Clinical strains of HCMV encode 20 putative ORFs within a region of the genome termed ULb′ that are postulated to encode functions related to persistence or immune evasion. We have previously identified ULb′-encoded pUL138 as necessary, but not sufficient, for HCMV latency in CD34+ hematopoietic progenitor cells (HPCs) infected in vitro. pUL138 is encoded on polycistronic transcripts that also encode 3 additional proteins, pUL133, pUL135, and pUL136, collectively comprising the UL133-UL138 locus. This work represents the first characterization of these proteins and identifies a role for this locus in infection. Similar to pUL138, pUL133, pUL135, and pUL136 are integral membrane proteins that partially co-localized with pUL138 in the Golgi during productive infection in fibroblasts. As expected of ULb′ sequences, the UL133-UL138 locus was dispensable for replication in cultured fibroblasts. In CD34+ HPCs, this locus suppressed viral replication in HPCs, an activity attributable to both pUL133 and pUL138. Strikingly, the UL133-UL138 locus was required for efficient replication in endothelial cells. The association of this locus with three context-dependent phenotypes suggests an exciting role for the UL133-UL138 locus in modulating the outcome of viral infection in different contexts of infection. Differential profiles of protein expression from the UL133-UL138 locus correlated with the cell-type dependent phenotypes associated with this locus. We extended our in vitro findings to analyze viral replication and dissemination in a NOD-scid IL2Rγc null-humanized mouse model. The UL133-UL138 NULL virus exhibited an increased capacity for replication and/or dissemination following stem cell mobilization relative to the wild-type virus, suggesting an important role in viral persistence and spread in the host. As pUL133, pUL135, pUL136, and pUL138 are conserved in virus strains infecting higher order primates, but not lower order mammals, the functions encoded likely represent host-specific viral adaptations.

Liver repopulation with bone marrow-derived hepatocytes (BMHs) can cure the genetic liver disease fumarylacetoacetate hydrolase (Fah) deficiency. BMHs emerge from fusion between donor bone marrow-derived cells and host hepatocytes. To use such in vivo cell fusion efficiently for therapy requires knowing the nature of the hematopoietic cells that fuse with hepatocytes. Here we show that the transplantation into Fah(-/-) mice of hematopoietic stem cells (HSCs) from lymphocyte-deficient Rag1(-/-) mice, lineage-committed granulocyte-macrophage progenitors (GMPs) or bone marrow-derived macrophages (BMMs) results in the robust production of BMHs. These results provide direct evidence that committed myelomonocytic cells such as macrophages can produce functional epithelial cells by in vivo fusion. Because stable bone marrow engraftment or HSCs are not required for this process, macrophages or their highly proliferative progenitors provide potential for targeted and well-tolerated cell therapy aimed at organ regeneration.

intestine ͉ tumorigenesis ͉ cell fusion ͉ tissue regeneration

Kinases are dysregulated in most cancer but the frequency of specific kinase mutations is low, indicating a complex etiology in kinase dysregulation. Here we report a strategy to rapidly identify functionally important kinase targets, irrespective of the etiology of kinase pathway dysregulation, ultimately enabling a correlation of patient genetic profiles to clinically effective kinase inhibitors. Our methodology assessed the sensitivity of primary leukemia patient samples to a panel of 66 small-molecule kinase inhibitors over 3 days. Screening of 151 leukemia patient samples revealed a wide diversity of drug sensitivities, with 70% of the clinical specimens exhibiting hypersensitivity to one or more drugs. From this data set, we developed an algorithm to predict kinase pathway dependence based on analysis of inhibitor sensitivity patterns. Applying this algorithm correctly identified pathway dependence in proof-of-principle specimens with known oncogenes, including a rare FLT3 mutation outside regions covered by standard molecular diagnostic tests. Interrogation of all 151 patient specimens with this algorithm identified a diversity of gene targets and signaling pathways that could aid prioritization of deep sequencing data sets, permitting a cumulative analysis to understand kinase pathway dependence within leukemia subsets. In a proof-of-principle case, we showed that in vitro drug sensitivity could predict both a clinical response and the development of drug resistance. Taken together, our results suggested that drug target scores derived from a comprehensive kinase inhibitor panel could predict pathway dependence in cancer cells while simultaneously identifying potential therapeutic options.