cells and effector T cells are necessary to facilitate diagnosis and develop effective treatments. Mycosis fungoides (MF), the most common CTCL, is considered to be a low-grade T-cell lymphoma 6 . The premycotic and mycotic phases can last several years, but in some cases the disease progresses very rapidly 7 . Due to the relatively Results CD4 + /CD8 + T cells are successfully isolated from a small amount of lesional blood. We obtained 200 to 300 µL of lesional blood from the wound site resulting from the skin biopsy. The cells were separated from the sample using a cell sorter. Approximately 3000 CD4 + T cells and 1000 CD8 + T cells were successfully isolated from 5 μL of peripheral blood (Fig. 1a). Although the lesional blood contained slightly fewer cells than the peripheral blood, we were able to collect a sufficient number of cells from lesional blood for the analyses (Fig. 1b,c). Cytometry by time-of-flight (CyTOF) revealed that the lesional blood contained more granulocytes, and fewer monocytes and B cells than the peripheral blood (Fig. 1d,e). These findings indicate that the cell populations in lesional blood might differ from that in peripheral blood. Isolated cells and sera from lesional blood were therefore further analyzed.
CD8 + CD45RO + T-cells in lesional blood negatively correlate with the mSWAT score. We obtained 14 biopsy specimens from the lesional skin of MF patients for immunohistochemical analysis (Table 1). All MF patients were assessed using the modified severity-weighted assessment tool (mSWAT), and skin biopsies were obtained upon admission (Fig. 2a). Although in the mass cytometry analysis of the initial 4 samples, the proportions of CD4 + and CD8 + T cells did not differ significantly between lesional blood and peripheral blood (Fig. 1e), flow cytometry analysis of the 14 additional samples using a paired t-test revealed that lesional blood contained a significantly greater proportion of CD4 + CD45RO + and CD8 + CD45RO + T cells (Fig. 2b,c). Furthermore, the proportion of CD8 + CD45RO + T cells in the lesional and peripheral blood negatively correlated with the mSWAT score (Fig. 2e). The proportion of CD4 + CD45RO + T cells weakly inversely correlated with the mSWAT score (Fig. 2d). Thus, assessment of lesional blood drops might reveal the phenotypic details of infiltrating cells.
CD8 + CD45RO + T-cells infiltrate MF lesions and negatively correlate with CTCL pathogenesis. The CD8 + tumor-infiltrating lymphocyte levels in patients with MF positively correlate with an improved survival rate and exert an antitumor effect 9 . Therefore, increased levels of CD8 + T cells in CTCL serve as a promising criterion for predicting patient survival and supporting treatment decisions, and inclusion of patients in randomized controlled trials 10 . Moreover, the partial activation of CD8 + cytotoxic T cells present in CTCL and their positive correlation with a better prognosis suggest that they have an important role in the antitumor response 11 . Tissue specimens showed a negative correlation between the mSWAT score and CD8 + CD45RO + T cells with less infiltration of effector CD8 + T cells in advanced cases (Fig. 2h,i). In contrast, CD4 + CD45RO + T cells in tissue specimens did not correlate with mSWAT scores (Fig. 2f,g), consistent with the results from the lesional blood. These infiltrating cells in tissue specimens, however, cannot be easily assessed by immunohistochemistry alone.
Chemokine profiles differ between lesional blood and peripheral blood. In sera simultaneously isolated from peripheral and lesional blood samples, the levels of chemokines such as CCL5, CCL11, CCL17, CCL22, and CXCL11 were significantly increased in the lesional blood compared with the peripheral blood (Fig. 2j). The increases in these chemokines are specific to MF lesions. CCL5, CCL17, and CCL22 are derived from keratinocytes; CCL22 and CXCL11 are derived from endothelial cells; and CCL11 is derived from macrophages in MF lesions 12 . In particular, levels of the CCR4 ligands CCL17 and CCL22 are upregulated in the epidermis and serum of patients with MF [13][14][15] . Malignant T cells expressing CCR4 are recruited by CCL17 and CCL22 14 . We found that sera from lesional blood revealed a specific chemokine environment for MF. The results demonstrated that the chemokine profile of lesional blood differs from that of peripheral blood, indicating that lesional blood drops are sufficient for obtaining a detailed chemokine profile of the lesion environment.
CD4 + CD45RO + T cells and CD8 + CD45RO + T cells from lesional and peripheral blood differ in RNA sequence and T-cell receptor repertoire analyses. We isolated CD4 + CD45RO + and CD8 + CD45RO + T cells from lesional and peripheral blood (n = 3) using a FACS Melody sorter (Becton Dickinson). RNA sequenc-  www.nature.com/scientificreports/ ing (RNA-seq) of the isolated cells was performed to analyze the transcriptome. CD4 + CD45RO + T cells in the lesional blood highly expressed genes relating to cancer progression and CTCL pathogenesis: RGS1, RDH10, HES1, DNAH9, ANK2, and SGK1 ( Fig. 3a) [16][17][18][19][20][21][22][23][24][25] . To further confirm the increases in the 51 highly expressed genes, the genes were enriched in the Jensen Disease library of Enrichr. Enrichr showed that the diseases were related to cancer, including "skin cancer" and "lymphoid leukemia" (Fig. 3b). All differentially expressed genes are listed in Table S1. Although we performed T-cell receptor (TCR) repertoire analysis only for 1 case, CD4 + CD45RO + T cells in the lesional blood exhibited a unique TCR repertoire, showing reduced diversity compared with the TCR repertoire in peripheral blood (Fig. 3c,d). Specific polymorphisms in TP53 and STAT3 are associated with CTCL malignancy 26 . In Cases 16 and 17, the frequency of TP53 polymorphism c.C98G (rs1042522) in CD4 + CD45RO + T cells was higher in the lesional blood than in the peripheral blood (Table S2). No TP53 polymorphism was detected in Case 12, and STAT3 mutations c.A1936T and c.A1480T were detected only in the lesional blood.
Stage progression and skewed TCR repertoires in CD8 + CD45RO + T cells. Biopsy specimens and lesional blood were collected from lesions at each stage (erythema, plaque, and tumor) from the same patient. A certain number of CD8 + CD45RO + cells was found in the erythema areas, but few were found in the plaque and tumor tissues in Case 10 (Fig. 4a,c,e; immunofluorescence staining). We isolated CD8 + CD45RO + T cells from the lesional blood and peripheral blood separately. The isolated cells were subjected to TCR repertoire analysis. The CD8 + CD45RO + T cells in the lesional blood showed unique TCR repertoires in lesions of each stage (Fig. 4b,d,f,g). We assume that CD8 + CD45RO + cells in lesional blood would recognize a tumor antigen in tumor and plaque tissue. A previous report indicated that malignant T-cell clones exhibit heterogeneity between skin lesions in the same patient 38 . Our results are consistent with previous findings that neoplastic T-cell clones vary in skin lesions. Furthermore, different TCR repertoires were present in tumor-stage lesions of the same patient (Fig. 4f), which may result from the generation of different neoplastic T-cell clones for the growth of tumor lesions. In another patient (Case 15), a biopsy was performed from an adjacent lesion that developed erythema and plaque. CD8 + CD45RO + cells were found in both the erythema and plaque areas (Fig. 4h,j), and repertoire   www.nature.com/scientificreports/ analysis showed that the same TCRs (TRAV1-2-TRAJ33) were increased (Fig. 4i,k,l). These findings indicate that CD8 + T cells with different TCR repertoires are directed toward each skin lesion, which is expected given the heterogeneity of malignant T cells in skin lesions.

Discussion
The present study revealed that the composition of lesional blood differs from that of peripheral blood, and lesional blood appears to reflect the condition of the skin's immune environment in lesional areas of MF patients.
Many techniques are used to analyze the cells and tissue environments of skin lesions, and assessment of lesional blood drops is a feasible method. Although assessment of lesional blood drops would not obviate the need for skin biopsy due to the different results, analysis of lesional blood could provide valuable information about infiltrating cells and the separated serum. This technique can also be used to maximize the retrieval of intact cells and blood when skin biopsy is performed or a tumor is resected. We assume that lesional blood would contain lesional resident cells, cells from capillary vessels, and blood overflowing from capillary vessels, allowing us to identify a unique pattern of cell populations and chemokines in the lesional blood.
Mass cytometry analysis showed differences in the cell populations between lesional and peripheral blood, and the chemokine assay revealed increases in specific chemokines in lesional blood. The different types of cells in the lesional blood may be due to the induction of chemokines. Memory T cells, CD45RO + T cells, were frequently observed in lesional blood, suggesting that these cells are involved in the lesion. Chemokines that are associated with MF were particularly abundant in lesional blood. CCL17 and CCL22 are ligands for CCR4, and these chemokines are strongly associated with the pathogenesis of MF 13,14 . Although we have not yet collected lesional blood from healthy areas, we expect to clarify the detailed pathogenesis of MF by comparing lesional blood from healthy and lesional areas in the same patient.
RNA-seq of CD4 + CD45RO + T cells and CD8 + CD45RO + T cells suggested that lesional blood contains malignant T cells and effector T cells, respectively. This notion was supported by the skewed TCR repertoire in lesional blood (Fig. 3c,g). In the present study, CD4 + CD45RO + T cells from the lesional blood highly expressed genes relating to cancer progression and CTCL progression. Among the differentially expressed genes in this study, expression of SGK1, HES1, and RDH10, which were identified in other studies conducting single cell RNA-seq (scRNA-seq) of CTCL skin tumors 24,25 , was increased. Although bulk RNA-seq was performed in this study, these genes were also identified in CD4 + CD45RO + T cells in lesional blood (Fig. 3a). It may thus be possible to isolate lymphocytes recruited by chemokines or present in the lesion area from lesional blood. A larger number of differentially expressed genes was detected in the CD4 + CD45RO + cells of peripheral blood because cells in the lesional blood vary between lesions and stages. In addition, there are different types of malignant cells 39 , which may be why fewer differentially expressed genes were detected in the CD4 + CD45RO + cells of lesional blood (Fig. 3a). CD8 + CD45RO + T cells in the lesional blood had a less inflammatory transcriptome than CD8 + CD45RO + T cells in the peripheral blood. Effector T cells express exhausted phenotypes characterized by the expression of the PD-1, ICOS, TIM-3, LAG-3, and CTLA-4 markers in lesional skin 33 . Although expression of these genes was not increased, weak expression of genes related to the inflammatory system and cell division suggest that CD8 + CD45RO + T cells in the lesional blood are a type of exhausted T cell 40 .
The clones of malignant cells are different in each skin lesion in patients with MF 41 . In the present study, we collected CD8 + CD45RO + cells from lesional blood and performed TCR repertoire analysis for each skin lesion. In Case 10, we performed a skin biopsy from 3 isolated skin lesions (i.e., erythema, plaque, and tumor) (Fig. 4a-g). Few CD8 + CD45RO + cells, however, were detected in the plaque and tumor tissues in Case 10 (Fig. 4c,e). A characteristic repertoire pattern of CD8 + CD45RO + cells was detected in the different skin lesions of Case 10 ( Fig. 4b,d,f). Case 15 showed a more skewed repertoire of the plaque tissue compared with an area of erythema from the same skin lesion (Fig. 4i,k). Although the lesional blood would not be a perfect representation of the cells in this lesion, we could indirectly show that different skin lesions have different clones of malignant cells. In other words, the TCR repertoire of CD8 + CD45RO + cells to malignant T cell antigens differed dependent on the skin lesion. Further studies with an increased number of cases are needed to investigate the details.
We performed the same procedure in psoriasis patients, and lesional blood in psoriasis exhibited a different pattern than lesional blood from MF (Fig. S1). The percentage of CD4 + T cells was higher in the lesional blood than in the peripheral blood (Fig. S1b). The CD4 + T cells of the lesional blood had a unique transcriptome, including IL36B, FABP7, NLRC4, and FOS. These genes lead inflammation and are detected in psoriatic lesions 42,43 . Interleukin (IL)-36 cytokines are a subgroup of the IL-1 cytokine family. IL-36B is highly detected in psoriatic lesions 44 . Chemokine receptors are G-protein coupled and T cells in lesional blood are induced by chemokines, therefore the top GO term is considered to "positively regulate the G protein coupled receptor signaling pathway". The pattern of chemokine increases in psoriasis patients differed from that in MF patients (Fig. S1e). The pattern of chemokine increases in blood from psoriatic lesions are produced mainly by keratinocytes 45 . These chemokines are related to T cell and neutrophil trafficking 46,47 , and are specific to the pathogenesis of psoriasis. Therefore, we demonstrated that lesional blood obtained from psoriatic lesions can also reveal the psoriatic lesion environment.  www.nature.com/scientificreports/ In conclusion, we successfully developed a method for assessing the immune environment of CTCL from small amounts (drops) of lesional blood. Lesional blood drops contained sufficient numbers of isolated cells and amounts of sera for the assessment. The cells and sera isolated from lesional blood can also be used for further gene expression analysis of target cells. Our findings provide new insight into the pathogenesis of MF, a rare cutaneous lymphoma. The technique described in this study could be further applied to evaluate other skin inflammatory diseases, such as atopic dermatitis and psoriasis, as well as to assess the efficacy of treatments for these diseases.

Methods
Patients. We recruited 19 patients with MF (mean age: 67.10 ± 13.91 years; 9 women, 10 men) from the Department of Dermatology at the Nagoya City University. Exclusion criteria were: 1) age under 20 years, 2) HTLV-1 positive status, and 3) pregnant. The institutional review board of the Nagoya City University Graduate School of Medical Sciences approved the study (approval number: #60-18-0101). Written informed consent was obtained from the patients. All the experimental protocols adhered to relevant ethical guidelines for involving humans. The patients' profiles are described in Table 1. Samples were obtained from the first 14 patients who visited our department from December 2018 to December 2019 and used for tissue staining, flow cytometry analyses, and chemokine assays. Some samples were subjected to RNA-seq (Cases 4 and 12) and TCR repertoire analyses (Cases 4, 10, 11, and 14). Samples obtained from Cases 15 through 19 were used for RNA-seq or TCR repertoire analyses (Table 1).
Lesional blood collection. Skin biopsies were regularly obtained from the patients for diagnosis and assessment of treatment efficacy. After administering local anesthesia using lidocaine without epinephrine to the skin lesion, a punch biopsy was performed at a depth of 1 to 3 mm to avoid reaching the fatty layer. Oozing blood from the wounded area was collected as quickly as possible into an Eppendorf tube containing anticoagulant to avoid clotting. As anticoagulants, we used 5 μL of 100 U mL −1 heparin sodium (TERUMO) for flow cytometry analysis, and 3 μL of 0.5 mol L −1 EDTA (Invitrogen) for the RNA-seq and TCR repertoire analyses. Approximately 200 to 300 μL of blood was collected from each lesion area using a P20 Pipetman (Gilson). We also collected 30 to 50 μL of lesional blood without anticoagulant to obtain serum. After collecting the lesional blood, we performed another punch biopsy in the same lesion area, again at a sufficient depth to obtain a skin sample. Serum was collected and stored at − 80 °C until analysis. We collected peripheral blood from the patient's arm and treated the peripheral blood in the same manner as the lesional blood.
Mass cytometric immunoassay. Blood cells were stained for mass cytometry after hemolysis using a Maxpar Human Peripheral Blood Phenotyping Panel kit (Fluidigm). Hemolyzed peripheral blood and lesional blood samples were resuspended in 1 mL phosphate-buffered saline and incubated for 5 min at room temperature with 1 mL of Cisplatin-108Pt (Fluidigm). The cells were washed using Maxpar Cell Staining Buffer (Fluidigm) and centrifuged, and the supernatant discarded; the pellets were then resuspended in 50 μL of the same buffer and 50 μL of a prepared cocktail of titrated Maxpar metal-conjugated antibodies was added (Fluidigm). After incubating for 15 min at room temperature, the cells were washed twice and fixed with 2% paraformaldehyde. The stained cells were analyzed by St. Luke's MBL Corp using cytometry by time-of-flight. Mass cytometry data were analyzed using Cytobank (https:// www. cytob ank. org/). Cell sorting. CD4 + CD45RO + and CD8 + CD45RO + T cells were sorted using the BD FACSMelody Cell Sorter (BD Bioscience). The sorted T cells were collected for the TCR repertoire and RNA-seq analyses as described below.
RNA-seq analysis. The CD4 + CD45RO + and CD8 + CD45RO + T cells were prepared as described above. The T cells were lysed with TRIzol reagent (Thermo Fisher Scientific) and stored at − 80 °C. The lysates were sent to Genewiz Japan Corp for RNA-seq and related analyses. In brief, RNA was extracted with chloroform/isopropanol and recovered from the supernatants using RNA Clean and Concentrator-5 columns (ZymoResearch) following the manufacturer's instructions. The RNA purity was assessed with an Agilent 2100 Bioanalyzer. The RNA was subjected to library preparation with the TaKaRa  www.nature.com/scientificreports/ expressed genes were counted using the DESeq2 package in R (version 3.6.3). Up-and downregulated genes were defined as those (i) differentially expressed in peripheral and lesional blood cells with a P-value < 0.05, and (ii) having a greater than twofold change in the average normalized number of peripheral and lesional blood cells. Gene ontology analysis and enrichment analysis using the Jensen DISEASES dataset of differentially expressed genes was performed using the Enrichr webtool (https:// maaya nlab. cloud/ Enric hr/) and Metascape (https:// metas cape. org/) [48][49][50] .
The T cells were lysed with Isogen-LS (NIPPON GENE) and stored at − 80 °C. The lysates were sent to Repertoire Genesis Inc. (Ibaraki, Japan) for next-generation sequencing, which was performed as previously described 51 . Briefly, total RNA was converted to complementary DNA (cDNA) with the SuperScript reverse transcriptase (Invitrogen). Double-stranded cDNA was synthesized and ligated with a 5′ adaptor oligonucleotide, then cut with the SphI restriction enzyme. Next, the double stranded cDNA was amplified through polymerase chain reaction using primers specific for the adaptor and TCRα constant region. The sequencing was performed with the Illumina MiSeq paired-end platform (2 × 300 bp). Data processing was performed with the repertoire analysis software developed by Repertoire Genesis Inc 52 . TCR sequences were assigned with a dataset of reference sequences from the international ImMunoGeneTics information system database (http:// www. imgt. org). The percentage of sequence reads with TRAV, TRAJ, TRBV, and TRBJ genes was calculated. The Circos plots were produced using the Circos software package 53 . The Shannon-Weaver index shows the diversity and is defined as follows.
Statistics. Statistical analyses were performed using GraphPad Prism 7. All numerical data are summarized using mean ± standard deviation. Paired or unpaired Student t-tests were used to determine the significance of differences between groups, unless otherwise indicated in the figure legend. P-values < 0.05 were considered statistically significant.

Data availability
All RNA-seq data sets have been deposited in the DNA Data Bank of Japan under accession numbers DRA010717 (http:// www. ddbj. nig. ac. jp/ intro-e. html). All data are available from the corresponding author upon reasonable request.