L-3-Phosphoserine phosphatase (PSPH) regulates cutaneous squamous cell carcinoma proliferation independent of L-serine biosynthesis

Abstract

Background: L-3-Phosphoserine phosphatase (PSPH) is a highly conserved and widely expressed member of the haloacid dehalogenase superfamily and the rate-limiting enzyme in l-serine biosynthesis. We previously found Psph expression to be uniquely upregulated in a α6β4 integrin transgenic mouse model that is predisposed to epidermal hyperproliferation and squamous cell carcinoma (SCC) formation implicating a role for Psph in epidermal homeostasis.

Objective: We examined the status of PSPH in normal skin epidermis and skin tumors along with its sub-cellular localization in epidermal keratinocytes and its requirement for squamous cell carcinoma (SCC) proliferation.

Methods: First, an immunohistochemical study was performed for PSPH in normal skin and skin cancer specimens and in cultured keratinocytes. Next, biochemical analyses were performed to confirm localization of PSPH and to identify candidate binding proteins. Finally, proliferation and apoptosis studies were performed in human SCC and normal keratinocytes, respectively, transduced with vectors encoding small hairpin RNAs targeting PSPH or overexpressing a phosphatase-deficient PSPH mutant.

Results: PSPH is expressed throughout the proliferative layer of the epidermis and hair follicles in rodent and human skin and is highly induced in SCC. In keratinocytes, PSPH is a cytoplasmic protein that primarily localizes to endosomes and is present primarily as a homodimer. Knock down of PSPH dramatically diminished SCC cell proliferation and cyclin D1 levels in the presence of exogenous of l-serine production suggesting a non-canonical role for PSPH in epithelial carcinogenesis.

Conclusions: Psph is highly induced in proliferative normal keratinocytes and in skin tumors. PSPH appears to be critical for the proliferation of SCC cells; however, this phenomenon may not involve the phosphoserine metabolic pathway.

Fig. 1

PSPH is expressed in the epidermis of murine and human skin. A. Bar graph depicting relative Psph levels quantified by RT-PCR analysis on epidermal RNA samples isolated from acetone- and TPA-treated Wt and α6β4 transgenic mouse skin. Psph levels are normalized by expression of Gapdh. B. Immunoblot validation of monoclonal Psph antibody 5B2 in primary murine, MK-EGFP and MK-Psph-EGFP, and human, huKC1 and huKC2, keratinocyte cultures. Arrow points to an apparent 25 kD protein in all cells (PSPH; lower) or 52 kD fusion protein (PSPH-EGFP; upper) only observed in MK-PSPH-EGFP cells. C. Immunoblot detection of Psph in lysates generated from murine epidermis (Ep), liver (Li), brain (Br), and lung (Lu) tissues. Tubulin (lower panel) levels were assessed to confirm protein loading. D–H. Immunohistochemical detection of Psph in adult murine skin (DAB staining) (D) and immunofluorescent co-labeling of PSPH (F) and Krt14 (G) in human foreskin. No labeling was observed in streptavidin (E) or IgG (H) control-stained sections. Arrows point to strong Psph labeling in the basal layer of the epidermis (D, F). Panels F–G represent the same field of view. I–N. Immunofluorescent co-labeling of Psph and Krt14 (I–K) or Psph and Ki67 (L–N) in primary murine keratinocyte cultures. Panels I–K and L–N represent the same field of view. Abbreviations: epid, epidermis; derm, dermis; sg, sebaceous gland; hf, hair follicle; SA, streptavidin; Wt, wild-type. Scale bar: 50 μm (DE); 20 μm (F–H); 10 μm (I–N).

Fig. 2

Sub-cellular localization of Psph in mouse keratinocytes. A–F. Primary mouse keratinocytes were co-labeled with Psph monoclonal antibody 5B2 and markers for the following sub-cellular structures: endoplasmic reticulum (G), lysosomes (H), nuclear membrane (I), mitochondria (J), trans-Golgi (K) and endosomes (L). M–R. Merged panels for each co-labeled group were generated to confirm the localization of Psph. Scale bar: 10 μm.

Fig. 3

Cytoplasmic Psph functions as a homodimer. A. GFP immunoblot analysis of cytosolic (C) and nuclear (N) lysates from murine keratinocytes (MK) transduced with pBabe-EV, pBabe-EGFP or pBabe-Psph-EGFP retroviral vectors. Arrows point to 52 kD Psph-EGFP fusion protein (upper) or EGFP (lower). Lower blot shows immunodetection of tubulin only in cytosolic lysates of each group. B. Rb (top), PSPH (middle) and tubulin (bottom) immunoblot analysis of nuclear lysates derived from human liver (Li) and lung (Lu) tissue. C. Silver-stained acrylamide gel of anti-HA and anti-FLAG tandem affinity purification in HEK293 cells transfected with pcDNA3 empty vector (1) or pcDNA3-HA-PSPH-FLAG vector (2). Bands marked with arrows were identified by LC-MS/MS to be HA-PSPH-FLAG (upper) and endogenous PSPH (lower). D. Immunoblot analysis showing detection of input levels of HA-PSPH-FLAG expression (arrow) in pcDNA3-HA-PSPH-FLAG transfected (2) but not in pcDNA3 empty vector transfected (1) cells used above in A. E. Psph immunoblot analysis of lysates from primary keratinocytes overexpressing EGFP (1) or Psph-EGFP fusion protein (2) either input (left panel) or immunoprecipitated with EGFP antibodies (right panel). Arrows point to detection of the Psph-EGFP fusion protein (upper) or endogenous Psph (lower). F. Psph immunoblot analysis of whole cell lysates from HEK293 cells subjected to PAGE under non-reducing (NR) and reducing (R) conditions. Abbreviations: MW, molecular weight ladder.

Fig. 4

Effect of PSPH on keratinocyte proliferation. A. PSPH immunoblot analysis in pre-confluent (20–40%) and confluent (80–100%) cultures of huSCC13 (1) and huSCC35 (2) cells. GAPDH immunoblot indicates protein loading. B. Bar graph depicting relative densitometric units of Psph protein levels normalized to GAPDH levels (from A) in pre-confluent and confluent huSCC13 keratinocyte cultures. Error bars represent the average of duplicate experiments. C. Psph immunoblot analysis showing knock down of PSPH in human SCC keratinocytes transduced with a Dox-inducible shRNA targeting PSPH (shPSPH) or a non-silencing control (shNS). Actin (lower panel) levels were assessed to confirm protein loading. D. Bar graph depicting relative densitometric units of Psph protein levels normalized to actin levels (from C) in shNS and shPSPH transduced human huSCC13 and huSCC35 (immunoblot not shown) keratinocyte cultures treated or untreated with Dox. E. Colony forming assay showing Rhodamine A-stained cultures of shNS and shPSPH huSCC13 keratinocytes cultured for 14 days with Dox. F. Bar graph depicting the quantification of colonies > 4 mm in size in shNS and shPSPH transduced huSCC13 cells. Error bars represent the standard deviation of three replicate plates. G. Bar graph depicting the quantification of colonies > 4 mm in size in shNS and shPSPH transduced huSCC35 cells (colony assay plates not shown). Error bars represent the standard deviation of three replicate plates.

Fig. 5

PSPH is induced in murine and human skin tumors. A. Cyclin D1 and cleaved PARP immunoblot analysis in shNS and shPSPH huSCC13 keratinocytes treated (+) or untreated (−) with Dox. Tubulin (lower panel) levels were assessed to confirm protein loading. B. Bar graph depicting enzymatic phosphatase activity on para-nitrophenyl-phosphate by recombinant GST, Psph-GST or Psph^Asn20-GST proteins. Error bars represent the standard deviation of three replicate assays. C. Bar graph depicting the percentage of Annexin V+ cells as determined by FACS analysis in murine keratinocytes transduced with pBabe-EGFP, pBabe-Psph-EGFP or pBabe-Psph^Asn20-EGFP retroviral vectors. Error bars represent the standard deviation of three replicate assays. D. PSPH immunoblot analysis of whole cell lysates from normal murine epidermis (Epid), benign papillomas (Pap) and SCCs. E. PSPH immunoblot analysis of lysates derived from 2 separate cultures (PK24, PK115) of primary normal and 7 separate cultures (huSCC13, huSCC35, huSCC39, huSCC73, huSCC161, huSCC169, huSCC194) of SCC human keratinocytes. Actin (D) or tubulin (E) levels were assessed to indicate protein loading. F–H. Immunolocalization of Psph in murine papilloma (F) and cutaneous SCC (G) tissue sections. Arrow heads point to areas containing atypical cells in G. Panel H represents a secondary antibody-stained control serial section of panel G. Abbreviations: kc, keratinocytes; st, stroma. Scale bar: 50 μm.