Hematopoietic RIPK1 deficiency results in bone marrow failure caused ...

Hematopoietic RIPK1 deficiency results in bone marrow failure caused by apoptosis and RIPK3-mediated necroptosis Justine E. Rodericka, Nicole Hermancea, Matija Zelica, Matthew J. Simmonsb, Apostolos Polykratisc,d, Manolis Pasparakisc,d, and Michelle A. Kellihera,1 a

Department of Cancer Biology, University of Massachusetts Medical School, Worcester, MA 01605; bBIND Biosciences, Inc., Cambridge, MA 02139; and Institute for Genetics, Centre for Molecular Medicine and dCologne Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases, University of Cologne, 50674 Cologne, Germany c

Edited by Tak W. Mak, The Campbell Family Institute for Breast Cancer Research at Princess Margaret Cancer Centre, Ontario Cancer Institute, University Health Network, Toronto, ON, Canada, and approved August 12, 2014 (received for review May 31, 2014)

Receptor-interacting serine/threonine-protein kinase 1 (RIPK1) is recruited to the TNF receptor 1 to mediate proinflammatory signaling and to regulate TNF-induced cell death. RIPK1 deficiency results in postnatal lethality, but precisely why Ripk1−/− mice die remains unclear. To identify the lineages and cell types that depend on RIPK1 for survival, we generated conditional Ripk1 mice. Tamoxifen administration to adult RosaCreERT2Ripk1fl/fl mice results in lethality caused by cell death in the intestinal and hematopoietic lineages. Similarly, Ripk1 deletion in cells of the hematopoietic lineage stimulates proinflammatory cytokine and chemokine production and hematopoietic cell death, resulting in bone marrow failure. The cell death reflected cell-intrinsic survival roles for RIPK1 in hematopoietic stem and progenitor cells, because Vav-iCre Ripk1fl/fl fetal liver cells failed to reconstitute hematopoiesis in lethally irradiated recipients. We demonstrate that RIPK3 deficiency partially rescues hematopoiesis in Vav-iCre Ripk1fl/fl mice, showing that RIPK1-deficient hematopoietic cells undergo RIPK3-mediated necroptosis. However, the Vav-iCre Ripk1fl/fl Ripk3−/− progenitors remain TNF sensitive in vitro and fail to repopulate irradiated mice. These genetic studies reveal that hematopoietic RIPK1 deficiency triggers both apoptotic and necroptotic death that is partially prevented by RIPK3 deficiency. Therefore, RIPK1 regulates hematopoiesis and prevents inflammation by suppressing RIPK3 activation. tumor necrosis factor

16). RIPK1 also is recruited to the Toll-like receptor adapter TRIF via the Rip homotypic interaction motif (RHIM) to mediate NFκB activation (17) and, under conditions of caspase-8 inhibition, initiates necroptosis (14, 16). Necrostatin-1 (Nec-1), an allosteric RIPK1 inhibitor, inhibits necroptosis induced by TNF or the TLR3 ligand poly I:C and abolishes the formation and activation of an RIPK1/3 complex (13–16, 18). Although the molecular details whereby RIPK1 initiates necroptosis are unclear, RIPK3 and the pseudo kinase MLKL appear to be required (2). Genetic studies in mice have revealed cross-regulation between the apoptotic and necroptotic pathways. For example, the FADD/caspase-8/FLICE-like inhibitory protein long form (FLIPL) complex regulates RIPK1 and RIPK3 activity during development, because the embryonic lethality associated with a caspase-8 deficiency is completely rescued by the absence of RIPK3 (19, 20). Similarly, RIPK1 deficiency rescues FADDassociated embryonic lethality (21). Thus, in the absence of FADD or caspase-8, embryos succumb to RIPK1- and RIPK3dependent necroptosis. However, Fadd−/−/Ripk1−/− mice, die perinatally (21, 22), as do Ripk1−/− mice, revealing that RIPK1 has prosurvival roles beyond the regulation of the FADD/ caspase-8/FLIPL complex.

| programmed necrosis | hematopoietic failure

Significance Receptor-interacting serine/threonine-protein kinase 1 (RIPK1) is involved in TNF signaling and interacts with the related RIPK3 to regulate cell death and inflammation. RIPK1 has kinase-independent prosurvival and kinase-dependent prodeath functions. To identify the lineages that depend on RIPK1 for survival, we generated conditional Ripk1 mice. Acute Ripk1 deletion results in rapid death of the animal caused by extensive cell death in the intestinal and hematopoietic lineages. A hematopoietic RIPK1 deficiency stimulates proinflammatory cytokine/chemokine production and cell death, resulting in bone marrow failure. Hematopoietic failure is partially rescued by a RIPK3 deficiency, indicating that RIPK1-deficient hematopoietic cells undergo RIPK3-mediated necroptosis. These findings show that in the hematopoietic lineage RIPK1 suppresses RIPK3 activity and suggest that RIPK-dependent necroptosis may contribute to human bone marrow failure syndromes.

T

he proinflammatory cytokine TNF stimulates receptorinteracting serine/threonine-protein kinase 1 (RIPK1) ubiquitination, NFκB and MAPK activation, and induction of apoptosis or necroptosis (1, 2). TNF signaling via TNF receptor 1 (TNFR1) is highly regulated and results in the recruitment of several adapter proteins including TNFR1-associated death domain (TRADD) protein, the E3 ubiquitin ligases cellular inhibitor of apoptosis protein-1 and -2 (cIAP1/2), and TNFRassociated factor 2 (TRAF2) or 5, and the serine threonine death domain-containing kinase RIPK1 (complex I) (1). We have demonstrated that the kinase activity of RIPK1 is not required for NFκB activation (3); rather, RIPK1 is modified by the addition of Lys63-linked and linear polyubiquitin chains (3–6). Polyubiquitinated RIPK1 then recruits NEMO/IκB kinase-γ (IKKγ) to mediate IKK activation and TAK1/TAB2/3 to mediate MAPK activation, resulting in antiapoptotic and proinflammatory gene expression (7, 8). Deubiquitination of RIPK1 by cylindromatosis (CYLD) results in the formation of a cytosolic complex containing TRADD, Fas-associated death domain protein (FADD), caspase-8, and RIPK1 (complex IIa) (2). Caspase-8 cleaves and inactivates RIPK1 and CYLD and stimulates apoptosis (9–11). In the absence of caspase-8 or the presence of caspase inhibitors, TNF family members and potentially other ligands stimulate the kinase activity of RIPK1 to induce necroptosis (9, 11– 14436–14441 | PNAS | October 7, 2014 | vol. 111 | no. 40

Author contributions: M.A.K. designed research; J.E.R., N.H., and M.Z. performed research; N.H., M.J.S., A.P., and M.P. contributed new reagents/analytic tools; J.E.R. and N.H. analyzed data; and M.A.K. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1

To whom correspondence should be addressed. Email: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1409389111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1409389111

Roderick et al.

Floxed Ripk1 Allele

2

Deleted Allele Exon3

B

RosaCreERT2

3 2

4

5

4

5

RosaCreERT2 Ripk1fl/fl

C

RosaCreERT2 RosaCreERT2 Ripk1fl/fl

Ileum

40

Weight (g)

30

*

**

20

Colon

10 0

Males

Thymus

Absolute Cell Number

D

Spleen

2x10 5 1x10 5

** 0

L-S-K

E

H&E

Females


3x10 5

Total Colony Number

100

***

**

***

***

L-SK LT-HSC ST-HSC MPP


80 60 40

*

20 0

20x

Fig. 1. Acute ablation of RIPK1 results in intestinal epithelial and hematopoietic cell loss. (A) Schematic depiction of the gene-targeting strategy used to generate the conditional Ripk1 allele. (B) Representative H&E sections of intestines and hematopoietic organs from the indicated genotypes. (C) Total body weight of RosaCreERT2 (n = 10) and RosaCreERT2 Ripk1fl/fl (n = 9) mice. (D) Graph depicting absolute numbers of HSPCs from RosaCreERT2 (n = 6) and Rosa-CreERT2 Ripk1fl/fl (n = 6) mice. (E) Colony-forming assay of bone marrow from Rosa-CreERT2 (n = 2) and RosaCreERT2 Ripk1fl/fl (n = 2) mice. All data were collected 24 h after the third tamoxifen injection. Error bars represent SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Hematopoietic cell death observed upon acute ablation of RIPK1 may be a secondary effect resulting from the inflammation induced by cell death in the intestine and potentially other lineages. Consistent with this possibility, serum proinflammatory chemokine and cytokine levels were increased significantly in tamoxifen-treated RosaCreERT2 Ripk1fl/fl mice but not in controls (Fig. S2). To exclude the potential confounding effects of a proinflammatory microenvironment on HSPC survival, we used the Vav-iCre transgene to delete Ripk1 in hematopoietic cells. Half of the Vav-iCre Ripk1fl/fl mice exhibited histopathological features associated with complete RIPK1 deficiency and died during the postnatal period, likely as the result of reported Vav-iCre activity in the germ line (28). The remaining Vav-iCre Ripk1fl/fl mice survived beyond the postnatal period (average survival 35 d) but appeared smaller and weighed significantly less than Vav-iCre Ripk1fl/+ littermate controls (Fig. 2 A and B). Ripk1 deletion was detected in bone marrow, spleen, and thymus of these mice but not in nonhematopoietic tissues (Fig. S3A). The Vav-iCre Ripk1fl/fl mice developed marked pancytopenia and anemia with an average 38-fold decrease in bone marrow cellularity (Fig. 2 C–E and Fig. S3). Histological examination revealed hypocellular bone marrow and extensive cell death in the thymus and spleen of Vav-iCre Ripk1fl/fl mice compared with Vav-iCre Ripk1fl/+ controls (Fig. 2 D and E and Fig. S3E). Lineage PNAS | October 7, 2014 | vol. 111 | no. 40 | 14437

CELL BIOLOGY

Results To identify the cell types and lineages in the adult dependent on RIPK1 for survival, we mated conditional Ripk1 mice to the tamoxifen-inducible Rosa26CreERT2 transgenic mice (27) (Fig. 1A and Fig. S1A). Administration of three daily injections of 1 mg of tamoxifen resulted in efficient RIPK1 depletion in RosaCreERT2 Ripk1fl/fl mice and in the rapid induction of pathology in the intestinal and hematopoietic lineages (Fig. 1 and Fig. S1). RosaCreERT2 Ripk1fl/fl mice treated with tamoxifen developed diarrhea and weight loss and were killed 24 h after the third tamoxifen injection (Fig. 1 B and C). Histological analysis revealed loss of bone marrow cellularity and erosion of villus structures with pronounced loss of intestinal epithelial cells in the ileum and colon (Fig. 1B) Increased numbers of dying cells that stained positive for cleaved caspase-3 were detected in the thymus, spleen, and small and large intestines of RosaCreERT2 Ripk1fl/fl mice, indicating that RIPK1 deficiency results in caspase-dependent apoptosis in these lineages (Fig. S1 E and F). Cell death in thymus and spleen resulted in significant decreases in overall thymic and splenic cellularity (Fig. S1H), features observed in mice with complete RIPK1 deficiency (22). Cell death and/or inflammation was not evident in the other organs examined (Fig. S1G), nor was there evidence of anemia, indicating that intestinal epithelial cell loss is likely the major contributing factor in the death of the tamoxifen-treated RosaCreERT2 Ripk1fl/fl mice. Genomic PCR confirmed Ripk1 deletion in the intestines and the hematopoietic organs of tamoxifen-treated RosaCreERT2 Ripk1fl/fl mice (Fig. S1 B–D). Tamoxifen-treated RosaCreERT2 mice did not develop intestinal or hemato-pathology. RIPK1 deficiency resulted in decreases in the absolute number of lymphoid and myeloid cells (Fig. S1 I–K). The Lin−Sca1+cKit+ (L−SK) hematopoietic stem and progenitor cells (HSPCs) and the Lin− Sca1− cKithigh (L− S− K) myeloidenriched progenitors also were depleted significantly (ninefold) in tamoxifen-treated RosaCreERT2 Ripk1fl/fl mice as compared with tamoxifen-treated RosaCreERT2 controls (Fig. 1D). The LSK population was analyzed further by staining with CD34 and fetal liver kinase 2 (Flk2) antibodies to quantify the longand short-term hematopoietic stem cells (HSCs) and multipotent progenitors (MPPs), all of which were reduced significantly in RosaCreERT2 Ripk1fl/fl mice (Fig. 1D). To examine stem and progenitor cell function, we plated bone marrow from tamoxifen-treated RosaCreERT2 Ripk1fl/fl or RosaCreERT2 control mice in an in vitro hematopoietic colony assay and observed >75% reduction in progenitor activity in the RIPK1-deficient cultures (Fig. 1E). Collectively, these data show that acute ablation of RIPK1 in adult mice results in intestinal epithelial and hematopoietic cell death, revealing crucial prosurvival roles for RIPK1 in these lineages.

FRT Site

A

Bone marrow

We have demonstrated that complete RIPK1 deficiency results in increased TNF-induced cell death that can be rescued, in part, by the absence of the TNFR1 (22, 23). However, Ripk1−/−Tnfr1−/− animals still succumb (23), indicating that other death ligands/ pathways contribute to the RIPK1-associated lethality. Consistent with this hypothesis, RIPK3 deficiency recently has been shown to rescue the perinatal lethality observed in Ripk1−/−Tnfr1−/− mice (24, 25). Similarly, combined caspase-8 and RIPK3 deficiency also rescues the RIPK1-associated lethality (24–26). Collectively, these genetic studies in mice reveal that the perinatal death of Ripk1−/− mice reflects TNF-induced apoptosis and RIPK3-mediated necroptosis. The nature of the ligand(s) or the trigger(s) of RIPK3mediated necroptosis in vivo remain unclear. However, Ripk1−/− MEFs are prone to necroptosis induced by poly I:C or by treatment with type I or type II IFN (24, 25), suggesting that these pathways contribute. Although these studies reveal a regulatory role for RIPK1, the multiorgan cell death and inflammation observed in the complete and compound RIPK1-knockout strains have made it difficult to discern the specific tissues that require RIPK1 for survival.

Weight (g)

20 15

***

10

80 60

0 0

0

0

0

30 20

***

40

50

Vav-iCre Ripk1fl/+ Vav-iCre Ripk1fl/fl

1 108

***

10

0

0

30

Days

*** ***

Thymus Spleen

Spleen

B.M.

Bone Marrow

Vav-iCre Ripk1fl/+

Thymus

***

20

40

Cellularity

5

10

2 108

Vav-iCre Ripk1fl/fl

E

10

Hematocrit (%)

5

RBC (M/uL)

WBC (K/uL)

***

D

50

15

10

p=.0004

20

5

C 15


40

H&E

10x

Fig. 2. Deletion of RIPK1 in the hematopoietic lineage results in acute lethal bone marrow failure. (A) Macroscopic images of indicated genotypes (Left) and total body weight when killed (Right). (B) Kaplan–Meier survival curve of Vav-iCre Ripk1fl/fl mice (n = 14) and Vav-iCre Ripk1fl/+ (n = 11) controls (average latency 35 d, P < 0.0004). (C) Blood cell counts and hematocrits for Vav-iCre Ripk1fl/+ (n = 11) and Vav-iCre Ripk1fl/fl (n = 14) mice. (D) Graph showing thymus, spleen, and bone marrow cellularity for Vav-iCre Ripk1fl/+ (n = 11) and Vav-iCre Ripk1fl/fl (n = 14) mice. (E) Representative histological images of hematopoietic organs stained with H&E. Error bars represent SEM. ***P < 0.001.

analysis by flow cytometry confirmed decreases in the absolute numbers of lymphoid cells and myeloid cells in Vav-iCre Ripk1fl/fl mice compared with littermate controls (Fig. S3 B–D). Despite the hematopoietic cell loss, some of the Vav-iCre Ripk1fl/fl mice exhibited mild to moderate inflammation characterized by focal areas of granulocyte infiltration in skin or liver (Fig. S3F). These histologic findings indicate that hematopoietic RIPK1 deficiency can result in tissue inflammation. These mouse genetic studies using hematopoietic or inducible Cre recombinase reveal crucial roles for RIPK1 in hematopoietic cell survival. To reveal the kinetics of hematopoietic cell death, we examined 2-wk-old Vav-iCre Ripk1fl/fl mice for evidence of hematopoietic deficiency. We detected no significant decreases in peripheral blood counts or in the cellularity of the bone marrow or spleen, although modest decreases in the absolute numbers of T and B cells were evident in thymus, spleen, and bone marrow (Fig. S4). However, statistically significant decreases in absolute number of HSPCs and in colony-forming activity were observed (Fig. 3 A and C). HSPC number and colony formation in Vav-iCre Ripk1fl/fl mice were reduced further at 5 wk (Fig. 3 B and C). The hematopoietic cell loss observed in these mice was associated with increases in serum chemokines including interferon gamma-induced protein-10 (IP-10), keratinocyte chemoattractant (KC), monocyte chemoattractant protein-1 (MCP-1), and monokine induced by γIFN (MIG), and proinflammatory cytokines (TNF-α, IFN-γ, IL-6, and G-CSF) (Fig. S5). The anti-inflammatory cytokine IL-10 also was up-regulated, whereas the IFN-γ–regulated chemokine LPS-induced CXC chemokine (LIX) (the mouse ortholog of CXCL5) was 14438 | www.pnas.org/cgi/doi/10.1073/pnas.1409389111

significantly down-regulated in Vav-iCre Ripk1fl/fl mice; these features also are observed in human patients with bone marrow failure (29–31). These data suggest that RIPK1 deficiency sensitizes HSPCs to cytokine-mediated cell death. To test this hypothesis, we isolated bone marrow from 14-d-old Vav-iCre Ripk1fl/fl mice and added TNF, TNF-related apoptosis-inducing ligand (TRAIL), or type I or type II IFN to the hematopoietic colony assays. Colony number and size were reduced significantly when the RIPK1-deficient progenitors were cultured with TNF or with type I or type II IFN but not when cultured with TRAIL; these findings indicate that, in addition to TNF, RIPK1-deficient hematopoietic progenitors also are sensitive to cell death induced by type I or type II IFN (Fig. 3D). RIPK1 deficiency had clear effects on all the hematopoietic lineages examined (Figs. 2 and 3 and Fig. S3). Mice deficient in RIPK1 also were significantly smaller than littermate controls (Fig. 2A), and most of these mice exhibited focal areas of inflammation (Fig. S3). To determine whether RIPK1 has a hematopoietic cell-intrinsic survival function, we transplanted embryonic day 14.5 fetal liver cells from Vav-iCre Ripk1fl/fl mice and littermate controls into lethally irradiated syngeneic recipients and monitored the relative contributions of donor and host cells to hematopoiesis (Fig. 4A). Mice transplanted with RIPK1-deficient fetal liver cells exhibited significant reductions in the HSPC population as early as 2 wk posttransplantation (Fig. 4B). Eight weeks after transplantation, the RIPK1-deficient HSPC populations were nearly undetectable in the transplanted mice (Fig. 4C), confirming that RIPK1 has a cell-intrinsic survival function in HSPC. Although cleaved caspase-3+ hematopoietic cells were detected in the thymus and spleen, dying cells that were negative for cleaved caspase-3 were observed in the RIPK1-deficient bone marrow and spleen, suggesting that some hematopoietic cells die from necrosis rather than apoptosis. We hypothesized that the HSPC death in Vav-iCre Ripk1fl/fl mice could reflect RIPK3dependent necroptosis. To address the potential role of RIPK3, we generated and analyzed Vav-iCre Ripk1fl/fl Ripk3−/− mice. In contrast to Ripk1−/−Ripk3−/− double-knockout mice, which die during the postnatal period (24–26), Vav-iCre Ripk1fl/fl Ripk3−/− mice develop normally and reach adulthood without showing any macroscopic or histologic abnormalities (Fig. 5 A, B, and E). Remarkably, RIPK3 deficiency rescued the pancytopenia, anemia, and bone marrow hypocellularity associated with

A 3x10 5

Vav-iCre Ripk1fl/+

2x10 5

fl/fl

1x10 5

C Total Colony Number

25

100

Vav-iCre Ripk1

***

3x10 4 2x10 4

**

1x10 4 0

L-S-K

*

***


B 1x10 6

Vav-iCre Ripk1fl/+

8x10 5

Vav-iCre Ripk1fl/fl

6x10 5 8x10 4 4x10 4 0

* *** L-S-K

*

**

***


D

Normalized Colony Number

B




Vav-iCre Vav-iCre Ripk1 f l / + Ripk1 fl/f l

Event Free Survival

A

80


60 40

**

20 0

1.5

*** Day 14

Day 35


1.0

**

0.5

*** 0.0

Vehicle TRAIL

IFN-γ

** IFN-α

TNF-α

fl/fl

Fig. 3. Bone marrow failure arises in Vav-iCre Ripk1 mice as the result of HSPC loss. (A and B) Graphs depicting absolute numbers of HSPCs in Vav-iCre Ripk1fl/+ (n = 5) and Vav-iCre Ripk1fl/fl (n = 6) mice at 14 d old (A) or in Vav-iCre Ripk1fl/fl (n = 3) and Vav-iCre Ripk1fl/+ (n = 3) mice at the time of disease (B). (C) Graph depicting hematopoietic colony potential at 14 d old or at time of disease (n = 4 for each genotype). (D) Graph showing total colony number in the presence of vehicle, TRAIL, IFN-γ, IFN-α, or TNF-α for Vav-iCre Ripk1fl/+ (n = 3) and VaviCre Ripk1fl/fl (n = 3) mice. Total colony number was normalized to vehicle-treated bone marrow. Error bars represent SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Roderick et al.

mediated necroptosis (14, 33), we found that RIPK1 is not required to execute necroptotic hematopoietic cell death in vivo. Rather, our data suggest that RIPK1 functions in the hematopoietic lineage to prevent RIPK3-initiated necroptosis. Consistent with the notion that necroptosis constitutes an inflammatory form of cell death, we found serum cytokines and chemokines (TNF-α, IFN-γ, IL-6, IP-10, KC, MCP-1, MIG, IL-10, and G-CSF) up-regulated in Vav-iCre Ripk1fl/fl mice (Fig. S5); however, cytokine/chemokine levels were significantly reduced in Vav-iCre Ripk1fl/fl Ripk3−/− mice, in which necroptotic death was prevented (Fig. 5 F and G). Furthermore, Vav-iCre Ripk1fl/fl Ripk3−/− HSPCs no longer appeared as sensitive to γ- or α-IFN–induced cell death but retained sensitivity to TNF-induced apoptosis (Fig. 5H). These in vitro colony data suggest that RIPK1-deficient HSPCs undergo TNF-induced apoptosis and RIPK3-mediated necroptosis potentially triggered by TNF and/or type I or type II interferons. Analysis of the LSK population in Vav-iCre Ripk1fl/fl Ripk3−/− mice revealed that the absence of RIPK3 did not completely rescue the absolute numbers of HSPC or in vitro colony formation (Fig. 5I and Fig. S6F), indicating that hematopoietic stem and progenitor repopulating capability may be compromised. To test this notion directly, we transplanted bone marrow from VaviCre Ripk1fl/fl Ripk3−/− mice and littermate controls into lethally irradiated syngeneic recipients and monitored survival. In contrast to controls, five of six of the mice transplanted with

WT (CD45.1)

5

Vav-iCreRipk1fl/+

C 1.5x10

fl/fl

Vav-iCreRipk1 1x10 5

5x10 4 0

* * L-S-K

*

*

L-SK LT-HSC ST-HSC

**

6

Vav-iCreRipk1fl/+ Vav-iCreRipk1fl/fl

1.0x10 6 5.0x10 5 5x10 4

* 0

*

L-S-K

MPP

**

*

*

L-SK LT-HSC ST-HSC

MPP

Fig. 4. RIPK1-deficient HSPCs fail to reconstitute lethally irradiated recipients. (A) Experimental design. (B and C) Graphs showing absolute number of donorderived, CD45.2+ HSPCs from mice transplanted with Vav-iCre Ripk1fl/+ or VaviCre Ripk1fl/fl fetal liver cells at 2 wk (B) and 8 wk (C) posttransplantation. Error bars represent SEM. *P < 0.05, **P < 0.01.

hematopoietic RIPK1 deficiency (Fig. 5 C–E). These genetic data reveal that, in the absence of RIPK1, hematopoietic cells undergo RIPK3-mediated necroptosis. Thus, in contrast to published studies showing that RIPK1 is essential for TNF- and Fas-mediated (9, 12, 13, 24, 25, 32) as well as for TLR3/4-

A

Vav-iCre Ripk1f l / + Ripk3-/-

B

Vav-iCre Ripk1fl/f l Ripk3-/-

C ***

25

***

Vav-iCre Ripk1 fl/+ Vav-iCre Ripk1 fl/fl

***

***

15

15

60

15 10

***

*** RBC (M/uL)

WBC (K/uL)

Weight (g)

20 10

5

10

5

5

Vav-iCre Ripk1fl/fl

Thymus

5000

4000

(pg/mL)

3000

*

2000 1000

Bone Marrow

Normalized Total Colony Number

H

H&E

20x

1.5

IP-10

KC

MCP-1

MIG

2000

4 107 2 107

0

0

40000

*

*

***

30000

***

150 100

20000 10000

IFN-α

IFN-γ

TNF-α

TNF-α

IL-6

IL-10

IL-1α

Vav-iCre Vav-iCre Ripk1fl/flRipk3-/-

2 105 1 105

*** L-S-K

*

*

*


G-CSF

Vav-iCre Ripk1 fl/+ Ripk3 -/Vav-iCre Ripk1 fl/fl Ripk3 -/-

100

3 105

0

IL-1β

J

Ripk1fl/+Ripk3-/-

4 105

**

Control TRAIL

0

IFN-γ

I

Vav-iCre Ripk1fl/+Ripk3-/Vav-iCre Ripk1fl/flRipk3-/-

0.5

0

LIX

1.0

0.0

20

6 107

50


Spleen

0

8 107

Vav-iCre Ripk1fl/flRipk3-/-

6000

4000

***

***

1 108

Vav-iCre Ripk1fl/fl

G

Vav-iCre Ripk1fl/flRipk3-/-

***

Percent survival

F

(pg/mL)

E

Vav-iCre Ripk1fl/f l Ripk3-/-

***

40

0

0

0

Vav-iCre Ripk1f l / + Ripk3-/-

D

Vav-iCre Ripk1 fl/+Ripk3 -/Vav-iCre Ripk1 fl/fl Ripk3 -/-

CELL BIOLOGY


B 2x10

Hematocrit (%)

Vav-iCre Ripk1fl/fl (CD45.2)


E 14.5 Fetal Liver

Cellularity

IR 1100 rads

Vav-iCre Ripk1fl/+ (CD45.2)

(pg/mL)

A

p=0.008

80 60 40 20 0 0

20

Days

40

60

Fig. 5. RIPK3 deficiency partially rescues the lethal bone marrow failure of Vav-iCre Ripk1fl/fl mice. (A and B) Macroscopic images (A) and total body weight (B) of indicated genotypes at age 5 wk. (C) Blood cell counts and hematocrits of diseased Vav-iCre Ripk1fl/fl mice (n = 14) compared with age-matched VaviCre Ripk1fl/fl Ripk3 −/− mice (n = 5) and littermate controls Vav-iCre Ripk1fl/+ (n = 11) and Vav-iCre Ripk1fl/+Ripk3 −/− (n = 3). (D) Bone marrow cellularity for Vav-iCre Ripk1fl/+ (n = 11), Vav-iCre Ripk1fl/fl (n = 14), Vav-iCre Ripk1fl/+ Ripk3 −/− (n = 3), and Vav-iCre Ripk1fl/fl Ripk3 −/− (n = 5) mice. (E) Representative images of hematopoietic organs stained with H&E. (F and G) Graphs showing serum chemokine (F) and cytokine (G) levels for Vav-iCre Ripk1fl/fl (n = 10) and Vav-iCre Ripk1fl/fl Ripk3 −/−(n = 5) mice. (H) Colony-forming assay of bone marrow from 5-wk-old Vav-iCre Ripk1fl/+ Ripk3−/− (n = 3) and Vav-iCre Ripk1fl/fl Ripk3−/− (n = 5) mice cultured in the presence of vehicle, TRAIL, IFN-γ, IFN-α, or TNF-α. Total colony number was normalized to vehicle-treated bone marrow. (I) Absolute numbers of HSPCs were determined by flow cytometry in 5-wk-old Vav-iCre Ripk1fl/+ Ripk3−/−(n = 3) and Vav-iCre Ripk1fl/fl Ripk3−/− (n = 5) mice. (J) Kaplan–Meier survival curve of lethally irradiated recipients transplanted with Vav-iCre Ripk1fl/+ Ripk3−/− (n = 5) or Vav-iCre Ripk1fl/fl Ripk3−/− (n = 6) bone marrow cells. Error bars represent SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (Student’s t test in B, C, H, and I; Mann–Whitney test in F and G).

Roderick et al.

PNAS | October 7, 2014 | vol. 111 | no. 40 | 14439

Vav-iCre Ripk1fl/fl Ripk3−/− bone marrow cells died (Fig. 5J). Although RIPK3 deficiency reduced inflammation and prevented bone marrow failure in Vav-iCre Ripk1fl/fl mice, hematopoietic precursors lacking both RIPK1 and RIPK3 lacked longterm multilineage repopulating activity. Based on our in vitro colony data, we speculate that Vav-iCre Ripk1fl/fl Ripk3−/− HSPCs remain sensitive to TNF-induced apoptosis and this sensitivity compromises HSPC function in transplanted mice. Discussion We demonstrate that hematopoietic RIPK1 deficiency results in HSPC loss and subsequent pancytopenia, anemia, and bone marrow failure. An absence of RIPK3 reduces the inflammation and hematopoietic cell death, and consequently normal numbers of erythroid, lymphoid, and myeloid cells are observed in Vav-iCre Ripk1fl/flRipk3−/− mice. Collectively, these data indicate that RIPK1-deficient hematopoietic cells undergo RIPK3-mediated necroptosis, showing that RIPK1 is not absolutely required for necroptotic death in vivo. These findings are supported by transplants with Ripk1−/− Ripk3−/− progenitors (26) and in vitro studies demonstrating that necroptosis can be RIPK1 independent (34–36). Importantly, we provide evidence that pancytopenia and bone marrow failure reflect cell autonomous effect(s), because mice reconstituted with Vav-iCre Ripk1fl/fl fetal liver cells recapitulate the HSPC and hematopoietic lineage loss observed in Vav-iCre Ripk1fl/fl mice. These data demonstrate that RIPK1 loss in hematopoietic cells results in RIPK3 activation and induction of necroptosis, revealing clear anti-inflammatory roles for RIPK1 in the hematopoietic lineage. How RIPK1 regulates RIPK3 activity is unclear; however, steady-state hematopoiesis appears normal in kinase-inactive Ripk1D138N/D138N mice (37), suggesting that the kinase activity of RIPK1 is not required to regulate RIPK3 activity. Moreover, genetic rescue appears specific to the hematopoietic and skin lineages (26, 38), because RIPK3 deficiency has marginal effects on the overall survival of RIPK1-deficient mice (24–26). In both the inducible and hematopoietic lineage knockout models, the HSPC population appears particularly dependent on RIPK1 for survival. Twenty-four hours after the final tamoxifen injection, we observe an 8- to 10-fold decrease in HSPCs with more modest reductions in lineage-restricted cells (Fig. 1D and Fig. S1 I–K). Consistently, the absolute numbers of HSPCs are reduced in 2-wk-old Vav-iCre Ripk1fl/fl mice and become nearly undetectable by age 5 wk. Our data suggest that hematopoietic RIPK1 deficiency is sufficient to stimulate necroptosis and cytokine/chemokine production and/or release; however, these mice develop only mild tissue inflammation before succumbing to hematopoietic failure. Therefore, it remains unclear from these studies whether hematopoietic RIPK1 deficiency is sufficient to trigger systemic inflammation. What triggers hematopoietic necroptosis in Vav-iCre Ripk1fl/fl mice also is unknown. Hematopoietic cell death may reflect unregulated RIPK3 activity and be independent of receptor signaling (as suggested in ref. 39). Alternatively, necroptosis may be triggered by infection, stress, and/or injury or by the transition from the fetal liver to the bone marrow microenvironment. It also is possible that hematopoietic RIPK1 deficiency stimulates the degradation of prosurvival proteins cIAP1/2 and TRAF2, thereby making cell death more likely. In addition to mediating survival, the IAPs have been shown to regulate TNF production; thus RIPK1 deficiency may promote cell death and induce inflammation via effects on cIAP1/2 protein stability (40–42). Increased chemokine and cytokine production also are features associated with the rapid HSPC death observed in the RosaCreERT2 Ripk1fl/fl model (Fig. S2). Hence we propose that hematopoietic RIPK1 deficiency results in RIPK3-mediated necroptosis, DAMP release, and inflammation. Consistent with this model, RIPK3 deficiency significantly reduces serum cytokine and chemokine levels and inhibits the hematopoietic cell death observed in Vav-iCre Ripk1fl/fl mice (Fig. 5 D–G and Fig. S6). 14440 | www.pnas.org/cgi/doi/10.1073/pnas.1409389111

In addition to TNF, RIPK1-deficient MEF are sensitive to necroptosis induced by type I or type II IFN treatment (24, 25, 43). We find that TNF-α, IFN-α, and IFN-γ, but not TRAIL, reduce the hematopoietic colonies derived from RIPK1-deficient progenitors (Fig. 3D). TNF-α and IFN-γ levels are up-regulated in patients with bone marrow failure (44–46) and in both mouse models of RIPK1 deficiency, suggesting that TNF-α and IFN-γ may trigger apoptosis and/or RIPK3-dependent necroptosis in vivo. Our in vitro studies reveal that Vav-iCre Ripk1fl/fl Ripk3−/− progenitors remain sensitive to TNF-induced apoptosis but appear to be protected from IFN-mediated necroptosis (Fig. 5H). In contrast to the lineage-restricted cells, which are rescued in Vav-iCre Ripk1fl/fl Ripk3−/− mice, the HSPCs remain significantly reduced, suggesting that the RIPK1- and RIPK3-deficient progenitors undergo apoptosis. In fact, our transplantation studies demonstrate that these progenitors are unable to reconstitute hematopoiesis fully in irradiated mice (Fig. 5J). Therefore, RIPK3 deficiency reduces inflammation but rescues only short-term hematopoiesis in mice reconstituted with RIPK1-deficient HSPCs. During transplantation the Vav-iCre Ripk1fl/fl Ripk3−/− bone marrow cells likely encounter high levels of TNF-α and other cytokines as well as irradiation-induced changes in the bone marrow microenvironment (47, 48), which may trigger apoptosis. Consistent with this hypothesis, Rickard and colleagues (26) demonstrate that Ripk1−/−Ripk3−/−Casp8−/− HSPCs exhibit wild-type hematopoietic repopulating capability. TNF-α and IFN-γ overproduction are features associated with human bone marrow failure syndromes such as Fanconi anemia (FA) and aplastic anemia. FA patients exhibit a profound defect in HSPC that is present before the onset of clinical bone marrow failure (49). In addition to interstrand cross-linking agents, FA patient cells also are hypersensitive to IFN-γ– and TNFinduced cell death (50), raising the intriguing possibility that the progressive HSPC elimination observed in these patients reflects, in part, TNF- and/or IFN-γ–induced RIPK-dependent necroptosis. It has been speculated that FA patients may benefit from treatments aimed at inhibiting proinflammatory cytokine production. Thus, our findings may have translational implications for these patients as highly selective RIPK1 and RIPK3 inhibitors are being developed. Materials and Methods Mice. Ripk1 conditional mice (Ripk1fl/fl) (38) were crossed with Ripk3−/− mice (a generous gift from Vishva Dixit, Genentech, San Francisco), Vav-iCre (28) or ROSA26-CreERT2 (27) mice (obtained from Jackson Laboratory). When mice became moribund, they were weighed and killed humanely. Complete blood counts and hematocrits were performed on a Hemavet 950FS analyzer (Drew Scientific). For bone marrow transplantation studies, recipient C57BL/6 CD45.1 mice received 11 Gy of total body irradiation in a split dose (550 rads) with a 4-h rest between doses using a Cesium-137 irradiator. Irradiated recipients were reconstituted by i.v. injection of 2 × 106 E14.5 fetal liver cells or bone marrow cells. Recipients were maintained on medicated water and monitored daily for signs of failed engraftment. For Vav-iCre Ripk1fl/fl transplants mice were killed at 2 and 8 wk after reconstitution. All animal procedures used in this study were approved by The University of Massachusetts Medical School Institutional Animal Care and Use Committee. Histology. Tissues were fixed in 10% formalin (Fisher Scientific), and bone marrow was decalcified in Cal-Rite (Richard Allen Scientific) for 48 h. Samples were stained with H&E or with a cleaved caspase-3 antibody (Cell Signaling) at a 1:200 dilution. Images were taken at 10–20× magnification on an Olympus BX41 microscope using an Evolution MP 5.0 Mega-Pixel Camera (MediaCybernetics) and QCapture Pro software (QImaging). Colony-Forming Assay. Bone marrow cells were seeded in MethoCult medium M3434 (STEMCELL Technologies), and total colony number was determined following the manufacturer’s protocol. Ligands were added at the time of plating at the following concentrations: mTRAIL, 200 ng/mL (Pepro Tech); mTNF-α, 10 ng/mL (RandD); mIFN-α 100 units/mL (PBL Interferon Source); mIFN-γ, 10 ng/mL (Pepro Tech).

Roderick et al.

Statistical Measures. Statistical analyses were performed using GraphPad Prism software, version 6.0. Kaplan–Meier survival curves were analyzed using a log rank test with a 95% confidence interval. A two-sided P < 0.05 was considered statistically significant for Student’s t tests and nonparametric Mann–Whitney tests.

Serum Cytokines. Serum cytokines were measured using a 12-plex protein/ peptide multiplex analysis (Luminex Technology) conducted by the National Mouse Metabolic Phenotyping Center at the University of Massachusetts Medical School. Chemokines and cytokines that were below the level of detection were assigned the value of zero.

ACKNOWLEDGMENTS. This work was supported by National Institutes of Health/National Institute of Allergy and Infectious Diseases Grant AI075118 (to M.A.K.). M.P. received funding from European Research Council Grant 2012-ADG_20120314, Deutsche Forschungsgemeinschaft Grants SFB670, SFB829, and SPP1656, European Commission Seventh Framework Program for Research and Technological Development Grants 223404 (Masterswitch) and 223151 (InflaCare), and Deutsche Krebshilfe Grant 110302. J.E.R. was supported by Postdoctoral Fellowship 125087-PF-13-247-01-LIB from the American Cancer Society.

1. Chen ZJ (2012) Ubiquitination in signaling to and activation of IKK. Immunol Rev 246(1):95–106. 2. Vanden Berghe T, Linkermann A, Jouan-Lanhouet S, Walczak H, Vandenabeele P (2014) Regulated necrosis: the expanding network of non-apoptotic cell death pathways. Nat Rev Mol Cell Biol 15(2):135–147. 3. Lee TH, Shank J, Cusson N, Kelliher MA (2004) The kinase activity of Rip1 is not required for tumor necrosis factor-alpha-induced IkappaB kinase or p38 MAP kinase activation or for the ubiquitination of Rip1 by Traf2. J Biol Chem 279(32):33185–33191. 4. Ikeda F, et al. (2011) SHARPIN forms a linear ubiquitin ligase complex regulating NF-κB activity and apoptosis. Nature 471(7340):637–641. 5. Wertz IE, et al. (2004) De-ubiquitination and ubiquitin ligase domains of A20 downregulate NF-kappaB signalling. Nature 430(7000):694–699. 6. Gerlach B, et al. (2011) Linear ubiquitination prevents inflammation and regulates immune signalling. Nature 471(7340):591–596. 7. Ea CK, Deng L, Xia ZP, Pineda G, Chen ZJ (2006) Activation of IKK by TNFalpha requires site-specific ubiquitination of RIP1 and polyubiquitin binding by NEMO. Mol Cell 22(2):245–257. 8. Wu CJ, Conze DB, Li T, Srinivasula SM, Ashwell JD (2006) Sensing of Lys 63-linked polyubiquitination by NEMO is a key event in NF-kappaB activation [corrected]. Nat Cell Biol 8(4):398–406. 9. Chan FK, et al. (2003) A role for tumor necrosis factor receptor-2 and receptorinteracting protein in programmed necrosis and antiviral responses. J Biol Chem 278(51):51613–51621. 10. Lin Y, Devin A, Rodriguez Y, Liu ZG (1999) Cleavage of the death domain kinase RIP by caspase-8 prompts TNF-induced apoptosis. Genes Dev 13(19):2514–2526. 11. O’Donnell MA, et al. (2011) Caspase 8 inhibits programmed necrosis by processing CYLD. Nat Cell Biol 13(12):1437–1442. 12. Holler N, et al. (2000) Fas triggers an alternative, caspase-8-independent cell death pathway using the kinase RIP as effector molecule. Nat Immunol 1(6):489–495. 13. Cho YS, et al. (2009) Phosphorylation-driven assembly of the RIP1-RIP3 complex regulates programmed necrosis and virus-induced inflammation. Cell 137(6): 1112–1123. 14. He S, Liang Y, Shao F, Wang X (2011) Toll-like receptors activate programmed necrosis in macrophages through a receptor-interacting kinase-3-mediated pathway. Proc Natl Acad Sci USA 108(50):20054–20059. 15. He S, et al. (2009) Receptor interacting protein kinase-3 determines cellular necrotic response to TNF-alpha. Cell 137(6):1100–1111. 16. Kaiser WJ, et al. (2013) Toll-like receptor 3-mediated necrosis via TRIF, RIP3, and MLKL. J Biol Chem 288(43):31268–31279. 17. Meylan E, et al. (2004) RIP1 is an essential mediator of Toll-like receptor 3-induced NF-kappa B activation. Nat Immunol 5(5):503–507. 18. Degterev A, et al. (2008) Identification of RIP1 kinase as a specific cellular target of necrostatins. Nat Chem Biol 4(5):313–321. 19. Kaiser WJ, et al. (2011) RIP3 mediates the embryonic lethality of caspase-8-deficient mice. Nature 471(7338):368–372. 20. Oberst A, et al. (2011) Catalytic activity of the caspase-8-FLIP(L) complex inhibits RIPK3-dependent necrosis. Nature 471(7338):363–367. 21. Zhang H, et al. (2011) Functional complementation between FADD and RIP1 in embryos and lymphocytes. Nature 471(7338):373–376. 22. Kelliher MA, et al. (1998) The death domain kinase RIP mediates the TNF-induced NF-kappaB signal. Immunity 8(3):297–303. 23. Cusson N, Oikemus S, Kilpatrick ED, Cunningham L, Kelliher M (2002) The death domain kinase RIP protects thymocytes from tumor necrosis factor receptor type 2-induced cell death. J Exp Med 196(1):15–26. 24. Dillon CP, et al. (2014) RIPK1 blocks early postnatal lethality mediated by caspase-8 and RIPK3. Cell 157(5):1189–1202. 25. Kaiser WJ, et al. (2014) RIP1 suppresses innate immune necrotic as well as apoptotic cell death during mammalian parturition. Proc Natl Acad Sci USA 111(21):7753–7758. 26. Rickard JA, et al. (2014) RIPK1 regulates RIPK3-MLKL-driven systemic inflammation and emergency hematopoiesis. Cell 157(5):1175–1188.

27. Ventura A, et al. (2007) Restoration of p53 function leads to tumour regression in vivo. Nature 445(7128):661–665. 28. de Boer J, et al. (2003) Transgenic mice with hematopoietic and lymphoid specific expression of Cre. Eur J Immunol 33(2):314–325. 29. Feng X, et al. (2011) Cytokine signature profiles in acquired aplastic anemia and myelodysplastic syndromes. Haematologica 96(4):602–606. 30. Schnyder-Candrian S, Strieter RM, Kunkel SL, Walz A (1995) Interferon-alpha and interferon-gamma down-regulate the production of interleukin-8 and ENA-78 in human monocytes. J Leukoc Biol 57(6):929–935. 31. Justo GA, Bitencourt MA, Pasquini R, Castelo-Branco MT, Rumjanek VM (2013) Increased IL10 plasmatic levels in Fanconi anemia patients. Cytokine 64(2):486–489. 32. Lin Y, et al. (2004) Tumor necrosis factor-induced nonapoptotic cell death requires receptor-interacting protein-mediated cellular reactive oxygen species accumulation. J Biol Chem 279(11):10822–10828. 33. Ma Y, Temkin V, Liu H, Pope RM (2005) NF-kappaB protects macrophages from lipopolysaccharide-induced cell death: the role of caspase 8 and receptor-interacting protein. J Biol Chem 280(51):41827–41834. 34. Feoktistova M, et al. (2011) cIAPs block Ripoptosome formation, a RIP1/caspase-8 containing intracellular cell death complex differentially regulated by cFLIP isoforms. Mol Cell 43(3):449–463. 35. Moujalled DM, et al. (2013) TNF can activate RIPK3 and cause programmed necrosis in the absence of RIPK1. Cell Death Dis 4:e465. 36. Upton JW, Kaiser WJ, Mocarski ES (2010) Virus inhibition of RIP3-dependent necrosis. Cell Host Microbe 7(4):302–313. 37. Polykratis A, et al. (2014) Cutting edge: RIPK1 kinase inactive mice are viable and protected from tnf-induced necroptosis in vivo. J Immunol 193(4):1539–1543. 38. Dannappel M, et al. (2014) RIPK1 maintains epithelial homeostasis by inhibiting apoptosis and necroptosis. Nature, 10.1038/nature13608. 39. Orozco S, et al. (2014) RIPK1 both positively and negatively regulates RIPK3 oligomerization and necroptosis. Cell Death Differ. 40. Gentle IE, et al. (2011) In TNF-stimulated cells, RIPK1 promotes cell survival by stabilizing TRAF2 and cIAP1, which limits induction of non-canonical NF-kappaB and activation of caspase-8. J Biol Chem 286(15):13282–13291. 41. Kearney CJ, et al. (2013) Inhibitor of apoptosis proteins (IAPs) and their antagonists regulate spontaneous and tumor necrosis factor (TNF)-induced proinflammatory cytokine and chemokine production. J Biol Chem 288(7):4878–4890. 42. Kim JY, et al. (2011) TNFα induced noncanonical NF-κB activation is attenuated by RIP1 through stabilization of TRAF2. J Cell Sci 124(Pt 4):647–656. 43. Thapa RJ, et al. (2013) Interferon-induced RIP1/RIP3-mediated necrosis requires PKR and is licensed by FADD and caspases. Proc Natl Acad Sci USA 110(33):E3109–E3118. 44. Dufour C, et al. (2003) TNF-alpha and IFN-gamma are overexpressed in the bone marrow of Fanconi anemia patients and TNF-alpha suppresses erythropoiesis in vitro. Blood 102(6):2053–2059. 45. Milsom MD, et al. (2009) Ectopic HOXB4 overcomes the inhibitory effect of tumor necrosis factor-alpha on Fanconi anemia hematopoietic stem and progenitor cells. Blood 113(21):5111–5120. 46. Young NS, Scheinberg P, Calado RT (2008) Aplastic anemia. Curr Opin Hematol 15(3): 162–168. 47. Pearl-Yafe M, et al. (2010) Tumor necrosis factor receptors support murine hematopoietic progenitor function in the early stages of engraftment. Stem Cells 28(7): 1270–1280. 48. Weill D, Gay F, Tovey MG, Chouaib S (1996) Induction of tumor necrosis factor alpha expression in human T lymphocytes following ionizing gamma irradiation. J Interferon Cytokine Res 16(5):395–402. 49. Ceccaldi R, et al. (2012) Bone marrow failure in Fanconi anemia is triggered by an exacerbated p53/p21 DNA damage response that impairs hematopoietic stem and progenitor cells. Cell Stem Cell 11(1):36–49. 50. Pang Q, Andreassen PR (2009) Fanconi anemia proteins and endogenous stresses. Mutat Res 668(1-2):42–53.

Roderick et al.

PNAS | October 7, 2014 | vol. 111 | no. 40 | 14441

CELL BIOLOGY

Flow Cytometry. Single-cell suspensions were stained with cell-surface antibodies for myeloid (Gr-1 and CD11B) and lymphoid (CD3, CD4, CD8, B220) markers. For LSK analysis, bone marrow cells were stained with a biotin lineage mixture, Sca-1, c-Kit, CD34, and Flk2. To distinguish between donor and host hematopoietic cells in the transplant studies, an anti-CD45.2 antibody was added to the LSK staining mixture. All samples were run on a BD LSRII flow cytometer (BD Bioscience) and analyzed using FlowJo software (Tree Star). A complete list of antibodies including clone numbers is given in Table S1.