Involvement of Macrophages in the Pathogenesis of Familial Amyloid Polyneuropathy and Efficacy of Human iPS Cell-Derived Macrophages in Its Treatment

We hypothesized that tissue-resident macrophages in familial amyloid polyneuropathy (FAP) patients will exhibit qualitative or quantitative abnormalities, that may accelerate transthyretin (TTR)-derived amyloid deposition. To evaluate this, we examined the number and subset of tissue-resident macrophages in heart tissue from amyloid-deposited FAP and control patients. In both FAP and control patients, tissue-resident macrophages in heart tissue were all Iba+/CD163+/CD206+ macrophages. However, the number of macrophages was significantly decreased in FAP patients compared with control patients. Furthermore, the proportion of intracellular TTR in CD14+ monocytes was reduced in peripheral blood compared with healthy donors. Based on these results, we next examined degradation and endocytosis of TTR in human induced pluripotent stem (iPS) cell-derived myeloid lineage cells (MLs), which function like macrophages. iPS-MLs express CD163 and CD206, and belong to the inhibitory macrophage category. In addition, iPS-MLs degrade both native and aggregated TTR in a cell-dependent manner in vitro. Further, iPS-MLs endocytose aggregated, and especially polymerized, TTR. These results suggest that decreased tissue-localized macrophages disrupt clearance of TTR-derived amyloid deposits, leading to progression of a pathological condition in FAP patients. To improve this situation, clinical application of pluripotent stem cell-derived MLs may be useful as an approach for FAP therapy.

Introduction study, we examined the phenotype of tissue-resident macrophages in heart tissue from FAP patients and controls. We found that tissue-resident macrophages are CD163 or CD206 positive, with a lower number in FAP patients compared with control patients. In addition, the frequency of TTR uptake in CD14 + monocytes derived from peripheral blood mononuclear cells (PBMC) was decreased in FAP patients compared with healthy donors (HD). Furthermore, we found that iPS-MLs degrade native and aggregated TTR, and endocytose aggregated TTR in vitro. These results suggest that tissue-resident CD163 + or CD206 + macrophages are involved in FAP pathogenesis, and iPS-ML therapy may be a novel and rational approach for the treatment of FAP.

Patients and blood samples
Autopsy specimens were evaluated from 16 non-LT FAP ATTR V30M patients and 11 control patients without heart disease, who were autopsied at Kumamoto University Hospital between 1987 and 2012. The specimens were prepared from formalin-fixed, paraffin-embedded tissue blocks and subsequently stored. Clinical information on these patients is shown in Table 1. For flow-cytometry analysis, PBMC were separated from blood of 15 non-LT ATTR FAP patients (13 V30M, one Y114C, and one I107V) and 15 HD. For TTR staining experiments, CD14 + cells were purified from PBMC of six non-LT ATTR V30M FAP patients and seven HD. These samples were collected at Kumamoto University Hospital between 2012 and 2016. Written informed consent was obtained from HD and patients for participation in this study, and all experiments using human samples were conducted in accordance with The Declaration of Helsinki and the approval of the Institutional Review Board of Kumamoto University (Permit Number: 1087).

Generation of aggregated TTR
Recombinant human wild-type and V30M (mutated) TTR were purchased from Wako (Osaka, Japan). To generate aggregated wild-type or mutated TTR, each TTR was incubated at pH 4.0 for 24 h [33].

Congo red and immunohistochemistry staining
Formalin-fixed heart and kidney samples were serially sectioned at a thickness of 4 μm and adhered to microscope slides. Slides were stained with haematoxylin-eosin (HE) and alkaline Congo red. For immunohistochemistry staining, tissue samples were deparaffinized and treated with methanol containing 0.3% H 2 O 2 for 30 min to inactivate endogenous peroxidase [34]. Normal goat serum (Dako, Glostrup, Denmark) or Blocking-One (Nacalai Tesque, Kyoto, Japan) was used to block nonspecific background staining. Monoclonal mouse antihuman CD68 antibody (Clone: G-M1; Dako), monoclonal mouse anti-human CD163 antibody (Clone: 10D6; Novocastra, Newcastle upon Tyne, UK), monoclonal mouse anti-human CD206 antibody (Clone: 5C11; Acris Antibodies, Herford, Germany), polyclonal rabbit anti-human Iba1 antibody (Wako), and polyclonal goat anti-human CD3-ε antibody (Clone: M-20; Santa Cruz Biotechnology, CA, USA) were used as primary antibodies. The secondary antibodies were polyclonal goat anti-rabbit (Dako), polyclonal goat anti-mouse (Dako), and polyclonal rabbit anti-goat antibody conjugated with horseradish peroxidase (HRP) (Dako). Reactions were visualized using the DAB Liquid System (Dako). Five visual fields from each stained section were randomly chosen and the number of each positive cell type counted by two observers. The average count number per stained section was calculated. For double-immunohistochemical staining, a mixture of anti-human Iba1 antibody and either anti-human CD163 or CD206 antibody was used as the primary antibody solution. Simple Stain Mouse MAX-PO (Nichirei, Tokyo, Japan) was reacted and visualized by DAB, and then Simple Stain RAT MAX-PO was reacted and visualized by Fast Blue BB (Sigma-Aldrich, St Louis, MO, USA).
Separation of CD14 + monocytes from peripheral mononuclear blood cells PBMC were isolated from human blood using Ficoll-Paque (GE Healthcare, Buckinghamshire, UK). PBMC were incubated with FcR blocking regent (Miltenyi Biotec, Bergish Gladbach, Germany) on ice for 10 min, and then CD14 + monocytes purified using the MACS cell sorting system (Miltenyi Biotec).

Enzyme-linked immunosorbent assay
After culturing iPS-MLs or SH-SY5Y cells with the appropriate TTRs for 3 days, TTR concentration in each culture supernatant was measured by ELISA. Briefly, 96-well plates were coated overnight at 4°C with supernatants collected in carbonate buffer. The next day, plates were blocked with coating buffer containing 0.5% gelatin for 1 h at room temperature. To detect TTR, polyclonal rabbit anti-human TTR antibody (Dako; diluted 1:1000) was used as the primary antibody, and polyclonal goat anti-rabbit IgG antibody conjugated with HRP (Dako; diluted 1:5000) as the secondary antibody, which were sequentially incubated for 1 h at room temperature. After incubation with SureBlue™ TMB Microwell Peroxidase Substrate (KPL, Gaithersburg, MD, USA), absorbance was detected at 450 nm.

Western blotting analysis
After iPS-MLs or SH-SY5Y cells were cultured with the appropriate TTRs for 3 days, culture supernatants were mixed with sample buffer (Bio-Rad Laboratories, Hercules, CA, USA) containing 2-mercoptoetanol, and boiled at 95°C for 5 min. Resultant samples were separated by SDS-PAGE and transferred to PVDF membranes (Bio-Rad). Polyclonal rabbit anti-human TTR antibody (Dako; diluted 1:1000) was used as the primary antibody, and polyclonal goat anti-rabbit antibody conjugated with HRP (Dako; diluted 1:5000) as the secondary antibody. The reaction was visualized using ECL prime reagents (GE Healthcare, WI, USA), and detected using a LAS-4000EPUVmini (GE Healthcare).

TTR staining in CD14 + monocytes and iPS-MLs
To stain CD14 + monocytes, iPS-MLs and SH-SY5Y cells were cultured with the abovedescribed TTRs, and re-suspended at a density of 2 × 10 5 cells/mL in DMEM medium (Gibco).
Next, cells were adhered to slide glass using Cytospin, and fixed in 4% formalin phosphate buffer. After proteolytic activation for 10 min by proteinase K (Dako), samples were treated with methanol containing 0.3% H 2 O 2 for 20 min to remove endogenous peroxidase activity. PBS containing 5% skimmed milk was used to block nonspecific background staining. Polyclonal rabbit anti-human TTR antibody (Dako) was diluted (1:100) in Dako REAL Antibody Diluent (Dako), and used as the primary antibody. Envision System-HRP Labelled Polymer Anti-Rabbit (Dako) was used as the secondary antibody. The reaction was visualized using the Dako Envision + kit/HRP (DAB) (Dako). TTR + cells in CD14 + monocytes, iPS-MLs and SH-SY5Y cells were counted. Briefly, five visual fields were randomly chosen from each stained section, and the TTR + cell number in each visual field determined. The average number of TTR + cells per visual field was calculated.

MTS assay
After culturing iPS-MLs (1× 10 5 cells/well) with the appropriate TTRs for 3 days, cell viability was measured by MTS assay. Briefly, a Cell Titer 96 Aqueous One Solution Cell Proliferation Assay kit (Promega, Madison, WI, USA) was used in each well of the cultured plate. After incubation for 1 h at 37°C, absorbance in each well was measured at 490 nm.

Statistical analysis
For the ELISA assay, the Tukey post-hoc test after a separate two-way repeated-measures analysis of variance (ANOVA) was used. The pairwise t-test with Bonferroni correction after oneway ANOVA was used for the MTS assay. This analysis was performed using R version 3.2.1. For immunohistochemistry staining of heart tissue, repeated count immunostained cells were analyzed using the generalized Poisson mixed model. The proportion of repeated count TTR + iPS-MLs, SH-SY5Y cells and CD14 + monocytes were analyzed using the mixed model. These analyses were performed using SAS Version 9.4. A value of p < 0.05 was considered statistically significant.

Histopathological characteristics of FAP ATTR V30M patients
The characteristics of FAP patients employed in this study are demonstrated in Table 1. To investigate the condition of macrophages in FAP, we analyzed the number of tissue-resident macrophages in the heart, which is one of the most TTR-derived amyloid fibril-laden organs. Moreover, inflammation causes recruitment of inflammatory cells, including macrophages, and affects the number and polarity of endogenous tissue-resident macrophages, although this process rarely occurs in the heart [35]. By performing HE and anti-CD3 staining, we first found that both control-and FAP-derived heart tissue do not contain migrating inflammatory cells such as T cells (Fig 1A-1C and 1J-1L, and S1 Fig). Next, heart tissue from control and FAP patients was stained with Congo red, as Congo red polarization confirms amyloid deposition. Although there was no amyloid deposition in control patients, mild or severe amyloid deposition was observed in heart tissue from all FAP patients (Fig 1D-1I and 1M-1R). Additionally, tissue destruction and myocardial cell death were observed, coincident with areas of severe amyloid deposition (data not shown).
Decreased tissue-resident macrophages in heart tissue from FAP patients To detect heart-tissue resident macrophages, we performed immunohistochemistry using Iba and CD68 antibodies, which are both well-known macrophage markers [36,37]. The number of Iba1-and CD68-positive macrophages in FAP-derived heart tissue was decreased compared with control patients (Fig 2A-2C and S2 Fig). Tissue-resident macrophages are reported to be inhibitory macrophages [24]. Therefore, we investigated the phenotype of tissue-resident macrophages in control and FAP patients. To identify these macrophages, we immunohistochemically stained heart tissue using the inhibitory macrophages markers, CD163 and CD206 (Fig  2A and 2B). The number of both CD163-and CD206-positive cells was decreased in FAP patients compared with controls (Fig 2D and 2E). Additionally, we confirmed that all Iba1-positive cells were identical to CD163-or CD206-positive cells by double immunohistochemical staining (Fig 2F and S2 Fig). Taken together, these results suggest that tissue-resident macrophages are decreased in FAP patients.

Reduced intracellular TTR in CD14 + monocytes from FAP patients
Monocytes are reported to partly differentiate into tissue-resident macrophages [38], therefore, we focused our attention on blood monocytes. CD14 + monocytes are well known to subdivide into CD14 high CD16 -, CD14 high CD16 + , and CD14 low CD16 + monocyte subsets [39]. Thus, we investigated the proportion of these three subsets between HD and FAP patients. There were no significant differences in the proportion of these three subsets or CD163 expression in each subset between both groups (S3 Fig). Next, isolated CD14 + monocytes from HD and FAP patients were applied to cytospin preparations and stained with an anti-TTR antibody. Consequently, CD14 + monocytes from FAP patients were shown to have lower intracellular TTR immunoreactivity compared with HD (Fig 3A and 3B). Overall, these results suggest that CD14 + monocytes may affect TTR clearance, with a decreased ability in FAP patients.

Expression of CD163 and CD206 on iPS-MLs
We observed a lower number of inhibitory macrophages in heart tissue from FAP patients compared with control patients. Further, inhibitory macrophages are reported to have antiinflammatory effects and induce tissue repair [24]. These results suggest that decreased tissueresident inhibitory macrophages in FAP patients may exacerbate TTR-derived amyloid deposition. Consequently, it is possible that inhibitory macrophage transfer may be useful as a therapeutic approach for FAP. In this regards, we have previously generated iPS-MLs that function as macrophage-like myeloid lineage cells, and shown that they exhibit efficient β-amyloid phagocytic activity in a mouse model of Alzheimer's disease [32]. Thus here, we determined if iPS-MLs also have phagocytic capacity for TTR-derived amyloid fibrils. First, we examined the phenotype of iPS-MLs by flow cytometry. After confirming cell morphology and hematopoietic and myeloid markers, such as CD45, CD33, CD14, and CD11b (Fig 4A and 4B), we detected expression of CD163 and CD206 on iPS-MLs (Fig 4B). Our results suggest that iPS-MLs belong to the inhibitory macrophage class.

In vitro degradation of native and aggregated TTR by iPS-MLs
We next determined if iPS-MLs reduce TTR in vitro. To exclude the influence of serum, which naturally contains wild-type TTR, iPS-MLs or SH-SY5Y cells (as controls) were cultured in the presence of native wild-type or mutated TTR under serum-free conditions. After 3 days, TTR in the culture supernatant was quantified by ELISA. Concentration of both native wild-type and mutated TTR decreased in a cell number-dependent manner in iPS-MLs compared with SH-SY5Y cells (Fig 5A and 5C). This tendency was similar in iPS-ML culture in the presence of both wild-type-and mutated-derived aggregated TTR (Fig 5B and 5D).
Additionally, we confirmed degraded TTR in culture supernatants by western blotting. Truncated TTR was detected in iPS-ML culture conditions with all types of TTR (Fig 5E-5H), and possibly reflects TTR degradation owing to iPS-MLs. Under native wild-type and mutated TTR culture conditions, neither iPS-MLs nor SH-SY5Y cells decreased full-length TTR in a cell number-dependent manner (Fig 5E and 5G). In contrast, compared with native TTR culture conditions, wild-type-and mutated-derived aggregated TTR culture conditions resulted in the appearance of polymerized TTR fibrils. Furthermore, iPS-MLs significantly degraded polymerized TTR fibrils and truncated TTR regardless of TTR type. These phenomena were dependent on iPS-ML number (Fig 5F and 5H). These results suggest that iPS-MLs degrade both wildtype and mutated TTR, with a greater degradation capability for polymerized TTR fibrils.

Proliferated cell viability of iPS-MLs in the presence of wild-type TTR
Next, to determine the toxicity of TTR on iPS-MLs, we performed the MTS assay. Viability of iPS-MLs cultured in the presence of native or aggregated wild-type TTR increased compared with those cultured without TTR (S4 Fig).In contrast, mutated and mutated-derived aggregated TTR did not influence iPS-MLs (S4 Fig).

Intracellular uptake of TTR in iPS-MLs
To determine if iPS-MLs phagocytize aggregated TTR, iPS-MLs and SH-SY5Y cells cultured with native or aggregated TTR were stained with an anti-TTR antibody. Although native wild- type and mutated-derived aggregated TTR was slightly internalized by iPS-MLs, intake of mutated TTR was (Fig 6A and 6B). In addition, intracellular TTR was predominantly identified in iPS-MLs cultured with aggregated wild-type TTR compared with other TTR-culture conditions. Conversely, SH-SY5Y cells showed no intracellular TTR immunoreactivity under any culture conditions (Fig 6A and 6C). Taken together, these results indicate that iPS-MLs mainly endocytose wild-type-derived aggregated TTR.

Discussion
To examine tissue-resident macrophages in FAP patients, we chose heart tissue as an organ of TTR deposition for the following two reasons. The first being that heart tissue is one of the most TTR-derived amyloid fibril-laden organs [2]. Indeed, cardiac amyloidosis in FAP patients progresses regardless of LT [21,22]. Additionally, senile systemic amyloidosis, which is a nonhereditary disease, causes TTR-derived cardiac amyloidosis by idiopathic deposition of wildtype TTR [40]. In contrast, no or only mild TTR-derived amyloid deposits are found in the perivascular lesions of liver, spleen, and lymph nodes in FAP patients [41,42]. In particular, although hepatocytes produce the majority of systemic TTR, they are not usually affected by this [43]. In this regards, we focused on the heart, with fewer TTR-derived amyloid fibrils in FAP patients and relatively abundant tissue-resident macrophages. Kupffer cells are present among hepatocytes, and spleen or lymph nodes also store macrophages [44,45]. To determine  [46]. Accordingly, we investigated whether or not tissueresident macrophages are decreased in heart tissue, which has relatively fewer macrophages and higher TTR deposition.
Our second reason was that heart tissue inflammation (e.g., myocarditis) is rare. Under inflammatory conditions, inflammatory cells (including macrophages) are easily recruited into the inflammatory site, and macrophages shift in polarity to inflammatory macrophages, and thereby release inflammatory cytokines [38,47]. In contrast, under recovery conditions, inhibitory macrophages are dominant and have anti-inflammatory effects [24,47]. Furthermore, inhibitory macrophages are involved in the removal of debris, angiogenesis, and tissue remodeling [24]. Because we aimed to identify tissue-resident macrophages in as physiological a condition as possible, we chose heart tissue, which is comparatively insulated from the influence of inflammation.
Based on these findings, we hypothesized that FAP patients would display quantitative or qualitative abnormalities in tissue-resident macrophages, especially inhibitory macrophages, and these abnormalities may accelerate TTR-derived amyloid deposition in some organs.
In this present study, our analysis of heart tissue showed a significantly lower number of tissue-resident macrophages in non-LT FAP patients compared with control patients. This suggests that quantitative reduction of tissue-resident macrophages partly triggers TTR deposition in heart tissue, which originally had fewer macrophages compared with the liver, spleen, and lymph nodes.
Tissue-resident macrophages originate from embryonic precursors derived from the yolk sac and can self-replenish. These macrophages are known as the inhibitory phenotype [23]. In contrast, macrophages can be partly developed from circulating blood monocytes [26,38], and monocyte-derived macrophages include both inflammatory and inhibitory phenotypes [26,38]. We found that tissue-resident macrophages in heart tissue from FAP patients were of the inhibitory phenotype. Therefore, the inhibitory macrophages we detected are more likely derived from the yolk sac rather than differentiated monocytes. In addition, qualitative abnormalities in decreased inhibitory macrophages may result in inadequate care for this regional organ, and consequently, TTR deposition may be accelerated. However, the cause of the macrophages decrease in heart tissue from FAP patients is not obvious. Further research and approaches are needed.
As described above, tissue-resident macrophages, at least in part, differentiate from monocytes [26,38]. CD14 + monocytes can be divided into three distinct subsets: CD14 high CD16 -, CD14 high CD16 + , and CD14 low CD16 + monocytes, which have different capacities for modulating inflammatory responses [48,49]. Accordingly, we next determined if there are differences in the frequency of each monocyte subset between HD and FAP patients, but found they were similar between both groups. In addition, CD163 was also equally expressed on each subset. In contrast, the intracellular TTR proportion in CD14 + monocytes from FAP patients was decreased compared with HD. These results suggest that FAP patients have insufficient TTR processing ability in the periphery. However, there was no difference in CD163 expression between HD and FAP patients. Moreover, because CD163 is a hemoglobin scavenger receptor [50], it is not directly involved in TTR uptake. It was recently reported that administration of the interleukin-1 antagonist, anakinra, induced inhibition of TTR deposition at the sciatic nerve in a FAP model mouse [51]. Moreover, interleukin-6 expression is highly expressed in chondrocytes in the presence of mutated V122I TTR [52]. These results suggest the possibility that mutated TTR produces pro-inflammatory cytokines, which may affect both tissue-resident macrophages and peripheral monocytes. Further studies are needed to clarify the mechanism of decreased inhibitory tissue-resident macrophages, and reduced endocytosis of intracellular TTR in peripheral monocytes of FAP patients.
We have recently shown that iPS-MLs are a potential therapeutic agent in the treatment of Alzheimer's disease [31,32], which causes accumulation of misfolded β-amyloid. FAP is also a disorder characterized by deposition of misfolded TTR, therefore we investigated the possibility that iPS-MLs are a useful tool for FAP. We found that iPS-MLs are inhibitory macrophages and degrade aggregated, especially polymerized TTR, in a cell number-dependent manner. Furthermore, intracellular TTR was detected in iPS-MLs cultured with wild-type-derived TTR compared with mutated-derived TTR, and iPS-MLs exhibit more phagocytic ability for aggregated TTR. In addition, iPS-MLs cultured with wild-type-derived TTR showed increased cell viability. In contrast, the propensity of iPS-MLs for mutated-derived TTR was not as great, and endocytosis of this TTR was poor. Further, mutated-derived TTR did not affect cell viability of iPS-MLs. These results suggest that iPS-MLs, which show improved cell viability with wildtype TTR, are able to more readily endocytose this type and do not function well with mutated TTR. However, truncated TTR appeared in iPS-MLs cultured with all types of TTR. These results indicate that iPS-MLs degrade both native and aggregated TTR.
Taken together, we have demonstrated the intrinsic function of iPS-MLs to be the inhibitory phenotype, leading to degradation and incorporation of TTR in vitro. However, we have not yet shown if iPS-MLs actually reduce TTR deposition in vivo. Therefore, although further investigations are required, considering the recent advances in human iPS cell-related technologies, pluripotent stem cell-derived myeloid lineages may be a promising means for the treatment of FAP. iPS-MLs (1× 10 5 cells/well) were cultured with native wild-type, mutated, wild-type-derived, and mutated-derived aggregated TTR. After 3 days, the viability of each group was evaluated by the MTS assay. Data were analyzed using the pairwise t-test with Bonferroni correction after one-way ANOVA, with ÃÃ p < 0.01 indicating a significant difference. Data are representative of four independent experiments. (TIF)