MHC Class II Regulates Stemness and Differentiation of Normal and Malignant B Cells

MHC class II regulates stemness and differentiation of normal and malignant B cells

Julia Merkenschlager1, Urszula Eksmond1, Jan Attig1, George R. Young2, Carla

Nowosad3, Pavel Tolar3 and George Kassiotis1,4,*

1Retroviral Immunology, 2Retrovirus-Host Interactions and 3Immune Receptor

Activation Laboratory, The Francis Crick Institute, 1 Midland Road, London NW1

1AT, UK.

4Department of Medicine, Faculty of Medicine, Imperial College London W2 1PG,

UK.

*Correspondence: Dr George Kassiotis, E-mail: [email protected], T: 44

(0) 20379 61483

SUMMARY

Peptide presentation by membrane-bound major histocompatibility complex class II (MHC II) proteins governs CD4+ T-cell development, survival and effector response. However, the MHC II α and β chains demonstrate signalling capacity with their engagement triggering biochemical cascades that alter the in vitro function or survival of MHC II-expressing cells1,2. These observations suggest a possible, as yet undefined role for MHC II in immune-cell physiology or pathology. Here we show that MHC II expression cell-autonomously modifies the balance between stemness and differentiation in B-cell precursors, as well as malignant B cells. Initiation of MHC II expression early during bone marrow B- cell development limited the occupancy of cycling compartments by promoting differentiation, thus regulating the numerical output of B cells. MHC II deficiency preserved stem-cell characteristics and prolonged proliferation of pro-

B cells in vivo, whereas ectopic MHC II expression was sufficient to promote hematopoietic stem-cell (HSC) differentiation in vitro. Moreover, MHC II restrained growth of murine B-cell leukaemias in vitro and in vivo, independently of CD4+ T-cell surveillance. Notably, acquisition of stem-cell features in MHC II deficient B-cell leukaemias included expression of fibroblast activation protein

(FAP), previously thought restricted to tumour stromal fibroblasts3,4. Amongst human B-cell malignancies, lack of MHC II expression, particularly in Hodgkin lymphoma cell-lines was accompanied not only by pronounced FAP expression, but also by presentation of a multipotent bone marrow stromal-cell (BMSC) phenotype, indicating a B-cell–BMSC transition.

Major histocompatibility complex class II (MHC II) is best known for presenting antigenic peptides to T cells, but may also transmit reverse signals capable of modulating diverse cellular processes. To date, MHC II has been reported to alter immune-cell responsiveness to acute stimuli, such as lipopolysaccharide (LPS) and other microbial products5,6. Further, its ligation has been reported to promote immune-cell adhesion, cytokine production and in some instances apoptosis1,2.

Despite this, the contribution of cell-autonomous MHC II signalling to immune cell development, function or pathology, remains to be understood.

In light of this and to investigate cell-autonomous functions of MHC II, we employed

Cre-mediated ablation of a conditional H2-Ab1 allele (H2-Ab1c), encoding the MHC

II Aβ chain. Here, incomplete and ectopic Cre activity in on-target and off-target cell- types, creates a degree of H2-Ab1 deletion mosaicism, permitting comparison between MHC II-expressing and MHC II-deleted cells within the same host.

Dendritic-cell (DC)-targeted Cre expression (CD11c-Cre driver; Itgaxcre) expectedly abolished MHC II expression in DCs, but surprisingly also in the majority of B cells

(Fig. 1a, b). This was in contrast to the reported activity of this Cre driver7.

Unexpectedly high proportions of B cells also lost MHC II expression when Cre expression was highly restricted to classical DCs using the zDC-Cre driver (Zbtb46cre)

(Fig. 1a, b). Furthermore, targeting Cre expression to early or immature B-cell lineages using the mb1-Cre (Cd79acre) or the CD23-Cre (Fcer2acre) driver, respectively, abolished MHC II expression in nearly all B cells, as well as in a considerable proportion of DCs (Fig. 1a, b).

Differences in the relative efficiency of Cre-mediated recombination in distinct conditional alleles notwithstanding, we considered whether the unexpectedly high frequency of apparent off-target loss of MHC II in the Cre drivers tested, was

3 accentuated by aberrant lineage commitment or selective outgrowth of initially rare

MHC II-deleted cells. To probe this possibility and to circumvent the inflammatory disease that develops in ItgaxcreH2-Ab1c mice8, we set up competitive bone marrow chimeras between H2-Ab1c and wild-type (WT) mice (Extended Data Fig. 1a-d).

Interestingly, bone marrow from ItgaxcreH2-Ab1c or Fcer2acreH2-Ab1c donors produced preferentially B cells, at the expense of T cells and myeloid cells (Extended

Data Fig. 1a-d), indicating skewed development towards the B-cell lineage.

Importantly, this skewing in development, also drove the proportional outgrowth of

MHC II-deleted cells over WT competitors within the B-cell lineage (Fig. 1c, d).

Indeed, several subsets of mature B cells were significantly enriched in MHC II- deleted B cells, with the exception of germinal centre B cells, (Fig. 1c, d), which were exclusively MHC II-positive, likely due to their dependency on CD4+ T-cell help for survival. Comparable results were obtained when H2dlAb1-Ea mice, constitutively deleted in all MHC II-encoding genes, were used as donors (Extended Data Fig. 1e and 2a), arguing against a secondary effect of Cre-mediated deletion or of expression of unpaired MHC class II chains9.

Next, we explored how the observed advantage of MHC II-deleted cells related to the kinetics of MHC II expression during B-cell development. Surface expression of

MHC II in WT cells was first detectable by flow cytometry in Hardy’s fraction10 (Fr)

A cells, gradually increasing both in terms of frequency and intensity in subsequent Fr

B to F (Fig. 1e, f). This was in keeping with prior reports, indicating that initiation of

H2-Ab1 transcription occurs as early as Fr A cells11,12. Microscopically, MHC II proteins were detected at the plasma membrane weakly in WT Fr A cells and more prominently in Fr B or C cells (Fig. 1g), further confirming MHC II expression at these early developmental fractions. Consistent with the kinetics of MHC II

4 expression, the overrepresentation of MHC II-deficient B cells in mixed bone marrow chimeras manifested between Fr A and D (Fig. 1h and Extended Data Fig. 2b). In contrast, MHC II deficiency did not affect the starting frequency of hematopoietic stem cells (HSCs) with in vitro B-cell potential (Extended Data Fig. 3). Collectively, these findings suggest that MHC II affects progression through the B-cell developmental stage, at which it is first expressed.

Low level MHC II expression in developing B cells is considered sufficient to mediate antigen presentation to T cells, albeit with reduced efficiency13. To begin to investigate how MHC II expression might influence B-cell development, we first explored a possible role for T cells bearing MHC II-restricted T-cell receptors

(TCRs). However, H2-Ab1c deleted progenitors still outcompeted WT counterparts in mixed bone marrow chimeras generated in H2dlAb1-EaRag1-/- recipients, where CD4+

T-cell development is preluded (Extended Data Fig. 4). Furthermore, competitive in vitro differentiation assays revealed that Fr A cells purified from ItgaxcreH2-Ab1c or

Fcer2acreH2-Ab1c donors numerically outcompeted those from WT donors in subsequent developmental stages (Fig. 2a). Collectively, these experiments suggested that MHC II expression in early B-cell precursors influences their developmental transition from Fr A to D cell-intrinsically.

To further investigate possible cell-intrinsic effects of MHC II, we compared the transcriptional profiles of WT and ItgaxcreH2-Ab1c Fr D pre-B cells (Fig. 2b, c).

Within WT Fr D pre-B cells, expression of surface MHC II proteins segregated cells according to maturation status, as indicated by expression of multiple MHC II-related genes, immunoglobulin genes and other maturation markers (e.g. Ccr6, Fcrl5 and

Il9r) (Fig. 2b and Table S1). These findings suggested that MHC II-expressing Fr D B

5 cells were more advanced in their differentiation, compared with MHC II-negative counterparts, further supported by comparison with splenic B cells (Extended Data

Fig. 5), as well as by publicly-available expression data (Extended Data Fig. 6a).

In contrast to MHC II-expressing WT Fr D cells and despite transcription of immunoglobulin genes, expression of many expressed maturation markers was absent from the respective ItgaxcreH2-Ab1c Fr D subset (Fig. 2b), indicating stunted differentiation. This was also evident in the ItgaxcreH2-Ab1c Fr D population as a whole, which was transcriptionally distinct from WT counterparts (Fig. 2c and Table

S2). Furthermore, specific to ItgaxcreH2-Ab1c Fr D cells were a set of genes, including

Cox6a2, Plcb4 and Erdr1 (Fig 2c), which were shared with common lymphoid progenitor (CLP) or Fr A cells (Extended Data Fig. 6b), suggesting retention of precursor traits in MHC II-deleted pre-B cells. Indeed, comparison of genes differentially expressed between ItgaxcreH2-Ab1c and WT Fr D B cells with precompiled expression profiles of developing B cells confirmed the advanced differentiated state of MHC II-expressing WT Fr D cells and importantly, revealed that the respective ItgaxcreH2-Ab1c Fr D subset exhibited a transcriptional signature comparable with those of earlier precursors (CLP and Fr A) (Fig. 2e and Extended

Data Fig. 7). Additionally, whereas WT Fr D cells typically exit the extensive cell- division that characterises the earlier Fr B and C, and become quiescent14,15,

ItgaxcreH2-Ab1c Fr D maintained a proliferative signature, with the top pathways differentially upregulated being cell-cycle-related (Fig. 2d). This suggested that outgrowth of H2-Ab1c deleted progenitors, as observed in the competitive assays, might be achieved by prolonged proliferation. These differences were restricted to pre-B cells, as follicular B cells from ItgaxcreH2-Ab1c and WT donors were transcriptionally comparable (Fig. 2c). Thus, expression of MHC II at the earliest B-

6 cell developmental stage appeared to promote B-cell differentiation at the expense of proliferation.

Transition through the distinct maturational checkpoints of B-cell development is driven by well-characterised signals, including cytokines and chemokines (CXCL12,

SCF and IL-7), and signalling initially by the pro-BCR and pre-BCR and ultimately the mature BCR14,15, which may have been affected by MHC II deficiency. These signals additionally support cellular proliferation of B-cell precursors, particularly pronounced at Fr B and C cells14,15. The early advantage of MHC class II-deleted B cells was still evident in competition between Rag1-/- B cells and H2dlAb1-EaRag1-/- B cells, which were again overrepresented in such mixed bone marrow chimeras

(Extended Data Fig. 8), despite a block in differentiation at Fr C, owing to the lack of a functional BCR. Similarly, transcription of Il7r, as well as known targets of IL-7 receptor signalling, such as Ebf1 (early B-cell factor) and Cd79a (mb-1, encoding

Igα), and transcription of Kit, Cxcr4 and Flt3, encoding receptors for SCF (stem cell factor), CXCL12 and Flt3L, respectively, were broadly comparable between

ItgaxcreH2-Ab1c and WT Fr D B-cell precursors (Extended Data Fig. 9). These data argued against an IL-7-driven mechanism or a requirement for BCR-initiated signals.

B-cell development is also sensitive to host-extrinsic factors, such as infection or inflammation, which can skew bone marrow development in favour of myeloid cells, at the expense of B cells16. MHC II deficiency has been reported to enhance and reduce responsiveness of mature B cells and DCs to Toll-like receptor (TLR) 4 agonists5,6. Nevertheless, responsiveness to ligation of the two TLRs highly expressed in B-cell precursors, namely TLR2 and TLR4, was minimally affected by MHC II deficiency (Extended Data Fig. 10). These results raised the possibility that, instead of

7 modifying responsiveness to known B-cell differentiation signals, MHC II was promoting precursor-cell differentiation directly. To explore this possibility, we measured the efficiency of differentiation of HSCs ectopically expressing MHC II.

Strikingly, in competitive in vitro B-cell differentiation assays, MHC II-transduced

H2dlAb1-Ea HSCs exhibited considerably enhanced ability to differentiate, compared with H2dlAb1-Ea HSCs or WT HSCs, which did not yet endogenously express MHC II

(Fig. 3a, b). Thus, ectopic MHC II expression was sufficient to promote B-cell differentiation of HSCs.

Depending on the cell-type, MHC II cross-linking in vitro has been previously reported to trigger two major signalling cascades1,2. One involves elevation of cellular cAMP levels and translocation of protein kinase C (PKC), and requires the MHC II β chain cytoplasmic domain1,2. Another leads to Src kinase activation and Ca2+ mobilisation, entirely independently of the cytoplasmic domains of either MHC II chains1,2. However, an alternative way by which MHC II molecules may exert cell- intrinsic effects is through the cytoplasmic domain of the Aα chain1,2. This domain controls the efficiency of translational diffusion of membrane-bound MHC II molecules, likely through interaction with the cytoskeleton17, but is considered redundant for the two described signalling cascades triggered by MHC II cross- linking. To distinguishing between the two possibilities, we constructed a version of

MHC II, lacking the cytoplasmic domain of the Aα chain (Fig. 3d). Despite achieving surface expression comparable to the WT AαAβ heterodimer (Fig. 3e), counterpart lacking the Aα cytoplasmic domain was unable to enhance B-cell development of transduced HSCs (Fig. 3f). This requirement for the Aα cytoplasmic domain argued that MHC II expression promotes and consolidates HSC differentiation through a

8 mechanism involving altered transport and/or recycling of MHC II molecules that was distinct from the known MHC II signalling cascades.

The loss of MHC II is a frequent feature of B-cell malignancies and is associated with poor prognosis18. MHC II-deleted B-cell precursors retain stem-cell characteristics and overexpress genes associated with cancer. For example, gain-of-function mutations in Plcb4 (Phospholipase b4), a gene involved in cellular senescence, are a cause of uveal melanoma in humans19, and mouse-specific Erdr1 (Erythroid differentiation regulator 1), expressed in CD34+ HSCs and transformed B cells, functions as a stress-related survival factor20,21. Moreover, many B-cell tumours rely on aberrant activation signals22, which may be modified by MHC II. We therefore hypothesised that MHC II loss in transformed B cells may confer a cell-intrinsic fitness advantage, in addition to escape from CD4+ T-cell immune surveillance.

To explore this, we examined a series of precursor B-cell leukaemia cell-lines, originally developed in IL-7-overexpressing mice23. Individual cell-lines were arrested at distinct stages of B-cell development and exhibited variable MHC II expression (Extended Data Fig. 11a, b). Also variable was expression of genes that characterised MHC II-deleted non-transformed pre-B cells, with the exception of

Erdr1, which was highly expressed in all cell-lines (Extended Data Fig. 11c).

Sublines of A1, F6 or G7 leukaemic cells selected by cell-sorting for high MHC II expression quickly became MHC II-negative upon in vivo growth in lymphocyte- deficient hosts (Fig. 4a), signifying a possible growth advantage of MHC II-negative cells, independently of CD4+ T-cell immune surveillance.

To prevent the natural loss of MHC II expression in leukaemic cells grown in vivo and to introduce MHC II expression in those cells that were previously negative, such

9 as the B3 cells, we transduced them with H2-Ab1 and H2-Aa transgenes. Upon transplantation at equal ratios into lymphocyte-deficient hosts, MHC II-negative B3,

F6 or G7 leukaemic cells quickly outcompeted their MHC II-transduced counterparts, although in vivo growth of D5 leukaemic cells remained unaffected (Fig. 4b).

Moreover, MHC II-negative F6 and B3 leukaemic cells could outpace their MHC II- expressing equivalents in vitro, either in a feeder cell-dependent or independent manner (Fig. 4c, d).

Notably, transduction of F6 cells with MHC II also lead to their phenotypic progression to Fr C (Fig. 4e), further supporting a role for MHC II in promoting differentiation. Consistent with reduced in vitro and in vivo growth, MHC II- transduced F6 cells exhibited distinct changes in gene expression, including upregulated expression of additional MHC II-related genes, immunoglobulin genes and other maturation markers (Fig. 4f).

Enhanced differentiation of MHC II-transduced F6 cells was associated with reduced expression of a smaller set of genes, including some involved in the IFN response and in metabolism (Fig. 4f). Particularly surprising was the modulation of Fap expression, encoding the dipeptidyl peptidase and endopeptidase fibroblast activation protein

(FAP). FAP is most prominently expressed in stromal fibroblasts of almost all human epithelial tumours, but generally not in healthy tissue making it a tractable target for multiple cancer therapies4,24. FAP may also be involved in embryonic development and tissue remodelling, and its immune targeting uncovered expression in multipotent bone marrow stromal cells (BMSCs)25. Assessment of publicly-available data of physiological murine B-cell development revealed Fap expression in earliest CD34- and CD34+ long-term HSCs, but not at later stages (Extended Data Fig. 12), highlighting Fap expression as a stem-cell feature. In addition, Fap expression has not

10 been reported in transformed lymphocytes. To examine this and to investigate a wider correlation between FAP and MHC II, we analysed publicly-available data from a range of human B-cell leukaemia cell-lines. Particularly pronounced was FAP expression in Hodgkin lymphoma (HL) cell-lines, compared with those from other B- cell malignancies (Fig. 5a). Importantly, FAP expression in the HL cell-line panel was bimodal and strongly anti-correlated with expression of human leukocyte antigen

(HLA) II genes, encoding MHC II proteins in human (Fig. 5a, b). A similar anti- correlation was observed in other B-lymphoid or myeloid cell-lines, albeit FAP- expressing clones were the minority (Extended Data Fig. 13).

Consistent with elevated FAP mRNA levels in HL cell-lines that had downregulated

HLA II gene expression, FAP protein levels inversely correlated with MHC II protein expression (Fig. 5c). Surprisingly, in addition to presenting fibroblastoid morphology,

FAP-expressing HL cell-lines coexpressed many of the markers that define multipotent BMSCs, but not those that are specific to B cells (Fig. 5d). Furthermore, this transcriptional distinction appear gradual and in proportion with FAP expression, with L-1236 cells exhibiting an intermediate profile (Fig. 5d). These results indicated that lack of MHC II expression in HL cell-lines is accompanied not only by expression of FAP, but also by apparent transition to a multipotent BMSC phenotype.

MHC II is of paramount importance in the initiation of adaptive immune responses and the strongest genetic determinant of susceptibility to autoimmune and infectious diseases26-28. Although antigen presentation to CD4+ T cells is the primary function of

MHC II, accumulating evidence for signalling capacity of MHC II molecules suggests additional cell-intrinsic functions. The data presented here demonstrate that MHC II participates in the differentiation programme of HSCs committed to the B-cell

11 lineage. Not only is its expression progressively upregulated during B-cell development, MHC II in turn provides differentiation cues, highlighting a potential positive feedback look consolidating B-cell maturation. Notably, MHC II expression in pro-B cells appears developmentally regulated and occurs during adult, but not foetal or neonatal murine B-cell development11,12. These findings implicate MHC II in

B-cell tolerance and support a model whereby MHC II-negative foetal or neonatal pro-B cells produce mature B1 B cells, whereas MHC II-expressing adult pro-B cells, which also express the MHC II receptor CD4, are progenitors of conventional B2 B cells11,12.

MHC II expression is induced in a variety of hematopoietic and non-hematopoietic cells during inflammation, but is often lost following transformation, particularly in certain B-cell malignancies18,29-31. Although traditionally considered an escape mechanism from CD4+ T cell immunosurveillance, our findings suggest that loss of

MHC II expression in transformed cells may additionally provide a cell-intrinsic fitness advantage by virtue of reduced differentiation. Among B-cell malignancies,

HL is characterised by the most dramatic loss of B-cell phenotype32. The results presented here suggest a transition of malignant B cells to multipotent BMSC phenotype, akin to the widely-recognised epithelial – mesenchymal cell transition33.

Selection of cells retaining of stem-cell properties, as well as escaping immunosurveillance following loss of MHC II would undoubtedly accelerate cancer progression. Critically, the unexpected expression of FAP in cancers that have lost

MHC II, uncovers a new therapeutic opportunity.

AUTHOR CONTRIBUTIONS

J.M. designed, performed and analysed most of the experiments with input from U.E.

J.A. and G.R.Y. assisted in the analysis of RNA-seq data. C.N. and P.T. performed the large-scale fluorescence imaging. G.K. conceived and supervised the project and, together with J.M., wrote the manuscript.

ACCESSION NUMBERS

RNA-seq data have been deposited in the ArrayExpress database at EMBL-EBI

(www.ebi.ac.uk/arrayexpress) under accession numbers E-MTAB-6763 and E-

MTAB-6765.

ACKNOWLEDGMENTS

We wish to thank Prof. Amanda Fisher for providing the murine bone marrow B-cell leukaemia cell-lines and Dr. Jakob Loschko and Prof. Michel C. Nussenzweig for providing the Zbtb46creH2-Ab1c mice. We are grateful for assistance from the Flow

Cytometry, Biological Resource, Cell Services, and High Throughput Screening facilities at the Francis Crick Institute. This work benefited from data assembled by the ImmGen and CCLE consortia. This work was supported by the Francis Crick

Institute which receives its core funding from Cancer Research UK (FC001099), the

UK Medical Research Council (FC001099), and the Wellcome Trust (FC001099).

REFERENCES

1 Wade, W. F. et al. Structural compartmentalization of MHC class II signaling function. Immunol. Today 14, 539-546 (1993). 2 Al-Daccak, R., Mooney, N. & Charron, D. MHC class II signaling in antigen- presenting cells. Curr. Opin. Immunol. 16, 108-113 (2004). 3 Fearon, D. T. The carcinoma-associated fibroblast expressing fibroblast activation protein and escape from immune surveillance. Cancer. Immunol. Res. 2, 187-193 (2014). 4 Jiang, G. M. et al. The application of the fibroblast activation protein alpha- targeted immunotherapy strategy. Oncotarget 7, 33472-33482 (2016). 5 Rodo, J., Goncalves, L. A., Demengeot, J., Coutinho, A. & Penha-Goncalves, C. MHC Class II Molecules Control Murine B Cell Responsiveness to Lipopolysaccharide Stimulation. J. Immunol. 177, 4620-4626 (2006). 6 Liu, X. et al. Intracellular MHC class II molecules promote TLR-triggered innate immune responses by maintaining activation of the kinase Btk. Nat. Immunol. 12, 416-424 (2011). 7 Caton, M. L., Smith-Raska, M. R. & Reizis, B. Notch-RBP-J signaling controls the homeostasis of CD8- dendritic cells in the spleen. J. Exp. Med. 204, 1653-1664 (2007). 8 Loschko, J. et al. Absence of MHC class II on cDCs results in microbial- dependent intestinal inflammation. J. Exp. Med. 213, 517-534 (2016). 9 Labrecque, N., Madsen, L., Fugger, L., Benoist, C. & Mathis, D. Toxic MHC class II beta chains. Immunity 11, 515-516 (1999). 10 Hardy, R. R., Carmack, C. E., Shinton, S. A., Kemp, J. D. & Hayakawa, K. Resolution and characterization of pro-B and pre-pro-B cell stages in normal mouse bone marrow. J. Exp. Med. 173, 1213-1225 (1991). 11 Lu, L. S. et al. Identification of a germ-line pro-B cell subset that distinguishes the fetal/neonatal from the adult B cell development pathway. Proc. Natl. Acad. Sci. U. S. A. 99, 3007-3012 (2002). 12 Hayakawa, K., Tarlinton, D. & Hardy, R. R. Absence of MHC class II expression distinguishes fetal from adult B lymphopoiesis in mice. J. Immunol. 152, 4801-4807 (1994). 13 Lombard-Platet, S., Fisher, A. G., Meyer, V. & Ceredig, R. Expression of functional MHC class II molecules by a mouse pro-B cell clone. Dev. Immunol. 4, 85-92 (1995). 14 Busslinger, M. Transcriptional control of early B cell development. Annu. Rev. Immunol. 22, 55-79 (2004). 15 Clark, M. R., Mandal, M., Ochiai, K. & Singh, H. Orchestrating B cell lymphopoiesis through interplay of IL-7 receptor and pre-B cell receptor signalling. Nat. Rev. Immunol. 14, 69-80 (2014). 16 Cain, D., Kondo, M., Chen, H. & Kelsoe, G. Effects of acute and chronic inflammation on B-cell development and differentiation. J. Invest. Dermatol. 129, 266-277 (2009). 17 Wade, W. F., Freed, J. H. & Edidin, M. Translational diffusion of class II major histocompatibility complex molecules is constrained by their cytoplasmic domains. J. Cell. Biol. 109, 3325-3331 (1989). 18 Thibodeau, J., Bourgeois-Daigneault, M. C. & Lapointe, R. Targeting the MHC Class II antigen presentation pathway in cancer immunotherapy. Oncoimmunology 1, 908-916 (2012).

19 Johansson, P. et al. Deep sequencing of uveal melanoma identifies a recurrent mutation in PLCB4. Oncotarget 7, 4624-4631 (2016). 20 Dormer, P., Spitzer, E., Frankenberger, M. & Kremmer, E. Erythroid differentiation regulator (EDR), a novel, highly conserved factor I. Induction of haemoglobin synthesis in erythroleukaemic cells. Cytokine 26, 231-242 (2004). 21 Dormer, P., Spitzer, E. & Moller, W. EDR is a stress-related survival factor from stroma and other tissues acting on early haematopoietic progenitors (E- Mix). Cytokine 27, 47-57 (2004). 22 Rickert, R. C. New insights into pre-BCR and BCR signalling with relevance to B cell malignancies. Nat. Rev. Immunol. 13, 578-591 (2013). 23 Fisher, A. G., Burdet, C., Bunce, C., Merkenschlager, M. & Ceredig, R. Lymphoproliferative disorders in IL-7 transgenic mice: expansion of immature B cells which retain macrophage potential. Int. Immunol. 7, 415-423 (1995). 24 Hamson, E. J., Keane, F. M., Tholen, S., Schilling, O. & Gorrell, M. D. Understanding fibroblast activation protein (FAP): substrates, activities, expression and targeting for cancer therapy. Proteomics. Clin. Appl. 8, 454- 463 (2014). 25 Tran, E. et al. Immune targeting of fibroblast activation protein triggers recognition of multipotent bone marrow stromal cells and cachexia. J. Exp. Med. 210, 1125-1135 (2013). 26 Jones, E. Y., Fugger, L., Strominger, J. L. & Siebold, C. MHC class II proteins and disease: a structural perspective. Nat. Rev. Immunol. 6, 271-282 (2006). 27 Unanue, E. R., Turk, V. & Neefjes, J. Variations in MHC Class II Antigen Processing and Presentation in Health and Disease. Annu. Rev. Immunol. 34, 265-297 (2016). 28 Trowsdale, J. & Knight, J. C. Major histocompatibility complex genomics and human disease. Annu. Rev. Genomics. Hum. Genet. 14, 301-323 (2013). 29 Steidl, C. et al. MHC class II transactivator CIITA is a recurrent gene fusion partner in lymphoid cancers. Nature 471, 377-381 (2011). 30 Kambayashi, T. & Laufer, T. M. Atypical MHC class II-expressing antigen- presenting cells: can anything replace a dendritic cell? Nat. Rev. Immunol. 14, 719-730 (2014). 31 Johnson, D. B. et al. Melanoma-specific MHC-II expression represents a tumour-autonomous phenotype and predicts response to anti-PD-1/PD-L1 therapy. Nat. Commun. 7, 10582 (2016). 32 Kuppers, R. The biology of Hodgkin's lymphoma. Nat. Rev. Cancer 9, 15-27 (2009). 33 Nieto, M. A., Huang, R. Y., Jackson, R. A. & Thiery, J. P. EMT: 2016. Cell 166, 21-45 (2016).

FIGURE LEGENDS

Figure 1 | Initiation of MHC II expression shapes B-cell development potential. a, b, MHC II expression (a) and frequency (±s.e.m.) of MHC II- cells (b) in gated splenic DCs and B cells from mice of the indicated genotype. Symbols represent individual mice. *p<0.001 between WT and all other genotypes by one-way ANOVA. c, Mean change (±s.e.m.) from input frequency of CD45.2+ donor cells in splenic T cells or transitional (T), marginal zone (MZ), follicular (FO) and germinal centre

(GC) B cells from chimeras between CD45.1+ WT control and CD45.2+ test genotype donors (WT : WT, n=12; WT : ItgaxcreH2-Ab1c, n=6; WT : Fcer2acreH2-Ab1c, n=5). * p<0.01; **p<0.001 between WT and other genotypes by t-tests. d, MHC II expression in splenic FO and overlaid GL7+ GC B cells of host and donor origin in, distinguished by CD45 allotype expression. e, f, MHC II expression (e) and mean frequency

(±s.e.m.) of MHC II+ cells (f) in Fr A-F bone marrow B cells from mice of the indicated genotype (WT, n=3; ItgaxcreH2-Ab1c, n=5; Fcer2acreH2-Ab1c, n=6). g,

MHC II expression (green), detected by immunofluorescence staining in WT Fr A-C bone marrow B cells. ItgaxcreH2-Ab1c Fr A bone marrow B cells are included as staining control. h, Mean change (±s.e.m.) from input frequency of CD45.2+ cells in

Fr A-F bone marrow B cells from chimeras between CD45.1+ WT control and

CD45.2+ test genotype donors (WT : WT, n=9; WT : ItgaxcreH2-Ab1c, n=6; WT :

Fcer2acreH2-Ab1c, n=5). *p<0.001 between WT and other genotypes by t-tests.

Figure 2 | MHC II deficiency delays B-cell development progression. a, In vitro development of purified Fr A ItgaxcreH2-Ab1c B-cell precursors into pre-B

(pre; B220+CD2+IgD-IgM-) and immature B cells (imm; B220+CD2+IgD-IgM+), and

16 mean change (±s.e.m.) from input frequency of CD45.2+ cells in Fr A, pre-B cells and immature B cells following in vitro development of purified Fr A cells in competitive cultures between CD45.1+ WT control and CD45.2+ test genotype input Fr A cells

(WT : WT, n=6; WT : ItgaxcreH2-Ab1c, n=2; WT : Fcer2acreH2-Ab1c, n=8). *p<0.02 between WT and other genotypes by t-tests. b, Expression of differentially expressed genes (≥2-fold, p≤0.05) between MHC II- and MHC II+ Fr D WT B-cell precursor samples (left) and of selected MHC II, immunoglobulin and B-cell differentiation genes in MHC II- or MHC II+ WT and ItgaxcreH2-Ab1c Fr D cells (right). c,

Expression of differentially expressed genes (≥2-fold, q=0.05) between WT and

ItgaxcreH2-Ab1c Fr D B-cell precursor samples (top) and of the top 7 genes overrepresented in ItgaxcreH2-Ab1c Fr D cells (bottom). WT and ItgaxcreH2-Ab1c FO

B cells are also included for comparison. d, Functional pathways revealed by enrichment analysis on genes differentially expressed (≥2-fold, q=0.05) between WT and ItgaxcreH2-Ab1c Fr D cells. e, Correlation of the transcriptional signature of

ItgaxcreH2-Ab1c Fr D cells with those of CLP and Fr A cells, quantified with

ImmQuant.

Figure 3 | Ectopic MHC II expression is sufficient to drive B-cell development. a, Competitive in vitro differentiation assay between CD45.1+ WT HSCs transduced with GFP, CD45.2+ H2dlAb1-Ea HSCs transduced with GFP and CD45.2+ H2dlAb1-Ea

HSCs transduced with MHC II Ab1 and Aa chains. MHC II expression is shown 2 days after transduction of the respective HSCs, distinguished by GFP and CD45 allotype expression. b, Expression of B-cell lineage markers B220, CD24 and PB-1 and percentage of lineage-negative (Lin-) and Fr A-C cells in each of the HSCs donors described in (a), following growth in B-cell media (n=2). c, Amino acid sequence of

17 the transmembrane and intracellular domains of the MHC II Aa chain with the complete sequence (Aa) and with truncation of the intracellular domain (AaΔC). d,

MHC II expression in H2dlAb1-Ea HSCs transduced with GFP or with MHC II Ab1 and either Aa or AaΔC chains. e, Expression of B-cell lineage markers B220, IgM, CD2,

CD24 and PB-1 in each of the HSCs donors described in (e), following growth in B- cell media separately (n=2).

Figure 4 | MHC II controls growth and differentiation of transformed B cells. a, Frequency of MHC II- leukaemic cells recovered from Rag1-/- recipients that had received A1, F6 or G7 cells, sorted on the basis of MHC II expression prior to inoculation. b, F6 cells, transduced with either GFP of MHC II Ab1 and Aa chains were mixed in equal ratios and transferred into Rag1-/- recipients. GFP and MHC II expression in mixtures of F6 cells before (pre) and after (post) transfer in Rag1-/- recipients and frequency of GFP+ MHC II- leukaemic cells recovered from Rag1-/- recipients that had received mixtures of B3, D5, F6 or G7 cells transduced either GFP of MHC II Ab1 and Aa chains. *p<0.001 by t-tests. In (a and b), symbols represent individual recipient mice and group means (±s.e.m.) are also shown. c, B3 and F6 cells were transduced either with GFP or MHC II Ab1 and Aa chains. GFP+ and MHC

II+ cells from the same cell line were mixed at equal ratios and cocultured. Mean change (±s.e.m.) from starting frequency of MHC II- cells over time in such cocultures (B3, n=2; F6, n=4). *p=0.03; **p=0.08 by one-way ANOVA on ranks. d,

Growth rate, measured with the IncuCyte live-cell analysis system, of B3 or F6 cells transduced either with GFP or MHC II Ab1 and Aa chains. Symbols represent separate cultures. *p<0.001 by two-tailed Mann-Whitney Rank Sum tests. e, CD24 and BP-1 expression in parental F6 cells and those transduced either with GFP or

MHC II Ab1 and Aa chains. f, Expression of differentially expressed genes (≥2-fold, q=0.05) between MHC II- and MHC II+ F6 cells (left), expression of selected MHC II, immunoglobulin and B-cell differentiation genes in the same cells (middle), and expression of selected genes downregulated in MHC II+ F6 cells (right).

Figure 5 | Multipotent BMSC features in human B-cell malignancy. a, FAP expression in cell lines derived from the indicated B-cell leukaemias in RNA- seq data obtained from CCLE (HL, Hodgkin lymphoma; DLBCL, Diffuse large B- cell lymphoma; PCM, Plasma-cell myeloma; BL, Burkitt lymphoma; BCL un., B-cell lymphoma unspecified; B-ALL, Acute lymphoblastic B-cell leukaemia). b,

Expression of FAP and HLA II genes in the indicated HL cell lines in RNA-seq data obtained from CCLE. c, FAP and HLA II expression, determined by flow cytometry, in the indicated HL cell lines. Grey dots denote unstained control samples. d,

Expression of BMSC-specific and B-cell-specific genes in the indicated HL cell lines in RNA-seq data obtained from CCLE.

METHODS

Mice.

Inbred C57BL/6J (B6) and CD45.1+ congenic B6 (B6.SJL-Ptprca Pep3b/BoyJ) mice were originally obtained from The Jackson Laboratory. CD45.1+ CD45.2+ B6 mice were obtained by intercrossing B6 and CD45.1+ congenic B6 mice. Mice carrying a conditional H2-Ab1 allele (H2-Ab1c)1 were crossed to mice expressing Cre in DCs under the Itgax (CD11c-Cre driver; Itgaxcre)2 or the Zbtb46 promoter (zDC-Cre driver; Zbtb46cre)3, or in early and late B-cell lineages, under the Cd79a (mb1-Cre;

Cd79acre)4 and Fcer2a (CD23-Cre; Fcer2acre)5 promoter, respectively. Mice constitutively lacking all conventional MHC II genes (H2dlAb1-Ea)6, Rag1-deficient

(Rag1-/-) mice7 and Rag2-deficient (Rag2-/-) mice8, have been previously described.

All mouse strains were on the B6 genetic background and maintained at the Francis

Crick Institute’s animal facilities. All animal experiments were approved by the ethical committee of the Francis Crick Institute, and conducted according to local guidelines and UK Home Office regulations under the Animals Scientific Procedures

Act 1986 (ASPA).

Competitive bone marrow chimeras.

Bone marrow cell suspensions were prepared by flushing the bone cavities of femurs and tibiae from donor mice with air-buffered Iscove's Modified Dulbecco's Media

(IMDM). In certain experiments, T cells were depleted from bone marrow cell suspensions by immunomagnetic negative selection, using the EasySep system

(StemCell Technologies). Recipient mice were preconditioned by intraperitoneal injection of the myeloablative drug Busilvex (Pierre Fabre Ltd.), at a dose of 10 mg

1 per kg of body weight, dissolved in phosphate-buffered saline (PBS). They then received an intravenous injection of donor bone marrow-cell mixtures (a total of 107 cells in IMDM), 24 hours after myeloablation. The degree of reconstitution was assessed in recipient mice 8-10 weeks post bone marrow transplantation. The ratio of the two competing donor types in neutrophil or Fr A pre-B-cell progeny of donor bone marrow cells was taken to reflect the input ratio.

Flow cytometry and cell sorting.

Single-cell suspensions were stained for 20 min at room temperature or at 4°C with directly-conjugated antibodies to surface antigens, listed in Table S3. Immune-cell subtypes were identified using gating strategies depicted in Extended Data Fig. 14.

Multi-color cytometry were performed on LSRFortessa X-20 (BD Biosciences) flow cytometers and analysed with FlowJo v10 (Tree Star Inc.) analysis software. Cell populations of interest were purified (>98% purity) by cell sorting, performed on a

FACSAria Fusion flow cytometer (BD Biosciences) or MoFlo cell sorters (Dako-

Cytomation).

Large-scale fluorescence imaging.

Pro-B cells were enriched by cell sorting from WT or H2dlAb1-Ea bone marrow samples and were then washed and resuspended in 50 μl of warm Hanks' balanced salt solution

(HBSS), supplemented with 0.1% bovine serum albumin (BSA), before added to imaging chambers, pre-coated poly-lysine as previously described9. Cells were subsequently fixed in 2% paraformaldehyde (PFA) and stained with surface markers in order to distinguish Fr A-C. For intracellular staining, cells were then permeabilized using the FoxP3 fixation/permeabilization kit (eBioscience) on ice,

2 washed with permeabilization buffer, and then incubated with FITC-labelled anti-

Mouse MHC Class II (I-A/I-E) antibodies (Table S3). Non-specific binding of antibodies was blocked by incubation with 5% (v/v) mouse serum (Jackson

ImmunoResearch). All imaging was performed using a motorized Olympus IX81 microscope with a 100× oil objective (Olympus), controlled through Metamorph software (Molecular Devices). Images were processed and analysed by a user-guided pipeline implemented in Matlab (Mathworks) tethered to ImageJ by the MIJ plugin

(http://bigwww.epfl.ch/sage/soft/mij/), as previously described9.

Hematopoietic progenitor colony-forming unit (CFU) assay.

Hematopoietic progenitor cells in the bone marrow were quantified using the

MethoCult system (StemCell Technologies), according to the manufacturer’s instructions. Briefly, bone marrow cells were suspended in methylcellulose- containing media, supplemented with specific growth factors and cytokines that promote the growth of either myeloid progenitors (MethoCult GF M3434) or B-cell progenitors (MethoCult M3630). The total number or colonies, as well as blast sizes were scored on days 7 or 10, depending on the progenitor assay.

Competitive in vitro differentiation of Fr A pro-B cells.

Fr A pro-B cells were purified by cell sorting from CD45.1+ control (donor 1; always

WT) and CD45.2+ test genotype (donor 2; WT, ItgaxcreH2-Ab1c or Fcer2acreH2-Ab1c) bone marrow. Fr A cells from the two types of donor were then mixed at equal ratios and seeded onto confluent monolayers of B16.14 or ST2 stroma-cell lines10, pre- treated with mitomycin C, and cultured for 7-10 days in IMDM (Sigma-Aldrich), supplemented with 5% fetal bovine serum (Gibco, Life Technologies), 2 mM L-

3 glutamine, 100 U penicillin, and 0.1 mg/ml streptomycin. Differentiation of Fr A input cells was then assessed by flow cytometric analysis of cells at the end of the culture period.

Plasmids, transfections and retroviral transductions.

Plasmids containing the cDNA sequences of mouse H2-Aa and H2-Ab1, separated by an internal ribosome entry site (IRES) in the pRV retroviral vector, were synthesised and sequenced by GENEWIZ Inc. The codon encoding the first amino acid residue of the Aa chain cytoplasmic domain was mutation to a stop codon (CGA > TGA) using the QuikChange site-directed mutagenesis kit (Agilent Technologies). Retroviral particles were produced by transfection of pRV plasmids encoding MHC II chains or

GFP into the Platinum-E (Plat-E) retroviral packaging cell line, using the GeneJuice transfection reagent (Novagen). Plat-E culture supernatants containing viral stocks were harvested two days after the initial transfection and used for transduction of target cells. Cells were mixed with supernatants containing viral stocks in the presence of 4 μg/ml polybrene (Sigma-Aldrich) and spun at ~315×g for 45 min, before resuspended in IMDM for further culture. Where indicated, sublines homogenously expressing the transduced gene product were established by cell sorting based on that expression (either GFP or MHC II).

In vitro differentiation of purified HSCs.

HSCs were enriched from WT or H2dlAb1-Ea bone marrow-cell suspensions by immunomagnetic negative selection using the EasySep mouse hematopoietic progenitor-cell isolation kit (StemCell Technologies). Isolated HSCs were then rested for 4-6 hours in Roswell Park Memorial Institute (RPMI), supplemented with 10%

FCS and IL-3, IL-6 and SCF (all at 50ng/ml), before being retrovirally transduced.

For transduction, cells were resuspended in Plat-E culture supernatants containing viral stocks and spun onto RetroNectin plates (Clontech) at ~1,500×g for 2 hours.

Transduced cells were identified by expression of either GFP or MHC II 24-48 hours post transduction and were further enriched by cell sorting before being seeded into the relevant in vitro differentiation cultures. For B-cell development, 100,000-300,000 cells were added onto monolayers of irradiated B16.14 stromal cells, in RPMI supplemented with 10% FCS and SCF (20ng/ml) and IL-7 (10ng/ml). Five days later,

HSC progenitor cells were re-seeded onto fresh monolayers of irradiated of B16.14 cells and FLT3L (20ng/ml) was further added to the media. For DC development, transduced HSCs were cultured without feeder cells in IMDM supplemented with 5%

FCS and SCF (10ng/ml), FLT3L (100ng/ml) and GM-CSF (20ng/ml). Differentiation of DCs or B cells in these cultures was assessed at regular intervals by flow cytometry, between days 5 and 24 of differentiation.

Transcriptional profiling by RNA sequencing

Transcriptional profiles of primary Fr D and follicular B cells or F6 sublines were obtained by RNA sequencing (RNAseq). Fr D and follicular B cells from WT and

ItgaxcreH2-Ab1c donor mice were first purified by cell sorting (>98%). RNA was then extracted using the RNeasy Mini QIAcube or Micro kits (Qiagen) following the manufacturer’s instructions and subjected to RNA sequencing (GENEWIZ Inc.).

Reads were assessed for quality and contamination with FastQC v0.11.2

(http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and pre-processed with

Trimmomatic v0.32 (http://www.usadellab.org/cms/?page=trimmomatic) to remove the identified adapters, low quality read tails and to filter for length, and were then

5 assessed for differential expression with DESeq2 v1.6.3

(http://bioconductor.org/packages/release/bioc/html/DESeq.html) within R v3.1.3.

Data for all samples were normalized and log-transformed and were then analysed using Qlucore Omics Explorer 3.3 (Qlucore). Pathway analyses were performed using

The Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.8

(https://david.ncifcrf.gov/home.jsp) and g:Profiler (https://biit.cs.ut.ee/gprofiler)11.

Processed RNAseq data for expression of selected genes throughout murine B-cell development were obtained from the Immunological Genome Project Consortium

(ImmGen) (http://www.immgen.org)12. Transcriptional profiles of primary Fr D cells were compared with precompiled signatures of immune-cell types, using the

ImmQuant deconvolution software (http://csgi.tau.ac.il/ImmQuant). Transcriptional profiling of human hematopoietic cancer-cell lines was carried out on processed

RNAseq data obtained from the Cancer Cell Line Encyclopedia (CCLE) (available at https://portal.gdc.cancer.gov) or raw RNAseq data obtained, with permission, from the Genentech13 (available at European Genome-phenome Archive accession number:

EGAD00001000725). Data were analysed using the Qlucore Omics Explorer 3.3.

Expression analyses by quantitative real-time reverse transcription-based PCR.

The level of transcription of selected genes in pro-B-cell leukaemia lines was quantified by quantitative real-time reverse transcription-based PCR (qRT-PCR).

RNA preparation and DNase treatment were carried out using the RNeasy Mini

QIAcube kit (Qiagen) and cDNA produced with the High Capacity Reverse

Transcription kit (Applied Biosystems) with an added RNase-inhibitor (Promega

Biosciences). A final clean-up was performed with the QIAquick PCR purification kit

(Qiagen). RNase-free water (Qiagen) and later nuclease-free water (Qiagen) was used

6 throughout the protocol. Purified cDNA was then used as template for the amplification of target gene transcripts with SYBR Green PCR Master Mix (Applied

Biosystems), run on the ABI Prism SDS 7000 and 7900HT (Applied Biosystems) cyclers, with primers listed in Table S4. Data are plotted as expression of the target transcript, relative to expression of Hprt in the same sample, derived using the algorithm: value = 2^(CT value of Hprt – CT value of target).

TLR agonist stimulation.

WT and H2dlAb1-Ea bone marrow cells were stimulated overnight with the indicated doses of the TLR4-agonist LPS (from Salmonella minnesota R595, Axxora) or the

TRL2 agonist Pam3CSK4 (Axxora). Responses to stimulation were quantified by

CD69 expression 18-24 hours after stimulation, in Fr A-C cells distinguished by flow cytometry.

Cell-lines and growth.

Murine pro-B-cell leukaemia lines were established from spontaneous leukaemias in mice overexpressing IL-7 and have been previously described10. Human Hodgkin lymphoma cell-lines Hs 611.T (CVCL_0823), L-1236 (CVCL_2096) and Hs 616.T

(CVCL_0825) were obtained from the American Type Culture Collection (ATCC) and TO 175.T (CVCL_3806) was obtained from the German Collection of

Microorganisms and Cell Cultures (DSMZ). Cell-lines were mycoplasma-free and authenticated by short tandem repeat (STR) DNA profiling. Cell growth rates were quantified using the IncuCyte live-cell analysis system (Essen BioScience) following manufacturer’s protocols. F6 cells were grown without feeder cells, whereas B3 cells were grown either with or without B16.14 feeder cells with comparable results.

Statistical analyses.

Statistical comparisons were made using SigmaPlot 13 (Systat Software Inc.).

Parametric comparisons of normally-distributed values that satisfied the variance criteria were made by unpaired Student's t-tests or one-way ANOVAs. Data that did not pass the variance test were compared with non-parametric two-tailed Mann-

Whitney Rank Sum test or ANOVA on ranks tests. Calculation of correlation coefficients was performed using Excel 2016. Analysis of processed RNAseq data, hierarchical clustering and heat-map production was with Qlucore Omics Explorer 3.3

(Qlucore, Lund, Sweden).

Methods references

1 Hashimoto, K., Joshi, S. K. & Koni, P. A. A conditional null allele of the major histocompatibility IA-beta chain gene. Genesis 32, 152-153 (2002). 2 Caton, M. L., Smith-Raska, M. R. & Reizis, B. Notch-RBP-J signaling controls the homeostasis of CD8- dendritic cells in the spleen. J. Exp. Med. 204, 1653-1664 (2007). 3 Loschko, J. et al. Absence of MHC class II on cDCs results in microbial- dependent intestinal inflammation. J. Exp. Med. 213, 517-534 (2016). 4 Hobeika, E. et al. Testing gene function early in the B cell lineage in mb1-cre mice. Proc. Natl. Acad. Sci. U. S. A. 103, 13789-13794 (2006). 5 Kwon, K. et al. Instructive role of the transcription factor E2A in early B lymphopoiesis and germinal center B cell development. Immunity 28, 751-762 (2008). 6 Madsen, L. et al. Mice lacking all conventional MHC class II genes. Proc. Natl. Acad. Sci. U. S. A. 96, 10338-10343 (1999). 7 Mombaerts, P. et al. RAG-1-deficient mice have no mature B and T lymphocytes. Cell 68, 869-877 (1992). 8 Monroe, R. J. et al. RAG2:GFP knockin mice reveal novel aspects of RAG2 expression in primary and peripheral lymphoid tissues. Immunity 11, 201-212 (1999). 9 Nowosad, C. R., Spillane, K. M. & Tolar, P. Germinal center B cells recognize antigen through a specialized immune synapse architecture. Nat. Immunol. 17, 870-877 (2016). 10 Fisher, A. G., Burdet, C., Bunce, C., Merkenschlager, M. & Ceredig, R. Lymphoproliferative disorders in IL-7 transgenic mice: expansion of immature B cells which retain macrophage potential. Int. Immunol. 7, 415-423 (1995). 11 Reimand, J. et al. g:Profiler-a web server for functional interpretation of gene lists (2016 update). Nucleic. Acids. Res. 44, W83-89 (2016). 12 Heng, T. S. & Painter, M. W. The Immunological Genome Project: networks of gene expression in immune cells. Nat. Immunol. 9, 1091-1094 (2008). 13 Klijn, C. et al. A comprehensive transcriptional portrait of human cancer cell lines. Nat. Biotechnol. 33, 306-312 (2015).

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Extended Data Figure 1 | Skewed B-cell development of MHC II-deficient HSCs. a, Competitive bone marrow chimeras were setup by reconstituting myeloablated CD45.1+CD45.2+ WT (WT) hosts with mixtures of CD45.1+ control (donor 1; always WT) and CD45.2+ test genotype (donor 2; WT, ItgaxcreH2-Ab1c, Fcer2acreH2-Ab1c or H2dlAb1-Ea) bone marrow. b-e, B220 and MHC II expression and mean percentage of splenic T cells, B cells and myeloid cells originating from the host or each of the two types of donor, in competitive chimeras between WT and WT donors (n=5) (b); WT and ItgaxcreH2-Ab1c donors (n=8) (c); WT and Fcer2acreH2-Ab1c donors (n=7) (d) or WT and H2dlAb1-Ea donors (n=5) (e).

abWT : WT WT : H2dlAb1-Ea 40 40

* ** 20 ** 20 * ** ** ** ** *

0 0 cells, cells, change input from cells, change input from change cells,  

-20 -20 % of CD45.2 % of

-40 CD45.2 of % -40 T3 T1 T2 E F MZ C D FO A B GC

T cells Hardy’s Fr B cells

Extended Data Figure 2 | Competitive fitness of H2dlAb1-Ea HSCs. a, Mean change (±s.e.m.) from input frequency of CD45.2+ donor cells in splenic T cells or transitional (T), marginal zone (MZ), follicular (FO) and germinal centre (GC) B cells from chimeras between CD45.1+ WT control and CD45.2+ H2dlAb1-Ea donors (n=5). b, Mean change (±s.e.m.) from input frequency of CD45.2+ donor cells in Fr A-F bone marrow B cells from chimeras between CD45.1+ WT control and CD45.2+ H2dlAb1-Ea donors (n=10). *p<0.05; **p<0.005 by t-tests in both a and b.

WT ItgaxcreH2-Ab1c Fcer2acreH2-Ab1c

100 100

75 75 S

50 50

25 25 M/L Colony size distribution, % distribution, size Colony Number colonies CFU-B of Number 0 0

Extended Data Figure 3 | Comparable numbers of pre-B cell progenitors in MHC II-deficient bone marrow. Number of colony-forming B-cell (CFU-B) colonies (left) and colony size distribution (right) grown out of WT, ItgaxcreH2-Ab1c or Fcer2acreH2-Ab1c bone marrow cells in methylcellulose-based medium. Symbols represent replicate cultures. S, small; M/L, medium/large.

Extended Data Figure 4 | CD4+ T-cell-independent competitive advantage of MHC II-deficient B-cell precursors. a, Competitive bone marrow chimeras were setup by reconstituting myeloablated CD45.2+ Rag1-/- or MHC II-deficient Rag1-/- H2dlAb1-Ea hosts with a mixture of equal proportions of CD45.1+ WT control (donor 1) and CD45.2+ ItgaxcreH2-Ab1c (donor 2) T-cell-depleted bone marrow. Rag1-/- HSCs are unable to generate T or B cells and lack of MHC II expression in Rag1-/- H2dlAb1-Ea hosts prevents CD4+ T cell development from donor HSCs. b, CD4 and CD8 expression in the B220- fraction of spleen cells (non-B cells) from the indicated host. c, Mean frequency (±s.e.m.) of CD45.2+ ItgaxcreH2-Ab1c donor cells in TCRβ+ thymocytes or B220+ splenic B cells in Rag1-/- (n=4) or Rag1-/- H2dlAb1-Ea hosts (n=6). d, MHC II expression in CD45.1+ WT and CD45.1- ItgaxcreH2-Ab1c donor splenic B cells from the indicated host.

Extended Data Figure 5 | Advanced differentiation of MHC II+ Fr D WT B-cell precursor. a, Expression of differentially expressed genes (≥2-fold, p≤0.05) between MHC II- and MHC II+ Fr D WT B-cell precursors and hierarchical clustering with MHC II+ FO B-cells from the same mice. b, Expression of the top 7 B-cell differentiation-related genes in MHC II- and MHC II+ Fr D cells, in comparison with FO B cells.

pro B - CLP pro B - Fr A pro B - Fr B/C Fr E Splenic B cells a 40 Ackr2 20

0 b 400 100 Aicda Cox6a2 200 50

0 0 60 200 Ccr6 Itgam 30 100

0 0 10 40 Crip3 Pla2g4c 5 20

0 0 400 40 Fcrl5 Plcb4 200 20

0 0 60 40 Il9r Itgad

Expression Value Normalized by DESeq2 Normalized Value Expression 30 20

0 by DESeq2 Normalized Value Expression 0 60 10 Pltp Gm21887 30 5

0 0

Extended Data Figure 6 | Stemness and differentiation gene expression throughout B-cell development. a, b, Expression level of selected B-cell differentiation (a) and stem-cell (b) genes in the indicated stages of B-cell development. Data were obtained fom ImmGen. Gm21887 is the X-linked syntenic copy of the Y-linked Erdr1 gene.

Extended Data Figure 7 | Retention of stemness features in MHC II-deficient B- cell precursors. Correlation of the transcriptional signature of WT and ItgaxcreH2- Ab1c Fr D cells with pre-compiled signatures of the defined stages of B-cell development. Cell-type quantities were calculated using the ImmQuant deconvolution software.

Extended Data Figure 8 | Competitive advantage of MHC II-deficient B-cell precursors, independently of BCR signalling. Competitive bone marrow chimeras were setup by reconstituting myeloablated CD45.1+ CD45.2+ WT hosts with a mixture of equal proportions of CD45.1+ Rag2-/- control (donor 1) and either CD45.2+ Rag1-/- or Rag1-/- H2dlAb1-Ea bone marrow (donor 2). Mean change (±s.e.m.) from input frequency of CD45.2+ donor cells in Fr A-C bone marrow B cells from chimeras between Rag2-/- and Rag1-/- donors (n=4) and between Rag2-/- and Rag1-/- H2dlAb1-Ea donors (n=18). *p<0.02 by t-tests.

Extended Data Figure 9 | Expression of genes essential for B-cell development is unaffected by MHC II deficiency. Expression of genes known to promote B-cell development in WT and ItgaxcreH2-Ab1c Fr D B-cell precursors, determined by RNA- seq in this study. WT and ItgaxcreH2-Ab1c FO B cells are also included for comparison.

WT H2dlAb1-Ea

Fr A Fr B Fr C 100

cells, % cells, 50 

CD69 25

0 -2 -1 0 1 -2 -1 0 1 -2 -1 0 1 PamCys3, Log10 mg/ml

100

cells, % 50 

CD69 25

0 -2 -1 0 1 -2 -1 0 1 -2 -1 0 1 LPS, Log10 mg/ml

Extended Data Figure 10 | TLR agonist responsiveness of B-cell precursors is unaffected by MHC II deficiency. Mean frequency (±s.e.m.) of CD69+ cells according the dose of TLR2 agonist PamCys3 or TLR4 agonist lipopolysaccharide (LPS) of Fr A, B or C WT (n=2) or H2dlAb1-Ea (n=2) B-cell precursors.

Extended Data Figure 11 | Characteristics of murine B-cell precursors leukaemia cell lines. a, Developmental stage of the indicated cell lines, according to expression of B220, IgM, CD2, CD24 and PB-1. b, B220 and MHC II expression in the indicated cell lines. c, Expression of genes specific to MHC II-deleted non-transformed pre-B cells, assessed by qRT-PCR.

Fap LTHSC CD34 10 LTHSC CD34 STHSC MMP2 MMP3 MMP4 5 pro B - CLP pro B - Fr A pro B - Fr B/C Fr E Expression Value Expression 0 Splenic B cells Normalized DESeq2by Normalized

Extended Data Figure 12 | Fap expression throughout B-cell development. Expression level of Fap in the indicated stages of B-cell development. Data were obtained from ImmGen.

HL ‐ CCLE B‐cell and myeloid ‐ CCLE B‐cell and myeloid ‐ Genentech

Hs 616.T 6 6 6 2 2 TO 175.T 2 SCC-3 Hs 751.T 4 4 4

L-1236 HD-MY-Z U-937 KARPAS-1106P mRNA, Log mRNA, mRNA, Log mRNA, 2 2 Log mRNA, 2 U-937 PL-21 FAP FAP SU-DHL-1 FAP SU-DHL-1 0 0 0 0204060 0204060 0204060 HLA II mRNA, Log2 HLA II mRNA, Log2 HLA II mRNA, Log2

Extended Data Figure 13 | Anti-correlation of FAP and HLA II gene expression. Expression of FAP mRNA plotted against the combined expression of all HLA II genes in HL cell-lines from CCLE (left), other B-lymphoid or myeloid cell-lines for CCLE (middle) or other B-lymphoid or myeloid cell-lines for Genentech (right). Cell- lines positive for FAP expression are indicated. Spearman's Rank correlation coefficient for HL cell-lines is -0.923.

Extended Data Figure 14 | Flow cytometric gating of immune cell populations. a, Gating of splenic B cells and DCs. b, Gating of splenic transitional (T) 1, T2 T3, marginal zone (MZ) and follicular (FO) B cells. c, Gating of splenic germinal centre (GC) B cells. d, Gating of bone marrow Fr A-F B cells.