Objectives: Tertiary lymphoid organs are formed at sites of chronic inflammation and are thought to contribute to the immune response. Here, we aimed to characterize the structure and function of tertiary lymphoid organs in a model of murine kidney allotransplant to understand their role in alloimmunity.
Materials and Methods: We transplanted 4 C57BL/6 mouse kidneys (isograft group) and 17 DBA/2 mouse kidneys into C57BL/6 mouse recipients. Three DBA/2-to-C57BL/6 transplant mice that rejected their grafts acutely (before 10 days posttransplant) were excluded from the study. The 14 surviving DAB2 grafts were retrieved at day 45 posttransplant and evaluated histologically. The presence of antibody-secreting cells and circulating levels of donor-specific antibodies were also evaluated.
Results: We found that tertiary lymphoid organs can be associated with a beneficial response in a kidney allotransplant model. Characterization of B-cell subsets within tertiary lymphoid organs in mouse kidney allografts revealed naive, plasma, and memory B cells, which were mostly grouped within or in close proximity of tertiary lymphoid organs. Staining for intracellular immunoglobulin G showed that many of the B cells within tertiary lymphoid organs were capable of producing antibodies. Although allospecific antibodies were found in the serum of recipient mice and were deposited in the transplanted kidneys, graft function was not affected in this model.
Conclusions: B cells within tertiary lymphoid organs are functional and contribute to the humoral arm of the alloresponse. However, tertiary lymphoid organs are not necessarily associated with graft rejection, suggesting that protective mechanisms are at play.
Key words : Alloantibodies, Allotransplantation, Renal transplant
Rejection of transplanted organs requires encounters and cooperation of different cell types from both donors and recipients. Lymphoid organs such as draining lymph nodes and spleen provide a niche for this process to take place. In addition, tertiary lymphoid organs (TLOs), which are formed at the site of chronic inflammation, may also play a role. After transplant, there is constant immunologic engagement between donor and recipient cells. This can be seen in both chronic rejection1 and tolerance.2 Tertiary lymphoid organs in donor grafts contain different cell populations such as T, B, and dendritic cells.3 It is quite likely that B cells within TLOs can contribute to the alloimmune response.
In the past decade, B-cell involvement in post-transplant immunology has been mostly directed to avoid the production of donor-specific antibodies (DSAs). Recently, it has become obvious that poor graft outcome is not necessarily associated with the presence of DSAs in circulation.4 Furthermore, the presence of higher levels of different alloantibody isotypes can be associated with different graft outcomes, with higher levels of subclasses that have poor complement-fixing abilities resulting in better tissue performance.5
One of the key mechanisms through which regulatory B cells operate is interleukin 10 (IL-10) secretion and direct contact with CD4-positive T cells.6,7 Tertiary lymphoid organs are lymph node-like structures commonly found at sites of chronic inflammation such as infections and autoimmunity.8,9 In renal transplant patients, the presence of TLO has been linked to ongoing chronic rejection.10,11 However, not all renal grafts undergoing chronic rejection contain these B- and T-cell-rich nodules,9 suggesting that their involvement in adverse immunologic responses in the field of organ transplantation may be much more complex than previously thought. Furthermore, it has been reported that formation of TLOs and organization of lymphocytic infiltrates in lymphoid nodules are associated with superior graft function,12 suggesting that formation of TLOs is not necessarily associated with poor outcomes.
Lymphoid neogenesis within the donor graft may indicate that transplanted tissue is not only the target of the alloimmune response but may also act as a site where this response can develop locally. Molecular studies in murine allograft models have revealed that expression of biomarkers of tolerance could be found in the graft and in secondary lymphoid organs.13 Another report suggested that B-cell tolerance is broken down specifically in TLO, with more alloreactive antibodies detected within the graft than in circulation.14 All of these studies point toward a complex immunologic role for lymphocytic infiltrates within renal grafts and highlight a need to investigate the mechanisms of rejection and tolerance in models that promote TLO formation.
Here, we investigated the profile of B cells and the presence of antibody-secreting cells within the TLO in a mouse kidney allograft model of tolerance to further understand the contribution of B cells in TLOs to the alloresponse. We described immuno-globulin G (IgG)-secreting plasma cells and the deposition of IgG antibodies within the allografts. In addition, we performed isotypic profiling of circulating DSAs to elucidate its link with B-cell infiltrates in the graft and graft function.
Materials and Methods
Eight- to 12-week-old female donor DBA/2 (H-2d) and recipient C57BL/6 (H-2b) mice were purchased from Harlan (Oxon, UK). Mice were housed in specific pathogen-free facilities and used in accordance with the Animals (Scientific Procedures) Act 1986 of the United Kingdom.
Murine kidney transplant and graft function measurements
Murine kidney transplant was performed as described by Han and associates.15 The left native kidney was removed at the time of transplant. The remaining native kidney was removed 7 days later. Whole blood samples were taken at regular intervals for alloantibody assay and blood urea nitrogen (BUN) measurement using Infinity Urea (Thermo Fisher Scientific, Middletown, CT, USA) according to the manufacturer’s instructions.
Kidney grafts were retrieved 45 days posttransplant. Half of the explants were fixed in periodate-lysine-paraformaldehyde solution, and the other half were fixed in 10% neutral buffered formalin solution. Periodic acid-Schiff staining was carried out on 5-μm-thick paraffin sections. The area of the TLO was measured in medium-powered fields (×160) of graft cross sections using Lucia G software (Nikon, Tokyo, Japan). Results were normalized against the total area of the kidney and expressed as micrometer squared per medium-powered field.
Periodate-lysine-paraformaldehyde-fixed and frozen tissue sections were defrosted and fixed with ice-cold acetone for 5 minutes. Sections were subsequently treated with avidin/biotin block (Vector Laboratories Ltd., Cambridge, UK) and exposed to primary antibodies against CD19 (clone 1D3; BD Pharmingen, Oxford, UK); CD138 (clone281-2; BioLegend, London, UK); polyclonal IgG (eBiosciences, Hatfield, UK); polyclonal IgG1, IgG2b, and IgG3 (Southern Biotech, Birmingham, AL, USA); polyclonal IgG2c (Bethyl, Montgomery, AL, USA); IL-10 (clone JESS-2A5; AbDSerotec, Kidlington, UK); and polyclonal C3 (LifeSpan Bioscience, Seattle, WA, USA). Appropriate second-ary antibodies were next applied when 2-stage staining was required, which was followed by washing and streptavidin-conjugated fluorochromes. The second set of antibodies (if used) was applied to tissue sections to allow visualization of double-stained cells. Sections were then washed and mounted in Vectashield mounting medium with 4′,6-diamidino-2-phenylindole (Vector Laboratories). Slides were viewed with A1R Si confocal microscope or fluorescent microscope (Nikon), and positive cells were counted in at least 5 random high-powered fields (hpf; ×600).
Alloantibody in sera
The presence of IgG DSAs was determined by the ability of serum to bind to donor splenocytes. Binding was detected by flow cytometry, using goat fluorescein isothiocyanate-conjugated anti-mouse IgG (Sigma; Dorset, UK; 5897) and IgG1, IgG2b, IgG2c, and IgG3.
All continuous data are expressed as means ± standard error of the mean, and P < .05 was taken as significant. P values were estimated using two-way and one-way analysis of variance with post hoc Bonferroni intergroup comparison. Correlation analysis was performed using linear regression function. All computations were performed using GraphPad Prism 6 (GraphPad Software, Inc. La Jolla, CA, USA).
Allotransplant model of tolerance and tertiary lymphoid organ formation
To characterize B-cell subsets in TLOs from trans-planted kidneys, we used an established model of TLO formation in murine kidney allografts. Of 17 DBA/2 mouse kidneys transplanted into fully major histocompatibility complex-mismatched C57BL/6 recipients, 14 (82.4%) survived until termination of the experiment on day 45. DBA/2 grafts were performing well at day 45 posttransplant in this model, with long-term survival likely.14 The 3 grafts (17.6%) with acute rejection succumbed within 10 days posttransplant. Recipients of syngeneic C57BL/6 grafts had normal BUN results for the duration of the experiment, as did most recipients of DBA/2 allografts, except for 3 that had elevated BUN levels at time of retrieval (Figure 1, A and B). In addition, in grafts that survived until day 45 posttransplant, extensive formation of TLO could be seen (Figure 1C). There was a strong positive correlation between graft function and TLO area (Figure 1D).
B-cell subsets and antibody-producing cells within tertiary lymphoid organs
To characterize the subpopulations of B cells infiltrating the kidney allografts, tissue sections were stained for CD19 (B-cell marker), CD138 (plasma cell marker), and surface IgG (present on memory B cells and plasma cells) and intracellular IgG (plasma cells).
Dense populations of CD19-positive cells could be seen within the epicentral blood vessels of TLOs and across the whole thickness of TLOs. CD138- and IgG-positive mature plasma cells were mostly localized on the periphery of TLOs and not in the immediate vicinity of the blood vessel (Figure 2A), which could suggest local differentiation of B cells.
To investigate whether infiltrating cells within the kidney grafts were capable of producing antibodies, confocal microscopy was performed to look for the presence of intracellular IgG, which signifies antibody-producing cells. Intracellular IgG-positive cells could be seen, mainly in the periphery of TLOs (Figure 2B). On average, 10 intracellular IgG-positive cells/hpf could be seen within the periphery of the allograft TLO (Figure 2C).
Although isografts had few B cells (0.87 ± 0.36/hpf), plasma cells (0.53 ± 0.21/hpf), and IgG-positive cells (0.2 ± 0.1/hpf), the DBA/2 allografts retrieved 45 days posttransplant from C57BL/6 recipients had 47.5 ± 6.77/hpf infiltrating B cells, 25.46 ± 5.46/hpf plasma cells, and 11.25 ± 2.25/hpf IgG-positive cells, reaching P = .003, P < .001, and P ≤ .001, respectively. Allografts also contained significantly more surface IgG-positive, CD138-negative cells compared with isografts, which showed 2.63 ± 0.52/hpf and 0.25 ± 0.14/hpf, respectively (P = .003). This population could contain memory B cells.
To detect the presence of IL-10-producing B cells, sections of allografts and isografts were stained with anti-IL-10 and anti-CD19 antibodies. We could find distinct IL-10-positive cells only in the allografts and only within TLOs or their immediate vicinity (Figure 2, D-F).
To establish whether the numbers of infiltrating B cells had any effect on graft performance, we plotted total numbers of CD19-positive, CD138-positive, and IgG-positive cells against BUN levels. Although we did not find a correlation between graft function and the number of infiltrating CD19-, CD138-, or IgG-positive cells (Figure 3, A, C, and E), we did observe a strong correlation between lower BUN measurements and higher percentages of these cell types within the TLO rather than outside of the lymphoid aggregates (Figure 3, B, D, and F) (r2 = 0.7423, P = .028 for CD19-positive; r2 = 0.9553, P = .0008 for CD138-positive; and r2 = 0.8275, P = .0119 for IgG-positive cells).
We observed no specific polarization toward one or a group of IgG subclasses within IgG-positive cells, with a trend that did not reach overall signi-ficance, showing high numbers of IgG2c-positive cells (Figure 3G). Allografts contained on average 4.25 ±1.73/hpf IgG1-positive, 2.33 ± 0.68/hpf IgG2b-positive, 6.17 ± 1.41/hpf IgG2c-positive, and 1.58 ± 0.48/hpf IgG3-positive cells, whereas isografts contained significantly less (less than 1/hpf IgG-positive cell). We also found that, although all IgG subclasses were present within the grafts, only the percentage of IgG2c-positive cells within TLOs could be correlated with better graft performance (r2 = 0.7023, P = .0372) (Figure 3H).
Pattern of antibody deposition in donor grafts
Immunoglobulin G staining of kidney allograft tissues showed not only cellular staining but also distinct staining of lamina propria of some of the epithelial cells in the renal tubules. This IgG staining was mostly seen in allografts and not in isografts (Figure 4, A and B). To confirm complement activation, we stained the section with C3 (Figure 4, C and D).
Noncellular staining levels of deposited IgG1, IgG2b, Ig2c, and IgG3 were analyzed in 20 random high-powered fields. An arbitrary semi-quantitative scoring system was used to quantify noncellular deposition of immunoglobulins on the renal tubules (Figure 4A). Tubular staining was evaluated based on the percentage of tubular circumference positive for IgG. Three points were given to tubules that had 100% of their circumference stained, 2 points to those with more than 50% positivity, and 1 point to those with less than 50% positivity along the circumference. Cross sections of tubules that showed no IgG staining were scored as 0 points. All arbitrary scores were normalized against total number of tubules within high-powered fields.
We observed no correlation between total IgG deposits in allografts and graft function, although amount of noncellular IgG1staining was positively correlated (r2 = 0.7668, P = .0222) with total area of the graft cross section occupied by TLO (Figure 4, E and F).
We also stained sections of DBA/2 kidneys transplanted into fully mismatched recipients for 4 different IgG isotypes: IgG1, IgG2b, IgG2c, and IgG3. No polarization toward one of the isotypes or even toward the isotypes that were more or less able to fix complement and therefore promote antibody-mediated rejection was seen. There was also no detectable link between the deposition in the allografts of the individual isotypes and graft function or TLO presence (data not shown).
Level of circulating donor-specific antibody is not associated with graft
The alloantibody levels were measured weekly in sera obtained from transplant recipients. The mean fluorescence intensity of alloantibody titers was plotted against time, and area under the curve was measured to assess alloantibody presence in the circulation over the duration of the study.
The good renal allograft function shown in our model of DBA/2-to-C57BL/6 transplant was not simply due to immune ignorance, as 12 of 13 recipients (many of which had low BUN levels at time of graft retrieval) developed high DSA titers as early as 2 weeks posttransplant (Figure 5, A and B). Sera from isograft recipients showed no detectable levels of C57BL/6-reactive antibodies. There was no distinguishable polarization toward specific IgG isotypes within these sera, with all subclasses being present at comparable levels in DBA/2 kidney recipients (Figure 5C). Linear regression analysis showed that elevated levels of IgG1 and IgG2b isotypes, but not IgG2c or IgG3 isotypes and total IgG/IgM, could be linked with worse graft function (Figure 5, D and E).
In this study, we focused on B-cell and humoral responses within TLOs to further understand their contribution to the alloimmune response. Our mouse model showed that TLOs form reliably over time and that their presence correlates with good graft function. The role of TLOs as structures rich in immunologically active cells in antibody production is unknown. Allografts, however, have been previously proven to be the source of alloantibodies.10 Tertiary lymphoid organs are sites where infiltrating inflammatory cells are trapped by defective lymphatic drainage; thus, they produce huge amounts of cytokines and growth factors, while dendritic cells and B cells constantly present antigens.16 It is yet not known whether TLO responses mirror a systemic response that secondary lymphoid organs initiate or whether they have specific features that determine graft outcome. Because molecular and immunologic markers of tolerance are predominantly expressed in trans-planted organs, this site-specific lymphoid neogenesis has recently attracted increasing attention in the field of allotransplantation.6
Although historically formation of TLO in trans-planted organs has been linked to chronic rejection, the presence of new lymphoid tissue can potentially generate regulatory and protective immune responses that allow graft accommodation.12 In clinical studies, lymphoid aggregates were found not only in rejecting allografts but also in those devoid of histologic signs of rejection.17 Experimental models of autoimmunity and infection suggest that the development of tertiary lymphoid tissue correlates with a marked limitation of tissue damage. Animal models of transplant tolerance where TLOs are formed are an excellent opportunity to investigate the function of TLOs in a “regulatory” environment. Here, we found a strong correlation between the amount of lymphoid tissue and better function of fully mismatched DBA/2 kidney grafts in C57BL/6 recipients. In humans, biopsies are mostly taken from organs that are undergoing rejection. Therefore, the existence of TLOs in grafts with good function may go unnoticed. The model presented here allows an insight into the natural (without the influence of immunosup-pression) tissue occurrences that promote TLO formation.
Tertiary lymphoid organs are local, com-partmentalized structures where antigen-driven maturation and T-cell-dependent activation of naive B cells occur. This creates the possibility that naive alloimmune B cells can be educated within the TLO, resulting in a suppressive or inflammatory response. The hypothesis of local activation is also supported by the finding that a humoral reaction can be initiated within the graft.10 Because human renal allograft tolerance has been associated with the maintenance of a healthy B-cell compartment, we wanted to investigate B cells and antibody pro-duction within TLOs more closely in a model of renal transplantation. Although a negative correlation between the number of intrarenal B cells and graft survival has been suggested,18 in our model, high numbers of B and plasma cells were found in kidneys of recipients that had low and high BUN results. Despite observations of negative effects of infiltrating lymphocytes on human kidney grafts, a correlation has been shown between large numbers of infiltrating B and T cells and better graft survival.14 Many clinical studies have reported that the presence of CD20-positive rich infiltrates does not correlate with decreased renal graft survival, C4d deposition, or DSA development.19-21 In our model, graft function was associated with B-cell infiltration only if they were organized into TLOs, as evident by the distinct link between organization of CD19-, CD138-, and IgG-positive cells within TLO and low, stable BUN levels.
The presence of IgG- and CD138-positive cells only in the periphery of TLOs found in our histologic specimens could suggest functionality of the nodules and the possibility that B cells differentiate within the TLO, an observation that has also been previously highlighted.14 This phenomenon could also be seen in histologic analyses of large cohorts of clinical samples, in which noncluster-forming, but not cluster-forming B-cell presence, was shown to be correlated with antibody-mediated rejection markers.12,18 It is possible that the presence of large numbers of immune cells does not adversely affect the graft if they are organized within the TLO.
Operationally tolerant individuals have also been reported to retain normal numbers of regulatory T and B cells, whereas patients with chronic rejection have significantly lower numbers of both.22,23 Secretion of IL-10 continues to be a vital component in the identification of regulatory B cells.24,25 We were able to show that the presence of IL-10-positive, CD19-positive B cells could potentially represent a subset of IL-10-producing B cells.
B-cell tolerance has been proven to be broken in intragraft TLOs by the presence of higher levels of autoantibodies within the graft than in the blood, with the graft appearing to be the sole possible source of autoantibodies in some patients.14 These interesting results suggest that antibodies could be produced within the graft and influence graft outcome. Using this model of allotransplant, we investigated whether different graft outcomes could be linked with varying levels of DSAs in sera or IgG deposited in the graft. Because complement fixation is strongly associated with the ability of antibody to cause antibody-mediated rejection,26 we wanted to check whether a regulatory profile of circulating or tissue-bound IgG was established based on IgG isotype polarization in a similar way to that reported by Le Texier and associates.27 The potency of C57BL/6 mouse IgG subclasses to activate the classical complement pathway decreases in the following order: IgG2b > IgG2c > IgG3 > IgG1. Humans with noncomplement-fixing antibodies have a prognosis similar to those without DSAs.28 In our model, C57BL/6 recipients of DBA/2 kidneys developed high levels of DSAs in sera, and the alloantibody titers were not associated with poor graft performance or the amount of IgG deposited in the kidney. We also could see no polarization toward Th1 (IgG2b and IgG2c) or Th2 (IgG1 and IgG3) isotypes in sera or grafts. Most of the DBA/2 kidney recipients developed high levels of DSAs as early as 2 weeks posttransplant, which remained high until the end of the experiment. At the same time, complement activation markers could be seen in peritubular capillaries of all allografts but not in isografts. The amount of noncellular IgG deposition in allografts did not have an effect on renal function, but it correlated with the presence of TLOs.
We were unable to establish whether the antibodies deposited in kidney grafts were produced locally and which antigens they were binding to. Some reports have suggested that auto- and donor antigens are targets for antibodies produced within the graft.10,14 In the clinical setting, high DSA levels are linked to C4d deposition on peritubular capillaries, but staining of tubular epithelium has been linked to autoantigen stimulation in animal models.16 When we marked the distribution of IgG deposition within the grafts in relation to TLOs, we could see no dependence of IgG presence on close proximity to the B-cell aggregates (data not shown).
In conclusion, our data suggest that TLOs within donor kidney allografts are capable of supporting the B-cell response but are not necessarily associated with poor graft outcome.
DOI : 10.6002/ect.2017.0261
From the MRC Centre for Transplantation, King’s College London School of
Medicine at Guy’s, King’s and St. Thomas’ Hospitals, London, United Kingdom
Acknowledgements: The authors have no conflicts of interest to declare. The authors acknowledge the NIKON Imaging Centre at King’s College London for help with confocal microscopy and MRC Centre grant MR/J006742/1 for supporting this work.
Current address of A. K. Nowocin: National Institute for Biological Standards and Control (NIBSC), Medicines and Healthcare Regulatory Agency, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, UK
Corresponding author: Wilson Wong, MRC Centre for Transplantation, 5th Floor, Tower Wing, Guy’s Hospital, London SE1 9RT, UK
Phone: +44 2071881522
Figure 1. Tertiary Lymphoid Organ Formation in DBA2 Mouse Kidney Grafts Transplanted Into C57BL/6 Mouse Recipients
Figure 2. Histological Analysis of B Cells Within Tertiary Lymphoid Organs
Figure 3. B Cells Influence Allograft Performance When Organized Into Tertiary Lymphoid Organs
Figure 4. Immunoglobulin G and C3 Deposition in Grafted Kidneys
Figure 5. Donor-Specific Antibody Production in Allograft Recipients